The Microbiology and Metal Attenuation in a Natural Wetland Impacted by Acid Mine Drainage

The microbiology and metal attenuation in a natural wetland impacted by acid mine drainage

A thesis submitted to the University of Manchester for the degree of Doctor of Philosophy in the Faculty of Science and Engineering

2018

Oscar E. Aguinaga Vargas

School of Earth and Environmental Sciences

TABLE OF CONTENTS

ABSTRACT ...... 6 DECLARATION ...... 7 COPYRIGHT STATEMENT ...... 8 ACKNOWLEDGMENTS ...... 9

Chapter 1: GENERAL INTRODUCTION ...... 10 1.1 ACID MINE DRAINAGE ...... 10 1.2 ENVIRONMENTAL IMPLICATIONS OF ACID MINE DRAINAGE ... 13 1.2.1 Presence of trace metals ...... 13 1.2.2 Environmental impact ...... 14 1.3 ACID MINE DRAINAGE REMEDIATION ...... 18 1.3.1 Acid mine drainage remediation by wetlands ...... 19 1.3.1.1 Physicochemical processes ...... 19 1.3.1.2 Role of plants ...... 21 1.3.1.3 Role of microorganisms ...... 23 1.4 CASE STUDIES OF ACID MINE DRAINAGE REMEDIATION BY WETLANDS ...... 24 1.4.1 Constructed wetlands ...... 24 1.4.2 Natural wetlands ...... 28 1.5 PARYS MOUNTAIN ...... 30 1.6 HYPOTHESES, AIMS AND OBJETIVES OF THE THESIS ...... 39

Chapter 2: CHARACTERIZATION OF ACID MINE DRAINAGE POLLUTION AND MICROBIAL COMMUNITY SHIFTS IN THE SOUTHERN AND NORTHERN AFON GOCH ...... 42 2.1 ABSTRACT ...... 43 2.2 INTRODUCTION ...... 44 2.3 MATERIALS AND METHODS ...... 46 2.3.1 Study site and sample collection ...... 46 2.3.2 Water and sediments quality analysis ...... 48 2.3.3 DNA extraction ...... 52 2.3.4 16S rRNA gene sequencing ...... 52 2.3.5 Statistical analysis of environmental data ...... 53

2.3.6 Sequence data analysis ...... 54 2.4 RESULTS ...... 56 2.4.1 Environmental characterization of the SAG and NAG ...... 56 2.4.2 Microbial community structure from MiSeq data ...... 62 2.4.3 Relative abundance of known bacteria related to S and Fe metabolism ...... 69 2.4.4 Prediction of microbial taxonomic-derived metabolism ...... 75 2.5 DISCUSSION ...... 76 2.5.1 Microbial structure related to AMD pollution levels ...... 76 2.5.2 Prediction of metabolic potential ...... 86 2.6 CONCLUDING REMARKS ...... 87

Chapter 3: ANALYSIS OF BACTERIA ACTIVITY AND METAL DISTRIBUTION ALONG THE SOUTHERN AFON GOCH WETLAND ...... 89 3.1. ABSTRACT ...... 90 3.2 INTRODUCTION ...... 91 3.3 MATERIALS AND METHODS ...... 94 3.3.1 Field site locations and sampling ...... 94 3.3.2 In situ water analysis and total aqueous metal ...... 96 3.3.3 Metal analysis in sediments...... 96 3.3.4 Measurement of C and N along core depths ...... 97 3.3.5 DNA and RNA extraction ...... 98 3.3.6 Construction of 16S rRNA libraries ...... 98 3.3.7 Quantification of RNA transcripts ...... 99 3.3.8 Bioinformatic analysis ...... 100 3.3.9 Statistical analysis ...... 101 3.4 RESULTS ...... 103 3.4.1 Phylogenetic analysis of 16S rRNA libraries from W2 surface sediments ...... 103 3.4.2 Measurement of pH in water samples ...... 107 3.4.3 Sulphur content and speciation in water samples ...... 108 3.4.4 Iron content and speciation in wetland water samples .... 108 3.4.5 Trace metal levels in water samples along the wetland ... 111 3.4.6 Metal distribution along sediment depths ...... 111 3.4.7 Expression of dsrA and soxB genes ...... 118 3

3.4.8 Expression of 16S rRNA genes from Fe oxidizing bacteria118 3.4.9 Expression of soxB and F. myxofaciens genes in different plant stand in W2 sediment ...... 121 3.4.10 ITRAX analysis ...... 122 3.4.11 Total and dissolved organic carbon and nitrogen concentrations along core depths ...... 127 3.5 DISCUSSION ...... 130 3.5.1 Surface sediment bacteria and Fe and S speciation along the wetland ...... 130 3.5.2 Metal distribution along sediment depths ...... 131 3.5.3 Expression of S metabolic genes ...... 134 3.5.4 Expression of Fe bacteria ribosomal genes ...... 135 3.6 CONCLUDING REMARKS ...... 137 3.7 SUPLEMENTARY DATA ...... 139

Chapter 4: EFFECT OF OXYGEN AND ORGANIC SUBSTRATES ON BACTERIAL ACTIVITY AND METAL ATTENUATION IN INCUBATED WETLAND SEDIMENTS ...... 143 4.1 ABSTRACT ...... 144 4.2 INTRODUCTION ...... 145 4.3 MATERIALS AND METHODS ...... 148 4.3.1 Sediment sampling ...... 148 4.3.2 Experimental design ...... 149 4.3.3 Measurement of environmental variables ...... 151 4.3.4 DNA extraction and shotgun metagenomic sequencing ... 152 4.3.5 Bioinformatic analysis ...... 153 4.3.6 Statistical analysis ...... 154 4.4 RESULTS ...... 154 4.4.1 Changes in water chemistry during microcosm incubations ...... 154 4.4.2 General characteristics of the sediment metagenomes ... 177 4.4.3 Comparison of taxonomy profiles ...... 177 4.4.4 Comparison of functional profiles ...... 184 4.4.5 Sulphate metabolism analysis ...... 187 4.5 DISCUSSION ...... 198 4.5.1 Changes in S and metal chemistry ...... 198

4.5.2 Differences in bacterial taxonomy and metabolism ...... 201 4.6 CONCLUDING REMARKS ...... 207

Chapter 5: GENERAL DISCUSSION AND CONCLUSIONS ...... 210 5.1 SUMMARY AND KEY FINDINGS ...... 210 5.2 RESEARCH IMPLICATIONS AND RELEVANCE ...... 216 5.3 FUTURE DIRECTIONS ...... 223

6 REFERENCES ...... 225

Word count: 44,049

ABSTRACT

A natural wetland that has historically received acid mine drainage

(AMD) from the abandoned metal mine at Parys Mountain (Anglesey,

UK) has proved efficient in the removal of Fe and other trace metals.

This study aims to evaluate the microbial and physicochemical mechanisms involved in the metal attenuation process observed along the wetland by assessing the role of the microbial communities and microbial – derived changes in metal chemistry that leads to an improvement in the water quality.

A combination of molecular microbiology approaches, metal analytical techniques and microcosm experiments show that the wetland retains the diversity and metabolic structure of sediment bacteria communities despite the high acidity and metal concentration. In the middle of the wetland, increased bacterial activity related to Fe and S oxidation coincided with a removal of sulphate and metals from the water column. Incubation experiments of wetland sediments showed that the presence of bacteria generated higher levels of sulphide and particulate metals in the water column and the abundance of metabolic pathways related to AMD remediation was influenced by organic matter through increase in bacterial sulphate reduction activity. The findings of this study suggest that sediment bacteria are a key component that contributes to the immobilisation of Fe and trace metals observed in adapted natural wetlands to AMD pollution.

DECLARATION

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

I. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes.

II. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance Presentation of Theses Policy You are required to submit your thesis electronically Page 11 of 25 with licensing agreements which the University has from time to time. This page must form part of any such copies made.

III. The ownership of certain Copyright, patents, designs, trademarks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions.

IV. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see http://documents.manchester.ac.uk/DocuInfo.aspx?DocID=244 2 0), in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations (see http://www.library.manchester.ac.uk/about/regulations/) and in The University’s policy on Presentation of Theses.

ACKNOWLEDGMENTS

I gratefully acknowledge receipt of a PhD scholarship, funded by The National Fund for Scientific, Technological Development and Technological Innovation of Peru (FONDECYT).

I want to express my gratitude and appreciation to my supervisor Dr Jon Pittman and co-supervisor Dr Keith White for giving me the opportunity to work in this project and for their continuous motivation, help and guidance. Special thanks to Dr Andrew Dean for his crucial help during my first days in the lab and for his expertise throughout this work. I also want to thank my advisor Dr Franciska De Vries for her valuable comments during key milestones of this research.

I am sincerely thankful to Mariela Aguilera, Congyu Hou, Yingnan Hao, James Wakelin and Harold Garner for their assistance during field work and all the hard work implied. I would also like to acknowledge all the people from different laboratories at the Faculty of Science and Engineering that provided technical support. In particular to Paul Lythgoe and Deborah Ashworth for their help with crucial analyses performed in this work.

I would like to thank all the people from Plant Science that provided me a comfortable place to carry out my research. Special gratitude to Javiera, Helena, Amirul, Jeanette, Tom, Owen, Alejandra, Stefano, Matthew and Cecilia for their help at different moments during my PhD.

I am extremely grateful to my family for their love and constant support while dealing with the distance. This work is dedicated to them.

Chapter 1: GENERAL INTRODUCTION

1.1 ACID MINE DRAINAGE

Since the Stone Age, man has mined nearly 1150 million tons of Cu,

Pb, Co, Zn, Cd and Cr and this activity has been increasing over the centuries (Sheoran and Sheoran, 2006). Metal and coal mining from abandoned and currently active mines has released large amounts of contaminants such as trace metals, cyanides and sulphides

(Azapagic, 2004). Effluents released from mines with potentially toxic levels of acidity and metal ions are known as acid mine drainage

(AMD). The generation of AMD requires three factors: a sulphide mineral, an oxidant and a solvent. Therefore, when sulphide rocks are exposed to the environment and there is a water source such as rainfall or groundwater, AMD generation will occur. Iron pyrite is involved in the vast majority of these scenarios as is commonly found with other reduced minerals and oxides. Iron pyrite is also an important component in many coal deposits (Doulati Ardejani et al.,

2010) and for this reason, AMD is commonly produced due to anthropogenic activities related to both metal and coal mining. There are complex and diverse series of chemical reactions involved in the generation of AMD. However these can be summarised as two key reactions (Marchand et al., 2010). The first is the oxidation of iron pyrite:

+2 -2 + 2FeS2 + 7O2 + 2H2O = 2Fe + 4SO4 + 4H (1)

There is a significant release of protons which produces the acidity characteristic of AMD with the pH ranging between 1.5 and 3 (Bond et al., 2000, Gonzalez-Toril et al., 2003, Ferris et al., 2004). The second and subsequent reaction is the oxidation of Fe+2 to Fe+3:

+2 + +3 4Fe + O2 + 4H = 4Fe + 2H2O (2)

The production of ferrihydrite gives the characteristic red/brown colouration of AMD streams. In addition, Fe+3 ions can also oxidize more iron pyrite with the consequent solubilisation of further Fe+2:

+3 +2 -2 + FeS2 + 14Fe + 8H2O = 15Fe + 2SO4 + 16H (3)

The reactions described above occur spontaneously but Eq. (1) and

(2) can also be catalysed by microbial activity. The most commonly encountered bacteria responsible for catalysing these reactions are chemolithotrophic acidophile species such as Acidithiobacillus ferrooxidans, Acidithiobacillus thiooxidans and Leptospirillum ferrooxidans (Harrison Jr, 1984, Kelly and Wood, 2000). Under oxic and acidic conditions, these microorganisms use Fe and S compounds from pyrite and other similar minerals as electron donors for their aerobic respiration (Baker and Banfield, 2003). Iron oxidizing bacteria

(i.e. A. ferrooxidans) oxidize Fe+2 to Fe+3 and the S oxidizers (i.e. A. thioxidans) oxidize S compounds to sulphuric acid. Both processes take place in the external membrane of the cell. The final effect of these activities is the biological enhancement of pyrite and Fe+2 oxidation. Many studies suggest that an initial spontaneous oxidation of pyrite by primary Fe+3 ions as described in Eq. (3) seems to be 11 required for the microbial oxidation to occur (Curutchet et al., 1995,

Tributsch, 2001, Rodríguez et al., 2003, Yu et al., 2008). Current models suggest that bacterial extracellular polymeric substances

(EPS) bind Fe+3 ions and concentrate them at the mineral/cell interface (Sand et al., 2001, Kinzler et al., 2003, Vera et al., 2013).

Therefore the attachment of microbes to the mineral creates a suitable microenvironment that enables oxidation to occur. The main microbial oxidation takes place during the release of soluble Fe+2 and

S compounds. Soluble Fe+2 is used as an electron donor as part of the electron chain reaction carried in the bacterial membrane in which molecular oxygen is the electron acceptor. In this process, cytochrome c (encoded by the Cyc2 gene) located in the outer membrane, acts as the primary acceptor of Fe+2 electrons (Yarzábal et al., 2002), which are subsequently transferred to a periplasmatic protein called rusticyanin (Cobley and Haddock, 1975). This protein transfers the electrons to a second cytochrome that passes them through the rest of the chain reaction and ATP production machinery

(Appia-Ayme et al., 1998). Therefore the bacterial-mediated oxidation of Fe+2 generates a constant amount of Fe+2 necessary for the continuous solubilisation of sulphide minerals and the generation of AMD. Thiosulphate is the major S compound released during pyrite oxidation; it is also used by S oxidizers as an electron donor using c- and a-type cytochromes and catalysed by thiosulphate oxidase enzymes (Aleem, 1965). Sulphuric acid produced due to thiosulphate

12 oxidation contributes to the acidity necessary for the main chemical and microbiological reactions involved in AMD.

1.2 ENVIRONMENTAL IMPLICATIONS OF ACID MINE

DRAINAGE

1.2.1 Presence of trace metals

Iron pyrite is not the only sulphide mineral involved in AMD. There are other metal/metalloid containing compounds such as arsenopyrite, chalcopyrite, sphalerite, galena, molybdenite and pentlantite (Gray, 1998). These compounds usually contain chemically bound and non - ferrous metals that are released when these minerals are oxidised and acidity is produced. Levels of acidity generation vary depending on the protons that can be released from sulphide minerals. For example, experimental weathering of different sulphide compounds has shown that pyrite and spharelite produced more acidity than galena and chalcocite (Jennings et al., 2000).

In the case of Au rich pyrite, Au can constitute approximately 44% of the total arsenopyrite while silver (Ag) has been found in concentrations around 800 - 850 ppm in chalcopyrite and galena

(Cabri, 1992). Such ores are characterised by large but variable amounts of iron pyrite but also may contain enough valuable metal to make their extraction economically viable. During and after the mining of thes e ores, metals such as Cu and Zn will therefore be found in the mine waste. For example, sampling of mine waste areas of an disused old Spanish Pb-Zn mine revealed concentrations of

28,453 ppm Pb, 70,000 ppm Zn, 21 ppm Cd and 308 ppm Cu

(Rodríguez et al., 2009).

The degree of AMD pollution varies depending on the acidity and type and concentration of the metal. For example, the inactive underground mine at Iron Mountain in California presents a massive exposed sulphide ore body which generates AMD with extreme low pH (0.5 to 1) (Nordstrom and Alpers, 1999). Gold mining operations in South Africa have produced several mine tailings with AMD streams that present pH levels ranging from 2.3 to 2.5 in drainages near the tailing and pH values <5 in drainage away the tailing area (Tutu et al., 2008). Metal concentrations depend on the site geology and Fe is usually the most abundant metal. For example, an abandoned mine in Italy contained Fe, Cu, Pb, Zn and Mn concentrations of 717 ppm,

11 ppm, 0.02 ppm, 30.5 ppm, and 14 ppm, respectively (Benvenuti et al., 1997) while an abandoned Pb-Zn sulphide mine in Turkey generated AMD with different concentrations of Fe (87 - 464 ppm),

Mn (2 – 11 ppm), Zn (5 - 300 ppm) depending on the sample site within the water column (Aykol et al., 2003).

1.2.2 Environmental impact

By the year 1989, approximately 72,000 ha of lakes and reservoirs and 19,300 km of streams and rivers worldwide were polluted by

AMD (Johnson and Hallberg, 2005) and the U.S. Environmental

Protection Agency concluded that 4590 km of streams were acidic due to AMD in South-eastern United States (Herlihy et al., 1990). By

1995, almost 20,000 km of streams and rivers had been damaged by

AMD in the southern United States (Ziemkiewicz et al., 2003). It has been estimated that 193 tonnes of Zn, 18.5 tonnes of Pb, 0.64 tonnes Cd, 19.1 tonnes of Cu, 551 tonnes Fe, 72 tonnes Mn and 5.1 tonnes As are released from abandoned mines to streams in England and Wales every year (Mayes et al., 2010). In Africa, destruction of large natural vegetated areas due to AMD (pH ~ 2.3) and high concentration of Al (Bell et al., 2001) has been reported. Discharge loads of dissolved Fe and Al ranging from 20 to 1000 kg/hour has been observed in coal mines from New Zealand (Davies et al., 2011).

Acidification of freshwater produces ionic and acid-base disturbance in the homeostasis of many organisms causing organ failures and harmful changes in ion flux rates (Perry et al., 2003). Metal speciation is also influenced by low pH causing an increase in metal mobility and adsorption on biological surfaces, therefore increasing the toxicity in microbiota (Franklina et al., 2000, Wilde et al., 2006) and macrobiota (Mason et al., 2000).

AMD–affected freshwater ecosystems often show evidence of considerable loss of biota richness and diversity. Photosynthetic species sensitive to acidity are replaced by more tolerant species such as Chlamydomonas acidophila (Dean et al., 2019), certain filamentous and single-celled green algae (Chlorophyta) and single- celled protist Euglenozoa (Bray et al., 2009). Metal deposits on freshwater streams also contribute to the reduction of phytoplankton

15 and periphyton biomass due to toxicity (Niyogi et al., 1999).

Therefore, there is a substantial decrease in primary production in the food chain (Niyogi et al., 2002). Invertebrate species are severely affected in their diversity and abundance. In many cases only a few species remain (e.g. coleoptera, neuropteran, and the Diperan chironomidae) (Winterbourn, 1998, Dean et al., 2019). Loss of species is mainly because their physiological limits cannot resist the shifts in pH and metal concentrations (Rainbow, 2002). Loss of habitats is also a deleterious factor due to deposition of precipitates in AMD streams with high insoluble metals (i.e. Fe) which smother the sediment and clog the interstices thereby reducing oxygen availability to the infauna (McKnight and Feder, 1984).

Bioaccumulation of metals in fish mainly occurs in the liver, gill, kidney, skin and intestines and then transferred into the food chain

(Liu et al., 2012). Trace metals from AMD streams can diminish mucus production and number of globet cells on the gill surface of fish such as carp (Bols et al., 2001) followed by damage in kidney tissues (Vinodhini and Narayanan, 2008). As in invertebrates, ochre deposition reduces fish hatching (Brenner and Cooper, 1978, Steffens et al., 1993) and metal ions can penetrate egg shells causing disturbance in embryonic development (Jezierska et al., 2009). Fish are therefore commonly absent from severely impacted AMD affected ecosystems. Fluvially transported sediments are a significant driver of river metal flux, and the fluvial system is particularly important for the transport and dispersal of metals in terrestrial and aquatic 16 environments. Therefore, fluvial metal transport allows these contaminants to reach distant and less polluted ecosystems (Hogsden and Harding, 2011).

Cadmium exposure generates oxidative stress by inducing reactive oxygen species (ROS) (Wang et al., 2008). Lead can inhibit or mimic essential ions such as Ca, Fe and Zn, interfering with normal functions such as neuronal and muscular activity (Marasinghe Wadige et al., 2014). The metalloid As can affect cellular respiration by uncoupling oxidative phosphorylation (Lage et al., 2006).

Agricultural land is affected when the irrigation water is contaminated with AMD pollution. Zn and Cu are often found associated with organic matter in the soil and As and Cu can be detected in high concentration within crop tissues (Lin et al., 2005). Soil contamination from mining is a major problem in many parts of the world, including China where very large areas are contaminated (Su,

2014). Assessment of trace metal concentrations in a

Agricultural soil irrigated with acid mine wastewater in the surroundings of the Dabaoshan mine in South China revealed high levels of Cu, Zn, Pb and Cd (Zhuang et al., 2009). Crop analysis in the same area evidenced metal levels above the permissible limits in rice and vegetables (Lin et al., 2005). The same effect was observed in rice from long-term mine impacted paddy soils in Macedonia

(Rogan et al., 2009). Soil analysis near AMD discharges in the

Bolivian highlands revealed maximum trace metal concentrations of

Cu, Pb, Fe, Zn, Mn and Cd in agricultural zones and exceeded guidelines levels of Cd, Pb and Zn in potatoes (Garrido et al., 2009), especially Cd for which the dry weight concentration in potato tubers was 1.2 ppm (Oporto et al., 2007).

1.3 ACID MINE DRAINAGE REMEDIATION

The most common method for AMD remediation is the use of active treatment plants where addition of alkaline material (e.g. Ca oxides,

Ca carbonates, Mg oxides, Na hydroxides) promote the neutralization of the pH and accelerates the oxidation of Fe ions (Coulton et al.,

2003). The sludge produced has to be neutralized and high volumes of waste are produced (Macías et al., 2017). Passive treatments require less maintenance and include the use of anoxic limestone drains to reduce acidity and to maintain Fe in its reduced state, avoiding the production of oxides that can diminish neutralization efficiency (Kleinmann et al., 1998). The design of this method varies depending on the net acidity, Fe+3, Al and dissolved oxygen of the effluent (Ziemkiewicz et al., 2003). Passive treatments also include the complete or partial use of biological agents to remediate AMD. For example, successive alkalinity–producing systems incorporate organic compost to the anoxic limestone drains technology to generate biological sulphate reduction (Kepler and McCleary, 1994).

Bioreactors using sulphate-reducing bacteria is a fully biological technology that produce a more stable sludge and generate less operation costs and minimal energy consumption (Neculita et al.,

2007). Other bioreactor technologies include the use of Fe oxidizing 18

(Long et al., 2003) and sulphidogenic bacteria (Johnson, 2000).

When long-term solutions are required, the use of constructed wetland that mimic the AMD attenuation observed in AMD–impacted natural wetland is utilized (Sheoran and Sheoran, 2006).

1.3.1 Acid mine drainage remediation by wetlands

Wetlands are permanent or temporary inundated ecosystems with characteristic flora and fauna (micro- and macroinvertebrates, wildfowl). They play important roles in the environment such as water storage, photosynthetic carbon production and habitat diversity

(Knight, 1992). Wetlands are however vulnerable to land reclamation for agriculture (Healy and Hickey, 2002). Several studies have demonstrated the efficiency of wetland systems as biological filters for pollution control (Knight, 1992, Baker, 1992, Brix, 1994, Haberl et al., 2003, Babatunde et al., 2008). AMD pollution can also be remediated by wetlands via a number of biological and physicochemical mechanisms.

1.3.1.1 Physicochemical processes

Deposition can be defined as the process in which particles are added to the soil or sediment via gravity as they lose kinetic energy in the fluid. Metals can be deposited in a wetland due to the decrease of water flow rates caused by the large surface area of the wetland

(Ranieri and Young, 2012) and by the presence of vegetation

(discussed below) (Teuchies et al., 2013). Wetlands therefore act as a sink in which metals are removed from the mine drainage. Density

19 is a crucial factor in this process, so metals need to have sufficient density to deposit (Marchand et al., 2010). Deposited metals are in a particulate state or adsorbed to particulate metals (in particularly hydrated Fe oxide) and/or organic matter plus other kinds of suspended material (Cheng et al., 2002).

Precipitation refers to the formation of a solid in a solution or in association with another solid. Wetland systems usually provide the conditions for metals to precipitate, forming insoluble compounds such as oxides, oxyhydroxides, hydroxides during oxidation

(Hafeznezami et al., 2012) and sulphide compounds due to metal reduction (Wu et al., 2013). Those insoluble compounds are then trapped on and in the substrate which leads to the separation of metals from the water column. Redox potential, pH and anion concentrations are important variables influencing metal precipitation

(Frohne et al., 2011). Oxygen can be provided by plants both in the water column and rhizosphere (Armstrong et al., 2000), and sulphide is mainly produced by microbial reduction of sulphate and sulphuric acid in AMD streams (Webb et al., 1998) as discussed below.

Adsorption is another important process involved in metal removal.

Metal ions can be adsorbed to sediments, organic matter or eukaryotic and prokaryotic cell walls of dead and living organisms

(Machemer and Wildeman, 1992). The process can be based on physical or chemical reactions (Marchand et al., 2010), and metals

20 may compete for adsorption sites (Seo et al., 2008) and with other ions such as Ca, Mg and K (Echeverria et al., 1998).

1.3.1.2 Role of plants

In a wetland system, the interaction of aqueous metals with plants is extremely difficult to predict as plants modify metal speciation, redox conditions, pH, organic matter content and microbial activity, which influences the mobilization or immobilization of metals (Jacob and

Otte, 2003). However, it is believed that the presence of plants in wetlands increases the efficiency of metal removal from the water column in a process called phytoremediation (Rai, 2008).

Plants can create suitable conditions for the physicochemical process outlined above to occur (Ali et al., 2013). Mine drainage passing through vegetated areas have decreased velocity so stream flow is reduced, and mobilization of metals is diminished due to deposition of metal and an increase in residence time, which allows more interaction of the metals with substances enhancing metal precipitation; both processes are known as phytostabilization (Weis and Weis, 2004).

A more active role of wetland plants involves the release of oxygen through the roots which induces chemical changes in the rhizosphere and a subsequent immobilization of metals (Yang et al., 2010). The oxygenated sediment layers surrounding the roots can create ferric hydroxide precipitates (Fe plaques) in the root surface (Crowder and

Macfie, 1986) and these plaques also sequester other trace metals

21 from the water through adsorption and co-precipitation (Tripathi et al., 2014).

Plant decay, including dead root tissues and the release of organic acids, increases the organic matter content of the sediment (Jacob and Otte, 2003). Organic matter can strongly bind to metals and reduce their mobility (Twardowska and Kyziol, 2003). Organic matter can reduce metal mobility by decreasing redox potential and can influence metal mobilization through bacteria growth (Kashem and

Singh, 2001). Moreover, plants growth and abundance can change total concentration of C and N in soils (Liao et al., 2008) and therefore, variations in these nutrients can alter the diversity and function of several bacteria phyla (Cruz-Martínez et al., 2009).

Wetland plants also have the ability to uptake and store metals in their tissues (McGrath et al., 2002). Wetland plants such as

Phragmites, Typha, Azolla, and Eichhornia are the most common plants with metal phytoextraction properties (Rai, 2008). Metals entering the plant can be translocated to above-ground tissues such as the leaves and stems (Weis and Weis, 2004) but they mainly remain in the roots (Dean et al., 2013). However, plant metal uptake represents a minor component of the metal removal from AMD in wetlands, ranging from 0.5 to 1.5% of total removal (August et al.,

2002, Dean et al., 2013).

1.3.1.3 Role of microorganisms

Prokaryotes have an important role in AMD remediation within wetlands. The main remediation processes mediated by bacteria involve production of alkalinity and generation of reducing conditions in which metals can precipitate. Because of the high concentration of bacteria, sulphate reduction is the process that contributes the most to metal removal in AMD-impacted environments, including wetlands

(Johnson and Hallberg, 2005).

There are around 60 genera containing 220 species of sulphate reducing bacteria (SRB) so far identified (Muyzer and Stams, 2008).

Desulfovibrio, Desulfobulbus, Desulfotomaculum, and

Desulfosporosinus spp. are the main taxa found within wetland ecosystems (Pester et al., 2012). In AMD affected wetlands, the sulphate reduction to hydrogen sulphide (H2S) catalysed by SRB generates the metal (e.g. Zn, Cu and Cd) precipitates with H2S forming high insoluble sulphides in a reaction described as follows

(Johnson and Hallberg, 2005):

+2 + M + H2S  MS + 2H (4)

Under similar anoxic and reductive conditions, some groups of microorganisms known as metal reducing bacteria (MRB), can reduce metals, mostly Cr(VI), Mn(IV), Fe(III), U(IV) and Se(IV) by using them as sole electron acceptors (Tebo and Obraztsova, 1998). Thus both SRB and MRB reduce the mobility of the metals by precipitation as the insoluble sulphide.

Other microbial process such as denitrification (nitrate reduction) and methanogenesis may collaborate in metal removal by creating reducing and alkaline environments but they are less importance in most wetlands due to low availability of the substrates required

(Johnson et al, 2005).

Oxidation mediated by bacteria are also important processes, especially in the formation of biogenic Fe(III) oxides capable of absorbing other trace metals (Kappler and Straub, 2005). Bacteria able to oxidise Fe can be divided into acidophiles (responsible for

AMD generation) and neutrophils that transform Fe(II) to Fe(III) through different mechanisms such as coupled to nitrate reduction or photosynthesis (Hedrich et al., 2011). The use of neutrophilic iron- oxidizing bacteria can be a promising strategy to remove metals from mine drainages (Langley et al., 2009). Mn oxides are very strong oxidants that can also precipitate toxic metals (Miyata et al., 2007).

Although these oxides occur naturally they are also produced as a result of bacteria activity (Tebo et al., 2004).

1.4 CASE STUDIES OF ACID MINE DRAINAGE REMEDIATION

BY WETLANDS

1.4.1 Constructed wetlands

Constructed wetlands are a passive treatment method considered as an alternative to chemical AMD remediation. The aim is to replicate the remediation mechanisms of natural wetlands in a controlled

24 environment (Sheoran and Sheoran, 2006). Aerobic constructed wetland systems are designed for enhancing the interaction of oxygen with metals, therefore inducing the precipitation of metals as hydroxides, oxyhydroxides and oxides while anaerobic constructed wetlands, also known as compost wetlands, are amended with organic substrates for enhancing microbial sulphate reduction and the subsequent precipitation of metals as sulphides (Gazea et al., 1996).

In many cases, constructed wetlands are also augmented with limestone to increase alkalinity-mediated metal precipitation (Scholz and Lee, 2005).

Plants are also used in constructed wetland for metal removal. For example, the formation of Fe oxides contributed to the retention of

86% of total Fe in wetland plants such as Sphagnum sp. and Typha latifolia (Wieder et al., 1990). In addition, Henrot et al. (1990) studied the retention of Fe and Mn and its correlation with microbial activity, pH, biomass and metal concentration in laboratory microcosms amended with AMD. Results showed that precipitation of metal oxides was the dominant process for remediation, resulting in the removal of 62% of total Fe and the addition of an antiseptic

(formaldehyde) severely decrease Fe-oxide precipitation, suggesting that the process was microbial-mediated (Henrot and Wieder, 1990).

In another study, a constructed wetland in the Idaho Springs-Central

City mining district of Colorado with a substrate of compost, manure and decomposed wood was used during two years for remediation of

AMD with high concentrations of Fe, Mn, Cu and Zn. Machemer et al.

(1992) showed that during the first 80 days, Fe was completely removed and Mn and sulphate were removed up to 10% and 20% respectively. Sulphate-reducing microorganisms were isolated from the substrate and 0.5 mmol/L of sulphide was measured in the water outflow, thus suggesting that precipitation of metal sulphides induced by microbial sulphate reduction was the main process involved

(Machemer and Wildeman, 1992). A larger (0.39 ha) constructed wetland including anaerobic cells for stimulating microbial sulphate precipitation for remediation of a AMD stream in south-eastern Ohio,

USA was analysed by Mitsch et al (1998). Fe decreased by an average of 170 mg/L to 32 mg/L (80.8%), 14% of Fe was found in the sediment surface and 7.2% in the upper 30 cm. Aluminium concentration was reduced from 83 mg/l to 56 mg/l (32.7%) from inflow to outflow. Manganese levels were low and they changed little during the process (inflow: 5.86 mg/L, outflow: 5.52 mg/L). The wetland inflow had a sulphate concentration of 1672 mg/l which decreased along the system to 1216 mg/l (27%). The pH of the influent varied from 2.73 to 3.08 and only increased to 3.51 in the effluent; therefore there was no significant decrease in acidity (Mitsch and Wise, 1998). In the majority of cases, long-term constructed wetlands can show a decrease in efficiency with time. For example, a wetland treating AMD from an abandoned underground mine in

Kentucky, USA was examined by Barton et al. (1999). During the first year the pH increased from 3.05 to 7.20 but decreased after 3 years

26 due to a depletion of the limestone placed on top of the substrate.

The high pH achieved during the first years allowed the reduction of

Fe and Al levels (95%), plus Mn and sulphate (85%). However, metal reduction decreased during the following years to an average of 22% and reached a critical point when metal concentrations in the wetland were higher than in the inflow. Accumulation of precipitate layers and dead vegetation also seemed to contribute to the process failure. The analysis indicates that the constructed wetland was only efficient during the first 6 months of treatment, after that period the lack of alkalinity and metal overload were the main failure factors (Barton and Karathanasis, 1999). New technologies have however been developed to increase the long term efficiency of constructed wetlands. Reducing and alkalinity producing systems (RAPS) are based on a mixture of limestone and organic matter for promoting sulphate reducing microorganisms. Adequate permeability and filtration is necessary and the design includes a large cross – sectional area and high head pressures. RAPS have been shown to continually treat AMD for more than 20 years without the need of substrate replacement or augmentation (Jarvis et al., 2002) but regular flushes are needed.

Successive alkalinity producing systems (SAPS), which are similar to

RAPS but use vertical flow where the metal precipitates accumulate on coupled aerobic pounds, thereby avoiding clogging, have also yielded promising results for long term treatment (Nairn and Mercer,

2000). 27

In summary, constructed wetlands have proved to be efficient for removing metals from the water column with around 20% to 60% of metal removed as it is possible to create a suitable environment for the proliferation of sulphate reducing microorganisms and plant species. However, surveillance to assess degree of clogging and a staged renewal of substrate is necessary in order to maintain alkalinity during long-term treatment of AMD.

1.4.2 Natural wetlands

Efficiency of natural wetlands varies depending on physicochemical characteristics, microbial and macrobiota ecology (Webb et al., 1998,

Vymazal et al., 2007). May et al. (2001) compared the metal accumulation in plants and sediments between two wetlands constructed by the Tenessee Valley Authority in the south-eastern

USA and a natural wetland nearby, all of them receiving AMD. The results showed that at near neutral pH, constructed wetlands have a

Fe and Mn removal efficiency of 98% and 79%, respectively, whereas the natural wetland was only able to remove 10% of Fe and 40% of

Mn (Mays and Edwards, 2001). However, when there are ideal conditions for maintaining lower pH, some natural wetlands can provide optimal results. For example, in a natural wetland in eastern

Finland receiving AMD, decrease in Cu, Fe, Ni, Zn and S concentrations has been observed at pH values ranging from 2.8 to

3.5 (Vymazal et al., 2007). These results are consistent with the natural attenuation processes in acid streams involving the adsorption of metalloids such as As onto Fe mineral precipitates (Asta 28 et al., 2010). Therefore, precipitation by microbial induced metal sulphides may not always be the main factor for removal of contaminants in natural wetlands. In some cases, natural wetlands can generate anoxic and alkaline sediments with favourable redox potentials for bacterial sulphate reduction; however detailed analysis has shown that this process is not predominant due to insufficient decomposable organic substrates (Tarutis et al., 1992).

Natural variations such as season and water flows are important in natural wetlands efficiency. August et al. (2002) studied during one year a natural wetland receiving AMD from an abandoned mine in

Colorado, USA. During the whole year, 95% of Fe was removed by metal deposition in the form of Fe oxides, with no significant seasonal variations; however, removal of Mn was much lower and subject to seasonal differences due to Mn uptake by plants during growing season but Mn release following plant death at the end of the season.

The balance of metal concentrations during the year of study revealed the loss of metal accumulation capacity and net release of metals, therefore the wetland was acting as a metal source rather than a metal sink (August et al., 2002).

Palmer et al. (2014) performed adsorption experiments for estimating the long term potential of Ni, Sb and As accumulation by a natural wetland receiving mine effluent and drainage waters over six years and concluded that maximal adsorption had not yet been reached; therefore there is a potential for long term treatment. However,

29 metal concentration in the sediment had already reached toxic levels and field studies involving hydrological and environmental analysis suggest that factors such as metal leaching, dilution of mine waters, temperature and redox conditions may need to be critically assessed to ensure safe long term treatment (Palmer et al., 2015).

To date, studies have examined the processes involved in trace metal remediation by wetlands but little is known about the relative importance of each process (such as metal oxidation/reduction mediated by bacteria and metal adsorption onto sediments and organic matter), the interaction between them and the effect of natural conditions such as seasonal changes in temperature and plant growth. A better understanding of the reasons for changes in metal retention, in particular the decline in retention efficiency, during long term wetland treatment is also required.

