Assessment of Anaerobic Bioremediation Potential in Hydrocarbon Contaminated Aquifers in Alberta, Canada

University of Calgary PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

2019-04-30 Assessment of Anaerobic Bioremediation Potential in Hydrocarbon Contaminated Aquifers in Alberta, Canada

Kharey, Gurpreet Singh

Kharey, G. S. (2019). Assessment of Anaerobic Bioremediation Potential in Hydrocarbon Contaminated Aquifers in Alberta, Canada (Unpublished master's thesis). University of Calgary, Calgary, AB. http://hdl.handle.net/1880/110251 master thesis

University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. Downloaded from PRISM: https://prism.ucalgary.ca UNIVERSITY OF CALGARY

Assessment of Anaerobic Bioremediation Potential in Hydrocarbon Contaminated Aquifers in

Alberta, Canada

Gurpreet Singh Kharey

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF MASTER OF SCIENCE

GRADUATE PROGRAM IN BIOLOGICAL SCIENCES

CALGARY, ALBERTA

APRIL, 2019

Anaerobic biodegradation of hydrocarbon fuels (mono-aromatic and short chain alkane hydrocarbon) was demonstrated under both field and microcosm experiments, with detection of signature metabolites, fumarate addition genes, and microbial community composition to determine the potential of hydrocarbon contaminated sites for bioremediation and assess the evidence of its past and/or present occurrence. It was determined here that the potential for bioremediation is unique to both the contaminated sites, but also to the hydrocarbon load present.

Quantification of the fumarate addition genes using a newly designed mixture of primers was done. These qPCR assays for assA and bssA abundances concluded that an upper limit to gene abundance is present according to hydrocarbon concentrations, of approximately 5 ppm hydrocarbon. Monitoring hydrocarbon loss from groundwater collected at these sites, paired with fumarate addition gene abundances concluded that the contaminated sites in question have the potential for bioremediation, but is limited with higher hydrocarbon concentrations.

Keywords: anaerobic hydrocarbon degradation, fumarate addition, qPCR, signature

metabolite, hydrocarbon, contamination.

ii Acknowledgements

If I were left to achieve all that I have by myself, I’m afraid I would not have made it past

Kindergarten. Heartbroken that playtime was over, and my toys would need to be put away, I refused. I continue to refuse to this day. To thank all those who have had a hand in keeping me on-track and allowing me to grow into the person I am today would be a thesis unto itself, because I am truly grateful for all those that helped me to be where I am, and where I will go from here.

Firstly, I thank the Almighty for everything; all glory is yours, no one shall know my name. I thank my parents, for they had more influence over my successes than I did. For always supporting me, pushing me to be better, and giving me their experience, love, and counsel, even when they didn’t understand what I do. Absolutely none of this would have been even a dream if it wasn’t for my parents. Because of them, and for them, this dream became a reality.

I thank my supervisor, Dr. Lisa Gieg, for believing and taking a chance on a skinny, awkward kid that had no idea what to do in life and telling me that I’m doing better than I think.

Thank you for always supporting and teaching me that I can succeed. Also, thanks to my committee members, Dr. Raymond Turner and Dr. Casey Hubert for their support and expertise in making my work better. Thank you to Dr. Fredrick Biddle and Dr. David Hansen for giving me the opportunity to teach in my undergraduate, the true motivator for me to pursue a Graduate degree. Thank you to the United Farmers of Alberta (UFA) for providing samples and related information on the sites in this study, and to NSERC for funding the research.

Huge thanks to my lab-mates/science-friends, past and present, for all the things they have taught me about science and life, and just being amazing people in general (and putting up with my insanity/aggressive humility). There’s just too many to acknowledge (in no order): to

iii Gabrielle for doing the initial groundwater analyses and helping me get my project on its feet; to

Corynne for teaching me all the molecular biology techniques; to Courtney and Carolina for teaching me how to be a good scientist; to Nicole and Danika for being my most trusted advisors in all things hydrocarbon and beyond; to Matija for your priceless advice in all things qPCR and beyond; to the lab Blackboard for being there to absorb all my frustrations and sparking my interest in art; to Steven for the countless hours spent in lab as my student; and to all the others for being such great friends and making my Masters the best years of my academic career.

Huge thanks to my non-science friends: Sumare, Harman, Farwa, Kelsey, and all the others, for keeping me motivated to achieve, inspiring me, keeping me sane, and pushing me to be a better man every day. I thank you all for your years of support, trust, and inspiration; for celebrating my successes, and your unmatched counsel when I struggled. I cannot put into words the impact each one of you has had on my academic and personal lives.

I would be half the man I am today without you all, and you all will make me a much better man than I can imagine in the future.

I am eternally grateful.

iv Dedication

To my parents,

Sardar Jagtar Singh and Sardarni Gurdeep Kaur

without you, there is nothing.

v Table of Contents

Abstract ...... ii Acknowledgements ...... iii Dedication ...... v Table of Contents ...... vi List of Tables ...... ix List of Figures and Illustrations ...... xii List of Symbols, Abbreviations and Nomenclature ...... xvi Epigraphs ...... xvii

Chapter 1 : Literature Review ...... 1 1.1 Introduction ...... 1 1.2 Physical and Chemical Remediation Approaches ...... 2 1.3 Bioremediation ...... 4 1.4 Anaerobic Hydrocarbon Biodegradation ...... 6 1.5 Diagnostic Hydrocarbon Metabolites to Indicate In Situ Anaerobic Bioremediation ...... 11 1.6 Anaerobic HC Biodegradation Genes – Advantages and Limitations ...... 13 1.7 Use of Microcosms to Study Hydrocarbon Biodegradation ...... 20

Chapter 2 : Research Objectives and Goals ...... 23

Chapter 3 : Analysis of Field Sites for Evidence of Anaerobic Biodegradation ...... 26 3.1 Introduction ...... 26 3.2 Methods: ...... 28 3.2.1 Site Description...... 28 3.2.1.1 Site A...... 28 3.2.1.2 Site B...... 30 3.2.2 Sample collection...... 31 3.2.3 Hydrocarbon and Electron Acceptor Analyses ...... 32 3.2.4 Biomass Collection...... 33 3.2.5 DNA Extraction...... 33 3.2.6 16S rRNA Gene Sequencing Sample Preparation and Analysis ...... 33 3.2.7 assA and bssA PCR Presence/Absence Assay ...... 35 3.3 Metabolite Extraction and Analysis ...... 36 3.4 Results and Discussion ...... 37 3.4.1 Site A ...... 37 3.4.1.1 Hydrocarbon concentrations and potential electron acceptors...... 37 3.4.2 Site B ...... 40 3.4.2.1 Hydrocarbon concentrations and potential electron acceptors...... 40 3.4.3 Metabolic and gene evidence for in situ HC biodegradation...... 43 3.4.3.1 Site A signature metabolites and FAE gene presence...... 43 3.4.3.2 Site B signature metabolites and FAE gene presence...... 48 3.4.3.3 Electron acceptor concentrations in field sites...... 52 3.4.4 16S rRNA gene sequencing using Illumina Mi-Seq...... 56 3.4.4.1 Site A microbial community composition...... 56

vi 3.4.4.2 Site B microbial community composition...... 61 3.5 Conclusions ...... 64

Chapter 4 : Primer design for quantification of assA and bssA on field samples from Site A and Site B ...... 65 4.1 Introduction: ...... 65 4.2 Methods: ...... 66 4.2.1 Designing assA and bssA primers that capture diversity ...... 66 4.2.2 qPCR Thermocycling Conditions: ...... 70 4.2.3 Sample Preparation/DNA Extractions ...... 72 4.2.4 qPCR Standards for absolute quantification ...... 72 4.2.5 Illumina MiSeq sequencing of unknown amplicons and Sanger sequencing of standards ...... 74 4.3 Results and Discussion: ...... 77 4.3.1 Newly designed bssA qPCR primer mix ...... 77 4.3.1.1 Site A bssA detection...... 77 4.3.1.2 Site A bssA quantification...... 80 4.3.1.3 Site B bssA detection...... 83 4.3.1.4 Site B bssA quantification...... 85 4.3.2 Newly designed assA qPCR primer mix...... 88 4.3.2.1 Site A assA detection...... 88 4.3.2.2 Site A assA quantification...... 89 4.3.2.3 Site B assA detection ...... 91 4.3.2.4 Site B assA quantification...... 93 4.3.3 Effect of different polymerases on FAE gene detection ...... 95 4.3.4 assA and bssA qPCR amplicon sequencing using Illumina MiSeq...... 97 4.3.4.1 assA sequencing results...... 97 4.3.4.2 bssA sequencing results...... 101 4.4 Conclusions ...... 105

Chapter 5 : Microcosm experiments monitoring hydrocarbon loss and assA and bssA gene abundance ...... 107 5.1 Introduction: ...... 107 5.2 Methods: ...... 108 5.2.1 Microcosm Experimental Design...... 108 5.2.1.1 Hydrocarbon monitoring ...... 109 5.2.1.2 HC Degradation Stoichiometric Reaction Equations under tested conditions...... 110 5.2.2 DNA Extraction ...... 111 5.2.3 16S rRNA gene sequencing for microbial community composition ...... 112 5.2.4 qPCR Analysis for assA and bssA genes ...... 112 5.3 Results and Discussion: ...... 112 5.3.1 Microcosm Well C01-04 – Site A, Low Contamination...... 112 5.3.1.1 Hydrocarbon degradation ...... 112 5.3.1.2 FAE gene abundance in C01-04 microcosms...... 116 5.3.1.3 Microbial community composition of C01-04 microcosms...... 122

vii 5.3.1.4 Summary of microcosms prepared with C01-04 groundwater, low contamination ...... 124 5.3.2 Microcosm Well C02-08 – Site A, High Contamination ...... 125 5.3.2.1 Hydrocarbon degradation ...... 125 5.3.2.2 FAE gene analysis/quantification ...... 128 5.3.2.3 Microbial community composition ...... 129 5.3.3 Summary of Microcosms prepared with C02-08 groundwater, high contamination ...... 130 5.3.4 Microcosm Well REC 12 – Site B, Low Contamination...... 131 5.3.4.1 Hydrocarbon degradation...... 131 5.3.4.2 FAE gene abundance...... 137 5.3.4.3 Microbial community composition of REC 12 groundwater sample.143 5.3.5 Summary of Microcosms prepared with REC 12 groundwater, low contamination ...... 145 5.3.6 Microcosm Well REC 34 – Site B, High Contamination ...... 146 5.3.6.1 Hydrocarbon degradation...... 146 5.3.6.2 FAE gene abundance...... 150 5.3.6.3 Microbial community composition...... 151 5.3.6.4 Summary of Site B microcosm REC 34, higher contamination. ..152 5.4 Conclusions ...... 153

Chapter 6 : Conclusions and Future Directions...... 156

Chapter 7 : References ...... 160

Chapter 8 : Appendix ...... 177 8.1 Media Recipe ...... 179

viii List of Tables

Table 1: Site A hydrocarbon concentrations. Samples were collected Sept 9, 2016, and analyzed by AGAT Laboratories. Shaded cells represent concentrations higher than the Alberta Guidelines for hydrocarbons in groundwater...... 38

Table 2: Site B hydrocarbon concentrations. Samples were collected Sept 9, 2016 and analyzed by AGAT Labs. Shaded cells represent concentrations higher than the Alberta Guidelines for hydrocarbons in groundwater...... 41

Table 3: Signature metabolites and FAE gene presence in Site A. Hydrocarbon analysis was done in 2016 only. ‘2016’ and ‘2017’ denoted presence of the metabolite in that year. ‘assA’ and ‘bssA’ denote detection of that FAE gene. Blank spaces indicate no detection. Dashed spaces indicate wells were not sampled in that year. A comparison of Taq and KAPA polymerases in FAE gene detection is also shown...... 45

Table 4: Signature metabolites and FAE gene presence in Site B. Hydrocarbon analysis was done in 2016 only. ‘2016’ and ‘2017’ denoted presence of the metabolite in that year. ‘assA’ and ‘bssA’ denote detection of that FAE gene. Blank spaces are no detection. Bolded well names were not sampled in 2017. Dashed spaces indicate wells were not sampled in that year. A comparison of Taq and KAPA polymerases in FAE gene detection is also shown...... 49

Table 5: Primers used and designed in this study. Melting temperature Tm calculated via IDT OligoAnalyzer 3.1 online tool using default qPCR parameters. ‘Literature primers’ are those which were not used for qPCR purposes, but were used in non-qPCR assays (this also includes 8543r, which was also used in qPCR assays). Primers designed but ultimately omitted from primer mix are shaded...... 68

Table 7: Comparison of bssA presence in Site A between literature primers (comparing amplification via Fermentas Taq and KAPA Hi-Fi) and newly-designed bssA qPCR primer mixture. “bssA” denotes presence, blank space denotes absence. Dashed spaces indicate wells were not sampled in that year...... 79

Table 8: Comparisons of approximate reported assA and bssA abundances reported in literature. Note that abundances reported in this study are up to 2 log greater than other studies, in both detection limit and in highest reported abundances...... 83

Table 9: Comparison of bssA presence in Site B between literature primers (comparing amplification via Fermentas Taq and KAPA Hi-Fi) and newly-designed bssA qPCR primer mixture. “bssA” denoted presence, blank space denotes absence. Wells in bold- face and shaded cells were not sampled in 2017. Dashed spaces indicate wells were not sampled in that year...... 85 ix Table 10: Comparison of assA presence in Site A between literature primers (comparing amplification via Fermentas Taq and KAPA Hi-Fi) and newly-designed assA qPCR primer mixture. “assA” denotes presence, blank space denotes absence. Dashed spaces indicate wells were not sampled in that year...... 89

Table 11: Comparison of assA presence in Site B between literature primers (comparing amplification via Fermentas Taq and KAPA Hi-Fi) and newly-designed assA qPCR primer mixture. “assA” denotes presence, blank space denotes absence. Wells in bold- face and shading were not sampled in 2017. Dashed spaces indicate wells were not sampled in that year...... 93

Table 14: Comparison of assA gene detection sing literature primers 7757af/8543r46 and newly designed qPCR primer mix. ‘assA’ denotes detection of the gene. ‘NSP’ denotes presence of non-specific bands’ either larger or smaller than expected amplicon size. Blank spaces mean no detection...... 117

Table 15: Comparison of bssA gene detection using literature primers 7772f/8546r and newly designed qPCR primer mix. ‘bssA’ denotes detection of the gene. ‘NSP’ denotes presence of non-specific bands’ either larger or smaller than expected amplicon. Blank spaces mean no detection...... 122

Table 19: Comparison of assA gene detection using literature primers 7757af/8543r and newly designed qPCR primers. ‘assA’ denotes detection of the gene. ‘NSP’ denotes

x presence of non-specific bands’ either larger or smaller than expected amplicon. Blank spaces mean no detection...... 138

Table 20: Comparison of bssA gene detection using literature primers 7772f/8546r and newly designed qPCR primers. ‘bssA’ denotes detection of the gene. ‘NSP’ denotes presence of non-specific bands’ either larger or smaller than expected amplicon. Blank spaces mean no detection...... 143

Table 22: Summary of major findings from each microcosm discussed in this Chapter...... 155

Table 23: Sequences used to compile Multiple Sequence Alignment (MSA) for assA and bssA primer design. Sequences were assembled in the Benchling online tool. Note, some sequences were trimmed for improved alignment ...... 177

Table 24: Compilation of identified non-specific amplicons from newly designed assA primer mix ...... 178

Table 25: Compilation of identified non-specific amplicons from newly designed bssA primer mix ...... 178

xi List of Figures and Illustrations

Figure 1: Generalized schematic of fumarate addition to octane and toluene, and subsequent metabolic pathways of anaerobic HC biodegradation...... 10

Figure 2: Site map of Site A, with hydrocarbon concentrations measured in 2016 (by AGAT Laboratories). Also included are concentrations of nitrate, sulfate, and iron (II) in 2016 and 2017, as measured in this thesis work. Nitrate and sulfate analyses were performed using HPLC. Iron (II) analysis was done using a modified colorimetric assay71. ‘N/S’ means Not Sampled. Intial hydrocarbon spill is believed to be in the area marked ‘AST NEST’. Groundwter flow direction is marked with arrows...... 29

Figure 3: Site B map outlining positions of sampled wells. Each sampled well is accompanied by a blowout listing hydrocarbon concentration in 2016, and electron acceptor concentrations for 2016 and 2017. Nitrate and sulfate analyses were performed using HPLC. Soluble iron (II) analysis was done using a modified colorimetric assay71. ‘N/S’ means Not Sampled. Groundwater flow is eastward...... 31

Figure 4: Concentrations of hydrocarbons benzene, toluene, ethylbenzene, xylenes (BTEX), and C6 – C10 alkanes in each sampled well of Site A in 2016. Sampling and hydrocarbon analysis were done by AGAT Laboratories...... 40

Figure 5: Concentrations of the hydrocarbons benzene, toluene, ethylbenzene, and xylenes (BTEX) and C6 – C10 alkanes in each sampled well of Site B in 2016. Analysis was done by AGAT Laboratories...... 43

Figure 6: Concentrations of various electron acceptors in Site A between 2016 and 2017. Nitrate (A), sulfate (B), and iron (II) (C) are reported in mM. Ranking of wells is presented in numerical order...... 54

Figure 7: Concentrations of various electron acceptors in Site B between 2016 and 2017. Nitrate (A), sulfate (B), and iron (II) (C) were analyzed and are reported in mM. Ranking of wells is alphabetical...... 56

Figure 8: Microbial community composition of Site A in 2016. Amplification of the 16S rRNA gene using primers 926f/1392r 16S degenerate primers targeting the V6-V8 hypervariable regions of 16S rRNA gene. All taxa whose abundances are below 5% were removed and summed into “Other (Below 5% Relative Abundance)”. Wells are in order of decreasing total HC concentration; uncontaminated wells are C02-07 and C02- 06...... 59

Figure 9: Microbial community composition of Site A in 2017. Amplification of the 16S rRNA gene using primers 926f/1392r 16S degenerate primers targeting the V6-V8 hypervariable regions of 16S rRNA gene. All taxa whose abundances are below 5% were removed and summed into “Other (Below 5% Relative Abundance)”. Wells are in order of decreasing total HC concentration; uncontaminated well is C02-06...... 60

xii Figure 10: Microbial community composition of Site B in 2016. Amplification of the 16S rRNA gene using primers 926f/1392r 16S degenerate primers targeting the V6-V8 hypervariable regions of 16S rRNA gene. All taxa whose abundances are below 5% were removed and summed into “Other (Below 5% Relative Abundance)”. Wells are in order of decreasing total HC concentration; uncontaminated well is S14-49B...... 62

Figure 11: Microbial community composition of Site B in 2017. Amplification of the 16S rRNA gene using primers 926f/1392r 16S degenerate primers targeting the V6-V8 hypervariable regions of 16S rRNA gene. All taxa whose abundances are below 5% were removed and summed into “Other (Below 5% Relative Abundance)”. Wells are in order of decreasing total HC concentration, uncontaminated well is S14-49B...... 63

Figure 12: Sample standard curve of assA quantification. assA amplicons amplified using 7757af/8543r were used as standard DNA46. This standard curve was prepared from Site B 2016 field sample quantification...... 71

Figure 13: Sample standard curve of bssA quantification. bssA amplicons amplified using 7772af/8546r were used as standard DNA24. This standard curve was prepared from Site A 2017 field sample quantification...... 73

Figure 14: Quantifications of the bssA gene in Site A for 2016 (red circles) and 2017 (black squares). Linear regression trendlines are matched in colour. Only those samples whose amplicons matched the length of standard amplicons are plotted here. It is worth noting that other samples can comparable quantification values, arising from non-specific binding (data not shown). Error bars represent standard deviation of technical replicates (n=3)...... 82

Figure 15: Quantifications of the bssA gene in Site B for 2016 (red circles) and 2017 (black squares). Linear regression trendlines are matched in colour. Only those samples whose amplicons matched the length of standard amplicons are plotted here. It is worth noting that other samples have comparable quantification values, arising from non-specific binding (data not shown). Error bars represent standard deviation of technical replicates (n=3)...... 86

Figure 16: Quantifications of the assA gene in Site A for 2016 (black squares) and 2017 (red circles). Linear regression trendlines are matched in colour. Only those samples whose amplicons matched the length of standard amplicons are plotted here. It is worth noting that other samples can comparable quantification values, arising from non-specific binding (data not shown). Note the break in x-axis between 1 and 4 ppm alkane concentrations. Error bars represent standard deviation of technical replicates (n=3)...... 90

Figure 17: Quantifications of the assA gene in Site B for 2016 (red circles) and 2017 (black squares). Linear regression trendlines are matched in colour. Only those samples whose amplicons matched the length of standard amplicons are plotted here. It is worth noting that other samples can comparable quantification values, arising from non-specific binding (data not shown). Error bars represent standard deviation of technical replicates (n=3)...... 94

xiii Figure 18: Maximum Likelihood phylogram of sequenced assA and bssA amplicons. Nucleic acid sequences were aligned with ClustalX (30 iterations). Tree-building via PhyML 3.0 Maximum Likelihood using TN93 substitution model, fast likelihood based method aLRT for branch supporting, and SPR (Subtree Pruning and Re-Grafting) tree improvement model86,87,99. The outgroup is pyruvate formate lyase (pfl) from sulfate- reducing strain D. alkenivorans strain AK-01...... 99

Figure 19: Monitoring C01-04 sourced microcosms for hydrocarbons (benzene, toluene, and octane) over a 111-day period under various electron accepting conditions (nitrate (A), sulfate (B), iron (III) (C), ‘All 3’ (III) (D) no added electron acceptor ‘No EA’ (E), and sterile controls (F)). Note that benzene and toluene concentrations are on the left y-axis, and octane concentrations are on the right y-axis. All microcosms, including sterile controls, were re-amended with BTO on day 66. Analyses were done by GC-FID. Error bars represent standard deviation of 4 test replicates, and 2 sterile control replicates...... 114

Figure 20: Electron acceptor concentrations in Site A C01-04 sourced groundwater microcosms on Day 1 and Day 90 of incubation. Nitrate (A), Sulfate (B), and Iron (II) (C) are shown across all experimental conditions (nitrate, sulfate, iron, nitrate+sulfate+iron (all 3), no added electron acceptor (No EA), and kill controls). Error bars represent standard deviation of 4 replicates. Multiple t-tests were done for HC amended cultures. p<0.005 is denoted by ‘*’...... 116

Figure 21: Quantification of assA gene in Site A sourced C01-04 well groundwater. Initial Day 0 is shown as “T=0”, all other quantifications are on Day 90. Units are in copies of assA per mL of microcosm. Error bars represent standard deviation of 4 replicates...... 119

Figure 22: Microbial community composition of C01-04 microcosms shown to the lowest taxonomic level. The ‘No EA’ cultures did not have a No HC condition. ‘Other (Below 5%)’ contain identified taxa present under 5% relative sequence abundance...... 124

Figure 23: Monitoring of C02-08 sourced microcosms for hydrocarbons (benzene, toluene, and octane) over a 134-day period under various electron accepting conditions (nitrate (A), sulfate (B), iron (III) (C), ‘All 3’ (III) (D) no added electron acceptor ‘No EA’ (E), and sterile controls (F)). Note that benzene and toluene concentrations are on the left y- axis, and octane concentrations are on the right y-axis. Analyses were done by GC-FID. Error bars represent standard deviation of 4 replicates, sterile control for 2 replicates...... 126

Figure 24: Electron acceptor concentrations in Site A C02-08 sourced groundwater microcosms on Day 1 and Day 90 of incubation. Nitrate (A), Sulfate (B), and Iron (II) (C) are shown across all experimental conditions (nitrate, sulfate, iron, nitrate+sulfate+iron (‘All 3’), no added electron acceptor (No EA), and kill controls). Error bars represent standard deviation of 4 replicates. Multiple t-tests were done for HC amended cultures. p<0.005 is denoted by ‘*’...... 128

Figure 25: Microbial community composition of C02-08 microcosms. ‘No HC’ cultures were not sequenced as no culture existed with no HC added. ‘Other (Below 5%)’ contain identified taxa present under 5% relative read abundance...... 129

xiv Figure 26: Monitoring REC 12 sourced microcosms for hydrocarbons (benzene, toluene, and octane) over a 121-day period under various electron accepting conditions (nitrate (A), sulfate (B), iron (III) (C), ‘All 3’ (III) (D) no added electron acceptor ‘No EA’ (E), and sterile controls (F)). Note that benzene and toluene concentrations are on the left y-axis, and octane concentrations are on the right y-axis. Analyses were done by GC-FID. Error bars represent standard deviation of 4 replicates, sterile control for 2 replicates...... 133

Figure 27: Electron acceptor concentrations in Site B REC 12 sourced groundwater microcosms on Day 1 and Day 90 of incubation. Nitrate (A), Sulfate (B), and Iron (II) (C) are shown across all experimental conditions (nitrate, sulfate, iron, nitrate+sulfate+iron (all 3), no added electron acceptor (No EA), and kill controls). Error bars represent standard deviation of 4 replicates. Multiple t-tests were done for HC amended cultures. p<0.005 is denoted by ‘*’...... 136

Figure 28: Quantification of assA gene in Site B sourced REC 12 well groundwater. Initial Day 0 is shown as “T=0”, all other quantifications are on Day 90. Units are in copies of assA per mL of microcosm. Error bars represent standard deviation of 4 replicates...... 139

Figure 29: Melt peak chart of assA amplicons from REC 12 microcosms. Thick black = 'All 3'; grey = Nitrate, Iron, No EA; dark grey = sulfate. One replicate curve of three shown. Curve was generated using Bio-Rad CFX Manager software...... 141

Figure 30: Microbial community composition of REC 12 Microcosms. No EA cultures did not have a No HC condition. ‘Other (Below 5%)’ contain identified taxa present under 5% relative sequence abundance...... 145

Figure 31: Monitoring REC 34 sourced microcosm for hydrocarbons (benzene, toluene, and octane) over a 120-day period under various electron accepting conditions (nitrate (A), sulfate (B), iron (III) (C), ‘All 3’ (III) (D) no added electron acceptor ‘No EA’ (E), and sterile controls (F)). Note that benzene and toluene concentrations are on the left y-axis, and octane concentrations are on the right y-axis. Analyses were done by GC-FID. Error bars represent standard deviation of 4 replicates, sterile control for 2 replicates...... 148

Figure 32: Electron acceptor concentrations in Site B REC 34 sourced groundwater microcosms on Day 1 and Day 90 of incubation. Nitrate (A), Sulfate (B), and Iron (II) (C) are shown across all experimental conditions (nitrate, sulfate, iron, nitrate+sulfate+iron (‘All 3’), no added electron acceptor (No EA), and kill controls). Error bars represent standard deviation of 6 replicates, as HC-free cultures were not possible. Multiple t-tests were done for HC amended cultures. p<0.005 is denoted by ‘*’...... 150

Figure 33: Microbial community composition of REC 34 microcosms. Note no sequencing for ‘No HC’ was done due to no HC-free sample. ‘Other (Below 5%)’ contain identified taxa present under 5% relative sequence abundance...... 152

xv List of Symbols, Abbreviations and Nomenclature

Symbol/Abbreviations Definition ASS Alkylsuccinate Synthase Ass Alkylsuccinate synthase operon assA Alkylsuccinate synthase alpha subunit gene AST Aboveground Storage Tank BSS Benzylsuccinate Synthase Bss Benzylsuccinate synthase operon bssA Benzylsuccinate synthase alpha subunit gene BTEX Benzene, Toluene, Ethylbenzene, and Xylenes BTO Mixture of Benzene, Toluene, and Octane dsDNA Double-stranded DNA EA Electron Acceptor FAE Fumarate Addition Enzyme G+C Guanidine + Cytosine content GC Gas Chromatography GC-MS Gas Chromatography Mass Spectrometry gDNA Genomic DNA GRE Glycyl Radical Enzyme GRE-AE Glycyl Radical Enzyme – Activating Enzyme GW Groundwater HC Hydrocarbon(s) HPLC High Performance Liquid Chromatography NCBI National Center for Biotechnology Information NMS Naphthylsuccinate Synthase enzyme nmsA Naphthylmethylsuccinate synthase alpha subunit gene No HC No Added Hydrocarbon(s) NTC No Template Control PCR Polymerase Chain Reaction qPCR Quantitative PCR Tm Melting Temperature UST Underground Storage Tank

xvi Epigraphs

There are plenty of others willing to call you a failure […] Don't you ever say it of yourself […]

Do you understand? You care about something, you fight for it. You hit a wall, you push through

it. There's something you need to know about failure, Tintin, you can never let it defeat you.

Captain Archibald Haddock in The Adventures of Tintin (movie) (2011)

“If you’re not having fun, you’re doing something wrong.”

“Humor is reason gone mad.”

“Those are my principles, if you don’t like them…well, I have others.”

“That's three quotes? Add another quote and make it a gallon.”

Groucho Marx

Can’t stop runnin’ till I run it, man I feel this in my stomach, I’ve been training my whole life for

everything that I got comin’

‘Run It’ by Villa (2018)*

Uchiyan ne gallan tere yaar diyan†

‘So High’ by Sidhu Moosewala (2017)

Hold on to your butts

Samuel L. Jackson as Ray Arnold in Jurassic Park (movie) (1993)

* Quoted with permission from artist † Your friend is doing big things (rough translation from Punjabi)

xvii 1

Chapter 1: Literature Review

1.1 Introduction

The persistence of hydrocarbons in the environment as a result of failing infrastructure, spills, and industrial accidents is a long lasting, expensive and time-consuming issue. Events involving the uncontrolled release of hydrocarbons into the environment can range from a cataclysmic scale (i.e. Deepwater Horizon, Exxon Valdez, pipeline ruptures, train derailments, etc.) to a smaller, localized scale of leaking gasoline station tanks, overfill spills, improper disposal, and release during transfer between receptacles1–3. Generally, the hydrocarbons involved in the ‘cataclysmic’ releases are in the form of crude oil, a complex mixture of unrefined hydrocarbons removed from the Earth’s subsurface. More common and largely unreported are smaller leaks/releases involving gasoline and diesel, which are refined petroleum fractions of specified complexity. Hydrocarbons, compounds composed of hydrogen and carbon chains or rings of varying lengths, are the primary components of fuels such as gasoline and diesel. The composition of gasoline and diesels fuels are C6 – C10 hydrocarbons or C10 – C25, respectively, existing as linear, cyclic, or aromatic compounds. The ubiquity of gasoline as a fuel source for machinery and vehicles makes hydrocarbon spills an on-going and ever-present hazard to the environment and human health (as reviewed by Hilpert et al 2015)3. Our primary source of contact for gasoline/diesel are at gasoline/diesel stations, where their respective hydrocarbon tenders are stored underground in metal tanks. After exposure to moisture over time, these underground metal tanks can corrode and inevitably leak hydrocarbons into the environment. Once introduced to groundwater, the contaminants can be carried far from the source (due to groundwater flow) and potentially enter drinking water sources, bodies of water,

2 or housing infrastructure. Exposure to these chemicals may lead to the development of chronic illnesses, such as nasal and renal cancers, leukemia, and lower quality of life3,4.

Remediation technologies to remove spilled gasoline/diesel hydrocarbons (HC) range from physical to chemical to microbiological methods and have specific conditions where they are most effective1. The remediation of contaminated soils and groundwater require different methods and technologies and has been reviewed in-depth by Gan et al. (2009), Gitipour et al.

(2018) and others, including advantages, disadvantages, and implementation cost1,5–7. A brief summary of some soil and groundwater remediation methods are described below.

1.2 Physical and Chemical Remediation Approaches

Soil washing involves the excavation of contaminated soils and ‘washing’ the soil with water (and optional added solvents) as a method of dissolving the organic contaminants. These methods are not suitable for certain types of soil or where excavation is not possible1,8.

Additionally, the residual solvent needs to be cleaned before disposal. This is similar to landfarming where excavated soils are spread out and instead of washing with a solvent, the soil is amended with nutrients and air, stimulating aerobic bioremediation through microbial growth.

Both technologies have been used for decades and have been found to remove up to 90% of contaminants. However, both methods cannot be used where excavation is not possible and is not suitable for certain dense soils (e.g. clayey soils)1,5.

In situ Soil Vapor Extraction is a method in which the non-soluble components of hydrocarbon fuels are first volatilized with air injection and subsequently vacuumed for destruction above ground1,8. An added benefit of this technology is the formation of an oxygen rich environment that can stimulate aerobic bioremediation (microbial breakdown of contaminants using oxygen as a terminal electron acceptor)9,10. However, this technology cannot

3 be used near groundwater or in denser soils (soils with pore sizes <5 μm), due to restricted groundwater flow1,8. Other remediation systems ranging from the simpler (as briefly introduced here) to the outlandish (e.g., Vitrification, using higher sources of energy to convert the contaminated soil to glass in situ) are reviewed in depth by Khan et al. (2004)1.

Contamination of soils is generally localized, however, if the contamination plume reaches a groundwater aquifer, the contamination can spread quickly due to the groundwater flow away from the source. This possibility for a dynamic contamination event poses unique and urgent problems. For groundwater remediation, the most popular remediation technology has been

‘pump-and-treat’, which requires removal of the contaminated groundwater for remediation above ground. This method is generally chastised for its inefficiency of completely removing the contaminated groundwater, and disadvantages include the length of time (upwards of 50 years), high cost of implementation, and inability to remediate contaminants to acceptable drinking water guidelines1. Although a rather logical method to remove contaminants, the dynamic nature of groundwater does not allow for such simple methods in most cases. Other methods involving the implementation of aeration or vacuum system wells (bioslurping, horizontal wells) or the erection of underground reactive walls/barriers (which either sorb or react with contaminants) require a large investment of capital and depending on the size and nature of the contaminants, and can be relatively long term (decades) treatments with high maintenance costs.

Generally, the remediation systems and technologies require a large investment of capital, specific soil/groundwater conditions, and would require the contaminated areas to be readily accessible for excavation/well drilling. However, these methodologies are limited in use due to the infrastructure around which gasoline stations are built (roads, sidewalks, buildings, etc.) which cannot be removed under most circumstances. With the contaminant plume having the

4 ability to travel hundreds of metres away from the point source, remediating potentially square kilometres worth of contaminated soil and groundwater is not cost/time effective. As such, many remediation methods cannot be used efficiently in urban areas.

1.3 Bioremediation

The implementation of other methods that do not require large capital investments or the use of large expensive machinery/processes that can ultimately increase pollution elsewhere

(incineration, volatilization into the air, etc.) needs to be done. Within the last few decades, interest in the exploitation of the vastly diverse metabolisms of microorganisms present natively in the groundwater to degrade contaminating hydrocarbons in situ has piqued the interest of industry.

