Viral Mineralization and Geochemical Interactions

by

Jennifer E. Kyle

A thesis submitted in conformity with the requirements

for the degree of Doctor of Philosophy

Graduate Department of Geology

University of Toronto

© Copyright by Jennifer E. Kyle 2009 Viral Mineralization and Geochemical Interactions

Jennifer E. Kyle

Doctor of Philosophy

Department of Geology University of Toronto

2009

Abstract

Viruses are ubiquitous biological entities whose importance and role in aquatic habits is beginning to take form. However, several habitats have undergone limited to no examination with viral-geochemical parameters minimally examined and viral-mineral relationships in the natural environment and the role of mineralization on viral-host dynamic completely lacking. To further develop knowledge on the presence and abundances of viruses, how viruses impact aquatic systems, and how viral-host interactions can be impacted under mineralizing conditions, viruses were examined under a variety of habitats and experimental conditions. Water samples were collected from the deep subsurface (up to 450 m underground) and acid mine drainage

(AMD) systems in order to determine the presence, abundance, and viral-geochemical relationships within the systems. Samples were also collected from a variety of freshwater habitats, which have undergone limited examination, to determine viral-geochemical and viral- mineral relationships. Lastly, bacteriophage-host dynamics were examined under authigenic mineral precipitation to determine how mineralization impacts this relationship.

Results reveal that not only are viruses present in the deep subsurface and AMD systems, but they are abundant (up to 107 virus-like particles/mL) and morphogically diverse. Viruses are also the strongest predictor of prokaryotic abundance in southern Ontario freshwater systems

ii where potential nutrients are rich. Geochemical variables, such as pH and Eh, were shown to have negative impacts of viral abundance indicting that AMD environments are detrimental for free viruses (i.e. not particle associated).

Direct evidence of viral-mineral interactions was found using transmission electron microscopy as viral particles were shown attached to iron-bearing mineral phases (determined through elemental analysis). In addition, evidence of viral participation in mineralization events was found in both AMD and freshwater environments where inverse correlations were noted between viral abundance and jarosite saturation indices (r = -0.71 and r = -0.33, respectively), and goethite saturation indices were also noted to be the strongest predictor of VLP abundance in freshwater habitats explaining 78% of the variability in the data. Lastly, iron precipitation and/or metal ion binding to bacterial surfaces greatly reduced phage replication (~98%) revealing bacterial mineralization has a protective benefit strongly hindering viral replication.

iii Statement of Authorship

The work presented in this thesis represents my research ideas and efforts in collaboration with ideas and suggestions by my supervisor Grant Ferris as acknowledged by the co-authorship in the publications and manuscripts that resulted from this work. Chapters 2 and 3 are prefaced by a reference in which the manuscript has been published. Permission from the publishers and co-authors to include the manuscripts in my thesis has been received. Chapter 4 represents work that has been submitted.

Chapter 2: Viruses in Granitic Groundwater from 69 to 450 m Depth of the Äspö Hard

Rock Laboratory, Sweden.

Jennifer E. Kyle, Hallgerd S. C. Eydal, F. Grant Ferris, and Karsten Pedersen.

Jennifer Kyle and Hallgerd Eydal planned the field session, the type of analyses to be conducted, and collected the samples. Jennifer Kyle prepared and analyzed sample on the transmission electron microscopy and Hallgerd Eydal conducted microbial counts. Jennifer Kyle conducted data interpretation with the assistance of Karsten Pedersen. Jennifer Kyle and Karsten Pedersen wrote a majority of the manuscript with input from Hallgerd Eydal and Grant Ferris. Published:

ISME (2008), vol. 20, pp. 571-574

Chapter 3: Virus Mineralization at Low pH in the Rio Tinto, Spain

Jennifer E. Kyle, Karsten Pedersen, and F. Grant Ferris

Jennifer Kyle developed the idea, designed the sampling protocol, and performed the analysis.

Bob Harris at the University of Guelph assisted with energy dispersive spectroscopy on the transmission electron microscope and obtaining the image shown in figure 3.

iv Jennifer Kyle collected all samples with the assistance of Grant Ferris, and performed in situ physiochemical measures with the assistance of Grant Ferris and Karsten Pedersen. Jennifer

Kyle conducted data interpretation with the assistance of Grant Ferris. Jennifer Kyle wrote the paper with input from Grant Ferris and editing and suggestions by Karsten Pedersen. Published:

Geomicrobiology Journal (2008), vol. 25, pp. 338-345.

Chapter 4:

Jennifer E. Kyle and F. Grant Ferris

Jennifer Kyle developed the idea, designed the sampling protocol, collect and prepared all samples, performed all the analyses with the exception of the following: Dan Mathers performed

ICP-AEOS on 2 of the 3 sampling sessions from Sudbury, and Wendi Abi at the University of

Ottawa performed dissolved carbon analysis. Joe Fyfe and Robin Armstrong assisted with field sampling locations located on Xstrata Canada property. Jennifer Kyle performed data interpretation and wrote the manuscript with input from Grant Ferris. Submitted: Applied and

Environmental Microbiology.

Chapter 5:

Jennifer E. Kyle

Jennifer Kyle developed the idea, designed the sampling protocol, collect and prepared all samples, performed all the analyses, and interpreted data. Jennifer Kyle wrote the manuscript with editing performed by Grant Ferris.

v Acknowledgements

I would first like to thank my supervisor, Grant Ferris. I recall walking into our first meeting with references in hand, ideas written on scraps of paper, crossing my fingers that he would at least entertain the doctoral research pitch I was about to give. From our first meeting,

Grant has been supportive, encouraging, and enthusiastic of letting me explore an idea I had developed, an idea that other people disregarded. Grants continuous belief in my research and continuous support has been greatly appreciated and is a quality I hope to possess towards students and colleagues in the future.

To past and current members of the Microbial Geochemistry Lab, I would like you for your advice, assistance, and good times. It has been a pleasure working with each of you and I look forward to future collaborations. So thank you Samantha Smith, Andy Mitchell, Chris

Omelon, Rachel James, Kerry Evans-Tokaryk, and Chris Kennedy (who was invaluable towards the end of my research).

To my father, I appreciate and thank you for the many offers to assist in any way possible. It is comforting to know that there is always someone that I can count on for anything and everything. I would especially like to thank my mom. The early morning, midday, and late evening phone calls helped keep me going these past four years. There is no other person who has been more supportive. Emotionally, this PhD thesis belongs to my mom just as much as myself. I think my mom was more frustrated when experiments did not go as hoped and more excited when they did than I.

Lastly, I would like to thank Justin. A person whose passion for life and love for science is something to be desired.

vi Table to Contents

Abstract ii

Statement of Authorship iv

Acknowledgements vi

Table of Contents vii

List of Figures xiii

List of Tables xiv

Chapter 1: Introduction 1

1.1 Research statement 1

1.2 General description of bacteriophages 2

1.2.1 Phage morphology 2

1.2.2 Locality 3

1.2.3 Phage abundance 4

1.3 Bacteriophage replication 5

1.3.1 Attachment 5

1.3.2 Penetration and genome injection 5

1.3.3 Synthesis of phage components 6

1.3.4 Assembly 7

1.3.5 Release 7

1.4 Phage attachment and mineral sorption to bacterial surfaces 8

1.4.1 Bacterial mineralization 8

1.4.2 Receptor sites for phage attachment and mineralization 9

vii 1.4.2.1 Gram negative bacteria 9

1.4.2.2 Gram positive bacteria 10

1.5 Prokaryotic Habitats 10

1.5.1 The deep subsurface 10

1.5.2 Acid mine drainage 11

1.6 Role of viruses in biogeochemistry 12

1.7 Viral-Mineral interactions 12

1.7.1 Mineral and viral surface charges 13

1.7.2 Factors effecting viral-mineral sorption 13

1.8 Viral preservation 15

1.9 References 15

Chapter 2: Viruses in Granitic Groundwater From 69 to 450 m Depth of the Äspö Hard Rock Laboratory, Sweden 22

2.1 Abstract 24

2.2 Short communication 25

2.3 Acknowledgements 29

2.4 References 29

2.5 Figure legends 32

Chapter 3: Viral Mineralization at Low pH in the Rio Tinto, Spain 35

3.1 Abstract 37

3.2 Introduction 38

viii 3.3 Methods and Materials 39

3.3.1 Site description and sample collection 39

3.3.2 Prokaryotic and viral abundance 40

3.3.3 Transmission electron microscopy 41

3.3.4 X-Ray diffraction 41

3.3.5 Geochemical and Statistical Calculations 42

3.4 Results 42

3.4.1 Microbial abundance and physiochemical correlations 42

3.4.2 Viral diversity 43

3.4.3 Viral-inorganic particle association 43

3.5 Discussion 44

3.6 Acknowledgements 50

3.7 References 50

3.8 Table legends 55

3.9 Figure legends 57

Chapter 4: Geochemistry of Virus – Prokaryote Interactions in Freshwater and Acid Mine Drainage Environments, Ontario, Canada 60

4.1 Abstract 62

4.2 Introduction 63

4.3 Methods and Materials 64

4.3.1 Site description and sample collection 64

4.3.2 Viral and prokaryote abundance and viral imaging 65

ix 4.3.3 Aqueous chemistry 66

4.3.4 Statistical and geochemical data analysis 66

4.4 Results 67

4.4.1 Aqueous chemistry 68

4.4.2 Viral-prokaryote abundance and geochemical relationships 68

4.4.3 Predictors of prokaryotic abundance 69

4.4.4 Viral-mineral correlations 69

4.5 Discussion 70

4.5.1 Relationships between viruses, prokaryotes, and geochemical variables 70

4.5.2 Viral-mineral correlations 73

4.6 Acknowledgements 77

4.7 References 77

4.8 Table legends 85

4.9 Figure legends 89

Chapter 5: Bacterial-phage interactions and authigenic mineral precipitation 92

5.1 Abstract 92

5.2 Introduction 93

5.3 Methods and Materials 95

5.3.1 Site Characterization and sample collection 95

5.3.1.1 Viral and prokaryotic abundance and viral imaging 95

5.3.1.2 Aquatic chemistry 96

x 5.3.1.3 Isolation of IOB phage 96

5.3.1.4 Isolation of IOB phage from Longvac 99

5.3.2 Bacillus subtilis – SPβc2 mineralization experiments 99

5.3.2.1 Obtaining lysate 100

5.3.2.2 Plaque assay 101

5.3.2.3 Experiment 1: Bacillus subtilis with iron plus phage 101

5.3.2.4 Experiment 2: Lysogen plus iron 103

5.3.2.5 Experiment 3: Phage with iron plus Bacillus subtilis 104

5.4 Results 104

5.4.1 AMD site characterization 104

5.4.1.1 IOB phage isolation using foreign cultures 105

5.4.1.2 IOB phage isolation 105

5.4.1.3 IOB phage isolation using Longvac cultures 105

5.4.2 Bacillus subtilis – SPβc2 mineralization experiments 106

5.4.2.1 Experiment 1: Bacillus subtilis with iron plus phage 106

5.4.2.2 Experiment 2: Lysogen plus iron 106

5.4.2.3 Experiment 3: Phage with iron plus Bacillus subtilis 107

5.5 Discussion 107

5.5.1 IOB phage isolation 107

5.5.2 Bacillus subtilis – SPβc2 experiment 110

5.6 Conclusions 114

5.7 Acknowledgements 114

5.8 References 115

xi 5.9 Table legends 120

5.10 Figure legends 122

Chapter 6: Synthesis and Future Work 128

6.1 Synthesis 128

6.1.1 Viruses in extreme environments 128

6.1.2 Viral control of prokaryotic abundance 129

6.1.3 Viral-mineral and viral-geochemical interactions 129

6.1.4 Role of bacterial mineralization in phage replication 130

6.2 Future Work 131

6.2.1 Environmental phage therapy in acid mine drainage 132

6.2.2 Phage-host dynamics under mineralizing conditions 132

6.2.3 Viral mineralization and preservation 132

Appendix I: Exact microbial and VLP counts for Rio Tinto and Ontario samples 134

Appendix II: Evidence of strength of multiple regression model in Chapter 4 where 137 the dependent variable is prokaryotic abundance.

Appendix II: Evidence of strength of multiple regression model in Chapter 4 where 140 the dependent variable is VLP abundance.

xii List of Figures

Fig. 1.1: Common bacteriophage morphotypes. 3

Fig. 2.1a: The relation between the total number of cells (TNC) and the number of virus like particles (VLP) in groundwater from Äspö hard rock laboratory. 33

Fig. 2.1b: The relation between the average of 10log number of VLP, depth, and amount of chloride in groundwater from Äspö hard rock laboratory. 33

Fig. 2.2: TEM of VLPs from Äspö hard rock laboratory. 34

Fig. 3.1: Map of Rio Tinto sampling sites. 58

Fig. 3.2: TEM micrograph of common phage morphotypes found in the Rio Tinto. 58

Fig. 3.3: High-resolution TEM micrograph of a Myoviridae phage. 59

Fig. 3.4: TEM micrographs of RT-066 with inorganic, iron-bearing mineral phases attached to the phages. 59

Fig. 4.1: Map of southern Ontario with sample locations. 90

Fig. 4.2: TEM micrographs of common VLPs found in southern Ontario surface waters. 90

Fig. 4.3: TEM micrograph of possible VLPs sorbed to inorganic material from an AMD site. 91

Fig. 5.1: TEM image of mineralized Bacillus subtilis after 30 min incubation with iron and SPβc2. 123

Fig. 5.2: TEM image of lysogen with minimal mineralization (a) and extensive mineralization (b). 124

Fig. 5.3: SPβc2 surrounds cells partially surrounded by ESP (a-d) noted using TEM. 125

Fig. 5.4: SPβc2 surrounds dividing Bacillus subtilis cell in a lysogenic culture. 126

Fig. 5.5: TEM image of spherical (spore?) noted after 20 hours of incubation of lysogen with iron. 127

Fig. 5.6: TEM image of lysogen after 2 hours of iron incubation. 127

xiii List of Tables

Table 3.1: Physiochemical characteristics and microbial abundances of Rio Tinto water samples 55

Table 3.2: Correlation indices of Rio Tinto chemical constituents with microbial abundance and physiochemical characteristics. 55

Table 3.3: Major dissolved ion concentration of Rio Tinto water samples 56

Table 3.4: Saturation index values from geochemical modelling of Rio Tinto water samples. 56

Table 4.1: Prokaryote, virus, physiochemical, and geochemical concentrations determined for each sample location. 85

Table 4.2: Spearman rank correlation coefficients of microbial and physiochemical constituents. 87 Table 4.3: Multiple regression analysis with prokaryote abundance as the dependent variable. 88

Table 4.4: Simple linear regression analysis with VLP abundance as dependent variable. 88

Table 4.5: Pearson correlation coefficient of mineral saturation indices verses pH and microbial constituents. 88

Table 5.1: Geochemical constituents and prokaryotic and viral abundances of AMD waters. 120

Table 5.2: Mean values of results for the mineralized bacteria plus phage microcosms. 120

Table 5.3: Mean values of measurements conducted in the lysogen plus iron microcosms. 121

Table 5.4: Results of Bacillus subtilis plus iron over time. 121

xiv Chapter 1

Introduction

1.1. Research Statement

Viruses are small, infectious, intracellular parasites that require a host organism to replicate. As a biological entity that the scientific community debates whether or not is living, viruses are dynamic, important members of any biological community. The existence of viruses, more specifically bacteriophages (viruses that infect prokaryotic microorganisms), within natural environments has only come to light within the past couple of decades with their ubiquity only recently established. The ecological role of bacteriophages (herein referred to as phages) in aquatic environments, especially marine environments, is beginning to be understood; however, notably lacking from the literature is the role that geochemical and mineralogical variables have on viral populations. More specifically (i) the presence and abundance of viruses in extreme biogeochemical ecosystems, such as the deep subsurface and acid mine drainage (AMD), (ii) the influence of geochemical variables on viral abundances in unexamined or minimally examined aquatic environments, (iii) the role of minerals on viral dynamics, and (iv) the role of mineralization on phage-host relationships.

The goal of this research is to (i) identify the presence and abundance of viruses in the deep subsurface and AMD environments, (ii) determine viral-geochemical and viral-mineral interactions within AMD environments, (iii) determine viral-geochemical and viral-mineral interactions within freshwater environments, and (iv) examine phage-host dynamics under natural and experimental mineralizing conditions.

The importance of this research is multifaceted and ranges from expanding our knowledge of aquatic microbial geochemistry to the potential discovery of novel microbial

1 biosignatures (i.e. microfossils). In addition, this research has bioremediation implications where phages could be used for environmental phage therapy (i.e. the use of phages to eliminate environmental problems created or propagated by prokaryotes, such as AMD). Also, as AMD environments are possible Earth analogs to past processes on Mars, gaining an understanding of viral influences and/or viral-geochemical and viral-mineral interactions would enhance our knowledge of current systems with the ability to apply this information to past events (i.e. “the present is the key to the past”).

1.2 General Description of Bacteriophages

1.2.1 Phage Morphology

Phages are typically 30-100 nm in diameter and comparatively simple biological entities as they are all composed of a protein head (called a capsid) that contains the viral genome. Most known bacteriophages have double stranded (ds)DNA. The vast majority of phages (96 % of those studied thus far; Ackermann 2007) are tailed (Figure 1), which is used to attach to the host bacterial cell and channel the viral genome into the host. Phage receptors used in host attachment are commonly located at the tail tip. Phages without tails attach and inject their genetic material through the expression of proteins located within the capsid (Kutter et al. 2005).

Classification of phages are based mainly on phage morphology (Fig. 1; Ackermann 2007) and its host genus; however, as more bacteriophages are isolated from nature, classification is moving towards nucleic acid sequencing (Ackermann 2006).

2

Figure 1. Common bacteriophage morphotypes. Phage morphotypes are divided into four groups containing a variety of viral groups (families).

1.2.2 Locality

Phages are seemly ubiquitous and present wherever there is a potential host (i.e. prokaryotes). They have been discovered all over the world in a variety of environments including marine (Suttle 2005), estuarine (Cochran and Paul 1998; Hewson et al. 2001), lacustrine (Maranger and Bird 1995), sediments (Maranger and Bird 1996; Ricciardi-Rigault et al. 2000), and soils (Ashelford et al. 2003). In addition, abundant phage populations have been described from many extreme environments including deep-sea hydrothermal vents (Geslin et al.

2003; Ortmann and Suttle 2005), high temperature terrestrial hot springs (Rachel et al. 2002),

Antarctic waters (Guixa-Boixereu et al. 2002) and perennial lakes (Lisle and Priscu 2004), and

3 hypersaline lakes (Oren et al. 1997; Jiang et al. 2004). Some work has been conducted within acid mine drainage environments (Ward et al. 1993; Allen et al. 2007), but viral abundances and morphological descriptions have not been reported. No reports of viruses within the deep subsurface have, to our knowledge, been previously published.

1.2.3 Phage Abundance

Viruses are the most abundant biological entity on Earth (total 1030 to 1032 viruses; Suttle

2005), composing ~ 94 % of the nucleic-acid-containing particles in the oceans; however, due to their small size they only comprise ~ 5 % of the total oceanic biomass (Suttle 2007). Current estimates suggest there is at least one virus for every living organism (Flint et al. 2000). Viral abundance is commonly found to be one to two orders of magnitude greater than that of the prokaryotic population, although this is not always the case (Alonso et al. 2001). Typical viral abundances within the systems mentioned above range from 105 to 108 phage particles/mL (see references above).

Viral abundance is believed to be governed by the abundance and productivity of the host organisms. For example, viral and bacterial abundances are typically greater in marine sediments vs. the water column, oxic vs. anoxic waters (Ricciardi-Rigault et al. 2000), and during the summer vs. the winter (Cochran and Paul 1998).

The most common cause of viral decay is UV radiation from sunlight. Given that the viral genome is only protected by the viral capsid, UV radiation only has a few nanometers of biopolymer to penetrate before reaching and mutating the genome. Additional decay factors include heat and hydrolytic enzymes (Fuhrman 1999).

4 1.3 Bacteriophage replication

Viruses require a host’s metabolism to replicate. Phages can only reproduce through infecting and taking over the host’s biosynthetic machinery. In order for viral particles to be replicated, 5 distinct steps are required: (i) viral attachment to host cell, (ii) penetration of the host cell envelope to enable passage for viral genetic material (injection), (iii) synthesis of viral components (i.e. proteins and nucleic acids), (iv) assembly of viral particles, and (v) release of progeny into the environment.

1.3.1. Attachment

The phage-host cell relationship is specific as phages typically attach to prokaryotes of a specific species, or sometimes only to a specific strain of an individual species, that contain complementary receptor sites. Receptor sites are typically found on the cell surface, such as in lipopolysaccharides (LPS) for gram negative bacteria, and teichoic acids in gram positive bacteria; however, some are located on external capsules, flagella, or pili (Lindberg 1973). For phage attachment, initial contact between a phage and its host bacteria is governed by chance

(Flint et al. 2000). The rate of attachment follows second order kinetics as dA/dt = k[V][H], where [V] and [H] is the concentration of phages and host cell bacteria, respectively, and k is the rate constant (Flint et al. 2000). When the host density and/or phage concentration is high, the likelihood of a phage encountering its host is greater.

1.3.2. Penetration and genome injection

Once a phage irreversibly attaches to its host receptors, viral DNA can then be injected into the host cell although the mechanism of DNA injection varies between phages (Kutter et al.

5 2005). Commonly, an enzymatic reaction occurs at the tail tip causing the formation of a small hole within the cell envelope creating a small channel for the viral genome to be ejected from the capsid into the host cell (Guttman et al. 2005). For SPP1 phage (host Bacillus subtilis), once the tail is in position to transfer the genome, conformational changes occur along the tail structure signaling the release of the genome within the capsid into its host (Plisson et al. 2007).

1.3.3. Synthesis of phage components

Once the viral DNA is injected into the host cell, one of two replication cycles can take place depending of the types of phage infecting the cell. Lytic phages immediately takeover of host’s cellular metabolism resulting in a lytic replication cycle. Whereas temperate phages incorporate the viral DNA into the host genome called a lysogenic replication cycle.

In the lytic cycle, phages immediately take over the host’s replication mechanics. Early viral genes are transcribed for the production of viral enzymes used to stop host DNA synthesis and of products that assist in protecting the phage genome from host attack (Gutterman et al.

2005; Kutter et al. 2005). Intermediate viral genes transcribe for the production of viral DNA, and late genes produce viral components such as the capsid and tail (Gutterman et al. 2005).

A lysogenic cycle differs in that the viral genome becomes integrated into the host cell genome (called a prophage) instead of automatically taking over the cell. In a lysogen (a bacterium containing a prophage), both the viral and host genomes are replicated as the host cell continues to grow and divide, which in turn creates additional infected cells. A prophage may be induced into a lytic cycle by inducing agents such as environmental stress (changes in temperature, pH, salinity, UV radiation, and/or introduction of harmful pollutants). When this occurs, the viral genome is excised from the host genome and a lytic cycle, as described above,

6 begins. Lysogeny is thought to be a common replication cycle as this strategy enables phages to survive when host density is low or environmental conditions are not favorable for free phages.

In addition, lysogeny can be beneficial for the host cell because, once a host is infected, the host becomes immune to infection by the same or related viruses (Fuhrman 1999). Lysogens also contain viral genes that may contribute to the survival of the host population through increasing the host’s virulence factor (Chibani-Chennoufi et al. 2004). In a process called phage conversion, phage encoded virulence genes can covert a nonpathogenic bacteria to a virulent strain or increase the virulence of a particular stain (Boyd 2005). For example, many common bacterial diseases seen in humans (i.e. cholera, botulism, diphtheria, toxic shock syndrome) are the result of phage encoded toxins (see Boyd 2005) causing the illness.

1.3.4. Assembly

Once the genes for the viral components (i.e. capsids, nucleic acids, tail structures) are transcribed proteins are then able to form. The genome is packaged into the capsids at which time the tail, if applicable, attaches to the capsid (Gutterman et al. 2005); however in a recent study, a phage, Acidianus two-tailed virus (ATV), was found to develop its tail after being released from its’ host Acidianus convivator (Häring et al. 2005). In addition, multiple virions

(viral particles within the host cell) are commonly produced within one host cell.

1.3.5. Release

The final step in viral replication is the release of newly formed virions into the surrounding environment. Viral enzyme(s) break down the cell envelope structure causing the cell to break down and burst open (i.e. lysis) enabling the release of newly assembled viral

7 particles into the surrounding environment. The number of viral particles produced per host cell is referred to as the burst size. Burst sizes can range from just a few viral particles up into the hundreds. Multiple factors can affect the burst size including the length of time from infection to assembly (referred to as the latent period), host growth conditions, and the size of the host cell.

1.4 Phage Attachment and Mineral Sorption to Bacterial Surface

1.4.1 Bacterial Mineralization

Bacterial surfaces are prone to mineralization as the cell surface usually has an overall net electronegative charge (at ~ pH 7), attracting mineral forming ions in the surrounding environment. This negative charge is due to the carboxyl and phosphoryl, and to a lesser extent hydroxyl groups, found in the cell wall. Abundant carboxyl groups found within peptidoglycan layer of the cell wall are responsible for most of the cells negative charge. Phosphoryl groups found in teichoic and teichuronic acids in gram positive bacteria, and phosphoryl and carboxyl groups in lipopolysaccharides (LPS) in gram negative bacterial, contribute additionally to this negative charge. Positively charged groups also exist, such as amine groups in teichoic acids, and amino groups in peptidoglycan (Beveridge and Murray 1980), but they are less abundant.

This net negative charge of the bacterial surface causes the bacterial cell to act as a passive nucleation site for metal deposition and mineral formation. Commonly, mineral formation is unintended and uncontrolled by the organism as the cell interacts with ions in the surrounding environment and metabolic byproducts excreted by the organism itself (Frankel and Bazylinski

2003). Mineral formation in this instance is induced upon the cell. Initial minerals are usually hydrous and poorly ordered, but over time they lose water and become more crystalline (Fortin et al. 1997). Metal cations in solution are able to directly bind to anionic functional groups, but

8 anionic ions, such as silicate, must bind to either amine groups (depending on pH) or form metal ion bridges with carboxyl and/or phosphate groups to bind at the cell surface (Mera and

Beveridge 1993).

1.4.2 Receptor sites for phage attachment and mineralization

1.4.2.1 Gram negative bacteria

Gram negative bacterial phage receptor sites are found in LPS, membrane proteins, and phospholipids (Lindberg 1973). After phage attachment and subsequent infection, a lysogenic conversion induced by a temperate phage will often alter the surface characteristics to deter additional infection and/or reduce attachment rate of related phage species. If an unrelated phage succeeds in injecting their viral genome into an already infected cell, the original phage genome encodes for repressor proteins that block transcription of the new phage genome (Kuttle and

Sulakelidze 2005).

For attachment to occur the presence of a particular molecule is often not sufficient enough for attachment, but how the molecule is assembled and attached to neighbouring molecules is crucial. Variations within and connections between molecules enables different phages to attach (Lindberg 1973).

