600 J 00 Ascorbic acid reduction of active chlorine prior to- determining Ames mutagenicity of chlorinated natural organic matter (NOM)| Edward T. Urbansky* and Kathleen M. Schenck United States Environmental Protection Agency (EPA), Office of Research and Development, National Risk Management Research Laboratory, Water Supply and Water Resources Division. Cincinnati, Ohio 45268 USA. E-mail: Urbanskv.Edward@EPA.gov; Schenck.Kathleen@EPA.gov; Fax: +1 513 569 7658; Tel: +1 513 569 7655 Received 15th November 1999, Accepted 28th January 2000 Published on the Web 29th February 2000 M2P>y potable water disinfection byproducts (DBFs) that result from the reaction of natural organic matter (NOM) with oxidizing chlorine are known or suspected to be carcinogenic and mutagenic. The Ames assay is routinely used to assess an overall level of mutagenicity for all compounds in samples from potable water supplies or laboratory studies of DBF formation. Reduction of oxidizing disinfectants is required since these compounds can kill the bacteria or react with the agar, producing chlorinated byproducts. When mutagens are collected by passing potable water through adsorbing resins, active chlorine compounds react with the resin, producing undesirable mutagenic artifacts. The bioanalytical and chemoanalytical needs of drinking water DBF studies required a suitable reductant. Many of the candidate compounds failed to meet those needs, including 2,4-hexadienoic (sorbic) £cid, 2,4-pentanedione (acetylacetone), 2-butenoic (crotonic) acid, 2-butenedioic (maleic and fumaric) acids and buten-2-ol (crotyl alcohol). Candidates were rejected if they (1) reacted too slowly with active chlorine, (2) formed ftiutagenic byproducts, or (3) interfered in the quantitation of known chlorination DBFs. L-Ascorbic acid reacts- rapidly and stoichiometrically with active chlorine and has limited interactions with halogenated DBFs. ID this work, we found no interference from L-ascorbic acid or its oxidation product (dehydroascorbic acid) s£i mutagenicity assays of chlorinated NOM using Salmonella typhimurium TA100, with or without metabolic activation (S9). This was demonstrated for both aqueous solutions of chlorinated NOM and concentrates derived from the involatile, ether-extractable chlorinated byproducts of those solutions. Aim of investigation The study of disinfection byproducts (DBFs) in water intended for human consumption continues to be a concern for the EPA. In particular, the potential presence of carcinogens and mutagens remains a focus in the assessment of the risks associated with chlorinated potable water supplies. To measure this potential, the Ames assay1'2 for mutagenicity has been applied to field studies of DBF formation in drinking water and laboratory studies of chlorinated natural organic matterf (NOM). Many halogenated byproducts have been identified and may be directly quaatitated.3"5 However, because of the nature of NOM, a considerable number of uncharacterized and potentially genotoxic compounds may result from its reaction with oxidizing (active) chlorine compounds used for disinfec- tion. It is neither possible nor practical to identify and quantitate all of these. Consequently, the Ames assay provides a convenient assessment of mutagenicity as a bulk property and allows for whatever synergy or additivity naturally exists. This laboratory has generally extracted DBFs from potable .waters using polymeric adsorbent resins, specifically, Amber- • 'tlite® XAD-2 and XAD-8 (Rohm and Haas, Philadelphia, PA, USA).2'4'6 This approach has been criticized .because the fThis paper is the work product of United States government employees engaged in their official duties. As such, it is in the public domain and exempt from copyright restrictions. © US government. Jin keeping with the current convention, we have chosen the term natural organic matter as opposed to such terms as humic acid, humate, fulvic acid, or fulvate. NOM is often poorly characterized and substantially variable from one source water to another. Because