Aerobic degradation of hydrocarbons is well known and is routinely implemented in other bioremediation strategies (e.g., landfarming, aeration of soils in situ)1,8,9,11,12, but anaerobic degradation is relatively new, with intense study emerging in the last few decades13,14. Recent focus has been on the understanding of anaerobic hydrocarbon degradation. This is because when hydrocarbons are introduced into a groundwater system, the subsequent increase in aerobic microbial activity rapidly consumes the oxygen, causing an anoxic (or micro-oxic) environment to form15–17. Hydrocarbon degradation activity in the now anoxic plume thus must rely on anaerobic metabolism, which uses other terminal electron acceptors such as nitrate, sulfate, iron

(III), or carbon dioxide. Groundwater systems are heterogeneous depending on the associated geology and geochemistry, influencing groundwater movement, the types of carbon sources, and terminal electron acceptors available (biogeochemistry). Focus on in situ bioremediation of contaminated groundwater systems has been on anaerobic HC degradation to cater to that environment.

The benefits of using biological processes to remediate hydrocarbon contamination are:

(1) that the investment of capital is substantially less (other physical/chemical methods can reach into the thousands of dollars per square yard), (2) limited infrastructure is needed (i.e. drilling monitoring wells) and (3) is a recognized method for remediation1,9. Allowing biological processes to attenuate the contaminants also means that the remediation mechanism is travelling with the contaminants in a groundwater system and there is no secondary remediation required

(i.e. cleaning pulled water or volatilized contaminants) as the toxic compounds are mineralized

(completely broken down) to CO2 or incorporated into biomass. However, a disadvantage to using bioremediation is the rate at which contaminants are removed/degraded; this can be substantially slower than using physical/chemical methods18. Physical removal of contaminants removes large volumes of contaminated material, either by excavation or pulling groundwater, while biological processes take longer due to limitations in nutrients, functioning under non- optimal temperature conditions, etc10. In contaminated sites where bioremediation is feasible, it may be the most attractive option from a capital standpoint. In most cases, using more than one technology, i.e. combining physical/chemical methods to remove most of the contaminants, followed by bioremediation to remove the remaining contaminants, is the most lucrative option.

Unfortunately, there is no singular method available to our knowledge, which has wide applicability, high remediation efficiency, and low operational/monitoring costs.

Bioremediation is a fundamental part of many remediation methods, but its effective use is reliant on accurate testing and monitoring for both the initial potential for bioremediation, and to monitor whether bioremediation is indeed occurring. Part of the challenges in establishing this biodegradation potential is the diversity of hydrocarbons present which may be themselves detrimental to microbial life. As such, microbes that can degrade these toxic compounds are in a

6 state of flux between life and death, and are limited by the concentrations of toxic compounds present and the environment they are found (~2 mM total HC19, ~300 μM benzene; Gieg, personal communication). Within the last 15 years, the ‘plume fringe concept’ was introduced showing that anaerobic degradation of hydrocarbons in a groundwater aquifer is most active at the edges of the plume, (either the leading edge, or the top and/or bottom of the water edge) where the concentration of the contaminants are lower, and bioavailability of electron acceptors is high20–22. Activity in the ‘body’ of the plume is lower due to limited resource availability and higher contamination concentration. Thus, studies into the intricacies of anaerobic hydrocarbon bioremediation are vital to efficiently implement this strategy. The use of bioremediation has been implemented on a number of hydrocarbon contaminated sites, which are also listed in part in Khan et al. (2004)1. Indeed, many studies assessing environmental sites are focused on determining potential for using bioremediation to increase the understanding of its effective use2,23–27.

1.4 Anaerobic Hydrocarbon Biodegradation

Hydrocarbons are intrinsically recalcitrant compounds, and resistant to breakdown under environmental conditions due to the stability of the C-C bonds. Aerobically, this is achieved by

O2 and oxygenases which produce a highly reactive oxygen radical that can overcome this stability. The attachment of a more reactive moiety, or ‘activation’ of the hydrocarbon, is a necessary step for microbial/enzymatic degradation to occur28. Anaerobically, carboxylation, hydroxylation, methylation, and fumarate addition have been characterized as activation strategies for monoaromatics, alkanes, and polyaromatic compounds. Carboxylation, hydroxylation, and methylation have been reported for the activation of unsubstituted aromatics like benzene, the most recalcitrant of the BTEX (benzene, toluene, ethylbenzene, and xylenes)

7 compounds29. Conversion of benzene to toluene (methylation), benzoate (carboxylation), or phenol (hydroxylation) allows for degradation to occur under pathways which exist for those compounds. Fumarate addition on the other hand has been characterized for TEX and alkanes.

This is the most well-studied mechanism of hydrocarbon activation under anaerobic conditions, and will be the focus of this work. This mechanism actually involves the addition of a hydrocarbon to fumarate, but is referred to in this thesis using the more common term ‘fumarate addition’.

To date, three fumarate addition enzymes (FAEs) have been characterized; alkylsuccinate synthase (ASS, catalyzing fumarate addition to n-alkanes), benzylsuccinate synthase (BSS, catalyzing fumarate addition to alkylated mono-aromatics), and naphthylmethylsuccinate (NMS, catalyzing fumarate addition to alkylated polyaromatics)30–32. Unlike their namesake of fae, referring to fairies, these enzymes do not belong to the mythical realm and their activity is well documented and studied. The ASS and BSS enzymes belong to the glycyl radical enzyme (GRE) family, which have a characteristic radical at a highly conserved glycine residue. The radical forms from the activity of another activating enzyme called glycyl radical activating enzymes

(GRE-AE). The radical on the GRE acts to abstract a hydrogen atom from the substrate molecule

(in this case hydrocarbons), forming a radical and allowing electron transfer between the substrate hydrocarbon and fumarate, regenerating the glycine radical in the process33,34. These enzymes are irreversibly inactivated with exposure to oxygen (however a protein YfiD present in

Escherichia coli has been shown to reverse this inactivation in the related GRE, pyruvate formate lyase)34,35. It has also been reported that the rate limiting step of these enzymes is the fumarate addition reaction itself, and is higher for BSS due to the loss of resonance

8 stabilization36. The activity of ASS and BSS are the focus of this thesis (as polyaromatic compounds are not used in this study).

The ASS enzyme catalyzes the addition of fumarate across its double bond to the C2 carbon of the alkane chain30, while the BSS enzyme attacks the terminal carbon of a methylated monoaromatics (toluene, o-, m-, and p- xylene) or the beta carbon of ethylated monoaromatics

(ethylbenzene). The fumarate addition reaction is summarized in Figure 1. Further degradation of the succinate molecules leads to the production of benzylsuccinyl-CoA and alkylsuccinyl-CoA.

Benzylsuccinyl-CoA is then further degraded via a modified beta-oxidation pathway to form benzoyl-CoA, with an eventual ring-opening step. Further degradation of the benzyl/alkylsuccinate-CoA structure of both toluene/xylene and alkane derived metabolites occurs via a modified beta-oxidation forming a benzoyl-CoA (toluene/xylenes) or an alkyl CoA- ester type compounds. Subsequent steps in the degradation of mono-aromatic signature metabolites are known to converge to produce benzoate37. Downstream degradation of benzoate is dependent on other microorganisms and other metabolic processes, including syntrophic methanogenic activity38. Further degradation of the alkyl-CoA product involves a proposed carbon skeleton rearrangement, followed by removal of a carboxyl moiety, leaving a fatty acid

39 which is then degraded by beta oxidation ultimately to CO2 (Figure 1) . The degradation pathway for ethylbenzene has been proposed to follow a similar degradation pathway not catalyzed by BSS and produce alkyl-moieties more similar to alkane degradation than toluene degradation40. An in-depth overview of the anaerobic degradation of aromatics has been reviewed by Gieg and Toth37, and an in-depth review of anaerobic n-alkane degradation is given by Agrawal and Gieg (2013)39. These anaerobic degradation mechanisms are relatively well

9 studied, with key steps and metabolites identified. Thus, the anaerobic degradation activity

(metabolites and genes) can be detected and used as a method to assess ongoing activity.

Figure 1: Generalized schematic of fumarate addition to octane and toluene, and subsequent metabolic pathways of anaerobic HC biodegradation.

1.5 Diagnostic Hydrocarbon Metabolites to Indicate In Situ Anaerobic Bioremediation

The use of fumarate addition products, like benzylsuccinate, as diagnostic markers for anaerobic hydrocarbon degradation has been demonstrated in numerous field studies30,41,42.

Diagnosing hydrocarbon degradation using these succinate molecules is possible for 4 main reasons: first, the products of fumarate additions are unique to anaerobic degradation; second, they are not produced by other industrial means (eliminating false positives); third, they do not accumulate in the environment (they are produced transiently and consumed rapidly) and can be stabilized through sample acidification to be detected using gas chromatography-mass spectrometry (GC-MS); and fourth the metabolites have an explicit relationship to the parent hydrocarbon31. Thus, these compounds are also known colloquially as ‘signature metabolites.’

The presence of the initial activation molecule, alkylsuccinates (for alkanes) or benzylsuccinates

(for toluene), or 2-, 3-, or 4-methylbenzylsuccinate (for o-, m-, and p- xylene) have been shown to be diagnostic indicators for the in situ activity of anaerobic hydrocarbon degradation42,43.

The transient nature of these metabolites may make reliable detection difficult as they are produced and readily consumed thereafter. However, the main difficulties are the low concentrations of the produced metabolites (often present in nanomolar concentrations) and the limited detection capabilities of GC-MS and LC-MS (liquid chromatography-mass spectrometry) instruments to detect these compounds, in addition to the difficulties in accurately identifying signature metabolites during analyses27,39,44,45. Limitations in instrumentation and the antecedent knowledge and direction needed to accurately identify the class of metabolites can hinder accurate diagnosis. Hydrocarbon analysis showing the presence of, for example toluene, does not necessitate its signature metabolite, benzylsuccinate, will be detected, and vice versa. Previous studies have shown that although an explicit link between precursor and metabolite is present,

12 the physical detection of either is not guaranteed in field studies41,44,45. Further degradation of the metabolites may be actively occurring as metabolite flux, thus the metabolites would not be detected in this instance. The lack of detection of these signature metabolites does not demonstrate that anaerobic degradation activity is not occurring, thus their detection is useful in having another level of evidence to support that activity. Also, production of these signature metabolites does not necessitate the further and compete degradation to CO2 by the producing microorganism. For example, fumarate addition to p-xylene to 4-methylbenzylsuccinate can be carried out by Thauera aromatica, but is not capable of degrading it further, but the

Alphaproteobacteria Magnetospirillum has a benzoyl-CoA reductase which can use 4- methylbenzylsuccinate as a substrate16. As another example, ASS has been shown to catalyze fumarate addition to aromatic hydrocarbons as well, under non-energy producing mechanisms where the formation of the succinate intermediate and subsequent formation of the benzoyl-CoA is a dead-end product. In this case certain ASS encoding organisms do not have the required mechanisms for further growth supporting catalysis16. Differences in microbial community composition therefore have an influence on the metabolite detection, where communities lacking benzoate degraders may have readily detectable metabolites, while communities with that capability may have none. Thus, the need for further evidence to support metabolite data (or used in lieu of the lack of metabolite data) is needed and has been the focus of recent biomarker studies46,47.

Detection of the products/genes of other activation mechanisms (carboxylation, hydroxylation, methylation) can be detected using the same methods as fumarate addition (GC-

MS, genetic assays etc.). However, many of these compounds do not fulfill the 4 rules of being

’signature’, as they have commercial origins, do not have an explicit relation to the parent

13 hydrocarbon (can be produced from many hydrocarbons), or are not unique to hydrocarbon degradation. Hence, they are less studied and give less conclusive evidence of anaerobic degradation activity. Phenols (hydroxylated benzene), toluene (methylated benzene), and benzoate (carboxylated benzene) are common constituents of gasoline and intermediates in non- fumarate addition reactions related to toluene degradation. The detection of these compounds gives limited information, primarily on the composition of the hydrocarbon contamination, rather than degradation activity. Thus, detection of truly signature metabolites, as previously described, paint a clearer picture of whether anaerobic hydrocarbon degradation is actually occurring in situ.

1.6 Anaerobic HC Biodegradation Genes – Advantages and Limitations

Investigating contaminated environments for the presence of the genes encoding ASS and

BSS has also been used successfully as a biomarker for anaerobic hydrocarbon degradation potential via fumarate addition25,27,45–50. Studies, including this one, focus on the detection of the assA and bssA genes as biomarkers for degradation potential. These assays only measure the potential for fumarate addition as the mere presence of a gene does not necessitate its expression or the activity of the translated protein. However, combining the presence of the FAE genes with signature metabolite detection can provide more solid evidence that in situ anaerobic HC degradation is or has occurred at a field site.

Focus of primer design has been on a conserved region encoding the catalytically active alpha subunits of both ASS and BSS enzymes, assA and bssA23,24. Effective capture of the diversity of assA and bssA has plagued gene detection studies where 16S rRNA gene sequencing inconsistently detect known alkane or aromatic degraders, despite detection of the FAE genes, or

14 vice versa (FAE genes are not detected despite identification of known hydrocarbon degrading taxa)50,51. A similar issue is linked with the detection of the FAE genes and the signature metabolites that their translated proteins produce. Detection of the assA or bssA gene may occur, but the produced metabolite may have been transported away from the sampling area through groundwater movement. To combat this issue of missing diversity, numerous studies have attempted to add to the known diversity of FAE genes through design and modification of published primers. Initially, primer sets were designed specific to families of microorganisms described as mono-aromatic or alkane degraders, focusing on sulfate and nitrate reducing organisms (or syntrophic Smithella spp. in the case of assA) as these organisms are the most prevalent/enriched in many microbial cultures24,47,51,52. Callaghan et al. (2010) designed widely popular primers for the detection of assA, a group of 9 primers targeting different regions and groups of the assA gene (with a primer pair targeting bssA)47. A drawback to this is the inefficient use of reagents to individually probe a sample for species-specific assA or bssA sequences. In contrast, using a single primer pair from a group loses potential diversity.

Degenerate primers were designed by von Netzer et al which had the capability of detecting assA and bssA sequences from different microbial groups46. These primers, FAE-B, FAE-KM, and

FAE-N, were designed to target the assA and bssA genes in a strict or general sense, with the additional detection of nmsA genes using the general sense primers (FAE-B)46. Again, these primers were designed as a test to determine if degenerate primers could capture the diversity of assA and bssA present in various environmental samples. As was demonstrated by von Netzer et al. (2013)46, the comparison between primers designed by Winderl et al. (2007)24 (7772f/85646r) and their own (which were designed based off 7772f/8546r) revealed that the coverage across known species tested did not overlap. The Winderl et al.24 primers detected bssA heavily among

15 the Proteobacteria but not in Clostridia, while the von Netzer et al.46 primers detected genes in the Clostridia, but not the Proteobacteria. No single primer pair designed to date has captured assA or bssA diversity confidently (according to their designers), of known and putative anaerobic hydrocarbon degraders. Although these designed primers also came in different variations specific for clusters of assA genes (and modified from primers designed by Winderl et al.24) they increased the PCR assay efficiency from upwards of 9 primers down to 2 or 324,46,47 .

The design of more universally applicable FAE gene primers is frustrated by the diversity among these enzymes, resulting from the diverse taxa and environments in which they are found.

To address the diversity of the assA sequence, an in-depth analysis of the diversity of the assA sequences in different sampling locations globally, and different types of naturally occurring hydrocarbon seeps (methane, natural gas, and hydrocarbon) was performed by Stagars et al (2016)48. They showed that the diversity of the assA sequence was dependent on the substrate hydrocarbons present at the respective environmental sites. Methane seep assA sequences were divergent from gas (in contrast to methane gas by also containing ethane, butane, propane, etc.) and hydrocarbon seep assA, which the authors had predicted, as the methane seeps have limited HC species (predominantly methane) while gas and HC seeps vary substantially

48 (upwards to C18 and beyond) . These authors argued that despite the analysis of assA sequences across the globe, the capture of all the possible diversity was incomplete. With only a handful of characterized alkane degraders identified to a specific genus, and a multitude of uncultured alkane degraders as found in genetic databases, the capture of all assA will be an ongoing challenge. Indeed, the authors had found clusters of assA present in their study samples that had no cultured representatives, a finding echoed by all newly designed primer studies.

Owing to the challenge of gene capture are the differences in enzyme motifs, where communities of alkane degraders have evolved to highly specific alkane substrates (e.g.

Azoarcus sp. HxN1 is specific for C6 - C8 alkanes, while Desulfatibacillum aliphaticivorans str.

53,54 CV2803 is specific for C13 – C18 alkanes) . Thus, depending on the type of contamination present, the hydrocarbon degrading community may be tailored, enzymatically, and may be divergent from other hydrocarbon degrading communities. Because the primary focus of primers is on the subunit encoding the active site, this change in substrate specificity may be exacerbating this issue. A similar analysis assaying the global diversity of bssA genes based on available substrates and the variance in sequence therein has not been done to the author’s knowledge.

Despite the limitations in diversity capture of FAE gene primers, these primers still accurately identify FAE genes. Numerous studies have characterized fuel contaminated sites for bioremediation potential through FAE gene detection25,26,45,46,49,50,55–57. However, comparisons with other primer sets in interrogating the same environmental sample have not always overlapped in detection, as was demonstrated by von Netzer et al. (2013)46. Indeed, nearly all sites studied have had the assA and bssA genes present using some primer sets, while detection of the signature metabolites (or metabolites signifying the other degradation pathways) have not always been seen27,44,45. While the limitation in ‘true’ FAE gene capture is a lingering issue, the application of these primer sets to determine potential activity seems to be effective46,48,50,58,59.

However, while using this gene detection information as a diagnostic tool is and has been implemented2,25,50, the apparent gap in complete understanding of the genetic foundation of fumarate addition mechanisms gives an incomplete picture of the processes occurring in a hydrocarbon plume, and possibly inaccurate monitoring of bioremediation. As it stands, genetic

17 assays into the diversity of the assA and bssA genes to detect their presence and, by proxy the potential for their activity, is one of the best tools available to assay for bioremediation.

Detection of the FAE genes in a contaminated system gives valuable information for the potential of anaerobic bioremediation. However, correlations to the detection and activity of FAE harbouring organisms is completely missed. Hence, quantifying the FAE genes and investigating a correlation between the abundance of FAE genes over time related to hydrocarbon concentration changes would give an unprecedented view of bioremediation potential and potentially an unseen correlation between detection and activity in ‘real-time’. Owing to the limitation in diversity capture, few primer sets have been designed for the quantification of the assA or bssA genes. Those that have been designed have been done so with a limited detection scope, either specific for sulfate-reducing or nitrate-reducing strains or specific to substrate- specific enrichments20,27,46,52,60,61. Stagars et al. (2016)48 designed 4 fluorescence FISH probes which would allow for quantification of different clusters of assA sequences identified in their study, however the prevalence of non-specific probe annealing reinforced the difficulty in accurately detecting FAE genes. To assay the activity through genetic means, assessing the transcription of the Ass and Bss genes to mRNA would need to done. Oberding and Gieg

(2018)60 undertook RT-qPCR (reverse transcriptase quantitative PCR) to determine the activity and change in assA expression when exposed to paraffinic n-alkane octacosane (C28H58) as substrate. That study did show a significant increase in the expression of assA when exposed to the substrate (compared to controls)60. As was discussed, the assA genes present in an octacosane-degrading culture may have a differing sequence to that of, for example, an octane- degrading culture. An RNA-SIP study was done by Fowler et al. (2014) looking at the changes in bssA abundance over time in a toluene-degrading culture, and found no significant change over

18 the incubation period61. They too had expressed concern of limited primer capture in accurately quantifying the bssA gene. The primer set used was tailored for Desulfosporosinus which was found in culture, but may not have been the only bssA present/active in degradation or in the microbial community61. To the author’s knowledge, these were two out of a handful of studies looking at expression changes of a FAE gene.

Quantitative PCR (qPCR) uses DNA intercalating dyes (SYBR Green) to quantify the fluorescence of SYBR Green-bound double-stranded DNA (dsDNA) during the PCR reaction.

As amplicon numbers increase with each progressing cycle, measurements of the increasing fluorescence are taken. These measurements can then be used to back calculate the starting quantity of the target gene in a DNA sample. Generally, primers producing shorter amplicons are favoured in qPCR reactions, where longer amplicons decrease the reaction efficiency (expend more dNTPs, fluorescence increases dependent on length rather than abundance, etc.)62. Widely used qPCR bssA primers have been designed exclusively from Betaproteobacterial bssA sequences, namely, Thauera aromatica strains K172 and T1, and Azoarcus sp. strain T51. As such the sequences from Alphaproteobacteria, Clostridia, and Deltaproteobacteria may not be directly assessed, a limitation acknowledged by the authors51. In that inaugural study by Beller et al (2002)51 the use of bssA quantification as a measure of activity was not reflected across all microcosm setups. Addition of ethanol as a degradation stimulant yielded bssA abundances that did not match the toluene degradation observed. This was attributed to limited knowledge of possible bssA, a decreased sensitivity in detection, or a different degradation pathway51. In contrast, a study done by Kazy et al (2010) using the same primers on microcosm studies saw that bssA abundance was correlated to toluene degradation under denitrifying conditions, up to

108 copies/g58. Similar abundances were also seen in a field site biostimulation study done by

Müller et al (2017) using the same qPCR primers under sulfate- and iron-reducing conditions57.

The differences in bssA abundance results in three different studies using the same primers attests to the variability of bssA genes and possible tailoring of enzymes to substrate contaminants. It may be that the primers were better suited for the Kazy et al58 and Müller et al57 sites/cultures than they were from Beller’s study site51, despite the fact that both Kazy et al58 and

Beller et al51 prepared microcosms under nitrate-reducing conditions. Interestingly, Müller et al. field sites57 were under sulfate- and iron-reducing conditions, but used primers designed for denitrifying bssA strains.

Generally, in contaminated aquifers, the gene abundance of bssA genes has been found to be between 103 to 108 copies/g sediment or copies/L water 20,26,57,58,63. These gene abundances were evaluated using primers specifically designed to target bssA sequences from a single cluster, e.g. sulfate reducers (Deltaproteobacteria and Clostridia) or nitrate reducers

(Betaproteobacteria). The large difference in the abundances of the bssA gene, ranging across 5 orders of magnitude, is dependent upon experimental differences, such as the primers used, the type of sample, toluene degradation activity, and the extent to which Bss-harbouring microorganisms were enriched. Studies relating the abundance of bssA to 16S rRNA gene abundances found in some cases that while the 16S rRNA gene abundances were relatively consistent, the abundances of the bssA genes varied across orders of magnitude51,57. This issue of variable quantification of FAE genes has been grandfathered over time with frequent use of demonstrably limited primers. This limitation on the detection of the FAE genes in an environmental sample is part of the struggle and the heterogenic nature of these studies.

Although the limitation in known and captured diversity of FAE genes is an issue, it must be stressed that the detection of the FAE genes, like signature metabolites, is a sound and tested

20 method to determine bioremediation potential20,57,58,63. Primer sets discussed above have their limitations, however they have detected (to varying extents) FAE genes in environmental samples. The main issue, which may not have been clear, is the frustration in developing universal assA and bssA primers capable of capturing greater FAE gene diversity.

1.7 Use of Microcosms to Study Hydrocarbon Biodegradation

As has been discussed in detail above, the diverse nature of substrates, environments, and the FAE harbouring organisms themselves confound the development of a universal assay for hydrocarbon-degradation activity detection. The use of molecular methods in assaying environmental samples has its own difficulty, as has been described above, with the nature of groundwater dynamics. Apparent reductions in hydrocarbon concentrations at field sites may suggest that degradation has occurred, when in reality the plume has simply travelled away from

(e.g., via dilution) the sampling well. Khan et al (2004) notes that laboratory testing of contaminated sites to assess both the potential for bioremediation and to monitor activity is required1. Microcosm based studies where environmental samples are tested under a litany of experimental conditions to better assess the most viable remediation conditions are extremely valuable. Experimentation to determine the expected rates at which degradation may occur can be gleaned from this method. By studying biodegradation in laboratory conditions, knowledge about the FAE genes, signature metabolites, degradation pathways, the organisms responsible and isolation of those organisms can be done, with the aim to better understand biodegradation and better assess it in natural environments.

The advantages of using this laboratory, microcosm-based assays are the in-depth analysis of the activity and the luxury of accurate monitoring of contaminant loss over time, which would otherwise be impossible to do using in situ environmental assays. Disadvantages

21 are that the setup of these experiments is highly variable from differing conditions between in situ and ex-situ environments. Nevertheless, microcosm studies provide valuable insight to the potential for hydrocarbon biodegradation of a contaminated site.

From laboratory microcosm studies, well studied alkane-degrading species have been isolated and/or characterized, such as Desulfatibacillum alkenivorans PV2803, Desulfatibacillum aliphaticivorans CV2803, and Desulfoglaeba alkenexedens AK01. Initial studies into alkane degradation and elucidation of genes and metabolic pathways were found using these isolates30,54,64,65. Alkane degraders are primarily members of the Deltaproteobacteria and those characterized are sulfate reducers or syntrophic species. Primarily, assA sequences fall into the

Deltaproteobacteria class, with some known alkane degraders falling into Archaea and

Clostridial clades, which suggests horizontal gene transfer may have occurred66. Other cultures of alkane degraders can be present under methanogenic conditions, such as syntrophic Smithella spp.60,67. BSS harbouring microorganisms are more diverse, residing across the Proteobacteria

(Alpha, Beta, and Delta) and Firmicutes. Well studied Bss harbouring taxa include but are not limited to, Thauera aromatica, Geobacter metallireducens, Desulfobacula toluolica, Azoarcus spp., and Georgfuchsia toluolica13,23,68. To our knowledge, no characterized species isolate undergoes toluene (or methylated mono-aromatics) degradation under methanogenic conditions, however many uncharacterized cultures have been described61,69,70. Knowledge of these isolates from microcosm enrichment studies allowed for the continued characterization of the fumarate addition or other metabolic pathways.

The detection of such hydrocarbon-degrading organisms in field samples has not been found to be consistent. Generally, the presence of these organisms is limited to the detection of the class or order to which they belong to (i.e. Deltaproteobacteria rather than

Desulfatibacillum). Therefore, conclusions towards the presence of known hydrocarbon degraders from these data such as 16S rRNA gene sequencing, are limited and not cannot be taken as diagnostic. Although profiling the microbial community using 16S rRNA gene sequencing is a tested microbiological method, in this application the information that it gives cannot be taken as concrete evidence for biodegradation potential. In contrast, the detection of metabolites and genes are as a result of the direct activity and potential, respectively, and reaction of the tested process itself, while profiling the microbial community gives information on the types of taxa enriched when challenged with hydrocarbons.

In this thesis work, a combination of diagnostic processes was done to characterize the potential for bioremediation at two hydrocarbon-contaminated sites. Approaches of signature metabolite detection (for plausible confirmation of anaerobic degradation activity), assaying for the FAE genes (to evaluate the genetic potential for anaerobic hydrocarbon degradation), assaying the microbial community composition (to identify if any standout taxa seem to be correlated with degradation activity), FAE gene quantifications (in attempt to correlate the degradation activity with elevated FAE gene abundance), and finally evaluating the degradation activity using microcosms (monitoring the degradation activity in real-time) were combined.

Using all these methods will give a multifaceted view of degradation potential and possible activity at hydrocarbon-contaminated sites. Further, an attempt at designing broad-range primers compatible with qPCR to detect and quantify assA and bssA genes in multiple hydrocarbon- degrading microorganisms in a single assay was undertaken. Owing to the difficulties in assessing in situ hydrocarbon degradation as has been discussed above, a multifaceted approach is necessary.

Chapter 2: Research Objectives and Goals

As was discussed in the previous section, a multifaceted approach into determining the potential for bioremediation at hydrocarbon contaminated sites is essential towards gaining a better understanding of this process. In this M.Sc. thesis research, I used a multifaceted approach

– analysis of signature metabolites, electron acceptor monitoring, microbial community analysis, and FAE gene detection/quantification, in order to better understand anaerobic HC degradation in fuel-contaminated aquifers. A microcosm study monitoring HC-degradation activity under different electron accepting conditions was also performed. This research was broken into 3 main

Objectives, with specific goals for each:

1. Assessment of bioremediation potential of two hydrocarbon contaminated sites in

Alberta, Canada. This was done through analysis of signature metabolites to determine

whether anaerobic hydrocarbon degradation has occurred. This was linked with FAE

gene detection assays using literature primers, in order to determine if a co-occurrence

trend between signature metabolites and FAE genes was present. Lastly, microbial

community analysis was done using 16S rRNA gene sequencing, to determine whether a

trend was evident between detected taxa and signature metabolite detection. In combining

these data, with the presence of hydrocarbon degrading taxa, FAE genes, and the

products of fumarate addition, I aimed to ascertain whether these contaminated sites had

the potential for bioremediation. The results of this Objective are presented in Chapter 3.

2. Design and testing of qPCR compatible primers for capturing FAE genes using samples

from hydrocarbon contaminated sites, as described in Objective 1. An attempt to design

qPCR compatible primers with the objective of maximizing diversity capture was

undertaken. To achieve this, a novel mixture of forward primers specific for different

groups of FAE genes was designed for both assA and bssA genes, with a single reverse

primer for each primer mix. Each forward primer in the mix was designed such that they

bind the same region, producing identical length amplicons. This may be the first attempt

at such a primer mixture for the detection and quantification of FAE genes. Using these

primers, the environmental samples characterized through Objective 1 were interrogated

to determine their efficacy of quantification and detection. These were compared with

established primers (from available literature) to compare detection efficacy. FAE gene

abundances were compared with HC concentrations to determine whether a positive

correlation could be made. These amplicons were also sequenced to further verify their

amplification ability. The results of this Objective are presented in Chapter 4.

3. To determine the bioremediation potential of the environmental samples from Objective

1, microcosm studies were set up to determine hydrocarbon biodegradation activity of

native groundwater microbial communities. Samples were chosen to reflect higher and

lower contamination scenarios to assess whether differences in contaminant

concentration would influence biodegradation potential. These microcosms were

challenged with a known concentration of hydrocarbons and amended with different

electron acceptors separately and in combination to assess the response of the microbial

community. Quantification and detection of the FAE genes using newly designed primers

from Objective 2 were carried out, to further test their efficacy and see if a trend existed

between degradation activity and FAE gene abundance. These microcosm tests were

done to give an idealized view of hydrocarbon degradation potential of these

contaminated sites, and the results are presented in Chapter 5.

After all that work, I’d say this research was pretty successful. I hope you agree that it was too.

Enjoy.

Chapter 3: Analysis of Field Sites for Evidence of Anaerobic Biodegradation

3.1 Introduction

Monitoring the potential for bioremediation of hydrocarbon (HC) contaminants in groundwater systems has been explored in many studies under both aerobic and anaerobic conditions, as has been summarized in recent reviews16,21,39. Hydrocarbon-degrading organisms are frequently described based on the hydrocarbons that they biodegrade, e.g. alkane degraders

(Desulfoglaeba alkenexedens, Desulfatibacillum aliphaticivorans, Smithella SCADC, etc.) or aromatic HC degraders (Thauera aromatica, Georgfuchsia toluolica, Azoarcus sp etc.).

However, mere detection of these organisms does not guarantee they are actually biodegrading in situ, as these organisms may be dormant or using other carbon sources. Thus, further studies into assessing in situ HC degradation activity or the potential for activity are needed, and can be based in detecting biodegradation genes or the diagnostic metabolites. Under anaerobic conditions, genes that encode for fumarate addition reactions are good candidates for detecting potential anaerobic biodegradation capacity in the field. If the genes encoding the enzymes responsible for this enzymatic reaction are not detected, hydrocarbon degradation activity cannot be expected, but if they are, then biodegradation may occur. To date, three enzymes involved in fumarate addition (fumarate addition enzymes; FAE) have been characterized, based on the substrate hydrocarbon; ASS/MAS, alkylsuccinate synthase/methyl alkylsuccinate synthase, which catalyzes the addition of fumarate to alkanes; BSS, benzylsuccinate synthase, which catalyzes the addition of fumarate to (m)ethylated monoaromatics; and NMS, naphthyl methylsuccinate synthase, that adds fumarate to the methyl group of polyaromatic rings.

Detection of the genes encoding these enzymes is focused on the alpha subunit, the active site of the enzymes. Should these genes be detected in the environmental samples, the potential for

27 hydrocarbon degradation exists. However, as was discussed for the presence of organisms, mere presence of these genes in situ does not necessarily denote activity. Thus, detection also of the products of fumarate addition gives even stronger evidence that the anaerobic hydrocarbon- degrading organisms present are active, have the necessary genes, and are actively producing metabolites characteristic of hydrocarbon degradation. Fortunately, such unique metabolites have been characterized, and their detection is diagnostic of anaerobic hydrocarbon degradation41.

These metabolites, dubbed ‘signature metabolites’ are succinates produced directly from fumarate addition to hydrocarbons, through the activity of the enzymes mentioned above. These succinates are not produced under any other process (natural or industrial) and are transient (and thus do not accumulate). Detection of these metabolites along with associated genes and known

HC-degrading taxa offers strong evidence of ongoing anaerobic hydrocarbon degradation.

Here, we assessed the bioremediation potential of 2 hydrocarbon-contaminated field sites in Alberta, Canada, designated Site A and Site B. Both sites are historical gasoline and diesel service stations which over time had leaked hydrocarbons from degrading storage tanks.

Assessment of whether these contaminated aquifers have the capability of remediating themselves via hydrocarbon metabolism by the native microbial community was done by analyzing groundwater for the presence of anaerobic biodegradation genes and diagnostic metabolites. Assessing the microbial community composition of these contaminated sites was also done to see if known hydrocarbon degraders were present, and if certain taxa were more prevalent in wells with higher contamination. We hypothesized that genetic and metabolic markers for anaerobic hydrocarbon degradation would be detected in both Site A and Site B, which would indicate that bioremediation is feasible and ongoing process.

3.2 Methods:

3.2.1 Site Description.

3.2.1.1 Site A.

Located approximately 130 kms south east of Calgary, Site A is an inactive gasoline and diesel fuel distribution site. This site was active as a bulk fuel distribution site until 2003 (start time is unknown). In 1992, a leak from an Aboveground Storage Tank (AST) was discovered.

Landfarming was undertaken in 1995, followed by in situ bioremediation efforts in 2007

(reported as biostimulation). Specifics on protocols were not reported and resulting remediation efforts were inconclusive. All infrastructure has been removed. Geology of this site is silty clay to clayey silt, with the groundwater depth at ~1 m below ground surface. Figure 2 shows a map of Site A, along with measured hydrocarbon concentrations and potential electron acceptors. The hydrocarbon plume is travelling east of the site, as marked with arrows (Figure 2).

