9:: (r vii: .1. r u u r . .5 A g. » . h fin... .. slzzitv‘). r a, 3‘. rCWxIx.S: ‘ . a355,, _ , .45.": .. ,, ‘ , , . . f _ , 1.: Ti—lEElS .5999 llllllllllllllllllilHillllll‘lllllllllllll’lllllllllllllll 3 1293 018 This is to certify that the dissertation entitled Contaminant Biodegradation in the Presence of a Sorbent presented by Denise Patricia Kay has been accepted towards fulfillment of the requirements for Ph. D. degreein Crop & Soil Sciences Major profear Date fl‘iz7" 7 7 MSU i: an Affirmative Action/Equal Opportunity Institution 0- 12771 LIBRARY MIchIgan State University PLACE IN RETURN Box to remove this checkout from your record. To AVOID FINES retum on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 1/” mm.“ CONTAMINANT BIODEGRADATION IN THE PRESENCE OF A SORBENT By Denise Patricia Kay A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PI—HLOSOPHY Department of Crop and Soil Sciences 1 999 ABSTRACT CONTAMINANT BIODEGRADATION IN THE PRESENCE OF A SORBENT By Denise Patricia Kay Complex interactions between a contaminant, the soil solution, organic matter, clay and other soil constituents can alter contaminant availability to bacteria and affect degradation. As a result laboratory experiments which describe the biodegradation of compounds in liquid culture cannot necessarily be extrapolated to the same compound in soil. We used an experimental matrix comprised of two sorbates, three sorbents, and four strains of bacteria to elucidate some of the factors controlling bioavailability of sorbed contaminants. The chosen experimental matrix allowed us to assess the bioavailability of sorbed contaminant in twenty-four unique systems. These systems represented a wide range of sorption and desorption extents, a non-ionic and an anionic contaminant with different mechanisms of sorption, and four different bacterial strains with different adhesion abilities. Bioavailability was assessed by comparing the measured rate of degradation in the presence of a sorbent to the rate predicted by the independently described processes of desorption, and degradation in the solution phase. The results showed that bioavailability varies with sorbent type, organism, and substrate. There were nine combinations in which desorption and solution-phase degradation accurately described the system. There were ten combinations in which the sorbed contaminant may have been directly available. In these systems evidence suggests that bacteria present on the sorbent surface provided the necessary concentration gradient to facilitate desorption into the thin aqueous layer surrounding the cell where uptake proceeded rapidly without diffusion into the bulk solution. There were also five combinations in which the presence of a sorbent inhibited solution phase degradation. The adhesion abilities of each of the four bacterial strains to both hydrophobic and hydrophilic surfaces were determined and related to the bioavailability results. The results showed that the relative ability to adhere to surfaces was independent of surface type, method of measurement, and the presence or absence of low levels of sorbed carbon sources. Availability of contaminant to high-adhesion organisms was affected by the presence of a sorbent, either enhancement or inhibition of bioavailability, more frequently than for low adhesion organisms. Depletion of substrate in the presence of a sorbent was accurately predicted from depletion of solution phase contaminant and desorption in twice as many cases for low adhesion organisms than high adhesion organisms. The tremendous variation in bioavailability in the twenty-four systems studied indicates that caution should be exercised when generalizing any single conclusion regarding the bioavailability of sorbed compounds. Copyright by DENISE PATRICIA KAY 1 999 DEDICATION This body of work is dedicated to my parents, Patricia Ann Kay and Frederick John Kay. Their unwavering love and support have let me believe that anything is possible. My mother once told me a story about myself. She said that when I was very young and she was trying to teach me my ABCs we would stand in the kitchen and she would say all the letters. I would not ask any questions and I would not repeat the series of letters. My mother persisted. After incorporating songs and games and still not hearing even “A..B..C” come out of my mouth she began to become a bit concerned but she didn’t give up. She tells me that one day I finally marched down the stairs from my room, recited all twenty-six letters from A to Z and grinned from ear to ear. In some ways I feel that I have changed very little from those days. For better or worse I seem to have approached the acquisition of my Ph.D. in the same manner that I learned my ABCs. Whether it was insecurities or pride that stopped me from asking many questions, I will never know, but I am proud to have completed the task in my own time and my own way. My alternate title for this thesis is “Bioavailability from A to Z”. Thanks Mom and Dad for never faltering in your belief that I could do it. ACKNOWLEDGMENTS I am grateful to all the members of my committee, both permanent and transient, who guided me through this growth process. Thank you Dr. Stephen Boyd, Dr. Thomas Voice, Dr. James Tiedje, Dr. Terrence Marsh, Dr. Larry Fomey, and Dr. Sharon Anderson for your time attention and patience. Thanks to Dr. Hauke Harms and the members of his laboratory at EAWAG in Switzerland for teaching me the adhesion assay technique and providing numerous unique opportunities for personal growth. I am most indebted to my friends along this journey, especially Dr. Matthew Zwiemik, who have made each load bearable and each obstacle surmountable with their smiles, friendship, and encouragement. vi TABLE OF CONTENTS LIST OF TABLES .................................................................................................... x LIST OF FIGURES ................................................................................................. xii LIST OF SYMBOLS AND ABBREVIATIONS ..................................................... xv CHAPTER 1 GENERAL INTRODUCTION .................................................................................. 1 Literature Cited .............................................................................................. 8 CHAPTER 2 BIOAVAILABILITY OF SORBED CONTAMINANTS ........................................ 13 Abstract ....................................................................................................... 13 Introduction ................................................................................................. 14 Theory ......................................................................................................... 16 Materials and Methods ................................................................................. 20 Organisms ........................................................................................ 20 Stock Solutions ................................................................................ 21 Establishing First-Order Kinetics ..................................................... 21 Silica Slurries ................................................................................... 22 Sorption Kinetics ............................................................................. 23 Desorption Kinetics .......................................................................... 23 Bioavailability Assays ...................................................................... 24 Distribution Coefficients .................................................................. 24 Modeling .......................................................................................... 25 2,4-D Mineralization in the Presence of Non-Sorbing Ottawa Sand.. 26 Results and Discussion ................................................................................ 27 Conclusions ................................................................................................. 33 Acknowledgements ..................................................................................... 34 Literature Cited ............................................................................................ 34 CHAPTER 3 CONSISTENCIES IN BACTERIAL ADHESION MEASUREMENTS AND THEIR RELATIONSHIP TO BIOAVILABILITY .............................................................. 38 Abstract ....................................................................................................... 38 Introduction ................................................................................................. 39 Materials and Methods ................................................................................. 4O Organisms ........................................................................................ 4O Column Adhesion Assay .................................................................. 41 Stock Solutions ................................................................................ 42 Batch Adhesion Assay ..................................................................... 42 Results and Discussion ................................................................................ 44 vii Conclusions ................................................................................................. 49 Acknowledgements ..................................................................................... 50 Literature Cited ............................................................................................ 50 CHAPTER 4 BIOAVAILABILITY RELATED TO MECHANISM OF SORPTION ................... 52 Abstract ....................................................................................................... 52 Introduction ................................................................................................. 53 Materials and Methods ................................................................................. 55 Organisms ........................................................................................ 55 Stock Solutions ................................................................................ 56 Silica Slurries ................................................................................... 56 Sorption Kinetics ............................................................................. 57 Desorption Kinetics .......................................................................... 57 Distribution Coefficients .................................................................. 58 Pore Volume .................................................................................... 59 Results ......................................................................................................... 60 Discussion ................................................................................................... 63 Conclusions ................................................................................................. 70 Acknowledgements ..................................................................................... 71 Literature cited ............................................................................................ 71 CHAPTER 5 SORBENT IMPACT ON SOLUTION pH AND SOLUBLE CARBON AFFECT SOLUTION PHASE DEGRADATION OF 2,4-DICHLOROPHENOXYACETIC ACID ...................................................................................................................... 73 Abstract ....................................................................................................... 73 Introduction ................................................................................................. 74 Materials and Methods ................................................................................. 76 Organisms ........................................................................................ 76 Stock Solutions ................................................................................ 77 Silica Slurries and Solutions ............................................................. 77 Depletion Assays ............................................................................. 78 Modeling .......................................................................................... 79 Results and Discussion ...................................................... 80 Conclusions ................................................................................................. 84 Acknowledgements ..................................................................................... 85 Literature Cited ............................................................................................ 85 APPENDIX LESSONS LEARNED FROM EVALUATION OF BIOAVAILABILITY OF 2,4- DICHLOROPHENOXYACETIC ACID SORBED TO SOIL ................................. 88 Introduction ................................................................................................. 88 Materials and Methods ................................................................................. 89 viii Soil Characterization ........................................................................ 89 Partition Coefficients ....................................................................... 89 Organisms ........................................................................................ 90 Cell Adhesion to Surfaces ................................................................ 91 Biodegradation Kinetics ................................................................... 91 Bioavailability Assays ...................................................................... 93 Results and Discussion ................................................................................ 94 Conclusions ............................................................................................... 104 Acknowledgements ................................................................................... 1 05 Literature Cited .......................................................................................... 105 ix LIST OF TABLES Table 1. The linear sorption coefficient (Kd) and fraction of solute in equilibrium (feq) values for each combination of sorbent and contaminant ........................................................................................ 29 Table 2. Bioavailability of sorbed 2,4-D as indicated by the ratio of the measured rate of degradation in the presence of sorbent to the rate predicted by the Bioavailability Factor Model. ................................... 30 Table 3. Bioavailability of sorbed 2,4-Dme as indicated by the ratio of the measured rate of degradation in the presence of sorbent to the rate predicted by the Bioavailability Factor Model. ................................... 31 Table 4. Bacterial adhesion in the presence and absence of 2,4-D and 2,4- Dme ................................................................................................... 48 Table 5. Relationship between bacterial adhesion and bioavailability of sorbed compounds .............................................................................. 49 Table 6. Kd and feq values for each combination of sorbent and contaminant....62 Table 7. Total pore volume and retained volume for each silica type. ................ 62 Table 8. Kd and Kd* for each combination of sorbent and contaminant ............. 68 Table 9. pH and soluble carbon measurements for PBS and silica supematants ....................................................................................... 80 Table 10. Relative rates of solution phase 2,4-D depletion in silica supernatants ....................................................................................... 81 Table 11. Properties of soils used in this study ................................................... 95 Table 12. Fraction of cells remaining attached to spherical beads in an evenly packed flow through column. ............................................................. 96 Table 13. Specific initial mineralization rate of 2,4-D for each strain and soil type combination. ............................................................................... 97 xi Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. LIST OF FIGURES First model for bioavailability of sorbed compounds. ...................................... 2 Second model for bioavailability of sorbed compounds. .................................. 3 Third model for bioavailability of sorbed compounds. ..................................... 3 Fourth model for bioavailability of sorbed compounds. ................................... 3 The contribution of desorption to the solution phase substrate concentration during substrate depletion in the presence of a sorbent can be predicted using the bioavailability factor (Bf). C/Co represents the substrate solution concentration at any time normalized to the initial concentration. 18 Calculation, from a desorption isotherm, of the fraction of sorbed substrate that is in desorption equilibrium. Fractions of sorbed substrate in equilibrium (feq), non-equilibrium (fneq) and irreversibly bound (firr) are based on the initial sorbed concentration (S0) and the equilibrium sorbed concentration (Seq) predicted from the distribution coefficient with the assumption of complete reversibility. Distinction between equilibrium and non-equilibrium portions was made at the point where there was a distinct observable change in the desorption rate. This occurred in less than 1 min for all systems. ........................................................................... 19 Figure 7. Initial mineralization rates of 2,4-D in the presence (0) and absence (A) of non-sorbing quartz sand. Figure 7A shows data for bacterial strain FB-4 and is representative of strains RASC and TFD6. The presence of non- sorbing sand does not effect the rate of mineralization of solution phase 2,4-D for these three strains. Figure 7B shows the data for bacterial strain JMP134. The presence of non-sorbing sand may decrease the rate of mineralization of solution phase 2,4-D for JMP134. ..................................... 32 Figure 8. The abilities of four strains to attach to Teflon (A) and glass (B) surfaces were measured by determining the fraction of cells that flowed through a column packed with spherical beads. Reversibility of adhesion was ascertained by changing the column influent from the cell suspension to a xii cell-free buffer and then to distilled water. All data points represent the mean of three measurements. Error bars represent one standard deviation. .. 45 Figure 9. Percent adhesion of four bacterial strains as determined by two methods on Figure 10. Figure 11. Figure 12. Figure 13. Figure 14. Figure 15. Figure 16. six surfaces. Adhesion measurement systems were (method/surface) A- filter/control B-filter/uncoated silica C-filter/ethyl coated silica D- filter/phenyl coated silica E-column/Teflon F-coltunn/glass. Each bar represents the mean of three measurements. Error bars represent one standard deviation. ....................................................................................... 46 2,4-Dme desorption from uncoated silica. Desorption for this sorbate- sorbent combination is representative of all combinations used in this study. ........................................................................................................... 61 Frequency of enhancement (white bar) and inhibition (black bar) of contaminant degradation rate relative to rate predicted by BFM using Kd“ .. 69 Normalized rates of solution phase 2,4-D depletion by four bacterial strains at pH 7.0, 6.0, and 5.0. Rates were normalized to the rate measured at pH 7.0 to allow comparison between strains. ............................................ 82 Enhancement of solution phase 2,4-D depletion rate is demonstrated by a value greater than one for the ratio of the measured rate of depletion to the rate of depletion predicted by the Bioavailability Factor Model. This ratio is shown relative to sorbed 2,4-D concentration for two experimental designs ......................................................................................................... 84 Sorption isotherms for 2,4—D on five soil types. ........................................... 95 Soil slurry mineralization rates. The 2,4-D mineralization rate measured for each soil slurry was normalized to the 2,4-D mineralization rate predicted for the given aqueous concentration of 2,4-D in that slurry. The normalized initial mineralization rate has a value greater than one if the measured mineralization rate in the soil slurry exceeded the rate predicted from the aqueous concentration. Mean values of triplicate normalized initial mineralization rates are represented in the bar graph. Hashed bars represent values less than one ....................................................................... 98 Soil-slurry supernatant mineralization rates. The 2,4-D mineralization rate measured for each supernatant was normalized to the 2,4-D xiii mineralization rate predicted for the given aqueous concentration of 2,4-D in that slurry. Hashed bars represent values less than one. ......................... 100 Figure 17. The effect of soil slurry minus the effect of soil-slurry supernatant on 2,4- D mineralization rates. Hashed bars represent values less than one ............ 102 xiv 2,4-D 2,4-Dme MSM PBS LIST OF SYMBOLS AND ABBREVIATIONS 2,4 Dichlorophenoxyacetic acid 2,4-Dichlorophenoxyacetic acid methyl ester Minimal salts media Phosphate buffered saline Distribution coefficient (ml g'l) Fraction of sorbed compound in equilibrium with solution phase Fraction of sorbed substrate in non-equilibrium Fraction of sorbed substrate irreversibly bound for condition and time scale of these studies Bioavailability Factor Model Bioavailability factor Substrate solution concentration (ug ml'l) Substrate solution concentration at time 0 (ug ml'l) Equilibrium sorbed concentration predicted from the distribution coefficient with the assumption of complete reversibility Initial sorbed concentration Half-velocity concentration First —order degradation rate coefficient Ratio of sorbent to water XV Chapter 1 GENERAL INTRODUCTION In the soil environment, the concentration of contaminant in the bulk solution can be reduced by sorption of the contaminant to soil and by movement of the contaminant into micropores, potentially limiting the bioavailability, that is, the availability of these compounds for biological degradation. The sorption of poorly water-soluble organic compounds in soils occurs predominately by partitioning into the soil organic carbon phase when the organic carbon content is above 0.1%, hence the extent of sorption is directly proportional to the soil organic carbon content (Chiou, et a1. 1979, Chiou, et a1. 