1.5 PARYS MOUNTAIN

Parys Mountain is an abandoned copper mine located on the Island of

Anglesey, in Wales, UK that continuously releases mine water to the southern Afon Goch (SAG) (‘Red River’ in Welsh; also known as Afon

Goch Dulas) and the northern Afon Goch (NAG) (Figure 1.1a). The

SAG has a length of 12 km and is fed by old drainage ponds immediately below the mountain and enters the Irish Sea via the estuary known as Dulas Bay (Younger and Potter, 2012). At 2.2 km downstream from the source the SAG enters a natural wetland that has shown long term efficiency for metal remediation (Dean et al.,

2013). In order to prevent flooding, in 2003 drainage operations were performed in Parys Mountain in which much of the subterranean mine drainage was diverted to an existing adit creating a significantly enhanced AMD stream entering the NAG. Before this deviation, the natural AMD stream of the NAG showed metal values below the UK drinking water standards (except for Fe and Cd) (unpublished data from Natural Resources Wales). Therefore, this river has been receiving additional pollution (generating high concentrations of all metals monitored) for approximately 15 years and as there is no wetland surrounding the NAG, no evidence of improvement on the water quality of NAG has been observed. The presence of a long-term polluted well-adapted wetland system and a short-term non-adapted river without a wetland, both contaminated by the same AMD source, create the ideal study site to understand the fundamentals of natural

AMD mechanisms. Figure 1.1 shows the location of Parys Mountain, the SAG wetland and the NAG.

Early studies of the pollution in the SAG revealed a very acidic stream in the upper reaches (pH < 3) compared to the tributary rivers (pH ~

6.8) and high concentrations of soluble Fe (110 ppm), Cu (10.6 ppm), Mn (7.7 ppm) and Zn (21.7 ppm) in the main stream water, and high concentrations of leachable Fe (971 ppm), Cu (65 ppm), Mn

(103 ppm) and Zn (157 ppm) in the sediments (Foster et al., 1978).

A more extensive study performed during a 14-month period showed that despite the diversion of mine-water to the NAG adit resulting in decrease in metal flux entering the SAG wetland, the water quality 31 showed the same acidity and demonstrated that metal concentrations of Fe, Al, Mn, Cu and Zn can reach dissolved levels of 259 ppm, 167 ppm, 49 ppm, 60 and 42 ppm, respectively (Boult et al., 1994).

c a

Figure 1.1. View of Parys Mountain, the surrounding rivers and the wetland. (a) Location of Parys Mountain, the SAG wetland and the NAG. The source of AMD that enters the SAG and the NAG is indicated. The river paths are highlighted in blue. (Sources: Google, DigitalGobe. Imagery Date: 6/1/2009. (b) View of the wetland in summer 2016 showing one of the ponds and the vegetation receiving discharges from the SAG. (c) View of the unpolluted NAG approximately 2 km. upstream of the adit. (d) View of the polluted NAG approximately 2 km downstream of the adit.

The precipitates that settled where the SAG enters Dulas Bay were analysed in detail to examine possible metal fixation in the sediments. Results revealed ochre with high concentrations of Cu

(13,000 ppm) and Zn (770 ppm) in the form of oxides on top of the sediments, and a sulphate reduction process located immediately below the surface formed an anoxic black mud largely composed of

Fe, Cu and Zn sulphides (Parkman et al., 1996). Analysis of deeper sediment showed that Mn and As were also fixed as sulphides in layers between 16 and 30 cm depth and corroborate the sorption of metal onto Fe, but also Mn oxides and oxydroxides, in the red brown oxic upper layers (Whiteley and Pearce, 2003).

Microbiological analysis by plating diluted samples of surface water in solid medium, revealed that the slight decrease of acidity (0.6 pH units) detected as the SAG stream flows away from the source allows a downstream increase of total acidophilic heterotrophic bacteria which seem to have an important role in metal immobilization.

Concentrations of Fe2+ was the main factor for shifts in the populations of Fe oxidizing bacteria such as Acidithiobacillus ferrooxidans predominant at the source and Leptosphirillum ferrooxidans more abundant further down the SAG (Walton and

Johnson, 1992). Molecular analysis of pore water and biofilm samples suggested that the dominance of L. ferrooxidans in the stream was due to its being a obligate aerobic while A. ferrooxidans which is facultative anaerobe was also detected in underground water samples with other acidophilic but obligate heterotrophs of the genus 34

Acidiphilium, Acidisphaera and Acidobacterium (Coupland and

Johnson., 2004). Further molecular analysis revealed that an Fe and

S oxidizing heterotroph related to the actinomycete group was the dominant organism, suggesting that the less polluted conditions of the SAG allows the presence of microbes with adaptable metabolism

(Bryan et al., 2004) compared with other more severely AMD- impacted streams (Bond et al., 2000, Gonzalez-Toril et al., 2003,

Ferris et al., 2004). Fluorescence in situ hybridization analysis of macroscopic biofilms downstream the discharge source of Parys

Mountain concluded that the microbial communities were dominated by β-Proteobacteria and terminal restriction fragment length polymorphism analysis evidenced that these populations have a very low overall diversity (Hallberg et al., 2006). These results are consistent with previous studies of the microbial ecology from abandoned Cu mines in temperate environments where the predominance of a few adaptable acidophilic heterotrophs is a common characteristic (Johnson et al., 2001).

The effect of plant populations on the metal distribution of the Parys

Mountain wetland has also been assessed. Previous studies demonstrated that the release of oxygen from the roots of wetland plants produces ferric plaques in the rhizosphere where metal ions are adsorbed, therefore decreasing the soluble metal concentration

(Crowder and Macfie, 1986, Hansel et al., 2001, Jiang et al., 2009).

However, this effect may not be present in the SAG wetland, where surface and pore water from sites colonized by plants revealed an 35 increase in Fe, Mn, Cu and S concentrations, therefore the interaction of plants with the metals is unclear (Batty et al., 2006).

Dean et al. (2013) studied the efficiency of the SAG wetland to remediate AMD over 14 years from 1997 to 2010. Results showed that the wetland causes metal retention despite the flow rate and despite the concentration of metal inflow. The mine drainage diversion that occurred during the collection of data (2003) allowed the study of the wetland during two regimes and showed that despite flow variations, the wetland was able to retain from 55% to 97% of

Fe. Analysis of biota evidenced the presence of a complex community of unknown bacteria species phylogenetically related to acidophilic bacteria. The study suggested that a combination of metal sequestration, metal sedimentation and alkalinisation promoted by a community of plants and bacteria species was likely to occur in the wetland. Further analysis including effect of plant roots in metal speciation and bacteria activity along the wetland was proposed in order to validate the complementary role of plants and bacteria in the remediation process.

It is clear that bacteria play an important role in AMD remediation by constructed and natural wetlands, including the SAG wetland.

Therefore it is important to examine microbial community diversity and metabolism in these systems in order to understand and improve the remediation processes involved. Analysis of microorganisms from

Parys Mountain effluents using traditional microbial methods and

36 molecular approaches has demonstrated the existence of common

AMD generating acidophilic Fe oxidizing bacteria in the spoil soil on

Parys Mountain (Bryan et al., 2004), the effluent entering the wetland

(Walton and Johnson, 1992) and in the NAG adit (Coupland and

Johnson., 2004). Molecular analysis of environmental DNA from sediment samples along the wetland have detected strains similar to heterotrophic bacteria with metabolism involved in Fe and S cycling

(Dean et al., 2013), suggesting changes in the microbial community due to the passage of the SAG through the wetland. However, diversity, abundance and metabolism of the bacterial community along the wetland remain unknown. Moreover, taxonomic and metabolic traits that can explain the role of bacteria in remediation process along the SAG wetland but not in the polluted NAG have never been evaluated.

The pH at the upper reaches close to the effluent from Parys

Mountain is below 3 and most metals remain in solution, however, the pH rise and a significant precipitation and deposition of ferrihydrite occurs downstream, indicating that along the approximately 11 km stream, the SAG acts as a sink for contaminants (Boult et al., 1994). A more detailed analysis of the wetland encompassing approximately the first 2 km of the stream, showed that the initial dissolved concentrations of Fe (~20 mg/L), Zn

(~10 mg/L) and Cu (~3 mg/L) decrease to less than 2 mg/L at 2500 m from the source (Dean et al., 2013). Moreover, at the first 1000 m approximating to the middle reaches of the wetland, a marked fall of 37

~ 50% in dissolved Fe, Zn and Cu was observed (Dean et al., 2013).

Metal sulphides immediately below the sediment surface where the

SAG reaches the sea in the Dulas Bay estuary confirms sulphate reduction (Parkman et al., 1996) but there is not data describing the biochemical process in the upper reaches of the river. Moreover, the reasons for the unexpected decrease of soluble metals in the middle reaches of the wetlands remain unclear.

It is known that wetland plants influence metal mobility through redox and pH changes in the rhizosphere, but whether the use of plants in constructed wetlands is favourable for this technology remains unclear due to the various variables involved (Jacob and

Otte, 2003). For example, oxygen release from plants increase biomass of microaerophilic bacteria that can enhance metal precipitation (Emerson et al., 1999). However, anaerobic environments facilitate the formation of metal sulphide than can easily be removed from the water column (Jong and Parry, 2003). On the other hand, organic matter released from plants can influence reducing conditions and therefore precipitate metals and can also bind to trace metals through chelation and complexation processes

(Grybos et al., 2009). However, organic matter is also needed for microorganisms involved in Fe/Mn oxidation (Ghiorse, 1984) and sulphate reduction (Lovley and Phillips, 1986).

1.6 HYPOTHESES, AIMS AND OBJECTIVES OF THE THESIS

Based on the existing background knowledge described in the previous section, the following hypotheses were formulated:

Hypothesis 1: The existence of the wetland in the SAG allows the presence of a more diverse bacterial community compared to the

NAG, which permits a number of metabolic processes related to the removal of trace metals from the water column.

In respect to this hypothesis, the following aim and objectives have been formulated:

Aim: To compare the taxonomy, diversity and metabolic activity of the bacterial communities from sediments of the SAG and the NAG and in relation to the pH and metal distribution in both systems.

Objectives:

• Characterization of pH, conductivity, Fe, S and trace metals

along the SAG and the NAG.

• Illumina-based sequencing of 16S rRNA genes from sediments

with varying levels of AMD pollution and with or without

wetland plants, within the SAG and the NAG.

• Analysis of diversity and taxonomic traits of the data obtained

from the 16S rRNA gene sequencing.

• Prediction of metabolic profiles of the sampled microbial

communities based metabolic inference techniques from 16S

rRNA sequence libraries.

Hypothesis 2: Bacterial activity related to Fe, S and trace metals precipitation increase from the beginning to the middle reaches of the

SAG wetland and facilitates the decrease in soluble metals in the water column and the continuous retention of metals in the surface

(oxic) and bottom (anoxic) layers of the sediments.

For this hypothesis, the following aims and objectives have been formulated:

Aim: To identify bacteria-mediated mechanisms of AMD attenuation in the water column and metal immobilization in different sediment layers along the SAG wetland.

Objectives:

• Identify difference in surface bacteria populations between

different vegetation types in the middle of the wetland.

• Measurement of Fe and S and trace metals partition in the

water column along the wetland.

• Measurement of Fe and S and trace metals partition in top and

bottom layers from sediments along the wetland.

• Measurement of Fe and S bacteria-mediated transformation in

top and bottom layers from sediments along the wetland.

• Analysis of metal distribution, mobility, C and N concentrations

in different core depths along the wetland.

Hypothesis 3: Oxygen and organic matter provided from wetland plants enhance bacteria activity related to metal transformation and reduce levels of soluble Fe and trace metals in the water column.

For this hypothesis, the following aims and objectives have been formulated:

Aim: To understand the effect of environmental factors provided by plants, specifically carbon and oxygen, on sediment bacterial activity and metal behavior under controlled exposure to AMD.

Objectives:

• Incubation of microcosms with, respectively, natural and

sterilized sediments treated with carbon (organic acids) and

oxygen (air in-flow).

• Measurement of Fe and S speciation and Zn and Cu partition

during an incubation period of 88 days.

• Metagenomic sequencing of DNA extracted from the natural

sediment microcosm.

• Comparative analysis of microbial taxonomic and metabolic

traits from the two microcosms.

• Analysis of functional genes related to Fe, Zn, Cu and S

metabolism.

Chapter 2: CHARACTERIZATION OF ACID MINE DRAINAGE POLLUTION AND MICROBIAL COMMUNITY SHIFTS IN THE SOUTHERN AND NORTHERN AFON GOCH

Aguinaga, O. E., McMahon, A., White, K. N., Dean, A. P., & Pittman, J. K.

The author analysed environmental and 16S rRNA gene sequencing data, performed metabolic predictions, interpretation and writing up the manuscript. Andrew Dean and Anna McMahon performed sample collection. Andrew Dean prepared 16S libraries for sequencing. Andrew Dean, Keith White and Jon Pittman provided full guidance and manuscript review. Data from this chapter was published in the following peer – review paper:

Aguinaga, O. E., McMahon, A., White, K. N., Dean, A. P., & Pittman, J. K. (2018). Microbial community shifts in response to acid mine drainage pollution within a natural wetland ecosystem. Frontiers in Microbiology, 9: 1445.

This chapters contains some additional data not present in the original published version.

2.1 ABSTRACT

Natural wetlands have been demonstrated to play an important role in remediating pollutants such us acid mine drainage (AMD) from abandoned mine sites. However, many aspects of the microbial processes that contribute to AMD remediation by wetlands remain unclear. To understand these processes, DNA analysis from bacterial communities in an AMD-polluted-river-wetland system was performed. Next-generation sequencing of bacterial 16S rRNA gene from river and wetland sediment samples identified variation in bacterial community structure and diversity on the basis of water/sediment chemistry parameters and wetland plant presence.

Metabolic reconstruction analysis allowed prediction of relative abundance of microbial metabolic pathways and revealed differences between sites that cluster on the basis of the severity of AMD pollution. Global metabolic activity was predicted to decrease in polluted river sediments in contrast to wetland and un-polluted sites, indicating a metabolic stress response to AMD pollution. This study confirms that wetlands maintain the diversity and metabolic structure of sediment microbial communities exposed to high levels of acidity and metal pollution, and suggests that such diversity is critical for the remediation action of the wetland.

2.2 INTRODUCTION

Microorganisms can mediate geochemical process such as element cycling in an attempt to adapt to different environments (Madsen,

2011). The characterization of these microorganisms in AMD polluted environments is necessary for an understanding of the impact of acidity and elevated metal concentrations on microbial ecology and the geochemical changes, including how such activity reverses the effect of AMD (Chen et al., 2015a, Méndez-García et al., 2015, Huang et al., 2016)

Next-generation sequencing and metagenomic tools have allowed improvements in the identification and quantification of bacteria taxa in AMD polluted water and sediment samples (Amaral-Zettler et al.,

2011, Kuang et al., 2013, Brantner and Senko, 2014, Liu et al.,

2014). These studies have shown a high dominance of acidophilic taxa and a reduced overall species richness and diversity. Moreover, it is been demonstrated that physicochemical characteristics, in particular pH, can predict changes in the community structure of bacteria in different AMD environments (Kuang et al., 2013, Liu et al.,

2014). Different metabolic activities such as oxidation and reduction of Fe, S, C, N and metals transportation have been characterized in

AMD environment and linked to specific bacteria taxa (Méndez-García et al., 2015). The evaluation of genes expression and genomic-based metabolic predictions has linked AMD – related metabolic processes with community structure and environmental quality of AMD streams

(Chen et al., 2015a, Hua et al., 2015, Kuang et al., 2016).

Analysis of microbial communities structure have been performed in

AMD water, sediments and biofilms (Huang et al., 2016). However, very few studies have used high-throughput sequencing approaches to examine sediment bacteria associated with wetland ecosystem

(Diaby et al., 2015), despite the potential for natural and constructed wetlands to remediate AMD (Sheoran and Sheoran, 2006, Mayes et al., 2009, Dean et al., 2013).

The AMD impacted Northern Afon Goch (NAG) river system provides an ideal study site to quantify and contrast the microbial community composition in river sediments of an impacted ecosystem without a wetland to a long-term impacted ecosystem with a substantial natural wetland. This wetland surrounds the Southern Afon Goch

(SAG) and which drains the same disused mine site, is predicted to have adapted to the AMD pollution. To the author’s knowledge, this study is the first direct comparison of a natural wetland and a river without a wetland that received the same AMD pollution. This comparison will elucidate the differences in bacterial composition and allow a prediction of changes in metabolic activities in wetland and river sediment sites. It is hypothesized that the wetland maintains a level of bacteria diversity which allows the presence of different metabolic traits that modify the bacteria habitat and lead to an improvement of the water quality along the wetland.

2.3 MATERIALS AND METHODS

2.3.1 Study site and sample collection

Water and sediments samples were collected from June 2010 to

October 2014 at 3 locations along the wetland surrounding the SAG and at 5 sites along the NAG (Figure 2.1). The NAG rises to the west side of Parys Mountain and runs approximately 5.5 km in length with the Dyffryn Adda adit from the Parys Mountain mine joining at approximately 2.5 km along the river. The river then runs north through the town of Amlwch and enters the Irish Sea through an industrial site next to the port. The river course is largely canalized.

Five sites were sampled (Figure 2.1) two upstream of the adit (site

N1 and N2), the Dyffryn Adda adit itself (site NA), and two sites downstream of the adit (N3 and N4). Sampling was carried out on four (at site N1) or six (sites N2–N4 and NA) occasions on June 2010

(all sites), July 2010 (all sites), August 2013 (all sites), October 2013

(not site N1), March 2014 (all sites), and October 2014 (not site N1).

The SAG is 11 km in length and runs south of Parys Mountain

(Figure 2.1). Approximately 500 m below the mine source at the

Mona adit, the river flows through a natural wetland of 0.1 km2.

The river then runs east and enters the Irish Sea at Dulas Bay.∼ Three sites were sampled within the wetland (sites S1–S3). Site S1 is an entry point for AMD runoff at the start of the wetland, while sites S2 and S3 are in the middle and at the end of the wetland, respectively

(Figure 2.1). Sampling was carried out on four (site S2) or seven

(sites S1 and S3) occasions on June 2010 (all sites), July 2010 (all 46 sites), November 2011 (all sites), August 2013 (all sites), October

2013 (not site S2), March 2014 (all sites), and October 2014 (not site

S2).

Sample sites were selected on the NAG in order to characterize the unpolluted (N1, N2) and AMD – impacted (N3, N4) sections of the riverNAG plus the AMD source (NA) that discharges between sites N2 and N3. Samples on the SAG along the wetland (S1, S2, S3) were chosen to examine compare the AMD attenuation process along the wetland in length and elucidate possible mechanisms. Samples were taken during predominant spring, summer and autumn seasons to coincide with the main growing season of the wetland and hence obtain samples during maximal wetland productivity.

Triplicate sediment and water samples for analysis of pH, conductivity

(as a measure of conductive ion concentrations), and dissolved, sediment, and particulate metals were collected at each site on each sampling occasion. Samples were processed within 5 h following collection.

Sediment samples for DNA extraction were collected in March 2014 and included an unpolluted wetland (UV) at the Cefni Reservoir located at approximately 13km. from Parys Mountain. Samples were taken from non-vegetated sediments at the NAG sites and from sediment surrounding Juncus sp. roots from sites S1, S2, S3 and UW.

A non-vegetated sediment sample from site S1 (S1R), and a sediment sample surrounding cottongrass (Eriophorum angustifolium)

47 roots from site S2 (S2C) were also included. Table 2.1 shows the location and description of each sample site included in this study.

2.3.2 Water and sediments quality analysis

Physicochemical parameters and metal content were analysed according to a previous study of the SAG wetland (Dean et al., 2013).

A YSI 556 probe (Xylem Analytics, UK.) was used to measure water pH and conductivity. For analysis of filtered metals a known volume of water was filtered through a 0.45 µm cellulose acetate filter and the filtrate stored in an acid-washed polypropylene bottle.

Figure 2.1. Sample sites within the Parys Mountain river catchment in Anglesey, Wales, United Kingdom. Sites on the SAG are marked S1– S3, with the wetland areas shaded in green. Sites on the NAG are marked N1–N4. NA is the location of the adit that discharges AMD to the river. Map modified from Dean et al, (2013).

Filtered samples were acidified to 2% with ultra-pure nitric acid to ensure the metals remained in solution. The pre-weighed filters were retained for analysis of metal particulates. Sediment samples were collected using a plastic scoop and sealed in a plastic bag until analysis. On return to the laboratory the filters containing suspended particulates and the sediment samples were air-dried at 80°C for

48h. The dried sediments were then passed through a 250 µm mesh stainless steel sieve to homogenize particles for digestion. The filter papers and 100 mg of disaggregated sediments were digested in 67% ultra-pure nitric acid for 24 h at 100°C. Digests were then diluted to

2% nitric acid in deionized Milli-Q water (Millipore, UK) and stored at

5°C prior to analysis. Water samples and digests were analysed by inductively coupled plasma atomic emission spectroscopy (ICP-AES) using a Perkin-Elmer Optima 5300 for Al, As, Cd, Cu, Fe, Mn, Pb and

Zn. The spectroscope was calibrated using an internal standard, which was a matrix matched serial dilution of Specpure multi element plasma standard solution 4 (Alfa Aesar, USA). Anions were not measured as this chapter was focused on the distribution of trace metals. Additional limitations were the availability of such as methods to analyse changes in speciation during transport return to the lab in the case of sulphate and other sulphur species.

Table 2.1. Description and location of each sample site in this study.

Sample Description Coordinates

53°15'57.8" N, UW Unpolluted wetland at Cefni Reservoir 4°21'0.6" W

Upper reach of SAG wetland immediately 53°23'13.5" N, S1/S1R below the river source and receiving AMD 4°19'31.9" W from Parys Mountain.

Middle of the SAG wetland at 53°22'54.7" N, S2/S2C approximately 0.8 km. from the river 4°19'42.4" W source.

End of the SAG wetland at approximately 53°22'17.1" N, S3 2 km. from the river source. 4°20'12.4" W

Approximately 2km. upstream of the adit 53°22'52.8" N, N1 without AMD impact. 4°22'02.0" W

Approximately 0.4 km. upstream of the 53°23'37.5" N, N2 adit without AMD impact 4°21'22.4" W

Adit in the NAG where AMD from Parys 53°23'42.3" N, NA Mountain enter the NAG. 4°21'02.9" W

Approximately 0.15 km. downstream of 53°23'47.2" N, N3 the adit polluted by AMD 4°21'03.6" W

Approximately 2km downstream of the 53°24'47.6" N, N4 adit polluted by AMD 4°20'21.7" W

2.3.3 DNA extraction

Environmental DNA (eDNA) was extracted from 100 mg of each sediment sample (3 independent samples per site) using a Powersoil

DNA isolation kit (MoBio Laboratories, USA) and quantified using a

Nano-drop 3300 (Thermo-Scientific, USA).

2.3.4 16S rRNA gene sequencing

The V3 and V4 region of the 16S rRNA gene create a single amplicon that was amplified using the extracted eDNA samples as template as previously described (Klindworth et al., 2013). The full sequences of amplicon primers including the overhang adapter sequences for compatibility with Illumina index and sequencing adapter were as follow: Forward Primer: 5'- TCG TCG GCA GCG TCA GAT GTG TAT

AAG AGA CAG CCT ACG GGN GGC WGC AG – 3’ (forward overhang:

5′-TCG TCG GCA GCG TCA GAT GTG TAT AAG AGA CAG) and Reverse

Primer: 5’- GTC TCG TGG GCT CGG AGA TGT GTA TAA GAG ACA GGA

CTA CHV GGG TAT CTA ATC C – 3’ (reverse overhang: 5′-GTC TCG

TGG GCT CGG AGA TGT GTA TAA GAG ACA G),The following were added to 96-well PCR white plates (Bio-Rad, UK): 12.5 µL of KAPA

HiFi HotStart Ready Mix (KAPA, USA), 5 µL of each forward and reverse primers (1 µM) and 2.5 uL of eDNA, (5 ng/µL) as template.

Plates were run on thermal cycler a under the following conditions:

An initial denaturation of 95 ̊C for 3 minutes and 25 cycles of 95 ̊C for

30 seconds, 55 ̊C for 30 seconds and 72 C for 30 seconds, including

⁰ 52 a final step of 72 ̊C for 5 minutes in the last cycle. Following purification using AMPure XP beads (Beckman Coulter, UK) index PCR and addition of Nextera sequence adapters was performed using a

Nextera XT Index kit (Illumina Inc., USA) according to the manufacturer’s instructions. Amplicons were sequenced by Illumina

MiSeq Next Generation sequencing (Faculty of Life Sciences Genomics

Analysis Facility, University of Manchester). Sequence data were deposited in the European Nucleotide Archive, BioProjectID:

PRJEB23187, accession numbers: ERS1983433 (site UW),

ERS1983434 (site S1), ERS1983435 (site S2), ERS1983436 (site S3),

ERS1983438 (site S2C), ERS1983439 (site N1), ERS1983440 (site

N2), ERS1983441 (site NA), ERS1983442 (site N3), ERS1983443

(site N4), ERS1983444 (site S1R).

2.3.5 Statistical analysis of environmental data

Principal component analysis (PCA) of environmental data was performed using the R vegan package v.2.4.2. In all cases the distribution of eigenvalue variation across PC demonstrated that this sample-to-variable ratio was found to be sufficient for PC consistency

(Jung and Marron, 2009). PC1 and PC2 were chosen for ordination plot as they represented 73% of the variance. The environmental data (except pH) was log-transformed and then standardized prior to

PCA. Statistical comparison of environmental data was performed using one-way ANOVA (p < 0.05) and Tukey’s multiple comparison post hoc test using GraphPad Prism. Hierarchical 53 clustering (HC) of environmental data was performed in R using

Euclidean distance and statistical comparison of nodes performed by a Similarity Profile (SIMPROF) test (Clarke et al., 2008).

2.3.6 Sequence data analysis

FASTQ files containing raw de-multiplexed sequence reads of the 16S rRNA V3-V4 region were trimmed and paired and further filtered using QIIME v.1.9.0 (Caporaso et al., 2010). Chimeric sequences were identified and removed by UCHIME v.4.2 (Edgar et al., 2011) before OTU picking. OTUs were de novo picked at 97% similarity using UCLUST (Edgar, 2010).

Taxonomic classification of the representative sequences was performed using the Greengenes v.13.8 97% OTU dataset and using the Naive Bayes machine learning classifier, which was trained using reference V3-V4 region sequence reads that were amplified as described above (Werner et al., 2011). Taxonomy classification of the representative OTU sequences of SAG wetland sites S1, S2, S2C and

S3 was performed using VSEARCH consensus taxonomy classification and the larger SILVA 16S rRNA gene reference database version 128

(Pruesse et al., 2007). Known IOB (Fe oxidizing bacteria), IRB (Fe reducing bacteria), SOB (S oxidizing bacteria) and SRB (S reducing bacteria) were selected and their relative abundances were compared.

In order to evaluate the microbial alpha diversity in each sample,

Shannon–Weiner diversity index (H), Pielou’s evenness index

(H/Hmax), and Chao1 species richness estimate values were

54 calculated using QIIME. Before the calculations, samples were resampled to an even depth of 310,833 sequences per sample, the sequence depth size of the smallest sample (N2). Rarefaction curves were also generated by QIIME.

Hierarchical clustering (HC), non-metric multidimensional scaling

(NMDS), which is an ordination method based on distances to visualize similarities of individuals according to specific variables, and multivariate statistical analysis of assigned taxa dataset obtained using the Greengenes database were carried out by R vegan package

(version 2.4.2.). Species relative abundance was square-root- transformed and a distance matrix based on Bray-Curtis dissimilarity was obtained for HC and NMDS. Complete linkage was used as agglomerative clustering method and Similarity Profile (SIMPROF) test (Clarke et al., 2008) was used for identify genuine groups between samples in HC. NMDS was performed to cluster sites and species in the same plot using HC distance matrix and 999 permutations. Samples were classified, according to their environmental characterization, in “unpolluted” and “polluted” groups. Analysis of similarity (ANOSIM) (Clarke, 1993) was used to evaluate species assemblage difference between those groups and

Similarity Percentage (SIMPER) (Clarke, 1993) was used to identify the species discriminating these groups and their contribution. The

BIOENV procedure is a dissimilarity- based method that uses correlations of abundance / count data with sample sites (Clarke and

Ainsworth, 1993). BIOENV was used to obtain the best subsets of

55 environmental variables explaining the species assemblage. Metabolic structure was predicted from the 16S library of each sample site with the PAPRICA (version 0.4.0) metabolic inference pipeline (Bowman and Ducklow, 2015). PAPRICA creates a phylogenetic tree based on

16S rRNA genes from completed genomes and infers number and abundance of metabolic traits. For PAPRICA predictions, libraries were subsampled to 310,833 reads, the size of the smallest library.

Abundance of metabolic pathways and enzymes were normalized according to the estimated number of 16S rRNA copies for each sample. Abundance of enzymes and metabolic pathways were normalised according to the estimated number of 16S rRNA copies for each sample. For generating relative log2 fold-change abundance values, normalised enzyme abundance values were converted to logarithmic scale (base 2) after adding a value of 10 to allow consideration of zero values. HC of normalized metabolic pathway abundances was performed using the Euclidean distance method and heatmaps were created using R vegan package v.2.4.2.

2.4 RESULTS

2.4.1 Environmental characterization of the SAG and NAG

Metal content, pH and conductivity of sample sites from the 2 rivers are shown in Figure 2.2. Physicochemical parameters from the unpolluted wetland (UW) are shown in Table 2.2. The SAG and NAG had markedly different environmental profiles. In N1 and N2, the water was characterised by a neutral pH and low conductivity before

56 reaching the adit . In contrast, highly acidic (pH 2.3 – 2.5) and high conductivity water was observed downstream of the adit. The conditions change slightly at the last NAG site as pH increases slightly to almost 3 and conductivity decreases to ~ 1 mS. The pattern in the

SAG is characterized by a continuous increase of pH from 2.4 to 5.7 and decrease of conductivity from 7.81 mS to 2.36 mS along the wetland (Figure 2.2d and e). Metal distribution does not present such a marked pattern as pH and conductivity. Lower dissolved metals along the SAG were recorded when compared to the NA (Figure 2.2a) below the adit. In contrast, S2 on the SAG contained the highest levels of particulate Fe, Zn, Cu, Mn and Pb (Figure 2.2b). Total Cd,

Mn, Zn and Pb in sediments were higher in the NAG. Total Fe in sediments is the only metal that shows a similar pattern to pH and conductivity, with higher values downstream of the adit and continuous decrease along the SAG (Figure 2.2c).

Hierarchical clustering (HC), based on the physicochemical parameters for each sample sites, shows that sites impacted by AMD cluster together within the lower reach of the wetland (S2) and another group at the upper reach of the SAG (S1) and the AMD- impacted section of the NAG downstream of the adit (Figure 2.3a). To correlate environmental factors to the samples assemblage observed, a PCA was performed and the plot showed that dissolved water concentration of most of the metals was correlated with sites NA, N3 and to a lesser extent with N4. In contrast, the less polluted sites N1,

N2 and S3 correlated with low pH and high concentration of Al, Mn 57 and Zn in the sediments. S2 was positioned separately from the other sample sites and correlated with particulate metal levels in the water column (Figure 2.3b).

58 a

59 b

60 c

d e

Figure 2.2. Water chemistry (pH and conductivity) and total metal concentration in water (dissolved and particulate) and sediment samples from sites on the NAG and the SAG. Data are from 4 – 7 samples taken between June 2010 and October 2014. Average values of (a) dissolved metals, (b) particulate metals, (c) total metals in sediments, (d) pH and (e) conductivity. Boxes show the 25th and 75th percentiles, the black line within the boxes shows the median values. Whisker bars show the minimum and maximum values. Boxes that do not share lowercase letters are significantly different (p < 0.05). Data were analysed by one-way ANOVA with Tukey posthoc.

2.4.2 Microbial community structure from MiSeq data

DNA for 16S sequencing and sequence library generation were extracted from sediment samples along the SAG and NAG. In addition, a 16S rRNA gene library of a sediment sample from an unpolluted wetland nearby (UW) was also included in the analysis.

Each site contained 367,239 – 1,473,986 sequences with a range of

OTUs (Operational Taxonomic Units) between 3,678 and 21,430

(Table 2.3).

Table 2.2. Selected physicochemical parameters of water (pH, conductivity and dissolved metals) and total metals in sediment samples at the unpolluted wetland site (site UW) at Cefni Reservoir, Anglesey, UK. Data are mean values (± SEM) from 3 samples. BDL, below detectable limit.

Parameter Value

pH 7.5 ± 0.1

Conductivity 0.49 ± 0.05 mS Dissolved Al 0.12 ± 0.02 mg L-1

Dissolved As BDL

Dissolved Cd BDL

Dissolved Cu 0.08 ± 0.04 mg L-1

Dissolved Fe 0.29 ± 0.04 mg L-1

Dissolved Mn 0.43 ± 0.19 mg L-1

Dissolved Pb BDL

Dissolved Zn 0.14 ± 0.03 mg L-1

Sediment Al 16.06 ± 0.92 mg g-1

Sediment As 0.01 ± 0.002 mg g-1

Sediment Cd 0.90 ± 0.06 μg g-1

Sediment Cu 0.01 ± 0.001 mg g-1

Sediment Fe 32.54 ± 1.85 mg g-1

Sediment Mn 4.11 ± 0.45 mg g-1

Sediment Pb 0.02 ± 0.001 mg g-1

Sediment Zn 0.15 ± 0.004 mg g-1

Figure 2.3. Discrimination of sites on the basis of physicochemical parameters. (a) Hierarchical clustering plot showing the assemblage of each sample site using similarities environmental parameters as a distance matrix. Sites with non-significant clustering (p <0.05) as determined by SIMPROF test are indicated with red lines. (b) PCA plot illustrating the discrimination of sites based on physicochemical parameters and their correlation with each environmental factor analysed.

The adit (site NA) had the lowest OTU number whereas the equally highly polluted site N4 had the highest number of total OTUs (Table

2.3). Likewise, both the Shannon-Weiner diversity index and Chao’s species richness index indicates that NA had low bacterial diversity, while the while the Pielou’s index score indicated that this site had low community evenness. In contrast, unpolluted sites UW, N1, N2 and S3 were the most diverse. In particular, the Shannon-Weiner diversity score indicates that site S3 diversity is substantially greater compared to sites S1 and S2 at the upper reaches of the wetland. Yet the impacted wetland sites S1 and S2 maintained relatively high diversity with scores of 5.73 and 6.56, respectively compared to the

AMD impacted river sites NA, N3 and N4 with scores of 4.29, 3.36, and 5.63, respectively. This pattern was also seen with the Chao’s species richness scores (Table 2.3).

Excluding the unassigned group, the unpolluted sites UW, S3, N1 and

N2 had between 1039 and 1060 unique assigned taxa at the genus level, while taxa number was substantially reduced at the polluted sites. The most severely impacted site (NA) only contained 463 taxonomic labels (Table 2.2). Within the impacted wetland, site S1 on the upper reaches of the SAG, showed the lowest diversity and was similar to diversity in sediments at S1 site not associated with wetland plants (S1C). Site S2 at the middle of the wetland was less diverse than at the end of the wetland (S3). No marked difference in diversity was observed in sediments from the middle part of the wetland associated with Juncus sp. (S2) and E. angustifolium (S2C). 65

Relative abundances of all taxa at phylum level show that

Proteobacteria was dominant in all of the sites, especially in NA, N3 and N4, where this phylum accounted for 49.6%, 60.9% and 64.1% of the total taxa abundance, respectively (Figure 2.4a). Examination of Proteobacteria in more detail revealed dominance by an unclassified Xanthomonadaceae, particularly at site NA (17.5%), but was also abundant at site S1. An unclassified Methylophilales was extremely abundant (46.2%) at site N3, and a Gallionella sp. was abundant (16.8%) at site N4 (Figure 2.4).

Hierarchical clustering of sites regarding species similarity was performed (Figure 2.5a). Non-impacted sites (UW, N1 and N3) and the remediated lower reach of the SAG below the wetland (S3) were clustered separately from the other sites. Clustering indicates that polluted sites N3, N4, S1, S2 and NA were grouped together, suggesting that AMD may explain the clustering pattern. SIMPROF test showed no significant separation (p<0.05) within S1 and S2, and within N3 and N4.

ANOSIM shows that there is a significant separation between the two groups (R = 0.61, p = 0.016). SIMPER analysis indicates that the unclassified Methylophilales taxa, which was dominant at sites NA, N3 and N4 (Figure 2.5b), was the highest ranked contributor to the assemblage of the sites. Top 10 taxa with mayor contribution to the microbial structure are detailed in Table 2.4

Table 2.3. Bacterial diversity estimates and summary result of 16S rDNA sequence analysis for each sample site. (a) Assigned taxa number indicates the number of assigned taxa for each site at the genus level (excluding unassigned taxa). (b) Total OTUs indicates the sum of all OTUs obtained at each site. (c) Total sequence indicates the number of end-paired sequences obtained after trimming and chimeras removal.