3.2.1.2 Site B.

Located approximately 350 kms north west of Calgary, Site B is a decommissioned fuel distribution site which operated from the 1950s to the 1990s. Hydrocarbon contamination

(specifically gasoline) occurred from an AST located in the centre of the site, ~30 years prior to this report. A second Underground Storage Tank (UST) cluster was located in the northeast corner. Hydrocarbon release has also been reported from this storage unit, although further details are not available. Thus, two hydrocarbon plumes are present at this site. Geology of the site is silty to silty clay, with groundwater depth at ~5 m below ground surface. Numerous, undisclosed remediation attempts since 2000 have been made at this site. Figure 3 shows a map of Site B, along with measured hydrocarbon concentrations and potential electron acceptors. The hydrocarbon plume is travelling eastwards towards a developed area marked ‘Condo Property’ and ‘RV Lot’ (Figure 3).

3.2.2 Sample collection.

Samples were collected twice from each sampling site; Sept 9, 2016 (Site A and Site B),

April 13, 2017 (Site B only), and July 14, 2017 (Site A only) by AGAT Laboratories personnel.

Well sampling was done by initially removing 6 L of well water to purge the wells. Two 1 L

32 volumes of water sample were then taken, completely filling 1 L glass jars to minimize air. One volume was left unchanged, while the second volume was acidified to pH 2 with hydrochloric acid (concentration unknown) using pre-measured acid stocks (individual ‘packs’ for each sample) by the sampling technician on-site. All sampling jars were placed on ice and transported to the University of Calgary in coolers. Sampling of a few wells from Site A was observed by the author. Upon arrival, all sample jars were kept at 4oC until processed.

3.2.3 Hydrocarbon and Electron Acceptor Analyses

Initial groundwater sample analysis was done by AGAT Laboratories on behalf of

Stantec Consulting. Hydrocarbon concentrations, potential electron acceptor concentrations

(nitrate, sulfate, iron, manganese), and pH analyses were provided. A re-analysis of electron acceptors (nitrate, sulfate, and iron (II)) was done to remain consistent with further experiments done in the lab and are the values presented in this study. Sampling for electron acceptor analysis was done by taking 500 μL of volume from concentrated biomass, re-centrifuging the sampled volume, and storing at -20oC until further use. Sulfate and nitrate concentrations were monitored by liquid chromatography on a Waters HPLC with 2 separate detectors (Waters 432

Conductivity Detector (for sulfate), Waters 2489 UV/Visible Detector (for nitrate)), Waters 1515

Isocratic HPLC Pump, and a Waters 2707 Autosampler) with a Waters IC-PAK Anion High

Capacity column (Waters, Milford, MA). The method for anion determination was set at a flow rate of 2.0 mL/min (2000 psi) with an acetonitrile buffer (240 mL acetonitrile, 40 mL butanol, 40

72 mL borate/gluconate buffer, 1680 mL Mili-Q H2O) mobile phase (20 min method length) .

Sample volume injected was 50 μL. Monitoring for soluble iron (II) was done using a modified spectrophotometric method from Lovely et al. (1987)71. Standards of known concentrations of all

33 electron acceptors were run with each analytical assay. Hydrocarbons, manganese, and pH were not re-measured in the lab.

3.2.4 Biomass Collection.

Non-acidified samples (1 L) were centrifuged at 10,000 rpm for 10 minutes using a

Beckman Coulter Avanti J-E series centrifuge unit. Collected pellets were then resuspended in approximately 10 mL of supernatant and stored at -80oC in 15 mL Falcon tubes.

3.2.5 DNA Extraction.

DNA from pellet biomass was extracted using the MP Bio FastDNA Extraction Kit for

Soil (MP Biomedicals, Solon, OH) as per the manufacturer’s instructions, using 1.6 mL of centrifuged sample and eluted in 75 µL of provided elution buffer. All extractions were carried out in duplicate using identical methods. Quantification of extracted gDNA (genomic DNA) was done using a Qubit 2.0 fluorometer (sample volume – 1 μL).

3.2.6 16S rRNA Gene Sequencing Sample Preparation and Analysis

The 16S rRNA gene was amplified from extracted gDNA. Microbial community composition was assessed using Illumina MiSeq sequencing of the V6-V8 hypervariable region.

Amplification was done using primers Illumina926f

(TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGAAACTYAAAKGAATTGRCGG) and

Illumina1392r

(GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGACGGGCGGTGTGTRC). These were chosen as they were well established primers used in our lab for HC field studies73.

Two different polymerases were used in amplicon preparation, Fermentas Taq and KAPA

Hi-Fi polymerases (Fermentas: ThermoFisher, Waltham, MA; KAPA: Roche, Basel,

Switzerland). Initially, Fermentas Taq was the preferred polymerase used for amplification in lab. KAPA Hi-Fi, a higher fidelity polymerase, was introduced later to aid in amplifying samples with low DNA concentration and/or samples with PCR inhibitors present. This polymerase switch allowed for an opportunity to compare the PCR results using different polymerases. First- round PCR reactions contained: 12.5 µL polymerase mastermix (Fermentas Taq or KAPA Hi-

Fi), 0.5 µL each forward and reverse primer (200 nM), 1, 2, 5, or 10 µL of DNA template, with remaining volume of PCR-grade water up to 25 μL final volume. Thermocycling conditions were as follows: (Fermentas Taq) 95oC for 3 min, 25 cycles of 95oC for 40 sec, 55oC for 2 min,

72oC for 1 min, followed by a 7 min elongation step at 72oC; (KAPA Hi Fi) 95oC for 3 min, 35 cycles of 98oC for 20 sec, 65oC for 15 sec, 72oC for 15 sec, followed by a 10 min elongation step at 72oC. Second round PCR reactions to attach Illumina NexteraXT indexes were done in replicates of 50 µL reactions using either Fermentas Taq polymerase or KAPA Hi-Fi polymerase as follows: 25 µL polymerase mastermix, 1 µL each forward and reverse indexing primer (200 nM), 10 µL of purified first round DNA amplicon, and 13 µL PCR-grade water. Thermocycling conditions were as follows: 95oC for 3 min, 7 cycles of 95oC for 30 sec, 55oC for 30 sec, 72oC for 30 sec, followed by a 5 min elongation step at 72oC. Purifications of the first and second round PCR amplicons were done using AMPure Magnetic beads (or a Qiagen PCR Purification kit) (AMPure: Beckman Coulter, Brea, CA; Qiagen, Hilden, Germany).

Samples prepared using different combinations of amplification and purification protocols, e.g., Fermentas/AMPure vs. KAPA/Qiagen was based on ease of preparations. For example, for some samples that could not be amplified using Fermentas Taq, KAPA and Qiagen

35 or AMPure was used for sample purification of <10 samples instead. Most 2016 samples were prepared using Taq/AMPure, while the remaining samples used KAPA Hi-Fi polymerase, with either AMPure or Qiagen purification kits. Samples were then normalized to 2 ng/µL and pooled into a sequencing library. Samples were loaded onto the Illumina MiSeq sequencing platform

(2x300) kit at the International Microbiome Centre (University of Calgary). Resulting sequencing reads were assembled using PEAR 0.9.6 (50 bp overlap, 350 bp truncation) and given taxonomic rank through MetaAmp Version 2.0 using the SILVA 132 database at a species 97% similarity cutoff74,75. Microbial community analyses were done using a 5% read abundance cut-off. This was done to limit the amount of less abundant reads, ultimately simplifying the analysis to the most highly abundant taxa.

3.2.7 assA and bssA PCR Presence/Absence Assay

Assaying field samples for the presence of the assA and bssA genes through PCR was done using literature primer sets and thermocycling conditions24,27,46,47. Primer sets used to interrogate all field samples were: assA = 7757af/8543r (von Netzer et al. (2013)); bssA = 7772f/8546r

(Winderl et al. (2007)), which were taken from literature24,46. These primers performed best from other literature primer sets tested in the lab (data not shown)27,46,47. The protocol for the assay was as follows: 12.5 µL of ThermoFisher 2x Fermentas PCR MasterMix, 10.5 µL PCR-grade

H2O, 0.5 µL each of forward and reverse primer (200 nM), and 1 µL of gDNA template. The identical assay was also done using KAPA Hi-Fi polymerase using the identical protocol.

Confirmation of amplification was done using gel electrophoresis (1% agarose) by comparing with a positive control of a known assA or bssA harbouring organism, D. alkenivorans AK-01

36 and Thauera aromatica, respectively. Imaging was done with Bio-Rad GelDoc Imager XR+

(running on Image Lab Software Suite).

3.3 Metabolite Extraction and Analysis

Acid preserved 1 L volumes of groundwater (~ pH 2) were extracted for organics using ethyl acetate as the organic solvent. For each 1 L of groundwater, 200 mL volume of ethyl acetate was used, split into 3 separate extraction repetitions. The ethyl acetate extractant was filtered through filter paper containing anhydrous sodium sulfate to collect residual water. The collected ethyl acetate layer was concentrated to approximately 100 µL using a KA RV-10 rotovap apparatus, heating the solvent to 60oC under vacuum. The organics present were then derivatized using N, O-Bis(trimethylsilyl)trifluoroacetamide (BSTFA), replacing the hydroxyl groups on organic acids with trimethylsilyl groups to aid in chromatographic analysis. These prepared samples were analyzed on an Agilent Technologies 7890A Gas Chromatograph equipped with a 5975C inert XL mass spectrometry detector (MSD) with triple-axis detector (GC-

MS). The injection inlet was held isothermally at 280°C. Components were separated on an HP-1 capillary column (19091Z-115E, 50 m × 320 μm × 0.52 μm film thickness, Agilent Technologies,

Inc.) using a 10:1 split ratio and helium as a carrier gas. The oven temperature was held at 100°C for

5 min before the temperature was increased at a rate of 5 °C/min to 300°C then held for 20 min. The mass spectrometer scan parameters were set between 50 and 700 mass units, with 2.28 scans/second.

The sample was auto-injected using 1 µL of sample. Resulting chromatographic peaks were then compared with standards prepared and run previously in-lab, under identical GC-MS analysis conditions.

3.4 Results and Discussion

3.4.1 Site A

3.4.1.1 Hydrocarbon concentrations and potential electron acceptors.

Table 1 shows the Alberta Guidelines as well as HC concentrations measured for all wells sampled, shaded values indicating HC concentrations above guidelines. The Alberta guidelines for the allowed concentration of hydrocarbons in non-drinking water groundwater are: benzene = 0.005 ppm, toluene = 0.024 ppm, ethylbenzene = 0.0016 ppm, xylenes (all isomers) =

0.02 ppm, and alkanes = 2.2 ppm76. Two wells showed the highest total hydrocarbon contamination at 10.416 ppm (C03-10) and 10.383 ppm (C02-08) with 8 wells showing below 5 ppm total hydrocarbons (C01-01, C01-04, C02-06 (0 ppm), C02-07 (0 ppm), C03-11, C03-12,

C03-13, MW07, and MW23).

C6 - C10 Total Benzene Toluene Ethylbenzene Xylenes Well Name Alkanes Hydrocarbon (ppm) (ppm) (ppm) (ppm) (ppm) (ppm) Guideline 0.005 0.0024 0.0016 0.02 2.2 C01-01 0.281 0.0014 <0.0005 0.001 0.2 0.5 C01-04 2.45 0.0153 0.0912 0.0081 0.5 3.1 C02-06 <0.0005 <0.0003 <0.0005 <0.0005 <0.1 <0.1 C02-07 <0.0005 <0.0003 <0.0005 <0.0005 <0.1 <0.1 C02-08 2.24 0.072 0.86 1.71 5.5 10.3 C03-10 7.95 0.128 1.26 0.178 0.9 10.4 C03-11 0.131 <0.0003 <0.0005 <0.0005 <0.1 0.1 C03-12 0.098 <0.0003 0.0018 0.0015 0.1 0.2 C03-13 0.0123 <0.0003 0.013 0.0102 0.4 0.4 C03-14 0.24 0.0976 1.8 0.0102 4.4 6.5 MW07 <0.0005 <0.0003 <0.0005 0.0102 <0.1 <0.1 MW23 <0.0005 <0.0003 <0.0005 0.0102 <0.1 <0.1 Trip Blank <0.0005 <0.0005 <0.0005 <0.0005 <0.1 <0.1

Figure 4 summarizes the HC concentration in wells at Site A in graphical form. Although a majority of wells have hydrocarbon concentrations lower than the Alberta Guidelines stipulate, the few wells that do have higher concentrations pose a risk and necessitate remediation.

However, since the sampling locations are limited, it cannot be inferred that the peak contamination at Site A resides around wells C03-10 and C02-08; higher contaminated areas may be present in the large, unsampled area labelled ‘WAREHOUSE’, ‘PUMPS’ and ‘GAS

PUMPS’ (Figure 2). Approximately 90 m separates the two highly contaminated wells C02-08 and C03-10, which are close in proximity to relatively uncontaminated wells. Well C02-07

‡ Benzene > 0.005 ppm. Toluene > 0.0024 ppm. Ethylbenzene > 0.0016 ppm. Xylenes > 0.02 ppm. C6-C10 Alkanes > 2.2 ppm.

(uncontaminated) and C03-13 (total HC, 0.4 ppm) surround C02-08 (total HC, 10.3 ppm), while

C01-04 (total 3.1 ppm) is approximately 15 m away from C03-10 (total HC, 10.4 ppm) (Figure

2). This suggests that the geology in this area is dictating the direction and flow of the hydrocarbon plume where closely located wells are not evenly contaminated.

Referring to the geographical location of the wells, wells C02-06 and C02-07 are upstream of the contamination plume and are considered to be background wells with 0 ppm HC.

The surrounding area has no background hydrocarbon contamination/leakage from another source. As summarized in Figure 2, the potential electron acceptors measured the groundwater samples shows sulfate to have the highest concentrations, suggesting that this can serve as an important EA (electron acceptor) at the site.

With the presence of these hydrocarbons, it was expected that the coinciding signature metabolites from ASS and BSS activity should be present if anaerobic hydrocarbon degrading organisms are present and active.

3.4.2 Site B

3.4.2.1 Hydrocarbon concentrations and potential electron acceptors.

As referenced in Section 3.4.1.1 (Site A hydrocarbon contamination and potential electron acceptors) the Alberta Guidelines of Hydrocarbon contamination of non-drinking water groundwater is easily surpassed when dealing with hydrocarbon spills in small locations, as summarized in Table 2, Figure 3 & Figure 5.

Site B is a more highly contaminated site than Site A, with maximum total hydrocarbon concentrations reaching 34.9 ppm (REC 34) with 7 wells above 10 ppm, (ISCO-3-C, ISCO-4-C,

ISO49, REC 11, REC 24, REC 31, and REC 34) (Figure 5). Of the 12 wells sampled in Site B,

41 only S14-49B was uncontaminated with 0 ppm total hydrocarbons. All wells surpassed the 0.005 ppm benzene guidelines, the 0.024 ppm toluene guideline, the 0.0016 ppm ethylbenzene guideline, and the 0.02 ppm xylenes guideline (except well ISCO-3-C at 0.0051 ppm). In 2016, all wells except ISCO-3-B and REC 26 were above the 2.2 ppm alkane guideline (Table 2).

C6 - C10 Total Benzene Toluene Ethylbenzene Xylenes Well Name Alkanes Hydrocarbon (ppm) (ppm) (ppm) (ppm) (ppm) (ppm) Guideline 0.005 0.0024 0.0016 0.02 2.2 ISCO-3-B 1.65 0.0713 0.108 0.0454 < 0.1 3.8 ISCO-3-C 9.88 0.557 0.0051 0.0051 3.1 23.9 ISCO-4-C 9.26 0.92 0.0323 0.0323 2.8 23.2 ISO 49 0.297 0.87 0.223 4.46 1.3 13.2 REC 11 0.771 0.732 0.0408 4.96 2.7 15.7 REC 12 2.35 0.0299 0.0489 0.0432 2.2 7.2 REC 24 8.39 0.377 0.297 0.675 2.8 22.2 REC 26 0.391 0.0523 0.0421 0.403 0.2 2.0 REC 31 7.89 0.867 0.394 0.979 5.5 25.7 REC 34 2.23 3.83 0.325 6.11 9.9 34.9 S14-49B <0.0005 <0.0003 <0.0005 <0.0005 < 0.1 <0.1 S14-7R 1.26 0.125 0.0365 0.0754 1.2 4.2 Trip Blank <0.0005 <0.0003 <0.0005 <0.0005 < 0.1 <0.1

The contamination plume is not as straightforward as it is in Site A, where uncontaminated wells were largely upstream of the groundwater flow. In Site B the uncontaminated well S14-49B is located within a few metres of REC 34 (total hydrocarbon: 34.9 ppm) and ISO49 (total hydrocarbon: 13.2 ppm) (Figure 3). Wells that are within a few metres of

§ Benzene > 0.005 ppm. Toluene > 0.0024 ppm. Ethylbenzene > 0.0016 ppm. Xylenes > 0.02 ppm. C6-C10 Alkanes > 2.2 ppm.

42 each other have vastly different hydrocarbon profiles. ISCO-3-C, ISCO-3-B, and ISCO-4-C are clustered within a few metres and have a 20 ppm difference in total hydrocarbons (ISCO-3-B =

3.8 ppm, ISCO-3-C = 23.9 ppm, and ISCO-4-C = 23.2 ppm). Wells REC 31 and REC 12 are also dissimilar despite close proximity, having an 18.6 ppm difference (REC 31 = 25.7 ppm and REC

12 = 7.2 ppm) in total hydrocarbon contamination. Furthermore, the hydrocarbon concentrations in ‘upstream’ wells, for example REC 11, (the east-most well) and assumed to be most removed from the contamination source) is the sixth highest contaminated well at 15.7 ppm total hydrocarbons. This non-linear distribution of hydrocarbons makes it difficult to predict trends on the potential biodegradation activity as the plume travels through the site.

Benzene, toluene, and alkanes are the largest contaminants in these sites, which poses a high health risk, but also may be inhibitory to microbial growth and activity. This would be most prevalent in the highest contaminated wells (Chu and Gieg, unpublished data). The situation in

Site B is more urgent due to the plume moving towards and into a condo property (marked

“CONDO PROPERTY”; Figure 3). Of the potential EA measured in lab, sulfate and iron (II)

2- were of highest concentrations, which suggests that they (SO4 and Fe (II)) may serve as potential EA at this site.

In summary, Site B is highly contaminated, containing hydrocarbon concentrations well above the Alberta Guidelines for non-drinking groundwater. Despite this, we may still expect to see diagnostic biodegradation genes and metabolites if HC-degrading communities are present and active.

Figure 5: Concentrations of the hydrocarbons benzene, toluene, ethylbenzene, and xylenes (BTEX) and C6 – C10 alkanes in each sampled well of Site B in 2016. Analysis was done by AGAT Laboratories.

3.4.3 Metabolic and gene evidence for in situ HC biodegradation.

3.4.3.1 Site A signature metabolites and FAE gene presence.

3.4.3.1.1 Aromatic hydrocarbons, benzylsuccinates, and bssA.

Table 3 summarizes the metabolites and genes detected from Site A for both the 2016 and 2017 sampling years. For aromatic derived signature metabolites in Site A, only the ortho

(2016 C02-08, and 2016 and 2017 C03-14) and meta (2017 C03-14) isomers of xylene were detected, namely o-methylbenzylsuccinate and m-methylbenzylsuccinate, respectively77,37. This suggests that these 2 xylene isomers were being biodegraded at the site. The activity of benzylsuccinate synthase, responsible for the anoxic activation of toluene and other methylated

44 aromatics, like xylene, has been implicated in co-metabolism of non-toluene methylated aromatics42. It is possible that since benzylsuccinate (signature metabolite of toluene) was not detected, it may have been metabolized further into benzoyl-CoA, while the degradation of xylene by co-metabolic reactions activity may be slower, resulting in metabolite detection. The absence of detection of o-methylbenzylsuccinate in 2017 suggests that this metabolite has either been degraded, its precursor o-xylene was not present, or the metabolite was below detectable limits.

The point of contamination is located near well C03-14 (Figure 2), and as such more metabolites than those derived only from o- and m- xylenes were expected. Toluene concentrations in this well were almost 9-fold greater than xylene, and yet only xylene-derived metabolites were detected. In well C02-08, which has a higher concentration of xylenes, only xylene-derived metabolites were detected in 2016. Again, this may be due to the dynamic nature of both metabolites and of groundwater flow, where metabolites are rapidly consumed, and activation of hydrocarbons may be much slower. These findings suggest that the presence of a precursor hydrocarbon does not necessitate detection of the resulting metabolite in environmental systems. As such, the abundance of the ‘parent’ compound of a signature metabolite would be in higher abundance than the produced ‘daughter’ signature metabolite. As the signature metabolite may be further degraded, it’s accumulation and subsequent detection would be less likely. However, at this site, metabolites were detected only in the highest contaminated wells, which may suggest activity is slower in more contaminated wells or are present at concentrations too low for detection in lower contaminated wells. The lack of consistent detection of the signature metabolite and the parent compound was also noted by

Jobelius et al.44. These conclusions have been found throughout the signature metabolite detection assays presented in this work.

SITE A C02-08 C03-14 C03-10 C01-04 C03-13 C01-01 C03-12 C02-06 C02-07 C03-11 MW 07 MW23 TRIP BLANK Aromatic Hydrocarbons Toluene (ppm) 0.072 0.098 0.128 0.015 < 0.0003 0.001 < 0.0003 < 0.0003 < 0.0003 < 0.0003 < 0.0003 < 0.0003 <0.0003 Ethylbenzene (ppm) 0.861 1.800 1.260 0.091 0.013 < 0.0005 0.002 < 0.0005 < 0.0005 < 0.0005 < 0.0005 < 0.0005 <0.0005 Xylenes (ppm) 1.710 0.010 0.178 0.008 0.010 0.001 0.002 < 0.0005 < 0.0005 < 0.0005 0.010 0.010 <0.0005 Alkylbenzylsuccinates Benzylsuccinate Ethylbenzylsuccinate o -, m -, p - 2017/2016 2017/2016 methylbenzylsuccinate Gene Presence bssA 2016 (Taq) bssA bssA bssA bssA 2017 (Taq) bssA 2016 (KAPA) bssA bssA bssA bssA bssA bssA bssA bssA bssA bssA 2017 (KAPA) bssA bssA bssA bssA bssA bssA bssA Alkanes C6 - C10 Alkanes (ppm) 5.5 4.4 0.9 0.5 0.4 0.2 0.1 0 0 0 0 0 0 Alkylsuccinates C5 2017/2016 2017/2016 2016 C5 w/ unsaturation 2016 2016 C6 C6 w/ unsaturation 2016 2016 2017/2016 2017/2016 2016 2016 2017/2016 C7 2017 2017/2016 2017/2016 C7 w/ unsaturation 2017/2016 2017/2016 2017/2016 2016 2016 2017/2016 C8 2017/2016 2017/2016 2017/2016 C8 w/ unsaturation 2016 2016 2017/2016 2016 2017/2016 2016 C9 2016 2017/2016 C9 w/ unsaturation 2016 2017 2017/2016 2017/2016 2017 Gene Presence assA 2016 (Taq) assA assA assA assA assA assA assA 2017 (Taq) assA assA 2016 (KAPA) assA assA assA assA assA assA 2017 (KAPA) assA assA assA assA assA

In the context of this study, a ‘correlation’ is defined as the case where both the diagnostic metabolite and gene were detected in the same groundwater well sample. Correlations between detecting the benzylsuccinate metabolites and the bssA gene are not apparent at Site A.

Assaying bssA presence using different polymerases (Taq vs KAPA) suggests that the polymerase used can substantially affect the detection of target genes (this will be discussed in more detail in Chapter 4). Using Taq, bssA was detected in only 3 wells in 2016 (C02-08, C03-

10, and C01-04). But using KAPA, 10 wells had detectable bssA present (3/10 wells detected in

2016 (C01-01, C02-07 (uncontaminated), and MW23). Even with the higher fidelity KAPA polymerase, the detection of the bssA linked to produced aromatic signature metabolites does not improve. That being said, the produced signature metabolites are transient and thus actively degraded, whereas the microorganisms (and thus their genetic material) are more likely to be persistent, so it is much more likely that the genes are an artifact of previous degradation activity in lower contaminated wells. Even then, the presence of the gene is a strong indicator for the potential for fumarate addition. The movement of groundwater may be influencing the distribution of the genes and the metabolites, wherein more active biodegradation may not be where samples were collected.

3.4.3.1.2 Alkanes, alkylsuccinates, and assA.

Table 3 also summarizes the detection of the alkylsuccinate signature metabolites and the assA gene. All wells with an alkane concentration higher than 0.1 ppm had detectable signature metabolites present in both 2016 and 2017. The 3 most common alkane lengths with detected metabolites were (1) unsaturated C6 alkanes (detected in 6 wells in 2016, 2 wells in 2017), (2) unsaturated C7 detected in 2016 in 5, and in 2017 in 3 wells, and (3) unsaturated C8 alkylsuccinates detected in 2016 in 6 wells, and detected in 2 wells in 2017. Other detected metabolites were present in less than 5 wells, from C5 to C9 (saturated and unsaturated), excluding those mentioned above. Unsaturated alkanes refer to those that are non-linear, for example cyclic alkanes or branched cyclic alkanes. As was mentioned with benzylsuccinate metabolites, the appearance and disappearance of a single metabolite type from one year to the next is an indication of either on-going degradation activity, or due to the movement of groundwater removing metabolites from sampled groundwater areas.

The presence of the assA gene in these sampled wells showed higher co-occurrence with metabolite presence (than was seen with benzylsuccinates and bssA presence). In 2016, the assA gene was detected in all wells with alkane concentrations above 0.1 ppm (with the exception of well C01-04 (0.5 ppm))**. This was also seen in 2017, with the exception of 2 wells, C03-10 (0.9 ppm alkanes) and C01-01 (0.2 ppm alkanes). Detection of assA across sampling years was more consistent using KAPA compared to Taq polymerase, where KAPA detected assA in 5 wells in

2017, compared to detection in only 1 well using Taq polymerase. However, almost all wells where metabolites were detected also saw the detection of assA, independent of the sampling year. Overall, there is some correlation of assA and associated metabolite detection across sampling years. As mentioned for bssA/benzylsuccinates, the presence of the assA gene and detection of the associated metabolites is a strong indicator for anaerobic alkane degradation potential in Site A.

Interestingly, there is no overlap in metabolite detection for alkylsuccinates and benzylsuccinates, with only 2 wells showing both classes of metabolites (C02-08 and C03-14, the highest contaminated wells). It is most likely, as was mentioned above, that metabolites were metabolized and thus not detected consistently across all wells. Overlap of the assA and bssA genes is more consistent, where 7 of 8 HC > 0 ppm wells revealed the presence of both assA and bssA genes (using KAPA polymerase). Only 2 of 5 wells with detected bssA also had bssA detected using Taq polymerase. This suggests that the alkane and aromatic degraders were both present in the groundwater and their activity is likely a consequence of the HC concentrations present.

** well C03-13 was not sampled for molecular assays in 2016

In summary, the linking of signature metabolite detection to fumarate addition gene detection at Site A shows that they are not perfectly correlated. However, at least one or the other

(metabolite or gene) has been detected in HC-containing groundwater. This trend was most prevalent with alkylsuccinate signature metabolites and the assA gene presence. In most cases, sampled groundwaters showed genetic potential for anaerobic degradation, but did not show the metabolic evidence (e.g., metabolites themselves) of that reaction. However, since the nature of the metabolites is transient due to groundwater flow, the detection (or lack thereof) does not necessarily equate to no activity. A lack of metabolite detection could also be due to analytical limitations, or possibly because the contaminant being no longer present due to further degradation.

3.4.3.2 Site B signature metabolites and FAE gene presence.

3.4.3.2.1 Aromatic hydrocarbons, benzylsuccinates, and bssA.

At Site B, aromatic HC-derived signature metabolites were more abundant in 2016, mainly deriving from xylenes and toluene. o-Xylene, m-xylene, and toluene-derived metabolites were detected in 3 wells (Table 4). Detection in both 2016 and 2017 occurred in REC 34, the only well with >1 ppm toluene. This suggests, as in Site A, that higher concentrations of precursor hydrocarbons may result in metabolite production over a longer period of time. Thus, the degradation activity is ongoing and detectable over time. Only one well, REC 34, contained detectable ethylbenzylsuccinate (fumarate addition metabolite from ethylbenzene (0.325 ppm), in 2016. However, there is no HC threshold above which metabolites are consistently present in wells with an HC concentration greater than x ppm. This inconsistency in the detection of metabolites was also seen in other studies2, wherein with similar HC profiles did not have similar

49 metabolite detection profiles. As was mentioned in Section 3.4.3.1, the appearance and disappearance of the metabolites is most likely due to degradation of the succinate metabolites, or their removal (migration) from sampling well/areas over time due to groundwater flow.

SITE B REC-34 REC-31 ISCO-03-C ISCO-04-C REC-24 REC-11 ISO-49 REC-12 S14-7R ISCO-03-B REC-26 S14-49B TRIP BLANK Aromatic Hydrocarbons Toluene (ppm) 3.83 0.867 0.557 0.92 0.377 0.732 0.87 0.0299 0.125 0.0713 0.0523 <0.0003 <0.0003 Ethylbenzene (ppm) 0.325 0.394 0.0051 0.0323 0.297 0.0408 0.223 0.0489 0.0365 0.108 0.0421 <0.0005 <0.0005 Xylenes (ppm) 6.11 0.979 0.0051 0.0323 0.675 4.96 4.46 0.0432 0.0754 0.0454 0.403 <0.0005 <0.0005 Alkylbenzylsuccinates Benzylsuccinate 2017/2016 2016 2017/2016 Ethylbenzylsuccinate 2016 o -, m -, p - 2017/2016 2016 methylbenzylsuccinate Gene Presence bssA 2016 (Taq) bssA bssA bssA bssA bssA 2017 (Taq) bssA bssA bssA 2016 (KAPA) bssA bssA bssA bssA bssA 2017 (KAPA) bssA bssA bssA bssA bssA bssA bssA Alkanes C6 - C10 Alkanes (ppm) 22.4 15.6 13.5 13 12.5 9.2 7.3 4.7 2.7 1.9 1.1 <0.1 <0.1 Alkylsuccinates C5 2017/2016 2017 2016 2016 2017/2016 2016 2017 2016 2017/2016 C5 w/ unsat 2017/2016 2017/2016 2016 2016 2017/2016 2016 2017/2016 C6 2017 C6 w/ unsat 2017/2016 2017/2016 2016 2016 2017/2016 2017 2017/2016 2016 C7 2016 C7 w/ unsat 2016 2017/2016 2017 2017 C8 C8 w/ unsat C9 C9 w/ unsat 2016 Gene Presence assA 2016 (Taq) assA assA assA assA assA assA assA assA assA 2017 (Taq) assA assA assA assA assA assA 2016 (KAPA) assA assA assA assA assA assA assA assA 2017 (KAPA) assA assA assA assA assA assA assA assA

The presence of the bssA genes in these wells co-occurred with metabolite presence less, as was also seen for Site A (Section 3.4.3.1.1). Of the 3 metabolite-containing wells in 2016, only 2 also harboured the bssA gene (ISO49 and REC 34), both of which had o-xylene and/or toluene derived metabolites (benzylsuccinate and o-methylbenzylsuccinate). The bssA gene was detected in 4 wells in 2016 and 2 wells in 2017 using Taq polymerase. Using KAPA polymerase,

4 wells had bssA detection in 2016, and 7 in 2017. Only 4 wells had repeated detection from

2016. Thus, the polymerase used for gene detection did have an impact on the detection of genes in a field sample. Furthermore, the all three wells which had benzylsuccinate metabolites detected (REC 34, ISCO-4-C, REC 11, and ISCO-3-B), all had detected bssA (with little consistency across genes and polymerases used (Table 4)). All other wells with detected bssA did not show corresponding metabolite detection.

The appearance and disappearance of the bssA gene suggests that groundwater movement is responsible for transporting the organisms to and away from the sampling sites. Due to the heterogeneity of environmental samples it would be unreasonable to expect complete correlation of genes and metabolites. However, the detection of bssA in wells with TEX concentrations >0 ppm, and with some metabolite detection, these data provide evidence for anaerobic aromatic

HC degradation at Site B.

3.4.3.2.2 Alkanes, alkylsuccinates, and assA.

Compared to aromatic concentrations, alkane concentrations are 2 – 20-fold higher than

TEX concentrations in Site B wells. Of the 12 wells sampled, metabolites were detected in 10 wells, all having alkane concentrations > ~2 ppm (Table 4). The most common metabolites were saturated and unsaturated C5 (9 wells, 3 repeats in 2017), and unsaturated C6 alkylsuccinates (8 wells). The presence of diverse alkylsuccinates gives strong evidence for the active anaerobic degradation of alkanes at Site B.

As was the case with the Site A, there is some co-occurrence with the concentrations of alkanes to the presence of alkylsuccinate metabolites. All wells with alkane concentrations above

2 ppm have metabolites present. This suggests that an alkane concentration greater than ~2 ppm are more likely to have metabolites detected. More wells had detected metabolites in 2016 than

2017, which suggests that either active biodegradation of metabolites was occurring, or groundwater movement removed the metabolites from the sampling area.

Using Taq polymerase, 8 of 12 wells revealed the presence of the assA gene in 2016

(ISCO-3-B, ISCO-3-C, ISO49, REC 11, REC 12, REC 26, REC 34, and S14-7R), while microbes in 5 of 10 wells harboured the gene in 2017 (ISO49, REC 12, REC 26, REC 31, and

S14-7R). Of these wells, 4 wells had repeated detection from 2016 to 2017: ISO49, REC 12,

REC 26, and S14-7R. In 2016, using KAPA polymerase, 7 of 12 wells harboured the assA gene

(failing to detect assA in REC 34 and ISCO-3-C which were detected using Taq) but having additional detection in wells ISCO-4-C and REC 24. Detection of assA using KAPA in 2017 was similar to Taq (in 6 of 10 wells), with novel detection in REC 24. The ability of Taq and KAPA polymerase to amplify the target was not as expected. In many cases, the ‘lower fidelity’ polymerase amplified the bssA gene in wells where KAPA failed, and vice versa. Thus, it seems the higher fidelity polymerases are not infallible. Interestingly, consistent detection of assA was seen in wells with alkane concentrations in the 1 – 7 ppm range but, not in higher contaminated wells. Like the metabolites, there is no clear defining characteristic of any well that determined whether the genes will be present. Wells with consistent assA detection across sampling years do not necessarily have alkylsuccinates present (e.g., REC 12). Relating the presence of alkylsuccinates with the presence of benzylsuccinates, again there is little correlation. This suggests that activity of anaerobic alkane degradation does not guarantee aromatic degradation will be present in that well as well, or vice versa. Only 3 wells have both classes of metabolites present, REC 34 which has the highest TEX concentrations (10.265 ppm;) and alkane concentrations (22.4 ppm), ISO49 (TEX = 5.533 ppm ; alkanes = 7.3 ppm), and well S14-7R

(TEX = 0.2369 ppm; alkanes = 2.7 ppm). Again, there were no trends in HC concentrations that

52 would favour the detection of either metabolites or genes. However, it is interesting that alkane degradation and aromatic degradation is seen, yet do not completely overlap, as was seen in Site

A. Again, as mentioned in Section 3.4.3.1, groundwater movement and metabolic activity, rather than hydrocarbon concentration, may govern the appearance and disappearance of genes and metabolites. However, the failure to detect both genes and metabolites in this and other wells/sites may be due also to limitations in molecular and/or analytical techniques/instrumentation (e.g. primer sets, polymerases, detection limits for succinates, etc.).