1983). The sorption of water-soluble or ionic organic compounds is controlled to a much lesser extent by partitioning into soil organic matter. Sorption of these compounds is more likely due to adsorptive surface reactions and accumulation in micropores. The availability of these sorbed organic compounds to bacteria depends in part on desorption of the compound into the bulk solution. Desorption has been described by several kinetic models including the two-box and radial-diffusion models (Karickhoff 1980, Van Genuchten and Wierenga 1976, Wu and Gschwend 1986). In the two-box model, desorption is conceptualized as occurring by rapid release from an easily accessible exterior followed by a kinetically limited desorption from a less accessible interior. The latter process may be conceptualized as intraorganic-matter diffusion, and contaminants within this domain would be biologically unavailable. According to the radial diffusion model, contaminants diffuse through pore fluids in the interstices of soil particles, and this process is retarded by microscale partitioning between fluids and intra- 1 particle solids. Organic compounds dissolved in the soil solution may occupy soil micropores smaller than one micrometer which, on average, constitute 50% of the soil pore space (Hassink, et a1. 1993). Under such conditions bioavailability is dependent on diffusion out of the micropores because size exclusion limits bacteria from directly accessing substrate in these pores (Rijnaarts, et al. 1990, Scow and Hutson 1992). Voice and Xhao (1997) developed four conceptual models to describe the possible combinations of mechanisms controlling the bioavailability of sorbed compounds. The simplest model assumes that only dissolved compounds are available to bacteria for degradation. The availability of sorbed compounds is therefore limited by the rate of desorption (Figure 1). sorbed dis olved P cells Figure 1. First model for bioavailability of sorbed compounds. The second model increases the complexity by integrating the effects bacteria may have on desorption by processes such as surfactant production or physical occlusion of the sorbent surface (Figure 2). sorbed enhance/suppress dissolved ycells Figure 2. Second model for bioavailability of sorbed compounds. The third model includes the possibility that bacteria have direct access to sorbed phase compounds although the cells do not affect the rate of desorption. This system may not be limited by desorption or mass transfer (Figure 3). sorbed dissolved >cells Figure 3. Third model for bioavailability of sorbed compounds. Finally the fourth and most complex model combines the features of the second and third models (Figure 4). sorbed dissolved b cells Figure 4. Fourth model for bioavailability of sorbed compounds. Reports found in the literature do not allow generalized conclusions about the bioavailability of sorbed compounds. Numerous examples demonstrate reduced bioavailability when a contaminant sorbs to soil (Al Bashir, et al. 1994, Mihelcic and Luthy 1991, Moyer, et a1. 1972, Ogram, et a1. 1985, Radosevich, et a1. 1997, Rijnaarts, et a1. 1990, Scow and Alexander 1992, Scribner, et al. 1992, Shelton and Doherty 1997, Shimp and Young 1988, Smith, et a1. 1992, Steen, et a1. 1980, Weber and Coble 1968, Weissenfels, et al. 1992). Much of this research has indicated that sorbed contaminants are inaccessible to bacteria, and that desorption into bulk solution is a prerequisite for biodegradation (Apajalahti and Salkinoja-Salonen 1984, Ogram, et a1. 1985, Robinson, et a1. 1990, Shelton and Parkin 1991). However experiments that suggest direct access to sorbed contaminants by bacteria have also been reported (Amador and Alexander 1988, Calvillo and Alexander 1996, Crocker, et al. 1995, Efroymson and Alexander 1991, Efroymson and Alexander 1994, Griffith and Fletcher 1991, Guerin and Boyd 1992, Guerin and Boyd 1993, Guerin and Boyd 1997, Harms and Zehnder 1995, Marshman and Marshall 1981, Ortega-Calvo, et al. 1997, Ortega-Calvo and Saiz-Jimenez 1998, Tang, et a1. 1998). This research has suggested that the accumulation of a contaminant at the soil- water interface in concentrations higher than in the bulk solution may enhance the availability to microorganisms that are adhered to the interface. Using sorbents of different dimensions, which manifested different desorption kinetics, Crocker et al. (1995) evaluated the effect of contaminant desorption rate on the bioavailability of naphthalene. They found that for the soil isolate Alcaligenes sp. strain NP-Alk, sorbed naphthalene was available only upon its desorption into solution. Desorption was very rapid from clay aggregates of < 0.25 mm, and hence the pool of sorbed naphthalene was readily available for biodegradation. For naphthalene sorbed to larger clay aggregates, desorption was slow and naphthalene mineralization was desorption rate limited. Ogram et a1. (1985) evaluated the bioavailability of soil-sorbed 2,4-D by fitting empirical data to mathematical models and concluded that sorbed 2,4-D was completely protected from biological degradation by the bacterial species F lavobacterium sp. strain FB—4, but that solution phase 2,4-D was available to both free and attached organisms. The systems described by Crocker et al. and Ogram et a1. illustrate the first model of bioavailability as described above. Harms and Zehnder (1995) showed that bacteria attached to granule surfaces could be efficiently supplied with substrate by promoting desorption from the granule via substrate consumption. Guerin and Boyd (1993) concluded that there was differential bioavailability of soil-sorbed naphthalene to two bacterial species. One organism, Pseudomonas putida strain 17484, appeared to have direct access to sorbed naphthalene while another organism, NP-Alk, could only access aqueous phase naphthalene. In this concept it is not only the sorption of the contaminant to the soil which influences the bioavailability, but also the adhesion of microorganisms with the capability to degrade the contaminant in areas of localized high concentrations. In further studies Guerin and Boyd (1997) also concluded that Pseudomonas putida strain 17484 had direct access to naphthalene which was sorbed to a variety of different soil types and synthetic sorbents. The more complex models for bioavailability of sorbed compounds are required to represent the systems described by Guerin and Boyd, and Harms and Zhender although the studies required assumptions that weaken our ability to differentiate between the models. Given the complexity of the dynamic interactions that are possible between contaminant, sorbent and bacteria it is likely that the best overall descriptive model for the bioavailability of sorbed compounds in the natural environment will include multiple mechanisms as in the fourth model presented earlier. In any particular system one or more processes such as slow desorption, or biosurfactant production may dominate the interactions. The possible combinations of conditions and processes that can result in enhanced, inhibited or absence of bioavailability of solution and sorbed phase contaminants is endless and seems to bear out the wide array of results found in the literature. This does not mean that it is hopeless to try to understand the bioavailability of sorbed compounds. However, because it is a highly dynamic system a single generalization that sorbed contaminants either are or are not biologically available can not be supported. It is more useful to focus on particular conditions or processes that dominate the biological availability of sorbed compounds. We have chosen 2,4-dichlorophenoxyacetic acid (2,4-D) and its methyl ester (2,4- Dme) as model compounds to study the complex relationships that control the bioavailability of sorbed compounds. As Guerin and Boyd stated (1992), to develop even the simplest model for bioavailability of sorbed compounds on theoretical grounds would require apriori knowledge of unique physical and biological parameters including mass transfer kinetics, aggregate size, bacterial attachment to the sorbent and the proportion of sorbed solute residing in the dissolved and labile and non-labile sorbed phases. The experiments described in this dissertation take advantage of a model system that makes independent determination of these unique properties possible. The sorbents used in the model system are silica particles of relatively uniform dimensions and known surface properties. Selected hydrophobic surface coatings provided a range of sorption extent, desorption extent, and desorption kinetics. Four strains which can use 2,4- dichlorophenoxyacetic acid or its methyl ester, the model contaminants, as sole carbon and energy sources have been selected. The adhesion abilities of each of these strains to both hydrophobic and hydrophilic surfaces were determined. This model system was used to independently determine the rate and extent of desorption, the rate of degradation of solution phase contaminant, and other parameters critical for defining the system functions. A modified version of the Bioavailability Factor Model (Zhang, et a1. 1998) was used to determine if these independent processes could describe degradation of the contaminant in the presence of a sorbent. There were three main objectives for this dissertation work: 0 To determine if biodegradation of a contaminant in the presence of a sorbent can be modeled by the sum of the independently described processes of desorption from the sorbent and degradation of solution phase contaminant. 0 To determine if bacterial adhesion to the sorbent surface influences bioavailability of the sorbed contaminant. 0 To determine how bioavailability of sorbed contaminants is related to the contaminant sorption mechanism. Understanding the conditions that control bioavailability of soil sorbed organic compounds is important for both environmental risk assessment and application of bioremediation strategies. There is a growing trend in risk assessment towards the concept that potentially toxic contaminants in the soil can be considered nonhazardous if they are shown to be unavailable to biosphere organisms (Chaney, et al. 1996, O‘Connor, et al. 1991). In addition, successful application of bioremediation strategies relies on a clear understanding of the factors that control bioavailability of soil sorbed contaminants. In fact it has been shown in many cases that bioremediation is not limited by the ability of microorganisms to degrade the contaminating compound but by the ability of the microorganisms to access the contaminant (Al Bashir, et al. 1994, Hatzinger and Alexander 1995, Hatzinger and Alexander 1997, Robinson, et al. 1990, Scow and Hutson 1992). Complex interactions between the contaminant, the soil solution, organic matter, clay and other soil constituents can alter contaminant bioavailability. As a result laboratory experiments which describe the biodegradation of compounds in liquid culture cannot necessarily be extrapolated to the same compound in soil. Hence, a better understanding of the factors which control bioavailability of soil bound contaminants will help toxicologists make critical decisions regarding remediation standards, and result in more successful and less costly bioremediation applications (Cowan, et al. 1995). Literature Cited Al Bashir, B., J. Hawari, R. Samson, and R. Leduc. 1994. Behavior of nitrogen- substituted naphthalenes in flooded soil: Part II. Effect of bioavailability on biodegradation kinetics. Water Res. 28(8): 1827-1833. Amador, J. A., and M. Alexander. 1988. Effect of humic acids on the mineralization of low concentrations of organic compounds. Soil Biol. Biochem. 20(2): 185-191. Apajalahti, J. H. A., and M. S. Salkinoja-Salonen. 1984. Absorption of pentachlorophenol (PCP) by bark chips and its role in microbial PCP degradation. Microb. Ecol. 10:359-367. Calvillo, Y. M., and M. Alexander. 1996. Mechanisms of microbial utilization of biphenyl sorbed to polyacrylic beads. Appl. Microbiol. Biotechnol. 45(3):383-390. 8 Chaney, R. L., J. A. Ryan, and G. A. O'Connor. 1996. Organic contaminants in municipal biosolids: risk assessment, quantitative pathways analysis, and current research priorities. The Science of the Total Environment. 185:187-216. Chiou, C. T., L. J. Peters, and V. H. Freed. 1979. A physical concept of soil-water equilibria for nonionic organic compounds. Science. 206:831-832. Chiou, C. T., P. E. Porter, and D. W. Schmedding. 1983. Partition equilibria of nonionic organic compounds between soil organic matter and water. Environ. Sci. Technol. 17:227-231. Cowan, C. E., D. J. Versteeg, R. J. Larson, and P. J. Kloepper-Sams. 1995. Integrated approach for environmental assessment of new and existing substances. Regulat. Toxicol. Pharmacol. 21:3--1. Crocker, F. H., W. F. Guerin, and S. A. Boyd. 1995. Bioavailability of naphthalene sorbed to cationic surfactant-modified smectite clay. Environ. Sci. Technol. 29(12):2953- 2958. Efroymson, R. A., and M. Alexander. 1991. Biodegradation by an Arthrobacter spp. of hydrocarbons partitioned into an organic solvent. Appl. Environ. Microbiol. 57(5): 1441- 1447. Efroymson, R. A., and M. Alexander. 1994. Role of partitioning in biodegradation of phenantlrrene dissolved in nonaqueous-phase liquids. Environ. Sci. Technol. 28:1172- 1179. Griffith, P. C., and M. Fletcher. 1991. Hydrolysis of protein and model dipeptide substrates by attached and nonattached marine Pseudomonas sp. strain NCIMB 2021. Appl. Environ. Microbiol. 57(8):2186-2191 . Guerin, W. F., and S. A. Boyd. 1992. Differential bioavailability of soil-sorbed naphthalene to two bacterial species. Appl. Environ. Microbiol. 58(4):]142-1152. Guerin, W. F., and S. A. Boyd. 1993. Bioavailability of sorbed naphthalene to bacteria: Influence of contaminant aging and soil organic carbon content. SSSA Special Publication(3 2): 197-208. Guerin, W. F., and S. A. Boyd. 1997. Bioavailability of naphthalene associated with natural and synthetic sorbents. Water Res. 31(6): 1504-1512. Harms, H., and A. J. B. Zehnder. 1995. Bioavailability of sorbed 3-chlorodibenzofuran. Appl. Environ. Microbiol. 61(1):27-33. Hassink, J., L. A. Bouwman, K. B. Zwart, and L. Brussaard. 1993. Relationships between habitable pore space, soil biota and mineralization rates in grassland soils. Soil Biol. Biochem. 25:47-55. Hatzinger, P. B., and M. Alexander. 1995. Effect of aging of chemicals in soil on their biodegradability and extractability. Environ. Sci. Technol. 29:537-545. Hatzinger, P. B., and M. Alexander. 1997. Biodegradation of organic compounds sequestered in organic solids or in nanopores within silica particles. Environ. Toxicol. Chem. 16:2215-2221. Karickhoff, S. W. 1980. Sorption kinetics of hydrophobic pollutants in natural sediments, p. 193-205. In R. A. Baker (ed.), Contaminants and Sediments, vol. 2. Ann Arbor Press, Ann Arbor, MI. Marshman, N. A., and K. C. Marshall. 1981. Bacterial growth on proteins in the presence of clay minerals. Soil Biol. Biochem. 13:127-134. Mihelcic, J. R., and R. G. Luthy. 1991. Sorption and microbial degradation of naphthalene in soil-water suspensions under denitrification conditions. Environ. Sci. Technol. 25(1):169-177. Moyer, J. R., R. J. Hance, and C. E. McKone. 1972. The effect of adsorbents on the rate of degradation of herbicides incubated with soil. Soil Biol. Biochem. 4:307-311. O'Connor, G. A., R. L. Chaney, and J. A. Ryan. 1991. Bioavailability to plants of sludge-bome toxic organics. Rev. Environ. Contarn. Toxicol. 121: 129-155. Ogram, A. V., R. E. Jessup, and P. S. C. Rao. 1985. Effects of sorption on biological degradation rates of (2,4-dichlorophenoxy)acetic acid in soils. Appl. Environ. Microbiol. 49:582-587. Ortega-Calvo, J. J., M. Lahlow, and C. Saiz-Jimenez. 1997. Effect of organic matter and clays on the biodegradation of phenanthrene in soils. Int. Biodeterior. Biodegrad. 40(2-4): 101-106. Ortega-Calvo, J. J., and C. Saiz-Jimenez. 1998. Effect of humic fractions and clay on biodegradation of phenanthrene by a Pseudomonasfluorescens strain isolated from soil. Appl. Environ. Microbiol. 64(8):3123-3126. Radosevich, M., S. J. Traina, and O. H. Tuovinen. 1997. Atrazine mineralization in laboratory-aged soil microcosms inoculated with s-triazine-degrading bacteria. J. Environ. Qual. 26(1):206-214. Rijnaarts, H. H. M., A. Bachmann, J. C. Jumelet, and A. J. B. Zehnder. 1990. Effect of desorption and intraparticle mass transfer on the aerobic biomineralization of alpha- 10 hexachlorocyclohexane in a contaminated calcareous soil. Environ. Sci. Technol. 24(9):]349-1354. Robinson, K. G., W. S. Farmer, and J. T. Novak. 1990. Availability of sorbed toluene in soils for biodegradation by acclimated bacteria. Water Res. 24(3):345-350. Scow, K. M., and M. Alexander. 1992. Effect of diffusion on the kinetics of biodegradation: experimental results with synthetic aggregates. Soil Sci. Soc. Am. J. 56:128-134. Scow, K. M., and J. Hutson. 1992. Effect of diffusion and sorption on the kinetics of biodegradation: Theoretical considerations. Soil Sci. Soc. Am. J. 56(1):]19-127. Scribner, S. L., T. R. Benzing, S. Sun, and S. A. Boyd. 1992. Desorption and bioavailability of aged simazine residues in soil from a continuous corn field. J. Environ. Qual. 21(1):115-120. Shelton, D. R., and M. A. Doherty. 1997. A model describing pesticide bioavailability and biodegradation in soil. Soil Sci. Soc. Am. J. 61:1078-1084. Shelton, D. R., and T. B. Parkin. 1991. Effect of moisture on sorption and biodegradation of carbofuran in soil. J. Agric. Food Chem. 39:2063-2068. Shimp, R. J., and R. L. Young. 1988. Availability of organic chemicals for biodegradation in settled bottom sediments. Ecotoxicol. Environ. Saf. 15:31-45. Smith, S. C., C. C. Ainsworth, S. J. Traina, and R. J. Hicks. 1992. Effect of sorption on the biodegradation of quinoline. Soil Sci. Soc. Am. J. 56(3):737-746. Steen, W. C., D. F. Paris, and G. L. Baughman. 1980. Effect of sediment sorption on microbial degradation of toxic substances, vol. 1. Ann Arbor Sci. Publ., Ann Arbor, MI. Tang, W. C., J. C. White, and M. Alexander. 1998. Utilization of sorbed compounds by microorganisms specifically isolated for that purpose. Appl. Microbiol. Biotechnol. 49:117-121. Van Genuchten, M. T., and P. J. Wierenga. 1976. Mass transfer studies in sorbing porous medial. Analytical solutions. Soil Sci. Soc. Am. J. 40(4):473-479. Voice, T. C., X. Zhao, and M. A. Maraqa. 1997. Presented at the 214th American Chemical Society National Meeting, Las Vegas, Nevada. Weber, J. B., and H. D. Cable. 1968. Microbial decomposition of diquat adsorbed on montrnorillonite and kaolinite clays. J. Agric. Food Chem. l6(3):475-478. ll Weissenfels, W. D., H. J. Klewer, and J. Langhoff. 1992. Adsorption of polycyclic aromatic hydrocarbons (PAHs) by soil particles: Influence on biodegradability and biotoxicity. Appl. Microbiol. Biotechnol. 36(5):689-696. Wu, S. C., and P. M. Gschwend. 1986. Sorption kinetics of hydrophobic organic compounds to natural sediments and soils. Environ. Sci. Technol. 20:717-725. Zhang, W., E. J. Bouwer, and W. P. Ball. 1998. Bioavailability of hydrophobic organic contaminants: Effects and implications of sorption-related mass transfer on bioremediation. Ground Wat. Monitor. Remed. 18(1):126-128. 12 Chapter 2 BIOAVAILABILITY OF SORBED CONTAMINANTS Abstract A better understanding of the factors that control the bioavailability of soil bound contaminants will result in more effective remediation decisions. We evaluated the bioavailability of sorbed contaminant for twenty-four unique combinations of bacterial species, sorbent, and contaminant. This was done by comparing the measured rate of degradation in the presence of a sorbent to the rate predicted by the independently described processes of desorption, and degradation in the solution phase. The results showed that bioavailability varies with sorbent type, organism, and substrate. There were nine combinations in which desorption and solution-phase degradation accurately described the system. There were ten combinations in which the sorbed contaminant may have been directly available. In some of these cases the rates of depletion from solution were nearly three times the rate predicted from desorption and degradation of only solution phase contaminant. In these systems evidence suggests that bacteria present on the sorbent surface could have provided the necessary concentration gradient to facilitate desorption from the micro-pores into the thin aqueous layer surrounding the cell where uptake proceeded rapidly without diffusion into the bulk solution. There were also five combinations in which the presence of a sorbent significantly inhibited solution phase degradation. The tremendous variation in bioavailability in the twenty-four systems 13 studied here indicates that caution should be exercised when generalizing any single conclusion regarding the bioavailability of sorbed compounds. Introduction Successful application of bioremediation strategies relies on a clear understanding of the factors that control bioavailability of soil-sorbed contaminants. Complex interactions between the contaminant, the soil solution, organic matter, clay and other soil constituents can alter contaminant bioavailability. As a result caution should be exercised when extrapolating compound biodegradation in liquid culture to in-situ bioremediation. In fact it has been shown in many cases that bioremediation is not limited by the ability of microorganisms to degrade the contaminating compound but by the ability of the microorganisms to access the contaminant (A1 Bashir, et al. 1994, Hatzinger and Alexander 1995, Hatzinger and Alexander 1997, Robinson, et al. 1990, Scow and Hutson 1992). In-situ bioremediation may be subject to kinetic limitations similar to conventional pump and treat processes for the cleanup of contaminated soils if the rate of degradation of sorbed contaminants is desorption limited. In contrast, if organisms adhered to particle surfaces have access to sorbed contaminants in-situ bioremediation would potentially be more efficient than conventional remediation methods. Hence, a better understanding of the factors which control bioavailability of soil bound contaminants will result in more effective remediation decisions (Cowan, et al. 1995) Numerous examples found in literature suggest reduced bioavailability when a contaminant is sorbed (Al Bashir, et al. 1994, Mihelcic and Luthy 1991, Moyer, et al. 1972, Ogram, et al. 1985, Radosevich, et al. 1997, Rijnaarts, et al. 1990, Scow and 14 Alexander 1992, Scribner, et al. 1992, Shelton and Doherty 1997, Shimp and Young 1988, Smith, et al. 1992, Steen, et al. 1980, Weber and Coble 1968, Weissenfels, et al. 1992). Ogram et a1. (1985) and Shelton and Doherty (1997) concluded that soil sorbed 2,4-D was completely protected from biological degradation based on the fit of empirical data to mathematical models. Rijnaarts et al. (1990), and Mihelcic and Luthy (1991) found that solute diffusion through micropores limited biodegradation of organic chemicals. Al Bashir et al. (1994) showed that naphthalene mineralization in soil Slurries was directly pr0portional to substrate aqueous concentrations. Experiments which suggest direct access to sorbed contaminants by bacteria have also been reported (Amador and Alexander 1988, Calvillo and Alexander 1996, Crocker, et al. 1995, Efroymson and Alexander 1991, Efroymson and Alexander 1994, Griffith and Fletcher 1991, Guerin and Boyd 1992, Guerin and Boyd 1993, Guerin and Boyd 1997, Harms and Zehnder 1995, Marshman and Marshall 1981, Ortega-Calvo, et al. 1997, Ortega-Calvo and