In terms of mineralization, the outer surface of gram negative bacteria also interacts with metal cations in solution. The phosphoryl groups in LPS and phospholipids are the most reactive electronegative sites in the outer membrane capable of binding metal cations (Ferris and

Beveridge 1986). This would be a major site for metal cation interaction and mineral formation.

Components within LPS molecules contain a free carboxyl, and repeating residues can be substituted by anionic groups making these reactive sites for metal cation binding (Ferris 1989).

9 1.4.2.2 Gram positive bacteria

Common phage receptors are found in teichoic acids of gram positive bacterial cell walls.

The diversity of residues and the linkages within and between units in teichoic acids specifies which phages can attach. Teichoic acids are also active sites for mineral formation as the negatively charged phosphate groups attracts ions from the surrounding environment. For gram positive bacteria the negative charge is concentrated on the outer surface (Doyle 1989) where teichoic acids are available for interaction. Carboxyl groups in peptidoglycan undergo more metal deposition compared to phosphodiester groups of teichoic acids possibly due to greater accessibility (Beveridge and Murray1980).

1.5 Prokaryotic Habitats

As discussed above, viruses are believed to be present wherever potential hosts exists.

Two environments that have undergone no or limited investigation in terms of viral ecology are the deep subsurface and acid mine drainages (AMD).

1.5.1 The Deep Subsurface

Microbial diversity within subsurface environments is largely based on host rock mineralogy and the aquatic geochemistry of fracture waters (Sahl et al. 2008). The deep subsurface is unique in that it is largely a photosynthesis-independent system as organic carbon derived from photosynthetic processes is quickly depleted near the Earths surface (Pedersen and

Karlsson 1995; Pedersen 1999). One hypothesis for the origins of life on Earth is that life originated in the deep subsurface as radiation from the Sun would have been detrimental to any life forming on the Earths surface. One of the most extensively examined deep subsurface

10 environments is located in southeastern Sweden.

Äspö Hard Rock Laboratory (HRL) in Sweden is a research site that investigates subsurface storage of nuclear waste in crystalline granitic bedrock. At the HRL, total prokaryotic abundance ranges from 103 to 107 cells/mL where microbes such as sulfate and iron reducers, homoacetogens, and acetoclastic and autotrophic methanogens comprise a large percentage of the total number of prokaryotes (Pedersen 1996, 1999). It is suggested that the deep granitic biosphere is hydrogen driven with hydrogen produced by geologic processes

(Pedersen 1999). Until this investigation, the presence of viruses in the deep subsurface had not been considered.

1.5.2 Acid Mine Drainage

Acid mine drainages are characterized by low pH (~ < 4.0) and a high content of dissolved metals (i.e. Fe, Cu, and Zn). They are commonly found in mining environments where sulfide minerals are being excavated for ore production. The exposure of sulfide minerals (most commonly pyrite) to the atmosphere and water results in chemical iron oxidation of sulfides and the production of acidic conditions, commonly depicted by the following equations (Nordstrom and Southam 1997; Baker and Banfield 2003, respectively):

[Eq.1] FeS2 + 15/4O2(atm) + 7/2H2O  Fe(OH)3 + 2H2SO4

2+ 2- + [Eq.2] FeS2 + 3.5O2(atm) + H2O  Fe + 2SO4 + 2H

Chemical (i.e. abiotic) oxidation of ferrous to ferric iron is very slow in acidic environments (pH of less than 4.0), but is catalyzed by a factor of five (Singer and Stumm 1970)

11 by iron oxidizing bacteria (IOB). Biotic oxidation of ferrous iron results in the production of ferric iron, which in turn is able to act as an oxidant of pyrite creating a closed feedback loop where upon ferric iron oxidizes pyrite:

2+ + 3+ [Eq.3] Fe + 0.5O2 + 2H  Fe + H2O (biotic; IOB)

3+ 2+ 2- + [Eq.4] 14Fe + FeS2 + 8H2O  15Fe + 2SO4 + 16H

Although acidophilic (acid-loving) microorganisms do not initiate AMD, they play a large role in the propagation and maintenance (ferric iron acts as a buffering agent (see Eq. 4;

Fernández-Remolar et al. 2004)) of acidic (and therefore metal-rich) aquatic environments.

1.6 Role of Viruses in Biogeochemistry

Viruses are acknowledged as catalyst in biogeochemical cycles in aquatic environments

(Suttle 2005). The nature of viral activity results in the production of dissolved organic matter

(DOM) from particulate organic matter (POM) as lytic phages typically burst host cells for their release. This results in a trophic level transfer of nutrients (C, P, N) from organisms that graze on bacterial cells to heterotrophic bacteria that can utilize DOM (see Middelboe et al. 1996;

Wilhelm and Suttle 1999; Middelboe and Jorgensen 2006; Riemann et al. 2009). Also POM to

DOC conversion keeps nutrients within the given community for longer periods of time instead of sinking to further depths in the water column (Suttle 2005). In addition, viral lysis releases nutrients that may be limiting within the environment (i.e. N, P, Fe; Fuhrman 1999).

1.7 Viral-Mineral Interactions

12 Viral-inorganic particle interactions have undergone examination for determining conditions upon which viral particles are removed and released from and into aquatic systems for the purpose of understanding the transport of infective agents in drinking water. These studies along with an understanding of colloidal stability (DLVO theory) has provided a great amount of knowledge on viral-mineral interactions.

1.7.1 Mineral and Viral Surface Charges

The total net surface charge for a mineral is a result of (1) the permanent structural charge of the mineral, (2) the net proton charge, (3) the inner-sphere complex, and (4) the outer-sphere complex (Stumm and Morgan 1996). The inner- and outer-sphere complexes, referred to as the electric double layer, begins at the mineral surface and extends into the bulk solution. The inner- sphere complex, herein referred to as the Stern layer, is considered a fixed layer next the mineral surface that contains counterions that partially neutralize the surface charge. The outer-sphere complex, herein referred to as the Gouy layer, is next to the Stern layer and extends outward into solution. The thickness of the Gouy layer is strongly dependent upon the pH and ionic strength of the bulk solution, and determines the force and distance in which particles interact (Gerba

1984).

Viral particles are dominantly composed of proteins that are composed of amino acids, some of which (glutamic acid, aspartic acid, histidine, and tyrosine) contain functional groups, such as carboxyl and amine groups, that can become ionized (Gerba 1984; Loveland et al. 1996) giving the viral surface a net charge.

1.7.2 Factors effecting viral-mineral sorption

13 Electrostatic and van der Waals forces are the major forces influencing viral-mineral sorption with the solution pH, isoelectric point of both the viral particles, and the inorganic materials acting as a major determinant in the interaction between the two particles (Gerba 1984;

Loveland et al. 1996; Dowd et al. 1998; Guan et al. 2003). Viral particles typically have pHIEP

(the pH at which the protein carries no net charge; IEP is the isoelectric point) values ranging from 3-11 (Gerba 1984). When the pH of the aquatic environment is below the pHIEP of the virus, the virus will have a net positive surface charge resulting in viral attachment to negatively charged mineral surfaces. As the pH of solution increases above the pHIEP of the virus, repulsive electrostatic forces between the viral particle and inorganic surface will result in detachment of the virus. Typically, low pH values result in irreversible viral attachment with higher pH values favoring free (i.e. unattached) viral particles (Gerba 1984; Schulze-Makuch et al. 2003). Also, minerals with higher pHIEP typically act as better viral adsorbants (Murray and Laband 1979) as there is a greater probability that the mineral will have a positive charge in near neutral aquatic systems (Gerba et al. 1984).

In addition to pH of solution and pHIEP of the involved particles, ionic strength influences viral-mineral sorption. High ionic strength solutions decrease the thickness of the Gouy layer enabling viral particles to closely approach mineral surfaces. Even if particles contain like charges, attractive van der Waals forces enable viral attachment (Gerba 1984). Low pH values also have the same effect as high ionic strength solutions. In low ionic strength solutions, electrostatic forces dominate over van der Waals inhibiting attachment between like charges.

Other factors that influence viral sorption are the presence of cations and dissolved organic matter (DOM) (Gerba 1984). Cations, such as Ca2+, have been shown to increase viral adsorption through the formation of a cation bridge enabling two negatively charged particles to

14 attach. However, this is not always the case. Zhuang and Jin (2003) found that divalent cations increased viral transport as the cations decreased viral-mineral attractive electrostatic interactions by interacting with the negatively charged viral particles.

The presence of DOM tends to result in greater viral abundance in solution as DOM has been found to hinder viral mineral attachment (Ryan et al. 1999; Blanford et al. 2005; Foppen et al. 2006). Dissolved organic matter tends to out compete viral particles in mineral sorption as both commonly posses similar net charges at similar pH values.

1.8 Viral Preservation

To date there is no known evidence of viruses within the rock record, which is believed to be due to their small size (30-100 nm) and lack of a unique chemical and isotopic signature.

For these reasons, viruses have either not been considered or disregarded in terms of viral preservation and the development of microfossils. However, one study conducted by Daughney et al. (2004) found that when viruses are subjected to increased amount of iron, iron becomes associated with the viral particles causing distinction between viral particles and inorganic iron oxides difficult.

1.9 References

Ackermann HW. 2007. 5500 phages examined in the electron microscope. Arch. in Virol.

152:227-243.

Allen EE, Tyson GW, Whitaker RJ, Detter JC, Richardson PM, Banfield JF. 2007. Genome

dynamics on a natural archaeal population. PNAS 104:1883-1888.

Anesio AM, Hollas C, Granéli W, Laybourn-Parry J. 2004. Influence of humic substances on

15 bacterial and viral dynamics in freshwaters. Appl. Environ. Microbiol. 70:4848-4854.

Alonso MC, Jimenez-Gomez- F, Rodriguez J, Borrego JJ. 2001. Distribution of virus-like

particles in an oligotrophic marine environment (Alboran Sea, Western Mediterranean).

Microb. Ecol. 42:407-415.

Ashelford KE, Day MJ, Fry JC. 2003. Elevated abundance of bacteriophage infecting bacteria in

soil. Appl. Environ. Microb. 69:285-289.

Baker BJ, Banfield JF. 2003. Microbial communities in acid mine drainage. FEMS Microbiol.

Ecol. 44:139-152.

Beveridge TJ, Murray GE. 1980. Sites of metal deposition in the cell wall of Bacillus subtilis. J.

Bacteriol. 141:876-887.

Blanford WJ, Brusseau ML, Jim Yeh TC, Gerba CP, Harvey R. 2005. Influence of water

chemistry and travel distance on bacteriophage PRD-1 transport in a sandy aquifer. Water

Res. 39:2345-2357.

Boyd EF. 2005. Bacteriophages and bacterial virulence. In: Kutter E, Sulakelidze A., editors.

Bacteriophages: biology and applications. Baton Raton, FL: CRC Press. P 223-266.

Chibani-Chennoufi S, Bruttin A, Dillmann ML, Brüssow H. 2004. Phage-host dynamics: an

ecological perspective. J. Bacteriol. 186:3677-3686.

Cochran PK, Paul JH. 1998. Seasonal abundance of lysogenic bacteria in a subtropical estuary.

Appl. Environ. Microbiol. 64:2308-2312.

Daughney CJ, Châtellier X, Chan A, Kenward P, Fortin D, Suttle CA, Fowle DA. 2004.

Adsorption and precipitation of iron from seawater on a marine bacteriophage (PWH3A-P1).

Marine Chem. 91:101-115.

Dowd SE, Pillai SD, Wang S, Corapcioglu MY. 1998. Delineating the specific influence of virus

16 isoelectric point and size on virus adsorption and transport through sandy soils. Appl.

Environ. Microbiol. 64:405-410.

Doyle RJ. 1989. How cell walls of gram-positive bacteria interact with metal ions. In: Beveridge

TJ, Doyle RJ, editors. Metal Ions and Bacteria. New York, NY: John Wiley & Sons. P 275-

294.

Fernández-Remolar DC, Gómez-Elvira J, Gómez F, Sebastian E, Martíin J, Manfredi JA, Torres

J, Gonzáles Kesler C, Amils R. 2004. The Tinto River, an extreme acidic environment under

control of iron, as an analog to the Terra Meridiani hematite site of Mars. Planetary and

Space Sci. 52:239-248.

Ferris FG, Beveridge TJ. 1986. Physiochemical roles of soluble metal cations in the outer

membrane of Escherichia coli K-12. Can. J. Microbiol. 32: 594-601.

Ferris FG. 1989. Metallic ion interactions with the outer membrane of gram-negative bacteria.

In: Beveridge TJ, Doyle RJ, editors. Metal Ions and Bacteria. New York, NY: John Wiley &

Sons. P 295-323.

Flint, S.J., Enquist, L.W., Krug, R.M., Racaniello, V.R., and Skalka, A.M. 2000. Principles of

virology: Molecular biology, pathogenesis, and control. ASA Press, Washington, D.C.

Foppen JWA, Okletey S, Schijven JF. 2006. Effect of goethite coating and humic acid on the

transport of bacteriophage PRD1 in columns of saturated sand. J. Contamin. Hydrol.

85:287-301.

Frankel RB, Bazylinski DA. 2003. Biologically Induced Mineralization by Bacteria. In Dove

PM, De Yoreo JJ, Weiner S, editors. Biomineralization Vol 54. Rev Min Geochem.

Washington, DC: Mineralogical Society of America. P 95-114.

Fuhrman JA. 1999. Marine viruses and their biogeochemical and ecological effects. Nature

17 399:541-548.

Gerba CP. 1984. Applied and theoretical aspects of virus adsoption to surfaces. Adv. Appl.

Microbiol. 30:133-168.

Geslin C, Le Romancer M, Gaillard M, Erauso G, Prieur D. 2003. Observation of virus-like

particles in high temperature enrichment cultures from deep-sea hydrothermal vents. Res.

Microbiol. 154:303-307.

Guan H, Schulze-Makuch D, Schaffer S, Pillai SD. 2003. The effect of critical pH on virus fate

and transport in saturated porous medium. Ground Water 41:701-708.

Guixa-Boixereu N, Vaqué D, Gasol JM, Sánchez-Cámara J, Pedrós-Alió C. 2002. Viral

distribution and activity in Antarctic waters. Deep Sea Res. II 49:827-845.

Gutterman B, Raya R, Kutter E. 2005. Basic phage biology. In: Kutter E, Sulakelidze A.,

editors. Bacteriophages: biology and applications. Baton Raton, FL: CRC Press. P 29-66.

Häring M, Rachel R, Peng X, Garrett RA, Prangishvili D. 2005. Viral diversity in hot springs of

Pozzuoli, Italy, and characterization of a unique archaeal virus, Acidianus bottle-shaped

virus, from a new family, the Ampullaviridae. J. Virol. 79:9904-9911.

Hewson I, O’Neil JM, Fuhrman JA, Dennison WC. 2001. Virus-like particle distribution and

abundance in sediments and overlying waters along eutrophication gradients in two

subtropical estuaries. Limnol. Oceanogr. 46:1734-1746.

Jiang S, Steward G, Jellison R, Chu W, Choi S. 2004. Abundance, distribution, and diversity of

viruses in alkaline, hypersaline Mono Lake, California. Microb. Ecol. 47:9-17.

Kutter E, Raya R, Calson K. 2005. Molecular mechanisms of phage infection. In: Kutter E,

Sulakelidze A., editors. Bacteriophages: biology and applications. Baton Raton, FL: CRC

Press. P 165-222.

18 Lindberg AA. 1973. Bacteriophage receptors. Ann. Rev. Microbiol. 27:205-241.

Lisle JT, Priscu JC. 2004. The occurrence of lysogenic bacteria and microbial aggregates in the

lakes of McMurdo Dry Valleys, Antarctica. Microbial. Ecol. 47:427-439.

Loveland JP, Ryan JN, Amy GL, Harvey RW. 1996. The reversibility of virus attachment to

mineral surfaces. Colloids and Surfaces 107:205-221.

Maranger R, Bird DF. 1995. Viral abundance in aquatic systems: a comparison between marine

and fresh waters. Mar. Ecol. Prog. Ser. 121:217-226.

Mera MU, Beveridge TJ. 1993. Mechanism of silicate binding to the bacterial cell wall in

Bacillus subtilis. J. Bacteriol. 175:1936-1945.

Middelboe M, Jørgensen NOG, Kroer N. 1996. Effects of viruses on nutrient turnover and

growth efficiency of noninfected marine bacterioplankton. Appl. Environ. Microbiol.

62:1991-1997.

Middelboe M, Jørgensen NOG. 2006. Viral lysis of bacteria: an important source of dissolved

amino acids and cell wall components. J. Mar. Biol. Ass. UK 86:605-612.

Murray JP, Labond SJ. 1979. Degradation of poliovirus by adsorption on inorganic surfaces.

Appl. Environ. Microbiol. 37:480-486.

Nordstrom DK, Southam G. 1997. Geomicrobiology of sulide mineral oxidation. In: Banfield JF,

Nealson KH, editors. Geomicrobiology: Interactions between Microbes and Minerals Vol.

35. Washington, DC: Mineralogical Society of America. P 361-390.

Oren A, Bratback G, Heldal M. 1997. Occurrence of virus-like particles in the Dead Sea.

Extremophiles 1:143-149.

Ortmann AC, Suttle CA. 2005. High abundances of viruses in a deep-sea hydrothermal vent

system indicates viral mediated microbial mortality. Deep Sea Res. I 52:1515−1527.

19 Pedersen K, Karlsson F. 1995. Investigations of subterranean microorganisms: their

importance for performance assessment of radioactive waste disposal. In: SKB Technical

Report 95-10. Swedish Nuclear Fuel and Waste Management Co., Stockholm. P 222

Pedersen K. 1996. Investigations of subterranean bacteria in deep crystalline bedrock

and their importance for the disposal of nuclear waste. Can. J. Microbiol. 42:382-391.

Pedersen K. 1999. Evidence for a hydrogen-driven, intra-terrestrial biosphere in deep granitic

rock aquifers. In: Bell CR, Brylinksky M, Johnson-Green P, editors. Microbial Biosystems:

New Frontiers, Proceedings of the 8th International Symposium on Microbial Ecology.

Plisson C, White HE, Auzat I, Zafarani A, São-Jose C, Lhuillier S, Tavares P, Orlova EV. 2007.

Structure of bacteriophage SPP1 tail reveals trigger for DNA ejection. The EMBO J 26:3720-

3728.

Rachel R, Bettstetter M, Hedlund BP, Häring M, Kessler A, Stetter KO, Prangishvili D. 2002.

Remarkable morphological diversity of viruses and virus-like particles in hot terrestrial

environments. Arch. Virol. 147:2419-2429.

Ricciarde-Rigault M, Bird DF, Prairie YT. 2000. Changes in sediment viral and bacterial

abundances with hypolimnetic oxygen depletion in a shallow eutrophic Lac Brome (Quebec,

Canada). Can. J. Fish. Aquat. Sci. 56:1284-1290.

Riemann L, Holmfeldt K, Titelman J. 2009. Importance of viral lysis and dissolved DNA for

bacterioplankton activity in a P-limited estuary, Northern Baltic Sea. Microbial Ecol. 57:286-

294.

Ryan JN, Elimelech M, Ard RA, Harvey RW, Johnson PR. 1999. Bacteriophage PRD1 and silica

colloid transport and recovery in an iron oxide-coated sand aquifer. Environ. Sci. Technol.

33:63-73.

20 Sahl JW, Schmidt R, Swanner ED, Mandernack KW, Templeton AS, Kieft TL, Smith RL,

Sanford WE, Callaghan RL, Mitton JB, Spear JR. 2008. Subsurface microbial diversity in

deep-granitic-fracture water in Colorado. Appl. Environ. Microbiol. 74:143-152.

Schulze-Makuch D, Guan H, Pillai SD. 2003. Effects of pH and geological medium on

bacteriophage MS2 transport in a model aquifer. Geomicrobiol. J. 20:73-84.

Stumm W, Morgan JJ. 1996. Aquatic Chemistry: Chemical equilibria and rates in natural waters.

Third Edition. New York, NY: John Wiley & Sons Inc.

Suttle CA. 2005. Viruses of the sea. Nature 437:356-361.

Suttle CA. 2007. Marine viruses- major players in the global ecosystem. Nature Rev. Microbiol.

5:801-812.

Ward TE, Bruhn DF, Shean ML, Watkins CS, Bulmer D, Winston V. 1993. Characterization of

a new bacteriophage which infects bacteria of the genus Acidiphilium. J. Gen. Virol.

74:2419-2425.

Wilhelm SW, Suttle CA. 1999. Viruses and nutrient cycles in the sea. Bioscience 49:781-788.

Zhuang J, Jin Y. 2003. Virus retention and transport through Al-oxide coated sand columns:

effects of ionic strength and composition. J. Contaminant Hydrol. 60:193-209.

21 Chapter 2

This chapter is reprinted with the permission from the publisher Nature Publishing Group and co-authors in: ISME (2008), vol. 20, pp. 571-574. Viruses in granitic groundwater from 69 to

450 m depth of the Äspö hard rock laboratory, Sweden. By: J. E. Kyle, H. S. C. Eydal, F. G.

Ferris, and K. Pedersen.

22 Viruses in Granitic Groundwater from 69 to 450 m Depth of the Äspö

Hard Rock Laboratory, Sweden

Jennifer E. Kyle1*, Hallgerd S. C. Eydal2, F. Grant Ferris1, and Karsten Pedersen2

1Department of Geology, University of Toronto, Earth Sciences Centre, Toronto, Ontario, Canada M5S 3B1

2 Department of Cell and Molecular Biology, Göteborg University, SE 405 30 Göteborg, Sweden

*Corresponding author

Jennifer E. Kyle Email: [email protected] Phone: (416) 978-0661; Fax: (416) 978-3938

Running Title: Viruses in Deep Granitic Groundwater

Subject Category: Microbial ecology and functional diversity in natural habitats

Keywords: Bacteriophage; Biosphere; Groundwater; Prokaryotes; Subsurface; Viruses

23 Abstract

The objectives for this study were to determine if viruses exist in deep granitic groundwater and to analyse their abundance and morphological diversity. Fluorescent microscopy counts on ten groundwater samples ranging from 69 to 450 m depth were in the range of 104 to 106 total number of prokaryotic cells (TNC) mL−1 and 105 to 107 virus-like particles (VLP) mL−1. A good positive correlation of VLP with TNC (r = 0.91, p=0.0003) was found with an average VLP/TNC ratio of 12. Transmission electron microscopy revealed four distinct bacteriophage groups (polyhedral, tailed, filamentous, and pleomorphic) with at least seven phage families of which some are known to be lytic. Our results suggest the presence of viruses in deep granitic groundwater to 450 m depth. If they are active and lytic, they would constitute an important group of predators that could control numbers of microorganisms in the analysed groundwater.

24 Short communication

Prokaryotes colonize the deep subsurface to a depth of at least 3.3 km (Amend and Teske

2005; Lin et al. 2006). Deep intraterrestrial microbial life is investigated to understand the diversity of life of Earth, the evolution and potential origin of life in the deep underground and the tolerances of intraterrestrial life to extreme environmental conditions (Fredrickson and

Balkwill, 2006). Applied aspects, for example the impact of microbial activity on deep intraterrestrial storage of spent nuclear fuel, are also important (Pedersen 2002). To completely understand the ecology of microorganisms and their impact on the surrounding environment, consideration needs to be given to the smallest member of microbial communities, viruses.

Groundwater samples were obtained in November 2006 from ten boreholes along the Äspö HRL tunnel, ranging from 69 to 450 m depth. The samples were analysed for numbers of virus-like particles (VLP), total number of prokaryotic cells (TNC) and chloride. In addition, one sample from each borehole was observed with transmission electron microscopy (TEM) and the viral morphological diversity of the samples was registered. To our knowledge this is the first investigation of viruses in a deep intraterrestrial, fractured hard rock environment.

Äspö HRL is located on the island of Äspö near Oskarshamn, Sweden and comprises a

3.6 km long tunnel that spirals down from the surface to a depth of 460 m in granitic bedrock

(Pedersen, 2001). It is a deep research facility that investigates the geological storage of spent nuclear fuel (Pedersen, 2002). The sampled boreholes were collected under in situ borehole pressure, as described elsewhere (Pedersen, 2001). Samples used for determining abundance and

TEM imaging were collected in four sterile 50-mL polypropylene tubes and immediately preserved with 0.02 µm filtered 37% acid-free formaldehyde to a final concentration of 2%.

Samples were stored at 4 ˚C until further analysis. Three of the samples were used for

25 determining TNC and VLP; they were stained with SYBR Gold, according to the methods of

Nobel and Furhman (1998) and Chen et al. (2001) and analysed using an epifluorescence microscope (Leica DMR HD). At least 300 microbial and 400 viral particles were counted per filter in up to 30 fields, except in borehole KJ0052F03 at 447 m depth where 10 mL of sample per filter resulted in approximately 150 VLP and TNC counted in 30 fields. Each field counted was 0.01 mm2 in size. Transmission electron microscopy (Philips 201 TEM operating at 60 kV) was used to image the viral particles in the fourth sample of every sample set. Samples were filtered through a 0.2 µm syringe filter and centrifuged (RC5B Plus Superspeed centrifuge) at 19

000 rpm for 2 hours. All of the water except for 20−50 µL was removed; 20 µL of sample was then transferred onto formvar- and carbon-coated copper grids for 25−30 minutes, and then stained with 1% uranyl acetate for 60 seconds, after which excess sample was wicked from the grid using filter paper. Grids were stored in the dark until being viewed on the TEM.

A good positive correlation of VLP with TNC (Figure 1a) was found. The VLP/TNC ratios ranged from 1.1 up to 18.0 with an average ratio of 12. The groundwater in the boreholes at 69 m was young, comprising a mixture of groundwater aged from months to years (Banwart et al, 1994), and has been demonstrated to harbour significant microbial activity (Banwart et al,

1996). The VLP abundance and TNC were among highest in these boreholes (Figure 1a), reflecting recent contact with shallow (0−5 m), microbiologically diverse and active groundwater

(Banwart et al, 1996). However, deep groundwater from 300 and 415 m showed similar numbers.

The age and origin of the groundwater surrounding the Äspö HRL tunnel have been studied and found generally to correlate with salinity (Laaksoharju et al, 1999b). Typically, a large part of the groundwater at a depth of 500 m is approximately 7,000 years old; i.e., it went

26 underground at the end of the last Fennoscandian glaciation (Laaksoharju et al, 1999a). The amount of chloride in the investigated Äspö groundwater, which represents the salinity, was not well correlated with depth (Figure 1b), owing to the very heterogeneous character of the aquifers in the rock. The VLP numbers showed a good exponential correlation with chloride (Figure 1b) as did TNC (not shown). There was no correlation between VLP, TNC or chloride with depth.

The inverse exponential relationships between chloride and VLP and chloride and TNC (Figure

1b), may be due to electrostatic phenomena. A high ionic strength decreases the electrostatic double layer which increases the chance that viruses and prokaryotes are trapped in the secondary attraction trough (Marshall, 1976) and the resulting attached virus-microbe ecosystems will not be revealed by groundwater samples. Alternatively, old saline groundwater that stands isolated (Laaksoharju et al, 1999b) may be less favourable for microbial growth and viral activity, compared to more diluted groundwater. Future sampling and analysis of both biofilms and groundwater are required to fully understand the observed decrease of TNC and

VLP in groundwater with high salinity.

Transmission electron microscopy exposed a diverse suite of viral morphologies (Figure

2). In a total of 252 examined viruses, four different morphological groups were identified, including polyhedral, tailed, filamentous, and pleomorphic shapes. At 69 m underground, 12 viral sub-groups were represented (135 observations), compared to only one at 447 m (29 observations). Numbers of tailed viruses (Figure 2a, c, j−l) represented 43 % of the viral morphotypes detected (110 observations), while numbers of polyhedral viruses (Figure 2b, m) represented 31 % of the morphotypes (78 observations) except at 447 m where they represented

100 % (29 observations). Of the tailed viruses, Siphoviridae (Figure 2a, j, k) were the most common, followed by Podoviridae (Figure 2l) and then Myoviridae (Figure 2c). Spherical

27 capsids were more common than helical capsids in all of the tailed viruses. Tail lengths and shapes also varied, along with the presence and/or absence of a base plate. Filaments were not found on any of the viruses examined. The hosts of Siphoviridae and Myoviridae morphotypes are typically found infecting the bacterial hosts (Prangishvili et al, 2006a), although tailed phages have also been commonly found infecting hosts in the archaeal domain, Euryarchaeota

(Prangishvili, 2006b). Filamentous Inoviridae were found at a depth of 69 m (21 observations), where they occurred radiating from a central point and attached to each other along their outer ends by thin filaments (Figure 2d). The pleomorphic viruses in the Äspö groundwater were represented by archaeal types (Figure 2e, f, i), most of which were fusiform archaeal viruses

(Figure 2e, i) (12 observations), and Guttaviridae (Figure 2f) (2 observations). Archaeal virus abundance decreased with depth, as only Fuselloviridae viruses were noted at or below 300 m.

Of the archaeal viruses, Salterprovirus (Figure 2e) were the most abundant archaeal virus noted at 69 m (4 observations). The viral diversity was consequently large in the shallow samples and it decreased somewhat with increasing salinity.

Viruses are dependent on active and growing host microorganisms for their multiplication. The number of VLP has been demonstrated to be significantly related to bacterial turnover in samples from deep Mediterranean sediments (Danovaro et al, 2002), to bacterial activity in sediments from Nivå Bay in Denmark (Middelboe et al, 2003), and to the number of host cells in the Adriatic Sea aquatic system (Corinaldesi et al, 2003). High VLP/TNC ratios of about ten, like those observed here (the average was 12), are consequently indicative of viruses actively infecting microorganisms that also must be metabolically active. This confirms earlier obtained energy source assimilation data (Pedersen and Ekendahl, 1992) and recent ATP analysis data (Eydal and Pedersen, 2007), both of which suggested that the investigated

28 microorganisms were in a state of growth. A predator−prey relationship may be present in deep groundwater that then contains active and growing microorganisms, continuously predated by viruses to observed steady state numbers in the range of 104 to 106 cells mL1 (Pedersen 2001) just as it is in many surface environments (Wiggins and Alexander1985).

Acknowledgements

This research was made possible by generous support from the Swedish Foundation for

International Cooperation in Research Higher Education (STINT), the Swedish Nuclear Fuel and

Waste Management Co. (SKB), and the Swedish Science Research Council (VR). The authors would like to thank the personnel at the Äspö Hard Rock Laboratory for their general support during our sampling and analysis field work. Sara Eriksson and Lotta Hallbeck are acknowledged for valuable comments on the manuscript.

References

Amend JP, Teske A (2005) Expanding frontiers in deep subsurface microbiology. Palaeogeogr.

Palaeoclimatol. Palaeoecol. 219: 131−155.

Banwart S, Gustafsson E, Laaksoharju M, Nilson AC, Tullborg EL, Wallin B (1994) Large-scale

intrusion of shallow water into a granite aquifer. Water Resour. Res. 30: 1747−1763.

Banwart S, Tullborg EL, Pedersen K, Gustafsson E, Laaksoharju M, Nilson AC et al. (1996)

Organic carbon oxidation induced by large-scale shallow water intrusion into a vertical

fracture zone at the Äspö Hard Rock Laboratory (Sweden). J. Contam. Hydrol. 21: 115−125.

Chen F, Lu JR, Binder BJ, Liu YC, Hodson RE (2001) Application of digital image analysis and

flow cytometry to enumerate marine viruses stained with SYBR gold. Appl. Environ.

29 Microbiol. 67: 539−545.

Corinaldesi C, Crevatin E, Del Negro P, Marini M, Russo A, Fonda-Umani S et al. (2003)

Large-scale spatial distribution of virioplankton in the Adriatic Sea: Testing the trophic state

control hypothesis. Appl. Environ. Microbiol. 69: 2664−2673.

Danovaro R, Manini E, Dell'Anno A (2002) Higher abundance of bacteria than of viruses in deep

Mediterranean sediments. Appl. Environ. Microbiol. 68: 1468−1472.

Eydal HSC, Pedersen K (2007) Use of an ATP assay to determine viable microbial biomass in

Fennoscandian Shield groundwater from depths of 3−1000 m. J. Microbiol. Methods

70: 363-373.

Fredrickson, JK, Balkwill, DL (2006) Geomicrobiological processes and biodiversity in the deep

terrestrial subsurface. Geomicrobiol. J. 23:345-356.

Laaksoharju M, Shårman C, Shårman E (1999a) Multivariate mixing and mass balance (M3)

calculations, a new tool for decoding hydrogeochemical information. Appl. Geochem. 14:

861-871.

Laaksoharju M, Tullborg EL, Wikberg P, Wallin B, Smellie J (1999b) Hydrogeochemical

conditions and evolution at Äspö HRL, Sweden. Appl. Geochem. 147: 835−859.

Lin LH, Wang PL, Rumble D, Lippmann-Pipke J, Boice E, Pratt LM. et al. (2006) Long-term

sustainability of a high-energy low-diversity crustal biome. Science 314: 479−482.

Marshall KC (1976) Interfaces in microbial ecology. Harward University Press, Cambridge pp 1-

156.

Middelboe M, Glud RN, Finster K (2003) Distribution of viruses and bacteria in relation to

diagenetic activity in an estuarine sediment. Limno.Oceanogr. 48: 1447−1456.

Noble RT, Fuhrman JA (1998) Use of SYBR Green I for rapid epifluorescence

30 counts of marine viruses and bacteria. Aquat. Microb. Ecol. 14: 113-118.

Pedersen ,K (2001) Diversity and activity of microorganisms in deep igneous rock aquifers of

the Fennoscandian Shield. In: Fredrickson, JK, Fletcher, M (Eds.) Subsurface Microbiology

and Biogeochemistry. Applied and Ecological Microbiology, Wiley-Liss Inc, New York, pp

97−139.

Pedersen, K (2002) Microbial processes in the disposal of high level radioactive waste 500 m

underground in Fennoscandian shield rocks. In: Keith-Roach, MJ, Livens, FR, (Eds.)

Interactions of Microorganisms with Radionuclides. Radioactivity in the environment 2,

Elsevier, Amsterdam, pp 279−311.

Pedersen K, Ekendahl S (1992) Incorporation of CO2 and introduced organic compounds by

bacterial populations in groundwater from the deep crystalline bedrock of the Stripa mine. J.

Gen. Microbiol. 138: 369−376.

Prangishvili D, Forterre P, Garrett RA (2006a) Viruses of the Archaea: a unifying view. Nature

4: 837−848.

Prangishvili D, Garret RA, Koonin EV (2006b) Evolutionary genomics of archaeal viruses:

unique viral genomes in the third domain of life. Virus Res. 117: 52-67.

Wiggins BA, Alexander M (1985) Minimum bacterial density for bacteriophage replication:

implications for significance of bacteriophages in natural ecosystems. Appl. Environ.

Microbiol. 49: 19-23.

31 Figure legends

Figure 1a. The relation between the total number of cells (TNC) and the number of virus like particles (VLP) in groundwater from ten different boreholes distributed along the Äspö hard rock laboratory tunnel at depths from 69 m down to 450 m. Three independent analyses were done for each borehole. Dashed lines show 95% confidence intervals. The least-squares regression line for

VLP versus TNC is shown (10Log(VLP)= 1.30 × 10Log(VLP) – 0.62; r=-0.91, p=0.00001, n=30).

Figure 1b. The relation between the average (n=3) of 10log number of virus like particles (VLP), depth and amount of chloride in groundwater from ten different boreholes distributed along the

Äspö hard rock laboratory tunnel at depths from 69 m down to 450 m. Numbers close to the symbols indicate sample depth. Dashed lines show 95% confidence intervals. The least-squares regression line for 10Log(VLP) versus chloride is shown (Chloride= - 2654 × 10Log(VLP) +

20172; r=-0.90, p=0.0004, n=10).

Figure 2. Transmission electron micrographs of viruses from Äspö hard rock laboratory groundwater. Viral morphotypes found near depths of 69 m (a−h), 294 m (i, j), 415 m (k, l), and

447 m (m) are shown: a, Siphoviridae (B1); b, polyhedral virus with base plate; c, Myoviridae with base plate; and d, Inoviridae connected by filaments around the outer ends (arrows). Two polyhedral viruses are also shown: e, Salterprovirus; f, Guttaviridae; g, polyhedral virus with spike-like protrusions; h, polyhedral virus (STIV-like); i, Fuselloviridae with twinned tail; j,

Siphoviridae (B1) with curved tail; k, Siphoviridae (B1) with straight tail; and m, polyhedral virus. Scale bar is 125 nm, except in a and d where it is 250 nm.

32

Figure 1.

33

Figure 2.

34 Chapter 3

This chapter is reprinted with the permission from the publisher Taylor and Francis and co- authors in: Geomicrobiology Journal (2008), vol. 25, pp. 338-345. Virus mineralization at low pH in the Rio Tinto, Spain. By: J. E. Kyle, K. Pedersen, and F. G. Ferris.

35 Virus Mineralization at Low pH in the Rio Tinto, Spain

Jennifer E. Kyle1*, Karsten Pedersen2, and F. Grant Ferris1

1Department of Geology, University of Toronto, Earth Sciences Centre, Toronto, Ontario, Canada, M5S 3B1

2Göteborg University, Department of Cell and Molecular Biology, Microbiology Section, Box 462, SE-405 30 Göteborg, Sweden

*Corresponding author email: [email protected]

36 Abstract

Water and sediment samples were collected from the Rio Tinto in south-western Spain to assess (1) the presence and diversity of viruses in an acid mine drainage system and (2) determine if relationships occur between geochemical parameters and viral abundance.

Epifluroescence microscopy and transmission electron microscopy revealed that viruses are not only present, but geochemical evidence and multivariate statistical analyses suggest that viruses in the Rio Tinto participate in mineralization processes. Viral capsids and tails occurred with iron-bearing minerals sorbed to their surfaces, at times with mineralization so extensive that differentiating between viral and inorganic particles using microscopy was difficult. Moreover, a strong inverse relationship between viral abundance and jarosite saturation state (Pearson correlation coefficient r = -0.71) was observed implying that viruses were removed from suspension owing to ongoing mineral precipitation (i.e., decreasing number of viruses with increasing rates of mineral precipitation, as inferred from saturation state). Viral-mineral interactions may additionally impact virus-host relationships as a weak correlation was found between viral and prokaryotic abundance, a relationship that is usually found to be highly correlated. Viral abundance and pH were strongly correlated (pearson correlation coefficient r =

0.94) indicating viral sensitivity to low pH conditions.

37 Introduction

Investigations concerning the presence and diversity of bacteriophages in acidic environments have focused primarily on terrestrial hot springs (Rice et al. 2001; Rachel et al.

2002; Häring et al. 2005; Ortmann et al. 2006). Fewer studies have taken place in acid mine drainage (AMD) systems; however, recent genomic studies of prokaryotic species from AMD suggest that lysogenic prophages may be involved in lateral gene transfer (Allen et al. 2007). In addition, Ward et al. (1993) characterized an AMD temperate phage (φAc1) that was shown to be more stable at near neutral pH than in the acidic environment from which host bacterial strains were isolated. Given phage instability at low pH (Ward et al. 1993), the authors suggested lysogeny would be a common reproductive strategy, as phages exposed to acidic conditions after host cell lysis would have to find a new host cell quickly before becoming inactivated. In contrast, a lipid-containing virus (NS11) was shown to not only tolerate low pH environments, but to have an optimal replication cycle at a pH of 3.5 (Sakaki and Oshima 1976).

These general observations are consistent with viral population analyses of acidic, high temperature, terrestrial hot springs where phages commonly exhibit lysogenic replication cycles, and most of the lysogenic phages infect archaeal species (Rice et al. 2001;Ortmann et al. 2006).

The low pH (1.5 to 3.1) and high dissolved metal concentrations of the Rio Tinto in southwestern Spain stand as a classic example for AMD environments (Figure 1; Gonzáles-Toril et al. 2003). There have been numerous studies on the geochemistry and microbiology of the

Rio Tinto (López-Archilla and Amils 2001; Gonzáles-Toril et al. 2003; Ferris et al. 2004), but the presence and diversity of bacteriophages has not been addressed explicitly. More recently, the Rio Tinto has become a focal point of considerable interest for astrobiology as it may provide insight into life that may have once survived on Mars (Fernández-Remolar et al. 2005; Knoll et

38 al. 2005). This is because mineralogical data gathered from the NASA Rover missions found iron-bearing sulfate minerals on the surface of Mars that are similar to those forming in the Rio

Tinto drainage system (Buckby et al., 2003; Fernández-Remolar et al., 2004). The presence of these minerals has made the Rio Tinto a possible Earth analog to past processes that occurred on

Mars, particularly in the Meridiani Planum region (Fernández-Remolar et al. 2005).

Microbial identification in the Rio Tinto conducted by Gonzáles-Toril et al. (2003) found that 80% of prokaryotic diversity is composed of Acidithiobacillus ferrooxidans, Leptospirillum ferrooxidans, and Acidiphilium. Archaeal species, such as Ferroplasma acidiphilum, are also found but in low abundance (Gonzáles-Toril et al. 2003). Interestingly, eukaryotic microorganisms dominate, comprising over 65% of the river’s biomass (López-Archilla and

Amils 2001). To date, viral analyses have not been conducted in the Rio Tinto. Total bacterial counts conducted before this current study ranges from 105 – 107 cells/mL (Gonzáles-Toril et al.,

2003), which is above the lower limits of viral replication (in non-acidic environments; Wiggins and Alexander, 1985). In an attempt to further characterize the biotic diversity within the Rio

Tinto, the presence of viruses and their abundance and diversity was studied in this investigation.

In addition, geochemical analyses and multivariate statistics were applied to ascertain the extent and nature of correlations between virus populations and AMD geochemistry.

Methods and Materials

Site Description and Sample Collection

The drainage basin for the Rio Tinto is within the Iberian Pyrite Belt, which is the largest volancogenic massive sulfide deposit in the world (LeBlanc et al. 2000). Water and sediment

39 samples were collected in 50 mL and 60mL sterile falcon tube and nalgene bottles, respectively, in June of 2006 from the source waters of the Rio Tinto and downstream along the river to the town of Niebla (Fig. 1), approximately 65 km from the source. Duplicates of each sample were taken, one of which was preserved in a final concentration of 2 % (v/v) glutaraldehyde.

Additional water samples were also collected and filtered through a 0.2 µm syringe filter

(cellulose acetate filter, Sarturius) for chemical analysis. Samples were stored in a cooler during field work and then in a fridge until analysis. Temperature, pH, and Eh measurements were taken directly in the water at each sample site in the field. The temperature of the water was measured using a pIONeer 10 portable pH meter equipped with a pHC5977 combination pH electrode (pH range 0–14, ± 0.5 at zero; temperature range –10 to 110°C, ± 0.3°C) (Radiometer,

Stockholm, Sweden). Redox was measured using the same pH meter, but equipped with a

MC3187Pt combined platinum electrode with an Ag/AgCl reference system, range –2000 to

2000 mV (± 0.01% of reading; calibrated using hexacyanoferrate II/III redox buffer)

(Radiometer).

Prokaryotic and viral abundances

Samples (2 mL) collected for prokaryotic and viral counts (particles that were viral-sized) were stained with SYBR Green I and examined using an epifluroescence microscope (Nikon

Microphot-FXA). The staining procedure followed that of Noble and Fuhrman (1998). Briefly,

1 to 3 mL of sample was filtered onto a 0.02 µm Anopore membrane filter. The filter was then placed onto a drop of 2.5 % SYBR Green I and stored in the dark. After 15 minutes, excess stain was removed from the filter and set aside to dry for an additional 15 minutes. A drop of antifade solution was placed onto a cover slip which was then inverted over the filter. Samples were

40 viewed under blue excitation.

Transmission Electron Microscopy

Glutaraldehyde preserved water samples were filtered using sterile 0.2 µm syringe filter, and centrifuged (Sorvall RC5B Plus) at 19 000 rpm for 2 hours. Since a pellet was not noticeable in the samples, most of the water was extracted, except for 20-40 µL of sample at the bottom of the tube. The remaining sample was mixed by gently pipetting the solution up and down. 20 µL of sample was then transferred onto 300 mesh formvar, carbon-coated (providing extra support and strength under the high accelerating voltage of the electron beam) copper grids for 25-30 minutes. Samples were subsequently stained with 10 µL of 1% uranyl acetate for 60 seconds. Excess liquid was then wicked off the grid with filter paper before viewing with a

Philips 201 transmission electron microscope (TEM).

Sediment samples were collected and prepared in a similar manner as the water samples except that prior to filtration, samples were incubated in 0.01 M sodium pyrophosphate

(Na4P2O7) for 1 hour to disrupt ionic bonds between the viruses and mineral surfaces (Maranger and Bird 1996). Samples were then shaken for 45 seconds, filtered through 0.2 µm pore sizes syringe filters, and processed as described above.

Analytical TEM was conducted using a Philips CM 10 TEM equipped with a Sapphire energy dispersive spectrometer (EDS). Elemental analysis was conducted at 80 kV using a 200 nm spot size on virus capsids and inorganic particles associated with the viruses. Samples were angled at 25˚ towards the detector to improve analyses.

X-Ray Diffraction

41 Bulk mineralogy of the sediment samples was determined using an X-Ray Diffractometer

(XRD; Philips PW1830) with a Cu Ka radiation target source. Sediment samples were dried at

60ºC, and ground into a fine powder using a mortar and pestle. Specimens were then mounted onto powder holders or fixed on glass slides using an acetone slurry depending on the amount of sample available. Samples were scanned at 1 s per step at a continuous scan rate of 0.02º2Θ/s.

Geochemical and Statistical Calculations

Chemical constituents of water samples measured using inductively coupled plasma atomic emission spectroscopy (ICP-AES; Perkin Elmer Optima 5300DV), and an ICP- high resolution sector field mass spectrometer (Thermo Finningan ELEMENT). The measured concentrations were entered into PHREEQC (USGS version 2.13.2) to obtain speciation and saturation indices of potential mineral precipitates within the Rio Tinto. Principle component analysis (STATISTIC 6.1) was then used to evaluate relationships between the chemical and microbial constituents, the physiochemical properties, and the calculated saturation indices of the collected water samples.

Results

Microbial Abundance and Physiochemical correlations

The observed virus-like particle (VLP) abundance (103-106 VLP/mL) was found to be similar to prokaryotic abundance (104-106 cells/mL; Table 1), although not significantly correlated with each other (r = 0.38; Table 2); however, a strong positive correlation was found between VLPs and pH (r = 0.94), and a strong inverse relationship between VLPs and Eh (r = -

0.89). The source water of the Rio Tinto contained greatest abundance of VLPs and also had the

42 highest pH (3.6) and lowest Eh (343 mV) value of the water samples collected. Prokaryotes, on the other hand, show a weak relationship with pH (r = 0.53), and no significant relationship with

Eh (r = -0.25). Among dissolved chemical substances, dissolved iron and sulfate were the most abundance species with concentrations up to 120 mM and 295 mM, respectively (Table 3).

Viral Diversity

The viral diversity was minimal when compared to other aquatic systems (i.e. Rachel et al. 2002). Polyhedral viruses dominated composing approximately half of the total viral diversity (66 polyhedral viruses out of a total of 127 viral observations). Tailed phages comprised most of the remaining diversity (based on morphological observations) as

Siphoviridae were the most common of the tailed phages (Fig. 2a) followed by Myoviridae (Fig.

2b, 3). One TEM micrograph of a Myoviridae revealed the intricate detail of the tail sheath, with a glimpse inside the tip of the tail that is surrounded by short spikes (i.e., base-plate pins; Fig. 3).

Archaeal morphotypes, such as Fuselloviridae and Guttaviridae, were rarely noted.

Viral-Inorganic Particle Association

Within two of the samples (Berrocal and from the area in the town of Nerva), phages were found attached to inorganic particles (Fig. 4). EDS showed that these particles were dominated by iron-bearing phases (Fe Kα, Kβ peaks at 6.4 and 7.0 KeV, respectively ) with trace amounts of sulphur and potassium in some samples suggesting that an Fe-hydroxy sulfate phase

(i.e. jarosite) was present. Detrital clays were also present as trace amounts of aluminium and silica were found in most samples. Phages were noted sorbed to these iron-bearing mineral phases (Fig. 4 a-d), at times linking two isolated phages together (Fig. 4a). The iron-bearing

43 phases were also found sorbed to the tails of some phages (Fig. 4 d, e); however, at times it was difficult to distinguish if the mineralization was occurring on the capsid or the tail and/or neck region due to the position of the iron-bearing minerals (Fig. 4 c, d). The interaction between the phages and minerals was frequently so extensive it was challenging to differentiate the phages from the iron minerals (Fig. 4f).

Geochemical calculations performed using PHREEQC indicated that sulfate, iron-, and manganese-bearing minerals were oversaturated at most sites (Table 4). The implication is that jarosite minerals and some iron oxides actively precipitate from solution. Bulk mineral analysis of the sediment samples by XRD confirmed that the iron bearing minerals were dominated by iron hydroxyl sulfates, including jarosite and mikasaite. Iron oxides, such as 2-line ferrihydrite and goethite were also found.

Principle component analysis using the calculated saturation indices revealed that jarosite was the only mineral with a strong VLP correlation (r = -0.66 to -0.71). As the saturation index for jarosite increased (i.e. the greater potential to precipitate from solution) the viral abundance decreased. This viral-mineral relationship was also noted with Al(OH)SO4 (r = -0.53). In addition to jarosite (r = -0.72 to - 0.75), prokaryotes exhibited a strong inverse relationship with iron (oxyhydr)oxides (r = -0.69 to -0.72), and Al(OH)SO4 (r = -0.72).

Discussion

This is the first study to document the morphological diversity of viruses within the Rio

Tinto drainage basin. Viral morphotypes provide some clues into potential viral hosts and viral replications cycles (Suttle 2007), which is important to understanding microbial community dynamics. Polyhedral viruses, for example, are lytic viruses that could belong to any of the three

44 domains of life as it is a common morphotype so it is difficult to differentiate between potential hosts based on viral morphology alone. The tailed viruses, also lytic, typically represent bacteriophages that infect bacterial hosts, although Siphoviridae and Myoviridae have occasionally been found infecting Euryarchaeota species (Prangishvili et al. 2006). Many

Siphoviridae phages are lysogenic until an environmental stress prompts an induction of a lytic event (Suttle 2005, 2007). In this regard, the predominance of Siphoviridae phages within the

Rio Tinto AMD system indicates that the prokaryotic hosts may have been stressed upon sample collection. Lastly, pleomorphic phages are found infecting archaeal species (i.e. fusiform morphotype Fuselloviridae). Pleomorphic phages were rarely noted within the Rio Tinto possibly due to (a) phage sensitivity to storage conditions, and/or (b) the lack of potential hosts for phage replication as archaeal species represent a small fraction of the prokaryotic population within the Rio Tinto (Gonzáles-Toril et al. 2003).

Viral abundances in neutral pH environments are known to correlate with prokaryotic abundance such that viral abundance is typically found to be one to two orders of magnitude greater that the prokaryote abundance (Maranger and Bird 1996; Oren et al. 1997; Jiang et al.

2004; Ortmann and Suttle 2005). This was not the case in the acidic Rio Tinto. The weak correlation between VLP and prokaryotes within the Rio Tinto may be due to a number of factors including (1) viral degradation, (2) host abundance, (3) physiochemical condition of the water, (4) lysogenic life cycle, and/or (5) the large percentage of eukaryotic organisms within the

Rio Tinto.

Previous studies have reported that viral degradation may be rapid after sample collection with up to 75% of the VLP lost after 17 days when samples are fixed with formaldehyde and stored at 4˚C (Wen et al. 2004; Helton et al. 2006). In this context, VLP abundances reported

45 here might be underestimated considering the length of time between sample collection and analysis (ca. 2 weeks transit time). At the same time, an inadequate abundance of bacterial hosts may explain the low viral-prokaryote relationship. Wiggins and Alexander (1985) emphasize that the host population density must be large enough to sustain a vital abundance of viral replication.

As viral attachment is initially governed by chance collisions, higher concentrations of host cells and viral particles will result in a greater probability of attachment and viral replication. Viruses that are free (i.e. not particle associated) within the Rio Tinto waters would have a limited amount of time to attach and infect a host cell as they are subjected to low pH conditions that may be detrimental to the survival of the viral particles (Ward et al. 1993).

Viruses that are not free (i.e. particle associated which is common in low pH environments) are unlikely to infect host cells as they are removed from solution. In addition, lysogeny has been suggested as a common life cycle for viruses in acidic environments as the viral particles would only be subjected to acidic conditions when the temperate phage is induced into a lytic cycle (i.e. induction may occur when the host cell is under environmental stress). The VLP-prokaryotic abundance relationship reported here may also be biased as this relationship is under the assumption that all of the VLPs within the system are from prokaryotic hosts. In this regard, it is interesting to note that López-Archilla and Amils (2001) reported that more than 65% of the biomass in the Rio Tinto is from eukaryotic organisms (i.e. algae, fungi, and protists). Since these organisms are susceptible to viral infections, this means that eukaryotic viruses are likely included in our total VLP fluorescent microscopic counts.

Our TEM results and statistical analyses show that viruses undergo mineralization, and may be removed from suspension through jarosite precipitation. In addition, bacteriophages

46 attach to the same cellular components on the bacterial surface (Lindberg 1973; Yu et al. 1981;

Heller 1992) that promote the mineralization of bacterial cells (Ferris and Beveridge 1986; Ferris

1989; Fein et al. 1997). The question arises that if a host cell and viruses are undergoing mineralization (e.g., bacterial mineralization has been shown to occur in the Rio Tinto by Ferris et al. 2004), are receptor sites on the virus and host cell wall blocked preventing infections? If so, would this inhibition of viral infection cause decrease VLP abundance? Certainly, the negative bacterial-mineral correlations from PCA analyses (i.e. jarosite, iron (oxyhydr)oxides,

Al(OH)SO4) are indicative of bacterial mineralization implying that viral receptor sites might be blocked or at least partially obstructed to infective viruses.

