vvEPA United States Environmental Protection Agency Office ol Research and Development Washington DC 20460 EPA600/R-98/162 November 1998 www.epa.gov/ncerqa DRINKING WATER PROGRESS REVIEW WORKSHOP FOR THE 1995-1998 SCIENCE TO ACHIEVE RESULTS f. ,, *"•„# % C, -~& ^ fr - *, -v^vJ.^4 ^-^ *• * ^ . (STAR) GRANTS December 8-9,1998 Arlington, Virginia rf* »* .s^a., — * tur-fz.; * v •£•-«»* I - ~**m ------- ------- Table of Contents Introduction vii Section 1. Arsenic Grants Arsenic Contribution From Dietary Sources 3 Edo D. Pellizzari, Reshan Fernando, James H. Raymer, Olujide Akiribo, C. Andrew Clayton A Dose-Response and Susceptibility Investigation of Skin Keratoses and Hyperpigmentation Caused by Arsenic in Drinking Water 4 Allan H. Smith, Reina Haque, D.N. Guha Mazumder, Binay K. De, 'Nilirna Ghosh, Soma Mitra, David Kalman Arsenic-Glutathione Interactions and Skin Cancer 6 Elizabeth T. Snow Arsenicals, Glutathione Reductase and Cellular Redox Status 7 Miroslav Styblo Carcinogenicity of Sodium Arsenite in p53+'~ Male Mouse on a Choline-Deficient Diet 9 Raymond Tice, Glenda Moser, Thomas Goldsworthy Section 2. Microbial Pathogens Grants Mechanisms of Inactivation of Viruses in Groundwater 13 Maria E. Alvarez, Suresh Pillai Meaningful Detection of Known and Emerging Pathogens in Drinking Water 15 Gerard A. Cangelosi Cryptosporidiumparvum Volunteer Study: Infectivity, Illness, and Immunity 16 Cynthia L. Chappell, Pablo C. Okhuysen, Herbert L. DuPont, Charles R. Sterling, Walter Jakubowski Development of a Quantitative Cell Culture-Based Infectivity Assay for Cryptosporidium parvum in Source and Finished Water 18 Ricardo De Leon, Paul A. Rochelle Cultural and Enhanced RT-PCR Methods for Waterborne Caliciviruses 20 G. Shay Font Virus Filtration and Water Reuse: From Microscale Phenomena to Field-Scale Observations 21 Stanley B. Grant National Estimate of the Occurrence of Waterborne Disease 22 Deborah Levy Studies of the Infectivity of Norwalk and Norwalk-Like Viruses 23 Christine L. Moe PCR-Based Detection of Cytopathogenic and Noncytopathogenic Viruses in Water 25 Ian L. Pepper, Kelly A. Reynolds, Charles P. Gerba Detection and Occurrence of Human Caliciviruses in Drinking Water 26 MarkD. Sobsey 111 ------- Table of Contents (continued) Genomic Database for Cryptosporidium Species 28 Steve J. Upton, Christine C. Dykstra, Byron L. Blagburn Understanding Risk Factors to Cryptosporidium parvum: Studies in Gnotobiotic Pigs 29 Lucy A. Ward, SukHan Cheung, Yunfei Wang, C.K. Nielsen Section 3. Disinfection By-Products Grants Development of Biomarkers, in Peripheral Blood, for Exposure to and Effects of'the Water Disinfectant By-Products Haloacetonitroles 33 Ahmed E. Ahmed Metabolic Fate of Halogenated Disinfection By-Products In Vivo and Their Relation to Biological Activity 34 LM. Ball Evaluation of the Efficacy of a New Secondary Disinfectant Formulation By Using Hydrogen Peroxide and Silver and the Formation of Disinfection By-Products Resulting From Interactions With Conventional Disinfectants 35 Stuart Batterman Integrated Approach for the Control of Cryptosporidium parvum Oocysts and Disinfection By-Products in Drinking Water Treated With Ozone and Chloramines 37 Benito J. Marinas, Roger A. Minear The Use of Ozonation and FBT for the Control of THM Precursors in Drinking Water 38 Susan J. Masten Genotoxicity and Occurrence Assessment of Disinfection By-Product Mixtures in Drinking Water 39 Roger A. Minear, MichaelJ. Plewa Molecular Weight Separation and HPLC/MS/MS Characterization of Previously Unidentified Drinking Water Disinfection By-Products 41 Roger A. Minear, Sylvia Barrett Direct Quantisation of Haloacetic Acids By Surface Enhanced Raman Scattering 42 Michael J. Natan Health Risk of Trihalomethanes Found in Drinking Water: Carcinogenic Activity and Interactions 43 Michael A. Pereira Assessment of Human Dietary Ingestion Exposures to Water Disinfection By-Products via Food 44 James H. Raymer, Gerald G. Akland, Edo D. Pellizzari, C. Andrew Clayton, Doris J. Smith Analysis of Organic By-Products From the Use of Ozone/Chlorine and Ozone/Chloramines in Drinking Water Treatment 45 David A. Reckhow Overview of Disinfection By-Products Research and Preliminary ICR Findings 47 Susan D. Richardson Physiologically Based Pharmacokinetic Modeling of Haloacid Mixtures in Rodents and Humans 48 Irvin R. Schultz IV ------- Table of Contents (continued) Development of a New, Simple, Innovative Procedure for the Analysis of Bromate and Other Oxyhalides at Sub-ppb Levels in Drinking Water 50 Howard Weinberg Inhalation and Dermal Exposure to Disinfection By-Products of Chlorinated Drinking Water 51 Clifford P. Weisel, Jeffrey Laskin Kinetic-Based Models for Bromate Formation in Natural Waters 52 Paul Westerhoff Mechanistic-Based Disinfectant and Disinfectant By-Product Models 53 Paul Westerhoff, David Reckhow, Gary Amy, Zaid Chowdhury Mechanisms and Kinetics of Chloramine Loss and By-Product Formation in the Presence of Reactive Drinking Water Distribution System Constituents 55 Richard L. Valentine Index 57 ------- ------- Introduction The mission of the United States Environmental Protection Agency (EPA) is to protect public health and to safeguard and improve the natural environment—air, water, and land. Achieving this mission requires applying sound science to assessing environmental problems and evaluating possible solutions. The National Center for Environmental Research and Quality Assurance (NCERQA) at EPA is committed to providing the best products in high-priority areas of scientific research through significant support for long-term research. One high-priority research program identified by the EPA Office of Research and Development (ORD) is Drinking Water. The Safe Drinking Water Act mandates that EPA identify and regulate drinking water contaminants that may have any adverse health effects, and that are known or anticipated to occur in public water systems. EPA regulations addressing requirements of the Act require disinfection of surface water and certain groundwater supplies. Scientific evidence suggests that exposure to chemical by-products formed during the disinfection process may be associated with adverse health effects. Reducing the amount of disinfectant or altering the disinfection process may decrease by-product formation; however, these practices may increase the potential for microbial contamination. EPA's challenge is to balance the health risks caused by exposure to microbial pathogens with the health risks caused by exposure to disinfection by-products (DBFs). To meet this challenge, the Agency has developed comprehensive research plans for microbial contaminants/disinfection by-products and for arsenic in drinking water. These plans will form the basis for prioritizing research needs for the Agency's drinking water program. Research described in this program review has been funded through EPA's Science to Achieve Results (STAR) Program, which is managed by NCERQA. Grants were awarded through open, peer-reviewed competition of proposals submitted in response to Requests for Application (RFA) for each of the years from 1995 to 1998. The research being supported by these grants, along with work conducted in the EPA laboratories and other outside research institutions, is addressing key research questions identified in the microbial/disinfection by-products and arsenic research plans. Some key questions are: What data are required to assess arsenic exposure in human populations? What are the health risks caused by exposure to microbial pathogens, and what methods are needed to measure occurrence of pathogens in drinking water? What are the health effects associated with exposure to DBFs, and how can we better characterize the risk posed by exposure to multiple or complex mixtures of DBFs? This program review will allow investigators to interact with one another and to discuss the progress and findings of their research with EPA and other interested parties. If you have any questions regarding the program, please contact the program manager, William G. Stelz, at 202-564-6834 or stelz.william@epa.gov. vn ------- ------- Section 1. Arsenic Grants ------- ------- Arsenic Contribution From Dietary Sources Edo D. Pellizzari, Reshan Fernando, James H. Raymer, Olujide Akinbo, and C. Andrew Clayton Research Triangle Institute, Research Triangle Park, NC This project's objective is to enhance the un- derstanding of human exposures to arsenic (As), hi speciated forms, especially with regard to the dietary pathway. The contribution of diet to arsenic exposure is an important piece of information to consider as reductions hi the allowable drinking water concen- trations are contemplated. This study is investigating the distribution of toxic As species in composite food samples from EPA Region 5 National Human Exposure Assessment Study (NHEXAS) and food, water, and urine samples from the NHEXAS Children's Study. Data from these probability-based samples will generate information about exposure to As species of all residents of Region 5 and for children in two rural counties and one urban county hi Minnesota. The specific objectives of Phase I are: (1) to refine, optimize, and validate a speciation method for arseno-betaine (AsB), arsenocholhie (AsC), monomethyl arsonic acid (MMA), dimethyl arsinic acid (DMA), arsenite [As(III)], and arsenate [As(IV)] hi food (dup- licate diets); and (2) to optimize and validate a spec- iation method for As hi urine, and to set up a speciation method for As hi drinking water. The objectives of Phase II are to: (1) estimate the levels of speciated (and total) As hi selected foods and estimate the effects of food preparation methods on these levels, including the cooking of foods hi water contaminated with As; (2) identify food types associated with elevated levels of speciated (and total) As based on food diaries and analysis of samples to be collected during a mini-market basket survey; (3) estimate the food exposure dis- tributions for speciated forms of As, using measure- ments on archived duplicate-diet food samples from the NHEXAS Region 5 Study; (4) estimate the food and drinking water exposure distributions, and urine dose distributions, for speciated forms of As, using measure- ments on archived samples from the NHEXAS Chil- dren's Study; and (5) estimate correlations/associations among urine/food/drinking water levels of speciated As, using measurements on archived samples from the NHEXAS Children's-Study. Speciation is accomplished using ion chromato- graphy hi conjunction with inductively coupled plasma/ mass spectrometry (IC-ICP-MS). Aspects of the re- search include optimizing the separation and recovery of As species from the matrix, preservation of chemical species during extraction and analysis, and stability of the As species during storage prior to extraction. Three food extraction methods are being evaluated for their ability to isolate As species from duplicate diet samples. These methods are: (1) MeOH-Water-CHCl3, (2) MeOH-Water, and (3) enzymatic digestion hi water (Trypsin-(NH4)2CO3). A mass balance approach is being used to evaluate the extraction methods for As species. To challenge and evaluate the performance of the extraction schemes, two standard duplicate diet composite samples were prepared, one high (49%) and one low (6%) in calories of fat, each composite con- taining a food known to have As present. In addition, high- and low^fat duplicate diet composite samples were prepared, but without (or very low) As-containing foods. The latter composites will be used to examine storage- stability of As species and as quality control samples during the analysis of NHEXAS duplicate diet samples. Cooking experiments began to examine the impact of cooking on As concentrations and species hi foods and to determine whether As may be transferred to foods eaten when foods are cooked hi water con- taminated with As. In these experiments, the uncooked food is being analyzed for total As. If any As is measured, it is speciated. Then the food (e.g., shrimp, tuna, chicken) is subjected to broiling, grilling, frying, etc., to determine whether the species of As are transformed from one form to another. Foods also are being cooked in water containing a range of As 'concentrations typically found hi the environment. The cooked food and residual water are being analyzed to permit mass balance. The conversions of As com- pounds during boiling and hi the absence of foods also will be evaluated to improve understanding of the system. Using an anion-exchange separation procedure and detection by ICP-MS, baseline resolution of AsB, AsC, As(III), DMA, MMA, and As(IV) was achieved. The chromatographic conditions to achieve this separa- tion consisted of a Hamilton PRP-X100 Anion exchange column and gradient elution with mobile phase A consisting of 5mM Na2CO3/NaHCO3 (aq), pH7:MeOH, 94:6, v/v and Mobile phase B containing 25mM NazCOj/NaHCO;, (aq), pH 7:MeOH, 94:6, v/v, at a flow rate of ImL mhi"1. Measurement of levels as low as 25 pg for each species (on column) was achieved. The three methods were used to extract As species from 2g of each composite food fortified with As-containing foods, the low-fat and high-fat food composites. The major component hi each sample was AsB. Also, the intensity of the AsB peak hi ah* three extracts were comparable. The preliminary findings are that extraction efficiencies of these three methods appear similar with these duplicate diet samples. The remaining research includes completing the method development and validation for As species hi food, optimizing the As species method for urine, analyzing duplicate diet samples, performing cooking experiments, determining the fate of As in the foods, and data analyses. ------- A Dose-Response and Susceptibility Investigation of Skin Keratoses and Hvperpigmentation Caused by Arsenic in Drinking Water Allan II. Smith, Reina Hague University of California, Berkeley, CA D.N. Guha Mazumder, Binay K. De, Nilima Ghosh, Soma Mitra Institute of Post Graduate Medical Education and Research, Calcutta, India David Kalman University of Washington, Seattle, WA Keratoses and hyperpigmentation are hallmark dermal signs of arsenic (As) toxicity. Keratoses are hard raised lesions that appear on the palms and soles. Hyperpigmentation is marked by raindrop-shaped diffuse dark spots on the trunk and limbs. The first detailed assessment of the dose-response relationship of arsenic-induced keratoses and hyperpigmentation is under way. This project's objective is to determine if susceptibility varies by As methylation capability and nutritional factors such as methionine and cysteine. As methylation will. be assessed by urinary assays. Nutritional status will be determined by blood measurements of key macronutrients and micronutrients as well as by analysis of a dietary questionnaire. A case-control study has commenced that takes advantage of the largest population-based survey con- ducted in an As-affected area of West Bengal, India (see Figure 1). The landmark cross-sectional survey, con- ducted between 1995 and 1996, included more than 7,000 participants, 400 of whom were found to have As-induced skin lesions. Approximately 280 of the 400 individuals with skin lesions were exposed to drinking water containing less than 500 mg/L of inorganic As. These 280 individuals comprise the case group for the present investigation. The control group consists of lesion-free individuals randomly selected from the cross- sectional survey database matched on age and sex. Data obtained from personal interviews and chemical analyses of drinking water samples will be used to assess As exposure. The interview consists of questions about lifetime residential history, water sources at work, and fluid consumption. The clinical exam involves various dermatologic, neurologic, res- piratory, and hepatic endpoints. A dietary question- naire supplemented with results of blood assays will be used to ascertain the participants' nutritional status. Urinary assays will be used to determine As methylation efficiency. The interviews and sample collection commenced hi June 1998. The field team reached their target rate of five interviews per week. To date, more than 100 par- ticipants have .been seen. Approximately 50 urine and blood samples have been transported to the United States for analyses. By building on an existing study, the present investigation was specifically designed to identify the shape