Iron Redox Process in Clay Minerals and Its Environmental Significance

MIAMI UNIVERSITY The Graduate School

Certificate for Approving the Dissertation

We hereby approve the Dissertation

Linduo Zhao

Candidate for the Degree

DOCTOR OF PHILOSOPHY

______Hailiang Dong, Director ______John Rakovan, Reader ______Jonathan Levy, Reader ______Christopher Gorski, Reader ______Richard Edelmann, Graduate School Representative ABSTRACT

IRON REDOX PROCESS IN CLAY MINERALS & ITS ENVIRONMENTAL SIGNIFICANCE

by Linduo Zhao

The importance of microbial nitrate-dependent Fe(II) oxidation to iron biogeochemistry is well recognized. Past research has focused on oxidation of aqueous Fe2+ and structural Fe(II) in oxides, carbonates, and phosphate, but the importance of structural Fe(II) in phyllosilicates in this reaction is only recently studied. However, the effect of clay mineralogy on the rate and the mechanism of the reaction, and subsequent mineralogical end products are still poorly known.

The objective of the first research was to study the coupled process of microbial oxidation of Fe(II) in clay mineral nontronite (NAu-2), and nitrate reduction by Pseudogulbenkiania species strain 2002, and to determine mineralogical changes associated with this process. Bio-oxidation experiments were conducted using Fe(II) in microbially reduced nontronite as electron donor and nitrate as electron acceptor to investigate cell growth on this process. The bio-oxidation extent under growth and nongrowth conditions reached 67% and 57%, respectively. Over the same time period, nitrate was completely reduced. Abiotic oxidation by nitrite partly accelerated Fe(II) oxidation rate under the growth condition. The oxidized Fe(III) largely remained in the nontronite structure, but secondary minerals such as vivianite, ferrihydrite, and magnetite formed depending on specific experimental conditions.

The objective of the second research was to study microbially mediated redox cycles of Fe in nontronite (NAu-2). During the reduction phase, structural Fe(III) in NAu-2 served as electron acceptor, lactate as electron donor, AQDS as electron shuttle, and dissimilatory Fe(III)-reducing bacterium Shewanella putrefaciens CN32 as mediator. During the oxidation phase, biogenic Fe(II) served as electron donor and nitrate as electron acceptor. Nitrate-dependent Fe(II)-oxidizing bacterium Pseudogulbenkiania sp. strain 2002 was added as mediator in the same media. For all three cycles, structural Fe in NAu-2 was able to reversibly undergo three redox cycles without significant dissolution. Fe(II) in bioreduced samples occurred in two distinct environments, at edges and in the interior of the NAu-2 structure.

The overall objective of the third study was to study biological nitrate-dependent Fe(II) oxidation in illite IMt-1 and the effects of bio-oxidation on clay mineral transformation. Our data demonstrated that Pseudogulbenkiania sp. strain 2002 was able to couple oxidation of structural Fe(II) in IMt-1 with reduction of nitrate to N2 with nitrite as a transient intermediate. Fe(II)-oxidizing bacteria caused clay mineral structure change, and facilitated the illite→kaolinite and illite→smectite transformations. The biogenic smectite is a transient phase.

IRON REDOX PROCESS IN CLAY MINERALS & ITS ENVIRONMENTAL SIGNIFICANCE

A DISSERTATION

Presented to the Faculty of

Miami University in partial

fulfillment of the requirements

for the degree of

Doctor of Philosophy

Department of Geology and Environmental Earth Science

Linduo Zhao

The Graduate School Miami University Oxford, Ohio

2015

Dissertation Director: Hailiang Dong

Linduo Zhao

2015

TABLE OF CONTENTS

Chapter 1: Introduction ...... 1

References ...... 5

Chapter 2: Biological oxidation of Fe(II) in reduced nontronite coupled with nitrate reduction by Pseudogulbenkiania sp. Strain 2002 ...... 9 Abstract ...... 10 Body Text...... 11 References ...... 34 Tables and Figures ...... 42

Chapter 3: Biological redox cycling of iron in nontronite and its potential application in nitrate removal...... 56 Abstract ...... 57 Body Text...... 58 References ...... 73 Tables and Figures ...... 79

Chapter 4: Biological oxidation of Fe(II) in illite coupled with nitrate reduction and its role in clay mineral transform ...... 99 Abstract ...... 100 Body Text...... 101 References ...... 114 Tables and Figures ...... 117

Chapter 5: Conclusions and further recommendations ...... 129

iii

LIST OF TABLES

Chapter 2: Biological oxidation of Fe(II) in reduced nontronite coupled with nitrate reduction by Pseudogulbenkiania sp. Strain 2002 ...... 9

1- 77 K Fitting and calculated Mossbauer parameters...... 42

Chapter 3: Biological redox cycling of iron in nontronite and its potential application in nitrate removal...... 56

1- Bio-reduction and bio-oxidation extents and initial rates ...... 79

2- Fe L3 EELS peak position, the L3/L2 peak intensity ratios, and the comparisons of Fe(II) contents between the EELS and the chemical results ...... 80

3- Calculated Mössbauer parameters for the room temperature spectrum of the 2nd bioreduced NAu-2 sample with AQDS ...... 81

Chapter 4: Biological oxidation of Fe(II) in illite coupled with nitrate reduction and its role in clay mineral transformation ...... 99

1- Fe(II) concentration before and after bio-oxidation, bio-oxidation extents and rates 117

LIST OF FIGURES

Chapter 2: Biological oxidation of Fe(II) in reduced nontronite coupled with nitrate reduction by Pseudogulbenkiania sp. Strain 2002 ...... 9

1- Time-course changes of concentrations of Fe(II) in reduced NAu-2, nitrate, nitrite, protein (a proxy for biomass), and N2 produced from the nitrate-dependent Fe(II) oxidation using strain 2002 ...... 46

2- X-ray diffraction patterns for bioreduced NAu-2 sample (i.e., onset of bio-oxidation) and for those oxidized by strain 2002 for various times (10, 30 and 60 days) under the growth (A) and the nongrowth conditions (B) ...... 47

3- Comparison of XRD patterns between abiotic control (e.g., bioreduced by strain CN32 cells but never oxidized by strain 2002 cells) and microbially oxidized samples after 60 days of inoculation of strain 2002 cells under the growth condition ...... 48

4- X-ray diffraction patterns for the NAu-2 sample after reduction by CN32 cells for 2 weeks and oxidation by strain 2002 cells under the growth condition for 6 months. Vivianite and magnetite were detected in this sample ...... 49

5- Secondary electron microscopic (SEM) images & energy dispersive spectroscopy (EDS) spectrum of NAu-2 samples after nitrate-dependent Fe(II)-oxidation of reduced NAu-2 by strain 2002 under the growth condition for 30 days ...... 50

6- SEM images of vivianite from the NAu-2 sample following reduction by CN32 cells for 2 weeks and oxidation by strain 2002 cells under the growth condition for 6 months ...... 51

7- Transmission electron microscopic (TEM) image of a strain 2002 cell with cell encrustation under the growth condition for 6 months ...... 52

8- TEM image of ferrihydrite from the nitrate-dependent Fe(II)-oxidation experiment with strain 2002 cells for 60 days ...... 53

9- TEM images of secondary mineral and NAu-2 from the nitrate-dependent Fe(II) oxidation experiment with strain 2002 for 6 months...... 54

10- 77-K Mössbauer spectra for different NAu-2 samples ...... 55

Chapter 3: Biological redox cycling of iron in nontronite and its potential application in nitrate removal...... 56

1- Time-course changes of concentrations of Fe(II) in NAu-2, nitrate, and nitrite through three redox cycles with bicarbonate (gray symbols) and PIPES (black symbols) buffers...... 86

2- EELS spectra of bioreduced and bio-oxidized NAu-2 samples showing L3 and L2 edges of structural Fe in NAu-2 as a function of increased Fe redox cycle...... 87

3- Mössbauer spectra for the second bioreduced NAu-2 sample in the presence of AQDS at (A) RT and (B) 77 K, showing an increase in Fe(II) content from 14% at RT to 30% at 77 K...... 88

4- Second derivative FTIR spectra of OH-stretching and OH-bending bands in NAu-2 through three redox cycles...... 89

5- Schematic diagrams of Fe(III) and Fe(II) sites in (A) unaltered, (B) biologically reduced (with AQDS), and (C) biologically reoxidized NAu-2...... 90

6- Macroscopic color changes of NAu-2 through 3 Fe redox cycles. Bio-reduced (A, C, & E) and bio-oxidized (B, D, & F) NAu-2 from the 1st, 2nd, and 3rd cycles, respectively .....91

st 7- Time-course change of N2 gas through the 1 redox cycle with bicarbonate and PIPES buffers...... 92

8- X-ray diffraction patterns for NAu-2 samples through three microbial redox cycles in bicarbonate buffer ...... 93

nd 9- SEM & TEM images of biogenic albite (NaAlSi3O8) from the 2 bio-reduced NAu-2 sample ...... 94

10- TEM images of the bio-oxidized NAu-2 samples from the 1st and 2nd redox cycle, which were selected as representatives for 6 samples from three cycles...... 95

3+ 11- Calibration curve between the integral intensity ratio of L3/L2 and the Fe /total Fe ratio using 4 standard minerals ...... 96 vi

12- Comparison of two room-temperature (RT) Mössbauer spectra for unaltered and bioreduced NAu-2 showing the emergence of a distinct Fe(II) doublet and broadening of the central doublet after bio-reduction of structural Fe(III) in NAu-2, and a tentative model fit of the RT spectrum for the bio-reduced NAu-2 showing various contributions of Fe species ...... 97

13- Mössbauer spectra of biologically reduced NAu-2 in absence of AQDS at room temperature and 77 K showing a broad signal without any distinct Fe(II) doublet at RT and distinct doublets due to Fe(III) (82%) and Fe(II) (18%) at 77K, respectively ...... 98

Chapter 4: Biological oxidation of Fe(II) in illite coupled with nitrate reduction and its role in clay mineral transformation ...... 99

1- Time-course changes of concentrations of Fe(II) in IMt-1 and NAu-2, nitrate, and nitrite produced from the nitrate-dependent Fe(II) oxidation using strain 2002 ...... 120

2- X-ray diffraction patterns of air-dried original IMt-1 sample showing illite, air-dried bio-oxidized IMt-1 samples after 3 months of incubation showing smectite and kaolinite, and air-dried bio-oxidized IMt-1 samples after 2 years of incubation showing only kaolinite...... 121

3- X-ray diffraction patterns of air-dried and ethylene glycolated bio-oxidized IMt-1 sample after 3 months of incubation, showing that 14 Å d-spacing peak was present in air-dried sample, but absent in ethylene glycolated sample ...... 122

4- Second derivative FTIR spectra of OH-stretching bands in the abiotic control and bio- oxidized samples after 1 year of incubation ...... 123

5- SEM images showing neoformed biogenic kaolinite and smectite as a result of bio- oxidation of structural Fe(II) in 1Mt-1 after 3 months ...... 124

6- SEM images of neoformed biogenic kaolinite from the bio-oxidized IMt-1 sample after 2 years of incubation...... 125

7- HRTEM images of the neoformed biogenic smectite produced in bio-oxidized IMt-1 sample ...... 126

vii

8- HRTEM images of bio-oxidized IMt-1 samples after 2 years of incubation, showing the neoformed biogenic kaolinite particles ...... 127

9- HRTEM images showing structure changes of IMt-1 illite as a result of bio-oxidation for 1 month ...... 128

viii

DEDICATION

I dedicate this to my mother and friends.

ACKNOWLEDGEMENTS

I would like to express my sincere gratitude to many people who have helped me during my Ph.D. study. The first and the foremost is my advisor Dr. Hailiang Dong. Hailiang took me on as his Ph.D. student in 2010. His knowledge and enthusiasm as a scientist has been invaluable to my research. Throughout the five years here at Miami, Hailiang has consistently shown a confidence in my ability to do research. I am grateful to his persistence, patience, and guidance.

Thank you to everyone who served and has served on my committees: Jonathan Levy, John Rakovan, Jason Rech, Mark Kreleler, Richard Edelmann, Andre Sommer, and Christopher Gorski. I would also like to thank John Morton for his assistance in labs. I thank Richard Edelmann and Matthew Duley in Electron Microscopy Facility for training and helping me on all electron microscopies.

The friendship and insightful conversations provided by my fellow graduate students have been great. I’d like to thank all of my fellow graduate students. During many years at Miami, I made lots of great friendships with lots of different people. I would especially like to thank Brandon Briggs, Deng Liu, Jing Zhang, Qiuyuan Huang, Rajesh Singh, Michael Bishop, Li Zhang, Shreya Srivastava, and Alex Kugler, for their advice and encouragement. Finally, I would like to thank my mother for always giving me encouragement, listening to me and believing in me.

These projects were supported by the National Science Foundation (to Hailiang Dong), and Clay Mineral Society (to Linduo Zhao).

CHAPTER 1 INTROCUTION Iron-bearing clay minerals are ubiquitous in natural sediments, soils, and sedimentary rocks (Dong et al., 2009; Stucki and Kostka, 2006). Studies have shown that the physical and chemical properties of iron-bearing clay minerals, such as layer charge, layer expandability, and surface area, are influenced by changes in the Fe valence state (Jaisi et al., 2008a; Dong et al., 2009; Bishop et al., 2011; Dong, 2012). Such redox modified iron-bearing clay minerals hold a potential for removal of environmental contaminants, such as nitrate (Ernstsen, 1996; Ernstsen et al., 1998; Shelobolina et al., 2012a,b), and can be obtained chemically or biologically, i.e., by dithionite or phyllosilicate-Fe redox cycling organisms. Prior studies have confirmed the presence and the active roles of dissimilatory iron reducing bacteria (DIRB) and nitrate-dependent Fe(II)-oxidizing bacteria (NDIOB) in various redox active environments where there are abundant iron- bearing clay minerals, suggesting that these bacteria make important contributions to redox cycling of Fe, N, and C in nature (Konhauser et al., 2011).

In the last two decades, the discovery of nitrate-dependent Fe(II)-oxidizing bacteria (NDIOB) (Straub et al., 1996; Weber et al., 2006a) in diverse natural environments has promoted extensive research in iron biogeochemistry and its relation to nitrogen cycling. These bacteria can utilize the following reaction (1) to provide energy to support their growth. They have been observed in various environments, including lakes, streams, hydrothermal vents, wetlands, and aquifer sediments (Hafenbradl et al., 1996; Straub et al., 1996; Chaudhuri et al., 2001; Weber et al., 2001, 2006a; Edwards et al., 2003; Shelobolina et al., 2003a, 2012b). Although many pure cultures of Fe(II)-oxidizing bacteria capable of coupling nitrate reduction to Fe(II) oxidation have been isolated, only strain 2002 has shown the ability to autotrophically grow by Fe(II) oxidation (Weber et al., 2006b, 2009) under a circumneutral-pH condition, and only Strain 2002 (Weber et al., 2006b, 2009) and strain G2 (Shelobolina et al., 2003b) is capable of oxidizing Fe(II) in the absence of an organic co-substrate (e.g., acetate).

2+ − + 10퐹푒 + 2푁푂3 + 24퐻2푂 → 10퐹푒(푂퐻)3 + 18퐻 + 푁2 (1)

With nitrate as electron acceptor, the Fe(II)-oxidizing microorganisms are able to oxidize different forms of Fe(II), including aqueous Fe2+ (Straub et al., 1996; Lacket al., 2002; Weber et al., 2006b), mineral-bound Fe(II) (Chaudhuri et al., 2001; Weber et al., 2001), and structural Fe(II) in phyllosilicates (Shelobolina et al., 2012a; Shelobolina et al., 2012b). The coupled reaction between nitrate reduction and Fe(II) oxidation can produce a transient accumulation of nitrite (Hafenbradl et al., 1996; Straub et al., 1996;

Shelobolina et al., 2003a; Weber et al., 2006b), N2 gas (Benz et al., 1998; Weber et al., 2006b,c), and secondary mixed-valence Fe-bearing minerals; under certain conditions, ammonium (NH4) and nitrous oxide (N2O) may be produced by the coupled reaction between nitrate and Fe(II) (Weber et al., 2006b,c). A variety of secondary Fe(III)-bearing minerals can form during biologically mediated Fe(II) oxidation, including ferrihydrite (Lack et al., 2002; Miot et al., 2009b), magnetite, hematite (Chaudhuri et al., 2001), goethite (Miot et al., 2009b; Larese-Casanova et al., 2010), lepidocrocite and green rust (Chaudhuri et al., 2001; Lack et al., 2002), and Fe(III) phosphates (Larese-Casanova et al., 2010). The types of mineral products formed depend on geochemical conditions, bacterial strains involved, and the oxidation rates of ferrous iron.

To date, there are only two studies where NDIOB has been demonstrated to oxidize structural Fe(II) in phyllosilicates. Shelobolina et al. (2012a) showed that structural Fe(II) in trioctahedral phyllosilicate biotite can be microbially oxidized when coupled with nitrate reduction. Since the extent of Fe(II) oxidation was small and nitrate was not completely removed, the products of Fe(II) oxidation were not definitively identified by either X-ray diffraction (XRD) or Mössbauer spectroscopy. In a separate study (Shelobolina et al., 2012b), the same authors studied microbially catalyzed nitrate- dependent Fe(II) oxidation in illite–smectite mixed layer phyllosilicates of a natural soil. Because smectite and illite usually exhibit drastically different Fe(III) reduction extent and rate (Dong et al., 2009; Dong, 2012), it is likely that these two minerals have different Fe(II) oxidation extent and rate as well, however, it is currently unknown how different these two important end members of dioctahedral phyllosilicates, i.e., smectite and illite, are in their Fe(II) oxidation extents and rates.

Previous research has proven that most Fe in phyllosilicate minerals can undergo biotic/abiotic reduction followed by abiotic oxidation without extensive solubilization (Stucki, J.W., 2011). Two recent studies (Shelobolina et al., 2012a, 2012b) demonstrated that, as discussed above, biological oxidation of structural Fe(II) in phyllosilicates can be coupled with nitrate reduction. Together, these studies suggest that structural Fe in phyllosilicates has the potential to undergo multiple redox cycles with important implications for environmental processes including continuous removal of nitrate in groundwater aquifers. Given the important role of phyllosilicates in sustaining a dynamic Fe redox cycle, much has been studied about the effects of the oxidation state of structural Fe on the chemical and physical properties of phyllosilicates (Stucki, J.W., 2011). However, most research has been conducted on chemical redox cycling of Fe in clay minerals (Yang, J., 2010; Ribeiro et al., 2009; Lee, et al., 2006), biological redox cycling of Fe, despite its significance in the biogeochemical cycles of Fe, N, and C (Dong, et al., 2009; Stucki, J.W., 2011), has not received equivalent attention.

Illite is common in consolidated sedimentary rocks at depth (Moore, et al., Oxford University Press), where anaerobic microbial community and nitrate reduction zone are likely to be co-present with illite. Therefore, oxidation of structural Fe(II) in illite can be coupled with nitrate reduction by indigenous microbial community in these areas. In addition, previous studies have reported that illite could transform to I/S mixed-layer clay by leaching interlayer K+ under acidic or circumneutral environments (Rimmer, et al., 1982). However, microbial activity has never been considered as an environmental variable in promoting illite-to-smectite transformation. Previous study has shown microbially mediated smectite-illite reaction that was achieved through reduction of structural Fe(III) (Kim, et al., 2003), it is possible that bio-oxidation of structural Fe(II) in illite can reverse the process, i.e., illite-to-smectite transformation. Together, nitrate- dependent Fe(II)-oxidizing bacteria may have a potential in promoting illite-smectite transformation and thereby close the microbial loop of clay mineral-microbe interactions, however, this pathway has never been explored. Additional studies are needed to better understand microbe-illite interaction.

The goal of this PhD research is to develop a comprehensive understanding of Fe biogeochemistry associated with clay minerals, specifically focusing on Fe(II) oxidation in clay mineral coupled with nitrate reduction by NDIOB, microbially mediated redox cycling of structural Fe in nontronite (NAu-2), and the effect of iron-oxidizing bacteria on clay mineral transformation. This research, therefore, is focused to address the following fundamental questions:

(1) What are the extent and rate of microbial oxidation of structural Fe(II) in iron-rich smectite (nontronite, NAu-2) and illite (IMt-1)?

(2) How do extent and rate of Fe reduction and oxidation change with increased Fe redox cycle in nontronite?

(3) How do mineralogical and structural/chemical properties of NAu-2 change as a result of Fe redox cycles mediated by DIRB and NDIOB?

(4) What is the clay mineral transformation associated with the microbially mediated Fe(II) oxidation in clays?

REFERENCES

Benz, M., Brune, A., and Schink, B., 1998. Anaerobic and aerobic oxidation of ferrous iron at neutral pH by chemoheterotrophic nitrate-reducing bacteria. Arch. Microbiol. 169, 159-165.

Bishop, M. E., Dong, H. L., Kukkadapu, R. K., Liu, C. X., and Edelmann, R. E., 2011 Bioreduction of Fe-bearing clay minerals and their reactivity toward pertechnetate (Tc-99). Geochim. Cosmochim. Acta 75, 5229-5246.

Chaudhuri, S. K., Lack, J. G., and Coates, J. D., .2001. Biogenic Magnetite Formation through Anaerobic Biooxidation of Fe(II). Appl. Environ. Microbiol. 67, 2844- 2848.

Dong, H. L., Jaisi, D. P., Kim, J., and Zhang, G. X., 2009. Microbe-clay mineral interactions. Am. Mineral. 94, 1505-1519.

Dong, H. L., 2012. Clay-Microbe Interactions and Implications for Environmental Mitigation. Elements 8, 113-118.

Edwards, K. J., Bach, W., and Rogers, D. R., 2003. Geomicrobiology of the ocean crust: A role for chemoautotrophic Fe-bacteria. Biol. Bull. 204, 180-185.

Ernstsen, V., 1996. Reduction of nitrate by Fe2+ in clay minerals. Clays Clay Miner. 44, 599-608.

Ernstsen, V., Gates, W. P., and Stucki, J. W., 1998. Microbial reduction of structural iron in clays - A renewable source of reduction capacity. J Environ Qual 27, 761-766.

Hafenbradl, D., Keller, M., Dirmeier, R., Rachel, R., Roβnagel, P., Burggraf, S., Huber, H., and Stetter, K. O., 1996. Ferroglobus placidus gen. nov., sp. nov., a novel hyperthermophilic archaeum that oxidizes Fe2+ at neutral pH under anoxic conditions. Arch. Microbiol. 166, 308-314.

Jaisi, D. P., Liu, C. X., Dong, H. L., Blake, R. E., and Fein, J. B., 2008a. Fe2+ sorption onto nontronite (NAu-2). Geochim. Cosmochim. Acta 72, 5361-5371.

Kim, J.W., Dong, H., Seabaugh, J., Newell, S., and Eberl, D.D., 2004. Role of microbes in the smectite-to-illite reaction. Science. 303(6): 830-832.

Konhauser, K. O., Kappler, A., and Roden, E. E., 2011. Iron in Microbial Metabolisms. Elements 7, 89-93.

Lack, J. G., Chaudhuri, S. K., Chakraborty, R., Achenbach, L. A., and Coates, J. D., 2002. Anaerobic biooxidation of Fe(II) by Dechlorosoma suillum. Microb. Ecol. 43, 424-431.

Larese-Casanova, P., Haderlein, S. B., and Kappler, A., 2010. Biomineralization of lepidocrocite and goethite by nitrate-reducing Fe(II)-oxidizing bacteria: Effect of pH, bicarbonate, phosphate, and humic acids. Geochim. Cosmochim. Acta 74, 3721-3734.

Lee, K.; Kostka, J. E.; Stucki, J. W., 2006. Comparisons of structural Fe reduction in smectites by bacteria and dithionite: an infrared spectroscopic study. Clays and Clay Minerals, 54 (2), 195-208.

Miot, J., Benzerara, K., Morin, G., Bernard, S., Beyssac, O., Larquet, E., Kappler, A., and Guyot, F., 2009a. Transformation of vivianite by anaerobic nitrate-reducing iron- oxidizing bacteria. Geobiology 7, 373-384.

Miot, J., Benzerara, K., Morin, G., Kappler, A., Bernard, S., Obst, M., Ferard, C., Skouri- Panet, F., Guigner, J. M., Posth, N., Galvez, M., Brown, G. E., and Guyot, F., 2009b. Iron biomineralization by anaerobic neutrophilic iron-oxidizing bacteria. Geochim. Cosmochim. Acta 73, 696-711.

Moore, D. M., and Reynolds, R. C. Jr., 1997. X-ray diffraction and the identification and analysis of clay minerals. Oxford University Press.

Ribeiro, F. R.; Fabris, J. D.; Kostka, J. E.; Komadel, P.; Stucki, J. W., 2009. Comparisons of structural iron reduction in smectites by bacteria and dithionite: II. A variable- temperature Mössbauer spectroscopic study of Garfield nontronite. Pure and Applied Chemistry. 81 (8), 1499-1509.

Rimmer, S. M., 1982. Origin of an underclay as revealed by vertical variations in mineralogy and chemistry. Clays and Clay Minerals. 30(6), 422-430

Shelobolina, E. S., O'Neill, K., Finneran, K. T., Hayes, L. A., and Lovley, D. R., 2003a. Potential for in situ bioremediation of a low-pH, high-nitrate uranium- contaminated groundwater. Soil & Sediment Contam. 12, 865-884.

Shelobolina, E. S., Vanpraagh, C. G., and Lovley, D. R., 2003b. Use of ferric and ferrous iron containing minerals for respiration by Desulfitobacterium frappieri. Geomicrobiol. J. 20, 143-156.

Shelobolina, E., Xu, H., Konishi, H., Kukkadapu, R., Wu, T., Blöthe, M., and Roden, E. 2012a. Microbial Lithotrophic Oxidation of Structural Fe(II) in Biotite. Appl. Environ. Microbiol. 78, 5746-5752.

Shelobolina, E., Konishi, H., Xu, H., Benzine J., Xiong, M., Wu, T., Blöthe, M., and Roden, E. 2012b. Isolation of phyllosilicate-iron redox cycling microorganisms from an illite-smectite rich hydromorphic soil. Front. Microbiol. 3:134. doi: 10.3389/fmicb.2012.00134.

Straub, K. L., Benz, M., Schink, B., and Widdel, F., 1996. Anaerobic, nitrate-dependent microbial oxidation of ferrous iron. Appl. Environ. Microbiol. 62, 1458-1460.

Stucki, J. W.; Kostka, J. E., 2006. Microbial reduction of iron in smectite. Comptes Rendus Geoscience, 338 (6-7), 468-475.

Stucki, J. W., 2011. A review of the effects of iron redox cycles on smectite properties. Comptes Rendus Geoscience. 343 (2-3), 199-209.

Weber, K. A., Picardal, F. W., and Roden, E. E., 2001. Microbially catalyzed nitrate- dependent oxidation of biogenic solid-phase Fe(II) compounds. Environ. Sci. Technol. 35, 1644-1650.

Weber, K. A., Achenbach, L. A., and Coates, J. D., 2006a. Microorganisms pumping iron: anaerobic microbial iron oxidation and reduction. Nat. Rev. Microbiol. 4, 752-764.

Weber, K.A., Pollock, J., Cole, K.A., O'Connor, S.M., Achenbach, L.A., Coates, J.D., 2006b. Anaerobic nitrate-dependent iron(II) bio-oxidation by a novel lithoautotrophic betaproteobacterium, strain 2002. Appl. Environ. Microbiol. 72, 686-694.

Weber, K. A., Hedrick, D. B., Peacock, A. D., Thrash, J. C., White, D. C., Achenbach, L. A., and Coates, J. D., 2009. Physiological and taxonomic description of the novel autotrophic, metal oxidizing bacterium, Pseudogulbenkiania sp strain 2002. Appl. Microbiol. Biotechnol. 83, 555-565.

Yang, J., 2010. Effects of redox cyclingsof iron in nontronite on reduction of technetium. M.S. Dissertation, Miami University, Oxford, OH.

CHAPTER 2 Biological oxidation of Fe(II) in reduced nontronite coupled with nitrate reduction by Pseudogulbenkiania sp. Strain 2002

Linduo Zhao1, Hailiang Dong1,2,*, Ravi Kukkadapu3, Abinash Agrawal4,

Deng Liu5, Jing Zhang1, Richard E. Edelmann6

1Department of Geology and Environmental Earth Science, Miami University, OH 45056, USA 2Geomicrobiology Laboratory, State Key Laboratory of Geobiology and Environmental Geology, China University of Geosciences, Beijing 100083, China 3EMSL, Pacific Northwest National Laboratory, Richland, WA 99352, USA 4Department of Earth & Environmental Sciences, Wright State University, Dayton, OH 45435, USA 5State Key Laboratory of Geobiology and Environmental Geology, China University of Geosciences, Wuhan 430074, China 6Center for Advanced Microscopy & Imaging, Miami University, Oxford, OH 45056, USA

Department of Geology Miami University Oxford, OH 45056 Tel: 513-529-2517 Fax: 513-529-1542 Email: [email protected]

Geochimica et Cosmochimica Acta 2013, 119, pp. 231–247

ABSTRACT The importance of microbial nitrate-dependent Fe(II) oxidation to iron biogeochemistry is well recognized. Past research has focused on oxidation of aqueous Fe2+ and structural Fe(II) in oxides, carbonates, and phosphate, but the importance of structural Fe(II) in phyllosilicates in this reaction is only recently studied. However, the effect of clay mineralogy on the rate and the mechanism of the reaction, and subsequent mineralogical end products are still poorly known. The objective of this research was to study the coupled process of microbial oxidation of Fe(II) in clay mineral nontronite (NAu-2), and nitrate reduction by Pseudogulbenkiania species strain 2002, and to determine mineralogical changes associated with this process. Bio-oxidation experiments were conducted using Fe(II) in microbially reduced nontronite as electron donor and nitrate as electron acceptor in bicarbonate-buffered medium under both growth and nongrowth conditions to investigate cell growth on this process. The extents of Fe(II) oxidation and nitrate reduction were measured by wet chemical methods. X-ray diffraction (XRD), scanning and transmission electron microscopy (SEM and TEM), and 57Fe-Mössbauer spectroscopy were used to observe mineralogical changes associated with Fe(III) reduction and Fe(II) oxidation in NAu-2. The bio-oxidation extent under growth and nongrowth conditions reached 67% and 57%, respectively. Over the same time period, nitrate was completely reduced under both conditions to nitrogen gas (N2), via an intermediate product nitrite. Abiotic oxidation by nitrite partly accelerated Fe(II) oxidation rate under the growth condition. The oxidized Fe(III) largely remained in the nontronite structure, but secondary minerals such as vivianite, ferrihydrite, and magnetite formed depending on specific experimental conditions. The results of this study highlight the importance of iron-bearing clay minerals in the global nitrogen cycle with potential applications in nitrate removal in natural environments. Key words: iron oxidation, nitrate-dependent Fe(II) oxidizing bacteria, nontronite

1. INTRODUCTION The discovery of nitrate-dependent Fe(II)-oxidizing bacteria (Straub et al., 1996; Weber et al., 2006a) in diverse natural environments has promoted extensive research in iron biogeochemistry and its relation to nitrogen cycling. Along with dissimilatory iron reducing bacteria (DIRB) and other iron-oxidizing bacteria, the nitrate-dependent Fe(II)- oxidizing microorganisms make important contributions to Fe redox cycling in nature (Konhauser et al., 2011). In ancient Fe(II)-rich Precambrian environment and in modern soils and aquatic sediments (Konhauser et al., 2002; Weber et al., 2006 a,b,c, 2009; Coby et al., 2011; Shelobolina et al., 2012 a,b), the redox transition between the Fe(II) and Fe(III) valence states plays an important role in providing energy to support microbial growth and in coupling the biogeochemical cycles of Fe, C, and N. Although both iron reduction and oxidation processes are thought to play an important role in sedimentary environments (Blothe and Roden, 2009; Konhauser et al., 2011), iron oxidation has received less attention than iron reduction in the last 20 years. The microbial communities involved in N and Fe redox cycling are present in various environments, including lakes, streams, hydrothermal vents, wetlands, and aquifer sediments (Hafenbradl et al., 1996; Straub et al., 1996; Chaudhuri et al., 2001; Weber et al., 2001, 2006a; Edwards et al., 2003; Shelobolina et al., 2003a, 2012b). Many pure cultures of Fe(II)-oxidizing bacteria capable of coupling nitrate reduction to Fe(II) oxidation have been isolated; for example, Desulfitobacterium frappieri strain G2 (Shelobolina et al., 2003b), Acidovorax sp. strain 2AN (Chakraborty et al., 2011), Acidovorax sp. strain BoFeN1 (Kappler et al., 2005), Dechlorosoma suillum strain PS (Chaudhuri et al., 2001; Lack et al., 2002), Pseudogulbenkiania sp. strain 2002 (Weber et al., 2006b, 2009), hyperthermophilic archaeon Ferroglobus placidus (Hafenbradl et al., 1996), Acidovorax ebreus strain TPSY (Byrne-Bailey et al., 2010), and Bradyrhizobium spp. and strains of Cupriavidus necator and Ralstonia solanacearum (Shelobolina et al., 2012b). Among these, only strain 2002 has shown the ability to autotrophically grow by Fe(II) oxidation (Weber et al., 2006b, 2009) under a circumneutral-pH condition. Strain 2002 (Weber et al., 2006b, 2009) and strain G2 (Shelobolina et al., 2003b) are capable of oxidizing Fe(II) in the absence of an organic co-substrate (e.g., acetate).

With nitrate as electron acceptor, the Fe(II)-oxidizing microorganisms are able to oxidize different forms of Fe(II), including aqueous Fe2+ (Straub et al., 1996; Lack et al., 2002; Weber et al., 2006b) and mineral-bound Fe(II) (Chaudhuri et al., 2001; Weber et al., 2001). The coupled reaction between nitrate reduction and Fe(II) oxidation can produce a transient accumulation of nitrite (Hafenbradl et al., 1996; Straub et al., 1996;

Shelobolina et al., 2003a; Weber et al., 2006b), N2 gas (Benz et al.,1998; Weber et al., 2006b,c), and secondary Fe(II)- and mixed-valence Fe-bearing minerals; under certain conditions, ammonium (NH4) and nitrous oxide (N2O) may be produced by the coupled reaction between nitrate and Fe(II) (Weber et al., 2006b,c). In most cases, nitrate is completely reduced in a few weeks. Nitrate-dependent oxidation of aqueous Fe2+ has been described as follows (Straub et al., 1996): 2+ − + 10퐹푒 + 2푁푂3 + 24퐻2푂 → 10퐹푒(푂퐻)3 + 18퐻 + 푁2 A variety of secondary Fe(III)-bearing minerals may form during biologically mediated Fe(II) oxidation, depending on the source of Fe(II) (e.g., aqueous vs. solid phase). These mineral products range from poorly crystalline materials such as ferrihydrite (Lack et al., 2002; Miot et al., 2009b), to crystalline minerals including magnetite, hematite (Chaudhuri et al., 2001), goethite (Miot et al., 2009b; Larese- Casanova et al., 2010), lepidocrocite and green rust (Chaudhuri et al., 2001; Lack et al., 2002), and Fe(III) phosphates (Larese-Casanova et al., 2010). The types of mineral products formed depend on geochemical conditions, bacterial strains involved, and the oxidation rates of ferrous iron. Senko et al. (2005) and Lack et al., (2002) showed that the formation of either ferrihydrite, green rust, or magnetite was a function of the rate of Fe(II) oxidation. Pantke et al. (2011) demonstrated the formation of green rust during nitrate-dependent Fe(II) oxidation, which is usually considered as a precursor for the formation of goethite (Kappler et al., 2005), hematite, and magnetite (Chaudhuri et al., 2001; Pantke et al., 2011). Ferrous iron occurs in a number of forms including aqueous Fe2+, structural Fe(II) in oxides and carbonates, and clay minerals. Despite the ubiquitous presence of iron-bearing phyllosilicates in soils, sediments and sedimentary rocks, it is surprising to note that most studies have focused on microbial reduction of Fe(III) in clay minerals (Dong et al., 2009; Dong, 2012), but not so much on Fe(II) oxidation except for two studies

(Shelobolina et al., 2012a,b). Studies have shown that the physical and chemical properties of Fe-bearing clay minerals, such as layer charge, layer expandability, surface area, are influenced by microorganisms through changes in the Fe valence state (Jaisi et al., 2008a; Dong et al., 2009; Bishop et al., 2011; Dong, 2012). Such redox modified clay minerals containing structural Fe(II) hold a potential for removal of environmental contaminants, including nitrate (Ernstsen, 1996; Ernstsen et al., 1998; Shelobolina et al., 3- + 2012a,b); aqueous phosphate (PO4 ) (Violante and Pigna, 2002), ammonium (NH4 ) (Weber et al., 2006c), arsenic (As) (Hohmann et al., 2010), technetium (Tc) (Jaisi et al., 2009; Bishop et al., 2011), and uranium (U) (Zhang et al., 2011). Two recent studies have recently reported nitrate-dependent oxidation of structural Fe(II) in phyllosilicates. Shelobolina et al. (2012a) showed that structural Fe(II) in trioctahedral phyllosilicate biotite can be microbially oxidized when coupled with nitrate reduction. Since the extent of Fe(II) oxidation was small and nitrate was not completely removed, the products of Fe(II) oxidation were not definitively identified by either X-ray diffraction (XRD) or Mössbauer spectroscopy. Transmission electron microscopy (TEM) revealed precipitation of poorly-crystalline Fe(III) oxyhydroxides. In a separate study (Shelobolina et al., 2012b), the same authors studied microbially catalyzed nitrate- dependent Fe(II) oxidation in dioctahedral phyllosilicates, but the exact nature of the phyllosilicates was not determined because the reduced clay size fractions of a natural soil were used as a source of Fe(II). XRD and TEM identified illite–smectite mixed layers in the size fractions. Because smectite and illite usually exhibit drastically different Fe(III) reduction extent and rate (Dong et al., 2009; Dong, 2012), it is likely that these two minerals have different Fe(II) oxidation extent and rate as well, however, it is currently unknown how different these two important end members of dioctahedral phyllosilicates are in their Fe(II) oxidation extents and rates. In light of the significant report that Fe(II)-bearing clay minerals can influence nitrogen and iron biogeochemical cycles, as well as their potential applications in removing nitrate contamination from natural environments, further studies are needed, especially using pure, well- characterized clay mineral such as smectite. Furthermore, it is important to determine the mechanism of nitrate-dependent Fe(II) oxidation in phyllosilicates and to identify mineral end products.

The objective of the present study was therefore to fill the above knowledge gap by studying a coupled reaction between reduction of nitrate and oxidation of Fe(II) in a microbially reduced iron-rich smectite, nontronite (NAu-2), by an autotrophic Fe(II) oxidizer Pseudogulbenkiania sp. strain 2002. Specifically, the aims of the study were to determine: (1) the extent and rate of microbial oxidation of structural Fe(II) in an iron- rich smectite (nontronite); (2) the effect of this Fe(II) oxidation on nitrate removal; (3) mineralogical changes associated with this nitrate dependent Fe(II) oxidation; and (4) the role of microbial growth condition on the reaction kinetics. Nontronite (NAu-2), as a well-characterized and well-studied Fe-rich smectite variety, was used as a structural Fe source. In its structure, one octahedral sheet is sandwiched between two tetrahedral sheets (Stucki and Roth, 1977). Prior to nitrate-dependent Fe(II)- oxidation experiment, structural Fe(III) in nontronite was first biologically reduced (Yang et al., 2012), and the resulting Fe(II)-bearing nontronite was then investigated for microbial oxidation with nitrate as the sole electron acceptor. The stoichiometry of Fe(II) oxidation and nitrate reduction were quantified. Various intermediate products of nitrate reduction were measured. Mineralogical changes associated with Fe(III) reduction and nitrate-dependent Fe(II) oxidation were determined by using XRD, scanning electron microscopy (SEM), high resolution transmission electron microscopy (HRTEM), and Mössbauer spectroscopy. The results of this study expand our understanding of the iron redox cycle associated with clay minerals with important implications for microbiological removal of nitrate contaminant in the environment. 2. MATERIALS AND METHODS 2.1 Mineral, media, and experimental setup 2.1.1. Preparation of microbially reduced NAu-2 Nontronite NAu-2, from Uley Mine in South Australia, was purchased from the Source Clays Repository of the Clay Minerals Society (Chantilly, Virginia, USA). The total Fe content of NAu-2 was 24%, of which 0.6% was Fe(II) (Keeling et al., 2000). The formula of this material is M0.72(Si7.55Al0.45)(Fe3.83Mg0.05)O20(OH)4, where M represents the interlayer cation, such as Na, Ca or K (Keeling et al., 2000; Gates et al., 2002). NAu- 2 consists of Fe(III) in both octahedral (92%) and tetrahedral sites (~8%) (Gates et al., 2002; Jaisi et al., 2005). The NAu-2 sample was size-fractionated prior to use as

14 previously described (Liu et al., 2011). The 0.02-0.5 µm size fraction was separated, dried, and made into 5 g/L clay suspension in double distilled water (ddH2O) followed by autoclaving. Published Mössbauer spectroscopy data indicated that there was no Fe(II) in the 0.02-0.5 µm size fraction (Yang et al., 2012). Biological reduction of structural Fe(III) in NAu-2 was performed using Shewanella putrefaciens CN32 with lactate as electron donor as previously described (Jaisi et al., 2007). After Fe reduction reached its maximum (30.8%), lactate was removed from the reduced clay suspension by washing it five times with sterile and anoxic (80:20 N2:CO2 purged) sodium bicarbonate buffer (29.76 mM NaHCO3 and 1.34 mM KCl, pH 7.0) and centrifugation at 1,000g for 10 min. The pH of the final suspension was adjusted to 7.0 with NaOH, and the nontronite concentration was adjusted to 6.3-6.8 g/L containing 392- 420 mg/L (7.0-7.5 mM) structural Fe(II). The nontronite suspension was then transferred to 120 mL serum bottles and pasteurized in an 80 ̊C water bath for 30 min immediately followed by cold water for 10 min. This pasteurization procedure was repeated three times followed by five washes with anoxic and sterile bicarbonate buffer to ensure that CN32 cells were killed and the nontronite suspension was sterilized. The sterility of the pasteurized nontronite suspension was confirmed by lack of strain CN32 cell growth on plates, as consistent with our previous results (Jaisi et al., 2008b, 2009). A complete removal of lactate and other organic carbon was confirmed by high-performance liquid chromatography (HPLC). 2.1.2. Preparation of Fe(II)-oxidizing bacterial culture Strain 2002 (ATCC BAA-1479; DSM 18807; kindly provided by Karrie A. Weber, University of Nebraska-Lincoln) is routinely cultured in an anoxic freshwater basal medium with acetate (10 mM) and nitrate (10 mM) as electron donor and acceptor, respectively, as previously described (Weber et al., 2006b). The inoculated medium was periodically checked for growth by measuring optical density at 600 nm on a spectrophotometer. Approximately 24 h after inoculation at 36 ̊C, when the culture reached the log phase, cells were harvested by centrifugation. To remove extra acetate and nitrate, cells were washed with sterile and anoxic (80:20 N2:CO2 purged) bicarbonate buffer three times by centrifugation (6000g for 10 min). After the final wash, cells were re-suspended in sterile and anoxic bicarbonate buffer as a stock solution.

2.1.3. Microbial oxidation of Fe(II) in microbially reduced Nontronite After the microbially reduced NAu-2 was transferred into 120 mL serum bottles (a final concentration 6.3-6.8 g/L) in an anaerobic glove box, bottles were divided into growth and nongrowth treatments with two replicates for each treatment. To the growth treatment, sterile and anoxic NaH2PO4 (5.4 mM), trace minerals (10 ml/L), and Wolfe vitamins (10 ml/L) (Weber et al., 2009) were added to examine the effects of cell growth on Fe(II) oxidation and nitrate reduction. In the nongrowth treatment, an equivalent volume of anoxic bicarbonate buffer was used to replace the NaH2PO4, trace minerals, and vitamins. The serum bottles were sealed with blue butyl rubber stoppers and aluminum crimps. Washed log-phase strain 2002 cells were then added into the 8 bottles (final concentration 1×10 cells/mL). A small volume (~0.17 mL) of a KNO3 stock solution (1 M conc.) was added as electron acceptor to the bottles to achieve a final concentration of 9-10 mM. The headspace of each bottle was purged with CO2 for 15 min. Abiotic controls were the same as the treatments except that they were devoid of strain 2002 cells. The total volume of the treatment solution was 17 mL and the headspace was 103 mL. 2.1.4. Reduction of nitrate by strain 2002 in absence of external electron donor Because the reaction stoichiometry revealed that the amount of nitrate reduced greatly exceeded the stoichiometric requirement for the amount of Fe(II) oxidized under both growth and non-growth conditions (below), one supplementary experiment was conducted to check the possibility of nitrate reduction by strain 2002 cells. A growth medium that contained 29.76 mM bicarbonate buffer, 5.4 mM NaH2PO4, 10 ml/L Wolfe vitamins, and 10 ml/L trace minerals (Weber et al., 2009) was prepared in 25 mL Balch tubes. All the bottles were made anoxic by purging with N2:CO2 mix gas (80:20) for 0.5 h, sealed with blue butyl rubber stoppers and aluminum caps, and autoclaved. Strain 2002 cells were grown to the log-phase and washed. To test nitrate reduction by energy reserves in active cells and biomass in inactive cells, various combinations of active and inactive cell concentrations were used. A fraction of the live culture was inactivated by autoclaving for 30 min. Three separate treatments were prepared: (1) inactive cells only (at two different concentrations: 2×105 and 1×106 cells/mL); (2) active cells only (0– 2×107 cells/mL); and (3) active (constant conc. at 2×105 cells/mL) + inactive cells (0–

7 3×10 cells/mL). A KNO3 stock solution was added into each bottle to achieve a final concentration of 10 mM. After 10 h, samples were taken and measured for nitrate concentration by the salicylate method as described below. 2.1.5. Abiotic oxidation of Fe(II) in microbially reduced NAu-2 by nitrite Since a certain amount of nitrite accumulated (up to 2 mM) under the growth condition and may abiotically oxidize Fe(II), another supplementary experiment was conducted with microbially reduced NAu-2 (final conc. 3 g/L) as electron donor and sterilized nitrite (NaNO2, final concentration 8.7 mM) as electron acceptor under both growth and nongrowth conditions in 25-mL Balch tubes. The bioreduced NAu-2 was sterilized by repeated pasteurization (3 times each at 20 min) and the sterility was confirmed by plating. Nitrite and Fe(II) concentrations were monitored over time by the sulphanilamide method and 1,10-phenanthroline method, respectively as described below. To explore the possibility of Fe(II) oxidation by nitrite during the Fe(II) measurement, the same experiment was repeated but under the acidic and boiling conditions that are required for the 1,10-phenanthroline method (Amonette and Templeton, 1998). 2.2. Analysis

2.2.1. Measurements of Fe(II), nitrate, nitrite, N2, and protein concentrations The reaction progress was monitored by collecting samples from incubation bottles over time for the quantification of total Fe(II), nitrate, nitrite, N2, and total cell protein during the experiment. Total Fe(II) concentration was analyzed by the 1,10- phenanthroline method (Amonette and Templeton, 1998). Nitrate and nitrite samples were collected from the supernatants following centrifugation (10,000g for 10 min) and their concentrations were determined by the sodium salicylate procedure (Yang et al., 1998) and the sulphanilamide method (Revanasiddappa and Kumar, 2003), respectively.

The amount of N2 in headspace of the serum bottles was quantified by a Hewlett Packard 5890 series II Gas Chromatographic (GC) system equipped with a thermal conductivity detector. Due to the difficulty in enumerating cells in the presence of clay particles, cell protein concentration was measured to evaluate relative changes in biomass over the course of the bio-oxidation experiment. However, cell protein concentration may not accurately represent biomass because protein content per cell could vary due to metabolic changes associated with specific chemical conditions. Nonetheless, we assume protein

17 concentration was an approximation of cell biomass based on its calibration to DAPI cell count. This calibration suggested that cell protein concentration and DAPI cell counts co- varied (Fig. EA-1) according to the following equation: 푌 = 0.0529푋 + 0.075, 푅2 = 0.95 where Y = DAPI cell counts (cells/mL), and X = protein concentration (µg/L). Time-course protein concentration was measured. Briefly, a cell-nontronite suspension of 0.1 mL in volume was washed by ddH2O three times. After addition of 0.2 M NaOH, cell suspension was boiled for 10 min to release protein from cells and centrifuged to harvest total protein from the supernatant. Protein fraction in the supernatant was quantified with the Bradford assay (Bradford, 1976) using bovine serum albumin (BSA) as a standard. 2.2.2. XRD Clay mineral smear mounts were prepared on petrographic slides and dried at 30 ̊C in a glove box incubator. All slides were kept in a desiccator containing ethylene glycol until XRD analysis in a humidity-controlled laboratory (Moore and Reynolds, 1997). The sue of ethylene glycolwas to expand smectite interlayers so that smectite can be distinguished from non-expandable clays such as illite. Powder XRD patterns were collected utilizing a Scintag X1 X-ray powder diffractometer with CuKa radiation and a fixed slit scintillation detector at 1400 W (40 kV, 35 mA). Scans were collected over a range of 2-70̊ two-Theta, but XRD patterns only up to 30̊ two-Theta are shown in the results. The MDI Jade 7 software was used for identification of mineral phases. The software utilizes International Center for Diffraction Data Powder Diffraction File Database (ICDD PDF-2, Sets 1–46, 1996) as a reference source. 2.2.3. SEM The spatial associations between the reduced NAu-2 and Fe(II)-oxidizing cells, and the morphology and particle size of various minerals were examined by scanning electron microscopic analysis. Clay suspensions were mounted onto glass cover slips, which were pretreated with poly-L-lysine for 10 min. After fixation with 2% paraformaldehyde and 2.5% glutaraldehyde, the samples on the cover slips were sequentially dehydrated through grade series of ethanol followed by critical point drying with a Tousimis Samdro-780A Critical Point Dryer (CPD) (Dong et al., 2003b). The cover slips were

18 mounted onto SEM stubs via clear double-sided sticky tape and gold-coated. The samples were observed with a Zeiss Supra 35 variable pressure (VP) SEM with EDAX Genesis 2000 X-ray energy dispersive spectroscopy (SEM/EDS) using 7-10 keV accelerating voltage and8.5 mm working distance. The EDS spectra provided a primary means for mineral composition. 2.2.4. TEM The nontronite samples that were oxidized for 10 days and 6 months under the growth condition were selected for TEM observations. The samples were diluted by a factor of 50, and pipetted onto 300 mesh copper grids with carbon-coated nitrocellulose membrane. The grids were dried overnight in an anaerobic glove box. TEM imaging and analysis were performed with a JEOL JEM-2100 LaB6 TEM/STEM with a 200 keV accelerating voltage. The bright-field imaging mode (TEM BF) was used to studythe morphology of NAu-2, secondary minerals, and cells. TEM images were recorded using a Gatan Orius SC200D camera attached on a Gatan 863 Tridiem GIF Post-Column Energy Filter EELS/EFTEM (Gatan Image Filter). For mineral identification, selected area electron diffraction (SAED) patterns (the Gatan Orius SC200D camera) and EDS spectra (Bruker AXS Microanalysis Quantax 200 with 4030 SDD detector) were obtained. 2.2.5. Mössbauer spectroscopy To study Fe speciation associated with microbial reduction of Fe(III) and oxidation of Fe(II) in nontronite, Mössbauer spectroscopy was performed for NAu-2 that was pristine, microbially reduced (reduction extent 30.8%), and microbially oxidized for 10 days under the growth condition. Mössbauer spectra of the samples were collected using either a WissEl Elektronik (Germany) or Web Research Company (St. Paul, MN) instruments that included a closed-cycle cryostat SHI-850 obtained from Janis Research Company, Inc. (Wilmington, MA), a Sumitomo CKW-21 He compressor unit (Wilmington, MA), and an Ar-Kr proportional counter detector with the WissEl setup or a Ritverc (St. Petersburg, Russia) NaI detection system. 57Co/Rh sources (50-mCi to 75-mCi, initial strength) were used as the gamma energy sources. With the WissEl setups, the transmitted counts were stored in a multichannel scalar (MCS) mode as a function of energy (transducer velocity) using a 1024-channel analyzer. The data were folded to 512 channels to provide a flat background and a zero-velocity position corresponding to the

19 center shift (CS) of a metal Fe foil at room temperature (RT). Calibration spectra were obtained with a 25-µm -thick Fe foil (Amersham, United Kingdom) placed in the same position as the samples to minimize any geometry errors. The Mössbauer data were modeled with the RecoileTM software (University of Ottawa, Canada) using a Voigt- based structural fitting routine (Rancourt and Ping, 1991). Sample preparation (type of sample holder, etc.) was identical to the procedures reported previously (Peretyazhko et al., 2012). 3. RESULTS 3.1. Microbial oxidation of Fe(II) in bioreduced NAu-2 The oxidation of Fe(II) in bioreduced NAu-2 was coupled with reduction of nitrate under both growth and nongrowth conditions by strain 2002. In abiotic control, Fe(II) concentration slightly decreased within 20 days under both conditions. With phosphate, trace minerals, and vitamins in the growth medium as nutrients to promote cell growth, Fe(II) concentration decreased from 7.5 mM to 3.5 mM in the first 2.5 days (Fig. 1A). From 2.5 to 12 day, a small amount of nitrite accumulated from nitrate reduction (Fig. 1C), and this nitrite could have accelerated Fe(II) reduction. Our supplementary experiment confirmed that a certain amount of Fe(II) was indeed chemically oxidized by nitrite under the acidic and boiling condition that is required for the Fe(II) measurement (Amonette and Templeton, 1998). Therefore, the apparent dip in Fe(II) concentration from 2.5 to 12 day was ascribed to chemical oxidation by nitrite during the Fe(II) measurement. By day 17, the measured Fe(II) concentration (2.5 mM) corresponded to 67% extent of Fe(II) oxidation (Fig. 1A). Under the nongrowth condition, the maximum extent of Fe(II) oxidation (57%) was reached by 7.5 day (Fig. 1B) and chemical oxidation of Fe(II) by nitrite may be insignificant because of low nitrite concentration (Fig. 1D). Over the same time period, nitrate concentration in the growth and the nongrowth treatments decreased from 9 and 10 mM to nearly 0 mM, respectively (Fig. 1C and D). Lack of nitrate reduction in the abiotic control suggested that nitrate reduction in both treatments could not be caused by any remaining CN32 cells or any cell debris that may have been carried over from bioreduction. Nitrate concentration decreased at a slower rate in the growth treatment than in the nongrowth treatment.

A close examination of the reaction stoichiometry revealed that the amount of nitrate reduced under both growth and nongrowth conditions greatly exceeded the stoichiometric requirement for the amount of Fe(II) oxidized. According to the reaction stoichiometry (Eq. 1), 1 mM and 0.8 mM nitrate should have been consumed under the growth and nongrowth conditions, respectively (Growth condition: 5 mM Fe(II) oxidized; nongrowth condition: 4 mM Fe(II) oxidized). However, the measured consumption of nitrate was 9 mM and 10 mM, respectively. To explore the reasons for the observed non-stoichiometric ratio of Fe(II) oxidation over nitrate reduction, a supplementary experiment was set up to examine biological reduction of nitrate using either energy reserve within active cells (i.e., glycogen) or biomass of inactive cells. The results showed that a small amount of nitrate (9%) was reduced by inactive cell biomass during the tested concentration range (105 to 106 cells/mL). In the presence of active or active + inactive cells, nitrate reduction was observed (Fig. EA-2). When active cells were present alone, the extent of reduction was proportional to active cell concentration. When active and inactive cells were both present, the extent of reduction was positively correlated with inactive (electron donor) cell concentration when active cell concentration was kept constant (Fig. EA-2). A certain amount of transient nitrite accumulated under the growth condition, reaching up to 2 mM by day 12, but quickly decreased to 0 mM (Fig. 1C). Under the nongrowth condition, a smaller amount of nitrite (<0.8 mM) accumulated at the beginning (0-2.5 days), but again quickly decreased to 0 mM (Fig. 1D). Nitrogen gas (N2) was produced from nitrate-dependent Fe(II) oxidation. The amount of N2 produced from each inoculated serum bottle under the growth condition was 0.08 mmol (Fig. 1), which was approximately a half the amount of nitrate consumed (10 mmol/L nitrate concentration × 0.017 L total volume of each bottle = 0.17 mmol) according to the reaction (Eq. 1). However, under the nongrowth condition, the amount of N2 produced was not stoichiometric to the amount of nitrate consumed, suggesting a possible production of other products such as N2O that was not measured in this study. Ammonia was not produced from nitrate reduction, which was consistent with a previous study of using strain 2002 to oxidize aqueous Fe2+ (Weber et al., 2006b).

Over the course of the Fe(II) oxidation experiment, cell growth was monitored by measuring total protein concentration. As expected, a higher cell protein concentration was measured under the growth condition (Fig. 1E) than under the nongrowth condition (Fig. 1F). Moreover, under the growth condition, cell protein concentration increased rapidly at the beginning of the experiment, slightly decreased during the time of nitrite accumulation, then slowly increased until 17 day (Fig. 1E). Cell protein concentration under the nongrowth condition decreased at the beginning that corresponded to the period where a small amount of nitrite accumulated, and then remained at a low level throughout the duration of the experiment (Fig. 1F). Because nitrite is likely inhibitory to microbial activity (Bancroft et al., 1979), the apparent decrease in protein concentration as nitrite accumulated (Fig. 1E and F in comparison with Fig. 1C and D) may be due to cell death, nongrowth, or metabolic change associated with nitrite accumulation. Because of the possibility of Fe(II) oxidation by nitrite, a second supplementary experiment was conducted to test this chemical reaction. Our data showed that at neutral pH and room temperature, there was little abiotic oxidation of Fe(II) by nitrite under the non-growth condition, but there was a certain amount of Fe(II) oxidation under the growth condition (up to 30%). However, the chemical Fe(II) oxidation rate by nitrite was nine times slower than the biotic rate, e.g., 0.24 mmol L-1 day-1 vs. 2.09 mmol L-1 day-1, consistent with the observation by Shelobolina et al. (2012b). 3.2. Mineralogical changes associated with nitrate-dependent Fe(II) oxidation 3.2.1. XRD results Bioreduced NAu-2 showed two broad peaks corresponding to (001) and (002) reflections of nontronite (Fig. 2). Bio-oxidation of Fe(II) in reduced NAu-2 led to a broadened and shifted (001) peak with low intensity, suggesting some changes of the nontronite structure induced by strain 2002 cells. With further incubation under the growth condition, the spacing of the d(001) peak shifted from two-theta 6.5 ̊ (i.e., 15.01 Å) at 10 days to 5.5 ̊ (17.28 Å) at 60 days. The d(001) peak became slightly sharper with an extended time of Fe(II) oxidation (up to 60 days). Under the nongrowth condition, similar changes were observed. Mineralogical changes were compared between abiotic control (no strain 2002 cells) and biologically oxidized nontronite samples after 60 days of incubations (Fig. 3). Under

22 the growth condition, vivianite was present in abiotic control, but absent in biologically oxidized sample. As expected, vivianite was not detected in either abiotic or biotic samples under the nongrowth condition due to lack of phosphate. With extended incubation for 6 months, vivianite was detected again under the growth condition by XRD (Fig. 4). When structural Fe(II) in bioreduced NAu-2 was chemically oxidized by nitrite (i.e., supplementary experiment 2), no vivianite, ferrihydrite, or magnetite was detected under both growth and nongrowth conditions. 3.2.2. SEM and TEM observations SEM observations of the NAu-2 sample oxidized for 30 days under the growth condition revealed a number of cells in association with NAu-2 particles (Fig. 5A and B). NAu-2 particles were identified by their flaky and wavy shape and Si, Al, O, and Fe peaks from an EDS spectrum (Fig. 5C). This composition appeared to be different from the original NAu-2 (Keeling et al., 2000; Liu et al., 2011) which contained less Fe and some amounts of K and Ca. The presence of vivianite in abiotic control (e.g., bioreduced but not oxidized) and in the 6-month oxidized nontronite samples under the growth condition were confirmed by SEM observations (Fig. 6). TEM images showed that under the growth condition many strain 2002 cells became encrusted by solid materials after Fe(II) oxidation (Fig. 7A). Based on EDS analysis, the encrustations around the cells were mainly composed of O, P, Fe, Si, Na, K and Ca (Figs. 7B and C). Some peaks (such as Si and Na) may be from adjacent NAu-2 particles. A larger P peak in the thicker crust in comparison to the thinner crust suggested some Fe- and P-bearing materials in the crust. A trace amount of ferrihydrite was found under TEM in the NAu-2 sample oxidized for 60 days under the growth condition (Fig. 8). The identification of ferrihydrite was based on the amorphous nature of particles (Fig. 8A), concentric polycrystalline rings with characteristic d-spacings of 0.47 nm, 0.30 nm, 0.24 nm, and 0.19 nm for ferrihydrite (Fig. 8B), and an EDS spectrum showing Fe and O peaks only (Fig. 8C) (Kukkadapu et al., 2003). Two lines of the 6-line ferrihydrite were absent, possibly due to their low intensities. Presence of residual nontronite and newly formed magnetite were observed by TEM in the oxidized NAu-2 samples that were incubated for 6 months under the growth

23 condition. Indication of nontronite and magnetite were first suggested by characteristic morphologies (layers for nontronite and nano-particles for magnetite) (Fig. 9A and B). Qualitative EDS and SAED patterns further confirmed magnetite (Fig. 9C and D) and nontronite (Fig. 9F and G). An HRTEM lattice-fringe image of nontronite showed partially collapsed layer spacings of 1.09 nm and 1.23 nm (Fig. 9E). No illite was observed in either bioreduced or re-oxidized NAu-2 sample. Under the nongrowth condition, no vivianite was formed, but smaller amounts of magnetite and ferrihydrite were observed (relative to the growth condition). When Fe(II) was chemically oxidized by nitrite under the growth condition, a small amount of vivianite was present under TEM but no magnetite or ferrihydrite. Chemical Fe(II) oxidation under the nongrowth condition resulted in similar mineralogy but no vivianite. 3.2.3. Mössbauer spectroscopy Structural Fe(III) in the pristine nontronite (i.e., unreduced) occurred in two octahedral Fe(III)-oct environments (Fig. 10A), as consistent with our previous study (Yang et al., 2012). Liquid nitrogen (77 K) Mössbauer spectra confirmed that the percent of Fe(II) (out of total solid Fe) was 31% after the initial reduction experiment (Fig. 10B) but decreased to 18% after re-oxidation for 10 days under the growth condition (Fig. 10C), which were generally consistent with the relative proportion of Fe(III) and Fe(II) determined by the chemical method (31% and 13%, respectively). Spectral modeling further indicated that both the Fe(III) sites were reduced, albeit to different extents extents, as noted in our previous studies (Jaisi et al., 2005;Yang et al., 2012). Reoxidation partially restored the structure, which was evident from the Fe(III) Mössbauer spectral parameters that are similar to those for the pristine (Table 1) and completely air-oxidized samples (Yang et al., 2012). 4. DISCUSSION 4.1. Microbial oxidation of structural Fe(II) in reduced NAu-2 coupled with nitrate reduction Our data unambiguously demonstrated that Pseudogulbenkiania sp. strain 2002 was able to couple oxidation of structural Fe(II) in NAu-2 with reduction of nitrate to N2 with

24 nitrite as a transient intermediate. In this process, strain 2002 was able to gain energy for growth with no need for organic carbon such as acetate (Weber et al., 2006b, 2009). The stoichiometry (molar ratio) of the amount of nitrate reduced to the amount of Fe(II) oxidized under both growth (2:1) and nongrowth conditions (8.5:3.5) greatly exceeded the theoretical values according to Eq. 1 (1:5). Similar observations were made for nitrate-dependent microbial oxidation of Fe(II) in solid minerals (Benz et al., 1998; Weber et al., 2001). Nitrate loss by sorption to NAu-2 and cell surface was unlikely because the ratio of the amount of N2 produced to the amount of nitrate consumed was close to the stoichiometric ratio for the growth experiment. One speculation is that some other N species may have been produced from nitrate reduction (Straub et al., 1996; Weber et al., 2001). An alternative explanation is heterotrophic nitrate reduction coupled with oxidation of dead bacterial biomass (Weber et al., 2001). This explanation is supported by our supplementary experiment, where our data showed nitrate reduction. These data suggest that chemical reduction of nitrate by cell components from inactive cells was possible (only up to ~9%), and heterotrophic nitrate reduction was mostly enzymatic. When active cells were present alone, nitrate reduction could have been achieved using stored energy reserves (i.e., glycogen). When active and inactive cells were added together, nitrate reduction could have happened using both stored energy reserve from active cells and released organic matter from inactive cells. Either scenario could have resulted in a greater ratio of nitrate reduction to Fe(II) oxidation than the stoichiometric requirement. Different levels of nitrite were accumulated under the growth (up to 2 mM) and the nongrowth (up to 1 mM) conditions. Under the growth condition, our observed nitrite level was much higher than that observed by Weber et al. (2006b) (i.e., 2 mM vs. <0.05 mM). Under the nongrowth condition, the amount of nitrite accumulation observed in this study was similar to that observed by Weber et al. (2006b). A main difference between the growth and nongrowth experiments was the nutrient level. Under the nutrient-limited nongrowth condition, a large fraction of inoculated cells may have experienced cell lysis/decomposition which would result in a lower biomass (Fig. 1F). The released organic matter from lysed/decomposed cell biomass may have served as electron donor for nitrate reduction. In this case, organic matter would outcompete Fe(II) as electron

25 donor to reduce nitrate. The redox couple“CO2/organic matter” has a low reduction potential of -290 mV (Kappler et al., 2005). In contrast, under the growth condition, cell biomass increased over time, suggesting that cell lysis/decomposition was unlikely, and little organic matter release was expected. In this case, Fe(II), not organic matter, may have been the main electron donor to reduce 2+ nitrate. The reduction potential for the redox couple “Fe(OH)3/Fe ” was much higher than the “CO2/organic matter” couple (Kappler et al., 2005), approximately 0 mV (Thauer et al., 1977). This difference in the system reduction potential between the growth and non-growth conditions may have accounted for their difference in nitrite accumulation. Kappler et al. (2005) observed nitrite formation during nitrate-dependent Fe(II) oxidation but not during nitrate-dependent acetate oxidation and suggested that entry of electrons at different redox potentials could change the relative rates of the enzymatically catalyzed reductions of different nitrogen species. Thus, our results are consistent with those by Kappler et al. (2005) in showing lower nitrite accumulation under the nongrowth condition. It is possible that the lower redox potential under the nongrowth condition could have accelerated reduction rates of nitrate by circumventing the nitrite accumulation as intermediate and resulting in direct N2 production. 4.2. Abiotically oxidation of reduced NAu-2 by nitrite The amounts of nitrite accumulated under the growth (2 mM) and the nongrowth (up to 1 mM) conditions could have chemically reacted with Fe(II) in the reduced NAu-2 and resulted an enhancement of Fe(II) oxidation. Thermodynamically, the chemical oxidation of Fe2+ by nitrite reduction is favorable (Picardal, 2012). Indeed, previous studies have shown enhanced extents of chemical oxidation of solid-phase Fe(II) (Coby and Picardal, 2005) due to its reaction with nitrite (Weber et al., 2001; Rakshit et al., 2005) under neutral pH. However, our supplementary experimental results showed no chemical oxidation of structural Fe(II) in nontronite by nitrite under the nongrowth condition. A small amount of chemical Fe(II) oxidation by nitrite under the growth condition was likely due to the presence of vivianite, consistent with a previous study (Miot et al., 2009a). In contrast, under the nongrowth condition, there was no vivianite and structural Fe(II) in NAu-2 was not easily oxidized. However, in our biotic experiments, there was no vivianite formation during the active Fe(II) oxidation period (0-17 day and 0-7.5 days

26 under the growth and nongrowth condition, respectively) and therefore chemical Fe(II) oxidation by nitrite was deemed unlikely. Therefore, we concluded that the observed Fe(II) oxidation under both growth and nongrowth conditions should be largely biological. A similar observation was made by Shelobolina et al. (2012b) who observed a negligible amount of chemical Fe(II) oxidation by nitrite, and the chemical oxidation rate was three to six times lower than the biotic oxidation rate. Although chemical Fe(II) reduction by nitrite was insignificant relative to biotic Fe(II) oxidation at neutral pH and room temperature, our results showed that chemical Fe(II) oxidation rate at acidic pH and boiling temperature, a condition required for Fe(II) measurement, was much higher. This observation was consistent with previous studies (Picardal, 2012; Klueglein and Kappler, 2013) and call for caution. This artificial enhancement of Fe(II) oxidation by nitrite could have accounted for a rapid Fe(II) oxidation at 12 day under the growth condition and at 2.5 day under the nongrowth condition. Fortunately, because of small amounts of nitrite and its rapid consumption, this effect on the overall extent of Fe(II) oxidation should be small. However, future studies call for a modified Fe(II) measurement method, where sulfamic acid can be used (Klueglein and Kappler, 2013) in place of sulfuric acid in the 1,10-phenanthroline method. Our additional experiments have confirmed the validity of this modified method in measuring Fe(II) concentration in the presence of nitrite (Zhao et al., unpublished data). 4.3. Mineralogical changes associated with nitrate-dependent microbial Fe(II) oxidation A series of mineralogical changes occurred following the sequence of original NAu-2 → bioreduction of NAu-2 → Re-oxidation. In our experiment, Fe(II)-oxidizing bacteria partially altered the NAu-2 structure, as evidenced by the decreased intensity and broadening of d(001) peak of inoculated NAu-2 in the first 10 days of Fe(II) oxidation. However, the width of d(001) peak decreased from 10 to 60 days suggesting oxidative removal of poorly-crystalline/smaller NAu-2 particles. A similar observation was observed previously, where the d(001) peak became sharper after three cycles of biotic reduction-abiotic oxidation of NAu-2 (Yang et al., 2012). Following bioreduction, a certain fraction of NAu-2 apparently dissolved and released Fe2+ from its structure, consistent with previous observations in our laboratory (Jaisi et al., 2008a). This released Fe2+ could be aqueous or sorbed onto residual NAu-2

27 particle surfaces. Regardless of which form, this Fe2+ should be available to react with 3- PO4 in the growth medium to precipitate vivianite (Fig. 3). Vivianite is a typical mineral formed as a result of bioreduction of structural Fe(III) in smectites (Dong et al., 2003a) in phosphate-bearing medium. However, vivianite was only observed in abiotic control (i.e., reduced NAu-2), but not in the sample reoxidized for 60 days. This observation suggests that in the biotic treatment vivianite was either reoxidized preferentially relative to structural Fe(II) in NAu-2 after its formation or never formed, perhaps because aqueous or surface-sorbed Fe2+ could be rapidly oxidized by strain 2002 cells or removal of phosphate from aqueous solution by bacterial uptake. It has been observed that, in presence of Fe(II)-oxidizing cells, vivianite could be oxidized by nitrite that is produced by Fe(II)-oxidizing cells (Miot et al., 2009b). The data obtained in this study do not allow differentiation between these two possible mechanisms for the absence of vivianite in biotic treatments. Under the growth condition, oxidation of vivianite and aqueous or sorbed Fe2+ led to the formation of ferrihydrite and magnetite as reaction products from nitrate–dependent Fe(II) oxidation, as evidenced by XRD and TEM data (Figs. 8 and 9). Greenish gray mixed valence Fe(II)–Fe(III) hydroxides, known as green rust, was observed in previous microbial Fe(II) oxidation experiment (Chaudhuri et al., 2001), but it was not observed in our study. The reason may be related to the rate of Fe(II) oxidation. The initial rate of Fe(II) oxidation under both the growth and nongrowth conditions was approximately 4 mg L-1 h-1. However, in the Chaudhuri’s experiment (2001), the Fe(II) oxidation rate was 2.6 mg L-1 h-1. A previous study (Lack et al., 2002) observed that rapid chemical Fe(II) oxidation caused ferrihydrite formation while low rates led to green rust formation. The observation of ferrihydrite in the present investigation versus green rust formation in the previous report (Chaudhuri et al., 2001) is apparently consistent with this interpretation. Ferrihydrite is an unstable phase and is believed to be a precursor of magnetite in the presence of aqueous Fe2+ (Hansel et al., 2005). Previous studies on nitrate-dependent Fe(II) oxidation reported that ferrihydrite (Straub et al., 1996) or green rust (Lack et al., 2002) usually leads to more crystalline iron oxide formation with extended incubation time. Indeed, magnetite was observed in our samples that were inoculated for 6 months

28 under both the growth and nongrowth conditions. The apparent lack of ferryhydrite in our XRD or Mössbauer analysis is presumably due to its low abundance. Vivianite formed again in the growth medium following an extended incubation time (6 months) (Fig. 6). This vivianite could have formed from Fe(II) in nontronite that was not oxidized. Although the exact pathway is not certain, it is possible that structural Fe(II) in residual nontronite (~33% Fe(II) remaining at the end of oxidation under the growth condition) could have reacted with Fe(III) in ferrihydrite, magnetite, or any surface adsorbed Fe(III) via intervalence electron transfer to form Fe(II), which would react with aqueous phosphate to form vivianite. Inter-valence electron transfer has been observed between structural Fe(III) in nontronite and surface-adsorbed Fe(II) (Schaefer et al., 2011), apparently the reverse reaction, i.e., structural Fe(II) and solid-phase or adsorbed Fe(III) may be possible as well. Under the nongrowth condition no vivianite formation was expected because of lack of phosphate. Lower abundances of ferrihydrite and magnetite under the nongrowth condition may be due to the lower extent of Fe(II) oxidation relative to the growth condition (e.g., 57% vs. 67%). Different from both growth and nongrowth biotic experiments, in chemically oxidized nontronite, no formation of ferrihydrite and magnetite, but emergence of unidentified peaks on XRD pattern suggest different mineralogical transformations, but definitive evidence must await further characterization. The differences in mineral transformations between biotic (this study) and abiotic (Yang et al., 2012) Fe(II) oxidation by air following the initial bioreduction by CN32 cells are also noteworthy. In the investigation involving abiotic Fe(II) oxidation (Yang et al., 2012), no secondary minerals such as vivianite, ferrihydrite, and magnetite were observed. A comparison between these two studies highlights the role of microorganisms involved in the oxidation process. The role of microorganisms in the secondary mineral transformation is likely achieved through their influences on the extent and rate of Fe(II) oxidation and/or solution chemistry. 4.4. Mechanisms of Fe(II) oxidation in reduced NAu-2 Three possible mechanisms have been proposed for the nitrate-dependent oxidation of structural Fe(II) in clay minerals (Shelobolina et al., 2012a), as follows: (1) direct electron transfer from structural Fe(II) in nontronite to nitrate in solution with or without

29 mediation by an extracellular enzyme; (2) indirect electron transfer involving aqueous phase Fe3+ in two steps: initial dissolution of structural Fe(II) in nontronite followed by oxidation of aqueous Fe2+ to aqueous Fe3+ by nitrate. Once a certain amount of aqueous Fe3+ forms, electron transfer between aqueous Fe3+ and structural Fe(II) in nontronite could possibly continue; (3) indirect electron transfer from structural Fe(II) to nitrate, mediated by surface-sorbed or solid Fe(III). The two ‘indirect’ electron-transfer mechanisms are similar; in the third mechanism the electron-transfer from structural Fe(II) in nontronite to nitrate is through surface-sorbed Fe(III) (either as an ion or solid- phase Fe(III)). Previous studies have shown inter-valence electron transfer from structural Fe(III) in oxides (Williams and Scherer, 2004) or nontronite (Schaefer et al., 2011) to surface- adsorbed Fe(II), it is possible that the reverse reaction, i.e., from structural Fe(II) in nontronite to surface-adsorbed Fe3+ ions or solid-phase Fe(III) is also likely. Aqueous chemistry and mineralogical data support the second and third mechanisms. The formation of vivianite in bioreduced NAu-2, but not in the abiotic control (Fig. 3) suggests that a certain amount of NAu-2 may have been reductively dissolved releasing Fe2+ after bioreduction. If the released Fe2+ was rapidly oxidized to Fe3+ by nitrate in aqueous solution, the aqueous Fe3+ could in turn oxidize structural Fe(II) in NAu-2, and thus would support the second mechanism. If the released Fe2+ was not rapidly oxidized 3+ 3- to Fe , but instead it combined with PO4 to form vivianite This vivianite was later oxidized by strain 2002 cells to form ferrihydrite and magnetite (Figs. 8 and 9). Given the experimental evidence for this mineralogical transformation sequence the third mechanism seems more likely. Ferrihydrite and magnetite could have served as solid phase Fe(III) oxidant to oxidize structural Fe(II) in NAu-2. The observation that vivianite formed over an extended period of incubation time (6 months) (Figs. 4 and 6) supports this mechanism, i.e., structural Fe(II) in residual nontronite at the end of oxidation experiment continued to transfer electrons to structural Fe(III) in ferrihydrite and magnetite to form Fe(II) which then could combine with phosphate to form vivianite. Because electron transfer between solid Fe(III) and solid Fe(II) is difficult to proceed, it is likely that this transfer may have only taken place with sorbed phase Fe(III) at the interface between ferrihydrite/magnetite coating and underlying NAu-2.