For tailed phages, the receptor binding sites on the phages themselves are generally located along the tail fibers (Heller, 1992). If this binding region is blocked due to mineralization, the virus would be rendered inactive (i.e. unable to infect host cells). However, it is possible that some tailed phages, especially Siphoviridae, may be at an advantage when compared to other viruses within AMD systems, owing to their flexible tail. A thin, supple appendage would be more likely to penetrate between mineral particles at the host cell surface enabling viral infections. Depending on the degree of bacterial mineralization (i.e. thickness and/or continuance of mineralization at the cell surface), phages without tails (polyhedral phages), short tails (Podoviridae), and/or contractile tails (Myoviridae) may be at a selective disadvantage for phage attachment and/or DNA injection due to the potential inability to access cellular receptor sites (i.e., virus diameters may not fit between mineralized areas on bacterial cell surfaces preventing viral attachment).

Bacteriophages and other viruses are known to interact with minerals, especially iron oxyhydroxides (You et al. 2005; Templeton et al. 2006). Past studies examining the attachment

47 and release of phages to iron oxides have suggested that the viral capsid, which is composed of proteins containing reactive carboxyl and amine groups (Gerba 1984; Daughney et al. 2004), would be the primary site of phage-mineral contact; however, the TEM images captured in this study confirm that the tail serves additionally as a site of phage-mineral interaction. As phage tails are composed of proteins, this is not unexpected. Still, whether viral-particle associations in the Rio Tinto system result simply from binding of iron-bearing mineral phases to viral capsids

(and tails) or heterogeneous nucleation and precipitation in association with viral particles remains to be determined.

Currently, there is no known evidence of viruses in the geologic record, which is attributed to their small size and lack of a unique chemical or isotopic signature. This study, however, provides strong evidence of viral involvement in mineralization events, although the extent of involvement is currently unknown. In another study, Daughney et al. (2004) showed that when marine phage PWH3a-P1 (a Myoviridae) was exposed to increased concentration of dissolved iron in solution, individual viruses became difficult to distinguish as the viral particles and newly formed iron oxides (i.e. lepidocrocite and goethite) tended to clump together. These observations suggest that the formation of viral particle-mineral aggregates is not only common in nature, but additionally that a potential exists for mineralized viruses to be incorporated into sediments, and in time, possibly into the geological record.

There are terraces of ironstone outcrops along the Rio Tinto which are composed of three units that transform mineralogically with age from poorly ordered goethite with minor amounts of hydronium jarosite in the youngest terrace to hematite in the oldest terrace (Fernández-

Remolar et al. 2003, 2005). Mineralogical studies conducted on modern precipitates and sediments noted that nano-sized ferric-oxides, including nano-sized goethite, are found within

48 freshly precipitated sediments (Fernández-Remolar et al. 2005). It is these nano-sized ferric- oxides and hydroonium jarosite that occur in the youngest terrace. Fossil evidence of bacteria, algae, fungi, and plant materials have been preserved by thin coatings of goethite in this terrace, which is estimated to be to have formed less than 11 000 years old ago (Fernández-Remolar et al.

2005). Goethite becomes more ordered within the intermediate and older terraces as the presence of sulfate salts disappears. Hematite dominates in the oldest terrace which is believed to have formed around 2 million years ago (Fernández-Remolar et al. 2005). Given the interactions between viral particles and iron-bearing minerals, viruses were probably incorporated into these sediments, but whether they are identifiable particularly after aging and subsequent mineral transformation reactions remains to be investigated.

The action of viruses from infection to lysis contributes substantially to the dynamics of community structure. Within any thriving aquatic ecosystem, viruses are a significant member of the community as viral lysis of host cells liberate particulate and dissolved organic matter, releasing bioavailable carbon and other essential nutrients, such as nitrogen and phosphorous, into the ecosystem (Fuhrman 1999; Wilhelm and Suttle 1999; Middelboe and Jørgensen 2006).

This release of nutrients may benefit non-infected hosts, as well as other organisms (Middelboe et al. 1996, Proctor and Fuhrman 1990). In regards to the possibility of ancient life on Mars, apart from being able to withstand low pH and a high concentration of dissolved metals, a major challenge suggested by Knolls et al. (2005) for Martian microbial communities would be a source of bioavailable nitrogen and phosphorous. An interesting thought is that perhaps viruses and their intrinsic capacity to release essential nutrients back into ecosystems provided a modicum of sustenance for ancient life on Mars.

49 Acknowledgements

This research was made possible by support from the Swedish Foundation for

International Cooperation in Research, Ontario Graduate Scholarship, and National Sciences and

Engineering Research Council of Canada (NSERC). We would like to thank Robert Harris for his assistance on the TEM at the University of Guelph, and the two anonymous reviewers for their comments leading to the improvement of this manuscript.

References

Allen EE, Tyson GW, Whitaker RJ, Detter JC, Richardson PM, Banfield JF. 2007. Genome

dynamics on a natural archaeal population. PNAS 104:1883-1888.

Buckby T, Black S, Coleman ML, Hodson ME. 2003. Fe-sulphate-rich evaporative mineral

precipitates from the Rio Tinto, southwest Spain. Mineralogical Magazine 67:263-278.

Daughney CJ, Châtellier X, Chan A, Kenward P, Fortin D, Suttle CA, Fowle DA. 2004.

Adsorption and precipitation of iron from seawater on a marine bacteriophage (PWH3A-P1).

Marine Chemistry 91:101-115.

Fein JB, Daughney CJ, Yee N, Davis TA. 1997. A chemical equilibrium model for metal

adsorption onto bacterial surfaces. Geochim Cosmochim Acta 61:3319-3328.

Fernández-Remolar DC, Gómez-Elvira J, Gómez F, Sebastian E, Martíin J, Manfredi JA, Torres

J, Gonzáles Kesler C, Amils R. 2004. The Tinto River, an extreme acidic environment under

control of iron, as an analog to the Terra Meridiani hematite site of Mars. Planetary and

Space Sci 52:239-248.

Fernández-Remolar DC, Rodriguez N, Gómez F. 2003. Geological record of an acidic

environmental driven by iron hydrochemistry: The Tinto River system. J Geophys Res

50 108:16-1 – 16-15.

Fernández-Remolar DC, Morris RV, Gruener JE, Amils R., Knoll AH. 2005. The Rio Tinto

Basin, Spain: Mineralogy, sedimentary geobiology, and implications for interpretation of

outcrop rocks at Meridiani Planum, Mars. Earth and Planetary Science Letters 240:149-167.

Ferris FG. 1989. Metallic ion interactions with the outer membrane of gram-negative bacteria.

In: Beveridge TJ, Doyle RJ, editors. Metal Ions and Bacteria. New York, NY: John Wiley &

Sons. P 295-323.

Ferris FG, Beveridge TJ. 1986. Physiochemical roles of soluble metal cations in the outer

membrane of Escherichia coli K-12. Can J Microbiol 32:594-601.

Ferris FG, Hallbeck L, Kennedy CB, Pedersen K. 2004. Geochemistry of acidic Rio Tinto

headwaters and role of bacteria in solid phase metal partitioning. Chem Geol 212:291-300.

Fuhrman JA. 1999. Marine viruses and their biogeochemical and ecological effects. Nature

399:541-548.

Gerba CP. 1984. Applied and theoretical aspects of virus adsoption to surfaces. Adv Appl

Microbiol 30:133-168.

Gonzáles-Toril E, Llobet-Brossa E, Casamayor EO, Amann R, Amils R. 2003. Microbial

ecology of an extreme acidic environment, the Tinto River. Appl Environ Microbio 69:4853-

4865.

Häring M, Rachel R, Peng X, Garrett RA, Prangishvili D. 2005. Viral diversity in hot springs of

Pozzuoli, Italy, and characterization of a unique archaeal virus, Acidianus bottle-shaped

virus, from a new family, the Ampullaviridae. J Virol 79:9904-9911.

Heller KJ. 1992. Molecular interaction between bacteriophage and the gram-negative cell

envelope. Arch Micriol 158:235-248.

51 Helton RR, Liu L, Wommack KE. 2006. Assessment of factors influencing direct enumeration of

viruses within estuarine sediments. Appl Environ Microbiol 72:4747-4774.

Jiang S, Steward G, Jellison R, Chu W, Choi S. 2004. Abundance, distribution, and diversity of

viruses in alkaline, hypersaline Mono Lake, California. Microb Ecol 47:9-17.

Kargel JS. 2004. Proof for water, hints of life? Science 306:1689-1691.

Knoll AH, Carr M, Clark B, Des Marais DJ, Farmer JD, Fischer WW, Grotzinger JP, McLennan

SM, Malin M, Schröder C, Squyres S, Tosca NJ, Wdowiak T. 2005. An astrobiological

perspective on Meridiani Planum. Earth and Planetary Sciences Letters 240:179-189.

Leblanc M, Morales JA, Borrego J, Elbaz-Poulichet F. 2000. 4500-year-old mining pollution in

southwestern Spain: long-term implications for modern mining pollution. Econ Geol 95:655-

662.

Lindberg AA. 1973. Bacteriophage receptors. Ann Rev Microbiol 27:205-241.

López-Archilla AI, Martin L, Amils R. 2001. Microbial community composition and ecology of

an acidic aquatic environment: The Tinto River, Spain. Microb Ecol 41:20-35.

Maranger R, Bird DF. 1996. High concentrations of viruses in the sediments of Lac Gilbert,

Québec. Microb Ecol 31:141-151.

Middelboe M, Jørgensen NOG, Kroer N. 1996. Effects of viruses on nutrient turnover and

growth efficiency of noninfected marine bacterioplankton. Appl Environ Microb 62:1991-

1997.

Middelboe M, Jørgensen NOG. 2006. Viral lysis of bacteria: an important source of dissolved

amino acids and cell wall components. J Mar Biol Ass UK 86:605-612.

Noble RT, Fuhrman JA. 1998. Use of SYBR Green I for rapid epifluroescence counts of marine

viruses and bacteria. Aquat Mcirob Ecol 14:113-118.

52 Oren A, Bratback G, Heldal M. 1997. Occurrence of virus-like particles in the Dead Sea.

Extremophiles 1:143-149.

Ortmann AC, Suttle CA. 2005. High abundances of viruses in a deep-sea hydrothermal vent

system indicates viral mediated microbial mortality. Deep Sea Res I 52:1515−1527.

Ortmann AC, Wiedenheft B, Douglas T, Young M. 2006. Hot crenarchaeal viruses reveal deep

evolutionary connections. Nature Rev Microb 4:520-528.

Prangishvili D, Garret RA, Koonin EV. 2006. Evolutionary genomics of archaeal viruses: unique

viral genomes in the third domain of life. Virus Res 117:52-67.

Proctor LM, Furhman JA. 1990. Viral mortality of marine bacteria and cyanobacteria. Nature

343:60-62.

Rachel R, Bettstetter M, Hedlund BP, Häring M, Kessler A, Stetter KO, Prangishvili D. 2002.

Remarkable morphological diversity of viruses and virus-like particles in hot terrestrial

environments. Arch Virol 147:2419-2429.

Rice G, Stedman K, Snyder J, Wiedenheft B, Willits D, Brumfield S, McDermott T, Young MJ.

2001. Viruses from extreme thermal environments. PNAS 98:13341-13345.

Sakaki Y, Oshima T. 1976. A new lipid-containing phage infecting acidophilic thermophilic

bacteria. Virol 75:256-259.

Suttle CA. 2005. Viruses of the sea. Nature 437:356-361.

Suttle CA. 2007. Marine viruses- major players in the global ecosystem. Nature Rev Microbiol

5:801-812.

Templeton MR, Andrews RC, Hofmann R. 2006. Impact of iron particles in groundwater on the

UV inactivation of bacteriophages MS2 and T4. J Appl Microbiol 101:732-741.

Ward TE, Bruhn DF, Shean ML, Watkins CS, Bulmer D, Winston V. 1993. Characterization of

53 a new bacteriophage which infects bacteria of the genus Acidiphilium. J Gen Virol 74:2419-

2425.

Wen K, Ortmann AC, Suttle CA. 2004 Accurate estimation of viral abundance by

epifluroescence microscopy. Appl Environ Microb 70:3862-3867.

Wiggins BA, Alexander M. 1985. Minimum bacterial density for bacteriophage replication:

implications for significance of bacteriophages in natural ecosystems. Appl Environ Microb

49:19-23.

Wilhelm SW, Suttle CA. 1999. Viruses and nutrient cycles in the sea. Bioscience 49:781-788.

You Y, Han J, Chiu PC, Jin Y. 2005. Removal and inactivation of waterborne viruses using

zerovalent iron. Environ. Sci. Technol 39:9263-9269.

Yu F, Yamada H, Mizushima S. 1981. Role of lipopolysaccharide in the receptor function for

bacteriophage TuIb in Escherichia coli. J Bacteriol 148:712-715.

Zhuang J, Jin Y. 2003. Virus retention and transport through Al-oxide coated sand columns:

effects of ionic strength and composition. J Contaminant Hydrol 60:193-209.

54 Table 1. Physiochemical characteristics and microbial abundance of sample sites downstream from source water.

Sample Distance Temp pH Eh Abundance (mL 104) VPL/Prokaryote site Downstream (ºC) (mV) Ratio (km) Prokaryotic VLP Source 0 18 3.6 343 25.1 102.0 4.1 Ravine a 0.1 23 2.2 632 19.0 62.8 3.3 Ravine b 0.2 29 2.4 527 6.5 2.3 2.4 Train Stop 4.6 24 2.6 474 16.0 24.8 1.5 Berrocal 21.0 24 2.4 608 5.8 1.1 0.2 Valverde 53.8 24 2.6 611 18.0 0.9 0.05 Niebla 62.3 24 2.7 579 44.2 19.6 0.4

Table 2. Correlation indices of Rio Tinto chemical constituents with microbial abundance and physiochemical characteristics.

pH Eh VLP Prokaryote VLP/Prokaryote Ionic Strength pH 1.00 -0.86 0.94 0.53 0.65 -0.69 Eh -0.86 1.00 -0.89 -0.25 -0.70 0.28 VLP 0.94 -0.89 1.00 0.38 0.86 -0.45 Prokaryote 0.53 -0.25 0.38 1.00 0.04 -0.72 VLP/Prok. 0.65 -0.70 0.86 0.04 1.00 -0.01 Al -0.61 0.26 -0.34 -0.63 0.15 0.95 Ca -0.47 0.03 -0.44 -0.36 -0.42 0.60 Cu -0.13 -0.15 -0.09 -0.12 -0.08 0.26 Fe -0.64 0.33 -0.38 -0.69 0.10 0.95 Mg -0.63 0.17 -0.42 -0.68 -0.06 0.97 Mn -0.47 -0.03 -0.36 -0.53 -0.20 0.76 Na -0.38 0.29 -0.53 0.13 -0.74 0.00 Pb 0.05 0.23 -0.18 0.33 -0.48 -0.62 SO4 -0.68 0.34 -0.41 -0.71 0.07 0.98 Zn -0.17 -0.11 -0.14 -0.17 -0.13 0.28 Ionic strength -0.69 0.28 -0.45 -0.72 -0.01 1.00

55 Table 3. Major dissolved ion concentration from ICP analysis of 0.22 µm filtered Rio Tinto water samples.

Site Concentration (mM) + + 2+ 2+ 3+ 3+ 4+ 2+ 2- Na K Ca Mg Al Fe Mn Cu Zn SO4 Source 0.3 3.1 0.2 0.4 0.3 0.1 0.01 0.001 0.02 1.4 Ravine a 0.4 n.d. 1.6 35.8 64.6 120.1 0.8 0.1 0.1 294.7 Ravine b 1.6 n.d. 6.8 36.7 29.2 79.9 1.9 0.1 0.3 187.1 Train Stop 2.5 n.d. 6.3 39.1 47.8 43.2 2.2 6.3 4.5 150.8 Berrocal 2.4 1.5 3.6 17.4 17.8 24.0 0.9 2.2 2.0 90.4 Valverde 1.4 1.6 1.0 3.3 3.3 2.6 0.2 0.4 0.3 15.4 Niebla 2.1 2.1 2.2 2.9 2.4 1.0 0.1 0.3 0.2 12.2 *n.d. = not detected

Table 4. Saturation index values from geochemical modelling of Rio Tinto water samples.

Site Saturation Index K- H- Na- Ferrihydrite Lepidocrocite Goethite Magnetite Al(OH)SO4 Jarosite Jarosite Jarosite Source 7.7 5.0 4.5 -0.06 3.5 4.0 1.5 -0.2 Ravine a - 11.5 9.8 -0.2 3.3 4.2 9.0 1.2 Ravine b - 12.3 11.7 0.8 4.3 5.2 12.0 1.4 Train - 12.1 11.7 0.8 4.3 5.2 11.5 1.7 Stop Berrocal 13.2 11.9 11.4 0.6 4.2 5.0 5.1 1.2 Valverde 9.3 8.3 7.3 -0.8 2.7 3.6 0.7 0.1 Niebla 6.7 7.1 6.3 -1.2 2.3 3.2 -0.4 0.03

56 Figure Legends

Figure 1. Map of Rio Tinto sampling sites () and nearby towns ().

Figure 2. Transmission electron micrograph of common phage morphotypes found in the Rio

Tinto. (a) Siphoviridae and (b) radiating cluster of Myoviridae.

Figure 3. High-resolution TEM micrograph of a Myoviridae phage. Note the textured pattern on the tail, spikes on the tail tip (black arrows), and what appears to be the inner side of the tail near the tip (white arrow).

Figure 4. Transmission electron micrographs of RT-066 with inorganic, iron-bearing mineral phases attached to the phages. The inorganic material is attached to (a) the capsid of Myoviridae

(top) and Siphoviridae (bottom), connecting the two separated phages; (b) capsid of

Siphoviridae; (c) capsid of an icosahedral (?) phage; (d) capsids and/or tails; and (e) tail of

Siphoviridae (B2) phage. At times, a cluster of phages with attached inorganics result in a nano- sized mix of phages and iron-dominated particles making the phages difficult to distinguish from the minerals.

57

Figure 1.

Figure 2.

58

Figure 3.

Figure 4.

59 Chapter 4

This chapter has been submitted to Applied and Environmental Microbiology. Geochemistry of

Virus-Prokaryote Interactions in Freshwater and Acid Mine Drainage Environments, Ontario,

Canada. By: Jennifer E. Kyle and F. Grant Ferris.

60 Geochemistry of Virus - Prokaryote Interactions in Freshwater and Acid

Mine Drainage Environments, Ontario, Canada

Jennifer E. Kyle*, and F. Grant Ferris

Department of Geology, University of Toronto, Earth Sciences Centre, Toronto, Ontario,

Canada, M5S 3B1

Corresponding author contact information:

Jennifer Kyle, Department of Geology, University of Toronto, Earth Sciences Centre, Toronto,

Ontario, Canada, M5S 3B1

Phone: 1-416-978-0661

Fax: 1-416-978-3938

Email: [email protected]

Running Title: Geochemistry of Virus-Prokaryote Interactions

61 Abstract

An extensive water sample survey was conducted in southern Ontario, Canada across a variety of freshwater and acid mine drainage (AMD) systems in order to further understand the role of viruses in aquatic environments. Samples were evaluated in terms of virus and prokaryote abundances, physiochemistry (i.e. pH, temperature, nitrate, phosphate, dissolved organic carbon), and geochemistry (i.e. dissolved elemental concentrations). Backwards step- wise multiple regression analysis found that VLP (virus-like particle) abundance, phosphate, pH, sulfate, and magnesium are predictors of prokaryotic abundance with the model describing 90 % of the variability in the data (R2 = 0.90). VLP abundance was found to be the strongest predictor of prokaryotic abundance suggesting that viruses exert strong control over prokaryotic abundance. The only statistically significant (p < 0.05) predictors of VLP abundance were mineral saturation indices of goethite (α-FeO(OH); R2 = 0.78) although moderate Pearson component analysis correlations (r) were noted with ferrihydrite, jarosite, and pyrolusite.

Considering the pHIEP of the minerals, this indicates that as pH increases viruses detach from goethite. This relationship along with an inverse relationship, using Spearman rank order correlations (rs) between jarosite ((H3O)Fe3(SO4)2(OH)6) saturation indices and VLP abundance

(rs = -0.33), indicates that viral inactivation through mineral attachment may be a contributor to the moderate relationship between VLP and prokaryotic abundance (rs = 0.45). AMD environments (low pH, high Eh, high dissolved iron) are correlated with low VLP abundances although no relationship was noted with prokaryotic abundance suggesting oxidizing AMD environments are detrimental to viruses and/or that viruses are being partially removed possibly through mineral attachment.

62 Introduction

Given that viruses are the most numerous biological entity on Earth ( >1030 viruses,

Suttle 2007) and they are seemingly ubiquitous, the role viruses play in aquatic environments requires close examination to understand the ecology and biogeochemistry of aquatic microbial communities. Some of these roles have been investigated such that viruses are known to be major players in prokaryote mortality, which can benefit microbial communities through the release of nutrients, such as P and N, and dissolved organic matter (DOM) (10, 27, 38, 51). The action of viruses can also create niches for less dominant organisms through the lysis of more productive, dominant members of the community (44). In addition, viruses mediate bacterial evolution through genetic exchange and through structural changes in bacterial surfaces to prevent infection. There is some evidence that biological parameters (i.e. bacterial abundance and levels of chlorophyll-a) play a dominant role in predicting viral abundance in freshwater environments (22, 25, 26, 36); however, this is not always the case. Clasen et al. (2008) examined 16 lakes in Wisconsin, USA, and British Columbia, and northwestern Ontario, Canada and found that only 39% viral abundance variability was explained by bacterial, cyanobacterial, and chlorophyll-a abundance.

Besides biological parameters, nutrients sources (i.e. DOM, phosphate), and some physiochemical factors (i.e. temperature, oxygen) have also shown strong relationships with viral abundance. For example, eutrophic environments been shown to have greater viral abundances than oligotrophic (11,17, 49), as do oxic verses anoxic environments (36). Viral abundances have been shown to be seasonally dependent with greater viral abundance in warmer months (3,

52). In order to gain a greater understanding of virus-prokaryote interactions in aquatic systems, a detailed biogeochemical survey was undertaken across a wide variety of habitats (i.e.

63 freshwater lakes, rivers, wetlands, acid mine drainage (AMD)). Among these habitats, few investigations have focused on rivers (41), wetlands (18), or AMD (1) with even fewer considering more than one type of environment at a time.

Methods and Materials

Site description and sample collection

Surface water samples were collected from acid mine drainage (AMD) sites in the summer of 2007 and 2008 and from circum-neutral pH environments in the fall of 2008 throughout southern Ontario, Canada. The bedrock in the region is representative of three geological regimes (Figure 1): (1) the Sudbury Igneous Complex containing abundant metal sulfide ore deposits where the AMD samples were collected, (2) the western St. Lawrence

Lowlands dominated by carbonate rocks and shale (~ 440-470 million years old), and (3) the

Grenville Province of the Canadian Shield (~0.9-1.6 billion years ago) dominated by metamorphic rocks such as gneiss and schist. Individual sample sites were selected based on obtaining samples from various types of aquatic habitats (i.e. lakes, rivers, tributaries, wetlands) and across regions with different bedrock geology to obtain as broad a range of biogeochemical conditions as possible. A total of 47 samples were collected; 5 from AMD, 18 from lakes (4 from Lake Ontario), 14 from rivers, and 10 from wetlands. In situ measurement of temperature, pH and Eh were conducted at each sample site. Water samples were collected and prepared for analysis as described below. Samples collected for direct epifluorescent counting and transmission electron microscopic imaging were preserved with a final concentration of 2.5 %

(v/v) aqueous glutaraldehyde. Additional samples were filtered using a 0.2 µm syringe filter into

60 mL polypropylene bottles and amber EPA certified 40 mL glass vials to measure chemical

64 constituents and dissolved organic carbon (DOC), respectively, so that there was no headspace.

Water was also collected in 500 mL polypropylene bottles with no head space and stored at 4ºC to determine dissolved abundances of iron, sulfate, nitrate, phosphate, and alkalinity, and in

50mL falcon tubes for turbidity analysis. All samples were stored at 4ºC in the dark until analysis.

Viral and prokaryotic abundance and viral imaging

Preserved samples (0.3 to 0.8 mL) were stained with SYBR Green I within 8 hours (to one week for the AMD samples) of collection according to Noble and Fuhrman (1998). Briefly, samples were filtered onto 0.02 µm Anopore membrane filter and then filters were placed onto a

100 µL drop of 2.5 % SYBR Green I and stored in the dark. After 15 minutes, excess stain was removed from the filters and then the filters were stored an additional 20 minutes in the dark to dry. A drop of antifade solution was placed on cover slips, which were then inverted onto the samples. Samples were stored at -20ºC until viewed on a epifluorescence microscope (Nikon

Microphot-FXA) under blue excitation light.

A representative suite of samples were imaged by transmission electron microscopy

(TEM) to determine the morphology of virus particles in the samples. This involved filtration of preserved samples through 0.2 µm syringe filters and centrifugation of the filtrate (Sorvall RC5B

Plus, rotor SS-34) at 43150 x g for 2 hours at 10°C. Most of the supernatant was discarded, except for approximately 50 µL of sample at the bottom of the tube which was mixed by gently pipetting the solution up and down. 20 µL of sample was transferred onto 300 mesh formvar, carbon-coated (providing extra support and strength under the high accelerating voltage of the electron beam) copper grids for 20 minutes. Samples were subsequently stained with 10 µL of

65 1% uranyl acetate (0.02 µm filtered) for 45 seconds. Excess liquid was then wicked off the grid with filter paper before viewing with a Philips 201TEM.

Aqueous Chemistry

Analyses were conducted within 10 hours of sample collection using HACH spectrophotometric assays for dissolved concentrations of total iron, ferrous iron, sulfate, nitrate, and phosphate. For these analyses, samples were filtered using 0.2µm syringe filters into 25 mL glass cuvettes for analysis. Alkalinity was measured within 20 hours of sample collection by titration (HACH alkalinity titrator). For inductively coupled plasma atomic optical emission spectroscopy (ICP-AOES Perkin Elmer Optima 7300DV), filtered samples were acidified (final pH ~3.0) with trace metal grade nitric acid within 8 hours of sample collection. Samples for measurement of DOC concentrations were shipped to G.G. Hatch Isotope Laboratories at the

University of Ottawa and analyzed on an OI Analytical Aurora Model 1030W TOC Analyzer.

Turbidity was measured using a HACH 2100N Turbidity Meter equipped with 25 mL glass cuvettes.

Statistical and Geochemical Data Analysis

Multiple regression analyses were conducted using the general linear models module of

STATISTICA 6.1 to determine predictors of prokaryotic and viral abundance. Regression analyses were performed in a backwards step-wise fashion to identify and remove independent variables that were not statistically significant (based p-values > 0.05 and low beta (β) coefficients (i.e. no to little contribution to the model). Once the independent predictor variables in the regression model were all statistically significant (p < 0.05), the model was examined for

66 multicollinearity (i.e. two or more variables with low tolerance and high variance inflation factors). If multicollinearity was noted, one variable was dropped from the model and then the model was reevaluated. Variables that did not affect the strength of the model were removed.

When the model exhibited a strong correlation (R2 > 0.90) that was statistically significant (p <

0.05), with no evidence of multicollinearity and normally distributed residuals, the model was accepted. One exception to the above was made as backwards step-wise removal of mineral saturation indices (R2 > 0.75) resulted in a single predictor as strong signs of multicollinearity existed between variables.

Nonparametric Spearman rank order correlations (rs) were used to evaluate viruses- prokaryote relationships and aqueous geochemistry owing to the non-gaussian distribution of the geochemical data, which is commonly encountered in natural systems.

To evaluate potential mineral-virus-prokaryote relationships, mineral saturation indices were calculated for each sample by entering the corresponding geochemical data into PHREEQC

(version 2.13.2, USGS). A mineral saturation index (SI) is defined as SI = log (IAP/Keq), where

IAP is the ion activity production and Keq is the equilibrium constant. Minerals with near- saturated (slightly negative SI) and supersaturated values (positive SI) that have the potential to precipitate in the studied environments were used in principle component analysis (PCA) to evaluate correlations with corresponding virus-prokaryote data. In this case, PCA was used instead of Spearman rank order correlations because the virus-prokaryote data and mineral saturation indices were normally distributed, and PCA offers stronger results than Spearman correlations.

Results

67 Aqueous chemistry

All values and averages for microbial and geochemical data collected in this study are reported in Table 1. Most of the geochemical parameters had concentrations that spanned 2 to 3 orders of magnitude. The AMD samples notably had the lowest pH and highest elemental concentrations with the exception of K and Na, which were the second highest. The highest K and Na, and lowest Eh values were found in a small urban pond in Toronto that was covered with aquatic plants with a strong SOx smell.

Samples collected from the St. Lawrence Lowlands had higher concentrations of all elements with the exception of Fe, and higher pH values when compared to the silicate- dominated Grenville Province. No trend was noted for temperature, pH, turbidity or iron in the non-AMD samples. However, lakes generally contained less sulfate, nitrate, and DOC, and rivers generally had higher Eh values and lower phosphate then the other non-AMD sites.

Viral and prokaryotic abundance and geochemical relationships

Viral abundance was found to be at least one order of magnitude greater than the prokaryotic abundance (n = 47) with the exception of one sample from AMD (Table 1). A number of different morphological types of viruses were identified by TEM. Tailed (especially

Myoviridae and Siphoviridae) and polyhedral morphotypes were the most common, with the occasional Fuselloviridae morphotype (Figure 2).

Spearman rank order correlations noted statistically significant (p < 0.05) relationships with moderate strength correlations for a number of parameters (Table 2). A positive correlation between prokaryotes and VLP abundance (rs = 0.49), and negative correlation between prokaryotes and VPR (viral - prokaryotic ratio; rs = -0.42) was found. Additional noteworthy

68 correlations revealed that at low pH and high Eh (i.e. acid mine drainage environment) there is a decrease in VLP abundance (rs = 0.58 and rs = -0.43, respectively) and VPR (rs = 0.59 and rs = -

0.35, respectively). VLP and virus-prokaryote ratio (VPR) were also noted to be negatively correlated with total dissolved iron (rs = -0.49). No significant relationships were noted between viral abundance and temperature or potential nutrient sources (i.e. DOC, phosphate, and nitrate).

A weak relationship was noted between VPR and temperature (rs = -0.38). No significant or strong geochemical relationship was noted with prokaryote abundance, with the exception of turbidity (rs = 0.31).

Predictors of prokaryotic abundance

Multiple regression analysis revealed that viruses were the most influential predictor of prokaryotic abundance (Table 3). Additional predictors included pH, sulfate, phosphate, and magnesium. The strength of the model was very strong with 90 % of the variability explained

(R2 = 0.90, p < 0.007). All relationships were positively correlated with the exception of magnesium, which is negatively correlated. Despite a strong correlation between sulfate and magnesium (rs = 0.82), multicollinearity was not noted within the model. Iron was also noted as a contributor to prokaryotic abundance but because of strong multicollinearity with sulfate, iron was removed from the model (as sulfate behaved more conservatively across the range of pH sampled, 2.5-9.0).

Viral-Mineral correlations

Mineral saturation indices were the only strong contributors of viral abundance. Simple regression analysis suggested the apparent saturation index (SI) of goethite (α-FeO(OH)) was a

69 good predictor of viral abundance (R2 = 0.78; p < 0.001; Table 4). Mineral SI values had moderate relationships with VLP abundance and VPR (goethite, ferrihydrite (Fe(OH)3), jarosite

((H3O)Fe3(SO4)2(OH)6), and pyrolusite (MnO2); Table 5) and a weak relationship between prokaryotic abundance and jarosite. Transmission electron microscopy (Figure 3) revealed

VLPs attached to inorganic material found within an AMD sample. Many of the mineral saturation indices in the studied environments are strongly correlated with pH (Table 5).

Discussion

Relationships between viruses, prokaryotes, and geochemical variables

Viral abundance is typically greater than prokaryotic abundance in marine environments

(28, 43, 53, 54) and the same trend has been shown in freshwater environments (21, 26, 36, 41).

Our study agrees with the general trend of greater viral abundance compared to prokaryotes. The mean VPR values reported in this study (x = 14, ranging from 0.9 to 50) is consistent with the mean values reported in other studies within Canada; VPRx = 22.5 (26) and VPRx = 2.9 (6).

Multiple regression analysis revealed that 90% of the variability in prokaryotic abundance can be explained by five variables, with the most influential predictor being viral abundance. This suggests that viruses exhibit great control over prokaryotic abundance. Viruses are known to be major contributors in prokaryote mortality, as are grazing protozoa (11, 36, 48).

Multiple reports have discussed viral-induced prokaryote mortality (i.e. 43, 46) and the attendant increase in prokaryote diversity (29, 45) and input of nutrients (10, 27, 38, 43); however, direct studies linking viral abundance as predictors of prokaryotic abundance is limited. Our results stress the important role of viruses in regulating the prokaryotic abundance in freshwater aquatic systems and AMD environments. Specifically, it is hypothesized that viruses will allow the total

70 prokaryotic population to reach a particular level before viral-induced lysis reduces the population to a lower density. Maranger and Bird (1995) suggested a similar theory in which viruses were implicated in maintaining lower bacterial population densities by lysis of the more active members, reducing the competition for limiting nutrients (i.e., phosphorous). As the total prokaryotic population in an ecosystem is comprised of many bacterial species, with only a few species being dominant, Thingstad and Lignell’s (1997) “kill the winner” theory may be applicable if the dominant population is killed through viral lysis to create niches for less dominant organisms of the community.

Thingstad’s (2000) model of hierarchical top-down control of prokaryote diversity implied that viral lysis of a particular host was a critical variable in controlling diversity, after limiting nutrient availability and size-selective predation. In our investigation, phosphate was found to be a factor in regulating prokaryotic abundance, although to a lesser extent than viral abundance (based on beta coefficients). These observations suggest, that viral abundance has a strong influence on prokaryotic abundance, perhaps even more so than common nutrients (i.e. phosphate, nitrate, DOC). In fact, no strong prokaryotic or viral relationships were noted with potential nutrients, as has been reported previously (2, 17, 42). Given the high concentrations of nitrate and DOC within our systems, prokaryotic growth and resulting viral production seemed to be weakly controlled only by phosphate, which is known to be a common limiting nutrient in freshwater systems.

In addition to phosphate, pH, sulfate, and magnesium were noted to be predictors of prokaryotic abundance, and all of which, with the exception of magnesium, are known to affect bacterial growth. Bacterial species are sensitive to the pH range in which they grow, and sulfate is commonly used in anaerobic bacterial metabolism as a terminal electron acceptors. In

71 addition, sulfate is produced by bacteria oxidation of sulfide minerals. These processes are common in wetland and AMD environments, respectively. Moreover, sulfur compounds are essential nutrients for the synthesis of amino acids and coenzyme A. The predictive role magnesium plays in prokaryotic abundance is less obvious, but it is known to an important constituent of bacterial cell membranes and serves as a cofactor for many enzymes (33).

Although not a predictor, turbidity was weakly associated with prokaryotic abundance (rs

= 0.31). As turbidity is not correlated with DOC, it is likely that prokaryotes may be associated with suspended inorganic materials, possibly acting as sites for microbial growth.

The impact of grazers on prokaryotic mortality within our study is not known but the moderately positively relationship between virus and prokaryote abundance (rs = 0.45) raises the possibility that some of the viruses belong to hosts other than prokaryotes, such as phytoplankton

(6, 23) which could be prokaryotic predators. A moderate virus-prokaryote relationship could also be a result of high incidences of lysogeny in some ecosystems, and/or viral inactivation through mineral attachment (Fig. 3). Cochran and Paul (1998) found high levels of inducible prophages year round except from November to February, suggesting that prophage induction is greater in warmer months. This could be due to either (i) greater sunlight (i.e. UV) exposure (5) which can cause up to 5% loss in phage viability per hour in surface waters (55), and/or (ii) the higher surface water temperature in the warmer months promoting bacterial growth leading to greater phage production (5). The samples for our study were collected during fall months

(October to early November) when water temperatures and UV radiation (i.e. later sunrise and earlier sunset) are declining, which may result in increased incidences of lysogeny through the reduction in strong inducing agents (i.e. UV and temperature).

The virus and prokaryote abundance relationship with each other and with VPR (Table 3)

72 suggests the possibility that the burst size (number of viral particles produced from viral induced lysis of one host cell) is increasing as the host density decreases within our freshwater systems.

In order to maintain viral survival and the replication of more viral particles would be required, as the chances of encountering a host cell would decrease under low host density conditions (13,

19, 32). Experimentally, the minimum host density for viral replication reported to be as low as

102 cells/mL (19) and as high at 104 cells/mL (50); however, Kokjohn et al. (1991) states that chance of virus-host interactions may not only be depended on the host density, but also the motility of the host cell as motile cells would be able to cover greater distances increasing the probability of encountering a viral particle.

The VPR is also significantly correlated with temperature (rs = -0.38) suggesting that either viral abundance decreases or prokaryotic abundance increases as temperature increases.

Both circumstances could be occurring, as temperature is a strong growth parameter for prokaryotes but if the increase in temperature is a result of greater sunlight then the UV radiation could be detrimental to free viruses.

Viral-Mineral correlations

Acid mine drainage environments have a measurable effect on viral abundance. Our study, along with another conducted in the Rio Tinto, Spain (20), found that these harsh environments (low pH, highly oxidizing, and high concentration of dissolved iron) are characterized by low viral abundance although prokaryotic abundance remains unaffected.

Limited work as been conducted on viruses from acidic environments, most of which has occurred in terrestrial hot springs (4, 15, 31, 37), and less conducted in AMD environments (1,

20, 47). One possible explanation for the low viral abundance under acidic and strongly

73 oxidizing conditions is offered by Ward et al. (1993). The authors studied a temperate phage,

φAc1, isolated from an AMD, in which the phage was not stable under the conditions it was found, but was stable closer to near neutral pH. The authors suggested that lysogeny would be a common replication cycle under the harsh geochemical conditions associated with AMD, which would be detrimental to free viral particles after short periods of exposure.

Another explanation of low viral abundance in acidic environments is that the viruses are being removed from solution through attachment to actively precipitating iron-bearing minerals.

Within aquatic environments, virus transport has been found to be restricted through (i) inactivation or loss of infective capability (i.e. UV radiation), (ii) irreversible attachment to mineral surfaces, and (iii) reversible attachment to mineral surfaces (24). Commonly, pH strongly influences virus adsorption with lower pH values favoring adsorbed viruses (12). The isoelectric point (IEP) of both the virus and inorganic surface can also impact the interaction

(12). Specifically, minerals with higher a pHIEP, such as iron oxides (i.e. pHIEP of 7-8) are more likely to have a net positive surface charge in most natural environments (12) and will tend to be a better adsorbent of viruses then minerals with low pHIEP, such as quartz (pHIEP of 2.5) (9, 12,

16, 24, 39, 56). Typically, viral surfaces possess net negative charges in most natural environments (pH 6-8) leading to electrostatic attractive forces with mineral surfaces (9, 14, 24,

40). However, hydrophobic and van der Waals forces can also come into play when both surfaces are negatively charged. The former can becomes important at higher pH values and the latter can occur at higher ionic strengths and pH values due to the compression of the diffuse double layer (12). Within this system, attractive electrostatic forces are believed to be responsible for the removal of VLPs at low pH. Jarosite has a very low pHIEP (Christopher

Weisener, personal communication) enabling the surface to be negatively charged at lower pH

74 values. Many of the viruses studied to date have pHIEP of greater than 4 ((i.e. MS2 has an pHIEP of 3.9; PRD1, pHIEP of 4.2; QB pHIEP of 5.3; ϕX174 pHIEP of 6.6, (9)); published values may vary depending on the composition of the phage suspension) causing the surface to be positively charged below this value. Attractive electrostatic interactions may exist between jarosite and viruses within the system leading to the removal of viruses through mineral attachment. This jarosite-virus interaction has been reported, in one other study (20). The authors found a much stronger inverse relationship between VLP abundance and jarosite saturation (r = -0.71) compared to our study (r = -0.33).

In addition to removal through electrostatic forces, virus may be becoming entrained during the precipitation of nanoparticulate mineral phases (8, 20). This behavior was observed in

AMD samples collected for the present study (Figure 3) and in another investigation by Kyle et al. (2008) where VLP-mineral attachment was clearly evident in transmission electron micrographs. Although, the precise mineralogy of the nanoparticulate inorganic phase(s) is not known, hydrous ferric oxides and ferric hydroxysulfate minerals commonly precipitate as poorly ordered nanoparticulate solids before evolving to more crystalline morphotypes such as goethite and jarosite (i.e., Ostwald ripening).

In addition to evidence of virus-mineral interactions occurring in low pH environments, linear regression analysis found that 78% of the variability in viral abundance is explained by the apparent saturation indices of goethite (i.e., α-FeO(OH)), with additional positive relationships with ferrihydrite (i.e., Fe(OH)3), and pyrolusite (i.e., MnO2). Experiments conducted by

Loveland et al. (1996) using phage PRD1, quartz (pHIEP 2.5) and iron-coated quartz surfaces

(pHIEP approximately 5.1) found the attachment edges for 50% of virus attachment to be 2.5-3.5 pH units above the pHIEP of the mineral (attachment edge for PRD1 on quartz to be 6, and for

75 PDR1 on iron-coated quartz to be 7.5). As the pH increased above the attachment edge, detachment increases dramatically as both surfaces become increasingly more negative. Once the pH reached 8-9, complete detachment occurred. Loveland et al. (1996) results may explain what is occurring within our system. A majority of the samples collected within southern Ontario have a pH value of 7.5 or greater, which may possibly account for the influence of goethite on VLP abundance and the positive correlations between VLP abundance and the hydrous iron oxides

(goethite and ferrihydrite). Manganese oxides have been shown to have a pHIEP of 4-4.5 (35), which if Loveland et al. (1996) results are applied, pyrolusite may have an attachment edge anywhere from 6.5-8.

Geochemical variables, as well as biological parameters, are important in understanding viral and prokaryotic abundances in the natural environment. When common nutrients are not limiting (i.e. nitrate, DOC, and possibly phosphate), viral abundance was shown to have the strongest influence on prokaryotic abundance, more so than geochemical variables that promote growth (i.e., pH, phosphate, sulfate, and magnesium). Viral abundance, on the other hand, was influenced by minerals that are able to irreversibly (i.e., jarosite) and reversibly (i.e., goethite) sorb viral particles, and the geochemical conditions that promote mineral precipitation. Viral inactivation through mineral attachment could be one reason for the low viral abundance associated within AMD environments, although the harsh conditions alone could be detrimental to the viruses. The moderate relationship noted between VLP and prokaryotic abundance (rs =

0.49) may be due to these viral-mineral interactions but may also be due to viruses belonging to hosts other than prokaryotes, and/or increase lysogeny, which is known to be common in cooler months.

76 Acknowledgement

This work was supported by a National Science and Engineering Research Council of

Canada Discovery Grant (FGF) and Postgraduate Scholarship (JEK), as well as a Geological

Society of America Student Research Grant (JEK).

We would like to thank Xstrata Canada, and in particular Joe Fyfe and Robin Armstrong for their assistance in collecting AMD samples. Also, Wendy Abdi and Patricia Wickham for analyzing the DOC samples, and Dan Mathers at Analyst, University of Toronto for analyzing the AMD samples collected in 2008 on the ICP-AOES.

References

1. Allen, E. E., G. W. Tyson, R. J. Whitaker, J. C. Detter, P. M. Richardson, and J. F.

Anesio, A. M., C. Hollas, W. Granéli, and J. Laybourn-Parry. 2004. Influence of

humic substances on bacterial and viral dynamics in freshwaters. Appl. Environ.

Microbiol. 70:4848-4854.

2. Auguet, J. C., H. Montanié, D. Delmas, H. J. Hartmann, and V. Huet. 2005. Dynamic

of virioplankton abundance and its environmental control in the Charente Estuary

(France). Microb. Ecol. 50:337-349.

3. Auguet, J. C., H. Montanié, H. J. Hartmann, P. Lebaron, E. O. Casamayor, P.

Catala, and D. Delmas. 2009. Potential effect of freshwater virus on the structure and

activity of bacterial communities in the Marennes-Oléron Bay (France). Microb. Aquatic

Sys . 57:295-306.

4. Bize, A., X. Peng, M. Prokofeva, K. MacLellan, S. Lucas, P. Forterre, R. A. Garrett,

E. A. Bonch-Osmolovskaya, and D. Prangishvili. 2008. Viruses in acidic geothermal

77 environment of the Kamchatka Peninsula. Res. Microbiol. 159:358-366.

5. Brüssow, H., and E. Kutter. 2005. Phage Ecology, p. 129-163. In K. Kutter and A.

Sulakvelidze (eds.), Bacteriophages: biology and applications. CRC Press, Boca Raton,

Florida.

6. Clasen, J. L., S. M. Brigden, J. P. Payet, and C. A. Suttle. 2008. Evidence that viral

abundance across oceans and lakes is driven by different biological factors. Freshwater

Biology. 53:1090-1100.

7. Cochran, P. K., and J. H. Paul. 1998. Seasonal abundance of lysogenic bacteria in a

subtropical estuary. Appl. Environ. Microbiol. 64:2308-2312.

8. Daughney, C., X. Chateller, A. Chan, P. Kenward, D. Fortin, C. A. Suttle, and D. A.

Fowle. 2004. Adsorption and precipitation of iron from seawater on a marine

bacteriophage (PWH3A-P1). Marine Chemistry. 91:101-115.

9. Dowd, S. E., S. D. Pillai, S. Wang, and M. Y. Corapcioglu. 1998. Delineating the

specific influence of virus isoelectric point and size on virus adsorption and transport

through sandy soils. Appl. Environ. Microbiol. 64:405-410.

10. Fuhrman, J. A. 1999. Marine viruses and their biogeochemical and ecological effects.

Nature 399:541-548.

11. Fuhrman, J. A. and R. T. Noble. 1995. Viruses and protists cause similar bacterial

mortality in coastal seawater. Limnol Oceanog. 40:1236-1242.

12. Gerba, C. P. 1984. Applied and theoretical aspects of virus adsoption to surfaces. Adv.

Appl. Microbiol. 30:133-168.

13. Gons, H. J., H. L. Hoogveld, S. G. H. Simis, and M. Tijdens. 2006. Dynamic modeling

of viral impact on cyanobacterial populations in shallow lakes: implications of burst size.

78 J. Mar. Biol. Ass. U.K. 86:537-542.

14. Guan, H., D. Schulze-Makuch, S. Schaffer, and S. D. Pillai. 2003. The effect of

critical pH on virus fate and transport in saturated porous medium. Ground Water.

41:701-708.

15. Häring M., R. Rachel, X. Peng, R.A. Garrett, and D. Prangishvili. 2005. Viral

diversity in hot springs of Pozzuoli, Italy, and characterization of a unique archaeal virus,

Acidianus bottle-shaped virus, from a new family, the Ampullaviridae. J. Virol. 79:9904-

9911.

16. Harvey, R. W., and J. N. Ryan. 2004. Use of PRD1 bacteriophage in groundwater viral

transport, inactivation, and attachment studies. FEMS Microbiol. Ecol. 49:3-16.

17. Hewson, I., J. M. O’Neil, J. A. Fuhrman, and W. C. Dennison. 2001. Virus-like

particle distribution and abundance in sediments and overlying waters along

eutrophication gradients in two subtropical estuaries. Limnol and Oceanogr. 46:1734-

1746.

18. Jackson, E.F. and C. R. Jackson. 2008. Viruses in wetland ecosystems. Freshwater

Biology. 53:1214–1227.

19. Kokjohn, T. A., G. S. Sayler, and R. V. Miller. 1991. Attachment and replication of

Pseudomonas aeruginosa bacteriophages under conditions simulating aquatic

environments. J. General Micrbiol. 137:661-666.

20. Kyle, J. E., K. Pedersen, and F. G. Ferris. 2008. Virus mineralization at low pH in the

Rio Tinto, Spain. Geomicrobiol. 25:338-345.

21. Laybourn-Parry, J., W. A. Marshall, and N. J. Madan. 2007. Viral dynamics and

patterns of lysogeny in saline Antarctic lakes. Polar Biol. 30:351-358.

79 22. Lisle, J. T., and J. C. Priscu. 2004. The occurrence of lysogenic bacteria and microbial

aggregates in the lakes of McMurdo Dry Valleys, Antarctica. Microbial Ecol. 47:427-

439.

23. López-Archilla , A. I., L. Martin, and R. Amils. 2001. Microbial community

composition and ecology of an acidic aquatic environment: The Tinto River, Spain.

Microbial Ecol. 41:20-35.

24. Loveland, J. P., J. N. Ryan, G. L. Amy, and R. W. Harvey. 1996. The reversibility of

virus attachment to mineral surfaces. Colloids and Surfaces. 107:205-221.

25. Lymer, D., E. S. Lindström, and K. Vrede. 2008. Variable importance of viral-induced

bacterial mortality along gradients of trophic status and humic content in lakes.

Freshwater Biology 53:1101-1113.

26. Maranger, R., and D. F. Bird. 1995. Viral abundance in aquatic systems: a comparison

between marine and fresh waters. Mar. Ecol. Prog. Ser. 121:217-226.

27. Middelboe, M., and N. O. G. Jørgensen. 2006. Viral lysis of bacteria: an important

source of dissolved amino acids and cell wall components. J. Mar. Biol. Ass. UK.

86:605-612.

28. Middelboe, M., T. G. Nielsen, and P. K. Bjørnsen. 2002. Viral and bacterial production

in the North Water: in situ measurements, batch-culture experiments and characterization

and distribution of a virus-host system. Deep-Sea Res. II. 49:5063-5079.

29. Miki, T., T. Nakazawa, T. Yokokawa, and T. Nagata. 2008. Functional consequences

of viral impacts on bacterial communities: a food-web model analysis. Freshwater

Biology. 53:1142-1153.

30. Noble R.T., and J. A. Fuhrman. 1998. Use of SYBR Green I for rapid epifluroescence

80 counts of marine viruses and bacteria. Aquat. Microb. Ecol. 14:113-118.

31. Ortmann, A. C., B. Wiedenheft, T. Douglas, and M. Young. 2006. Hot crenarchaeal

viruses reveal deep evolutionary connections. Nature Rev. Microb. 4:520-528

32. Parada, V., G. J. Herndl, and M. G. Weinbauer. 2006. Viral burst size of

heterotrophic prokaryotes in aquatic systems. J. Mar. Biol. Ass. U.K. 86:613-621.

33. Prescott, L. M., J. P. Harley, and D.A. Klein. 2002. Microbiology, 5th edition.

McGraw-Hill Companies, Inc., New York, N.Y.

34. Ram, A. S. P., D. Boucher, T. Sime-Ngando, D. Debroas, and J. C. Romagoux. 2005.

Phage bacteriolysis, protistan bacterivory potential, and bacterial production in a

freshwater reservoir: coupling with temperature. Microbial Ecol. 50:64-72.

35. Rao, G. S. 1959. Adsorption of Amine on Manganese Dioxide. M.S. Thesis.

Massachusetts Institute of Technology, Boston.

36. Ricciarde-Rigault, M., D. F. Bird, and Y. T. Prairie. 2000. Changes in sediment viral

and bacterial abundances with hypolimnetic oxygen depletion in a shallow eutrophic Lac

Brome (Quebec, Canada). Can. J. Fish. Aquat. Sci. 56:1284-1290.

37. Rice, G., K. Stedman, J. Snyder, B. Wiedenheft, D. Willits, S. Brumfield, T.

McDermott, and M. J. Young. 2001. Viruses from extreme thermal environments.

PNAS 98:13341-13345.

38. Riemann, L., K. Holmfeldt, and J. Titelman. 2009. Importance of viral lysis and

dissolved DNA for bacterioplankton activity in a P-limited estuary, Northern Baltic Sea.

Microbial Ecol. 57:286-294.

39. Ryan, J. N., M. Elimelech, R. A. Ard, R. W. Harvey, and P. R. Johnson. 1999.

Bacteriophage PRD1 and silica colloid transport and recovery in an iron oxide-coated

81 sand aquifer. Environ. Sci. Technol. 33:63-73.

40. Schulze-Makuch, D., H. Guan, and S. D. Pillai. 2003. Effects of pH and geological

medium on bacteriophage MS2 transport in a model aquifer. Geomicrobiol. J. 20:73-84.

41. Slovácková, H., and B. Marsálek. 2008. Virioplankton and microbial comities in two

Czech rivers (Svrakta and Morava River). Aquat. Sci. 70:282-291.

42. Stopar, D., A. Cerne, M Zigman, M. Poljsak-Prigatelj, and T. Turk. 2003. Viral

abundance and a high proportion of lysogenic suggests that viruses are important

members of the microbial community in the Gulf of Trieste. Microbial Ecol. 47:1-8.

43. Suttle, C. A. 2007. Marine viruses- major players in the global ecosystem. Nature Rev.

Microbiol. 5:801-812.

44. Thingstad, T. F. 2000. Elements of a theory for the mechanisms controlling abundance,

diversity, and biogeochemical role of lytic bacterial viruses in aquatic systems. Limnol.

Oceanogr. 45:1320-1328.

45. Thingstad, T. F., R. Lignell. 1997. A theoretical approach to the question of how

trophic interactions control carbon demand, growth rate, abundance, and diversity.

Aquat. Microb. Ecol. 13:19-27.

46. Tijdens, M., D. B. van de Waal, H. Slovackova, H. L. Hoogveld, and H. J. Gons.

2008. Estimates of bacterial and phytoplankton mortality caused by viral lysis and

microzooplankton grazing in a shallow eutrophic lake. Freshwater Biology. 54:1126-

1141.

47. Ward, T.E., D. F. Bruhn, M. L. Shean, C. S. Watkins, D. Bulmer, and V. Winston.

1993. Characterization of a new bacteriophage which infects bacteria of the genus

Acidiphilium. J. Gen. Virol. 74:2419-2425.

82 48. Weinbauer, M. G., and P. Peduzzi. 1995. Significance of viruses verses heterotrophic

nanoflafellates for controlling bacterial abundance in the Northern Adriatic Sea. J.

Plankton Res. 17:1851-1856.

49. Weinbauer, M. G., D. Fuks, P. Peduzzi. 1993. Distribution of viruses and dissolved

DNA along a coastal trophic gradient in the northern Adriatic Sea. Appl. Environ.

Microbiol. 59:4074-4080.

50. Wiggins, B. A., and M. Alexander. 1985. Minimum bacterial density for bacteriophage

replication: implications for significance of bacteriophages in natural ecosystems. Appl.

Environ. Microbiol. 49:19-23.

51. Wilhelm, S. W., and Suttle, C. A. 1999. Viruses and nutrient cycles in the sea.

Bioscience 49:781-788.

52. Wilhelm, S. W., and A. R. Matteson. 2008. Freshwater and marine virioplankton: a

brief overview of commonalites and differences. Freshwater Biology. 53:1076-1089.

53. Wilhelm, S. W., S. M. Brigden, and C. A. Suttle. 2002. A dilution technique for the

direct measurement of viral production: a comparison in stratified and tidally mixed

coastal waters. Microbial Ecol. 43:168-173.

54. Winter, C., A. Smit, G. J. Herndl, and M. G. Weinbauer. 2004. Impact of

virioplankton on archaeal and bacterial community richness as assessed in seawater batch

cultures. Appl. Environ. Microbiol 70:804-813.

55. Wommack, K. E., R. T. Hill, T. A. Muller, and R. R. Colwell. 1996. Effects of

sunlight on bacteriophage viability and structure. Appl. Environ. Microbiol. 62:1336-

1341.

56. You, Y., J. Han, P. C. Chiu, and Y. Jin. 2005. Removal and inactivation of waterborne

83 viruses using zerovalent iron. Environ. Sci. Technol. 39:9263-9269.

84 1 Table 1. Prokaryote, virus, physiochemical, and geochemical concentrations determined for each sample location. All values are mg/L unless 2 otherwise reported. 3

Prok. VLP Temp Eh Turbidity Alk (as 2+ - 3- Location 5 5 VPR pH Fetotal Fe SO4 NO3 PO4 DOC Al Ca Cu K Mg Mn Na Si Zn (x10 /mL) (x10 /mL) (˚C) (mV) (NTU) CaCO3).

Sudbury Igneous Complex

Fecunis 1 0.63 2.73 4.32 2.91 14 664 n.d n.d 1.60 0.41 900 n.d n.d n.d 7.50 243.20 2.70 3.80 58.20 1.20 167.00 36.40 1.27

Fecunis 2 3.35 3.07 0.92 4.00 14 662 n.d n.d 1.28 0.22 937.5 1.70 0.00 1.80 4.50 277.00 3.70 25.00 52.90 1.89 140.00 20.90 0.50

Longvac 1 6.84 16.21 2.37 2.45 14 660 n.d n.d 1170.00 196.00 2400 n.d n.d n.d 18.40 209.30 0.66 3.55 54.50 2.39 14.32 39.50 1.30

Longvac 2 1.05 7.35 7.00 3.60 14 660 n.d n.d 294.00 105.00 1800 69.60 0.01 4.60 18.90 136.90 0.63 13.10 37.00 2.15 12.30 28.00 0.40

Longvac 3 4.69 9.66 2.06 2.50 14 660 n.d n.d 210.00 44.00 1400 62.40 0.10 5.00 14.80 143.00 0.49 5.11 40.80 2.52 7.40 30.70 0.40

St. Lawrence

Lowlands

Cedarvale 1 1.42 37.99 26.79 8.43 13 299 n.d n.d 0.04 0.02 60 1.20 0.12 5.81 0.00 151.69 0.01 3.16 37.00 0.05 121.83 6.96 0.01

Cedarvale 2 1.79 45.39 25.41 8.38 13 312 n.d n.d 0.06 0.05 76 1.50 0.12 8.75 0.02 239.21 0.00 3.35 56.60 0.09 228.69 11.40 0.02

Cedarvale 3 7.52 105.72 14.04 7.30 13 50 17.70 220 0.25 0.14 60 1.50 1.03 30.67 0.01 166.23 0.01 25.87 45.54 0.81 81.74 11.99 0.01

Cedarvale 4 6.86 45.03 6.56 7.21 13 -88 62.70 226 0.13 0.07 47 3.00 1.18 23.64 0.02 180.41 0.00 10.90 46.67 0.37 84.09 13.18 0.02 Lake Ontario 3.20 27.42 8.55 8.57 15.5 430 1.70 57.5 0.01 0.00 32 1.40 0.25 3.17 0.01 46.46 0.00 1.58 13.63 0.00 18.81 0.13 0.01 Sailing club Lake Ontario Sunnyside 2.80 42.18 15.05 8.22 17 358 0.90 72.2 0.05 0.01 33 1.80 0.12 3.49 0.01 54.53 0.01 1.95 15.73 0.02 28.03 0.60 0.01 Park Lake Ontario 3.86 37.46 9.70 8.37 15.5 225 3.00 65.8 0.02 0.00 29 1.80 0.11 2.78 0.01 49.96 0.01 1.46 13.49 0.00 14.93 0.50 0.02 Port Hope Lake Ontario Prince 3.33 31.94 9.60 9.09 18 302 46.40 54.8 0.07 0.00 29 1.20 0.04 4.56 0.01 40.43 0.01 1.64 12.81 0.01 14.45 0.13 0.01 Edward Point Boyne River 1.38 41.06 29.75 8.48 5 491 0.43 214.4 0.02 0.01 32 3.40 0.05 6.65 0.02 112.11 0.01 2.32 29.14 0.00 33.55 4.09 0.02 Osprey 1.81 17.50 9.67 7.77 3.5 435 5.69 41 0.23 0.12 3 1.00 0.05 25.53 0.00 46.25 0.01 1.09 15.86 0.00 0.62 2.26 0.02 Wetland Beaver River 0.81 21.66 26.74 8.04 9 420 0.65 148 0.06 0.01 13 2.20 0.02 4.63 0.02 91.82 0.01 1.05 36.84 0.00 6.70 3.38 0.02 Nottawasaga 1.23 23.02 18.72 8.23 10 422 5.57 46 0.01 0.00 16 1.60 0.13 2.40 0.01 34.91 0.01 0.77 10.62 0.00 5.20 0.69 0.01 Bay Nottawasaga 1.76 30.62 17.40 8.05 7 347 4.92 106 0.08 0.01 26 2.10 0.07 6.15 0.02 110.44 0.01 1.79 25.48 0.05 18.40 5.61 0.02 River Willow 0.96 26.75 27.86 8.25 9 429 1.50 102 0.09 0.04 1 1.50 0.05 5.22 0.02 89.20 0.01 1.60 19.37 0.03 29.63 6.23 0.02 Creek Minesing 1.12 31.46 28.09 7.77 8 230 1.39 133 0.10 0.01 24 1.00 0.12 7.14 0.02 107.04 0.01 2.48 24.43 0.04 20.85 8.05 0.02 Swamp Highland 0.72 36.50 50.55 8.25 9 496 1.36 137 0.12 0.02 70 0.80 0.07 4.42 0.02 135.42 0.01 2.87 28.36 0.05 108.47 2.85 0.02 Creek Rouge River 3.64 61.06 16.77 8.56 9 482 3.42 114 0.07 0.02 61 1.90 0.09 5.31 0.02 121.36 0.01 3.36 25.00 0.01 84.88 2.73 0.02 1

85 Rouge River 4.21 38.43 9.13 8.30 10 477 11.50 107 0.03 0.01 56 1.00 0.05 5.51 0.02 121.08 0.01 3.42 24.59 0.03 72.56 3.36 0.02 2 Rouge River 2.89 56.88 19.68 8.28 8 403 5.25 121 0.03 0.00 55 1.20 0.02 5.41 0.02 120.42 0.01 3.25 24.17 0.03 71.61 3.42 0.02 3 Lake Scugog 2.51 33.67 13.41 8.24 9 501 3.93 94 0.07 0.00 7 1.90 0.05 37.23 0.02 68.30 0.01 1.53 11.29 0.00 23.32 3.23 0.02 East Cross 1.81 18.17 10.04 7.84 6.5 292 0.78 235 0.15 0.05 2 1.50 0.14 69.05 0.02 115.76 0.01 0.77 16.62 0.19 14.95 7.30 0.02 Creek Scugog River 2.63 39.41 14.98 8.05 10 412 1.85 159 0.08 0.01 15 2.30 0.04 46.45 0.02 85.33 0.01 1.78 13.02 0.01 15.94 3.00 0.02 Sturgeon 2.74 39.04 14.25 8.06 8.5 392 2.23 38 0.05 0.01 4 1.30 0.05 12.92 0.01 23.94 0.01 0.48 3.18 0.00 4.46 2.87 0.02 Lake

Grenville

Province

Lake 3.46 72.80 21.04 8.45 9 407 3.44 97 0.02 0.00 24 1.80 0.08 6.19 0.01 53.09 0.01 2.05 11.55 0.00 25.93 2.01 0.02 Couchinching Severn River 4.91 59.60 12.14 7.94 9 407 1.15 37 0.19 0.05 16 1.00 0.12 6.69 0.01 38.65 0.01 1.36 8.47 0.02 17.63 2.46 0.02

Muskoka Bay 3.12 41.60 13.33 7.46 11 283 5.52 18 0.38 0.06 6 1.60 0.14 5.70 0.01 20.25 0.01 1.36 3.79 0.26 27.08 2.52 0.02 Muskoka 4.00 29.50 7.38 7.33 10 392 0.54 6.5 0.10 0.02 1 1.30 0.10 4.97 0.01 3.81 0.01 0.20 1.25 0.01 2.91 1.72 0.02 River Black River 2.25 34.10 15.16 7.21 5 392 0.94 3.5 0.30 0.04 2 0.30 0.10 6.84 0.04 3.37 0.01 0.30 1.17 0.02 1.36 2.77 0.01

Kahshe Lake 3.35 26.35 7.87 7.48 10 400 5.52 6.6 0.25 0.05 1 1.20 0.05 6.68 0.01 4.64 0.01 0.14 1.16 0.04 2.05 1.01 0.02 Sparrow 3.46 53.22 15.38 8.21 9 399 0.70 72.1 0.03 0.00 19 1.20 0.11 5.78 0.01 44.84 0.01 1.70 10.16 0.01 22.32 2.07 0.02 Lake Lake Bernard 1.35 19.82 14.68 7.49 9 340 1.53 13 0.07 0.00 5 1.50 0.04 3.83 0.00 9.88 0.01 0.94 2.17 0.02 13.74 3.91 0.01

Horn Lake 2.25 20.23 8.99 7.09 10 472 0.69 1.3 0.04 0.01 2 1.30 0.04 3.39 0.00 2.60 0.01 0.09 0.51 0.05 0.86 0.32 0.02 South Horn 1.52 22.60 14.87 5.85 5.5 418 0.44 0.3 0.39 0.18 3 2.10 0.05 10.35 0.11 3.55 0.01 0.08 0.92 0.05 1.32 3.56 0.01 Lake Rd Cecebe Lake 3.13 22.33 7.13 6.99 9 455 3.43 4.9 0.39 0.06 3 1.70 0.01 8.00 0.04 4.77 0.01 0.42 1.41 0.03 3.64 2.52 0.01 Magnetawan 2.77 24.60 8.88 6.99 7 446 1.43 4.3 0.34 0.06 2 1.70 0.10 7.71 0.04 4.94 0.01 0.35 1.46 0.03 2.56 3.35 0.02 River wetland near 3.35 24.93 7.44 6.48 6.5 422 0.75 4.4 0.43 0.13 3 2.00 0.02 8.56 0.08 3.28 0.01 0.21 1.19 0.20 1.18 4.31 0.01 Mayfield Lake Mary Lake 1.53 14.52 9.49 6.83 10 460 0.46 4.1 0.18 0.04 2 1.90 0.07 6.49 0.03 4.48 0.01 0.29 1.38 0.00 4.53 2.47 0.01

Stony Lake 6.81 89.22 13.10 8.41 9 495 0.46 69 0.02 0.00 1 1.10 0.01 22.37 0.01 40.42 0.01 0.68 5.22 0.00 7.32 2.11 0.02

York River 2.42 15.86 6.55 7.85 9 463 1.59 24 0.32 0.05 6 2.50 0.09 10.32 0.01 15.91 0.01 1.12 2.56 0.03 8.94 2.94 0.02 Little Mississippi 1.37 18.84 13.75 8.05 9 467 1.42 28 0.15 0.07 1 1.90 0.14 10.30 0.01 16.65 0.01 0.63 3.83 0.01 4.17 2.42 0.02 River Denbigh 5.09 38.88 7.64 8.12 9 430 2.13 10 0.02 0.00 5 1.20 0.10 21.34 0.01 38.71 0.01 0.88 6.54 0.02 28.50 2.24 0.02 Lake Upper Mazinaw 1.30 10.62 8.17 7.88 10 472 0.65 17 0.03 0.01 3 1.30 0.04 5.84 0.00 11.11 0.01 0.55 2.42 0.00 4.16 2.47 0.02 Lake Little Skootamatta 3.20 45.10 14.09 7.17 10 460 0.78 16 0.12 0.04 2 1.30 0.06 6.94 0.04 4.62 0.01 0.40 1.36 0.03 3.60 2.44 0.01 River Mean value 2.88 33.59 13.89 7.34 10.11 415 5.28 71.43 35.05 7.23 173 4.38 0.12 10.68 1.35 75.88 0.16 3.17 17.94 0.27 33.93 6.44 0.10

86 4 n.d. means not determined.

5

6 Table 2. Spearman rank correlation coefficients of microbial and physiochemical constituents. 7 - 3- Prokaryote VLP VPR pH Temperature Eh turbidity alkalinity Fetotal SO4 NO3 PO4 DOC Prokaryote 1.000 VLP 0.448 1.000 VPR -0.419 0.520 1.000 pH -0.001 0.579 0.588 1.000 Temp 0.247 -0.074 -0.384 -0.035 1.000 Eh -0.110 -0.429 -0.347 -0.208 0.028 1.000 turbidity 0.309 0.218 -0.113 0.227 0.282 -0.297 1.000 alkalinity 0.001 0.432 0.487 0.561 0.046 -0.156 0.268 1.000 Fetotal -0.009 -0.486 -0.486 -0.818 0.005 0.199 -0.002 -0.348 1.000 SO4 0.102 0.105 0.009 0.117 0.481 0.108 0.424 0.639 0.019 1.000 - NO3 -0.107 -0.268 -0.165 -0.242 0.109 0.174 -0.062 0.136 0.251 0.180 1.000 3- PO4 0.165 0.286 0.133 0.139 0.179 -0.489 0.179 0.174 -0.072 0.134 -0.059 1.000 DOC 0.153 0.267 0.055 -0.030 -0.520 -0.203 0.039 0.063 0.119 -0.400 0.035 0.107 1.000 Mg -0.038 0.053 0.115 0.168 0.346 0.077 0.352 0.929 0.012 0.822 0.176 0.172 -0.218 8 9 Bold type indicates correlations coefficient are significantly correlated (p < 0.05) 10

87 Table 3. Multiple regression analysis with prokaryote abundance as the dependent variable. R2 = 0.90, n = 46, F-value = 78.21, p-level = 0.00.

Independent Parameter Std Error p-level Mean Std Dev Beta coefficient variables estimates (β) pH 0.140 0.048 0.006 7.54 1.29 0.33 2- SO4 0.002 0.000 0.000 108.60 351.96 0.26 3- PO4 3.728 0.888 0.000 0.12 0.22 0.28 VLP 0.048 0.010 0.000 34.63 20.11 0.59 Mg2+ 0.037 0.013 0.009 16.27 15.69 -0.25

Table 4. Regression analysis with VLP abundance as dependent variable, n = 40, F-value = 134.02, p-level = 0.00.

Independent R2 p-level Mean Std Dev Beta variable coefficient (β) Goethite 0.78 0.00 5.71 1.02 0.89

Table 5. Pearson correlation coefficient of mineral saturation indices verses pH and microbial constituents, n = 47

Mineral Mineral Formula Prokaryote VLP VPR pH abundance abundance Ferrihydrite Fe(OH)3 -0.036 0.388 0.301 0.753 Goethite α-FeO(OH) 0.010 0.413 0.285 0.744 H-Jarosite (H3O)Fe3(SO4)2(OH)6 0.237 -0.326 -0.495 -0.906 K-Jarosite KFe3(SO4)2(OH)6 0.278 -0.212 -0.423 -0.831 Na-Jarosite NaFe3(SO4)2(OH)3 0.275 -0.204 -0.402 -0.834 Pyrolusite MnO2 -0.142 0.433 0.484 0.893

88 Figure Legends

Figure 1. Map of southern Ontario with solid circles (•) representing 1-4 sample locations, and stars () representing large, nearby cities. Water samples were collected from three major regions composed of different geologic formations (i.e. bedrock types). The lower zone in which Toronto resides, is called the St. Lawrence Lowlands and is dominated by carbonate rocks and shale. Central southern Ontario is composed of silicate rocks of the

Grenville Province, and the upper zone, called the Sudbury-Igneous Complex which is composed of many ore deposits (i.e. sulfides). A map of Canada reveals where southern

Ontario is located within the country.

Figure 2. Transmission electron micrographs of common VLPs found in southern Ontario surface waters. All tailed phage morphotypes were noted (a-g), with Myoviridae commonly revealing more complex tail tips (a-c). Siphoviridae morphotypes were also noted (d, e), as well as Podoviridae (right side in f, g). Fuselloviridae (left side in f), a morphotypes commonly noted in Archaea, were rarely noted. Polyhedral morphotypes were also quite common (h, arrow notes what appears to be tail fibers). Scale bar is 100 nm.

Figure 3. Transmission electron micrograph of possible VLPs (arrows) sorbed to inorganic material from an AMD site, Longvac. VLP spheres range from 33-40 nm in diameter. Scale bar is 100 nm.

89

Figure 1

Figure 2

90

Figure 3

91 Chapter 5

Bacterial-Phage Interactions and Authigenic Mineral Precipitation

5.1 Abstract

Bacterial-phage interactions were examined under mineralizing conditions where authigenic mineral precipitation was occurring. Iron-oxidizing bacteria (IOB), which actively undergo iron mineralization due to their metabolism, were isolated from an acid mine drainage (AMD) environment in an attempt to isolate an IOB phage. In addition, experiments were conducted using Bacillus subtilis and a temperate phage, SPβc2 under iron saturated conditions (0.1 mM ferric iron, near neutral pH) to induce bacterial mineralization.

Although a phage could not be isolated for an IOB, B. subtilis mineralization resulted in a substantial decrease in phage replication (~ 98%). In addition, iron addition to lysogenic cultures did not induced viral lysis despite iron precipitation at cell surfaces. If the B. subtilis-SPβc2 results were applied to natural environments, bacterial mineralization would be advantageous to bacterial hosts as it protects against phage attachment and subsequent viral lysis. Moreover, lysogeny would be advantageous to phages in mineralizing environments as the precipitation of minerals on host bacteria drastically hinders phage replication. Distinctive patterns of microbial and phage associated mineral precipitates were not noted within the experiments.

92 5.2 Introduction

Interactions between bacteria and bacteriophages have been examined in a wide range of environments including the deep terrestrial subsurface and acid mine drainage systems

(Kyle et al. 2008a,b). While authigenic mineral precipitation is common in many aquatic environments, interactions between bacteriophages and their hosts under mineralizing conditions have not been examined. This is an intriguing biogeochemical relationship to investigate as (i) bacteriophage attach to the same components in bacterial cell walls that attract dissolved mineral-forming elements, in the surrounding environment, (ii) long term viral infection of a bacterial cell (i.e., lysogeny) causes structural and compositional changes to the cell surface where phages and dissolved ions bind, possibly altering the reactivity of the attachment-sorption site, and (iii) the cell may become stressed during surface associated mineral formation possibly causing the induction of a lysogenic cell.

Extensive investigations have been conducted on bacterial mineralization

(Lowenstam 1981; Ferris et al. 1988; Beveridge 1989; Westall et al. 1995; Phoenix et al.

2000; Toporski et al. 2002; Chan et al. 2004). Under most natural aquatic environments where pH values range from 6 – 8, bacterial surfaces are characterized by a net negative charge due to the deprotonation of functional groups (dominantly carboxyl and phosphoryl groups) located within the cell wall and external sheaths. Once these charged functional groups attract dissolve metal cations (i.e. Fe3+, Si4+, Ca2+) from solution, they can serve as a heterogeneous nucleation site for mineral precipitation (Fortin et al. 1997). Minerals such as silica and ferrihydrite have been shown to form on bacteria, at times entombing and preserving cells as microfossils (Westall et al. 1995; James and Ferris 2004). Through these studies we have gained an understanding of how bacteria can be structurally preserved by

93 authigenic mineral precipitation and incorporated into the rock record as microfossils.

The sorption of ferrous iron and precipitation of hydrous ferric oxides by bacteriophage was reported by Daughney et al. (2004); however, no studies, to our knowledge, have considered what happens to bacteria-phage interactions under geochemical conditions that promote authigenic mineral precipitation. An additional issue is that there is no known evidence of viruses in the rock record, owing presumably to their small size (30-

100 nm) and lack of a unique biologic signature (i.e. lipid or isotopic). For these reasons, investigating the mineralization of phages and the potential impact of phages on bacterial mineralization will not only lead to a greater understanding of bacteria-phage dynamics, but also provide new insight on how phages are apt to influence bacterial microfossil formation and whether their interaction with bacteria may yield a mineral biosignature that can be recognized in the rock record.

In this investigation, two studies were conducted to examine phage-host dynamics under mineralizing conditions. The first involved an attempt to isolate a phage that belongs to an iron oxidizing bacteria (IOB) from an acid mine drainage (AMD) environment as bacteria from these systems are known to naturally undergo mineralization (commonly ferric hydroxides) due to the bacterium’s metabolism. The second involved use of a well characterized bacterium, Bacillus subtilis 186, that has been studied extensively in mineral precipitation experiments (Beveridge and Murray 1980; Ferris et al. 1988; Warren and Ferris

1998; Châtellier and Fortin 2004; Wightman and Fein 2005) and a known temperate phage,

SPβc2. It is hypothesized that the mineralization of host and/or phage will impact this parasitic relationship.

94 5.3 Methods and Materials

5.3.1 Site characterization and sample collection

Samples were collected from two acid mine drainage (AMD) environments, Fecunis and Longvac tailings ponds, located near Sudbury, Ontario, Canada in August of 2007 and in the summer of 2008. For both sites, physiochemical parameters (pH, Eh, and dissolved iron, sulfate, nitrate, phosphate) were measured in situ. Additional samples were filtered using a

0.2 µm syringe filter into 60 mL nalgene bottles and amber EPA certified 40 mL glass vials so that dissolved elemental constituents and dissolved organic carbon (DOC), respectively, could be measured. There was no head space left in the containers and they were stored in a cooler at 4°C. Samples collected for direct epifluorescent counting of viral and prokaryotic abundance and transmission electron microscopic imaging were preserved to a final concentration of 2.5 % (v/v) aqueous glutaraldehyde.

5.3.1.1 Viral and prokaryotic abundance and viral imaging

Preserved samples (0.3 to 0.8 mL) were stained with SYBR Green I within one week of collection according to the method of Noble and Fuhrman (1998). Briefly, samples were filtered onto 0.02 µm Anopore membrane filter, which were then placed onto a 100 µL drop of 2.5 % SYBR Green I and stored in the dark. After 15 minutes, excess stain was removed from the filters and they were stored an additional 20 minutes in the dark to dry. A drop of antifade solution was placed on cover slips, which were then inverted onto the samples.

Samples were stored at -20ºC until viewed on a epifluorescence microscope (Nikon

Microphot-FXA) under blue excitation light.

Samples were imaged by transmission electron microscopy (TEM) to confirm the

95 presence of viruses and determine the morphology of virus particles. This involved filtration of preserved samples through 0.2 µm syringe filters and centrifugation of the filtrate (Sorvall

RC5B Plus, rotor SS-34) at 43150 x g for 2 hours at 10°C. Most of the supernatant was discarded, except for approximately 50 µL of sample at the bottom of the tube, which was mixed by gently pipetting the solution up and down. 20 µL of sample was transferred onto

300 mesh formvar, carbon-coated (providing extra support and strength under the high accelerating voltage of the electron beam on the TEM) copper grids and the TEM specimens were allowed to sit for 20 minutes. Samples were subsequently stained with 10 µL of 1% uranyl acetate (0.02 µm filtered) for 45 seconds. All excess liquid was then wicked off the grid with filter paper and the samples were air dried thoroughly before viewing with a Philips

201 TEM operating at 80 kV.

5.3.1.2 Aqueous Chemistry

Analyses were conducted within 24 hours of sample collection using HACH spectrophotometric assays for dissolved concentrations of total iron, ferrous iron, sulfate, nitrate, and phosphate. For these analyses, samples were filtered using 0.2µm syringe filters into 25 mL glass cuvettes for analysis. Dissolved elemental constituents were analyzed using inductively coupled plasma atomic optical emission spectroscopy (ICP-AOES Perkin Elmer

Optima 7300DV). Samples for measurement of DOC concentrations were shipped to G.G.

Hatch Isotope Laboratories at the University of Ottawa and analyzed on an OI Analytical

Aurora Model 1030W TOC Analyzer.

5.3.1.3 Isolation of IOB phage

96 In August of 2007, triplicate cultures of Acidithiobacillus ferrooxidans and

Leptospirillum ferrooxidans (donated by Doug Gould, CANMET) and 11 IOB isolated from the Rio Tinto, Spain (J. E. Kyle, unpublished data) were inoculated with 4 mL of 0.2 µm filtered AMD water at each site. Control cultures were not infected and blanks consisted of growth medium only. The cultures, which had been growing for 8 days at room temperature

(~22°C) on a shaker at 100 rpm, were growing in 50 mL serum bottles containing 30 mL of liquid FeSo medium (Johnson 1995) which was composed of the following (/L): 25 mM

FeSO4 ⋅ 7H2O, 2.5 mM K2S4O6, 1.8g (NH4)SO4, 0.7 g MgSO4 ⋅ 7H2O, and 0.35 g tryptic soy broth (TSB). The pH of the medium was adjusted to 2.3 using 25 % HSO4. This medium normally contains 0.7% agarose, but it was excluded as phages are more easily isolated using liquid cultures.

At the same time, enrichment cultures of IOB from Fecunis and Longvac were grown by inoculating FeSo medium with 1 mL of unfiltered water. The cultures were incubated at room temperature and shaken at 100 rpm until growth was noted then stored at 4 °C.

