Transport and activity of dissimilatory metal-reducing bacteria in porous media for the remediation of heavy metals and chlorinated hydrocarbons by Robin Gerlach A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Engineering Montana State University © Copyright by Robin Gerlach (2001) Abstract: Dissmiilatory metal-reducing bacteria (DMRB) are capable of reducing a wide range of metals and the injection of DMRB is considered for the cleanup of contaminated soil and groundwater. The success of such an approach, among other factors, relies upon the bacterial transport through porous media and the contaminant transformation by DMRB. Long term starvation of Shewanella algae BrY, the model DMRB used in this research, allows for significantly improved transport through quartz sand porous media columns. However, the reasons for the improved transport remain unclear. Changes in cell size, net electrostatic charge, hydrophobicity, buoyant density, and effective diffusion coefficient do not provide a sufficient explanation and this question will have to remain the focus of future research. In addition, it became evident that the employed mathematical model based on the colloid filtration theory is not capable of appropriately describing scale and physiology dependent effects on bacterial transport through porous media. A complete understanding of the processes and factors influencing the porous media transport of starved bacteria is still lacking and will remain as the focus of future research. The reduction of chromium [Cr(VI)], a compound of significant environmental concern at many Department of Energy, Department of Defense, and Environmental Protection Agency Superfund sites, can be facilitated by DMRB. Starved S. algae BrY cells can be resuscitated into an actively metabolizing state in the presence of ferric iron [Fe(III)], which is abundant in the environment. Active S. algae BrY cells can either directly reduce Cr(VI) or produce surface reactive ferrous iron [Fe(II)]. Fe(II) chemically reduces and precipitates Cr(VI) eliminating it from contaminated water. S. algae BrY cells can also potentially contribute towards the long term reactivity of zero valent iron subsurface barriers. The microbial reduction of Fe(III) allows for the removal or activation of surface associated corrosion products and results in increased transformation rates of carbon tetrachloride, another widespread environmental contaminant. The results summarized in this dissertation indicate that the injection of starved DMRB is a promising technology for the remediation of subsurface environments contaminated with heavy metals and chlorinated hydrocarbons. TRANSPORT AND ACTIVITY OF DISSIMILATORY METAL-REDUCING BACTERIA IN POROUS MEDIA FOR THE REMEDIATION OF HEAVY METALS AND CHLORINATED HYDROCARBONS by Robin Gerlach A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Engineering I MONTANA STATE UNIVERSITY Bozeman, Montana April 2001 © COPYRIGHT by Robin Gerlach 2001 All rights reserved Ii APPROVAL of a dissertation submitted by Robin Gerlach This dissertation has been read by each member of the dissertation committee and has been found to be satisfactory regarding content, English usage, format, citations, bibliographic style, and consistency, and is ready for submission to the College of Graduate Studies. /a - Dr. Donald A. Rabem (Signature) j Cl (Date) Approved for the College of Graduate Studies Dr. Bruce R. McLeod (Date) Ni STATEMENT FOR PERMISSION TO USE In presenting this dissertation in partial fulfillment of the requirements for a doctoral degree at Montana State University, I agree that the Library shall make it available to borrowers under the rules of the Library. I further agree that copying of this dissertation is allowable only for scholarly purposes, consistent with “fair use” as prescribed in the U.S. Copyright Law. Requests for extensive copying or reproduction of this dissertation should be referred to Bell & Howell Information and Learning, 300 North Zeeb Road, Ann Arbor, Michigan 48106, to whom I have granted “the exclusive right to reproduce and distribute my 'dissertation in- and from microform along with the non-exclusive right to reproduce and distribute my abstract in any form in whole or in, part.” Signature L lg; 2oot iv ACKNOWLEDGEMENTS I thank my advisors Al Cunningham and Frank Caccavo Jr. for their support and interest in my research project. Their guidance, encouragement, and breadth of knowledge have contributed significantly to the completion of this dissertation and my development as a researcher. Thanks also to the remaining committee members Anne Camper, Marty Hamilton, and Warren Jones who were always available for helpful discussions. My appreciation goes to all students, staff, and faculty of the Center for Biofilm Engineering (CBE) at Montana State University (MSU) who made the past four years of my life extremely interesting and enriching. Special thanks goes to John Neuman, our laboratory manager, without whom the CBE would only be half of what it is now and to Kathy Lange, one of the most sincere people I have ever met. Following former and current CBE students deserve special mentioning: Stephanie Douthit, Patricia Stoots, Julie Eyre, Chris Wend, and John Komlos who have all played an important role in my scholarly and personal development. ' , The assistance of following MSU undergraduate students is acknowledged: Jennifer Nyman, Kara Boettcher, and Christina McKinstry. I also want to thank the Characklis family: Bill Characklis for creating a vision and founding the CBE in Bozeman, Montana; Nancy Characklis and John Austin for adopting me as their surrogate son and for always being there for me; and lastly Greg Characklis (and Joe Hughes) who encouraged me to seek out the CBE for my dissertation work. May I also not forget the care and support which I received from Al Cunningham’s wife Sam. My parents, however, deserve the biggest thanks. Without their love and care, their children and grandchildren would have not become the caring and intelligent people they are today. This work was supported by the Center for Biofilm Engineering at Montana State UniversityBozeman, a National Science Foundation supported Engineering Research Center (NSF Cooperative Agreement. EEC-8907039). The hospitality of the Microbiology Department at the University of New Hampshire and the Department of Chemical Engineering at Washington State University is acknowledged. TABLE OF CONTENTS 1. INTRODUCTION...........................................................................................................••........................I Bacterial Transport Through Porous Media for Bioaugmentation (Research Goal I ) ............................ 4 Parameters Influencing Bacterial Transport Through Porous Media............................................... 5 Influence of Solution Chemistry on Bacterial Transport Through Porous Media..................... 5 Influence of Porous Media Properties and Hydraulic Conditions on Bacterial Transport Through Porous M edia............................................................... - ....................................-6 Influence of Cell Properties on Bacterial Transport through Porous Media.............................7 Implications for Bioaugmentation Strategies......................... ..........................................................9 Bacterial Starvation as a Transport Enhancement Strategy............................................................ 10 Permeable Reactive Subsurface Barriers and Microbial Metal-Reduction (Research Goal 2)...............12 Research Goals...................................................................................................................................... 16 Objectives....................................................................................................... 17 2. ENHANCING TRANSPORT OF SHEWANELLA ALGAE BrY THROUGH POROUS MEDIA BY STARVATION...... ........................................................................................................... 18 Abstract.............. ....................................................................................................................................18 Introduction.... ....................................................................................................................................... 18 Quantification of Bacteria in Porous Media...................................................................................19 Materials and Methods.... .......................................................................................................................21 Strain and Culturing Methods........................................................................ 21 Transport studies............................................................................................................................. 21 Porous Media Sampling.................................................................................................................. 23 Modeling........................................................................................................................................ 23 Results and Discussion..........................................................................................................................26 Short (30 cm) Column Experiments.............................................................................................. 26 Breakthrough of Bacteria......................................................................................................... 27 Distribution of Bacteria in the Pore Water............................................................................. 28 Sorption of Bacteria Along the Flowpath................................ ................................................29 Recovery of Bacteria from Columns....................................... ................................................31 Large (3 m) Column Experiments................................................................................................. 33 Breakthrough of Bacteria....................................................................................................... 34 Distribution of Bacteria in the Pore Water........................................ ......................................35 Sorption of Bacteria Along the Flowpath............................................................................... 37 Recovery of Bacteria from Columns...................................................................................... 38 Modeling and Scale Up Issues...................................................................................................... 39 Influence of Influent Cell Concentration and Fluid Velocity........................................................40 Conclusions......................................... :.................................................................................................41 Additional Research Needs.................... ........................................................................................ 42 3. CHANGE OF TRANSPORT-RELATED CELL PROPERTIES DURING STARVATION OF SHEWANELLA ALGAE BrY......... .................................................................................................. 45 Abstract....................... .......................................................................................................................... 45 Introduction............................................................................................................................................ 45 Materials and Methods.........................................................................................................................48 Strains and Culturing Methods......................................................................................................48' vi Cell Quantification and Cell Imaging.............................................................................................49 Hydrophobicity.................... .....................1.................................................................................... 49 Zeta Potential and Diffusion Coefficient......................................................................................... 50 Buoyant Density.............................................................................................................................. 50 Adhesion Assay............................................................................................................................... 50 Transportability Through Porous Media......................................................................................... 51 Results and Discussion............................................................................. :............................................ 51 Cell Quantification.......................................................................................................................... 52 Hydrophobicity and Zeta Potential.................................................................................................. 53 Diffusion Coefficient and Buoyant Density.................................................................................... 55 Modeling Results............................................................................................................................ 55 Influence of Diffusion Coefficient and Buoyant Density, Calculation of Tj.............................55 Calculation of Collision Efficiencies a ................... 57 Applicability of Filtration Model to 3 m Column Results...............................................................58 Batch Adhesion Assays as a Predictive Tool.................................................................................. 59 Conclusions.............................................................................................................................................61 4. BIOGEOCHEMICAL ELIMINATION OF CHROMIUM(VI) FROM CONTAMINATED WATER...................................................................................................................................................64 Abstract........................ 64 Introduction........................................................................................... 64 Materials and M e th o d s............................. ................................................... :...................................... 67 Experimental............................. 67 Test Organism and Culture Methods....................................................................................... 67 Amorphous Iron Coating.......................................................................................................... 68 Batch Experiments, Iron Reduction.........................................................................................68 Batch Experiments, Chromium Reduction...............................................................................69 Regeneration of Ferrous Iron in the Presence of Chromium Precipitates and Repeated Chromium Reduction........................................................................................................ 69 Column Experiments................................................................................... 70 Analytical........................................................................................................................................ 71 Iron Quantification............................................................. i....................................................71 Chromium Quantification........................................................................................................ 72 Results and Discussion....!...................................................................................................................... 73 Microbial Fe(III) Reduction............................................................................................................ 73 Cr(VI) Reduction by Microbially Produced Fe(II) ...................................................... 75 Regeneration of Surface Reactive Ferrous Iron in the Presence of Chromium Precipitates...........78 Iron and Chromium Reduction in Columns.................................................................................... BI Conclusions.............................................................................................................................................85 5. DISSIMILATORY IRON-REDUCING BACTERIA CAN INFLUENCE THE REDUCTION OF CARBON TETRACHLORIDE BY IRON METAL........:.............................................................87 Abstract...................................................................................................................................................87 Introduction.............................................................................................................................................87 Materials and Methods..................... 89 Organism and Culture Conditions................................................................................................... 89 CT Reduction...................................... 90 Iron Reduction........................................... ..................................................................................... 9 1 Adhesion of BrY to Iron........................................... 91 Analytical Methods....................................................................................... '•............................... 91 Results and Discussion................... .-...................................................................................................... 92 Conclusions......................................................................................................................................... -.95 vii 6. CONCLUSIONS.....................................................................................................................................97 Implications for Field Applications....................................................................................... REFERENCES CITED....................................................................................... 99 102 I viii LIST OF TABLES Table PaSe 2.1. Summary of parameters for bacterial transport experiments in 30 cm porous media columns filled with 40 mesh quartz sand....................... 26 2.2. Normalized breakthrough concentrations, absolute, and relative recoveries of cells from 30 cm columns........................................................................................................ 33 2.3. Su m m ary of parameters for bacterial transport experiments in 3 m porous media columns filled with 40 mesh quartz sand............ :................... .•................................................................. -34 2.4. Normalized breakthrough concentrations, absolute, and relative recoveries of cells from 3 m columns.................................................................................................................................... 39 3.1. Parameters used for the calculation of normalized breakthrough concentrations in 30 cm columns as shown in Figure 3.1............................................................................................ 3.2. Estimated single collector efficiencies (rj) and collision efficiency factors (a) for starved and vegetative S. algae BrY cells, 30 cm columns....................................... ....................... 3.3. Estimated single collector efficiencies (77) and collision efficiency factors (a) for starved and vegetative S. algae BrY cells, 3 m columns................ .................................................. I 4.1. Adhesion of Y algae BrY cells to HEPES-pretreated iron and Fe(O)................................ ix LIST OF FIGURES Figure 1 Page LI. Examples of permeable reactive barrier designs.............................................................................13 1.2. Reactive Media employed in permeable reactive barriers....................................................... ....14 2.1. Normalized breakthrough concentration (c/c0) of starved and vegetative cells in the effluent of 30 cm long sand..,...................... *..................................................................................27 2.2. Normalized concentrations of starved and vegetative cells of S. algae BrY in the pore water of 30 cm long columns..........................................................................................................29 2.3. Sorption of starved and vegetative cells of S. algae BrY to quartz sand along the flowpath of 30 cm long columns......................................................................'.............................................. 31 2.4. Normalized breakthrough concentration of starved and vegetative cells of Shewanella algae BrY in the effluent of 3 m long columns...............................................................................35 2.5. Normalized concentrations of starved and vegetative.cells of S. algae BrY in the pore water of 3 m long columns....................... ...................................................................................... 37 2.6. Sorption of starved and vegetative cells of S. algae BrY to quartz sand along the flowpath of 3 m long columns........................................................................................................................38 3.1. Predicted, normalized, steady-state breakthrough concentrations through 30 cm quartz sand columns................................................................................................................................... 47 3.2. UV fluorescence images o f Shewanella algaeBiY-gfp cells.........................................................52 3.3. Concentration of total cells, culturable cells, and absorbance at 600 nm during starvation of Shewanella algae BrY................................................................................................................ 53 3.4. Change in hydrophobicity index and zeta potential during starvation of Shewanella algae BrY.................................................................................................................................................. 54 3.5. Change of apparent buoyant density and apparent diffusion coefficient during starvation of Shewanella algae BrY................................................................................................................ 55 3.6. Change in adhesivity and transportability during starvation of Shewanella algae BrY..................59 3.7. Plot of fractional recovery from 30 cm quartz sand columns over the fraction of cells adhered in batch experiments......................................................................................................... -60 4.1. PTFE, stainless steel column assembly for studying microbial Fe(III)-reduction and Cr(VI) reduction in flow through systems....................................................................................... 71 4.2. Production of Fe(II) from amorphous Fe(III) oxyhydroxide-coated sand by starved and vegetative cells of S. algae BrY...................................................................................................... 74 X 4.3. Cr(VI) transformation in vials containing microbially reduced Fe-coated sand and active S. algae BrY, microbially reduced Fe-coated sand (autoclaved to eliminate microbial activity), and oxidized Fe-coated sand........................................................................................... 76 4.4. Direct Cr(VI) transformation by active S. algae BrY in the absence of sand, in the presence of quartz sand, and in the presence of iron-coated sand.................................................. 78 4.5. Mass ofFe(II) over time in vials containing iron oxyhydroxide-coated sand as determined by separate analysis of aqueous and solid phase in the vials.......................................................... 80 4.6. Cr(VI) transformation in vials with initially exhausted Cr(VI)-reduction capacity after reinoculation and incubation with S. algae B rY ............................................................................ 81 4.7. Normalized breakthrough curves of a fluorescein tracer and its corresponding Cf(VI) breakthrough data through columns containing biologically produced Fe(II)............................... 83 4.8. Processes governing the fate of Cr(VI) in environments containing surface associated iron and dissimilatory metal-reducing bacteria (DMKB)................................................................84 5.1. Carbon tetrachloride degradation and chloroform production by passivated Fe(O), passivated Fe(O) with active S. algae BrY.cells,, and active S. algae BrY cells alone.................... 93 5.2. Production of Fe(II) by passivated Fe(O), passivated Fe(O) with active S. algae BrY cells, and passivated Fe(O) with heat-killed S. algae BrY cells............................................................... 94 xi ABSTRACT Dissimilatory metal-reducing bacteria (DMRB) are capable of reducing a wide range of metals and the injection of DMRB is considered for the cleanup of contaminated soil and groundwater. The success of such an approach, among other factors, relies upon the bacterial transport through porous media and the contaminant transformation by DMRB. Long term starvation of Shewanella algae BrY, the model DMRB used in this research, allows for significantly improved transport through quartz sand porous media columns. However, the reasons for the improved transport remain unclear. Changes in cell size, net electrostatic charge, hydrophobicity, buoyant density, and effective diffusion coefficient do not provide a sufficient explanation and this question will have to remain the focus of future research. In addition, it became evident that the employed mathematical model based on the colloid filtration theory is not capable of appropriately describing scale and physiology dependent effects on bacterial transport through porous media. A complete understanding of the processes and factors influencing the porous media transport of starved bacteria is still lacking and will remain as the focus of future research. The reduction of chromium [Cr(VI)], a compound of significant environmental concern at many Department of Energy, Department of Defense, and Environmental Protection Agency Superfund sites, can be facilitated by DMRB. Starved S. algae BrY cells can be resuscitated into an actively metabolizing state in the presence of ferric iron [Fe(III)], which is abundant in the environment. Active S. algae BrY cells can either directly reduce Cr(VI) or produce surface reactive ferrous iron [Fe(II)]. Fe(II) chemically reduces and precipitates Cr(VI) eliminating it from contaminated water. S. algae BrY cells can also potentially contribute towards the long term reactivity of zero valent iron subsurface barriers. The microbial reduction of Fe(III) allows for the removal or activation of surface associated corrosion products and results in increased transformation rates of carbon tetrachloride, another widespread environmental contaminant. The results summarized in this dissertation indicate that the injection of starved DMRB is a promising technology for the remediation of subsurface environments contaminated with heavy metals and chlorinated hydrocarbons. I CHAPTER I INTRODUCTION Every year, large quantities of organic and inorganic compounds are released into the environment as the result of human activities. These releases can be deliberate or accidental. Due to their toxicity and recalcitrance in the environment, many of these compounds can accumulate to unacceptable levels and pose a severe risk to humans and the environment. Davis (1972) was one of the first to suggest the application of bacteria in the remediation of contaminated groundwater. The use of indigenous or exogenous bacteria for the cleanup of contaminated groundwater has since become a widely applied approach. In situ bioremediation has become one of the preferred technologies for the cleanup of contaminated sites since it avoids the need for excavation and subsequent disposal of hazardous waste. While the bacterial degradation of fuel hydrocarbons and polyaromatic hydrocarbons is well documented (Alexander, 1999; Crawford and Crawford, 1996), the use of bacteria for the cleanup of soils and groundwater contaminated with chlorinated hydrocarbons (U.S.EPA, 2000) and metals (Evanko and Dzombak, 1997; McCullough et al., 1999) remains the focus of much research. The increased need and interest for biological metal and radionuclide remediation is partially based on more than 50 years of using nuclear energy for both peaceful and military purposes. Many Department of Energy (DOE) and Department of Defense (DOD) sites are contaminated with hazardous and radioactive waste. These sites create a legacy that is estimated to cost the DOE alone more than US $ 60 billion over the next 10 years (McCullough et al., 1999). The contamination at many DOE facilities is spread over large areas and deep below the surface (McCullough et al., 1999). Conventional treatment strategies, such as pump and treat or excavation with subsequent treatment, might not be practical, economical, or advisable due to the long term liability that might be created. Thus, the development of alternative strategies for the treatment of DOE sites can potentially save billions of dollars. 2 Two subsurface remediation strategies have received much attention over the recent years and have demonstrated potential for the development of economically viable, long term, low maintenance cleanup strategies. 1. The use of permeable reactive barriers (PRBs). PRBs established downstream of contaminated sites have proven to effectively prevent off-site migration of chlorinated hydrocarbons, metals, and radionuclides (Scherer et ah, 2000; U.S.EPA, 1998; Yuyun and Allen, 1999). 