thiols, alcohols, amines, and other moieties can contribute significantly to the composition of NOM, such terms as humic acid are misleading and no more descriptive. residual active chlorine reacts with the resin and produces a small contribution to the overall mutagenicity of the sample.7 While this contribution is small, it is nonetheless a source of debate since the artifactual mutagenic compounds remain unidentified. Because DBFs continue to form in the distribution system, the kinetics of DBF formation and thus variation of mutagenicity with time are of interest. However, attempts to monitor the Ames mutagenicity as a function of time have been hampered by the obvious complication that oxidizing chlorine compounds are bactericidal. Dechlorination with common reductants, e.g., S2Osz~ or SOs2", caused measurable loss of chlorinated organic compounds and mutagenicity.7"10 Experi- mentally, it is difficult to neutralize residual oxidant without adding a small excess of reducing agent, and residual reductant concentrations below 30 uM were shown to affect results.8 In 1989, Croue and Reckhow showed that sulfite rapidly attacked trichloroethanenitrile (Cl3CC=N), dibromoethanenitrile (Br2CHC=N), and trichloronitromethane (chloropicrin, C13CNO2).8 In 1985, Wilcox and .Denny demonstrated that thiosulfate, sulfite, weta-bisulfite, sulfur dioxide, and biotin all reduced the Ames mutagenicity of chlorinated water sample, although biotin did not exert this effect when added to a solvent extract.9 As early as 1980, Cheh et al. found that adding sulfite to reduce residual chlorine led to a loss of mutagenicity and suggested that post-chlorination addition of sulfite might be used to degrade mutagenic compounds in the treatment process.10 Due to the lack of a satisfactory reducing agent, our laboratory was resigned to accepting a mutagenic artifact resulting from reaction of halogen oxidants with organic matter, i.e., XAD® resins or agar. Attempts to find workable chlorine-scavenging agents focused on a variety of alkenes, ------- such as 2,4-hexadienoic (sorbic) acid, butenoic (crotonic) acid, (jE)-butenedioic (fumaric) acid, (Z)-butenedioic (maleic) acid, buten-2-ol (crotyl alcohol), and 2,4-pentanedione (acetylace- tone, better represented as the enol, 4-hydroxypent-3-ene-2- one), or on carboxaldehydes, such as dextrose. Nonetheless, all of these were rejected on account of either insufficiently facile reaction with active chlorine, mutagenicity of reaction products, or both. L-Ascorbic acid, better known as vitamin C, is known to react stoichiometrically with iodine1' and rapidly with hypo- chlorous acid.'2 In reaction with halogens or hypohalous acids, ascorbic acid is oxidized to dehydroascorbic acid, which exists in several forms and can itself decompose.13 Accordingly, L- ascorbic acid was deemed a possible choice for scavenging chlorine in potable water samples or quenching large excesses of active chlorine in studies on DBF formation. The characteristics ascribed to an ideal reducing agent were described previously.14 It has been shown that L-ascorbic acid interacts minimally with halogenated hydrocarbons (EPA Method 551.1AB analytes)14 and halocarboxylates (EPA Method 552.1 analytes).15 Both of these classes of compounds are found as DBFs and can be determined directly as opposed to relying on a bioassay for evidence of mutagenic activity. In drinking water, one of the mutagens of greatest activity is (Z)- 2-chloro-3-dichloromethyl-4-oxo-2-butenoic acid (Mutagen X),6 and so any interaction with Mutagen X would proscribe use of a particular reducing agent. Fortunately, Fukui et al. demonstrated that exposure to L-ascorbic acid at 5 mM for up to 1 h resulted in no loss of MX.16 L-Ascorbic acid had no effect on mutagenicity of a number of materials, including coal dust, tobacco, diesel emissions, and fried beef.17 A number of investigations have demonstrated a reduction in Ames mutagenicity due to L-ascorbic acid; however, all of these involved nitric oxide, nitro- or Af-nitrosp-compounds,18"21 aflatoxin,22'23 