of the dose-response curve for arsenic and skin lesions, and to detect a possible threshold. This project also will identify potential susceptibility factors, and thereby reduce the uncertainty in generalizing risk to other populations. Because keratoses and hyperpigmen- tation are the most common and earliest occurring endpoints of chronic ingestion, and because evidence suggests that these skin lesions are biomarkers of sub- sequent cancer risks, this project will significantly con- tribute to a fuller understanding of the long-term health effects of consuming water containing low levels of As. Interviewing and sample collection will continue in India, and data entry and analyses of urine and blood samples will continue hi the United States. Approxi- mately 400 individuals will be recruited by December 1999 (200 cases and 200 controls). ------- KERATOSES 50-99 100-149 150-199 200-349 350-499 As Level in Ttabewell Drinking Water (ug/1) 10.7 500-799 >800 25.0 T <50 HYPERPIGMENTATION 100-149 150-199 200-349 350-499 Arsenic Level in Tubewell Drinking Water (ug/1) 22.7 500-799 >800 Figure 1. Prevalence of keratoses and hyperpigmentation per 100 for females and males in West Bengal, India, 1995-1996 Cross Sectional Survey (Adapted from Guha Mazumder et al., InternationalJournal of Epidemiology, 1998). ------- Arsenic-Glutathione Interactions and Skin Cancer Elizabeth T. Snow Nelson Institute of Environmental Medicine, New York University School of Medicine, Tuxedo, NY This project's objective is to test the hypothesis that arsenic (As)-induced cancer is the result of changes in cellular redox control mediated by altered glutathione (GSH) levels by exploring the effect of As on gluta- thione-regulating enzymes in human keratinocytes in vitro, and in mice in vivo. It is expected that exposure to physiologically relevant, low doses of As will result hi the activation of enzymes such as p21-ras, glutathione S- transferase (GST), glutathione reductase (GR), y-glut- amyl-cysteine synthetase (y-GCS), and glutathione per- oxidase (GPx) due to changes in cellular phosphoryla- tion and/or redox status, and will thereby potentiate the induction of cellular stress responses. Subsequent to an oxidative stress response in certain cell types, such as keratinocytes, As can induce hyperproliferation and in- hibit DNA repair. This project's approach is three-fold: (1) Cultured human keratinocytes will be treated with low doses of As, and the relative activity of several key enzymes, GST, y-GCS, GR, GPx, andp21-ras, will be examined. (2) The activity of these enzymes will be assayed both hi purified form and in extracts from untreated and As- treated cells. The role of GSH and As(GS)x complexes in mediating these responses will be examined by measuring cellular GSH and GSSG levels, by assessing the formation of As(GS)x complexes, and by modulating GSH levels. (3) The role of GSH hi As carcinogenesis will be evaluated by examining the effect of varied GSH levels on the rate of papilloma induction hi a mouse skin tumorigenesis model using normal mice and mice that overexpress human GPx. 25Or Preliminary findings have shown that physio- logically relevant concentrations of As, in a variety of forms, do not directly inhibit GSH metabolizing en- zymes. However, low concentrations of inorganic As can cause significant changes hi cellular GSH levels and hi the relative activity of a variety of GSH metabolizing enzymes, either in cultured human keratinocytes (see Figure 1) or hi the epidermis of mice exposed via their drinking water. These results show that even low, relatively nontoxic concentrations of As can directly modulate cellular redox levels which, hi turn, may alter cellular signalling and other aspects of intermediary metabolism and thereby contribute to the carcinogenic process. However, contrary to earlier hypotheses, this project's findings indicate that these changes are not brought about by the direct inhibition of GSH-related enzymes by inorganic As or As metabolites. Most enzymes are quite insensitive to physiological concentrations of As. Inhibition is only seen at high As concentrations or with complexes found at very low concentrations hi skin cells. It is proposed that As produces alterations hi the activity of these redox-regulating enzymes by initiating changes hi the regulation of a variety of stress-response genes. The means whereby As initiates these changes is not yet established. The next step is to evaluate the molecular reg- ulation of these enzymes hi cultured human keratino- cytes and to continue ongoing carcinogenesis experi- ments to evaluate the role of redox control processes hi the production of skin cancer. 5 7.5 10 12.5 Arsenic(lll) Figure 1. Increased GSH, y-GCS activity, and cystine uptake in human keratinocytes after 24 hours As(HI). ------- Arsenicals, Glutathione Reductase and Cellular Redox Status Miroslav Styblo Department of Pediatrics, University of North Carolina, Chapel Hill, NC This project's goal is to examine two related aspects of the mode of action of inorganic arsenic (iAs) as a carcinogen and toxin, including: (1) interactions between metabolites of iAs and glutathione reductase (GR), a key enzyme in redox metabolism of glutathione (GSH); and (2) effects of the As-GR interactions on the ratio of GSH to glutathione disulfide (GSSG), an indicator of cellular redox status. This project integrates studies at three levels of biological complexity: purified enzyme, cultured cells, and intact animals. The experimental approach includes: (1) examination of the interactions of iAs and its meta- bolites with GR purified from human erythrocytes; (2) examination of the effects of arsenicals on GR activity, GSH/GSSG ratio and intracellular peroxides in cultured human cells as compared with patterns of arsenic metabolism hi cells; and (3) examination of the in vivo effects of arsenicals on GR activity and GSH/GSSG ratio in tissues of mice. During the first year of the project, metabolic patterns for arsenicals have been examined in human cell lines to be used in future experiments. These cell lines have been derived from tissues that are known to be a maui site for metabolism of iAs (liver) or targets for its carcinogenic effects (skin and urinary bladder). Primary hepatocytes isolated from rats have been used as a positive metabolic control. Among the cell lines examined, rat primary hepatocytes had the greatest capacity for methylation of arsenite (iAs111). The methylation capacity was sig- nificantly lower in the human cell lines (primary hepatocytes > epidermal keratinocytes > SV-40 trans- formed urinary bladder [Urotsa] cells). In primary rat hepatocytes, dimethylarsenic (DMAs) was the major metabolite exported from cells to culture media. The total methylation yield increased hi the range of 0.1 to 4 mM iAs1". At 10 to 20 mM iAs1", inhibition of meth- ylation occurred. In cell lines with low methylation capacity, iAs and/or monomethylarsenic (MAs) accum- ulated hi the cells, suggesting that complete methylation (dimethylation) is a precondition for clearance of As from cells. Cytotoxicities of iAs and its metabolites also have been examined. Trivalent methylated arsenicals (MAs"1 and DMAs"1) were more acutely toxic than iAs1" for all cell lines examined (see Table 1). GSH (2.5 mM) completely protected cells against toxicity of triValent arsenicals. Addition of 2 ^M sodium selenite reversibly inhibited methylation and increased retention of iAs1" in cells. Presence of GSH hi culture media dramatically enhanced the inhibitory effects of selenite (see Figure 1). Selenite potentiated the cytotoxicity of iAs1", MAs111, and DMAs1" and interfered with the protective effect of GSH (see Figure 2). These results indicate that: (1) the methylation of iAs hi cultured cells yields potentially toxic trivalent methylated metabolites; (2) the high methylation capa- city did not protect cells from cytotoxic effects of trivalent arsenicals; and (3) the methylation capacity is concentration dependent and affected by the presence of GSH and selenite. The effects of iAs and its methylated metabolites on GR activity and GSH/GSSG ratios in animal and human cell lines are currently being studied. In parallel, the relationship between As exposure and induction of oxidative stress in cultured cells is being examined. Interactions between arsenicals and GR isolated from human erythrocytes will be examined during the second year of the project followed by in vivo studies in mice. MTTAssav Cell Line Methylation Capacity (pmol iAsnl/106 cells/hr) Estimated IC50 iAs1" MAsmO Human Keratinocytes 0.1 1.4 0.3 'Values ()jM) for Arsenicals MAsnlI2 DMAs"1! Rat Hepatocytes 370 Human Keratinocytes 0.1 Urotsa ND >5 2.6 »5 3.3 »5 1.2 2.5 2.3 1.6 2.6 »5 3.5 Neutral Red Assay 0.7 0.4 ICso is defined as a concentration of an arsenical that results in 50% decrease in viability of cells over 24-hr incubation period. Table 1. Methylation capacities and susceptibility of cells to toxic effects of trivalent arsenicals. ------- -GSH +GSH o E .0. CO W Q 40 30 20 10 0 0/C 4 (MM) Figure 1. Inhibition of methylation of 0.1 u,M lAs"1 in primary rat hepatocytes by selenite in the presence or absence of GSH (2.5mM). Total amounts of methylated metabolites (MAs and DMAs) in cell culture (cells and medium) after 8-hour incubation. 100' O •S «M _CONTROL_ CONTROL-t-Se' -Ss lAs»(10) MAsra(1) DMAs'" (10) Arsenical (uM) o O •5 Z%t CONTROL r1 +Se 0.4 0.4 10 10 Concentration of MAs"1 (pM) -GSH +GSH Figure 2. Effects of selenite on acute toxicity of trivalent arsenicals in primary rat hepatocytes: (a) effects of 2 |iM selenite or cytotoxicity of iAs™ (10 uM) MAsm(l uM), and DMAs"1 (10 uM); (b) protective effects of 2.5 mM GSH against cytotoxicity of MAs"1 in the presense and in the absence of 2 uM selenite. Cytotoxicity was examined after 24-hour incubation using MTT assay. ------- Carcinogenicity of Sodium Arsenite in p53+/~ Male Mouse on a Choline-Deficient Diet Raymond Tice, Glenda Moser, and Thomas Goldsworthy Integrated Laboratory Systems, Research Triangle Park, NC Arsenic is a multisite human carcinogen. There are no known experimental model systems in which arsenite is tumorigenic. This project's objectives are to: (1) evaluate the carcinogenic ability of sodium arsenite (AS) in drinking water as a function of hepatic methyl donor status by comparing tumorigenicity in hetero- zygous p53+'~ mice on a choline-deficient (CD) diet as compared with mice on a choline-sufficient (CS) diet; (2) determine the cocarcinogenic ability of sodium arsenite with the skin carcinogen 4-vinyl- 1-cyclo-hexene diepoxide (VCD) or the bladder carcinogen 2-meth- oxy-5-methylaniline (p-cresidine); (3) extend acute genotoxicity data; and (4) correlate hepatic methylation activity to nutritional status and tumorigenicity. A 28-day dose-finding study was done in male C57B1/6 mice on CD and CS diets. The tumor study was done in male p53+/~ mice on CD and CS diets. In the 28-day study, groups of five mice 8 weeks of age were exposed to 0.005, 0.01, or 0.02 percent AS in drinking water. There were no treatment-related deaths, clinical observations, or significant histopathological changes in liver or skin after AS exposure. Body weight gain was decreased hi mice on CD and CS diets exposed to 0.02 percent AS. There were no differences in water, AS, or food consumption hi mice on the CD diet as compared with mice on the CS diet. As a function of water consumption and concentration of As, there was a dose-dependent, but not linear, increase in AS consumption and decrease hi AS consumption with time of AS exposure. Based on the results from the 28-day study, beginning at 8 weeks of age, groups of at least 20 male p53+l~ mice on a CD and CS diet were exposed to 0.005 percent AS with or without concurrent VCD- or /j-cresidine exposure. At 6 months of treat- ment, all mice except VCD-exposed were euthanized. Mortality was less than 5 percent hi all groups, and there were no consistent clinical findings associated with treatment. Although there was no difference hi body weight gain between mice on the CD and CS diets, there was a decrease hi body weight gain in p-cresidine and AS/p-cresidirie-exposed mice. Water consumption was decreased hi arsenic-exposed mice, while food con- sumption was decreased in ^-cresidine-exposed mice. The only gross findings at necropsy were an increased thickening of bladder and an increase in bladder tumors hi p-cresidine-exposed mice on a CD diet. No necropsy data is available on VCD-exposed mice because they will be followed for an additional 4 months. These data suggest 0.005 percent AS hi the drink- ing water for 6 months did not increase the incidence of gross tumors in.p53*'~ mice on a CD or CS diet. p53*'~ mice on a CD diet are more susceptible to ^-cresidine- induced bladder tumors than^55+/ mice on a CS diet. Histolopathological evaluations will be done on hematoxylin-eosin-stained liver, skin, bladder, feet, and lung to detect microscopic changes. Total arsenic bur- den of the fur and arsenical speciation of urine, blood, liver, skin, lung, and kidney will be determined to detect differences hi tissue deposition of arsenic hi mice on CD and CS diets. Micronuclei quantitation and single-cell gel electrophoresis will be used to detect DMA damage in blood, brain, liver, lung, kidney, testis, skin, bladder, and epididymis of mice exposed to ar- senic for 6 months under different nutritional conditions. This information will be used to link exposure- dose-responses for arsenite toxicity and carcinogenicity. ------- ------- Section 2. Microbial Pathogens Grants ------- ------- Mechanisms of Inactivation of Viruses in Groundwater Maria E. Alvarez El Paso Community College, El Paso, TX Suresh Pillai Texas A&M University Research Center, El Paso, TX Extensive information exists on the inactivation kinetics of a variety of viruses in surface and ground- water. However, little is known about the molecular mechanisms of viral inactivation. The possibility that viruses may become reactivated when environmental conditions change has not been studied in detail. An understanding of the structural and compositional changes of viral particles as inactivation proceeds is critical to fully establish the quality of water supplies and the efficiency of disinfection protocols. This project's objectives are to: (1) determine the mechanisms of inactivation of MS-2 bacteriophage and poliovirus in groundwater, and (2) determine whether the viral particles that have entered the reversibly in- activated state could become infectious, and to identify the environmental factors that could promote or retard the process. Microcosm studies were performed on ground- water collected from the Rio Grande floodplain. Viral concentration was determined by the plaque assay method. Microcosm samples showing complete inacti- vation of seeded MS-2 phage at 27°C when incubated at 4°C showed up to three logs of viral reactivation after the temperature shift (see Figure 1). No reactivation was detected when poliovirus was studied under similar conditions, Indicating that the reactivation phenomenon may be unique to the bacteriophage system or that the sensitivity of the plaque assay for poliovirus does not allow detection of reactivated viruses. Structural analy- ses of viral particles as inactivation proceeds hi ground- water show the presence of virus particles with different sedimentation coefficients. These results support the hypothesis that virus inactivation proceeds in a stepwise process, which involves the rearrangement of viral components that may eventually lead to the ejection of the viral genome from the capsid. Because groundwater components interfere with protein analyses of inactivated viruses, subsequent studies were conducted hi buffered systems using chlorine as the inactivating agent. Viruses with sedimentation coefficients inter- mediate between intact and empty capsids were observed when MS-2 bacteriophage and polioviruses were inactivated with chlorine in the presence of Mg++ ions. The protein components of viruses inactivated by chlorine showed no difference hi protein components, as compared with intact viruses, when analyzed by SDS- PAGE and chromatofocusing. Studies also were con- ducted on the stability of RNA in groundwater. The re- sults indicate that naked viral RNA can persist for extended periods of time in groundwater in the presence of indigenous heterotrophic microbial populations. Viral RNA does not appear to be the primary target for viral inactivation in groundwater. Whether RNA associated with viral capsids behave in a different manner has yet to be determined. The results also indicate that reverse transcriptase-polymerase chain reaction (RT-PCR) analyses of groundwater samples could be used to detect viral RNA sequences not contained within capsids. The data obtained thus far indicate that viral inactivation hi groundwater is due to minor rearrangements of the capsid protein that may interfere with the adsorption or penetration of the virus or viral genome into the host cell. Future studies are aimed at increasing the sensitivity of the assay to detect minor changes hi viral structure as inactivation proceeds. 13 ------- MS2 Bacteriophage Poliovirus 024 8 9 Days Days Figure 1. Inactivation of viruses in groundwater at 27°C. 14 ------- Meaningful Detection of Known and Emerging Pathogens in Drinking Water Gerard A. Cangelosi University of Washington and Seattle Biomedical Research Institute, Seattle, WA Providers of safe drinking water must balance the conflicting needs of controlling microbial contamination and minimizing health risks associated with disinfection by-products. Essential to this endeavor are micro- biological monitoring methods that are practical, meaningful, and adaptable to newly discovered or emerging pathogens. Many such pathogens are difficult or impossible to cultivate in laboratory media, and polymerase chain reaction (PCR) detection of their genetic material hi water can be of uncertain sig- nificance because of the detection of dead cells or their remnants. The project's goal is to overcome these problems by using standard molecular methods to detect two nonstandard nucleic acid analytes, rRNA precursors (pre-rRNA) and bromodeoxyuridine-labeled DNA (BrdU-DNA) (see Figure 1). Both of these analytes can be detected in species-specific fashion, and both are diagnostic of cells capable of nucleic acid synthesis and growth (i.e., viable cells). Preliminary data have demonstrated the potential utility of pre-rRNA and BrdU-DNA assays in clinical diagnostic laboratories. The assays will be modified for water supply analysis and studied to answer questions regarding the survival in drinking water of two bacterial pathogens, Myco- bacterium avium and Helicobacterpylori. Bacterial model systems (Escherichia coli, M. avium, and/or H. pylori) and water samples from the Seattle Department of Public Utilities will be used to test the feasibility and utility of measuring bacterial pre- rRNA and BrdU-DNA in water supplies. The assays will be used to study the starvation kinetics of M. avium in drinking water and the resistance of this pathogen to disinfection. Also, it will be determined whether the assays can distinguish replicative (viable) from non- replicative forms of H. pylori hi water. The assays are designed to overcome the most significant challenge associated with genotypic de- tection of microorganisms in environmental samples, namely the false-positive detection of residual nucleic acid from dead cells or contaminants. If successful, these methods can help drinking water providers to know where, when, and how to control emerging pathogens in'the water supply. These tools also will yield long-awaited answers on the role, if any, of drinking water in the transmission of M. avium and H. pylori. Total (mature) Pre-iRNA rKNA Dividing cells Non-dividing (stationary phase) cells Death- phase cells Figure 1. Slot blot detection of pre-RNA and mature rRNA pools in dividing, stationary-phase, and dying cells of Mycobacterium smegmatis. 15 ------- Cryptosporidium parvum Volunteer Study: Infectivity. Illness, and Immunity Cynthia L. Chappell, Pablo C. Okhuysen, and Herbert L. DuPont University of Texas, School of Public Health and Medical School, Houston, TX Charles R. Sterling University of Arizona, Tucson, AZ Walter Jakubowski National Exposure Research Laboratory, U.S. Environmental Protection Agency, Cincinnati, OH Cryptosporidium parvum causes diarrheal disease in both immunocompetent and immunocompromised individuals. Until recently, information concerning this pathogen in humans has come from outbreak situations, case studies, and infections in travelers. In 1993, a study of Cryptosporidium infectivity and natural history of the infection in healthy adult volunteers was instituted at the University of Texas Health Science Center in Houston. This project's goal is to establish the infec- tivity of diverse C. parvum isolates in healthy adults and to evaluate the effect of preexisting serum antibody on infectivity. These studies have generated information that can be used in risk assessment of waterborne transmission. Volunteers from 18 to 55 years of age are enrolled only after they have undergone a complete history and physical examination as well as a battery of tests to ensure that they are in excellent health. Importantly, all volunteers are proven HIV-negative with normal T-cell subsets and no immunodeficiencies. After challenge, volunteers are monitored daily for the first 14 days and three times per week for an additional 4 weeks. Active surveillance of volunteers' households and/or other close contacts for diarrheal illness is maintained throughout the study period. When diarrhea occurs in the subjects or their contacts, a complete enteric workup is performed. No secondary trans- missions have been documented to date. All C. parvum oocysts are used within 6 weeks of calf production and are tested for viability by excystation rate and mouse infectivity. All of the isolates used in these studies belong to the genotype 2 (animal) subgroup of C. parvum as defined by multi-locus analysis. Fecal oocyst excretion is detected using the direct immuno- fluorescence assay (Merifluor kit, Meridian Diag- nostics). To date, three geographically diverse isolates have been studied for their infectivity hi volunteers who had no evidence by ELISA of previous exposure. Challenge doses ranging from 10 to 1 million oocysts have been given, and infectivity has been documented by the presence of fecal oocysts and/or a diarrheal illness characteristic of cryptosporidiosis. The dose necessary to cause infection in 50 percent of the volunteers (ID50) varied with isolate and ranged from approximately 10 to 1,100 oocysts (see Figure 1). Interestingly, the isolate with the lowest ID50 also was the most virulent when assessed for illness attack rate (86 percent vs. 50-55 percent). However, the onset, duration, and severity of illness did not differ among infected individuals. Protective immunity was studied in 17 volunteers who had high levels of serum IgG before challenge. In this group, both infection and illness were correlated with dosage levels exceeding 5,000 oocysts. Indeed, the ID50 was significantly increased to approximately 1800 oocysts, a 20-fold increase over the ID50 in serologically negative individuals. Subjects receiving dosage levels similar to those that might be associated with waterborne exposure were protected from infection and illness. Of those who did become infected and/or ill, fewer volunteers shed detectable oocysts. The data reveal that significant variability in infectivity and virulence occurs among C. parvum genotype 2 oocysts. Individuals with serum antibodies to C. parvum are less susceptible to subsequent exposure, especially with low numbers of oocysts. Future studies will be designed to compare infectivity and virulence of genotype 1 oocysts and to examine the crossprotection between genotype 1 and genotype 2 C. parvum. Also, studies on the infectivity of nonparvum species of Cryptosporidium are planned. 16 ------- Cumulative % infected 110 100 90 80 70 60 50 40 30 20 J pj t / / / / / / ~ — / / / / / / f f / * / f / 1 rf • -»-TAMU 012345 Challenge dose (log) Figure 1. JD50's for various isolates in volunteers. Reference of method: Reed and Muench, AmJ Hyg 27:493, 1938. 17 ------- Development of a Quantitative Cell Culture-Based Infectivity Assay for Cryptosporidium parvum in Source and Finished Water Ricardo De Leon and Paul A. Rochelle Metropolitan Water District of Southern California, La Verne, CA This project's goal is to develop a quantitative cell culture infectivity assay and a rapid molecular screening assay for Cryptosporidium parvum in finished and source water concentrates. Specific objectives include: (1) optimization of cell culture for infection with low levels of oocysts; (2) development of a molecular detection assay targeting messenger RNA from a C. parvum-spscific heat shock protein gene; (3) develop- ment of a quantitative in situ nucleic acid-based assay for detecting infections; (4) development of sample cleanup procedures that are compatible with cell culture and recovery of infectious oocysts; and (5) evaluation and validation of the method with environmental samples. The general approach involved recovery and purification of C. parvum oocysts from environmental water samples using a combination of filtration, centri- fugation, and oocyst purification, and inoculation of these oocysts onto monolayers of human cells grown on adapted microscope slides. Following incubation, infec- tious foci are detected using in situ hybridization. While developing the in situ hybridization detection method, an intermediate detection method was employed. This in- volved detection of C. parvum-speciRc gene transcripts by reverse transcriptase-polymerase chain reaction (RT- PCR) on messenger RNA extracted from infected cell cultures (see Figures la and Ib). Cell culture growth conditions were optimized to permit detection of infection with low levels of oocysts, and PCR primers were used for the specific detection of C. parvum. Using RT-PCR, infections were detected in human cells inoculated with £ 10 oocysts. A variety of immunomagnetic separation (IMS) methods were developed and evaluated as sample cleanup procedures. Oocyst recoveries of up to 100 percent were obtained, and it was demonstrated that IMS did not affect oocyst infectivity. Considerable effort was expended on col- laboration with the U.S. Environmental Protection Agency in the development and validation of Draft Method 1622, which is a new procedure for the recovery and purification of Cryptosporidium oocysts from environmental water samples. A protocol was developed that allowed the in vitro infectivity assay to be performed directly on oocysts recovered using this new procedure. The results clearly demonstrated that oocysts recovered using Draft Method 1622 retained their infectivity. The results demonstrated that in vitro cell culture combined with molecular detection methods provides a practical method for determining the infectivity of waterborne C. parvum. The overall method of oocyst recovery, cell infection, and molecular detection of infections is robust, sensitive, and specific for C. parvum. Specificity is an important consideration be- cause there are a variety of species of Cryptosporidium that may be found in water but only C. parvum is recognized as a human pathogen. In addition, it is essential that methods used to recover and purify oocysts from environmental samples do not adversely affect the infectivity of oocysts. Consequently, the demonstration that oocysts recovered using Draft Method 1622 retain their infectivity is an important finding. Future research will include continued develop- ment of the quantitative aspects of the cell culture infectivity assay with the primary focus on colorimetric in situ hybridization, comparison of different cell lines for their ability to support low levels of infection, and evaluation of the final method with environmental water samples. Immunomagnetic separation Primary detection method Intermediate detection method Figure la. Outline of approach. 18 ------- 1234567 Control: Present in all human cells C. parvum specific 1. Molecular size standards 2. Untreated oocysts 3. Oocysts recovered by immunomagnetic separation 4. Oocysts recovered using USEPA Draft Method 1622 5. Cells inoculated with sample recovered from an unspiked environmental water concentrate 6. Cells inoculated with inactivated oocysts 7. Uninfected cells Figure Ib. Detection of infectious Cryptosporidium parvum by reverse transcriptase-polymerase chain reaction on messenger RNA extracted from infected cells. 19 ------- Cultural and Enhanced RT-PCR Methods for Waterborne Caliciviruses G. Shay Font National Exposure Research Laboratory, U.S. Environmental Protection Agency, Cincinnati, OH The Biohazard Assessment Research Branch of the U.S. Environmental Protection Agency's National Exposure Research Laboratory is developing reverse transcriptase-polymerase chain reaction (RT-PCR) detection techniques for Norwalk and other calici- viruses. This project's goal has been to develop univer- sal RT-PCR procedures for processing water environ- mental samples for these nonculturable human enteric viruses. Human caliciviruses are divided into two major structural groups, each with many distinct members. Norwalk virus and other small round structured viruses such as Snow Mountain, Southampton, Hawaii, Taunton, and Bristol agents belong to one structural group; the Sapporo-like human caliciviruses belong to a second group. Members of the second group are more closely related genetically to animal caliciviruses than are members of the first group and have a classic "Star of David" morphology like the animal calicivi- ruses. Norwalk virus can be directly amplified using RT-PCR. However, RT-PCR is subject to inhibition from environmental inhibitors concentrated along with viruses during the processing procedure. These inhibi- tors affect both the reverse transcriptase enzyme that produces the DNA copy of an RNA virus genome and the DNA polymerase that amplify the DNA copy during PCR. To date, most procedures that have been devel- oped to remove inhibitors have not worked with all water sample types, or sample volumes have been too small to be used for exposure assessments. This project's objective is the detection of calici- viruses using cultural methods and RT-PCR. Powerful molecular and immunological screening methods for virus replication hi cell lines have been developed. These methods are being used to develop a cultural method for caliciviruses. A method to directly detect caliciviruses in environmental samples also has been developed. The approach removes organic and inor- ganic inhibitors by combining virus purification and concentration methods with chemical and solvent treatments. It results in an 800-fold concentration of the portion of the sample used for molecular assays, which greatly Increases the amount of sample assayed per RT- PCR reaction. The approach has been successfully eval- uated on more than 300 field samples from different geographical regions; 1 percent of the samples was positive for Norwalk virus. No information is available on the levels of caliciviruses in environmental waters. This information is important because viral gastroenteritis is a common illness in the United States. Although this disease can be caused by several virus types, it is estimated that Norwalk virus is responsible for 20-40 percent of the outbreaks of illness hi adults and in children more than 2 years of age. Primer sets to detect additional caliciviruses are currently being developed. After their development and initial testing, they will be incorporated into the method and the final method will be field evaluated again. 20 ------- Vims Filtration and Water Reuse: From Microscale Phenomena to Field-Scale Observations Stanley B. Grant Department of Civil and Environmental Engineering, University of California, Irvine, CA Many communities in the United States rely on groundwater to supply their drinking water needs. To maintain a steady-state groundwater supply, water dis- tricts increasingly use urban runoff and/or reclaimed sewage to recharge then" groundwater aquifers. An important human health concern associated with this practice is the possibility that human enteric viruses present in the recharge water may be transported through the subsurface to production wells, enter water distribution systems, and cause outbreaks of gastro- intestinal disease. This project's goal is to increase un- derstanding of the biogeochemical features of recharge basins that most influence the removal of viruses from recharge water by physicochemical filtration. The viruses that have been studied thus far include several that infect bacteria (bacteriophage) and a recombinant analog of the important human pathogen, Norwalk virus. Each of these viruses has its own uni- que electrostatic "signature" (as determined by capillary microelectrophoresis), and the electrostatic signature of each vims strongly influences its removal from water by physicochemical filtration. These results have important implications relative to: (1) the suitability of bacteriophage as indicators for viruses of genuine human health concern (like Norwalk virus), and (2) the optimal siting and operation of re- charge basins employing recycled sewage and/or urban runoff. 21 ------- Deborah Levy Centers for Disease Control and Prevention National Estimate of the Occurrence of Waterborne Disease (Abstract to be provided) 22 ------- Studies of the Infectivity of Norwalk and Norwalk-Like Viruses Christine L. Moe Deptartment of Epidemiology, University of North Carolina, Chapel Hill, NC This project's objective is to develop an under- standing of the risks associated with exposure to waterborne human caliciviruses as a function of dose and host susceptibility factors. This study will deter- mine the infectious dose of two important human caliciviruses (HuCVs), a prototype Genogroup I virus (Norwalk virus [(NV]), and a prototype Genogroup II virus (Snow Mountain Agent [SMA]), which are recog- nized as major waterborne pathogens. The specific ob- jectives are to: (1) identify the dose range of NV and SMA (ID10, ID50 and IDg,,) hi human volunteers with various levels of preexisting antibodies; (2) examine the immune response (serum and secretory antibodies) and determine the characteristics of volunteers who are susceptible to infection; and (3) evaluate the fit of several mathematical models of dose-infectivity to the data. Two double-blinded human challenge studies are proposed to determine the dose-infectivity relationships for NV and SMA. These studies will build on a pilot NV dose-ranging study previously supported by the U.S. Environmental Protection Agency. The NV Low Dose Study, with 40 subjects, will focus on infectivity hi the critical low-dose region of the dose-response curve and will provide more accurate information on the risks of NV infection associated with low levels of virus typical of the concentrations in water. The SMA Dose- Ranging Study, with 45 subjects, will be conducted in three rounds. Each round will have 15 subjects ran- domized to one of three doses that approximate the ID10, ID50, and ID^. In both studies, subjects will be moni- tored for gastrointestinal symptoms for 5 days and will return for Day 8, 14, and 21 followup visits. Stool specimens will be assayed for NV or SMA RNA by reverse transcription-polymerase chain reaction (RT- PCR). NV and SMA serum antibodies and secretory antibodies will be measured by enzyme immunoassay. Infection will be defined as excretion of NV or SMA or seroconversion. The outcomes of interest in these studies are both symptomatic and asymptomatic NV and SMA infection. Symptomatic infection is of concern because of the disease burden on the population, the effect on absen- teeism and the impact on the health care system. In terms of public health protection, asymptomatic in- fection also is of concern because of the potential for secondary transmission and the consequences of these infections for the immunocompromised population, including infants and the elderly. Several mathematical models of dose-infectivity (probit, log-linear, and beta- Poisson) will be evaluated. The results of the pilot dose-ranging study indicated that antibody status of the subject prior to dosing was a significant predictor of infection. Crude analyses (that did not control for host factors) did not indicate a simple dose-infectivity relationship. The best fit to the data was with a two-population beta-Poisson model that modeled the dose-infectivity relationship for subjects with preexisting NV IgG separately from those without preexisting NV IgG (see Figure 1). Subjects with preexisting NV IgG were infected at lower doses, and infection was related to dose. Subjects without pre- existing NV IgG tended to be younger and were pro- bably a mix of susceptible individuals who had not been previously exposed to Norwalk virus and individuals who were resistant to Norwalk virus infection. The data indicated that the latter case was more common because most of these individuals did not become infected. Preliminary estimates of risk of enteric virus in- fection and illness from drinking water using this new NV dose-response data are higher than previous esti- mates of annual risk of waterborne viral disease based on rotavirus dose-response data and using similar assumptions and models. Different dose-infectivity models yielded different risk estimates, and the range of these estimates reflects the uncertainty associated with infectivity at. low doses. Our preliminary results are consistent with observations that NV is highly in- fectious hi water; ingestion of untreated water is associated with significant risk of illness; and many waterborne outbreaks of gastroenteritis are due to human caliciyiruses. Currently, the first 20 subjects in the NV Low Dose Study are being dosed. This data will be used to refine our models of the dose-response relationship observed in the pilot dose-ranging study. The infectivity of NV and SMA hi humans will be defined in terms of RT-PCR detectable units because this is the only available method that can measure HuCVs in both clinical and environmental samples. This study will provide impor- tant date on the relationship between viral dose and susceptibility to infection (as measured by preexisting serum antibody levels and other host factors), clinical symptoms, and immune response. The results of these studies will help to estimate the risk of NV and SMA infection and gastroenteritis associated with exposure to contaminated water and establish safe exposure limits for HuCVs hi water to reduce waterborne disease. 23 ------- Q Antibody - positive O Antibody - negative Antibody Pos ' Antibody Neg »•• •**"""""" Two-population beta-Poisson model Figure 1. Norwalk virus dose response. 24 ------- PCR-Based Detection of Cytopathogenic and Noncytopathogenic Viruses in Water Ian L. Pepper, Kelly A. Reynolds, and Charles P. Gerba University of Arizona, Tucson, AZ This study focused on a new concept for virus detection utilizing a biological amplification step followed by an enzymatic amplification step. This is known as integrated cell culture (ICC) polymerase chain reaction (PCR) or ICC/PCR. This strategy combines both cultural and molecular techniques for rapid detection of viable human viruses, in large equivalent volume concentrates, without the limitations of toxicity or inhibition. The combined methodology provides the first reliable method for practical analysis and direct monitoring of environmental samples for viral pathogens posing a significant threat to public health. ICC/PCR allows for detection of infectious viruses in days compared to the weeks necessary with cell culture alone. ICC/PCR was able to detect 2.8 plaque-forming units (pfu) of poliovirus in 24 hours compared with 72 hours with conventional cell culture alone. Likewise, ICC/PCR detected 1 pfu/flaskof the noncytopathogenic virus, HAV, in 3 days compared with 14 days for cell culture. Noncytopathogenic strains of rotavirus also could be detected with ICC/PCR. This new meth- odology was able to detect low concentrations of polio- virus, that were not detected by cell culture until a third passage had been undertaken, approaching 6 weeks of incubation. Therefore, ICC/PCR was able to overcome the potential for false negative results. ICC/PCR detection of virus in environmental samples, including sewage concentrates, required 24 hours as compared with 10 days with conventional cell culture. The ICC/PCR methodology allows for detection of infectious viruses in hours to days compared to the days to weeks necessary with cell culture alone, over- coming the major flaws of cell culture and direct PCR methodologies. Another advantage to ICC/PCR is that no new method has been created, but rather two well used techniques are now being combined in a new way, enhancing the known strengths, while eliminating the known weaknesses of each method. What does this new technology mean for en- vironmental health, service, and regulatory agencies in the future? It demonstrates that a reliable method now exists for routine analysis of viruses in the environ- ment. Previously, bacterial indicator organisms have been used to determine water quality with respect to fecal contamination and potential public health impacts. These organisms do not correlate well with the presence of viruses, but a rapid, reliable method has not been available for direct virus testing, until the advent of ICC/PCR. With improved, viable virus detection sensitivity and reduced assay times (ICC/PCR requires 24-72 hours versus 5-14 days, or more, with 'con- ventional cell culture), ICC/PCR is the future for effective environmental virus monitoring. Even with samples that are suitable for direct PCR amplification monitoiring, having low inhibitory compounds and sufficiently high levels of target organisms, subsequent use of ICC/PCR would be useful to evaluate the viable nature of the target, with minimal cost and time involvement. 25 ------- Detection and Occurrence of Human Caliciviruses in Drinkini MarkD. Sobsey Department of Environmental Sciences and Engineering, University of North Carolina, Chapel Hill, NC This project's objectives are to: (1) develop improved methods to recover, concentrate, and purify human caliciviruses (HuCVs) from water; (2) develop new and improved molecular methods using reverse transcriptase-polymerase chain reaction (RT-PCR) and oligoprobing (OP; gene probe) to amplify and detect these HuCVs; (3) apply these methods to the detection of field HuCVs in environmental sewage and water samples, thereby determining HUCV occurrence; and (4) attempt to cultivate Norwalk virus (NV) in dif- ferentiated intestinal and other cell cultures and develop a practical HuCV infectivity assay system. The following methods will be further optimized and evaluated for HuCV concentration from water: adsorption to and elution from electropositive, micro- porous, Virosorb 1MDS filters, precipitation by polyethylene glycol (PEG) and possibly other pre- cipitating agents, immunocapture using polyclonal anti- bodies (human serum immune globulin) bound to paramagnetic beads, organic solvent (chloroform or fluorocarbon) extraction, and molecular exclusion by gel (Sephadex) chromatography and centrifugal ultrafil- tration. Improved methods to amplify the nucleic acid of HuCVs concentrated from water will use specific primers and probes and conditions for RT-PCR amplification and nucleic acid hybridization that will be developed in this study. In general, RT-PCR amplifi- cation and nonradioactive oligoprobe hybridization of HuCVs will be done according to previously described procedures, but will involve the selection and testing of new and unproved RT and PCR primers for amplifi- cation of genomic targets and new and improved nonradioactive oligoprobes for hybridization detection of the resulting amplicons. The methods developed in this study for concentration, purification, RT-PCR amplification and oligoprobe hybridization detection of HuCVs will be applied to field samples of water. Initially, HuCVs in water samples implicated in HuCV gastroenteritis will be detected. Later, HuCV detection will be applied to geographically representative surface and groundwaters of the United States. Some will be fecally contaminated from identifiable sources of sewage or other human fecal waste and others will be from relatively uncontaminated or "pristine" sources that are drinking waters or the protected raw (surface or ground) sources of drinking water supplies. Attempts will be made to cultivate Norwalk virus in primary cell and organ cultures, diploid cell strains, and continuous cell lines. This project will develop new methods to detect human caliciviruses and other enteric viruses hi water that will be immediately applied to field samples of water as soon as they are developed. The research will produce one or more detailed protocols or stepwise procedures, including specification of the needed ma- terials and full descriptions of methods, used to recover HuCVs as well as other enteric viruses in water and then detect them by RT-PCR amplification and oligoprobe hybridization. This project also may provide a method for propagation and infectivity assay of Norwalk virus and possibly other HuCVs. Human caliciviruses are important sources of waterborne disease hi the United States and elsewhere (see Figure 1). To better assess, prevent, and control the health risks from these viruses, unproved methods are needed for their detection in water and wastewater. 26 ------- Figure 1. Electron photomicrograph of human caliciviruses. 27 ------- Genomic Database for Crvptosporidium Species Steve J, Upton Division of Biology, Kansas State University, Manhattan, KS Christine C. Dykstra and Byron L. Btagburn Auburn University, Auburn, AL This project's objective is to acquire various iso- lates and species of Cryptosporidium from the environ- ment, and produce gDNA libraries to be deposited into the American Type Culture Collection (ATCC). In this project, oocysts are acquired from various sources, purified from feces on CsCl gradients, and surface ster- ilized using 10 percent Clorox bleach. Oocysts are then washed three times using phosphate buffered saline and centrifugation, and the DNA is extracted from each iso- late. The DNA is digested with EcoRl and/or SauSa. Fragments from the EcoRl digestions are ligated into pBluescript II plasmid vector and the plasmids used to transform competent Escherichia coli XLl-Blue cells. Fragments cut with SauSa are ligated into a bacterio- phage vector (LambdaZap express). Prior to submiss- ion of the libraries to the ATCC, quality control checks are performed. These include: (1) determination of the percentage of recombinants; (2) determination of the average insert size; and (3) probes for the presence of hsp70, the 18s rRNA subunit, and/or actin inserts. To date, enough isolates/species have been acquired so that more than 20 independent libraries are either available or soon will be (see Table 1). The acquisition of additional isolates and species is continuing. Once a large number of genomic libraries representing a variety of strains and species of Crypto- sporidium are available to researchers, investigators will be able to sequence and compare any region of the ge- nome desired. In addition, investigators will no longer need to rely solely on the sequences submitted by mul- tiple laboratories into the national databases to make comparisons. This should allow for the development of many different types of specific, rather than random, primers for developing diagnostic tests. Species C. batteyi C. meleagridis C. baileyi C.muris C.parvum C,parvum Cparvum C-parvum C.parvum Cparvum C.parvum C-parvum C.parvum C.parvum C.parvum C.parvum C-parvum C. parvum C. serpentis C.sp.% C.sp.% C.sp.% C. sp. t Origin Alabama California Georgia Virginia Alabama Colombia Colorado Florida Kansas New York Spain Massachusetts Iowa Iowa Louisiana Peru Texas Iowa Kansas Idaho Idaho California Alabama Isolate AuCbl Fresno none yet 108735 AuCpl Colombia 6304321 AuCp2 KSU-1 Cornell Galicia GCH1 Iowa(A) Iowa® Louisiana Peru TAMU UCP KSU-2 95-400 VS1742 VMTRC49 AuCml ATCC No. * * 87666 * * * * 87439 * * 87665 87667 87668 * * * * 87664 * * * * Enzyme SauSa EcoRl SauSa EcoRl SauSa EcoRl EcoRl SauSa EcoRl EcoRl EcoRl EcoRl EcoRl EcoRl EcoRl SauSa SauSa EcoRl EcoRl EcoRl EcoRl EcoRl SauSa Hosl Original Host chicken chicken turkey rock hyrax calf calf human calf calf human calf human calf calf calf human human calf corn snake beef steer dairy cow dairy cow dairy cow t Data Last Passage Host chicken chickenf turkeyt mouse calf calf N/P calf calf N/P calf calf calf calf calf calf calf calf corn snake N/P N/P N/P N/P *ATCC number not yet acquired. tCurrently under propagation. JBovine abomasal Cryptosporidium sp. N/P not passaged. sometimes termed C. muris-like. Table 1. Isolates and species of Ciytosporidium associated with EPA grant #R825148-01-0. 