Shannon – Pielou´s Chao’s species Assigned taxa Total OTUs Total sequences Sites Weiner diversity evenness index richness number (a) (b) (c) index UW 7.35 0.58 22,592 1053 16,968 79,019 S1 5.73 0.45 9,865 721 5,863 367,239 S1R 5.71 0.45 10,395 732 6,670 434,428 S2 6.56 0.52 19,624 851 11,881 434,428 S2C 7.22 0.57 17,743 888 12,217 609,736 S3 7.92 0.63 24,969 1039 18,315 609,955 N1 8.01 0.63 25,675 1152 17,386 435,309 N2 7.15 0.57 16,862 1060 11,256 310,833 NA 4.29 0.34 5,283 463 3,678 542,593 N3 3.36 0.27 16,491 1102 13,582 1,106,448 N4 5.63 0.45 23,115 1009 21,438 1,473,986

Figure 2.4. Relative abundance of bacterial taxa following OTU taxonomic assignment. All taxa at phylum level (a) and Proteobacteria at species level (b) for each sample site. Selected taxa of high abundance in multiple samples are labelled, with ranges of relative taxa abundance given in parentheses.

NMDS ordination based on the same criteria showed the same assemblage for polluted and unpolluted sites. Taxonomic assignments were also included in the ordination and show a higher number of taxa related with the unpolluted sites. Top 5 taxa obtained on the

SIMPER analysis are also indicated in Figure 2.5b. BIO-ENV analysis showed that pH and conductivity were of the combination of environmental variables that better explains the microbial community composition (Table 2.5).

2.4.3 Relative abundance of known bacteria related to S and

Fe metabolism

For this analysis, samples S1, S2C (E. angustifolium stand) and S2 were used. Site S1 shows low abundance of bacteria related S and Fe metabolism (SOB in particular, Figure 2.6a and b). S2 shows lower relative abundance than S2C in all the bacteria groups except for IOB

(Figure 3.7a and b). SOB and IOB were dominated by unclassified

Bradyrhixobium and unclassified Ferrovum in S2, respectively (Figure

2.6b and Figure 2.7b). S2C show higher abundance of bacteria taxa for all groups except IOB when compared with all the other sites. The end of the wetland (S2) shows similar pattern than the wetland site close to the source (S1) except for IRB. S3 showed higher abundance of IRB taxa than S2 and S1 but lower than S2C.

Figure 2.5. Discrimination of sites on the basis of bacterial community structure. (a) Hierarchical clustering plot of samples based on species similarity. Red lines indicate no statistical evidence (p<0.05) for sub- structures within samples according to SIMPROF test. (b) Two dimensional NMDS plot of sites based on species similarity showing a separation between unpolluted sites (UW, S3, N1, N2), polluted sites (S1, S2, N3, N4) and the highly polluted mine adit site (NA) on the basis of NMDS1. Assigned taxa assemblage is also shown (grey circles) and the top 5 taxa with the most driving contribution to sites assemblage based on SIMPER analysis are indicated and listed in red.

Table 2.4. Top 10 taxa from SIMPER analysis that best explain the assemblage of the bacterial community structure obtained from relative abundance of taxonomic assignments. All taxa belong to the Kingdom Bacteria. Percentage of contribution is shown.

Rank Taxa %

p__Proteobacteria;c__Betaproteobacteria;o__Methylophilales;f__ 1 1.63 ;g__;s__

p__Proteobacteria;c__Gammaproteobacteria;o__Xanthomonadal 2 1.4 es;f__Xanthomonadaceae;g__;s__

p__Acidobacteria;c__Acidobacteriia;o__Acidobacteriales;f__Acido 3 1 bacteriaceae;g__;s__

4 p__Proteobacteria;c__Betaproteobacteria;o__;f__;g__;s__ 0.88

p__Actinobacteria;c__Acidimicrobiia;o__Acidimicrobiales;f__;g__ 5 0.85 ;s__

;p__Proteobacteria;c__Alphaproteobacteria; 6 0.76 o__Rhodospirillales;f__Acetobacteraceae;g__;s__

p__Proteobacteria;c__Betaproteobacteria;o__Gallionellales;f__G 7 0.71 allionellaceae;g__Gallionella;s__

8 p__WPS-2;c__;o__;f__;g__;s__ 0.71

p__Bacteroidetes;c__Cytophagia;o__Cytophagales;f__Cytophaga 9 0.67 ceae;g__;s__

p__Verrucomicrobia;c__[Pedosphaerae];o__[Pedosphaerales];f_ 10 0.65 _auto67_4W;g__;s__

Table 2.5. Sets of environmental variables from BIO-ENV analysis. Top 10 set of variables which best explain the assemblage of the bacterial community based on relative abundances of taxonomic assignments. Dis: dissolved values. Par: particulate values. Sed: sediment values. Rho: Spearman’s rank correlation coefficient.

Variables Rho

Conductivity, Dis(Fe), Dis(Al), Dis(Cu), Dis(Mn), Dis(As), 0.854 Par(Fe), Sed(As)

pH, Dis(Fe), Dis(Al), Dis(Mn), Dis(As), Dis(Pb), Par(Fe), 0.854 Sed(Fe), Sed(As)

pH, Conductivity, Dis(Fe), Dis(Zn), Dis(Cu), Dis(Mn), 0.850 Dis(As), Dis(Pb), Par(Fe), Sed(Fe), Sed(As)

pH, Conductivity, Dis(Zn), Dis(Al), Dis(Mn), Dis(As), 0.849 Dis(Pb), Par(Fe), Sed(Fe), Sed(As)

Conductivity, Dis(Zn), Dis(Cu), Dis(As), Dis(Pb), Par(Fe), 0.845 Sed(As)

pH, Dis(Fe), Dis(Mn), Dis(As), Dis(Pb), Par(Fe) 0.841

pH, Dis(Cu), Dis(As), Dis(Pb), Par(Fe) 0.836

pH, Conductivity, Dis(Fe), Dis(Zn), Dis(Al), Dis(Cu), 0.835 Dis(Mn), Dis(As), Dis(Pb), Par(Fe), Sed(Fe), Sed(As)

pH, Dis(As), Dis(Pb) 0.812

Dis(Fe), Dis(Al), Dis(As), Dis(Pb) 0.811

Figure 2.6. Relative abundance of of known S reducing / oxidizing bacteria in surface sediment samples along the Southern Afon Goch wetland. Bars showed relative abundance (OUT count divided total) of SRB (a) and SOB (b) taxa for each site following OTUs taxonomic assignment using SILVA database.

Figure 2.7. Relative abundance of known Fe reducing / oxidizing bacteria in surface sediment samples along the Southern Afon Goch wetland. Bars showed relative abundance (OUT count divided total) of IRB (a) and IOB (b) taxa for each site following OTUs taxonomic assignment using SILVA database.

2.4.4 Prediction of microbial taxonomic-derived metabolism

Using phylogeny for inferring metabolic prediction by the PAPRICA method (Bowman and Ducklow, 2015) a score of predicted abundance of metabolic pathways and enzymes was obtained for each sample sites. HC based on similarity of metabolic pathways showed similar clustering than the HC obtained by taxa similarity

(Figure 2.8a). NA, N3 and N4 clustered different from the other sites

(p<0.001) but did not form a distinct cluster (p>0.05). A heatmap of all the metabolic pathways predicted showed a decreased overall abundance in NA, N3 and N4 (Figure 2.8b). A second heatmap of specific metabolic pathways related to bacteria metabolism observed in AMD environments, showed the same decreased abundance at NA,

N3 and N4 (Figure 2.8c).

A score of predicted abundance of enzyme classes was also obtained by the PAPRICA methods. The overall changes in abundance between unpolluted and polluted sites on the SAG (site S1 relative to site UW) and NAG (site NA relative to site N1) was performed (Figure 2.9). A relatively strong positive correlation (R2 >0.4) was observed in the predicted enzyme abundance change between NAG and SAG. This suggests that AMD entering S1 and NA produce a similar pattern in the predicted metabolic profile.

A second analysis comparing changes between the site of highest source of AMD pollution (site S1 or NA) and the furthest downstream site (site S3 or N4) on the two rivers was performed. This provides an

75 evaluation of metabolic potential change moving through the wetland

(x-axis) or moving down the NAG (y-axis) (Figure 2.9). The analysis showed no significant correlations (all R2 <0.1). This suggests an important differences in the metabolic potential profiles on the sites depending if the AMD flow pass through the wetland or the river.

Enzyme classes EC1, EC2, EC3, EC4 and EC6 showed changes specific to the northern river and no significant change through the wetland.

In contrast, a few enzymes, especially of the EC1 oxidoreductase and

EC2 transferase classes, showed increased abundance through the wetland but are not predicted to increase in abundance between site

NA and N4 (Figure 2.10).

2.5 DISCUSSION

2.5.1 Microbial structure related to AMD pollution levels

The phylum Proteobacteria was present in all the sediment samples and was the most abundant phylum recorded in other studies of streams polluted with AMD (Bruneel et al., 2006, Chen Hong, 2015,

Sun et al., 2015b).

76 a

77 c

Figure 2.8. Metabolic prediction analysis and discrimination of sites on the basis of metabolic potential. (a) Hierarchical clustering of sample sites based on metabolic pathway similarity. Sites with non-significant clustering (p <0.05) as determined by SIMPROF test are indicated with red lines. Heat maps showing the predicted abundance of (b) all prokaryotic metabolic pathways (b) and selected metabolic pathways related to AMD element cycling.

Figure 2.9. Predicted changes in enzyme abundance in response to AMD pollution. Scatter plots of Enzyme Commission (EC) enzyme reaction classes showing the relationship between wetland (southern Afon Goch) and non-wetland (northern Afon Goch) log2 fold-change values for polluted (site S1 or NA) versus unpolluted (site UW or N1) sites. Each dot corresponds to log2 fold-change in abundance of a predicted enzyme reaction. The linear regression best fit line is shown.

79 a

Figure 2.10. Predicted changes in enzyme abundance in response to the presence of a wetland. (a) Scatter plots of EC enzyme reaction classes showing the relationships between wetland and non-wetland log2 fold-change values for pollution source (site S1 or NA) versus downstream sites (site S3 or N4). Each dot corresponds to log2 fold- change in abundance of a predicted enzyme reaction. The linear regression fit line is shown. (a) Scatter plot for all enzyme classes. The red shaded quadrant indicates enzymes that are significantly increased in abundance specifically at site S3 but not at site N4, relative to the pollution source sites. The enzymes of each class that are increased in abundance at site S3 are shown (right plot).

The phylum Proteobacteria includes many species from acidic environments with metabolism related to metal immobilization and includes S - and Fe -oxidizing autotrophs such as the genus

Acidithiobacillus (Kelly and Wood, 2000, Baker and Banfield, 2003), sulphate reducing bacteria such as Desulfovibrio sp. (Lentini et al.,

2012, Wu et al., 2013) and bacteria resistant to metals, including

Pseudomonas sp. (Silver and Phun, 1996, Tang et al., 2011).

When sites were analysed at the genus level (Figure 2.4b), it is evident that the polluted sites on the NAG showed dominance of specific taxa that is not seen in the other samples. A bacterium of the

Methylophilales order was the most abundant taxa in the NAG sites affected by AMD (NA, N3 and N4). Studies have reported that some species of Methylophilales have a metabolism with high adaptation capacity and can use many substrates in the respiration process

(Lanclos et al., 2016). SIMPER analysis indicated that the unclassified

Methylophilales taxa, which was dominant at sites NA, N3 and N4 was the highest ranked contributor to the assemblage of the sites. A

Methylophilales taxa associated with Cu tolerance was previously reported as one of the most abundant taxa within a microbial mat involved in natural remediation of trace metal-contaminated mine water (Drewniak et al., 2016), while the same taxa was also abundant in streams exposed to alkaline mine drainage (Bier et al.,

2014).

The order Gallionellales was the dominant taxon in N4. Gallionella sp. is a bacterium with well-known Fe oxidizing activity in microaerophilic 81 environments (Emerson and Moyer, 1997, Mitsunobu et al., 2012) capable of generating biogenic Fe oxides in the sediments (Ferris,

2005, Kikuchi et al., 2014). Interestingly, this bacterium has been reported only in water with a circumneutral pH (6.5 – 7.5) (Gault et al., 2012). No marked dominance of specific taxa in the AMD impacted wetland sites is observed. Only the most impacted wetland site (S1) shows dominance of a bacterium from the order

Xanthomonadales (18.0%). The order Xanthomonadales has been previously observed in Fe(III)-rich sediments (Senko et al., 2008) and iron-oxidizing acidophile communities (Jones et al., 2015).

Dominant taxa found in this study are similar to metagenomics analysis of bacterial communities from an AMD low-temperature (6-

10 ̊C) stream draining the Kristineberg mine in Sweden, where metagenomic reads showed that Acidithiobacillus sp. (19.2% -

25.5%), Acidobacteria-like sp. (0.9% - 5.4%), and Gallionellaceae - like species (0.2%-3.9%) were the most abundant (Liljeqvist et al.,

2015). A previous study performed a metagenomic and transcriptomic analysis of bacterial communities in four AMD streams and concluded that Fe oxidizing bacteria from the genera

Acidithiobacillus, Leptospirillum and Acidiphilium were the most abundant taxa and have the highest transcriptional activity (Chen et al., 2015a).

Dominance of specific taxa decreased and microbial diversity increased in S2 and S3 compared to S1 on the SAG. The pattern in the relative abundance of bacteria in S3 was similar to that observed

82 in the unpolluted upper reaches of the NAG at N1 and N2, which indicates a recovery in microbial diversity as the SAG flows through the wetland. Metagenomic comparison of bacterial community from unpolluted and polluted sites has been previously studied with different results. Metagenomic analysis by Illumina sequencing of microbial structure along a river with an AMD pollution gradient showed changes in bacterial dominance pattern (Sun et al., 2015b) similar to this study. However, sediments from a historically mining- impacted stream showed different microbial composition but similar diversity when compared to a relative pristine stream in Luanda,

Angola (Reis et al., 2013).

Several studies of AMD environments across the world are allowing the identification of the environmental factors than explain the structure and diversity of species in microbial communities AMD systems (Méndez-García et al., 2015, Huang et al., 2016). Analysis of samples from mine tailings with pH values from 1.9 to 4.1 (Kuang et al., 2013) and 2.2 to 7.3 (Liu et al., 2014) suggest that pH is the main influence on species diversity. Moreover, in studies of terrestrial soil, the pH is also considered the major parameter affecting the structure of the bacterial community (Fierer and Jackson, 2006,

Lauber et al., 2009). MDS analysis of species similarity in this study showed that sites with higher pH (N1, N2 and S3) clustered together.

PCA ordination of environmental variables (Figure 2.3b) showed the correlation between higher pH and sites N1, N2 and S3 indicating that pH is an important factor determining the microbial structure. Similar

83 conclusions were drawn from a metagenomic analysis performed in

AMD environments from several mine sites across Southeast China with different physicochemical parameters as multivariable analysis concluded that pH was the main predictor of bacterial abundance and diversity (Kuang et al., 2013). Pyrosequencing analysis of a microbial community from a drainage in the Pb-Zn Carnoules mine in southern

France, showed a correlation of dominant genera such as Ferrovum and Gallionella with high concentrations of As, Fe and sulphate in the mine drainage (Volant et al., 2014). This study has shown that Fe and trace metals are also correlated with sample sites with a similar dominant taxa pattern. BIO-ENV analysis found that pH alone was not the key environmental variable for explaining the variation in community structure between the sites. Thus conductivity and metal concentration characteristics are strongly correlated with species assemblage, indicating that factors other than pH may also be an important on the bacterial assemblage of AMD-impacted sites.

Bacteria related to Fe and S metabolism was further analysed along the SAG due to the rapid increase of particulate metals in the middle of the wetland. It is known that oxidation/reduction of Fe and S can mediate the precipitation of other metals in AMD remediation systems

(Machemer and Wildeman, 1992, Webster et al., 1998) For this analysis, a larger database than Greengenes with higher number of taxa labels was used in order to identify key taxa. SRB

(Desulfatirhabdium and Desulfolobus) detected in the sediment of S1 and S1C (Figure 2.6a) belong to the family Desulfobulbaceae, which

84 thrive in environments rich in hydrogen sulphide (Pfeffer et al., 2012,

Larsen et al., 2015). Low abundance of SOB is observed in S1.

However, S2 (mainly in E. angustifolium stand) and lower wetland shows evidence dominance by Bradyrhizobium and Sulfuricurvum.

Previous studies have demonstrated the importance of Sulfuricurvum in the transformation of reduced S compounds such as elemental S and thiosulphate (Han et al., 2012). These bacteria maintain low

2− levels of total S by oxidizing intermediate compounds into SO4 that can be easily metabolized by other bacteria and plants. The presence of these bacteria suggest possible changes in S speciation that can mediate metals co-precipitation with reduced S compounds.

Abundance of IRB and IOB increase in surface sediments downstream from the AMD source. Increase in IRB is mainly due to the dominant

Geotrix and Geobacter (Figure 2.6b). Previous studies have revealed that the anaerobic chemoorganotrophic metabolism of Geotrix fermentans allow the use of Fe3+ as electron receptors (Nevin and

Lovley, 2002). Geobacter possess a more complex metabolism as it is capable of both oxidizing and reducing Fe (Childers et al., 2002).

The dominant IOB were the obligate Fe2+ oxidizing Gallionella and

Ferrovum even though Gallionella is commonly found in circumneutral environments where the pH is around 6 (Emerson and Weiss, 2004).

Surprisingly, while abundance of Gallionella are similar between E. angustifolium and Juncus sp. stand from the middle wetland, abundance of Ferrovum is much higher in the Juncus sp. stand and

85 almost absent in E. angustifolium sediments. This is evidence that wetland plant species can facilitate the proliferation specific AMD related taxa despite the short distance between different plant stand.

2.5.2 Prediction of metabolic potential

The uptake and accumulation of dissolved metals by wetland plants is a minor component of the total metals retained in the Afon Goch wetland (Dean et al., 2013) and suggests that the primary bioremediation action is due to microbial-derived enzymatic activities. The changes in the microbial structure along the wetland and the river and the changes in the metal fraction (increased in particulate metals) in the middle reaches of the wetlands, suggest differences in metabolic traits. To quantify the microbial-derived

‘metabolic potential’ (Kuang et al., 2016) of each site , we used the

PAPRICA metabolic inference prediction method by phylogenetic placement as an alternative to methodologies such as Tax4Fun and

PICRUST, which have some limitations in environmental microbiome analysis (Bowman and Ducklow, 2015, Koo et al., 2017). PICRUST predictions have observed enrichment of microbial functions related to energy production such as methanogenesis and nitrogen fixation in

AMD irrigated paddy soils,(Wang et al., 2018). In contrast, a decrease in predicted abundance of S and C cycling has been observed in mine tailings using PAPRICA (Li et al., 2017).

PAPRICA showed decreased predicted abundance of overall metabolic pathways and important biochemical reactions mediated by bacteria and related to AMD adaptation such as S, As, C, H and N cycling 86

(Méndez-García et al., 2015) (Figure 2.8a and b). The sites with the most decrease abundance where the same with lower taxa diversity.

This is the first study analysing AMD impacted environments by

PAPRICA.

Using > 2 and < -2 log2 fold-difference as the threshold indicating a significant change in enzyme abundance, the plots predict that there are a number of significant increases and decreases in enzyme abundance through the wetland (Figure 2.9a). No change was observed along the northern river and, likewise, predicted changes in enzyme abundance along the northern river that are not seen through the SAG wetland. Interestingly, EC1 oxidoreductases increased in abundance through the wetland but are not predicted to increase abundance between site NA and N4 (Figure 2.9c). Reduction of organic matter in waterlogged systems is a main bacterial mechanisms for carbon cycling (Gabriel et al., 2017) . Decomposition of organic matter is also an important mechanism for the generation of biogenic Fe oxides (Fortin and Langley, 2005). Iron oxidation is a main process for precipitation of trace metals in natural remediation system (Johnson and Hallberg, 2005, Sheoran and Sheoran, 2006).

2.6 CONCLUDING REMARKS

In this chapter, a taxonomic and taxonomic-derived metabolic characterization of the microbial community from a long-term AMD river that flows through a polluted-adapted wetland was compared against a less adapted non-vegetated river. Sediments from the non- 87 vegetated river were characterised by a reduced microbial diversity and dominance of specific taxa, hence a reduced metabolic activity was predicted. This evidences the stress produced by acidity and high metal concentrations. In contrast, a more stable microbial diversity and predicted metabolic functions were detected in sediments along the wetland, which suggest the key role of bacteria in the observed reduction of acidity and removal of dissolved metals at the end of the wetland. Further analysis in this chapter showed that the distribution of Fe and S oxidizing/reducing bacteria changes depending on the wetland site and plant species. To obtain new insights about the remediation mechanisms, evaluation of metal retention and mobilization mediated by Fe and S species should be approached. In order to further understand the role of bacteria in the remediation process, a more detailed analysis of Fe and S speciation and possible correlations with the activity of key bacterial genes along the SAG wetland is needed.

Chapter 3: ANALYSIS OF BACTERIA ACTIVITY AND METAL DISTRIBUTION ALONG THE SOUTHERN AFON GOCH WETLAND

Aguinaga, O. E., Wakelin, J., White, K. N., Dean, A. P., & Pittman, J. K.

The author analysed environmental parameters and bacterial activity, interpretation and writing up the manuscript. The author and James

Wakelin performed sample collection. Jon Pittman, Keith White and

Andrew Dean provided full guidance and manuscript review.

3.1. ABSTRACT

A natural wetland surrounding the Southern Afon Goch receives acid mine drainage from an abandoned metal mine at Parys Mountain

(Anglesey, UK). The wetland has shown to markedly reduce acidity and metal contamination of nearby coastal waters. Evaluation of Fe,

S and trace metal partitioning in the water column and analysis of metal distribution with sediment depth were performed to understand the remediation mechanisms. In order to understand the role of bacteria in the metal attenuation process, transformations of Fe and

S mediated by bacteria was analysed. Increased expression of bacterial soxB (S oxidation) gene and 16S rRNA gene from obligate

Fe oxidizing F. myxofaciens were observed in the same location within the wetland, were significant reduction of sulphate and Fe+2 levels was detected. Generation of particulate Fe and reduction of dissolved trace metals also occurs. High resolution X-ray fluorescence scans of sediment cores revealed similar patterns of Zn, Cu and S distribution with sediment depth. This study reveals that the distribution and speciation of Fe and S that mediates the immobilization of trace metals onto the wetland sediments is markedly influenced by bacterial activity.

3.2 INTRODUCTION

In Chapter 2, it was shown that the microbial diversity was reduced in the upper reaches of the SAG close to the AMD source but that this loss of diversity was recovered at the end of the wetland. As a result, metabolic activities were predicted to increase along the wetland.

Furthermore, environmental analysis in Chapter 2 suggests that the wetland remediation process is mediated in part by precipitation of metals in the middle of the wetland, as seen by the variations of particulate metal levels and a subsequent immobilization of metals at the end of the wetland. Due to the capacity of Fe and S oxidizing/reducing bacteria to modify environmental characteristics such as the distribution and speciation of other metals in the environment (Machel, 1989, Blgham et al., 1990, Fredrickson et al.,

1998), an evaluation of Fe and S oxidizing/reducing bacteria in

Chapter 2 was performed and showed differences in the taxonomic structure depending on the wetland site and plant species.

This Chapter further investigates the role of microorganisms by measuring the activity of specific bacteria and genes related to Fe and S metabolism and the nutrients available in surface and bottom sediment layers. Moreover, the metal chemistry in the wetland was further analysed by complementing the particulate and dissolved values previously obtained with a more detailed evaluation of Fe and

S speciation along the wetland. Mechanisms of metal deposition and retention onto the sediments were also further investigated with high resolution measurement of metal distribution along sediment cores. 91

Enzymatic mechanisms of Fe oxidation and reduction remain unclear.

The metabolic pathways and enzymes that have been identified, varies between species and are also used for the metabolism of other compounds. For example, the protein rusticyanin in acidophilic autotrophic bacteria (Cox and Boxer, 1978) is not present in neutrophilic or phototropic Fe oxidizing bacteria. Many types of cytochromes are involved in Fe oxidation but they are also in charge of oxidizing other molecules, such as oxygen, as part of electron transport chains in membranes (Weber et al., 2006, Bird et al.,

2011). The strategy used in this study was to quantify the abundance of bacteria which are obligate Fe oxidizing or reducing bacteria and that can survive under different environmental conditions.

In contrast to Fe metabolism, there are specific genes related to S oxidation and reduction such as the previously characterized soxB gene, which encodes a protein essential for thiosulphate bacterial oxidation and subsequent S oxidation in various phylogenetic groups of bacteria (Epel et al., 2005); and the dsrA gene, which is a member of the family of sulphite reductases in charge of bacteria dissimilatory sulphate reduction (Muyzer and Stams, 2008). These functional genes were also quantified to obtain insights into the S metabolism within bacteria.

High resolution of S and metal concentrations was assessed by sediment cores scanning by X-ray fluorescence (ITRAX). This method was used for the first time in mine pollution studies to measure stable

92 earth elements in estuarine sediments receiving drainage from an operational mine (Rodríguez-Germade et al., 2015). ITRAX core scanning was also used for evaluation of trace metal distribution in polluted harbour sediments impacted by industrial activities

(Rodríguez-Germade et al., 2014, Croudace et al., 2015). In all cases, ITRAX technology has proven to be an efficient and fast approach for monitoring and detecting sediment pollution. Results of bacteria activity, total metals and nutrients in surface and bottom layers from the sediments were correlated with the continuous values obtained by ITRAX within the first 50 cm of the wetland sediments.

Marked shifts in metal distribution and bacteria populations can rapidly occur within short distance in AMD impacted environments

(Valkanas and Trun, 2018). The SAG wetland is approximately 2 km in length and perhaps surprisingly, most of the dissolved metals are removed from the water column and pH approaches neutral values in approximately the first km of the system (between the upper and middle reaches of the wetland; see Figure 3.1).

The combination of biological and geochemical measurements described in this Chapter were designed to detect the rapid changes in metal distribution and bacteria composition in the SAG wetland and to elucidate the possible biological and geochemical mechanisms behind the efficiency of the system to remediate AMD within a short distance. Therefore, the following aim and objectives were formulated:

Aim:

To identify possible mechanisms of AMD attenuation in the water column and metal immobilization in different sediment layers along the SAG wetland.

Objectives:

• Measurement of Fe and S and trace metals partitioning in the

water column along the wetland.

• Measurement of Fe and S and trace metals partitioning in the

top and bottom layers of sediments along the wetland.

• Measurement of Fe and S bacteria-mediated transformation in

the top and bottom layers of sediments along the wetland.

• Analysis of metal distribution, C and N concentrations in

different core depths along the wetland.

3.3 MATERIALS AND METHODS

3.3.1 Field site locations and sampling

Water and sediment samples were taken from the SAG wetland on

September 2017. Three locations along the wetland were selected according to physicochemical parameters in the water column along the wetland. Site W1 was located at ~500 m from the source, where the attenuation process starts. Site W2 (same as S2 in chapter 2) was located where decrease in acidity and dissolved metals in the water column was observed (Figure 3.1). Site W3 is located in the

94 lower reaches of the wetland (Figure 3.2) where AMD pollution is significant reduced.

Figure 3.1. Localization of sampling sites along the Southern Afon Goch wetland. (a) Values of dissolved Fe, Zn and Cu, pH and conductivity (Modified from Dean et al., 2013) and localization of sample sites. (b) Map of the SAG wetland and sampling sites selected for water and sediment cores. Map data obtained from Google, DigitalGobe. Imagery Date: 3/24/2017.

Surface water samples (~2 cm depth) were taken and processed as previously described (Chapter 2). Sediment core samples (~ 50 cm depth) were obtained using a Russian corer (Van Walt, UK). A depth of 50 cm was sufficient to obtain the oxic (including the rhizosphere) and the anoxic zones of the sediment. Sediment subsamples from

95 different depths and were taken in 50 mL polypropylene tubes for metal analysis and DNA and RNA extraction.

3.3.2 In situ water analysis and total aqueous metal

In situ water measurements of pH were performed using a portable pH meter (Hanna Instruments, UK). In situ water analysis of Fe2+, sulphate and sulphide concentrations was carried out by using the

1,10-Phenanthroline, Methylene Blue and SulfaVer 4 methods, respectively (Rice et al., 2012), employing a DR900 Multiparameter

Portable Colorimeter (Hach, USA). Dissolved and particulate metals in the water column were analysed as previously described (Chapter 2).

3.3.3 Metal analysis in sediments

Non-destructive, high resolution X-ray fluorescence (XRF) spectrometry was performed using an ITRAX core scanner (School of

Environment, Education and Development, University of Manchester) to measure element concentrations along nine sediment core samples

(3 replicates for sites W1, W2 and W3). Prior to analysis, each core was prepared by ensuring that the surfaces were completely flat using a roller. The X-rays used to irradiate the cores were generated by a 3kW Mo-tube. A step size of 1 mm and a count time at each step of 20 seconds were selected.

Data for Fe, S, Zn, Cu, Mn, Al, As, Pb and Si was obtained from the scans and expressed as total counts per second (CPS). To transform

96 these values into concentrations (mg/g), the total content of each metal from selected 10 mm fractions was obtained using ICP-

AES. Subsequently, CPS data was correlated with ICP-AES results in calibration curves and a linear regression was calculated for each element (Supplementary Figure S3.1). Sediment samples obtained from the cores at the top (0-20cm) and bottom (20-50cm) layers were also analysed by ICP–AES. ICP-AES measurements for ITRAX data calibration and sediment samples were performed as previously described (Chapter 2).

3.3.4 Measurement of C and N along core depths

Water-extractable C and N were analysed after shacking 5 g of sediment from each core at 10 cm intervals in 35 mL of Milli-Q water for 10 min using an orbital shaker. Extractants were filtered using

Whatman no.1 filters (Camlab, UK). Total and inorganic C concentrations were obtained using a Shimadzu 5000 A TOC analyser

(Shimadzuk Limited, UK). Dissolved organic C (DOC) was determined by subtracting inorganic C values from total C in the extracts. Total

N, dissolved nitrate and ammonium were determined by an

AutoAnalyser 3 HR (Seal Analytical, UK). Dissolved organic N (DON) was determined by subtracting nitrate and ammonium values from total N in the extracts.

For analysis of total C and N content, sediment samples, again from each core at 10 cm intervals, were dried at 80°C for 24 hours, 97 disaggregated using a Mixer Mill MM 400 (Retsch, Germany), weighted and analysed by combustion using an Elemental Vario EL elemental analyser (Elementar Analysensysteme, Germany).

3.3.5 DNA and RNA extraction

DNA and RNA were extracted for construction of 16S rRNA libraries and quantification of RNA transcripts, respectively. DNA was extracted from surface sediment (0.2 g) from different plant stands at W2, using a Powersoil DNA isolation kit (Qiasen, USA). To stabilize microbial RNA in sediment samples, LifeGuard soil preservation solution (Qiagen, USA) was added following the manufacturer’s instructions. RNA was extracted from the five replicate (4 g) sediment subsamples from the top and bottom layers of the cores. A

RNeasy PowerSoil Total RNA kit (Qiagen, USA) was used to extract the RNA. To remove genomic DNA, RNA samples were treated with

DNAse I (New England Biolabs, UK) following the manufacturer’s instruction. Nucleotides were quantified using a Nano-drop 3300

(Thermo-Scientific, USA).

3.3.6 Construction of 16S rRNA libraries

16S rRNA sequences from extracted DNA were amplified using the My

Taq™ Red Mix (Bioline, USA) and primers Eubac27F and 1492R

(DeLong, 1992). PCR cycling conditions were as follows: 95 °C for 1 min, 35 cycles of 95 °C for 15 s, 50 °C for 15 s and 72 °C for 45 s. A final extension at 72 °C for 5 min was added. PCR products were ligated into pGEM® T-Easy plasmid vector (Promega, USA) using T4

DNA Ligase and 2X Rapid Ligation Buffer. The reaction was incubated at 4°C overnight. XL 1-Blue supercompetent cells (Agilent

Technologies, USA) were transformed with the plasmid. Cells were plated on LB - ampicillin/IPTG/X-Gal agar plates and incubated at

37°C overnight. Colonies were selected from each plate, inoculated in

LB – ampicillin broth and incubated at 37°C overnight. Plasmids from each culture were purified using the QIAPrep Spin Miniprep Kit

(Qiagen, Germany) and Sanger sequenced by GATC-Biotech.

3.3.7 Quantification of RNA transcripts

Reverse transcription of extracted RNA samples was performed using

SuperScript II Reverse Transcriptase (Thermo-Scientific, USA) following the manufacturer’s instructions. Random hexamers

(Thermo-Scientific, USA) were used as primers for synthesis of cDNA.

Gene expression analysis by quantitative polymerase chain reaction

(qPCR) using cDNA samples was performed for six bacterial genes related to Fe oxidizing bacteria and bacterial S metabolism.

Description of target genes and primers used for quantification of

RNA transcripts are displayed in Table 3.1. Serial 1 in 10 dilutions of cDNA from known RNA concentration samples were used as templates for generating a standard curve for each set of primers.

Concentration of RNA transcripts were calculated by absolute quantification.

The following was added to 96-well PCR plates (Applied Biosystems,

UK): 10 µL of SensiFAST SYBR Hi-Rox mix (2x), 0.8 µL of each forward and reverse primers (10µM), 0.5 µL of cDNA, volume was made up to 20 µL with nuclease-free water. Plates were run on a

Step One Plus Real Time PCR system (Applied Biosystem, UK) with

SYBR Green Rox detection program. Analysis was performed on five independent biological replicates for each sample site. Reactions using RNA as templates were included as a control of possible genomic DNA contamination. Negative controls using no template were also included.

3.3.8 Bioinformatic analysis

A phylogenetic tree based on the 16S rRNA sequences from the clone libraries was constructed using the neighbour-joining method (Saitou and Nei, 1987) with 500 bootstrap replicates using MEGAN 7 (Kumar et al., 2016). For this construction, additional 16S rDNA sequences were included:

100

• Ferrovum myxofaciens (EF133508) and 2 related clones

(EU360493 and EU360502) used in the design of the “Ferrovum

and relatives” set of primers (Heinzel et al., 2009).

• Gallionella ferruginea (L07897) used in the design of the

“Gallionella and relatives” set of primers (Heinzel et al., 2009).

• Thiobacillus denitrificans (NR025358) and Thiobacillus thioparus

(GU967679) which genomes contain the sequence targeted by

the set of primers specific for the soxB gene (Tourna et al.,

2014)

• Known bacteria used in a previous phylogenetic construction of

clones from the Afon Goch wetland (Dean et al., 2013).

3.3.9 Statistical analysis

To evaluate significant differences among mean values of metal concentrations and among in situ water parameters and to evaluate significant differences in the expression of the different genes evaluated, one-way analysis of variance (ANOVA) and post-hoc test using Turkey’s method was carried out using GraphPad Prism 7.

Differences between means were considered significant if p<0.05.

101

Table 3.1. Description and sequences of primers used for quantification of RNA transcripts

Target Amplicon description Primer sequence (5’ – 3’) Reference

Fragment of the 16S rRNA gene 341F: CCTACGGGAGGCAGCAG (Muyzer et al., Total bacteria from all bacteria 518R:ATTACCGCGGCTGCTGG 1993)

Fragment of the 16S rRNA gene Gallionella 122f: ATATCGGAACATATCCGGAAGT (Heinzel et al., specific to bacteria related to and relatives 384r: GGTATGGCTGGATCAGGC 2009) Gallionella ferruginea

Fragment of the 16S rRNA gene Ferrovum and 643f: ACTGGCAAGCTAGAGTCTGT (Heinzel et al., specific to bacteria related to relatives 847r: TCGCGTTAGCTTCGTTACTGA 2009) Ferrovum myxofaciens

Geobacter Fragment of the 16S rRNA gene Geo561F: GCGTGTAGGCGGTTTBTTAA (Cummings et al., spp. specific to Geobacter spp. Geo858R: TCAATACCCGCAACACCTAG 2003)

Dissimilatory sulphite Fragment of the dsr gene from S DSR-F: ACSCACTGGAAGCACGCCGG (Kondo et al., reductase reducing bacteria (221 bp) DSR-R: GTGGMRCCGTGCAKRTTGG 2004) (dsrA) Sulphur SoxF: ATCGGYCAGGCYTTYCCSTA oxidizing Fragment of the SoxB gene from (Tourna et al., SoxR: MAVGTGCCGTTGAARTTGC enzyme known S oxidizing bacteria. 2014)

(soxB)

102

3.4 RESULTS

3.4.1 Phylogenetic analysis of 16S rRNA libraries from W2 surface sediments

16S rRNA libraries were created from sediments surrounding different plant species (Juncus sp., Phragmites australis and Eriophorum angustifolium) as well as river sediment without plants to evaluate differences in bacteria diversity within the middle W2 site. A total of

21 different clones were obtained from all samples. BLAST analysis showed that most of the sequences aligned with uncultured bacteria from acidic metal polluted sites such as acid mine drainage, acid lakes and wastewater treatment plants (Table 3.2).

