Citation
Elucidating the impact of metals on the metabolic profiles and taxonomic structure of bacterial communities within the Chattanooga fen

Material Information

Title:
Elucidating the impact of metals on the metabolic profiles and taxonomic structure of bacterial communities within the Chattanooga fen
Creator:
Foster, Kelsey
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English

Thesis/Dissertation Information

Degree:
Master's ( Master of science)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Integrative Biology, CU Denver
Degree Disciplines:
Biology
Committee Chair:
Roane, Timberley M.
Committee Members:
Mosier, Annika C.
Phiel, Christopher

Notes

Abstract:
Fens, marshy wetlands with an accumulation of partially decomposed organic matter known as peat, are estimated to store up to a third of the Earth’s soil organic carbon. The Chattanooga Fen, located in the San Juan National Forest in southwestern Colorado, is anthropogenically and endogenously impacted by elevated concentrations of toxic metals. The anthropogenic source of metals to the Fen is Acid Mine Drainage (AMD), acidic metal-rich effluent formed by oxidized metals from the nearby Gold Finch Mine. The endogenous metal impact on the Fen is a result of subsurface acid rock drainage (ARD) formed from the natural oxidation of sulfur-bearing minerals. While the entirety of the Fen is naturally impacted by ARD-associated metals, the geographical terrain bisects the Fen into AMD-affected and AMD-unaffected portions. The chemical profiles of ARD and AMD are distinctly different from each other in terms of dissolved metal concentrations, pH, and anions. This study aimed to elucidate the impact of AMD in shaping the taxonomic diversity and carbon usage of the microbial communities within this rare, uncharacterized Fen. Sediment cores and water samples were collected in July and September 2016 at the outflow of the Gold Finch mine, along an AMD-gradient, and from naturally metal-impacted sediments unimpacted by AMD. Illumina high-throughput sequencing of extracted 16S rDNA from sediments provided the taxonomic structure of the bacterial communities within the different areas of the Fen. Carbon source usage was evaluated via BIOLOG EcoPlates and qPCR of three glycoside hydrolase (GH) genes. Data suggests that differences in both taxonomic structure and carbon source utilization can be explained by the presence of AMD. Sediments unaffected by AMD had a higher relative abundance of Acidobacteria, while Deltaproteobacteria was dominant in AMD-affected sediments. The communities within the AMD-affected sediments also utilized a higher proportion of carboxylic acids on the culture-based EcoPlates, while the unaffected samples displayed a greater usage of carbohydrates in terms of resulting optical density values. However, there were two commonalities between AMD-affected and unaffected sediments which was the use of xylose and cellobiose. The average number of carbon sources utilized by unaffected samples was 35.5% greater than affected; the lowest number of substrates being used was observed with the direct acid mine effluent samples. The abundance of all three GH genes was also greater in all AMD-unaffected samples compared to AMD-affected samples. The understanding of how metal source impacts the diversity and functional potential of microorganisms within the Chattanooga Fen, and other similar fen systems, will be essential in influencing future management decisions.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
Copyright Kelsey Foster. Permission granted to University of Colorado Denver to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

Downloads

This item has the following downloads:


Full Text
ELUCIDATING THE IMPACT OF METALS ON THE METABOLIC PROFILES AND
TAXONOMIC STRUCTURE OF BACTERIAL COMMUNITIES WITHIN THE CHATTANOOGA
FEN
by
KELSEY FOSTER
B.S., University of Colorado Denver, 2014
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Science Biology Program
2017


©2017
KELSEY FOSTER
ALL RIGHTS RESERVED


This thesis for the Master of Science degree by Kelsey Foster has been approved for the Biology Program
by
Timberley M. Roane, Chair Annika C. Mosier Christopher Phiel
December 16, 2017


Foster, Kelsey (M.S., Biology Program)
Elucidating the Impact of Metals on the Metabolic Profiles and Taxonomic Structure of Bacterial Communities within the Chattanooga Fen
Thesis directed by Associate Professor Timberley M. Roane
ABSTRACT
Fens, marshy wetlands with an accumulation of partially decomposed organic matter known as peat, are estimated to store up to a third of the Earth’s soil organic carbon. The Chattanooga Fen, located in the San Juan National Forest in southwestern Colorado, is anthropogenically and endogenously impacted by elevated concentrations of toxic metals. The anthropogenic source of metals to the Fen is Acid Mine Drainage (AMD), acidic metal-rich effluent formed by oxidized metals from the nearby Gold Finch Mine. The endogenous metal impact on the Fen is a result of subsurface acid rock drainage (ARD) formed from the natural oxidation of sulfur-bearing minerals. While the entirety of the Fen is naturally impacted by ARD-associated metals, the geographical terrain bisects the Fen into AMD-affected and AMD-unaffected portions. The chemical profiles of ARD and AMD are distinctly different from each other in terms of dissolved metal concentrations, pH, and anions. This study aimed to elucidate the impact of AMD in shaping the taxonomic diversity and carbon usage of the microbial communities within this rare, uncharacterized Fen. Sediment cores and water samples were collected in July and September 2016 at the outflow of the Gold Finch mine, along an AMD-gradient, and from naturally metal-impacted sediments unimpacted by AMD. Illumina
IV


high-throughput sequencing of extracted 16S rDNA from sediments provided the taxonomic structure of the bacterial communities within the different areas of the Fen. Carbon source usage was evaluated via BIOLOG EcoPlates and qPCR of three glycoside hydrolase (GH) genes. Data suggests that differences in both taxonomic structure and carbon source utilization can be explained by the presence of AMD. Sediments unaffected by AMD had a higher relative abundance of Acidobacteria, while Deltaproteobacteria was dominant in AMD-affected sediments. The communities within the AMD-affected sediments also utilized a higher proportion of carboxylic acids on the culture-based EcoPlates, while the unaffected samples displayed a greater usage of carbohydrates in terms of resulting optical density values. However, there were two commonalities between AMD-affected and unaffected sediments which was the use of xylose and cellobiose. The average number of carbon sources utilized by unaffected samples was 35.5% greater than affected; the lowest number of substrates being used was observed with the direct acid mine effluent samples. The abundance of all three GH genes was also greater in all AMD-unaffected samples compared to AMD-affected samples. The understanding of how metal source impacts the diversity and functional potential of microorganisms within the Chattanooga Fen, and other similar fen systems, will be essential in influencing future management decisions.
Approved: Timberley M. Roane
v


ACKNOWLEDGEMENTS
There are many people I would like to thank for their involvement in this project. First off, thank you to my advisor Dr. Timberley Roane for not only giving me the opportunity to work in her lab and her guidance every step of the way, but for her continued support and kindness. She opened my eyes to the wonderful world of microbiology and introduced me to so many new opportunities, all the while setting an amazing example of howto conduct yourself as a woman in the scientific field.
Thank you to my committee members, Dr. Annika Mosier and Dr. Christopher Phiel for all of your assistance throughout this experience. Dr. Phiel gave me my first experience working in the lab, and I could not be more grateful. Dr. Mosier has been more than willing to go out of her way to advise me on so many lab techniques and has also become someone I greatly admire.
I would also like to thank Jeff Boon for all of your assistance in the metal analysis. I appreciate all of the time you spent assisting in the project.
Thank you to Sladjana Subotic and Andrew Boddicker for not only helping me through so many tough days in the lab, but for continuing to be amazing friends to me during some of the toughest periods of my life. The both of you have become my family, and I will never be able to put into words what your friendships means to me. I also appreciate you both along with Pedja Stojanovic for baring though the freezing cold to collect samples with me, all the while keeping me laughing even though I could not feel my hands.
VI


Thank you to my parents, sister, brother, nephews, and nieces for continuing to motivate me throughout my collegiate career. I could not have done any of this without you all in my corner. You have made every step worth-while, and I hope I have made you proud. You all are the reason I have become the person I am today.
To Matthew Wayne, Sheldon Herman, Lucy Diamond, and Kyrie Jane, thank you for taking care of me through everything. You have given me more love and support than I ever could have imagined. I love you with everything I have.
Finally, thank you to all of my friends and colleagues who have all helped me keep pushing through in one way or another. Munira Lantz, Anna Scopp, Bhargavi Ramanathan, and Gabrielle Rietz, I will always be thankful for your friendship. You have all helped me in everything from teaching techniques to qPCR to reminding me that laughter is always the best medicine. You are wonderful human beings, and I hope you know what a large part you played in my graduate career.
VII


Table of Contents
CHAPTER
I. PROJECT OVERVIEW AND OBJECTIVES...........................................1
II. THE CHATTANOOGA FEN......................................................3
Sources of Metal Impact....................................................5
Impact and Toxicity of Metals on Bacteria...............................10
Microbiology of Peatland Systems..........................................12
III. METHODS BACKGROUND.....................................................16
16S rDNA High-Throughput Sequencing.......................................16
Metabolic Profiling: BIOLOG EcoPlates.....................................21
Quantification of Glycoside Hydrolase Genes...............................23
IV. CONDUCTED STUDY........................................................26
Introduction..............................................................26
Methods...................................................................27
Sample Collection.......................................................27
DNA Extraction, PCR Amplification, and Next-Generation Sequencing.......29
Computational Analyses..................................................31
BIOLOG EcoPlate Preparation and Reading.................................31
DNA Extraction and Amplification from EcoPlates.........................32
viii


Quantitative PCR.....................................................33
Results and Discussion.................................................34
Site Chemistry.......................................................34
Taxonomic Structure of Bacterial Communities within the Chattanooga Fen.39
Carbon Source Usage..................................................51
Quantitative PCR.....................................................57
V. CONCLUSIONS AND FUTURE DIRECTIONS.....................................63
SUPPLEMENTAL INFORMATION.................................................66
REFERENCES...............................................................71
IX


CHAPTER I
PROJECT OVERVIEW AND OBJECTIVES
The Chattanooga Fen, located in the Colorado Mineral Belt is both endogenously and anthropogenically metal-impacted, providing a rare opportunity to examine the impact of AMD on a naturally metal- impacted site. A natural geological structure, the Chattanooga Fen is formed by acidic, metal-laden groundwater. The natural metal impact within the Fen is known as acid rock drainage (ARD) formed from the oxidation of sulfur bearing minerals within the bedrock beneath the ponds formed at the Fen. The Chattanooga Fen is also bisected by the Gold Finch Mine, creating acid mine drainage (AMD) affected and unaffected regions. No longer active, acid mine drainage from the Mine continues to introduce additional metals to the affected areas of the Fen. Whether endogenous (ARD) or anthropogenic (AMD), metals in the environment pose serious ecological risk and may affect a variety of cellular processes (Nies, 1999). However, the response of endogenous ARD, metal-adapted microbial communities to AMD exposure within this ecosystem have yet to be studied. The presented study aims to evaluate the impact of AMD from the Gold Finch Mine on the indigenous metal- adapted microbial communities. In order to characterize the bacterial community taxonomic structure, alpha and beta diversity metrices were performed on the resulting sequences obtained from sediment samples. Metabolic profiling was used to inform the use of quantitative PCR to identify differences in the expression of three glycoside hydrolase genes, as indicators of AMD impact. The specific objectives of the study conducted were to:
1


• Objective 1: Determine the taxonomy of bacterial communities within the Chattanooga Fen from ARD impacted sediment unaffected by AMD and affected by AMD
• Objective 2: Examine the carbon source usage of the bacterial communities within the Fen
• Objective 3: Taxonomically characterize which bacteria were utilizing common carbon sources (xylose and cellobiose) between AMD- affected and AMD-unaffected sediment communities within an ARD impacted site
• Objective 4: Quantify the amount of three glycoside hydrolase genes present in AMD-affected and AMD-unaffected ARD samples which encode enzymes necessary to degrade xylose and cellobiose
2


CHAPTER II
THE CHATTANOOGA FEN
Fens are oligotrophic, minerotrophic peatlands, which are groundwater-fed, nutrient-poor marshy wetlands characterized by an accumulation of partially decomposed organic matter known as peat (Lin et al., 2012). Peatland environments contribute about 20% of the total annual methane emissions to the atmosphere. These unique peatland ecosystems are large carbon sinks, containing about one-third of all soil organic carbon (SOC) in the world, despite only covering around 5% of the Earth’s surface (Andersen etal., 2013). This can be explained by the imbalance between net primary production by plants and decomposition by microorganisms due to the consistently water saturated environment and consequent anoxia. The Chattanooga Fen, part of the Colorado Mineral Belt, is located in southwestern Colorado near the old township of Chattanooga (Figure 2.1).
Figure 2.1. Map depicting the location of the fen in the San Juan Mountains between Ouray and Silverton (left) and a topographical map of the Chattanooga Fen with an arrow depicting the location of the ponds where sediment samples were taken in the conducted study
3


This largest Fen in the San Juan Mountains sits at an elevation of 10,500 feet located downhill from the Gold Finch Mine, and is fed by the Animas River Watershed (Cooper et al., 2008) (Figure 2.2). Through carbon dating, this Fen is predicted to be around 600 years old with up to 3 meters of accumulated peat. The environment is now home to a high diversity of plant species including the rare arctic peat moss, Sphagnum balticum, even though its main range is in Canada, more than 2,000 km away (Cooper et al., 2008). The water which saturates the soil in the Chattanooga Fen, as well as the pooled waters which form the ponds, has a net acidity, with a pH ranging from 3.8 to 6.4, and is influenced by the bedrock with which the groundwater comes into contact (Chimner and Cooper, 2006; Sackett, 2015). This is a trait unique to the Chattanooga Fen. Most fens are non-iron fens and are pH neutral to alkaline. Bogs, another type of wetland, are traditionally acidic, though they are fed by rainwater. Acidic, iron fens are unique to Colorado and form in areas where the groundwater that feeds them has a naturally low pH due to weathering of iron pyrite (thus the name "iron fens”) and its oxidation to sulfuric acid.
An inactive mine, the Gold Finch Mine, is located above the Chattanooga Fen. The Gold Finch Mine is approximately 140 years old and was once used for mining gold and molybdenum (Stanton et. al, 2008). Below the mine, a portal made of concrete was built, and acid mine drainage, originating from within the mine, exits the mine and enters into a portion of the Fen. The Chattanooga Fen is also naturally metal-impacted in its entirety by acid rock drainage (ARD) which has formed from the dissolution of bedrock which has come into contact with the acidic groundwater source (Figure 2.2). Due to the rarity of iron-fens, as well as the metal contamination from both
4


anthropogenic (the Gold Finch Mine) and endogenous sources, the Chattanooga Fen is a prime opportunity for the following comparative study into the effect of AMD on previously ARD-impacted ecosystems.
Figure 2.2. The constructed portal below the inactive Gold Finch Mine above the Chattanooga Fen where the copper- colored acid mine drainage enters portions of the Fen
Sources of Metal Impact
Metal contamination in the sediments of the Chattanooga Fen arose from both endogenous and anthropogenic sources. The anthropogenic source is a result of mining from the Gold Finch Mine which resulted in acid mine drainage (AMD). AMD is the formation and movement of acidic water rich in heavy metals (Johnson and Hallberg, 2003). AMD in the Chattanooga Fen resulted from mining practices exposing pyrite, an iron sulfide known as Fool’s Gold, which reacted with water and air to form sulfuric acid and dissolved iron via a series of oxidation reactions (Baker and Banfield, 2003). This reaction is summarized as:
2FeS2(s) + 702(g) + 2FkO(l) = 2Fe2+(aq) + 4S042-(aq) + 4H+(aq)
5


Metal sulfide oxidation resulted in increased acidity, which resulted in dissolution of metal sulfides. Although metals tend to become more soluble with decreasing pH, ferric iron (Fe3+) can precipitate out of solution at pH values between 2.3 and 3.5 as iron hydroxide (Fe(OH)3) or jarosite (KFe3+3(0H)6(S04)2), resulting in the copper to orange-colored precipitate which can be observed at many AMD sites, including the Chattanooga Fen (Figure 2.2) (Verplanck, 2008). AMD has a net acidity which can further dissolve exposed ore material, releasing metals into the environment. AMD can be generated at both abandoned and active mines, including the Gold Finch Mine, and is a serious source of water pollution that can kill aquatic life, restrict stream use, and damage water supplies. As of 2008, the U.S. Environmental Protection Agency has estimated that in Colorado alone, approximately 1,300 miles of streams are affected by AMD; the negative impact of which on wildlife has been well documented (Barry et al., 2000; David, 2003; Savinov et al., 2003; Verplanck, 2008).
A natural, or endogenous, source of metals to the Chattanooga Fen is known as acid rock drainage (ARD). This is also acidic water rich in metals, but this type is formed over time from the natural oxidation or dissolution of sulfide minerals exposed to water and air as a result of endogenous processes (Hogsden and Harding, 2012). This source of natural metal impact can be observed in many Colorado watersheds, including the Animas River Watershed (Hauff et al., 2003). The pools of water within the Chattanooga Fen are impacted by ARD via the welling of naturally acidic groundwater through metal-rich sedimentary rocks which form the bedrock below the peat formation (Chimner and Cooper, 2007) (Figure 2.3). Although AMD is often a more significant concern than ARD due to the increased surface area of minerals exposed to weathering
6


and oxidative conditions, both AMD and ARD result in the introduction of metal-rich,
acidic effluent into the Fen. While metals are ubiquitous in nature, the Chattanooga Fen is unique in that it is impacted by ARD throughout as well as AMD in portions which receive mine effluent from the inactive Gold Finch Mine.
Fens receive both surface and subsurface
outflows. As a result, fens tend to reflect the chemistry of the underlying geology. The naturally acidic water within the Chattanooga Fen dissolves parents rock material and forms Acid Rock Drainage.
Acid Mine Drainage
Here's a look at what AMD Is and how it affects
the surrounding environment.
©During mining, pyrite is exposed to oxygen.
©Ground water seeps into the mine.
© Oxygen, water and pyrite react to form suHuric acid and in turn dissolve metals from the rocks.
© Water drains out of the mine.
©> Dissolved metals react with oxygen and fall out of solution into the stream water, turning a bright color.
@1 Aquatic animals and plants are killed by the drainage.
A -*
Source: Western Pennsyk/ania Conservancy
Steve Thomas/Post-Gazette
Figure 2.3. Schematic of the formation or ARD (left) and AMD (right). The Chattanooga Fen is impacted by both ARD from the naturally acidic groundwater which feeds the Fen as well as AMD because of the inactive Gold Finch Mine
The Chattanooga Fen is bisected by the terrain into areas affected and unaffected by AMD, while the entirety is metal-impacted via the acidic ground waters which feed the Fen (Figure 2.4). The sediments near AMD or ARD contaminated sites are often a repository for chemicals such as metals, which is why they were the samples taken for this study. Once metals enter the sediment, they cannot be degraded, and therefore, they are a persistent environmental hazard (Bernhardt et al., 2012; Savinov et al.,
2003). The long- term impact of a high concentration of metals, from either AMD or ARD, on microbial community structure is not well understood. Metals within the sediments of the Chattanooga Fen force microorganisms to adapt to the presence of metals, noting that microorganisms are more responsive to changing conditions than
7


plants and animals in the same area due to their high surface area to volume ratio and metabolic flexibility (Giller et al., 1998).
It should be noted that the entirety of the Chattanooga Fen is ARD-impacted. However, only portions of the Fen are AMD-impacted. Therefore, in the current study affected samples refer to AMD-affected ARD sediment samples, and unaffected samples refer to AMD-unaffected ARD sediment samples.
8


Affected
PondC
Affected Pond B
Constructed ' Cold Finch Mine Portal
Affected Pond A
AMD Effluent
Affected Pond 1
Affected Pond 2
Figure 2.4. Diagram of the terrain of the Chattanooga Fen. The endogenously metal-impacted Fen is bisected into portions affected and unaffected by AMD where sediment samples were taken as seen in the bottom photograph of the diagram
9


Impact and Toxicity of Metals on Bacteria
As in other metal-impacted systems, metals from both AMD and ARD in the Chattanooga Fen are expected to be toxic to cells. Once metals enter the cell via passive transport proteins or substrate specific transporters driven by ATP hydrolysis, the cell can be severely damaged. This is due to the strong ionic nature of metals which allows them to bind many cellular ligands having a bactericidal or bacteriostatic effect on the cell (Banfalvi, 2011). Metals can damage nucleic acids, disrupt protein structure and function, induce genetic mutations, inhibit membrane fluidity and function, and induce oxidative stress (Mikiya, 1992; Silver and Phung, 2009). It has been shown that morphology can also be affected, and certain species have been seen to convert from rod shaped (bacilli) to spheres (cocci) (Rouch etal., 1995). This result suggests that the processes which regulate cell division and cell wall synthesis may be affected.
In response to metals in their environment, bacteria have developed several metal resistance/tolerance mechanisms in order to keep metal concentrations below a physiologically toxic level (Figure 2.5). Possible mechanisms developed by bacteria in the Chattanooga Fen include extracellular metal exclusion mechanisms and intracellular metal reduction mechanisms. The genes encoding these mechanisms are often found on plasmids and can be transferred via horizontal gene transfer, or resistance can be obtained through mutations to the genome (Nies, 1999)
10


Figure 2.5. Metal resistance mechanisms of microorganisms. Extracellular metal exclusion mechanisms are denoted by a, b, c; Intracellular metal removal mechanisms are denoted by d and e; Intracellular metal reactivity reduction denoted by f and g (Gadd, 2010)
Extracellular mechanisms include (Figure 2.5a, b, and c):
• Extracellular polysaccharide substance (EPS) production: Groups of bacteria can produce sticky EPS in the formation of biofilms and they are less susceptible to metals entering the cell
• Metal chelation: Metal chelating compounds will bind metal ions and prevent diffusion into the cell (e.g. siderophores)
• Precipitation: Mineralization of metals as salts and reduced metals
11


Intracellular mechanisms include (Figure 2.5d, e, f, and g):
• Metallothionein production: Metallothioneins are metal binding stress response proteins that protect against elevated levels of metals within the cell
• Efflux systems: Efflux is the active transport of metal cations out of the cytoplasm in order to reduce intracellular metal concentrations
• Enzymatic detoxification: The utilization of reduction-oxidation mechanisms to decrease chemical reactivity of metals within the cell
• Volatilization: Reduces cytosolic bioavailable concentrations of metals via enzymatic reduction
Microbiology of Peatland Systems
Despite the importance of peatland ecosystems in the global carbon cycle, no known microbiological analyses have been performed at the Chattanooga Fen. In fact, the microbial composition of iron-fens and fens in general remains greatly understudied. Culture-dependent methods have traditionally been used to examine microbial communities within these ecosystems, however, this method of identification requires that organisms were cultivable and easily identifiable in the laboratory. Many previous studies on fens and specifically iron-fens in southwestern Colorado have
12


utilized culturing techniques to explore the microorganismal additions to the formation of AMD and ARD (Baath and Anderson, 2003; Preston et al. 2012; Stanton et al., 2008).
In previous studies identifying dominate bacterial species with non-iron fens, nearly one half of all of the colonies grown on nutrient agar plates after one to two weeks of incubation were members of the fast growing Betaproteobacteria, including many from the genus Burkholderia (Gilbert and Mitchell, 2006; Reiche et ah, 2008). Other colonies commonly cultured from non-metal impacted peatland environments included organisms belonging to Alphaproteobacteria, Actinobacteria, Bacteriodetes, and Gammaproteobacteria (Gilbert and Mitchell, 2006; Reiche et ah, 2008; Loy et ah, 2004). Cultivation of organisms is no longer a requirement for examining microbial communities due the advent of modern molecular biology techniques. This has renewed interest in examining microbial community responses to environmental disturbances such as the metal-impact in the Chattanooga Fen. Lin et. al (2012) utilized culturing-independent methods to examine the differences in microbial communities within a bog (rainwater fed peatland) compared to those within a fen system (groundwater fed peatland) in the Glacial Lake Agassiz Peatland of northwestern Minnesota. This study found that bacterial richness in the fen was almost twice as high as the bog (estimated using Chaol, an alpha diversity metric). Acidobacteria was dominant in the bog system, while Firmicutes dominated the fen. Another study examined slightly acidic Canadian fen soil (pH of 4.5) in the lab, and determined that Acidobacteria, Nitrospirales, Alpha-, Gamma-, Deltaproteobacteria, and Cyanobacteria were regular inhabitants of the fen soil (Andersen et al., 2013). 16S rRNA based stable isotope probing from another study determined that active xylose and glucose
13


fermenting bacteria were present in an acidic, methane-emitting fen in Canada. These organisms included Acidaminococcaceae, Aeromonadaceae, Clostdriciaceae, Enterobacteriaceae, and Pseuodomonadaceae (Hamberger etal., 2008).
There have been few studies using culture-independent methods to determine the taxonomy of the microbial communities within fens, especially not within metal-impacted fens such as the Chattanooga Fen. The taxonomy of microbial communities within metal-impacted systems in general have also been understudied, but the organisms which seem to dominate endogenously and anthropogenically metal-impacted sites remain consistent in these studies (Table 2.1). The presence of some of bacteria such as Sulfobacillus and Acidobacteria can be explained by the presence of increased sulfur concentrations and acidic environments. These organisms also show differences in function, as it has been observed that both iron and sulfur reduction is increased in AMD and ARD sites while carbon degradation was observed to be decreased overall in both site types (Chen et al., 2016). The study presented here utilized molecular techniques such as high-throughput sequencing in attempt to understand the underlying bacterial communities present in the Chattanooga Fen and to determine if the resulting Fen communities resembled those of a fen-like or metal-impacted system.
14