Saiz-Jimenez 1998, Tang, et al. 1998). Guerin and Boyd (1993), using kinetic mineralization assays, concluded that there was differential bioavailability of soil-sorbed naphthalene to two bacterial species. One organism, Pseudomonas putida strain 17484, appeared to have direct access to the pool of sorbed naphthalene while another organism, NP-Alk, could only access aqueous phase naphthalene. In further studies Guerin and Boyd (1997) and Crocker et al. (1995) concluded that Pseudomonas putida strain 17484 had direct access to naphthalene which was sorbed to a variety of different sorbents. Efroymson and Alexander (1994) demonstrated that the rate of mineralization exceeded the rate of partitioning of phenanthrene from NAPL to the water 15 phase. Tang et al. (1998) made the same conclusion regarding the mineralization of phenanthrene sorbed to polyacrylic beads. The experiments described herein were designed to determine the bioavailability of sorbed contaminants by directly and independently accounting for the rates and extents of both sorption and desorption in relation to the biodegradation of sorbed contaminants. A modified version of the Bioavailability Factor Model (Zhang, et al. 1998) was used to elucidate the conditions governing bioavailability in our model systems. The results showed ten combinations in which the sorbed contaminant may have been directly available. Nine combinations in which desorption and solution phase degradation accurately described the system. Five combinations in which the presence of the sorbent apparently inhibited solution phase degradation. Theory In liquid-batch culture first-order substrate depletion kinetics are observed at substrate concentrations below KS (Alexander and Scow 1989, Guerin and Boyd 1993). Rate of depletion is directly proportional to the bioavailable concentration under first- order conditions. Here, kb is the first-order biodegradation rate coefficient: -dC _ k C (It b (1) When a sorbent is present and the rate of desorption is fast relative to the rate of degradation, desorption will make a significant contribution to the solution phase substrate concentration. The impact of fast desorption on the solution phase substrate 16 concentration can be predicted using the Bioavailability Factor Model (Zhang, et, al. 1998) (Figure 5). —dC The bioavailability factor, Bf, is used to account for the impact of desorption on the solution phase substrate concentration and is defined as follows when sorption/desorption processes are fast relative to degradation rates: 3,: 1 1+KdR s/w (3) Here, Kd is the substrate distribution coefficient and Rs/w is the ratio of sorbent to water (wt. / vol.) in the batch system. Bf, as described by Zhang, assumes complete reversibility of sorbed substrate (Figure 5). l7 1.0 3 0.8 3 . \\ o 0.6“ . o - \ i ' Sorbent present 0 0.4 i l \ 0.2 1‘ s i Sorbent absent I 00 i — A —— —- — _— fl ___m __nff,'“_;_7 —___f ___T_ __1_‘ 0 2 4 6 8 10 Figure 5. The contribution of desorption to the solution phase substrate concentration during substrate depletion in the presence of a sorbent can be predicted using the bioavailability factor (Bf). C/Co represents the substrate solution concentration at any time normalized to the initial concentration. The assumption regarding complete sorption reversibility did not always hold true. In the present study there were no cases where the rate of desorption was slow relative to the rate of depletion from solution. However, most of the experimental systems had a fraction of sorbed contaminant that was irreversibly sorbed for the time scale and conditions of our study. Bf can be modified by the fraction of the sorbed contaminant that is in desorption equilibrium, (feq), to account for these cases: 1 Bf = 1"l'.’::ql(r1R.r/w (4) 18 The sorption distribution coefficient and desorption isotherm are used to experimentally determine feq (Figure 6). ,3 So ----------- - — --------- 1"- E f“ 8 '8 I § .--:_¥ ) fneq m 3 firr seq--- _ Time Figure 6. Calculation, from an example desorption isotherm, of the fraction of sorbed substrate that is in desorption equilibrium. Fractions of sorbed substrate in equilibrium (feq), non-equilibrium (fneq) and irreversibly bound (fir-r) are based on the initial sorbed concentration (So) and the equilibrium sorbed concentration (Seq) predicted from the distribution coefficient with the assumption of complete reversibility. Distinction between equilibrium and non-equilibrium portions was made at the point where there was a distinct observable change in the desorption rate. This occurred in less than 1 min for all systems. The Bioavailability Factor Model as modified and presented here requires a priori knowledge of the first-order biodegradation rate coefficient for solution phase substrate, sorption extent, desorption rate and desorption extent. When these factors are known the rate of substrate depletion from solution in the presence of a sorbent can be predicted with the assumption that only solution-phase substrate is available for degradation. It is then possible to test the assumption that only solution phase substrate is available for degradation by comparing the rate of substrate depletion from solution as predicted by the Bioavailability Factor Model to the actual rate measured in a model system. 19 The working hypothesis tested in this study was that degradation of sorbed contaminant is limited by the rate at which it desorbs into solution by mass action, i.e. that only solution phase contaminant is available for biodegradation. The hypothesis was supported if biodegradation in the presence of a sorbent was successfully modeled by the sum of the independently described processes of desorption, and degradation of solution phase contaminant as determined by the Bioavailability Factor Model. The hypothesis was rejected if the rate of biodegradation in the presence of a sorbent was different than the rate predicted by combined desorption and degradation. Sum of squares reduction test was used to determine if there was a significant difference between the two rates. Materials and Methods Organisms Four strains of proteobacteria that are able to use 2,4-dichlorophenoxyacetic acid (2,4-D) or 2,4-dichlorophenoxyacetic acid methyl ester (2,4-Dme) as a sole carbon and energy source were selected for this study. Two of the strains, designated RASC and TFD6, were classified by McGowan (1994) based on 16S rDNA sequence analysis as being Burkholderia sp. Don and Pemberton (1981) previously classified the third strain, JMP134, as Alcaligenes eutrophus. The fourth strain, FB4, was classified as F lavobacterium sp. by Ogram et al. (1985). Each of these organisms internally metabolizes 2,4-D. Cells were grown until early stationary phase in a minimal salts media ([MSM] 1419.6 mg NazHPO4, 1360.9 mg KH2P04, 0.3 mg (NH4)ZSO4, 50 mg MgSO4 7H20, 5.88 mg CaC12 2H20, 3.2 mg EDTA disodium salt, 2.78 mg FeSO4 71120, 1.15 mg ZnSO4 7H20, 1.69 mg MnSO4 H20, 0.375 mg CuSO4 51120, 0.233 mg 20 Co(NO3)2 6H20 and 0.1236 mg (NH4)6M07024 4H20 per liter of distilled water) that contained 400 mg L'1 2,4-D. Cells to be used as inoculum were rinsed once and resuspended in phosphate buffered saline ([PBS] 8.5 g NaCl, 0.6 g Nag P04, 0.3 g KH2PO4, per liter of distilled water, pH 7.0). Stock Solutions Sterile stock solutions of 14C-labeled 2,4-D (Sigma, 296 pure, 12.8 mCi mmol’l) and 2,4-Dme (Sigma, 296 pure, 13.6 mCi mmol'l) were prepared in PBS. The solutions contained approximately 200,000 dpm ml'1 ring-U-[14C] 2,4-D or 2,4-Dme (Sigma, 12.8 mCi mmol'l) with a final concentration of 10 mg L'l. The solutions were filter sterilized and stored in five 100 ml portions in light shielded bottles at 4°C. Establishing F irst-Order Kinetics 2,4-D degradation rates were measured in solutions of PBS with 2,4-D concentrations of 0.33, 0.67, 1.0, and 1.5 pg ml‘1 with 20,000 dpm ml'l ring-U-[MC] 2,4-D and cell concentrations of approximately 1X108 CFU ml’l. Depletion of solution phase ring-U-[14C] 2,4-D was monitored over time. Twenty ml serum vials containing 10 m1 of labeled 2,4-D solution were sampled at 6, 13, 20, 27, 34, and 41 min after inoculation. Sampling consisted of taking 0.4 ml solution by syringe. This volume was transferred to a microconcentrator tube (Gellman, Nanosep MF Microconcentrators, 0.2um pore size). The tube was immediately centrifuged (Ependorf Centrifuge 5415C) at 11750 RCF for 1 nrin to separate the cells from the solution; 40 pl of 2N HCl were present at the bottom of the microconcentrator tube and effectively drove off any carbon 21 dioxide present in the solution sample. After centrifugation the filter cup was separated from the tube and the cap was cut off the tube. The filter cup and tube were dropped into separate scintillation vials containing 10 ml of scintillation cocktail fluid. Activity was counted on a Packard 1500 Tri-Carb Liquid Scintillation Analyzer for 5 min. Solution phase 14C was used to determine concentrations of 2,4-D that were plotted versus time. Linear regression was used to estimate the rate of depletion for each initial solution concentration. 2,4-D is readily intracellularly metabolized by the microorganisms used in this study (Don and Pemberton 1981, Fulthorpe, et al. 1995, Ogram, et al. 1985). The metabolites proceed via an ortho-cleavage pathway to the formation of TCA cycle intermediates (Pieper, et a1. 1989, Pieper, et al. 1988). As a sole carbon source present in the solution, accumulation of metabolites was not anticipated. This assumption was confirmed with HPLC analysis. Silica Slurries Uncoated, ethyl coated and phenyl-coated silica were obtained from the J .T. Baker company. The irregularly shaped particles have a size of 40 pm with 60 A pores. According to the J .T. Baker Company, surface coverage on the ethyl and phenyl-coated silicas is essentially complete. To create the silica Slurries the solids were wetted with PBS in a Speed Vac (DNA 110 Speed Vac, Savant Instruments, Inc., Farmingdale, NY). Silica (2.5 g) was weighed into a sterile 10 m1 Falcon tube and 5 ml PBS were carefully layered on top of the silica. The tube was placed in a Speed Vac where gentle centrifugation and vacuum where applied simultaneously for 1 min. A disposable sterile loop was used to resuspend the silica. The slurry was then transferred to a 25 ml glass column with a glass fiit. The silica was rinsed with 200 ml of PBS in 10 m1 aliquots. 22 The aliquots were pulled through the silica in a downflow motion with a syringe connected to the bottom of the column. The solution level was never allowed below the silica surface to prevent drying of the silica. After rinsing, the silica was transferred to a 20 m1 serum vial. PBS and stock solutions containing radio-labeled 2,4-D or 2,4-Dme were added to achieve the appropriate slurry density and contaminant concentrations. Teflon stoppers were crimp capped onto the vials. The vials were laid on a platform rocker and mixed well enough to prevent the silica from settling. Sorption Kinetics Sorption kinetics were measured to determine the contact time necessary to reach an appropriate pseudo-equilibrium for the bioavailability experiments. Batch Slurries of silicas were used to measure decreases in solution phase concentrations of contaminant over time. Batch Slurries of 10 ml were sampled at time intervals of 0.25, 0.5, 1, 3, 6, 10, 18, 24, and 48 hr. Samples (0.4 ml) of homogenous slurry were transferred to microconcentrator tubes and solution phase concentrations were determined as previously described. Desorption Kinetics The rate of desorption of each contaminant from each sorbent type was required for use in the Bioavailability Factor Model. A desorption gradient was established by adding PBS to silica Slurries after pseudo-equilibrium sorption with a contaminant had been established. Change in solution phase contaminant concentration was measured over time. The slope of the initial linear portion of the desorption curve was used to establish the rate of desorption. Batch sorbent Slurries were established with initial solution contaminant concentrations of 1.0 mg L'l. After sorption reached pseudo- 23 equilibrium additional PBS was added to the Slurries to induce desorption. The batches were then be sampled at time intervals of 20, 40, 55, 70, 90, 150, 210, 270, 330 and 960 seconds to determine solution phase concentrations as described previously. Bioavailability Assays Bioavailability assays determined the rate of contaminant depletion from solution in the presence of a sorbent. Depletion of solution phase ring-U-[14C] 2,4-D or 2,4-Dme at cell densities of approximately 1X108 CFU ml‘1 in a silica slurry was monitored over time. Control 20 m1 serum vials contained 10 ml PBS and contaminant at 1.0 mg L'1 with approximately 20,000 dpm ml‘1 but no silica. Experimental serum vials contained 10 ml PBS, contaminant at 1.0 mg L'1 with approximately 20,000 dpm ml'l, and 2.5 g silica. Vials were sampled at 6, 13, 20, 27, 34, and 41 min after inoculation. Sampling proceeded as described in the “Establishing First Order Kinetics” section. Distribution Coefficients The sorbent-water distribution coefficients for 2,4-D and 2,4-Dmc were determined for each sorbent type integrally with the bioavailability assays. Silica Slurries were prepared as described in a previous section of methods. The final Slurries contained 2.5 g of silica and 10 ml of PBS with a contaminant concentration of 1 mg L'1 that included the radio-labeled tracer at 20000 dpm ml‘l. The final contaminant concentrations were achieved by putting 1 ml of the appropriate stock solution into each slurry. The same volume of stock solution was also put into control vials that contained 9 m1 PBS and no sorbent. Three one m1 volumes of each stock solution were put directly into scintillation vials containing 10 m1 of scintillation cocktail to determine the total 24 amount of contaminant in the slurry and control systems. The sorbent Slurries and controls were placed on a rocking platform for 45 minutes at a rate high enough to keep the silica suspended. After sorption, a homogenous 0.4 m1 sample of each slurry was transferred to a microconcentrator tube (Gellman, Nanosep MF Microconcentrators, 0.2pm pore size). The microconcentrator tube without the filter portion had been pre- weighed. The tubes were immediately centrifuged (Ependorf Centrifuge 5415C) at 11750 RCF for 1 min to separate the solution from the sorbent. After centrifugation the filter cup was separated from the vial. The vial was weighed with the solution sample . and then the cap was cut off. The filter cup and vial were dropped into separate scintillation vials containing 10 ml of scintillation cocktail fluid. Activity was counted on a Packard 1500 Tri-Carb Liquid Scintillation Analyzer for 5 min. A second slurry sample was transferred to a pre-weighed Ependorf tube. The tube was reweighed with the sample then oven dried at 40° C for 72 hrs before the dried sample was weighed. Solution concentration was calculated from the amount of compound in the centrifuged solution sample and the volume of sample as determined gravimetrically. Sorbed concentration was calculated from the amount of compound in the centrifuged silica sample and the dry weight of silica as determined gravimetrically. Each Kd value is the mean of four independent determinations. Modeling In brief, results from the partition coefficient and desorption kinetic studies were used to determine the proportions of substrate in equilibrium, non-equilibrium, and irreversible sorption sites. These proportions were then used to calculate a bioavailability factor for the particular contaminant/sorbent combination. The bioavailability factor was 25 used to model the expected rate of substrate disappearance from solution when a sorbent is present according to the Bioavailability Factor Model. This calculated expected rate of substrate disappearance from solution, which takes desorption extents and rates into consideration, was compared to the rates measured in the presence of a sorbent to determine if the system was desorption limited. Sum of squares reduction test was used to determine if there was a significant difference between the two rates. 2, 4-D Mineralization in the Presence of Non-Sorbing Ottawa Sand Mineralization kinetics were determined using bioavailability assays similar to those developed by Guerin and Boyd (1992). Mineralization rates of 2,4-D in sand Slurries were directly measured. The sand was pre-sterilized in bulk with gamma irradiation and 15 g were measured into serum bottles in a sterile hood. Seventy-five ml of PBS with 1 pg ml'1 2,4-D, a fraction of which was ring-U-[14C] 2,4-D, was added to each vial and allowed to equilibrate with the sand. Control vials contained 75 m1 PBS including contaminant but no sand. Each vial was inoculated with approximately 108 CFU ml"1 of cells in early stationary phase and crimp sealed with a Teflon lined septum. After inoculation, the bottles were placed at a slant on a platform shaker. At sampling intervals of up to 6 hr the septum of each bottle was pierced with a syringe and 1 ml of solution and 1 ml of headspace was removed. Mineralization of ring-U-[14C] 2,4-D was measured by trapping 14C02 from the samples. Samples were added by syringe to a sealed test tube that contained 1 m1 2N HCl at the bottom, and a cup that contained a small piece of filter paper soaked in 0.25 ml 1N KOH suspended from the top. Tubes sat overnight to allow maximum recovery of 14C02. Filter papers and 2.5 ml ethanol used 26 to rinse the cups were transferred to 7.5 ml scintillation counting fluid. Activity was counted on a Packard 1500 Tri-Carb Liquid Scintillation Analyzer for 5 min. Initial mineralization rate was determined from the linear portion of the mineralization curve that is established in the first hour of the assay. Results and Discussion The model system described here was used to determine if the independently measured processes of contaminant desorption, and degradation from the solution phase can describe degradation of the contaminant in the presence of a sorbent. We evaluated the bioavailability of sorbed contaminant in model systems by comparing the measured rate of degradation in the presence of a sorbent to the rate predicted by the Bioavailability Factor Model. The sorbents used in the model system were silica particles of relatively uniform dimensions and known surface properties. Selected hydrophobic surface coatings provided a range of sorption and desorption extents. Four strains that can use 2,4- dichlorophenoxyacetic acid (2,4-D) or its methyl ester (2,4-Dme), the model contaminants, as sole carbon and energy sources were selected. Each strain internally metabolizes the model contaminants to final products of carbon dioxide and water without the release of intermediate metabolites. Using both 2,4-D and 2,4-Dme as model contaminants enabled the exploration of factors that control the bioavailability of both an an-ionic and a non-ionic contaminant. The 2,4-D molecule has both a hydrophobic ring portion and an acetic acid side chain that is predominately in the anionic form at pH 7.0. Hydrogen bonding with the silica surface and diffusion into micropores were likely the major forms of sorption for 2,4-D 27 on uncoated silica. 2,4-D sorption to the coated silica is expected to also include some hydrophobic interactions. The addition of a methyl group to form 2,4-Dme removes the charge associated with the acetic acid moiety and sorption occurs predominately via hydrophobic interactions. The model contaminants and sorbents were chosen to provide a range of sorption extents. This was accomplished by increasing the hydrophobicity of the silica coating. The linear distribution coefficient, Kd, describes the extent of contaminant sorption; these values ranged from 1.56 ml g'1 to 0.31 ml g'1 (Table 1). Sorption/desorption processes were fast relative to degradation rates for all sorbent-contaminant combinations. Sorption pseudo-equilibrium was established in each case in less than fifteen minutes. The fraction of sorbed contaminant that desorbed as solution phase concentrations were depleted is referred to as the fraction in equilibrium (feq). These values ranged from 1 to 0 as determined from the desorption isotherm and Kd values (Table 1). This is the broadest possible range of feq values. An feq value of 1 represents a system in which all the sorbed contaminant is in instantaneous equilibrium with the bulk solution. An feq value of 0 represents a system in which all the sorbed contaminant is irreversibly sorbed. In all cases where desorption occurred it was very fast relative to the rate of depletion from solution. The fraction of sorbed contaminant that was not in desorption equilibrium was irreversibly sorbed for the time scale and conditions of our study. 28 Table 1. The linear sorption coefficient (Kd) and fraction of solute in equilibrium (feq) values for each combination of sorbent and contaminant Sorbent and Contaminant Type Uncoated silica Ethyl coated silica Phenyl coated silica 2,4-Dme 2,4-D 2,4-Dme 2,4-D 2,4-Dme 2,4-D Kd (ml/g) 1.56 3:0.03 1.37 21:0.02 0.71 21:0.09 0.31 i003 1.14 $0.07 0.69 $0.06 f eq 0.44 i001 1.00 i000 0.00 3:0.00 0.33 3:0.28 0.11 i004 0.56 $0.01 Kd values represent the mean of four measurements i one standard deviation. feq values represent the mean of two measurements :1: one standard deviation. For the bioavailability assays it was important to work under conditions which resulted in first-order kinetics for depletion of the contaminant. This was accomplished by using contaminant concentrations below KS and a bacterial population that was not multiplying. The inoculum was in early stationary phase, there was not enough carbon in solution to provide for growth and the assays were short-term. Independent assays of solution-phase-depletion kinetics established that all four strains of bacteria demonstrated first-order degradation kinetics for model substrate concentrations up to at least 1.5 mg L' 1. The bioavailability assays compared the measured rate of degradation in the presence of a sorbent to the rate predicted by the Bioavailability Factor Model (BFM). The Bioavailability Factor Model uses the independently determined processes of desorption, and degradation in the solution phase to predict the rate of degradation in the presence of a sorbent when only solution phase contaminant is available to the bacteria. The bioavailability of sorbed contaminant was assessed in twenty-four unique systems. 29 These systems represented a wide range of sorption and desorption extents, a non-ionic and an an-ionic contaminant with different mechanisms of sorption, and four different strains. The results showed ten combinations in which the ratio of the measured to the predicted rate of depletion was significantly greater than one. Sorbed contaminant may have been directly available in these systems. There were nine combinations in which desorption and solution phase degradation accurately described the system and five combinations in which the presence of the sorbent apparently inhibited solution phase degradation (Tables 2 and 3). Table 2. Bioavailability of sorbed 2,4-D as indicated by the ratio of the measured rate of degradation in the presence of sorbent to the rate predicted by the Bioavailability Factor Model. Strain Sorbent type FB4 TFD6 JMP134 RASC Uncoated silica 1.27“ 1.48" 1.51“ 1.10 Ethyl coated silica 0.37“ 0.62* 1.21* 0.97 Phenyl coated silica 0.84 0.78 1.61 ** 1.11* ** p S 0.01 * p 5 0.025 Level of significance as determined with the sum of squares reduction test for the difference between the measured and predicted rates. 30 Table 3. Bioavailability of sorbed 2,4-Dme as indicated by the ratio of the measured rate of degradation in the presence of sorbent to the rate predicted by the Bioavailability Factor Model. Strain Sorbent type FB4 TF D6 JMP134 RASC Uncoated silica 1.46M 1.50“ 2.67" 1.16 Ethyl coated silica 0.24* * 0.46* * 1.47* 0.91 Phenyl coated silica 1.08 0.77* 1.25 1.13 ** p S 0.01 * p 5 0.025 Level of significance as determined with the sum of squares reduction test for the difference between the measured and predicted rates. When the ratio of the rate of depletion measured in the presence of a sorbent to the rate predicted by the BFM is equal to one there is not significant evidence to conclude that sorbed contaminants available, they are likely unavailable. In these cases desorption is expected to be the rate limiting process. The ratio of the measured to the predicted rate of depletion does not always equal one. A ratio less than one indicates that the presence of the sorbent inhibited solution phase degradation of the contaminant. Guerin et. al. (1997) suggested that this may result from adhesion of the bacteria to the silica occluding a portion of the cell membrane and thereby limiting contaminant uptake. However, this assumes that uptake is the rate limiting step in the degradation of the contaminant. A second possibility is that the presence of the sorbent results in a second carbon source that results in the inhibition of degradation of the target contaminant (Daugherty and Karel 1994, Kim and Maier 1986, Papanastasiou and Maier 1982). Two conclusions are possible when the ratio is greater than one. Sorbed contaminants may be directly available to the bacteria via increased localized 31 concentrations, or the presence of the sorbent stimulated the metabolism of the bacteria. The latter possibility could occur if adhesion of the microorganism to the sorbent surface resulted in a change in the allocation of energy to various cell processes. However, this change in energy allocation is more likely to result in the production of exopolysacharides than increased rates of substrate metabolism. Adhesion stimulated metabolism was excluded as a contributing mechanism in experiments that compared the rates of 2,4-D metabolism in the presence and absence of non-sorbing quartz sand. No increase in rates of metabolism were observed even though bacterial adhesion was greater than sixty percent for each of the strains (Figure 7). x10"12 x10'12 Initial 2,4-D mineralization rate - — . — + - 0 i ~- i 0.0 0.5 1.0 0.0 0.5 1.0 2,4-D solution concentration 2,4-D solution concentration (ug ml'l) (ug ml'l) Figure 7. Initial mineralization rates of 2,4-D in the presence (0) and absence (A) of non-sorbing quartz sand. Figure 7A shows data for bacterial strain FB-4 and is representative of strains RASC and TFD6. The presence of non-sorbing sand does not effect the rate of mineralization of solution phase 2,4-D for these three strains. Figure 7B shows the data for bacterial strain JMP134. The presence of non-sorbing sand may decrease the rate of mineralization of solution phase 2,4-D for JMP134. The possibility that sorbed contaminants may be directly available to the bacteria via increased localized concentrations would result in enhanced degradation rates in the presence of a sorbent. This would be possible if desorption or mass transfer is increased 32 beyond that measured in the abiotic systems. Our experiments showed the greatest enhancement of degradation in systems with the highest Kd and fraction of sorbed substrate in equilibrium (feq). This was the case for sorption of either contaminant to uncoated silica. Uncoated silica lacks an organic phase which eliminates the possibilities for hydrophobic interactions, leaving accumulation in micro-pores and surface reactions as the predominant sorptive processes. The relatively high Kd and feq can be taken as an indication that the contaminant accumulated in the micropores and on the surface but was readily desorbed when a concentration gradient was present. Depletion of surface adsorbed 2,4-D by bacteria present on the sorbent surface could have provided the necessary concentration gradient to facilitate desorption and mass transfer from the micro-pores into the thin aqueous layer surrounding the cell where uptake and degradation proceeded rapidly without diffusion of the contaminant into the bulk solution. The availability of contaminant accumulated on the surface and in micro-pores in greater concentrations than in the bulk solution could account for the enhanced rate of contaminant degradation observed when uncoated silica is present as a sorbent. Conclusions The tremendous variation in bioavailability in the model systems studied here indicates that caution should be exercised when generalizing any single conclusion regarding the bioavailability of sorbed compounds. These results show that bioavailability varies with sorbent type, organism, and substrate. There is evidence that some systems can be predicted from desorption and solution phase degradation while others have rates of depletion from solution nearly three times the rate predicted from desorption and degradation of only solution phase contaminant. Finally there are also 33 cases in which the presence of a sorbent significantly inhibits degradation of even solution phase contaminant. Acknowledgements This research was supported by NSF grant DEB9120006, USDA NRICGP No. 94-37107-03 86, and the Michigan Agricultural Experiment Station. Literature Cited Al Bashir, B., J. Hawari, R. Samson, and R. Leduc. 1994. Behavior of nitrogen- substituted naphthalenes in flooded soil: Part II. Effect of bioavailability on biodegradation kinetics. Water Res. 28(8): 1827-1833. Alexander, M., and K. M. Scow. 1989. Kinetics of biodegradation in soil, p. 243-269. In B. L. Sawhney and K. Brown (ed.), Reactions and Movement of Organic Chemicals in Soils. Soil Science Society of America and American Society of Agronomy, Madison, WI. Amador, J. A., and M. Alexander. 1988. Effect of humic acids on the mineralization of low concentrations of organic compounds. Soil Biol. Biochem. 20(2): 1 85-191. Calvillo, Y. M., and M. Alexander. 1996. Mechanisms of microbial utilization of biphenyl sorbed to polyacrylic beads. Appl. Microbiol. Biotechnol. 45(3):383-390. Cowan, C. E., D. J. Versteeg, R. J. Larson, and P. J. Kloepper—Sams. 1995. Integrated approach for environmental assessment of new and existing substances. Regulat. Toxicol. Pharmacol. 21:3-31. Cracker, F. H., W. F. Guerin, and S. A. Boyd. 1995. Bioavailability of naphthalene sorbed to cationic surfactant-modified smectite clay. Environ. Sci. Technol. 29(12):2953- 2958. Daugherty, D. D., and S. F. Karel. 1994. Degradation of 2,4-dichlorophenoxyacetic acid by pseudomonas cepacia DBOl(pR0101) in a dual-substrate chemostat. Appl. Environ. Microbiol. 60(9):3261-3267. Don, R. H., and J. M. Pemberton. 1981. Properties of six pesticide degradation plasmids isolated from Alcaligenes paradoxus and Alcaligenes eutrophus. J. Bacteriol. l45(2):681-686. 34 Efroymson, R. A., and M. Alexander. 1991. Biodegradation by an Arthrobacter spp. of hydrocarbons partitioned into an organic solvent. Appl. Environ. Microbiol. 57(5):1441- 1447. Efroymson, R. A., and M. Alexander. 1994. Role of partitioning in biodegradation of phenanthrene dissolved in nonaqueous-phase liquids. Environ. Sci. Technol. 28:1172- 1179. Fulthorpe, R. R., C. McGowan, O. V. Maltseva, W. E. Holben, and J. M. Tiedje. 1995. 2,4-Dichlorophenoxyacetic acid-degrading bacteria contain mosaics of catabolic genes. Appl. Environ. Microbiol. 61(9):3274-3281. Griffith, P. C., and M. Fletcher. 1991. Hydrolysis of protein and model dipeptide substrates by attached and nonattached marine Pseudomonas sp. strain NCIMB 2021. Appl. Environ. Microbiol. 57(8):21 86-2191 . Guerin, W. F., and S. A. Boyd. 1992. Differential bioavailability of soil-sorbed naphthalene to two bacterial species. Appl. Environ. Microbiol. 58(4):1142-1152. Guerin, W. F., and S. A. Boyd. 1993. Bioavailability of sorbed naphthalene to bacteria: Influence of contaminant aging and soil organic carbon content. SSSA Special Publication(32): 197-208. Guerin, W. F., and S. A. Boyd. 1997. Bioavailability of naphthalene associated with natural and synthetic sorbents. Water Res. 31(6): 1504-1512. Harms, H., and A. J. B. Zehnder. 1995. Bioavailability of sorbed 3-chlorodibenzofuran. Appl. Environ. Microbiol. 61(1):27-33. Hatzinger, P. B., and M. Alexander. 1995. Effect of aging of chemicals in soil on their biodegradability and extractability. Environ. Sci. Technol. 29:537-545. Hatzinger, P. B., and M. Alexander. 1997. Biodegradation of organic compounds sequestered in organic solids or in nanopores within silica particles. Environ. Toxicol. Chem. 16:2215-2221. Kim, C. J., and W. J. Maier. 1986. Acclirnation and biodegradation of chlorinated organic compounds in the presence of alternate substrates. J. Water Pollut. Control Fed. 58:157-164. Marshman, N. A., and K. C. Marshall. 1981. Bacterial growth on proteins in the presence of clay minerals. Soil Biol. Biochem. 13:127-134. McGowan, C. 1994. Interspecies gene transfer in the evolution of 2,4- dichlorophenoxyacetic acid degrading bacteria. Ph.D. thesis. Michigan State University, East Lansing. 35 Mihelcic, J. R., and R. G. Luthy. 1991. Sorption and microbial degradation of naphthalene in soil-water suspensions under denitrification conditions. Environ. Sci. Technol. 25(1): 169-177. Moyer, J. R., R. J. Hance, and C. E. McKone. 1972. The effect of adsorbents on the rate of degradation of herbicides incubated with soil. Soil Biol. Biochem. 4:307-311. Ogram, A. V., R. E. Jessup, and P. S. C. Rao. 1985. Effects of sorption on biological degradation rates of (2,4-dichlorophenoxy)acetic acid in soils. Appl. Environ. Microbiol. 49:582-587. Ortega-Calvo, J. J., M. Lahlow, and C. Saiz-Jimenez. 1997. Effect of organic matter and clays on the biodegradation of phenanthrene in soils. Int. Biodeterior. Biodegrad. 40(2-4): 101-106. Ortega-Calvo, J. J., and C. Saiz-Jimenez. 1998. Effect of humic fractions and clay on biodegradation of phenanthrene by a Pseudomonasfluorescens strain isolated from soil. Appl. Environ. Microbiol. 64(8):3123-3126. Papanastasiou, A. C., and W. J. Maier. 1982. Kinetics of biodegradation of 2,4- dchlorophenoxyacetate in the presence of glucose. Biotechnol. Bioeng. 24:2001-2011. Pieper, D. H., K. Engesser, and H. Knackmuss. 1989. Regulation of catabolic pathways of phenoxyacetic acids and phenols in Alcaligenes eutrophus JMP 134. Archives of Microbiology. 151:365-371. Pieper, D. H., W. Reinke, K.-H. Engesser, and H.-J. Knackmuss. 1988. Metabolism of 2, 4-dichlorophenoxyacetic acid, 4-chloro-2-methylphenoxyacetic acid and 2- methylphenoxyacetic acid by Alcaligenes eutrophus JMP 134. Arch Microbiol. 150:95- 102. Radosevich, M., S. J. Traina, and O. H. Tuovinen. 1997. Atrazine mineralization in laboratory-aged soil microcosms inoculated with s-triazine-degrading bacteria. J. Environ. Qual. 26(1):206-214. Rijnaarts, H. H. M., A. Bachmann, J. C. Jumelet, and A. J. B. Zehnder. 1990. Effect of desorption and intraparticle mass transfer on the aerobic biomineralization of alpha- hexachlorocyclohexane in a contaminated calcareous soil. Environ. Sci. Technol. 24(9): 1349-1354. Robinson, K. G., W. S. Farmer, and J. T. Novak. 1990. Availability of sorbed toluene in soils for biodegradation by acclimated bacteria. Water Res. 24(3):345-350. 36 Scow, K. M., and M. Alexander. 1992. Effect of diffusion on the kinetics of biodegradation: experimental results with synthetic aggregates. Soil Sci. Soc. Am. J. 56:128-134. Scow, K. M., and J. Hutson. 1992. Effect of diffusion and sorption on the kinetics of biodegradation: Theoretical considerations. Soil Sci. Soc. Am. J. 56(1):] 19-127. Scribner, S. L., T. R. Benzing, S. Sun, and S. A. Boyd. 1992. Desorption and bioavailability of aged simazine residues in soil from a continuous corn field. J. Environ. Qual. 21(1):] 15-120. Shelton, D. R., and M. A. Doherty. 1997. A model describing pesticide bioavailability and biodegradation in soil. Soil Sci. Soc. Am. J. 61: 1078-1084. Shimp, R. J., and R. L. Young. 1988. Availability of organic chemicals for biodegradation in settled bottom sediments. Ecotoxicol. Environ. Saf. 15:31-45. Smith, S. C., C. C. Ainsworth, S. J. Traina, and R. J. Hicks. 1992. Effect of sorption on the biodegradation of quinoline. Soil Sci. Soc. Am. J. 56(3):737-746. Steen, W. C., D. F. Paris, and G. L. Baughman. 1980. Effect of sediment sorption on microbial degradation of toxic substances, vol. 1. Ann Arbor Sci. Publ., Ann Arbor, MI. Tang, W. C., J. C. White, and M. Alexander. 1998. Utilization of sorbed compounds by microorganisms specifically isolated for that purpose. Appl. Microbiol. Biotechnol. 49:117-121. Weber, J. B., and H. D. Coble. 1968. Microbial decomposition of diquat adsorbed on montrnorillonite and kaolinite clays. J. Agric. Food Chem. l6(3):475-478. Weissenfels, W. D., H. J. Klewer, and J. Langhoff. 1992. Adsorption of polycyclic aromatic hydrocarbons (PAHs) by soil particles: Influence on biodegradability and biotoxicity. Appl. Microbiol. Biotechnol. 36(5):689-696. Zhang, W., E. J. Bouwer, and W. P. Ball. 1998. Bioavailability of hydrophobic organic contaminants: Effects and implications of sorption-related mass transfer on bioremediation. Ground Wat. Monitor. Remed. 18(1): 126-128. 37 Chapter 3 CONSISTENCIES IN BACTERIAL ADHESION MEASUREMENTS AND THEIR RELATIONSHIP TO BIOAVILABILITY Abstract Biodegradation of localized zones of high contaminant concentration on sorbent surfaces may be related to bacterial adhesion. Unfortunately it is difficult to quantify bacterial adhesion in natural systems as evidenced by the lack of published methods. The experiments described here show that laboratory assays of bacterial adhesion in model systems could establish relative adhesion abilities for four strains. The relative ability of the four strains to adhere to surfaces was independent of the five surface types tested, which ranged from very hydrophilic, silica, to very hydrophobic, Teflon. Adhesion ability was also independent of the absence or presence of low levels of sorbed carbon sources and of the two assay methods used to quantify adhesion. The consistency of relative adhesion abilities allowed each strain to be qualitatively designated as a high or low adhesion strain. This qualitative designation was then related to the results of bioavailability assays from previous studies. High adhesion organisms were affected by the presence of a sorbent, either enhancement or inhibition of bioavailability, in seven of the twelve systems evaluated. Low adhesion organisms were only affected in two of the twelve systems. Depletion of substrate in the presence of a sorbent was accurately predicted by the sum of depletion from solution and desorption in twice as many cases for low adhesion organisms than high adhesion organisms. 38 Introduction A growing number of research reports on bioavailability have concluded that the availability of sorbed compounds to bacteria is system dependent. As more sophisticated methods to directly assess the availability of sorbed compounds have been developed our opportunities to identify the specific conditions which control bioavailability have increased. Guerin and Boyd (l992)showed that the bioavailability of sorbed naphthalene was organism specific. They measured the mineralization of naphthalene sorbed to whole soils. The results suggested dramatically different bioavailability of sorbed naphthalene to two strains. Further studies by Guerin and Boyd (1997) showed that the bioavailability of sorbed naphthalene was also dependent on the sorbent type. Harms and Zehnder (1995) used column studies to demonstrate that bacteria attached to porous Teflon granules could be efficiently supplied with desorbing 3-chlorodibenzofuran. Their results suggested that bacterial adhesion to the sorbent surface could have an important influence on the degradation of sorbed compounds. Tang et al. (1998) reported recently that the rate of degradation of sorbed phenanthrene by bacteria isolated by enrichment on sorbed substrate was faster than degradation by the wild type strain and faster than the rate of phenanthrene desorption determined in the absence of bacteria. These studies indicate the important role that bacterial adhesion to the sorbent surface can have on the biodegradation of sorbed compounds. The lack of published methods and our own research difficulties have taught us that it is very difficult to measure the extent of bacterial adhesion in natural systems. Column studies with soils can be biased by pore clogging rather than true adhesion. Fine silt and organic matter clogging of the filter membrane can restrict the use of filtration to separate free cells from 39 soil. Attempts to separate free cells on a density basis have also been hampered by the very heterogeneous nature of soils. Therefore, we have attempted to use laboratory adhesion assays in model systems to study the relationship between bacterial adhesion and bioavailability of sorbed compounds. The experiments described herein used two different methods to determine the extent of bacterial adhesion to surfaces ranging from very hydrophilic, glass, to very hydrophobic, Teflon. The first method was based on the column adhesion assay developed by Rijinaarts et a1. (1993), where adhesion to glass and Teflon beads was quantified in flow through columns. Adhesion reversibility was also determined by rinsing the column with cell- free buffer and finally cell-free distilled water. The second method quantified bacterial adhesion in batch systems. The batch systems were comprised of hydrophilic and hydrophobic silica Slurries. Free cells were efficiently separated from cells attached to the silica particles via filtration. The objective of this study were to determine the adhesion of four bacterial isolates from Michigan agricultural soils on five surface types ranging from hydrophilic to hydrophobic. The adhesion assay results were used to classify each bacterial strain as either a high or low adhesion strain. This qualitative adhesion assessment was then related to the results of bioavailability assays from previous experiments. Materials and Methods Organisms Four strains of proteobacteria that are able to use 2,4-dichlorophenoxyacetic acid (2,4-D) or 2,4-dichlorophenoxyacetic acid methyl ester (2,4-Dme) as a sole carbon and energy source were selected for this study. Two of the strains, designated RASC and 40 TF D6, were classified by McGowan (1994) based on 168 rDNA sequence analysis as being Burkholderia sp. Don and Pemberton (1981) previously classified the third strain, JMP134, as Alcaligenes Eutrophus. The fourth strain, FB4, was classified as F lavobacterium sp. by Ogram et al. (1985). Each of these organisms internally metabolizes 2,4-D. Cells were grown until early stationary phase in a minimal salts media ([MSM] 1419.6 mg NazHPO4, 1360.9 mg KH2P04, 0.3 mg (NH4)ZSO4, 50 mg MgSO4 7H20, 5.88 mg CaC12 2H20, 3.2 mg EDTA disodium salt, 2.78 mg FeSO4 71120, 1.15 mg ZnSO4 71120, 1.69 mg MnSO4 H20, 0.375 mg CuSO4 5H20, 0.233 mg Co(NO3)2 6H20 and 0.1236 mg (NH4)6Mo7024 4H20 per liter of distilled water) that contained 400 mg L'1 2,4-D. Column Adhesion Assay The ability of each of the four strains to attach to a hydrophobic (Teflon) and a hydrophilic (glass) surface was measured by determining the fraction of cells that remained attached to spherical beads once they had been added to an evenly packed flow through column. The methods were adopted from the column experiments described by Rijnaarts et al. (1993). Briefly, 10 cm glass columns (1.0 cm i.d.) were loaded with glass or Teflon beads submerged in phosphate buffered saline ([PBS] 8.5 g NaCl, 0.6 g Na2 P04, 0.3 g KH2P04, per liter of distilled water, pH 7.0). Cells were grown as described above, washed and then diluted to an optical density at 280 nm of 0.60 in PBS. These suspensions were supplied to the vertical down-flow columns by a peristaltic pump at a constant flow rate of approximately 19 ml h'1 for 80 nrin. The effluent of each column was collected in S-min fractions. The optical density of the effluent fractions was 41 measured using a spectrophotometer set at a wavelength of 280 nm to determine the proportion of cells that remained adhered to the column. After 12 pore volumes the influent was changed to cell-free PBS for 7 pore volumes and finally to cell free distilled water for 9 pore volumes to measure detachment. The assays were done synchronously in triplicate with a shared influent cell suspension. Stock Solutions Sterile stock solutions of 14C-labeled 2,4-D (Sigma, 296 pure, 12.8 mCi mmol'l) and 2,4-Dme (Sigma, 296 pure, 13.6 mCi mmol'l) were prepared in PBS. The solutions contained approximately 200,000 dpm ml'l ring-U-[l 4C] 2,4-D or 2,4-Dme with a final concentration of 10 mg L'l. The solutions were filter sterilized and stored in five 100-m] portions in light shielded bottles at 4°C. Batch Adhesion Assay The second set of adhesion assays was done in batch-silica Slurries. Three types of silica were used: hydrophilic uncoated silica, and hydrophobic silicas with phenyl and ethyl coatings. A11 silicas were obtained from the J .T. Baker company. The irregularly shaped particles have a size of 40 pm with 60 A pores. According to the J .T. Baker Company, surface coverage with the hydrophobic functional groups was complete on the coated silicas. To create the silica Slurries the solids were wetted with PBS in a Speed Vac (DNA 110 Speed Vac, Savant Instruments, Inc., Farmingdale, NY). Silica (0.375 g) was weighed into a sterile 2 ml Ependorf tube and 1 ml PBS was carefully layered on top. The tubes were placed in a Speed Vac where gentle centrifugation and vacuum were applied simultaneously for 1 min. A disposable sterile loop was used to resuspend the 42 silica, then 350 pl PBS and 150 pl stock solutions containing 2,4-D or 2,4-Dme were added to achieve the appropriate slurry density (1 :4 w/v) and contaminant concentrations. All tubes were placed on a vertical Roto-Torque rotator at a moderate speed (Cole Panner Instrument Company, Model 7637) for at least 30 minutes to allow sorption of the contaminant to reach psuedo-equilibrium. A filtration method was used to determine the adhesion of each of the four strains to the silica particles. Adhesion was quantified by separating free cells from those attached to silica with a Nuclepore 8.0 pm polycarbonate filter. The cells were grown, as described above, in MSM with 400 mg L'1 2,4-D that also included ring-U-[14C] 2,4-D (Sigma, 296 pure, 12.8 mCi mmol'l) at 250,000 dpm ml'l. In this manner the cells incorporated a significant portion of the radiolabel that allowed their proportions in various compartments of the silica slurry system to be quantified. When the cells were in early stationary phase, approximately 10 ml of culture was centrifuged at 7649 RCF (Sorvall RC-SB Refiigerated Superspeed Centrifuge with 88-34 rotor) for 10 min and then washed twice in PBS. The final inoculum was resuspended in PBS. This suspension was serial diluted and plated to determine the cell density in the assays. Inoculum (15 pl) was added to Ependorf tubes that contained the silica slurry or PBS plus 2,4-D or 2,4-Dme. The cell density in the silica Slurries was in the order of 107 CFU ml ' 1. Inoculum (15 pl) was also directly added to three scintillation vials containing 10 ml of scintillation fluid so the total radioactivity in the systems was known. All tubes were placed back on the rotator for 10 min afier inoculation. To separate free cells from those attached to silica the entire contents of one tube were filtered through a 25 mm Nuclepore polycarbonate filter membrane with 8.0 pm 43 pores. The tube was then rinsed with two 1 ml aliquots of PBS that were also poured onto the filter. After the filtrate was collected in a scintillation vial, 8 ml of scintillation fluid was added; if a gel did not form 5 more ml of scintillation fluid was added to the via]. The filter paper with the associated silica and the empty Ependorf tube were put into separate scintillation vials with 10 ml scintillation fluid to complete the mass balance. Activity was counted on a Packard 1500 Tri-Carb Liquid Scintillation Analyzer for 5 min. Results and Discussion The column adhesion assays showed that all of the organisms adhered to a greater extent on Teflon than to glass beads. In all cases the adhesion was irreversible when the columns were flushed with cell-free buffer. Only a very small fraction of cells were released when the columns were rinsed with distilled water (Figure 8). 