The construction of a phylogenetic tree based on these clones and other known bacteria (Figure 3.3) showed that 2 of the 5 clones from

E. angustifolium sediments (clone numbers C22 and C27) were grouped with the Fe oxidizers F. myxofaciens and G. ferruginea.

Known bacteria that express the gene soxB (such as T. denitrificans and T. thioparus) also clustered together with E. angustifolium clones.

103

Table 3.2. GenBank database BLAST analysis of clones from 16S rRNA library of sediment samples from the Southern Agon Goch wetland. P: clones from P. australis stand. E: clones from E. angustifolium stand. J: clones from Juncus spp. stand R: clones from river samples. aAccording to the information provided on clone description in the Genbank database.

16S Top hit % identity Isolation source clon of the top hit a e (Accession Number/ (% Description a) coverage) P11 GQ264470.1/Uncultured 96% Simulated low level bacterium clone (97%) waste site P12 HQ598721.1/Uncultured 98% Woodland soil Acidobacteria bacterium (96%) P13 AJ536881.1/Uncultured 99% Uranium mining bacterium waste pile, soil (94%) sample P18 HE604030.1/Uncultured 97% Acidic lignite mine bacterium lake (100%) P110 DQ223220.1/Uncultured 98% Subsurface water bacterium (96%) E21 99% Bulk soil of reed AB240242.1/Uncultured bed bioreactor bacterium (96%) E22 FJ538157.1/Uncultured beta 97% Paddy field soil proteobacterium (97%) E23 HQ599082.1/Uncultured 99% Woodland soil Acidobacteria bacterium (96%) E24 JX222236.1/Uncultured 98% Subsurface aquifer bacterium sediment (97%) E27 HM243885.1/Uncultured 96% Middle sediment bacterium clone from Honghu Lake (94%) R33 AJ292586.1/Uncultured 93% Polychlorinated eubacterium biphenyl-polluted (93%) soil

R34 DQ659229.1/Uncultured 90% Acid mine drainage bacterium clone (93%)

104

R36 KC541501.1/Uncultured 99% River sediment bacterium (96%) 92% Lake sediment J1 AB661524.1/denitrifying methane oxidizing bacteria (96%) Pelagic iron-rich HE604033.1/Uncultured 91% J2 aggregates bacterium (97%)

Suboxic freshwater- 90% J3 DQ676398.1/Acidobacteria pond sediment (94%)

AB821113.1/Uncultured 93% J4 Forest soil bacterium (97%) Minerotrophic acidic 2+ 92% J6 GU134911.1/Fe oxidizing fen enrichment culture (93%)

Gasoline-polluted 97% J7 JQ919773.1/Uncultured soil Bacteroidetes bacterium (96%)

95% Surface layer – J8 EU644258.1/Uncultured bacterium (96%) Tundra soil Pesticides waste 90% J9 GU477345.1/Rhodo- water pseudomonas sp. (96%)

105

Figure 3.3. Phylogenetic tree based on clone libraries of 16S rRNA from different stand of plants in the middle of the Southern Afon Goch wetland. Clones obtained from sediments from E. angustifolium, Juncus sp., P. australis and river samples are named in green, red, orange and blue, respectively. Lowercase letters next to the names represent 16S rRNA sequence used for the design of the “Ferrovum and relatives” set of primers (a), “Gallionella and relatives” set of primers (b) and sequences from soxB containing species (c). Sequences from other known bacteria were also included. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (500 replicates) are shown next to the branches. The scale represents the number of base substitutions per site.

106

3.4.2 Measurement of pH in water samples

In situ measurements of pH in surface waters from W1, W2 and W3 show a significant increase in pH (Figure 3.4) from the upper reaches

(pH ~2.2) to the end of the wetland (pH ~ 5). A similar decrease in acidity was observed in 2010 and 2011 (Figure 3.1a).

c 6 b

4 pH a

0 W 1 W 2 W 3

Figure 3.4. Average values of pH from water samples along the Southern Afon Goch wetland. Plots show the average values based on 5 replicates for each site. Error bars show standard deviation. Bars that do not share lowercase letters are significantly different (p < 0.05). Data were analysed by one-way ANOVA with Tukey posthoc.

107

3.4.3 Sulphur content and speciation in water samples

Dissolved S concentrations were significantly higher at the upper site

W1 compared to W2 and W3 (Figure 3.4a). In contrast, concentrations of particulate S showed no significant difference along

the three wetland sites (Figure 3.4b). Measurements of sulphate and sulphide showed a decrease in both compounds along the wetland.

Sulphate levels in W2 and W3 were significant lower than in W1

(Figure 3.4c), while sulphide concentration in W3 was significant lower than in W1 (Figure 3.4d).

3.4.4 Iron content and speciation in wetland water samples

Dissolved Fe concentrations were significantly higher at W1 compared to W2 and W3 (Figure 3.5a). An increase of Fe particulate concentration was observed in the middle site W2 and was significantly higher compared to W3 (Figure 3.5b). Measurements of

Fe2+ showed a decrease along the wetland but no significant differences between sample sites. In contrast, Fe3+ concentration in

W3 was significantly lower (p<0.05) when compared to W1 and W2

(Figure 3.5c).

108

a b

80 a 25

20 60

15 40 10

20 b 5 mg dissolved S / L Smg / dissolved

b L S / mg particulate

0 0 W1 W2 W3 W1 W2 W3

c d

40 0.25 a a

0.20 30

0.15 ab 20 0.10 b 10 mg sulphate / L / mg sulphate b b L / mg sulphide 0.05

0 0.00 W1 W2 W3 W1 W2 W3

Figure 3.4. Sulphur content and speciation in water samples from the Southern Afon Goch wetland. In situ analysis showing of dissolved (a) and particulate (b) values of S and total concentration of sulphate (c) and sulphide (d). Plots show the average values based on 5 replicates for each site. Boxes show the 25th and 75th percentiles, the black line within the boxes shows the median values. Whisker bars show the minimum and maximum values. Boxes that do not share lowercase letters are significantly different (p < 0.05). Data were analysed by one-way ANOVA with Tukey posthoc.

109

a b

15 200

a a 150 10

100

ab 5 b b 50 b mg dissolved Fe / L Fe / mg dissolved m g particulate Fe L /

0 0 W1 W2 W3 W1 W2 W3

15 2 * Fe 3 a Fe 10 * a m g L / 5 b

0 W1 W2 W3

Figure 3.5. Iron content and speciation in water samples from the Southern Afon Goch wetland. In situ analysis showing of dissolved (a) and particulate (b) values of Fe and total concentration of Fe2+ and Fe3+ (c). Plots show the average values based on 5 replicates for each site. Boxes show the 25th and 75th percentiles, the black line within the boxes shows the median values. Whisker bars show the minimum and maximum values. Boxes that do not share lowercase letters are significantly different (p < 0.05) than boxes from other sites. Boxes with asterisk are significant different than their grouped box in the sample site (p<0.05). Data were analysed by one-way ANOVA with Tukey posthoc.

110

3.4.5 Trace metal levels in water samples along the wetland

A significant difference in dissolved concentrations of Zn, Mn and Al in the upper site W1 compared to W2 and W3 was observed (Figure

3.6a, c, and d). No significant difference between the middle and lower sites W2 and W3 was observed for any of the metals analysed.

Concentrations of Cu, As and Pb showed no significant difference at any site along the wetland (3.6b, e and f). No significant variation in particulate trace metal levels at any site were observed (Figure 3.7).

3.4.6 Metal distribution along sediment depths

The metal distribution from different depths within the middle of the wetland (W2) was analysed. Cores were taken from different plant stands to detect possible difference in metal distribution according to plant species.

Core sampling revealed three different layers based on macroscopic observation (Figure 3.8). Subsamples from iron-rich top (0-10 cm) and middle (10-20 cm) layers along with a black mud bottom layer of

20 to 30 cm showed difference in Fe distribution compared to S and trace metals. No middle layer was detected in un-vegetated river samples.

Fe concentrations were significantly lower in the bottom layers compared to surface and middle layers. In contrast, S and trace metals show higher concentrations in the bottom layers. The presence of different plant species do not result in significant differences in any of the elements analysed in the sediments.

111

a b

6 5 a

4 3

b b 2 2

1 mg dissolved Zn / L Zn / mg dissolved mg dissolved Cu / L Cu / mg dissolved

0 0 W1 W2 W3 W1 W2 W3

c d

4 5 a a 4 3

2 b b b b 2

1 1 m g dissolved Al / L mg dissolved Mn / L Mn / mg dissolved

0 0 W1 W2 W3 W1 W2 W3

e f

15 15

10 10

5 5 m g dissolved As / L mg dissolved Pb / L Pb / mgdissolved

0 0 W1 W2 W3 W1 W2 W3

Figure 3.6. Concentrations of dissolved trace metals in water samples from the Southern Afon Goch wetland. Plots show the average values of dissolved Zn(a), Cu(b), Mn(c), Al(d), As(e) and Pb(f) based on 5 replicates for each site. Boxes show the 25th and 75th percentiles, the black line within the boxes shows the median values. Whisker bars show the minimum and maximum values. Boxes that do not share lowercase letters are significantly different (p < 0.05). Data were analysed by one-way ANOVA with Tukey posthoc.

112

a b

0.6 0.4

0.3 0.4

0.2

0.2 0.1 mg particulate Zn / L Zn / mg particulate mg particulate Cu / L Cu / mgparticulate

0.0 0.0 W1 W2 W3 W1 W2 W3

c d

0.60 5

4 0.55

3 0.50 2

0.45 1 mg particulate Al / L / Al mg particulate mg particulate Mn / L Mn / mg particulate

0.40 0 W1 W2 W3 W1 W2 W3

e f

0.70 0.9

0.65 0.8

0.60 0.7

0.55 0.6 mg particulate As / L As / mg particulate mg particulate Pb / L Pb / mg particulate

0.50 0.5 W1 W2 W3 W1 W2 W3

Figure 3.7. Concentrations of particulate trace metals in water samples from the Southern Afon Goch wetland. Plots show the average values of dissolved Zn(a), Cu(b), Mn(c), Al(d), As(e) and Pb(f) based on 5 replicates for each site. Boxes show the 25th and 75th percentiles, the black line within the boxes shows the median values. Whisker bars show the minimum and maximum values. Data were analysed by one-way ANOVA with Tukey posthoc.

113

Figure 3.8. Sediments layers observed from core sampling. The surface layer contains larger soil particles and plant debris. The middle layer is characterized by a red-brown colour with compacted ochre and the bottom layer was mainly formed by black mud.

Sediment cores from W1, W2 and W3 were divided into a top layer

(0-20 cm) and bottom layer (20-50 cm) for further validation of the elements distribution. The analysis showed that S levels in the top layer were significantly lower than the bottom layer at W1 and W3

(Figure 3.10a). In contrast, Fe showed a significantly lower concentration in the top layer only in W2 (Figure 3.10b). There was a significant reduction in Fe concentrations in the W3 top layer when compared with the top layers of W1 and W2.

114

a b 250 150 a a a b b a b 200 a c a b c b e b 100 b 150 c d b e e 100 c c m gS g / mg Fe / g mg/ Fe 50 c e 50 d a a a a a c c c a a c 0 0 R E J P E J P R E J P R E J P E J P R E J P

Surface M id d le B ottom Surface M id d le B ottom

c d

40 b 40

b 30 30

20 20 b b mg g Zn / mgg Cu / 10 10

a a a a a a a a a a a a a a 0 0 R C J P C J P R C J P R E J P E J P R E J P

Surface M id d le B ottom Surface M id d le B ottom

e f 30 0.20 a b b b a 0.15 b b 20 c

a b a b c 0.10 a a a a

m g Al / g c 10 mg Mn /g a 0.05 a a a a a a a a 0 0.00 R E J P E J P R E J P R E J P E J P R E J P Surface M id d le B ottom Surface M id d le B ottom

g h 0.6 15

b 0.4 10

a a a c a g mg Pb / m g A s / g 0.2 5 d d a a a a a a a d a b b b b b b c

0.0 0 R E J P E J P R E J P R E J P E J P R E J P

Surface M id d le B ottom Surface M id d le B ottom

Figure 3.9. Concentrations of Fe, S and trace metals with sediment depths in the middle of the Southern Afon Goch wetland. Plots show the average values of total Fe(a), S(b), Cu(c), Zn(d), Mn(e), Al(f), As(g) and Pb(h) based on 5 replicates for each site. Boxes show the 25th and 75th percentiles, the black line within the boxes shows the median values. Whisker bars show the minimum and maximum values. Data were analysed by one-way ANOVA with Tukey posthoc. R: river, C: E. angustifolium stand, J: Juncus sp. stand, P: P. australis stand.

115

a b

50 200

40 * 150 a a

30 100

20 b

50 10 * * mgsediment gdry S / mg Fe / g dry sediment mgdry g / Fe 0 0 W1 W2 W3 W1 W2 W3

0-20cm 20-50cm

Figure 3.10. Concentrations of total S (a) and Fe (b) in top and bottom sediment layers obtained by ICP-AES. Plots show the average values based on 5 replicates for each site. Boxes show the 25th and 75th percentiles, the black line within the boxes shows the median values. Whisker bars show the minimum and maximum values. Boxes that do not share lowercase letters are significantly different than boxes from the sample layer in other sites (p < 0.05). Boxes with an asterisk are significantly different than their grouped box in the sample site (p<0.05). Data were analysed by one-way ANOVA with Tukey posthoc.

Surface core concentrations of Cu, Mn, Al and Pb were significantly lower than in the bottom layer at W1 only (Figure 3.11). Zn was present at higher concentrations in the upper layers of samples from all three sites while As showed no significant difference at any site.

No significant reduction in trace metals levels were observed in the top layers along the wetland. Cu was the only trace metal that showed significant changes in the bottom layer with lower concentrations at W2 and W3 compared to W1.

116

a b

40 25 a

20 30

15 20 b 10

10 * 5 b * * * mg Zn / g dry sediment dry g mg/ Zn mg Cu / g dry sediment mgdry g Cu / 0 0 W1 W2 W3 W1 W2 W3

c d

0.8 40

0.6 30

0.4 20

0.2 10 * mg Al / g dry sediment dry g mg/ Al mg Mn / g dry sediment mgdry g Mn / * 0.0 0 W1 W2 W3 W1 W2 W3

e f

1.5 6

1.0 4

0.5 2 mgsediment gdry As / mg Pb / g dry sediment mggdry Pb / * 0.0 0 W1 W2 W3 W1 W2 W3

0-20cm 20-50cm

Figure 3.11. Concentrations of total Zn (a), Cu (b), Mn (c), Al (d), As (e) and Pb (f) in surface and bottom sediment layers obtained by ICP-AES. Plots show the average values based on 5 replicates for each site. Boxes show the 25th and 75th percentiles, the black line within the boxes shows the median values. Whisker bars show the minimum and maximum values. Boxes that do not share lowercase letters are significantly different than boxes from the sample layer in other sites (p < 0.05). Boxes with asterisk are significant different than their grouped box in the sample site (p<0.05). Data were analysed by one-way ANOVA with Tukey posthoc.

117

3.4.7 Expression of dsrA and soxB genes

The mRNA transcript abundance of the dsrA gene (essential for bacterial dissimilatory sulphate reduction) and the soxB gene

(involved in the bacterial S oxidation mechanism) were quantified by qPCR using specific primers.

The upper sediment layer at W2 showed a significant increase of soxB expression compared to the bottom layer, and to the other two sites.

No significant differences were observed in the bottom layers along the wetland. There was no significant difference between surface and bottom layers in the expression of dsrA. However, the amount of dsrA transcript in the top layers of W1 and W2 were significantly higher than in W3 (Figure 3.12)

3.4.8 Expression of 16S rRNA genes from Fe oxidizing bacteria

No significant partitioning in the expression of Geobacter sp. 16S rRNA was observed in the W2 and W3 cores (Figure 3.13b) but a significant reduction was seen in the bottom layer of W1 compared to

W2 and W3. Levels of “Gallionella and relatives” 16S rRNA showed no significant differences in any of the layers or sites (Figure 3.13c) while “Ferrovum and relatives” 16S rRNA gene showed a significant increase in the top layer of W2 compared to the rest of the samples

(Figure 3.13d). Abundance of total bacteria 16S rRNA gene expression was significantly higher in surface compared to bottom layers in W1 and W2 (Figure 3.13a).

118

a b

20 15

* a b 15 a 10 b

10 a a

5 5 transcripts / g wet sediment / transcripts transcripts / g wet sediment / transcripts 10 10 0 0 lo g lo g W1 W2 W3 W1 W2 W3 soxB dsrA

0-20cm 20-50cm

Figure 3.12. Abundance of RNA transcripts from bacteria genes related to S metabolism bacteria in sediments from the Southern Afon Goch wetland. Bars show average values based on 5 replicates for each site of soxB (a) and dsrA (b) in surface and bottom sediment layers. Error bars show the standard deviation. Bars that do not share lowercase letters are significantly different than boxes from the sample layer in other sites (p < 0.05). Bars with asterisk are significant different than their grouped box in the sample site (p<0.05). Data were analysed by one-way ANOVA with Tukey posthoc.

119

a b

20 15

* * 15 b * b 10 a 10

5 5 transcripts / g wet sediment / transcripts transcripts / g wet sediment / transcripts 10 10 0 0 lo g lo g W1 W2 W3 W1 W2 W3 Total bacteria 16S rRNA Geobacter spp. 16S rRNA

c d 10 20

8 15 *

6 10 4

5 2 transcripts / g wet sediment / transcripts transcripts / g wet sediment / transcripts 10 10 0 0 lo g lo g W1 W2 W3 W1 W2 W3 "Gallionella and relatives" 16S rRNA "Ferrovum and relatives" 16S rRNA

0-20cm 20-50cm

Figure 3.13. Abundance of 16S rRNA transcripts from total and Fe reducing/oxidizing bacteria in sediments from the Southern Afon Goch wetland. Bars show average values based on 5 replicates for each site of total bacteria (a) Geobacter spp. (b) Gallionella and relatives (c) and Ferrovum and relatives (d) in surface and bottom sediment layers. Error bars show the standard deviation. Bars that do not share lowercase letters are significantly different than boxes from the sample layer in other sites (p < 0.05). Bars with asterisk are significant different than their grouped box in the sample site (p<0.05). Data were analysed by one-way ANOVA with Tukey posthoc.

120

3.4.9 Expression of soxB and F. myxofaciens genes in different plant stand in W2 sediment

The increased abundance of the functional gene soxB and the

Ferrovum and relatives 16S rRNA gene that was observed in W2 upper sediments was further analysed in sediments associated with different plant species and a river sample. No significant difference in abundance was observed in soxB (Figure 3.14a). However, Ferrovum and relatives showed a significant decreased abundance in the river sample compared to the vegetated sediments, yet there was no significant difference between the different plant species (Figure

3.14b).

a b

15 20

15 10 * 10

5 5 transcripts / g wet sediment wet g / transcripts transcripts / g wet sediment wet g / transcripts 10 0 10 0 log R P E J log R P E J soxB "Ferrovum and relatives" 16S rRNA

Figure 3.14. Abundance of RNA transcripts from soxB gene and 16S rRNA from “Ferrovum and relatives” in river samples and different plant stand in the middle of the Southern Afon Goch wetland. Bars show average values based on 5 replicates for each site of soxB (a) and “Ferrovum and relatives” (b) in surface sediments. P: P. australis stand. E: E. angustifolium stand. J: Juncus spp. stand R: river sample without vegetation. Error bars show the standard deviation. Bars with asterisk are significant different (p<0.05). Data were analysed by one-way ANOVA with Tukey posthoc.

121

3.4.10 ITRAX analysis

ITRAX core scan analysis showed Fe values between 50 ppm and 100 ppm along the core depths in all the samples. W1 and W2 showed peaks of Fe in the upper layers (Figure 3.15a and b). Values of S were higher in bottom layer of W3 and in upper layers of W1. Levels of trace metals varied depending on the element. Al showed a stable profile (~30 ppm) along depths, while Mn, As and Pb showed considerable variations along depths and between sites (Figure 3.15a b and b). S, Zn and Cu showed similar profiles that were different from the Fe profile. Peaks of the S, Zn and Cu could be observed in identical positions in upper (~200 mm in W1), middle (~300 mm in

W2) and bottom depths (~400 mm in W3) (Figure 3.16).

a mg Fe / g dry sediment

W1 W2 W3

0 50 100 150 0 50 100 150 0 50 100 150

0 0 0

100 100 100

200 200 200

300 300 300 Core depth (mm) depth Core

4 0 0 4 0 0 4 0 0

5 0 0 5 0 0 5 0 0

122 b mg S / g dry sediment

W1 W2 W3

0 50 100 150 0 50 100 150 0 50 100 150

0 0 0

100 100 100

200 200 200

300 3 0 0 3 0 0 Core depth (mm) depth Core

4 0 0 4 0 0 4 0 0

5 0 0 5 0 0 5 0 0 c mg Zn / g dry sediment

W1 W2 W3

0 50 100 150 200 0 50 100 150 200 0 50 100 150 200

0 0 0

100 100 100

200 200 200

3 0 0 3 0 0 3 0 0 Core depth (mm) depth Core

4 0 0 4 0 0 4 0 0

5 0 0 5 0 0 5 0 0

123 d mg Cu / g dry sediment

W1 W2 W3

0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100

0 0 0

100 100 100

200 200 200

3 0 0 3 0 0 3 0 0 Core depth (mm) depth Core

4 0 0 4 0 0 4 0 0

5 0 0 5 0 0 5 0 0 e mg Mn / g dry sediment

W1 W2 W3

0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5

0 0 0

100 100 100

200 200 200

3 0 0 3 0 0 3 0 0 Core depth (mm) depth Core

4 0 0 4 0 0 4 0 0

5 0 0 5 0 0 5 0 0

124 f mg Al / g dry sediment

W1 W2 W3

0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50

0 0 0

100 100 100

200 200 200

3 0 0 3 0 0 3 0 0 Core depth (mm) depth Core

4 0 0 4 0 0 4 0 0

5 0 0 5 0 0 5 0 0 g mg As / g dry sediment

W1 W2 W3

0 1 2 3 4 0 1 2 3 4 0 1 2 3 4

0 0 0

100 100 100

200 200 200

3 0 0 3 0 0 3 0 0 Core depth (mm) depth Core

4 0 0 4 0 0 4 0 0

5 0 0 5 0 0 5 0 0

125

h mg Pb / g dry sediment

W1 W2 W3

0 50 100 150 200 0 50 100 150 200 0 50 100 150 200

0 0 0

100 100 100

200 200 200

3 0 0 3 0 0 3 0 0 Core depth (mm) depth Core

4 0 0 4 0 0 4 0 0

5 0 0 5 0 0 5 0 0

Figure 3.15. Profiles of Fe, S and trace metal concentrations along sediment depths using ITRAX core scanning. Profiles show concentration (mg/g) of Fe (a), S (b) and trace metals (c-h) at intervals of 1mm depth from 3 replicates at W1 (left), W2 (middle) and W3 (right) using ITRAX core scanning. Lines represent the mean of 3 replicates of independent sediment cores. The upper values of the standard deviation is shown as a dotted line.

126

mg / g dry sediment

W1 W2 W3

0 50 100 150 0 50 100 150 0 50 100 150 0 0 0

100 100 100

200 200 200

300 300 300 Core depth (mm) depth Core

4 0 0 4 0 0 4 0 0

5 0 0 5 0 0 5 0 0

Figure 3.16. Overlapped profiles of Fe, S, Cu and Zn along sediment depths using ITRAX core scanning. Co-localization is observed between S and the trace metals in all the samples. Lines represent the mean of 3 replicate sediment cores.

3.4.11 Total and dissolved organic carbon and nitrogen concentrations along core depths

Concentrations of DOC (~ 1 mg/g) show no marked difference along depth. DON levels increase at 20cm in W1 but there is no significant variation (~0.01 – 0.1) along depths (Figure 3.17). Total C and N show increased concentrations in W1, W2 and W3 at 20, 30 and 40 cm depths respectively (Figure 3.18).

127

Dissolved Organic C (mg/g) Dissolved Organic N (mg/g)

0 0

10 10

20 20 W1

30 30 Core depth (cm) depth Core

40 40

50 50

0 .0 0 .5 1 .0 1 .5 2 .0 2 .5 0 .0 0 .1 0 .2

0 0

10 10

20 20 W2

30 30 Core depth (cm) depth Core 40 40

50 50

0 .0 0 .5 1 .0 1 .5 2 .0 2 .5 0 .0 0 .1 0 .2

0 0

10 10

20 20 W3

30 30 Core depth (cm) depth Core

40 40

50 50

0 .0 0 .5 1 .0 1 .5 2 .0 2 .5 0 .0 0 .1 0 .2

Figure 3.17. Dissolved organic C and N along sediment depths in the Southern Afon Goch wetland. Lines represent the mean of 3 replicates of independent sediment cores. Error bars show standard deviations.

128

% Total C % Total N

0 0

10 10

20 20 W1 30 30 Core depth (cm) depth Core 40 40

50 50

0 2 0 4 0 6 0 8 0 0 1 2 3 4

0 0

10 10

20 20 W2 30 30 Core depth (cm) depth Core 40 40

50 50

0 2 0 4 0 6 0 8 0 0 1 2 3 4

0 0

10 10

20 20

W3 30 30

Core depth (cm) depth Core 40 40

50 50

0 2 0 4 0 6 0 8 0 0 1 2 3 4

Figure 3.18. Percentage of total N and C along sediment depths in the Southern Afon Goch wetland. Lines represent the mean of 3 replicates of independent sediment cores. Error bars show standard deviations.

129

3.5 DISCUSSION

3.5.1 Surface sediment bacteria and Fe and S speciation along the wetland

Concentrations of total dissolved S reduced rapidly in parallel with sulphate (Figure 3.4). Sulphate is the main S compound produced during the oxidation of pyrite and subsequent generation of AMD

(Evangelou and Zhang, 1995). The reduction in dissolved S suggests that most of the S that is being removed is in the form of sulphate.

The key process facilitating the removal of sulphate from the water column is its reduction and transformation to hydrogen sulphide

(Akcil and Koldas, 2006). However, sulphide levels were much lower than the other S compounds measured. One explanation is that other

S compounds such as sulphite, thiosulphate and elemental S that are also susceptible to chemical and biological oxidation (Auernik and

Kelly, 2008, Amouric et al., 2009) are present in significant quantities. Measurement of sulphide is important as this compound is capable of binding to trace metals and co-precipitating (Machemer and Wildeman, 1992). However, large amounts of sulphide were not be detected in the water column due to a rapid precipitation of metal sulphides close to the source in the upper reaches of the wetland.

Clones obtained from the E. angustifolium stand (W2) were phylogenetically related to known soxB (S oxidizing gene) containing species. For example clones E22 and E27 were grouped with T. denitrifians and T. thioparus, while E24 was clustered together with

D. desulfuricans. All these bacteria are known for their capacity to

130 oxidise S. Sulphur oxidizing (SOB) and reducing (SRB) bacteria were previously analysed in Chapter 3 and recorded differences in the abundances of these bacteria, depending on the plant species. The results from the present Chapter confirm this pattern and suggest that the potential bacteria–mediated S oxidation is due to the presence of the E. angustifolium stand.

The reduction of dissolved Fe levels when the AMD enters the middle of the wetland (W2) occurs as the particulate Fe concentration increases. This can be explained by a transition from soluble Fe+2 to aggregates of Fe3+ in the form of Fe3+ oxides and hydroxides

(Johnson and Hallberg, 2005). In situ analysis of Fe species show that, even though the difference between Fe2+ and Fe3+ concentrations decrease as the water column moves through the wetland, amounts of Fe3+ are always higher (Figure 3.5c). Moreover, the increase in particulate Fe in W2 does not correlate with differences in the Fe3+ levels. This suggests that other mechanisms are involved in Fe aggregation and soluble Fe removal from the water column. In addition, no clones obtained from W2 sediments were identified as known or related Fe oxidizing bacteria.

3.5.2 Metal distribution along sediment depths

Two approaches for evaluating total metal and S distribution with core depth involved high resolution ITRAX scans to obtain continuous

131 measurements and ICP-AES analysis of the core divided into a top and bottom layer.

No difference in Fe concentration between the top and bottom layers was observed in sediment cores from site W1 (Figure 3.10b).

However, in W2, Fe values were higher in the top. It is probable that the elevated concentration of particulate Fe from the water column in the middle wetland is contributing to an elevated concentration of total Fe in the first 20 cm of the sediment due to precipitation from the water column.

Peaks in the Fe ITRAX profile observed in W1 and W2 also indicate elevated Fe in the first 20 cm compared to the lower depths (Figure

3.15a). Levels of Fe in the first 10 cm of W2 were higher than in W1 and W3. The elevated Fe at the surface may be due to ochreous precipitates of Fe oxy-hydroxydes (Galán et al., 2003).

Levels of S are significantly lower in the first 20 cm compared to bottom layers in cores from W1 and W3 (Figure 3.10a). ITRAX profiles confirm this distribution in W2 and W3 where higher peaks of

S are detected in the 30 cm and 40 cm respectively (Figure 3.15b).

However W1 showed higher peaks in the top layer. Deposition of sulphate onto the sediments in the form of the iron- oxyhydroxysulphate mineral schwertmannite mediated by low pH values and high Fe concentrations has been observed in river sediments affected with AMD (Chen et al., 2015b). Moreover, analysis of S compounds in paddy field soil cores impacted by AMD showed

132 that water-soluble sulphate concentrations are higher at 20 cm and

30 cm depth compared to the surface (Yang et al., 2016).

An increase of metal concentration in the ITRAX profile of Zn, Cu, and

Pb in the 10-20cm segment is observed, and ICP-AES top/bottom layer measurement confirmed this distribution. However, layer measurements of Al, Mn and As showed different patterns compared to the ITRAX data. Previous studies have demonstrated that the

ITRAX profile need to be carefully interpreted in sediments with different water content (Tjallingii et al., 2007) and large variations in organic matter and carbon concentration (Chawchai et al., 2016).

Therefore different concentrations of C and N can play an important role in explaining the difference between the ITRAX and ICP-AES results. Previous studies have demonstrated that mobilization of Cr,

Hg, Cu and As can be influenced by dissolved organic matter concentrations in the soil solution from polluted wetlands (Kalbitz and

Wennrich, 1998). In this study, no significant variations of DOC and

DON values along the sediments were observed. Therefore no correlation of DOC and DON with metal patterns along sediment depths is evidenced. Accumulation of total C and N is observed in bottom layers in W1 and W2 in a similar pattern as S in W2 and Zn in

S3. However, in order to confirm possible mobilization of metals mediated by non-dissolved organic matter, experiments regarding direct extraction of metals from organic sediment fractions is required

(Kaasalainen and Yli-Halla, 2003).

133

In the case of Al, this element at the limit of ITRAX detection due to its low atomic weight and CPS can be underestimated as previously observed (Rothwell et al., 2006).

When Fe, Zn, Cu and S ITRAX profiles overlapped, Zn, Cu and S displayed the same pattern and peaks were observed in the same depths (Figure 3.16). However, the Fe ITRAX profile shows a different distribution. This suggests that the mobility of these trace metals is modulated by S. It is likely to be due to immobilization of Cu and Zn in sulphide compounds, considering the elevated level of reduced S compounds in AMD environments (Yang et al., 2016).

3.5.3 Expression of S metabolic genes

An increased activity of S oxidation in the middle of the wetland was detected by measuring the expression of the soxB gene (Figure

3.12a). This result coincides with the elevated abundance of known

SOB in the middle of the wetland. Even though difference in SRB abundance was previously detected (Chapter 2), no significant variation in the expression of dissimilatory sulphate reduction was observed as measured by the expression of dsrA gene, suggesting that the expression of this mechanism is constant along the wetland.

A similar increased expression of the soxB gene has been previously associated with the oxidation of thiosulphate to sulphate using the

SoxCD enzyme complex in a terrestrial sulphidic spring (Headd and

Engel, 2013). This indicates the need to measure other S compounds

134 beside sulphate and sulphide to further understand the partitioning of

S oxidation expression along the wetland.

No difference in soxB gene expression was detected between sediments from different plant species or from unvegetated sediments (Figure 3.14a). It is known that the soxB gene encodes part of a periplasmic thiosulphate oxidizing complex that is widespread among different phylogenetic groups (Friedrich et al.,

2001, Petri et al., 2001). This diversity has also been observed in thiosulphate oxidizing bacteria expressing the soxB gene in rhizosphere soil from different plant species (Ghosh et al., 2006,

Anandham et al., 2008, Li et al., 2014). Ubiquitous expression of soxB can thus explain the similar expression levels observed in different wetland plant species and non-vegetated areas.

3.5.4 Expression of Fe bacteria ribosomal genes

This study found that DNA from Gallionella show similar abundance in the middle and lower reaches of the wetland. RNA expression of the group “Gallionella and relatives” found no differences between the sites and depth in the sediments (Figure 3.18c). These results suggest there is no substantial modification of Gallionella abundance in the wetland, despite the variations of pH and soluble Fe2+.

Gallionella ferruginea was previously detected in acid water (pH 2.7 to 3,4) from mine tailings (Bruneel et al., 2006) and in biofilms from

135 mine waste with high concentration of trace metals (Kim et al.,

2002).

DNA analysis show that Geobacter spp. abundance increase in surface sediments along the wetland. RNA expression from Geobacter spp. showed higher abundance in the bottom layer (Figure 3.13b).

Anaerobic Fe metabolism from bacteria related to the Geobacter taxa was previously observed (Coates et al., 2001) and utilization of Fe minerals in anaerobic sediments by these bacteria has be described

(Adams et al., 2007).

The highest variation in rRNA expression was observed in the case of

“Ferrovum and relatives”. As for 16S RNA gene analysis, the highest abundance of rRNA transcripts was observed in the middle wetland

(Figure 3.13d). Strains of Ferrovum myxofaciens capable of catalyzing the oxidative dissolution of pyrite were previously isolated from Parys Mountain (Johnson et al., 2013). This present study has shown that Ferrovum can also thrive in the surrounding wetland under less extreme conditions and its abundance coincides with an improvement in the water quality in the middle area of the wetland.

Even though no specific functional gene related to Fe metabolism was measured, the analysis of different obligate Fe metabolising bacteria suggest that Fe oxidation is an important mechanism along the wetland.

136

3.6 CONCLUDING REMARKS

The distribution and mobilization of metals as part of the attenuation of AMD were analysed in more detail in this chapter and confirm that variation in metal concentration with depth and potential availability is part of a complex dynamic of metal chemistry in wetlands

(Dunbabin and Bowmer, 1992, Sheoran and Sheoran, 2006). It was observed that there was a partitioning of Fe in the surface sediments while S and trace metals accumulated in deeper layers of the sediment. ITRAX technology was used for the first time in assessing the metal retention capacity of a wetland in this study. However, the timeframe within these metals were deposited onto the wetland sediments remain unclear. For this, ITRAX analysis can be coupled with isotope measurements to obtain information on historical metal deposition in the SAG. Analysis using ITRAX scans and 210Pb and

137Cs measurements were successfully applied to understand the chronology of metal contamination in former mining landscapes

(Miller et al., 2015).

Chapter 2 highlights the important role of bacteria in the SAG wetland remediation process. New insight into the remediation mechanisms occurring along the SAG wetland was provided by an examination of the sediment microbiology and its relation with metal distribution and speciation along the wetland. A higher S oxidation activity, an increase in the Fe oxidizing Ferrovum taxa in Juncus spp. sediments and abundance of S metabolizing bacteria in E. angustifolium sediments were observed where the AMD stream reaches the middle 137 wetland at a very short distance from the source. These spatial changes were co-localized with the presence of more diverse vegetation and abundance (Site W2) plus changes in trace metal chemistry and Fe and S speciation that lead to an a attenuation of the

AMD. The first evidence that SAG vegetation could influence bacteria- mediated AMD attenuation was obtained in this part of the study. The positive contribution of plants to microbial-mediated remediation of

AMD is likely to include the provision of oxygen to facilitate aerobic metabolism and organic matter as a substrate, and this is explored in

Chapter 4.

