Genes and Microbes Impacting the Geochemistry of Arsenic Mobilised Aquifers in Bangladesh and Cambodia

A thesis submitted to The University of Manchester for the degree of

Doctor of Philosophy

in the Faculty of Science and Engineering

2017

Edwin Thadheu Gnanaprakasam

School of Earth and Environmental Sciences

0 Table of Contents

List of Figures …………………………………………………………………………………………………………….. 6 List of Tables ……………………………………………………………………………………………………………… 9 List of Abbreviations………………………………………………………………………………………………….. 11 Thesis Abstract …………………………………………………………………………………………………………. . 13 Declaration ……………………………………………………………………………………………………………….. 14 Copyright Statement …………………………………………………………………………………………………. 15 Acknowledgements ………………………………………………………………………………………………….. 16 About the Author ……………………………………………………………………………………………………… 18 Chapter 1 : Introduction ...... 19 1.1. Project context ...... 19 1.2. Aims and objectives ...... 20 1.3. Thesis structure ...... 22 1.4. Paper status and author contributions ...... 25 Chapter 2 : Literature Review ...... 27 2.1. Arsenic general introduction ...... 27 2.2. Arsenic geochemistry ...... 30 2.2.1. Arsenic minerals and compounds ...... 30 2.2.2. Arsenic speciation ...... 31 2.3. Arsenic bio-geochemical cycle ...... 33 2.4. Geomicrobiology of arsenic mobilisation...... 35 2.4.1. Extracellular interaction ...... 35 2.4.2. Intracellular interaction ...... 36 2.4.2.1. Methylation and demethylation ...... 36 2.4.2.2. As(III) oxidation ...... 38 2.4.2.3. As(V) reduction...... 38 2.5. Role of other elements in the arsenic biogeochemical cycle ...... 40 2.6. Metagenomics of arsenic ...... 41 2.6.1. 16S rRNA gene based community analysis ...... 42 2.6.2. Functional genes analysis (targeted amplification and sequencing) ...... 44 2.6.3. Whole genome sequencing (WGS) based community analysis and gene annotation ...... 50 2.7. Metatranscriptomic analysis of arsenic mRNAs ...... 51 2.8. Metaproteomic analysis of arsenic proteins ...... 52 2.9. Summary ...... 54

1 Chapter 3 : Methodology ...... 55 3.1. Introduction ...... 55 3.2. Sampling methods ...... 55 3.3. General methods involved in the molecular analysis...... 57 3.3.1. DNA extraction ...... 57 3.3.2. RNA extraction ...... 58 3.3.3. Protein extraction ...... 58 3.3.4. Primer design ...... 59 3.3.5. Thermal cycler and polymerase chain reaction (PCR) ...... 59 3.3.6. Gel electrophoresis and documentation ...... 59 3.3.7. DNA and RNA quantification ...... 60 3.3.8. SequalPrep™ normalisation of PCR products ...... 62 3.3.9. Pooling ...... 62 3.3.10. Next generation sequencing ...... 62 3.3.10.1. Roche/454 (Pyrosequencing) ...... 63 3.3.10.2. MiSeq and HiSeq (Illumina) ...... 65 3.3.11. High performance computing (HPC) ...... 66 3.3.12. Bioinformatics tools ...... 66 3.4. Metagenomic methods ...... 69 3.4.1. 16S rRNA gene based community analysis ...... 70 3.4.1.1. Roche 454 pyrosequencing and QIIME pipeline ...... 70 3.4.1.2. MiSeq and the Amplicon analysis pipeline ...... 72 3.4.2. WGS based metagenomic analysis ...... 74 3.5. Metatranscriptomics methods ...... 76 3.6. Proteomic methods ...... 77 3.7. Geochemical methods ...... 78 3.7.1. pH and Eh ...... 78 3.7.2. Diffuse reflectance spectrum: ...... 79 3.7.3. X-Ray fluorescence ...... 79 3.7.4. EXAFS spectra for iron and arsenic ...... 79 3.7.5. Analysis of trace elements ...... 81 3.8. Statistical methods ...... 82 3.8.1. O2-PLS method (Orthogonal Projection to Latent Structure) ...... 82 3.8.2. Spearman’s rank correlation ...... 83

2 Chapter 4 : Paper I - The microbial community structure and arsenic biogeochemistry in two arsenic impacted aquifers in Bangladesh…………………………… ...... 84 4.1. Introduction ...... 85 4.2. Materials and methods ...... 89 4.2.1. Study site ...... 89 4.2.2. Sample collection ...... 90 4.2.3. Sediment chemistry ...... 91 4.2.3.1. X-Ray fluorescence ...... 91 4.2.3.2. EXAFS spectra for iron and arsenic ...... 92 4.2.4. Aqueous chemistry ...... 94 4.2.5. Microbial Ecology ...... 94 4.2.5.1. 16S rRNA Based Bacterial Community Analysis ...... 94 4.2.5.2. Functional Gene Analyses ...... 95 4.2.6. Statistical Analyis: ...... 95 4.3. Results ...... 96 4.3.1. Site F...... 96 4.3.1.1. Sediment chemistry ...... 96 4.3.1.2. Aqueous chemistry ...... 98 4.3.1.3. Microbial ecology ...... 98 4.3.1.3.1. 16S rRNA gene based community analysis ...... 98 4.3.1.3.2. Iron-reducing bacteria ...... 100 4.3.1.3.3. Arsenic-metabolising bacteria ...... 101 4.3.1.3.4. Nitrate/Nitrite-metabolising bacteria ...... 102 4.3.1.3.5. Sulphate-reducing bacteria ...... 102 4.3.1.3.6. PCR confirmation of functional genes ...... 103 4.3.2. Site B ...... 103 4.3.2.1. Sediment chemistry ...... 103 4.3.2.2. Aqueous chemistry ...... 105 4.3.2.3. Microbial ecology ...... 105 4.3.2.3.1. 16S rRNA community analysis ...... 105 4.3.2.3.2. Iron-metabolising bacteria ...... 107 4.3.2.3.3. Arsenic-metabolising bacteria ...... 107 4.3.2.3.4. Nitrate/nitrite-metabolising bacteria ...... 108 4.3.2.3.5. Sulphate reducing bacteria ...... 108 4.3.2.3.6. PCR confirmation of functional genes ...... 109 4.3.2.4. Correlation between the microbial and mineral interphase...... 109 4.4. Discussion ...... 111 Supporting Information …………………………………………………………………………………………… 116

3 Chapter 5 : Paper II - Application of two block latent variable regression analysis (O2-PLS) of microbial and geochemical data to identify potential arsenic cycling microbes in Cambodian aquifers ...... 124 5.1. Introduction ...... 125 5.2. Materials and methods ...... 128 5.2.1. Study site ...... 128 5.2.2. Sample collection ...... 129 5.2.3. Geochemistry ...... 129 5.2.4. Molecular (DNA) analysis ...... 130 5.2.4.1. DNA extraction and library preparation for MiSeq ...... 130 5.2.4.2. 16S rRNA gene analysis pipeline ...... 131 5.2.4.3. The analysis of arrA gene ...... 131 5.2.5. Uni- and multi-variate analyses ...... 132 5.2.5.1. O2-PLS method (Two-way Orthogonal Projection to Latent Structure) ..... 132 5.2.5.2. Spearman’s rank correlations ...... 133 5.3. Results ...... 133 5.3.1. Geochemistry ...... 133 5.3.2. Molecular ecology ...... 135 5.3.3. arrA marker gene analysis ...... 139 5.3.4. Uni- and multi-variate analyses of geochemistry and microbes...... 140 5.3.4.1. Spearman’s rank correlations ...... 140 5.3.4.2. O2-PLS model ...... 141 5.4. Discussion ...... 144 Supporting Information ………………………………………………………………………………………. ... 148 Chapter 6 : Paper III: Metagenomic and metatranscriptomic studies reveal the diversity of microbes and their functional genes that could influence the arsenic redox cycle in aquifers in Bangladesh ...... 156 6.1. Introduction ...... 157 6.2. Materials and Methods ...... 161 6.2.1. Site description ...... 161 6.2.2. Sample collection and processing ...... 162 6.2.3. Metagenomics Methods ...... 163 6.2.3.1. 16S rRNA based community analysis ...... 163 6.2.3.2. Whole genome sequencing (WGS) based metagenome analysis ...... 164 6.2.4. Metatranscriptomics methods ...... 165 6.2.5. Geochemical methods ...... 166 6.2.5.1. Analysis of trace elements in sediment ...... 166 6.2.5.2. Analysis of trace elements in water ...... 167

4 6.3. Results ...... 167 6.3.1. XRF and ICP-MS based As, Fe and S quantification ...... 167 6.3.2. 16S rRNA based microbial community analysis (Site F and Site B) ...... 169 6.3.3. 16S rRNA gene based community analysis of two contrasting aquifers ...... 173 6.3.4. WGS based Metagenome ...... 174 6.3.4.1. Reads, scaffolds and bins ...... 174 6.3.4.2. Prokaryotic communities ...... 175 6.3.4.3. Arsenic Genes ...... 176 6.3.5. Arsenic related metatranscriptomics ...... 178 6.4. Discussion ...... 179 Supporting Material ………………………………………………………………………………………...... 183 Chapter 7 : Conclusions and future work directions ...... 192 7.1. Conclusions ...... 192 7.2. Future work directions ...... 194 List of References ……………………………………………………………………………………..………………196 Appendix I: Supporting Information-2 for Chapter 4 ...... 226 Appendix-II: A laboratory experiment on As(V) respiration and dissimilatory arsenate reductase protein expression by Shewanella sp. ANA -3 …………………………235

5 List of Figures Chapter 1 : Introduction ...... 19 Figure 1-1. The context and approaches of the research presented in the thesis ...... 20 Chapter 2 : Literature Review ...... 27 Figure 2-1. The geographic occurrence of groundwater arsenic pollution………………...... 29 Figure 2-2. The Eh-pH diagram for As at 25oC and one atmosphere pressure with total arsenic 10-5 mol. l-1 and total sulphur 10-3 mol.L-1 ...... 32 Figure 2-3. Arsenic bio-geochemical cycle mediated by microorganisms………………….. .. 33 Figure 2-4. Mechanisms of the microbial transformations of arsenic in the environment. Source: (Lloyd and Oremland, 2006) ...... 39 Figure 2-5. Schematic depiction of the various aspects of metagenomic analysis ...... 42 Figure 2-6. Arsenic gene clade found in different dissimilatory arsenic respiring bacteria ...... 45 Chapter 3 : Methodology ...... 55 Figure 3-1. Summary of instruments, techniques and bioinformatics tools used in this project…………………………………………………………………………………………………………………… 55 Figure 3-2. Field photos on sediment and water sampling for metaomic analysis…… .... 56 Figure 3-3. Bioanalyser-Agilent high density nucleic acid chips………………………………… ... 61 Figure 3-4. The next generation sequencing (NGS) machines used in the project…… ... 63 Figure 3.5. An illustration of a cycle in a pyrosequencing reaction…………………………… ... 64 Figure 3-6. An illustration of the steps involved in sequencing by means of the synthesis method used in Illumina MiSeq and HiSeq ...... 65 Figure 3-7. A diagrammatic depiction of the primer structure used for the pyrosequencing PCR ...... 70 Figure 3-8. The workflow of the 16S rRNA gene based community analysis using the 454 sequencing platform and QIIME downstream analysis……………………………… ...... 71 Figure 3-9. A diagrammatic depiction of the primer structure used for the MiSeq indexing PCR…………………………………………...... 73 Figure 3-10. The workflow of the 16S rRNA gene based community analysis using the MiSeq sequencing platform and amplicon analysis pipeline ...... 74 Figure 3-11. The work flow of the Whole Genome Sequence (WGS) analysis using the HiSeq sequencing platform and metagenomic pipeline ...... 75 Figure 3-12. The schematic workflow of the metatranscriptomic analysis using the HiSeq sequencing platform ...... 76 Figure 3-13. Proteomics workflow used to analyse the expression of Arsenate Reductase (ARR) in Shewanella sp. ANA-3 ...... 78 Figure 3-14. The O2-PLS method used for the multivariate analysis of geochemical and molecular data……………………………………………………………………………………………………… . 83

6 Chapter 4 : Paper I - The microbial community structure and arsenic biogeochemistry in two arsenic impacted aquifers in Bangladesh ………………………………………………………… ...... 84 Figure 4-1. The geolocation of the study sites with the arsenic concentration………… .... 90 Figure 4-2. Chemical and microbial ecology of sediments and water fr0m Site F aquifer ...... 97 Figure 4-3. Chemical and microbial ecology analysis of sediments and water from Site B aquifer……………………………………………………………………………………………………………….. 104 Figure 4-3. Top Spearman rank correlation (p=0.01) plots relating microbial communities to mineral species ...... 110 Figure 4-S1-1. PCR based confirmation of functional genes………………………………………… . 117 Figure 4-S1-3. Top 9 correlation of bacteria with arsenic at site F…………………………….. ... 121 Figure 4-S1-3. Top 9 correlation of bacteria with arsenic at site F…………………………….. ... 121 Figure 4-S1-4. Top 9 correlation of bacteria with Fe and As at site B………………………….. 122 Figure 4-S1-5. Top 9 correlation of bacteria with arsenic at site B…………………………….. . 122 Figure 4-S1-6. Top 9 correlation of bacteria with Fe and As at site F and B……………… .. 123 Figure 4-S1-7. Top 9 correlation of bacteria with Arsenic at site F and site B…………… .. 123 Chapter 5 : Paper II - Application of two block latent variable regression analysis (O2-PLS) of microbial and geochemical data to identify potential arsenic cycling microbes in Cambodian aquifers ...... 124 Figure 5-1. Sampling transects of T-Sand (along the Bassac river) (Sample sites LR-01 through LR-09) and T-Clay (along the Mekong river) (Sample sites LR-10 through LR-14) ...... 128 Figure 5-2. A schematic diagram of the orthogonal projection to latent structure (O2-PLS) model used to analyse the correlation between geochemistry and microbes in the aquifer system ...... 132 Figure 5-3. Box chart representing summary geochemical statistics on a logarithmic scale for the composition of Cambodian groundwater and surface water ...... 134 Figure 5-4. Bar diagrams summarising the number of reads and OTUs obtained in each sample in 16S rRNA community analysis using MiSeq ...... 135 Figure 5-5. Weighted principle coordinate analysis (PCoA) plot of betadiversity among the samples ...... 136 Figure 5-6. Percent abundance of bacteria belonging to different phyla and classes, and the micro molar concentration of arsenic ...... 137 Figure 5-7. A bar diagram representing the total number of prokaryotes correlating to various geochemical parameters through Spearman’s rank correlation ...... 140 Figure 5.8. Joint component scores plot showing the similarity between geochemical and microbial data ...... 141 Figure 5.9. The joint loadings of geochemistry and microbial blocks ...... 142 Figure 5-10. Arsenic prediction using the PLS-R-RFE algorithm ...... 143

7 Figure 5-S-1. Sampling transects of T-Sand (along the Bassac river) and T-Clay (along the Mekong river) ...... 148 Figure 5-S-2. 16S rRNA gene based community analysis workflow…………………………… .. 149 Figure 5-S-3. A diagrammatic depiction of the primer structure used for the MiSeq indexing PCR ...... 150 Figure 5-S-4. Gel documentation of PCR based amplification of arrA gene. The second stage of nested PCR yielded 625 base pair amplified product………………………. .. 150 Chapter 6 : Paper III: Metagenomic and metatranscriptomic studies reveal the diversity of microbes and their functional genes that could influence the arsenic redox cycle in aquifers in Bangladesh ...... 156 Figure 6-1. The sampling sites of Site F, Site B and CW-CAT ...... 161 Figure 6-2. As, Fe and S concentration in sediment and groundwater samples ...... 168 Figure 6-3. Bar diagrams summarising the number of reads and OTUs obtained in each sample in the 16S rRNA community analysis using MiSeq ...... 170 Figure 6.4. Weighted and unweighted principle coordinate analysis (PCoA) plots of betadiversity among the sediment and water samples from Site-F and Site-B ...... 171 Figure 6-5. Relative abundance of bacteria belonging to different phyla and classes…………...... 172 Figure 6-6. Close relatives of the bacterial species and their relative abundance in the two contrasting aquifers CAT and BA14-3, revealed through 16S rRNA gene based community analysis ...... 173 Figure 6-7. Phylogenetic tree for samples BA14-3 and CAT ...... 176 Figure 6-8. Number of bins containing translated genes ...... 177 Figure 6-S-1. The workflow of the 16S rRNA gene based community analysis using the MiSeq sequencing platform and amplicon analysis pipeline...... 183 Figure 6-S-2. Workflow for the whole genome sequencing (WGS) based gene and community analysis using the HiSeq sequencing platform and metagenomic pipeline ...... 184 Figure 6-S-3. Schematic workflow of the metatranscriptomic analysis using the HiSeq sequencing platform ...... 185

8 List of Tables

Chapter 2 : Literature Review ...... 27 Table 2-1. Functional genes involved in the arsenic redox cycle and in arsenic metabolism ...... 46 Table 2-2. Targeted amplification of the arrA gene from the metagenome……………… ... 47 Table 2-3. Targeted amplification of the As(III) oxidase gene from the metagenome. .. 48 Table 2-4. Targeted amplification of the As(V) reductase gene (detox) from the metagenome ...... 49 Table 2-5. Arsenic related transcriptomic works ...... 51 Table 2-6. Respiratory arsenate reductase from different bacteria…………………………… ... 53 Chapter 3 : Methodology ...... 55 Table 3-1. The differences between MiSeq and HiSeq2500………………………………………… .. 66 Table 3-2. The bioinformatics tools used in this research………………………………………….. .. 67 Chapter 4 : Paper I - The microbial community structure and arsenic biogeochemistry in two arsenic impacted aquifers in Bangladesh………………… ...... 84 Table 4-S1-1. Chemical and molecular ecology analysis of sediments and water from the Site F aquifer*……………………………………………………………………………………………………….. .118 Table 4-S1-2. Chemical and molecular ecology analysis of sediments and water from the Site B aquifer*……………………………………………………………………………………………………….. .119 Table 4-S1-3. Correlations between mineralogy and molecular data……………………….. .. 120 Chapter 5 : Paper II - Application of two block latent variable regression analysis (O2-PLS) of microbial and geochemical data to identify potential arsenic cycling microbes in Cambodian aquifers………………………………………………………...... 124 Table 5-S-1. The extract of relevant data of geochemistry previously published by Richards, Magnone, et al., (2017)………………………………………………………………………………… . 151 Table 5-S-2. The extract of relevant data of arsenic speciation previously published by Richards, Magnone, et al., (2017)…………………………………………………………………………… . 152 Table 5-S-3. Significant microbial variables correlating arsenic based on PLS-RFE…. .. 153 Table 5-S-4. Microbes representing arrA genes in sample LR14-30…………………………….. 154 Chapter 6 : Paper III: Metagenomic and metatranscriptomic studies reveal the diversity of microbes and their functional genes that could influence the arsenic redox cycle in aquifers in Bangladesh ...... 156 Table 6-S-1. Summary of the geochemical analysis for Site F used in the analysis….. ... 186 Table 6-S-2. Summary of the geochemical analysis for Site B used in the analysis…. ... 186 Table 6-S-3. Two contrasting aquifers………………………………………………………………………… 186 Table 6-S-4. Total percentage of reads representing different families of prokaryotes from a total of 32 samples………………………………………………………………………………………….. . 187

9 Table 6-S-5. Account of reads, scaffolds and bins for two contrasting aquifer samples……...... 188 Table 6-S-6. Detailed account of genomic bins for BA14-3…………………………………………..189 Table 6-S-7. Detailed account of genomic bins for CAT…………………………………………… .. 190 Table 6-S-8. Arsenic realted transcript level expression matching translated genomic bins………………………………………………………………………………………………………………………………. 191

10 List of Abbreviations AS(III) Arsenite As(V) Arsenate bp Base Pair cDNA Complementary DNA CTAB Cetyl Trimethyl Ammonium Bromide DARB Dissimilatory Arsenate Reducing Bacteria DARM Dissimilatory Arsenate Reducing Microbes DARP Dissimilatory Arsenate Reducing Prokaryotes DIR Dissimilatory Iron Reduction DMA Dimethyl Arsenate DNA Deoxyribose Nucleic Acid dNTPs Deoxy Nucleoside Triphosphate EDTA Ethylene Diamine Tetra Acetic Acid Eh Redox Potential EXAFS Extended X-ray Absorption Fine Structure FPKM Reads per Kilobase of Exon per Million Reads Mapped ICP-AES Inductively Coupled Plasma Atomic Emission Spectroscopy ICP-MS Inductively Coupled Plasma Mass Spectrometry LDS Lithium Dodecyl Sulphate MMA Monomethyl Arsenate mRNA Messenger RNA (Transcript) NCBI National Centre for Biotechnology Information NGS Next Generation Sequencing O2-PLS Orthogonal Projection of Latent Structure OSC Orthogonal Signal Correction OTU Operational Taxonomic Unit PCR Polymerase Chain Reaction pH Potential of Hydrogen ppb Parts per Billion ppm Parts per Million RNA Ribose Nucleic Acid rRNA Ribosomal RNA SDS Sodium Dodecyl Sulphate SNP Single Nucleotide Polymorphism SOB Sulphite Oxidising Bacteria SRB Sulphate Reducing Bacteria Tris-HCl Trisaminomethane Hydrochloride WGS Whole Genome Sequence WHO World Health Organisation XANES X-ray Absorption Near Edge Spectroscopy XRF X-ray Fluorescence Spectroscopy

11 Genes and protein abbreviations Protein Genes abbreviation Protein arrA ArrA Respiratory arsenate reductase α- subunit arrB ArrB Respiratory arsenate reductase b-subunit arrC ArrC Respiratory arsenate reductase membraneous subunit arrD ArrD Respiratory arsenate reductase arrR ArrR Arsenic responsive repressor arrS ArrS Sensor histidine kinase arsC ArsC Arsenate reductase arsN ArsN Acetyl transferase arsR ArsR Arsenic responsive repressor gstB GstB Glutathione S-transferase ACR1 ACR1 Transcriptional regalatory protein ACR2 ACR2 Arsenate reductase; aioA AioA As(III) oxidase, large subunit aioB AioB As(III) oxidase, small subunit aioC AioC c-type cytochrome aioD AioD Molybdenum cofactor aioS/arxS AioS/ArxS Histidine kinase aioX/arxX AioX/ArxX As(III) binding protein aioR/arxR AioR/ArxR Transcriptional regulator arsH ArsH Organoarsenical oxidase arxA ArxA As(III) oxidase arxC ArxC Quinol oxidoreductase subunit arxD ArxD Molybdoenzyme chaperon aroS/aroR AroS/AroR Histidine kinase moeA MoeA Molybdenum cofactor biosynthesis protein arsA ArsA As(III)-pump ATPase arsB ArsB As(III)-pump protein arsD ArsD Arsenical metallochaperone ACR3 ACR3 As(III) permease arsJ ArsJ Organoarsenical efflux permease arsP ArsP Efflux system pgpA PgpA P-glycoprotein-related protein

12 Thesis Abstract Arsenic in aquifers poisons more than 100 million people in Asia alone, as aquifers remain the primary source of water for drinking and farming. Previous studies have suggested a link between the mobilisation of arsenic in aquifers and biochemical processes. As a result of the complex interaction of microbes with arsenic bearing minerals, the relatively immobile arsenate [As(V)] is reduced to labile and more soluble arsenite [As(III)] in aquifers, resulting in elevated concentrations of the metalloid. The numerous microbial communities capable of multiple-metabolic activities colonising these arsenic impacted aquifers mean that the exact mechanism of arsenic mobilisation in aquifers remains poorly understood. To resolve this ambiguity, this study undertakes a combination of metaomic, geochemical, and statistical analyses of 75 aqueous and sediment samples (three sample sets) from 3 transects with arsenic impacted aquifers in Bangladesh and Cambodia. Key geochemical and physical properties including arsenic speciation, iron speciation, mineral and elemental compositions, pH and Eh were recorded using the state-of-the art techniques of XANES, XRF, ICP-MS and other in situ techniques. Next generation sequencing (NGS) platforms such as MiSeq, HiSeq, Nextseq and Pyrosequencing, were used to sequence and analyse DNA and RNA extracted from field samples, allowing characterisation the extent bacterial communities, including any arsenic related genes and transcripts found in these arsenic impacted aquifers. The biogeochemical findings suggest that direct As redox transformations are central to arsenic fate and transport, and that there is a residual reactive pool of both As(V) and Fe(III) in deeper sediments that could be released by microbial respiration in response to hydrologic perturbation, such as increased groundwater pumping that introduces reactive organic carbon to depth. The main findings of this molecular investigation are (i) the most abundant bacterial species belonging to the families of Comamonadaceae, Moraxellaceae, Rhodocyclaceae, Gallionellaceae etc, not known for dissimilatory arsenic reduction, might possess arrA genes and thus have the potential to mobilise arsenic through dissimilatory arsenate reduction; (ii) the bacterial community structure revealed through 16S rRNA gene based sequencing and analysis, resembles the family level community structure revealed through the WGS based community analysis; (iii) although arsenic resistant genes are found in many organisms, they are transcribed only in a few organisms; (iv) the application of O2-PLS analyses may be useful for not only identifying novel organisms associated with key biogeochemical process, but also has clear potential to predict the physical/chemical environment in situ associated with microbial samples via community profiling. In conclusion, the results obtained from this study help establish the identity of microorganisms potentially playing a role in arsenic mobilisation in aquifers, and help decipher the underpinning mechanisms. This deeper level of understanding will in turn help to better target measures that can be applied to arsenic mitigation.

13 Declaration

The author of this thesis declares that no portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning.

14 Copyright Statement i. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes. ii. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with the licensing agreements which the University has from time to time. This page must form part of any such copies made. iii. The ownership of certain Copyright, patents, designs, trademarks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions. iv. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations and in The University’s policy on Presentation of Theses. See

http://documents.manchester.ac.uk/DocuInfo.aspx?DocID=2442 http://www.library.manchester.ac.uk/about/regulations/

15 Acknowledgements

First and foremost, I would like to extend my gratitude from the bottom of my heart to Jon Lloyd, my research supervisor, for walking beside me in my shoes over the course of the entire research, Dave Polya, my co-supervisor, for the exchange of enlightening talks in the realms of philosophy, theology and politics, and Brian Mailloux (Barnard College, Columbia University), my research mentor, a friend and a 500mg caffeine capsule, for guiding me via Skype every Thursday at 3 pm for the past three years. Without the three of you, I cannot imagine completing my research.

Ben Bostick and Lex van Geen (Lamont Doherty Earth Observatory, Columbia University) deserve my special thanks for their collaborative work and guidance in the Bangladesh project, and their help with my manuscripts.

I owe my special gratitude to the research groups headed by Stephen Eyre (Musculoskeletal research group, UoM), Roy Goodacre, (Metabolomic research group, MIB), Mike Wilkins (Geomicrobiology group, Ohio State University) and Jennifer Cavet (Bacterial Pathogenicity group, UoM) for allowing me to use their facilities to complete my projects.

My gratitude mixed with admiration goes to Christopher Boothman, who initiated me into lab works and was always willing to get things sorted out in the different labs I was rattling around. I am in gratitude to the following colleagues: Robert Danczack (Ohio State University) for sharing his knowledge on genome analysis and being ever willing to help me, Abdelkhalik Mosa (Computer Science Department, UoM) for introducing me to Python, Mauro Tutino (Musculoskeletal research group, UoM) for teaching me to use the microbial community analysis pipeline, Yun Xu (MIB, UoM) for helping me with the statistical modelling, Athanasios Rizoulis and Madhura Castelino for guiding me through adopting DNA sequencing for my research and Laura Richards for helping me with the geochemical data for the second paper.

My office mates for over three years, Hannah Roberts (a friendly and systematic leader) and Josh Weatherill (a hardworking brother and friend), deserve an exceptional ‘thank you’ for sharing the overlapping space amicably, bidding the

16 biscuits and teabags generously and showering me with research ideas voluntarily. I express my sincere thanks to Mike Turner for always being available to help me with a smile on his face, Clare Robinson for always inspiring me with her smile and humility.

Rick Eckersley, Raj Shanmugam, Dharani Shanmugam, Sylvia Connolly, Mike Mockridge, Peter Lillis, Mark Dowd, Ken Vance, Domenic Federico and Selin George for their visits, encouraging words, moral support and brotherly care for the past three years.

I am really indebted to Peter Francis, SJ, Dermot Preston, SJ and Maria Christie, SJ, John Murphy, SJ, and other Jesuit friends who stood by me during this time which I reckon to be the most turbulent period of my personal life.

I would like to thank my mother, 7 siblings and their families for their encouragement and moral support in my research and their disturbing silence in my life-changing decisions. With a special note, I would like to thank Dominic Nicholas and Leema Rose (elder brother and sister) for their academic guidance and moral support in my pursuit of knowledge and meaning in life.

David Lillis, who helped me with a smooth transition from a tax free religious life to the wider world of taxes and insurances by allowing me to be his business partner in cinema and properties, deserves my countless thanks. Thank you for your fatherly care, business guidance and above all your tremendous patience.

Finally, I would like to dedicate this thesis to the people of Guyana, South America, who showered me with love and made me feel accepted in every walk of life.

17 About the Author

The author graduated from St. Xavier’s Autonomous College, Palayamkottai, India, with 80% in B.Sc. Zoology with a Chemistry minor and subsequently completed his M.Sc. in Biotechnology at the Loyola Autonomous College, Chennai, India with 73% in 2002. After completing two more ecclesiastical degrees in B.Ph. (Bachelor’s in Philosophy) and B.Th. (Bachelor’s in Theology), he was ordained a Jesuit priest in 2007. He taught biochemistry for undergraduate students at the University of Guyana, Guyana, South America, before he enrolled for his PhD at the University of Manchester, where the research reported in this thesis was undertaken.

18 Chapter 1 : Introduction

1.1. Project context

Arsenic poisoning of groundwaters in Bangladesh alone has been declared, “the largest mass poisoning of a population in history” by the World Health Organisation (WHO) (Argos et al., 2010). More widely, in excess of 100 million people in Asia are subjected to health risks by arsenic contaminated water as these aquifers remain the primary source of water for drinking and farming (Nordstrom, 2002; Smedley & Kinniburgh, 2002; Sharma et al., 2014). The arsenic contamination in water is not anthropogenic but biogeogenic in nature (Islam et al., 2005; Huang, 2014).

Previous studies have suggested a link between microbial processes and the mobilisation of arsenic in aquifers. As a result of the complex interaction of microbes with the arsenic-bearing minerals in the sediment, the relatively immobile arsenate [As(V)] is reduced to labile and more soluble arsenite [As(III)] in aquifers, resulting in elevated concentrations of arsenic in groundwater (Lloyd & Oremland, 2006). Different bacteria and archaea colonising these arsenic impacted aquifers possess genes that regulate the proteins mediating the arsenic redox cycle. Among these, dissimilatory arsenate reducing microbes (DARMs) with the arrA gene could play a vital role in mobilising arsenic in aquifers (Krafft & Macy, 1998; Malasarn et al., 2004; Handley et al., 2009).

Focusing on field sites in Bangladesh and Cambodia, this multidisciplinary study using state-of-the-art geological, geochemical, metagenomic and metatranscriptomic techniques, aims to illustrate the complex interplay between hydrology, mineralogy and microbiology that underpins arsenic release in aquifer sediments. A more complete understanding of these processes will help underpin more accurate predictions of at risk aquifers and inform mitigation procedures, for example linked to interventions that may stimulate or suppress key microbial processes that immobilise or mobilise arsenic into groundwaters. Figure 1-1 outlines the context and approaches adopted in this study.

19

Figure 1-1. The context and approaches of the research presented in the thesis.

1.2. Aims and objectives

The aim of this project is to gain a detailed understanding of the genes and microbes involved in the arsenic redox cycle and their link to the geochemistry of the arsenic impacted aquifers. In this way, the project aims to identify the microbes (based on DNA markers), arsenic related genes (within the metagenome) and their expression (metatranscriptome) in field samples where a comprehensive set of geochemical parameters have been recorded, including arsenic and Fe speciation. Additionally, this work includes a laboratory-based study of Shewanella sp. ANA-3 in order to monitor the expression of the arrA gene in an augmented arsenic environment.

Four main areas of research were formulated and investigated using a range of metagenomic, metatranscriptomic, proteomic, mineralogical, spectroscopic and geochemical techniques, including various next generation sequencing (NGS) platforms; pyrosequencing, MiSeq, HiSeq and Nextseq. The genomic data generated through the NGS techniques were analysed using various genomic tools with advanced algorithms. The correlation between the data sets of microbes and geochemistry were analysed using techniques including a latent variable regression

20 model that eliminates orthogonal structures (mutually irrelevant data) in the data set in order to minimise the errors in the results. Thus, the primary objective of this research was to investigate the mechanism underpinning microbial-mediated arsenic mobilisation in aquifers based on the following studies.

(i) Field-based study of the microbial community structure and genes in arsenic impacted sediment and water samples. If arsenic is mobilised by the microbes in the environment, the arsenic-enriched niche should host a specific community of microbes that are associated with arsenic. Using the 16S rRNA marker gene and SNP (Single Nucleotide Polymorphism) markers (only selected samples) based community analysis, I have identified the dominant microbes found in these arsenic impacted aquifers. The arrA gene was targeted in a selected sample in order to identify whether these genes belonged to the organisms characterised for dissimilatory arsenate reduction. In addition, a binning based whole genome analysis was conducted on two geochemically contrasting aquifers to enumerate the difference in the genes and microbes, specific to each environment.

(ii) Metatranscriptomic analysis of selected water samples. The putative functional genes were identified from DNA samples using their open reading frames and thus the genes could be unexpressed or expressed. To investigate whether the genes of interest were transcribed in the arsenic- rich subsurface environments, the mRNA transcripts of two geochemically contrasting environmental samples were analysed. The mapping of the reference genome obtained from the binning based genomic analysis showed evidence for the expression of genes related to arsenic reduction.

(iii) Investigating the geochemistry of the sediment and water samples. Arsenic is naturally found in various mineral forms and its stability is highly influenced by the geological and geochemical properties of arsenic- bearing minerals. Studying the mineralogical and geochemical properties of these arsenic impacted aquifers is key to understanding the ‘microbes and their metabolisms’, which in turn influence the fate of arsenic in these

21 systems. Using the state-of-the-art techniques including XRF, IPC-MS, ICP-AES and XANES, the geochemical factors were recorded and used as background data set for identifying any correlations with the DNA and RNA analysis.

(iv) Statistical analysis: Two data sets comprising microbial community composition and geochemical measurements with hundreds of mutually excluding variables posed a significant challenge when targeting the correlations between microbial and geochemical factors. To reduce this statistical regression error, O2-PLS method, a two-block (microbes- geochemistry) latent variable regression method with an integral orthogonal signal correction, was employed. Using the Kennard-Stone algorithm in the method, a model was built to predict the concentration of arsenic from the microbial community structure and vice versa.

Note: Protein extraction from the environmental sediments and waters was attempted. Since the yields of extracts were below levels of detection this line of investigation was not pursued, and efforts were focused on metagenomic and metatranscriptomic techniques.

1.3. Thesis structure

The major part of this thesis comprises three research papers covering the four main areas of research detailed above. These papers are preceded by a review of the relevant literature and a detailed description of the methodologies used. The thesis concludes with a summary and a discussion of future work directions.

Chapter 1: Introduction. This chapter introduces the overall structure of the thesis in the format of a chapter based summary.

Chapter 2: Literature Review. This chapter contains information on elemental arsenic, arsenic bearing minerals, and arsenic biogeochemistry, including arsenic-microbe interactions. The main thrust of the literature review covers arsenic metaomics, which addresses environmental genes, mRNA and proteins related to arsenic.

22 Chapter 3: Methods and Principles. This chapter describes the important metaomic methods, instrumentation and bioinformatics tools and moves on to the various geochemical methods and instrumentation. The principle behind each of the methods/instrumentation/bioinformatics tools is described in brief.

Chapter 4: Research Paper I. This chapter contains the paper titled, “The microbial community structure and arsenic biogeochemistry in two arsenic impacted aquifers in Bangladesh.” This paper investigates correlations between 16S rRNA gene based community structure obtained using the pyrosequencing platform and geochemical parameters including arsenic and Fe speciation in two arsenic impacted anoxic aquifers in Bangladesh. A PCR-based search for the functional genes encoding dissimilatory arsenate reductases (arrA) and dissimilatory sulphate reductases (dsr) was carried out on all the samples. These analyses showed that the arsenic-rich sediments were colonized by diverse bacterial communities implicated in both dissimilatory Fe(III)- and As(V)-reduction, while correlation analyses implicated the involvement of phylogenetic groups not normally associated with As mobilisation.

Chapter 5: Research Paper II. This chapter comprises the second paper titled as “Application of two block latent variable regression analysis (O2-PLS) of microbial and geochemical data to identify potential arsenic cycling microbes in Cambodian aquifers.” All 30 water samples obtained from two arsenic impacted transects along the Mekong and Bassoc Rivers in Cambodia were analysed for the microbial community structure based on the 16S rRNA gene using the MiSeq sequencing platform. A pyrosequencing based arrA gene analysis was also conducted to investigate the presence of genes that encode the dissimilatory arsenate reduction. Finally, O2-PLS, a two block (geochemistry-microbes) latent variable regression analysis approach, was used to extract the most significant microbes correlating the arsenic concentration. Of the 47 microbes extracted through this analysis, 63% of the microbes had

23 arsenic related genes. Using the most significant microbes, a statistical model was created to predict the presence of possible microbes that can be found in the arsenic impacted environment. This manuscript is drafted for submission to “Scientific Reports”.

Chapter 6: Research Paper III. This chapter presents the paper titled “Metagenomic and metaproteomic studies reveal the diversity of microbes and their functional genes that influence the arsenic redox cycle in aquifers in Bangladesh.” This paper presents the microbial community structure and the arsenic related functional genes in two arsenic impacted aquifers using the state-of-the-art next generation sequencing platforms; MiSeq and HiSeq. 16S rRNA gene based microbial community analysis, binning based whole genome analysis targeting arsenic related microbes and genes, and the mapping of the transcripts for gene expression analysis are the highlights of this paper. The metagenomic and metatranscriptomic data were interpreted in the light of the geochemical data obtained from analysing the subset of the samples. This comparative study reveals a ubiquitous presence of arsenic metabolising genes and proteins including dissimilatory arsenate reductase gene and its protein (arr and Arr), detox arsenate reductase gene and its protein (ars and Ars) and arsenite oxidase gene and its protein (aio and Aio) in both contrasting aquifers.

Chapter 7: Conclusions and Future Work. This chapter comprises a summary of the work presented in the thesis and a discussion of possible future work directions.

Appendices: Appendix I is the Supporting Information 2 for Chapter 4 (Paper 1). Appendix II is an arsenic gene expression study on Shewanella sp. ANA- 3. This chapter summarises work initiated on As-related gene expression in a model organism capable of dissimilatory arsenate reduction in an anoxic environment. Shewanella sp. ANA-3 was grown in a medium containing 10 mM arsenate and fumarate each as the sole electron acceptors, and the proteome of late log phase cultures was analysed using mass spectrometry to know whether arrA gene was expressed in arsenate medium.

24 1.4. Paper status and author contributions

Chapter 4: Research Paper I. “The microbial community structure and arsenic biogeochemistry in two arsenic impacted aquifers in Bangladesh” submitted to mBio on the 28th of July 2017 and conditionally accepted on the 11th of September 2017. E.T. Gnanaprakasam : Principal author; preparation of the manuscript, 16S rRNA gene based community analysis, functional gene analysis, data analysis. J.R. Lloyd : Project design, project guide, manuscript editing. B. C. Bostick : Assisted with XANES, ICP-MS and XRF. A. van Geen : Manuscript feedback. C. Boothman : DNA extraction, taught library preparation for pyrosequencing and QIIME analysis. I. Choudhury : Assisted with the sample collection. K.M. Ahmed : Assisted with the sample collection and project design. B.J. Mailloux : Principal investigator, sample collection, correlation analysis.

Chapter 5: Research Paper II. “Application of the tw0-block latent variable regression analysis (O2-PLS) of metagenomic and geochemical data to identify potential arsenic cycling microbes in Cambodian aquifers.” E.T. Gnanaprakasam : Principal author; library preparation for MiSeq sequencing, microbial community analysis, arrA gene analysis. Y. Xu : Assisted with the O2-PLS analysis. M. Tutino : Guided me through the bioinformatics pipeline established to analyse 16S rRNA. L.A. Richards : Assisted with the sample collection, and geochemical analysis including ICP-MS, ICP-A B.E. van Dongen : Assisted with the sample collection, D.A. Polya : Project design, editing the manuscript R. Goodacre : Advice and guidance for the application of the statistical model. J.R. Lloyd : Principal Investigator, project advisor, editing the manuscript.

25 Chapter 4: Research Paper III. “Metagenomic and metaproteomic studies reveal the diversity of microbes and functional genes that could influence the arsenic redox cycle in the arsenic impacted aquifers in Bangladesh.” E.T. Gnanaprakasam : Principal author; preparation of the manuscript, 16S rRNA gene based community analysis, binning based metagenomic analysis and data analysis. B.J. Mailloux : Sample collection, geochemical data analysis, and project design. R. Danczak : Metatranscriptomic analysis and taught me how to use the WGS analysis pipeline. B. C. Bostick : Aided with ICP-MS and XRF data analysis. A. van Geen : Assisted with the sample collection and project design. M. Wilkinson : Helping to interpret metagenomic data. J.R. Lloyd : Principal investigator, project guide, editing manuscript.

26 Chapter 2 : Literature Review

2.1. Arsenic general introduction

Although the toxic metalloid arsenic had been discovered by the people of early civilisations, the element was first isolated and characterised in 1250 AD by Albert Magnus, a German alchemist and Dominican Friar. The discovery of the element arsenic attracted further scientific curiosity prompted by its role in ‘alchemy and poisoning’. Arsenic has been recorded in the historic annals for its notoriety as a ‘king of poisons’, used by the power mongers to murder kings in order to inherit their thrones (inheritance powder) (Nriagu, 2001), and by oppressed wives to eliminate their inconvenient husbands (Manna of St Nicholas of Bari) (Frankenberger Jr, 2001). Over the course of the last century, its antimicrobial properties have been recognised and exploited in medicine and related industries. However, in the beginning of the 21st century, it was discovered that it was not only humans but microbes, as well, that extensively manipulate arsenic under different environmental conditions.

Arsenic is placed within the series of elements in group V of the periodic table together with nitrogen, phosphorus, antimony and bismuth. As a metalloid, with its electronic configuration [Ar]3d10 4s2 4p3, it has 5.60 kJ mol-1 as the sum of the first three ionisation enthalpies and displays the properties of metals and non-metals. Arsenic halide and alkyl compounds behave as electron donors due to the presence of lone pairs of electrons in the outer shell. There is little evidence for the existence of cationic species of arsenic under aqueous conditions. Instead, it is endowed with oxyanions (Cotton and Wilkinson, 1988), which makes it a generous electron donor or acceptor. Arsenic can be highly reactive with certain key organic and inorganic compounds resulting in biochemical changes in living beings (Al Lawati et al., 2012).

Humans are highly susceptible to the reactive nature of arsenic, which is often transported through drinking water or consuming food grown in arsenic contaminated land. The biochemical interaction of arsenic in the human body produces a clinical condition called ‘arsenicosis’ (Jomova et al., 2011). This medical condition is caused by the chronic exposure of body cells to arsenic typically from 5 to 20 years, and in its early stages it often manifested by skin lesions, hyperkerotosis

27 and hyperpigmentation (McCarty et al., 2007), and may lead to organ failure and cancer in the later stages. These symptoms are caused by the interference of arsenate [As(V)] or arsenite [As(III)] in various biochemical reactions. By binding to the sulfhydryl groups of pyruvate dehydrogenase in the glycolytic pathway (Bergquist et al., 2009) and other enzymes mediating gluconeogenesis and fatty acid oxidation (Miller et al., 2002), arsenite decreases the synthesis of acetyl coenzyme A, which results in a decreased production of ATP. 3-Arsenopyruvate inhibits the phosphoenolpyruvate mutase that mediates the biosynthesis of C-P bonds in the living cells (Chawla et al., 1995). In general, arsenic interferes with the metabolic pathways in human beings and causes apoptosis (Akay et al., 2004). Subsequently, the carcinogenic, teratogenic, and mutagenic effects of arsenic (Bencko, 1977) make it a notorious poison. Jomova et al., 2011 have comprehensively reviewed the adverse effects of arsenic in human beings.

The World Health Organization (WHO) has defined the provisional maximum tolerable limit of arsenic as 10 µg/L or 10 ppb (Guidelines for Drinking Water Quality: Recommendations, vol. 1., 1993; WHO, 2011). Arsenic pollution may be classified as either geogenic, caused by natural processes, e.g. volcanic and geothermal activities (Webster and Nordstrom, 2003), biogenic, caused by microbial activities, or anthropogenic (Bertin et al., 2011; Slyemi and Bonnefoy, 2012; Drewniak and Sklodowska, 2013; Huang, 2014), caused by agriculture, industrial and mining activities (Garelick et al., 2008). The overall danger of arsenic contamination is summarised in various review articles and books (Cullen and Reimer, 1989; Frankenberger Jr, 2001; Smedley and Kinniburgh, 2002; Garelick et al., 2008; Ravenscroft et al., 2009; Bowell et al., 2014; Vithanage, 2017).

Arsenic pollution has been reported in more than 70 countries on 6 continents (Ravenscroft et al., 2009). Arsenic contaminated aquifers pose the greatest threat to human health mainly because the aquifers remain the only source of drinking water and farming to many in the world. Figure 2-1 depicts the geographic occurrence of groundwater arsenic pollution, where the arsenic concentration exceeds the permitted limit.

28

Figure 2-1. The geographic occurrence of groundwater arsenic pollution. (adapted from (Barringer and Reilly, 2013). The green markings are the areas (states, provinces, countries) where the arsenic concentration exceeds 10 µg/L or 10 ppb. The red highlight is the arsenic hotspot where the most people suffer from arsenic poisoning. The red arrows A and B denote the sampling sites of Bangladesh and Cambodia, respectively.

In the last century, arsenic pollution in drinking water was reported in several South American countries including Argentina, Chile, Peru, Colombia and Mexico. More recently, such pollution has been discovered in many Asian countries as well, notably Taiwan, Bangladesh, Cambodia, India, China, Vietnam, Myanmar and Pakistan (Mukherjee et al., 2006; Bundschuh et al., 2012).

In these Asian countries, the water source for drinking and irrigation is widely obtained through bore-wells drilled into the aquifers. The aquifers are confined or unconfined in nature often containing both largely impermeable aquitards and permeable sediments. These sediments contain arsenic-bearing minerals and the arsenic from these may be released into water by various mechanisms (Stute et al., 2007; Datta et al., 2009; Dhar et al., 2011; Polya and Middleton, 2017). This study places an emphasis on these aquifers, where the release of arsenic takes place.

29 2.2. Arsenic geochemistry

In general, arsenic occurs in mineral forms rather than its elemental form. Arsenic is speciated depending on environmental conditions (physical and chemical) (Cullen and Reimer, 1989). Furthermore, in the ecosystem, when arsenic is cycled through various stages, it exhibits the properties of its compounds, rather than its elemental properties. Therefore it is essential to understand how arsenic occurs in nature and how it is cycled in the environment and this necessarily requires an understanding of the behaviour of bacteria that play an important role in this cycling (Lloyd et al., 2011).

2.2.1. Arsenic minerals and compounds

Though the average overall crustal abundance of arsenic is only 2.5 mg.kg-1 , its average concentration in the upper continental crust is higher, 5.7 mg.kg-1 (Hu and Gao, 2008). The concentration of arsenic in sedimentary rocks is normally in the range between 5 and 10 mg.kg-1 (Webster and Nordstrom, 2003). The lowest concentrations of arsenic have been reported in sand and sandstones. The average concentration of arsenic in sandstones is around 4 mg.kg-1 (Smedley and Kinniburgh, 2002). The sediments in the Gangetic river basin have arsenic concentrations in the range of 1.2 mg.kg-1 to 26 mg.kg-1 (Jin et al., 2003). Arsenic with a concentration as high as 2,100 mg.kg-1 has been recorded in the sediments from the area of hot brines in the Red Sea (National Research Council (US) Committee on Medical and Biological Effects of Environmental Pollutants, 1977).

There are more than 200 minerals (Garelick et al., 2008) that can host arsenic either within their mineral structure or as a an adsorbed component (Cullen and Reimer, 1989; Frankenberger Jr, 2001; Smedley and Kinniburgh, 2002; Garelick et al., 2008; Ravenscroft et al., 2009; Bowell et al., 2014; Bhattacharya et al., 2016). Of these minerals, arsenopyrite (FeAsS), is the most abundant arsenic-containing mineral, while other important minerals include arsenolite (As2O3), cobaltite (CoAsS), proustite (Ag3AsS3), olivenite (Cu2OHAsO4), realgar (AsS), scorodite (FeAsO4.2H2O) and biotite (K(Mg,Fe)3(AlSi3O10)(F,OH)2) (Smith et al., 1998). Most of these are arsenic ore minerals or their secondary products. These arsenic minerals are

30 commonly found in close association with the mineralisation of transition and other metals, such as P, Sb, Au, Ag, Pb, Cd, Mo and W (Smedley and Kinniburgh, 2002).

In rock-forming minerals, arsenic tends to be associated with sulphides or iron minerals. Arsenopyrite is relatively common in many ores, with arsenic concentrations up to 120,000 mg/kg recorded (Pichler et al., 1999). Arsenic is present in the crystal structure of many sulphide minerals as a substitute for sulphur. In an aerobic environment, pyrite, being unstable, oxidises to iron oxides with the release of large amounts of sulphate. Arsenic is not necessarily released by this process because the formation of Fe(II)-bearing minerals can concomitantly immobilise arsenic in the solid-phase (Handley et al., 2013).

Many oxide minerals and hydrous metal oxides show the presence of high arsenic concentrations where arsenic is either part of the minerals or adsorbed on the surface. The adsorption of arsenate to hydrous iron oxides is particularly strong, even at very low arsenic concentrations (Goldberg and Johnston, 2001).

In sediments, metal oxides, specifically those of iron, aluminium and manganese, are the major minerals that to which arsenic (both arsenate and arsenite) binds. Moreover, 50 percent of the iron in freshwater sediments occurs in the form of iron oxides, of which about 20 percent is highly reactive with arsenic. Additionally, in quantity, aluminium oxides are capable of influencing arsenic adsorption. Studies show that below a pH of 7.5, aluminium hydroxides are as effective as iron hydroxides for adsorbing As(V) (Edwards, 1994). The degree of adsorption onto non-iron minerals is generally minor compared to the iron oxides in sediments.

2.2.2. Arsenic speciation

The mobility and toxicity of arsenic depends on its molecular formulation or species. Various factors including the hydrogen ion concentration (pH), oxidation potential (Eh) and temperature (T) control the speciation in an environment. With its electronic configuration [Ar]3d10 4s2 4p3, arsenic mainly occurs in four chemical species including as +5, +3, 0 and -3 (Cullen and Reimer, 1989). However, in an aqueous environment, it is most commonly found in +5 and +3 forms where +5 is generally referred to as the “oxidized” state and +3 as the “reduced state”. The

31 oxidized pentavalent arsenic is termed arsenate [As(V) or (HO)2AsO(OH)] and the reduced trivalent arsenic is termed arsenite [As(III) or (HO)2As(OH)]. Both As(V) and As(III) have inorganic and organic forms. The most common pentavalent (As(V)) inorganic form of arsenic is arsenic acid, whereas the most common organic forms are methylarsonic acid or methylarsonate and dimethylarsinic acid or dimethylarsinate. The most common trivalent (As(III)) inorganic form of arsenic is the arsenous acid, while its most common organic forms are the methylarsonous acid and dimethylarsinous acid (Bentley and Chasteen, 2002).

Furthermore, depending on the pH conditions, the As(V) speciates into 2− different ionic forms including the hydrogen arsenate ion [HAsO4 ]; arsenic acid 4− 3− [H3AsO4]; the dihydrogen arsenate ion [H2AsO ] and the arsenate ion [AsO4 ] (Ferguson and Gavis, 1972). Moreover, the trivalent arsenite has the anionic forms 3− 2− including ortho-arsenite [AsO3 ] a ion of arsenous acid, meta-arsenite[AsO ]n - a 4– 5– polymeric chain anion, pyro-arsenite[As2O5 ] and a polyarsenite [As3O7 ] (Cullen and Reimer, 1989). The Eh and pH dependence of the species predominance fields of common aqueous species are shown in Figure 2-2, which also shows the fields of stability of realgar (AsS and orpiment (As2S3) (Ferguson and Gavis, 1972).

Figure 2-2. The Eh-pH diagram for As at 25oC and one atmosphere pressure with total arsenic 10-5 mol. l-1 and total sulphur 10-3 mol.L-1 . (Source of the figure: (Ferguson and Gavis, 1972)). Solid species are enclosed in parenthesis in the cross-hatched area, which indicates a solubility rate below 10-5.3 mol.L-1.

32 The speciation of arsenic is biologically significant because (i) the arsenic uptake by organisms depends on the valence state of the element; and also (ii) bacterial arsenic interaction depends on the valence state of arsenic (Kruger et al., 2013; Yamamura and Amachi, 2014).

2.3. Arsenic bio-geochemical cycle

Biotic and abiotic processes mediate the global arsenic cycle. The specific processes involved in this cycle include geothermal activities (López et al., 2012), redox reactions (Pederick et al., 2007a), desorption (Herbel and Fendorf, 2006), precipitation (Kirk et al., 2004) and volatilisation (Cullen and Reimer, 1989). Figure 2- 3 depicts the various processes (most of them mediated by microbes) involved in the arsenic cycle.

Figure 2-3. Arsenic bio-geochemical cycle mediated by microorganisms.

In the microbial eco-system, the naturally mineralised As(V) is reduced to As(III) by bacteria by means of dissimilatory metal reduction. The reduced As(III) is highly labile and mobile in aquatic environments. Therefore, the reduced As(III) may be transported to microbes, algae, plants and animals resulting in the formation of

33 arseno-sugars and methylated arsenicals. These accumulated organic arsenicals undergo oxidation to close the bio-geo-arsenic cycle. Though the cycle appears to be driven by biotic intervention, it is in fact influenced by many other physical and chemical parameters including redox. Arsenic cycling can depend on electron shuttles that accelerate redox process in a system. In the reduction process, As(V) gets reduced to As(III) by gaining two electrons, whereas in the case of oxidation, the process is reversed. The maximum redox potential of As(V)/As(III) is +135 mV (Oremland and Stolz, 2003). One of the electron suppliers in this redox process consists of various bacterial species that respire arsenic.

In view of the objectives of this study, the bacterial activities involved in the arsenic cycle in a sedimentary aquifer environment alone will be discussed in this instance. Bacteria become an integral part of the bio-geochemical arsenic cycle in the sedimentary aquifers as they have an active role in the mobilisation of arsenic. Arsenic as As(V) enters into the bacterial cells through the phosphate transporters (Pit) and specific phosphate transporters (Pst) in the bacterial cell wall, which is used for the phosphate uptake under normal circumstances (Harold and Baarda, 1966; Rosenberg et al., 1977; Rosen, 1999). The analysis of arsenic gene islands reveals the functional relationship between arsenate and phosphate (Li et al., 2013). After being transported into the cell, in a methylation cascade, some bacteria can convert As(V) into monomethylarsonic acid (MMAv) or dimethylarsinic acid (DMAv) presumably in order to detoxify the system. Some prokaryotes also produce volatile methylated arsines (Ji and Silver, 1992; Silver and Phung, 2005; Wang et al., 2014).

In an alternative mechanism, As(V) is reduced to As(III) by the arsenate reductase protein (ArsC) (Mukhopadhyay and Rosen, 2002a; Mukhopadhyay et al., 2002). The reduced As(III) is pumped out of the cell by the specific efflux pump, ArsB. Through this ATP dependent mechanism, bacterial cells detoxify the system. But some bacterial cells gain energy through a process called dissimilatory-arsenate- reduction in which As(V) is utilised as an electron acceptor in place of oxygen under anaerobic conditions. In a reduced state, As(III) is discharged into the aqueous environment (Macy et al., 1996; Newman and Beveridge, 1997; Krafft and Macy, 1998; Switzer Blum et al., 1998; Afkar et al., 2003; Silver and Phung, 2005).

34 In closing the cycle, arsenite oxidising bacteria oxidise As(III) to As(V) in an enzymatic process using the arsenite oxidase (Aso presently known as Aio) in a coupled process where oxygen or nitrate is reduced in anaerobic conditions (Rhine et al., 2007).

2.4. Geomicrobiology of arsenic mobilisation

Many mineralogical and geochemical studies have been conducted on various aquifer sediments in the pursuit of understanding the different mechanisms behind arsenic mobilisation (Berg et al., 2001; Zheng et al., 2004; Horneman et al., 2004; Islam et al., 2004; Gault et al., 2005; Rowland et al., 2007; Stute et al., 2007; Héry et al., 2008; Datta et al., 2009; Rizoulis, Lawati, et al., 2014; Sun, Bostick, et al., 2016) Recently, a geomicrobiological approach to arsenic mobilisation has taken the centre stage in arsenic research. This approach studies in detail the intra- and extra-cellular interaction between different arsenic species and various bacterial species.

2.4.1. Extracellular interaction

The surface interaction between bacteria and metals or metallic compounds plays an important role in deciding the fate of the bacteria-metallic interaction (Handley et al., 2009; Hohmann et al., 2011). Various parameters including ionic strength, hydrophobicity and surface energy as well general bacterial growth parameters, such as temperature and hydrogen ion concentration in the medium influence the interaction between bacteria and metals/metallic compounds (Li and Logan, 2004). In this interaction, the bacterial extracellular polymeric substance (EPS matrix) may act as a sieve for cations, anions, apolar compounds and other particles in an aqueous environment. The stickiness of the matrix traps and accumulates the surface particles that come in contact with bacteria. In general, while the polar species of heavy metals accumulate on the bacterial cell wall, hydrophobic compounds like benzene, toluene and xylene get trapped in the matrix (Flemming and Wingender, 2010). In this interaction, both the biosorption of bacteria and the adsorption of chemicals play an equal role in any further biochemical changes.

The recent studies on the biosorption of arsenic species [As(V), As(III), MMAIII] with bacteria have broadened the understanding of the extracellular

35 interactions of bacteria with arsenic (Yan et al., 2010; Prasad et al., 2011; Giri et al., 2013). While certain studies prove that the biofilms increase As(V) absorption (Prieto et al., 2013), other studies support the fact that the sorption of arsenic to bacterial cell surface is in the form of an electrostatic interaction involving hydroxyl, amide and amino groups on the surface of the microorganism. This surface interaction is temperature and pH dependent (Yan et al., 2010; Prasad et al., 2011; Giri et al., 2013). Moreover, microbial attachment to the minerals increases the solubility of As(V) due to competitive adsorption between As(V) and Fe(III)-oxyhydroxide to the bacterial cell wall (Huang et al., 2010).

2.4.2. Intracellular interaction

Once the arsenic enters the bacterial cell, it undergoes various biochemical changes in accordance with the intracellular environment. On the basis of the different biochemical modifications within the cell, arsenic undergoes methylation, demethylation, oxidation and reduction in order to either detoxify the system or to supply energy to the system.

2.4.2.1. Methylation and demethylation

Initially, the studies on the arsenic methylation process have been conducted on different species of fungi rather than on bacteria. Recently the mechanism was studied in various aerobic and anaerobic bacteria including Proteus sp., Corynebacterium sp., Flavobacterium sp., Escherichia coli, Pseudomonas sp., Achromobacter sp., Alcaligenes sp. (Shariatpahani et al., 1983), Pseudomonas putida, Xanthomonas sp., Klebsiella oxytoca (Maeda et al., 1990), Flavobacterium-cytophaga spp. (Turpeinen et al., 2002), Methanobrevibacter smithiia (Meyer et al., 2008), Clostridium collagenovorans, Desulfovibrio gigas, Desulfovibrio vulgaris and Methanobacterium formicicum (Michalke et al., 2000).

These studies in bacteria demonstrate that the methylation process is the same as in fungi (Kuehnelt and Goessler, 2003). In this process, As(V) in the cell is reduced to As(III) before a methyl group is added in an oxidative coupled reaction (Dombrowski et al., 2005) catalysed by a family of the As(III) S-adenosylmethionine methyltransferase enzyme (ArsM) (Qin et al., 2006a). Moreover, while anaerobic

36 bacteria use methylcobalamine as the electron donor in the process, the aerobic bacteria use glutathione and other thiol-containing compounds to produce the methylated arsenic species including monomethyl arsenite (MMAIII), dimethyl arsenate (DMAV) and trimethyl arsine oxide (Kräutler, 1990; Stupperich, 1993). The extracellular studies in the growth medium showed that the monomethylarsonic acid (MMAv) was an intermediate in this methylation process but, due to its low permeability, was not pumped out of the cell in appreciable quantities. For this reason, there is a considerable difference in the concentration of MMAv in the medium compared to DMAIII (Huang and Matzner, 2007). According to Challenger’s pathway, the monomethylarsonous acid (MMAIII) and dimethylarsinous acid (DMAIII) are also intermediates of the arsenic methylation. Of these methylated arsenic species, the gaseous trimethylarsine oxides (TMAv) displayed the greatest mobility (Mukai et al., 1986). The methylated arsenic, due to its lower adsorption affinity, becomes more mobile than its inorganic counterparts (Huang and Matzner, 2007). In lower reducing conditions, the arsenic methylation is augmented and the mobilisation is increased (Frohne et al., 2011) but in higher reduction conditions in sediments, the iron oxides undergo reductive dissolution, with the reduction of As(V) to As(III). The resistance to arsenic is achieved through the process of methylation.

Arsenic demethylation is associated with methylated arsenic (MMA and DMA) in the aqueous phase. Although it is known that methylated arsines in the gaseous phase undergo a rapid photoxidative degradation (Mestrot et al., 2011), the role of bacterial interaction with the gaseous arsines is not known. In the aqueous phase, demethylation occurs under both oxic and anoxic conditions but the process occurs faster in the oxic environment than in the anoxic environment (Huang and Matzner, 2007). In arsenic contaminated soil, the mixed culture of Streptomyces and Burkholderia species reduced the monomethylarsonic acid (MMAv) to monomethylarsonous acid (MMAIII). The results demonstrated that this process unfolds in two steps (Yoshinaga et al., 2011). In the case of Mycobacterium neoaurum, both the monomethylarsonic acid and the monomethylarsonous acid were reduced to a mixture of As(V) and As(III) (Lehr et al., 2003). The process of methylation was

37 postulated to help remediate arsenic pollution while, conversely, the removal of organic moieties increased the toxicity of arsenic.

2.4.2.2. As(III) oxidation

In this mechanism, arsenite-oxidising bacteria oxidise As(III) to As(V) in an enzymatic process catalysed by arsenite oxidase (Aio, formerly known as Aso) in a coupling process where oxygen (or nitrate in anaerobic conditions) is used as the electron acceptor. This process is conducted by a wide range of bacteria including Thermus aquaticus and Thermus thermophiles (Gihring et al., 2001), Alcaligenes faecalis (Anderson et al., 1992), Hydrogenophaga sp. str. NT-14 (Hoven and Santini, 2004), Acidithiobacillus ferrooxidans (Duquesne et al., 2003) and Desulfotomaculum auripigmentum (Newman et al., 1997). Because As(V) is often less mobile and has more affinity to mineral surfaces, oxidation of As(III) has been widely muted as a partial remedy for the As(III) polluted environment (Newman et al., 1997; Smedley and Kinniburgh, 2002).

2.4.2.3. As(V) reduction

The mobilisation of arsenic in the aqueous phase is often attributed to the reduction of As(V). Many bacteria that mediate this process are listed in various reviews (Páez-Espino et al., 2009; Tsai et al., 2009; Zhu et al., 2013, 2017; Huang, 2014; Yamamura and Amachi, 2014; Andres and Bertin, 2016). These bacteria reduce As(V) either through the energy-driven detoxification system (Rosen, 1999; Saltikov et al., 2003; Lloyd and Oremland, 2006) or through the respiratory system, where the bacteria gain energy (Newman and Beveridge, 1997; Macy et al., 2000; Oremland and Stolz, 2003; Saltikov et al., 2005; Malasarn et al., 2008). However, only As(V)-respiring bacteria are implicated in the As mobilisation.

A considerable number of bacteria including Escherichia coli, Staphylococcus aureus and Staphylococcus xylosis known for arsenic resistance rather than arsenic respiration have been studied (Tamaki and Frankenberger Jr, 1992; Cervantes et al., 1994; Ji et al., 1994). In both mechanisms, As(V) is reported to enter the cells through Pit or Pst (phosphate transporters) on the cell wall. In the detoxification system, the reduction of As(V) to As(III) is catalysed by a cytoplasmic enzyme called arsenate

38 reductase (ArsC), to which an electron is supplied by glutathione or ferredoxin. The reduced As(III) is pumped out of the cell through a membrane efflux complex called ArsB or Acr3 (Rosen, 1999).

In the case of arsenic respiration, the arsenic respiring microbes use As(V) as a terminal electron acceptor in the anaerobic respiratory chain. These microbes are categorised under the “Dissimilatory As(V) Reducing Prokaryotes (DARPs)”. A few DARPs studies to date are of Chrysiogenes arsenatis (Krafft and Macy, 1998), Sulfospirillum arsenophilum (Stolz and Oremland, 1999), Sulfospirillum barnesii (Zobrist et al., 2000), Shewanella sp. ANA-3 (Saltikov and Newman, 2003), Bacillus selenitireducens strain SF1 (Yamamura et al., 2005), Desulfitobacterium hafniense strain DCB-2 (Niggemyer et al., 2001), and Geobactor uranireducens Rf4 (Giloteaux et al., 2012), Alkaliphilus ormelandi (Fisher et al., 2008). In most circumstances, the DARPs use acetate or lactate as electron donors. Chrysinogenes arsenatis in particular uses acetate as the electron donor, which eliminates the possibility that ATP is formed through phosphorylation (Krafft and Macy, 1998). These bacteria possess a gene cluster (in some, it is putative) including arrA that regulates the expression of

Figure 2-4. Mechanisms of the microbial transformations of arsenic in the environment. Source: (Lloyd and Oremland, 2006). the respiratory As(V) reductase (Arr) which, in turn, catalyses arsenic [As(V)] reduction in the respiratory chain. As an overview, Lloyd and Oremland, 2006 have

39 pictorially represented the mechanisms of microbial transformations of arsenic in the environment (Figure 2-4). The detailed account of the functional genomics related to As(V) reduction will be discussed later in the sections investigating metagenomics, metatranscriptomics and metaproteomics.

2.5. Role of other elements in the arsenic biogeochemical cycle

In the environment, where various minerals and elements coexist with a diverse microbial population, there is no possibility for the mineral cycle to occur in a closed system (Tamaki and Frankenberger Jr, 1992). The cycling of a particular element is often coupled with the cycling of other elements causing a domino effect. For example, the arsenic biogeochemical cycle is often coupled with the cycling of other elements such as C, Fe, S and N (Al Lawati et al., 2012; Héry et al., 2014; Rizoulis, Lawati, et al., 2014; Hug et al., 2016; Zhang et al., 2017). Thus, the microbial communities colonising these environments are capable of mediating more than one element. This complex system makes it difficult to compartmentalise arsenic biogeochemical cycling from the cycling of other elements (Zahid Hassan et al., 2015).

One of the important cycles associated with arsenic biogeochemical cycling is Fe biogeochemical cycling. The dissimilatory reduction of Fe(III) to Fe(II) can be energetically favourable for specialist anaerobic microorganisms including Geobacter species and can result in either the solubilisation of Fe(II) and/or transformations in the sediment Fe mineralogy (Islam et al., 2005; Herbel and Fendorf, 2006). Dissimilatory iron reducing bacteria (DIRB) can augment the reduction of Fe(III) to Fe (II) in the presence of acetate. In the coupling process, As(V) is eventually reduced to As(III) under reductive dissolution (Islam et al., 2004). Some microorganisms such as Shewanella sp. ANA-3, Sulfurospirillum barnessi, Geobacter lovleyi, Geobacter uraniireducens can reduce both Fe(III) and As(V) (Campbell et al., 2006; Kocar et al.,

2006).

Similarly, microbial sulphate reduction may influence arsenic solubility through the formation of insoluble arsenic sulphides, e.g. As2S3 (Rittle et al., 1995; O’Day et al., 2004; Omoregie et al., 2013). Sulphate reducing bacteria (SRB) can influence the arsenic redox cycle by generating hydrogen sulphide or elemental

40 sulphur from sulphate and the re-precipitation of arsenic as arsenic sulphide. It is an important process in mitigating arsenic mobilisation via the precipitation of poorly soluble arsenic-bearing sulphide phases. In the reverse process, sulphur oxidising bacteria including Gallionella spp. are capable of immobilising arsenic by using free or arsenic bound sulphur as an electron donor to transform As(III) to As(V).

Nitrate reduction is a widespread respiratory process that can also influence the fate of arsenic under anaerobic conditions, via nitrite-driven oxidation of Fe(II) and As(III). In an anoxic environment, As(III) was present where nitrate was depleted, and As(V) was found when the nitrate concentration increased (Senn and Hemond, 2002; Omoregie et al., 2013).

Similarly, organic matter (OM) can influence the redox cycle of arsenic and may be rich in reactive groups including carboxylic, amino, sulfhydryl, hydroxyl, esteric, phenolic and nitroso. Low molecular weight OM including lactate, fumarate, malate, acetate, citrate and glycerol serve as the electron donor to the microbes for the conservation of energy, where As(V) is used as the electron acceptor (Niggemyer et al., 2001; Lear et al., 2007; Handley et al., 2009).

2.6. Metagenomics of arsenic

The term metagenomics is used to describe genetic studies of uncultured microbial communities from environmental samples, using sequencing and bioinformatics tool-based studies (Riesenfeld et al., 2004; Schloss and Handelsman, 2005). The 16S rRNA marker gene based community analysis, marker genes based functional group analysis and whole genome sequencing based gene annotation and community analysis are three major parts of metagenomic studies (Ge et al., 2013). In all cases, the DNA of the microorganisms is extracted straight from the environment samples including sediment, soil and water, or these environmental samples are enriched further in microcosm experiments (Figure 2-5 ). Initially, the 16S rRNA gene based community analysis heavily relied on enrichment- based microcosm experiments (Rowland et al., 2002; Gault et al., 2005; Omoregie et al., 2013; Rizoulis, Wafa M. Al Lawati, et al., 2014; Sun, Chillrud, et al., 2016). With the

41 advancements in ‘omic techniques’ including low cost next generation sequencing (NGS) and bioinformatics with advanced algorithms, most studies currently analyse

Figure 2-5. Schematic depiction of the various aspects of metagenomic analysis. the DNA samples obtained directly from the environment, whereas community analysis and functional gene analysis are done through whole genome sequencing platforms. The main advantage of applying metaomics to direct environmnet samples is that it captures all microbial communities as they are in the environmnet, whereas the enrichment cultures favour the growth of only a few bacterial communities that disproportionately outgrow the rest of the communities that could have significant role in arsenic mobilisation. Various metagenomic tools and techniques used in this study are detailed in Chapter 3: Methods.

2.6.1. 16S rRNA gene based community analysis

Identifying the microbes colonising the arsenic environment is a vital step in understanding the process of arsenic reduction. Since 1994, various studies (summarised in various reviews) (Ahman, 1994; Dowdle et al., 1996; Macy et al., 1996; Malasarn et al., 2004; Yamamura and Amachi, 2014; Andres and Bertin, 2016; Zhu et al., 2017) have identified the microbes and noted the presence of gene islands related to arsenic metabolism and tolerance. In order to identify these bacterial colonies living in arsenic impacted environments, a region- or a few regions- or the full length-

42 of 16S rRNA gene, known for its ‘conserved and hyper-variable regions’, have been sequenced and analysed extensively (Héry et al., 2008; Al Lawati et al., 2012; Rizoulis, Lawati, et al., 2014; Z. Hassan et al., 2015). The 16S rRNA genes were initially retrieved through RFLP (Restriction Fragment Length Polymorphism) based molecular cloning methods. As sequencing technologies developed, enrichment and cloning independent community analysis has been conducted in recent studies.

The 16S rRNA gene based community analysis of various environmental samples from arsenic impacted aquifers (sediment and water samples) has shown the presence of different bacterial and archaeal species related to arsenic cycling. Most of the arsenic metabolising bacteria found in the arsenic-polluted environments belong to different bacterial phyla including α-, - and - Proteobacteria, Actinobacteria, Firmicutes and Cyanobacteria (Saltikov et al., 2003; Hoeft et al., 2004; Oremland et al., 2005; Kulp et al., 2006; Song et al., 2009). These phyla encompass various genera of bacteria including Chrysiogenes spp. (Krafft and Macy, 1998), Sulfurospirillum spp. (Héry et al., 2014), Bacillus spp. (Afkar et al., 2003), Desulfitobacterium spp. (Pérez- Jiménez et al., 2005), Geobacter spp. (Héry et al., 2014), Anaeromyxobacter sp. (Dong et al., 2014) and Shewanella sp. (Drewniak et al., 2015), that are identified as representatives of dissimilatory arsenate reducing bacteria. These representatives are known to possess an arsenic gene island that regulates the metabolism of arsenic.

Though arsenic related microbes are present in the arsenic impacted aquifers, their relative abundance falls below 1% in general (Lear et al., 2007; Yamamura and Amachi, 2014). This could be due to the fact that the microbes mediating carbon, nitrogen, oxygen, phosphorus, iron and sulfur cycles outnumber the microbes involved in arsenic cylcing. So, the low abundance of arsenic respiring microbes are justifiable. But, this begs the question as to whether the known arsenic metabolising bacteria alone are capable of mobilising arsenic in the environment or if other arsenic metabolising bacteria, which are not implicated in the arsenic mobilisation, play a vital role. The community structures in researched arsenic impacted aquifers do not follow a definite pattern (Gremion et al., 2003; Sarkar et al., 2013; Buse et al., 2014; Hug et al., 2014; Mirza et al., 2014). In turn, this poses additional challenges to implicating any specific community in the process of arsenic mobilisation in the

43 aquifers. In general, the microbial families that are revealed in these studies are soil or planktonic communities. The most abundant bacterial communities found in these arsenic studies belong to general soil or water microbes from the bacterial families including Comamonadaceae, Moraxellaceae, Pseudonocardiaceae, Methylophilaceae, Rhodocyclaceae, Bacillaceae, Micrococcaceae, Nitrospiraceae, Gallionellaceae, Xanthomonadaceae, Cyanobacteria, Rhizobiaceae, Clostridaceae and Geobacteraceae. (Bowman et al., 1997; Cummings D.E., 2002; Wauters et al., 2003; Lear et al., 2007; Cai et al., 2009a; Hamamura et al., 2013; Buse et al., 2014; Paul et al., 2015; Sanyal et al., 2016).

2.6.2. Functional genes analysis (targeted amplification and sequencing)

The arsenic gene island possesses an array of genes or putative genes that encode various enzymes involved in the arsenic redox cycle (Andres and Bertin, 2016; Zhu et al., 2017). Various lab-based studies have characterised the genes responsible for various functions. Table 2-1 summarises the different genes involved in the arsenic redox cycle. In various studies, these genes were targeted using specific primers in order to investigate the presence of these genes in the arsenic impacted environments so as to correlate the presence of genes to the arsenic redox cycle (Pérez-Jiménez et al., 2005; Kulp et al., 2006; Lear et al., 2007; Pederick et al., 2007b; Héry et al., 2008; Song et al., 2008; Escudero et al., 2013).

The clade of genes targeted in various studies are the arrAB gene clade. In all arsenic respiring bacteria, these two genes were found next to one another (Andres and Bertin, 2016). In Alkaliphilus oremlandii and Wolinella succinogenes, these genes are accompanied by arrC and arrD. In different bacteria, the genes are arranged differently in the arsenic operons. Figure 2-6 depicts the arrangement of the arrA gene clade in several arsenic respiring bacteria. In the environmental samples, the

44

Figure 2-6. Arsenic gene clade found in different dissimilatory arsenic respiring bacteria. (The genes in red depict arsenic respiring genes (arr) where as purple are arsenic resistant genes (ars). Adapted from (Yamamura and Amachi, 2014; Andres and Bertin, 2016)). arrA gene, which mediates the dissimilatory As(V) reduction has been targeted using specific primers. Although the microbial owners of these arrA genes were not found in the 16S rRNA gene based community analysis, the presence of these genes affirms the possibility of the low abundance of these microbes colonising arsenic-impacted environments. Table 2-2 lists some important studies that amplified the arrA genes in environmental DNA.

45 Table 2-1. Functional genes involved in the arsenic redox cycle and in arsenic metabolism (Adapted from Andres and Bertin, 2016; Zhu et al., 2017) Genes/ PCR Protein Protein/Function Reference Process Primers abbreviation As(V) Reduction (dissimilatory As(V) reduction) AS1F, AS1R ArrA Respiratory arsenate reductase α- subunit; catalyses the reduction of As(V) (Krafft and Macy, 1998; Afkar et al., arrA and AS2F 2003; Saltikov and Newman, 2003; arrB ArrB Respiratory arsenate reductase -subunit; catalyses the reduction of As(V) Kruger et al., 2013) ArrC Respiratory arsenate reductase membraneous subunit; involved in anchoring and (Duval et al., 2008; Van Lis et al., arrC electron transfer 2013) arrD ArrD Respiratory arsenate reductase; chaperon protein involved in the reduction of As(V) (Van Lis et al., 2013) arrR ArrR Arsenic responsive repressor; regulates the expression of the arr operon (Van Lis et al., 2013) arrS ArrS Sensor histidine kinase; regulates the expression of the arr operon (Van Lis et al., 2013) As(V) Reduction (detox reduction) arsC ArsC Arsenate reductase; catalyses the cytoplasmic reduction of As(V) (Mukhopadhyay and Rosen, 2002a) arsN Acy19/M13R ArsN Acetyl transferase; putative As(V) reductase (Chauhan et al., 2008) arsR ArsR Arsenic responsive repressor; regulates the expression of arr operon (Van Lis et al., 2013) gstB GstB Glutathione S-transferase B; catalyses the reduction of As(V) with reduced GSH (Chrysostomou et al., 2015) ACR1 M13F/M13R ACR1 (yeast) Transcriptional regalatory protein; regulation of the ACR gene (Piotr et al., 1997) ACR2 M13F/M13R ACR2 (yeast) Arsenate reductase; catalyses the cytoplasmic reduction of As(V) (Piotr et al., 1997) As(III) Oxidation aioA P8 and/ AioA As(III) oxidase, large subunit; catalyses the oxidation of As (III) (Anderson et al., 1992; Silver and Pc1544 Phung, 2005; Lett et al., 2012; Van aioB AioB As(III) oxidase, small subunit; catalyses the oxidation of As (III) Lis et al., 2013) aioC AioC c-type cytchrome; (Santini et al., 2007; Branco et al., 2009) aioD AioD Molybdenum cofactor; (Branco et al., 2009; Slyemi and Bonnefoy, 2012) aioS/arxS AioS/ArxS Histidine kinase; arsenic signal transduction (Kashyap et al., 2006; Sardiwal et al., 2010) aioX/arxX AioX/ArxX As(III) binding protein; involved in As(III) signalling and regulation (Liu et al., 2012; Zargar et al., aioR/arxR AioR/ArxR Transcriptional regulator; regulates the expression of aio/arx operon 2012)(Liu et al., 2012; Zargar et al., 2012) arsH ArsH Organoarsenical oxidase (Xue et al., 2014; J. Chen, Bhattacharjee, et al., 2015) arxA ArxA As(III) oxidase; catalyses the oxidation of As (III) arxC ArxC Quinol oxidoreductase subunit; involved in oxidation of As (III) (Richey et al., 2009) arxD ArxD Molybdoenzyme chaperon; involved in oxidation of As(III) aroS/aroR AroF/AroR AroS/AroR Histidine kinase; involved in the As(III) dependent transcriptional regulation (Sardiwal et al., 2010)

46 moeA MoeA Molybdenum cofactor biosynthesis protein; synthesises the molybdenum cofactor Arsenic transport arsA ArsA As(III)-pump ATPase; catalytic subunit of an oxyanion-translocating ATPase (Lin et al., 2006) arsB ArsB As(III)-pump protein; helps extrude As(III) from the cell (Tisa and Rosen, 1990) arsD ArsD Arsenical metallochaperone; transfers trivalent metalloids to ArsA ACR3 ACR3 As(III) permease; helps extrude As(III) from the cell (Piotr et al., 1997) arsJ ArsJ Organoarsenical efflux permease; helps extrude organoarsenicals from the cell (Chen et al., 2016) arsP ArsP Efflux system; helps extrude trivalent organoarsenicals from the cell (J. Chen, Madegowda, et al., 2015) pgpA PgpA P-glycoprotein-related protein; recognises and transports thiol-metal connjugates (Grondin et al., 1997)

Table 2-2. Targeted amplification of the arrA gene from the metagenome

Primers used Particulars of the gene amplified Reference ArrAfwd/ ArrArev ~160-200 bp amplicon tested positive for arrA gene. (Malasarn et al., 2008) The genomic arrA gene (~2000 bp) from Desulfosporosinus sp strain Y5, (Pérez-Jiménez et al., arrPJ0100F/ arrPJ2100R was amplified. 2005) ~500bp amplified products closely matched B. arseniciselenatis, and HAArrA-D1F /HAArrA-G2R (Kulp et al., 2006) Shewanella sp. Strain ANA-3. 165bp arrA genes were amplified. ArrAfwd, ArrArev 625bp arrA genes were amplified and its 72.2% amino acid sequences (Lear et al., 2007) AS1F, AS1R and AS2F resembled. Sulfurospirillum barnesii. The amplified arrA genes (~625) resembled the arrA genes of 5 bacterial AS1F and AS1RAS2F and AS1R (Pederick et al., 2007a) genes that respire arsenic. ArrAUF1 and ArrAUR3 852 bp fragment of the arrA was amplified. (Fisher et al., 2008) ~630 bp amplicons from the sediment sample showed the presence of arrA AS1F and AS1R AS2F and AS2R (Song et al., 2009) gene. ArrAUF1 and ArrAUR3- Functional gene arrA was amplified from the sediments. (Héry et al., 2010) AS2F/AS1R arrA genes (~160-200bp, 625bp and 500bp) were amplified in all the 3 different primer sets (Escudero et al., 2013) samples.

47 While the ‘arsenic gene island arrA’ regulates the reduction of As(V) in DARPs, another ‘arsenic gene island asoAB’ regulates the oxidation of As(III) in arsenic oxidising bacteria. The asoA and asoB genes encode the larger subunits of arsenite oxidase. Located upstream of asoB are approximately 15 genes, considered to regulate arsenic resistance and metabolism. Downstream of asoA are 6 genes involved in arsenic resistance and metabolism (Silver and Phung, 2005). The genes encoding these two subunits of arsenite oxidase, formally known as aoxB-aoxA/aroA-aroB/asoA-asoB are presently designated as aioA and aioB (Lett et al., 2012). The proteins coded by aioAB are distantly related to the proteins coded by arrA. Table 2-3 gives an account of various studies analysing the ‘aioAB gene island’ formerly known as the ‘asoAB gene island’. This gene island includes the genes, aioA and aioB that code for two subunits of arsenite oxidase.

Table 2-3. Targeted amplification of the As(III) oxidase gene from the metagenome. P8 and The newly found ULPAs1 strain had a gene assemblage similar to (Weeger et al., Pc1544 other As(III) oxidising bacteria. 1999) Gene island for the As(III) oxidase has been mapped. (Silver and Phung, 2005) 2 sets of aioA genes were amplified and analysed. (Inskeep et al., primers 2007) 6 novel As(III) oxidising bacteria were isolated from industrial soil (Rhine et al., 2007) in New Jersey using newly designed primers that will be reliable markers for the As(III)oxidisers (aoxAB). aroA-like genes were identified and their expression studies were (Hamamura et al., performed on Aquificalse spp. 2009) 5 pairs of 5 aoxB genes encoding arsenite oxidase and 51 arsenite transporter (Cai et al., 2009a) degenerat genes were amplified using degenerative primers. ive primers Degenera aoxB gene synonymous to aroA and asoA was amplified from (Handley et al., tive Marinabacter santoriniensis 2009) primers The new isolates from France possessed a diversity of aoxB genes (Heinrich-Salmeron that encode As(III) oxidase large subunit. et al., 2011) 17 sets of The aioA sequence from X14 corresponded to the aioA gene from (Delavat et al., 2012) specific Thiomonas sp. (989bp) primers The aio genes include two subunits of the arsenite oxidase enzyme (Lett et al., 2012) (aioA and aioB), a sensor histidine kinase (aioS), a transcriptional regulator (aioR) and an oxyanion binding protein (aioX). The genome analysis of Thiomonas arsenitoxidans showed the (Slyemi and presence of the As(III) oxidase gene cluster (aioBA cluster). This Bonnefoy, 2012) gene cluster is reported to be different from the others, as these resembled metaloregulator genes. 2 sets of A fragment of the 500bp aioA gene was amplified in many samples. (Escudero et al., degenerat The phylogeny resembled other arsenite oxidase genes. 2013) e primers

Besides the metabolic oxidation and reduction of arsenic, certain microbes have evolved to detoxify arsenic through the cytoplasmic reduction of As(V). The detoxifying system is entirely different from the dissimilatory arsenate reduction system, and encoded by the ars operon. Genomic studies in bacteria have revealed the presence of ars operons in various bacteria such as Alcaligenes sp., Bacillus sp., Microbacterium sp. etc.(Achour-Rokbani et al., 2010). Silver and Phung (Silver and Phung, 2005) reported the organisation of the ars operon, which consists of the arsC, arsB, arsA and arsD genes respectively, coding for the cytoplasmic arsenate reductase, arsenite membrane pump, ATPase subunit and arsenic operon regulator, whereas, Achour-Rokbani et al. (Achour et al., 2007)have reported a different organisation of the genes in the operon. Table 2-4 provides a few important genomic studies on the ars operon.

Table 2-4. Targeted amplification of the As(V) reductase gene (detox) from the metagenome.

3 sets of primers The arsenic operon in Bacillus subtilis had (Sato and Kobayashi, genes in the order arsR, ORF2, arsB and 1998) arsC. The transcriptomics and proteomic studies had confirmed the gene arrangement. The molecular analysis revealed the (Saltikov et al., 2003) presence of four genes arsDABC. arsR is a repressor which was not found 5kb upstream of arsD and 1kb downstream of arsC. The phylogenetic analysis of the bacterial (Jackson and Dugas, and archaeal arsC gene sequences shows the 2003) common origin of the arsenite reductase. The arsRBCC operon and arsC gene were analysed using genomics and the operon (Li and Krumholz, and gene were cloned in E. coli to study the 2007) expression pattern. The expression studies confirmed the arsenic resistance by reduction. 3 sets of primers The genomic analysis showed the presence (Achour et al., 2007) of arsB and ACR3-like genes. These genes confirmed the phosphate transporter genes in bacteria. The ars gene cluster was analysed and the (Achour-Rokbani et expression patterns were studied in a host al., 2010) system. The 16S rRNA and the specific gene analysis (Villegas-torres et al., of the sample showed the presence of 2011)

49 phylogenetic communities having the arsC gene. The arsC genes are clustered on a clade. 9 sets of primers 300-400bp gene coding arsenate reductase (Escudero et al., 2013) matched the GeneBank sequences coding for ArsC. 2 pairs of The acr3 gene that codes for transport (Escudero et al., 2013) degenerate primers enzyme is present in Actinobacteria and α- proteobacteria in water and in the sediment samples from Chile.

2.6.3. Whole genome sequencing (WGS) based community analysis and gene annotation

It is estimated that 98% of bacterial communities in the environment cannot be cultivated in the laboratory (Wade, 2002). Thus, a culture based investigation to unlock the ‘central dogma of genetics’ in most bacteria still remains impracticable. The combination of low cost DNA technologies, bioinformatics tools with advanced and highly tested algorithms have made it possible to study microbes that are not cultivable in the laboratory (Segata et al., 2013). Whole genome sequencing (shotgun metagenome) based gene annotation and community analysis is novel in investigating bacteria colonising arsenic impacted environments (Xiao et al., 2016). Before the advent of the shotgun metagenome sequencing analysis, most studies depended on cloning based sequencing and community analysis, which was bound to various limitations including time, cost, accuracy and technology. In other words, many studies were limited to the 16S rRNA gene based community analysis and target gene amplification. Therefore, the metagenome data we have so far are mainly from the 16S rRNA gene or from specific functional genes.

Recently, the WGS has experienced an innovation in terms of generating a large database of genomic libraries, which makes it possible to acquire the total genome of a bacterium. Two arsenic resistant bacteria and an arsenic resistant gene (arsN) have been discovered from the metagenomic library (Chauhan et al., 2008). Shotgun sequencing based WGS analysis has revealed the presence of arsenic metabolising genes found in bacteria whose

50 DNA was extracted from a low arsenic environment (Xiao et al., 2016). WGS based metagenome analysis of the arsenic contaminated environments is still scarce. Furthermore WGS based gene annotation techniques are effective only in the light of metatranscriptomics and metaproteomics through which the expression of the genes is confirmed (Hultman et al., 2015).

2.7. Metatranscriptomic analysis of arsenic mRNAs

“The “metatranscriptome” refers to the subset of genes expressed within a microbial community at a certain point in time” (Warnecke and Hess, 2009). The expression pattern depends on the environmental factors that the bacteria are subjected to. At times, according to the environmental factors, genes are turned on or off (Trapnell et al., 2012). Therefore, profiling the mRNA is an effective way to define the functions of the genes. In the metatranscriptomic process, samples obtained from the environment are used to extract total RNA from which mRNA is enriched using different methods such as magnetic bead capturing, preferential polyadenylation of the mRNA and preferential digestion of the rRNA. Through reverse transcription, enriched mRNA is transcribed into cDNA using random priming or priming with poly-dT primers. The sequenced cDNA is subjected to computational genome analysis (Carvalhais and Schenk, 2013).

Very few studies have been conducted on the transcriptomics of arsenic in cultures, and metagenomic studies on the genes and communities associated with arsenic are in their infancy. Owing to the difficulties in processing highly unstable mRNA from the environment, transcripts are analysed from cultured samples. Table 2-5 presents some transcriptomic works related to arsenic. Table 2-5. Arsenic related transcriptomic works

Bacteria Transcripts Author Shewanella arsC transcript that encodes the cytoplasmic arsenic (Saltikov et al., 2003) ANA-3 reductase was extracted and analysed. 5 bacteria arrA transcripts were extracted from 5 different bacteria. (Malasarn et al., 2004) Variovorax sp. mRNA coded by aroA like gene was analysed. (Inskeep et al., 2007) Marinobacter aoxB (Handley et al., santoriniensis 2009) Geobacter arrA (Giloteaux et al., lovleyi 2012)

51 One of the main objectives of this study is to analyse the environmental mRNA transcribed by the arrA genes. To the best of my knowledge, only one study has conducted ‘whole metatranscriptomic analysis’ of arsenic contaminated samples (Edwardson and Hollibaugh, 2017). Owing to the difficulties in getting mRNA from soil and water, only two samples were analysed for metatranscriptomics.

2.8. Metaproteomic analysis of arsenic proteins

The direct analysis of the proteins obtained from environmental samples is referred to as metaproteomics. 2D and 3D gel electrophoresis techniques, used to separate environmental protein samples, have been replaced recently by Mass Spectrometry (MS) analytical methods (Maron et al., 2007). The separated proteins are subjected to a proteomic database search to identify the amino acid sequence in the peptides. And through reverse genomics, the coding genes are traced back in order to analyse the microbial community possessing the relevant genes. Using these techniques, the presence of various arsenic proteins in an environment could be identified and characterised. From pure culture, many proteins have been isolated and characterised (Krafft and Macy, 1998; Afkar et al., 2003; Malasarn et al., 2008). The following proteins play important roles in the transport and biochemical change of arsenic in microbial cells.

The respiratory arsenate reductase (ARR), both isolated and characterised from the periplasm of Chrysiogenes arsenatis (Krafft and Macy, 1998), has two subunits namely ArrA and ArrB of 87 kDa and 29 kDa in size, respectively. Since the enzyme uses arsenate alone as the electron acceptor, the gene is switched on only in the presence of arsenate. As the enzyme contains molybdenum along with iron and sulphur, it is placed under the dimethyl sulfoxide (DMSO) reductase family of molybdenum-containing oxidoreductases that includes other terminal reductases used in microbial respiration (Malasarn et al., 2008). The presence of iron and sulphur in the centre acts as prosthetic groups for the enzymatic reactions. The same enzyme is isolated and characterised from other bacteria including Shewanella sp.

52 ANA-3 (Malasarn et al., 2004) and Bacillus selenitireducens (Afkar et al., 2003). Though the size and the N-terminal sequences vary in the subunits of the enzymes, their enzyme activity correlates to one another. Table 2-6 presents a few important works on arsenate reductase enzymes from different bacterial species.

Table 2-6. Respiratory arsenate reductase from different bacteria

Bacteria Proteins and subunits Location Reference Chrysiogenes ARR* (123Kda) (Krafft and Macy, Periplasmic arsenatis ArrA (87Kda) and ArrB (29Kda) 1998) Heterodimer (150 KDa) composed of two Bacillus Membrane (Afkar et al., 2003) subunits ArrA (110 KDa) and ArrB (34 selenitireducens bound kDa) arrA product (respiratory arsenate (Saltikov and Periplasmic reductase) 95.2kDa and Newman, 2003) Shewanella ARR (131Kda) ANA-3 ArrA (95Kda) and ArrB (27Kda) Periplasmic (Malasarn et al., 2008)

15 amino acid sequence Respiratory (Fisher et al., 2008) Alkaliphilus sp. Arsenate Reductase The clade of the arsenate respiratory reductase closely resembles the (Duval et al., 2008) polysulfide reductase as opposed to the arsenite oxidase. The Arr activity was reversible and performed the activity of both the (Richey et al., 2009) reductase and oxidase. Cytoplasmic arsenate reductase (ArsC), isolated and characterised from the Staphylococcus aureus arsenic-resistant plasmid p1258 was a monomer having a molecular mass of 14.5 Kda (Ji et al., 1994). The mechanism of arsenate reduction in this detoxification system has been studied in another ArsC isolated from Escherichia coli plasmid R773, which reduced arsenate to detoxify the system (Rosen, 1999). ArsC from E. coli needs reduced glutathione and glutaredoxin for the arsenate reductase activity. The single catalytic cysteine residue, Cys12, forms a covalent thiolate-arsenate complex (Mukhopadhyay and Rosen, 2002b). At present, glutaredoxin supports the reduction of enzyme bound thiolate-arsenite complex. When reduced, As(III) is pumped out by transport proteins.

The arsenite oxidase (AioAB formerly known as ArsAB) isolated from the periplasm of Alcaligenes faecalis consisted of a monomer of 85 kDa

53 containing one molybdenum, five iron and inorganic sulphide (Anderson et al., 1992). The arsenite oxidase that contains a molybdopterin and [Fe-S] cluster possesses an iron in the redox-active centre. A 2D model of the arsenite oxidase is drafted to illustrate the possible binding sites of the enzyme to form the enzyme substrate complex (Silver and Phung, 2005). In this model, the arsenite oxidase has two subunits, ArsA, an 88 kDa molybdopterin subunit and a 14 kDa Rieske subunit.

To the best of my knowledge no metaproteomic work related to the arsenic contaminated aquifers has been reported to date. Even the continuous efforts in this study to extract proteins from the sediments were futile. So the proteomics done in this study was based on Shewanella sp. ANA-3, a model organism that respires arsenic.

2.9. Summary

Arsenic contamination in aquifers affects more than a 100 million people in Asia alone. The predominant form of arsenic in the aquifer sediments is typically As(V) whereas in water it is predominately As(III), which is generally more mobile and reactive than As(V). The reduction from As(V) to As(III) is mediated by many microbial communities including dissimilatory arsenic reducing microbes that colonise these arsenic impacted aquifers. Though various studies have confirmed As(V) reduction is a microbially mediated process in these aquifers, the exact mechanism remains a scientific mystery. Analysing the microbial communities and their functional genes in these environments through metagenomics, metatranscriptomics and metaproteomics methods could shed light on the exact mechanism of arsenic mobilisation in these aquifers: and this may help mitigate arsenic pollution.

54 Chapter 3 : Methodology

3.1. Introduction

From sampling to analysis of the downstream processes, many experimental techniques and bioinformatics tools have been used in this research. This chapter briefly describes the methods and the principles behind the various techniques and tools used in this research. Certain parts of the methods are repeated in the forthcoming chapters to support the publication format. Figure 3-1 presents the various techniques and bioinformatics tools used in this thesis.

Figure 3-1. Summary of instruments, techniques and bioinformatics tools used in this project.

3.2. Sampling methods

The samples and subsamples were collected, stored and transported using different methods for the different branches of the studies. The water samples were analysed both in the field and in the laboratory depending on the experiments. For the metaomic works, a filtering technique was used to concentrate microbial cells from the tube wells. In Bangladesh, the water samples were collected and filtered using submersible plastic pumps

55 (Groundwater Essentials), a 10” filter housing (Cole-Parmer), and 0.2-μm Serial Nylon filters (Cole-Parmer). Water was pumped through the filters for between 24 and 72 h at a rate of ∼4.2 L·min−1 and a pressure of 5 psi. The filters were capable of collecting only the planktonic bacterial communities (Mailloux et al., 2013). In Cambodia, the samples were collected using the hand filtering system using manual vacuum pumps. Upon filtering, the filters were snap frozen and transported to the laboratories.

For the geochemical analysis including As, other trace elements, and major cations, water samples were collected in 30-ml or 60-ml acid-cleaned HDPE bottles and acidified to 1% HCl (Fisher Optima) immediately after collection and without filtration (Zheng et al., 2005). Samples for anions were collected in nanopore-washed 30 ml HDPE bottles without filtration. The pH and electrical conductivity of the groundwater was measured using a pH/Eh meter (Orion 210A) and a conductivity/temperature meter (Orion 105A+) with waterproof probes that were calibrated on the day of sampling (Zheng et al., 2005; Dhar et al., 2008). Figure 3-2 helps visualise the field set-up for collecting both water and sediment samples.

Figure 3-2. Field photos on sediment and water sampling for metaomic analysis.

56

Sediment cores were collected using the “hand-flapper” method, a manual drilling method often used to install wells (Horneman et al., 2004; van Geen et al., 2006). Upon coring, the sediment samples were split into aliquots for different types of analysis. The samples used for XAS analysis were saturated with glycerol (approximately 1:1 v/v) as a preservative, and stored at - 20o C prior to the analysis. Glycerol is added in order to prevent oxidation by slowing oxygen diffusion, and because it fixes bacteria to prevent microbiological alteration. In this case, spectra were collected using a thin ( about 2 mm thick) film of wet sediment without further treatment. The sample collection and synchrotron access were coordinated to minimise the lag time (about 2 weeks) between sampling and analysis. The samples used for XRF were used immediately in the field (without treatment) and samples for other analyses were stored in the refrigerator prior to the analysis. From each depth, approximately 5 g of sediment was incised aseptically in an anaerobic o bag flushed with N2 and stored anaerobically at -20 C for molecular (DNA/RNA) investigations. The samples were hermetically sealed and transported to the respective laboratories for analysis purposes. The diffuse spectral reflectance of fresh cores was also measured in the field as a proxy for the Fe(II)/Fe(II+III) ratio in the acid-leachable Fe fraction (Horneman et al., 2004).

3.3. General methods involved in the molecular analysis

For recovering, confirming, purifying, quantifying, characterising and sequencing the nucleic acids and peptides, many different molecular techniques are frequently used. This part of the chapter describes the different areas of molecular work implemented in this research.

3.3.1. DNA extraction

Extracting the DNA from water and sediment samples is done in three steps, where in the first step the cell lysis is mediated through a combination of bead-beating and detergent lysing (SDS and EDTA) that breaks down the

57 fatty acids and lipids of the cell wall. In the second step, the inhibitors including organic and inorganic compounds (humic substance, cell debris and proteins) are removed by precipitating them with potassium acetate. In the third step, DNA is mobilised to a silica support in alkaline pH, and followed by the elusion of the bound DNA with TE buffer solutions (Ogram et al., 1987; Yeates et al., 1998). In this research, the DNA from the sediment samples was extracted using PowerSoil- and PowerWater- DNA extraction kits (MOBIO Laboratories, Carlsbad, CA, USA).

3.3.2. RNA extraction

RNA unlike DNA is highly unstable and susceptible to degradation by ribonucleases. Therefore, extracting RNA from the sediment and water with low biomass is challenging. With the combination of bead beating with detergent (CTAB) and enzymatic lysis methods, the cells are disrupted in the first step. In the following step, a phenol: chloroform (liquid-liquid) extraction is applied in order to separate the low dense nucleic acids from the other organic molecules. The separated nucleic acids were purified by washing and precipitating twice using isopropanol and ethanol respectively. In the final step, the DNA was digested with DNase to obtain RNA alone which was subjected to mRNA enrichment (rRNA depletion) in the following step (Frias- Lopez et al., 2008; Urich et al., 2008; Gilbert, 2010).

3.3.3. Protein extraction

Protein recovery from the water and soil samples poses a big challenge due to its complex matrix. To a large extent, proteomics still depends either on microcosm based proteomics or on pure culture based proteomics. Having failed to recover proteins from the soil and water samples in this research, we extracted protein from Shewanella sp. ANA-3. In the extraction process, the cells were harvested from the mid log phase of the bacterial culture and were centrifuged at 3220 g for 20 minutes and then heat-shocked with an LDS buffer at 90 degrees for 5 minutes, and the supernatant run on the SDS-PAGE gel to separate the proteins. The gel was dispensed in a selective buffer and

58 concentrated using a centrifuge (Jin & Manabe, 2005; Mallick & Kuster, 2010; Georges et al., 2014).

3.3.4. Primer design

Designing an appropriate primer set is vital for any PCR-based DNA amplification process. Many tools with advance algorithms are available online for the primer design. In this research, Primer-BLAST, a web interface, was used to design new primer sets for amplifying the arrA gene. The tool works on the basis of finding the primer flanking region (region adjacent to the 5' end of the target gene) and then they are searched against the nucleotide database for the potential targets (Ye et al., 2012). While designing the primers, parameters including the GC content, primer length and temperature boundaries were considered. The designed primers were synthesised by Integrated DNA Technologies (IDTdna, BVBA, Leuven, Belgium).

3.3.5. Thermal cycler and polymerase chain reaction (PCR)

The polymerase chain reaction is the pivotal step in which a region of double stranded DNA is replicated exponentially over multiple cycles, using dNTPs and the DNA polymerase enzyme. The thermal cycler used to conduct the polymerase chain reactions exploits the principle of temperature modulations (three different temperatures for denaturing, annealing and extension) in each cycle (Weier & Gray, 1988; Stirling & Bartlett, 2003). Though there are different thermal cyclers in the laboratory where the research is conducted, we used Techne TC-300 thermal cyclers for this research.

3.3.6. Gel electrophoresis and documentation

The separation of macromolecules in an electric field is referred to as electrophoresis. When the electric field is applied in an electrode tank with buffer, the charged molecules travel from the cathode (-) to the anode (+). Depending on the molecular size and charge, the molecules travel through the pores of the gel resulting in separation by size (Vesterberg, 1993). Agarose and

59 polyacrylamide gels are two main types of gels used for separating the micro- and macro molecules (DNA, RNA, peptides). The pore size of the matrix can be defined by altering the percent of gel used.

The agarose gel electrophoresis is the simplest and most effective method for separating and identifying 0.5- to 25-kb long fragments of DNA. The principle behind this method is the separation by molecular size, where the separation matrix is agarose gel. With the aid of DNA binding dyes including ethidium bromide or commercial dyes such as SYBR Green and Picogreen, the separated DNA is visualised (Vogelstein & Gillespie, 1979).

SDS-PAGE (Sodium Dodecyl Sulphate – Polyacrylamide Gel Electrophoresis) is an electrophoresis method, which uses polyacrylamide as the matrix and sodium dodecyl sulphate as the anionic detergent that creates a net negative charge within a wide pH range (Shapiro et al., 1967). The charge to mass ratio based migration is the basic principle behind this method. The SDS imparts a net negative charge and destroys the complex secondary and tertiary structure of the proteins, and the polyacrylamide gels restrain larger molecules from migrating as fast as smaller molecules. When electricity is applied, the negatively charged peptides start migrating towards the positive charge and the rate of travel is influenced by the size of the molecule and the sieve of the gel. As a result, different sized molecules get separated in the gel. The proteins in general are stained with a dye such as Coomassie Brilliant Blue. For this research, a commercially available instant SDS-PAGE gel (Novex, Life technologies) was used to separate the proteins.

Gel Documentation (Gel doc) relies on the principle of UV mediated fluorescence and imaging. Upon exposure to the ultra violet radiation emitted by the UV light transilluminator in the Gel Doc, the fluorophores get excited and emit light, which in turn is imaged by charge coupled cameras. GEL DOC 2000, BIO-RAD was the system used in this research.

3.3.7. DNA and RNA quantification For nucleic acid quantification, we used a combination of spectrophotometers and flu0rometers.

60 Nucleic acid quantification using the Qubit 3.0: Qubit exploits the fluorometric principle. A fluorescent dye binds to the nucleic acid, and the difference in fluorescence between bound and unbound dye is several orders of magnitude. The dye intercalated between the bases in DNA or RNA is excited when exposed to UV. The fluorescent signal is picked up by the detector. The amount of fluorescence signal from this mixture is directly proportional to the concentration of nucleic acid in the solution. Nucleic acid quantification using a bioanalyser (Agilent 2100 Bioanalyzer): This works on the combined principles of gel electrophoresis and fluorometry (charge-to-mass ratios and excitement-emission). Each chip within the machine contains an interconnected set of gel-filled channels that allow molecular sieving of nucleic acids. A series of electrodes control the sample movement within the chip. These make contact with the samples when the instrument lid is closed. Each electrode is connected to an independent power supply, providing maximum control and flexibility (Panaro et al., 2000). Figure 3.3 shows the high sensitivity DNA chip and its internal structure.

Figure 3-3. Bioanalyser-Agilent high density nucleic acid chips.

61 3.3.8. SequalPrep™ normalisation of PCR products

The sole objective of this normalisation process is to purify the PCR products amplified for MiSeq sequencing. Under low pH conditions, the positive surface charge of the ChargeSwitchR coating binds the negatively charged nucleic acid backbone. Proteins and other contaminants (such as short oligonucleotide primers) are not bound and are simply washed away.

3.3.9. Pooling

Pooling of a library is a term used for mixing equal concentrations (4nM) of DNA (in MiSeq) from different samples (Kozich et al., 2013; Schirmer et al., 2015). The pooling of samples enables the proportional synthesis of reads from each sample during sequencing. The library concentration in each sample is calculated using the following formula:

퐶표푛푐푒푛푡푟푎푡푖표푛 푖푛 ng/μl 6 퐷푁퐴 퐶표푛. 푖푛 푛푀 = g ∗ 10 660 ∗ 푎푣푒푟푎푔푒 푙푖푏푟푎푟푦 푠푖푧푒 mol If the molarity of each sample library exceeds 4nM, they are diluted with 10mM Tris-HCl buffer.

3.3.10. Next generation sequencing

The newer sequencing methods other than the Sanger sequencing (first-generation sequencing) technology are referred to as next-generation sequencing (NGS). Most of the next generation methods rely on the principle referred to as ‘sequencing by synthesis’, in which all four nucleotides labelled with a distinctive fluorophore are added simultaneously with the DNA polymerase. The nucleotides emitting fluorescence while being added during the synthesising of new strands are captured with an aid of charge coupled device cameras followed by the conversion of the intensity of fluorescent signals into flowgrams, which reveal the underlying DNA sequence.

The detailed narration of these techniques is published in various research articles and reviews (Hert et al., 2008; Mardis, 2008; Wall et al., 2009; Chistoserdova, 2010; Metzker, 2010; Shendure & Lieberman Aiden, 2012; Kozich et al., 2013). In these methods, a concise description of sequencing

62 machines including Roche/454, MiSeq and HiSeq-Illumina and the principle behind these techniques used in this research alone are given. Figure 3-4 presents the picture of sequencing machines used in this research.

Figure 3-4. The next generation sequencing (NGS) machines used in the project.

3.3.10.1. Roche/454 (Pyrosequencing)

Pyrosequencing is built on the principle of ‘sequencing by synthesis’, in which the synthesis and sequencing of a DNA fragment take place simultaneously, as illustrated in Figure 3-5 . In pyrosequencing, an indexing PCR step adds adapters to the DNA fragments that bind to beads coated with streptavidin and complementary adapters and primers (Pourmand et al., 2002; Nyrén, 2007). The concentration of mixing the fragments with the beads gave a high probabilistic chance of only 1 fragment of DNA to one bead and each bead with the reaction mixture is emulsified in a bubble, where more than a million copy of DNA fragments are synthesised.

63

Figure 3.5. An illustration of a cycle in a pyrosequencing reaction.

Upon completion of the polymerase reaction, each bead hybridised with more than a million copies of identical DNA fragments is transferred to a well, which contains a mixture of primers, adenosine 5’ phosphosulfate (APS), luciferin and the enzymes including DNA polymerase, ATP sulfurylase, luciferase and apyrase. In every cycle, one of four dNTPs is added to the well. If the added nucleotide is complementary to the nucleotide in the DNA fragment, the polymerisation reaction takes place resulting in the release of inorganic pyrophosphate (PPi, which in turn is converted to ATP by the ATP sulfurylase enzyme in the presence of APS. The ATP supplies energy for the conversion of luciferin to oxyluciferin mediated by the enzyme luciferase resulting in the emission of visible light. The intensity of light captured by the camera is converted to an electropherogram from a pyrogram. The unused nucleotides and ATPs are degraded in every cycle by the enzyme called apyrase (Ronaghi, 2001; Pourmand et al., 2002; Nyrén, 2007; Roesch et al., 2007; Mardis, 2008). The pyrosequencing platform was withdrawn from the market in 2016.

64 3.3.10.2. MiSeq and HiSeq (Illumina)

The Illumina sequencing platforms of MiSeq and HiSeq exploit the principle of “sequencing by synthesis” in common with pyrosequencing. In this sequencing method, the target DNA is cleaved into small fragments (75-300 bp) and the fragments are ligated with ligands. The tagged fragments are

Figure 3-6. An illustration of the steps involved in sequencing by means of the synthesis method used in Illumina MiSeq and HiSeq. (Figure adapted from the Illumina manual). hybridised to the complementary adapter sequences mobilised in the wells in the flow cell. When four different fluorophore tagged nucleotides are added to the wells, the DNA polymerase enzyme adds the complementary nucleotides to the elongation strand. In the process, the colour and the intensity of the light emitted is captured by the CCD camera and the images are interpreted into sequences. Once the cycle is completed, the wells are washed and the new cycle starts and repeats until the number of programmed cycles is completed. Figure 3-6 illustrates the important steps involved in sequencing DNA using the Illumina platform.

65 Though the principle and steps involved in DNA sequencing are the same for both MiSeq and HiSeq, quantitatively there are a few differences as shown below (Table 3-1).

Table 3-1. The differences between MiSeq and HiSeq2500

MiSeq HiSeq2500

No of Flow Cells 1 2 Lanes/FC 1 lane/FC 2 or 8 lanes

Max. Clusters/run 25 M 600 M

Max. Read Length 2x300 bp 2x150 bp

Max. Output/run 15 Gb 1500 Gb

Max Read Number 25 million 5 billion

Max. Hrs/run ~56 hrs ~6 days

Cost ~ £70,000 ~ £700,000

3.3.11. High performance computing (HPC)

High performance computing (HPC) is the use of parallel processing for running advanced programmes that handle enormous data efficiently, reliably and quickly. The HPCs, in general use functions above a teraflop or 1012 floating point operations per second (Younge et al., 2011). Most tools used to analyse the genomic data in this thesis were installed on a Linux operating system used on these servers. The huge output of sequences (~50 GB data) generated by NGS technologies demand parallel and speedy processes which are vital to compare and contrast the databases. For the downstream processing of the genomic data in this research, we used two HPC systems offered by the University of Manchester and the Ohio State University. The pipelines were used for the 16S rRNA gene based community analysis and whole genome sequence based gene annotation.

3.3.12. Bioinformatics tools

Bioinformatics tools are key to omics studies, which connect the tree of life through genes and their functions. In this research many bioinformatics

66 tools were used in down-streaming processes. Considering the enormity of the various tools with advanced algorithms, a synopsis of the main tools that I used is tabulated below in Table 3-2.

Table 3-2. The bioinformatics tools used in this research

Tool Description Reference It is an automated phylogenetic inference tool which AMPHORA is used for genome tree construction and (Wu & Scott, 2012) 2 metagenomics phylotyping. BLAST stands for Basic Local Alignment Search Tool and is used to find regions of local similarity between sequences. The programme is the basic tool (Altschul et al., 1990; BLAST to analyse genes and proteins. In this study, it is Ye et al., 2012) used to infer the functional and evolutionary diversity between the sequences, design primers, and tracing a species through an OTU. The Illumina BaseSpace sequence hub is a cloud based genomics analysis and storage platform. In BaseSpace this research it is used to transfer the HiSeq reads over the cloud. Bowtie2 is used for aligning sequencing reads to long reference sequences. It is an ultrafast and memory (Langmead & Bowtie2 efficient tool for aligning sequences from 50bp to Salzberg, 2012) 100bp of length. It supports gapped, local and paired-end alignment modes. The Consensus Assessment of Sequence and Variation (CASAVA) is an Illumina pipeline used to assess the quality of the sequence and to align the CASAVA (Hosseini et al., 2010) amplicons against the reference genome. In this research, the bcl2fastq conversion tool in the CASAVA pipeline was used. It is a multiple sequence alignment software used to align the DNA and amino acid sequences. The best ClustalW2 match for the selected sequences is calculated along (Chenna et al., 2003) with the identities, similarities and differences thereof. Cufflinks is mainly used for the RNA-Seq analysis. While it can function as a reference-based de novo Cufflinks transcriptome assembler, it can also identify novel (Trapnell et al., 2012) transcripts in the data by examining their alignments to the genome. Cutadapt is used to remove adapter sequences, Cutadapt linkers and primers from the high throughput (Martin, 2009) sequencing reads. The expectation maximisation iterative reconstruction of genes from environment (EMIRGE) method is used to reconstruct the small subunit ribosomal RNA gene (SSU gene) for EMIRGE (Miller et al., 2011) bacterial taxonomy and identification. The algorithm capitalises on the probabilistically possible full length SSU sequences to assess the completeness of the genome.

67 FastQC provides visual summary graphs and tables of the sequence data quality. It furthermore points FastQC (Andrews, 2010) out where there is any problem with the data or which part of the data has the problem. It is designed to create phylogenetic trees for publications. The advantage of this software is that FigTree the tree can be collapsed in various level to present (Rambaut, 2009) the diversity and the interconnectedness of the microbes the researcher is interested in. Geneious is a commercial software used for primer design, DNA alignment, mapping sequences etc. In Geneious this study it has been mainly used to visualise the (Kearse et al., 2012) phylogenetic trees and to visualise the gene arrangements in an operon. The iterative de Bruijn graph De Novo assembler is used to align short read sequencing data with a highly uneven sequencing depth. In each iteration, IDBA-UD short and low depth contigs are removed to (Peng et al., 2012) reconstruct longer contigs with higher accuracy. In this study IDBA-UD was used to assemble the paired end sequences generated by HiSeq2500. The Kyoto Encyclopaedia of Genes and Genomes (Ogata et al., 1999; (KEGG) Automatic Annotation Server (KAAS) is KAAS or Kanehisa, 2002; used to annotate genes in complete genomes based KEGG- Moriya et al., 2005, on the best-hit information. In this study, more than KAAS 2007; Kanehisa 90 bins were analysed for functional genomes and Laboratories, 2017) their pathways. The Molecular Evolutionary Genetics Analysis (MEGA) software is a desktop application which was MEGA regularly used in this research to align and blast 16S (Kumar et al., 2008) rRNA based OTUs. Also the arrA nucleotides were translated into amino acids for the peptide blast. MetaBAT is a genome binning software which integrates empirical probabilistic distances of MetaBAT (Kang et al., 2015) genome abundance and tetranucleotide frequency to bin the scaffolds belonging to a microorganism. It is an alternative software used for multiple sequence alignment of protein and nucleotide (Robert C Edgar, MUSCLE sequences. When the sequences are below 1MB, it is 2004; Robert C. easy to align the desired sequences with high Edgar, 2004) accuracy. It is (PAired eND Assembler) an open source tool specifically designed to assemble paired end reads rapidly and with the correction of most errors. The PANDAseq (Masella et al., 2012) speciality of this programme lies in error masking on simulated data. In this research, this programme is used to assemble the paired end reads. Pfam is a large collection of protein multiple sequence alignments database. Pfam is extensively (Bateman, 2002; Finn Pfam used in the gene annotation workflow in the et al., 2016) metagenomic, metatranscriptomic and proteomic work flow. It is an open source software for an online based PredictProt prediction of protein structure and function. This (Yachdav et al., 2014) ein tool is used in this research to see the homology of the protein sequences in the arsenic proteomic

68 experiment. The PROkaryotic Dynamic programming Gene- finding Algorithm is a prokaryotic gene recognition and translation initiation site identification tool, Prodigal extensively used in this research for gene annotation. (Hyatt et al., 2010) Though Prodigal is an integral part of the pipeline, individual sequences were manually annotated using Prodigal. QIIME expanded as ‘Quantitative Insights Into (Caporaso, Microbial Ecology’ is an open-source software Kuczynski et al., 2010; QIIME pipeline based on the PyCogent toolkit. In this Caporaso, Lauber et research, this workflow was used to analyse 16S al., 2010) rRNA gene data generated by 454. Scaffold is a proteome software used to visualise and validate complex peptides generated by MS/MS. In Scaffold this research, Scaffold is used to contrast the peptide (Searle, 2010) sequences from Shewanella sp. ANA-3 grown in media with two different electron acceptors. It is used to trim sequences on the basis of the sliding window, length and Phred score. It is (Joshi& Fass, 2011; programmed to trim both single and paired-end Sickle Del Fabbro et al., reads and different quality scoring formats. It is 2013) specially programmed to consider the quality of the reads that decreases at both ends. It is an open-source software used to assemble single or multicellular reads. The algorithms in the (Bankevich et al., SPAdes programme are specifically written for a de nova 2012) assembly of the reads, which is beneficial for characterising any novel microorganism. It is a database containing a collection of protein families featuring curated multiple sequence TIGRFAMs alignments, designed to support the automated (Haft, 2001) functional identification of proteins by sequence homology. The UPARSE pipeline, a commercial pipeline, is a method for generating clusters of OTUs representing UPARSE the 16S rRNA gene. The clustering method is (Edgar, 2010, 2013) USEARCH implemented as the cluster_otus command in USEARCH, but the method itself is the UPARSE- OTU algorithm. VSEARCH is an alternative open source pipeline to the UPARSE pipeline. The programme is mainly VEARCH (Rognes et al., 2016) used for OTU picking from the assembled 16S rRNA gene sequences.

3.4. Metagenomic methods

Metagenomic analysis includes two major approaches, namely the 16S rRNA gene based community analysis, which is conducted in order to obtain a panoramic view of the microbes present in the samples and the whole genome sequencing WGS based gene and community analysis to enumerate the genes and functions in a particular microbe. In this research, two samples were

69 subjected to the WGS based metagenome analysis and the rest of the samples for the 16S rRNA gene based community analysis.

3.4.1. 16S rRNA gene based community analysis

Depending on the sequencing platform and the bioinformatics tools chosen in the downstream processing, 16S rRNA marker gene based community analysis was conducted by means of two distinctive workflows in this research. The first workflow was based on high through-put sequences generated by the Roche 454 pyrosequencing technique and QIIME pipeline, whereas the second work flow was based on reads generated by the sequences generated by the MiSeq sequencing platform and the downstream pipeline generated exclusively to handle MiSeq reads, by the University of Manchester.

3.4.1.1. Roche 454 pyrosequencing and QIIME pipeline

The workflow followed for the research in Chapter 4 is described in the flowchart given in Figure 3-8. The workflow starts with the DNA from the sediment samples being extracted using the PowerSoil DNA extraction kit (MOBIO Laboratories, Carlsbad, CA, USA). Upon extraction of the DNA, the V1-V2 hypervariable region of the bacterial 16S rRNA gene was amplified using tagged fusion universal bacterial primers 27F (Lane, 1991) and 338R (Hamady et al., 2008) synthesised by IDTdna (Integrated DNA Technologies, BVBA, Leuven, Belgium). The fusion forward primer consists of 454 Life Sciences - Lib-L-Primer A, a 4-mer key, a 10-mer barcode (MID) and the forward primer 27F and the fusion reverse primer consists of a 454 Life Sciences-Lib-L-Primer B, a 4-mer key and the reverse primer 338R (Figure 3-7).

Figure 3-7. A diagrammatic depiction of the primer structure used for the pyrosequencing PCR.

70 The V1-V2 hypervariable region of the 16S rRNA gene was amplified in 50-μL volume reactions using 0.5 μL (2.5 units) FastStart High Fidelity DNA polymerase (Roche Diagnostics GmbH, Mannheim, Germany), 1.8 mM MgCl2, 200 μM of each dNTP, 0.4 μM of each forward and reverse fusion primers. The PCR was performed in a thermal cycler (Techne TC-300 ) with the following PCR conditions programmed: initial denaturing step at 95 °C for 2 min, 35 cycles of 95 °C for 30 s (denaturing), 55 °C for 30 s (annealing), 72 °C for 45 s (extension), and a final elongation step at 72 °C for 5 min. The PCR products were loaded onto an agarose gel for gel electrophoresis (90 volts for 45 minutes). Then the gel was transferred to geldoc (GEL DOC 2000, BIO-RAD) and bands of the correct fragment size (~410 bp) were excised, purified using a QIAquick gel extraction kit (QIAGEN, GmBH, Hilden, Germany), and eluted in 30 μL of DNAse free H2O. The purified PCR products were quantified using

Figure 3-8. The workflow of the 16S rRNA gene based community analysis using the 454 sequencing platform and QIIME downstream analysis. an Agilent 2100 Bioanalyzer (Agilent Technologies, Inc., Santa Clara, CA, USA), was used to quantify the purified libraries and pooled so that the mixture contained equal amounts of DNA from each sample. The emulsion PCR and the pyrosequencing run were performed at the University of

71 Manchester sequencing facility, using a Roche 454 Life Sciences GS Junior system (Rizoulis et al., 2014a).

The amplicon processing workflow was carried out using the software MacQIIME 1.8.0. The processing started with the collection of sequences, the 454-machine generated quality score file, which contained a score for each base in each sequence included in the fasta file and the tab-delimited mapping file containing information of 15 samples analysed in Chapter 4. In the first step using the fasta, mapping, and quality files the command line split_libraries.py split libraries with the specifications of –l 300 (minimum sequence length of 300bp), -L 400 (maximum sequence length of 400bp), -M 2 (maximum primer mismatch), --reverse_primer_mismatches 2, -z of truncate_remove. In the following step, using the pick_rep_set.py command, the operational taxonomic units (OTUs) were identified by clustering sequences with a 97% identity threshold. This was followed by the execution of the command pick_rep_set.py in which one representative sequence was assigned to each OTU and the rest of the sequences were eliminated in the following steps. In the assign_taxonomy.py step, the taxonomy for each OTU was assigned using the ribosome database project (RDP) with the confidence threshold of 0.90. The OTU heatmaps were produced using the command make_otu_heatmap_html.py and the summary of the taxonomy and their plots were produced using summarize_taxa_through_plots.py. The default programme for aligning the representatives in QIIME is uclust but for this work clustal was used to align the sequences. Using the filter_aligmnet.py and make_phylogeny.py, the phylogenetic trees were built. Using the QIIME command line, alpha and beta diversities, PCoA plots were created for the visualisation. The OTUs were manually and individually blasted to tabulate the scores.

3.4.1.2. MiSeq and the Amplicon analysis pipeline

The DNA library preparation for the MiSeq sequencer was done using the protocols published previously (Kozich et al., 2013). The DNA was extracted from the filters containing biomass using the MoBio Power Water

72 DNA Isolation Kit (MoBio Laboratories, Inc., Carlsbad, CA, USA). The extracted DNA was quantified using the Invitrogen™ Qubit™ 3.0 Fluorometer using the highly sensitive Invitrogen™ Qubit™ quantitation assays. The V4 region of the 16S rRNA gene was amplified using dual indexed primers (150x2) (Figure 3-9). The V4 hypervariable region of the 16S rRNA gene was amplified in 50-μL volume reactions using 0.5 μL (2.5 units) FastStart High Fidelity

Figure 3-9. A diagrammatic depiction of the primer structure used for the MiSeq indexing PCR. DNA polymerase (Roche Diagnostics GmbH, Mannheim, Germany), 5 μL of buffer with MgCl2, 1 μL of dNTPs, 2 μL (10 μM) of each forward and reverse fusion primers. The PCR was performed in a thermal cycler (Techne TC-300 ) with the following PCR conditions programmed: initial denaturing step at 95 °C for 2 min, 35 cycles of 95 °C for 30 s (melting), 55 °C for 30 s (annealing), 72 °C for 2 minutes (extension), and a final elongation step at 72 °C for 7 min (Kozich et al., 2013).

The libraries were quantified again to see whether all the samples had amplified products required for library preparation. The normalisation was performed using the SequalPrep™ Normalization Plate Kit (Thermo Fisher Scientific) in order to obtain an equal quantity of the DNA library, which involves the three steps of binding, washing and elution. The normalised library was sized and quantified using high sensitivity DNA chips and Agilent Bioanalyser 2100 (Agilent Technologies, Inc., Santa Clara, CA, USA). A uniform concentration of the 4 nM DNA library of each sample was achieved by pooling the samples on the basis of their concentration. The pooled library was denatured along with PhiX before loading on to the Illumina Miseq next generation sequencer. MiSeq Reagent Kits v2 chemistry was used for sequencing.

The fastq files generated by the sequencer were analysed using the pipeline established to analyse the microbial communities using the 16S rRNA

73 marker gene. The pipeline contains the tools including cutadapt, sickle, FastQC, Spades, Pandaseq, Vsearch and QIIME. The alpha diversity was analysed using the R statistical calculation package vegan and in QIIME using the rarefied data. The Principal Coordinate Analysis (PCoA) was used to compare groups of samples based on phylogenetic or count-based distance metrics. The OTUs of interest were blasted manually to tabulate the identity and similarity. The objectives of using each tool have been previously explained in Table 3-2. The entire workflow is presented in Figure 3-10.

Figure 3-10. The workflow of the 16S rRNA gene based community analysis using the MiSeq sequencing platform and amplicon analysis pipeline.

3.4.2. WGS based metagenomic analysis

The workflow (Figure 3-11) for sequencing 101x2 paired-end shotgun reads using HiSeq 2500 started with 1 ng of gDNA, extracted from the filters containing biomass using the MoBio Power Water DNA Isolation Kit (MoBio Laboratories, Inc., Carlsbad, CA, USA). Following the manufacturer’s instruction provided along with the Nextera XT DNA library prep kit, the library was prepared. In the first step, the DNA was tagmented using the Tagment DNA Buffer (TD) and the tagmentation was assessed by running 1 μL of the sample on the High Sensitivity DNA chip in an Agilent Bioanalyzer 2100.

74 The indexing PCR reagents including indices i5 and i7, Nextera PCR Master Mix (NPM) added along the tagmented DNA and the PCR was carried out in the thermal cycler (Techne TC-300) with the programmed PCR conditions of preheating (72oC for 3 minutes), initial denaturing (95oC for 30 s), melting (95oC for 10 s), annealing (55oC for 30 s), extension (72oC for 30 s) and final elongation (72oC for 5 s). The library was cleaned using AMPure XP beads in the resuspension buffer and normalised in a bead based binding and elusion steps. The normalised library was once again quantified using the Agilent Bioanalyzer 2100 before it was pooled and loaded on the sequencing cartridge (HiSeq SBS kit V4). The sequencing was done in the HiSeq 2500 system at the University of Manchester.

Figure 3-11. The work flow of the Whole Genome Sequence (WGS) analysis using the HiSeq sequencing platform and metagenomic pipeline. The shotgun reads generated by the HiSeq2500 system were demultiplexed using bcl2fastq (v.2.17.14) before they were subjected to quality control using Sickle. The quality controlled sequences were assembled using IDBA-UD and the coverage calculations were tabulated using Bowtie2. MetaBAT, a binning tool was used to bin the scaffolds to the respective bins.

75 The assembled scaffolds were annotated using prodigal, Pfam, TIGRFAM, ProSiteProfiles, Usearch and Amphora databases and tools. The bins were manually cleaned on the basis of the GC content and coverage depth. Once the bins were cleaned once again the genes were annotated and the path ways scored.

3.5. Metatranscriptomics methods

A metatranscriptomic analysis has been conducted as given in the schematic diagram in Figure 3-12. As previously explained in Section 3.3.2, the total RNA was extracted from the filters containing biomass from two aquifers identified in this study as CAT and BA13.6. The total RNA was quantified using a Nanodrop (260 nm : 280 nm) and a Qubit 3.0. The RNA was enriched using the mRNA enrichment kit (MICROBExpressTM Kit) in three steps of annealing, bead-binding and recovering, given in the manufacturer’s instruction. From

Figure 3-12. The schematic workflow of the metatranscriptomic analysis using the HiSeq sequencing platform.

76 the enriched mRNA, cDNA was prepared following the manufacturer’s instruction of the Apollo 324 PrepX mRNA library protocol (IntegenX), which is based on directional RNA adaptor ligation (Lau et al., 2016). Sample CAT was indexed with a 6-mer of “ATCACG”, BA14-3 was indexed with another 6- mer of “CGATGT”. In the following steps, the cDNA library was amplified and quantified using Agilent Bioanalyzer 2100. The quantified library was sequenced using the Illumina HiSeq 2500 platform with Illumina’s Truseq Rapid SBS chemistry.

The sequenced reads were filtered for high-quality reads using the tools in the galaxy pipeline at Princeton University. The reads below the Phred quality score of 30 (Q30) were discarded in the Quality Control (QC) step followed by demultiplexing the reads and removing the adapters and primers. Once again using Sickle, quality control was performed with the same parameters. The quality controlled reads were assembled using Bowtie2 and the genes were called using prodigal. The software Cufflink was run to count transcript abundances and calculate FPKM (Reads per kilobase of exon per million reads mapped). The genes were aligned against the metagenomics bins to find whether the organisms were active in the samples analysed.

3.6. Proteomic methods

Having failed to recover proteins from the soil and water samples in this research, I worked with a workflow (Figure 3-13) that supports an extraction and analysis of proteins from Shewanella sp. ANA-3. In the extraction process, the cells (harvested from the mid log phase of the bacterial culture) were centrifuged at 3220 g for 20 minutes and heat-shocked with LDS buffer at 90 degrees for 5 minutes and run on the SDS-PAGE to separate the proteins. The gel was dispensed in a selective buffer and concentrated using a centrifuge. The peptides were ionized and fragmented before they were analysed by liquid chromatography—MS/MS using an Ultimate 3000 Rapid Separation LC (RSLC; Dionex Corporation) coupled to a LTQ Velos Pro (Thermo Fisher Scientific) mass spectrometer. The data produced were searched again in the UniProt bacterial database. The data were validated

77 using the proteome software Scaffold (Marsili et al., 2008; Fresquet et al., 2015).

Figure 3-13. Proteomics workflow used to analyse the expression of Arsenate Reductase (ARR) in Shewanella sp. ANA-3.

3.7. Geochemical methods

Various chemical parameters were analysed in this study to characterise the chemical environment in which the arsenic related microbes express their genes. The largest part of the chemical analysis was done at Columbia University, NY and the raw data were analysed to suit our metagenomic research.

3.7.1. pH and Eh

pH and Eh were recorded in-situ during the sample collection using a multimeter equipped with relevant probes/sensors and flow cells (Professional Plus Series Portable Multimeter, YSI) (Richards, Magnone, et al., 2017).

78 3.7.2. Diffuse reflectance spectrum

The sediment samples collected were subjected to the diffuse reflectance spectrum in the field using the handheld spectrophotometer CM- 2002 (Minolta Corp., USA). The sediment samples were placed in a clear polyethylene cling-wrap (Glad®, USA). The reading was recorded three times in a sequence without moving the instrument and finally the average value was taken for the plot (Horneman et al., 2004; Zheng et al., 2005; Dhar et al., 2008).

3.7.3. X-Ray fluorescence

The elemental composition of the sediment was determined by means of X-ray Fluorescence (XRF) spectroscopy using an InnovX Delta Premium hand-held XRF analyser using triplicate measurements and a collection period of approximately 75 seconds. This device calculates environmental concentrations based on the total fluorescence at three energy levels in order to efficiently discriminate between elements such as lead and arsenic with similar emission lines. The analytical precision based on statistical counting errors is approximately ±1-2 mg kg-1 and ±300 mg kg-1 for arsenic and Fe, respectively. Under the same conditions, the analytical precision of the repeated measurements of the uniform standard reference materials collected independently over days to weeks are within 5% of each other and within 6% of the certified values. Since XRF only analyses a small quantity of the sample, the measured concentrations can vary more significantly when repeated on heterogeneous natural materials, and the analytical accuracy under the experimental conditions is thus somewhat lower, about 10-20% for any single measurement.

3.7.4. EXAFS spectra for iron and arsenic

Approximately 0.1 g of sediment from each depth was analysed for Fe and As using extended X-ray absorption fine structure (EXAFS) at the Stanford Synchrotron Radiation Laboratory (SSRL) on Beamlines 4-1 and 4-3, equipped with a 13- and 32-element Ge detector, respectively. The beamlines were

79 configured with a Si(220) monochromator with a phi angle of 90 degrees. Soller slits were installed to minimise the effects of scattered primary radiation. The beam was detuned at least 50% to reject higher-order harmonic frequencies and prevent detector saturation. Scans were calibrated to Fe K- edge of As(V) (11874.0 eV for sodium arsenate) and Fe metal (7112 eV) using a sample placed between the second and third ionization chambers. A subset of each sample was sealed in Kapton tape and mounted between the first and second ionization chambers, and sample spectra were obtained in the fluorescence mode in combination with a 6 μx Ge filter (for arsenic analysis) or Mn filter (for Fe analysis).

Spectra were processed in SIXpack (Webb, 2005) unless mentioned otherwise. Arsenic K-edge XANES spectra were normalised and fit using linear combination fitting (LCF) using orpiment (As2S3), arsenite (as adsorbed arsenite, As(III)), and arsenate (as adsorbed arsenate, As(V)) (Sun et al., 2016). Other arsenic oxidation states were considered but were not needed for any fits. Adsorbed references were created by adsorbing either sodium arsenate or sodium arsenate solutions on 1 g/L ferrihydrite solutions to a solid concentration of 1,000 mg arsenic/kg Fe concentration.

Fe mineralogy was determined using LCF fitting of EXAFS spectra with those of the reference spectra of the commonly occurring sediment minerals hematite, goethite, ferrihydrite, magnetite, mackinawite, siderite, biotite, and hornblende) (Sun et al., 2016). Averaged Fe K-edge EXAFS spectra were normalised with linear pre-edge and quadratic post-edge functions. Normalised spectra were converted to k3-weighted chi functions with a threshold energy (E0) of 7124 eV. Reference spectra to be included in linear combination fitting were selected based on minerals known to be present or commonly present in sediments, and also included minerals often found using other methods (for example, by diffraction or magnetic susceptibility). Principal component analysis (PCA) was also applied to sample k3-weighted spectra to select reference spectra; target transforms were used to compare significant components against a spectral library of Fe mineral references to

80 ensure that only references that were deemed statistically viable were included. Only target transforms having SPOIL values < 6 were considered as potential references (Beauchemin et al., 2002; Strawn & Baker, 2009). In the end, all spectra were fitted with a single set of reference spectra, although individual samples are likely to contain minerals not represented in this reference set. This approach was adopted because it quantifies Fe mineralogy in a manner that is internally consistent and representative of the oxidation state, but it means that the fractional mineral concentrations of some minerals are combined and represented with a smaller set of similar reference spectra that are not easily differentiated. For example, the Fe(II) and Fe(III) silicates all have sufficiently similar spectra that it is hard to resolve them and fits involving several of these minerals concurrently are thus unreliable. Thus, we limit fits to representative and common Fe(II) silicates such as biotite and Fe(II,III) silicates such as hornblende. As a result, care must be used in interpreting the mineralogical data as indicative of the relative concentration of specific minerals, in particular for Fe silicates. Least-squares LCF was done on k3-weighted EXAFS spectra over a k-range of 2 to ~13 Å−1, to determine the relative fraction of each of the 8 reference compounds that contributes to the sample spectra. Uncertainties reported by SIXpack include the Monte Carlo- based error propagation from fitting, spectral noise in sample and reference spectra, and similarities between the reference spectra (Webb, 2005). The fractions and uncertainties were reported in the unit of % of total Fe, and if needed, converted to concentrations by multiplying bulk Fe concentrations and expressed in the unit of mg Fe per kg sediment, i.e., mg kg-1.

3.7.5. Analysis of trace elements

Using an Axiom single-collector instrument (Thermo Elemental, Germany), the concentrations of As, P, Fe, Mn, S, Ca, Mg, K, Na and 33 other trace elements in groundwater samples were measured at Lamont–Doherty Earth Observatory with a reproducibility typically <5% by high-resolution inductively-coupled plasma mass spectrometry (HR ICP-MS) (Cheng et al., 2004; Dhar et al., 2008). Protocols that were followed to ensure the accuracy

81 and precision of the data included: (1) two NIST standard reference materials (1640 and 1643E, Trace element in natural water), and an internal laboratory consistency standard (LDEO tapwater spiked with analyte elements) were included with each run. Results for these standards were always within 5% of the certified values after calibration of the instrument with separate standards at the beginning and end of each run (2) whenever possible, timeseries samples from the same well were analyzed within the same run of 30 samples, which usually improved the reproducibility to <3%; (3) At least 2 samples were reanalyzed between two consecutive runs for the same well to ensure consistency between runs (Dhar et al., 2008).

3.8. Statistical methods

A multivariate analysis was carried out to find correlations between the geochemical and microbial data. Two different kinds of multivariate analyses were conducted to narrow down our search for microbes that correlated with the key chemical parameters of pH, Eh, arsenic and Fe.

3.8.1. O2-PLS method (Orthogonal Projection to Latent Structure)

The orthogonal projection to latent structure (O2-PLS) analysis is a two-block (X-Y) latent variable regression (LVR) method with an integral orthogonal signal correction (OSC) filter (Trygg & Wold, 2002, 2003). This method exploits the statistical process of scaling the orthogonal components and joint components to find the correlation between geochemistry and molecular data as depicted in Fig 3-14. Procrustes analysis was conducted with the assumption that there could many different factors influencing the pattern of geochemistry, and the microbial action could be just one of them but the most significant one. Since the microbial data was very sparse (~60% of it are 0s), the variables in the microbial data irrelevant to geochemistry were removed. The O2-PLS links two data blocks together by applying an orthogonal signal correction (OSC) to each of the data block to remove variance orthogonal (independent) to the other block and then build a PLS model on what remained.

82

Figure 3-14. The O2-PLS method used for the multivariate analysis of geochemical and molecular data.

3.8.2. Spearman’s rank correlation

Spearman’s rank correlations between geochemical and microbial data were analysed using SciPy, a python based ecosystem, with a critical value of < 0.01 of Spearman’s rank correlation coefficient. The geochemical parameters were taken as independent variables and the 16S rRNA marker gene based microbial species were taken as dependent variables. The geochemical parameters that correlated with the microbial data were tabulated and plotted. Likewise, arsenic was correlated against the other geochemical parameters to study the possible geochemical factors influencing the arsenic release.

83 Chapter 4 : Paper I - The microbial community structure and arsenic biogeochemistry in two arsenic impacted aquifers in Bangladesh

Edwin T Gnanaprakasamb, Jonathan R Lloydb, Christopher Boothmanb, Kazi Matin Ahmedc, Imtiaz Choudhuryc, Benjamin C. Bostickd, Alexander van Geend, Brian J. Maillouxa*

a. Environmental Science Department, Barnard College, NY, USA. b. School of Earth and Environmental Sciences and Williamson Research Centre for Molecular Environmental Science, The University of Manchester, Manchester, UK. c. Department of Geology, University of Dhaka, Dhaka, Bangladesh. d. Lamont-Doherty Earth Observatory, Columbia University, Palisades, NY, USA.

Abstract

Long-term exposure to trace levels of arsenic (As) in shallow groundwater used for drinking and irrigation puts millions of people at risk of chronic disease. Although microbial processes are implicated in mobilising arsenic from aquifer sediments into groundwater, the precise mechanism remains ambiguous. The goal of this work was to use, for the first time, a comprehensive suite of state-of-the-art molecular techniques in order to better constrain the relationship between indigenous microbial communities and the iron and arsenic mineral phases present in sediments at two well- characterised arsenic impacted aquifers in Bangladesh. At both sites, arsenate (As(V)) was the major species of As present in sediments at depths with low aqueous As concentrations, while most sediment As was arsenite (As(III)) at depths with elevated aqueous As concentrations. This is consistent with a role for the microbial As(V) reduction in mobilising arsenic. The 16S rRNA gene analysis indicated that the arsenic-rich sediments were colonised by diverse bacterial communities implicated in both dissimilatory Fe(III)- and As(V)- reduction, while the correlation analyses involved phylogenetic groups not

84 normally associated with As mobilisation. Findings suggest that direct As redox transformations are central to arsenic fate and transport, and that there is a residual reactive pool of both As(V) and Fe(III) in deeper sediments that could be released by microbial respiration in response to hydrologic perturbation, such as increased groundwater pumping that introduces reactive organic carbon to depth.

Keywords: Arsenic, XANES, XRF, geomicrobiology, groundwater, 16S rRNA gene, PCR, pyrosequencing, arrA gene

4.1. Introduction

Arsenic (As), is a major threat to the lives of millions of people whose primary water source for drinking and farming is constituted by arsenic contaminated aquifers in Bangladesh (Karim, 2000; Charlet and Polya, 2006). Although the World Health Organization (WHO) guideline value for arsenic in water is 10 µg L-1, the arsenic concentrations reported in Bangladesh aquifers range from <0.5 to 2,500 µg L-1 (BGS/MML, 1999; Smedley and Kinniburgh, 2002). Previous studies have also shown that arsenite (As(III)) accounts for 90% of the soluble arsenic concentrations in ground water, whereas arsenate (As(V)) constitutes only 10% (Zheng et al., 2005; Rowland et al., 2008), and in sediments, the ratio of As(V) to As(III) varies as a function of the mineralogical properties and microbial activities (Breit et al., 2001; Horneman et al., 2004; Zheng et al., 2005; Dhar et al., 2008). Thus, there is a consensus that arsenic is naturally released from sediments to water under microbially-induced reducing conditions within the aquifer sediments (Rowland et al., 2002; Islam et al., 2004; Mailloux et al., 2013). Furthermore, previous studies suggest that the biogeochemical cycling of iron and arsenic, specifically the reduction of mineral assemblages containing Fe(III) and sorbed arsenic(V), plays a critical role in the mobilisation of arsenic in aquifers (Cherry et al., 1979; Cummings et al., 2000; Nickson et al., 2000; Kocar et al., 2006; Rizoulis, Wafa M. Al Lawati, et al., 2014; Z. Hassan et al., 2015). Although

85 most of these studies have been based on ‘lab microcosm’ incubations, the precise mechanisms of arsenic release in situ remain poorly constrained. In order to understand this complex process in aquifers, it is necessary to correlate various factors including the quantity of arsenic both in the sediment and in water, the relative abundance of arsenic and the iron speciation and mineralogy in both phases, as well as the relevant microbial communities associated with arsenic hotspots. The aim of this multidisciplinary study is to address the paucity of data available that directly link microbial ecology to the in situ geochemistry of arsenic impacted aquifers, thereby identifying the dominant biogeochemical processes driving arsenic mobilisation.

Microbial arsenic reduction occurs through a variety of pathways. Soluble As(V) can be reduced directly to As(III) by microbes during the intracellular detoxification process or can be used to conserve energy for growth via dissimilatory As(V) reduction (Ji and Silver, 1992; Zobrist et al., 2000; Saltikov and Newman, 2003). The ‘AB’ gene cluster in bacteria, containing the ars and arr genes, is used for these detoxification and energy conserving processes, respectively. During detoxification, the intracellular reduction of As(V) mediated by the ArsC protein is a pre-requisite to the efficient export of As(III) from the cell (Rosen, 1999; Mukhopadhyay and Rosen, 2002a). In the case of the dissimilatory arsenic reduction, As(V) is used as the terminal electron acceptor under anoxic conditions, mediated by the terminal arsenate reductase, Arr, a molybdoprotein located in the periplasm of Gram-negative bacterial cells. Both of these processes produce As(III), which is usually the dominant form of aqueous arsenic in contaminated aquifers (Krafft and Macy, 1998; Afkar et al., 2003; Lloyd and Oremland, 2006). Although a wide range of organisms carry the arsenic resistance operon, including many that are not implicated in arsenic mobilisation, a narrow distribution of organisms can respire As(V) (Malasarn et al., 2004; Saltikov et al., 2005; Yamamura and Amachi, 2014). Laboratory incubations (or ‘microcosms’) using sediments supplied with the addition of C-13 labelled carbon sources have also suggested that the expression of arrA could be an

86 important factor in controlling the high concentrations of arsenic in aquifers (Lear et al., 2007; Héry et al., 2008; Dhar et al., 2011; Al Lawati et al., 2012; Rizoulis, Lawati, et al., 2014). However, only a few studies have examined the arsenic reduction by microorganisms in the aquifers with geogenic arsenic via the direct analysis of the field samples, which lack exogenous carbon sources and exhibit lower rates of metabolism (Gault et al., 2005; Desoeuvre et al., 2016).

The dissimilatory reduction of Fe(III) to Fe(II) can also be energetically favourable for specialist anaerobic microorganisms including the Geobacter species and can result in either the solubilisation of Fe(II) and/or transformations in the sediment Fe minerology (Islam et al., 2005; Herbel and Fendorf, 2006). Generally, Fe(III) minerals have more sorption sites for As(III) and As(V) compared to Fe(II)-bearing minerals. Under reducing conditions, where Fe(III) minerals are dissolved and Fe2+ is produced and Fe(II)-bearing minerals form, the total number of sorption sites for all iron minerals present is expected to be lower (and thus more As is expected to remain in the solution). These sorption sites also favour As(V) binding, which can result in As(III) dominating in the solution (Bostick and Fendorf, 2003; Dixit and Hering, 2003; Hohmann et al., 2011). Sediment extractions show that iron and arsenic are widely correlated in aquifers, indicating that Fe-(hydr)oxides play a critical role in controlling arsenic solubility (Berg et al., 2001). Extractions with phosphate that target weakly bound arsenic fractions have shown that this surface-bound fraction may be the critical pool of arsenic that governs arsenic mobility, and much of this arsenic may be associated with specific iron minerals. Recent improvements have focused on understanding the relationship between Fe and arsenic speciation using sequential extractions (Jung and Zheng, 2006; Huang et al., 2010), but interpreting such results is often difficult. Other techniques, such as X-ray absorption spectroscopy (XAS) are useful to study this relationship, but have been applied to only a few aquifer materials, and even fewer that are systematically related in time and space (Herbel and Fendorf, 2006; Datta et al., 2009; Ying et al., 2013). To date,

87 and to the best of our knowledge, only one study (Gault et al., 2005) has measured arsenic and Fe speciation in aquifer sediments alongside a preliminary characterisation of the extant sediment microorganisms. This information is needed to help determine the roles of the arsenic and Fe redox processes in controlling aquifer arsenic levels in situ, and given the improvements in the techniques available for geomicrobiological studies since this initial 2005 paper, the work described in this communication is timely.

Other redox processes could also affect aqueous arsenic levels. For example, As(III) and Fe(II) are re-oxidised in the presence of nitrate or oxygen, producing As(V)-sorbing particulate hydrous Fe(III)oxides and oxyhydroxides (Senn and Hemond, 2002; Omoregie et al., 2013). Similarly, the microbial sulphate reduction may influence arsenic solubility through the formation of insoluble arsenic sulphides, e.g. As2S3 (Rittle et al., 1995; O’Day et al., 2004; Omoregie et al., 2013).

In this study, we hypothesised that if arsenate reduction is the only process leading to arsenic release, the sediment will remain rich in As(V) and the groundwater will contain As(III) that is mobilised during reduction. If Fe(III) reduction alone is responsible for the release of arsenic, then once again the sediment will be dominated by As(V), with As(V) dominant in the groundwater. If both the reduction of As(V) and subsequent sorption to reactive Fe minerals are important for controlling arsenic mobilisation, then the sediment will accumulate As(III) until the reactive Fe(III) minerals are reduced, and the sorption sites for As(III) are removed. In this scenario, the sediment will become enriched in As(III). Key to differentiating these processes is the sampling mineralogy and aqueous levels in the environment at a sufficiently high resolution in order to observe gradients in As and Fe speciation, and correlating metal speciation with microbiological data. The objective of this cross-disciplinary work was, for the first time, to use a broad range of cutting edge molecular techniques in order to (i) determine the speciation of arsenic and iron with depth in the shallow aquifer sediments collected from Bangladesh across an arsenic gradient, (ii) analyse

88 the geochemistry of these two sites and (iii) compare arsenic speciation with the composition of the bacterial communities, including locating zones where arsenic-respiring bacteria are present, and identifying their potential role in controlling arsenic solubility. This study was conducted in three parts. First, we examined the hydrogeology of the sediments and water samples from two contrasting sites (Site F and Site B), followed by analyses of the composition of bacterial communities (and key functional genes that could impact on the As speciation). Finally we explored correlations between the microbial community compositions and the data obtained from the hydro-geological analyses of the sediment cores and relevant water samples. This study therefore illustrates the complex interplay between hydrology, mineralogy and microbiology that underpins arsenic release in aquifer sediments. This work is important, as a more complete understanding of these processes could contribute to more accurate predictions and inform mitigation procedures, for example linked to interventions that may stimulate or suppress key microbial processes that immobilise or mobilise arsenic into ground waters.

4.2. Materials and methods

4.2.1. Study site

The study sites are located in Araihazar Upazila, Bangladesh approximately 25 km east of the capital Dhaka as marked in Figure 4-1. The sites are located in Lashkardi village (Site F) (23.774 °N, 90.606 °E) and Baylarkandi village (Site B), (23.780N, 90.640E) . Tube wells were previously installed at these arsenic impacted sites in 2001 in order to monitor the temporal variability (Zheng et al., 2005; Dhar et al., 2008). Site F is a sandy site with sand extending to the surface. The recharge rate is 0.5 m y-1 and the maximum arsenic concentration of 200 g L-1 is reached at 26 m depth (Stute et al., 2007). Site F drilling was performed 5 m from the tube wells in a hollow that was 1.6 m lower than the tube wells. Site B is capped by approximately 6 m of the fine-grained silt and clay that overlies the sandy aquifer. The recharge rate is 0.08 m y-1 and the maximum arsenic concentration of 460 g L-1 is reached at 13.7 m depth (Stute et al., 2007). Assuming a porosity of 0.25 at each

89 site the vertical groundwater velocity is 2.0 m y-1 at site F and 0.32 m y-1 at site B. While the water samples were collected from the existing tube wells installed in 2001, the sediment samples were acquired from a fresh coring at the same sites adjacent to the existing wells.

Figure 4-1. The geolocation of the study sites with the arsenic concentration. Site B is located at Baylarkandi village, where as Site F is located near the community borewell at Lashkardi village. Both the sites are in Araihazar Upazilla in Bangladesh. The study areas are located in the low-lying (3–6 m elevation) Meghna River flood plain to the east of Dhaka. The colours in the figure shows the arsenic concentration distribution. (blue: 10g/L, green: 11–50 g/L and red: 51–100 g/L. 4.2.2. Sample collection Sediment cuttings and cores were collected using the ‘hand-flapper’ method, a manual drilling method often used to install wells (Horneman et al., 2004; van Geen et al., 2006). Cuttings were collected from the sediment as it exited the drill pipe. Intact sediment cores were collected at approximately 1.52 m (5 feet) intervals to 18.3 m (60 ft) at site B and 26.0 m (80 ft) at site F, by lowering a 1″ or 2″ ID gravity corer inside the drill pipe and into the sediment that had not yet been drilled. Upon coring, the sediment samples were split into aliquots for different types of analyses. The samples for the XAS analysis

90 were saturated with glycerol (approximately 1:1 v/v) as a preservative, and stored at -20o C prior to analysis. Glycerol is added to prevent oxidation by slowing the oxygen diffusion, and because it fixes bacteria in order to prevent microbiological alteration. In this case, spectra were collected using a thin (a few mM) thick film of wet, sediments without further treatment. The sample collection and synchrotron access were coordinated in order to minimise the lag time (about 2 weeks) between sampling and the analysis. The samples used for the XRF were used immediately in the field (without treatment) and the samples for the other analyses were stored in the refrigerator prior to analysis. From each depth, approximately 5 g of sediment was incised o aseptically in an anaerobic bag fluxed with N2 and stored anaerobically at -20 C for molecular investigation. The samples were hermetically sealed and transported to the respective laboratories for analysis purposes. The diffuse spectral reflectance of fresh cuttings was also measured in the field as a proxy for the Fe(II)/Fe(II+III) ratio in the acid-leachable Fe fraction (Horneman et al., 2004).

Groundwater samples were collected from 5 tube-wells from Site F and 3 tube-wells from Site B. A battery-driven submersible pump (Whale SuperPurger) was used in order to pump the water from the wells at a rate of about 2 L/min. The samples for arsenic, other trace elements and other cations were collected in 30 ml acid-cleaned HDPE bottles, as described previously (Zheng et al., 2005; Dhar et al., 2008).

4.2.3. Sediment chemistry

4.2.3.1. X-Ray fluorescence

The elemental composition of the sediment was determined by means of X-ray Fluorescence (XRF) spectroscopy using an InnovX Delta Premium hand-held XRF analyser using triplicate measurements and a collection time of approximately 75 seconds. This device calculates environmental concentrations based on total fluorescence at three energy levels in order to efficiently discriminate between elements, such as lead and arsenic with

91 similar emission lines. The analytical precision based on statistical counting errors is approximately ±1-2 mg kg-1 and ±300 mg kg-1 for arsenic and Fe, respectively. Under the same conditions, the analytical precision of the repeated measurements of uniform standard reference materials collected independently over days to weeks are within 5% of each other and are within 6% of the certified values. Since the XRF only samples a small quantity of the sample, the measured concentrations can vary more significantly when repeated on heterogeneous natural materials, and analytical accuracy under the experimental conditions is thus somewhat lower, about 10-20% for any single measurement.

4.2.3.2. EXAFS spectra for iron and arsenic

Approximately 0.1 g of sediment from each depth was analysed for Fe and arsenic EXAFS analysis was performed at the Stanford Synchrotron Radiation Laboratory (SSRL) on Beamlines 4-1 and 4-3, equipped with a 13- and 32-element Ge detector, respectively. The beamlines were configured with a Si(220) monochromator with a phi angle of 90 degrees. Soller slits were installed in order to minimise the effects of the scattered primary radiation. The beam was detuned at least 50% to reject higher-order harmonic frequencies and prevent detector saturation. Scans were calibrated to Fe K- edge of As(V) (11874.0 eV for sodium arsenate) and Fe metal (7112 eV) using a sample placed between the second and third ionisation chambers. A subset of each sample was sealed in Kapton tape and mounted between the first and second ionisation chambers, and sample spectra were obtained in fluorescence mode in combination with a 6 μx Ge filter (for the analysis of arsenic) or Mn filter (for the analysis of Fe).

Spectra were processed in SIXpack (Webb, 2005) unless mentioned otherwise. Arsenic K-edge XANES spectra were normalised and fitted using linear combination fitting (LCF) using orpiment (As2S3), arsenite (as adsorbed arsenite, As(III)), and arsenate (as adsorbed arsenate, As(V)) (Sun, Chillrud, et al., 2016). Other arsenic oxidation states were considered but were not required for any fits. Adsorbed references were created by adsorbing either

92 sodium arsenate or sodium arsenate solutions on 1 g/L ferrihydrite solutions to a solid concentration of 1000 mg arsenic/kg Fe concentration.

Fe mineralogy was determined using the LCF fitting of the EXAFS spectra with those of the reference spectra of the commonly occurring sediment minerals (hematite, goethite, ferrihydrite, magnetite, mackinawite, siderite, biotite, and hornblende) (Sun, Chillrud, et al., 2016). Averaged Fe K- edge EXAFS spectra were normalised using linear pre-edge and quadratic post-edge functions. Normalised spectra were converted to k3-weighted chi functions with a threshold energy (E0) of 7124 eV. The reference spectra to be included in linear combination fitting were selected based on the minerals known to be present or commonly present in sediments, and also included minerals often found using other methods (for example, by diffraction or magnetic susceptibility). The principal component analysis (PCA) was also applied on sample k3-weighted spectra in order to select reference spectra. Target transforms were used to compare significant components against a spectral library of Fe mineral references in order to ensure that only references that were deemed statistically viable were included. Only target transforms having SPOIL values < 6 were considered as potential references (Beauchemin et al., 2002; Strawn and Baker, 2009). In the end, all spectra were fit with a single set of reference spectra, although individual samples likely contain minerals not represented in this reference set. This approach was implemented because it quantifies Fe mineralogy in a manner that is internally consistent and representative of the oxidation state, but it means that the fractional mineral concentrations of some minerals are combined and represented with a smaller set of similar reference spectra that are not easily differentiated. For example, the Fe(II) and Fe(III) silicates all have sufficiently similar spectra, which make it hard to resolve them, and consequently fits involving several of these minerals concurrently are rendered unreliable. Thus, we limit fits to representative and common Fe(II) silicates such as biotite and Fe(II,III) silicates such as hornblende. As a result, care must be used in interpreting the mineralogical data as indicative of the relative concentration

93 of specific minerals, in particular for Fe silicates. A least-squares LCF was implemented on the k3-weighted EXAFS spectra over a k-range of 2 to ~13 Å−1, in order to determine the relative fraction of each of the 8 reference compounds that contribute to the sample spectra. Uncertainties reported by SIXpack include the Monte Carlo-based error propagation from fitting, spectral noise in the sample and reference spectra, and similarities between the reference spectra (Webb, 2005). The fractions and uncertainties were reported in the unit of % of total Fe, and if needed, converted to concentrations by multiplying bulk Fe concentrations and expressed in the unit of mg Fe per kg sediment, i.e., mg kg-1.

4.2.4. Aqueous chemistry

High-resolution inductively coupled plasma mass spectrometry (ICP- MS) was used to analyse soluble arsenic, Fe and other elements in well water (Cheng et al., 2004; Dhar et al., 2008). The accuracy and the detection limits (<0.1 μg/L) were verified against reference standards NIST1640A and NIST1643.

4.2.5. Microbial Ecology

4.2.5.1. 16S rRNA Based Bacterial Community Analysis

The culture-independent-community analysis was conducted using 16S rRNA gene sequencing protocols. The total DNA was isolated from the 1 gram samples of sediment using the PowerSoilTM DNA Isolation Kit (MoBio Laboratories, Inc., Carlsbad, CA, USA), following the manufacturer’s protocol. The quantity of DNA in the extracts was measured at 260nm using a Nanodrop ND-1000 spectrophotometer (Thermo Scientific). Pyrosequencing used Roche’s ‘FastStart High FidelityPCR System’ with the forward primer 27F (Lane, 1991) and the reverse primer 338R (Hamady et al., 2008) that target the V1-V2 hypervariable region of the 16S rRNA gene in bacteria. Sequencing was performed using a Roche 454 Life Sciences GS Junior System at the sequencing facility at the University of Manchester. The 454 pyrosequencing reads obtained in fasta format were analysed using QIIME version 1.8.0 using the

94 parameters of 97% sequence similarity in OTU picking and the 80% confidence threshold in taxonomic classification.

4.2.5.2. Functional Gene Analyses

In addition to the microbial community analysis using 16S rRNA gene primers, arsenate-respiring bacteria were targeted with a combination of functional gene primers that target the arrA gene, including AS1F/AS1R-AS2R (Song et al., 2009), HA ArrA-D1F/HA Arra-G2R (Kulp et al., 2006), ArrAUF1/ArrAUR3 (Fisher et al., 2008), ArrAfwd/ArrArev and CHArrAfwd/CHArrArev (Malasarn et al., 2004). The primer-set Dsr1F/Dsr4R was also used to amplify the 1.9 kb functional gene (dsr) coding for the dissimilatory sulphite reductase (DSR) conserved amongst all known sulphate- reducing prokaryotes (Wagner et al., 1998). The distribution of iron reducers was estimated using a PCR based amplification using 584F/840R primer-set specific for the 16S rRNA genes of Fe(III)-reducing bacteria of the family Geobacteraceae (Cummings et al., 2003). The primer details and the PCR conditions are provided in the supporting information-1 (Text 4-S1-1) as the PCR for Dissimilatory Arsenate Reducing Bacteria (arrA gene), the PCR for Dissimilatory Suphate Reducing Bacteria (dsr gene) and the PCR for the Geobactereraceae 16S rRNA gene.

4.2.6. Statistical Analyis:

Correlations were determined between the independent parameters of geochemistry and the dependent parameters of the percent species relative abundance using the Spearman’s rank correlation with a critical significance level of 0.01 in order to minimise the effect of the outliers. The number of significant correlations was summed in order to understand controls on species abundance and to relate microbiological abundance to the mineralogical and solution composition.

95 4.3. Results

Key data are presented in Figure 4-2 and Figure 4.-3 in order to provide a synopsis of the contrasting biogeochemistry of the two sample sites and their associated microbiology. Table 4-S1 and Table 4- S2 (Supporting Information- 1) show the comprehensive data from the various mineralogical and molecular investigations conducted on samples from Site F and Site B in this study.

4.3.1. Site F

4.3.1.1. Sediment chemistry

Site F is uniformly sand from an approximate 30 m below ground surface (bgs) to the surface and is characterised by a relatively rapid recharge rate of 0.5 m y-1. The arsenic concentration of the 15 sediment samples was 2-3 mg kg-1, with some values as low as 1 and as high as 6 mg kg-1 at a depth of 18.4 m (Fig. 1A). The variation in arsenic levels mirrored changes in solid-phase Fe and Mn. The Fe sediments were highest at the surface and at depth but were typically 10-20 g kg-1 and ranged from 9.3 g kg-1 to 35.3 g kg-1. Manganese concentrations followed a similar trend to Fe in relation to depth, with the lowest concentration of 174 mg kg-1 observed at 16.9 m bgs and the highest of 753 mg kg-1 observed at a depth of 26 m bgs.

Solid phase Fe and arsenic speciation was determined at 17 depths ranging from 3.1 m to 26 m bgs. At Site F, arsenic speciation changed smoothly as a function of depth. From the surface to 18.4 m bgs, the solid phase arsenic gradually decreased from almost entirely As(V) except to 58% As(V) at 12.3 m bgs. From 19.9 m to 26 m bgs the percentage of As(V) rapidly decreased, with appreciable As2S3 found only in this transitional zone, reaching 38% of the total arsenic at 19.9m bgs. Below 19.9 m bgs, the sediment was dominated by As(III) with less than 20% As(V) (Fig. 1B). The iron mineralogy of authigenic

96

Figure 4-2. Chemical and microbial ecology of sediments and water fr0m Site F aquifer. (A) Sediment arsenic and iron concentration, and reflectance curve (blue, red and green) along the sediment depths with the presence of dissimilatory arsenic respiring bacteria (brown) and iron-reducing members of the Geobacteraceae (lines); (B) As(V), As(III) and AS2S3 species in the sediment (XANES) at site F; (C) Arsenic X-ray absorption near edge structure (XANES) spectra along various depths of sediments. (D) Fraction of Fe phases in relation to the Fe-minerals along various depths. (E) As, Fe concentration (blue and green) in wells (Aqueous Phase) and ground water Tritium Age (red); (F) 16S rRNA gene analysis for microbial communities (bar diagram; percentage of each class of the Bacterial Kingdom).

97 Fe minerals also changed gradually as a function of depth. At the surface, most Fe was present as a combination of ferrihydrite and goethite, and similar quantities of Fe in silicates. Above 10 m, ferrihydrite represented as much as 50% of the total Fe and decreased with depth. Goethite consistently represented 20-30% of the total Fe at all depths. The Fe(II,III) silicates, modelled with biotite and horneblende, were widely variable but showed no trends with depth and together represented about half of the total Fe. By 30 m bgs, the abundance of Fe(III) decreased, particularly when only considering the abundance of non-silicate Fe species, and Fe(II)-bearing minerals were detected though they were still relatively minor species (Figure 4-2-D).

4.3.1.2. Aqueous chemistry

The aqueous chemistry of 5 well-samples from site F showed the presence of arsenic ranging from 0.5 µg L-1 at shallow depths to 203.24 µg L-1 at 26 m bgs, far in excess of the 10 µg L-1 WHO. The pH of the water samples ranged between 6.3 and 7.0, and the dissolved oxygen concentration remained below the levels of detection, providing suitable conditions for the neutrophilic anaerobic metal-reducing bacteria implicated in arsenic mobilisation. The soluble Fe concentrations paralleled these changes, increasing from 0.008 mg L-1 (0.15 µM) near the surface to 17 mg L-1 (300 µM) at 26 m bgs. There was a clear relationship between soluble arsenic and Fe concentrations in the wells of each depth (Figure 4-2-E). In contrast, the sulphur concentrations in the wells decreased with depth, from 2.1 mg L-1 (65 µM) at 6.2 m to 0.13 mg L-1 (4.2 µM) at 26 m.

4.3.1.3. Microbial ecology

4.3.1.3.1. 16S rRNA gene based community analysis

A total of 70,907 16S rRNA gene sequences were obtained for Site F by processing pyrosequencing data through the QIIME pipeline. The number of sequences obtained for each depth varied from 5,788 to 10,426, and the alpha

98 rarefaction analyses suggested that the number of distinct species detected at each depth ranged between 308 at 16.9 m and 729 at 10.8 m (Figure 4-2-F).