In summary, attempts to positively correlate the presence of FAE genes, produced signature metabolites, and concentrations of HC falls short, as no clear defining trends were present. Because of site variation, it is prudent to examine groundwater samples for both diagnostic genes and metabolites to ensure that one or the other is captured as evidence for site potential for anaerobic HC degradation. However, the detection of metabolites and/or genes in both sites give strong evidence for both the potential for anaerobic hydrocarbon degradation in these sites (gene presence), but also that degradation activity is occurring over the sampling period (metabolite presence).

3.4.3.3 Electron acceptor concentrations in field sites.

Site A and Site B maps (Figure 2 and Figure 3, respectively) summarized the HC and potential electron acceptor concentrations in the groundwater samples. The following sections presents the

EA data in more detail.

3.4.3.3.1 Site A

Figure 6A shows that one uncontaminated well contained ~ 4 -7 mM nitrate (well C02-

07) while nitrate was comparatively low in all contaminated wells. Sulfate concentrations were comparatively higher in 4 wells (5 – 10 mM), including in low/uncontaminated wells C02-07,

MW07, suggesting that sulfate in the groundwater flowing into the site can serve as an EA

(Figure 6B). Fe (II), the product of microbial Fe (III) reduction, was also detected between 1 – 2 mM in a handful of wells (Figure 6C). These data suggest that the site is characterized by a mix of these electron accepting conditions, with background electron accepting conditions a mixture of nitrate- and sulfate-reducing (as judged by concentrations in background, uncontaminated wells). The dominant organisms present in this site are expected to be nitrate- and sulfate- reducing, with possible iron-reducing microorganisms.

3.4.3.3.2 Site B.

Figure 7 shows the concentrations of nitrate (A), sulfate (B), and iron (II) (C) in Site B.

Nitrate and sulfate concentrations for almost all samples collected from in Site B are near 0 mM.

The anomalous higher peaks seen for ISO49 and ISCO-3-C is due to the accidental incubation of

55 concentrated biomass at room temperature for an unknown length of time (>2 months)††. The wells implicated in this erroneous incubation are ISO49, REC 34, ISCO-3-C, and S14-49B.

Thus, these data are not likely to reflect the true nitrate and sulfate concentrations. Well ISCO-3-

B had elevated Fe (II) concentrations, which is suggesting that iron-reducing organisms may also be present. From these data, iron-reducing and methanogenic organisms are primarily expected.

†† This was due to some dumb-bell who knocked the sample bag out of the -80oC freezer and taped it to the freezer door, leaving it at room temperature without finding out who the samples belong to, despite being labelled with the Lab name. #Freezergate #NeverForget

3.4.4 16S rRNA gene sequencing using Illumina Mi-Seq.

3.4.4.1 Site A microbial community composition.

To determine the microbial community composition of the collected GW samples,

Illumina Mi-Seq sequencing was performed, with taxonomic analysis done using the MetaAmp pipeline75. Figures 8 and 9 show the microbial community composition of Site A in 2016 and

2017, respectively.

The heterogeneous nature of these samples is reflected in the 16S rRNA gene sequencing results. The most abundant taxon in all samples in 2016 and 2017 was the genus Rhodoferax, present in all wells in the range of 3 – 44% (in 2016) and 5 – 62% (in 2017) relative abundance.

Members of this genus are known to have nitrate- and iron-reducing abilities, aligning with some of the terminal electron acceptor data (Section 3.4.3.3.1 and Figure 6). This taxon was also detected in microcosm in separate studies done by Aburto and Fahy, who saw the taxon disappear when anaerobic cultures were transferred to oxic conditions78,79. This transfer was not conducted in our studies, so it cannot be determined if Rhodoferax would respond similarly.

These researchers had also noted that cultures which had dominant Rhodoferax did not exhibit any benzene degradation, but did see benzene loss when other taxa like Hydrogenophaga increased in abundance79. Appearance of taxa other than Rhodoferax are inconsistent (i.e. greater contaminated wells do not have a particular taxon enriched relative to less or uncontaminated wells). This is interesting as it suggests that the microorganisms present in these field site well groundwaters are not culminating into a similar hydrocarbon-degrading community. Sulfate was also present in this field site (Figure 6), but no sulfate-reducing organisms were identified. Many of the taxa present in some of the contaminated well samples affiliate with known hydrocarbon degraders, such as Azoarcus, Geobacter, Polaromonas, some members of the Burkholderiaceae, and Rhodocyclaceae families, and Methanosaeta. However, none are unique to any well which has hydrocarbon concentrations above a certain value, nor are they dominant. Other taxa present are generally known to be soil microbes but not hydrocarbon degraders. Thus, these data suggest that the native microorganisms present in the contamination plume are not necessarily HC- degraders. It is possible that the taxa responsible for degradation may have been underrepresented due to PCR and sequencing biases, or they are less common in the environment.

However, in such heterogeneous samples not detecting these HC degrading organisms does not necessarily mean that HC degradation activity is not occurring. Evidence of anaerobic hydrocarbon degradation was explained in previous sections, thus the presence or absence of known hydrocarbon degrading taxa are not necessarily diagnostic of potential. As will be discussed in prepared microcosm experiments (Chapter 5), under some conditions the known HC degrading taxa present in these groundwaters are present and can be enriched.

3.4.4.2 Site B microbial community composition.

The microbial community analysis of Site B samples collected in 2016 and 2017 is summarized in Figures 10 and 11, respectively. As was seen in Site A, the most abundant taxon was the iron (III) and nitrate reducing Rhodoferax. Beyond this consistent detection, no other taxon was consistently detected in wells with ‘high’ or ‘low’ concentrations of hydrocarbons.

This heterogeneous distribution of organisms is similar to that seen in Site A. However, in 2017, the distribution of taxa are consistent across sampled wells (Figure 11). Again, Rhodoferax and other Burkholderiaceae are the most abundant, with Polaromonas being detected in wells with high HC concentrations in 2017 and in low to uncontaminated wells in 2016. Primarily,

Rhodoferax and other Burkholderiaceae are the most abundant in wells with HC concentrations

>5 ppm, with the uncontaminated wells reporting no Rhodoferax (or any Burkholderiaceae). In

2017, the uncontaminated well, S14-49B, is primarily composed of the methanogens,

Methanosaeta and Methanobacterium. Methanosaeta was detected in REC 12 (which was enriched in established microcosms; Chapter 5) and in ISCO-3-C. In 2016, neither of these taxa were detected in S14-49B, nor were they detected in REC 12, it was only detected in ISCO-3-C.

No trends can be concluded from this sequencing data, beyond identification of a few potential hydrocarbon-degrading taxa.

Linking identified taxa to assA and bssA genes that were detected in the groundwater samples for both Sites A & B should be examined in future studies and will be discussed further in Chapter 4.

Figure 10: Microbial community composition of Site B in 2016. Amplification of the 16S rRNA gene using primers 926f/1392r 16S degenerate primers targeting the V6-V8 hypervariable regions of 16S rRNA gene. All taxa whose abundances are below 5% were removed and summed into “Other (Below 5% Relative Abundance)”. Wells are in order of decreasing total HC concentration; uncontaminated well is S14-49B.

3.5 Conclusions

Assessment of the bioremediation potentials of contaminated sites A and B was done using three different methods, 16S rRNA gene sequencing, FAE gene detection analysis, and signature metabolite detection. These approaches showed that anaerobic alkane and mono-aromatic hydrocarbon biodegradation was occurring and continued to occur between yearly sampling

(2016 and 2017). Appropriate signature metabolites were detected in both sampling years which showed that some of the organisms at the sites were actively catalyzing fumarate addition reactions for hydrocarbon degradation. However, the detection was not as widespread as the hydrocarbon contamination plume suggested, likely due to the metabolically transient nature of the metabolites (metabolite flux) and less likely, movement of groundwater that may have transferred metabolites away from where they were produced. The genetic potential for anaerobic hydrocarbon degradation was seen at both sites as the assA and the bssA genes were detected across sampling years, providing evidence that microorganisms capable of fumarate addition were present. The microbial community analysis of these sites showed Rhodoferax as the dominant taxon, but no clear trends between levels of contamination and particular taxa could be deduced. The heterogeneous nature of the groundwater samples/aquifer systems seem to add to the complexity of the sequencing data. However, the collective metabolite and genetic evidence gleaned from this study gives unequivocal evidence that both sites have the potential for anaerobic hydrocarbon degradation and actively degrading HC at these sites.

Chapter 4: Primer design for quantification of assA and bssA on field samples from

Site A and Site B

4.1 Introduction:

Polymerase Chain Reaction (PCR) based assays to determine the presence of the assA

(encoding alkylsuccinate synthase; which catalyzes fumarate addition to alkanes) and bssA

(encoding benzylsuccinate synthase; which catalyzes fumarate addition to methyl/ethyl substituted aromatics) genes have been used for nearly two decades in both lab and field studies.

Assaying for the presence of these genes gives a proxy for the anaerobic bioremediation potential of hydrocarbon (HC) contaminated sites. For example, if these genes are detected in a given sample, the potential for bioremediation can be assumed. Compared to mere gene detection by

PCR, few studies have quantified anaerobic biodegradation genes using quantitative PCR

(qPCR). qPCR quantifies genes using alternating steps of target amplification and fluorescence measurement when added intercalating dyes bind to double stranded DNA, thereby quantifying

DNA concentrations with each thermocycle. Most previously designed primers sets have been used to target assA or bssA of specific terminal electron accepting taxa (i.e. sulfate reducers or nitrate reducers). Recently, primers designed to capture assA/bssA across electron accepting conditions have been explored with success, but are not compatible with qPCR methods, where the suggested amplicon lengths are <400 bp 24,46. Hence, quantification assays for these genes remain comparatively less developed, owing to the difficulty of confident diversity capture24,27,51.

A limitation of this assay is the lack of characterized assA/bssA sequences in genomic databases

(that is, sequences that are associated with sequenced genera/species). Many PCR (and qPCR) primer sets have been designed for different but specific ‘clades’ of assA/bssA ( e.g., nitrate or sulfate-reducers, or for syntrophs in methanogenic cultures, producing different length

66 amplicons)24,47,51. In order to investigate a field sample (that may be harbouring diverse clades) for these sequences, all or a combination of these primers sets need to be used, increasing the materials and time needed27,47.

In this study, I explored the potential for using a mixture of equimolar primers, each designed to target a different subset of assA/bssA sequences deposited in the NCBI database that represent the diversity of the assA/bssA sequences and that are compatible with qPCR methods.

These primers all bind to the same region and produce the same length amplicon, thereby making them compatible as a mixture for qPCR applications. Using these qPCR primers, HC- contaminated field samples (those described in Chapter 3) were interrogated to determine efficacy of quantification and to determine whether there were any correlations between gene abundance and HC concentrations. High-throughput sequencing was done to confirm whether these primers were able to perform comparably to those from literature. This primer mixture experiment was done as a proof-of-concept approach, as it may be the first to attempt to use a mixture of individually designed primers to capture greater diversity, going further than previous in silico analyses80. The nature of the primer mixes welcomes additions and modifications to primer mix composition, tailoring primers according to application. Using these new primer mixtures, we hypothesized that an increase in assA and bssA abundances would positively correlate with increased alkane and aromatic concentrations in the groundwater, respectively.

4.2 Methods:

4.2.1 Designing assA and bssA primers that capture diversity

Searching for ‘assA’ and ‘alkylsuccinate synthase’ in the NCBI database resulted in

~1,700 entries, from which 20 were selected (7 from organismal entries, 9 from non-taxonomic

67 clones). Similarly, a search for ‘bssA’ and ‘benzylsuccinate synthase’ resulted in ~16,000 entries, from which 17 sequences were selected (9 from organismal entries, 8 from non-taxonomic clones) (Table 23) (Chapter 8: Appendix). Initially a larger pool of sequences were selected from the NCBI database, in an attempt to encompass different sampling sites and research groups, but were removed due to repeated sequences or sequence homology.

Selected sequences were compiled into a Multiple Sequence Alignment (MSA) using

ClustalOmega aligning algorithm via the Benchling online tool, replacing duplicated sequences81,82. ClustalOmega was chosen at it does not reverse sequences and was found to align sequences best (compared to MAFFT, the other sequence alignment tool integrated in

Benchling) without introducing gaps. Regions of sequence homology were identified (i.e. regions that had fewest basepair mismatches), and forward primers were designed using the built-in Primer3 function. The primers generated by Primer3 did not align with the selected regions to target, thus they were adjusted slightly in length to bind to the intended region. This was done to ensure uniform produced amplicons, which would otherwise complicate qPCR analyses. Primers specific to the differing sequences in the MSA were made by modifying the new primer to be identical to sequences in the MSA that were not homologous with each other.

Name Sequence Tm (oC) Reference bssA forward (qPCR) bssOil GAA TCC CTG GTT ACA GGT CCA C 64.1 This study bssMys CAA TCC GTG GCA CAA CTG CAT G 66.3 This study bssSuf GAA TAC GTG GAG CGA CCC GCT C 68.1 This study bssWin CAA TCC GTG GCT TCA GGT TCA T 65 This study bssOri2 CAA TGC CTG CCA TAA CCC CAT C 65.5 This study bssMtl2 GAA CGG TTC GCA TAA CCC CAT C 65.2 This study bssA reverse (qPCR) bssHitr TCC TCG TAG CCT TCC CAG TT 64.6 This study assA forward (qPCR) assOri CTC CGC CAC GGC CAA CTG 67.4 This study assMsd CTC AGC CAC CGC CAA CTG 65 This study assEx CTC TGC GAC CGC GAA TTG 63 This study assSml TAG CGC CAC GGC CAA CTG 67 This study assA reverse (qPCR) 8543r TCG TCR TTG CCC CAY TTN GG 65.7 (von Netzer et al. 2013) Literature primers 7772f GAC ATG ACC GAC GCS ATY CT 65.1 (Winderl et al. 2007) 7757af TCG GAC GCG TGC AAC GAT CTG A 69.8 (von Netzer et al. 2013) 8546r TCG TCG TCR TTG CCC CAY TT 66.4 (von Netzer et al. 2013) - von Netzer et al. (2013)46 - Winderl et al. (2007)24

Ultimately, 6 forward bssA and 4 forward assA primers were designed that were specific for MSA sequences that were not homologous to each other but were more homologous than other regions of the assA or bssA gene (Table 5). These sequences spanned the majority of sulfate-reducing microorganisms (Table 23, Appendix). Primers for the Betaproteobacteria assA were not designed as this experiment was a proof-of-concept that a primer mix could work

69 effectively. Homo/hetero-dimerization among forward primers and the reverse primer was tested in silico using IDT OligoAnalyzer 3.1, discarding/redesigning any primer with ΔG < -7 kJ

(where more negative ΔG necessitates more energy required to go from dsDNA (double- stranded) to a ssDNA (single-stranded DNA)). Forward primers were then combined into an equimolar (20 nM) mix (separate mixtures were prepared for assA and bssA). The reverse primer used for assA was taken from literature, as sequence coverage in the MSA did not extend beyond the binding region of 8543r (the reverse primer used) (Table 5)46. The reverse primer for bssA was designed similarly to forward primers (as literature primers binding site was not reflected in the MSA across majority of sequences) and named bssHitr, which binds 48 bp upstream of 8546r

(reverse primer46), and so named as it ‘Hits’ (aligns with) the upstream region.

Unfortunately, only 4 forward primers were included in the bssA primer mix. Ori2 and

Mtl2 were removed due to intermittent issues with primer dimer formation, which is suspected to be an artifact from frequent freeze/thaw‡‡. I chose to discard these 2 primers as they had the greatest mismatches compared to the other 4 primers. Resulting amplicon lengths using the qPCR primer mix were for assA = 486 bp, and bssA = 141 bp. The amplicons derived from these primers were designed in such a way as to amplify a region within the amplicon from primers designed by Winderl et al. and von Netzer et al. as a method to ensure accurate amplification24,46.

Table 5 summarizes the primers used and designed in this study.

In addition to FAE gene quantification, FAE gene detection assays were also done, following the protocol described previously (Section 3.2.7). Following the conclusions found in

‡‡ As such, primers were re-diluted frequently and always loaded into reaction mixtures while on ice. They could have been kept in the primer mixture but due to repeated problems, this was decided against to save time and get the assay to work.

Chapter 3 regarding the influences of polymerases on gene detection, 3 polymerases were compared here to determine if differences are also seen when using the newly designed primer mixes. The polymerases used for these PCR assays were: Fermentas Taq, KAPA HiFi, and

BioLine SensiFast qPCR polymerases. Fermentas Taq & KAPA HiFi used with literature primers only, while BioLine was used for newly designed qPCR primers. The BioLine SensiFast polymerase is a SYBR green mastermix formulated for qPCR. These polymerases were chosen as they were the polymerases most used in lab for PCR assays (not including the SensiFast polymerase which was only used for qPCR). Previously Bio-Rad SsoFast SuperMix was used as the preferred qPCR polymerase in lab, however, due to cost limitations the BioLine SensiFast polymerase was adopted after tests determined the activity of both qPCR mixes to be similar

(data not shown).

4.2.2 qPCR Thermocycling Conditions:

The reaction volume used was 25 µL (consisting of 12.5 µL BioLine SensiFast, 10.5 µL

PCR-grade H2O, 0.5 µL each primer, 1 µL DNA template). Thermocycling conditions for assA qPCR was as follows: 95oC for 3 min, 39 cycles of 95oC for 15 sec, 62oC for 25 sec (each cycle ending with a plate read), 62oC for 2 min, followed by a melt curve (65oC to 95oC, increasing in

0.5oC increments, hold for 5 sec, then plate read). The annealing temperature for bssA was adjusted to 65oC following the same method. All qPCR was done using BioLine SensiFast No-

ROX MasterMix on a Bio-Rad CFX96 Real-Time Thermocycler controlled by Bio-Rad CFX

Manager software. All software analysis was performed using Single Threshold Cq

Determination with Baseline Subtracted Curve Fit settings.

All samples were analyzed using a 1% agarose gel electrophoresis method to validate the sizes of the resulting amplicons. Only those samples which had bands that matched the amplicon in the standard curve were taken as positive amplifications (assA = 486 bp; bssA = 141 bp). All other samples that had non-specific amplicons were disregarded. Amplification was done according MIQE guidelines83: standard curve R2 > 0.98 (for standard curve linearity), E = 80% -

120% (E = Efficiency of amplicon doubling each cycle, where 100% = 2x increase each cycle,

90% = 1.8x), NTC Cq > 38, ΔCq < 0.2 (NTC = no template control; Cq = Quantification cycle, the cycle at which amplification enters the logarithmic phase, the point at which abundance calculations are done). All samples that fell outside of these guidelines were removed or repeated. Standard curves and all samples were analyzed in triplicate, with 2 No-Template

Controls (NTCs) prepared for each assay.

4.2.3 Sample Preparation/DNA Extractions

Samples used in this study were taken from initial DNA extractions of Site A and B field samples from both 2016 and 2017. Protocols followed were as described in Section 3.2.4 &

Section 3.2.5.

4.2.4 qPCR Standards for absolute quantification

Standards for qPCR were prepared using the archetypal anaerobic toluene-degrader

Thauera aromatica (bssA) and anaerobic alkane degrader Desulfatibacillum alkenivorans AK-01

(assA) using literature primers46,20. Primers designed in this study were designed in such a way that the amplified region fell within the amplicon product of literature primers, which, theoretically, would ensure amplification when the literature primers were effective.

Amplification of the assA and bssA genes using the literature primers (77757af/8543r, assA46;

7772f/8546r, bssA20) was performed and appropriately sized bands were gel extracted using a

Qiagen Gel Extraction Kit (Qiagen: Hilden, Germany). Amplicons were quantified using a Qubit fluorometric system (2 μL DNA sample). Standard curves were prepared using a 1:10 dilution of amplicons from 107-101 copies/μL (assA) and 109-104 copies/μL for bssA. Standard curves were generated in triplicate.

Converting gene copies/μL to gene copies/L GW (groundwater) was based on the assumption that DNA extractions from soil/groundwater were 100% efficient. Calculating gene copies/L were based on an original groundwater sample volume of 1 L that was concentrated to

10 mL (a 100x concentration), therefore the amount of biomass in both volumes is identical, assuming 100% efficient concentration. Sampling 1.6 mL for DNA extraction from the concentrated 10 mL is equivalent to sampling 160 mL from 1 L, thus:

푐표푝푖푒푠 푛 푐표푝푖푒푠 × 75 휇퐿 푒푙푢푡푒푑 퐷푁퐴 = Equation 1 휇퐿 160 푚퐿 푒푥푡푟푎푐푡푒푑 퐺푊

푛 푐표푝푖푒푠 푥 푐표푝푖푒푠 = Equation 2 160 푚퐿 푒푥푡푟푎푐푡푒푑 퐺푊 1000 푚퐿 푡표푡푎푙 퐺푊

푛 푐표푝푖푒푠 푥 푐표푝푖푒푠 = ( ) 1000 푚퐿 푡표푡푎푙 퐺푊 Equation 2.1 160 푚퐿 푒푥푡푟푎푐푡푒푑 퐺푊

푐표푝푖푒푠 × 75 휇퐿 푥 푐표푝푖푒푠 = ( 휇퐿 ) 1000 푚퐿 푡표푡푎푙 퐺푊 Equation 1 + 2.1 160 푚퐿 푒푥푡푟푎푐푡푒푑 퐺푊

푥 푐표푝푖푒푠 푛 푐표푝푖푒푠 = × 468.75 Equation 3 퐿 푔푟표푢푛푑푤푎푡푒푟 휇퐿

Where: ‘n copies’ = quantified copies/μL of assA or bssA from extracted DNA;

‘x copies’ = converted quantification from copies/μL extracted DNA to copies/L GW.

Analysis of assA and bssA quantifications involved the analysis of the melting curve

(obtained by increasing the reaction temperature by 0.5oC and taking a fluorescence reading) where a fluorescent signal is emitted during intercalation of SYBR Green to the dsDNA form.

Melting temperatures are specific and dependent on the G+C content and length of the amplicon product, where the amplicons used for standards had a melting temperature of ~87.5 – 88oC

(bssA) and ~89 – 89.5oC (assA). Interestingly, not all environmental samples had melting temperatures that matched those of the standards, with melting temperatures ranging from 82 –

92oC. This vast range of observed melting temperatures was traced to both differing G+C content of amplicons (suspected) and longer amplicons arising from non-specific binding (as was determined through gel electrophoresis). Only those samples that matched the amplicon length from standards were plotted in Figure 14 to Figure 16 (Section 4.3).

4.2.5 Illumina MiSeq sequencing of unknown amplicons and Sanger sequencing of

standards

Those samples which matched the amplicon lengths of the standard curve amplicons were gel extracted (Qiagen Gel Extraction Kit) and all 3 replicates were pooled (total volume 75

μL). Primers containing the Illumina MiSeq adapter sequences (for addition of sequencing barcodes) were taken from 16S rRNA gene primers commonly used in our studies (926f/1392r)73 and added to the 5’ end of each primer constituting the primer mix for both assA and bssA (Table

6). MiSeq primer mix was prepared to the same concentration as the qPCR primer mix (20 μM).

Thermocycling conditions were as follows: 95 oC – 3 min, 16 cycles of 95oC – 30 sec, 62oC – 15 sec, 72oC – 15 sec, final extension at 72oC – 5 min using KAPA Hi-Fi polymerase. All samples were analyzed in replicates of 2 or 3, depending on the concentration of purified amplicon.

Illumina Nextera XT Indexes were assigned and attached using PCR. Thermocycling conditions are as follows: 95oC – 3 min, 8 cycles of 95oC – 30 sec, 55oC – 30 sec, 72oC – 30 sec, final extension at 72oC – 5 min. Reactions were in 50 μL volumes containing 25 μL ThermoFisher

Fermentas Taq polymerase, 10 μL purified MiSeq adapter attached amplicon, 1 μL of forward and reverse barcoding primer (100 nM), and 13 μL PCR grade water. These second-round PCR

75 amplicons were gel purified the same way as first round. Samples were then normalized and pooled into a 2 ng/μL library and submitted to International Microbiome Centre (University of

Calgary) for Illumina sequencing (via 2 x 300 bp sequencing kit) of assA sequences.

Preparation of bssA amplicons was done in an identical manner to that of assA, with the following changes: second-round PCR amplification (attaching Nextera XT Indexes) was done using KAPA Hi-Fi polymerase in single replicates of 25 μL reactions (12.5 μL KAPA Hi-Fi polymerase, 5 μL DNA, 1 μL primers (200 nM), 6.5 μL PCR H2O. Thermocycling conditions were as follows: 95oC – 3 min, 8 cycles of 95oC – 30 sec, 60oC – 15 sec, 72oC – 15 sec, final extension at 72oC – 5 min. Sequencing was done at The University of Calgary Cumming School of Medicine Centre for Genomics and Informatics, using the Illumina MiSeq 150x2 kit.

Sequences were then assembled using PEAR 0.9.6 (50 bp overlap, assembly length set to 0 (does not truncate)) and analyzed through MetaAmp using the ‘non-16S’ option (which does not assign taxonomy; reads truncated to 150 bp)74,75. BLASTn analysis, alignment, and tree building were performed as described above for assA.

Table 6: Designed Illumina MiSeq adapter primers for assA and bssA qPCR primers. Adapter sequences are attached to the 5’ end. “MS” denotes MiSeq. Note, that MSbssOri2 and MSMtl2 (shaded) were not used in the final primer mixture, as they were excluded from the non-MiSeq primer mixture (included in table for reference). Name Sequence (Nextera Adapter Sequence in bold) Length (bp) bssA forward MSbssSuf TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGAATACGTGGAGCGACCCGCTC 55 MSbssWin TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCAATCCGTGGCTTCAGGTTCAT 55 MSbssMys TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCAATCCGTGGCACAACTGCATG 55 MSbssOil TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGAATCCCTGGTTACAGGTCCAC 55 MSbssOri2 TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCAATGCCTGCCATAACCCCATC 55 MSbssMtl2 TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGAACGGTTCGCATAACCCCATC 55 bssA reverse MSbssHitr GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGTCCTCGTAGCCTTCCCAGTT 54 assA forward MSassOri TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCTCCGCCACGGCCAACTG 51 MSassMsd TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCTCAGCCACCGCCAACTG 51 MSasSml TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGTAGCGCCACGGCCAACTG 51 MSassEx TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCTCTGCGACCGCGAATTG 51 assA reverse MS8543r GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGTCGTCRTTGCCCCAYTTNGG 54

Resulting sequences were assembled using PEAR 0.9.6 (50 bp overlap, 350 bp truncation

(the default setting)) and analyzed through MetaAmp using the ‘non-16S’ option (which does not assign taxonomy) at 97% similarity74,75. The sequences were analyzed using MetaAmp as a primer sequence trimming and quality control step75. Resulting ‘cleaned’ sequences were analyzed through BLASTn84. Sequences matching assA annotated sequences or whole genomes from putative alkane-degrading microorganisms were selected, disregarding any sequence which returned matches with <70% coverage and identity. Sequences that passed this quality control and selected for analysis were 14 of 86 for assA and 16 of 65 for bssA. Selected sequences were then compiled into a Multiple Sequence Alignment (MSA) in UGENE software and aligned using ClustalW to determine similarity to other sequenced reads85. ClustalW was chosen as it gave the least gaps in aligned sequences when alignments were performed compared to using other alignment algorithms (MUSCLE, ClustalOmega, and T-Coffee) (data not shown). A

Maximum Likelihood dendrogram was built using approximate Likelihood Ratio Test (aLRT) tree branch support (to estimate correct placement of internal branches) and a combination of

Nearest Neighbor Interchanges (NNI) and Sub-tree pruning and Regrafting (SPR) tree building methods (for initial tree building when choosing which possible tree is the most likely result from given nucleotide data)86,87. Using this combination of tree building/refining, highly computationally demanding bootstrap calculations are unnecessary86,87. These methods were chosen as Maximum Likelihood analysis does not cluster the sequences by assumed homology thereby removing slight diversity among sequences (assuming each sequence is unique and determining best placement within a changing distribution), which was deemed to be paramount in this proof-of-concept application.

Sanger sequencing of standard curve amplicons was done through Eurofins (facility in KY,

USA), and resulting sequences were analyzed through a BLASTn search84. Both assA and bssA

BLASTn searches returned sequences at 100% match submitted by original literature primer authors (data not shown).

The hydrocarbon concentrations used to plot qPCR results as a function of hydrocarbon concentrations were given in Table 1 (Site A) & Table 2 (Site B) (Section 3.4).

4.3 Results and Discussion:

4.3.1 Newly designed bssA qPCR primer mix

4.3.1.1 Site A bssA detection.

Table 7 shows the detection of the bssA gene in Site A field samples using literature primers with Taq and KAPA polymerases and newly designed qPCR primers with BioLine polymerase. Microorganisms in 4 well samples from this site harboured the bssA gene in both

2016 and 2017 (C02-08, C03-12, C01-04, and C03-10). For bssA, in most wells where literature primers detected bssA using Taq, the newly designed primers do as well, detecting the gene in 2 of 3 wells, with detection in 2 additional wells (Table 7). However, assaying with KAPA polymerase detected bssA in 4 uncontaminated well samples, which were not detected using Taq polymerase or with qPCR primers and BioLine polymerase. There is an element of uncertainty, as will be mentioned later, that non-specific amplicons of similar length when using KAPA made accurate identification of the correct length amplicon difficult. There is a chance false-positives are reported here. However, in some cases bssA was detected in previously undetected well samples in Site A using BioLine polymerase and qPCR primers, where it was previously undetected using literature primers and Taq polymerase. More consistent detection was seen in some wells over sampling years using qPCR primers across all used polymerases (i.e. C03-10,

C01-04, C01-01, and C03-12) (Table 7). The difference in detection ability using the same literature primers and different polymerases may be a key contributing factor to the success of the newly designed primers. If the polymerase has greater fidelity (e.g. such as KAPA polymerase) and/or ability to amplify a target, more target detection is likely, as was demonstrated in detection differences between Taq and KAPA in detecting bssA. However, the efficacy of the newly designed qPCR primers plus a qPCR polymerase, the results suggest that the qPCR primers have similar bssA capture to literature primers using high-fidelity polymerases. The outstanding detection of bssA in uncontaminated well groundwater using

KAPA polymerase may be a limitation in the design of the primer mix, false-positives, or microbes present in the GW ecosystem that possess that gene.

Detection of the bssA gene using newly designed primers was successful but did not match the detection using literature primers and KAPA polymerase. Of the 4 wells with TEX

79 concentrations >0.1 ppm, qPCR primers detected the bssA gene in only 2 (C03-10 (1.6 ppm), and

C01-04 (0.1 ppm)). The qPCR primers failed to amplify the bssA gene in the highest TEX contaminated wells, C02-08 and C03-14. Literature primers detected the bssA gene in 3 wells in

2016 and all 4 wells in 2017 with KAPA, which included the highest contaminated wells, C02-

08 and C03-14. The bssA gene was not detected in 2017 using Taq polymerase but had the same detection pattern as KAPA in 2016 for well samples with HC concentrations >0.1 ppm (3 wells).

Inability to detect the bssA gene in the higher contaminated wells using qPCR primers is most likely due to limitations in primer design, where homologs of bssA may be present, or the literature primers are amplifying non-specific targets, thereby skewing the comparison between literature and newly designed qPCR primers.

Literature PCR primers Literature PCR primers Designed qPCR primers Total TEX Site A bssA (Fermentas Taq) (KAPA Hi-Fi) (BioLine) (ppm) Well 2016 2017 2016 2017 2016 2017 2.6 C02-08 bssA bssA bssA 1.9 C03-14 bssA 1.6 C03-10 bssA bssA bssA bssA bssA 0.1 C01-04 bssA bssA bssA bssA bssA <0.1 C03-13 <0.1 C01-01 bssA bssA <0.1 C03-12 bssA bssA bssA bssA <0.1 MW 07 bssA <0.1 MW23 bssA bssA <0.1 C02-06 <0.1 C02-07 bssA <0.1 C03-11 bssA bssA <0.1 TRIP BLANK

In contrast, in wells with lower TEX contamination, of the 8 wells with TEX concentrations < 0.1 ppm, newly designed qPCR primers detected the bssA gene in 2 wells (C01-

01 and C03-12), while detection using literature primers and KAPA polymerase detected the gene in 6 wells. The bssA gene was not detected in any wells with <0.1 ppm TEX using Taq polymerase.

Results from this bssA detection assay suggest that the literature primers with KAPA polymerase results in the most positive detection (10 well samples), while qPCR primers with qPCR polymerase is comparable but only detects the gene in 4 of the 10. These data show that the presence of the bssA gene is not dependent on the concentration of TEX. This is most likely from the movement of groundwater transporting these bss harbouring microorganisms across the site. Possible influence of PCR inhibitors co-extracted with the gDNA may be reflected in the presence/absence across used polymerases.

4.3.1.2 Site A bssA quantification.

Quantification of the bssA genes in Site A samples are summarized in Figure 14 and plotted against total BTEX concentrations. False positives (those with non-specific amplicons) and zero value quantifications were eliminated from the figure. As was reflected in Table 7, only

4 wells had quantifiable bssA with the qPCR primers used. Wells C03-12 and C03-10 had the most bssA L-1 (C03-12 2016 = 3.84x1010 copies/L, 2017 = 4.08x109 copies/L; C03-10 2016 =

4.48x109 copies/L, 2017 = 3.18x109 copies/L). These newly designed qPCR primers gave values approximately 2 orders of magnitude greater than any other abundance obtained from other groups and field sites20,50,26 (Table 8). This discrepancy may be due a number of factors; first, these newly designed primers may be more effective in capturing the bssA genes present

(however, tests regarding detection ability on isolates other than Thauera aromatica should be conducted to determine efficacy, for example Georgfuchsia and Azoarcus); second, abundances may be inflated due to non-specific amplification (which was seen during gel electrophoresis validation of newly designed qPCR primers and non-qPCR literature primers), and third, the abundances may accurately reflect the abundances of bssA, while literature qPCR studies may be under-quantified (a simultaneous quantification of 16S rRNA genes may serve as a comparison to literature abundances). Interestingly, the bssA gene abundance is not directly linked to BTEX concentrations, i.e. the presence of higher BTEX concentrations does not mean more bssA genes, as C03-10 has nearly 20x more BTEX despite having comparable bssA abundance to well C03-

12 (Figure 14). Also, wells C03-12 and C01-01 groundwaters both have <0.1 ppm TEX and yet

C03-12 has nearly 2 – orders of magnitude higher bssA abundances.