138

3.7 SUPLEMENTARY DATA

Calibration curves from ITRAX results

Fe 120

100

60 Título del eje 40

0 0 500 1000 1500 2000 y = 0.0379x + 34.03 mg / g R² = 0.862

S 70 60

40 CPS 30

0 0 0.2 0.4 0.6 0.8 1 1.2 mg / g y = 50.044x - 2.5272 R² = 0.8503

139

Zn 10 9 8 7 6 5 CPS 4 3 2 1 0 0 5 10 15 20 y = 0.3663x25 + 0.276430 mg / g R² = 0.9815

Cu 60

30 CPS

0 0 50 100 150 200 250 300 y = 0.1813x + 4.6082 mg / g R² = 0.9561

140

Mn 0.25

0.2

0.15

0.1 CPS

0.05

0 0 0.2 0.4 0.6 0.8 1

-0.05 y = 0.2764x - 0.045 mg / g R² = 0.9056

Al 14

8 CPS 6

0 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 y = 72.297x - 9.5783 mg / g R² = 0.8557

141

As 1.2

0.8

0.6

0.4 CPS

0.2

0 0 1 2 3 4 5 6 -0.2

-0.4 mg / g

y = 0.2568x - 0.4763 R² = 0.8363

Pb 5 4.5 4 3.5 3 2.5 CPS 2 1.5 1 0.5 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 mg / g y = 3.1294x + 0.4763 R² = 0.8605

Figure S3.1 Calibration graphs for each element analysed by ITRAX core scans. The plots show known total metal values previously obtained by ICP-AES (X axis) and the respective CPS values (Y axis). Linear regression lines, regression equations and R2 values are also shown.

142

Chapter 4: EFFECT OF OXYGEN AND ORGANIC SUBSTRATES ON BACTERIAL ACTIVITY AND METAL ATTENUATION IN INCUBATED WETLAND SEDIMENTS

Aguinaga, O. E., White, K. N., Dean, A. P., & Pittman, J. K.

The author performed sediment incubation experiments, analysis of metagenomic data, interpretation and writing up the manuscript. Jon

Pittman, Keith White and Andrew Dean provided full guidance and manuscript review.

143

4.1 ABSTRACT

Plants and microorganisms have the ability to modify metal speciation and distribution in acid mine drainage (AMD) impacted wetlands, which can generate a mechanism of metal attenuation and therefore, a passive treatment of AMD. However it is a very complex mechanism as multiple physicochemical, biological and environmental variables are involved. In this study, surface sediments from the Afon Goch wetland, which is a well - adapted remediating system of an AMD stream from the abandoned copper mine of Parys Mountain, were incubated in artificial AMD and the effect of bacteria presence and addition of organic matter and oxygen (as an attempt to mimic presence of vegetation) in the metal attenuation process was evaluated. Measurements of metal partitioning in incubated sediments and metagenomic analysis of sediment bacteria communities concluded that addition of organic matter in the form of a mix of free volatile acid enhanced the remediation efficiency though the generation of higher levels of particulate Fe, Zn and Cu. The incubated sediments where these changes were detected also showed higher relative abundance of functional genes related to bacterial transformation of S compounds. Oxygenated and non-oxygenated treatments show no marked difference in bacteria activity. This study confirmed the importance of organic matter in the AMD remediation observed the Afon Goch wetland. Further studies including plants are proposed in order to further understand the role of vegetation in the remediation process.

144

4.2 INTRODUCTION

In Chapter 2, evidence for the enhancement of microbial diversity by vegetated wetland sediments was presented. In Chapter 3 it was shown that the more diverse and abundant plant community in the middle of the wetland was co-localized with higher Fe and S bacterial oxidation activity.

The presence of plants have been proposed to alter the speciation and behaviour of metals including within acid mine drainage environments by the production of organic matter that adsorbs and immobilize metal ions and alters their bioavailability characteristics

(Idaszkin et al., 2017). An increase in organic compounds in soil following root exudation due to metal stress has been observed in vegetated environments (Dong et al., 2007, Javed et al., 2018).

These organic compounds can also act as carbon substrates for bacterial growth thereby enhancing bacterial enzymatic activities that enhance AMD mitigation (Gibert et al., 2004). Plants also release oxygen via radial oxygen loss into the rhizosphere (Colmer, 2003) creating aerobic conditions that affect the water chemistry in waterlogged soils (Engelaar et al., 1995) and increase mobilization of trace metals (Wang et al., 2011, Yang et al., 2014). Moreover, oxygen release from plant roots creates the ideal microenvironment in the rhizosphere for deposition of S compounds through dissimilatory S reduction by microorganisms (Wiessner et al., 2017)

145

However, metal mobilization or retention by wetland plants is a complex and dynamic process. Organic matter and oxygen release can vary depending on plant species, the degree of root tissue sloughing, the surrounding microbiota and seasonal fluctuations

(Jacob and Otte, 2003). For example, a study that evaluated the effect of plants on metal behavior in a wetland receiving AMD over a period of 1 year concluded that the complexity of chemical and biological processes may prevent the complete characterization of the AMD attenuation processes (Batty et al., 2006).

Experiments using artificial AMD have been previously carried out to understand processes such as sulphate reduction activity (La et al.,

2003), bacterial activity (Morales et al., 2005) and metal recovery

(Luo et al., 2014) in AMD environments. Microcosm AMD experiments under controlled conditions have previously been used to elucidate mechanisms related to Fe and Mn retention (Henrot and Wieder,

1990), acidity attenuation (Frömmichen et al., 2004) and the role of microbial populations (Becerra et al., 2009). Various microbial and substrate amendments have been used to improve the efficiency of

AMD remediation in microcosms (Jin et al., 2008). Comparative metagenomic analysis has been previously performed in AMD environments to understand the taxonomy of unculturable microorganisms (Schloss and Handelsman, 2005), dominance of specific taxa (Hua et al., 2015) and functional traits related to bacterial adaption (Liljeqvist et al., 2015), resistance genes (Mirete et al., 2007) and metal transformation (Bonnefoy and Holmes, 2012). 146

Metagenomics has also been used for understanding microbial changes in incubated soil microcosms artificially amended with AMD

(Brantner and Senko, 2014) and to evaluate prokaryotic diversity and bacteria – mediated As oxidation in incubated AMD experiments

(Tardy et al., 2018).

The aim of this study was to evaluate the response to plant-derived factors by adding organic matter and oxygen to AMD wetland sediments submerged in an artificial AMD solution. By mimicking the provision of these two components of wetland plants, we intended to further understand how plants influence the chemistry and microbiology of AMD impacted wetland systems. In parallel, metal and S transformation was compared with sterilized (irradiated) sediments to evaluate the role of microorganisms in the AMD processes. The role of the microbial community and metabolic processes in metal and S behavior was also assessed by a comparison of the two treatments. Therefore, the following aim and objectives were formulated:

Aim:

To understand the effect of environmental factors provided by plants, specifically carbon and oxygen, on sediment bacterial activity and metal behavior under controlled exposure to AMD.

147

Objectives:

• Incubation of microcosm with, respectively, natural and

sterilized sediments treated with carbon (organic acids) and

oxygen (air in-flow).

• Measurement of Fe and S speciation and Zn and Cu partition

during the incubation period of 88 days.

• Metagenomic sequencing of DNA extracted from the natural

sediment microcosm.

• Comparative analysis of microbial taxonomic and metabolic

traits from the two microcosms.

• Analysis of functional genes related to Fe, Zn, Cu and S

metabolism.

4.3 MATERIALS AND METHODS

4.3.1 Sediment sampling

Sediment samples (n = 10) were obtained from the middle of the

SAG wetland (Site S2 in Chapter 2). The aim was to sample the microbial community from the water / sediment interface and also the site of Fe and trace metal precipitation (see Chapter 2).

Sediments (approximately 5 g each) surrounding Juncus sp. were collected to approximately 1 cm depth using a plastic scoop and sealed in a plastic bag. On return to the laboratory, samples were mixed together and then refrigerated at 4 C until used for incubation experiments. ⁰

148

4.3.2 Experimental design

4.3.2.1 Sediment irradiation

In order to inhibit microorganism activity without affecting the sediment properties, 2.5 kg of the sediments were split into 250 g aliquots, placed in polyethylene bags, and irradiated at a dose of 40 kGy using a 60Co gamma irradiator (Dalton Cumbria Facility,

University of Manchester). Inhibition of microorganisms through irradiation is important in order to prevent any metabolic function that could lead to metals transformation and hence, assure that only abiotic processes were present in the irradiated sediments.

Comparisons with non – irradiated sediments (where both biotic and abiotic processes can influence metals transformation) were carried out.

4.3.2.2 Microcosm set up

Microcosms consisting in 5 g of non-irradiated sediments (NS) and irradiated sediments (IS) submerged in a metal solution with similar metal and pH conditions to the Afon Goch wetland water column were prepared in 50 mL open tubes. A total of 20 mL of the simplified artificial AMD solution (pH = 3.0, 10 ppm Fe, 5 ppm Zn and 2 ppm

Cu) was added to each tube. Previous sediment incubation experiments detected changes in Fe+2 oxidation rates after 55 days incubation (Senko et al., 2011) and addition of complex carbon sources has been previously investigated over 60 days (Adams et al., 149

2007). In this study a total of 40 (20 NS and 20 IS) microcosms were incubated over 88 days at a constant temperature of 10 °C.

Water level was maintained with distilled water due to evaporation loss when needed. Count of colony forming units on LB media agar was performed to monitor changes in bacterial number in the sediments.

4.3.2.3 Treatments

The microcosms were grouped in 3 different treatments and a control

(5 replicates per group) for both the NS and IS as shown below.

• Oxygenated (OX): Filtered air through 0.45 µm pore size

cellulose membrane (Camlab, UK) was continuously introduced

into the water column.

• Added Carbon (AC): A mixture of volatile free acids (VFA)

(Sigma, UK) was regularly inoculated in the microcosm to

provide an organic substrate. The VFA content of a pilot-scale

constructed wetland showed that acetic acid values ranged

between 7.91 and 13.91 mg/L (Huang et al., 2005). Therefore

in this study a final concentration of 10 mg/L of each acid from

the mix (acetic, formic, propionic, isobutyril, butyric, isovaleric,

valeric, isocaproic, hexanoic and n-heptanoic acid) was added

in the water column by inoculating aliquots every 10th day.

• Oxygenated + added carbon (COX): Microcosms were aerated

and the volatile free acid mix was added, as described above.

• Control microcosms (C): No aeration or added carbon. 150

The conceptual diagram below illustrates the 3 treatments and the control experiments used in this study:

4.3.3 Measurement of environmental variables

The pH was monitored using a portable pH meter (HANNA

Instruments, UK) and maintained in the range between 3.2 and 3.6 in the water column of each microcosm by adding H2SO4 when needed.

Dissolved oxygen in the water was measured every 7 days using an optical oxygen probe (Pyro Science, Germany) and dissolved organic 151 carbon (DOC) in the water was measured every 20 days using an

AutoAnalyser 3 HR (Seal Analytical, UK) as described in Chapter 3.

Concentration of Fe, Zn and Cu in the water column and sediments of each microcosm were analysed by ICP-AES as described in Chapter 2.

Dissolved and particulate metals in the water column were analysed every 5 days while total metals in the sediments were measured at the end of the experiment.

4.3.4 DNA extraction and shotgun metagenomic sequencing

At the end of the incubation period (day 88), total DNA was extracted from 0.5 g of sediments from each non-irradiated microcosm using the PowerSoil DNA Extraction kit (MoBio Laboratories, USA) according to the manufacturer’s instructions. DNA was not extracted from the irradiated sediments.

DNA libraries were prepared by the Leeds Institute of Molecular

Medicine Next Generation Sequencing Facility, University of Leeds,

UK, using a NEBNext Ultra DNA Library Prep Kit for Illumina and a

NEBNext Multiplex Oligos for Illumina kit (New England Biolabs, USA), both according to the manufacturer’s instructions. Paired–end sequencing of the whole metagenome of each library was performed using Illumina HiSeq3000 with 2 x 150 bp read length (Leeds

Institute of Molecular Medicine Next Generation Sequencing Facility).

Forward and reverse end reads from each site were paired using the

PEAR (version 0.9.8) merger software (Zhang et al., 2013). Paired

152 sequences were then trimmed using Trimmomatic (version 0.38)

(Bolger et al., 2014) for removing low-quality reads. FASTQ files from trimmed sequences were uploaded to the MG-RAST server (Meyer et al., 2008) for processing and annotation.

4.3.5 Bioinformatic analysis

Taxonomic assignment of 16S rRNA gene sequence reads from the metagenomic sequencing was performed using the Greengenes

Database v05, 2013 (DeSantis et al., 2006) for comparison with the taxonomy analysis carried out in Chapter 2. Microbial alpha diversity in each sample, Shannon–Weiner diversity index (H), Pielou’s evenness index (H/Hmax), and Chao1 species richness estimate values were calculated using the PAST software (version 3.20)

(Hammer et al., 2001). For evaluating the affiliation of taxonomic groups to specific metabolic activities, the RefSeq Release 89 (Pruitt et al., 2005) was used. The data was compared using a maximum e‐ value of 1e−5, a minimum identity of 97%, and a minimum alignment length cutoff of 15.

Functional classification of the metagenomic data was performed through the annotation pipeline from MG-RAST using the SEED subsystem database (Overbeek et al., 2013). The data was classified using a maximum e‐value of 1e−5, a minimum identity of 60%, and a minimum alignment length of 15 amino acids for proteins (default

153 settings). For the taxonomic and functional classifications, the best hit results were used for further analysis.

4.3.6 Statistical analysis

Non-metric multidimensional scaling (NMDS) plots of taxonomic and functional data was performed using the R vegan package v.2.4.2. Abundance of taxonomic and metabolic hits were normalized by the total number of copies of the specific rRNA or functional gene.

Statistical comparisons between different metabolic profiles was performed using one–way ANOVA (p < 0.05) and Bonferroni correction was applied for multiple comparisons using STAMP version

2.1.3 (Parks et al., 2014).

4.4 RESULTS

4.4.1 Changes in water chemistry during microcosm incubations

Concentrations of sulphate (~ 50 mg/L at day 1) showed a significant marked decrease during the first 10 days of incubation in all of the treatments (> 50%) from the NS microcosms (Figure 1a). The added carbon (AC) treatment showed significantly the largest decrease after

20 days of incubation (~ 2 mg/L). At day 10, sulphate concentrations in the OX treatment were significantly higher than the other treatments, but values were similar to the control over the subsequent days (Figure 1a). A less marked reduction of sulphate (~ 154

20%) was observed in the irradiated (IS) microcosms (Figure 1b).

The aerated IS microcosm with added carbon (COX) showed a significant reduction of sulphate within day 40 and day 60 compared to day 0 and to the other treatments. Sulphate levels in treatments

AC and COX were significantly lower than in control and OX treatments between day 30 and 40 and between day 65 and 75

(Figure 1b).

155

a Non - irradiated sediments

40 / L 4 -2

20 mg SO

0 0 20 40 60 80 Days

b Irradiated sediments

60 / L 4 -2 40

mg SO 20

0 0 20 40 60 80 Days

Figure 1. Changes in sulphate concentrations over the microcosm incubation period of 88 days from non-irradiated (a) and irradiated (b) sediments. Values show the mean of 5 replicates. Error bars show standard deviation.

156

A significant increase in sulphide levels from ~ 2 mg/L to ~ 5 mg/L was observed in the AC and COX treatments after 60 and 70 days, respectively, in the NS microcosms compared with OX treatment and the control and compared with values at day 0 (Figure 2a). In the IS microcosm, no significant changes in sulphide concentrations were observed (~0.5 mg/L) in the control and OX treatment or COX and

AC treatments (~1.5 mg/L) compared to day 0 (Figure 2b). However, sulphide values in the added carbon (AC) treatment were significantly higher than in the OX and COX treatment but not in the control, throughout all of the time points.

Initial values of dissolved S (~ 100 mg / L) decreased after the first

20 incubation days (~ 50 mg / L) in all the treatments of the NS microcosms (Figure 3a). At the end of the incubations, dissolved S values from all treatment except COX were significantly lower than at day 0. A less marked reduction in dissolved S values was observed in

IS microcosms (Figure 3b). COX treatment showed significantly lower values between days 45 and 55.

157

a Non - irradiated microscosms

4 / L 2-

mg S 2

0 0 20 40 60 80

Days

b Irradiated microcosms

2.5

2.0

/ L 1.5 2-

1.0 mg S

0.5

0.0 0 20 40 60 80 Days

Figure 2. Changes in sulphide concentrations over the microcosm incubation period of 88 days from non-irradiated (a) and irradiated (b) sediments. Values show the mean of 5 replicates. Error bars show standard deviation.

158

a Non - irradiated sediments

150

100

50 mg dissolved S / L Smg / dissolved

0 0 20 40 60 80 Days

b Irradiated sediments

150

100

50 mg dissolved S / L Smg / dissolved

0 0 20 40 60 80 Days

Figure 3. Changes in dissolved S concentrations over the microcosm incubation period of 88 days from non-irradiated (a) and irradiated (b) sediments. Values show the mean of 5 replicates. Error bars show standard deviation.

159

In the NS microcosm, AC and COX treatments showed a significant increase in particulate S after 65 and 74 incubation days (Figure 4a).

No significant variations were observed in particulate S concentrations in the control and OX treatments over the incubation period. Concentrations of particulate S were significantly higher in the

IS microcosm subject to AC and COX than the control and OX treatment after 50 days incubation (Figure 4b). However no significant change with time was observed when compared with the initial values at day 0.

In the NS microcosms, initial values of Fe+2 of around 15 mg/L decreased in the control and all treatments to below 10 mg/L from day 40 onwards (Figure 5a). AC and COX treatments showed significantly lower Fe+2 values after 60 days incubation. In the IS microcosm, only the OX treatment showed significantly lower concentrations of Fe+2 between day 20 and 60 although all treatments showed significantly lower levels compared to the control after 50 days of incubation (Figure 5b). Significantly higher values in

Fe+3 of above 6 mg/L were observed in all treatments compared with the control after 60 days compared to day 0 in the NS microcosms

(Figure 6a). Treatments AC and COX showed the highest values compared to OX and the control. In the IS microcosms, no significant changes in Fe+3 values in the control (~3 mg/L) were observed over along the incubation period (Figure 6b). Treatments OX, AC and COX were significant higher than the control after day 60.

160

a Non - irradiated sediments

5 mg particulate S / L S / mg particulate

0 0 20 40 60 80

Days

b Irradiated sediments

2 mg particulate S / L S / mg particulate

0 0 20 40 60 80

Days

Figure 4. Changes in particulate S concentrations over the microcosm incubation period of 88 days from non-irradiated (a) and irradiated (b) sediments. Values show the mean of 5 replicates. Error bars show standard deviation. 161

a Non - irradiated sediments

15 / L +2 10 mg Fe 5

0 0 20 40 60 80 Days

b Irradiated sediments

15 / L

-2 10 mg S 5

0 0 20 40 60 80 Days

Figure 5. Changes in Fe+2 concentrations over the microcosm incubation period of 88 days from non-irradiated (a) and irradiated (b) sediments. Values show the mean of 5 replicates. Error bars show standard deviation.

162

a Non - irradiated sediments

15 / L +3 10 mg Fe 5

0 0 20 40 60 80

Days

b Irradiated sediments

10 / L +3

5 mg Fe

0 0 20 40 60 80 Days

Figure 6. Changes in Fe+3 concentrations over the microcosm incubation period of 88 days from non-irradiated (a) and irradiated (b) sediments. Values show the mean of 5 replicates. Error bars show standard deviation.

163

In the NS microcosms, a significant decrease in dissolved Fe (10 – 15 mg/L) was observed in all the treatments compared to day 0 (2 – 10 mg/L) (Figure 7a). No significant difference during the incubation was observed in the IS microcosms (Figure 7b). At the end of the experiments, only control showed significant higher of dissolved Fe at the end of the experiments in the IS microcosms.

In the NS microcosms, a significant increase in particulate Fe (2 – 5 mg/L) was observed in all the treatments compared to day 0 (6 – 12 mg/L) (Figure 8a). Treatment AC and COX showed significant higher values at day 67 and 80. In the IS microcosms, OX treatment show higher values (~ 6 mg/L) between days 20 and 40 compared to the other treatments (> 4 mg/L) (Figure 8b). At the end of the experiments, the control showed significant lower levels of particulate

Fe at the end of the experiments.

The concentration of dissolved Zn showed a significant decrease in comparison to day 0 values in the AC and COX treatments containing non-irradiated sediment (NS) to below ~ 5 mg/L by the end of the incubation period (Figure 9a). No reduction was observed in the control and OX treatments. In the IS microcosm, a marked drop (~

50%) was observed in dissolved Zn in all treatments at day 10

(Figure 9b).

164

a Non - irradiated sediments

5 m g dissolved FeL /

0 0 20 40 60 80 Days

b Irradiated sediments

5 m g dissolved FeL /

0 0 20 40 60 80

Days

Figure 7. Changes in dissolved Fe concentrations over the microcosm incubation period of 88 days from non-irradiated (a) and irradiated (b) sediments. Values show the mean of 5 replicates. Error bars show standard deviation.

165

a Non - irradiated sediments

5 mg particulate Fe / L Fe / mg particulate

0 0 20 40 60 80

Days

b Irradiated sediments

5 mg particulate Fe / L Fe / mg particulate

0 0 20 40 60 80 Days

Figure 8. Changes in particulate Fe concentrations over the microcosm incubation period of 88 days from non-irradiated (a) and irradiated (b) sediments. Values show the mean of 5 replicates. Error bars show standard deviation.

166

a Non - irradiated sediments

5 mg dissolved Zn / L Zn / mg dissolved

0 0 20 40 60 80

Days

b Irradiated sediments

5 mg dissolved Zn / L Zn / mg dissolved

0 0 20 40 60 80 Days

Figure 9. Changes in dissolved Zn concentrations over the microcosm incubation period of 88 days from non-irradiated (a) and irradiated (b) sediments. Values show the mean of 5 replicates. Error bars show standard deviation.

167

Values of particulate Zn in COX treatment were significantly lower than the other treatments between day 43 and 56. In the NS microcosms, particulate levels of Zn showed a significant increase for all the treatments compared to day 0. COX treatment show higher values of particulate Zn but no significant difference was observed between treatments at the end of the experiment (Figure 10a). In the

IS microcosm, the control and the OX treatments show no significant difference in particulate Zn concentration along the incubation period.

Only the COX treatment show significantly higher values at day 44 and day 62 (Figure 10b).

In the NS microcosms, a significant drop of dissolved Cu values from day 0 (~ 3.8 mg/L) to the end of the incubation (> 2 mg/L) was observed (Figure 11a). Values in the control and OX treatment showed no significant difference at day 0 and at the end of the incubation. However, significant difference in the final values of dissolved Cu in the AC and COX treatments compared to the control and OX treatment was detected. In the IS microcosm, the AC treatment showed the largest reduction of dissolved Cu compared to day 0 but no significant difference between treatments was observed at the end of the experiment. In the NS microcosms, particulate Cu levels showed a significant increase in all treatments compared to day

0 but no significant difference was observed between treatments at the end of the incubation (Figure 11b).

168

a Non - irradiated sediments

5 mg particulate Zn / L / Zn mg particulate

0 0 20 40 60 80 Days

b Irradiated sediments

5 mg particulate Zn / L / Zn mg particulate

0 0 20 40 60 80 Days

Figure 10. Changes in particulate Zn concentrations over the microcosm incubation period of 88 days from non-irradiated (a) and irradiated (b) sediments. Values show the mean of 5 replicates. Error bars show standard deviation.

169

a Non - irradiated sediments

2 mg dissolved Cu / L Cu / mg dissolved

0 0 20 40 60 80

Days

b Irradiated sediments

1 mg dissolved Cu / L Cu / mg dissolved

0 0 20 40 60 80 Days

Figure 11. Changes in particulate Cu concentrations over the microcosm incubation period of 88 days from non-irradiated (a) and irradiated (b) sediments. Values show the mean of 5 replicates. Error bars show standard deviation.

170

a Non - irradiated sediments

1 mg particulate Cu L / mg particulate

0 0 20 40 60 80

Days

b Irradiated sediments

2.0

1.5

1.0

0.5 mg particulate Cu L / mg particulate

0.0 0 20 40 60 80 Days

Figure 12. Changes in particulate Cu concentrations over the microcosm incubation period of 88 days from non-irradiated (a) and irradiated (b) sediments. Values show the mean of 5 replicates. Error bars show standard deviation.

171 a

Figure 13. PCA plots showing the correlations between water parameter and the different treatments in non – irradiated (a) and irradiated (b) sediments.

172

In the NS microcosms, particulate levels of Cu showed a significant increase for all the treatments compared to day 0. COX and AC treatments showed higher values of particulate cu but no significant different was observed betwee n treatments at the end of the experiment (Figure 12a). In the IS microcosm, the control and the

OX treatment showed no significant difference of particulate Cu along the incubation period. Only the COX treatment showed significantly higher values during day 43 (Figure 12b).

The PCA plots (Figure 13) show some differences between irradiated and non – irradiated sediments. In the non – irradiated sediments,

Fe+2 correlated with OX while in the irradiated sediment it correlated with the control. In contrast, sulphate correlated with OX in the irradiated sediments while it correlated with the control in the non – irradiated experiments. Stronger correlations of particulate S and Zn with the treatments were carbon was added (AC and COX) was evidenced in the non – irradiated sediments. Moreover, sulphide production was correlated with the control and OX treatment in the irradiated sediments while in the non – irradiated sediments it correlated with the carbon amended treatments.

There were no significant differences in S and metal concentrations in the sediments between the start and finish of the experiment for both sets of sediments or between treatments for both sets of sediments

(Figure 14). A decrease in the concentration of dissolved organic carbon (DOC) over time was observed in the AC and COX treatments

173 in the NS microscosm although there was only a small change in the

AC treatments in the IS microcosms (Figure 15b). The control and the OX treatment in the NS microcosm showed lower values (~1 mg/L) compared to the IS microcosms (~ 5 mg/L). Levels of dissolved oxygen (DO) in the aerated treatments (COX and OX) were

~ 20% in the NS and IS microcosms while levels of DO in the AC treatment were lower in the NS (~3%) compared to the IS microcosms (> 6%) (Figure 15a).

174

Figure 14. Total sediment Fe, S, Zn, and Cu concentrations in microcosm experiments. Concentrations at day 0 of incubation period (a) and at the end of the incubation in the NS microcosm (b) and IS microcosm (c) are shown. Values displayed are the average of 5 replicates. Boxes show the 25th and 75th percentiles, the line within the boxes shows the median values. Whisker bars show the minimum and maximum values.

175

Non - irradiated sediments Irradiated sediments

25 25

20 20

15 15

10 10

5 5 % dissolved oxygen % dissolved % dissolved oxygen % dissolved

0 0 0 20 40 60 80 0 20 40 60 80 Days Days

Non - irradiated sediments Irradiated sediments

15 15

10 10

5 5 DOC(mg/g) DOC(mg/g)

0 0 0 20 40 60 80 0 20 40 60 80 Days Days

Figure 15. Levels of dissolved oxygen (DO) and dissolved organic carbon (DOC) during the incubation experiments. Average values of dissolved oxygen (a) and organic carbon (b) from 5 replicates from non-irradiated and irradiated microcosms are shown.

176

4.4.2 General characteristics of the sediment metagenomes

DNA isolated from non-irradiated sediment microcosms were used for whole genome sequencing. DNA was extracted from sediments from the three treatments (AC, COX and OX) and the control (C) without any amendment. After sequencing and subsequent quality trimming, a total of 12,465,283 – 16,722,052 sequences with an average length of 134 ± 55 - 166 ± 61 bp were obtained from each sample (Table

1). Between ~0.09% and 0.2% of sequences from the total number of trimmed sequences were identified as ribosomal RNA gene sequence reads. The vast majority of sequences (~ 92 – 96%) were associated with predicted proteins with known (30% - 49%) and unknown function (43% - 55%). These statistics considered eukaryotic, bacteria and archaea sequences. However, eukaryotic sequences were discarded in the taxonomic analysis.

4.4.3 Comparison of taxonomy profiles

In order to compare the bacterial and archaea community structure after the incubation period of the non-irradiated sediment with the composition of the natural wetland sediments analysed in Chapter 2, the Greengenes database was used as described in Chapter 2. The total number of OTUs varied between 15,699 (AC2) and 19,995

(AC3). The mean value of the Shannon-Weiner diversity index was calculated for each treatment (Table 2). Control sediments (C) gave the highest value (4.05) followed by COX (3.97), OX (3.89) and AC

177

(3.86). Sediments were taken from site S2 where the Shannon diversity was 6.56 (Chapter 2). This indicates a reduction of diversity during the incubation period in all treatment and in a less extent in the control group. The same pattern was observed for the Pielou’s evenness index. The initial value of 0.52 from site S2 was higher than the values of all the microcosm sediments (0.14 – 0.21). Chao1 species richness indicator was higher in sediments from the COX treatment (average of 383.4) and lower in the control group (average of 283.4). A NMDS analysis was performed based on species similarity. All of the replicates were grouped by treatment, with just one OX treatment sample (OX5) as an outlier, being closer to the

Control (Figure 15). Sediments with added carbon (AC and COX) formed a closer cluster than the Control and OX sediments. No difference in taxa clustering in relation to the different treatments was observed. At the phylum level, Proteobacteria was the dominant phylum in all the samples (Figure 16). Bateriodetes and Firmicutes were also a dominant phylum that varied according to the treatment.

Acidobacteria was reduced in all the COX treatments. Thermotoaga showed a marked reduction in all of the treatments when compared to the natural sediments. The dominant phylum Proteobacteria was further analysed at the species level (Figure 17). An uncultured delta

Proteobacteria was the dominant species. Diversity was maintained in all of the samples despite the treatment as confirmed with the alpha diversity calculations described above.

178

Table 1. Summary of metagenomic sequencing results obtained from sediments from the 3 treatments (AC, COX and OX) and the control sediments (C).

Sample Total Sequences Average rRNA Predicted Predicted Sequences after length genes proteins proteins quality (bp) with with control known unknown function function C1 15,187,692 12,998,516 148 ± 60 12,797 3,980,301 7,256,330

C2 16,944,116 14,573,813 134 ± 55 16,442 4,524,170 7,775,108

C3 16,394,422 14,168,476 155 ± 55 16,999 5,227,855 7,833,266

C4 15,980,467 13,778,468 146 ± 54 15,988 4,669,253 7,821,150

C5 14,581,044 12,721,827 146 ± 53 16,484 5,075,977 6,470,911

OX1 15,237,131 13,060,623 163 ± 58 26,283 6,374,637 5,886,754

OX2 16,166,051 13,720,778 161 ± 61 27,284 6,441,872 6,041,679

OX3 15,794,828 13,731,492 156 ± 57 21,774 5,872,757 6,774,206

OX4 17,892,234 15,028,810 159 ± 56 31,444 7,502,463 6,568,371

OX5 15,423,604 13,292,520 154 ± 53 16,937 5,279,888 7,171,279

AC1 16,066,804 13,918,086 160 ± 56 22,966 5,919,552 7,102,710

AC2 15,136,527 13,069,032 139 ± 56 17,520 4,424,701 6,832,826

AC3 19,809,799 16,722,052 146 ± 51 22,998 6,349,145 9,146,707

AC4 18,615,814 15,728,142 151 ± 54 20,269 5,959,192 8,525,177

AC5 15,042,444 13,057,009 165 ± 62 15,438 4,758,333 7,225,564

COX1 14,871,114 12,948,456 165 ± 60 28,889 5,578,418 6,481,229

COX2 15,344,307 13,162,812 166 ± 61 29,188 5,705,510 6,532,685

COX3 14,397,919 12,465,283 165 ± 59 27,614 5,377,912 6,299,825

COX4 15,897,450 13,805,134 157 ± 56 30,417 5,807,646 7,028,207

COX5 15,620,878 13,702,534 162 ± 58 29,960 5,856,207 6,885,504

179

Table 2. Summary of bacterial community diversity parameters for sediments from the 3 treatments (AC, COX and OX) and the control sediments (C). (a) Assigned taxa number indicates the number of distinct assigned taxa at genus level (excluding unassigned taxa) for each metagenome. (b) Total OTUs is the sum of all OTUs (97% identity) for each site.

Site Shannon- Peilou´s Chao1 Assigned Total Weiner evenness species taxa OTUs (b) diversity index richness number(a) index C1 3.83 0.14 316 4063 16,293 C2 4.11 0.26 233 3121 18,047 C3 4.38 0.28 277 2791 19,218 C4 3.79 0.16 278 3912 17,982 C5 4.15 0.20 313 4010 16,185 Mean 4.05 0.21 283.4 OX1 3.98 0.14 369 9971 19,658 OX2 3.82 0.19 244 3400 18,301 OX3 3.92 0.15 340 5717 17,533 OX4 3.71 0.12 345 8175 18,540 OX5 4.02 0.16 346 7880 18,735 Mean 3.89 0.15 328.8 AC1 3.59 0.12 299 6342 17,915 AC2 3.74 0.13 323 5785 15,699 AC3 4.33 0.23 332 4305 19,995 AC4 3.75 0.14 297 5022 18,599 AC5 3.85 0.14 333 7081 16,907 Mean 3.85 0.15 316.8 COX1 4.09 0.15 389 9440 17,723 COX2 4.14 0.16 389 9320 17,970 COX3 3.81 0.11 390 8897 17,538 COX4 3.94 0.13 383 9161 18,890 COX5 3.84 0.13 366 8422 18,516 Mean 3.96 0.14 383.4

180

Figure 15. Ordination analysis based on taxonomic profiles. Two- dimensional NMDS plot of sites based on species similarity. Five replicates for control sediments (C) and each treatment (AC, OX and COX) are displayed. Assigned taxa assemblage is also shown (grey open circles).

181

Figure 16. Relative abundance of bacteria and archaea phyla in the microcosm sediments. Five replicate from control sediment (C) and 3 treatments (oxygenated (OX), added carbon (AC) and oxygenated plus added carbon (COX)) are shown. Taxonomic assignment was performed using Greengenes v.13.8 database at 97% identity. All phyla levels are displayed.

182

Figure 17. Relative abundance of bacteria species of the Proteobacteria phylum in the microcosm sediments. Five replicates of control sediment (C) and 3 treatments (oxygenated (OX), added carbon(AC) and oxygenated plus carbon amended (COX) are shown. Taxonomic assignment was performed using Greengenes v.13.8 database at 97% identity. Labels of the 5 most abundant species are displayed.

183

4.4.4 Comparison of functional profiles

The SEED Subsystems is a hierarchical organization of functional genes. Each hiearchy is a group of genes retrieved from complete genomes exclusively which are known to share the same function

(Overbeek et al., 2013). The SEED subsystem annotation performed by MG-RAST showed 4 categories (Level 1, Level 2, Level 3 and

Function, which is the functional annotation of a specific sequence).

NMDS was performed for functional genes using the SEED

Subsystems database at the function level (lowest hierarchy) (Figure

17). Samples were also clustered by treatments the same as for the

16S rRNA gene analysis. Replicates from the COX treatment were clustered in a much more defined group. Comparisons of the different functional profiles at Level 1 (highest hierarchy) show similar relative abundance between treatments (Figure 19), with the exception of the

Motility and Chemotaxis functional group, where the control sediments show less abundance than the others and the Nitrogen metabolism group, where the Control and AC show slightly less abundance but higher than OX treatment (apart from OX5).

Because the simplified artificial AMD solution used for the incubation experiments only contained the metals Fe, Zn and Cu, the abundance of genes related to bacterial transportation of Fe, Zn and Cu was evaluated in order to understand the bioavailability of these metals between different treatments. ABC transporter diagrams related to

Fe, Zn and Cu were obtained from the KEEG Pathway Maps

(www.genome.jp). Genes related to Fe+2 transport (EfeB and EfeU), 184

Fe+3 hydroxamate ABC transporter, Zn ABC transporter and the

Zn/Cu/Mn ECF transporter were compared. The two genes coding the

Zn ABC transporter showed the highest variation according to the p value. Genes related to Cu showed no significant differences. A significant increase in the relative abundance of genes related to Fe+2 metabolism was only observed in AC and COX treatments when all were compared to the control (Figure 20).

Figure 18. Ordination analysis based on functional profiles. Two- dimensional NMDS plot of sites based on function similarity. Five replicates for control sediments (C) and each treatment (AC, OX and COX) are displayed.

185

Figure 19. Comparison of the relative abundance of the main functional categories of the SEED Subsystem database for each microcosm treatment.

Figure 20. Comparative functional profile of metabolic traits related to Fe and Zn metabolism between the different microcosm treatments. Only features that are significant different (p < 0.05) are displayed. Features are sorted by p value.

186

4.4.5 Sulphate metabolism analysis

Sulphur was the other element added in the simplified artificial AMD apart from Fe and trace metals. It is a main component of natural

AMD streams and can mediate metal transformation (Herlihy and

Mills, 1985). Therefore, S bacteria metabolism was also analysed.

The S metabolism diagram obtained from the KEEG Pathway Maps

(www.genome.jp) was used for this analysis (Figure 21). A total of 24 enzymes obtained from the metagenomic analysis belonged to 8 metabolic pathways of the sulphur metabolism diagram. Table 3 shows the list of 24 genes coding enzymes from key reactions used in the analysis of sulphur metabolism.

187

Table 3. List of metabolic pathways and enzymes used in the analysis of bacterial S metabolism.

Metabolic pathway Enzyme Sulphate adenyltransferase • Sulphate adenylyltransferase subunit 1 • Sulphate adenylyltransferase subunit 2 • Sulphate adenylyltransferase, dissimilatory-type Adenylylsulphate reductase • Adenylylsulphate reductase alpha-subunit • Adenylylsulphate reductase beta-subunit Dissimilatory sulphite reductase • Dissimilatory sulphite reductase (desulfoviridin) Sulphite reductase • Sulphite reductase, dissimilatory-type gamma subunit • Putative sulphite reductase, gamma subunit • Sulphite reductase [NADPH] flavoprotein alpha-component • Sulphite reductase [NADPH] hemoprotein beta-component • Sulphite reductase alpha subunit • Sulphite reductase beta subunit Ferredoxin – sulphite reductase • Ferredoxin--sulphite reductase • Ferredoxin--sulphite reductase, actinobacterial type • Ferredoxin--sulphite reductase, bacillial type Hydrogenase / sulphur reductase • Hydrogenase/sulphur reductase, alpha subunit • Hydrogenase/sulphur reductase, beta subunit • Hydrogenase/sulphur reductase, delta subunit • Hydrogenase/sulphur reductase, gamma subunit Thiosulphate reductase • Thiosulphate reductase cytochrome B subunit • Thiosulphate reductase electron transport protein phsB Tetrathionate reductase • Tetrathionate reductase subunit A • Tetrathionate reductase subunit B • Tetrathionate reductase subunit C 188

Figure 21. Metabolic pathways obtained from metagenomic analysis and present in the S transformation map. (1) Sulphate adenyltransferase, (2) Adenylylsulphate reductase, (3) Dissimilatory sulphite reductase, (4) Sulphite reductase, (5) Ferredoxin – sulphite reductase, (6) Hydrogenase / sulphur reductase, (7) Thiosulphate reductase, (8) Tetrathionate reductase. Map modified from Kanehisa Laboratories (www.kegg.jp).

189

Relative abundance values of the selected genes shown in Table 3 from the control sediments were compared against the other treatments.

Differences were ranked according to the p-value obtained from the

ANOVA. Comparison between the Control and OX sediments showed that 10 genes were significantly different (Figure 22a). The genes coding for the sulphite reductase alpha subunit and the sulphate adenylyltransferase subunit 2 showed the largest differences in abundance. A total of 19 genes were significantly different between N and C sediments (Figure 22b). The gene coding for the sulphate adenylyltransferase dissimilatory-type enzyme had the lowest p value and was significantly higher in the Control sediments. Genes related to sulphite reduction were significantly higher in the AC treatment.

Comparisons between the Control and COX sediments showed 15 genes with significant differences (Figure 22c). The gene coding for the

Sulphate adenylyltransferase subunit 1 showed the lowest p-value. The abundance of 7 genes were greater in the COX sediments, including genes related to sulphite and tetrathionate reduction.

Multiple comparisons of all the genes were performed. Results showed that thiosulphate reduction was the reaction with the highest variation followed by a group of genes related to adenlylysuplhate reduction which were more abundant in the Control and OX sediments. A second group of genes related to sulphite reduction (including dissimilatory type) with significant variations between treatments were more

190 abundant in the AC and COX sediments (Figure 23). The RefSeq database was used to analyse the taxonomic affiliations of the genes coding for the top 5 enzymes with the lowest corrected p values. The genes analysed were those coding for the following enzymes:

Thiosulphate reductase electron transport protein, Ferredoxin-sulfite reductase, Adenylylsulphate reductase alpha-subunit, Sulphate adenylyltransferase, dissimilatory-type and Dissimilatory sulfite reductase (desulfoviridin). The first gene was only detected in the genera Salmonella and Wolinella in all samples (Figure 24a). The gene coding for the Ferredoxin-sulfite reductase that was more abundant in the AC and COX sediments, showed a dominance of the Geobater genera (Figure 24b). The gene coding the Adenylylsulphate reductase alpha-subunit (more abundant in the Control and OX sediments) only showed a dominance in Thiobacillus in the OX sediments (Figure 24c).

In the case of the sulphur dissimilatory reations, the gene coding for the sulphate adenylyltransferase also had a dominant taxon (Thioalkavibrio) in the OX sediments (Figure 24d) and the gene coding for the dissimilatory sulfite reductase (desulfoviridin) showed dominance in

Candidatus koribacter in the Control and OX sediments and in the genera Geobacter (in particular in the species Geobacter uraniireducens) in the AC and OX sediments (Figure 24e).

191 a

192

Figure 22. Comparative functional profile of metabolic traits related to sulphur metabolism. Comparison of sulphur metabolism genes from control sediments (C) against OX (a), AC (b) and COX (c) sediments. Only features that are significant different (p < 0.05) are displayed. Features are sorted by corrected p values.

193

Figure 23. Multiple comparison of the sulphur metabolism genes between different sediment types. Genes were ranked by the corrected p values (Bonferroni test for multiple comparisons). The colours show the proportion (abundance value divided by the total abundance in all the sediment types) for each gene.

194 a

195

196

Figure 24. Taxonomic affiliation of the 5 key enzymes coding genes for sulphur metabolism with the lowest corrected p-values between treatments. Relative abundances of taxa from (a) Thiosulphate reductase electron transport protein phsB (b) Ferredoxin--sulfite reductase (c) Adenylylsulphate reductase alpha-subunit (d) Sulphate adenylyltransferase, dissimilatory-type and (e) Dissimilatory sulfite reductase (desulfoviridin), alpha and beta subunits is shown.

197

4.5 DISCUSSION

4.5.1 Changes in S and metal chemistry

No marked differences between microcosm treatments with non- irradiated sediments (NS microcosms) were observed in the levels of sulphate and dissolved S over the incubation periods of 88 days. In contrast, levels of sulphide and particulate S were significantly higher in the added carbon treatments (AC and COX) compared to OX and the

Control, respectively. Previous studies have used organic substrates to enhance sulphate reduction in laboratory (Drury, 1999), mesocosms

(Koschorreck et al., 2007) and constructed wetlands (Vile and Wieder,

1993). It is known that bacterial dissimilatory sulphate reduction uses sulphate as an electron acceptor for oxidation of organic carbon (Berner,

1984). Moreover, organic carbon has been described as an essential energy source for sulphate reducing microorganisms in AMD environments (Cook and Kelly, 1989) where supply of degradable carbon compounds is limited (Peine and Peiffer, 1996).

These results indicate that addition of organic matter enhances the remediation process due to increased production of sulphide capable of binding to metals and more particulate S that facilitates precipitation of metals from the water column. Moreover, precipitation of Zn, Cu and Pb has been achieved by the degradation of VFA and subsequent production of biogenic sulphide in bench-scale systems (Alvarez et al.,

2007). Microcosms containing irradiated sediments (IS microcosms)

198 show a similar pattern but with concentrations of sulphide (> 2 mg/L) lower than in the non-irradiated sediments where sulphide concentrations range between 4 and 6 mg/L. It is therefore likely that bacterial metabolism enhances production of sulphide as part of the energy production mechanism.

Increased remediation capacity is also indicated by the relative amounts of Fe+2 and Fe+3 in the NS microcosms. The AC and COX treatments showed the largest reduction in Fe+2 and higher production of Fe+3, which suggests an increased conversion rate in the presence of organic matter. Previous studies has observed that organic matter increases the amount of soluble Fe+2 by promoting the formation of nano-sized Fe oxy-hydroxides, which are more susceptible to microbial Fe reduction

(Pédrot et al., 2011). Dissimilatory Fe reducing bacteria are known to facilitate this solublization process (Zachara et al., 2001), which will also release adsorbed metals to the water column. However, biological Fe oxidation is a more complex mechanism. For example, addition of organic matter can enhance bacterial nitrate-dependent Fe oxidation

(Klueglein and Kappler, 2013) mediated by heterotrophic Actinobacteria

(Kanaparthi et al., 2013) while fulvic acid can act as a Fe oxidation catalyzer following oxygen consumption (Voelker and Sulzberger, 1996).

In the IS microcosms, oxygenation (OX) generated sporadic increase of

Fe +3 during and reached higher values than the C treatment in the NS microcosm. This suggests that chemical and biological Fe+2 oxidation are 199 important mechanisms in the remediation process and their efficiency will depend on the oxygen levels and bacterial activity in the system. A previous study has shown that biological Fe+2 oxidation enhances Fe redox cycling at the aerobic/anaerobic interface at circumneutral pH

(Sobolev and Roden, 2001). Moreover, some species of Gallionella and

Leptotrix have been identified as being responsible for this process

(Emerson and Weiss, 2004, Emerson and Moyer, 1997). The increased levels of Fe+3 in the aerated IS microcosm is not expected considering that Fe+2 oxidation is mainly mediated by bacteria in acid environments

(Weber et al., 2006). Other factors such as initial organic matter content from the sampled sediment should be considered to elucidate these changes.

Levels of Zn and Cu showed expected behavior in relation to the observed transformation of Fe and S. Particulate levels of Zn and Cu correlate with greater Fe+2 oxidation and sulphide production rates. The

AC and COX treatments show greater conversion of dissolved Zn and Cu to their particulate forms in both the NS and IS microcosms, which suggests that organic matter binds and aggregates to these metals.

Colloids and other organic/metals complexes formed by organic matter and Fe species have been detected in other environments (Baalousha,

2009, Sharma et al., 2010). No difference in sediment levels of S and metals was observed despite the generation of particulate compounds as the initial concentration of ions in the water column was insufficient

200 to produce changes in the sediments. For example, if all of the Fe were deposited onto the sediments, the initial concentration will only increase by 0.2 mg/L. Lower levels of DO and DOC in the NS microcosms compared to the irradiated sediments indicate significant consumption of oxygen and organic matter by bacteria and confirm the efficacy of irradiation for inactivating microbial metabolism.

4.5.2 Differences in bacterial taxonomy and metabolism

Diversity was maintained in the bacterial communities within the microcosm sediments, although every treatment clustered separately in the taxonomy ordination analysis. However, differences were not seen at the phylum level. Proteobacteria, Actinobacteria and Acidobacteria were the dominant phyla as seen in other AMD environments (Aguinaga et al., 2018). Relative abundance of Thermotogae decreased in all treatments with respect to the control. The phylum Thermotogae comprise species that have been previously detected in AMD impacted environments (Wang et al., 2016) and differential tolerance to oxygen has been described (Tosatto et al., 2008). Therefore, changes in the concentration of oxygen in the microcosms by the addition of air and organic matter may have influenced the abundance of this phylum in all the treatments. The Greengenes database was used to analyse the taxa composition of the samples in order to make comparison with the taxonomic data obtained from the original natural wetland sediments as 201 described in Chapter 2. Data presented in Chapter 2 indicates that phylum composition was not affected by AMD and the wetland protected species diversity despite AMD pollution. Therefore, the conserved diversity in all microcosm treatments was expected as the sediments were obtained from the AMD-impacted wetland.

When the phylum Proteobacteria was analysed in detail, the genera

Geobacter was again observed as a dominant species and the diversity pattern observed was similar to those observed in the AMD impacted wetland sediments examined in Chapter 2. Higher abundance of

Geobacter was observed in the microcosms with amended VFA. Some

Geobacter species utilize VFA for the oxidation of Fe+3 compounds

(Coates et al., 2001). Performance of Geobacter–mediated microbial fuel cells can also be modulated by addition VFA (Yang et al., 2015). These studies suggest that release of organic acids by wetlands can promote the proliferation of the genera Geobacter. However, the Greengenes database was not efficient for the identification the other dominant species and was not capable of accounting for the taxonomic differences that explain the NMDS clustering. For this, the RefSeq database was used and taxonomic affiliation of different metabolic traits was performed and is described later in this discussion.

In Chapter 2, the prediction of taxonomy-derived metabolisms using

PAPRICA indicated differences in the wetland sites compared to the river samples despite high levels of AMD. Ordination analysis of all of the 202 metabolic traits at the functional level of the SEED subsystems database in this Chapter shows that the COX treatment that was intended to mimic the effect of the wetland by artificially increasing oxygen and organic matter levels formed an isolated group. Therefore, this supports the metabolic predictions observed in Chapter 2.

Overall metabolic differences were evaluated by the analysis of the primary metabolic and functional categories. Differences were observed in the “Mobility and Chemotaxis” and “Nitrogen metabolism” categories.

A more in detailed analysis was performed by evaluating genes related to Fe, Zn, Cu and S metabolism, because those were the elements in the simplified artificial AMD solution.

The genes coding the Zn ABC transporter showed greater variation when

Fe/Zn/Cu metabolic traits were evaluated. The OX and AC treatment show higher abundance of Zn ABC transporter genes but the COX treatment showed lower abundance of these genes when compared to N treatment. This suggests a greater Zn bioavailability in the N treatment due to higher concentration of dissolved Zn in the water column compared to the control microcosm. The presence of Zn ABC transporters has been previously detected in taxa isolated from AMD environments (Mardanov et al., 2016) and changes in Zn levels modulating the transcription of Zn ABC transporter was been previously described in rhizosphere environments (Lim et al., 2013). Only genes related to Fe+2 transport into the cell were significantly higher in the 203 control than in the treatments with added carbon. This suggests that organic matter plays a role in reducing the bioavailability of Fe+2 to the microorganisms.

Differences in the sulphur metabolism genes were investigated between treatments. First, treatments AC, OX and COX were compared with the no treatment group (N). From the 24 selected genes, OX had the lowest number of significantly different genes (10) when compared with N. It is known that anaerobic bacteria can continue to reduce sulphur compounds to obtain energy (Dolla et al., 2006) in the presence of oxygen, and that they can thrive in anaerobic microhabitats within oxygenated environments (Perry, 1995). Therefore, less marked variation in metabolism between oxygenated and non-oxygenated treatment is expected.

Significant differences between genes related to dissimilatory sulphate reduction were only detected when the AC and COX treatments were compared against N. In both treatments, the abundance of these genes was higher than in N. This suggests a potential role for organic matter in dissimilatory sulphate reduction (DSR) activity. Previous work has demonstrated the role of VFA such as acetate in modulating DSR activity in aquifers (Chapelle and Lovley, 1992) and VFA have been used in reactors to enhance sulphate reduction activity (Liu et al., 2015). DSR is important in remediation processes as bacteria are capable of rapidly changing the pH and enhancing metal precipitation through generation 204 of reduced sulphur compounds in constructed wetlands (Vile and

Wieder, 1993, Mitsch and Wise, 1998).

After single comparisons against the Control, multiple comparisons between all treatments were performed and the taxa affiliated to the top

5 traits were analysed. The enzyme thiosulphate reductase was the trait with highest variation between treatments. Thiosulphate is a reduced S compound previously detected in AMD environments involved in the bacterial energy production though reduction of S (Schippers and

Sand, 1999). Only the two genera Salmonella and Wolinella were affiliated to the gene coding this enzyme, with no marked different in the taxa distribution between treatments. The second enzyme with higher variation was Ferredoxin-dependent sulphite reductase. This enzyme is capable of the reduction of sulphite described mainly in plants

(von Arb and Brunold, 1985). In this study, the genera

Anaeromyxobacter (in the Control and OX treatment) and Geobacter (in

AC and COX) had the higher number of copies of this gene. Previous studies have detected Anaeromyxobacter in AMD impacted environments (Sun et al., 2015a) and the metal-reducing capacity of members of this phylum have been reported (Treude et al., 2003).

Sulphite reductase activity has been previously observed in anaerobic respiration of Geobacter (Chin et al., 2004). This difference in taxa dominance between the amended VFA treatments (AC and COX) and the non-amended VFA treatments was also observed in the case of the

205 dissimilatory sulphite reductase (desulfoviridin) enzyme. This membrane-bound enzyme was first reported in Desulfovibrio desulfuricans as part of the DSR mechanism (Steuber and Kroneck,

1998). The expression of desulfovirin has also been observed in AMD environments under controlled lab conditions (Alazard et al., 2010). In this case, Geobacter was again the dominant taxon affiliated with this enzyme in the AC and COX treatment and Candidatus koribacter

(phylum Acidobacteria) was the dominant taxon in the OX and the

Control. Candidatus koribacter has been previously reported as dominant species in unpolluted soils (Žifčáková et al., 2016). Reduced abundance of this taxon was observed in mining impacted environments

(Rojas et al., 2016) but DSR activity of this taxon has not been reported before. The other 2 metabolic traits from the top 5 were adenylylsulphate reductase and dissimilatory sulphate adenylyltransferase. In both cases, marked dominance of a specific taxa was only observed in the OX treatment. Both enzymes are related to conversion of sulphate to adenosine 5’-phosphosulphate as part of the

S intake in bacteria (Muyzer and Stams, 2008). The dominant taxa were

Thiobacillus for the adenylylsulphate reductase and Thioalkalivibrio for the dissimilatory sulphate adenylyltransferase Thiobacillus members such as T. denitrificans (Beller et al., 2006) and T. thioparus (Lyric and

Suzuki, 1970) are bacteria with previously reported adenylylsulphate

206 reductase activity. Thioalkalivibrio has been observed in mining leachates (Borkowski, 2013).

4.6 CONCLUDING REMARKS

This study aimed to evaluate the effect of oxygen and organic matter resulting from the presence of plants on the remediation of AMD in wetland sediment. The supply of these components is not always constant and depends on many variables such as plant decay rate and seasonal variations in light regime and temperature. In this study, artificial addition of organic matter in the form of VFA and oxygen to the system was carried out in order to minimize the number of variables and thereby allow changes in water chemistry and microbial communities to be evaluated.

Water chemistry shows a marked difference between treatments with and without addition of organic matter. Increased remediation efficiency in the form of higher levels of particulate Fe, Zn and Cu was observed in the organic matter treatment. The addition of VFA maintained DOC values of between 5 and 10 mg/L compared to >5 mg/L in treatments where no VFA was added. This difference in organic matter content may play an important role in the remediation efficiency of the system. At the microbial level, increased DOC resulted in changes in the bacterial composition in the sediments. Changes in the bacterial community is

207 reflected in the greater abundance of genes related to dissimilatory sulphate reduction in the VFA treatments that in turn were correlated with higher sulphide levels observed in the water column. The genera

Geobacter was the dominant taxon where these genes were detected and is a key taxon responsible for metal removal though reduction of sulphur compounds. Oxygenated and non-oxygenated treatments show no marked difference in the bacterial community, indicating that organic matter is a greater driver of remediation efficiency than dissolved oxygen. Variations detected in the abundance of specific genes should be complemented with expression analysis though transcriptomic studies in order to confirm the changes in bacteria activity proposed in this study. Sulphide generation and sulphate depletion was analysed in detail in this study due to the importance of sulphur metabolism in AMD remediation. However, other mechanisms that facilitate the deposition and fixation of trace metals under acid conditions should be addressed for an overall understanding of mine waste remediation, including the role of refractile organic matter such a humic substances in the immobilization of trace metals (Potysz et al., 2017).

Previous studies have characterized the use of different organic substrates for passive AMD remediation (Gibert et al., 2004) and oxygen release (Armstrong et al., 2000) and other root exudates in the form or dissolved organic compounds (Bais et al., 2006) have been measured in wetland systems. In this experiment changes in the water chemistry and

208 bacterial activity though artificial oxygenation and addition of organic matter was achieved. These changes allowed the quantification of metal attenuation efficiency. However, experiments using vegetated sediments are needed in order to understand the effect of other variables involved.

209

Chapter 5: GENERAL DISCUSSION AND CONCLUSIONS

5.1 SUMMARY AND KEY FINDINGS

The Southern (SAG) and Northern (NAG) Afon Goch rivers receive AMD effluent from Parys Mountain. The AMD that is received by the SAG comes from old drainage ponds below Parys Mountain that in turn receives leachate from the old spoil heaps. At 2.2 km from the source, the SAG enters a natural wetland. In 2003, AMD discharges entering the

NAG were significantly increased while the adit that also discharged into the SAG ceased to operate due to the lowering of the water table.

Previous work has detected effective metal remediation along the SAG wetland (Dean et al., 2013) while no signs of quality improvement was observed in the NAG.

In Chapter 2, differences in metal chemistry and microbiology that were detected between the SAG and NAG are described. A significant increase in particulate metals was detected in the middle of the SAG wetland when compared both with other sites along the wetland and with the

NAG. Concentrations of dissolved metal along the SAG wetland were similar to an unpolluted section upstream the AMD source of the NAG.

However, the polluted section of the NAG showed significantly higher concentrations of dissolved metals. Analysis of 16S rRNA sequences within sediment samples showed that the wetland maintains bacterial

210 diversity. Despite the levels of AMD pollution, this diversity was similar to that observed in sediment from an unpolluted wetland nearby. In contrast, sediments from the polluted section of the NAG showed diversity loss and dominance of specific taxa. Prediction of taxonomic– derived metabolic traits showed that the SAG wetland maintains the relative abundance of overall metabolic traits and promotes specific biochemical pathways related to AMD remediation and metal cycling.

The NAG showed a decrease in the relative abundance of these predicted metabolic traits. Analysis of predicted abundance of specific enzymes also showed different profiles depending whether the AMD flows though the SAG wetland or the NAG.

In order to further understand the mechanisms related to metal attenuation in the SAG wetland, a more detailed analysis of metal distribution and bacterial activity in sediments and the water column along the SAG wetland was performed, as described in Chapter 3. In the water column, a significant decrease of dissolved S in the form of sulphate was observed along the wetland. Levels of Fe+3 were always higher than Fe+2; however the Fe+3/Fe+2 ratio decreased along the stream while particulate Fe levels increased and dissolved Fe decreased in the middle of the wetland, suggesting deposition of Fe+3 onto the sediment. A significant decrease in dissolved concentrations of Zn, Mn and Al in the upper reaches of the wetland close to the AMD source was observed when compared with the middle of the wetland. In the

211 sediments, higher concentrations of trace metals were detected in deeper layers (~25 cm depth) while Fe concentrations were higher in surface sediments. Distribution of S, Zn and Cu showed co-localization of these elements along sediment. Higher concentration of S, Zn and Cu were detected in deeper layers as the distance from the AMD source increased. The metal attenuation observed in the water column coincided with changes in bacterial activity within surface sediments from the same location. Significantly higher expression of the bacterial soxB gene that encodes periplasmic oxidation of thiosulphate (a reduced

S compound) and significant increase in expression of the 16S rRNA gene of F. myxofaciens (an acidophilic obligate iron-oxidizing bacteria) were observed.

The middle of the wetland also showed higher abundance and diversity of vegetation. Moreover, experiments in Chapter 2 indicate that composition of Fe and S oxidizing/reducing bacteria communities varied depending on the surrounding plant species. Further analysis was attempted in order to understand the role of plants in the environmental and microbial changes previously described. However, due to the large number of variables involved in plant physiology and their interaction with sediments, wetland plants from the field were unable to survive within controlled mesocosm environments, therefore a lab–based experiment was designed as outlined in Chapter 4 using surface sediments from the middle of the wetland. Addition of free volatile acids

212 and oxygen as an attempt to mimic plant exudates into microcosms containing natural and irradiated sediments over 88 days was performed and metal chemistry and bacterial activity was evaluated. Natural sediments showed higher reduction of sulphate, increased concentration of sulphide and higher concentrations of particulate metals when compared with irradiated sediments where bacteria were removed.

Within the natural sediments, those with added carbon showed greater differences in these variables when compared with oxygenated and control sediments. Genes coding for proteins from the Zn ABC transporter and Fe+2 transport peroxidase were significantly reduced in oxygenated sediments with added carbon. Genes coding for key enzymes related to S metabolism such as dissimilatory sulphate reductase, and thiosulphate reductase were significantly higher in natural sediments with added carbon. Addition of carbon also increased the relative abundance of Geobacter sp., which has increased abundance of the S metabolism genes previously evaluated.

The following conceptual model (Figure 5.1) illustrate the main AMD remediation mechanisms that were identified and characterized in this thesis:

213

214

Figure 5.1. Conceptual model that illustrate all the main AMd remediation processes elucidated in this study.

215

5.2 RESEARCH IMPLICATIONS AND RELEVANCE

Studies previous to this work have analysed the processes involved in trace metals attenuation such as precipitation of metal sulphides (Gazea et al., 1996) and alkalinity – mediated metal precipitation (Scholz and

Lee, 2005) in constructed wetland. Other studies have described natural metal attenuation mechanisms such as decrease in metal solubility (Vymazal et al., 2007), generation of particulate metals (Asta et al., 2010) and long – term deposition of metals in natural wetlands

(August et al., 2002, Dean et al., 2013). However, little is known about the relative importance of each process in metal removal and the interaction between each process that can affect remediation efficiency.

In this study, the relationship between bacteria diversity, levels of AMD pollution and presence of wetland has shown first insights into the interaction between biotic and abiotic factors. In addition, laboratory controlled experiments performed in this work were able to quantify the importance of specific processes related to bacteria activity and presence of organic matter.

In an attempt to understand the relationship between the chemistry and the microbiology of the Afon Goch wetland–river system, Walton and

Johnson (1992) evaluated the iron oxidizing and heterotrophic bacteria at regular intervals along the metal attenuation gradient in the SAG wetland and concluded that abundance of iron oxidizing bacteria decreased with distance while abundance of acidophilic heterotrophic

216 bacteria increased. In the NAG adit, Coupland and Johnson (2004) showed a dominance of A. ferrooxidans (~45%) and presence of other acidophilic species such as Ferrimicrobium acidiphilum and Gallionella– like organisms in water samples. After a continuous monitoring of bacteria abundance over 9 years, a later study concluded that diversity increased with time even though no improvement in the water quality was observed (Kay et al., 2013). This thesis confirms with respect to the above – mentioned Afon Goch microbiology studies, the presence of

Gallionella–related organisms which constitute nearly 17% of the total bacteria in the NAG but reveals that Methylophilales is the dominant

(45% of total OTUs) taxa downstream the adit. Species from the order

Methylophilales have proven broad adaptation to mine water environments and high capacity to used different substrate as energy source (Drewniak et al., 2016, Lanclos et al., 2016). Therefore, this taxa is an interesting candidate for the potential detection of new microbial – mediate metal transformation mechanisms. In addition, the marked differences in Acidithiobacillus abundance (up to 6.1% in sediments detected in this thesis) between water and sediments samples imply that difference in bacteria composition depending on the specific microhabitat within the NAG should also be considered in future studies.

An important component of this study was a comparison of the bacteria taxonomic composition in the SAG wetland and the NAG that revealed the central role of wetlands in maintaining bacteria diversity in the

217 presence of AMD pollution. To the best of the author´s knowledge, this is the first 16S rRNA analysis comparing a natural wetland and river system impacted by AMD. Characterization of microbial communities in constructed wetlands have been previously attempted in order to understand performance issues. For example, increased abundance of

SOB compared to SRB has been associated with inefficient sulphate removal in passive AMD treatments that include artificial wetland technology (Valkanas and Trun, 2018) and dominance of Firmicutes and

Chloroflexi taxa have been reported in poorly constructed wetland cells lacking sulphate reduction activity (Miller et al., 2018). Therefore, the new knowledge obtained in this study regarding detailed taxonomic composition and changes along the SAG natural wetland, which is a proven efficient system, can contribute to improving the performance of constructed wetland technologies.

Evidence of sulphate reduction in the Afon Goch was first obtained when metal sulphides were detected below sediment surface nearby the Dulas

Bay estuary (Parkman et al., 1996). Later, 16S rRNA sequences from microorganisms phylogeneticaly related to known SRB were reported within surface sediments of the SAG wetland (Dean et al., 2013). This thesis produced the first evidence for the co-mobilization of S compounds and trace metals in sediments from the SAG wetland as a similar pattern in the distribution profiles of S, Zn and Cu with sediment depth (~ first 50 cm depth) was detected using ITRAX technology. To

218 the best of the author´s knowledge, this is the first time ITRAX analysis was used to assess wetland metal attenuation efficiency. Moreover, incubation experiments of SAG wetland sediments performed in this study, detected the presence of bacterial genes related to dissimilatory sulphate reduction and its association with an increased generation of sulphide (~2 fold), particulate Zn (~3 fold) and particulate Cu (~4 fold) under controlled laboratory conditions.

Further attempts to understand the metal attenuation process observed in the SAG wetland have demonstrated how the system can remove metals through precipitation of ferryhidrate due to changes in pH along the river (Boult et al., 1994), and that metal deposition can be influenced by sediment biofilms as suggested by transmission electron microscopy (Boult, 1996). This thesis suggests that apart from pH, Fe+3 precipitation can be influenced by obligate iron-oxidizing bacteria since a marked expression of 16S rRNA genes of Ferrovum and related bacteria were observed in wetland sites where more particulate Fe was detected.

Moreover, this study indicates that the genera Ferrovum, Geotrix and

Gallionella are the dominant taxa inhabiting the biofilm as confirmed by taxonomic affiliation of OTUs from surface sediments. Metal transformation in aquatic environments have been previously described and well–studied in these bacteria (Nevin and Lovley, 2002, Weber et al., 2006, Johnson et al., 2013); therefore the detection of these microorganisms in the water/sediment interface facilitates future studies

219 aiming to further understand the wetland remediation mechanism. This work also reveals that the relative abundance of the above-mentioned taxa changes according to the plant species colonizing the sediments.

Several studies have evaluated the use of different wetland plant species for phytoremediation, including uptake and bioaccumulation of metals in AMD environments (Wieder et al., 1990, Stoltz and Greger,

2002, Karathanasis and Johnson, 2003, Liu et al., 2007). This study suggests that the role of different plant species should also be evaluated in terms of their ability to promote the growth of a diverse number of bacteria taxa with different metal cycling mechanisms, and confirms that the role of plants in AMD remediation is related to their influence on the microbial assemblage rather than direct metal removal as previously assumed in mesocosm wetland experiments (Collins et al., 2004).

Attempts to replicate wetland conditions using living plants isolated from the SAG wetland failed during this study due to plant death and germination inefficiency. Therefore, addressing some of the research questions remains a challenge such as if enhanced metal attenuation in incubated sediments due to artificially added organic matter is comparable with a potential release of organic compounds from vegetation in the wetland, or if plant age and decomposition rate are important drivers of the remediation process. Answering these questions could also help to understand the effectiveness of the remediation as most removal of trace metals such as Cu and Zn and increase of pH

220 occurs over a relatively short distance despite the large amounts of acidity and metals entering the system.

This study has revealed that natural wetlands contain microbial communities adapted to AMD and hence include metabolic traits that can oxidize or reduce trace metals in the water column and immobilize them in the substrate. The constant supply of organic matter from wetland plants provides a suitable environment for the microbial communities to operate over the long term. High resolution metal analysis in the SAG showed that metals are not only removed from the water column but also are retained in the deeper sediment layers, therefore generating uncontaminated water and sediments at the end of the wetland. Hence, natural wetlands can be seen as an effective approach to the remediation of metal-contaminated water and sediments over the long term.

Results from this study suggest that engineered wetlands should aim to enhance microbial diversity through provision of organic material, including direct addition during the construction phase as there will be a delay in organic matter release from the newly planted vegetation.

Analysis of metal distribution in water and sediments in artificial wetlands should be performed in parallel with microbial analysis in order to quantify efficiency. In addition, high resolution metal profiles in the sediments should be carried out to provide information regarding metal interactions and immobilization. 221

With all these considerations, the goal of significantly increasing the efficiency of engineered wetlands becomes more realistic. However, many challenges remain, including decreasing the acidity of mine water to near-neutral values between pH 5 – 7. Some engineered wetlands have accomplished this increase through various improvements such as by adding organic matter (such as compost mixtures, peat and sawdust)

(Kelman Wieder, 1990, Gibert et al., 2005) and alkali - generating minerals (Kalin et al., 2006). The SAG wetland was able to increase pH from 2 to 5 in the first 100 metres, which demonstrates the high efficiency of the system over a short distance and hence reduces the need to artificially increase the neutralization capacity. High dissolved metal loads entering the system will need to be effectively immobilized in order to reduce metal levels in the water below environmental standards for engineered wetland to be considered efficient. The SAG was able to remove dissolved Fe from 55% (before the 2003 diversion of much of the contaminated run-off) to 97% (after lowering of the water) (Dean et al., 2013)and this study), and ITRAX scans demonstrated that more than 60% of precipitated Fe was retained in the first 150 m of wetland sediment. Therefore, artificial wetlands should not only remove dissolved metals from AMD but should provide efficient retention into the substrate and avoid the risk of metals being released again into the water. Land use availability is also an important limitation for artificial wetlands technology. For example, according to calculations

222 based on pH and Fe loads from previous wetlands in the U.S.

(Kleinmann, 1990), an AMD source such as the NAG which receives approximately 708,714 g Fe / day (Dean et al. unpublished) and average pH value of 2, will require a constructed wetland with an area of

70 ha in order to provide efficient remediation. This unrealistically large area illustrates the need for more efficient engineering designs and, in many cases, pre – treatment based on abiotic mechanisms such as chemical neutralization (Whitehead and Prior, 2005).

5.3 FUTURE DIRECTIONS

From the key findings described above, future approaches should be focused on the analysis of metal complexation such as S compounds and organic matter in the water column, water/sediment interface and within sediment depths in order to further understand possible sulphide – mediated metal immobilization or metal adsorption to organic carbon mechanisms along the wetland and therefore, to replicate these mechanisms in new/improved technologies for passive AMD remediation such as constructed wetland. Measurement of mRNA transcripts should also be performed in order to validate if the abundance of metabolic pathways related to metal attenuation from fresh and incubated sediments are being expressed in the microbial communities. As a first approach, expression of relevant genes with significant differences detected in this study could be screened for differences in mRNA 223 expression by RT-PCR. Finally, experiments including living plants such as vegetated mesocosms or in situ isolation of plant stands should be addressed in order to evaluate the metal chemistry and microbial activity in vegetated sediments, including over extended periods of time where the influence of more variables (plant decay, plants exudates, oxygen loss) can be evaluated.

224

6 REFERENCES

ADAMS, L., HARRISON, J., LLOYD, J., LANGLEY, S. & FORTIN, D. 2007. Activity and diversity of Fe (III)-reducing bacteria in a 3000-year-old acid mine drainage site analogue. Geomicrobiology Journal, 24, 295- 305. AGUINAGA, O. E., MCMAHON, A., WHITE, K. N., DEAN, A. P. & PITTMAN, J. K. 2018. Microbial community shifts in response to acid mine drainage pollution within a natural wetland ecosystem. Frontiers in Microbiology, 9, 1445. AKCIL, A. & KOLDAS, S. 2006. Acid Mine Drainage (AMD): causes, treatment and case studies. Journal of Cleaner Production, 14, 1139-1145. ALAZARD, D., JOSEPH, M., BATTAGLIA-BRUNET, F., CAYOL, J.-L. & OLLIVIER, B. 2010. Desulfosporosinus acidiphilus sp. nov.: a moderately acidophilic sulfate-reducing bacterium isolated from acid mining drainage sediments. Extremophiles, 14, 305-312. ALEEM, M. 1965. Thiosulfate oxidation and electron transport in Thiobacillus novellus. Journal of Bacteriology, 90, 95-101. ALI, H., KHAN, E. & SAJAD, M. A. 2013. Phytoremediation of heavy metals—concepts and applications. Chemosphere, 91, 869-881. ALVAREZ, M. T., CRESPO, C. & MATTIASSON, B. 2007. Precipitation of Zn (II), Cu (II) and Pb (II) at bench-scale using biogenic hydrogen sulfide from the utilization of volatile fatty acids. Chemosphere, 66, 1677- 1683. AMARAL-ZETTLER, L. A., ZETTLER, E. R., THEROUX, S. M., PALACIOS, C., AGUILERA, A. & AMILS, R. 2011. Microbial community structure across the tree of life in the extreme Rio Tinto. ISME Journal, 5, 42- 50. AMOURIC, A., APPIA-AYME, C., YARZABAL, A. & BONNEFOY, V. Regulation of the iron and sulfur oxidation pathways in the acidophilic Acidithiobacillus ferrooxidans. Advanced Materials Research, 2009. Trans Tech Publ, 163-166. ANANDHAM, R., INDIRAGANDHI, P., MADHAIYAN, M., RYU, K. Y., JEE, H. J. & SA, T. M. 2008. Chemolithoautotrophic oxidation of thiosulfate and phylogenetic distribution of sulfur oxidation gene (soxB) in rhizobacteria isolated from crop plants. Research in Microbiology, 159, 579-589. APPIA-AYME, C., BENGRINE, A., CAVAZZA, C., GIUDICI-ORTICONI, M.-T., BRUSCHI, M., CHIPPAUX, M. & BONNEFOY, V. 1998. Characterization and expression of the co-transcribed cyc1 and cyc2 genes encoding the cytochrome c 4 (c 552) and a high-molecular-mass cytochrome c from Thiobacillus ferrooxidans ATCC 33020. FEMS Microbiology Letters, 167, 171-177. ARMSTRONG, W., COUSINS, D., ARMSTRONG, J., TURNER, D. W. & BECKETT, P. M. 2000. Oxygen Distribution in Wetland Plant Roots and Permeability Barriers to Gas-exchange with the Rhizosphere: a Microelectrode and Modelling Study with Phragmites australis. Annals of Botany, 86, 687-703. ASTA, M. P., AYORA, C., ROMÁN-ROSS, G., CAMA, J., ACERO, P., GAULT, A. G., CHARNOCK, J. M. & BARDELLI, F. 2010. Natural attenuation of arsenic in the Tinto Santa Rosa acid stream (Iberian Pyritic Belt, SW Spain): the role of iron precipitates. Chemical Geology, 271, 1-12. AUERNIK, K. S. & KELLY, R. M. 2008. Identification of components of electron transport chains in the extremely thermoacidophilic crenarchaeon Metallosphaera sedula through iron and sulfur compound oxidation transcriptomes. Applied and environmental microbiology, 74, 7723-7732. AUGUST, E., MCKNIGHT, D., HRNCIR, D. & GARHART, K. 2002. Seasonal variability of metals transport through a wetland impacted by mine drainage in the Rocky Mountains. Environmental Science & Technology 36, 3776-86. AYKOL, A., BUDAKOGLU, M., KUMRAL, M., GULTEKIN, A. H., TURHAN, M., ESENLI, V., YAVUZ, F. & ORGUN, Y. 2003. Heavy metal pollution and acid drainage from the abandoned Balya Pb-Zn sulfide Mine, NW Anatolia, Turkey. Environmental Geology, 45, 198-208. AZAPAGIC, A. 2004. Developing a framework for sustainable development indicators for the mining and minerals industry. Journal of Cleaner Production, 12, 639-662. BAALOUSHA, M. 2009. Aggregation and disaggregation of iron oxide nanoparticles: influence of particle concentration, pH and natural organic matter. Science of the Total Environment, 407, 2093-2101.