Isolation of IOB from Fecunis and Longvac enrichment cultures was accomplished by plating on FeSo medium solidified with 0.7 % w/v agarose. The agarose was washed in distilled water for 30 minutes, then centrifuged at 10 000 rpm for 7 minutes before sterilization to remove any soluble constituents. Individual colonies were selected based on differing colour and colony morphology and then transferred three times by streaking onto solidified FeSo medium to obtain a pure culture. One colony of IOB was isolated from

Fecunis and three were isolated from Longvac (herein referred to Longvac A, B, and C).

Each of the four cultures were then transferred into liquid FeSo medium until only one morphotype was noted under a light microscope. The samples were stored at 4°C.

97 In June of 2008, cultures of Fecunis, Longvac A, B, and C IOB were infected with

AMD water from the location of where the cultures had been isolated. Before sample collection, the growth rate of the cultures was monitored such that at 7th day of growth the

OD600 was between 0.3 to 0.6, consistent with exponential growth. Triplicate cultures that were growing for 7 and 9 days were brought into the field and infected with 1/10th the volume of the cultures (3 mL into 30 mL culture). All infected cultures were shaken at 100 rpm at room temperature. Approximately 8 hours later the 9 days old cultures appear to have cleared and contained debris at the bottom of the vials. A subsample from each culture was preserved in glutaraldehyde to a final concentration of 2.5 % (v/v), and another subsample from 2 vials of each culture was filtered through a 0.2 µm syringe filter into a 15 mL falcon tube and stored at 4°C (to create a phage stock if phages were present). The following day, mitomycin C (a DNA damaging agent that is known to induce viral lysis) was added to a final concentration of 0.5 µg/mL into one vial for each culture in an attempt to induce potential lysogens with the culture. Mitomycin C treated vials were returned to the shaker for 24 hours, after which Longvac B-9-mito (9 represents the 9 day old cultures, and mito represents mitomycin C treated) cultures contain debris at the bottom of the vial. A 15 mL subsample from Longvac B-9-mito was filtered through a 0.2 µm syringe filter into a 15 mL falcon tube and stored at 4°C, and another subset was preserved in glutaraldehyde. As no noticeable changes had occurred in the untreated samples (no addition of mitomycin C), the cultures were preserved and then stored at 4°C.

Five days post-infection, phage stock from Longvac A9, B9, C9, and Longvac B9- mito cultures were injected (1/10th of final volume) into 4 day old liquid cultures in an attempt to propagate and isolate an IOB phage. In addition, a phage titer was conducted in

98 0.9 mL of liquid FeSo medium in which 0.1 mL of phage stock dilution (up to 10-8) and 0.1 mL Longvac B-9 cultures were added.

Preserved samples of Longvac A9, B9, and C9, Longvac B7, and Longvac B9-mito cultures were stained for TEM using the same protocol described above.

5.3.1.4 Isolation of IOB phage from Longvac

A total of 45 replicate cultures of Longvac B9 were prepared for the August 2008 field session. Thirty cultures were infected with AMD water from Longvac, and 10 were infected with water collected below the sediment/water interface (inserted syringe into sediment and withdrew water). Five of the cultures remained as controls (i.e., not infected).

After 1 day of incubation, mitomycin C (0.5 µg/mL final concentration) was added to 10 of the cultures (1 from the sediment samples) and incubated for an additional 4 days.

5.3.2 Bacillus subtilis – SPβc2 mineralization experiments

Bacillus subtilis is a gram positive, aerobic, chemoorganotroph that was originally isolated from a soil environment. This bacterium was chosen for the experiments as it has been used in numerous mineralization experiments (Ferris et al. 1988; Châtellier et al. 2001;

Yee et al. 2004; Wightman and Fein 2005) and its well characterized cell surface contains functional groups known to bind dissolved metals (Beveridge and Murray 1980; Daughney et al. 1998; Cox et al. 1999).

Temperate phage SPβc2 is a Siphoviridae that uses its tail to bind to the host cell

(Fink and Zahler 2005). This phage is temperature sensitive so when the lysogen is exposed to a brief heat shock (52 °C), the repressor function enabling lysogeny is destroyed and the

99 prophage enters lytic growth cell (Fink and Zahler 2005). The size of the phage (~70nm capsid and 150 nm long tail) and life cycle makes the phage easier to view under the transmission electron microscope (TEM) and the lysogen easier to manipulate during the experiment. Both Bacillus subtilis 168 and Bacillus subtilis 168 lysogen containing prophage SPβc2 were donated from the Bacillus Genetic Stock Center (BGSC ID 1A100 and

BGSC ID 1L5, respectively).

5.3.2.1 Obtaining Lysate

SPβc2 viral particles were obtained by protocols outlined in Hemphill 1990. Briefly, the lysogen was put onto the edge of plates that contained modified M medium broth (MMB) with 1.5 % agar. The inoculated plates were incubated at 37 °C overnight. The MMB contained the following ( /L): 10 g Bacto-tryptone, 5 g yeast extract, 5 g NaCl, 0.005 M

MgCl2, and 0.0001 M MnCl2. The next day, an individual colony was inoculated into 2 mL

MMB and grown overnight at 37 °C and shaken at 200 rpm. The following day, 100 mL of

MMB medium was inoculated with 2 mL of overnight culture then shaken at 200 rpm at 37

°C until the optical density (OD600) reached 0.3 (OD600 analyzed on a HACH spectrophotometer at 600 nm wavelength). When the OD600 = 0.3, the lysogen was submerged into a water bath at 52 °C for 5 minutes then the culture was returned to the incubator and shaken at 200 rpm at 37 °C until the cells lysed (OD600 of 0.02). To isolate the viral particles, the culture was centrifuged at 6300 rpm for 10 minutes at 15 °C. A phage stock was produced through filtering the supernatant to remove cellular debris using a sterile

0.2 µm syringe filter into a sterile serum vial covered in aluminum foil and stored at 4 °C.

100 5.3.2.2 Plaque Assay

To determine the viral abundance in the phage stock, a plaque titer was conducted using an agar overlay method (0.5% agar for top layer and 1.5% agar for bottom layer).

Bacillus subtilis cells were grown up overnight in 2 mL of MMB then transferred into

100mL medium the following morning. When the cells reached an OD600 of 0.85, 0.1 mL of the culture was added to 1.5 mL centrifuge tubes containing 0.1 mL of phage dilutions tube

(dilutions 10-4 to 10-8) and placed in an incubator at 37 °C for 20 minutes (to allow for phage adsorption and infection). The contents of the centrifuge tubes were then added to 3 mL of liquid top agar in a water bath at 52 °C for 5 minutes. The top agar solution was then poured on top of plates containing bottom agar and swirled to obtain a level top agar coverage. The plates were then inverted and incubated at 37 °C overnight and plaques were counted the following morning. Duplicates for each dilution were made. The 10-7 phage dilution plates contained 62 and 72 plaque forming units (PFU) revealing the phage stock contained 7.2 x

109 phages/mL (see Equation 1).

[Eq.1] phages/mL = (number of PFUs )(1 / volume of phage stock added)(dilution factor)

5.3.2.3 Experiment 1: Bacillus subtilis with iron plus phage

This experiment was conducted to determine if the mineralization of a host cell prevents and/or hinders phage attachment and replication. B. subtilis was grown overnight in

2 mL of MMB then transferred into 2000mL flask containing 200mL medium the following day. When the cells reached an OD600 of 0.6, a 50 mM stock solution of FeCl3·6H2O was added to the microcosm to a final concentration of 0.1 mM of iron. However a major

101 problem occurred as the MMB medium removed large amounts of iron. Samples taken for

TEM revealed iron was not forming on the bacterial cells. To overcome this problem, once

B. subtilis reached an OD600 of 0.6, the cells were washed with modified artificial ground water (AGW; containing 0.0403 mM KNO3, 0.448 mM MgSO4⋅7H2O, 1.75 mM CaCl2, 1.1 mM NaHCO3; 0.0623 mM KHCO3; Mitchell and Ferris 2005) at a pH of 7.4 Artificial groundwater was used as divalent cations, i.e. calcium and magnesium ions, are known to assist in adsorption for some phages and phage replication (Ackermann and DuBow 1987;

Moebus 1987).

For these experiments, the culture was centrifuged at 6000 rpm for 10 mins at 20 °C to pellet the cells. The supernatant was discarded and replaced with AGW, re-centrifuged, then resuspended in AGW. To confirm the cells were able to continue to grow and susceptible to SPβc2 infection, a plaque assay was conducted (as described above). Full lawns with plaques were noted the next day.

Once the culture was washed and resuspended in 100 mL AGW, FeCl3·6H2O stock was slowly added to a final concentration of 0.1 mM of iron. 20 mL of AGW suspended cells were then transferred into tripicate 250 mL sterile pyrex bottles. The cultures remained at room temperature and were not shaken for up to 90 minutes after iron addition. Plaque assays and measurements of pH and total dissolved iron were conducted at 0, 45, and 90 minutes after iron addition. The experiments were not conducted over a greater length of time as reports have shown that most iron precipitates on cellular surfaces within two hours

(Wightman and Fein 2005). The pH of the microcosm was measured before and after iron addition. Total dissolved iron was measured using a HACH spectrophotometry (model

DR/2500). Samples were collected before and after iron addition and after 30 minutes of

102 incubation for TEM. For this, 20 µL of sample was dropped on a 300 mesh carbon, formvar- coated copper grid for 5 minutes. Then 20 µL of 1 % uranyl acetate was added for 60 seconds. Excess liquid was then wicked off the grid with filter paper before viewing with a

Philips 201 TEM as noted above. Control experiments with no iron addition were also conducted. Duplicates for each plaque assay were plated.

5.3.2.4 Experiment 2: Lysogen plus iron

This experiment was conducted to determine if the mineralization of a lysogen would induce viral lysis. Washed B. subtilis cells were used for this experiment as centrifugation of the lysogen induced lysis. Harvested and washed B. subtilis cells were resuspended in 120 mL of AGW then divided amongst 5 microcosms; duplicates of the controls (no iron addition) and triplicates of the lysogen with iron. 1 mL of SPβc2 stock was added to each microcosm, both of which were then incubated at 37 °C for 20 mins. FeCl3·6H2O stock was added to the lysogenic solution to a final concentration of 0.1 mM of iron. Immediately after iron addition, sample was withdrawn and measured for pH, OD600, and total dissolved iron.

After 2 hours of incubation at room temperature, unshaken, the same measurements were conducted and a subsample was taken for TEM, half of which were stained with 1% uranyl acetate. Also, 0.1 mL from the control and lysogenic solution was added to top agar to conduct a plaque assay. Duplicates of each were plated. The microcosms were sampled and plated again 20 hours later, with the exception of one lysogen-iron microcosm which was analyzed after 17 hours.

An additional control microcosm containing washed B. subtilis and iron was also monitored (underwent procedure above without SPβc2 addition).

103

5.3.2.5 Experiment 3: Phage with iron plus Bacillus subtilis

This experiment was conducted to determine if phage mineralization prevented and/or hindered phage attachment and subsequent infection. Harvested and washed B. subtilis cells were resuspended in 24 mL of AGW then infected with 1mL of phage stock for a final phage

8 concentration of 2.9 x 10 phages/mL. FeCl3·6H2O stock was added to phage solution to a final concentration of 0.1 mM of iron and incubation at 4 °C (to prevent viral degradation due to temperature and UV exposure) to allow for phage-iron interaction. After 30 and 60 minutes of incubation, a phage assay (without subsequent dilutions) using the phage-iron with washed B. subtilis cells at an OD600 of 0.7 was conducted. Dissolved total iron and pH of the phage-iron solution were measured and a sample was stained (as noted above) for

TEM. Iron added to a control microcosm (final concentration of 0.1 mM) containing 24 mL of AGW and 1 mL of MMB was also monitored.

5.4 Results

5.4.1 AMD Site characterization

Viral abundance, prokaryotic abundance, physiochemical, and geochemical results are shown in Table 1. Prokaryotic abundances ranged from 0.63 to 6.84 x 105 cells /mL, and

VLP abundance from 2.73 to 105.72 x 105 VLPs /mL. Transmission electron micrographs revealed diverse range of morphotypes with spherical and tailed phages being the most common. Samples collected from Longvac had greater concentrations of DOC and dissolved abundances of iron, sulfate, nitrate, and aluminum, and Fecunis had greater concentrations of copper. Otherwise, the geochemistry was similar for both Longvac and Fecunis.

104 5.4.1.1 IOB phage isolation using foreign cultures

The use of foreign cultures to isolate an IOB phage was unsuccessful. The infected cultures did not appear visually different than the control cultures, with no evidence of cell lysis (i.e., clearing of culture and/or cellular debris on the bottom of the vial). To verify a phage was not isolated, 2 of the 3 infected vials were preserved to a final concentration of 2.5

% (v/v) glutaraldehyde. The preserved samples were then filtered through a 0.2 µm syringe filter and centrifuged at 19 000 rpm for 2 hours at 10 °C, and then stained for TEM.

Transmission electron microscopy did not reveal any VLP morphologies, with the exception a few spherical shapes though the size and shapes were inconsistent with each other.

5.4.1.2 IOB phage isolation

After two days of growth of the re-infected cultures, no noticeable differences were noted between the infected cultures and the controls (not infected) for both the mitomycin C treated and untreated samples. The samples were examined under the light microscope in which no significant difference was noted between the infected and control cultures in the number of cells present. After an addition 3 days of incubation, the infected cultures continued to be identical to the controls, both of which contained noticeable biofilm growth and no evidence of cellular debris within the vials. Some spherical shapes were noted using

TEM although there was a lack of consistency in size and shape.

5.4.1.3 IOB phage isolation using Longvac cultures

None of the infected Longvac cultures appeared to have lysed, even with mitomycin

C addition. All samples formed biofilms after 5 days of growth, post infection, and appeared

105 the same as the controls.

5.4.2 Bacillus subtilis – SPβc2 mineralization experiments

5.4.2.1 Experiment 1: Bacillus subtilis with iron plus phage

Plaque assay results revealed a 2 orders of magnitude decrease in PFUs (plaque forming units) between the non mineralized and mineralized cells (Table 2). Total dissolved iron dramatically decreased in the presence of cells almost immediately coming out of solution (Table 2) and white aggregates formed within the solution.

Only minor amounts of iron precipitation were suspected to occur at the cell surface

(Fig. 1a) with rare occurrences of heavy mineralization around the cell (Fig. 1b). No phage particles were noted within either microcosm.

5.4.2.2 Experiment 2: Lysogen plus iron

Complete coverage of bacterial growth for the plaque assay was not attained, which made it impossible to determine the number of PFUs. Although most of the plates were covered (>95%) the lawns contained numerous connect blank areas which left a heavily speckled appearance indicative of the absence of phage plaques.

For both the control (without iron addition) and the lysogen, pH increased over time and OD600 decreased (Table 3). Total dissolved iron for both the nonmineralized and mineralized microcosms decreased over the first 2 hours and then increased back to the original optical density almost 1 day later. Results of the microcosm containing B. subtilis plus iron is reported in table 4, revealing a similar trend as with the lysogenic cultures except

OD600 did not decrease as much.

Bacterial mineralization was not evident using the TEM. Some cells appeared coated

106 (Figure 2), although individual mineral particles were not noted. Most of the cells contained an extensive network of interconnecting exopolymeric substances (EPS) strands that were amplified in appearance when stained with uranyl acetate. Phage particles were noted after

17 hours of lysogen incubation with iron. Of the few remaining bacterial cells noted in this microcosm, most were surrounded by SPβc2 particles (Figure 3a, c, d). Cells with partial or extensive EPS around the cells contained greater phage concentrations where there was direct access to the cell (Figure 3c). SPβc2 particles surrounding cells were also noted in unstained samples (Figure 4). The phages were only found around bacteria and not in the surrounding areas.

For the 20 hour incubated lysogen plus iron microcosms, large (~ 1 micron), dark, spherical structures with connected filaments (Fig. 5) were abundant. Also, many B. subtilis cells contained easily stained (with uranyl acetate) material adjacent to the cells, most of which contained SPβc2 particles (Fig. 6).

5.4.2.3 Experiment 3: Phage with iron plus Bacillus subtilis

This preliminary experiment revealed that after one hour of iron incubation with

SPβc2 a small amount of iron (0.8 mg/L total) was removed from solution with no significant change in pH (6.51 – 6.59). The control microcosm either removed the same or greater amounts of iron from solution. The plaque assays resulted in incomplete lawns containing numerous bacterial colonies, which make PFU determination impossible.

No evidence of SPβc2 particles were noted using the TEM.

5.5 Discussion

107 5.5.1 IOB phage isolation

Efforts at isolating an IOB phage were unsuccessful although the experimental approach shows promise. Infecting Rio Tinto IOB isolates with Sudbury filtrate was shown to be unsuccessful. Though the initial attempts at infecting IOB isolates from Sudbury resulted in cell death, however subsequent attempts at infection did not.

A few possibilities exist as for why an IOB phage could not be isolated.

First of all, phages within AMD systems may be species and/or strain specific. It is well known that many bacteriophages are species-specific with recent studies highlighting bacteriophages that are also often strain-specific (Chibani-Chennoufi et al. 2004; Holmfeldt et al. 2007). The bacterial species isolated from the Rio Tinto maybe of a different species and/or strain that those isolated from Sudbury, although the latter is more probable. Research conducted by Gonzáles-Toril et al. (2003) in the Rio Tinto found that 80 % of the prokaryotic community consisted of Acidithiobacillus ferrooxidans, Leptospirillum ferrooxidans, and

Acidiphilium. Acidithiobacillus ferrooxidans and L. ferrooxidans are commonly found in acidic aquatic mining environments around the world, including Ontario, Canada (Schrenk et al. 1998; Baker and Banfield 2003; Bernier and Warren 2005; Mahmoud et al. 2005).

Although molecular analysis was not conducted to identify the IOB isolates from this study, it would not be unexpected to find the same dominant IOB species but different strains in the

Rio Tinto and Sudbury.

Secondly, the bacterial cultures may have mutated during the attempt to isolate an

IOB. In the natural environment, bacterial hosts and phages would co-evolve where hosts would be able to develop resistance to infection followed by phage mutation once again leading to bacterial susceptibility. In the absence of infectious phages, bacterial species

108 isolated using enrichment cultures may have altered to the point that receptor sites are no longer complimentary for phage attachment and infection. It is also possible that the phages mutated within the AMD environment. Given that most bacteriophages are dsDNA viruses and DNA has low error replication rates due to error-correcting DNA polymerase (Flint et al.

2000), this scenario is unlikely.

Thirdly, the growth medium for the IOB cultures may not have contained the necessary constituents required for (i) the expression of complementary receptor sites for phage attachment, and/or (ii) trace elements required for or phage attachment and replication.

One of the most common bacterial species found in AMD, A. ferrooxidans is both an iron and sulfur oxidizing bacteria. When grown using different substrates (ferrous iron, elemental sulfur, and pyrite), A. ferrooxidans cells grown using Fe2+ as an electron donor were found have a lower isoelectric point (pHIEP of 2.0) and lower protein content then the other two substrates (Sharma et al. 2003). These observations would seem to be favorable for phage infection as the ferrous iron medium would result in the deprotonation of functional groups at lower pH values and the production of proteinaceous material may act as a barrier to surface receptor sites. The extent in which that proteinaceous material would prohibit infection is predicted to be minimal as many phages contain polysaccharide depolymerases that degrade

EPS layers surrounding cells enabling direct access to the cell surface (i.e. Hughes et al.

1998; Deveau et al. 2002).

Minor growth constituents may also influence bacterial growth and/or the ability of phages to attach and/or penetrate the host cell. Divalent cations, such as Ca2+ and Mg2+, are commonly required for phage attachment and sometimes for viral penetration (Paranchych

1966; Steensma and Blok 1979; Moebus 1987). In fact, Siphoviridae phage commonly

109 require Ca2+ for bacterial attachment (Steensma and Blok 1979). The medium used in the study was basic in that it did not contain a source of calcium or trace elements. Required trace elements may be been added during infection (filtration sample water into cultures); however, trace elements would have been diluted 10 fold. Commonly when A. ferrooxidans cells are cultured, the medium includes the addition of calcium nitrate, and for Leptospirillum sp., the addition of trace elements (i.e. Mn, Zn, Co, Cu). If phages require at least one of these elements at greater concentrations then available for phage replication, then infection and isolation of an IOB phage would not have been possible. Given that the identity of the

IOB cultures are unknown, it can at least be acknowledged that the sensitivity of phage-host interactions and the influence of growth conditions on the surface reactivity and characteristics may be critical.

Lastly, natural mineralization of bacterial surfaces through the precipitation of ferric hydroxides may have blocked receptor sites located on the bacterial surfaces preventing viral attachment. This could be one of the few advantages bacterial mineralization (in addition to protection from UV damage, dehydration, predation; see Phoenix and Konhauser 2008).

5.5.2 Bacillus subtilis-SPβc2 experiments

Iron addition to B. subtilis cultures drastically hindered SPβc2 replication although it did not induce viral lysis. Within 30 minutes of iron addition to B. subtilis suspensions most of the iron was removed (approximately 97 %) and the number of SPβc2 plaques decreased by two orders of magnitude (approx. 98.5 %). Although iron precipitation was rapid, noticeable amount of iron precipitates were not noted at the magnification examined on the

TEM for the majority of cells. This is likely due to the low bacteria to iron ratio where iron

110 is being distributed amongst many cells consequently resulting in less noticeable mineralization at the cell surface. Where mineralization was suspected the particles were less than 50 nm in size (Fig. 1a) however isolated cases of heavy mineralization were noted (Fig.

1b). Three scenarios are considered to explain these results: (i) most SPβc2 could not attach to the cell as iron precipitates and/or Fe(OH)2 ions bond to cellular receptor sites, (ii) SPβc2 attached to the cell but is unable to penetrate and/or inject its’ DNA due to a decrease in metabolic activity of the host cell, and/or (iii) SPβc2 particles became mineralized once iron was added to the microcosm. Of these three options, the first seems the most probable and latter the least, although a combination of processes could be occurring.

The detection of iron on bacterial cells without evidence of precipitates has been reported in Warren and Ferris (1998). The paucity of visible iron precipitates in this study could be the result of nanoparticle formation at the cell surface and/or the binding of Fe(OH)2 ions. Iron oxide nanoparticles (2-3 nm in size), such as ferrihydrite, are known to form on bacterial surfaces and EPS closely associated with the cell as Fe(OH)2 ions undergo

+ + precipitation (i.e. Fe(OH)2  FeOOH + H ; Banfield et al. 2000; Chan et al. 2004). Given the sensitivity of phage-host interactions the blockage of receptor sites due to precipitates and/or nonessential ion binding would prevent attachment and therefore infection.

If SPβc2 attachment is occurring then SPβc2 replication was impeded. Metabolic activity of host cells is important for phage replication. The proton motive force existing across cellular membrane has been shown to be critical for the injection of viral genome into host cells for most DNA phages (Guttman et al. 2005). The cells were likely stressed due to the lack of nutrients and aeration causing cellular processes to be hampered, preventing

SPβc2 replication.

111 The last scenario involves the removal of SPβc2 particles due to SPβc2 mineralization. Although possible, this scenario seems unlikely as total dissolved iron abundances (0.08-0.15 mg/L) were at concentrations measured in many natural environments

(J.E. Kyle, unpublished data). Preliminary experiments containing SPβc2 particles and iron revealed that SPβc2 is not a dominant site for mineral formation as only small amounts of iron (0.8mg/L) precipitated from solution (lower than or similar to the control). The iron is likely precipitating out of solution due to the pH of solution (> 6.5) and not due to a phage template. Incomplete bacterial lawns noted during the plaque assay demonstrated a decrease in B. subtilis growth, which is like due to SPβc2 infection and replication. However, if

SPβc2 particles did act as templates for mineral formation, this did not inhibit phage replication cycles.

The prevention of SPβc2 replication within a host cell would be an advantage of bacterial mineralization as one the greatest factors in bacterial mortality, viruses, (Thingstad and Lignell 1997; Tijdens et al. 2008) would be eliminated.

The question of why bacterial mineralization occurs and its potential benefits has been addressed with result that revealed that it protects the cells from UV light (Phoenix and

Konhauser 2008), inhibits autolytic enzymes (Ferris et al. 1986), and localizes iron oxides for easily accessible metabolic needs (Chan et al. 2004). This study indicates that bacterial mineralization would also protect cells against viral replication resulting in host lysis.

Iron precipitation and minor bacterial mineralization did not induce viral lysis.

Although lysogenic cells died (decrease in OD600 and number of cells noted using TEM), growth within the control microcosm (lysogen no iron) declined at the same rate as the iron addition microcosms. Viral lysis was likely caused by the stress induced upon the cells due to

112 the lack of nutrients (suspended in 96% AGW) and required growth conditions (temperature and aeration), not iron precipitation at the cell surface. The extensive production of exopolymeric substances (EPS) around cells, the appearance of spore-like structures (Fig. 5), and cell lysis with expulsion of SPβc2 particles (Fig. 6) may be a result of environmental stress. The degree of EPS production seemed to be an important factor in phage attachment.

Cellular regions with more extensive EPS appeared to have less SPβc2 particles near or attached to the surface than areas with no or partial EPS coverage with direct access to the cell surface (Fig. 3). As EPS has been shown to not be a reliable preventative mechanism against viral infection (as some phages are known to produce polysaccharide depolymerases;

Hughes et al. 1998; Deveau et al. 2002), it is possible that EPS has undergone iron precipitation inhibiting phage enzymes from obtaining access to the cell surface.

Exopolymeric substances have been shown to act as sites for mineral formation (Chan et al.

2004).

Unique microbial and/or phage biosignatures were not noted in this experiment.

Although iron precipitation influenced phage-host replication, no morphological signatures were evidence probably due to the minimal mineralization occurring at the cell surface.

The amount of dissolved iron in both the lysogen and B. subtilis-only microcosms were found initially to decrease then increase to almost original levels after approximately one day of incubation. The initial decrease in iron is partly due to ferric hydroxide precipitation; however over time iron becomes associated with dissolved organic matter

(DOM) created through viral lysis of host cells. Also noted was an increase in pH over the course of the experiment. This would be caused by the production of carbonic acid due to the chemoorganotrophic metabolism of B. subtilis.

113 It is interesting to note that the destruction of cells and resulting creation of cellular fragments (with a greater surface area) did not enhance mineralization as previously unexposed cellular functional groups are exposed to the mineralizing solution. Although iron seems to be associated with DOM, cell fragments did not act as an obvious site for mineral formation and growth, at least in the time provided in the experiment.

5.6 Conclusions

The prevention of SPβc2 replication due to iron precipitation lends strength to the hypothesis that bacterial mineralization dramatically hinders phage infection through the blockage of cellular receptor sites. In addition, partial bacterial mineralization does not cause a sufficient cellular stress to induce viral lysis. These findings suggests that within a mineralizing environment phage infection and replication rates would decrease a couple orders of magnitude making lysogeny a favorable replication cycle as bacterial mineralization does not cause viral lysis. If these results were applied to the AMD system in the study, IOB would significantly benefit from metabolically induced mineralization, and temperate phages would significantly benefit from lysogeny as finding a suitable host and achieving a successful replication cycle would be problematic. As all living organisms are believed to have an associated virus, phages must have adopted a response to ensure phage replication despite mineralized bacterial surfaces, and in this case, lysogeny may be that response.