2. The metabolism of dissimilatory metal-reducing bacteria (DMRB). DMRB have demonstrated potential for the remediation of groundwater contaminated with chlorinated hydrocarbons, heavy metals, and radionuclides (Amonette et al., 2000; McCullough et al, 1999; Workman et al., 1997). Dissimilatory metal-reducing bacteria (DMRB) gain energy for growth by coupling the oxidation of organic compounds or hydrogen to the dissimilatory reduction of ferric iron [Fe(III)] and other oxidized metals. DMRB can enzymatically reduce large amounts of a wide range of metal ions, including Fe(III), Cr(VI) and U(VI) (Harding, 1997; Lovley, 1993; Lovley, 1995b). Although the direct enzymatic reduction of heavy metals by DMRB potentially allows for the development of remediation strategies, the indirect reduction of soluble contaminants such as Cr(VI), U(VI), and carbon tetrachloride might be more important and feasible in subsurface environments (Fredrickson and Gorby, 1996a; Harding, 1997; McCullough et al., 1999). DMRB can reduce a wide range of ferric iron minerals to produce redox reactive ferrous iron [Fe(II)] (Caccavo, Jr. et al., 1992). Ferrous iron can chemically react with soluble contaminants including Cr(VI), U(VI), and carbon tetrachloride (Amonette et al., 2000; Eary and Rai, 1988; McCullough et al., 1999). Over the last decade, it has become apparent that the biological and geochemical processes described above can govern the fate of metals and radionuclides in the subsurface. Consequently, a number of governmental agencies have begun to support research addressing the fate of heavy metals and radionuclides in the subsurface. The DOE, for instance, established the Natural and Accelerated 3 Bioremediation Research (NABIR) program in 1997. The NABIR program focuses on the development of a strong theoretical foundation for the assessment and remediation of DOE sites contaminated with mixtures of metals and radionuclides. The high interest in the biogeochemical processes involved also recently led to a five day symposium on “Chemical-Biological Interactions in Contaminant Fate”, organized by the Division of Environmental Chemistry at the 220th American Chemical Society National Meeting in Washington D.C. (August 20-24, 2000). The majority of the presentations focused on the influence of iron minerals and dissimilatory metal-reducing bacteria on the fate of chlorinated hydrocarbons, heavy metals, and radionuclides. The presentations clearly demonstrated the importance of metal-reducing bacteria in contaminated subsurface environments. The idea of utilizing DMRB to establish and maintain permeable reactive barriers (PRB) in the subsurface is intriguing. DMRB, present in the subsurface or injected into the subsurface, could be stimulated to produce redox reactive ferrous iron from indigenous or injected ferric iron. The produced ferrous iron would chemically react with chlorinated hydrocarbons, Cr(VI), U(VI), and other contaminants. However, reports of field trials evaluating the use of DMRB for the establishment of redox reactive subsurface barriers are lacking. Thus, this research project is designed to improve the basis for future field demonstrations involving the use of DMRB for environmental cleanup. The major research goals of this dissertation are: 1. The development of a strategy for the enhancement of bacterial transport through porous media, and 2. The assessment of using dissimilatory metal-reducing bacteria for the in situ bioremediation of heavy metals and chlorinated hydrocarbons. These two m ain research goals are reflected in the structure of this dissertation. The first part of the introduction reviews the processes and parameters governing bacterial transport through porous media. This review is the basis for Chapters 2 and 3, which evaluate bacterial starvation as a promising strategy for the bacterial transport enhancement (Chapter 2), and document the changes in transport-related bacterial cell properties during long term starvation (Chapter 3). The second part of the introduction provides an 4 overview of the current literature on permeable reactive subsurface barriers and microbial metal-reduction. This overview provides the background for the Chapters 4 and 5, which describe the potential use of DMRB to establish and maintain redox reactive zones in the subsurface capable of the continuous precipitation of .Cr(VI) (Chapter 4) and the potential of DMRB to influence the performance of zero valent iron redox reactive subsurface barriers (Chapter 5). Bacterial T ransport Through Porous Media for Bioaugmentation (Research Goal I) The transform ation rates achievable by stimulation of the indigenous microflora at contaminated subsurface sites are often insufficient to develop economically viable remediation technologies. An insufficient number of indigenous microorganisms capable of degrading the compounds of interest or the inability to effectively stimulate the indigenous microflora are commonly encountered reasons for low transformation rates (Vogel, 1996; Walter, 1997). Therefore, bioaugmentation, the injection of I microorganisms which can biotransform environmental contaminants, has been proposed and implemented (Dybas et al., 1998; Ellis et ah, 2000; McCullough et al., 1999; Salanitro et al., 2000; Steffan et al., 1999). Unfortunately, the injection of previously cultivated bacteria into subsurface formations can cause excessive biofilm growth in injection wells. The resulting clogging of well screens poses a major problem in the development of subsurface bioaugmentation strategies (Molz et al., 1986; Nelson et al., 1985, Malusis et al., 1997; Maxwell and Baqai, 1995). Thus, for the successful development of subsurface bioaugmentation strategies, the ability to avoid excessive adhesion, limit biofilm growth, and to facilitate bacterial transport through porous media needs to be improved. Much research has focused on the development of strategies for the enhancement of bacterial transport through porous media. However, a promising technology applicable to a wide range of bacterial strains and to different hydrogeological conditions has yet to be developed. A bacterial strain or consortium chosen for subsurface bioaugmentation should have the following characteristics. The microorganisms should easily obtain regulatory, approval for injection. The microorganisms should be easily transported to avoid the plugging of injection wells and readily distributed 5 over large areas. The cells should yet be adhesive enough to distribute themselves along the flowpath. They should be compatible with the indigenous microflora, i.e. they should not negatively influence the existing microflora, but should be able to survive and actively metabolize the compounds of interest. Bacterial strains or consortia for bioaugmentation should be inexpensive to obtain and easy to culture. The following literature review on bacterial transport through porous media will justify the choice of bacterial starvation as a means for enhancing bacterial transport through porous media. Parameters Influencing Bacterial Transport Through Porous Media Bacterial transport is typically governed by aqueous phase advective movement coupled with retardation by adhesion onto surfaces and straining or trapping in interstitial pores. Bouwer et al. (2000), Mills (1997), Lawrence and Hendry (1996), and Harvey (1991) provide excellent reviews of the factors influencing bacterial transport through porous media. The numerous parameters influencing bacterial transport and adhesion in porous media can be grouped into three major categories, I) solution chemistry, 2) porous media characteristics and hydraulic conditions, and 3) properties of the bacterial cells. These categories are discussed in the following paragraphs. Influence of Solution Chemistry on Bacterial Transport Through Porous Media. A wide range of solute characteristics has been reported to influence bacterial adhesion to surfaces and transport through porous media. An increase in ionic strength has been correlated with increasing attachment (Pontes et al„ 1991; Gannon et al., 1991b; Jewett et al., 1995; Mills et al., 1994; Scholl et al., 1990; Tan et al., 1994) and is usually attributed to the compression of the electrostatic double layer in the presence of high ion concentrations (Das and Caccavo Jr., 2000; Deshpande and Shonnard, 1999; Mills et al., 1994). Scholl and Harvey (1992), Kinoshita et al. (1993), and Jewett et al. (1995) report the effect of changes in pH values on bacterial transport and attachment, however, no uniform results are found. Increased attachment with decreasing pH is reported by Scholl and Harvey (1995) and Kinoshita et al. (1993), while Jewett et al. (1995) state that changes in pH from 5.5 to 7 do not affect bacterial transport through porous media 6 columns. Temperature has also been shown to influence bacterial transport, however the effect seems to be case dependent. While Bellamy et al. (1985) describe higher cell removal at higher temperatures, Sarkar et al. (1994) describe lower removal at higher temperatures. Johnson and Logan (1996) investigated the influence of dissolved and sediment organic matter (DOM, SOM) on bacterial transport through porous media columns and describe an increase in travel distance in the presence of SOM. This increase in travel distance is attributed to an increase in negative surface charge of the iron-coated quartz sand through the addition of SOM. Like SOM, surfactants or dispersants can also result in decreased attachment and therefore facilitate the transport of bacteria through porous media. This is most likely due to decreased cell porous media interactions in the presence of surface active compounds. However, the activity or viability of the bacteria may be influenced negatively by the addition of such compounds (Goldberg et al., 1990; Gross and Logan, 1995b; Jackson et al., 1994; Paul and Jeffrey, 1984; Sarkar et al., 1994). Influence of Porous Media Properties and Hydraulic Conditions on Bacterial Transport Through Porous Media. A number of porous media characteristics and hydraulic conditions have been reported to influence bacterial adhesion and transport through porous media. For instance, the pore size, grain size, and their distributions can influence bacterial transport through porous media (Fontes et al., 1991; Marlow et al., 1991; Sharma and Mclnemey, 1994; Smith et al., 1985; Wood and Ehrlich, 1978). Sharma and McInemey (1994) report that penetration rates of non-motile strains increase linearly with the size of the glass beads used, however, penetration rates of motile strains become independent with bead diameters of 398 |tm or larger. Presumably, the lower specific surface area provided by larger grains and the higher hydraulic conductivity of the coarse grained porous medium allow for increased transport (Fontes et al., 1991; Marlow et al., 1991). Wood and Ehrlich (1978) and Smith et al. (1985) attribute enhanced bacterial transport to preferred flowpaths, which are provided by fractures or macropores present in the native material. The soil mineralogy (e.g. presence of Fe-minerals) and the organic matter content can influence the porous media surface charge and surface hydrophobicity, respectively. Both, the surface charge and the surface hydrophobicity, can influence bacterial adhesion to surfaces and transport through porous media. 7 The presence of positively charged surfaces like Fe-minerals in soils is reported to lead to increased attachment of bacterial cells, which carry a net negative charge at neutral pH values (Johnson and Logan, 1996; McCaulou et ah, 1994; Mills et ah, 1994; Scholl and Harvey, 1992; Scholl et ah, 1990). Hydrophobic surfaces have been described to increase the number of cells adhering in batch and column systems (Absolom et ah, 1983; Fletcher and Loeb, 1979; McCaulou et ah, 1994) however, hydrophilic cells might show decreased attachment to hydrophobic surfaces (Paul and Jeffrey, 1985). In addition to the surface chemistry, the physical characteristics, such as surface roughness, can significantly influence attachment regardless of the prevailing chemistry (Geesey and Costerton, 1979; Mueller, 1996). Thus, it is not surprising that hydrodynamic conditions have also been found to play a very important role in bacterial adhesion to surfaces. Rijnaarts et ah (1993) and Scheuerman et ah (1998) suggest that the hydrodynamics can become much more important in view of bacterial attachment than the physicochemical conditions. The interstitial fluid velocity can vary widely when cells are actively injected into the subsurface. The interstitial fluid velocity depends on the pumping rates during injection and the radial distance from the point of injection. Most authors report an increase in breakthrough and dispersal of bacteria at higher interstitial fluid velocities (Gannon et ah, 1991b; Marlow et ah, 1991; Sarkar et ah, 1994; Tan et ah, 1994), although Camesano and Logan (1998) report higher dispersal at low pumping velocities and Gross et ah (1995) observe no influence of fluid velocity on bacterial transport through porous media. Tnflnenne of Cell Properties on Bacterial Transport through Porous Media. A number of bacterial cell properties have an influence on bacterial adhesion and bacterial transport through porous media. The cell surface charge, often measured as zeta potential, has been shown to influence bacterial adhesion to surfaces. Cell adhesion has been correlated mostly inversely but also directly with surface net electrostatic charge (Gilbert et ah, 1991; Sharma et ah, 1985; VanLoosdrecht et ah, 1987a). Cell hydrophobicity has been reported to influence bacterial adhesion to surfaces and transport through porous media (Caccavo, Jr. et ah, 1997; DeFlaun et ah, 1999; Fletcher and Marshall, 1982; VanLoosdrecht et ah, 1987b) and separate 8 mechanisms have been found for the bacterial adhesion to hydrophilic and hydrophobic surfaces (Paul and Jeffrey, 1985). Both, the hydrophobicity and net electrostatic charge, are influenced by the type of surface molecules present on the outside of the cells. Proteins and (lipo)polysaccharides are believed to influence bacterial adhesion to surfaces and transport through porous media but it cannot yet be predicted how certain surface molecules influence bacterial transport and adhesion (Abbott et al., 1983; Caccavo, Jr., 1999; Caccavo, Jr. et al., 1997; Costerton et al., 1978; Fletcher, 1976; Fletcher and Floodgate, 1973; Rijnaarts et al., 1996; Williams and Fletcher, 1996). Cell motility and chemotaxis can have an influence on transport through porous media. While the influence of bacterial motility on bacterial transport through porous media seems to be mostly negligible at flow velocities typically encountered in aquifers (Barton and Ford, 1995; Barton and Ford, 1997; Camper et al., 1993), bacterial motility is suggested to result in increased dispersal of cells (Camesano and Logan, 1998) and in increased penetration under static conditions (Jenneman et al., 1985; Reynolds et al., 1989). Cell appendages, such as flagella and pill, have also been shown to become important during initial bacterial attachment to surfaces. This influence is attributed to an increase in collisions due to the greater effective diameter, an increase in effective diffusivity, and to the ability of the bacterial appendages to reach across the boundary layer between the liquid and the solid phase (Busscher et al., 1990a; Busscher et al., 1990b; Busscher et al., 1992; Mueller, 1996; O'Toole and Kolter, 1998a; O'Toole and Kolter, 1998b; Piette and Idziak, 1991). Das and Caccavo suggest that the flagellum acts as a motility-independent adhesin (Das and Caccavo, unpublished results). Other bacterial cell properties correlated with transport through porous media include the cell size and cell shape. Smaller and more spherical cells are commonly thought to be transported better through porous media than larger and elongated cells (Fontes et al., 1991; Gannon et al., 1991a; Weiss et al., 1995). Harvey et al. (1997) show that the buoyant density of bacterial cells can influence the sedimentation rates of bacteria in porous media environments, but it is considered negligible at ambient groundwater flow velocities (Corapcioglu and Haridas, 1984). 9 Many of the cell properties listed above are influenced by the physiological state of the bacteria and can be significantly different for the same bacterium depending on the environmental conditions (Grasso et al., 1996; VanLoosdrecht et al., 1987a). Thus, controlling the physiology of a bacterial strain before injection offers great potential for the development of effective bacterial transport strategies. Tmnlications for Bioausmentation Strategies The review of the current literature implies that bacterial transport through porous media is influenced by a combination of numerous parameters. The importance of any single parameter is specific to any given situation and cannot be predicted easily. Camper et al. (1993) and Harvey (1997) state that bacterial transport experiments should be performed with the aquifer material of interest and as close as possible to the expected conditions in the field since the determination of individual characteristics of cells or the porous medium will not allow an accurate prediction. A good overview on how to design and standardize bacterial transport experiments is given by Harvey (1997). The examination of the parameters influencing bacterial adhesion to surfaces and transport through porous media makes evident that only a very limited number of parameters can be effectively manipulated in field scale applications. Attempts to change the solution chemistry or the physicochemical properties of the porous medium are likely to become an economical and technological challenge. Thus, manipulating parameters falling into these two categories might be extremely difficult or not advisable. In addition to the economical and technological challenges, lowering the ionic strength or changing the pH of groundwater could lead to undesirable changes in the groundwater chemistry potentially resulting in increased contaminant mobility. Manipulating the physicochemical properties o f the aquifer by for instance hydraulically or pneumatically fracturing the aquifer is difficult to control over large areas and might result in increased contaminant transport through zones with high hydraulic conductivity. Manipulating and controlling the bacterial inoculum and the interstitialfluid velocity appear to have the greatest potential o f economical and technological success in field scale subsurface bioaugmentation strategies. 10 Bacteria are commonly cultured in large amounts before injection in bioaugmentation projects. The cultivation step allows for the manipulation of the inoculum in many different ways. Strains with a decreased tendency to adhere to surfaces can be selected or bacteria can be harvested at a specific growth rate. Procedures on how to select for such strains have been described in Caccavo et al. (1997) and DeFlaun et al. (1990, 1999). Unfortunately, these procedures can be time consuming and might not be successful for every bacterial strain or situation. Transport enhancement strategies which are potentially applicable to any bacterial strain and consortium are more desirable. The following paragraph will indicate that a potentially applicable strategy involves the use of starved bacteria. Bacterial Starvation as a Transport Enhancement Strategy Starvation is believed to be a common survival mechanism for bacteria which cannot form spores or cysts (Hood and MacDonell, 1987; Kjelleberg et al., 1993b; Tabor et al., 1981). Starvation of bacteria can result in radical size reduction and a rapid decrease in metabolic activity until the bacteria approach complete dormancy (Kjelleberg, 1993a). Starved bacteria can survive for years in the absence of nutrients (Amy and Haldeman, 1997; Kjelleberg, 1993a) and the improved transport of metabolically dormant cells is believed to contribute towards the presence of bacteria in the deep subsurface (Fredrickson and Onstott, 1996b; Lappin-Scott and Costerton, 1990). Starved bacteria can be resuscitated relatively rapidly by the addition of suitable nutrients (Amy and Morita, 1983; Cunningham et al., 1997; Kjelleberg, 1993a; Novitsky and Morita, 1978), making bacterial starvation a potentially effective means of facilitating transport of through porous media for a number of engineering applications (Bouwer et al., 2000; Cunningham et al., 1997; Cusack et al., 1992; Gerlach et al., 1998; Lappin-Scott and Costerton, 1992; Lappin-Scott et al., 1988a; Lappin-Scott et al., 1988b; MacLeod et al., 1988; Sharp et al., 1999). The injection, resuscitation, and subsequent plugging of high permeability zones in oil bearing formations using starved bacteria is demonstrated in the literature (Cusack et al., 1992; Lappin-Scott and Costerton, 1990; Lappin-Scott et al., 1988a; Shaw et al., 1985). The use of starved bacteria results in significant improvements in secondary oil recovery since the bacteria travel deeper into the high 11 permeability zones and form bacterial plugs upon resuscitation, which force the injected water through the low permeability zones to recover residual oil more efficiently. Starved bacteria have also been used in a series of laboratory and field scale experiments demonstrating the possibility to form hydraulic groundwater barriers. Starved cells derived from bacteria known to produce large amounts of extracellular polysaccharides (EPS) were injected into porous media and resuscitated. The bacteria are reported to form thick biofilms upon resuscitation and to decrease the porous media permeability dramatically (Cunningham, 2000; Cunningham et al., 1997). It is well known that a number of changes occur during long term starvation of non spore-forming bacteria. A large amount of KNA and protein appear to be degraded rapidly at the onset of starvation (Kaplan and Apirion, 1975b; Pine, 1965), however synthesis of new proteins appears to be required for long term survival of starving cells (Groat et al., 1986; Hengge-Aronis, 1993; Kaplan and Apirion, 1975a; Kjelleberg et al., 1993b; Reeve et al., 1984). These proteins are believed to be part of a general stress response, often involving RpoS, and allow bacteria to become more stress resistant and more efficient scavengers (Cappelier et al., 2000; Fischer et al., 1998; Foster and Spector, 1995; Matin, 1991; Matin et al., 1989; Spector, 1998; Spector et al., 1999; Svensater et al., 2000; Teich et al., 1999; Zinser and Kolter, 1999; Zinser and Kolter, 2000). Increased stress resistance is verified by the observation that starved cells survived better under a range of stress conditions than their vegetative counterparts (Albertson et al., 1990; Hartke et al., 1994; Nystrom et al., 1988; Nystrom et al., 1990; Nystrom et al., 1992) and that slow growing cells are better adapted for starvation stress survival than their fast growing counterparts (Muller and Babel, 1996). In the section “Influence of Cell Properties on Bacterial Transport through Porous Media” cell surface properties were discussed, which are known to influence bacterial attachment to surfaces and transport through porous media. It was discussed that the growth state of bacteria and the presence of nutrients can influence attachment. Thus, it is not surprising that bacterial starvation can significantly influence the attachment of bacteria to surfaces. Short term starvation of bacteria can result in an increased tendency to attach to surfaces (Dawson et al., 1981; Kjelleberg, 1984) whereas long term starvation (weeks 12 to months) may decrease bacterial attachment and enhance bacterial transport through porous media (Bouwer et ah, 2000; Cusack et ah, 1992; Gerlach et ah, 1998; Lappin-Scott and Costerton, 1992; LappinScott et ah, 1988a; Lappin-Scott et ah, 1988b; MacLeod et ah, 1988; Sharp et ah, 1999). It is one goal of this research to develop and evaluate strategies for the enhancement of bacterial transport through porous media, which are potentially applicable to field relevant scales, a wide range of bacterial strains, and different hydrogeological conditions. Ideally, such a strategy should allow to culture a large number of bacteria inexpensively, which are mobile enough to transport over large distances and yet adhesive enough to attach to soil particles. This would allow establishing a sessile population capable of transforming dissolved and sorbed contaminants. Long term nutrient starvation appears to provide such a widely applicable strategy for the enhancement of bacterial transport through porous media. The motivation for using the DMRB, Shewanella algae BrY, for these studies is explained in the following section. Permeable Reactive Subsurface Barriers and Microbial Metal-Reduction !Research Goal 2) The use of permeable subsurface treatment zones (or' barriers) established downstream of contaminated areas provides potential for preventing the off-site migration of contaminants without severely influencing the natural groundwater flow. Permeable reactive barriers (PRBs) can provide long term control of a contamination problem, if their reactivity is maintained over extended periods of time. PRBs are commonly constructed by excavating trenches perpendicular to the groundwater flow direction and refilling these trenches with the reactive material of choice. These systems represent semi-passive approaches, which minimize the exposure of operating personnel to potentially hazardous compounds. Permeable reactive barriers can be made in a number of different designs and consist of many different materials (Figure 1.1 and Figure 1.2, Gavaskar et ah, 1998; Sacre, 1997; Scherer et ah, 2000; U.S.EPA, 1997; U.S.EPA, 1998; Yuyun and Allen, 1999). 13 Reactive I Cell Funnel Reactive Cell Plume Plume ' Pea Gravel (a) Continuous reactive barrier configuration Reactive v Cellx Plume Pea Gravel (b) Funnel-and-gate configuration Funnel Caisson Gate I Plume Caisson Gate 2 Funnel (c) Multiple caisson gates Figure 1.1. Reactive Pea Medium I Gravel Reactive Medium 2 (d) Serial reactive media Examples of permeable reactive barrier designs (from Gavaskar et al., 1998). The most commonly used material is zero valent iron (Fe(O)). Fe(O) has been employed in permeable reactive subsurface barriers which are capable of remediating a wide range of contaminants including chlorinated organics (R-Cl, Eykholt and Davenport, 1998, Gatpagar et al., 1997, Gillham and O'Hannesin, 1994, Helland et al., 1995, Johnson et al., 1996, Johnson et al., 1998, Roberts et al., 1996, Sayles et al., 1997, Siantar et al., 1996), heavy metal and radionuclide oxyanions, such as chromate (CrO42', Gould, 1982, Cantrell et al., 1995) and oxycations, such as uranyl (UO22*) (Fiedor et al., 1998), nitroaromatics (R-NO2, Cao et al., 1999, Agrawal and Tratnyek, 1996), and nitrate (NO3", Huang et al., 1998, Till et al., 1998, Siantar et al., 1996). The corrosion of Fe(O) provides the electrons necessary for the reduction of these contaminants according to the following half reactions. Fe(O) -»Fe2+ + 2e~ (Eq.1.1) Fe2' ->Fe3' + e- (Eq.1.2) The electrons provided react with the different contaminants according to following half reactions. R-Cl + 2 e~ + H+ CrO42" + 3 e" + 8 H+ UO22' + 2 e~ R-H + CE -> Cr3' + 4 H2O -> UO2 (Eq. 1.3) (Eq. 1.4) (Eq. 1.5) 14 R-NO2 + 6 e" + 6 H+ NO)" + 8 e" + 9 H+ R-NH2 + 2 H2O (Eq. 1.6)' -> NH3 + 3 H2O microbes (2 %) peat (6 %) (Eq. 1.7)12 chitosan beads (2 %) hydrogen sulfide (2 %) other (17 %) Figure 1.2. Reactive Media employed in permeable reactive barriers, adapted from Sacre (1997). The simplicity, low maintenance requirements, and robustness of the approach have made Fe(O) PRBs one of the preferred technologies for the remediation of groundwater contaminated with chlorinated hydrocarbons and heavy metals. However, no reliable data on the long term performance of Fe(O) PRBs are available yet. Laboratory studies simulating long term performance of Fe(O) indicate that contaminant degradation rates can either decrease (Devlin et al., 1998, Gatpagar et al., 1997) or more rarely increase (Holland et al., 1995) 1 Other potential reaction products include nitroso and hydroxylamino compounds. (Klausen et al., 1995) 2 Other potential reaction products are nitrogen gas and nitrite (Till et al., 1998) ■ 15 as a function of time. This temporal effect on contaminant degradation rates is usually attributed to the development of corrosion products that passivate the metal surface and influence electron transport to the contaminant (Helland et al, 1995; Johnson and Tratnyek, 1994; Odziemkowski et ah, 1998). A wide range of iron corrosion products has been found in laboratory and field studies. The corrosion products commonly found in the presence of contaminated groundwater include iron (hydroxides such as goethite (Pratt et al., 1997), magnetite (Bonin et al., 1998; Odziemkowski and Gillham, 1997; Odziemkowski et al., 1998), maghemite (Peterson et al., 1997), hematite (Pratt et al., 1997), mixed chromium-iron-hydroxides (Powell et al., 1995a, Blowes et al, 1997b), green rust (Odziemkowski and Gillham, 1997; Odziemkowski and Gillham, 1997), and siderite .(Mackenzie et al., 1995). The corrosion product distribution appears to be dependent on the type of iron used, contaminant present, geochemical conditions, and time of exposure to contaminated groundwater. Fe(O) PRBs are usually established using trench and fill methods to place the iron into the subsurface. However, the installation can become uneconomical at depths of more than 12 to 15 meters and alternative methods for establishing permeable reactive zones in the subsurface have been proposed (Yuyun and Allen, 1999). Since injection wells can be installed much deeper than trenches the injection of colloidal Fe(O) particles to establish redox reactive zones has been proposed (Cantrell and Kaplan, 1997a; Cantrell et al., 1997b; Cantrell et al., 1995). A group at Pacific Northwest National Laboratory (PNNL) has demonstrated and patented a method using dithionite (S2O42") as a chemical reductant (AmOnette et al., 1998; Amonette et al., 1994; Istok et al., 1999; Scott et al., 1998). The dithionite is injected into the subsurface in a pH 11 bicarbonate buffer and reduces indigenous ferric iron to ferrous iron, which chemically reacts with heavy metals, radionuclides, and chlorinated hydrocarbons. This technology has been demonstrated at DOS's Hanford site in Richland, WA and at an army base in Fort Lewis, WA (Beets, 1998; Fruchter et al., 1997): Alternatively, indigenous ferric iron could be reduced by microorganisms. Dissimilatory metalreducing bacteria (DMRB) are widely distributed in both pristine and contaminated terrestrial, aquatic and subsurface environments and can reduce many iron minerals found in the environment and on passivated i? 16 Fe(O) (Caccavo, Jr. et al., 1992; Fredrickson and Gorby, 1996a; Heijman et al., 1995; Heijman et al., 1993; Kostka and Nealson, 1995; Lovley, 1995b; Nealson and Saffarini, 1994; Roden and Zachara, 1996). DMRB can also reduce a large number of heavy metals and radionuclides enzymatically and therefore offer a potentially useful mechanism for the remediation of contaminated subsurface environments (Beets, 1998; Fredrickson and Gorby, 1996a; Fruchter et al., 1997; Lovley, 1995a; Lovley, 1995b; Lovley, 1997). Alternatively, DMRB can produce ferrous iron from indigenous ferric iron minerals in the subsurface, which in turn chemically reduces carbon tetrachloride (CL) to chloroform (CF) and Cr(VI) to Cr(III) (Amonette et al., 2000; Caccavo, Jr. et al., 1996; Gorby et al., 1994; Gorby et al., 1995). In summary, the potential use of DMRB in subsurface remediation has gained considerable interest from engineers, microbiologists, and geochemists. DMRB are known to influence the geochemistry of subsurface environments and have the potential to be part of successful strategies for the bioremediation of heavy metal and radionuclide contaminated subsurface environments. DMRB are also likely to contact existing subsurface remediation installations such as Fe(O) PRBs and to influence their performance. Research Goals This research is designed to evaluate the use of dissimilatory metal reducing bacteria (DMRB) for the bioaugmentation of contaminated subsurface environments. Two major research needs are identified based on the literature review. 1. The transport of DMRB through porous media has to be enhanced and injection well clogging avoided. 2. The applicability of DMRB to mediate the precipitation of heavy metals and the dechlorination of chlorinated hydrocarbons has to be evaluated. This dissertation addresses these needs by utilizing the dissimilatory metal-reducing bacterium, Shewanella algae BrY, as the model organism. The following four objectives are addressed in the four main chapters of this dissertation. 17 Objectives 1. Starvation of Shewanella algae BrY is evaluated as a strategy for the enhancement of bacterial transport through porous media. 2. The reasons for the improved transport of starved S. algae BrY bacteria are elucidated by monitoring changes in transport related cell properties during nutrient starvation. 3. The possibility of using S. algae BrY to establish and maintain redox reactive zones for the elimination of chromium(VI) from contaminated groundwater is tested. 4. The potential influence of S. algae BrY on existing remediation systems based on zero valent iron is evaluated. 