44 Q . o l A ' " 8 1.0 T —o—RASC Z’ 1 Cell suspension Cell-free buffer + JMP134 o '_ 'nfl t 'nfl t E? 0.8 i '1 “en ‘ “en —B—TFD6 2 l +FB4 _ c3 . . \ l Drstrlled water 8 influent ‘5 l D 33‘ _.;—_.._._L 717-1.; 40 50 60 S 8 —o—RASC g +JMP134 g. —a—TFD6 2 —o—FB4 8 D O 25. CE: 5's .. _ 33 Evil... 40 50 60 Total Flow (ml) Figure 8. The abilities of four strains to attach to Teflon (A) and glass (B) surfaces were measured by determining the fraction of cells that flowed through a column packed with spherical beads. Reversibility of adhesion was ascertained by changing the column influent from the cell suspension to a cell-free buffer and then distilled water. All data points represent the mean of three measurements. Error bars represent one standard deviation. 45 The filtration-adhesion assays measured adhesion of the four strains in batch- silica Slurries. The mass balance showed excellent recovery of 1"'C in the filtration- adhesion assays (98% :1: 7). The high recovery demonstrated that the 14C incorporated into the cells during growth was not lost to C02 during the adhesion assay. Adhesion to the uncoated silica was lower than adhesion to either of the hydrophobically coated silicas for all four strains. Strains TFD6 and FB4 attached to a greater extent than strains JMP134 and RASC in each of the slurry types. These results are consistent with the results of the column-adhesion assays (Figure 9). 120 i— - w 100 i ___JIJMP134 _ 1 IRASC a I 801 aTFD6 ? a 3513; y a) 7/ a=_ a .- xx // — // // // // _ // a A A/ g D E F Adhesion measurement system Figure 9. Percent adhesion of four bacterial strains as determined by two methods on six surfaces. Adhesion measurement systems were (method/surface) A- filter/control, B-filter/uncoated silica, C-filter/ethyl coated silica, D-filter/phenyl coated silica, E-column/Teflon, F-column/glass. Each bar represents the mean of three measurements. Error bars represent one standard deviation. 46 Adhesion was slightly higher in the control Ependorf tubes that contained just PBS than in the tubes containing uncoated silica slurry for two of the four strains. This would be possible if adhesion to the walls of the tube was greater than adhesion to the uncoated silica particles. Mixing may have disrupted bacterial adhesion to the tube walls when the uncoated-silica slurry was in the tube. The presence of sorbed carbon sources has the potential to affect bacterial adhesion to the sorbents however this was not the case in this experimental system. Variations in the experimental set-up of the silica Slurries revealed that adhesion ability was independent of the absence or presence of low levels of sorbed carbon sources (Table 4). Adhesion was not significantly different for strain TFD6 when 2,4-D was present or absent. Adhesion was not significantly different for strains JMP134, RASC, and F B4 when either 2,4-D or the methyl ester were present. 47 Table 4. Bacterial adhesion in the presence and absence of 2,4-D and 2,4-Dme Strain Surface Carbon source JMP134 RASC TFD6 FB4 Uncoated silica None NDb ND 25 i3 ND Uncoated silica 2,4-D ] 1i] 9 i8 21 i6 28 i9 Uncoated silica 2,4-Dme 12:]:1 5 3:2 ND 26 i4 Ethyl coated silica None ND ND 92 i2 ND Ethyl coated silica 2,4-D 59 i3 76 i2 92 i8 96 2t] Ethyl coated silica 2,4-Dme 58 i3 74 i1 ND 95 i1 Phenyl coated silica None ND ND 94 i2 ND Phenyl coated silica 2,4-D 54 i2 65 i1 95 i5 95 i1 Phenyl coated silica 2,4-Dme 56 i2 63 i2 ND 95 :l:1 a Values represent the mean % adhesion for three measurements A: one standard deviation. b ND means no data. The information in Figure 9 and Table 4 was used to qualitatively designate two general classes of bacterial adhesion. Strains TF D6 and FB4 were designated high adhesion strains. Strains JMP134 and RASC were designated low-adhesion strains. The qualitative designation for bacterial strain adhesion ability was then related to the results of the bioavailability assays presented in Chapter 2 of this dissertation. The high adhesion organisms, TF D6 and F B4, demonstrated potential use of sorbed substrate in four of the twelve systems tested. Depletion from solution in the presence of a sorbent could be predicted by desorption and solution phase degradation in five of the systems tested. The high adhesion organisms exhibited inhibition of even solution phase substrate 48 in three of the tested systems. The low adhesion organisms demonstrated two cases of enhanced depletion, ten cases with no effect and no cases of inhibition. Thus, high adhesion organisms were affected by the presence of a sorbent, either enhancement or inhibition, in seven of the twelve systems. Low adhesion organisms were only affected in two of the twelve. Depletion of substrate in the presence of a sorbent was accurately predicted by the sum of depletion from solution and desorption in twice as many cases for low adhesion organisms than high adhesion organisms (Table 5). Table 5. Relationship between bacterial adhesion and bioavailability of sorbed compounds Low Adhesion Strains High Adhesion Strains Sorbent Substrate JMP134 RASC TFD6 FB4 Uncoated Silica 2,4-D 0a 0 + + Uncoated Silica 2,4-Dme + 0 + + Phenyl-coated Silica 2,4-D + 0 0 0 Phenyl-coated Silica 2,4-Dme 0 0 0 0 Ethyl-coated Silica 2,4-D 0 0 0 - Ethyl-coated Silica 2,4-Dme 0 0 - - a Symbols: 0, substrate depletion was predicted from desorption and solution phase depletion rates; +, substrate depletion was faster than predicted; -, substrate depletion was slower than predicted. Conclusions Variations in the silica slurry adhesion assays suggested that the relative ability of the four strains to adhere to surfaces was independent of the five surface types tested, which ranged from very hydrophilic, silica, to very hydrophobic, Teflon, and of the two 49 assay methods used to quantify adhesion. The consistencies in these bacteria] adhesion assays offered a basis for generalization of relative adhesion abilities to a vary range of surface types. Until practical methods are developed for quantification of adhesion in natural soils it may be possible to measure adhesion in a simple model system and estimate the relative adhesion in soils or other complex systems. We were able to relate our findings on the relative adhesion abilities of four strains to the results of bioavailability assays. The results suggest that bacterial adhesion to sorbent surfaces plays an important role in the bioavailability of sorbed materials. Acknowledgements This research was supported by NSF grant DEB9120006, USDA NRICGP No. 94—37107-03 86, and the Michigan Agricultural Experiment Station. Special thanks are offered to the laboratory of Hauke Harms at EAWAG, Switzerland, for teaching the column adhesion assay technique to Denise Kay. Literature Cited Don, R. H., and J. M. Pemberton. 1981. Properties of six pesticide degradation plasmids isolated from Alcaligenes paradoxus and Alcaligenes eutrophus. J. Bacteriol. l45(2):681-686. Guerin, W. F., and S. A. Boyd. 1992. Differential bioavailability of soil-sorbed naphthalene to two bacterial species. Appl. Environ. Microbiol. 58(4):]142-1152. Guerin, W. F., and S. A. Boyd. 1997. Bioavailability of naphthalene associated with natural and synthetic sorbents. Water Res. 31(6):1504-1512. Harms, H., and A. J. B. Zehnder. 1995. Bioavailability of sorbed 3-chlorodibenzofuran. Appl. Environ. Microbiol. 61(1):27-33. McGowan, C. 1994. Interspecies gene transfer in the evolution of 2,4- dichlorophenoxyacetic acid degrading bacteria. Ph.D. thesis. Michigan State University, East Lansing. 50 Ogram, A. V., R. E. Jessup, and P. S. C. Rao. 1985. Effects of sorption on biological degradation rates of (2,4-dichlorophenoxy)acetic acid in soils. Appl. Environ. Microbiol. 49:582-5 87. Rijnaarts, H. H. M., W. Norde, E. J. Bouwer, J. Lyklema, and A. J. B. Zehnder. 1993. Bacterial adhesion under static and dynamic conditions. Appl. Environ. Microbiol. 59:3255-3265. Tang, W. C., J. C. White, and M. Alexander. 1998. Utilization of sorbed compounds by microorganisms specifically isolated for that purpose. Appl. Microbiol. Biotechnol. 49:117-121. 51 Chapter 4 BIOAVAILABILITY RELATED TO MECHANISM OF SORPTION Abstract Sorption of organic compounds reduces their solution-phase concentrations, potentially rendering those compounds unavailable for biological degradation. Sorption is usually the result of combined contributions of various mechanisms. We have studied sorption and desorption kinetics and extent in model systems and related the results to a previous study of bioavailability in the same systems. The experimental matrix was comprised of two sorbates, three sorbents, and four strains of bacteria. One sorbate, 2,4- dichlorophenoxyacetic acid, is predominately in the anionic form at pH 7.0. The second sorbate is the methyl ester of the same compound and remains uncharged in the model system. The simplest sorbent, uncoated silica, was chosen to represent a porous mineral surface. Silica coated with either phenyl or ethyl functional groups was used to simulate natural sorbents that contain organic phases. Sorption and desorption kinetics in all systems were fast. They could be considered instantaneous relative to degradation rates. Sorption and desorption extents varied. The product of the sorption extent and fraction in desorption equilibrium represents the portion of the sorbed contaminant that is potentially immediately bioavailable. This portion was directly related to the occurrence of enhanced rates of contaminant degradation in the presence of a sorbent relative to the rate predicted by the Bioavailability Factor Model. The portion was highest when uncoated silica was the sorbent. We propose that the high portion of directly available contaminant 52 and the enhanced rates of degradation are the result of accumulation in micropores being the predominate mechanism of sorption for uncoated silica. Introduction In the soil environment, the concentration of contaminant in the bulk solution can be reduced by sorption to soil. The extent of sorption is generally determined by measuring the loss of a test organic compound from solution and then performing a mass balance to calculate the amount of organic compound sorbed. The extent of sorption can then be expressed by the distribution coefficient, Kd, which represents the ratio of the amount of chemical sorbed per mass of solid sorbent to its aqueous concentration. The distribution coefficient represents the extent of sorption but does not offer any insight to the mechanism of sorption. Sorption is usually the result of combined contributions of various mechanisms. This is demonstrated in complex natural soil and sediment systems that are heterogeneous mixtures of minerals, inorganic precipitates and organic matter. In the presence of a complex sorbent the mechanisms of sorption may include (1) direct chemical bonding or surface complex formation, (2) electrostatic interactions involving association with the diffuse double layer, and (3) exclusion from the aqueous solution due to hydrophobic effects (Schwarzenbach, et al. 1992, Stumm 1992, Westall 1987). The mechanisms of sorption will vary with the sorbent and the sorbate. The sorption of poorly water-soluble non-ionic organic compounds in soils is typically dominated by hydrophobic interactions. Sorption occurs primarily by partitioning into the soil organic carbon phase, and hence the extent of sorption is directly proportional to the soil organic carbon content (Chiou, et al. 1979, Chiou, et al. 1983). 53 The extent of sorption of non-ionic compounds can be acceptably predicted from the octanol / water partition coefficient of the compound and the organic content of the sorbent. The hydrophobic partitioning concept has been successful in describing sorption of neutral, hydrophobic organic compounds to solid materials containing amounts of natural organic material greater than 0.1% (Karickhoff 1984). The sorption of ionic compounds often can not be successfully predicted from the octanol / water partition coefficient and the organic carbon content of the sorbent. The sorption of ionic organic compounds is strongly influenced by solution pH and ionic strength (Schellenberg, et al. 1984). In addition to surface charge, surface functional groups, organic concentration, natural organic matter and temperature have all been shown to influence sorption (Schwarzenbach, et a]. 1992, Stone, et al. 1993, Stumm 1992) The availability of sorbed organic compounds to bacteria may depend on the mechanism of sorption. We have chosen a model system to explore the effects of various sorption mechanisms on bioavailability. The experimental matrix is comprised of two sorbates, three sorbents, and four strains of bacteria. One sorbate, 2,4- dichlorophenoxyacetic acid (2,4-D), is predominately in the anionic form at pH 7.0. The second sorbate is the methyl ester of the same compound (2,4-Dme) and remains uncharged in the model system. The simplest sorbent, uncoated silica, was chosen to represent a porous mineral surface. Silica coated with either phenyl or ethyl functional groups was used to simulate natural sorbents that contain organic phases. The four strains of bacteria used in the studies were isolated from Michigan agricultural soils and are capable of using either 2,4-D or the methyl ester as sole carbon and energy sources. 54 Sorption and desorption kinetics and extents were established for each combination of sorbent and sorbate. These results indicated the predominant mechanisms of sorption. Conclusions regarding the mechanisms of sorption were then related to the results of the bioavailability assays presented in Chapter 2 of this thesis. Materials and Methods Organisms Four strains of proteobacteria that are able to use 2,4-dichlorophenoxyacetic acid (2,4-D) or 2,4-dichlorophenoxyacetic acid methyl ester (2,4-Dme) as a sole carbon and energy source were selected for this study. Two of the strains, designated RASC and TFD6, were classified by McGowan (1994) based on 16S rDNA sequence analysis as being Burkholderia sp. Don and Pemberton (198]) previously classified the third strain, JMP134, was as Alcaligenes Eutrophus. The fourth strain,-F B4, was classified as F lavobacterium sp. by Ogram et a1. (1985). Each of these organisms internally metabolizes 2,4-D. Cells were grown until early stationary phase in a minimal salts media ([MSM] 1419.6 mg Na2HP04, 1360.9 mg KH2PO4, 0.3 mg (NH4)ZSO4, 50 mg MgSO4 7H20, 5.88 mg CaC12 2H20, 3.2 mg EDTA disodium salt, 2.78 mg FeSO4 7H20, 1.15 mg ZnSO4 7H20, 1.69 mg MnSO4 H20, 0.375 mg CuSO4 5H20, 0.233 mg Co(NO3)2 6H20 and 0.1236 mg (NH4)6Mo7024 4H20 per liter of distilled water) that contained 400 mg L'1 2,4-D. 55 Stock Solutions Sterile stock solutions of 14C-labeled 2,4-D (Sigma, 296 pure, 12.8 mCi mmol-l) and 2,4-Dme (Sigma, 296 pure, 13.6 mCi mmol'l) were prepared in phosphate buffered saline ([PBS] 8.5 g NaCl, 0.6 g Na2 P04, 0.3 g KH2P04, per liter of distilled water, pH 7.0). The solutions contained approximately 200,000 dpm ml'1 ring-U-[14C] 2,4-D or 2,4-Dme in addition to unlabeled compound for a final concentration of 10 mg L‘l. The solutions were filter sterilized and stored in five 100 ml portions in light shielded bottles at 4°C. Silica Slurries Uncoated, ethyl-coated and phenyl-coated silica were obtained from the J. T. Baker company. The irregularly shaped particles have a size of 40 pm with 60 A pores. According to the J. T. Baker company, surface coverage on the ethyl and phenyl-coated silicas is essentially complete. To create the silica Slurries the solids were wetted with PBS in a Speed Vac (DNA 110 Speed Vac, Savant Instruments, Inc., F armingdale, N.Y.). Silica (2.5 g) was weighed into a sterile 10 ml Falcon tube and 5 ml PBS were carefully layered on top. The tube was placed in a Speed Vac where gentle centrifugation and vacuum were applied simultaneously for 1 min. A disposable sterile loop was used to resuspend the silica. The slurry was then transferred to a 25 ml glass column with a glass frit. The silica was rinsed with 200 ml of PBS in 10 ml aliquots. The aliquots were pulled through the silica in a down-flow motion with a syringe connected to the bottom of the column. The solution level was never allowed below the silica surface to prevent drying of the silica. After rinsing, the silica was transferred to a 20 ml serum vial. PBS 56 and stock solutions containing radiolabeled 2,4-D or 2,4-Dme were added to achieve the appropriate slurry density (1 :4 w/v) and contaminant concentrations. The vials were laid on a platform rocker and mixed well enough to prevent the silica from settling. Sorption Kinetics Sorption kinetics were measured to determine the contact time necessary to reach an appropriate psuedo-equilibrium. Batch Slurries of silicas were used to measure decreases in solution phase concentrations of contaminant over time. Batch Slurries were sampled at time intervals of 0.25, 0.5, 1, 3, 6, 10, 18, 24, and 48 hr. Samples (0.4 ml) of homogenous slurry were transferred to microconcentrator tubes (Gellman, Nanosep MF Microconcentrators, 0.2pm pore size). The tubes were immediately centrifuged (Ependorf Centrifuge 5415C) at 11750 RCF for 1 min to separate the solution from the sorbent. After centrifugation the filter cup was separated from the tube. The tube cap was cut off. The filter cup and tube were dropped into separate scintillation vials containing 10 ml of scintillation cocktail fluid. Activity was counted on a Packard 1500 Tri-Carb Liquid Scintillation Analyzer for 5 min. Desorption Kinetics A desorption gradient was established by adding PBS to silica Slurries after psuedo-equilibrium sorption had been established. Change in solution phase contaminant concentration was measured over time. The slope of the initial linear portion of the desorption curve was used to establish the rate of desorption. Batch sorbent Slurries were established with initial solution contaminant concentrations of 1.0 mg L‘l. After sorption reached pseudo-equilibrium additional PBS was added to the Slurries to induce desorption. The batches were then sampled at time intervals of 20, 40, 55, 70, 90, 150, 57 210, 270, 330 and 960 seconds to determine solution phase concentrations as described previously. Distribution Coeflicients The sorbent-water distribution coefficients for 2,4-D and 2,4-Dme were determined for each sorbent type by a batch equilibration method. Silica Slurries were prepared as described above. The final Slurries contained 2.5 g of silica and 10 ml of PBS with a contaminant concentration of 1 mg L'1 that included the 14C-labeled tracer at 20,000 dpm ml'l. The final contaminant concentrations were achieved by putting 1 ml of the appropriate stock solution into each slurry. The same volume of stock solution was also put into control vials that contained 9 ml PBS and no sorbent. Three 1 ml volumes of each stock solution were put directly into scintillation vials containing 10 ml of scintillation cocktail to determine the total amount of contaminant in the slurry and control systems. The sorbent Slurries and controls were placed on a rocking platform for 45 minutes at a rate high enough to keep the silica suspended. After sorption, a homogenous 0.4 ml sample of each slurry was transferred to a microconcentrator tube (Gellman, Nanosep MF Microconcentrators, 0.2 pm pore size). The microconcentrator tube without the filter portion had been pre-weighed. The tubes were immediately centrifuged (Ependorf Centrifuge 5415C) at 11750 RCF for 1 min to separate the solution from the sorbent. After centrifugation the filter cup was separated from the via]. The vial was weighed with the solution sample and then the cap was cut off. The filter cup and via] were dropped into separate scintillation vials containing 10 ml of scintillation cocktail fluid. Activity was counted on a Packard 1500 Tri-Carb Liquid Scintillation Analyzer for 5 min. A second slurry sample was transferred to a pre-weighed Ependorf 58 tube. The tube was reweighed with the sample then oven dried at 40°C for 72 hrs before the dried sample was weighed. Solution concentration was calculated from the amount of compound in the centrifuged solution sample and the volume of sample as determined gravimetrically. Amount of sorbed contaminant was calculated by mass balance and used with the dry weight of silica to determine the sorbed concentration. Each Kd value is the mean of four independent determinations. Pore Volume The total pore volume in of each type of silica was determined during the silica rinsing procedure. The glass column was pre-weighed before the 2.5 g of silica was transferred into it as described in the “Silica Slurry” section. The column was weighed again with the solution level even with the silica level in the column. The total pore volume occupied by the solution was determined gravimetrically. The volume of solution retained with the silica after a slurry sample had been centrifuged in a microconcentrator tube at 11750 RCF for 1 min could have been calculated gravimetrically. Unfortunately not all of the pertinent data was collected. The volume of solution retained by the silica was instead calculated from a mass balance as described below. A mass balance of the radiolabeled compound in the slurry was created to determine the volume of solution retained with the silica afier centrifugation. Data from the distribution coefficient determination included: DS Amount of compound in the centrifuged silica sample (pg) MS Mass of the dried silica in the sample (g) D1 Amount of compound in the centrifuged solution sample (pg) V1 Volume of solution after sample centrifugation (ml) T Total amount of compound in the system (pg) 59 From these data a sorbed concentration that incorporates the solution retained after centrifugation can be calculated from the ratio of DS to MS. Likewise the solution concentration can be calculated from the ratio of D1 to V1. The total amount of compound (T), total mass of silica (M), and the total volume of solution (V) in the system are already known. Given this information the mass balance can be written and solved for the volume of solution retained with the silica (Vr) as in equations 1 and 2. Ds Dl T=—M+— V—V Ms Vl( r) (I) Vr—— —M+—V-T _ Dl Ms V1 V1 Ds Dl ( ) (2) Results Sorption kinetic experiments showed that sorption psuedo-equilibrium was established for all combinations of sorbate and sorbent in less than fifteen minutes. Sorption contact time was set to 45 minutes in all subsequent experiments. Desorption kinetics were also very rapid and could be considered instantaneous for our pru'poses (Figure 10). 