Table 2.1. Taxonomy of bacteria which dominate endogenously and anthropogenically metal-impacted sites. Endogenous impact refers to ARD-impacted sites while anthropogenic refers to AMD-impacted sites
Taxon Phylum Source of Metal Impact
Acidiphilium (genus) Alphaproteobacteria Both
Acidisphaera (genus) Alphaproteobacteria Both
Acidithiobacillus (genus) Proteobacteria Both
Acidobacteria (phylum) Acidobacteria Endogenous
Acidocella (genus) Alphaproteobacteria Both
Acinetobacter (genus) Gammaproteobacteria Endogenous
Actinobacteria (phylum) Actinobacteria Endogenous
Gallionellaceae (family) Betaproteobacteria Both
Gemmatimonas (genus) Gemmatimonadetes Anthropogenic
Legionella (genus) Gammaproteobacteria Anthropogenic
Leptospirillum (genus) Nitrospirae Anthropogenic
Sphingomonas (genus) Sulfobacillus (genus) Alphaproteobacteria Firmicutes Anthropogenic Anthropogenic
15


CHAPTER III
METHODS BACKGROUND 16S rDNA High-Throughput Sequencing
In this study, high-throughput sequencing of a bacterial identification marker gene, the 16S rRNA gene, was used to taxonomically characterize the bacterial communities associated with the Chattanooga Fen (Riesenfeld, Schloss, and Handelsman 2004). Upwards of 25-million paired-end reads per run from the Illumina MiSeq sequencer allows for massively parallel sequencing of DNA fragments. Targeted sequencing strategies are often used to sequence segments of the 16S rRNA gene (a gene that has been highly conserved in Prokaryotes throughout evolution) from mixed-DNA samples in order to taxonomically characterize bacterial communities. To allow for taxonomic identification in bacterial census studies, regions within the 16S rRNA gene whose sequence varies from species to species, known as variable regions, are targeted for sequencing and allow for taxonomic identification (Figure 3.1). A general overview of the next-generation 16S rRNA gene amplicon sequencing process is detailed below (Figure 3.2).
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 bp
â–¡ V3 V4 V5 V6 â–¡
CONSERVED REGIONS unspecific applications
VARIABLE REGIONS: group or species-specific applications
Figure 3.1. Representation of the 16S rRNA gene including conserved and variable regions. The hypervariable region 4 was amplified to be sequenced in the presented study.
Image adapted from: www.biology.stackexchange.com
16


Environmental sample
Nucleic acid extraction/purification
16S rRNA sequencing
_JWV^ _AAAr jwr _Mn/“ _yw\r~ jwr —N\N~ JWT
PCR amplify 16S rRNA gene
i
Sequence
l
Group sequences into OTUs
Compare OTU sequences to databases
i
Identification of:
• Species
• Relative abundance of species within sample
>,[1
Figure 3.2: Broad procedural overview of 16S rRNA gene amplicon sequencing using next-generation sequencing technologies. Operation taxonomic units (OTUs) refer to DNA sequences >97% similar (similarity threshold is user-defined). Image adapted from: www.neb.com
Illumina DNA sequencing technology allows for multiple DNA molecules in a mixed sample to be sequenced concurrently (Caporaso et al. 2011). The hypervariable V4 regions of 16S rRNA genes present in extracted DNA are amplified using specific forward and reverse primers. The forward primer anneals to the conserved region of the 16S rRNA upstream of the V4 region and contains the 5’ Illumina adaptor sequence (which hybridizes to the flow cell during sequencing). Included is also a forward and reverse primer pad and primer linker used to link the Illumina adaptor with the
17


forward or reverse primer and for annealing of sequencing primers during sequencing. The primer pads and linkers do not have homology with the region of the 16S rDNA adjacent to the V4 region. The reverse primer, containing a 12-nucleotide molecular barcode, known as a Golay barcode, contains the 3’ complement of the Illumina adaptor, and is used to identify to which sample a specific sequence read belongs and allows for multiple samples to be sequenced in a single sequencing run (Figure 3.3). In an attempt to reduce reaction-specific PCR biases, these reactions are performed in triplicate. Triplicate reactions are pooled and cleaned and concentrated to remove genomic DNA, excess primers, dNTPs, and other reaction components, leaving only the amplicon of interest. All samples are then pooled in equimolar concentrations, creating a sequencing library (a library is a sample containing multiple sequences).
18


V4
index 806R >
SBSF (515F) p 46 bp overlap -| 150 bp index
P5 r

A P7
I 1 150 bp 254 bp SBSR : (806R)
Figure 3.3. V4 region is amplified using F515 and 806R primers. Paired 150 base pair sequencing gives a 254 bp fragment with a 46 bp overlap. Figure adapted from Integrated DNA Technologies
The sequencing library is denatured using sodium hydroxide, diluted in hybridization buffer, and loaded into the reagent cartridge prior to the actual sequencing process. The diluted, denatured library is then dispensed onto the MiSeq sequencing flow cell, and the Illumina adaptors of the amplicon strands hybridize to the oligonucleotides bound to the flow cell’s surface (Figure 3.4). After hybridization, a process called ‘bridge amplification’ occurs and DNA strands are amplified. This results in the formation of up to 25 million unique clusters, each containing thousands of copies of a single amplicon. Following this step, DNAtemplates are copied using fluorescently-labeled nucleotides which are then interpreted by a detector as specific nucleotides (Figure 3.4; The Illumina HiSeq 2000 Sequencing Technology 2015).
19


Figure 3.4. Cluster generation on the Illumina MiSeq sequencing flow cell. The DNA molecules are amplified using a process called bridge amplification, which results in the generation of up to 25 million unique clusters. Source: Illumina Inc.
Figure 3.5. DNA clusters are sequenced simultaneously using sequencing-by-synthesis technology. Fluorescently labeled nucleotides are incorporated one by one into the synthesis and a detector interprets the wavelength of light each fluorophore emits as a specific nucleotide. Source: Illumina Inc.


Bioinformatic and statistical tools, such as Quantitative Insights Into Microbial Ecology (QIIME) and R software can be used to analyze sequencing data generated from high-throughput 16S rRNA gene sequencing (Caporaso et al. 2010; R Core Team 2014). In order to quality filter sequencing reads, assign taxonomy to reads, and perform phylogenetic and statistical analyses, such as calculating alpha- and beta-diversity indices (diversity within samples and among samples), QIIME can be used. R software is a powerful statistical and programming language used to analyze virtually all data types. For sequencing data, R may be used to calculate multivariate statistics, including principal component analysis and generation of heatmaps and other plots to observe patterns amongst samples.
Metabolic Profiling: BIOLOG EcoPlates
BIOLOG EcoPlates were created as a sensitive and reliable monitor of environmental change, and they can be used for community level physiological profiling. EcoPlates were used here to examine carbon source usage (Figure 3.6).
Figure 3.6 Representation of a BIOLOG EcoPlate post inoculation and incubation with a sediment sample. The wells will turn a purple color when electrons released by the organisms within the well reduce the included tetrazolium dye due to metabolism of the included carbon substrate
21


Created for microbial community studies in ecological systems, EcoPlates measure the ability of microorganisms to metabolize 31 different carbon sources (Figure 3.7). Included in each of the 96 wells of the EcoPlate, along with the designated carbon source, is a tetrazolium dye. As carbon source utilization occurs, the released electrons will reduce the tetrazolium dye, resulting in a purple color. The optical density (OD 590) of the each well in the plate is then read using a microplate reader generating a metabolic fingerprint, or a pattern of carbon usage. This type of profiling can be useful in distinguishing spatial and temporal changes in microbial communities, noting that the technique is restricted to profiling a subset of the cultural microbial community. The technique has been used to monitor wastewater treatment plants, soils, and industrial waste sludge, along with many other sites being monitored for microbial community changes. This appears to be the first study to utilize the BIOLOG EcoPlates to monitor microbial communities associated with an iron-fen system. The BIOLOG EcoPlates provided potential metabolic properties to monitor using quantitative PCR. In this study, these properties were the degradation of xylose and cellobiose via glycoside hydrolase (GH) genes.
22


BiOLOG
Microbial Community Analysis
EcoPlateâ„¢
A1 Water A2 0-Methyl-D- Glucoside A3 D-Galactonic Acid y-Lactone A4 L-Arginine A1 Water A2 P-Methyl-D- Glucoside A3 D-Galactonic Acid y-Lactone A4 L-Arginine A1 Water A2 P-Methyl-D- Glucoside A3 D-Galactonic Acid y-Lactone A4 L-Arginine
B1 Pyruvic Acid Methyl Ester B2 D-Xylose B3 D- Galacturonic Acid B4 L-Asparagine B1 Pyruvic Acid Methyl Ester B2 D-Xylose 33 D- Salacturonic Acid B4 L-Asparagine B1 Pyruvic Acid Methyl Ester B2 D-Xylose 33 3- Galacturonic Acid B4 L-Asparagine
C1 Tween 40 C2 i-Erythritol C3 2-Hydroxy Benzoic Acid C4 L- Phenylalanine C1 Tween 40 C2 i-Erythritol C3 2-Hydroxy Benzoic Acid C4 L- Phenylalanine C1 Tween 40 C2 i-Erythritol C3 2-Hydroxy Benzoic Acid C4 L- Phenylalanine
D1 Tween 80 D2 D-Mannitol D3 4-Hydroxy Benzoic Acid D4 L-Serine D1 Tween 80 D2 D-Mannitol D3 4-Hydroxy Benzoic Acid D4 L-Serine D1 Tween 80 D2 D-Mannitol D3 4-Hydroxy Benzoic Acid D4 L-Serine
E1 o- Cyclodextrin E2 N-Acetyl-D- Glucosamine E3 T- Hydroxybutyric Acid E4 L-Threonine E1 a- Cyclodextrin E2 N-Acetyl-D- Glucosamine E3 Y- Hydroxybutyric Acid E4 L-Threonine E1 a- Cyclodextrin E2 N-Acetyl-D- Glucosamine E3 Y- Hydroxybutyric Acid E4 L-Threonine
F1 Glycogen F2 D- Glucosaminic Acid F3 Itaconic Acid F4 Glycyl-L-Glutamic Acid F1 Glycogen F2 D- Glucosaminic Acid F3 Itaconic Acid F4 Glycyl-L-Glutamic Acid F1 Glycogen F2 D- Glucosaminic Acid F3 Itaconic Acid F4 Glycyl-L-Glutamic Acid
G1 D-Cellobiose G2 Glucose-1 -Phosphate G3 a-Ketobutyric Acid G4 Phenylethyl- amine G1 D-Cellobiose G2 Glucose-1-Phosphate G3 a-Ketobutyric Acid G4 Phenylethyl- amine G1 D-Cellobiose G2 Glucose-1 -Phosphate G3 a-Ketobutyric Acid G4 Phenylethyl- amine
H1 a-D-Lactose H2 D,L-a- Glycerol Phosphate H3 D-Malic Acid H4 Putrescine H1 o-D-Lactose H2 D,L-o- Glycerol Phosphate H3 D-Malic Acid H4 Putrescine H1 a-D-Lactose H2 D,L-a- Glycerol Phosphate H3 D-Malic Acid H4 Putrescine
Figure 3.7 Carbon sources present in each well of the BIOLOG EcoPlate. Xylose and cellobiose wells are highlighted. In the present study, EcoPlates were utilized to evaluate the effect of AMD on microbial carbon source usage by previously ARD-impacted communities
Quantification of Glycoside Hydrolase Genes
Quantitative PCR (qPCR) utilizes the linearity of DNA amplification to determine the amount of a known sequence within a sample. In this study, qPCR was utilized to determine the amount of three glycoside hydrolase genes involved in the degradation of cellobiose and xylose based on the results of EcoPlate analysis.
The present study measured DNA generation with the use of SYBR green I, a fluorescent reporter included in the PCR reaction which will fluoresce when bound to dsDNA PCR products. During each cycle of the PCR, DNA amplification is monitored. The reactions are characterized by the cycle at which the fluorescence first rises above the set threshold. If the starting material is abundant, amplification, and, in turn,
23


fluorescence from the SYBR green I, is observed in earlier cycles. If the amount of initial DNA is low, then amplification and fluorescence will occur in later cycles. By using multiple dilutions of a known amount of standard DNA, a standard curve is generated of log concentration against quantification cycle. The quantification cycle is the point at which fluorescence becomes measurable above the background (Figure 3.8). The amount of fluorescence and the quantification cycle can then be correlated to the generated standard curve, and the amplified product can be quantified. Here, primers were utilized to quantify the abundance of three glycoside hydrolase (GH) genes involved in the degradation of cellobiose and xylose.
Action of SYBR Green I Dye
imiiniimmHiiimiiniimmmmiminim
denaturation
imiaiimmmaaumimiimmuiiiAiiiiiiii
•. •
• • «-
I.Dycin solution emits low fluorescence
1. f. .Cl. i -
,iii 1.1:
2. Emission of the fluorescence by binding
r â–  itt â–  HnnTTmmTiTTTm rTriifiiTnimmi m
IIIIil11-1
luimimiiiimiumumimmimiiiTiiililllj
pnmr anneal, Detector
miiiiijimiHiuiiiiiiiniiiiminiiiiuiiMMM
.....j. ,,,1111111111111111n)ir.
extension
ffiMlinillhniyhihVrhiUiihhiiWihi' 11
a umuiiiii: 11111111111A11111
TTTTl ii- d
Figure 3.8. Diagram of qPCR utilizing SYBR green I. When SYBR green is in solution, it emits low fluorescence, but when it binds to dsDNA, it will fluoresce and the qPCR machine will detect the fluorescence (Image adapted from www.sigmaaldrich.com)
24


Studies which quantified GH6, GH48, and GH10 in environmental samples are extremely limited. To the best of our knowledge, this is the first study to quantify these genes in both AMD and ARD impacted sediments.
25


CHAPTER IV
CONDUCTED STUDY Introduction
The impact of acid mine drainage (AMD) on previously metal-exposed microbial communities is addressed in this study. The Chattanooga Fen is a wetland system that is both naturally metal impacted by endogenous acid rock drainage (ARD) and, in some areas, impacted by AMD from the nearby Gold Finch Mine. The fen is physically bisected into two distinct areas, each with ARD exposure, but only one of which has the additional exposure to AMD. To explore how the long-term release of AMD into the Fen has impacted the previously metal adapted bacterial communities, this study utilized high-throughput sequencing to determine taxonomic structure, metabolic profiling methods to evaluate carbon source utilization, and quantitative PCR to quantify genes involved in carbon degradation.
There have been few previous studies examining the bacterial communities within fens, and even fewer which have been conducted on acidic fens. The microbiological studies on acidic peatlands have mainly been performed in Canada and the northern U.S., though these are not metal impacted by two different sources like the Chattanooga Fen (Andersen et al, 2013; Lin et al, 2012). The studies which have been done relied heavily on culturing, a limited technique because only 1 to 10% of all bacteria are cultivable in a laboratory setting. Modern, molecular methods such as high-throughput sequencing are now being utilized to more thoroughly characterize a variety of environments, including wetlands and metal-impacted systems. The presented study aimed to characterize the bacterial communities within a unique,
26


previously metal-adapted system, where portions of the system were additionally impacted by the presence of acid mine drainage.
Methods
Sample Collection
In order to characterize the bacterial communities within the Chattanooga Fen as well as their metabolic profiles, sediment samples were collected along an AMD gradient within the site that is also ARD-impacted (from AMD-unaffected ponds formed within the fen to AMD-affected ponds and straight from the Gold Finch AMD effluent). This was accomplished in July and September of 2016 via the collection of 2” X 3-10” sediment cores. Duplicate cores were taken at each site, and one was placed on dry ice to preserve nucleic acids while the other was stored on ice to maintain cell viability (Figure 4.1). There were five AMD-affected samples taken in both July and September (denoted as CFA and CFMine samples). Four AMD-unaffected sediment samples were collected in July, and five AMD-unaffected samples were collected in September (denoted as CFU samples) (Table 4.1). At the time of sampling, temperature, pH, conductivity, and dissolved oxygen were measured in the field from each sampling site within the Fen at the time of sample collection using a Thermo Scientific Orion 5-Stat Multiparameter Meter Kit.
27


Table 4.1. List of sediment samples collected from the Chattanooga Fen during July and September of 2016.
Site Abbreviation Site Description
CFMine-July Chattanooga Fen from sediments directly impacted by AMD flowing from the Gold Finch Mine addit.
CFAA-July The smallest, copper-orange colored pond sampled. This pond sits closest to the direct AMD sediment, and AMD could be seen trickling into the pond. Large build-up of peat beneath the sediment.
CFAB-July Orange pond slightly lower than and east of the CFA1 pond. Large build up of peat below the sediment.
CFAl-July The largest of affected ponds sampled. Sits just south of the CFAB pond. Large build-up of peat below the sediment.
CFA2-July Orange pond just south-east of and below CFA1. Large build-up of peat.
CFMine-Sept Chattanooga Fen from sediments directly impacted by AMD flowing from the Gold Finch Mine addit.
CFAB-Sept The pond was smaller than when the July sample was taken. Orange color and closest to the mine outflow (the CFAA-July site was nearly gone]. Still a large build-up of peat below the sediment.
CFAC-Sept Orange pond just east of the CFAB pond, though smaller than CFAB. Large build-up of peat below the sediment.
CFAl-Sept The largest of affected ponds sampled. Sits just south of the CFAB pond. Large build-up of peat below the sediment.
CFA2-Sept Orange pond just south-east of and below CFA1. Large build-up of peat.
CFUA-July Northeast of the mine outflow, greenish ponds unimpacted by AMD due to the fact that it sits higher than the mine portal. The highest of all of the sampled ponds.
CFUB-July The greenish colored pond closest to the CFUA pond. Larger than the CFUA pond. The deepest of the unaffected ponds sampled.
CFUl-July The largest unaffected pond sampled. Sits beneath a row of trees up the hillside from the affected ponds
CFU2-July Slightly smaller pond just east of the CFU1 pond. Greenish color with multiple trees lining the pond.
CFUB-Sept The greenish colored pond up the hillside from the mine outflow. The deepest of the unaffected ponds sampled, though slightly shallower than the July sample.
CFUl-Sept The largest unaffected pond sampled. Sits beneath a row of trees up the hillside from the affected ponds.
CFU2-Sept Slightly smaller pond just east of the CFU1 pond. Greenish color with multiple trees lining the pond.
CFU3-Sept The greenest of the sampled ponds, and the furthest northeast of the mine outflow.
CFU4-Sept The smallest pond sampled. Isolated, greenish pond just south of the CFU3 pond.
To measure water chemistry, 250 mL water samples co-located with the sediment samples were also collected. Water was filtered through a Nalgene 0.45 pm cellulose nitrate membrane filter (Thermo Scientific, Waltham, MA), acidified to pH <2.0
28


with ultrapure 1:1 HNO3, and stored in a 4°C cold room in the dark for further analysis. Dissolved metals were quantified by the Shared Analytical Laboratory, an analytical chemistry service laboratory on the University of Colorado Denver downtown campus, using a Thermo Jarrell Ash ICAP 61 inductively-coupled plasma optical emission spectrometer (Thermo Jarrell Ash Corporation, Franklin, MA).
Figure 4.1. Photographs of sediment samples being collected via coring in unaffected and affected ponds.
DNA Extraction, PCR Amplification, and Next-Generation Sequencing
The collected cores were cut using sterile PVC cutters, and the top 2 inches of sediment was placed in sterile tubes. Total genomic DNA was extracted from 10.0 ± 0.1 grams of mechanically homogenized sediment from each tube using the MO BIO PowerMax Soil DNA Isolation Kit (MO BIO Laboratories, Inc., Carlsbad, CA) following manufacturer’s protocol. 16S rRNA genes from each sample were amplified in triplicate, pooled, and cleaned. The V4 hypervariable region of the 16S rRNA gene was amplified using primers F515 (5’ GTG CCA GCM GCC GCG GTA A 3’) and 806R (5’ GGA CTA CHV GGG TWT CTA AT 3’), each containing 5’ overhangs necessary for Illumina
29


high-throughput sequencing, and the reverse primer containing a 12-nucleotide sample-specific Golay barcode sequence (Caporaso etal. 2012). The polymerase chain reaction (PCR) reaction mixture contained 10 pL 5 PRIME Hot Master Mix (5 PRIME, Gaithersburg, MD) (final reaction concentrations: 0.5 U Taq DNA polymerase, 22.5 mM KC1,1.25 mM Mg2+, and 100 pM of each dNTP), 200 nmol/L of each primer, 200 nanograms (IX) bovine serum albumen (BSA) (New England BioLabs, Inc., Ipswich, MA), molecular biology grade water (Thermo Fisher Scientific, Inc., Waltham, MA) and template DNA to a total volume of 25 pL. Reactions were subjected to a 3-minute denaturation step at 94°C, 30 cycles of a 45 second denaturation at 94°C, a 60-second annealing step at 50°C, and a 90-second extension step at 72°C, followed by a ten-minute final extension step at 72°C.
Successful amplification was verified via agarose gel electrophoresis (1% agarose gel) and visualized with ethidium bromide to confirm that the PCR product size obtained coincided with expected product size. Successfully amplified reactions were pooled and purified using the ZYMO RESEARCH DNA Clean & Concentrator (ZYMO RESEARCH, Irvine, CA), following the included protocol. DNA concentrations from cleaned reaction concentrates were quantified using the Qubit Broad-Range dsDNA Assay Kit and Qubit 3.0 Fluorometer (Thermo Fisher Scientific, Inc., Waltham, MA), following the included protocol. DNA from each quantified sample was pooled in equimolar ratios, concentrated, and re-quantified using Qubit. The library of pooled samples, along with aliquots of the forward, reverse, and index sequencing primers, were sent to the University of Colorado Anschutz Medical Campus Genomics and Microarray Core Facility for Illumina high-throughput sequencing.
30