Although the reasons for the disparity are not clear, groundwater flow rates may be a reason for the comparable bssA abundances among disparately contaminated wells, despite the fact that the sampling wells are on opposite sides of the site (Figure 2). Tracer studies have shown that microbes can travel distances of approximately 9 m in 20 days in silty groundwater aquifers88, thus there is precedence for microbial transport in aquifer environments. The field sites in this study have been contaminated for at least 10 years, thus, established transport of native organisms enriched on the aromatic contaminants and travelling the relatively short distance across the field site is almost guaranteed and may be one factor in explaining the similarity of bssA abundances between wells C03-10 (1.6 ppm TEX) and C03-12 (<0.1 ppm

TEX). Activity in the plume leading to the sampling well may have had an influence on aromatic

HC degraders, causing an enrichment of BSS encoding organisms in areas where nutrients and/or aromatics may be more concentrated, thus degrading the aromatics and proliferating, followed by the movement of the groundwater and sampled across site.

Thus, in this field site (Site A), there was no correlation between concentration of mono- aromatic compounds and the detection/abundance of bssA. As will be discussed in later section

(4.3.1.4) pertaining to Site B, different sites can have bssA presence/absence linked with TEX presence/absence.

Reported Highest Reported Targeted FAE Targeted Electron Accepting Detection Limit Sample Abundance Cited Primer Reference Reference gene Conditions (copies/L or g) (copies/L or g) ~ log 5 ~ log 9 Sulfate-reducers and methanogens This Study assA not reported ~ log 6 Methanogenic paraffin degrading Oberding and Gieg (2018) ~ log 3 ~ log 8 Sulfate-reducers and methanogens Aitken et al. (2013) ~ log 5 ~ log 10 Nitrate- and sulfate-reducers This Study ~ log 2 ~ log 8 Sulfate-reducers Winderl et al. (2008) Pilloni et al. (2019) Winderl et al. (2007) non-qPCR ~ log 4 ~ log 8 Nitrate- and sulfate-reducers Muller et al. (2017) primers 7772f/8546r ~ log 4 ~ log 4 Sulfate-reducers and methanogens Beller et al. (2008) Oka et al. (2011) bssA ~ log 2 ~ log 3 Nitrate-reducers Beller et al. (2002) Oka et al. (2011) ~ log 1 ~ log 8 Nitrate-reducers Beller et al. (2002) Kazy et al. (2010) ~ log 3 ~ log 8 Sulfate-reducers Beller et al. (2008) ~ log 3 ~ log 7 Deltaproteobacterial "F1" Winderl et al. (2008) ~ log 2 ~ log 8 Nitrate-reducers Beller et al. (2002) Da Silva and Alvarez (2004) ~ log 3 ~ log 8 Nitrate-reducers Beller et al. (2002)

4.3.1.3 Site B bssA detection.

Detection of the bssA gene in Site B is more consistent than Site A in terms of total wells with bssA detected in groundwater, across polymerases and primers (Table 9). Of the 12 sampled wells, 6 well samples had detectable bssA using Taq, 7 with KAPA, and 9 with newly designed qPCR primers and BioLine qPCR polymerase (regardless of sampling year). Under all

PCR/qPCR parameters, bssA was detected in all wells with >0.1 ppm TEX (exceptions: ICSO-4-

C which did not have bssA detected in groundwaters sampled in either 2016 nor 2017, and well

S14-49B, which had bssA detected in 2017 despite a TEX concentration <0.1 ppm). However, at

Site B there was still inconsistent detection of the bssA gene in each of the wells across polymerases and primers. For example, only 3 of 7 wells in 2016 matched detection with Taq and qPCR (REC 34, REC 11, and ISCO-3-B), and only 1 of 9 well matched detection between

KAPA and qPCR (S14-7R) in 2016. In 2017, Taq and qPCR detection of bssA matched in 1 of 5 wells (ISCO-3-C); KAPA and qPCR matched in 3 of 8 wells (ISO49, ISCO-3-C, and S14-7R).

The efficacy (detection in samples) of the bssA qPCR primer mix in detection in Site B seems to be better than what was seen in Site A, however, comparing detection between KAPA and qPCR is difficult due to possible false-positives in the KAPA assays (as was discussed earlier, Section

4.3.1.1). The detection of the bssA gene was not consistent across any primer or polymerase, as only a maximum of 3 wells had consistent bssA detection using all polymerases. Interestingly, there is inconsistent detection between Taq and KAPA, using the same primer set. One would expect that the detection using KAPA would be increased relative to Taq (due to higher fidelity, less inhibitor effects etc.), and yet there are wells which showed bssA detection with Taq and not with KAPA, and vice versa. A possible explanation could be the differences in the PCR reagents and PCR conditions, which only highlights the effect of PCR reagents in PCR assays.

Differences in detection between the qPCR and KAPA assays is likely due to the differences in primer diversity, where some bssA sequences (or homologs) are missed, in addition to the aforementioned PCR effects.

Literature PCR primers Literature PCR primers Designed qPCR primers Total TEX Site B bssA (Fermentas Taq) (KAPA Hi-Fi) (BioLine) (ppm) Well 2016 2017 2016 2017 2016 2017 10.3 REC-34 bssA bssA bssA 5.7 REC-11 bssA bssA 5.6 ISO-49 bssA bssA bssA bssA 2.2 REC-31 bssA 1.3 REC-24 bssA bssA 1.0 ISCO-04-C 0.6 ISCO-03-C bssA bssA bssA bssA 0.5 REC-26 bssA bssA 0.2 S14-7R bssA bssA bssA bssA 0.2 ISCO-03-B bssA bssA 0.1 REC-12 bssA bssA <0.1 S14-49B bssA <0.1 TRIP BLANK

4.3.1.4 Site B bssA quantification.

Quantification of the bssA gene in Site B plotted against BTEX concentration shows that the minimum abundance of bssA, regardless of BTEX concentration, is approximately 1x106 copies/L (Figure 15). Only one well, S14-7R, had detection both 2016 and 2017, which saw a

138-fold increase in bssA copies/L from 2016 to 2017. What is most prevalent in Figure 15 is (as was also seen in Site A) that the abundance of bssA was not linked to the concentration of

BTEX, as wells with similar BTEX concentrations had at least 2 orders of magnitude difference in bssA abundance.

Also, there is no correlation across all wells with similar abundance and concentrations of a single aromatic compound. For example, groundwater from ISO49 and REC 11 had similar xylene and toluene concentrations (Xylenes = 4.46 and 4.96 ppm, Toluene = 0.87 and 0.732 ppm, respectively), but the bssA abundance differed by 2 orders of magnitude. Wells REC 31 and ISCO-3-C had similar bssA abundances and total BTEX concentrations, however ISCO-3-C had negligible xylene and ethylbenzene (0.0051 ppm for both), while only the toluene concentration was similar (REC 31 = 0.867 ppm, ISCO-3-C = 0.557 ppm). These toluene concentrations are similar to that of REC 11, which has bssA abundances of 3 orders of

87 magnitude less than ISCO-3-C and REC 31. Thus, there is no clear contributing HC compound concentration where similar concentrations dictate similar gene abundance. While these higher bssA abundance wells are all localized in the same area (west of the ‘CONDO PROPERTY’ and

‘RV LOT’ (Figure 3), the other wells in that area do not have similar abundances. Thus, the activity/growth of BSS encoding organisms are not uniform in the groundwater, which is likely due to the heterogeneous nature of environmental samples. DNA extraction variance between samples may also be a contributing factor. Due to the heterogeneity of field sites/samples, the former seems most likely20,89,90.

In Site B, bssA was not detected nor quantified in uncontaminated wells. This was also seen in the field study conducted by Beller et al. (2008)50. Other groups have detected background abundances (e.g. in uncontaminated samples) of bssA to be approximately 104 copies/L, while the lowest abundances seen in this study is 105 copies/L (Table 7)26,51.

Background bssA levels in this system are most likely below the detection limit for this assay/primer design, which is 105 copies/L, and does not necessarily mean bssA is non-existent in uncontaminated areas (as shown for our site A analyses where high abundances (~log 8 copies/L) were measured in lower contaminated wells. There also seems to be an increase of almost 2 orders of magnitude from 2016 abundances to 2017. As no HC quantification data was available for field sites in 2017, this increase cannot be attributed to a corresponding decrease in

HC concentrations, nor can it be attributed to increases in benzylsuccinate signature metabolites detected in 2017, as fewer wells had detected benzylsuccinates from 2016 to 2017 (Table 4).

In summary, the quantification of the bssA gene in field sites suggests that bssA gene abundances are independent of aromatic HC concentrations under these conditions. This

88 observation may be due to heterogeneity in groundwater flow and variation in the contamination plume, or to other microbial or environmental factors which have not been explored here.

4.3.2 Newly designed assA qPCR primer mix.

4.3.2.1 Site A assA detection.

Table 10 outlines the detection of the assA gene in Site A using Taq, KAPA polymerases, and the newly designed qPCR primers (with BioLine polymerase). Taq polymerase amplified assA in 6 of 12 well groundwater samples, KAPA polymerase amplified assA in 7 of 12 well samples (with overlaping detection with Taq in 5 wells), and newly designed qPCR primers amplified the gene in 9 of 12 well samples (with 6 well samples overlapping with KAPA polymerase). All wells in Site A had alkane concentrations of less than 6 ppm, thus the possible effect alkane concentrations (as will be seen in Site B; Section 4.3.2.3) are not expected here.

Detection of the assA gene was more consistent across the primer and polymerase combinations.

Only one sampled groundwater, from well C01-04, did not have assA amplified by Taq or qPCR, but was detected using KAPA polymerase. Detection of the assA gene also occurred in 3 wells with alkane concentrations <0.1 ppm (C02-07, C03-11, and MW07) using the qPCR assay.

Overall, the performance of the newly designed qPCR assay was similar to that of the literature primers using both Taq and KAPA polymerases. The assA gene was detected consistently in wells with alkane concentrations >0.1 ppm.

Table 10: Comparison of assA presence in Site A between literature primers (comparing amplification via Fermentas Taq and KAPA Hi-Fi) and newly-designed assA qPCR primer mixture. “assA” denotes presence, blank space denotes absence. Dashed spaces indicate wells were not sampled in that year.

Literature PCR primers Literature PCR primers Designed qPCR primers Total Alkanes Site A assA (Fermentas Taq) (KAPA Hi-Fi) (BioLine) (ppm) Well 2016 2017 2016 2017 2016 2017 5.5 C02-08 assA assA assA assA assA assA 4.4 C03-14 assA assA assA assA assA 0.9 C03-10 assA assA assA assA assA 0.5 C01-04 assA assA 0.4 C03-13 assA assA 0.2 C01-01 assA assA assA 0.1 C03-12 assA assA assA assA <0.1 C02-06 <0.1 C02-07 assA assA <0.1 C03-11 assA assA <0.1 MW 07 assA assA <0.1 MW23 <0.1 TRIP BLANK

4.3.2.2 Site A assA quantification.

The presence of the assA gene, as mentioned above, suggests ASS encoding organisms were present natively in the groundwater/soil environment. Quantification of assA shows that the highest alkane-contaminated wells (C02-08 (5.5 ppm) and C03-14 (4.4 ppm)) have assA concentrations in 107copies/L (C02-08) and 108 copies/L (C03-14) (Figure 16). Wells with alkane concentrations <0.1 ppm have assA quantifications between 105 and 106 copies/L. It is most likely that the 105 copies/L groundwater is reflecting the limit of detection or non-specific amplicons of the qPCR assay, rather than the actual abundances of assA in ‘background’ wells, which may be lower. However, false positives were eliminated from analyses, through verification via gel electrophoresis against a positive control (as described in Section 3.2.7).

Well C03-10 has similar assA abundance as C03-14 and C02-08, despite having a 5-fold lower alkane concentration (0.9 ppm). However, the general trend in assA abundances seen in

Site A suggests that the higher the alkane concentration, the greater the assA abundances, to a limit of log 8 assA copies/L at ~5 ppm alkanes. This trend was not seen for bssA abundances

(both Site A and B) (Sections 4.3.1.2 & 4.3.1.4). This observation supports the hypothesis that increased alkane concentrations result in higher assA abundances, however it seems that a threshold for assA abundance is linked to alkane concentrations above ~5 ppm. Comparing abundances across sites in wells with similar alkane concentrations, the abundances are within

91 the same order of magnitude, ~107 copies/L in wells with ~5 ppm alkanes. These data are very exciting because consistent trends in abundance are seen in both Sites A and B.

In summary, the detection of the assA gene using newly designed qPCR primers is similar to that when using literature primers (both using Taq or KAPA polymerases). The quantification of the assA gene correlates with alkane concentrations to an upper threshold of ~5 ppm, where abundances do not increase thereafter.

4.3.2.3 Site B assA detection

Table 11 summarizes the detection of assA using literature primers 7757af/8543r46 using

Taq and KAPA polymerases, and the qPCR primer mix designed in this study. Wells with microorganisms harbouring the assA gene were somewhat consistent across the polymerases and primers, where Taq amplified assA in 9 of 12 wells, KAPA in 10 of 12 wells, and newly designed qPCR primers in 11 of 12 wells (including uncontaminated well S14-49B).

Performance of the polymerases did not seem to have an effect in well samples with alkane concentrations <10 ppm, where all wells samples had detectable assA. However, in wells with >10 ppm alkanes, the detection ability between polymerases does not overlap as well. Taq amplified the assA gene in 3 of 5 wells (1 in 2017), KAPA in 5 of 5 wells (with only one well overlapping with Taq detections), and the newly designed primers amplified assA in 4 of 5 wells

(2 wells overlapping with Taq, and 4 wells overlapping with KAPA). This suggests that the concentrations of alkanes >10 ppm has an effect on PCR efficiency. As discussed earlier, the difficulty in distinguishing true-positives from false-positives was also seen here. Some well samples, when amplicons were run on a 1% agarose gel, had non-specific products, where

“larger” in ISCO-3-C (2016) under qPCR primers refers to an amplicon not the same length as

92 the standards (Table 11). These types of ‘non-standard’ bands were common in both PCR and qPCR assays, with literature and newly designed qPCR primers. Identification of this suspected non-target amplicon through was not done. Melt curve analysis in the qPCR protocol using newly designed primers showed many curves not matching with the melt curve for standards.

This suggests that the primers are either amplifying non-target DNA (which will be briefly discussed later) or that the primers in the mix are forming complex dimers/trimers. These unknown amplicons introduce a level of difficulty when working with this primer mix. Although amplification and quantification occur (and as will be discussed in Section 4.3.4.2, wherein the amplicons were verified to be bssA) this primer mixture is not perfect, but it was found to be adequate. With the exploratory nature of the use of a primer mix, perfection cannot be expected.

In comparison with the few assA quantification studies, the overall abundances are similar, ~ 8 log per unit sample in higher contaminated samples (Table 8)27,60. To the author’s knowledge, very few studies have quantified assA gene abundances in environmental samples, much less than for bssA (Table 8). Further studies of assA abundances in contaminated aquifers would give insight to the trends of assA in response to more diverse short chain alkane contaminants.

In summary, the detection of the assA gene using newly designed qPCR primers is comparable to literature primers, using either Taq or a high-fidelity polymerase. However, the primer mixture is not infallible, and the occurrence of non-specific amplification is an issue.

Table 11: Comparison of assA presence in Site B between literature primers (comparing amplification via Fermentas Taq and KAPA Hi-Fi) and newly-designed assA qPCR primer mixture. “assA” denotes presence, blank space denotes absence. Wells in bold-face and shading were not sampled in 2017. Dashed spaces indicate wells were not sampled in that year.

Literature PCR primers Literature PCR primers Designed qPCR primers Total Alkanes Site B assA (Fermentas Taq) (KAPA Hi-Fi) (BioLine) (ppm) Well 2016 2017 2016 2017 2016 2017 22.4 REC-34 assA assA assA assA 15.6 REC-31 assA assA assA 13.5 ISCO-03-C assA assA larger 13 ISCO-04-C assA assA 12.5 REC-24 assA assA assA 9.2 REC-11 assA assA 7.3 ISO-49 assA assA assA assA assA assA 4.7 REC-12 assA assA assA assA assA assA 2.7 S14-7R assA assA assA assA assA assA 1.9 ISCO-03-B assA assA assA 1.1 REC-26 assA assA assA assA assA assA <0.1 S14-49B assA <0.1 TRIP BLANK

4.3.2.4 Site B assA quantification.

Quantification of assA in Site B samples (Figure 17) shows a range of almost 4 orders of magnitude difference, where the lowest and highest gene abundance was seen in wells with alkane concentrations below 1 ppm (REC 26, and S14-49B (uncontaminated)).

The uncontaminated well S14-49B has the assA gene present most likely reflects background assA abundance, or from microorganisms transported via groundwater flow as this well is located near well ISO-49, which has assA near 1x107 copies/L. Of the 16 wells (both years) which had positive assA quantification, 9 harboured 107 copies/L, spanning alkane concentrations from 4.7 ppm (REC 12) to 22.4 ppm (REC 34).

However, it seems that the assA abundances do not pass 107 copies/L GW when alkane concentrations are >5 ppm, but reach 108 copies/L GW at alkane concentrations <5 ppm. This is interesting, as it suggests that a threshold of growth and/or activity exists, where alkane

95 concentrations > 5 ppm are limiting to growth of ASS harboring organisms in some way, similar to what was observed in Site A. However, in wells where assA was detected in both years, abundances were within an order of magnitude, which may suggest that either the alkane degrading community in the well areas are somewhat stable/limited in growth by nutritional/hydrocarbon factors limit movement. As in Figure 3 (Site B Map) those wells that have similar assA abundances despite drastically different alkane concentrations are all clustered together on the west side of ‘CONDO PROPERTY’ (wells REC 12, ISO49, REC 11, REC 24,

REC 31, and REC 34) (Figure 17). It may be that the cement foundation of the property is limiting nutrient availability or limiting microbial movement/densities. Well REC 26 (109 copies/L GW) is in an open area central to the site, and wells ISCO-3-B and ISCO-4-C (both 106 copies/L GW) are more north in what is assumed to be an open lot with little to no underground structures.

4.3.3 Effect of different polymerases on FAE gene detection

Addressing the differences in polymerase sensitivity in this study was important to understand what effects the polymerase itself was making on the positive and/or negative FAE gene presences. Taq polymerase, isolated from Thermus aquaticus, is a DNA polymerase I with no exonuclease activity, and is susceptible to inactivation by phenol91. KAPA Hi-Fi polymerase is listed as a B Class polymerase (DNA Polymerase II) with 3’ – 5’ exonuclease activity92. The type and class of the BioLine SensiFast polymerase is unknown. Both the KAPA and the

BioLine polymerases are considered High Fidelity (HiFi) polymerases. A higher fidelity polymerase is defined as a polymerase that has exonuclease proofreading activity (usually 3’ –

5’), which decreases error rates during DNA replication. This exonuclease activity is most

96 relevant in sequencing applications, rather than in presence/absence assays. As was discussed earlier, the polymerase used had a substantial effect on FAE gene detection, where increased detection was generally seen using KAPA and BioLine polymerases but did not always detect the FAE genes in the same well samples as did Taq polymerase. These differences in the detection may be due to the sensitivity of the polymerase to low abundance targets (where the target gene is below a sensitivity threshold of the polymerase itself and is less likely to be amplified) or from the presence of PCR inhibitors. These differences in polymerases have been found in forensic studies primarily focused on polymerase comparisons using different sample types with possible PCR inhibitors present (humic acids, metal ions, phenol, blood, bone, etc.), and undoubtedly in other studies as well93–96. To the author’s knowledge, these comparisons in polymerase activity have not been explored in hydrocarbon-contaminated systems.

As was seen in the previous sections discussing the FAE gene detections in these sampled field sites, the differences in the amplification ability of the bssA gene between Taq polymerase,

KAPA polymerase, and BioLine polymerase is most likely due to higher sensitivity and resistance to PCR inhibitors of the latter 2 polymerases (as has been tested in the lab using samples previously unable to be amplified using Taq polymerase, Section 3.2.6). Generally, humic acids have been found to inhibit PCR through both polymerase binding and sequences specific DNA binding, which causes a shift in the amplicon lengths produced94,97. Protein mutations on Taq polymerase have been shown to increase resistance to some PCR inhibitors, however, to the authors knowledge, the specific mechanism for each inhibitor type has not been determined93,94,97. This is in part due to the diversity of polymerases used and possible PCR inhibitors as well. Differences in the FAE gene detection that was seen in this study was most likely due to the polymerase’s resistance to inhibitors, or differences in the inhibitor

97 concentrations found in each sample (e.g., from differences in water chemistries around a sampling well). Thus, the differences seen in polymerase sensitivity may be due to both the presence of PCR inhibitors and to the enzyme structure of high-fidelity polymerases.

4.3.4 assA and bssA qPCR amplicon sequencing using Illumina MiSeq.

4.3.4.1 assA sequencing results.

Illumina MiSeq sequencing of assA amplicons from field samples (this Chapter) and microcosm experiments (Chapter 5) samples was performed. After read preparation and quality control (as described in Section 4.2.5), a BLASTn search of the 86 sequences returned 6 which matched NCBI entries annotated as assA with >70% identity and coverage§§. Positive assA reads were then compiled into a Maximum Likelihood tree with related NCBI matches and sequences from literature (Figure 18). Clustering of the various sequences shows that 2 main clusters of assA sequences are present in these samples; those from sulfate-reducers (like Desulfatibacillum alkenivorans, and Desulfatibacillum aliphaticivorans) and methanogenic/syntrophic cultures

(with Smithella sp.)24,30,46,54 ***.

Addressing the efficacy of the primers constituting the primer mix, some BLASTn results matched the microorganisms the primers in the mixture were designed to specifically target. For example, primer ‘assOri’ was designed from D. alkenivorans (Ori = Original; D. alkenivorans was set as the ‘archetypal’ organism, as it was available in lab)98, and primer ‘assSml’ was designed to capture Smithella SCADC sequences (Sml = Smithella) and sequences closely

§§ Sequence matches with <70% coverage and identity were not annotated as assA sequences, and thus disregarded. *** matches to characterized assA (AK-01 and Smithella SCADC) were accompanied with ‘cloned’ and/or ‘uncultured’ assA.

98 related to the assA genes from these taxa were found (Figure 18). It is likely that the directed design of the primers in the primer mix may be influencing the types of sequences seen. The other primers in the primer mix, ‘assMsd’ and ‘assEx’ were designed from an annotated masD sequence and Desulfoglaeba alkenexedens, respectively. An assA sequence similar to that of D. alkenexedens was detected in these samples (Figure 18).

It is interesting that sequences associated with methanogenic (syntrophic) hydrocarbon degrading culture and sites activity were detected, despite no methanogenic activity seen in the sequenced samples (discussed in Chapter 5). Some samples, as those in REC 12 microcosms (as will be discussed later in Chapter 5) did harbor methanogenic taxa present, however methane was not detected. Thus, it may be that the assA sequences attributed to these syntrophic cultures may not necessarily be directly involved in methanogenic hydrocarbon-degrading activity. assA gene sequences associated with methanogenic cultures were not detected when using the

7757af/8543r primers designed by von Netzer et al.46, which suggests that this primer mix may have an advantage in detection of more diverse assA sequences (and the reverse primer, 8543r, was used as reverse primer for our primer mix). The assA primers designed by von Netzer46 were used to investigate diverse sites, including sites described as being methanotrophic. That study did not detect any methanogenic activity/taxa as was the case in this study. In that case, assA or assA-like sequences were retrieved.

Figure 18: Maximum Likelihood phylogram of sequenced assA and bssA amplicons. Nucleic acid sequences were aligned with ClustalX (30 iterations). Tree-building via PhyML 3.0 Maximum Likelihood using TN93 substitution model, fast likelihood based method aLRT for branch supporting, and SPR (Subtree Pruning and Re-Grafting) tree improvement model86,87,99. The outgroup is pyruvate formate lyase (pfl) from sulfate- reducing strain D. alkenivorans strain AK-01.

100

The performance of the primer mix in comparison to literature assA primers (namely von

Netzer’s 7757af/8543r)46 shows that the capture of diversity of assA sequences is improved, with detection of deltaproteobacterial taxa and taxa implicated in syntrophic alkane degradation. This new primer mix is an improvement on current PCR-based assays, which require the use of many primers, each producing unique amplicons, to assay for FAE genes. In this study, we demonstrated that a primer mix, with primers mindfully designed to target certain diverse clades of assA, each returning amplicons of identical length, can be used effectively for both PCR and qPCR applications.

4.3.4.1.1 Non-specific amplification using assA primer mix

BLASTn matches not annotated as assA (but had coverage and identity above 70%) are described in Table 24 (Chapter 8: Appendix). Almost all are not involved in HC degradation, but rather in protein catabolism, ribosome, or other metabolic functions.

Many of the organisms harbouring the mismatches are known as aerobic/microaerophilic and are present in HC contaminated sites (Variovorax (aerobic, aromatics), Polaromonas

(aerobic, aromatics), Azoarcus (facultative, aromatic and alkane), Massilia (aerobic, PAH) etc.)

53,100–103. This suggests that either the assA gene is present in these genomes and are uncharacterized, or that non-specific binding is amplifying non-assA genes from those organisms. One instance of an ‘aromatic hydrocarbon degrading protein’ was seen in the

BLASTn searches from Sphingobium sp. (an aerobic putative PAH degrader), which shows that many sequences annotated in silico may not be annotated as FAE genes. Two of these sequences are annotated as formyltransferases which may have similar active site motifs which led to amplification. This non-specific binding has an additional consequence of inflating gene

101 abundances in quantification assays, as was discussed earlier. When designing primers to survey diversity across clades, non-specific biding is to be expected. Previous groups have not used next-generation sequencing platforms for their assA or bssA identifications, and thus such non- specific binding have not been reported24,27,46,47,52.

Whether or not the assA gene has been correctly annotated in submitted genomes (or annotated as glycine radical or acyltransferases) cannot be controlled for. Searching for ‘alkyl’,

‘alkyl succinate’, ‘formate’ (for pyruvate formate lyase), ‘glycine’, ‘glycine radical’, or ‘glycyl’

(for glycine radical) in genomes of these organisms in NCBI resulted in no matches. Thus, homologous genes and/or missed annotations to other enzymatic functions is possible.

4.3.4.2 bssA sequencing results.

High throughput sequencing of amplified bssA gene sequences in field samples using newly designed bssA primer mix resulted in 65 sequences after quality control (as was described in Section 4.2.5). BLASTn search yielded 11 OTUs with matches to bssA sequences. Tree- building was combined with assA sequences and followed the same methods. Maximum

Likelihood analysis shows that 2 main clusters of bssA formed, one clustering within Clostridial species Desulfosporosinus, and the other clustering with alpha and betaproteobacteria, such as

Thauera, Azoarcus, and Magnetospirillum. No bssA sequences were retrieved from deltaproteobacteria. Such clustering across betaproteobacterial species like Azoarcus and

Thauera, suggests that these sequences are very similar, and are also similar to Magnetospirillum spp. bssA sequences. This clustering suggests that the bssA gene present in these genera are very similar, if not identical, which was also seen during alignment of these sequences during tree- building (data not shown)104. Recently, horizontal gene transfer via integrative and conjugative

102 elements (ICE) in Azoarcus CIB was determined to radiate both aerobic and anaerobic BTX degradation genes to genera previously unable to utilize these compounds66. Horizontal gene transfer was hypothesized to be the contributing factor of this clustering by Winderl et al.24 as well. Another explanation is that the close relatedness of diverse bssA sequences could be a consequence of the short amplicon lengths used in the alignment. As the amplicons were trimmed to 150 bp, the amount of nucleotide variation in such a short fragment would be underrepresented, thus giving the illusion of tight clustering†††. The latter explanation is most probable. Shorter amplicons are a consequence of qPCR primer design (amplicons are suggested to be <400 bp in length). In this case the bssA amplicon was 141 bp in length, thereby decreasing the sequence variation that amplicon can encompass. Alignment of shorter sequences to longer sequences in a somewhat conserved region of a gene makes assigning diversity and variation of these amplicons difficult. Had the amplicons been longer, it would be possible that the diversity of these sequences and the large cluster within the alpha and betaproteobacteria would be more widespread, perhaps into the deltaproteobacteria.

The tut operon found in Thauera aromatica was also a dominant match in many of the sequences. The tut operon is a homolog of the bss operon and is reported to be very similar in function and regulation to bss, therefore detection of that operon would be expected105. However,

Maximum Likelihood analysis suggests these sequences do not differ substantially. The assA and bssA genes seem to be closely related, as shown by the branching of the Desulfotomaculum bssA sequencs from the ass/mas cluster. This close association of bssA and assA has been shown before, including a close relationship to the assA gene of D. alkenivorans AK-0164,104. bssA

††† Both horizontal gene transfer and short amplicon lengths could also be attributed to assA sequencing results as well104.

103 primers designed by Winderl et al24. were also found to amplify the assA gene present in AK-01 during this study (data not shown).

Each primer in the primer mix was designed to target a specific group of bssA genes. In the MSA, many sequences that did not cluster were from uncultured clones. Primers were made from those sequences as well in an effort to capture that diversity as well. Primers ‘bssMys’,

‘bssWin’, and ‘bssOil’ were designed from such sequences (in NCBI: Mys = von Netzer submitted sequence, Win = Winderl submitted sequence, Oil = Uncultured sequence sample name with ‘OIL003’). The lack of characterized bssA sequences aligning to specific genera hinders direct targeting of ‘under-represented’ bssA in mixed communities. Primers bssOri, bssMtl, and bssSuf were designed to target characterized taxa (Ori = Thauera aromatica K172,

Mtl = Geobacter metallireducens, Suf = Desulfotomaculum)‡‡‡. Surprisingly, sequences similar to Thauera were captured despite omission of bssOri. Had bssOri and bssMtl remained in the primer mix, it is the belief of the author that there would be greater captured diversity, perhaps into deltaproteobacteria with Geobacter specific primer bssMtl. Assays with these omitted primers included should be done in the future.

Comparing the efficacy of these newly designed bssA primer mix to the von Netzer46 primers (upon which they were based) shows that they do not capture the same diversity. von

Netzer et al.46 demonstrated that their primers captured sequences within the deltaproteobacteria

(related sequences to Desulfobacula toluolica and Geobacter daltonii). Both sequences are included in Figure 18 for reference. Primers designed by Winderl et al24 (from which bssA qPCR primers were based) amplified bssA genes from deltaproteobacteria, with a greater emphasis on

‡‡‡ Primers bssOri and bssMtl were excluded from final primer mix as described in Section 4.2.1

104 suspected nmsA genes, rather than bssA when used by von Netzer et al46. Alpha and betaproteobacterial bssA sequences were underrepresented using von Netzer primers, so the newly designed qPCR primer mix for bssA seems to capture more sequences in that clade.

However, as was discussed earlier, the short amplicon/sequence length of our sequences may be inhibiting the alignment resolution (clustering those sequences together due to similarities in a shorter length sequence).

These data show that the newly designed qPCR primers for bssA did capture a diversity of taxa from alpha and betaproteobacteria, and Clostridia. In this case, the bssA primer mix is a proof-of-concept for a customized primer mix for FAE genes. These data suggest that the use of a primer mix to target specific diversity can work, if prior knowledge allows a directed and purposeful design. With more characterized bssA sequences available in databases, adding to the primer mix (and re-designing previously designed primers) will aid in directed amplification and quantification.

4.3.4.2.1 Non-specific amplification using bssA primer mix

Sequences that did not return bssA matches were compiled into Table 25 (Chapter 8:

Appendix). Compared to non-specific matches using assA primer mix, bssA non-matches are limited. The only ones were acyl-CoA dehydrogenases (which is involved in β-oxidation of fatty acids), a 30S rRNA gene (smaller subunit of prokaryotic ribosome) and surprisingly, an assA match. Similarities in the active site of the acyl-CoA dehydrogenase and the assA is likely contributing to the cross-amplification. This result is surprising, but not-unexpected, as the design of literature primers by Winderl et al.24 uses AK-01, an assA harbouring microorganism, as the archetypal organism. The fact that only one sequence returned an assA sequence, while the

105 same samples sequenced using assA qPCR primers did not return bssA sequences shows that there is little to no cross amplification of the assA and bssA genes using these primer sets. The bssA primer mix does seem to have some cross amplification, but that is deemed to be minimal.

Amplification of non-target genes were also reported by Winderl et al.24 showing low similarity to ABC transporter ATP-binding proteins. Again, as was explained previously, primers designed for diversity have a caveat of non-specific amplification, which was highlighted using next- generation sequencing.

4.4 Conclusions

Newly-designed primers for the improved detection and quantification of the assA and bssA genes in Site A and Site B showed great promise. These newly designed primers showed comparable sensitivity and amplification ability to literature primers, with some differences. The type of polymerase used with literature primers was seen to make a difference in the detection of the assA and bssA genes in field samples. Quantifications of the assA and bssA genes were successful, suggesting assA abundance is limited by an alkane concentration threshold of ~5 ppm. Quantifications of bssA suggested that bssA abundances are not correlated to aromatic concentrations found in field samples. However, the abundances of these genes are similar to those found in other studies (within ~ 2-orders of magnitude compared to published literature values) and thus can be used as a quantitative indicator for bioremediation potential. Sequencing of the assA and bssA qPCR amplicons yielded matches to known HC degraders such as

Desulfatibacillum alkenivorans AK-01, Smithella SCADC, Thauera aromatica, and Azoarcus spp. etc. Matches to as of yet uncharacterized assA and bssA clones/genomes, suggest that diversity of the assA/bssA sequences warrants further exploration. Hopefully these sequences

106 will be characterized in the future, showing the true depth of assA and bssA that literature and the new primer sets capture. These data paint a very promising picture for the use of these primers in diagnostic applications for determining the bioremediation potential of organisms present in HC contaminated sites.

107

Chapter 5: Microcosm experiments monitoring hydrocarbon loss and assA and bssA

gene abundance

5.1 Introduction:

Monitoring the loss of hydrocarbons in situ can be challenging due to the dynamics of groundwater movement. Thus, testing samples from contaminated environments in the laboratory for their ability to biodegrade hydrocarbons to assess the potential for biodegradation at a given site is pivotal to realizing bioremediation potential. Such an approach is part of the toolbox to understand to understand whether HC biodegradation is possible at a given site1,8,106.

A combination of methods to assess anaerobic hydrocarbon degradation in laboratory experiments can be done by monitoring changes in hydrocarbon concentrations, electron acceptor concentrations, FAE gene abundance/detection, and microbial community composition can provide a clear picture of the response the native microbial community has to a hydrocarbon load58,79,107–109. Assessment of bioremediation potential from these data would give a multifaceted and thorough understanding of plausible activity that may occur in situ.

In this study, we assessed the ability of microbial communities present in two fuel contaminated sites to biodegrade hydrocarbons under various electron accepting conditions.

Microcosms with an added mixture of benzene, toluene, and octane were prepared and analyzed.