60 Solution concentration (ug ml'l) 0 5 10 15 20 Time in minutes Figure 10. 2,4-Dme desorption from uncoated silica. Desorption for this sorbate- sorbent combination is representative of all combinations used in this study. Measured Kd values ranged from 0.39 to 1.25 ml g‘l (Table 6); 2,4-Dme was sorbed to a greater extent than 2,4-D for all sorbents tested. The fraction of sorbed contaminant that desorbed when solution phase concentrations were reduced is the fraction in equilibrium (feq). F eq values ranged from 1 to 0 as determined from the desorption isotherm and Kd values (Table 6). (See Chapter 2 for an explanation of feq determination.) In all cases where desorption occurred it was very fast relative to the rate of depletion from solution. The fraction of sorbed contaminant that was not in desorption equilibrium was irreversibly sorbed for the duration and conditions of our study. 61 Table 6. Kd and feq values for each combination of sorbent and contaminant Sorbent and Contaminant Type Uncoated silica Ethyl-coated silica Phenyl-coated silica 2,4-Dme 2,4-D 2,4-Dme 2,4-D 2,4-Dme 2,4-D Kd (ml/g) 0.73 £0.07 0.53 i018 0.76 $0.10 0.39 21:0.07 1.25 :I:0.13 0.77 $0.08 feq 0.44 3:00] 1.00 £0.00 0.00 i000 0.33 :I:0.28 0.1] 3:004 0.56 :L-0.01 Kd values represent the mean of four measurements 3: one standard deviation. feq values represent the mean of two measurements i one standard deviation. The total pore volume of a column containing 2.5 g of the ethyl or phenyl-coated silica was approximately 2 ml. The total pore volume of a column containing 2.5 g of uncoated silica was more than twice this amount (Table 7). The volume of solution retained by the silica after high-speed centrifugation was also calculated. The volume retained by the uncoated silica was equal to nearly half of the total pore volume. There was no measurable retained volume in the ethyl and phenyl coated silicas (Table 7). Table 7. Total pore volume and retained volume for each silica type. Uncoated silica Ethyl coated silica Phenyl coated silica Total pore volume (ml / 2.5 g) 4.7 2.1 2.2 Rimmed VOI‘m‘e 2.1 :1: 0.28 -0.24 :l: 0.16 -0.16 s 0.13 (ml / 2.5g) Total pore volume values represent a single measurement. Retained volume values represent the volume of solution retained by the silica after centrifugation at 11750 RCF for one minute. These values are the mean of eight calculated values :t one standard deviation. 62 Discussion Previous experiments on bioavailability were designed to test the hypothesis that degradation of sorbed contaminant is limited by the rate at which it desorbs into solution by mass action. If the kinetics of desorption are determined separately from the kinetics of degradation of solution phase contaminant, then the theoretical maximum degradation rate, assuming sorbed contaminant is unavailable to bacteria, can be predicted. If the contaminant degradation rate in the presence of a sorbent is greater than that predicted by the two independently measured processes then one possible conclusion is that bacteria at the sorbent surface facilitate degradation. Other conclusions including the stimulation of metabolism by attachment to a surface are discussed and eliminated in Chapter 2 of this dissertation. The highly significant (p S 0.01) result for six of the twenty-four bioavailability assays was that the contaminant degradation rate in the presence of a sorbent is greater than that predicted by the two independently measured processes. The bioavailability experiments are described in detail in Chapter 2. However, a review of the core of the bioavailability assays is important for relating mechanisms of sorption to bioavailability. In simple liquid batch culture, first-order degradation kinetics are observed at substrate concentrations below KS (Alexander and Scow 1989, Horvath 1972). Under first-order conditions biodegradation rate is directly proportional to the bioavailable concentrations. Therefore a decrease in the instantaneous degradation rate can be expected when the bioavailable concentration decreases. The addition of a sorbent to the solution would reduce the solution phase concentration. However, substrate can be expected to desorb as solution phase degradation proceeds. The bioavailability 63 experiments tested whether degradation of sorbed contaminant was limited by the rate at which it desorbs into solution by mass action. The results of the bioavailability assays showed that, in some cases, the rate of contaminant degradation exceeded the rate predicted from solution phase concentrations even when desorption was accounted for. More specifically, the following model describes the degradation of a contaminant in the presence of a sorbent: T = total mass of contaminant in system (1) T = M1 + M2 + M3 M1 = mass of contaminant sorbed to silica M2 = mass of contaminant in solution (2) M1 k1 M2 k2 M3 —> —> M3 = mass of contaminant as C02 k3 T The first-order rate constant which describes desorption of the contaminant from the sorbent into solution is k1. The desorption rate constant was determined by allowing contaminant to sorb, then reducing the bulk solution contaminant concentration and measuring the concentration of contaminant in the solution over time. The frrst-order rate constant that describes the degradation of solution phase contaminant is kg. The degradation rate constant was determined in a liquid-batch system by measuring the disappearance of solution phase contaminant over time. The rate constant that describes the degradation of sorbent associated contaminant is k3. This rate was not independently measured. When desorption is much slower than degradation (k1<>k2) equilibrium conditions prevail. In both latter cases the rate of degradation in the presence of a sorbent is influenced by desorption of substrate into the bulk solution. Therefore RI and k2 must be determined independently in order to evaluate the possibility of direct access to sorbed contaminant by bacteria. The rate of desorption in all of the model systems tested was much faster than the rate of solution phase degradation (k]>>k2). Equilibrium conditions prevailed in all cases. The relatively short sorption contact time likely limited sorption mechanisms to surface related phenomena. However, sorption and desorption extents varied. Sorption extent is expressed by the distribution coefficient, Kd- The model contaminants were chosen to provide a range of sorption extents. This was accomplished by the increasing hydrophobicity of the silica coating. Kd was greatest for both sorbates on the phenyl- coated silica and nearly as high for 2,4-Dme on both uncoated and ethyl coated silica. Kd was lowest for 2,4-D sorption to uncoated and ethyl coated silica (Table 6). Desorption extent is expressed by feq which represents the fraction of the sorbed compound that is in desorption equilibrium. There were no cases of desorption non- equilibrium in our model systems. The fraction of sorbed compound that did not instantaneously desorb as solution phase concentrations were depleted was irreversibly bound for the time scale and conditions of our experiments. Feq values ranged from 0 to 65 1. This is the broadest possible range of feq values. An feq value of 1 represents a system in which all the sorbed contaminant is in instantaneous equilibrium with the bulk solution. An feq value of 0 represents a system in which all the sorbed contaminant is irreversibly sorbed. Feq values were always higher for the anionic 2,4-D than for the non-ionic 2,4-Dme on each sorbent type. Feq values were highest for the uncoated silica and lowest for the ethyl-coated silica. Desorption has been described by several kinetic models including the two box and radial diffusion models (Karickhoff 1980, Van Genuchten and Wierenga 1976, Wu and Gschwend 1986). In the two box model, desorption is conceptualized as occurring by rapid release from an easily accessible exterior followed by a kinetically limited desorption from a less accessible interior. The latter process may be conceptualized as intraorganic matter diffusion. Contaminants within this domain would likely be biologically unavailable. According to the radial diffusion model, contaminants diffuse through pore fluids in the interstices of soil particles, and this process is retarded by microscale partitioning between fluids and intraparticle solids. Organic compounds dissolved in the soil solution may occupy soil micropores smaller than one micrometer, too small for bacteria to enter. The irregularly shaped silica particles used as model sorbents in these experiments have a size of 40 pm with 60 A pores. The volume of interstitial fluid in columns containing 2.5 g of any of the sorbents should be nearly equal due to the shared average dimensions of the sorbent particles. Determination of the total pOre volume of a column packed with each of the three sorbent types showed that uncoated silica had more than twice the total pore volume of the two coated silicas. If the volume of interstitial 66 fluid is assumed to be nearly equal based on the reasons given above then the greater total pore volume in the uncoated silica column suggests the presence of intra-particle pores accessible by the aqueous solution. These pores may have been occluded when the uncoated silica was covered with the hydrophobic coatings or the presence of the coatings may prevent aqueous solution from entering the 60 A pores. In addition calculations of the volume of solution retained by the uncoated silica after high-speed centrifugation accounted for approximately half of the total pore volume. The coated silicas retained no measurable solution after centrifugation. These data support the concept that the uncoated silica retains solution in intraparticle pores that is not retained b the coated silicas (Table 6). The Kd values as calculated in Table 7 do not consider compound in retained solution sorbed. However, compound that is in intraparticle pores of the sorbent is inaccessible to bacteria because of size exclusion. This renders it distinctly different from compound in the bulk solution. Table 8 presents a modified Kd (Kd*) that incorporates the compound retained in solution after centrifugation as sorbed. This is a more appropriate Kd for assessing bioavailability because contaminant in solution in these pores in not directly bioavailable. The Kd values calculated in Table 7 can be modified as shown in equation 3. D1 (Ds+Vr0—V-l—)éfl Kd* = Ms VI (3) 67 Uncoated silica is the only sorbent type that has a measurable volume of solution retained after centrifugation. Therefore uncoated silica is the only sorbent that has significantly different values for Kd and the retained-volume-modified Kd“. A large portion of the sorption of either contaminant in the uncoated silica is due to contaminant in solution that is retained by the uncoated silica. Sorption on the coated silicas, especially of 2,4-Dme, was likely limited to hydrophobic interactions with the surface coatings. Table 8. Kd and Kd* for each combination of sorbent and contaminant Sorbent and Contaminant Type Uncoated silica Ethyl-coated silica Phenyl-coated silica 2,4-Dme 2,4-D 2,4-Dme 2,4-D 2,4-Dme 2,4-D Kd (ml g_1) 0.73 4007 0.53 3:0.18 0.76i0.10 0.39s007 1.25 e013 0,773,003 Kd* (mlg_1) 15610.03 1.37:0.02 0.711009 0.311003 1.14s0.07 0.69i0.06 Kd“ is the previously calculated Kd modified to incorporate contaminant in the retained solution as sorbed. Kd“ values represent the mean of four measurements i one standard deviation. According to the two box model of contaminant sorption to soil, there is a labile sorbed phase in instantaneous equilibrium with the aqueous phase (Van Genuchten and Wierenga 1976). Reduction of surface contaminant concentrations via degradation promotes the desorption of contaminant held within the sorbent aggregate. Potentially this process could efficiently provide contaminant to bacteria on the sorbent surface without the desorbing contaminant ever being completely diluted in the bulk solution. 68 This mechanism could account for rates of solution phase depletion greater than the rates predicted by the BFM. In this scenario the systems that have the greatest amount of sorbed contaminant in desorption equilibrium are most likely to result in enhanced rates of contaminant depletion. Figure 1] demonstrates that this trend was observed in the bioavailability assays. The frequency of enhanced degradation rate increases as the amount of sorbate in desorption equilibrium increases. 4 '2; t “5 i 5, 21 c: d) :3 8‘ E l ! othfiw--- . .- . Desorneq. (ug) 1.02 0.49 0.33 0.10 0.09 0.00 Sorbent: Uncoated Uncoated Phenyl ctd Phenyl ctd Ethyl ctd Ethyl ctd Sorbate: 2,4-D 2,4-Dme 2,4-D 2,4-Dme 2,4-D 2,4-Dme Figure 11. Frequency of enhancement (white bar) and inhibition (black bar) of contaminant degradation rate relative to the rate predicted by the BFM using Kd*. In six of the twenty four bioavailability assays the rate of contaminant degradation was underestimated from the rate of desorption and bulk solution phase degradation. In these cases the rate of degradation in the presence of the sorbent exceeded the theoretical maximum predicted by the rates of desorption and degradation of solution phase 69 contaminant at a highly significant level (p S 0.01). Uncoated silica was the sorbent in five of the six cases of enhanced degradation rate (Figure 11). We propose that contaminant present in the solution retained by uncoated silica was readily available to bacteria on the silica surface as indicated by the high amount of contaminant in desorption equilibrium. In contrast, the bioavailability of contaminant sorbed to hydrophobically coated silica was limited by the rate of desorption into bulk solution. The rate of contaminant degradation in the presence of a sorbent was accurately predicted by the sum of desorption and solution phase degradation in twelve of the sixteen bioavailability assays that involved contaminant sorption to hydrophobically coated silicas. In these cases the presence of the sorbent had no effect on degradation that could not be accounted for by desorption and solution phase degradation (Figure 11). There were also three bioavailability assays that resulted in degradation rates significantly (p S 0.01) slower than what was predicted if only bulk phase contaminant was available. These three cases all involved ethyl-coated silica as a sorbent (Figure 11). Conclusions The results of the bioavailability assays in conjunction with what we learned about the sorption mechanisms lead us to propose that reversibly sorbed contaminant present in uncoated silica pores may be available to bacteria without ever being completely dissolved in the bulk solution. The bioavailability of contaminant sorbed by hydrophobic interactions on the coated silica surfaces was not greater than the rate predicted by the independent processes of desorption and solution phase degradation. 70 Acknowledgements This research was supported by NSF grant DEB9120006, USDA NRICGP No. 94-371 07-03 86, and the Michigan Agricultural Experiment Station. Literature cited Alexander, M., and K. M. Scow. 1989. Kinetics of biodegradation in soil, p. 243-269. In B. L. Sawhney and K. Brown (ed.), Reactions and Movement of Organic Chemicals in Soils. Soil Science Society of America and American Society of Agronomy, Madison, WI. Chiou, C. T., L. J. Peters, and V. H. Freed. 1979. A physical concept of soil-water equilibria for nonionic organic compounds. Science. 206:831-832. Chiou, C. T., P. E. Porter, and D. W. Schmedding. 1983. Partition equilibria of nonionic organic compounds between soil organic matter and water. Environ. Sci. Technol. 17:227-231. Don, R. H., and J. M. Pemberton. 1981. Properties of six pesticide degradation plasmids isolated from Alcaligenes paradoxus and Alcaligenes eutrophus. J. Bacteriol. l45(2):681-686. Horvath, R. S. 1972. Microbial co-metabolism and the degradation of organic compounds in nature. Bacteriol. Rev. 36(2): 146-155. Karickhoff, S. W. 1980. Sorption kinetics of hydrophobic pollutants in natural sediments, p. 193—205. In R. A. Baker (ed.), Contaminants and Sediments, vol. 2. Ann Arbor Press, Ann Arbor, MI. Karickhoff, S. W. 1984. Organic pollutant sorption in aquatic systems. Environ. Toxicol. Chem. 4:469-479. McGowan, C. 1994. Interspecies gene transfer in the evolution of 2,4- dichlorophenoxyacetic acid degrading bacteria. Ph.D. thesis. Michigan State University, East Lansing. Ogram, A. V., R. E. Jessup, and P. S. C. Rao. 1985. Effects of sorption on biological degradation rates of (2,4-dichlorophenoxy)acetic acid in soils. Appl. Environ. Microbiol. 49:582-5 87. Schellenberg, K., C. Leuenberger, and R. P. Schwarzenbach. 1984. Sorption of chlorinated phenols by natural sediments and aquifer materials. Environ. Sci. Technol. l8(9):652-657. 7] Schwarzenbach, R. P., P. M. Gschwend, and D. M. Imboden. 1992. Environmental Organic Chemisny. John Wiley & Sons, New York. Stone, A. T., A. Torrents, J. Smolen, V. D., and J. Hadley. 1993. Adsorption of organic compounds possessing ligand donor groups at the oxide/water interface. Environ. Sci. Technol. 27:895-909. Stumm, W. 1992. Chemistry of the Solid-Water Interface: Processes at the Mineral- Water and Particle-Water Interface in Natural Systems. John Wiley & Sons, New York. Van Genuchten, M. T., and P. J. Wierenga. 1976. Mass transfer studies in sorbing porous medial. Analytical solutions. Soil Sci. Soc. Am. J. 40(4):473-479. Westall, J. C. 1987. Adsorption mechanisms in aquatic surface chemistry. In W. Stumm (ed.), Aquatic Surface Chemistry: Chemical Processes at the Particle-Water Interface. Wiley, New York. Wu, S. C., and P. M. Gschwend. 1986. Sorption kinetics of hydrophobic organic compounds to natural sediments and soils. Environ. Sci. Technol. 20:717-725. 72 Chapter 5 SORBENT IMPACT ON SOLUTION pH AND SOLUBLE CARBON AFFECT SOLUTION PHASE DEGRADATION OF 2,4-DICHLOROPHENOXYACETIC ACID Abstract Environmental conditions affecting biodegradation of a given compound include pH, ionic strength, soluble carbon, biomass, nutrients and temperature. The objective of this study was to evaluate how silica sorbents affect solution phase depletion of 2,4- dichlorophenoxyacetic acid (2,4-D) vis-a-vis their effects on solution properties. First we independently determined the affect of three sorbents on solution pH and soluble carbon. Second, the rate of solution phase depletion in each sorbent supernatant was compared to rate of depletion in phosphate buffered saline (PBS). In addition the affect of solution pH on rate of depletion was separately determined for four strains of bacteria. Finally a bioavailability assay was used to demonstrate a method to isolate the effects of bioavailability of sorbed contaminants versus indirect effects of the sorbent on solution phase depletion. The pH of PBS was essentially unchanged in the ethyl and phenyl- coated silica supernatants but it decreased appreciably after contact with uncoated silica. All four strains of bacteria showed increased rates of solution phase 2,4-D degradation with decreasing pH over a range of 7.0-5.0. The soluble carbon in the uncoated-silica supernatant was not different from the PBS. Ethyl and phenyl-coated silica supematants had elevated levels of soluble carbon. Solution phase depletion of 2,4-D was effected by the properties imparted to the PBS by all of the silica types. 73 Introduction The bioavailability experiments described in earlier chapters have offered evidence that sorbed contaminants may be available for biological degradation at a rate faster than desorption into bulk solution would permit. These conclusions are based on comparisons of rates of contaminant depletion in the presence of a sorbent to those predicted based on the contaminant mass initially in solution plus that which desorbed during the time course of the assay. Rates of contaminant desorption and depletion from solution in the absence of sorbent were measured independently. Experiments were carefully controlled to insure that the composition of the aqueous phase in sorbent-free controls was as similar as possible to that of the silica Slurries. This condition was essential for determining the bioavailability of sorbed compounds. Degradation of a compound in the presence of a solid matrix can be affected by factors other than sorption such as pH, ionic strength, soluble carbon and nutrients. Bacteria vary widely in their pH tolerance (Doetch and Cook 1973). Given that other conditions remain suitable, most bacteria can survive in a range of 6 to 9 pH (Atlas 1995). The extreme examples Show bacteria surviving at pH values as low as 0, Thiobacillus thioxidans, and as high as 13, Plectonema nostocorum (Ingledew 1990; Kroll 1990). Furthermore, pH changes can impact overall cellular activity by modifying ' the kinetic properties of the enzymes. Enzymes are affected by pH through alterations in protein structure which influences stability as well as catalytic reactivity (Phillips 1994). Corbin and Upchurch (Corbin and Upchurch 1967) showed that 2,4-D had a maximum rate of detoxification at pH 5.3 in soil. Detoxification occurred at a much slower rate in pH 7.5 soils. 