Response of Microorganisms to an Accidental Gasoline Spillage in an Arctic Freshwater Ecosystem

Transcription

1 APPLIED AND ENVIRONMENTAL MICROBIOLOGY, June 1977, p Copyright ) 1977 American Society for Microbiology Vol. 33, No. 6 Printed in U.S.A. Response of Microorganisms to an Accidental Gasoline Spillage in an Arctic Freshwater Ecosystem A. HOROWITZ AND R. M. ATLAS* Department of Biology, University of Louisville, Louisville, Kentucky Received for publication 31 January 1977 The response of microorganisms to an accidental spillage of 55,000 gallons of leaded gasoline into an Arctic freshwater lake was studied. Shifts in microbial populations were detected after the spillage, reflecting the migration pattern of the gasoline, enrichment for hydrocarbon utilizers, and selection for leadedgasoline-tolerant microorganisms. Ratios of gasoline-tolerant/utilizing heterotrophs to "total" heterotrophs were found to be a sensitive indicator of the degree of hydrocarbon contamination. Respiration rates were elevated in the highly contaminated area, but did not reflect differences between moderately and lightly contaminated areas. Hydrocarbon biodegradation potential experiments showed that indigenous microorganisms could extensively convert hydrocarbons to CO2. In situ measurement of gasoline degradation showed that, if untreated, sediment samples retained significant amounts of gasoline hydrocarbons including "volatile components" at the time the lake froze for the winter. Nutrient addition and bacterial inoculation resulted in enhanced biodegradative losses, significantly reducing the amount of residual hydrocarbons. Enhanced biodegradation, however, resulted in the appearance of compounds not detected in the gasoline. Since the contaminated lake serves as a drinking water supply, treatment to enhance microbial removal of much of the remaining gasoline still may be advisable. On August 5, 1976, an accidental gasoline spillage of leaded MOGAS was detected at the Naval Arctic Research Laboratory located at Barrow, Alaska (Cdr. R. H. Schaus, personal communication). The source of the spillage was found to be a break in a pipeline buried in a gravel pad at the permafrost level. It is not known how long the pipe had been broken before the spill was detected. Based on the amount of gasoline missing from the storage tanks, it was estimated that 55,000 gallons of MOGAS had leaked into the environment. Several similar spillages of refined oils of this or greater magnitude have occurred in association with construction of the trans-alaska pipeline (A. Carson, paper presented at the 27th Annual Alaska Science Conference, Fairbanks, August 1976). The spilled gasoline moved through the gravel pad over the permafrost and entered a nearby lake. Some of the gasoline was observed to enter the lake through the sediment, bubbling up into the water column. The contaminated lake supplies the drinking water for the base; the water inlet is located at the end of the lake opposite the source of contamination, an approximately 250-m distance. Immediate steps 1252 were initiated to try to physically contain and recover the spilled gasoline; an earthen dam was constructed to contain the gasoline in a shallow portion of the lake, and wells were dug in the gravel for gasoline to seep into, so that it could be pumped back into storage tanks. It was estimated that half of the spilled gasoline was recovered by this process. Shortly after detection of the spillage, we began studying the response of the microbial community to the contaminating gasoline. These studies included determination of changes in levels of microbial populations and the ability of indigenous microorganisms to degrade the hydrocarbon components of the spilled MOGAS. MATERIALS AND METHODS Sampling. Three areas of the contaminated lake were chosen for study. The first area (HC) was heavily contaminated. It was near the gravel pad and consisted of small, shallow pools of water that were completely covered with gasoline. The second area (MC) showed moderate visible contamination. Gasoline was visibly seeping from the sediment, and the water surface was partially covered with gasoline. The surface gasoline film was greatly affected by wind. The third, a lightly contaminated area (LC),

2 VOL. 33, 1977 was beyond the earthen dam built to contain the spill and showed no visible gasoline contamination at the beginning of our study. One month after the spill had been detected, the sediment in this area showed visible gasoline contamination, but no surface slick was observed. Water and sediment samples were collected weekly starting on August 16th in each area. Samples were collected in sterile glass containers and returned to the laboratory immediately for analysis. The temperatures of the samples at the times of collection were between 0 and 5 C. Enumeration of microorganisms. Viable "total" heterotrophic microorganisms were enumerated in duplicate by surface spreading serial dilutions on Trypticase soy agar. Gasoline-utilizing and MOGAStolerant microorganisms were enumerated by surface spreading