Computational Analyses
Phylogenetic analyses were performed using the QIIME (Quantitative Insights Into Microbial Ecology) software (Caporaso etal. 2010). Using the join_paired_ends.py command, paired end sequenced were joined, which uses the fastq-join method and a specified minimum overlap score of 30 (Aronesty 2011). Joined sequences were filtered using the split_libraries_fastq.py command, and sequences with a Phred quality score of 30 were kept, which corresponds to a sequencing error rate of 0.1%. Operational taxonomic units (OTUs) were picked using the pick_open_reference.py where reads were clustered together based on 97% or greater sequence similarity. This command also assigned taxonomy to OTU’s using the August 2013 Greengenes bacterial and archaeal 16S rRNA database (DeSantis 2006). Chimeric sequences were identified using DECIPHER Find Chimeras Web tool and were removed from the dataset prior to further analysis (Wright 2012). Alpha diversity was performed in QIIME using the alpha_diversity.py command. PCoA plots for the sediment data were generated using the jackknifed_beta_diversity.py command in QIIME.
BIOLOG EcoPlate Preparation and Reading
BIOLOG EcoPlates (BIOLOG, Hayward, CA) were utilized to examine the metabolic profile diversity within the cultural portion of microbial communities from the Fen sediment samples. The EcoPlates are 96-well plates with 21 different carbon sources, each repeated three times on the plate with three negative control wells. The core sediment samples which had been placed on ice for viability were immediately cut using autoclaved PVC cutters 1 inch from the top of the sediment so the sediment could be reached with sterile tools. An autoclaved spatula was then used to remove the top 2
31


inches of sediment from each of the viability cores, and the sediment was placed into sterile tubes. From these tubes, 10 grams of sediment was mixed with 90 mL of 0.145 M sodium chloride solution made with sterile deionized water in a 250 mL flask. The flasks were then placed on a shaker for at 200 RPM for 20 minutes at 20°C. Once the shaking had completed, the flasks were put into a 4°C refrigerator for 30 minutes to allow the sediment particles to settle. Using a multichannel pipette, 150 pL of liquid from each flask was placed into wells of the EcoPlates. Once the entire plate was inoculated, the plate was covered with sealing tape, the lid was replaced, and the plates were stored in a 25°C incubator.
Every 24 hours post inoculation, a BioTek Gen5 microplate reader was used to measure the absorbance (590nm) of each well in the 96-well plate.
Average well color development was calculated for each triplicate in each plate and then averaged. Average well color development (AWCD) was calculated using the following equation:
AWCD= 2(C-R)/31 C= OD590nm
R= OD590nm of the control well DNA Extraction and Amplification from EcoPlates
Xylose and cellobiose were identified as substrates (carbon sources) of interest. DNA was then extracted from each of the BIOLOG EcoPlates from both the xylose and cellobiose wells. To accomplish this, 50 pL of culture from each of the three xylose wells
32


was placed into an extraction tube from the MO BIO PowerMax Soil DNA Isolation Kit (MO BIO Laboratories, Inc., Carlsbad, CA) along with 1 pL of molecular grade water in order to increase the volume to the required amount. The manufacturers protocol was then followed, and library preparation was completed in the same manner as the total 16S rDNA libraries discussed above. The process was repeated for the cellobiose wells of each EcoPlate. The resulting libraries were sent to the Anschutz Medical Campus Genomics and Microarray Core Facility for Illumina high-throughput sequencing. Quantitative PCR
Utilizing the StepOnePlus Real- Time PCR system, the abundance of three glycoside hydrolase genes (GH10, GH6, GH48) were quantified. The primers used were X10-F (5’-CTACGACTGGGAYGTNIBSAAYGA-3’) andX10-R (5’-
GTGACTCTGGAWRCCIABNCCRT-3’) for the GH10 gene (Wang eta/., 2010); cell2F (5’-ACCTGCCCRCCGYGACT-3’) and cell2R (5’- GAGSGARTCSGGCTCRAT-3’) where R = A or G; Y = C or T; S = G or C for the GH6 gene (Merlin et al, 2014); and GH48_F8 (5’-GCCADGHTBGGCGACTACCT-3’) and GH48_R5 (5’- CGCCCCABGMSWWGTACCA-3’) for the GH48 gene (de Menezes et al., 2015). qPCR conditions and reactions were carried out as previously described for GH10 (Wang et al., 2010), for GH6 (Merlin et al., 2014), and for GH48 gene (de Menezes etal., 2015). The reactions were performed in a 25 pL mixture using 12.5 pL of PowerUp SYBR Green Master Mix, 2 ng of extracted DNA diluted to 1 ng/pL of molecular grade water, 0.5 pL of each primer, and 9.5 pL of molecular grade water.
Standard curves for each gene were generated using synthesized gene fragments from Integrated DNA Technologies (Coralville, IA). The standard gene copies for the
33


assays ranged from 60 to 6.08 X 109 copies for GH6 and GH48 and from 35 to 3.5 X 109 for GH10. For all samples and standards, triplicate reactions were carried out. In all triplicate measurements, the standard deviation was less than 10%. To check for specificity of amplification, melt curves were generated for each SYBR assay. The resulting qPCR assays were also run on through gel electrophoresis to confirm amplification of the expected gene. PCR efficiencies ranged from 72.8 - 80.9% for GH10 reactions, 73.7 - 82.7% for GH6 reactions, and 71.6 - 88.3% for GH48 reactions. The correlation coefficients (R2) was greater than 0.99 for all assays.
Results and Discussion
Site Chemistry
The chemistry of the water from the AMD-affected sites were distinct from the AMD-unaffected sites in terms of pH, dissolved oxygen (DO), and conductivity (Table 4.2). AMD-affected samples had an average pH of 5.36, while unaffected samples had an average pH of 4.59 (student’s t-test p-value=0.009). The most basic pH values came directly from the effluent from the Gold Finch Mine, the source of AMD impact to portions of the Fen. This may be due to the fact that the mine drainage flows through a constructed concrete portal which sits below the Gold Finch Mine, which raises the pH of the outflow or because the water flows through mineral ore material with a higher carbonate content. The ground water which feeds the Fen is naturally acidic (which in turn dissolves the parent rock material and forms the ARD within the Fen), and the unaffected sites do not receive the flow of raised pH water from the mine.
34


The amount of dissolved oxygen (the amount of gaseous 02 in the water) was greater in unaffected sites compared to the affected samples (student’s t-test p-value= 0.03). Dissolved oxygen often relates to water quality, and it is essential to support aquatic life. This includes bacteria and other microorganisms which depend on DO to decompose organic material at the bottom of a body of water, such as the ponds at the Chattanooga Fen, and, in turn, contribute to elemental cycling. The highest DO levels came from the unaffected ponds in September which may be due to the fact that oxygen dissolves easier in cooler temperatures than warm temperatures. When the September samples were collected, there was a steady snowfall, and the temperatures were significantly lower than when the July samples were collected. The AMD- affected samples had significantly higher conductivity values compared to the unaffected samples (student’s t-testp-value=0.003) (Table 4.2).
Conductivity (the measure of a solution’s ability to conduct electricity) was significantly higher in AMD-affected sites compared to AMD-unaffected sites (Table 4.2). Conductivity can change depending on temperature, which would explain why the July samples had greater conductivity values compared to the September samples because warmer temperatures tend to increase conductivity due to the increased dissolution of metals. Changes in conductivity can also be related to the presence of pollutants, such as AMD, which affects organisms within the site because aquatic life can tolerate certain conductivity ranges. Pollutants such as AMD can increase the concentration of ions, thus increasing the conductivity of the water. The geology of an environment and the material through which the ground water flows through also influences the conductivity of a solution. Therefore, the bedrock which the groundwater
35


flows through to feed the mine outflow as well as the affected sites may flow through materials which better ionize when washed into water compared to the unaffected sites. The fact that the bedrock which the water flows through both from the mine and the groundwater differs may also explain the differences in metal concentrations between affected and unaffected sites since affected sites had a greater amount of calcium, iron, magnesium, manganese, and zinc (Table 4.3).
After analysis of dissolved metal concentrations, it was found that the acid mine drainage sample along with the AMD-affected samples had elevated concentrations of calcium, iron, magnesium, manganese, and zinc compared to the AMD-unaffected samples (student t-test p-values<0.05) (Table 4.3). Each of these metals have been shown in previous studies to be elevated in AMD-impacted sites, though these studies were not on metal-impacted fen systems, such as the Chattanooga Fen (Chen et al., 2016). The unaffected samples generally had a higher concentration of aluminum compared to the affected sites. However, the CFA1 sample taken from September was an outlier for the affected samples, and had a concentration of aluminum closer to the unaffected samples. If the outlier is removed, the concentration of aluminum within the unaffected samples is greater compared to AMD-affected samples (student’s t-test p-value= 0.004).
It should be noted that the CFMine samples were included with the AMD-affected samples in these analyses. However, to determine whether the Gold Finch Mine samples were skewing the analyses, the mine samples were removed from analysis. Results still showed similar differences between AMD-affected and AMD-unaffected sites (student’s t-test p-values <0.05).
36


Table 4.2. Chemistry parameters measured in the field for each sampling site
Sample Temp (°C) pH Dissolved Oxygen (mg/L) Conductivity (pS/cm)
CFMine-July 5.8 6.09 5.07 468
CFAA-July 15.1 4.95 7.1 440
CFAB-July 18.5 4.57 4.2 403
CFAl-July 14.8 5.17 5.91 472
CFA2-July 16.1 5.41 5.12 370
CFMine-Sept 5.7 6.1 2.25 434
CFAB-Sept 7.7 6.66 1.12 225
CFAC-Sept 6.8 5.37 0.89 320
CFAl-Sept 7.5 4.38 15.71 143
CFA2-Sept 6.4 4.85 13.03 378
CFUA-July 13.1 5.2 3.86 237
CFUB-July 16.4 4.44 3.82 230
CFUl-July 16.1 4.28 3.34 233
CFU2-July 18.1 4.46 3.83 214
CFUl-Sept 7.3 4.51 21.12 119
CFU2-Sept 6.7 4.78 25.57 113
CFU3-Sept 5.8 4.45 23.25 232
CFU4-Sept 6.6 4.61 15.94 108
CFUB-Sept 6.1 4.65 14.3 185
37


Metal Concentration (ppb)
Sample A1 Ba Ca Cd Cu Fe Mg Mn Na Ni Pb Zn
CFMine-July 1356.08 24.7554 47406.8303 137.655 134.545 8337.66 5215.98 1922.48 591.528 253.655 1147.99 383.326
CFAA-July 1947.43 32.4253 58575.4905 196.406 175.704 1301.13 6988.16 2310.95 2029.28 348.536 1713.03 469.893
CFAB-July 1624.83 25.7293 49692.8299 131.769 127.652 1507.38 6501.63 1954.6 2191.81 225.012 1153.48 340.53
CFAl-July 884.324 14.8154 57838.1922 39.3975 68.1976 277.429 5048.68 2273.72 779.061 65.0974 295.974 351.536
CFA2-July 1778.05 27.1363 54850.2381 165.771 151.688 817.835 6912.82 1543.03 1672.97 279.777 1445.63 297.358
CFMine-Sept 746.819 13.0216 69231.8965 32.4508 61.4696 10718.7 5938.28 2665.12 2548.12 57.4851 270.505 427.475
CFAB-Sept 38.1353 4.09384 57754.4337 5.17716 37.5752 5743.01 4812.39 2113.93 3110.72 10.0978 69.8665 38.7059
CFAC-Sept 773.685 14.1418 50410.8062 6.63118 44.466 259.409 4768.76 1250.09 2760.66 7.82862 95.2006 205.423
CFAl-Sept 6210.34 43.8914 22369.1431 178.17 171.131 3827.46 5691.02 829.679 1879.25 319.969 1540.75 454.052
CFA2-Sept 853.98 14.4447 60927.9423 29.2992 64.9206 541.466 5338.13 2412.53 2310.58 55.3496 269.685 396.249
CFUA-July 2917.75 25.5529 19101.2672 58.8709 76.8158 266.046 4186.64 611.9 1291.65 94.0518 478.472 238.677
CFUB-July 3779.38 35.0579 18736.0746 147.426 140.327 593.685 5159.65 618.038 1660.47 244.397 1293.34 281.417
CFUl-July 2686.68 30.1068 17573.4299 84.5531 95.3062 898.79 4056.6 544.811 1204.14 141.714 710.468 251.47
CFU2-July 2263.24 27.0817 16644.9891 60.1088 78.1416 635.64 3559.91 521.992 897.832 93.1841 488.214 222.985
CFUB-Sept 2053.94 29.6405 13454.7205 <5 30.9573 445.335 2438.79 364.37 2710.65 <50 <50 352.546
CFUl-Sept 1503.88 11.7711 14904.4547 <5 22.4995 348.462 2536.38 461.982 910.334 <50 <50 80.6132
CFU2-Sept 1533.3 14.0017 15801.138 <5 43.2225 1923.79 2967.85 488.941 1135.37 <50 <50 89.1303
CFU3-Sept 3515.83 23.3788 19725.7103 48.127 70.9978 156.725 4403.11 640.754 1797.99 75.2251 431.12 283.389
CFU4-Sept 1444.38 13.1105 14778.0652 <5 29.264 537.756 2367.3 437.497 1960.52 <50 <50 129.834
Table 4.3. Dissolved metal concentrations from water samples collected above sediment samples taken from the Chattanooga Fen in July and September of 2016. Concentrations are given in parts per billion (ppb). The concentrations of calcium, iron, magnesium, manganese, and zinc are greater in AMD-affected sites
(JO
oo


Taxonomic Structure of Bacterial Communities within the Chattanooga Fen
The taxonomic composition of the AMD-affected sites was also unique from the AMD-unaffected sites within the Chattanooga Fen. However, QIIME analyses revealed seven phyla to be common and most abundant within the Chattanooga Fen. These seven phyla included Acidobacteria, Actinobacteria, Bacteroidetes, Chloroflexi, Cyanobacteria, Proteobacteria, and Verrucomicrobia. The sediment samples which came directly from the outflow of AMD was dominated by Proteobacteria (an average of 77.05% relative abundance). In the affected samples, the relative abundance percentage of Bacteroidetes and Actinobacteria were higher compared to the unaffected samples (student’s t-test p-values of 0.04 and 0.0003, respectively). The unaffected samples had greater relative abundance percentage of Acidobacteria and Verrucomicrobia compared to affected samples (student’s t-test p-values of 0.0001 and 0.03, respectively) (Table 4.4 and Figure 4.2).
Table 4.4. Top seven most relatively abundant phyla shown to dominate within the sediment samples collected from the Chattanooga Fen. The average percent relative abundance is shown for each phyla from mine, affected, and unaffected samples, the standard deviations are included in parentheses
Mine Samples Affected Samples Unaffected Samples
Acidobacteria 3.27(0.35) 7.81(4.29) 20.71(6.68)
Actinobacteria 3.13(1.17) 2.39(1.74) 1.18(088)
Bacteroidetes 8.28(2.49) 9.82(3.73) 3.87(1.71)
Chloroflexi 0.2(0) 4.1(1.2) 4.46(11.51)
Cyanobacteria 0.58(0.21) 18.36(12.7) 19.47(8.78)
Proteobacteria 77.05(5.13) 34.9(10.75) 30.39(4.29)
3.3(0.25) 3.33(2.056) 4.83(0.84)
Verrucomicrobia
39


While the phyla Proteobacteria was common to the Gold Finch Mine AMD, to the AMD-affected, and to the AMD-unaffected sediment samples, the classes which encompass the phylum differed. Alphaproteobacteria was the dominant class in the unaffected samples (an average of 11.56% relative abundance compared to 6.15%, student’s t-test p-value of 0.01). Deltaproteobacteria had an average relative abundance percentage of 14.15% in the AMD-affected and Gold Finch Mine samples compared to an average relative abundance of 6.5% in the unaffected samples (student’s t-test p-value of 0.005). Betaproteobacteria also comprised a large amount of the AMD-affected and Gold Finch Mine samples (average relative abundance of 21.29% compared to 8.32% in the unaffected samples).
The dominant classes of Proteobacteria within the AMD-affected samples from the Chattanooga Fen were Betaproteobacteria and Deltaproteobacteria, which are both compromised of Gram negative bacteria. Though metabolically diverse, many of the organisms within the Betaproteobacteria class, including Burkholderiales and Methylophiales which were found in all affected samples, obtain their energy from inorganic compounds. However, another order common to all affected samples was the order Nitrosomonadales which includes organisms that perform nitrification and are important in the global nitrogen cycle (Mendez-Garccia et al, 2015).
Many of the Deltaproteobacteria which were found more frequently in the AMD-affected sites compared to the unaffected sites can survive in unfavorable environments (such as AMD). One of the orders found in all of the affected sites was Myxococcales. These gliding organisms will form fruiting bodies in order to survive until nutrients are more plentiful. There were also several orders found in all of the affected samples that
40


are sulfate-reducing bacteria, including Desulfobacterales and Desulfovibrionales. Considering the sulfate-rich environments these organisms live in, there presence is to be expected.
After analysis, it was determined that in the Gold Finch Mine samples, bacteria of the genus Galionella of the class Betaproteobacteria were dominant (representing 50.13% of the organisms from the July mine sample and 55.41% of the organisms from the September mine sample). Gallionella species are iron-oxidizing, chemolithotrophic bacteria that have been found in a variety of different aquatic habitats. These bacteria play an important part in the oxidation of iron (Jones et al, 2015).
The next most abundant organisms in the Gold Finch Mine samples came from the Bacteriovoracaceae family (8.51% of the organisms in the July sample and 13.84% of the organisms in the September mine sample) (Figure 4.3). Within this family are extremely unique, Gram negative organisms. These bacteria are commonly found in enteric environments as well as river water. They are known as bacterial predators because they can attach to the outer membrane of Gram negative bacteria, enter the periplasmic space, and use hydrolytic enzymes to feed on the host cell biopolymers (Chen etal, 2012).
In AMD-unaffected samples within the Chattanooga Fen, the phylum Acidobacteria was found in greater relative abundance (an average of 22.54% compared to 6.55% in AMD-affected samples and 3.23% in Gold Finch Mine Samples) than the AMD-affected sites, particularly the orders Acidobacteriales and Soilbacterales. Both of these contain acidophilic bacteria naturally found within soil and sediment. However, even though Acidobacteria are widespread within terrestrial ecosystems,
41


they remain largely uncharacterized and uncultured. The only cultured Acidobacteria are heterotrophic organisms which are aerobic or facultatively anaerobic.
A common group of organisms found in the sediment coming from affected and unaffected samples, but only found in less than 0.2% abundance in the sediment samples which came directly from the acid mine drainage, was the order Stramenopiles which are of the phylum Cyanobacteria (an average of 17.9% ± 13.1% and 19.79% ± 9.98%, relative abundance, respectively) (Figure 4.3). Cyanobacteria are photo synthetic bacteria which generally do not live in flowing waters, which may explain why they were not found in the mine effluent sediments.
42


â– Iine-July
i Unaiiigned; Other
ip_Crcnarchacota ip_Euryarchacota I p_ [Parva rch ac ota) [Other
tp_
tp__AC1
[p__AD3
tp_Acidobactcria
tp__Actinobactcria
tp_Armatimo nadetc s
tp__BH180-139
tp__BHC1
tp„ B a c tcro i dete s tp__Caldiscrica
tp__Chlamydiae
tp__Chlorobi
tp__ChlorofIcxi
tp__Cyanobactcria
tp_Elusimicrobia
tp__FBP
tp__FCPU426
tp__Fibrobacteres
tp_Firmiajtcs
tp_Fusobarteria
tp__GN02
tp__GN04
tp__GOUT A4
tp_Gcmmatimonadctcs
tp__H-l 78
tp__Myd24-12
tp__KSB3
tp__LCP-89
tp__LD1
tp_Lcotisphacrac
tp__MAT-CR-M4-BO/
tp__MVP-21
tp__MVS-104
tp__NC10
tp__NKB19
tp Nilrospirae
tp__0D1
tp_OPl
tp_OPll
tp__0P3
tp_0P8
tp__0P9
tp__PAUC34F
tp_Planctomycetcs tp_Por bacteria tp__Pro tcob ac tcri a
tp__SC 4
tp__SRI
tp__Spirocha etes
tp__TA06
tp__TM6
tp__TM7
tp TPP 58_____________
Figure 4.2. Relative abundance of the bacteria within each collected sediment sample (represented at the phylum level)
43


f_Bacteriovoracaceae g_Gallionella g_Gallionella Unassigned g_Gallionella g_Gallionella g_Anae r o myxob acte r f_ Sinobacteraceae c_ Betaproteobacteria_ o_ Pedosphaerales f_ Sinobacteraceae f_Gallionellaceae f_Koribacteraceae Unassigned c_Betaproteobacteria; o_SBlal4 g_Geothrix o_Bacteroidales o_Bacteroidales g_Desulfobacca o_Stramenopiles f_Desulfobacteraceae Unassigned
c_ Deltaproteobacteria;o_MBNT15 o_Stramenopiles c_Anaerolineae; o_A31 c_ Betaproteobacteria Acidobacteria; c_OS-K g_Candidatus Koribacter; o_Stramenopiles c_WPS-2
a
*
CO
D
u.
U
a a
x x
D D
LL U.
u u
a a
x x
H
D D
ll ll
U U
X X X X a a 1 a X X X X Q.
5 fN D .3 H D ,3 CO D 3^ i D X 6 X CO X fN Ji H 5 IN < .3 H < .3 < 3 i X i
LL LL LL LL < < < < LL LL LL LL c
u u u U LL u LL u ll U LL U U U U U 2 LL U
x
3
c
i
u
Figure 4.3. Heatmap of the top 30 OTUs from each sediment sample. All of the included OTUs had a minimum of 30,000 reads
44


Alpha diversity, which measures the diversity within individual samples, was also measured using Chaol, Shannon Index, Faith’s phylogenetic diversity, and Observed OTUs metrices. This was performed for affected and unaffected samples as well as within the two time points (July and September sampling). The Shannon Index measured species diversity, while Chaol and Observed OTU richness measure the species richness, and Faith’s phylogenetic diversity correlates phylogenetic relationships with species richness. There was no determined correlation between AMD-affected and AMD-unaffected samples overall nor between AMD-affected and AMD-unaffected July versus September samples, in terms of any of the alpha diversity measurements performed (Table 4.5). The lack of correlation implies that AMD did not have an impact on alpha diversity among the samples. Recall that these samples are endogenously metal exposed, and, as such, have presumably already adapted to the presence of potentially toxic metals. Unlike other studies that have found AMD negatively impacts alpha diversity, this was not observed here (Chen et al., 2016).
45


Table 4.5. Alpha diversity metrics for all samples of the same material grouping where Chaol estimate, Observed OTU richness, Faith’s phylogenetic diversity, and Shannon Index were performed and a t-test was done to test if the mean value for each group was significantly different. The data shown is the average of all samples belonging to each group with the standard deviation in parentheses________________________________________
Chaol Shannon Index Faith's PD Observed OTUs
Affected Samples 1401.77(582.17) 7.63(1.35) 152.74(48.99) 1818.09(605.24)
Unaffected Samples 1560.32 (559.34) 7.65(0.667) 126.67(27.95) 1533.68(411.72)
July Samples 1921.523(531.27) 7.64(1.09) 153.53(42.33) 1877.33(538.32)
September Samples 1176.69(214.34) 7.63(1.07) 128.56(38.96) 1508.81(481.72)
t-test p-value Affected/Unaffected 0.285 0.964 0.201 0.278
t-test p-value July/September Samples 0.318 0.988 0.223 0.155
To evaluate the impact of anthropogenic metal impact on the taxonomic composition of the collected sediments from the Chattanooga Fen, affected and unaffected samples were compared. Principal coordinate analysis (PCoA) plots of weighted and unweighted UniFrac values were generated. PCoA plots are a form of multidimensional scaling which aid in the visualization of similarities between individuals in a dataset. In this case, the PCoA plots created utilized UniFrac, which is a distance metric used for comparing biological communities. The weighted variant of UniFrac is quantitative and takes into account the relative abundance of observed organisms, while the unweighted UniFrac variant only takes into account presence or absence of the organisms. Using the weighted UniFrac values, Axis 1 explained 42.35% of the variation arrayed along this axis, while Axis 2 explained 20.32% of the observed variation. Therefore, in total between both the first and second principal coordinates,
46


62.67% of the variance was explained, and the affected and unaffected groups cluster together (Figure 4.4). Using the unweighted UniFrac values, Axis 1 explained 17.69%, while Axis 2 explained 15.07%. Therefore, when using the unweighted UniFrac variant, which does not take relative abundance into account, 32.76% of the variance was explained (Figure 4.4). ANOSIM statistical testing was then used to compare the variation in species composition and abundance between AMD-affected and AMD-unaffected samples, also known as beta diversity. The resulting R-statistics from both weighted and unweighted UniFrac values were positive numbers, suggesting there was similarity in bacterial communities between the samples within the given groupings (AMD-affected and AMD-unaffected) (Table 4.6).
ANOSIM statistical testing was used to determine if there was a temporal influence on the bacterial community variation in the samples. The R-statistics for both weighted and unweighted UniFrac variants were extremely close to 0 (R=0.0126 and 0.0186), implying that there was not a difference in community structure between July and September samples (Table 4.7).
Alpha diversity metrices showed no correlation between condition (AMD-affected or unaffected) or month of sample collection and diversity within samples (Table 4.5). However, the constructed PCoA plots based off of weighted UniFrac variants showed that the first two axis explained 62.67% of the variance in taxonomic composition between AMD-affected and AMD-unaffected sites (Figure 4.4). This could be due to the fact that the AMD flows from the mine to the affected samples, and organisms are brought into the affected sites along with organisms that were already present in the endogenously metal-impacted Fen. There is one AMD-affected sample
47


which sits closer to the AMD-unaffected samples then the AMD-affected samples, which is CFA1. This site was considered an AMD-affected site because of its proximity to the Gold Finch Mine and location on the hillside, but the fact that it groups more closely with the AMD-unaffected samples warrants further investigation.
Differences in chemistry (pH, DO, conductivity) as well as dissolved metal concentrations (AMD-affected samples had elevated levels of calcium, iron, magnesium, manganese, and zinc compared to AMD-unaffected samples) were noted earlier. These differences justified the groupings of AMD-affected and AMD-unaffected sites, and these differences were also observed in the bacterial community analyses. This result suggests that AMD impacts the abundance of bacterial communities present within the ARD sediments of the Chattanooga Fen.
48


PC2- Percent variation explained 20.32%
PCoA • PCI VS PC2
PCoA - PCI vs PC2
020
015
010
005
000
-0.05
-0.10
-0.15
-0.20
-025
1 1 1 ■ • • • 1 * a P ■ ■ •
•
• ■
- ••
•
•
■ 1 1 1 • 1
4 -0.3 -0.2 -0 1 00 01 0
025
020 -
N?
er-
rs
o 015 in
i.»L
005 •
>
c
OJ
u
L.
0)
Cl
0.05 â– 
IN
u
0. 010
015
' 1 1 1 1 i
•
•
' '
• ■
• ■
• •

• •
•
I*
• ■
â– 
» ■
•
a a
• i 1
â–  i i â–  1 i
PCI - Percent variation explained 42 35%
-0 20 -0 15 -0.10 -0 05 0 00 0 05 010 0
PCI • Percent variation explained 17.69%
Affected Samples
Unaffected Samples
Figure 4.4. Principle Component Analysis of weighted (left) and unweighted (right) UniFrac distances for AMD-affected and unaffected samples. The more similar the bacterial communities are in terms of taxonomy, the closer the samples will cluster within the planes of the plot. The AMD-affected samples are in closer proximity to each other than to the AMD-unaffected samples. The affected sample which is the closest to the unaffected samples in £ both the weighted and unweighted graph is CFA1 which is the site that sits closest to the unaffected sampling sites


Table 4.6. Resulting ANOSIM values when comparing affected and unaffected samples using both weighted and unweighted UniFrac values in order to determine the effect of AMD on the bacterial community structures and abundance. The R-stat and p-values show that the samples within AMD-affected samples are more similar to eachother than to the bacterial communities within the AMD-unaffected samples
ANOSIM R-stat p-value
Affected/Unaffected [weighted] 0.304 0.01
Affected/Unaffected (unweighted] 0.415 0.001
Table 4.7. Resulting ANOSIM values when comparing July and September samples using both weighted and unweighted UniFrac values in order to determine the effect of month of the bacterial community structures and abundance. The R-stat and p-values show that the samples from July and September are all similar to each other. This implies that time, within the context of the two sampling periods, did not influence the bacterial community structure
ANOSIM R-stat p-value
July/September (weighted] 0.0126 0.356
July/September (unweighted] 0.0186 0.352
50


Carbon Source Usage
Following incubation of the AMD-affected and unaffected samples in the EcoPlates, the AMD-affected samples utilized all six of the different carboxylic acids present in the plates (Table 4.8). However, each of the AMD-affected samples had a lower usage of carbon sources compared to the unaffected samples using an average of 64.5% of the total carbon sources available (standard deviation of 9.6%). The AMD-unaffected samples utilized all of the carbon sources to some extent. Carbohydrates had the largest OD readings by the organisms within the AMD-unaffected sediments, which correlates to heavy usage of the carbon sources by the microorganisms within the wells. Xylose and cellobiose, in particular, had the highest OD readings among all of the sample types (an average OD reading of 2.36 for AMD-affected samples and 2.42 for AMD-unaffected samples), and so were selected for further analysis.
51


Table 4.8. EcoPlate analysis of microbial carbon substrate utilization. Xylose and cellobiose are highlighted because they were the only substrates used to the maximum in all sediment samples
OD < 0.5: -; 0.5 < OD < 1.0: +; 1.0 < OD < 1.5: ++; 1.5 < OD < 2.0: +++, 2.0 < OD: ++++
Guild Carbon Source Affect ed-luly Unaffected- ly Affected- Sept Unaffected- Sept
Amine Phenylethylamine + + - + +
Putrescine + + + + +
Amino Acid L-Arginine + + + + + +
L-Asparagine - + - ++
L-Phenylalanine + + + + - +
L-Serine - + + - + + +
L-Threonine + + + + + + +
Glycyl-L-Glutamic Acid + + + + + + +
Carbohydrate B-Methyl-D- Glucoside + + + + - + +
D-Galactonic Acid y-Lactone - + + + + +++
D-Xylose ++++ + + + + + + + + + + + +
i-Erythritol - + + + + + +
D-Mannitol ++ + + + + + + + + +
N-Acetyl-D- Glucosamine " + + + - + + + +
D-Cellobiose ++ + + + + + + + + + + + + + +
Glucose-1-Phosphate + + + + - + +
u-D-Lactose + + + + + + + + + + +
D,L-a-Glycerol Phosphate + + + + - + +
Carboxylic Acid Pyruvic Acid Methyl Ester ++ + + + + + +
D-Galacturonic Acid + + + + + + + + + +
y-I lydroxybutyric Acid ++ + + + + +
D-Glucosaminic Acid + + + + + + + + +
Itaconic Acid ++ + + + + + +
cx-Ketobutyric Acid ++ + + + + + + +
D-Malic Acid ++ + + + + + + + + +
Phenolic Compounci 2-Hydroxy Benzoic Acid + + + + - + +
4- Hydroxy Benzoic Acid ++ + + + + + +
Polymer Tween 40 + + + + +
Tween 80 + + + + + + + +
a-Cyclodextrine + + + + + +
Glycogen + + + + ++ ++
52


Average well color development (AWCD) was used to correlate the darker purple color (greater usage of the carbon source within the well) with the amount of carbon sources being used by the organisms within each sample (Figure 4.5). AWCD data showed that the AMD-unaffected samples from both July and September utilized more of the included carbon sources within the EcoPlates to a greater degree (average OD readings of 1.97) compared to the microorganisms within the AMD-affected sediment samples (average OD reading of 1.21; t-testp-value=0.01).
The results of the EcoPlate analyses revealed that the AMD-affected and AMD-unaffected samples from the Chattanooga Fen contained organisms which differed from each other in not only the carbon substrates utilized but in AWCD as well, signifying a difference in amount of the carbon sources used (Table 4.8 and Figure 4.5). The affected samples showed greater substrate usage within the carboxylic acids while the unaffected samples showed greater carbohydrate usage.
The observed results were similar to a recent EcoPlate to determine the ecotoxicological impact of metal exposed dairy sewage sludge on the microbial communities (Gryta et al., 2014). Through their analyses, it was shown that metal exposure resulted in deterioration of overall microbial activity and diversity of substrate utilization. In non-metal exposed samples, all substrates were utilized compared to only two substrates being utilized to the maximum potential by the metal exposed sludge communities. This may have been related to a reduction in catabolic functions. It is important to note that the study above examined metal impact on previously unexposed communities.
53


In the present study, however, there were two common plant sugars utilized to the maximum degree (>2.5 OD) by all sediment samples; xylose and cellobiose. The fact that both were utilized by all samples led to the following objective of the project which was to determine the abundances of three glycoside hydrolase genes which are involved in the degradation of both types of plant sugars.
Cellobiose is a disaccharide which can be obtained from the enzymatic hydrolysis of cellulose, the main structural component of plant cell walls (35-50% of a plant’s dry weight) which constitutes a large part of soil organic matter. Factors which control the rate of cellulose degradation are not well understood. Enzymes which degrade cellulose are described as glycoside hydrolases (GH) in at least ten GH families (GH 1, 3, 5, 6, 8, 9,12, 44, 45, and 48). Within these GH families, there are three types of proteins that are active on (3-1,4 glycosidic bonds. There are endocellulases that are active on internal bonds, exocellulases that degrade the polymer from its extremities, and (3-glucosidases that produce glucose from cellobiose (Merlin etal, 2014). There are three GH families which contain enzymes with only endoglucanase and/or exoglucanase activities; GH6, GH45, and GH48). Endoglucanase and exoglucanase are involved in the early steps of cellulose degradation which is when the polymer is broken down to cellobioses (made up of two glucose molecules). Microorganisms actively expressing endoglucanases and/or exoglucanases are called "cellulose degraders”, and use cellulose as a carbon and energy source for growth and are widely distributed within the environment (de Menezes etal., 2015).
54


Xylan is the major component of hemicellulose, the second most abundant polysaccharide in plant cell walls behind cellulose. A large variety of cooperatively acting enzymes are required to completely hydrolyze the linear polysaccharide xylan into xylose. These are essential enzymes for microorganisms for their ability to degrade plant material into usable nutrients. The enzymes include endo- l,4-(3-D-xylanase, (3-D-xylosidase, a-D-glucuronidase acetyl xylanesterase, and several others. Xylanases within GH10 are the most abundant in the environment, substrate specific to xylan, more common in bacteria than fungi, and are larger than the other GH xylanases (e.g. GH 5, 7, 8,11). Xylanases are also studied because of their potential applications in agriculture and industry. For example, they have been used in the papermaking process for bleaching of wood pulp and for food additives for poultry. In the future, they could also be used in the generation of biofuel from unusable plant material (Wang et al, 2010).
55


Average Well Color Development (OD)
1
0.8
0.6
0.4
0.2
0
-0.2
Average Well Color Development
M 48 72 96 120
Hours Post Inoculation
UNAFFECTED-July AFFECTED-July - — - UNAFFECTED-September - — - AFFECTED-September
Ln
09
Figure 4.5. Average Well Color Development (AWCD) of the metabolized substrates within the BIOLOG EcoPlates for affected and unaffected samples at specified incubation times (n=5 for AFFECTED-July, AFFECTED-September, and UNAFFECTED-September; n=4 for UNAFFECTED-July).


In order to identify the organisms utilizing the xylose and cellobiose, DNA was extracted and sequenced from the xylose and cellobiose wells of the EcoPlates. There were six different OTUs that were the most common from both the xylose and cellobiose wells as well as between affected and unaffected samples. These OTUs included Oxalobacteraceae, Enterobacteriaceae, Gammaproteobacteria, Pseudomonas, Paenibacillus, and Cellulomonadaceae. The diversity within AMD-affected versus AMD-unaffected samples, xylose versus cellobiose, and July versus September samples were not different. This reflects that the heterotrophic, culturable organisms within all of these samples were similar whether AMD-affected or unaffected (Supplemental Figures 1-4, Supplemental Table 1). This suggests little impact of AMD on the culturable organisms utilizing xylose and cellobiose from the Chattanooga Fen.
Quantitative PCR
Common sugars utilized in all of the BIOLOG EcoPlates, both affected and unaffected, were the plant sugars cellobiose and xylose. Enzymes which are responsible for the degradation of both sugars are members of the super-family of glycoside hydrolases. Because of this, this study aimed to determine the amount of three glycoside hydrolase genes present in affected and unaffected sediment samples collected from the Chattanooga Fen.
Glycoside hydrolases are a large and complex group of enzymes and the fact that, in many GH families, there is a presence of multiple substrate specificities, the design of molecular tools (e.g. qPCR primers) for an in-depth investigation into their environmental role has been challenging. However, for GH6 and GH48, all of the
57


functionally characterized enzymes have been shown to target the degradation of cellulose. A majority of the enzymes function as cellobiohydrolases and endoglucanases, major components of multienzyme cellulolytic systems often acting in synergy with other cellulases to depolymerize cellulose.
Glycoside Hydrolase 6 and Glycoside Hydrolase 48
Quantitative PCR for the glycoside hydrolase 6 (GH6) gene was found in higher
abundance in AMD-unaffected samples compared to affected with an average of 3.28 x 105 copies per gram of sediment (standard deviation of 264129.4) and 1.10 x 104 copies per gram of sediment (standard deviation of 11566.3), respectively (Figure 4.6A). The Gold Finch Mine sample collected during September had values below the detectable limit, and, therefore, was not included in these analyses. When a zero is used for the September mine sample, the average copy per gram of sediment of GH6 drops to 9816.5. The two groups were statistically different from each other, both when the zero is not included and when it is included (t-test p-value= 0.002 without the zero included and 0.001 when the zero is included). There was one outlier in the unaffected group (914078 copies/gram of sediment), which came from CFUl-July. The July Gold Finch Mine sample had the lowest amount of GH6 copies per gram of sediment (806 copies/gram of sediment) of the samples that amplified, which implies similarity to the AMD-affected samples.
Interestingly, the GH6 primers were previously utilized to quantify the amount of GH6 cellulase gene present in different soils, the peatland environment with a pH of 6.43 averaged 1.89E+08 copies per gram of soil (Merlin et al., 2014). This study found that soil acidity decreased the abundance of the GH6 gene, implying acid toxicity. The
58


abundance of GH6 in the present study implies acid toxicity in both the AMD-affected and unaffected samples.
The results of the glycoside hydrolase 48 (GH48) gene abundance were similar to those of the GH6 gene. The affected samples had a significantly lower amount of the GH48 gene compared to the unaffected samples (t-test p-value= 0.0006) (Figure 4.6B). The affected samples had an average of 263928 copies per gram of sediment (standard deviation of 202486) compared to the unaffected sample which had an average of 2173712 copies per gram of sediment (standard deviation of 1249504). The mine samples from both July and September had amounts of GH48 that were below the detectable limit, therefore, they were not included in the analysis. The largest amount of GH48 came from CFUA-July with 37427058 copies per gram of sediment.
According to quantitative PCR analysis, both GH6 and GH48 genes were found in higher amounts in the AMD-unaffected samples. This implies an AMD impact on the types of genes being used for the degradation of cellulose or the presence of genes available for cellulose degradation, despite the identification of similar degrading bacterial species found in the EcoPlates (Supplemental Figures 3 and 4). While a study addressing the impact of AMD on gene abundance in metal-impacted wetland systems could not be found, a study looking at the effect of AMD-exposure on soil microbial communities found a corresponding decrease in the abundance of the GH6 gene in acidic soils. Interestingly, the levels of GH6 found in this study in both the AMD-affected and unaffected samples were lower than those found in the soil study implying not only a pH effect but also a potential metal effect.
59


Glycoside Hydrolase 10
When qPCR was performed for the GH10 gene, the results were again similar to the results for GH6 and GH48. The unaffected samples had a significantly larger amount of the GH10 gene per gram of sediment compared to the affected samples (t-test p-value= 0.0002) (Figure 4.6C). The average copy per gram of sediment for unaffected samples was 5853261 (standard deviation= 1963014). Both the mine samples were again below the detectable limit; therefore, those samples were not included in the analyses. The affected samples had an average of 1393280 copies of GH10 per gram of sediment (standard deviation= 1465770). The lowest amount of GH10 within affected samples when the mine samples are not included was from CFAA-July (387261 copies/gram of sediment). The greatest amount of GH10 copies came from CFUl-July (8586215 copies/gram of sediment).
A recent study which examined the amount of GH10 in alpine tundra soil in the Tianshan Mountains of China determined that the main xylanotic bacteria to produce xylanases in this environment were Actinobacteria, Proteobacteria, Verrucomicrobia, and Firmicutes, all of which were present in the current study. The GH10 copy numbers have not been previously studied in fen environments, or acidic fen environments, but there have been several xylanases from GH families 10 and 11 isolated from acidophilic microorganisms (Collins et al., 2005).
All of the glycoside hydrolase gene abundances which were measured in this study (GH6, GH48, and GH10) were different between AMD-affected and AMD-unaffected samples. The decreased abundance of these genes in the AMD- affected sites
60


could result in the higher observed peat material in the AMD-affected versus unaffected sites, or, perhaps, the organisms which survive in the AMD-impact have different sequences for these genes then what is generally observed, and the primers utilized in this study were unable to amplify these GH genes within those samples. The results of this aspect of the conducted study show that AMD may impact the degradation of carbon, which is essential monitor in this type of ecosystem because peatlands are known to be large carbon sinks.
61


GH48 Quantities
Affected Unaffected
Condition
Figure 4.6. qPCR results from GH6(A), GH48(B), and GH10(C) amplification. For all amplified GH genes, the AMD-unaffected samples had a greater abundance of the gene present compared to the AMD-affected samples. For GH6 qPCR, the September CFMine sample did not amplify, and was not included in analysis (AMD-affected and AMD-unaffected n=9). For GH48 and GH10, the mine samples did not amplify, and were not included in the analysis (AMD-affected n=8, AMD-unaffected, n=9)
62


CHAPTER V
CONCLUSIONS AND FUTURE DIRECTIONS
Fens are essential carbon sinks which remain largely understudied. The Chattanooga Fen is a rare, acidic iron fen which has yet to be microbiologically characterized. This unique ecosystem provided the prime opportunity to explore the taxonomic structure and metabolic profiles of bacterial communities which have previous, long-term metal exposure (ARD), and how some of these communities have adapted to the presence of AMD flowing into portions of the Fen from the Gold Finch Mine.
This study found that the bacterial communities which live in the Chattanooga Fen are as unique as the ecosystem in which they survive and distinct from each other depending on their location within the Fen. The bacterial communities that live directly in the acidic, mine effluent that impacts a portion of the Fen include Gallionella and Bacteriovoracacea which are unique in that they have adapted to utilize nutrients available to them in their environment (i.e. iron and other Gram negative organisms). The AMD flowing into lower elevation ponds within the Fen is also contributing other Beta- and Deltaproteobacteria to the already metal-adapted bacterial communities. The ponds unaffected by AMD, with the lowest pH include Acidobacteria which thrive in acidic environments.
Not only is the taxonomy distinct between AMD-affected and AMD-unaffected bacterial communities within the Chattanooga Fen, but their metabolic profiles are also
63


unique. The heterotrophic microorganisms within AMD-unaffected ponds utilized all 31 of the provided carbon sources to a greater degree within the BIOLOG EcoPlates compared to the affected communities. However, there were two common sugars utilized (xylose and cellobiose) in all plates inoculated with samples taken from the Fen. This led to the final portion of the conducted study which was to determine the abundance of three glycoside hydrolase genes (GH6, GH48, and GH10) that aid in the degradation of xylose and cellobiose. The AMD-unaffected samples contained a significantly greater amount of all three glycoside hydrolase genes compared to the AMD-affected samples. These findings are a possible explanation for the increased build-up of peat within AMD-affected sites compared to AMD-unaffected sites at the Fen.
The impact of AMD on the previously-metal adapted bacterial communities within the Chattanooga Fen was reflected in bacterial community taxonomy, carbon source usage, and abundance of glycoside hydrolase genes involved in the degradation of xylose and cellobiose. The results of this study were similar to previous studies examining the impact of AMD on non-ARD impacted bacterial communities in terms of organisms present (i.e. Deltaproteobacteria in AMD sediments) and ARD impacted communities (i.e. Acidobacteria within the ARD impacted AMD-unaffected sediments of the Fen). AMD impact was also observed in carbon source usage and potential carbon source degradation (glycoside hydrolase gene abundance). This study showed that AMD negatively affects the usage of organic carbons as well as its potential degradation. This information may influence future management decisions of the Fen because monitoring the degeneration of carbon is essential in this large carbon sink.
64


Though the results of this study provided much needed insight into the microbiology of the Chattanooga Fen, there are still several future studies which could be conducted in examining this unique system. Additional aspects of the microbiology of the Chattanooga Fen that remain to be addressed include examining the potential impact of AMD on other groups of microorganisms, such as the fungi as major players in organic carbon decomposition, and conducting a more detailed analysis of gene expression of the Fen microbial communities through the use of transcriptomics.
Comparative studies amongst iron fen could provide valuable information regarding the consistency of AMD-impacts, as well as ecosystem level implications of fen deterioration due to their importance as carbon sinks.
65


SUPPLEMENTAL INFORMATION
o.o
^.Pseudomonas
c_Gammaproteobacteria
c.Gammaproteobacteria
f.Enterobacteriaceae
^.Pseudomonas
g_Pseudomonas
f.Enterobacteriaceae
f.Enterobacteriaceae
^.Pseudomonas
^.Pseudomonas
f.Oxalobacteraceae
f Cellulomonadaceae
Up UO up up up T* 7» T* up
X X X X X X M 5 X (N X X
00 D u. U D U. U cf) D u. U (N D u. U H D LL U Jj D u. U D u. U D u. U D u. U U < u. U
Lrt tA
X CD X < X f-i < X 00 < 1 x (N < X >-4 < up X 2 u.
LL u LL. u LL u LL u LL u LL U LL U X & c U
i
u
Supplemental Figure 1. Top twelve OTUs from 16s rDNA amplified out of the xylose wells of the EcoPlate
66


g_Pseudomonas
c.Gammaproteobacteria
g_Pseudomonas
^.Pseudomonas
^.Pseudomonas
c.Gammaproteobacteria
f_Enterobacteriaceae
f.Enterobacteriaceae
g_Pseudomonas
f.Oxalobacteraceae
^.Pseudomonas
f Cellulomonadaceae
«/l to in to -p -p -p -p m to to to *p
u CD D U D u n D U LL U U. U u_ U LL u LL u U U u u LL u LL u U LL u u u u u
Supplemental Figure 2. Top twelve OTUs from 16s rDNA amplified out of the cellobiose wells of the EcoPlate
67


0.0
g_Pseudomonas; s_umsongensis
f_Cellulomonadaceae g_______Paenibacillus
f_Oxalobacteraceae
f_Enterobacteriaceae
g_Pseudomonas
g_Pseudomonas
c_Gammaproteobacteria
g_Pseudomonas
f_Enterobacteriaceae
f_Enterobacteriaceae
c_Gammaproteobacteria
f_Enterobacteriaceae
g_Pseudomonas
g_Pseudomonas
l/l CO CO CO
u X U X
u u CQ CQ
< < < <
LL LL LL LL
u U u u
CO l/) CO CO -p -p ~p ~p —p
1 u >< (J >< u X u X u X u X
fM fM 1 CD CQ < < rsi (N T—1 tH
< < < < < < < < < < < <
LL LL LL LL LL LL LL LL LL LL LL LL
U U U u U U U U U u U U
7 T u x
a) cu qi
c £ £
n ll Life U u