or peroxides.24 These compounds do not result from chlorination except for trichloronitromethane. Peroxides would not be expected to survive in a raw water, and certainly not after chlorination. If the other compounds were in fact present in a potable water supply, we would expect to find them in both raw and finished water as they are not DBFs. Any reduction in mutagenicity due to interactions should therefore be observed in tests of both raw and finished water and the effects should cancel out. Experimental procedure§ Reagents Standard chlorine solutions. Commercially available solu- tions of 5.0% w/w aqueous sodium hypochlorite (NaOCl) were standardized by iodometric analysis, kept under refrigeration at 7+2°C, and used without modification. Standard thiosulfate solution. A solution of 0.10 M Na2S2O3 was prepared by dissolving the pentahydrate salt in doubly deionized water. The solution was preserved with 0.1 % w/w Na2COa to prevent disproportionation and a few drops of HPLC grade CHC13 to prevent growth of Thiobacillus thiopants. Standard iodine solution. Into a 1000 ml volumetric flask, 2.85 g KIO3, 600ml of doubly deionized water, and 170ml 0.25 M H2SO4 were added. A mass of 37.7 g KI was dissolved in a minimum of doubly deionized water. While stirring magnetically, the KI solution was slowly added to the>iodic acid solution. Doubly deionized water was added to volume. The following comproportionation [eqn. (1)] leads to a §Mention of specific manufacturers or brand names should not be construed as endorsement by the United States government. standard solution that contains primarily triiodide, I3 : IOJ- + 81 -+6H •*•->• 3Ij- + 3H2O (I) The excess iodide (~0.1 M) ensures the solubility and limits the volatility of the I2 by forming the triiodide ion. Some precipitation of I2 was initially observed, but the iodine dissolved with 2h of continuous stirring. Total oxidant: Mr = M + [V] = 0.0400 M. Stock NOM solution. A mass of 120 g humic acid (Fluka, Buchs, Switzerland) was combined with 800 ml doubly deionized water with constant stirring. The pH was adjusted to 7.0 with dropwise addition of 50% w/w sodium hydroxide, and the mixture was stirred for 3 h. The mixture was divided into 250ml polypropylene tubes and ce'ntrifuged for 20min. The centrifugates were recombined, and the pH was readjusted to 7.0. Double deionized water was added to bring the final volume to 1000 ml, and the mixture was permitted to settle overnight. The mixture was divided and centrifuged. again. The clear liquid portion was decanted into a low actinic glass bottle and stored at 4 °C. Triplicate 3 g samples of stock NOM solution were weighed into tared vials and dried overnight at 11,0°C; the remaining solid was 8.63% (by mass) of the initial sample mass. The dried solid was collected and sent, to Microanalysis, Inc. (Wilming- ton, DE, USA) for elemental analysis. Found: 49.7 % C, 3.4 % H, 3.4 % N, 16.0% ash. Calculation of the total organic carbon concentration from these values gives [C]org= 42.9 gl"1 =3.57 M. Working NOM solution. Stock NOM solution was diluted with doubly deionized water to produce a solution containing 2.1 gdl"1 C. The chlorine demand of this solution was determined using five replicates as follows: Into 40 ml PTFE- lined screw-cap glass vials, a 10.00 ml aliquot of working NOM solution was dispensed. Subsequently, 3.00 ml of standard chlorine solution and 3.0 ml of 0.25 M H2SO4 were added. The vials were capped and shaken for several minutes, then placed into a constant temperature bath at 45 °C for 3.5h. When cooled to room temperature, a 10.00ml portion (excess) of standard thiosulfate solution was added to each vial. Vials were recapped and repeatedly inverted for about 10 min. Vial contents were quantitatively transferred to Phillips beakers and diluted to about 70 ml with deionized water. To each, 7 ml of 9M H2SO4 and 2g of KI were added. While stirring magnetically, the excess thiosulfate was back-titrated with standard iodine solution to a starch-iodide end point. To reduce foaming and