74 The presence of additional soluble carbon can also impact the depletion of a target compound. In some cases, additional substrates have enhanced cell growth or enhanced adaptation to utilization of the target compound in mixed culture experiments, thus lowering the final concentration of the target compound after treatment (Bouwer and McCarty 1985; Kim and Maier 1986; Rittrnann 1987; Bouwer, Mercer et al. 1988). In other cases, the presence of alternative substrates has inhibited or repressed the biodegradation pathway or led to the selection of strains that lack the necessary pathways for degradation (F rick, Crawford et al. 1987; Topp, Crawford et al. 1988). Daugherty and Karel (Daugherty and Karel 1994) showed that the rate of 2,4—D degradation was affected by the presence of succinate. Repression of 2,4-D degradation was the dominant factor at low ratios of 2,4-D to succinate. However, the specific rate of 2,4-D degradation remained at values equivalent to or higher than the specific rate for cells grown on 2,4-D alone when the ratio of 2,4-D to succinate was high. Similar effects were seen in studies on the effect of glucose on 2,4-D degradation (Papanastasiou and Maier 1982) and of nutrient broth on 2,4-D degradation in mixed cultures (Kim and Maier 1986). The experiments described here were designed to explore how properties of the solution imparted by the sorbent can effect solution phase depletion of 2,4- dichlorophenoxyacetic acid (2,4-D). First we independently determined the effect of three sorbents on solution pH and soluble carbon. Second, the rate of solution phase depletion in each sorbent supernatant was compared to rate of depletion in standard phosphate buffered saline (PBS). In addition the affect of solution pH on rate of depletion was independently determined for four strains of bacteria over a range of pH 75 and subsequently compared to the experimental data. Finally a bioavailability assay was used to demonstrate a method to isolate the effects of bioavailability of sorbed contaminants versus indirect effects of the sorbent on solution phase depletion. We found that biodegradation of sorbed phase contaminant can have as large an impact on overall in-situ bioremediation as changes in the solution pH. Materials and Methods Organisms Four strains of proteobacteria that are able to use 2,4-dichlorophenoxyacetic acid (2,4-D) as a sole carbon and energy source were selected for this study. Two of the strains, designated RASC and TF D6, were classified by McGowan (McGowan 1994) based on 16S rDNA sequence analysis as being Burkholderia sp. Don and Pemberton (Don and Pemberton 1981) previously classified the third strain, JMP134, was as Alcaligenes Eutrophus. The fourth strain, F B4, was classified as F lavobacterium sp. by Ogram et al. (Ogram, Jessup et al. 1985). Each of these organisms internally metabolizes 2,4-D. Cells were grown until early stationary phase in a minimal salts media ([MSM] 1419.6 mg NazHPO4, 1360.9 mg KH2P04, 0.3 mg (NH4)ZSO4, 50 mg MgSO4 7H20, 5.88 mg CaC12 2H20, 3.2 mg EDTA disodium salt, 2.78 mg FeSO4 7H20, 1.15 mg ZnSO4 7H20, 1.69 mg MnSO4 H20, 0.375 mg CuSO4 5H20, 0.233 mg Co(NO3)2 6H20 and 0.1236 mg (NH4)6Mo7024 4H20 per liter of distilled water) that contained 400 mg L-1 2,4-D. 76 Stock Solutions Sterile stock solutions of 14C-labeled 2,4-D (Sigma, 296 pure, 12.8 mCi mmol'l) were prepared in phosphate buffered saline ([PBS] 8.5 g NaCl, 0.6 g Nag P04, 0.3 g KH2P04, per liter of distilled water, pH 7.0). The solutions contained approximately 200,000 dpm ml-1 nng-U-[14C] 2,4-D in addition to unlabeled compound for a final concentration of 10 mg L'l. The solutions were filter sterilized and stored in five 100 ml portions in light shielded bottles at 4°C. Silica Slurries and Solutions Uncoated, ethyl-coated and phenyl-coated silica were obtained from the J. T. Baker company. The irregularly shaped particles have a size of 40 pm with 60 A pores. According to the J .T. Baker Company, surface coverage on the ethyl and phenyl-coated silicas is complete. To create the silica Slurries the solids were wetted with PBS in a Speed Vac (DNA 110 Speed Vac, Savant Instruments, Inc., Farmingdale, NY.) Enough silica for one slurry was weighed into a sterile 10 ml Falcon tube and 5 ml PBS was carefully layered on top . The tube was placed in a Speed Vac where gentle centrifilgation and vacuum were applied simultaneously for 1 min. A disposable sterile loop was used to resuspend the silica. The silica was transferred to a 20 ml serum vial. Silica Slurries were established with varying silica to solution ratio and varying amount of total 2,4-D for depletion assays. PBS and stock solution containing radio- labeled 2,4-D were added to the wetted silica to achieve the appropriate slurry density and contaminant concentrations. Teflon stoppers were crimp capped onto the vials. The vials were laid on a platform rocker and mixed well to prevent the silica from settling. 77 After 45 minutes the Slurries were used for depletion assays as described in the following section of methods. Solution from silica Slurries was collected for pH and soluble carbon measurements and for sorbent free depletion assays. PBS was added to the wetted silica achieve a slurry density of 1:4 w/v. The vials were laid on a platform rocker and mixed well to prevent the silica from settling. After 45 minutes the Slurries were filtered through a 0.2 pm filter and the solution was retained. Total carbon and pH of the slurry solution were determined. Total carbon was determined on a Dorhrmann DC-190 High- Temperature TOC Analyzer. The appropriate volume of stock solution containing radio- labeled 2,4-D was added to the filtered solution for a final 2,4-D concentration of 1 mg L' 1. This solution was then used in the depletion assays as described in the following section of methods. The final set of samples used in the depletion assays contained PBS with no sorbent. The PBS was adjusted to a pH of 5.0, 6.0 or 7.0. These samples were used to determine the effect of pH on solution phase 2,4-D depletion rates. Depletion Assays Depletion assays determined the rate of contaminant depletion from solution in the presence or absence of a sorbent. Depletion of solution phase ring-U-[14C] 2,4-D at cell densities of approximately 1X108 CFU ml‘1 was monitored over time. Control 20 ml serum vials contained 10 ml PBS and contaminant at 1.0 mg L'1 with approximately 20,000 dpm ml"1 but no silica. Experimental serum vials were established as described in the previous section of methods. Vials were sampled at 6, 13, 20, 27, 34, and 41 min 78 after inoculation. Sampling consisted of taking 0.4 ml solution by syringe. This volume was transferred to a microconcentrator tube (Gelhnan, Nanosep MF Microconcentrators, 0.2pm pore size). The tube was immediately centrifuged (Ependorf Centrifuge 5415C) at 11750 RCF for l min to separate the cells from the solution. 40 pl of 2N HCl were present at the bottom of the microconcentrator tube and effectively drove off any carbon dioxide present in the solution sample. After centrifugation the filter cup was separated from the tube and the cap was cut off the tube. The filter cup and via] were placed in separate scintillation vials containing 10 ml of scintillation cocktail fluid. Activity was counted on a Packard 1500 Tri-Carb Liquid Scintillation Analyzer for 5 min. Solution phase concentrations of 2,4-D were plotted versus time. Linear regression was used to estimate the rate of depletion for each initial solution concentration. Modeling In brief, results from previous partition coefficient and desorption kinetic studies were used to determine the proportions of substrate in equilibrium, non-equilibrium, and irreversible sorption sites. These proportions were then used to calculate a bioavailability factor for the particular contaminant/sorbent combination. The bioavailability factor was used to model the expected rate of substrate disappearance from solution when a sorbent is present according to the Bioavailability Factor Model (see Chapter 2). This calculated expected rate of substrate disappearance from solution, which takes desorption extents and rates into consideration, was compared to the rates measured in the presence of a sorbent to determine if the system was desorption limited. 79 Results and Discussion Phosphate buffered saline (PBS) was used to wet each of the silica sorbents and make the silica Slurries. The pH of PBS was 6.9. The pH was essentially unchanged in the ethyl and phenyl-coated silica supematants but it decreased appreciably afier contact with uncoated silica (Table 9). This pH change was likely due to the dissociation of hydrogen ions from the uncoated-silica surface. Table 9 shows that there was some carbon in the PBS sample. The soluble carbon in the uncoated-silica supernatant was not different from the PBS. However, ethyl and phenyl-coated silica supematants had elevated levels of soluble carbon. The low level of carbon measured in the PBS may have been obtained from the scintillation vials used to transport the samples. Regardless, all samples were transported in the same type of via] so relative comparisons are meaningful. Table 9. pH and soluble carbon measurements for PBS and silica supematants PBS Uncoated-silica Ethyl-coated Phenyl-coated supernatant silica supernatant silica supernatant pH 6.92 5.76 6.89 6.99 Soluble carbon (ppm) 76 76 103 240 All values represent the mean of two samples. Rates of 2,4-D depletion from solution by each of four strains were measured in PBS and each of the silica supematants. Solution phase depletion of 2,4-D was effected by the properties imparted to the PBS by all of the silica types (Table 10). The rate of 2,4-D depletion from uncoated silica supernatant was significantly higher than the rate in PBS for three of the four strains. Ethyl-coated silica supernatant resulted in enhancement 80 for one strain and no significant effect for the others. Phenyl-coated silica supernatant resulted in one case of significant inhibition and one case of enhancement. Table 10. Relative rates of solution phase 2,4-D depletion in silica supematants Strain Supernatant source FB4 TFD6 JMP134 RASC Uncoated silica 1.72m": 1.13 1.66* 1.58“ Ethyl-coated silica 1.18 0.99 1.57M 1.07 Phenyl-coated silica 1.06 0.69" 1.86M 0.96 a Ratio of the measured rate of depletion from solution in the silica supernatant to the rate in PBS. ** p S 0.01 * p 5 0.025 Level of significance for the difference between the rate measured in the supernatant and in PBS. To separate the effect of solution pH from other supernatant effects the rate of solution phase 2,4-D depletion was measured in PBS at pH 5.0, 6.0 and 7.0 for all four strains. All four strains of bacteria showed increased rates of solution phase 2,4-D degradation with decreasing pH (Figure 12). The solution phase degradation rate at pH 5.0 was as much as three hundred percent greater than the rate at pH 7.0 for strain JMP134 or as little as twenty five percent greater for strain TFD6. 81 o " 0— F B4 E 3'0 l -A— TFD6 ,,/E’ g 25 1 43— JMP /o '0 I{£::f'vg_.-»-~A"" «3 ”0* RASC / g 2.0 ’ ”wiggl-I ”’- "/ /// "O /.., A____/_’_ ___’_V#_.. e 1 , / V" l 5 AM - - - ~-—— A m p - -f I“- '0 l :1; __ ::#.:~:;—— ——~"' W T T T # flfi g 1 0 F- '7‘3 l E 0.5 -1 o 1 Z 1 0.0 '- -~ . — , _ fl 7 6 5 Solution pH Figure 12. Normalized rates of solution phase 2,4-D depletion by four bacterial strains at pH 7.0, 6.0, and 5.0. Rates were normalized to the rate measured at pH 7.0 to allow comparison between strains. The design of the bioavailability assay experiments can be adjusted to reveal effects on solution phase depletion rates that are due to sorbent impact on the solution rather than availability of sorbed substrate. This is an alternative to independent measurement of the effects of solution pH and soluble carbon. This approach was tested with the model system of uncoated silica, 2,4-D and FB4. In the first experiment the total amount of substrate in the system was varied. In the second experiment the silica to solution ratio was varied. AS a consequence the resulting solution phase and sorbed phase 2,4-D concentrations varied in both experiments. This experimental design made it possible to determine if the changes in 2,4-D depletion rate were the result of the final 82 equilibrium concentrations or the manner in which these concentrations were achieved (i.e. the increased silica to solution ratio or the increased total amount of 2,4-D). If the enhancement of the 2,4-D depletion rate had been due to the final equilibrium concentrations rather than the manner in which these concentrations were achieved then the results of the two experiments would have been similar. Figure 13 shows that the two experiments did not have similar trends. Therefore the enhancement of the 2,4-D depletion rate was not solely dependent on the sorbed phase concentration. When the total amount of 2,4-D in the system was varied the degree of rate enhancement remained relatively consistent. However, as the silica to solution ratio was increased the extent of rate enhancement correspondingly increased. In this model system the increasing ratio of uncoated silica to solution caused a decrease in the resulting solution pH. The bacterial strain F B4 showed enhanced rates of solution phase 2,4-D depletion at decreased pH. If the bioavailability assay had been done at only one silica to solution ratio without strict attention to the composition of the sorbent-free control solution the enhancement of the 2,4-D depletion rate could have easily been attributed to bioavailability of sorbed 2,4-D. 83 o Varying total 2,4-D 1 A Varying silica to soln. ratio Ratio of measured to predicted rate of solution phase 2,4-D depletron DJ 0 01.__——-.—— . . . . . 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Sorbed 2,4-D concentration 03pm) Figure 13. Enhancement of solution phase 2,4-D depletion rate is demonstrated by a value greater than one for the ratio of the measured rate of depletion to the rate of depletion predicted by the Bioavailability Factor Model. This ratio is shown relative to sorbed 2,4-D concentration for two experimental designs. Conclusions Many studies on the bioavailability of sorbed compounds exist in the literature (Moyer, Hance et al. 1972; Steen, Paris et a1. 1980; Shimp and Young 1988; Mihelcic and Luthy 1991; Guerin and Boyd 1992; Scribner, Benzing et al. 1992; Weissenfels, Klewer et al. 1992; Al Bashir, Hawari et al. 1994; Radosevich, Traina et al. 1997; Ortega- Calvo and Saiz-Jimenez 1998). Unfortunately some of them did not control for the impact that different sorbents can have on the composition of the associated solutions and the resulting effect on the overall degradation rate. This study demonstrates that a lack of strict controls to ensure that the composition of the sorbent free control solution is as 84 close as possible to the solution associated with the sorbent can lead to misguided conclusions about bioavailability of sorbed contaminants. Mathematical modeling is used to understand and optimize processes of in-situ biodegradation. These models rely on comprehensive descriptions of the processes that occur in dynamic natural environments. Solution pH is recognized as an important factor and incorporated in these models. We measured the noteworthy effect that a pH shift can have on the degradation rate of 2,4-D. In comparison to previous bioavailability experiments we found that biodegradation of sorbed phase contaminant can have as large an impact on overall in-situ bioremediation as changes in the solution pH. This potentially sizeable contribution to bioremediation success Should be incorporated into future mathematical models. Acknowledgements This research was supported by NSF grant DEB9120006, USDA NRICGP No. 94-37107-03 86, and the Michigan Agricultural Experiment Station. Literature Cited Al Bashir, B., J. Hawari, R. Samson and R. Leduc 1994. Behavior of nitrogen- substituted naphthalenes in flooded soil: Part II. Effect of bioavailability on biodegradation kinetics. Water Res. 28(8): 1827-1833. Atlas, R. M. 1995. Bacterial Growth and Reproduction: Effects of Acidity and pH. Principles of Microbiology. J. M. Smith and R. J. Callanan. St. Louis, Mosby-Year Book Inc.: 346-347. Bouwer, E., J. Mercer, M. Kavanaugh and F. DiGiano 1988. Coping with groundwater contamination. J. Water Pollut. Control Fed. 60: 1415-1427. Bouwer, E. J. and P. L. McCarty 1985. Utilization rates of trace halogenated organic compounds in acetate-grown biofilms. Biotechnol. Bioeng. 27: 1564-1571. 85 Corbin, F. T. and R. P. Upchurch 1967. Influence of pH on detoxification of herbicides in soil. Weeds 15: 370-377. Daugherty, D. D. and S. F. Karel 1994. Degradation of 2,4-dichlorophenoxyacetic acid by pseudomonas cepacia DB01(pR0101) in a dual-substrate chemostat. Appl. Environ. Microbiol. 60(9): 3261-3267. Doetch, R. N. and T. M. Cook 1973. Introduction to Bacteria and their Ecobiology. Baltimore, University Park Press. Don, R. H. and J. M. Pemberton 1981. Properties of six pesticide degradation plasmids isolated from Alcaligenes paradoxus and Alcaligenes eutrophus. J. Bacteriol. 145(2): 681-686. Frick, T. D., R. L. Crawford, M. Martinson, T. Chresand and G. Bateson 1987. Microbiological cleanup of groundwater contamination by pentachlorophenol. Environmental Biotechnology. G. S. Omenn. New York, Plenum Press: 173-191. Guerin, W. F. and S. A. Boyd 1992. Differential bioavailability of soil-sorbed naphthalene to two bacterial species. Appl. Environ. Microbiol. 58(4): 1142-1152. Ingledew, W. J. 1990. Acidophiles. Microbiology of Extreme Environments. C. Edwards. New York, McGraw-Hill: 33-54. Kim, C. J. and W. J. Maier 1986. Acclimation and biodegradation of chlorinated organic compounds in the presence of alternate substrates. J. Water Pollut. Control Fed. 58: 157-164. Kroll, R. G. 1990. Alkalophils. Microbiology of Extreme Environments. C. Edwards. New York, McGraw-Hill: 55-92. McGowan, C. 1994. Interspecies gene transfer in the evolution of 2,4- dichlorophenoxyacetic acid degrading bacteria. East Lansing, Michigan State University. Mihelcic, J. R. and R. G. Luthy 1991. Sorption and microbial degradation of naphthalene in soil-water suspensions under denitrification conditions. Environ. Sci. Technol. 25(1): 169-177. Meyer, J. R., R. J. Hance and C. E. McKone 1972. The effect of adsorbents on the rate of degradation of herbicides incubated with soil. Soil Biol. Biochem. 4: 307-311. Ogram, A. V., R. E. Jessup and P. S. C. Rao 1985. Effects of sorption on biological degradation rates of (2,4-dichlorophenoxy)acetic acid in soils. Appl. Environ. Microbiol. 49: 582-587. 86 Ortega-Calvo, J. J. and C. Sail-Jimenez 1998. Effect of humic fractions and clay on biodegradation of phenanthrene by a Pseudomonasfluorescens strain isolated from soil. Appl. Environ. Microbiol. 64(8): 3123-3126. Papanastasiou, A. C. and W. J. Maier 1982. Kinetics of biodegradation of 2,4- dchlorophenoxyacetate in the presence of glucose. Biotechnol. Bioeng. 24: 2001-2011. Phillips, A. T. 1994. Enzyme Activity: pH. Methods for General and Molecular Bacteriology. P. Gerhardt and W. A. Wood. Washington, American Society for Microbiology: 579-580. Radosevich, M., S. J. Traina and O. H. Tuovinen 1997. Atrazine mineralization in laboratory-aged soil microcosms inoculated with s-triazine-degrading bacteria. J. Environ. Qual. 26(1): 206-214. Rittmann, B. E. 1987. Aerobic biological treatment. Environ. Sci. Technol. 21: 128-136. Scribner, S. L., T. R. Benzing, S. Sun and S. A. Boyd 1992. Desorption and bioavailability of aged simazine residues in soil from a continuous corn field. J. Environ. Qual. 21(1): 115-120. Shimp, R. J. and R. L. Young 1988. Availability of organic chemicals for biodegradation in settled bottom sediments. Ecotoxicol. Environ. Saf. 15: 31-45. Steen, W. C., D. F. Paris and G. L. Baughman 1980. Effect of sediment sorption on microbial degradation of toxic substances. Ann Arbor, MI, Ann Arbor Sci. Publ. Topp, E., R. L. Crawford and R. S. Hanson 1988. Influence of readily metabolizable carbon on pentachlorophenol metabolism by a pentachlorophenol degrading F lavobacterium sp. Appl. Environ. Microbiol. 54: 2452-2459. Weissenfels, W. D., H. J. Klewer and J. Langhoff 1992. Adsorption of polycyclic aromatic hydrocarbons (PAHs) by soil particles: Influence on biodegradability and biotoxicity. Appl. Microbiol. Biotechnol. 36(5): 689-696. 87 Appendix LESSONS LEARNED FROM EVALUATION OF BIOAVAILABILITY OF 2,4- DICHLOROPHENOXYACETIC ACID SORBED TO SOIL Introduction The potential impact that sorption limited bioavailability may have on the success of bioremediation technologies has been stressed throughout this thesis. Some of the important factors affecting bioavailability of sorbed contaminants have been elucidated with the use of a simplified model system that included silica sorbents. The design of these experiments came after attempts to evaluate the bioavailability of 2,4-D sorbed to complex natural soils. This appendix presents those findings and the valuable lessons learned from those experiments. The 2,4-D mineralization rate in soil Slurries, with a portion of 2,4-D sorbed to soil, was determined and compared to the mineralization rate predicted from soil free controls with an equal aqueous concentration of 2,4-D. The results demonstrated that 2,4-D mineralization rates in soil Slurries exceeded the rates predicted by the aqueous phase 2,4-D concentration in seventeen of twenty cases. However further considerations of the methods revealed features of the experimental design that could benefit from improvement. These features included quantification of 2,4-D degradation via depletion from solution rather than generation of C02. Desorption influences on the solution phase concentration during the assay were not accounted for. In these experiments there were not proper controls to account for the properties of the slurry solution imparted by the 88 soils including altered pH and soluble organic carbon content. The experiments are excellent demonstrations of factors which, if accounted for, will allow more definitive conclusions regarding the bioavailability of sorbed contaminants to biodegradative bacteria. For prosperity’s sake the lessons learned are included here. Although it was difficult to draw unequivocal conclusions regarding the bioavailability of sorbed 2,4-D in these soil Slurries, some trends were visible. 