onto Bushnell Haas agar containing 0.5% (vol/vol) emulsified MOGAS as sole carbon source (1). Gasoline-utilizing microorganisms were also enumerated by surface spreading onto Bushnell Haas agar with no added carbon source and incubating in a chamber with an atmosphere containing volatile gasoline hydrocarbons (1 ml of MOGAS in a petri dish/i-liter chamber volume). Gasoline-utilizing microorganisms enumerated on the volatile MICROBIAL RESPONSE TO ARCTIC GASOLINE SPILLAGE 1253 hydrocarbons need not be tolerant of lead, which organisms enumerated on emulsified MOGAS medium must tolerate. MOGAS for these enumerations was obtained from the physical recovery wells in the gravel pads. Analysis of the MOGAS by atomic adsorption and gas-liquid chromatography (Fig. 3A) showed it to be partially weathered and to contain 1,800,ug of lead per ml. Incubation for all enumerations was at 10 C for 7 days. Counts for water samples are expressed per milliliter and for sediment samples as per gram of wet weight. Statistical analyses on the counts were performed by an analysis of variance procedure. Respiration. Respiration rates of sediment samples were measured in duplicate with a Gilson respirometer. Seven-gram portions of sediment samples and 5 ml of 0.2% acetate, 0.05% NH4NO3, and 0.006% Na2HPO4 were added to the flasks. The flasks were incubated at 10 C with 80 oscillations per min. Hydrocarbon biodegradation potential. Tengram sediment samples were incubated in Biometer flasks (Bellco) for 4 days at 10 C with a 10-ml solution of 10 mm NH4NO3 and 0.5 mm Na2HPO4 and 25,ul of Prudhoe crude oil spiked with either [14C]hexadecane (specific activity, 4,iCi/ml) or ['4C]pristane (specific activity, 2 uci/ml). Pristane and hexadecane are representative branched- and straight-chain alkanes that are available 14C radiolabeled. The addition of the ['4C]hexadecane in crude oil is a modification of the method of Walker and Colwell (10) using "4CO2 hydrocarbons for measuring hydrocarbon biodegradation activity. "4CO2 was trapped in NaOH in the flask side arm. After incubation, sediment was separated from microbial cells by filtration through Whatman no. 1 paper, and the microbial cells were recovered on 0.2-,umpore size membrane filters (Millipore). Some cells were undoubtedly absorbed on the sediment particles and not recovered by this procedure. Residual hydrocarbons were recovered by extraction with diethyl ether. The extracts were placed on 70- to 230- mesh silica gel 60 columns (5 by 1 cm, Merck) with 1 cm of Na2SO4 on top of the column, and the paraffinic fraction containing the untransformed radiolabeled hexadecane or pristane eluted with hexane. The efficiency of recovery for hexadecane was 89% and 60% for pristane. Biodegradation of MOGAS. Sediment collected in the HC area was divided into 30-g portions. The sediment portions were placed in open glass containers, and the chambers were replaced in situ. Some of the sediment portions were amended with NH4NO3 and Na2HPO4, 10 and 0.5 mm final concentrations, respectively. In addition to nutrient supplementation, some of the sediment portions were inoculated with 108 viable cells of a mixture of two hydrocarbon-utilizing bacteria, strains ADA-M20 and ADA-M27. Both organisms were gram-negative, nonmotile rods isolated by the sequential enrichment method on oil depleted by bacterium RAG- 1 (4) from estuarine water at Pt. Barrow, Alaska. Both organisms could grow at temperatures from 5 to 25 C and salt concentrations from 0.5 to 10% NaCl, and both could metabolize several carbon sources other than hydrocarbons. Other sediment portions were neither supplemented nor inoculated. After 0, 3, and 5 weeks of incubation, two replicate samples for each treatment were acidified, and the residual hydrocarbons were extracted with diethyl ether. The ether extracts were dried with sodium sulfate and evaporated under nitrogen. The extracted material was weighed and analyzed by gas-liquid chromatography. Gas-liquid chromatography was performed with a Hewlett-Packard model 5830 flame ionization gas chromatograph with 10% Apiezon L columns (2 m by 3 mm, outer diameter), N2 carrier at a flow rate of 30 ml/min, programmed at 40 C isothermal for 8 min, 8 C/min to 250 C, and 250 C isothermal for 20 min. RESULTS Changes in levels of microbial populations. Numbers of viable counts of microorganisms were significantly greater in samples from the more heavily contaminated areas than from the less heavily contaminated areas (Fig. 1 and 2). Populations in the HC area were greater than in the MC area, and populations in the MC area were greater than in the LC area at the 97% confidence level. In sediment (Fig. 1), microbial counts were high in the HC area throughout the sampling period. In the LC area, there was an initial drop, probably due to toxicity, followed by a rise of three orders of magnitude during 4 weeks in sediment counts, followed by a severe drop in the last sample. The initial drop and rise corresponded with the influx of MOGAS into the LC area. The MC area sediment counts showed an intermediate pattern.