improve visibility of the end point, a few drops of propanone were added just prior to equivalence. Chlorine demand is calculated as the ratio of chlorine consumed relative to organic carbon (molar basis): A[Cl->]r/ A[C]org = A([C12] + [HOC1] + [CIO -])/A[C]org =0.99. Preparation of mutagenic samples Chlorinated NOM. To a 1 000 ml volumetric flask, the following were added: 60.0 ml of stock NOM solution, 100ml of phosphate buffer (0.50 M NaH2PO4 + 0.50 M Na2HPO4), and 141 ml of stock chlorine solution (0.708 M); this represents 94% of the calculated chlorine demand. Doubly deionized water was added to bring the volume to 1000 mi. The mixture was permitted to react for 24 h; iodometric titrimetry showed -90% of the active chlorine (HOC1) had been consumed. Consequently, most of the DBFs should have been formed. The post-chlorination solution pH was 6.59. The solution was divided into 100ml portions. Some were reserved without further treatment; however, within another 24 h, no chlorine could be detected in the unreduced aliquots by starch-iodide. Although this means that additional chlorinated byproducts were formed, we assume that the major contribu- 162 J. Environ. Monit., 2000. 2, 161-163 ------- tors to mutagenicity are derived from the 90% of the active chlorine that had already reacted. To other portions, ascorbic acid was added at 2.0 gdl~'; this resulted in immediate reduction of the remaining 10% of the active chlorine (0.07 M HOC1) as indicated by starch-iodide. Volumes of both solutions (unreduced and reduced) were collected and frozen at —15 °C pending mutagenicity testing. Extraction and concentration. Three unreduced and three ascorbic acid-reduced (100ml) portions of the aqueous solution were extracted with equal volumes of pesticide residue analysis grade tert-butyl methyl ether (MTBE). Extracts were rotary evaporated under house vacuum at 60 °C to remove the bulk of the MTBE. Solid residues were dried for 72 h at 45 °C to remove any traces of MTBE and then redissolved in 10.0 ml of HPLC grade dimethyl sulfoxide (DMSO). Assuming that partitioning strongly favors the MTBE, this step should concentrate the involatile extractable constituents by a factor of ~10 relative to the aqueous solutions above. The DMSO extracts were collected and frozen at — 15°C pending mutagenicity testing. Ames mutagenicity assay Mutagenic activity was determined hi Salmonella typhimuriwn strain TA100 with and without metabolic activation, using the standard plate method of Maron and Ames.1 Genetic markers for TA100 were verified before use. Spontaneous and positive control (l.Oug NaN3 plate"1 and l.Oug 2-aminoanthracene plate"1) responses and solvent controls were incorporated in each assay. L-Ascorbic acid and unhalogenated NOM were tested separately and found to be non-mutagenic. The samples were assayed using five dose levels, with duplicate plates per dose. A repeat assay was done on a separate day. For the aqueous solutions, the highest dose was 1000 ul; for the DMSO concentrates, the highest dose was 200 \ii. The liver homo- genate (S9) was prepared from Aroclor 1254-induced male, Sprague-Dawley rats (Molecular Toxicology, Inc., Boone, NC, USA) and the S9 cofactor mix was prepared as described in Maron and Ames.1 The S9 concentration in the S9 mix was 5% v/v, and 500 ul of the mix was added per plate. Results and discussion Mutagenicity (revertants per unit volume dose) was expressed as the average slope, calculated by covariance analysis. There are no significant differences between the slopes of the aqueous chlorinated NOM with and without ascorbic acid (P=0.1069) or between the slopes of the DMSO concentrates with and without ascorbic acid (P- 0.7486) based on covariance analysis of the without S9 data (Table 1). These results therefore suggest that ascorbic acid does not reduce the observed mutagenicity of chlorinated NOM, whether the aqueous solution is tested directly or a concentrated extract is tested. Taking into account the intrinsic variability associated with the Ames assay, the level of mutagenicity observed in the Table 1 Ames mutagenicity (revertants uF1) for chlorinated NOM samples Sample type HAsc added? Without S9 With S9 Aqueous solutions DMSO concentrates N Y N Y 0.20±0.01" 0.17 + 0.01 1.53 + 0.08 1.50 + 0.06 0.07 + 0.01'' 0.09 + 0.01A 0.25 ±0.05'' 0.46 + 0.04'' "Standard error of the average slope of the dose-response data deter- mined by covariance analysis. ''The highest dose used in the assays did not produce a two-fold increase above background (solvent con- trol value), although there were marginal dose-related increases in the number of revertants observed. concentrates is about ten-fold that of the aqueous solutions. This would seem to confirm our assumption that the mutagenic compounds are both involatile and MTBE-extractable. In the presence of S9, the levels of activity in each of the samples are reduced, consistent with results from previous studies of chlorination byproducts.4 It appears that ascorbic acid can satisfactorily meet both the chemoanalytical and bioassay needs associated with monitor- ing DBF formation, especially in a laboratory study. None- theless, this work does not demonstrate that ascorbic acid is suitable under all conditions and for all water supplies. Investigators wishing to use this compound to scavenge residual chlorine are cautioned to carefully evaluate the potential matrix interactions. Certainly, ascorbic acid is the most promising candidate we have found thus far, and we anticipate it will prove useful hi future investigations of chlorination byproduct formation. Acknowledgements We acknowledge Mano Sivaganesan for assistance in statistical analyses and Betty Merriman for assistance in preparing the stock NOM solution and in carrying out the bioassay. References 1 D. M. Maron and B. N. Ames, Mutat. Res., 1983,113, 173. 2 K. M. Schenck, J. R. Meier, H. P. Ringhand and F. C. Kopfler, Environ. Sci. Techno!., 1990, 24, 863. 3 J. R. Meier, R. B. Knohl, W. E. Coleman, H. P. Ringhand, J. W. Munch, W. H. Kaylor, R. P. Streicher and F. C. Kopfler, Mutat. Res., 1987, 189, 363. 4 K. M. Schenck Patterson, B. W. Lykins, Jr. and S. D. Richardson, /. Water Supply Res. Technol.—Aqua, 1995, 44, 1. 5 S. D. Richardson, in Encyclopedia of Environmental Analysis and Remediation, ed. R. A. Meyers, Wiley, New York, 1998, pp. 1398- 1421. 6 K. M. Schenck, L. J. Wymer, B. W. Lykins, Jr. and R. M. Clark, Chemosphere, 1998, 37, 451. 7 A. G. Sweeney and A. M. Cheh, J. Chromatogr., 1985,325,95, and references cited therein. 8 J.-P. Croue and D. A. Reckhow, Environ. Sci. Techno!., 1989, 23, 1412. 9 P. Wilcox and S. Denny, in Water Chlorination, Environmental Impact, and Health Effects, ed. R. L. Jolley, R. J. Bull, W. P. Davis, S. Katz, M. H. Roberts, Jr. and V. A. Jacobs, Lewis, Chelsea, MI, 1985, vol. 5, pp. 1341. 10 A. M. Cheh, J. Skochdopole, P. Koski and L. Cole, Science, 1980, 207, 90. II D. C. Harris, Quantitative Chemical Analysis, Freeman, New York, 3rd edn., 1991. p. 413. K2 L. K. Folkes, L. P. Candeias and P. Wardman, Arch. Biochem. Biophvs., 1995, 323, 120. 13 J. Hvoslef and B. Pedersen, Acta Chem. Scand. Ser. B, 1979, 33, 503. 14 E. T. Urbansky, J. Environ. Monit., 1999. 1, 471, and references cited therein. 15 E. T. Urbansky, D. M. Freeman and F. J. Rubio, submitted for publication. 16 S. Fukui, S. Ogawa, T. Motozuka and Y. Hanasaki, Chenwsphere, 1991,23, 761. 17 T. Ong, W. Z. Whong. J. D. Stewart and H. E. Brockman. Mutat. Res., 1989. 222, 19. 18 S.-C. Park, Korean J. Biochem., 1980, 12. 9. 19 E. P. Norkus and W. A. Kuenzig, Curcinogenexis. 1985, 6. 1593. 20 S. Bala and I. S. Graver. Mutat. Res.. 1989, 222, 141. 21 P. L. Arroyo, V. Hatch-Pieott, H. F. Mower and R. V. Cooney. Mutai. Res., 1992.281, 193. 22 R. K. Bhattacharya, A. R. Francis and T. K. Shetty. Mulai. Res.. 19S7, 188, 121. 23 W.-Z. Whong, J. Stewart, H. E. Brockman and T.-m. Ong. Teratog., Carcinug., Mutagen., 1988, 8, 215. 24 M. Minnunni. U. Wolleb, O. Mueller, A. Pfeifer and H. U. Aeschbacher, Mutat. Res., 1992, 269, 193. Paper a909046k J. Environ. Monit.. 2000. 2, 161-163 163 ------- |