- -2.4
- -2.8
- -3.2
- -3.6
Supplemental Figure 3. Top 15 OTUs from the 16S rDNA extracted from the xylose and cellobiose wells within EcoPlates inoculated with affected samples. The samples which have an X were extracted from xylose wells and the samples which include a C were extracted from cellobiose wells. The samples which were collected in July are denoted with a J at the end, and the samples which were collected in September are denoted by an S at the end
68


0.0
f_Enterobacteriaceae
f_Enterobacteriaceae
f_Enterobacteriaceae
g_Pseudomonas c_Gammaproteobacteria c_Gammaproteobacteria g_Pseudomonas f_Enterobacteriaceae g_Pseudomonas g_Pseudomonas; s_umsongensis g_Pseudomonas
g_Pseudomonas g__Paenibacillus
f_Oxalobacteraceae f Cellulomonadaceae
CO 00 00 (/} W tfi V, CO 00 00 •-p ►7* ►7* ►7' *-p >-p
1 u 1 X 1 u 1 X i U 1 X i U 1 X 1 u 1 X u X u >5 u X u X
OQ OQ ro 00 CN H CQ DQ < <. CN CN H rH
D D D D 0 D D D D D D D D D D D
Uh u. u. u- tL. Uh u- t-L b u. do tL. u. u. u. u. do
u U u u u U u u u u u U u u u u u U
-0.4
-0.8
-1.2
-1.6
-2.0
-2.4
-2.8
-3.2
-3.6
Supplemental Figure 4. Top 15 OTUs from the 16S rDNA extracted from the xylose and cellobiose wells within EcoPlates inoculated with unaffected samples. The samples which have an X were extracted from xylose wells and the samples which include a C were extracted from cellobiose wells. The samples which were collected in July are denoted with a J at the end, and the samples which were collected in September are denoted by an S at the end
69


Supplemental Table 1. Alpha diversity metrics for all samples of the same material grouping where Chaol estimate, Observed OTU richness, Faith’s phylogenetic diversity, and Shannon Index were performed and a t-test was done to test if the mean value for each group was significantly different. The data shown is the average of all samples belonging to each group with the standard deviation in parentheses
Chaol Shannon Index Faith’s PD Observed OTUs
Affected Samples 90.96(60.18) 3.71(1.25) 3.10(1.53) 31.16(13.59)
Unaffected Samples 84.57(51.55) 3.59(1.43) 3.37(1.83) 30.22(13.78)
July Samples 96.95(48.27) 3.61(1.33) 3.67(1.71) 31.94(13.85)
September Samples 79.81(61.62) 3.69(1.36) 2.85(1.56) 29.6(13.44)
Xylose Samples 88.60(66.02) 3.37(1.46) 3.04(1.74) 28.54(14.15)
Cellobiose Samples 87.27(44.62) 3.93(1.15) 3.43(1.61) 32.88(12.84)
t- test p-value Affected/Unaffected 0.736 0.798 0.632 0.838
f-test p-value July/September Samples 0.362 0.862 0.134 0.609
t- test Xylose/Cellobiose Samples 0.944 0.205 0.490 0.342
70


REFERENCES
Andersen R, Chapman S, Artz R. 2013. Microbial communities in natural and disturbed peatlands: A review. Soil Biology and Biochemistry 57:979-994.
Andersen R, Grasset M, Thormann L, Rochefort AJ, Francez. 2010. Changes in microbial structure and function following Sphagnum peatland restoration. Soil Biology & Biochemistry. 42: 291-301.
Andersson S, Dalhammar G, Kuttuva Rajarao G. 2011. Influence of microbial interactions and EPS/polysaccharide composition on nutrient removal activity in biofilms formed by strains found in wastewater treatment systems. Microbiol. Res. 166:449-57.
Artz RRE, Baird AJ, Belya LR, Comas X, Reeve AS, Slater LD. 2009. Microbial community structure and carbon substrate use in Northern peatlands. Geophysical Monograph Series. 184.
Baker BJ, Banfield JF. 2003. Microbial communities in acid mine drainage. FEMS Microbiol. Ecol. 44:139-52
Banfalvi G. 2011. Heavy Metals, Trace Elements and Their Cellular Effects. Cellular Effects of Heavy Metals 3-28.
Bernhardt ES, Lutz BD, King RS, Fay JP, Carter CE, Helton AM. 2012. How many mountains can we mine? Assessing the regional degradation of Central Appalachian rivers by surface coal mining. Environ Sci Technol 46: 8115-8122.
Barry, K. L., J. A. Grout, C. D. Levings, B. H. Nidle, and G. E. Piercey. 2000. "Impacts of acid mine drainage on juvenile salmonids in an estuary near Britannia Beach in Howe Sound British Columbia." Canadian Journal of Fisheries and Aquatic Sciences 57(10): 2031-2043.
Braud A, Hoegy F, Jezequel K, Lebeau T, Schalk IJ. 2009. New insights into the metal specificity of the Pseudomonas aeruginosa pyoverdine-iron uptake pathway. Environ. Microbiol. 11:1079-91.
Bruins MR, Kapil S, Oehme FW. 2000. Microbial resistance to metals in the environment. Ecotoxicol. Environ. Saf. 45:198-207.
Chen, H., Young, S., Berhane, T. K., & Williams, H. N. (2012). Predatory bacteriovorax communities ordered by various prey species. PLoS ONE, 7(3).
Chen, L. xing, Huang, L. nan, Mendez-Garcia, C., Kuang, J. bang, Hua, Z. shuang, Liu, J., & Shu, W. sheng. (2016). Microbial communities, processes and functions in acid mine drainage ecosystems. Current Opinion in Biotechnology.
71


Chimner RA, Cooper DJ. 2006. A Report on the Restoration Potential of Chattanooga Fen, San Juan’s, Colorado.
Cooper DJ, Nydick K, Lemly J. 2008. Final Report: Regional Assessment of Fen Distribution, Condition, and Restoration Needs, San Juan Mountains.
David CP. 2003. Heavy metal concentrations in growth bands of corals: A record of mine-tailings input through time (Marinduque, Philipines). Marine Pollution Bulletin 46:187-196.
de Menezes, A. B., Prendergast-Miller, M. T., Poonpatana, P., Farrell, M., Bissett, A., Macdonald, L. M.,... Thrall, P. H. (2015). C/N ratio drives soil actinobacterial cellobiohydrolase gene diversity. Applied and Environmental Microbiology, 81(9), 3016-3028.
Duggan, J. 2009. Longevity of Acid Rock Drainage: Mineralogical and Chemical Comparison of Mine- Waste Piles and Post- Glacial Talus Rock Producing Acidic Solutions. Colorado State University Department of Geosciences. Presentation given during Portland GSA Annual Meeting.
Elliott P, Ragusa S, Catcheside D. 1998. Growth of sulfate-reducing bacteria under acidic conditions in an upflow anaerobic bioreactor as a treatment system for acid mine drainage. Water Research 32:3724-3730.
Fisk M, Ruether K, Yavitt J. 2003. Microbial activity and functional composition among northern peatland ecosystems. Soil Biology & Biochemistry. 35: 591-602.
Frank KL, Rogers DR, Olins HC, Vidoudez C, Girguis PR. 2013. Characterizing the distribution and rates of microbial sulfate reduction at Middle Valley hydrothermal vents. The ISMEJournal ISMEJ. 7(7):1391-1401.
Gadd GM. 2010. Metals, minerals and microbes: geomicrobiology and bioremediation. Microbiology 156:609-43.
Giller KE, Witter E, Mcgrath SP. 1998. Toxicity of heavy metals to microorganisms and microbial processes in agricultural soils: a review. Soil Biology and Biochemistry. 10-11:1389-1414.
Giyta, A., Frqc, M., & Oszust, K. (2014). The Application of the Biolog EcoPlate Approach in Ecotoxicological Evaluation of Dairy Sewage Sludge. Applied Biochemistry and Biotechnology, 174(4), 1434-1443.
Hauff PL, Prosh, EC, Coulter DW, Sares, MA, Bird DA, Peters DC, and Henderson FB, III. 2003. Evaluating the impact of natural acid drainage on Colorado watersheds using hyperspectral remote sensing, Proceedings of the tenth international conference on tailings and mine waste. 3-7.
72


Hirsch PR, Mauchline TH, Clark IM. 2010. Culture-independent molecular techniques for soil microbial ecology. Soil Biol. Biochem. 42:878-887.
Hodson ME. 2012. Effects of Heavy Metals and Metalloids on Soil Organisms. Environmental Pollution Heavy Metals in Soils 141-160.
Hogsden KL, Harding JS. 2012. Anthropogenic and natural sources of acidity and metals and their influence on the structure of stream food webs. Environ. Pollut. 162:466-74.
Johnson DB, Hallberg KB. 2003. The microbiology of acidic mine waters. Research in Microbiology 154:466-473.
Jones, D. S., Kohl, C., Grettenberger, C., Larson, L. N., Burgos, W. D., & Macalady, J. L. (2015). Geochemical niches of Iron-oxidizing acidophiles in acidic coal mine drainage. Applied and Environmental Microbiology, 81(4), 1242-1250.
Kamika I, Momba MN. 2013. Assessing the resistance and bioremediation ability of selected bacterial and protozoan species to heavy metals in metal-rich industrial wastewater. BMC Microbiology BMC Microbiol 13-28.
Lin, X, Green S, Tfaily M, Prakash O, Konstantinidis, KT, Corbett JE, Chanton JP, Cooper WT, and Kostka JE. 2012. Microbial Community Structure and Activity Linked to Contrasting Biogeochemical Gradients in Bog and Fen Environments of the Glacial Lake Agassiz Peatland." Applied and Environmental Microbiology 78.19. 7023-031.
Mendez-Garcia, C., Pelaez, A. I., Mesa, V., Sanchez, J., Golyshina, O. V., & Ferrer, M. (2015). Microbial diversity and metabolic networks in acid mine drainage habitats. Frontiers in Microbiology.
Merlin, C., Besauiy, L., Niepceron, M., Mchergui, C., Riah, W., Bureau, F.,... Bodilis, J. (2014). Real-time PCR for quantification in soil of glycoside hydrolase family 6 cellulase genes. Letters in Applied Microbiology, 59(3), 284-291. Wang, G., Wang, Y., Yang, P., Luo, H., Huang, H., Shi, P., ... Yao, B. (2010). Molecular detection and diversity of xylanase genes in alpine tundra soil. Applied Microbiology and Biotechnology, 87(4), 1383-1393.
Mikiya Hiroki. 1992. Effects of heavy metal contamination on soil microbial population. Soil Science and Plant Nutrition. 38:1,141-147.
Monsieurs P, Moors H, Van Houdt R, Janssen PJ, Janssen A, Coninx I, Mergeay M, Leys N. 2011. Heavy metal resistance in Cupriavidus metallidurans CH34 is governed by an intricate transcriptional network. Biometals 24:1133-51.
Myers B, Webster KL, Mclaughlin JW, Basibko N. 2011. Microbial Activity across a Boreal Peatland Nutrient Gradient: The Role of Fungi and Bacteria. Wetlands Ecology and Management. 20: 77-88.
73


Nies DH. 1999. Microbial heavy-metal resistance. Appl. Microbiol. Biotechnol. 51:730-50.
Quillet L, Besaury L, Popova M, Paisse S, Deloffre J, Ouddane B. 2011. Abundance, Diversity and Activity of Sulfate-Reducing Prokaryotes in Heavy Metal-Contaminated Sediment from a Salt Marsh in the Medway Estuary (UK). Marine Biotechnology Mar Biotechnol 14:363-381.
Ren N, Xie T, Xing D. 2009. Composition of extracellular polymeric substances influences the autoaggregation capability of hydrogen-producing bacterium Ethanoligenens harbinense. Bioresour. Technol. 100:5109-13.
Rouch DA, Lee BT, Morby AP. 1995. Understanding cellular responses to toxic agents: a model for mechanism-choice in bacterial metal resistance. J. Ind. Microbiol. 14:132-41.
Sackett, J. 2015. Comparative Microbial Ecology of Sediment- Associated Microbial Communities from Anthropogenically and Endogenously Metal Impacted Systems. Master’s Thesis.
Silver S. 1996. Bacterial resistances to toxic metal ions - a review. Gene. 179:9-19.
Silver S, Phung, LT. 2009 . Heavy metals, bacterial resistance. In Encyclopedia of Microbiology, pp. 220-227 . Edited by M. Schaechter. Oxford: Elsevier.
Sims A, Zhang Y, Gajaraj S, Brown PB, Hu Z. 2013. Toward the development of microbial indicators for wetland assessment. Water Res. 47:1711-25.
Swan BK, Martinez-Garcia M, Preston CM, Sczyrba A, Woyke T, Lamy D, Reinthaler T, Poulton NJ, Masland ED, Gomez ML, Sieracki ME, DeLong EF, Herndl GJ, Stepanauskas R 2011. Potential for chemolithoautotrophy among ubiquitous bacteria lineages in the dark ocean. Science 333 6047: 1296-300
Verplanck, P.L., ed., 2008, Understanding contaminants associated with mineral deposits: U.S. Geological Survey Circular 1328, p. 96
Wang Y, Shi J, Wang H, Lin Q, Chen X, Chen Y. 2007. The influence of soil heavy metals pollution on soil microbial biomass, enzyme activity, and community composition near a copper smelter. Ecotoxicol. Environ. Saf. 67:75-81
Zhu J, Zhang J, Li Q, Han T, Xie J, Hu Y, Chai L. 2013. Phylogenetic analysis of bacterial community composition in sediment contaminated with multiple heavy metals from the Xiangjiang River in China. Mar. Pollut. Bull. 70:134-9.
74


Full Text

PAGE 1

ELUCIDATING THE IMPACT OF METALS ON THE METABOLIC PROFILES AND TAXONOMIC STRUCTURE OF BACTERIAL COMMUNITIES WITHIN THE CHATTANOOGA FEN by KELSEY FOSTER B.S., Uni versity of Colorado Denver, 2014 A thesis submitted to the Faculty of the Gradu ate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Science Biology Program 2017

PAGE 2

i i © 2017 KELSEY FOSTER ALL RIGHTS RESERVED

PAGE 3

iii This thesis for the Master of Science degree by Kelsey Foster has been approved for the Biology Program b y Timberley M. Roane, Chair Annika C. Mosier Christopher Phiel December 16 , 2017

PAGE 4

iv Foster, Kelsey (M.S. , Biology Program ) Elucidating the Impact of Metals on the Metabolic Profiles and Taxonomic Struc ture of Bacterial Communities within the Chattanooga Fen Thesis directed by Associate Professor Timberley M. Roane ABSTRACT Fens, marshy wetlands with an accumulation of partially decomposed organic matter known as peat, are estimated to store up to a thi carbon. The Chattanooga Fen, located in the San Juan National Forest in southwestern Colorado, is anthropogenically and endogenously impacted by elevated concentrations of toxic metals. The anthropogenic source of metals to t he Fen is Acid Mine Drainage (AMD), acidic metal rich effluent formed by oxidized metals from the nearby Gold Finch Mine . The endogenous metal impact on the Fen is a result of subsurface acid rock drainage (ARD) formed from the natural oxidation of sulfur bearing minerals. While the entirety of the Fen is naturally impacted by ARD associated metals, the geographical terrain bisects the Fen into AMD affected and AMD unaffected portions. The chemical profiles of ARD and AMD are distinctly different from each other in terms of dissolved metal concentrations, pH, and anions. This study aimed to elucidate the impact of AMD in shaping the taxonomic diversity and carbon usage of the microbial communities within this rare, uncharacterized Fen. Sediment cores and wa ter samples were collected in July and September 2016 at the outflow of the Gold Finch mine, along an AMD gradient, and from naturally metal impacted sediments unimpacted by AMD. Illumina

PAGE 5

v high throughput sequencing of extracted 16S rDNA from sediments prov ided the taxonomic structure of the bacterial communities within the different areas of the Fen. Carbon sourc e usage was evaluated via BIOLOG EcoPlates and qPCR of three glycoside hydrolase (GH) genes. Data suggests that differences in both taxonomic struc ture and carbon source utilization can be explained by the presence of AMD. Sediments unaffected by AMD had a higher relative abundance of Acidobacteria, while Deltaproteobacteria was dominant in AMD affected sediments. The communities within the AMD affec ted sediments also utilized a higher proportion of carboxylic acids on the culture based EcoPlates, while the unaffected samples displayed a greate r usage of carbohydrates in terms of resulting optical density values . However, there were two commonalities between AMD affected and unaffected sediments which was the use of xylose and cellobiose. The average number of carbon sources utili zed by unaffected samples was 35.5 % greater than affected; the lowest number of substrates being used was observed with the direct acid mine effluent sample s . The abundance of all three GH genes was also greater in all AMD unaffected samples compared to AMD affected samples. The understanding of how metal source impacts the diversity and functional potential of microorganisms w ithin the Chattanooga Fen, and other similar fen systems, will be essential in influencing future management decisions. Approved: Timberley M. Roane

PAGE 6

vi ACKNOWLEDGEMENTS There are many people I would like to thank for their involvement in this project. First off, thank you to my ad visor Dr. Timberley Roane for not only giving me the opportunity to work in her lab and her guidance every step of the way, but for her continued support and kindness. She opened my eyes to the wonderful world of microbiology and introduced me to so many n ew opportunities, all the while setting an amazing example of how to c onduct yourself as a woman in the scientific field. Thank you to my committee members, Dr. Annika Mosier and Dr. Christopher Phiel for all of your assistance throughout this experience . Dr. Phiel gave me my first experience working in the lab, and I could not be more grateful. Dr. Mosier has been more than willing to go out of her way to advise me on so many lab techniques and has also become someone I greatly admire. I would also lik e to thank Jeff Boon for all of your assistance in the metal analysis. I appreciate all of the time you spent assisting in the project. Thank you to Sladjana Subotic and Andrew Boddicker for not only helping me through so many tough days in the lab, but for continuing to be amazing friends to me during some of the toughest periods of my life. The both of you have become my family, and I will never be able to put into words what your friendships means to me. I also appreciate you both along with Pedja Stoj anovic for baring though the freezing cold to collect samples with me, all the while keeping me laughing even though I could not feel my hands.

PAGE 7

vii Thank you to my parents, sister, brother, nephews, and nieces for continuing to motivate me throughout my coll egiate career. I could not have done any of this witho ut you all in my corner . You have made every step worth while, and I hope I have made you proud. You all are the reason I have become the person I am today. To Matthew Wayne, Sheldon Herman, Lucy Diam ond, and Kyrie Jane, thank you for taking care of me through everything. You have given me more love and support than I ever could have imagined. I love you with everything I have. Finally, thank you to all of my friends and colleagues who have all helpe d me keep pushing through in one way or another. Munira Lantz, Anna Scopp, Bhargavi Ramanathan, and Gabrielle Rietz, I will always be thankful for your friendship. You have all helped me in everything from teaching techniques to qPCR to reminding me that l aughter is always t he best medicine. You are wonderful human beings, and I hope you know what a large part you played in my graduate career.

PAGE 8

viii Table of Contents CHAPTER I . PROJECT OVERVIEW AND OBJECTIVES ................................ ................................ ............................ 1 II . THE CHATTANOOGA FEN ................................ ................................ ................................ ..................... 3 Sources of Metal Impact ................................ ................................ ................................ .......................... 5 Impact and Toxicity of Metals on Bacteria ................................ ................................ ............... 10 Microbiology of Peatland Systems ................................ ................................ ................................ ... 12 III . METHODS BACKGROUND ................................ ................................ ................................ ................. 16 16S rDNA High Throughput Sequencing ................................ ................................ ...................... 16 Metabolic Profiling: BIOLOG EcoPlates ................................ ................................ .......................... 21 Q uantification of Glycoside Hydrolase Genes ................................ ................................ .............. 23 IV . CONDUCTED STUDY ................................ ................................ ................................ ............................ 26 Introduction ................................ ................................ ................................ ................................ .............. 26 Methods ................................ ................................ ................................ ................................ ...................... 27 Sample Collection ................................ ................................ ................................ ............................... 27 DNA Extraction, PCR Amplification, and Next Generation Sequencing ........................ 29 Computational Analyses ................................ ................................ ................................ .................. 31 BIOLOG EcoPlate Preparation and Reading ................................ ................................ ............. 31 DN A Extraction and Amplification from EcoPlates ................................ ............................... 32

PAGE 9

ix Quantitative PCR ................................ ................................ ................................ ................................ 33 Results and Discussion ................................ ................................ ................................ ......................... 34 Site Chemistry ................................ ................................ ................................ ................................ ..... 34 Taxonomic Structure of Bacterial Communities within the Chattanooga Fen ........... 39 Carb on Source Usage ................................ ................................ ................................ ........................ 51 Quantitative PCR ................................ ................................ ................................ ................................ 57 V . CONCLUSIONS AND FUTURE DIRECTIONS ................................ ................................ ................. 63 SUPPLEMENTAL INFORMATION ................................ ................................ ................................ ......... 66 REFERENCES ................................ ................................ ................................ ................................ ................ 71

PAGE 10

1 CHAPTER I PROJECT OVERVIEW AND OBJECTIVES The Chattanooga Fen, located in the Colorado Mineral Belt is both endogenously and anthropogenically metal impacted , providing a rare opportunity to examine the impact of AMD on a naturally metal impacted site . A natural geological structure, the Chattan ooga Fen is formed by acidic, metal laden ground water. The natural metal impact within the Fen is known as acid rock drainage (ARD) formed from the oxidation of sulfur bearing minerals within the bedrock beneath the ponds formed at the Fen. The Chattanooga Fen is also bisect ed by the Gold Finch Mine, creating aci d mine drainage (AMD) affected and unaffected regions. No longer active , acid mine drainage from the Mine continues to introduce additional metals to the affected areas of the Fen. Whether endogenous ( ARD ) or anthropogenic (AMD), metals in the environment pose serious ecological risk and may affect a variety of cellular processes (Nies, 1999) . However, the response of endogenous ARD , metal adapted microbial communities to AMD exposure within this ecosystem have yet to be studied . The presented study aims t o evaluate the impact of AMD from the Gold Finch Mine on the indigenous metal adapted microbial communities. In order to characterize the bacterial community taxonomic structure , alpha and beta diversity metrices were performed on the resulting sequences obtained from sediment samples. Metabolic profiling was us ed to inform the use of quantitative PCR to identify differences in the expression of three glycoside hydrolase genes, as indicators of AMD impact. The specific objecti ves of the study conducted wer e to:

PAGE 11

2 Objective 1: Determine the taxonomy of bacterial communities within the Chattanooga Fen from ARD impacted sediment unaffected by AMD and affected by AMD Objective 2: Examine the carbon source usage of the bacterial communities within the Fen Object ive 3: Taxonomically characterize which bacteria were utilizing common carbon sources (xylose and cellobiose) between AMD affected and AMD un affected sediment communities within an ARD impacted site Objective 4: Quantify the amount of three glycoside hydr olase genes present in AMD affected and AMD unaffected ARD samples which encode enzymes necessary to degrade xylose and cellobiose

PAGE 12

3 CHAPTER II THE CHATTANOOGA FEN Fens are oligotrophic, minerotrophic peatlands, which are groundwater fed, nutrient poor marshy wetlands characterized by an accumulation of partially decomposed organic matter known as peat (Lin et al., 2012). Peatland environments contribute about 20% of t he t otal annual methane emissions to the atmosphere. These uniq ue peatland ecosystems a re large carbon sinks, containing about one third of all surface (Andersen et al., 2013). This can be explained by the imbalance between net primary production by plants and decomposition by microorganisms due to the consistently water saturated environment and consequent anoxia. The Chattanooga Fen, part of the Colorado Mineral Belt, is located in southwest ern Colorado near the old township of Chattanooga (Figure 2 .1) . Figure 2 .1. Map depicting the location of the fen in the San Juan Mountains between Ouray and Silverton (left) and a topographical map of the Chattanooga Fen with an arrow depicting the location of the ponds where sediment samples were taken in the condu cted study