18 CHAPTER2 ENHANCING TRANSPORT OF SHEWANELLA ALGAE BrY THROUGH POROUS MEDIA BY STARVATION Abstract The successful bioaugmentation of contaminated subsurface environments relies on the effective transport and distribution of bacteria through porous media. This chapter provides evidence that starvation of the dissimilatory metal-reducing bacterium Shewanella algae BrY results in a significant enhancement of bacterial transport. Intermediate scale column studies (30 cm and 3 m length) were conducted to compare the transport of starved and freshly cultivated (vegetative) cells of S. algae BrY through quartz sand columns. Plate counts were utilized to estimate bacterial numbers and activity. The fractional recovery (number of cells recovered in the effluent compared to number of cells injected) and the normalized breakthrough concentration were statistically significantly greater for starved cells in both 30 cm and 3 m long columns. The distribution of starved cells sorbed to the sand along the flowpath was in addition more homogeneous for starved cells. It was concluded that bacterial starvation facilitates the transport and distribution of S. algae BrY through porous media. Introduction . The literature review in Chapter I suggests that nutrient starvation is potentially a reliable way to control the physiological state of bacteria and thus many transport-related parameters, including the cell size, shape, surface charge, hydrophobicity, buoyant density, type and amount of cell surface molecules, the presence of pili and flagella, and thus the motility and chemotaxis of cells. Starvation of bacteria results in radical size reduction and a rapid decrease in metabolic activity until the bacteria approach complete 19 dormancy (Kjelleberg, 1993a). Starved bacteria can survive for years in the absence of nutrients but can be resuscitated relatively rapidly by the addition of suitable nutrients (Amy and Haldeman, 1997; Kjelleberg, 1993a). Short term starvation (hours to days) of bacteria can result in increased attachment to surfaces (Dawson, et al, 1981; Kjelleberg, 1984) while long term starvation (weeks to months) can decrease attachment and enhance bacterial transport through porous media significantly (Bouwer et al., 2000; Gerlach et al., 1998; Lappin-Scott and Costerton, 1992; Sharp et al., 1999). The injection of starved bacteria has been suggested to promote plugging of subsurface formations for enhanced secondary oil recovery (Cusack et al., 1992; Lappin-Scott et al., 1988a; Lappin-Scott et al., 1988b; MacLeod et al., 1988) and for groundwater containment (Cunningham et al., 1997). This chapter describes the results of transport experiments and the derivation of parameters which quantitatively describe the transport of starved and vegetative cells of Shewanella algae BrY through porous media. Quantification of Bacteria in Porous Media The quantification of bacteria in groundwater and soil poses one of the greatest challenges in microbiology. There is a lack of. established methods for the reliable quantification of bacterial cell numbers and activity in porous media, which are applicable to field scale investigations. The methods employed in laboratory and field scale investigations include traditional plate counts, most probable number (MPN) methods, direct microscopic counts, the use of genetically engineered bacteria containing fluorescent markers (e.g. green fluorescent protein (gfp), Burlage et al., 1996) or an antibiotic resistance marker (Berry and Hagedom, 1991), radiolabeled cells (DeFlaun et al., 1999; Gross et al., 1995a), ferromagnetic antibody aided concentration (Johnson et al., 2001; Zhang et al., 1999), and 13Cdabeled cells together with isotope selective mass spectrometric measurements (DeFlaun et al., 1997). Some of these methods provide powerful tools for the quantification of bacteria in soil columns in the laboratory but are not practical for field scale applications due to regulatory concerns (i.e. radiolabeled cells or antibiotic markers). The use of ferromagnetic antibodies and 13C labeling allow for the selective enumeration of cells in groundwater but cannot effectively assess the number or activity of sorbed cells. Recent developments in 20 molecular biology have significantly improved our ability to selectively detect and quantify bacteria in environmental samples. However, affordable methods to reliably quantify bacterial activity in porous media are still under development. Bacterial plate counts have been used in a number of research projects evaluating the bacterial transport through porous media (Alexander et al, 1991). Plate counts of attached bacteria require the desorption of bacteria before culturing, which is commonly achieved by shaking, blending, or sonicating a soil sample in the presence of a buffered solution. The desorption of bacteria from soils is an important step since it represents a trade-off between complete desorption of bacteria and bacterial injury or decay due to stress factors during the desorption procedure. Stress due to buffer constituents, such as surfactants, of agitation procedures can significantly influence the outcome of the subsequent culturing step. Stevenson (1958) already reported that increasing the sonication time during an ultrasound aided desorption procedure increases the number of bacteria recovered from soils. However, extensive sonication can lead to a sharp decrease in viable cell numbers most likely due to heating and disruption of the cell integrity by ultrasound. It is well known that bacterial plate counts underestimate the number of bacterial cells especially when attempting to quantify the number or activity of bacteria sorbed to surfaces (Wagner et al., 1993; Zuberer, 1994). However, since plate counts provide a lower estimate of bacterial numbers and activity, they were utilized in this research. It can be assumed that the reported numbers commonly underestimate the number of active bacteria since plate counts do not account for viable but not culturable (VBNC) cells. Experiments comparing the transport of starved and vegetative cells of Shewanella algae BrY were performed on two different scales. First, short columns (30 cm length) were utilized to develop a test system for the relatively rapid assessment of bacterial transport in porous media. Larger scale experiments involving columns of 3 m length were utilized to test the applicability of the parameters derived from the short column experiments and to test the starvation transport enhancement strategy on a more field relevant scale. 21 Materials and Methods Strain and Culturing Methods Cultures of Shewanella algae BrY (formerly Shewanella alga BrY, Caccavo, Jr. et al., 1992) were maintained on tryptic soy agar (40 g L"1; Difco, Detroit, MI) at room temperature. Cells were aerobically grown to the late exponential, early stationary phase in tryptic soy broth (30 g L'1; Difco, Detroit, MI) at room temperature for 15 hours on a rotary shaker at 150 rpm. The cells were harvested by centrifugation (5860 x g, 20 minutes, 4 °C), washed twice in phosphate buffered saline solution (PBS; 8.5 g L"1 NaCl, 0.96 g L"1K2HPO4, 0.61 g L"1KH2PO4, pH 7.0), and resuspended in PBS. Washed S. algae BrY cells were either used directly (vegetative state) or starved by aseptically stirring on a magnetic stir plate at room temperature as described by Caccavo et al. (1996). Starved cells were harvested after 7 weeks of starvation by centrifugation and resuspended in PBS before incorporation into the transport experiments. Transport studies Cells of S. algae BrY, starved or vegetative, were injected into porous media columns filled with 40 mesh quartz sand (Unimin Corp., Emmet, ID). The columns were established using commercially available polyvinylchloride (PVC) pipe. Columns of two different dimensions were used, 3 m long and 2.54 cm in diameter, and 30 cm long and 1.5 cm in diameter. The column inlets and outlets were layered with nylon mesh, and pea gravel to prevent settling of sand into the inlet tubing and washout of sand through the outlet tubing. All columns were equipped with sample ports at the influent and effluent points and a number of columns were equipped with additional sample ports evenly distributed along the length of the column. The columns were packed wet to avoid the inclusion of air pockets and facilitate reproducible packing. The water level was maintained just below the top of the column during packing by periodically draining part of the water by gravity. The columns were always packed by slowly pouring the porous 22 medium into the water filled column to the desired settled bed height. The columns were operated upflow using a constant head tank or a peristaltic pump for 30 cm and 3 m long columns, respectively. The pore volume and hydrodynamic dispersion coefficient of each column were determined utilizing Na-fluorescein (Aldrich, Milwaukee, WI) as a tracer and by fitting the experimental data to an analytical solution of the advection dispersion equation described below (Van Genuchten and Alves, 1982). The interstitial fluid velocity (v) in the bacterial transport experiments was calculated using the measured flowrate, the cross sectional area of the column, and the porosity, which was estimated from the tracer studies. The bacterial concentrations were expressed as fractions of their influent concentrations in each experiment and were plotted as a function of the number of pore volumes eluted from the columns. In addition, the fractional recovery was calculated from the influent and effluent data. The fractional recovery was defined as the number of bacteria in the effluent of the column (integral of the effluent concentration ' over time) divided by the total number of bacteria injected into the column (integral of the influent concentration over time). The columns were prepared for inoculation by flushing with 2 pore volumes 0.5 % NaOCl (dilute commercial bleach), I pore volume sterile PBS, 2 pore volumes 0.01 M sodium thiosulfate to eliminate residual chlorine and 2 pore volumes sterile PBS. A commercial kit (EACH, Loveland, CO) was used to ensure residual chlorine concentrations below the detection limit (0.05 mg L"1) in the effluent of the columns. Plate counts routinely performed on effluent samples after this treatment did not show any growth. During the transport experiments, the columns were inoculated with approximately two pore volumes of starved or vegetative cells of S. algae BrY suspended in PBS. The transport of bacteria was monitored by sampling the influent, effluent, and the pore water along the flowpath of the columns for culturable cells. Aliquots (0.1 mL and I mL for 30 cm and 3 m columns, respectively) were taken using sterile, disposable I mL syringes equipped with 22 gauge needles. Dilutions in PBS were performed immediately and colony forming units were determined on tryptic soy agar (40 g L"1; Difco, Detroit, MI; 23 incubation for 24 hours at 30°C). The reported values are the arithmetic means of the three appropriate plate counts. Porous Media Sampling The distribution of sorbed cells throughout the columns was estimated at the end of the experiments by sectioning the columns and analyzing the sand cores for culturable cells after desorption from the sand. The 30 cm columns were divided into six sections of each five centimeter length while the 3 m long columns were sectioned into ten 30 cm long sections. The desorption of cells was accomplished by diluting a desorption solution, described by Camper et al. (1985), 1:5 in PBS. Higher concentrations of the desorption solution appeared to inhibit growth of S. algae BrY cells on TSA plates. An aliquot (5 ml) of this solution was added to I g of sand in a test tube, which was vortexed for I second. The test tube was placed on a horizontal shaker (150 rpm) for 30 min and then vortexed again for 3 seconds. The supernatant was sampled immediately after coarse particles had settled and plate counts on tryptic soy agar were performed after dilution in PBS. Modeling Bacterial transport through porous media has often been described using clean bed filtration models. Yao et al. (1971) suggested a set of equations describing the removal of suspended particles in a porous media bed. Improved and alternative clean bed filtration models were developed subsequently (Logan et al., 1993; Logan et al., 1995; Rajagopalan and Tien, 1976; Tien and Payatakes, 1976). Logan et al. (1995) summarized, clarified, and compared clean bed filtration model equations and proposed the following steady state solution. c — = exp 3 (1- 0 ) -arjx 2 (Eq. 2.1) 24 In Equation 2.1, c is the effluent. cell concentration (cells length"3), C0 the influent cell concentration (cells length"3), dc is the diameter of the spherical collectors (length), 0 the porosity, a the collision efficiency factor (i.e. the number of collisions of suspended bacteria with collectors which result in adhesion of a bacterium compared to the number of total collisions), Tj is the single collector efficiency (i.e. the rate at which particles strike a single collector divided by the rate at which particles flow towards the collector), and x is the distance into the porous medium (length). Assuming spherical collectors, the overall single collector efficiency (77) can be calculated as follows (Logan et al., 1995). . n = A N ~! + ^ n 2r+ N g (Eq. 2.2) The three terms involving NPe, Nr, and N0 represent the removal processes based on diffusion, interception, and gravitational settling, respectively (Logan et al., 1995). They can be calculated from (Eq 2.3) N pe = (Eq 2.4) [pp-p)gd2p (Eq 2.5) IS jiiv where dp is the diameter of the bacteria (length), Dp the particle diffusion coefficient (length2 time"1), Ps and p are the densities of the bacteria and the fluid (mass length"3), respectively, g is the gravitational constant (length time"2), jl the fluid viscosity (mass length1 time *), and v the interstitial fluid velocity (length time"1). Dp can be calculated from JcT D = 0 .1 0 5 1 -— — Pd, (2.6) with k the Boltzman constant, T the absolute temperature (Logan et al., 1995). The steady state solution (Equation 2.1) is based on the application of the advection dispersion equation (Equation 2.7) assuming first order removal of particles from the suspension. 25 dc d 2c dc , -= ^ -Y -V -— • dx dx (Eq. 2.7) ^att D in Equation 2.7 is the hydrodynamic dispersion coefficient (length2 time"1). Equation 2.7 describes the filtration process using a simple first order reaction term with katt (time"1) as rate coefficient. Additional terms for cell death or lysis can be included. However, cell death and lysis were considered negligible in this study since plate counts in the influent remained relatively constant and commonly varied less than 10 % during the time of the transport experiment. . Dispersion can commonly be neglected when calculating particle collision efficiencies in porous media beds at steady state (Unice and Logan, 2000). Thus, Equation 2.7 becomes at steady state. dc (Eq 2.8) c v dx Integration yields an equation containing a first order attachment coefficient in distance (kaU/v, length"1), which can be estimated from the suspended cell concentration profile over the length of the column. — = e x p (-^ x ) C0 ' (E q 2 .9 ) V The comparison of Equations 2.1 and 2.9 finally reveals that (Eq. 2.10) 2 Jcatt was estimated independently from two different sets of data; from suspended cell concentration profiles over the length of the column after 2 pore volumes and by fitting an analytical solution of Equation 2.7 (Equation 2.11, Van Genuchten and Alves, 1982, p. 60) to the effluent concentration profiles over time observed in the transport experiments. The analytical solution (Equation 2.11) was also utilized to model the experimental results. c(x,t)=Y (v-u)x exp v 2D gy/c y x —ut ijo t (v + + exp v m) x 2D g%/c / x+ut 24Ft (Eq. 2.11) 26 where \0.5 (Eq. 2.12) U - V v Results and Discussion Short GO cm') Column Experiments , Triplicate experiments utilizing 30 cm long columns were performed for both starved and vegetative cells. The dispersion coefficient and porosity which were estimated from tracer studies varied between 0.63 cm12 m in' and 2.83 cm2min'1, and 46 % and 50 %, respectively. The interstitial fluid velocity varied between 1.84 cm min ' and 3.09 cm min ' and the average influent concentration between 6.86 x IO7 CFU m l/' and 2.65 x IO8 CFUm l/' (Table 2.1). Table 2.1. ' Su m m ary of parameters for bacterial transport experiments in 30 cm porous media columns filled with 40 mesh quartz sand katt kan6 ■ 7.79 x IO7 0.0018 n.d.3 1.86 2.65 x IO8- 0.0008 0.0005 0.46 2.41 1.43 x IO8 0.0014 0.0011 2.83 0.48 2.51 8.75 x IO7 0.0029 n.d. 1.98 0.50 2.27 6.86 x IO7 0.0023 0.0020 2.32 0.50 1.84 9.84 x IO7 0.0022 0.0017 D1 Q2 Vi starved 0.63 0.50 3.09 cells 1.21 0.50 1.67 vegetative cells Co45 . 1Dispersion coefficient [cm2 min"'], estimated from tracer studies 2porosity, estimated from tracer studies 3 interstitial fluid velocity [cm min '] 4 influent cell concentration [CPU m l/'] 5first order attachment coefficient [s '] estimated using Equation 2.7 6 first order attachment coefficient [s '] estimated using Equation 2.9 7 n.d.: not determined 27 Breakthrough of Bacteria. Starved cells of Shewanella algae BrY were detected in the effluent of the columns at higher normalized concentrations than vegetative cells (Figure 2 .1). The average normalized breakthrough concentration for starved cells was 38 ± 6 % compared to 13 ± 1% for vegetative cells. The higher breakthrough of starved cells is also reflected in greater fractional recoveries (20 ± 3%) compared to vegetative cells (6 ± I %). A multiple regression analysis was performed to evaluate the effects of bacterial physiology (i.e. starved vs. vegetative cells), influent cell concentration, and interstitial fluid velocity on the fractional recovery and normalized breakthrough concentration of Shewanella algae BrY after transport through 30 cm quartz sand columns. The regression analysis revealed that, even after correcting for the potential influence of influent cell concentration and interstitial fluid velocity, the physiological state of the cells significantly influenced the transport behavior of cells in the column experiments. The resulting Pvalues of P = 0.006 and P = 0.005 for fractional recovery and normalized breakthrough concentration, respectively, clearly indicated the statistically significant difference between starved and vegetative cells of S. algae BrY. 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 pore volumes eluted Figure 2.1. Normalized breakthrough concentration (c/c0) of starved ( ♦ , ■, A) and vegetative (O , □, A ) cells of Shewanella algae BrY in the effluent of 30 cm long quartz sand columns and corresponding model fits (Equation 2.7, solid lines). 28 The corresponding model fits provided additional information about the transport behavior of . starved and vegetative Shewanella algae BrY cells (Figure 2.1). Vegetative cells appeared to break through slightly earlier (i.e. at 0.95 and 0.91 pore volumes) than predicted by the model in two columns. The third column, however, showed a delayed breakthrough (at 1.14 pore volumes) for vegetative cells in comparison to the model predictions. Starved cells were detected in the effluent approximately at (one column) or after the predicted times (two columns, at 1.08 and 1.14 pore volumes). Since the experiments were not designed to determine the exact time of breakthrough, it was difficult to derive definitive statements about the time of breakthrough. An early breakthrough of cells was maybe due to size dependent exclusion since the effective porosity for particles can be significantly lower than for solutes (Bales et al., 1989; Bitton et al., 1974; Morley et al., 1998). Since an early breakthrough was observed more frequently for the larger vegetative cells, pore size exclusion may have influenced the transport of S. algae BrY in the quartz sand columns. Distribution of Bacteria in the Pore Water. The enhanced transport of starved S. algae BrY cells was also reflected in the profiles of suspended cell concentrations for starved and vegetative cells after the injection of 2 pore volumes of cell suspension (Figure 2.2). A total of only four columns was considered for this analysis because plate counts for the third set of columns were unsuccessful. Using a least square method to fit Equation 2.9 to the experimental data, first order sorption coefficients (Arozz) of 5.3 x IO"4 s"1 and 1.1 x IO"3 s"1 were obtained for starved cells (Revalues: 0.580 and 0.870) and 2.0 x IO"3 s"1 and 1.7 x IO'3 s"1 for vegetative cells (R2-Values: 0.930 and 0.932). Despite the low regression coefficients for the starved cell concentration profiles, the smaller Aozz values provided supporting evidence that starved cells of S. algae BrY transport better through the quartz sand columns than their vegetative counterparts. A steady state model allowing for a bimodal distribution of first order sorption coefficients (Bolster et al., 2000) was applied to fit the experimental data but did not result in obvious improvements (data not shown). The Aozz values estimated using Equation 2.9 were in the same range, but consistently lower than the Aozz values derived from fitting the analytical solution of Equation 2.7 to the experimental data (see Table 2.1). These 29 differences might be due to the uncertainties in estimating the steady state breakthrough concentration from the experimental data (Figure 2.1), which was necessary for the calculation of katt using Equation 2.7. In addition, Equation 2.9 assumes steady state conditions, which might not have been achieved completely after two pore volumes in the 30 cm long columns. Since steady state could not be assumed, the ka„ values derived from fitting Equation 2.7 were used for subsequent calculations. The kalt values were used to predict the breakthrough of starved and vegetative cells through the 3 m long columns. 0.6 ■ distance into column [cm] Figure 2.2. Normalized concentrations of starved ( ♦ , ▲) and vegetative (□, O) cells of S. algae BrY in the pore water along the flowpath of 30 cm long quartz sand columns after injection of approximately 2 pore volumes of cell suspension. Model fits (Equation 2.9) are represented by the solid lines. Sorption of Bacteria Along the Flowpath. The distribution of sorbed cells throughout the 30 cm columns at the end of each experiment further supported the hypothesis that starved cells were better transported, less retarded, and more homogeneously distributed throughout the columns (Figure 2.3). The 30 number of cells desorbed per g of sand in each 5 cm section was normalized to the total number of cells injected to allow comparison between the different columns. It was evident that vegetative cells sorb more strongly than starved cells within the first few cm of the columns. The relative number of vegetative cells (CFU) detected per g of sand decreased from 7.17 x IO"3 ± 2.41 x IO"3 g"1 in the first 5 cm section to 3.63 x IO"3 ± 4.82 x IO"4 g"1 (51 % of sorbed bacteria detected in first 5 cm of column) at 15 to 20 cm into the column and then remained constant. The quantification limit of this method was approximately LS x l O4 CFU g'1 sand or approximately IO"6 g"1. Starved cells appeared to be more homogeneously distributed along the flowpath. The normalized concentration of starved cells decreased only slightly from 3.22 x IO"3 ± 2.24 x 10"4 g"1 within the first 5 cm of the column to 2.49 x IO 3 ± 3.97 x IO^g"1 (77 %) between 15 to 20 cm into the column and remained constant through the remainder the column. The large error bars obtained for sorbed vegetative cells in the influent region of the columns suggested a relatively high variability in the behavior of the three different vegetative cell cultures. Differences in the transport behavior of freshly cultivated cells were also noted by Johnson and Logan (1996) who reported significant day to day variations in transport behavior of equivalently treated cell cultures. The surface coverage o f the quartz sand with cells was less than 0.5 %. This estimate was established under the assumption that all cells, which were not detected in the effluent of the columns, adhered to the first 5 cm of the porous ,media column. An equivalent surface area for vegetative cells was calculated using the cell dimensions (2.2 mm x 0.63 mm) given by Caccavo et al. (1996). The total surface area covered by vegetative cells was then compared to the total surface area of the quartz sand in the first five centimeter of the column (15 g x 0.168 m2 g"1, as determined by surface area analysis using nitrogen gas). 31 1.0% 0 . 8% 0.4% 0 .2% 0 .0% distance into column [cm] Figure 2.3. Sorption of starved ( ♦ ) and vegetative (■) cells of S. algae BrY to quartz sand along the flowpath of 30 cm long columns. The sorbed cell concentrations were normalized to the total number of cells injected into each column. The averages of each three different columns and their corresponding standard errors of the mean are plotted. Recovery of Bacteria from Columns. The normalized effluent concentrations, fractional recoveries, total recoveries of cells from the experimental systems, and the distribution of the recovered cells in the effluent, sorbed to sand, and present in the pore water are listed in Table 2.2. As mentioned above, the normalized breakthrough concentration and fractional recoveries were significantly higher for starved cells than for vegetative cells. Total recoveries of cells from the experimental systems were limited using the desorption and plate count methods described. Between 37 % and 69 % of the total number of cells injected into each column were recovered by analyzing the effluent, the sand, and the pore water separately. Total recoveries were around 40 % with the exception of one case in which 69 % of the vegetative cells added were accounted for. No clear difference was observed between 32 the ability to recover starved or vegetative cells (2-sample t-test P = 0.482). Such low and inconsistent recoveries were reported by other researchers who used plate counts for the enumeration of bacteria in porous media transport experiments (Alexander et ah, 1991; Camper et ah, 1993; Fontes et al., 1991; Tan et ah, 1994). Since a decay of cells in the feed reservoir was not observed (the commonly observed variations remained within 10 % of the average influent concentration), the inability to recover all bacteria from the porous media columns might have been due to incomplete recovery of bacterial cells sorbed to the quartz sand. Cells may have been injured during the desorption procedure and were thus not recoverable using the plate count method employed. Bacterial decay or a change of bacterial cells into the viable but not culturable state during the passage through the porous media column might have also caused incomplete recoveries. The relative distribution of the recovered cells to the three phases (cells in the effluent, sorbed to sand, and in the pore water) appeared to be significantly different for starved and vegetative cells. While 36 ± 6 % of the recovered starved cells were detected in the effluent, only 15 ± 3 % of the recovered vegetative cells were detected there (2-sample t-test P = 0. 031). Relative recoveries of sorbed bacteria were higher for vegetative cells (72 ± 3 % compared to 47 ± 8 % for starved cells; P = 0.039). No significant difference was observed for the relative number of bacteria recovered in the pore water (17 + 4 % and 13 ± I % for starved and vegetative cells, respectively; P = 0.261). These distributions further supported the hypothesis that starved cells of S. algae BrY were better transported, less retarded, and more homogeneously distributed through porous media. 33 Table 2.2. Normalized breakthrough concentrations, absolute, and relative recoveries of cells from 30 cm columns. total replicate normalized fractional breakthrough recovery recovered concentration in effluent fraction of recovered cells detected on sand 1 fraction of fraction of recovered recovered cells cells detected in detected in effluent1 pore water 1 Starved I 34% 17% 37% 50% 31% 19% Cells 2 45% 22% 42% 37% 43% 20% 3 35% 20% 48% 53% 36% 12% vegetative I 14% 6% 43% 75% 12% 13% Cells 2 12% 5% 69% 70% 18% 12% 3 13% 7% 40% 71% 15% 14% 1 fractions calculated based on the total recovery of cells from the system. I.e. distribution of recovered cells to the three phases: Sorbed (detected on sand), transported through column (detected in effluent), and present in pore water at the end of the experiment (detected in pore water)____________________ Large (3 mi Column Experiments In order to develop economically viable bioaugmentation strategies, it is necessary to predict and control the transport of bacteria through porous media at field relevant scales. Intermediate scale experiments, utilizing 3 m long (2,54 cm diameter) columns were performed in duplicate to evaluate scale dependent effects on bacterial transport and to test the ability of the filtration model to predict bacterial transport at larger scales. A summary of the column parameters is given in Table 2.3. 34 Table 2.3. Summary of parameters for bacterial transport experiments in 3 m porous media columns filled with 40 mesh quartz sand D1 e2 3.85 0.49 cells 1.98 vegetative cells starved ^ Co4 katt katt 3.98 2.57 x IO7 0.000552 n.d.7 0.47 3.89 1.40 x IO8 0.000453 0.0006 3.64 0.50 3.98 3.86 x IO7 3.02 0.48 3.43 7.60 x IO7 . 0.0029 0.00091 n.d.7 0.0027 1Dispersion coefficient [cm2 min"1], estimated from tracer studies 2porosity, estimated from tracer studies 3 interstitial fluid velocity [cm min'1] 4 influent cell concentration [CPU mL"1] 5first order attachment coefficient [s"1] estimated using Equation 2.7 6 first order attachment coefficient [s'1] estimated using Equation 2.9 7 n.d.: not determined__________________________________ Breakthrough of Bacteria. The trends in concentrations of culturable cells in the effluent of the 3 m long columns were in agreement with the 30 cm scale results (Figure 2.4). However, the breakthrough occurred at lower concentrations since the bacteria had to pass through a longer column causing greater removal. The breakthrough of starved S. algae BrY cells occurred at higher concentrations (10 ± 3 %) than for their vegetative counterparts (0.5 ± 0.7 %) and the fractional recoveries were 4 ± I % for starved and 0.2 ± 0.3 % for vegetative cells. Despite the obvious visual difference between the breakthrough curves of starved and vegetative cells, two sample t-tests indicated only a low statistical significance for the observed differences in the normalized breakthrough concentration (P = 0.136) and fractional recovery (P — 0.079) of starved and vegetative cells. The low statistical significance was mainly due to the limited number of samples available (n = 4) and the high variability in normalized breakthrough concentrations and fractional recovery for vegetative cells. Multiple regression analyses could not be performed due to the limited number of samples available. It should also be noted that a steady state was possibly not completely achieved after two pore volumes in the 3 m long columns. The effluent cell concentrations appeared to still 35 increase when the experiments were terminated after two pore volumes of bacterial cell suspension were injected. The model fits (solid lines in Figure 2.4) indicated that both starved and vegetative cells tended to break through slightly later than the model predicted. It is possible that retardation of cells occurred in the columns, which would explain the delayed breakthrough (Gannon et al., 1991b; Harvey et al., 1989; McCaulou et al., 1994). Retardation can occur independently from the attenuation of cells (Harvey et al., 1989) and it might be that retardation, although it might have occurred in the smaller scale columns, was only observable consistently in the larger scale columns. 0.1 % 0.001 % 0 . 00001% pore volumes eluted Figure 2.4. Normalized breakthrough concentration of starved ( ♦ , ■) and vegetative (O , □) cells of Shewanella algae BrY in the effluent of 3 m long quartz sand columns and corresponding model fits (Equation 2.7, solid lines). Distribution of Bacteria in the Pore Water. Concentration profiles of suspended cells after two pore volumes were acquired for two columns, one for starved cells, one for vegetative cells (Figure 2.5). The kal, value obtained for starved cells using Equation 2.9 was 6.2 x IO^ s ' (R2 = 0.964) and 2.7 x IO3S 1 36 (R2 = 0.998) for vegetative cells. In contrast to the 30 cm long columns, the Icatt values estimated using Equation 2.9 were higher than the corresponding katt values derived from fitting the analytical solution of Equation 2.7 to the experimental data (see Table 2.3). The difference between the first order sorption coefficients estimated with Equation 2.9 and Equation 2.7 was already discussed for the 30 cm long columns and might have been due to incomplete recovery of bacteria in the pore water or the effluent of the columns. Although the katt values derived for 30 cm and 3 m long columns appeared to fall in a similar range, the katt values derived from the short columns could not predict the breakthrough concentrations in the 3 m long columns. Using the average katt derived from the. 30 cm long columns, the predicted normalized breakthrough concentration for 3 m long columns was 0.01 % and 0.0000009 % for starved and vegetative cells, respectively. These predicted concentrations did not agree with the normalized breakthrough concentrations of starved (10 %) and vegetative cells (0.5 %) observed in 3 m column ■ experiments. Due to the exponential relationship of Equation 2.9 slight changes in katt resulted in significant changes in the effluent concentration. Scale dependent effects on bacterial transport through porous media were described by Harvey et al. (1993) and Martin et al. (1996). Harvey et al. (1993) stated the importance of small and large scale variability of heterogeneities on the results of bacterial transport, Martin et al. (1996) observed decreases in the collision efficiency factor (<% also called sticking efficiency), and thus decreasing Icatl with increasing distance into the porous media columns. These observations indicated that the estimation of katt and a from laboratory scale experiments do not necessarily allow the accurate prediction of field scale results. In addition, a number of researchers recently suggested the existence of subpopulatiohs with different transport behaviors in monoclonal cultures (Albinger et al., 1994). The presence of a subpopulation of sticky bacteria resulted in a higher than predicted number of bacteria adhering to the porous media grains close to the injection point while the less sticky subpopulation traveled extremely well through the porous medium, resulting in higher than expected effluent concentrations. 