225

BABATUNDE, A., ZHAO, Y., O'NEILL, M. & O'SULLIVAN, B. 2008. Constructed wetlands for environmental pollution control: a review of developments, research and practice in Ireland. Environment International, 34, 116-126. BAIS, H. P., WEIR, T. L., PERRY, L. G., GILROY, S. & VIVANCO, J. M. 2006. The role of root exudates in rhizosphere interactions with plants and other organisms. Annual Review of Plant Biology, 57, 233- 266. BAKER, B. J. & BANFIELD, J. F. 2003. Microbial communities in acid mine drainage. FEMS Microbiology Ecology, 44, 139-152. BAKER, L. A. 1992. Introduction to nonpoint source pollution in the United States and prospects for wetland use. Ecological Engineering, 1, 1-26. BARTON, C. & KARATHANASIS, A. 1999. Renovation of a failed constructed wetland treating acid mine drainage. Environmental Geology, 39, 39-50. BATTY, L. C., BAKER, A. J. & WHEELER, B. 2006. The effect of vegetation on porewater composition in a natural wetland receiving acid mine drainage. Wetlands 26, 40-48. BECERRA, C., LÓPEZ-LUNA, E., ERGAS, S. & NÜSSLEIN, K. 2009. Microcosm-based study of the attenuation of an acid mine drainage-impacted site through biological sulfate and iron reduction. Geomicrobiology Journal 26, 9-20. BELL, F., BULLOCK, S., HÄLBICH, T. & LINDSAY, P. 2001. Environmental impacts associated with an abandoned mine in the Witbank Coalfield, South Africa. International Journal of Coal Geology, 45, 195-216. BELLER, H. R., CHAIN, P. S., LETAIN, T. E., CHAKICHERLA, A., LARIMER, F. W., RICHARDSON, P. M., COLEMAN, M. A., WOOD, A. P. & KELLY, D. P. 2006. The genome sequence of the obligately chemolithoautotrophic, facultatively anaerobic bacterium Thiobacillus denitrificans. Journal of Bacteriology 188, 1473-1488. BENVENUTI, M., MASCARO, I., CORSINI, F., LATTANZI, P., PARRINI, P. & TANELLI, G. 1997. Mine waste dumps and heavy metal pollution in abandoned mining district of Boccheggiano (Southern Tuscany, Italy). Environmental Geology, 30, 238-243. BERNER, R. A. 1984. Sedimentary pyrite formation: an update. Geochimica et cosmochimica Acta, 48, 605- 615. BIER, R. L., VOSS, K. A. & BERNHARDT, E. S. 2014. Bacterial community responses to a gradient of alkaline mountaintop mine drainage in Central Appalachian streams. The Isme Journal, 9, 1378. BIRD, L. J., BONNEFOY, V. & NEWMAN, D. K. 2011. Bioenergetic challenges of microbial iron metabolisms. Trends in Microbiology, 19, 330-340. BLGHAM, J. M., SCHWERTMANN, U., CARLSON, L. & MURAD, E. 1990. A poorly crystallized oxyhydroxysulfate of iron formed by bacterial oxidation of Fe(II) in acid mine waters. Geochimica et Cosmochimica Acta, 54, 2743-2758. BOLGER, A. M., LOHSE, M. & USADEL, B. 2014. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics, 30, 2114-2120. BOLS, N. C., BRUBACHER, J. L., GANASSIN, R. C. & LEE, L. E. 2001. Ecotoxicology and innate immunity in fish. Developmental & Comparative Immunology, 25, 853-873. BOND, P. L., DRUSCHEL, G. K. & BANFIELD, J. F. 2000. Comparison of Acid Mine Drainage Microbial Communities in Physically and Geochemically Distinct Ecosystems. Applied and Environmental Microbiology, 66, 4962-4971. BONNEFOY, V. & HOLMES, D. S. 2012. Genomic insights into microbial iron oxidation and iron uptake strategies in extremely acidic environments. Environmental Microbiology, 14, 1597-1611. BORKOWSKI, A. 2013. Geomicrobiology of Acid Mine Drainage in the weathering zone of pyrite-bearing schists in the Rudawy Janowickie Mountains (Poland). Geological Quarterly, 57, 601-612, doi: 10.7306/gq. 1115. BOULT, S. 1996. Fluvial metal transport near sources of acid mine-drainage: relationships of soluble, suspended and deposited metal. Mineralogical Magazine, 60, 325-335. BOULT, S., COLLINS, D. N., WHITE, K. & CURTIS, C. 1994. Metal transport in a stream polluted by acid mine drainage—the Afon Goch, Anglesey, UK. Environmental Pollution, 84, 279-284. BOWMAN, J. S. & DUCKLOW, H. W. 2015. Microbial Communities Can Be Described by Metabolic Structure: A General Framework and Application to a Seasonally Variable, Depth-Stratified Microbial Community from the Coastal West Antarctic Peninsula. PLOS ONE, 10, e0135868. 226

BRANTNER, J. S. & SENKO, J. M. 2014. Response of soil-associated microbial communities to intrusion of coal mine-derived acid mine drainage. Environmental Science & Technology, 48, 8556-8563. BRAY, J. P., BROADY, P. A., NIYOGI, D. K. & HARDING, J. S. 2009. Periphyton communities in New Zealand streams impacted by acid mine drainage. Marine and Freshwater Research, 59, 1084-1091. BRENNER, F. J. & COOPER, W. L. 1978. Effect of suspended iron hydroxide on the hatchability and embryonic development of the coho salmon. The Ohio Journal of Science, 78, 34 - 38. BRIX, H. 1994. Use of constructed wetlands in water pollution control: historical development, present status, and future perspectives. Water Science and Technology, 30, 209-24. BRUNEEL, O., DURAN, R., CASIOT, C., ELBAZ-POULICHET, F. & PERSONNÉ, J.-C. 2006. Diversity of microorganisms in Fe-As-rich acid mine drainage waters of Carnoules, France. Applied and Environmental Microbiology, 72, 551-556. BRYAN, C. G., HALLBERG, K. B. & JOHNSON, D. B. 2004. Microbial populations in surface spoil at the abandoned Mynydd Parys copper mines. In Mine Water 2004–Proceedings of the International Mine Water Association Symposium (Jarvis AP, Dudgeon BA & Younger PL, eds) 107-112. CABRI, L. J. 1992. The distribution of trace precious metals in minerals and mineral products. Mineralogical Magazine, 56, 289-308. CAPORASO, J. G., KUCZYNSKI, J., STOMBAUGH, J., BITTINGER, K., BUSHMAN, F. D., COSTELLO, E. K., FIERER, N., PENA, A. G., GOODRICH, J. K., GORDON, J. I., HUTTLEY, G. A., KELLEY, S. T., KNIGHTS, D., KOENIG, J. E., LEY, R. E., LOZUPONE, C. A., MCDONALD, D., MUEGGE, B. D., PIRRUNG, M., REEDER, J., SEVINSKY, J. R., TURNBAUGH, P. J., WALTERS, W. A., WIDMANN, J., YATSUNENKO, T., ZANEVELD, J. & KNIGHT, R. 2010. QIIME allows analysis of high-throughput community sequencing data. Nature Methods, 7, 335-6. CHAPELLE, F. H. & LOVLEY, D. R. 1992. Competitive exclusion of sulfate reduction by Fe (lll)‐reducing bacteria: a mechanism for producing discrete zones of high‐iron ground water. Groundwater, 30, 29-36. CHAWCHAI, S., KYLANDER, M. E., CHABANGBORN, A., LÖWEMARK, L. & WOHLFARTH, B. 2016. Testing commonly used X‐ray fluorescence core scanning‐based proxies for organic‐rich lake sediments and peat. Boreas, 45, 180-189. CHEN HONG, Y. S., YI XING, YANG LI 2015. Illumina MiSeq sequencing investigation on the contrasting soil bacterial community structures in different iron mining areas. Environmental Science and Pollution Research, 22, 10788–10799. CHEN, L.-X., HU, M., HUANG, L.-N., HUA, Z.-S., KUANG, J.-L., LI, S.-J. & SHU, W.-S. 2015a. Comparative metagenomic and metatranscriptomic analyses of microbial communities in acid mine drainage. ISME Journal, 9, 1579-1592. CHEN, M., LU, G., GUO, C., YANG, C., WU, J., HUANG, W., YEE, N. & DANG, Z. 2015b. Sulfate migration in a river affected by acid mine drainage from the Dabaoshan mining area, South China. Chemosphere, 119, 734-743. CHENG, S., GROSSE, W., KARRENBROCK, F. & THOENNESSEN, M. 2002. Efficiency of constructed wetlands in decontamination of water polluted by heavy metals. Ecological Engineering, 18, 317-325. CHILDERS, S. E., CIUFO, S. & LOVLEY, D. R. 2002. Geobacter metallireducens accesses insoluble Fe (III) oxide by chemotaxis. Nature, 416, 767. CHIN, K.-J., ESTEVE-NÚNEZ, A., LEANG, C. & LOVLEY, D. R. 2004. Direct correlation between rates of anaerobic respiration and levels of mRNA for key respiratory genes in Geobacter sulfurreducens. Applied Environmental Microbiology, 70, 5183-5189. CLARKE, K. & AINSWORTH, M. 1993. A method of linking multivariate community. Marine ecology progress series, 92, 205-219. CLARKE, K. R. 1993. Non-parametric multivariate analyses of changes in community structure. Australian Journal of Ecology, 18, 117-143. CLARKE, K. R., SOMERFIELD, P. J. & GORLEY, R. N. 2008. Testing of null hypotheses in exploratory community analyses: similarity profiles and biota-environment linkage. Journal of Experimental Marine Biology and Ecology, 366, 56-69. COATES, J. D., BHUPATHIRAJU, V. K., ACHENBACH, L. A., MCLNERNEY, M. & LOVLEY, D. R. 2001. Geobacter hydrogenophilus, Geobacter chapellei and Geobacter grbiciae, three new, strictly anaerobic, dissimilatory Fe (III)-reducers. International Journal of Systematic and Evolutionary Microbiology, 51, 581-588. 227

COBLEY, J. G. & HADDOCK, B. A. 1975. The respiratory chain of Thiobacillus ferrooxidans: the reduction of cytochromes by Fe2+ and the preliminary characterization of rusticyanin a novel ‘blue’copper protein. FEBS Letters, 60, 29-33. COLLINS, B., MCARTHUR, J. V. & SHARITZ, R. R. 2004. Plant effects on microbial assemblages and remediation of acidic coal pile runoff in mesocosm treatment wetlands. Ecological Engineering, 23, 107-115. COLMER, T. 2003. Long‐distance transport of gases in plants: a perspective on internal aeration and radial oxygen loss from roots. Plant, Cell & Environment 26, 17-36. COOK, R. & KELLY, C. 1989. Sulfur cycling and fluxes in temperate dimictic lakes. Sulfur cycling of the continents, 145 - 188. COULTON, R., BULLEN, C. & HALLETT, C. 2003. The design and optimisation of active mine water treatment plants. Land Contamination and Reclamation, 11, 273-280. COUPLAND, K. & JOHNSON., D. B. 2004. Geochemistry and microbiology of an impounded subterranean acidic water body at Mynydd Parys, Anglesey, Wales. Geobiology, 2, 77-86. COX, J. C. & BOXER, D. H. 1978. The purification and some properties of rusticyanin, a blue copper protein involved in iron(II) oxidation from Thiobacillus ferro-oxidans. Biochemical Journal, 174, 497-502. CROUDACE, I. W., ROMANO, E., AUSILI, A., BERGAMIN, L. & ROTHWELL, R. G. 2015. X-ray core scanners as an environmental forensics tool: a case study of polluted harbour sediment (Augusta Bay, Sicily). Micro-XRF studies of sediment cores. Springer. CROWDER, A. & MACFIE, S. 1986. Seasonal deposition of ferric hydroxide plaque on roots of wetland plants. Canadian Journal of Botany, 64, 2120-2124. CRUZ-MARTÍNEZ, K., SUTTLE, K. B., BRODIE, E. L., POWER, M. E., ANDERSEN, G. L. & BANFIELD, J. F. 2009. Despite strong seasonal responses, soil microbial consortia are more resilient to long-term changes in rainfall than overlying grassland. The ISME Journal, 3, 738. CUMMINGS, D., SNOEYENBOS-WEST, O., NEWBY, D., NIGGEMYER, A., LOVLEY, D., ACHENBACH, L. & ROSENZWEIG, R. 2003. Diversity of Geobacteraceae species inhabiting metal-polluted freshwater lake sediments ascertained by 16S rDNA analyses. Microbial Ecology, 46, 257-269. CURUTCHET, G., POGLIANI, C. & DONATI, E. 1995. Indirect bioleaching of covellite by Thiobacillus thiooxidans with an oxidant agent. Biotechnology Letters, 17, 1251-1256. DAVIES, H., WEBER, P., LINDSAY, P., CRAW, D. & POPE, J. 2011. Characterisation of acid mine drainage in a high rainfall mountain environment, New Zealand. Science of the Total Environment, 409, 2971- 2980. DEAN, A. P., HARTLEY, A., MCINTOSH, O. A., SMITH, A., FEORD, H. K., HOLMBERG, N. H., KING, T., YARDLEY, E., WHITE, K. N. & PITTMAN, J. K. 2019. Metabolic adaptation of a Chlamydomonas acidophila strain isolated from acid mine drainage ponds with low eukaryotic diversity. Science of the Total Environment, 647, 75-87. DEAN, A. P., LYNCH, S., ROWLAND, P., TOFT, B. D., PITTMAN, J. K. & WHITE, K. N. 2013. Natural wetlands are efficient at providing long-term metal remediation of freshwater systems polluted by acid mine drainage. Environmental Science & Technology, 47, 12029-12036. DELONG, E. F. 1992. Archaea in coastal marine environments. Proceedings of the National Academy of Sciences, 89, 5685-5689. DESANTIS, T. Z., HUGENHOLTZ, P., LARSEN, N., ROJAS, M., BRODIE, E. L., KELLER, K., HUBER, T., DALEVI, D., HU, P. & ANDERSEN, G. L. 2006. Greengenes, a chimera-checked 16S rRNA gene database and workbench compatible with ARB. Applied and Environmental Microbiology, 72, 5069-5072. DIABY, N., DOLD, B., ROHRBACH, E., HOLLIGER, C. & ROSSI, P. 2015. Temporal evolution of bacterial communities associated with the in situ wetland-based remediation of a marine shore porphyry copper tailings deposit. Science of The Total Environment, 533, 110-121. DOLLA, A., FOURNIER, M. & DERMOUN, Z. 2006. Oxygen defense in sulfate-reducing bacteria. Journal of Biotechnology, 126, 87-100. DONG, J., MAO, W., ZHANG, G., WU, F. & CAI, Y. 2007. Root excretion and plant tolerance to cadmium toxicity-a review. Plant, Soil & Environment, 53, 193. DOULATI ARDEJANI, F., JODEIRI SHOKRI, B., BAGHERI, M. & SOLEIMANI, E. 2010. Investigation of pyrite oxidation and acid mine drainage characterization associated with Razi active coal mine and coal washing waste dumps in the Azad shahr–Ramian region, northeast Iran. Environmental Earth Sciences, 61, 1547-1560. 228

DREWNIAK, L., KRAWCZYK, P. S., MIELNICKI, S., ADAMSKA, D., SOBCZAK, A., LIPINSKI, L., BUREC-DREWNIAK, W. & SKLODOWSKA, A. 2016. Physiological and Metagenomic Analyses of Microbial Mats Involved in Self-Purification of Mine Waters Contaminated with Heavy Metals. Frontiers in Microbiology, 7. DRURY, W. J. 1999. Treatment of acid mine drainage with anaerobic solid-substrate reactors. Water Environment Research, 71, 1244-1250. DUNBABIN, J. S. & BOWMER, K. H. 1992. Potential use of constructed wetlands for treatment of industrial wastewaters containing metals. Science of the Total environment, 111, 151-168. ECHEVERRIA, J., MORERA, M., MAZKIARAN, C. & GARRIDO, J. 1998. Competitive sorption of heavy metal by soils. Isotherms and fractional factorial experiments. Environmental Pollution, 101, 275-284. EDGAR, R. C. 2010. Search and clustering orders of magnitude faster than BLAST. Bioinformatics, 26, 2460- 2461. EDGAR, R. C., HAAS, B. J., CLEMENTE, J. C., QUINCE, C. & KNIGHT, R. 2011. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics, 27, 2194-2200. EMERSON, D. & MOYER, C. 1997. Isolation and characterization of novel iron-oxidizing bacteria that grow at circumneutral pH. Applied and Environmental Microbiology, 63, 4784-4792. EMERSON, D. & WEISS, J. V. 2004. Bacterial iron oxidation in circumneutral freshwater habitats: findings from the field and the laboratory. Geomicrobiology Journal, 21, 405-414. EMERSON, D., WEISS, J. V. & MEGONIGAL, J. P. 1999. Iron-oxidizing bacteria are associated with ferric hydroxide precipitates (Fe-plaque) on the roots of wetland plants. Applied and Environmental Microbiology, 65, 2758-2761. ENGELAAR, W., SYMENS, J., LAANBROEK, H. & BLOM, C. 1995. Preservation of nitrifying capacity and nitrate availability in waterlogged soils by radial oxygen loss from roots of wetland plants. Biology and Fertility of Soils 20, 243-248. EPEL, B., SCHÄFER, K.-O., QUENTMEIER, A., FRIEDRICH, C. & LUBITZ, W. 2005. Multifrequency EPR analysis of the dimanganese cluster of the putative sulfate thiohydrolase SoxB of Paracoccus pantotrophus. JBIC Journal of Biological Inorganic Chemistry, 10, 636-642. EVANGELOU, V. & ZHANG, Y. 1995. A review: pyrite oxidation mechanisms and acid mine drainage prevention. Critical Reviews in Environmental Science and Technology, 25, 141-199. FERRIS, F., HALLBECK, L., KENNEDY, C. & PEDERSEN, K. 2004. Geochemistry of acidic Rio Tinto headwaters and role of bacteria in solid phase metal partitioning. Chemical Geology, 212, 291-300. FERRIS, F. G. 2005. Biogeochemical Properties of Bacteriogenic Iron Oxides. Geomicrobiology Journal, 22, 79- 85. FIERER, N. & JACKSON, R. B. 2006. The diversity and biogeography of soil bacterial communities. Proceedings of the National Academy of Sciences of the United States of America, 103, 626-631. FORTIN, D. & LANGLEY, S. 2005. Formation and occurrence of biogenic iron-rich minerals. Earth-Science Reviews, 72, 1-19. FOSTER, P., HUNT, D. & MORRIS, A. 1978. Metals in an acid mine stream and estuary. Science of the Total Environment, 9, 75-86. FRANKLINA, N. M., STAUBERA, J. L., MARKICHC, S. J. & LIMB, R. P. 2000. pH-dependent toxicity of copper and uranium to a tropical freshwater alga (Chlorella sp.). Aquatic Toxicology, 48, 275-289. FREDRICKSON, J. K., ZACHARA, J. M., KENNEDY, D. W., DONG, H., ONSTOTT, T. C., HINMAN, N. W. & LI, S.-M. 1998. Biogenic iron mineralization accompanying the dissimilatory reduction of hydrous ferric oxide by a groundwater bacterium. Geochimica et Cosmochimica Acta, 62, 3239-3257. FRIEDRICH, C. G., ROTHER, D., BARDISCHEWSKY, F., QUENTMEIER, A. & FISCHER, J. 2001. Oxidation of reduced inorganic sulfur compounds by bacteria: emergence of a common mechanism? Applied and environmental microbiology, 67, 2873-2882. FROHNE, T., RINKLEBE, J., DIAZ-BONE, R. A. & DU LAING, G. 2011. Controlled variation of redox conditions in a floodplain soil: Impact on metal mobilization and biomethylation of arsenic and antimony. Geoderma, 160, 414-424. FRÖMMICHEN, R., WENDT-POTTHOFF, K., FRIESE, K. & FISCHER, R. 2004. Microcosm studies for neutralization of hypolimnic acid mine pit lake water (pH 2.6). Environmental Science & Technology, 38, 1877-1887. GABRIEL, K. N., MEDVEDEFF, C. A. & KELLER, J. K. 2017. Research Article: Microbial organic matter reduction in a peatland soil: the importance of water-table level. BIOS, 88, 39-45.

229

GALÁN, E., GÓMEZ-ARIZA, J. L., GONZÁLEZ, I., FERNÁNDEZ-CALIANI, J. C., MORALES, E. & GIRÁLDEZ, I. 2003. Heavy metal partitioning in river sediments severely polluted by acid mine drainage in the Iberian Pyrite Belt. Applied Geochemistry, 18, 409-421. GARRIDO, A., CONDORI, J., STROSNIDER, W. & NAIRN, R. Acid mine drainage impacts on irrigation water resources, agricultural soils, and potatoes in Potosí, Bolivia. Proceedings of the 26th National Conference of the American Society of Mining and Reclamation, Billings, 2009. 486-499. GAULT, A. G., LANGLEY, S., IBRAHIM, A., RENAUD, R., TAKAHASHI, Y., BOOTHMAN, C., LLOYD, J. R., CLARK, I. D., FERRIS, F. G. & FORTIN, D. 2012. Seasonal Changes In Mineralogy, Geochemistry and Microbial Community of Bacteriogenic Iron Oxides (BIOS) Deposited in a Circumneutral Wetland. Geomicrobiology Journal, 29, 161-172. GAZEA, B., ADAM, K. & KONTOPOULOS, A. 1996. A review of passive systems for the treatment of acid mine drainage. Minerals Engineering, 9, 23-42. GHIORSE, W. 1984. Biology of iron-and manganese-depositing bacteria. Annual Reviews in Microbiology, 38, 515-550. GHOSH, W., MANDAL, S. & ROY, P. 2006. Paracoccus bengalensis sp. nov., a novel sulfur-oxidizing chemolithoautotroph from the rhizospheric soil of an Indian tropical leguminous plant. Systematic and Applied Microbiology, 29, 396-403. GIBERT, O., DE PABLO, J., CORTINA, J. L. & AYORA, C. 2004. Chemical characterisation of natural organic substrates for biological mitigation of acid mine drainage. Water Research, 38, 4186-4196. GIBERT, O., DE PABLO, J., CORTINA, J. L. & AYORA, C. 2005. Municipal compost-based mixture for acid mine drainage bioremediation: Metal retention mechanisms. Applied Geochemistry, 20, 1648-1657. GONZALEZ-TORIL, E., LLOBET-BROSSA, E., CASAMAYOR, E. O., AMANN, R. & AMILS, R. 2003. Microbial Ecology of an Extreme Acidic Environment, the Tinto River. Applied and Environmental Microbiology, 69, 4853-4865. GRAY, N. 1998. Acid mine drainage composition and the implications for its impact on lotic systems. Water Research, 32, 2122-2134. GRYBOS, M., DAVRANCHE, M., GRUAU, G., PETITJEAN, P. & PÉDROT, M. 2009. Increasing pH drives organic matter solubilization from wetland soils under reducing conditions. Geoderma, 154, 13-19. HABERL, R., GREGO, S., LANGERGRABER, G., KADLEC, R. H., CICALINI, A.-R., DIAS, S. M., NOVAIS, J. M., AUBERT, S., GERTH, A. & THOMAS, H. 2003. Constructed wetlands for the treatment of organic pollutants. Journal of Soils and Sediments, 3, 109. HAFEZNEZAMI, S., KIM, J.-L. & REDMAN, J. 2012. Evaluating removal efficiency of heavy metals in constructed wetlands. Journal of Environmental Engineering, 138, 475-482. HALLBERG, K. B., COUPLAND, K., KIMURA, S. & JOHNSON, D. B. 2006. Macroscopic streamer growths in acidic, metal-rich mine waters in North Wales consist of novel and remarkably simple bacterial communities. Applied and Environmental Microbiology, 72, 2022-2030. HAMMER, Ø., HARPER, D. & RYAN, P. 2001. PAST-Palaeontological statistics. Palaeontological Museum, University of Oslo, Oslo, Norway, 25, 2009. HAN, C., KOTSYURBENKO, O., CHERTKOV, O., HELD, B., LAPIDUS, A., NOLAN, M., LUCAS, S., HAMMON, N., DESHPANDE, S. & CHENG, J.-F. 2012. Complete genome sequence of the sulfur compounds oxidizing chemolithoautotroph Sulfuricurvum kujiense type strain (YK-1 T). Standards in genomic sciences, 6, 94. HANSEL, C., FENDORF, S., SUTTON, S. & NEWVILLE, M. 2001. Characterization of Fe Plaque and Associated Metals on the Roots of Mine-Waste Impacted Aquatic Plants. Environmental Science & Technology, 35, 3863-8. HARRISON JR, A. P. 1984. The acidophilic thiobacilli and other acidophilic bacteria that share their habitat. Annual Reviews in Microbiology, 38, 265-292. HEADD, B. & ENGEL, A. S. 2013. Evidence for niche partitioning revealed by the distribution of sulfur oxidation genes collected from areas of a terrestrial sulfidic spring with differing geochemical conditions. Applied and Environmental Microbiology, 79, 1171-1182. HEALY, M. & HICKEY, K. R. Historic land reclamation in the intertidal wetlands of the Shannon estuary, western Ireland. The 7th International Coastal Symposium, ICS 2002; Northern Ireland, 2002. Coastal Education and Research Foundation (CERF). HEDRICH, S., SCHLÖMANN, M. & JOHNSON, D. B. 2011. The iron-oxidizing proteobacteria. Microbiology, 157, 1551-1564. 230

HEINZEL, E., JANNECK, E., GLOMBITZA, F., SCHLÖMANN, M. & SEIFERT, J. 2009. Population Dynamics of Iron- Oxidizing Communities in Pilot Plants for the Treatment of Acid Mine Waters. Environmental Science and Technology, 43, 6138–6144. HENROT, J. & WIEDER, R. K. 1990. Processes of iron and manganese retention in laboratory peat microcosms subjected to acid mine drainage. Journal of Environmental Quality, 19, 312-320. HERLIHY, A. T., KAUFMANN, P. R., MITCH, M. E. & BROWN, D. D. 1990. Regional estimates of acid mine drainage impact on streams in the mid-Atlantic and southeastern United States. Water, Air, and Soil Pollution, 50, 91-107. HERLIHY, A. T. & MILLS, A. L. 1985. Sulfate reduction in freshwater sediments receiving acid mine drainage. Applied and Environmental Microbiology, 49, 179-186. HUA, Z.-S., HAN, Y.-J., CHEN, L.-X., LIU, J., HU, M., LI, S.-J., KUANG, J.-L., CHAIN, P. S., HUANG, L.-N. & SHU, W.-S. 2015. Ecological roles of dominant and rare prokaryotes in acid mine drainage revealed by metagenomics and metatranscriptomics. The ISME Jnoournal 9, 1280 - 1294. HUANG, L.-N., KUANG, J.-L. & SHU, W.-S. 2016. Microbial ecology and evolution in the acid mine drainage model system. Trends in Microbiology, 24, 581-593. HUANG, Y., ORTIZ, L., AGUIRRE, P., GARCÍA, J., MUJERIEGO, R. & BAYONA, J. M. 2005. Effect of design parameters in horizontal flow constructed wetland on the behaviour of volatile fatty acids and volatile alkylsulfides. Chemosphere, 59, 769-777. IDASZKIN, Y. L., CAROL, E. & MARÍA DEL PILAR, A. 2017. Mechanism of removal and retention of heavy metals from the acid mine drainage to coastal wetland in the Patagonian marsh. Chemosphere, 183, 361-370. JACOB, D. L. & OTTE, M. L. 2003. Conflicting processes in the wetland plant rhizosphere: metal retention or mobilization? Water, Air and Soil Pollution: Focus, 3, 91-104. JAVED, M. T., AKRAM, M. S., HABIB, N., TANWIR, K., ALI, Q., NIAZI, N. K., GUL, H. & IQBAL, N. 2018. Deciphering the growth, organic acid exudations, and ionic homeostasis of Amaranthus viridis L. and Portulaca oleracea L. under lead chloride stress. Environmental Science Pollution Research, 25, 2958-2971. JEZIERSKA, B., ŁUGOWSKA, K. & WITESKA, M. 2009. The effects of heavy metals on embryonic development of fish (a review). Fish physiology and biochemistry, 35, 625-640. JIANG, F. Y., CHEN, X. & LUO, A. C. 2009. Iron plaque formation on wetland plants and its influence on phosphorus, calcium and metal uptake. Aquatic Ecology, 43, 879-890. JIN, S., FALLGREN, P. H., MORRIS, J. M. & GOSSARD, R. B. 2008. Biological source treatment of acid mine drainage using microbial and substrate amendments: microcosm studies. Mine Water and the Environment, 27, 20-30. JOHNSON, D. 2000. Biological removal of sulfurous compounds from inorganic wastewaters. Environmental technologies to treat sulfur pollution: Principles and engineering, 175-205. JOHNSON, D. B. & HALLBERG, K. B. 2005. Acid mine drainage remediation options: a review. Science of the total environment, 338, 3-14. JOHNSON, D. B., HALLBERG, K. B. & HEDRICH, S. 2013. Uncovering a microbial enigma: isolation and characterization of the streamer-generating, iron-oxidizing acidophilic bacterium,“Ferrovum myxofaciens”. Applied and Environmental Microbiology, AEM. 03230-13. JOHNSON, D. B., ROLFE, S., HALLBERG, K. B. & IVERSEN, E. 2001. Isolation and phylogenetic characterization of acidophilic microorganisms indigenous to acidic drainage waters at an abandoned Norwegian copper mine. Environmental Microbiology, 3, 630-637. JONES, D. S., KOHL, C., GRETTENBERGER, C., LARSON, L. N., BURGOS, W. D. & MACALADYA, J. L. 2015. Geochemical Niches of Iron-Oxidizing Acidophiles in Acidic Coal Mine Drainage. Applied and Environmental Microbiology, 81, 1242-1250. JONG, T. & PARRY, D. L. 2003. Removal of sulfate and heavy metals by sulfate reducing bacteria in short- term bench scale upflow anaerobic packed bed reactor runs. Water research, 37, 3379-3389. JUNG, S. & MARRON, J. S. 2009. PCA consistency in high dimension, low sample size context. Ann. Statist., 37, 4104-4130. KAASALAINEN, M. & YLI-HALLA, M. 2003. Use of sequential extraction to assess metal partitioning in soils. Environmental Pollution, 126, 225-233. KALBITZ, K. & WENNRICH, R. 1998. Mobilization of heavy metals and arsenic in polluted wetland soils and its dependence on dissolved organic matter. Science of the Total Environment, 209, 27-39. 231

KALIN, M., FYSON, A. & WHEELER, W. N. 2006. The chemistry of conventional and alternative treatment systems for the neutralization of acid mine drainage. Science of The Total Environment, 366, 395- 408. KANAPARTHI, D., POMMERENKE, B., CASPER, P. & DUMONT, M. G. 2013. Chemolithotrophic nitrate- dependent Fe (II)-oxidizing nature of actinobacterial subdivision lineage TM3. The ISME Journal, 7, 1582. KAPPLER, A. & STRAUB, K. L. 2005. Geomicrobiological cycling of iron. Reviews in Mineralogy and Geochemistry, 59, 85-108. KARATHANASIS, A. & JOHNSON, C. 2003. Metal removal potential by three aquatic plants in an acid mine drainage wetland. Mine Water and the Environment, 22, 22-30. KASHEM, M. & SINGH, B. 2001. Metal availability in contaminated soils: I. Effects of floodingand organic matter on changes in Eh, pH and solubility of Cd, Ni and Zn. Nutrient Cycling in Agroecosystems, 61, 247-255. KAY, C. M., ROWE, O. F., ROCCHETTI, L., COUPLAND, K., HALLBERG, K. B. & JOHNSON, D. B. 2013. Evolution of microbial “streamer” growths in an acidic, metal-contaminated stream draining an abandoned underground copper mine. Life, 3, 189-210. KELLY, D. P. & WOOD, A. P. 2000. Reclassification of some species of Thiobacillus to the newly designated genera Acidithiobacillus gen. nov., Halothiobacillus gen. nov. and Thermithiobacillus gen. nov. International Journal of Systematic and Evolutionary Microbiology, 50, 511-516. KELMAN WIEDER, R. 1990. Metal cation binding to Sphagnum peat and sawdust: Relation to wetland treatment of metal-polluted waters. Water, Air, and Soil Pollution, 53, 391-400. KEPLER, D. A. & MCCLEARY, E. C. J. B. O. M. S. P. S. B.-. 1994. Successive alkalinity-producing systems (SAPS) for the treatment of acidic mine drainage. 1, 185-194. KIKUCHI, S., MAKITA, H., TAKAI, K., YAMAGUCHI, N. & TAKAHASHI, Y. 2014. Characterization of biogenic iron oxides collected by the newly designed liquid culture method using diffusion chambers. Geobiology, 12, 133-45. KIM, J., KIM, S. & TAZAKI, K. 2002. Mineralogical characterization of microbial ferrihydrite and schwertmannite, and non-biogenic Al-sulfate precipitates from acid mine drainage in the Donghae mine area, Korea. Environmental Geology, 42, 19-31. KINZLER, K., GEHRKE, T., TELEGDI, J. & SAND, W. 2003. Bioleaching—a result of interfacial processes caused by extracellular polymeric substances (EPS). Hydrometallurgy, 71, 83-88. KLEINMANN, R., HEDIN, R. & NAIRN, R. 1998. Treatment of mine drainage by anoxic limestone drains and constructed wetlands. Acidic Mining Lakes. Springer. KLEINMANN, R. L. P. 1990. Acid mine water treatment using engineered wetlands. International journal of mine water, 9, 269-276. KLINDWORTH, A., PRUESSE, E., SCHWEER, T., PEPLIES, J., QUAST, C., HORN, M. & GLOCKNER, F. O. 2013. Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Research, 41, e1. KLUEGLEIN, N. & KAPPLER, A. 2013. Abiotic oxidation of Fe (II) by reactive nitrogen species in cultures of the nitrate‐reducing Fe (II) oxidizer A cidovorax sp. BoFeN1–questioning the existence of enzymatic Fe (II) oxidation. Geobiology 11, 180-190. KNIGHT, R. L. 1992. Ancillary benefits and potential problems with the use of wetlands for nonpoint source pollution control. Ecological Engineering, 1, 97-113. KONDO, R., NEDWELL, D. B., PURDY, K. J. & SILVA, S. Q. 2004. Detection and enumeration of sulphate- reducing bacteria in estuarine sediments by competitive PCR. Geomicrobiology Journal, 21, 145- 157. KOO, H., HAKIM, J. A., MORROW, C. D., EIPERS, P. G., DAVILA, A., ANDERSEN, D. T. & BEJ, A. K. 2017. Comparison of two bioinformatics tools used to characterize the microbial diversity and predictive functional attributes of microbial mats from Lake Obersee, Antarctica. Journal of Microbiological Methods, 140, 15-22. KOSCHORRECK, M., BOZAU, E., FRÖMMICHEN, R., GELLER, W., HERZSPRUNG, P. & WENDT-POTTHOFF, K. 2007. Processes at the sediment water interface after addition of organic matter and lime to an acid mine pit lake mesocosm. Environmental Science & Technology, 41, 1608-1614.