PAGE 13

4 This largest Fen in the San Juan Mountains sits at an elevation of 10,500 feet located downhill from the Gold Finch Mine , and is fed by the Animas River Watershed (Cooper et al., 2008) (Figure 2 .2 ). Through carbon dating, this F en is predic ted to be around 600 years old with up to 3 meters of accumulated peat. T he environment is now home to a high diversity of plant species including the rare arctic peat moss, Sphagnum balticum , even though its main range is in Canada, more than 2,000 km away (C ooper et al., 2008). The water which saturates the soil in the Chattanooga Fen, as well as the pooled waters which form the ponds, has a net acidity, with a pH ranging from 3.8 to 6.4 , and is influenced by the bedrock with which the gro undwater comes into contact (Chimner and Cooper, 2006; Sackett, 2015). This is a trai t unique to the Chattanooga Fen. Most fens are non iron fens and are pH neutral to alkaline . Bogs, another type of wetland , are traditionally acidic, though they are fed by rainwater. Acidic, i ron fens are unique to Colorado and form in areas where the groundwater that feeds them has a naturally low pH due to weathering of iron pyrite and its oxidation to sulfuric acid. An inactive mine, the Gold Finch Mine, is loc ated above the Chattanooga Fen. The Gold Finch Mine is approximately 140 years old and was once used for mining gold and molybdenum (Stanton et. al, 2008) . Below the mine, a portal made of concrete was built, and acid mine drainage, originating from within the mine, exits the mine and enters into a portion of the Fen. The Chattanooga Fen is also naturally metal impacted in its entirety by acid rock drainage (ARD) which has formed from the dissolution of bedrock which has come into contact with the acidic gr oundwater source (Figure 2.2) . Due to the rarity of iron fens , as well as the metal contamination from both

PAGE 14

5 anthropogenic (the Gold Finch Mine) and endogenous sources, the Chattanooga Fen is a prime opportunity for the following comparative study into the effec t of AMD on previously ARD impacted ecosystems . Figure 2 .2 . The constructed portal below the inactive Gold Finch Mine above the Chattanooga Fen where the copper colored acid mine drainage enters portions of the Fen Sources of Metal Impact Metal contamination in the sediments of the Chattanooga Fen arose from both endogenous and anthropogenic sources. The anthropogenic source is a result of mining from the Gold Finch Mine which resulted in acid mine drainage (AMD). AMD is the formation and moveme nt of acidic water rich in heavy metals (Johnson and Hallberg, 2003). AMD in the Chattanooga Fen resulted from mining practices exposing pyrite, an iron old, which reacted with water and air to form sulfuric acid and dissolved iron via a series of oxidation reactions (Baker and Banfield, 2003). This reaction is summarized as: 2FeS 2 (s) + 7O 2 (g) + 2H 2 O(l) = 2Fe 2+ (aq) + 4SO 4 (aq) + 4H + (aq)

PAGE 15

6 Metal sulfide oxidation resulted in increased acidity, which resulted in dissolution of metal sulfides. Although metals tend to become more soluble with decreasing pH, ferric iron (Fe 3+ ) can precipitate out of solution at pH values betwee n 2.3 and 3.5 as iron hydroxide (Fe(OH) 3 ) or jarosite ( KFe³ (OH) (SO ) colored precipitate which can be observed at many AMD sites , including the Chattanooga Fen (Figure 2.2) (Ver planck, 2008) . AMD has a net acidity wh ich can further dissolve exposed ore material, releasing meta ls into the environment. AMD can be generated at both abandoned and active mines , including the Gold Finch Mine, and is a serious source of water pollution that can kill aquatic life, restrict st ream use, and damage water supplies. As of 2008, t he U.S. Environmental Protection Agency has estimated that in Colorado alone, approximately 1,300 miles of streams are affected by AMD; the negative impact of which on wildlife has been well documented (Bar ry et al., 2000; David, 2003; Savinov et al., 2003 ; Verplanck, 2008 ). A natural, or endogenous, source of metals to the Chattanooga Fen is known as acid rock drainage (ARD). This is also acidic water rich in metals, but this type is formed over time from t he natural oxidation or dissolution of sulfide minerals exposed to water and air as a result of endogenous processes (Hogsden and Harding, 2012). This source of natural metal impact can be observed in many Colorado watersheds , including the Animas River Wa tershed (Hauff et al., 2003). The pools of water wit hin the Chattanooga Fen are impacted by ARD via the welling of naturally acidic groundwater through metal rich sedimentary rocks which form the bedrock below the peat formation ( Chimner and Cooper, 2007) (Figure 2.3). Although AMD is often a more sig nificant concern than ARD due to the increased surface area of minerals exposed to weathering

PAGE 16

7 and oxidative conditions, both AMD and ARD result in the introduction of metal rich, acidic effluent into the Fen . W hile metals are ubiquitous in nature, the Chattanooga Fen is uniqu e in that it is impacted by ARD throughout as well as AMD in portions which receive mine effluent from the inactive Gold Finch Mine . Figure 2 .3 . Schematic of the formation or ARD (left) and AMD (right) . The Chattanooga Fen is impacted by both ARD from the naturally acidic groundwater which feeds the Fen as well as AMD because of the inactive Gold Finch Mine The Chattanooga Fen is bisected by the terrain into areas affected and unaffected by AMD, while the entirety is metal impacted via the acidic ground waters which feed the Fen (Figure 2.4). The sediments near AMD or ARD contaminated sites are often a repository for chemicals such as metals , which is why they were the samples taken for t his study. Once metals enter the sediment , they cannot be degraded, and therefore, they are a persistent environmental hazard (Bernhardt et al., 2012; Savinov et al., 2003). The long term impact of a high concentration of metals , from either AMD or ARD, o n microbial community structure is not well understood. Metals within the sediments of the Chattanooga Fen force microorganisms to adapt to the presence of metals, noting that microorganisms are more responsive to changing conditions than

PAGE 17

8 plants and animal s in the same area due to their high surface area to volume ratio and metabolic flexibility (Giller et al., 1998). It should be noted that the entirety of the Chattanooga Fen is ARD impacted. However, only portions of the Fen are AMD impacted. Therefore, in the current study affected samples refer to AMD affected ARD sediment samples, and unaffected samples refer to AMD unaffected ARD sediment samples.

PAGE 18

9 Figure 2.4. Diagram of the terrain of the Chattanooga Fen. The en d ogenously metal impacted Fen is bise cted into portions affected and unaffected by AMD where sediment samples were taken as seen in the bottom photograph of the diagram

PAGE 19

10 Impact and Toxicity of Metals on Bacteria As in other metal impacted systems, metals from both AMD and ARD in the Chattanoo ga Fen are expected to be toxic to cells. Once metals enter the cell via passive transport proteins or substrate specific transporters driven by ATP hydrolysis, the cell can be severely damaged. This is due to the strong ionic nature of metals which allows them to bind many cellular ligands having a bactericidal or bacteriostatic effect on the cell ( Bánfalvi , 2011). Metals can damage nucleic acids, disrupt protein structure and function, induce genetic mutations, inhibit membrane fluidity and function, and induce oxidative stress (Mikiya, 1992; Silver and Phung, 2009). It has been shown that morphology can also be affected, and certain species have been seen to convert from rod shaped (bacilli) to spheres (cocci) (Rouch et al., 1995). This result suggests th at the processes which regulate cell division and cell wall synthesis may be affected. In response to metals in their environment, bacteria have developed several metal resistance/tolerance mechanisms in order to keep metal concentrations below a physiolo gically toxic level (Figure 2. 5 ). Possible mechanisms developed by bacteria in the Chattanooga Fen include extracellular metal exclusion mechanisms and intracellular metal reduction mechanisms. The genes encoding these mechanisms are often found on plasmid s and can be transferred via horizontal gene transfer, or resistance can be obtained through mutations to the genome (Nies, 1999)

PAGE 20

11 Figure 2.5. Metal resistanc e mechanisms of microorganisms. Extracellular metal exclusion mechanisms are denoted by a , b, c; Intracellular metal removal mechanisms are denoted by d and e; Intracellular metal reactivit y reduction denoted by f and g (Gadd, 2010) Extracellular mechanisms include (Figure 2.5a, b, and c) : Extracellular polysaccharide substance (EPS) produc tion: Groups of bacteria can produce sticky EPS in the formation of biofilms and they are less susceptible to metals entering the cell Metal chelation: Metal chelating compounds will bind metal ions and prevent diffusion into the cell (e.g. siderophores) P recipitation: Mineralization of metals as salts and reduced metals a) b) c) d) e) f) g)

PAGE 21

12 Intracellular mechanisms include (Figure 2.5d, e, f , and g) : Metallothionein production: Metallothioneins are metal binding stress response proteins that protect against elevated levels of metals within the cell Efflux systems: Efflux is the active transport of metal cations out of the cytoplasm in order to reduce intracellular metal concentrations Enzymatic detoxification: The utilization of reduction oxidation mechanisms to decrease ch emical reactivity of metals within the cell Volatilization: Reduces cytosolic bioavailable concentrations of metals via enzymatic reduction Microbiology of Peatland Systems Despite the importance of peatland ecosystems in the global carbon cycle, no know n microbiological analyses have been performed at the Chattanooga Fen . In fact, the microbial composition of iron fens and fens in general remains great ly understudied. Culture dependent methods have traditionally been used to examine microbial communities within these ecosystems , however, this m ethod of identification requires that organisms were cultivable and easily identifiable in the laboratory. Many previous studies on fens and specifically iron fens in southwestern Colorado have

PAGE 22

13 utilized culturing t echniques to explore the microorganismal additions to the formation of AMD and ARD ( Baath and Anderson, 2003; Preston et al. 2012; Stanton et al., 2008) . In previous studies identifying dominate bacterial species with non iron fens, nearly one half of al l of the colonies grown on nutrient agar plates after one to two weeks of incubation were members of the fast growing Betaproteobacteria, including many from the genus Burkholderia (Gilbert and Mitchell, 2006; Reiche et al., 2008) . Other colonies commonly cultured from non metal impacted peatland environments included organisms belonging to Alphaproteobacteria, Actinobacteria, Bacteriodetes, and Gammaproteobacteria (Gilbert and Mitchell, 2006; Reiche et al., 2008; Loy et al., 2004). Cultivation of organisms is no longer a requirement for examining microbial communities due the advent of modern molecular biology techniques. This has renewed interest in examining microbial community responses to environmental disturbances such as the metal impact in the Chatta nooga Fen . Lin et. al (2012) utilized culturing independent methods to examine the differences in microbial communities within a bog (rainwater fed peatland) compared to those within a fen system (groundwater fed peatland) in the Glacial Lake Agassiz Peatl and of northwestern Minnesota. This study found that bacterial richness in the fen was almost twice as high as the bog (estimated using Chao1, an alpha diversity metric). Acidobacteria was dominant in the bog system, while Firmicutes dominated the fen . Ano ther study examined slightly acidic Canadian fen soil (pH of 4.5) in the lab, and determined that Acidobacteria, Nitrospirales, Alpha , Gamma , Deltaproteobacteria, and Cyanobacteria were regular inhabitants of the fen soil (Andersen et al., 2013) . 16S rRN A based stable isotope probing from another study determined that active xylose and glucose

PAGE 23

14 fermenting bacteria were present in an acidic, methane emitting fen in Canada. These organisms included Acidaminococcaceae, Aeromonadaceae, Clostdriciaceae, Enterob acteriaceae, and Pseuodomonadaceae (Hamberger et al., 2008). There have been few studies using culture independent methods to determine the taxonomy of the microbial communities within fens, especially not within metal impacted fens such as the Chattanoo ga Fen. The taxonomy of microbial communities within metal impacted systems in general have also been understudied, but the organisms which seem to dominate endogenously and anthropogenically metal impacted sites remain cons istent in these studies (Table 2 .1). The presence of so me of bacteria such as Sulfobacillus and Acidobacteria can be explained by the presence of increased sulfur concentrations and acidic environments. These organisms also show differences in function, as it has been observed that both iron and sulfur reduction is increased in AMD and ARD sites while carbon degradation was observed to be decreased overall in both site types (Chen et al., 2016). The study presented here utilized molecular techniques such as high throughput sequencing in a ttempt to understand the underlying bacter ial communities present in the Chattanooga Fen and to determine if the resulting Fen communities resembled those of a fen like or meta l impacted system.

PAGE 24

15 Ta ble 2 .1. Taxonomy of bacteria which dominate endogenous ly and anthropogenically metal impacted sites. Endogenous impact refers to ARD impacted sites while anthropogenic refers to AMD impacted sites

PAGE 25

16 CHAPTER III METHODS BACKGROUND 16S rDNA High Throughput Sequencing In this study, high throughput sequencing of a bacterial identification marker gene, the 16S rRNA gene , was used to taxonomically characterize the bacterial communities associated with the Chattanooga Fen (Riesenfeld, Schloss, and Handelsman 2004) . Upwards of 25 million paired end reads per run from the Illumina MiSeq sequencer al lows for massively parallel sequencing of DNA fragments. Targeted sequencing strategies are often used to sequence segments of the 16S rRNA gene (a gene that has been highly conserved in Prokaryotes throughout evolution) from mixed DNA samples in order to taxonomically characterize bacterial communities. To allow for taxonomic identification in bacter ial census studies, regions within the 16S rRNA gene whose sequence varies from species to species, known as variable regions, are targeted for sequencing and allow for taxonomic identification (Figure 3.1) . A general overview of the next generation 16S rRNA gene amplicon sequencing process is detailed be low (Figure 3.2 ) . Figure 3.1 . Representation of the 16S rRNA gene including conserved and variable regions . The hypervariable region 4 was amplified to be sequenced in the presented study . Image adapted from: www.biology.stackexchange.com

PAGE 26

17 Figure 3.2 : Broad procedural overview of 16S rRNA gene amplicon seq uencing using next generation sequencing technologies. Operation taxonomic units (OTUs) refer to defined). Image adapted from: www.neb.com Illumina DNA sequencing technology allows for multiple DNA molecules in a m ix ed sample to be sequenced concurrent ly (Caporaso et al. 2011) . T he hypervariable V4 regions of 16 S rRNA genes present in extracted DNA are amplified using specific forward and reverse primers. The forward primer anneals to the conserved region of the 16S rRNA upstream of the V4 region and (which hybridizes to t he flow cell during sequencing). Included is also a forward and reverse primer pad and primer linker used to link the Illumina adaptor with the

PAGE 27

18 forward or revers e primer and for annealing of sequencing primers during sequencing. The primer pads and linkers do not have homology with the region of the 16S rD NA adjacent to the V4 region . The reverse primer , containing a 12 nucleotide molecular barcode , known as a Go lay barcode , ement of the Illumina adaptor, and is used to identify to which sample a specific sequence read belongs and allows for multiple samples to be sequenced in a sin gle sequencing run (Figure 3.3 ). In an attempt to reduce reac tion specific PCR biases, these r eactions are performed in triplicate . Triplicate reactions are pooled and cleaned and concentrated to remove genomic DNA, excess primers, dNTPs, and other reaction components, leaving only the amplicon of interest. All sa mples are then pooled in equimolar concentrations, creating a sequencing library (a library is a sample containing multiple sequences) .

PAGE 28

19 Figure 3.3 . V4 region is amplified using F515 and 806R primers. Paired 150 base pair sequencing gives a 254 bp fragment with a 46 bp overlap. Figure adapted fr om Integrated DNA Technologies T he sequencing library is denatured using sodium hydroxide, diluted in hybridization buffer, and loaded into the reagent cartridge prior to the actual sequencing process . The diluted, denatured library is then dispensed onto the MiSeq sequencing flow cell , and the Illumina adaptors of the amplicon strands hybridize to the 3.4 ). After hybridization, a DNA strands are amplified. This results in the formation of up to 25 million unique clusters, each containing thousands of copies of a single amplicon . Following this step, DNAtemplates are copied using fluorescently labeled nucleotides which are then interpreted by a detector as specific nucleotides (Figure 3.4 ; The Illumina HiSeq 2000 Sequencing Technology 2015).

PAGE 29

20 Figure 3.4 . Cluster generation on the Illumina MiSeq sequencing flow cell. The DNA molecules are amplified u sing a process called bridge amplification, which results in the generation of up to 25 m illion unique clusters. Source: Illumina Inc. F igure 3.5 . DNA clusters are sequenced simultaneously using sequencing by synthesis technology . Fluorescently label ed nucleotides are incorporated one by one into the synthesis and a detector interprets the wavelength of light each fluorophore e mits as a specific nucleotide. Source: Illumina Inc.

PAGE 30

21 Bioinformatic and statistical tools, such as Quantitative Insights Into M icrobial Ecology (QIIME) and R software can be used t o analyze sequencing data generated from high throughput 16S rRNA ge ne sequencing (Caporaso et al. 2010; R Core Team 2014) . In order to quality filter sequencing reads, assign taxonomy to reads, and perform phylogenetic and statistical analyses, such as calculating alpha a nd beta diversity indices (diversity within samples and among samples ) , QIIME can be used . R software is a powerful statistical and programming language used to analyze virtually all data types. For sequencing data, R may be used to calculate multivariate statistics, including principal component analysis and generation of heatmaps and other plots to observe patterns among st samples. Metabolic Profiling: BIOLOG EcoPlates BIOLOG EcoPlates were created as a sensitive and reliable monitor of environmental cha nge , and they can be used for community level physiological profiling . EcoPlates were used here to examine carbon source usage (Figure 3.6 ). Figure 3.6 Representation of a BIOLOG EcoPlate post inoculation and incubation with a sediment sample. The wells will turn a purple color when electrons released by the organisms within the well reduce the included tetrazolium dye due to metabolism of the included carbon substrate

PAGE 31

22 Created for microbial community studies in ecological systems, EcoPlates measure the a bility of microorganisms to metabolize 31 different carbon sources (Figure 3.7 ). Included in each of the 96 wells of the EcoPlate , along with the designated carbon source , is a tetrazolium dye. As carbon source utilization occurs , the released electrons wi ll reduce the tetrazolium dye, resulting in a purple color . The optical each well in the plate is then read using a microplate reader generating a metabolic fingerprint, or a pattern of carbon usage. This type of profiling can be us eful in distinguishing spatial and temporal changes i n microbial communities, noting that the technique is restricted to profiling a subset of the cultural microbial community. The technique has been used to monitor wastewater treatment plants, soils, and industrial waste sludge, along with many other sites being monitored for microbial community changes. This appears to be the first study to utilize the BIOLOG EcoPlates to monitor microbial communities associated with an iron fen system. The BIOLOG EcoPlat es provided potential metabolic properties to monitor using quantitative PCR. In this study, these properties were the degradation of xylose and cellobiose via glycoside hydrolase (GH) genes.

PAGE 32

23 Figure 3.7 Carbon sources present in each well of the BI OLOG EcoPlate. Xylose and cellobiose wells ar e highlighted. In the present study, EcoPlates were utilized to evaluate the effect of AMD on microbial carbon source usage by previously ARD impacted communities Quantification of Glycoside Hydrolase Genes Qua ntita tive PCR (qPCR) utilizes the linearity of DNA amplification to determine the amount of a known sequence within a sample. In this study, qPCR was utilized to determine the amount of three glycoside hydrolase genes involved in the degradation of cellobi ose and xylose based on the results of EcoPlate analysis. The present study measured DNA generation with the use of SYBR green I , a fluorescent reporter included in the PCR reaction which will fluoresce when bound to dsDNA PCR products . During each cycle of the PCR, DNA amplification is monitored. The reactions are characterized by the cycle at which the fluorescence first rises above the set threshold. If the starting material is abundant, amplification, and, in turn,

PAGE 33

24 fluore scence from the SYBR green I , is observed in earlier cycles. If the amount of initial DNA is low, then amplification and fluorescence will occur in later cycles. By using multiple dilutions of a known amount of standard DNA, a standard curve is generated of log concentration against qu antification cycle. The quantification cycle is the point at which fluorescence becomes measurable above the background (Figure 3.8) . The amount of fluorescence and the quantification cycle can then be correlated to the generated standard curve, and the am plified product can be quantified. Here, primers were utilized to quantify the abundance of three glycoside hydrolase (GH) genes involved in the degradation of cellobiose and xylose. Figure 3.8. Diagram of qPCR utilizing SYBR green I. When SYBR green is in solution, it emits low fluorescence, but when it binds to dsDNA, it will fluoresce and the qPCR machine will detect the fluorescence (Image adapted from www.sigmaaldrich.com)

PAGE 34

25 Studies which quantified GH6, GH48, and GH10 in environmental samples are e xtremely limited. To the best of our knowledge, this is the first stud y to quantify these genes in both AMD and ARD impacted sediments.

PAGE 35

26 CHAPTER IV CONDUCTED STUDY Introduction The impact of acid mine drainage (AMD) on previously metal exposed microbial communities is addressed in this study. The Chattanooga Fen is a wetland system that is both naturally metal impacted by endogenous acid rock drainage (ARD) and, in some areas, impacted by AMD from the nearby Gold Finch Mine. The fen is physically bisecte d into two distinct areas, each with ARD exposure , but only one of which has the addition al exposure to AMD. To explore how the long term release of AMD into the Fen has impacted the previously metal adapted bacterial communities, this study utilized high throughput sequencing to determine taxonomic structure, metabolic profiling methods to evaluate carbon so urce utilization, and quantitative PCR to quantify genes involved in carbon degradation. There have been few previous studies examining the bacterial communities within fens, and even fewer which have been conducted on acidic fens. The microbiological studies on acidic peatla nds have mainly been performed in Canada and the northern U.S., though these are not metal im pacted by two different sources like the Chattanooga Fen (Andersen et al, 2013; Lin et al, 2012). The studies which have been done relied heavily on culturing, a limited technique because only 1 to 10% of all bacteria are cultivable in a laboratory settin g. Modern, molecular methods such as high throughput sequencing are now being utilized to more thoroughly characterize a variety of environments, including wetlands and metal impacted systems. The presented study aimed to characterize the bacterial communi ties within a unique,

PAGE 36

27 previously metal adapted system, where portions of the system were additionally impacted by the presence of acid mine drainage. Methods Sample Collection In order to characterize the bacterial communities within the Chattanooga Fen a s well as their metabolic profiles, sediment samples were collected along an AMD gradient within the site that is also ARD impacted (from AMD unaffected ponds formed within the fen to AMD affected ponds and straight from the Gold Finch AMD effluent). This was accomplished in July and September of 2016 via th sediment core s . Duplicate cores were taken at each site, and one was placed on dry ice to preserve nucleic acids while the other was stored on ice to mai ntain cell viability (F igure 4.1 ). There were five AMD affected samples taken in both July and September (denoted as CFA and CFMine samples). Four AMD unaffected sediment samples were collected in July, and five AMD unaffected samples were collected in September (denoted as CFU samples) (Table 4.1 ). At the time of sampling, temperature, pH, conductivity, and dissolved oxygen were measured in the field from each sampling site within the Fen at the time of sample collection using a Thermo Scientific Orion 5 Stat Multiparameter Met er Kit.