If the third mechanism were true, the cell crust material (Fig. 7A) around cell surface would be of benefit to the electron transfer process. If the crust material is made of Fe(III), this Fe(III) could facilitate electron transfer from structural Fe(II) of NAu-2 to nitrate in solution by serving as a solid-state electron shuttle. That is, electrons from structural Fe(II) within NAu-2 could be used to first reduce Fe(III) in this material to Fe(II); the resulting Fe(II) then delivers electrons to nitrate. Indeed, a beneficial role of this cell- encrusting material was proposed by Kappler et al. (2005) based on their observations of no obvious hazardous effects of this material to strain BoFeN1, and a more favorable thermodynamic calculation of Fe3+ in close proximity to cells than aqueous Fe2+. Furthermore, Emerson and Revsbech (1994) and Hallberg and Ferris (2004) observed a fibrous extracellular material excreted by cells. This material not only served as nucleation sites for the formation of iron oxides, but also were able to help transfer electron from Fe(II) to cells. Whereas the observations of Kappler et al. (2005), Emerson and Revsbech (1994) and Hallberg and Ferris (2004) strengthened our explanation, there was opposing evidence as well, where aerobic respiration rate decreased when Acinetobacter spp. were encrusted by iron (Macrae and Celo, 1975). 4.5. Nitrate removal with redox cycling of Fe in clay minerals Microbial oxidation of Fe2+ in aqueous (Weber et al., 2006b), oxide (Chaudhuri et al., 2001), and clay-associated (Weber et al., 2001; Shelobolina et al., 2012b) form coupled with nitrate reduction provides a prospect of applying N–Fe redox cycles to remove nitrate contaminant in soils and groundwater systems. The biological nitrate-dependent Fe(II) oxidation has been demonstrated in a wide variety of natural systems, many of which are in freshwater sediments (Straub et al., 1996; Chaudhuri et al., 2001; Shelobolina et al., 2003a; Weber et al., 2009). Our evidence showed that, when using Fe(II) in clay mineral nontronite as an electron donor, the oxidized Fe, i.e., Fe(III), may have largely remained in the structure of nontronite with only a small amount of Fe(III) oxide formation. Unlike iron oxides, this solid-state Fe redox cycling in clay minerals offers an interesting possibility that this pool of Fe may be a renewable source for continued removal of nitrate. For example, Fe(III) in NAu-2 can be re-reduced by Fe- reducing bacteria with addition of organic carbon (possibly from organic contaminants). A potential advantage of using clay minerals and Fe redox cycling microorganisms is to

31 continually remove nitrate with no hazadous byproducts. Coby et al. (2011) have shown the feasibility of applying microflora from freshwater river floodplain sediments amended with synthetic Fe(III) oxide (nanocrystalline goethite) to remove nitrate and acetate alternately. Additional experiments such as repeated cycles of Fe(II) oxidation and Fe(III) reduction of clay minerals would further address the issue of renewably removing nitrate. In addition to nitrate, the fate and mobility of heavy metals may be influenced by Fe redox cycling of clay minerals as well. For example, Hohmann et al. (2010) and Bishop et al. (2011) demonstrated that mineral transformations associated with Fe redox cycling has important implications for the mobility of toxic metals, such as As and Cr, due to precipitation and adsorption mechanisms. Fe redox cycles could be manipulated to gain maximal benefits in environmental remediation and site cleanup. Moreover, microbial Fe(II) oxidation not only plays an important role in the modern Fe, N, C and metal biogeochemical cycles, but may have been an important process in the early Earth’s history when there was no gaseous O2 in the atmosphere. It is increasingly recognized that the formation of the banded iron formation is heavily mediated by anaerobic Fe(II) oxidizing bacteria (Konhauser et al., 2011). However, it remains unclear if clay minerals were involved in this process. Our study implies that clay minerals may have been important players in nitrate-dependent, microbial Fe(II) oxidation process in an ancient, but important geological process. 5. CONCLUSIONS This comprehensive investigation has demonstrated that Pseudogulbenkiania sp. strain 2002 was able to couple oxidation of Fe(II) in reduced NAu-2 with reduction of nitrate to N2, via an intermediate product nitrite. The Fe(II) oxidation occurred largely in solid state with only a small amount of ferrihydrite and magnetite formation. The ratio of nitrate reduced to Fe(II) oxidized was in excess of the stoichiometric ratio, and direct biological reduction by strain 2002 cells was the likely reason. The chemical data, Mössbauer spectroscopy, and SEM/TEM observations suggested that the mechanisms for the nitrate-dependant oxidation of structural Fe(II) in clay mineral nontronite were complex, involving both direct (possibly enzymatic oxidation) and an indirect oxidation through surface-sorbed or solid Fe(III). The results of this study highlight the importance

32 of iron-bearing clay minerals in the global nitrogen cycle with potential applications for nitrate removal in soils and groundwater. ACKNOWLEDGMENTS The work was supported by a grant from National Science Foundation (EAR 1148039). A portion of the research was performed using EMSL, a national scientific user facility sponsored by the Department of Energy’s Office of Biological and Environmental Research located at Pacific Northwest National Laboratory. The JEOL 2100 TEM used in this study was supported by NSF grant EAR-0722807. We are grateful to four anonymous reviewers whose comments significantly improved the quality of the manuscript.

REFERENCES Amonette, J. E. and Templeton, J. C., 1998. Improvements to the quantitative assay of nonrefractory minerals for Fe(II) and total Fe using 1,10-phenanthroline. Clays Clay Miner. 46, 51-62.

Bancroft, K., Grant, I. F., and Alexander, M., 1979. Toxicity of NO2: Effect of Nitrite on Microbial Activity in an Acid Soil. Appl. Environ. Microbiol. 38, 940-944.

Benz, M., Brune, A., and Schink, B., 1998. Anaerobic and aerobic oxidation of ferrous iron at neutral pH by chemoheterotrophic nitrate-reducing bacteria. Arch. Microbiol. 169, 159-165.

Bishop, M. E., Dong, H. L., Kukkadapu, R. K., Liu, C. X., and Edelmann, R. E., 2011. Bioreduction of Fe-bearing clay minerals and their reactivity toward pertechnetate (Tc-99). Geochim. Cosmochim. Acta 75, 5229-5246.

Blothe, M. and Roden, E. E., 2009. Microbial Iron Redox Cycling in a Circumneutral-pH Groundwater Seep. Appl. Environ. Microbiol. 75, 468-473.

Bradford, M.M. 1976. A rapid and sensitive method for the quantization of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem, 1/2, 248-254.

Byrne-Bailey, K. G., Weber, K. A., Chair, A. H., Bose, S., Knox, T., Spanbauer, T. L., Chertkov, O., and Coates, J. D., 2011. Completed Genome Sequence of the Anaerobic Iron-Oxidizing Bacterium Acidovorax ebreus Strain TPSY. J.Bacteriol. 192, 1475-1476.

Canfield, D. E., Glazer, A. N., and Falkowski, P. G., 2010. The evolution and future of earth's nitrogen cycle. Science 330, 192-196.

Chakraborty, A., Roden, E. E., Schieber, J., and Picardal, F., 2011. Enhanced Growth of Acidovorax sp Strain 2AN during Nitrate-Dependent Fe(II) Oxidation in Batch and Continuous-Flow Systems. Appl. Environ. Microbiol. 77, 8548-8556.

Chaudhuri, S. K., Lack, J. G., and Coates, J. D., 2001. Biogenic Magnetite Formation through Anaerobic Biooxidation of Fe(II). Appl. Environ. Microbiol. 67, 2844- 2848.

Coby, A. J., Picardal, F., Shelobolina, E., Xu, H. F., and Roden, E. E., 2011. Repeated Anaerobic Microbial Redox Cycling of Iron. Appl. Environ. Microbiol. 77, 6036- 6042.

Coby, A. J. and Picardal, F. W., 2005. Inhibition of NO3- and NO2- reduction by microbial Fe(III) reduction: Evidence of a reaction between NO2- and cell surface-bound Fe2+. Appl. Environ. Microbiol. 71, 5267-5274.

Dong, H., Kostka, J. E., and Kim, J., 2003a. Microscopic evidence for microbial dissolution of smectite. Clays Clay Miner. 51, 502-512.

Dong, H., Kukkadapu, R. K., Fredrickson, J. K., Zachara, J. M., Kennedy, D. W., and Kostandarithes, H. M., 2003b. Microbial Reduction of Structural Fe(III) in Illite and Goethite. Environ. Sci. Technol. 37, 1268-1276.

Dong, H. L., Jaisi, D. P., Kim, J., and Zhang, G. X., 2009. Microbe-clay mineral interactions. Am. Mineral. 94, 1505-1519.

Dong, H. L., 2012. Clay-Microbe Interactions and Implications for Environmental Mitigation. Elements 8, 113-118.

Edwards, K. J., Bach, W., and Rogers, D. R., 2003. Geomicrobiology of the ocean crust: A role for chemoautotrophic Fe-bacteria. Biol. Bull. 204, 180-185.

Emerson, D. and Revsbech, N. P., 1994. Investigation of an Iron-Oxidizing Microbial Mat Community Located near Aarhus, Denmark - Field Studies. Appl. Environ. Microbiol. 60, 4022-4031.

Ernstsen, V., 1996. Reduction of nitrate by Fe2+ in clay minerals. Clays Clay Miner. 44, 599-608.

Ernstsen, V., Gates, W. P., and Stucki, J. W., 1998. Microbial reduction of structural iron in clays - A renewable source of reduction capacity. J Environ Qual 27, 761-766.

Evgenya, S., Hiromi, K., Huifang, X., Jason, B., Mai Yia, X., Tao, W., Marco, B., and Eric, R., 2012. Isolation of Phyllosilicate–Iron Redox Cycling Microorganisms from an Illite–Smectite Rich Hydromorphic Soil. Front. Microbiol. 3.

Gates, W. P., Slade, P. G., Manceau, A., and Lanson, B., 2002. Site occupancies by iron in nontronites. Clays Clay Miner. 50, 223-239.

Hafenbradl, D., Keller, M., Dirmeier, R., Rachel, R., RoÃŸnagel, P., Burggraf, S., Huber, H., and Stetter, K. O., 1996. Ferroglobus placidus gen. nov., sp. nov., a novel hyperthermophilic archaeum that oxidizes Fe2+ at neutral pH under anoxic conditions. Arch. Microbiol. 166, 308-314.

Hallberg, R. and Ferris, F. G., 2004. Biomineralization by Gallionella. Geomicrobiol. J. 21, 325-330.

Hansel, C. M., Benner, S. G., and Fendorf, S., 2005. Competing Fe(II)-Induced Mineralization Pathways of Ferrihydrite. Environ. Sci. Technol. 39, 7147-7153.

Hohmann, C., Winkler, E., Morin, G., and Kappler, A., 2010. Anaerobic Fe(II)-Oxidizing Bacteria Show As Resistance and Immobilize As during Fe(III) Mineral Precipitation. Environ. Sci. Technol. 44, 94-101.

Holloway, J. M., Dahlgren, R. A., Hansen, B., and Casey, W. H., 1998. Contribution of bedrock nitrogen to high nitrate concentrations in stream water. Nature 395, 785- 788.

Jaisi, D. P., Kukkadapu, R. K., Eberl, D. D., and Dong, H. L., 2005. Control of Fe(III) site occupancy on the rate and extent of microbial reduction of Fe(III) in nontronite. Geochim. Cosmochim. Acta 69, 5429-5440.

Jaisi, D. P., Dong, H. L., and Liu, C. X., 2007b. Influence of biogenic Fe(II) on the extent of microbial reduction of Fe(III) in clay minerals nontronite, illite, and chlorite. Geochim. Cosmochim. Acta. 71, 1145-1158.

Jaisi, D. P., Liu, C. X., Dong, H. L., Blake, R. E., and Fein, J. B., 2008a. Fe(2+) sorption onto nontronite (NAu-2). Geochim. Cosmochim. Acta 72, 5361-5371.

Jaisi, D.P., Dong, H., and Morton, J.P. 2008b. Partitioning of Fe(II) in reduced nontronite (NAu-2) to reactive sites: reactivity in terms of Tc(VII) reduction. Clays Clay Miner. 56, 175-189.

Jaisi, D. P., Dong, H. L., Plymale, A. E., Fredrickson, J. K., Zachara, J. M., Heald, S., and Liu, C. X., 2009. Reduction and long-term immobilization of technetium by Fe(II) associated with clay mineral nontronite. Chem. Geol. 264, 127-138.

Kappler, A., Schink, B., and Newman, D. K., 2005. Fe(III) mineral formation and cell encrustation by the nitrate-dependent Fe(II)-oxidizer strain BoFeN1. Geobiology 3, 235-245.

Keeling, J. L., Raven, M. D., and Gates, W. P., 2000. Geology and characterization of two hydrothermal nontronites from weathered metamorphic rocks at the Uley Graphite Mine, South Australia. Clay Clay Miner. 48, 537-548.

Klueglein, N. and Kappler, A. 2013. Abiotic oxidation of Fe(II) by reactive nitrogen species in cultures of the nitrate-reducing Fe(II) oxidizer Acidovorax sp BoFeN1- quesitoning the existence of enzymatic Fe(II) oxidation. Geobiology, 11, 180-190.

Konhauser, K. O., Hamade, T., Raiswell, R., Morris, R. C., Ferris, F. G., Southam, G., and Canfield, D. E., 2002. Could bacteria have formed the Precambrian banded iron formations? Geology 30, 1079-1082.

Konhauser, K. O., Kappler, A., and Roden, E. E., 2011. Iron in Microbial Metabolisms. Elements 7, 89-93.

Kukkadapu, R. K., Zachara, J. M., Fredrickson, J. K., Smith, S. C., Dohnalkova, A. C., and Russell, C. K., 2003. Transformation of 2-line ferrihydrite to 6-line ferrihydrite under oxic and anoxic conditions. Am. Mineral. 88, 1903-1914.

Lack, J. G., Chaudhuri, S. K., Chakraborty, R., Achenbach, L. A., and Coates, J. D., 2002. Anaerobic biooxidation of Fe(II) by Dechlorosoma suillum. Microb. Ecol. 43, 424-431.

Liu, D., Dong, H. L., Bishop, M. E., Wang, H. M., Agrawal, A., Tritschler, S., Eberl, D. D., and Xie, S. C., 2011. Reduction of structural Fe(III) in nontronite by methanogen Methanosarcina barkeri. Geochim. Cosmochim. Acta 75, 1057-1071.

Macrae, I. C. and Celo, J. S., 1975. Influence of Colloidal Iron on Respiration of a Species of Genus Acinetobacter. Appl. Microbiol. 29, 837-840.

Moore., D. M. and Robert C. Reynolds, J., 1997. X-Ray Diffraction and the Identification and Analysis of Clay Minerals. Oxford University Press, USA, 1997, New York.

Pantke, C., Obst, M., Benzerara, K., Morin, G., Ona-Nguema, G., Dippon, U., and Kappler, A., 2011. Green Rust Formation during Fe(II) Oxidation by the Nitrate-

Reducing Acidovorax sp. Strain BoFeN1. Environme Sci. Technol. 46, 1439- 1446.

Peretyazhko, T. S., Zachara, J. M., Kukkadapu, R. K., Heald, S. M., Kutnyakov, I. V., Resch, C. T., Arey, B. W., Wang, C. M., Kovarik, L., Phillips, J. L., and Moore, − D. A., 2012. Pertechnetate (TcO4 ) reduction by reactive ferrous iron forms in naturally anoxic, redox transition zone sediments from the Hanford Site, USA. Geochim. Cosmochim. Acta 92, 48-66.

Picardal, F., 2012. Abiotic and microbial interactions during anaerobic transformations of Fe(II) and NOx. Front. Microbiol. 3.

Picardal, F. W., Zaybak, Z., Chakraborty, A., Schieber, J., and Szewzyk, U., 2011. Microaerophilic, Fe(II)-dependent growth and Fe(II) oxidation by a Dechlorospirillum species. FEMS Microbiol. Lett. 319, 51-57.

Rakshit, S., Matocha, C. J., and Haszler, G. R., 2005. Nitrate reduction in the presence of wüstite. J. Environ. Qual. 34, 1286-1292.

Rancourt, D. G. and Ping, J. Y., 1991. Voigt-based methods for arbitrary-shape static hyperfine parameter distributions in Mössbauer spectroscopy. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms 58, 85-97.

Revanasiddappa, H. D. and Kumar, T. N. K., 2003. Spectrophotometric determination of trace amounts of nitrites in water and soil samples. Chem. Anal-Warsaw 48, 759- 765.

Schaefer, M.V., Gorski, C.A., and Scherer, M.M. 2011. Spectroscopic evidence for interfacial Fe(II)-Fe(III) electron transfer in a clay mineral. Environ. Sci. Technol., 45, 540-545.

Senko, J. M., Dewers, T. A., and Krumholz, L. R., 2005. Effect of oxidation rate and Fe(II) state on microbial nitrate-dependent Fe(III) mineral formation. Appl. Environ. Microbiol. 71, 7172-7177. 39

Shelobolina, E. S., Vanpraagh, C. G., and Lovley, D. R., 2003b. Use of ferric and ferrous iron containing minerals for respiration by Desulfitobacterium frappieri. Geomicrobiol. J. 20, 143-156.

Straub, K. L., Benz, M., Schink, B., and Widdel, F., 1996. Anaerobic, nitrate-dependent microbial oxidation of ferrous iron. Appl. Environ. Microbiol. 62, 1458-1460.

Stucki, J. W. and Roth, C. B., 1977. Oxidation-Reduction Mechanism for Structural Iron in Nontronite. Soil Sci. Soc. Amer. J 41, 808-814.

Violante, A. and Pigna, M., 2002. Competitive sorption of arsenate and phosphate on different clay minerals and soils. Soil Sci. Soc. Amer. J. 66, 1788-1796.

Weber, K. A., Picardal, F. W., and Roden, E. E., 2001. Microbially catalyzed nitrate- dependent oxidation of biogenic solid-phase Fe(II) compounds. Environ. Sci. Technol. 35, 1644-1650.

Weber, K. A., Achenbach, L. A., and Coates, J. D., 2006a. Microorganisms pumping iron: anaerobic microbial iron oxidation and reduction. Nat. Rev. Microbiol. 4, 752-764.

Weber, K.A., Urrutia, M.M., Churchill, P.F., Kukkadapu, R.K., Roden, E.E., 2006c. Anaerobic redox cycling of iron by freshwater sediment microorganisms. Environ. Microbiol. 8, 100-113.

Williams, A. G. B. and Scherer, M. M., 2004. Spectroscopic Evidence for Fe(II)−Fe(III) Electron Transfer at the Iron Oxide−Water Interface. Environ. Sci. & Technol. 38, 4782-4790.

Yang, J., Kukkadapu, R. K., Dong, H., Shelobolina, E. S., Zhang, J., and Kim, J., 2012. Effects of redox cycling of iron in nontronite on reduction of technetium. Chem. Geol.291, 206-216.

Yang, J. E., Skogley, E. O., Schaff, B. E., and Kim, J. J., 1998. A simple spectrophotometric determination of nitrate in water, resin, and soil extracts. Soil Sci. Soc. Amer. J. 62, 1108-1115.

Zhang, G. X., Burgos, W. D., Senko, J. M., Bishop, M. E., Dong, H. L., Boyanov, M. I., and Kemner, K. M., 2011. Microbial reduction of chlorite and uranium followed by air oxidation. Chem. Geol. 283, 242-250.

Table 1. 77 K Fitting and calculated Mossbauer parameters

______2 1 2 3 4 (5) 6 7 8 9 Sample Phase o or o  (mm/s) <> or <> sd <> or Phase, %

mm/s mm/s or H (Tesla) mm/s mm/s mm/s ______Pristine [Fe(III)-oct]-A 0.52 1.24 0.2 1.4 0.52 1.24 0.21 30 [Fe(III)-oct]-B 0.47 0.3 0.39 0.47 0.4 0.29 70

Bioreduced Fe(III)-A 0.48 1.24 0.18* 0.8 0.48 1.24 0.18 6 Fe(III)-B 0.48 0.45 0.26* 0.48 0.46 0.24 64 Fe(II) 1.23 2.73 0.28 1.23 2.73 0.28 31

Reoxidized Fe(III)-A 0.5 1.2 0.27 0.6 0.5 1.2 0.27 17 Fe(III)-B 0.48 0.39 0.24 0.48 0.4 0.22 66 Fe(II) 1.22 2.67 0.3 1.22 2.67 0.3 18 1Spectral component; 2center shift; 3quadrupole splitting or quadrupole shift ; 4std dev of the componet; 5reduced chi square 6average center shift;7average quadrupole or average quadrupole shift; 8standard deviation;9spectral percent; * value frozen for modeling: Lorentzian half widths at half maximum (HWHM) of all elemental Lorentzians in all elemetal doublet and sextets were 0.097 mm/sec; No coupling was allowed between CS, QS or e and average Bhf; the A+/A- areas of doublet are fixed at 1; A1/A3 and A2/A3 areas ae fixed at 2 and 3

FIGURE CAPTION

Figure 1. Time-course changes of concentrations of Fe(II) in reduced NAu-2, nitrate, nitrite, protein (a proxy for biomass), and N2 produced from the nitrate-dependent Fe(II) oxidation using strain 2002. Figures on the left (A, C, E) represent the growth condition, and the three figures on the right (B, D, F) represent the nongrowth condition.

Figure 2. X-ray diffraction patterns for bioreduced NAu-2 sample (i.e., onset of bio- oxidation) and for those oxidized by strain 2002 for various times (10, 30 and 60 days) under the growth (A) and the nongrowth conditions (B).

Figure 3. Comparison of XRD patterns between abiotic control (e.g., bioreduced by strain

CN32 cells but never oxidized by strain 2002 cells) and microbially oxidized samples after 60 days of inoculation of strain 2002 cells under the growth condition. Peaks labeled with “V” denote vivianite.

Figure 4. X-ray diffraction patterns for the NAu-2 sample after reduction by CN32 cells for 2 weeks and oxidation by strain 2002 cells under the growth condition for 6 months.

Vivianite and magnetite were detected in this sample.

Figure 5. Secondary electron microscopic (SEM) images & energy dispersive spectroscopy (EDS) spectrum of NAu-2 samples after nitrate-dependent Fe(II)-oxidation of reduced NAu-2 by strain 2002 under the growth condition for 30 days. A. SEM image of Pseudogulbenkiania sp.. B. SEM image of strain 2002 cells in association with wavy

43 NAu-2 clay particles. C. Qualitative SEM-EDS spectrum of NAu-2 showing a typical nontronite composition. Au is derived from gold coating.

Figure 6. SEM images of vivianite from the NAu-2 sample following reduction by CN32 cells for 2 weeks and oxidation by strain 2002 cells under the growth condition for 6 months. A. SEM image of vivianite. B. Qualitative SEM-EDS spectrum of vivianite. Au and Na are derived from gold coating and salt precipitation when the growth medium was dried.

Figure 7. Transmission electron microscopic (TEM) image of a strain 2002 cell with cell encrustation under the growth condition for 6 months. A. A strain 2002 cell with encrustation on its surface. B & C. TEM/EDS spectra of Fe precipitates from the thick

(black arrow) and the thin (white arrow) crust of the cell in Fig. 7A, respectively.

Figure 8. TEM image of ferrihydrite from the nitrate-dependent Fe(II)-oxidation experiment with strain 2002 cells for 60 days. A. Ferrihydrite aggregates. B & C. The selected area electron diffraction (SAED) pattern and the TEM/EDS pattern for the area

(black arrow) identified in Fig. 8A. Fig. 8B reveals four d-spacings 0.47 nm, 0.30 nm,

0.24 nm, and 0.18 nm, which are characteristic for 6-line ferrihydrite. Two other lines for ferrihydrite were not observed, likely due to low intensity. The small Si peak in the EDS spectrum may have originated from adjacent NAu-2, and Cu and Cr peaks may have come from the TEM support grid.

44 Figure 9. TEM images of secondary mineral and NAu-2 from the nitrate-dependent Fe(II) oxidation experiment with strain 2002 for 6 months. A. TEM image shows the presence of NAu-2 and magnetite in close association. B. TEM image of magnetite. C & D.

Qualitative TEM-EDS spectrum and SAED pattern of magnetite from Fig. B, showing its typical elemental composition and diffraction pattern. The Si peak was likely from the adjacent nontronite. E. High resolution transmission electron microscopic lattice-fringe image of nontronite showing 1.09 nm -1.23 nm layer spacing. F & G. The corresponding qualitative TEM-EDS and SAED pattern for Fig. E.

Figure 10. 77-K Mössbauer spectra for different NAu-2 samples: A: Pristine; B: microbially reduced (30.8%); C: re-oxidized for 10 days. Black dots are experimental data and the color curves are model fits.

Fig. 1.

Fig. 2.

Fig. 3.

Fig. 4.

Fig. 5

Fig. 6.

Fig. 7.

Fig. 8.

Fig. 9.

54 o Experimental _____ Modeled ______Fe(III)-A, 30% Fe(III)-A, 6% Fe(III)-A, 17% ______Fe(III)-B, 70% Fe(III)-B, 64% Fe(III)-B 66% _____ A) Pristine Fe(II) 31% _____ Fe(II), 18% B) Bioreduced C) Oxidized

-4 -3 -2 -1 0 1 2 3 -44 -3 -2 -1 0 1 2 3 -44 -3 -2 -1 0 1 2 3 4 Velocity (mm/sec) Fig. 10.

55 CHAPTER 3 Biological Redox cycling of iron in nontronite and its potential application in nitrate removal

Linduo Zhao1, Hailiang Dong*1,2, Ravi K. Kukkadapu3, Qiang Zeng2, Richard E.

Edelmann4, Martin Pentrák5, and Abinash Agrawal6

1. Department of Geology and Environmental Earth Science, Miami University, OH 45056, USA. 2. Geomicrobiology Laboratory, State Key Laboratory of Biogeology and Environmental Geology, China University of Geosciences, Beijing 100083, China. 3. Environmental Molecular Sciences Laboratory, Pacific Northwest National Laboratory, Richland, WA 99354, USA. 4. Center for Advanced Microscopy & Imaging, Miami University, Oxford, OH 45056, USA. 5. Department of Natural Resources and Environmental Sciences, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA 6. Department of Earth & Environmental Sciences, Wright State University, Dayton, OH 45435, USA

Department of Geology Miami University Oxford, OH 45056 Tel: 513-529-2517 Fax: 513-529-1542 Email: [email protected]

Environmental Science and Technology

56 ABSTRACT Biological redox cycling of structural Fe in phyllosilicates is an important but poorly understood process. The objective of this research was to study microbially mediated redox cycles of Fe in nontronite (NAu-2). During the reduction phase, structural Fe(III) in NAu-2 served as electron acceptor, lactate as electron donor, AQDS as electron shuttle, and dissimilatory Fe(III)-reducing bacterium Shewanella putrefaciens CN32 as mediator in bicarbonate- and PIPES-buffered media. During the oxidation phase, biogenic Fe(II) served as electron donor and nitrate as electron acceptor. Nitrate-dependent Fe(II)- oxidizing bacterium Pseudogulbenkiania sp. strain 2002 was added as mediator in the same media. For all three cycles, structural Fe in NAu-2 was able to reversibly undergo three redox cycles without significant dissolution. Fe(II) in bioreduced samples occurred in two distinct environments, at edges and in the interior of the NAu-2 structure. Nitrate reduction to nitrogen gas was coupled with oxidation of edge-Fe(II) and part of interior- Fe(II) under both buffer conditions, and its extent and rate did not change with Fe redox cycles. These results suggest that biological redox cycling of structural Fe in phyllosilicates is a reversible process and has important implications for biogeochemical cycles of carbon, nitrogen, and other nutrients in natural environments. Key words: redox cycling, NAu-2, Shewanella putrefaciens CN32, Pseudogulbenkiania sp. strain 2002

57 1. INTRODUCTION

In the last two decades, significant progress has been made toward understanding the biogeochemical mechanisms to mediate anaerobic nitrate-dependent Fe(II) oxidation (Weber et al., 2006; Shelobolina et al., 2012; Miot et al., 2009; Kappler et al., 2005; Larese-Casanova et al., 2010). This reaction can be generally described as follows: 2+ − + 10퐹푒 + 2푁푂3 + 24퐻2푂 → 10퐹푒(푂퐻)3 + 18퐻 + 푁2 In this reaction, either aqueous Fe2+ or oxide-Fe(II) has been used (Weber et al., 2006; Miot et al., 2009; Kappler et al., 2005; Larese-Casanova et al., 2010; Chaudhuri et al., 2001; Weber et al., 2001). However, the role of structural Fe in phyllosilicates in this reaction has received little attention, despite their ubiquitous presence in soils and sediments and various Fe contents (Shelobolina et al., 2012; Zhao et al., 2013; Dong et al., 2009, 2012; Peretyazhko et al., 2012). Previous research has proven that the structural Fe in phyllosilicates can undergo redox cycling either abiotically or biotically. Biologically, structural Fe can be oxidized by nitrate-dependent Fe(II)-oxidizing bacteria (NDIOB) (Zhao et al., 2013; Shelobolina et al., 2012) and reduced by dissimilatory Fe(III)-reducing bacteria (DIRB) (Dong et al., 2009; Kappler et al., 2005; Lovley et al., 2004) under the temperature and geochemical conditions typical of natural soil and sedimentary environments (Kappler et al., 2005; Lovley et al., 2004; Holmes et al., 2007; Weber et al., 2009). Most Fe in phyllosilicate minerals can undergo biotic/abiotic reduction followed by abiotic oxidation without extensive solubilization (Stucki, 2011). A more recent study demonstrated that biological oxidation of structural Fe(II) in bioreduced NAu-2 can be coupled with nitrate reduction (Zhao et al., 2013). Together, these studies suggest that structural Fe in phyllosilicates has the potential to undergo multiple redox cycles with important implications for environmental processes including continuous removal of nitrate in groundwater aquifers. Given the important role of phyllosilicates in sustaining a dynamic Fe redox cycle, much has been studied about the effects of the oxidation state of structural Fe on the chemical and physical properties of phyllosilicates (Stucki, 2011). Previous studies of dithionite-reduced smectites revealed significant changes in the smectite structure as a result of Fe redox processes (Gates et al., 1993; Stucki et al., 2006). The extent and

58 reversibility of these structural changes showed dependence on the level of reduction achieved by dithionite (Stucki et al., 2006; Gates et al., 1996). Although extensive research has been conducted on chemical redox cycling of Fe in clay minerals (Yang 2010; Ribeiro et al., 2009; Lee et al., 2006), biological redox cycling of Fe, despite its significance in the biogeochemical cycles of Fe, N, and C (Dong et al., 2009; Stucki 2011), has not received equivalent attention. Prior studies have confirmed the presence and the active roles of both phyllosilicate-Fe(III) reducing and phyllosilicate-Fe(II) oxidizing organisms in various redox active environments (Shelobolina et al., 2012; Dong et al., 2012; Dong et al., 2009), such as groundwater seeps (Blöthe et al., 2009), plant root zones (Weiss et al., 2007), and sediment-water interface (Weber et al., 2009), where both Fe-rich clay minerals and nitrate contaminant are likely to be present. With common copresence of Fe redox-cycling microorganisms, organic carbon source (possibly from organic contaminants) and nitrate, coupled redox cycles of Fe, C, and N would have occurred. In light of general lack of research on biological redox cycling of Fe in phyllosilicates and the importance of Fe-bearing clay minerals to the biogeochemical cycles of carbon, nitrogen, and other nutrients, further studies are needed. The overall objective of the present study was therefore to fill the above knowledge gap by studying one Fe-rich smectite, nontronite NAu-2, for its ability to undergo microbially mediated redox cycles, particularly by DIRB and NDIOB. Specifically, the objectives of the present study were to determine: (1) the changes in the extent and rate of Fe reduction and oxidation with increased Fe redox cycle, (2) mineralogical and structural/chemical changes of NAu-2 as a result of these Fe redox cycles, and (3) the effects of microbial Fe redox cycling on nitrate removal. 2. MATERIALS AND METHODS

2.1 Mineral, Media, and Experimental Setup.

Nontronite NAu-2 was used for all redox cycling experiments. The formula of NAu-2 3+ 2+ 3+ is (K0.01Na0.30Ca0.15)(Al0.55Fe 3.27Fe 0.06Mg0.12)(Si7.57Al0.15Fe 0.28)O20(OH)4M0.72 (Keeling et al., 2000; Gates et al., 2002). The total Fe content of NAu-2 is 24%, of which 99.4% is Fe(III) (Mark et al., 2004). Fe(III) occurs in both tetrahedral (8%) and octahedral sites (92%) (Gates et al., 2002; Schaefer et al., 2011). NAu-2 was size-

59 fractionated as previously described (Liu et al., 2011), and the 0.02-0.5 μm size fraction was used. To study the effects of different medium composition on biological redox cycling of Fe in NAu-2 and mineral transformation, sodium bicarbonate and PIPES buffers were used for these experiments. The 0.02-0.5 μm size fraction of NAu-2 was made into 40 mL clay suspensions in bicarbonate (29.76 mM NaHCO3 and 1.34 mM KCl at pH 7.0) and PIPES buffers (1,4-piperazinediethanesulfonic 10 mM at pH 7.0) with NAu-2 concentrations of 3.66 and 3.80 g/L, respectively. The NAu-2 suspensions were then transferred to 120 mL serum bottles, and they were purged with N2/CO2 (80:20) gas mix. The bottles were sealed with thick butyl rubber stoppers and aluminum crimps and autoclaved at 121 ̊C for 60 min (Zhao et al., 2013). Duplicate bottles were set up for these experiments.

Bioreduction experiment was performed using one strain of DIRB Shewanella putrefaciens CN32, with lactate (20 mM) as the sole electron donor, structural Fe(III) in NAu-2 as the sole electron acceptor, and anthraquinone-2,6-disulfonate (AQDS) (0.2 mM) as electron shuttle (Lovley et al., 1998). These high levels of lactate and AQDS were designed to probe important Fe redox process in nontronite but would not be present in natural environment. The experiment was set up as previously described (Zhao et al., 2013; Jaisi et al., 2007). The final cell concentration in the NAu-2 suspension was ∼108 cells/mL. After total Fe(II) concentration stabilized at its maximum, lactate, AQDS, and cells were removed from reduced NAu-2 suspension by five washes with sterile and anoxic (80:20 N2/ CO2 purged) bicarbonate buffer or PIPES buffer followed by centrifugation at 10 000 g for 20 min. Washed NAu-2 was then resuspended in 120 mL serum bottles with the same anoxic buffers followed by a pH adjustment to 7.0-7.5 with NaOH or HCl as necessary and a pasteurization step as previously described (Zhao et al., 2013). A complete removal of lactate and other organic carbon from bioreduced NAu-2 suspension was confirmed with high-performance liquid chromatography (HPLC) measurement. Bio-oxidation experiment was then performed using anaerobic autotrophic Fe-oxidizing bacterium Pseudogulbenkiania sp. strain 2002 (ATCC BAA-1479; DSM 18807) (Weber et al., 2009) with nitrate (10 mM) as the sole electron acceptor and biogenic Fe(II) in NAu-2 as the sole electron donor (Lovley et al., 2004). The experiment was set up as described in our previous paper (Zhao et al., 2013). Strain 2002 was

60 routinely cultured (Weber et al., 2009) to the exponential phase and added into bioreduced and washed NAu-2 suspension to achieve a final cell concentration of ∼108 cells/mL. The headspace of each serum bottle was purged with CO2 gas (a carbon source) using a sterile needle right after strain 2002 inoculation. After Fe(II) oxidation reached its maximum coupled with nitrate reduction, the same washing procedure was applied (e.g., centrifugation followed by resuspension, pH adjustment, and pasteurization). Three Fe redox cycles of biological dissimilatory Fe(III) reduction and nitrate-dependent Fe(II) oxidation were performed. Abiotic controls were the same as the treatments except that they were devoid of iron-reducing or iron-oxidizing bacteria. To reveal the mechanism of Fe(III) bioreduction, NAu-2 was also bioreduced without AQDS under the same conditions. All manipulations of washing NAu-2 or pH adjustment were carried out under strictly anoxic conditions within a glovebox (96% N2, 4% H2, Coy Laboratory Products, Grass Lake, MI, USA).

2.2 Analysis.

The reduction and oxidation reaction progress was monitored by collecting samples from incubation bottles over time for the quantification of Fe(II), total Fe, nitrate, nitrite,

NO, N2O, and N2 during the Fe redox cycling experiments. The procedures for these measurements are included in the Supporting Information. X-ray diffraction (XRD) was performed on clay samples to detect mineralogical changes of NAu-2 as it went through redox cycles. Powder XRD patterns were collected and analyzed as previously described (Zhao et al., 2013). The mineralogical changes of NAu-2 after redox cycles were also characterized by scanning and transmission electron microscopy (SEM and TEM) observations of clay samples collected at the end of each bio-oxidation and bioreduction cycle. The changes in the oxidation state of Fe in NAu-2 (as it went through multiple redox cycles) were examined by electron energy loss spectroscopy (EELS). The speciation of Fe(II) and Fe(III) in the second cycle bioreduced and bio-oxidized NAu-2 samples was also determined with Mössbauer spectroscopy. The bioreduced sample without AQDS was also analyzed with Mössbauer spectroscopy to better understand Fe(III) bioreduction mechanism. The structural changes of NAu-2 after redox cycling, such as changes in the position of OH-stretching (νOH) and OH-bending (δOH) vibration

61 bands, were analyzed by Fourier transform infrared spectroscopy (FTIR) in the middle- infrared region. The second derivative absorption spectra (SDS) were used for a better resolution of the individual bands in FTIR spectra (Mark et al., 2004). Sample preparation and operation procedures for XRD, SEM, TEM, EELS, and Mössbauer and FTIR spectroscopy are described in the Supporting Information.

3. RESULTS

Shewanella putrefaciens was able to reduce structural Fe(III) in nontronite (Figure 1). At the end of Fe(III) bioreduction, the oxidation of biogenic Fe(II) coupled with nitrate reduction commenced. Fe(II), nitrate, and nitrite concentration profiles did not show any significant difference between bicarbonate and PIPES buffers. In each Fe redox cycle, the color changed from dark green to light green upon oxidation and then back to dark green upon another round of reduction (Figure 6).

During the bioreduction phase of all three cycles, Fe(II) concentration increased rapidly within the first 2 days and gradually reached the maximum within 7 days (Figure 1). The overall extent of Fe(III) bioreduction, that is, the ratio of the net amount of Fe(II) production divided by the amount of Fe(III) at the beginning of that cycle (eq 1), decreased with each successive Fe redox cycle in both bicarbonate and PIPES buffers (Table 1). In contrast, the Fe(II) bio-oxidation extent, calculated in a similar way (eq 2), increased with each redox cycle under both buffer conditions:

퐹푒(퐼퐼) −퐹푒(퐼퐼) 푅푒푑푢푐푡푖표푛 푒푥푡푒푛푡 = 푓푖푛푎푙 푖푛푖푡푖푎푙 (Eq. 1) 퐹푒(퐼퐼퐼)푡표푡푎푙

퐹푒(퐼퐼) −퐹푒(퐼퐼) 푂푥푖푑푎푡푖표푛 푒푥푡푒푛푡 = 푖푛푖푡푖푎푙 푓푖푛푎푙 (Eq. 2) 퐹푒(퐼퐼)푡표푡푎푙

The initial rates of Fe(II) bio-oxidation (calculated for the first 24 h, eq 3) from all three cycles were lower than the bioreduction rates (also for the first 24 h, eq 3) (Table 1).

|퐹푒(퐼퐼) −퐹푒(퐼퐼) | 퐼푛푖푡푖푎푙 푟푒푑푢푐푡푖표푛 (표푥푖푑푎푡푖표푛) 푟푎푡푒 = 푓푖푛푎푙 푖푛푖푡푖푎푙 (Eq. 3) 훥푡푖푚푒

Although excess nitrate was added relative to Fe(II) (assuming a stoichiometric molar - - Fe/NO3 ratio of 5 if the N2 is the NO3 reduction product) (Zhao et al., 2013), Fe(II) in

62 NAu-2 was never exhausted in any cycle, even after a prolonged incubation (several months). Instead, nitrate was completely reduced to N2 gas in 5 days (Figure 7), and residual Fe(II) remained. NO or N2O was not detected at any time point, possibly because of their reactive transient nature. A certain amount of nitrite was produced as a transient intermediate in both bicarbonate and PIPES buffers. Nitrite concentration usually reached a maximum of around 2.5-3.0 mM after 1-2 days. In the second cycle with bicarbonate buffer, nitrite was not fully reduced. However, an addition of freshly cultured cells after 140 days completely reduced the residual nitrite (Figure 1B). The observed stoichiometry (molar ratio) of the amount of nitrate reduced to the amount of Fe(II) oxidized under bicarbonate buffer (4.8, 4.3, and 3.3 for the first, second, and third redox cycle, respectively) and PIPES buffer (2.7, 2.9, and 1.9 for the first, second, and third redox cycle, respectively) condition greatly exceeded the theoretical values of 1:5 according to eq R1. Similar observations were made in other studies on nitrate-dependent microbial oxidation of Fe(II) in solid minerals (Weber et al., 2001). One possibility is heterotrophic nitrate reduction coupled with oxidation of dead bacterial biomass or energy reserves (i.e., glycogen) stored in cells (Zhao et al., 2013).

XRD showed that NAu-2 was the dominant mineral throughout the three redox cycles with three broad peaks corresponding to d001, d002, and d005 reflections of nontronite (Figure 8). Neither bioreduction of Fe(III) nor bio-oxidation of Fe(II) in NAu-2 led to any broadened or shifted peaks, which suggests little change of the nontronite structure after three redox cycles. XRD results also showed that all bioreduced and bio-oxidized NAu-2 samples contained albite (NaAlSi3O8). The identification of albite was based on a peak with d-spacing of d002 = 0.32 nm (the black arrow in Figure 8) that corresponds to the most intense peak of standard albite (PDF 00-010-0393). Other peaks of albite were absent, possibly due to their low intensities. The existence of biogenic albite was further confirmed by SEM and TEM images, chemical composition from energy-dispersive X- ray spectroscopy (EDS) analysis, and selected area electron diffraction pattern of the second bioreduced NAu-2 sample (Figure 9). Albite formation in the NAu-2 samples suggests that the NAu-2 structure went through partial dissolution during Fe redox cycling.

63 High-resolution TEM images and lattice-fringe images of the bio-oxidized NAu-2 samples from the first and second redox cycles showed a coexistence of expandable, wavy, 1.3 nm layers and nonexpandable, lath-shaped, 1.0-1.1 nm layers (Figure 10). EDS analysis indicated that these bio-oxidized NAu-2 samples were mainly composed of O, Fe, Al, Si, and Fe (Figure 10A,D), and these elements were characteristic chemical composition for NAu-2. Some peaks in the spectrum of the first bio-oxidized NAu-2 sample, such as S, Cl, K, Ca, and Mn, may have been derived from the trace minerals in the growth medium. As the number of redox cycles increased, the amounts of these elements gradually diminished, likely due to successive washes. The Fe:O ratio of the second bio-oxidized NAu-2 was lower than the same ratio from the first cycle, which suggests preferential removal of Fe from the NAu-2 structure, possibly due to some redox-triggered dissolution.

EELS spectra confirmed the Fe oxidation state change as a result of Fe redox cycle (Figure 2). By using a calibration curve established in this study (Figure 11), the ferric to total Fe ratio in redox-cycled NAu-2 samples deduced from the L3/L2 integral intensity ratio (Figure 2) was comparable to the chemical results determined by the 1,10- phenanthroline method (Table 2). The slight difference between the EELS and the chemical results could be ascribed to the partial reduction or oxidation of structural Fe in NAu-2 under the TEM electron beam. Another possibility is that Fe in the NAu-2 structure may be heteregenously reduced or oxidized by bacteria. The energy positions of

L2 and L3 edges for redox cycled NAu-2 samples were nearly reversible upon three redox cycles (Figure 2). The Fe L3 edge in all reduced NAu-2 samples displayed the maximum at less than 709 eV, whereas all oxidized NAu-2 samples displayed the maximum at around 710 eV.

Bioreduction of NAu-2 resulted in the emergence of a distinct Fe(II) doublet and a nondistinct central broad feature at room temperature (RT, Figure 12A). A model fit to this spectrum was tentative (Figure 12B), and the nature of the broad feature was not apparent. However, assignment of the distinct Fe(II) doublet at RT was possible and revealed that there was ∼14% Fe(II) (Figure 3A). Upon cooling from RT to 77 K, the content of the distinct Fe(II) doublet increased to 30% Fe(II) (Figure 3B). This difference

64 in Fe(II) content between RT and 77 K could be due to two possibilities. First, a fraction of Fe(II) was not responsive to Mössbauer effect at RT, due to higher thermal energies and little or lower recoilless fraction of Fe(II) species, as in the case of Fe2+ in the interlayer of montmorillonite (Diamant et al., 1982). Second, there was an existence of

Fe(II) as Fe(II)x-Fe(III)y moieties that behaved differently at RT and 77 K. At RT, electron exchange between Fe(II)-Fe(III) in these moieties could have occurred at a rate higher than the Mössbauer detection time that could lead to a continuum of Fe species with oxidation states intermediate between Fe(II) and Fe(III) (such as Fe2x+ as suggested by the broad feature in Figure 3A). However, at 77 K, this electron exchange may be slower than the Mössbauer detection time, and as a result, Fe(II) and Fe(III) species in these moieties would have become distinct (Figure 3B). This second explanation was adopted in this paper for several reasons: (1) the broad feature in the RT spectrum for bioreduced NAu-2 had a center shift (CS) of 0.8 mm/s (Figures 12B, Table 3), which was between values for Fe2+ (0.3 mm/s) and Fe3+ (1.2 mm/s) (Murad et al., 2004). In addition, the large standard deviation of quadruple splitting (QS) (0.62 mm/sec, Table 3) was in agreement with existence of a continuum of Fe(II) species with different coordination environments. (2) Such a broad feature with similar CS and large standard deviation of QS and average oxidation state of 2.3 was also noted in Fe2+-sorbed nontronite (Schaefer et al., 2011). (3) The hypothesis of electron hopping between Fe(II) and Fe(III) was supported by a theoretical study (Alexandrov et al., 2013), where an electron hopping rate higher than 107/s at RT was noted for nontronite.

Thus, according to this model, the Fe(II) in bioreduced NAu-2 sample exists in two different forms: (1) Fe(II) in a relatively “rigid” environment (or with relatively higher recoilless fraction) as a distinct Fe(II) domain (14%) where it could be detected by Mössbauer spectroscopy, even at RT, and is typical of Fe(II) in the stable Fe(II)-clays (Kukkadapu et al., 2001), and (2) transient Fe(II) that supposedly rapidly exchanges electron with neighboring Fe(III) in an Fe(II)-Fe(III) domain (16%) where it was invisible to Mössbauer spectroscopy at RT but could be detected at 77 K. The sum of these two forms of Fe(II) gave rise to an Fe(II)/total Fe ratio of 30% for the second bioreduction cycle (Figure 3B), which agreed well with the chemically determined ratio of 32.7% (Table 2).

65 Spectral modeling of the bio-oxidized NAu-2 sample showed a distinct Fe(II) doublet only at both RT and 77 K (Figure 3C,D). In other words, the second form of Fe(II) from the Fe(II)−Fe(III) domain was preferentially oxidized. The 11.5% difference between Mössbauer (7%) and chemical (18.5%) results for the second oxidized NAu-2 sample (Table 2) could be ascribed to partial oxidation during the drying and shipping processes for the Mössbauer measurement.

FTIR spectra (Figure 4) for the redox-cycled NAu-2 samples collected after 20 days of incubation displayed distinct changes in the position of OH-stretching (νOH) and OH- bending (δOH) vibration bands. In the OH-stretching region, while unaltered NAu-2 sample showed one absorption peak at 3566 cm−1 (Figure 4A), all bioreduced samples exhibited one peak at 3556 (Figure 4B) or ∼3540 cm−1 (Figure 4C,D), which can be attributed to the Fe(II)-Fe(III)-OH environment (Lee et al., 2006). Upon oxidation, this peak largely diminished but restored after the next round of bioreduction. This observation was consistent to the Mössbauer spectroscopy result and suggests that the redox active Fe was located largely in the Fe(II)-Fe(III) domain. In addition, the band at ∼3643 cm−1 (Figure 4A), which could be attributed to Al-Al-OH environment (Lee et al., 2006), shifted to 3641 cm−1 upon the first redox cycle (Figure 4B). However, upon the second bioreduction, this band shifted to a lower wavenumber (3634-3636 cm-1) (Figure 4C,D) but shifted back to 3641 cm−1 upon subsequent oxidation (Figure 4C,D). This change in the Al-Al-OH bonding suggests that Fe valence change affected not only Fe bonding, but also the overall structure of NAu-2, consistent with a previous study (Lee et al., 2006). The principal feature in the OH-bending vibration region was a strong band at ∼820 cm−1 in the unaltered NAu-2 sample, which could be assigned to Fe-Fe-OH (Figure 4E) (Lee et al., 2006). This band shifted to 818 cm−1 after the first redox cycle (Figure 4F). However, during the second and the third cycle, the band shifted to a lower wavenumber (816-817 cm−1) after bioreduction, but upon bio-oxidation, it shifted back to 818-819 cm−1 (Figures 4G,H). A similar observation was made for the Al-Fe-OH band at 867 cm−1 (Lee et al., 2006). In the first redox cycle, this band stayed at 867 cm−1 (Figure 4E,F). However, during the second cycle, this band shifted to a lower wavenumber (860 cm−1) upon reduction, but the next reoxidation shifted this band back to the same position (868 cm−1) as for the unaltered NAu-2 (Figure 4G). After the third bioreduction, two

66 bands at 874 and 863 cm−1 were observed, while after the third biooxidation, the band at 867 cm−1 emerged (Figure 4H). In addition, the vibration band found in the 782-797 cm−1 region was at a slightly higher wavenumber (783, 789, and 797 cm−1) for the bioreduced samples than for the oxidized samples (782, 784, and 796 cm−1) (Figures 4E-H). The assignment of the peaks in this region was uncertain. Some researchers attributed it to Fe- Mg-OH (Russell et al., 1979; Stucki et al., 1976). In summary, these FTIR data suggest that other than the redox cycle of Fe, the rest of the structure was slightly distorted upon bioreduction but nearly fully restored upon subsequent reoxidation. Among the three cycles, structural changes were more obvious in the second and third cycles than in the first cycle.

4. DISCUSSION

4.1 Bioreduction of Fe(III) in NAu-2 through multiple redox cycles.

The overall decreased Fe(III) bioreduction extents with each successive Fe redox cycle is consistent with our previous study (Yang et al., 2012) and may be explained by several reasons. One possibility is change of NAu-2 physical and chemical properties (i.e., the Fe(II)/Fe(III) ratio, the layer charge, crystal structure, and mineralogy) across three redox cycles (Stucki 2011; Stucki et al., 2006). The formation of biogenic albite might be hindering access to some of the Fe in NAu-2 and thus might result in lower extents of bioreduction in subsequent cycles. An alternative explanation is the heterogeneous nature of NAu-2 particles. Natural NAu-2 particles vary in size and crystallinity (Jaisi et al., 2007), and therefore, they are subjected to different extents and rates of bioreduction. The structural Fe(III) in small or poorly crystalline NAu-2 particles should be preferentially reduced because of their larger surface area (Roden et al., 1996). The presence of these heterogeneities in the original NAu-2 sample would thus result in a higher extent of bioreduction in the first cycle with consequent reductive dissolution and structural changes, as evidenced by biogenic albite formation (Figure 9) and release of major elements of NAu-2 (Si, Al, Fe) to aqueous solution (Dong 2012; Dong et al., 2009; Liu et al., 2011). Our previous studies have shown that when nontronite was bioreduced and air reoxidized three times (Yang et al., 2012; Glasser 2014), most reductive dissolution occurred during the first cycle, and subsequent redox cycles did not result in any further

67 dissolution. Because the first redox cycle dissolved small and poorly crystalline particles, and residual large and well-crystalline particles were not really affected; the FTIR analysis of these residual NAu-2 particles did not show any major structural changes (Figure 4).

Starting from the second cycle, Fe redox cycling should have occurred, largely in solid state, to the structural Fe in relatively larger and well-crystalline NAu-2 particles. Because there would be a minimal amount of dissolution from this point onward, structural changes were largely reversible as evidenced by reversible band shifts of Fe- Fe-OH and Al-Fe-OH vibrations in the mid infrared region (Figure 4). However, the reduction extent still gradually decreased in the second and third cycles (Table 1), likely because redox-cycled NAu-2 became less expandable over time, as evidenced by an increasing number of 1.1 nm layer spacings in successive cycles (Figure 10). The less- expandable nature of NAu-2 would hinder subsequent bioreduction, as reduction extent has previously been observed to be proportional to the NAu-2 interlayer expandability (Bishop et al., 2014; Zhang et al., 2013).

The mechanism of bioreduction for structural Fe(III) in phyllosilicates has been largely elusive (Dong et al., 2009). The direction of electron transfer pathway for phyllosilicates has been postulated to be both parallel (i.e., via edges) and perpendicular (i.e., via basal planes) to NAu-2 layers (Dong et al., 2009). The parallel reduction mechanism from clay edges has been previously proposed for Garfield nontronite by Ribeiro et al (Ribeiro et al., 2009). In this model, bacterial reduction of Fe(III) to Fe(II) is believed to proceed from clay edges (Figure 5A,B) to the structural interior and a reduction front is created at the Fe(II)−Fe(III) redox boundary. The presence of this Fe(II)−Fe(III) domain in our study is manifested in the Mössbauer data (i.e., edge Fe(II)−Fe(III) domain in Figure 3B). This Fe(II)−Fe(III) domain can be tentatively assigned to NAu-2 edge sites based on a comparison of Mössbauer spectra between bioreduced NAu-2 samples with and without AQDS. In the absence of AQDS, bioreduced NAu-2 exhibited a broad feature at RT without a distinct Fe(II) doublet (Figure 13A), and this broad feature was similar to the one for bioreduced NAu-2 with AQDS (Figure 3A). Upon cooling to 77 K, this broad feature transformed to a distinct

68 Fe(II) doublet (Figure 13B), and its content (18%) was similar to the second form of Fe(II) in the AQDS, bioreduced sample (16%, Figure 3B). The existence of this form of Fe(II), that is, an Fe(II)−Fe(III) domain, in no-AQDS, bioreduced NAu-2 sample suggests that it was derived from NAu-2 edge sites because bioreduction of NAu-2 without AQDS should have occurred via a direct contact between bacteria and NAu-2, which was presumably only possible at edge sites via a reduction front (Dong et al., 2009; Ribeiro et al., 2009). In addition to the Fe(II)−Fe(III) domain, our Mössbauer spectra also revealed an Fe(II) domain within the interior of the NAu-2 structure (Figures 3B and 5B). In the Ribeiro’s model, such a domain was believed to be produced by chemical reduction via basal planes. The reason for its presence in a biologically reduced NAu-2 is intriguing, and it may be due to the presence of electron shuttle (AQDS) in the bioreduction medium. Indeed, this interior Fe(II) domain was only present when AQDS facilitated electron transfer in the direction of perpendicular to NAu-2 layers (i.e., the distinct Fe(II) doubletin Figure 3A), and in the absence of AQDS, such interior Fe(II) domain was absent (i.e., lack of a distinct Fe(II) doublet in the RT spectrum of bioreduced NAu-2, Figure 13A) (Jaisi et al., 2005). Nonetheless additional experiments to reduce NAu-2 to different extents (with and without AQDS) are currently in progress and may provide further insights into NAu-2 bioreduction mechanisms.

4.2 Bio-oxidation of Fe(II) in NAu-2 through multiple redox cycles.

The mechanism for bio-oxidation of structural Fe(II) in NAu-2 appears fundamentally different from that for bioreduction of structural Fe(III). RT and 77 K Mössbauer spectra of the bio-oxidized sample during the second cycle suggest a complete oxidation of Fe(II) in the Fe(II)−Fe(III) domain but a limited amount of oxidation of interior Fe(II) (Figure 3). This suggests that bio-oxidation occurred preferentially through edges (Figure 5C). A likely reason is that the electron shuttle AQDS was not present in the biooxidation experiment, and therefore, electron transfer via basal plane was unlikely to occur. The different electron transfer mechanisms between bioreduction and bio-oxidation are also supported by the fact that the initial bioreduction rates were significantly faster than the initial bio-oxidation rates (Table 1) because two electron transfer pathways for bioreduction (both parallel and perpendicular to NAu-2 layers, due to the presence of

69 AQDS) were expected to be more efficient than one electron transfer pathway for bio- oxidation (parallel to NAu-2 layers only, due to the absence of AQDS).

Specific mechanisms for the electron transfer from Fe(II) to nitrate in aqueous solution remain elusive, but previous research has invoked indirect electron transfer mechanisms that may be mediated by surfaced-sorbed aqueous Fe3+ or solidphase Fe(III) (Zhao et al., 2013; Shelobolina et al., 2012; Kappler et al., 2005). Initially, Fe(II) in the Fe(II)−Fe(III) domain at NAu-2 edges could directly transfer electrons to aqueous nitrate to form Fe3+. The surface-sorbed Fe3+ could in turn oxidize its adjacent Fe(II) via intervalence electron transfer (Schaefer et al., 2011; Neumann et al., 2013). If initial oxidation of Fe(II) resulted in solid-form Fe(III) (i.e., Fe(III) oxides), this solid Fe(III) could also oxidize its adjacent Fe(II) via intervalence electron transfer (Schaefer et al., 2011; Neumann et al., 2013). However, no biogenic Fe(III) minerals were observed by TEM or Mössbauer spectroscopy, which suggests that the latter mechanism is unlikely.

4.3 Biological redox cycling of Fe in phyllosilicates and environmental implications.

Our results demonstrated that the structural Fe in NAu-2 was able to undergo multiple redox cycles that were mediated by Fe(III)-reducing and Fe(II)-oxidizing bacteria. These results expand previous single-cycle studies using phyllosilicates (Zhao et al., 2013; Shelobolina et al., 2012; Shelobolina et al., 2012) and demonstrate that biological redox cycling of Fe in these minerals can be sustained over longer time. This sustainable nature of phyllosilicates is due to the nature of solid-state Fe redox cycling. Although our data suggest a small amount of reductive/oxidative dissolution, this type of dissolution only occurred in the first cycle with some irreversible changes in cation exchange capacity and specific surface area (Stucki 2011). Some biogenic minerals, such as albite, silica, siderite, vivianite, and illite, would be consistent with this model (Kim et al., 2004; Jaisi et al., 2007; Dong et al., 2003; Zhang et al., 2007; Zhang et al., 2007). However, once fine or poorly crystalline NAu-2 particles were preferentially dissolved, the residual NAu-2 particles became fairly resistant to further dissolution (Yang et al., 2012; Bishop et al., 2014), as evidenced by the nearly constant intensity and width of d(001) peak of NAu-2 after the second and third bioreduction and biooxidation cycles (Figure 8). Furthermore, the alternate presence and absence of Fe(III)-Fe(II)-OH environment, as detected by

70 FTIR in the bioreduced and bio-oxidized samples, respectively, further supports the reversible nature of Fe oxidation state in NAu-2. Similar observations were made by previous researchers on smectites; the observed νOH and δOH changes in peak position and intensity were largely reversible when the reduced smectite was reoxidized chemically (Lee et al., 2006). This sustainable nature of phyllosilicates, upon Fe redox cycling, is in contrast to Fe oxides, where a single cycle of reduction and oxidation usually dissolves the mineral and renders the redox cycle of Fe nonsustainable. For example, one study recently reported multiple anaerobic microbial redox cycles of synthetic goethite (α- FeOOH) (Coby et al., 2011), and a large amount of amorphous Fe materials was generated after redox cycling, which suggests that there was extensive dissolution after the Fe redox cycle of the goethite.

The reversible and solid-state nature of Fe redox cycling in phyllosilicates has important environmental implications. Iron-bearing phyllosilicates are dominant minerals in soils and sediments, and they play important roles in organic carbon retention (Keil et al., 2014), and nutrient cycling (Stucki 2011), largely due to their fine grain size and large surface area. Our results suggest that oscillating redox condition in natural environment, such as wetting and drying of agricultural soils, should not release organic carbon, trace metals, and nutrients because these minerals tend to remain in solid state. Our past studies have demonstrated that organic molecules intercalated into interlayer space of nontronite can indeed be protected, apparently because Fe redox cycling occurs largely in solid state (Zhang et al., 2007; Zhang et al., 2014). Our results also offer insights to nitrate remediation in the environment. Biological denitrification is a major pathway that can be accomplished by certain metal-reducing bacteria in the presence of organic matter (Krause et al., 1997). However, our approach may be important to remove nitrate under two conditions: (1) when certain metal-reducing bacteria are not able to reduce nitrate enzymatically and (2) when Fe-reducing bacteria and organic matter are not present at the same time or location as nitrate. In this case, nontronite can be reduced by Fe-reducing bacteria, and then reduced nontronite can be moved to a different location or used at a later time to remove nitrate. However, it should be recognized that some geochemical conditions used in this experiment (such as 20 mM lactate and 0.2 mM AQDS) do not exactly simulate natural conditions but instead are designed to probe important Fe redox

71 processes. Further studies are needed to examine these environmental processes under more realistic conditions. Coupled reduction and oxidation of structural Fe in phyllosilicates has important implications for microbial ecology as well. Previous studies have shown that certain bacteria can gain energy for growth from reduction of Fe(III) (Senn et al., 2002) and oxidation of Fe(II) (Lack et al., 2002) in phyllosilicates. By combining these two modes of energy generation, this study has shown that Fe(III)- reducing and Fe(II)-oxidizing bacteria can possibly form a syntrophic relationship to mutually support each other’s growth via clay minerals as an environmental medium. These two types of microorganisms may be spatially and temporally separated but are connected via clay minerals. Because of the abundance of Fe-bearing clay minerals in soils and sediments, it is expected that this mode of syntrophic metabolism may be an important pathway in both modern and ancient environments.

ACKNOWLEDGMENTS

The work was supported by a grant from National Science Foundation (NSF) (EAR 1148039). A portion of the research was performed using EMSL, a national scientific user facility sponsored by the Department of Energy’s Office of Biological and Environmental Research located at Pacific Northwest National Laboratory. The JEOL 2100 TEM used in this study was supported by NSF Grant No. EAR-0722807. We are grateful to the associate Editor and three anonymous reviewers whose comments significantly improved the quality of the manuscript.

72 REFERENCES

Alexandrov, V., Neumann, A., Scherer, M.M., and Rosso, K.M. Electron exchange and conduction in nontronite from first-principles. The Journal of Physical Chemistry. 2013, 117, 2032-2040.

Bishop, M. E.; Glasser, P.; Dong, H.; Arey, B.; Kovarik, L. Reduction and immobilization of hexavalent chromium by microbially reduced Fe-bearing clay minerals. Geochim. Cosmochim. Acta 2014, 133 (0), 186-203.

Blöthe, M.; Roden, E. E. Microbial iron redox cycling in a circumneutral-pH groundwater seep. Appl Environ Microbiol 2009, 75 (2), 468-473.

Chaudhuri, S. K.; Lack, J. G.; Coates, J. D. Biogenic magnetite formation through anaerobic biooxidation of Fe(II). Appl Environ Microbiol 2001, 67 (6), 2844-2848.

Coby, A. J.; Picardal, F.; Shelobolina, E.; Xu, H. F.; Roden, E. E. Repeated Anaerobic Microbial Redox Cycling of Iron. Appl Environ Microbiol Microbiology 2011, 77 (17), 6036-6042.

Diamant, A., Pasternak, M., and Banin, A. Characterization of adsorbed iron in montmorillonite by Mössbauer spectroscopy. Clays & Clay Minerals 1982, 30, 63- 66.

Dong, H.; Kostka, J. E.; Kim, J. Microscopic evidence for microbial dissolution of smectite. Clays and Clay Minerals 2003, 51 (5), 502-512.

Dong, H.; Jaisi, D. P.; Kim, J.; Zhang, G. Microbe-clay mineral interactions. American Mineralogist 2009, 94 (11-12), 1505-1519.

Dong, H. L. Clay-Microbe interactions and implications for environmental mitigation. Elements 2012, 8 (2), 113-118.

Gates, W. P.; Wilkinson, H. T.; Stucki, J. W. Swelling properties of microbially reduced ferruginous smectite. Clays and Clay Minerals 1993, 41 (3), 360-364.

73 Gates, W. P.; Stucki, J. W.; Kirkpatrick, R. J. Structural properties of reduced Upton montmorillonite. Physics and Chemistry of Minerals 1996, 23 (8), 535-541.

Gates, W. P.; Slade, P. G.; Manceau, A.; Lanson, B. Site occupancies by iron in nontronites. Clays and Clay Minerals 2002, 50 (2), 223-239.

Glasser P.A. Kinetics and mechanisms of Cr6+ reduction by structural Fe(II) in clay minerals. M.S. Dissertation, Miami University, Oxford, OH, 2014.

Holmes, D. E.; O'Neil, R. A.; Vrionis, H. A.; N'Guessan, L. A.; Ortiz-Bernad, I.; Larrahondo, M. J.; Adams, L. A.; Ward, J. A.; Nicoll, J. S.; Nevin, K. P.; Chavan, M. A.; Johnson, J. P.; Long, P. E.; Lovley, D. R. Subsurface clade of Geobacteraceae that predominates in a diversity of Fe(III)-reducing subsurface environments. Isme Journal 2007, 1 (8), 663-677.

Jaisi, D. P.; Dong, H. L.; Liu, C. X. Kinetic analysis of microbial reduction of Fe(III) in nontronite. Environ SciTechnol 2007, 41 (7), 2437-2444.

Kappler, A.; Benz, M.; Schink, B.; Brune, A. Electron shuttling via humic acids in microbial iron(III) reduction in a freshwater sediment. Fems Microbiology Ecology 2004, 47 (1), 85-92.

Kappler, A.; Schink, B.; Newman, D. K. Fe(III) mineral formation and cell encrustation by the nitrate-dependent Fe(II)-oxidizer strain BoFeN1. Geobiology 2005, 3 (4), 235-245.

Keeling, J. L.; Raven, M. D.; Gates, W. P. Geology and characterization of two hydrothermal nontronites from weathered metamorphic rocks at the Uley Graphite Mine, South Australia. Clays and Clay Minerals 2000, 48 (5), 537-548.

Kim, J.; Dong, H.; Seabaugh, J.; Newell, S. W.; Eberl, D. D. Role of microbes in the smectite-to-illitereaction. Science 2004, 303 (5659), 830-832.

Kukkadapu R. K.; Zachara J. M.; Smith S. C.; Fredrickson J. K.; Liu C. Dissimilatory bacterial reduction of Al-substituted goethite in subsurface sediments. Geochim. Cosmochim. Acta2001, 65, 2913–2924.

74 Jaisi, D. P.; Kukkadapu, R. K.; Eberl, D. D.; Dong, H. Control of Fe(III) site occupancy on the rate and extent of microbial reduction of Fe(III) in nontronite. Geochim. Cosmochim. Acta 2005, 69 (23), 5429-5440.

Jaisi, D. P.; Dong, H. L.; Kim, J.; He, Z. Q.; Morton, J. P. Nontronite particle aggregation induced by microbial Fe(III) reduction and exopolysaccharide production. Clays and Clay Minerals 2007, 55 (1), 96-107.

Keil, R.G., Mayer, L.M. Mineral matrices and organic matter, Treatise on Geochemistry 2nd Edition. 2014, 337-359.

Krause, B., Nealson, K.H. Physiology and enzymology involved in denitrification by Shewanella putrefaciens. Appl Environ Microbiol. 1997,63(7), 2613-8.

Lack, J. G.; Chaudhuri, S. K.; Kelly, S. D.; Kemner, K. M.; O'Connor, S. M.; Coates, J. D. Immobilization of radionuclides and heavy metals through anaerobic bio- oxidation of Fe(II). Applied and Environmental Microbiology 2002, 68 (6), 2704- 2710.

Larese-Casanova, P.; Haderlein, S. B.; Kappler, A. Biomineralization of lepidocrocite and goethite by nitrate-reducing Fe(II)-oxidizing bacteria: Effect of pH, bicarbonate, phosphate, and humic acids. Geochim. Cosmochim. Acta 2010, 74 (13), 3721-3734.

Lee, K.; Kostka, J. E.; Stucki, J. W. Comparisons of structural Fe reduction in smectites by bacteria and dithionite: an infrared spectroscopic study. Clays and Clay Minerals 2006, 54 (2), 195-208.

Liu, D.; Dong, H. L.; Bishop, M. E.; Wang, H. M.; Agrawal, A.; Tritschler, S.; Eberl, D. D.; Xie, S. C. Reduction of structural Fe(III) in nontronite by methanogen Methanosarcina barkeri. Geochim. Cosmochim. Acta 2011, 75 (4), 1057-1071.

Lovley, D. R.; Fraga, J. L.; Blunt-Harris, E. L.; Hayes, L. A.; Phillips, E. J. P.; Coates, J. D. Humic substances as a mediator for microbially catalyzed metal reduction. Acta Hydrochimica Et Hydrobiologica 1998, 26 (3), 152-157.

75 Lovley, D. R.; Holmes, D. E.; Nevin, K. P. DissimilatoryFe(III) and Mn(IV) reduction. Adv Microb Physiol 2004, 49, 219-286.

Mark, H.; Workman, J. Chemometrics in spectroscopy-Derivatives in spectroscopy update. Spectroscopy 2004, 19 (12), 110-112.

Miot, J.; Benzerara, K.; Morin, G.; Bernard, S.; Beyssac, O.; Larquet, E.; Kappler, A.; Guyot, F. Transformation of vivianite by anaerobic nitrate-reducing iron-oxidizing bacteria. Geobiology 2009, 7 (3), 373-384.

Murad E. and Cahsion J. Mössbauer spectroscopy of environmental materials and industrial utilization. 2004. Kluwer Academic Publisher.

Neumann, A.; Olson, T. L.; Scherer, M. M.Spectroscopic evidence for Fe(II)-Fe(III) electron transfer at clay mineral edge and basal sites. Environ SciTechnol 2013, 47 (13), 6969-6977.

Peretyazhko T. S.;Zachara J. M.;Kukkadapu R. K.;Heald S. M.; Kutnyakov I. V.; Resch C. T.; Arey B. W.; Wang C. M.; Kovarik L.; Phillips J.L.; Moore D. − A.Pertechnetate (TcO4 ) reduction by reactive ferrous iron forms in naturally anoxic, redox transition zone sediments from the Hanford Site, USA.Geochim. Cosmochim. Acta 2012, 92, 48–66

Roden, E. E.; Zachara, J. M. Microbial reduction of crystalline iron(III) oxides: influence of oxide surface area and potential for cell growth. Environ SciTechnol 1996, 30 (5), 1618-1628.

Russell, J. D.; Goodman, B. A.; Fraser, A. R. Infrared and Mössbauer studies of reduced nontronites. Clays and Clay Minerals 1979, 27 (1), 63-71.

Schaefer, M. V.; Gorski, C. A.; Scherer, M. M. Spectroscopic evidence for interfacial Fe(II)-Fe(III) electron transfer in a clay mineral. Environ SciTechnol 2011, 45 (2), 540-545.

76 Shelobolina, E.; Xu, H. F.; Konishi, H.; Kukkadapu, R.; Wu, T.; Blothe, M.; Roden, E. Microbial lithotrophicoxidation of structural Fe(II) in biotite. Appl Environ Microbiol. 2012, 78 (16), 5746-5752.

Shelobolina E.; Konishi H.; Xu H.; Benzine J.; Xiong M.; Wu T.; Blöthe M.; Roden E. Isolation of phyllosilicate–iron redox cycling microorganisms from an illite– smectiterich hydromorphic soil. Frontiers in Microbiology 2012, 3.

Senn, D. B.; Hemond, H. F. Nitrate controls on iron and arsenic in an urban lake. Science 2002, 296 (5577), 2373-2376.

Stucki, J. W.; Roth, C. B. Interpretation of infrared-spectra of oxidized and reduced nontronite. Clays and Clay Minerals 1976, 24 (6), 293-296.

Stucki, J. W.; Kostka, J. E. Microbial reduction of iron in smectite. Comptes Rendus Geoscience 2006, 338 (6-7), 468-475.

Stucki, J. W. A review of the effects of iron redox cycles on smectite properties. Comptes Rendus Geoscience 2011, 343 (2-3), 199-209.

Ribeiro, F. R.; Fabris, J. D.; Kostka, J. E.; Komadel, P.; Stucki, J. W. Comparisons of structural iron reduction in smectites by bacteria and dithionite: II. A variable- temperature Mössbauer spectroscopic study of Garfield nontronite. Pure and Applied Chemistry 2009, 81 (8), 1499-1509.

Weber, K. A.; Picardal, F. W.; Roden, E. E. Microbially catalyzed nitrate-dependent oxidation of biogenic solid-phase Fe(II) compounds. Environ SciTechnol 2001, 35 (8), 1644-1650.

Weber, K. A.; Achenbach, L. A.; Coates, J. D. Microorganisms pumping iron: anaerobic microbial iron oxidation and reduction. Nat. Rev. Microbiol. 2006, 4 (10), 752-764.

Weber, K. A.; Hedrick, D. B.; Peacock, A. D.; Thrash, J. C.; White, D. C.; Achenbach, L. A.; Coates, J. D. Physiological and taxonomic description of the novel autotrophic, metal oxidizing bacterium, Pseudogulbenkianiasp strain 2002. Applied Microbiology and Biotechnology 2009, 83 (3), 555-565.

77 Weiss, J. V.; Rentz, J. A.; Plaia, T.; Neubauer, S. C.; Merrill-Floyd, M.; Lilburn, T.; Bradburne, C.; Megonigal, J. P.; Emerson, D. Characterization of neutrophilic Fe(II)-oxidizing bacteria isolated from the rhizosphere of wetland plants and description of Ferritrophicum radicicola gen. nov. sp. nov., and Sideroxydans paludicola sp. nov. Geomicrobiology Journal 2007, 24 (7-8), 559-570.

Yang, J. Effects of redox cyclingsof iron in nontronite on reduction of technetium. M.S. Dissertation, Miami University, Oxford, OH, 2010.

Yang, J. J.; Kukkadapu, R. K.; Dong, H. L.; Shelobolina, E. S.; Zhang, J.; Kim, J. Effects of redox cycling of iron in nontronite on reduction of technetium. Chemical Geology 2012, 291, 206-216.

Zhang, G. X.; Dong, H. L.; Kim, J. W.; Eberl, D. D. Microbial reduction of structural Fe3+ in nontronite by a thermophilic bacterium and its role in promoting the smectite to illite reaction. American Mineralogist 2007, 92 (8-9), 1411-1419.

Zhang, G. X.; Kim, J. W.; Dong, H. L.; Sommer, A. J. Microbial effects in promoting the smectite to illite reaction: Role of organic matter intercalated in the interlayer. American Mineralogist 2007, 92 (8-9), 1401-1410.

Zhang, J.; Dong, H. L.; Liu, D.; Agrawal, A. Microbial reduction of Fe(III) in smectite minerals by thermophilic methanogen Methanothermobacter thermautotrophicus. Geochim. Cosmochim. Acta 2013, 106, 203-215.

Zhang, J., Dong, H., Zeng, Q., Agrawal, A.. The role of Fe(III) bioreduction by methanogens in the preservation of organic matter in smectite. Chemical Geology 2014, 389, 16-28.

Zhao, L.; Dong, H.; Kukkadapu, R.; Agrawal, A.; Liu, D.; Zhang, J.; Edelmann, R. E. Biological oxidation of Fe(II) in reduced nontronite coupled with nitrate reduction by Pseudogulbenkiania sp. Strain 2002. Geochim. Cosmochim. Acta 2013, 119 (15), 231-247.

78 TABLE CAPTION

Table 1. Bio-reduction and bio-oxidation extents and initial rates

Extent (%) Rate (mmol·g-1·d-1)

PIPES Bic. PIPES Bic.

Reduced 26.3±0.1 27.4±0.2 0.53±0.0 0.77±0.0

Cycle 1 Oxidized 45.6±1.3 47.1±0.9 0.19±0.0 0.28±0.0

Reduced 23.0±0.3 18.3±0.8 0.49±0.1 0.63±0.1 Cycle 2 Oxidized 66.4±4.1 49.4±4.1 0.17±0.0 0.32±0.1

Reduced 21.3±0.6 14.2±0.8 0.32±0.1 0.60±0.1 Cycle 3 Oxidized 75.5±3.5 73.5±3.6 0.29±0.0 0.29±0.1

NOTE: PIPES = PIPES buffer; Bic. = Bicarbonate buffer. Extent (%) is the final extent of bio-reduction of structural Fe(III) or bio-oxidation of structural Fe(II) in NAu-2, e.g., at the end of 10 days, that were calculated according to the Eq. 1 & 2 in the main text. Rate (mmol·g-1·d -1) is the initial rate of bio-reduction or bio-oxidation (e.g., first 24 hours), that was calculated according to the Eq. 3 in the main text. Errors represent two standard deviations from two replicate experimental treatments.

79 Table 2. Fe L3 EELS peak position, the L3/L2 peak intensity ratios, and the comparisons of Fe(II) contents between the EELS and the chemical results.