232

KUANG, J.-L., HUANG, L.-N., CHEN, L.-X., HUA, Z.-S., LI, S.-J., HU, M., LI, J.-T. & SHU, W.-S. 2013. Contemporary environmental variation determines microbial diversity patterns in acid mine drainage. ISME Journal, 7, 1038-1050. KUANG, J., HUANG, L., HE, Z., CHEN, L., HUA, Z., JIA, P., LI, S., LIU, J., LI, J., ZHOU, J. & SHU, W. 2016. Predicting taxonomic and functional structure of microbial communities in acid mine drainage. ISME Journal, 10, 1527-1539. KUMAR, S., STECHER, G. & TAMURA, K. 2016. MEGA7: molecular evolutionary genetics analysis version 7.0 for bigger datasets. Molecular Biology and Evolution, 33, 1870-1874. LA, H.-J., KIM, K.-H., QUAN, Z.-X., CHO, Y.-G. & LEE, S.-T. 2003. Enhancement of sulfate reduction activity using granular sludge in anaerobic treatment of acid mine drainage. Biotechnology Letters, 25, 503- 508. LAGE, C. R., NAYAK, A. & KIM, C. H. 2006. Arsenic ecotoxicology and innate immunity. Integrative and Comparative Biology, 46, 1040-54. LANCLOS, V. C., HENSON, M. W., PITRE, D. M. & THRASH, J. C. 2016. Draft Genome Sequence of Strain LSUCC0135, an Early Diverging Member of the Order Methylophilales in the Phylum Betaproteobacteria. Genome Announcements, 4, e01231-16. LANGLEY, S., GAULT, A., IBRAHIM, A., TAKAHASHI, Y., RENAUD, R., FORTIN, D., CLARK, I. & FERRIS, F. 2009. Sorption of strontium onto bacteriogenic iron oxides. Environmental Science & Technology, 43, 1008-14. LARSEN, S., NIELSEN, L. P. & SCHRAMM, A. 2015. Cable bacteria associated with long‐distance electron transport in N ew E ngland salt marsh sediment. Environmental Microbiology Reports, 7, 175-179. LAUBER, C. L., HAMADY, M., KNIGHT, R. & FIERER, N. 2009. Pyrosequencing-based assessment of soil pH as a predictor of soil bacterial community structure at the continental scale. Applied and Environmental Microbiology, 75, 5111-5120. LENTINI, C. J., WANKEL, S. D. & HANSEL, C. M. 2012. Enriched Iron(III)-Reducing Bacterial Communities are Shaped by Carbon Substrate and Iron Oxide Mineralogy. Frontiers in Microbiology, 3, 404. LI, X., RUI, J., XIONG, J., LI, J., HE, Z., ZHOU, J., YANNARELL, A. C. & MACKIE, R. I. 2014. Functional potential of soil microbial communities in the maize rhizosphere. PLOS One, 9, e112609. LI, Y., SUN, Q., ZHAN, J., YANG, Y. & WANG, D. 2017. Soil-covered strategy for ecological restoration alters the bacterial community structure and predictive energy metabolic functions in mine tailings profiles. Applied Microbiology and Biotechnology, 101, 2549-2561. LIAO, C., PENG, R., LUO, Y., ZHOU, X., WU, X., FANG, C., CHEN, J. & LI, B. 2008. Altered ecosystem carbon and nitrogen cycles by plant invasion: a meta‐analysis. New phytologist, 177, 706-714. LILJEQVIST, M., OSSANDON, F. J., GONZÁLEZ, C., RAJAN, S., STELL, A., VALDES, J., HOLMES, D. S. & DOPSON, M. 2015. Metagenomic analysis reveals adaptations to a cold-adapted lifestyle in a low- temperature acid mine drainage stream. FEMS Microbiology Ecology, 91. LIM, C. K., HASSAN, K. A., PENESYAN, A., LOPER, J. E. & PAULSEN, I. T. 2013. The effect of zinc limitation on the transcriptome of P seudomonas protegens Pf‐5. Environmental Microbiology, 15, 702-715. LIN, C., LU, W. & WU, Y. 2005. Agricultural soils irrigated with acidic mine water: acidity, heavy metals, and crop contamination. Soil Research, 43, 819-826. LIU, F., NI, H.-G., CHEN, F., LUO, Z.-X., SHEN, H., LIU, L. & WU, P. 2012. Metal accumulation in the tissues of grass carps (Ctenopharyngodon idellus) from fresh water around a copper mine in Southeast China. Environmental Monitoring and Assessment, 184, 4289-4299. LIU, J., DONG, Y., XU, H., WANG, D. & XU, J. 2007. Accumulation of Cd, Pb and Zn by 19 wetland plant species in constructed wetland. Journal of Hazardous Materials, 147, 947-953. LIU, J., HUA, Z.-S., CHEN, L.-X., KUANG, J.-L., LI, S.-J., SHU, W.-S. & HUANG, L.-N. 2014. Correlating microbial diversity patterns with geochemistry in an extreme and heterogeneous environment of mine tailings. Applied and Environmental Microbiology, 80, 3677-3686. LIU, Y., ZHANG, Y. & NI, B.-J. 2015. Evaluating enhanced sulfate reduction and optimized volatile fatty acids (VFA) composition in anaerobic reactor by Fe (III) addition. Environmental Science & Technology, 49, 2123-2131. LONG, Z.-E., HUANG, Y., CAI, Z., CONG, W. & OUYANG, F. 2003. Biooxidation of ferrous iron by immobilized Acidithiobacillus ferrooxidans in poly (vinyl alcohol) cryogel carriers. Biotechnology Letters, 25, 245- 249.

233

LOVLEY, D. R. & PHILLIPS, E. J. 1986. Organic matter mineralization with reduction of ferric iron in anaerobic sediments. Applied and Environmental Microbiology, 51, 683-689. LUO, H., LIU, G., ZHANG, R., BAI, Y., FU, S. & HOU, Y. 2014. Heavy metal recovery combined with H2 production from artificial acid mine drainage using the microbial electrolysis cell. Journal of Hazardous Materials 270, 153-159. LYRIC, R. M. & SUZUKI, I. 1970. Enzymes involved in the metabolism of thiosulfate by Thiobacillus thioparus. II. Properties of adenosine-5′-phosphosulfate reductase. Canadian Journal of Biochemistry 48, 344- 354. MACHEL, H. G. 1989. Relationships between sulphate reduction and oxidation of organic compounds to carbonate diagenesis, hydrocarbon accumulations, salt domes, and metal sulphide deposits. Carbonates and Evaporites, 4, 137. MACHEMER, S. D. & WILDEMAN, T. R. 1992. Adsorption compared with sulfide precipitation as metal removal processes from acid mine drainage in a constructed wetland. Journal of Contaminant Hydrology, 9, 115-131. MACÍAS, F., PÉREZ-LÓPEZ, R., CARABALLO, M. A., CÁNOVAS, C. R. & NIETO, J. M. 2017. Management strategies and valorization for waste sludge from active treatment of extremely metal-polluted acid mine drainage: A contribution for sustainable mining. Journal of Cleaner Production, 141, 1057- 1066. MADSEN, E. L. 2011. Microorganisms and their roles in fundamental biogeochemical cycles. Current Opinion in Biotechnology, 22, 456-464. MARASINGHE WADIGE, C. P., TAYLOR, A. M., MAHER, W. A., UBRIHIEN, R. P. & KRIKOWA, F. 2014. Effects of lead-spiked sediments on freshwater bivalve, Hyridella australis: linking organism metal exposure- dose-response. Aquatic Toxicology, 149, 83-93. MARCHAND, L., MENCH, M., JACOB, D. L. & OTTE, M. L. 2010. Metal and metalloid removal in constructed wetlands, with emphasis on the importance of plants and standardized measurements: A review. Environmental Pollution, 158, 3447-61. MARDANOV, A. V., PANOVA, I. A., BELETSKY, A. V., AVAKYAN, M. R., KADNIKOV, V. V., ANTSIFEROV, D. V., BANKS, D., FRANK, Y. A., PIMENOV, N. V. & RAVIN, N. V. 2016. Genomic insights into a new acidophilic, copper-resistant Desulfosporosinus isolate from the oxidized tailings area of an abandoned gold mine. FEMS Microbiology Ecology 92. MASON, R. P., LAPORTE, J. & ANDRES, S. 2000. Factors controlling the bioaccumulation of mercury, methylmercury, arsenic, selenium, and cadmium by freshwater invertebrates and fish. Archives of Environmental Contamination and Toxicology, 38, 283-97. MAYES, W. M., BATTY, L. C., YOUNGER, P. L., JARVIS, A. P., KOIV, M., VOHLA, C. & MANDER, U. 2009. Wetland treatment at extremes of pH: A review. Science of the Total Environment, 407, 3944-3957. MAYES, W. M., POTTER, H. A. & JARVIS, A. P. 2010. Inventory of aquatic contaminant flux arising from historical metal mining in England and Wales. Science of the Total Environment, 408, 3576-83. MAYS, P. A. & EDWARDS, G. S. 2001. Comparison of heavy metal accumulation in a natural wetland and constructed wetlands receiving acid mine drainage. Ecological Engineering, 16, 487-500. MCGRATH, S. P., ZHAO, J. & LOMBI, E. 2002. Phytoremediation of metals, metalloids, and radionuclides. Advances in Agronomy, 75, 1-56. MCKNIGHT, D. M. & FEDER, G. L. 1984. The ecological effect of acid conditions and precipitation of hydrous metal oxides in a Rocky Mountain stream. Hydrobiologia, 119, 129-138. MÉNDEZ-GARCÍA, C., PELÁEZ, A. I., MESA, V., SÁNCHEZ, J., GOLYSHINA, O. V. & FERRER, M. 2015. Microbial diversity and metabolic networks in acid mine drainage habitats. Frontiers in Microbiology, 6. MEYER, F., PAARMANN, D., D'SOUZA, M., OLSON, R., GLASS, E. M., KUBAL, M., PACZIAN, T., RODRIGUEZ, A., STEVENS, R. & WILKE, A. 2008. The metagenomics RAST server–a public resource for the automatic phylogenetic and functional analysis of metagenomes. BMC Bioinformatics 9, 386. MILLER, H., CROUDACE, I. W., BULL, J. M., COTTERILL, C. J., DIX, J. K. & TAYLOR, R. N. 2015. Modern pollution signals in sediments from Windermere, NW England, determined by micro-XRF and lead isotope analysis. Micro-XRF Studies of Sediment Cores. Springer. MILLER, R. K., KIRKHAM, R. M. & KLEINSCHMIT, A. J. 2018. Bioinformatics Analysis of Microbial Abundance and Diversity in Acid Mine Drainage from the Solomon Mine Near Creede, Colorado.

234

MIRETE, S., DE FIGUERAS, C. G. & GONZÁLEZ-PASTOR, J. E. 2007. Novel nickel resistance genes from the rhizosphere metagenome of plants adapted to acid mine drainage. Applied and Environmental Microbiology, 73, 6001-6011. MITSCH, W. J. & WISE, K. M. 1998. Water quality, fate of metals, and predictive model validation of a constructed wetland treating acid mine drainage. Water Research, 32, 1888-1900. MITSUNOBU, S., SHIRAISHI, F., MAKITA, H., ORCUTT, B. N., KIKUCHI, S., JORGENSEN, B. B. & TAKAHASHI, Y. 2012. Bacteriogenic Fe(III) (oxyhydr)oxides characterized by synchrotron microprobe coupled with spatially resolved phylogenetic analysis. Environmental Science & Technology, 46, 3304-11. MIYATA, N., TANI, Y., SAKATA, M. & IWAHORI, K. 2007. Microbial manganese oxide formation and interaction with toxic metal ions. Journal of Bioscience and Bioengineering, 104, 1-8. MORALES, T. A., DOPSON, M., ATHAR, R. & HERBERT JR, R. B. 2005. Analysis of bacterial diversity in acidic pond water and compost after treatment of artificial acid mine drainage for metal removal. Biotechnology Bioengineering, 90, 543-551. MUYZER, G. & STAMS, A. J. 2008. The ecology and biotechnology of sulphate-reducing bacteria. Nature Reviews Microbiology, 6, 441-454. MUYZER, G., WAAL, E. D. & UITTERLINDEN, A. 1993. Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S rRNA. Applied and Environmental Microbiology, 59. NAIRN, R. W. & MERCER, M. N. 2000. Alkalinity generation and metals retention in a successive alkalinity producing system. Mine Water and the Environment, 19, 124-133. NECULITA, C.-M., ZAGURY, G. J. & BUSSIÈRE, B. 2007. Passive treatment of acid mine drainage in bioreactors using sulfate-reducing bacteria. Journal of Environmental Quality, 36, 1-16. NEVIN, K. P. & LOVLEY, D. R. 2002. Mechanisms for accessing insoluble Fe (III) oxide during dissimilatory Fe (III) reduction by Geothrix fermentans. Applied and Environmental Microbiology, 68, 2294-2299. NIYOGI, D. K., LEWIS JR, W. M. & MCKNIGHT, D. M. 2002. Effects of stress from mine drainage on diversity, biomass, and function of primary producers in mountain streams. Ecosystems, 5, 554-567. NIYOGI, D. K., MCKNIGHT, D. M. & LEWIS, W. M. 1999. Influences of water and substrate quality for periphyton in a montane stream affected by acid mine drainage. Limnology and Oceanography, 44, 804-809. NORDSTROM, D. K. & ALPERS, C. N. 1999. Negative pH, efflorescent mineralogy, and consequences for environmental restoration at the Iron Mountain Superfund site, California. Proceedings of the National Academy of Sciences, 96, 3455-3462. OPORTO, C., VANDECASTEELE, C. & SMOLDERS, E. 2007. Elevated cadmium concentrations in potato tubers due to irrigation with river water contaminated by mining in Potosí, Bolivia. Journal of Environmental Quality 36, 1181 - 1186. OVERBEEK, R., OLSON, R., PUSCH, G. D., OLSEN, G. J., DAVIS, J. J., DISZ, T., EDWARDS, R. A., GERDES, S., PARRELLO, B. & SHUKLA, M. 2013. The SEED and the Rapid Annotation of microbial genomes using Subsystems Technology (RAST). Nucleic Acids Research, 42, D206-D214. PALMER, K., RONKANEN, A.-K. & KLØVE, B. 2015. Efficient removal of arsenic, antimony and nickel from mine wastewaters in Northern treatment peatlands and potential risks in their long-term use. Ecological Engineering, 75, 350-364. PARKMAN, R., CURTIS, C., VAUGHAN, D. & CHARNOCK, J. 1996. Metal fixation and mobilisation in the sediments of the Afon Goch estuary—Dulas Bay, Anglesey. Applied Geochemistry, 11, 203-210. PARKS, D. H., TYSON, G. W., HUGENHOLTZ, P. & BEIKO, R. G. 2014. STAMP: statistical analysis of taxonomic and functional profiles. Bioinformatics, 30, 3123-3124. PÉDROT, M., LE BOUDEC, A., DAVRANCHE, M., DIA, A. & HENIN, O. J. 2011. How does organic matter constrain the nature, size and availability of Fe nanoparticles for biological reduction? Colloid and Interface Science 359, 75-85. PEINE, A. & PEIFFER, S. 1996. Neutralisation processes in acid mine lake sediments. Advances in Limnology, 261 - 267. PERRY, K. 1995. Sulfate‐reducing bacteria and immobilization of metals. Marine Georesources & Geotechnology, 13, 33-39. PERRY, S. F., SHAHSAVARANI, A., GEORGALIS, T., BAYAA, M., FURIMSKY, M. & THOMAS, S. L. 2003. Channels, pumps, and exchangers in the gill and kidney of freshwater fishes: their role in ionic and acid-base regulation. Journal of Experimental Zoology Part A: Comparative Experimental Biology, 300, 53-62. 235

PESTER, M., KNORR, K.-H., FRIEDRICH, M. W., WAGNER, M. & LOY, A. 2012. Sulfate-reducing microorganisms in wetlands–fameless actors in carbon cycling and climate change. Frontiers in Microbiology, 3. PETRI, R., PODGORSEK, L. & IMHOFF, J. F. 2001. Phylogeny and distribution of the sox B gene among thiosulfate-oxidizing bacteria. FEMS Microbiology Letters, 197, 171-178. PFEFFER, C., LARSEN, S., SONG, J., DONG, M., BESENBACHER, F., MEYER, R. L., KJELDSEN, K. U., SCHREIBER, L., GORBY, Y. A. & EL-NAGGAR, M. Y. 2012. Filamentous bacteria transport electrons over centimetre distances. Nature, 491, 218. POTYSZ, A., GRYBOS, M., KIERCZAK, J., GUIBAUD, G., FONDANECHE, P., LENS, P. N. & VAN HULLEBUSCH, E. D. 2017. Metal mobilization from metallurgical wastes by soil organic acids. Chemosphere, 178, 197- 211. PRUESSE, E., QUAST, C., KNITTEL, K., FUCHS, B. M., LUDWIG, W., PEPLIES, J. & GLÖCKNER, F. O. 2007. SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Research, 35, 7188-7196. PRUITT, K. D., TATUSOVA, T. & MAGLOTT, D. R. 2005. NCBI Reference Sequence (RefSeq): a curated non- redundant sequence database of genomes, transcripts and proteins. Nucleic Acids Research, 33, D501-D504. RAI, P. K. 2008. Heavy metal pollution in aquatic ecosystems and its phytoremediation using wetland plants: an ecosustainable approach. International Journal of Phytoremediation, 10, 133-160. RAINBOW, P. S. 2002. Trace metal concentrations in aquatic invertebrates: why and so what? Environmental Pollution, 120, 497-507. RANIERI, E. & YOUNG, T. M. 2012. Clogging influence on metals migration and removal in sub-surface flow constructed wetlands. Journal of Contaminant Hydrology, 129-130, 38-45. REIS, M. P., BARBOSA, F. A. R., CHARTONE-SOUZA, E. & NASCIMENTO, A. M. A. 2013. The prokaryotic community of a historically mining-impacted tropical stream sediment is as diverse as that from a pristine stream sediment. Extremophiles, 17, 301-309. RICE, E., BAIRD, R., EATON, A., CLESCERI, L. & BRIDGEWATER, L. E. 2012. Standard methods for the analysis of water and wastewater. Washington (DC): American Public Health Association. RODRÍGUEZ-GERMADE, I., RUBIO, B. & REY, D. 2014. XRF scanners as a quick screening tool for detecting toxic pollutant elements in sediments from Marín harbour in the Ría de Pontevedra (NW Spain). Marine Pollution Bulletin, 86, 458-467. RODRÍGUEZ-GERMADE, I., RUBIO, B., REY, D. & BORREGO, J. 2015. Detection and monitoring of REEs and related trace elements with an Itrax™ core scanner in the Ría de Huelva (SW Spain). Water, Air, & Soil Pollution, 226, 137. RODRÍGUEZ, L., RUIZ, E., ALONSO-AZCÁRATE, J. & RINCÓN, J. 2009. Heavy metal distribution and chemical speciation in tailings and soils around a Pb–Zn mine in Spain. Journal of Environmental Management, 90, 1106-1116. RODRÍGUEZ, Y., BALLESTER, A., BLÁZQUEZ, M. L., GONZÁLEZ, F. & MUÑOZ, J. A. 2003. Study of bacterial attachment during the bioleaching of pyrite, chalcopyrite, and sphalerite. Geomicrobiology Journal, 20, 131-141. ROGAN, N., SERAFIMOVSKI, T., DOLENEC, M., TASEV, G. & DOLENEC, T. 2009. Heavy metal contamination of paddy soils and rice (Oryza sativa L.) from Kočani Field (Macedonia). Environmental geochemistry and health, 31, 439-451. ROJAS, C., GUTIERREZ, R. M. & BRUNS, M. A. 2016. Bacterial and eukaryal diversity in soils forming from acid mine drainage precipitates under reclaimed vegetation and biological crusts. Applied Soil Ecology, 105, 57-66. ROTHWELL, R. G., HOOGAKKER, B., THOMSON, J., CROUDACE, I. W. & FRENZ, M. 2006. Turbidite emplacement on the southern Balearic Abyssal Plain (western Mediterranean Sea) during Marine Isotope Stages 1–3: an application of ITRAX XRF scanning of sediment cores to lithostratigraphic analysis. Geological Society, London, Special Publications, 267, 79-98. SAITOU, N. & NEI, M. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Molecular Biology and Evolution, 4, 406-425. SAND, W., GEHRKE, T., JOZSA, P.-G. & SCHIPPERS, A. 2001. (Bio) chemistry of bacterial leaching—direct vs. indirect bioleaching. Hydrometallurgy, 59, 159-175. SCHIPPERS, A. & SAND, W. 1999. Bacterial leaching of metal sulfides proceeds by two indirect mechanisms via thiosulfate or via polysulfides and sulfur. Applied and Environmental Microbiology, 65, 319-321. 236

SCHLOSS, P. D. & HANDELSMAN, J. 2005. Metagenomics for studying unculturable microorganisms: cutting the Gordian knot. Genome Biology 6, 229. SCHOLZ, M. & LEE, B. H. 2005. Constructed wetlands: a review. International Journal of Environmental Studies, 62, 421-447. SENKO, J. M., BERTEL, D., QUICK, T. J. & BURGOS, W. D. 2011. The influence of phototrophic biomass on Fe and S redox cycling in an acid mine drainage-impacted system. Mine Water and the Environment, 30, 38-46. SENKO, J. M., WANJUGI, P., LUCAS, M., BRUNS, M. A. & BURGOS, W. D. 2008. Characterization of Fe(II) oxidizing bacterial activities and communities at two acidic Appalachian coalmine drainage- impacted sites. ISME Journal, 2, 1134-1145. SEO, D. C., YU, K. & DELAUNE, R. D. 2008. Comparison of monometal and multimetal adsorption in Mississippi River alluvial wetland sediment: batch and column experiments. Chemosphere, 73, 1757-1764. SHARMA, P., OFNER, J. & KAPPLER, A. 2010. Formation of binary and ternary colloids and dissolved complexes of organic matter, Fe and As. Environmental Science & Technology, 44, 4479-4485. SHEORAN, A. & SHEORAN, V. 2006. Heavy metal removal mechanism of acid mine drainage in wetlands: a critical review. Minerals Engineering, 19, 105-116. SILVER, S. & PHUN, L. T. 1996. Bacterial heavy metal resistance: new surprises. Annual Reviews in Microbiology, 50, 753-89. SOBOLEV, D. & RODEN, E. E. 2001. Suboxic deposition of ferric iron by bacteria in opposing gradients of Fe (II) and oxygen at circumneutral pH. Applied and Environmental Microbiology, 67, 1328-1334. STEFFENS, W., MATTHEIS, T. & RIEDEL, M. 1993. Field observations on the production of rainbow trout (Oncorhynchus mykiss) under high concentrations of water-borne iron. Aquatic Sciences-Research Across Boundaries, 55, 173-178. STEUBER, J. & KRONECK, P. M. 1998. Desulfoviridin, the dissimilatory sulfite reductase from Desulfovibrio desulfuricans (Essex): new structural and functional aspects of the membranous enzyme. Inorganica Chimica Acta 275, 52-57. STOLTZ, E. & GREGER, M. 2002. Accumulation properties of As, Cd, Cu, Pb and Zn by four wetland plant species growing on submerged mine tailings. Environmental and Experimental Botany, 47, 271-280. SU, C. 2014. A review on heavy metal contamination in the soil worldwide: situation, impact and remediation techniques. Environmental Skeptics and Critics, 3, 24. SUN, M., XIAO, T., NING, Z., XIAO, E. & SUN, W. 2015a. Microbial community analysis in rice paddy soils irrigated by acid mine drainage contaminated water. Applied Microbiology and Biotecnology, 99, 2911-2922. SUN, W., XIAO, T., SUN, M., DONG, Y., NING, Z., XIAO, E., TANG, S. & LI, J. 2015b. Diversity of the Sediment Microbial Community in the Aha Watershed (Southwest China) in Response to Acid Mine Drainage Pollution Gradients. Applied and Environmental Microbiology, 81, 4874-4884. TANG, Y.-S., WANG, L., JIA, J.-W., FU, X.-H., LE, Y.-Q., CHEN, X.-Z. & SUN, Y. 2011. Response of soil microbial community in Jiuduansha wetland to different successional stages and its implications for soil microbial respiration and carbon turnover. Soil Biology and Biochemistry, 43, 638-646. TARDY, V., CASIOT, C., FERNANDEZ-ROJO, L., RESONGLES, E., DESOEUVRE, A., JOULIAN, C., BATTAGLIA- BRUNET, F. & HÉRY, M. 2018. Temperature and nutrients as drivers of microbially mediated arsenic oxidation and removal from acid mine drainage. Applied Microbiology and Biotechnology, 102, 2413-2424. TARUTIS, W. J., UNZ, R. F. & BROOKS, R. P. 1992. Behavior of sedimentary Fe and Mn in a natural wetland receiving acidic mine drainage, Pennsylvania, USA. Applied Geochemistry, 7, 77-85. TEBO, B. M., BARGAR, J. R., CLEMENT, B. G., DICK, G. J., MURRAY, K. J., PARKER, D., VERITY, R. & WEBB, S. M. 2004. Biogenic manganese oxides: properties and mechanisms of formation. Annual Review of Earth and Planetary Sciences, 32, 287-328. TEBO, B. M. & OBRAZTSOVA, A. Y. 1998. Sulfate-reducing bacterium grows with Cr (VI), U (VI), Mn (IV), and Fe (III) as electron acceptors. FEMS Microbiology Letters, 162, 193-199. TEUCHIES, J., JACOBS, S., OOSTERLEE, L., BERVOETS, L. & MEIRE, P. 2013. Role of plants in metal cycling in a tidal wetland: implications for phytoremidiation. Science of the Total Environment, 445-446, 146- 54.

237

TJALLINGII, R., RÖHL, U., KÖLLING, M. & BICKERT, T. 2007. Influence of the water content on X‐ray fluorescence core‐scanning measurements in soft marine sediments. Geochemistry, Geophysics, Geosystems, 8. TOSATTO, S. C., TOPPO, S., CARBONERA, D., GIACOMETTI, G. M. & COSTANTINI, P. 2008. Comparative analysis of [FeFe] hydrogenase from Thermotogales indicates the molecular basis of resistance to oxygen inactivation. International Journal of Hydrogen Energy 33, 570-578. TOURNA, M., MACLEAN, P., CONDRON, L., O'CALLAGHAN, M. & WAKELIN, S. A. 2014. Links between sulphur oxidation and sulphur-oxidising bacteria abundance and diversity in soil microcosms based on soxB functional gene analysis. FEMS Microbiology Ecology, 88, 538-549. TREUDE, N., ROSENCRANTZ, D., LIESACK, W. & SCHNELL, S. 2003. Strain FAc12, a dissimilatory iron-reducing member of the Anaeromyxobacter subgroup of Myxococcales. FEMS Microbiology Ecology 44, 261- 269. TRIBUTSCH, H. 2001. Direct versus indirect bioleaching. Hydrometallurgy, 59, 177-185. TRIPATHI, R. D., TRIPATHI, P., DWIVEDI, S., KUMAR, A., MISHRA, A., CHAUHAN, P. S., NORTON, G. J. & NAUTIYAL, C. S. 2014. Roles for root iron plaque in sequestration and uptake of heavy metals and metalloids in aquatic and wetland plants. Metallomics, 6, 1789-1800. TUTU, H., MCCARTHY, T. & CUKROWSKA, E. 2008. The chemical characteristics of acid mine drainage with particular reference to sources, distribution and remediation: The Witwatersrand Basin, South Africa as a case study. Applied Geochemistry, 23, 3666-3684. TWARDOWSKA, I. & KYZIOL, J. 2003. Sorption of metals onto natural organic matter as a function of complexation and adsorbent–adsorbate contact mode. Environment International, 28, 783-791. VALKANAS, M. M. & TRUN, N. J. 2018. A seasonal study of a passive abandoned coalmine drainage remediation system reveals three distinct zones of contaminant levels and microbial communities. MicrobiologyOpen, e00585. VERA, M., SCHIPPERS, A. & SAND, W. 2013. Progress in bioleaching: fundamentals and mechanisms of bacterial metal sulfide oxidation—part A. Applied Microbiology and Biotechnology, 97, 7529-7541. VILE, M. A. & WIEDER, R. K. 1993. Alkalinity generation by Fe (III) reduction versus sulfate reduction in wetlands constructed for acid mine drainage treatment. Water, Air and Soil Pollution: Focus, 69, 425-441. VINODHINI, R. & NARAYANAN, M. 2008. Bioaccumulation of heavy metals in organs of fresh water fish Cyprinus carpio (Common carp). International Journal of Environmental Science & Technology, 5, 179-182. VOELKER, B. M. & SULZBERGER, B. 1996. Effects of fulvic acid on Fe (II) oxidation by hydrogen peroxide. Environmental Science & Technology, 30, 1106-1114. VOLANT, A., BRUNEEL, O., DESOEUVRE, A., HÉRY, M., CASIOT, C., BRU, N., DELPOUX, S., FAHY, A., JAVERLIAT, F., BOUCHEZ, O., DURAN, R., BERTIN, P. N., ELBAZ-POULICHET, F. & LAUGA, B. 2014. Diversity and spatiotemporal dynamics of bacterial communities: physicochemical and other drivers along an acid mine drainage. FEMS Microbiology Ecology, 90, 247-263. VON ARB, C. & BRUNOLD, C. 1985. Ferredoxin‐sulfite reductase and ferredoxin‐nitrite reductase activities in leaves of Pisum sativum: Changes during ontogeny and in vitro regulation by sulfide. Physiologia Plantarum, 64, 290-294. VYMAZAL, J., ŠVEHLA, J., KRÖPFELOVÁ, L. & CHRASTNÝ, V. 2007. Trace metals in Phragmites australis and Phalaris arundinacea growing in constructed and natural wetlands. Science of the Total Environment, 380, 154-162. WALTON, K. C. & JOHNSON, D. B. 1992. Microbiological and chemical characteristics of an acidic stream draining a disused copper mine Environmental Pollution, 76, 169-75. WANG, H., GUO, C., YANG, C., LU, G., CHEN, M. & DANG, Z. 2016. Distribution and diversity of bacterial communities and sulphate‐reducing bacteria in a paddy soil irrigated with acid mine drainage. Journal of Applied Microbiology, 121, 196-206. WANG, H., ZENG, Y., GUO, C., BAO, Y., LU, G., REINFELDER, J. R. & DANG, Z. 2018. Bacterial, archaeal, and fungal community responses to acid mine drainage-laden pollution in a rice paddy soil ecosystem. Science of The Total Environment, 616-617, 107-116. WANG, L., YAN, B., LIU, N., LI, Y. & WANG, Q. 2008. Effects of cadmium on glutathione synthesis in hepatopancreas of freshwater crab, Sinopotamon yangtsekiense. Chemosphere, 74, 51-6.

238

WANG, M., CHEN, A., WONG, M., QIU, R., CHENG, H. & YE, Z. J. E. P. 2011. Cadmium accumulation in and tolerance of rice (Oryza sativa L.) varieties with different rates of radial oxygen loss. Environmental Pollution, 159, 1730-1736. WEBB, J., MCGINNESS, S. & LAPPIN-SCOTT, H. 1998. Metal removal by sulphate-reducing bacteria from natural and constructed wetlands. Journal of Applied Microbiology, 82, 240-8. WEBER, K. A., ACHENBACH, L. A. & COATES, J. D. 2006. Microorganisms pumping iron: anaerobic microbial iron oxidation and reduction. Nature Reviews Microbiology, 4, 752. WEBSTER, J. G., SWEDLUND, P. J. & WEBSTER, K. S. 1998. Trace Metal Adsorption onto an Acid Mine Drainage Iron(III) Oxy Hydroxy Sulfate. Environmental Science & Technology, 32, 1361-1368. WEIS, J. S. & WEIS, P. 2004. Metal uptake, transport and release by wetland plants: implications for phytoremediation and restoration. Environment International, 30, 685-700. WERNER, J. J., KOREN, O., HUGENHOLTZ, P., DESANTIS, T. Z., WALTERS, W. A., CAPORASO, J. G., ANGENENT, L. T., KNIGHT, R. & LEY, R. E. 2011. Impact of training sets on classification of high-throughput bacterial 16s rRNA gene surveys. ISME Journal, 6, 94. WHITEHEAD, P. G. & PRIOR, H. 2005. Bioremediation of acid mine drainage: an introduction to the Wheal Jane wetlands project. Science of The Total Environment, 338, 15-21. WHITELEY, J. D. & PEARCE, N. J. G. 2003. Metal distribution during diagenesis in the contaminated sediments of Dulas Bay, Anglesey, N. Wales, UK. Applied Geochemistry, 18, 901-913. WIEDER, R. K., LINTON, M. N. & HESTON, K. P. 1990. Laboratory mesocosm studies of Fe, Al, Mn, Ca, and Mg dynamics in wetlands exposed to synthetic acid coal mine drainage. Water, Air, and Soil Pollution, 51, 181-196. WIESSNER, A., KUSCHK, P., NGUYEN, P. M. & MÜLLER, J. A. 2017. The sulfur depot in the rhizosphere of a common wetland plant, Juncus effusus, can support long-term dynamics of inorganic sulfur transformations. Chemosphere, 184, 375-383. WILDE, K. L., STAUBER, J. L., MARKICH, S. J., FRANKLIN, N. M. & BROWN, P. L. 2006. The effect of pH on the uptake and toxicity of copper and zinc in a tropical freshwater alga (Chlorella sp.). Archives of Environmental Contamination and Toxicology, 51, 174-85. WINTERBOURN, M. J. 1998. Insect faunas of acidic coal mine drainages in Westland, New Zealand. New Zealand Entomologist, 21, 65-72. WU, S., KUSCHK, P., WIESSNER, A., MÜLLER, J., SAAD, R. A. & DONG, R. 2013. Sulphur transformations in constructed wetlands for wastewater treatment: a review. Ecological Engineering, 52, 278-289. YANG, C., LU, G., CHEN, M., XIE, Y., GUO, C., REINFELDER, J., YI, X., WANG, H. & DANG, Z. 2016. Spatial and temporal distributions of sulfur species in paddy soils affected by acid mine drainage in Dabaoshan sulfide mining area, South China. Geoderma, 281, 21-29. YANG, J., MA, Z., YE, Z., GUO, X. & QIU, R. 2010. Heavy metal (Pb, Zn) uptake and chemical changes in rhizosphere soils of four wetland plants with different radial oxygen loss. Journal of Environmental Sciences, 22, 696-702. YANG, J., TAM, N. F.-Y. & YE, Z. 2014. Root porosity, radial oxygen loss and iron plaque on roots of wetland plants in relation to zinc tolerance and accumulation. Plant and Soil, 374, 815-828. YANG, N., HAFEZ, H. & NAKHLA, G. 2015. Impact of volatile fatty acids on microbial electrolysis cell performance. Bioresource Technology 193, 449-455. YARZÁBAL, A., BRASSEUR, G., RATOUCHNIAK, J., LUND, K., LEMESLE-MEUNIER, D., DEMOSS, J. A. & BONNEFOY, V. 2002. The high-molecular-weight cytochrome c Cyc2 of Acidithiobacillus ferrooxidans is an outer membrane protein. Journal of Bacteriology, 184, 313-317. YOUNGER, P. & POTTER, H. A. B. 2012. Parys in Springtime: Hazard Management and Steps Towards Remediation of the UK's Most Polluted Acidic Mine Discharge. In 9th International Conference on Acid Rock Drainage (ICARD); Ottawa, Canada. YU, R.-L., TAN, J.-X., PENG, Y., JING, S., OUYANG, X.-J. & DAI, Y.-J. 2008. EPS-contact-leaching mechanism of chalcopyrite concentrates by A. ferrooxidans. Transactions of Nonferrous Metals Society of China, 18, 1427-1432. ZACHARA, J. M., FREDRICKSON, J. K., SMITH, S. C. & GASSMAN, P. L. 2001. Solubilization of Fe (III) oxide- bound trace metals by a dissimilatory Fe (III) reducing bacterium. Geochimica et Cosmochimica Acta 65, 75-93. ZHANG, J., KOBERT, K., FLOURI, T. & STAMATAKIS, A. 2013. PEAR: a fast and accurate Illumina Paired-End reAd mergeR. Bioinformatics, 30, 614-620. 239

ZHUANG, P., MCBRIDE, M. B., XIA, H., LI, N. & LI, Z. 2009. Health risk from heavy metals via consumption of food crops in the vicinity of Dabaoshan mine, South China. Science of the Total Environment, 407, 1551-1561. ZIEMKIEWICZ, P. F., SKOUSEN, J. G. & SIMMONS, J. 2003. Long-term performance of passive acid mine drainage treatment systems. Mine Water and the Environment, 22, 118-129. ŽIFČÁKOVÁ, L., VĚTROVSKÝ, T., HOWE, A. & BALDRIAN, P. 2016. Microbial activity in forest soil reflects the changes in ecosystem properties between summer and winter. Environmental Microbiology, 18, 288-301.

240