3 1254 HOROWITZ AND ATLAS APPL. ENVIRON. MICROBIOL. LC MC HC - U 6 z SAMPLING DAT E FIG. 1. Enumeration of microbial populations from sediment samples. Symbols: *, Trypticase soy agar; A, emulsified gasoline-bushnell Haas agar; *, gasoline vapors-bushnell Haas agar. Downloaded from In water (Fig. 2), levels of microbial populations were generally two orders of magnitude higher in the HC area than in the MC area and two orders of magnitude higher in the MC area than in the LC area. Counts of gasoline utilizers on MOGAS vapors were higher than on emulsified MOGAS. Organisms enumerated on emulsified MOGAS must be tolerant to lead. The ratios of gasoline utilizers to total heterotrophs are shown in Table 1. The ratios of gasoline utilizers enumerated on MOGAS vapors to total heterotrophs were generally greater than one. The ratios of gasoline utilizers enumerated on emulsified MOGAS to total heterotrophs were generally less than one. During the sampling period, these later ratios were low in the LC area water, gradually rose from low to high in the LC area sediment and MC area water, and were high throughout the MC area sediment, HC area water, and HC area sediment. Respiration. The rates of oxygen consumption were similar in sediment from the LC and MC areas (Table 2). Rates of respiration were much higher in the HC area than in the other areas. In the HC area, there was a rise in the rates of oxygen consumption over a 1-month interval during the sampling period, followed by an abrupt decrease in the last sample. The lake began to freeze at the last sampling time. Hydrocarbon biodegradation potential. Only the first sample collected in the LC area showed limited ability to convert branched- and straight-chain alkanes to C02; all other samples showed high conversion rates to CO2 (Table 3). In general, the rates of conversion to CO2 were similar for hexadecane and pristane. Relatively little radiolabeled hydrocarbon was incorporated into the recovered cells. Somewhat less radiolabeled hydrocarbon was incorporated by cells in the HC area samples. After arn initial peak, there was a general decrease in the rates of "4CO2 production in all samples. Degradation of MOGAS. Weight losses of MOGAS from sediment samples collected in the HC area and incubated in situ with various treatments are shown in Table 4. The MOGAS had already been exposed for several weeks and had undergone some degradative losses prior to the beginning of this experiment. Further natural abiotic and biodegradative processes resulted in an 85% loss after 3 weeks, with an additional 5% loss during the next 2 weeks. Thus, after 5 weeks of exposure, 10%o of the on November 14, 2018 by guest

4 VOL. 33, 1977 MICROBIAL RESPONSE TO ARCTIC GASOLINE SPILLAGE 1255 LC MC HC L. 6 z * 8-/23 U31 9/7 S14 9,21 SAMPLING DATE FIG. 2. Enumeration of microbial populations from water samples. Symbols: *, Trypticase soy agar; A, emulsified gasoline-bushnell Haas agar; M, gasoline vapors-bushnell Haas agar. TABLE 1. Ratios ofgasoline-utilizing microorganisms to heterotrophic microorganisms from areas of differing degrees of MOGAS contaminationa Gasoline-utilizing microorganisms/heterotrophic microoganisms from: Type of Date HC MC LC sample GA/TSA BA-G/TSA GA/TSA BA-G/TSA GA/TSA BA-G/TSA Water 8/ / / / / / Sediment 8/ / / / / / a See Materials and Methods for description of sampling areas. Abbreviations: GA, emulsified gasoline- Bushnell Haas agar; TSA, Trypticase soy agar; BA-G, gasoline vapors-bushnell Haas agar.

