Spectro-Microscopic Studies of Microbial Selenium and Iron Reduction in a Metal Contaminated Aquifer

By

Sirine Constance Fakra

A dissertation submitted in partial satisfaction of the

requirements for the degree of

Doctor of Philosophy

in

Earth and Planetary Science

in the

GRADUATE DIVISION

of the

UNIVERSITY OF CALIFORNIA, BERKELEY

Committee in charge:

Professor Jillian F. Banfield, Chair

Professor James Bishop

Professor Céline Pallud

Spring 2015

Spectro-Microscopic Studies of Microbial Selenium and Iron Reduction in a Metal Contaminated Aquifer

© 2015

By

Sirine Constance Fakra

Abstract

Spectro-microscopic studies of microbial selenium and iron reduction in a metal contaminated aquifer

by

Sirine C. Fakra

Doctor of Philosophy in Earth and Planetary Science

University of California, Berkeley

Professor Jillian F. Banfield, Chair

Redox-sensitive metal contaminants in subsurface environments can be reduced enzymatically or indirectly by microbial activity to convert them from soluble mobile (toxic) to comparatively insoluble, relatively immobile (less bioavailable) forms. The broad purpose of the research presented in this dissertation was to acquire a deep understanding of selenium and iron microbial reduction and immobilization in the subsurface and to characterize in detail the nature of the bioreduction products. To this end, biofilms formed during a biostimulation experiment in a metal-contaminated aquifer adjacent to the Colorado River in Colorado, USA were studied. Biofilms develop in a wide variety of natural settings and the aqueous chemical conditions within biofilms are strongly affected by the presence of extracellular polymers that potentially confer biofilm cells with a greater tolerance to heavy metals than planktonic cells.

This thesis integrates field and laboratory experimental methods to provide 2D and 3D ultrastructural information, 2D chemical speciation and community membership via metagenomics methods. In addition, physiological information was obtained via characterization of an isolated bacterium and insights related to the product structure and stability were achieved by chemical synthesis-based studies. In this dissertation, an apparatus permitting correlative cryogenic spectro-microscopy was developed (Appendix I) and applied to determine in detail the cell-mineral relationships and the speciation of selenium in the biofilms (Chapter 1). The research involved integration of both cryogenic electron microscopy and X-ray absorption spectroscopy datasets on the same sample region to document the size, structure and distribution of bioreduction products. Because many of the microbial species in the mine tailings-contaminated aquifer are novel and difficult to cultivate in the laboratory, part of the research involved phylogenic analyses of the biofilm organisms via analysis of 16S rRNA genes. A novel betaproteobacterium of the genus Dechloromonas (Dechloromonas selenatis strain RGW, Chapter 2) was isolated from the Rifle site and shown to be capable of reducing selenate to red amorphous elemental Se0. This isolate was also capable of reducing toxic arsenate. Chapter 3 investigates further the stability of elemental selenium colloids at ambient pressure as a function of temperature and particle size. The last chapter (Chapter 4) focuses on the distribution and speciation of iron in the Rifle aquifer

1

during a biostimulation experiment. The combined results demonstrate the importance of both clays and cell-associated ferric iron oxyhydroxide aggregates for growth of planktonic iron-reducing bacteria.

These insights provide fundamental information about organisms that mediate selenium, iron and arsenic biogeochemical transformations in the subsurface and the nature of the product phases. The data may help to identify substrate amendment regimes for sustained Se remediation. Following short-term acetate addition to the aquifer, selenium remained immobile for at least one year, suggesting the acetate amendment approach has significant potential for bioremediation of selenium, in addition to uranium and vanadium as previously studied. Although focused on selenium and iron bio-reduction, the instrumentation and approaches developed here are generally applicable for accurate determination of cell-mineral interactions and metal speciation and can be further extended to constrain aquifer-scale reactive transport models in a wide range of environments.

2

To my brother

i

TABLE OF CONTENTS

ACKNOWLEDGEMENTS

INTRODUCTION

CHAPTER 1. Correlative cryogenic spectro-microscopy to investigate selenium bioreduction products.

CHAPTER 2. Dechloromonas selenatis, a Betaproteobacterium from a contaminated aquifer that reduces selenate to amorphous selenium.

CHAPTER 3. Size and temperature-dependent crystallization of elemental selenium.

CHAPTER 4. Iron speciation analysis indicates the use of clays and iron oxyhydroxides by planktonic- and biofilm-associated Fe-reducing bacteria.

APPENDIX I. Microprobe cryogenic apparatus for correlative spectro-microscopy.

APPENDIX II. Supplemental materials for Chapter 1.

ii

ACKNOWLEDGEMENTS

As a part-time employee and graduate student, there are many people who have helped me these past years and without whom this PhD would simply not have been possible. First, to my graduate advisor Jillian Banfield who believed in me more than I did. She has been an incredible mentor. Her stimulating ideas and her vision are truly inspiring. To Howard Padmore who has made this whole journey possible, I am deeply and forever grateful. Both Jillian and Howard have supported me in more ways than I can count.

I would like to thank my committee members, Jim Bishop and Céline Pallud, for taking the time to learn about my research and provide insightful feedback.

I am indebted to the Advanced Light Source’s director Roger Falcone and the Berkeley Lab Learning Institute for financing me and for supporting a program where full time employees can earn a doctorate degree.

I am very grateful to members of the Banfield’s group in particular Birgit Luef for teaching me the art of cryo-plunging samples and cryogenic electron microscopy, and who helped me a great deal with cryo-TEM data. I want to thank Sean Mullin, Cindy Castelle and Laura Hug who taught me the basics of microbiology and have greatly helped me with phylogenetic trees, Ken Williams for introducing me to the Rifle site and answering my tons of questions patiently. I would like to acknowledge Denise Schines for providing me with confocal data. Thanks to Roseann Csencsits, Kelly Wrighton, Luis Comolli, Kim Handley and Tyler Arbour for stimulating discussions over the years, as well as Margie Winn for administrative support.

A big thanks to my Berkeley Lab colleagues, Matthew A. Marcus and Tony Warwick who always encouraged me and have been great mentors over the years; Tolek Tyliszczak, Mary K. Gilles and David K. Shuh who have always supported me; Jeff Kortright, Tony Young, Andrew Westphal and Anna Butterworth who lent me some important pieces of equipment at crucial times. A large thank you to Paul Baker and his team at Instec Inc. for help with the microprobe cryo-stage.

Last, to my parents, who have always encouraged and supported me the best way they could. To my beautiful and smart brother, you are always on my mind, this one is for you.

Finally to my husband, I love you so very much. I would not have made it without your unconditional love and support.

iii

Introduction

1.1 Introduction and motivation

One of the most active current research in environmental science focuses on the

bioremediation of contaminated environments. Centuries of anthropogenic activities

have led to the accumulation of metal and metalloid contaminants into the environment,

posing a direct threat to ecosystems and human health. By contrast to organic

contaminants, heavy metals are not biodegradable and remain in the environment1.

Their toxicity mostly depend on their forms, with the general rule that the more soluble

they are, the more toxic2.

Selenium, present in trace amounts in rock-forming minerals is a major

environmental contaminant, present in the porphyry copper deposits of the western

United States and around the world3. Globally, the largest fluxes in the Se cycle are

from land into the marine system along aquatic pathways. Natural trace Se

contamination occurs mostly through geochemical processes, such as erosion of soils

and weathering of rocks (e.g. black shales)4, 5. However, the anthropogenic release is

by far the major contributor to the Se cycle, releasing up to 88,000 tons of Se per year5.

Oil refining, combustion of fossil fuels, drainage from mines, and agriculture represent

the primary sources of contamination6.

The biogeochemical cycling of selenium7, is still not well defined but is

predominantly governed by microorganisms which play a crucial role in oxidation,

reduction, methylation, and volatilization. Se oxyanions (selenate and selenite) which

iv

dominate in aqueous systems can be reduced by microbes (enzymatically, or indirectly)

to comparatively insoluble, immobile and non-toxic forms (e.g. Se0). The research

described in this dissertation stems from the Rifle Field Study and explores the potential

for the stimulation of microorganisms at reducing and controlling the mobility of Se

(and Fe) in the subsurface.

1.2 Contribution of this thesis and outline of the dissertation

Determining accurately the distribution and forms of associated metals, as well as

understanding the functioning of the microbial community present in contaminated

systems are key to formulate strategies to bio-remediate affected areas. The

corresponding geo-microbiological materials are complex heterogeneous multi-phase

systems, generally aggregated, poorly crystalline and poorly concentrated, rending

their analyses by traditional techniques (e.g. X-ray diffraction) challenging. One of the

major contribution brought by this doctoral thesis is the development and application

of a cryogenic apparatus allowing the correlation of electron microscopy and X-ray

absorption spectroscopy datasets to precisely decipher the distribution and speciation

of metals and metalloids in the Rifle contaminated aquifer.

This dissertation is organized as publishable units (chapters) with the following

outline. Chapters 1 and 2 describes the application of a novel cryogenic apparatus

(described in Appendix I) to study selenate bioreduction products in intact mixed

bacterial biofilms and in a newly isolated organism, respectively, obtained from the

Rifle aquifer. Chapter 3 investigates further the stability of red amorphous elemental

v

selenium colloids. Chapter 4 focuses on the distribution and speciation of iron in the

aquifer during a biostimulation experiment and demonstrates the importance of the

mineral colloidal fraction (phyllosilicates and oxides) for the growth of planktonic iron-

reducing bacteria.

References

1. Lovley, D., Bioremediation of organic and metal contaminants with dissimilatory metal reduction. Journal of Industrial Microbiology 1995, 14, (2), 85-93. 2. Lovley, D. R., Dissimilatory Metal Reduction. Annual Review of Microbiology 1993, 47, (1), 263-290. 3. Dungan, R. S.; Frankenberger, W. T., Microbial Transformations of Selenium and the Bioremediation of Seleniferous Environments. Bioremediation Journal 1999, 3, (3), 171-188. 4. Frankenberger, J. W. T.; Arshad, M., Bioremediation of selenium-contaminated sediments and water. BioFactors 2001, 14, (1), 241-254. 5. W.T. Frankenberger, J., and S.M. Benson, Selenium in the environment. Marcel Dekker, Inc. ed.; Taylor & Francis: N.Y.C., N.Y., 1994. 6. Santolo, O. H. M. a. G. M., Kesterson Reservoir-Past, Present, and Future: An Ecological Risk Assessment. In Selenium in the Environment, W.T. Frankenberger, J., and S.M. Benson, Ed. Marcel Dekker: 1994; pp 69-117. 7. Shrift, A., A Selenium Cycle in Nature? Nature 1964, 201, (4926), 1304-1305.

vi

CHAPTER 1

Correlative Cryogenic Spectro-Microscopy to Investigate

Selenium Bioreduction Products

1

Abstract

Accurate mapping of the composition and structure of minerals and associated biological

materials is critical in geomicrobiology and environmental research. Here we developed an

apparatus allowing correlation of cryogenic transmission electron microscopy (TEM) and

synchrotron hard X-ray microprobe (SHXM) datasets to precisely determine the distribution,

valence state and structure of selenium in biofilms sampled from a contaminated aquifer near Rifle,

CO, USA. Results were replicated in the laboratory via anaerobic selenate-reducing enrichment

cultures. 16S rRNA analyses of field-derived biofilm indicated the dominance of

Betaproteobacteria from the Comamonadaceae family, and uncultivated members of the

Simplicispira genus. The major product in field and culture-derived biofilms is ~25-300 nm red

amorphous Se0 aggregates of colloidal nanoparticles. Correlative analyses of the cultures provided

direct evidence for microbial dissimilatory reduction of Se(VI) to Se(IV) to Se0. Extended X-ray

absorption fine structure spectroscopy showed red amorphous Se0 with a first shell Se-Se

interatomic distance of 2.339 ± 0.003 Å. Complementary scanning transmission X-ray microscopy

revealed that these aggregates are strongly associated with a protein-rich biofilm matrix. These

findings have important implications for predicting the stability and mobility of Se bioremediation

products and understanding of Se biogeochemical cycling. The approach, involving correlation of

cryo-SHXM and cryo-TEM datasets from the same specimen area, is broadly applicable to

biological and environmental samples.

Introduction

2

Deciphering the roles of microbial processes in biogeochemical transformations is

important for many bio-mineral systems and environmental research in general. Characterization

of field samples is necessary because relevant pure cultures or characterized mixed microbial

communities can be difficult to obtain, and can miss critical aspects that are key to environmental

processes. Damage to biological materials can occur by shrinkage, breakage or loss of subcellular

and extracellular structures. Furthermore, examination of dry or freeze-dried samples can lead to

denaturation of proteins, alteration of the mineral structure and loss of spatial resolution at the cell

surface-nanoparticle interface, yielding incorrect interpretations1.

Most bio-mineral samples are complex, of variable chemical composition at the nano- and

micro-scales, and often contain poorly-ordered materials in low concentration. Synchrotron hard

X-ray microprobes (SHXM) are ideally suited to characterize such systems, and provide metal

distribution through X-ray fluorescence mapping (µXRF), valence state, local atomic structure via

X-ray absorption spectroscopy (µXAS) and phase by X-ray diffraction (µXRD)2. X-ray damage

through metal photo-reduction/oxidation and amorphization has been well documented and is

especially acute in organics-containing samples3-5. Selenium (Se) compounds in particular are

prone to X-ray damage as illustrated by the breaking of the Cγ—Se bond in selenomethionine

(SeMet)6, selenate photoreduction7 and selenite photo-oxidation8 or just simply reported such as

in the case of Se-contaminated biological wastewater samples9. Additionally, organic-bound

metals and low-Z containing molecules are often difficult to analyze at room temperature because

thermal vibrations damp the amplitude of extended X-ray absorption fine structure (EXAFS)

oscillations, especially at high photo-electron wavenumbers. These issues are exacerbated in dilute

and/or poorly ordered bio-minerals where longer beam exposure is usually required for adequate

3

signal-to-noise ratios. To alleviate these effects, measurements are most often carried out at LN2

(77 K) or less frequently at LHe (4 K) temperatures. In the case of selenium, prior Se K-edge

cryo-EXAFS studies of amorphous and crystalline Se10 demonstrated that the Debye-Waller

thermal disorder component is only marginally reduced below 100 K, suggesting that the use of

LHe (versus LN2) would not significantly improve the EXAFS signal. Lastly, cryogenic analyses

should allow better detection of any potentially volatile selenides produced.

There is a strong incentive to link cryogenic SHXM with cryogenic transmission electron

microscopy (cryo-TEM11). Cryo-TEM provides ultra-structural information at 2-4 nm resolution12.

Whole frozen hydrated intact bacteria embedded in amorphous (vitreous) ice can be analyzed,

eliminating artifacts associated with traditional fixation and dehydration methods or sectioning1.

The sample is generally cooled with LN2 (vs. LHe13), and low dose imaging mode is used to

minimize electron damage and preserve sample structures in a “near-native” state. However, cryo-

TEM provides limited chemical information on thin samples (≲ 750 nm) within a narrow field of

view (100’s nm). On the other hand, SHXM is effective at characterizing, at the micron scale,

millimeter-scale areas of poorly concentrated samples (10’s of ppm). Ultimately correlating cryo-

TEM with cryo-SHXM datasets of the same sample region allows to link sub-nm-scale structural

information to crucial chemical speciation data. Here we use this approach to investigate the

distribution and speciation of selenium in biofilms from an unconfined aquifer adjacent to the

Colorado River, near Rifle, CO, USA. The shallow groundwater has residual metal contamination

(U, V, As and Se) at tens of µM levels due to past ore-milling activities. These levels exceed U.S.

E.P.A. drinking water standards (100 nM for Se). To date, the potential for acetate amendments

into the subsurface to stimulate uranium bioreduction has been extensively studied14-17 whereas

4

microbiological and geochemical processes controlling selenium mobility, thus toxicity at the site

and others similar, are poorly understood.

Se occurs predominantly in four oxidation states (VI, IV, 0 and –II). Se oxyanions, selenate

(Se(VI), SeO42-) and selenite (Se(IV), SeO3

2−), are toxic at ppm concentrations and known to

bioaccumulate in the food chain causing significant ecological damage18. The biogeochemical

cycle of Se in nature is not well defined19 but is predominantly governed by microorganisms20. An

important part of this cycle is the dissimilatory21 reduction of Se oxyanions (DSeR) by anaerobes22.

These microorganisms couple the oxidation of organic matter (or H2) to the reduction of Se

oxyanions, forming either relatively insoluble non-toxic Se0 or reactive and toxic selenide Se(-II).

Dissimilatory selenate reduction to Se0 is a major sink for Se oxyanions in anoxic environments22,

23. Although phylogenetically diverse selenite-reducing bacteria have been well characterized,

relatively little is known about selenate-reducing bacteria. On the other hand, the stability,

reactivity and bioavailability of Se0 colloids are still not well understood, and likely depend

strongly on the size, morphology and allotropic form of Se. A handful of studies have reported

microbial reduction of Se oxyanions to Se0, mostly as red amorphous24-28 (primarily chains) and

red crystalline monoclinic29 (Se8 rings). However, few provide direct evidence of the structure of

the Se allotropes produced, especially from field-preserved samples.

Previously at the Rifle site, microbial reduction of Se oxyanions was detected during a

biostimulation experiment30, but many questions remained about the form(s) and distribution of

the products. Here we developed a cryo-stage that allows the transfer of cryogenically preserved

Se-rich biofilms between a TEM and a SHXM, enabling essentially artifact-free ultra-structural

biological and chemical information from the same sample region. We used this approach to

5

examine in detail the spatial distribution and chemical speciation of selenium in samples obtained

through field experimentation at the Rifle site and cultivation.

Materials and Methods

Additional materials and methods can be found in Appendix II.

Biofilm samples from the Rifle site. The samples were collected during the “Super 8”

uranium bio-stimulation field experiment17 (August-September 2010) at the Rifle Integrated Field

Research Challenge site adjacent to the Colorado River (Western Colorado, USA). The site is

located on a relatively low-lying alluvial terrace created by a floodplain meander of the Colorado

River and is described extensively elsewhere14, 16. The shallow, unconfined aquifer consists of

alluvial sands, silts, and gravels; details on the geochemistry and mineralogy can be found in prior

studies15, 16. As previously described17, the groundwater was amended with sodium acetate and

injected into the subsurface at various depths. Acetate (CH3COO-) served as a carbon source and

electron donor over the course of the 25-day amendment period. Biofilm samples (CG02 well, 4

m depth, 5 mM acetate) were collected 16 days after injection of acetate to the anoxic aquifer. No

geochemical sampling exist at the acetate injection well CG02 and the closest down-gradient

monitoring well CD01 is used as reference. For all analyses except CLSM, samples were collected

by scraping the biofilm off the injection tubing (polyethylene, HDPE) used to circulate the acetate-

amended groundwater (pH~7.2). The biofilms were uniformly distributed across the tubing, with

minimal O2 (from diffusion across the tubing wall) and low nitrate level (~10 µM average).

6

Samples were flash-frozen directly in the field, to preserve their physical and chemical integrity,

using a portable cryo-plunger31 and procedures described in Appendix II.

2D cryo-TEM Cryo-TEM images were acquired with a JEOL–3100-FFC electron

microscope (JEOL Ltd, Akishima, Tokyo, Japan) equipped with a FEG electron source operating

at 300 kV, an Omega energy filter (JEOL), an LN2-cooled sample transfer stage (80 K), and a

Gatan 795 4K×4K CCD camera (Gatan Inc., Pleasanton, CA, USA) mounted at the exit of an

electron decelerator held at a voltage of 200 to 250 kV. Survey of the grids and selection of suitable

targets were performed in low dose defocused diffraction mode. Images were recorded at different

magnifications with a pixel size of 0.56 and 0.701 nm at the specimen. Several images were

recorded with a 2K×2K CCD camera instead, with a pixel size of 0.69, 0.92 and 1.2 nm at the

specimen. Underfocus values ranged from 12± 0.5 µm to 15± 0.5 µm, and energy filter widths

were typically around 30 eV. Over 100 and 70 images were recorded to evaluate the morphology

and size of cells and colloidal particles in field and culture-derived biofilms respectively.

X-ray microprobe. Micro-focused X-ray fluorescence (µXRF) mapping, X-ray

diffraction (µXRD) and Se K-edge X-ray absorption spectroscopy data were collected at the

Advanced Light Source (ALS) bending magnet beamline 10.3.2 (2.4- 17 keV) with the storage

ring operating at 500 mA and 1.9 GeV32. All data were recorded at 95 K using a cryo-stage

described below. Maps and µXRF spectra were collected at 13 keV with a beam spot size ranging

from 2×2 μm to 5×5 μm, and counting times up to 200 ms/pixel. Fluorescence emission counts

were recorded using a seven-element Ge solid-state detector (Canberra) and XIA electronics. Se

K-edge µXANES spectra were recorded in fluorescence mode by continuously scanning the Si

(111) monochromator (Quick XAS mode) from 160 eV below up to 407 eV above the edge 7

(12500-13067 eV, i.e. up to k = 10 Å-1). EXAFS spectra were recorded up to 740 eV above the

edge (12500-13400 eV, i.e. up to k ~ 13.7 Å-1). Spectra were calibrated using the white line of a

red amorphous Se standard set at 12660 eV. All data were processed using LabVIEW custom

software and standard procedures described elsewhere33. To rapidly survey the valence state of

selenium in the samples, valence state scatter plots were generated from XANES data using a

spectral database of Se compounds (see Table S1 and Appendix II). Further, least-squares linear

combination fitting of the XANES spectra was performed as previously described34. Micro-

EXAFS spectra of cultures were reduced with 2 ( )k kχ weighting, out to k = 12 Å-1 and analyzed

via shell-by-shell fitting using Artemis35, 36. Only the first shell was fitted, as other shells were not

visible enough for accurate analysis. The structure of trigonal Se0 as described by Keller and

coworkers37 was used to create FEFF6l input files from which to extract Se-Se paths out to 3.5 Å

and fit the experimental t-Se EXAFS spectrum at 95 K. Fit of the culture data were performed in

q space (2-12 Å) using a Kaiser-Bessel window (1.4-2.5 Å). Details on µXRD analyses can be

found in the Appendix II.

Correlative cryogenic SHXM and TEM. A custom X-ray microprobe cryo-stage (-190

to +150 °C, ±0.1°C precision) was designed and built in collaboration with Instec Inc, to fit the

geometry of the beamline and allow cryo-transfer of flash-frozen samples (Figures 1, S1 and

Appendix I). The apparatus allows cryo-XRF/XAS measurements to be performed in fluorescence

and/or transmission modes. XRD measurements in transmission, are used to check for ice

contamination and quality of the transfer. The apparatus consists of four parts: a stage (CLM77K),

a sample loading frame (SLF) that can accommodate two round sample grid boxes, a grid holder

tongue (GHT) and a temperature controller (mk1000). Prior to any cold experiment, the CLM77K

8

stage is heated to +110°C for 5 minutes then purged with dry N2 gas to remove any moisture

trapped. The CLM77K is cooled to -190 °C (83 K) using a pressurized LN2 tank. During cooling,

the sample chamber and all windows are purged with dry N2 gas to prevent water vapor

condensation and frost. The SLF, placed in a LN2 bath in a Styrofoam container, is used to support

the GHT during sample loading/unloading. Using the SLF, a single cryo-TEM grid (or Si3N4

window), mounted in a JEOL 3100 TEM cartridge, can be loaded onto the GHT for correlative

analyses. Alternatively, another GHT designed to accommodate up to three TEM grids (or

windows) can also be used. In either configuration, the spring loaded GHT cover snaps closed over

the cryo-TEM grids, keeping them thermally insulated. Once the CLM77K is cooled and stable at

-190 ± 0.1°C, the GHT containing the cryo-samples, is then quickly inserted and locked into the

CLM77K, where the sample temperature reaches 95 K. Thermocouples located in the GHT and

CLM77K, and connected to the mk1000 are used to continuously monitor (every second) the

temperature of the sample and the stage respectively. Once the cryo-microprobe measurements

(~12 h maximum duration) are complete, the GHT with samples is cryo-transferred back to the

SLF in an LN2 bath and samples are subsequently stored in LN2.

Results and discussion

Geochemistry

Biofilm samples (referred as ‘Biofilm CG02’) were collected 16 days after the start of

acetate amendment in the subsurface (Figure S2), during the Fe(III) reduction period38. Acetate

injection into the aquifer resulted in a rapid decrease of soluble Se; the minimum concentration

was reached after 7 days. The concentration of dissolved Se (Se oxyanions) just prior to acetate 9

injection on day 1 was 0.81 µM, versus 0.03 µM on day 15 (closest time point to sample

collection), suggesting that ~96% of dissolved Se was converted into a solid phase. Soluble Se

concentration remained well below the U.S. EPA limit for Se in drinking water (0.1µM) past

biostimulation experiment. Sulfate concentrations remained stable indicating minimal sulfate

reduction and sulfide production during the sampling period. These trends are consistent with prior

acetate biostimulation experiments16 on un-amended portions of the Rifle aquifer as was the case

here.

Microbial community composition

Phylogenetic analyses of the 16S rRNA gene sequences recovered from Biofilm CG02

show an abundance of organisms from the Betaproteobacteria class (72% of the community

sequences, Figures S3 and S4). These bacteria are often found dominant in freshwaters and

inferred to play an important role in the nitrogen cycle, including nitrate respiration.

Comamonadaceae is the most abundant family (Figure S3C) and uncultivated members of the

Simplicispira genus were the most abundant identified organisms (Figures S3D and S4B). It is

well known from prior studies at this site that acetate enriches for members of the

Comamonadaceae family, specifically the Simplicispira39 genus. In fact, this result was shown

previously in well CD01, the same groundwater reference well used in the current study40. The

facultative anaerobe Simplicispira strain BDI41 (motile, weakly-curved rod) a nitrate and vanadate

reducer, was isolated from this site. We tested this isolate for selenate or selenite reduction to Se0.

The organism was cultivated both in nitrate-free and nitrate-amended (2 mM) bicarbonate

freshwater medium inoculated with acetate (5, 10 mM) and selenate (5 mM) or selenite (1, 2 and

5 mM). No red (or grey/black/brown) Se0 precipitates were formed, indicating that strain BDI does

10

not carry out this transformation in solutions containing mM concentrations of these Se oxyanions.

A large number of uncultivated members of the Comamonadaceae family was detected, some of

which could contribute to the reduction of Se oxyanions. However, we could not test this

hypothesis because to date only the BDI strain has been isolated from this site. Some members of

the Comamonadaceae family such as Comamonas sp. are known selenite reducers42. Identified

genus from this family, Hydrogenophaga sp. (H2-oxidizer) are rod-shaped, autotrophic

denitrifiers43 that have been correlated with nitrate and selenate fluxes44. These organisms likely

contribute to selenate reduction because some strains have been reported to reduce selenate and

selenite to selenide45. Moreover, Hydrogenophaga sp. have been previously detected in Se-rich

biofilms from the Rifle site30.

The second most abundant family detected was Rhodocyclaceae (Figures S3C and S4C),

which contains many denitrifying bacteria. Among identified genera from this family, rod-shaped

Zoogloea sp. (9%), named the “living glue” are known for their characteristic gelatinous

extracellular polysaccharide matrix. This genus contains strains capable of denitrification46 and of

selenite reduction to elemental Se47. Zoogloea sp. have also been previously detected in Se-rich

biofilms from Rifle30. Ferribacterium sp.48, a Fe(III)-reducer, and Dechloromonas RCB form a

monophyletic group (Figure S4C), thus an accurate taxonomic affiliation cannot be obtained. Both

genera combined represent 5% of the bacterial community. Dechloromonas sp. are well-known

perchlorate and nitrate-reducing bacteria49, with some strains capable of oxidizing Fe(II)50. This

genus is commonly found or used to enhance selenate reduction in bioreactors51, 52 and has been

repeatedly detected at the Rifle site over the years30, 53-57.

11

Bacteria from the Chloroflexi phylum (5%) are members of the Anaerolineaceae family.

These organisms could contribute to selenate reduction because a variety of selenate reductases

have been identified in the genomes of Anaerolineaceae bacteria from Rifle native sediments

(http://ggkbase.berkeley.edu/). Some members of Bacteroidetes, the second most abundant

phylum, could also play a part in the reduction of Se oxyanions since selenate reductases have

been identified in the genomes of Bacteroidetes obtained from the Rifle site58. These organisms

are known to degrade complex organic compounds, notably polysaccharides and proteins59. Prior

studies at this site showed that acetate enriches for these organisms, and that they could possibly

couple acetate oxidation to nitrogen reduction58. Additionally, 16S rRNA analyses identified

members of bacterial candidate phyla (CP), WS6 and OD1 (6% total), previously reported at

Rifle56, 60, 61. Members of these CP, inferred to be obligate fermenters, have a small genome61 and

cell size60, and reportedly lack respiration capabilities57, 61 though some members of OD1 might

be capable of reducing sulfur56. Finally, acetate-oxidizing Fe(III)-reducing Rhodoferax/Albidiferax

sp.62 (Comamonadaceae) and Sulfuricurvum sp.63 (ε-proteobacteria), a nitrate-reducer sulfur-

oxidizer, were found in this study and have previously been detected at the Rifle site.

Organisms from unidentified genera and uncultivated Simplicispira members account for

nearly three-quarter of the bacterial community. Their potential role in Se oxyanions reduction is

unknown and warrants further investigation. These results suggest uncultivated members of the

Comamonadaceae family may play an important role in Se oxyanions reduction at this location.

Furthermore, most bacteria from biofilm CG02 exhibit pili (Figure S5), suggesting possible

interspecies12 electron transfer could possibly contribute or even drive the reduction of Se

oxyanions, as well as denitrification in the biofilm.

12

Selenium and carbon distribution and speciation in the field-derived biofilms

3D confocal microscopy revealed a 2 to 12 µm thick biofilm with several haystacks of self-

organized cells (Movie S1). On the surface of the haystack, the cells range in size from 0.75 to 2.7

µm (Movie S2). Within the core and bottom of the stack, cells are similar in length, but with an

oval shape. 2D cryo-TEM data (Figures 2A and S5) revealed weakly curved rods, rod-, oval- and

round-shaped cells. All cells exhibited a Gram-negative cell wall with an inner membrane, an outer

membrane and a peptidoglycan layer within a periplasmic space. Few cells exhibited an S-layer;

many had visible polar flagella and/or pili. Most cells contained cytoplasmic granules. STXM-

derived C K-edge XANES spectra of granules exhibits a major peak at 288.4 eV, attributed to

carboxyls or esters64, 65, likely associated with polyhydroxyalkanoates (polyesters), carbon,

electron and energy storage polymers commonly produced by bacteria under limited nutrients.

Carbon storage granules have been widely reported in organisms we detected by 16S r-RNA, such

as Simplicispira sp.39, Dechloromonas sp.66, Hydrogenophaga sp.67 and Zoogloea sp.46, 68.

Further, a prior study69 on members of the Comamonadaceae family showed that a N-poor medium

tends to stimulate the production of polyhydroxyalkanoates, whereas a C-poor environment leads

to the production of polyphosphates. A vast majority of colloidal nanoparticles (CNPs) were found

as extracellular aggregates, or as few electron dense particles on cell surfaces (Figure 2A inset).

SEM imaging and EDS spectra of these particles showed Se, with traces of Si and Ca from the

background (Figure S6). The CNPs ranged from 60 to 300 nm in size, with a distribution centered

at ~160 nm (Figure S7A). STXM data at C K- and Se L2,3 edges showed abundant Se-rich

aggregates associated with a thick protein-rich biofilm matrix (Figure 3A) containing EPS (peak

at 288.6 eV, π∗(C=O) transition of carboxyl group in acidic polysaccharides) and carbonates, with

13

a signature π∗(C=O) peak at ~ 290.3 eV (Figure S8). By contrast, only a few CNPs are associated

with each individual cell surface (Figure 3B), consistent with TEM data. Se L2,3 edges XANES

spectra of CNPs compared to model compounds are shown in Figure 3C. Spectra of Se standards

are consistent with prior reports70, 71. The spectra of CNPs closely match that of red amorphous

Se0, however the allotropic form of Se cannot be determined.

By contrast, most Se compounds can easily be classified by K-edge XANES, due to a clear

shift (2-3 eV) of the edge position (‘white line’) depending on Se oxidation state (Figure S9A).

This can be explained by dipole selection rules where 1s core electrons are excited into the

unoccupied 4p electronic states and the edge position shifts towards higher energy as the oxidation

state increases. Cryo-XRF mapping and XANES were performed on twenty nine Se-rich areas

from several cryo-TEM grids to get enough statistics. Se valence plots and least-square linear

combination fitting (LSQF) of the spectra both indicate that red amorphous Se0 is the major end

product (Figures 4 and S9B, Table S2).

Selenium and carbon distribution and speciation in enrichment cultures

Field results were replicated in the lab using a piece of Biofilm CG02 as inoculum.

Anaerobic selenate-respiring organisms were enriched using acetate (electron donor and carbon

source) and selenate as the sole electron acceptor. Red precipitates were observed after 5 days,

suggesting selenate reduction to Se0.

Correlated cryogenic SHXM/ TEM analyses of a 19-day-old culture in bicarbonate

medium (Figures 5 and S10) show a cluster of bacteria associated with ~100 ± 60 nm diameter

Se0 CNPs and traces of Se(IV). Organo-selenium and inorganic selenides were not detected in this

14

region, suggesting that reduction proceeds in a two-step reaction from Se(VI) to Se(IV) to Se(0).

Considering the low X-ray dose applied to the frozen samples (see Appendix II), photo-reduction

is an unlikely explanation for the formation of these compounds. The corresponding µ-XRD

spectrum (Figure S10B) does not show a match to crystalline Se0 (either hexagonal or

monoclinic). Cultures in bicarbonate medium (“Bicarb”) or in groundwater artificial medium

(“GWA”) were further analyzed 60 days after the addition of selenate and acetate to the medium

to determine the stable end products. Both Se valence plots and LSQF of the XANES spectra

indicate red amorphous Se0 is the major end product in all culture samples (Figure 4), regardless

of the medium used. Cryo-µEXAFS of Se0 CNP aggregates from a 19-day-old bicarbonate culture

(Figure 6, Table S3) shows that the first shell (Se-Se bond) lies at 2.339 ± 0.003 Å, consistent

with previous reports for amorphous Se72,73. Only the first shell could be fitted by a shell-by-shell

method, the second shell was not sufficiently visible to fit due to lack of structural order, thus the

Se-Se-Se bond angle could not be determined. There is a transient accumulation of selenite during

selenate reduction in the culture samples that is not observed in the field data, as evidenced both

in the valence plot and fitting results (Figure 4). 2D cryo-TEM data reveal electron-dense particles

nucleated on cell surfaces (Figure 2) and extensive extracellular aggregates ~130 ± 70 nm formed

(Figure S7B). STXM data shows that Se0 CNPs located near cell surfaces are coated with a

protein-rich organic layer (Figure S11). 2D Cryo-TEM and STXM reveal relatively large granules

on most bacteria in cultures (Figure 2).

Although a vast majority of the data indicated red amorphous Se0, a few regions analyzed

outside thick biofilm regions in the 19-day-old culture show crystallization, as evidenced by

diffraction contrast and the presence of planar defects, likely twin planes (Figure S12). It is

15

interesting that crystallization occurred over just 19 days at 25 °C, given that transformation of

bulk red amorphous Se to grey trigonal Se should occur at appreciable rates only above ~50 °C74.

Nonetheless, in both cultures and field samples red amorphous Se was stable for months.

Evidence for dissimilatory selenate reduction (DSeR)

Cryo-spectromicroscopy and 16S rRNA gene sequence analysis of field and culture

samples strongly suggest the presence of DSeR organisms. In that process, selenate, used as the

sole terminal electron acceptor, is sequentially reduced to selenite and Se0. The presence of organic

Se is minor, and can be attributed to either cell lysis, selenate-tolerant bacteria or potentially

selenite to selenide reducing bacteria. When using sulfate-containing media (GWA), few sulfate-

reducing anaerobes could also be capable of reducing µM amounts of selenate, although they

generally do not couple this reduction to growth27, 28, 75. Moreover, the ability of sulfate respirers

to reduce selenate (or selenite) is greatly constrained by the availability of sulfate. Se oxyanions

are thermodynamically predicted to be reduced prior to sulfate22 according to their respective redox

potentials (+0.44 V for SeO42-/SeO3

2-, +0.21 V for SeO32-/Se0, -0.22 V for SO4

2-/H2S and -0.52 V

for SO42-/SO3

2-). At the shallow depth (4 meters) where biofilms were collected, the sulfate

concentration (8 mM, Figure S2) is at a level that usually precludes selenate reduction by sulfate

reducers21, 76. By contrast, DSeR microorganisms can reduce mM amounts of selenate to Se0,

consistent with our observations. In part due to similar potentials for the SeO42-/ SeO3

2- (+0.44 V)

and NO3-/NO2

- (+0.42 V) redox couples, microbial reduction of nitrate and selenate often occur

close together77, 78, consistent with the presence of denitrifying bacteria we find by 16S rRNA gene

sequencing analyses. 16

Selenate reduction is mediated by either a soluble periplasmic selenate reductase (SerABC)

or a nitrate reductase or via a dissimilatory sulfate-reducing pathway. A variety of selenate

reductases have been identified in the genomes of bacteria from the Rifle site

(http://ggkbase.berkeley.edu/). Selenite is usually imported into the cytoplasm where it is reduced

to Se0 via a membrane-associated reductase, followed by rapid expulsion of Se particles via a

membrane efflux pump75. Selenite is reduced by reacting with proteins in a “Painter-type”

reaction, suggested as a general microbial detoxification reaction to Se oxyanions79.

Transient accumulation of selenite in cultures

Selenite can be dissimilatory reduced to Se0 with massive uptake of selenite or assimilatory

reduced to organoselenium (e.g., SeMet) and selenides. Dissimilatory reduction of selenite by

anaerobic bacteria generally produces abundant extracellular Se0 particles29, consistent with our

observations. Selenite is a transitory intermediate in reduction of selenate, being produced and

reduced concomitantly. The accumulation of selenite observed during growth on selenate,

regardless of the medium used, could occur because the microbial community reduces selenate

faster than selenite. Alternatively, Se oxidizers may re-oxidize Se0 to selenite. The first hypothesis

is the most likely, as reduction of selenite to red Se0 by the cultures was minimal, and occurred

very slowly (over weeks). Further, oxidation of Se0 by Se-oxidizers is unlikely as we did not detect

organisms known to oxidize Se. Even if they were present, which we cannot rule out considering

the large pool of unidentified organisms, the process is generally very slow 80. More importantly

in our cultures, cells are under different geochemical conditions than in the field and are subjected

to concentrations of Se oxyanions orders of magnitude higher (5 mM versus ~1 µM), selectively

enriching for few members of the community.

17

Association of red amorphous Se CNPs with proteins

Our results show that Se0 CNPs are strongly associated with proteins in the biofilm81, 82.

Particles may be surface stabilized from dissolution or phase transformation when embedded in

the protein-rich biofilm matrix, as suggested by prior research on biogenic and synthetic Se0

nanoparticles in the presence of proteins83, 84. Previous studies have suggested the presence of

surface-associated proteins on Se0 produced by selenite reducers85. More generally, prior studies

have shown that biogenic (and synthetic) selenium nanoparticles can be associated with a plenitude

of high-affinity proteins83, 86. Proteins, peptides, and amino acids could be released after cell

death87 and scavenged by hydrophobic elemental selenium surfaces. Alternatively, bacteria may

also excrete Se-binding proteins24. Finally, Se0 particles found outside cells could have been

released through cell lysis28, as we have observed in old culture samples. Any of these processes

would lead to extracellular aggregation of Se0 nanoparticles, preventing entombment of cells. The

aggregation of Se particles likely affects selenium mobility and transport88, as evidenced by prior

work showing that aggregation induced by extracellular metal-binding polypeptides and proteins

plays an important role in constraining the dispersion of nanoparticles in the environment89.

Conclusions

Many anthropogenic activities (e.g., agriculture, petroleum refining, mining, glass and

pigment manufacturing) generate Se contaminated wastewaters. Existing treatment technologies,

based on chemical co-precipitation or adsorption, are rather inefficient, especially for selenate, and

too expensive for practical industrial use. Bioremediation represents an attractive alternative

approach, but strongly relies on determining accurately the chemical speciation and distribution of

18

the bio-reduction products in order to be directly applicable to a diversity of anaerobic soil and

groundwater environments contaminated with selenium.

The community of Se oxyanions reducers detected here includes several genera previously

detected in Se-rich biofilms collected under similar conditions at the Rifle site30. However, the

distribution patterns and allotropic form of selenium found here are clearly distinct. The prior study

suggested the presence of cells encrusted with red monoclinic Se0, but using our novel correlative

cryogenic spectro-microscopy apparatus, we demonstrate that the main product is extracellular red

amorphous Se0 captured in a protein-rich biofilm matrix and that only few particles are associated

with each cell surface. Both the newly identified protein coating and extensive particle aggregation

are expected to reduce re-oxidation rates, thereby minimizing the rapid re-release of aqueous Se

to the environment.

Acknowledgments

This material is partially based upon work supported through the Lawrence Berkeley

National Laboratory’s Sustainable Systems Scientific Focus Area. The U.S. D.O.E. Office of

Science, Office of Biological and Environmental Research funded the work under contracts DE-

SC0004733 and DE-AC02-05CH11231. Part of the equipment was funded by the LBL EFRC

Center for Nanoscale Control of Geologic CO2. S. F. thanks Paul Baker at Instec Inc. for his help

with the microprobe cryo-stage, Sue Spaulding for lab support, Tolek Tylizscack for support at

ALS beamline 11.0.2, Mary Gilles and Steve Kelly for sharing their ESEM and Ken H. Downing

(LBL) for providing cryo-TEM infrastructure. S.F. is grateful to D. Strawn, A. Ryser, E.A.H.

19

Pilon-Smits and J.L. Freeman for sharing their selenium standard spectra. The Advanced Light

Source is supported by the Office of Basic Energy Sciences, Office of Science, U.S. D.O.E.

Contract No. DE-AC02-05CH11231.

Additional experimental details, figures, and tables as referenced in the text can be found

in Appendix II.

References

1. Dohnalkova, A. C.; Marshall, M. J.; Arey, B. W.; Williams, K. H.; Buck, E. C.; Fredrickson, J. K., Imaging Hydrated Microbial Extracellular Polymers: Comparative Analysis by Electron Microscopy. Applied and Environmental Microbiology 2011, 77, (4), 1254-1262. 2. Manceau A., M. A. M., N. Tamura, Quantitative speciation of metals in soils and sediments. Applications of synchrotron radiation 2002; Vol. 49. 3. Howells, M. R.; Hitchcock, A. P.; Jacobsen, C. J., Introduction: Special issue on radiation damage. Journal of Electron Spectroscopy and Related Phenomena 2009, 170, (1–3), 1-3. 4. Fayard, B.; Salomé, M.; Takemoto, K.; Kihara, H.; Susini, J., Some practical considerations about the effects of radiation damage on hydrated cells imaged by X-ray fluorescence microscopy. Journal of Electron Spectroscopy and Related Phenomena 2009, 170, (1–3), 19-24. 5. Cody, G. D.; Brandes, J.; Jacobsen, C.; Wirick, S., Soft X-ray induced chemical modification of polysaccharides in vascular plant cell walls. Journal of Electron Spectroscopy and Related Phenomena 2009, 170, (1–3), 57-64. 6. Holton, J., XANES measurements of the rate of radiation damage to selenomethionine side chains. Journal of Synchrotron Radiation 2007, 14, (1), 51-72. 7. George, G. N.; Pickering, I. J.; Pushie, M. J.; Nienaber, K.; Hackett, M. J.; Ascone, I.; Hedman, B.; Hodgson, K. O.; Aitken, J. B.; Levina, A.; Glover, C.; Lay, P. A., X-ray-induced photo-chemistry and X-ray absorption spectroscopy of biological samples. Journal of Synchrotron Radiation 2012, 19, (6), 875-886. 8. Huggins, F. E.; Sanei, H., Synchrotron-radiation-induced oxidation of selenite to selenate in coal-derived fly ash. Journal of Synchrotron Radiation 2011, 18, (3), 530-533. 9. Lenz, M.; van Hullebusch, E. D.; Farges, F.; Nikitenko, S.; Corvini, P. F. X.; Lens, P. N. L., Combined Speciation Analysis by X-ray Absorption Near-Edge Structure Spectroscopy, Ion Chromatography, and Solid-Phase Microextraction Gas Chromatography−Mass Spectrometry To Evaluate Biotreatment of Concentrated Selenium Wastewaters. Environmental Science & Technology 2010, 45, (3), 1067-1073. 10. Kolobov, A. V.; Oyanagi, H.; Tanaka, K.; Tanaka, K., Structural study of amorphous selenium by in situ EXAFS: Observation of photoinduced bond alternation. Physical Review B 1997, 55, (2), 726-734. 11. Milne, J. L. S.; Subramaniam, S., Cryo-electron tomography of bacteria: progress, challenges and future prospects. Nat Rev Micro 2009, 7, (9), 666-675.

20

12. Comolli, L. R.; Banfield, J. F., Inter-species interconnections in acid mine drainage microbial communities. Frontiers in Microbiology 2014, 5, 367. 13. Comolli, L. R.; Downing, K. H., Dose tolerance at helium and nitrogen temperatures for whole cell electron tomography. Journal of Structural Biology 2005, 152, (3), 149-156. 14. Anderson, R. T.; Vrionis, H. A.; Ortiz-Bernad, I.; Resch, C. T.; Long, P. E.; Dayvault, R.; Karp, K.; Marutzky, S.; Metzler, D. R.; Peacock, A.; White, D. C.; Lowe, M.; Lovley, D. R., Stimulating the In Situ Activity of Geobacter Species To Remove Uranium from the Groundwater of a Uranium-Contaminated Aquifer. Applied and Environmental Microbiology 2003, 69, (10), 5884-5891. 15. Campbell, K. M.; Kukkadapu, R. K.; Qafoku, N. P.; Peacock, A. D.; Lesher, E.; Williams, K. H.; Bargar, J. R.; Wilkins, M. J.; Figueroa, L.; Ranville, J.; Davis, J. A.; Long, P. E., Geochemical, mineralogical and microbiological characteristics of sediment from a naturally reduced zone in a uranium-contaminated aquifer. Applied Geochemistry 2012, 27, (8), 1499-1511. 16. Williams, K. H.; Long, P. E.; Davis, J. A.; Wilkins, M. J.; N'Guessan, A. L.; Steefel, C. I.; Yang, L.; Newcomer, D.; Spane, F. A.; Kerkhof, L. J.; McGuinness, L.; Dayvault, R.; Lovley, D. R., Acetate Availability and its Influence on Sustainable Bioremediation of Uranium-Contaminated Groundwater. Geomicrobiology Journal 2011, 28, (5-6), 519-539. 17. Bao, C.; Wu, H.; Li, L.; Newcomer, D.; Long, P. E.; Williams, K. H., Uranium Bioreduction Rates across Scales: Biogeochemical Hot Moments and Hot Spots during a Biostimulation Experiment at Rifle, Colorado. Environmental Science & Technology 2014, 48, (17), 10116-10127. 18. Santolo, O. H. M. a. G. M., Kesterson Reservoir-Past, Present, and Future: An Ecological Risk Assessment. In Selenium in the Environment, W.T. Frankenberger, J., and S.M. Benson, Ed. Marcel Dekker: 1994; pp 69-117. 19. Shrift, A., A Selenium Cycle in Nature? Nature 1964, 201, (4926), 1304-1305. 20. W.T. Frankenberger, J., and S.M. Benson, Selenium in the environment. Marcel Dekker, Inc. ed.; Taylor & Francis: N.Y.C., N.Y., 1994. 21. Lovley, D. R., Dissimilatory Metal Reduction. Annual Review of Microbiology 1993, 47, (1), 263-290. 22. Oremland, R. S.; Hollibaugh, J. T.; Maest, A. S.; Presser, T. S.; Miller, L. G.; Culbertson, C. W., Selenate Reduction to Elemental Selenium by Anaerobic Bacteria in Sediments and Culture: Biogeochemical Significance of a Novel, Sulfate-Independent Respiration. Applied and Environmental Microbiology 1989, 55, (9), 2333-2343. 23. Steinberg, N. A.; Oremland, R. S., Dissimilatory Selenate Reduction Potentials in a Diversity of Sediment Types. Applied and Environmental Microbiology 1990, 56, (11), 3550-3557. 24. Debieux, C. M.; Dridge, E. J.; Mueller, C. M.; Splatt, P.; Paszkiewicz, K.; Knight, I.; Florance, H.; Love, J.; Titball, R. W.; Lewis, R. J.; Richardson, D. J.; Butler, C. S., A bacterial process for selenium nanosphere assembly. Proceedings of the National Academy of Sciences 2011, 108, (33), 13480-13485. 25. Hunter, W.; Manter, D., Reduction of Selenite to Elemental Red Selenium by Pseudomonas sp. Strain CA5. Curr Microbiol 2009, 58, (5), 493-498. 26. Kessi, J.; Ramuz, M.; Wehrli, E.; Spycher, M.; Bachofen, R., Reduction of Selenite and Detoxification of Elemental Selenium by the Phototrophic BacteriumRhodospirillum rubrum. Applied and Environmental Microbiology 1999, 65, (11), 4734-4740. 27. Tomei, F. A.; Barton, L. L.; Lemanski, C. L.; Zocco, T. G., Reduction of selenate and selenite to elemental selenium by Wolinella succinogenes. Canadian Journal of Microbiology 1992, 38, (12), 1328-1333. 28. Tomei, F.; Barton, L.; Lemanski, C.; Zocco, T.; Fink, N.; Sillerud, L., Transformation of selenate and selenite to elemental selenium byDesulfovibrio desulfuricans. Journal of Industrial Microbiology 1995, 14, (3-4), 329-336.

21

29. Oremland, R. S.; Herbel, M. J.; Blum, J. S.; Langley, S.; Beveridge, T. J.; Ajayan, P. M.; Sutto, T.; Ellis, A. V.; Curran, S., Structural and Spectral Features of Selenium Nanospheres Produced by Se-Respiring Bacteria. Applied and Environmental Microbiology 2004, 70, (1), 52-60. 30. Williams, K. H.; Wilkins, M. J.; N'Guessan, A. L.; Arey, B.; Dodova, E.; Dohnalkova, A.; Holmes, D.; Lovley, D. R.; Long, P. E., Field evidence of selenium bioreduction in a uranium-contaminated aquifer. Environmental Microbiology Reports 2013, 5, (3), 444-452. 31. Comolli, L. R.; Duarte, R.; Baum, D.; Luef, B.; Downing, K. H.; Larson, D. M.; Csencsits, R.; Banfield, J. F., A portable cryo-plunger for on-site intact cryogenic microscopy sample preparation in natural environments. Microscopy Research and Technique 2012, 75, (6), 829-836. 32. Marcus, M. A.; MacDowell, A. A.; Celestre, R.; Manceau, A.; Miller, T.; Padmore, H. A.; Sublett, R. E., Beamline 10.3.2 at ALS: a hard X-ray microprobe for environmental and materials sciences. Journal of Synchrotron Radiation 2004, 11, (3), 239-247. 33. Kelly, S. D., Hesterberg, D., & Ravel, B, Analysis of Soils and Minerals Using X-ray Absorption Spectroscopy. In Mineralogical Methods, Drees, A. L. U. L. R., Ed. Soil Science Society of America: Madison, WI, 2008; Vol. Part 5, p 367. 34. Bañuelos, G. S.; Fakra, S. C.; Walse, S. S.; Marcus, M. A.; Yang, S. I.; Pickering, I. J.; Pilon-Smits, E. A. H.; Freeman, J. L., Selenium Accumulation, Distribution, and Speciation in Spineless Prickly Pear Cactus: A Drought- and Salt-Tolerant, Selenium-Enriched Nutraceutical Fruit Crop for Biofortified Foods. Plant Physiology 2011, 155, (1), 315-327. 35. Newville, M., IFEFFIT: interactive XAFS analysis and FEFF fitting. J. Synchrotron Radiation 2001, 8, (2), 322-324. 36. Ravel, B.; Newville, M., ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT. J. Synchrotron Radiation 2005, 12, (4), 537-541. 37. Keller, R.; Holzapfel, W. B.; Schulz, H., Effect of pressure on the atom positions in Se and Te. Physical Review B 1977, 16, (10), 4404-4412. 38. Luef, B.; Fakra, S. C.; Csencsits, R.; Wrighton, K. C.; Williams, K. H.; Wilkins, M. J.; Downing, K. H.; Long, P. E.; Comolli, L. R.; Banfield, J. F., Iron-reducing bacteria accumulate ferric oxyhydroxide nanoparticle aggregates that may support planktonic growth. ISME J 2013, 7, (2), 338-350. 39. Grabovich, M.; Gavrish, E.; Kuever, J.; Lysenko, A. M.; Podkopaeva, D.; Dubinina, G., Proposal of Giesbergeria voronezhensis gen. nov., sp. nov. and G. kuznetsovii sp. nov. and reclassification of [Aquaspirillum] anulus, [A.] sinuosum and [A.] giesbergeri as Giesbergeria anulus comb. nov., G. sinuosa comb. nov. and G. giesbergeri comb. nov., and [Aquaspirillum] metamorphum and [A.] psychrophilum as Simplicispira metamorpha gen. nov., comb. nov. and S. psychrophila comb. nov. International Journal of Systematic and Evolutionary Microbiology 2006, 56, (3), 569-576. 40. Yelton, A. P.; Williams, K. H.; Fournelle, J.; Wrighton, K. C.; Handley, K. M.; Banfield, J. F., Vanadate and Acetate Biostimulation of Contaminated Sediments Decreases Diversity, Selects for Specific Taxa, and Decreases Aqueous V5+ Concentration. Environmental Science & Technology 2013, 47, (12), 6500-6509. 41. Yelton, A. P. Metagenomic and Cultivation-Based Analysis of Novel Microorganisms and Functions in Metal-Contaminated Environments. UC Berkeley, Berkeley, 2012. 42. Zheng, S.; Su, J.; Wang, L.; Yao, R.; Wang, D.; Deng, Y.; Wang, R.; Wang, G.; Rensing, C., Selenite reduction by the obligate aerobic bacterium Comamonas testosteroni S44 isolated from a metal-contaminated soil. BMC microbiology 2014, 14, 204. 43. Willems, A.; Busse, J.; Goor, M.; Pot, B.; Falsen, E.; Jantzen, E.; Hoste, B.; Gillis, M.; Kersters, K.; Auling, G.; De Ley, J., Hydrogenophaga, a New Genus of Hydrogen-Oxidizing Bacteria That Includes Hydrogenophaga flava comb. nov. (Formerly Pseudomonas flava), Hydrogenophaga palleronii (Formerly Pseudomonas palleronii), Hydrogenophaga pseudoflava (Formerly Pseudomonas pseudoflava and “Pseudomonas carboxydoflava”), and Hydrogenophaga taeniospiralis (Formerly Pseudomonas taeniospiralis). International Journal of Systematic Bacteriology 1989, 39, (3), 319-333.

22

44. Lai, C.-Y.; Yang, X.; Tang, Y.; Rittmann, B. E.; Zhao, H.-P., Nitrate Shaped the Selenate-Reducing Microbial Community in a Hydrogen-Based Biofilm Reactor. Environmental Science & Technology 2014, 48, (6), 3395-3402. 45. Ranjard, L.; Prigent-Combaret, C.; Favre-Bonté, S.; Monnez, C.; Nazaret, S.; Cournoyer, B., Characterization of a novel selenium methyltransferase from freshwater bacteria showing strong similarities with the calicheamicin methyltransferase. Biochimica et Biophysica Acta (BBA) - Gene Structure and Expression 2004, 1679, (1), 80-85. 46. Dugan, P.; Stoner, D.; Pickrum, H., The Genus Zoogloea. In The Prokaryotes, Dworkin, M.; Falkow, S.; Rosenberg, E.; Schleifer, K.-H.; Stackebrandt, E., Eds. Springer New York: 2006; pp 960-970. 47. Srivastava, N.; Mukhopadhyay, M., Biosynthesis and structural characterization of selenium nanoparticles mediated by Zooglea ramigera. Powder Technology 2013, 244, (0), 26-29. 48. Cummings, D. E.; Caccavo Jr, F.; Spring, S.; Rosenzweig, R. F., Ferribacterium limneticum, gen. nov., sp. nov., an Fe(III)-reducing microorganism isolated from mining-impacted freshwater lake sediments. Arch Microbiol 1999, 171, (3), 183-188. 49. Coates, J. D.; Chakraborty, R.; Lack, J. G.; O'Connor, S. M.; Cole, K. A.; Bender, K. S.; Achenbach, L. A., Anaerobic benzene oxidation coupled to nitrate reduction in pure culture by two strains of Dechloromonas. Nature 2001, 411, (6841), 1039-1043. 50. Chakraborty, A.; Picardal, F., Neutrophilic, nitrate-dependent, Fe(II) oxidation by a Dechloromonas species. World J Microbiol Biotechnol 2013, 29, (4), 617-623. 51. Zhang, Y.; Frankenberger Jr, W. T., Supplementing Bacillus sp. RS1 with Dechloromonas sp. HZ for enhancing selenate reduction in agricultural drainage water. Science of The Total Environment 2007, 372, (2–3), 397-405. 52. Chung, J.; Ryu, H.; Abbaszadegan, M.; Rittmann, B. E., Community structure and function in a H(2)-based membrane biofilm reactor capable of bioreduction of selenate and chromate. Appl Microbiol Biotechnol 2006, 72, (6), 1330-1339. 53. Mouser, P. J.; N’Guessan, A. L.; Elifantz, H.; Holmes, D. E.; Williams, K. H.; Wilkins, M. J.; Long, P. E.; Lovley, D. R., Influence of Heterogeneous Ammonium Availability on Bacterial Community Structure and the Expression of Nitrogen Fixation and Ammonium Transporter Genes during in Situ Bioremediation of Uranium-Contaminated Groundwater. Environmental Science & Technology 2009, 43, (12), 4386-4392. 54. N'Guessan, A. L.; Elifantz, H.; Nevin, K. P.; Mouser, P. J.; Methe, B.; Woodard, T. L.; Manley, K.; Williams, K. H.; Wilkins, M. J.; Larsen, J. T.; Long, P. E.; Lovley, D. R., Molecular analysis of phosphate limitation in Geobacteraceae during the bioremediation of a uranium-contaminated aquifer. ISME J 2009, 4, (2), 253-266. 55. Handley, K. M.; Wrighton, K. C.; Piceno, Y. M.; Andersen, G. L.; DeSantis, T. Z.; Williams, K. H.; Wilkins, M. J.; N'Guessan, A. L.; Peacock, A.; Bargar, J.; Long, P. E.; Banfield, J. F., High-density PhyloChip profiling of stimulated aquifer microbial communities reveals a complex response to acetate amendment. FEMS Microbiology Ecology 2012, 81, (1), 188-204. 56. Wrighton, K. C.; Thomas, B. C.; Sharon, I.; Miller, C. S.; Castelle, C. J.; VerBerkmoes, N. C.; Wilkins, M. J.; Hettich, R. L.; Lipton, M. S.; Williams, K. H.; Long, P. E.; Banfield, J. F., Fermentation, Hydrogen, and Sulfur Metabolism in Multiple Uncultivated Bacterial Phyla. Science 2012, 337, (6102), 1661-1665. 57. Wrighton, K. C.; Castelle, C. J.; Wilkins, M. J.; Hug, L. A.; Sharon, I.; Thomas, B. C.; Handley, K. M.; Mullin, S. W.; Nicora, C. D.; Singh, A.; Lipton, M. S.; Long, P. E.; Williams, K. H.; Banfield, J. F., Metabolic interdependencies between phylogenetically novel fermenters and respiratory organisms in an unconfined aquifer. ISME J 2014, 8, (7), 1452-1463. 58. Handley, K. M.; VerBerkmoes, N. C.; Steefel, C. I.; Williams, K. H.; Sharon, I.; Miller, C. S.; Frischkorn, K. R.; Chourey, K.; Thomas, B. C.; Shah, M. B.; Long, P. E.; Hettich, R. L.; Banfield, J. F.,

23

Biostimulation induces syntrophic interactions that impact C, S and N cycling in a sediment microbial community. ISME J 2013, 7, (4), 800-816. 59. Thomas, F.; Hehemann, J.-H.; Rebuffet, E.; Czjzek, M.; Michel, G., Environmental and gut Bacteroidetes: the food connection. Frontiers in Microbiology 2011, 2. 60. Luef, B.; Frischkorn, K. R.; Wrighton, K. C.; Holman, H.-Y. N.; Birarda, G.; Thomas, B. C.; Singh, A.; Williams, K. H.; Siegerist, C. E.; Tringe, S. G.; Downing, K. H.; Comolli, L. R.; Banfield, J. F., Diverse uncultivated ultra-small bacterial cells in groundwater. Nat Commun 2015, 6. 61. Kantor, R. S.; Wrighton, K. C.; Handley, K. M.; Sharon, I.; Hug, L. A.; Castelle, C. J.; Thomas, B. C.; Banfield, J. F., Small Genomes and Sparse Metabolisms of Sediment-Associated Bacteria from Four Candidate Phyla. mBio 2013, 4, (5). 62. Finneran, K. T.; Johnsen, C. V.; Lovley, D. R., Rhodoferax ferrireducens sp. nov., a psychrotolerant, facultatively anaerobic bacterium that oxidizes acetate with the reduction of Fe(III). International Journal of Systematic and Evolutionary Microbiology 2003, 53, (3), 669-673. 63. Handley, K. M.; Bartels, D.; O'Loughlin, E. J.; Williams, K. H.; Trimble, W. L.; Skinner, K.; Gilbert, J. A.; Desai, N.; Glass, E. M.; Paczian, T.; Wilke, A.; Antonopoulos, D.; Kemner, K. M.; Meyer, F., The complete genome sequence for putative H-2- and S-oxidizer Candidatus Sulfuricurvum sp., assembled de novo from an aquifer-derived metagenome. Environmental Microbiology 2014, 16, (11), 3443-3462. 64. Boyce, C. K.; Cody, G. D.; Feser, M.; Jacobsen, C.; Knoll, A. H.; Wirick, S., Organic chemical differentiation within fossil plant cell walls detected with X-ray spectromicroscopy. Geology 2002, 30, (11), 1039-1042. 65. Cody, G. D.; Ade, H.; Wirick, S.; Mitchell, G. D.; Davis, A., Determination of chemical-structural changes in vitrinite accompanying luminescence alteration using C-NEXAFS analysis. Organic Geochemistry 1998, 28, (7–8), 441-455. 66. Salinero, K.; Keller, K.; Feil, W.; Feil, H.; Trong, S.; Di Bartolo, G.; Lapidus, A., Metabolic analysis of the soil microbe Dechloromonas aromatica str. RCB: indications of a surprisingly complex life-style and cryptic anaerobic pathways for aromatic degradation. BMC Genomics 2009, 10, (1), 351. 67. Yoon, S. C.; Choi, M. H., Local Sequence Dependence of Polyhydroxyalkanoic Acid Degradation in Hydrogenophaga pseudoflava. Journal of Biological Chemistry 1999, 274, (53), 37800-37808. 68. Saito, T.; Saegusa, H.; Miyata, Y.; Fukui, T., Intracellular degradation of poly(3-hydroxybutyrate) granules of Zoogloea ramigera I-16-M. FEMS Microbiology Letters 1992, 103, (2-4), 333-338. 69. Spring, S.; Wagner, M.; Schumann, P.; Kämpfer, P., Malikia granosa gen. nov., sp. nov., a novel polyhydroxyalkanoate- and polyphosphate-accumulating bacterium isolated from activated sludge, and reclassification of Pseudomonas spinosa as Malikia spinosa comb. nov. International Journal of Systematic and Evolutionary Microbiology 2005, 55, (2), 621-629. 70. J. Madwid, R. A., R. Blyth, I. Coulthard, C.J. Doonan, D. Liu, R. Hoffmeyer, M.J. Pushie, T. Regier, J. Ruszkowski, S.P. Singh, D. Thavarajah, C.I.E. Wiramanaden, S.I. Yang, L. Zhang, G.N. George, I.J. Pickering Selenium L-edge spectroscopy at the SGM Beamline as a tool for environmental selenium speciation; Canadian Light Source: 2008. 71. Bugaris, D. E.; Copping, R.; Tyliszczak, T.; Shuh, D. K.; Ibers, J. A., La2U2Se9: An Ordered Lanthanide/Actinide Chalcogenide with a Novel Structure Type. Inorganic Chemistry 2010, 49, (5), 2568-2575. 72. Majid, M.; Bénazeth, S.; Souleau, C.; Purans, J., XAFS study of interchain and intrachain order in Se1-xTex glasses: Nearest neighbors. Physical Review B 1998, 58, (10), 6104-6114. 73. Yang, S. I.; Lawrence, J. R.; Swerhone, G. D. W.; Pickering, I. J., Biotransformation of selenium and arsenic in multi-species biofilm. Environmental Chemistry 2011, 8, (6), 543-551. 74. Minaev, V. S.; Timoshenkov, S. P.; Kalugin, V. V., Structural and phase transformations in condensed selenium. Journal of Optoelectronics and Advanced Materials 2005, 7, (4), 1717-1741.

24

75. Losi, M. E.; Frankenberger, W. T., Reduction of Selenium Oxyanions by Enterobacter cloacae SLD1a-1: Isolation and Growth of the Bacterium and Its Expulsion of Selenium Particles. Applied and Environmental Microbiology 1997, 63, (8), 3079-84. 76. Zehr, J. P.; Oremland, R. S., Reduction of Selenate to Selenide by Sulfate-Respiring Bacteria: Experiments with Cell Suspensions and Estuarine Sediments. Applied and Environmental Microbiology 1987, 53, (6), 1365-1369. 77. Oremland, R. S.; Blum, J. S.; Bindi, A. B.; Dowdle, P. R.; Herbel, M.; Stolz, J. F., Simultaneous Reduction of Nitrate and Selenate by Cell Suspensions of Selenium-Respiring Bacteria. Applied and Environmental Microbiology 1999, 65, (10), 4385-4392. 78. Steinberg, N. A.; Blum, J. S.; Hochstein, L.; Oremland, R. S., Nitrate Is a Preferred Electron Acceptor for Growth of Freshwater Selenate-Respiring Bacteria. Applied and Environmental Microbiology 1992, 58, (1), 426-428. 79. Harrison, J. J.; Ceri, H.; Turner, R. J., Multimetal resistance and tolerance in microbial biofilms. Nat Rev Micro 2007, 5, (12), 928-938. 80. Dowdle, P. R.; Oremland, R. S., Microbial Oxidation of Elemental Selenium in Soil Slurries and Bacterial Cultures. Environmental Science & Technology 1998, 32, (23), 3749-3755. 81. Flemming, H. C.; Wingender, J., The biofilm matrix. Nature reviews. Microbiology 2010, 8, (9), 623-33. 82. Sutherland, I. W., The biofilm matrix – an immobilized but dynamic microbial environment. Trends in Microbiology 2001, 9, (5), 222-227. 83. Dobias, J.; Suvorova, E. I.; Bernier-Latmani, R., Role of proteins in controlling selenium nanoparticle size. Nanotechnology 2011, 22, (19), 195605. 84. Kaur, G.; Iqbal, M.; Bakshi, M. S., Biomineralization of Fine Selenium Crystalline Rods and Amorphous Spheres. The Journal of Physical Chemistry C 2009, 113, (31), 13670-13676. 85. Pearce, C. I.; Pattrick, R. A. D.; Law, N.; Charnock, J. M.; Coker, V. S.; Fellowes, J. W.; Oremland, R. S.; Lloyd, J. R., Investigating different mechanisms for biogenic selenite transformations: Geobacter sulfurreducens, Shewanella oneidensis and Veillonella atypica. Environmental Technology 2009, 30, (12), 1313-1326. 86. Lenz, M.; Kolvenbach, B.; Gygax, B.; Moes, S.; Corvini, P. F. X., Shedding light on selenium biomineralization: proteins associated with bionanominerals. Applied and Environmental Microbiology 2011. 87. Bayles, K. W., The biological role of death and lysis in biofilm development. Nat Rev Micro 2007, 5, (9), 721-726. 88. Buchs, B.; Evangelou, M. W. H.; Winkel, L. H. E.; Lenz, M., Colloidal Properties of Nanoparticular Biogenic Selenium Govern Environmental Fate and Bioremediation Effectiveness. Environmental Science & Technology 2013, 47, (5), 2401-2407. 89. Moreau, J. W.; Weber, P. K.; Martin, M. C.; Gilbert, B.; Hutcheon, I. D.; Banfield, J. F., Extracellular Proteins Limit the Dispersal of Biogenic Nanoparticles. Science 2007, 316, (5831), 1600-1603.

25

Figure 1. SHXM cryo-stage allowing correlative cryogenic TEM/microprobe measurements. A)

CLM77K stage with the JEOL 3100 TEM cartridge sample transfer tongue inserted. B) View of

the cartridge grid holder tongue installed on the mounting frame. C) View of the stage with the

sample transfer tongue inserted. The sample is oriented at 45 degrees to the incident beam, micro-

XRD is performed in transmission mode. More details can be found in Appendices I and II.

26

Figure 2. Cryo-TEM images of anaerobes, colloidal nanoparticles and aggregates from (A)

Biofilm CG02 and (B) Selenate-reducing enrichment culture (19-day-old) grown in bicarbonate

medium. IM = inner membrane, OM= outer membrane, F= flagellum, PH=

polyhydroxyalkanoates, CNP= colloidal nanoparticles, Au= gold fiducial particles on a carbon

coated lacey Formvar film. Scale bars are 200 nm.

27

Figure 3. STXM measurements on biofilm CG02: A) Carbon (in red) and selenium (in green)

bicolor-coded distribution map. B) Se map (in green) over-layed with an absorption contrast image

at 280 eV. C) Se L2,3 XANES spectra of Se colloidal particles compared to standards Se0, organic

Se, Se(IV) and Se(VI). Abundant extracellular Se0 CNPs are associated with the thick biofilm, but

only few Se colloidal nanoparticles (CNPs) associated with each cell.

28

Figure 4. Se valence state scatter plot. Standard compounds are shown in open black squares,

biofilm CG02 in red (29 spots), selenate-reducing enrichment cultures grown in bicarbonate

medium (dark blue, 18 spots) or in groundwater artificial medium (light blue, 23 spots). The Se(-

II,-I) group includes organic Se standards. See Table S1 for the list of standards used. See methods

and Appendix II for further details. Results of least-square linear combination fitting of all sample

XANES spectra are summarized in Table S2. Both analyses indicate that Se0 is the main product

in field-derived and culture samples.

29

Figure 5. Correlative cryogenic spectro-microscopy on a 19-day-old selenate-reducing enrichment

culture grown in bicarbonate medium. A) Low dose cryo-TEM image (3.5 nm pixels, 7.2 x 7.2

µm2) of a cluster of bacteria and associated Se CNP aggregates. B) Cryo-µXRF Se distribution

map at 13keV (2 µm beam, 1 µm pixels). The green box represents the entire TEM region. C) Low

dose cryo-TEM image at higher magnification of the blue box area in panel A. D) Se K-edge

XANES collected in the light green box. Best fit is obtained using 87% red amorphous Se and

13% sodium selenite standards. Corresponding µXRF spectrum exhibited Se with traces of Ca; µ-

XRD showed no evidence for crystalline Se0 (Figure S10). Red amorphous Se0 aggregates (~ 100

± 60 nm) is the main product of selenate reduction in this region of the sample, with minor presence

of Se(IV), suggesting a 2-step reduction process. Scale bars are 2 µm (A-B) and 200 nm (C).

30

Figure 6. Cryo-µEXAFS of a 19-day-old selenate enrichment culture grown in bicarbonate

medium A) in k-space, compared to red amorphous Se0 and hexagonal grey Se0 (Se foil) standards.

B) Shell-by-shell fitting analysis of the 1st shell with Se-Se interatomic distance at 2.339 ± 0.003

Å. The experimental spectrum is shown in blue, the fit (performed in q space) is in red and the

residual in green. Fitting results are summarized in Table S3.

31

CHAPTER 2

Dechloromonas selenatis, a Betaproteobacterium from a

contaminated aquifer that reduces selenate to amorphous selenium

32

Abstract

In order to assess the potential for selenate Se (VI) bio-reduction in an aquifer, we enriched for

organisms that can grow by coupling acetate oxidation to selenate reduction and surveyed for

selenate reductase genes. The inoculum was obtained from a uranium, vanadium, selenium and

arsenic-contaminated aquifer near Rifle, Colorado, USA. Enrichment cultures growing on acetate

and selenate were dominated by organisms related to Ferribacterium and Dechloromonas. From

these enrichments, we isolated a strain that shares 98.9% 16S rRNA gene sequence identity with

Dechloromonas aromatica RCB and grows anaerobically by oxidizing acetate and reducing

selenate. We refer to this isolate as Dechloromonas selenatis strain RGW99. Additionally, we

detected a wide variety of selenate reductases in the genomes of organisms related to

Dechloromonas spp. from this site. Cryogenic 2D transmission electron microscopy (TEM) and

STXM revealed that Se nanoparticles are either associated with cell surfaces or in colloidal

aggregates in the culture medium. 3D electron tomography revealed that these colloids do not

form inside the cytoplasm but rather originate in the cell membrane. The 1.87 ± 0.48 µm long, 555

± 62 nm diameter slightly bent rod-shaped cells contain cytoplasmic polyhydroxyalkanoates

(PHA) storage granules that comprise up to ~30% of the cell volume. PHA may contribute to

survival of D. selenatis under conditions of fluctuating nutrient availability, as occur in

groundwater systems, potentially sustaining their contribution to selenate reduction. Scanning

transmission X-ray microscopy (STXM) analyses at SeL2,3- and C K-edges detected Se0

nanoparticles near cell surfaces and trapped in the extracellular polysaccharides substances.

Cryogenic X-ray microprobe and SEM analyses showed that the end product of selenate reduction

by D. selenatis RGW99 is red amorphous Se0 (240 ± 66 nm). Se K-edge XANES spectroscopy

identified Se (IV) as an intermediate product, suggesting that the reaction proceeds in a step-wise 33

fashion via Se(IV). Overall, this study establishes a role for D. selenatis RGW99 in selenate

reduction in the Rifle aquifer and provides insight into the nature and stability of the bioreduction

products.

Introduction

Microbiological and geochemical processes controlling selenium (Se) mobility, and hence its

toxicity remain poorly understood. The speciation of Se is mainly controlled by the pH and redox

conditions of the environment. Se occurs predominantly in four oxidation states (VI, IV, 0 and –

II) in nature. The oxyanions, selenate (Se(VI), SeO42-) and selenite (Se(IV), SeO3

2−), are toxic at

relatively low concentrations (ppm) . Selenate is the most mobile Se species in groundwater due

to its high solubility and low sorption affinity at near neutral pH, and is the predominant species

in natural waters.

The biogeochemical cycle of Se in nature is not well defined1 but is predominantly

governed by microorganisms2 from all three domains of life. Dissimilatory selenate reduction

(DSeR)3 to Se0 is a major sink for Se oxyanions in anoxic sediments4, 5. In this pathway,

anaerobes2 couple the oxidation of organic matter or H2 to the reduction of Se oxyanions,

potentially forming relatively insoluble and non-toxic Se0. To date, the vast majority of studies

have suggested that microbial reduction of Se oxyanions produces Se0 either as the red monoclinic6

(Se8 rings) or red amorphous7-11 (primarily chains) phases. However, in most reports the Se phases

(amorphous vs. crystalline and if crystalline, the polymorph) and cell-mineral interactions are

rarely rigorously determined. The precise determination of selenate bio-reduction products and

34

associated microbes are crucial because it in turn impacts knowledge of the reactivity,

bioavailability and mobility of Se in the environment.

Here, we enriched for organisms able to grow on acetate with selenate as the sole electron

acceptor and investigated the distribution and speciation of selenium bio-reduction products. The

experiments used an inoculum from an aquifer near Rifle, CO, USA that is contaminated with U,

V, As and Se due to past ore-milling activities. From these enrichments, a species related to

Dechloromonas aromatica strain RCB was isolated. Dechloromonas are members of the

Rhodocyclaceae (Betaproteobacteria) that are found in many soil and aquifer environments12. This

microbe has been shown to degrade aromatic hydrocarbons, reduce perchlorate and nitrate, and

oxidize Fe(II) and H2S13, 14. Members of this genus are common in bioreactor communities used

for selenate reduction 15, 16 and have been added to consortia to enhance bioreactor performance.

Dechloromonas spp. have been repeatedly detected in the Rifle aquifer17-22 and a prior study17

suggested that Dechloromonas are major players in Se oxyanion reduction at this site. However to

date, direct proof of selenate bio-reduction by Dechloromonas spp. has been lacking. Here we

show that a Dechloromonas strain can reduce selenate. Surveys of the genetic potential of bacteria

in microbial communities from groundwater and sediment at the site indicate that many bacteria

from this genus, along with a variety of organisms from other lineages, likely reduce selenate.

Information about the structural characteristics of selenate bioreduction products provide

important constrains for predicting the stability and reactivity of Se in aquifer settings.

Materials and Methods

35

Cultivation conditions. For enrichment cultivation, we sampled groundwater from the

CG02 well in the aquifer at the Old Rifle site, near Rifle, Colorado, CO, USA. Cultures were

established either in 10 mL carbonate-buffered freshwater medium (sulfate and nitrate free) and

inoculated with sodium selenate or sodium selenite (5 mM, Sigma Aldrich) and sodium acetate

(10 mM, Sigma Aldrich). Cultures were sparged with N2:CO2 (80:20) and sealed with butyl rubber

stoppers and aluminum crimp seals, as previously described23. All enrichment cultures were

incubated at room temperature in the dark. The selenate-reducing enrichments were the primary

samples analyzed.

16S ribosomal RNA gene sequencing. DNA was extracted from culture samples using

the PowerSoil DNA Extraction Kit ((MoBio Laboratories, Inc., CA, USA) according to

manufacturer’s specifications, except that the initial lysing step was accomplished by vortexing at

maximum speed for 2 minutes and then incubating at 65 ˚C for 30 minutes, with an additional

minute of vortexing every 10 minutes. DNA was amplified using 27f and 1492r primers over a

temperature gradient, and PCR products were cleaned up using the MoBio UltraClean PCR Clean-

Up Kit. The forward and reverse sequences were overlapped for generate a 1,143 bp sequence

that was classified by placement in a phylogenetic tree.

16S rRNA gene phylogenetic analysis. 16S rRNA gene sequences were aligned using

SSU-Align24 along with the best-hits of these sequences in the SILVA database (version 115 of

non-redundant SILVA)25, a curated set of reference bacterial sequences, and a set of archaea

sequences and a set of Gammaproteobacteria to serve as a phylogenetic root. Sequences with ≥800

bp aligned were used to infer a maximum-likelihood phylogeny using RAxML26 with the

36

GTRCAT model of evolution and to construct neighbor-joining trees27. Bootstrap analyses were

based on 100 re-samplings.

Selenate reductase protein analysis. We investigated 55 metagenomic datasets from the

Rifle site to identify sequences that could potentially be classified as selenate reductases. The

predicted protein sequences were aligned with reference sequences for biochemically

characterized proteins including the following families: arsenite/arsenate reductases, polysulfide

reductases, nitrite/nitrate reductases, tetrathionate reductases, perchlorate and selenate reductases.

Proteins considered likely to be capable of reducing selenium were those that clustered with the

reference sequence from Thauera. Where possible, the sequences were linked to genomes to

enable identification of the organisms potentially involved in selenate reduction.

Cryo-plunging of samples for synchrotron and TEM analyses. Culture samples were

flash-frozen in the laboratory. Aliquots of 5 µL sample solution were deposited onto TEM Cu-

grids (200 mesh, lacey carbon coated formvar, Ted Pella Inc.) and were manually blotted with

filter paper (Grade 1 filter paper, Whatman®). TEM grids were pre-treated by glow-discharge to

improve sample deposition onto the grids. Colloidal gold particles (10 nm, BBInternational,

Cardiff, UK) were put on some of the TEM grids and allowed to dry prior to sample addition.

Samples were cryo-plunged in either liquid ethane (for TEM and synchrotron characterization) or

liquid nitrogen (for some synchrotron analyses). All samples were placed in grid boxes and stored

in a LN2 tank until analysis. Further details on the portable cryo-plunger and procedures employed

can be found elsewhere28.

Synthesis of Se colloidal nanoparticles

37

Red amorphous selenium particles were synthesized at room temperature, following a

protocol modified from Jeong et al.29. Stocks of selenous acid (70 mM) and hydrazine (0.35 M)

were prepared in ethylene glycol (anhydrous, 99.8%, Sigma-Aldrich CAS# 107-21-1). 5 mL

ethylene glycol was added in a clean open flask and stirred for 5-10 minutes on a stirring plate, 1.6

mL selenous acid (99.999% trace metals basis, Aldrich CAS# 7783-00-8) was added, then 240 µL

hydrazine (Hydrazine hydrate, reagent grade, N2H4 50-60 %, Sigma-Aldrich CAS# 10217-52-4).

After reacting for 1.5 h, the samples were rinsed three times in MQ and spun at 15,000 rpm at 4

°C for 3 min. Samples were subsequently stored in glass vials in the fridge. Samples were generally

prepared and analyzed by SEM right prior to synchrotron analyses. The particle size distribution

was estimated from SEM measurements of 150 particles.

2D cryo-TEM. Cryogenic TEM images were acquired on a JEOL–3100-FFC electron

microscope (JEOL Ltd, Akishima, Tokyo, Japan) equipped with a FEG electron source operating

at 300 kV, an Omega energy filter (JEOL), cryo-transfer stage, and a Gatan 795 4Kx4K CCD

camera (Gatan Inc., Pleasanton, CA, USA) mounted at the exit of an electron decelerator30 held at

a voltage of 200 to 250 kV. The stage was cooled with liquid nitrogen to 80 K during acquisition

of all data sets. The survey of the grids and the selection of suitable targets were performed in low

dose defocused diffraction mode. Images were recorded at different magnifications with a pixel

size ranging from 0.224 nm to 0.701 nm at the specimen. Underfocus values ranged from 0 µm ±

0.5 µm to 12 µm ± 0.5 µm, and energy filter widths were typically around 30 eV.

Additional cryo-TEM images and cryo-electron diffraction patterns were collected with a

FEI CM200 with FEG, at 200kV. For analyses on this microscope, a 5 µL droplet of culture sample

was deposited onto a TEM grid and air-dried right prior to analysis.

38

3D cryo-TEM

Seven tomographic data sets were acquired on Dechloromonas selenatis cells.

Tomographic tilt series were acquired under low dose conditions, typically over an angular range

between +65° and -65°, ± 5° with increments of 2°. Sixty four images were typically recorded for

each tilt series, acquired semi-automatically with the program Serial-EM

(http://bio3d.colorado.edu/; Kremer et al 1996) adapted to JEOL microscopes. For tilt series data

sets, all images had a pixel size of 0.56 nm at the specimen. Underfocus values ranged from 10 to

12 µm ± 0.5 µm, and energy filter widths were approximately 30 eV. For all data sets the dose

used per complete tilt series ranged from 75 to 120 e-/Å2. All tomographic reconstructions were

performed with the Imod software31 (http://bio3d.colorado.edu/). ImageJ 1.38x (Collins 2007) was

used for analysis of the 2D image projections. All movies were created using the VideoMach

software.

SEM/ EDS

Scanning electron microscopy and EDS data were collected at 5 keV and 15 keV using an

FEI/Philips XL 30 FEG-SEM, equipped with an EDX spectrometer (EDAX, Inc). The specimens

were tilted 4 degrees toward the X-ray detector to optimize the X-ray detection geometry. Working

distance varied between 7.5 and 10 mm. Collection time for EDS was 500 seconds for each area.

The lateral resolution of the microscope was 10 nm at 15 kV.

Energy Dispersive Spectroscopy (EDS). Additional elemental analyses were carried out

in TEM mode in the JEOL 2100-F Field-Emission Analytical TEM equipped with Oxford INCA

Energy Dispersive Spectroscopy (EDS) X-ray detection system at the Molecular Foundry at LBL.

39

EDS spectra were acquired for 60 or 120 live seconds with the probe diameter indicated on the

images at 200 kV.

X-ray microprobe. Micro-focused X-ray fluorescence (µXRF) mapping, X-ray

diffraction (µXRD) and Se K-edge X-ray absorption spectroscopy data were collected at ALS

bending magnet beamline 10.3.2 (2.4- 17 keV) with the storage ring operating at 500 mA and 1.9

GeV 32. All data were recorded at 95 K using a cryostage described below. Maps and µXRF spectra

were collected at 13 keV with a beam spot size ranging from 2×2 μm to 5×5 μm, and counting

times up to 200 ms/pixel. Fluorescence emission counts were recorded for S, K, Cl, Ca, Mn, Fe,

Cu, Zn, Au and Se using a seven-element Ge solid-state detector (Canberra) and XIA electronics.

Se K-edge µXANES spectra were recorded in fluorescence mode by continuously scanning the Si

(111) monochromator (Quick XAS mode) from 160 eV below up to 407 eV above the edge

(12500-13067 eV, i.e. up to k = 10 Å-1). EXAFS spectra were recorded up to 740 eV above the

edge (12500-13400 eV, i.e. up to k ~ 13.7 Å-1). Spectra were calibrated using the white line of a

red amorphous Se standard set at 12660 eV. All data were processed using LabVIEW custom

software and standard procedures described elsewhere33. Least-squares linear combination fitting

of the XANES spectra was performed as previously described34 using an updated spectral database

of Se compounds. Additionally, Se valence state scatter plots were generated from XANES data

as described below. Cryogenic µEXAFS spectra of cultures were reduced with 2 ( )k kχ weighting,

out to k = 12 Å-1. Spectra were further analyzed via shell-by-shell fitting using Artemis35, 36. Only

the first shell was fitted, as other shells were not visible enough for accurate analysis. The structure

of trigonal Se (0) as described by Keller and coworkers37 was used to create FEFF6l input files

from which to extract Se-Se paths out to 3.5 Å. and to fit the experimental t-Se EXAFS spectrum

40

at 95 K. A Kaiser-Bessel window (1.4-2.5 Å) was used to Fourier transform and fit the data in q

space.

Micro-diffraction patterns were collected at 95 K in transmission mode with a CCD camera

(Bruker SMART6000) at 17 keV (λ = 0.729 Å) using a beam spot size of 12×7 μm and exposure

times of 240 sec. A background pattern was also recorded nearby the region of interest. Calibration

of the camera distance was obtained using an alumina (α-Al2O3) powder standard and Fit2D

software38. Fit2D was also used to obtain one-dimensional XRD profiles from the radial

integration of 2D patterns. XRD peaks were indexed using Jade 9 software (Materials Data Inc.)

and the ICDD PDF-4+ database. Additional cards from Mincryst for red monoclinic Se0, grey

trigonal-Se0 and ice at were added to the database.

Selenium valence state scatter plot. We used a method similar in concept to that

described in Marcus et al.39 (for Fe species) to evaluate the valence state of Se. Normalized

XANES spectra were processed using a new custom Matlab program. Spectra were reduced to two

variables, a and b, defined as the normalized XANES absorption values at 12664.25 eV and

12667.80 eV respectively. These energies were determined empirically to obtain the best

separation between chemical families in the (a,b) scatter plots. Selenate (Se(VI)), selenite Se(IV),

elemental Se(0) and Se(-II) selenide/Se(-I)/Organic Se families are found to separate, as evidenced

with Se standards of known valence, plotted in black. Using this method, unknown experimental

XANES data could be at first sight classified according to valence state. Least square linear

combination fitting is still required to confirm the identity and valence state of the unknown. All

standards (see Table S1 in chapter 1), were measured at ALS beamline 10.3.2, either at room

41

temperature with samples mounted on kapton tape, or at 95 K using a cryo-stage, described in the

next section, with samples mounted on Si3N4 windows (TEM windows, SiMPore Inc.).

Scanning Transmission X-ray Microscopy (STXM). STXM measurements were

performed on the soft X-ray beamline 11.0.2 (150-2000 eV) of the Advanced Light Source,

Lawrence Berkeley National Lab, Berkeley, CA40. This microscope employs a Fresnel zone plate

lens (25 nm outer zones) to focus a monochromatic soft X-ray beam onto the sample. The sample

is raster-scanned in 2D through the fixed beam and transmitted photons are detected via a phosphor

scintillator-photomultiplier assembly. X-ray images recorded at energies just below and at the

elemental absorption edge of interest (Se L3 and C K) were converted into optical density (OD)

images and used to derive elemental maps (OD = ln (I0/I), where I0 is the incident X-ray intensity

and I is the transmitted beam intensity through the sample. Chemical maps were obtained by taking

the difference of OD images at 280, 288.2 eV (proteins); 289, 290.3 eV (carbonates); and 1425,

1440 eV (Se). Image sequences (‘stacks’) recorded at energies spanning the Se L2,3-edges (1420-

1520 eV) and C K-edge (280-320 eV) were used to obtain XANES spectra from regions of interest.

Se L2,3-edges XANES spectra are sensitive to the oxidation state and the local bonding

environment of Se. Se spectra were compared with a spectral library of model compounds (see

Table S1). The main Se L3 resonance of the red amorphous Se (0) standard was set to 1435 eV and

used for relative energy calibration of the spectra. At least two different sample regions were

analyzed for each element. The theoretical spectral and spatial resolutions during measurements

were +/-100 meV and 30 nm respectively. The photon energy was calibrated at the C K-edge using

the Rydberg transition of gaseous CO2 at 292.74 eV (C 1s→ 3s (ν = 0)). All measurements were

performed at ambient temperature under He at pressure < 1 atm. All data were processed with the

42

aXis2000 software (http://unicorn.mcmaster.ca/aXis2000.html). Note that this STXM does not

have cryo-capabilities and frozen samples were thawed and air-dried at room temperature just prior

to measurements.

Results and Discussion

Anaerobic selenate-respiring organisms were enriched using acetate as the electron donor

and carbon source and selenate as the sole electron acceptor (see methods). A biofilm developed

and adhered to the walls of the vials. Red precipitates were visually observed after two days,

suggesting selenate reduction to Se0. The isolate was obtained from ten serial dilutions of these

enrichments. In addition to growth on selenate and selenite (see below), we tested for growth on

arsenate. The isolate did not grow on 5 mM arsenate (data not shown), but significant arsenite

production was detected in a 1 day-old culture prior to its demise (Fig. S1).

16S rRNA gene sequence phylogeny

Phylogenetic analysis of the 16S rRNA gene sequence of the biomass revealed that the

isolate is closely related to Ferribacterium (98.3% gene similarity) and Dechloromonas species

(98.9% gene similarity to Dechloromonas RCB). The placement of the isolate RGW99 in both the

neighbor-joining (Fig. 1) and the maximum likelihood trees (Fig. 2) indicates that it is most closely

related to Dechloromonas aromatica, but distinct from it. Given this, we propose this organism

to be a new species within the Rhodocyclaceae family of the Betaproteobacteria. Since the most

closely related identified organism is Dechloromonas aromatica RCB, we propose the name

Dechloromonas selenatis RGW99 for the isolate described here, based on its demonstrated ability 43

to reduce selenate. Both Ferribacterium41, an Fe reducer and Dechloromonas have been

previously detected at the Rifle site17, 18 19-22, including in acetate-amended Rifle groundwater

samples42,22. Dechloromonas sp.12 are well-known perchlorate- and nitrate-reducing bacteria13, 43.

D. aromatica was reported as the first known organism capable of degrading benzene

anaerobically12 and has been shown since to be able to degrade a variety of complex hydrocarbon

compounds44. This genus is widespread in contaminated environments45, 46 and has been shown to

promote the reduction of selenate in bioreactors15, 16, 47.

Ultrastructural description

Cryo-TEM data (Fig. 3 and 4) reveal slightly bent rod-shaped cells with a Gram-negative

cell wall consisting of an inner membrane, an outer membrane and a peptidoglycan layer within a

periplasmic space, measured to vary between 20 and 40 nm. A single polar flagellum (e.g. Fig.

3A) and sometimes pili can be observed. From 2D cryo-TEM images and 3D electron tomograms,

the cells were found to be 1.87 ± 0.48 µm long and 555 ± 62 nm in diameter. Cells contain

organelles with sizes and shapes consistent with ribosomes (Fig. 4). In addition, most cells

exhibited many (often > 10) cytoplasmic granules that range in size from 50 to 300 nm and occupy

about ~ 30% of the cell volume. STXM-derived C K-edge XANES spectra of granules (not shown)

exhibit a major resonance at 288.4 eV, associated with carboxyls or esters48, 49. These granules are

likely carbon and energy storage polymers in the family of polyhydroxyalkanoates (PHA,

polyesters). PHAs have received increasing interest over the past decades due their potential as a

renewable source of biodegradable thermoplastics. These cytoplasmic organelles are often

produced by bacteria under unbalanced nutrient conditions when the carbon source is in excess50,

51. The granules may help the cells survive under low nutrient conditions52, 53. A wide variety of

44

prokaryotes and eukaryotes are known to be able to produce PHAs54. These polyester inclusions,

discovered almost 90 years ago in Bacillus megaterium, have been previously extensively studied

in bacteria such as Ralstonia spp.51, 55, Pseudomonas spp.56, Bacillus spp.57, 58. The presence of

carbon storage granules in D. aromatica has been previously reported13, 59-61. Furthermore, 2D

cryo-TEM images (e.g. Fig. 3A inset) show that these granules are coated with a 15-20 nm thick

layer and all have 4 nm-wide filamentous structures protruding from them. The filaments either

extend towards the cell wall and some traverse or interconnect granules. It is known that PHA

granules are coated with a large number of proteins (mostly phasins62)63, 64 and possibly minor

traces of phospholipids54. Interestingly, these filaments have a diameter similar to that of phasin

PhaP1, but the analysis required to confirm this is beyond current instrument capabilities. Further,

few but more electron dense granules, possibly composed of polyphosphates, were also observed.

These are often located near the poles of the cells. Lastly, electron dense Se particles and

aggregates of particles were found associated with cell surfaces or dispersed in the medium (e.g.

Fig. 3B). Electron tomograms (Movies S1 and S2) show that these particles do not nucleate inside

the cytoplasm of the cells but are rather connected to the inner membrane through the perisplasmic

space.

Selenium and carbon distribution and speciation

EDS spectra of the particles imaged by SEM showed the presence of only Se (Fig. 5; Cu

peaks were associated with the Cu TEM grid; C and O peaks originated from the formvar support

from the grid). Particles had a size distribution centered at ~240 nm in a 3 month-old culture (Fig.

6), compared to ~100 nm in a 1 day-old sample, suggesting particle aggregation over time.

Additional particle aggregation is minimal after 8 days. STXM imaging revealed a few particles

45

in contact with the cell surfaces or trapped in the EPS matrix (Fig. 7), consistent with TEM data.

C K-edge spectra show EPS, with a main absorption peak at 288.8 eV associated with the π∗(C=O)

transition of carboxyl group in acidic polysaccharides, and carbonates, originating from

bicarbonate in the culture medium, with a signature π∗(C=O) peak at ~ 290.3 eV. Se L2,3 edge

XANES spectra of particles (not shown) compared with spectra from model compounds identified

Se0, however the allotropic form of Se0 could not be determined with STXM.

Using previously described methods (see methods and Chapter 1), Se valence plots and

least-square linear combination (LSQ) fitting of the Se K-edge XANES spectra show that red

amorphous Se0 is the major end product, and that selenite is an intermediate species that persists

throughout reduction. Cryogenic XRF mapping and XANES were performed on 27 Se-rich regions

from several cryo-TEM grids at different time points during cultivation, spanning from 1 to 85

days. Se valence plots (Fig. 8), LSQ fitting of the spectra (Table 1) and cryogenic electron

diffraction patterns (Fig. 5C) indicate that red amorphous Se0 is the major end product. The results

suggest that reduction proceeds from Se(VI) to Se(IV) to Se(0). Both cryogenic µXRD and

electron diffraction patterns do not match crystalline Se0 (either grey hexagonal or red monoclinic)

(Fig. 3).

We compared cryogenic µEXAFS spectra of Se0 aggregates from 1.5 and 3-month old

cultures of D. selenatis RGW99 and our chemically synthesized Se0 with the crystal structure of

hexagonal Se, as determined by Keller et al.37 (Fig. 11 and Table 2). The EXAFS spectra were

fitted using the Artemis software36, 65 and only the first shell could be fitted by a shell-by-shell

method. The second shell was not sufficiently visible to obtain an accurate fit due to lack of

structural order, thus the Se-Se-Se bond angle could not be determined. The 1st shell (Se-Se bond) 46

lies at 2.348 ± 0.002 Å for synthetic Se0, 2.346 ± 0.002 Å for 1.5 month-old D. selenatis and 2.350

± 0.002 Å for the 3 month-old D. selenatis. These inter-atomic distances are shorter than the bond

length in hexagonal Se0 at 2.374 Å, and are consistent with prior reports for amorphous Se66,67.

The coordination numbers (N) and sigma square (σ2) values for the synthetic Se, 1.5- and 3 month-

old D. selenatis samples are very similar. Even though the 3 month-old D. selenatis sample exhibits

the longest bond length among the three samples, the inter-atomic distances have error bars that

are wide enough so that the Se-Se bond length in the 1.5, 3 month-old and synthetic Se0 cannot be

distinguished. These results suggest that the end product of selenate reduction by D. selenatis

remains stable for at least several months.

Organisms potentially capable of selenate reduction at the Rifle site

Based on analysis of metagenomic data (see the Rifle database at http://ggkbase.berkeley.edu/),

we find that a wide variety of organisms may be capable of selenate reduction at the site (Fig. 11).

In addition to sequences from organisms closely related to Dechloromonas (e.g., RAAC16_31_9),

we detected selenate reductase-like genes in draft genomes from a variety of other bacteria. For

example, genes clustered with the reference Thauera sequence were identified in an organism

affiliated with candidate phylum NC10 (gwa2_scaffold_28025_4), several novel

Betaproteobacteria (RBG_19_scaffold_2003_4; RBG_19_scaffold_1677_5;

gwa2_scaffold_1639_12), Burkholderia (rifcsplowo2_02_sub10_scaffold_37863), a

Gammaproteobacterium (RBG_16_scaffold_32990_7) and several novel (unclassified) bacteria

(RBG_16_scaffold_9846; 16ft_4_scaffold_10874; rbg_19ft_combo_scaffold_1584). At this time,

the scale of the selenate reductase group is difficult to define due to the lack of phylogenetically

47

diverse reference sequences. The detection of the selenate-reductase-like sequences may indicate

a wide distribution of this function in the subsurface.

Conclusion

Dechloromonas selenatis RGW99 isolated from groundwater is capable of growth using

acetate as the sole electron donor and selenate as the sole electron acceptor. Spectroscopic and

microscopic data indicate that growth by this organism results in production of ~250 nm diameter

red amorphous Se0 aggregates that are both associated with the surfaces of cells and extracellular

acid polysaccharides. These products persist over timescales of months. Sequence data indicate

that this organism belongs to a group of related species that likely contribute to selenate

bioreduction in this and other environments.

Acknowledgments

This material is partially based upon work supported through the Lawrence Berkeley

National Laboratory’s Sustainable Systems Scientific Focus Area. The U.S. Department of Energy

(DOE), Office of Science, Office of Biological and Environmental Research funded the work

under contracts DE-SC0004733 (University of California) and DE-AC02-05CH11231 (LBL). Part

of the equipment used for this study was funded by the Center for Nanoscale Control of Geologic

CO2, an LBL Energy Frontier Research Center. S.F. thanks Sue Spaulding for lab support, Tolek

Tylizscack (LBL) for support at ALS 11.0.2, Mary K. Gilles (LBL) and Anna Butterworth (SSL)

for use of their SEM, Ken H. Downing (LBL) for providing cryo-TEM infrastructure and Dan

48

Strawn (U. Idaho) and Elizabeth Pilon-Smits (U. Colorado) for sharing selenium standard spectra.

Work at the Advanced Light Source and the Molecular Foundry is supported by the Office of Basic

Energy Sciences, Office of Science of the U.S. Department of Energy under Contract No. DE-

AC02-05CH11231.

49

Figure 1: Neighbor-joining tree based on 16S rRNA gene sequences showing the phylogenetic

relationships of Dechloromonas selenatis (colored in red) to other organisms from the

Rhodocyclaceae family. Bootstrap values (percentages of 100 replicates) are given at nodes; only

values >50% are shown. Gammaproteobacteria were used as the outgroup.

50

Figure 2. Maximum-likelihood 16S rRNA gene phylogeny. Sequences with ≥ 800 bp aligned were

used to infer a maximum-likelihood phylogeny using RAxML with the GTRCAT model of

evolution and 100 bootstrap re-samplings. The red star represent the 16S rRNA sequence

recovered in this study.

51

Figure 3. 2D cryo-TEM images of Dechloromonas selenatis RGW99 and electron dense colloidal

particles and aggregates. A) Low dose image of a bacterium showing a polar flagellum and Se

particles. Inset shows 4 nm diameter filaments (indicated by red arrows) protruding out of a PHA

granule. B) Electron dense Se particles and aggregates in contact with cell surfaces. Micro tubular-

like structures and vesicular products can be observed. Ten nm gold colloidal nanoparticles are

used as fiducials. C) High resolution cryogenic image of dry Se nanoparticles and D)

corresponding cryogenic electron diffraction pattern showing no crystallinity. Scale bars are 200

nm unless specified otherwise.

52

Figure 4. 3D cryo-TEM one-voxel-thick slices of the reconstruction showing A) 60 to 250 nm

diameter PHA granules in a D. selenatis RGW99 cell. B) Se0 particles appear to be connected to

the periplasmic space toward the inner membrane. See full 3D reconstruction movies S1-S2.

53

Figure 5. A) TEM images of Se0 particles and B) Corresponding energy dispersive spectrum from

this region showing that the particles contain Se only. Cu peaks derive from the Cu grid, C and O

peaks are associated with the formvar and carbon coating. C) Electron diffraction pattern of region

shown in panel A, evidencing no crystallinity.

54

Figure 6. Size distribution of the selenium particles in 1 day-old (A) 8 days-old (B) and 1 month-

old (C) D. selenatis RGW99 cultures. Distributions were obtained from SEM measurements of

134, 153 and 320 particles respectively. SEM images of red amorphous Se0 produced by D.

selenatis RGW99 (D, E) and by chemical synthesis (F). Size distribution of the 193 ± 23.2 nm

synthetic selenium particles (G).

55

Figure 7. STXM measurements of a 1.5 month-old culture of D. selenatis RGW99. A) STXM

image at 280 eV, showing bacteria and few Se particles (indicated by blue arrows). B) Chemical

map of the area in A) showing the distribution of protein (in red), extracellular polymeric

substances (in green) and carbonate (in blue). Particles not attached to cell surfaces are trapped in

carbonate-rich extracellular polysaccharides substances. C) C K-edge spectra of cells and

extracellular substances (in black) compared to standards (in grey). Vertical dash line refers to

proteins (in red, 288.2 eV), acidic polysaccharides (in green, 288.8 eV) and carbonates (in blue,

290.4 eV). Scale bar is 1 micron.

56

Figure 8. A) Se valence scatter plot showing reference compounds of known valence in black, and

D. selenatis RGW99 data points from several time points (1 day to 85 days-old). Accumulation of

selenite is evidenced. Se0 is the major end product. Note that over time, the product changes from

a mixture of selenate and selenite to a mixture of elemental selenium and selenite.

57

Figure 9. Micro-EXAFS at 95 K of 1.5 and 3 month-old cultures of D. selenatis RGW99. A)

Spectra in k-space, compared to standards: hexagonal Se0 (Se foil), biogenic red amorphous Se0

and synthetic 193 nm red amorphous Se0. B) Overlay of the Fourier transformed spectra of Se

aggregates from a 3 month-old culture with synthetic red amorphous Se0, shows a close match and

indicates high stability of the biogenic end product.

58

Figure 10. Micro-EXAFS at 95 K of Dechloromonas selenatis RGW99, compared to synthetic

193 nm red amorphous Se0. Shell-by-shell co-refinement fitting analysis of the 1st shell with Se-

Se shell interatomic distances is reported in Table 2. Experimental spectra are shown in blue, fits

in red and the Kaiser-Bessel window used for the fit in green.

59

Figure 11. Maximum-likelihood tree showing the phylogeny of selenate reductase and other

related Mo-bearing proteins. Sequences with ≥ 400 bp aligned were used to infer a maximum-

likelihood phylogeny using RAxML with the GTRCAT model of evolution and 100 bootstrap re-

samplings (only values >50% are shown). Putative selenate reductase sequences recovered from

the genomes of bacteria from the Rifle site are shown in red. Alanine aminotransferase complex

III were used as the outgroup.

60

Figure S1. Se K-edge XANES at 95 K of a 1 day-old D. selenatis RGW99 culture in bicarbonate

medium supplemented with 5 mM sodium arsenate and 10 mM acetate. Results of LSQ linear

combination fitting using sodium arsenate, As (V) and sodium arsenite, As (III) standards indicate

partial reduction of As (V) to As (III).

61

Table 1. Summary of results of LSQ linear combination fitting of XANES spectra on D.

selenatis RGW99 cultures. Standards compounds employed are listed in Table S1 of chapter 2.

Sample Red am. Se

(%)

Selenite

(%)

Selenate

(%)

Other

(%)

Cultures

Bicarb (n=27) 62.6 31.1 2.6 3.7

ND= not detected, n= number of spots measured.

Table 2. EXAFS co-refinement fitting. Fits were performed in q space, in the k=2-12 Å-1 range,

with k-weight = 2, using a Kaiser-Bessel window and dk=1. R range used was 1.4-2.5 Å.

3 month-old D. selenatis Shell N S0

2 σ2 E0 delr Reff R ================================================================================= Se-Se 2.000 0.957 0.00232 -0.711 -0.02501 2.37470 2.34969 Synthetic Se 193 nm Shell N S0

2 σ2 E0 delr Reff R ================================================================================= Se-Se 2.000 0.957 0.00227 -0.711 -0.02627 2.37470 2.34843 1.5 month old D. selenatis Shell N S0

2 σ2 E0 delr Reff R ================================================================================= Se-Se 2.000 0.957 0.00195 -0.711 -0.02890 2.37470 2.34580

Note: σ2 = mean-square disorder of neighbor distance, R = distance to neighboring atom, N = coordination number of

neighboring atom, ∆E0 = shift in threshold energy E0, S02 = amplitude reduction term.

62

References

1. Shrift, A., A Selenium Cycle in Nature? Nature 1964, 201, (4926), 1304-1305. 2. Oremland, R. S.; Blum, J. S.; Culbertson, C. W.; Visscher, P. T.; Miller, L. G.; Dowdle, P.; Strohmaier, F. E., Isolation, Growth, and Metabolism of an Obligately Anaerobic, Selenate-Respiring Bacterium, Strain SES-3. Applied and Environmental Microbiology 1994, 60, (8), 3011-3019. 3. Lovley, D. R., Dissimilatory Metal Reduction. Annual Review of Microbiology 1993, 47, (1), 263-290. 4. Oremland, R. S.; Hollibaugh, J. T.; Maest, A. S.; Presser, T. S.; Miller, L. G.; Culbertson, C. W., Selenate Reduction to Elemental Selenium by Anaerobic Bacteria in Sediments and Culture: Biogeochemical Significance of a Novel, Sulfate-Independent Respiration. Applied and Environmental Microbiology 1989, 55, (9), 2333-2343. 5. Steinberg, N. A.; Oremland, R. S., Dissimilatory Selenate Reduction Potentials in a Diversity of Sediment Types. Applied and Environmental Microbiology 1990, 56, (11), 3550-3557. 6. Oremland, R. S.; Herbel, M. J.; Blum, J. S.; Langley, S.; Beveridge, T. J.; Ajayan, P. M.; Sutto, T.; Ellis, A. V.; Curran, S., Structural and Spectral Features of Selenium Nanospheres Produced by Se-Respiring Bacteria. Applied and Environmental Microbiology 2004, 70, (1), 52-60. 7. Debieux, C. M.; Dridge, E. J.; Mueller, C. M.; Splatt, P.; Paszkiewicz, K.; Knight, I.; Florance, H.; Love, J.; Titball, R. W.; Lewis, R. J.; Richardson, D. J.; Butler, C. S., A bacterial process for selenium nanosphere assembly. Proceedings of the National Academy of Sciences 2011, 108, (33), 13480-13485. 8. Hunter, W.; Manter, D., Reduction of Selenite to Elemental Red Selenium by Pseudomonas sp. Strain CA5. Curr Microbiol 2009, 58, (5), 493-498. 9. Kessi, J.; Ramuz, M.; Wehrli, E.; Spycher, M.; Bachofen, R., Reduction of Selenite and Detoxification of Elemental Selenium by the Phototrophic BacteriumRhodospirillum rubrum. Applied and Environmental Microbiology 1999, 65, (11), 4734-4740. 10. Tomei, F. A.; Barton, L. L.; Lemanski, C. L.; Zocco, T. G., Reduction of selenate and selenite to elemental selenium by Wolinella succinogenes. Canadian Journal of Microbiology 1992, 38, (12), 1328-1333. 11. Tomei, F.; Barton, L.; Lemanski, C.; Zocco, T.; Fink, N.; Sillerud, L., Transformation of selenate and selenite to elemental selenium byDesulfovibrio desulfuricans. Journal of Industrial Microbiology 1995, 14, (3-4), 329-336. 12. Coates, J. D.; Chakraborty, R.; Lack, J. G.; O'Connor, S. M.; Cole, K. A.; Bender, K. S.; Achenbach, L. A., Anaerobic benzene oxidation coupled to nitrate reduction in pure culture by two strains of Dechloromonas. Nature 2001, 411, (6841), 1039-1043. 13. Salinero, K.; Keller, K.; Feil, W.; Feil, H.; Trong, S.; Di Bartolo, G.; Lapidus, A., Metabolic analysis of the soil microbe Dechloromonas aromatica str. RCB: indications of a surprisingly complex life-style and cryptic anaerobic pathways for aromatic degradation. BMC Genomics 2009, 10, (1), 351. 14. Chakraborty, A.; Picardal, F., Neutrophilic, nitrate-dependent, Fe(II) oxidation by a Dechloromonas species. World J Microbiol Biotechnol 2013, 29, (4), 617-623. 15. Zhang, Y.; Frankenberger Jr, W. T., Supplementing Bacillus sp. RS1 with Dechloromonas sp. HZ for enhancing selenate reduction in agricultural drainage water. Science of The Total Environment 2007, 372, (2–3), 397-405. 16. Chung, J.; Ryu, H.; Abbaszadegan, M.; Rittmann, B. E., Community structure and function in a H(2)-based membrane biofilm reactor capable of bioreduction of selenate and chromate. Appl Microbiol Biotechnol 2006, 72, (6), 1330-1339. 17. Williams, K. H.; Wilkins, M. J.; N'Guessan, A. L.; Arey, B.; Dodova, E.; Dohnalkova, A.; Holmes, D.; Lovley, D. R.; Long, P. E., Field evidence of selenium bioreduction in a uranium-contaminated aquifer. Environmental Microbiology Reports 2013, 5, (3), 444-452.

63

18. Mouser, P. J.; N’Guessan, A. L.; Elifantz, H.; Holmes, D. E.; Williams, K. H.; Wilkins, M. J.; Long, P. E.; Lovley, D. R., Influence of Heterogeneous Ammonium Availability on Bacterial Community Structure and the Expression of Nitrogen Fixation and Ammonium Transporter Genes during in Situ Bioremediation of Uranium-Contaminated Groundwater. Environmental Science & Technology 2009, 43, (12), 4386-4392. 19. N'Guessan, A. L.; Elifantz, H.; Nevin, K. P.; Mouser, P. J.; Methe, B.; Woodard, T. L.; Manley, K.; Williams, K. H.; Wilkins, M. J.; Larsen, J. T.; Long, P. E.; Lovley, D. R., Molecular analysis of phosphate limitation in Geobacteraceae during the bioremediation of a uranium-contaminated aquifer. ISME J 2009, 4, (2), 253-266. 20. Handley, K. M.; Wrighton, K. C.; Piceno, Y. M.; Andersen, G. L.; DeSantis, T. Z.; Williams, K. H.; Wilkins, M. J.; N'Guessan, A. L.; Peacock, A.; Bargar, J.; Long, P. E.; Banfield, J. F., High-density PhyloChip profiling of stimulated aquifer microbial communities reveals a complex response to acetate amendment. FEMS Microbiology Ecology 2012, 81, (1), 188-204. 21. Wrighton, K. C.; Thomas, B. C.; Sharon, I.; Miller, C. S.; Castelle, C. J.; VerBerkmoes, N. C.; Wilkins, M. J.; Hettich, R. L.; Lipton, M. S.; Williams, K. H.; Long, P. E.; Banfield, J. F., Fermentation, Hydrogen, and Sulfur Metabolism in Multiple Uncultivated Bacterial Phyla. Science 2012, 337, (6102), 1661-1665. 22. Wrighton, K. C.; Castelle, C. J.; Wilkins, M. J.; Hug, L. A.; Sharon, I.; Thomas, B. C.; Handley, K. M.; Mullin, S. W.; Nicora, C. D.; Singh, A.; Lipton, M. S.; Long, P. E.; Williams, K. H.; Banfield, J. F., Metabolic interdependencies between phylogenetically novel fermenters and respiratory organisms in an unconfined aquifer. ISME J 2014, 8, (7), 1452-1463. 23. Coates, J. D.; Phillips, E. J.; Lonergan, D. J.; Jenter, H.; Lovley, D. R., Isolation of Geobacter species from diverse sedimentary environments. Applied and Environmental Microbiology 1996, 62, (5), 1531-6. 24. Nawrocki, E. P. Structural rna homology search and alignment using covariance models. Washington University, 2009. 25. Quast, C.; Pruesse, E.; Yilmaz, P.; Gerken, J.; Schweer, T.; Yarza, P.; Peplies, J.; Glöckner, F. O., The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Research 2013, 41, (D1), D590-D596. 26. Stamatakis, A., RAxML Version 8: A tool for Phylogenetic Analysis and Post-Analysis of Large Phylogenies. Bioinformatics 2014. 27. Saitou, N.; Nei, M., The neighbor-joining method: a new method for reconstructing phylogenetic trees. Molecular biology and evolution 1987, 4, (4), 406-25. 28. Comolli, L. R.; Duarte, R.; Baum, D.; Luef, B.; Downing, K. H.; Larson, D. M.; Csencsits, R.; Banfield, J. F., A portable cryo-plunger for on-site intact cryogenic microscopy sample preparation in natural environments. Microscopy Research and Technique 2012, 75, (6), 829-836. 29. Jeong, U.; Xia, Y., Synthesis and Crystallization of Monodisperse Spherical Colloids of Amorphous Selenium. Advanced Materials 2005, 17, (1), 102-106. 30. Downing, K. H.; Mooney, P. E., A charge coupled device camera with electron decelerator for intermediate voltage electron microscopy. Review of Scientific Instruments 2008, 79, (4), -. 31. Kremer, J. R.; Mastronarde, D. N.; McIntosh, J. R., Computer Visualization of Three-Dimensional Image Data Using IMOD. Journal of Structural Biology 1996, 116, (1), 71-76. 32. Marcus, M. A.; MacDowell, A. A.; Celestre, R.; Manceau, A.; Miller, T.; Padmore, H. A.; Sublett, R. E., Beamline 10.3.2 at ALS: a hard X-ray microprobe for environmental and materials sciences. Journal of Synchrotron Radiation 2004, 11, (3), 239-247. 33. Kelly, S. D., Hesterberg, D., & Ravel, B, Analysis of Soils and Minerals Using X-ray Absorption Spectroscopy. In Mineralogical Methods, Drees, A. L. U. L. R., Ed. Soil Science Society of America: Madison, WI, 2008; Vol. Part 5, p 367.

64

34. Bañuelos, G. S.; Fakra, S. C.; Walse, S. S.; Marcus, M. A.; Yang, S. I.; Pickering, I. J.; Pilon-Smits, E. A. H.; Freeman, J. L., Selenium Accumulation, Distribution, and Speciation in Spineless Prickly Pear Cactus: A Drought- and Salt-Tolerant, Selenium-Enriched Nutraceutical Fruit Crop for Biofortified Foods. Plant Physiology 2011, 155, (1), 315-327. 35. Newville, M., IFEFFIT: interactive XAFS analysis and FEFF fitting. J. Synchrotron Radiation 2001, 8, (2), 322-324. 36. Ravel, B.; Newville, M., ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT. J. Synchrotron Radiation 2005, 12, (4), 537-541. 37. Keller, R.; Holzapfel, W. B.; Schulz, H., Effect of pressure on the atom positions in Se and Te. Physical Review B 1977, 16, (10), 4404-4412. 38. Hammersley, A. P.; Svensson, S. O.; Hanfland, M.; Fitch, A. N.; Hausermann, D., Two-dimensional detector software: From real detector to idealised image or two-theta scan. High Pressure Research 1996, 14, (4-6), 235-248. 39. Marcus, M. A.; Westphal, A. J.; Fakra, S. C., Classification of Fe-bearing species from K-edge XANES data using two-parameter correlation plots. J Synchrotron Radiat 2008, 15, (Pt 5), 463-8. 40. Kilcoyne, A. L.; Tyliszczak, T.; Steele, W. F.; Fakra, S.; Hitchcock, P.; Franck, K.; Anderson, E.; Harteneck, B.; Rightor, E. G.; Mitchell, G. E.; Hitchcock, A. P.; Yang, L.; Warwick, T.; Ade, H., Interferometer-controlled scanning transmission X-ray microscopes at the Advanced Light Source. J Synchrotron Radiat 2003, 10, (Pt 2), 125-36. 41. Cummings, D. E.; Caccavo Jr, F.; Spring, S.; Rosenzweig, R. F., Ferribacterium limneticum, gen. nov., sp. nov., an Fe(III)-reducing microorganism isolated from mining-impacted freshwater lake sediments. Arch Microbiol 1999, 171, (3), 183-188. 42. Kerkhof, L. J.; Williams, K. H.; Long, P. E.; McGuinness, L. R., Phase Preference by Active, Acetate-Utilizing Bacteria at the Rifle, CO Integrated Field Research Challenge Site. Environmental Science & Technology 2011, 45, (4), 1250-1256. 43. Achenbach, L. A.; Michaelidou, U.; Bruce, R. A.; Fryman, J.; Coates, J. D., Dechloromonas agitata gen. nov., sp. nov. and Dechlorosoma suillum gen. nov., sp. nov., two novel environmentally dominant (per)chlorate-reducing bacteria and their phylogenetic position. International Journal of Systematic and Evolutionary Microbiology 2001, 51, (2), 527-33. 44. Chakraborty, R.; O'Connor, S. M.; Chan, E.; Coates, J. D., Anaerobic Degradation of Benzene, Toluene, Ethylbenzene, and Xylene Compounds by Dechloromonas Strain RCB. Applied and Environmental Microbiology 2005, 71, (12), 8649-8655. 45. Spain, A. M.; Krumholz, L. R., Nitrate-Reducing Bacteria at the Nitrate and Radionuclide Contaminated Oak Ridge Integrated Field Research Challenge Site: A Review. Geomicrobiology Journal 2011, 28, (5-6), 418-429. 46. Haller, L.; Tonolla, M.; Zopfi, J.; Peduzzi, R.; Wildi, W.; Poté, J., Composition of bacterial and archaeal communities in freshwater sediments with different contamination levels (Lake Geneva, Switzerland). Water Research 2011, 45, (3), 1213-1228. 47. Zhao, H.-P.; Van Ginkel, S.; Tang, Y.; Kang, D.-W.; Rittmann, B.; Krajmalnik-Brown, R., Interactions between Perchlorate and Nitrate Reductions in the Biofilm of a Hydrogen-Based Membrane Biofilm Reactor. Environmental Science & Technology 2011, 45, (23), 10155-10162. 48. Boyce, C. K.; Cody, G. D.; Feser, M.; Jacobsen, C.; Knoll, A. H.; Wirick, S., Organic chemical differentiation within fossil plant cell walls detected with X-ray spectromicroscopy. Geology 2002, 30, (11), 1039-1042. 49. Cody, G. D.; Ade, H.; Wirick, S.; Mitchell, G. D.; Davis, A., Determination of chemical-structural changes in vitrinite accompanying luminescence alteration using C-NEXAFS analysis. Organic Geochemistry 1998, 28, (7–8), 441-455.

65

50. Dohnalkova, A. C.; Marshall, M. J.; Arey, B. W.; Williams, K. H.; Buck, E. C.; Fredrickson, J. K., Imaging Hydrated Microbial Extracellular Polymers: Comparative Analysis by Electron Microscopy. Applied and Environmental Microbiology 2011, 77, (4), 1254-1262. 51. Beeby, M.; Cho, M.; Stubbe, J.; Jensen, G. J., Growth and Localization of Polyhydroxybutyrate Granules in Ralstonia eutropha. Journal of Bacteriology 2012, 194, (5), 1092-1099. 52. Anderson, A. J.; Dawes, E. A., Occurrence, metabolism, metabolic role, and industrial uses of bacterial polyhydroxyalkanoates. Microbiological Reviews 1990, 54, (4), 450-472. 53. Madison, L. L.; Huisman, G. W., Metabolic Engineering of Poly(3-Hydroxyalkanoates): From DNA to Plastic. Microbiology and Molecular Biology Reviews 1999, 63, (1), 21-53. 54. Pötter, M.; Steinbüchel, A., Biogenesis and Structure of Polyhydroxyalkanoate Granules. In Inclusions in Prokaryotes, Shively, J., Ed. Springer Berlin Heidelberg: 2006; Vol. 1, pp 109-136. 55. Pötter, M.; Madkour, M. H.; Mayer, F.; Steinbüchel, A., Regulation of phasin expression and polyhydroxyalkanoate (PHA) granule formation in Ralstonia eutropha H16. Microbiology 2002, 148, (8), 2413-2426. 56. Galán, B.; Dinjaski, N.; Maestro, B.; de Eugenio, L. I.; Escapa, I. F.; Sanz, J. M.; García, J. L.; Prieto, M. A., Nucleoid-associated PhaF phasin drives intracellular location and segregation of polyhydroxyalkanoate granules in Pseudomonas putida KT2442. Molecular Microbiology 2011, 79, (2), 402-418. 57. McCool, G. J.; Fernandez, T.; Li, N.; Cannon, M. C., Polyhydroxyalkanoate inclusion-body growth and proliferation in Bacillus megaterium. FEMS Microbiology Letters 1996, 138, (1), 41-48. 58. Valappil, S. P.; Peiris, D.; Langley, G. J.; Herniman, J. M.; Boccaccini, A. R.; Bucke, C.; Roy, I., Polyhydroxyalkanoate (PHA) biosynthesis from structurally unrelated carbon sources by a newly characterized Bacillus spp. Journal of Biotechnology 2007, 127, (3), 475-487. 59. Liu, Z.; Wang, Y.; He, N.; Huang, J.; Zhu, K.; Shao, W.; Wang, H.; Yuan, W.; Li, Q., Optimization of polyhydroxybutyrate (PHB) production by excess activated sludge and microbial community analysis. Journal of Hazardous Materials 2011, 185, (1), 8-16. 60. Chourey, K.; Nissen, S.; Vishnivetskaya, T.; Shah, M.; Pfiffner, S.; Hettich, R. L.; Löffler, F. E., Environmental proteomics reveals early microbial community responses to biostimulation at a uranium- and nitrate-contaminated site. PROTEOMICS 2013, 13, (18-19), 2921-2930. 61. Oshiki, M.; Onuki, M.; Satoh, H.; Mino, T., PHA-accumulating microorganisms in full-scale wastewater treatment plants. Water science and technology : a journal of the International Association on Water Pollution Research 2008, 58, (1), 13-20. 62. Steinbüchel, A.; Valentin, H. E., Diversity of bacterial polyhydroxyalkanoic acids. 1995; Vol. 128, p 219-228. 63. Jendrossek, D., Polyhydroxyalkanoate Granules Are Complex Subcellular Organelles (Carbonosomes). Journal of Bacteriology 2009, 191, (10), 3195-3202. 64. Jendrossek, D.; Pfeiffer, D., New insights in the formation of polyhydroxyalkanoate granules (carbonosomes) and novel functions of poly(3-hydroxybutyrate). Environmental Microbiology 2014, 16, (8), 2357-2373. 65. Newville, M., IFEFFIT : interactive XAFS analysis and FEFF fitting. Journal of Synchrotron Radiation 2001, 8, (2), 322-324. 66. Majid, M.; Bénazeth, S.; Souleau, C.; Purans, J., XAFS study of interchain and intrachain order in Se1-xTex glasses: Nearest neighbors. Physical Review B 1998, 58, (10), 6104-6114. 67. Yang, S. I.; Lawrence, J. R.; Swerhone, G. D. W.; Pickering, I. J., Biotransformation of selenium and arsenic in multi-species biofilm. Environmental Chemistry 2011, 8, (6), 543-551.

66

CHAPTER 3

Size and temperature-dependent crystallization

of elemental selenium

67

Abstract

To date, the stability of spherical red amorphous Se colloids that are commonly

biogenically produced in many anoxic environments is not well understood. The red amorphous

allotropic form of Se is a thermodynamically metastable phase, stable at month time scales as we

have previously shown. Temperature is one of the key parameters affecting their stability. Here,

we investigated changes in the red amorphous allotropic form of Se0 under atmospheric pressure

as a function of temperature and size of the colloidal nanoparticles. A synthetic red amorphous Se0

previously shown to be indistinguishable from microbial product (Chapter 2) was used as the

starting phase material. Three particle sizes ranging from 500 down to 65 nm with narrow size

distributions were investigated. We find that the temperature at which red amorphous Se

transforms into trigonal (hexagonal) Se varies as a function of the particle size. We modeled the

transition temperature with a 1/R dependence where R is the radius of the nanoparticles and find

our results to be consistent with prior amorphous-solid phase transition studies. These findings

illustrate the importance of particle size on energetics of the amorphous phase to further understand

the stability, reactivity and bioavailability of biogenic Se0 colloids in Se contaminated

environments prone to seasonal changes and arid climates.

Introduction

To date, most studies have suggested that microbial reduction of Se oxyanions produces

red amorphous elemental Se1-5. However the stability of these products, especially in arid climates,

such as in California central valley where Se contamination is most prevalent and where

temperatures often exceed 30° (glass phase transition of Se0), is poorly understood. Selenium is a

68

metalloid from the chalcogen family that exhibits a range of unique physical and chemical

properties. It is an excellent photoconductor, used in rectifiers, photocopiers, solar cells,

photodetection and sensor applications and is also potentially a key tracer of planetary processes6.

In addition to various amorphous/vitreous forms, five crystalline polymorphs exist under standard

pressure and temperature conditions (300K, 1 atm): trigonal (grey, hexagonal), α-, β- and γ-

monoclinic Se (Se8 rings), and rhombohedral Se (Se6 chains). Amorphous selenium is known to

exhibit a glass-transition temperature at around 31 °C7 and a phase transition to a hexagonal

polymorph in the range ~70-80 °C for bulk Se7. At normal pressure, trigonal (hexagonal) Se is a

p-type semiconductor and consists of infinite helical chains of Se atoms along the c axis8, whereas

photosensitive semiconductor amorphous Se mostly contains packed coiled disordered chains9, 10.

Prior studies of structural and phase transformations of elemental Se have mostly relied on IR-,

Raman- and UV-photoelectron spectroscopy as well calorimetry measurements and diffraction.

Here we investigate the transition of red amorphous Se into hexagonal Se in colloidal particles

(499 down to 65 nm) as a function of temperature at constant pressure (atmospheric) and identical

reaction times using in situ Se K-edge X-ray absorption near-edge structure (XANES)

spectroscopy and X-ray diffraction. The particle diameters investigated here reflect those reported

for most microbially produced red amorphous Se and are therefore environmentally relevant.

The rate of transformation of an amorphous to crystalline material generally depends on

four main parameters: the structure and thus energetics of the reactant and product phases, time,

pressure and the temperature. In this study, the reaction time (10 minutes) and pressure (1

atmosphere) were kept constant and the particle size and temperature varied. In order to properly

evaluate the solid-solid phase transition in elemental Se as a function of temperature, care was

69

taken to produce synthetic particles at room temperature with a narrow size distribution and

controlled shape. We find that the temperature range during which the transition from red

amorphous Se to the thermodynamically stable hexagonal phase occurs, varies as a function of the

particle size. The experimental transition temperature is interpreted primarily in terms of a

thermodynamic effect. Kinetic factors cannot be ruled out however are inferred to be minor in

view of our experimental conditions and observations.

Materials and Methods

Synthesis of Se colloidal nanoparticles

Red amorphous selenium particles (499 nm and 193 nm in diameter) were synthesized at

room temperature, following a protocol modified from Jeong et al.11. Stocks of selenous acid (70

mM) and hydrazine (0.35 M) were prepared in ethylene glycol (anhydrous, 99.8%, Sigma-Aldrich

CAS# 107-21-1). Prior to experiments, all glassware was cleaned by autoclaving and rinsing with

autoclaved MilliQ water. 20 mL ethylene glycol was put in an open flask kept in the dark (wrapped

in aluminum foil) and stirred for ~ 5 minutes on a stirring plate, then 5 mL of hydrazine was added

(Hydrazine hydrate, reagent grade, N2H4 50-60 %, Sigma-Aldrich CAS# 10217-52-4). That

mixture was stirred for ~ 10 minutes. Variable amount of selenous acid (99.999% trace metals

basis, Aldrich CAS# 7783-00-8) was finally added (see Table 1). The reduction of selenous acid

with hydrazine follows the reaction:

H2SeO3(EG) + N2H4(EG) → Se (↓) + N2 (↑) + 3H2O

70

After reacting for 1.5 h, the samples were rinsed three times in MQ and spun at 15,000 rpm

at 4 °C for 3 min. Samples were subsequently stored in glass vials in a refrigerator (5 °C). This

synthesis did not allow production of particles of size smaller than 193 nm. Smaller particles (65

nm) were produced using another recipe also performed at room temperature and modified from

Lin et al.12. A stock of 1.05 M sodium thiosulfate pentahydrate in ethylene glycol and a stock of

0.07 M selenous acid in ethylene glycol were prepared. 1 mL selenous acid was added to 1 mL

thiosulfate then poly(sodium 4-styrenesulfonate) (PSS, Sigma-Aldrich CAS# 25704-18-1) was

added. After reacting for 1 h, the samples were rinsed three times in MQ and spun at 15,000 rpm

at 4 °C for 3 min. Samples were subsequently stored in glass vials in the fridge. Syntheses were

generally performed right prior to synchrotron and SEM analyses. For all samples, one microliter

of solution was deposited onto a 3 mm diameter Si3N4 window (TEM windows, SiMPore Inc.) and

air-dried.

SEM/ EDS

Scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS) data were

collected at 5 keV and 15 keV using a FEI/Philips XL 30 FEG-SEM, equipped with an EDX

spectrometer (EDAX, Inc). The specimens were tilted 4° toward the X-ray detector to optimize

the X-ray detection geometry. The working distance varied between 7.5 and 10 mm. Collection

time for EDS was 500 seconds for each area. The lateral resolution of the microscope was 10 nm

at 15 kV. Se0 was conductive enough so that coating of the samples prior to examination was not

required. Thus, there is no overestimation of particle size due to a coating. The particle size

distribution was estimated by manually measuring > 100 particles in SEM images using the Fiji

software, and then further analyzed with the Origin software.

71

X-ray microprobe. All measurements made in dark conditions. A 1 µL droplet of sample

solution was deposited onto Si3N4 windows (50 nm thick membrane, TEM windows, SiMPore

Inc.) and air-dried. Micro-focused X-ray diffraction (µXRD) and Se K-edge X-ray absorption

spectroscopy data were collected at the Advanced Light Source (ALS) bending magnet beamline

10.3.2 (2.4- 17 keV) with the storage ring operating at 500 mA and 1.9 GeV13. All data were

collected using the custom microprobe Instec stage described in Chapter 1 and Appendix I.

µXRF spectra were collected at 13 keV with a beam spot size of 12×5 μm, and counting times of

30 sec. Fluorescence emission counts were recorded with a seven-element Ge solid-state detector

(Canberra) and XIA electronics. Se K-edge µXANES spectra were recorded in fluorescence mode

by continuously scanning the Si (111) monochromator (Quick XAS mode) from 160 eV below up

to 407 eV above the edge (12500-13067 eV, i.e., up to k = 10 Å-1. All data were collected under

the same conditions: one 10 mn long QXAS scan was collected every 5 °C, from room temperature

(25 °C, inside the hutch) up to 100 °C, using a ramp of 1 °C/ minute. Spectra were calibrated using

the white line of a red amorphous Se standard set at 12660 eV. All data were processed using

LabVIEW custom software and standard procedures described elsewhere14.

Micro-diffraction patterns were collected in transmission mode with a CCD camera

(Bruker SMART6000) at 17 keV (λ = 0.729 Å) using a beam spot size of 12×7 μm and exposure

times of 240 sec. A background pattern was also recorded nearby the region of interest. Diffraction

data were also recorded on the fly as XANES data were being collected so as to follow the

transition more rapidly. Calibration of the camera distance was obtained using an alumina (α-

Al2O3) powder standard and Fit2D software15. Fit2D was also used to obtain one-dimensional

XRD profiles from the radial integration of 2D patterns. XRD peaks were indexed using Jade 9

72

software (Materials Data Inc.) and the ICDD PDF-4+ database. Additional standards from

Mincryst for red monoclinic Se0, grey trigonal-Se0 were added to the database.

Results and Discussion

The SEM imaging revealed colloids with a relatively mono-dispersed size, for the three

particle sizes investigated, 499 nm, 193 nm and 65 nm (see Figure 1 and Table 1). Energy

dispersive spectroscopy (EDS), performed on all samples, showed only peaks for Se and no

contamination. The samples were heated from 25 °C up to 100 °C, in 5 °C steps, at a constant rate

of 1 °C/min (Figures 2-4 and S1). Once a preset temperature had been reached, the sample was

kept at that temperature for 10 minutes, during which a Se K-edge XANES scan was acquired.

As the temperature increases, a post K-edge shoulder feature (indicated by a red vertical

dashed-line) can be observed to increase in the X-ray absorption near-edge structure (XANES)

spectra as crystallization occurs. This result is consistent with a prior study of amorphous Se at

high temperature (McLeod 2009). Each XANES spectrum from the temperature series was fitted

in a first approximation to the two end-members of the series, i.e. using spectra collected at room

temperature (RT) and 100 °C. From these least-square linear combination fits, proportions of

elemental red amorphous Se and hexagonal Se were extracted. Figure 5 shows the percentage of

hexagonal Se formed as a function of temperature for the three particle sizes. We arbitrarily define

an initiation temperature (Tc) as the temperature at which 50% of the original red amorphous Se

was converted into hexagonal Se. A plot of this initiation temperature (Tc) as a function of particle

size (Figure 6) shows that Tc decreases as the particle size decreases. This trend was confirmed

73

by measuring an extra set of data on 61± 22 nm particles (data not shown) giving a similar Tc

value to that of the 65 nm particles.

Our measurements cannot ascertain unambiguously if the observed transformation is

kinetically or thermodynamically driven. In fact, the method used here does not allow

determination of the rate of reaction nor the order of the reaction. In general, to do so, one would

need to determine the percentage of nanoparticles transformed after a certain time. If this

percentage is a linear function of time, the reaction is of the first order and in this case, the rate of

reaction is simply the slope of the percentage transformed per unit time. Solid-solid phase

transitions with a first order reaction rate have been previously observed in semiconductors such

as in 4 to 10 nm CdS nanocrystals and are believed to exemplify the transformation pathways for

solid-solid phases changes between crystalline nanophases16. Polycrystalline or disordered bulk

systems exhibit in general a more complex behavior (not a first order reaction) because the

transformation nucleates randomly at surface or defect sites in the system (grain boundaries for

instance). A similar complex behavior has been reported in the case of transformation involving

polymorphic nanophases. As an illustration, a prior study has focused on the transformation

kinetics of 4 to 6 nm polymorphic TiO2 between 465 °C and 525 °C from anatase to rutile phases.

Conventional kinetics analysis failed reportedly due to the non-inclusion of particle size effects17.

However, grain coarsening and growth of the particles from 4-6 nm up to >25 nm occurred during

these and other similar kinetic experiments18. To date, the question whether first order rate

reactions occur in transformations involving amorphous phases in nanoparticles with a stable size

has not been answered.

74

Prior to the set of experiments we tested on one sample whether the XANES spectrum

would change if the scan was 10 min-long versus 20 min-long and found no significant changes

of the spectral shape. From this observation we concluded that equilibrium was reached within 10

minutes and thus we used 10 min-long scans for all data series. We therefore infer that, as a first

approximation, the observed transitions during our experiments are primarily driven by

thermodynamics. Considering this, we can qualitatively understand our data based on a

thermodynamic argument and an analogy to melting. Traditionally, phase transitions and more

generally phase diagrams of the material have to be considered in the pressure-temperature-particle

size space assuming equilibrium between the different phases. Buffat and Borel19 used this

approach almost forty years ago to predict and observe the melting temperature of gold

nanoparticles as a function of size. Starting with general thermodynamic equilibrium conditions,

they deduce a simple scaling law for the melting temperature that only involve measurable

parameters such a particle diameter, latent heat of fusion, densities and surface energies. In

particular, they showed that the melting temperature has an inverse dependence on the particle size

that correlates with the surface to volume ratio, T(R) = T∞ – A/R where R is the particle size. Since

then, a similar size dependence trend of the melting temperature has been observed for

semiconductor CdSe20 and metallic Bi21 nanoparticles.

A similar scaling law between critical (transition) temperature and particle size (T ~ 1/R)

has been experimentally observed for Cu2S nanosolids and theoretically explained based on a

thermodynamic argument22. The irreversible solid-solid phase transition from amorphous

(metastable) Se to hexagonal (stable) Se is more complex than melting and solid-solid transitions

discussed above because of the absence of any phase equilibrium for the metastable state, i.e it is

75

not possible to define a transition temperature based on thermodynamics. The discussions on

melting point above can apply strictly to reversible solid-solid phase transitions at the equilibrium

state, but not for irreversible transitions. Similarly, it is also not possible to define any phase

diagram for irreversible phase transitions.

Despite these difficulties, Satoh and coworkers23 have recently argued that based on

similarities between melting and the initiation of a transition process between a metastable phase

to a stable phase, one can hypothesize that the transition initiation temperature depends on the

particle size in a similar manner to melting, i.e.,

𝑇𝑇(𝑅𝑅) = 𝑇𝑇𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 −𝐴𝐴𝑅𝑅, (1)

where R the is radius of the nanocrystals, T(R) is the size-dependent transition initiation

temperature, Tbulk is the transition initiation temperature in bulk, and A is a constant. The transition

initiation temperature for irreversible solid-solid phase transitions is predicted to decrease

inversely with the size. This trend has been indeed been observed for the solid-solid transition

from amorphous anatase to rutile TiO2 with sizes from 1.6 to 2.5 nm, a very limited size range.

The scaling law predicted by equation (1) appears consistent with our data (inset of Figure

6) over a range of sizes that extends from 65 to ~ 500 nm (slope A = 398.3 ± 19.5 °C. nm). The

amorphous to hexagonal transition in Se nanoparticles appears to be a unique model system that

may test the extent of the validity of the scaling law over an unprecedented broad size range,

despite the fact that more data points are needed, especially towards smaller sizes and that potential

coarsening effects need to be determined. The theoretical justification of why this transformation

follows the simple scaling law is unclear. However, it is remarkable that the same scaling law

76

seems to apply for melting, solid-solid phase transitions and amorphous-to-solid phase transitions.

It may be an indication that the amorphous-to-crystalline phase transition in Se nanoparticles

shares some common physics with the other type of phase transitions and may be described in a

theoretical framework proper to thermodynamics equilibrium, even though the thermodynamic

state of an amorphous system is not properly defined.

Conclusion

This study provided the first evidence for the role of the size of amorphous particles on the

stability of an important mineral product of bioreduction. Specifically, a dependence of the

apparent phase transition temperature, Tc, for red amorphous to hexagonal Se was documented.

This effect is attributed primarily to the influence of particle size on the structure and energetics

of the amorphous phase. However, transformation kinetics cannot be ruled out because the reactant

is metastable. Further investigations using smaller nanoparticles will be required to confirm the

trends observed here and refine a theoretical model for this solid-solid transition with a starting

amorphous phase. In particular, the amorphous to hexagonal Se phase transition may constitute a

novel model system where data such as the one obtained here could be compared to the predictions

of newly developed models.

Acknowledgments

77

The U.S. Department of Energy (DOE), Office of Science, Office of Biological and

Environmental Research funded the work under contracts DE-SC0004733 (University of

California) and DE-AC02-05CH11231 (LBL). Part of the equipment used for this study was

funded by the Center for Nanoscale Control of Geologic CO2, an LBL Energy Frontier Research

Center. S.F. thanks Sue Spaulding for lab support and Mary K. Gilles (LBL) for use of her SEM.

Work at the Advanced Light Source is supported by the Office of Basic Energy Sciences, Office

of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.

78

Figure 1: Average particle size for the three red amorphous Se syntheses. “Si” refers to the

synthesis number.

79

Figure 2: A, B) SEM images of the 499 ± 19.3 nm particles. C) Size distribution obtained from

SEM measurements D) Change of crystallinity evidenced in Se K-edge XANES spectra (see

vertical red dash line) as a function of temperatures from 30 °C up to 100 °C by step of 5°C.

80

Figure 3. A, B) SEM images of the 193 ± 23.2 nm particles. C) Size distribution obtained from

SEM measurements D) Change of crystallinity evidenced (see vertical red dash line) in Se K-edge

XANES spectra as a function of temperatures from room temperature 25 up to 100 °C, by step of

5°C.

81

Figure 4. A, B) SEM images of the 65 ± 12 nm particles. C) Size distribution from SEM

measurements D) Change of crystallinity evidenced (see vertical red dash line) in Se K-edge

XANES spectra as a function of temperatures from room temperature 25 °C up to 100 °C, by step

of 5°C.

82

Figure 5. Percentage of hexagonal Se formed as a function of temperature, at a constant time rate,

derived from LSQ linear combination fitting of the XANES spectra. A critical temperature (Tc)

was defined as the temperature where 50% (extrapolated) of the original red amorphous Se got

converted into hexagonal Se.

83

Figure 6. Plot of the critical temperature (Tc), defined as the temperature where 50%

(extrapolated) of the original red amorphous Se has been converted into hexagonal Se, as a function

of particle size. The Tc values were extracted from Figure 5. Tc is observed to decrease as the

particle size decreases, a trend line (dashed) is shown based on prior studies as discussed in the

text. The inset shows Tc plotted as a function of 1/R where R is the radius of the particles. A line

fit is shown with slope A = 398.3 ± 19.5 °C.nm and an R-squared value of 0.9976.

84

Figure S1. Micro- X-ray diffraction patterns recorded at 17 keV on the 193 nm red amorphous Se sample at (B) room temperature 25 °C, (C) 100 °C and (D) 150 °C. A background pattern (no sample) is shown in (A) for comparison.

85

Table 1. Summary of synthesis conditions. “Si” refers to the synthesis number. S and H refer to selenous acid and hydrazine respectively.

Molar ratio

S/H

H/S added (mL) Color start/end Size (nm)

S1 0.1 25/2.5 Light red/red 499 ± 19.3

S2 6.66 0.24/1.6 Yellow/dark orange 193 ± 23.2

S3 N.A. N.A. Yellow/yellow 65 ± 12

Table 2. Particle mean size, standard deviation and variation coefficients. “Si” refers to the

synthesis number.

Sample Mean Std Dev CV

S1 503.5 19.3 3.8

S2 193 23.2 12.1

S3 64.5 7.8 12.1

86

References

1. Debieux, C. M.; Dridge, E. J.; Mueller, C. M.; Splatt, P.; Paszkiewicz, K.; Knight, I.; Florance, H.; Love, J.; Titball, R. W.; Lewis, R. J.; Richardson, D. J.; Butler, C. S., A bacterial process for selenium nanosphere assembly. Proceedings of the National Academy of Sciences 2011, 108, (33), 13480-13485. 2. Hunter, W.; Manter, D., Reduction of Selenite to Elemental Red Selenium by Pseudomonas sp. Strain CA5. Curr Microbiol 2009, 58, (5), 493-498. 3. Kessi, J.; Ramuz, M.; Wehrli, E.; Spycher, M.; Bachofen, R., Reduction of Selenite and Detoxification of Elemental Selenium by the Phototrophic BacteriumRhodospirillum rubrum. Applied and Environmental Microbiology 1999, 65, (11), 4734-4740. 4. Tomei, F. A.; Barton, L. L.; Lemanski, C. L.; Zocco, T. G., Reduction of selenate and selenite to elemental selenium by Wolinella succinogenes. Canadian Journal of Microbiology 1992, 38, (12), 1328-1333. 5. Tomei, F.; Barton, L.; Lemanski, C.; Zocco, T.; Fink, N.; Sillerud, L., Transformation of selenate and selenite to elemental selenium byDesulfovibrio desulfuricans. Journal of Industrial Microbiology 1995, 14, (3-4), 329-336. 6. König, S.; Luguet, A.; Lorand, J.-P.; Wombacher, F.; Lissner, M., Selenium and tellurium systematics of the Earth’s mantle from high precision analyses of ultra-depleted orogenic peridotites. Geochimica et Cosmochimica Acta 2012, 86, (0), 354-366. 7. Minaev, V. S.; Timoshenkov, S. P.; Kalugin, V. V., Structural and phase transformations in condensed selenium. Journal of Optoelectronics and Advanced Materials 2005, 7, (4), 1717-1741. 8. Cherin, P.; Unger, P., The crystal structure of trigonal selenium. Inorganic Chemistry 1967, 6, (8), 1589-1591. 9. Corb, B. W.; Wei, W. D.; Averbach, B. L., Atomic models of amorphous selenium. Journal of Non-Crystalline Solids 1982, 53, (1–2), 29-42. 10. Jóvári, P.; Delaplane, R. G.; Pusztai, L., Structural models of amorphous selenium. Physical Review B 2003, 67, (17), 172201. 11. Jeong, U.; Xia, Y., Synthesis and Crystallization of Monodisperse Spherical Colloids of Amorphous Selenium. Advanced Materials 2005, 17, (1), 102-106. 12. Lin, Z.-H.; Chris Wang, C. R., Evidence on the size-dependent absorption spectral evolution of selenium nanoparticles. Materials Chemistry and Physics 2005, 92, (2–3), 591-594. 13. Marcus, M. A.; MacDowell, A. A.; Celestre, R.; Manceau, A.; Miller, T.; Padmore, H. A.; Sublett, R. E., Beamline 10.3.2 at ALS: a hard X-ray microprobe for environmental and materials sciences. Journal of Synchrotron Radiation 2004, 11, (3), 239-247. 14. Kelly, S. D., Hesterberg, D., & Ravel, B, Analysis of Soils and Minerals Using X-ray Absorption Spectroscopy. In Mineralogical Methods, Drees, A. L. U. L. R., Ed. Soil Science Society of America: Madison, WI, 2008; Vol. Part 5, p 367. 15. Hammersley, A. P.; Svensson, S. O.; Hanfland, M.; Fitch, A. N.; Hausermann, D., Two-dimensional detector software: From real detector to idealised image or two-theta scan. High Pressure Research 1996, 14, (4-6), 235-248. 16. Jacobs, K.; Zaziski, D.; Scher, E. C.; Herhold, A. B.; Paul Alivisatos, A., Activation Volumes for Solid-Solid Transformations in Nanocrystals. Science 2001, 293, (5536), 1803-1806. 17. Gribb, A. A.; Banfield, J. F., Particle size effects on transformation kinetics and phase stability in nanocrystalline TiO2. American Mineralogist 1997, 82, (7), 717-728. 18. Zhang, H.; Banfield, J. F., Kinetics of Crystallization and Crystal Growth of Nanocrystalline Anatase in Nanometer-Sized Amorphous Titania. Chemistry of Materials 2002, 14, (10), 4145-4154. 19. Buffat, P.; Borel, J. P., Size effect on the melting temperature of gold particles. Physical Review A 1976, 13, (6), 2287-2298.

87

20. Goldstein, A. N.; Echer, C. M.; Alivisatos, A. P., Melting in Semiconductor Nanocrystals. Science 1992, 256, (5062), 1425-1427. 21. Olson, E. A.; Efremov, M. Y.; Zhang, M.; Zhang, Z.; Allen, L. H., Size-dependent melting of Bi nanoparticles. Journal of Applied Physics 2005, 97, (3), 034304. 22. Li, Y.; Qi, W.; Li, Y.; Janssens, E.; Huang, B., Modeling the Size-Dependent Solid–Solid Phase Transition Temperature of Cu2S Nanosolids. The Journal of Physical Chemistry C 2012, 116, (17), 9800-9804. 23. Satoh, N.; Nakashima, T.; Yamamoto, K., Metastability of anatase: size dependent and irreversible anatase-rutile phase transition in atomic-level precise titania. Sci. Rep. 2013, 3.

88

CHAPTER 4

Iron speciation analysis indicates the use of clays and iron

oxyhydroxides by planktonic- and biofilm-associated

Fe-reducing bacteria

89

Abstract

Iron-reducing bacteria (FeRB) play key roles in the anaerobic carbon cycle. Dissimilatory

Fe(III) reduction represents an important means of mineralization of organic matter when oxygen,

nitrate and sulfate are depleted in the environment. Here we determined the chemical speciation

of iron minerals in planktonic and biofilm samples from an aquifer in Rifle, Colorado, USA that

was amended with acetate to stimulate growth of Fe-reducing bacteria (FeRB). Prior work on

planktonic consortia revealed dominance by Geobacter spp., a well-known group of FeRB. On the

other hand, the biofilm consortium was devoid of Geobacter species and dominated by β-

Proteobacteria, including Fe reducers Ferribacterium, Rhodoferax/Albidiferax sp. For planktonic

and biofilm samples collected near the peak time of Fe reduction, X-ray microprobe measurements

showed that Fe (II, III)-rich micron scale materials are primarily composed of clay minerals and

Fe oxyhydroxides. Scanning transmission X-ray microscopy data of FeRB in the planktonic

consortium indicate that the cell surfaces are associated with Fe (II, III)-rich aggregates of

nanoparticles. These results suggest that iron reducers from planktonic and biofilm consortia use

Fe (III) originating from both Fe oxyhydroxides and phyllosilicates for growth.

Introduction

Considerable research has been conducted to understand the functioning of indigenous microbial

communities and their relevance for biogeochemical cycles and bioremediation in aquifer

settings1-6. Microbial transformations of Fe-bearing minerals play important roles in the iron cycle

and affect Fe mobility and distribution in the environment7, 8. The electrons that Fe(III)-respiring

bacteria (FeRB) pass to ferric iron-bearing compounds can derive from oxidation of sediment-

associated organic matter9. Thus, FeRB play important roles in the carbon cycle. In fact, in some

90

anaerobic settings, the respiration of Fe(III) can be a more important mechanism of mineralization

of a wide range of carbon compounds than the respiration using sulfate10. Strictly anaerobic and

facultative Fe-reducing bacteria can also use H2 as the electron donor for growth, thus these

organisms also impact the hydrogen cycle11.

One approach used to investigate FeRB growth in the environment is to stimulate

subsurface microorganisms by acetate addition1. Geobacter spp.12 have been typically reported as

the dominant planktonic FeRB near the peak of iron reduction in field experiments in acetate-

amended groundwater1, 13. On the other hand, Rhodoferax species have been shown to compete

with Geobacter species during the Fe reduction phase following acetate amendment14.

Organisms in circumneutral pH environments that grow using ferric iron as the electron

acceptor face the challenge of delivery of electrons to a solid product. One option available to the

cells is to adhere directly to mineral surfaces. For example, Geobacter species are known to

directly contact Fe(III) oxide surfaces during growth by Fe(III) reduction15. However, this strategy

is not possible for planktonic cells. Recently, we showed that cells can accumulate aggregates of

ferric iron oxyhydroxide nanoparticles onto their surfaces to enable growth in groundwater and

laboratory cultures (Luef, Fakra et al.16). Because nanoparticles have large reactive surface areas,

they sorb a variety of toxic metals, including V, As, U. Reduction of iron in these aggregates will

liberate surface-associated ions into groundwater. Thus, in addition to influencing the carbon and

hydrogen cycles, FeRB can significantly impact many other trace element and nutrient cycles.

To date, very few studies have examined in detail the Fe(III)-rich minerals that are

available for use by planktonic and biofilm-associated FeRB in natural systems. Such

investigations are required to understand the reactions that support growth of these organisms and

determine the consequences for biogeochemical cycles and bioremediation. Bulk-scale chemical 91

measurements are not sensitive enough to fine particulate materials that are key to microbial

metabolism. Thus, here we used scanning transmission X-ray microscopy and X-ray microprobe

measurements to identify and determine the distribution and valence state of iron in minerals

associated with the surfaces of FeRB cells and present in groundwater and biofilms.

Materials and Methods

Bio-stimulation experiment at the Old Rifle site

A 25-day acetate amendment field experiment was conducted to stimulate Fe(III)-reducing

microorganisms in an aquifer adjacent to the Colorado River near Rifle, Colorado, USA, starting

August 23rd 2010. Acetate-amended (50 mM, Sigma Aldrich, St Louis, MO, USA) groundwater

was injected into the subsurface at rates designed to achieve an aquifer concentration of 5 mM.

Description of the biostimulation experiment can be found in Chapter 1 and references therein.

Groundwater from well CD-01 was sampled 8 days after acetate injection on August 31st, 2010.

Biofilms were collected 16 days after acetate injection on September 8th, 2010.

Sample preparation

As previously described in Chapter 1, biofilm and planktonic groundwater samples were

collected with a portable cryo-plunger instrument17. Aliquots of 5 µL groundwater were deposited

either onto 200 mesh lacey carbon-coated formvar Cu grids (Ted Pella Inc., Redding, CA, USA)

or TEM nickel finder grids (Maxtaform Finder Grid Style H7, 63 mm pitch, 400 mesh; SPI

Supplies, West Chester, PA, USA). In addition, groundwater from well CD-01 sampled 8 days

after the start of acetate amendment was filtered onto a 0.2-mm hydrophilic polyethersulfone filter

92

(Pall Corporation, Port Washington, NY, USA). The filter was flash frozen in liquid nitrogen and

stored at -80 °C until synchrotron measurements.

X-ray microprobe. All samples were mounted on Si3N4 windows (50 nm thick membrane,

TEM windows, SiMPore Inc.) and all data were collected at room temperature. Micro-focused X-

ray diffraction (µXRD) and Fe K-edge X-ray absorption spectroscopy data were collected at ALS

bending magnet beamline 10.3.2 (2.4 - 17 keV) with the storage ring operating at 500 mA and 1.9

GeV18. µXRF spectra were collected at 10 keV with a beam spot size of 12×5 μm, and counting

times of 30 sec. Fluorescence emission counts were recorded for K, Cl, Ca, Si, S, Ti, Fe, Mn, V,

Ni and Cu (from the support grid) with a seven-element Ge solid-state detector (Canberra) and

XIA electronics. Fe K-edge µXANES spectra were recorded in fluorescence mode by

continuously scanning the Si (111) monochromator (Quick XAS mode) from 100 eV below up to

300 eV above the edge (7011-7415 eV). Fe spectra were calibrated using a Fe foil with the first

derivative set at 7110.75 eV. All data were processed using LabVIEW custom software and

standard procedures described elsewhere19.

Fe valence plots were generated according to a method fully described in Marcus et al. 20

and Westphal et al.21, using an updated iron XAS database containing a variety of clays mineral

standards.

Micro-diffraction patterns were collected in transmission mode with a CCD camera

(Bruker SMART6000) at 17 keV (λ = 0.729 Å) using a beam spot size of 12×7 μm and exposure

time of 240 sec. Calibration of the camera distance was obtained using an alumina (α-Al2O3)

powder standard and Fit2D software22. Fit2D was also used to obtain one-dimensional XRD

93

profiles from the radial integration of 2D patterns. XRD peaks were indexed using Jade 9 software

(Materials Data Inc.) and the ICDD PDF-4+ database.

STXM

STXM analyses were performed on the following samples: (1) Flash-frozen groundwater

samples (8 days after the start of acetate amendment) on cryo-TEM grids; (2) Filtered groundwater

deposited onto a Si3N4 window (Silson, Ltd, UK). (3) Flash-frozen Se-rich biofilms on cryo-TEM

grids. All samples were air-dried, just prior to measurements. STXM measurements were carried

out on beamline 11.0.2 of the Advanced Light Source, Lawrence Berkeley National Laboratory23

. This microscope employs a Fresnel zone plate lens (25nm outer zones) to focus a monochromatic

X-ray beam onto the sample. The sample is raster-scanned in 2D through the fixed beam and

transmitted photons are detected via a phosphor scintillator–photomultiplier assembly. The image

contrast relies on core electron excitation by X-ray absorption24, 25. X-ray images recorded at

energies just below and at the Fe L3 absorption edge were converted into optical density (OD)

images and used to derive elemental maps. Image sequences (‘stacks’) recorded at energies

spanning the Fe L2,3 edges (700–735 eV) were used to obtain near-edge X-ray absorption fine

structure (NEXAFS) spectra from regions of interest. Fe L2,3 edges NEXAFS spectra are sensitive

to Fe oxidation state and local bonding environment26. The relative amplitude of the two absorption

peaks at the Fe L3 edge is roughly indicative of the relative proportions of Fe(II) (at 707.8 eV) and

Fe(III) (at 709.5 eV) present in the region of interest27. Iron NEXAFS spectra were compared with

spectral libraries of reference compounds. Fe(III) standards included 2-line ferrihydrite,

lepidocrocite, goethite, akaganeite, FeCl3 and hematite ( -Fe2O3). Fe(II) references included

pyrite, pyrrhotite, FeCl2, biotite, siderite and vivianite. Mixed-valence Fe standard was magnetite

94

(Fe3O4). Fe L2,3 -edges spectra (shape, intensity and energy shift) are sensitive to the oxidation

state. The interpretation is limited by the relevance of our standards and other published spectra.

All measurements were performed at ambient temperature under helium at pressure <1 atm. The

theoretical spatial and spectral resolutions were 30 nm and ±0.1 eV, respectively. Beam-induced

damage was checked and not observed under our experimental conditions. The main Fe L3

resonance of the 2-line ferrihydrite standard set at 709.5 eV was used for relative calibration of the

Fe spectra. All data were processed with the aXis2000 software (AP Hitchcock, an IDL-based

analytical package, http://unicorn. mcmaster.ca).

Results

Bulk geochemistry

Injection of the acetate into well CD-01 resulted in a rapid increase in Fe(II) in the

groundwater (Fig. 1). Sulfate concentrations stayed stable during the experiment, indicating

minimal sulfide production during the sampling period. These trends are consistent with prior

measurements at the Rifle site3. Table 1 summarizes the concentration of the major cations in the

groundwater of well CD-01 at the two time points when biofilm and planktonic samples were

collected.

Valence state of iron in Fe-bearing minerals associated with planktonic cells

For planktonic samples, which were collected slightly prior to the peak of iron reduction, based on

geochemical measurements of aquifer fluids, the valence state of iron from many aggregates

associated with bacteria with various cell morphologies was investigated by STXM at the Fe L2,3

edges. Absorption contrast imaging and elemental mapping at the Fe L3-edge (Fig. 2) reveal many 95

Fe-rich aggregates attached to the surfaces of bacterial cells. Corresponding Fe L2,3 X-ray

absorption near–edge structure (XANES) average spectra of these aggregates (Fig. 3) show they

are composed of mixed-valence Fe(II)–Fe(III). The Fe(II)/Fe(III) ratio varies from one aggregate

to another, but most have a high proportion of Fe(III) and traces of Ca and V16. An example of

cell surface-associated Fe rich aggregates is shown in Fig. 4. The presence of iron oxyhydroxides

is evidenced by the LSQ fit of Fe K-edge spectrum collected in this region. No iron sulfides or

iron carbonates were detected via STXM or X-ray microprobe measurements in all the sample

locations investigated (Fig. 5).

Speciation of iron in particles from groundwater samples

Fe K-edge XANES spectroscopy performed on 26 iron-rich regions of particulate material from

the groundwater samples revealed dominance by mixed Fe(II, III)-rich species (Fig. 5). From the

valence plot, only three analysis points fall into the region defined for Fe(II), based on

measurements of standards. Both valence plots and least-square linear combination fitting (LSQ),

indicate that the vast majority of data points fall close to data points from known Fe oxyhydroxides

and phyllosilicate standards (Table 2). Fig.6 shows such an example of the presence of both

phyllosilicates (illite) and FeOOH (ferrihydrite). LSQ XANES fitting revealed the presence of Fe

oxides and phyllosilicates such as illite, montmorillonite, chlorite and clays. Note that aluminum

and magnesium cannot be detected with the microprobe, as they are below the cutoff of the Si(111)

monochromator. Micro-XRD data confirms the presence of phyllosilicates (not shown) on many

of the spots investigated. No iron sulfides or iron carbonates were detected (Fig.5). Micro-XRF

spectra reveal most iron rich spots contain also calcium and vanadium, as also evidenced by STXM

96

(Fig. 7). The majority of XANES spectra of the Fe and V rich particles fit to a combination of 2-

line ferrihydrite-sorbed vanadate (V(V)) and a V(V) ore standard (see example in Fig. 8).

Speciation of iron in the biofilm-associated particles

Eleven iron-rich spots were investigated by Fe K-edge XANES. From the valence plot, most of

spots fall between the Fe(III) and Fe(II) valence families (Fig. 9). Thus, we conclude that the

particles are mixed valence Fe(II, III) minerals. Micro-XRF spectra (Fig. 10) indicate Fe-bearing

particles containing Ca, Ti/Ba, Si and K, consistent with the presence of phyllosilicates. Further,

LSQ fitting revealed the presence of Fe oxides and silicates (Fig.11). XRD performed on the 11

spots investigated, confirms the presence of phyllosilicates.

Discussion

Nano- and micron-scale measurements provided detailed insight into the minerals present

in groundwater and biofilms, some of which could sustain growth of FeRB. We consistently

detected hydrous Fe oxhydroxides, illite, biotite, chlorite (ripidolite), and minerals from the

smectite family. In both groundwater and biofilm samples, these minerals typically contained

traces of Ti and/or Ba and V. Illite is a hydrous layer silicate that is usually derived from the

weathering of muscovite. Smectite minerals are expandable clays with very high capacity to hold

water and exchangeable cations. The Ti was probably present in the octahedral sheet of biotite that

weathered to smectite, and is now probably present as TiO2 nanoparticles. Ba and Ti have been

previously reported in biotites28, substituting for interlayer K.

97

A prior study29 conducted at the Old Rifle site showed that sediments are dominated by

quartz and feldspars, with lesser amounts of amphiboles and clay. Interestingly, this study found

illite as the dominant phase, with smectite, chlorite and kaolinite also present. Another prior study30

showed the presence of the clinochlore, illite and smectite, as well as Fe oxides such as hematite

and goethite. Moreover, Campbell et al.29 detected several types of Fe(II)/(III)-(oxyhydr)oxides

(crystalline, poorly ordered and Al-substituted) and phyllosilicates in the sediment. We find that

Fe(III) oxides and Fe(III)-rich clays are the dominant Fe species in the groundwater at these two

time points, suggesting that the bacteria may utilize these two sources of Fe(III). Notably, we

determined that the aggregates of iron oxyhydroxides associated with cell surfaces contain both

Fe(II) and Fe(III), but that Fe(III) dominates aggregates collected prior to the peak of Fe(III)

reduction. In contrast, specific aggregates in the biofilm sample, collected at the peak of iron

reduction, contain a mixture of Fe(II) and Fe(III). Overall, solution geochemical data and spectro-

microscopy analyses, connect growth of FeRB with increasing conversion of Fe(II) to Fe(III) over

the course of the Fe-reduction episode.

Conclusion

We conclude that FeRB growing in groundwater that was amended by acetate make use of

both Fe oxyhydroxide nanoparticle aggregates and Fe-rich clay minerals to sustain their growth.

The study provided a detailed insight into the extent of redox transformation of initially Fe(III)-

rich minerals in the early and middle phase of iron reduction in a groundwater system. Detection

98

of predominantly Fe(III)-rich minerals prior to the peak reduction period and mixed valence

materials at the peak supports the conclusion that the materials characterized were important in

sustaining growth of the FeRB.

Acknowledgments

This material is partially based upon work supported through the Lawrence Berkeley

National Laboratory’s Sustainable Systems Scientific Focus Area. The U.S. Department of Energy

(DOE), Office of Science, Office of Biological and Environmental Research funded the work

under contracts DE-SC0004733 (University of California) and DE-AC02-05CH11231 (LBL).

Work at the Advanced Light Source is supported by the Office of Basic Energy Sciences, Office

of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.

99

Figure 1. Concentration of Fe(II) and acetate as a function of time. Vertical grey dash-lines

indicate the start and end of the 25-day acetate injection period. The red vertical line indicates the

time when groundwater samples (planktonic cells) were collected on 08/31/2010, either cryo-

plunged on site on TEM grids or on frozen filters. The blue vertical line indicates the time when

biofilm samples were collected on 09/08/2010.

100

Figure 2. Five examples of paired STXM images at Fe L3 edge (709.5 eV) and Fe distribution

maps for planktonic cells. Fe-rich aggregates are associated with cells of various morphologies,

suggesting that multiple organism types accumulate Fe-bearing particles for growth.

Corresponding spectra are shown in Figure 4. Arrows point to Fe aggregates.

101

Figure 3. Fe L2,3 XANES spectra of Fe particles found on planktonic cells displayed in Figure 3,

compared to Fe (II) standards, siderite and Fe(II) phosphate; and Fe(III) standards, ferrihydrite and

Fe(III) phosphate.

102

Figure 4. Visible light micrograph (VLM) of a bacterium and associated Fe particles in a groundwater sample. Corresponding low dose TEM image showing mineral particles near the cells. Bicolor-coded µXRF distribution map of the area shown in the visible light micrograph, showing iron in red and calcium in green. Corresponding Fe K-edge µXANES spectrum of the area shown in the TEM image. The best LSQ fit matches primarily with a very poorly-ordered iron oxyhydroxide standard whose structure has been described elsewhere31.

103

Figure 5. Iron state valence plot showing data points for the groundwater samples in red and pure

valence standards in black. In blue are the mixed valence standards. In green is the known cell-

associated iron spot shown in Figure 4.

104

Figure 6. Visible light micrograph of particle T7 (and others) from a groundwater sample. Fe K-

edge µXANES spectrum of T7 and best LSQ fit obtained for a mixture of phyllosilicate and iron

oxyhydroxide.

105

Figure 7. STXM measurements on a typical Fe- and Ca-rich particle from a groundwater sample. STXM absorption image of the particle along with iron, calcium and vanadium distribution maps. Fe L2,3 XANES reveals that the particles are mixed Fe(II,III), enriched in Fe(III) and contain traces of vanadium. Intensity scale bars are in optical density units.

106

Figure 8. Microprobe measurements of a groundwater sample showing Fe, V and Ca-rich particles. Typical V K-edge spectrum obtained from a Fe and V rich particle, fit to a combination of 76% 2-line ferrihydrite-sorbed vanadate V(V), and 24% V(V) ore standard.

107

Figure 9. A) Tricolor-coded micro X-ray fluorescence elemental distribution map of the biofilm

sample showing calcium in red, iron in green and potassium in blue. Pixel size is 3 µm. Open

yellow circles show the 11 Fe-rich regions investigated both by Fe K-edge µXANES and µXRD.

B) Iron state valence plot showing the 11 biofilm data points in red and iron standards in black.

108

Figure 10. µXRF spectra of the 11 Fe-rich spots in the biofilm sample, collected 100 eV above the Fe K-edge, showing most of them contain Si, K, Ca and Fe as well as other trace elements.

109

Figure 11. Examples of Fe K-edge XANES fits of Fe-rich regions of biofilm and groundwater samples, showing the presence of iron oxyhydroxides as well as silicates.

110

Table 1. Concentration of the major cations (> 100 µg/L) in the groundwater of well CD-01 on the days of sample collection. The groundwater (planktonic cells) and biofilm samples were collected on 8/31/2010 and 9/8/2010 respectively.

111

Table 2: Fitting results for the groundwater samples

T6 spot1: 60% illite, 40% Montmorillonite.

spotM3: 57% hematite, 43% stx1 clay.

Spot H4: 77% augite, 23% Chlorite (Ripidolite).

Spot T7: 42% illite, 58% ferrihydrite 2L.

Spot G2: 76% augite, 24% biotite.

Spot T5s1: 100% biogenic FeOOH.

112

References

1. Anderson, R. T.; Vrionis, H. A.; Ortiz-Bernad, I.; Resch, C. T.; Long, P. E.; Dayvault, R.; Karp, K.; Marutzky, S.; Metzler, D. R.; Peacock, A.; White, D. C.; Lowe, M.; Lovley, D. R., Stimulating the In Situ Activity of Geobacter Species To Remove Uranium from the Groundwater of a Uranium-Contaminated Aquifer. Applied and Environmental Microbiology 2003, 69, (10), 5884-5891. 2. Bao, C.; Wu, H.; Li, L.; Newcomer, D.; Long, P. E.; Williams, K. H., Uranium Bioreduction Rates across Scales: Biogeochemical Hot Moments and Hot Spots during a Biostimulation Experiment at Rifle, Colorado. Environmental Science & Technology 2014, 48, (17), 10116-10127. 3. Williams, K. H.; Long, P. E.; Davis, J. A.; Wilkins, M. J.; N'Guessan, A. L.; Steefel, C. I.; Yang, L.; Newcomer, D.; Spane, F. A.; Kerkhof, L. J.; McGuinness, L.; Dayvault, R.; Lovley, D. R., Acetate Availability and its Influence on Sustainable Bioremediation of Uranium-Contaminated Groundwater. Geomicrobiology Journal 2011, 28, (5-6), 519-539. 4. Williams, K. H.; Wilkins, M. J.; N'Guessan, A. L.; Arey, B.; Dodova, E.; Dohnalkova, A.; Holmes, D.; Lovley, D. R.; Long, P. E., Field evidence of selenium bioreduction in a uranium-contaminated aquifer. Environmental Microbiology Reports 2013, 5, (3), 444-452. 5. Yelton, A. P.; Williams, K. H.; Fournelle, J.; Wrighton, K. C.; Handley, K. M.; Banfield, J. F., Vanadate and Acetate Biostimulation of Contaminated Sediments Decreases Diversity, Selects for Specific Taxa, and Decreases Aqueous V5+ Concentration. Environmental Science & Technology 2013, 47, (12), 6500-6509. 6. Stucker, V. K.; Williams, K. H.; Robbins, M. J.; Ranville, J. F., Arsenic geochemistry in a biostimulated aquifer: An aqueous speciation study. Environmental Toxicology and Chemistry 2013, 32, (6), 1216-1223. 7. Henry Lutz Ehrlich, D. K. N., Geomicrobiology of Iron. In Geomicrobiology, Fifth Edition, Press, C., Ed. CRC Press: 2008; pp 279-345. 8. Lovley, D. R., Dissimilatory Metal Reduction. Annual Review of Microbiology 1993, 47, (1), 263-290. 9. Lovley, D. R.; Phillips, E. J. P., Novel mode of microbial energy metabolism: organic carbon oxidation coupled to dissimilatory reduction of iron or manganese. Appl Environ Microbiol 1988, 54, 1472-1480. 10. Lovley, D. R.; Phillips, E. J. P., Organic matter mineralization with reduction of ferric iron in anaerobic sediments. Appl Environ Microbiol 1986, 51, 683-689. 11. Ehrlich, H. L.; Newman, D. K., In 2009; p 606. 12. Caccavo, F.; Lonergan, D. J.; Lovley, D. R.; Davis, M.; Stolz, J. F.; McInerney, M. J., Geobacter sulfurreducens sp. nov., a hydrogen-oxidizing and acetate-oxidizing dissimilatory metal-reducing microorganism. Appl Environ Microbiol 1994, 60, 3752-3759. 13. Wilkins, M. J.; VerBerkmoes, N. C.; Williams, K. H.; Callister, S. J.; Mouser, P. J.; Elifantz, H., Proteogenomic monitoring of Geobacter physiology during stimulated Uranium bioremediation. Appl Environ Microbiol 2009, 75, 6591-6599. 14. Zhuang, K.; Izallalen, M.; Mouser, P.; Richter, H.; Risso, C.; Mahadevan, R.; Lovley, D. R., Genome-scale dynamic modeling of the competition between Rhodoferax and Geobacter in anoxic subsurface environments. ISME J 2011, 5, (2), 305-316. 15. Childers, S. E.; Ciufo, S.; Lovley, D. R., Geobacter metallireducens accesses insoluble Fe(III) oxide by chemotaxis. Nature 2002, 416, 767-769. 16. Luef, B.; Fakra, S. C.; Csencsits, R.; Wrighton, K. C.; Williams, K. H.; Wilkins, M. J.; Downing, K. H.; Long, P. E.; Comolli, L. R.; Banfield, J. F., Iron-reducing bacteria accumulate ferric oxyhydroxide nanoparticle aggregates that may support planktonic growth. ISME J 2013, 7, (2), 338-350.

113

17. Comolli, L. R.; Duarte, R.; Baum, D.; Luef, B.; Downing, K. H.; Larson, D. M.; Csencsits, R.; Banfield, J. F., A portable cryo-plunger for on-site intact cryogenic microscopy sample preparation in natural environments. Microscopy Research and Technique 2012, 75, (6), 829-836. 18. Marcus, M. A.; MacDowell, A. A.; Celestre, R.; Manceau, A.; Miller, T.; Padmore, H. A.; Sublett, R. E., Beamline 10.3.2 at ALS: a hard X-ray microprobe for environmental and materials sciences. Journal of Synchrotron Radiation 2004, 11, (3), 239-247. 19. Kelly, S. D., Hesterberg, D., & Ravel, B, Analysis of Soils and Minerals Using X-ray Absorption Spectroscopy. In Mineralogical Methods, Drees, A. L. U. L. R., Ed. Soil Science Society of America: Madison, WI, 2008; Vol. Part 5, p 367. 20. Marcus, M. A.; Westphal, A. J.; Fakra, S. C., Classification of Fe-bearing species from K-edge XANES data using two-parameter correlation plots. J Synchrotron Radiat 2008, 15, (Pt 5), 463-8. 21. Westphal, A. J.; Fakra, S. C.; Gainsforth, Z.; Marcus, M. A.; Ogliore, R. C.; Butterworth, A. L., Mixing Fraction of Inner Solar System Material in Comet 81P/Wild2. The Astrophysical Journal 2009, 694, (1), 18. 22. Hammersley, A. P.; Svensson, S. O.; Hanfland, M.; Fitch, A. N.; Hausermann, D., Two-dimensional detector software: From real detector to idealised image or two-theta scan. High Pressure Research 1996, 14, (4-6), 235-248. 23. Kilcoyne, A. L.; Tyliszczak, T.; Steele, W. F.; Fakra, S.; Hitchcock, P.; Franck, K.; Anderson, E.; Harteneck, B.; Rightor, E. G.; Mitchell, G. E.; Hitchcock, A. P.; Yang, L.; Warwick, T.; Ade, H., Interferometer-controlled scanning transmission X-ray microscopes at the Advanced Light Source. J Synchrotron Radiat 2003, 10, (Pt 2), 125-36. 24. Stöhr, J., NEXAFS Spectroscopy. 1992, 25. 25. Kirz, J.; Jacobsen, C.; Howells, M., Soft X-ray microscopes and their biological applications. Quarterly reviews of biophysics 1995, 28, (1), 33-130. 26. de Groot, F. M. F., X-ray absorption and dichroism of transition metals and their compounds. J Electron Spectroscopy 1994, 67, 529-622. 27. van Aken, P. A.; Liebscher, B., Quantification of ferrous/ ferric ratios in minerals: new evaluation schemes of Fe L2,3 electron energy-loss near-edge spectra. Phys Chem Miner 2002, 29, 188-200. 28. Shaw, C. S. J.; Penczak, R. S., Barium and titanium-rich biotite and phlogopite from the western and eastern gabbro, Coldwell alkaline complex, northwestern Ontario. The Canadian Mineralogist 1996, 34, (5), 967-975. 29. Campbell, K. M.; Kukkadapu, R. K.; Qafoku, N. P.; Peacock, A. D.; Lesher, E.; Williams, K. H.; Bargar, J. R.; Wilkins, M. J.; Figueroa, L.; Ranville, J.; Davis, J. A.; Long, P. E., Geochemical, mineralogical and microbiological characteristics of sediment from a naturally reduced zone in a uranium-contaminated aquifer. Applied Geochemistry 2012, 27, (8), 1499-1511. 30. Komlos, J.; Peacock, A.; Kukkadapu, R. K.; Jaffé, P. R., Long-term dynamics of uranium reduction/reoxidation under low sulfate conditions. Geochimica et Cosmochimica Acta 2008, 72, (15), 3603-3615. 31. Toner, B. M.; Santelli, C. M.; Marcus, M. A.; Wirth, R.; Chan, C. S.; McCollom, T.; Bach, W.; Edwards, K. J., Biogenic iron oxyhydroxide formation at mid-ocean ridge hydrothermal vents: Juan de Fuca Ridge. Geochimica et Cosmochimica Acta 2009, 73, (2), 388-403.

114

Appendix I

Microprobe cryogenic apparatus for correlative

spectro-microscopy

To analyze selenium bio-reduction products, I co-designed a cryogenic apparatus

in collaboration with Paul Baker at Instec Inc. I implemented and further modified the

instrument at the X-ray microprobe beamline 10.3.2 of the Advanced Light Source (LBL).

This setup permits the transfer at cryogenic temperature of flash-frozen samples collected

either directly on the field (Rifle site) or in the laboratory (cultivations) and allows

correlated cryogenic electron and X-ray analyses of the same sample region.

1.1 Specifications

The cryostage was custom designed to:

• Fit the geometry of the microprobe beamline 10.3.2. • Accommodate up to three samples mounted on 3 mm diameter Si3N4 windows

and TEM grids. • Accommodate a TEM JEOL cartridge for correlative measurements on the

JEOL 3100 cryo-TEM at the Donner Lab (LBL). • Provide a fast cooling rate and stable temperature below the devitrification

temperature (~120 K). • Be mechanically stable with low vibration (<500 nm amplitude) and light

enough to be scanned in 2 directions (< 4 pounds). • Have appropriate window covers that efficiently transmit incident and emitted

X-rays in the range 2.3-17 keV (such as kapton). • Allow X-ray diffraction and X-ray absorption measurements in transmission

mode.

115

Mechanical stability of the cryogenic setup over hours-long periods was also highly

desired so as to be able record full cryogenic µEXAFS spectra from poorly concentrated

samples requiring many scans (typically ~ 8 scans, ~ 35 mn/scans, thus ~ 5 h).

1.2 Concept drawings

The apparatus consists of four parts: a stage (CLM77K), a sample loading frame (SLF),

a grid holder tongue (GHT) and a temperature controller (mk1000). Below are the concept

drawing of the CLM77K stage. Figure 1 shows a top view of the stage (in grey) mounted

on the optical table, located inside the beamline 10.3.2 hutch. The CLM77K is mounted

onto a rotation stage with bracket connected to an X-Y translation stage. In this

configuration the sample in the CLM77K can be scanned in 2 dimensions, at 45° to the

incident fixed beam, with the fluorescence signal emitted from the sample collected at 90°

to the incident beam (usual configuration at this endstation). Dashed blue lines show the

incoming, transmitted and outgoing X-ray beams.

116

Figure 1. Top view of the CLM77K stage and other components at the X-ray microprobe endstation. Note that the CCD detector used for X-ray diffraction analyses is not shown.

Figure 2. Close-up view of the CLM77K stage and other components at the X-ray microprobe endstation.

117

Figure 3. Concept drawings of the CLM77K stage. A) Side view of the stage mounted vertically, near the Io ionization chamber and the fluorescence detector. The grid holder tongue (GHT) is inserted in the stage. Note the thermocouple embedded in the grid holder/tongue used to monitor the sample temperature. The red line shows the incoming beam exiting the Io chamber, the blue line shows the outgoing beam; red and blue lines cross at the focused sample position. B) CLM77K stage with the inserted GHT designed to accommodate up to three cryo-TEM grids. C) Top view of the stage and Io chamber. The hollow feature is designed to avoid hitting the ion chamber.

1.3 Description

The CLM77K is cooled to -190 °C (83 K) using a pressurized LN2 tank. During

cooling, the sample chamber and all windows are purged with dry N2 gas to prevent water

vapor condensation and frost. The support loading frame (SLF) is used to support the GHT

during sample loading/unloading and is placed in a LN2 bath in a Styrofoam container

118

during sample transfer from a grid box to the SLF to subsequently the stage (see Figures

4 and 5).

Figure 4: View of the sample loading frame (SLF) sitting in a LN2 bath, along with the grid holder tongue designed to host up to 3 cryo-TEM grids. The SLF can accommodate up to 2 grid boxes.

Figure 5: A) View of the grid holder tongue (GHT) accommodating 3 TEM grids. Inset shows a visible light microscope image of the sample mounted on a TEM grid using that GHT, the cross hair shows the X-ray beam position. B-C) show the Cu grid covers used to bolt the grids into place in the SLF.

119

Alternatively, a single cryo-TEM grid (or Si3N4 window), mounted in a JEOL 3100 TEM

cartridge, can be loaded onto the GHT for correlative analyses, as seen in Figure 6.

Figure 6. A) View of the JEOL 3100 TEM cartridge loaded onto the grid holder tongue. The clamp snaps it into place. B) Visible light microscope image of the sample mounted on a Si3N4 window using that GHT. White cross hair shows the X-ray beam position. In either configuration, the spring loaded GHT cover snaps closed over the cryo-TEM grids

to keep them thermally insulated. Once the CLM77K is cooled and stable at -190 ± 0.1 °C,

the GHT containing the cryo-samples, is then quickly inserted and locked into the

CLM77K, where the sample temperature reaches 95 K (see Figure 7). Thermocouples

located in the GHT and CLM77K, and connected to the mk1000 are used to continuously

monitor (every second) the temperature of the sample and the stage respectively. Once the

cryo-microprobe measurements (~12 h maximum duration) are complete, the GHT with

samples is cryo-transferred back to the SLF in an LN2 bath and samples are subsequently

stored in LN2.

120

Figure 7. Temperature profile of the stage and sample. The stage is heated (no sample tongue) to remove moisture, then cooled down (takes about 15 min to reach equilibrium at 95 K). The tongue with the frozen sample is then inserted in the stage. Microprobe measurements lasted about 9 h in this example before transferring the sample back.

121

Appendix II

Supplemental Materials for

Chapter 1

Other Supplementary Material for Chapter 1 includes movie S1.

Materials and Methods

Enrichment cultures. For enrichment cultivation, we sampled pieces of the tubing used to deliver

acetate to the subsurface, in which a biofilm had grown onto the inner surface and preserved these

pieces at -80 °C. Pieces of biofilm were thawed and added anaerobically to fresh medium in a

glove box. Cultures were established either in 10 mL carbonate-buffered freshwater medium

(sulfate and nitrate free) or in groundwater artificial medium (GWA, contains sulfate) and

inoculated with sodium selenate or sodium selenite (5 mM, Sigma Aldrich) and sodium acetate or

sodium L- lactate (10 mM, Sigma Aldrich). Cultures were sparged with N2:CO2 (80:20) to remove

dissolved oxygen and sealed with butyl rubber stoppers and aluminum crimp seals, as previously

described1. All cultures were incubated at room temperature in the dark. Only cultures grown

anaerobically using acetate as the carbon source and electron donor and selenate as the electron

122

acceptor were further analyzed. Samples were flash-frozen in the lab then stored in LN2 until

analyses.

Groundwater geochemical measurements. Groundwater samples were collected at a depth of

5m below ground surface in the monitoring well, filtered through 0.45 μm Polyvinylidine Fluoride

(PVDF) syringe filters (Merck Millipore, Ireland), acidified to a pH<2 with trace metals grade

nitric acid, and analyzed for total selenium concentrations via inductively coupled plasma mass

spectroscopy (ICP-MS; Elan DRCII, Perkin−Elmer, CA). Acetate and bromide were measured

with a Dionex ICS-2100 chromatograph with a Dionex AS-18 column (Thermofisher Scientific,

USA).

Cryo-plunging of samples for synchrotron and TEM analyses. Field-derived biofilm and

enrichment culture samples were flash-frozen directly in the field or in the laboratory, respectively.

Samples were either mounted on TEM Cu-grids (200 mesh, lacey carbon coated formvar, Ted

Pella Inc.) or on Si3N4 windows (TEM windows, SiMPore Inc.). TEM grids and Si3N4 windows

were pre-treated by glow-discharge to improve sample deposition, and were pre-loaded with

colloidal gold particles (10 nm, BBInternational, Cardiff, UK) serving as fiducials. Aliquots of 5

µL of fresh biofilm sample solution were deposited onto the grids. The grids were then manually

blotted with filter paper (Grade 1 filter paper, Whatman®) and plunged into liquid ethane at LN2

temperature, using a portable cryo-plunger. All samples were placed in grid boxes and stored in a

LN2 tank until analysis. Further details on the cryo-plunger instrument and procedures can be

found elsewhere2.

Selenium valence state scatter plot. We used a method similar in concept to that described in

Marcus et al.3 (for Fe species) to evaluate the valence state of Se. Normalized XANES spectra 123

were processed using a new custom Matlab program. Spectra were reduced to two variables, a and

b, defined as the normalized XANES absorption values at 12664.25 eV and 12667.80 eV

respectively. These energies were determined empirically to lead to the best separation between

chemical families in (a,b) scatter plots. Selenate (Se(VI)), selenite Se(IV), elemental Se(0) and

Se(-II) selenide/Se(-I)/Organic Se families are found to separate, as evidenced with Se standards

of known valence, plotted in black (Fig. S9). Using this method, unknown experimental XANES

data could be at first sight classified according to valence state. Least square linear combination

fitting is still required to confirm the identity and valence state of the unknown. All standards (see

Table S1), were recorded at ALS beamline 10.3.2 either at room temperature, with samples

mounted on kapton tape, or at 95 K using a cryo-stage, described in the next section, with samples

mounted on Si3N4 windows (TEM windows, SiMPore Inc.).

Micro-X-ray diffraction. Data were collected in transmission with a CCD camera (Bruker

SMART6000) at 17 keV (λ = 0.729 Å) using a beam spot size of 12×7 μm and exposure times of

240 sec. A background pattern was also recorded nearby the region of interest. Calibration of the

camera distance was obtained using an alumina (α-Al2O3) powder standard and Fit2D software4.

Fit2D was also used to obtain one-dimensional XRD profiles from the radial integration of 2D

patterns. XRD peaks were indexed using Jade 9 software (Materials Data Inc.) and the ICDD PDF-

4+ database. Additional cards from Mincryst for red monoclinic Se0, grey trigonal-Se0 and

hexagonal ice were added to the database.

Confocal laser scanning microscopy (CLSM). CLSM was performed on a cross section of the

tubing with biofilms attached on the inner surface. Samples were fixed fixed in 2.5 %

Paraformaldehyde (final concentration). Confocal images were acquired on a Carl Zeiss Inc. LSM

124

710 Zen 2010 Black (Carl Zeiss MicroImaging Inc., Thornwood, NY, USA), equipped with Argon

488 nm laser. A 63x CPlanApo water immersion objective was used to acquire z-stack images at

1024x1024, 8 bit resolution. The pinhole was set to 1 Airy unit, yielding an optical slice thickness

of 0.8 um. Step size was set to Nyquist sampling, at 0.4 um. Samples were stained with SYTO BC

(Invitrogen product #S34855) and mounted in water in a glass depression slide with a coverslip

measuring 173 um in thickness. All 3-D renderings were performed using Imaris software

(Bitplane AG, Zurich, Switzerland).

16S ribosomal RNA gene sequencing. DNA was extracted from biofilm samples using the

PowerSoil DNA Extraction Kit ((MoBio Laboratories, Inc., CA, USA) according to specifications

of the manufacturer, except that the initial lysing step was accomplished by vortexing at maximum

speed for 2 minutes and then incubating at 65 ˚C for 30 minutes, with an additional minute of

vortexing every 10 minutes. DNA was amplified using 27f and 1492r primers over a temperature

gradient, and PCR products were cleaned up using the MoBio UltraClean PCR Clean-Up Kit. The

amplified DNA was then used to create 16S rRNA gene clone libraries with the pCR 2.1-TOPO

vector according to manufacturer specifications and electrocompetent cells. Generated sequences

from successful insertions were trimmed to remove Phred quality scores ≤20, with forward and

reverse sequences assembled via Phrap. Each sequence was then BLASTed against the SILVA

SSURef108 database.

16S rRNA gene phylogenetic analysis. 16S rRNA gene sequences were aligned using SSU-

Align5 along with the best-hits of these sequences in the SILVA database (version 115 of non-

redundant SILVA)6, a curated set of reference bacterial sequences, and a set of archaea sequences

(Fig. 3B) and a set of Gammaproteobacteria (Fig. 3C and 3D) to serve as a phylogenetic root.

125

Sequences with ≥800 bp aligned were used to infer a maximum-likelihood phylogeny using

RAxML7 with the GTRCAT model of evolution (Fig. S4A) and to construct neighbor-joining

trees8 (Fig. S4B and S4C). Bootstrap analyses were based on 100 re-samplings.

Scanning Transmission X-ray Microscopy (STXM). STXM measurements were performed on

the soft X-ray beamline 11.0.2 (150-2000 eV) of the Advanced Light Source, Lawrence Berkeley

National Lab, Berkeley, CA9. This microscope employs a Fresnel zone plate lens (25 nm outer

zones) to focus a monochromatic beam onto the sample. The sample is scanned in 2 dimensions

through the fixed beam and transmitted photons are detected with a phosphor scintillator-

photomultiplier assembly. Image contrast relies on core electron excitation by X-ray absorption10,

11. X-ray images recorded at energies just below and at the elemental absorption edge of interest

(Se L3 and C K) were converted into optical density (OD) images and used to derive elemental

maps (OD = ln (I0/I), where I0 is the incident X-ray intensity and I is the transmitted beam intensity

through the sample). Chemical maps were obtained by taking the difference of OD images at 280,

288.2 eV (proteins); 289, 290.3 eV (carbonates); and 1425, 1440 eV (Se). Image sequences

(‘stacks’) recorded at energies spanning the Se L2,3-edges (1420-1520 eV) and C K-edge (280-320

eV) were used to obtain XANES spectra from regions of interest. Fitting of the stacks were

performed using the stack fit option in aXis 2000 and relevant standards. Se L2,3-edges XANES

spectra are sensitive to the oxidation state and the local bonding environment of Se. Se spectra

were compared to model compounds (see Table S1, asterix annotation). The main Se L3 resonance

of the amorphous red Se0 standard was set to 1435 eV and used for relative energy calibration of

the spectra. At least two different sample regions were analyzed for each element. The theoretical

spectral and spatial resolutions during measurements were +/-100 meV and 30 nm respectively.

126

The photon energy was calibrated at the C K-edge using the Rydberg transition of gaseous CO2 at

292.74 eV (C 1s→ 3s (ν = 0)). All measurements were performed at ambient temperature under

He at pressure < 1 atm. All data were processed with the IDL-based aXis2000 software package

(http://unicorn.mcmaster.ca/aXis2000.html). This STXM did not have cryogenic capabilities and

flash-frozen samples were thawed and air-dried at room temperature just prior to measurements.

SEM/ EDS. Scanning electron microscopy and energy dispersive spectroscopy data were collected

at 5 keV and 15 keV using an FEI/Philips XL 30 FEG-SEM, equipped with an EDX spectrometer

(EDAX, Inc). The specimens were tilted 4 degrees toward the X-ray detector to optimize the X-

ray detection geometry. Working distance varied between 7.5 and 10 mm. Collection time for EDS

was 500 seconds for each area. The lateral resolution of the microscope was 10 nm at 15 kV.

Supplementary text

X-ray radiation dose rate estimate. We define the dose rate as:

𝐷𝐷𝐷𝐷 = 𝐸𝐸/𝑝𝑝ℎ × 𝑒𝑒 × 𝐹𝐹𝐷𝐷 × 𝜇𝜇/ρ where:

E/ph= energy (eV) per photon with E = 13 keV and 𝑒𝑒 = 1.6 × 10−19 J

𝐹𝐹𝐷𝐷 = 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹 𝑑𝑑𝑒𝑒𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 =𝑓𝑓𝐹𝐹𝐹𝐹𝐹𝐹 𝑓𝑓𝐹𝐹𝐹𝐹𝐹𝐹

𝑓𝑓𝐹𝐹𝐹𝐹𝐹𝐹 𝑑𝑑𝑝𝑝𝑠𝑠𝑑𝑑 𝑑𝑑𝑑𝑑𝑠𝑠𝑒𝑒=

109

15 × 6 × 10−12 = 1.11 1019 𝑝𝑝ℎ 𝑑𝑑−1𝑚𝑚−2

𝜇𝜇 ρ⁄ = 𝑚𝑚𝑚𝑚𝑑𝑑𝑑𝑑 𝑚𝑚𝑑𝑑𝑑𝑑𝑒𝑒𝑑𝑑𝐹𝐹𝑚𝑚𝑑𝑑𝑑𝑑𝑠𝑠𝑑𝑑 𝑐𝑐𝑠𝑠𝑒𝑒𝑓𝑓𝑓𝑓𝑑𝑑𝑐𝑐𝑑𝑑𝑒𝑒𝑑𝑑𝑑𝑑 ~ 3.5 𝑐𝑐𝑚𝑚2 𝑔𝑔−1 𝑓𝑓𝑠𝑠𝑓𝑓 𝐹𝐹𝑑𝑑𝑙𝑙𝐹𝐹𝑑𝑑𝑑𝑑 𝑤𝑤𝑚𝑚𝑑𝑑𝑒𝑒𝑓𝑓 𝑚𝑚𝑑𝑑 13 𝑘𝑘𝑒𝑒𝑘𝑘

Hence, 𝐷𝐷𝐷𝐷 = 8.08 × 103 𝐽𝐽 𝑘𝑘𝑔𝑔−1 𝑑𝑑−1 = 8.08 103 Gy 𝑑𝑑−1

127

Micro-XRF mapping used dwell times of 50 ms/pixel typically, hence an X-ray dose on each pixel

of ~103 Gy. By comparison, Pereiro-López et al.12 have estimated the critical radiation dose

required to image frozen hydrated biological material to be ~109 Gy while Fayard et al.13 have

shown that LN2 cooled blood cells exposed to X-ray doses up to ~ 1010 Gy remained well

preserved, free of easily visible structural changes and mass loss.

128

Figure S1. A) Front (beam) view of the sample cryo-transfer tongue accommodating 3 cryo-TEM

grids (or Si3N4 windows) B) Temperature profiles of the cryo-stage and of the frozen hydrated

sample prior, during and after microprobe measurements, showing the evolution of temperature

during the transfer of the frozen sample and subsequent measurements once the temperature

reaches equilibrium. C) Backside view of the stage, oriented at 45° to incident beam with the

sample on the rotation axis and in focus position (regular configuration). Micro-XRD is performed

in transmission mode and used to check the quality of the cryotransfer.

129

Figure S2. Concentrations of soluble Se, acetate, and sulfate as a function of time in reference

well CD01 during the Super 8 experiment at the Rifle site. Solid vertical lines indicate the start

and end of acetate injection into the subsurface. Biofilm samples were collected 16 days after the

start of acetate amendment (vertical dash line). The decrease in acetate concentration between days

18 and 24 is related to a 6-day gap between two periods of acetate injection. Inset: Location of the

down-gradient monitoring well CD01 and acetate injection well CG02 where the biofilms were

collected. Further details on the well layout and the Super 8 experiment can be found elsewhere14.

130

Figure S3. Relative abundance within the Rifle CG02 biofilm microbial community at the A)

Phyllum- B) Class- C) Family- and D) Genus-levels. The number of sequences affiliated with each

group was divided by the total number of sequences (OTUs).

131

Figure S4A. Maximum-likelihood 16S rRNA gene phylogenic tree. Sequences with ≥800 bp

aligned were used to infer a maximum-likelihood phylogeny using RAxML with the GTRCAT

model of evolution and 100 bootstrap re-samplings. Red stars represent the 16S rRNA sequences

recovered in this study.

132

Figure S4B. Neighbor-joining tree based on 182 16S rRNA gene sequences showing the

phylogenic relationships of the organisms from the Comamonadaceae family sampled in this study

(colored in red). Bootstrap values (percentages of 100 replicates) are given at nodes; only values

>50% are shown. Gammaproteobacteria were used as an outgroup.

133

Figure S4C. Neighbor-joining tree based on 111 16S rRNA gene sequences showing the

phylogenetic relationships of the organisms from the Rhodocyclaceae family sampled in this study

(colored in red). Bootstrap values (percentages of 100 replicates) are given at nodes; only values

>50% are shown. Gammaproteobacteria were used as an outgroup.

134

Figure S5. Cryogenic transmission electron micrographs of representative organisms present in

biofilm CG02 with approximate dimensions A) 2 µm x 0.63 µm, B) 1.1 µm x 0.78 µm, C) 1.6

µm x 0.65 µm, D) 570 nm diameter, E) 570 nm diameter, F) 1.2 µm x 0.75 µm, G) 1.3 µm x 0.8

µm, H) 550 nm diameter. Most cells contain granules, some have a visible polar flagellum (A, B,

E, G), a distinct S-layer (G, D, E, H). Most of them exhibit electron dense particles. Numerous

thin pili-like structures (~ 3-6 nm diameter) are found and pointed by blue arrows. F= flagellum,

OM= outer membrane, IM= inner membrane, S= S-layer. B= bacteriophage. Au= gold fiducial

nanoparticles (10 nm). Scale bars are 200 nm.

135

Figure S6. Scanning electron microscopy and cryogenic microprobe measurements of Rifle CG02

biofilms. A, B) SEM images at 15 keV of CNPs embedded in the biofilm matrix. C) EDS spectrum

at 15 keV of a colloidal nanoparticle shown in A). Cu fluorescence emission peaks were associated

with the Cu grid and the C and O peaks originated from the biofilm matrix and the formvar on the

grid. D) Cryo-XRF Se distribution map of a piece of biofilm recorded at 13 keV with 2 µm pixels.

C) Se K-edge XANES spectrum collected on the yellow circled area. LSQ linear combination

fitting indicates 100% amorphous red Se.

136

Figure S7. Size distribution of CNPs in A) Rifle CG02 biofilms (16-days-old) and B) Bicarbonate

cultures (19-days-old). A total of 199 and 422 particles were respectively measured for A) and B),

from SEM and TEM images. Gaussian fits of the distributions are also shown.

137

Figure S8. A-C) STXM derived C K-edge spectra of a cell (from granule free regions), granules

and extra-cellular material and corresponding tricolor maps obtained from fitting the carbon stack

(i.e. image sequence) pixel by pixel. Proteins (at 288.2 eV) are coded in red, granules

(carboxyls/esters at 288.4 eV) in green, and lacy formvar grid in blue. Standards shown for

comparison are: calcite (CaCO3), Lipopolysaccharides from Escherichia coli EH100 (LPS), 1,2-

Dipalmitoyl-sn-glycero-3-phosphoethanolamine (PE lipid), bovine serum albumin (BSA) protein,

alginate (acidic polysaccharide), agarose (neutral polysaccharide), and DNA standards. All spectra

were normalized at 300 eV. Indicative dash lines are at 288.2 eV (proteins) and 290.3 eV

(carbonates). D) Corresponding nitrogen map shows that the granules are nitrogen-poor.

138

Figure S9. A) Cryo-Se K-edge XANES spectra of a subset of standards (see Table S1). Variables

a and b represent normalized XANES absorption values at 12664.25 and 12667.8 eV and are used

to generate the valence state scatter plot, shown in panel B). A total of 36 standards were used and

measured either at room temperature (open squares) or at 95 K (filled squares) are shown. The

Se(-II, -I) group includes organic Se compounds.

139

Figure S10. Correlated cryogenic measurements on a 19-day-old selenate and acetate enrichment

culture grown in bicarbonate medium A) Size distribution of Se0 CNPs. B) µXRD profile on the

area shown in Figure 5, matched ice and Cu (from the grid) and showed no evidence for crystalline

Se0 . The µXRF spectrum collected at 17keV on that spot showed Se with traces of Ca. The Cu

peak is attributed to the Cu grid and Ar comes from air.

140

Figure S11. STXM measurements at C K-edge and Se L2,3 of a 50-day-old selenate-reducing

enrichment culture grown in GWA medium. A) Absorption image at 280 eV of a cell and

associated Se0 CNPs. B-C) Tricolor-coded maps showing the CNPs coated with proteins.

Distribution maps of Selenium (D), proteins (E) and extracellular (carbonate-rich) polymeric

substances (EPS). G) C K-edge XANES spectrum of the cell and EPS, compared with standards:

sodium bicarbonate (NaHCO3, present in GWA medium), 1,2-Dipalmitoyl-sn-glycero-3-

phosphoethanolamine (PE lipid), bovine serum albumin (BSA) protein, alginate (acidic

polysaccharide), agarose (neutral polysaccharide), and DNA standards. All spectra were

normalized at 300 eV. Indicative vertical dash lines are at 288.2 (proteins), 288.6 (acidic

polysaccharides) and 290.3 eV (carbonates). H) Se L2,3 XANES spectrum of CNPs shown in inset,

indicate Se0. Scale bars (A-F) are 500 nm.

141

Figure S12. In cultures red CNPs Se0 (A) coalesce, (B) aggregate and (C) slowly crystallize over

time.

Movie S1: CLSM of field-derived biofilms. Cells were detected by staining their DNA with SYTO

BC (Invitrogen product #S34855). 2D Z-slice confocal movie showing “haystacks” of cells. The

biofilm ranges in thickness from 2-12 µm. Cells range in length from 0.75 to 2.7 µm. Scale bar is

3µm.

142

Table S1. List of selenium standard compounds used in valence plot and LSQ linear combination fitting analyses.

Compound Formula Temperature/ Source

Crystal system

Organic Seleno-diglutathione C20H32N6O12S2Se RT, Freeman 200615

Seleno-cystathionine C7H14N2O4Se RT, Freeman 200615

γ-glutamyl-methyl-seleno-cysteine C9H16N2O5Se RT, Freeman 200615 Se-Methyl-seleno-cysteine C4H9NO2Se RT, Freeman 200615 Se-ethionine C6H13NO2Se RT, Ryser 200516 Seleno-DL-cystine* C6H12N2O4Se2 98K, SA CAS # 2897-21-4* RT, Freeman 200615 RT, Ryser 200516 Seleno-DL-methionine* C5H11NO2Se 98K, SA CAS # 1464-42-2* RT, Ryser 200516 Se-(Methyl)seleno-cysteine hydrochloride* C4H9NO2Se·HCl 98K, SA CAS # 863394-07-4*

Se(-II)

Berzelianite Cu1.8Se RT, Ryser 200516 Cubic

Copper selenide CuSe RT, Ryser 200516 N.D. Zinc selenide ZnSe RT, Ryser 200516 Cubic

Se (-I) Krutaite CuSe2 RT, Ryser 200516 Cubic

Penroseite (Ni,Co,Cu)Se2 RT, Ryser 200516 Cubic

Se (0)

Red Se from selenate Se RT, Microbial, Strawnγ N.D.

Red Se from selenite Se RT, Microbial, Ryser 200516 Monoclinic

Red Se 193 nm Se RT and 98K, Synthetic Amorphous

Red Se from selenite* Se RT and 98K, microbial* Amorphous

Grey Se* Se RT and 98K, synthetic* Trigonal

Se foil Se RT and 98K, Exafs Materials Trigonal

Selenium shot 2-6 mm, 99.999% Se RT and 98K, AA CAS # 7782-49-2 Amorphous

Selenium black 99+ Se RT, EMD Millipore # 034-001-00-2 Trigonal

Se (IV) Chalcomenite CuSeO3H2O RT, Ryser 200516 Orthorhombic

Mandarinoite Fe3+2Se3O9•6(H2O) RT, Ryser 200516 Monoclinic

Sodium selenite* Na2SeO3 RT and 98K, SA CAS # 10102-18-8* Tetragonal

RT, Freeman 200615

RT, Ryser 200516

Calcium selenite Ca.H2O3Se 98K, P&B CAS # 13780-18-2 Monoclinic

Copper(II) selenite hydrate CuO3Se•xH2O 98K, AA CAS # 10214-40-1 Orthorhombic

143

Zinc selenite ZnSeO3 98K, AA CAS # 13597-46-1 Orthorhombic

Se dioxide SeO2 98K, SA CAS # 7446-08-4 Trigonal

Selenium tetrachloride SeCl4 RT, SA CAS # 10026-03-6 Monoclinic

Se(VI)

Calcium selenate CaSeO4 RT, MP Bio CAS # 14019-91-1 Monoclinic Sodium selenate* Na2SeO4 98K, SA CAS # 13410-01-0* Orthorhombic RT, Freeman 200615 RT, Ryser 200516 Barium selenate BaSeO4 98K, AA CAS # 7787-41-9 Orthorhombic Potassium selenate K2SeO4 98K, AA CAS # 7790-59-2 Orthorhombic Zinc selenate pentahydrate ZnSeO4•5H2O 98K, AA CAS # 13597-54-1 Triclinic Cupric selenate H10O9CuSe RT, MP Bio CAS # 10031-45-5 Triclinic Neodymium selenate O12Se3Nd2 98K, MP Bio 05227729 N.D. Magnesium selenate O4MgSe 98K, MP Bio CAS # 14986-91-5 Monoclinic

N.D.= not determined; RT= room temperature; γunpublished, SA= Sigma-Aldrich, AA= Alfa Aesar, P&B= Pfaltz & Bauer.*Standards also measured on STXM at room temperature.

Table S2. Summary of results of LSQ linear combination fitting of XANES spectra on field-derived biofilms and enrichment cultures. Standards compounds employed are listed in Table S1.

Sample Grey hex. Se Red am. Se Selenite Selenate Org. Se Other

Biofilm CG02 (n=29) ND 96.6 3.4 ND ND ND

Culture GWA (n=23) ND 42.5 35.8 4.3 8.7 8.7

Culture Bicarb (n=18) 5.6 48.5 34.7 5.6 ND 5.6

ND= not detected, n= number of spots measured.

Table S3. EXAFS fitting parameters (see Fig.6).

Sample Shell N S02 σ2 (Å2) ∆E0 (eV) R (Å)

19d bicarb* Se-Se 2.00 1.011 ± 0.067 0.00282 ± 0.0004 -2.446 ± 0.898 2.33897±0.0034

Note: *19-day-old enrichment culture in bicarbonate medium. σ2 = mean-square disorder of neighbor distance, R = distance to neighboring atom, N = coordination number of neighboring atom, ∆E0 = shift in threshold energy E0, S02 = amplitude reduction term.

144

References

1. Coates, J. D.; Phillips, E. J.; Lonergan, D. J.; Jenter, H.; Lovley, D. R., Isolation of Geobacter species from diverse sedimentary environments. Applied and Environmental Microbiology 1996, 62, (5), 1531-6. 2. Comolli, L. R.; Duarte, R.; Baum, D.; Luef, B.; Downing, K. H.; Larson, D. M.; Csencsits, R.; Banfield, J. F., A portable cryo-plunger for on-site intact cryogenic microscopy sample preparation in natural environments. Microscopy Research and Technique 2012, 75, (6), 829-836. 3. Marcus, M. A.; Westphal, A. J.; Fakra, S. C., Classification of Fe-bearing species from K-edge XANES data using two-parameter correlation plots. J Synchrotron Radiat 2008, 15, (Pt 5), 463-8. 4. Hammersley, A. P.; Svensson, S. O.; Hanfland, M.; Fitch, A. N.; Hausermann, D., Two-dimensional detector software: From real detector to idealised image or two-theta scan. High Pressure Research 1996, 14, (4-6), 235-248. 5. Nawrocki, E. P. Structural rna homology search and alignment using covariance models. Washington University, 2009. 6. Quast, C.; Pruesse, E.; Yilmaz, P.; Gerken, J.; Schweer, T.; Yarza, P.; Peplies, J.; Glöckner, F. O., The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Research 2013, 41, (D1), D590-D596. 7. Stamatakis, A., RAxML Version 8: A tool for Phylogenetic Analysis and Post-Analysis of Large Phylogenies. Bioinformatics 2014. 8. Saitou, N.; Nei, M., The neighbor-joining method: a new method for reconstructing phylogenetic trees. Molecular biology and evolution 1987, 4, (4), 406-25. 9. Kilcoyne, A. L.; Tyliszczak, T.; Steele, W. F.; Fakra, S.; Hitchcock, P.; Franck, K.; Anderson, E.; Harteneck, B.; Rightor, E. G.; Mitchell, G. E.; Hitchcock, A. P.; Yang, L.; Warwick, T.; Ade, H., Interferometer-controlled scanning transmission X-ray microscopes at the Advanced Light Source. J Synchrotron Radiat 2003, 10, (Pt 2), 125-36. 10. Stöhr, J., NEXAFS Spectroscopy. 1992, 25. 11. Kirz, J.; Jacobsen, C.; Howells, M., Soft X-ray microscopes and their biological applications. Quarterly reviews of biophysics 1995, 28, (1), 33-130. 12. Pereiro, E.; Nicolas, J.; Ferrer, S.; Howells, M. R., A soft X-ray beamline for transmission X-ray microscopy at ALBA. J Synchrotron Radiat 2009, 16, (Pt 4), 505-12. 13. Fayard, B.; Salomé, M.; Takemoto, K.; Kihara, H.; Susini, J., Some practical considerations about the effects of radiation damage on hydrated cells imaged by X-ray fluorescence microscopy. Journal of Electron Spectroscopy and Related Phenomena 2009, 170, (1–3), 19-24. 14. Bao, C.; Wu, H.; Li, L.; Newcomer, D.; Long, P. E.; Williams, K. H., Uranium Bioreduction Rates across Scales: Biogeochemical Hot Moments and Hot Spots during a Biostimulation Experiment at Rifle, Colorado. Environmental Science & Technology 2014, 48, (17), 10116-10127. 15. Freeman, J. L.; Zhang, L. H.; Marcus, M. A.; Fakra, S.; McGrath, S. P.; Pilon-Smits, E. A. H., Spatial Imaging, Speciation, and Quantification of Selenium in the Hyperaccumulator Plants Astragalus bisulcatus and Stanleya pinnata. Plant Physiology 2006, 142, (1), 124-134. 16. Ryser, A.; Strawn, D.; Marcus, M.; Johnson-Maynard, J.; Gunter, M.; Moller, G., Micro-spectroscopic investigation of selenium-bearing minerals from the Western US Phosphate Resource Area. Geochemical Transactions 2005, 6, (1), 1.

145