37 O 50 100 150 200 250 300 distance into column [cm] Figure 2.5. Normalized concentrations of starved ( ♦ ) and vegetative (■) cells of S. algae BrY in the pore water along the flowpath of 3 m long quartz sand columns after injection of approximately 2 pore volumes of cell suspension. Model fits (Equation 2.9) are represented by the solid lines. Sorption of Bacteria Along the Flowpath. The distribution of sorbed cells throughout the columns at the end of each experiment (Figure 2.6) showed the same trend as for the 30 cm column experiments. Vegetative cells appeared to attach more strongly to the quartz sand close to the injection point than starved cells. The normalized cell concentrations of vegetative cells decreased rapidly with distance into the column. Between 120 cm and 150 cm into the column, the relative concentration of vegetative cells sorbed to sand had dropped to approximately 0.5 % (9.31 x 10"6 g ') of the concentration detected in the first 30 cm of the column (1.71 x IO 3 g '). The relative concentration of starved cells sorbed to sand between 120 cm and 150 cm into the column (6.35 x IO"5 g"1) was at 9 % of the concentration detected in the first 30 cm (7.28 x IO"4 g"1). The relative sorbed cell concentrations in the last 30 cm of the column (270 cm to 300 cm section) was 2.75 x IO"6 g"1(0.2 % of sorbed cell concentration in first 30 cm of column) and 1.54 x 105 g ' 38 (2 %) for vegetative and starved cells, respectively. As in the short column experiments, large standard errors were observed for sorbed vegetative cells, indicating higher culture to culture variability for vegetative cells. As in the 30 cm columns, the surface coverage of the quartz sand with vegetative cells was less than 0.5 %. distance into column [cm] Figure 2.6. Sorption of starved ( ♦ ) and vegetative (■) cells of S. algae BrY to quartz sand along the Oowpath of 3 m long columns. The sorbed cell concentrations were normalized to the total number of cells injected into each column. The averages of each two different columns and their corresponding standard errors of the mean are plotted. Recovery of Bacteria from Columns. Table 2.4 lists the normalized effluent concentrations, fractional recoveries, total recoveries of cells from the experimental systems, and the distribution of the recovered cells in the effluent, sorbed to sand, and present in the pore water. 39 Table 2.4. Normalized breakthrough concentrations, absolute, and relative recoveries of cells from 3 m columns total replicate normalized fractional breakthrough recovery recovered concentration in effluent fraction of recovered cells detected on sand 1 Fraction of recovered cells detected in effluent1 fraction of recovered cells detected in pore water 1 starved I 8% 4% 51% 91.5% 7% 0.97% cells 2 12% 5% 38% 86.3% 13% 1.15% vegetative I 0.004% 0.0006% 37% 99.9% ■ 0.002% 0.11% cells 2 1.0% 0.4% 55% 98.7% 0.7% 0.63% 1 fractions calculated based on the total recovery of cells from the system. I.e. distribution of recovered cells to the three phases: Sorbed (detected on sand), transported through column (detected in effluent), and present in pore water at the end of the experiment (detected in pore water)____________________ The normalized breakthrough concentration and fractional recoveries were compared above (“Breakthrough of Bacteria"). The total recoveries of cells from the 3 m long columns fall in the same range as for the 30 cm columns; between 37 % and 55 % of the total number of cells injected into each column were recovered. The relative distribution of recovered cells to the three phases (effluent, sorbed to the sand, and pore water) also followed the same pattern that was described for the 30 cm long columns. Due to the limited number of total samples (n = 4) and the high variability in the behavior of the vegetative cell cultures, significance levels of two sample t-tests remained low. The fraction of recovered vegetative cells ,sorbed to the quartz sand (99 ± I %) was higher than for starved cells (89 ± 4 % ; P = 0.160). This was balanced by a higher fraction of recovered starved cells in the effluent (10 ± 4 %) compared to vegetative cells (0.4 ± 0.5 %; P = 0.163). The difference between the fraction of recovered starved and vegetative cells detected in the pore water was less significant than for the other two phases (P = 0.222). Modeling and Scale Up Issues In order to apply the. starvation transport enhancement strategy in the field, predictions about the transport behavior of starved cells must be made on field relevant scales. The first order attachment 40 coefficients (Icatt) estimated in this work and their observed decrease at larger scales agreed with the ' observations made by others (McCaulou et al., 1995; Schijven et ah, 1999). The scale dependency of Icath however, did not allow to predict the breakthrough of starved or vegetative cells through 3 m long columns based on 30 cm column results. A scale dependency of Icalt was described and is known to limit the applicability of the simple first order attachment terms to predict results on larger scales (Martin et al., 1996). In addition, the transport model used was derived for spherical collectors. The quartz sand used in this study was irregularly shaped and this might have contributed to the observed discrepancies between the experimental data and the modeling predictions at both, the 30 cm and the 3 m scale. Another potential explanation for the inability to model the experimental results using the filtration theory with first order attachment is population variability within the bacterial cultures. The presence of subpopulations with different transport behaviors and the observed scale dependent effects would require the application of transport models, which could potentially account for spatial and intrapopulation heterogeneities of transport related parameters. Detailed studies recently revealed the presence of subpopulations within monoclonal cultures (Albinger et al., 1994; Baygents et al., 1998; Bolster et al., 2000; Simoni et al., 1998) and thus multi-parameter models might be more appropriate to describe bacterial transport through porous media. However, for the development of a conceptual model describing population- and scale-dependent effects, very specific knowledge of the parameters influencing bacterial transport is required. This research, however, did not allow to reveal specific mechanisms that might be responsible for the inability to use short columns in order to predict bacterial transport at larger scales. Thus, the choice of a more appropriate transport model was not obvious and was not pursued in this research. Influence of Influent Cell Concentration and Fluid Velocity Additional improvements in bacterial transport through porous media could be achieved by optimizing other parameters which are known to influence bacterial transport and which can be controlled in field applications. Preliminary experiments evaluating the influence of influent cell concentration and interstitial fluid velocity on the transport of starved S. algae BrY were performed in 30 cm long columns. A 41 change in influent cell concentration from 2.28 x IO6 CFU ml'1to 1.27 x IO7 CFU ml"1resulted in a change in fractional recovery from 21 % to 17 % for the low and high cell concentration, respectively. A change in interstitial fluid velocity from approximately 3 cm min"1 to approximately 13 cm min"1 only led to an increase in fractional recovery from 17 % to 19 % for low and high fluid velocity, respectively. These data suggest that the control of influent cell concentration and fluid velocity can potentially influence the bacterial transport through porous media but their importance appears to be low compared to the influence of starvation. Conclusions A major problem that is preventing the widespread implementation of bioaugmentation strategies is the tendency of bacteria to strongly adhere to the surfaces in the immediate vicinity of the injection point. Biofouling of these areas can lead to a failure of the injection system and limited dispersion of the bacterial inoculum in the subsurface. The use of starved bacteria might overcome this problem since starved bacteria typically adhere less, transport farther, and distribute more homogeneously through porous media than vegetative cells. The promising potential of starved cells for bioaugmentation strategies is indicated by higher fractional recoveries (3.2 fold higher for 30 cin columns, 22.6 fold higher for 3 m columns), higher normalized breakthrough concentrations (2.9 and 19.9 fold), and lower first order attachment coefficients Qcatt, 1.8 fold lower for 30 cm columns, and 3.2 fold lower for 3 m columns). In addition to their improved transport behavior, starved cells are metabolically dormant. Their metabolic dormancy and lag phase before regaining full metabolic activity would allow for the simultaneous or subsequent injection of nutrients without the danger of immediate biofouling of the employed injection system. There is sufficient evidence to assume that transport enhancement by bacterial long term starvation is not limited to Shewanella algae BrY but is a rather general starvation response across many bacterial strains (Amy and Haldeman, 1997; Kjelleberg, 1993a; Lappin-Scott and Costerton, 1990). Thus, the starvation transport enhancement strategy could be applicable to a large number of bacterial strains with 42 a variety of metabolic capabilities. If successfully developed, the starvation transport enhancement strategy would provide the practitioner with a versatile tool for the development of bioaugmentation technologies. Additional Research Needs In order to better understand the transport behavior of S. algae BrY cells, the parameters governing the transport of starved bacteria have to be identified. According to the filtration theory (Equation 2.10) Jcalt is influenced by 77 and a, v, and 0. While the fluid velocity (v) and the porosity (0) are independent from the inoculum used, 77 and a account for properties of the bacteria. The single collector efficiency (77) can be calculated using Equations 2.2 through 2.6. The cell diameter (dp), diffusion coefficient (Dp), and buoyant density (pp) are the cell properties which can account for differences in 77 between starved and vegetative cells. The collision efficiency (a) incorporates all physicochemical interactions between the bacteria and the porous medium, and in addition any inaccuracies in the calculation of 77 (Martin et ah, 1996). Equations for the calculation of a, however are not available for bacteria. This indicates that the currently existing filtration models do not yet account for these physicochemical interactions and inaccuracies. Changes in cell shape, surface charge, hydrophobicity, and the presence of pili and flagella are likely to influence a and therefore influence bacterial attachment to surfaces and transport through porous media. It was considered important to monitor the change of several of these bacterial cell properties in an attempt to identify parameters responsible for the starvation transport enhancement. The results of these investigations and their implications are summarized in Chapter 3. Once a better understanding of the factors governing the transport of starved S. algae BrY cells has been gained, a model accounting for these factors must be developed. However, until the processes governing bacterial transport through porous media are sufficiently understood and appropriate models developed, transport experiments should be conducted at field relevant scales, with the bacteria and subsurface material of interest, and by mimicking field conditions as closely as possible. .43 Since there is not necessarily a clear relationship between attenuation and retardation of bacteria in porous media it is extremely difficult to extrapolate from column breakthrough data to the distribution of sorbed cells (Harvey et ah, 1989). Therefore, the development of more accurate methods for the quantification of sorbed bacteria and their activity is essential to develop a thorough understanding of bacterial transport through porous media. A more complete recovery of cells from porous media would allow to better estimate the distribution of cells along the flowpath and therefore allow to describe the adhesion processes better. Using plate counts, between 30 % and 60 % of the injected cells remained unaccounted for in this study. Such a low recovery allows for much speculation and can result in the development of mathematical models which, although they accurately describe the experimental data, might not be based on a thorough understanding of the processes involved. Since the activity of sorbed bacteria is the most important factor determining the success of bioaugmentation strategies, it would be desirable to develop procedures which allow for the enumeration and activity assessment of sorbed bacteria simultaneously. The recent development of molecular techniques has rapidly improved our ability to selectively detect and quantify bacteria in environmental samples. The most commonly used suite of molecular methods for the identification and quantification of bacterial numbers and activity consists of selective ribosomal KNA (rRNA) and DNA extraction, quantitative polymerase chain reaction (qPCR) and reverse transcriptase PCR (RT-PCR), denaturing gradient gel electrophoresis (DGGE) fingerprinting, and nucleic acid sequencing. The detection and quantification of rRNA and rDNA (DNA encoding for the rRNA) could be utilized for quantification and activity assessments. The rRNA content of cells can be correlated with bacterial activity while the amount of rDNA estimates bacterial numbers (Schafer et al., 2001). Consequently, the rRNA to rDNA-ratio can be used as an estimate for the specific activity of bacterial cells (Duineveld et al., 2001). Unfortunately, the techniques currently available are I) mostly designed for aqueous samples, 2) still fairly work intensive, and 3) require the use of expensive equipment and supplies (qPCR equipment and fluorescent labels). The current development of nucleic acid microarrays might provide an opportunity to develop these methods more 44 inexpensively. Successfully developed microarray technologies might allow for the inexpensive and selective enumeration and activity assessment of bacterial cells in environmental samples. 45 CHAPTER 3 CHANGE OF TRANSPORT-RELATED CELL PROPERTIES DURING STARVATION OF SHEWANELLA ALGAE BrY Abstract The enhanced transport of long term nutrient starved bacteria has often been attributed to the decrease in average cell size during starvation. However, according to the filtration theory, cell size alone is not sufficient to explain the difference in transportability between starved and vegetative cells of Shewanella algae BrY. This chapter documents the changes of a number of transport related cell properties o f Shewanella algae BrY during carbon and nitrogen starvation. The cell buoyant density decreased during starvation, the apparent diffusion coefficient and the cell surface charge (measured as zeta potential) increased but no change in hydrophobicity'was observed. According to the filtration model applied, the cell size, buoyant density, and diffusion coefficient of starved cells could partly explain the improved transport behavior of starved cells through porous media while the transport of vegetative cells was drastically overpredicted. The observed slight changes in cell surface charge and hydrophobicity were not sufficient to explain the difference between starved and vegetative cells. It was concluded that parameters beyond the ones investigated appear to be responsible for the differences in porous media transport of starved and vegetative S. algae BrY cells and that a better understanding of these parameters must be obtained before an appropriate mathematical model can be developed. Introduction Chapter 2 clearly demonstrated that starved Shewanella algae BrY cells are transported better through porous media than their vegetative counteiparts. However, explanations for the improved 46 transportability of starved cells are lacking. Changes in cell size and cell shape have been correlated with transport through porous media (Bitton et al, 1974; Fontes et ah, 1991; Gannon et ah, 1991a; Weiss et ah, 1995) and were suggested to be at least partly responsible for the improved transport of starved bacteria (Lappin-Scott and Costerton, 1992; Lappin-Scott et al., 1988b). However, based on the filtration theory, the differences in cell size between starved and vegetative cells of Shewanella algae BrY were not sufficient to explain the observed differences in bacterial transport. Caccavo et al. (1996) reported a change in cell size from approximately 2.2 pm x 0.63 pm (length x width) to more spherical cells of approximately 1.04 pm x 0.63 pm during starvation. Based on the one­ dimensional filtration theory (Equation 2.1), this change would lead to an only slightly increased breakthrough (c/c0) of starved cells. Figure 3.1 shows the steady state concentration in the effluent of 30 cm long columns for particles in the range from 0.1 pm to 5 pm. A list of the parameters used in these calculations is given in Table 3.1. According to these calculations, starved cells of S. algae BrY break through the 30 cm long porous rhedia columns at approximately 47.2 % of the influent cell concentration while vegetative cells are predicted to break through at 46.6 %. The experimental data summarized in Chapter 2 show slightly lower breakthrough concentrations for starved cells (38 %) and significantly lower concentrations for vegetative cells (13 %). Thus, parameters other than the cell size are likely to govern the transport of starved and vegetative S. algae BrY cells through porous media columns. Equation 2.1 suggests that, if the filtration theory and the parameters applied here are appropriate for the description of bacterial transport through porous media, changes in the single collector efficiency (77) or the collision efficiency factor (a) must be responsible for the observed changes in normalized breakthrough concentrations (c/c0). The single collector efficiency (77) is a measure for the rate at which particles strike porous media grains and is influenced by the size, diffusion coefficient, and buoyant density of the bacterial cells (see Equations 2.2 through 2.6). The collision efficiency factor (ti) is a measure for the number of collisions, which result in the adhesion of particles to the porous media grains. The collision efficiency factor is also referred to as sticking efficiency in bacterial transport experiments. In contrast to Tj there is no theoretical basis available for the calculation of (X, it allows to account for physicochemical 47 interactions between the bacteria and the porous media grains (Martin et ah, 1996) and should remain between zero and one (Logan et ah, 1995). Figure 3.1. Predicted, normalized, steady-state breakthrough concentrations through 30 cm quartz sand columns of particles with diameters between 0.1 pm and 5 pm according to the filtration theory (Equation 2.1, Logan et ah, 1995). The arrows indicate the calculated breakthrough concentrations for starved and vegetative cells of S. algae BrY, whose sizes were reported to be 1.04 pm and 2.2 pm, respectively (Caccavo, Jr. et ah, 1996). The curve was generated with the values listed in Table 3.1. The observed differences in porous media transport between starved and vegetative cells and the inaccurate model predictions discussed in Chapter 2 make it necessary to investigate the parameters which potentially influence 77 and a, and thus the transport and adhesion of bacteria in porous media. The diameter (dp), buoyant density (pp) and diffusion coefficient of the cells (Dp) are known to influence 77 (see Equations 2.2 through 2.6). Changes in cell hydrophobicity, net electrostatic charge, character and properties of surface molecules, and the presence of flagella and pili are likely to influence the ability of cells to adhere to surfaces after a collision has occurred, i.e. influence the collision efficiency or stickiness of cells (a). 48 Table 3.1. Parameters used for the calculation of normalized breakthrough concentrations in 30 cm columns as shown in Figure 3.1. Param eter Value Parameter Value P ' 998.2 kg nr3 V 2 cm min"1 Pp 1044 kg m"3 g 9.80665 m s"2 dc 0.5 mm 1 P 1.002 x IO"3 kg m"1 s" T 293.16 K 0 0.5 k 1.3048 x IO23 JK"1 a I2 4 0.1 |tm to 5 |rm X 0.30 m 1 effective filtration size (manufacturer information, Unimin Coip., Emmet, ID) 2 a= I indicates complete destabilization of the bacterial suspension, i.e. every collision of a bacterial cell with a porous media grain results in the removal of the bacterium from the cell suspension.______________________________________________ The changes in cell diameter during starvation of S. algae BrY were previously described by Caccavo et al. (1996). These data were used for the development of Figure 3.1. The research summarized in this chapter describes the changes of cell numbers (colony forming units, direct cell counts, and absorbance at 600 run), hydrophobicity, apparent diffusion coefficient, electrostatic charge, apparent buoyant density, adhesivity to quartz sand, and transportability through 30 cm quartz sand columns of Shewanella algae BrY over a starvation period of 7 weeks. The findings are expected to improve the understanding of bacterial transport enhancement by starvation and contribute towards the development of a model appropriately describing the transport of starved and vegetative bacteria through porous media. Materials and Methods Strains and Culturing Methods Cultures of Shewanella algae BrY were maintained and grown as described in Chapter 2. Starved cells were harvested weekly by centrifugation and resuspended in PBS for all assays, except, for microscopic direct counts, viable cell counts, and absorption measurements, which were performed with 49 samples taken directly from the starvation cultures. Shewanella algae BrY-gfp (gfp = green fluorescent protein) was graciously provided by Yuri Gorby (Pacific Northwest National Laboratory, Richland, WA). S. algae BrY-gfp was cultured and maintained in the presence of 20 mg L"1kanamycin. ■ Cell Quantification and Cell Imaging Cells were quantified using absorbance measurements at 600 nm, microscopic direct counts, and dilution plate counts using tryptic soy agar (TSA, Difco). Absorbance measurement were performed using a Shimadzu spectrophotometer (UV-2101 PC). Microscopic direct counts were performed after fixing cells in 4 % formaldehyde solution and staining cells on dark polycarbonate filters (Whatman, Clifton, NJ) with 50 mg L'1 acridine orange or 10 mg L"1 4,,6-diamidino-2-phenylindole (DAPI) solution for one to five minutes. The filters were examined using a Nikon Eclipse E800 epifiuorescence microscope with a mercury lamp UV source. The filter blocks used had an excitation/emission bandwidth of 450-490 nm/>515 nm and 340-380 nm/435-495 nm for acridine orange and DAPI, respectively. Electronic images of S. algae BrY-gfp cells were acquired using the same microscope with mercury lamp, the 450-490 nm/>515 nm filter block and a charge coupled device (CCD) camera (Princeton Instruments, Tenton, NJ). Cell sizes were estimated manually from the acquired images by measuring cell width and length and comparing it to the appropriate scale bar. Plate counts were performed after dilution in PBS by spreading 0.1 TnT. of bacterial cell suspension onto tryptic soy agar. Plates were incubated at 30 °C for 24 hours and colonies were counted using a 5x magnification glass. Hvdronhobicitv Changes in cell hydrophobicity were estimated using the bacterial adhesion to hydrocarbons (BATH) assay (Rosenberg, 1984; Rosenberg et al., 1980). In brief, cell suspensions (5 mL) with an absorbance (600 nm) between 0.120 and 0.180 were added to acid washed screw cap vials. Hexadecane (2 mL, Fisher, Pittsburgh, PA, previously filtered through a 0.2 pm membrane) was added, the vials were 50 closed, and vortexed for 30 seconds. The hexadecane and aqueous phases were allowed to separate for 15 minutes and an aliquot of the aqueous phase was transferred carefully to a cuvette using a glass pasteur pipette. The absorbances (600 run) of the treatments were compared to control treatments lacking hexadecane. All measurements were conducted in triplicate. The hydrophobicity index (HI) was calculated according to H I- A (AtOnm,initial - A (AMnm9Jinal (Eq. 3.1) A (AMnm9Initial . Zeta Potential and Diffusion Coefficient The zeta potential and the apparent diffusion coefficient of the cells was determined from cell suspensions in PBS at pH 7.0 and 25 °C using a Zetasizer 2c (Malvern Instruments). Zeta potentials and apparent diffusion coefficients were determined from the average of 10 measurements. Buoyant Density The apparent buoyant density was calculated using the density gradient centrifugation procedure described by Harvey et al. (1997). In brief, 2.5 m L o f a bacterial suspension was carefully layered on top of 40 mL of a Percoll/PBS solution at pH 7 (Sigma, St. Louis, MO) and the tubes were centrifuged at 15,000 x g for 60 minutes. Immediately after centrifugation, the position of the bacterial bands in the tubes was compared to bands formed by brightly colored beads of known density (DMB-10, Sigma, St. Louis, MO). Adhesion Assay Adhesion studies were performed as described previously (Caccavo, Jr. et al., 1997). In brief, cell suspensions in PBS were added to 2 mL mirocentrifuge vials containing 0.1 g of quartz sand (approximate surface area 0.0168 m2). The vials were placed on a horizontal shaker at 150 rpm for 30 minutes, centrifuged at 300 x g for 5 minutes, and the supernatant was sampled for total microscopic cell counts 51 using acridine orange. Cell concentrations between 1.17 x IO7 and 1.93 x IO8 cells per ml were used and the adhesion assays were performed in triplicate. The number of adhered cells was calculated by subtracting the number of unadhered cells from the number of total cells. Adhesion of cells to the tubes and cell lysis were accounted for with controls of adhesion assay buffer without quartz sand. Transportability Through Porous Media T r an sp o rt experiments were performed with vegetative cells, and cells starved for one, four, and seven weeks in the same manner as described in Chapter 2. Results and Discussion Figure 3.2 shows images of Shewanella algae BrY-gfp cells during long term starvation. Cells immediately after culturing and washing were between 1.8 pm and 3.0 pm in length and approximately 0.6 Iim in width. Cells fluoresced brightly when observed under ultraviolet (UV) light with the appropriate filter set. In the upper left comer of image (a) ongoing cell divisions were captured. Two cells were separating from each other at the end of a cell division cycle at the time the image was acquired. S. algae BrY-gfp cells shrunk over time and were approximately 1.5 pm in length and 0.6 pm in width after 7 weeks but still fluoresced due to the presence of gfp (image b). The cells appeared to be perfectly spherical with an approximate diameter of 0.6 pm after 6 months of starvation (image c). S. algae BrY-gfp cells still fluoresced after 6 months and could be resuscitated on tryptic soy agar (data not shown). Visual observation indicated that S. algae BrY-gfp cells were motile until the third week but mostly non motile at longer starvation times. 52 I l 10 |im Figure 3.2. UV fluorescence images of Shewanella algae BrY gfp cells immediately after culturing and washing (a), after 7 weeks of starvation (b), after 6 months of starvation (c). All images were obtained from live specimen without additional staining. Cell Quantification Viable cells from Shewanella algae BrY starvation cultures were estimated as colony forming units on tryptic soy agar. The concentration of viable cells slightly increased during the first week of starvation from initially 3.83 x IO9 ± 1.33 x IO8 CFU mL"1 to 4.93 x IO9 ± 4.84 x IO8 CFU mV1 (Figure 3.3). The concentration of cells decreased to 2.10 x IO8 ± 1.04 x IO7 CFU mL"1 after seven weeks of starvation. Initially total cell counts were approximately ten times higher (3.15 x IO10 ± 2.05 x IO9 cells mL"1) than the concentration of colony forming units. The total concentration of cells decreased slightly during the first week to 2.77 x IO10 ± 2.02 x IO9 cells mL 1and was approximately 2.5 times higher than the concentration of viable cells. The numbers further decreased over time and reached 4.98 x IO9 ± 1.05 x IO8 cells mL 1 after 7 weeks, which was approximately 25 times higher than the concentration of colony forming units at that time. The absorbance (600 run) of the cell suspensions decreased almost linearly from 1.22 ± 0.06 to 0.172 ± 0.01 over seven weeks of starvation. These results concur with the observations by Caccavo et al. (1996) who reported a decrease in viable cell numbers during starvation. The change in the 53 ratio of culturable to total cells from initially ten to approximately 25 after seven weeks of starvation indicated an increase in the fraction of cells that were physically intact but not culturable. Resuscitation methods other than plate counts on full strength tryptic soy agar might have allowed for a better recovery of S. algae BrY cells from long term starvation, however, the development of improved recovery methods was not focus of this research. 1.5 1.2 E 0.9 I E c S- 5 "O 0.6 I S S £ O 0.3 0.0 0 1 2 3 4 5 6 7 Time [weeks] Figure 3.3. Concentration of total cells (♦ ), culturable cells (■), and absorbance at 600 nm (▲) during starvation of Shewanella algae BrY. Error bars represent the standard error of the means (n = 3). Hvdrophobicitv and Zeta Potential The change in hydrophobicity and zeta potential during starvation is shown in Figure 3.4. The hydrophobicity index appeared to slightly increase during the first 3 weeks of starvation but decreased again starting in week 4. It remained between 27 % and 42 % until week 5 and dropped slightly to 18 ± 13 % after 7 weeks of starvation. The standard errors obtained using the BATH assay, however, were too 54 large to determine definite trends. An analysis of variance was performed to obtain pooled variances for a two sample t-test. The t-test failed to indicate a statistically significant difference between the initial and final hydrophobicity indices (P = 0.66). The zeta potential slightly increased over the 7 week period. Initially -39.5 ± 1.6 mV were measured, which agrees with the value of -37.7 ± 0.7 mV reported by Caccavo et al. (1997). The zeta potential remained at approximately -40 mV until week 4 but then increased to reach -33.3 ± 2.8 mV after 7 weeks. However, this change in zeta potential could not explain the observed transport improvement reported in Chapter 2. A change of net electrostatic charge to more positive values should have resulted in increased electrostatic interaction between the negatively charged quartz sand and the bacterial cells (Sharma et al., 1985). However, as shown in Chapter 2, the transport of S. algae BrY cells was improved and not inhibited by starvation. Thus, electrostatic interactions were not likely to be responsible for the improved transport of starved S. algae BrY cells through quartz sand at circumneutral pH. TL -30 100 % 5 50% time [weeks] Figure 3.4. Change in hydrophobicity index ( ♦ ) and zeta potential (■) during starvation of Shewanella algae BrY. Error bars represent the standard error of the means (n —3 and n=10 for hydrophobicity index and zeta potential, respectively). 