Ferrous/Total Fe (%) L3 Peak Ratio of EELS Chemical (eV) L3/L2 result result

Cycle Reduced: 708.2 5.7 33.0±5.1 27.4±0.2

1 Oxidized 710.9 6.8 21.7±1.3 15.9±0.2

Cycle Reduced: 708.7 5.5 35.1±2.7 32.7±0.8

2 Oxidized 710.0 7.0 20.7±3.5 18.5±1.1

Cycle Reduced: 709.0 5.0 47.8±8.3 33.5±0.8

3 Oxidized 709.9 7.2 16.4±2.0 12.1±0.7

80 Table 3. Calculated Mössbauer parameters for the room temperature spectrum of the 2nd bio-reduced NAu-2 sample with AQDS.

2 3 4 5 <Δ> σd <Δ> Phase Phase1 mm/s mm/s mm/s %

Fe(III)-A 0.36 0.64 0.16 18

Fe(III)-B 0.34 0.15 0.1 31

Fe(II) in the Fe(II)-Fe(III) domain (i.e., Fe2.5+) 0.80 0.61 0.62 38

Fe(II) in the Fe(II)-Fe(II) domain-a 1.14 2.67 0.23 6

Fe(II) in the Fe(II)-Fe(II) domain-b 1.02 2.4 0.29 7

1Spectral component; 2Center shift; 3Quadrupole splitting (Δ); 4Standard deviation of quadrupole splitting; 5Spectral component

Lorentzian half widths at half maximum (HWHM) of all element doublets was fixed at 0.097 mm/sec; No coupling was allowed between CS and Δ ; the A+/A- areas of doublet are fixed at 1

81 FIGURE CAPTION

Figure 1. Time-course changes of concentrations of Fe(II) in NAu-2, nitrate, and nitrite through three redox cycles with bicarbonate (gray symbols) and PIPES (black symbols) buffers. (A) Time-course change of Fe(II) concentration (mmol/g) (square). (B)

Time-course changes of nitrate (circle) and nitrite (cross) (mM) concentrations. Here, the bioreduction extent was defined as the ratio of the net amount of Fe(II) production over the amount of Fe(III) present at the beginning of that cycle and may not be proportional to the amount of Fe(II) shown in panel A. The extents of Fe(III) reduction and Fe(II) oxidation of abiotic control groups were lower than 1%, and for the purpose of clarity, these curves were omitted from the figures.

Figure 2. EELS spectra of bioreduced and bio-oxidized NAu-2 samples showing L3 and L2 edges of structural Fe in NAu-2 as a function of increased Fe redox cycle.

Figure 3. Mössbauer spectra for the second bioreduced NAu-2 sample in the presence of AQDS at (A) RT and (B) 77 K, showing an increase in Fe(II) content from 14% at RT to 30% at 77 K. This increase of 16% is likely caused by the capture of a certain fraction of Mössbauer invisible Fe(II) upon cooling from RT to 77 K because at 77 K, electron hopping between Fe(II) and Fe(III) in Fe(II)-Fe(III) moieties is expected to be slower than Mössbauer detection time. Panels C and D show similar Fe(II) doublet at RT and 77

K spectra, illustrating the disappearance of the Fe(II)-Fe(III) domain upon biooxidation.

Interior Fe(II) content also decreased from 14% to 7% upon bio-oxidation.

Figure 4. Second derivative FTIR spectra of OH-stretching and OH-bending bands in

NAu-2 through three redox cycles. Panels A, B, C, and D represent the OH-stretching

82 vibrational bands of the unaltered, first, second, and third redox cycled NAu-2 samples, respectively. Panels E, F, G, and H represent the OH-bending vibrational bands of the unaltered, first, second, and third redox cycled NAu-2 samples, respectively.

Figure 5. Schematic diagrams of Fe(III) and Fe(II) sites in (A) unaltered, (B) biologically reduced (with AQDS), and (C) biologically reoxidized NAu-2. Panel B illustrates an Fe(II)-Fe(III) domain created from NAu-2 edge through bioreduction (i.e., corresponding to the broad feature in Figure 3A), and an interior Fe(II) domain within the

NAu-2 structure (i.e., corresponding to the distinct Fe(II) doublet in Figure 3A) created via bioreduction through basal plane as mediated by AQDS. Panel C illustrates oxidation of all edge Fe(II) and a fraction of interior Fe(II) .

Figure 6. Macroscopic color changes of NAu-2 through 3 Fe redox cycles. Bio- reduced (A, C, & E) and bio-oxidized (B, D, & F) NAu-2 from the 1st, 2nd, and 3rd cycles, respectively.

st Figure 7. Time-course change of N2 gas through the 1 redox cycle with bicarbonate

(grey bars) and PIPES (black bars) buffers.

Figure 8. X-ray diffraction patterns for NAu-2 samples through three microbial redox cycles in bicarbonate buffer. Three broad peaks correspond to d001, d002, and d005 reflections of nontronite. No broadened or shifted peaks appear, suggesting little change of the nontronite structure after three redox cycles. The peak with d-spacing of d002 =

0.32 nm (black arrow) is identified as albite based on standard (PDF 00-010-0393)

nd Figure 9. SEM&TEM images of biogenic albite (NaAlSi3O8) from the 2 bio-reduced

NAu-2 sample. (A) SEM image of albite and qualitative SEM-EDS spectrum(B). C peak

83 is derived from C coating. (C) TEM image of albite with the corresponding qualitative

TEM-EDS (D) and selected area electron diffraction (SAED) pattern (E). The small Ca peak in the EDS spectrum may have originated from adsorbed cation, and Cu peaks came from the TEM support grid. Scale bars denote 400 nm in both figures.

Figure 10. TEM images of the bio-oxidized NAu-2 samples from the 1st and 2nd redox cycle, which were selected as representatives for 6 samples from three cycles. (A)

TEM image showing the 1st bio-oxidized NAu-2 sample, including two areas with different layer spacing and its qualitative TEM-EDS spectrum; (B&C) High resolution

TEM lattice fringe images from the enlarged squared & circled areas in Fig. A, showing

1.3 nm layer spacing and 1.0 nm layer spacing, respectively; (D) High resolution TEM lattice fringe image of the 2nd bio-oxidized NAu-2 showing 1.1 nm layer spacing, and its qualitative TEM-EDS spectrum. Scale bar denotes 80 nm in Fig. A, and 20nm in Fig. B,

C, &D.

Figure 11. Calibration curve between the integral intensity ratio of L3/L2 and the

Fe3+/total Fe ratio using 4 standard minerals.

Figure 12. (A) Comparison of two room-temperature (RT) Mössbauer spectra for unaltered and bio-reduced NAu-2 showing the emergence of a distinct Fe(II) doublet and broadening of the central doublet after bio-reduction of structural Fe(III) in NAu-2. The central doublet in unaltered NAu-2 is ascribed to structural Fe(III), but its broadening after bio-reduction is ascribed to presence of an Fe(II)-Fe(III) domain. (B) A tentative model fit of the RT spectrum for the bio-reduced NAu-2 showing various contributions of Fe species. Mössbauer parameters are listed in Table 3.

84 Figure 13. Mössbauer spectra of biologically reduced NAu-2 in absence of AQDS at room temperature (RT; A) and 77 K (B) showing a broad signal without any distinct

Fe(II) doublet at RT and distinct doublets due to Fe(III) (82%) and Fe(II) (18%) at 77K, respectively. Experimental condition was identical to the one in Fig. 3A & B. The data in this figure suggests the absence of a distinct Fe(II)domain (as noted in Fig. 3A & B) within the interior of clay when the sample was bio-reduced without AQDS

85 FIGURE

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

CHAPTER 4 Fe(II) oxidation in illite coupled with nitrate reduction and its role in

clay mineral transformation

Linduo Zhao1, Hailiang Dong*1,2, Richard E. Edelmann3, Qiang Zeng2

Department of Geology Miami University Oxford, OH 45056 Tel: 513-529-2517 Fax: 513-529-1542 Email: [email protected]

99 ABSTRACT In pedogenic processes, clay minerals continuously transform from pre-existing phases into other clay phases, in a sequence of intermediate interstratified clays. Previous studies have emphasized temperature, pressure, chemical composition, climatic condition, and time as geological variables for clay mineral transformations, whereas the role of microbes was not recognized until ten years ago. The overall objective of the third study was to study biological nitrate-dependent Fe(II) oxidation in illite IMt-1 and the effects of bio-oxidation on clay mineral transformation. Our data demonstrated that Pseudogulbenkiania sp. strain 2002 was able to couple oxidation of structural Fe(II) in

IMt-1 with reduction of nitrate to N2 with nitrite as a transient intermediate. Fe(II)- oxidizing bacteria caused clay mineral structure change, and facilitated the illite→kaolinite and illite→smectite transformations. The biogenic smectite is a transient phase.

100 1. INTRODUCTION

Iron-bearing clay minerals are ubiquitous in natural sediments, soils, and sedimentary rocks (Stucki and Kostka, 2006; Dong et al., 2009). In pedogenic processes, they continuously transform from pre-existing clay minerals into other clay phases, in a sequence of intermediate interstratified clays (Moore and Reynolds, 1997). Previous studies have emphasized temperature, pressure, chemical composition, climatic condition, and time as geological variables for clay mineral transformations (Bethke and Altaner, 1986; Dong et al., 1997; Huang et al., 1993; Pevear, 1993), whereas the role of microbes was not recognized until ten years ago (Kim et al., 2004). Microbes commonly thrive in sedimentary environments, and it is able to change clay’s physical and chemical properties by dissolving and precipitating pre-existing minerals, oxidizing and reducing its structural Fe (Dong et al., 2009; Dong, 2012; Stucki, 2011). Although extensive research has been conducted on the important role of dissimilatory iron-reducing bacteria in driving smectite to illite conversion (Kim et al., 2004; Dong et al., 2009; Deng et al., 2011), the effect of iron-oxidizing bacteria on clay mineral transformation, despite their common co-existence in nature and their significance in the biogeochemical cycles of Fe, N, and C (Dong et al., 2009; Stucki, 2011), has not received equivalent attention.

Among all Fe(II)-oxidizing bacteria, anaerobic nitrate-dependent Fe(II) oxidizing bacteria have received extensive attention in the last two decades (Weber et al., 2006; Shelobolina et al., 2012; Miot et al., 2009; Kappler et al., 2005; Larese-Casanova et al., 2010), and significant progress has been made toward understanding the biogeochemical mechanisms of this process. Such bacteria have been found in various environments, including lakes, streams, hydrothermal vents, wetlands, and aquifer sediments (Hafenbradl et al., 1996; Straub et al., 1996; Chaudhuri et al., 2001; Weber et al., 2001, 2006a; Edwards et al., 2003; Shelobolina et al., 2003a, 2012b). These bacteria can utilize the following reaction (1) to provide energy to support their growth (Weber et al., 2006). Also, in this reaction, both aqueous Fe2+ and oxide-Fe(II) can be biologically oxidized (Weber et al., 2006; Miot et al., 2009; Kappler et al., 2005; Larese-Casanova et al., 2010; Chaudhuri et al., 2001; Weber et al., 2001), and more recent studies have demonstrated biological oxidation of structural Fe(II) in phyllosilicates, including biotite (Shelobolina

101 et al., 2003), nontronite, and illite-smectite mixed layers (Shelobolina et al., 2012), when this process is coupled with nitrate reduction. Other phyllosilicates, such as illite, despite their ubiquitous presence and distinct differences in physical/chemical property from nontronite (Moore and Reynolds, 1997), have not received equivalent attention.

2+ − + 10퐹푒 + 2푁푂3 + 24퐻2푂 → 10퐹푒(푂퐻)3 + 18퐻 + 푁2 (1)

Microbe-illite interactions could have important implications for iron biogeochemistry and geological processes in nature. In contrast to smectite (including nontronite) that is usually present in surface soils and sediments at shallow depths, illite is more common in consolidated sedimentary rocks at depth (Moore and Reynolds, 1997), where anaerobic microbial community and nitrate reduction zone are likely to be co- present with illite. Therefore, oxidation of structural Fe(II) in illite can be coupled with nitrate reduction by indigenous microbial community in these areas. In addition, previous studies have reported that illite could transform to I/S mixed-layer clay by leaching interlayer K+ under acidic or circumneutral environments (Rimmer, et al., 1982). However, microbial activity has never been considered as an environmental variable in promoting illite-smectite transformation. Microbially mediated smectite-illite reaction has been achieved through reduction of structural Fe(III) (Kim et al., 2004; Dong et al., 2009), and it is possible that bio-oxidation of structural Fe(II) can reverse the process, i.e., illite- to-smectite transformation. Together, nitrate-dependent Fe(II)-oxidizing bacteria may have a potential in promoting illite-smectite transformation and thereby close the microbial loop of clay mineral-microbe interactions, however, this pathway has never been explored. Additional studies are needed to better understand microbe illite interaction.

The overall objective of the present study was therefore to fill the above knowledge gap by studying biological nitrate-dependent Fe(II) oxidation in illite IMt-1 and the effects of bio-oxidation on clay mineral transformation. Specifically, the present study was designed to determine: (1) the extent and rate of biological Fe(II) oxidation in illite when coupled with nitrate reduction; (2) the mechanisms of biological Fe(II) oxidation in illite; (3) clay mineral transformation associated with this microbial process.

102 2. MATERIAL AND METHODS

2.1 Mineral, media, and experimental setup

Illite (IMt-1) was purchased from the Source Clays Repository of the Clay Minerals Society (West Lafayette, IN, USA). The total Fe content in IMt-1 was 12.3%, of which 10% was Fe(II) (Bishop, et al., 2011). The formula of IMt-1 is: 3+ 2+ (Ca0.01Na0.08K1.58)(Al2.78Fe 0.67Fe 0.08Mg0.47)(Si6.89Al1.11)O20(OH)4. The IMt-1 sample was size-fractionated prior to use as previously described (Zhang, et al., 2012). The 0.5-2 µm size fraction was separated, dried, and made into clay suspension in double distilled water (ddH2O) followed by autoclaving (Zhao, et al., 2013).

To study the mechanism for bio-oxidation of Fe(II) in illite, the experiment was performed on two groups of IMt-1 sample: original IMt-1 and bio-reduced IMt-1. To obtain bio-reduced IMt-1 sample, IMt-1 was reduced by Shewanella putrefaciens CN32 with lactate as electron donor as previously described (Jaisi et al., 2007). After Fe reduction reached its maximum (10%), lactate and cells were removed from the reduced clay suspension by washing it five times with sterile and anoxic (80:20 N2:CO2 purged) sodium bicarbonate buffer (29.76 mM NaHCO3 and 1.34 KCl, pH 7.0). Bio-reduced IMt- 1 was then resuspended with the same sterile and anoxic bicarbonate buffer with illite concentration of 13.3 g/L (corresponding to 0.22 mmol/g or 4.01mM structural Fe(II)). Original IMt-1 was suspended by bicarbonate buffer with illite concentration of 10 g/L (corresponding to 0.12 mmol/g or 2.14 mM structural Fe(II)). The pH of the original and bio-reduced IMt-1 suspensions were adjusted to 7.0 with NaOH or HCl as necessary, and then they were transferred to 50 mL serum bottles. To maintain an anoxic condition for bio-reduced IMt-1, the above procedures (i.e., washing followed by resuspension and pH adjustment) were processed in an anaerobic glove box (96% N2, 4% H2). The original

IMt-1 suspensions were purged with N2/CO2 (80:20) gas mix to obtain an anoxic condition. Both original and bio-reduced IMt-1 suspensions were pasteurized as previously described (Zhao, et al., 2015).

Bio-oxidation experiment was conducted using anaerobic autotrophic Fe-oxidizing bacterium Pseudogulbenkiania sp. strain 2002 (ATCC BAA-1479; DSM 18807) with nitrate as the sole electron acceptor and Fe(II) in NAu-2 as the sole electron donor. Strain

103 2002 was routinely cultured to the exponential phase and added into both original and bio-reduced IMt-1 suspensions to achieve a final cell concentration of ∼108 cells/mL.

The headspace of each serum bottle was purged with CO2 gas (a carbon source) using a sterile needle right after strain 2002 inoculation. A small volume of KNO3 solution was added as electron acceptor to the bottles to achieve a final concentration of 6-7 mM. Abiotic controls were the same as the treatments except that they were devoid of strain 2002 cells.

2.2 Analysis

The reaction progress was monitored over time by collecting samples from incubation bottles for the quantification of total Fe(II), nitrate, nitrite, and N2 during the experiment. The 1,10-phenanthroline method was applied to analyze total Fe(II) concentration (Amonette and Templeton, 1998). Nitrate and nitrite samples were collected from the supernatant following centrifugation of IMt-1 suspension (10,000 g for 10 min) and their concentrations were determined using a Dionex 500X ion chromatograph equipped with an Ion Pac AS22 analytical column. Gas samples were taken from the head space of the experimental serum bottles. The amounts of N2, NO, and N2O within each serum bottle were quantified by a Hewlett Packard 5890 series II Gas Chromatographic (GC) system equipped with a thermal conductivity detector.

X-ray diffraction (XRD) was performed on original and bio-oxidized IMt-1 after 3 month and 2 years of incubation to detect mineralogical changes of IMt-1 after bio- oxidation. Powder XRD patterns were collected and analyzed as previously described.

The structural changes of IMt-1 upon bio-oxidation, such as changes in the position of OH-stretching (νOH) vibration bands, were analyzed by Fourier transform infrared spectroscopy (FTIR) in the middle-infrared region. Samples were prepared and analyzed as previously described.

Scanning and transmission electron microscopy (SEM and TEM) were performed on both original and bioreduced IMt-1 samples that were bio-oxidized for different amount of time. Only results for original IMt-1 samples were presented in this paper, because no distinct difference between original and bioreduced IMt-1 samples was observed. The

104 original IMt-1 samples that were bio-oxidized for 1, 3, 6 months, and 2 years were selected for SEM and TEM observations, to study the bio-oxidation mechanism and clay mineral transformation.

For SEM analyses, clay suspensions were mounted onto glass cover slips, which were pretreated with poly-L-lysine for 10 min. After dehydration through grade series of ethanol followed by critical point drying with a Tousimis Samdro-780A Critical Point Dryer (CPD) (Dong et al., 2003b), the cover slips with clay particles were mounted onto SEM stubs via clear double-sided sticky tape, and then carbon-coated. The samples were then observed with a Zeiss Supra 35 variable pressure (VP) SEMwith EDAX Genesis 2000 X-ray energy dispersive spectroscopy (SEM/EDS) using 4-7 keV accelerating voltage and 8.5 mm working distance. The EDS spectra provided a primary means for mineral composition.

For TEM analyses, the clay suspensions (0.2 mL) were diluted by anoxic double- distilled water (ddH2O) with a dilution factor of 200, and pipetted onto 300 mesh copper grids with carbon-coated nitrocellulose membrane. The grids were dried overnight in an anaerobic glove box (96% N2, 4% H2). TEM imaging and analysis were performed with a JEOL JEM-2100 LaB6 TEM with a 200 KeV accelerating voltage. Selected area electron diffraction (SAED) patterns (GatanOrius SC200D camera) and energy dispersive spectroscopy (EDS, Bruker AXS Microanalysis Quantax200 with 4030 SDD detector) were obtained.

3. RESULTS

In contrast to minor changes in the abiotic controls, Fe(II) concentration in the inoculated experimental bottles of both original and bio-reduced IMt-1 decreased steadily with time, indicating that strain 2002 was able to oxidize structural Fe(II) in IMt-1 (Figure 1). To determine the effects of clay composition and structure on Fe bio- oxidation, Fe(II) bio-oxidation data in this illite experiment were compared with the dataset from a similar experiment of Fe(II) bio-oxidation in nontronite (NAu-2) as previously reported (Zhao, et al., 2015). Both Fe(II) bio-oxidation extent and rate for the bio-reduced IMt-1 were higher than those for the original IMt-1 (Table 1). In comparison

105 with nontronite, the Fe(II) bio-oxidation extent and rate for both the original and bioreduced IMt-1 were much lower than those for bio-reduced NAu-2 (Table 1).

In both original and bio-reduced IMt-1 treatment groups, 6.5 mM nitrate was completely reduced to N2 gas in 7 days (Figure 1), but Fe(II) was never fully oxidized, despite the fact that excess nitrate was added relative to Fe(II) (assuming a stoichiometric - - molar Fe/NO3 ratio of 5 if the N2 is the NO3 reduction product). Nitrite was produced as a transient intermediate during bio-oxidation in both original and bio-reduced IMt-1 treatment groups, up to 0.38 mM and 0.26 mM, respectively. More nitrite accumulation was observed in bio-reduced NAu-2 (1.5 mM) than in original IMt-1 (0.5 mM). Similar to our previous study (Zhao, et al., 2013; Zhao, et al., 2015), neither NO nor N2O was detected at any time point, maybe due to their reactive transient nature. The observed molar ratio of nitrate reduced to Fe(II) oxidized in original and bio-reduced IMt-1 were 31.27 and 4.89, respectively. They both greatly exceeded the theoretical values of 0.2 according to equation 1. Similar observations were made in other studies on nontronite and Fe(II) minerals, and suggest that heterotrophic nitrate reduction coupled with oxidation of dead bacterial biomass or energy reserves (i.e., glycogen) stored in cells (Weber, et al., 2001; Zhao, et al., 2013; Zhao, et al., 2015)

XRD patterns for the bio-oxidized, original IMt-1 showed that IMt-1 was the dominant mineral throughout the experiment with four broad peaks corresponding to d001, d002, d003 and d005 reflections of illite (Figure 2). XRD results also revealed that air-dried, bio-oxidized IMt-1 sample after 3 months and 2 years of incubation contained 7Å and 3.5 Å peaks that correspond to kaolinite (Figure 2). However, the 14 Å d-spacing peak, which could correspond to the smectite interlayers in an air-dried sample, was only found in the 3 month sample, and then disappeared upon further incubation to 2 years. The sample was further treated with ethylene glycol to confirm the presence of smectite. Unexpectedly, this 14 Å disappeared, probably because the neoformed mineral was poor crystalline and its periodicity was destroyed upon intercalation by ethylene glycol. (Figure 3).

FTIR spectra (Figure 4) for the abiotic control and bio-oxidized, original IMt-1 samples collected after 1 year of incubation displayed distinct difference in the position

106 of OH-stretching (νOH) vibration bands. In the OH-stretching region, while IMt-1 abiotic control sample showed one absorption peak at 3644 cm-1, which could be attributed to Al-Al-OH environment, this peak shifted to a lower wave number 3628 cm-1 in the bio- oxidized sample. A previous study showed that a band near this wave number (~3670 cm- 1) was produced by Al-Al-OH that linked the tetrahedral and octahedral sheets within one kaolinite unit gave (Madejová, 2003). Also, an absorption band at 3670 cm-1 was observed in IMt-1 abiotic control sample but this band shifted to 3656 cm-1 in bio- oxidized sample. In addition, a shoulder at 3699 cm-1 emerged in IMt-1 upon bio- oxidation. Both 3656 cm-1and 3699 cm-1 bands correspond to the OH groups between the octahedral and tetrahedral sheets in different units of a kaolinite sample (Madejová, 2003). In summary, the difference in Al-Al-OH bonding and OH group between IMt-1 abiotic control and bio-oxidized samples indicated that the crystal structure of IMt-1 transformed towards a kaolinite-like structure upon bio-oxidization.

Biogenic smectite and kaolinite in bio-oxidized, original IMt-1 sample were further confirmed by SEM images and energy-dispersive X-ray spectroscopy (EDS) analysis (Figure 5 and 6). After 3 months of incubation, 2 types of clay particles were detected in bio-oxidized IMt-1 sample under SEM (Figure 5) that exhibited different morphologies from original illite (Figure 5A and particle a in Figure 5B): platy particles connected by intertwined wires (particle b in Figure 5B and Figure 5D) and particles with a wavy texture (particle c in Figure 5B and Figure 5E). Qualitative SEM-EDS analyses revealed that these particles have a kaolinite-like and smectite-like composition, respectively (Figure 5C and Figure 5F). Particle b in Figure 5B and particle in Figure 5D were tentatively identified as a growing biogenic kaolinite, considering its diagnostic morphology, a high Al/Si ratio, and certain concentrations of Fe, Na, Mg, and K. Similarly, particle c in Figure 5B and particle in Figure 5E were tentatively identified as a growing biogenic smectite.

After 2 years of incubation, stacks of platy particles with a hexagonal shape emerged in bio-oxidized IMt-1 sample (particle b in Figures 6A and 6B). This shape was consistent with the morphology of well crystallized natural kaolinite. Furthermore, SEM- EDS analysis indicated that these particles were mainly composed of O, Si, and Al, with

107 an Al: Si ratio of 0.8 (Figure 6C), which was close to the stoichiometric ratio for kaolinite. Peaks of Na and K may have been derived from adjacent illite particles (Particle a in Figure 6A), or cations adsorbed on the surface of kaolinite. These lines of evidence strongly suggest that biogenic kaolinite formed as a result of bio-oxidation of structural Fe(II) in IMt-1. No smectite-like particle was observed in the 2 year sample. These SEM observations were in good agreement with XRD results.

TEM observations further confirmed XRD and SEM results. TEM lattice fringe images of the original IMt-1 sample showed illite particles in well-defined packets, and TEM-EDS analysis revealed a typical composition for illite, with relatively abundant Fe, K, and a high Al/Si ratio (Figure 7A and 7F). These data suggest that illite was the only clay mineral present in the original IMt-1 sample with no evidence of dissolution. However, two types of clay minerals were observed in the bio-oxidized IMt-1 sample after 2 weeks and 6 months of incubations: (1) platy particles that were similar in morphology and composition to the original IMt-1 sample (Figure 7B and 7C); (2) wavy fringes with 1.4-nm spacing, which are typical of smectite (Figure 7D and 7E). These wavy fringes were more susceptible to beam damage than the straight 1.0-nm illite fringes, possibly due to its poorly-crystalline nature. More smectite-like fringes were observed in the 6-month sample than in 2-week sample (Figure 7B and 7C). Wavy smectite-type fringes showed progressively decreased amounts of K and Fe, and increased Al/Si ratio relative to the original illite (Figure 7F). TEM observation on bio- oxidized IMt-1 sample after 2 years of incubation displayed only straight and continuous fringes, in contrast to the extensive presence of wavy smectite-like fringes in 2 week and 6 month samples. These observations suggest a transient nature of smectite as a result of bio-oxidation.

