Vol. 47, No. 5APPLIED AND ENVIRONMENTAL MICROBIOLOGY, May 1984, p. 1005-1011 0099-2240/84/051005-07$02.00/0 Copyright (C 1984, American Society for Microbiology Comparison of Methods to Measure Acute Metal and Organometal Toxicity to Natural Aquatic Microbial Communitiest ROBERT B. JONAS,: CYNTHIA C. GILMOUR, DAPHNE L. STONER, MARGARET M. WEIR, AND JON H. TUTTLE* University of Maryland, Center for Environmental and Estuarine Studies, Chesapeake Biological Laboratory, Solomons, Maryland 20688-0038 Received 22 September 1983/Accepted 22 February 1984 Microbial communities in water from Baltimore Harbor and from the mainstem of Chesapeake Bay were examined for sensitivity to mercuric chloride, monomethyl mercury, stannic chloride, and tributyltin chloride. Acute toxicity was determined by measuring the effects of [3H]thymidine incorporation, [14C]glutamate incorporation and respiration, and viability as compared with those of controls. Minimum inhibitory concentrations were low for all metals (monomethyl mercury, <0.05 ,ug liter-1; mercuric chloride, <1 ,ug liter-1; tributyltin chloride, <5 ,ug liter-) except stannic chloride (5 mg liter-'). In some cases, mercuric chloride and monomethyl mercury were equally toxic at comparable concentrations. The Chesapeake Bay community appeared to be slightly more sensitive to metal stress than the Baltimore Harbor community, but this was not true for all treatments or assays. For culturable bacteria the opposite result was found. Thymidine incorporation and glutamate metabolism were much more sensitive indicators of metal toxicity than was viability. To our knowledge, this is the first use of the thymidine incorporation method for ecotoxicology studies. We found it the easiest and fastest of the three methods; it is at least equal in sensitivity to metabolic measurements, and it likely measures the effects on the greater portion of the natural community. Highly toxic heavy metals and organometals are common contaminants of natural waters (4, 11, 13). Sources of these substances include industrial and domestic wastewater, at- mospheric deposition, erosion, and even direct application, e.g., algicides and antifouling coatings. Research efforts have been directed toward determining "safe" levels of these substances which would not adversely affect the biota of aquatic environments. One experimental approach has been to measure concen- trations of toxic metals in a certain aquatic environment and then, by comparison with results of toxicity studies conduct- ed with standard laboratory organisms, to predict toxic impact on the natural communities within that environment. However, it is uncertain that standard laboratory organisms will respond to toxic metals in the same way as species belonging to the natural community. In another approach, indigenous organisms are removed from their natural habitat and tested under laboratory conditions. The latter experi- mental design has been used to investigate responses of microorganisms to heavy metals (3, 24). A serious problem with this method, which requires isolation and maintenance of the microorganisms on artificial media usually containing high nutrient concentrations, is that the toxicity tests are conducted under conditions bearing little relationship to the habitat from which the organisms were originally isolated. A better approach is to investigate toxic effects of metals on intact microbial communities. However, if these commu- nities are exposed to heavy metals in nutrient medium (10), (i) the medium composition would be expected to alter the physiochemical equilibrium of the metal species compared to the natural water, (ii) only lethal effects can normally be * Corresponding author. t Contribution no. 1511 of the Center for Evironmental and Estuarine Studies of the University of Maryland. t Present address: Department of Biology, George Mason Uni- versity, Fairfax, VA 22030. observed, and (iii) the selectivity of the nutrient medium limits observation to only that small part of the microbial community that can be cultured under the chosen nutrient conditions. From an ecological viewpoint, toxicity testing with micro- organisms should be conducted under conditions as similar as possible to those existing in the natural environment. For example, the effective concentration of the toxicant may be a function of the community composition, organic carbon concentration, salinity, pH, and other variables. Only a few investigators