Monitoring for hydrocarbon loss, electron acceptor changes, and assA and bssA gene abundances/detection was done to give thorough understanding of whether anaerobic hydrocarbon degradation is a viable option for remediation. Microbial community analysis was also performed. If the contaminated sites have the potential for anaerobic HC biodegradation, we expect to see hydrocarbon loss coupled with an increase in assA and bssA abundances and appropriate stoichiometric change in electron acceptor concentrations over time.

108

5.2 Methods:

5.2.1 Microcosm Experimental Design

Approximately 3 L of hydrocarbon-contaminated groundwater was received from two fuel contaminated sites. Groundwater from Site B wells REC 12 (lower contaminated) and REC

34 (higher contaminated), and Site A wells C01-04 (lower contaminated) and C02-08 (higher contaminated), (described in Section 3.2.1) were chosen as proxies for ‘higher’ and ‘lower’ contaminated wells respective of their sites, to determine whether differences in biodegradation activity was dependent on native hydrocarbon concentrations. Seventy millilitres (wells C02-08,

REC 12, and REC 34) or 50 mL (well C01-04; we received less groundwater volume from this well) of contaminated groundwater were dispensed into 160 mL serum bottles and capped with

Viton stoppers (The Chemours Company: The Netherlands) under anoxic conditions in a N2/CO2

(no H2) inflated anaerobic chamber. It is important to note that headspaces were not flushed with

N2 or N2/CO2 gas after dispensing in order to preserve the groundwater system (e.g. concentration of volatile hydrocarbons) as much as possible, e.g. to mimic ‘natural’ conditions.

An anaerobic minimal salts medium (Pfennig anaerobic medium, see Appendix 8.1) was prepared at a 10x concentration and was added to the microcosms, 0.7 mL for the 50 mL microcosms, and 1 mL for the 70 ml microcosms (omitting Pfennig I solution to reduce precipitation). Pfennig I was added to each microcosm separately (Appendix 8.1). This small amount of concentrated medium was added to microcosms to ensure that sufficient nutrients were available (e.g. nitrogen, phosphorus, potassium, trace metals and minerals, etc.) and to minimize the volume change in the microcosms. The electron acceptors nitrate (NaNO3), sulfate

(Na2SO4), or Fe (III) (FeO(OH)2) were added to the required volume during medium preparation.

109

The final concentration of each electron acceptor in the microsoms was 5 mM. Microcosms were either amended with nitrate (NaNO3), sulfate (Na2SO4), or Fe (III) (FeO(OH)2), All 3, or received no exogenous electron acceptors. Each condition was set up with six bottles, including two bottles that were designated hydrocarbon-free controls, to account for background electron- accepting processes. Notably, hydrocarbon-free controls could not be established for REC 34 and

C02-08, because these samples contained high background levels of HCs. Benzene, toluene, and n-octane (0.5 μL each) were spiked into each of the four test bottles. This amounted to 5.6 μmol benzene, 4.7 μmol toluene, and 3.1 μmol n-octane added to each hydrocarbon-amended microcosm. Partitioning of each hydrocarbon from aqueous phase to gaseous phase was calculated using Henry’s Law, using constants found in Eastcott et al.110. Solubility of benzene

(1780 mg/L), toluene (~520 mg/L), and octane (0.66 mg/L) were referenced from Heath et al.111.

Added volume of benzene and toluene are below solubility threshold in water at 25 oC (8.75 mg/L and 8.66 mg/L, respectively). Added octane was above solubility (7.04 mg/L). Theoretical loss of electron acceptors from complete anaerobic degradation of the added amounts of BTO

(Benzene, Toluene, and Octane) is summarized in Table 12. Sterile controls were set up using 2

HC-amended bottles with no added medium and autoclaved twice. All microcosms were incubated at room temperature in the dark.

5.2.1.1 Hydrocarbon monitoring

Over time, microcosms were monitored for hydrocarbon loss. Headspace analysis of hydrocarbon concentrations in each bottle was measured by Gas Chromatography – Flame

Ionization Detector (GC-FID) at day zero (before hydrocarbon spike) and day one (day following hydrocarbon spike), and approximately every 2 weeks thereafter, according to the

110 method described in Fowler et al (2012)112. Hydrocarbon standards for the BTO mixture were prepared analogous to the microcosms (in Milli-Q water). The first 3 readings were done using 1 mL plastic syringes which were inaccurate and gave highly varied results; subsequent measurements were done using a 100 μL gas-tight glass syringe.

5.2.1.2 HC Degradation Stoichiometric Reaction Equations under tested conditions.

The theoretical stoichiometric reactions for the biodegradation of benzene, toluene, and n-octane are shown below. Based on these, and the amounts of BTO added to each microcosm, theoretical electron acceptor concentrations that would be consumed were calculated and shown in Table 12.

Nitrate-Reducing Conditions113,114:

+ - C6H6 + 6 H + 6 NO3 → 6 CO2 + 3 N2 + 6 H2O

+ - - C7H8 + 0.2 H + 7.2 NO3 → 7 HCO3 + 3.6 N2 + 0.6 H2O

+ - - C8H18 + 2 H + 10 NO3 → 8 HCO3 + 5 N2 + 6 H2O

Sulfate-Reducing Conditions113:

2- - - + C6H6 + 3 H2O + 3.75 SO4 → 6 HCO3 + 3.75 HS + 2.25 H

2- - - + C7H8 + 3 H2O + 4.5 SO4 → 7 HCO3 + 4.5 HS + 2.5 H

2- - - + C8H18 + 6.25 SO4 → 8 HCO3 + 6.25 HS + 1.75 H + H2O

Iron (III)-Reducing Conditions55:

3+ - 2+ + C6H6 + 30 Fe + 18 H2O → 6 HCO3 + 30 Fe + 36 H

3+ - 2+ + C7H8 + 36 Fe + 21 H2O → 7 HCO3 + 36 Fe + 43 H

3+ - 2+ + C8H18 + 50 Fe + 24 H2O → 8 HCO3 + 50 Fe + 58 H

111

Table 12: Theoretical electron acceptor loss in microcosms amended with benzene, toluene, and octane (0.5 μL each) if HC were fully biodegraded. A distinction between complete degradation of benzene, toluene, and octane versus complete degradation of only toluene and octane is made. Microcosms sourced from Site A well C01-04 has a culture volume of 50 mL, all other microcosms contain 70 mL groundwater sample volume. Theoretical Electron Acceptor Loss (mM) in 70 mL volume in 50 mL volume Benzene + Benzene + Electron Toluene + Toluene + Toluene + Toluene + Acceptor Octane Octane Octane Octane Nitrate 4.44 1.92 6.22 2.68 Sulfate 2.78 1.20 3.89 1.68 Iron (III) 22.00 9.58 31.00 13.42

Sampling for electron acceptors was done on day 0 and on the final day, removing 500

μL of volume, centrifuging at 13,000 rpm for 5 minutes and keeping the volume at -20oC until analysis. Analytical methods for nitrate, sulfate, and iron (II) analyses were described in Section

3.2.3. t-Tests were performed for EA losses to determine if the changes seen in concentrations were significant across replicates, as denoted with asterisks.

5.2.2 DNA Extraction

Microcosms were sampled on day 0 and on the final day for DNA extraction. Only a single biological replicate was sampled for DNA extraction (due to budget limitations). The sampled volume (2 mL) was stored at -20oC until DNA extraction could be performed. The DNA extraction protocol was as described in Section 3.2.5. The extracted DNA was stored at -20oC until it could be used for 16S rRNA gene sequencing and FAE gene analysis.

112

5.2.3 16S rRNA gene sequencing for microbial community composition

Microbial community analysis was performed on extracted DNA from a single biological replicate for each electron-accepting condition. Microcosms REC 34 and C02-08 (both higher contaminated samples) were not sampled for the ‘no added hydrocarbon controls’, as the native hydrocarbon concentration was already elevated, thus no ‘no hydrocarbon controls’ existed. The

Illumina MiSeq 16S rRNA gene sequencing sample preparation protocol followed was identical to that described in Section 3.2.6.

5.2.4 qPCR Analysis for assA and bssA genes

Assaying microcosms for FAE gene abundances was done using the same extracted DNA samples as described above. Primers used to quantify the assA and bssA genes in microcosm samples are described in Chapter 4 (Table 5). Detection of the FAE genes was compared across

2 polymerases, KAPA HiFi and BioLine SensiFast polymerases, using newly designed qPCR primer mixes, and a literature primer set46 (designated ‘von Netzer primers’).

Calibration curves, qPCR reaction protocol, and thermocycling conditions are as described in Chapter 4, Sections 4.2.2 & 4.2.4. Abundances are reported as technical replicates of 3 replicates.

5.3 Results and Discussion:

5.3.1 Microcosm Well C01-04 – Site A, Low Contamination

5.3.1.1 Hydrocarbon degradation

Figure 19 shows hydrocarbon concentration measurements of benzene, toluene, and octane over 135 days in different electron acceptor-amended microcosms with unfiltered

113 groundwater from well C01-04. Well C0104 was the lower contaminated well of the two Site A microcosms (benzene = 2.45 ppm, toluene = 0.0153 ppm, alkanes = 0.5 ppm). Within 42 days of incubation, all 3 HC were depleted to nearly 100% in initial amendments under all electron accepting conditions, a trend that was not seen in other microcosms. In the sterile controls, a 40 –

50% loss of benzene and toluene was seen, along with 15.2% loss in octane, all of which can be assumed due to abiotic loss. In experimental microcosms, the remaining 50 – 60% HC was available for biotic consumption. Hydrocarbon degradation after re-amendment on C01-04 on day 66 showed a >90% toluene loss under nitrate- and sulfate-containing conditions, while iron- reducing conditions reached only 85.9% toluene degradation. Neither benzene nor octane saw losses greater than sterile controls over the same time period, despite seeing near 100% loss after the initial amendments under all nutrient treatments. Table 13 summarizes the HC losses measured in the incubations, with all 3 HC completely lost in the first incubation, and only toluene lost in the second incubation. Figure 20 shows the changes in electron acceptor concentrations during the first 90 days of incubation for all treatments (which spans initial amendment and re-amendment). The only significant changes in EA were measured in the Fe

(III) amended incubations (Figure 20C). No significant changes were measured in incubations relative to controls for other EA. It was initially unclear as to why substantial HC depletion did not correlate with EA changes, thus we hypothesized that the microcosms were not as anaerobic as we had initially thought. To test this, we analyzed for O2 concentrations after ~90 days because we had not flushed the headspace when the microcosms were initially dispensed (to preserve HC content).

114

115

Table 13: Percent HC loss in Site A sources microcosm C01-04, lower contaminated. Note C01-04 is split before and after reamendment at day 69 ('Post Reamendment'). Shading demarcates HC loss greater than sterile controls. Percent Loss on Day n Well (Days) EA Treatment Benzene Toluene Octane Nitrate 100.0 100.0 64.9 Sulfate 93.9 100.0 97.0 C01-04 Iron 100.0 100.0 100.0 (42) All 3 100.0 100.0 87.6 No EA 93.5 100.0 100.0 Sterile Control 41.4 49.9 15.2 Nitrate 20.3 100.0 36.9 C01-04 Sulfate 25.8 100.0 39.6 (69) Iron 28.6 85.9 40.0 Post All 3 -39.6 100.0 4.3 Reamendment No EA 1.6 100.0 18.0 Sterile Control 33.4 30.8 30.0

O2 measurements done at the end of the experiment (using the sterile control as a proxy measurement for time 0 O2 concentrations) suggested that indeed a microaerophilic environment was present in the headspace (~134 μM; 4.3 mg L-1), which most likely influenced both the initial rapid hydrocarbon degradation rate (which was not continued after re-amendment, as final

O2 measurements show 0 μM oxygen) and the less than expected changes in electron acceptor concentrations. The rapid rate at which the HC were initially degraded, with less-than-expected change in electron acceptors, and complete loss of O2 in the headspace supports that aerobic biodegradation occurred in these microcosms18.

116

5.3.1.2 FAE gene abundance in C01-04 microcosms.

5.3.1.2.1 assA gene detection and quantification.

Detection assays using PCR comparing von Netzer et al. primer set 7757af/8543r and newly designed qPCR primers showed that the von Netzer primers were only able to detect the assA gene in T=0, ‘All 3’ + HC, and ‘All 3’ No HC samples. Non-specific binding was seen in 7 of 10 samples, including all samples with assA detected. Detection using newly designed qPCR

117 primers showed detection of the assA gene in 2 of 5 HC amended samples (sulfate and ‘All 3’, assA was detected in ‘All 3’ using von Netzer primers) and in one non-HC amended sample (‘All

3’ which was detected using von Netzer primers). The gene was not detected in T=0 sample using these newly designed primers, but other non-specific bands were seen. This shows that detection ability of these newly designed primers is similar to that of literature primers off which they are based, but do not overlap entirely. Differences in the assA sequences specific to enrichments may have an influence on the discrepancies in detection, as primer sets are biased to certain sequences/groups (as explained in Chapter 4).

Literature PCR primers Designed qPCR primers HC C01-04 assA (KAPA Hi-Fi) (BioLine) Added? EA Condition Target Non-Specific Target Non-Specific T = 0 assA NSP NSP Nitrate Sulfate NSP assA NSP Yes Iron NSP 'All 3' assA NSP assA No EA NSP NSP Nitrate Sulfate No Iron NSP 'All 3' assA NSP assA NSP

Quantification of the assA gene in microcosms prepared from C01-04 GW showed that octane amendments had a significant impact on assA abundance, when comparing abundances between HC-amended and no-HC microcosms (Figure 21). Referring to Table 14 and the

118 detection of assA, assA was not present in T=0 samples when assayed with the newly designed primer mix. The abundances reported in Figure 21 reflect either abundances of assA undetected via gel electrophoresis, amplicons whose lengths have been altered by humic acid inhibiting

PCR, or non-specific amplicons. The presence of humic acids and other PCR inhibitors has been shown to change the length of amplicons through binding both the target DNA and the polymerase, with higher concentrations of humic acids shifting melt curves, indicative of variable lengths of amplicons94. PCR inhibitors were not removed during sample preparations to preserve DNA concentrations. It is hypothesized that removal of inhibitors may allow more accurate identification of positive vs. negative detections.

All nutrient amendments (except iron (III)) showed an increase of nearly 1 order of magnitude from initial abundances. Abundances in ‘No HC’ cultures did not change significantly from initial abundances, suggesting that increases in assA were due to HC amendment, not from nutrient amendments. Interestingly, the assA gene abundance decreased by ~0.5 log in ‘Iron

(III)’ amended cultures with added HC, compared to ‘No HC’ and ‘T=0’. Despite significant increase in Fe (II) from reduction of Fe (III) (Figure 20C), and substantial octane loss (Table 13), assA gene abundances decreased. We hypothesize that the primer mix was unable to accurately capture assA genes in these microcosms.

Furthermore, increases in assA gene abundances were seen for all other EA conditions, though there were no significant change in EA concentrations. Should anaerobic HC degradation occur, decreases in EA are expected. Initial HC losses were assumed to be under aerobic conditions, when O2 was depleted at an unknown time. Following O2 consumption, anaerobic processes would begin (as is seen in field studies115). For these microcosms the reamendment of

HC was assumed to be under anoxic conditions and it is during this time that the assA

119 abundances likely increased. Octane loss occurred at a slower rate than initial amendment

(Figure 19 & Table 13). When assessing the changes in EA, there is a general decrease in average nitrate and sulfate, which is hypothesized to change over a greater incubation period

(Figure 20). Thus, these increases in assA from an initially oxic culture may be indicative of a shift towards anoxic conditions, where octane degradation occurred slowly.

C01-04 Microcosm

assA Gene Abundance After 90 Days 5

4 HC No HC

copies/mL GW) copies/mL 3

Gene Abundance Gene

(log assA 2 0 te te n 3 A = a a ro ll E T itr lf I A o N Su N Electron Accepting Conditions

120

5.3.1.2.2 bssA gene detection and quantification.

Comparison of detection of the bssA gene between von Netzer et al. 7772f/8546r primers and newly designed bssA primer mix was conducted and summarized in Table 14. The bssA gene was detected in 3 EA treatments with added HC (nitrate, ‘All 3’, and No EA), and one with no added HC (iron (II)). Detection of the bssA gene using newly designed bssA primer mix resulted in 2 positives with added HC (‘All 3’ and No EA), and one with no added HC (nitrate). This suggests that the bssA gene can be detected using both primer sets and is consistently detected in

‘All 3’ and iron (III) amended cultures with HC. This shows, as was seen in Section 3.4.3.1 that each primer set has its own detection capabilities and may not capture all bssA (or assA) present.

As has been discussed throughout, the appearance of non-specific amplicons is not unique to the newly designed primer mix, as amplicons of varying length were detected when using the von

Netzer 7772f/8546r primers (data not shown).

Initial quantification assays for bssA in field samples (Sections 3.4.3.1.1 & 3.4.3.2.1) were done successfully, however when attempting to replicate those assays for microcosms, issues arose surrounding the behavior of the primers mix (i.e. formation of primer dimers).

Analysis of the melt curve initially suggested a contamination problem (as the melting temperature was above the ~75 oC generally accepted melting temperature for primer-dimers).

Subsequent troubleshooting involved multiple campaigns of: replacing all reagents and disposables, fresh re-dilution of primer mix before assays, re-ordering primers (from different suppliers), keeping primers and qPCR reactions on ice, cleaning of separate qPCR micropipettors, changing location of reaction preparation (to eliminate contamination), having someone else prepare reactions, and multiple emotional breakdowns. It was concluded that the contamination seen was the result of complexes formed by the primers themselves, forming

121

DNA structures more than typical primer dimers. This was extremely puzzling as the bssA gene quantification assays for field samples did not exhibit this issue. After approximately 140 test qPCR NTC reactions, a clean NTC was achieved. However, when transferring to the full assay, primer complexing re-appeared, rendering the assay un-useable. In lieu of time, the quantification of bssA genes in microcosms was abandoned. In hind-sight, Sanger sequencing of the amplicons my have revealed that the amplicons were primer dimers, thereby shortening the troubleshooting time. The reason the primer mix failed seems to be from handling and mixing, where it was found that slightly elevated temperatures from the fingers when handling the tubes expedited complex formation. Although this issue was frustrating and soul-crushingly painful, the nature of the primer mix and combining primers in this manner is open to such issues. This may be the reason it has not been attempted before (to the authors’ knowledge). However, as a proof-of-concept the primer mix works, but can be fraught with complications. The use of this primer mix for quantification was abandoned, however using the primers in presence/absence assays was done in lieu of quantification, as already described.

122

Literature PCR primers Designed qPCR primers HC C01-04 bssA (KAPA Hi-Fi) (BioLine) Added? EA Condition Target Non-Specific Target Non-Specific T = 0 bssA Nitrate bssA Sulfate NSP Yes Iron 'All 3' bssA NSP bssA NSP No EA bssA NSP bssA NSP Nitrate bssA NSP Sulfate No Iron bssA NSP NSP 'All 3' NSP

5.3.1.3 Microbial community composition of C01-04 microcosms.

Comparisons of 16S rRNA gene sequencing data between HC and No HC cultures show an enrichment in the relative sequence abundance of Burkholderiaceae across all cultures in the presence of added HC (Figure 22). The presence and enrichment of Burkholderiaceae strongly suggests that aerobic or microaerophilic mechanisms were dominant in the enrichments, supporting the detection of O2 in at least the first phase of the enrichments. Members of this taxon have been implicated in the efficient degradation of benzene, toluene, and octane using aerobic mechanisms116,117,118. Aerobic alkane degradation via members of Burkholderiaceae has been shown, but for n-alkane chains longer than C9, however it is likely these taxa also degraded

118,119 the C8 alkane added in our cultures . Interestingly, Burkholderiaceae were enriched in ‘HC’ cultures amended with sulfate and iron (III) (>80% of all reads, compared to <30% in no HC cultures), but not under solely nitrate-reducing conditions (no change was seen in ‘No HC’

123 cultures). However, Burkholderiaceae enrichment under strictly anaerobic conditions using nitrate, sulfate, and iron (III) on benzene substrate has not been reported (though Peptococcaceae was suggested as the primary benzene degrader under these conditions, acting syntrophically with Burkholderiaceae)117. Peptococcaceae was not detected in our cultures above 5% relative abundance. These data suggest that members of Burkoholderiaceae are either the main HC degraders present in these cultures, or a metabolite degrader serving in a syntrophic role with other bacteria.

In the No HC amended cultures, a different microbial community composition was present. Polaromonas, a genus in Burkholderiaceae, was generally enriched only in the No HC cultures (Figure 22). Polaromonas is an aerobic, putative naphthalene degrader, which can also degrade benzene and toluene, and is a suspected alkane degrader via cytochrome P450 hydroxylase100–102,120,121. OTUs aligning with Burkholderiaceae most likely harbour other putative HC degrading genera that were not resolved to the genus level. The ferric iron sequestering ability of Nitrosospira in ‘Iron No HC’ may be responsible for that taxon’s enrichment, as no literature suggests Nitrosospira having hydrocarbon degradation capabilities122. Rhodoferax is an Fe(III) and nitrate-reducing facultative anaerobe that can use fumarate, succinate, and benzoate as carbon sources123, and as such may play a role in metabolite degradation from hydrocarbon activation. Rhodoferax has been found in aerobic benzene cultures, but has not been implicated in benzene degradation despite its abundance79. The other taxa identified in ‘All 3 No HC’ are aerobic soil bacteria with no reported hydrocarbon degrading capabilities (Arenimonas, Terrimonas, and Lutibacter)124–126. Overall, the ‘No HC’ cultures were enriched in non-hydrocarbon degrading taxa; while HC-amended cultures saw an enrichment in Burkholderiaceae which are putative HC-degrading taxa.

124

UnresolvedBurkholderiaceae Polaromonas Nitrosospira Rhodoferax

No EA + HC Lutibacter All 3 No HC Microscillaceae OLB12 All 3 + HC Terrimonas Iron No HC Iron + HC Woesearchaeia (Genus) Sulfate No HC Arenimonas Sulfate + HC Nitrate No HC Candidatus Roizmanbacteria Nitrate + HC Latescibacteria (Genus) T=0 C01-04 Field Rhizobacter Pseudomonas 0 50 100 Anaerolineaceae Percent Relative Sequence Abundance GWD2-49-16 Chryseolinea Other (Below 5%)

5.3.1.4 Summary of microcosms prepared with C01-04 groundwater, low contamination

In conclusion, these microcosms exhibited both aerobic and anaerobic degradation of hydrocarbons, with enrichment of putative hydrocarbon degraders (namely, Burkholderiaceae).

After suspected depletion of oxygen and HC re-amendment (day 69), only degradation of toluene

(complete) and octane (~10% compared to controls) was seen under nitrate-, sulfate-, and iron

(III)-reducing conditions. These microcosms demonstrate the potential of HC degradation in these groundwater samples under different electron accepting conditions. Detection of the assA

(quantified) and bssA (not quantified) genes supports the notion that activity of anaerobic

125 hydrocarbon degradation occurred at some time during the incubation, most likely after the initial depletion of oxygen.

5.3.2 Microcosm Well C02-08 – Site A, High Contamination

5.3.2.1 Hydrocarbon degradation

Figure 23 shows the hydrocarbon degradation by well C02-08-associated microcosms, which was the higher contaminated of Site A wells (benzene = 2.24 ppm, toluene = 0.072 ppm, alkanes = 5.5 ppm) used in the microcosm study. No benzene was added, as preliminary GC-FID analysis showed benzene at concentrations similar or higher than planned amendments (data not shown).

Over the incubation period (134 days), very little benzene loss occurred (32.6% in the

‘All 3’ electron acceptor conditions, which is a slightly more than sterile controls). Toluene loss was seen relative to controls in all electron accepting conditions, with the lowest consumption in nitrate-amended (50.5%), and highest in the iron (III) and no electron acceptor conditions (89.4% and 91.4%, respectively) (Table 16). This suggests that the most active toluene degradation occurred under iron-reducing and possibly methanogenic conditions. Octane degradation occurred mainly in the iron (III), ‘All 3’, and No EA cultures which, like toluene, suggests iron- reducing or methanogenic activity (Table 16). However, despite the apparent HC losses seen, initial measurements taken for HC were inaccurate and may be inflating the observed losses (as explained in Section 5.3.1.1). Whether or not the loss of HC measured is due to the direct degradation of the HC contaminants or through a co-metabolism with other carbon sources native in the water sample cannot be determined from these data. Figure 24 shows the changes in electron acceptor concentration between day 1 and day 90.

126

Figure 23: Monitoring of C02-08 sourced microcosms for hydrocarbons (benzene, toluene, and octane) over a 134-day period under various electron accepting conditions (nitrate (A), sulfate (B), iron (III) (C), ‘All 3’ (III) (D) no added electron acceptor ‘No EA’ (E), and sterile controls (F)). Note that benzene and toluene concentrations are on the left y-axis, and octane concentrations are on the right y-axis. Analyses were done by GC-FID. Error bars represent standard deviation of 4 replicates, sterile control for 2 replicates.

127

Again, in these incubations, oxygen measurements (as measured in sterile control headspaces) nearly 100 days after termination showed 28 μM (0.9 mg L-1) oxygen, which would be considered micro-oxic conditions. Oxygen concentrations > 6 μM (0.19 mg L-1) have been reported to inhibit nitrate reduction and are suspected to inhibit sulfate and iron (III) reduction127.

Therefore, the micro-oxic environment of these cultures may have been enough to inhibit anaerobic respiration, which explains the lack of EA change corresponding to HC loss. There may be a slow degradation interplay across all electron accepting conditions, as sorption to sediments is an unlikely cause, relative to sterile controls (which had the same sediments). For all HC-amended cultures conditions, there were few significant changes in EA concentrations

(Figure 24), correlating with the minimal loss of HC between live and control incubations under all conditions (Figure 23).

Table 16: Percent hydrocarbon loss in Site A microcosm C02-08. Shading demarcates hydrocarbon loss greater than sterile controls. Although losses are apparent, initially there were challenges in measuring HC in these microcosms, therefore results are misleading. Percent Loss on Day n Well (Days) EA Treatment Benzene Toluene Octane Nitrate -0.5 50.5 -1.2 Sulfate 5.6 73.6 -42.2 C02-08 Iron 17.0 89.4 87.4 (134) All 3 32.6 82.2 89.7 No EA -122.1 91.4 90.8 Sterile Control 20.9 18.0 45.9

128

5.3.2.2 FAE gene analysis/quantification

Assays for assA or bssA gene abundances were not performed for this sample, as confident HC degradation did not occur to any significant extent.

129

5.3.2.3 Microbial community composition

No EA Rhodoferax All 3 Lutibacter Betaproteobacteriales Iron (Class) Sulfate Azoarcus Nitrate Candidatus Roizmanbacteria (Genus) T = 0 Other (Below 5%) C02-08 Field

0 50 100 Percent Relative Sequence Abundance

High throughput sequencing on these cultures revealed very little diversity in each culture condition and across EA conditions. Note that sequencing was not done on ‘No HC’ cultures as the high HC content natively present in the groundwater did not permit the establishment of proper HC-free controls. Electron acceptor additions or medium additions had no influence on the enrichment of any taxa, where taxa seen in the original field samples are present in virtually identical read abundances in the ‘enriched’ microcosms. The only exceptions are the emergence of Lutibacter in all cultures (except ‘No EA’) and the disappearance of Azoarcus (above 5% read abundance) in all cultures except ‘All 3’. As was mentioned above, the degradation of HC in these cultures may be misleading due to incorrect headspace sampling. The dominant taxon,

Rhodoferax, has not been shown to be directly involved in HC degradation79, but may be involved in downstream HC metabolite degradation. Its primary relevant carbon sources are benzoate, fumarate, and succinate, so degradation of intermediate products resulting from ASS

130 and BSS activity is likely128. Losses of toluene (> 70%) and octane (>80%) were observed in

Sulfate, Iron (III), All 3, and No EA cultures, all of which have unresolved

Betaproteobacteriales present (Figure 25)§§§, but again this loss data in uncertain. Members of the former Betaproteobacteria class have been well characterized as nitrate-reducing hydrocarbon degraders, including Burkholderiaceae as described in the previous section (Section

5.3.1.3)68,129,130. Interestingly, unresolved Betaproteobacteriales are not present in nitrate- amended cultures. Rhodoferax is also a genus in Betaproteobacteriales but has been resolved to the genus level. It is likely that other genera are present in the Betaproteobacteriales reads and may be implicated in direct HC degradation (or identity has been attributed to Rhodoferax incorrectly). Although these data do not clearly point to active HC degradation, subsequent transfers and maintenance to further enrich organisms and remove sediments may paint a clearer picture of the dynamics of higher contaminated wells in Site A and similar sites.

5.3.3 Summary of Microcosms prepared with C02-08 groundwater, high contamination

Hydrocarbon degradation was not convincingly demonstrated in these incubations.

Toluene losses appeared to have occurred, but may have been due to measurement errors. Few significant changes in electron acceptor concentrations or microbial community composition occurred during this incubation period. The elevated concentrations of hydrocarbons natively present in these groundwater samples is most likely the reason for this lack of microbial activity.

This is in support of the ‘plume fringe concept’ where lower microbial activity is seen where

§§§ SILVA 132 update came with a drastic re-arrangement and reclassification of the Proteobacteria from Parks et al. forming a new taxonomic database, Genome Taxonomy Database (GTDB), where Betaproteobacteria are now classified under Gammaproteobacteria140. Thus, all taxonomic ranks thereafter are skewed relative to taxonomy assigned using previous versions of SILVA or other non-SILVA using databases and literature. The class Betaproteobacteriales is thus non-existent in literature pre-2018.

131 hydrocarbon concentrations are higher21. However, a notion in that concept is the limited availability of electron acceptors tied with the higher concentration of hydrocarbons, due to higher microbial activity in the plume edge using resources. These data seem to suggest that hydrocarbon concentrations are the greater inhibiting factor, as ample EA were provided. It would be interesting to see if transferring these microcosms (i.e. removing the hydrocarbons to less than 200 μM) would result in increased microbial activity, thus giving evidence for a toxic/dormant threshold. Regardless, these microcosms suggest that a toxic/dormancy threshold, where increased hydrocarbon concentrations inhibit activity, is in effect.

5.3.4 Microcosm Well REC 12 – Site B, Low Contamination.

5.3.4.1 Hydrocarbon degradation.

Figure 26 shows the HC degradation in microcosms prepared from groundwater sample

REC 12 representing a ‘low-contaminated’ sample from Site B (2016 [HC]: benzene = 2.35 ppm, toluene = 0.0299 ppm, alkanes = 4.7 ppm). Over the 121-day incubation period no significant benzene was consumed, compared to sterile controls. The presence of toluene has been shown to inhibit the biodegradation of benzene in both aerobic and anaerobic cultures, where benzene loss would occur after toluene loss107,108. Toluene losses under nitrate-, sulfate-, and ‘All 3’-amended conditions were >98.7%, parallel to octane losses under the same conditions (Table 17). Little toluene or octane loss was seen under iron-reducing and methanogenic conditions, matching abiotic loss in sterile controls (Table 17). These data suggest that nitrate- and sulfate- reducers are the largest contributors to toluene and octane degradation in these microcosms. This is supported by the changes in EA concentrations from day 1 to day 90 (Figure 27). In nitrate and

HC-amended cultures (nitrate and ‘All 3’), nitrate decreased by more than 1.98 mM (nitrate) and

132

2.35 mM (All 3), which is significant compared to controls (Figure 27), but less than the theoretical nitrate requirement for the complete degradation of BTO (4.44 mM) (Table 18).

Changes in sulfate concentrations were similar to nitrate, with decreases greater than 1.19 mM in sulfate only amended cultures, and 0.62 mM in ‘All 3’ (sulfate decrease was similar in ‘No HC’ controls, which suggests sulfate reduction may have occurred on other carbon sources).

133

Figure 26: Monitoring REC 12 sourced microcosms for hydrocarbons (benzene, toluene, and octane) over a 121-day period under various electron accepting conditions (nitrate (A), sulfate (B), iron (III) (C), ‘All 3’ (III) (D) no added electron acceptor ‘No EA’ (E), and sterile controls (F)). Note that benzene and toluene concentrations are on the left y-axis, and octane concentrations are on the right y-axis. Analyses were done by GC-FID. Error bars represent standard deviation of 4 replicates, sterile control for 2 replicates.

134

Table 17: Percent degradation of each amended hydrocarbon in Site B sourced microcosms from REC 12. Percentages are calculated from calculated average concentrations of 4 replicates from Day 1 and Day n, where n = total incubation time in days. Shaded formatting denotes a percentage degradation greater than that of the sterile control. Percent Loss on Day n Well (Days) EA Treatment Benzene Toluene Octane Nitrate 23.0 98.7 87.9 Sulfate 36.1 99.2 99.6 REC 12 Iron 10.2 59.2 42.5 (121) All 3 34.3 99.7 99.3 No EA 30.9 61.1 60.9 Sterile Control 37.5 61.6 51.1

A significant increase in Fe (II) of 1.07 mM in Fe (III) amended cultures was measured

(Figure 27C), however this was far less than the expected 5 mM increase (for complete reduction of added 5 mM iron (III)) due to HC loss through iron reduction (see Table 18). However, HC loss beyond abiotic loss in sterile controls was not seen in Fe (III)-only amended cultures. Thus, the changes in Fe (II) concentrations seen are possibly due to non-HC degrading organisms, or very slow growing HC degraders.

In total, EA loss in ‘All 3’ cultures was 3.17 mM (2.35 mM nitrate, 0.62 mM sulfate, and

0.2 mM iron (III)) which was equivalent to total sulfate loss in sulfate only cultures, and more than the total nitrate loss in nitrate only cultures. Linking the loss of toluene and octane and a corresponding (but non-proportional) loss of EA strongly suggests that anaerobic HC degradation under nitrate, sulfate, iron (III) did occur under single EA and ‘All 3’ treatments.

Taking theoretical losses of degradation of only toluene and octane, the only HCs which showed degradation compared to controls, the EA losses reach close to theoretically-predicted values,

(single amendments: nitrate = 103.4%, sulfate = 99.8%, Iron (II) = 11.2%; and in ‘All 3’ nitrate

= 122.7%, sulfate 51.8% and iron (II) = -2.1%) (Table 18). This shows that the degradation of

135 toluene and octane is reflected in EA concentrations changes, rather than the degradation of benzene, which was not seen.

The discrepancy in theoretical values and actual values loss could be due to initial use of oxygen present in the cultures (as was seen for the Site A sample C01-04, Section 5.3.1.1), which changed to anaerobic nitrate reducing conditions after O2 depletion. However, initial oxygen concentrations in REC 12 microcosms were negligible (0.14 mg L-1 (4.55 μM)) and was below anaerobic respiration inhibition levels of some microorganisms127. These cultures can thus be considered anoxic. As with all other microcosms in this study, final oxygen concentrations are negligible. It is most likely that oxygen was initially present, it was consumed very quickly, yielding anoxic conditions earlier, so that substantial anaerobic electron acceptor changes could be seen. Alternatively, it could be that aerobic organisms were performing the initial aromatic and alkane activation steps (e.g. ring opening and/or hydroxylation steps, which are faster than anaerobic fumarate addition)18, leaving further degradation to anaerobic microorganisms that became enriched after oxygen depletion or in anoxic biofilms/regions of the microcosm. This interplay between aerobic and anaerobic degradation has been reported previously78,131 and are very likely mechanisms present in the other microcosms reported in this study.