HRTEM images of bio-oxidized IMt-1 samples after 2 years of incubation showed a pseudo-hexagonal crystal (Figure 8A) with 0.7 and 1.0 nm lattice fringe spacing (Figure 8B and 8D) , which matched the dominant spacings from XRD pattern (Figure 2), and corresponded to d001 reflections of kaolinite and illite, respectively. Furthermore, fringes with 1.7 nm d-spacing were observed, and they suggest a mixed-layer illite-kaolinite. This mixed-phase was not observed in XRD pattern, possibly due to its low abundance.

108 Both 0.7 and 1.0 nm fringes were commonly observed in other areas of this sample (Figure 8D). TEM-EDS analysis of both areas revealed an increasing Al:Si ratio (1:1.7/1:1.9) (Figure 8C and 8E) in contrast to the ratio of 1:2.5 for the original IMt-1 sample (Figure 7F), consistent with contribution of kaolinite to the overall chemical composition. Cations Mg and K in the spectrum may have been derived from adjacent illite particles or illite-kaolinite mixed-layer in the sample.

HRTEM images of the original IMt-1 samples and bio-oxidized IMt-1 sample after 1 month incubation revealed illite structural alteration during bio-oxidation (Figure 9). The original IMt-1 samples showed typical tetrahedral-octahedral-tetrahedral layers (TOT) structure with 1 nm layer spacing (Figure 9A and 9B). However, HRTEM images of bio- oxidized IMt-1 sample after 1 month of incubation showed that a packet of illite with 1 nm layer spacing (Figure 9C) displayed a dissolution and structure change (Figure 9B, C and E) with an increased Si/Al ratio as opposed to the original IMt-1 sample (Figure 9D). In Figure 9E, the lateral transition occurred from the TOT structure (left white arrow) to the structure going through dissolution (middle black arrow) and further to the structure with changed tetrahedral layers.

4. DISCUSSION

4.1 Mechanisms of Fe(II) oxidation in illite

Our data demonstrated that Pseudogulbenkiania sp. strain 2002 was able to couple oxidation of structural Fe(II) in IMt-1 with reduction of nitrate to N2 with nitrite as a transient intermediate (Figure 1). The observation of non-stoichiometry of oxidized Fe(II) to reduced nitrate, as well as nitrite accumulation in IMt-1 treatment group are consistent with our previous study, and can be explained by heterotrophic nitrate reduction coupled with oxidation of energy reserves (i.e., glycogen) stored in cells. The stored energy in cells would outcompete Fe(II) as electron donor to denitrify, due to more negative redox 3+ 2+ potential (CO2/organic matter -290 mV vs Fe(OH) /Fe 0 mV) (Kappler, et al., 2005), thus leading to non-stoichiometrically high reduction of nitrate. Especially when there is less structural Fe(II) in IMt-1 than in NAu-2, the system in IMt-1 treatment group is unlikely to cause intermediate (i.e. nitrite) during denitrification, since the electron

109 transfer at more negative redox potential could expedite the relative reduction rates of nitrate into final nitrogen species.

The mechanism of bio-oxidation of structural Fe(II) in phyllosilicates has been poorly known. Our previous data on bio-oxidation of nontronite suggest that the electron transfer between strain 2002 and nontronite occurs via clay edges through direct contact between clays and bacterium (Ribeiro, et al., 2009; Zhao, et al., 2015). In comparison to nontronite, the bio-oxidation extent of Fe(II) in IMt-1 illite is much lower (Table 1). This type of difference between illite and nontronite has been observed in terms of bioreduction extent and rate (Bishop et al., 2014). In order for bio-oxidation to take place, there should be a direct contact between clay minerals and bacterium because strain 2002 is not known to produce electron shuttles or nanowires. Both clays and bacterium have negative layer charge, and illite owns higher negative layer charge but lower surface area than smectite (Moore and Reynolds, 1997). As a result, lower areal contact is expected to occur between strain 2002 cells and illite relative to smectite, which would eventually lead to a lower Fe(II) bio-oxidation extent in IMt-1 than in nontronite NAu-2.

In contrast to bioreduction where upon direct contact, both interlayer space and particle edges were involved in electron transfer process, and layer expandability is a main factor in determining the rate and extent of Fe bioreduction in both smectite and illite, the interlayer space (Bishop, et al., 2014) is not directly involved in Fe(II) oxidation. The non-expandable nature of illite interlayer should not be responsible for the lower bio- oxidation extent of illite relative to smectite. Our previous data suggest that bio-oxidation of Fe(II) to Fe(III) in nontronite proceeded from edges to interior structure (Zhao, et al., 2015), so bio-oxidation extent should not be affected by interlayer expandability. Two additional lines of evidence support this speculation: 1) Neither IMt-1 nor NAu-2 showed a complete oxidation of structural Fe(II) (Figure 1), probably because much of the Fe(II) domain within the clay structure could not be reached by oxidizing bacteria. 2) Our data indicate a lower Fe(II) bio-oxidation extent of original IMt-1 relative to bio-reduced IMt- 1 (Table 1). This difference is likely due to less Fe(II) on the edges of the original IMt-1 relative to the bioreduced IMt-1, because bioreduction of Fe(III) to Fe(II) would produce Fe(II) on clay edges. If Fe bio-oxidation takes places from clay edges, oxidation extent of

110 bio-reduced IMt-1 should be higher than that for the original IMt-1, which is supported by our experimental data.

4.2 Microbially mediated clay mineral transformation

In our experiment, Fe(II)-oxidizing bacteria facilitated the illite→kaolinite (Figure 2, 4, 5, 6, and 8) and illite→smectite transformations (Figure 2, 3, and 7). The biogenic smectite is a transient phase (Figure 7): it emerged after 3 and 6 months of bio-oxidation, but disappeared with extended incubation for 2 years. The disappearance of the biogenic smectite in the 2-year sample suggests that the smectite either transformed into kaolinite or dissolved with longer incubation. Our aqueous chemistry data for the 2 year sample revealed no significant increases of Si, Al, and Fe in solution (data not shown), and TEM and XRD data indicated no precipitates of biogenic silica and/or Fe/Al (oxy)hydroxides. These lines of evidence support the possibility of smectite→kaolinite conversion. Indeed, in nature, smectite has been found to be unstable in humid areas, and would transform into kaolinite through mixed-layer K/S.

Specific mechanisms for the illite→kaolinite or illite→smectite transitions as a result of microbial activity remain elusive. Based on the previous research on the microbially catalyzed smectite→illite reaction, as well as our HRTEM observation of structure dissolution and tetrahedral layers change upon bio-oxidation (Figure 9), a combination of dissolution-precipitation and solid-state mechanisms may have been responsible for the illite→kaolinite reaction. In this mechanism, illite transformed to kaolinite through several steps:

1) Initial dissolution of the illite structure involves removal of a tetrahedral sheet from each TOT unit, forming a TO structure but still with 1.0 nm spacing. TEM observations revealed that some illite particles were going through dissolution at the beginning phase of bio-oxidation (Figure 9). Dissolution of clay minerals by microbial activities has been observed in many studies with various types of secondary mineral formation. However, this is the first time that a TOT unit of illite with structural change on tetrahedral sheet, as well as going through a dissolution upon Fe(II) bio-oxidation was observed. This observation is consistent with the result of a previous study, in which an illite crystal

111 experienced along-layer dissolution upon bio-reduction (Dong, et al., 2003). An inequivalent charge distribution of two tetrahedral sheets in one illite TOT unit might be the cause. It has been speculated that, within an illite TOT unit, the tetrahedral sheet that faces the outside world have a lower negative charge than the other tetrahedral sheet (Dong, et al., 2003). As a result, this tetrahedral sheet with the reduced layer charge might be more easily dissolved by microbial activity when comparing with the other sheet.

2) As a result of the first step, a gap between the octahedral sheet and interlayer space is created. These gaps may be connected laterally along the layers and would provide pathways for fluid to move through illite layers. Under favorable condition, i.e., the increased layer charge in the octahedral sheet, as a result of bio-oxidation of structural Fe(II), would expel interlayer cation to balance the charge. Subsequently, a TOT structure of illite would transform into a TO structure.

3) The removal of the tetrahedral sheet and the interlayer cations would cause a layer spacing collapse from 10 to 7 Å-spacing, characteristic layer spacing for kaolinite. SEM observations support this mechanism in which illite to kaolinite conversion occurs via different stages. The distinctly different morphologies of kaolinite particles, as well as an gradual increasing Al:Si ratio are observed from bio-oxidized IMt-1 samples after months and 2 years of incubation (Figure 5 and 6), and they may indicate an illite to kaolinite transformation process from dissolution and collapse to re-precipitation and crystallization.

The mechanism for the illite→smectite transition does not require the removal of a tetrahedral sheet, but instead just explusion of interlayer cations due to Fe(II) oxidation to Fe(III). As a result, the attractive force between two tetrahedral sheets sandwiching interlayer cations should decrease and interlayers would thus become expandable and wavy, which are typical of smectite. Eventually, smectite could transform into kaolinite through mixed-layer K/S by a solid-state mechanism as previously reported (Amouric and Olives, 1998).

4.3 Geological implication and future study

112 Microbes should be considered as an important geological variable for the study of clay mineral diagenesis. Various mixed-layered clay minerals are ubiquitous in soils and sediments, e.g. illite/smectite, chlorite/smectite, kaolinite/smectite, and etc., and the clay mineral conversion from one end member to another usually proceeds through mixed- layered clay intermediates. Our results suggest that nitrate-dependent Fe(II)-oxidizing bacteria can promote the illite to kaolinite conversion with smectite as a transient phase. This conversion is consistent with many mineral reactions in natural environments, e.g. conversion of smectite to kaolinite with advancement of a weathering front, or illite weathering to vermiculite at pH<6 (Han, et al., 2014). The conversion of illite to smectite has been observed in natural environments and this reaction is usually favored at pH>6, which is similar to the pH of this experiment. Iron-oxidizing bacteria are likely to play a substantial role in illite to kaolinite conversion in natural soil environments, such as those from Jiujiang, China, where sediments have formed along Yangzi River in subtropical area.

An early study demonstrated that microbial reduction of structural Fe(III) in smectite accelerated the smectite-illite reaction at room temperature in two weeks (Kim et al., 2004). This reaction typically requires 300-350oC, 100 megapascals, and 4 to 5 months in the absence of microbial activity. That study and all other follow-up studies are important because this smectite-to-illite reaction is typically used as an index to infer paleotemperature, based on the fact that smectite is a low temperature mineral, and with increased sediment burial and temperature, smectite should gradually transform to illite. However, with microbial activity, the inferred paleo-temperature will be inaccurate and may be much higher than actual temperature. In this study, we have demonstrated that microbial oxidation of structural Fe(II) in illite can reverse the reaction, e.g., illite to smectite transformation. In this case, the inferred paleo-temperature is again inaccurate, but in opposite way, i.e., the inferred temperature will be lower than actual temperature.

113 REFERENCE

Amonette, J. E. and Templeton, J. C., 1998. Improvements to the quantitative assay of

nonrefractory minerals for Fe(II) and total Fe using 1,10-phenanthroline. Clays

Clay Miner. 46, 51-62.

Bishop, M. E.; Glasser, P.; Dong, H.; Arey, B.; Kovarik, L. Reduction and

immobilization of hexavalent chromium by microbially reduced Fe-bearing clay

minerals. Geochim. Cosmochim. Acta 2014, 133 (0), 186-203.

Dong, H.; Jaisi, D. P.; Kim, J.; Zhang, G. Microbe-clay mineral interactions. American

Mineralogist 2009, 94 (11-12), 1505-1519.

Dong, H. L. Clay-Microbe interactions and implications for environmental mitigation.

Elements 2012, 8 (2), 113-118.

Jaisi, D. P.; Dong, H. L.; Liu, C. X. Kinetic analysis of microbial reduction of Fe(III) in nontronite. Environ SciTechnol 2007, 41 (7), 2437-2444.

Kappler, A.; Schink, B.; Newman, D. K. Fe(III) mineral formation and cell encrustation

by the nitrate-dependent Fe(II)-oxidizer strain BoFeN1. Geobiology 2005, 3 (4),

235-245.

Kim, J.; Dong, H.; Seabaugh, J.; Newell, S. W.; Eberl, D. D. Role of microbes in the

smectite-to-illitereaction. Science 2004, 303 (5659), 830-832.

Larese-Casanova, P.; Haderlein, S. B.; Kappler, A. Biomineralization of lepidocrocite

and goethite by nitrate-reducing Fe(II)-oxidizing bacteria: Effect of pH,

114 bicarbonate, phosphate, and humic acids. Geochim. Cosmochim. Acta 2010, 74

(13), 3721-3734.

Madejova J., 2003. FTIR techniques in clay mineral studies. Vibrational Spectroscopy.

31, 1-10.

Miot, J.; Benzerara, K.; Morin, G.; Bernard, S.; Beyssac, O.; Larquet, E.; Kappler, A.;

Guyot, F. Transformation of vivianite by anaerobic nitrate-reducing iron-oxidizing

bacteria. Geobiology 2009, 7 (3), 373-384.

Moore., D. M. and Robert C. Reynolds, J., 1997. X-Ray Diffraction and the Identification

and Analysis of Clay Minerals. Oxford University Press, USA, 1997, New York.

Ribeiro, F. R.; Fabris, J. D.; Kostka, J. E.; Komadel, P.; Stucki, J. W. Comparisons of

structural iron reduction in smectites by bacteria and dithionite: II. A variable-

temperature Mössbauer spectroscopic study of Garfield nontronite. Pure and

Applied Chemistry 2009, 81 (8), 1499-1509.

Shelobolina E.; Konishi H.; Xu H.; Benzine J.; Xiong M.; Wu T.; Blöthe M.; Roden E.

Isolation of phyllosilicate–iron redox cycling microorganisms from an illite–

smectiterich hydromorphic soil. Frontiers in Microbiology 2012, 3.

Shelobolina, E.; Xu, H. F.; Konishi, H.; Kukkadapu, R.; Wu, T.; Blothe, M.; Roden, E.

Microbial lithotrophicoxidation of structural Fe(II) in biotite. Appl Environ

Microbiol. 2012, 78 (16), 5746-5752.

Stucki, J. W.; Kostka, J. E. Microbial reduction of iron in smectite. ComptesRendus

Geoscience 2006, 338 (6-7), 468-475.

115 Stucki, J. W. A review of the effects of iron redox cycles on smectite properties.

ComptesRendus Geoscience 2011, 343 (2-3), 199-209.

Weber, K. A.; Picardal, F. W.; Roden, E. E. Microbially catalyzed nitrate-dependent

oxidation of biogenic solid-phase Fe(II) compounds. Environ SciTechnol 2001, 35

(8), 1644-1650.

Weber, K. A.; Achenbach, L. A.; Coates, J. D. Microorganisms pumping iron: anaerobic

microbial iron oxidation and reduction. Nat. Rev. Microbiol. 2006, 4 (10), 752-764.

Zhao, L.; Dong, H.; Kukkadapu, R.; Agrawal, A.; Liu, D.; Zhang, J.; Edelmann, R. E.

Biological oxidation of Fe(II) in reduced nontronite coupled with nitrate reduction

by Pseudogulbenkiania sp. Strain 2002. Geochim. Cosmochim. Acta 2013, 119

(15), 231-247.

116 TABLE CAPTION

Table 1. Fe(II) concentration before and after bio-oxidation, bio-oxidation extents and rates

Bio- Initial Fe(II) Final Fe(II) Bio-oxidation oxidation concentration concentration rate (5 days) extent

mmol/g % mmol/g·day

Illite Original 0.20 0.18 8.65 0.003

(IMt-1) Reduced 0.31 0.25 20.08 0.01

Nontronite Original

(NAu-2) Reduced 1.10 0.53 51.75 0.11

NOTE: Bio-oxidation extent was calculated according to the equation 퐹푒(퐼퐼) −퐹푒(퐼퐼) 푂푥푖푑푎푡푖표푛 푒푥푡푒푛푡 = 푖푛푖푡푖푎푙 10 푑푎푦푠 (1). Bio-oxidation rate was calculated 퐹푒(퐼퐼)푡표푡푎푙 |퐹푒(퐼퐼) −퐹푒(퐼퐼) | according to the equation 퐵푖표 − 표푥푖푑푎푡푖표푛 푟푎푡푒 = 5 푑푎푦푠 푖푛푖푡푖푎푙 (2). Bio- 훥푡푖푚푒 oxidation of reduced nontronite data was from the previous research.

117 FIGURE CAPTION

Figure 1. Time-course changes of concentrations of Fe(II) in IMt-1 and NAu-2, nitrate, and nitrite produced from the nitrate-dependent Fe(II) oxidation using strain 2002. The two Figures on the left (A & C) represent the Fe(II) concentration, and the two figures on the right (B & D) represent the nitrate and nitrite concentrations. The bio-oxidation data for reduced nontronite were from the previous research (Zhao et al., 2013) for a comparison purpose.

Figure 2. X-ray diffraction patterns of A) air-dried original IMt-1 sample showing illite

(I) only with four peaks corresponding to d001, d002, d003 and d005 reflections of standard illite, B) air-dried bio-oxidized IMt-1 samples after 3 months of incubation showing a 14

Å d-spacing peak that corresponds to d001 reflection of smectite (S), and 7 Å and 3.5 Å d- spacing peaks that correspond to d001 and d002 reflections of kaolinite (K); C) air-dried bio-oxidized IMt-1 samples after 2 years of incubation (C), showing only 7 Å and 3.5 Å d-spacing peaks.

Figure 3. X-ray diffraction patterns of air-dried (A) and ethylene glycolated (B) bio- oxidized IMt-1 sample after 3 months of incubation, showing that 14 Å d-spacing peak was present in air-dried sample, but absent in ethylene glycolated sample

Figure 4. Second derivative FTIR spectra of OH-stretching bands in the abiotic control (black) and bio-oxidized (gray) samples after 1 year of incubation.

Figure 5. SEM images showing neoformed biogenic kaolinite and smectite as a result of bio-oxidation of structural Fe(II) in 1Mt-1 after 3 months. (A) Original IMt-1 sample. (B) bio-oxidized IMt-1 sample after 3 months of incubation, showing illite particles (a), neoformed kaolinite (b), and smectite particles (c). (C) The corresponding qualitative SEM-EDS spectrum for the above particles.

Figure 6. SEM images of neoformed biogenic kaolinite from the bio-oxidized IMt-1 sample after 2 years of incubation. (A) SEM images of biogenic kaolinite particle with a pseudo-hexagonal shape (b) next to illite particles (a). (B) A magnified image of the

118 biogenic kaolinite particle from Figure A. (C) The corresponding qualitative SEM-EDS for particles a and b in image A, that are characteristic of illite and kaolinite, respectively.

Figure 7. HRTEM images of the neoformed biogenic smectite produced in bio-oxidized IMt-1 sample. (A) TEM image of the original IMt-1 sample, showing illite particle in well-defined pack. (B) TEM images of the bio-oxidized IMt-1 sample after 2 weeks of incubation, showing a packet of wavy smectite-like layers occurred in the illite matrix. (C) TEM images of the bio-oxidized IMt-1 sample after 6 months of incubation, displaying more abundant wavy smectite layers than the 2-week sample (B). (D & E) High resolution TEM lattice fringe images from the enlarged white and black squared areas in (C), showing both lath-shaped, 1.0 nm layers (arrow) and wavy layers (white line) (D), as well as wavy 1.4 nm fringes (E). (F) The corresponding qualitative TEM- EDS for (A), (B), and (C).

Figure 8. HRTEM images of bio-oxidized IMt-1 samples after 2 years of incubation, showing the neoformed biogenic kaolinite particles. (A) TEM image showing a stack of pseudo-hexagonal crystals with a diameter of 0.2 µm. (B) High resolution TEM lattice fringe image from the square area in (A), showing 0.7, 1.0, and 1.7 nm layer spacings corresponding to kaolinite, illite, mixed-layer illite-kaolinite, respectively. (C) The corresponding qualitative TEM-EDS for (A). (D) High resolution TEM lattice fringe images showing the co-existence of 0.7 and 1.0 nm layer spacings in other areas of the sample, corresponding to kaolinite and illite, respectively. (E) The corresponding qualitative TEM-EDS for (D)

Figure 9. HRTEM images showing structure changes of IMt-1 illite as a result of bio- oxidation for 1 month. (A) High resolution TEM lattice fringe image of the original IMt- 1 sample, showing 1 nm layer spacing. (B) High resolution TEM lattice fringe images from the enlarged square areas in (A), displaying tetrahedral-octahedral-tetrahedral sheets in illite structure. (C) High resolution TEM lattice fringe image of the bio-oxidized IMt-1 sample after 1 month of incubation, showing a packet of illite, with 1 nm layer spacing, but different clay structure (E) and the corresponding qualitative TEM-EDS (F) for (A) and (C).

119

Figure 1

120

Figure 2

121

Figure 3

122

Figure 4

123

Figure 5

124

Figure 6

125

Figure 7

126

Figure 8

127

Figure 9

128 CHAPTER 5

CONCLUSIONS AND FURTHER RECOMMENDATIONS

This dissertation consists of three main projects:

1. The comprehensive understanding of the coupled process of microbial oxidation of Fe(II) in clay mineral nontronite (NAu-2), and nitrate reduction by Pseudogulbenkiania species strain 2002, and to determine mineralogical changes associated with this process.

2. The investigation of microbially mediated redox cycles of Fe in nontronite (NAu-2), and its potential application in nitrate removal.

3. The study of microbially mediated nitrate-dependent Fe(II) oxidation in illite IMt-1 and the effects of bio-oxidation on clay mineral transformation.

CONCLUSIONS

The first subproject was designed to understand bio-oxidation of Fe(II) in reduced NAu-2 coupled with nitrate reduction by Pseudogulbenkiania sp. strain 2002. Multiple analytical techniques were employed to reach the following conclusions:

1. Pseudogulbenkiania sp. strain 2002 was able to couple oxidation of Fe(II) in reduced

NAu-2 with reduction of nitrate to N2, via an intermediate product nitrite.

2. The ratio of nitrate reduced to Fe(II) oxidized was in excess of the stoichiometric ratio of 5, and this can be explained by heterotrophic nitrate reduction coupled with oxidation of energy reserves (i.e., glycogen) stored in Pseudogulbenkiania sp. strain 2002 cells.

3. The Fe(II) oxidation occurred largely in solid state with only a small amount of ferrihydrite and magnetite formation.

4. The chemical data, Mössbauer spectroscopy, and SEM/TEM observations suggested that the mechanisms for the nitrate-dependant oxidation of structural Fe(II) in clay mineral nontronite were complex, involving both direct (possibly enzymatic oxidation) and indirect oxidation through surface-sorbed aqueous Fe3+or solid Fe(III).

129 The second subproject was to understand the microbially mediated redox cycles of Fe in nontronite (NAu-2). Three Fe redox cycles of biological dissimilatory Fe(III) reduction and nitrate-dependent Fe(II) oxidation were performed by dissimilatory Fe(III)-reducing bacterium Shewanella putrefaciens CN32 and Pseudogulbenkianiasp. strain 2002, respectively. Multiple analytical techniques were employed to reach the following conclusions:

5. Structural Fe in NAu-2 was able to undergo multiple microbially mediated redox cycles without significant dissolution.

6. Fe(II) in bioreduced samples occurred in two distinct environments, at edges and in the interior of the NAu-2 structure.

7. Electron transfer between strain 2002 and nontronite occurs via clay edges through direct contact between clays and bacterium. Nitrate reduction to nitrogen gas was coupled with oxidation of edge-Fe(II) and part of interior-Fe(II), and its extent and rate did not change with Feredox cycles.

The third subproject was to investigate the microbially mediated Fe(II) oxidation in illite coupled with nitrate reduction and its role in clay mineral transformation. Main results are summarized below:

8. Pseudogulbenkiania sp. strain 2002 was able to couple oxidation of structural Fe(II) in

IMt-1 with reduction of nitrate to N2 with nitrite as a transient intermediate

9. Higher negative layer charge and lower surface area of illite should be responsible for the lower bio-oxidation extent of illite relative to smectite.

10. Illite particles went through dissolution as a result of microbially mediated bio- oxidation.

11. Fe(II)-oxidizing bacteria facilitated the illite→kaolinite and illite→smectite transformations. The biogenic smectite is a transient phase during the clay mineral transformations.

FURTHER RECOMMENDATIONS

130 Although this research has made a number of achievements in furthering our understanding of microbially mediated Fe(II) oxidation in clay minerals coupled with nitrate reduction, and Fe redox cycling in clay minerals, there are, however, a few important questions remaining for future research:

1. What is the electron donor for biological reduction of nitrate that caused non- stoichiometric ratio of Fe(II) oxidation over nitrate reduction, energy reserve within activecells (i.e., glycogen) or biomass of inactive cells?

2. Are there NO and N2O produced as transient intermediates at the beginning phase of Fe(II) bio-oxidation?

3. Which location within cell does the Fe(III) mineral precipitate occur as a result of bio- oxidation of Fe(II) coupled with nitrate reduction? Cell surface, cell membrane, cytoplasm, or extracellularly ?

4. How does the mode of syntrophic metabolism, i.e., co-culture of Fe-reducing and - oxidizing bacteria, via clay minerals as an environmental medium, affect the nitrate removal?

5. Experiments to reduce NAu-2 to different extents (with and without AQDS) are needed for further insights into NAu-2 bioreduction mechanisms.

6. Can Pseudogulbenkiania sp. strain 2002 immobilize heavy metals, such as As?

7. What is the correlation between Fe(II)/Fe(III) ratio and illite/smectite ratio in clay minerals?

131