have incorporated this rationale into their experimental designs. Albright et al. (1) and Albright and Wilson (2) used the heterotrophic potential technique (16, 26) to evaluate the impact of a variety of metallic salts on[14C]glucose assimilation and respiration by an intact hetero- trophic microbial community from a freshwater pond. They attempted to mimic the pond conditions as closely as possi- ble by adding the metal salts solutions directly to freshly collected water samples, which were subsequently incubat- ed at in situ temperature. Vaccaro et al. (23) also used the kinetic approach (16, 26) to evaluate copper toxicity to microbial communities in Sannich Inlet water enclosed in a large-volume controlled ecosystem pollution experiment (CEPEX) mesocosm. Copper was added directly to the enclosures, and both acute and long-term responses were determined by measuring [14C]glucose metabolism in small samples at various time intervals. In some instances repli- cate samples from the mesocosms were amended with additional copper to determine possible changes in commu- nity sensitivity to the metal. A problem with the kinetic approach to measurement of heterotrophic potential is that only potential metabolic rates at higher than natural sub- strate concentrations are measured, although calculations derived from these measurements often give accurate deter- minations of turnover time in eutrophic systems. Values approaching natural metabolic rates can be deter- mined directly by using a radiolabeled substrate at true 1005 APPL. ENVIRON. MICROBIOL. tracer concentrations (7, 25). The tracer method has been used by Sunda and Gillespie (20) to evaluate the metabolic response of a marine bacterium to cupric ion stress and by Jonas (Ph.D. thesis, University of North Carolina, Chapel Hill, 1981) to assess the effect of cupric ion on salt marsh- estuarine microbial communities. In the tracer technique, the specific substrate is normally chosen on the basis of its probable utilization by the largest possible portion of the natural microbial community. Never- theless, even with experimental evidence that the above is true, it is unlikely that all members of the natural community are capable of transporting or metabolizing the specific substrate. Furthermore, there is no reason to assume that utilization of any carbon and energy source is necessarily the most sensitive indicator of toxicity. Therefore, we have investigated the effect of heavy metals and organometals on the incorporation of the radiolabeled nucleic acid precursor,[methyl-3H]thymidine. We hypothesized that changes in thymidine incorporation should be a better indicator of acute metal toxicity than other measurements because thymidine is more universally utilized than carbon and energy source metabolites (6) and thus could reflect toxic effects exerted upon systems other than macromolecular synthesis, e.g., poisoning of any cell function could conceivably inhibit cell growth as estimated by methyl thymidine incorporation. To test this hypothesis, we simultaneously tested the influence of heavy metals and organometals on [methyl-3H]thymidine incorporation and on glutamic acid metabolism by the micro- bial community as well as on viability of the culturable portion thereof. MATERIALS AND METHODS Sampling. Water samples (10 liters each) were collected ca. 10 cm below the surface with a 10% (vol/vol) HCI-rinsed polyethylene bucket or polypropylene bottles. Five separate samples were collected sequentially at each of two sites in Chesapeake Bay. A relatively pristine site was located alongthe mainstem of the Chesapeake Bay approximately 200 meters east of the mouth of the Patuxent River (Patuxentbuoy no. 1, 38019' N, 76022' W). Samples were collected at this site on 1 December 1982 when the surface water temperature was 10?C and the salinity 17%o. The second site was offshore from Hawkins Point in Baltimore Harbor,39018' N, 76035' W. Samples were collected from an ore dock on 20 December 1982 at a water temperature of 60C and 2%o salinity. Samples were returned to the laboratory in an insulated chest as rapidly as possible (<2 h). To provide a uniform sample representative of each site (R. B. Jonas, Ph.D. thesis), 750-ml portions of each of the five samples werefiltered through a 100-,um Nitex screen and mixed in a sterile, HCI-rinsed, 4-liter polypropylene beaker from which 10-ml subsamples were removed to appropriate polypropyl- ene incubation vessels for metal treatments. In situ water temperatures were maintained throughout. Metal treatments. The following metals or organometals were used over a range of added concentrations: mercuric mercury (Hg) was added as HgCl2 (1 to 100 ,ug ofHg liter-'); monomethylmercury (MeHg) as CH3HgCl (0.05 to 5.0 ,ug of Hg liter-'); stannic tin (Sn) as SnCl4 (5 to 100 mg of Snliter-1); and tributyltin (TBT) as [CH3(CH2)3]3SnCl (5 to 500 ,ug liter-'). Sn and TBT were obtained from Alfa Products(Danvers, Mass.), MeHg was from ICN Pharmaceuticals Inc. (Plainview, N.Y.), and Hg was from Fisher Scientific Co. (Fairlawn, N.J.). The metals were diluted in sterile distilled water (Hg and Sn) or absolute ethanol (MeHg and TBT) so that the addition of 100 ,ul of metal solution to a water sample provided the desired final concentration. Treatments were initiated by metal addition followed imme- diately by radiotracer addition where appropriate. Solvent controls consisted of water amended only with 100 ,u of distilled water or with 100 RI of absolute ethanol. [methyl-3H]thymidine incorporation. The rate of [methyl- 3H]thymidine incorporation was determined by a method similar to that of Fuhrman and Azam (5, 6). For each metal treatment and solvent control, approximately 0.5 to 1.0 ,uCi of [methyl-3H]thymidine (ICN) was added to each of four replicate 10-ml water samples contained in 50-ml polypropyl- ene screw-cap tubes. One of the four, treated with formalde- hyde (2% [vol/vol] final concentration), served as an abiotic control. After [methyl-3H]thymidine addition, the samples were incubated for exactly 59 min at in situ temperature with rotary shaking at 120 rpm. The incubation tubes were then chilled in an ice bath for 1 min followed by the addition of 10 ml of cold (4?C) 10% (wt/vol) trichloroacetic acid to termi- nate the reaction and precipitate macromolecules. The 1-h incubation period was chosen on the basis of preliminary experiments, which gave linear rates of thymidine incorpo- ration with untreated water samples over periods up to 2 h. After trichloroacetic acid addition, the samples were mixed well and held for a minimum of 15 min in an ice bath, and the particulate radioactive material was collected on 0.2-pum porosity Nuclepore filters (Nuclepore Corp., Pleasanton, Calif.). The filters were washed twice with 2-ml portions of cold 5% trichloroacetic acid, and radioactivity was assayed by liquid scintillation counting (Packard Instrument Co., Inc., Rockville, Md.; model 3330) in Aquasol II (New England Nuclear Corp., Boston, Mass.). Quench correction was by the channels ratio method. Mean disintegrations per minute for the three replicates were computed, and values for abiotic controls were subtracted before comparing treat- ment effects. [methyl-3H]thymidine incorporation in metal treated samples was compared with that of the appropriate solvent controls; organometals were compared with ethanol solvent controls, and metals were compared with distilled water solvent controls. Glutamic acid metabolism. The effect of metals on incorpo- ration and respiration of glutamic acid was determined in a manner analagous to that described above by a modification of the tracer level technique of Williams and Askew (25). Approximately 0.1 ,uCi of L-[U-14C]sodium glutamate (ICN) was added to each 10-ml sample contained in a 20-ml screw- capped polypropylene vial fitted with a polypropylene cup suspended from the interior surface of the cap. The cup contained a filter paper wick and 0.2 ml of hyamine hydrox- ide (New England Nuclear) to trap 14CO2 (12). After incuba- tion for 1 h at in situ temperatures, 0.3 ml of 0.4 N H2SO4 was injected through a silicone rubber-sealed port in the vial cap to stop activity and to liberate 14CO2 from the water. The samples were then shaken on a rotary shaker at 120 rpm for 1 h at room temperatrure to permit 14CO2 absorption. 14CO2 trapping efficiency was 86% by this procedure. The filter paper wicks were removed to scintillation vials for counting, and the cups were rinsed with several volumes of Aquasol. Particle-associated radioactivity was collected from the acidified water and assayed as described above, except that the filters were washed with 5 ml of Nelson salt solution (14) of the appropriate salinity. Preliminary experiments with untreated Chesapeake Bay water indicated that rates of glutamate incorporation and respiration were linear over a minimum of 1.5 h of incuba- tion. Acid treatment of the water to remove 14C02 resulted 1006 JONAS ET AL. ACUTE METAL TOXICITY METHODS 1007 in a 37% decrease in particle-associated radioactivity com- pared with samples filtered immediately after incubation. This loss has been attributed to the release of unmodified radiolabeled substrate pools (8). Therefore, particle-associ- ated radioactivity measured in these experiments likely represents true incorporation rather than accumulation of glutamate into intracellular pools (22). Viability testing. The bactericidal effects of the metals and organometals on the culturable fraction of the microbial community were assessed as in the radiotracer experiments. Metal solutions or appropriate solvents were added to single 10-ml samples contained in polypropylene tubes. After 1 h of incubation at the in situ temperature, portions of each sample were diluted in bicarbonate buffered (pH 7.7) Nelson salts of the appropriate salinity. Appropriately diluted sam- ples were then spread onto half-strength Nelson agar plates at pH 7.7 and in situ salinity. Colonies were enumerated after incubation for 10 days at 22 + 2?C. The data for these experiments are expressed as CFU. RESULTS Thymidine incorporation, glutamate metabolism, and CFU. Preliminary experiments indicated that thymidine incorpo- ration, glutamate metabolism, and CFU were the same in raw water as in raw water with 100 ,ul of distilled water. Likewise, there were no significant differences between values for samples amended only with distilled water or ethanol at either sampling site (Table 1). Control values for all parameters remained the same when measured before or after the 1-h time period required for the initiation of all the different metal treatments. Therefore, comparisons made between metal treatments and their respective controls are valid. Thymidine incorporation, glutamate metabolism, and numbers of culturable microorganisms were uniformly high- er in Baltimore Harbor water than in mid-Bay water (Table 1). Turnover time of thymidine was threefold faster in Baltimore Harbor than in the mid-Bay and compared well with the threefold greater number of culturable microorga- nisms in the Baltimore Harbor water. On the other hand, glutamate metabolism was considerably higher in Baltimore Harbor than would be predicted from comparison of thymi- dine turnover and CFU between the two sites. The major difference was in the rate of glutamate incorporation. Gluta- mate incorporation and respiration by the mid-Bay microbial community represented approximately equal portions of glutamate metabolism (velocity of metabolism/velocity of incorporation, 2.1), whereas in Baltimore Harbor water, incorporation significantly exceeded respiration (velocity of metabolism/velocity of incorporation, 1.7). These results suggest differences not only in the composition of the two communities but also in their physiological state. Toxicity of Hg and MeHg. The sensitivity of the microbial communities to added Hg at the two sites was approximately the same whether determined by thymidine incorporation or glutamate incorporation and respiration (Fig. 1A and B). Thymidine incorporation was the most sensitive indicator of Hg toxicity at both sites. Although the degree of inhibition was the same at most Hg concentrations, the data suggest that the community at the mid-Bay site was somewhat more sensitive to Hg (-90% inhibition at 1 ,ug of Hg liter-) than that at Baltimore Harbor (-60% inhibition at 1 p,g of Hg liter-1). Glutamate incorporation and respiration were inhib- ited equally as the Hg concentration was increased. Although thymidine incorporation and glutamate metabo- lism detect either bactericidal or bacteriostatic effects, CFU responses clearly indicated that Hg was bactericidal (Fig. 1A =xEx X CLE 14 D"l El o o 0 P S 0 " 0 ;; < -500 TBT Baltimore Harbor <5 <5 5-10 50-100 Despite the similarity of results with the thymidine and glutamate methods in this study, thymidine incorporation, a measure of cell growth, integrates many metabolic and biosynthetic pathways and thus could generally be used to detect a wider range of cell damage than metabolism of any single carbon and energy source. Except in the case of MeHg treatment of Baltimore Harbor water, the CFU technique was a significantly less sensitive indicator of metal toxicity. Therefore, it would be unwise to rely on this method alone to determine potential toxicity of heavy metals and organometals to natural micro- bial communities. Nevertheless, inhibition of thymidine incorporation or glutamate metabolism does not indicate whether the mode of action of these compounds was bacte- riostatic or bactericidal. It is clear from the CFU data that all four compounds were bactericidal, at least to the Baltimore Harbor community (Fig. 1 and 2). The sensitivity of the CFU test might be increased by lengthening metal exposure times, but previous data for copper toxicity indicates that 1 h is sufficient for development of the full bactericidal effect on CFU (Jonas, Ph.D. thesis). The mid-Bay culturable microbial community was consis- tently less sensitive to the bactericidal effects of the metals and organometals tested than was the Baltimore Harbor community. This result contrasts with thymidine incorpo- ration and glutamate metabolism determinations, which indi- cated that the microbial communities at both sites exhibited roughly similar sensitivities to the test compounds. The comparative insensitivity of the culturable portion of the mid-Bay community to the two organometals is striking. The explanation for these observations may lie simply in the fact that the culturing technique selects only a portion of the total metabolically active community and the metal sensitivity of the culturable community differs from that of the whole. For example, it is possible that a large part of the Baltimore Harbor culturable microbial community consisted of zymog- enous bacteria derived from upland sources. Alternatively, the Baltimore Harbor community may have been stressed in its natural state by metals or other toxic compounds, and that small additional stress, i.e., metal treatment, may have caused a relatively greater inhibitory effect than in the less- contaminated mid-Bay. The data do not provide a choice between these alternatives, but the approximately 1:1 ratio of assimilation and respiration of glutamate at both sites (Table 1) suggests that neither community was severely stressed. Rather it seems likely that the culturable communi- ties at the two sites differed significantly in their respective relationship to the entire community. The data reported here indicate that some of the com- pounds tested are considerably more toxic to microorga- nisms than previously reported. Hallas and Cooney (9), using a technique in which metals are incorporated into solidified media, reported that the most sensitive bacterial isolates from Chesapeake Bay were inhibited by less than 1 mg ofTBT tin liter-1 and by about 12 mg of SnCl4 tin liter-'. Our data indicate that the MIC in the intact community was less than 5 jig liter-' for TBT and less than 5 mg liter-' for Sn. Singh and Bragg (17) reported that 2 mg of TBT liter-' inhibited amino acid metabolism in Escherichia coli by 50%. In water disinfection applications, approximately 500 to 1,000 ,ug of TBT liter-1 was needed to inhibit Leginonella pneumophila (19). As little as 1 ,ug of Hg liter-' reduced metabolic activity of the Cheaspeake Bay and Baltimore Harbor communities by 50 to 90%, and 100 ppb was completely bacteriostatic (Fig. 1). In a freshwater pond 100 ppb of added Hg had no effect on CFU density and reduced glucose metabolism by only about 85% (1). Similarly, Singleton and Guthrie (18) found that in mixed freshwater and brackish water ecosystems 40 ,ug of Hg liter-' had no effect on culturable bacteria. The data presented here indicate that intact microbial communi- ties in Chesapeake Bay are extremely sensitive to Hg. The actual MIC may be considerably lower than 1 ,ug liter-1, the lowest concentration tested (Fig. 1). Because the exposure environment can markedly influ- ence metal speciation and toxicity (20, 21), the quantitative relationships between metal concentration and community inhibition are somewhat uncertain. However, our data are conservative in the sense that the active form of an inorganic metal (usually the free ion activity) is less than the total concentration of that metal. Organometals are probably also toxic as the covalently bound free cation form and may exist in aquatic systems primarily in that state. Therefore, the MIC for organometals are also likely to be conservatively high. At comparable concentrations (Fig. 1) MeHg was no more toxic than Hg. This suggests that methylation of Hg to the more volatile MeHg may be an effective means of protecting microorganisms from the toxicant. This suggestion is sup- ported by results of Pan-Hou and Imura (15), which show that mercury methylation by Clostridium cochleanium T-2C is a primary mechanism of protection against mercuric mercury in this anaerobe. LITERATURE CITED 1. Albright, L. J., J. W. Wentworth, and E. M. Wilson. 