5.7 Acknowledgments

This work was supported by a National Science and Engineering Research Council of

114 Canada Discovery Grant (FGF) and Postgraduate Scholarship (JEK), as well as a Geological

Society of America Student Research Grant (JEK).

We would like to thank Xstrata Canada, and in particular Joe Fyfe and Robin

Armstrong for their assistance in collecting AMD samples. Also, Wendy Abdi and Patricia

Wickham at the University of Ottawa for analyzing the DOC samples, and Dan Mathers at

Analyst, University of Toronto for analyzing the ICP-AOES samples collected in 2008. We would especially like to thank Dan Ziegler at the Bacillus Genetic Stock Centre for the generous donation and assistance with the Bacillus subtilis cultures.

5.9 References

Ackermann HW, DuBow MS. 1987. Natural groups of bacteriophages. In: Viruses of

Prokaryotes, vol II. Boca Raton, FL: CRC Press.

Baker BJ, Banfield JF. 2003. Microbial communities in acid mine drainage. FEMS

Microbiology Ecology 44:139-152.

Bernier L, Warren LA. 2005. Microbially driven acidity generation in a tailings lake.

Geobiology 3:115-133.

Beveridge TJ. 1989. Role of cellular design in bacterial metal accumulation and

mineralization. Ann Rev Microbiol 43:147-171.

Beveridge TJ, Murray GE. 1980. Sites of metal deposition in the cell wall of Bacillus

subtilis. Journal of Bacteriology 141:876-887.

Châtellier X, Fortin D. 2004. Adsorption of ferrous ions onto Bacillus subtilis cells. Chem

Geol 212:209-228.

Chan SC, De Stasio G, Welch SA, Girasole M, Frazer BH, Nesterova MV, Fakra S, Banfield

115 JF. 2004. Microbial polysaccharides template assembly of nanocrystal fibers. Science

303:1656-1658.

Chibani-Chennoufi S, Bruttin A, Dillmann ML, Brüssow H. 2004. Phage-host dynamics: an

ecological perspective. Journal of Bacteriology 186:3677-3686.

Cox JS, Smith DS, Warrant LA, Ferris FG. 1999. Characterizing heterogeneous bacterial

surface functional groups using discrete affinity spectra for proton binding.

Environmental Science and Technology 33:4514-4521.

Daughney CJ, Fein JB, Yee N. 1998. A comparison of the thermodynamics of metal

adsorption onto two common bacteria. Chem Geol 144:161-176.

Daughney CJ, Châtellier X, Chan A, Kenward P, Fortin D, Suttle CA, Fowle DA. 2004.

Adsorption and precipitation of iron from seawater on a marine bacteriophage (PWH3A-

P1). Marine Chemistry 91:101-115.

Deveau H, Van Calsteren MR, Moineau S. 2002. Effect of exopolysaccharides on phage-host

interactions in Lactococcus lactis. Applied and Environmenal Microbiology 68:4364-

4369.

Hughes KA, Sutherland IW, Jones MV. 1998. Biofilm susceptibility to bacteriophage attack:

the role of phage-borne polysaccharides depolymerase. Microbiology 144:3039-3047.

Ferris FG, Fyfe WS, Beveridge TJ. 1988. Metallic ion binding by Bacillus subtilis:

implications for the fossilization of microorganisms. Geology 16:149-152.

Fink, P.S. and Zahler, S.A. 2006. Temperate bacteriophages of Bacillus subtilis. In The

Bacteriophages. Edited by R. Calendar. P 557-571.

Flint, S.J., Enquist, L.W., Krug, R.M., Racaniello, V.R., and Skalka, A.M. 2000. Principles

of virology: Molecular biology, pathogenesis, and control. ASA Press, Washington, D.C.

116 Fortin D, Ferris FG, Beveridge TJ. 1997. Surface-mediated mineral development by

bacteria. In: Ribbe PH, editor. Geomicrobiology: Interactions between Microbes and

Minerals Vol. 35. Washington, DC: Mineralogical Society of America. P 161-177.

Gutterman B, Raya R, Kutter E. 2005. Basic phage biology. In: Kutter E, Sulakelidze A.,

editors. Bacteriophages: biology and applications. Baton Raton, FL: CRC Press. P 29-66.

Gonzáles-Toril, E., Llobet-Brossa, E., Casamayor, E.O., Amann, R., and Amils, R. 2003.

Microbial ecology of an extreme acidic environment, the Tinto River. Appl Environ

Microbio., 69:4853-4865.

Hemphil HE. 1990. Bacteriophages. In: Harwood CR, Cutting SM, Molecular biological

methods for Bacillus. John Wiley and Sons. P 451-478.

Holmfeldt K, Middelboe M, Nybroe O, Riemann L. 2007. Large variabilities in host strain

susceptibility and phage host range govern interactions between lytic marine phages and

their Flavobacterium hosts. Applied and Environmental Microbiology 73:6730-6739.

James RE, Ferris F.G. 2004. Evidence for microbial-mediated iron oxidation at a

neutrophilic groundwater springs. Chemical Geology 212:301-311.

Johnson DB. 1995. Selective solid media for isolating and enumerating acidophilic bacteria.

Journal of Microbiological Methods 23:205-218.

Kyle JE, Eydal HSC, Ferris FG, Pedersen K. 2008. Viruses in granitic groundwater from 69

to 450 m depth of the Äspö hard rock laboratory, Sweden. ISME Journal 2:571-574.

Kyle JE, Pedersen K, and Ferris FG. 2008. Virus mineralization at low pH in the Rio Tinto,

Spain. Geomicrobiology 25:338-345.

Lowenstam HA. 1981. Minerals formed by organisms. Science 211:1126-1130.

Mahmoud KK, Leduc LG, Ferroni GD. 2005. Detection of Acidithiobacillus ferrooxidans in

117 acid mine drainage environments using fluorescent in situ hybridization (FISH). Journal

of Microbiological Methods 61:33-45.

Mitchell AD, Ferris FG. 2005. The coprecipitation of Sr into calcite precipitates induced by

bacterial ureolysis in artificial groundwater: temperature and kinetic dependence.

Geochemica et Cosmochimica Acta 69:4199-4210.

Moebus K. 1987. Ecology of marine bacteriophages. In: Goyle SM, Gerba CP, Bitton B,

editors. Phage ecology. New York, NY: John Wiley & Sons. P 137-156.

Noble RT, Fuhrman JA. 1998. Use of SYBR Green I for rapid epifluroescence counts of

marine viruses and bacteria. Aquatic Microbial Ecology 14:113-118.

Paranchych W. 1966. Stages in phage R17 infection: The role of divalent cations. Virology

28:90-99.

Phoenix VR, Adams DG, Konhauser KO. 2000. Cyanobacteria viability during

hydrothermal mineralization. Chemical Geology 169:329-338.

Phoenix VR, Konhauser KO. 2008. Benefits of bacterial mineralization. Geobiology 6:303-

308.

Schrenk MO, Edwards KJ, Goodman RM, Hamers RJ, Banfield JF. 1998. Distribution of

Thiobacillus ferrooxidans and Leptospirillum ferrooxdians: Implications for generation

of acid mine drainage. Science 279:1519-1522.

Sharma PK, Das A, Hanumantha Rao K, Forssberg KSE. 2003. Surface characterization of

Acidithiobacillus ferrooxidans cells gown under different conditions. Hydrometallurgy

71:285-292.

Steensma HY, Blok J. 1979. Effect of calcium ions on the infection of Bacillus subtilis by

bacteriophage SF 6. Journal of General Virology 42:305-314.

118 Thingstad TF, Lignell R. 1997. A theoretical approach to the question of how trophic

interactions control carbon demand, growth rate, abundance, and diversity. Aquatic

Microbial Ecology 13:19-27.

Tijdens M, van de Waal DB, Slovackova H, Hoogveld HL, Gons HJ. 2008. Estimates of

bacterial and phytoplankton mortality caused by viral lysis and microzooplankton grazing

in a shallow eutrophic lake. Freshwater Biology 54:1126-1141.

Toporski J, Steele A, Westall F, Thomas-Keprta K, McKay D. 2002. The simulated

silicification of bacteria- new clues to the modes and timing of bacterial preservation and

implications for the search for extraterrestrial microfossils. Astrobiology 2:1-26.

Warren LA, Ferris FG. 1998. Continuum between sorption and precipitation of Fe(III) on

microbial surfaces. Environmental Science and Techology 32:2331-2337.

Westall F, Boni L, Guerzoni E. 1995. The experimental silicification of microorganisms.

Palaeontology 38:495-528.

Wightman PG, Fein JB. 2005. Iron adsorption by Bacillus subtilis bacterial cell walls.

Chemical Geology 216:177-189.

Yee N, Fowle DA, Ferris FG. 2004. A donnan potential model for metal sorption onto

Bacillus subtilis. Geochim Cosmochim Acta 68:3657-3664.

119 Table 1. Geochemical constituents and prokaryotic and viral abundances of AMD waters. All values are reported in mg/L unless otherwise stated. Fecunis 1 Fecunis 2 Fecunis 3 Longvac 1 Longvac 2 Longvac 3 Prokaryote* 0.63 1.39 3.35 6.84 1.05 4.69 VLP* 2.73 0.42 3.07 16.21 7.35 9.66 VPR 4.32 0.31 0.92 2.37 7.00 2.06 pH 2.91 3.40 4.00 2.45 3.60 2.50 Eh (mV) 664 678 662 660 660 660 Fetotal 1.60 2.28 1.28 1170 294 210 Fe2+ 0.41 0.35 0.22 196 105 44 2- SO4 900 600 937.5 2400 1800 1400 - NO3 n.d 2.30 1.70 n.d 69.60 62.40 3- PO4 n.d 0.00 0.00 n.d 0.01 0.10 DOC n.d 1.80 n.d n.d 4.60 5.00 Al 7.54 5.28 4.51 18.36 18.91 14.81 Ca 243.17 190.44 277.00 209.32 136.93 143.00 Cu 2.74 4.25 3.71 0.66 0.63 0.49 K 3.80 18.34 25.00 3.55 13.10 5.11 Mg 58.23 36.94 52.90 54.50 37.02 40.80 Mn 1.20 1.31 1.89 2.39 2.15 2.52 Na 166.99 162.24 140.00 14.32 12.32 7.39 Ni 11.96 11.24 14.2 12.76 13.2 12.10 Si 36.40 16.66 20.90 39.48 27.97 30.70 Zn 1.27 0.36 0.50 1.30 0.40 0.40 n.d. means not determined * abundance x 105 /mL

Table 2. Mean values of results for the mineralized bacteria plus phage microcosms. t = 0 min t = 45 min t = 90 min Control (no Fe) PFUs 491 319 149 pH 6.6 n.d. n.d. Fe addition PFUs 6 6 1 pH 5.8 6.0 6.0 Fetotal (mg/L) 0.15 0.08 0.10 PFUs = plaque forming units n.d. means not determined

120 Table 3. Mean values of measurements conducted in the lysogen plus iron microcosms. t = 0 hrs t = 2 hrs t = 17 hrs Control (noFe) pH 6.92 7.16 7.45 OD600 0.71 0.71 0.36 Fe addition pH 6.77 7.03 7.54 OD600 0.76 0.77 0.39 Fetotal (mg/L) 4.05 2.02 3.88

Table 4. Results of Bacillus subtilis plus iron over time t = 0 hrs t = 2 hrs t = 17 hrs pH 6.57 6.63 7.45 OD600 0.83 0.81 0.52 Fetotal (mg/L) 4.58 4.15 4.60

121 Figure legend

Figure 1. TEM image of mineralized Bacillus subtilis after 30 min incubation with iron and

SPβc2. The extent of visible mineralization was minimal. Image is not stained. Scale bars are 500 nm.

Figure 2. TEM image of lysogen with minimal mineralization (arrows indicates inorganic particles; 17 hours incubation; a) and extensive mineralization (2 hours incubation; b).

Images are not stained. Scale bar is 500 nm.

Figure 3. SPβc2 surrounds cells partially surrounded by ESP (a-d) noted using TEM. Image

(b) is a close up of SPβc2 particles (arrows) found in (a). Phages tend to surround cells with less EPS (c). All images are stained with uranyl acetate. Scale bar is 500 nm for (a, d), 250 nm for (b), and 1 µm for (c).

Figure 4. SPβc2 surrounds dividing Bacillus subtilis cell in a lysogenic culture. Image is not stained. Scales bar is 500 nm.

Figure 5. TEM image of spherical (spore?) noted after 20 hours of incubation of lysogen with iron. Image is stained with uranyl acetate. Scale bar is 500 nm.

Figure 6. TEM image of lysogen after 2 hours of iron incubation. Dark stain next to cell contains SPβc2 particles (arrow).

122

Figure 1.

123

Figure 2.

124

Figure 3.

125

Figure 4.

126

Figure 5.

Figure 6.

127 Chapter 6

Synthesis and Future Work

6.1 Synthesis

This research further expands our knowledge of aquatic ecology and microbial geochemistry through the discovery and description of viruses in the deep subsurface and in acid mine drainage. This research also provides important information on the influence that geochemical variables (i.e. pH, chloride) exert on viral abundance. Additional discoveries that resulted from this research further our understanding of viral-geochemical interactions, viral-mineral interactions, the protective benefit of bacterial mineralization, phage-host dynamics, and the role phages play in the field of geomicrobiology and viral ecology.

6.1.1 Viruses in extreme environments

This research has revealed that viruses (more specifically bacteriophages) are present and abundant in the deep subsurface and AMD environments (up to 107 virus-like particles/mL), two environments that previously went unexamined in terms of virology. VLP abundance is strongly correlated with prokaryotic abundance (r = 0.91) in the deep subsurface but not in AMD environments of the Rio Tinto (r = 0.38).

In terms of viral morphology, polyhedral and tailed morphotypes dominated both environments, with polyhedral morphotypes more common in AMD environments.

Pleomorphic (i.e. Archaeal morphotypes) phages are more abundant in the deep subsurface with morphotype diversity decreasing with depth (these morphotypes were rare in AMD).

The overall viral morphotype diversity within the deep subsurface decreases somewhat with increasing salinity.

128

6.1.2 Viral control of prokaryotic abundance

Another important outcome of this research is the discovery that viral abundance is the strongest predictor of prokaryotic abundance and not parameters that strongly influence the prokaryotic growth (i.e. phosphate and pH) in freshwater and AMD environments in southern Ontario. This suggests that viruses exert significant control over the total host density possibly preventing uncontrollable growth in nutrient-rich environments. This viral control is suggested to occur in the deep subsurface where prokaryotic abundances are kept at a steady state with numbers in the range of 104 to 106 cells/mL.

6.1.3 Viral-mineral and viral-geochemical interactions

This research revealed for the first time, to our knowledge, (i) electron microscopic of viral-mineral interactions in natural aquatic environments, (ii) viral participation in mineralization events, and (iii) strong geochemical relationships with viruses within AMD and freshwater environments.

Transmission electron microscope photomicrographs coupled with EDS of samples collected from the Rio Tinto in Spain show viral particles attached to iron-bearing minerals.

X-ray diffraction of the sediments revealed that jarosite is the dominant iron-bearing mineral, although iron oxides, such as ferrihydrite and goethite are also noted. Geochemical and statistical calculations of the chemical, physical, and microbial constituents of Rio Tinto and

Ontario survey samples found that viral abundance is negatively correlated with jarosite saturation states (r = -0.71 and rs = -0.33, respectively) revealing that as the saturation state of jarosite increases, viral abundance decreases. In addition, regression analysis on Ontario

129 water samples indicated goethite saturation indices are the strongest predictor of viral abundance and explained 78% of the variability in the data. This suggests that as pH increases viral particles remain attached to goethite until attractive electrostatic forces are overcome by repulsive forces. Moderate saturation index correlations with VLPs are also noted with Al(OH)SO4 in the Rio Tinto, and with ferrihydrite and pyrolusite in the Ontario water samples. Strong correlations are also noted between viral abundance and geochemical parameters such as pH and Eh in the Rio Tinto (r = 0.94 and r = -0.89, respectively) and

Ontario (rs = 0.58 and rs = -0.43, respectively) samples revealing the negative impact of

AMD environments on viral abundance.

In the deep subsurface, correlations were noted between VLP abundance and chloride concentration; however, this may be an artifact of an inverse relationship existing between chloride and prokaryotic abundance, as VLP-prokaryotic correlations are strong (r = 0.91).

Whether or not VLP abundance has a true correlation with chloride, both VLPs and prokaryotes are likely be influenced by the ionic strength of the water. Increased ionic strengths would increase VLP and prokaryotic attachment to inorganic surfaces as the electrostatic double layer would decrease. Also, old saline groundwater may be less favorable for prokaryotic growth (and subsequent phage production).

6.1.4 Role of bacterial mineralization in phage replication

Iron precipitation and/or the binding Fe(OH)2 ions to host cellular surfaces was found to drastically hinder phage attachment and subsequent replication. Bacterial mineralization and/or the binding of metal ions to receptor sites would be advantageous such that it provides almost complete protection towards viral infection and replication. In addition, bacterial

130 mineralization does not seem to be a strong environmental stressor as lysogenic cells were not induced into a lytic cycle. Given phages are seemly present wherever bacteria are present, phages appear to have developed a response that enables the replication of progeny where host cells are undergoing mineralization (i.e. acid mine drainage, terrestrial hot springs). The response may be to undergo a lysogenic replication cycle.

Bacterial mineralization and lysogenic replication cycles are suggested as potential causes of low VLP-prokaryotic abundance correlations in the Rio Tinto (r = 0.38). Inverse prokaryotic abundance correlations with mineral saturation indices of jarosite (r = -0.72 to -

0.75), iron oxyhydroxides (r = -0.69 to -0.72), and Al(OH)SO4 (r = -0.72) indicates that bacterial mineralization and/or bacterial attachment to precipitating mineral phases is likely occurring. The results above provide evidence that if bacterial surface receptors were blocked, phage attachment and subsequent replication would be strongly inhibited. As a method of phage survival lysogenic replication cycles may be critical when there is a lack of host receptor sites and where there are acidic and strongly oxidizing AMD conditions.

6.2 Future Work

The results of this research offer multiple exciting avenues in which further research could be pursued. Potential avenues for future investigation include (i) using phages as a potential bioremediation technique in environmentally problematic areas, (ii) further examining phage-host dynamics under mineralizing conditions, and (iii) examining viral participation in mineralization events to determine the likelihood of viral preservation within the rock record. The exploration of viral-mineral and viral-geochemical interactions is almost endless as this field only contains less than a handful of scientists.

131

6.2.1 Environmental phage therapy in acid mine drainage

Given the parasitic nature of phage-host relationships, the potential to use phages as a bioremediation technique where upon bacteria are causing or propagating an environmental problem exists. One investigative study would be to study the potential to use bacteriophages to remediate AMD. It is well known that iron oxidizing bacteria (IOB) propagate the problem of AMD through the oxidation of ferrous iron and the generation of hydrogen ions, leading to further acidification and increased levels of dissolved metals in the water. If one could isolate a phage that infects and destroys IOB, this may be a potential avenue to extinguish a causative agent in AMD. Attempts at isolating an IOB phage could be repeated with the use of different growth mediums to determine if the presence of divalent cations and/or trace elements assists in the phage attachment and replication.

6.2.2 Phage-host dynamics under mineralizing conditions

Extension of the Bacillus subtilis- SPβc2 experiments could be conducted using different final iron concentrations to determine (i) the minimal amount of iron is required to hinder phage replication on mineralized B. subiltis, (ii) whether the degree of iron precipitation onto a lysogen results in viral induced lysis, and (iii) how to induce phage mineralization and the resulting phage-host interaction. All experiments could be monitored using a TEM (whole mounts and thin sections made from resin embedded samples) to identify unique morphological biosignatures are produced from phage-host interactions under authengenic mineral precipitation.

6.2.3 Viral mineralization and preservation

132 The research and results shown in this thesis clearly reveal that viruses are participating in mineralization events. The extent of this participation and long term preservation potential of viral-mineral interaction has as yet to be determined. Experiments could be conducted using progressive greater concentration of iron (used to form iron hydroxides, i.e. ferrihydrite) and silica (used to form opal-A) additions to phage stocks over time and under different geochemical conditions (i.e. pH). Samples would be preserved and embedded to enable thin sectioning though viral-mineral aggregates. This would enable high resolution, in depth examination of the interface between the viral capsid and mineral surface, and how this interface may differ over time.

High resolution microscopy of samples collected throughout the duration of the experiment could be used to determine if unique morphological signatures are produced and how the morphological signatures develop through time. Thin sections made from recent rock deposits formed under conditions in which viruses are known to undergo mineral sorption (i.e. AMD) could be examined for similar morphological signatures noted during the experiment.

133 Appendix I

Exact microbial and VLP counts for Rio Tinto and Ontario samples

Table 1. Values of VLP and prokaryotic abundance of Rio Tinto samples measured using epifluorescence microscopy. Mean value of duplicate measurements reported in chapter 3 unless only one value attained.

Sample Prokaryote Abundance (x105/mL) VLP Abundance (x105/mL) Measurement 1 Measurement 2 Measurement 1 Measurement 2 Source 2.5 n.d. 10.2 n.d. Ravine a 1.9 n.d. 6.3 n.d. Ravine b 0.6 n.d. 0.2 n.d. Train Stop 1.8 1.5 2.0 2.7 Berrocal 5.8 5.9 0.2 0.05 Valverde 1.8 2.0 0.3 1.5 Niebla 3.7 5.6 1.8 2.1 n.d. means not determined

Table 2. Values of VLP and prokaryotic abundance of Ontario water samples measured using epifluorescence microscopy. Mean value of duplicate measurements reported in chapter 4 unless only one value attained.

Sample Prokaryote Abundance (x105/mL) VLP Abundance (x 105/mL) Measurement 1 Measurement 2 Measurement 1 Measurement 2 Sudbury Igneous Complex

Fecunis 1 0.4 0.4, (1.3) 0.8 0.8, (1.4) Fecunis 2 1.2 1.2, (1.1) 0.4 0.5, (0.4) Fecunis 3 2.7 3.1, (2.7) 2.2 1.9, (1.6) Longvac 1 5.6 6.8, (9.5) 11.7 17.9, (23.5) Longvac 2 1.6 1.0, (0.6) 8.8 7.7, (5.5) Longvac 3 2.9 4.7, (4.9) 6.4 9.5, (12.1)

St. Lawrence Lowlands

Cedarvale1 1.4 1.4 38.2 37.3

134 Cedarvale2 1.4 2.4 34.4 61.1 Cedarvale3 8.8 3.9 192.4 81.5 Cedarvale4 4.1 8.8 30.0 59.0 Lake Ontario Sailing Club 3.2 n.d. 27.4 n.d. Sunnyside Park Lake Ontario 2.4 3.3 36.8 49.8 Port Hope 4.0 3.6 27.6 46.9 Prince Edward Point 3.3 n.d. 31.9 n.d. Boyne River 1.4 n.d. 41.1 n.d. Osprey Wetland 1.7 1.3 20.2 15.4 Beaver River 0.9 0.7 19.4 23.9 Nottawasaga 1.5 20.1 Bay 1.0 26.1 Nottawasaga River 1.3 2.1 27.1 34.1 Willow Creek 1.0 n.d. 26.8 n.d. Minesing 1.1 24.4 Swamp 0.9 42.1 Highland Creek 0.5 0.9 41.9 30.9 Rouge River 1 3.6 n.d. 60.3 n.d. Rouge River 2 1.7 6.1 33.7 43.1 Rouge River 3 4.2 1.6 72.4 41.3 Lake Scugog 2.8 2.1 35.6 31.7 East Cross Creek 2.4 1.0 12.9 25.2 Scugog River 2.9 2.3 38.9 40.0 Sturgeon Lake 2.4 3.1 45.9 31.6

Grenville Province

Lake Couchinsing 4.6 2.4 83.8 59.7 Severn River 3.9 4.9 55.7 61.3 Muskoka Bay 3.1 n.d. 41.8 n.d. Muskoka River 3.9 4.1 25.7 33.3 Black River 2.5 2.0 36.6 31.5 Kahshe Lake 2.3 3.5 20.5 34.5 Sparrow Lake 3.3 3.6 47.1 59.3 Lake Bernard 1.2 1.4 21.5 18.3 Horn Lake 1.8 2.5 18.2 22.4 S Horn Lake Rd 1.2 1.8 22.9 22.2 Cecebe Lake 4.3 1.9 23.2 21.3 Magnetawan River 4.0 1.9 31.9 19.7

135 Wetland near Mayfield Lake 2.8 3.8 22.7 27.0 Mary Lake 1.5 n.d. 14.5 n.d. Stony Lake 6.8 n.d. 89.2 n.d. York River 2.0 2.7 17.4 14.5 Little Mississippi n.d. n.d. River 1.2 18.8 Denbigh Lake 4.1 6.0 26.3 52.8 Upper Mazinaw Lake 1.3 n.d. 10.6 n.d. Little Skootamatta River 2.8 3.4 44.6 45.1 n.d. means not determined values in brackets means triplicate sample counted

136 Appendix II

Evidence of strength of multiple regression model in Chapter 4 where the dependent variable

is prokaryotic abundance.

Figure Legend

Figure 1. Histogram of raw residuals of results for multiple regression analysis where the dependent variable is prokaryotic abundance.

Figure 2. Normal probability plot of raw residuals of results for multiple regression analysis where the dependent variable is prokaryotic abundance.

Figure 3. Graph of predicted vs. observed values of results for multiple regression analysis where the dependent variable is prokaryotic abundance

137 25

20

15 No. 10 of obs.

5

0 -5 -4 -3 -2 -1 0 1 2 3 X <= Category Boundary Figure 1.

3.0

2.5 .99 2.0 .95 1.5

1.0 .75 0.5 .55 0.0 .35 -0.5

Expected -1.0 .15 -1.5 .05

Normal -2.0 .01 -2.5 Value -3.0 -4 -3 -2 -1 0 1 2 3 Residual Figure 2.

138 10

9

8

7

6

5

Predicted 4

3 Values 2

1

0 0 1 2 3 4 5 6 7 8 9 O bserved Values Figure 3.

139 Appendix III

Evidence of strength of multiple regression model in Chapter 4 where the dependent variable

is VLP abundance.

Figure Legend

Figure 1. Histogram of raw residuals of results for multiple regression analysis where the dependent variable is VLP abundance.

Figure 2. Normal probability plot of raw residuals of results for multiple regression analysis where the dependent variable is VLP abundance.

Figure 3. Graph of predicted vs. observed values of results for multiple regression analysis where the dependent variable is VLP abundance.

140 14

12

10

8

No. 6 of obs. 4

2

0 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 X <= Category Boundary Figure 1.

3.0

2.5 .99 2.0 .95 1.5

1.0 .75 0.5 .55 0.0 .35 -0.5

Expected -1.0 .15

-1.5 .05 Normal -2.0 .01 -2.5 Value

-3.0 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 Residual Figure 2.

141 45

40

35

30

25 Predicted

20 Values

15

10 -20 0 20 40 60 80 100 120 Observed Values Figure 3.

142