136

137

Table 18: Comparison of theoretical electron acceptor changes with empirical changes. Note, shading represents percent theoretical changes >25% in Benzene+Toluene+Octane, and >50% in only Toluene+Octane. Benzene, Toluene, Octane Degradation Toluene, Octane Degradation Culture Electron Well Theoretical Actual Loss Theoretical Theoretical Actual Loss Theoretical Group Acceptor Loss (mM) (mM) Loss (%) Loss (mM) (mM) Loss (%) Nitrate 4.44 2.0 44.6 1.916 2.0 103.4 Single EA Sulfate 2.78 1.2 43.0 1.197 1.2 99.8 Iron (II) 22 1.1 4.9 9.582 1.1 11.2 REC 12 Nitrate 4.44 2.4 52.9 1.916 2.4 122.7 All 3 Mix Sulfate 2.78 0.6 22.3 1.197 0.6 51.8 Iron (II) 22 -0.2 -0.9 9.582 -0.2 -2.1

5.3.4.2 FAE gene abundance.

5.3.4.2.1 assA gene detection and quantification.

Detection of the FAE genes in groundwater sourced from the Site B REC 12 well was compared using literature primers 7757af/8543r46 with KAPA HiFi polymerase, and newly designed primers with BioLine SensiFast qPCR polymerase. Gel electrophoresis analysis of amplicons supports that the assA gene is found in almost all microcosm samples, and evidence of non-specific amplicons was also seen (Table 19). Repeating the detections assay with von Netzer et al. primers 7757af/8543r also showed the same result, except in 2 samples (No EA + HC and sulfate No-HC). Thus, in these microcosms, the assA gene is consistently detected across these tested primers. It is most promising that the assA gene is detected in almost all microcosms, supporting the anaerobic hydrocarbon (octane) loss described previously.

138

Table 19: Comparison of assA gene detection using literature primers 7757af/8543r and newly designed qPCR primers. ‘assA’ denotes detection of the gene. ‘NSP’ denotes presence of non-specific bands’ either larger or smaller than expected amplicon. Blank spaces mean no detection.

Literature PCR primers Designed qPCR primers HC REC 12 assA (KAPA Hi-Fi) (BioLine) Added? EA Condition Target Non-Specific Target Non-Specific T = 0 assA assA Nitrate assA assA NSP Sulfate assA assA NSP Yes Iron assA assA NSP 'All 3' assA assA NSP No EA assA Nitrate assA assA NSP Sulfate assA No Iron assA assA 'All 3' assA assA NSP

REC 12 Microcosm assA Gene Abundance After 90 Days 139 6

5 HC No HC

copies/mL GW) copies/mL 4

Gene Abundance Gene

(log assA 3 0 te te n 3 A = a a ro ll E T itr lf I A o N Su N Electron Accepting Conditions

Using the newly designed qPCR primer mix assaying assA gene abundance compared to field abundances, only ‘All 3’ showed any significant increase (‘T=0’ = log 4 copies/L, ‘All 3’ = log 5 copies/L). All other EA amendments had decreased assA abundance after 90 days, with the lowest in sulfate and No EA microcosms (~log 3 copies/L). Nitrate and sulfate ‘No HC’ amendments had higher assA abundance than cultures with added HC. Abundance in ‘All 3’ cultures was 10-fold greater than all other treatments, which is surprising as electron acceptor decreases, and percent HC loss were matched or exceeded by single nutrient cultures. Both nitrate and sulfate amended cultures with added HC have assA abundances less than ‘No HC’ abundances. This result is unexpected and contrary to trends seen in other treatments and in

140 microcosms from Site A (Section 5.3.1.2.1), where HC amended have greater assA abundance due to a substrate (octane) dependent response. Gene abundance of assA in these microcosms shows that iron (III) amended cultures have a greater assA abundance than nitrate and sulfate, both of which had greater octane loss (87.9% and 99.6%, respectively; compared to iron amended loss at 42.5%, less than sterile controls) (Figure 28 & Table 17). This result is surprising as iron (III) cultures saw no HC loss beyond sterile controls, and yet have greater assA abundances. The assA quantification is counter to expectations, especially in nitrate and sulfate microcosms where higher assA abundances were expected. It is likely, given these data, that the increase in assA abundance is due to amplification of an assA homolog, as melting temperatures and melt curves are slightly different between ‘All 3’ (melting temp = 88.5oC, single peak) and nitrate/iron/No EA (melting temp 89.5oC); sulfate has presence of non-specific binding (melt temps at 85 and 88oC) (Figure 29). Non-target binding is unlikely in this case, as that usually implies amplicon products of longer/shorter lengths (and higher or lower melting temperatures, respectively) which was seen only in sulfate microcosms when amplicons were checked via gel electrophoresis (data not shown). Illumina MiSeq sequencing was inconclusive in determining the extent of non-specific binding (Section 4.3.4). Another explanation of this shift in the melting curve is a probable action of humic acids, which have been found to bind DNA, shifting the melting curve in qPCR applications94. Heterogeneous ‘solid’ particles present in each microcosm may have differing humic acid concentrations, affecting the melt curve. Another explanation is that a different, non-fumarate addition HC degradation mechanism may be functioning. Here,

HC would still be degraded, but an increase in FAE would not occur. However, considering the abundances of T=0, HC-amended, and No HC-amended cultures were within the same order of magnitude, the REC 12 groundwater sample may already have been enriched for anaerobic

141 hydrocarbon degradation. In this case, the addition of nutrients and extra EA may not influence

FAE gene abundance.

5.3.4.2.2 bssA gene abundance and detection using bssA primer mix

As was described in Section 5.3.1.2.2, the complications faced during quantification of the bssA gene using the newly designed qPCR primer mix in C01-04 sourced microcosms were also seen here. An explanation is given in that section. The bssA primer mix was used to detect, rather than quantify, the bssA gene.

Comparison of detection capability of von Netzer et al. 7772f/8546r46 and newly designed bssA primer mix (as was done in Section 5.3.1.2.2 bssA for Site A groundwater sourced microcosm C01-04) was done for REC 12 microcosms and is summarized in Table 20. Using the

142 newly designed primer mix, the bssA gene was detected in all HC-amended treatments (except in

No EA cultures) and in the ‘no HC amended’ cultures with iron and nitrate added (nitrate, iron, and ‘All 3’). Literature primers were able to consistently detect the bssA gene in all HC un- amended cultures, but not in cultures with added HC, despite the degradation activity seen

(Figure 26). This gives further evidence for the apparent anaerobic nature of these microcosms.

Detection of bssA using the newly designed bssA primer mix suggests these new primers perform comparably to the literature primers, as was found in previous detection assays. However, the failure to detect the bssA gene in HC-amended cultures using literature primers is interesting. It may be that a difference bssA sequence was enriched under those conditions, that are not detected using the literature primers, but were detected using the newly designed qPCR primers.

Sequencing of those bssA amplicons specifically was not done but would provide information as to what types of bssA are being enriched that are eluding those primers. Consistent detection of the bssA gene across the electron-accepting conditions suggests that these microcosms were generally enriched for anaerobic hydrocarbon degradation, as was seen in the hydrocarbon degradation observed (Figure 26 & Table 17). This finding contrasts with the microcosms from

Site A sourced well C01-04, which was suspected of aerobic degradation. Thus, such bssA gene offers evidence that the REC 12 sourced microcosms were fostering anaerobic mechanisms. The ability of the newly designed primers to consistently capture the bssA gene in a demonstrably anaerobic microcosm is very promising.

143

Literature PCR primers Designed qPCR primers HC REC 12 bssA (KAPA Hi-Fi) (BioLine) Added? EA Condition Target Non-Specific Target Non-Specific T = 0 bssA NSP bssA Nitrate bssA bssA NSP Sulfate bssA Yes Iron bssA bssA NSP 'All 3' NSP bssA No EA Nitrate bssA NSP bssA Sulfate bssA No Iron bssA bssA NSP 'All 3' bssA bssA

5.3.4.3 Microbial community composition of REC 12 groundwater sample.

Illumina sequencing of the 16S rRNA gene and taxonomy assignment through MetaAmp

(as described in Section 3.2.6) showed an enrichment of Methanosaeta in REC 12 microcosms.

Methanosaeta, an archaeal methanogen, was enriched during the preparation phase where bottle were kept at 4oC under native conditions, as is seen in the differences between ‘REC 12 Field’ and ‘T = 0’ (Figure 30). Regardless of the subsequent EA addition, the dominance of

Methanosaeta persisted, even under known methanogenesis inhibiting electron acceptor conditions (iron (III), nitrate, and sulfate). Methane measurements were conducted for these microcosms, and methane was detected in only sulfate (8.14 µM ± 14.1), ‘All 3’ (49.7 µM ±

63.1), and No EA (15.34 µM ± 15.5) conditions. These data suggest that the methanogenic activity of these cultures are not well established and for an unexplained reason the inhibitory effect of the iron (III) and sulfate in `Sulfate and ‘All 3’ conditions had an inconsistent effect on

144 the production of methane (as seen by standard deviations that are greater than averages). The concentrations of produced methane were low, and errors suggest that the methanogenic community/production is not consistent across all the replicates (n = 3 for methane measurements, n = 1 for Illumina sequencing) (data not shown). Methane was not detected in

‘Nitrate’ and ‘Iron (III)’ microcosms, which does suggest that methanogenic activity was inhibited, even though methanogenic taxa were of high relative abundance. However, the detection of methane in the ‘All 3’ microcosm suggests that the growth of either Methanosaeta or other taxa overcame that inhibitory effect (data not shown). As Figure 30 shows, the enrichment of acetotrophic Methanosaeta may play a key role in the degradation of alkanes as an acetate-using methanogen in partnership with syntrophic alkane-degraders. Deltaproteobacteria were not detected above 5% read abundance under any nutrient condition. As such, the dominance of acetotrophic Methanosaeta may be a direct result of acetate production through

HC degradation via fumarate addition. However, the taxa directly involved in the HC-activation process may have an uncharacterized assA gene or is part of ‘Other (Below 5%)’ in microcosms, or, as mentioned earlier, a HC-degrading pathway other than fumarate addition may be occurring. Other taxa present in these cultures, such as Rhodoferax and Polaromonas, may be playing a role in the degradation of by-products from anaerobic HC activation78,79. Another interesting taxon present in the microcosms was identified as Candidatus Roizmanbacteria, present in ~10% read abundance in ‘All 3 HC’ cultures which was seen to be the only microcosm with increased assA abundance (Figure 30). Candidatus Roizmanbacteria, an uncharacterized hot springs/soil taxon, has not been characterized metabolically (though has a

CRISP-Cas system132). Whether the increased assA abundance is due to Candidatus

Roizmanbacteria cannot be determined at this time.

145

Methanosaeta Rhodoferax No EA + HC Polaromonas All 3 No HC All 3 + HC Candidatus Iron No HC Roizmanbacteria (Genus) Iron + HC Betaproteobacteriales Sulfate No HC (Class) Sulfate + HC Nitrate No HC Geobacter Nitrate + HC Azovibrio T = 0 Latescibacteria (Genus) REC 12 Field Dysgonomonadaceae 0 50 100 (Family) Percent Relative Sequence Abundance Other (Below 5%)

5.3.5 Summary of Microcosms prepared with REC 12 groundwater, low contamination

Microcosms prepared from Site B REC 12 groundwater were capable of degrading toluene and octane under nitrate- and/or sulfate- reducing conditions (nitrate, sulfate, and ‘All

3’). Benzene degradation was not seen during the 121-day incubation. These conditions also saw a stoichiometric change in electron acceptors, which suggests that anaerobic processes are responsible for the degradation activity. Some change in iron (II) concentrations occurred, but hydrocarbon loss was not seen greater than abiotic loss in sterile controls. Surprisingly, FAE gene abundance did not reflect the degradation of octane seen in single electron acceptor- amended cultures, where abundances were similar to that of ‘No HC’ controls. FAE genes in

‘All 3’ saw the largest increase relative to ‘No HC’ controls. Methanogenic activity of these cultures was inconsistent, however detection of a methanogenic taxon may suggest that some

146 activity may be occurring, perhaps in pockets of the culture volume where electron acceptors that inhibit methanogenesis are less abundant. These data strongly suggest that anaerobic hydrocarbon degradation is occurring and can be stimulated under nitrate and/or sulfate reducing conditions. Monitoring of FAE gene abundance in response to hydrocarbon load did not reflect hydrocarbon losses and electron acceptor changes, however that is most likely a limitation of the primer mix itself. These microcosms seem to be the most reflective of anaerobic hydrocarbon degradation of all the microcosm experiments established in this study.

5.3.6 Microcosm Well REC 34 – Site B, High Contamination

5.3.6.1 Hydrocarbon degradation.

Figure 31 shows HC degradation from REC 34 sourced water from Site B (2017) over

120 days. For all REC 34 microcosms, no benzene, toluene, or octane were amended at any time, thus there is no difference between ‘HC’ and No HC’. Very little degradation occurred of any

HC under any electron accepting condition at levels greater than the abiotic losses seen for sterile controls. Hydrocarbon loss greater than sterile controls only occurred under No EA

(methanogenic) conditions. From these degradation data, it suggests that methanogenic mechanisms for degradation of octane, toluene, and benzene may have been active and may have been inhibited by the addition of other electron acceptors133–135, as was seen in REC 12.

However, methanogenic taxa were not detected in under any EA condition (Figure 33). Only octane saw loss greater than sterile controls under nitrate-reducing (54%) and iron-reducing

(42.8%) conditions (Table 21).

Percent loss seen under ‘No EA’ conditions in Table 21 is highly suspect as upon closer inspection of Figure 31E, initial HC concentrations are abnormally high during the first 3

147 measurement times and drop drastically thereafter; therefore, percent loss calculated from those values are of low confidence. This observation is the same for all nutrient conditions tested, where initial HC concentrations are very high, decrease drastically, and fail to decrease thereafter. An explanation is given for this effect in Methods Section 5.2.1.1. Initial concentrations of benzene and toluene in these microcosms were 2 – 3-fold greater than other microcosms in this study (100 – 300 μM each HC), while octane concentrations were similar (~

3 μM). Although previous studies have demonstrated inhibitory BTEX concentrations upwards of 2 mM BTX19, it is possible that concentrations >300 μM BTO (in addition to other compounds) could also impede microbial activity during this incubation period. Thus, we do not believe that anaerobic hydrocarbon degradation occurred in these microcosms during this incubation period.

148

Changes in electron acceptors were negligible across all nutrient treatments, with no significant changes after 90 days (Figure 32). As there was no difference in the HC and No HC cultures as was the setup in the previous microcosms, EA data is presented as HC-amended with

149

6 replicates. Only nitrate concentrations changed significantly in both Nitrate and ‘All 3’ cultures. This may reflect some loss of octane seen in nitrate cultures, but is more likely from other, non-hydrocarbon degradation mechanisms. No other change in electron acceptor concentration was seen. O2 measurements showed that these microcosms were generally anoxic

(oxygen = 0.02 ± 0.14 mg/L (n = 3)). This supports that observation that anaerobic biodegradation did not occur in a significant way in these microcosms.

Table 21: Percent degradation of each amended hydrocarbon in Site B sourced microcosm REC 34. Percentages are calculated from calculated average concentrations of 4 replicates from Day 1 and Day n, where n = total incubation time in days. Shaded formatting denotes a percentage degradation greater than that of the sterile control. Percent Loss on Day n Well (Days) EA Treatment Benzene Toluene Octane Nitrate 22.1 42.0 54.0 Sulfate 9.2 18.8 22.5 REC 34 Iron 21.1 38.3 42.8 (120) All 3 12.3 25.0 19.3 No EA 67.4 72.8 75.9 Sterile Control 31.8 40.4 24.2

150

5.3.6.2 FAE gene abundance.

As HC biodegradation activity was not seen in these cultures, assA and bssA quantification assays were not done.

151

5.3.6.3 Microbial community composition.

Illumina sequencing of these cultures showed that methanogenic taxa are not present. As such, despite the apparent degradation of octane, organismal support for methanogenic activity is not present in these microcosms. It is not likely that methanogenic taxa were missed during sequencing as they were detected in REC 12 microcosms. Similar read abundances of

Rhodoferax were seen in field samples, and in all microcosm conditions, which can utilize both nitrate and iron (III), however no change in those EAs was seen (Figure 33). After preparatory incubation under 4oC there was an enrichment of Dojkabacteria which did not change after EA incubation (function of Dojkabacteria is unknown). Changes in microbial community composition was only seen under nitrate reducing conditions with detection of suspected HC degraders in Betaproteobacteriales (which includes Burkholderiaceae, Nitrosomonadaceae, and

Rhodocyclaceae). Other taxa were also detected: Dysgonomonadaceae (aerobic polysaccharide degrader136), and Xanthomonadaceae (aerobic taxa seen in PAH degradation137,138). Aerobic organisms were detected in these cultures, despite anoxic conditions. This may be a result of enrichment from the native field community and not necessarily a product of possible slight HC degradation.

152

No EA Rhodoferax All 3 Dojkabacteria WS6 (Genus) Iron Betaproteobacteriales (Class) Sulfate Dysgonomonadaceae Nitrate (Family) T = 0 Xanthomonadaceae (Family) REC 34 Field Geobacter Burkholderiaceae (Family) 0 50 100 Other (Below 5%) Percent Relative Sequence Abundance

5.3.6.4 Summary of Site B microcosm REC 34, higher contamination.

These data show that higher concentrations of HC (~300 μM total BTO) inhibited hydrocarbon degradation activity to a large extent (aerobic or anaerobic). This was reflected in the lack of electron acceptor concentration changes, or proliferation/detection of putative hydrocarbon degraders. This microcosm experiment was somewhat disappointing in that degradation activity was not seen, however it suggests that HC concentrations above 100 μM (as in other microcosms) is too high to see any HC degradation under these conditions, which is a concentration much lower than previously reported19. It is most unlikely that degradation would occur under field conditions. This microcosm served as an example of the upper limit of potential bioremediation assays, where sites with hydrocarbon concentrations are too high are not appropriate for this method of remediation.

153

5.4 Conclusions

As this microcosm study yielded highly variable results in terms of HC biodegradation potential and the 2 study sites, a summary of the major results of each microcosm setup is given in Table 22. These data suggest that hydrocarbon degrading communities are present and active in Site A and Site B, and are demonstrably more active in the lower contaminated groundwater samples (C01-04 and REC 12). Activity of these communities appears to be linked to the hydrocarbon concentrations present at these sites, where BTO concentrations above ~400 μM were inhibitory and/or show decreased degradation of toluene and octane. While these inhibitory effects are seen at HC concentrations lower than reported in literature (total BTX concentration

~150 mg/L, equivalent to ~1.9 mM of benzene)19, it is possible that some inhibition is occurring with BTO compounds and other HCs as well. These hydrocarbon degradation experiments under laboratory conditions also exemplified the need for testing of contaminated sites, as not every site has the same hydrocarbon degrading potential or capabilities18,107. This was demonstrated in the differences in the degradation between microcosms from Site A (C01-04) and Site B (REC

12). Both microbially active microcosms showed significant HC loss, with different enriched microbial communities. Electron acceptor utilization at these sites was diverse with cultures using nitrate, sulfate, and/or iron (III). C01-04 demonstrated both aerobic and anaerobic (Fe (III)- reduction) hydrocarbon degradation, while REC 12 showed primarily anaerobic (nitrate-, sulfate-

, and iron (III)-reduction) and methanogenic hydrocarbon degradation. Also, both the assA and bssA genes (explicitly anaerobic) were detected and quantified under these initially aerobic microcosms (C01-04 showed increases in assA abundances when challenged with HC, while

REC 12 showed decreases in abundance (despite being under more anoxic conditions)). This suggests that anerobic processes flourished after oxygen depletion. The mixed community of

154 aerobic and anaerobic organisms suggests that an interplay of aerobic and anaerobic processes can exist in contaminated aquifers.

Laboratory testing of higher and lower contaminated groundwater samples showed that under the conditions tested, HC degradation is a feasible and active process which may be influenced with additional EA and limiting nutrients. The use of assA gene quantifications to monitor the extent of HC degradation potential can be used as a method to investigate a contaminated site and to monitor the activity of anaerobic hydrocarbon degradation in laboratory microcosms. Quantifying the bssA gene was not done due to some difficulty and should be explored in the future. Finally, though some results in this microcosm study were not fully understood (or seemed paradoxical), the study has led to the establishment of new BTO enrichment cultures that are worth further studying (e.g., further propagating) to better understand processes of the anaerobic degradation of these model hydrocarbons.

155 REC34 REC12 C02-08 C01-04 WellSource

Site B Site B Site A Site A Site

= ** = N/A asapplicable not

Table Added: High Added: Low Added: High Added: Low Contamination

None n-Octane Toluene Benzene n-Octane Toluene n-Octane Toluene Benzene

bssA

genequantified not was : Summary : major fromof microcosm findings each Octane onlyOctane BTO onlyOctane Octane + Toluene Octane + Toluene TO - amendment 2nd BTO - amendment 1st Degraded HC greater Degraded Iron Nitrate EA No EA No 3' 'All Sulfate Nitrate measurements initial inaccurate from Likely conditions All conditions All

thanabiotic controls

assay

was not done not was Nitrate (III)Fe Sulfate Nitrate Nitrate (III)Fe O Utilized

EA N/A Yes N/A Yes assA 'All 3' 'All Iron Sulfate Nitrate 3' 'All Sulfate amended?

inHC-

detected detected N/A Slightincrease time 0 from time 0 from log ~1 decrease N/A log ~1 increase in HC- 'All 3' 'All Iron Sulfate Nitrate EA No 3' 'All Sulfate Nitrate assA

quantified? discussed in this Chapter. this in discussed N/A Yes N/A Yes bssA 'All 3' 'All Iron Sulfate Nitrate EA No 3' 'All amended? inHC- detected detected * N/A N/A N/A N/A quantified?

bssA bssA

enrichment HC-amended No inWS6 all Dojkabacteria Rhodoferax enrichment HC-amended No cultures Rhodoferax and Methanosaeta taxaseen enrichment No of HC Polaromonas inHC-amended Burkholderiaceae Enrichedtaxa

cultures and and inall in No Slight anaerobic Yes, Slight anaerobic and aerobic Yes, bioremediation bioremediation Potential for for Potential established?

156

Chapter 6: Conclusions and Future Directions

The work of this thesis was to evaluate the bioremediation potential of two hydrocarbon contaminated sites via anaerobic hydrocarbon degradation. This was done through three main objectives; the analysis of field site samples (Chapter 3), detection and quantification of FAE genes using literature and newly-designed primer sets (Chapter 4), and the establishment and monitoring of hydrocarbon-containing microcosms (Chapter 5). These three objectives allowed a multifaceted view of the hydrocarbon bioremediation potential of these sites.

Chapter 3 discussed the use of signature metabolite detection and the detection of the FAE genes in assessing the potential for bioremediation in collected field samples from two field sites

(Site A and Site B) over a two-year period. Detection of appropriate signature metabolites gave unequivocal evidence for previous or on-going anaerobic hydrocarbon degradation. Paired with the detection of both FAE genes, assA and bssA, the potential for bioremediation was established. However, a co-occurrence of the FAE gene and the appropriate signature metabolite was not always present. assA was consistently detected in field samples when alkylsuccinate signature metabolites were detected, while the detection of both bssA and benzylsuccinates was not as consistent. However, it was speculated that the transient nature of the metabolite and the dynamics of groundwater flow influences the detection of both signature metabolite and the FAE gene. Analysis of the microbial community composition was inconclusive with respect to enrichment of certain HC-degrading taxa. Further work in quantifying the signature metabolite concentrations present in these field samples may give insight into microbial activity in higher versus lower contaminated samples.

In Chapter 4, qPCR primer design to quantify the FAE genes in the field samples (Chapter

3) was conducted. This work was done as a proof-of-concept for the design of forward primer

157 mixes in order to capture a greater diversity of FAE gene sequences, while having qPCR compatibility. Quantification of the FAE genes was done successfully for the field samples. The work also sought to determine whether a link could be established between field hydrocarbon concentrations and FAE genes abundances. For the assA gene, there appeared to be a link to alkane concentrations, wherein the gene was detected when up to ~5 ppm alkanes were present, past which the assA abundance did not increase. This shows that higher alkane concentrations do not, at least in these samples/fields, result in higher assA abundances. In contrast, such a trend was not seen between bssA abundances and alkylbenzene (TEX) concentrations. Furthermore, assA and bssA amplicon sequencing from these primer mixtures suggested that the captured diversity is on-par with (non-qPCR) primer sets frequently used in literature. Future work should be done in furthering/optimizing the use of a primer mixture in capturing diversity, and in designing primer mixture guidelines to minimize the possibilities of primer dimerization.

However, as a proof-of-concept for the use of a primer mixture as presented here, this work was successful in detecting and quantifying the FAE genes in the field samples under study.

In Chapter 5, microcosms were established to monitor bioremediation under different EA conditions (nitrate, sulfate, iron (III), ‘All 3’ (nitrate + sulfate + iron (III)), and no added EA), using two field samples (highly contaminated and low contaminated), from each study site.

These microcosms were monitored for the loss of added HCs (benzene, toluene, and octane), changes in EA, changes in microbial community composition, and the abundances of the FAE genes. It was found that the microcosms prepared from the lower contaminated samples were most microbially active, showing significant HC loss over the incubation period. Additionally, it was speculated that aerobic hydrocarbon degradation mechanisms were initially active and yielded to anaerobic mechanisms after oxygen was depleted. This phenomenon was seen most

158 predominantly in the lower contaminated groundwater sample from Site A (C01-04). This was further supported by the enrichment of aerobic putative HC-degrading taxa and minimal EA changes over the incubation period. Interestingly, increases in assA gene abundances were reported, despite plausible initial aerobic degradation. Anaerobic mechanisms were found to be more dominant in the lower contaminated groundwater sample from Site B (REC 12), which was supported by an enrichment of methanogenic taxa, with significant changes in EA concentrations. assA quantifications in these microcosms did not show substantial enrichment of assA, despite evidence of anaerobic hydrocarbon degradation activity. Quantification of bssA was not done due to difficulties with that assay. In contrast to the microcosms prepared from the lower contaminated groundwater from both sites, microcosms prepared from the higher contaminated groundwater samples failed to show any evidence of anaerobic hydrocarbon degradation. It was concluded that higher concentrations of HC inhibit HC degradation. Future work in the maintenance and transfer of the successful microcosms (e.g., established from the lower contaminated samples) should be done. Establishment of possible anaerobic or aerobic benzene/toluene/octane cultures would be beneficial, to determine which taxa are responsible for the degradation seen in the lower contaminated microcosms, and if any such HC-degrading taxa can be enriched in the samples which came from higher contaminated samples. Repeating the bssA quantification assay to determine if a trend is present between bssA abundance and EA conditions would complete this study (which was not done in lieu of methodological challenges encountered and time).

As a whole, the work presented here was done to determine whether two hydrocarbon contaminated sites located in Alberta, Canada have the potential to be bioremediated. Multiple approaches were used in effort to give a comprehensive outlook on anaerobic hydrocarbon

159 degradation, combining microbial community analyses, signature metabolite analyses, and the detection and quantification of FAE genes. Combining these methods into laboratory microcosms with different electron accepting conditions was done to demonstrate/evaluate the biodegradation activity in real time. Multifaceted analyses of hydrocarbon degradation activity should be done more frequently to assess hydrocarbon contaminated sites for their bioremediation potential. Although studies in anaerobic hydrocarbon degradation have been done previously, nuances in environments and advancement in methodologies allows for better informed implementation of bioremediation in real-world applications. This work was done in efforts to add to that understanding.

Thank you.

160

Chapter 7: References

1. Khan, F. I., Husain, T. & Hejazi, R. An overview and analysis of site remediation

technologies. J. Environ. Manag. 71, 95–122 (2004).

2. Kimes, N. E. et al. Metagenomic analysis and metabolite profiling of deep-sea sediments

from the Gulf of Mexico following the Deepwater Horizon oil spill. Front. Microbiol. 4,

(2013).

3. Hilpert, M., Mora, B. A., Ni, J., Rule, A. M. & Nachman, K. E. Hydrocarbon release

during fuel storage and transfer at gas stations: environmental and health effects. Curr.

Environ. Heal. Rep. 2, 412–422 (2015).

4. Lynge, E. et al. Risk of cancer and exposure to gasoline vapors. Am. J. Epidemiol. 145,

449–58 (1997).

5. Gan, S., Lau, E. V. & Ng, H. K. Remediation of soils contaminated with polycyclic

aromatic hydrocarbons (PAHs). J. Hazard. Mater. 172, 532–549 (2009).

6. Gitipour, S., Sorial, G. A., Ghasemi, S. & Bazyari, M. Treatment technologies for PAH-

contaminated sites: a critical review. Environ. Monit. Assess. 190, (2018).

7. Logeshwaran, P., Megharaj, M., Chadalavada, S., Bowman, M. & Naidu, R. Petroleum

hydrocarbons (PH) in groundwater aquifers: An overview of environmental fate, toxicity,

microbial degradation and risk-based remediation approaches. Environ. Technol. Innov.

10, 175–193 (2018).

8. Nadim, F., Hoag, G. E., Liu, S., Carley, R. J. & Zack, P. Detection and remediation of soil

and aquifer systems contaminated with petroleum products: An overview. J. Pet. Sci. Eng.

26, 169–178 (2000).

9. Chandra, S., Sharma, R., Singh, K. & Sharma, A. Application of bioremediation

161

technology in the environment contaminated with petroleum hydrocarbon. Ann.

Microbiol. 63, 417–431 (2013).

10. O’Brien, P. L., DeSutter, T. M., Casey, F. X. M., Wick, A. F. & Khan, E. Evaluation of

soil function following remediation of petroleum hydrocarbons—a review of current

remediation techniques. Curr. Pollut. Reports 3, 192–205 (2017).

11. Das, N. & Chandran, P. Microbial degradation of petroleum hydrocarbon contaminants:

An overview. Biotechnol. Res. Int. 2011, (2011).

12. Bombach, P., Richnow, H. H. & Kästner, M. Current approaches for the assessment of in

situ biodegradation. Appl. Microbiol. Biotechnol. 86, 839–852 (2010).

13. Biegert, T., Fuchs, G. & Heider, J. Evidence that anaerobic oxidation of toluene in the

denitrifying bacterium Thauera aromatica is initiated by formation of benzylsuccinate

from toluene and fumarate. Eur. J. Biochem. 238, 661–668 (1996).

14. Beller, H. R. & Spormann, A. M. Benzylsuccinate formation as a means of anaerobic

toluene activation by sulfate-reducing strain PRTOL1. Appl. Environ. Microbiol. 63,

3729–3731 (1997).

15. Vogt, C., Kleinsteuber, S. & Richnow, H. H. Anaerobic benzene degradation by bacteria.

Microb. Biotechnol. 4, 710–724 (2011).

16. Rabus, R. et al. Anaerobic microbial degradation of hydrocarbons: From enzymatic

reactions to the environment. J. Mol. Microbiol. Biotechnol. 26, 5–28 (2016).

17. Meckenstock, R. U. et al. Biodegradation : Updating the Concepts of Control for

Microbial Cleanup in Contaminated Aquifers. Environ. Sci. Technol. 49, 7073–7081

(2015).

18. Morgan, P., Lewis, S. T. & Watkinson, R. J. Biodegradation of benzene, toluene,

162

ethylbenzene and xylenes in gas-condensate-contaminated ground-water. Environ. Pollut.

82, 181–190 (1993).

19. Alvarez, P. J. J. & Vogel, T. M. Substrate interactions of benzene, toluene, and para-

xylene during microbial degradation by pure cultures and mixed culture aquifer slurries.

Appl. Environ. Microbiol. 57, 2981–2985 (1991).

20. Winderl, C., Anneser, B., Griebler, C., Meckenstock, R. U. & Lueders, T. Depth-resolved

quantification of anaerobic toluene degraders and aquifer microbial community patterns in

distinct redox zones of a tar oil contaminant plume. Appl. Environ. Microbiol. 74, 792–

801 (2008).

21. Meckenstock, R. U. et al. Biodegradation: Updating the concepts of control for microbial

cleanup in contaminated aquifers. Environ. Sci. Technol. 49, 7073–7081 (2015).

22. Lueders, T. The ecology of anaerobic degraders of BTEX hydrocarbons in aquifers.

FEMS Microbiol. Ecol. 93, 1–35 (2016).

23. Kane, S. R., Beller, H. R., Legler, T. C. & Anderson, R. T. Biochemical and genetic

evidence of benzylsuccinate synthase in toluene-degrading, ferric iron-reducing

Geobacter metallireducens. Biodegradation 13, 149–154 (2002).

24. Winderl, C., Schaefer, S. & Lueders, T. Detection of anaerobic toluene and hydrocarbon

degraders in contaminated aquifers using benzylsuccinate synthase (bssA) genes as a

functional marker. Environ. Microbiol. 9, 1035–1046 (2007).

25. Johnson, J. M., Wawrik, B., Isom, C., Boling, W. B. & Callaghan, A. V. Interrogation of

Chesapeake Bay sediment microbial communities for intrinsic alkane-utilizing potential

under anaerobic conditions. FEMS Microbiol. Ecol. 91, 1–14 (2015).

26. Oka, A. R., Phelps, C. D., Zhu, X., Saber, D. L. & Young, L. Y. Dual biomarkers of

163

anaerobic hydrocarbon degradation in historically contaminated groundwater. Environ.

Sci. Technol. 45, 3407–3414 (2011).

27. Aitken, C. M. et al. Evidence that crude oil alkane activation proceeds by different

mechanisms under sulfate-reducing and methanogenic conditions. Geochim. Cosmochim.

Acta 109, 162–174 (2013).

28. Beller, H. R. & Spormann, A. M. Analysis of the novel benzylsuccinate synthase reaction

for anaerobic toluene activation based on structural studies of the product. J. Bacteriol.

180, 5454–5457 (1998).

29. Foght, J. Anaerobic biodegradation of aromatic hydrocarbons: Pathways and prospects. J.

Mol. Microbiol. Biotechnol. 15, 93–120 (2008).

30. Cravo-laureau, C., Grossi, V., Raphel, D., Matheron, R. & Hirschler-réa, A. Anaerobic n-

alkane metabolism by a Desulfatibacillum aliphaticivorans strain CV2803 T. Appl.

Environ. Microbiol. 71, 3458–3467 (2005).