28 ------- Understanding Risk Factors to Crvvtosvoridium parvum: Studies in GnotoMotic Pigs Lucy A. Ward, Suk Han Cheung, Yunfei Wang, C.K. Nielsen The Ohio State University, Ohio Agricultural Research and Development Center, Wooster, OH This project's goal is to increase knowledge and understanding of risks associated with cryptosporidiosis. A gnotobiotic pig model is being used to pursue the following objectives: (1) assess and compare the patho- genesis (infectivity and virulence) of Cryptosporidium parvum in neonatal versus older gnotobiotic pigs; (2) evaluate the susceptibility and clinical responses of immunosuppressed gnotobiotic pigs to C. parvum; and (3) characterize the humoral (B cell), cellular (T cell), and cytokine immune responses in gnotobiotic pigs with cryptosporidiosis. Detailed comparative pathogenesis and immunity studies are being conducted in gnotobiotic pigs of different ages and immune status following inoculation with the human C. parvum isolate, GCH1 (from Saul Tzipori, Tufts University School of Veterinary Medi- cine, Grafton, MA) and a second C. parvum isolate (designated OH or Ohio) obtained from an Ohio Agri- cultural Research and Development Center (OARDC) laboratory worker in 1997. The median lethal dose, infectious dose, and diarrhea! dose for these two C. parvum isolates in neonatal gnotobiotic pigs have been established. As few as five GCH1 oocysts infected neonatal gnotobiotic pigs with a duration of shedding of greater than 10 days with little diarrhea. The median diarrheal doses were 730 and 6,900 oocysts for GCH1 and OH, respectively, resulting in moderate diarrheal disease of 5 to 10 day duration. Doses of > 5 x 107 GCH1 oocysts were lethal with onset of oocyst shedding and diarrheal disease observed by 24 hours post- inoculation. In contrast, although less than 50 OH oocysts infected neonatal pigs, doses of ;> 108 OH were not lethal. It is assumed that environmental isolates vary in viability, infectivity, and virulence, although little is known of this aspect. The results demonstrated that the GCH1 isolate was more virulent (approximately 10 times) in terms of infectious, diarrheal, and lethal doses compared with the OH isolate in gnotobiotic pigs. The median diarrheal doses are significantly lower than the suckling mouse model (104 to 105 oocysts) and more consistent with human and nonhuman primate studies. Further, pre- liminary studies suggest that the infectivity of either of the C. parvum isolates is not affected by the age of the host, although host age appears to influence the severity of subsequent disease. In summary, the findings sug- gest that the onset of oocyst shedding, severity of diarrheal disease, number of oocysts shed, and duration of shedding is dose- and age-dependent and notable differences exist in terms of infectivity and virulence between the GCH1 and OH C. parvum isolates in neo- natal gnotobiotic pigs. Future studies will further establish the sig- nificance of age and immune status as determinants of risk to cryptosporidiosis. These studies also will include basic immuhological studies to assess the role of the host's immune response and the risk to crypto- sporidiosis. 29 ------- ------- Section 3. Disinfection By-Products Grants ------- ------- Development of Biomarkers, in Peripheral Blood, for Exposure to and Effects of the Water Disinfectant Bv-Products Haloacetonitroles Ahmed E. Ahmed Department of Pathology, University of Texas Medical Branch, Galveston, TX Drinking waters are contaminated with a mixture of halogenated hydrocarbons that are disinfection by- products. Among those are haloacetonitriles (HAN). This project's goal is to develop unique biomarkers, in peripheral blood, for HAN exposure and cellular injury that may result from HAN-induced alkylative or oxidative damage to cellular molecules. The water disinfectant by-product dichloro- acetonitrile (DCAN) is a mutagen that is known to in- duce DNA strand breaks in cultured human lympho- blasts. The effect of DC AN in cultured mouse peritoneal macrophages (MPMs, RAW 264.7) was explored. MPMs were exposed to various concentrations (100 /tM- 400 fM.) of DCAN for 4 hours. The phagocytic activation of MPMs was characterized by the production of reactive oxygen intermediates (ROI) and secretion of TNF-a. The ratios of intracellular reduced glutathione (GSH) and oxidised glutathione (GSSG) were assessed and used as an indicator of oxidative stress. Electro- phoretic detection of DNA degradation and light micro- scopy were used for the characterization of DCAN- induced apoptosis. Lactate dehydrogenase (LDH) leak- age and trypan blue exclusion were used as markers of the necrotic effects of DCAN. Following treatment with DCAN, GSSG was increased (135% of control, P<0.05). DCAN activa- tion of MPMs was observed by elevated levels of ROI (190-250% of control, P<0.05) and increased secretion of TNF-a (450% of control, P<0.05). Electrophoresis of DNA of treated MPMs indicated a dose-dependent increase in the degradation of genomic DNA. This in- formation, combined with morphological studies, in- dicated that exposure of MPMs to lower levels of DCAN (100 /*M and 200 /*M) incites apoptic cell death. At 400 fiM DCAN, cellular necrosis was observed as indicated by increased LDH leakage and decreased viability (45% of control, P<0.05). These studies suggest that the dose-dependent DCAN-induced apoptosis or necrosis in MPMs is due to the disturbance in GSH/GSSG ratio and initiation of ROI-mediated oxidative damage mechanisms. These stu- dies should provide a basis for the development of bio- markers for regulatory guidelines and policies governing the tolerancelevels for chronic human exposure to HAN. Future steps will focus on the characterization and quantitative determination of alkylative (DNA and hemoglobin adducts) and oxidative damage to cellular macromolecules. 33 ------- Metabolic Fate of Halogenated Disinfection Bv-Products In Vivo and Their Relation to Biological Activity L.M. Ball University of North Carolina, Chapel Hill, NC Halogenated by-products of drinking water disin- fection are of concern because of their widespread human consumption and the current uncertainty over their health effects. Haloacids as a class are the most abundant of the halogenated disinfection by-products. Many of these compounds are suspect carcinogens and, in addition, may form macromolecular adducts that could be useful as biomarkers of exposure to disin- fectants. This project's goal is to elucidate mechanisms of biological activity and disposition of haloacids. Two specific approaches will be followed: one directed towards DBFs that are highly abundant, although less potent biologically, and the other directed towards DBFs that appear to be highly biologically ac- tive, though present at low levels. The first approach involves the identification and quantitation of macro- molecular adducts formed by the haloacetic acids, chlor- oacetic acid, dichloroacetic acid, from their brominated analogs—bromoacetic and dibromoacetic acid—and from the mixed haloacetic acid, bromochloroacetic acid. Structural identification and quantitation of DNA adducts will be approached by chromatographic and mass spectrometric techniques. Quantities of adducts sufficient for detailed structural elucidation are often difficult or impossible to obtain by direct modification of DNA; hence, direct chemical synthesis will be under- taken to provide standards for investigation of adduct formation in vivo. The second approach will involve investigation of the metabolism and disposition of MX, to identify the chemical nature and mechanism of forma- tion of metabolites of the potent bacterial mutagen (Z)- 2-chloro-3-(dichloromethyl)-4-oxobutenoic acid, more commonly known in its ring-closed form as 3-chloro- 4-(dichloromethyl)-5-hydroxy-2(5H)-furanone or MX (for Mutagen "X")- Structural identification of MX metabolites, and their tissue distribution in the rat, will be carried out to gain an understanding of the nature of the active species present at potential target organs, and the levels and duration of internal exposure. Synthesis and characterization of putative DNA adducts from haloacetic acids is in progress. Gaining knowledge of the chemical structures of macromolecular adducts formed will provide information on routes of genotoxic activity. In addition, this information consti- tutes the foundation for development of assays to measure rates of adduct formation and loss, and the development of biomarkers of exposure and effect. Future steps involve the comparison of techniques for sensitivity and specificity in detecting putative adducts in biological matrices. 34 ------- Evaluation of the Efficacy of a New Secondary Disinfectant Formulation By Using Hydrogen Peroxide and Silver and the Formation of Disinfection Bv-Prodncts Resulting From Interactions With Conventional Disinfectants Stuart Batterman School of Public Health, University of Michigan, Ann Arbor, MI This project's objectives are to address two critical issues associated with the use of a new secondary disinfectant formulation using hydrogen peroxide and silver. The issues are: (1) the efficacy of the formu- lation to provide long-term residual disinfection, in- cluding the control of coliform bacteria, bacterial regrowth and slime/biofilm control; and (2) the identification and quantification of disinfection by- products (DBFs) resulting from interactions with con- ventional chlorine-based and oxidant-based disinfec- tants. The project encompasses laboratory studies and field demonstrations to evaluate the efficacy of the proposed alternative disinfectant in a range of source waters and utility system characteristics. Secondary objectives include investigations related to taste and odor. The proposed secondary disinfectant is one of the few nonchlorine based disinfectants that can provide long-term residual disinfection in drinking water systems. By combining two or more disinfection agents, it is possible to lower concentrations of each component, reduce exposures, minimize DBF forma- tion, and thus minimize the health risks associated with disinfection. The approach consists of three major compo- nents: (1) laboratory evaluation of microbial disinfec- tion efficacy, including optimal doses of primary and secondary disinfectants; (2) laboratory evaluation of DBF formation potential resulting from interactions with various primary disinfectants; and (3) field demonstra- tion of the disinfectant to provide "real world" results. Inactivation performance for hydrogen peroxide ranged between 0.18 and 0.65 logs for 5 and 30 ppm, respectively, and between 0.54 and 2.87 logs for silver at 5 and 30 ppb, respectively. The combined disinfectant was more potent than each of its components alone. This synergistic effect may be attributed to combined stress conditions, exerted by a combination of two disinfectants, rather than to the formation of highly active species that are known to evolve during the reaction of hydrogen peroxide and multiple valence transition metal ions [such as Fe(H)/Fe(ffl) and Cu(I)/Cu(H)]. (Results are based on several model wild-type indicator organisms, E. coli - B, E. coli 4100/MC, and E. coli - K12, and 1 hour contact time.) The addition of the secondary disinfectant drama- tically reduced DBF levels when using chlorine as a primary disinfectant (see Figure 1). These reductions averaged 72+9 percent for trihalomethanes (THM) and 67 ±11 percent for haloacetic acids (HAA) across three types of water and two chlorination levels. (Experiments were performed at 25°C, pH = 7, 10 min contact time for chlorine, secondary disinfectant at 30 mg/L for H2O2 and 30 Mmg/L for Ag+, and DPBs determinations after 24 hr.) This reduction in by-product formation is most likely due to the reaction of Cl~ by H2O2. Followup mechanistic studies will be conducted to characterize this process. A screening level health and ecological risk as- sessment evaluated risks due to the silver that would be discharged from wastewater treatment systems if the new disinfectant formulation was adopted. The assessment uses a probabilistic three trophic-level (freshwater algae, aquatic invertebrates, and trout/carp) simulation model to simulate the bioaccumulation of silver in freshwater aquatic species. Each level considers uptake from water, food or sediment, with first-order uptake and elimination kinetics, typical feeding rates, and metal assimilation. Most of the model parameters, derived from the literature, were considered lognormally distributed. Silver discharges were simulated in three scenarios: river (high dilution, low turbidity), clear-stream (low dilution, low turbidity), and turbid stream (low dilution, high turbidity). Although some model inputs are uncertain; many predictions were comparable to field results. Although the highest (^O"1 percentile) silver concentrations in the clear-stream scenario may pose a developmental risk to some invertebrates, a chronic risk to early survival and development of trout,, and a chronic risk of argyria to subsistence fishers, predictions under most conditions show that the bioaccumulation of sEver is unlikely to result in toxic effects to aquatic wildlife and humans consuming fish. Results of this research regarding residual disinfection efficacy, disinfection by-products, optimum dosages, and risks of the new disinfectant will be compared with that of conventional disinfectants. This information is significant in exposure and risk assessments of pathogens and DBFs that support future policies and decisions regarding the most appropriate disinfection approach. Ongoing and planned future activities include: (1) further laboratory tests on disinfection efficacy of the new formulation, including viral inactivation studies; (2) further tests on DPB formation potential using the new formulation with other primary disinfectants, such as chlorine dioxide; (3) a comparative analysis of candidate disinfectants under a range of conditions; and (4) and full- scale field tests. 35 ------- ppb O H2O2/Ag Flgure 1. Reduction in THMs and HAAs as a result of addition of secondary disinfectant. Mixed ground and surface water used in the local city supply. DBFs resulting after 24 hours. Foreground - with secondary disinfectant; rear - Cl alone. Basic water parameters: Br- - 0.081mg/L, TOC = 3.1mg/L. Initial Cl = 5mg/L; residual Cl = 1.85mg/L, pH = 6.85 (buffered). TTHM = total THMs; THAAs = total HAAs. 36 ------- Integrated Approach for the Control of Ctyptosporidium parvum Oocysts and Disinfection Bv-Products in Drinking Water Treated With Ozone and Chloramines Benito J. Marinas and Roger A. Minear Department of Civil and Environmental Engineering, University of Illinois at Urbana-Champaign, Urbana, IL This project's goal is to develop process design recommendations for the simultaneous control of Cryptosporidium parvum (C. parvum) oocysts and disinfection by-products (DBFs) in natural waters treated with ozone and chloramines. Because the main ob- jective of the study is to develop an integral control strategy, the scope of work focuses on a limited number of selected DBFs (bromate, formaldehyde, and cyan- ogen halides) associated with the ozone/chloramines disinfection strategy. This project includes experimental tasks designed for the simultaneous study of C. parvum oocyst in- activation and selected DBF formation/decomposition in natural waters treated with ozone and chloramines using both batch and flow-through reactors. An integrated predictive model will be developed and calibrated with these experimental results. The model will be used to determine optimum process design, and verified hi full- scale systems using fluorescent-dyed polystyrene micro- spheres as nonbiological surrogate indicators for C. parvum oocyst inactivation. Tasks recently initiated include setting up ex- perimental and analytical systems, and formulation of a modeling approach. It is anticipated that this research project will result in a better and more integral understanding of the kinetics of DBF formation/decomposition and C. parvum oocyst inactivalion in water supply systems using ozone and chloramines as primary and secondary disinfectants. A predictive model incorporating this information will be developed for the overall optimization of both C. parvum and DBF formation control. 37 ------- The Use of Ozonation and FBT for the Control of THM Precursors in Drinking Water Susan J. Masten Department of Civil and Environmental Engineering, Michigan State University, East Lansing, MI A biological fluidized bed treatment system (FBT) in combination with ozonation was used for the control of trihalomethane (THM) precursors in Huron River water. The average annual levels of total organic carbon (TOC), trihalomethane formation potential (THMFP), and turbidity were 6.5 mg/L, 420 mg/L, and 3.3 MTU, respectively. The major water quality para- meters that were monitored included turbidity, pH, alkalinity, TOC, biodegradable organic carbon (BDOC), ultraviolet absorption measured at 254 nm (UV-254), THMFP, apparent molecular weight (AMW) distri- bution, and humic and nonhumic fractions of natural organic matter (NOM). Ozone dose, hydraulic retention time, and the concentration of dissolved ozone controlled both the destruction of organic carbon and the production of low- molecular weight organic compounds and biodegradable organic matter during ozonation of Huron River water. The study showed that biodegradable organic matter in Huron River water consisted of rapidly and slowly biodegrading fractions ("fast" and "slow" BDOC), which agreed with findings of other researchers who used different source waters. The biodegradability of water and biodegradation efficiency were characterized by the maximum biodegradation rate of "fast" BDOC (Rmax), the minimum empty bed contact time (EBCT) that was required to eliminate "fast" BDOC (EBCTmin), and by the minimum concentration of "slow" BDOC that remained in the water after biodegradation at EBCTmin (BDOCmin). The study showed that FBT was more efficient than biofiltration hi terms of NOM removal and EBCT. Among the processes investigated (single-pass ozonation/FBT, ozonation/FBT with recycle, single-pass FBT/ozonation with biofiltration, recirculating FBT/ ozonation with biofiltration), the recirculating FBT/ ozonation process followed by biofiltration was most efficient in terms of the removal of NOM relative to ozone consumption and biodegradation rate. At an ozone dose of 1 mg/mg C and total EBCT of 40-50 minutes, the removal of NOM was comparable with that achieved at the Ann Arbor Water Treatment Plant, which uses lime softening, flocculation/sedimentation, ozonation and granular activated carbon (GAC) filtra- tion to treat Huron River water. For a design capacity of 1 million gallons per day, the cost of treatment by the FBT/ozonation process followed by biofiltration was estimated to be at least 40 percent lower man that of a conventional flocculation/sedimentation process with ozonation and GAC filtration. 38 ------- Genotoxicity and Occurrence Assessment of Disinfection Bv-Product Mixtures in Drinking Water Roger A. Minear and Michael J. Plewa Department of Civil Engineering, University of Illinois at Urbana-Champaign, Urbana, IL This project addresses the questions of whether the single-cell gel electrophoresis (SCO COMET) assay using transgenic mammalian cells, rather than bacterial genotoxicity tests, can predict human risks from drinking water disinfection by-products (DBFs) and if it is sufficiently responsive to evaluate DBFs produced from different disinfection processes and under different conditions. The study is founded on the premise that mammalian cells may be more representative of direct cytotoxic and genotoxic effects from DBFs on humans than the "traditional" bacterial assays. This project's objectives are to: (1) calibrate Salmonella typhimurium and transgenic mammalian cells genotoxicity assays using regulated DBFs; (2) compare the relative genotoxicities of chlorinated versus bro- minated DBFs; (3) compare the relative genotoxicities of chlorination by-products (CBPs) versus brominated ozonation by-products (OBPs); (4) compare the relative genotoxicities of DBFs derived from singular versus combination (sequential) use of ozonation and chlori- nation; and (5) provide a DBF occurrence database for extrapolating genotoxicity results to current disinfection practice. Representative DBFs will be produced from organic matter isolated from a series of representative source waters used for drinking water supplies using both chlorination and ozonation in laboratory reactors under a range of disinfection conditions. Selective conditions will allow differential evaluation of bro- minated DBFs via ozonation of bromide containing waters and also provide information on the relation- ship of toxicity to DBF molecular weight. Bulk DBFs will be analyzed for toxicity and mutation induction in S. typhimurum using a suspension test n S9 (mam- malian microsomal metabolic activation). The same DBFs will be analyzed with transgenic Chinese hamster lung cells n S9 using the SCO COMET method to detect direct genomic DNA damage. The cytotoxicity of each sample will be determined using a microtiter well technique. The relative genetic damage per unit mass of each DBF sample will be calculated and rank ordered. The level of correlation of the DBF-induced genetic damage will be correlated both qualitatively and quantitatively with the Salmonella mutation assay and the mammalian cell SCO COMET assay. The relative genotoxic risk for specific DBFs or DBF samples will be determined. The project team has developed and calibrated a rapid, semiautomated, microplate-based cytotoxicity assay for S. typhimurium. This method provided a mea- surement of relative cytotoxicity for each DBF. The DBF standards have been evaluated and rank-ordered (see Figure 1). A novel microplate cytotoxicity assay also was developed for mammalian cells. Reverse mutation at the hisG46 and hisD3052 target genes was analyzed for each DBF with and without mammalian microsomal metabolic activation. This assay is being calibrated with DBF standards. DBFs from each of the four disinfection processes on the same natural organic matter (NOM) source were compared, and the effect of bromide on the total quantities of DBFs and their distribution was examined. DBFs analyzed using the (2,3,4,5,6-pentafluorobenzyl)-hydroxylamine(PFBHA) gas chromatographic method included nonhalogenated aldehydes. In addition, procedures developed in our laboratories were used to determine the distribution of total organic halogen (TOX) between chlorine- and bromine-containing species and to evaluate the fraction of the total organic chlorine and total organic bromine that was accounted for by known and measurable compounds. The DBF standards were assayed with S. typhimurium tester strains under suspension conditions for both cytotoxicity (repression of growth relative to the negative control) and for mutation induction at the hisD3052 and hisG46 gene targets. This is the first tune that these agents have been analyzed in such a quantita- tive manner. Cytotoxicity is usually ignored when conducting Salmonella plate incorporation tests. By using our novel microplate cytotoxicity method, the highest concentration of DBF to be the %C1/2 value was established. Concentrations then declined for the con- centration response mutagenicity experiments. These data demonstrate the relationship between cytotoxicity and mutagenicity of the DBF standards. For future Salmonella cytotoxicity and muta- genicity studies, the DBF mixture generated from known water sources will be analyzed and compared to the results from our DBF standards. Future research will focus on defining the methods for analysis of the DBF standards using SCO electrophoresis in mammalian cells. In addition, a genotoxic potency for these agents in these different assays will be determined, and their rank order and relative sensitivity will be compared. 39 ------- E 8 100 in Q O 1 75 I 50 25 r MX BromoawHcAdd Bromofbrm DibrociMMCetlcAcId Trlbromoacdlc Acid Chlcrofbrm Ethind DlmethylsuKoxldo 10-3 10-2 10-1 10° 101 Test Agent Concentration (mM) Figure 1. Comparative cytotoxicity of water DBFs measured as a repression of the growth rate as compared with the concurrent negative control in S. typhimuri strain TA100. 40 ------- Molecular Weight Separation and HPLC/MS/MS Characterization of Previously Unidentified Drinking Water Disinfection Bv-Prodiicts Roger A. Minear Department of Civil Engineering, University of Illinois at Urbana-Champaign, Urbana, IL Sylvia Barrett Metropolitan Water District of Southern California, Los Angeles, CA This project's goal is to develop new approaches for better characterizing disinfection by-product (DBF) molecular weight profiles by using tandem mass spec- trometry (MS/MS) techniques. The project will include examination of the differences in DBFs that result from different water disinfection processes. The underlying hypothesis is that new approaches are needed for this assessment, and tandem mass spectrometry, coupled with prior separations, offers promise in that regard. A prerequisite to making such procedures meaningful is the development of preseparation pro- cedures that will simplify the mass spectral data. The MS/MS system affords easy assessment of molecular weight in the first stage followed by generation of specific chemical structural data on the mass selected species distributions via measurement of related frag- ments from the selected ion. The MS/MS system, hence, has its own separation capabilities. This project is directed at enhancing these capabilities for complex DBF mixtures with preselection by molecular size sep- arations using ultrafiltration (UP) membranes and size exclusion chromatography (SEC). These preselected molecular size fractions then would be followed by other high performance liquid chromatographic (HPLC) techniques. It is expected that more than molecular weight information will be developed from the planned ex- periments. The use of several disinfection processes on the same samples also will provide additional insight as to how DBFs differ as a result of variations in the disinfection process. The specific HPLC separation processes, such as size exclusion chromatography and reverse phase chromatography, are expected to provide additional clues regarding the chemical characteristics of unidentified DBFs. The research will provide infor- mation on the molecular distribution of the DBFs col- lectively and specific information on the halogenated components distribution. IQiere are no findings to date; however, given current expectations that DBFs with molecular weights exceeding 5,000 Daltons will be of little toxicological significance, the expected results from the proposed research will provide a means for assessing the potential differences in disinfection processes for generating greater or lesser contributions to the lower molecular weight DBF fractions. This information affords an op- portunity to improve risk assessment using the mol- ecular weight criterion and better information of the location of the halogens in the less than 5,000 Dalton components of DBFs. 41 ------- Direct Quantitation of Haloacetic Acids Bv Surface Enhanced Raman Scattering MichaelJ. Natan Department of Chemistry, The Pennsylvania State University, University Park, PA There are five reasons why the surface enhanced Raman scattering (SERS) is potentially a powerful tool for environmental analysis. First, Raman spectra com- prise fingerprint-like collections of molecular vibrations that can uniquely identify low molecuar weight species. Second, water is a weak Raman scatterer, and is thus the preferred solvent for Raman measurements. Third, SERS, the process whereby molecules in close proximity to suitably roughened noble metal surfaces experience enormous increases in Raman cross section, has been shown to have single-molecule sensitivity. Fourth, Raman is a scattering process, and is particularly well- suited to analysis of solid interfaces. Finally, Raman instrumentation has been advanced to the stage where briefcase-sized instruments are commercially available. This project has focused on two areas. The first is the development of SERS detection for capillary electro- phoresis (CE). Known for its high separation efficiency, short separation time, and ease of sample handling, CE is rapidly becoming a powerful separation technique used in an extremely broad range of analytes. However, most detection methods that have been used (e.g., fluore- scence) require postcolumn derivatization with fluorescent reporters. A method in which the CE eluate can be directly deposited onto a SERS substrate that is rastered underneath the tip has been developed. In this way, separated compounds are deposited at different (known) positions. The path of the capillary is then retraced with a Raman epi-illumination probe, allowing the Raman spectra of the unlabelled analytes to be acquired. The second area concerns the design, fabrication, and testing of new, highly enhancing three-layer SERS substrates. The three-layer substrates comprise a layer of colloidal gold, a layer of chemically deposited silver (Ag), and a layer of evaporated Ag. These substrates have been shown to be highly reproducible from both intra- and intersubstrate perspectives, and have been shown to allow subpicogram detection of pesticides. 42 ------- Health Risk of Trihalomethanes Found in Drinking Water: Carcinogenic Activity and Interactions Michael A. Pereira Department of Pathology, Medical College of Ohio, Toledo, OH This project's objective is to evaluate the mechanisms through which the trihalomethanes (THMs) and chlorinated acetic acids (dichloroacetic acid, DCA, and trichloroacetic acid, TCA) cause liver cancer in B6C3F1 mice and, for the THMs, colon cancer in F344 rats. In mouse liver, the carcinogenic activity of the THM, DCA, and TCA has been proposed to result from a nongenotoxic mechanism involving their ability to in- duce cell proliferation. One of the mechanisms con- trolling the expression of the mRNA for the immediate- early protooncogenes associated with cell proliferation is DNA methylation. Methylation of the 5-position in cytosine is a normal occurrence in DNA and controls the synthesis of mRNA, including the expression of the mRNA protooncogenes. Hence, decreased methylation of their genes is expected to increase their expression. The THMs, DCA, TCA, and trichloroethylene (another environmental contaminant and a parent compound of DCA and TCA) have been determined to decrease the level of 5-methylcytosine in liver DNA and in DCA- and TCA-induced liver tumors. Upon termination of exposure, liver rumors promoted by DCA regressed and the level of DNA methylation returned to control level, while those promoted by TCA continued to progress and maintained the reduced level of DNA methylation. Furthermore, chloroform administered hi the drinking water did not affect DNA methylation corresponding to its lack of carcinogenic activity when administered by this route. Because chloroform followed by DCA and TCA are the most prevalent chlorine disinfection by- products, a study is being conducted to determine the interaction of chloroform with the two chlorinated acetic acids when administered hi the drinking water. The study will determine whether chloroform affects the po- tency of the tumor-promoting activity of the chlorinated acetic acids. , It is predicted that chloroform will de- crease the carcinogenic activity of DCA and TCA, sug- gesting that their common occurrence hi drinking water would be less hazardous than exposure to them alone. In the NTP bioassay hi rats, bromodichloro- methane and bromoform administered by gavage hi corn oil were carcinogenic hi the colon. Four THMs ad- ministered hi drinking water were evaluated for then- ability to induce and/or promote azoxymethane (AOM) induced aberrant crypt foci (ACF) hi the colon of F344 rats. ACF are putative precancerous lesions that appear to predict an increased risk of colon cancer. None of the four THMs either induced ACF or promoted AOM-induced ACF. Presently, the ability of the THM to induce cell proliferation hi the colon of F344 rats is being studied as a function of route of administration (i.e., in drinking water or by gavage hi corn oil). This study also attempts to determine whether bromodichloromethane induced colon cancer by a genotoxic mechanism. Rats are being administered bro- modichloromethane to develop tumors that will be evaluated for mutations hi the K-ras oncogene, other oncogenes, and tumor suppresser genes. This study will determine whether the mechanism of colon carcino- genesis by bromodichloromethane is genotoxic and thus suggest appropriate hazard assessment for its environ- mental (drinking water) exposure. 43 ------- Assessment of Human Dietary Ingestion Exposures to Water Disinfection By-Products via Food James H. Raymer, Gerald G. Akland, Edo D. Pellizzari, C. Andrew Clayton, and Doris J. Smith Research Triangle Institute, Research Triangle Park, NC This project's objective is to estimate the magnitude of exposure to disinfection by-products (DBFs) in drinking water via their ingestion after uptake into food during cooking. The tests have shown that foods can become contaminated with chemicals in the water used in the home during food preparation (e.g., cooking). The magnitude of this contamination process has not been studied. This research will specifically ad- dress the uptake of compounds known to arise from the process of water disinfection (ozonation in conjunction with a secondary process such as chloramination), including nonhalogenated aldehydes, ketones and acids, trihalomethanes, haloacetic acids, bromate, chloropic- rin, and haloacetonitriles. The main hypotheses to be tested are: (1) foods prepared using contaminated water become contaminated; (2) food is a significant source of DBF exposure; (3) DBF concentrations in food can be predicted with knowledge of DBF concentrations in tap water and foods consumed; and (4) dietary exposures of children are higher than for an adult living hi the same household. Analytical methods will be developed for ozonation DBFs in foods and beverages. Controlled, laboratory experiments will be conducted to determine how selected DBFs, especially those produced during ozonation, are adsorbed by food during the cooking process. Other DBF methods for foods and beverages for halogenated compounds (currently under develop- ment) will be brought in as needed. The relation between DBF concentrations hi water and in foods following cooking hi contaminated water will be modeled. This project will specifically address food items commonly eaten by children, hi addition to food items eaten by adults, so that estimates of ingestion of the study compounds to this subgroup can be de- termined. A field study will be conducted hi two cities each having different factors hi water disinfection (e.g., secondary disinfection, bromide concentration, high dissolved organic carbon) to test the validity of the model for predicting potential exposures and to estimate the human exposures to DBFs from food and water. The major benefit is that current risk assessments assume that ingestion consists of adding the con- centration levels of the specific compounds known to exist in the drinking water (multiplied by the normal assumption of volume of drinking water consumed) to the levels of the specific compounds that exist hi the food, as determined by the Food and Drug Ad- ministration hi various surveys. However, this project will attempt to verify that the estimates of exposure related to ingestion actually underestimate the actual exposures, and to estimate the amount of under- estimation for the two population groups of interest, children and adults, based on actual measurements of in- home prepared foods and drinking water. Furthermore, this research will provide new information on dietary exposure to a specific set of compounds that is currently not available. The risk assessment process will be improved hi that a greater understanding of DBF ex- posure via food will be obtained. Confirmation of the significance of this exposure pathway for DBFs will lead to more accurate risk assessments, and provide evidence for judging the extent to which children might be included hi the portion of the population that has higher total exposures. 44 ------- Analysis of Organic By-Products From the Use of Ozone/Chlorine and Ozone/Chloramines in Drinking Water Treatment David A. Reckhow Department of Civil and Environmental Engineering, University of Massachusetts, Amherst, MA This project's goal is to test, develop, and refine new and emerging analytical methods for nonvolatile organic disinfection by-products (DBFs), and to identify new ozonation and mixed ozonation/chloramination by- products. Special emphasis will be placed on new aqu- eous-phase derivatization agents. Because new deriva- tization and extraction techniques will be used, it is expected that some new DBFs will be discovered over the course of performing this research. In addition, the results from this study will fill in some of the gaps in our knowledge of differences between-DBFs hi plant effluents and DBFs at exit points in a distribution system. This information will be valuable for deter- mining research needs related to water quality in distri- bution systems. The methods studied in this research largely employ the use of new or existing, but little-used, derivatizing agents, especially those that can be used directly in aqueous samples. Early phases of this re- search are being conducted on pure solutions of known disinfection by-products, and laboratory-treated model waters. However, the work has quickly progressed to field samples from several New England water treatment facilities that use ozone as a primary disinfectant. The main thrust of this project has been to apply some of the new and emerging techniques for aqueous-phase deriv- atization to "tag" nonvolatile DBFs and render them more hydrophobic and/or more easily detected. In some cases, the derivatized compounds are subsequently transferred to an organic phase, where they are analyzed directly by gas chromatography (GC) or further derivatized prior to analysis. Where possible, GC with electron capture detection and gas chromato- graphy-mass spectrometry (GC/MS) with negative chemical ionization are being used. For compounds without strong electron affinities or for highly polar derivatives, LC/MS with electrospray is being used. A search for new DBFs is parallelling the search for new analytical methods. A comprehensive resin ad- sorption procedure is being employed that was devel- oped in the geochemistry field. This is providing the platform from which some of the new and emerging analytical technologies for identifying new nonvolatile DBFs can be used. Progress to date covers three general areas: (1) new DBF method development, (2) persistence of known DBFs in drinking water distribution systems, and (3) sample preparation with gross characterization of natural organic matter (NOM). Work on chloroformate derivatization with GC/MS has yielded good results with a wide range of poly functional organic compounds. Clean chromatograms and mass spectra have been obtained for several dozen authentic standards using hexylchloroformate as a derivatizing reagent. Linear standard curves have been developed and extended down to the high ppb range. Unfortunately, the project has not yet been able to demonstrate the higher levels of sensitivity with GC/MS that would be needed for direct analysis of water samples (i.e., the low ppb or high ppt range). Plant sampling and DBF analysis have shown that many aldehydes and ketoacids can persist well into distribution systems, when biologically active filtration is not practiced. There has been success in extracting and isolating the more hydrophilic organic constituents of treated drinking water. Gross characterizations show this material to be relatively unreactive with chlorine and low in hydroxyaromatic content (by pyrolysis GC/MS with and without methylation). Ozonation by-products (OBPs) are only partly removed through treatment and distribution in nonfiltration plants (see Figure 1). Some compounds appear to be lost in transmission/distribution, which suggests biodegradation. Others are more persistent and survive largely unattenuated to the consumers' tap. This suggests that specific ozone by-product analysis may be a sensitive test for water systems or segments of water systems that have uncontrolled biodegradation. This project is continuing with methods development for the aqueous-phase derivatization and GC/MS approach. The focus is on improving method sensitivity as well as on sample isolation and con- centration techniques. In addition, work is in progress on adapting the LC/MS methods for use with drinking water matrices. The most promising method is one developed for marine pore water that uses 2-nitro- phenylhydrazine with phase transfer catalysts. Water treatment plant sampling will continue (a new set every, 1-2 weeks) independent of these method development studies. Samples will be analyzed for known by-products (e.g., aldehydes, ketoacids). As new methods become available, they will be applied as well. 45 ------- Sampling date: 1 Oct 98 Formaldehyde Glyoxal Me-Glyoxal Glyoxalic Acid Pyruvic Acid Figure 1. Following OBPs in a nonfiltration ozonation system. 46 ------- Overview of Disinfection By-Products Research and Preliminary ICR Findings Susan D. Richardson National Exposure Research Laboratory, U.S. Environmental Protection Agency, Athens, GA Chlorine, ozone, chlorine dioxide, and chlor- amine are currently the major disinfectants being used to disinfect drinking water. Following the discovery of chloroform as a chlorination disinfection by-product (DBF) in 1974, researchers began to identify disin- fection by-products. Of the disinfectants, chlorination by-products have been the most studied, and, as a result, more than 300 chlorine DBFs have been reported in the literature (Richardson, 1998). Fewer studies have been carried out for the alternative disinfectants; however, there is some information known about their by- products. The U.S. Environmental Protection Agency (EPA) has been interested in identifying the previously unidentified DBFs with the goal of determining if there are any potentially harmful DBFs present. Researchers at EPA's National Exposure Research Laboratory in Athens, GA,'have been carrying out research to ac- complish this: goal, with an emphasis on DBFs from alternative disinfectants and on polar DBFs. Research- ers at the EPA's National Exposure Research Labora- tory in Cincinnati, OH, have been carrying out methods- related research to improve the identification of polar aldehydes, improve the extraction of polar compounds (to enable the isolation and identification of polar DBFs), and to lower the detection limits for bromate, an animal carcinogen that will be regulated under Stage 2 of the DBF Rule. Preliminary Information Collection Rule (ICR) data are now available. These data, col- lected at treatment plants across the United States, include quantitative occurrence data for aldehydes, bromate, and cyanogen chloride. References Richardson, S.D. 1998. Drinking water disinfection by-products. The Encyclopedia of Environmental Analysis & Remediation, Robert A. Meyers, editor, vol. 3, pp. 1398-1421, John Wiley & Sons, New York. 47 ------- Physiologically Based Pharmacokinetic Modeling of Haloacid Mixtures in Rodents and Humans Irvin R. Schultz Batelle Memorial Institute, Pacific Northwest Division, Richland, WA This project's goals are twofold: (1) characteri- zation of the comparative pharmacokinetics of chloro, bromo, and mixed chlorobromo haloacids (HAs) in rodents, and (2) development of a physiologically based pharmacokinetic (PBPK) model that can accurately predict the tissue distribution and elimination of HAs during chronic oral exposure in mice, rats, and humans (see Figures la and Ib). Groups of mice and rats were given either intravenous injections or gavage doses of an HA and the blood concentration-tune profile and urinary excretion characterized in individual animals. Additional groups of mice and rats were pretreated with a di-HA for 14 days at various drinking water concentrations, and the subsequent effect's of these treatments on metabolism were measured. These experiments used both in vivo (mice: blood elimination after intravenous injection; rats: CO2 formation after gavage dosing) and in vitro methods to quantify metabolic activity. The blood elimination and oral bioavailability of the HAs were characterized using compartmental and statistical moment pharmacokinetic methods. The in vitro metabolism was measured in liver homogenates and determination of the Michaelis-Menten constants (Vmax and ATm) of HAs determined using nonlinear least-squares regression methods. These data sets will be used to test and validate a PBPK model for di- and tri-HAs in the rat and mouse. Later, the PBPK model will be extended to humans and used to predict tissue levels of haloacetates during chronic, low dose exposure to HAs resulting from drinking water consumption. After the mice received intravenous dosing, their blood concentrations of all HAs declined hi a mono- or biexponential manner with a short distributive phase. There was a remarkable similarity in the extent of extra- vascular distribution among HAs, while pronounced differences existed hi the rate of excretion. The most dramatic structural feature that influenced the disposi- tion of HAs was substitution of a halogen for a hy- drogen. All di-HAs had elimination half-lives of less than 2 hours hi contrast to the tri-HAs, where the blood elimination half-life varied from 0.6 to 8 hours. The urinary excretion of all di-HAs was low and accounted for less than 3 percent of the dose hi contrast to the tri- HAs where urinary excretion accounted for at least 30 percent of me dose. All di-HAs are more rapidly meta- bolized than tri-HAs. Pretreatment with a specific di- HA greatly diminished the in vitro metabolism of all the di-HAs. These results indicate that large differences exist hi the excretion rate of HAs, which are apparently attributable to differences hi both metabolism and uri- nary elimination. The metabolism of di-HAs decreases during prolonged administration, which suggests that previous pharmacokinetic studies of HAs using control animals may be biased because our results indicate that elimination of di-HAs is reduced during chronic exposures. Therefore, estimation of target organ con- centrations of these compounds during chronic dosing may be higher than previously thought. This project will establish the linearity of HA pharmacokinetics at low doses, and the threshold drin- king water concentration that causes inhibition of metabolism for the di-HAs. These results will be in- corporated into the working PBPK model for HAs to predict tissue dosimetry during low exposure rates. 48 ------- • Inspired Air I— IC02 T • A A. [L 1 (EAaledAir) Inspired Air L- Di-HA treatment Figure la. Schematic view of the HA PBPK model. 80 60 - EQ 40-1 o O 20 - 0 J O- DCA Pretreated CD -Control •,• PBPK Model Predictions 120 CD 80 < 40 O Q 0 J T I 1 i 1 0 2 4 6 8 10 Time (Hr) 0 5 10 15 20 25 Time (hr) Figure Ib. Examples of model prediction in rats. Left panel: CO2 formation from oral doses of I4C-DCA. Right panel: Liver concentrations of DCA after gavage dosing. Pretreated rats were exposed to DCA up to 14 days prior to challenge doses. Observed data for DCA in liver is from Evans, O.B. Biochem Pharmacol 1982; 31(19):3124- 3126. 49 ------- Development of a New, Simple, Innovative Procedure for the Analysis of Bromate and Other Oxyhalides at Sub-ppb Levels in Drinking Water Howard Weinberg Department of Environmental Sciences and Engineering, University of North Carolina, Chapel Hill, NC The oxyhalide disinfection by-product bromate is most commonly associated with the ozonation of drink- ing water containing bromide and has been identified as a potential source of kidney tumors. However, accurate exposure assessment studies at the low concentrations typically found in drinking water have been hindered by the absence of sensitive analytical procedures for detection and quantitation. Health effects studies sug- gest that bromate should be regulated at 0.5/tg/L or less in drinking water. Accordingly, an analytical method that can be used routinely is needed to quantify this contaminant with great sensitivity and selectivity. This project is employing a relatively unconven- tional approach to the analysis of chemical components of drinking water. Ion chromatographic (1C) separation with no pretreatment followed by a postcolumn reaction to produce tribromide (Br3~) from bromate is applied to the analysis of a variety of aqueous samples. The tri- bromide ion is detected by ultraviolet absorbance at 267 run. This method is very sensitive for bromate with a limit of quantitation of 0.2jig/L and also is very selective. Common anions typically separated by 1C exhibit no interference, even at the levels normally found in drinking water. The method also has been optimized for similar sensitive quantitation for the oxyhalides iodate and chlorite; with the use of recently available higher capacity chromatographic columns, even lower detection levels are predicted. The meth- odology is being applied to a wide variety of drinking waters throughout the United States, including those disinfected by water utilities as well as those produced commercially and bottled. A series of controlled experiments has been performed that employ a synthetic aquatic matrix con- taining a mixture of anions together with reconstituted natural organic matter as a source of organic carbon, all dissolved in deionized water. These studies were done to determine the method's applicability to a wide range of drinking waters varying hi total organic carbon (TOC), natural bromide, and treatment. Ozonation of this water at varying ozone to TOC doses, different levels of alkalinity, pH, and bromide provided an opportunity to investigate correlations between bromate (at sub-ppb levels) and transferred ozone dose in these waters (see Figure 1). Furthermore, the analytical me- thod has been applied in practice to the determination of bromate hi a pilot ozonation plant treating a surface water with low-level bromide (< 50/ig/L) at a variety of ozone to TOC doses. The results of quality-controlled analysis of these samples indicate the applicability of this analytical method to the determhiation of sub-ppb, levels of bromate that are indeed likely to be found in ozonated waters that contain low bromide. The availability of this relatively simple meth- odology will provide the tools for assessing exposure to bromate in drinking water that has hitherto been im- peded by the lack of sensitivity of existing method- ologies. This analytical methodology will provide the U.S. Environmental Protection Agency (EPA) and the water quality monitoring community with the ability, using existing analytical equipment with simple add-on accessories, to fulfill the goals of the Information Col- lection Rule and criteria for future drinking water regu- lations by achieving the required sensitivity very easily in oxyhalide analysis. The next steps are to further simplify the meth- odology, share split sampling with EPA Cincinnati to compare and validate different methods, and to begin a comprehensive survey of U.S. drinking waters contain- ing low levels of bromide. 1.0 mA -0.5 Bromate .(0.185 |ig/L) 234 Time (minutes) Figure 1. Ion chromatographic measurement of bormate in ozonated surface water at various ozone to TOC doses (mg/mg) - TOC = 2.5 mg/L; 0 (upper), 0.25 (middle), and 0.5 (bottom). 50 ------- Inhalation and Dermal Exposure to Disinfection By-Products of Chlorinated Drinking Water Clifford P. Weisel and Jeffrey Luskin Environmental and Occupation Health Sciences Institute, Piscataway, NJ This project's goal is to determine the inhalation and dermal exposure to disinfection by-products (DBFs) of chlorination from the multiple water uses that occur within a home. The exposures will be determined by: (1) mea- surements of the concentration of the DBFs or their metabolites in urine or breath following known expo- sures; (2) estimation of dermal transport coefficients using in vitro measurements across excised skin under controlled conditions; and (3) measurement of size and number distribution of residential water aerosols gen- erated by typical water uses that result in inhalation ex- posures. During the first year, biomarker measurement methodology is being optimized. The targeted DBFs are haloacetic acids, haloketones, chloral hydrate, andhalo- acetonitriles. Subjects will be exposed to the DBFs via inhalation to volatile and nonvolatile emissions from water and by dermal absorption by direct contact to water in showers and baths. The DBFs will be added directly to purified water that is used for showering or bathing at the upper range of concentrations measured hi drinking water systems. Tune series breath levels of volatile DBFs will be measured to assess the uptake and release rate of the DBFs in the body. The methodology for haloacetic acids and halo- ketones analyses hi urine has been optimized. A 5 mL urine sample is acidified with concentrated sulfuric acid, extracted twice with ethyl ether, centrifuged, the ether evaporated, the halogenated acids and ketones methy- lated at 60 °C for 1 hour with 10 percent sulfuric acid/methanol mixture in methyl tertiary butyl ether (MTBE), sodium sulfate is added and the MTBE is analyzed by gas chromatography/electron capture de- tection (GC/ECD). A similar method is used for the water analysis, except that MTBE is used initially in place of ethyl ether. The mean percent recoveries based on replicate analyses of spiked samples were between 80 and 100 percent for all compounds except dibromoacetic acid from urine, which was 67 percent. The detection limits were all less than 0.2 ;ig/L from water and 0.5 Hg/L from urine, except bromochloroacetic acid from urine, which had a value of 1.2 jtg/L. A coeluting peak to methyl bromochloroacetate has been found in the urine extracts. Changes to the chromatographic conditions, including dual column chromatography, are being considered to resolve this problem. This project will provide data on the total DBF exposure and dose for a series of compounds that have not been previously evaluated. Fundamental data nec- essary for drinking water exposure models from a variety of water uses will be obtained, including particle size and number distributions. The preliminary results demonstrated that the proposed biomarker mea- surement methodologies have the necessary sensitivity to be used for the planned exposures at environmental levels. Human studies are to begin. Initially, ingestion studies will be done to provide a baseline on the excre- tion pattern for the DBFs when maximal metabolism will occur (see Figure 1). The DBF spiking system for the shov/er experiments and the dermal immersion system are being developed. o & £ § ! & -2000 -1000 0 1000 Elapsed Time (min) 2000 Figure 1. Temporal change in urinary excretion rates of TCAA during the ingestion exposure study. Exposure occurred at zero minute; concentration of TCAA in water was 25.5 volume of water was 500 mL; and amount of TCAA ingested was 12.75 ftg. 51 ------- Kinetic-Based Models for Bromate Formation in Natural Waters Paul Westerhoff Department of Civil and Environmental Engineering, Arizona State University, Tempe, AZ Ozone (O3) is an effective disinfectant, but it can form by-products (e.g., bromate). Bromate forms via oxidation of naturally occurring bromide through a series of steps (see Figure 1). There is a need to develop tools to understand and predict bromate (BrO3~) formation while still achieving high levels of microbial disinfection. The central hypothesis is that a kinetic- based understanding of natural organic matter (NOM) reactions with hydroxyl (HO) radicals and aqueous bromine (HOBr/OBr~) over a range of temperatures is necessary to develop mechanistic-based models for bromate formation in bulk waters. This project's objectives include: (1) developing a comprehensive database of BrO3~, O3, and HO radi- cal concentrations; (2) determining rates of reaction bet- ween HOBr and OBr~ and NOM; (3) calibrating and verifying a BrO3~ formation mechanistic-based model that includes NOM; (4) simulating BrO3~ control measures necessary to meet proposed and future MCLs; and (5) linking the numerical BrO3~ formation model with hydraulic and CT disinfection models. A mechanistic-based, numerical, kinetic BrO3~ formation program will be developed (see Figure 2). The program links an oxidant module for predicting O3 and HO radical concentrations with a BrO3" formation module. The model employs a set of bromide oxidation reactions that have been previously developed by the investigator, and calibrated against bromine and BrO3~ formation hi NOM-free water; NOM reactions now will be incorporated. The oxidant module will be calibrated against experimental O3 decay data (e.g., simple first- order decay) and HO radical concentrations (calculated from the disappearance of a HO radical probe com- pound during ozonation). Predicted BrO3~ levels will be calibrated and verified against an internal database that accounts for synergistic effects of key parameters (bromide, pH, ozone dose, temperature, inorganic car- bon, and ammonia) on ozone decay, HQ radical con- centrations, and BrO3~ formation, and an external U.S. Environmental Protection Agency database. The project will start on December 15, 1998, so there are no preliminary findings to date. f °3 !< .NOM , (Oxidation) HC Or 03 \ HO BrO/* r BrO ' NOM (Substitution) ;ano-bromine compounds Figure 1. Pathways for bromate formation. Data Input DOC, Br, pH, NH3, Temp O3 decay rate parameter Oxidant Modular O3 (+ NOM) -» nHO (+ products), k, Bromate Formation Modular Reactions Provided in Tatjle 1 Temperature Correction Relationship (Hydraulic Module i | CSTRs in Series : I CT Calculation Module j Predicted Bromate & Ozone Concentrations Figure 2. Framework for proposed model. 52 ------- Mechanistic-Based Disinfectant and Disinfectant By-Product Models Paul Westerhoff Department of Civil and Environmental Engineering, Arizona State University, Tempe, AZ David Reckhow Department of Civil and Environmental Engineering, University of Massachusetts, Amherst, MA Gary Amy . University of Colorado, Boulder, CO Zaid Chowdhury Malcolm-Pimie, Inc., Phoenix, AZ The water industry faces new challenges hi understanding and controlling disinfection by-product (DBF) formation as health concerns demonstrate a need for more stringent regulatory DBF requirements. Mechanistic tools for understanding and predicting the rate and extent of DBF formation are required to facilitate the evaluation of DBF control alternatives. A mechanistic-based numerical model will be de- veloped for chlorine decay and regulated trihalomethane (THM) and haloacetic acid (HAA) formation derived from (free) chlorination (see Figure 1); the model framework will allow future modifications for other DBFs and chloramination. Predicted chlorine residual and DBF results will be compared against predictions from several other quasimechanistic models. A signifi- cant improvement in prediction accuracy over existing empirical models is anticipated. Several modeling hy- potheses are proposed as a basis for a mechanistic-based model for disinfectant decay and DBF formation. The central modeling hypothesis is that a two-site reaction mechanism can be used to predict disinfectant decay hi the presence of natural organic matter (NOM). It as- sumes that NOM contains both slow and fast dis- infectant-reacting and DBF-forming sites. NOM site densities and concentrations are related to the con- centration, size, structure, and functionality of NOM. This project's model also will include fitted rate constants that are a function of pH and temperature. A series of distribution functions, based upon the predicted ratios of free-bromine to free-chlorine, will be used to estimate each of the four THM species (THM4) and each of the nine HAA species (HAA9). DBF experimental data from completed projects conducted by the investigators and other researchers will be integrated into a single unified database. Existing empirical models and newly developed numerical models will initially be calibrated with our unified data- base. Additional experimental data will be collected be- cause prior databases lack complete documentation of NOM characteristics before and during disinfection addition. Controlled laboratory disinfection and DBF formation studies will be conducted using water col- lected at several points through different water treatment plants, including raw, coagulated, softened, and pre- oxidized (ozone and/or chlorine dioxide) waters; thus, the waters represent a wide range of water qualities and NOM characteristics. Experiments will investigate the effects of pH and temperature, and NOM, bromide, and free-chlorine concentrations. DBF hydrolysis studies also will be conducted. The project will start on December 15, 1998, so there are no preliminary findings to date. 53 ------- Organic DBP Formation j Reaction Module (OFRM) J | Disinfectant Consumption | Reaction Module ^ (DCRM) J User-Interface & Numerical Solution Module Inorganic Reaction Module (IRM) CI2(aq, + H20 -» HOCI+H++Cr pK=-3.3 HOCI^H* + OCr pK.=7.5 HOCI+Br-5- HOBr+OCr k=2950M'1s'1 HOBr<->H*+OBr' DBP Stability Reaction Module (DSRM) DBP(x)+OH' -» DBP'(x) or X kH Hydroly5ls DBP(x)+{B} ->DBP'(x) or X kBiodegrad. where "x" is halogen and "B" is biomass Figure 1. Schematic illustration of computer program. 54 ------- Mechanisms and Kinetics of Cfaloramine Loss and By-Product Formation in the Presence of Reactive Drinking Water Distribution System Constituents Richard' L. Valentine Department of Civil and Environmental Engineering, University of Iowa, Iowa City, I A The fate of chloramines in drinking water distribution systems, the nature of the reactions and by-products involved, as well as the factors that influence these processes are largely unknown. Recent exploratory work at the University of Iowa indicates the potential importance of several ubiquitous reactive distribution system components. These substances include natural organic matter (NOM), reduced iron and manganese, iron oxides, bromide, nitrite, and oxygen. This project's goal is to enhance our under- standing of the influence of these reactive substances on: (1) the fate of monochloramine and the nature of inorganic reaction products, (2) the kinetics of mono- chloramine loss, and (3) the formation of selected organic disinfection by-products (DBFs). Results also will be used to (4) extend existing mechanistic chlora- mine reaction models to include the effects of these reactive substances. This project's overall approach is to study chlora- mine loss and DBF formation to provide a com- prehensive picture of the fate of chloramines in the presence of NOM, bromide, nitrite, complexed and uncomplexed reduced iron and manganese, iron oxides, and oxygen. Studies will be conducted using collected and laboratory prepared water in batch reactors to assess the influence of aqueous phase constituents. Reactions involving whole pieces of authentic pipe deposit material also will be studied in a recirculating tubular reactor to better mimic the fluid flow in pipes. Constituent con- centrations will be systematically varied as will be monochloramine concentration, pH, and the Cl/N ratio. Variations may include investigations of the effect of prechlorination and preozonation of the waters. Mono- chloramine concentrations will be measured over a period of at least 5 days. Samples will be taken for an- alysis of selected DBFs. Mass and redox balances will be made to try to account for the oxidizing potential of monochloramine. Results will be analyzed hi the context of an existing mechanistic description of the reactions responsible for the autodecomposition of monochlo- ramine. This project will enhance the understanding of the management of risks associated with microbial patho- gens caused by an inability to maintain an adequate disinfectant residual, and to risks associated with the formation of DBFs. Information gained will facilitate improvements hi water quality and water plant op- eration by providing water plant managers new informa- tion on the fate and effects of chloramines. New models will provide guidance hi chloramination practices and assessment of the significance of chloramine loss and DBF formation potential. 55 ------- ------- Index Note: The year of the grant award appears in parentheses following the Principal Investigator's name. Arsenic Grants Pellizzari, E.D. (1997), 3 Smith, A.H. (1997), 4 Snow, E.T. (1997), 6 Styblo, M. (1997), 7 Tice, R. (1997), 9 Microbial Pathogens Grants Alvarez, M.E. (1995), 13 Cangelosi, G.A. (1998), 15 Chappell, C.L. (1995), 16 De Leon, R. (1996), 18 Fout, G.S. (current work for EPA/NERL - Cincinnati), 20 Grant, S.B. (1995), 21 Levy, D. (current work for CDC), 22 Moe, C.L. (1997), 23 Pepper, I.L. (1995), 2^5 Sobsey, M.D. (1995), 26 Upton, S.J. (1996), 28 Ward, L.A. (1997), 29 Disinfection By-Products Grants Ahmed, A.E. (1997), 33 Ball, L.M. (1997), 34 Batterman, S. (1996), 35 Marinas, B.J. (1998), 37 Masten, S.J. (1998), 38 Minear, R.A. (1997), 39, 41 Natan, M.J. (1996), 42 Pereira, M.A. (1996), 43 Raymer, J.H. (1998), 44 Reckhow, D.A. (1996), 45 Richardson, S.D. (current work for EPA/NERL), 47 Schultz, I.R. (1997), 48 Weinberg, H. (1997), 50 Weisel, C.P. (1997), 51 Westerhoff, P. (1998), 52, 53 Valentine, R.L. (1998), 55 57 ------- ------- ------- ------- |