2,4-D mineralization rates in soil Slurries exceeded the rates predicted by the aqueous phase 2,4- D concentration for the strains that demonstrated the greatest adherence to surfaces, and in the soils with the lowest organic carbon content. Materials and Methods Soil Characterization The five soils used in this study were all obtained from southern Michigan. The A and B horizons of Capac and Colwood soil series were used to reduce variability in most soil characteristics while obtaining relatively large differences in soil organic carbon. The soils were air-dried and passed through a 2-mm sieve. Percent sand, silt and clay were determined by the Michigan State University Soil Testing Laboratory (East Lansing, MI). Percent carbon was measured on a Dohnnann DC-190 High-Temperature TOC Analyzer. Partition Coeflicients The soil-water partition coefficients for 2,4-D were determined for each soil type by a batch equilibration method. Five gram portions of air-dried soil were measured into a 25-ml glass Corex tube in a sterile hood. The soil was presterilized with gamma 89 irradiation. Six different solutions containing 1000 to 5000 dpm ml'1 ring-U-[14C] 2,4-D (Sigma, 12.8 mCi mmol’l) and total concentrations of 2,4-D ranging from 0.01 to 4 pg ml'1 were prepared in phosphate buffered saline ([PBS]; 8.5 g of NaCl, 0.3 g of KH2P04, 0.6 g of Na2HP04 per liter of distilled water, pH 7.0). One ml of each solution was added to 7.5-ml scintillation cocktail and counted for 5 min on a Packard 1500 Tri-Carb Liquid Scintillation Analyzer to determine initial l4C activity in each 2,4- D solution. Twenty-five ml of these solutions were added to the soils and to controls with no soil. The soil Slurries and controls were placed on a rotating shaker for approximately 18 h, which was long enough to obtain steady-state for the sorption reactions. The tubes were then centrifuged at 7649 x g for 20 min. The activity in the supernatant was determined by adding l-ml aliquots to 7.5 ml scintillation fluid and counting the activity for 5 min. The quantity of 2,4-D sorbed to the soil was determined by calculating the difference between the activity initially in solution and the activity remaining in solution after equilibration. It was determined that there was no significant sorption to the Corex tubes. Distribution coefficients (Kd) for each soil were determined from the sorption isotherms by linear regression. Organisms Five strains that were proteobacteria and able to use 2,4-dichlorophenoxyacetic acid (2,4-D) as a sole carbon and energy source were selected for this study. The strains were chosen to represent a range of abilities to adhere to surfaces. Two of the strains, designated RASC and TFD6, were classified by McGowan (1994) based on 16S rDNA sequence analysis as being Burkholderia sp. and TFD41 was determined to be Ralstonia 90 sp. (Alcaligenes sp.) most closely related to Ralstonia eutropha. The fourth strain, JMP134, was previously classified as Alcaligenes eutrophus by Don and Pemberton (1981). The fifth strain, F B4, was classified as F lavobacterium sp. by Ogram et al. (1985). Cell Adhesion to Surfaces The ability of each of the five strains to attach to a hydrophobic (Teflon) and a hydrophilic (glass) surface was measured by determining the fraction of cells that remained attached to spherical beads once they had been added to an evenly packed flow through column. The methods were adopted from the column experiments described by Rijnaarts et al. (1990). Briefly, 10 cm glass columns (1.0 cm i.d.) were loaded with glass or Teflon beads submerged in PBS. Cultures were prepared as described in the following section and diluted in PBS to an optical density at 280 nm of 0.60. These suspensions were supplied to the vertical down-flow columns by a peristaltic pump at a constant flow rate of approximately 19 ml h'1 for 80 min. The effluent of each column was collected in 5 min fractions. The Optical density of the effluent fractions was measured to determine the proportion of cells which remained adhered to the column. After 80 nrin the influent was changed to cell-free PBS for 50 min to measure detachment. The data presented was obtained from assays done in triplicate. Biodegradation Kinetics To determine an appropriate concentration for the bioavailability assays, the kinetics of 2,4-D mineralization in liquid media were determined for each organism at various 2,4-D concentrations. It was important to work at 2,4-D concentrations for which all strains exhibited first order mineralization kinetics. 2,4-D mineralization rates were 91 measured in solutions of PBS containing 5000 dpm ml'1 ring-U-[14C] 2,4-D and total 2,4-D concentrations of 0.01, 0.05, 0.1, 0.5, 1.0, 5.0, and 10 pg ml'l. Seventy five ml of each 2,4-D solution were distributed to 150 ml serum bottles. The serum bottles were inoculated with 0.75 ml of a dense cell suspension for final cell concentrations of approximately 1X108 CFU ml‘1 in the serum bottle. To prepare the inocula bacterial cultures were grown in a minimal-salts media ([MSM] 1419.6 mg Na2HP04, 1360.9 mg KH2PO4, 0.3 mg (NH4)2SO4, 50 mg MgSO4 7H20, 5.88 mg CaClz 2H20, 3.2 mg EDTA disodium salt, 2.78 mg FeSO4 7H20, 1.15 mg ZnSO4 7H20, 1.69 mg MnSO4 H20, 0.375 mg CuSO4 5H20, 0.233 mg Co(NO3)2 6H20 and 0.1236 mg (NH4)6M07024 41120 per liter of distilled water) that contained 400 pg ml-1 2,4-D. After reaching early stationary phase the cells were rinsed once and resuspended in PBS. After inoculation, the bottles were immediately sealed with a Teflon lined septum and aluminum crimp cap and placed at a slant on a platform shaker. At sampling intervals of up to 6 h the septum of each bottle was pierced with a syringe and 1 ml of solution and 1 ml of headspace were removed. Mineralization of ring-U-[14C] 2,4-D was measured by trapping 14C02 from the samples. Samples were added by syringe to a sealed test tube that contained 1 ml 2N HCl at the bottom, and a cup that contained a small piece of filter paper soaked in 0.25 ml 1N KOH suspended from the top. Tubes sat overnight to allow maximum recovery of 14C02. Filter papers and 2.5 m1 ethanol used to rinse the cup were transferred to 7.5 ml scintillation counting fluid. Activity was counted on a Packard 1500 Tri-Carb Liquid Scintillation Analyzer for 5 min. Initial mineralization rate was determined from the linear portion of the mineralization curve which was established in 92 the first hour of the assay. For all the strains 3 linear relationship between 2,4-D aqueous concentration and initial mineralization rates was observed at concentrations below 1 pg ml-1 2,4-D. The mineralization rates with cells free in solution were compared to the mineralization rates in the presence of non-sorbing quartz sand for each strain. The solids to solution ratio used here and in the bioavailability assays was 15:75 (wtzvol). Rates of 2,4-D mineralization did not differ for any of the strains except TFD41 which showed decreased 2,4-D degradation rates in the presence of quartz sand. TFD41 was removed from the study. Bioavailability Assays Bioavailability assays were performed to directly assess the bioavailability of soil- sorbed 2,4-D. Soil free controls were used to establish the initial mineralization rate of aqueous phase 2,4-D for each strain. Three concentrations, 1.0, 0.67, 0.33 pg rnl'1 , were used to establish the linear relationship between initial mineralization rate and 2,4-D aqueous concentration. This relationship could then be used to predict the initial mineralization rates expected in soil Slurries at solution concentrations less than 1.0 pg ml'1 , under the assumption that only aqueous phase 2,4-D was available for biodegradation. One pg ml‘1 2,4-D was used as the initial solution concentration in the bioavailability assays. The actual solution 2,4-D concentration for each soil slurry at the sorption steady state was calculated from the previously established soil isotherm. To directly measure the mineralization rate in the soil Slurries 15 g of air-dried soil was measured into a serum bottle. The soil was pre-sterilized in bulk with gamma irradiation and measured into the serum bottles in a sterile hood. Seventy five ml of PBS 93 with 1 pg ml“l 2,4-D, a fraction of which was ring-U-[14C] 2,4-D, was added to each vial and allowed to equilibrate with the soil for approximately 20 hr. Control vials contained sterilized non-sorbing Ottawa sand (12 mesh) and 75 ml PBS with 2,4-D. Each vial was inoculated with approximately 108 CF U ml'1 of cells in early stationary phase and crimp sealed with a Teflon lined septum. The methods of culture preparation and 2,4-D mineralization rate determination were as described in the previous section of methods and were determined for each combination of the four strains and the five soil types. Results and Discussion The five soils used in this study had organic carbon contents ranging from 0.36 to 5.82 percent (Table 11). 2,4-D demonstrated a linear sorption isotherm on all five soil types in the concentration range tested (Figure 14). As expected, the 2,4-D sorption isotherms showed that the extent of 2,4-D sorption to the soils increased as organic carbon content of the soil increased (Grover 1973, Liu and Cibes-Viade 1973). The linear distribution coefficients (Kd) ranged from 1.48 ml g'1 in the high organic carbon Colwood soil, to 0.18 ml g'1 in the low organic carbon Capac B-horizon soil (Table 11). 94 Table 11. Properties of soils used in this study. Soil Series Horizon %oca %Sand %Silt %Clay de pHc Colwood A 5 .82 41 35 24 1.48 6.7 Capac A 2.07 44 36 20 0.74 6.2 Oshtemo A 0.78 69 15 16 0.14 5.8 Colwood B 0.42 42 26 36 0.23 6.2 Capac B 0.36 60 20 20 0.18 6.3 a Percent organic carbon by dry weight. b 2,4-D sorption coefficient (m1 g'l). 9 pH of soil slurry in PBS. 5.00 4.50 (CD1 A) A 4.00 2’ g 3.50 D 3.00 V. 2.50 4 (Cap A) 13 2.00 (D g 1 .50 ‘0 1 .00 (Col B) . Cap B) 0.50 (Osh A) 0.00 ‘ 0.00 1. 00 2.00 3.00 4.00 5.00 Dissolved 2,4-D (pg/ml) Figure 14. Sorption isotherms for 2,4-D on five soil types. The strains used in this study were characterized for their adhesion properties. This was used as an indication of their propensities to attach to soils. The assays measured adhesion of the bacterial species to hydrophobic (Teflon) and hydrophilic 95 (glass) surfaces, and showed that all of the strains adhered to the hydrophobic matrix to a greater extent than to the hydrophilic matrix (Table 12). The rank order of species with increasing ability to adhere to a surface remained the same for adherence to Teflon and glass. There was no detectable detachment of any strain from either surface type when the columns were flushed with ten pore volumes of cell free buffer. Table 12. Fraction of cells remaining attached to spherical beads in an evenly packed flow through column. Strain Bead type RASC JMP134 TFD6 FB4 Teflon 0.48 0.51 0.71 0.71 Glass 0.00 0.04 0.29 0.62 Mineralization rates were determined for each combination of the four strains and the five soil types to understand some of the complex relationships between bioavailability, bacterial adhesion to surfaces, and soil organic carbon content. The 2,4-D mineralization rate in soil Slurries, with a portion of 2,4-D sorbed to soil, was directly assessed and compared to the predicted mineralization rate in a solution with an equal aqueous concentration of 2,4-D. The specific initial mineralization rates for all bioavailability assays are presented in Table 13. 96 Table 13. Specific initial mineralization rate of 2,4-D for each strain and soil type combination. Specific initial mineralization rate (pg 106cells'l min'l) Soil type Col A Cap A Cap B % Organic carbon 5.82 2.07 0.36 Strain RASC **110i5 166st1] "22335 JMP134 ** 29:1:2 43i3 ** 36i1 ** 75i2 TFD6 713:3 94i2 1043c3 FB4 692t2 **1 17:1:4 **208i12 **144:1:2 * Slurry mineralization rate is outside the 90% prediction interval of the control line. ** Slurry mineralization rate is outside the 95% prediction interval of the control line. Values given represent the mean of three rate measurements and one standard deviation. To assist in the comparison of 2,4-D mineralization rates among the four strains, the 2,4-D mineralization rate measured for each soil slurry was normalized to the 2,4-D mineralization rate predicted for the given aqueous concentration of 2,4-D in that slurry. The normalized initial mineralization rate has a value greater than one if the measured mineralization rate in the soil slurry exceeded the rate predicted from the initial solution phase 2,4-D concentration (Figure 15). 97 1.0 " _ JMP134 0.0 -. Normalized initial mineralization rate RASC flacterial species in 7‘ / order of increasing .3\ COIA adhesion abilities Soil type in order of increasinfg\> % organic carbon Figure 15. Soil slurry mineralization rates. The 2,4-D mineralization rate measured for each soil slurry was normalized to the 2,4-D mineralization rate predicted for the given aqueous concentration of 2,4-D in that slurry. The normalized initial mineralization rate has a value greater than one if the measured mineralization rate in the soil slurry exceeded the rate predicted from the aqueous concentration. Mean values of triplicate normalized initial mineralization rates are represented in the bar graph. Hashed bars represent values less than one. Figure 15 shows that in seventeen out of twenty cases the rate of 2,4-D mineralization in the presence of a sorbent exceeded the rate predicted from the initial aqueous phase concentration. One possible conclusion is that bacteria present at the soil surface experience localized zones of higher substrate concennation than in the bulk solution. Unfortunately this conclusion can not be made unequivocally. In large part this is because the soil free controls that were used to establish the expected rates of solution phase 2,4-D mineralization were done in plain PBS. The pH of PBS was buffered at 7.0. However after contact with the soils the pH ranged from 5.8 to 6.7 (Table 11). The 98 buffering capacity of the soil was apparently stronger than that of the buffered solution. This effect was not noticed until after the bioavailability assay data had been gathered. The experiments in Chapter 5 demonstrated how decreased pH resulted in increased rates of solution phase degradation of 2,4-D by these strains (Figure 12). In addition the soluble organic carbon content of the PBS in the soil Slurries was visibly different than unaltered PBS. As discussed in Chapter 5 this can have positive or negative effects on the solution phase degradation rates. To account for any effects that the soil supernatant may have had on the bioavailability assays independent of the soil solids, the effect of the soil-slurry supernatant on the mineralization rate of 2,4-D was measured. Soil slunies were prepared as in previous assays however, the slurry was centrifuged and the soil free supernatant was used rather than the entire slurry. Rates of 2,4-D mineralization in the soil supematants were compared to rates in PBS. This demonstrated the effects of pH and soluble carbon that were independent of the presence of the solids (Figure 16). Figure 16 demonstrates that an appreciable enhancement effect can be measured in the absence of the soil solids. Enhancement in these cases can not be due to bacteria present at the soil surface experiencing localized zones of high substrate concentration. Rather, properties of the soil-slurry supernatant have altered the rate of solution phase 2,4-D mineralization. 99 Normalized initial mineralization rate H O / Bacterial species in order of increasing -~\z\% adhesion abilities Supernatant type in order of increasing % organic carbon Figure 16. Soil-slurry supernatant mineralization rates. The 2,4-D mineralization rate measured for each supernatant was normalized to the 2,4-D mineralization rate predicted for the given aqueous concentration of 2,4-D in that slurry. Hashed bars represent values less than one. The results of the supernatant assays were used to correct the results of the soil assays so that effects due to just the 2,4-D sorbed to solids could be evaluated. Both sets of assays were done in conjunction with unaltered PBS controls. This offered the basis for comparison between the assays with the assumption that supernatant effects remained the same in the absence and presence of soil solids. Enhancement or inhibition in the assays was expressed as a fiaction of the rate of depletion in PBS. Effects measured in the supernatant assays were subtracted from effects measured in the soil-slurry assays to determine the effects of just the soil solids. Figure 17 shows the normalized rates of 2,4- D mineralization in the soil slurries minus the effects of the soil-slurry supematants. The trends relating increased bioavailability to decreased soil organic carbon and increased 100 bacterial adhesion are still visible in this figure although eight rather than seventeen of the cases have mineralization rates exceeding those expected from solution phase mineralization rates. As the bacterial ability to adhere to Teflon or glass increased the mineralization rates increased (Figure 17). Strain FB4, utilized in previous studies by Ogram et al. (1985), demonstrated the greatest ability to adhere to surfaces. For this organism, mineralization rates were higher than predicted from the aqueous phase 2,4-D concentration in four of the five soils. RASC with the least ability to adhere to surfaces did not demonstrate enhanced mineralization rates in any of the five soil types. A second visible trend relates soil organic carbon content to enhanced mineralization rates. Greer and Shelton (1992) showed that rates of 2,4-D degradation were higher in low organic carbon soils than in high organic carbon soils. Similarly, the data from this study show that the potential mineralization of soil sorbed 2,4-D decreases as soil organic carbon content increases. None of the strains exhibited elevated 2,4-D mineralization rates in the soil with the highest organic carbon content (Figure 17). 101 Normalized initial mineralization rate 2.0 1.5 "0 ,-::::: FB4 05 .. VTFD6 / JMP134 0.0 RASC [xi/Bacterial species in /’ order of increasing \ .-_ X 7 3» ColA adhesion abilities Soil type in order of incfeas'in‘g‘Xr-g % organic carbon Figure 17. The effect of soil slurry minus the effect of soil-slurry supernatant on 2,4-D mineralization rates. Hashed bars represent values less than one. The relationship between the extent of sorption (reflected by sorption coefficients Kd) and the kinetics of desorption may help explain the observation that contaminants sorbed to soils with high organic matter contents appear to be less available than they are in low organic matter content soils. Brusseau and Rao (1989) have made the empirical observation that the sorption coefficient has an inverse log linear relationship to the sorption kinetics. That is, as the tendency towards more extensive sorption increases, the kinetics of desorption decrease. Their observations seem consistent with our results in so far as higher organic matter contents will manifest larger sorption coefficients and hence slower desorption rates which lead to lower bioavailability of the sorbed contaminant pool. Intra-organic matter diffusion has been suggested as a rate limiting mechanism in 102 the desorption of poorly water soluble organic compounds from soils (Brusseau, et al. 1991). The three-dimensional hydrophobic matrix of organic matter provides a sink for the diffusive mass transfer of sorbate into the interior of organic matter; desorption from the hydrophobic matrix is limited by the rate of migration of 2,4-D from the interior of the matrix to the soil-water interface. It seems plausible that organic matter may be present in more particulate forms in the high organic content soils, thereby providing a greater diffusion path length and slower desorption rates. Organic matter in the low organic content soils may be present to a larger degree in thinner surface coatings which manifest faster desorption rates. Furthermore, in low organic carbon content soils a greater portion of the sorbed 2,4-D may be more surface localized, and our results indicate that surface localized 2,4-D may be available for degradation by bacteria that attach to the surface. Finally, bioavailability can be examined from different vantage points. Biodegradation of an organic compound can be described in two general steps, uptake and biological transformation. If bioavailability is used to assess the complete removal of a contaminant from the soil environment then quantification of bioavailability via production of C02 is justified. This method is restricted to systems that demonstrate first-order carbon dioxide production rates. If bioavailability is defined as the removal of a contaminant from the soil solids and solution via uptake by cells then there are advantages to measuring disappearance of the parent compound. Measuring depletion of the parent compound removes the assumption or necessity to test if uptake is the rate limiting step rather potential enzyme rate limitations or release of intermediates that must be re-absorbed before final 103 transformation to carbon dioxide. If either of these processes occurred the bioavailability would be underestimated by correlation with C02 production. In addition measurement of depletion of the parent compound increases the applicability of the bioavailability assays to systems that may not demonstrate first-order carbon dioxide productions rates but do demonstrate first-order contaminant uptake by the bacterial cells. For these experiments, where bioavailability was assessed via evolution of carbon dioxide, bioavailability may have been overestimated if rapid desorption occurred during the initial portion of the bioavailability assay. In the case of 2,4-D it is probable that rapid desorption did occur. Figure 10 in Chapter 4 shows that desorption was extremely rapid for all sorbate-sorbent combinations used in the silica studies. Desorption of 2,4-D from the soil can alter the solution concentration in less than 1 min. The first sample taken for the bioavailability assays that measured C02 production from soil Slurries was after ten minutes. The predicted rate of C02 production was based on the solution concentration prior to inoculation. This may underestimate the rate of C02 production and overestimate the bioavailability if desorption is rapid compared to mineralization. Conclusions The lessons learned from this set of experiments formed the stepping stones for the development of the improved silica assays and bioavailability models. Sorbent-free controls must account for properties imparted to the solution by the sorbent. In particular changes in pH and soluble organic carbon should be determined and controlled for. The degradation of 2,4-D is better monitored via depletion from solution than production of carbon dioxide as the former is more closely related to bioavailability than the latter. 104 Finally rapid desorption must be accounted for to make strong conclusions regarding the direct availability of sorbed compounds. Acknowledgements This research was supported by NSF grant DEB9120006, USDA NRICGP No. 94-37107-03 86, and the Michigan Agricultural Experiment Station. Special thanks are offered to the laboratory of Hauke Harms at EAWAG, Switzerland, for teaching the column adhesion assay technique to Denise Kay. Literature Cited Brusseau, M. L., R. E. Jessup, and P. S. C. Rao. 1991. Nonequilibrium sorption of organic chemicals: elucidation of rate-limiting processes. Environ. Sci. Technol. 25: 134- 142. Brusseau, M. L., and P. S. C. Rao. 1989. The influence of sorbate-organic matter interactions on sorption nonequilibrium. Chemosphere. 18(9-10): 1691-1706. Don, R. H., and J. M. Pemberton. 1981. Properties of six pesticide degradation plasmids isolated from Alcaligenes paradoxus and Alcaligenes eutrophus. J. Bacteriol. l45(2):681-686. Greer, L. E., and D. R. Shelton. 1992. Effect of inoculant strain and organic matter content on kinetics of 2,4-dichlorophenoxyacetic acid degradation in soil. Appl. Environ. Microbiol. 58(5): 1459-1465. Grover, R. 1973. The adsorptive behavior of acid and ester forms of 2,4-D on soils. Weed Res. 13:51-58. Liu, L. C., and H. R. Cibes-Viade. 1973. Adsorption of fluometuron, prometryne, Sencor, and 2,4-D by soils. Univ. Puerto Rico J. Agric. 57:286-293. McGowan, C. 1994. Interspecies gene transfer in the evolution of 2,4- dichlorophenoxyacetic acid degrading bacteria. Ph.D. thesis. Michigan State University, East Lansing. Ogram, A. V., R. E. Jessup, and P. S. C. Rao. 1985. Effects of sorption on biological degradation rates of (2,4-dichlorophenoxy)acetic acid in soils. Appl. Environ. Microbiol. 49 :582-5 87. 105 Rijnaarts, H. H. M., A. Bachmann, J. C. Jumelet, and A. J. B. Zehnder. 1990. Effect of desorption and intraparticle mass transfer on the aerobic biomineralization of alpha- hexachlorocyclohexane in a contaminated calcareous soil. Environ. Sci. Technol. 24(9):]349-1354. 106 "llllllllllllllllf