PAGE 37

28 Table 4.1 . List of sediment samples collected from the Chattanooga Fen during July and September of 2016 . To measure water chemistry, 250 mL water samples co located with the sediment sam ples were also collected. W ater was filtered through a Nal gene 0.45 µm cellulose nitrate membrane filter (Thermo Scientific, Waltham, MA), acidified to pH <2.0

PAGE 38

29 with ultrapure 1:1 HNO 3 , and stored in a 4°C cold room in the dark for further analysis. Dissolved metals were quantified by the Shared Analytical Laborat ory, an analytical chemistry service laboratory on the University of C olorado Denver downtown campus, using a Thermo Jarrell Ash ICAP 61 inductively coupled plasma optical emission spectrometer (Thermo Jarrell Ash Corporation, Franklin, MA). Figure 4.1 . Photographs of sediment samples being collected via coring in unaffected and affected ponds. DNA Extraction, PCR Amplification, and Next Generation Sequencing The collected cores were cut using sterile PVC cutters, and the top 2 inches of sediment was placed in sterile tubes. Total genomic DNA was extracted from 10.0 ± 0.1 grams of mechanically homogenized sediment from each tube using the MO BIO PowerMax Soil DNA Isolation Kit (MO BIO Laboratories, Inc., Carlsbad, CA) following 16S rRNA genes from each sample were amplified in triplicate, pooled, and cleaned. The V4 hypervariable region of the 16S rRNA gene was

PAGE 39

30 high throughput sequencing, and the reverse primer containing a 12 nucleotide sample specific Golay barcode sequence (Caporaso et al. 2012) . The polymerase chain Gaithersburg, MD) (final reaction concentrations: 0.5 U Taq DNA polymerase, 22.5 mM KCl, 1.25 mM Mg 2+ mer, 200 nanograms (1X) bovine serum albumen (BSA) (New England BioLabs, Inc., Ipswich, MA), molecular biology grade water (Thermo Fisher Scientific, Inc., Waltham, MA) and minute de naturation step at 94°C, 30 cycles of a 45 second denaturation at 94°C, a 60 second annealing step at 50°C, and a 90 second extension step at 72°C, followed by a ten minute final extension step at 72°C. Successful amplification was verified via agarose ge l electrophoresis (1% agarose gel) and visualized with ethidium bromide to confirm that the PCR product size obtained coincided with expected product size. Successfully amplified reactions were pooled and purified using the ZYMO RESEARCH DNA Clean & Conce ntrator (ZYMO RESEARCH, Irvine, CA) , following the included protocol. DNA concentrations from cleaned reaction concentrates were quantified using the Qubit Broad Range dsDNA Assay Kit and Qubit 3.0 Fluorometer (Thermo Fisher Scientific, Inc., Waltham, MA) , following the included protocol. DNA from each quantified sample was pooled in equimolar ratios, concentrated, and re quantified us ing Qubit. The library of pooled samples, along with aliquots of the forward, reverse, and index sequencing primers, were sent to the University of Colorado Anschutz Medical Campus Genomics and Microarray Core Facility for Illumina high throughput sequencing.

PAGE 40

31 Computational Analyses Phylog enetic analyses were performed using the QI IME (Quantitative Insights Into Microbial Eco logy) software (Caporaso et al. 2010). Using the join_paired_ends.py command , paired end sequenced were joined , which uses the fastq join method and a specified minimum overlap score of 30 (Aronesty 2011). Joined sequences were filtered using the s plit_lib raries_fastq.py command, and sequences with a Phred quality score of 30 were kept, which corresponds to a sequencing error rate of 0.1%. Ope rational taxonomic units (OTUs) were picked using the pick_open_reference.py where reads were clustered together bas ed on 97% or greater sequence similarity. This command the August 2013 Greengenes bacterial and archaeal 16S rRNA database (DeSantis 2006). Chimeric sequences were identified using DECIPHER Find Chimeras Web tool and w ere removed from the dataset prior to further analysis (Wright 2012). Alpha diversity was performed in QIIME using the alpha_diversity.py command. PCoA plots for the sediment data were generated using the jackknifed_beta_diversity.py command in QIIME. BIOL OG EcoPlate Preparation and Reading BIOLOG EcoPlates (BIOLOG, Hayward, CA) were utilized to examine the metabolic profile diversity within the cultural portion of microbial communities from the Fen sediment samples. The EcoPlates are 96 well plates with 2 1 different carbon sources, each repeated three times on the plate with three negative control wells. The core sediment samples which had been placed on ice for viability were immediately cut using autoclaved PVC cutters 1 inch from the top of the sediment so the sediment could be reached with sterile tools . An autoclaved spatula was then used to remove the top 2

PAGE 41

32 inches of sediment from each of the viability cores, and the sediment was placed into sterile tubes. From these tubes, 10 grams of sediment was mi xed with 90 mL of 0.145 M sodium chloride solution made with sterile deionized water in a 250 mL flask. The flasks were then placed on a shaker for at 200 RPM for 20 minutes at 20°C. Once the shaking had completed, the flasks were put into a 4°C refrigerat or for 30 minutes to allow the sediment particles to settle. Using a multichannel pipette, 150 µL of liquid from each flask was placed into wells of the EcoPlates. Once the entire plate was inoculated, the plate was covered with sealing tape, the lid was r eplaced, and the plates were stored in a 25°C incubator. Every 24 hours post inoculation, a BioTek Gen5 microplate reader was used to measure the absorbance (590nm) of each well in the 96 well plate. Average well color development was calculated for each triplicate in each plate and then averaged. Average well color development (AWCD) was calculated using the following equation: R)/31 DNA Extraction and Amplification from EcoPlates Xylose and cellobiose were identified as substrates (carbon sources) of interest. DNA was then extracted from each of the BIOLOG EcoPlates from both the xylose and cellobiose wells. To accom plish this, 50 µL of culture from each of the three xylose wells

PAGE 42

33 was placed into an extraction tube from the MO BIO PowerMax Soil DNA Isolation Kit (MO BIO Laboratories, Inc., Carlsbad, CA) along with 1 µL of molecu lar grade water in order to increase the volume to the required amount. The manufacturers protocol was then followed, and library preparation was completed in the same manner as t he t otal 16S rDNA libraries discussed above. The process was repeated for the cellobiose wells of each EcoPlate. The resulting libraries were sent to the Anschutz Medical Campus Genomics and Microarray Core Facility for Illumina high throughput sequencing. Quantitative PCR Utilizing the StepOnePlus Real Time PCR system, the abunda nce of three glycoside hydrolase genes (GH10, GH6, GH48) were quantified. The primers used were CTACGACTGGGAYGTNIBSAAYGA GTGACTCTGGAWRCCIABNCCRT et al ACCTGCCCRCCGYGACT GAGSGARTCSGGCTCRAT G; Y = C or T; S = G or C for the GH6 gene (Merlin et al GCCADGHTBGGCGACTACCT CGCCCCABGMSWWGTACCA the GH48 gene (de Menezes et al ., 2015). qPCR condition s and reactions were carried out as previously described for GH10 (Wang et al ., 2010), for GH6 (Merlin et al ., 2014), and for GH48 gene (de Menezes et al ., 2015). The reactions were performed in a 25 µL mixture using 12.5 µL of PowerUp SYBR Green Master Mi x, 2 ng of extracted DNA diluted to 1 ng/µL of molecular grade water, 0.5 µL of each primer, and 9.5 µL of molecular grade water. Standard curves for each gene were generated using synthesized gene fragments from Integrated DNA Technologies (Coralville, I A) . The stan dard gene copies for the

PAGE 43

34 for GH10. For all samples and standards, triplicate reactions were carried out. In all triplicate measurements, the standard deviation was less th an 10%. To check for specificity of amplification, melt curves were generated for each SYBR assay. The resulting qPCR assays were also run on through gel electrophoresis to confirm amplification of the expected gene. PCR efficiencies ranged from 72.8 80. 9% for GH10 reactions, 73.7 82.7% for GH6 reactions, and 71.6 88.3% for GH48 reactions. The correlation coefficients (R²) was greater than 0.99 for all assays. Results and Discussion Site Chemistry The chemistry of the water from the AMD affected sit es were distinct from the AMD unaffected sites in terms of pH, dissolved oxygen (DO), and conductivity (Table 4 .2). AMD affected samples had an average pH of 5.36, while unaffected samples had an t test p value=0.009). The mos t basic pH values came directly from the effluent from the Gold Finch Mine, the source of AMD impact to portions of the Fen. This may be due to the fact that the mine drainage flows through a constructed concrete portal which sits below the Gold Finch Mine , which raises the pH of the outflow or because the water flows through mineral ore material with a higher carbonate content . The ground water which feeds the Fen is naturally acidic (which in turn dissolves the parent rock material and forms the ARD withi n the Fen), and the unaffected sites do not receive the flow of raised pH water from the mine.

PAGE 44

35 greater in unaffected sites compared to the affected samples t test p value= 0.03). Dissolved oxygen often relates to water quality, and it is essential to support aquati c life. This includes bacteria and other microorganisms which depend on DO to decompose organic material at the bottom of a body of water, such as the ponds at the Chattanooga Fen, and, in turn, contribute to elemental cycling. The highest DO levels came f rom the unaffected ponds in September which may be due to the fact that oxygen dissolves easier in cooler temperatures than warm temperatures. When the September samples were collected, there was a steady snowfall, and the temperatures were significantly l ower than when the July samples were collected. The AMD affected samples had significantly higher conductivity values compared to the unaffected t test p value=0.003) (Table 4.2) . conduct electricity) was significantly higher in AMD affected sites compared to AMD unaffected sites (Table 4.2) . Conductivity can change depending on temperature, which would explain why the July samples had greater conductivity values compared to the Sep tember samples because warmer temperatures tend to increase conductivity due to the increased dissolution of metals . Changes in conductivity can also be related to the presence of pollutants, such as AMD, which affects organisms within the site because aquatic life can tolerate certain conductivity ranges. Pollutants such as AMD can increase the concentratio n of ions, thus increasing the conductivity of the water. The geology of an environment and the material through which the ground water flows through also influence s the conductivity of a solution. Therefore, the bedrock which the groundwater

PAGE 45

36 flows through to feed the mine outflow as well as the affected sites may flow through materials which better ionize when washed into water compared to the unaffected sites. The fact that the bedrock which the water flows through both from the mine and the groundwater differs may also explain the differences in metal concentrations between affected and unaffected sites since affected sites had a greater amount of calcium, iron, magnesium, manganese, and zinc (Table 4.3 ). After analysis of dissolved metal concentrations, it was found that the acid min e drainage sample along with the AMD affected samples had elevated concentrations of calcium, iron, magnesium, manganese, and zinc compared to the AMD unaffected samples (student t test p values<0.05) (Table 4.3). Each of these metals have been shown in previous studies to be elevated in AMD impacted sites, though these studies were not on metal impacted fen systems, such as the Chattanooga Fen (Chen et al., 2016). The unaffected samples generally had a higher concentration of aluminum compared to the affected sites. However , the CFA1 sample taken from September was an outlier for the affected samples, and had a concentration of aluminum closer to the unaffected samples. If the outlier is removed, the concentration of aluminum within the unaffected samples is greater compared to AMD t test p value= 0.004). It should be n oted that the CFMine samples were included with the AMD affected samples in these analyses. However, to determine whether the Gold Finch Mine samples were skewing the analyses, the mine samples were removed from analysis. Results still showed similar differences between AMD affected and AMD unaffected t test p values <0.05).

PAGE 46

37 Table 4.2 . Chemistry p arameters measured in the field for each sampling site

PAGE 47

38 Table 4.3 . Dissolved metal concentrations from water samples collected above sediment samples taken from the Chattanooga Fen in July and September of 2016. Conc entrations are given in parts per billion (ppb). The concentrations of calcium, iron, magnesium, manganese, and zinc are greater in AMD affected sites

PAGE 48

39 Taxonomic S tructure of Bacterial Communities within the Chattanooga Fen The taxonomic composition of the AMD affected sites was also unique from the AMD unaffected sites within the Chattanooga Fen. However, QIIME analyses revea l ed seven phyla to be common and most ab undant within the Chattanooga Fen. These seven phyla included Acidobacteria, Actinobacteria, Bacteroidetes, Chloroflexi, Cyanobacteria, Proteobacteria, and Verrucomicrobia. The sediment samples which came directly from the outflow of AMD was dominated by P roteobacteria (an average of 77.05% relative abundance). In the affected samples, the relative abundance percentage of Bacteroi detes and Actinobacteria were higher compared to the unaffected samples t test p values of 0.04 and 0.0003, respective ly). The unaffected samples had great er relative abundance percentage of Acidobacteria and Verrucomicrobia compared test p values of 0.0001 and 0.03, respect ively) (Table 4.4 and Figure 4.2 ). Table 4.4 . Top seven most re latively abundan t phyla shown to dominate within the sediment samples collected from the Chattanooga Fen. The average percent relative abundance is shown for each phyla from mine, affected, and unaffected samples, the standard deviati ons are included in parenthese s

PAGE 49

40 While the phyla P roteobacteria was common to the Gold Finch Mine AMD , to the AMD affected, and to the AMD unaffected sediment samples, the c lasses which encompass the phylum differed . Alphaproteobacteria was the dominant class in the unaffected samples (an average of 11.56% relative abundance compared to 6.15 % , test p value of 0.01). Deltaproteobacteria had an average relative abundance percentage of 14.15% in the AMD affected and Gold Finch Mine samples compared to an average relative abunda nce of 6.5 % test p value of 0.005). Betaproteobacteria also comprised a large amount of the AMD affected and Gold Finch Mine samples (average relative abundance of 21.29% compared to 8.32% in the unaffected samples). The dominant classes of Proteobacteria within the AMD affected samples from the Chattanooga Fen were Betaproteobacteria and Deltaproteobacteria, which are both compromised of Gram negative bacteria. Though metabolically diverse, many of the organisms with in the Betaproteobacteria class, including Bur k holderiales and Methylophiales which were found in all affected samples, obtain their energy from inorganic compounds. However, another order common to all affected samples was the order Nitrosomonadales which includes organisms that perform nitrification and are important in the global nitrogen cycle ( Mendez Garccia et al, 2015) . Many of the Deltaproteobacteria which were found more frequently in the AMD affected sites compared to the unaffected sites can survive in unfavorable environmen ts (such as AMD). One of the orders found in all of the affected sites was Myxococcales. These gliding organisms will form fruiting bodies in order to survive until nutrients are more plentiful. There were also several orders found in all of the affected s amples that

PAGE 50

41 are sulfate reducing bacteria, including D esulfobacterales and Desulfovib rionales. Considering the sulfate rich environments these organisms live in, there presence is to be expected. After analysis, it was determined that in the Gold Finch M ine samples, bacteria of the genus Galionella of the class Betaproteobacteria were dominant (representing 50.13% of the organisms from the July mine sample and 55.41% of the organisms from the September mine sample) . Gallionella species are iron oxidizing , chemoli thotrophic bacteria that have been found in a variety of different aquatic habitats. These bacteria play an important part in the oxidation of iron (Jones et al, 2015) . The next most abundant organisms in the Gold Finch Mine samples came from the Bacteriovoracaceae f amily ( 8.51% of the organisms in the July sample and 13.84% of the organisms in the September mine sample) (Figure 4.3 ) . Within this family are extremely unique, Gram negative organisms. These bacteria are commonly found in enteric environments as well as river water. They are known as bacterial predators because they can attach to the outer membrane of Gram negative bacteria, enter the periplasmic space, and use hydrolytic enzymes to feed on the host cell biopolymers (Chen et al, 2012) . In AMD unaffected samples within the C hattanooga Fen, the phylum Acidobacte ria was found in greater relative abundance (an average of 22.54% compared to 6.55% in AMD affected samples and 3.23% in Gold Finch Mine Samples) than the AMD affected sites, particularly the orders Acidobacteriales and Soilbacterales. Both of these contain acidophilic bacteria naturally found within soil and sediment. However, even though Acidobacteria are widespread within terrestrial ecosystems,

PAGE 51

42 they remain largely uncharacterized and uncultured. The only cultured Aci dobacteria are heterotrophic organisms which are aerobic or facultatively anaerobic. A common group of organisms found in the sediment coming from affected and unaffected samples, but only found in less than 0.2 % abundance in the sediment samples which ca me directly from the acid mine drainage, was the order Stramenopil es which are of the phylum Cyan obact eria (an average of 17.9% ± 13.1% and 19.79% ± 9.98%, r elative abundance , respectively) (Figure 4.3 ). Cyanobacteria are photosynthetic bacteria which gene rally do not live in flowing waters, which may explain why they were not found in the mine effluent sediments.

PAGE 52

43 Figure 4.2 . Relative abundance of the bacteria within each collected sediment sample (represented at the phylum level)

PAGE 53

44 Figure 4.3 . Heatma p of the top 30 OTUs from each sediment sample . All of the included OTUs had a minimum of 30,000 reads

PAGE 54

45 Alpha diversity, which measures the diversity within individual samples, was and Observed OTUs metrices. This was performed for affected and unaffected samples as well as within the two time points (July and September sampling). The Shannon Index measured species diversity, while Chao1 and Observed OTU richness measure the species relationships with species richness. There was no determined correlation between AMD affected and AMD unaffected samples overall nor between AMD affected and AMD unaffected July versus Se ptember samples, in terms of any of the alpha diversity m easurements performed (Table 4.5 ). The lack of correlation implies that AMD did not have an impact on alpha diversity among the samples. Recall that these samples are endogenously metal exposed, and, as such, have presumably already adapted to the presence of potentially toxic metals. Unlike other studies that have found AMD negatively impacts alpha diversity, this was not observed here (Chen et al., 2016).

PAGE 55

46 Table 4.5 . Alpha diversi ty metrics for all samples of the same material g rouping where Chao1 estimate, Observed OTU richne ty, and Shannon Index were performed and a t test was done to test if the mean value for each group was significantly different. The data shown is the average of all samples b elonging to each group with the standard deviation in parentheses Shannon Index Obser ved OTUs Affected Samples 1 401.77( 582.17) 7.63(1.35) 152.74(48.99) 1818.09(605.24) Unaffected Samples 1 560.32 ( 559.34) 7.65(0.667) 126.67(27.95) 1533.68(411.72) July Samples 1 921.523( 531.27) 7.64(1.09) 153.53(42.33) 1877.33(538.32) September Samples 1 176.69( 214.34) 7.63(1.07) 128.56(38.96) 1508.81(481.72) t test p value Affected/Unaffected 0.285 0.964 0.201 0.278 t test p value July/September Samples 0.3 18 0.988 0.223 0.155 To evaluate the impact of anthropogenic metal impact on the taxonomic composition of the collected sediments from the Chattanooga Fen, affected and unaffected samples w ere compared. Princip al coordinate analysis (PCoA) plots of weighted and unweighted UniFrac values were generated. PCoA plots are a form of multidimensional scaling which aid in the visualization of similarities between individuals in a dataset. In this case, the PCoA plots created utilized UniFrac , which is a distance metric used for comparing biological communities. The weighted variant of UniFrac is quantitative and takes into account the relative abundance of observed organisms, while the unweighted UniFrac variant only takes into account presence or absence of the organisms. Using the weighted UniFrac values, Axis 1 explained 42.35 % of th e variation arra yed along this axis, while Axis 2 explained 20.32% of the observed variation . Therefore, in total between bot h the first and secon d principal coordinates,

PAGE 56

47 62.67% of the variance was explained, and the affected and unaffected groups cluster together (Figure 4.4 ). Using the unweighted UniFrac values, Axis 1 explained 17.69%, while Axis 2 explained 15.07%. Therefore, when using the unwe ighted UniFrac variant, which does not take relative abundance into account, 32.76% of the va riance was explained (Figure 4. 4 ). ANOSIM statistical testing was then used to compare the variation in species composition and abundance between AMD affected and AMD unaffected samples , also known as beta diversity . The resulting R statistic s from both weighted and unweighted UniFrac values were positive nu mbers, suggesting there was similarity in bacterial communities between the samples within the given groupings ( AMD affected and AMD unaffected) (Table 4.6 ). ANOSIM statistical testing was used to determine if there was a temporal influence on the bacterial community variation in the samples. The R statistics for both weighted and unweighted UniFrac variants wer e extremely close to 0 (R=0.0126 and 0.0186 ), imply ing that there wa s not a difference in community structure between July and September samples (Table 4.7 ). Alpha diversity metrices showed no correlation between condition ( AMD affected or unaffected) o r month of sample collection and diversity within samples (Table 4.5). However, the constructed PCoA plots based off of weighted UniFrac variants showed that the first two axis explained 62.67% of the variance in taxonomic composition between AMD affected and AMD unaffected sites (Figure 4.4). This could be due to the fact that the AMD flows from the mine to the affected samples, and organisms are brought into the affected sites along with organisms that were already present in the endogenously metal impact ed Fen. There is one AMD affected sample

PAGE 57

48 which sits closer to the AMD unaffected samples then the AMD affected samples, which is CFA1 . This site was considered an AMD affected site because of its proximity to the Gold Finch Mine and location on the hillsid e, but the fact that it groups more closely with the AMD unaffected samples warrants further investigation. Differences in chemistry (pH, DO, conductivity) as well as dissolved metal concentrations (AMD affected samples had elevated levels of calcium, iro n, magnesium, manganese, and zinc compared to AMD unaffected samples) were noted earlier. These differences justified the groupings of AMD affected and AMD unaffected sites, and these differences were also observed in the bacterial community analyses. This result suggests that AMD impacts the abundance of bacterial communities pr esent within the ARD sediments of the Chattanooga Fen.

PAGE 58

49 Figure 4.4 . Principle Component Analysis of weighted (left) and unweighted (right) UniFrac distances for AMD affected and unaffected samples. The more similar the bacterial communities are in terms of taxonomy, the closer the samples will cluster within the planes of the plot. The AMD affected samples are in closer proximity to each other than to the AMD una ffected samples. The affected sample which is the closest to the unaffected samples in both the weighted and unweighted graph is CFA1 which is the site that sits closest to the unaffected sampling sites

PAGE 59

50 Table 4.6 . Resulting ANOSIM values when comparing affected and unaffected samples using both weighted and unwe ighted UniFrac values in order to determine the effect of AMD on the bacterial com munity structures and abundance . The R stat and p values show that the samples within AMD affected samples are more similar to eachother than to the bacterial communities wit hin the AMD unaffected samples Table 4.7 . Resulting ANOSIM values wh en comparing July and September samples using both weighted and unweighted UniFrac values in order to determine the effect of month of the bacterial community structures and abundance . The R stat and p values show that the samples from July and September are all similar to each other. This implies that time , within the context of the two sampling periods, did not influence the bacterial community structure

PAGE 60

51 Carbon Source Usage Follo wing incubation of the AMD affected and unaffected sample s in the EcoPlates, th e AMD affected samples utilized all six of the different carboxylic acids present in the plates (Table 4.8 ). However, each of the AMD affected samples had a lower usage of carbo n sources compared to the unaffected samples using an average of 64.5% of the total carbon sources available (standard deviation of 9.6%) . The AMD unaffected samples utilized all of the carbon sources to some extent. Carbo hydrates had the largest OD readin gs by the organisms within the AMD unaffected sediments , which correlates to heavy usage of the carbon sources by the microorganisms within the wells . Xylose and cellobiose , in particular, had the highest OD readings among all of the sample types (an avera ge OD reading of 2.36 for AMD affected samples and 2.42 for AMD unaffected samples) , and so were selected for further analysis.

PAGE 61

52 Table 4. 8. EcoPlate analysis of microbial carbon substrate utilization . Xylose and cellobiose are highlighted because they we re the only substrates used to the maximum in all sediment samples OD < 0.5:

PAGE 62

53 Average well color development (AWCD) was used to correlate the darker purple color (greater usage of the carbon source within the well) with the amount of carbon sou rces being used by the organisms within each sample (Figure 4.5) . AWCD data showed that the AMD unaffected samples from both July and September utilized more of the included carbon sources within the EcoPlates to a greater degree (average OD readings of 1. 97) compared to the microorganisms within the AMD affected sediment samples ( average OD reading of 1.21; t test p value=0.01). The results of the EcoPlate analyses revealed that the AMD affected and AMD unaffected samples from the Chattanooga Fen containe d organisms which differed from each other in not only the carbon substrates utilized but in AWCD as well, signifying a difference in amount of the carbon sour ces used (Table 4.8 and Figure 4.5 ). The affected samples showed greater substrate usage within t he carboxylic acids while the unaffected samples showed greater carbohydrate usage. The observed results were similar to a recent EcoPlate to determine the e cotoxicological impact of metal exposed dairy sewage sludge on the microbial communities (Gryta et al., 2014) . Through their analyses, i t was shown that metal exposure resulted in deterioration of overall microbial activity and diversity of substrate utilization. In non metal exposed samples, all substrates were utilized compared to only two substrates being utilized to the maximum potential by the metal exposed sludge communities. This may have been related to a reduction in catabolic functions. It is important to note that the study above examined metal impact on previously unexposed communities.