31. Beller, H. R. Metabolic indicators for detecting in situ anaerobic alkylbenzene

degradation. Biodegradation 11, 125–139 (2000).

32. Meckenstock, R. U. & Mouttaki, H. Anaerobic degradation of non-substituted aromatic

hydrocarbons. Curr. Opin. Biotechnol. 22, 406–414 (2011).

33. Funk, M. A., Marsh, E. N. G. & Drennan, C. L. Substrate-bound structures of

benzylsuccinate synthase reveal how toluene is activated in anaerobic hydrocarbon

degradation. J. Biol. Chem. 290, 22398–22408 (2015).

34. Shisler, K. A. & Broderick, J. B. Glycyl radical activating enzymes: Structure,

mechanism, and substrate interactions. Arch. Biochem. Biophys. 546, 64–71 (2014).

35. Wagner, A. F. V. et al. YfiD of Escherichia coil and Y061 of bacteriophage T4 as

164

autonomous glycyl radical cofactors reconstituting the catalytic center of oxygen-

fragmented pyruvate formate-lyase. Biochem. Biophys. Res. Commun. 285, 456–462

(2001).

36. Bharadwaj, V. S., Vyas, S., Villano, S. M., Maupina, C. M. & Dean, A. M. Unravelling

the impact of hydrocarbon structure on the fumarate addition mechanism – a gas-phase ab

initio study. Phys. Chem. Chem. Phys. 17, 4054–4066 (2015).

37. Gieg, L. M. & Toth, C. R. Signature metabolite analysis to determine in situ anaerobic

hydrocarbon biodegradation. In: Handbook of Hydrocarbon and Lipid Microbiology,

Springer, 30 pp., (2017).

38. McInerney, M. J., Sieber, J. R. & Gunsalus, R. P. Syntrophy in anaerobic global carbon

cycles. Curr. Opin. Biotechnol. 20, 623–632 (2009).

39. Agrawal, A. & Gieg, L. M. In situ detection of anaerobic alkane metabolites in subsurface

environments. Front. Microbiol. 4, 1–11 (2013).

40. Kniemeyer, O., Fischer, T., Wilkes, H., Glo, F. O. & Widdel, F. Anaerobic degradation of

ethylbenzene by a new type of marine sulfate-reducing bacterium. Appl. Environ.

Microbiol. 69, 760–768 (2003).

41. Gieg, L. M. & Suflita, J. M. Detection of anaerobic metabolites of saturated and aromatic

hydrocarbons in petroleum-contaminated aquifers. Environ. Sci. Technol. 36, 3755–3762

(2002).

42. Beller, H. R., Ding, W. H. & Reinhard, M. By-products of anaerobic alkylbenzene

metabolism useful as indicators of in-situ bioremediation. Environ. Sci. Technol. 29,

2864–2870 (1995).

43. Elshahed, M. S., Gieg, L. M., McInerney, M. J. & Suflita, J. M. Signature metabolites

165

attesting to the in situ attenuation of alkylbenzenes in anaerobic environments. Environ.

Sci. Technol. 35, 682–689 (2001).

44. Jobelius, C. et al. Metabolites indicate hot spots of biodegradation and biogeochemical

monitoring well gradients in a high-resolution monitoring well. Environ. Sci. Technol. 45,

1–12 (2011).

45. Zhou, L. et al. Analyses of n-alkanes degrading community dynamics of a high-

temperature methanogenic consortium enriched from production water of a petroleum

reservoir by a combination of molecular techniques. Ecotoxicology 21, 1680–1691 (2012).

46. von Netzer, F. et al. Enhanced gene detection assays for fumarate-adding enzymes allow

uncovering of anaerobic hydrocarbon degraders in terrestrial and marine systems. Appl.

Environ. Microbiol. 79, 543–552 (2013).

47. Callaghan, A. V et al. Diversity of benyzl- and alkylsuccinate synthase genes in

hydrocarbon-impacted environments and enrichment cultures. Environ. Sci. Technol.

7287–7294 (2010).

48. Stagars, M. H., Emil Ruff, S., Amann, R. & Knittel, K. High diversity of anaerobic

alkane-degrading microbial communities in marine seep sediments based on (1-

methylalkyl)succinate synthase genes. Front. Microbiol. 6, 1–17 (2016).

49. Stagars, M. H., Mishra, S., Treude, T., Amann, R. & Knittel, K. Microbial community

response to simulated petroleum seepage in Caspian sea sediments. Front. Microbiol. 8,

1–16 (2017).

50. Beller, H. R. et al. Comparative assessments of benzene, toluene, and xylene natural

attenuation by quantitative polymerase chain reaction analysis of a catabolic gene,

signature metabolites, and compound-specific isotope analysis. Environ. Sci. Technol. 42,

166

6065–6072 (2008).

51. Beller, H. R., Kane, S. R., Legler, T. C. & Alvarez, P. J. J. A real-time polymerase chain

reaction method for monitoring anaerobic, hydrocarbon-degrading bacteria based on a

catabolic gene. Environ. Sci. Technol. 36, 3977–3984 (2002).

52. Gittel, A. et al. Ubiquitous presence and novel diversity of anaerobic alkane degraders in

cold marine sediments. Front. Microbiol. 6, (2015).

53. Ehrenreich, P., Behrends, A., Harder, J. & Widdel, F. Anaerobic oxidation of alkanes by

newly isolated denitrifying bacteria. Arch. Microbiol. 173, 58–64 (2000).

54. Cravo-Laureau, C., Matheron, R., Cayol, J. L., Joulian, C. & Hirschler-Réa, A.

Desulfatibacillum aliphaticivorans gen. nov., sp. nov., an n-alkane- and n-alkene-

degrading, sulfate-reducing bacterium. Int. J. Syst. Evol. Microbiol. 54, 77–83 (2004).

55. Kim, S. J. et al. Metabolic versatility of toluene-degrading, iron-reducing bacteria in tidal

flat sediment, characterized by stable isotope probing-based metagenomic analysis.

Environ. Microbiol. 16, 189–204 (2014).

56. Berdugo-Clavijo, C. & Gieg, L. M. Conversion of crude oil to methane by a microbial

consortium enriched from oil reservoir production waters. Front. Microbiol. 5, 1–10

(2014).

57. Müller, J. B. et al. Combined iron and sulfate reduction biostimulation as a novel

approach to enhance BTEX and PAH source-zone biodegradation in biodiesel blend-

contaminated groundwater. J. Hazard. Mater. 326, 229–236 (2017).

58. Kazy, S. K., Monier, A. L. & Alvarez, P. J. J. Assessing the correlation between anaerobic

toluene degradation activity and bssA concentrations in hydrocarbon-contaminated aquifer

material. Biodegradation 21, 793–800 (2010).

167

59. Kleindienst, S. et al. Diverse sulfate-reducing bacteria of the Desulfosarcina/

Desulfococcus clade are the key alkane degraders at marine seeps. ISME J. 8, 2029-2044

(2014).

60. Oberding, L. K. & Gieg, L. M. Methanogenic paraffin biodegradation: alkylsuccinate

synthase gene quantification and dicarboxylic acid production. Appl. Environ. Microbiol.

84, 1–14 (2018).

61. Fowler, S. J., Manefield, M. & Gieg, L. M. Identification of toluene degraders in a

methanogenic enrichment culture. FEMS Microbiol. Ecol. 89, 625–636 (2014).

62. Debode, F., Marien, A., Janssen, É., Bragard, C. & Berben, G. The influence of amplicon

length on real-time PCR results. Biotechnol. Agron. Soc. Env. 21, 3–11 (2017).

63. Pilloni, G. et al. Dynamics of hydrology and anaerobic hydrocarbon degrader

communities in a tar-oil contaminated aquifer. Microorganisms 7, 1–15 (2019).

64. Callaghan, A. V., Wawrik, B., Ní Chadhain, S. M., Young, L. Y. & Zylstra, G. J.

Anaerobic alkane-degrading strain AK-01 contains two alkylsuccinate synthase genes.

Biochem. Biophys. Res. Commun. 366, 142–148 (2008).

65. Davidova, I. A., Duncan, K. E., Choi, O. K. & Suflita, J. M. Desulfoglaeba alkanexedens

gen. nov., sp. nov. an n-alkane-degrading, sulfate-reducing bacterium. Int. J. Syst. Evol.

Microbiol. 56, 2737–2742 (2006).

66. Zamarro, M. T., Martín-Moldes, Z. & Díaz, E. The ICEXTD of Azoarcus sp. CIB, an

integrative and conjugative element with aerobic and anaerobic catabolic properties.

Environ. Microbiol. 18, 5018–5031 (2016).

67. Toth, C. R. A. & Gieg, L. M. Time course-dependent methanogenic crude oil

biodegradation: Dynamics of fumarate addition metabolites, biodegradative genes, and

168

microbial community composition. Front. Microbiol. 8, 1–16 (2018).

68. Weelink, S. A. B. et al. A strictly anaerobic betaproteobacterium Georgfuchsia toluolica

gen. nov., sp. nov. degrades aromatic compounds with Fe(III), Mn(IV) or nitrate as an

electron acceptor. FEMS Microbiol. Ecol. 70, (2009).

69. Beller, H. R. & Edwards, E. A. Anaerobic toluene activation by benzylsuccinate synthase

in a highly enriched methanogenic culture. Appl. Environ. Microbiol. 66, 5503–5505

(2000).

70. Edwards, E. A. & Grbic-Galic, D. Anaerobic degradation of toluene and o-xylene by

methanogenic consortium. Appl. Environ. Microbiol. 60, 313–322 (1994).

71. Lovley, D. R. & Phillips, E. J. P. Rapid assay for microbially reducible ferric iron in

aquatic sediments. Appl. Environ. Microbiol. 53, 1536–1540 (1987).

72. Fida, T. T. et al. Synergy of sodium nitroprusside and nitrate in inhibiting the activity of

sulfate reducing bacteria in oil-containing bioreactors. Front. Microbiol. 9, 1–11 (2018).

73. An, D. et al. Metagenomics of hydrocarbon resource environments indicates aerobic taxa

and genes to be unexpectedly common. Environ. Sci. Technol. 47, 10708–10717 (2013).

74. Zhang, J., Kobert, K., Flouri, T. & Stamatakis, A. PEAR: A fast and accurate Illumina

Paired-End reAd mergeR. Bioinformatics 30, 614–620 (2014).

75. Dong, X. et al. Fast and simple analysis of MiSeq amplicon sequencing data with

MetaAmp. Front. Microbiol. 8, 1–8 (2017).

76. Alberta Environment and Parks (AEP). Alberta Tier 1 Soil and Groundwater Remediation

Guidelines. Land and Forestry Policy Branch, Policy Division (2016). doi:ISBN 978-1-

4601-2695-0

77. Krieger, C. J., Beller, H. R., Reinhard, M. & Spormann, A. M. Initial reactions in

169

anaerobic oxidation of m-xylene by the denitrifying bacterium Azoarcus sp. strain T. J.

Bacteriol. 181, 6403–6410 (1999).

78. Aburto, A. et al. Mixed aerobic and anaerobic microbial communities in benzene-

contaminated groundwater. J. Appl. Microbiol. 106, 317–328 (2009).

79. Fahy, A., McGenity, T. J., Timmis, K. N. & Ball, A. S. Heterogeneous aerobic benzene-

degrading communities in oxygen-depleted groundwaters. FEMS Microbiol. Ecol. 58,

260–270 (2006).

80. Müller, A. L., Kjeldsen, K. U., Rattei, T., Pester, M. & Loy, A. Phylogenetic and

environmental diversity of DsrAB-type dissimilatory (bi)sulfite reductases. ISME J. 9,

1152–1165 (2015).

81. Sievers, F. et al. Fast, scalable generation of high-quality protein multiple sequence

alignments using Clustal Omega. Mol. Syst. Biol. 7, (2011).

82. Benchling [Biology Software]. Retrieved from https://benchling.com 300 (2017).

83. Bustin, S. A. et al. MIQE précis: Practical implementation of minimum standard

guidelines for fluorescence-based quantitative real-time PCR experiments. BMC Mol.

Biol. 11, (2010).

84. Altschul, S. F. et al. Gapped BLAST and PSI-BLAST: a new generation of protein

database search programs Stephen. Nucleic Acids Res. 25, 3389–3402 (1997).

85. Thompson, J. D., Higgins, D. G. & Gibson, T. J. CLUSTALW: improving the sensitivity

of progressive multiple sequence alignment through sequence weighting, position-specific

gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673–4680 (1994).

86. Guindon, S. et al. New algorithms and methods to estimate maximum-likelihood

phylogenies: Assessing the performance of PhyML 3.0. Syst. Biol. 59, 307–321 (2010).

170

87. Hordijk, W. & Gascuel, O. Improving the efficiency of SPR moves in phylogenetic tree

search methods based on maximum likelihood. Bioinformatics 21, 4338–4347 (2005).

88. Bitton, G. & Harvey, R. W. Transport of pathogens through soils and aquifers. In:

Environmental Microbiology, pp. 103–124 (1992).

89. Bear, J. & Cheng, A. H. D. Modeling Groundwater Flow and Contaminant Transport.

Springer (2010).

90. Knaebel, D. B. & Crawford, R. L. Extraction and purification of microbial DNA from

petroleum‐contaminated soils and detection of low numbers of toluene, octane and

pesticide degraders by multiplex polymerase chain reaction and Southern analysis. Mol.

Ecol. 4, 579–592 (1995).

91. ThermoFisherScientific. Product Information; Taq DNA Polymerase (recombinant)

#EP0402. 1–2 (2015).

92. KAPABiosystems. Technical Data Sheet KAPA HiFi HotStart ReadyMix PCR Kit. Tech.

Data Sheet (2017).

93. Kermekchiev, M. B., Kirilova, L. I., Vail, E. E. & Barnes, W. M. Mutants of Taq DNA

polymerase resistant to PCR inhibitors allow DNA amplification from whole blood and

crude soil samples. Nucleic Acids Res. 37, 1–14 (2009).

94. Opel, K. L., Chung, D. & Mccord, B. R. A study of PCR inhibition mechanisms using real

time PCR. J. Forensic Sci. 55, 25–33 (2010).

95. Cating, R. A., Hoy, M. A. & Palmateer, A. J. A comparison of standard and high-fidelity

PCR: evaluating quantification and detection of pathogen DNA in the presence of orchid

host tissue. Plant Dis. 96, 480–485 (2012).

96. Hedman, J., Nordgaard, A., Rasmusson, B., Ansell, R. & Rådström, P. Improved forensic

171

DNA analysis through the use of alternative DNA polymerases and statistical modeling of

DNA profiles. Biotechniques 47, 951–958 (2009).

97. Albers, C. N., Jensen, A., Bælum, J. & Jacobsen, C. S. Inhibition of DNA polymerases

used in q-PCR by structurally different soil-derived humic substances. Geomicrobiol. J.

30, 675–681 (2013).

98. Callaghan, A. V. et al. The genome sequence of Desulfatibacillum alkenivorans AK-01: A

blueprint for anaerobic alkane oxidation. Environ. Microbiol. 14, 101–113 (2012).

99. Tamura, K. & Nei, M. Estimation of the number of nucleotide substitutions in the control

region of mitochondrial DNA in humans and chimpanzees. Mol. Biol. Evol. 10, (1993).

100. Jeon, C. O., Park, W., Ghiorse, W. C. & Madsen, E. L. Polaromonas naphthalenivorans

sp. nov., a naphthalene-degrading bacterium from naphthalene-contaminated sediment.

Int. J. Syst. Evol. Microbiol. 93–97 (2004).

101. Sun, W., Xie, S., Luo, C. & Cupples, A. M. Direct link between toluene degradation in

contaminated-site microcosms and a Polaromonas strain. Appl. Environ. Microbiol. 76,

956–959 (2010).

102. Xie, S., Sun, W., Luo, C. & Cupples, A. M. Novel aerobic benzene degrading

microorganisms identified in three soils by stable isotope probing. Biodegradation 22, 71–

81 (2011).

103. Iffis, B., St-Arnaud, M. & Hijri, M. Bacteria associated with arbuscular mycorrhizal fungi

within roots of plants growing in a soil highly contaminated with aliphatic and aromatic

petroleum hydrocarbons. FEMS Microbiol. Lett. 358, 44–54 (2014).

104. Khelifi, N. et al. Anaerobic oxidation of long-chain n-alkanes by the hyperthermophilic

sulfate-reducing archaeon, Archaeoglobus fulgidus. ISME J. 8, 2153–2166 (2014).

172

105. Coschigano, P. W. Transcriptional analysis of the tutE tutFDGH gene cluster from

Thauera aromatica strain T1. Appl. Environ. Microbiol. 66, 1147–1151 (2000).

106. NRC. In situ bioremediation: when does it work? (National Academy Press, 1993).

107. Borden, R. C., Hunt, M., Shafer, M. B. & Barlaz, M. A. Anaerobic biodegradation of

BTEX in aquifer material. Environ. Res. Br. 1–9 (1997).

108. B Da Silva, M. L. & J Alvarez, P. J. Enhanced anaerobic biodegradation of benzene-

toluene-ethylbenzene-xylene-ethanol mixtures in bioaugmented aquifer columns. Appl.

Environ. Microbiol. 70, 4720–4726 (2004).

109. Aburto, A. & Peimbert, M. Degradation of a benzene-toluene mixture by hydrocarbon-

adapted bacterial communities. Ann. Microbiol. 61, 553–562 (2011).

110. Eastcott, L., Shiu, W. Y. & Mackay, D. Environmentally relevant physical-chemical

properties of hydrocarbons: A review of data and development of simple correlations. Oil

Chem. Pollut. 4, 191–216 (1988).

111. Heath, J. S., Koblis, K. & Sager, S. L. Review of chemical, physical, and toxicologic

properties of components of total petroleum hydrocarbons. J. Soil Contam. 2, 1–25

(1993).

112. Fowler, S. J., Dong, X., Sensen, C. W., Suflita, J. M. & Gieg, L. M. Methanogenic toluene

metabolism : community structure and intermediates. Environ. Microbiol. 14, 754–764

(2012).

113. Heider, J., Spormann, A. M., Beller, H. R. & Widdel, F. Anaerobic bacterial metabolism

of hydrocarbons. FEMS Microbiol. Rev. 22, 459–473 (1999).

114. Burland, S. M. & Edwards, E. A. Anaerobic benzene biodegradation linked to nitrate

reduction. Appl. Environ. Microbiol. 65, 529–533 (1999).

173

115. Bauer, R. D. et al. Enhanced biodegradation by hydraulic heterogeneities in petroleum

hydrocarbon plumes. J. Contam. Hydrol. 105, 56–68 (2009).

116. Lünsmann, V. et al. In situ protein-SIP highlights Burkholderiaceae as key players

degrading toluene by para ring hydroxylation in a constructed wetland model. Environ.

Microbiol. 18, 1176–1186 (2016).

117. van der Zaan, B. M. et al. Anaerobic benzene degradation under denitrifying conditions :

Peptococcaceae as dominant benzene degraders and evidence for a syntrophic process.

Environ. Microbiol. 14, 1171–1181 (2012).

118. Kadnikov, V. V. et al. Composition of the microbial communities of bituminous

constructions at natural oil seeps at the bottom of Lake Baikal. Microbiology 82, 373–382

(2013).

119. Mohanty, G. & Mukherji, S. Biodegradation rate of diesel range n-alkanes by bacterial

cultures Exiguobacterium aurantiacum and Burkholderia cepacia. Int. Biodeterior.

Biodeg. 61, 240–250 (2008).

120. Mattes, T. E. et al. The genome of Polaromonas sp. strain JS666: Insights into the

evolution of a hydrocarbon- and xenobiotic-degrading bacterium, and features of

relevance to biotechnology. Appl. Environ. Microbiol. 74, 6405–6416 (2008).

121. Scheps, D., Malca, S. H., Hoffmann, H., Nestl, B. M. & Hauer, B. Regioselective x-

hydroxylation of medium-chain n-alkanes and primary alcohols by CYP153 enzymes

from Mycobacterium marinum and Polaromonas sp. strain JS666. Org. Biomol. Chem. 1,

6727–6733 (2011).

122. Norton, J. M. et al. Complete genome sequence of Nitrosospira multiformis , an ammonia-

oxidizing bacterium from the soil environment. Appl. Environ. Microbiol. 74, 3559–3572

174

(2008).

123. Risso, C. et al. Genome-scale comparison and constraint-based metabolic reconstruction

of the facultative anaerobic Fe(III)-reducer Rhodoferax ferrireducens. BMC Genomics 10,

447 (2009).

124. Ros, M., Rodríguez, I., García, C. & Teresa, M. Bacterial community in semiarid

hydrocarbon contaminated soils treated by aeration and organic amendments. Int.

Biodeterior. Biodeg. 94, 200–206 (2014).

125. Militon, C., Jézéquel, R., Gilbert, F., Corsellis, Y. & Sylvi, L. Dynamics of bacterial

assemblages and removal of polycyclic aromatic hydrocarbons in oil-contaminated coastal

marine sediments subjected to contrasted oxygen regimes. Environ. Sci. Pollut. Res.

15260–15272 (2015).

126. Zhang, S., Wang, Q., Wan, R. & Xie, S. Changes in bacterial community of anthracene

bioremediation in municipal solid waste composting soil. J. Zhejiang Univ. 12, 760–768

(2011).

127. Wilson, L. P. & Bouwer, E. J. Biodegradation of aromatic compounds under mixed

oxygen / denitrifying conditions : a review. J. Ind. Microbiol. Biotechnol. 116–130 (1997).

128. Finneran, K. T., Johnsen, C. V & Lovley, D. R. Rhodoferax ferrireducens sp. nov., a

psychrotolerant, facultatively anaerobic bacterium that oxidizes acetate with the reduction

of Fe (III). Int. J. Syst. Evol. Microbiol. 53, 669–673 (2003).

129. Leuthner, B. et al. Biochemical and genetic characterization of benzylsuccinate synthase

from Thauera aromatica: A new glycyl radical enzyme catalysing the first step in

anaerobic toluene metabolism. Mol. Microbiol. 28, 615–628 (1998).

130. Sun, W., Sun, X. & Cupples, A. M. Presence, diversity and enumeration of functional

175

genes (bssA and bamA) relating to toluene degradation across a range of redox conditions

and inoculum sources. Biodegradation 25, (2014).

131. Borden, R. C., Daniel, R. A., LeBrun IV, L. E. & Davis, C. W. Intrinsic biodegradation of

MTBE and BTEX in a gasoline-contaminated aquifer. Water Resour. Res. 33, 1105–1115

(1997).

132. Chen, L.-X. et al. Candidate Phyla Radiation Roizmanbacteria from hot springs have

novel, unexpectedly abundant, and potentially alternatively functioning CRISPR-Cas

systems. bioRxiv 448639 (2018). doi:doi.org/10.1101/448639

133. Bodegom, P. M. Van, Scholten, J. C. M. & Stams, A. J. M. Direct inhibition of

methanogenesis by ferric iron. FEMS Microbiol. Ecol. 49, 261–268 (2004).

134. Kluber, H. D. & Conrad, R. Inhibitory effects of nitrate, nitrite, NO and N2O on

methanogenesis by Methanosarcina barkeri and Methanobacterium bryantii. FEMS

Microbiol. Ecol. 25, (1998).

135. Abram, J. W. & Nedwell, D. B. Inhibition of methanogenesis by sulphate reducing

bacteria competing for transferred hydrogen. Microbiology 92, 89–92 (1978).

136. Murakami, T. et al. Metagenomic analyses highlight the symbiotic association between

the glacier stone fly Andiperla willinki and its bacterial gut community. Environ.

Microbiol. 20, 4170–4183 (2018).

137. Gutierrez, T. et al. Polycyclovorans algicola gen. nov., sp. nov., an aromatic-hydrocarbon-

degrading marine bacterium found associated with laboratory cultures of marine

phytoplankton. Appl. Environ. Microbiol. 79, 205–214 (2013).

138. Jayamani, I. & Cupples, A. M. Stable isotope probing and high-throughput sequencing

implicate Xanthomonadaceae and Rhodocyclaceae in ethylbenzene degradation. Environ.

176

Eng. Sci. 32, 240–250 (2015).

139. Mcinerney, M. J., Bryant, M. P. & Pfennig, N. Anaerobic bacterium that degrdes fatty

acids in syntrophic assocition with methanogens. Arch. Microbiol. 135, 129–135 (1979).

140. Parks, D. H. et al. A standardized bacterial taxonomy based on genome phylogeny

substantially revises the tree of life. Nat. Biotechnol. 36, (2018).

177

Chapter 8: Appendix

assA sequences name (as per NCBI) Accession Number bssA sequences name (as per NCBI) Accession Number Desulfatibacillum alkenivorans AK-01, complete genome CP001322.1 Desulfobacula toluolica Tol2 complete genome FO203503.1 Sulfate-reducing bacterium AK-01 alkylsuccinate synthase (assA1) DQ826035.1 Geobacter metallireducens benzylsuccinate synthase alpha subunit AF441130 gene, complete cds (bssA) gene, partial cds; and benzylsuccinate synthase beta subunit Sulfate-reducing bacterium AK-01 alkylsuccinate synthase (assA2) DQ826036 Thauera(bssB) gene, aromatica complete tdiSR cds and bssDCAB operons for benzylsuccinate AJ001848.3 gene, complete cds synthase and a two-component regulatory system Desulfoglaeba alkanexedens alkylsuccinate synthase (assA) gene, GU453656.1 Geobacter metallireducens GS-15, complete genome CP000148.1 complete cds Uncultured prokaryote clone Passaic_River_OTU2 alkylsuccinate GU453657 Thauera sp. MZ1T, complete genome CP001281.2 synthase (assA) gene, partial cds Uncultured prokaryote clone Gowanus_Canal_OTU1 GU453659 Desulfobacula toluolica strain DSM 7467 benzylsuccinate synthase EF123663 alkylsuccinate synthase (assA) gene, partial cds alpha subunit (bssA) gene, partial cds Uncultured prokaryote clone Fort_Lupton_OTU3 alkylsuccinate GU453664 Desulfotomaculum sp. Ox39 benzylsuccinate synthase alpha subunit EF123665.1 synthase (assA) gene, partial cds (bssA) gene, partial cds Uncultured prokaryote clone Arthur_Kill_OTU1 alkylsuccinate GU453666 Sulfate-reducing bacterium TRM1 benzylsuccinate synthase alpha EF123667 synthase (assA) gene, partial cds subunit (bssA) gene, partial cds Uncultured bacterium clone BGM02 alkylsuccinate synthase alpha JX219367 Uncultured bacterium clone B49bss_021 benzylsuccinate synthase EF123670 subunit (assA) gene, partial cds alpha subunit (bssA) gene, partial cds Uncultured bacterium clone BGM24 alkylsuccinate synthase alpha JX219368 Uncultured bacterium clone Pb312_80 benzylsuccinate synthase EF123703 subunit (assA) gene, partial cds alpha subunit (bssA) gene, partial cds Uncultured bacterium clone M-NAPH012 alkylsuccinate synthase KC464263.1 Uncultured prokaryote clone Fort_Lupton_OTU1 benzylsuccinate GU453672 alpha subunit (assA) gene, partial cds synthase (bssA) gene, partial cds Uncultured bacterium clone M-OIL045 alkylsuccinate synthase KC464317 Uncultured bacterium clone F3A10 benzylsuccinate synthase alpha JX219282.1 alpha subunit (assA) gene, partial cds subunit (bssA) gene, partial cds Smithella sp. enrichment culture clone SCADC alkylsuccinate KF824850 Uncultured bacterium clone F5A30 benzylsuccinate synthase alpha JX219323 synthase alpha subunit (assA) gene, complete cds subunit homologue (bssA) gene, partial cds Bacterium enrichment culture clone residual oil-degrading OTU1 KU094062 Uncultured bacterium clone M-CON001 benzylsuccinate synthase KC463949.1 alkylsuccinate synthase alpha subunit gene, partial cds alpha subunit (bssA) gene, partial cds Bacterium enrichment culture clone octadecane-degrading OTU1 KU094063 Uncultured bacterium clone M-OIL003 benzylsuccinate synthase KC464029 alkylsuccinate synthase alpha subunit gene, partial cds alpha subunit (bssA) gene, partial cds Desulfatibacillum aliphaticivorans partial masD gene for LN868321.1 Desulfosporosinus sp. enrichment culture clone S2S-F11 1.697-4 KJ398020 methylalkyl succinate synthase, strain CV2803 benzylsuccinate synthase (bssA) gene, partial cds Desulfatibacillum alkenivorans partial masD gene for methylalkyl LN868322 Uncultured bacterium clone OTU1 benzylsuccinate synthase alpha KX148522.1 succinate synthase, strain PF2803 subunit (bssA) gene, partial cds Uncultured Desulfatibacillum sp. partial masD gene for 1-methyl LN879422 alkyl succinate synthase, strain Propane60-GuB Uncultured bacterium partial masD gene for 1-methylalkyl LT546441 succinate synthase, clone OTU_1 uncultured bacterium partial assA gene for alkylsuccinate synthase, LT907865.1 clone HeM4

178

Table 24: Compilation of identified non-specific amplicons from newly designed assA primer mix

Non-target matches; assA Accesion Number Annotation coverage % ID % Hydrogenophaga sp. PBC, complete genome CP017311.1 30S ribosomal protein S17 58 84 Sphingobium sp. TKS chromosome 1, complete sequence CP005083.1 excinuclease ABC subunit A 72 78 Roseomonas gilardii strain U14-5 chromosome 1, complete sequence CP015583.1 excinuclease ABC subunit A 83 77 Sphingobium sp. SCG-1 plasmid unnamed, complete sequence CP026373.1 aromatic hydrocarbon degradation protein 100 97 Pseudomonas veronii strain R02, complete genome CP018420.1 UDP-4-amino-4-deoxy-L-arabinose formyltransferase 99 99 Melaminivora sp. SC2-7 chromosome, complete genome CP027792.1 transketolase 81 85 Variovorax paradoxus S110 chromosome 1, complete sequence CP001635.1 transketolase 78 85 Gemmatimonas aurantiaca T-27 DNA, complete genome AP009153.1 replicative DNA helicase 88 74 Pseudomonas brenneri strain BS2771 genome assembly, chromosome: I LT629800.1 UDP-4-amino-4-deoxy-L-arabinose formyltransferase 100 99 Mesorhizobium sp. M7D.F.Ca.US.005.01.1.1 chromosome CP034453.1 ABC transporter 100 83 Jatrophihabitans sp. GAS493 genome assembly, chromosome: I LT907982.1 Dyp-type peroxidase family 94 81 Polaromonas naphthalenivorans CJ2, complete genome CP000529.1 aminotransferase, class I and II; 90 79 Leptolyngbya sp. O-77 DNA, complete genome AP017367.1 Ferredoxin-dependent glutamate synthase 1 61 78 Burkholderia plantarii strain PG1 chromosome 1, complete sequence CP002580.1 ribosomal protein L2 95 80 Collimonas fungivorans Ter331, complete genome CP002745.1 Phosphoribosylformylglycinamidine synthase 100 86 Blastococcus saxobsidens DD2 complete genome FO117623.1 Pyruvate dehydrogenase E1 component 90 78 Streptomyces sp. ZFG47 chromosome, complete genome CP030073.1 pyruvate dehydrogenase (acetyl-transferring) homodimeric type 79 79 Agromyces aureus strain AR33, complete genome CP013979.1 peroxidase 69 80 Massilia violaceinigra strain B2 chromosome CP024608.1 urocanate hydratase 97 84 Azoarcus sp. DN11 chromosome, complete genome CP021731.1 aspartate ammonia-lyase 100 87 Azospirillum humicireducens strain SgZ-5 plasmid pYZ2 CP028903.1 aspartate ammonia-lyase 100 86 Thauera sp. K11 chromosome, complete genome CP023439.1 aspartate ammonia-lyase 97 85

Table 25: Compilation of identified non-specific amplicons from newly designed bssA primer mix Non-target matches; bssA Accesion Number Annotation coverage % ID % Ottowia oryzae strain KADR8-3 chromosome, complete genome CP027666.1 acyl-CoA dehydrogenase 80 82 Acidovorax avenae subsp. avenae strain QHB1 chromosome CP028300.1 acyl-CoA dehydrogenase 83 82 Bradyrhizobium sp. S23321 DNA, complete genome AP012279.1 RNA polymerase beta subunit 40 92 Desulfatibacillum aliphaticivorans partial masD gene, strain CV2803 LN868321.1 alkylsuccinate synthase subunit A 63 100 Desulfatibacillum alkenivorans AK-01, complete genome CP001322.1 alkylsuccinate synthase subunit A 63 100

179

8.1 Media Recipe

‘Pfennig’ Freshwater Minimal Salts Medium *** (from McInerney et al. 1979)139 Pfennig I 5.0 mL Pfennig II 5.0 mL Wolin Metals 1.0 mL Balch Vitamins 1.0 mL Resazurin 0.1 mL of 0.1% stock NaHCO3 0.35 g MilliQ H2O 100 mL Electron Accpetors: NaNO3 NaSO4 FeO(OH)3 Protocol: 1. Add all media components (minus Pfennig I and bicarbonate) 2. pH = 7.1 – 7.2 3. Boil to remove O2 4. Add electron acceptor 5. Bubble with N2/CO2 gas (90/10) 6. Add bicarbonate when cool 7. Dispense into serum bottles, continue bubbling. Seal with Viton stoppers and crimp with aluminum seals. 8. Add Pfennig I 9. Add reducing agent (sterile sodium sulfide) • 0.1 mL/100 mL of 2.5% stock (nitrate reducing conditions) • 2 mL/100 mL of 2.5% stock (sulfate, iron (III), and methanogenic conditions) Pfennig I /100 mL MilliQ H2O K2HPO4 10.0 g Pfennig II /100 mL MilliQ H2O MgCl2 * 6 H2O 6.6 g NaCl 8.0 g NH4Cl 8.0 g CaCl2 * 2 H2O 1.0 g Wolin Metals /100 mL MilliQ H2O EDTA 0.5 g MgSO4 * 6 H2O 3.0 g MnSO4 * H2O 0.5 g NaCl 1.0 g CaCl2 * 2 H2O 0.1 g ZnSO4 * 7 H2O 0.1 g FeSO4 * 7 H2O 0.1 g CuSO4 * 7 H2O 0.01 g Na2MoO4 * 2 H2O 0.01 g H3BO4 0.01 g

180

Na2SeO4 0.005 g NiCl2 * 6 H2O 0.003 g Balch Vitamins /100 mL MilliQ H2O Biotin 2.0 mg Folic acid 2.0 mg Pyridoxine-HCl 10.0 mg Thiamine-HCl 5.0 mg Riboflavin 5.0 mg Nicotinic acid 5.0 mg D-L-calcium pantothenate 5.0 mg Vitamin B12 0.1 mg p-aminobenzoic acid 5.0 mg lipoic acid 5.0 mg mercaptoethanesulfonic acid 5.0 mg

*** Note, concentrations are reported as per 100 mL, 10X concentrated as normal