1972. Technique for measuring metallic salt effects upon the indige- nous heterotrophic microflora of a natural water. Water Res. 6:1589-1596. 2. Albright, L. J., and E. M. Wilson. 1974. Sub-lethal effects of several metallic salts organic compounds combinations upon the heterotrophic microflora of a natural water. Water Res. 8:101- 105. 1010 JONAS ET AL. ACUTE METAL TOXICITY METHODS 1011 3. Boney, A. D. 1970. Sub-lethal effects of mercury on marine algae. Mar. Pollut. Bull. 2:69-71. 4. Forstner, V., and G. T. W. Wittmann. 1979. Metal pollution in the aquatic environment. Springer-Verlag, Berlin. 5. Fuhrman, J. A., and F. Azam. 1980. Bacterioplankton second- ary production estimates for coastal waters of British Columbia, Antarctica, and California. Appl. Environ. Microbiol. 39:1085- 1095. 6. Fuhrman, J. A., and F. Azam. 1982. Thymidine incorporation as a measure of heterotrophic bacterioplankton production in marine surface waters: evaluation and field results. Mar. Biol. 66:109-120. 7. Gocke, K. 1977. Comparisons of methods for determining the turnover times of dissolved organic compounds. Mar. Biol. 42:131-141. 8. Griffiths, R. P., F. J. Hanus, and R. Y. Morita. 1974. The effects of various water-sample treatments on the apparent uptake of glutamic acid by natural marine microbial populations. Can. J. Microbiol. 20:1261-1266. 9. Hallas, L. E., and J. J. Cooney. 1981. Effects of stannic chloride and organotin compounds on estuarine microorganisms. Dev. Ind. Microbiol. 22:529-535. 10. Hallas, L. E., and J. J. Cooney. 1981. Tin and tin-resistant microorganisms in Chesapeake Bay. Appl. Environ. Microbiol. 41:466-471. 11. Helz, G. 1976. Trace element inventory for the Northern Chesa- peake Bay. Geochim. Cosmochim. Acta 40:573-580. 12. Hobbie, J. E., and C. C. Crawford. 1969. Respiration correction for bacterial uptake of dissolved organic compounds in natural waters. Limnol. Oceanogr. 14:528-532. 13. Ketchum, B. H., V. Zitko, and D. Saward. 1975. Aspects of heavy metal and organohalogen pollution in aquatic ecosys- tems, p. 75-90. In A. D. McIntyre and C. F. Mills (ed.), Ecological toxicology research-effects of heavy metal and organohalogen compounds. Plenum Publishing Corp., New York. 14. Nelson, J. D., and R. R. Colwell. 1975. The ecology of mercury- resistant bacteria in Chesapeake Bay. Microbial Ecol. 1:191- 218. 15. Pan-Hou, H. S., and N. Imura. 1982. Physiological role of mercury-methylation in Clostridium cochlearium T-2C. Bull. Environ. Contam. Toxicol. 29:290-297. 16. Parsons, T. R., and J. D. Strickland. 1962. On the production of particulate organic carbon by heterotrophic processes in seawa- ter. Deep-Sea Res. 8:211-222. 17. Singh, A. P., and P. D. Bragg. 1979. The action of tributyltin chloride on the uptake of proline and glutamine by intact cells of Escherichia coli. Can. J. Biochem. 57:1376-1383. 18. Singleton, F. L., and R. K. Guthrie. 1977. Aquatic bacterial populations and heavy metals. I. Composition of aquatic bacte- ria in the presence of copper and mercury salts. Water Res. 11:639-642. 19. Soracco, R. J., and D. H. Pope. 1983. Bacteriostatic and bactericidal modes of action of bis(tributyltin) oxide on Legion- ella pneumophila. Appl. Environ. Microbiol. 45:48-57. 20. Sunda, W. G., and P. A. Gillespie. 1979. The responses of a marine bacterium to cupric ion and its use to estimate cupric ion activity. J. Mar. Res. 37:761-777. 21. Sunda, W. G., and R. R. L. Guillard. 1976. The relationship between cupric ion activity and the toxicity of copper to phytoplankton. J. Mar. Res. 34:511-529. 22. Tuttle, J. H. 1980. Organic carbon utilization by resting cells of thiosulfate-oxidizing marine heterotrophs. Appl. Environ. Mi- crobiol. 40:516-521. 23. Vaccaro, R. F., F. Azam, and R. E. Hodson. 1977. Response of natural marine bacterial populations to copper: controlled eco- system pollution experiment. Bull. Mar. Sci. 27:17-22. 24. Walker, C. W., Jr., and C. W. Houston. 1981. Toxicity of cadmium to bacteria. Biotechnol. Lett. 3:437-442. 25. Williams, P. J. L., and C. Askew. 1968. A method of measuring the mineralization by micro-organisms of organic compounds in seawater. Deep-Sea Res. 15:365-375. 26. Wright, R. T., and J. E. Hobbie. 1965. The uptake of organic solutes in lake water. Limnol. Oceanogr. 10:22-28. VOL. 47, 1984