PAGE 63

54 In the present study, however, there were two common plant sugars utilized to the maximum degree (>2.5 OD) by all sediment samples; xylose and cellobiose. The fact that both were utilized by all samples led to the following objective of the project which was to determine the abundances of three glycoside hydrolase genes which are involved in the degradation of both types of plant sugars. Cellobiose is a disaccharide which can be obtained from the enzymatic hydrolysis of cellulose, the main structural componen t of plant cell walls (35 50% of a control the rate of cellulose degradation are not well understood. Enzymes which degrade cellulose are described as glycoside hydrol ases (GH) in at least ten G H families (GH 1, 3, 5, 6, 8, 9, 12, 44, 45, and 48). Within these GH families, there are three types of 1,4 glycosidic bonds. There are endocellulases that are active on internal bonds, exocellulases that degrade the polymer from its extremities, glucosidases that produce glucose from cellobiose (Merlin et al, 2014 ) . There are three GH families which contain enzymes with only endoglucanase and/or exoglucanase activities; GH6, GH45, and GH48). Endoglucanase and exoglucanase are involved in th e early steps of cellulose degradation which is when the polymer is broken down to cellobioses (made up of two glucose molecules). Microorganisms a carbon and energy source for growth and are widely distributed within the environment ( de Menezes et al ., 2015 ) .

PAGE 64

55 Xylan is the major component of hemicellulose, the second most abundant polysaccharide in plant cell walls behind cellulose. A large variety of cooperatively acti ng enzymes are required to completely hydrolyze the linear polysaccharide xylan into xylose. These are essential enzymes for microorganisms for their ability to degrade plant material into usable nutrients. The enzymes include endo 1,4 D D x D glucuronidase acetyl xylanesterase, and several others. Xylanases within GH10 are the most abundant in the environment, substrate specific to xylan, more common in bacteria than fungi, and are larger than the other GH xylanases (e.g. GH 5, 7 , 8, 11). Xylanases are also studied because of their potential applications in agriculture and industry. For example, they have been used in the papermaking process for bleaching of wood pulp and for food additives for poultry. In the future, they could also be used in the generation of biofuel from unusable plant material (Wang et al, 2010) .

PAGE 65

56 Figure 4.5 . Average Well Color Development (AWCD) of the metabolized substrates within the BIOLOG EcoPlates for affected and unaffected samples at specified incubation times (n=5 for AFFECTED July, AFFECTED September, and UNAFFECTED September; n=4 for UNAFFECTED July).

PAGE 66

57 In order to identify the organisms utilizing the xylose and cellobiose , DNA was extracted and sequenced from the xylose and cellobiose wells of the EcoPlat es. There were six different OTUs that were the most common from both the xylose and cellobiose wells as well as between affected and unaffected samples. These OTUs included Oxalobacteraceae, Enterobacteriaceae, Gammaproteobacteria, Pseudomonas, Paenibacillus, and Cellulomo nadaceae . The d iversity within AMD affected versus AM D unaffected samples, xylose versus cellobiose, and July versus September samples were not different. This reflects that the heterotrophic, culturable organisms within all of these samples were similar w hether AMD affected or unaffected (Supplemental Figures 1 4, Supplemental Table 1). This suggests little impact of AMD on the culturable organisms utilizing xylose and cellobiose from the Chattanooga Fen. Quantitative PCR Common sugars utilized in all of the BIOLOG EcoPlates, both affected and unaffected, were the plant sugars cellobiose and xylose. Enzymes which are responsible for the degradation of both sugars are members of the super family of glycoside hydrolases. Because of this, this study aimed to determine the amount of three glycoside hydrolase genes present in affected and unaffected sediment samples collected from the Chattanooga Fen. Glycoside hydrolases are a large and complex grou p of enzymes and the fact that, in many GH families, there is a presence of multiple substrate specificities, the design of molecular tools (e.g. qPCR primers) for an in depth investigation into their environmental role has been challenging. However, for GH6 and GH48, all of the

PAGE 67

58 functionally characterized enzymes ha ve been shown to target the degradation of cellulose. A majority of the enzymes function as cellobi ohydrolases and endoglucanases, major components of multie nzyme cellulolytic systems often act ing in synergy with other cellulases to depolymerize cellulose. Glycoside Hydrolase 6 and Glycoside Hydrolase 48 Quantitative PCR for the glycoside hydrolase 6 (GH6) gene was found in higher abundance in AMD unaffected samples compared to affected with an average of 3 . 28 x 10 copies per gram of sediment (standard deviat ion of 264129.4) and 1 .10 x 10 copies per gram of sediment (standard deviation of 11566.3), respectively (Fi gure 4.6 A ). The Gold Finch Mine sample collec ted during September had values below the detectable limit, and, therefore, was not included in these analy ses. When a zero is used for the September mine sample, the average copy per gram of sediment of GH6 drops to 9816.5. The two groups were statistically different from each other, both when the zero is not included and when it is included (t test p value= 0.002 without the zero included and 0.001 when the zero is included). There was one outlier in the unaffected group (914078 copies/gram of sediment), which ca me from CFU1 July. The July Gold Finch Mine sample had the lowest amount of GH6 copies per gram of sediment (806 copies/gram of sediment) of the samples that amplified, which implies similarity to the AMD affected samples. Interestingly, the GH6 primers were previously utilized to quantify the amount of GH6 cellulase gene present in different soil s , th e peatland environment with a pH of 6.43 averaged 1.89E+08 copies per gram of soil (Merlin et al., 2014) . This study found that soil acidity decreased the abundance of the GH6 gene, implying acid toxicity. The

PAGE 68

59 abundance of GH6 in the present study implies acid toxicity in both the AMD affected and unaffected samples. The results of the glycoside hydrolase 48 (GH48) gene abundance were similar to those of the GH6 gene . The affected samples had a significantly lower amount of the GH48 gene compared to the un affected samples ( t test p value= 0.0006) (Figure 4.6B ) . The affected samples had an average of 263928 copies per gram of sediment (standard deviation of 202486) compared to the unaffected sample which had an average of 2173712 copies per gram of sediment (standard deviation of 1249504). The mine samples from both July and September had amounts of GH48 that were below the detectable limit, therefore, they were not included in the analysis. The largest amount of GH48 came from CFUA July with 37427058 copies per gram of sediment. According to quantitative PCR analysis, both GH6 and GH48 genes were found in higher amounts in the AMD unaffected samples. This implies an AMD impact on the types of genes being used for the degradation of cellulose or the presence of genes available for cellulose degradation, despite the identification of similar degrading bacterial species found in the EcoPlates (Supplemental Figures 3 and 4) . While a study addressing the impact of AMD on gene abundance in metal impacted wetland systems could not be found, a s tudy looking at the effect of AMD exposure on soil microbial communities found a corresponding decrease in the abundance of the GH6 gene in acidic soils. Interestingly, the levels of GH6 found in this study in both the AMD affected and unaffected samples w ere lower than those found in the soil study implying not only a pH effect but also a potential metal effect.

PAGE 69

60 Glycoside Hydrolase 10 When qPCR was performed for the GH10 gene, the results were again similar to the results for GH6 and GH48. The unaffecte d samples had a significantly larger amount of the GH10 gene per gram of sediment compared to the affected samples ( t test p value= 0.0002) (Figure 4.6C ). The average copy per gram of sediment for unaffected samples was 5853261 (standard deviation= 196301 4). Both the mine samples were again below the detectable limit; therefore, those samples were not included in the analyses. The affected samples had an average of 1393280 copies of GH10 per gram of sediment (standard deviation= 1465770). The lowest amount of GH10 within affected samples when the mine samples are not included was from CFAA July (387261 copies/gram of sediment). The greatest amount of GH10 copies came from CFU1 July (8586215 copies/gram of sediment). A recent study which examined the amoun t of GH10 in alpine tundra soil in the Tianshan Mountains of China determined that the main xylanotic bacteria to produce xylanases in this environment were Actinobacteria, Proteobacteria, Verrucomicrobia, and Firmicutes , all of which were present in the c urrent study. The GH10 copy numbers have not been previously studied in fen environments, or acidic f en environments, but there have been several xylanases from GH families 10 and 11 isolated from acidophilic microorganisms (Collins et al., 2005). All of the glycoside hydrolase gene abundances which were measured in this study (GH6, GH48, and GH10) were di fferent between AMD affected and AMD unaffected samples. The decreased abundance of these genes in the AMD affected sites

PAGE 70

61 could result in the higher ob served peat material in the AMD affected versus unaffected sites, or, perhaps, the organisms which survive in the AMD impact have different sequences for these genes then what is generally observed, and the primers utilized in this study were unable to amp lify these GH genes within those samples. The results of this aspect of the conducted study show that AMD may impact the degradation of carbon, which is essential monitor in this type of ecosystem because peatlands are known to be large carbon sinks.

PAGE 71

62 Figure 4.6 . qPCR results from GH6(A), GH48(B), and GH10(C) amplification. For all amplified GH genes, the AMD unaffected samples had a greater abundance of the gene present compared to the AMD affected samples. For GH6 qPCR, the September CFMine sample di d not amplify, and was not included in analysis (AMD affected and AMD unaffected n=9). For GH48 and GH10, the mine samples did not amplify, and were not included in the analysis (AMD affected n= 8, AMD unaffected, n=9)

PAGE 72

63 CHAPTER V CONCLUSIONS AND FUTURE DI RECTIONS Fens are essential carbon sinks which remain largely understudied. The Chattanooga Fen is a rare, acidic iron fen which has yet to be microbiologically characterized. This unique ecosystem provided the prime opportunity to explore the taxonomic s tructure and metabolic profiles of bacterial communities which have previous, long term metal exposure (ARD) , and how some of these communities have adapted to the presence of AMD flowing into portions of the Fen from the Gold Finch Mine. This study foun d that the bacterial communities which live in the Chattanooga Fen are as unique as the ecosystem in which they survive and distinct from each other depending on their location within the Fen. The bacterial communities that live directly in the acidic, min e effluent that impacts a portion of the Fen include Gallionella and Bacteriovoracacea which are unique in that they have adapted to utilize nutrients available to them in the ir environment (i.e. iron and other Gram negative organisms). The AMD flowing int o lower elevation ponds within the Fen is also contributing other Beta and Deltaproteobacteria to the already metal adapted bacterial communities. The ponds unaffected by AMD, with the lowest pH include Acidobacteria whi ch thrive in acidic environments . Not only is the taxonomy distinct between AMD affected and AMD unaffected bacterial communities within the Chattanooga Fen, but their metabolic profiles are also

PAGE 73

64 unique. The heterotrophic microorganisms within AMD unaffected ponds utilized all 31 of the p rovided carbon sources to a greater degree within the BIOLOG EcoPlates compared to the affected communities. However, there were two common sugars utilized (xylose and cellobiose) in all plates inoculated with samples taken from the Fen. This led to the fi nal portion of the conducted study which was to determine the abundance of three glycoside hydrolase genes (GH6, GH48, and GH10) that aid in the degradation of xylose and cellobiose. The AMD unaffected samples contained a significantly greater amount of al l three glycoside hydrolase genes compared to the AMD affected samples. These findings are a possible explanation for the increased build up of peat within AMD affected sites compared to AMD unaffected sites at the Fen. The impact of AMD on the previousl y metal adapted bacterial communities within the Chattanooga Fen was reflected in bacterial community taxonomy, carbon source usage, and abundance of glycoside hydrolase genes involved in the degradation of xylose and cellobiose. The results of this study were similar to previous studies examining the impact of AMD on non ARD impacted bacterial communities in terms of organisms present (i.e. Deltaproteobacteria in AMD sediments) and ARD impacted communities (i.e. Acidobacteria within the ARD impacted AMD un affected sediments of the Fen). AMD impact was also observed in carbon source usage and potential carbon source degradation (glycoside hydrolase gene abundance). This study showed that AMD negatively affects the usage of organic carbons as well as its pote ntial degradation. This information may influence future management decisio ns of the Fe n because monitoring the degeneration of carbon is essential in this large carbon sink.

PAGE 74

65 Though the results of this study provided much needed insight into the microbio logy of the Chattanooga Fen, there are still several future studies which could be conducted in examining this unique system. Additional aspects of the microbiology of the Chattanooga Fen that remain to be addressed include examining the potential impact o f AMD on other groups of microorganisms, s uch as the fungi as major players in organic carbon decomposition, and conducting a more detailed analysis of gene expression of the Fen microbial communities through the use of transcriptomics. Comparative studi es amongst iron fen could provide valuable information regarding the consistency of AMD impacts, as well as ecosystem level implications of fen deterioration due to their importance as carbon sinks.

PAGE 75

66 SUPPLEMENTAL INFORMATION Supplemental Figure 1. Top twelve OTUs from 16s rDNA amplified out of the xylose wells of the EcoPlate

PAGE 76

67 Supplemental Figure 2 . Top twelve OTUs from 16s rDNA amplified out of the cellobiose wells of the EcoPlate

PAGE 77

68 Supplemental Figure 3 . Top 15 OTUs from the 16S rDNA extracted fro m the xylose and cellobiose wells within EcoPlates inoculated with affected samples. The samples which have an X were extracted from xylose wells and the samples which include a C were extracted from cellobiose wells. The samples which were collected in Ju ly are denoted with a J at the end, and the samples which were collected in September are denoted by an S at the end

PAGE 78

69 Supplemental Figure 4. Top 15 OTUs from the 16S rDNA extracted from the xylose and cellobiose wells within EcoPlates inoculated wit h un affected samples. The samples which have an X were extracted from xylose wells and the samples which include a C were extracted from cellobiose wells. The samples which were collected in July are denoted with a J at the end, and the samples which were co llected in September are denoted by an S at the end

PAGE 79

70 Supplemental Table 1. Alpha diversity metrics for all samples of the same material grouping where Chao1 estimate, Observed OTU richnes y, and Shannon Index were performe d and a t test was done to test if the mean value for each group was significantly different. The data shown is the average of all samples belonging to each group with the s tandard deviation in parentheses

PAGE 80

71 REFERENCES Andersen R, Chapman S, Artz R . 2013. Microbial communities in natural and disturbed peatlands: A review. Soil Biology and Biochemistry 57 :979 994. Andersen R, Grasset M, Thormann L, Rochefort AJ, Francez. 2010. Changes in microbial structure and function following Sphagnum peatland restorati on . Soil Biology & Biochemistry. 42: 291 301 . Andersson S, Dalhammar G, Kuttuva Rajarao G. 2011. Influence of microbial interactions and EPS/polysaccharide composition on nutrient removal activity in biofilms formed by strains found in wastewater treatment systems. Microbiol. Res. 166:449 57. Artz RRE, Baird AJ, Belya LR, Comas X, Reeve AS, Slater LD. 2009. Microbial community structure and carbon substrate use in Northern peatlands . Geophysical Monograph Series. 184. Baker BJ, Banfield JF. 2003. Microbial communities in acid mine drainage. FEMS Microbiol. Ecol. 44:139 52 Bánfalvi G. 2011. Heavy Metals, Trace Elements and Their Cellular Effects. Cellular Effects of Heavy Metals 3 28. Bernhardt ES, Lutz BD, King RS, Fay JP, Carter CE, Helton AM. 2012. How many mountains can we mine? Assessing the regional degradation of Central Appalachian rivers by surface coal mining. Environ Sci Technol 46 : 8115 8122. Barry, K. L., J. A. Grout, C. D. Levings, B. H. Nidle, and G. E. Piercey. 2000. "Impacts of acid mine dr ainage on juvenile salmonids in an estuary near Britannia Beach in Howe Sound British Columbia." Canadian Journal of Fisheries and Aquatic Sciences 57(10): 2031 2043. Braud A, Hoegy F, Jezequel K, Lebeau T, Schalk IJ. 2009. New insights into the metal spec ificity of the Pseudomonas aeruginosa pyoverdine iron uptake pathway. Environ. Microbiol. 11:1079 91. Bruins MR, Kapil S, Oehme FW. 2000. Microbial resistance to metals in the environment. Ecotoxicol. Environ. Saf. 45:198 207. Chen, H., Young, S., Be rhane, T. K., & Williams, H. N. (2012). Predatory bacteriovorax communities ordered by various prey species. PLoS ONE , 7 (3). Chen, L. xing, Huang, L. nan, Méndez García, C., Kuang, J. liang, Hua, Z. shuang, Liu, J., & Shu, W. sheng. (2016). Microbial communities, processes and functions in acid mine drainage ecosystems. Current Opinion in Biotechnology .

PAGE 81

72 Chimner RA, Cooper DJ. 2006. A Report on the Restorati on Potential of Chattanooga Distribution, Condition, and Restoration Needs, San Juan Mountains. David CP. 2003. Heavy metal concentrations in growth ba nds of corals: A record of mine tailings input through time (Marinduque, Philipines). Marine Pollution Bulletin 46:187 196. de Menezes, A. B., Prendergast Miller, M. T., Poonpatana, P., Farrell, M., Bissett, A., /N ratio drives soil actinobacterial cellobiohydrolase gene diversity. Applied and Environmental Microbiology , 81 (9), 3016 3028. Duggan, J. 2009. Longevity of Acid Rock Drainage : Mineralogical and Chemical Comparison of Mine Waste Piles and Post Glacial Talus Rock Producing Acidic Solutions. Colorado State University Department of Geosciences. Presentation given during Portland GSA Annual Meeting. Elliott P, Ragusa S, Catcheside D . 1998. Growth of sulfate reducing bacteria under acidic conditions in an upflow anaerobic bioreactor as a treatment system for acid mine drainage. Water Research 32 :3724 3730. Fisk M, Ruether K, Yavitt J. 2003. Microbial activity and functional composition among northern peatland ecosystems . Soil Biology & Biochemistry. 35: 5 91 602. Frank KL, Rogers DR, Olins HC, Vidoudez C, Girguis PR. 2013. Characterizing the distribution and rates of microbial sulfate reduction at Middle Valley hydrothermal vents. The ISME Journal ISME J . 7(7):1391 1401. Gadd GM. 2010. Metals, minerals and microbes: geomicrobiology and bioremediation. Microbiology 156:609 43. Giller KE, Witter E, Mcgrath SP. 1998. Toxicity of heavy metals to microorganisms and microbial processes in agricultural soils: a review. Soil Biology and Biochemistry . 10 11:1389 141 4. in Ecotoxicological Evaluation of Dairy Sewage Sludge. Applied Biochemistry and Biotechnology , 174 (4), 1434 1443. Hauff PL , Prosh, EC , Coulter DW, Sares, MA , Bir d D A , Peters D C, and Henderson FB, III. 2003. Evaluating the impa ct of natural acid drainage on Colorado watersheds using hyperspectral remote sensing, Proceed ings of the tenth international confere nce on tailings and mine waste. 3 7.

PAGE 82

73 Hirsch PR, Mauchline TH, Clark IM. 2010. Culture independent molecular techniques for soil microbial ecology. Soil Biol. Biochem. 42:878 887. Hodson ME. 2012. Effects of Heavy Metals and Metalloids on Soil Organisms. Environmental Pollution Heavy Metals in Soils 141 160. Ho gsden KL, Harding JS. 2012. Anthropogenic and natural sources of acidity and metals and their influence on the structure of stream food webs. Environ. Pollut. 162:466 74. Johnson DB, Hallberg KB . 2003. The microbiology of acidic mine waters. Research in Mi crobiology 154 :466 473. Jones, D. S., Kohl, C., Grettenberger, C., Larson, L. N., Burgos, W. D., & Macalady, J . L. (2015). Geochemical niches of Iron oxidizing acidophiles in acidic coal mine drainage. Applied and Environmental Microbiology , 81 (4), 1242 1250. Kamika I, Momba MN. 2013. Assessing the resistance and bioremediation ability of selected bacterial and protozoan species to heavy metals in metal rich industrial wastewater. BMC Microbiology BMC Microbiol 13 28. Lin, X, Green S, Tfaily M, Prakash O, Konstantinidis, KT, Corbett JE, Chanton JP , Cooper WT, and Kostka JE. 2012. Microbial Community Structure and Activity Linked to Contrasting Biogeochemical Gradients in Bog and Fen Environments of the Glacial Lake Agassiz Peatland." A pplied and Environmental Microbiology 78.19. 7023 031. 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 . (2014). Real time PCR for quantification in soil of glycoside hydrolase family 6 cellulase genes. Letters in Applied Microbiology , 59 (3), 284 291. Wang, G., Wang, Y., Yang, P., Luo, genes in alpine tundra soil. Applied Microbiology and Biotechnology , 87 (4), 1383 1393. Mikiya Hiroki. 1992. Effects of heavy metal contamination on soil microbial population. Soil Science and Plant Nutrition. 38:1, 141 147. Monsieurs P, Moors H, Van Houdt R, Janssen PJ, Janssen A, Coninx I, Mergeay M, Leys N. 2011. Heavy metal resistance in Cupriav idus metallidurans CH34 is governed by an intricate transcriptional network. Biometals 24:1133 51. Myers B, Webster KL, Mclaughlin JW, Basiliko N. 2011. Microbial Activity across a Boreal Peatland Nutrient Gradient: T he Role of Fungi and Bacteria. Wetlands Ecology and Management. 20: 77 88.

PAGE 83

74 Nies DH. 1999. Microbial heavy metal resistance. Appl. Microbiol. Biotechnol. 51:730 50. Quillet L, Besaury L, Popova M, Paissé S, Deloffre J, Ouddane B . 2011. Abundance, Diversity and Activity of Sulfate Reducing Prok aryotes in Heavy Metal Contaminated Sediment from a Salt Marsh in the Medway Estuary (UK). Marine Biotechnology Mar Biotechnol 14 :363 381. Ren N, Xie T, Xing D. 2009. Composition of extracellular polymeric substances influences the autoaggregation capabil ity of hydrogen producing bacterium Ethanoligenens harbinense. Bioresour. Technol. 100:5109 13. Rouch DA, Lee BT, Morby AP. 1995. Understanding cellular responses to toxic agents: a model for mechanism choice in bacterial metal resistance. J. Ind. Microbio l. 14:132 41. Sackett, J. 2015. Comparative Microbial Ecology of Sediment Associated Microbial Communities from Anthropogenically and Endogenously Metal Impacted Systems. Silver S. 1996. Bacterial resistances to toxic metal ions a revi ew. Gene. 179:9 19. Silver S, Phung, L T. 2009 . Heavy metals, bacterial resistance. In Encyclopedia of Microbiology , pp. 220 227 . Edited by M. Schaechter. Oxford: Elsevier. Sims A, Zhang Y, Gajaraj S, Brown PB, Hu Z. 2013. Toward the development of microb ial indicators for wetland assessment. Water Res. 47:1711 25. Swan BK, Martinez Garcia M, Preston CM, Sczyrba A, Woyke T, Lamy D, Reinthaler T, Poulton NJ, Masland ED, Gomez ML, Sieracki ME, DeLong EF, Herndl GJ, Stepanauskas R 2011. Potential for chemolit hoautotrophy among ubiquitous bacteria lineages in the dark ocean. Science 333 6047: 1296 300 Verplanck, P.L., ed., 2008, Understanding contaminants associated with mineral deposits: U.S. Geological Survey Circular 1328, p. 96 Wang Y, Shi J, Wang H, Lin Q, Chen X, Chen Y. 2007. The influence of soil heavy metals pollution on soil microbial biomass, enzyme activity, and community composition near a copper smelter. Ecotoxicol. Environ. Saf. 67:75 81 Zhu J, Zhang J, Li Q, Han T, Xie J, Hu Y, Chai L. 2013. Ph ylogenetic analysis of bacterial community composition in sediment contaminated with multiple heavy metals from the Xiangjiang River in China. Mar. Pollut. Bull. 70:134 9.