5 1256 HOROWITZ AND ATLAS ether-extractable components of the MOGAS remained in the sediment. Treatment with nutrient supplementation increased degradative losses, and only 4% of the ether-extractable compounds remained after 5 weeks. Inoculation with oil-utilizing bacteria in addition to nutrient supplementation resulted in more rapid degradative losses, with only 4% of the extractable MOGAS components remaining after 3 weeks. Gas chromatographic analysis of the extractable compounds showed marked changes in the relative concentrations of MOGAS components (Fig. 3). As expected, there was generally a decrease in the lighter components, with more extensive degradation. Unexpectedly, however, extensive degradation resulted in the appearance of a number of compounds of greater retention times that were not detectable in MO- TABLE 2. Sampling date Respiration: oxygen consumption in sediment samplesa Oxygen consumed (,Il/h per g) in sediment samples from: HC MC LC 8/ / / / / a See Materials and Methods for description of sampling areas. GAS at the beginning of the experiment, even when the undegraded MOGAS was analyzed at higher concentrations. These new compounds were the major peaks detected after 5 weeks of exposure with nutrient addition and inoculation. No effort was made to identify these compounds. DISCUSSION Microbial populations in an Arctic freshwater ecosystem were found to respond rapidly to the presence of contaminating MOGAS. There was a rapid population shift to high numbers of gasoline hydrocarbon-utilizing and leaded MO- GAS-tolerant organisms. The ratios of hydrocarbon-utilizing to total heterotrophic microorganisms varied with the degree of MOGAS contamination and appears to be a useful, sensitive indicator of environmental hydrocarbon con- TABLE 4. Degradation of spilled MOGAS Treatment TABLE 3. Hydocarbon biodegradation potentiala % Wt loss in samples exposed for: 3 weeks w weeks None Ammonium nitrate and phos phate Ammonium nitrate and phos phate + inoculation Conversion rate (cpm, x 1,000) of: APPL. ENVIRON. MICROBIOL. Determination Date Hexadecane Pristane HC MC LC HC MC LC 14C2 produced 8/ / / / / / C incorporated 8/ into cells 8/ / / / Residual 14C 8/ paraffin 8/ / / / / a See Materials and Methods for description of sampling areas.

6 VOL. 33, 1977 MICROBIAL RESPONSE TO ARTIC GASOLINE SPILLAGE 1257 w i c j ~ ~~~~~~,, Iiiiiiii,.I B A gil liii, 1i, A-.1 a I1 dl,,1ill 1,1.j1,,,,. F 2.~~~~~~~~~~~~~1 d 1lli 1l.l1 I,l,Ġj RETENTION TIME (MIN) FIG. 3. Relative percentages of compounds in degraded MOGAS samples determined by gas-liquid chromatography. (A) MOGAS leaking from broken pipe. (B) MOGAS extracted from sediment at start of experiment. (C) MOGAS extracted from untreated sediment after 3 weeks ofexposure. (D) MOGAS extracted from nutrient-treated sediment after 3 weeks. (E) MOGAS extracted from nutrient-treated-plus-inoculated sediment after 3 weeks. (F) MOGAS extracted from untreated sediment after 5 weeks. (G) MOGAS extracted from nutrient-treated sediment after 5 weeks. (H) MOGAS extracted from nutrient-treated-plus-inoculated sediment after 5 weeks. (a) Retention time for standard 2-methylpentane. (b) Retention time for standard toluene. (c) Retention time for standard n-octane. (d) Retention time for standard ethylbenzene. (e) Retention time for standard m-xylene. (f) Retention time for standard n-nonane. (g) Retention time for standard n- decane. (h) Retention time for standard naphthalene. (i) Retention time for standard 1-methylnaphthalene. (j) Retention time for standard phenylnonane. (k) Retention time for standard n-hexadecane. (1) Retention time for standard pristane. tamination for accidental spillages. Counts of hydrocarbon utilizers after exposure to MO- GAS exceeded conventional plate counts for total heterotrophs, indicating enrichment for hydrocarbon-utilizing populations. High numbers of hydrocarbon utilizers and high percentages of total heterotrophic populations that can utilize hydrocarbons have previously been reported as indicators of chronic oil pollution in temperate areas (1, 8). Jamison et al. (6) also found the presence of many hydrocarbon-utilizing microbial species after a gasoline spillage into groundwater in a temperate region. The observation of higher numbers of microorganisms enumerated on MOGAS vapors than on media containing emulsified MOGAS probably reflects the toxicity of lead and possibly other components in MOGAS to hydrocarbon-utilizing microorganisms. In the HC area, the difference was less between counts on emul-