55 Diffusion Coefficient and Buoyant Density The apparent diffusion coefficient and the apparent buoyant density both changed slightly during starvation (Figure 3.5). The diffusion coefficient increased from 2.7 x IO"9 ± 1.11 x IO"10 cm2 s ' to 4.79 x IO 9 ± 1.43 x IO"'9 cm2 s ' after seven weeks and the apparent buoyant density decreased from 1.044 g cm"3 to 1.020 g cm"3. I I I C O ■a I 1.02 time [weeks] Figure 3.5. Change of apparent buoyant density ( ♦ ) and apparent diffusion coefficient (■) during starvation of Shewanella algae BrY. Error bars represent the standard error of the means (n = 3 and n=l0 for buoyant density and diffusion coefficient, respectively). Modeling Results Influence of Diffusion Coefficient and Buoyant Density. Calculation of Tl. The measured values for the cell diameter (1.04 ]im and 2.2 pm, Caccavo et al. 1996), buoyant density (1.020 g cm"3 and 1.044 g cm"3), and diffusion coefficient (2.7 x 109 cm2 s ' and 4.79 x 109 cm2 s ') of starved and vegetative cells 56 (respectively) were used to calculate the theoretical breakthrough concentrations according to Equations 2.1 through 2.5. All other values were used as listed in Table 3.1. The observed increase of the apparent diffusion coefficient and the decrease in buoyant density alone were not sufficient to explain the differences in breakthrough concentrations of starved and vegetative cells. The breakthrough concentrations calculated for starved cells (45.4 %) were slightly closer to the experimentally observed breakthrough (38 %) after accounting for buoyant density (pp) and diffusion coefficient (Z)p) indicating that pp and Dp have a slight influence on the transport behavior of bacterial cells. However, the model failed to predict the breakthrough of vegetative cells. A normalized breakthrough concentration of 42.0 % was predicted and only 13 % were observed in the 30 cm column experiments summarized in Chapter 2. , The determination of buoyant density (pp) and diffusion coefficient (Dp) also allowed the calculation of the single collector efficiencies (77) for starved and vegetative cells. However, the differences in 7j between starved (7.62 x IO-4 ± 1.30 x IO"4, see Table 3.2 for listing) and vegetative cells (9.37 x IO"4 ± 1.24 x 10"4) could only account for part of the differences in breakthrough concentrations and were not statistically significantly different (two sample t-test, P = 0.190). Thus, if all other parameters of the model were correct, changes in the collision efficiency factor (a) must have been responsible for the differences between starved and vegetative cells. Table 3.2. Estimated single collector efficiencies (77) and collision efficiency factors (a) for starved and vegetative S. algae BrY cells, 30 cm columns. starved vegetative n a % a 0.000637 1.846 0.000794 2.660 0.000896 1.006 0.001008 2.494 0.000752 1.284 0.001008 2.418 57 Calculation of Collision Efficiencies a. The estimation of 77 for each column allowed the calculation of a values for starved and vegetative cells using the first order attachment coefficients estimated in Chapter 2 (Table 2.1) and Equation 2.10. The or values obtained for starved cells (1.38 ± 0.43) were statistically significantly lower than for vegetative cells (2.55 ± 0.12, P = 0.049). This difference in a values, indicated that differences in the sticking efficiencies (ti) were partly responsible for the observed differences in porous media transport of starved and vegetative cells, assuming that all other values used for the calculations, such as cell size were correct (see Table 3.2 for list of orvalues). The sticking efficiency (or) was introduced into the filtration theory to account for physicochemical interactions between the particles (bacteria) and the porous media grains (Martin et al., 1996). However, since there is no theoretical basis available for the calculation of or, it is commonly estimated from filtration experiments. It, thus, also indirectly serves to account for any inaccuracies in the filtration theory. Since the filtration model was derived for spherical collectors, such inaccuracies can derive from the application of the filtration model to porous media grains that are irregularly shaped like, the quartz sand used in this study. According to the filtration theory (Logan et al., 1995), values for a should remain between zero and one. However, values of a greater than one have been reported to be necessary to accommodate for differences in model predictions and experimental data (Baygents et al., 1998; Logan et al., 1995; Martin et al., 1996). It was concluded that the filtration theory, as applied here, was not completely appropriate for describing the experimental data from bacterial transport experiments (Chapter 2) unless values of greater than one were assigned to the sticking efficiency (ti). The cell surface hydrophobicity and the net electrostatic charge could have potentially influenced the physicochemical interaction between bacteria and porous media grains, and therefore a. However, as discussed above, no significant changes in cell surface hydrophobicity were observed using the BATH assay. However, since the BATH assay measurements were associated with high standard errors, future research should evaluate other methods that allow to estimate cell surface hydrophobicity more accurately. Hydrophobic interaction chromatography utilizing hydrophobic resins such as phenyl- or octyl-sepharose is 58 potentially suitable for this purpose. The observed changes in zeta potential to less negative values during starvation cannot explain the increased transport of starved bacteria either. A change of cell surface charge to less negative values should have resulted in increased cell quartz sand interactions and thus decreased transportability. Applicability of Filtration Model to 3 m Column Results The measurement of the buoyant density (Pp) and diffusion coefficient (Dp) were used to predict the transport through the 3 m long columns. However, as discussed in Chapter 2, the model drastically underpredicted the breakthrough concentrations for both starved and vegetative cells. While breakthrough concentrations of 0.04 % were predicted for starved cells, 10 % ± 3 % were observed in the experiments described in Chapter 2. The breakthrough of vegetative cells was also underpredicted; a normalized effluent concentration of 0.017 % was predicted but 0.4 % ± 0.5 % were observed. While t] was not significantly different for starved and vegetative cells (starved: 5.37 x 10"4 ± 0; vegetative 5.85 x 10"4 ± 2.88 x IO"5, see Table 3.3), a values were lower for starved cells (0.46 + 0.06) than for vegetative cells (1.38 ± 0.87). However, a two sample t-test failed to indicate a statistically significant difference between the a values of starved and vegetative cells (P = 0.374) due to the high variability observed in the transportability of vegetative cells. The a values calculated for the 3 m columns were significantly smaller than the ones calculated for 30 cm long columns (see Table 3.2). A decrease of collision efficiency factors (a) with increasing distance into a porous medium was also observed by Martin et al. (1996) and is a further indication for scale dependent effects on bacterial transport. Table 3.3. Estimated single collector efficiencies (if) and collision efficiency factors (d) for starved and vegetative S. algae BrY cells, 3 m columns. vegetative starved n Ctf V a 0.000537 0.502 0.000505 1.995 0.000537 0.412 0.000606 0.767 59 Batch Adhesion Assays as a Predictive Tool The tendency of S. algae BrY cells to adhere to quartz sand in batch systems was monitored in order to provide a tool that would potentially allow to screen bacterial strains for their applicability to the starvation transport enhancement strategy. The fraction of cells adhered to quartz sand in batch experiments and the fractional recovery of cells after passage through 30 cm quartz sand columns are shown in Figure 3.6. The adhesion assay showed a significant decrease in the fraction of cells adhered to quartz sand over the first two weeks of starvation (t-test with pooled standard errors, P = 0.05) and the fraction attached remained below 10 % in average during the seven weeks of starvation. It should be noted that the adhesion assay yielded high standard errors and negative adhesion values after two and five weeks, i.e. a higher cell number was observed in the treatments containing quartz sand than in the controls. It is obvious, though, that there was an inverse relationship between the number of cells adhered to quartz sand and the fractional recovery of cells (Figure 3.6). 100% 80 % . , 0 ) o C «) - 20 % time [weeks] Figure 3.6. Change in adhesivity (fraction of cells adhered to quartz sand, ♦ ) , and transportability (fractional recovery of viable cells after passage through 30 cm of quartz sand, □) during starvation of Shewanella algae BrY. Error bars represent the standard error of the means. 60 A linear relationship was suggested after the average fractional recovery from 30 cm porous media columns was plotted against the average fraction of cells adhered to quartz sand in the batch adhesion studies (Figure 3.7). The regression analysis resulted in a correlation coefficient of R2 = 0.9421 and a slope of (fractional recovery) x (fraction adhered) ' = -0.2365. Thus, a decrease in the fraction of cells adhered to quartz sand in batch experiments resulted in an increase in fractional recovery during transport through 30 cm quartz sand columns. These results imply that it might be possible to use less time- and cost-intensive batch experiments to qualitatively predict the transport behavior of cells through porous media. However, to obtain quantitative results, the reproducibility of adhesion assays will have to be improved. The acquisition of sorption isotherms would potentially allow for better quantitative assessments. However, the acquisition of sorption isotherms for bacterial cultures is expected to be approximately as work intensive as column studies and therefore would have only little potential as a screening tool. In addition, batch adhesion experiments are not likely to emulate the potential influence of hydrodynamics and kinetics on bacterial adhesion in porous media columns. fractional recovery = -0.2365 x (fraction adhered in batch) + 0.2413 R2 = 0.9421 ™ 0.1 fraction adhered in batch Figure 3.7. Plot of fractional recovery from 30 cm quartz sand columns over the fraction of cells adhered in batch experiments. 61 Conclusions The research summarized in this chapter demonstrated that, according to the filtration theory, changes in cell size, buoyant density, and diffusion coefficient could partly explain the enhanced transport of starved cells. The remaining differences between starved and vegetative cells had to be accounted for by modifying the collision efficiency factor (a), which had to take on values of greater than one. Although values of a greater than one have.been reported in the literature, they are conceptually not appropriate to use in the filtration model. Thus, the applicability of the filtration theory for bacterial transport through porous media in its cunrently applied form has to be questioned and the development of a mathematical model capable of appropriately describing bacterial transport through porous media is necessary. Such a model should be capable of appropriately describing the observed scale- and physiology-dependent changes influencing the transport and attachment behavior of bacterial cells, which were described in Chapter 2 and this chapter. The observed changes in cell hydrophobicity and zeta potential were not sufficient to explain the differences in sticking efficiency between starved and vegetative cells. Research elucidating the change of other potentially transport related cell properties is ongoing. Currently scanning and transmission electron microscopy (SEM, TEM) are utilized to observe the fate of flagella during long term starvation of S. algae BrY. The SEM and TEM work is performed by Alice Dohnalkova at the Environmental and Molecular Science Laboratory (EMSL) located at Pacific Northwest National Laboratory (PNNL, Richland, WA). Caccayo et al. (1992) reported that vegetative cells of Shewanella algae BrY have a single polar flagellum. The presence of flagella could significantly increase the effective diameter of vegetative cells and therefore result in a greater removal of cells according to the filtration theory. Long term starvation might result in the loss of the flagellum and could therefore provide an explanation for the transport differences observed for starved and vegetative cells. Preliminary results utilizing a flagella deficient mutant (S. algae BrY-NE, which was kindly provided by Amitabha Das from PNNL), however indicate that the absence of the flagellum, though it potentially increases the transportability of vegetative cells, does not result in the same 62 increase in transportability as for starved cells (data not shown). Mueller (1996) discussed the influence of flagella and cell motility on the attachment of Pseudomonas fluorescens to the surface of microscopic flow cells. The higher adsorption rate coefficient of the flagellated, motile strain was attributed to superior transport properties. The non-motile strain however showed higher sticking efficiencies suggesting that the flagellum was not necessary for attachment but increased the specific adsorption rate coefficient. Das and Caccavo in contrast suggest that the flagellum acts as a motility-independent adhesin for Shewanella algae BrY (Das and Caccavo, unpublished results). Mueller (1996) however states that the hydrodynamic conditions can significantly influence the attachment of bacteria and it is likely that the hydrodynamic conditions in porous media are different from microscopic flow cells. Thus, the change of other parameters might be responsible for the improved transport behavior of starved cells. A more thorough characterization of the cell surface of S. algae BrY during starvation might provide additional information about the change of transport related cell surface properties. Information about protein and carbohydrate patterns on the cell surface could potentially be obtained using X-ray Photoelectron Spectroscopy (XPS) and Time of Flight Secondary Ion Mass Spectroscopy (ToF SIMS) as described by Caccavo et al. (1997). In addition, an alternative method should be applied to verify the hydrophobicity results obtained with the BATH assay (e.g’ hydrophobic interaction chromatography). In view of potential field applications of the starvation transport enhancement strategy, the starvation time should be optimized to produce cells with optimal transport properties in the least amount of time. The fractional recoveries observed for S. algae BrY (Figure 3.6) suggest that similar transport efficiencies can be achieved after four and seven weeks of starvation. Thus, four weeks of starvation might be sufficient to achieve a significant improvement in bacterial transport. Seven weeks of starvation time were initially chosen because previous work by Caccavo et al. (1996) indicated an almost constant cell size after seven weeks of starvation. A thorough observation of the results by Caccavo et al. (1996) indicate that after five weeks of starvation the iron reductase activity reached levels that were below the detection limit of the assay utilized. One can therefore speculate that, if the iron reductase activity is an indicator for the overall metabolic 63 activity and if metabolic activity is necessary for irreversible attachment, the decrease in enzyme activity might also be responsible for the decreased attachment of starved S. algae BrY cells to the quartz sand in porous media columns. Additional research is necessary to test this hypothesis. 64 CHAPTER4 BIOGEOCHEMICAL ELIMINATION OF CHROMILM(VI) FROM CONTAMINATED WATER Abstract Ferrous iron [Fe(II)] reductively transforms heavy metals in contaminated groundwater and the bacterial reduction of indigenous ferric iron [Fe(III)] to Fe(II) has been proposed as a means of establishing redox reactive barriers in the subsurface. The reduction of Fe(III) to Fe(II) can be accomplished by stimulation of indigenous dissimilatory metal-reducing bacteria (DMRB) or injection of DMRB into the subsurface. The microbially produced Fe(II) can chemically react with contaminants such as Cr(VI) to form insoluble Cr(IH)-precipitates. In batch and column experiments, the DMRB Shewanella algae BrY reduced surface-associated ferric iron to ferrous iron, which chemically reduced highly soluble Cr(VI) to insoluble Cr(HI). Once the chemical Cr(VI)-reduction capacity of the Fe(H)ZFe(HI) couple in the experimental systems was exhausted, the addition of S. algae BrY allowed for the repeated reduction of Fe(III) to Fe(H), which again reduced Cr(VI) to Cr(III). The research presented herein indicates that a biological process using DMRB allows the establishment of a biogeochemical cycle that facilitates chromium precipitation. Such a system could provide a means for establishing and maintaining remedial redox reactive zones in Fe(III)-bearing subsurface environments. Introduction Chromium(VI) compounds have been used extensively in the manufacture of alloys, the electroplating industry, the manufacture of dyes and pigments, the preservation of wood, in the leather 65 tanning industry, and as corrosion inhibitors in conventional and nuclear power plants (Hayes, 1997; Langard, 1980). Due to leakage, poor storage, and improper disposal practices, Cr(VI) has become one of the most frequently detected contaminants at hazardous waste sites (National Research Council, 1994; Watts, 1997). The spills are often extensive and the frequent presence of co-contaminants such as chlorinated aliphatics, radionuclides, metals, and organic solvents makes Cr(VI) remediation an environmental challenge (James, 1994; Watts, 1997; Pagilla and Canter, 1999). Cr(VI) is highly soluble and therefore easily transported with the ground, water (Bartlett and James, 1988). It is toxic and mutagenic to humans and other organisms (Committee on Biologic Effects of Atmospheric Pollutants, 1974; Gibb, 1999; Keith and Telliard, 1979; Watts, 1997). In its reduced trivalent state [Cr(III)], chromium forms insoluble hydroxides (Hug and Laubscher, 1997; Richard and Bourg, 1991) and remains immobile under most environmentally-relevant pH and redox conditions (Blowes et al., 1997a; James, 1996; Lantz, 1992; Pohland et al., 1993; Richard and Bourg, 1991). Cr(III) is less toxic than Cr(VI) and is neither readily taken up by organisms or plants nor transferred through the food chain (Hayes, 1997; National Research Council, 1994). Traditional remediation technologies such as excavation with subsequent treatment or Pump and Treat have often proven to be expensive and can create a long term liability for the responsible party (Mackay and Cherry, 1989; Seaman et al., 1999). Technologies to treat large volumes of soils and groundwater inexpensively are lacking. Passive treatment systems, such as subsurface barriers, offer an alternative to traditional cleanup technologies. Subsurface barriers minimize the need for expensive operating equipment and minimize human exposure to contaminants. Treatment systems relying on geochemical processes have attracted increased attention in recent years. Permeable reactive subsurface barriers made from iron metal have been successfully used for the elimination of a variety of contaminants including chlorinated solvents (O'Hannesin and Gillham, 1998), heavy metals (Cantrell et al., 1995; Gould, 1982), and radionuclides (Fiedor et al., 1998). The iron metal is usually introduced into the aquifer using trench and fill methods. However, the installation of trenches can become uneconomical at depths of more than approximately 12 m (Scott et al., 1998; Fortner, 1995). Thus, 66 alternative technologies allowing for the establishment of redox reactive zones in the subsurface have been proposed and implemented in the field. These technologies involve the injection of small Fe(O) particles under high pressure or after hydraulic or pneumatic fracturing of the aquifer (Landis and Vidumsky, 1999; Cantrell et al., 1997b). In situ redox manipulation has also been suggested for sites where trench and fill methods are impractical (Seaman et al., 1999). Cr(VI) can be reduced by ferrous iron, sulfides, and natural organic compounds (Wittbrodt and Palmer, 1996) and the reduction of Cr(VI) in several field studies has been attributed to the presence of ferrous iron (Anderson et al., 1994; Henderson, 1994; Pagilla and Canter, 1999; Patterson et al., 1997). Industrially, the addition of ferrous iron as a reductant represents a common treatment scheme for Cr(VI) contaminated water (Eary and Rai, 1988) and the use of ferrous iron in subsurface barriers for the remediation of chromium has been proposed (Blowes et al., 1997a; Powell et al., 1995b). During a field demonstration, at the Department of Energy (DOE) site in Hanford (WA), an inorganic reductant (sodium dithionite, NazSzO^ was injected into the aquifer and allowed to reduce indigenous Fe(III)-minerals to Fe(II) (Scott et al., 1998). The dithionite-containing carbonate-bicarbonate buffer (pH 11) was injected into the subsurface and reduced between 60% and 100% of the Fe(III) present in the Hanford sediments. The resulting reductive capacity of the reduced aquifer material was calculated to be sufficient for the treatment of 51 to 85 pore volumes (or 7 to 12 years) of groundwater at I ppm Cr(VI) and 9 ppm dissolved oxygen (Scott et al., 1998). Dissimilatory Metal-Reducing Bacteria (DMRB) have been found to use a wide range of iron minerals as electron acceptors at ambient groundwater pH (Caccavo, Jr. et al., 1992; Fredrickson and Gorby, 1996a; Roden and Zachara, 1996). DMRB are widely distributed in pristine and contaminated terrestrial, aquatic, and subsurface environments (Lovley, 1995b). They gain energy for growth by coupling the oxidation of organics or H2 to the dissimilatory reduction O fF e (III) minerals and other metals. Gorby et al. (1994) and Amonette et al. (2000) demonstrated that DMRB can reduce Fe(III) oxides, which can in turn reduce carbon tetrachloride (CT) to chloroform (CF) and Cr(VI) to Cr(III) according to the following reactions. 67 2 Fe2+ + CCl4 + H+ -> 2 Fe3+ + CHCl3 + CF (Eq 4.1) 3 Fe2+ + CrO42" + 5 H+ -> 3 Fe3++ Cr(OH)3J / + H2O (Eq 4.2) Depending on the Fe(H)ZCr(VI) ratio and reaction time, reaction (2) will result in the precipitation of pure chromium-hydroxides or mixed iron-chromium-hydroxides with iron and chromium in different ratios (i.e. FexCr3.x(OH)3, x=0,l,2,3) (Cary et al., 1977; Powell et al., 1995b)._____________________________________________________ If DMKB can repeatedly generate surface reactive ferrous iron from indigenous iron minerals in the subsurface, they could potentially be used to establish in situ bioreactors or permeable reactive barriers. A viable remediation strategy based on this approach must provide continuous reduction and precipitation of chromium from contaminated groundwater. The continuous reduction of Cr(VI) relies upon the renewal of Fe(II) through microbial reduction of the Fe(IH) present. To the author’s knowledge the ability of DMRB to repeatedly generate surface reactive Fe(II) for the chemical reduction of Cr(VI) to Cr(HT) in the presence of iron-chromium precipitates has not yet been demonstrated. Laboratory-scale model systems were used to evaluate the potential for such a process. Materials and Methods Experimental . Test Organism and Culture Methods. Shewanella algae BrY (Caccavo, Jr. et al,, 1992) was used as the model DMRB in these studies. S. algae BrY was maintained on tryptic soy agar (40 g L"1, Difco Laboratories) at ambient temperature and grown to the late log, early stationary phase in tryptic soy broth (30 g L"1, Difco Laboratories) at 150 rpm for 15 hours at ambient temperature. Cultures were centrifuged at 5,860 x g for 20 minutes at 4 0C, washed twice in phosphate buffered saline solution (PBS) to remove nutrients, and resuspended in the same buffer (Caccavo, Jr. et al., 1996). S. algae BrY cells immediately after this treatment were defined as vegetative cells and were used after resuspension and dilution in 68 oxygen-free minimal medium. Starved cells were obtained by aseptically stirring cell suspensions on magnetic stir plates at room temperature. Cells were harvested after 7 weeks. Centrifugation and resuspension in oxygen-free minimal medium served to remove cell debris from the starvation cultures (Caccavo, Jr. et al., 1996). Starved cells were used in parts of this study because it was previously demonstrated that starved cells are transported more effectively through porous media than vegetative cells and would therefore have an advantage in subsurface bioaugmentation with DMRB (Chapter 2; Bouwer et al., 2000). Amorphous Iron Coating. Artificially iron-coated sand was obtained by soaking 200 g of quartz sand (40 mesh, Unimin Corp., Emmet, ID) for several days in twice its volume of I N HNO3, subsequent washing with distilled water until the pH of the rinsewater exceeded 5.0, and mixing with 400 mL of a suspension of amorphous iron(III)oxyhydroxide (Lovley and Phillips, 1986). This suspension was prepared by dissolving 13.8 g FeCl3 in 400 mL of deionized water and slowly adjusting the pH to 7.0 using NaOH. A layer of clear liquid formed after the mixture of sand and iron(IH)oxyhydroxide was briefly shaken and allowed to settle. The clear liquid was decanted after approximately 12 hours of settling time and replaced with an equal volume of 0.1 mM NaCl at pH 7.0. The entire overlying suspension was decanted after an additional 12 hours, leaving only the sand. The above procedure was repeated four additional times over the course of 96 hours. The coated sand was washed to remove loosely attached iron, allowed to air-dry, and passed through a 2 mm sieve to break up larger aggregates. The coating procedure yielded a coating of amorphous iron(III)oxyhydroxide of about 1.3 + 0.12 mg Fe(HI) per gram of sand. No Fe(II) was detected on the sand using the analytical method described below. Batch Experiments. Iron Reduction. Oxygen free minimal medium containing NaHCO3, 2.5; NH4Cl, 1.5; KH2PO4, 0.6; KC1, 0.1 (in grams per L of deionized water); vitamins 10 mL per L, and trace minerals 10 mL per L (Balch et al., 1979; Lovley and Phillips, 1988) amended with 10 mM lactate was prepared under an N2=CO2 (80:20) gas atmosphere (Balch and Wolfe, 1976). Oxygen free minimal medium (9 mL) and I mL of bacterial culture, either starved or vegetative, were added to test tubes containing I g 69 of iron-coated sand and the vials were crimp sealed using thick butyl rubber stoppers (Bellco Glass, Vineland, NJ). All manipulations were performed under an oxygen-free N2/CO2 purge. The resulting bacterial concentrations were approximately 1.5 x IO7 colony forming units (CPU) per mL. Control vials were established lacking either bacteria or lactate and all vials were incubated statically in the dark. Triplicate sets of vials were sacrificed periodically to determine the amount of ferrous iron in the supernatant and on the sand separately. Batch Experiments. Chromium Reduction. A subset of vials containing biologically generated Fe(II) was autoclaved to eliminate microbial activity and thus to distinguish between chemical and biological Cr(VI)-reduction. Cr(VI) (0.15 mg) was injected from a K2CrO4 stock solution to a final concentration of 15 mg L"1. Cr(VI) concentrations in the aqueous phase were monitored over time as described below. Vials were sacrificed at the end of each experiment and analyzed for Cr(VI) and total Cr in the aqueous and solid phase separately. Controls containing oxidized iron oxyhydroxide coated sand were established to account for abiotic losses of Cr(VI). Experiments evaluating the ability of S. algae BrY to reduce Cr(VI) directly were also performed. Cr(VI) was injected into vials containing vegetative cells of S. algae BrY cells in oxygen free minimal medium supplemented with 10 mM lactate and Cr(VI) concentrations were monitored oyer time. Regeneration of Ferrous Iron in the Presence of Chromium Precipitates and Repeated Chromium Reduction. A set of vials containing microbially reduced iron was sterilized by autoclaving and then repeatedly injected with a Cr(VI) 'solution until their reduction capacity for Cr(VI) was exhausted. The capacity of these vials to reduce Cr(VI) was regarded as exhausted if 48 hours after the last injection of 0.05 mg Cr(VI), more than 0.05 mg/vial (=5 mg L"1) Cr(VI) were still detected. Vials were sacrificed for analysis of Fe(II), total Fe, Cr(VI), and total Cr after exhaustion of the Cr(VI) reduction capacity. A subset of the remaining vials was reinoculated with S. algae BrY to evaluate the potential for regeneration of reduced iron through the addition of DMRB. Reinoculated vials were sacrificed over time for iron analysis. 70 Vials that showed evidence of Fe(II) production were autoclaved and their ability to repeatedly reduce Cr(VI) was tested as described above. Column Experiments. Porous media columns were established to evaluate the microbially mediated geochemical elimination of Cr(VI) from contaminated waters in flow through systems (Figure 4.1). Polytetrafluoroethene (PTFE) columns, 5 cm in length, fitted with stainless steel nuts and ferrules, and PTFE and Tygon® tubing (Cole Palmer) were established. Each column was filled with seven grams of iron-coated sand. A PTFE grid was placed at the inlet and outlet o f each column to prevent washout of sand into the outlet or settling of sand into the inlet tubing. The columns were operated upflow with a peristaltic pump at flowrates of 4 mL min"1and up to 15 columns were operated simultaneously via a manifold. Each column was inoculated with approximately 6 x IO8 CFU of starved S. algae BrY and the columns were operated in semi-batch mode. Approximately 20 mL of oxygen-free minimal medium supplemented with 10 mM lactate was pumped through each column every two days. Columns were sacrificed over time, and the sand was analyzed for Fe(II) and total Fe. After evidence OfFe(II) production, a Cr(VI) solution was injected concurrently with a tracer (fluorescein) at a flowrate of 1.16 mL min"1, resulting in a residence time of approximately 6 min in the columns. Inlet and outlet concentrations of Cr(VI) and the tracer concentrations were monitored over time. The columns were sacrificed at the conclusion of the experiment and analyzed for both total Cr and Cr(VI). A subset of columns was reinoculated with starved S. algae BrY cells after depletion of the Cr(VI) reduction capacity and the ability of the columns to reduce Cr(VI) was assessed as described above after allowing adequate time for iron reduction. 71 r*— : . =4 T — I Sam pling ports Fe-coated sand O2Tree N2ZCO2 (80/20) gas Z Stainless steel, PTFE column assembly 0.2 pm filters Injection ports ------------ O — Tygoif tubing Valve Manifold —Oxygen-free minimal medium + 10 mM lactate Figure 4.1. PTFE, stainless steel column assembly for studying microbial Fe(III)-reduction and Cr(VI) reduction in flow through systems. Each column contained 7 g of iron coated sand and up to 15 columns were operated simultaneously. A reduced copper catalyst at 400 °C served to remove residual oxygen from the nitrogen/carbon dioxide gas mixture. Analytical Iron Quantification. Amorphous iron-coated sand was analyzed for both total iron and ferrous iron [Fe(II)]. The method described by Lovley and Phillips (Lovley and Phillips, 1987; Lovley and Phillips, 1988) was modified as follows. Instead of 0.5 M HC1, 2.5 M HCl was used, which proved to be sufficient for the solubilization of all Fe from the sand after incubation on a horizontal shaker for 30 minutes at 150 rpm. Fe(II) was determined by adding 5 mL of 2.5 N hydrochloric acid to I g of iron coated sand. Total iron was quantified by adding 5 mL of 0.25 N hydroxylamine solution in 2.25 N hydrochloric acid to I g of iron coated sand. Aliquots of the acid extracts were added to 5 mL of a solution of I g ferrozine (Sigma, St. 72 Louis, MO) in I L 750 mM HEPES buffer at pH 7 and the absorbance was determined spectrophotometrically at 562 nm. Chromium Quantification. Aqueous Cr(VI) was quantified using diphenylcarbazide (Urone, 1955). A 500 (jL sample was added to 10 mL 0.2 N sulfuric acid and vortexed. A diphenylcarbazide (DPC) solution (0.5 mL, 0.25% of DPC in 100% acetone) was added and the absorbance was read at 540 nm. Cr(VI) was extracted from the sand using a modified procedure described by Vitale et al. (1994). A 0.28 M Na2CO3ZO-S M NaOH solution (5 mL) was added to I g of sand, mixed, and heated to 90-95 °C for 60 minutes. The vials were swirled repeatedly during the heating process and deionized water was added, if needed, to replace losses due to evaporation. The vials were allowed to cool to room temperature and the total volume was filtered through a 0.2 pm nylon membrane (Fisher Scientific, Pittsburgh, PA). The pH was adjusted to 7.5 using concentrated HNO3 and 0.5 mL of the DPC solution was added. The pH was adjusted to I with concentrated H2SO4 and the total volume of the sample was adjusted to 10.5 mL with 0.2 N H2SO4 before the absorbance was read at 540 nm. Aqueous phase total Cr was determined after filtration using a modified standard method (Anonymous 1992). One drop of methyl orange indicator solution was added to I mL of sample. Concentrated NH4OH was added until the solution turned yellow, diluted H2SO4 (1:2 from concentrated H2SO4) was added until the indicator turned red, and the total volume was adjusted to approximately 5 mL. Boiling chips were added, the vial was closed with a rubber stopper, and a 20 gauge needle was inserted as exhaust; the solution was heated to a boil. One drop of a KMnO4 solution (40 g L"1) was added and boiling was continued. If fading occurred, another drop of KMnO4 was added and boiling was continued for at least 2 minutes after addition of the last drop of KMnO4. Sodium azide (NaN3, 5 g L"1) was added in 0.1 mL increments and boiling continued until the color disappeared. The vials were allowed to cool after addition of I drop of concentrated H3PO4. Concentrated H2SO4 was used to adjust the pH to 1.0 and the whole content of each vial was filtered through a 0.2 pm nylon filter. The volume was adjusted to 10 mL using 0.2 N H2SO4 and 0.5 mL of DPC solution were added before measuring the absorbance at 540 nm. 73 Total Cr from the sand was determined analogously substituting the I mL liquid sample with I g of Fe-coated sand. Results and Discussion Microbial FetIID Reduction S. algae BrY can use Fe(III) as electron acceptor to grow in the absence of oxygen with lactate as the electron donor. Starved and vegetative cells of S. algae BrY reduced surface associated ferric iron [Fe(III)] to ferrous iron [Fe(II)] over time (Figure 4.2). The reduction was faster in vials inoculated with vegetative bacteria than in those inoculated with starved bacteria. Vegetative cells reduced approximately 90% (1.19 ± 0.09 mg) of the ferric iron available on the sand (1.3 ± 0.12 mg) within 24 hours. Starved cells exhibited a slower rate of Fe(III) reduction and reduced approximately 80% (1.04 ± 0.02 mg) of the ferric iron within 10 days. Control vials lacking either lactate or S. algae BrY cells showed no evidence of Fe(II) production, indicating that the observed reduction was due to electron donor-dependent microbial activity. It has been shown that starved cells of S. algae BrY reduce soluble ferric pyrophosphate slower than vegetative cells (Caccavo, Jr. et a l, 1996). The data presented here are consistent with these results. Although starved cells of S. algae BrY reduce surface-associated Fe(III) at a slower rate than vegetative cells, the amount of Fe(II) produced by starved cells after 10 days is only slightly lower than the amount produced by vegetative cells. Most of the Fe(II) produced (> 95%) remained associated with the sand surfaces as observed by separate analyses of the supernatant and the sand (data not shown). The fact that most of the reduced iron remains surface-associated is important in view of a possible field application since surface-associated Fe(II) is more reactive with dissolved contaminants than dissolved Fe(II) (Kriegman-King and Reinhard, 1994). Furthermore, in aquifers dissolved Fe(II) would be transported out of the system over time with the flowing groundwater minimizing the efficacy of this technology. 74 Tim e [hours] Figure 4.2. Production of Fe(II) from amorphous Fe(III) oxyhydroxide-coated sand by starved ( • ) and vegetative cells (■) of S. algae BrY. Every system contained I g of sand coated with 1.3 ± 0.12 mg of Fe(III) initially. Controls lacking either lactate (A) or S. algae BrY (D) did not produce significant amounts of Fe(II). Error bars represent the standard error of the means (n=3). The injection of DMRB into the subsurface and the stimulation of DMRB in the subsurface to enhance contaminant transformation has been proposed (Caccavo, Jr. et al., 1996, Gorby et al., 1994). Here it was demonstrated that starved S. algae BrY cells, which can be transported through porous media better than vegetative cells (Chapter 2; Bouwer et al., 2000), can reduce surface-associated Fe(III) and produce surface-associated, reactive Fe(II). Thus zones of surface reactive Fe(II) could be established in contaminated aquifers by stimulating naturally-occurring DMRB or by bioaugmenting the contaminated subsurface with starved DMRB. 75 CrfVrD Reduction by Microbiallv Produced FeCID The reduction of Cr(VI) by Fe(II) has been documented using soluble Fe(II) (Seaman et ah, 1999) and surface associated Fe(II) in natural aquifer material (Caccavo, Jr. et ah, 1996; Gorby et ah, 1994; Kriegman-KingandReinhard, 1994). Cr(VI) concentrations decreased immediately after injection in the experimental vials containing biologically produced, surface-associated Fe(II) (Figure 4.3). The reaction of Fe(II) with Cr(VI) occurred so fast in these vials that at the first sampling event after 30 seconds, the concentration of Cr(VI) was already significantly lower than the control concentration of 15 mg L"1. Vials that were sterilized immediately before the injection to eliminate microbial activity showed the same initial decrease, providing evidence that chemical processes are governing the reduction of Cr(VI) by Fe(II) initially. Control vials containing oxidized iron oxyhydroxide-coated sand did not show a significant decrease in Cr(VI) over the duration of the experiment, indicating that sorption of Cr(VI) to the iron-coated sand was negligible. The initial increase in Cr(VI) concentrations in the control vials (see Figure 4.3) can be attributed to incomplete mixing before the first sampling event. Since the reaction of Cr(VI) with Fe(II) was expected to occur rapidly, the systems were sampled as soon as possible after addition of Cr(VI), i.e. after 30 seconds. Thus, there was only a limited time available for mixing and the resulting incomplete mixing might account for initially lower Cr(VI) concentrations. Based on the stoichiometry of reaction (2), the vials contained sufficient Fe(II) to completely reduce the Cr(VI) added (0.15 mg, resulting in an initial concentration of 15 mg L'1). However, 11 minutes after the injection of Cr(Vl), the reduction of 15 mg L'1 Cr(VI) remained incomplete. This incomplete conversion of Cr(VI) could be due to mass transport limitations within the experimental systems or the presence OfFe(II) sites which were not reactive with Cr(VI). Powell et al. (1995) stated that the reactivity of surface-associated Fe(II) can vary depending on the mineralogy. Although no mineralogical analyses were performed during this research, the white color of the precipitate during biological iron reduction and the high carbonate content of the buffer suggest that siderite (FeCO3) formed at least partly on the sand grains. Siderite has been described as only moderately reactive with dissolved phase contaminants relative 76 to other iron minerals (Agrawal and Tratnyek, 1996; Blowes et al., 1997a). This may explain the fraction of moderately and non reactive Fe(II) in the batch vials. Diffusion limitations (Agrawal and Tratnyek, 1996) and the accumulation of chromium precipitates can also limit Cr(VI) reduction by surface reactive Fe(II). h 150- 2 4 hours Tim e [min] Figure 4.3. Cr(VI) transformation in vials containing (■) microbially reduced Fe-coated sand and active S. algae BrY, (# ) microbially reduced Fe-coated sand (autoclaved to eliminate microbial activity), and (A) oxidized Fe-coated sand. The reaction of Fe(II) with Cr(VI) occurred so fast that at the first sampling event, the concentration of Cr(VI) was already significantly lower in vials containing Fe(II) than in control vials. The target concentration of 15 mg L'1 is indicated by (O). Error bars represent the standard error of the means (n = 3). The vials were analyzed for total chromium and Cr(VI) at the conclusion of each experiment. The majority (> 95%) of the chromium recovered from the sand was recovered as Cr(III) (data not shown) demonstrating that Cr(VI) was transformed to Cr(III) in the presence of biologically produced surface associated Fe(II). 77 Complete conversion of Cr(VI) within 24 hours was only observed in the vials containing active S. algae BrY cells. This decrease suggests either direct reduction of Cr(VI) by S. algae BrY or an indirect mechanism in which S. algae BrY reproduced ferric iron resulting from reaction (2) to ferrous iron, which then again reduced Cr(VI) to Cr(III). Direct microbial reduction of Cr(Vl) by metal-reducing bacteria and soil indigenous microbial populations has been reported (Bader et al., 1999; Bopp and Ehrlich, 1988; Turick et al., 1998). It was demonstrated here that direct microbial reduction of Cr(VI) by active S. algae BrY is possible, although significantly slower than the chemical reduction with surface reactive Fe(II). Regardless of whether surfaces were available (quartz sand or iron-coated sand) or not, active S. algae BrY cells did not significantly reduce 5 mg L"1 Cr(VI) within 11 minutes (Figure 4.4). However, the concentration of Cr(VI) dropped to below I mg L"1 in all vials containing S. algae BrY after 24 hours, indicating slow direct microbial reduction of Cr(VI) by S. algae BrY. As mentioned previously, at the first sampling event after 30 seconds, the target concentration of 5 mg L"1 Cr(VI) was not observed consistently in the vials, presumably due to incomplete mixing. The concentrations stabilized to approximately 5 mg L"1 Cr(VI) after 1-2 minutes into the experiment and repeated mixing. Compared to the systems containing microbially produced Fe(II) (Figure 4.3), the direct microbial reduction of Cr(VI) was much slower. These observations concur with Caccavo et al. (1996), Eary and Ray (1988), and Fendorf and Li (1996) who reported that the reduction of Cr(VI) by dissolved and surface associated ferrous iron is rapid compared to the direct microbial reduction. Reactions with surface associated Fe(II) usually occur within minutes to hours at environmentally relevant pH values and temperatures. These reactions can require stoichiometric or excess amounts of ferrous iron (Davis and Olsen, 1995; Eary and Rai, 1988; Seaman et al., 1999). Fe(II) has been found to reduce Cr(VI) at pH values of less than 10 (Palmer and Wittbrodt, 1991), and the rate of reduction increases with decreasing pH (Powell et al., 1995b; Eary and Rai, 1991). Because the pH of ground water rarely exceeds 10, the reaction should be feasible in field situations and has been proposed for the treatment of industrial waste water and groundwater containing Cr(VI) (Eary and Rai, 1988; Seaman et al., 1999). 78 I O 24 hours Time [min] Figure 4.4. Direct Cr(VI) transformation by active S. algae BrY (▲) in the absence of sand, (D) in the presence of quartz sand, and ( • ) in the presence of iron-coated sand. The time between injection of bacteria and Cr(VI) (minutes) was not sufficient for significant Fe(II) production in the vials. The target concentration of 5 mg L"1 is indicated by (O). Error bars represent the standard error of the means (n = 3). Regeneration of Surface Reactive Ferrous Iron in the Presence of Chromium Precipitates These experiments were designed to determine whether surface reactive Fe(II) can be regenerated in the presence of Fe(III)Cr(III)precipitates formed during the reduction of Cr(VT) by Fe(II). Surface reactive Fe(II) was produced initially by inoculating vials containing amorphous iron(oxy)hydroxide coated sand with starved S. algae BrY. Figure 4.5 shows the total mass of Fe(II) detected in the vials over the duration of the experiment. More than 90% (1.19 ± 0.10 mg) of the ferric iron available in the vials (1.3 ± 0.12 mg) was reduced to Fe(II) in the absence of Cr(VI), 20 days after the inoculation with starved S. algae BrY. The Fe(II) containing vials were autoclaved after 20 days to eliminate microbial activity and repeatedly injected with a Cr(VI) solution until their reduction capacity for Cr(VI) was exhausted. Only 79 approximately 40% (0.15 mg Cr(VI) per vial) of the theoretically reducible Cr(VI) was reduced over a period of twelve days (day 20 to day 32). Destmctive analysis of triplicate vials revealed that the majority of the chromium detected on the sand (> 95%) was present as Cr(III) suggesting that Cr(VI) was reduced to Cr(IH) which precipitated (data not shown). The incomplete reduction of Cr(VI) indicates the presence of slowly or non reactive fractions of Fe(II). ' However, the conversion of Cr(VI) to Cr(III) could only account for approximately 80% of the decrease in ferrous iron concentration (from 1.19 ± 0.10 mg to 0.78 + 0.10 mg) indicating that ferrous iron was possibly oxidized by compounds other than Cr(VI). Despite the fact that syringes were flushed with oxygen-free gas before sampling, trace amounts of oxygen could have been injected into the vials and lead to the oxidation of Fe(II) to Fe(IH). After the Cr(VI) reduction capacity of the vials was exhausted (day 32), a subset of the remaining vials was reinoculated with S. algae BrY to evaluate the possible regeneration of surface reactive ferrous iron through the addition of DMRB. Fe(H) increased from 0.78 ± 0.10 mg per vial to 1.10 ± 0.07 mg per vial in reinoculated vials only between day 32 and day 55 (Figure 4.5). Vials, which were not reinoculated, did not show an increase but rather a slight decrease to 0.73 ±0.15 mg Fe(II), which might be due to slow reactions of Cr(VI) with Fe(II) in these vials. Cr(VI) was still detectable in these vials at day 55 while no Cr(VI) was detected in reinoculated vials. Although the amount of total Fe(H) in the reinoculated vials (1.10 ± 0.07 mg) was lower than the amount of Fe(II) produced during the initial incubation in the absence of Cr-precipitates (1.19 ± 0.10 mg) it is clear that S. algae BrY was able to re-reduce a significant amount of the Fe(IH) which was generated during the prior injection of Cr(VI). 80 Tim e [days] Figure 4.5. Mass of Fe(II) over time in vials containing iron oxyhydroxide-coated sand as determined by separate analysis of aqueous and solid phase in the vials. ( • ) after inoculation with active S. algae BrY (day 0 - 20), (A) after autoclaving and repeated chromium injection, (■) after autoclaving, repeated chromium injection, and reinoculation with S. algae BrY (day 32 - 55). Error bars represent the standard error of the means (n = 3). Reinoculated vials were examined for their ability to reduce Cr(Vl) a second time after production of Fe(II). Cr(VI) was injected to a concentration of 15 mg L ' and Cr(VI) concentrations were monitored over time. Cr(VI) concentrations in reinoculated vials decreased immediately after injection (Figure 4.6) and only approximately I mg L"1 Cr(VI) was detected after 30 seconds. The rate of Cr(Vl) reduction in these vials was even faster than during the first cycle (compare to Figure 4.3) where approximately 5 mg L 1 remained after 11 minutes. No decrease in Cr(VI) was observed in vials which were not reinoculated. The final concentration of approximately 18 mg L"1in these vials is the result of residual Cr(VI) in the vials before the injection of Cr(VI) from the stock solution. These results suggest that S. algae BrY can regenerate the reduction capacity of iron minerals by producing surface reactive Fe(II) from Fe(III) even in the presence of Cr(III)-precipitates. It is likely that 81 the rate and extent of Fe(II) production in natural systems, before and after precipitation of Cr(III), will depend on the type of iron mineral present, the concentration and ability of DMRB to reduce the Fe-Crmineral phases after precipitation of Cr, and on the time allowed for Fe-reduction. Tim e [min] Figure 4.6. Cr(VI) transformation in vials with initially exhausted Cr(VI)-reduction capacity (S) after reinoculation and incubation with S. algae BrY, (■) without reinoculation. The Cr(VI) spike increased the initial Cr(VI) concentration in the vials by 15 mg L"1 (indicated by O). Error bars represent the standard error of the means (n = 3). Iron and Chromium Reduction in Columns In order to develop the basis for a viable restoration technology, Cr(VI) must be reductively precipitated by surface reactive Fe(II) in flow through systems at short contact times. Column studies were designed to more closely approximate field situations, in which the contact time of Cr(VI) with biologically produced surface-associated Fe(II) is limited. Columns inoculated with starved S. algae BrY were fed in semi-batch mode and were sacrificed periodically to analyze for Fe(II). 82 Fe(II) was produced over time (data not shown), albeit at a significantly slower rate than in the batch experiments described above. Starved S. algae BrY reduced more than 95% of the Fe(III) present in the columns after 2 months as compared to 80% after 7 to 10 days in batch experiments. The production of ferrous iron in the columns indicated that starved S. algae BrY colonized the Fe-coated sand and utilized the surface-associated Fe(III) as electron acceptor. After evidence of Fe(II) production, a Cr(VI) solution (6.6 ±0.1 mg L"1) was injected into the columns concurrently with a tracer (Fluorescein, Na-salt). Influent and effluent concentrations of the tracer and Cr(VI) were monitored over time. The data were expressed as fractions of their influent concentration and were plotted as a function of the number of pore volumes eluted from each column. Figure 4.7 shows the tracer breakthrough curves of three different columns with their corresponding model fits (Fetter 1993) and their corresponding Cr(VI) breakthrough data. It is evident that one of the columns had a higher dispersivity likely due to differences in packing of the sand. However, despite this difference in tracer breakthrough it is evident that the breakthrough of Cr(VI) in all three columns was clearly retarded as compared to fluorescein. Statistical examination of the breakthrough curves using a two sample xg-test (Gibbons, 1971, pp. 131-137) clearly indicates that in all three columns containing microbially produced Fe(II), the breakthrough of Cr(VI) was retarded (xg-test, P value = 0.05). The mass of Cr(VI) retarded in the columns was calculated on the basis of the breakthrough curves for each column, approximating complete breakthrough of Cr(VI) after four pore volumes- (see Figure 4.7). The mass of Cr(VI) in the effluent of the column (integral of the effluent concentration over time) was compared with the mass of Cr(VI) injected into the column (integral of the influent concentration over time). As in the batch experiments only a fraction of the stoichiometrically possible mass of Cr(VI) was retarded in each column, however, the fraction of chromium retarded in the columns was even smaller (I + 0.7% (n=3) of the total reduction capacity for Cr(VI)) than for the batch studies (approximately 40% of the total reduction capacity for Cr(VI)). 83 pore volumes eluted Figure 4.7. Normalized breakthrough curves of a fluorescein tracer (■, A, ♦ ) (corresponding model fits indicated by solid lines) and its corresponding Cr(VI) breakthrough data (D, A , O ) through columns containing biologically produced Fe(II). The influent Cr(VI) concentration was 6.6 ± 0.1 mg L"1 for every column. It is possible that the short contact time in the columns (approximately 6 minutes) is not sufficient for more extensive Cr(VI) reduction to occur. Mass transport limitations, the presence a slowly or non reactive Fe(II)-fraction (such as siderite), and the accumulation of chromium precipitates on surface reactive sites could be responsible for the slow and incomplete reaction. The sand was analyzed for total chromium and Cr(VI) by sacrificing triplicate columns after the injection of Cr(Vf). More than 95% of the chromium was recovered as Cr(III) (data not shown), indicating that the retardation of chromium was due to reductive precipitation within the columns. Three different processes are likely to be responsible for the reductive precipitation of chromium. Cr(Vf) may react chemically with surface reactive Fe(II) or active S. algae BrY cells may directly reduce Cr(VI) to precipitate Cr(III). These processes in addition to DMRB mediated biogeochemical cycling of Fe(III) back to Fe(II) (process 2 in Figure 4.8) may explain the attenuated increase of Cr(VI) towards the 84 end of the experiment. All three processes together with the limited reduction capacity of the columns would cause the slow but continued reductive precipitation of chromium within the columns and would result in Cr(VI) effluent concentrations that approach the influent concentration asymptotically. It is likely that all three processes occur within the columns and therefore it is surprising that of the total reduction capacity of surface associated Fe(II) for Cr(VI), only approximately 1% is utilized after approximately 4 pore volumes and that Cr(VI) concentrations in the effluent are almost equal to the influent concentration. QjMRtT) e Cr(VI) -> Cr(III) Fe(Il)sur ©F e(III) sur (^dmreT) Figure 4.8. Processes governing the fate of Cr(VI) in environments containing surface associated iron and dissimilatory metal-reducing bacteria (DMRB). (I) Direct Cr(VI)-reduction by DMRB; (2) chemical reduction of Cr(VI) by surface associated ferrous iron (Fe(H)sur) and biological re-reduction (regeneration) of surface associated ferric iron (Fe(III)sur) to Fe(II)sur by DMRB. The remaining columns were reinoculated after depletion of the Cr(VI) reduction capacity to evaluate the possibility of regenerating the reactivity of the columns by injection of S. algae BrY. Approximately two months after reinoculation, Fe(II) production in these columns was evident and Cr(VI) was injected a second time concurrently with fluorescein. The normalized breakthrough curves were very 85 similar to those plotted in Figure 4.7 (data not shown) indicating that the reduction capacity of the columns was regenerated by the reinoculation of S. algae BrY.. Conclusions The results presented here suggest the possibility of artificially establishing a biogeochemical cycle (Figure 4.8) in Cr(Vl) contaminated subsurface environments. DMKB reduce ferric iron [Fe(III)] to ferrous iron [Fe(II)], which in turn reduces Cr(VI) to Cr(III) while Fe(H) is oxidized (back) to Fe(HI). DMRB can then re-reduce the Fe(III) to Fe(H) to close the biogeochemical cycle. The use of DMRB potentially allows the establishment and maintenance of redox reactive zones (permeable subsurface barriers) in Fe(IH)-bearing subsurface -environments. The applicability of this technology to any given field site will depend on our ability to produce surface reactive Fe(II) from the iron minerals present in the aquifer. This can be achieved by stimulating indigenous microorganisms or, if microbial iron reduction cannot be sufficiently supported, by our ability to augment the existing populations with metal-reducing bacteria. The use of starved metal-reducing bacteria has potential for such an approach. Starved bacteria are transported better through porous media (Chapter 2; Bouwer et al, 2000) and, as shown here, can produce surface reactive Fe(II). However, like in most biofemediation strategies, there are limitations and concerns that have to be considered. High initial Cr(VI) concentrations or the presence of other toxic compounds might slow down or prevent biological iron and chromium reduction. High or low pH values as well as a lack of electron donor can also hinder biological iron reduction and, subsequently, chromium reduction. The process described herein does not remove chromium from the aquifer but immobilizes it by precipitation. Although Fe(III)Cr(IH)hydroxides and pure Cr(HI)Iiydroxide have very low solubilities and are unlikely to result in chromium concentrations above the drinking water limit for chromium, the reoxidation of Cr(IH) compounds is a concern. Although microbially mediated reoxidation of Cr(IH) is not a proven concern in systems containing natural soil material (James, 1994, Bader et al., 1999), manganese oxides have been found to quickly oxidize Cr(IH) in soils (Rai and Zachara, 1988, Bartlett and James, 86 1979). Therefore, the application of oxidants such as'potassium permanganate or hydrogen peroxide, commonly used in certain remediation schemes, can potentially lead to a remobilization of immobilized chromium. Future research is needed to address the identification of mineral phases developing during the microbial reduction of iron (hydr)oxides and iron chromium (hydr)oxides. The stability of these precipitates will be assessed and the recycling, frequency of the ferric/ferrous iron redox couple in the presence of Cr(VI) will be evaluated in batch and flow through systems. 87 CHAPTER 5 . DISSIMILATORY IRON-REDUCING BACTERIA CAN INFLUENCE THE REDUCTION OF CARBON TETRACHLORIDE BY IRON METAL Reproduced with permission from Gerlach, R., Cunningham, A.B. & Caccavo, F. Jr. (2000): Dissimilatory Iron-Reducing Bacteria Can Influence the Reduction of Carbon Tetrachloride by Iron Metal. Environmental Science and Technology 34 (12): 2461-2464. Copyright 2000 American Chemical Society. Abstract Little is known about the long term performance of zero valent iron (Fe(O)) subsurface barriers. Groundwater exposure induces corrosion processes that can passivate the Fe(O) surface and decrease barrier reaction rates. This chapter presents evidence that dissimilatory iron-reducing bacteria (DIRB) can stimulate the rate of carbon tetrachloride (CT) transformation in the presence of corroded iron. The DERB, Shewanella algae BrY, adhered to the corroded Fe surfaces that showed little or no capacity to transform CT. The addition of BrY to these systems with decreased CT transformation rates resulted in increased ferrous iron concentrations and increased CT transformation to chloroform (CF). The results suggest that DIRB can have an influence on the long term performance of Fe(O) barriers. Introduction Permeable, reactive, subsurface barriers of zero valent iron (Fe(O)) have been successfully employed for the remediation of contaminated groundwater. Fe(O) is effective in degrading or transforming a wide range of contaminants including chlorinated organics (Eykholt and Davenport, 1998; Gatpagar et al., 1997; Gillham and O'Hannesin, 1994; Helland et al., 1995; Johnson et al., 1998; Johnson et al., 1996; Roberts et al., 1996; Sayles et al., 1997; Siantar et al., 1996), heavy metal oxyanions (Cantrell et al., 1995; Gould, 1982) and oxycations (Fiedor et al., 1998), nitroaromatics (Agrawal and Tratnyek, 1996; Cao et al., r 88 1999), and nitrate (Huang et al., 1998; Siantar et a l, 1996; Till et al., 1998). The corrosion of Fe(O) provides the electrons necessary for the reduction of these contaminants. Laboratory studies simulating long term performance of Fe(O) have indicated that contaminant degradation rates can either decrease (Devlin et al., 1998; Gatpagar et al., 1997) or more rarely increase (Helland et al., 1995) as a function of time. This temporal effect on contaminant degradation rates is usually attributed to the development of corrosion products that develop on the metal surface (Helland et al., 1995; Johnson and Tratnyek, 1994; Scherer et al., 1998). Johnson and Tratnyek (1994), Scherer et al. (1998), and Odziemkowski et al. (1998) have proposed that the accumulation of corrosion products on Fe(O) surfaces either inhibits contaminant access to the metal surface or forms new sites for contaminant adsorption, reaction, and catalysis. A decrease in reaction rates can be due to the accumulation of thick layers of amorphous corrosion products on Fe(O) surfaces, which may function as a physical barrier between dissolved contaminants and the underlying reactive sites. The accumulation of highly redox reactive corrosion products such as green rust, however, can also significantly contribute to the reduction of contaminants in Fe(O) systems (Erbs et al., 1999). A wide range of Fe(O) corrosion products has been found in laboratory and field studies, and corrosion product distribution appears to be dependent on the type of Fe(O) used, the contaminant present, geochemical conditions, and the time of exposure to contaminated groundwater. The corrosion products commonly found in the presence of contaminated groundwater include goethite (ct-FeOOH; Pratt et al., 1997), magnetite (Fe3O4; Bonin et al., 1998; Odziemkowski and Gillham, 1997; Odziemkowski et al.; 1998; Peterson et al., 1997), maghemite Cy-Fe2O3; Peterson et al., 1997), hematite (Ctr-Fe2O3; Pratt et al., 1997), mixed chromium-iron-hydroxides .(Blowes et al., 1997b; Powell et al., 1995a), green rust (mixed Fe(II)-Fe(III)-hydroxy salts; Odziemkowski and Gillham, 1997), and siderite (FeCO3; Mackenzie et al., 1997). Many of these iron minerals are used by dissimilatory iron-reducing bacteria (DIRB) as electron acceptors (Caccavo, Jr. et al., 1992; Fredrickson and Gorby, 1996a; Heijman et al., 1995; Heijman et al., 1993; Kostka and Nealson, 1995; Roden and Zachara, 1996). DIRB are widely distributed in both pristine 89 and contaminated terrestrial, aquatic, and subsurface environments (Lovley, 1995b) and are thus likely to contact Fe(O) banners in situ. DIRB gain energy for growth by coupling the oxidation of organic matter to the dissimilatory reduction OfFe(III) minerals. Gorby et al. (1994) demonstrated that DIRB can reduce structural Fe(HI) oxides, which can in turn reductively dechlorinate carbon tetrachloride (CT) to chloroform (CF). The bacterial reduction of Fe(III) to Fe(II) allowed for the following reaction to occur: 2 Fe2++ CCl4 + H+ 2 Fe3+ + CHCl3 + Cf (Eq. 5.1) DIRB have also been shown to influence the corrosion of mild steel by solubilizing the protective ferric iron film on the passivated metal surface (Obuekwe et a l.,1981a; Obuekwe et al., 1981b). A decrease in corrosion layer thickness due to solubilization of corrosion products in Fe(O) systems could also lead to increased reaction rates of iron metal with chlorinated aliphatics (Devlin et al., 1998; Gatpagar et al., 1997; Scherer et al., 1998). It was hypothesized that DIRB can influence the performance of Fe(O) subsurface barriers by adhering to corroded iron and reducing Fe(III) corrosion products to Fe(II)-compounds. The reduction of Fe(IH) to Fe(II) can lead to the formation of surface reactive Fe(II) species or result in the removal of passivating ferric precipitates on the Fe(O) surface. Either one of these processes will potentially increase the reductive transformation of carbon tetrachloride (CT). Materials and Methods Organism and Culture Conditions Shewanella algae BrY (Caccavo, Jr. et al., 1992) was used as a model DIRB in these studies. S. algae BrY was maintained on tryptic soy agar (40 g D'1, Difco Laboratories) at ambient temperature. S. algae BrY cultures were grown in tryptic soy broth (30 g L"1, Difco Laboratories) at 28 °C and 150 rpm for 15 hr. Cultures were centrifuged at 5,520 x g for 20 minutes at 4 °C, washed once in anaerobic 10 mM 90 HEPES (N-[2-hydroxyethyl]piperazme-N’-[2-ethanesulfonic acid]) buffer (pH 7.0) amended with 2.5 g liter1 NaCl and suspended in the same buffer. All buffers were made anaerobic by boiling and cooling under a constant stream of O2-Ifee N2 (Balch and Wolfe, 1976). All manipulations were performed under an O2-Ifee N2 purge. CT Reduction The ability of S. algae BrY to influence CT reduction by passivated Fe(O) was examined. Fe(O) powder (20 mg, 100 mesh, Fisher Scientific) was added to 10 ml of aerobic HEPES buffer (pH 7.0), sealed in culture tubes under an aerobic headspace and vigorously shaken on a wrist action shaker for approximately 40 hours. This method Was found to be efficient for pre-exhausting the CT-reducing capacity of the Fe(O) powder. Following pretreatment, the tubes were purged with O2-Ifee N2 for 15 minutes by lowering a stainless-steel canula into the bottom of each tube and letting the anaerobic gas percolate through the aqueous phase. Anaerobic sodium lactate (final concentration 10 mM) and either heat-killed or active cells of S. algae BrY (final concentration of IO10 cells ml"1) were added to the tubes using degassed syringes and needles. The tubes were then filled to their maximum capacity (27.5 ml) with anaerobic HEPES buffer to minimize loss of CT into the headspace. The gas-sparging canula was slowly removed from the tube as a PTFE-Iined silicon septum was inserted. Control tubes containing either heatkilled or active S. algae BrY cells without iron powder were also established. The experiments were initiated by injecting 0.5 ml of a CT (Sigma) stock solution into the tubes. The stock solution was made by adding I ml of pure phase CT to 40 ml of deionized water. Excess aqueous phase was allowed to escape from the tube by allowing it to equilibrate with a N2 filled syringe. Aqueous phase samples (0.1 ml) were taken with a gas-tight syringe through the PTFE lined septum immediately after CT addition and periodically over the duration of the experiment. At each time-point, the sampling-syringe was flushed with N2 and approximately the same volume O f N 2 was injected into each tube. The sample was injected into I ml of hexane and allowed to equilibrate for at least 10 minutes on a horizontal shaker (150 rpm). The hexane phase was then used for gas chromatographic analysis. The total headspace volume at the end of the 91 experiment was less than 5 % of the vessel volume, ensuring that less than 4 % of CT remaining and less than 0.5 % of CF produced in the tubes were in the headspace (using the Henry constants given by Gossett, 1987). All CT reduction treatments were conducted in triplicate. Iron Reduction The ability of S. algae BrY to use Fe(O) corrosion products as a terminal electron acceptor was examined in an experiment analogous to that described above, omitting the CT and replacing the PTFElined septa with blue butyl rubber stoppers. The initial cell concentration was 1.60 x IOld cells ml"1. Sampling was performed by removing aqueous phase-aliquots, with a disposable syringe and 22-gauge needle. Samples were injected directly into 0.5 N HCl for ferrous iron analysis. All iron reduction treatments were conducted in triplicate. Adhesion of BrY to Iron Adhesion studies were performed as described previously (Caccavo, Jr. et al., 1997) with the exception that the Fe(III) oxide-bicarbonate buffer was replaced with Fe(O)-HEPES buffer. Anaerobic tubes of HEPES-pretreated iron were prepared as described above. A final concentration of 4.2 x IO8 cells ml"1 was used and adhesion assays were performed in triplicate. Analytical Methods CT, CF, and dichloromethane (DCM) were analyzed using a HP 6890 gas chromatograph (GC) equipped with an electron capture detector. An aliquot (10 pi) of hexane phase of the liquid-liquid extracts was injected onto a 60/80 Carbopack B/l % SP-1000 column (8 ft x 1/8 in). The GC was operated under the following conditions: Injector temperature of 275 0C, initial temperature of 100 °C, increased at 15 0C min"1 up to 200 °C, detector temperature of 250 °C, and Na flow of 40 ml min"1. The GC was controlled by HP Chem Software, Vers. A.04.01. This method permitted for the detection of concentrations of less than 92 IOjlg r1of CT, CF, or DCM. Ferrous iron analysis was performed with Ferrozine as described by Lovley and Phillips (1987). Results and Discussion A number of studies have suggested that contact between DIKB cells and insoluble Fe(III) minerals is a necessary prerequisite for Fe(III) mineral reduction (Arnold et al., 1988; Caccavo, Jr. et al., 1992; Caccavo, Jr. et al., 1996; Lovley and Phillips, 1988; Tugel et al., 1986). The ability of S. algae BrY to catalyze reactions on the surface of Fe(O) particles would thus be predicated on the ability of S. algae BrY cells to adhere to those particles. The results in Table 4.1 demonstrate that 82 % of the S. algae BrY cells adhered to 2 mg mV1 of HEPES-pretreated iron, while cells were unable to adhere to untreated Fe(O). When 20 mg ml"1 of HEPES-pretreated iron were provided, more than 99 % of the cells adhered while only 10 % adhered to 20 mg ml'1 of untreated Fe(O). The difference in the ability of S. algae BrY to adhere to each of these forms of Fe(O) is most likely due to the presence of Fe(III) oxides on the surface of the corroded iron. Previous studies have demonstrated that DIKB preferentially adhere to Fe(III) coated surfaces (Caccavo, Jr. et al., 1997; Grantham and Dove, 1996). The results of the adhesion experiments suggest that DERB would be more likely to catalyze reactions on the surface of corroded iron than on the surface of untreated Fe(O). Table 4.1. Adhesion of S. algae BrY cells to HEPES-pretreated iron and Fe(O) treatment % adhered cells 2 mg HEPES-pretreated iron 82.2 + 6.2 2 mg Fe(O) -5.4 ± 8.2 20 mg HEPES-pretreated iron 99.3 + 5.8 20 mg Fe(O) 10.6 ± 6.3 1 numbers represent the means of three replicate assays ± the standard errors of those means 93 Following adhesion, active S. algae BrY cells were capable of restoring the capacity of HEPESpretreated iron to transform CT to CF. After 20 hours of incubation CT concentrations were statistically significantly lower (t-test, P-value 5 0.025) in treatments with active S. algae BrY than in controls containing HEPES-pretreated iron without cells or cells without iron (Figure 5.1A). In fact, neither S. algae BrY nor HEPES-pretreated iron caused significant transfoimation of CT. No direct CT transformation by BrY was observed, which agrees with previous work by Gorby et al. (1994) and Workman et al. (1997). Workman et al. (1997) were able to stimulate CT transformation in the presence of BrY through the addition of Vitamin B12. However, the levels of Vitamin B12 used by Workman et al. (1997) were much higher than normally observed in natural systems and are unlikely to have existed in these experiments. 40 60 Time [hours] Figure 5.1. 0 20 40 60 80 100 Time [hours] (A) Carbon tetrachloride degradation and (B) chloroform production by (■) passivated Fe(O), ( • ) passivated Fe(O) with active S. algae BrY cells and (▲) active S. algae BrY cells alone. Error bars represent the standard error of the means (n = 3). Treatments containing both HEPES-pretreated iron and active cells also accumulated more CF than the controls (Figure 5. IB) (Ps0.0003). Approximately 60 % of the added CT were converted to CF in tubes containing both HEPES-pretreated iron and S. algae BrY. No dichloromethane (DCM) or other transformation products were detected in any of the setups. 94 Active cells of S. algae BrY were capable of generating ferrous iron with lactate as an electron donor and HEPES-pretreated iron as a sole terminal electron acceptor (Figure 5.2). These experiments were conducted in the absence of CT since the presence of CT could have resulted in instantaneous re-oxidation of microbially produced Fe(II) to Fe(III), thus masking the production of Fe(II) by DIRB. The sum of dissolved and colloidal Fe(II) was measured in these systems. Dissolved Fe(II) is reported to be a weak reductant for CT transformation to CF (Doong and Wu, 1992) while surface associated Fe(II) is thought to be a stronger reductant (Kriegman-King and Reinhard, 1992; Kriegman-King and Reinhard, 1994; Matheson and Tratnyek, 1994). Using conventional extraction methods, however, it is difficult to measure the quantity of adsorbed Fe(II) selectively, because the grains of Fe(O) provide a large reservoir of extractable metal ions (Johnson et al., 1998). Figure 5.2 clearly shows that, over a period of ca. 6 days, the treatments with active S. algae BrY cells resulted in significantly higher Fe(II) concentrations than either control (P^ 0.027). Time [hours] Figure 5.2. Production of Fe(II) by (■) passivated Fe(O), (# ) passivated Fe(O) with active S. algae BrY cells and (▲) passivated Fe(O) with heat-killed S. algae BrY cells. Error bars represent the standard error of the means (n = 3). 95 Conclusions The influence of microorganisms on tire performance of Fe(O) subsurface barriers has only recently been recognized. Methanogenic and denitrifying bacteria have been shown to have a positive effect on the degradation of CT and CF by Fe(O) (Novak et al., 1998; Till et ah, 1998; Weathers et al., 1997) . It is reported here for the first time that DIKB can influence the transformation of a chlorinated aliphatic in the presence of passivated iron. The results indicate that DERB can have a positive influence on the performance of zero valent iron by restoring the reactivity of passivated iron to transform CT to CF. The work described herein did not allow for the identification of the exact reaction mechanism. It is possible that DIRB restore the reactivity of corroded iron by producing surface-bound reactive Fe(II) sites through microbial reduction of Fe(III) precipitates or by reductive dissolution of Fe(III) corrosion products. Reductive dissolution of ferric iron precipitates would lead to a decrease in corrosion layer thickness, which could increase the mass transport of contaminants to surface reactive sites or facilitate the electron transport from the Fe(O) to reaction sites on the corrosion product-covered surface of the iron grains (Scherer et al., 1998). Thin passive films composed of magnetite, maghemite, and other mixed valent (hydr)oxides have been modeled as semiconducting FeaO3 doped with Fe(II) (Schmuki et al., 1995). Thicker corrosion layers tend to be amorphous with fewer semiconductor characteristics (Scherer et al., 1998) . The reductive dissolution of amorphous iron (oxy)hydroxides from the grain surface would lead to a decrease in corrosion layer thickness and could result in enhanced electron transport from the underlying iron metal to surface reactive sites. More research is needed to identify the mechanisms by which DIRB influence the performance of Fe(O). Specifically designed experiments involving surface analytical equipment should allow for the identification of the reaction mechanisms. Future research should also address the role of iron-reducing bacteria in pilot and field scale barriers. DIKB could be isolated from core samples and their effects could be evaluated under more field realistic conditions (alkaline pH, field-weathered iron material). 96 A thriving DIRB population in Fe(O) barriers could also have a negative effect on the reactivity of Fe(O). The development of thick biofilms could inhibit mass transport of dissolved contaminant to the iron surface (Grady and Lim, 1980; Revsbech and Jorgensen, 1986), thus lowering the overall reactivity of Fe(O) barriers. Although currently not observed in field applications, excessive microbial growth within subsurface barriers could potentially lead to a decrease in porosity, which could result in hydraulic failure of the barrier. DIRB activity in Fe(O) barriers could also result in enhanced dissolution of reactive corrosion products thus decreasing the overall reactivity of Fe(O) subsurface barriers. 97 CHAPTER 6 CONCLUSIONS The research summarized in this dissertation was designed to evaluate the use of dissimilatory metal-reducing bacteria (DMRB) for the bioaugmentation of subsurface environments contaminated with heavy metals and chlorinated hydrocarbons. Two important aspects determining the success or failure of bioaugmentation strategies were evaluated including the effective delivery of DMRB into the subsurface and the biogeochemical influence of DMRB on the fate of contaminants. The four objectives listed in the Introduction were addressed in four separate but interrelated research projects, which are summarized in the four main chapters of this dissertation. This chapter summarizes the findings, discusses their implications, and identifies future research needs. Chapter 2 compared the porous media transport of starved and vegetative cells of Shewanella algae BrY, the model DMRB used in this dissertation. Bacterial long term starvation had been suggested as 1a transport enhancement strategy previously to this work, but a quantitative comparison had not been attempted. Based on the fractional recovery of cells (i.e. the total number of bacteria transported through a porous media column divided by the total number of cells injected) starved cells were transported better through porous media than vegetative cells. A 3.2 fold increase in fractional recovery was observed in 30 cm long columns for S. algae BrY cells, which were previously starved for 7 weeks. The longer (3 m) columns showed an approximately 23 fold increase in fractional recovery for starved cells. In addition, starved cells were more homogeneously sorbed to the quartz sand along the flowpath while vegetative cells tended to preferentially sorb to sand close to the influent side of the columns. The preferential sorption of vegetative cells to the vicinity of the injection point could potentially cause the clogging of injection wells in field situations, a well known problem in the application of bioaugmentation technologies. 98 The research described in Chapter 3 was designed to evaluate potential reasons for the transport enhancement by starvation. A number of cell properties known to influence bacterial transport through porous media was monitored over seven weeks of starvation and an existing mathematical model, based on the colloid filtration theory, was used to evaluate their importance. Changes in cell size, buoyant density, diffusion coefficient, hydrophobicity, and net electrostatic charge could only partly explain the transport enhancement by starvation. However, a comparison of the results of batch adhesion assays with the fractional recoveries from 30 cm long columns suggested that less labor intensive batch experiments might provide an effective screening tool. Such a tool would allow to screen bacterial strains for their starvation transport enhancement potential without the need to perform labor intensive column studies. Chapters 2 and 3 suggest additional experimental work, which will improve the understanding of physiological changes and mechanisms responsible for the transport enhancement by bacterial starvation. The influence of flagella, the change in surface characteristics of the cells, and the relationship between metabolic activity and bacterial transport through porous media should be evaluated. Once the transport mechanisms are understood, efforts should be focused on the development of an appropriate mathematical model. The employed model based on the filtration theory was not easily adapted to accurately describe scale- and physiology-dependent effects on bacterial transport through porous media. Methods for the reliable, selective, and quantitative assessment of bacterial numbers and activity in porous media should be developed in order to facilitate the model development process. Chapter 2 suggests a method based on the use of molecular techniques, which, if successfully developed, could provide the necessary information in the future. In view of potential field applications, the starvation time should be optimized. The results summarized in this dissertation suggest that for Shewanella algae BrY four weeks of starvation result in bacteria that transport approximately as well as bacteria starved for seven weeks. However, the concentration of bacteria in the starvation cultures were approximately five fold greater after four weeks than after seven weeks, which would allow to produce larger numbers of bacteria from the same initial volume of cells. 99 Chapters 4 and 5 focused on the ability of Shewanella algae BrY cells to directly and indirectly influence the fate of heavy metals and chlorinated hydrocarbons in the subsurface. Chapter 4 described the resuscitation of starved S. algae BrY into actively metabolizing cells in the presence of surface associated iron minerals. Active S. algae BrY cells reduced chromium(VI) directly and in addition allowed the establishment of a biogeochemical cycle involving surface reactive iron which facilitated the precipitation of chromium(VI) from contaminated water. A research project, based on these findings is currently addressing the potential feasibility of this technology by characterizing the mineral phases developing during iron and chromium reduction and by evaluating the stability of the chromium precipitates under different redox conditions. Chapter 5 described the potential influence of DMRB on the performance of zero valent iron (Fe(O)) permeable reactive barriers (PRBs). DMRB are ubiquitous in the environment and therefore likely to come in contact with existing PRBs. However, the potential effects of DMRB on Fe(O) PRBs had not been evaluated previously to this work. The microbial reduction of corrosion products, which accumulated on the Fe(O) surface by S. algae BrY cells was presumably responsible for the increase in reaction rates of Fe(O) with carbon tetrachloride. Chapter 5 was published in “Environmental Science and Technology” and is reprinted in this thesis with permission of the American Chemical Society. This publication has stimulated widespread interest and a number of funding requests and research projects in the scientific community. Drs. Michelle Scherer and Pedro Alvarez at the University of Iowa are currently investigating the influence of DMRB on the mineralogy and reactivity of Fe(O) in the presence of chlorinated aliphatics. Research groups at the Oakridge National Laboratory and MSE Technology Applications (Butte, MT) are interested in the effect of DMRB on the performance of Fe(O) PRBs for uranium elimination from groundwater. Implications for Field Applications The author of this dissertation is convinced that the research summarized herein contributes towards an improved understanding of bacterial transport through porous media. Continued research in this 100 area will contribute towards the development of a widely applicable technology for subsurface bioaugmentation. There is ample evidence that long term nutrient starvation of bacteria enhances the transport of a number of different bacterial strains and the knowledge about starvation responses suggests that starvation could be a generally applicable strategy for transport enhancement. The development of a generally applicable transport enhancement strategy would allow for the use of any bacterium of interest. Indigenous and exogenous bacterial strains and potentially whole consortia could be used. The bacteria could be starved after isolation and enrichment, and injected into contaminated subsurface environments to accelerate bioremediation processes. The ultimate success of bioaugmentation strategies will be based on the survival and activity of the injected microorganisms. The survival and activity of these organisms is influenced by many factors including nutrient availability, moisture content, physical and chemical characteristics of the environment (pH, toxicity, electron acceptors, etc.), predation, parasitism, and competition with the indigenous microbial population. Future research should, therefore, focus on further improvement of bacterial transport strategies and on the development of environmentally benign, however competitive and metabolically active inocula. Research focusing on the optimization of fluid velocity and inoculum concentration for bacterial transport and the use of site groundwater or artificial groundwater for starvation might lead to further improvements of the starvation transport enhancement strategy. Future research should evaluate the bacterial transport, bacterial activity, and long term stability of chromium precipitates on bench- and meso-scale. The long term stability of chromium precipitates will be addressed in bench scale experiments in a collaborative research project between the Idaho National Environmental and Engineering Laboratory (INEEL) and the Center for Biofilm Engineering at Montana State University. In view of a potential field scale application alternative carbon sources should be evaluated. An excess of lactate was used in the studies described. The type and amount of carbon source added to any subsurface environment should be optimized to minimize costs and the danger of excessive microbial growth close to the injection well. The transport of starved Shewanella algae BrY through porous 101 media, their resuscitation in the presence of iron minerals, and the reaction of ferrous iron with Cr(VI) should be evaluated on a larger scale. A meso-scale subsurface testing facility such as the VEGAS in Stuttgart (Germany) would be ideal for this purpose. 102 REFERENCES CITED Anonymous. 1992. "3500-Cr Chromium." In: A. E. Greenberg, L. S. Clesceri, and A. D. Eaton (Eds.), Standard methods fo r the examination o f water and wastewater. American Public Healtih Association, Washington, D.C. pp. 3-58 - 3-60. Abbott A., P. R. Rutter, and R. C. W. Berkeley. 1983. "The influence of ionic strength, pH and a protein layer on the interaction of Streptococcus mutans and glass surfaces." J.Gen.Microbiol. 129:439445. Absolom D. R., R V. Lamberti, A. Policova, W. Zingg, C. J. VanOss, and A. W. Neumann. 1983. "Surface thermodynamics of bacterial adhesion." Appl.Environ.Microbiol. 46:90-97. Agrawal A., P. G. Tratnyek. 1996. "Reduction of nitro aromatic compounds by zero-valent iron metal." Environ.Sci.Technol. 50:153-160. Albertson N. H., T. Nystrom, and S. Kjelleberg. 1990."Starvation-induced modulations in binding proteindependent glucose transport by the marine Vibrio sp. Sl4." FEMS Microbiol. Lett. 55:205-209. Albinger O., B. K. Biesemeyer, R. G. Arnold, and B. E. Logan. 1994. "Effect of bacterial heterogeneity on adhesion to uniform collectors by monoclonal populations." FEMS Microbiol.Lett. 124:321-326. Alexander M. 1999. Biodegradation and bioremediation. 2nd ed. Academic Press, San Diego. Alexander, M., Wagenet, R. J, Baveye, J. T., Gannon, J. T., Mingelgrin, U, and Tan, Y. 1991. "Movement of Bacteria Through Soil and Aquifer Sand." EPA/600/S2-91/010. Center for Environmental Research Information Cincinnati, OH. Amonette, I. E., Fmchter, J. S., Gorby, Y. A., Cole, C. R., Cantrell, K. J, and Kaplan, D. I. 1998. "Method of removing oxidized contaminants from water." US Patent 5783088. Amonette, J. E., Szecsody, I. E., Schaef, H. T., Templeton, J. C., Gorby, Y. A., and Fmchter, J. S. 1994. Abiotic reduction of aquifer materials by dithionite: A promising in-situ remediation technology. In Gee, G. W. and Wing, N. R.: Proceedings of the 33rd Hanford Symposium on Health and the Environment. In-situ remediation: Scientific basis for current and future technologies. 2:pp. 851881. Amonette J. E., D. J. Workman, D. W. Kennedy, J. S. Fmchter, and Y. A. Gorby. 2000. "Dechlorination of Carbon Tetrachloride by Fe(II) Associated with Goetbite." Environ.Sci.Technol. 34:4606-4613. Amy P. S., D. L. Haldeman. 1997. The Microbiology o f the Terrestrial Deep Subsurface. CRC Press, Boca Raton, FE. Amy P. S., R. Y. Morita. 1983. "Starvation-Survival Patterns of Sixteen Freshly Isolated Open-Ocean Bacteria." Appl.Environ.Microbiol. 45:1109-1115. 103 Anderson L. D., D. B. Kent, and J. A. Davis. 1994. "Batch Experiments Characterizing the Reduction of Cr(VI) Using Suboxic Material from a Mildly Reducing Sand and Gravel Aquifer." Environ.Sci.Technol. 25:178-185. • • Arnold R. G., T. J. DiChnstina, and M..R. Hoffmann. 1988. "Reductive dissolution of Fe(III) oxides by Pseudomonas sp. 200." Biotechnol.Bioeng. 32:1081-1096. Bader J. L., G. Gonzalez, P. C. Goodell, A. S. AM, and S. D. Pillai.' 1999. "Aerobic Reduction of Hexavalent Chromium in Soil by Indigenous Microorganisms." Bioremediation Journal 3:201211. Balch W. E., G. E. Fox, L. J. Magrum, C. R. Woese, and R. S. Wolfe. 1979. "Methanogens: reevaluation of a unique biological group." Microbiological Reviews 43:260-296. Balch W. E., R. S. Wolfe. 1976. "New approach to the cultivation of methanogenic bacteria: 2mercaptoethanesulfonic acid (HS-CoM)-dependent growth of Meihanobacterium ruminantium in a pressurized atmosphere." Appl.Environ.Microbiol. 32:781-791. Bales R. C., C. P. Gerba, G. H. Grondin, and S. C. Jensen. 1989. "Bacteriophage transport in sandy soil and fractured tuff." Appl.Environ.Microbiol. 55:2061-2067. Bartlett R. I., B. R. James. 1979. "Behavior of Chromium in Soils: III. Oxidation." J.Environ.Engin. 3:3135. Bartlett R. J., B. R. James. 1988. "Mobility and Bioavailability of Chromium in Soils." In: J. O. Nriagu, and E. Nieboer (Eds.), Chromium in the 'Natural and Human Environments. John Wiley and Sons, New York. pp. 267-304. Barton J. W., R. M. Ford. 1995. "Determination of Effective Transport Coefficients for Bacterial Migration in Sand Columns." Appl.Environ.Microbiol. 67:3329-3335. Barton J. W., R. M. Ford. 1997. "Mathematical model for Characterization of Bacterial Migration through Sand Cores." Biotechnol.Bioeng. 53:487-496. Baygents J. C., J. R. J. Glynn, O. Albinger, B. K. Biesemeyer, K. L. Ogden, and R. G. Arnold. 1998. "Variation of surface charge density in monoclonal bacterial populations: implications for transport through porous media." Environ.Sci. Technol. 32:1596-1603. Beets K. S. 1998. "Novel barrier remediates chlorinated solvents." Environ.Sci.Technol. 32:495A. Bellamy W. D., D. W. Hendricks, and G. S. Logsdon. 1985. "Slow sand filtration theory: Influences of selected process variables." J.Am. Water Well Assoc. 72:62-66. Berry D. F., C. Hagedom. 1991. "Soil and Groundwater Transport of Microorganisms." Biotech. 75:57-73. Bitton G., N. Lahav, and Y. Henis. 1974. "Movement and retention of Klebsiella aerogenes in soil columns." Plant and Soil 40:373-380. Blowes D. W., C. I. Ptacek, and J. L. Jambor. 1997a. "In-Situ Remediation of Cr(VI)-Contaminated Groundwater Using Permeable Reactive Walls: Laboratory Studies." Environ.Sci.Technol. 37:3348-3357. ; 104 Blowes D. W., C. J. Ptacek, and J. L. Jambor. 1997b. "Insitu remediation of Cr(VI)-contaminated groundwater using permeable reactive walls: Laboratory studies." Environ.Sci.Technol. 57:33483357. Bolster C. H., A. L. Mills, G. M. Homberger, and J. S. Herman. 2000. "Effect of Intra-Population Variability on the Long-Distance Transport of Bacteria." Ground Water 55:370-375. Bonin P. M., M. S. Odziemkowski, and R. W. Gillham. 1998. "Influence of chlorinated solvents on the polarization and corrosion behavior of iron in borate buffer." Corrosion Science 40:1391-1409. Bopp L. H., H. L. Ehrlich. 1988. "Chromate Resistance and Reduction in Pseudomonas fluorescens strain LB300." Archives o f Microbiology 750:426-431. Bouwer E. J., H. H. M. Rijnaarts, A. B. Cunningham, and R. Gerlach. 2000. "Biofilms in Porous Media." In: I. D. Bryers (Ed.), Biofilms II: Process Analysis and Applicatiom. Wiley-Liss, Inc., New York, pp. 123-158. Burlage R. S., Z. K. Yang, and T. Mehlhom. 1996. "A Transport for Green Fluorescent Protein Transcriptional Fusions: Application for Bacterial Transport Experiments." Gene 775:53-58. Busscher H. L, M. N. Bellon-Fontaine, N. Mozes, H. C. Van der Mei, J. Sjollema, O. Cerf, and P. G. Rouxhet. 1990a. "Deposition of Leuconostoc mesenteroides and Streptococcus thermophilus to solid substrata in a parallel plate flow cell." Biofouling 2:55-63. Busscher H. J., M. N. Bellon-Fontaine, I. Sjollema, and H. C. Van der Mei. 1990b. "Relative Importance of Surface Free Energy as a Measure of Hydrophobicity in Bacterial Adhesion to Solid Surfaces." In: R. I. Doyle, and M. Rosenberg (Eds.), Microbial Cell Surface Hydrophobicity. American Society for Microbiology Press, Washington, D C. pp. 335-359. Busscher H. J., M. M. Cowan, and H. C. Van der Mei. 1992. "On the relative importance of specific and non-specific approaches to oral microbial adhesion." FEMS Microbiol. Rev. 55:199-210. Caccavo F., Jr. 1999. "Protein-Mediated Adhesion of the Dissimilatory Fe(HI)-Reducing Bacterium Shewanella alga BrY to Hydrous Ferric Oxide." AppLEnvironMicrobiol. 65:5017-5022. , Caccavo F., Jr., N. Birger Ramsing, and J. W. Costerton. 1996. "Morphological and metabolic responses to , starvation by the dissimilatory metal-reducing bacterium Shewanella alga BrY." Appl.Environ.Microbiol. 62:4678-4682. Caccavo F., Jr., R. P. Blakemore, and D. R. Lovley. 1992. "A hydrogen-oxidizing, Fe(HI)-reducing microorganism from the Great Bay Estuary, New Hampshire." Appl.Environ.Microbiol: 55:32113216. Caccavo F., Jr., B. Frolund, F. van Ommen Kloeke, and P. H. Nielsen. 1996. "Deflocculation of activated sludge by the dissimilatory Fe(III)-reducing bacterium Shewanella alga BrY." Appl.Environ.Microbiol. 62:1487-1490. Caccavo F., Jr., P. C. Schamberger, K. Keiding, and P. H. Nielsen. 1997. "Role of hydrophobicity in adhesion of the dissimilatory Fe(III)-reducing bacterium Shewanella alga to amorphous Fe(IH) oxide." Appl.Environ.Microbiol. 65:3837-3843. 105 Camesano T. A., B. E. Logan. 1998. "Influence of fluid velocity and cell concentration on the transport of motile and nomnotile bacteria in porous media." Environ.Sci.Technol. 52:1699-1708. Camper A. K., J. T. Hayes, P. Sturman, W. L. Jones, and A. B. Cunningham. 1993. "Effects of motility and adsorption rate coefficient on transport of bacteria through saturated porous media." Appl.Environ.Microbiol. 59:3455-3462. Camper A. K., M. W. LeCheyaliier, S. C. Broadaway, and G. A. McFeters. 1985. "Evaluation of procedures to desorb bacteria from granular activated carbon." J.Microbiol.Meth. 5:187-198. Cantrell K. J., D. I. Kaplan. 1997a. "Zero-valent iron emplacement in sand columns." JEnviron.Engin. 725:499-505. ' Cantrell K. J., D. I. Kaplan, and T. J. Gilmore. 1997b. "Injection of colloidal Fe0 particles in sand with shear-thinning fluids." JEnviron.Engin. /25:786-791. Cantrell K. I., D. I. Kaplan, and T. W. Wietsma. 1995. "Zero-Valent Iron for the In Situ Remediation of Selected Metals in Groundwater." J.Haz.Materials 42:201-212. Cao I., L. Wei, Q. Huang, L. Wang, and S. Han. 1999. "Reducing degradation o f azo dyes by zero-valent iron in aqueous solution." Chemosphere 58:565-571. Cappelier J. M., A. Rossero, and M. Federighi. 2000. "Demonstration of a protein synthesis in starved Campylobacter jejuni cells." Int J. Food Microbiol. 55:63-67. Cary E. E., W. H. Allaway, and O. E. Olson. 1977. "Control of Chromium Concentration in Food Plans 2. Chemistry of Chromium in Soils and Its Availability to Plants." J.Agric.Food Chem. 25:305-309. Committee on Biologic Effects of Atmospheric Pollutants. 1974. Chromium. National Academy of Sciences, Washington, D.C. Corapciogiu M. Y., A. Haridas. 1984. "Transport and Fate of Microorganisms in Porous Media: A Theoretical Investigation." J.Hydrol. 72:149-169. Costerton J. W., G. G. Geesey, and K.-J. Cheng. 1978. "How bacteria stick." SclAm. 258:86-95. Crawford R. L., D. L. Crawford. 1996. Bioremediation principles and applications. Cambridge University Press, Cambridge, New York. Cunningham, A. B. 2000. Subsurface Biofilm Barriers: An Emerging Technology for Containment and Remediation of Contaminated Groundwater. Keynote Presentation at the Hazardous Waste Research 2000 Conference in Denver, CO, 23-25 May 2000. Cunningham A. B., B. Warwood, P. Sturman, K. Horrigan, G. James, J. W. Costerton, and R. Hiebert. 1997. "Biofilm processes in porous media - practical applications." In: P. S. Amy, and D. L. Haldeman (Eds.), The Microbiology o f the Terrestrial Deep Subsurface. CRC Press, Boca Raton, FE. pp. 325-344. Cusack F., S. Singh, C. McCarthy, J. Grecco, M. De Rocco, D. Nguyen, H. M. Lappin-Scott, and J. W. Costerton. 1992. "Enhanced oil recovery - three dimensional sandpack simulation of ultramicrobacteria resuscitation in reservoir formations." J. Gen.Microbiol. /58:647-655. 106 Das A., F. Caccavo Jr. 2000. "Dissimilatory Fe(III) Oxide Reduction by Shewanella alga BrY Requires Adhesion." Curr.Microbiol. 40:344-347. Davis A., R. L. Olsen. 1995. "The geochemistry of chromium migration and remediation in the subsurface." Ground Water 55:759-768. Davis, J. B. 1972. "The Mitigation of Petroleum Products in Soil Ground Water: Principles of Countermeasures." Publication 4149. American Petroleum Institute Dawson M. P., B. Humphrey, and K. C. Marshall. 1981. "Adhesion: A Tactic in the Survival Strategy of a Marine Vibrio During Starvation." Curr.Microbiol. <5:195-198. DeFlaun M. F., C. Murray, W. Holben, T. Scheibe, A. Mills, T. Ginn, T. GrifFm, E. Majer, and J. Wilson. 1997. "Preliminary observations on bacterial transport in a coastal plain aquifer." FEMS Microbiol. Rev. 20:473-487. DeFlaun M. F., S. R. Oppenheimer, S. Streger, C. W. Condee, and M. Fletcher. 1999. "Alterations in Adhesion, Transport, and Membrane Characteristics in an Adhesion-Deficient Pseudomonad." Appl.Environ.Microbiol. (55:759-765. DeFlaun M. F., A. S. Tanzer, A. L. McAteer, B. M. Marshall, and S. B. Levy. 1990. "Development of adhesion assay and characterization of an adhesion-deficient mutant of Pseudomonas fluorescens." • Appl.Environ.Microbiol. 56:112-119. Deshpande P. A., D. R. Shonnard. 1999. "Modeling the effects of systematic variation in ionic strength on the attachment kinetics of Pseudomonas fluorescens UPER-I in saturated sand columns." Wat.Resourc.Res. 55:1619-1627. Devlin J. F., J. Klausen, and R. P. Schwarzenbach. 1998. "Kinetics of nitroaromatic reduction on granular iron in recirculating batch experiments." Environ.Sci.Techhol. 52:1941-1947. Doong R.-A., S.-C. Wu. 1992. "Reductive dechlorination of chlorinated hydrocarbons in aqueous solutions containing ferrous iron and sulfide ions." Chemosphere 24:1063-1075. Duineveld B. M., G. A. Kowalchuk, A. Keijzer, J. D. van Elsas, and J. A. van Veen. 2001. "Analysis of Bacterial Communities in the Rhizosphere of Chrysanthemum via Denaturing Gradient Gel Electrophoresis of PCR-Amplified 16S rRNA as Well as DNA Fragments Coding for 16S rRNA." Appl.Environ.Microbiol. 67:172-178. Dybas M. J., M. Barcelona, S. Bezborodnikov, S. Davies, L. Forney, H. Heuer, O. Kawka, T. Mayotte, T. L. Sepulveda, K. Smalla, M. Sneathen, J. Tiedje, T. Voice, D. C. Wiggert, M. E. Witt, and C. S. Criddle. 1998. "Pilot-scale evaluation of bioaugmentation for in-situ remediation of a carbon tetrachloride-contaminated aquifer." Environ.Sci.Technol. 52:3598-3611. Eary L. E., D. Rai. 1988. "Chromate Removal from Aqueous Wastes by Reduction with Ferrous Ion." Environ.Sci. Technol. 22:972-977. Eary L. E., D. Rai. 1991. "Chromate Reduction by Subsurface Soils uncler Acidic Conditions." Soil Sci.Soci.AmJ. 55:676-683. 107 Ellis D. E., E. J. Lutz, J. M. Odom, R. J. Buchanan, C. L. Bartlett, M. D. Lee, M. R. Harkness, and K. A. DeWeerd. 2000. "Bioaugmentation for accelerated in situ anaerobic bioremediation." Environ.Sci. TechnoL 34:2254-2260. Erbs M., H. C. Bruun Hansen, and C. E. Olsen. 1999. "Reductive dechlorination of carbon tetrachloride using iron(II) iron(III) hydroxide sulfate (green rust)." Environ.Sci.Technol. 55:307-311. Evanko, Cynthia R. and Dzombak, David A. 1997. "Technology Evaluation Report: Remediation of Metals-Contaminated Soils and Groundwater." TE-97-01. Ground-Water Remediation Technologies Analysis Center Pittsburgh, PA. Eykholt G. R., D. T. Davenport. 1998. "Dechlorination of chloroacetaniline herbicides alachlor and metolachlor by iron metal." Environ.Sci. Technol. 32:1482-1487. Fendorf S. E., G. C. Li. 1996. "Kinetics of chromate reduction by ferrous iron." Environ.Sci.Technol 50:1614-1617. Fiedor J. N., W. D. Bostick, R. J. Jarabek, and J. Farrell. 1998. "Understanding the mechanism of uranium removal from groundwater by zero-valent iron using x-ray photoelectron spectroscopy." Environ.Sci.Technol. 52:1466-1473. Fischer D., A. Teich, P. Neubauer, and R. Hengge-Aronis. 1998. "The general stress sigma factor sigmaS of Escherichia coli is induced during diauxic shift from glucose to lactose." JBacteriol 180:62036206. Fletcher M. 1976. "The effects of proteins on bacterial attachment to polystyrene." J.Gen.Microbiol. 94:400-404. Fletcher M., G. D. Floodgate. 1973. "An electron-microscopic demonstration of an acidic polysaccharide . involved in the adhesion of a marine bacterium to solid surfaces." J.Gen.Microbiol. 74:325-334. Fletcher M., G. L. Loeb. 1979. "Influence of Substratum Characteristics on the Attachment of a Marine Pseudomonad to Solid Surfaces." Appl.Environ.Microbiol. 57:67-72. Fletcher M., K. C. Marshall. 1982. "Bubble Contact Angle Method for Evaluating Substratum Interfacial Characteristics and Its Relevance to Bacterial Attachment." Appl.Environ.Microbiol. 44:184-192. Fontes D. E., A. L. Mills, G. M. Homberger, and J. S. Herman. 1991. "Physical and Chemical Factors Influencing Transport of Microorganisms Through Porous Media." Appl.Environ.Microbiol. 57:2473-2481. Fortner B. 1995. "Wall of iron purifies contaminated groundwater." Water Environment and Technology 7:14-17. Foster J. W., M. P. Specton 1995. "How Salmonella survive against the odds." Annu.Rev.Microbiol 49:145174. Fredrickson I. K., Y. A. Gorby. 1996a. "Environmental processes mediated by iron-reducing bacteria." Current Opinion in Biotechnology 7:287-294. Fredrickson J. K., T. C. Onstott 1996b. "Microbes Deep Inside the Earth." ScientiflcAmerica 275:68-73. 108 Frachter, J. S., Cole, C. R., Williams, M. D., Vermeul, V. R., Teel, S. S., Amonette, J. E., Szecscody, J. E., and Yabusaki, S. B. 1997. Creating a subsurface permeable treatment barrier using in situ redox manipulation.".Pacific Northwest National Laboratory Richland, WA. . Gannon J. T., V. B. Mamlal, and M. Alexander. 1991a. "Relationship Between Cell Surface Properties and Transport of Bacteria Through Soil." Appl.Environ.Microbiol. 57:190-193. Gannon J. T., Y. Tan, P. Baveye, and M. Alexander. 1991b. "Effect of Sodium Chloride on Transport of Bacteria in a Saturated Aquifer Material." Appl.Environ.Microbiol. 57:2497-2501. Gatpagar J., E. Grulke, T. Tsang, and D. Bhattacharyya. 1997. "Reductive dehalogenation of trichloroethylene using zero-valent iron." EnvironmentalProgress /5:137-143. Gavaskar A. R., N. Gupta, B. M. Sass, R. J. Janosy, and D. O'Sullivan. 1998. Permeable Barriers fo r Groundwater Remediation; Design, Construction, and Monitoring. Battelle Press, Columbus, Richland. Geesey G. G., J. W. Costerton. 1979. "Microbiology of a northern river: Bacterial distribution and relationship to suspended sediment and organic carbon." Can.J.Microbiol. 25:1058-1062. Gerlach, R., Cunningham, A. B., and Caccavo, F., Jr. 1998. Formation of Redox-Reactive Subsurface • Barriers Using Dissimilatory Metal-Reducing Bacteria. In Ericksen, L. E. and Rankin, M. M.: Proceedings of the 1998 Conference on Hazardous Waste Research - Bridging Gaps in Technology and Culture. May 18-21, 1998. Snowbird, UT. pp. 209-223. Gibb H. J. 1999. "Epidemiologic Study of Chromate Production Workers: A Summary." Plating and Surface Finishing 86:19-19. Gibbons J. D. 1971. Nonparametric statistical interference. Mc-Graw Hill, New York. Gilbert P., D. J. Evans, I. G. Duguid, and M. R. W. Brown. 1991. "Surface characteristics and adhesion of Escherichia coli and Staphylococcus epidermidis." J.Appl.Bacteriol. 77:72-77. Gillham R. W., S. O'Hannesin. 1994. "Enhanced Degradation of Halogenated Aliphatics by Zero-Valent Iron." Ground Water 32:958-967. Goldberg S., Y. Konis, and M. Rosenberg. 1990. "Effect of Cetylpyridinium chloride on microbial adhesion to hexadecane and polystyrene." Appl.Environ.Microbiol. 55:1678-1682. Gorby, Y. A., Amonette, S. E., and Fruchter, J. S. 1994. Remediation of Contaminated Subsurface Materials by a Metal-Reducing Bacterium. Proceedings of the 33rd Hanford Symposium on Health and the Environment. In-situ remediation: Scientific basis for current and future technologies, pp. 233-247. Gorby, Y. A., Workman, D. J., Amonette, S. E., and Fruchter, J. S'. 1995. Transformation of carbon tetrachloride by biogenic Fe(II). Emerging Technologies in Hazardous Waste Management VH. p. 593. Gossett J. M. 1987. "Measurement of Henry's Law constants for Ci and C2 chlorinated hydrocarbons." Environ.Sci. Technol. 21:202-208. 109 Gould J. P. 1982. "The kinetics of hexavalent chromium reduction by metallic iron." Wat.Res. 75:871-877. Grady C. P. L., H. C. Lim. 1980. Biological wastewater treatment theory and applications. Marcel Dekker New York. Grantham M. C., P. M. Dove. 1996. "Investigation of bacterial-mineral interactions using Fluid Tapping Mode™ atomic force microscopy." Geochimica et Cosmochimica Acta 74:2473-2480. Grasso D., B. F. Smets, K. A. Strevett, B. D. Machinist, C. J. van Oss, R. F. Giese, and W. Wu. 1996. "Impact of physiological state on surface thermodynamics and adhesion of Pseudomonas aeruginosa.”Environ.Sci.Technol. 50:3604-3608. Groat R. G., J. E. Schultz, E. Zychlinsky, A. Bockman, and A. Matin. 1986. "Starvation proteins in Escherichia coir, kinetics of synthesis and role in starvation survival." J. Bacterial. 765:486-493. Gross M. I., O. Albinger, D. G. Jewett, B. E. Logan, R. C. Bales, and R. G. Arnold. 1995a. "Measurement of bacterial collision efficiencies in porous media." Wat. Res. 29:1151-1158. Gross M. I., B. E. Logan. 1995b. "Influence of Different Chemical Treatments on Transport of Alcaligenes paradoxus in Porous Media." Appl.Environ.Microbiol. 67:1750-1756. Harding G. L. 1997. "Bioremediation and the dissimilatory Reduction of Metals." In: C. J. Hurst, G. R. Knudsen, M. J. Mclnemey, L. D. Stetzenback, and M. V. Walter (Eds.), Manual o f Environmental ■Microbiology. ASM Press, Washington D.C. pp. 806-810. Hartke A., S. Bouche, X. Gansel, P. Boutibonnes, and Y. Auffray. 1994. "Starvation-Induced Stress Resistance in Lactococcus lactis subsp. Iactis IL1403." Appl.Environ.Microbiol. 60:3474-3478. Harvey R. W. 1991. "Parameters involved in modeling movement of bacteria in groundwater." In: C. J. Hurst (Ed.), Modeling the Environmental Fate o f Microorganisms. American Society for Microbiology, Washington D.C. pp. 89-114. Harvey R. W. 1997. "In situ and laboratory methods to study subsurface microbial transport." In: C. J. Hurst, G. R. Knudsen, M. J. Mclnemey, L. D. Stetzenback, and M. V. Walter (Eds.), Manual o f Environmental Microbiology. ASM Press, Washington D.C. pp. 586-599. • Harvey R. W., L. H. George, R. L. Smith, and D. R. LeBlanc. 1989. "Transport of microspheres and indigenous bacteria through a sandy aquifer: Results of natural- and forced-gradient tracer experiments." Environ.Sci.Technol. 25:51-56. Harvey R. W., N. Kinner, D. MacDonald, D. W. Metge, and A. Bunn. 1993. "Role of Physical Hererogeneity in the Inteipretation of Small-Scale Laboratory and Field Observations of Bacteria, Microbial-Sized Microsphere and Bromide Transport Through Aquifer Sediments." Wat.Resourc.Res. 29 :2713-2721. Harvey R. W., D. W. Metge, N. Kinner, and N. Mayberry. 1997. "Physiological considerations in applying laboratory-determined buoyant densities to predictions of bacterial and protozoan transport in groundwater: results of in-situ and laboratory tests." Environ.Sci.Technol. 57:289-295. Hayes R. B. 1997. "The Carcinogenicity of Metals in Human." Cancer Causes and Control 5:371-385. no Heijman C. G., E. Grieder, C. Holliger, and R. P. Schwarzenbach. 1995. "Reduction of Nitroaromatic Compounds Coupled to Microbial Iron Reduction in Laboratory Aquifer Columns." Environ.Sci. Technol. 29:775-783. Heijman C. G., C. Holliger, M. A. Glaus, R. P. Schwarzenbach, and J. Zeyer. 1993. "Abiotic Reduction of 4-chloronitrobenzene, to 4-chloroaniline in a dissimilatory iron-reducing enrichment culture." AppLEnvironMicrobiol. 59:4350-4353. Helland B. R., P. J. J. Alvarez, and J. L. Schnoor. 1995. "Reductive dechlorination of carbon tetrachloride with elemental iron." J Haz. Mat. 47:205-216. Henderson T. 1994. "Geochemical Reduction of Hexavalent Chromium in the Trinity Sand Aquifer." Ground Water 52:477-486. Hengge-Aronis R. 1993. "Survival of hunger and stress: the role of rpoS in early stationary phase gene regulation in E. coli." Cell 72:165-168. Hood M. A., M. T. MacDonelL 1987. "Distribution of ultramicrobacteria in a gulf coast estuary and induction of ultramicrobacteria." Microbiol.Ecol. 74:113-127. Huang C.-P., H.-W. Wang, and P. C. Chiu. 1998. "Nitrate reduction by metallic iron." Waf.Res. 32:22572264. Hug S. J., H. Laubscher. 1997. "Iron(III) Catalyzed Photochemical Reduction of Chromium(VI) by Oxalate and Citrate in Aqueous Solutions." Environ.Sci.Technol. 37:160-170. Istok J. D., J. E. Amonette, C. R. Cole, J. S. Fruchter, M. D. Humphrey, J. E. Szecscody, S. S. Teel, V. R. Vermeul, M. D. Williams, and S. B. Yubasaki. 1999. "In Situ Redox Manipulation by Dithionite Injection: Intermediate-Scale Laboratory Experiments." Ground Water 37:884-889. Jackson A., D. Roy, and G. Breitenbeck. 1994. "Transport of a bacterial suspension through a soil matrix using water and an anionic surfactant." Wat. Res. 23:943-949. James B. R. 1994. "Hexavalent Chromium Solubility and Reduction in Alkaline Soils Enriched with Chromite Om Processing Residue." J. Environ. Qual. 23:227-233. James B. R. 1996. "The Challenge of Remediating Chromium-Contaminated Soil." Environ.Sci.Technol. 30:248A-251A. Jenneman G. E., M. J. Mclnemey, and R. M. Knapp. 1985. "Microbial Penetration through NutrientSaturated Berea Sandstone." Appl.Environ.Microbiol. 50:383-391. Jewett D. G., T. A. Hilbert, B. E. Logan, R. G. Arnold, and R. C. Bales. 1995. "Bacterial transport in laboratory columns and filters: Influence of ionic strength and pH on collision efficiency." Wat.Res. 29:1673-1680. Johnson T. L., W. Fish, Y. A. Gorby, and P. G. Tratnyek. 1998. "Degradation of Carbon Tetrachloride by Iron Metal: Complexation Effects on the Oxide Surface." J.Contam.Hydrol. 29:379-398. Johnson T. L., M. M. Scherer, and P. G. Tratnyek. 1996. "Kinetics of halogenated organic compound degradation by iron metal." Environ.Sci.Technol. 30:2634-2640. Ill Johnson, T. L and Tratnyek, P. G. 1994. A column study of geochemical factors affecting reductive dechlorination of chlorinated solvents by zero-valent iron. In Gee, G. W. and Wing, N. R.: Proceedings of the 33rd Hanford Symposium on Health and the Environment. In-situ remediation: Scientific basis for current and future technologies, pp. 931-947. Johnson W. P., B. E. Logan. 1996. "Enhanced Transport of Bacteria in Porous Media by Sediment-Phase and Aqueous-Phase Natural Organic Matter." Wat.Res. 30:923-931. Johnson W. P., P. Zhang, M. E. Fuller,’ T. D. Scheibe, B. J. Mailloux, T. C. Onstott, M. F. DeFlaun, S. S. Hubbard, J. Radtke, W. P. Kovacic, and W. Holben. 2001. "Ferrographic Tracking of Bacterial Transport in the Field at the Narrow Channel Focus Area, Oyster, VA." Environ.Sci Technol 35:182-191. Kaplan R., D. Apirion. 1975a. "Decay of ribosomal ribonucleic acid in Escherichia coli cells starved for various nutrients." J. Biol. Chem. 250:3174-3178. Kaplan R., D. Apirion. 1975b. "The fate of ribosomes in Escherichia coli cells starved for a carboni source." J. Biol. Chem. 250 :1854-1863. Keith L. H., W. A. Telliard. 1979. "Priority Pollutants I-a perspective view." Environ.Sci.Technol. 73:416423. Kinoshita T., R. C. Bales, M. T. Yahya, and C. P. Gerba. 1993. "Bacteria transport in a porous medium: Retention of Bacillus and Pseudomonas on silica surfaces." Wat. Res. 27:1295-1301. Kjelleberg S. 1984. "Effects of Interfaces on Survival Mechanisms of Copiotrophic Bacteria in LowNutrient Habitats." In: M. J. King, and C. A. Reddy (Eds.), Current Perspectives in Microbial Ecology. American Society for Microbiology, Washington, D.C. pp. 151-159. Kjelleberg S. 1993a. Starvation in Bacteria. Plenum Press, New York. Kjelleberg S., N. Albertson, K. Flardh, L. Holmquist, A. Jouper-Jaan, R. Marouga, J. Ostling, B. Svenblad, and D. Weichart. 1993b. "How do non-differentiating bacteria adapt to starvation?" Antonie Van Leeuwenhoek 63-32>?>-34\. Klausen I , S. P. Trdber, S. B. Haderlein, and R. P. Schwarzenbach. 1995. "Reduction of substituted nitrobenzenes by Fe(II) in aqueous mineral suspensions." Environ.Sci.Technol. 29:2396-2404. Kostka J. E., K. H. Nealson. 1995. "Dissolution and reduction of magnetite by bacteria." Environ.Sci. Technol. 29:2535-2540. Knegman-King M., M. Reinhard. 1992. "Transformation of carbon tetrachloride in the presence of sulfide, biotite, and vermiculite." Environ.Sci.Technol. 26:2198-2206. Knegman-King M., M. Reinhard. 1994. "Transformation of carbon tetrachloride by pyrite in aqueous solution." Environ.ScLTechnol. 25:692-700. Landis, R. and Vidumsky, John 1999. "Recent Advancements in PRB Emplacement." In Situ Permeable Barriers: Application and Deployment Training Session. United States Environmental Protection Agency Boston, MA. June 22-23, 1999. 112 Langard S. 1980. "Chromium." In: H. A. Waldron (Ed.), Metals in the Environment. Academic Press Inc., London, pp. 111-132. Lantz W. L. 1992. "Remediation of Soil and Groundwater at a Superfund Site." Water Environment & Technology 12-15. Lappin-Scott H. M., I. W. Costerton. 1990. "Starvation and penetration of bacteria in soils and rocks." Experientia 4<5:807-812. Lappin-Scott H. M., I. W. Costerton. 1992. Curr. Opin. Biotechnol 5:283-285. Lappin-Scott H. M., F. Cusack, F. A. MacLeod, and I. W. Costerton. 1988a. "Nutrient Resuscitation and Growth of Starved Cells in Sandstone Cores: A Novel Approach to Enhanced Oil Recovery." ApplEnvironMicrobiol. 54:1373-1382. Lappin-Scott H. M., F. Cusack, F. A. MacLeod, and J. W. Costerton. 1988b. "Starvation and Nutrient Resuscitation of Klebsiella pneumoniae isolated from oil well waters." JApplBacteriol <54:541549. Lawrence J. R., M. I. Hendry. 1996. "Transport of bacteria through geologic media." CanJMicrobiol 42:410-422. Logan B. E., T. A. Hilbert, and R. G. Arnold. 1993. "Removal of bacteria in laboratory filters: Models and experiments." Wat. Res. 27:955-962. Logan B. E., D. G. Jewett, R. G. Arnold, E. J. Bouwer, and C. R. OMelia. 1995. "Clarification of clean-bed filtration models." J.Environ.Engin. /21:869-873. Lovley D. R. 1993. "Dissimilatory Metal Reduction." Annu. RevMicrobiol 47:263-290. Lovley D. R. 1995a. "Bioremediation of organic and metal contaminants with dissimilatory metalreduction." J. Indust. Microbiol. /4:85-93. Lovley D. R. 1995b. "Microbial Reduction of Iron, Manganese, and Other Metals." Advances in Agronomy 54:175-231. Lovley D. R. 1997. "Microbial Fe(HI) reduction in Subsurface environments." FEMS Microbiol. Rev. 20:305-313. Lovley D. R., E. J. P. Phillips. 1986. "Organic blatter Mineralization with Reduction of Ferric Iron in Anerobic Sediments." ApplEnvironMicrobiol 57:683-689. Lovley D. R., E. J. P. Phillips. 1987. "Rapid assay for microbially reducible ferric iron in aquatic sediments." ApplEnvironMicrobiol 53:1536-1540. Lovley D. R., E. J. P. Phillips. 1988. "Novel mode of Microbial Energy Metabolism: Organic Carbon Oxidation Coupled to Dissimilatory Reduction of Iron or Manganese." ApplEnvironMicrobiol 54:1472-1480. Mackay D. M., J. A. Cherry. 1989. "Groundwater Contamination: Pump-and-Treat Remediation." Environ.Sci.Technol. 23:630-636. 113 Mackenzie, P. D., Baghel, S. S., Eykholt, G. R., Homey, D. P., Salvo, J. I , and Sivavec, T. M. 1995. Extended Abstracts, 209th National Meeting, American Chemical Society, Division of Environmental Chemistry. 35:pp. 796-799. Mackenzie, P. D., Sivavec, T. M., and Homey, D. P. 1997. Extending hydraulic lifetime or iron walls. 2nd International Containment Technology Conference, pp. 781-787. . MacLeod F. A., H. M. Lappin-Scott, and J.. W. Costerton. 1988. "Plugging of a Model Rock System by Using Starved Bacteria." Appl.Environ.Microbiol. 54:1365-1372. Malusis M. A., D. J. Adams, K. F. Reardon, C. D. Shackleford, D. C. Mosteller, and A. W. Bourquin. 1997. "Microbial transport in a pilot-scale biological treatment zone." In: B. C. Alleman, and A. Leeson (Eds.), Bioremediation. Battelle Press, Columbus, OH. pp. 559-564. Marlow H. I., K. L. Duston, M. R. Wiesner, M. B. Tomson, J. T. Wilson, and C. H. Ward. 1991. "Microbial transport through porous media: The effects of hydraulic conductivity and injection velocity." J. Haz. Mat. 25:65-74. Martin M. I., B. E. Logan, and W. P. Johnson. 1996. "Scaling bacterial filtration rates in different sized porous media." J.Environ.Engin. 722:407-415. Matheson L. I., P. G. Tratnyek. 1994. "Reductive Dehalogenation of Chlorinated Methanes by Iron Metal." Environ.Sci.Technol. 25:2045-2053. Matin A. 1991. "The molecular basis of carbon-starvation-induced general resistance in Escherichia coli." Mol.Microbiol. 5:3-10. Matin A., E. A. Auger, P. H. Blum, and J. E. Schultz. 1989. "Genetic basis of starvation survival in nondifferentiating bacteria." Annu.Rev.Microbiol. 45:293-316. Maxwell C. R., H. A. Baqai. 1995. "Remediation of petroleum hydrocarbons by inoculation with laboratory-cultured microorganisms." In: R. E. Hinchee, J. K. Frederickson, and B. C. Alleman (Eds.), Bioaugmentation for Site Remediation. Battelle Press, Columbus, OH. pp. 129-134. McCaulou D. R., C. B. Bales, and R. G. Arnold. 1995. "Effect of Temperature-Controlled Motility on Transport of Bacteria and Microspheres Through Saturated Sediment." Wat.Resourc.Res. 57:271280. McCaulou D. R., R. C. Bales, J. F. McCarthy, and R. G. Arnold. 1994. "Use of short-pulse experiments to study bacteria transport through porous media." J. Contam.Hydrol. 75:1-14. McCullough, Julie, Hazen, T. C, Benson, Sally M., Melting, F. B., and PalmiSano, Anna C. 1999. "Bioremediation of metals and radionuclides ... what it is and how it works." LBNL 42595. Natural and Accelerated Bioremediation Research Program (NABIR) of the Office of Biological and Environmental Research of the U.S. Department of Energy's Office of Science Mills A. L. 1997. "Movement of Bacteria in the Subsurface." In: P. S. Amy, and D. L. Haldeman (Eds.), The Microbiology o f the Terrestrial Deep Subsurface. CRC Press, Boca Raton, Florida, pp. 225243. 114 Mills A. L , J. S. Herman, G. M. Homberger, and T. H. DeJesus. 1994. "Effect of Solution Ionic Strength and Iron Coatings on Mineral Grains on the Sorption of Bacterial Cells to Quartz Sand." ApplEnvironMicrobiol. <50:3300-3306. Molz F. J., M. A. Widdowson, and L. D. Benefield. 1986. "Simulation of microbial growth dynamics coupled to nutrient and oxygen transport in porous media." Wat. Resourc. Res. 22:1207-1216. Morley L. M., G. M. Homberger, A. L. Mills, and J. S. Herman. 1998. "Effects of transverse mixing on transport of bacteria through heterogeneous porous media." Wat. Resourc. Res. 54:1901-1908. Mueller R. F. 1996. "Bacterial Transport and Colonization in Low Nutrient Environments." Water Res. 30:2681-2690. Muller R. H., W. Babel. 1996. "Measurement of.Growth at Very Low Rates ((mu) >= 0), an Approach To Study the Energy Requirement for the Survival of Alcaligenes eutrophus. IMP 134." Appl.Environ.Microbiol. 62:147-151. National Research Council. 1994. Alternatives fo r Ground Water Cleanup. National Academy Press, Washington, D.C. Nealson K. H., D. Saffarini. 1994. "Iron and Manganese in Anaerobic Respiration: Environmental Significance, Physiology, and Regulation." Annu.Rev.Microbiol. 48:311-343. Nelson C. H., J. A. Robinson, and W. G. Characldis. 1985. "Bacterial adsorption to smooth surfaces: Rate, extent, and spatial pattern." Biotechnol Bioeng. 227:1662-1667. Novak P. J., L. Daniels, and G- F. Parkin. 1998. "Enhanced dechlorination of carbon tetrachloride and ■chloroform in the presence of elemental iron and Methanosarcina barkeri, Methanosarcina thermophila, Methanosaeta concilia." Environ.Sci.Technol. 32:1438-1443. Novitsky J. A., R. Y. Morita. 1978. "Starvation induced barotolerance as a survival mechanism of a psychrophilic marine vibrio in the waters of the Arctic Convergence." Mar.Biol. 49:7-10. Nystrom T., N. Albertson, and S. Kjelleberg. 1988. "Synthesis of membrane and periplasmic proteins during starvation of a marine Vibrio sp." J Gen.Microbiol. 734:1645-1651. Nystrom T., K. Flardh, and S. Kjelleberg. 1990. "Responses to multiple-nutrient starvation in marine Vibrio sp. strain CCUG 15956.”JBacteriol 772:7085-7097. Nystrom T., R. M. Olsson, and S. Kjelleberg. 1992. "Survival, stress resistance, and alterations in protein expression in the marine vibrio sp. strain S14 during starvation for different individual nutrients." Appl.Environ.Microbiol. 58:55-65. O1Hannesin S., R. W. Gillham. 1998. "Long-Term Performance of an In Situ "Iron Wall" for Remediation ofVOCs." Groundwater 36:164-172. O'Toole G. A., R. Kolter. 1998a. "Flagellar and twitching motility are necessary for Pseudomonas aeruginosa biofihn development." Mol. Microbiol. 30:295-304. 115 O'Toole G. A., R. Kolter. 1998b. "Initiation of biofihn formation in Pseudomonas fluorescens WCS365 proceeds via multiple, convergent signalling pathways: a genetic analysis." Mol. Microbiol 25:449-461. Obuekwe C. O., D. W. S. Westlake, J. A. Plambeck, and F. D. Cook. 1981a. "Corrosion of mild steel in cultures of ferric iron reducing bacterium isolated from crude oil - I. Polarization characteristics." . Corrosion 57:461-467. Obuekwe C. O., D. W. S. Westlake, I. A. Plambeck, and F. D. Cook. 1981b. "Corrosion of mild steel in cultures of ferric iron reducing bacterium isolated from crude oil - 2. Mechanisms of anode depolarization." Corrosion 57:632-637. Odziemkowski, M. S. and Gillham, R. W. 1997. Surface redox reactions on commercial grade granular iron (Steel) and their influence on the reductive dechlorination of solvent. Micro Raman spectroscopy studies. Extended Abstracts, 213th National Meeting, American Chemical Society, Division of Environmental Chemistry. 37:pp. 177-180. Odziemkowski M. S., T. T. Schumacher, R. W. Gillham, and E. J. Reardon. 1998. "Mechanism of oxide film formation on iron in simulating groundwater solutions: Raman spectroscopic studies." Corrosion Science 40:371-389. Pagilla K. R., L. W. Canter. 1999. "Laboratory Studies on Remediation of Chromium-Contaminated Soils." J.Environ.Engin. /25:243-248. Palmer C. D., P. R. Wittbrodt. 1991. "Processes Affecting the Remediation of Chromium-Contaminated Sites." Environmental Health Perspectives 92:25-40. Patterson R. R., S. Fendorf, and M. Fendorf. 1997. "Reduction of Hexavalent Chromium by Amorphous Iron Sulfide." Environ.Sci.Technol. 57:2039-2044. Paul I. H., W. H. Jeffrey. 1984. "The Effect of Surfactants on the Attachment of Estuarine and Marine Bacteria to Surfaces." Can.J.Microbiol. 57:224-228. Paul J. H., W. H. Jeffrey. 1985. "Evidence for separate adhesion mechanisms for hydrophilic and hydrophobic surfaces in Vibrio proteolytica." Appl.Environ.Microbiol. 50:431-437. Peterson M. L., G. E. Brown, Jr., G. A. Parks, and C. L. Stein. 1997: "Differential redox and sorption of Cr(IIW I) on natural silicate and oxide minerals: EXAFS and XANES results." Geochimica et CosmochimicaActa 67:3399-3412. Piette J. P., E. S. Idziak. 1991. "Role of flagella in adhesion of Pseudomonas fluorescens to tendon slices." Appl.EnvironMicrobiol. 57:1635-1639. Pine M. J. 1965. "Heterogeneity of protein turnover in Escherichia coli." Biochim.Biophys.Acta 104:439456. Pohland F. G., W. H. Cross, and J. P. Gould. 1993. "Metal Speciation and Mobility as Influenced by Landfill Disposal Practices." In: H. E. Allen, E. M. Perdue, and D. S. Brown (Eds.), Metals in Groundwater. Lewis Publishers, Boca Raton, pp. 419-419. 116 Powell R. M., R. W. Puls, S. K. Hightower, and D. A. Sabatini. 1995a. "Coupled Iron Corrosion and Chromate Reduction: Mechanisms for Subsurface Remediation." Environ.Sci.Technol 29-19131922. Powell R. M., R. W. Puls, S. K. Hightower, and D. A. Sabatini. 1995b. "Coupled Iron Corrosion and Chromate Reduction: Mechanisms for Subsurface Remediation." Environ.Sci.Technol 29-19131922. Pratt A. R., D. W. Blowes, and C. I. Ptacek. 1997. "Products of Chromate Reduction on Proposed Subsurface Remediation Material." Environed.Technol. 37:2492-2498. Rai D., I. M. Zachara. 1988. Chromium reactions in geologic materials. Electric Power Research Institute, Palo Alto, CA. Rajagopalan R., C. Tien. 1976. "Trajectory analysis of deep-bed filtration with the sphere-in-cell porous media model." AIChE Journal 22:523-533. Reeve C. A., P. S. Amy, and A. Matin. 1984. "Role of protein synthesis in the survival of carbon-starved Escherichia coli K-12." J. Bacterial. 760:1041-1046. Revsbech N. P., B. B. Jorgensen. 1986. "Microelectrodes: their use in microbial ecology." Adv.Microb.Ecol. 9:293-352. Reynolds P. J., P. K. Sharma, G. E. Jenneman, and M. I. Mclnemey. 1989. "Mechanisms of microbial movement in subsurface materials." Appl.Environ.Microbiol. 55:2280-2286. Richard F. C., A. C. M. Bourg. 1991. "Aqueous Geochemistry of Chromium: A Review." Wat. Res. 25:807-816. Rijnaarts H. M., W. Norde, E. J. Bouwer, J. Lyklema, and A. I. B. Zehnder. 1993. "Bacterial Adhesion under Static and Dynamic Conditions." Appl.Environ.Microbiol. 59:3255-3265. Rijnaarts H. M., W. Norde, E. I. Bouwer, J. Lyklema, and A. J. B. Zehnder. 1996. "Bacterial Deposition in Porous Media: Effects of Cell-Coating, Substratum Hydrophobicity, and Electrolyte Concentration." Environ.Sci.Technol. 30:2877-2883. Roberts A. L., L. A. Totten, W. A. Arnold, D. R. Burris, and T. J. Campbell. 1996. "Reductive elimination of chlorinated ethenes by zero-valent metals." Environ.Sci.Technol. 30:2654-2659. Roden E. E., I. M. Zachara. 1996. "Microbial Reduction of Crystalline Iron(III) Oxides: Influence of Oxide Surface Area and Potential for Cell Growth." Environ.Sci.Technol. 30:1818-1628. Rosenberg M. 1984. "Bacterial adherence to hydrocarbons: a useful technique for studying cell surface hydrophobicity." FEMSMicrobiol. Let. 22:289-295. Rosenberg M., D. Gutnick, and E. Rosenberg. 1980. "Adherence of bacteria to hydrocarbons: A simple method for measuring cell-surface hydrophobicity." FEMSMicrobiol. Let. 9:29-33. Sacre, Jeffrey A. 1997. "Treatment walls: A status update." TP-97-02. Ground-Water Remediation Technology Analysis Center Pittsburgh, PA. 117 Salamtro J. P., P. C. Johnson, and G. E. Spinnler. 2000. "Field-scale demonstration of enhanced MTBE bioremediation through aquifer bioaugmentation and oxygenation." Environ.Sci.Technol. 34:41524162. Sarkar A. K., G. Georgiou, and M. M. Sharma. 1994. "Transport of Bacteria in Porous Media: I. An Experimental Investigation," Biotechnol.Bioeng. 44:489-497. Sayles G. D., G. You, M. Wang, and M. J. Kupferle. 1997. "DDT, DDD, and DDE dechlorination by zerovalent iron." Environ.Sci.Technol. 37:3448-3454. Schafer H., L. Bernard, C. Courties, P. Lebaron, P. Servais, R. Pukall, E. Stackebrandt, M. Troussellier, T. Guindulain, J. Vives-Rego, and G. Muyzer. 2001. "Microbial community dynamics in Mediterranean nutrient-enriched seawater mesocosms: changes in the genetic diversity of bacterial populations." FEMSMicrobiol Ecol 34:243-253. Scherer M. M., B. A. Balko, and P. G. Tratnyek. 1998. "The role of oxides in reduction reactions at the metal-water interface." In: D. L. Sparks, and T. J. Grundl (Eds.), Mineral-Water Interfacial Reactions - Kinetics and Mechanisms. American Chemical Society, pp. 301-322. Scherer M. M., S. Richter, R. L. Valentine, and P. J. J. Alvarez. 2000. "Chemistry and microbiology of permeable reactive barriers for in situ groundwater clean up." Critical Reviews in Environmental Science and Technology 30:363-411. Scheuermann T. R., A. K. Camper, and M. A. Hamilton. 1998. "Effects of Substratum Topography on Bacterial Adhesion." Journal o f Colloid and Interface Science 208:23-33. Schijven J. F., W. Hoogenboezem, S. M. Hassanizadeh, and J. H. Peters. 1999. "Modeling removal of bacteriophages MS2 and PRDl by dune discharge at Castricum, Netherlands." Wat. Resourc. Res. 35:1101-1111. Schmuki P., M. Buchler, S. Viftanen, H. Bdhni, R. Muller, and L. I. Gauckler. 1995. "Bulk Metal Oxides as a Model for the Electronic Properties of Passive Films." Journal o f the electrochemical society 742:3336-3342. Scholl M. A., R. W. Harvey. 1992. "Laboratory investigations on the role of sediment surface and groundwater chemistry in transport of bacteria through a contaminated sandy aquifer." Environ.Sci.Technol. 26:1410-1417. Scholl M. A., A. L. Mills, J. S. Herman, and G. M. Homberger. 1990. "The influence of mineralogy and solution chemistry on the attachment of bacteria to representative aquifer material." J.Contam.Hydrol. 6:321-336. Scott M. J., F. B. Metting, J. S. Fruchter, and R. E. Wildung. 1998. "Research investments pays off." Soil & Groundwater Cleanup October:6-l3. Seaman I. C., P. M. Bertsch, and L. Schwallie. 1999. "In Situ Cr(VI) Reduction within Coarse-Textured, Oxide-Coated Soil and Aquifer Systems Using Fe(II) Solutions." Environ.Sci.Technol. 33:938944. Sharma M. M., Y. I. Chang, and T. F. Yen. 1985. "Reversible and irreversible surface charge modifications of bacteria for facilitating transport through porous media." Colloids.Surf: 76:193-206. 118 Shanna P. K., M. J. Mclnemey. 1994. "Effect of grain size on bacterial penetration, reproduction and metabolic activity in porous media glass bead chambers." Appl.Environ.Microbiol. 60:1481-1486. Sharp, R. R., Gerlach, R., and Cunningham, A. B. 1999. Bacterial Transport Issues Related to Subsurface Biobarriers. In Leeson, A. L. and Alleman, B. C.: Engineered approaches for in situ bioremediation of chlorinated solvent contamination, Proceedings from the Fifth In Situ and OnSite Bioremediation Symposium, held April 19-22,1999 in San Diego, CA. pp. 211-216. Shaw I. C., B. Bramhill, N. C. Wardlaw, and I. W. Costerton. 1985. "Bacterial Fouling of a Model Core System." Appl.Environ.Microbiol. 49:69.3-701. Siantar D. P., C. G. Schreier, C.-S. Chou, and M. Reinhard. 1996. "Treatment of l,2-dibromo-3chloropropane and nitrate contaminated water with zero-valent iron or hydrogen/palladium catalyst." Wat.Res. 30:2315-2322. Simoni S. F., H. Harms, T. N. Bosma, and A. I. B. Zehnder. 1998. "Population heterogeneity affects transport of bacteria through sand columns at low flow rates." Environ.Sci.Technol. 32:2100-2105. Smith M. S., G. W. Thomas, R. E. White, and D. Ritonga. 1985. "Transport of Escherichia coli through intact and disturbed soil columns." J.Environ.Qual. /4:87-91. Spector M. P. 1998. "The starvation-stress response (SSR) of Salmonella." Adv.Microb Physiol. 40:233279. Spector M. P., d. P. Garcia, S. M. Bearson, A. Mahmud, M. Magut, B. B. Finlay, G. Dougan, J. W. Foster, and M. J. Fallen. 1999. "The ipoS-dependent starvation-stress response locus stiA encodes a nitrate reductase (narZYWV) required for carbon-starvation-inducible thermotolerance and acid tolerance in Salmonella typhimurium." Microbiology 745:3035-3045. Steffan R. J., K. L. Sperry, and M. T. Walsh. 1999. "Field-scale evaluation of in situ bioaugmentation for ' remediation of chlorinated solvents in groundwater." Environ.Sci.Technol. 33:2771-2781. \ Stevenson I. L. 1958. "The effect of sonic vibration on the bacterial plate count of soil." Plant Soil 10:1-18. Svensater G., B. Sjogreen, and I. R. Hamilton. 2000. "Multiple stress responses in Streptococcus mutans and the induction of general and stress-specific proteins." Microbiology 146:107-117. Tabor P. S., K. Ohwada, and R. R. Colwell. 1981. "Filterable marine bacteria found in the deep sea: Distribution, taxonomy, and response to starvation." Microbiol.Ecol. 7:67-83. Tan Y., I. T. Gannon, P. Baveye, and M. Alexander. 1994. "Transport of bacteria in an aquifer sand: Experiments and model simulations." Wat.Resourc.Res. 30:3243-3252. Teich A., S. Meyer, H. Y. Lin, L. Andersson, S. Enfors, and P. Neubauer. 1999. "Growth rate related concentration changes of the starvation response regulators sigmaS and ppGpp in glucose-limited fed-batch and continuous cultures of Escherichia coli." Biotechnol. Prog. 35:123-129. Tien C., A. C. Payatakes. 1976. "Advances in deep-bed filtration." AIChE Journal 25:737-759. Till B. A., L. J. Weathers, and P. J. I. Alvarez. 1998. "Fe(0)-supported autotrophic denitrification." Environ.Sci.Technol. 32:634-639. 119 Tugel J. B., M. E. Hines, and G. E. Jones. 1986. "Microbial iron reduction by enrichment cultures isolated from estuarine sediments." Appl.Environ.Microbiol. 52:1167-1172. Turick C. E., C. Graves, and W. A. ApeL 1998. "Bioremediation potential of Cr(VI)-contaminated soil using indigenous microorganisms." Bioremediation Journal 2:1-6. U.S.EPA 1997. "Permeable reactive subsurface barriers for the interception and remediation of chlorinated hydrocarbon and chromium(VI) plumes in ground water." EPAJ600/F-97/008. National Risk Management Research Laboratory, Office of Research and Development Ada, OK. U.S.EPA 1998. "Permeable Reactive Barrier Technologies for Contaminant Remediation." EPA/600/R98/125. U.S. EPA Washington D.C. U.S.EPA 2000. "Engineered approaches to in situ bioremediation of chlorinated solvents: Fundamentals and field applications." EPA 542-R-00-008. U.S. Environmental Protection Agency, Office of Solid Waste and Emergency Response, Technology Innovation Office Washington D.C. Unice K. E., B. E. Logan. 2000. "Insignificant Role of Hydrodynamic Dispersion on Bacterial Transport." J.Environ.Engin. 726:491-500. Urone P. F. 1955. "Stability of Colorimetric Reagent for Chromium, s-Diphenylcarbazide, in Various Solvents." Analytical Chemistry 1354-1355. Van Genuchten, M. T. and Alves, W. J. 1982. "Analytical Solutions of the One-Dimensional Convective Dispersive Solute Transport Equation." Technical Bulletin 1661. US Department of Agriculture VanLoosdrecht M. C. M., J. Lyklema, W. Norde, G. Schraa, and A. J. B. Zehnder. 1987a. "Electrophoretic mobility and hydrophobicity as a measure to predict the initial steps of bacterial adhesion." Appl.Environ.Microbiol. 53:1898-1901. VanLoosdrecht M. C. M., J. Lyklema, W. Norde, G. Schraa, and A. J. B. Zehnder. 1987b. "The role of bacterial cell wall hydrophobicity in adhesion." Appl.Environ.Microbiol. 53:1893-1897. Vitale R. J., G. R. Mussoline, J. C. Petura, and B. R. James. 1994. "Hexavalent Chromium Extraction from Soils: Evaluation of an Alkaline Digestion Method." J. Environ. Qua!. 23:1249-1256. Vogel T. M. 1996. "Bioaugmentation as a soil bioremediation approach." Curr.Opin.Biotechnol. 7:311316. • Wagner M.„ R. Amann, H. Lemmer, and K. H. Schleifer. 1993. "Probing activated sludge with oligonucleotides specific for proteobacteria: inadequacy of culture-dependent methods for describing microbial community structure." AppLEnvironMicrobiol. 59:1520-1525. Walter M. V. 1997. "Bioaugmentation." In: C. J. Hurst, G. R. Knudsen, M. J. Mclnemey, L. D. Stetzenback, and M. V. Walter (Eds.), Manual o f Environmental Microbiology. ASM Press, Washington D.C. pp. 753-757. Watts R. J. 1997. Hazardous Wastes: Sources. Pathways, Receptors. JohnWiley & Sons, Inc., New York. 120 885. —s:z=LT.====.==acz:c 30:2470-2477. Wood W. W., G. G. Ehrlich. 1978. "USe of bakefs yeas, to .race microbial movement in gmnnd water." Ground Water 16:398-403. tetrachloride." Environ.Sci.Technol. 37:2292-2297. Zhang P.. W. P. Johnson, and R. Rowland. 1999. "Bacterial backing using Petrographic separation." Environ.Sci.Technol. 33:2456-2460. ~ ‘5 ^ T i = M 5 S r — ' , i i ” S X C K iS s r “■“ ■'■“ *■■"— ■ MONTANA STATE UNIVERSITY 762 BOZEMAN 035 942 b