The sediment depths except 10.8 m and 16.9 m were dominated by an organism most closely related to Acinetobacter sp. TTH0-4 (99.0% identity) (Zhang et al., 2016). The Acinetobacter species are normally associated with oxic environments, being non fermentative, aerobic bacteria that decompose large organic molecules (Garrity et al., 2010), although the use of stored polyphosphate as an intracellular energy source can support metabolism in environments undergoing redox transitions (Zehnder and Van Groenestijn, 1990). Although the Acinetobacter species have been detected previously in arsenic contaminated sites in Assam (Ghosh and Sar, 2013) and the Hubei Province in China (Cai et al., 2009b), they have not been recognised as As(V) (or indeed Fe(III)) respiring bacteria, even if Acinetobacter sp. TTH0-4 is known to carry the genes coding for arsenic resistance proteins (ArsC and ArsH). Of the total community composition, a close relative of Acinetobacter sp. TTH0-4 comprised 21% of the bacterial population at a depth of 7.7 m, 11% at 13.8 m, 4% at 16.9 m, 11% at 19.9 m, 27% at 22.9 m and 11% at 26 m. Other dominant species found in almost all depths included organisms most closely related to known nitrate-reducing bacteria, such as Massilia brevitalea (98.4% identity) (Zul et al., 2008), Psychrobacter glacincola (97.4% identity) (Bowman et al., 1997), and Arthrobacter defluvii (100% identity) (Kim et al., 2008). The close relative of Massilia brevitalea, was the most abundant species comprising 20% of the bacterial population at a depth of 10.8 m, whereas Psychrobacter glacincola LMG21273 dominated the sediment, constituting 41% of the bacterial population at 13.8 m. A close relative of Massilia aurea AP13 (100% identity), a Gram-negative betaproteobacterium capable of starch degradation (Gallego et al., 2006) was found in all depths except 7.7 m. The highest concentration of sediment arsenic in the present study was at 16.9 m and 26 m, where a close relative of the Massilia aurea AP13 formed 28% and 24% respectively of the total bacterial population.

99 The 16S rRNA gene sequences were very closely aligned with Arthrobacter humicola (100% identity), a Gram-positive Actinobacterium, (Kageyama et al., 2008) ranging from a high abundance of 18% of the total microbial community at a depth of 10.8 m to a low abundance of 1.8% at 26 m. Comamonas aquatica (99% identity), a Gram-negative betaproteobacterium (Wauters et al., 2003) was found only in a relatively high arsenic concentrated sediment at 16.9 m, constituting 12% of the bacterial population in this case. While certain species of Arthrobacter are known for nitrate reduction, the Comamonas species are known for the oxidative carbohydrate metabolism with oxygen as an electron acceptor. A more detailed account of the bacterial representatives in each sample, their relative abundance and identity are presented in the Supporting Information-2 as Appendix-1 (Bangladesh Site F – Microbial Community Summary), and are consistent with a broad range of processes including both aerobic and anaerobic metabolism.

The blast search for OTUs representing less than 1% of the microbial community indicated the presence of a diverse range of bacteria potentially colonising the sediments using many contrasting forms of metabolic activity. Of the various representatives of microbes, the key bacteria potentially involved in the metabolism of iron, arsenic and nitrate are discussed below.

4.3.1.3.2. Iron-reducing bacteria

Under anoxic conditions, the reduction of Fe(III) has been implicated in the mobilisation of arsenic in aquifer sediments, either through the reductive dissolution of As-bearing Fe(III) oxides, and/or supporting the growth of organisms able to respire Fe(III) and also As(V) to the more mobile As(III) (Islam et al., 2004; Kocar et al., 2006). At least four relatives to known Fe(III)-reducing bacteria were found at all sediment depths in Site F in a low abundance (<1%). Rhodobacter capsulatus (98% identity), a dissimilatory Fe(III)-reducing bacterium (Dobbin et al., 1996) was detected at depths of 7.7m, 13.8m and 16.9m with a composition of below 0.2% of the total population for each depth. Clostridium butyricum EG3 (98% identity), another organism associated with dissimilatory Fe(III) reduction (Park et al., 2001), was

100 also found at depths of 19.9 m (0.01%), 22.9 m (0.04%) and 26 m (0.02%). Other bacteria well known to respire Fe(III) were also present in a low abundance across various depths, including close relatives to the Geobacter spp., which are also known to respire As(V) (Héry et al., 2008). Given the potential link between the Geobacter species and the respiration of both Fe(III) and As(V) (Dobbin et al., 1996; Pederick et al., 2007b), PCR amplification using the 584F/840R primer-set specific for the 16S rRNA genes of the members of the Geobacteraceae family, was also used in this study. PCR products were obtained in 6 samples out of the 8 tested, confirming the widespread presence of these Fe(III)-reducing bacteria in the iron rich sediments (Figure S1-1-B, Supporting Information-1).

4.3.1.3.3. Arsenic-metabolising bacteria

At least 8 bacterial species closely related to those known to metabolise arsenic were identified at an abundance below 1% across various depths in the aquifer sediments at Site F, and included respiratory arsenate reducers (carrying the arr gene) implicated in mobilising arsenic (Héry et al., 2008), detoxifying arsenate reducers (carrying ars genes) and close relatives to the aerobic arsenite oxidisers encoding the aio genes. For example, a close relative to Psudomonas putida str. WB (100% identity), a dissimilatory arsenic- respiring bacterium (carrying both arrA and arsC genes, (Freikowski et al., 2010)) was found at depths of 7.7, 16.9, and 26 m comprising less than 0.1% of the population. An organism most closely related to Geobacter lovleyi (93% identity), a dissimilatory Fe(III)-reducing bacterium, which is also implicated in respiring As(V) (Sung et al., 2006), was found at depths of 7.7, 10.8, and 16.9 m at a relative abundance of below 0.1%. A close relative to the arsenate- respiring bacterium Bacillus selenatarsenatis (95% identity) (Switzer Blum et al., 1998; Richey et al., 2009), was also found at depths of 7.7, 10.8, and 16.9 m, forming a community composition below 0.1%. A close relative to Hydregenophaga sp. CL3 (98% identity) (Rhine et al., 2007), a bacterium containing the aroAB genes responsible for oxidising As(III) was found at depths of 7.7, 10.8, and 16.9 m comprising less than 0.1% of the total bacterial

101 population, while a close relative of the Sinorhizobium sp. (99% identity), another As(III)-oxidising bacterium with the aioA gene was also found (Hamamura et al., 2013) at these same depths below 0.1%.

4.3.1.3.4. Nitrate/Nitrite-metabolising bacteria

Nitrate reduction is a widespread respiratory process that can also influence the fate of arsenic under anaerobic conditions, via the nitrite-driven oxidation of Fe(II) and As(III) (Senn and Hemond, 2002; Omoregie et al., 2013). As previously mentioned, close relatives to known nitrate-reducing bacteria including Massilia brevitalea, Psychrobacter glacincola, Arthrobacter defluvi and Pseudomonas antartica were found in high abundance (2% - 40%) across all depths, indicating the widespread potential for this form of metabolism in the sediments at Site F. In addition, nitrifying bacteria were detected, such as a close relative to Nitrospora sp. (100% identity; found in depths 7.7, 13.8, 16.9, and 22.9 m and comprising below 0.2% of the community structure), helping close the nitrogen redox cycle (Daims et al., 2015). Two OTUs very closely related to the Bradyrhizobium species (100%), a genus of nitrifying bacteria that are known to convert nitrogen into ammonium were found in low abundance at the same depths (Gussin et al., 1986).

4.3.1.3.5. Sulphate-reducing bacteria

Microbial sulphate reduction can be an important process in mitigating arsenic mobilisation via the precipitation of poorly soluble arsenic-bearing sulfide phases. Sulfate-reducing bacteria were detected at various depths at Site F, including examples affiliated with the Genera Desulfosporosinus (Focardi et al., 2010) at 10.8 m (0.05% relative abundance) and 16.9 m (0.03%), Desulfobacca (Loy et al., 2004) at 13.8 and 16.9 m (0.02% each) and the family Syntrophobacteraceae (Loy et al., 2004) at depths of 7.7 (0.1%) and 10.8 m (1.25%), respectively.

102 4.3.1.3.6. PCR confirmation of functional genes

To help further identify the sites of colonisation of the As(V)-reducing bacteria, several primer sets were also used to target the alpha subunit of the arrA gene required for the dissimilatory arsenate reduction. Of the 5 different sets of primers used, only the nested PCR using primers AS1F, AS1R and AS2F amplified the gene successfully from sediments and positive controls, in this case using a semi-nested PCR technique. The arrA gene was detected in the samples collected at depths of 7.7, 10.8, 16.9, and 26 m. Figure 4-S1-1-A in the Supporting Information-1 accounts for the sample depths, where the arrA genes were amplified. The dsr gene found in dissimilatory sulphate-reducing bacteria, which can precipitate arsenic as sulfide phases (Wagner et al., 1998; Omoregie et al., 2013), was also amplified successfully using the primer sets (Dsr1F/Dsr4R) from sediment depths of 7.7 and 10.8 m. (Figure 4-S1-1-C, Supporting Information-1).

4.3.2. Site B

4.3.2.1. Sediment chemistry

Site B is capped with a fine grained silt and clay layer for the first 6 m and has a slower recharge rate of 0.08 m y-1. The arsenic concentration in 8 sediment samples varied from 1 mg kg-1 at 7.6m to 6 mg kg-1 at 16.8 m. From 11 to 14 m, the arsenic concentration remained stable at 3 mg kg-1. The Fe concentration in the sediment was highest at 7.6 m, reaching 30.2 g kg-1, while the lowest concentration of 11.2 g kg-1 was recorded at a depth of 11 m. The concentration of Mn ranged from 178 mg kg-1 at 11 m to 612 mg kg-1 at 7.6 m, correlating well with the concentrations of Fe measured at the same depths (Figure 4-3-A).

103

Figure 4-3. Chemical and microbial ecology analysis of sediments and water from Site B aquifer. A) Sediment arsenic and iron concentration, and reflectance curve (blue, red and green) along the sediment depths with the presence of dissimilatory arsenic-reducing bacteria (brown) and iron-reducing members of the Geobacteraceae (lines); (B) As(V), As(III) and AS2S3 species in the sediment (XANES) at site B; (C) Arsenic X-ray absorption near edge structure (XANES) spectra along various depths of sediments. (D) Fraction of Fe phases in relation to the Fe-minerals along various depths. (E) As, Fe concentration (blue and green) in wells (Aqueous Phase) and ground water Tritium Age (red); (F) 16S rRNA gene analysis for microbial communities (bar diagram: percentage of each class of the Bacterial Kingdom).

104 Solid phase arsenic speciation was determined by XANES at 13 depths on cuttings and core samples from 1.5 m to 18.2 m below the ground surface (bgs). From 1.5 to 6 m bgs, the solid phase As was dominated by As(V). The percentage of As(V) decreased below 6 m bgs, and As2S3 reached a maximum in the core samples at 7.6 m bgs. Below 7.6 m bgs, the sediment was dominated by As(III) except at 12.2 and 15.5 m bgs, where As(V) was present (Figure 4-3-B). The mineralogy of authigenic Fe minerals showed some broad similarities with those from Site F, with ferrihydrite and goethite levels declining from 3 to 7.6 m bgs, indicating Fe(III) reduction, with siderite then detected at 10.97 and 13.7 m bgs. The Fe(II,III) silicates, modelled with biotite and hornblende, also remained stable throughout the depth profile (Figure 4- 3-D).

4.3.2.2. Aqueous chemistry

The aqueous chemistry was analysed in 3 water samples from the wells at Site B. At a well depth of 7.3 m (BW0-3), the soluble arsenic concentration was 24 µg L-1, rising to 290 µg L-1 and 460 µg L-1 at depths of 10.8 m (BW10-8) and 14.3 m (BW14-3) respectively, far in excess of the WHO guideline of 10 µg L-1. The Fe concentrations in water samples from the wells ranged from 15 mg kg-1 (270 µM) at a well depth of 7.3 m to 20 mg kg-1 (350 µM) at depths of 14.3 m, which correlated well with the Fe concentrations in the corresponding sediment samples, and was proportional to the arsenic concentrations detected in the water (Figure 4-3-E).

4.3.2.3. Microbial ecology

4.3.2.3.1. 16S rRNA community analysis

The total number of processed 16S ribosomal RNA gene sequences obtained from Site B by pyrosequencing was 55,759. The highest number of reads was obtained at a depth of 11 m (9,794 reads), and the lowest number was from a depth of 7.6 m (6,097 reads). The highest diversity of species was observed at a depth of 7.6 m with 1,215 distinct OTUs, whereas the lowest

105 number of distinct OTUs was observed at a depth of 7.6 m (cut) with 74 species (Figure 4-3-F).

At Site B, similarly to Site F, the Acinetobacter species were also relatively common, with a close relative to Acinetobacter sp. A32 (99.5% identity), another arsenic resistant gammaproteobacterium (Achour et al., 2007) found in all depths except at 7.6 m, with an abundance ranging from 3% at a depth of 13.7 m to 17% at 10.7 m. Uniquely at 16.8 m, both Acinetobacter sp. A32 (99.0% identity) and Brevundimonas sp. A21-66 (99.3% identity), another alphaproteobacterium reported to be found in arsenic contaminated soil and capable of reducing nitrate (Kavitha et al., 2009), were recorded correlating with the highest concentration of arsenic that we detected in the sediments from this study (6 mg kg-1). A close relative of Planococcus sp. A08 (99.7% identity), an arsenic-resistant Bacillus reported to grow in concentrations of up to 640 mM arsenate, and up to 14 mM arsenite, was detected at 10.7 m at a relative abundance of 1.4% (Achour et al., 2007).

Massilia species, known for nitrate reduction, were detected in relatively high abundance at four depths of 7.6, 10.7, 14, and 16.8 m. A close relative to Massilia sp. 51Ha (98.0% identity) (Bassas-Galia et al., 2012) was the most abundant bacterium at a depth of 7.6 m with an abundance of 9.4%. Massilia brevitalea (97.4% identity) was one of the most abundant species at 7.6 and 14 m with an abundance of 2.7 and 43%, respectively. These betaproteobacteria are known to denitrify, in common with the close relative to the gammaproteobacterium Psychrobacter glacincola (98.7% identity), which was found in abundance at depths of 7.6, 10.7, and 11 m with a composition of 3.0, 25 and 39%, respectively. A close relative of Herbaspirillum sp. P-64 (98.0% identity), a betaproteobacterium (Shrestha et al., 2007) was also found in all depths except 7.6 m, with an abundance of 1.0% at 7.6 m, 20% at 10.7 m, 4% at 11 m, 33% at 13.7 m, 19% at 14 m and 63% at 16.8 m. Some species of Herbaspirillum are known to encode the arsC detoxification arsenate reductase (Govarthanan, 2015) and also encode respiratory and assimilatory nitrate reductases (NAR and NAS) (Bonato et al., 2016). Also

106 found in these sediment samples were common soil bacteria including the Duganella, Chryseobacterium, Exiguobacterium, Arthrobacter, and Trichococcus species. A summary of different representatives of the bacterial communities and their relative abundance found in site B is presented in the Supporting Information-2 (Bangladesh Site B – Microbial Community Summary) as Appendix 1.

The blast search for OTUs detected at below 1% of the sequences revealed the presence of diverse bacteria related to different metabolic activities in the sediment at site B, again including those potentially involved in the metabolism of iron, arsenic and nitrate.

4.3.2.3.2. Iron-metabolising bacteria

Close relatives to known Fe(III)-reducing bacteria were also found at all sediment depths at Site B. In common with Site F, the Rhodobacter species were found at all depths except 10.7 m. The abundance varied from 0.02% at 7.6 m to 0.07% at 16.8 m. A close relative of the Fe(III)-reducer Pelobacter carbinolicus (98% identity) (Lovley et al., 1995), was also detected at a low relative abundance of below 1% at depths of 7.6 m and 14 m. A close relative to the Gallionella capsiferriformans strain ES-2 (99% identity) (Emerson et al., 2013), an Fe(II)-oxidising bacterium which potentially closed the Fe redox cycle, was detected at an abundance of 0.03-0.04% at depths of 7.6, 10.7, and 16.8 m, respectively. The presence of Geobacter related sequences was also noted, and Geobacteraceae specific 16S rRNA gene primers showed the presence of these organisms in 5 of the 7 Site B samples that were analysed (. Figure 4-S1-1-B, Supporting Information-1).

4.3.2.3.3. Arsenic-metabolising bacteria

A range of potential arsenic metabolising bacteria were identified, including an organism most closely related to Geobacter lovleyi (92% identity), the dissimilatory As(V) (and Fe(III))-respiring) bacterium that was detected at 7.6 m and 16.8 m with an abundance below 0.1% and noted in the key samples at Site F. A close relative of arrA and arsC carrying Pseudomonas putida (99%

107 identity) (Fernández et al., 2016) was also detected in samples from the same depths of 7.6 and 16.8 m with an abundance below 1%. Arthrobacter aurescens (100% identity), a heterotrophic arsenate-reducing bacterium (Macur et al., 2004) was also found at all depths. Stenotrophomonas sp. MM7 (100% identity), an arsenite-oxidising bacterium (Bahar et al., 2012) that has the potential to detoxify As(V) via the Ars system, was also detected at depths of 14 and 16.8 m with an abundance below 0.1%. Hydrogenophaga sp. CL3 (100% identity), another ‘detoxifying’ arsenite-oxidising bacterium (Rhine et al., 2007) was also detected at depths of 10.7, 13.7 and 16.8 m with a low abundance of below 1%.

4.3.2.3.4. Nitrate/nitrite-metabolising bacteria

In common with Site F, there was evidence of the widespread colonisation of Site B sediments with nitrogen cycling bacteria. As previously noted, close relatives of Psychrobacter glacincola (97.8% identity) and Massilia brevitalea (97.4% identity), two nitrate-reducing bacteria were found across various depths in high abundance of up to 50%, indicating an active metabolism related to nitrate reduction (Bowman et al., 1997; Zul et al., 2008). Rhodobacter spp (99% identity), as previously mentioned, also implicated in Fe(III)-reduction but well known to respire nitrate, were also commonly found at a low abundance of below 1% at depths of 7.6 and 13.7 m. A close relative to nitrite-oxidising Bradyrhizobium sp. (99%) (Gussin et al., 1986) was detected at below 1% of the community at 7.6, 10.7, 14 m and 16.8 m. A close relative to Nitrospira moscoviensis (99% identity), another nitrite-oxidising bacterium (Ehrich et al., 1995), was also detected at these depths at an abundance of below 1% of the bacterial community.

4.3.2.3.5. Sulphate reducing bacteria

Sulphate-reducing bacteria were also detected in low abundance at Site B, typically below 1% of the total bacterial population and affiliated with the deltaproteobacteria and Nitrospira. Representatives of the families implicated in sulfate reduction, including Desulfarculaceae and Desulfobulbaceae, were

108 found at depths of 7.6 and 16.8 m (0.03% and 0.2%, respectively), while members of the Desulfovibrionaceae were found at 7.6, 13.7, and 16.8 m (0.03, 0.06, and 0.03%), and the Desulfuromonadaceae at 7.6 m (0.1%) (Meyer and Kuevert, 2007). Sulphate-reducing bacteria from the Syntrophobacteraceae (Loy et al., 2004) were found in very low abundance of 1.4, 0.1 and 0.2%, respectively at 7.6, 13.7, and 16.8 m.

4.3.2.3.6. PCR confirmation of functional genes

The primer sets (AS1F/AS2F-AS1R) targeting the arrA gene required for dissimilatory arsenic-respiring bacteria, amplified the target gene successfully in sediment samples collected at depths of 7.6 and 16.8 m, where the arsenic concentrations in the sediment were 5 and 6 mg/kg, respectively. In all other depths, the arsenic respiring genes were not amplified. . Figure 4-S1-1-A in the Supporting Information-1 shows the sample depths where the arrA genes were amplified. Finally, the primer sets (Dsr1F/Dsr4R) targeting the dsr gene conserved in dissimilatory sulphate-reducing bacteria amplified the target gene in sediment depths of 7.6, 13.7, and 16.8 m. (Figure 4-S1-1-C, Supporting Information-1).

4.3.2.4. Correlation between the microbial and mineral interphase

Given the complexity of the microbial communities detected by next generation sequencing, and the comprehensive geochemical and mineralogical data sets collected from our two field sites, Spearman’s rank correlations were used in order to identify key phylogenetic groups that were associated to the changes in mineralogy and arsenic solubility at our field site. A synopsis of these results (Figure 4-4) shows the top correlations at Site F and B between relative abundance and As(III), Fe mineralogy and the concentrations of aqueous arsenic. A more detailed analysis is included in the Supporting Information-1 including Table 4-S1-3 and the correlation figures from Figure 4-S1-2 to from Figure 4-S1-7.

109

Figure 4-4. Top Spearman rank correlation (p=0.01) plots relating microbial communities to mineral species. (A,D&G) Arsenic (III) correlation with microbial communities at site B, Site F and combined sites (B,E&H) Fe (III) correlation with microbial communities at site B, Site F and combined sites ; (C,F&I) Aqueous As Correlation with microbial communities.

At site F, 133 out of 502 (26%) species were correlated with a p<0.01 to a mineralogical or redox sensitive parameter for a total of 228 correlations, with 8 species with 2 correlations and 25 species with 3 correlations. Of these, 85% of the correlations were to a solid phase, 5% to a liquid phase parameter, and 9% to both. In the solid phase, most correlations were to As(III), As(V), and sediment reflectance, with fewer correlations to Fe minerals (Table 4-S1-3-A and Figure 4-S1-2).

At site B, 129 out of 561 (23%) of the species were correlated with a p<0.01 to a mineralogical or redox sensitive parameter for a total of 189 correlations with 55 species with 2 correlations and 14 species with 3 correlations. A total of 85% of the correlations were to a solid phase, 13 % to a liquid phase parameter, and 2% to both. In the solid phase, most correlations

110 were the biotite, ferrihydrite, and goethite with fewer correlations to the arsenic phases (Table 4-S1-3-A and 4-S1-4).

When the sites were combined and only the species present at both sites were considered, 73 out of 274 (26%) of the species were correlated with a p<0.01 to a mineralogical or redox sensitive parameter for a total of 139 correlations with 33 species with 2 correlations and 3 species with 3 correlations. Forty-five percent of the correlations were to a solid phase, 18% to a liquid phase parameter, and 37% to both. Combining both sites increased the percentage of aqueous phase correlations in comparison to the solid phase. The highest number of correlations were to aqueous Fe, As(V) (solid phase), and aqueous arsenic. Figure 4-4 presents the top 9 correlation plots that depict the microbial community correlations with key mineralogical and geochemical species. Of particular interest is the correlation of As(III) and the Gram-positive bacteria affiliated with the order Clostridiales, which normally associated with the anaerobic degradation of organic matter, but were not previously noted to be able to respire As(V). It may be that their role in arsenic release is due to their ability to degrade complex organic matter, which drives the microbial As(V) reduction by more specialised metal-reducing bacteria, but a more direct involvement warrants further investigation. Focusing on data from Site F (Figure 4-4), similar hypotheses could be tested for close affiliates of the Bacteriodes species, correlated with high arsenic concentrations. Thus, these correlations might suggest previously unidentified physiological traits, and this concept could be extended to the identification of potential Fe(III)-reducing bacteria affiliated with the genera Sphingomonas and Rhizobium, that were correlated with ferrihydrite and goethite, respectively.

4.4. Discussion

Two shallow, high arsenic aquifer sites were analysed in order to determine sediment phase arsenic speciation and the corresponding microbial assemblages with depth across an arsenic gradient. The depth gradients do not trace one groundwater flow path from the surface to depth but instead are

111 used as a proxy to represent the evolution of groundwater. The two sites were chosen for the contrasting hydrology and geology. Site F has sand to the surface with rapid recharge, irrigation pumping nearby and a relatively deep arsenic maximum at 18.4 m, while site B is capped by a thick clay, has slower recharge rates and has a shallower and higher arsenic maximum. At both sites the solid phase arsenic was dominated by As(V) near the surface where the groundwater is low in arsenic. At depths where the groundwater was elevated in arsenic, the sediment was predominantly As(III). This consistent trend across these sites suggests that this redox reaction plays a role in controlling the observed groundwater arsenic levels.

These results are broadly consistent with previous studies (Bostick and Fendorf, 2003; Dixit and Hering, 2003; Hohmann et al., 2011; Jung et al., 2012) which suggest that the number of sorption sites for arsenic will be diminished, where Fe(III) minerals are respired at depth and converted to sorbed Fe(II) and ferrous minerals, and the binding of any arsenic present will be skewed towards the accumulation of any residual As(V), rather than As(III), which would be expected to accumulate in the aqueous phase. The conversion of As(V) to As(III) at depth is clear from our XAS data at both sites, and the reduction of Fe(III) minerals with depth is particularly striking for Site B. At both sites, the solid phase arsenic was dominated by As(V) near the surface where the groundwater is low in arsenic. At depths where the groundwater was elevated in arsenic, the sediment was predominantly As(III). This consistent trend across these sites suggests this redox reaction plays a role in controlling the observed groundwater arsenic levels.

The microbial ecology of the two contrasting sediments was studied by means of high throughput 16S rRNA gene pyrosequencing, and functional gene probing via PCR. These approaches confirmed a rich diversity of organisms potentially able to catalyse a wide range of biogeochemical reactions that could impact on both arsenic speciation and solubility. 16S rRNA gene profiling revealed that known representatives of both dissimilatory Fe(III)- and As(V)-reducing bacteria were widespread throughout the

112 sediment profiles at both sites, including those affiliated with the family Geobacteraceae, implicated in the microcosm studies in mobilising As from other sedimentary settings (Lloyd, 2003; Islam et al., 2004; Osborne et al., 2015). This widespread distribution of organisms that could potentially respire As(V), producing the more mobile As(III) detected at depth, was further mirrored by the prevalence of the arrA detected throughout the profiles at both sites. Interestingly, there was also evidence for the presence of bacteria across the profiles, that given the right geochemical conditions, could attenuate arsenic mobilisation, including sulphate-reducing bacteria that are able to precipitate arsenic as As2S3 (Newman and Beveridge, 1997; Omoregie et al., 2013) although their detection did not always coincide with the presence of arsenic sulphides in the sediments. This was especially true at Site F, although at Site B (which had a lower recharge rate), the dsr gene was successfully amplified at two depths where As2S3 was detected by XANES. Even at these potential ‘hotspots’, they did not dominate the arsenic speciation. Nitrate- reduction can also be invoked in the control of arsenic speciation, being coupled to the oxidation and Fe(II) and As(III), resulting in the precipitation of As(V)-bearing Fe(III) oxides (Omoregie et al., 2013). Once more, nitrate reducers were present throughout the samples analysed, although it was clear that with depth As(III) dominated in the solid phase and the levels of soluble arsenic increased.

Overall, the extant organisms and the biogeochemical processes that they catalyse seem to be comparable for both study sites, despite their contrasting hydrologies. As(V) and Fe(III) are reduced at depth, most likely leading to their release to groundwater, and also in As(III) dominating in the solid phase. Competing biogeochemical processes, such as denitrification and sulphate reduction (that can capture arsenic), and natural attenuation via sorption, are clearly not capable of preventing the accumulation of concentrations of arsenic well above the WHO guideline. Although this study cannot unequivocally identify the causative organisms or identify the mechanisms of arsenic release, the Geobacter species known to reduce both

113 Fe(III) and As(V) and implicated in arsenic mobilisation in previous studies were detected throughout and could play a critical role in arsenic release. In addition, approximately 25% of the species were correlated to a solid phase arsenic species or a Fe mineral indicating that a significant fraction of the microbial community is associated with these critical phases. The Spearman rank correlations identified new phylogenetic groups that could be linked directly or indirectly with As(V) reduction and mobilisation, and identifying their potential role in such processes warrants further investigation. However, given the complex microbial communities identified, and the relatively low abundance of known metal-reducing bacteria, the organisms causing these problems are likely to constitute a relatively minor component of the microbial communities. The mobilisation of arsenic in carbon stimulated microcosms and pure culture lab experiments (Gault et al., 2005; Mailloux et al., 2013; Rizoulis, Lawati, et al., 2014) can be rapid (in the order of weeks). It is likely that under in situ conditions, these processes take longer to deliver the arsenic into the aqueous phase, consistent with higher arsenic waters that are several decades old (Mailloux et al., 2013). This would also be consistent with lower organic loadings of bioavailable organic material expected at depth, and the relatively low abundance of metal-reducing bacteria in the microbial communities detected. The precise nature of the organic matter fuelling metal reduction at depth clearly requires further investigation, as does the potential role of other potential electron donors, such as ammonium, in such processes (Francis et al., 2003).

Mechanisms at play at the two sites could include the release of As(V) during the reductive dissolution of As(V)-bearing Fe(III) oxides, followed by the reduction of the liberated As(V) by the periplasmic Arr system, and then the re-sorption of As(III) (with some remaining in the solution). Other mechanisms that could play a role include the direct solid phase reduction of As(V), noted in pure culture studies (Oremland, 2013), and the mobilisation of some of the resultant As(III). Although this is not consistent with the direct role of a periplasmic enzyme system (mediated by Arr), the extracellular

114 reduction of metals via humics, secreted flavins and nanowires protruding from the cell surface is becoming an accepted paradigm (Marsili et al., 2008) and electron transfer by means of mineral assemblages could also play a role. It should also be noted that the solid phase biomineral capture of both As(III) and As(V) is also well known from laboratory studies, e.g. to Fe(II) minerals (Zobrist et al., 2000; Islam et al., 2005; Coker et al., 2006) and could be invoked in situ. The microbial processes controlling arsenic fate in complex aquifer systems, such as those operating at these Bangladeshi field sites, clearly warrant further study and will benefit from the application of new techniques that will allow the identification of active microbes mediating arsenic release in situ, and the genes involved. This is an accepted limitation of the targeted DNA-focused approaches that we have used in this study, which do not exclusively target the active microbial community in situ. This could be addressed by encompassing transcriptomic profiling techniques, which is a goal for our future studies on these systems. Metagenomic analyses could also help identify the key metabolic processes that drive the system, including autotrophy and the heterotrophic metabolism of the electron donors, and the role of coupled and competing processes, such as Fe(III) reduction and methanogenesis, respectively.

115 Supporting Information Paper I - The microbial community structure and arsenic biogeochemistry in two arsenic impacted aquifers in Bangladesh

Text 4-S1-1. PCR Primers and Conditions

1. PCR for Dissimilatory Arsenate Reducing Microbes ( arrA gene)(Song et al., 2009)

AS1F 5’ – CGA AGT TCG TCC CGA THA CNT GG – 3’ AS1R 5’ – GGG GTG CGG TCY TTN ARY TC – 3` Initial Denaturation step at 94˚C for 5 minutes 94˚C for 30 Seconds (melting) 35 Cycles: 50˚C for 30 Seconds (annealing) 72˚C for 1 Minute (extension) Final Extension step at 72˚C for 5 minutes

Semi-Nested PCR: uses 1ul of PCR product from initial reaction I in a 50ul reaction. AS2F 5’ – GTC CCN ATB ASN TGG GAN RAR GCN MT – 3’ AS1R 5’ – GGG GTG CGG TCY TTN ARY TC – 3`

Initial Denaturation step at 94˚C for 2 minutes 94˚C for 30 Seconds (melting) 30 Cycles: 55˚C for 30 Seconds (annealing) 72˚C for 1 Minute (extension) Final Extension step at 72˚C for 5 minutes Shewanella ANA-3 as +ve control

2. PCR for Dissimilatory Sulphate Reducing bacteria (dsr)(Wagner et al., 1998) Dsr1F 5` - ACS CAC TGG AAG CAC G- 3` Dsr4R 5` - GTG TAG CAG TTA CCG CA- 3` Initial Denaturation step at 94 ˚C for 5 minutes 94˚C for 40 Seconds (melting) 20 Cycles: 65˚C (decreasing by 0.5˚C per cycle) for 40 seconds (annealing) (Touchdown) 72˚C for 2 minutes (extension)

94˚C for 40 Seconds (melting) 20 Cycles: 55˚C for 40 Seconds (annealing) 72˚C for 2 minutes (extension) Final Extension step at 72˚C for 5 minutes Desulfovibrio desulfuricans as +ve control

116

3. PCR for Geobactereraceae 16S rRNA gene(Cummings D.E., 2002) GEO564F 5` -AAG CGT TGT TCG GAW TTA T- 3` GEO840R 5` -GGC ACT GCA GGG GTC AAT A- 3` Initial Denaturation step at 94˚C for 4 minutes 94˚C for 30 Seconds (melting) 35 Cycles: 58˚C for 30 Seconds (annealing) 72˚C for 30 Seconds (extension) Final Extension step at 72˚C for 3 minutes Geobacter sulfurreducens as +ve control

Figure 4-S1-1. PCR based confirmation of functional genes. (A) Gel picture showing the amplification of arrA gene in the samples; (B) Gel picture showing the PCR based amplification of Geobacter specific 16S rRNA gene. (C) Gel picture showing the PCR based amplification of dsr gene.

117 Table 4-S1-1. Chemical and molecular ecology analysis of sediments and water from the Site F aquifer*

* ‘Blank cells’ in the table denote the absence of analysis for the respective sample, ‘√’ denotes the positive results based on PCR products of respective 16S rRNA gene, Geobacter specific 16S rRNA gene, and arrA gene for arsenate reducers and dsr gene for sulphate reducers; ‘x’ denotes the negative results or unamplified products . ** The standard error for As, Fe and Mn were ≤ 1, ≤ 99 and ≤ 12, respectively.

118 Table 4-S1-2. Chemical and molecular ecology analysis of sediments and water from the Site B aquifer*

* ‘Blank cells’ in the table denote the absence of analysis for the respective sample, ‘√’ denotes the positive results based on PCR products of respective 16S rRNA gene, Geobacter specific 16S rRNA gene, and arrA (arsenic reductase) and dsr (sulphate reductase) functional genes and ‘x’ denotes the negative results or unamplified products . ** The standard error for As, Fe and Mn were ≤ 1, ≤ 99 and ≤ 12, respectively.

119

Table 4-S1-3. Correlations between mineralogy and molecular data

Table S-4.3.A. Percent of microbes that correlate different geochemical factors in both sold and aqueous phases

(Spearman p<0.01) Site B Site B and F Site F (%) (%) (%) Table S-4.3.B. Number of microbial correlations in solid and aqueous phase Reflectance 5.6 0.4 2.6 AsIII (ppm) 6.8 4.3 5.8 Solid Liquid Both

AsV (ppm) 5 0.2 8 Site F 86 5 9

As2S3 (ppm) 0.2 0.9 0.7 Site B 85 4 1 Siderite (ppm) 3.2 0.4 2.2 Site B & F 60 22 18 Goethite (ppm) 0.4 5 2.9 Hematite (ppm) 2.4 0.5 2.9 Magnetite (ppm) 3.2 0.4 1.8 Table S-4.3.C. Number of microbial correlations to 1or more geochemical Mackinawite (ppm) 2.4 0.4 1.1 variables Biotite (ppm) 1.4 10.3 0.7 1 variable 2 Variables 3 Variables

Hornblende (ppm) 3.8 2.3 2.2 Solid 73 53 0 Site F Ferrihydrite (ppm) 1.4 5 0 Liquid 3 2 14 Aqueous As 3.2 0 3.3 Solid 81 6 25 Site B Aqueous Fe 3.2 0 7.7 Liquid 2 4 0 Aqueous Sulphur 0.2 0.9 4 Solid 38 19 3 Site B & F Aqueous Mn 3.2 0.9 1.1 Liquid 18 13 0

120

Figure 4-S1-3. Top 9 correlation of bacteria with arsenic at site F.

Figure 4-S1-3. Top 9 correlation of bacteria with arsenic at site F.

121

Figure 4-S1-4. Top 9 correlation of bacteria with Fe and As at site B.

Figure 4-S1-5. Top 9 correlation of bacteria with arsenic at site B.

122

Figure 4-S1-6. Top 9 correlation of bacteria with Fe and As at site F and B.

Figure 4-S1-7. Top 9 correlation of bacteria with Arsenic at site F and site B.

123 Chapter 5 : Paper II - Application of two block latent variable regression analysis (O2-PLS) of microbial and geochemical data to identify potential arsenic cycling microbes in Cambodian aquifers

Edwin T. Gnanaprakasama, Laura A. Richardsa, Yun Xub, Mauro Tutinoc, Royston Goodacreb, Bart E. van Dongena, David A. Polyaa, Jonathan R. Lloyda

a. School of Earth and Environmental Sciences and Williamson Research Centre for Molecular Environmental Science, University of Manchester, Manchester, UK. b. School of Chemistry and Manchester institute of Biotechnology, University of Manchester, Manchester, UK. c. School of Biological Sciences, Division of Musculoskeletal & Dermatological Sciences, University of Manchester, UK.

Abstract

Arsenic in aquifers poisons more than 100 million people in Asia alone, as aquifers remain a major source of water for drinking and irrigation. Previous studies have suggested a link between microbial processes and the mobilisation of arsenic in aquifers. As a result of the complex interaction between microbes and arsenic- bearing minerals, the relatively immobile arsenate [As(V)] is reduced to labile and more soluble arsenite [As(III)] in aquifers, resulting in elevated aqueous concentrations of arsenic. The complex microbial communities capable of multiple metabolic activities colonising these arsenic-impacted aquifers has made it challenging to identify the exact mechanism of arsenic mobilisation in the aquifers. To resolve this ambiguity, this study undertook a statistical analysis of microbial and geochemical data-sets derived from analysing 30 groundwater and surface water samples from two transects along dominant groundwater flow-paths in an arsenic- impacted aquifer in Cambodia. Initial analyses sought to identify correlations between key geochemical parameters including pH, Eh, As, Fe, S, N and dissolved organic carbon (DOC), followed by cross-correlations with 16S rRNA gene-based community data obtained from the MiSeq sequencing platform. In addition, the functional diversity of arsenate-respiring bacteria was explored by PCR amplification and sequencing of arsenate reductase gene (arrA). Geochemical and

124 16S rRNA data were used in a two block latent variable regression method, with an integral orthogonal signal correction filter (O2-PLS), to construct a statistical model that identified the most significant bacteria correlating with key geochemical data, including arsenic levels in the groundwaters. The results identified 62 microbes that have a significant correlation to arsenic in two Cambodian transects, with 68% most closely related to known to carry the ars operon. The potential impact of those organisms on arsenic biogeochemistry is discussed, alongside future research directions including culture-dependant and metagenomic studies that are required to clarify further the role of those organisms in controlling As speciation in environmental systems.

Key words: Arsenic, metagenome, 16S rRNA gene, arrA gene, biogeochemistry, O2- PLS method.

5.1. Introduction

The health of more than 100 million people is threatened by drinking or farming with arsenic-contaminated water from shallow aquifers. Surpassing the guideline limit of 0.13 μM (10 μg·L−1) set by the World Health Organisation ( WHO, 2011), these waters contain a high concentration of geogenic arsenic. The precise mechanism of arsenic mobilisation in aquifers remains unknown, although there is a partial consensus (Dixit and Hering, 2003) that the process of mobilisation is mediated by microbes in reducing conditions where arsenate-reducing bacteria reduce arsenate [As(V)] which sorbs strongly to Fe minerals, forming the more mobile arsenite [As(III)], via dissimilatory or detoxification processes (Krafft and Macy, 1998; Islam et al., 2004; Saltikov et al., 2005; Silver and Phung, 2005; Lloyd and Oremland, 2006). In the former process, dissimilatory arsenic-respiring microbes (DARM) respire arsenic in order to conserve energy. During this process, As(V) is reduced to As(III) in the periplasm of the Gram-negative bacterium cell by a dissimilatory arsenate reductase (Arr), encoded by the arr gene (Krafft and Macy, 1998; Afkar et al., 2003; Song et al., 2009). In the detoxification process, As(V) is reduced via ArsC (coded by the ars operon) in the cytoplasm ,and then pumped out of the cell as As(III) (Ji and Silver, 1992; Butcher et al., 2000). Other than arsC, the arsenic operon in the detox system contains arsR, arsD, arsA and arsB genes where

125 arsR and arsD genes control the expression of the operon and the arsA and arsB genes regulate the anion-transporting arsenite pump (Butcher et al., 2000; Ryan and Colleran, 2002; Andres and Bertin, 2016; Zhu et al., 2017). While the Arr system has been reported to reduce both absorbed and soluble As(V), the ArsC can reduce only soluble As(V) that reaches the cell (Zobrist et al., 2000; Macur et al., 2004). The microbial-mediated arsenic cycle is influenced by geochemical factors including pH and Eh, and the speciation of Fe, C, N and S (Andres and Bertin, 2016; Zhu et al., 2017). At circum-neutral pHs, when the environment is oxic, As(V) is the - dominant arsenic species in the form of H2AsO4 , whereas in anoxic conditions

As(III), in the form of As(OH)3 dominates (Masscheleyn et al., 1991). The pH also determines the bonding between arsenic and Fe. When the pH is neutral, the sorption of As(V) and As(III) on to Fe oxide is high compared to when the pH is acidic (pH 5-6), when the sorption of As(V) onto amorphous Fe oxide and goethite is higher than the sorption of As(III). At alkaline pH (pH 7-8) the bonding of As(III) to Fe oxides is greater than As(V) (Dixit and Hering, 2003). In addition, the microbial transformation of arsenic-bearing Fe minerals can also cause arsenic mobilisation through reductive dissolution. In this process, dissimilatory Fe(III)- reducing bacteria (DIRB) reduce Fe(III)-oxyhydroxides, resulting in the release of Fe(II) and As(V) from the ferric minerals (Cummings et al., 2000; Islam et al., 2004, 2005). Linked to this process, the delivery of organic matter (OM) will also influence the fate of the arsenic by supplying microbial communities with electron donors including organic acids and alcohols for the conservation of energy (Niggemyer et al., 2001; Lear et al., 2007; Handley et al., 2009). In addition, sulphate-reducing bacteria (SRB) transform Fe, sulphate and arsenic species leading to the formation of arsenic trisulphide As2S3 or arsenopyrite (FeAsS), which are poorly soluble under aquifer conditions (Newman and Beveridge, 1997). Likewise, sulphur-oxidising bacteria use arsenic-bound sulphur as the electron donor and reduce As(V) into As(III) (Hoeft et al., 2004; Hollibaugh et al., 2006). In anoxic environments, microbes are also capable of coupling As(III) oxidation with denitrification, thus immobilising arsenic through the formation of As(V), and also the formation of Fe(III) oxyhydroxides to which they may associate (Hoeft et al., 2007).

126 To explore the complex interplay between microbes in arsenic-impacted environments and the various geological and geochemical factors noted above, including pH, Eh, bioavailable OC, Fe, S, and N speciation (Andres and Bertin, 2016; Zhu et al., 2017), we selected study sites from two contrasting transects in an arsenic impacted aquifer in Cambodia. We initially analysed a suite of geological and geochemical parameters in these aquifers (Richards, Magnone, et al., 2017; Richards, Sültenfuß, et al., 2017), and then, using marker gene-based metagenomic analysis, including 16S rRNA gene and arsenic marker gene (arrA) analysis, we characterised the microbial communities found in groundwaters collected from the field sites. We used a two-way orthogonal projection of latent structure (O2-PLS) method (Trygg and Wold, 2002, 2003) to identify any significant correlations between the microbial species detected and the geochemical parameters in the samples. The O2-PLS method is relevant to this study because the microbial data generated through 16S rRNA gene sequencing has many insignificant residues (orthogonal structures) that are not relevant to the geochemical data. Filtering the orthogonal structures in a data-set improves the quality of the data. Furthermore, O2-PLS separately models the structured noise in the data-sets, meaning no rotation of the regression coefficients is needed, which allows for the prediction of the best correlations between the data-sets (Trygg and Wold, 2003). The present study is the first of its kind to employ the O2-PLS model to identify correlations between the microbes and geochemical parameters by eliminating the orthogonal factors that least influence the two data-sets. Thus, the objectives of this study were, (i) to investigate the microbial communities colonising high arsenic-impacted aquifers based on high-throughput 16S rRNA gene analyses using the Illumina MiSeq platform; (ii) to identify the key arsenic-reducing bacteria possessing the genes responsible for arsenic reduction based on arrA gene sequencing; (iii) to determine the correlation between the microbial communities and key geochemical parameters including arsenic, sulphur, sulphate, ammonium, nitrate, pH and Eh, using conventional principal component analysis (PCA) that includes residual factors and a more advanced O2-PLS model that uses a two block (X-Y) latent variable regression method with an integral orthogonal signal correction. The ultimate aim of this work is to obtain a clearer

127 picture of the microbes controlling arsenic speciation in SE Asian aquifers, and develop further the tools available to explore the complex interplay between microbial communities and biogeochemical processes in environmental systems.

5.2. Materials and methods

5.2.1. Study site

The samples were acquired from two arsenic-impacted transects along dominant groundwater flow-paths perpendicular to the Mekong and Bassoc rivers in the Kien Svay district of northern Kandal Province, Cambodia an area identified by various studies as having high groundwater arsenic concentrations (Polya et al., 2005; Pederick et al., 2007a; Polizzotto et al., 2008; Rowland et al., 2008; Héry et al., 2014; Rizoulis, Wafa M. Al Lawati, et al., 2014; Richards, Magnone, et al., 2017). The two transects, hereafter referred to as T-Sand and T-Clay, were contrasting sand- dominated and clay-dominated transects respectively.

Figure 5-1. Sampling transects of T-Sand (along the Bassac river) (Sample sites LR- 01 through LR-09) and T-Clay (along the Mekong river) (Sample sites LR-10 through LR-14). Adapted from Richards et al., 2017 and Uhlemann et al., 2017.

The samples were collected from bore-wells, which had been installed along these transects in 2013-2014 using manual rotary drilling (Richards et al., 2015). The position of the well clusters, sampling depths and the sediment types along these

128 two transects are presented in the schematic diagram in Figure 5-S-1, in the supporting information. The surface water samples were collected from ponds heavily affected by seasonal wetlands.

5.2.2. Sample collection

Water samples (30) were collected from the two different transects (Richards, Magnone, et al., 2017). Samples reported in this manuscript were collected post- monsoon in November – December 2014 for both geochemical and molecular analysis. Water samples for DNA extraction were collected and filtered on site. At each depth, about 500 ml each of a groundwater and surface water sample was collected using methods described previously (Richards, Magnone, et al., 2017) and placed into a one litre Duran bottle (cleaned between samples using AnalaR grade methanol ≥ 99.8 %, VWR, UK) and filtered using sterile 0.22 µm pore size membrane filter (diameter 47 mm) using hand-pumped filtration devices attached to a glass filtration manifold (VWR, UK). Samples were filtered after collection as quickly as logistically possible; typically this was within an hour of sample collection but in some cases was several days after collection (with water samples stored refrigerated in the dark until filtration). Upon filtering the filters containing the biomass were transferred using sterilized forceps into sterile 20 ml conical centrifuge tubes, sealed with Parafilm, and then stored frozen in Cambodia until transported as non-temperature controlled airfreight to University of Manchester, where they were stored at approximately -20o C until further sample processing and analysis.

5.2.3. Geochemistry

Geochemical analyses were carried out in-situ and in the laboratory at the Manchester Analytical Geochemistry Unit (MAGU) at the University of Manchester using methods and principles described previously (Richards, Magnone, et al., 2017; Richards, Sültenfuß, et al., 2017). pH and Eh were recorded in-situ during sample collection using a multimeter and flow cells (Professional Plus Series Portable Multimeter,YSI). A field spectrophotometer (Spectroquant Nova 60A, Merck, Germany) was used for in-situ analyses of sulphide, ammonium, nitrite and nitrate.

129 The analysis of chemical elements including arsenic and iron were conducted on subsamples using an inductively coupled plasma atomic emission spectrophotometer (ICP-AES, Perkin-Elmer Optima 5300 dual view) and inductively coupled plasma mass spectrometry (ICP-MS, Agilent 7500cx). Dissolved organic carbon (DOC) was analysed on subsamples using a total organic carbon (TOC) analyser located at ALS Environmental, UK. Arsenic speciation was determined after separation in the field using ion-exchange cartridges following a method developed by (Watts et al., 2010) as detailed by Richards et al. (2017), .

5.2.4. Molecular (DNA) analysis

5.2.4.1. DNA extraction and library preparation for MiSeq

The DNA library preparation for the MiSeq sequencer was done following previously published protocols (Kozich et al., 2013). DNA was extracted from the filters containing biomass using the MoBio Power Water DNA Isolation Kit (MoBio Laboratories, Inc., Carlsbad, CA, USA). The extracted DNA was quantified using the Invitrogen™ Qubit™ 3.0 Fluorometer using the highly sensitive Invitrogen™ Qubit™ quantitation assays. The V4 region of the 16S rRNA gene was amplified using dual indexed primers (150x2) (Figure 5-S-2). The PCR amplicons were quantified again to see whether all the samples had amplified products required for library preparation. The normalisation, which involves three steps (binding, washing and elution), was performed using the SequalPrep™ Normalization Plate Kit (Thermo Fisher Scientific) in order to obtain an equal quantity of each DNA library. The normalised library was sized and quantified using high sensitivity DNA chips and a Bioanalyser 2100 (Agilent), and had an average length of 398 bp in each sample. A uniform concentration of 4 nM DNA library of each sample was achieved using the following formula:

Concentration in ng/μl 6 DNA concentratrion in nM = g ∗ 10 660 ∗ average library size mol

The pooled library was denatured along with PhiX before loading on to the illumina Miseq next generation sequencer. A MiSeq Reagent Kits v2 chemistry was

130 used for sequencing. A 95% success rate in sequencing run was achieved with a cluster density of 720 k/mm2, with 88% of clusters passing filter, 86% of reads having >Q30 value and a 5% aligned amount of PhiX. There were no spikes in a corrected intensity plot and all indices were identified following index reads. The total output data in fastq format obtained for 30 samples was 6.8 GB.

5.2.4.2. 16S rRNA gene analysis pipeline

The fastq files generated by the sequencer were analysed using a pipeline established to analyse microbial community composition using 16S rRNA marker genes. The pipeline contains tools including Cutadapt (Martin, 2009), sickle (Joshi and Fass, 2011), FastQC (Andrews, 2010), SPAdes (Bankevich et al., 2012), PANDAseq (Masella et al., 2012), Vsearch(Rognes et al., 2016) and QIIME (Caporaso, Kuczynski, et al., 2010b). The alpha diversity was analysed using an R statistical calculation vegan package in QIIME using the rarified data. Principal Coordinate Analysis (PCoA) was used to compare groups of samples based on phylogenetic or count- based distance metrics. The objectives of using each tool in the pipeline to analyse the data were given in Figure 5-S-2. in the supporting information.

5.2.4.3. The analysis of arrA gene

The dissimilatory arsenate reductase gene (arrA gene) was targeted and amplified using semi-nested PCR using the primers As1F-As1R and AS2F-AS1R as published previously (Song et al., 2009). Pyrosequencing PCR was performed using the Fast Start High Fidelity PCR system (Williamson et al., 2013) (Text 5-S-1). The amplified sequences were run on a 10% gel to separate the fragments before it was incised and gel-extracted. The pyrosequencing run was performed at the University of Manchester sequencing facility, using a Roche 454 Life Sciences GS Junior system. The 454 pyrosequencing reads were subjected to denoising and chimera removal, followed by analysis using Qiime 1.8.0 (Caporaso, Lauber, et al., 2010). Since there were only 22 nonreplicated sequences the reads were manually blast searched in NCBI nucleotide database with the help of Mega7 aligning tool.

131 5.2.5. Uni- and multi-variate analyses

5.2.5.1. O2-PLS method (Two-way Orthogonal Projection to Latent Structure)

Two-way orthogonal projection to latent structure (O2-PLS) analysis is a two-block (X-Y) latent variable regression (LVR) method with an integral orthogonal signal correction (OSC) filter (Trygg and Wold, 2002, 2003). We used this method to analyse our data-sets because the microbial data-set was far more complex than that of the geochemistry data. There were 1357 variables in the microbial data-set compared to only 11 variables in the geochemistry data-set. Therefore, it was essential to filter out irrelevant information in the microbial data- set to be able to find the best possible correlation between two data-sets. As given in Figure 5.2, in O2-PLS model each data

Figure 5.2. A schematic diagram of the orthogonal projection to latent structure (O2-PLS) model used to analyse the correlation between geochemistry and microbes in the aquifer system. block is decomposed into two parts: orthogonal components and joint components. The orthogonal components capture the variances which are unrelated (i.e. orthogonal) to the other data block and the joint components are then built to maximise the co-variance between the two data blocks after the orthogonal components been removed. This process can minimise the interference from unrelated variance in each of the data block and best correlation can thereby be

132 determined. The correlation between the two data-sets was measured by using Procrustes analysis on the scores of the joint components (Andrade et al., 2004). Procrustes analysis is a commonly used technique to measure the similarity between two multivariate data-sets. Considering two data-sets as two patterns in high dimensional spaces, Procrustes analysis optimally matches the two patterns through three ‘admissible’ transformations: translation, rotation and re-scaling. The goodness of fit was expressed as Procrustes error that varies from 0 to 1, an error of 0 means a perfect matching while 1 means nothing in common between the two data- sets. The O2-PLS and Procrustes analysis were conducted using Matlab 2016a (MathWorks, MA, U.S.A.) 5.2.5.2. Spearman’s rank correlations

O2-PLS method assesses the correlation between the microbial and geochemistry data in multivariate level while Spearman’s rank correlation finds the most correlated part between the two data-sets down to univariate level. The analysis was carried out using SciPy, a python-based ecosystem, with a critical value of < 0.01 of Spearman’s rank correlation coefficient. Geochemical parameters were taken as independent variables and 16S rRNA marker gene-based microbial species were taken as dependent variables. The geochemical parameters that correlated with the microbial data were tabulated and plotted. Similarly, arsenic was correlated against the other geochemical parameters to study the possible geochemical factors associated with arsenic release.

5.3. Results

5.3.1. Geochemistry

The major and trace elemental parameters of the sampled ground and surface waters are reported in detail by Richards, Magnone, et al., (2017). So only those relevant to the current study are summarized in Figure 4., The relevant geochemical data pertaining to the samples used for molecular analysis in this study are also presented in Table 5-S-1 and Table 5-S-2 in the supporting information. In brief, both groundwaters and surface waters were generally dominated by calcium, magnesium and bicarbonate. Most waters were reducing (Eh range: -153 – +152 mV;

133 median: -54 mV); circum-neutral (pH range: 6.6 – 8.4; median: 7.0) and groundwater arsenic, generally dominated by As(III), ranged from 0.00 to 4.0 µM. Groundwater arsenic is highly heterogeneous throughout the study area, with higher concentrations generally, but not exclusively, found in the deeper, sandy

10 - HCO3 1 + NH4 As 0.1 assumed phenol Fe

0.01 Mn - NO3 Range (mM) Range - 1E-3 NO2 3- PO4 1E-4 2- SO4 S

HCO - NH + As DOC* Fe Mn NO - NO - PO 3- SO 2- S 3 4 3 2 4 4 Figure 5-3. Box chart representing summary geochemical statistics on a logarithmic scale for the composition of Cambodian groundwater and surface water. Boxes represent the 25 % and 75 % range; the line within the box represents the median; the square represents the mean; the x represents the 1 and 99 % range; and the straight lines indicate the maximum and minimum. Data represents a subset of the entire dataset reported in (Richards, Magnone, et al., 2017) relevant to the sampling season, locations and geochemical parameters reported in this manuscript.

aquifers rather than in shallow, clay dominated areas. The profiles of sulphate, Fe and dissolved oxygen were consistent with arsenic heterogeneity, and indicate differing redox zones with depth and across the field area. High As and Fe in reducing conditions is consistent with arsenic mobilization via reductive dissolution of Fe(III) (hydr)oxides supported by previous studies (Islam et al., 2004; Kocar et al., 2008; Lawson et al., 2013, 2016). While noting the significant variation in geochemical characteristics, even within the study area (and particularly while comparing the sand-dominated sequences in T-Sand to more clay-dominated

134 sequences in T-Clay), the groundwater chemistry was broadly typical of most arsenic-bearing groundwaters in Southeast Asia (BGS, 1999; Berg et al., 2001; McArthur et al., 2001; Smedley and Kinniburgh, 2002; Polizzotto et al., 2008; Rowland et al., 2008; Buschmann and Berg, 2009; Lawson et al., 2013; Sovann and Polya, 2014). 5.3.2. Molecular ecology The DNA extracted and sequenced from 30 filtered groundwater samples had a total of 4,885,138 reads with mean and median values of 162,837 and 159,601, respectively. The highest number of reads was generated for the sample LR05-15,

Figure 5-4. Bar diagrams summarising the number of reads and OTUs obtained in each sample in 16S rRNA community analysis using MiSeq. Cyan bars represent number of OTUs and green bars denote number of reads. Sample names of the form LR-XX-XX refer to the sample site (see Figure 5-1) and the depth in metres (Figure 5-S-1.). MEK-prefixed samples refer to Mekong River water; RAIN to rain water; SW to surface water; D to duplicate.

135 with 229,899 reads, whereas LR08-30 had the lowest number with 109,401 reads. The surface water samples had a relatively lower number of reads compared to the groundwater samples as shown in Figure 5-4. These reads had 1359 distinctive operational taxonomic units (OTUs), representing diverse bacterial communities in different samples. The highest bacterial diversity was found in sample LR09-06 with 805 OTUs, whereas the sample LR05-30 had the lowest diversity, with 370 OTUs. While T-clay and T-sand had similar mean diversities of 638 and 602 OTUs respectively, the surface and rainwaters had a considerable lower mean diversity of 529 OTUs. The PCoA analysis of rarefied OTU-table based on the Euclidean distance matrix, grouped the samples with the lower variance as shown in Figure 5-5. 80% of the samples formed five distinctive clusters; two clusters of samples from the T- Sand samples, two clusters comprising of the surface water and rainwater samples, and a cluster from the T-Clay samples. The other 20% of samples remained as outliers.

Figure 5-5. Weighted principle coordinate analysis (PCoA) plot of betadiversity among the samples.

136 Phylogenetic analyses, based on the 16S rRNA marker gene, showed (Figure 5-6) the dominance of bacteria belonging to the Proteobacteria phylum that encompasses many Gram-negative mineral cycling bacterial species (Bergey, 2005). The T-Sand samples in general (except LR05 and LR07), had a pronounced dominance of Betaproteobacteria whereas the T-Clay and surface water samples were dominated by Gammaproteobacteria and Firmicutes. The Betaproteobacteria detected were most close related to species belonging to the families Gallionellaceae, Comamonadaceae, Methylophiloaceae and Oxalobacteraceae, while the Gammaproteobacteria were dominated by bacterial species belonging to the families including Moraxellaceae, Methylococaceae, Xanthomonadaceae, Pseudomonadaceae and Enterobacteraceae.

Figure 5-6. Percent abundance of bacteria belonging to different phyla and classes, and the micro molar concentration of arsenic. The bacteria are presented in the stack diagram, and the arsenic concentration in the respective samples are presented in the line diagram.

137 Along the T-sand transect, the samples (except LR05 and LR07) with the highest levels of arsenic (~ 5 µM to 9 µM) were dominated by Betaproteobacteria, but along the T-clay transect all samples (except LR13) had more variations amongst the species detected at relative high abundance. About 80% of the species in sample LR13-30 belonged to the Gammaproteobacteria. While the rainwater was distinct in its community structure, and dominated by Actinobacteria, Thaumarchaeota and Alphaproteobacteria, the surface water samples (except Mek-SW-N) were dominated by Cyanobacteria and Actinobacteria. The constituents of the microbial communities were also grouped based on metabolic activities using a microbial metabolic database, showing the presence of microbes with the potential to control the speciation of As, Fe, N, and S. Among the arsenic related metabolic bacteria, dissimilatory As(V) and Fe(III) reducing bacteria (DARB) were represented by Geobacter sp. (100% identity) with the low abundance of <0.009% in 13 samples. There were also an abundance of organisms most closely related to bacteria that can reduce soluble As(V) via the Ars detoxification system including including Bacillus magaterium (99% identity), Steotrophomonas matophilia (100% identity), Deinococcus actinosclerus (100% identity), Exiguobacterium sp. MH3 (98% identity), Deinocoocus ficus (99% identity), Pseudomonas p. DRA525 (100% identity) identity, Pseudomonas balearica (100% identity), Paenibacillus sp. JDR-2 (99% identity), Fictibacillus arsenicus (99% identity) and Bacillus cereus (100% identity). In total abundance in each sample (confirmed by NCBI database searches of these bacteria using nucleotide blast tool). Close relatives to bacteria noted to carry Ars genes constituted from 0.01% to 43.8% of the total number of sequences detected. Close relatives to known arsenite-oxidising bacteria including Deinoccocus spp., Pseudomonas spp., Thiobacillus spp. and Stenotrophomonas spp. constituted up to 32% of the total abundance in the samples.

Close relatives to known sulfide oxidising bacteria with >98% identity to Sideroxydans lithotrophicus (99% identity), Comamonas aquatic (98% identity), Comamonas tetosteroni (100% identity), Comomonas aquatica (98% identity) and Sulfuricella denitrificans (98% identity) represented up to 20% of the total abundance. Further NCBI database searches confirmed that these reference bacteria

138 encode soxAB genes responsible for sulfide oxidation. Close relatives of Fe(II)- oxidising families of bacteria including Gallionellaceae and Comamonadaceae were found ubiquitously in high abundance (e.g. ~32% in LR09-9) in many samples, whereas the Fe(III)-reducing (and As(V)-reducing) Geobacteriaceae were found in much lower abundance throughout the sample set. Finally, close relatives of bacteria including Bacillus spp. and Pseudomonas spp., with the potential to denitrify were detected in high abundance.

5.3.3. arrA marker gene analysis

PCR-based arrA marker gene analysis showed the presence of the arrA gene in all samples except the rain samples. The pyrosequencing-based arrA gene analysis of a representative sample, LR14-30 with 4.2 µM (314.7µg/L) arsenic, which was significantly high than the 10 µg/L WHO provisional guidevalue for drinking water, but towards the middle range of our groundwater samples As values, suggested the presence of arsenate-respiring bacteria which were not apparent in our 16S rRNA gene-based community analysis. The NCBI database search of the arrA DNA sequences retrieved showed a high similarity (80-85%) to those characterised previously in bacteria including Sulfuritalea hydrogenivorans (84% identity; Watanabe et al., 2017), Geobacter lovleyi (87%; Sung et al., 2006), Alkalilimnicola ehrilichii MLHE-1 (85%; Hoeft et al., 2007) and Geobacter uranireducens (85%; Lear et al., 2007). The higher similarity (99%) of sequences was found with the uncultured gene sequences coding for the alpha subunit of the arsenate respiratory reductase (ArrA). Blast searching the amino acid sequences translated from of the retrieved arrA gene showed the presence of diverse dissimilatory arsenate reductase from different uncultured organisms with the maximum identity of 99%. Further manual blasting of amino acid sequences showed a relative identity of arrA sequences similar to that of characterised microbes containing Arr protein. Table 5- S-4 in the supporting material presents the arrA gene sequences, matching the genes and organisms in the NCBI database.

139 5.3.4. Uni- and multi-variate analyses of geochemistry and microbes

5.3.4.1. Spearman’s rank correlations Spearman’s rank correlations were used to identify the number of microbes that correlated to the key geochemical parameters in the samples studied. A synopsis of these results (Figure 5-7) shows the number of microbial species (based on relative abundance) correlating to various geochemical parameters.

Figure 5-7. A bar diagram representing the total number of prokaryotes correlating to various geochemical parameters through Spearman’s rank correlation. In total, there were 1057 correlations in which 578 out of 1357 species, represented by distinctive OTUs, correlated with various geochemical parameters. Though the number of correlations was high in number, 87% remained below an r2 value of 0.30, explaining the weak correlations between the two sets of data. The maximum number of microbial correlations was seen with the Eh and NH4, data- sets, with 181 correlations comprising 13 % of the total correlations. Among the “major element” variables, As and Fe had the largest number of correlations, with 148 (11%) and 121 (9%) species respectively. Among the geochemical parameters, 2 2 total soluble arsenic was positively correlated with Fe (r = 0.44), NH4 (r = 0.11) and Eh (r2=0.23), whereas the total sulphur and sulphate had negative correlations to arsenic.

140 5.3.4.2. O2-PLS model

The O2-PLS model employed in this analysis uses 2 joint components and 6 orthogonal components, comprised of 2 from the geochemistry block and 4 from the microbial block, in order to establish the best correlation between the two data blocks. The number of joint components and orthogonal components were determined using a cross validation procedure as described in Bouhaddani et al.(El Bouhaddani et al., 2016). The Procrustes analysis performed on the joint components scores shows a high similarity (Procrustes error = 0.0907) in the pattern between the geochemical and microbial data-sets as shown in Figure 5-8.

Figure 5.8. Joint component scores plot showing the similarity between geochemical and microbial data. Geochemical data are shown in red and microbial data are in blue. T1 &T2 are the geochemical scores of the joint components 1 and 2 in the geochemical data block, whereas U1 &U2 are the microbial scores of the joint components 1 and 2 in the microbial data

The loadings plot of the geochemical data block, as shown in Figure 5-9-A, showed the correlations between the variables within the geochemical data. The

141 first latent variable W1 suggested that the concentration of Fe was closely correlated with that of As, whereas the factors including pH, Eh were inversely correlated with the As concentration, the second latent variable W2 showed that total sulphur, sulphate and nitrite were inversely correlated with the concentrations of Fe and As. In addition, comparing the joint loadings of the geochemical block with those of the microbial block (Figure 5-9-B) can help identifying the most correlated (or inversely correlated) variables in the microbial data block with the concentrations of As.

A B

Figure 5.9. The joint loadings of geochemistry and microbial blocks. W1 and W2 are the geochemical loadings reflecting the inter-correlation amongst the geochemical variables, whereas C1 and C2 are the microbial loadings, reflecting the inter- correlation amongst the microbial variables. The numbers in Figure 10.B are the ids of microbial variables most likely correlating with the arsenic concentration

The O2-PLS model had provided a ‘full picture’ matching between the geochemical data and the corresponding microbial data. However, considering As as the main factor of interest, it is also useful to focus on As alone to establish a predictive model between the microbial data and the concentration of As. Such model would be most useful to find out which variables in the microbial data are mostly associated with the concentrations of As. Therefore, a partial least square regression (PLSR) was employed. The importance of each variable in the microbial data was assessed by employing a recursive feature elimination (RFE) procedure (Guyon et al., 2002). Because the microbial data-set contains a large number of variables and comparatively much smaller number of samples, it is important to

142 take extra caution to minimise the risk of over-fitting. In this study, we employed a two-step validation procedure. Primarily, 10 representative samples were selected using Kennard-Stone algorithm (Kennard and Stone, 1969)as the independent test- set. The remaining 20 samples were used for building PLS-R model and RFE feature ranking. PLS-R-RFE was conducted in an iterative way. The 20 samples were randomly split into a training set and an inner test-set using bootstrapping resampling method (Efron, 1981). A PLS-R model was built on the training set and applied to the test-set to get an estimation of error in prediction in terms of root mean square error (RMSE). Then the variable with least contribution to the model (the one with least magnitude in regression coefficient) was removed from the data and another PLS-R model was built and tested on the test-set. This procedure was repeated until only one variable left in the data. A rank list about the importance of the variables to the model was then compiled based on the order of the variables

A B

Figure 5-10. Arsenic prediction using the PLS-R-RFE algorithm. (A) shows the prediction against known arsenic concentration from the inner test-set obtained by bootstrapping resampling and (B) shows those of the independent test-set. been removed from the data, assuming that the least important variables will be removed first (high rank value) and the most important variables will be removed last (low rank value). In the rank list, rank 1 is the most important variable while rank 1357 is the least important one. The PLS-R-RFE processing was also repeated 100 times and the 100 rank lists were averaged to produce the final variable rank list. The final PLS-R model was built on all the 20 samples using the variables with the averaged rank no greater than 50 (the PLS-R-RFE procedure had suggested that where there were less 50 variables left in the data, the performance of the model

143 began to deteriorate rapidly (data not shown)) and there were 47 variables (Table 5- S-1) matched this criterion. This model was then applied to the independent test- set. In Figure 5-10-A, it can be seen that there was a good agreement between known and predicted concentration of As, especially at high concentration levels (>1 μM). More importantly, in Figure 5-10-B there is a good agreement between the known and predicted on the independent test-set on most of the samples, only three predictions showed significant deviation from the known. This suggested that it might be indeed possible to predict the concentration of As based on the microbial data. Although we must emphasise that the number of samples available to this study is very limited and the model is strictly exploratory. Given the complex nature of microbial data and delicate interactions with geochemistry parameters, a much larger data is required to build a reliable predictive model for practical use.

The microbial data variables most closely correlated with the arsenic concentrations were extracted, in order to predict the arsenic concentration from the microbial community structure using partial least square regression (PLSR) with recursive feature elimination (RFE). The significant variables of 42 bacteria from the microbial data-set correlating with arsenic given in Table 5-S-3 were analysed further using an NCBI database search of OTUs representing these bacteria. Of the 62 microbes extracted as the significant variables correlating with arsenic, 42 microbes (69%) had ars (intracellular detoxification arsenate reductase) genes. Most organisms had the arsR gene, which regulates the expression of ars detoxification operon. 52% of the significant bacterial communities belonged to the phylum of Proteobacteria.

5.4. Discussion

Diverse microbial communities were observed in 30 ground water samples derived from two high arsenic impacted transects along the Bassac and Mekong rivers in Cambodia. The vast diversity of the microbes colonising these high-arsenic aquifers could be steered by various abiotic factors including pH, Eh, DOC, and the chemical composition of minerals (Ferguson and Gavis, 1972; Smedley and Kinniburgh, 2002). The geological and geochemical factors including arsenic of the groundwaters that host these microbial colonies are altered or maintained in turn

144 by these microbial colonies, creating a unique environment that helps the microbes evolve to have a specific genome over time (Colbourn et al., 1975; Zheng et al., 2005; Couture and Van Cappellen, 2011; Richards, Magnone, et al., 2017). In this study, we identify a statistically significant link between the microbes with arsenic related genes and the arsenic redox cycle in the anoxic aquifers where arsenic is mobilised.

As demonstrated in previous studies, the 16S rRNA gene-based community analysis showed relative high abundance of similar bacteria specific to these arsenic- impacted aquifers (Paul et al., 2015; Zahid Hassan et al., 2015; Xiao et al., 2016). Bacteria belonging to the Fe(II)-oxidising family of Gallionellaceae were found in high abundance (up to 49%) in the suboxic depths, indicating the possibility of Fe oxidation at these depths (Emerson et al., 2013; Omoregie et al., 2013). In the high arsenic concentration range (1.65 µM – 9.57 µM or 124 µg/L – 717 µg/L), the arsenic- reducing bacteria containing cytoplasmic arsenate reductase gene (ars) belonging to the families of Comamonadaceae (up to 48% relative abundance) and Moraxellaceae (up to 75% relative abundance) were recorded. While the widespread Moraxellaceae family was dominated by members of the genus Acinetobacter, known detoxify arsenic via reduction processes (Cai et al., 2009a). The family Comamonadaceae was represented by Hydrogenophaga species, known for arsenite oxidation (vanden Hoven and Santini, 2004), Aquabacterium species, previously reported in high arsenic groundwaters (Paul et al., 2015), and close relatives to Tepidimonas and Inhella. The other bacterial genera recorded at relatively high abundance in this study include Gallionella, known for Fe(II)) oxidation (Emerson et al., 2013), Methylophilus (Buse et al., 2014), Methylotenera (Paul et al., 2015), Methylococcus , Methylomonas (Cummings D.E., 2002), Methylocaldum , Zooglea , Dechloromonas , Azonexus, Sulfuritalea (Watanabe et al., 2017), Bacillus, Paenibacillus, Nitrospira (Daims et al., 2015) and Thermomonas known for nitrogen or methane metabolisms.

Of the bacteria detected at relatively high abundance, only Helicobacteraceae and Bacillaceae, associated with dissimilatory arsenate reduction were found in many samples. In sample LR14-30, Bacillaceae family constituted 46% the total abundance. Though certain species of bacteria including those belonging to Bacillaceae are capable of many other metabolism (Bergey, 2010). Other organisms

145 such as members of the families Geobacteraceae and Shewanellaceae known for dissimilatory arsenate reduction were not represented in the 16S rRNA based community analysis (Islam et al., 2005; Malasarn et al., 2008; Shelobolina et al., 2008). However, PCR-based arrA marker gene amplification did show the presence of the arrA gene in all samples except the rainwater samples. The pyrosequencing- based arrA gene analysis of sample LR14-30 suggested the presence of arsenic- respiring bacteria which were not apparent in the 16S rRNA gene-based community analysis. A higher similarity (99%) of sequences was found with the uncultured gene sequences coding for the alpha subunit of the enzyme arsenate respiratory reductase. The blast search of the amino acid sequences of the arrA gene found matches with putative dissimilatory arsenate reductases from different uncultured organisms with the maximum identity of 99%. These results show that the arsenate reduction carried out by these organisms could be vital in the mobilisation of arsenic in the aquifers. Known As(V)-respirers including Geobacter, Sulfuritalea and Alkalilimnicola species were implicated, but further culture dependent or metagenomic analyses are warranted to identify the organisms carrying arr genes in these samples.

The spearman rank correlation shows that, among the geochemical parameters, 2 2 2 arsenic had positive correlations to Fe (r = 0.44), NH4 (r = 0.01) and Eh (r =0.23), whereas the total sulphur and sulphate had negative correlations to arsenic. These results are in line with previous studies that have emphasised the role of geochemical factors including Eh, Fe and N in controlling the redox cycle and fate of arsenic in aquifers (Senn and Hemond, 2002; O’Day et al., 2004; Islam et al., 2005; Herbel and Fendorf, 2006; Kocar et al., 2006; Héry et al., 2010; Omoregie et al., 2013).

Overall these data would seem to be consistent with the presence of microbes living in a dynamic redox zone, and capable of facilitating reductive (e.g. anaerobic nitrate-reducing) and oxidising (microaerophilic Fe(II)-oxidising) processes. There was ample evidence for the presence of bacteria able to oxidise As(III) or reduce As(V), including the retrieval of functional genes associated with respiratory As(V) reduction (arrA). However, from 16S rRNA gene surveys, the planktonic cultures were more likely dominated by organisms able to detoxify the high levels of soluble

146 arsenic detected in these Cambodian groundwater samples, via the Ars system. This was emphasised by the application of a O2-PLS analyses, which identified 62 distinct bacteria that correlated with arsenic when orthogonal factors were eliminated, and 42 (69%) of the closest relatives to these are known to encode ars genes. Extending these analyses to the sediment-attached microbial communities within these aquifers, where specialist anaerobes mediate direct electron transfer to Fe(III)/As(V) minerals, is an obvious extension to this work, and may well identify a contrasting group of Fe(III)- and As(V)-respiring prokaryotes. There is also a clear need to supplement these studies with metagenomic sequencing, and where possible transcriptomic analyses, to unequivocally identify the metabolic processes controlling arsenic speciation in at risk aquifers.

Finally, the application of O2-PLS analyses may be useful for not only identifying novel organisms associated with key biogeochemical process, but also has clear potential to predict the physical/chemical environment in situ associated with microbial samples via community profiling. This would have implications for the study of a wide range of environmental processes, in addition to supporting efforts to understand better, and mitigate against arsenic mobilisation in aquifers worldwide.

147

Supporting Information Paper II - Application of two block latent variable regression analysis (O2- PLS) of metagenomic and geochemical data to identify potential arsenic cycling microbes in Cambodian aquifers

Figure 5-S-1. Sampling transects of T-Sand (along the Bassac river) and T-Clay (along the Mekong river). Sample names of the form LR-XX-XX refer to the sample site (see Figure 1) and the depth in metres (Figure 5-S-1.). MEK-prefixed samples refer to Mekong River water; RAIN to rain water; SW to surface water; D to duplicate.

Text 5-S-1. Primers and PCR conditions for arrA gene amplification (Song et al., 2009) Stage-I AS1F 5’ – CGA AGT TCG TCC CGA THA CNT GG – 3’ AS1R 5’ – GGG GTG CGG TCY TTN ARY TC – 3` Initial Denaturation step at 94˚C for 5 minutes 94˚C for 30 Seconds (melting) 35 Cycles: 50˚C for 30 Seconds (annealing) 72˚ for 1 Minute (extension) Final Extension step at 72˚C for 5 minutes Stage-II (Nested PCR: uses 1ul of PCR product from initial reaction I in a 50ul reaction) AS2F 5’ – GTC CCN ATB ASN TGG GAN RAR GCN MT – 3’ AS1R 5’ – GGG GTG CGG TCY TTN ARY TC – 3` Initial Denaturation step at 94˚C for 2 minutes 94˚C for 30 Seconds (melting) 30 Cycles: 55˚C for 30 Seconds (annealing) 72˚C for 1 Minute (extension) Final Extension step at 72˚C for 5 minutes

148

Figure 5-S-2. 16S rRNA gene based community analysis workflow.

149

Figure 5-S-3. A diagrammatic depiction of the primer structure used for the MiSeq indexing PCR.

Figure 5-S-4. Gel documentation of PCR based amplification of arrA gene. The second stage of nested PCR yielded 625 base pair amplified product. –ve denotes negative control whereas +ve denotes the positive control Geobacter uraniireducens. Coloured in yellow (LR14-30) is used for the sequence based analysis of arrA gene. The cream coloured samples with LR15 ids are samples not included in this study.

150 Table 5-S-1. The extract of relevant data of geochemistry previously published by Richards, Magnone, et al., (2017).

Sample ID As Fe S SO4 Mn NH4 NO3 NO2 DOC Eh pH (µM) (mM) (mM) (mM) (mM) (mM) (mM) (mM) (mM) (mV) LR01-15 0.90 0.06 0.04 0.00 0.02 0.01 0.02 0.00 0.027 -79 6.9 LR01-30 6.75 0.17 0.04 0.00 0.01 0.06 0.01 0.00 0.030 -121 7.0 LR01-45 9.35 0.04 0.03 0.00 0.02 0.04 0.11 1.36 0.024 -134 7.2 LR02-15 1.71 0.06 0.11 0.06 0.01 0.05 0.06 0.00 0.014 -119 6.9 LR02-30 7.76 0.10 0.04 0.00 0.01 0.06 0.06 0.00 0.039 -143 7.1 LR05-SW 0.07 0.00 0.03 0.00 0.00 0.01 0.02 0.23 0.095 141 7.2 LR05-15 3.18 0.53 0.04 0.00 0.03 0.02 0.02 0.01 0.072 -153 7.0 LR05-30 9.21 0.12 0.03 0.00 0.01 0.10 0.11 0.00 0.049 -150 7.2 LR07-15_1 0.20 0.01 0.05 0.00 0.01 0.21 0.02 0.00 0.039 -34 6.7 LR07-15_2 0.20 0.01 0.05 0.00 0.01 0.21 0.02 0.00 0.039 -34 6.7 LR07-30_1 7.72 0.21 0.04 0.00 0.00 0.17 0.01 0.00 0.061 -22 7.0 LR07-30_2 7.72 0.21 0.04 0.00 0.00 0.17 0.01 0.00 0.061 -22 7.0 LR08-30 1.65 0.01 0.04 0.00 0.00 0.20 0.02 0.91 0.059 -43 6.9 LR09-6 0.06 0.02 0.15 0.06 0.02 0.01 0.00 0.00 0.017 -48 6.9 LR09-9 2.41 0.17 0.09 0.01 0.04 1.16 0.03 1.57 0.060 -108 6.9 LR09-30 9.57 0.14 0.04 0.00 0.01 0.60 0.01 0.00 0.066 -146 7.3 LR10-6 1.11 0.00 0.59 0.44 0.00 0.03 0.10 3.41 0.015 -11 6.9 LR10-15 7.36 0.18 0.16 0.07 0.01 1.98 0.00 0.00 0.040 -76 7.1 LR10-30 5.99 0.05 0.06 0.00 0.02 1.31 0.02 0.00 0.098 -83 7.5 LR12-30 6.23 0.12 0.03 0.00 0.02 0.31 0.02 0.00 0.041 -93 7.1 LR13-30 6.06 0.09 0.04 0.00 0.00 2.28 0.01 0.00 0.124 -79 6.9 LR14-SW 0.19 0.00 0.04 0.01 0.01 0.00 0.02 0.00 0.086 31 6.7 LR14-6 0.02 0.00 1.94 1.74 0.02 0.01 0.02 0.45 0.023 59 7.2 LR14-15 0.29 0.06 0.48 0.33 0.01 3.99 0.04 4.24 0.132 -54 7.1 LR14-30 4.20 0.12 0.05 0.00 0.00 3.41 0.04 0.94 0.164 -17 6.6 MEK-SW_N 0.02 0.00 0.05 0.04 0.00 0.00 0.07 0.21 0.038 152 7.1 MEK-SW_D1 0.07 0.00 0.05 0.02 0.00 0.00 0.02 0.16 0.047 69 7.4 MEK-SW_D2 0.07 0.00 0.05 0.02 0.00 0.00 0.02 0.16 0.047 69 7.4 RAIN_1 0.10 0.00 0.05 0.02 0.00 0.00 0.09 0.00 0.017 59 8.4 RAIN_2 0.10 0.00 0.05 0.02 0.00 0.00 0.09 0.00 0.017 59 8.4

151 Table 5-S-2. The extract of relevant data of arsenic speciation previously published by Richards, Magnone, et al., (2017).

Total ∑As As Total Total Species As (III) As (V) DMA MMA Sample ID (μg/L) As (μM) As (μM) (uM) (%) (%) (%) (%) LR01-15 67.42 0.90 0.90 0.80 0.97 0.00 0.02 0.01 LR01-30 505.72 6.75 6.70 6.40 0.96 0.00 0.02 0.01 LR01-45 700.52 9.35 9.30 8.30 0.93 0.03 0.02 0.02 LR02-15 128.11 1.71 1.70 1.60 0.93 0.03 0.02 0.02 LR02-30 581.39 7.76 7.80 6.90 0.95 0.02 0.02 0.01 LR05-SW 5.24 0.07 LR05-15 238.25 3.18 3.20 2.70 0.98 0.00 0.02 0.00 LR05-30 690.03 9.21 9.20 7.70 0.95 0.02 0.02 0.01 LR07-15_1 14.98 0.20 0.20 1.10 0.88 0.07 0.02 0.02 LR07-15_2 14.98 0.20 0.20 1.10 0.88 0.07 0.02 0.02 LR07-30_1 578.39 7.72 7.70 6.30 0.94 0.01 0.02 0.02 LR07-30_2 578.39 7.72 7.70 6.30 0.94 0.01 0.02 0.02 LR08-30 123.62 1.65 1.60 2.50 0.94 0.02 0.02 0.02 LR09-6 4.49 0.06 LR09-9 180.56 2.41 2.40 1.80 0.97 0.00 0.02 0.00 LR09-30 716.99 9.57 9.60 8.40 0.94 0.03 0.02 0.01 LR10-6 83.16 1.11 1.10 0.90 0.03 0.74 0.01 0.22 LR10-15 551.42 7.36 7.40 6.90 0.98 0.00 0.02 0.00 LR10-30 448.78 5.99 6.00 5.20 0.93 0.04 0.02 0.01 LR12-30 466.76 6.23 6.20 5.30 0.97 0.00 0.02 0.01 LR13-30 454.02 6.06 6.10 9.30 0.98 0.00 0.01 0.00 LR14-SW 14.23 0.19 LR14-6 1.49 0.02 0.00 0.00 0.50 0.28 0.09 0.13 LR14-15 21.72 0.29 0.30 0.30 0.83 0.13 0.02 0.02 LR14-30 314.67 4.20 4.20 4.30 0.99 0.00 0.01 0.00 MEK-SW_N 0.02

MEK-SW_D1 5.24 0.07 MEK-SW_D2 5.24 0.07 RAIN_1 7.49 0.10 RAIN_2 7.49 0.10 MEDIAN 128.11 1.68 5.10 4.75 0.95 0.02 0.02 0.01 MIN 1.49 0.02 0.00 0.00 0.03 0.00 0.01 0.00 MAX 716.99 9.57 9.60 9.30 0.99 0.74 0.09 0.22

152

Table 5-S-3. Significant microbial variables correlating arsenic based on PLS-RFE

Variable Max Total Query Description Identity E value Accession related gene ID score score cover

33 Pyrobaculum arsenaticum 85% 152 152 57% 6.00E-34 gi|145282035|CP000660.1 arsA

94 Methanosarcina sp. MTP4 81% 181 544 88% 2.00E-43 gi|805342629|CP009505.1 ars family

130 Desulfitobacterium hafniense 98% 417 2089 100% 8.00E-114 gi|219536331|CP001336.1 ars family

184 Paludibacter propionicigenes 97% 371 1115 100% 6.00E-100 gi|312441806|CP002345.1 arsCBR

192 Tuber borchii 98% 458 458 100% 4.00E-125 gi|7141344|AF233293.1

362 Methyloferula stellate 98% 464 464 100% 7.00E-127 gi|304654561|FR686345.1

368 Methylobacterium sp. 100% 487 487 100% 8.00E-134 gi|1193727477|KY945724.1

387 Ensifer adhaerens strain 99% 469 1409 100% 2.00E-129 gi|1036657752|CP015880.1 arsR family

445 Desulfuromonas sp. 96% 337 674 100% 2.00E-89 gi|1015516084|CP015080.1 arsR family

446 Desulfuromonas sp. 97% 446 893 100% 2.00E-122 gi|1015516084|CP015080.1 arsR family

466 Anaeromyxobacter sp. 99% 337 674 100% 2.00E-89 gi|152026452|CP000769.1

526 Desulfomonile tiedjei 89% 323 323 99% 2.00E-85 gi|390621545|CP003360.1 arsR family

540 Desulfomonile tiedjei 81% 202 202 99% 5.00E-49 gi|390621545|CP003360.1

558 Planctomycetes bacterium 99% 341 682 98% 8.00E-90 gi|1140105870|CP017641.1 arsD

564 Planctomycetes bacterium 97% 444 444 99% 4.00E-121 gi|336110679|JF488132.1

639 Thermodesulfovibrio yellowstonii 95% 267 267 100% 1.00E-68 gi|206741110|CP001147.1 arsR family

716 Uncultured bacterium 96% 429 429 100% 2.00E-116 gi|146429725|EF220541.1

717 Geoalkalibacter subterraneus 99% 239 717 98% 5.00E-60 gi|749207097|CP010311.1 arsR family

806 Mycobacterium haemophilum 98% 442 442 95% 2.00E-121 gi|944205708|CP011883.2 ars family

813 Blastococcus saxobsidens 96% 377 1132 100% 1.00E-101 gi|378781357|FO117623.1 arsBCR

823 Streptomyces puniciscabiei 98% 339 2034 98% 4.00E-90 gi|1069152781|CP017248.1 arsB

829 Arsenic-oxidizing bacterium A12 97% 431 431 98% 7.00E-118 gi|523408152|KC527603.1

835 Streptomyces glaucescens 98% 464 464 100% 7.00E-127 gi|388330003|JQ899216.1

903 Uncultured bacterium 98% 306 306 99% 2.00E-79 gi|1229031628|MF612744.1

905 Uncultured bacterium 98% 306 306 99% 2.00E-79 gi|1229031628|MF612744.1

906 Streptomyces vietnamensis 97% 291 2037 100% 1.00E-75 gi|751387084|CP010407.1 arsR family

916 Blastococcus saxobsidens 97% 377 1132 100% 1.00E-101 gi|378781357|FO117623.1 arsBCR

918 Amycolatopsis mediterranei 98% 300 1190 100% 2.00E-78 gi|532225686|CP003777.1 arsR family

936 Uncultured bacterium 97% 400 400 91% 9.00E-108 gi|306033116|HM598829.1

967 Acinetobacter baumannii 98% 452 2715 100% 3.00E-124 gi|1211343212|CP001182.2 ars

976 Acinetobacter baumannii 98% 452 2715 100% 3.00E-124 gi|1211343212|CP001182.2 ars

153 993 Neisseria weaver 95% 412 1648 100% 4.00E-112 gi|1028913308|LT571436.1 arsR family

1031 Rubrivivax gelatinosus 97% 446 1340 100% 2.00E-122 gi|381376528|AP012320.1 ars and acr

1042 Sideroxydans lithotrophicus 97% 421 843 94% 6.00E-115 gi|291582584|CP001965.1 arsB

1084 Cryobacterium sp. 97% 342 1028 100% 3.00E-91 gi|1209535967|CP021992.1 arsB

1132 Sideroxydans lithotrophicus 97% 421 843 94% 6.00E-115 gi|291582584|CP001965.1 arsB

1142 Arsenic-oxidizing bacterium A12 97% 431 431 98% 7.00E-118 gi|523408152|KC527603.1

1144 Arsenic-oxidizing bacterium 97% 431 431 98% 7.00E-118 gi|523408152|KC527603.1

1146 Rubrivivax gelatinosus 97% 446 1340 100% 2.00E-122 gi|381376528|AP012320.1 ars and acr

1191 Pandoraea thiooxydans 97% 437 874 98% 1.00E-119 gi|1127371944|CP014839.1 ars RB

1220 Methylococcus capsulatus 99% 394 789 100% 7.00E-107 gi|66270661|AE017282.2 arsR family

1234 Methylomonas denitrificans 99% 429 1288 100% 3.00E-117 gi|998900463|CP014476.1 arsB family

1240 Methylomonas methanica 96% 435 1305 100% 5.00E-119 gi|333805788|CP002738.1 arsRCD family

1297 Reinekea aestuarii 98% 231 231 98% 7.00E-57 gi|1041495346|KX453231.1

1312 Choricystis parasitica 99% 477 477 100% 6.00E-131 gi|698349348|KM462878.1

1315 Unidentified bacterium 97% 446 446 100% 1.00E-121 gi|224039082|FJ662723.1

1333 Prochlorococcus marinus 95% 398 797 97% 5.00E-108 gi|33772318|BX548175.1 arsR family

Table 5-S-4. Microbes representing arrA genes in sample LR14-30

Max Query Identit qlen seq Description E value Aession score cover y gth

arr1 Desulfotomaculum sp. strain TC-1 anaerobic arsenite oxidase (arxA) gene 108 28% 2.00E-19 78% gi|1134445597|KX242336.1 594

arr2 Desulfuromonas sp. WB3 dissimilatory arsenate reductase large subunit (arrA) gene 369 96% 4.00E-98 78% gi|831250863|KM452746.1 581

arr3 Sulfuritalea hydrogenivorans sk43H DNA, dissimilatory arsenate reductase (ArrA) gene 187 87% 1.00E-43 84% gi|572099409|AP012547.1 222

arr4 Geobacter lovleyi SZ,dissimilatory arsenate reductase (ArrA) gene 221 90% 1.00E-53 87% gi|189419341|CP001089.1 462

arr5 Geobacter uraniireducens clone ArrA21 dissimilatory arsenate reductase (ArrA) gene 241 71% 2.00E-59 78% gi|341868964|JF827112.1 579

arr6 Alkalilimnicola ehrlichii MLHE-1, dissimilatory arsenate reductase (ArrA) gene 89.1 37% 5.00E-14 85% gi|114225560|CP000453.1 249

arr7 Geobacter uraniireducens clone ArrA21 dissimilatory arsenate reductase (ArrA) gene 248 89% 7.00E-62 80% gi|341868964|JF827112.1 363

arr8 Alkalilimnicola ehrlichii MLHE-1, dissimilatory arsenate reductase (ArrA) gene 139 77% 7.00E-29 75% gi|114225560|CP000453.1 378

arr9 Geobacter uraniireducens clone ArrA21 dissimilatory arsenate reductase (ArrA) gene 465 86% 4.00E-127 83% gi|341868964|JF827112.1 582

arr10 Geobacter uraniireducens clone ArrA21 dissimilatory arsenate reductase (ArrA) gene 206 73% 3.00E-49 83% gi|341868964|JF827112.1 303

154 arr11 Planctomyces brasiliensis DSM 5305, dissimilatory arsenate reductase (ArrA) gene 106 79% 4.00E-19 74% gi|324966854|CP002546.1 313

arr12 Geobacter uraniireducens clone ArrA18 dissimilatory arsenate reductase (ArrA) gene 310 85% 3.00E-80 77% gi|341868960|JF827110.1 582

arr13 Geobacter uraniireducens clone ArrA53 dissimilatory arsenate reductase (ArrA) gene 89.1 61% 4.00E-14 80% gi|341868998|JF827129.1 204

arr14 Geobacter lovleyi SZ, dissimilatory arsenate reductase (ArrA) gene 166 94% 6.00E-37 75% gi|189419341|CP001089.1 405

arr15 Geobacter uraniireducens clone ArrA21 dissimilatory arsenate reductase (ArrA) gene 241 80% 2.00E-59 77% gi|341868964|JF827112.1 516

arr16 Sulfuritalea hydrogenivorans sk43H DNA, dissimilatory arsenate reductase (ArrA) gene 310 82% 3.00E-80 78% gi|572099409|AP012547.1 579

arr17 Sulfuritalea hydrogenivorans sk43H DNA, dissimilatory arsenate reductase (ArrA) gene 316 92% 5.00E-82 78% gi|572099409|AP012547.1 513 Halomonas chromatireducens strain AGD 8-3, dissimilatory arsenate reductase (ArrA)

139 57% 1.00E-28 76% gi|985012000|CP014226.1 arr18 gene 576

arr19 Geobacter uraniireducens clone ArrA21 dissimilatory arsenate reductase (ArrA) gene 450 78% 2.00E-122 84% gi|341868964|JF827112.1 585

arr20 Sulfuritalea hydrogenivorans sk43H DNA, dissimilatory arsenate reductase (ArrA) gene 225 79% 5.00E-55 82% gi|572099409|AP012547.1 330

arr21 Geobacter sp. OR-1 arrA gene for dissimilatory arsenate reductase 269 95% 3.00E-68 79% gi|429325169|AB769875.1 384

arr22 Sulfuricella denitrificans skB26 , dissimilatory arsenate reductase (ArrA) gene 100 83% 3.00E-17 71% gi|540607552|AP013067.1 435

arr23 Geobacter uraniireducens clone ArrA21 dissimilatory arsenate reductase (ArrA) gen 210 85% 2.00E-50 80% gi|341868964|JF827112.1 366

Acknowledgement The authors acknowledge the support of NERC via standard research grants NE/P01304X/1 andNE/J023833/1.Sebastian Uhlemann (British Geological Survey, UK and ETH Zurich, Switzerland) is thanked for producing the map shown on Figure 1. LAR acknowledges support from The Leverhulme Trust (ECF2015-657).

155 Chapter 6 : Paper III: Metagenomic and metatranscriptomic studies reveal the diversity of microbes and their functional genes that could influence the arsenic redox cycle in aquifers in Bangladesh

Edwin T Gnanaprakasamb, Brian J. Maillouxa, Robert Danczake, Michael J Wilkinse, Imtiaz Choudhuryc, Benjamin C. Bostickd, Alexander van Geend and Jonathan R Lloydb a. Environmental Science Department, Barnard College, NY, USA. b. School of Earth and Environmental Sciences and Williamson Research Centre for Molecular Environmental Science, The University of Manchester, Manchester, UK. c. Department of Geology, University of Dhaka, Dhaka, Bangladesh d. Lamont-Doherty Earth Observatory, Columbia University, Palisades, NY, USA. e. School of Earth Sciences & Department of Microbiology, The Ohio State University, Columbus, Ohio, USA.

Abstract

Arsenic, which is mobilised in aquifers and collected via wells for drinking and farming, causes major health issues for more than 100 million people in Asia alone. Metal-reducing bacteria have been implicated in the mobilisation of sorbed As(V) from Fe(III)-bearing aquifer materials via respiratory processes that liberate soluble Fe(II) and As(III), however the identity of the causative organisms and the precise mechanism of this process remain poorly understood. In this study, we use state-of-the-art metagenomic and metatranscriptomic techniques to analyse the microbes present in Bangladeshi aquifers, and arsenic metabolising genes that they encode. 16S rRNA gene based community analysis using the MiSeq sequencing platform targetted 20 sediment and 11 water samples from three arsenic impacted aquifers, and revealed vast diversity of microbes colonising these subsurface environments. In addition we chose two contrasting water samples (high arsenic and shallow versus low arsenic and deep) for whole genome sequencing (WGS)

156 based metagenomic and metaproteomic analyses, to investigate the presence of arsenic functional genes and their expression in the two contrasting aquifers. This comparative study reveals that irrespective of arsenic geochemistry, arsenic metabolising genes and proteins are ubiquitous, including the dissimilatory arsenate reductase gene and its protein (arr and Arr), the arsenate reductase gene and its protein associated with As(V) resistance (ars and Ars) and the respiratory arsenite oxidase gene and its protein (aio and Aio) in both contrasting aquifers. These arsenic cycling genes were identified in bins represented by families of bacteria including Comamonadaceae, Rhodocyclaceae, Moraxellaceae etc. that were abundant in the arsenic impacted aquifers studied. To the best of our knowledge, this is the first study using high-throughput sequencing and a metagenomic- metaproteomic approach in enumerating the genes (and their encoded proteins) involved in metabolism of arsenic in deep aquifer samples.

Key words: arsenic, metagenomics, metatranscriptomics, 16S rRNA gene, arrA gene, biogeochemistry, aquifer.

6.1. Introduction

Chronic exposure to arsenic concentrations exceeding the limit of 10 μg·L−1, set by the World Health Organization, causes serious health issues, including cardiovascular and neurological disorders, cancer and diabetes in human beings (Smedley and Kinniburgh, 2002; Kapaj et al., 2006; WHO, 2011). Human exposure to arsenic occurs through water from shallow aquifers in which arsenate [As(V)] bound to minerals is reduced to arsenite [As(III)]; the latter being more mobile and labile (Rowland et al., 2002; Oremland and Stolz, 2003; Sun, Quicksall, et al., 2016). Previous studies provided evidence to indicate that various microbes including arsenate-reducing bacteria colonising these aquifers may mediate these reduction processes, leading to the mobilisation of arsenic in aquifers (Ferguson and Gavis, 1972; Pederick et al., 2007b; Pearcy et al., 2011; Ying et al., 2013; Héry et al., 2014; Rizoulis, Wafa M. Al Lawati, et al., 2014; Whaley-Martin et al., 2016). Microbes reduce arsenate enzymatically for the conservation of energy through respiration or for detoxification purposes (Stolz and Oremland, 1999; Mukhopadhyay and Rosen, 2002a; Kruger et al., 2013). The dissimilatory respiratory processes in microbes are

157 mediated by an operon encoding dissimilatory arsenate reductase genes referred to as arr, whereas the detoxification process is regulated by an operon encoding the ars arsenate reductase gene system. Different genes involved in the expression and regulation of the arsenic reduction are summarised in various reviews (Yamamura and Amachi, 2014; Andres and Bertin, 2016; Zhu et al., 2017).

The arr genes code for a membrane arsenate reductase (Arr), which is responsible for the anaerobic respiration of As(V) with the periplasm of the Gram- negative cell. The arr genes have been identified in many domains of bacteria including the Beta-, Gamma-, Delta- and Epsilon-proteobacteria, Firmicutes, Chrysiogenetes and Defferibacters (Van Lis et al., 2013). Prominent families of bacteria having arr genes include the Geobacteriaceae, Shewanellaceae, Bacillaceae, Chrysiogenaceae, Peptococcaceae and Helicobacteraceae (Pérez-Jiménez et al., 2005). Although there are many arr genes identified through PCR-based gene targeting metagenomic studies, only in three bacteria, namely Chrysiogenes arsenatis (Krafft and Macy, 1998), Bacillus selenitireducens (Afkar et al., 2003) and Shewanella sp. ANA-3 (Malasarn et al., 2008), have gene expression studies been conducted and the proteins purified and characterised. To date, only two archaea, including Pyrobaculum arsenaticum and Pyrobaculum aerophilum, are known to reduce arsenate, although there is no evidence that they carry arr genes (Huber et al., 2000; Cozen et al., 2009).

The ars detoxification operon that regulates the reduction of As(V) to As(III) via the cytoplasmic arsenate reductase (Ars) is evolutionarily conserved across bacteria, archaea and eukaryotes (Mukhopadhyay and Rosen, 2002a). The cascade of “ars” genes, which is responsible for transport, reduction and extrusion of arsenic, is regulated by the arsenic-responsive repressor gene (arsR) (Van Lis et al., 2013). As(V) transported into the cytoplasm of the cell by phosphate transporters is reduced to As(III) by the arsenate reductase (ArsC) encoded by the arsC gene. The reduced As(III) is then pumped out of the cell by the ArsB protein (a transmembrane carrier pump) encoded by the arsB gene (Tisa and Rosen, 1990). Many domains of bacteria including the Firmicutes, Beta- and Gamma- proteobacteria are known to contain species capable of reducing As(V) via the detox

158 system (Amend et al., 2014). For example, Bacillus spp. (Guo et al., 2015), Exiguobacterium spp. (Anderson and Cook, 2004), Staphylococcus spp. (Corsini et al., 2010), Sinorhizobium spp. (Hamamura et al., 2013), Alcaligenes spp. (Bahar et al., 2012) , Citrobacter spp. , Halomonas spp. , Psychrobacter spp. (Bowman et al., 1997) and many of the Pseudomonas spp. (Chang et al., 2008) are known to possess the ars operon.

Besides catalysing As(V) reduction, microbes are capable of oxidising As(III) to As(V) in oxic and anoxic environments (Amend et al., 2014). Anaerobic As(III) oxidation is mediated by the ArxAB protein, which is in turn regulated by the arx clade of genes. In this anaerobic coupled reaction which occurs in the periplasm, As(III) is oxidised to As(V), liberating electrons that are used to reduce NO3- to NO2-. Similarly, aerobic As(III) oxidation is mediated through the arsenite oxidase (AioAB), which is in turn regulated by the aioAB genes (Anderson et al., 1992). This process is cataysed in a wide range of bacteria including Thermus aquaticus and Thermus thermophiles (Gihring et al., 2001), Alcaligenes faecalis (Anderson et al., 1992), Hydrogenophaga sp. str. NT-14 (Hoven and Santini, 2004), Acidithiobacillus ferrooxidans (Duquesne et al., 2003) and Desulfotomaculum auripigmentum (Newman et al., 1997). As(V) is less mobile and has more affinity to the mineral solids, and the oxidation of As(III) could therefore be a remedy for As(III) contaminated environment (Newman et al., 1997; Smedley and Kinniburgh, 2002).

Moreover, arsenic biogeochemical cycling is often coupled/influenced by other biogeochemical processes including steps in the Fe, N and S cycles. The dissimilatory reduction of Fe(III) to Fe(II) can be energetically favourable for specialist anaerobic microorganisms, including Geobacter species and can result in either the solubilisation of Fe(II) and/or transformations of Fe minerals in sediments (Islam et al., 2005; Herbel and Fendorf, 2006). Dissimilatory iron- reducing bacteria (DIRB) can catalyse the reduction of Fe(III) to Fe (II) in the presence of electron donors such as acetate. In the coupling process, As(V) is eventually reduced to As(III) leading to release of the As(III) (Islam et al., 2004). Some microorganisms, such as Shewanella sp. ANA-3, Sulfurospirillum barnessi, Geobacter lovleyi, Geobacter uraniireducens, can reduce both Fe(III) and As(V)

159 (Campbell et al., 2006; Kocar et al., 2006). Similarly, microbial sulphate reduction may influence arsenic solubility through the formation of insoluble arsenic sulphides, e.g. As2S3 (Rittle et al., 1995; O’Day et al., 2004; Omoregie et al., 2013). Sulphate-reducing bacteria (SRB) can influence the arsenic redox cycle by generating hydrogen sulphide, from elemental sulphur or sulphate and leading to the precipitation of arsenic as arsenic sulphides. This is an important process in mitigating arsenic mobilisation via the precipitation of poorly soluble arsenic- bearing sulphide phases.

In this context of a complex arsenic biogeochemical cycle, previous studies have sought to understand the mechanism of arsenic mobilisation in aquifers (Gault et al., 2005; Rowland et al., 2006, 2008; Song et al., 2009; Dhar et al., 2011; Pearcy et al., 2011; Andres and Bertin, 2016; Xiao et al., 2016). However, so far to the best of the author’s knowledge no study has employed highly advanced bioinformatics tools or whole genome sequencing (WGS) based metagenomic and metaproteomic analyses in order to understand the microbial controls on arsenic mobilisation in aquifers. In this comprehensive study, we have used state-of-the-art techniques including metagenomics and metatranscriptomics, using HiSeq and MiSeq sequencing technologies, in order to sequence and analyse the DNA and mRNA related to arsenic metabolism in aquifer samples. Primarily 20 sediment samples and 13 groundwater samples from 3 aquifers in Bangladesh were subjected to the 16S rRNA gene based community analysis and geochemical analysis for As, Fe, S and N, pH and Eh. From these 33 samples, two contrasting water samples (high arsenic and shallow VS and low arsenic and deep) were selected for WGS based metagenomic and metatranscriptomic analysis using environmental DNA and mRNA. This is the first study of its kind to analyse two contrasting aquifers based on combined metagenomic and metatranscriptomic work in order to identify the organisms and genes potentially controlling arsenic speciation in contaminated aquifer environments.

160 6.2. Materials and Methods

6.2.1. Site description

The study sites are located in Araihazar Upazila, Bangladesh approximately 25 km east of the capital Dhaka. The sites are located in the Lashkardi village (Site F, 23.774 °N, 90.606 °E), Baylarkandi village (Site B, 23.780N, 90.640E) and Araihazar (CW-CAT, 23.791°N, 90.660°E (Figure 6-1.). Tube wells in sites B and F, were previously installed at these arsenic impacted sites in 2001 in order to monitor the temporal variability (Zheng et al., 2005; Dhar et al., 2008), whereas community well CAT at site CW-CAT was previously installed for drinking (Mihajlov et al., 2016).

Figure 6-1. The sampling sites of Site F, Site B and CW-CAT. From Site F and Site B, 20 freshly cored sediment samples (10 each along depth) and 12 (6 each) water samples from monitoring tube wells were collected for the geochemical and community analyses. From Site B, well id BA14-3 and Site CW-Cat, well id CAT were chosen in order to filter planktonic communities from more than 10,000 gallons of water for the metagenomic and metatranscriptomic analyses. The red zones are wells having between 51- 100 ug/L arsenic whereas green indicates 11-50 ug/L and blue indicates 0-10 ug/L. The sampling sites are represented by the black boxes in the red zone.

161 Site F is a sandy site with sand extending to the surface. The recharge rate is 0.5 m y- 1 and the maximum arsenic concentration of 200 g L-1 is reached at a 26 m depth (Stute et al., 2007). Site F drilling was performed 5 m from the tube wells in a hollow that was 1.6 m lower than the tube wells. Site B is capped by approximately 6 m of fine-grained silt and clay that overlies the sandy aquifer. The recharge rate is 0.08 m y-1 and the maximum arsenic concentration of 460 g L-1 is reached at a 13.7 m depth (Stute et al., 2007). Assuming a porosity of 0.25 at each site, the vertical groundwater velocity is 2.0 m y-1 at site F and 0.32 m y-1 at site B. Well CAT is 237.7 m (~780 feet) deep with a characteristic of reducing whitish-grey sand. While 12 water samples were collected from the existing tube wells installed in 2001, 20 sediment samples were acquired from a fresh coring in January 2015 at the same sites adjacent to the existing wells.

6.2.2. Sample collection and processing

Sediment cores were collected using the “hand-flapper” method, a manual drilling method often used to install wells. Intact sediment cores were collected at approximately 1.52 m (5 feet) intervals to 21.3 m (70 feet) at site B and 27.5 m (85 feet) at site F, by lowering a 1″ or 2″ ID gravity corer inside the drill pipe and into the sediment that had not yet been drilled. Upon coring, the sediment samples were split into aliquots for different types of analyses. The samples for the XAS analysis were saturated with glycerol (approximately 1:1 v/v) as a preservative and stored at - 20o C prior to the analysis. Glycerol was added to prevent oxidation by slowing the oxygen diffusion, and because it fixes bacteria in order to prevent microbiological alteration. In this case, spectra were collected using a thin (a few mM) thick film of wet, sediments without further treatment. Sample collection and synchrotron access were coordinated to minimise the lag time (about 2 weeks) between sampling and the analysis. The samples used for the XRF were used immediately in the field (without treatment) and the samples for other analyses were stored in the refrigerator prior to the analysis. From each depth, approximately 5 g of sediments were incised aseptically in an anaerobic bag fluxed with N2 and stored anaerobically at -20o C for molecular investigation. The samples were hermetically sealed and transported to the respective laboratories for analysis purposes. The diffuse spectral

162 reflectance of fresh cuttings was also measured in the field as a proxy for the Fe(II)/Fe(II+III) ratio in the acid-leachable Fe fraction (Horneman et al., 2004).

Groundwater samples were collected from 6 tube-wells, each from Site F and Site B, respectively. A battery-driven submersible pump (Whale SuperPurger) was used to pump the water from the wells at the rate of about 2 L/min. The samples for arsenic, other trace elements and other cations were collected in 30 ml acid-cleaned HDPE bottles, as described previously (Zheng et al., 2005; Dhar et al., 2008).

The samples were labelled in the order of XX100-0 in which X represents the sample site (B or F), X1 denotes sediment (S) or water (A), 00-0 is the depth in meters e.g. BS14-3. While the figures and tables in the main manuscript are labelled as Figure 6-1 and Table 6.1, the figures and tables in the supporting information are labelled as Figure 6-S-1 or Table 6-S-1.

6.2.3. Metagenomics Methods

6.2.3.1. 16S rRNA based community analysis

The DNA was extracted from the filters containing biomass using the MoBio Power Water DNA Isolation Kit (MoBio Laboratories, Inc., Carlsbad, CA, USA). The extracted DNA was quantified using the Invitrogen™ Qubit™ 3.0 Fluorometer using the highly sensitive Invitrogen™ Qubit™ quantitation assays. The V4 region of the 16S rRNA gene was amplified using dual indexed primers (150x2).

The V4 hypervariable region of the 16S rRNA gene was amplified in 50-μL volume reactions using the 0.5 μL (2.5 units) FastStart High Fidelity DNA polymerase (Roche Diagnostics GmbH, Mannheim, Germany), 5 μL of buffer with MgCl2, 1 μL of dNTPs, 2 μL (10 μM) of each forward and reverse fusion primers. The PCR was performed in a thermal cycler (Techne TC-300 ) with the following PCR conditions programmed: initial denaturing step at 95 °C for 2 min, 35 cycles of 95 °C for 30 s (melting), 55 °C for 30 s (annealing), 72 °C for 2 minutes (extension), and a final elongation step at 72 °C for 7 min (Kozich et al., 2013).

The libraries were quantified again in order to determine whether all the samples had amplified products required for the library preparation. The

163 normalisation was performed using the SequalPrep™ Normalization Plate Kit (Thermo Fisher Scientific) in order to obtain an equal quantity of the DNA library, which involves the three steps of binding, washing and elution. The normalised library was sized and quantified using high sensitivity DNA chips and Agilent Bioanalyser 2100 (Agilent Technologies, Inc., Santa Clara, CA, USA). A uniform concentration of the 4 nM DNA library of each sample was achieved by pooling the samples on the basis of their concentration. The pooled library was denatured along with PhiX before loading on to the Illumina Miseq next generation sequencer. MiSeq Reagent Kits v2 chemistry was used for sequencing.

The fastq files generated by the sequencer were analysed using the pipeline established to analyse the microbial communities using the 16S rRNA marker gene. The pipeline contains the tools including cutadapt, sickle, FastQC, Spades, Pandaseq, Vsearch and QIIME (Martin, 2009; Andrews, 2010; Caporaso, Kuczynski, et al., 2010b; Joshi and Fass, 2011; Bankevich et al., 2012; Masella et al., 2012; Rognes et al., 2016). The alpha diversity was analysed using the R statistical calculation package vegan and in QIIME using the rarefied data. The Principal Coordinate Analysis (PCoA) was used to compare groups of samples based on the phylogenetic or count-based distance metrics. The OTUs of interest were blast searched manually against the NCBI data in order to tabulate the identity and similarity of the bacteria. The entire workflow used in order to conduct the 16S rRNA based community analysis is presented in Figure 6-S-1.

6.2.3.2. Whole genome sequencing (WGS) based metagenome analysis

The workflow for sequencing 101x2 paired end shotgun reads using HiSeq 2500 started with 1 ng of gDNA, extracted from the filters containing biomass using the MoBio Power Water DNA Isolation Kit (MoBio Laboratories, Inc., Carlsbad, CA, USA). Following the manufacturer’s instruction provided along with Nextera XT DNA library prep kit, the library was prepared. In the first step, the DNA was tagmented using the Tagment DNA Buffer (TD) and the tagmentation was assessed by running 1 μL of the sample on the High Sensitivity DNA chip in an Agilent Bioanalyzer 2100. The indexing of the PCR reagents including indices i5 and i7, Nextera PCR Master Mix (NPM) added along the tagmented DNA and the PCR was

164 implemented in the thermal cycler (Techne TC-300) with the programmed PCR conditions of preheating (72oC for 3 minutes), initial denaturing (95oC for 30 s), melting (95oC for 10 s), annealing (55oC for 30 s), extension (72oC for 30 s) and final elongation (72oC for 5 s). The library was cleaned using AMPure XP beads in the resuspension buffer and normalised in a bead based binding and elusion steps. The normalised library was once again quantified using the Agilent Bioanalyzer 2100 before it was pooled and loaded on the sequencing cartridge (HiSeq SBS kit V4). The sequencing was conducted in the HiSeq 2500 system at the University of Manchester. During all steps in DNA extraction and library preparaton, the proceduarl blanks were run to monitor the purity of the nucleic acids.

The shotgun reads generated by HiSeq2500 were demultiplexed using (bcl2fastq v.2.17.14) before they were subjected to quality control using Sickle. The quality controlled sequences were assembled using IDBA-UD and the coverage calculations were tabulated using Bowtie2. MetaBAT, a binning tool, was used to bin the scaffolds to the respective bins. The assembled scaffolds were annotated using the prodigal, Pfam, TIGRFAM, ProSiteProfiles, Usearch and Amphora databases and tools (Haft, 2001; Hyatt et al., 2010; Wu and Scott, 2012; Finn et al., 2016). The bins were manually cleaned on the basis of the GC content and coverage depth. Once the bins were cleaned, the genes were annotated again and the metabolic pathways were scored in an Excel sheet. The entire workflow used to analyse the WGS based metagenome analysis is presented in Figure 6-S-2.

6.2.4. Metatranscriptomics methods

A metatranscriptomic analysis has been conducted as indicated in the schematic diagram in Figure 6-S-3 in the supporting information. The total RNA was extracted from the filters containing biomass from two aquifers identified as CAT and BA14-3 in this study. The total RNA was quantified using Nanodrop (260nm:280nm) and Qubit 3.0. The RNA was enriched using the mRNA enrichment kit (MICROBExpressTM Kit) in the three steps of annealing, bead-binding and recovering, as per the manufacturer’s instructions. From the enriched mRNA, cDNA was prepared following the manufacturer’s instruction of the Apollo 324 PrepX mRNA library protocol (IntegenX), which is based on directional RNA adaptor

165 ligation (Lau et al., 2016). The sample CAT was indexed with a 6-mer of “ATCACG”, BA14-3 was indexed with another 6-mer of “CGATGT”. In the following steps, the cDNA library was amplified and quantified using Agilent Bioanalyzer 2100. The quantified library was sequenced using the Illumina HiSeq 2500 platform with Illumina’s Truseq Rapid SBS chemistry.

The sequenced reads were filtered for high-quality reads using the tools in the galaxy pipeline at Princeton University. The reads below the Phred quality score of 30 (Q30) were discarded in the Quality Control (QC) step followed by demultiplexing the reads and removing the adapters and primers. Once again using Sickle, quality control was performed with the same parameters. The quality controlled reads were assembled using Bowtie2 and the genes were called using prodigal. The software Cufflink was run to Running cufflinks in order to count transcript abundances and calculate FPKM (Reads per kilobase of exon per million reads mapped). The genes were aligned against the metagenomics bins in order to determine whether the organisms were active in the samples analysed.

6.2.5. Geochemical methods

6.2.5.1. Analysis of trace elements in sediment

The elemental composition of the sediment was determined by X-ray Fluorescence (XRF) spectroscopy using an InnovX Delta Premium hand-held XRF analyser using triplicate measurements and a collection time of approximately 75 seconds. This device calculates environmental concentrations based on the total fluorescence at three energy levels in order to efficiently discriminate between elements, such as lead and arsenic with similar emission lines. The analytical precision based on statistical counting errors is approximately ±1-2 mg kg-1 and ±300 mg kg-1 for arsenic and Fe, respectively. Under the same conditions, the analytical precision values of the repeated measurements of the uniform standard reference materials collected independently over days to weeks are within 5% of each other and within 6% of the certified values. Since the XRF uses a small quantity of sample, the measured concentrations can vary more significantly when repeated on heterogeneous natural materials, and the analytical accuracy under the

166 experimental conditions is thus somewhat lower, namely about 10-20% for any single measurement (Jung et al., 2012; Sun, Quicksall, et al., 2016).

6.2.5.2. Analysis of trace elements in water

Using an Axiom single-collector instrument (Thermo Elemental, Germany), the concentrations of As, P, Fe, Mn, S, Ca, Mg, K, Na and 33 other trace elements in the groundwater samples were measured at the Lamont–Doherty Earth Observatory with a reproducibility typically <5% by means of high-resolution inductively- coupled plasma mass spectrometry (HR ICP-MS) (Cheng et al., 2004; Dhar et al., 2008). The protocols that were followed to ensure the accuracy and precision of the data included: (1) two NIST standard reference materials (1640 and 1643E, Trace element in natural water), and an internal laboratory consistency standard (LDEO tap-water spiked with analyte elements) included with each run. The results for these standards were always within 5% of the certified values following the calibration of the instrument with separate standards at the beginning and end of each run; (2) whenever possible, time-series samples from the same well were analysed within the same run of 30 samples, which usually improved the reproducibility to <3%; (3) at least 2 samples were reanalysed between two consecutive runs for the same well to ensure consistency between the runs (Dhar et al., 2008).

6.3. Results

6.3.1. XRF and ICP-MS based As, Fe and S quantification

Site F was a sandy site with sand extending to the surface where as Site B was capped by approximately 6 m of fine-grained silt and clay that overlies the sandy aquifer. The concentrations of As, Fe and S were measured and are presented (Figure 6-2) here to facilitate the interpretation of the microbial and molecular data in the light of geochemistry. At Site-F, arsenic was recorded in the sediment in the range of 1 mg.kg-1 to 6 mg.kg-1 The As concentration peaked at depths 18.4 m and 26 m with 6 mg.kg-1 and 5 mg.kg-1, respectively. Consistent with Site-F, Site-B also had the same range of As concentrations in the sediment, ranging from 1 mg.kg-1 to 6 mg.kg-1. At depths 6.1 m and 16.8 m at Site-B, the As concentrations recorded were 5

167 mg.kg-1 and 6 mg.kg-1, respectively. The median arsenic concentration in both sites was 3 mg.kg-1. Site-B recorded the highest concentration of aqueous arsenic in

Figure 6-2. As, Fe and S concentration in sediment and groundwater samples. In the sediment, the unit is expressed in mg/Kg, whereas in groundwater the unit is expressed in ug/L. groundwaters at a depth of 14.3 m with 471 μg.L-1, whereas at Site-F the highest aqueous arsenic concentration was 214 μg.L-1 at a depth of 14.3 m. At Site-B, when the depth increased, the arsenic concentrations decreased consistently in the aqueous phase reaching 0.44 μg.L-1 at 90.6 m from 471 μg.L-1 at 14.3 m. Similarly, at Site-F, the concentration of arsenic dropped to 0.15 μg.L-1 from 214 μg.L-1 at a depth of 24.6 m.

The Fe concentration in the sediment samples remained consistent along the depths with the median concentration of 14,500 mg.kg-1 at Site-F and 26,523 mg.kg-1

168 at Site-B. The highest concentration of Fe was recorded at 26 m at Site-F with 35,300 mg.kg-1, whereas at Site-B it was 36,371 mg.kg-1 at 9.1 m. The aqueous Fe concentrations were proportional to the aqueous As concentrations along various depths at Site-F, but differed in the case of Site-B where As concentration remained consistent at various depths analysed. Contrary to the Fe concentration, the aqueous sulphur (S) concentration decreased as a function of depth at both sites (Table 6-S-1 and Table 6-S-2).

The two wells sampled for the WGS-based metagenome analyses had contrasting features of depth, As and S concentrations. Aquifer BA14-3 is a shallow 14.3 m well, contrastingly aquifer CAT is a deep 237.7 m well, which makes it completely anoxic. While the groundwater from the CAT site recorded 2.4 μg.L-1 of aqueous arsenic, BA14-3 had an aqueous arsenic concentration of 471 μg.L-1 . Similarly, the concentration of S was 123 μg.L-1 in BA14-3, while the CAT sample had 5402 μg.L-1 S. The Fe concentrations were 10170 μg.L-1 (BA14-3) and 8206 μg.L-1(CAT) (Table 6-S-3).

6.3.2. 16S rRNA based microbial community analysis (Site F and Site B)

MiSeq illumina based sequencing of the 16S rRNA gene from 32 samples generated a total of 5.8 million paired-end reads with a maximum of 389,981 reads (BA90-6) and a minimum of 129,344 reads (FS12-3). The groundwater samples yielded a median of 173,895 reads, which is 17,631 more reads in comparison to a median value of 156,264 reads from the sediment samples. The operational taxonomic units (OTUs) were more diverse in sediment samples than in the groundwater samples. The sediment samples had a median of 795 OTUs, comparison to a median value of 357 OTUs in the groundwater samples. The samples from the 9.1 m and 21.3 m depths at Site-B had 1039 and 1072 OTUs respectively, suggesting a constant diversity of microbes from shallow to deeper sediments. In the groundwater samples, taken at 14.3 m at Site-F and 90.6 m at Site- B, 811 OTUs and 771 OTUs, were noted respectively, suggesting a diverse community of anaerobic microbes in the reducing environments. Figure 6-3 presents the summary of reads and OTUs from the sediment and groundwater samples across various depths.

169

Figure 6-3. Bar diagrams summarising the number of reads and OTUs obtained in each sample in the 16S rRNA community analysis using MiSeq. Bars coloured in red and positioned on the left hand side are the number of reads and the black coloured bars on the right represent the OTUs

The principal coordinate analysis (PCoA) of the OTUs showed that the planktonic communities found in the groundwater samples were different from the communities that colonised the sediments. Figure 6-4 illustrates the grouping of the samples based of the communities of prokaryotes represented by distinctive OTUs. In the weighted plot, the aqueous samples are grouped together as one component and the sediment samples were grouped into three distinctive groups depicting the community variance among the sediment samples from two different sites. In the unweighted grouping of the prokaryotes (43.6% variation explained), the samples were grouped into four clusters, where two clusters were constituted of sediment samples grouped by sites and the other two were groundwater samples by sites. This distinctive grouping shows the variance among the communities based on the site and the type of sampling (sediment or water).

170

Figure 6.4. Weighted and unweighted principle coordinate analysis (PCoA) plots of betadiversity among the sediment and water samples from Site-F and Site-B. A. Unweighted plot; B. Weighted plot. The water samples are shaded in blue, green indicates the sediment samples, while the outliers comprising sediment and water samples are unshaded.

The 16S rRNA gene based phylogenetic analysis revealed a wide diversity of prokaryotes colonising these two aquifers. In total, 1531 prokaryotic species were represented by various OTUs in 32 sediment and groundwater samples. The 100% identity at the species level was represented by many uncultured and partial sequences and so family level confirmation (<96% identity) of the prokaryotes remained inevitable for many organisms. In the higher level of classification Proteobacteria outnumbered other phyla constituting over 90% of the total abundance in various samples including, FA18-3 (98 %), FA10-3 (96%), FS26-0 (94%), AND BA10-9 (90%) . Of the Proteobacteria, the Gamma- and Beta- proteobacteria had a higher percentage of abundance (<88%) in comparison to the Delta- and Epsilon-proteobacteria (<16%). After the Proteobacteria, representatives of Firmicutes, Actinobacteria and Acidobacteria were the next most highly represented in total abundance. Many bacteria thought to control arsenic biogeochemical cycling belong to these phyla and classes (Van Lis et al., 2013; Yamamura and Amachi, 2014; Andres and Bertin, 2016; Zhu et al., 2017). Figure 6-5 presents the overall distribution of the prokaryotes at the phylum and class level.

171

Figure 6-5. Relative abundance of bacteria belonging to different phyla and classes.

Of the 150 families of bacteria identified in this study, the Comamonadaceae constituted 12.9% of the total abundance followed by the Moraxellaceae with 12.8%. The other bacterial families that had more than 2% abundance in total are the Pseudonocardiaceae (8.8%), Methylophilaceae (4.9%), Rhodocyclaceae (4.1%),Bacillaceae 4.0%, Micrococcaceae (3.1%), Nitrospiraceae (3.1%), Gallionellaceae (2.3%) and Xanthomonadaceae (2.1%). These families are generally known for the driving various biogeochemical cycles of different elements including C, N, Fe, S, As, and P,. (Lloyd, 2003; Anderson et al., 2011; Wang et al., 2014). Bacterial families known to include dissimilatory arsenate-reducing bacteria (DARB) (Pérez-Jiménez et al., 2005) including the Bacillaceae (4.1%), Helicobacteraceae (1.7%), Clostridiaceae (0.7%), Peptococcaceae (0.6%), Geobacteriaceae (0.06%) and Shewanellaceae (0.01%), were distributed across many sediment and water samples (Yamamura and Amachi, 2014; Andres and Bertin, 2016; Zhu et al., 2017). Many bacterial families reported to have the ars gene clade used for As detoxification were

172 also found in this study, of which the Moraxellaceae were in a high relative abundance (12.8%), followed by the Bacillaceae (4.1%). As(III)-oxidising bacterial families detected at relatively high abundance included the Comamonadaceae (12.9%) and Rhodocyclaceae (4.1%). The Fe(II)-oxidising Gallionellaceae (2.2%), and the Fe(III)-reducing Geobacteriaceae (0.06%), the sulphite-oxidising Bradyrhizobiaceae (0.29) and Hyphomicrobiaceae (0.33%) are a few other bacterial families that were detected and could influence arsenic biogeochemical cycling. The full account of the families of prokaryotes found in this study are presented with their relative abundance in Table 6-S-4 in the supporting information.

6.3.3. 16S rRNA gene based community analysis of two contrasting aquifers

MiSeq illumina based sequencing of the 16S rRNA gene of two geochemically contrasting samples, BA14-3 and CAT generated 105,423 and 194,317 reads, respectively. The OTUs representing these reads were 911 OTUs in BA14-3 and 754 OTUs in the CAT sample. The database search of the OTUs representing more than 1% of the reads with >98% identity revealed a diverse community structure in both

Figure 6-6. Close relatives of the bacterial species and their relative abundance in the two contrasting aquifers CAT and BA14-3, revealed through 16S rRNA gene based community analysis.

173 aquifers (Figure 6-6). In the shallow aquifer of BA14-3, a close relative of Lysobacter ruishenii (99% identity) was the most abundant species detected with 6.92% of the total abundance and was followed by a close relative of the Curvibacter delicatus (100% identity) with 5.7% of total abundance and a relative of the Deinococcus daejeonensis (99% identity) at 3.3% total abundance. All the bacteria exceeding 1% of total abundance were mainly from the Xanthomonadaceae, Comamonadaceae, Deinococcaceae, Moraxellaceae and Rhodocyclaceae families. In the CAT deep aquifer, a close relative of Curvibacter delicatus belonging to the Comamonadaceae made up 12.2% of the total abundance, and was followed by close relatives of Sulfurisoma sediminicola (100% identity) and Dechloromonas hortensis (98%) (both from the Rhodocyclaceae ) with 6.7% and 6.4% of total abundance, respectively. Close relatives of the two species known for Fe(II) oxidation, namely Sideroxydans lithotrophicus (100% identity) and Ferriphaselus amnicola (100% identity), belonging to the Gallionellaceae family had a total abundance of 4.2%.

6.3.4. WGS based Metagenome

6.3.4.1. Reads, scaffolds and bins

The whole genome sequencing of two geochemically contrasting samples was done on the HiSeq, illumina platform and generated about 150 million base pairs. The average Phred score that stood above 35 for the forward and reverse reads in each case showed the high quality of reads obtained from sequencing. The assembling of these reads produced 330,975 scaffolds for the sample CAT sample and 376,313 scaffolds for the BA14-3 sample, with an N50 value (median length of scaffolds) of 1967 for CAT and 1491 for BA14-3. The MetaBAT based binning of these scaffolds generated 50 bins for CAT and 45 bins for BA14-3, respectively. The cut off percent of completeness for the bins based on single copy genes (SCG) was set as 70%, which reduced the number of bins to 25 (CAT) and 24 (BA14-3), respectively. In both samples together, only 14 bins had 100% completeness based on the SCGs. The details of the reads, scaffolds and bins are summarised in Table 6-S-5, in the supporting information.

174 6.3.4.2. Prokaryotic communities

On the basis of different phylogenetic markers, including the 16S rRNA gene (through EMIRGE), rpsC (gene for ribosomal protein) and SCG (single copy gene) markers, the scaffolds in different bins were designated to the respective bacteria. The assembled genome from both samples were used in order to construct the phylogenetic tree, which is presented in Figure 6.7. Considering the enormity of species, the tree was collapsed to different levels of phylogenetic hierarchy. BA14-3 was dominated by bacteria belonging to families including the Comamonadaceae, Oxalobacteraceae, Sphingomonadaceae, Rhodobacteraceae, Caulobacteraceae, Nitrospirae and Rubrobacteridae, whereas the CAT sample was dominated by the Gallionellaceae, Rhodocyclaceae, Nitrospirae and Lentisphaerae. Previous studies on arsenic contaminated aquifers too have reported the occurrence and abundance of same bacterial families (Héry et al., 2008; Paul et al., 2015; Z. Hassan et al., 2015; Xiao et al., 2016). At the family/order level, many bacteria identified through 16 rRNA gene based sequencing were found in both samples. At the genus level too, the bacteria identified in the 16S rRNA based community analysis were found in both samples. The family of the Comamonadaceae was represented by close relatives of Albidiferax ferrireducens, Curvibacter laceolatus and Ramlibacter tataouinensis in both samples. Out of the 18 scrutinised bins from the CAT sample, 4 bins were designated to bacteria belonging to the Gallionellaceae but at the genus level, only a close relative to Gallionella capsiformans was identified. Bacteria known for dissimilatory arsenate reduction were not found in any of the bins. The detailed account of the bins, along with the bacterial representation, genes and metabolic activities are given in Table 6-S-6 and Table 6-S-7 in the supporting information.

175 N

o m Samples

u

r a BA14-3

a b i K a

r a i

a r c e

t i CAT a e

s t i t e c

r

e c r a

e

r a

i t

b a b

l

b A c

l a a r i i Arsenic Functions a

e c c r n

h b b t e o a D e r t

p e n c r e ArsB e l a i n a m i a

d n a b

a

o A v e

C ArsC f

c o l

i o o

c G a i ArrA c a W i r a u r e ia i t r r M s e C t c e e AioA - t t y e T c a c c a a l h a b i a a y a n e b r b i k b e k o n r i m s n t N b a m f a c n F a A a b o r a a u z Z s a u n o b c s A A a t c s a a e M g t + o Y n K r e a i b d a r 1 a u M ia i r a c o B e t D t 3 e c ia r a r i b e A a t r w c ia m o a er a lk lb t tim a e ac S F rk b C o y e ri h n n B a lo ad e h ro e rg cc f t is a le es t S xi et es tes ma 2 no R ge P icro C M 3 ria WE te W bac jka Do Firmi cutes

Peregrinibacteria Gracilibacteria WOR-1 Absconditibacteria WOR-3 ogae Thermot Othe robia r simic Ga Elu mmap roteob acteria -2 OR hica W trop T mni hio O 6 C tric TM hro ha s ae Ec ma les ete ae rid to tia yc rid te thi les om cte ac X orh ct ba ob a od lan ro ri nt os P ub Co es ho pi R al ae m rac et e on ea yc rin ad e m te al no c e es ti ba a A c i e a l A n in i c r o r a C e pi g e li o th o n a a e g m O r ra e i t e O a n r c N n x P o i e ia e R a a m p c t r a h c o s c c s G is o e lo n o e e b O b a t co a t t a sa d a a d p c a t l o e c a e o ib a e h lio ri c t c tr r r d d e a y e e ic t ib a n c c r a S a a H r e e a e M A ix n i R l a la ce o r D y la l a Z 6 e p h e e m t a Z c s e i 0 i h iz e t c r e 4 E C o o a a a e N t S e R s b t s p a a m m b s p i i s A A c e i s p u R a e A a l e t R i h v l i l e c l m S r l a c i r h i e a a l o c l o n g a a s o h r e a o i s b o o n a r i b n i a d o g a c e s t G r r d n d g o l b i e e a o h o o d I f c p r i t R p o e e o i m l a o c p L o t a r o t b i V e c S r b b i t u n b o t a d s c o c c c e c o e b a a t i a e p c s i a a i a e a n k p n o a a t r c y c r v b i r a e e c e c i a a a e e x n r t t b b C e t r r o e b a i t o i l d a o t e t a d e u t i y f o s e i t r u e e l r t r a o l l c c A C b l o r i r c p s s a k a d r r l g M u i h a i e m a i l e i o i e p a m h h c o e m a a a n f s a b o o c s i s m e r D l a l s r r a h o e e R i e e t a e o n e e e u i t m v r s p t D D e f c h i e r r

t l a c c i a i S r y l T s a a u e o e d F N s n b B i

e a a i a

n D

t

e s Figure 6-7. Phylogenetic tree for samples BA14-3 and CAT. The bacterial hierarchy found in the samples is highlighted in cyan The blue squares represent the prokaryotes from sample BA14-3, whereas red represents the prokaryotes from sample CAT. The arsenic functional genes are highlighted in different shapes as given in the legend.

6.3.4.3. Arsenic Genes

In this study, we narrowed down our search for genes related to arsenic including arr (dissimilatory arsenate reductase gene), ars (cytoplasmic arsenate reductase gene) and aio (arsenite oxidation gene) in the bin based metagenomic analyis. Figure 6-8 presents the total number of bins in which these genes are found. BA14-3 had 10 bins with the arrA gene whereas CAT had 5 bins with the arrA gene.

176

Figure 6-8. Number of bins containing translated genes. BA14-3 and CAT are the sample IDs. Blue and purple bars represent the total number of bins with the respective genes; ‘Active Bins’ denotes the number of bins, where the arsenic genes are translated or active.

In BA14-3, these arrA genes were in the genome bins represented by bacteria most closely related to Albidiferax ferrirducens, Curvibacter lanceolatus and Ramilibacter tataouinensis from the family of Comamonadaceae, Novosphingobium, Novosphingobium aromaticivoran, Sphingobium japanicum from the family of Sphingomonadaceae and Rubrobacter xylanophilus from the family of Rubrobacteraceae. In CAT, the arrA genes were found in bins represented by the closest relatives of bacteria including, Ferribacterium limneticum Sulfurisoma sp. and Dechloromonas hortensis from the family of Rhodocyclaceae, Lentisphaera sp. from the family of Lentisphaeraceae and Brevundimonas sp. from the family of Caulobacteraceae. The ars genes that regulate the reduction of As(V) via detoxification were found ubiquitously in all bins with more than 70% completeness. The metabolic pathway search using the KASS-KEGG automatic annotation server showed that many of these bacteria, which had arsenic related

177 genes, also had genes related to N or S cycling as presented in Table 6-S-6 and Table 6-S-7.

6.3.5. Arsenic related metatranscriptomics

The metatranscriptomic analysis was done in this study primarily to investigate whether arsenic related genes found in the genomic bins were transcribed or active. HiSeq based cDNA sequencing generated 92,039,084 reads for BA14-3 and 69,378,192 reads for CAT. Bowtie2 (Langmead and Salzberg, 2012) and Cufflinks (Trapnell et al., 2012)aided assembly and transcript analysis revealed only 0.00001% arsenic related reads in the samples. Figure 6.8 shows that the arsenic related genes were transcribed in both the samples. Although the arrA gene responsible for dissimilatory arsenate reduction was found in 10 metagenomic bins in BA14-3 (the high arsenic aquifer), the analysis of mRNA showed no evidence of the arrA genes being transcribed. On the contrary CAT, (the low arsenic aquifer) which had 5 metagenomic bins with arrA gene had evidence of the gene being transcribed in two bins. The transcripts matched two metagenomic bins represented by close relatives to Ferribacterium limneticum (Rhodocyclaceae) and Lentisphaera sp. (Lentisphaeraceae). The cytoplasmic arsenate reductase gene arsB was active in both contrasting aquifers. In both aquifers, more than 20 bins were identified to have arsC genes but the genes were found active only in two bins each, in both aquifers. The active bins were represented by bacteria including Novosphingobium aromaticivoran (Sphingomonadaceae) and Ramilibacter tataouinensis (Comamonadaceae) in the BA14-3 aquifer samples and Gallionellaceae and Lentishaera sp. (Lentisphaeraceae) in the CAT aquifer samples. The arsB genes were active in both the aquifers with equal representation of two bins each by close relatives to Albidiferax ferrireducens and Ramilibacter tataouinensis (Comamonadaceae) in aquifer BA14-3 and Curvibacter sp. (Comamonadaceae) and Ferribacterium limneticum (Rhodocyclaceae). Although the arsenite oxidise gene (aio) was identified in 5 metagenomic bins in BA14-3 and 4 bins in CAT, only one bin in CAT was active and was represented by a close relative to Thermodesulphovibrio sp. (Nitrospiraceae). While Table 6-S-6 and Table 6-S-7 present the total number of bins and the representative bacteria found active in

178 both aquifers, Table 6-S-8 presents the details of the transcript scaffolds relating to the genomic bins and genes.

6.4. Discussion

The objective of this study was to investigate the diversity of arsenic related microbes and genes in two arsenic contaminated aquifers (Site B and Site F)(Zheng et al., 2004; Datta et al., 2009) in comparison to a low arsenic aquifer (CAT) (Mihajlov et al., 2016), in Bangladesh. To achieve this objective, the research was carried out in 4 stages (i) in the first stage the geochemistry of 33 sediment and groundwater samples from 2 arsenic impacted aquifers and the low arsenic negative control aquifer were analysed; (ii) in the second stage, a 16S rRNA gene based community analysis was carried out to determine which prokaryotes colonised these aquifers in different depths; (iii) in the third stage, on the basis of geochemical measurments, a high arsenic shallow aquifer (471 μg.L-1, 14.3 m ) and a low arsenic deep aquifer (2.4 μg.L-1, 237.7 m ) were chosen to investigate the difference in the arsenic related genes and microbial communities, based on WGS metagenome analysis; (iv) in the fourth stage, the mRNA from these two contrasting aquifers were analysed to investigate if these arsenic related genes were expressed in these two contrasting environments.

The geochemical analysis of sediment samples from two arsenic contaminated aquifers revealed that the sediment along various depths had a mean concentration of 3 mg.kg-1, a concentration critical enough to activate the expression of arsenic related genes (Saltikov et al., 2005). The previous studies of these sediments and groundwaters reveal that at shallow depths, the aqueous arsenic concentration is high and as the depth increases (>50 m) the aqueous arsenic concentration decreases (Zheng et al., 2004; Dhar et al., 2011; Mailloux et al., 2013; Li et al., 2015). Also previous studies revealed that the soluble arsenic species are mainly (~90%- 98%) As(III), and As(V) in the sediments (Cullen and Reimer, 1989; Sorg et al., 2014). In accordance with the previous studies, at shallow depths the aqueous arsenic concentration remained high in the range of 87.2 μg.L-1 and 471.4 μg.L-1, but beyond 50 m deep, all three sites recorded arsenic concentration in the range of 0.2 μg.L-1 and 2.5 μg.L-1 (Stute et al., 2007; Datta et al., 2009). The 16S rRNA gene based

179 community analysis of DNA from these depths revealed no visible change in the prokaryotic community structure with reference to depth or arsenic concentration. On the contrary, the principal coordinate analysis (PCoA) of the OTUs grouped all 33 samples under four different clusters based on site and sample types (sediment and groundwater). This shows that there is a statistical difference in the community structure based on sample types, not on the basis of arsenic concentration. The most abundant prokaryotic families found in all samples together were Comamonadaceae (12.94 % total abundance), Moraxellaceae (12.87 %), Pseudonocardiaceae (8.85 %), Methylophilaceae (4.96 %), Rhodocyclaceae (4.17 %), Bacillaceae 4.03%, Micrococcaceae (3.14%), Nitrospiraceae (3.14%), Gallionellaceae (2.26%) and Xanthomonadaceae (2.08%) (Héry et al., 2014; Z. Hassan et al., 2015; Desoeuvre et al., 2016). These families in general are known to catalyse various steps of the C, N, Fe, S, As, P, elemental cycles (Lloyd, 2003; Anderson et al., 2011; Wang et al., 2014) but are not implicated in arsenic biogeochemical cycling. With the exception of Bacillaceae, the families of DARPs known to reduce As(V) to conserve energy were not found in the most abundant families, but DARP families of Clostridiaceae (0.76%), Peptococcaceae (0.62%), Geobacteriaceae (0.06 %) and Shewanellaceae (0.01%) were found in low abundance (Pérez-Jiménez et al., 2005). This could be due to various reasons including (i) only a very few bacteria have been studied for dissimilatory arsenate reduction (Amend et al., 2014); (ii) the levels of As are significant from a health perspective but below that of other electron acceptors such as Fe (III) (Mirza et al., 2014); (iii) only a few prokaryotes are known to depend on arsenic for energy conservation compared to other bacteria which makes the DARPs a minority group of bacteria colonising these aquifers (Yamamura and Amachi, 2014).

The bacterial families reported to encode ars gene clade used for arsenic detoxification have been found in this study, in particular the families of Moraxellaceae (12.87%) and Bacillaceae (4.03%) (Cai et al., 2009a; Lara et al., 2012; Escudero et al., 2013; Guo et al., 2015). This shows the common presence of cytoplasmic arsenate reducing genes in many organisms as previously reported in many studies (Lin et al., 2006; Qin et al., 2006b; Escudero et al., 2013; Rahman and

180 Hassler, 2014; Paul et al., 2015). A relatively high abundance of As(III)oxidising bacterial families were detected including Comamonadaceae (12.94%) and Rhodocyclaceae (4.17%) (Amend et al., 2014). The Fe(II)-oxidising Gallionellaceae (2.26%) (Emerson et al., 2010), the Fe(III)-reducing Geobacteriaceae (0.06 %) (Lovley, 1991), sulphite-oxidising Bradyrhizobiaceae (0.29%) and Hyphomicrobiaceae (0.33%) are a few other bacterial families that could influence the arsenic biogeochemical cycle as reported previously. Most of these bacterial families are known for nitrogen cycling also (Anderson et al., 2011; Daims et al., 2015).

The two geochemically contrasting samples were analysed for their community structure, using both 16S rRNA gene based community analysis on MiSeq platform and the WGS metagenome based community analysis on HiSeq platform. Though two different sequencing platforms and analysis pipelines were used, the results showed that there is 90% similarity in the community structure in family level. It shows that the 16S rRNA based community analysis is an appropriate method to analyse the communities at the family level. At the species level the similarity was reduced by 40%, leaving ambiguity at the species level identity for most organisms (Riesenfeld et al., 2004; Kozich et al., 2013; Segata et al., 2013).

Binning based metagenomic analysis revealed the presence of many arsenic related genes in various bins. In our study, we hypothesised that the shallow-high arsenic aquifer will have more genes related to dissimilatory arsenate reduction compared to the deep-low arsenic aquifers. The results showed that there is no obvious difference between both aquifers on the basis of arsenic related genes. More studies are required to confirm (and quantify) this uniformity of occurrence of arsenic related genes in contrasting environments (L. Chen et al., 2015; Xiao et al., 2016).

The confirmation of gene expression through metatranscriptomic analysis showed that the arsenic related genes are active in both contrasting aquifers but in a very low abundance. Only 10 % of the genomic bins had evidence that the genes were expressed. This is especially surprising in the high arsenic aquifer where there was no evidence of arrA gene being expressed. In general, extracting mRNA from

181 aquifer samples is challenging. The difference in the bins could be attributed to the methodologies used in processing the mRNAs (L. Chen et al., 2015).

In conclusion, the results show that these aquifers are colonised by diverse microbial population with diverse arsenic genes with the potential to reduce mineral-bound As(V) to As(III) which readily dissolves in water. The main findings of this study are (i) the most abundant bacterial species belonging to the families of Comamonadaceae, Moraxellaceae, Rhodocyclaceae, Gallionellaceae etc, not known for dissimilatory arsenic reduction, might possess arrA genes and thus have the potential to mobilise arsenic through dissimilatory arsenate reduction; (ii) the bacterial community structure revealed through 16S rRNA gene based sequencing and analysis, resembles the family level community structure revealed through the WGS based community analysis; (iii) although arsenic resistant genes are found in many organisms, they are transcribed only in a few organisms. More metaomic studies on arsenic impacted aquifers will clarify the role of these and related organisms in mobilising arsenic in aquifers.

182 Supporting Material Paper III: Metagenomic and metatranscriptomic studies reveal the diversity of microbes and their functional genes that influence the arsenic redox cycle in the arsenic impacted aquifers in Bangladesh

Figure 6-S-1. The workflow of the 16S rRNA gene based community analysis using the MiSeq sequencing platform and amplicon analysis pipeline.

183

Figure 6-S-2. Workflow for the whole genome sequencing (WGS) based gene and community analysis using the HiSeq sequencing platform and metagenomic pipeline.

184

Figure 6-S-3. Schematic workflow of the metatranscriptomic analysis using the HiSeq sequencing platform.

185 Table 6-S-1. Summary of the geochemical analysis for Site F used in the analysis

SITE-F: SEDIMENT SAMPLES Unit FS03-1 FS06-2 FS09-2 FS12-3 FS15-3 FS18-4 FS21-4 FS24-5 FS26-0 FS27-5 As mg/L 1 3 3 1 3 6 1 3 5 3

Fe mg/L 33500 14100 16100 14900 12100 9700 9300 11800 35300 30000

SITE-F: GROUNDWATER SAMPLES FA05-0 FA10-3 FA14-3 FA18-6 FA24-6 FA56-0

As ug/L 2.7 87.2 101.4 186.1 214.3 0.2

Fe ug/L 172.60 1426.23 1823.5 312.81 6960.5 414.2

S ug/L 3637.9 4708.22 4109.9 3605.9 1489.9 BD*

Table 6-S-2. Summary of the geochemical analysis for Site B used in the analysis

SITE-B : SEDIMENT SAMPLES Unit BS03-1 BS06-1 BS09-1 BS12-2 BS13-7 BS15-2 BS16-8 BS18-3 BS19-8 BS21-3 As mg/L 3 5 3 3 4 3 6 4 3 3 Fe mg/L 30337 24238 36371 23137 30185 28809 18923 11273 24161 30507 SITE-B: GROUNDWATER SAMPLES BA07-3 BA10-9 BA14-3 BA19-4 BA28-1 BA90-6 As ug/L 23.5 268.6 471.14 388.08 254.00 0.44 Fe ug/L 11194.4 5508.1 10170.8 19201.1 14315.7 2282.6 S ug/L 18636.3 3317.4 123.02 49.90 30.64 69.10

Table 6-S-3. Two contrasting aquifers Unit BA14-3 CAT Depth meter 14.3 237.7

As ug/L (PPB) 471.1 2.5 Fe ug/L (PPB) 10170.8 8206.4 S ug/L (PPB) 123.1 5402.6

*BD = Below detectable limit

186 Table 6-S-4. Total percentage of reads representing different families of prokaryotes from a total of 32 samples.

% % % Family Abun.* Family Abun. Family Abun. Comamonadaceae 12.94% Acidimicrobiaceae 0.36% Eubacteriaceae 0.05% Moraxellaceae 12.87% Hyphomicrobiaceae 0.33% Heliobacteriaceae 0.05% Pseudonocardiaceae 8.85% Planctomycetaceae 0.32% Hyphomonadaceae 0.05% Methylophilaceae 4.96% Neisseriaceae 0.32% Nitrosomonadaceae 0.04% Rhodocyclaceae 4.17% Hydrogenophilaceae 0.32% Caldilineaceae 0.04% Bacillaceae 4.03% Chitinophagaceae 0.30% Methanobacteriaceae 0.03% Micrococcaceae 3.14% Bradyrhizobiaceae 0.29% Corynebacteriaceae 0.03% Nitrospiraceae 2.46% Family XI 0.28% Myxococcaceae 0.03% Gallionellaceae 2.26% Beijerinckiaceae 0.28% Peptostreptococcaceae 0.03% Xanthomonadaceae 2.08% Streptomycetaceae 0.27% Aurantimonadaceae 0.03% Uncultured bacterium II 1.89% Family XII 0.27% Rhodobiaceae 0.03% Sphingomonadaceae 1.87% Nocardiaceae 0.26% Bacteroidaceae 0.03% Helicobacteraceae 1.61% Ruminococcaceae 0.21% Cryomorphaceae 0.02% Chloroflexaceae 1.49% Family I 0.21% Cellulomonadaceae 0.02% Uncultured bacterium I 1.44% Syntrophaceae 0.20% Brevibacteriaceae 0.02% Unknown Bacterial Family 1.23% Burkholderiaceae 0.19% Nannocystaceae 0.02% A family of Caulobacteraceae 1.22% Gammaproteobacteria 0.19% Thermoactinomycetaceae 0.02% Oxalobacteraceae 1.19% Micromonosporaceae 0.16% Iamiaceae 0.02% A family of Anaerolineaceae 1.15% Deltaproteobacteria 0.16% Kineosporiaceae 0.02% Nocardiopsaceae 1.11% Lachnospiraceae 0.15% Solimonadaceae 0.02% Cyanobacteria 1.09% Leptospiraceae 0.14% Chromatiaceae 0.02% Candidate division OP11 0.99% Cytophagaceae 0.13% Sandaracinaceae 0.02% Rhizobiales 0.91% Phyllobacteriaceae 0.12% Brucellaceae 0.01% Porphyromonadaceae 0.89% Erysipelotrichaceae 0.12% Acetobacteraceae 0.01% Sporichthyaceae 0.88% Verrucomicrobiaceae 0.11% Gracilibacteraceae 0.01% Unknown Archaea Family 0.78% Solirubrobacterales 0.11% Lactobacillales 0.01% Clostridiaceae 0.76% Gemmatimonadaceae 0.11% Thermotogaceae 0.01% A family of Betaproteobacteria 0.75% Mycobacteriaceae 0.11% Rhodospirillales 0.01% Acidobacteriaceae 0.74% Alcaligenaceae 0.11% Brocadiaceae 0.01% Peptococcaceae 0.62% Family XVIII 0.11% Roseiflexaceae 0.01% Fusobacteriaceae 0.60% Staphylococcaceae 0.10% Xiphinematobacteraceae 0.01% Campylobacteraceae 0.57% Family XIII 0.10% Streptosporangiaceae 0.01% Pasteurellaceae 0.57% Solirubrobacteraceae 0.09% Actinomycetaceae 0.01% Rhodobacteraceae 0.51% Christensenellaceae 0.09% Methanoregulaceae 0.01% Microbacteriaceae 0.51% Planococcaceae 0.09% Veillonellaceae 0.01% Flavobacteriaceae 0.51% Erythrobacteraceae 0.09% Shewanellaceae 0.01% Intrasporangiaceae 0.49% Ignavibacteriaceae 0.08% Legionellaceae 0.01% Coriobacteriaceae 0.48% Holophagaceae 0.08% Haliangiaceae 0.01% Methylobacteriaceae 0.46% Promicromonosporaceae 0.07% Virgulinella fragilis 0.01% Methylococcaceae 0.43% Syntrophorhabdaceae 0.07% Polyangiaceae 0.01% Rhodospirillaceae 0.40% Bdellovibrionaceae 0.06% Halothiobacillaceae 0.01% Gaiellaceae 0.38% Cystobacteraceae 0.06% Thermomonosporaceae 0.01% Rikenellaceae 0.38% Syntrophobacteraceae 0.06% Methanosaetaceae 0.01% Bacteriovoracaceae 0.38% Geobacteraceae 0.06% Phycisphaeraceae 0.01% Crenotrichaceae 0.38% Methylocystaceae 0.05% * % Abun. = % Abundance

187 Table 6-S-5. Account of reads, scaffolds and bins for two contrasting aquifer samples Sample Preparation CAT BA14-3 Sample Preparation Kit Nextera XT Nextera XT Indexing Dual Dual Sequencing Sequencing Platform HiSeq HiSeq Lanes 2 samples in a lane 2 samples in a lane Sequencing Chemistry HiSeq V2 HiSeq V2 Cycles 101 x 2 101 x 2 Reads Raw R1 Reads - total bases 77,598,670 74,335,271 Raw R2 Reads- total bases 77,598,670 74,335,271 Raw R1 Reads - Average Quality Score 37 36.05 Raw R2 Reads- Average Quality Score 37 35.31 Trimmed R1 Reads - total bases 75,180,338 72,554,697 Trimmed R2 Reads - total bases 75,180,338 72,554,697 Trimmed R1 Reads - Average Quality Score 37 36.5 Trimmed R2 Reads - Average Quality Score 37 36.33 Scaffolds Total number of sequences 330,975 376,313 Total number of bps 353,223,240 355,055,601 Average sequence length 1067.22 bps. 943.51 bps. N50 1967 bps 1491 bps Range # reads (%) 0-100 310 (0.09%) 276 (0.07%) 100-500 173506 (52.42%) 212864 (56.56%) 500-1000 96092 (29.03%) 103313 (27.45%) 1000-5000 52555 (15.87%) 52318 (13.9%) 5000-10000 5388 (1.62%) 4801 (1.27%) 10000-20000 2004 (0.6%) 1751 (0.46%) 20000-50000 818 (0.24%) 717 (0.19%) 50000-100000 199 (0.06%) 190 (0.05%) 100000-500000 101 (0.03%) 82 (0.02%) 500000- 2 (0%) 1 (0%) GC and Coverage Depth Low GC 31.87 29.40 High GC 73.34 74.94 Averg GC 61.13 59.64 Lowest Coverage 11.79 7.60 Highest Coverage 173.99 401.87 Average Coverage 52.47 35.73 Median Coverage 27.39 23.9476 Number of Bins 50 45 BINS with 100% completeness based on the SCG 6 8 BINS ABOVE 70% completeness based on the SCG 17 18

188

Table 6-S-6. Detailed account of genomic bins for BA14-3

Comp- Mis- In CXXCH Anam Bin # Family id ArrA Aio Acr3 ArsBC ANR DNR DSR SOX lete bin 16S 5+ mox Bin 4 81% 35% Comamonadaceae Albidiferax ferriruducens Y 10 96 46 23 0 SOX Bin 6 100% 48% Sphingomonadaceae Sphingobium japonicum Y 1 27 27 5 NarB NirS AprAB Bin 8 97% 16% Intrasporangiaceae Instraporangium calvum Y 12 15 1 NasAB NarGHIJ SoxD Bin 9 94% 0% Sphingomonadaceae Sphingomonas wittichii Y 1 18 16 2 NirA Bin 10 100% 0% Chitinophagaceae Chitinophaga pinensis 12 12 1 NarGHIJ SoxD Bin 15 100% 0% Oxalobacteriaceae 0 NirK SoxD Bin 16 90% 10% Syntrophaceae Syntrophus aciditrophicus Y 13 7 5 NasAB SoxD Bin 17 100% 87% Sphingomonadaceae Novosphingobium Y 11 14 23 17 30 Bin 18 100% 33% Rubrobacteraceae Rubrobacter xylanophilus Y 6 14 37 22 52 Bin 19 100% 51% Sphingomonadaceae Novosphingobium aromaticivoran Y 14 11 37 45 7 NorBC Bin 22 97% 65% Sphingomonadaceae Sphingopyxis alaskensis Y 28 3 NirK Bin 24 74% 3% Cytophagaceae Niastella koreensis 23 13 1 NasAB NorBC SoxD Bin 27 100% 74% Caulobacteraceae Phenylobacterium zucineum Y 69 47 6 NorBC, SoxD Bin 30 100% 61% Rubrobacteraceae Rubrobacter xylanophilus Y 4 11 NasAB NarGHU NirK Bin 35 87% 10% Thermoanaerobacteraceae Ammonifex degensii 9 2 21 0 Sat Bin 36 84% 23% Comamonadaceae Curvibacter lanceolatus Y 9 3 0 Bin 37 77% 48% Comamonadaceae Ramlibacter tataouinensis Y 21 15 53 30 2 NarGHU SoxD Comp. denotes the completeness of each bin calculated on the basis of single copy genes (SCG); ID stands for the identity of bacteria based on the rpsC/ SCG /EMIRGE/Quicklook database. ‘In 16S’ is the presence of the same bacteria identified through the 16S rRNA gene based community analysis using the MiSeq platform (Y stands for Yes). ArrA=dissimilatory arsenate reductase; Aio=arsenite oxidase; Acr3 and ArsBC = Cytoplasmic arsenate reductase; CxxxCH are the cytrochrome motifs found in each bin. ANR = Anaerobic Nitrate reduction. NO3-->NH4, DNR=Denitrification=Nitrate-->N2, ana=Anammox=Anaerobic ammonia oxidation genes; Sox= Sulphate oxidation. The N and S metabolising genes were identified using the KEGG-KASS tool. The blank cells denote absence and the numbers denote the number of motifs found in the bins. Protein names in the cells denote the genes being highlighted in the KEGG-KASS pathways. In fluorescent green are the bins and genes found active.

189 Table 6-S-7. Detailed account of genomic bins for CAT

Bin # Comp. Misbin Family ID In 16S ArrA Aio Acr3 ArsBC CXXCH 5+ ANR DNR Anammox DSR SOX Bin 1 97% 81% Comamonadaceae Curvibacter sp. Y 2 NarGHIJ SoxD Bin 3 90% 87% Rhodocyclaceae Ferribacterium limneticum Y 7 2 3 3 164 SoxD Bin 4 76% 0% Rhodocyclaceae Dechloromonas hortensis Y 1 7 54 SoxD Bin 5 84% 0% Sphingomonadaceae Y 32 Bin 6 100% 0% Gallionellaceae Gallionella capsiformans Y 1 1 1 65 NorBC Bin 10 100% 0% Gallionellaceae Gallionella capsiformans Y 1 3 14 104 NorBC Bin 11 100% 68% Gallionellaceae Y 3 NorBC Bin 14 100% 100% Gallionellaceae Y 0 4 12 103 NorBC Bin 15 100% 0% Nitrospiraceae Thermodesulphovibrio sp. Y 2 1 12 62 NasAB NirK Bin 18 87% 0% Thiotrichales Thiothrix sp. 2 1 3 60 SoxD Bin 19 97% 77% Ectothiorhodospiraceae Ectothirhodospira sp. Bin 21 100% 100% Lentisphaeraceae Lentisphaera sp. 2 2 Bin 23 90% 10% Sphingomonadaceae Sphingopyxis sp. Y 4 Bin 24 97% 13% Beijerinckiaceae Beijerinckia sp. Y 0 0 11 4 NarGHU Bin 29 94% 3% Moraxellaceae Acinetobacter sp. Y 6 22 Bin 30 87% 68% Rhodocyclaceae Sulfurisoma sp. Y 1 11 Bin 37 90% 68% Sphingomonadaceae Sphingopyxis sp. Y 3 Bin 42 87% 45% Caulobacteraceae Brevundimonas sp. Y 1 6 Comp. denotes the completeness of each bin calculated on the basis of single copy genes (SCG); ID stands for the Identity of bacteria based on the rpsC/ SCG /EMIRGE/Quicklook database. ‘In 16S’ is the presence of the same bacteria identified through the 16S rRNA gene based community analysis using the MiSeq platform (Y stands for Yes). ArrA=dissimilatory arsenate reductase; Aio=arsenite oxidase; Acr3 and ArsBC = Cytoplasmic arsenate reductase; CxxxCH are the cytrochrome motifs found in each bin. ANR = Anaerobic Nitrate reduction. NO3-->NH4, DNR=Denitrification=Nitrate-->N2, ana=Anammox=Anaerobic ammonia oxidation genes; Sox= Sulphate oxidation. The N and S metabolising genes were identified using the KEGG-KASS tool. The blank cells denote absence and the numbers denote the number of motifs found in the bins. The protein names in the cells denote the genes being highlighted in the KEGG-KASS pathways. In fluorescent green are the bins and genes found active.

190

Table 6-S-8. Arsenic realted transcript level expression matching translated genomic bins

BA14-3 Bin locus length coverage FPKM FPKM_conf_lo FPKM_conf_hi Protein 4 scaffold_36_113:691-952 261 344.513 4.32E+06 3.64E+06 5.00E+06 ArsB 4 scaffold_194_59:4-503 499 29.7981 406392 305151 507634 ArsC 4 scaffold_194_60:0-146 146 166.94 3.30E+06 1.58E+06 5.02E+06 ArsB 4 scaffold_194_60:310-678 368 16.3474 207440 112125 302754 ArsB 4 scaffold_194_60:798-1041 243 117.667 1.52E+06 1.06E+06 1.98E+06 ArsB 4 scaffold_194_61:195-375 180 201.237 2.52E+06 1.53E+06 3.51E+06 ArsC 4 scaffold_868_14:189-334 145 404.818 4.98E+06 2.85E+06 7.10E+06 ArsC 4 scaffold_868_17:6-151 145 420.174 5.20E+06 3.02E+06 7.38E+06 ArsB 19 scaffold_824_33:0-145 145 366.494 4.52E+06 2.50E+06 6.55E+06 ArsC 37 scaffold_277_11:93-422 329 32.0283 497064 327027 667101 ArsC 37 scaffold_277_12:310-673 363 27.9431 371460 241467 501453 ArsB 37 scaffold_277_12:811-1027 216 30.3986 550878 218686 883070 ArsB 37 scaffold_277_13:210-428 218 99.0388 1.41E+06 884381 1.93E+06 ArsC CAT locus length coverage FPKM FPKM_conf_lo FPKM_conf_hi 1 scafffold_154_12:223-368 145 741.42 9.96E+06 6.79E+06 1.31E+07 ArsB 1 scafffold_154_12:423-568 145 332.484 4.48E+06 2.37E+06 6.60E+06 ArsB 1 scafffold_9036_1:964-1151 187 286.357 3.95E+06 2.74E+06 5.17E+06 ArsB 3 scafffold_111_114:723-902 179 175.731 2.38E+06 1.37E+06 3.40E+06 ArsB 3 scafffold_8105_5:47-332 285 50.601 749686 487333 1.01E+06 ArrA 14 scafffold_32_368:283-428 145 686.134 9.22E+06 6.15E+06 1.23E+07 ArsC 15 scafffold_74_56:1525-1702 177 186.298 2.83E+06 1.69E+06 3.96E+06 AioA 21 scafffold_667_4:157-295 138 217.581 3.06E+06 1.08E+06 5.04E+06 ArrA 21 scafffold_1259_1:40-185 145 1594.05 2.14E+07 1.68E+07 2.61E+07 ArrA 21 scafffold_3190_9:70-215 145 241.667 3.24E+06 1.44E+06 5.03E+06 ArsC 32 scafffold_5207_2:16-161 145 539.389 7.22E+06 4.54E+06 9.91E+06 ArsC Bin = metagenomic bins, FPKM = Fragments Per Kilobase of transcript per Million mapped reads. FPKM_conf_lo = Lower bound of the 95% confidence interval of the abundance of this isoform, as a fraction of the isoform abundance. That is, lower bound = FPKM * (1.0 - conf_lo); FPKM_conf_hi=Upper bound of the 95% confidence interval of the abundance of this isoform, as a fraction of the isoform abundance. That is, upper bound = FPKM * (1.0 + conf_lo).

191 Chapter 7 : Conclusions and future work directions

7.1. Conclusions

The primary objective of this project was to investigate the role of prokaryotes in the mobilisation of arsenic in aquifers in Bangladesh and Cambodia. Prior to this work, many microcosm and enrichment based studies had shown that the microbes are capable of reducing and mobilising arsenic directly or indirectly in arsenic bearing mineral assemblages. However, both the organisms involved and the underlying mechanisms have remained elusive. A combination of cutting edge molecular ecology techniques, alongside geochemical and mineralogical characterisations were used to address this knowledge gap and accomplish the objective of this study.

A molecular approach was taken to investigate the microorganisms colonising these aquifers and the arsenic functional genes that they encode. It has been argued that less than 2% of bacteria can be cultivated in laboratory media (Wade, 2002). For this reason, culture independent molecular approaches were selected, using a combination of state-of-art-techniques of high throughput next generation sequencing technologies and advanced bioinformatics algorithms to investigate the arsenic-related microbes and arsenic-functional genes found in sediment and groundwater samples in arsenic impacted aquifers.

The geochemical approach was also taken to characterise the sediments and groundwaters. Previous studies suggested that solid phase arsenic in these aquifers are dominated by As(V), whereas the aqueous phase is dominated by As(III). Arsenic speciation is in turn influenced by the biogeochemical cycling of other elements, with Fe, C, N, S speciation crucial as well as pH and Eh. For this reason, the geochemistry of these aquifers was analysed using advanced geochemical techniques including XANES, XRF, ICP-MS and ICP-AES.

The main hypotheses that underpin this thesis are:

As(V) is reduced to As(III) in two possible ways: (i) since sediment is dominated by As(V) and groundwater by As(III), the arsenic is

192 reduced in the sediment; (ii)As(V) could be reduced during Fe(III) reduction, and then reduced to As(III) in the aqueous phase.

Arsenic-reducing microbes will be found in greater abundance in sediments with As(V), and here arsenic functional genes such as arrA will be transcribed. On the contrary, arsenic-reducing microbes will not be found in such high relative abundance in arsenic free aquifers and the arsenic functional genes will not be transcribed.

There will be a statistical correlation between the microbial and geochemical data which will help identify the key organisms reacting to mobilised arsenic (and/or controlling its speciation and mobility).

In summary:

Analysing 38 sediment and groundwater samples from two contrasting aquifers from Bangladesh using state-of-the-art techniques of pyrosequencing, XANES, XRF and ICP-MS supported the view that direct As redox transformations are central to arsenic fate and transport, and that there is a residual reactive pool of both As(V) and Fe(III) in deeper sediments that could be released by microbial respiration in response to hydrologic perturbation, such as increased groundwater pumping that introduces reactive organic carbon to depth.

Analysing 30 arsenic contaminated groundwater samples from Cambodia using a combination MiSeq-illumina sequencing techniques, an amplicon analysis pipeline, ICP-MS, ICP-AES and O2-PLS (orthogonal projection of latent structures) methods showed that it is possible to find correlations between geochemical and microbial community data, and identified 61 microbes that have a significant correlation to arsenic in two Cambodian transects, with 68% most closely related to known to carry the ars operon required for resistance to As. The potential impact of these organisms on arsenic biogeochemistry is discussed, alongside future research directions including culture-dependant and metagenomic studies that are

193 required to clarify further the role of these organisms in controlling As speciation in environmental systems.

Analysing 33 sediment and ground water samples from three contrasting aquifers from Bangladesh in a subsequent study, here using 16S rRNA gene based community analysis of 33 samples (MiSeq platform), alongside WGS metagenome analysis of 2 contrasting samples (HiSeq) sequencing based metagenome and metatranscriptome analyses, XRF and ICP-MS based geochemical analysis, I showed that irrespective of arsenic geochemistry, arsenic metabolising genes and proteins are ubiquitous, including the dissimilatory arsenate reductase gene and its protein (arr and Arr), the arsenate reductase gene and its protein associated with As(V) resistance (ars and Ars) and the respiratory arsenite oxidase gene and its protein (aio and Aio) in both contrasting aquifers. These arsenic cycling genes were identified in bins represented by families of bacteria including Comamonadaceae, Rhodocyclaceae, Moraxellaceae etc. that were abundant in the arsenic impacted aquifers studied and have not been considered to play a role in As biogeochemistry previously.

7.2. Future work directions

As part of this study, DNA and mRNA were extracted and sequenced (WGS using Nextseq) from 5 samples (2 sediment and 3 groundwater samples). This data set is substantial, and complete bin analysis of the functional genes and transcripts through pipelines with advanced algorithms is an obvious future task.

An exclusive study that only targets archaea, in order to better constrain the role of Archaea in mobilisation of arsenic in aquifers also warrants further attention.

A multi-omic study comprising DNA, mRNA, proteins and metabolites as done previously in thermokarst bog soil (Hultman et al., 2015) could be done with arsenic contaminated aquifer samples.

194 A comparative study on rate of arsenic reduction could be done on model organisms capable of cytoplasmic arsenate reduction (reduction for detoxification or resistance) and dissimilatory arsenate reduction.

Arsenic mobilisation is a complex biogeochemical process involving many physical, chemical and molecular parameters mentioned in the literature of this thesis. An umbrella study of all these parameters on ‘sub-samples’ is vital in unravelling the mystery behind arsenic mobilisation. So, a mass sampling method to analyse all these parameters were planned in 2014 by one of my collaborators Lex van Geen. Including ‘multi-omic’ technique was suggested above as part of this umbrella study, and would help us get the right samples for understanding the role of microorganisms in arsenic redox cycling and coupling to other biogeochemical cycles. The team informed the possibility of advanced mass sampling technique in the following lines: “We will have the capability to drill and process samples 24 hrs/day. The limiting factor is likely to be the rate at which the cores can be processed under anaerobic conditions. The guiding principle will be to measure as many sediment and groundwater properties as possible on-site, especially in the case of labile properties - rather than trying to maximizing the length of core recovered. The mobile lab will be therefore be packed with an anaerobic chamber, a hydride-generation/atomic fluorescence analyzer for As, an ion chromatograph, a DIC/NPOC analyzer, at least one gas chromatograph, a spectrophotometer, and possibly other instruments. In a second container, whole cores will be logged for porosity etc. and split cores will be analyzed at high-resolution for reflectance, major elements by XRF, and magnetic susceptibility”. 19 March 2014 This work is still going ahead in the near future in SE Asia, and has the exciting addition of the inclusion of a sampling system that can freeze samples in situ, preserving RNA, metabolites and proteins for multi-omic studies. It is hoped that the techniques that I have developed and applied in my thesis will form the basis of the microbiological work on these important new samples.

195 List of References Achour-Rokbani, A., Cordi, A., Poupin, P., Bauda, P., and Billard, P. (2010) Characterization of the ars gene cluster from extremely arsenic-resistant Microbacterium sp. strain A33. Appl. Environ. Microbiol. 76: 948–55. Achour, A.R., Bauda, P., and Billard, P. (2007) Diversity of arsenite transporter genes from arsenic-resistant soil bacteria. Res. Microbiol. 158: 128–37. Afkar, E., Lisak, J., Saltikov, C., Basu, P., Oremland, R.S., and Stolz, J.F. (2003) The respiratory arsenate reductase from Bacillus selenitireducens strain MLS10. FEMS Microbiol. Lett. 226: 107–112. Ahmann, D., Roberts, A. L., Krumholz, L. R., & Morel, F. M. (1994). Microbe grows by reducing arsenic. Nature, 371(6500), 750-750. Akay, C., Thomas, C., and Gazitt, Y. (2004) Arsenic trioxide and paclitaxel induce apoptosis by different mechanisms. Cell Cycle 3: 324–334. Altschul, S.F., Gish, W., Miller, W., Myers, E.W., and Lipman, D.J. (1990) Basic local alignment search tool. J. Mol. Biol. 215: 403–10. Amend, J.P., Saltikov, C., Lu, G.-S., and Hernandez, J. (2014) Microbial arsenic metabolism and reaction energetics. Rev. Mineral. Geochemistry 79: 391–433. Anderson, C.R., Condron, L.M., Clough, T.J., Fiers, M., Stewart, A., Hill, R.A., and Sherlock, R.R. (2011) Biochar induced soil microbial community change: Implications for biogeochemical cycling of carbon, nitrogen and phosphorus. Pedobiologia (Jena). 54: 309–320. Anderson, C.R. and Cook, G.M. (2004) Isolation and characterization of arsenate- reducing bacteria from arsenic-contaminated sites in New Zealand. Curr. Microbiol. 48: 341–347. Anderson, G.L., Williams, J., and Hille, R. (1992) The purification and characterization of arsenite oxidase from Alcaligenes faecalis, a molybdenum-containing hydroxylase. J. Biol. Chem. 267: 23674–82. Andrade, J.M., Gómez-Carracedo, M.P., Krzanowski, W., and Kubista, M. (2004) Procrustes rotation in analytical chemistry, a tutorial. Chemom. Intell. Lab. Syst. 72: 123–132. Andres, J. and Bertin, P.N. (2016) The microbial genomics of arsenic. FEMS Microbiol. Rev. fuv050. Andrews, S. (2010) FastQC: a quality control tool for high throughput sequence data. software. Argos, M., Kalra, T., Rathouz, P.J., Chen, Y., Pierce, B., Parvez, F., et al. (2010) Arsenic exposure from drinking water, and all-cause and chronic-disease mortalities in Bangladesh (HEALS): A prospective cohort study. Lancet 376: 252–258. Bahar, M.M., Megharaj, M., and Naidu, R. (2012) Arsenic bioremediation potential of a new arsenite-oxidizing bacterium Stenotrophomonas sp. MM- 7 isolated from soil. Biodegradation 23: 803–812.

196 Bankevich, A., Nurk, S., Antipov, D., Gurevich, A.A., Dvorkin, M., Kulikov, A.S., et al. (2012) SPAdes: A New Genome Assembly Algorithm and Its Applications to Single-Cell Sequencing. J. Comput. Biol. 19: 455–477. Barringer, J.L. and Reilly, P.A. (2013) Arsenic in groundwater: a summary of sources and the biogeochemical and hydrogeologic factors affecting arsenic occurrence and mobility. In, Current perspectives in contaminant hydrology and water resources sustainability. InTech. Bassas-Galia, M., Nogales, B., Arias, S., Rohde, M., Timmis, K.N., and Molinari, G. (2012) Plant original Massilia isolates producing polyhydroxybutyrate, including one exhibiting high yields from glycerol. J. Appl. Microbiol. 112: 443–54. Bateman, A. (2002) The Pfam Protein Families Database. Nucleic Acids Res. 30: 276–280. Bencko, V. (1977) Carcinogenic, teratogenic, and mutagenic effects of arsenic. Environ. Health Perspect. 19: 179–82. Berg, M., Tran, H.C., Nguyen, T.C., Pham, H. V, Schertenleib, R., and Giger, W. (2001) Arsenic contamination of groundwater and drinking water in Vietnam: a human health threat. Environ. Sci. Technol. 35: 2621–6. Bergey, D.H. (2005) Bergey’s Manual of Systematic Bacteriology - Vol 2: The Proteobacteria Part A - Introductory Essays. Springer-Verlag New York Inc. 332. Bergey, D.H. (2010) Bergey’s Manual of Systematic Bacteriology - Vol 4: Bacteroidetes. Bergquist, E.R., Fischer, R.J., Sugden, K.D., and Martin, B.D. (2009) Inhibition by methylated organo-arsenicals of the respiratory 2-oxo-acid dehydrogenases. J. Organomet. Chem. 694: 973–980. Bertin, P.N., Heinrich-Salmeron, A., Pelletier, E., Goulhen-Chollet, F., Arsène- Ploetze, F., Gallien, S., et al. (2011) Metabolic diversity among main microorganisms inside an arsenic-rich ecosystem revealed by meta- and proteo-genomics. ISME J. 5: 1735–47. BGS/MML (1999) Arsenic contamination of ground water in Bangladesh: a review. Prepared for the Department of Public Health Engineering, Government of Bangladesh Dhaka. BGS, D. (1999) Groundwater studies for arsenic contamination in Bangladesh. Final report, rapid Investig. phase. Dep. Public Heal. Gov. Bangladesh, Mott MacDonald Br. Geol. Surv. Bhattacharya, P., Vahter, M., Jarsjö, J., Kumpiene, J., Ahmad, A., Sparrenbom, C., et al. (2016) Arsenic Research and Global Sustainability: Proceedings of the Sixth International Congress on Arsenic in the Environment (As2016), June 19-23, 2016, Stockholm, Sweden CRC Press. Black, A. and Craw, D. (2001) Arsenic, copper and zinc occurrence at the Wangaloa coal mine, southeast Otago, New Zealand. Int. J. Coal Geol. 45:

197 181–193. Bonato, P., Batista, M.B., Camilios-Neto, D., Pankievicz, V.C.S., Tadra-Sfeir, M.Z., Monteiro, R.A., et al. (2016) RNA-seq analyses reveal insights into the function of respiratory nitrate reductase of the diazotroph Herbaspirillum seropedicae. Environ. Microbiol. 18: 2677–2688. Bostick, B.C. and Fendorf, S. (2003) Arsenite sorption on troilite (FeS) and pyrite (FeS2). Geochim. Cosmochim. Acta 67: 909–921. El Bouhaddani, S., Houwing-Duistermaat, J., Salo, P., Perola, M., Jongbloed, G., and Uh, H.-W. (2016) Evaluation of O2PLS in Omics data integration. In, BMC bioinformatics. BioMed Central Ltd, p. 11. Bouvet, P.J.M. and Grimont, P. a. D. (1986) Taxonomy of the Genus Acinetobacter with the Recognition of Acinetobacter baumannii sp. nov., Acinetobacter haemolyticus sp. nov., Acinetobacter johnsonii sp. nov., and Acinetobacter junii sp. nov. and Emended Descriptions of Acinetobacter calcoaceticus a. Int. J. Syst. Bacteriol. 36: 228–240. Bowell, R., Alpers, C., Jamieson, H., Nordstrom, K., & Majzlan, J. (Eds.). (2014). Arsenic: Environmental geochemistry, mineralogy, and microbiology (Vol. 79). Walter de Gruyter GmbH & Co KG. Bowman, J.P., Nichols, D.S., and McMeekin, T. a. (1997) Psychrobacter glacincola sp. nov., a Halotolerant, Psychrophilic Bacterium Isolated from Antarctic Sea Ice. Syst. Appl. Microbiol. 20: 209–215. Branco, R., Francisco, R., Chung, A.P., and Morais, P.V. (2009) Identification of an aox system that requires cytochrome c in the highly arsenic-resistant bacterium Ochrobactrum tritici SCII24. Appl. Environ. Microbiol. 75: 5141– 5147. Breit, G.N., Whitney, J., Foster, A., Welch, A.H., Yount, J., and Sanzolone, R. (2001) Preliminary evaluation of arsenic cycling in the sediments of Bangladesh. In, U.S. Geol. Survey Workshop - Arsenic in the environment, Denver, CO, Feb. 21-22, 2001, wwwbrr.cr.usgs.gov/Arsenic/., pp. 21–22. Bučková, M., Puškarová, A., Chovanová, K., Kraková, L., Ferianc, P., and Pangallo, D. (2013) A simple strategy for investigating the diversity and hydrocarbon degradation abilities of cultivable bacteria from contaminated soil. World J. Microbiol. Biotechnol. 29: 1085–1098. Bundschuh, J., Litter, M.I., Parvez, F., Román-Ross, G., Nicolli, H.B., Jean, J.-S., et al. (2012) One century of arsenic exposure in Latin America: a review of history and occurrence from 14 countries. Sci. Total Environ. 429: 2–35. Buschmann, J. and Berg, M. (2009) Impact of sulfate reduction on the scale of arsenic contamination in groundwater of the Mekong, Bengal and Red River deltas. Appl. Geochemistry 24: 1278–1286. Buse, H. Y., Lu, J., Lu, X., Mou, X., & Ashbolt, N. J. (2014). Microbial diversities (16S and 18S rRNA gene pyrosequencing) and environmental pathogens within drinking water biofilms grown on the common premise plumbing

198 materials unplasticized polyvinylchloride and copper. FEMS microbiology ecology, 88(2), 280-295. Butcher, B.G., Deane, S.M., and Rawlings, D.E. (2000) The chromosomal arsenic resistance genes of Thiobacillus ferrooxidans have an unusual arrangement and confer increased arsenic and antimony resistance to Escherichia coli. Appl. Environ. Microbiol. 66: 1826–1833. Cai, L., Liu, G., Rensing, C., & Wang, G. (2009). Genes involved in arsenic transformation and resistance associated with different levels of arsenic- contaminated soils. BMC microbiology, 9(1), 4. Campbell, K.M., Malasarn, D., Saltikov, C.W., Newman, D.K., and Hering, J.G. (2006) Simultaneous microbial reduction of iron (III) and arsenic (V) in suspensions of hydrous ferric oxide. Environ. Sci. Technol. 40: 5950–5955. Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E.K., et al. (2010a) QIIME allows analysis of high- throughput community sequencing data Intensity normalization improves color calling in SOLiD sequencing. Nat. Publ. Gr. 7: 335–336. Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E.K., et al. (2010b) QIIME allows analysis of high-throughput community sequencing data (Supplementary). Nat. Methods 7: 335. Caporaso, J. G., Lauber, C. L., Walters, W. A., Berg-Lyons, D., Lozupone, C. A., Turnbaugh, P. J., ... & Knight, R. (2011). Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proceedings of the National Academy of Sciences, 108(Supplement 1), 4516-4522. Carro, L., Rivas, R., León-Barrios, M., González-Tirante, M., Velázquez, E., and Valverde, A. (2012) Herbaspirillum canariense sp. nov., Herbaspirillum aurantiacum sp. nov. and Herbaspirillum soli sp. nov., isolated from volcanic mountain soil, and emended description of the genus Herbaspirillum. Int. J. Syst. Evol. Microbiol. 62: 1300–1306. Carvalhais, L.C. and Schenk, P.M. (2013) Sample processing and cDNA preparation for microbial metatranscriptomics in complex soil communities. Methods Enzym. 531: 251–267. Cervantes, C., Ji, G., Ramirez, J., and Silver, S. (1994) Resistance to arsenic compounds in microorganisms. FEMS Microbiol. Rev. 15: 355–367. Chang, J.-S., Kim, Y.-H., and Kim, K.-W. (2008) The ars genotype characterization of arsenic-resistant bacteria from arsenic-contaminated gold–silver mines in the Republic of Korea. Appl. Microbiol. Biotechnol. 80: 155–165. Charlet, L. and Polya, D. a. (2006) Arsenic in shallow, reducing groundwaters in Southern Asia: An environmental health disaster. Elements 2: 91–96. Chaturvedi, P. and Shivaji, S. (2006) Exiguobacterium indicum sp. nov., a psychrophilic bacterium from the Hamta glacier of the Himalayan mountain ranges of India. Int. J. Syst. Evol. Microbiol. 56: 2765–2770. Chauhan, N.S., Ranjan, R., Purohit, H.J., Kalia, V.C., and Sharma, R. (2008)

199 Identification of genes conferring arsenic resistance to Escherichia coli from an effluent treatment plant sludge metagenomic library. FEMS Microbiol. Ecol. 67: 130–139. Chawla, S., Mutenda, E. K., Dixon, H. B., Freeman, S., & Smith, A. W. (1995). Synthesis of 3-arsonopyruvate and its interaction with phosphoenolpyruvate mutase. Biochemical Journal, 308(3), 931-935. Chen, J., Bhattacharjee, H., and Rosen, B.P. (2015) ArsH is an organoarsenical oxidase that confers resistance to trivalent forms of the herbicide monosodium methylarsenate and the poultry growth promoter roxarsone. Mol. Microbiol. 96: 1042–1052. Chen, J., Madegowda, M., Bhattacharjee, H., and Rosen, B.P. (2015) ArsP: a methylarsenite efflux permease. Mol. Microbiol. 98: 625–635. Chen, J., Yoshinaga, M., Garbinski, L.D., and Rosen, B.P. (2016) Synergistic interaction of glyceraldehydes‐3‐phosphate dehydrogenase and ArsJ, a novel organoarsenical efflux permease, confers arsenate resistance. Mol. Microbiol. 100: 945–953. Chen, L., Hu, M., Huang, L., Hua, Z., Kuang, J., Li, S., and Shu, W. (2015) Comparative metagenomic and metatranscriptomic analyses of microbial communities in acid mine drainage. ISME J. 9: 1579. Cheng, Z., Zheng, Y., Mortlock, R., and Van Geen, a (2004) Rapid multi-element analysis of groundwater by high-resolution inductively coupled plasma mass spectrometry. Anal. Bioanal. Chem. 379: 512–8. Chenna, R., Sugawara, H., Koike, T., Lopez, R., Gibson, T.J., Higgins, D.G., and Thompson, J.D. (2003) Multiple sequence alignment with the Clustal series of programs. Nucleic Acids Res. 31: 3497–3500. Cherry, J.A., Shaikh, A.U., Tallman, D.E., and Nicholson, R. V. (1979) Arsenic Species as an Indicator of Redox Conditions in Groundwater. Dev. Water Sci. 12: 373–392. Chistoserdova, L. (2010) Recent progress and new challenges in metagenomics for biotechnology. Biotechnol. Lett. 32: 1351–9. Chitpirom, K., Akaracharanya, A., Tanasupawat, S., Leepipatpiboon, N., and Kim, K.W. (2009) Isolation and characterization of arsenic resistant bacteria from tannery wastes and agricultural soils in Thailand. Ann. Microbiol. 59: 649– 656. Chrysostomou, C., Quandt, E.M., Marshall, N.M., Stone, E., and Georgiou, G. (2015) An Alternate Pathway of Arsenate Resistance in E. coli Mediated by the Glutathione S-Transferase GstB. ACS Chem. Biol. 10: 875–882. Coker, V.S., Gault, a G., Pearce, C.I., van der Laan, G., Telling, N.D., Charnock, J.M., et al. (2006) XAS and XMCD evidence for species-dependent partitioning of arsenic during microbial reduction of ferrihydrite to magnetite. Environ. Sci. Technol. 40: 7745–50. Colbourn, P., Alloway, B.J., and Thornton, I. (1975) Arsenic and heavy metals in

200 soils associated with regional geochemical anomalies in South-West England. Sci. Total Environ. 4: 359–363. Corsini, A., Cavalca, L., Crippa, L., Zaccheo, P., and Andreoni, V. (2010) Impact of glucose on microbial community of a soil containing pyrite cinders: role of bacteria in arsenic mobilization under submerged condition. Soil Biol. Biochem. 42: 699–707. Cotton, F. A., & Wilkinson, G. (1988). Advanced inorganic chemistry (Vol. 594, p. 1962). New York: Wiley. Couture, R.M. and Van Cappellen, P. (2011) Reassessing the role of sulfur geochemistry on arsenic speciation in reducing environments. J. Hazard. Mater. 189: 647–652. Cozen, A.E., Weirauch, M.T., Pollard, K.S., Bernick, D.L., Stuart, J.M., and Lowe, T.M. (2009) Transcriptional map of respiratory versatility in the hyperthermophilic crenarchaeon Pyrobaculum aerophilum. J. Bacteriol. 191: 782–794. Cullen, W.R. and Reimer, K.J. (1989) Arsenic speciation in the environment. Chem. Rev. 89: 713–764. Cummings, D.E., March, A.W., Bostick, B., Spring, S., Jr, F.C., Fendorf, S., et al. (2000) Evidence for Microbial Fe ( III ) Reduction in Anoxic , Mining- Impacted Lake Sediments Evidence for Microbial Fe ( III ) Reduction in Anoxic , Mining- Impacted Lake Sediments ( Lake Coeur d ’ Alene , Idaho ). Appl. Environ. Microbiol. 66: 154–162. Cummings, D.E., Snoeyenbos-West, O.L., Newby, D.T., Niggemyer, A.M., Lovley, D.R., Achenbach, L.A., and Rosenzweig, R.F. (2003) Diversity of Geobacteraceae species inhabiting metal-polluted freshwater lake sediments ascertained by 16S rDNA analyses. Microb. Ecol. 46: 257–269. Cummings D.E., S.-W.O.L. (2002) Diversity of Geobacteraceae Species Inhabiting Metal-Polluted Freshwater Lake Sediments Ascertained by 16S rDNA Analysis. Microb. Ecol. 44: 95–106. Daims, H., Lebedeva, E.V., Pjevac, P., Han, P., Herbold, C., Albertsen, M., et al. (2015) Complete nitrification by Nitrospira bacteria. Nature 528: 504–509. Datta, S., Mailloux, B., Jung, H.-B., Hoque, M. a, Stute, M., Ahmed, K.M., and Zheng, Y. (2009) Redox trapping of arsenic during groundwater discharge in sediments from the Meghna riverbank in Bangladesh. Proc. Natl. Acad. Sci. U. S. A. 106: 16930–16935. Delavat, F., Lett, M.-C., and Lievremont, D. (2012) Novel and unexpected bacterial diversity in an arsenic-rich ecosystem revealed by culture- dependent approaches. Biol. Direct 7: 28. Desoeuvre, A., Casiot, C., and Héry, M. (2016) Diversity and Distribution of Arsenic-Related Genes Along a Pollution Gradient in a River Affected by Acid Mine Drainage. Microb. Ecol. 71: 672–685. Dhar, R.K., Zheng, Y., Saltikov, C.W., Radloff, K.A., Mailloux, B.J., Ahmed, K.M.,

201 and Van Geen, A. (2011) Microbes enhance mobility of arsenic in pleistocene aquifer sand from Bangladesh. Environ. Sci. Technol. 45: 2648–2654. Dhar, R.K., Zheng, Y., Stute, M., van Geen, a., Cheng, Z., Shanewaz, M., et al. (2008) Temporal variability of groundwater chemistry in shallow and deep aquifers of Araihazar, Bangladesh. J. Contam. Hydrol. 99: 97–111. Dixit, S. and Hering, J.G. (2003) Comparison of arsenic(V) and arsenic(III) sorption onto iron oxide minerals: Implications for arsenic mobility. Environ. Sci. Technol. 37: 4182–4189. Dobbin, P.S., Warren, L.H., Cook, N.J., McEwan, A.G., Powell, A.K., and Richardson, D.J. (1996) Dissimilatory iron (III) reduction by Rhodobacter capsulatus. Microbiology 142: 765–774. Dombrowski, P.M., Long, W., Farley, K.J., Mahony, J.D., Capitani, J.F., and Di Toro, D.M. (2005) Thermodynamic analysis of arsenic methylation. Environ. Sci. Technol. 39: 2169–2176. Dong, G., Huang, Y., Yu, Q., Wang, Y., Wang, H., He, N., & Li, Q. (2014). Role of nanoparticles in controlling arsenic mobilization from sediments near a realgar tailing. Environmental science & technology, 48(13), 7469-7476. Dowdle, P.R., Laverman, A.M., and Oremland, R.S. (1996) Bacterial dissimilatory reduction of arsenic (V) to arsenic (III) in anoxic sediments. Appl. Environ. Microbiol. 62: 1664–1669. Drewniak, L., & Sklodowska, A. (2013). Arsenic-transforming microbes and their role in biomining processes. Environmental Science and Pollution Research, 20(11), 7728-7739. Drewniak, L., Stasiuk, R., Uhrynowski, W., & Sklodowska, A. (2015). Shewanella sp. O23S as a driving agent of a system utilizing dissimilatory arsenate- reducing bacteria responsible for self-cleaning of water contaminated with arsenic. International journal of molecular sciences, 16(7), 14409-14427. Duquesne, K., Lebrun, S., Casiot, C., Bruneel, O., Personné, J. C., Leblanc, M., ... & Bonnefoy, V. (2003). Immobilization of arsenite and ferric iron by Acidithiobacillus ferrooxidans and its relevance to acid mine drainage. Applied and Environmental Microbiology, 69(10), 6165-6173. Duval, S., Ducluzeau, A.-L., Nitschke, W., and Schoepp-Cothenet, B. (2008) Enzyme phylogenies as markers for the oxidation state of the environment: the case of respiratory arsenate reductase and related enzymes. BMC Evol. Biol. 8: 206. Edgar, R.C. (2004) MUSCLE: a multiple sequence alignment method with reduced time and space complexity. BMC Bioinformatics 5: 113. Edgar, R.C. (2004) MUSCLE: Multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32: 1792–1797. Edgar, R.C. (2010) Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26: 2460–2461.

202 Edgar, R.C. (2013) UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat. Methods 10: 996–998. Edwards, M. (1994). Chemistry of arsenic removal during coagulation and Fe-Mn oxidation. Journal of the American Water Works Association;(United States), 86(9). Edwardson, C.F. and Hollibaugh, J.T. (2017) Metatranscriptomic analysis of prokaryotic communities active in sulfur and arsenic cycling in Mono Lake, California, USA. ISME J. 1–14. Efron, B. (1981) Nonparametric estimates of standard error: the jackknife, the bootstrap and other methods. Biometrika 68: 589–599. Ehrich, S., Behrens, D., Lebedeva, E. V, Ludwig, W., and Bock, E. (1995) A New Obligately Chemolithoautotrophic, Nitrite-Oxidizing Bacterium, Nitrospira- Moscoviensis Sp-Nov and Its Phylogenetic Relationship. Arch. Microbiol. 164: 16–23. Emerson, D., Field, E.K., Chertkov, O., Davenport, K.W., Goodwin, L., Munk, C., et al. (2013) Comparative genomics of freshwater Fe-oxidizing bacteria: Implications for physiology, ecology, and systematics. Front. Microbiol. 4: 1– 17. Emerson, D., Fleming, E.J., and McBeth, J.M. (2010) Iron-oxidizing bacteria: an environmental and genomic perspective. Annu. Rev. Microbiol. 64: 561–583. Escudero, L. V, Casamayor, E.O., Chong, G., Pedrós-Alió, C., and Demergasso, C. (2013) Distribution of Microbial Arsenic Reduction, Oxidation and Extrusion Genes along a Wide Range of Environmental Arsenic Concentrations. PLoS One 8: e78890. Del Fabbro, C., Scalabrin, S., Morgante, M., and Giorgi, F.M. (2013) An extensive evaluation of read trimming effects on illumina NGS data analysis. PLoS One 8: 1–13. Ferguson, J.F. and Gavis, J. (1972) A review of the arsenic cycle in natural waters. Water Res. 6: 1259–1274. Fernández, M., Morel, B., Ramos, J.L., and Krell, T. (2016) Paralogous Regulators ArsR1 and ArsR2 of Pseudomonas putida KT2440 as a Basis for Arsenic Biosensor Development. Appl. Environ. Microbiol. 82: 4133–4144. Finn, R.D., Coggill, P., Eberhardt, R.Y., Eddy, S.R., Mistry, J., Mitchell, A.L., et al. (2016) The Pfam protein families database: Towards a more sustainable future. Nucleic Acids Res. 44: D279–D285. Fisher, E., Dawson, A.M., Polshyna, G., Lisak, J., Crable, B., Perera, E., et al. (2008) Transformation of inorganic and organic arsenic by Alkaliphilus oremlandii sp. nov. strain OhILAs. Ann. N. Y. Acad. Sci. 1125: 230–41. Flemming, H.-C. and Wingender, J. (2010) The biofilm matrix. Nat. Rev. Microbiol. 8: 623. Focardi, S., Pepi, M., Ruta, M., Marvasi, M., Bernardini, E., Gasperini, S., and

203 Focardi, S.E. (2010) Arsenic precipitation by an anaerobic arsenic-respiring bacterial strain isolated from the polluted sediments of Orbetello Lagoon, Italy. Lett. Appl. Microbiol. 51: 578–585. Francis, C.A., O’Mullan, G.D., and Ward, B.B. (2003) Diversity of ammonia monooxygenase (amoA) genes across environmental gradients in Chesapeake Bay sediments. Geobiology 1: 129–140. Frankenberger Jr, W.T. (2001) Environmental chemistry of arsenic CRC Press. Freikowski, D., Winter, J., and Gallert, C. (2010) Hydrogen formation by an arsenate-reducing Pseudomonas putida, isolated from arsenic-contaminated groundwater in West Bengal, India. Appl Microbiol Biotechnol 88: 1363–1371. Fresquet, M., Jowitt, T.A., Gummadova, J., Collins, R., O’Cualain, R., McKenzie, E.A., et al. (2015) Identification of a Major Epitope Recognized by PLA2R Autoantibodies in Primary Membranous Nephropathy. J. Am. Soc. Nephrol. 26: 302–313. Frias-Lopez, J., Shi, Y., Tyson, G.W., Coleman, M.L., Schuster, S.C., Chisholm, S.W., and Delong, E.F. (2008) Microbial community gene expression in ocean surface waters. Proc. Natl. Acad. Sci. U. S. A. 105: 3805–10. Fritz, I., Strömpl, C., Nikitin, D.I., Lysenko, A.M., and Abraham, W.R. (2005) Brevundimonas mediterranea sp. nov., a non-stalked species from the Mediterranean Sea. Int. J. Syst. Evol. Microbiol. 55: 479–486. Frohne, T., Rinklebe, J., Diaz-Bone, R.A., and Du Laing, G. (2011) Controlled variation of redox conditions in a floodplain soil: impact on metal mobilization and biomethylation of arsenic and antimony. Geoderma 160: 414–424. Gallego, V., Sánchez-Porro, C., García, M.T., and Ventosa, A. (2006) Massilia aurea sp. nov., isolated from drinking water. Int. J. Syst. Evol. Microbiol. 56: 2449–2453. Garelick, H., & Jones, H. (Eds.). (2008). Reviews of Environmental Contamination Volume 197: Arsenic Pollution and Remediation: An International Perspective (Vol. 197). Springer Science & Business Media. Garrity, G., Bell, J., and Lilburn, T. (2010) Family I. Pseudomonadaceae. Bergey´s Man. Syst. Bacteriol. vol 2B 323–379. Gault, A.G., Islam, F.S., Polya, D.A., Charnock, J.M., Boothman, C., Chatterjee, D., and Lloyd, J.R. (2005) Microcosm depth profiles of arsenic release in a shallow aquifer, West Bengal. Mineral. Mag. 69: 855–863. Ge, Y., Wang, D. Z., Chiu, J. F., Cristobal, S., Sheehan, D., Silvestre, F., ... & Wentao, H. (2013). Environmental omics: Current status and future directions. Journal of Integrated Omics, 3(2), 75-87. van Geen, a., Zheng, Y., Cheng, Z., Aziz, Z., Horneman, a., Dhar, R.K., et al. (2006) A transect of groundwater and sediment properties in Araihazar, Bangladesh: Further evidence of decoupling between As and Fe mobilization. Chem. Geol. 228: 85–96.

204 Georges, A. a, El-Swais, H., Craig, S.E., Li, W.K., and Walsh, D. a (2014) Metaproteomic analysis of a winter to spring succession in coastal northwest Atlantic Ocean microbial plankton. ISME J. 1–13. Ghosh, S. and Sar, P. (2013) Identification and characterization of metabolic properties of bacterial populations recovered from arsenic contaminated ground water of North East India (Assam). Water Res. 47: 6992–7005. Gihring, T.M., Druschel, G.K., McCleskey, R.B., Hamers, R.J., and Banfield, J.F. (2001) Rapid arsenite oxidation by Thermus aquaticus and Thermus thermophilus: field and laboratory investigations. Environ. Sci. Technol. 35: 3857–3862. Gilbert, J.A. (2010) Beyond the infinite -Tracking bacterial gene expression. Microbiol. Today 37: 82–85. Giloteaux, L., Holmes, D. E., Williams, K. H., Wrighton, K. C., Wilkins, M. J., Montgomery, A. P., ... & Long, P. E. (2013). Characterization and transcription of arsenic respiration and resistance genes during in situ uranium bioremediation. The ISME journal, 7(2), 370-383. Giri, A.K., Patel, R.K., Mahapatra, S.S., and Mishra, P.C. (2013) Biosorption of arsenic (III) from aqueous solution by living cells of Bacillus cereus. Environ. Sci. Pollut. Res. 20: 1281–1291. Goldberg, S. and Johnston, C.T. (2001) Mechanisms of Arsenic Adsorption on Amorphous Oxides Evaluated Using Macroscopic Measurements, Vibrational Spectroscopy, and Surface Complexation Modeling. J. Colloid Interface Sci. 234: 204–216. Gremion, F., Chatzinotas, A., and Harms, H. (2003) Comparative 16S rDNA and 16S rRNA sequence analysis indicates that Actinobacteria might be a dominant part of the metabolically active bacteria in heavy metal‐ contaminated bulk and rhizosphere soil. Environ. Microbiol. 5: 896–907. Grondin, K., Haimeur, A., Mukhopadhyay, R., Rosen, B.P., and Ouellette, M. (1997) Co‐amplification of the γ‐glutamylcysteine synthetase gene gsh1 and of the ABC transporter gene pgpA in arsenite‐resistant Leishmania tarentolae. EMBO J. 16: 3057–3065. Guidelines for Drinking Water Quality: Recommendations, vol.1. (1993) Geneva. Guo, H., Liu, Z., Ding, S., Hao, C., Xiu, W., and Hou, W. (2015) Arsenate reduction and mobilization in the presence of indigenous aerobic bacteria obtained from high arsenic aquifers of the Hetao basin, Inner Mongolia. Environ. Pollut. 203: 50–59. Gussin, G.N., Ronson, C.W., and Ausubel, F.M. (1986) Regulation of nitrogen fixation genes. Annu. Rev. Genet. 20: 567–91. Guyon, I., Weston, J., Barnhill, S., and Vapnik, V. (2002) Gene selection for cancer classification using support vector machines. Mach. Learn. 46: 389–422. Haft, D.H. (2001) TIGRFAMs: a protein family resource for the functional identification of proteins. Nucleic Acids Res. 29: 41–43.

205 Hamady, M., Walker, J.J., Harris, J.K., Gold, N.J., and Knight, R. (2008) Error- correcting barcoded primers for pyrosequencing hundreds of samples in multiplex. Nat. Methods 5: 235–237. Hamamura, N., Fukushima, K., and Itai, T. (2013) Identification of antimony- and arsenic-oxidizing bacteria associated with antimony mine tailing. Microbes Env. 28: 257–263. Hamamura, N., Macur, R.E., Korf, S., Ackerman, G., Taylor, W.P., Kozubal, M., et al. (2009) Linking microbial oxidation of arsenic with detection and phylogenetic analysis of arsenite oxidase genes in diverse geothermal environments B : NGB-BE. 11: 421–431. Handley, K.M., Héry, M., and Lloyd, J.R. (2009) Redox cycling of arsenic by the hydrothermal marine bacterium Marinobacter santoriniensis. Environ. Microbiol. 11: 1601–1611. Handley, K.M., McBeth, J.M., Charnock, J.M., Vaughan, D.J., Wincott, P.L., Polya, D. a., and Lloyd, J.R. (2013) Effect of iron redox transformations on arsenic solid-phase associations in an arsenic-rich, ferruginous hydrothermal sediment. Geochim. Cosmochim. Acta 102: 124–142. Harold, F.M. and Baarda, J.R. (1966) Interaction of arsenate with phosphate- transport systems in wild- type and mutant Streptococcus faecalis. J. Bacteriol. 91: 2257–2262. Hassan, Z., Sultana, M., van Breukelen, B.M., Khan, S.I., and Roling, W.F.M. (2015) Diverse arsenic- and iron-cycling microbial communities in arsenic- contaminated aquifers used for drinking water in Bangladesh. FEMS Microbiol. Ecol. 91: fiv026-fiv026. Hassan, Z., Sultana, M., Breukelen, B.M. Van, Khan, S.I., and Röling, W.F.M. (2015) Diverse arsenic- and iron-cycling microbial communities in arsenic- contaminated aquifers used for drinking water in Bangladesh. 1–42. Heinrich-Salmeron, A., Cordi, A., Brochier-Armanet, C., Halter, D., Pagnout, C., Abbaszadeh-Fard, E., et al. (2011) Unsuspected diversity of arsenite-oxidizing bacteria revealed by a widespread distribution of the aoxB gene in prokaryotes. Appl. Environ. Microbiol. AEM-02884. Herbel, M. and Fendorf, S. (2006) Biogeochemical processes controlling the speciation and transport of arsenic within iron coated sands. Chem. Geol. 228: 16–32. Hert, D.G., Fredlake, C.P., and Barron, A.E. (2008) Advantages and limitations of next-generation sequencing technologies: a comparison of electrophoresis and non-electrophoresis methods. Electrophoresis 29: 4618–4626. Héry, M., Van Dongen, B.E., Gill, F., Mondal, D., Vaughan, D.J., Pancost, R.D., et al. (2010) Arsenic release and attenuation in low organic carbon aquifer sediments from West Bengal. Geobiology 8: 155–68. Héry, M., Gault, A.G., Rowland, H. a. L., Lear, G., Polya, D. a., and Lloyd, J.R. (2008) Molecular and cultivation-dependent analysis of metal-reducing

206 bacteria implicated in arsenic mobilisation in south-east asian aquifers. Appl. Geochemistry 23: 3215–3223. Héry, M., Rizoulis, A., Sanguin, H., Cooke, D. A., Pancost, R. D., Polya, D. A., & Lloyd, J. R. (2015). Microbial ecology of arsenic‐mobilizing Cambodian sediments: lithological controls uncovered by stable‐isotope probing. Environmental microbiology, 17(6), 1857-1869. Hoeft, S.E., Blum, J.S., Stolz, J.F., Tabita, F.R., Witte, B., King, G.M., et al. (2007) Alkalilimnicola ehrlichii sp. nov., a novel, arsenite-oxidizing haloalkaliphilic gammaproteobacterium capable of chemoautotrophic or heterotrophic growth with nitrate or oxygen as the electron acceptor. Int. J. Syst. Evol. Microbiol. 57: 504–512. Hoeft, S.E., Kulp, T.R., Stolz, J.F., Hollibaugh JT, and Oremland, R.S. (2004) Dissimilatory arsenate\rreduction with sulfide as the electron donor: Experiments with Mono Lake water and \risolation of strain MLMS-1, a chemoautotrophic arsenate-respirer. Appl Env. Microbiol 70: 2741–2747. Hohmann, C., Morin, G., Ona-Nguema, G., Guigner, J.M., Brown, G.E., and Kappler, A. (2011) Molecular-level modes of As binding to Fe(III) (oxyhydr)oxides precipitated by the anaerobic nitrate-reducing Fe(II)- oxidizing Acidovorax sp. strain BoFeN1. Geochim. Cosmochim. Acta 75: 4699–4712. Hollibaugh, J.T., Budinoff, C., Hollibaugh, R. a., Ransom, B., and Bano, N. (2006) Sulfide oxidation coupled to arsenate reduction by a diverse microbial community in a soda lake. Appl. Environ. Microbiol. 72: 2043–2049. Horneman, a., van Geen, a., Kent, D. V., Mathe, P.E., Zheng, Y., Dhar, R.K., et al. (2004) Decoupling of As and Fe release to Bangladesh groundwater under reducing conditions. Part I: Evidence from sediment profiles. Geochim. Cosmochim. Acta 68: 3459–3473. Hosseini, P., Tremblay, A., Matthews, B.F., and Alkharouf, N.W. (2010) An efficient annotation and gene-expression derivation tool for Illumina Solexa datasets. BMC Res. Notes 3: 183. Hoven, R.N. and Santini, J.M. (2004) Arsenite oxidation by the heterotroph Hydrogenophaga sp . str . NT-14 : the arsenite oxidase and its physiological electron acceptor. 1656: 148–155. vanden Hoven, R.N. and Santini, J.M. (2004) Arsenite oxidation by the heterotroph Hydrogenophaga sp. str. NT-14: the arsenite oxidase and its physiological electron acceptor. Biochim. Biophys. Acta (BBA)-Bioenergetics 1656: 148–155. Hu, Z. and Gao, S. (2008) Upper crustal abundances of trace elements: A revision and update. Chem. Geol. 253: 205–221. Huang, J.-H. (2014) Impact of Microorganisms on Arsenic Biogeochemistry: A Review. Water, Air, Soil Pollut. 225: 1848. Huang, J.-H. and Matzner, E. (2007) Biogeochemistry of organic and inorganic

207 arsenic species in a forested catchment in Germany. Environ. Sci. Technol. 41: 1564–1569. Huang, J.H., Lazzaro, A., Zeyer, J., Kretzschmar, R., and Voegelin, A. (2010) Influence of arsenic surface adsorption and iron reduction on arsenic mobilization and reduction by Shewanella putrefaciens strain CN-32. In, Arsenic in Geosphere and Human Diseases; Arsenic 2010: Proceedings of the Third International Congress on Arsenic in the Environment (As-2010). CRC Press, p. 130. Huber, R., Sacher, M., Vollmann, A., Huber, H., and Rose, D. (2000) Respiration of arsenate and selenate by hyperthermophilic archaea. Syst. Appl. Microbiol. 23: 305–314. Hug, K., Maher, W.A., Stott, M.B., Krikowa, F., Foster, S., and Moreau, J.W. (2014) Microbial contributions to coupled arsenic and sulfur cycling in the acid- sulfide hot spring Champagne Pool, New Zealand. Front. Microbiol. 5. Hug, L.A., Thomas, B.C., Sharon, I., Brown, C.T., Sharma, R., Hettich, R.L., et al. (2016) Critical biogeochemical functions in the subsurface are associated with bacteria from new phyla and little studied lineages. Environ. Microbiol. 18: 159–173. Hultman, J., Waldrop, M. P., Mackelprang, R., David, M. M., McFarland, J., Blazewicz, S. J., ... & VerBerkmoes, N. C. (2015). Multi-omics of permafrost, active layer and thermokarst bog soil microbiomes. Nature, 521(7551), 208- 212. Hyatt, D., Chen, G.-L., LoCascio, P.F., Land, M.L., Larimer, F.W., and Hauser, L.J. (2010) Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinformatics 11: 119. Inskeep, W.P., Macur, R.E., Hamamura, N., Warelow, T.P., Ward, S. a, and Santini, J.M. (2007) Detection, diversity and expression of aerobic bacterial arsenite oxidase genes. Environ. Microbiol. 9: 934–43. Irlinger, F., Bimet, F., Delettre, J., Lefèvre, M., and Grimont, P. a D. (2005) Arthrobacter bergerei sp. nov. and Arthrobacter arilaitensis sp. nov., novel coryneform species isolated from the surfaces of cheeses. Int. J. Syst. Evol. Microbiol. 55: 457–62. Islam, F.S., Gault, A.G., Boothman, C., Polya, D. a, Charnock, J.M., Chatterjee, D., and Lloyd, J.R. (2004) Role of metal-reducing bacteria in arsenic release from Bengal delta sediments. Nature 430: 68–71. Islam, F.S., Pederick, R.L., Gault, A.G., Adams, L.K., Polya, D.A., Charnock, J.M., and Lloyd, J.R. (2005a) Interactions between the Fe ( III ) -Reducing Bacterium Geobacter sulfurreducens and Arsenate , and Capture of the Metalloid by Biogenic Fe ( II ). 71: 8642–8648. Islam, F.S., Pederick, R.L., Gault, A.G., Adams, L.K., Polya, D.A., Charnock, J.M., and Lloyd, J.R. (2005b) Interactions between the Fe (III)-reducing bacterium Geobacter sulfurreducens and arsenate, and capture of the metalloid by biogenic Fe (II). Appl. Environ. Microbiol. 71: 8642–8648.

208 Jackson, C.R. and Dugas, S.L. (2003) Phylogenetic analysis of bacterial and archaeal arsC gene sequences suggests an ancient, common origin for arsenate reductase. BMC Evol. Biol. 3: 18. Ji, G., Garber, E. a, Armes, L.G., Chen, C.M., Fuchs, J. a, and Silver, S. (1994) Arsenate reductase of Staphylococcus aureus plasmid pI258. Biochemistry 33: 7294–9. Ji, G. and Silver, S. (1992) Reduction of arsenate to arsenite by the ArsC protein of the arsenic resistance operon of Staphylococcus aureus plasmid pI258. Proc. Natl. Acad. Sci. U. S. A. 89: 9474–8. Jin, H., Wen, H., Yang, H., Liu, Z., Ren, Z., Che, G., and Zhao, Z. (2003) Arsenic poisoning of Bangladesh groundwater. Sci. Technol. 3: 16–18. Jin, Y. and Manabe, T. (2005) High-efficiency protein extraction from polyacrylamide gels for molecular mass measurement by matrix-assisted laser desorption/ionization-time of flight-mass spectrometry. Electrophoresis 26: 1019–1028. Jomova, K., Jenisova, Z., Feszterova, M., Baros, S., Liska, J., Hudecova, D., et al. (2011) Arsenic: toxicity, oxidative stress and human disease. J. Appl. Toxicol. 31: 95–107. Joshi, N.A. and Fass, J.N. (2011) Sickle: A sliding-window, adaptive, quality-based trimming tool for FastQ files (Version 1.33)[Software]. Jung, H.B., Bostick, B.C., and Zheng, Y. (2012) Field, experimental, and modeling study of arsenic partitioning across a redox transition in a bangladesh aquifer. Environ. Sci. Technol. 46: 1388–1395. Jung, H.B. and Zheng, Y. (2006) Enhanced recovery of arsenite sorbed onto synthetic oxides by L-ascorbic acid addition to phosphate solution: calibrating a sequential leaching method for the speciation analysis of arsenic in natural samples. Water Res. 40: 2168–2180. Kageyama, A., Morisaki, K., Omura, S., and Takahashi, Y. (2008) Arthrobacter oryzae sp. nov. and Arthrobacter humicola sp. nov. Int. J. Syst. Evol. Microbiol. 58: 53–6. Kämpfer, P., Fallschissel, K., and Avendaño-Herrera, R. (2011) Chryseobacterium chaponense sp. nov., isolated from farmed Atlantic salmon (Salmo salar). Int. J. Syst. Evol. Microbiol. 61: 497–501. Kämpfer, P., Vaneechoutte, M., Lodders, N., De Baere, T., Avesani, V., Janssens, M., et al. (2009) Description of Chryseobacterium anthropi sp. nov. to accommodate clinical isolates biochemically similar to Kaistella koreensis and Chryseobacterium haifense, proposal to reclassify Kaistella koreensis as Chryseobacterium koreense comb. nov. and emended d. Int. J. Syst. Evol. Microbiol. 59: 2421–8. Kanehisa, M. (2002) The KEGG databases at GenomeNet. Nucleic Acids Res. 30: 42–46. Kanehisa Laboratories (2017) KEGG: Kyoto Encyclopedia of Genes and Genomes.

209 http://www.kegg.jp/ (accessed 2015). 27: 29–34. Kang, D.D., Froula, J., Egan, R., and Wang, Z. (2015) MetaBAT, an efficient tool for accurately reconstructing single genomes from complex microbial communities. PeerJ 3: e1165. Kapaj, S., Peterson, H., Liber, K., and Bhattacharya, P. (2006) Human health effects from chronic arsenic poisoning–a review. J. Environ. Sci. Heal. Part A 41: 2399–2428. Karim, M.M. (2000) Arsenic in groundwater and health problems in Bangladesh. Water Res. 34: 304–310. Kashyap, D.R., Botero, L.M., Franck, W.L., Hassett, D.J., and McDermott, T.R. (2006) Complex regulation of arsenite oxidation in Agrobacterium tumefaciens. J. Bacteriol. 188: 1081–1088. Kavitha, S., Selvakumar, R., Sathishkumar, M., Swaminathan, K., Lakshmanaperumalsamy, P., Singh, A., and Jain, S.K. (2009) Nitrate removal using Brevundimonas diminuta MTCC 8486 from ground water. Water Sci. Technol. 60: 517–524. Kearse, M., Moir, R., Wilson, A., Stones-Havas, S., Cheung, M., Sturrock, S., et al. (2012) Geneious Basic: An integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics 28: 1647–1649. Kennard, R.W. and Stone, L.A. (1969) Computer aided design of experiments. Technometrics 11: 137–148. Kim, E.H., Jeong, H.-J., Lee, Y.K., Moon, E.Y., Cho, J.-C., Lee, H.K., and Hong, S.G. (2011) Actimicrobium antarcticum gen. nov., sp. nov., of the family Oxalobacteraceae, isolated from Antarctic coastal seawater. Curr. Microbiol. 63: 213–7. Kim, K.K., Lee, K.C., Oh, H.-M., Kim, M.J., Eom, M.K., and Lee, J.-S. (2008) Arthrobacter defluvii sp. nov., 4-chlorophenol-degrading bacteria isolated from sewage. Int. J. Syst. Evol. Microbiol. 58: 1916–21. Kirk, M.F., Holm, T.R., Park, J., Jin, Q., Sanford, R.A., Fouke, B.W., and Bethke, C.M. (2004) Bacterial sulfate reduction limits natural arsenic contamination in groundwater. Geology 32: 953–956. Kocar, B.D., Herbel, M.J., Tufano, K.J., and Fendorf, S. (2006) Contrasting effects of dissimilatory iron (III) and arsenic (V) reduction on arsenic retention and transport. Environ. Sci. Technol. 40: 6715–21. Kocar, B.D., Polizzotto, M.L., Benner, S.G., Ying, S.C., Ung, M., Ouch, K., et al. (2008) Integrated biogeochemical and hydrologic processes driving arsenic release from shallow sediments to groundwaters of the Mekong delta. Appl. Geochemistry 23: 3059–3071. Kozich, J.J., Westcott, S.L., Baxter, N.T., Highlander, S.K., and Schloss, P.D. (2013) Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the miseq illumina sequencing

210 platform. Appl. Environ. Microbiol. 79: 5112–5120. Krafft, T. and Macy, J.M. (1998) Purification and characterization of the respiratory arsenate reductase of Chrysiogenes arsenatis. Eur. J. Biochem. 255: 647–53. Kräutler, B. (1990) Chemistry of methylcorrinoids related to their roles in bacterial C1 metabolism. FEMS Microbiol. Lett. 87: 349–354. Kruger, M.C., Bertin, P.N., Heipieper, H.J., and Arsène-Ploetze, F. (2013) Bacterial metabolism of environmental arsenic - Mechanisms and biotechnological applications. Appl. Microbiol. Biotechnol. 97. Kuehnelt, D. and Goessler, W. (2003) Organoarsenic compounds in the terrestrial environment. Organomet. Compd. Environ. 223–275. Kulp, T.R., Hoeft, S.E., Miller, L.G., Saltikov, C., Murphy, J.N., Han, S., et al. (2006) Dissimilatory arsenate and sulfate reduction in sediments of two hypersaline, arsenic-rich soda lakes: Mono and Searles Lakes, California. Appl. Environ. Microbiol. 72: 6514–26. Kumar, S., Nei, M., Dudley, J., and Tamura, K. (2008) MEGA: A biologist-centric software for evolutionary analysis of DNA and protein sequences. Brief. Bioinform. 9: 299–306. Lane, D.J. (1991) 16S/23S rRNA sequencing, in Nucleic Acid Techniques in Bacterial Systematics. In, E. Stackebrandt,M.G. (ed). Wiley, New York, NY, pp. 115–1175. Langmead, B. and Salzberg, S.L. (2012) Fast gapped-read alignment with Bowtie 2. Nat. Methods 9: 357–9. Lara, J., Escudero González, L., Ferrero, M., Chong Díaz, G., Pedrós-Alió, C., and Demergasso, C. (2012) Enrichment of arsenic transforming and resistant heterotrophic bacteria from sediments of two salt lakes in Northern Chile. Extremophiles 16: 523–38. Lau, M.C.Y., Kieft, T.L., Kuloyo, O., Linage-Alvarez, B., van Heerden, E., Lindsay, M.R., et al. (2016) An oligotrophic deep-subsurface community dependent on syntrophy is dominated by sulfur-driven autotrophic denitrifiers. Proc. Natl. Acad. Sci. 113: E7927–E7936. Al Lawati, W.M., Rizoulis, A., Eiche, E., Boothman, C., Polya, D. a., Lloyd, J.R., et al. (2012) Characterisation of organic matter and microbial communities in contrasting arsenic-rich Holocene and arsenic-poor Pleistocene aquifers, Red River Delta, Vietnam. Appl. Geochemistry 27: 315–325. Lawson, M., Polya, D.A., Boyce, A.J., Bryant, C., and Ballentine, C.J. (2016) Tracing organic matter composition and distribution and its role on arsenic release in shallow Cambodian groundwaters. Geochim. Cosmochim. Acta 178: 160– 177. Lawson, M., Polya, D.A., Boyce, A.J., Bryant, C., Mondal, D., Shantz, A., and Ballentine, C.J. (2013) Pond-derived organic carbon driving changes in arsenic hazard found in Asian groundwaters. Environ. Sci. Technol. 47: 7085–

211 7094. Lear, G., Song, B., Gault, a G., Polya, D. a, and Lloyd, J.R. (2007) Molecular analysis of arsenate-reducing bacteria within Cambodian sediments following amendment with acetate. Appl. Environ. Microbiol. 73: 1041–8. Lehr, C.R., Polishchuk, E., Delisle, M.-C., Franz, C., and Cullen, W.R. (2003) Arsenic methylation by micro-organisms isolated from sheepskin bedding materials. Hum. Exp. Toxicol. 22: 325–334. Lett, M.-C., Muller, D., Lièvremont, D., Silver, S., and Santini, J. (2012) Unified nomenclature for genes involved in prokaryotic aerobic arsenite oxidation. J. Bacteriol. 194: 207–8. Li, B. and Logan, B.E. (2004) Bacterial adhesion to glass and metal-oxide surfaces. Colloids Surf. B. Biointerfaces 36: 81–90. Li, H., Li, M., Huang, Y., Rensing, C., and Wang, G. (2013) In silico analysis of bacterial arsenic islands reveals remarkable synteny and functional relatedness between arsenate and phosphate. Front. Microbiol. 4: 347. Li, P., Wang, Y., Dai, X., Zhang, R., Jiang, Z., Jiang, D., et al. (2015) Microbial community in high arsenic shallow groundwater aquifers in Hetao Basin of Inner Mongolia, China. PLoS One 10. Li, X. and Krumholz, L.R. (2007) Regulation of arsenate resistance in Desulfovibrio desulfuricans G20 by an arsRBCC operon and an arsC gene. J. Bacteriol. 189: 3705–11. Lin, Y.-F., Walmsley, A.R., and Rosen, B.P. (2006) An arsenic metallochaperone for an arsenic detoxification pump. Proc. Natl. Acad. Sci. 103: 15617–15622. Van Lis, R., Nitschke, W., Duval, S., and Schoepp-Cothenet, B. (2013) Arsenics as bioenergetic substrates. Biochim. Biophys. Acta - Bioenerg. 1827. Liu, G., Liu, M., Kim, E., Maaty, W.S., Bothner, B., Lei, B., et al. (2012) A periplasmic arsenite‐binding protein involved in regulating arsenite oxidation. Environ. Microbiol. 14: 1624–1634. Lloyd, J. and Oremland, R. (2006) Microbial transformations of arsenic in the environment: from soda lakes to aquifers. Elements 2: 85–90. Lloyd, J.R. (2003) Microbial reduction of metals and radionuclides. FEMS Microbiol. Rev. 27: 411–425. Lloyd, J.R., Gault, A.G., Héry, M., and Macrae, J.D. (2011) Microbial transformations of Arsenic in the subsurface. 77–90. López, D.L., Bundschuh, J., Birkle, P., Armienta, M.A., Cumbal, L., Sracek, O., et al. (2012) Arsenic in volcanic geothermal fluids of Latin America. Sci. Total Environ. 429: 57–75. Lovley, D.R. (1991) Dissimilatory Fe(III) and Mn(IV) reduction. Microbiol. Rev. 55: 259–287. Lovley, D.R., Phillips, E.J.P., Lonergan, D.J., and Widma, P.K. (1995) Fe(III) and S0 reduction by Pelobacter carbinolicus. Appl. Environ. Microbiol. 61: 2132–2138.

212 Loy, A., Küsel, K., Lehner, A., Drake, H.L., and Wagner, M. (2004) Microarray and functional gene analyses of sulfate-reducing prokaryotes in low-sulfate, acidic fens reveal cooccurrence of recognized genera and novel lineages. Appl. Environ. Microbiol. 70: 6998–7009. Macur, R.E., Jackson, C.R., Botero, L.M., McDermott, T.R., and Inskeep, W.P. (2004) Bacterial populations associated with the oxidation and reduction of arsenic in an unsaturated soil. Environ. Sci. Technol. 38: 104–11. Macy, J.M., Nunan, K., Iwigen, K.D., Dixon, D.R., Harbour, P.J., Cahill, M., and Sly, L.I. (1996) Bacterium Isolated from Gold Mine Wastewater. 1153–1157. Macy, J.M., Santini, J.M., Pauling, B. V, O’Neill, a H., and Sly, L.I. (2000) Two new arsenate/sulfate-reducing bacteria: mechanisms of arsenate reduction. Arch. Microbiol. 173: 49–57. Maeda, S., Ohki, A., Miyahara, K., Takeshita, T., and Higashi, S. (1990) Growth characteristics and arsenic metabolism of two species of arsenic‐tolerant bacteria. Appl. Organomet. Chem. 4: 245–250. Mailloux, B.J., Trembath-Reichert, E., Cheung, J., Watson, M., Stute, M., Freyer, G. a, et al. (2013) Advection of surface-derived organic carbon fuels microbial reduction in Bangladesh groundwater. Proc. Natl. Acad. Sci. U. S. A. 110: 5331–5. Malasarn, D., Keeffe, J.R., and Newman, D.K. (2008) Characterization of the arsenate respiratory reductase from Shewanella sp. strain ANA-3. J. Bacteriol. 190: 135–42. Malasarn, D., Saltikov, C.W., Campbell, K.M., Santini, J.M., Hering, J.G., and Newman, D.K. (2004) arrA Is a Reliable Marker for As (V) respiration. Science (80-. ). 4. Mallick, P. and Kuster, B. (2010) Proteomics: a pragmatic perspective. Nat. Biotechnol. 28: 695–709. Mardis, E.R. (2008) Next-generation DNA sequencing methods. Annu. Rev. Genomics Hum. Genet. 9: 387–402. Maron, P.-A., Ranjard, L., Mougel, C., and Lemanceau, P. (2007) Metaproteomics: a new approach for studying functional microbial ecology. Microb. Ecol. 53: 486–93. Marsili, E., Baron, D.B., Shikhare, I.D., Coursolle, D., Gralnick, J. a, and Bond, D.R. (2008) Shewanella secretes flavins that mediate extracellular electron transfer. Proc. Natl. Acad. Sci. U. S. A. 105: 3968–3973. Martin, M. (2009) Cutadaps. EmBnet.journal 47: 8–11. Masella, A.P., Bartram, A.K., Truszkowski, J.M., Brown, D.G., and Neufeld, J.D. (2012) PANDAseq: paired-end assembler for illumina sequences. BMC Bioinformatics 13: 31. Masscheleyn, P.H., Delaune, R.D., and Patrick, Jr., W.H. (1991) Effect of Redox Potential and pH on Arsenic Speciation and Solubility in a Contaminated

213 Soil. Environ. Sci. Technol. 25: 1414–1419. McArthur, J.M., Ravenscroft, P., Safiulla, S., and Thirlwall, M.F. (2001) Arsenic in groundwater: testing pollution mechanisms for sedimentary aquifers in Bangladesh. Water Resour. Res. 37: 109–117. McCarty, K.M., Chen, Y.-C., Quamruzzaman, Q., Rahman, M., Mahiuddin, G., Hsueh, Y.-M., et al. (2007) Arsenic methylation, GSTT1, GSTM1, GSTP1 polymorphisms, and skin lesions. Environ. Health Perspect. 115: 341–5. de Melo Oliveira, M.G., Abels, S., Zbinden, R., Bloemberg, G. V, and Zbinden, A. (2013) Accurate identification of fastidious Gram-negative rods: integration of both conventional phenotypic methods and 16S rRNA gene analysis. BMC Microbiol. 13: 162. Mestrot, A., Merle, J.K., Broglia, A., Feldmann, J., and Krupp, E.M. (2011) Atmospheric stability of arsine and methylarsines. Environ. Sci. Technol. 45: 4010–4015. Metzker, M.L. (2010) Sequencing technologies - the next generation. Nat. Rev. Genet. 11: 31–46. Meyer, B. and Kuevert, J. (2007) Phylogeny of the alpha and beta subunits of the dissimilatory adenosine-5??? -phosphosulfate (APS) reductase from sulfate- reducing prokaryotes - Origin and evolution of the dissimilatory sulfate- reduction pathway. Microbiology 153: 2026–2044. Meyer, J., Michalke, K., Kouril, T., and Hensel, R. (2008) Volatilisation of metals and metalloids: an inherent feature of methanoarchaea? Syst. Appl. Microbiol. 31: 81–87. Michalke, K., Wickenheiser, E.B., Mehring, M., Hirner, A. V, and Hensel, R. (2000) Production of volatile derivatives of metal (loid) s by microflora involved in anaerobic digestion of sewage sludge. Appl. Environ. Microbiol. 66: 2791–2796. Mihajlov, I., Stute, M., Schlosser, P., Mailloux, B.J., Zheng, Y., Choudhury, I., et al. (2016) Recharge of low‐arsenic aquifers tapped by community wells in Araihazar, Bangladesh, inferred from environmental isotopes. Water Resour. Res. 52: 3324–3349. Miller, C.S., Baker, B.J., Thomas, B.C., Singer, S.W., and Banfield, J.F. (2011) EMIRGE: reconstruction of full-length ribosomal genes from microbial community short read sequencing data. Genome Biol. 12: R44. Miller, W.H., Schipper, H.M., Lee, J.S., Singer, J., and Waxman, S. (2002) Mechanisms of Action of Arsenic Trioxide Mechanisms of Action of Arsenic Trioxide 1. 3893–3903. Mirza, B.S., Muruganandam, S., Meng, X., Sorensen, D.L., Dupont, R.R., and McLean, J.E. (2014) Arsenic(V) reduction in relation to Iron(III) transformation and molecular characterization of the structural and functional microbial community in sediments of a basin-fill aquifer in Northern Utah. Appl. Environ. Microbiol. 80: 3198–208.

214 Moriya, Y., Itoh, M., Okuda, S., and Kanehisa, M. (2005) KAAS: KEGG automatic annotation server. Genome Informatics 5: 2005. Moriya, Y., Itoh, M., Okuda, S., Yoshizawa, A.C., and Kanehisa, M. (2007) KAAS: An automatic genome annotation and pathway reconstruction server. Nucleic Acids Res. 35. Mukai, H., Ambe, Y., Muku, T., Takeshita, K., and Fukuma, T. (1986) Seasonal variation of methylarsenic compounds in airborne participate matter. Nature 324: 239–241. Mukherjee, A., Sengupta, M.K., Hossain, M.A., Ahamed, S., Das, B., Nayak, B., et al. (2006) Arsenic contamination in groundwater: a global perspective with emphasis on the Asian scenario. J. Health. Popul. Nutr. 24: 142–63. Mukhopadhyay, R. and Rosen, B.P. (2002a) Arsenate reductases in prokaryotes and eukaryotes. Environ. Health Perspect. 110: 745. Mukhopadhyay, R., Rosen, B.P., Phung, L.T., and Silver, S. (2002b) Microbial arsenic: from geocycles to genes and enzymes. FEMS Microbiol. Rev. 26: 311– 325. National Research Council (US) Committee on Medical and Biological Effects of Environmental Pollutants (1977) Arsenic: Medical and Biologic Effects of Environmental Pollutants. Newman, D.K. and Beveridge, T.J. (1997) Precipitation of Arsenic Trisul de by. Microbiology 63: 2022–2028. Newman, D.K., Beveridge, T.J., and Morel, F. (1997) Precipitation of arsenic trisulfide by Desulfotomaculum auripigmentum. Appl. Environ. Microbiol. 63: 2022–2028. Nickson, R.T., Mcarthur, J.M., Ravenscroft, P., Burgess, W.G., and Ahmed, K.M. (2000) Mechanism of arsenic release to groundwater, Bangladesh and West Bengal. Appl. Geochemistry 15: 403–413. Niggemyer, A., Spring, S., Stackebrandt, E., & Rosenzweig, R. F. (2001). Isolation and characterization of a novel As (V)-reducing bacterium: implications for arsenic mobilization and the genus Desulfitobacterium. Applied and Environmental Microbiology, 67(12), 5568-5580.Nordstrom, D.K. (2002) Public health. Worldwide occurrences of arsenic in ground water. Science (80-. ). 296: 2143–5. Nriagu, J. (2001) Arsenic Poisoning Through the Ages. In, Environmental Chemistry of Arsenic., pp. 1–26. Nyrén, P. (2007) The history of pyrosequencing. Methods Mol. Biol. 373: 1–14. O’Day, P. a, Vlassopoulos, D., Root, R., and Rivera, N. (2004) The influence of sulfur and iron on dissolved arsenic concentrations in the shallow subsurface under changing redox conditions. Proc. Natl. Acad. Sci. U. S. A. 101: 13703– 13708. Ogata, H., Goto, S., Sato, K., Fujibuchi, W., Bono, H., and Kanehisa, M. (1999)

215 KEGG: Kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 27: 29– 34. Ogram, A., Sayler, S., and Barkay, T. (1987) The extraction and purification of microbial D N A from sediments. 7: 57–66. Omoregie, E.O., Couture, R.-M., Van Cappellen, P., Corkhill, C.L., Charnock, J.M., Polya, D. a, et al. (2013) Arsenic bioremediation by biogenic iron oxides and sulfides. Appl. Environ. Microbiol. 79: 4325–35. Oremland, R. and J.F.S. (2013) The Ecology of Arsenic Author ( s ): Ronald S . Oremland and John F . Stolz Source: Science , New Series , Vol . 300 , No . 5621 ( May 9 , 2003 ), pp . 939-944. Published by American Association for the Advancement of Science Stable URL: http://www.jst. Science (80-. ). 300: 939–944. Oremland, R.S., Kulp, T.R., Blum, J.S., Hoeft, S.E., Baesman, S., Miller, L.G., and Stolz, J.F. (2005) A microbial arsenic cycle in a salt-saturated, extreme environment. Science 308: 1305–1308. Oremland, R.S. and Stolz, J.F. (2003) The ecology of arsenic. Science 300: 939–44. Osborne, T. H., McArthur, J. M., Sikdar, P. K., & Santini, J. M. (2015). Isolation of an arsenate-respiring bacterium from a redox front in an arsenic-polluted aquifer in West Bengal, Bengal Basin. Environmental science & technology, 49(7), 4193-4199. Páez-Espino, D., Tamames, J., de Lorenzo, V., and Cánovas, D. (2009) Microbial responses to environmental arsenic. Biometals 22: 117–30. Paisio, C.E., Talano, M.A., González, P.S., Busto, V.D., Talou, J.R., and Agostini, E. (2012) Isolation and characterization of a Rhodococcus strain with phenol- degrading ability and its potential use for tannery effluent biotreatment. Environ. Sci. Pollut. Res. 19: 3430–3439. Panaro, N.J., Yuen, P.K., Sakazume, T., Fortina, P., Kricka, L.J., and Wilding, P. (2000) Evaluation of DNA fragment sizing and quantification by the Agilent 2100 bioanalyzer. Clin. Chem. 46: 1851–1853. Park, H.S., Kim, B.H., Kim, H.S., Kim, H.J., Kim, G., Kim, M., et al. (2001) A Novel Electrochemically Active and Fe ( III ) -reducing Bacterium Phylogenetically Related to Clostridium butyricum Isolated from a Microbial Fuel Cell. Scanning 7: 297–306. Paul, D., Kazy, S. K., Gupta, A. K., Pal, T., & Sar, P. (2015). Diversity, metabolic properties and arsenic mobilization potential of indigenous bacteria in arsenic contaminated groundwater of West Bengal, India. PloS one, 10(3), e0118735. Pearcy, C. a., Chevis, D. a., Haug, T.J., Jeffries, H. a., Yang, N., Tang, J., et al. (2011) Evidence of microbially mediated arsenic mobilization from sediments of the Aquia aquifer, Maryland, USA. Appl. Geochemistry 26: 575–586. Pederick, R.L., Gault, a G., Charnock, J.M., Polya, D. a, and Lloyd, J.R. (2007a) Probing the biogeochemistry of arsenic: response of two contrasting aquifer

216 sediments from Cambodia to stimulation by arsenate and ferric iron. J. Environ. Sci. Health. A. Tox. Hazard. Subst. Environ. Eng. 42: 1763–74. Pederick, R.L., Gault, a G., Charnock, J.M., Polya, D. a, and Lloyd, J.R. (2007b) Probing the biogeochemistry of arsenic: response of two contrasting aquifer sediments from Cambodia to stimulation by arsenate and ferric iron. J. Environ. Sci. Health. A. Tox. Hazard. Subst. Environ. Eng. 42: 1763–74. Peng, Y., Leung, H.C.M., Yiu, S.M., and Chin, F.Y.L. (2012) IDBA-UD: A de novo assembler for single-cell and metagenomic sequencing data with highly uneven depth. Bioinformatics 28: 1420–1428. Pérez-Jiménez, J.R., DeFraia, C., and Young, L.Y. (2005) Arsenate respiratory reductase gene (arrA) for Desulfosporosinus sp. strain Y5. Biochem. Biophys. Res. Commun. 338: 825–829. Pichler, T., Veizer, J., and Hall, G.E.M. (1999) The natural input of extremely high arsenic concentrations into a coral reef ecosystem by hydrothermal fluids and its removal by Fe(III) oxyhydroxides. Environ. Sci. Technol. 33: 1373–1378. Piotr, B., Robert, W., Grzegorz, O., Andreu, G., and Stanislaw, U. (1997) Isolation of three contiguous genes, ACR1, ACR2 and ACR3, involved in resistance to arsenic compounds in the yeast Saccharomyces cerevisiae. Yeast 13: 819–828. Polizzotto, M.L., Kocar, B.D., Benner, S.G., Sampson, M., and Fendorf, S. (2008) Near-surface wetland sediments as a source of arsenic release to ground water in Asia. Nature 454: 505. Polya, D.A., Gault, A.G., Diebe, N., Feldman, P., Rosenboom, J.W., Gilligan, E., et al. (2005) Arsenic hazard in shallow Cambodian groundwaters. Mineral. Mag. 69: 807–823. Polya, D.A. and Middleton, D.R.S. (2017) Arsenic in drinking water: sources & human exposure. Best Pract. Guid. Control Arsen. Drink. Water 1. Pourmand, N., Elahi, E., Davis, R.W., and Ronaghi, M. (2002) Multiplex Pyrosequencing. Nucleic Acids Res. 30: e31. Prasad, K.S., Srivastava, P., Subramanian, V., and Paul, J. (2011) Biosorption of As (III) ion on Rhodococcus sp. WB-12: biomass characterization and kinetic studies. Sep. Sci. Technol. 46: 2517–2525. Prieto, D.M., Devesa-Rey, R., Rubinos, D. a, Díaz-Fierros, F., and Barral, M.T. (2013) Arsenate retention by epipsammic biofilms developed on streambed sediments: influence of phosphate. Biomed Res. Int. 2013: 591-634. Qin, J., Rosen, B.P., Zhang, Y., Wang, G., Franke, S., and Rensing, C. (2006) Arsenic detoxification and evolution of trimethylarsine gas by a microbial arsenite S-adenosylmethionine methyltransferase. Proc. Natl. Acad. Sci. U. S. A. 103: 2075–2080. Rahman, M. A., & Hassler, C. (2014). Is arsenic biotransformation a detoxification mechanism for microorganisms?. Aquatic toxicology, 146, 212-219. Rambaut, A. (2009) FigTree, version 1.3. 1. Comput. Progr. Distrib. by author,

217 website http//treebioedacuk/software/figtree/[accessed January 4, 2011]. Ravenscroft, P., Brammer, H., and Richards, K. (2009) Arsenic Pollution: A Global Synthesis, RGS-IBG Book Series, A John Wiley & Sons Ltd. Publ. United Kingdom. Reddy, G.S.N., Matsumoto, G.I., Schumann, P., Stackerbrandt, E., and Shivaji, S. (2004) Psychrophilic pseudomonads from Antarctica: Pseudomonas antartica sp. nov., Pseudomonas meridiana sp. nov. and Pseudomonas proteolytica sp. nov. Int. J. Syst. Evol. Microbiol. 54: 713–719. Rhine, E.D., Ní Chadhain, S.M., Zylstra, G.J., and Young, L.Y. (2007) The arsenite oxidase genes (aroAB) in novel chemoautotrophic arsenite oxidizers. Biochem. Biophys. Res. Commun. 354: 662–667. Richards, L.A., Magnone, D., van Dongen, B.E., Ballentine, C.J., and Polya, D.A. (2015) Use of lithium tracers to quantify drilling fluid contamination for groundwater monitoring in Southeast Asia. Appl. Geochemistry 63: 190–202. Richards, L.A., Magnone, D., Sovann, C., Kong, C., Uhlemann, S., Kuras, O., et al. (2017) Science of the Total Environment High resolution profile of inorganic aqueous geochemistry and key redox zones in an arsenic bearing aquifer in Cambodia. Sci. Total Environ. 590–591: 540–553. Richards, L.A., Sültenfuß, J., Ballentine, C.J., Magnone, D., van Dongen, B.E., Sovann, C., and Polya, D.A. (2017) Tritium Tracers of Rapid Surface Water Ingression into Arsenic-bearing Aquifers in the Lower Mekong Basin, Cambodia. Procedia Earth Planet. Sci. 17: 845–848. Richey, C., Chovanec, P., Hoeft, S.E., Oremland, R.S., Basu, P., and Stolz, J.F. (2009) Biochemical and Biophysical Research Communications Respiratory arsenate reductase as a bidirectional enzyme. Biochem. Biophys. Res. Commun. 382: 298–302. Riesenfeld, C.S., Schloss, P.D., and Handelsman, J. (2004) Metagenomics: genomic analysis of microbial communities. Annu. Rev. Genet. 38: 525–52. Rittle, K.A., Drever, J.I., and Colberg, P.J.S. (1995) Precipitation of Arsenic during Bacterial Sulfate Reduction. Geomicrobiol. J. 13: 1–11. Rizoulis, A., Lawati, A.W.M. Al, Pancost, B.R.D., Polya, D.A., A, A.B.E.V.D., and A, J.R.L. (2014) Microbially mediated reduction of Fe III and As V in Cambodian sediments amended with 13 C-labelled hexadecane and kerogen. 538–546. Rizoulis, A., Al Lawati, W.M., Pancost, R.D., Polya, D.A., van Dongen, B.E., and Lloyd, J.R. (2014) Microbially mediated reduction of Fe III and As V in Cambodian sediments amended with 13 C-labelled hexadecane and kerogen. Environ. Chem. 11: 538–546. Rizoulis, A., Al Lawati, W.M., Pancost, R.D., Polya, D. a., van Dongen, B.E., and Lloyd, J.R. (2014) Microbially mediated reduction of FeIII and AsV in Cambodian sediments amended with 13C-labelled hexadecane and kerogen. Environ. Chem. 11: 538. Rodrigues, D.F., Goris, J., Vishnivetskaya, T., Gilichinsky, D., Thomashow, M.F.,

218 and Tiedje, J.M. (2006) Characterization of Exiguobacterium isolates from the Siberian permafrost. Description of Exiguobacterium sibiricum sp. nov. Extremophiles 10: 285–294. Roesch, L.F.W., Fulthorpe, R.R., Riva, A., Casella, G., Hadwin, A.K.M., Kent, A.D., et al. (2007) Pyrosequencing enumerates and contrasts soil microbial diversity. ISME J. 1: 283–90. Rognes, T., Flouri, T., Nichols, B., Quince, C., and Mahé, F. (2016) VSEARCH: a versatile open source tool for metagenomics. PeerJ 4: e2584. Ronaghi, M. (2001) Pyrosequencing Sheds Light on DNA Sequencing Pyrosequencing Sheds Light on DNA Sequencing. 3–11. Rosen, B.P. (1999) Families of arsenic transporters. Trends Microbiol. 7: 207–12. Rosenberg, H., Gerdes, R.G., and Chegwidden, K. (1977) Two systems for the uptake of phosphate in Escherichia coli. J. Bacteriol. 131: 505–511. Rowland, H. a. L., Pederick, R.L., Polya, D. a., Pancost, R.D., Van Dongen, B.E., Gault, a. G., et al. (2007) The control of organic matter on microbially mediated iron reduction and arsenic release in shallow alluvial aquifers, Cambodia. Geobiology 5: 281–292. Rowland, H. a. L., Polya, D. a., Lloyd, J.R., and Pancost, R.D. (2006) Characterisation of organic matter in a shallow, reducing, arsenic-rich aquifer, West Bengal. Org. Geochem. 37: 1101–1114. Rowland, H. a L., Boothman, C., Pancost, R., Gault, a G., Polya, D. a, and Lloyd, J.R. (2002) The role of indigenous microorganisms in the biodegradation of naturally occurring petroleum, the reduction of iron, and the mobilization of arsenite from west bengal aquifer sediments. J. Environ. Qual. 38: 1598–607. Rowland, H. a L., Gault, A.G., Lythgoe, P., and Polya, D. a. (2008) Geochemistry of aquifer sediments and arsenic-rich groundwaters from Kandal Province, Cambodia. Appl. Geochemistry 23: 3029–3046. Ryan, D. and Colleran, E. (2002) Arsenical resistance in the IncHI2 plasmids. Plasmid 47: 234–240. Saltikov, C. W., Cifuentes, A., Venkateswaran, K., & Newman, D. K. (2003). The ars detoxification system is advantageous but not required for As (V) respiration by the genetically tractable Shewanella species strain ANA- 3. Applied and Environmental Microbiology, 69(5), 2800-2809. Saltikov, C.W. and Newman, D.K. (2003) Genetic identification of a respiratory arsenate reductase. Proc. Natl. Acad. Sci. U. S. A. 100: 10983–8. Saltikov, C.W., Wildman, R.A., and Newman, D.K. (2005) Expression Dynamics of Arsenic Respiration and Detoxification in. 187: 7390–7396. Santini, J.M., Kappler, U., Ward, S.A., Honeychurch, M.J., vanden Hoven, R.N., and Bernhardt, P. V (2007) The NT-26 cytochrome c 552 and its role in arsenite oxidation. Biochim. Biophys. Acta (BBA)-Bioenergetics 1767: 189–196. Sanyal, S. K., Mou, T. J., Chakrabarty, R. P., Hoque, S., Hossain, M. A., & Sultana,

219 M. (2016). Diversity of arsenite oxidase gene and arsenotrophic bacteria in arsenic affected Bangladesh soils. AMB Express, 6(1), 21. Sardiwal, S., Santini, J.M., Osborne, T.H., and Djordjevic, S. (2010) Characterization of a two-component signal transduction system that controls arsenite oxidation in the chemolithoautotroph NT-26. FEMS Microbiol. Lett. 313: 20–28. Sarkar, Angana, Sufia K. Kazy, and Pinaki Sar. "Characterization of arsenic resistant bacteria from arsenic rich groundwater of West Bengal, India." Ecotoxicology 22.2 (2013): 363-376. Sato, T., & Kobayashi, Y. (1998). The ars Operon in the skin Element of Bacillus subtilis Confers Resistance to Arsenate and Arsenite. Journal of bacteriology, 180(7), 1655-1661. Scheff, G., Salcher, O., and Lingens, F. (1984) Trichococcus flocculiformis gen. nov. sp. nov. A new gram-positive filamentous bacterium isolated from bulking sludge. Appl. Microbiol. Biotechnol. 19: 114–119. Schirmer, M., Ijaz, U. Z., D'Amore, R., Hall, N., Sloan, W. T., & Quince, C. (2015). Insight into biases and sequencing errors for amplicon sequencing with the Illumina MiSeq platform. Nucleic acids research, 43(6), e37-e37. Schloss, P.D. and Handelsman, J. (2005) Metagenomics for studying unculturable microorganisms: cutting the Gordian knot. Genome Biol. 6: 229. Searle, B.C. (2010) Scaffold: A bioinformatic tool for validating MS/MS‐based proteomic studies. Proteomics 6: 1265–1269. Segata, N., Boernigen, D., Tickle, T.L., Morgan, X.C., Garrett, W.S., and Huttenhower, C. (2013) Computational meta’omics for microbial community studies. Mol. Syst. Biol. 9: 666. Senn, D.B. and Hemond, H.F. (2002) Nitrate controls on iron and arsenic in an urban lake. Science 296: 2373–6. Shapiro, a L., Viñuela, E., and Maizel, J. V (1967) Molecular weight estimation of polypeptide chains by electrophoresis in SDS-polyacrylamide gels. Biochem. Biophys. Res. Commun. 28: 815–820. Shariatpahani, M., Anderson, A.C., Abdelghani, A.A., and Englande, A.J. (1983) Microbial metabolism of an organic arsenical herbicide. In, Biodeterioration 5: papers presented at the 5th International Biodeterioration Symposium, Aberdeen, September, 1981/edited by TA Oxley and S. Barry. Chichester [Sussex]: Wiley, c1983. Sharma, A.K., Tjell, J.C., Sloth, J.J., and Holm, P.E. (2014) Review of arsenic contamination, exposure through water and food and low cost mitigation options for rural areas. Appl. Geochemistry 41: 11–33. Shelobolina, E.S., Vrionis, H. a., Findlay, R.H., and Lovley, D.R. (2008) Geobacter uraniireducens sp. nov., isolated from subsurface sediment undergoing uranium bioremediation. Int. J. Syst. Evol. Microbiol. 58: 1075–1078.

220 Shendure, J. and Lieberman Aiden, E. (2012) The expanding scope of DNA sequencing. Nat. Biotechnol. 30: 1084–94. Shrestha, P.M., Noll, M., and Liesack, W. (2007) Phylogenetic identity, growth- response time and rRNA operon copy number of soil bacteria indicate different stages of community succession. Environ. Microbiol. 9: 2464–2474. Silver, S., & Phung, L. T. (2005). Genes and enzymes involved in bacterial oxidation and reduction of inorganic arsenic. Applied and Environmental Microbiology, 71(2), 599-608. Slyemi, D. and Bonnefoy, V. (2012) How prokaryotes deal with arsenic(†). Environ. Microbiol. Rep. 4: 571–86. Smedley, P.. and Kinniburgh, D.. (2002) A review of the source, behaviour and distribution of arsenic in natural waters. Appl. Geochemistry 17: 517–568. Song, B., Chyun, E., Jaffé, P.R., and Ward, B.B. (2009) Molecular methods to detect and monitor dissimilatory arsenate-respiring bacteria (DARB) in sediments. FEMS Microbiol. Ecol. 68: 108–17. Song, Y., Deng, S.P., Acosta-Martínez, V., and Katsalirou, E. (2008) Characterization of redox-related soil microbial communities along a river floodplain continuum by fatty acid methyl ester (FAME) and 16S rRNA genes. Appl. soil Ecol. 40: 499–509. Sorg, T.J., Chen, A.S.C., and Wang, L. (2014) Arsenic species in drinking water wells in the USA with high arsenic concentrations. Water Res. 48: 156–169. Sovann, C. and Polya, D.A. (2014) Improved groundwater geogenic arsenic hazard map for Cambodia. Environ. Chem. 11: 595–607. Stevens, H., Brinkhoff, T., Rink, B., Vollmers, J., and Simon, M. (2007) Diversity and abundance of Gram positive bacteria in a tidal flat ecosystem. Environ. Microbiol. 9: 1810–1822. Stirling, D. and Bartlett, J.M.S. (2003) PCR Protocols Humana Press. Stolz, J.F. and Oremland, R.S. (1999) Bacterial respiration of arsenic and selenium. FEMS Microbiol. Rev. 23: 615–627. Stupperich, E. (1993) Recent advances in elucidation of biological corrinoid functions. FEMS Microbiol. Rev. 12: 349–365. Stute, M., Zheng, Y., Schlosser, P., Horneman, A., Dhar, R.K., Datta, S., et al. (2007) Hydrological control of As concentrations in Bangladesh groundwater. Water Resour. Res. 43: 1–11. Sun, J., Bostick, B.C., Mailloux, B.J., Ross, J.M., and Chillrud, S.N. (2016) Effect of oxalic acid treatment on sediment arsenic concentrations and lability under reducing conditions. J. Hazard. Mater. 311: 125–133. Sun, J., Chillrud, S.N., Mailloux, B.J., Stute, M., Singh, R., Dong, H., et al. (2016) Enhanced and stabilized arsenic retention in microcosms through the microbial oxidation of ferrous iron by nitrate. Chemosphere 144: 1106–1115. Sun, J., Quicksall, A.N., Chillrud, S.N., Mailloux, B.J., and Bostick, B.C. (2016)

221 Arsenic mobilization from sediments in microcosms under sulfate reduction. Chemosphere 153:. Sung, Y., Sung, Y., Fletcher, K.E., Fletcher, K.E., Ritalahti, K.M., Ritalahti, K.M., et al. (2006) Geobacter lovleyi. Society 72: 2775–2782. Switzer Blum, J., Burns Bindi, A., Buzzelli, J., Stolz, J., and Oremland, R. (1998) Bacillus arsenicoselenatis, sp. nov., and Bacillus selenitireducens, sp. nov.: two haloalkaliphiles from Mono Lake, California that respire oxyanions of selenium and arsenic. Arch. Microbiol. 171: 19–30. Tamaki, S. and Frankenberger Jr, W.T. (1992) Environmental biochemistry of arsenic. In, Reviews of environmental contamination and toxicology. Springer, pp. 79–110. Tisa, L.S. and Rosen, B.P. (1990) Molecular characterization of an anion pump. The ArsB protein is the membrane anchor for the ArsA protein. J. Biol. Chem. 265: 190–194. Trapnell, C., Roberts, A., Goff, L., Pertea, G., Kim, D., Kelley, D.R., et al. (2012) Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. Nat. Protoc. 7: 562–578. Trygg, J. and Wold, S. (2003) O2-PLS, a two-block (X-Y) latent variable regression (LVR) method with an integral OSC filter. J. Chemom. 17: 53–64. Trygg, J. and Wold, S. (2002) Orthogonal projections to latent structures (O-PLS). J. Chemom. 16: 119–128. Tsai, S.-L., Singh, S., and Chen, W. (2009) Arsenic metabolism by microbes in nature and the impact on arsenic remediation. Curr. Opin. Biotechnol. 20: 659–67. Turpeinen, R., Pantsar-Kallio, M., and Kairesalo, T. (2002) Role of microbes in controlling the speciation of arsenic and production of arsines in contaminated soils. Sci. Total Environ. 285: 133–145. Urich, T., Lanzén, A., Qi, J., Huson, D.H., Schleper, C., and Schuster, S.C. (2008) Simultaneous assessment of soil microbial community structure and function through analysis of the meta-transcriptome. PLoS One 3. Vesterberg, O. (1993) A short history of electrophoretic methods. Electrophoresis 14: 1243–1249. Villegas-torres, M.F., Bedoya-reina, O.C., Salazar, C., Vives-florez, M.J., and Dussan, J. (2011) International Biodeterioration & Biodegradation Horizontal arsC gene transfer among microorganisms isolated from arsenic polluted soil. Int. Biodeterior. Biodegradation 65: 147–152. Vithanage, M. (2017) Arsenic Research and global Sustainability—As 2016 Bhattacharya et al.,(Eds) ISBN 978-1-138-02941-5. Vogelstein, B. and Gillespie, D. (1979) Preparative and analytical purification of DNA from agarose. Proc. Natl. Acad. Sci. U. S. A. 76: 615–619. Wade, W. (2002) Unculturable bacteria—the uncharacterized organisms that

222 cause oral infections. J. R. Soc. Med. 95: 81. Wagner, M., Roger, A.J., Flax, J.L., Gregory, A., Stahl, D.A., and Brusseau, G.A. (1998) Phylogeny of Dissimilatory Sulfite Reductases Supports an Early Origin of Sulfate Respiration Phylogeny of Dissimilatory Sulfite Reductases Supports an Early Origin of Sulfate Respiration. J. Bacteriol. 180:. Wall, P.K., Leebens-Mack, J., Chanderbali, A.S., Barakat, A., Wolcott, E., Liang, H., et al. (2009) Comparison of next generation sequencing technologies for transcriptome characterization. BMC Genomics 10: 347. Wang, P., Sun, G., Jia, Y., Meharg, A. a, and Zhu, Y. (2014) A review on completing arsenic biogeochemical cycle: Microbial volatilization of arsines in environment. J. Environ. Sci. 26: 371–381. Wang, X.H., Zhang, C., Yang, L.L., Li, S., Zhang, Y., and Gomes-Laranjo, J. (2011) Screening and identification of microbial strains that secrete an extracellular C-7 xylosidase of taxanes. World J. Microbiol. Biotechnol. 27: 627–635. Warnecke, F. and Hess, M. (2009) A perspective: metatranscriptomics as a tool for the discovery of novel biocatalysts. J. Biotechnol. 142: 91–5. Watanabe, T., Miura, A., Iwata, T., Kojima, H., and Fukui, M. (2017) Dominance of Sulfuritalea species in nitrate‐depleted water of a stratified freshwater lake and arsenate respiration ability within the genus. Environ. Microbiol. Rep. Watts, M.J., O’Reilly, J., Marcilla, A.L., Shaw, R.A., and Ward, N.I. (2010) Field based speciation of arsenic in UK and Argentinean water samples. Environ. Geochem. Health 32: 479–490. Wauters, G., De Baere, T., Willems, A., Falsen, E., and Vaneechoutte, M. (2003) Description of Comamonas aquatica comb. nov. and Comamonas kerstersii sp. nov. for two subgroups of Comamonas terrigena and emended description of Comamonas terrigena. Int. J. Syst. Evol. Microbiol. 53: 859–862. Webster, J.G. and Nordstrom, D.K. (2003) Geothermal arsenic. In, Arsenic in ground water. Springer, pp. 101–125. Weeger, W., Lièvremont, D., Perret, M., Lagarde, F., Hubert, J.C., Leroy, M., and Lett, M.C. (1999) Oxidation of arsenite to arsenate by a bacterium isolated from an aquatic environment. Biometals 12: 141–9. Weier HU; Gray JW. (1988) A Programmable System to Perform the Polymerase Chain Reaction. Dna 7: 441–447. Whaley-Martin, K.J., Mailloux, B.J., Van Geen, A., Bostick, B.C., Silvern, R.F., Kim, C., et al. (2016) Stimulation of Microbially Mediated Arsenic Release in Bangladesh Aquifers by Young Carbon Indicated by Radiocarbon Analysis of Sedimentary Bacterial Lipids. Environ. Sci. Technol. 50:. WHO, G. (2011) Guidelines for drinking-water quality. World Heal. Organ. 216: 303–304. Williamson, A.J., Morris, K., Shaw, S., Byrne, J.M., Boothman, C., and Lloyd, J.R. (2013) Microbial reduction of Fe(III) under alkaline conditions relevant to

223 geological disposal. Appl. Environ. Microbiol. 79: 3320–3326. Wu, M. and Scott, A.J. (2012) Phylogenomic analysis of bacterial and archaeal sequences with AMPHORA2. Bioinformatics 28: 1033–1034. Xiao, K.Q., Li, L.G., Ma, L.P., Zhang, S.Y., Bao, P., Zhang, T., and Zhu, Y.G. (2016) Metagenomic analysis revealed highly diverse microbial arsenic metabolism genes in paddy soils with low-arsenic contents. Environ. Pollut. 211. Xue, X.-M., Yan, Y., Xu, H.-J., Wang, N., Zhang, X., and Ye, J. (2014) ArsH from Synechocystis sp. PCC 6803 reduces chromate and ferric iron. FEMS Microbiol. Lett. 356: 105–112. Yachdav, G., Kloppmann, E., Kajan, L., Hecht, M., Goldberg, T., Hamp, T., et al. (2014) PredictProtein - An open resource for online prediction of protein structural and functional features. Nucleic Acids Res. 42: 337–343. Yamamura, S. and Amachi, S. (2014) Microbiology of inorganic arsenic: From metabolism to bioremediation. J. Biosci. Bioeng. xx: 1–9. Yamamura, S., Yamamoto, N., Ike, M., and Fujita, M. (2005) Arsenic extraction from solid phase using a dissimilatory arsenate-reducing bacterium. J. Biosci. Bioeng. 100: 219–222. Yan, L., Yin, H., Zhang, S., Leng, F., Nan, W., and Li, H. (2010) Biosorption of inorganic and organic arsenic from aqueous solution by Acidithiobacillus ferrooxidans BY-3. J. Hazard. Mater. 178: 209–217. Ye, J., Coulouris, G., Zaretskaya, I., Cutcutache, I., Rozen, S., and Madden, T.L. (2012) Primer-BLAST: a tool to design target-specific primers for polymerase chain reaction. BMC Bioinformatics 13: 134. Yeates, C., Gillings, M.R., Davison, A.D., Altavilla, N., and Veal, D.A. (1998) Methods for microbial DNA extraction from soil for PCR amplification. Biol. Proced. Online 1: 40–47. Ying, S.C., Masue-Slowey, Y., Kocar, B.D., Griffis, S.D., Webb, S., Marcus, M. a., et al. (2013) Distributed microbially- and chemically-mediated redox processes controlling arsenic dynamics within Mn-/Fe-oxide constructed aggregates. Geochim. Cosmochim. Acta 104: 29–41. Yoshinaga, M., Cai, Y., and Rosen, B.P. (2011) Demethylation of methylarsonic acid by a microbial community. Environ. Microbiol. 13: 1205–1215. Younge, A.J., Henschel, R., Brown, J.T., von Laszewski, G., Qiu, J., and Fox, G.C. (2011) Analysis of Virtualization Technologies for High Performance Computing Environments. Cloud Comput. (CLOUD), 2011 IEEE Int. Conf. Cloud Comput. 9–16. Zafiriadis, I., Ntougias, S., Mirelis, P., Kapagiannidis, a G., and Aivasidis, a (2012) Molecular characterization of denitrifying bacteria isolated from the anoxic reactor of a modified DEPHANOX plant performing enhanced biological phosphorus removal. Water Environ. Res. 84: 475–484. Zargar, K., Conrad, A., Bernick, D.L., Lowe, T.M., Stolc, V., Hoeft, S., et al. (2012)

224 ArxA, a new clade of arsenite oxidase within the DMSO reductase family of molybdenum oxidoreductases. Environ. Microbiol. 14: 1635–1645. Zehnder, A.J.B. and Van Groenestijn, J.W. (1990) Accumulation of polyphosphate by Acinetobacter sp.: Physiology, ecology, and application. In, Novel biodegradable microbial polymers. Springer, pp. 235–243. Zhang, G., Chen, T., Chang, S., Zhang, W., Wu, X., Wu, M., et al. (2016) Complete genome sequence of Acinetobacter sp. TTH0-4, a cold-active crude oil degrading strain isolated from Qinghai-Tibet Plateau. J. Biotechnol. 226: 54– 55. Zhang, Z., Moon, H.S., Myneni, S.C.B., and Jaffé, P.R. (2017) Effect of dissimilatory iron and sulfate reduction on arsenic dynamics in the wetland rhizosphere and its bioaccumulation in wetland plants (Scirpus actus). J. Hazard. Mater. 321: 382–389. Zheng, Y., van Geen, a., Stute, M., Dhar, R., Mo, Z., Cheng, Z., et al. (2005) Geochemical and hydrogeological contrasts between shallow and deeper aquifers in two villages of Araihazar, Bangladesh: Implications for deeper aquifers as drinking water sources. Geochim. Cosmochim. Acta 69: 5203–5218. Zheng, Y., Stute, M., Van Geen, a., Gavrieli, I., Dhar, R., Simpson, H.J., et al. (2004) Redox control of arsenic mobilization in Bangladesh groundwater. Appl. Geochemistry 19: 201–214. Zhu, Y.-G., Xue, X.-M., Kappler, A., Rosen, B.P., and Meharg, A.A. (2017) Linking genes to microbial biogeochemical cycling: lessons from arsenic. Environ. Sci. Technol. Zhu, Y.-G., Yoshinaga, M., Zhao, F.-J., and Rosen, B.P. (2013) Earth Abides Arsenic Biotransformations. Annu. Rev. Earth Planet. Sci. 42: 140307200240001. Zobrist, J., Dowdle, P.R., Davis, J. a., and Oremland, R.S. (2000) Mobilization of Arsenite by Dissimilatory Reduction of Adsorbed Arsenate. Environ. Sci. Technol. 34: 4747–4753. Zul, D., Wanner, G., and Overmann, J. (2008) Massilia brevitalea sp. nov., a novel betaproteobacterium isolated from lysimeter soil. Int. J. Syst. Evol. Microbiol. 58: 1245–51.

225 Appendix I: Supporting Information-2 for Chapter 4 : Paper I Chapter-4: Paper-I- The microbial community structure and arsenic biogeochemistry in two arsenic impacted aquifers in Bangladesh

Table SS-1. Phylo-Class profile of microbes at site F

Taxon FS07-7a FS10-8 FS13-8 FS16-9 FS16-9a FS19-9 FS22-9 FS26-0

Unclassified 3.17% 5.83% 1.54% 2.85% 0.96% 2.78% 1.21% 3.39% Acidobacteria 1.59% 9.99% 0.55% 1.03% 0.65% 0.10% 0.04% 0.16% Actinobacteria 25.38% 25.78% 23.15% 11.08% 13.39% 8.85% 22.17% 6.35% Bacteroidia 0.19% 0.87% 0.04% 1.02% 0.08% 1.54% 2.36% 3.21% Flavobacteriia 0.39% 0.41% 0.09% 5.09% 0.27% 4.04% 0.11% 3.67% Chloroflexi 1.74% 6.69% 0.41% 0.88% 0.49% 0.05% 0.03% 0.13% Bacilli 1.70% 1.58% 0.43% 2.11% 0.52% 4.21% 3.04% 6.45% Clostridia 0.20% 0.51% 0.11% 0.37% 0.27% 5.07% 5.56% 9.06% Gemmatimonadetes 0.27% 1.96% 0.14% 0.21% 0.15% 0.00% 0.02% 0.02% Nitrospira 0.39% 3.53% 0.09% 0.17% 0.19% 0.00% 0.01% 0.02% Alphaproteobacteria 4.73% 10.46% 9.45% 15.33% 5.52% 8.50% 1.59% 6.83% Betaproteobacteria 34.06% 23.99% 12.52% 53.53% 54.21% 27.05% 28.95% 26.79% Deltaproteobacteria 0.72% 4.80% 0.35% 0.60% 0.45% 0.04% 0.05% 0.12% Gammaproteobacteria 25.47% 3.60% 51.12% 5.73% 22.84% 37.78% 34.86% 33.79% Total 100.00% 100.00% 100.00% 100.00% 100.00% 100.00% 100.00% 100.00%

Fig. SS-1. SiteF-16S Microbial Community Analysis

226

Table SS-2. Microbial Community profile in each depth at Site F FS07-7a Sequences: 7087 OTU % Closest Match/ Class Accession % Match Reference Present (Identities) 2, 75 20.49% Acinetobacter sp. TTH0-4 CP012608.1 99.0% (Zhang et al., 2016) Gammaproteobacteria (583/583) 0 18.82% Massilia aurea (T); type strain:AP13 AM231588 100% (Gallego et al., 2006) Betaproteobacteria (302/302) 3, 95 12.49% Arthrobacter humicola (T); KV-653 AB279890 100% (Kageyama et al., 2008) Actinobacteria (308/308) 17 9.45% Massilia brevitalea (T); byr23-80 NR_044274 98.4% (Zul et al., 2008) Betaproteobacteria (299/304) 730, 32 6.14% Actimicrobium antarcticum strain KOPRI HQ699437 95.1% (Kim et al., 2011) 25157/ Betaproteobacteria (291/306) 1 4.50% Psychrobacter glacincola strain LMG AJ430829 97.4% (Bowman et al., 1997) 21273/ Gammaproteobacteria (295/303) 9 1.75% Arthrobacter defluvii (T); type strain: AM409361 100% (Kim et al., 2008) 4C1-a/ Actinobacteria (304/304) 27 1.28% Arthrobacter sp. YM-M-25 GU220063 97.1% (Black and Craw, 2001) Actinobacteria (271/279) 13 1.23% Exiguobacterium indicum (T); type NR_042347 99.7% strain: HHS 31/ Bacilli (327/328) 20 1.09% Arthrobacter arilaitensis (T); CIP 108037; AJ609628 96.7% Re117/ Actinobacteria (297/307) 7 1.04% Brevundimonas sp. A21-69 KF220438 98.9% (Ghosh and Sar, 2013) Alphaproteobacteria (275/278)

FS10-8 Sequences: 3742 OTU % Closest Match Accession % Match Reference Present (Identities) 17 19.88% Massilia brevitalea (T); byr23-80 NR_044274 98.4% (Zul et al., 2008) Betaproteobacteria (299/304) 3,95, 18.37% Arthrobacter humicola (T); KV-653 AB279890 100% (Kageyama et al., 2008) 388, 143 Actinobacteria (308/308) 0 6.79% Massilia aurea (T); type strain:AP13 AM231588 100% (Gallego et al., 2006) Betaproteobacteria (302/302) 32 4.14% Actimicrobium antarcticum strain KOPRI HQ699437 94.8% (Kim et al., 2011) 25157/ Betaproteobacteria (291/307) 1 3.07% Psychrobacter glacincola strain LMG AJ430829 97.4% (Bowman et al., 1997) 21273/ Gammaproteobacteria (295/303) 7 3.02% Brevundimonas sp. A21-69 KF220438 98.9% (Ghosh and Sar, 2013) Alphaproteobacteria (275/278) 321 1.84% Arthrobacter oxydans strain L27 KC934768 100% (Wang et al., 2011) Actinobacteria (305/305) 9 1.63% Arthrobacter defluvii (T); type strain: AM409361 100% (Kim et al., 2008) 4C1-a/ Actinobacteria (304/304) 58 1.42% Uncultured Nitrospira sp. clone S1-62 HQ674926 100% Nitrospira (315/315) 36 1.20% Bradyrhizobium sp. LmjM3 JX514883 99.6% Alphaproteobacteria (279/280) 78 1.10% uncultured Acidobacteria bacterium; S1- HQ674937 96.8% 7 (299/309) Gammaproteobacteria

FS13-8 Sequences: 6768 OTU % Closest Match Accession % Match Reference Present (Identities)

227 1 40.90% Psychrobacter glacincola strain LMG AJ430829 97.4% (Bowman et al., 1997) 21273/ Gammaproteobacteria (295/303) 75, 16, 5 11.34% Acinetobacter sp. TTH0-4 CP012608.1 99.0% (Zhang et al., 2016) Gammaproteobacteria (583/583) 4 6.93% Actimicrobium antarcticum strain KOPRI HQ699437 95.4% (Kim et al., 2011) 25157/ Betaproteobacteria (292/306) 11, 1348 6.40% Paracoccus marcusii (T) Y12703 99.6% Alphaproteobacteria (277/278) 10, 321 5.44% Arthrobacter oxydans strain L27 KC934768 100% Actinobacteria (305/305) 3, 95 4.58% Arthrobacter humicola (T); KV-653 AB279890 100% (Kageyama et al., 2008) Actinobacteria (308/308) 22 3.62% Herbaspirillum soli HQ830498 97.4% (Carro et al., 2012) Betaproteobacteria (298/306) 9 3.21% Arthrobacter defluvii (T); type strain: AM409361 100% (Kim et al., 2008) 4C1-a/ Actinobacteria (304/304) 15 1.45% Rhodococcus sp. CS1 JF521654 99.7% (Paisio et al., 2012) Actinobacteria (301/302) 24 1.33% Rhodococcus sp. CAS916i KF442764 99.7% (Paul et al., 2015) Actinobacteria (299/300) 7 1.05% Brevundimonas sp. A21-69 KF220438 98.9% (Ghosh and Sar, 2013) Alphaproteobacteria (275/278)

FS16-9 Sequences: 8472 OTU % Closest Match Accession % Match Reference Present (Identities) 0, 1195 37.72% Massilia aurea (T); type strain:AP13 AM231588 100% (Gallego et al., 2006) Betaproteobacteria (302/302) 8 12.29% Comamonas aquatica (T); type strain: AJ430344 99.0% (Wauters et al., 2003) LMG 2370/ Betaproteobacteria (288/291) 7 7.75% Brevundimonas sp. A21-69 KF220438 98.9% (Ghosh and Sar, 2013) Alphaproteobacteria (275/278) 1 4.24% Psychrobacter glacincola strain LMG AJ430829 97.4% (Bowman et al., 1997) 21273/ Gammaproteobacteria (295/303) 3, 143 3.90% Arthrobacter humicola (T); KV-653 AB279890 100% (Kageyama et al., 2008) Actinobacteria (308/308) 12 3.73% Uncultured Brevundimonas sp. clone KF379701 98.9% (Paul et al., 2015) BAS3146/ Alphaproteobacteria (275/278) 17 3.48% Massilia brevitalea (T); byr23-80 NR_044274 98.4% (Zul et al., 2008) Betaproteobacteria (299/304) 25 2.73% Chryseobacterium chaponense (T); Sa GU345045 100% (Kämpfer et al., 2011) 1147-06/ Flavobacteria (277/277) 13 1.49% Exiguobacterium indicum (T); type NR_042347 99.7% (Chaturvedi and Shivaji, strain: HHS 31/ Bacilli (327/328) 2006) 4 1.40% Actimicrobium antarcticum strain KOPRI HQ699437 95.4% (Kim et al., 2011) 25157/ Betaproteobacteria (292/306) 15 1.12% Rhodococcus sp. CS1 JF521654 99.7% (Paisio et al., 2012) Actinobacteria (301/302) 28 1.02% Chryseobacterium haifense strain HWG- JQ684226 97.0% (Kämpfer et al., 2009) A3/ Flavobacteria (295/304)

FS16-9a Sequences: 8779 OTU % Closest Match Accession % Match Reference Present (Identities) 0 28.16% Massilia aurea (T); type strain:AP13 AM231588 100% (Gallego et al., 2006) Betaproteobacteria (302/302) 4, 32 19.41% Actimicrobium antarcticum strain KOPRI HQ699437 95.4% (Kim et al., 2011) 25157/ Betaproteobacteria (292/306) 1 17.92% Psychrobacter glacincola strain LMG AJ430829 97.4% (Bowman et al., 1997) 21273/ Gammaproteobacteria (295/303)

228 17 6.85% Massilia brevitalea (T); byr23-80 NR_044274 98.4% (Zul et al., 2008) Betaproteobacteria (299/304) 75, 5, 76 4.38% Acinetobacter sp. TTH0-4 CP012608.1 99.0% (Zhang et al., 2016) Gammaproteobacteria (583/583) 10 3.34% Arthrobacter oxydans strain L27 KC934768 100% Actinobacteria (305/305) 7 1.51% Brevundimonas sp. A21-69 KF220438 98.9% (Ghosh and Sar, 2013) Alphaproteobacteria (275/278) 15 1.42% Rhodococcus sp. CS1 JF521654 99.7% (Paisio et al., 2012) Actinobacteria (301/302)

FS19-9 Sequences: 8755 OTU % Closest Match Accession % Match Reference Present (Identities) 0 21.68% Massilia aurea (T); type strain:AP13 AM231588 100% (Gallego et al., 2006) Betaproteobacteria (302/302) 6, 283, 13.36% Acinetobacter sp. W9-3 GQ497238.1 100% (Chitpirom et al., 2009) 43 Gammaproteobacteria (304/304) 2,75, 10.88% Acinetobacter sp. TTH0-4 CP012608.1 99.0% (Zhang et al., 2016) 1193 Gammaproteobacteria (583/583) 7 4.01% Brevundimonas sp. A21-69 KF220438 98.9% (Ghosh and Sar, 2013) Alphaproteobacteria (275/278) 1 3.24% Psychrobacter glacincola strain LMG AJ430829 97.4% (Bowman et al., 1997) 21273/ Gammaproteobacteria (295/303) 19 2.38% Acidovorax temperans strain C23 HQ259690 97.7% (Zafiriadis et al., 2012) Betaproteobacteria (295/302) 3 2.17% Arthrobacter humicola (T); KV-653 AB279890 100% (Kageyama et al., 2008) Actinobacteria (308/308) 20 1.78% Arthrobacter arilaitensis (T); CIP 108037; AJ609628 96.7% Re117/ Actinobacteria (297/307) 21 1.74% Pseudomonas antarctica (T); type strain: AJ537601 99.7% CMS 35/ Gammaproteobacteria (300/301) 16 1.50% Acinetobacter johnsonii (T); ATCC Z93440 97.1% 17909T/ Gammaproteobacteria (296/305) 17 1.48% Massilia brevitalea (T); byr23-80 NR_044274 98.4% (Zul et al., 2008) Betaproteobacteria (299/304) 12 1.42% Brevundimonas sp. A21-66 KF220439 98.9% (Ghosh and Sar, 2013) Alphaproteobacteria (275/278) 13 1.40% Exiguobacterium indicum (T); type NR_042347 99.7% (Chaturvedi and Shivaji, strain: HHS 31/ Bacilli (327/328) 2006) 14 1.38% Trichococcus flocculiformis (T); DSM AJ306611 99.7% 2094/ Bacill (304/305) 49 1.25% Acinetobacter sp. G3DM-29 EU037279 100% Gammaproteobacteria (304/304) 139 1.22% Acinetobacter sp. JN18_V65_E EF059531 98.7% Gammaproteobacteria (301/305)

FS22-9 Sequences: 9380 OTU % Closest Match Accession % Match Reference Present (Identities) 2, 75 27.22% Acinetobacter sp. TTH0-4 CP012608.1 99.0% (Zhang et al., 2016) Gammaproteobacteria (583/583) 3,95, 143 15.59% Arthrobacter humicola (T); KV-653 AB279890 100% (Kageyama et al., 2008) Actinobacteria (308/308) 17 13.20% Massilia brevitalea (T); byr23-80 NR_044274 98.4% (Zul et al., 2008) Betaproteobacteria (299/304) 0 8.39% Massilia aurea (T); type strain:AP13 AM231588 100% (Gallego et al., 2006) Betaproteobacteria (302/302) 1 6.79% Psychrobacter glacincola strain LMG AJ430829 97.4% (Bowman et al., 1997) 21273/ Gammaproteobacteria (295/303)

229 32, 4 4.85% Actimicrobium antarcticum strain KOPRI HQ699437 94.8% (Kim et al., 2011) 25157/ Betaproteobacteria (291/307) 29 1.39% Pseudomonas spinosa strain ATCC AB021387 94.7% 14606T/ Betaproteobacteria (288/304) 9 1.15% Arthrobacter defluvii (T); type strain: AM409361 100% (Kim et al., 2008) 4C1-a/ Actinobacteria (304/304) 14 1.10% Trichococcus flocculiformis (T); DSM AJ306611 99.7% 2094/ Bacill (304/305)

FS26 Sequences: 9543 OTU % Closest Match Accession % Match Reference Present (Identities) 0 24.16% Massilia aurea (T); type strain:AP13 AM231588 100% (Gallego et al., 2006) Betaproteobacteria (302/302) 6,283, 43 12.78% Acinetobacter sp. W9-3 AB758600 96.4% (Chitpirom et al., 2009) Gammaproteobacteria (293/304) 2,75, 11.94% Acinetobacter sp. A32 AM285014 96.4% (Achour et al., 2007) 1193 Gammaproteobacteria (295/306) 7 3.38% Brevundimonas sp. A21-69 KF220438 98.9% (Ghosh and Sar, 2013) Alphaproteobacteria (275/278) 14 3.33% Trichococcus flocculiformis (T); DSM AJ306611 99.7% (Scheff et al., 1984) 2094/ Bacill (304/305) 21 2.06% Pseudomonas antarctica (T); type strain: AJ537601 99.7% (Reddy et al., 2004) CMS 35/ Gammaproteobacteria (300/301) 3 1.83% Arthrobacter humicola (T); KV-653 AB279890 100% (Kageyama et al., 2008) Actinobacteria (308/308) 1 1.83% Psychrobacter glacincola strain LMG AJ430829 97.4% (Bowman et al., 1997) 21273/ Gammaproteobacteria (295/303) 49, 1007 1.65% Acinetobacter sp. G3DM-29 EU037279 100% Gammaproteobacteria (304/304) 19 1.63% Acidovorax temperans strain C23 HQ259690 97.7% (Zafiriadis et al., 2012) Betaproteobacteria (295/302) 18 1.54% Clostridium sp. 6-62 AB596881 98.0% Clostridia (292/298) 139 1.46% Acinetobacter sp. JN18_V65_E EF059531 98.7% Gammaproteobacteria (301/305) 16 1.36% Acinetobacter johnsonii (T); ATCC Z93440 97.1% 17909T/ Gammaproteobacteria (296/305) 12 1.14% Brevundimonas sp. A21-66 KF220439 98.9% (Ghosh and Sar, 2013) Alphaproteobacteria (275/278) 13 1.10% Exiguobacterium indicum (T); type NR_042347 99.7% (Chaturvedi and Shivaji, strain: HHS 31/ Bacilli (327/328) 2006)

230

Table SS-3. Phylo-Class profile of microbes at site B Taxon BS07-6 BS07-6a BS10-7a BS11-0 BS13-7 BS14-0 BS16-8 Unclassified 7.52% 0.58% 2.12% 0.17% 4.47% 1.41% 1.99% Acidobacteria 6.80% 0.01% 1.12% 0.03% 0.48% 1.11% 0.46% Actinobacteria 12.83% 75.02% 25.77% 43.03% 17.49% 11.17% 11.74% Alphaproteobacteria 15.73% 0.03% 5.05% 0.42% 2.51% 3.08% 14.87% Bacilli 3.94% 18.00% 5.65% 1.50% 21.05% 9.62% 1.29% Bacteroidia 0.36% 0.08% 0.18% 0.00% 4.12% 1.24% 1.61% Betaproteobacteria 19.43% 0.11% 13.82% 3.88% 29.81% 57.91% 33.86% Chloroflexi 4.41% 0.50% 0.92% 0.10% 7.72% 1.38% 0.62% Clostridia 0.97% 0.00% 0.38% 0.02% 4.91% 4.33% 5.17% Deltaproteobacteria 4.84% 0.39% 0.83% 0.06% 2.24% 0.97% 0.52% Flavobacteriia 6.33% 0.00% 0.79% 0.06% 0.41% 0.86% 1.55% Gammaproteobacteria 10.01% 5.09% 42.41% 50.70% 4.53% 6.06% 25.66% Gemmatimonadetes 1.86% 0.00% 0.24% 0.02% 0.04% 0.20% 0.14% Nitrospira 2.99% 0.16% 0.40% 0.01% 0.16% 0.34% 0.21% Saprospirae 1.99% 0.02% 0.32% 0.00% 0.08% 0.33% 0.31% Total 100.00% 100.00% 100.00% 100.00% 100.00% 100.00% 100.00%

Fig. SS-2. SiteF-16S Microbial Community Analysis

Table SS-4. Microbial Community profile in each depth at Site B BS07-6 Sequences: 5337 OTU % Closest Match Accession % Match Reference Present (Identities ) 10 9.42% Massilia sp. 51Ha FR865961 98.0% (Bassas-Galia et al., Betaproteobacteria (298/304) 2012) 906, 4.42% Duganella sp. P-125 AM412134 98.0% (Shrestha et al., 656 Betaproteobacteria (298/304) 2007)

231 18 3.71% Chryseobacterium anthropi strain KC866181 98.0% (de Melo Oliveira et Melo65 (288/294) al., 2013) Flavobacteria 2 3.05% Psychrobacter glacincola; NF1 AJ430829 98.7% (Bowman et al., Gammaproteobacteria (304/308) 1997) 1531 2.77% Massilia brevitalea (T); byr23-80 NR_044274 97.4% (Zul et al., 2008) Betaproteobacteria (296/304) 27 2.02% Novosphingobium capsulatum strain NR_025838 98.6% GIFU 11526/ Alphaproteobacteria (273/277) 30 1.82% Altererythrobacter marensis (T); type FM177586 96.8% strain: MSW-14/ Alphaproteobacteria (269/278) 23 1.74% Chryseobacterium haifense strain JQ684226 99.7% HWG-A3/ Flavobacteria (303/304) 28 1.61% Arthrobacter protophormiae strain AY211130 99.4% Mali 36/ Actinobacteria (306/308) 34 1.33% Chryseobacterium sp. MF10_Lo HF675174 97.4% Flavobacteria (296/304) 277 1.07% Herbaspirillum sp. P-64 AM411937 98.0% Betaproteobacteria (299/305) 36 1.01% Thermomonas brevis strain R-13291 NR_025578 100% Gammaproteobacteria (310/310)

BS07-6a Sequences: 5974 OTU % Closest Match Accession % Match Reference Present 0,22, 8, 71.62% Arthrobacter humicola (T); KV-653 NR_041546 99.4% (Kageyama et al., 42 Actinobacteria (306/308) 2008)

6 17.91% Exiguobacterium sibiricum (T); 255-15 CP001022 100% (Rodrigues et al., Bacilli (328/328) 2006) 941 3.75% Acinetobacter sp. A32 AM285014 99.3% (Achour et al., Gammaproteobacteria (302/304) 2007)

BS10-7a Sequences: 7149 OTU % Closest Match Accession % Match Reference Present 2, 560 25.13% Psychrobacter glacincola; NF1 AJ430829 98.7% (Bowman et al., Gammaproteobacteria (304/308) 1997) 277, 1 20.27% Herbaspirillum sp. P-64 AM411937 98.0% (Shrestha et al., Betaproteobacteria (299/305) 2007) 3,48, 16.49% Acinetobacter sp. A32 AM285014 99.7% (Achour et al., 941, Gammaproteobacteria (303/304) 2007) 1538 0,1374,8 8.13% Arthrobacter humicola (T); KV-653 NR_041546 99.4% (Kageyama et al., Actinobacteria (306/308) 2008) 14, 7 4.94% Arthrobacter sp. GWS-BW-H53M AY370617 100% (Stevens et al., Actinobacteria (306/306) 2007) 9, 145, 3.97% Arthrobacter arilaitensis (T); CIP AJ609628 99.7% (Irlinger et al., 2005) 11 108037; Re117 (305/306) Actinobacteria 5 2.76% Trichococcus flocculiformis (T); DSM Y17301 99.1% (Scheff et al., 1984) 2094/ Bacilli (323/326) 10 1.68% Massilia sp. 51Ha FR865961 98.0% (Bassas-Galia et al., Betaproteobacteria (298/304) 2012) 20 1.43% Planococcus sp. A08 AM284990 99.7% (Achour et al., Bacilli (315/316) 2007) 16 1.02% Paracoccus marcusii (T) Y12703 99.6% Alphaproteobacteria (274/275) 28 0.88% Arthrobacter protophormiae strain AY211130 99.4% Mali 36/ Actinobacteria (306/308)

232

BS11-0 Sequences: 9424 OTU % Closest Match Accession % Match Reference Present 2, 560 39.25% Psychrobacter glacincola; NF1 AJ430829 98.7% (Bowman et al., Gammaproteobacteria (304/308) 1997) 0, 222, 19.39% Arthrobacter humicola (T); KV-653 NR_041546 99.4% (Kageyama et al., 1374, 42 Actinobacteria (306/308) 2008) 14, 7 9.31% Arthrobacter sp. GWS-BW-H53M/ AY370617 100% (Stevens et al., Actinobacteria (306/306) 2007) OYU481, 9.29% Acinetobacter sp. A32 AM285014 99.0% (Achour et al., 941, 3 Gammaproteobacteria (301/304) 2007) 145, 11, 7.22% Arthrobacter arilaitensis (T); CIP AJ609628 97.7% (Irlinger et al., 2005) 9 108037; Re117 (299/306) Actinobacteria 277, 1 3.71% Herbaspirillum sp. P-64 AM411937 98.0% (Shrestha et al., Betaproteobacteria (299/305) 2007) 15 1.84% Herbaspirillum sp. SUEMI10 HQ830498 98.4% Actinobacteria (299/304) 28 1.22% Arthrobacter protophormiae strain AY211130 99.4% Mali 36/ Actinobacteria (306/308) 35 1.15% Arthrobacter sp. DSM 20546 X80742 98.4% Actinobacteria (301/306)

BS13-7 Sequences: 7375 OTU % Closest Match Accession % Match Reference Present 277 32.92% Herbaspirillum sp. P-64 AM411937 98.0% (Shrestha et al., Betaproteobacteria (299/305) 2007) 5 24.07% Trichococcus flocculiformis (T); DSM Y17301 99.1% (Scheff et al., 1984) 2094/ Bacilli (323/326) 0 13.65% Arthrobacter humicola (T); KV-653/ NR_041546 99.4% (Kageyama et al., Actinobacteria (306/308) 2008) 481, 3, 3.20% Acinetobacter sp. A32 AM285014 99.0% (Achour et al., 941 Gammaproteobacteria (301/304) 2007)

BS14-0 Sequences: 6400 OTU % Closest Match Accession % Match Reference Present 1531 43.70% Massilia brevitalea (T); byr23-80/ NR_044274 97.4% (Zul et al., 2008) Betaproteobacteria (296/304) 277 19.27% Herbaspirillum sp. P-64 AM411937 98.0% (Shrestha et al., Betaproteobacteria (299/305) 2007) 5 8.16% Trichococcus flocculiformis (T); DSM Y17301 99.1% (Scheff et al., 1984) 2094/ Bacilli (323/326) 0 7.57% Arthrobacter humicola (T); KV-653/ NR_041546 99.4% (Kageyama et al., Actinobacteria (306/308) 2008) 481, 3.48% Acinetobacter sp. A32 AM285014 99.0% (Achour et al., 941, 3 Gammaproteobacteria (301/304) 2007) 21, 52, 1.16% Clostridium sp. 6-62 AB596881 98.7% 376 Clostridia (293/297)

BS16-8 Sequences: 5781 OTU % Closest Match Accession % Match Reference Present 277, 1 65.34% Herbaspirillum sp. P-64 AM411937 98.0% (Shrestha et al., Betaproteobacteria (299/305) 2007) 481,941, 13.23% Acinetobacter sp. A32 AM285014 99.0% (Achour et al., 1538, 3 Gammaproteobacteria (301/304) 2007)

233 13 3.46% Acinetobacter johnsonii (T); ATCC Z93440 97.3% (Bouvet and 17909T/ Gammaproteobacteria (293/301) Grimont, 1986) 22 3.17% Acinetobacter calcoaceticus strain JX295977 99.7% (Bučková et al., MST_E1/ Gammaproteobacteria (303/304) 2013) 16 2.89% Paracoccus marcusii (T) Y12703 100% Alphaproteobacteria (278/278) 12 2.89% Brevundimonas mediterranea (T); AJ227801 98.2% (Fritz et al., 2005) V4.BO.10/ Alphaproteobacteria (278/283) 10 1.59% Massilia sp. 51Ha FR865961 98.0% Betaproteobacteria (298/304) 0 1.40% Arthrobacter humicola (T); KV-653/ NR_041546 99.4% Actinobacteria (306/308) 29 1.07% Brevundimonas sp. A21-66 KF220439 99.3% (Ghosh and Sar, Alphaproteobacteria (276/278 2013) 17 1.00% Arthrobacter sp. SaCR11 JX233514 98.7% Actinobacteria (300/304) 21 0.88% Clostridium sp. 6-62 AB596881 98.7% Clostridia (293/297) 26 0.73% Porphyrobacter donghaensis (T); SW- AY559428 99.3% 132/ Alphaproteobacteria (276/278)

Reference

Achour-Rokbani, A., Cordi, A., Poupin, P., Bauda, P., and Billard, P. (2010) Characterization of the ars gene cluster from extremely arsenic-resistant Microbacterium sp. strain A33. Appl. Environ. Microbiol. 76: 948–55.

Achour, A.R., Bauda, P., and Billard, P. (2007) Diversity of arsenite transporter genes from arsenic-resistant soil bacteria. Res. Microbiol. 158: 128–37.

Afkar, E., Lisak, J., Saltikov, C., Basu, P., Oremland, R.S., and Stolz, J.F. (2003) The respiratory arsenate reductase from Bacillus selenitireducens strain MLS10. FEMS Microbiol. Lett. 226: 107–112.

Ahman (1994) Microbe grows by reducing arsenic. 750.

Akay, C., Thomas, C., and Gazitt, Y. (2004) Arsenic trioxide and paclitaxel induce apoptosis by different mechanisms. Cell Cycle 3: 324–334.

Altschul, S.F., Gish, W., Miller, W., Myers, E.W., and Lipman, D.J. (1990) Basic local alignment search tool. J. Mol. Biol. 215: 403–10.

Amend, J.P., Saltikov, C., Lu, G.-S., and Hernandez, J. (2014) Microbial arsenic metabolism and reaction energetics. Rev. Mineral. Geochemistry 79: 391–433.

Anderson, C.R., Condron, L.M., Clough, T.J., Fiers, M., Stewart, A., Hill, R.A., and Sherlock, R.R. (2011) Biochar induced soil microbial community change: Implications for biogeochemical cycling of carbon, nitrogen and phosphorus. Pedobiologia (Jena). 54: 309–320.

Anderson, C.R. and Cook, G.M. (2004) Isolation and characterization of arsenate-reducing bacteria from arsenic-contaminated sites in New Zealand. Curr. Microbiol. 48: 341–347.

Anderson, G.L., Williams, J., and Hille, R. (1992) The purification and characterization of arsenite oxidase from Alcaligenes faecalis, a molybdenum-containing hydroxylase. J. Biol. Chem. 267: 23674–82.

Andrade, J.M., Gómez-Carracedo, M.P., Krzanowski, W., and Kubista, M. (2004) Procrustes

234 rotation in analytical chemistry, a tutorial. Chemom. Intell. Lab. Syst. 72: 123–132.

Andres, J. and Bertin, P.N. (2016) The microbial genomics of arsenic. FEMS Microbiol. Rev. fuv050.

Andrews, S. (2010) FastQC: a quality control tool for high throughput sequence data. software.

Argos, M., Kalra, T., Rathouz, P.J., Chen, Y., Pierce, B., Parvez, F., et al. (2010) Arsenic exposure from drinking water, and all-cause and chronic-disease mortalities in Bangladesh (HEALS): A prospective cohort study. Lancet 376: 252–258.

Bahar, M.M., Megharaj, M., and Naidu, R. (2012) Arsenic bioremediation potential of a new arsenite-oxidizing bacterium Stenotrophomonas sp. MM-7 isolated from soil. Biodegradation 23: 803–812.

Bankevich, A., Nurk, S., Antipov, D., Gurevich, A.A., Dvorkin, M., Kulikov, A.S., et al. (2012) SPAdes: A New Genome Assembly Algorithm and Its Applications to Single-Cell Sequencing. J. Comput. Biol. 19: 455–477.

Barringer, J.L. and Reilly, P.A. (2013) Arsenic in groundwater: a summary of sources and the biogeochemical and hydrogeologic factors affecting arsenic occurrence and mobility. In, Current perspectives in contaminant hydrology and water resources sustainability. InTech.

Bassas-Galia, M., Nogales, B., Arias, S., Rohde, M., Timmis, K.N., and Molinari, G. (2012) Plant original Massilia isolates producing polyhydroxybutyrate, including one exhibiting high yields from glycerol. J. Appl. Microbiol. 112: 443–54.

Bateman, A. (2002) The Pfam Protein Families Database. Nucleic Acids Res. 30: 276–280.

Bencko, V. (1977) Carcinogenic, teratogenic, and mutagenic effects of arsenic. Environ. Health Perspect. 19: 179–82.

Bentley, R. and Chasteen, T.G. (2002) Microbial Methylation of Metalloids : Arsenic , Antimony , and Bismuth Microbial Methylation of Metalloids : Arsenic , Antimony , and Bismuth. 66:.

Berg, M., Tran, H.C., Nguyen, T.C., Pham, H. V, Schertenleib, R., and Giger, W. (2001) Arsenic contamination of groundwater and drinking water in Vietnam: a human health threat. Environ. Sci. Technol. 35: 2621–6.

Bergey, D.H. (2005) Bergey’s Manual of Systematic Bacteriology - Vol 2: The Proteobacteria Part A - Introductory Essays. Springer-Verlag New York Inc. 332.

Bergey, D.H. (2010) Bergey’s Manual of Systematic Bacteriology - Vol 4: Bacteroidetes.

Bergquist, E.R., Fischer, R.J., Sugden, K.D., and Martin, B.D. (2009) Inhibition by methylated organo-arsenicals of the respiratory 2-oxo-acid dehydrogenases. J. Organomet. Chem. 694: 973–980.

Bertin, P.N., Heinrich-Salmeron, A., Pelletier, E., Goulhen-Chollet, F., Arsène-Ploetze, F., Gallien, S., et al. (2011) Metabolic diversity among main microorganisms inside an arsenic-rich ecosystem revealed by meta- and proteo-genomics. ISME J. 5: 1735–47.

BGS/MML (1999) Arsenic contamination of ground water in Bangladesh: a review. Prepared for the Department of Public Health Engineering, Government of Bangladesh Dhaka.

235 BGS, D. (1999) Groundwater studies for arsenic contamination in Bangladesh. Final report, rapid Investig. phase. Dep. Public Heal. Gov. Bangladesh, Mott MacDonald Br. Geol. Surv.

Bhattacharya, P., Vahter, M., Jarsjö, J., Kumpiene, J., Ahmad, A., Sparrenbom, C., et al. (2016) Arsenic Research and Global Sustainability: Proceedings of the Sixth International Congress on Arsenic in the Environment (As2016), June 19-23, 2016, Stockholm, Sweden CRC Press.

Black, A. and Craw, D. (2001) Arsenic, copper and zinc occurrence at the Wangaloa coal mine, southeast Otago, New Zealand. Int. J. Coal Geol. 45: 181–193.

Bonato, P., Batista, M.B., Camilios-Neto, D., Pankievicz, V.C.S., Tadra-Sfeir, M.Z., Monteiro, R.A., et al. (2016) RNA-seq analyses reveal insights into the function of respiratory nitrate reductase of the diazotroph Herbaspirillum seropedicae. Environ. Microbiol. 18: 2677– 2688.

Bostick, B.C. and Fendorf, S. (2003) Arsenite sorption on troilite (FeS) and pyrite (FeS2). Geochim. Cosmochim. Acta 67: 909–921.

El Bouhaddani, S., Houwing-Duistermaat, J., Salo, P., Perola, M., Jongbloed, G., and Uh, H.-W. (2016) Evaluation of O2PLS in Omics data integration. In, BMC bioinformatics. BioMed Central Ltd, p. 11.

Bouvet, P.J.M. and Grimont, P. a. D. (1986) Taxonomy of the Genus Acinetobacter with the Recognition of Acinetobacter baumannii sp. nov., Acinetobacter haemolyticus sp. nov., Acinetobacter johnsonii sp. nov., and Acinetobacter junii sp. nov. and Emended Descriptions of Acinetobacter calcoaceticus a. Int. J. Syst. Bacteriol. 36: 228–240.

Bowell, R., Alpers, C., Jamieson, H., Nordstrom, K., and Majzlan, J. (2014) Arsenic: Environmental geochemistry, mineralogy, and microbiology Walter de Gruyter GmbH & Co KG.

Bowman, J.P., Nichols, D.S., and McMeekin, T. a. (1997) Psychrobacter glacincola sp. nov., a Halotolerant, Psychrophilic Bacterium Isolated from Antarctic Sea Ice. Syst. Appl. Microbiol. 20: 209–215.

Branco, R., Francisco, R., Chung, A.P., and Morais, P.V. (2009) Identification of an aox system that requires cytochrome c in the highly arsenic-resistant bacterium Ochrobactrum tritici SCII24. Appl. Environ. Microbiol. 75: 5141–5147.

Breit, G.N., Whitney, J., Foster, A., Welch, A.H., Yount, J., and Sanzolone, R. (2001) Preliminary evaluation of arsenic cycling in the sediments of Bangladesh. In, U.S. Geol. Survey Workshop - Arsenic in the environment, Denver, CO, Feb. 21-22, 2001, wwwbrr.cr.usgs.gov/Arsenic/., pp. 21–22.

Bučková, M., Puškarová, A., Chovanová, K., Kraková, L., Ferianc, P., and Pangallo, D. (2013) A simple strategy for investigating the diversity and hydrocarbon degradation abilities of cultivable bacteria from contaminated soil. World J. Microbiol. Biotechnol. 29: 1085– 1098.

Bundschuh, J., Litter, M.I., Parvez, F., Román-Ross, G., Nicolli, H.B., Jean, J.-S., et al. (2012) One century of arsenic exposure in Latin America: a review of history and occurrence from 14 countries. Sci. Total Environ. 429: 2–35.

Buschmann, J. and Berg, M. (2009) Impact of sulfate reduction on the scale of arsenic

236 contamination in groundwater of the Mekong, Bengal and Red River deltas. Appl. Geochemistry 24: 1278–1286.

Buse, H.Y., Lu, J., Lu, X., Mou, X., and Ashbolt, N.J. (2014) Microbial diversities (16S and 18S rRNA gene pyrosequencing) and environmental pathogens within drinking water biofilms grown on the common premise plumbing materials unplasticized polyvinylchloride and copper. FEMS Microbiol. Ecol.

Butcher, B.G., Deane, S.M., and Rawlings, D.E. (2000) The chromosomal arsenic resistance genes of Thiobacillus ferrooxidans have an unusual arrangement and confer increased arsenic and antimony resistance to Escherichia coli. Appl. Environ. Microbiol. 66: 1826– 1833.

Cai, L., Liu, G., Rensing, C., and Wang, G. (2009a) Genes involved in arsenic transformation and resistance associated with different levels of arsenic-contaminated soils. BMC Microbiol. 9: 4.

Cai, L., Liu, G., Rensing, C., and Wang, G. (2009b) Genes involved in arsenic transformation and resistance associated with different levels of arsenic-contaminated soils. BMC Microbiol. 9: 4.

Campbell, K.M., Malasarn, D., Saltikov, C.W., Newman, D.K., and Hering, J.G. (2006) Simultaneous microbial reduction of iron (III) and arsenic (V) in suspensions of hydrous ferric oxide. Environ. Sci. Technol. 40: 5950–5955.

Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E.K., et al. (2010a) QIIME allows analysis of high- throughput community sequencing data Intensity normalization improves color calling in SOLiD sequencing. Nat. Publ. Gr. 7: 335–336.

Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E.K., et al. (2010b) QIIME allows analysis of high-throughput community sequencing data (Supplementary). Nat. Methods 7: 335.

Caporaso, J.G., Lauber, C.L., Walters, W.A., Berg-lyons, D., Lozupone, C.A., Turnbaugh, P.J., et al. (2010) Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample.

Carro, L., Rivas, R., León-Barrios, M., González-Tirante, M., Velázquez, E., and Valverde, A. (2012) Herbaspirillum canariense sp. nov., Herbaspirillum aurantiacum sp. nov. and Herbaspirillum soli sp. nov., isolated from volcanic mountain soil, and emended description of the genus Herbaspirillum. Int. J. Syst. Evol. Microbiol. 62: 1300–1306.

Carvalhais, L.C. and Schenk, P.M. (2013) Sample processing and cDNA preparation for microbial metatranscriptomics in complex soil communities. Methods Enzym. 531: 251–267.

Cervantes, C., Ji, G., Ramirez, J., and Silver, S. (1994) Resistance to arsenic compounds in microorganisms. FEMS Microbiol. Rev. 15: 355–367.

Chang, J.-S., Kim, Y.-H., and Kim, K.-W. (2008) The ars genotype characterization of arsenic- resistant bacteria from arsenic-contaminated gold–silver mines in the Republic of Korea. Appl. Microbiol. Biotechnol. 80: 155–165.

Charlet, L. and Polya, D. a. (2006) Arsenic in shallow, reducing groundwaters in Southern Asia: An environmental health disaster. Elements 2: 91–96.

237 Chaturvedi, P. and Shivaji, S. (2006) Exiguobacterium indicum sp. nov., a psychrophilic bacterium from the Hamta glacier of the Himalayan mountain ranges of India. Int. J. Syst. Evol. Microbiol. 56: 2765–2770.

Chauhan, N.S., Ranjan, R., Purohit, H.J., Kalia, V.C., and Sharma, R. (2008) Identification of genes conferring arsenic resistance to Escherichia coli from an effluent treatment plant sludge metagenomic library. FEMS Microbiol. Ecol. 67: 130–139.

Chawla, S., Mutenda, E.K., Dixon, H.B., Freeman, S., and Smith, a W. (1995) Synthesis of 3- arsonopyruvate and its interaction with phosphoenolpyruvate mutase. Biochem. J. 308 ( Pt 3: 931–5.

Chen, J., Bhattacharjee, H., and Rosen, B.P. (2015) ArsH is an organoarsenical oxidase that confers resistance to trivalent forms of the herbicide monosodium methylarsenate and the poultry growth promoter roxarsone. Mol. Microbiol. 96: 1042–1052.

Chen, J., Madegowda, M., Bhattacharjee, H., and Rosen, B.P. (2015) ArsP: a methylarsenite efflux permease. Mol. Microbiol. 98: 625–635.

Chen, J., Yoshinaga, M., Garbinski, L.D., and Rosen, B.P. (2016) Synergistic interaction of glyceraldehydes‐3‐phosphate dehydrogenase and ArsJ, a novel organoarsenical efflux permease, confers arsenate resistance. Mol. Microbiol. 100: 945–953.

Chen, L., Hu, M., Huang, L., Hua, Z., Kuang, J., Li, S., and Shu, W. (2015) Comparative metagenomic and metatranscriptomic analyses of microbial communities in acid mine drainage. ISME J. 9: 1579.

Cheng, Z., Zheng, Y., Mortlock, R., and Van Geen, a (2004) Rapid multi-element analysis of groundwater by high-resolution inductively coupled plasma mass spectrometry. Anal. Bioanal. Chem. 379: 512–8.

Chenna, R., Sugawara, H., Koike, T., Lopez, R., Gibson, T.J., Higgins, D.G., and Thompson, J.D. (2003) Multiple sequence alignment with the Clustal series of programs. Nucleic Acids Res. 31: 3497–3500.

Cherry, J.A., Shaikh, A.U., Tallman, D.E., and Nicholson, R. V. (1979) Arsenic Species as an Indicator of Redox Conditions in Groundwater. Dev. Water Sci. 12: 373–392.

Chistoserdova, L. (2010) Recent progress and new challenges in metagenomics for biotechnology. Biotechnol. Lett. 32: 1351–9.

Chitpirom, K., Akaracharanya, A., Tanasupawat, S., Leepipatpiboon, N., and Kim, K.W. (2009) Isolation and characterization of arsenic resistant bacteria from tannery wastes and agricultural soils in Thailand. Ann. Microbiol. 59: 649–656.

Chrysostomou, C., Quandt, E.M., Marshall, N.M., Stone, E., and Georgiou, G. (2015) An Alternate Pathway of Arsenate Resistance in E. coli Mediated by the Glutathione S- Transferase GstB. ACS Chem. Biol. 10: 875–882.

Coker, V.S., Gault, a G., Pearce, C.I., van der Laan, G., Telling, N.D., Charnock, J.M., et al. (2006) XAS and XMCD evidence for species-dependent partitioning of arsenic during microbial reduction of ferrihydrite to magnetite. Environ. Sci. Technol. 40: 7745–50.

Colbourn, P., Alloway, B.J., and Thornton, I. (1975) Arsenic and heavy metals in soils associated with regional geochemical anomalies in South-West England. Sci. Total Environ. 4: 359–

238 363.

Corsini, A., Cavalca, L., Crippa, L., Zaccheo, P., and Andreoni, V. (2010) Impact of glucose on microbial community of a soil containing pyrite cinders: role of bacteria in arsenic mobilization under submerged condition. Soil Biol. Biochem. 42: 699–707.

Cotton, F.A. and Wilkinson, G. (1988) Advanced inorganic chemistry Wiley New York.

Couture, R.M. and Van Cappellen, P. (2011) Reassessing the role of sulfur geochemistry on arsenic speciation in reducing environments. J. Hazard. Mater. 189: 647–652.

Cozen, A.E., Weirauch, M.T., Pollard, K.S., Bernick, D.L., Stuart, J.M., and Lowe, T.M. (2009) Transcriptional map of respiratory versatility in the hyperthermophilic crenarchaeon Pyrobaculum aerophilum. J. Bacteriol. 191: 782–794.

Cullen, W.R. and Reimer, K.J. (1989) Arsenic speciation in the environment. Chem. Rev. 89: 713–764.

Cummings, D.E., March, A.W., Bostick, B., Spring, S., Jr, F.C., Fendorf, S., et al. (2000) Evidence for Microbial Fe ( III ) Reduction in Anoxic , Mining-Impacted Lake Sediments Evidence for Microbial Fe ( III ) Reduction in Anoxic , Mining- Impacted Lake Sediments ( Lake Coeur d ’ Alene , Idaho ). Appl. Environ. Microbiol. 66: 154–162.

Cummings, D.E., Snoeyenbos-West, O.L., Newby, D.T., Niggemyer, A.M., Lovley, D.R., Achenbach, L.A., and Rosenzweig, R.F. (2003) Diversity of Geobacteraceae species inhabiting metal-polluted freshwater lake sediments ascertained by 16S rDNA analyses. Microb. Ecol. 46: 257–269.

Cummings D.E., S.-W.O.L. (2002) Diversity of Geobacteraceae Species Inhabiting Metal- Polluted Freshwater Lake Sediments Ascertained by 16S rDNA Analysis. Microb. Ecol. 44: 95–106.

Daims, H., Lebedeva, E.V., Pjevac, P., Han, P., Herbold, C., Albertsen, M., et al. (2015) Complete nitrification by Nitrospira bacteria. Nature 528: 504–509.

Datta, S., Mailloux, B., Jung, H.-B., Hoque, M. a, Stute, M., Ahmed, K.M., and Zheng, Y. (2009) Redox trapping of arsenic during groundwater discharge in sediments from the Meghna riverbank in Bangladesh. Proc. Natl. Acad. Sci. U. S. A. 106: 16930–16935.

Delavat, F., Lett, M.-C., and Lievremont, D. (2012) Novel and unexpected bacterial diversity in an arsenic-rich ecosystem revealed by culture-dependent approaches. Biol. Direct 7: 28.

Desoeuvre, A., Casiot, C., and Héry, M. (2016) Diversity and Distribution of Arsenic-Related Genes Along a Pollution Gradient in a River Affected by Acid Mine Drainage. Microb. Ecol. 71: 672–685.

Dhar, R.K., Zheng, Y., Saltikov, C.W., Radloff, K.A., Mailloux, B.J., Ahmed, K.M., and Van Geen, A. (2011) Microbes enhance mobility of arsenic in pleistocene aquifer sand from Bangladesh. Environ. Sci. Technol. 45: 2648–2654.

Dhar, R.K., Zheng, Y., Stute, M., van Geen, a., Cheng, Z., Shanewaz, M., et al. (2008) Temporal variability of groundwater chemistry in shallow and deep aquifers of Araihazar, Bangladesh. J. Contam. Hydrol. 99: 97–111.

Dixit, S. and Hering, J.G. (2003) Comparison of arsenic(V) and arsenic(III) sorption onto iron

239 oxide minerals: Implications for arsenic mobility. Environ. Sci. Technol. 37: 4182–4189.

Dobbin, P.S., Warren, L.H., Cook, N.J., McEwan, A.G., Powell, A.K., and Richardson, D.J. (1996) Dissimilatory iron (III) reduction by Rhodobacter capsulatus. Microbiology 142: 765–774.

Dombrowski, P.M., Long, W., Farley, K.J., Mahony, J.D., Capitani, J.F., and Di Toro, D.M. (2005) Thermodynamic analysis of arsenic methylation. Environ. Sci. Technol. 39: 2169–2176.

Dong, G., Huang, Y., Yu, Q., Wang, Y., Wang, H., He, N., and Li, Q. (2014) Role of Nanoparticles in Controlling Arsenic Mobilization from Sediments near a Realgar Tailing. Environ. Sci. Technol.

Dowdle, P.R., Laverman, A.M., and Oremland, R.S. (1996) Bacterial dissimilatory reduction of arsenic (V) to arsenic (III) in anoxic sediments. Appl. Environ. Microbiol. 62: 1664–1669.

Drewniak, L. and Sklodowska, A. (2013) Arsenic-transforming microbes and their role in biomining processes. Environ. Sci. Pollut. Res. 20:.

Drewniak, L., Stasiuk, R., Uhrynowski, W., and Sklodowska, A. (2015) Shewanella sp. O23S as a driving agent of a system utilizing dissimilatory arsenate-reducing bacteria responsible for self-cleaning of water contaminated with arsenic. Int. J. Mol. Sci. 16:.

Duquesne, K., Lebrun, S., Casiot, C., Bruneel, O., Leblanc, M., Morin, G., and Personne, J. (2003) Immobilization of Arsenite and Ferric Iron by Acidithiobacillus ferrooxidans and Its Relevance to Acid Mine Drainage Immobilization of Arsenite and Ferric Iron by Acidithiobacillus ferrooxidans and Its Relevance to Acid Mine Drainage.

Duval, S., Ducluzeau, A.-L., Nitschke, W., and Schoepp-Cothenet, B. (2008) Enzyme phylogenies as markers for the oxidation state of the environment: the case of respiratory arsenate reductase and related enzymes. BMC Evol. Biol. 8: 206.

Edgar, R.C. (2004) MUSCLE: a multiple sequence alignment method with reduced time and space complexity. BMC Bioinformatics 5: 113.

Edgar, R.C. (2004) MUSCLE: Multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32: 1792–1797.

Edgar, R.C. (2010) Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26: 2460–2461.

Edgar, R.C. (2013) UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat. Methods 10: 996–998.

Edwards, M. (1994) Chemistry of arsenic removal during coagulation and Fe-Mn oxidation. J. Am. Water Work. Assoc. States) 86:.

Edwardson, C.F. and Hollibaugh, J.T. (2017) Metatranscriptomic analysis of prokaryotic communities active in sulfur and arsenic cycling in Mono Lake, California, USA. ISME J. 1– 14.

Efron, B. (1981) Nonparametric estimates of standard error: the jackknife, the bootstrap and other methods. Biometrika 68: 589–599.

Ehrich, S., Behrens, D., Lebedeva, E. V, Ludwig, W., and Bock, E. (1995) A New Obligately Chemolithoautotrophic, Nitrite-Oxidizing Bacterium, Nitrospira-Moscoviensis Sp-Nov and Its Phylogenetic Relationship. Arch. Microbiol. 164: 16–23.

240 Emerson, D., Field, E.K., Chertkov, O., Davenport, K.W., Goodwin, L., Munk, C., et al. (2013) Comparative genomics of freshwater Fe-oxidizing bacteria: Implications for physiology, ecology, and systematics. Front. Microbiol. 4: 1–17.

Emerson, D., Fleming, E.J., and McBeth, J.M. (2010) Iron-oxidizing bacteria: an environmental and genomic perspective. Annu. Rev. Microbiol. 64: 561–583.

Escudero, L. V, Casamayor, E.O., Chong, G., Pedrós-Alió, C., and Demergasso, C. (2013) Distribution of Microbial Arsenic Reduction, Oxidation and Extrusion Genes along a Wide Range of Environmental Arsenic Concentrations. PLoS One 8: e78890.

Del Fabbro, C., Scalabrin, S., Morgante, M., and Giorgi, F.M. (2013) An extensive evaluation of read trimming effects on illumina NGS data analysis. PLoS One 8: 1–13.

Ferguson, J.F. and Gavis, J. (1972) A review of the arsenic cycle in natural waters. Water Res. 6: 1259–1274.

Fernández, M., Morel, B., Ramos, J.L., and Krell, T. (2016) Paralogous Regulators ArsR1 and ArsR2 of Pseudomonas putida KT2440 as a Basis for Arsenic Biosensor Development. Appl. Environ. Microbiol. 82: 4133–4144.

Finn, R.D., Coggill, P., Eberhardt, R.Y., Eddy, S.R., Mistry, J., Mitchell, A.L., et al. (2016) The Pfam protein families database: Towards a more sustainable future. Nucleic Acids Res. 44: D279–D285.

Fisher, E., Dawson, A.M., Polshyna, G., Lisak, J., Crable, B., Perera, E., et al. (2008) Transformation of inorganic and organic arsenic by Alkaliphilus oremlandii sp. nov. strain OhILAs. Ann. N. Y. Acad. Sci. 1125: 230–41.

Flemming, H.-C. and Wingender, J. (2010) The biofilm matrix. Nat. Rev. Microbiol. 8: 623.

Focardi, S., Pepi, M., Ruta, M., Marvasi, M., Bernardini, E., Gasperini, S., and Focardi, S.E. (2010) Arsenic precipitation by an anaerobic arsenic-respiring bacterial strain isolated from the polluted sediments of Orbetello Lagoon, Italy. Lett. Appl. Microbiol. 51: 578– 585.

Francis, C.A., O’Mullan, G.D., and Ward, B.B. (2003) Diversity of ammonia monooxygenase (amoA) genes across environmental gradients in Chesapeake Bay sediments. Geobiology 1: 129–140.

Frankenberger Jr, W.T. (2001) Environmental chemistry of arsenic CRC Press.

Freikowski, D., Winter, J., and Gallert, C. (2010) Hydrogen formation by an arsenate-reducing Pseudomonas putida, isolated from arsenic-contaminated groundwater in West Bengal, India. Appl Microbiol Biotechnol 88: 1363–1371.

Fresquet, M., Jowitt, T.A., Gummadova, J., Collins, R., O’Cualain, R., McKenzie, E.A., et al. (2015) Identification of a Major Epitope Recognized by PLA2R Autoantibodies in Primary Membranous Nephropathy. J. Am. Soc. Nephrol. 26: 302–313.

Frias-Lopez, J., Shi, Y., Tyson, G.W., Coleman, M.L., Schuster, S.C., Chisholm, S.W., and Delong, E.F. (2008) Microbial community gene expression in ocean surface waters. Proc. Natl. Acad. Sci. U. S. A. 105: 3805–10.

Fritz, I., Strömpl, C., Nikitin, D.I., Lysenko, A.M., and Abraham, W.R. (2005) Brevundimonas

241 mediterranea sp. nov., a non-stalked species from the Mediterranean Sea. Int. J. Syst. Evol. Microbiol. 55: 479–486.

Frohne, T., Rinklebe, J., Diaz-Bone, R.A., and Du Laing, G. (2011) Controlled variation of redox conditions in a floodplain soil: impact on metal mobilization and biomethylation of arsenic and antimony. Geoderma 160: 414–424.

Gallego, V., Sánchez-Porro, C., García, M.T., and Ventosa, A. (2006) Massilia aurea sp. nov., isolated from drinking water. Int. J. Syst. Evol. Microbiol. 56: 2449–2453.

Garelick, H., Jones, H., Dybowska, A., and Valsami-jones, E. (2008) Reviews of Environmental Contamination Volume 197 Springer New York, New York, NY.

Garrity, G., Bell, J., and Lilburn, T. (2010) Family I. Pseudomonadaceae. Bergey´s Man. Syst. Bacteriol. vol 2B 323–379.

Gault, A.G., Islam, F.S., Polya, D.A., Charnock, J.M., Boothman, C., Chatterjee, D., and Lloyd, J.R. (2005) Microcosm depth profiles of arsenic release in a shallow aquifer, West Bengal. Mineral. Mag. 69: 855–863.

Ge, Y., Wang, D.-Z., Chiu, J.-F., Cristobal, S., Sheehan, D., Silvestre, F., et al. (2013) Environmental OMICS: Current Status and Future Directions. J. Integr. OMICS 3:. van Geen, a., Zheng, Y., Cheng, Z., Aziz, Z., Horneman, a., Dhar, R.K., et al. (2006) A transect of groundwater and sediment properties in Araihazar, Bangladesh: Further evidence of decoupling between As and Fe mobilization. Chem. Geol. 228: 85–96.

Georges, A. a, El-Swais, H., Craig, S.E., Li, W.K., and Walsh, D. a (2014) Metaproteomic analysis of a winter to spring succession in coastal northwest Atlantic Ocean microbial plankton. ISME J. 1–13.

Ghosh, S. and Sar, P. (2013) Identification and characterization of metabolic properties of bacterial populations recovered from arsenic contaminated ground water of North East India (Assam). Water Res. 47: 6992–7005.

Gihring, T.M., Druschel, G.K., McCleskey, R.B., Hamers, R.J., and Banfield, J.F. (2001) Rapid arsenite oxidation by Thermus aquaticus and Thermus thermophilus: field and laboratory investigations. Environ. Sci. Technol. 35: 3857–3862.

Gilbert, J.A. (2010) Beyond the infinite -Tracking bacterial gene expression. Microbiol. Today 37: 82–85.

Giloteaux, L., Holmes, D.E., Williams, K.H., Wrighton, K.C., Wilkins, M.J., Montgomery, A.P., et al. (2012) Characterization and transcription of arsenic respiration and resistance genes during in situ uranium bioremediation. ISME J.

Giri, A.K., Patel, R.K., Mahapatra, S.S., and Mishra, P.C. (2013) Biosorption of arsenic (III) from aqueous solution by living cells of Bacillus cereus. Environ. Sci. Pollut. Res. 20: 1281–1291.

Goldberg, S. and Johnston, C.T. (2001) Mechanisms of Arsenic Adsorption on Amorphous Oxides Evaluated Using Macroscopic Measurements, Vibrational Spectroscopy, and Surface Complexation Modeling. J. Colloid Interface Sci. 234: 204–216.

Gremion, F., Chatzinotas, A., and Harms, H. (2003) Comparative 16S rDNA and 16S rRNA sequence analysis indicates that Actinobacteria might be a dominant part of the

242 metabolically active bacteria in heavy metal‐contaminated bulk and rhizosphere soil. Environ. Microbiol. 5: 896–907.

Grondin, K., Haimeur, A., Mukhopadhyay, R., Rosen, B.P., and Ouellette, M. (1997) Co‐ amplification of the γ‐glutamylcysteine synthetase gene gsh1 and of the ABC transporter gene pgpA in arsenite‐resistant Leishmania tarentolae. EMBO J. 16: 3057–3065.

Guidelines for Drinking Water Quality: Recommendations, vol.1. (1993) Geneva.

Guo, H., Liu, Z., Ding, S., Hao, C., Xiu, W., and Hou, W. (2015) Arsenate reduction and mobilization in the presence of indigenous aerobic bacteria obtained from high arsenic aquifers of the Hetao basin, Inner Mongolia. Environ. Pollut. 203: 50–59.

Gussin, G.N., Ronson, C.W., and Ausubel, F.M. (1986) Regulation of nitrogen fixation genes. Annu. Rev. Genet. 20: 567–91.

Guyon, I., Weston, J., Barnhill, S., and Vapnik, V. (2002) Gene selection for cancer classification using support vector machines. Mach. Learn. 46: 389–422.

Haft, D.H. (2001) TIGRFAMs: a protein family resource for the functional identification of proteins. Nucleic Acids Res. 29: 41–43.

Hamady, M., Walker, J.J., Harris, J.K., Gold, N.J., and Knight, R. (2008) Error-correcting barcoded primers for pyrosequencing hundreds of samples in multiplex. Nat. Methods 5: 235–237.

Hamamura, N., Fukushima, K., and Itai, T. (2013) Identification of antimony- and arsenic- oxidizing bacteria associated with antimony mine tailing. Microbes Env. 28: 257–263.

Hamamura, N., Macur, R.E., Korf, S., Ackerman, G., Taylor, W.P., Kozubal, M., et al. (2009) Linking microbial oxidation of arsenic with detection and phylogenetic analysis of arsenite oxidase genes in diverse geothermal environments B : NGB-BE. 11: 421–431.

Handley, K.M., Héry, M., and Lloyd, J.R. (2009) Redox cycling of arsenic by the hydrothermal marine bacterium Marinobacter santoriniensis. Environ. Microbiol. 11: 1601–1611.

Handley, K.M., McBeth, J.M., Charnock, J.M., Vaughan, D.J., Wincott, P.L., Polya, D. a., and Lloyd, J.R. (2013) Effect of iron redox transformations on arsenic solid-phase associations in an arsenic-rich, ferruginous hydrothermal sediment. Geochim. Cosmochim. Acta 102: 124–142.

Harold, F.M. and Baarda, J.R. (1966) Interaction of arsenate with phosphate-transport systems in wild- type and mutant Streptococcus faecalis. J. Bacteriol. 91: 2257–2262.

Hassan, Z., Sultana, M., van Breukelen, B.M., Khan, S.I., and Roling, W.F.M. (2015) Diverse arsenic- and iron-cycling microbial communities in arsenic-contaminated aquifers used for drinking water in Bangladesh. FEMS Microbiol. Ecol. 91: fiv026-fiv026.

Hassan, Z., Sultana, M., Breukelen, B.M. Van, Khan, S.I., and Röling, W.F.M. (2015) Diverse arsenic- and iron-cycling microbial communities in arsenic- contaminated aquifers used for drinking water in Bangladesh. 1–42.

Heinrich-Salmeron, A., Cordi, A., Brochier-Armanet, C., Halter, D., Pagnout, C., Abbaszadeh- Fard, E., et al. (2011) Unsuspected diversity of arsenite-oxidizing bacteria revealed by a widespread distribution of the aoxB gene in prokaryotes. Appl. Environ. Microbiol. AEM-

243 02884.

Herbel, M. and Fendorf, S. (2006) Biogeochemical processes controlling the speciation and transport of arsenic within iron coated sands. Chem. Geol. 228: 16–32.

Hert, D.G., Fredlake, C.P., and Barron, A.E. (2008) Advantages and limitations of next- generation sequencing technologies: a comparison of electrophoresis and non- electrophoresis methods. Electrophoresis 29: 4618–4626.

Héry, M., Van Dongen, B.E., Gill, F., Mondal, D., Vaughan, D.J., Pancost, R.D., et al. (2010) Arsenic release and attenuation in low organic carbon aquifer sediments from West Bengal. Geobiology 8: 155–68.

Héry, M., Gault, A.G., Rowland, H. a. L., Lear, G., Polya, D. a., and Lloyd, J.R. (2008) Molecular and cultivation-dependent analysis of metal-reducing bacteria implicated in arsenic mobilisation in south-east asian aquifers. Appl. Geochemistry 23: 3215–3223.

Héry, M., Rizoulis, A., Sanguin, H., Cooke, D. a, Pancost, R.D., Polya, D. a, and Lloyd, J.R. (2014) Microbial ecology of arsenic-mobilizing Cambodian sediments: lithological controls uncovered by stable-isotope probing. Environ. Microbiol.

Hoeft, S.E., Blum, J.S., Stolz, J.F., Tabita, F.R., Witte, B., King, G.M., et al. (2007) Alkalilimnicola ehrlichii sp. nov., a novel, arsenite-oxidizing haloalkaliphilic gammaproteobacterium capable of chemoautotrophic or heterotrophic growth with nitrate or oxygen as the electron acceptor. Int. J. Syst. Evol. Microbiol. 57: 504–512.

Hoeft, S.E., Kulp, T.R., Stolz, J.F., Hollibaugh JT, and Oremland, R.S. (2004) Dissimilatory arsenate\rreduction with sulfide as the electron donor: Experiments with Mono Lake water and \risolation of strain MLMS-1, a chemoautotrophic arsenate-respirer. Appl Env. Microbiol 70: 2741–2747.

Hohmann, C., Morin, G., Ona-Nguema, G., Guigner, J.M., Brown, G.E., and Kappler, A. (2011) Molecular-level modes of As binding to Fe(III) (oxyhydr)oxides precipitated by the anaerobic nitrate-reducing Fe(II)-oxidizing Acidovorax sp. strain BoFeN1. Geochim. Cosmochim. Acta 75: 4699–4712.

Hollibaugh, J.T., Budinoff, C., Hollibaugh, R. a., Ransom, B., and Bano, N. (2006) Sulfide oxidation coupled to arsenate reduction by a diverse microbial community in a soda lake. Appl. Environ. Microbiol. 72: 2043–2049.

Horneman, a., van Geen, a., Kent, D. V., Mathe, P.E., Zheng, Y., Dhar, R.K., et al. (2004) Decoupling of As and Fe release to Bangladesh groundwater under reducing conditions. Part I: Evidence from sediment profiles. Geochim. Cosmochim. Acta 68: 3459–3473.

Hosseini, P., Tremblay, A., Matthews, B.F., and Alkharouf, N.W. (2010) An efficient annotation and gene-expression derivation tool for Illumina Solexa datasets. BMC Res. Notes 3: 183.

Hoven, R.N. and Santini, J.M. (2004) Arsenite oxidation by the heterotroph Hydrogenophaga sp . str . NT-14 : the arsenite oxidase and its physiological electron acceptor. 1656: 148–155. vanden Hoven, R.N. and Santini, J.M. (2004) Arsenite oxidation by the heterotroph Hydrogenophaga sp. str. NT-14: the arsenite oxidase and its physiological electron acceptor. Biochim. Biophys. Acta (BBA)-Bioenergetics 1656: 148–155.

Hu, Z. and Gao, S. (2008) Upper crustal abundances of trace elements: A revision and update.

244 Chem. Geol. 253: 205–221.

Huang, J.-H. (2014) Impact of Microorganisms on Arsenic Biogeochemistry: A Review. Water, Air, Soil Pollut. 225: 1848.

Huang, J.-H. and Matzner, E. (2007) Biogeochemistry of organic and inorganic arsenic species in a forested catchment in Germany. Environ. Sci. Technol. 41: 1564–1569.

Huang, J.H., Lazzaro, A., Zeyer, J., Kretzschmar, R., and Voegelin, A. (2010) Influence of arsenic surface adsorption and iron reduction on arsenic mobilization and reduction by Shewanella putrefaciens strain CN-32. In, Arsenic in Geosphere and Human Diseases; Arsenic 2010: Proceedings of the Third International Congress on Arsenic in the Environment (As-2010). CRC Press, p. 130.

Huber, R., Sacher, M., Vollmann, A., Huber, H., and Rose, D. (2000) Respiration of arsenate and selenate by hyperthermophilic archaea. Syst. Appl. Microbiol. 23: 305–314.

Hug, K., Maher, W.A., Stott, M.B., Krikowa, F., Foster, S., and Moreau, J.W. (2014) Microbial contributions to coupled arsenic and sulfur cycling in the acid-sulfide hot spring Champagne Pool, New Zealand. Front. Microbiol. 5:.

Hug, L.A., Thomas, B.C., Sharon, I., Brown, C.T., Sharma, R., Hettich, R.L., et al. (2016) Critical biogeochemical functions in the subsurface are associated with bacteria from new phyla and little studied lineages. Environ. Microbiol. 18: 159–173.

Hultman, J., Waldrop, M.P., Mackelprang, R., David, M.M., McFarland, J., Blazewicz, S.J., et al. (2015) Multi-omics of permafrost, active layer and thermokarst bog soil microbiomes. Nature.

Hyatt, D., Chen, G.-L., LoCascio, P.F., Land, M.L., Larimer, F.W., and Hauser, L.J. (2010) Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinformatics 11: 119.

Inskeep, W.P., Macur, R.E., Hamamura, N., Warelow, T.P., Ward, S. a, and Santini, J.M. (2007) Detection, diversity and expression of aerobic bacterial arsenite oxidase genes. Environ. Microbiol. 9: 934–43.

Irlinger, F., Bimet, F., Delettre, J., Lefèvre, M., and Grimont, P. a D. (2005) Arthrobacter bergerei sp. nov. and Arthrobacter arilaitensis sp. nov., novel coryneform species isolated from the surfaces of cheeses. Int. J. Syst. Evol. Microbiol. 55: 457–62.

Islam, F.S., Gault, A.G., Boothman, C., Polya, D. a, Charnock, J.M., Chatterjee, D., and Lloyd, J.R. (2004) Role of metal-reducing bacteria in arsenic release from Bengal delta sediments. Nature 430: 68–71.

Islam, F.S., Pederick, R.L., Gault, A.G., Adams, L.K., Polya, D.A., Charnock, J.M., and Lloyd, J.R. (2005a) Interactions between the Fe ( III ) -Reducing Bacterium Geobacter sulfurreducens and Arsenate , and Capture of the Metalloid by Biogenic Fe ( II ). 71: 8642–8648.

Islam, F.S., Pederick, R.L., Gault, A.G., Adams, L.K., Polya, D.A., Charnock, J.M., and Lloyd, J.R. (2005b) Interactions between the Fe (III)-reducing bacterium Geobacter sulfurreducens and arsenate, and capture of the metalloid by biogenic Fe (II). Appl. Environ. Microbiol. 71: 8642–8648.

Jackson, C.R. and Dugas, S.L. (2003) Phylogenetic analysis of bacterial and archaeal arsC gene

245 sequences suggests an ancient, common origin for arsenate reductase. BMC Evol. Biol. 3: 18.

Ji, G., Garber, E. a, Armes, L.G., Chen, C.M., Fuchs, J. a, and Silver, S. (1994) Arsenate reductase of Staphylococcus aureus plasmid pI258. Biochemistry 33: 7294–9.

Ji, G. and Silver, S. (1992) Reduction of arsenate to arsenite by the ArsC protein of the arsenic resistance operon of Staphylococcus aureus plasmid pI258. Proc. Natl. Acad. Sci. U. S. A. 89: 9474–8.

Jin, H., Wen, H., Yang, H., Liu, Z., Ren, Z., Che, G., and Zhao, Z. (2003) Arsenic poisoning of Bangladesh groundwater. Sci. Technol. 3: 16–18.

Jin, Y. and Manabe, T. (2005) High-efficiency protein extraction from polyacrylamide gels for molecular mass measurement by matrix-assisted laser desorption/ionization-time of flight-mass spectrometry. Electrophoresis 26: 1019–1028.

Jomova, K., Jenisova, Z., Feszterova, M., Baros, S., Liska, J., Hudecova, D., et al. (2011) Arsenic: toxicity, oxidative stress and human disease. J. Appl. Toxicol. 31: 95–107.

Joshi, N.A. and Fass, J.N. (2011) Sickle: A sliding-window, adaptive, quality-based trimming tool for FastQ files (Version 1.33)[Software].

Jung, H.B., Bostick, B.C., and Zheng, Y. (2012) Field, experimental, and modeling study of arsenic partitioning across a redox transition in a bangladesh aquifer. Environ. Sci. Technol. 46: 1388–1395.

Jung, H.B. and Zheng, Y. (2006) Enhanced recovery of arsenite sorbed onto synthetic oxides by L-ascorbic acid addition to phosphate solution: calibrating a sequential leaching method for the speciation analysis of arsenic in natural samples. Water Res. 40: 2168–2180.

Kageyama, A., Morisaki, K., Omura, S., and Takahashi, Y. (2008) Arthrobacter oryzae sp. nov. and Arthrobacter humicola sp. nov. Int. J. Syst. Evol. Microbiol. 58: 53–6.

Kämpfer, P., Fallschissel, K., and Avendaño-Herrera, R. (2011) Chryseobacterium chaponense sp. nov., isolated from farmed Atlantic salmon (Salmo salar). Int. J. Syst. Evol. Microbiol. 61: 497–501.

Kämpfer, P., Vaneechoutte, M., Lodders, N., De Baere, T., Avesani, V., Janssens, M., et al. (2009) Description of Chryseobacterium anthropi sp. nov. to accommodate clinical isolates biochemically similar to Kaistella koreensis and Chryseobacterium haifense, proposal to reclassify Kaistella koreensis as Chryseobacterium koreense comb. nov. and emended d. Int. J. Syst. Evol. Microbiol. 59: 2421–8.

Kanehisa, M. (2002) The KEGG databases at GenomeNet. Nucleic Acids Res. 30: 42–46.

Kanehisa Laboratories (2017) KEGG: Kyoto Encyclopedia of Genes and Genomes. http://www.kegg.jp/ (accessed 2015). 27: 29–34.

Kang, D.D., Froula, J., Egan, R., and Wang, Z. (2015) MetaBAT, an efficient tool for accurately reconstructing single genomes from complex microbial communities. PeerJ 3: e1165.

Kapaj, S., Peterson, H., Liber, K., and Bhattacharya, P. (2006) Human health effects from chronic arsenic poisoning–a review. J. Environ. Sci. Heal. Part A 41: 2399–2428.

Karim, M.M. (2000) Arsenic in groundwater and health problems in Bangladesh. Water Res. 34:

246 304–310.

Kashyap, D.R., Botero, L.M., Franck, W.L., Hassett, D.J., and McDermott, T.R. (2006) Complex regulation of arsenite oxidation in Agrobacterium tumefaciens. J. Bacteriol. 188: 1081– 1088.

Kavitha, S., Selvakumar, R., Sathishkumar, M., Swaminathan, K., Lakshmanaperumalsamy, P., Singh, A., and Jain, S.K. (2009) Nitrate removal using Brevundimonas diminuta MTCC 8486 from ground water. Water Sci. Technol. 60: 517–524.

Kearse, M., Moir, R., Wilson, A., Stones-Havas, S., Cheung, M., Sturrock, S., et al. (2012) Geneious Basic: An integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics 28: 1647–1649.

Kennard, R.W. and Stone, L.A. (1969) Computer aided design of experiments. Technometrics 11: 137–148.

Kim, E.H., Jeong, H.-J., Lee, Y.K., Moon, E.Y., Cho, J.-C., Lee, H.K., and Hong, S.G. (2011) Actimicrobium antarcticum gen. nov., sp. nov., of the family Oxalobacteraceae, isolated from Antarctic coastal seawater. Curr. Microbiol. 63: 213–7.

Kim, K.K., Lee, K.C., Oh, H.-M., Kim, M.J., Eom, M.K., and Lee, J.-S. (2008) Arthrobacter defluvii sp. nov., 4-chlorophenol-degrading bacteria isolated from sewage. Int. J. Syst. Evol. Microbiol. 58: 1916–21.

Kirk, M.F., Holm, T.R., Park, J., Jin, Q., Sanford, R.A., Fouke, B.W., and Bethke, C.M. (2004) Bacterial sulfate reduction limits natural arsenic contamination in groundwater. Geology 32: 953–956.

Kocar, B.D., Herbel, M.J., Tufano, K.J., and Fendorf, S. (2006) Contrasting effects of dissimilatory iron (III) and arsenic (V) reduction on arsenic retention and transport. Environ. Sci. Technol. 40: 6715–21.

Kocar, B.D., Polizzotto, M.L., Benner, S.G., Ying, S.C., Ung, M., Ouch, K., et al. (2008) Integrated biogeochemical and hydrologic processes driving arsenic release from shallow sediments to groundwaters of the Mekong delta. Appl. Geochemistry 23: 3059–3071.

Kozich, J.J., Westcott, S.L., Baxter, N.T., Highlander, S.K., and Schloss, P.D. (2013) Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the miseq illumina sequencing platform. Appl. Environ. Microbiol. 79: 5112–5120.

Krafft, T. and Macy, J.M. (1998) Purification and characterization of the respiratory arsenate reductase of Chrysiogenes arsenatis. Eur. J. Biochem. 255: 647–53.

Kräutler, B. (1990) Chemistry of methylcorrinoids related to their roles in bacterial C1 metabolism. FEMS Microbiol. Lett. 87: 349–354.

Kruger, M.C., Bertin, P.N., Heipieper, H.J., and Arsène-Ploetze, F. (2013) Bacterial metabolism of environmental arsenic - Mechanisms and biotechnological applications. Appl. Microbiol. Biotechnol. 97:.

Kuehnelt, D. and Goessler, W. (2003) Organoarsenic compounds in the terrestrial environment. Organomet. Compd. Environ. 223–275.

247 Kulp, T.R., Hoeft, S.E., Miller, L.G., Saltikov, C., Murphy, J.N., Han, S., et al. (2006) Dissimilatory arsenate and sulfate reduction in sediments of two hypersaline, arsenic-rich soda lakes: Mono and Searles Lakes, California. Appl. Environ. Microbiol. 72: 6514–26.

Kumar, S., Nei, M., Dudley, J., and Tamura, K. (2008) MEGA: A biologist-centric software for evolutionary analysis of DNA and protein sequences. Brief. Bioinform. 9: 299–306.

Lane, D.J. (1991) 16S/23S rRNA sequencing, in Nucleic Acid Techniques in Bacterial Systematics. In, E. Stackebrandt,M.G. (ed). Wiley, New York, NY, pp. 115–1175.

Langmead, B. and Salzberg, S.L. (2012) Fast gapped-read alignment with Bowtie 2. Nat. Methods 9: 357–9.

Lara, J., Escudero González, L., Ferrero, M., Chong Díaz, G., Pedrós-Alió, C., and Demergasso, C. (2012) Enrichment of arsenic transforming and resistant heterotrophic bacteria from sediments of two salt lakes in Northern Chile. Extremophiles 16: 523–38.

Lau, M.C.Y., Kieft, T.L., Kuloyo, O., Linage-Alvarez, B., van Heerden, E., Lindsay, M.R., et al. (2016) An oligotrophic deep-subsurface community dependent on syntrophy is dominated by sulfur-driven autotrophic denitrifiers. Proc. Natl. Acad. Sci. 113: E7927– E7936.

Al Lawati, W.M., Rizoulis, A., Eiche, E., Boothman, C., Polya, D. a., Lloyd, J.R., et al. (2012) Characterisation of organic matter and microbial communities in contrasting arsenic-rich Holocene and arsenic-poor Pleistocene aquifers, Red River Delta, Vietnam. Appl. Geochemistry 27: 315–325.

Lawson, M., Polya, D.A., Boyce, A.J., Bryant, C., and Ballentine, C.J. (2016) Tracing organic matter composition and distribution and its role on arsenic release in shallow Cambodian groundwaters. Geochim. Cosmochim. Acta 178: 160–177.

Lawson, M., Polya, D.A., Boyce, A.J., Bryant, C., Mondal, D., Shantz, A., and Ballentine, C.J. (2013) Pond-derived organic carbon driving changes in arsenic hazard found in Asian groundwaters. Environ. Sci. Technol. 47: 7085–7094.

Lear, G., Song, B., Gault, a G., Polya, D. a, and Lloyd, J.R. (2007) Molecular analysis of arsenate- reducing bacteria within Cambodian sediments following amendment with acetate. Appl. Environ. Microbiol. 73: 1041–8.

Lehr, C.R., Polishchuk, E., Delisle, M.-C., Franz, C., and Cullen, W.R. (2003) Arsenic methylation by micro-organisms isolated from sheepskin bedding materials. Hum. Exp. Toxicol. 22: 325–334.

Lett, M.-C., Muller, D., Lièvremont, D., Silver, S., and Santini, J. (2012) Unified nomenclature for genes involved in prokaryotic aerobic arsenite oxidation. J. Bacteriol. 194: 207–8.

Li, B. and Logan, B.E. (2004) Bacterial adhesion to glass and metal-oxide surfaces. Colloids Surf. B. Biointerfaces 36: 81–90.

Li, H., Li, M., Huang, Y., Rensing, C., and Wang, G. (2013) In silico analysis of bacterial arsenic islands reveals remarkable synteny and functional relatedness between arsenate and phosphate. Front. Microbiol. 4: 347.

Li, P., Wang, Y., Dai, X., Zhang, R., Jiang, Z., Jiang, D., et al. (2015) Microbial community in high arsenic shallow groundwater aquifers in Hetao Basin of Inner Mongolia, China. PLoS One

248 10:.

Li, X. and Krumholz, L.R. (2007) Regulation of arsenate resistance in Desulfovibrio desulfuricans G20 by an arsRBCC operon and an arsC gene. J. Bacteriol. 189: 3705–11.

Lin, Y.-F., Walmsley, A.R., and Rosen, B.P. (2006) An arsenic metallochaperone for an arsenic detoxification pump. Proc. Natl. Acad. Sci. 103: 15617–15622.

Van Lis, R., Nitschke, W., Duval, S., and Schoepp-Cothenet, B. (2013) Arsenics as bioenergetic substrates. Biochim. Biophys. Acta - Bioenerg. 1827:.

Liu, G., Liu, M., Kim, E., Maaty, W.S., Bothner, B., Lei, B., et al. (2012) A periplasmic arsenite‐ binding protein involved in regulating arsenite oxidation. Environ. Microbiol. 14: 1624– 1634.

Lloyd, J. and Oremland, R. (2006) Microbial transformations of arsenic in the environment: from soda lakes to aquifers. Elements 2: 85–90.

Lloyd, J.R. (2003) Microbial reduction of metals and radionuclides. FEMS Microbiol. Rev. 27: 411–425.

Lloyd, J.R., Gault, A.G., Héry, M., and Macrae, J.D. (2011) Microbial transformations of Arsenic in the subsurface. 77–90.

López, D.L., Bundschuh, J., Birkle, P., Armienta, M.A., Cumbal, L., Sracek, O., et al. (2012) Arsenic in volcanic geothermal fluids of Latin America. Sci. Total Environ. 429: 57–75.

Lovley, D.R. (1991) Dissimilatory Fe(III) and Mn(IV) reduction. Microbiol. Rev. 55: 259–287.

Lovley, D.R., Phillips, E.J.P., Lonergan, D.J., and Widma, P.K. (1995) Fe(III) and S0 reduction by Pelobacter carbinolicus. Appl. Environ. Microbiol. 61: 2132–2138.

Loy, A., Küsel, K., Lehner, A., Drake, H.L., and Wagner, M. (2004) Microarray and functional gene analyses of sulfate-reducing prokaryotes in low-sulfate, acidic fens reveal cooccurrence of recognized genera and novel lineages. Appl. Environ. Microbiol. 70: 6998–7009.

Macur, R.E., Jackson, C.R., Botero, L.M., McDermott, T.R., and Inskeep, W.P. (2004) Bacterial populations associated with the oxidation and reduction of arsenic in an unsaturated soil. Environ. Sci. Technol. 38: 104–11.

Macy, J.M., Nunan, K., Iwigen, K.D., Dixon, D.R., Harbour, P.J., Cahill, M., and Sly, L.I. (1996) Bacterium Isolated from Gold Mine Wastewater. 1153–1157.

Macy, J.M., Santini, J.M., Pauling, B. V, O’Neill, a H., and Sly, L.I. (2000) Two new arsenate/sulfate-reducing bacteria: mechanisms of arsenate reduction. Arch. Microbiol. 173: 49–57.

Maeda, S., Ohki, A., Miyahara, K., Takeshita, T., and Higashi, S. (1990) Growth characteristics and arsenic metabolism of two species of arsenic‐tolerant bacteria. Appl. Organomet. Chem. 4: 245–250.

Mailloux, B.J., Trembath-Reichert, E., Cheung, J., Watson, M., Stute, M., Freyer, G. a, et al. (2013) Advection of surface-derived organic carbon fuels microbial reduction in Bangladesh groundwater. Proc. Natl. Acad. Sci. U. S. A. 110: 5331–5.

249 Malasarn, D., Keeffe, J.R., and Newman, D.K. (2008) Characterization of the arsenate respiratory reductase from Shewanella sp. strain ANA-3. J. Bacteriol. 190: 135–42.

Malasarn, D., Saltikov, C.W., Campbell, K.M., Santini, J.M., Hering, J.G., and Newman, D.K. (2004) arrA Is a Reliable Marker for As (V) respiration. Science (80-. ). 4.

Mallick, P. and Kuster, B. (2010) Proteomics: a pragmatic perspective. Nat. Biotechnol. 28: 695–709.

Mardis, E.R. (2008) Next-generation DNA sequencing methods. Annu. Rev. Genomics Hum. Genet. 9: 387–402.

Maron, P.-A., Ranjard, L., Mougel, C., and Lemanceau, P. (2007) Metaproteomics: a new approach for studying functional microbial ecology. Microb. Ecol. 53: 486–93.

Marsili, E., Baron, D.B., Shikhare, I.D., Coursolle, D., Gralnick, J. a, and Bond, D.R. (2008) Shewanella secretes flavins that mediate extracellular electron transfer. Proc. Natl. Acad. Sci. U. S. A. 105: 3968–3973.

Martin, M. (2009) Cutadaps. EmBnet.journal 47: 8–11.

Masella, A.P., Bartram, A.K., Truszkowski, J.M., Brown, D.G., and Neufeld, J.D. (2012) PANDAseq: paired-end assembler for illumina sequences. BMC Bioinformatics 13: 31.

Masscheleyn, P.H., Delaune, R.D., and Patrick, Jr., W.H. (1991) Effect of Redox Potential and pH on Arsenic Speciation and Solubility in a Contaminated Soil. Environ. Sci. Technol. 25: 1414–1419.

McArthur, J.M., Ravenscroft, P., Safiulla, S., and Thirlwall, M.F. (2001) Arsenic in groundwater: testing pollution mechanisms for sedimentary aquifers in Bangladesh. Water Resour. Res. 37: 109–117.

McCarty, K.M., Chen, Y.-C., Quamruzzaman, Q., Rahman, M., Mahiuddin, G., Hsueh, Y.-M., et al. (2007) Arsenic methylation, GSTT1, GSTM1, GSTP1 polymorphisms, and skin lesions. Environ. Health Perspect. 115: 341–5. de Melo Oliveira, M.G., Abels, S., Zbinden, R., Bloemberg, G. V, and Zbinden, A. (2013) Accurate identification of fastidious Gram-negative rods: integration of both conventional phenotypic methods and 16S rRNA gene analysis. BMC Microbiol. 13: 162.

Mestrot, A., Merle, J.K., Broglia, A., Feldmann, J., and Krupp, E.M. (2011) Atmospheric stability of arsine and methylarsines. Environ. Sci. Technol. 45: 4010–4015.

Metzker, M.L. (2010) Sequencing technologies - the next generation. Nat. Rev. Genet. 11: 31– 46.

Meyer, B. and Kuevert, J. (2007) Phylogeny of the alpha and beta subunits of the dissimilatory adenosine-5??? -phosphosulfate (APS) reductase from sulfate-reducing prokaryotes - Origin and evolution of the dissimilatory sulfate-reduction pathway. Microbiology 153: 2026–2044.

Meyer, J., Michalke, K., Kouril, T., and Hensel, R. (2008) Volatilisation of metals and metalloids: an inherent feature of methanoarchaea? Syst. Appl. Microbiol. 31: 81–87.

Michalke, K., Wickenheiser, E.B., Mehring, M., Hirner, A. V, and Hensel, R. (2000) Production of volatile derivatives of metal (loid) s by microflora involved in anaerobic digestion of

250 sewage sludge. Appl. Environ. Microbiol. 66: 2791–2796.

Mihajlov, I., Stute, M., Schlosser, P., Mailloux, B.J., Zheng, Y., Choudhury, I., et al. (2016) Recharge of low‐arsenic aquifers tapped by community wells in Araihazar, Bangladesh, inferred from environmental isotopes. Water Resour. Res. 52: 3324–3349.

Miller, C.S., Baker, B.J., Thomas, B.C., Singer, S.W., and Banfield, J.F. (2011) EMIRGE: reconstruction of full-length ribosomal genes from microbial community short read sequencing data. Genome Biol. 12: R44.

Miller, W.H., Schipper, H.M., Lee, J.S., Singer, J., and Waxman, S. (2002) Mechanisms of Action of Arsenic Trioxide Mechanisms of Action of Arsenic Trioxide 1. 3893–3903.

Mirza, B.S., Muruganandam, S., Meng, X., Sorensen, D.L., Dupont, R.R., and McLean, J.E. (2014) Arsenic(V) reduction in relation to Iron(III) transformation and molecular characterization of the structural and functional microbial community in sediments of a basin-fill aquifer in Northern Utah. Appl. Environ. Microbiol. 80: 3198–208.

Moriya, Y., Itoh, M., Okuda, S., and Kanehisa, M. (2005) KAAS: KEGG automatic annotation server. Genome Informatics 5: 2005.

Moriya, Y., Itoh, M., Okuda, S., Yoshizawa, A.C., and Kanehisa, M. (2007) KAAS: An automatic genome annotation and pathway reconstruction server. Nucleic Acids Res. 35:.

Mukai, H., Ambe, Y., Muku, T., Takeshita, K., and Fukuma, T. (1986) Seasonal variation of methylarsenic compounds in airborne participate matter. Nature 324: 239–241.

Mukherjee, A., Sengupta, M.K., Hossain, M.A., Ahamed, S., Das, B., Nayak, B., et al. (2006) Arsenic contamination in groundwater: a global perspective with emphasis on the Asian scenario. J. Health. Popul. Nutr. 24: 142–63.

Mukhopadhyay, R. and Rosen, B.P. (2002a) Arsenate reductases in prokaryotes and eukaryotes. Environ. Health Perspect. 110: 745.

Mukhopadhyay, R. and Rosen, B.P. (2002b) Arsenate reductases in prokaryotes and eukaryotes. Environ. Health Perspect. 110 Suppl: 745–748.

Mukhopadhyay, R., Rosen, B.P., Phung, L.T., and Silver, S. (2002) Microbial arsenic: from geocycles to genes and enzymes. FEMS Microbiol. Rev. 26: 311–325.

National Research Council (US) Committee on Medical and Biological Effects of Environmental Pollutants (1977) Arsenic: Medical and Biologic Effects of Environmental Pollutants.

Newman, D.K. and Beveridge, T.J. (1997) Precipitation of Arsenic Trisul de by. Microbiology 63: 2022–2028.

Newman, D.K., Beveridge, T.J., and Morel, F. (1997) Precipitation of arsenic trisulfide by Desulfotomaculum auripigmentum. Appl. Environ. Microbiol. 63: 2022–2028.

Nickson, R.T., Mcarthur, J.M., Ravenscroft, P., Burgess, W.G., and Ahmed, K.M. (2000) Mechanism of arsenic release to groundwater, Bangladesh and West Bengal. Appl. Geochemistry 15: 403–413.

Niggemyer, A., Spring, S., and Stackebrandt, E. (2001) Isolation and Characterization of a Novel As ( V ) -Reducing Bacterium : Implications for Arsenic Mobilization and the Genus Desulfitobacterium Isolation and Characterization of a Novel As ( V ) -Reducing

251 Bacterium : Implications for Arsenic Mobilization. Appl. Environ. Microbiol. 67: 5568– 5580.

Nordstrom, D.K. (2002) Public health. Worldwide occurrences of arsenic in ground water. Science (80-. ). 296: 2143–5.

Nriagu, J. (2001) Arsenic Poisoning Through the Ages. In, Environmental Chemistry of Arsenic., pp. 1–26.

Nyrén, P. (2007) The history of pyrosequencing. Methods Mol. Biol. 373: 1–14.

O’Day, P. a, Vlassopoulos, D., Root, R., and Rivera, N. (2004) The influence of sulfur and iron on dissolved arsenic concentrations in the shallow subsurface under changing redox conditions. Proc. Natl. Acad. Sci. U. S. A. 101: 13703–13708.

Ogata, H., Goto, S., Sato, K., Fujibuchi, W., Bono, H., and Kanehisa, M. (1999) KEGG: Kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 27: 29–34.

Ogram, A., Sayler, S., and Barkay, T. (1987) The extraction and purification of microbial D N A from sediments. 7: 57–66.

Omoregie, E.O., Couture, R.-M., Van Cappellen, P., Corkhill, C.L., Charnock, J.M., Polya, D. a, et al. (2013) Arsenic bioremediation by biogenic iron oxides and sulfides. Appl. Environ. Microbiol. 79: 4325–35.

Oremland, R. and J.F.S. (2013) The Ecology of Arsenic Author ( s ): Ronald S . Oremland and John F . Stolz Source : Science , New Series , Vol . 300 , No . 5621 ( May 9 , 2003 ), pp . 939-944 Published by : American Association for the Advancement of Science Stable URL : http://www.jst. Science (80-. ). 300: 939–944.

Oremland, R.S., Kulp, T.R., Blum, J.S., Hoeft, S.E., Baesman, S., Miller, L.G., and Stolz, J.F. (2005) A microbial arsenic cycle in a salt-saturated, extreme environment. Science 308: 1305– 1308.

Oremland, R.S. and Stolz, J.F. (2003) The ecology of arsenic. Science 300: 939–44.

Osborne, T.H., Mcarthur, J.M., Sikdar, P.K., and Santini, J.M. (2015) Isolation of an Arsenate- Respiring Bacterium from a Redox Front in an Arsenic-Polluted Aquifer in West Bengal, Bengal Basin.

Páez-Espino, D., Tamames, J., de Lorenzo, V., and Cánovas, D. (2009) Microbial responses to environmental arsenic. Biometals 22: 117–30.

Paisio, C.E., Talano, M.A., González, P.S., Busto, V.D., Talou, J.R., and Agostini, E. (2012) Isolation and characterization of a Rhodococcus strain with phenol-degrading ability and its potential use for tannery effluent biotreatment. Environ. Sci. Pollut. Res. 19: 3430– 3439.

Panaro, N.J., Yuen, P.K., Sakazume, T., Fortina, P., Kricka, L.J., and Wilding, P. (2000) Evaluation of DNA fragment sizing and quantification by the Agilent 2100 bioanalyzer. Clin. Chem. 46: 1851–1853.

Park, H.S., Kim, B.H., Kim, H.S., Kim, H.J., Kim, G., Kim, M., et al. (2001) A Novel Electrochemically Active and Fe ( III ) -reducing Bacterium Phylogenetically Related to Clostridium butyricum Isolated from a Microbial Fuel Cell. Scanning 7: 297–306.

252 Paul, D., Kazy, S.K., Gupta, A.K., Pal, T., and Sar, P. (2015) Diversity, metabolic properties and arsenic mobilization potential of indigenous bacteria in arsenic contaminated groundwater of West Bengal, India. PLoS One 10:.

Pearcy, C. a., Chevis, D. a., Haug, T.J., Jeffries, H. a., Yang, N., Tang, J., et al. (2011) Evidence of microbially mediated arsenic mobilization from sediments of the Aquia aquifer, Maryland, USA. Appl. Geochemistry 26: 575–586.

Pederick, R.L., Gault, a G., Charnock, J.M., Polya, D. a, and Lloyd, J.R. (2007a) Probing the biogeochemistry of arsenic: response of two contrasting aquifer sediments from Cambodia to stimulation by arsenate and ferric iron. J. Environ. Sci. Health. A. Tox. Hazard. Subst. Environ. Eng. 42: 1763–74.

Pederick, R.L., Gault, a G., Charnock, J.M., Polya, D. a, and Lloyd, J.R. (2007b) Probing the biogeochemistry of arsenic: response of two contrasting aquifer sediments from Cambodia to stimulation by arsenate and ferric iron. J. Environ. Sci. Health. A. Tox. Hazard. Subst. Environ. Eng. 42: 1763–74.

Peng, Y., Leung, H.C.M., Yiu, S.M., and Chin, F.Y.L. (2012) IDBA-UD: A de novo assembler for single-cell and metagenomic sequencing data with highly uneven depth. Bioinformatics 28: 1420–1428.

Pérez-Jiménez, J.R., DeFraia, C., and Young, L.Y. (2005) Arsenate respiratory reductase gene (arrA) for Desulfosporosinus sp. strain Y5. Biochem. Biophys. Res. Commun. 338: 825– 829.

Pichler, T., Veizer, J., and Hall, G.E.M. (1999) The natural input of extremely high arsenic concentrations into a coral reef ecosystem by hydrothermal fluids and its removal by Fe(III) oxyhydroxides. Environ. Sci. Technol. 33: 1373–1378.

Piotr, B., Robert, W., Grzegorz, O., Andreu, G., and Stanislaw, U. (1997) Isolation of three contiguous genes, ACR1, ACR2 and ACR3, involved in resistance to arsenic compounds in the yeast Saccharomyces cerevisiae. Yeast 13: 819–828.

Polizzotto, M.L., Kocar, B.D., Benner, S.G., Sampson, M., and Fendorf, S. (2008) Near-surface wetland sediments as a source of arsenic release to ground water in Asia. Nature 454: 505.

Polya, D.A., Gault, A.G., Diebe, N., Feldman, P., Rosenboom, J.W., Gilligan, E., et al. (2005) Arsenic hazard in shallow Cambodian groundwaters. Mineral. Mag. 69: 807–823.

Polya, D.A. and Middleton, D.R.S. (2017) Arsenic in drinking water: sources & human exposure. Best Pract. Guid. Control Arsen. Drink. Water 1.

Pourmand, N., Elahi, E., Davis, R.W., and Ronaghi, M. (2002) Multiplex Pyrosequencing. Nucleic Acids Res. 30: e31.

Prasad, K.S., Srivastava, P., Subramanian, V., and Paul, J. (2011) Biosorption of As (III) ion on Rhodococcus sp. WB-12: biomass characterization and kinetic studies. Sep. Sci. Technol. 46: 2517–2525.

Prieto, D.M., Devesa-Rey, R., Rubinos, D. a, Díaz-Fierros, F., and Barral, M.T. (2013) Arsenate retention by epipsammic biofilms developed on streambed sediments: influence of phosphate. Biomed Res. Int. 2013: 591634.

253 Qin, J., Rosen, B.P., Zhang, Y., Wang, G., Franke, S., and Rensing, C. (2006a) Arsenic detoxification and evolution of trimethylarsine gas by a microbial arsenite S- adenosylmethionine methyltransferase. 103:.

Qin, J., Rosen, B.P., Zhang, Y., Wang, G., Franke, S., and Rensing, C. (2006b) Arsenic detoxification and evolution of trimethylarsine gas by a microbial arsenite S- adenosylmethionine methyltransferase. Proc. Natl. Acad. Sci. U. S. A. 103: 2075–2080.

Rahman, M.A. and Hassler, C. (2014) Is arsenic biotransformation a detoxification mechanism for microorganisms? Aquat. Toxicol. 146:.

Rambaut, A. (2009) FigTree, version 1.3. 1. Comput. Progr. Distrib. by author, website http//treebioedacuk/software/figtree/[accessed January 4, 2011].

Ravenscroft, P., Brammer, H., and Richards, K. (2009) Arsenic Pollution: A Global Synthesis, RGS-IBG Book Series, A John Wiley & Sons Ltd. Publ. United Kingdom.

Reddy, G.S.N., Matsumoto, G.I., Schumann, P., Stackerbrandt, E., and Shivaji, S. (2004) Psychrophilic pseudomonads from Antarctica: Pseudomonas antartica sp. nov., Pseudomonas meridiana sp. nov. and Pseudomonas proteolytica sp. nov. Int. J. Syst. Evol. Microbiol. 54: 713–719.

Rhine, E.D., Ní Chadhain, S.M., Zylstra, G.J., and Young, L.Y. (2007) The arsenite oxidase genes (aroAB) in novel chemoautotrophic arsenite oxidizers. Biochem. Biophys. Res. Commun. 354: 662–667.

Richards, L.A., Magnone, D., van Dongen, B.E., Ballentine, C.J., and Polya, D.A. (2015) Use of lithium tracers to quantify drilling fluid contamination for groundwater monitoring in Southeast Asia. Appl. Geochemistry 63: 190–202.

Richards, L.A., Magnone, D., Sovann, C., Kong, C., Uhlemann, S., Kuras, O., et al. (2017) Science of the Total Environment High resolution pro fi le of inorganic aqueous geochemistry and key redox zones in an arsenic bearing aquifer in Cambodia. Sci. Total Environ. 590–591: 540–553.

Richards, L.A., Sültenfuß, J., Ballentine, C.J., Magnone, D., van Dongen, B.E., Sovann, C., and Polya, D.A. (2017) Tritium Tracers of Rapid Surface Water Ingression into Arsenic-bearing Aquifers in the Lower Mekong Basin, Cambodia. Procedia Earth Planet. Sci. 17: 845–848.

Richey, C., Chovanec, P., Hoeft, S.E., Oremland, R.S., Basu, P., and Stolz, J.F. (2009) Biochemical and Biophysical Research Communications Respiratory arsenate reductase as a bidirectional enzyme. Biochem. Biophys. Res. Commun. 382: 298–302.

Riesenfeld, C.S., Schloss, P.D., and Handelsman, J. (2004) Metagenomics: genomic analysis of microbial communities. Annu. Rev. Genet. 38: 525–52.

Rittle, K.A., Drever, J.I., and Colberg, P.J.S. (1995) Precipitation of Arsenic during Bacterial Sulfate Reduction. Geomicrobiol. J. 13: 1–11.

Rizoulis, A., Lawati, A.W.M. Al, Pancost, B.R.D., Polya, D.A., A, A.B.E.V.D., and A, J.R.L. (2014) Microbially mediated reduction of Fe III and As V in Cambodian sediments amended with 13 C-labelled hexadecane and kerogen. 538–546.

Rizoulis, A., Al Lawati, W.M., Pancost, R.D., Polya, D.A., van Dongen, B.E., and Lloyd, J.R. (2014) Microbially mediated reduction of Fe III and As V in Cambodian sediments amended with

254 13 C-labelled hexadecane and kerogen. Environ. Chem. 11: 538–546.

Rizoulis, A., Al Lawati, W.M., Pancost, R.D., Polya, D. a., van Dongen, B.E., and Lloyd, J.R. (2014) Microbially mediated reduction of FeIII and AsV in Cambodian sediments amended with 13C-labelled hexadecane and kerogen. Environ. Chem. 11: 538.

Rodrigues, D.F., Goris, J., Vishnivetskaya, T., Gilichinsky, D., Thomashow, M.F., and Tiedje, J.M. (2006) Characterization of Exiguobacterium isolates from the Siberian permafrost. Description of Exiguobacterium sibiricum sp. nov. Extremophiles 10: 285–294.

Roesch, L.F.W., Fulthorpe, R.R., Riva, A., Casella, G., Hadwin, A.K.M., Kent, A.D., et al. (2007) Pyrosequencing enumerates and contrasts soil microbial diversity. ISME J. 1: 283–90.

Rognes, T., Flouri, T., Nichols, B., Quince, C., and Mahé, F. (2016) VSEARCH: a versatile open source tool for metagenomics. PeerJ 4: e2584.

Ronaghi, M. (2001) Pyrosequencing Sheds Light on DNA Sequencing Pyrosequencing Sheds Light on DNA Sequencing. 3–11.

Rosen, B.P. (1999) Families of arsenic transporters. Trends Microbiol. 7: 207–12.

Rosenberg, H., Gerdes, R.G., and Chegwidden, K. (1977) Two systems for the uptake of phosphate in Escherichia coli. J. Bacteriol. 131: 505–511.

Rowland, H. a. L., Pederick, R.L., Polya, D. a., Pancost, R.D., Van Dongen, B.E., Gault, a. G., et al. (2007) The control of organic matter on microbially mediated iron reduction and arsenic release in shallow alluvial aquifers, Cambodia. Geobiology 5: 281–292.

Rowland, H. a. L., Polya, D. a., Lloyd, J.R., and Pancost, R.D. (2006) Characterisation of organic matter in a shallow, reducing, arsenic-rich aquifer, West Bengal. Org. Geochem. 37: 1101–1114.

Rowland, H. a L., Boothman, C., Pancost, R., Gault, a G., Polya, D. a, and Lloyd, J.R. (2002) The role of indigenous microorganisms in the biodegradation of naturally occurring petroleum, the reduction of iron, and the mobilization of arsenite from west bengal aquifer sediments. J. Environ. Qual. 38: 1598–607.

Rowland, H. a L., Gault, A.G., Lythgoe, P., and Polya, D. a. (2008) Geochemistry of aquifer sediments and arsenic-rich groundwaters from Kandal Province, Cambodia. Appl. Geochemistry 23: 3029–3046.

Ryan, D. and Colleran, E. (2002) Arsenical resistance in the IncHI2 plasmids. Plasmid 47: 234– 240.

Saltikov, C.W., Cifuentes, A., Venkateswaran, K., and Newman, D.K. (2003) The ars Detoxification System Is Advantageous but Not Required for As ( V ) Respiration by the Genetically Tractable Shewanella Species Strain ANA-3 The ars Detoxification System Is Advantageous but Not Required for As ( V ) Respiration by the Genetically.

Saltikov, C.W. and Newman, D.K. (2003) Genetic identification of a respiratory arsenate reductase. Proc. Natl. Acad. Sci. U. S. A. 100: 10983–8.

Saltikov, C.W., Wildman, R.A., and Newman, D.K. (2005) Expression Dynamics of Arsenic Respiration and Detoxification in. 187: 7390–7396.

Santini, J.M., Kappler, U., Ward, S.A., Honeychurch, M.J., vanden Hoven, R.N., and Bernhardt,

255 P. V (2007) The NT-26 cytochrome c 552 and its role in arsenite oxidation. Biochim. Biophys. Acta (BBA)-Bioenergetics 1767: 189–196.

Sanyal, S.K., Mou, T.J., Chakrabarty, R.P., Hoque, S., Hossain, M.A., and Sultana, M. (2016) Diversity of arsenite oxidase gene and arsenotrophic bacteria in arsenic affected Bangladesh soils. AMB Express 6:.

Sardiwal, S., Santini, J.M., Osborne, T.H., and Djordjevic, S. (2010) Characterization of a two- component signal transduction system that controls arsenite oxidation in the chemolithoautotroph NT-26. FEMS Microbiol. Lett. 313: 20–28.

Sarkar, A., Kazy, S.K., and Sar, P. (2013) Characterization of arsenic resistant bacteria from arsenic rich groundwater of West Bengal, India. Ecotoxicology 22:.

Sato, T. and Kobayashi, Y. (1998) The ars Operon in the skin Element of Bacillus subtilis Confers Resistance to Arsenate and Arsenite The ars Operon in the skin Element of Bacillus subtilis Confers Resistance to Arsenate and Arsenite. 180:.

Scheff, G., Salcher, O., and Lingens, F. (1984) Trichococcus flocculiformis gen. nov. sp. nov. A new gram-positive fdamentous bacterium isolated from bulking sludge. Appl. Microbiol. Biotechnol. 19: 114–119.

Schirmer, M., Ijaz, U.Z., D’Amore, R., Hall, N., Sloan, W.T., and Quince, C. (2015) Insight into biases and sequencing errors for amplicon sequencing with the Illumina MiSeq platform. Nucleic Acids Res. 43:.

Schloss, P.D. and Handelsman, J. (2005) Metagenomics for studying unculturable microorganisms: cutting the Gordian knot. Genome Biol. 6: 229.

Searle, B.C. (2010) Scaffold: A bioinformatic tool for validating MS/MS‐based proteomic studies. Proteomics 6: 1265–1269.

Segata, N., Boernigen, D., Tickle, T.L., Morgan, X.C., Garrett, W.S., and Huttenhower, C. (2013) Computational meta’omics for microbial community studies. Mol. Syst. Biol. 9: 666.

Senn, D.B. and Hemond, H.F. (2002) Nitrate controls on iron and arsenic in an urban lake. Science 296: 2373–6.

Shapiro, a L., Viñuela, E., and Maizel, J. V (1967) Molecular weight estimation of polypeptide chains by electrophoresis in SDS-polyacrylamide gels. Biochem. Biophys. Res. Commun. 28: 815–820.

Shariatpahani, M., Anderson, A.C., Abdelghani, A.A., and Englande, A.J. (1983) Microbial metabolism of an organic arsenical herbicide. In, Biodeterioration 5: papers presented at the 5th International Biodeterioration Symposium, Aberdeen, September, 1981/edited by TA Oxley and S. Barry. Chichester [Sussex]: Wiley, c1983.

Sharma, A.K., Tjell, J.C., Sloth, J.J., and Holm, P.E. (2014) Review of arsenic contamination, exposure through water and food and low cost mitigation options for rural areas. Appl. Geochemistry 41: 11–33.

Shelobolina, E.S., Vrionis, H. a., Findlay, R.H., and Lovley, D.R. (2008) Geobacter uraniireducens sp. nov., isolated from subsurface sediment undergoing uranium bioremediation. Int. J. Syst. Evol. Microbiol. 58: 1075–1078.

256 Shendure, J. and Lieberman Aiden, E. (2012) The expanding scope of DNA sequencing. Nat. Biotechnol. 30: 1084–94.

Shrestha, P.M., Noll, M., and Liesack, W. (2007) Phylogenetic identity, growth-response time and rRNA operon copy number of soil bacteria indicate different stages of community succession. Environ. Microbiol. 9: 2464–2474.

Silver, S. and Phung, L.T. (2005) Genes and Enzymes Involved in Bacterial Oxidation and Reduction of Inorganic Arsenic MINIREVIEW Genes and Enzymes Involved in Bacterial Oxidation and Reduction of Inorganic Arsenic. 71:.

Slyemi, D. and Bonnefoy, V. (2012) How prokaryotes deal with arsenic(†). Environ. Microbiol. Rep. 4: 571–86.

Smedley, P.. and Kinniburgh, D.. (2002) A review of the source, behaviour and distribution of arsenic in natural waters. Appl. Geochemistry 17: 517–568.

Smith, E.R.G., Naidu, R., and Alston, A.M. (1998) Arsenic in the soil environment.

Song, B., Chyun, E., Jaffé, P.R., and Ward, B.B. (2009) Molecular methods to detect and monitor dissimilatory arsenate-respiring bacteria (DARB) in sediments. FEMS Microbiol. Ecol. 68: 108–17.

Song, Y., Deng, S.P., Acosta-Martínez, V., and Katsalirou, E. (2008) Characterization of redox- related soil microbial communities along a river floodplain continuum by fatty acid methyl ester (FAME) and 16S rRNA genes. Appl. soil Ecol. 40: 499–509.

Sorg, T.J., Chen, A.S.C., and Wang, L. (2014) Arsenic species in drinking water wells in the USA with high arsenic concentrations. Water Res. 48: 156–169.

Sovann, C. and Polya, D.A. (2014) Improved groundwater geogenic arsenic hazard map for Cambodia. Environ. Chem. 11: 595–607.

Stevens, H., Brinkhoff, T., Rink, B., Vollmers, J., and Simon, M. (2007) Diversity and abundance of Gram positive bacteria in a tidal flat ecosystem. Environ. Microbiol. 9: 1810–1822.

Stirling, D. and Bartlett, J.M.S. (2003) PCR Protocols Humana Press.

Stolz, J.F. and Oremland, R.S. (1999) Bacterial respiration of arsenic and selenium. FEMS Microbiol. Rev. 23: 615–627.

Stupperich, E. (1993) Recent advances in elucidation of biological corrinoid functions. FEMS Microbiol. Rev. 12: 349–365.

Stute, M., Zheng, Y., Schlosser, P., Horneman, A., Dhar, R.K., Datta, S., et al. (2007) Hydrological control of As concentrations in Bangladesh groundwater. Water Resour. Res. 43: 1–11.

Sun, J., Bostick, B.C., Mailloux, B.J., Ross, J.M., and Chillrud, S.N. (2016) Effect of oxalic acid treatment on sediment arsenic concentrations and lability under reducing conditions. J. Hazard. Mater. 311: 125–133.

Sun, J., Chillrud, S.N., Mailloux, B.J., Stute, M., Singh, R., Dong, H., et al. (2016) Enhanced and stabilized arsenic retention in microcosms through the microbial oxidation of ferrous iron by nitrate. Chemosphere 144: 1106–1115.

Sun, J., Quicksall, A.N., Chillrud, S.N., Mailloux, B.J., and Bostick, B.C. (2016) Arsenic

257 mobilization from sediments in microcosms under sulfate reduction. Chemosphere 153:.

Sung, Y., Sung, Y., Fletcher, K.E., Fletcher, K.E., Ritalahti, K.M., Ritalahti, K.M., et al. (2006) Geobacter lovleyi. Society 72: 2775–2782.

Switzer Blum, J., Burns Bindi, A., Buzzelli, J., Stolz, J., and Oremland, R. (1998) Bacillus arsenicoselenatis, sp. nov., and Bacillus selenitireducens, sp. nov.: two haloalkaliphiles from Mono Lake, California that respire oxyanions of selenium and arsenic. Arch. Microbiol. 171: 19–30.

Tamaki, S. and Frankenberger Jr, W.T. (1992) Environmental biochemistry of arsenic. In, Reviews of environmental contamination and toxicology. Springer, pp. 79–110.

Tisa, L.S. and Rosen, B.P. (1990) Molecular characterization of an anion pump. The ArsB protein is the membrane anchor for the ArsA protein. J. Biol. Chem. 265: 190–194.

Trapnell, C., Roberts, A., Goff, L., Pertea, G., Kim, D., Kelley, D.R., et al. (2012) Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. Nat. Protoc. 7: 562–578.

Trygg, J. and Wold, S. (2003) O2-PLS, a two-block (X-Y) latent variable regression (LVR) method with an integral OSC filter. J. Chemom. 17: 53–64.

Trygg, J. and Wold, S. (2002) Orthogonal projections to latent structures (O-PLS). J. Chemom. 16: 119–128.

Tsai, S.-L., Singh, S., and Chen, W. (2009) Arsenic metabolism by microbes in nature and the impact on arsenic remediation. Curr. Opin. Biotechnol. 20: 659–67.

Turpeinen, R., Pantsar-Kallio, M., and Kairesalo, T. (2002) Role of microbes in controlling the speciation of arsenic and production of arsines in contaminated soils. Sci. Total Environ. 285: 133–145.

Urich, T., Lanzén, A., Qi, J., Huson, D.H., Schleper, C., and Schuster, S.C. (2008) Simultaneous assessment of soil microbial community structure and function through analysis of the meta-transcriptome. PLoS One 3:.

Vesterberg, O. (1993) A short history of electrophoretic methods. Electrophoresis 14: 1243– 1249.

Villegas-torres, M.F., Bedoya-reina, O.C., Salazar, C., Vives-florez, M.J., and Dussan, J. (2011) International Biodeterioration & Biodegradation Horizontal arsC gene transfer among microorganisms isolated from arsenic polluted soil. Int. Biodeterior. Biodegradation 65: 147–152.

Vithanage, M. (2017) Arsenic Research and global Sustainability—As 2016 Bhattacharya et al.,(Eds) ISBN 978-1-138-02941-5.

Vogelstein, B. and Gillespie, D. (1979) Preparative and analytical purification of DNA from agarose. Proc. Natl. Acad. Sci. U. S. A. 76: 615–619.

Wade, W. (2002) Unculturable bacteria—the uncharacterized organisms that cause oral infections. J. R. Soc. Med. 95: 81.

Wagner, M., Roger, A.J., Flax, J.L., Gregory, A., Stahl, D.A., and Brusseau, G.A. (1998) Phylogeny of Dissimilatory Sulfite Reductases Supports an Early Origin of Sulfate Respiration

258 Phylogeny of Dissimilatory Sulfite Reductases Supports an Early Origin of Sulfate Respiration. J. Bacteriol. 180:.

Wall, P.K., Leebens-Mack, J., Chanderbali, A.S., Barakat, A., Wolcott, E., Liang, H., et al. (2009) Comparison of next generation sequencing technologies for transcriptome characterization. BMC Genomics 10: 347.

Wang, P., Sun, G., Jia, Y., Meharg, A. a, and Zhu, Y. (2014) A review on completing arsenic biogeochemical cycle: Microbial volatilization of arsines in environment. J. Environ. Sci. 26: 371–381.

Wang, X.H., Zhang, C., Yang, L.L., Li, S., Zhang, Y., and Gomes-Laranjo, J. (2011) Screening and identification of microbial strains that secrete an extracellular C-7 xylosidase of taxanes. World J. Microbiol. Biotechnol. 27: 627–635.

Warnecke, F. and Hess, M. (2009) A perspective: metatranscriptomics as a tool for the discovery of novel biocatalysts. J. Biotechnol. 142: 91–5.

Watanabe, T., Miura, A., Iwata, T., Kojima, H., and Fukui, M. (2017) Dominance of Sulfuritalea species in nitrate‐depleted water of a stratified freshwater lake and arsenate respiration ability within the genus. Environ. Microbiol. Rep.

Watts, M.J., O’Reilly, J., Marcilla, A.L., Shaw, R.A., and Ward, N.I. (2010) Field based speciation of arsenic in UK and Argentinean water samples. Environ. Geochem. Health 32: 479–490.

Wauters, G., De Baere, T., Willems, A., Falsen, E., and Vaneechoutte, M. (2003) Description of Comamonas aquatica comb. nov. and Comamonas kerstersii sp. nov. for two subgroups of Comamonas terrigena and emended description of Comamonas terrigena. Int. J. Syst. Evol. Microbiol. 53: 859–862.

Webster, J.G. and Nordstrom, D.K. (2003) Geothermal arsenic. In, Arsenic in ground water. Springer, pp. 101–125.

Weeger, W., Lièvremont, D., Perret, M., Lagarde, F., Hubert, J.C., Leroy, M., and Lett, M.C. (1999) Oxidation of arsenite to arsenate by a bacterium isolated from an aquatic environment. Biometals 12: 141–9.

Weier HU; Gray JW. (1988) A Programmable System to Perform the Polymerase Chain Reaction. Dna 7: 441–447.

Whaley-Martin, K.J., Mailloux, B.J., Van Geen, A., Bostick, B.C., Silvern, R.F., Kim, C., et al. (2016) Stimulation of Microbially Mediated Arsenic Release in Bangladesh Aquifers by Young Carbon Indicated by Radiocarbon Analysis of Sedimentary Bacterial Lipids. Environ. Sci. Technol. 50:.

WHO, G. (2011) Guidelines for drinking-water quality. World Heal. Organ. 216: 303–304.

Williamson, A.J., Morris, K., Shaw, S., Byrne, J.M., Boothman, C., and Lloyd, J.R. (2013) Microbial reduction of Fe(III) under alkaline conditions relevant to geological disposal. Appl. Environ. Microbiol. 79: 3320–3326.

Wu, M. and Scott, A.J. (2012) Phylogenomic analysis of bacterial and archaeal sequences with AMPHORA2. Bioinformatics 28: 1033–1034.

Xiao, K.Q., Li, L.G., Ma, L.P., Zhang, S.Y., Bao, P., Zhang, T., and Zhu, Y.G. (2016) Metagenomic

259 analysis revealed highly diverse microbial arsenic metabolism genes in paddy soils with low-arsenic contents. Environ. Pollut. 211:.

Xue, X.-M., Yan, Y., Xu, H.-J., Wang, N., Zhang, X., and Ye, J. (2014) ArsH from Synechocystis sp. PCC 6803 reduces chromate and ferric iron. FEMS Microbiol. Lett. 356: 105–112.

Yachdav, G., Kloppmann, E., Kajan, L., Hecht, M., Goldberg, T., Hamp, T., et al. (2014) PredictProtein - An open resource for online prediction of protein structural and functional features. Nucleic Acids Res. 42: 337–343.

Yamamura, S. and Amachi, S. (2014) Microbiology of inorganic arsenic: From metabolism to bioremediation. J. Biosci. Bioeng. xx: 1–9.

Yamamura, S., Yamamoto, N., Ike, M., and Fujita, M. (2005) Arsenic extraction from solid phase using a dissimilatory arsenate-reducing bacterium. J. Biosci. Bioeng. 100: 219–222.

Yan, L., Yin, H., Zhang, S., Leng, F., Nan, W., and Li, H. (2010) Biosorption of inorganic and organic arsenic from aqueous solution by Acidithiobacillus ferrooxidans BY-3. J. Hazard. Mater. 178: 209–217.

Ye, J., Coulouris, G., Zaretskaya, I., Cutcutache, I., Rozen, S., and Madden, T.L. (2012) Primer- BLAST: a tool to design target-specific primers for polymerase chain reaction. BMC Bioinformatics 13: 134.

Yeates, C., Gillings, M.R., Davison, A.D., Altavilla, N., and Veal, D.A. (1998) Methods for microbial DNA extraction from soil for PCR amplification. Biol. Proced. Online 1: 40–47.

Ying, S.C., Masue-Slowey, Y., Kocar, B.D., Griffis, S.D., Webb, S., Marcus, M. a., et al. (2013) Distributed microbially- and chemically-mediated redox processes controlling arsenic dynamics within Mn-/Fe-oxide constructed aggregates. Geochim. Cosmochim. Acta 104: 29–41.

Yoshinaga, M., Cai, Y., and Rosen, B.P. (2011) Demethylation of methylarsonic acid by a microbial community. Environ. Microbiol. 13: 1205–1215.

Younge, A.J., Henschel, R., Brown, J.T., von Laszewski, G., Qiu, J., and Fox, G.C. (2011) Analysis of Virtualization Technologies for High Performance Computing Environments. Cloud Comput. (CLOUD), 2011 IEEE Int. Conf. Cloud Comput. 9–16.

Zafiriadis, I., Ntougias, S., Mirelis, P., Kapagiannidis, a G., and Aivasidis, a (2012) Molecular characterization of denitrifying bacteria isolated from the anoxic reactor of a modified DEPHANOX plant performing enhanced biological phosphorus removal. Water Environ. Res. 84: 475–484.

Zargar, K., Conrad, A., Bernick, D.L., Lowe, T.M., Stolc, V., Hoeft, S., et al. (2012) ArxA, a new clade of arsenite oxidase within the DMSO reductase family of molybdenum oxidoreductases. Environ. Microbiol. 14: 1635–1645.

Zehnder, A.J.B. and Van Groenestijn, J.W. (1990) Accumulation of polyphosphate by Acinetobacter sp.: Physiology, ecology, and application. In, Novel biodegradable microbial polymers. Springer, pp. 235–243.

Zhang, G., Chen, T., Chang, S., Zhang, W., Wu, X., Wu, M., et al. (2016) Complete genome sequence of Acinetobacter sp. TTH0-4, a cold-active crude oil degrading strain isolated from Qinghai-Tibet Plateau. J. Biotechnol. 226: 54–55.

260 Zhang, Z., Moon, H.S., Myneni, S.C.B., and Jaffé, P.R. (2017) Effect of dissimilatory iron and sulfate reduction on arsenic dynamics in the wetland rhizosphere and its bioaccumulation in wetland plants (Scirpus actus). J. Hazard. Mater. 321: 382–389.

Zheng, Y., van Geen, a., Stute, M., Dhar, R., Mo, Z., Cheng, Z., et al. (2005) Geochemical and hydrogeological contrasts between shallow and deeper aquifers in two villages of Araihazar, Bangladesh: Implications for deeper aquifers as drinking water sources. Geochim. Cosmochim. Acta 69: 5203–5218.

Zheng, Y., Stute, M., Van Geen, a., Gavrieli, I., Dhar, R., Simpson, H.J., et al. (2004) Redox control of arsenic mobilization in Bangladesh groundwater. Appl. Geochemistry 19: 201– 214.

Zhu, Y.-G., Xue, X.-M., Kappler, A., Rosen, B.P., and Meharg, A.A. (2017) Linking genes to microbial biogeochemical cycling: lessons from arsenic. Environ. Sci. Technol.

Zhu, Y.-G., Yoshinaga, M., Zhao, F.-J., and Rosen, B.P. (2013) Earth Abides Arsenic Biotransformations. Annu. Rev. Earth Planet. Sci. 42: 140307200240001.

Zobrist, J., Dowdle, P.R., Davis, J. a., and Oremland, R.S. (2000) Mobilization of Arsenite by Dissimilatory Reduction of Adsorbed Arsenate. Environ. Sci. Technol. 34: 4747–4753.

Zul, D., Wanner, G., and Overmann, J. (2008) Massilia brevitalea sp. nov., a novel betaproteobacterium isolated from lysimeter soil. Int. J. Syst. Evol. Microbiol. 58: 1245– 51.

261 Appendix-II: A laboratory experiment on As(V) respiration and dissimilatory arsenate reductase protein expression by Shewanella sp. ANA-3.

The mobilisation of arsenic in the aqueous phase is often attributed to the microbial mediated reduction of As(V). These bacteria reduce As(V) either through the energy-driven detoxification system (Rosen, 1999; Saltikov et al., 2003; Lloyd and Oremland, 2006) or through the respiratory system, where the bacteria gain energy via dissimilatory arsenate reductase (Arr) mediated electron transfer reaction (Newman and Beveridge, 1997; Macy et al., 2000; Oremland and Stolz, 2003; Saltikov et al., 2005; Malasarn et al., 2008).

The respiratory arsenate reductase (Arr), both isolated and characterised from the periplasm of Chrysiogenes arsenatis (Krafft and Macy, 1998), has two subunits namely ArrA and ArrB of 87 kDa and 29 kDa in size, respectively. Since the enzyme uses arsenate alone as the electron acceptor, the gene is switched on only in the presence of arsenate. As the enzyme contains molybdenum along with iron and sulphur, it is placed under the dimethyl sulfoxide (DMSO) reductase family of molybdenum-containing oxidoreductases that includes other terminal reductases used in microbial respiration (Malasarn et al., 2008). The presence of iron and sulphur in the centre acts as prosthetic groups for the enzymatic reactions. The same enzyme is isolated and characterised from other bacteria including Shewanella sp. ANA-3 (Malasarn et al., 2004) and Bacillus selenitireducens (Afkar et al., 2003). Though the size and the N-terminal sequences vary in the subunits of the enzymes, their enzyme activity correlates to one another (Saltikov et al., 2005). Having failed to recover proteins from the soil and water samples in this research, I worked with a workflow (Figure 1) that supports an extraction and analysis of proteins from Shewanella sp. ANA-3. My objective of this experiment was to learn the protein sequencing workflow in order to employ the method in metaproteomics.

As the first step, Shewanella sp. ANA-3 was grown in an anoxic medium containing 10 mM each of arsenate and fumarate as electron acceptors and, 10

262 mM acetate as the electron donor. The triplicate had arsenate, fumarate and a negative control with no electron acceptor. Every two hours the media containing cells were drawn to monitor the growth using spectrophotometer at OD600, and the samples were stored in duplicates to analyse the concentration of the electron acceptors and to analyse protein expression.

Figure 1. Proteomics workflow used to analyse the expression of Arsenate Reductase (ARR) in Shewanella sp. ANA-3

The cells grown in media containing arsenate and fumarate were separately centrifuged at 3220 g for 20 minutes and heat-shocked with LDS buffer and run on the SDS-PAGE to separate the proteins. The gel was dispensed in a selective buffer and concentrated using a centrifuge. The peptides were ionized and fragmented before they were analysed by liquid chromatography— MS/MS using an Ultimate 3000 Rapid Separation LC (RSLC; Dionex Corporation) coupled to a LTQ Velos Pro (Thermo Fisher Scientific) mass spectrometer. The data produced were searched again in the UniProt bacterial database for protein

263 identification. The data were validated using the proteome software Scaffold (Marsili et al., 2008; Fresquet et al., 2015). The arsenate and fumarate concentrations were analysed using ICP-MS.

The results showed that Shewanella sp. ANA-3 utilised both fumarate and arsenate to gain energy for the growth as shown in Figure 2. The protein analysis

Figure 2. Growth of Shewanella sp. ANA-3 using arsenate and fumarate as electron acceptors. A. the concentration of arsenate and fumarate at various time invervals. B. Growth of Shewanella sp. ANA-3 inferred by OD600. showed that dissimilatory arsenate reductase was expressed in the medium containing arsenate where as the medium with fumarate had no expression of the protein. Table-1 presents all peptides obtaines from MS/MS analysis. The highlighted row in the table shows that Shewanella sp. ANA-3 grown in arsenate medium has 6 peptide sequences containing, 95 kDa molybdopterin oxidoreductases (95% confidence) which was confirmed through peptide visualising tool Scaffold. It shows that the arrA gene can be switched on in arsenate concentration of 10 mM.

Table-1. The peptides isolated and analysed using MS/MS ()

Identified Proteins (109) Arsenate Fumarate Flavocytochrome c OS=Shewanella sp. (strain ANA-3) GN=Shewana3_3318 PE=4 7 16 SV=1 Major outer membrane lipoprotein, putative OS=Shewanella sp. (strain ANA-3) 5 5 GN=Shewana3_3077 PE=4 SV=1 Elongation factor Tu 1 OS=Shewanella sp. (strain MR-4) GN=tuf1 PE=3 SV=1 10 6 30S ribosomal protein S6 OS=Shewanella sp. (strain ANA-3) GN=rpsF PE=3 SV=1 4 3 Formate acetyltransferase OS=Shewanella sp. (strain ANA-3) GN=Shewana3_1555 6 7 PE=4 SV=1 Elongation factor G OS=Shewanella sp. (strain ANA-3) GN=fusA PE=3 SV=1 3 6

264 60 kDa chaperonin OS=Shewanella sp. (strain ANA-3) GN=groL PE=3 SV=1 10 3 30S ribosomal protein S5 OS=Shewanella xiamenensis GN=rpsE PE=3 SV=1 5 6 30S ribosomal protein S9 OS=Shewanella xiamenensis GN=rpsI PE=3 SV=1 2 3 30S ribosomal protein S4 OS=Shewanella oneidensis (strain MR-1) GN=rpsD PE=3 5 7 SV=1 Porin, Gram-negative type OS=Shewanella sp. (strain ANA-3) GN=Shewana3_0733 4 6 PE=4 SV=1 50S ribosomal protein L9 OS=Shewanella sp. (strain ANA-3) GN=rplI PE=3 SV=1 3 6 50S ribosomal protein L11 OS=Shewanella sp. (strain ANA-3) GN=rplK PE=3 SV=1 2 2 ATP synthase subunit alpha OS=Shewanella oneidensis (strain MR-1) GN=atpA PE=3 4 5 SV=1 50S ribosomal protein L15 OS=Shewanella xiamenensis GN=rplO PE=3 SV=1 4 4 30S ribosomal protein S1 OS=Shewanella sp. (strain ANA-3) GN=Shewana3_1977 6 6 PE=3 SV=1 50S ribosomal protein L24 OS=Shewanella xiamenensis GN=rplX PE=3 SV=1 5 7 Chaperone protein DnaK OS=Shewanella sp. (strain ANA-3) GN=dnaK PE=3 SV=1 6 4 50S ribosomal protein L3 OS=Shewanella xiamenensis GN=rplC PE=3 SV=1 3 7 Histone family protein DNA-binding protein OS=Shewanella sp. (strain ANA-3) 3 GN=Shewana3_2658 PE=3 SV=1 Molybdopterin oxidoreductase OS=Shewanella putrefaciens (strain CN-32 / ATCC 6 BAA-453) GN=Sputcn32_3826 PE=4 SV=1 Extracellular solute-binding protein, family 1 OS=Shewanella sp. (strain ANA-3) 6 6 GN=Shewana3_4077 PE=4 SV=1 Protein-export protein SecB OS=Shewanella xiamenensis GN=secB PE=3 SV=1 3 2 50S ribosomal protein L2 OS=Shewanella xiamenensis GN=rplB PE=3 SV=1 2 4 50S ribosomal protein L10 OS=Shewanella xiamenensis GN=rplJ PE=3 SV=1 4 3 Uncharacterized protein OS=Shewanella sp. (strain ANA-3) GN=Shewana3_1409 2 PE=4 SV=1 Flagellin domain protein OS=Shewanella sp. (strain ANA-3) GN=Shewana3_1333 5 3 PE=4 SV=1 50S ribosomal protein L1 OS=Shewanella sp. (strain ANA-3) GN=rplA PE=3 SV=1 3 3 30S ribosomal protein S3 OS=Shewanella xiamenensis GN=rpsC PE=3 SV=1 4 2 DNA-directed RNA polymerase subunit alpha OS=Shewanella xiamenensis GN=rpoA 2 2 PE=3 SV=1 50S ribosomal protein L14 OS=Shewanella sp. HN-41 GN=rplN PE=3 SV=1 3 2 Carbamoyl-phosphate synthase large subunit OS=Shewanella sp. (strain ANA-3) 3 4 GN=Shewana3_0968 PE=3 SV=1 50S ribosomal protein L18 OS=Shewanella xiamenensis GN=rplR PE=3 SV=1 3 6 30S ribosomal protein S8 OS=Shewanella xiamenensis GN=rpsH PE=3 SV=1 3 2 OmpA/MotB domain protein OS=Shewanella sp. (strain ANA-3) 3 4 GN=Shewana3_3146 PE=4 SV=1 Thioredoxin OS=Shewanella sp. (strain ANA-3) GN=Shewana3_0409 PE=3 SV=1 4 3 Pyruvate dehydrogenase E1 component OS=Shewanella sp. (strain ANA-3) 3 3 GN=Shewana3_0426 PE=4 SV=1 Trigger factor OS=Shewanella sp. (strain ANA-3) GN=tig PE=3 SV=1 2 4 ATP synthase gamma chain OS=Shewanella oneidensis (strain MR-1) GN=atpG PE=3 2 4 SV=1 50S ribosomal protein L17 OS=Shewanella xiamenensis GN=SXM_0251 PE=3 SV=1 3 2 ATP synthase subunit b OS=Shewanella sp. (strain ANA-3) GN=atpF PE=3 SV=1 2 5 30S ribosomal protein S12 OS=Shewanella sp. (strain ANA-3) GN=rpsL PE=3 SV=1 2 2 Dihydrolipoyl dehydrogenase OS=Shewanella sp. (strain ANA-3) 2 3 GN=Shewana3_0428 PE=3 SV=1 30S ribosomal protein S11 OS=Shewanella xiamenensis GN=rpsK PE=3 SV=1 2 50S ribosomal protein L6 OS=Shewanella baltica (strain OS678) GN=rplF PE=3 SV=1 2 Phosphoenolpyruvate synthase OS=Shewanella sp. (strain ANA-3) 3

265 GN=Shewana3_1736 PE=3 SV=1 50S ribosomal protein L22 OS=Shewanella xiamenensis GN=rplV PE=3 SV=1 2 Glyceraldehyde-3-phosphate dehydrogenase OS=Shewanella xiamenensis 4 GN=SXM_2018 PE=3 SV=1 Enolase OS=Shewanella xiamenensis GN=eno PE=3 SV=1 4 3 50S ribosomal protein L19 OS=Shewanella xiamenensis GN=rplS PE=3 SV=1 2 3 Chaperone protein HtpG OS=Shewanella oneidensis (strain MR-1) GN=htpG PE=3 5 2 SV=1 Outer membrane protein MtrB OS=Shewanella sp. (strain MR-4) 2 GN=Shewmr4_2512 PE=4 SV=1 50S ribosomal protein L5 OS=Shewanella frigidimarina (strain NCIMB 400) GN=rplE 3 PE=3 SV=1 50S ribosomal protein L4 OS=Shewanella xiamenensis GN=rplD PE=3 SV=1 4 MotA/TolQ/ExbB proton channel OS=Shewanella sp. (strain ANA-3) 4 GN=Shewana3_2625 PE=3 SV=1 Glutamine synthetase OS=Shewanella xiamenensis GN=SXM_3644 PE=3 SV=1 3 50S ribosomal protein L32 OS=Shewanella oneidensis (strain MR-1) GN=rpmF PE=3 2 2 SV=1 30S ribosomal protein S7 OS=Shewanella xiamenensis GN=rpsG PE=3 SV=1 3 50S ribosomal protein L23 OS=Shewanella xiamenensis GN=rplW PE=3 SV=1 2 2 DNA-directed RNA polymerase subunit beta OS=Shewanella piezotolerans (strain 2 3 WP3 / JCM 13877) GN=rpoB PE=3 SV=1 30S ribosomal protein S20 OS=Shewanella xiamenensis GN=SXM_2924 PE=4 SV=1 2 Cold shock domain-contain protein OS=Shewanella xiamenensis GN=SXM_2360 2 2 PE=4 SV=1 Dihydropteridine reductase OS=Shewanella sp. (strain ANA-3) GN=Shewana3_0876 2 3 PE=4 SV=1 50S ribosomal protein L15 OS=Halorhodospira halochloris str. A GN=rplO PE=3 SV=1 2 50S ribosomal protein L30 OS=Shewanella xiamenensis GN=rpmD PE=3 SV=1 3 Porphobilinogen deaminase OS=Shewanella sp. (strain ANA-3) GN=hemC PE=3 SV=1 3 6,7-dimethyl-8-ribityllumazine synthase OS=Shewanella xiamenensis GN=ribH PE=3 2 SV=1 Two-component response regulator OS=Shewanella xiamenensis GN=SXM_3290 2 PE=4 SV=1 Uncharacterized protein OS=Shewanella xiamenensis GN=SXM_1599 PE=4 SV=1 3 Peptidyl-dipeptidase Dcp. Metallo peptidase. MEROPS family M03A OS=Shewanella 2 sp. (strain ANA-3) GN=Shewana3_1408 PE=3 SV=1 50S ribosomal protein L7/L12 OS=Shewanella sp. (strain ANA-3) GN=rplL PE=3 SV=1 2 General secretion pathway protein D OS=Shewanella sp. (strain ANA-3) 2 2 GN=Shewana3_0151 PE=3 SV=1 Formate dehydrogenase alpha subunit OS=Shewanella sp. (strain ANA-3) 2 2 GN=Shewana3_3920 PE=4 SV=1 30S ribosomal protein S13 OS=Shewanella oneidensis (strain MR-1) GN=rpsM PE=3 2 SV=1 Alkyl hydroperoxide reductase subunit C-like protein OS=Plesiomonas shigelloides 2 302-73 GN=PLESHI_04112 PE=4 SV=1 Phosphate acetyltransferase OS=Shewanella xiamenensis GN=SXM_2467 PE=3 SV=1 2 FAD linked oxidase domain protein OS=Shewanella sp. (strain ANA-3) 2 GN=Shewana3_2905 PE=4 SV=1 Fructose-1,6-bisphosphate aldolase OS=Shewanella xiamenensis GN=SXM_0814 2 PE=4 SV=1 4Fe-4S ferredoxin, iron-sulfur binding domain protein OS=Shewanella sp. (strain 2 ANA-3) GN=Shewana3_2907 PE=4 SV=1 Alanine--tRNA ligase OS=Shewanella xiamenensis GN=alaS PE=3 SV=1 4 Elongation factor Ts OS=Shewanella sp. (strain ANA-3) GN=tsf PE=3 SV=1 2

266 Uncharacterized protein OS=Shewanella sp. (strain ANA-3) GN=Shewana3_1231 3 PE=4 SV=1 Glycine--tRNA ligase beta subunit OS=Shewanella xiamenensis GN=glyS PE=3 SV=1 3 ATPase OS=Shewanella xiamenensis GN=SXM_2957 PE=3 SV=1 4 Elongation factor P OS=Shewanella xiamenensis GN=efp PE=3 SV=1 4 TonB-dependent receptor OS=Shewanella sp. (strain ANA-3) GN=Shewana3_3715 2 PE=3 SV=1 3-oxoacyl-(Acyl-carrier-protein) reductase OS=Shewanella pealeana (strain ATCC 2 700345 / ANG-SQ1) GN=Spea_2494 PE=3 SV=1 30S ribosomal protein S2 OS=Shewanella xiamenensis GN=rpsB PE=3 SV=1 2 Outer membrane protein OmpH OS=Shewanella xiamenensis GN=SXM_1438 PE=3 2 SV=1 3-oxoacyl-(Acyl-carrier-protein) synthase I OS=Shewanella benthica KT99 2 GN=KT99_09963 PE=3 SV=1 OmpA/MotB domain protein OS=Shewanella sp. (strain ANA-3) 2 GN=Shewana3_0665 PE=3 SV=1 Succinyl-CoA ligase [ADP-forming] subunit beta OS=Shewanella xiamenensis 2 GN=sucC PE=3 SV=1 ArgR-regulated TonB-dependent receptor OS=Shewanella oneidensis (strain MR-1) 2 GN=SO_2907 PE=3 SV=1 30S ribosomal protein S21 OS=Shewanella violacea (strain JCM 10179 / CIP 106290 / 2 LMG 19151 / DSS12) GN=rpsU PE=3 SV=1 Threonine--tRNA ligase OS=Shewanella xiamenensis GN=thrS PE=3 SV=1 2 TPR repeat-containing protein OS=Shewanella sp. (strain ANA-3) 2 GN=Shewana3_2621 PE=4 SV=1 D-3-phosphoglycerate dehydrogenase OS=Shewanella xiamenensis GN=SXM_0746 2 PE=3 SV=1 Uncharacterized protein OS=Shewanella xiamenensis GN=SXM_3118 PE=4 SV=1 3 50S ribosomal protein L16 OS=Shewanella xiamenensis GN=rplP PE=3 SV=1 2 Biopolymer transport protein ExbD/TolR OS=Shewanella xiamenensis 2 GN=SXM_1602 PE=3 SV=1 Histone family protein DNA-binding protein OS=Shewanella xiamenensis 3 GN=SXM_0517 PE=3 SV=1 Glu/Leu/Phe/Val dehydrogenase dimerisation region OS=Shewanella halifaxensis 2 (strain HAW-EB4) GN=Shal_1711 PE=3 SV=1 Aspartate--tRNA ligase OS=Shewanella sp. (strain ANA-3) GN=aspS PE=3 SV=1 2 Putative agmatine deiminase OS=Shewanella baltica (strain OS155 / ATCC BAA- 2 1091) GN=aguA PE=3 SV=1 Peptidase Do. Serine peptidase. MEROPS family S01B OS=Shewanella sp. (strain 2 ANA-3) GN=Shewana3_0690 PE=4 SV=1 Decaheme cytochrome c OS=Shewanella sp. (strain ANA-3) GN=Shewana3_2676 2 PE=4 SV=1 Blue (Type 1) copper domain protein OS=Shewanella sp. (strain ANA-3) 2 GN=Shewana3_3478 PE=4 SV=1 30S ribosomal protein S10 OS=Kangiella koreensis (strain DSM 16069 / KCTC 12182 2 / SW-125) GN=rpsJ PE=3 SV=1

267 References

Achour-Rokbani, A., Cordi, A., Poupin, P., Bauda, P., and Billard, P. (2010) Characterization of the ars gene cluster from extremely arsenic-resistant Microbacterium sp. strain A33. Appl. Environ. Microbiol. 76: 948–55. Achour, A.R., Bauda, P., and Billard, P. (2007) Diversity of arsenite transporter genes from arsenic-resistant soil bacteria. Res. Microbiol. 158: 128–37. Afkar, E., Lisak, J., Saltikov, C., Basu, P., Oremland, R.S., and Stolz, J.F. (2003) The respiratory arsenate reductase from Bacillus selenitireducens strain MLS10. FEMS Microbiol. Lett. 226: 107–112. Ahman (1994) Microbe grows by reducing arsenic. 750. Akay, C., Thomas, C., and Gazitt, Y. (2004) Arsenic trioxide and paclitaxel induce apoptosis by different mechanisms. Cell Cycle 3: 324–334. Altschul, S.F., Gish, W., Miller, W., Myers, E.W., and Lipman, D.J. (1990) Basic local alignment search tool. J. Mol. Biol. 215: 403–10. Amend, J.P., Saltikov, C., Lu, G.-S., and Hernandez, J. (2014) Microbial arsenic metabolism and reaction energetics. Rev. Mineral. Geochemistry 79: 391–433. Anderson, C.R., Condron, L.M., Clough, T.J., Fiers, M., Stewart, A., Hill, R.A., and Sherlock, R.R. (2011) Biochar induced soil microbial community change: Implications for biogeochemical cycling of carbon, nitrogen and phosphorus. Pedobiologia (Jena). 54: 309–320. Anderson, C.R. and Cook, G.M. (2004) Isolation and characterization of arsenate- reducing bacteria from arsenic-contaminated sites in New Zealand. Curr. Microbiol. 48: 341–347. Anderson, G.L., Williams, J., and Hille, R. (1992) The purification and characterization of arsenite oxidase from Alcaligenes faecalis, a molybdenum-containing hydroxylase. J. Biol. Chem. 267: 23674–82. Andrade, J.M., Gómez-Carracedo, M.P., Krzanowski, W., and Kubista, M. (2004) Procrustes rotation in analytical chemistry, a tutorial. Chemom. Intell. Lab. Syst. 72: 123–132. Andres, J. and Bertin, P.N. (2016) The microbial genomics of arsenic. FEMS Microbiol. Rev. fuv050. Andrews, S. (2010) FastQC: a quality control tool for high throughput sequence data. software. Argos, M., Kalra, T., Rathouz, P.J., Chen, Y., Pierce, B., Parvez, F., et al. (2010) Arsenic exposure from drinking water, and all-cause and chronic-disease mortalities in Bangladesh (HEALS): A prospective cohort study. Lancet 376: 252–258. Bahar, M.M., Megharaj, M., and Naidu, R. (2012) Arsenic bioremediation potential of a new arsenite-oxidizing bacterium Stenotrophomonas sp. MM- 7 isolated from soil. Biodegradation 23: 803–812.

268 Bankevich, A., Nurk, S., Antipov, D., Gurevich, A.A., Dvorkin, M., Kulikov, A.S., et al. (2012) SPAdes: A New Genome Assembly Algorithm and Its Applications to Single-Cell Sequencing. J. Comput. Biol. 19: 455–477. Barringer, J.L. and Reilly, P.A. (2013) Arsenic in groundwater: a summary of sources and the biogeochemical and hydrogeologic factors affecting arsenic occurrence and mobility. In, Current perspectives in contaminant hydrology and water resources sustainability. InTech. Bassas-Galia, M., Nogales, B., Arias, S., Rohde, M., Timmis, K.N., and Molinari, G. (2012) Plant original Massilia isolates producing polyhydroxybutyrate, including one exhibiting high yields from glycerol. J. Appl. Microbiol. 112: 443–54. Bateman, A. (2002) The Pfam Protein Families Database. Nucleic Acids Res. 30: 276–280. Bencko, V. (1977) Carcinogenic, teratogenic, and mutagenic effects of arsenic. Environ. Health Perspect. 19: 179–82. Bentley, R. and Chasteen, T.G. (2002) Microbial Methylation of Metalloids : Arsenic , Antimony , and Bismuth Microbial Methylation of Metalloids : Arsenic , Antimony , and Bismuth. 66:. Berg, M., Tran, H.C., Nguyen, T.C., Pham, H. V, Schertenleib, R., and Giger, W. (2001) Arsenic contamination of groundwater and drinking water in Vietnam: a human health threat. Environ. Sci. Technol. 35: 2621–6. Bergey, D.H. (2005) Bergey’s Manual of Systematic Bacteriology - Vol 2: The Proteobacteria Part A - Introductory Essays. Springer-Verlag New York Inc. 332. Bergey, D.H. (2010) Bergey’s Manual of Systematic Bacteriology - Vol 4: Bacteroidetes. Bergquist, E.R., Fischer, R.J., Sugden, K.D., and Martin, B.D. (2009) Inhibition by methylated organo-arsenicals of the respiratory 2-oxo-acid dehydrogenases. J. Organomet. Chem. 694: 973–980. Bertin, P.N., Heinrich-Salmeron, A., Pelletier, E., Goulhen-Chollet, F., Arsène- Ploetze, F., Gallien, S., et al. (2011) Metabolic diversity among main microorganisms inside an arsenic-rich ecosystem revealed by meta- and proteo-genomics. ISME J. 5: 1735–47. BGS/MML (1999) Arsenic contamination of ground water in Bangladesh: a review. Prepared for the Department of Public Health Engineering, Government of Bangladesh Dhaka. BGS, D. (1999) Groundwater studies for arsenic contamination in Bangladesh. Final report, rapid Investig. phase. Dep. Public Heal. Gov. Bangladesh, Mott MacDonald Br. Geol. Surv. Bhattacharya, P., Vahter, M., Jarsjö, J., Kumpiene, J., Ahmad, A., Sparrenbom, C., et al. (2016) Arsenic Research and Global Sustainability: Proceedings of the Sixth International Congress on Arsenic in the Environment (As2016), June

269 19-23, 2016, Stockholm, Sweden CRC Press. Black, A. and Craw, D. (2001) Arsenic, copper and zinc occurrence at the Wangaloa coal mine, southeast Otago, New Zealand. Int. J. Coal Geol. 45: 181–193. Bonato, P., Batista, M.B., Camilios-Neto, D., Pankievicz, V.C.S., Tadra-Sfeir, M.Z., Monteiro, R.A., et al. (2016) RNA-seq analyses reveal insights into the function of respiratory nitrate reductase of the diazotroph Herbaspirillum seropedicae. Environ. Microbiol. 18: 2677–2688. Bostick, B.C. and Fendorf, S. (2003) Arsenite sorption on troilite (FeS) and pyrite (FeS2). Geochim. Cosmochim. Acta 67: 909–921. El Bouhaddani, S., Houwing-Duistermaat, J., Salo, P., Perola, M., Jongbloed, G., and Uh, H.-W. (2016) Evaluation of O2PLS in Omics data integration. In, BMC bioinformatics. BioMed Central Ltd, p. 11. Bouvet, P.J.M. and Grimont, P. a. D. (1986) Taxonomy of the Genus Acinetobacter with the Recognition of Acinetobacter baumannii sp. nov., Acinetobacter haemolyticus sp. nov., Acinetobacter johnsonii sp. nov., and Acinetobacter junii sp. nov. and Emended Descriptions of Acinetobacter calcoaceticus a. Int. J. Syst. Bacteriol. 36: 228–240. Bowell, R., Alpers, C., Jamieson, H., Nordstrom, K., and Majzlan, J. (2014) Arsenic: Environmental geochemistry, mineralogy, and microbiology Walter de Gruyter GmbH & Co KG. Bowman, J.P., Nichols, D.S., and McMeekin, T. a. (1997) Psychrobacter glacincola sp. nov., a Halotolerant, Psychrophilic Bacterium Isolated from Antarctic Sea Ice. Syst. Appl. Microbiol. 20: 209–215. Branco, R., Francisco, R., Chung, A.P., and Morais, P.V. (2009) Identification of an aox system that requires cytochrome c in the highly arsenic-resistant bacterium Ochrobactrum tritici SCII24. Appl. Environ. Microbiol. 75: 5141– 5147. Breit, G.N., Whitney, J., Foster, A., Welch, A.H., Yount, J., and Sanzolone, R. (2001) Preliminary evaluation of arsenic cycling in the sediments of Bangladesh. In, U.S. Geol. Survey Workshop - Arsenic in the environment, Denver, CO, Feb. 21-22, 2001, wwwbrr.cr.usgs.gov/Arsenic/., pp. 21–22. Bučková, M., Puškarová, A., Chovanová, K., Kraková, L., Ferianc, P., and Pangallo, D. (2013) A simple strategy for investigating the diversity and hydrocarbon degradation abilities of cultivable bacteria from contaminated soil. World J. Microbiol. Biotechnol. 29: 1085–1098. Bundschuh, J., Litter, M.I., Parvez, F., Román-Ross, G., Nicolli, H.B., Jean, J.-S., et al. (2012) One century of arsenic exposure in Latin America: a review of history and occurrence from 14 countries. Sci. Total Environ. 429: 2–35. Buschmann, J. and Berg, M. (2009) Impact of sulfate reduction on the scale of arsenic contamination in groundwater of the Mekong, Bengal and Red River deltas. Appl. Geochemistry 24: 1278–1286.

270 Buse, H.Y., Lu, J., Lu, X., Mou, X., and Ashbolt, N.J. (2014) Microbial diversities (16S and 18S rRNA gene pyrosequencing) and environmental pathogens within drinking water biofilms grown on the common premise plumbing materials unplasticized polyvinylchloride and copper. FEMS Microbiol. Ecol. Butcher, B.G., Deane, S.M., and Rawlings, D.E. (2000) The chromosomal arsenic resistance genes of Thiobacillus ferrooxidans have an unusual arrangement and confer increased arsenic and antimony resistance to Escherichia coli. Appl. Environ. Microbiol. 66: 1826–1833. Cai, L., Liu, G., Rensing, C., and Wang, G. (2009a) Genes involved in arsenic transformation and resistance associated with different levels of arsenic- contaminated soils. BMC Microbiol. 9: 4. Cai, L., Liu, G., Rensing, C., and Wang, G. (2009b) Genes involved in arsenic transformation and resistance associated with different levels of arsenic- contaminated soils. BMC Microbiol. 9: 4. Campbell, K.M., Malasarn, D., Saltikov, C.W., Newman, D.K., and Hering, J.G. (2006) Simultaneous microbial reduction of iron (III) and arsenic (V) in suspensions of hydrous ferric oxide. Environ. Sci. Technol. 40: 5950–5955. Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E.K., et al. (2010a) QIIME allows analysis of high- throughput community sequencing data Intensity normalization improves color calling in SOLiD sequencing. Nat. Publ. Gr. 7: 335–336. Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E.K., et al. (2010b) QIIME allows analysis of high-throughput community sequencing data (Supplementary). Nat. Methods 7: 335. Caporaso, J.G., Lauber, C.L., Walters, W.A., Berg-lyons, D., Lozupone, C.A., Turnbaugh, P.J., et al. (2010) Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Carro, L., Rivas, R., León-Barrios, M., González-Tirante, M., Velázquez, E., and Valverde, A. (2012) Herbaspirillum canariense sp. nov., Herbaspirillum aurantiacum sp. nov. and Herbaspirillum soli sp. nov., isolated from volcanic mountain soil, and emended description of the genus Herbaspirillum. Int. J. Syst. Evol. Microbiol. 62: 1300–1306. Carvalhais, L.C. and Schenk, P.M. (2013) Sample processing and cDNA preparation for microbial metatranscriptomics in complex soil communities. Methods Enzym. 531: 251–267. Cervantes, C., Ji, G., Ramirez, J., and Silver, S. (1994) Resistance to arsenic compounds in microorganisms. FEMS Microbiol. Rev. 15: 355–367. Chang, J.-S., Kim, Y.-H., and Kim, K.-W. (2008) The ars genotype characterization of arsenic-resistant bacteria from arsenic-contaminated gold–silver mines in the Republic of Korea. Appl. Microbiol. Biotechnol. 80: 155–165. Charlet, L. and Polya, D. a. (2006) Arsenic in shallow, reducing groundwaters in Southern Asia: An environmental health disaster. Elements 2: 91–96.

271 Chaturvedi, P. and Shivaji, S. (2006) Exiguobacterium indicum sp. nov., a psychrophilic bacterium from the Hamta glacier of the Himalayan mountain ranges of India. Int. J. Syst. Evol. Microbiol. 56: 2765–2770. Chauhan, N.S., Ranjan, R., Purohit, H.J., Kalia, V.C., and Sharma, R. (2008) Identification of genes conferring arsenic resistance to Escherichia coli from an effluent treatment plant sludge metagenomic library. FEMS Microbiol. Ecol. 67: 130–139. Chawla, S., Mutenda, E.K., Dixon, H.B., Freeman, S., and Smith, a W. (1995) Synthesis of 3-arsonopyruvate and its interaction with phosphoenolpyruvate mutase. Biochem. J. 308 ( Pt 3: 931–5. Chen, J., Bhattacharjee, H., and Rosen, B.P. (2015) ArsH is an organoarsenical oxidase that confers resistance to trivalent forms of the herbicide monosodium methylarsenate and the poultry growth promoter roxarsone. Mol. Microbiol. 96: 1042–1052. Chen, J., Madegowda, M., Bhattacharjee, H., and Rosen, B.P. (2015) ArsP: a methylarsenite efflux permease. Mol. Microbiol. 98: 625–635. Chen, J., Yoshinaga, M., Garbinski, L.D., and Rosen, B.P. (2016) Synergistic interaction of glyceraldehydes‐3‐phosphate dehydrogenase and ArsJ, a novel organoarsenical efflux permease, confers arsenate resistance. Mol. Microbiol. 100: 945–953. Chen, L., Hu, M., Huang, L., Hua, Z., Kuang, J., Li, S., and Shu, W. (2015) Comparative metagenomic and metatranscriptomic analyses of microbial communities in acid mine drainage. ISME J. 9: 1579. Cheng, Z., Zheng, Y., Mortlock, R., and Van Geen, a (2004) Rapid multi-element analysis of groundwater by high-resolution inductively coupled plasma mass spectrometry. Anal. Bioanal. Chem. 379: 512–8. Chenna, R., Sugawara, H., Koike, T., Lopez, R., Gibson, T.J., Higgins, D.G., and Thompson, J.D. (2003) Multiple sequence alignment with the Clustal series of programs. Nucleic Acids Res. 31: 3497–3500. Cherry, J.A., Shaikh, A.U., Tallman, D.E., and Nicholson, R. V. (1979) Arsenic Species as an Indicator of Redox Conditions in Groundwater. Dev. Water Sci. 12: 373–392. Chistoserdova, L. (2010) Recent progress and new challenges in metagenomics for biotechnology. Biotechnol. Lett. 32: 1351–9. Chitpirom, K., Akaracharanya, A., Tanasupawat, S., Leepipatpiboon, N., and Kim, K.W. (2009) Isolation and characterization of arsenic resistant bacteria from tannery wastes and agricultural soils in Thailand. Ann. Microbiol. 59: 649– 656. Chrysostomou, C., Quandt, E.M., Marshall, N.M., Stone, E., and Georgiou, G. (2015) An Alternate Pathway of Arsenate Resistance in E. coli Mediated by the Glutathione S-Transferase GstB. ACS Chem. Biol. 10: 875–882. Coker, V.S., Gault, a G., Pearce, C.I., van der Laan, G., Telling, N.D., Charnock,

272 J.M., et al. (2006) XAS and XMCD evidence for species-dependent partitioning of arsenic during microbial reduction of ferrihydrite to magnetite. Environ. Sci. Technol. 40: 7745–50. Colbourn, P., Alloway, B.J., and Thornton, I. (1975) Arsenic and heavy metals in soils associated with regional geochemical anomalies in South-West England. Sci. Total Environ. 4: 359–363. Corsini, A., Cavalca, L., Crippa, L., Zaccheo, P., and Andreoni, V. (2010) Impact of glucose on microbial community of a soil containing pyrite cinders: role of bacteria in arsenic mobilization under submerged condition. Soil Biol. Biochem. 42: 699–707. Cotton, F.A. and Wilkinson, G. (1988) Advanced inorganic chemistry Wiley New York. Couture, R.M. and Van Cappellen, P. (2011) Reassessing the role of sulfur geochemistry on arsenic speciation in reducing environments. J. Hazard. Mater. 189: 647–652. Cozen, A.E., Weirauch, M.T., Pollard, K.S., Bernick, D.L., Stuart, J.M., and Lowe, T.M. (2009) Transcriptional map of respiratory versatility in the hyperthermophilic crenarchaeon Pyrobaculum aerophilum. J. Bacteriol. 191: 782–794. Cullen, W.R. and Reimer, K.J. (1989) Arsenic speciation in the environment. Chem. Rev. 89: 713–764. Cummings, D.E., March, A.W., Bostick, B., Spring, S., Jr, F.C., Fendorf, S., et al. (2000) Evidence for Microbial Fe ( III ) Reduction in Anoxic , Mining- Impacted Lake Sediments Evidence for Microbial Fe ( III ) Reduction in Anoxic , Mining- Impacted Lake Sediments ( Lake Coeur d ’ Alene , Idaho ). Appl. Environ. Microbiol. 66: 154–162. Cummings, D.E., Snoeyenbos-West, O.L., Newby, D.T., Niggemyer, A.M., Lovley, D.R., Achenbach, L.A., and Rosenzweig, R.F. (2003) Diversity of Geobacteraceae species inhabiting metal-polluted freshwater lake sediments ascertained by 16S rDNA analyses. Microb. Ecol. 46: 257–269. Cummings D.E., S.-W.O.L. (2002) Diversity of Geobacteraceae Species Inhabiting Metal-Polluted Freshwater Lake Sediments Ascertained by 16S rDNA Analysis. Microb. Ecol. 44: 95–106. Daims, H., Lebedeva, E.V., Pjevac, P., Han, P., Herbold, C., Albertsen, M., et al. (2015) Complete nitrification by Nitrospira bacteria. Nature 528: 504–509. Datta, S., Mailloux, B., Jung, H.-B., Hoque, M. a, Stute, M., Ahmed, K.M., and Zheng, Y. (2009) Redox trapping of arsenic during groundwater discharge in sediments from the Meghna riverbank in Bangladesh. Proc. Natl. Acad. Sci. U. S. A. 106: 16930–16935. Delavat, F., Lett, M.-C., and Lievremont, D. (2012) Novel and unexpected bacterial diversity in an arsenic-rich ecosystem revealed by culture- dependent approaches. Biol. Direct 7: 28.

273 Desoeuvre, A., Casiot, C., and Héry, M. (2016) Diversity and Distribution of Arsenic-Related Genes Along a Pollution Gradient in a River Affected by Acid Mine Drainage. Microb. Ecol. 71: 672–685. Dhar, R.K., Zheng, Y., Saltikov, C.W., Radloff, K.A., Mailloux, B.J., Ahmed, K.M., and Van Geen, A. (2011) Microbes enhance mobility of arsenic in pleistocene aquifer sand from Bangladesh. Environ. Sci. Technol. 45: 2648–2654. Dhar, R.K., Zheng, Y., Stute, M., van Geen, a., Cheng, Z., Shanewaz, M., et al. (2008) Temporal variability of groundwater chemistry in shallow and deep aquifers of Araihazar, Bangladesh. J. Contam. Hydrol. 99: 97–111. Dixit, S. and Hering, J.G. (2003) Comparison of arsenic(V) and arsenic(III) sorption onto iron oxide minerals: Implications for arsenic mobility. Environ. Sci. Technol. 37: 4182–4189. Dobbin, P.S., Warren, L.H., Cook, N.J., McEwan, A.G., Powell, A.K., and Richardson, D.J. (1996) Dissimilatory iron (III) reduction by Rhodobacter capsulatus. Microbiology 142: 765–774. Dombrowski, P.M., Long, W., Farley, K.J., Mahony, J.D., Capitani, J.F., and Di Toro, D.M. (2005) Thermodynamic analysis of arsenic methylation. Environ. Sci. Technol. 39: 2169–2176. Dong, G., Huang, Y., Yu, Q., Wang, Y., Wang, H., He, N., and Li, Q. (2014) Role of Nanoparticles in Controlling Arsenic Mobilization from Sediments near a Realgar Tailing. Environ. Sci. Technol. Dowdle, P.R., Laverman, A.M., and Oremland, R.S. (1996) Bacterial dissimilatory reduction of arsenic (V) to arsenic (III) in anoxic sediments. Appl. Environ. Microbiol. 62: 1664–1669. Drewniak, L. and Sklodowska, A. (2013) Arsenic-transforming microbes and their role in biomining processes. Environ. Sci. Pollut. Res. 20:. Drewniak, L., Stasiuk, R., Uhrynowski, W., and Sklodowska, A. (2015) Shewanella sp. O23S as a driving agent of a system utilizing dissimilatory arsenate- reducing bacteria responsible for self-cleaning of water contaminated with arsenic. Int. J. Mol. Sci. 16:. Duquesne, K., Lebrun, S., Casiot, C., Bruneel, O., Leblanc, M., Morin, G., and Personne, J. (2003) Immobilization of Arsenite and Ferric Iron by Acidithiobacillus ferrooxidans and Its Relevance to Acid Mine Drainage Immobilization of Arsenite and Ferric Iron by Acidithiobacillus ferrooxidans and Its Relevance to Acid Mine Drainage. Duval, S., Ducluzeau, A.-L., Nitschke, W., and Schoepp-Cothenet, B. (2008) Enzyme phylogenies as markers for the oxidation state of the environment: the case of respiratory arsenate reductase and related enzymes. BMC Evol. Biol. 8: 206. Edgar, R.C. (2004) MUSCLE: a multiple sequence alignment method with reduced time and space complexity. BMC Bioinformatics 5: 113. Edgar, R.C. (2004) MUSCLE: Multiple sequence alignment with high accuracy

274 and high throughput. Nucleic Acids Res. 32: 1792–1797. Edgar, R.C. (2010) Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26: 2460–2461. Edgar, R.C. (2013) UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat. Methods 10: 996–998. Edwards, M. (1994) Chemistry of arsenic removal during coagulation and Fe-Mn oxidation. J. Am. Water Work. Assoc. States) 86:. Edwardson, C.F. and Hollibaugh, J.T. (2017) Metatranscriptomic analysis of prokaryotic communities active in sulfur and arsenic cycling in Mono Lake, California, USA. ISME J. 1–14. Efron, B. (1981) Nonparametric estimates of standard error: the jackknife, the bootstrap and other methods. Biometrika 68: 589–599. Ehrich, S., Behrens, D., Lebedeva, E. V, Ludwig, W., and Bock, E. (1995) A New Obligately Chemolithoautotrophic, Nitrite-Oxidizing Bacterium, Nitrospira- Moscoviensis Sp-Nov and Its Phylogenetic Relationship. Arch. Microbiol. 164: 16–23. Emerson, D., Field, E.K., Chertkov, O., Davenport, K.W., Goodwin, L., Munk, C., et al. (2013) Comparative genomics of freshwater Fe-oxidizing bacteria: Implications for physiology, ecology, and systematics. Front. Microbiol. 4: 1– 17. Emerson, D., Fleming, E.J., and McBeth, J.M. (2010) Iron-oxidizing bacteria: an environmental and genomic perspective. Annu. Rev. Microbiol. 64: 561–583. Escudero, L. V, Casamayor, E.O., Chong, G., Pedrós-Alió, C., and Demergasso, C. (2013) Distribution of Microbial Arsenic Reduction, Oxidation and Extrusion Genes along a Wide Range of Environmental Arsenic Concentrations. PLoS One 8: e78890. Del Fabbro, C., Scalabrin, S., Morgante, M., and Giorgi, F.M. (2013) An extensive evaluation of read trimming effects on illumina NGS data analysis. PLoS One 8: 1–13. Ferguson, J.F. and Gavis, J. (1972) A review of the arsenic cycle in natural waters. Water Res. 6: 1259–1274. Fernández, M., Morel, B., Ramos, J.L., and Krell, T. (2016) Paralogous Regulators ArsR1 and ArsR2 of Pseudomonas putida KT2440 as a Basis for Arsenic Biosensor Development. Appl. Environ. Microbiol. 82: 4133–4144. Finn, R.D., Coggill, P., Eberhardt, R.Y., Eddy, S.R., Mistry, J., Mitchell, A.L., et al. (2016) The Pfam protein families database: Towards a more sustainable future. Nucleic Acids Res. 44: D279–D285. Fisher, E., Dawson, A.M., Polshyna, G., Lisak, J., Crable, B., Perera, E., et al. (2008) Transformation of inorganic and organic arsenic by Alkaliphilus oremlandii sp. nov. strain OhILAs. Ann. N. Y. Acad. Sci. 1125: 230–41. Flemming, H.-C. and Wingender, J. (2010) The biofilm matrix. Nat. Rev.

275 Microbiol. 8: 623. Focardi, S., Pepi, M., Ruta, M., Marvasi, M., Bernardini, E., Gasperini, S., and Focardi, S.E. (2010) Arsenic precipitation by an anaerobic arsenic-respiring bacterial strain isolated from the polluted sediments of Orbetello Lagoon, Italy. Lett. Appl. Microbiol. 51: 578–585. Francis, C.A., O’Mullan, G.D., and Ward, B.B. (2003) Diversity of ammonia monooxygenase (amoA) genes across environmental gradients in Chesapeake Bay sediments. Geobiology 1: 129–140. Frankenberger Jr, W.T. (2001) Environmental chemistry of arsenic CRC Press. Freikowski, D., Winter, J., and Gallert, C. (2010) Hydrogen formation by an arsenate-reducing Pseudomonas putida, isolated from arsenic-contaminated groundwater in West Bengal, India. Appl Microbiol Biotechnol 88: 1363–1371. Fresquet, M., Jowitt, T.A., Gummadova, J., Collins, R., O’Cualain, R., McKenzie, E.A., et al. (2015) Identification of a Major Epitope Recognized by PLA2R Autoantibodies in Primary Membranous Nephropathy. J. Am. Soc. Nephrol. 26: 302–313. Frias-Lopez, J., Shi, Y., Tyson, G.W., Coleman, M.L., Schuster, S.C., Chisholm, S.W., and Delong, E.F. (2008) Microbial community gene expression in ocean surface waters. Proc. Natl. Acad. Sci. U. S. A. 105: 3805–10. Fritz, I., Strömpl, C., Nikitin, D.I., Lysenko, A.M., and Abraham, W.R. (2005) Brevundimonas mediterranea sp. nov., a non-stalked species from the Mediterranean Sea. Int. J. Syst. Evol. Microbiol. 55: 479–486. Frohne, T., Rinklebe, J., Diaz-Bone, R.A., and Du Laing, G. (2011) Controlled variation of redox conditions in a floodplain soil: impact on metal mobilization and biomethylation of arsenic and antimony. Geoderma 160: 414–424. Gallego, V., Sánchez-Porro, C., García, M.T., and Ventosa, A. (2006) Massilia aurea sp. nov., isolated from drinking water. Int. J. Syst. Evol. Microbiol. 56: 2449–2453. Garelick, H., Jones, H., Dybowska, A., and Valsami-jones, E. (2008) Reviews of Environmental Contamination Volume 197 Springer New York, New York, NY. Garrity, G., Bell, J., and Lilburn, T. (2010) Family I. Pseudomonadaceae. Bergey´s Man. Syst. Bacteriol. vol 2B 323–379. Gault, A.G., Islam, F.S., Polya, D.A., Charnock, J.M., Boothman, C., Chatterjee, D., and Lloyd, J.R. (2005) Microcosm depth profiles of arsenic release in a shallow aquifer, West Bengal. Mineral. Mag. 69: 855–863. Ge, Y., Wang, D.-Z., Chiu, J.-F., Cristobal, S., Sheehan, D., Silvestre, F., et al. (2013) Environmental OMICS: Current Status and Future Directions. J. Integr. OMICS 3:. van Geen, a., Zheng, Y., Cheng, Z., Aziz, Z., Horneman, a., Dhar, R.K., et al.

276 (2006) A transect of groundwater and sediment properties in Araihazar, Bangladesh: Further evidence of decoupling between As and Fe mobilization. Chem. Geol. 228: 85–96. Georges, A. a, El-Swais, H., Craig, S.E., Li, W.K., and Walsh, D. a (2014) Metaproteomic analysis of a winter to spring succession in coastal northwest Atlantic Ocean microbial plankton. ISME J. 1–13. Ghosh, S. and Sar, P. (2013) Identification and characterization of metabolic properties of bacterial populations recovered from arsenic contaminated ground water of North East India (Assam). Water Res. 47: 6992–7005. Gihring, T.M., Druschel, G.K., McCleskey, R.B., Hamers, R.J., and Banfield, J.F. (2001) Rapid arsenite oxidation by Thermus aquaticus and Thermus thermophilus: field and laboratory investigations. Environ. Sci. Technol. 35: 3857–3862. Gilbert, J.A. (2010) Beyond the infinite -Tracking bacterial gene expression. Microbiol. Today 37: 82–85. Giloteaux, L., Holmes, D.E., Williams, K.H., Wrighton, K.C., Wilkins, M.J., Montgomery, A.P., et al. (2012) Characterization and transcription of arsenic respiration and resistance genes during in situ uranium bioremediation. ISME J. Giri, A.K., Patel, R.K., Mahapatra, S.S., and Mishra, P.C. (2013) Biosorption of arsenic (III) from aqueous solution by living cells of Bacillus cereus. Environ. Sci. Pollut. Res. 20: 1281–1291. Goldberg, S. and Johnston, C.T. (2001) Mechanisms of Arsenic Adsorption on Amorphous Oxides Evaluated Using Macroscopic Measurements, Vibrational Spectroscopy, and Surface Complexation Modeling. J. Colloid Interface Sci. 234: 204–216. Gremion, F., Chatzinotas, A., and Harms, H. (2003) Comparative 16S rDNA and 16S rRNA sequence analysis indicates that Actinobacteria might be a dominant part of the metabolically active bacteria in heavy metal‐ contaminated bulk and rhizosphere soil. Environ. Microbiol. 5: 896–907. Grondin, K., Haimeur, A., Mukhopadhyay, R., Rosen, B.P., and Ouellette, M. (1997) Co‐amplification of the γ‐glutamylcysteine synthetase gene gsh1 and of the ABC transporter gene pgpA in arsenite‐resistant Leishmania tarentolae. EMBO J. 16: 3057–3065. Guidelines for Drinking Water Quality: Recommendations, vol.1. (1993) Geneva. Guo, H., Liu, Z., Ding, S., Hao, C., Xiu, W., and Hou, W. (2015) Arsenate reduction and mobilization in the presence of indigenous aerobic bacteria obtained from high arsenic aquifers of the Hetao basin, Inner Mongolia. Environ. Pollut. 203: 50–59. Gussin, G.N., Ronson, C.W., and Ausubel, F.M. (1986) Regulation of nitrogen fixation genes. Annu. Rev. Genet. 20: 567–91. Guyon, I., Weston, J., Barnhill, S., and Vapnik, V. (2002) Gene selection for cancer

277 classification using support vector machines. Mach. Learn. 46: 389–422. Haft, D.H. (2001) TIGRFAMs: a protein family resource for the functional identification of proteins. Nucleic Acids Res. 29: 41–43. Hamady, M., Walker, J.J., Harris, J.K., Gold, N.J., and Knight, R. (2008) Error- correcting barcoded primers for pyrosequencing hundreds of samples in multiplex. Nat. Methods 5: 235–237. Hamamura, N., Fukushima, K., and Itai, T. (2013) Identification of antimony- and arsenic-oxidizing bacteria associated with antimony mine tailing. Microbes Env. 28: 257–263. Hamamura, N., Macur, R.E., Korf, S., Ackerman, G., Taylor, W.P., Kozubal, M., et al. (2009) Linking microbial oxidation of arsenic with detection and phylogenetic analysis of arsenite oxidase genes in diverse geothermal environments B : NGB-BE. 11: 421–431. Handley, K.M., Héry, M., and Lloyd, J.R. (2009) Redox cycling of arsenic by the hydrothermal marine bacterium Marinobacter santoriniensis. Environ. Microbiol. 11: 1601–1611. Handley, K.M., McBeth, J.M., Charnock, J.M., Vaughan, D.J., Wincott, P.L., Polya, D. a., and Lloyd, J.R. (2013) Effect of iron redox transformations on arsenic solid-phase associations in an arsenic-rich, ferruginous hydrothermal sediment. Geochim. Cosmochim. Acta 102: 124–142. Harold, F.M. and Baarda, J.R. (1966) Interaction of arsenate with phosphate- transport systems in wild- type and mutant Streptococcus faecalis. J. Bacteriol. 91: 2257–2262. Hassan, Z., Sultana, M., van Breukelen, B.M., Khan, S.I., and Roling, W.F.M. (2015) Diverse arsenic- and iron-cycling microbial communities in arsenic- contaminated aquifers used for drinking water in Bangladesh. FEMS Microbiol. Ecol. 91: fiv026-fiv026. Hassan, Z., Sultana, M., Breukelen, B.M. Van, Khan, S.I., and Röling, W.F.M. (2015) Diverse arsenic- and iron-cycling microbial communities in arsenic- contaminated aquifers used for drinking water in Bangladesh. 1–42. Heinrich-Salmeron, A., Cordi, A., Brochier-Armanet, C., Halter, D., Pagnout, C., Abbaszadeh-Fard, E., et al. (2011) Unsuspected diversity of arsenite-oxidizing bacteria revealed by a widespread distribution of the aoxB gene in prokaryotes. Appl. Environ. Microbiol. AEM-02884. Herbel, M. and Fendorf, S. (2006) Biogeochemical processes controlling the speciation and transport of arsenic within iron coated sands. Chem. Geol. 228: 16–32. Hert, D.G., Fredlake, C.P., and Barron, A.E. (2008) Advantages and limitations of next-generation sequencing technologies: a comparison of electrophoresis and non-electrophoresis methods. Electrophoresis 29: 4618–4626. Héry, M., Van Dongen, B.E., Gill, F., Mondal, D., Vaughan, D.J., Pancost, R.D., et al. (2010) Arsenic release and attenuation in low organic carbon aquifer

278 sediments from West Bengal. Geobiology 8: 155–68. Héry, M., Gault, A.G., Rowland, H. a. L., Lear, G., Polya, D. a., and Lloyd, J.R. (2008) Molecular and cultivation-dependent analysis of metal-reducing bacteria implicated in arsenic mobilisation in south-east asian aquifers. Appl. Geochemistry 23: 3215–3223. Héry, M., Rizoulis, A., Sanguin, H., Cooke, D. a, Pancost, R.D., Polya, D. a, and Lloyd, J.R. (2014) Microbial ecology of arsenic-mobilizing Cambodian sediments: lithological controls uncovered by stable-isotope probing. Environ. Microbiol. Hoeft, S.E., Blum, J.S., Stolz, J.F., Tabita, F.R., Witte, B., King, G.M., et al. (2007) Alkalilimnicola ehrlichii sp. nov., a novel, arsenite-oxidizing haloalkaliphilic gammaproteobacterium capable of chemoautotrophic or heterotrophic growth with nitrate or oxygen as the electron acceptor. Int. J. Syst. Evol. Microbiol. 57: 504–512. Hoeft, S.E., Kulp, T.R., Stolz, J.F., Hollibaugh JT, and Oremland, R.S. (2004) Dissimilatory arsenate\rreduction with sulfide as the electron donor: Experiments with Mono Lake water and \risolation of strain MLMS-1, a chemoautotrophic arsenate-respirer. Appl Env. Microbiol 70: 2741–2747. Hohmann, C., Morin, G., Ona-Nguema, G., Guigner, J.M., Brown, G.E., and Kappler, A. (2011) Molecular-level modes of As binding to Fe(III) (oxyhydr)oxides precipitated by the anaerobic nitrate-reducing Fe(II)- oxidizing Acidovorax sp. strain BoFeN1. Geochim. Cosmochim. Acta 75: 4699–4712. Hollibaugh, J.T., Budinoff, C., Hollibaugh, R. a., Ransom, B., and Bano, N. (2006) Sulfide oxidation coupled to arsenate reduction by a diverse microbial community in a soda lake. Appl. Environ. Microbiol. 72: 2043–2049. Horneman, a., van Geen, a., Kent, D. V., Mathe, P.E., Zheng, Y., Dhar, R.K., et al. (2004) Decoupling of As and Fe release to Bangladesh groundwater under reducing conditions. Part I: Evidence from sediment profiles. Geochim. Cosmochim. Acta 68: 3459–3473. Hosseini, P., Tremblay, A., Matthews, B.F., and Alkharouf, N.W. (2010) An efficient annotation and gene-expression derivation tool for Illumina Solexa datasets. BMC Res. Notes 3: 183. Hoven, R.N. and Santini, J.M. (2004) Arsenite oxidation by the heterotroph Hydrogenophaga sp . str . NT-14 : the arsenite oxidase and its physiological electron acceptor. 1656: 148–155. vanden Hoven, R.N. and Santini, J.M. (2004) Arsenite oxidation by the heterotroph Hydrogenophaga sp. str. NT-14: the arsenite oxidase and its physiological electron acceptor. Biochim. Biophys. Acta (BBA)-Bioenergetics 1656: 148–155. Hu, Z. and Gao, S. (2008) Upper crustal abundances of trace elements: A revision and update. Chem. Geol. 253: 205–221.

279 Huang, J.-H. (2014) Impact of Microorganisms on Arsenic Biogeochemistry: A Review. Water, Air, Soil Pollut. 225: 1848. Huang, J.-H. and Matzner, E. (2007) Biogeochemistry of organic and inorganic arsenic species in a forested catchment in Germany. Environ. Sci. Technol. 41: 1564–1569. Huang, J.H., Lazzaro, A., Zeyer, J., Kretzschmar, R., and Voegelin, A. (2010) Influence of arsenic surface adsorption and iron reduction on arsenic mobilization and reduction by Shewanella putrefaciens strain CN-32. In, Arsenic in Geosphere and Human Diseases; Arsenic 2010: Proceedings of the Third International Congress on Arsenic in the Environment (As-2010). CRC Press, p. 130. Huber, R., Sacher, M., Vollmann, A., Huber, H., and Rose, D. (2000) Respiration of arsenate and selenate by hyperthermophilic archaea. Syst. Appl. Microbiol. 23: 305–314. Hug, K., Maher, W.A., Stott, M.B., Krikowa, F., Foster, S., and Moreau, J.W. (2014) Microbial contributions to coupled arsenic and sulfur cycling in the acid- sulfide hot spring Champagne Pool, New Zealand. Front. Microbiol. 5:. Hug, L.A., Thomas, B.C., Sharon, I., Brown, C.T., Sharma, R., Hettich, R.L., et al. (2016) Critical biogeochemical functions in the subsurface are associated with bacteria from new phyla and little studied lineages. Environ. Microbiol. 18: 159–173. Hultman, J., Waldrop, M.P., Mackelprang, R., David, M.M., McFarland, J., Blazewicz, S.J., et al. (2015) Multi-omics of permafrost, active layer and thermokarst bog soil microbiomes. Nature. Hyatt, D., Chen, G.-L., LoCascio, P.F., Land, M.L., Larimer, F.W., and Hauser, L.J. (2010) Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinformatics 11: 119. Inskeep, W.P., Macur, R.E., Hamamura, N., Warelow, T.P., Ward, S. a, and Santini, J.M. (2007) Detection, diversity and expression of aerobic bacterial arsenite oxidase genes. Environ. Microbiol. 9: 934–43. Irlinger, F., Bimet, F., Delettre, J., Lefèvre, M., and Grimont, P. a D. (2005) Arthrobacter bergerei sp. nov. and Arthrobacter arilaitensis sp. nov., novel coryneform species isolated from the surfaces of cheeses. Int. J. Syst. Evol. Microbiol. 55: 457–62. Islam, F.S., Gault, A.G., Boothman, C., Polya, D. a, Charnock, J.M., Chatterjee, D., and Lloyd, J.R. (2004) Role of metal-reducing bacteria in arsenic release from Bengal delta sediments. Nature 430: 68–71. Islam, F.S., Pederick, R.L., Gault, A.G., Adams, L.K., Polya, D.A., Charnock, J.M., and Lloyd, J.R. (2005a) Interactions between the Fe ( III ) -Reducing Bacterium Geobacter sulfurreducens and Arsenate , and Capture of the Metalloid by Biogenic Fe ( II ). 71: 8642–8648. Islam, F.S., Pederick, R.L., Gault, A.G., Adams, L.K., Polya, D.A., Charnock, J.M.,

280 and Lloyd, J.R. (2005b) Interactions between the Fe (III)-reducing bacterium Geobacter sulfurreducens and arsenate, and capture of the metalloid by biogenic Fe (II). Appl. Environ. Microbiol. 71: 8642–8648. Jackson, C.R. and Dugas, S.L. (2003) Phylogenetic analysis of bacterial and archaeal arsC gene sequences suggests an ancient, common origin for arsenate reductase. BMC Evol. Biol. 3: 18. Ji, G., Garber, E. a, Armes, L.G., Chen, C.M., Fuchs, J. a, and Silver, S. (1994) Arsenate reductase of Staphylococcus aureus plasmid pI258. Biochemistry 33: 7294–9. Ji, G. and Silver, S. (1992) Reduction of arsenate to arsenite by the ArsC protein of the arsenic resistance operon of Staphylococcus aureus plasmid pI258. Proc. Natl. Acad. Sci. U. S. A. 89: 9474–8. Jin, H., Wen, H., Yang, H., Liu, Z., Ren, Z., Che, G., and Zhao, Z. (2003) Arsenic poisoning of Bangladesh groundwater. Sci. Technol. 3: 16–18. Jin, Y. and Manabe, T. (2005) High-efficiency protein extraction from polyacrylamide gels for molecular mass measurement by matrix-assisted laser desorption/ionization-time of flight-mass spectrometry. Electrophoresis 26: 1019–1028. Jomova, K., Jenisova, Z., Feszterova, M., Baros, S., Liska, J., Hudecova, D., et al. (2011) Arsenic: toxicity, oxidative stress and human disease. J. Appl. Toxicol. 31: 95–107. Joshi, N.A. and Fass, J.N. (2011) Sickle: A sliding-window, adaptive, quality-based trimming tool for FastQ files (Version 1.33)[Software]. Jung, H.B., Bostick, B.C., and Zheng, Y. (2012) Field, experimental, and modeling study of arsenic partitioning across a redox transition in a bangladesh aquifer. Environ. Sci. Technol. 46: 1388–1395. Jung, H.B. and Zheng, Y. (2006) Enhanced recovery of arsenite sorbed onto synthetic oxides by L-ascorbic acid addition to phosphate solution: calibrating a sequential leaching method for the speciation analysis of arsenic in natural samples. Water Res. 40: 2168–2180. Kageyama, A., Morisaki, K., Omura, S., and Takahashi, Y. (2008) Arthrobacter oryzae sp. nov. and Arthrobacter humicola sp. nov. Int. J. Syst. Evol. Microbiol. 58: 53–6. Kämpfer, P., Fallschissel, K., and Avendaño-Herrera, R. (2011) Chryseobacterium chaponense sp. nov., isolated from farmed Atlantic salmon (Salmo salar). Int. J. Syst. Evol. Microbiol. 61: 497–501. Kämpfer, P., Vaneechoutte, M., Lodders, N., De Baere, T., Avesani, V., Janssens, M., et al. (2009) Description of Chryseobacterium anthropi sp. nov. to accommodate clinical isolates biochemically similar to Kaistella koreensis and Chryseobacterium haifense, proposal to reclassify Kaistella koreensis as Chryseobacterium koreense comb. nov. and emended d. Int. J. Syst. Evol. Microbiol. 59: 2421–8.

281 Kanehisa, M. (2002) The KEGG databases at GenomeNet. Nucleic Acids Res. 30: 42–46. Kanehisa Laboratories (2017) KEGG: Kyoto Encyclopedia of Genes and Genomes. http://www.kegg.jp/ (accessed 2015). 27: 29–34. Kang, D.D., Froula, J., Egan, R., and Wang, Z. (2015) MetaBAT, an efficient tool for accurately reconstructing single genomes from complex microbial communities. PeerJ 3: e1165. Kapaj, S., Peterson, H., Liber, K., and Bhattacharya, P. (2006) Human health effects from chronic arsenic poisoning–a review. J. Environ. Sci. Heal. Part A 41: 2399–2428. Karim, M.M. (2000) Arsenic in groundwater and health problems in Bangladesh. Water Res. 34: 304–310. Kashyap, D.R., Botero, L.M., Franck, W.L., Hassett, D.J., and McDermott, T.R. (2006) Complex regulation of arsenite oxidation in Agrobacterium tumefaciens. J. Bacteriol. 188: 1081–1088. Kavitha, S., Selvakumar, R., Sathishkumar, M., Swaminathan, K., Lakshmanaperumalsamy, P., Singh, A., and Jain, S.K. (2009) Nitrate removal using Brevundimonas diminuta MTCC 8486 from ground water. Water Sci. Technol. 60: 517–524. Kearse, M., Moir, R., Wilson, A., Stones-Havas, S., Cheung, M., Sturrock, S., et al. (2012) Geneious Basic: An integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics 28: 1647–1649. Kennard, R.W. and Stone, L.A. (1969) Computer aided design of experiments. Technometrics 11: 137–148. Kim, E.H., Jeong, H.-J., Lee, Y.K., Moon, E.Y., Cho, J.-C., Lee, H.K., and Hong, S.G. (2011) Actimicrobium antarcticum gen. nov., sp. nov., of the family Oxalobacteraceae, isolated from Antarctic coastal seawater. Curr. Microbiol. 63: 213–7. Kim, K.K., Lee, K.C., Oh, H.-M., Kim, M.J., Eom, M.K., and Lee, J.-S. (2008) Arthrobacter defluvii sp. nov., 4-chlorophenol-degrading bacteria isolated from sewage. Int. J. Syst. Evol. Microbiol. 58: 1916–21. Kirk, M.F., Holm, T.R., Park, J., Jin, Q., Sanford, R.A., Fouke, B.W., and Bethke, C.M. (2004) Bacterial sulfate reduction limits natural arsenic contamination in groundwater. Geology 32: 953–956. Kocar, B.D., Herbel, M.J., Tufano, K.J., and Fendorf, S. (2006) Contrasting effects of dissimilatory iron (III) and arsenic (V) reduction on arsenic retention and transport. Environ. Sci. Technol. 40: 6715–21. Kocar, B.D., Polizzotto, M.L., Benner, S.G., Ying, S.C., Ung, M., Ouch, K., et al. (2008) Integrated biogeochemical and hydrologic processes driving arsenic release from shallow sediments to groundwaters of the Mekong delta. Appl. Geochemistry 23: 3059–3071.

282 Kozich, J.J., Westcott, S.L., Baxter, N.T., Highlander, S.K., and Schloss, P.D. (2013) Development of a dual-index sequencing strategy and curation pipeline for analyzing amplicon sequence data on the miseq illumina sequencing platform. Appl. Environ. Microbiol. 79: 5112–5120. Krafft, T. and Macy, J.M. (1998) Purification and characterization of the respiratory arsenate reductase of Chrysiogenes arsenatis. Eur. J. Biochem. 255: 647–53. Kräutler, B. (1990) Chemistry of methylcorrinoids related to their roles in bacterial C1 metabolism. FEMS Microbiol. Lett. 87: 349–354. Kruger, M.C., Bertin, P.N., Heipieper, H.J., and Arsène-Ploetze, F. (2013) Bacterial metabolism of environmental arsenic - Mechanisms and biotechnological applications. Appl. Microbiol. Biotechnol. 97:. Kuehnelt, D. and Goessler, W. (2003) Organoarsenic compounds in the terrestrial environment. Organomet. Compd. Environ. 223–275. Kulp, T.R., Hoeft, S.E., Miller, L.G., Saltikov, C., Murphy, J.N., Han, S., et al. (2006) Dissimilatory arsenate and sulfate reduction in sediments of two hypersaline, arsenic-rich soda lakes: Mono and Searles Lakes, California. Appl. Environ. Microbiol. 72: 6514–26. Kumar, S., Nei, M., Dudley, J., and Tamura, K. (2008) MEGA: A biologist-centric software for evolutionary analysis of DNA and protein sequences. Brief. Bioinform. 9: 299–306. Lane, D.J. (1991) 16S/23S rRNA sequencing, in Nucleic Acid Techniques in Bacterial Systematics. In, E. Stackebrandt,M.G. (ed). Wiley, New York, NY, pp. 115–1175. Langmead, B. and Salzberg, S.L. (2012) Fast gapped-read alignment with Bowtie 2. Nat. Methods 9: 357–9. Lara, J., Escudero González, L., Ferrero, M., Chong Díaz, G., Pedrós-Alió, C., and Demergasso, C. (2012) Enrichment of arsenic transforming and resistant heterotrophic bacteria from sediments of two salt lakes in Northern Chile. Extremophiles 16: 523–38. Lau, M.C.Y., Kieft, T.L., Kuloyo, O., Linage-Alvarez, B., van Heerden, E., Lindsay, M.R., et al. (2016) An oligotrophic deep-subsurface community dependent on syntrophy is dominated by sulfur-driven autotrophic denitrifiers. Proc. Natl. Acad. Sci. 113: E7927–E7936. Al Lawati, W.M., Rizoulis, A., Eiche, E., Boothman, C., Polya, D. a., Lloyd, J.R., et al. (2012) Characterisation of organic matter and microbial communities in contrasting arsenic-rich Holocene and arsenic-poor Pleistocene aquifers, Red River Delta, Vietnam. Appl. Geochemistry 27: 315–325. Lawson, M., Polya, D.A., Boyce, A.J., Bryant, C., and Ballentine, C.J. (2016) Tracing organic matter composition and distribution and its role on arsenic release in shallow Cambodian groundwaters. Geochim. Cosmochim. Acta 178: 160– 177.

283 Lawson, M., Polya, D.A., Boyce, A.J., Bryant, C., Mondal, D., Shantz, A., and Ballentine, C.J. (2013) Pond-derived organic carbon driving changes in arsenic hazard found in Asian groundwaters. Environ. Sci. Technol. 47: 7085– 7094. Lear, G., Song, B., Gault, a G., Polya, D. a, and Lloyd, J.R. (2007) Molecular analysis of arsenate-reducing bacteria within Cambodian sediments following amendment with acetate. Appl. Environ. Microbiol. 73: 1041–8. Lehr, C.R., Polishchuk, E., Delisle, M.-C., Franz, C., and Cullen, W.R. (2003) Arsenic methylation by micro-organisms isolated from sheepskin bedding materials. Hum. Exp. Toxicol. 22: 325–334. Lett, M.-C., Muller, D., Lièvremont, D., Silver, S., and Santini, J. (2012) Unified nomenclature for genes involved in prokaryotic aerobic arsenite oxidation. J. Bacteriol. 194: 207–8. Li, B. and Logan, B.E. (2004) Bacterial adhesion to glass and metal-oxide surfaces. Colloids Surf. B. Biointerfaces 36: 81–90. Li, H., Li, M., Huang, Y., Rensing, C., and Wang, G. (2013) In silico analysis of bacterial arsenic islands reveals remarkable synteny and functional relatedness between arsenate and phosphate. Front. Microbiol. 4: 347. Li, P., Wang, Y., Dai, X., Zhang, R., Jiang, Z., Jiang, D., et al. (2015) Microbial community in high arsenic shallow groundwater aquifers in Hetao Basin of Inner Mongolia, China. PLoS One 10:. Li, X. and Krumholz, L.R. (2007) Regulation of arsenate resistance in Desulfovibrio desulfuricans G20 by an arsRBCC operon and an arsC gene. J. Bacteriol. 189: 3705–11. Lin, Y.-F., Walmsley, A.R., and Rosen, B.P. (2006) An arsenic metallochaperone for an arsenic detoxification pump. Proc. Natl. Acad. Sci. 103: 15617–15622. Van Lis, R., Nitschke, W., Duval, S., and Schoepp-Cothenet, B. (2013) Arsenics as bioenergetic substrates. Biochim. Biophys. Acta - Bioenerg. 1827:. Liu, G., Liu, M., Kim, E., Maaty, W.S., Bothner, B., Lei, B., et al. (2012) A periplasmic arsenite‐binding protein involved in regulating arsenite oxidation. Environ. Microbiol. 14: 1624–1634. Lloyd, J. and Oremland, R. (2006) Microbial transformations of arsenic in the environment: from soda lakes to aquifers. Elements 2: 85–90. Lloyd, J.R. (2003) Microbial reduction of metals and radionuclides. FEMS Microbiol. Rev. 27: 411–425. Lloyd, J.R., Gault, A.G., Héry, M., and Macrae, J.D. (2011) Microbial transformations of Arsenic in the subsurface. 77–90. López, D.L., Bundschuh, J., Birkle, P., Armienta, M.A., Cumbal, L., Sracek, O., et al. (2012) Arsenic in volcanic geothermal fluids of Latin America. Sci. Total Environ. 429: 57–75. Lovley, D.R. (1991) Dissimilatory Fe(III) and Mn(IV) reduction. Microbiol. Rev. 55:

284 259–287. Lovley, D.R., Phillips, E.J.P., Lonergan, D.J., and Widma, P.K. (1995) Fe(III) and S0 reduction by Pelobacter carbinolicus. Appl. Environ. Microbiol. 61: 2132–2138. Loy, A., Küsel, K., Lehner, A., Drake, H.L., and Wagner, M. (2004) Microarray and functional gene analyses of sulfate-reducing prokaryotes in low-sulfate, acidic fens reveal cooccurrence of recognized genera and novel lineages. Appl. Environ. Microbiol. 70: 6998–7009. Macur, R.E., Jackson, C.R., Botero, L.M., McDermott, T.R., and Inskeep, W.P. (2004) Bacterial populations associated with the oxidation and reduction of arsenic in an unsaturated soil. Environ. Sci. Technol. 38: 104–11. Macy, J.M., Nunan, K., Iwigen, K.D., Dixon, D.R., Harbour, P.J., Cahill, M., and Sly, L.I. (1996) Bacterium Isolated from Gold Mine Wastewater. 1153–1157. Macy, J.M., Santini, J.M., Pauling, B. V, O’Neill, a H., and Sly, L.I. (2000) Two new arsenate/sulfate-reducing bacteria: mechanisms of arsenate reduction. Arch. Microbiol. 173: 49–57. Maeda, S., Ohki, A., Miyahara, K., Takeshita, T., and Higashi, S. (1990) Growth characteristics and arsenic metabolism of two species of arsenic‐tolerant bacteria. Appl. Organomet. Chem. 4: 245–250. Mailloux, B.J., Trembath-Reichert, E., Cheung, J., Watson, M., Stute, M., Freyer, G. a, et al. (2013) Advection of surface-derived organic carbon fuels microbial reduction in Bangladesh groundwater. Proc. Natl. Acad. Sci. U. S. A. 110: 5331–5. Malasarn, D., Keeffe, J.R., and Newman, D.K. (2008) Characterization of the arsenate respiratory reductase from Shewanella sp. strain ANA-3. J. Bacteriol. 190: 135–42. Malasarn, D., Saltikov, C.W., Campbell, K.M., Santini, J.M., Hering, J.G., and Newman, D.K. (2004) arrA Is a Reliable Marker for As (V) respiration. Science (80-. ). 4. Mallick, P. and Kuster, B. (2010) Proteomics: a pragmatic perspective. Nat. Biotechnol. 28: 695–709. Mardis, E.R. (2008) Next-generation DNA sequencing methods. Annu. Rev. Genomics Hum. Genet. 9: 387–402. Maron, P.-A., Ranjard, L., Mougel, C., and Lemanceau, P. (2007) Metaproteomics: a new approach for studying functional microbial ecology. Microb. Ecol. 53: 486–93. Marsili, E., Baron, D.B., Shikhare, I.D., Coursolle, D., Gralnick, J. a, and Bond, D.R. (2008) Shewanella secretes flavins that mediate extracellular electron transfer. Proc. Natl. Acad. Sci. U. S. A. 105: 3968–3973. Martin, M. (2009) Cutadaps. EmBnet.journal 47: 8–11. Masella, A.P., Bartram, A.K., Truszkowski, J.M., Brown, D.G., and Neufeld, J.D. (2012) PANDAseq: paired-end assembler for illumina sequences. BMC

285 Bioinformatics 13: 31. Masscheleyn, P.H., Delaune, R.D., and Patrick, Jr., W.H. (1991) Effect of Redox Potential and pH on Arsenic Speciation and Solubility in a Contaminated Soil. Environ. Sci. Technol. 25: 1414–1419. McArthur, J.M., Ravenscroft, P., Safiulla, S., and Thirlwall, M.F. (2001) Arsenic in groundwater: testing pollution mechanisms for sedimentary aquifers in Bangladesh. Water Resour. Res. 37: 109–117. McCarty, K.M., Chen, Y.-C., Quamruzzaman, Q., Rahman, M., Mahiuddin, G., Hsueh, Y.-M., et al. (2007) Arsenic methylation, GSTT1, GSTM1, GSTP1 polymorphisms, and skin lesions. Environ. Health Perspect. 115: 341–5. de Melo Oliveira, M.G., Abels, S., Zbinden, R., Bloemberg, G. V, and Zbinden, A. (2013) Accurate identification of fastidious Gram-negative rods: integration of both conventional phenotypic methods and 16S rRNA gene analysis. BMC Microbiol. 13: 162. Mestrot, A., Merle, J.K., Broglia, A., Feldmann, J., and Krupp, E.M. (2011) Atmospheric stability of arsine and methylarsines. Environ. Sci. Technol. 45: 4010–4015. Metzker, M.L. (2010) Sequencing technologies - the next generation. Nat. Rev. Genet. 11: 31–46. Meyer, B. and Kuevert, J. (2007) Phylogeny of the alpha and beta subunits of the dissimilatory adenosine-5??? -phosphosulfate (APS) reductase from sulfate- reducing prokaryotes - Origin and evolution of the dissimilatory sulfate- reduction pathway. Microbiology 153: 2026–2044. Meyer, J., Michalke, K., Kouril, T., and Hensel, R. (2008) Volatilisation of metals and metalloids: an inherent feature of methanoarchaea? Syst. Appl. Microbiol. 31: 81–87. Michalke, K., Wickenheiser, E.B., Mehring, M., Hirner, A. V, and Hensel, R. (2000) Production of volatile derivatives of metal (loid) s by microflora involved in anaerobic digestion of sewage sludge. Appl. Environ. Microbiol. 66: 2791–2796. Mihajlov, I., Stute, M., Schlosser, P., Mailloux, B.J., Zheng, Y., Choudhury, I., et al. (2016) Recharge of low‐arsenic aquifers tapped by community wells in Araihazar, Bangladesh, inferred from environmental isotopes. Water Resour. Res. 52: 3324–3349. Miller, C.S., Baker, B.J., Thomas, B.C., Singer, S.W., and Banfield, J.F. (2011) EMIRGE: reconstruction of full-length ribosomal genes from microbial community short read sequencing data. Genome Biol. 12: R44. Miller, W.H., Schipper, H.M., Lee, J.S., Singer, J., and Waxman, S. (2002) Mechanisms of Action of Arsenic Trioxide Mechanisms of Action of Arsenic Trioxide 1. 3893–3903. Mirza, B.S., Muruganandam, S., Meng, X., Sorensen, D.L., Dupont, R.R., and McLean, J.E. (2014) Arsenic(V) reduction in relation to Iron(III)

286 transformation and molecular characterization of the structural and functional microbial community in sediments of a basin-fill aquifer in Northern Utah. Appl. Environ. Microbiol. 80: 3198–208. Moriya, Y., Itoh, M., Okuda, S., and Kanehisa, M. (2005) KAAS: KEGG automatic annotation server. Genome Informatics 5: 2005. Moriya, Y., Itoh, M., Okuda, S., Yoshizawa, A.C., and Kanehisa, M. (2007) KAAS: An automatic genome annotation and pathway reconstruction server. Nucleic Acids Res. 35:. Mukai, H., Ambe, Y., Muku, T., Takeshita, K., and Fukuma, T. (1986) Seasonal variation of methylarsenic compounds in airborne participate matter. Nature 324: 239–241. Mukherjee, A., Sengupta, M.K., Hossain, M.A., Ahamed, S., Das, B., Nayak, B., et al. (2006) Arsenic contamination in groundwater: a global perspective with emphasis on the Asian scenario. J. Health. Popul. Nutr. 24: 142–63. Mukhopadhyay, R. and Rosen, B.P. (2002a) Arsenate reductases in prokaryotes and eukaryotes. Environ. Health Perspect. 110: 745. Mukhopadhyay, R. and Rosen, B.P. (2002b) Arsenate reductases in prokaryotes and eukaryotes. Environ. Health Perspect. 110 Suppl: 745–748. Mukhopadhyay, R., Rosen, B.P., Phung, L.T., and Silver, S. (2002) Microbial arsenic: from geocycles to genes and enzymes. FEMS Microbiol. Rev. 26: 311– 325. National Research Council (US) Committee on Medical and Biological Effects of Environmental Pollutants (1977) Arsenic: Medical and Biologic Effects of Environmental Pollutants. Newman, D.K. and Beveridge, T.J. (1997) Precipitation of Arsenic Trisul de by. Microbiology 63: 2022–2028. Newman, D.K., Beveridge, T.J., and Morel, F. (1997) Precipitation of arsenic trisulfide by Desulfotomaculum auripigmentum. Appl. Environ. Microbiol. 63: 2022–2028. Nickson, R.T., Mcarthur, J.M., Ravenscroft, P., Burgess, W.G., and Ahmed, K.M. (2000) Mechanism of arsenic release to groundwater, Bangladesh and West Bengal. Appl. Geochemistry 15: 403–413. Niggemyer, A., Spring, S., and Stackebrandt, E. (2001) Isolation and Characterization of a Novel As ( V ) -Reducing Bacterium : Implications for Arsenic Mobilization and the Genus Desulfitobacterium Isolation and Characterization of a Novel As ( V ) -Reducing Bacterium : Implications for Arsenic Mobilization. Appl. Environ. Microbiol. 67: 5568–5580. Nordstrom, D.K. (2002) Public health. Worldwide occurrences of arsenic in ground water. Science (80-. ). 296: 2143–5. Nriagu, J. (2001) Arsenic Poisoning Through the Ages. In, Environmental Chemistry of Arsenic., pp. 1–26.

287 Nyrén, P. (2007) The history of pyrosequencing. Methods Mol. Biol. 373: 1–14. O’Day, P. a, Vlassopoulos, D., Root, R., and Rivera, N. (2004) The influence of sulfur and iron on dissolved arsenic concentrations in the shallow subsurface under changing redox conditions. Proc. Natl. Acad. Sci. U. S. A. 101: 13703– 13708. Ogata, H., Goto, S., Sato, K., Fujibuchi, W., Bono, H., and Kanehisa, M. (1999) KEGG: Kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 27: 29– 34. Ogram, A., Sayler, S., and Barkay, T. (1987) The extraction and purification of microbial D N A from sediments. 7: 57–66. Omoregie, E.O., Couture, R.-M., Van Cappellen, P., Corkhill, C.L., Charnock, J.M., Polya, D. a, et al. (2013) Arsenic bioremediation by biogenic iron oxides and sulfides. Appl. Environ. Microbiol. 79: 4325–35. Oremland, R. and J.F.S. (2013) The Ecology of Arsenic Author ( s ): Ronald S . Oremland and John F . Stolz Source : Science , New Series , Vol . 300 , No . 5621 ( May 9 , 2003 ), pp . 939-944 Published by : American Association for the Advancement of Science Stable URL : http://www.jst. Science (80-. ). 300: 939–944. Oremland, R.S., Kulp, T.R., Blum, J.S., Hoeft, S.E., Baesman, S., Miller, L.G., and Stolz, J.F. (2005) A microbial arsenic cycle in a salt-saturated, extreme environment. Science 308: 1305–1308. Oremland, R.S. and Stolz, J.F. (2003) The ecology of arsenic. Science 300: 939–44. Osborne, T.H., Mcarthur, J.M., Sikdar, P.K., and Santini, J.M. (2015) Isolation of an Arsenate-Respiring Bacterium from a Redox Front in an Arsenic-Polluted Aquifer in West Bengal, Bengal Basin. Páez-Espino, D., Tamames, J., de Lorenzo, V., and Cánovas, D. (2009) Microbial responses to environmental arsenic. Biometals 22: 117–30. Paisio, C.E., Talano, M.A., González, P.S., Busto, V.D., Talou, J.R., and Agostini, E. (2012) Isolation and characterization of a Rhodococcus strain with phenol- degrading ability and its potential use for tannery effluent biotreatment. Environ. Sci. Pollut. Res. 19: 3430–3439. Panaro, N.J., Yuen, P.K., Sakazume, T., Fortina, P., Kricka, L.J., and Wilding, P. (2000) Evaluation of DNA fragment sizing and quantification by the Agilent 2100 bioanalyzer. Clin. Chem. 46: 1851–1853. Park, H.S., Kim, B.H., Kim, H.S., Kim, H.J., Kim, G., Kim, M., et al. (2001) A Novel Electrochemically Active and Fe ( III ) -reducing Bacterium Phylogenetically Related to Clostridium butyricum Isolated from a Microbial Fuel Cell. Scanning 7: 297–306. Paul, D., Kazy, S.K., Gupta, A.K., Pal, T., and Sar, P. (2015) Diversity, metabolic properties and arsenic mobilization potential of indigenous bacteria in arsenic contaminated groundwater of West Bengal, India. PLoS One 10:.

288 Pearcy, C. a., Chevis, D. a., Haug, T.J., Jeffries, H. a., Yang, N., Tang, J., et al. (2011) Evidence of microbially mediated arsenic mobilization from sediments of the Aquia aquifer, Maryland, USA. Appl. Geochemistry 26: 575–586. Pederick, R.L., Gault, a G., Charnock, J.M., Polya, D. a, and Lloyd, J.R. (2007a) Probing the biogeochemistry of arsenic: response of two contrasting aquifer sediments from Cambodia to stimulation by arsenate and ferric iron. J. Environ. Sci. Health. A. Tox. Hazard. Subst. Environ. Eng. 42: 1763–74. Pederick, R.L., Gault, a G., Charnock, J.M., Polya, D. a, and Lloyd, J.R. (2007b) Probing the biogeochemistry of arsenic: response of two contrasting aquifer sediments from Cambodia to stimulation by arsenate and ferric iron. J. Environ. Sci. Health. A. Tox. Hazard. Subst. Environ. Eng. 42: 1763–74. Peng, Y., Leung, H.C.M., Yiu, S.M., and Chin, F.Y.L. (2012) IDBA-UD: A de novo assembler for single-cell and metagenomic sequencing data with highly uneven depth. Bioinformatics 28: 1420–1428. Pérez-Jiménez, J.R., DeFraia, C., and Young, L.Y. (2005) Arsenate respiratory reductase gene (arrA) for Desulfosporosinus sp. strain Y5. Biochem. Biophys. Res. Commun. 338: 825–829. Pichler, T., Veizer, J., and Hall, G.E.M. (1999) The natural input of extremely high arsenic concentrations into a coral reef ecosystem by hydrothermal fluids and its removal by Fe(III) oxyhydroxides. Environ. Sci. Technol. 33: 1373–1378. Piotr, B., Robert, W., Grzegorz, O., Andreu, G., and Stanislaw, U. (1997) Isolation of three contiguous genes, ACR1, ACR2 and ACR3, involved in resistance to arsenic compounds in the yeast Saccharomyces cerevisiae. Yeast 13: 819–828. Polizzotto, M.L., Kocar, B.D., Benner, S.G., Sampson, M., and Fendorf, S. (2008) Near-surface wetland sediments as a source of arsenic release to ground water in Asia. Nature 454: 505. Polya, D.A., Gault, A.G., Diebe, N., Feldman, P., Rosenboom, J.W., Gilligan, E., et al. (2005) Arsenic hazard in shallow Cambodian groundwaters. Mineral. Mag. 69: 807–823. Polya, D.A. and Middleton, D.R.S. (2017) Arsenic in drinking water: sources & human exposure. Best Pract. Guid. Control Arsen. Drink. Water 1. Pourmand, N., Elahi, E., Davis, R.W., and Ronaghi, M. (2002) Multiplex Pyrosequencing. Nucleic Acids Res. 30: e31. Prasad, K.S., Srivastava, P., Subramanian, V., and Paul, J. (2011) Biosorption of As (III) ion on Rhodococcus sp. WB-12: biomass characterization and kinetic studies. Sep. Sci. Technol. 46: 2517–2525. Prieto, D.M., Devesa-Rey, R., Rubinos, D. a, Díaz-Fierros, F., and Barral, M.T. (2013) Arsenate retention by epipsammic biofilms developed on streambed sediments: influence of phosphate. Biomed Res. Int. 2013: 591634. Qin, J., Rosen, B.P., Zhang, Y., Wang, G., Franke, S., and Rensing, C. (2006a) Arsenic detoxification and evolution of trimethylarsine gas by a microbial arsenite S-adenosylmethionine methyltransferase. 103:.

289 Qin, J., Rosen, B.P., Zhang, Y., Wang, G., Franke, S., and Rensing, C. (2006b) Arsenic detoxification and evolution of trimethylarsine gas by a microbial arsenite S-adenosylmethionine methyltransferase. Proc. Natl. Acad. Sci. U. S. A. 103: 2075–2080. Rahman, M.A. and Hassler, C. (2014) Is arsenic biotransformation a detoxification mechanism for microorganisms? Aquat. Toxicol. 146:. Rambaut, A. (2009) FigTree, version 1.3. 1. Comput. Progr. Distrib. by author, website http//treebioedacuk/software/figtree/[accessed January 4, 2011]. Ravenscroft, P., Brammer, H., and Richards, K. (2009) Arsenic Pollution: A Global Synthesis, RGS-IBG Book Series, A John Wiley & Sons Ltd. Publ. United Kingdom. Reddy, G.S.N., Matsumoto, G.I., Schumann, P., Stackerbrandt, E., and Shivaji, S. (2004) Psychrophilic pseudomonads from Antarctica: Pseudomonas antartica sp. nov., Pseudomonas meridiana sp. nov. and Pseudomonas proteolytica sp. nov. Int. J. Syst. Evol. Microbiol. 54: 713–719. Rhine, E.D., Ní Chadhain, S.M., Zylstra, G.J., and Young, L.Y. (2007) The arsenite oxidase genes (aroAB) in novel chemoautotrophic arsenite oxidizers. Biochem. Biophys. Res. Commun. 354: 662–667. Richards, L.A., Magnone, D., van Dongen, B.E., Ballentine, C.J., and Polya, D.A. (2015) Use of lithium tracers to quantify drilling fluid contamination for groundwater monitoring in Southeast Asia. Appl. Geochemistry 63: 190–202. Richards, L.A., Magnone, D., Sovann, C., Kong, C., Uhlemann, S., Kuras, O., et al. (2017) Science of the Total Environment High resolution pro fi le of inorganic aqueous geochemistry and key redox zones in an arsenic bearing aquifer in Cambodia. Sci. Total Environ. 590–591: 540–553. Richards, L.A., Sültenfuß, J., Ballentine, C.J., Magnone, D., van Dongen, B.E., Sovann, C., and Polya, D.A. (2017) Tritium Tracers of Rapid Surface Water Ingression into Arsenic-bearing Aquifers in the Lower Mekong Basin, Cambodia. Procedia Earth Planet. Sci. 17: 845–848. Richey, C., Chovanec, P., Hoeft, S.E., Oremland, R.S., Basu, P., and Stolz, J.F. (2009) Biochemical and Biophysical Research Communications Respiratory arsenate reductase as a bidirectional enzyme. Biochem. Biophys. Res. Commun. 382: 298–302. Riesenfeld, C.S., Schloss, P.D., and Handelsman, J. (2004) Metagenomics: genomic analysis of microbial communities. Annu. Rev. Genet. 38: 525–52. Rittle, K.A., Drever, J.I., and Colberg, P.J.S. (1995) Precipitation of Arsenic during Bacterial Sulfate Reduction. Geomicrobiol. J. 13: 1–11. Rizoulis, A., Lawati, A.W.M. Al, Pancost, B.R.D., Polya, D.A., A, A.B.E.V.D., and A, J.R.L. (2014) Microbially mediated reduction of Fe III and As V in Cambodian sediments amended with 13 C-labelled hexadecane and kerogen. 538–546. Rizoulis, A., Al Lawati, W.M., Pancost, R.D., Polya, D.A., van Dongen, B.E., and Lloyd, J.R. (2014) Microbially mediated reduction of Fe III and As V in

290 Cambodian sediments amended with 13 C-labelled hexadecane and kerogen. Environ. Chem. 11: 538–546. Rizoulis, A., Al Lawati, W.M., Pancost, R.D., Polya, D. a., van Dongen, B.E., and Lloyd, J.R. (2014) Microbially mediated reduction of FeIII and AsV in Cambodian sediments amended with 13C-labelled hexadecane and kerogen. Environ. Chem. 11: 538. Rodrigues, D.F., Goris, J., Vishnivetskaya, T., Gilichinsky, D., Thomashow, M.F., and Tiedje, J.M. (2006) Characterization of Exiguobacterium isolates from the Siberian permafrost. Description of Exiguobacterium sibiricum sp. nov. Extremophiles 10: 285–294. Roesch, L.F.W., Fulthorpe, R.R., Riva, A., Casella, G., Hadwin, A.K.M., Kent, A.D., et al. (2007) Pyrosequencing enumerates and contrasts soil microbial diversity. ISME J. 1: 283–90. Rognes, T., Flouri, T., Nichols, B., Quince, C., and Mahé, F. (2016) VSEARCH: a versatile open source tool for metagenomics. PeerJ 4: e2584. Ronaghi, M. (2001) Pyrosequencing Sheds Light on DNA Sequencing Pyrosequencing Sheds Light on DNA Sequencing. 3–11. Rosen, B.P. (1999) Families of arsenic transporters. Trends Microbiol. 7: 207–12. Rosenberg, H., Gerdes, R.G., and Chegwidden, K. (1977) Two systems for the uptake of phosphate in Escherichia coli. J. Bacteriol. 131: 505–511. Rowland, H. a. L., Pederick, R.L., Polya, D. a., Pancost, R.D., Van Dongen, B.E., Gault, a. G., et al. (2007) The control of organic matter on microbially mediated iron reduction and arsenic release in shallow alluvial aquifers, Cambodia. Geobiology 5: 281–292. Rowland, H. a. L., Polya, D. a., Lloyd, J.R., and Pancost, R.D. (2006) Characterisation of organic matter in a shallow, reducing, arsenic-rich aquifer, West Bengal. Org. Geochem. 37: 1101–1114. Rowland, H. a L., Boothman, C., Pancost, R., Gault, a G., Polya, D. a, and Lloyd, J.R. (2002) The role of indigenous microorganisms in the biodegradation of naturally occurring petroleum, the reduction of iron, and the mobilization of arsenite from west bengal aquifer sediments. J. Environ. Qual. 38: 1598–607. Rowland, H. a L., Gault, A.G., Lythgoe, P., and Polya, D. a. (2008) Geochemistry of aquifer sediments and arsenic-rich groundwaters from Kandal Province, Cambodia. Appl. Geochemistry 23: 3029–3046. Ryan, D. and Colleran, E. (2002) Arsenical resistance in the IncHI2 plasmids. Plasmid 47: 234–240. Saltikov, C.W., Cifuentes, A., Venkateswaran, K., and Newman, D.K. (2003) The ars Detoxification System Is Advantageous but Not Required for As ( V ) Respiration by the Genetically Tractable Shewanella Species Strain ANA-3 The ars Detoxification System Is Advantageous but Not Required for As ( V ) Respiration by the Genetically.

291 Saltikov, C.W. and Newman, D.K. (2003) Genetic identification of a respiratory arsenate reductase. Proc. Natl. Acad. Sci. U. S. A. 100: 10983–8. Saltikov, C.W., Wildman, R.A., and Newman, D.K. (2005) Expression Dynamics of Arsenic Respiration and Detoxification in. 187: 7390–7396. Santini, J.M., Kappler, U., Ward, S.A., Honeychurch, M.J., vanden Hoven, R.N., and Bernhardt, P. V (2007) The NT-26 cytochrome c 552 and its role in arsenite oxidation. Biochim. Biophys. Acta (BBA)-Bioenergetics 1767: 189–196. Sanyal, S.K., Mou, T.J., Chakrabarty, R.P., Hoque, S., Hossain, M.A., and Sultana, M. (2016) Diversity of arsenite oxidase gene and arsenotrophic bacteria in arsenic affected Bangladesh soils. AMB Express 6:. Sardiwal, S., Santini, J.M., Osborne, T.H., and Djordjevic, S. (2010) Characterization of a two-component signal transduction system that controls arsenite oxidation in the chemolithoautotroph NT-26. FEMS Microbiol. Lett. 313: 20–28. Sarkar, A., Kazy, S.K., and Sar, P. (2013) Characterization of arsenic resistant bacteria from arsenic rich groundwater of West Bengal, India. Ecotoxicology 22:. Sato, T. and Kobayashi, Y. (1998) The ars Operon in the skin Element of Bacillus subtilis Confers Resistance to Arsenate and Arsenite The ars Operon in the skin Element of Bacillus subtilis Confers Resistance to Arsenate and Arsenite. 180:. Scheff, G., Salcher, O., and Lingens, F. (1984) Trichococcus flocculiformis gen. nov. sp. nov. A new gram-positive fdamentous bacterium isolated from bulking sludge. Appl. Microbiol. Biotechnol. 19: 114–119. Schirmer, M., Ijaz, U.Z., D’Amore, R., Hall, N., Sloan, W.T., and Quince, C. (2015) Insight into biases and sequencing errors for amplicon sequencing with the Illumina MiSeq platform. Nucleic Acids Res. 43:. Schloss, P.D. and Handelsman, J. (2005) Metagenomics for studying unculturable microorganisms: cutting the Gordian knot. Genome Biol. 6: 229. Searle, B.C. (2010) Scaffold: A bioinformatic tool for validating MS/MS‐based proteomic studies. Proteomics 6: 1265–1269. Segata, N., Boernigen, D., Tickle, T.L., Morgan, X.C., Garrett, W.S., and Huttenhower, C. (2013) Computational meta’omics for microbial community studies. Mol. Syst. Biol. 9: 666. Senn, D.B. and Hemond, H.F. (2002) Nitrate controls on iron and arsenic in an urban lake. Science 296: 2373–6. Shapiro, a L., Viñuela, E., and Maizel, J. V (1967) Molecular weight estimation of polypeptide chains by electrophoresis in SDS-polyacrylamide gels. Biochem. Biophys. Res. Commun. 28: 815–820. Shariatpahani, M., Anderson, A.C., Abdelghani, A.A., and Englande, A.J. (1983) Microbial metabolism of an organic arsenical herbicide. In, Biodeterioration

292 5: papers presented at the 5th International Biodeterioration Symposium, Aberdeen, September, 1981/edited by TA Oxley and S. Barry. Chichester [Sussex]: Wiley, c1983. Sharma, A.K., Tjell, J.C., Sloth, J.J., and Holm, P.E. (2014) Review of arsenic contamination, exposure through water and food and low cost mitigation options for rural areas. Appl. Geochemistry 41: 11–33. Shelobolina, E.S., Vrionis, H. a., Findlay, R.H., and Lovley, D.R. (2008) Geobacter uraniireducens sp. nov., isolated from subsurface sediment undergoing uranium bioremediation. Int. J. Syst. Evol. Microbiol. 58: 1075–1078. Shendure, J. and Lieberman Aiden, E. (2012) The expanding scope of DNA sequencing. Nat. Biotechnol. 30: 1084–94. Shrestha, P.M., Noll, M., and Liesack, W. (2007) Phylogenetic identity, growth- response time and rRNA operon copy number of soil bacteria indicate different stages of community succession. Environ. Microbiol. 9: 2464–2474. Silver, S. and Phung, L.T. (2005) Genes and Enzymes Involved in Bacterial Oxidation and Reduction of Inorganic Arsenic MINIREVIEW Genes and Enzymes Involved in Bacterial Oxidation and Reduction of Inorganic Arsenic. 71:. Slyemi, D. and Bonnefoy, V. (2012) How prokaryotes deal with arsenic(†). Environ. Microbiol. Rep. 4: 571–86. Smedley, P.. and Kinniburgh, D.. (2002) A review of the source, behaviour and distribution of arsenic in natural waters. Appl. Geochemistry 17: 517–568. Smith, E.R.G., Naidu, R., and Alston, A.M. (1998) Arsenic in the soil environment. Song, B., Chyun, E., Jaffé, P.R., and Ward, B.B. (2009) Molecular methods to detect and monitor dissimilatory arsenate-respiring bacteria (DARB) in sediments. FEMS Microbiol. Ecol. 68: 108–17. Song, Y., Deng, S.P., Acosta-Martínez, V., and Katsalirou, E. (2008) Characterization of redox-related soil microbial communities along a river floodplain continuum by fatty acid methyl ester (FAME) and 16S rRNA genes. Appl. soil Ecol. 40: 499–509. Sorg, T.J., Chen, A.S.C., and Wang, L. (2014) Arsenic species in drinking water wells in the USA with high arsenic concentrations. Water Res. 48: 156–169. Sovann, C. and Polya, D.A. (2014) Improved groundwater geogenic arsenic hazard map for Cambodia. Environ. Chem. 11: 595–607. Stevens, H., Brinkhoff, T., Rink, B., Vollmers, J., and Simon, M. (2007) Diversity and abundance of Gram positive bacteria in a tidal flat ecosystem. Environ. Microbiol. 9: 1810–1822. Stirling, D. and Bartlett, J.M.S. (2003) PCR Protocols Humana Press. Stolz, J.F. and Oremland, R.S. (1999) Bacterial respiration of arsenic and selenium. FEMS Microbiol. Rev. 23: 615–627. Stupperich, E. (1993) Recent advances in elucidation of biological corrinoid

293 functions. FEMS Microbiol. Rev. 12: 349–365. Stute, M., Zheng, Y., Schlosser, P., Horneman, A., Dhar, R.K., Datta, S., et al. (2007) Hydrological control of As concentrations in Bangladesh groundwater. Water Resour. Res. 43: 1–11. Sun, J., Bostick, B.C., Mailloux, B.J., Ross, J.M., and Chillrud, S.N. (2016) Effect of oxalic acid treatment on sediment arsenic concentrations and lability under reducing conditions. J. Hazard. Mater. 311: 125–133. Sun, J., Chillrud, S.N., Mailloux, B.J., Stute, M., Singh, R., Dong, H., et al. (2016) Enhanced and stabilized arsenic retention in microcosms through the microbial oxidation of ferrous iron by nitrate. Chemosphere 144: 1106–1115. Sun, J., Quicksall, A.N., Chillrud, S.N., Mailloux, B.J., and Bostick, B.C. (2016) Arsenic mobilization from sediments in microcosms under sulfate reduction. Chemosphere 153:. Sung, Y., Sung, Y., Fletcher, K.E., Fletcher, K.E., Ritalahti, K.M., Ritalahti, K.M., et al. (2006) Geobacter lovleyi. Society 72: 2775–2782. Switzer Blum, J., Burns Bindi, A., Buzzelli, J., Stolz, J., and Oremland, R. (1998) Bacillus arsenicoselenatis, sp. nov., and Bacillus selenitireducens, sp. nov.: two haloalkaliphiles from Mono Lake, California that respire oxyanions of selenium and arsenic. Arch. Microbiol. 171: 19–30. Tamaki, S. and Frankenberger Jr, W.T. (1992) Environmental biochemistry of arsenic. In, Reviews of environmental contamination and toxicology. Springer, pp. 79–110. Tisa, L.S. and Rosen, B.P. (1990) Molecular characterization of an anion pump. The ArsB protein is the membrane anchor for the ArsA protein. J. Biol. Chem. 265: 190–194. Trapnell, C., Roberts, A., Goff, L., Pertea, G., Kim, D., Kelley, D.R., et al. (2012) Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. Nat. Protoc. 7: 562–578. Trygg, J. and Wold, S. (2003) O2-PLS, a two-block (X-Y) latent variable regression (LVR) method with an integral OSC filter. J. Chemom. 17: 53–64. Trygg, J. and Wold, S. (2002) Orthogonal projections to latent structures (O-PLS). J. Chemom. 16: 119–128. Tsai, S.-L., Singh, S., and Chen, W. (2009) Arsenic metabolism by microbes in nature and the impact on arsenic remediation. Curr. Opin. Biotechnol. 20: 659–67. Turpeinen, R., Pantsar-Kallio, M., and Kairesalo, T. (2002) Role of microbes in controlling the speciation of arsenic and production of arsines in contaminated soils. Sci. Total Environ. 285: 133–145. Urich, T., Lanzén, A., Qi, J., Huson, D.H., Schleper, C., and Schuster, S.C. (2008) Simultaneous assessment of soil microbial community structure and function through analysis of the meta-transcriptome. PLoS One 3:.

294 Vesterberg, O. (1993) A short history of electrophoretic methods. Electrophoresis 14: 1243–1249. Villegas-torres, M.F., Bedoya-reina, O.C., Salazar, C., Vives-florez, M.J., and Dussan, J. (2011) International Biodeterioration & Biodegradation Horizontal arsC gene transfer among microorganisms isolated from arsenic polluted soil. Int. Biodeterior. Biodegradation 65: 147–152. Vithanage, M. (2017) Arsenic Research and global Sustainability—As 2016 Bhattacharya et al.,(Eds) ISBN 978-1-138-02941-5. Vogelstein, B. and Gillespie, D. (1979) Preparative and analytical purification of DNA from agarose. Proc. Natl. Acad. Sci. U. S. A. 76: 615–619. Wade, W. (2002) Unculturable bacteria—the uncharacterized organisms that cause oral infections. J. R. Soc. Med. 95: 81. Wagner, M., Roger, A.J., Flax, J.L., Gregory, A., Stahl, D.A., and Brusseau, G.A. (1998) Phylogeny of Dissimilatory Sulfite Reductases Supports an Early Origin of Sulfate Respiration Phylogeny of Dissimilatory Sulfite Reductases Supports an Early Origin of Sulfate Respiration. J. Bacteriol. 180:. Wall, P.K., Leebens-Mack, J., Chanderbali, A.S., Barakat, A., Wolcott, E., Liang, H., et al. (2009) Comparison of next generation sequencing technologies for transcriptome characterization. BMC Genomics 10: 347. Wang, P., Sun, G., Jia, Y., Meharg, A. a, and Zhu, Y. (2014) A review on completing arsenic biogeochemical cycle: Microbial volatilization of arsines in environment. J. Environ. Sci. 26: 371–381. Wang, X.H., Zhang, C., Yang, L.L., Li, S., Zhang, Y., and Gomes-Laranjo, J. (2011) Screening and identification of microbial strains that secrete an extracellular C-7 xylosidase of taxanes. World J. Microbiol. Biotechnol. 27: 627–635. Warnecke, F. and Hess, M. (2009) A perspective: metatranscriptomics as a tool for the discovery of novel biocatalysts. J. Biotechnol. 142: 91–5. Watanabe, T., Miura, A., Iwata, T., Kojima, H., and Fukui, M. (2017) Dominance of Sulfuritalea species in nitrate‐depleted water of a stratified freshwater lake and arsenate respiration ability within the genus. Environ. Microbiol. Rep. Watts, M.J., O’Reilly, J., Marcilla, A.L., Shaw, R.A., and Ward, N.I. (2010) Field based speciation of arsenic in UK and Argentinean water samples. Environ. Geochem. Health 32: 479–490. Wauters, G., De Baere, T., Willems, A., Falsen, E., and Vaneechoutte, M. (2003) Description of Comamonas aquatica comb. nov. and Comamonas kerstersii sp. nov. for two subgroups of Comamonas terrigena and emended description of Comamonas terrigena. Int. J. Syst. Evol. Microbiol. 53: 859–862. Webster, J.G. and Nordstrom, D.K. (2003) Geothermal arsenic. In, Arsenic in ground water. Springer, pp. 101–125. Weeger, W., Lièvremont, D., Perret, M., Lagarde, F., Hubert, J.C., Leroy, M., and Lett, M.C. (1999) Oxidation of arsenite to arsenate by a bacterium isolated

295 from an aquatic environment. Biometals 12: 141–9. Weier HU; Gray JW. (1988) A Programmable System to Perform the Polymerase Chain Reaction. Dna 7: 441–447. Whaley-Martin, K.J., Mailloux, B.J., Van Geen, A., Bostick, B.C., Silvern, R.F., Kim, C., et al. (2016) Stimulation of Microbially Mediated Arsenic Release in Bangladesh Aquifers by Young Carbon Indicated by Radiocarbon Analysis of Sedimentary Bacterial Lipids. Environ. Sci. Technol. 50:. WHO, G. (2011) Guidelines for drinking-water quality. World Heal. Organ. 216: 303–304. Williamson, A.J., Morris, K., Shaw, S., Byrne, J.M., Boothman, C., and Lloyd, J.R. (2013) Microbial reduction of Fe(III) under alkaline conditions relevant to geological disposal. Appl. Environ. Microbiol. 79: 3320–3326. Wu, M. and Scott, A.J. (2012) Phylogenomic analysis of bacterial and archaeal sequences with AMPHORA2. Bioinformatics 28: 1033–1034. Xiao, K.Q., Li, L.G., Ma, L.P., Zhang, S.Y., Bao, P., Zhang, T., and Zhu, Y.G. (2016) Metagenomic analysis revealed highly diverse microbial arsenic metabolism genes in paddy soils with low-arsenic contents. Environ. Pollut. 211:. Xue, X.-M., Yan, Y., Xu, H.-J., Wang, N., Zhang, X., and Ye, J. (2014) ArsH from Synechocystis sp. PCC 6803 reduces chromate and ferric iron. FEMS Microbiol. Lett. 356: 105–112. Yachdav, G., Kloppmann, E., Kajan, L., Hecht, M., Goldberg, T., Hamp, T., et al. (2014) PredictProtein - An open resource for online prediction of protein structural and functional features. Nucleic Acids Res. 42: 337–343. Yamamura, S. and Amachi, S. (2014) Microbiology of inorganic arsenic: From metabolism to bioremediation. J. Biosci. Bioeng. xx: 1–9. Yamamura, S., Yamamoto, N., Ike, M., and Fujita, M. (2005) Arsenic extraction from solid phase using a dissimilatory arsenate-reducing bacterium. J. Biosci. Bioeng. 100: 219–222. Yan, L., Yin, H., Zhang, S., Leng, F., Nan, W., and Li, H. (2010) Biosorption of inorganic and organic arsenic from aqueous solution by Acidithiobacillus ferrooxidans BY-3. J. Hazard. Mater. 178: 209–217. Ye, J., Coulouris, G., Zaretskaya, I., Cutcutache, I., Rozen, S., and Madden, T.L. (2012) Primer-BLAST: a tool to design target-specific primers for polymerase chain reaction. BMC Bioinformatics 13: 134. Yeates, C., Gillings, M.R., Davison, A.D., Altavilla, N., and Veal, D.A. (1998) Methods for microbial DNA extraction from soil for PCR amplification. Biol. Proced. Online 1: 40–47. Ying, S.C., Masue-Slowey, Y., Kocar, B.D., Griffis, S.D., Webb, S., Marcus, M. a., et al. (2013) Distributed microbially- and chemically-mediated redox processes controlling arsenic dynamics within Mn-/Fe-oxide constructed aggregates. Geochim. Cosmochim. Acta 104: 29–41.

296 Yoshinaga, M., Cai, Y., and Rosen, B.P. (2011) Demethylation of methylarsonic acid by a microbial community. Environ. Microbiol. 13: 1205–1215. Younge, A.J., Henschel, R., Brown, J.T., von Laszewski, G., Qiu, J., and Fox, G.C. (2011) Analysis of Virtualization Technologies for High Performance Computing Environments. Cloud Comput. (CLOUD), 2011 IEEE Int. Conf. Cloud Comput. 9–16. Zafiriadis, I., Ntougias, S., Mirelis, P., Kapagiannidis, a G., and Aivasidis, a (2012) Molecular characterization of denitrifying bacteria isolated from the anoxic reactor of a modified DEPHANOX plant performing enhanced biological phosphorus removal. Water Environ. Res. 84: 475–484. Zargar, K., Conrad, A., Bernick, D.L., Lowe, T.M., Stolc, V., Hoeft, S., et al. (2012) ArxA, a new clade of arsenite oxidase within the DMSO reductase family of molybdenum oxidoreductases. Environ. Microbiol. 14: 1635–1645. Zehnder, A.J.B. and Van Groenestijn, J.W. (1990) Accumulation of polyphosphate by Acinetobacter sp.: Physiology, ecology, and application. In, Novel biodegradable microbial polymers. Springer, pp. 235–243. Zhang, G., Chen, T., Chang, S., Zhang, W., Wu, X., Wu, M., et al. (2016) Complete genome sequence of Acinetobacter sp. TTH0-4, a cold-active crude oil degrading strain isolated from Qinghai-Tibet Plateau. J. Biotechnol. 226: 54– 55. Zhang, Z., Moon, H.S., Myneni, S.C.B., and Jaffé, P.R. (2017) Effect of dissimilatory iron and sulfate reduction on arsenic dynamics in the wetland rhizosphere and its bioaccumulation in wetland plants (Scirpus actus). J. Hazard. Mater. 321: 382–389. Zheng, Y., van Geen, a., Stute, M., Dhar, R., Mo, Z., Cheng, Z., et al. (2005) Geochemical and hydrogeological contrasts between shallow and deeper aquifers in two villages of Araihazar, Bangladesh: Implications for deeper aquifers as drinking water sources. Geochim. Cosmochim. Acta 69: 5203–5218. Zheng, Y., Stute, M., Van Geen, a., Gavrieli, I., Dhar, R., Simpson, H.J., et al. (2004) Redox control of arsenic mobilization in Bangladesh groundwater. Appl. Geochemistry 19: 201–214. Zhu, Y.-G., Xue, X.-M., Kappler, A., Rosen, B.P., and Meharg, A.A. (2017) Linking genes to microbial biogeochemical cycling: lessons from arsenic. Environ. Sci. Technol. Zhu, Y.-G., Yoshinaga, M., Zhao, F.-J., and Rosen, B.P. (2013) Earth Abides Arsenic Biotransformations. Annu. Rev. Earth Planet. Sci. 42: 140307200240001. Zobrist, J., Dowdle, P.R., Davis, J. a., and Oremland, R.S. (2000) Mobilization of Arsenite by Dissimilatory Reduction of Adsorbed Arsenate. Environ. Sci. Technol. 34: 4747–4753. Zul, D., Wanner, G., and Overmann, J. (2008) Massilia brevitalea sp. nov., a novel betaproteobacterium isolated from lysimeter soil. Int. J. Syst. Evol. Microbiol. 58: 1245–51.

297

298