7 1258 HOROWITZ AND ATLAS sified MOGAS and MOGAS vapors than in other areas, suggesting that the microorganisms proliferating in this area had undergone selection for tolerance to MOGAS components as well as for ability to utilize hydrocarbons. Rates of respiration in the HC area also indicated that the microorganisms there were carrying out extensive metabolism in the presence of MOGAS. Microorganisms in the contaminated lake were found to be capable of degrading hydrocarbons in MOGAS. There have been relatively few reports on the ability of microorganisms to degrade gasoline (5, 6). There have been, however, a number of accidental gasoline and other refined-oil spillages into Arctic ecosystems associated with construction ofthe trans-alaskan pipeline as well as similar spillages in more temperate ecosystems. The hydrocarbon biodegradation potential experiments showed that the microorganisms exposed to MOGAS were capable of extensively converting contaminating hydrocarbons to CO2. The lower rate of 14C incorporation into the cells in the HC area and the decreasing activities with time of exposure to MOGAS in all three areas could be due to several reasons, including the diluting effect of the nonlabeled MOGAS hydrocarbons or increased concentrations of toxic MOGAS components after weathering and biodegradative losses. The lower activities could also be due to the production of toxic compounds from biodegradation (2) or to the depletion of an essential element such as iron (3). The in situ biodegradation experiments showed that abiotic weathering and biodegradation resulted in extensive losses from the spilled MOGAS, but when the lake began to freeze for the winter, 10o of the extractable material still remained in the untreated sediment. Nutrient addition effectively increased losses, and inoculation, together with nutrient supplementation, further enhanced degrada- APPL. ENVIRON. MICROBIOL. tive losses. Enhanced biodegradation resulted in only 3% extractable material left when the lake began to freeze; the residue had a very different compound distribution than MOGAS prior to extensive biodegradation. The residue after extensive biodegradation had a predominance of higher-retention-time compounds that may have been synthesized during biodegradation. The appearance of such higher-retentiontime compounds has previously been reported in laboratory experiments by Horowitz et al. (4), Pritchard et al. (7), and Walker and Colwell (9). The persistence and potential toxicity of these compounds is unknown. The addition of fertilizers would appear to be effective in the abatement of the effects of the gasoline spillage. Hydrocarbon analyses of water at the drinking water inlet after the spillage occasionally have shown hydrocarbon levels in excess of 1 mg/ml, and filtration steps have had to be taken to purify the drinking water (Cdr. R. H. Schaus, personal communication). Enhanced biodegradation can remove the MOGAS hydrocarbons remaining in the water and sediment, decreasing the hazard of continued use of this lake as a drinking water supply. ACKNOWLEDGMENTS The work reported in this paper was supported by Off'ice ofnaval Research contract no. N C We wish to thank the Naval Arctic Research Laboratory, Barrow, Alaska, for their generous assistance. LITERATURE CITED 1. Atlas, R. M., and R. Bartha Abundance, distribution and oil biodegradation potential of microorganisms in Raritan Bay. Environ. Pollut. 4: Atlas, R. M., and R. Bartha Inhibition by fatty acids ofthe biodegradation of petroleum. Antonie van Leeuwenhoek; J. Microbiol. Serol. 39: Dibble, J. T., and R. Bartha Effect of iron on the biodegradation of petroleum in seawater. Appl. Environ. Microbiol. 31: Horowitz, A., D. Gutnick, and E. Rosenberg Sequential growth of bacteria on crude oil. Appl. Microbiol. 30: Jamison, V. W., R. L. Raymond, and J. 0. Hudson Biodegradation of high-octane gasoline in groundwater. Dev. Ind. Microbiol. 16: Jamison, V. W., R. L. Raymond, and J. 0. Hudson Biodegradation of high-octane gasoline, p In J. M. Sharpley and A. M. Kaplan (ed.), Proc. 3rd Int. Biodeg. Symp. Applied Science Pub. Ltd., London. 7. Pritchard, P. H., R. M. Ventullo, and J. M. Sulfita The microbial degradation of diesel oil in multistage continuous culture systems, p In J. M. Sharpley and A. M. Kaplan (ed.), Proc. 3rd Int. Biodeg. Symp. Applied Science Pub. Ltd., London. 8. Walker, J. D., and R. R. Colwell Enumeration of petroleum-degrading microorganisms. Appl. Environ. Microbiol. 31: Walker, J. D., and R. R. Colwell Long-chain n- alkanes occurring during microbial degradation of petroleum. Can. J. Microbiol. 22: Walker, J. D., and R. R. Colwell Measuring the potential activity of hydrocarbon-degrading bacteria. Appl. Environ. Microbiol. 31: