A Study of the Role of Air-borne Particulates as the Cause of Unexplained Coliform Contamination in Drilled Wells. Marie T. Trest, Jon H. Standridge, Sharon M. Kluender, and Jeremy M. Olstadt University of Wisconsin State Laboratory of Hygiene Water Microbiology Unit Madison, Wisconsin William T. Rock Wisconsin Department of Natural Resources Private Water Supply Section Madison, Wisconsin INTRODUCTION Numerous studies published in recent years have described the problem of detecting total coliform bacteria in water supplies where no evidence of fecal contamination has been found (1-3). In late summer months, the incidence of coliform contamination of drilled wells in Wisconsin, as well as in other states, often rises to over 20% (4). Many researchers have offered biofilm formation as a possible explanation for this phenomenon. Beginning in 1987 LeChevallier published a series of studies demonstrating that bacterial growth in biofilm coated systems is common and has been responsible for positive coliform bacterial test results in water that was total coliform absent before it entered the biofilm inhabited system (1,2, 5-7). In one of these studies, LeChevallier, et al. documented that coliform bacteria introduced into an experimental biofilm coated system will colonize the biofilm causing subsequent positive coliform tests(1). Other researchers have demonstrated that coliforms adhering to biofilms may survive chlorination (8) or show increased bacterial resistance to chlorine disinfection in distribution systems (5,9,10). While a connection between biofilms and total coliform occurrence has been clearly established, the question of where these total coliforms originate from has been unanswered. During a large survey of wells in nine mid-western states, it was noted that point driven wells do not experience the same seasonal rise in total coliform occurrence as do drilled wells (11). One of the greatest differing factors between drilled and point driven wells is that point driven wells are constructed as sealed systems, while drilled wells allow air to enter the well through a screened opening in the cap. When water is pumped from a drilled well, the recharge from the aquifer often occurs at a slower rate than the pumping rate, resulting in a draw down of the water level of the well. As the water level draws down, air rushes into the well via the well cap. This finding suggested the possibility that coliform bacteria may be entering wells via an air-borne route. The goals of this project were to 1; determine if coliform organisms occur in air as viable bio-aerosols near well-heads, 2; determine if the coliforms recovered from well-head bio-aerosols are similar to those commonly found in contaminated
wells, and 3; determine if coliform bio-aerosols experimentally created near a well head are capable of artificially "infecting" a well. MATERIALS AND METHODS Well Sites. Samples were collected from 165 well sites located throughout Wisconsin. 96 of the wells were specifically chosen because they had recently experienced a total coliform positive test from the well water while the remaining 69 sites had coliform negative testing histories. Well site information was documented, with a description of recent weather events, activities near the wellhead, the ground surface near the well, including the existence of vegetation, presence of animals, feces, and enclosures over the well-head, or the presence of a well pit. Air sampling. A Single Stage/N6 Microbial Anderson Sampler was calibrated to 28.3 liters/min (1 ACFM) using a BIOS DryCal DC-1 Primary Air Flow Meter with Flow Cell DC-1HC. The sampler was operated using m-endo agar LES in a 100mm x 15mmpetri dish as the collection media(13). This method of recovering coliforms from the air proved more effective than liquid impinger sampling. The Anderson Sampler technology using selective media has been well documented as the sampling method of choice in studies similar to this one(14,15,16). The classification stage and base plate of the Anderson Sampler were decontaminated with a 70% Isopropyl Alcohol swab before each sample was taken and the alcohol was allowed to evaporate. A 15 minute (424.5 L) sample was collected. The samples were taken next to a well head with the vertical orientation of the inlet cone near the edge of the well cap to sample the air most likely to be pulled into a well during a drawn-down episode. Bacterial culture and identification. The m-endo plates were incubated for 24 hours at 35 C. Up to 6 typical coliform colonies from each m-endo plate were streaked to standard methods agar plates (13) and incubated for 24 hours at 35 C. One isolated colony from each SM plate was transferred with a sterile loop to ONPG reagent (1), and again incubated at 35 C for up to 24 hours to confirm the presence of coliform bacteria. 48 of these confirmed isolates (all from different well sites) were identified to species using API 20 E Test System strips (biomerieux Vitek, Inc., Hazelwood, MO). Well preparation for artificial contamination. A 6.65m deep, 5 cm diameter monitoring well was used for the well contamination experiments. The well was first tested for the presence of coliform bacteria on several occasions using Colilert-18 (IDEXX Laboratories, Inc., Westbrook, ME) reagent in Quanti-Trays (IDEXX Laboratories, Inc.). Bacteria were identified using the API 20 E System strips.
Due to coliform positive test results, the well was disinfected with liquid chlorine bleach and pumped for 3 hours to remove residual chlorine. Five gallons of water collected from a different well known to have significant biofilm problems was added to the experimental well and left for one week to encourage biofilm growth. The well was tested for coliform bacteria as outlined above after one and two weeks in addition to being screened for iron bacteria and heterotrophic plate counts to document the establishment of a biofilm(1). Artificial well contamination using bio-aerosols. Once the well was testing total coliform absent and biofilm-indicators were present, the well was exposed to a bio-aerosol of Enterobacter agglomerans using the following procedure. An apparatus was designed and fabricated utilizing a 5 gal Cubitainer (see Fig. 1). Sealed access ports were inserted into the Cubitainer for a Collision Nebulizer, an Anderson Sampler, and a well pump using plumbing hose connectors, two threaded washers, and silicone glue. Holes were bored through the Cubitainer, one washer was screwed onto the hose connector, and a layer of glue was applied to the washer. The connector was inserted into the Cubitainer and the second washer with a layer of glue was attached from the inside. A layer of silicone sealant was applied on the outer surface after the glue dried. The Cubitainer was secured to the 5 cm PVC pipe of the well by drilling a hole in a PVC pipe cap and silicone gluing and sealing the cap to the Cubitainer opening over a layer of cut plastic tubing. The water sample collection hose was lowered into the well and attached to the inside of the Cubitainer and the apparatus was attached with silicone sealant to the top of the well. Hoses clamps, and connectors were utilized to attach the Collision Nebulizer, the Anderson Sampler and the well pump. Air pressure for the nebulizer was generated by an air pump set at 20 psi, and the pressurized air was filtered through a 0.7µm Millipore type HC membrane Filter (Millipore Corporation, Bedford, MA) before entering the nebulizer. An environmental isolate of E. agglomerans was obtained from a well water sample (API 20E profile 1044153) and transferred only once to a nutrient agar slant. E. agglomerans was chosen because it was the most frequently identified coliform detected in the air samples in addition to occurring regularly among total coliform positive water samples. From the nutrient slant, a suspension of bacteria was created in stock phosphate buffer solution, and a dilution series to 10 6 was prepared (13). A heterotrophic plate count on standard plate count agar (13) was performed to verify the coliform concentration. A suspension of a concentration estimated to be a median 24-hour level of bacterial exposure to a drilled well was prepared based on a ten minute nebulizer run where 4.54mL of water were nebulized. This resulted in approximately 545 CFU nebulized into the 12 L of air in the Cubitainer.
Figure 1 Quality control. All new batches of m-endo plates, ONPG reagent, and Colilert reagent were tested for sterility and proper reactions in positive and negative controls, as was each batch of R2A and standard plate count agar. The Anderson sampler was calibrated on each day of use and all hosing for the well pump was autoclaved before each use. Incubator temperatures were monitored and recorded twice daily. RESULTS and DISCUSSION Of the165 collected air samples, 84 contained viable coliform organisms (51%), a number the researchers thought to be surprisingly high. Of the 96 wells with a previous history of coliform positivity, 59 of the air samples (61.5%) were coliform positive. These results contrast with the 69 wells with no recent coliform detections, where only 25 of the air samples (36%) were total coliform positive. Or, well-head locations where coliforms were detected in the air were 1.7 times as likely to have experienced coliform positivity problems. Conditions surrounding the wellhead location play a role in whether coliform bacteria were detected in the air samples. In samples where the well head was surrounded by vegetation (n=114), 68 contained coliform bacteria (59.6%), a remarkable increase from the average occurrence. Where the wellhead was surrounded by gravel, pavement, or cement (n=21), only 6 of the air samples were positive (28.6%), and in cases where the well head was in a building (n=19), 5 samples were coliform positive (26.3%).
Certain weather conditions also increased the likelihood of detecting coliforms in air samples. When a sample was collected within 3 hours of precipitation (n=10), 80% of the samples contained coliforms. In cases where the well head was surrounded by vegetation and there was a breeze (n=84), 63.1% of the air samples were positive. These data suggest that climatological conditions may increase the likelihood of vegetation-associated coliform aerosolization?) Other well surroundings were also found to influence the likelihood of finding aerosolized coliform bacteria. When there was evidence that a lawn was recently mowed, (N=11), 81.8% of collected air samples contained coliform organisms. At sites where there was evidence of pet ownership, i.e. presence of pet or fecal material(n=38), 22 samples were positive (57.9%). At the locations where fecal material was present in close proximity to the well, 73.3% of the air samples were positive (11 of 15), and interestingly, all 11 of these wells had histories of testing coliform positive. Furthermore, in places where fecal material was present or the well was near a barnyard housing cows or chickens (N=25), 60.0% of air samples were positive and in 14 of the 15 locations with positive air samples, the water had been coliform positive. Coliform identification data. A total of 48 coliform isolates from 48 different well sites were identified. This data is presented in the first two columns of Table 1. For comparison, column three of the table gives the coliform occurrence levels routinely experienced in Wisconsin wells (4). It can be concluded that the coliform population in aerosols collected near well-heads is very similar to the coliform population found in well water samples. Table 1. Comparison of the genera of coliforms identified from bio-aerosols near well-heads to the genera isolated from well waters Coliform genera From air samples From water samples Enterobacter 83% 53% Klebsiella 8% 8% Escherichia 2% 1% Citrobacter 0% 15% Serratia 0% 17% Other 7% 6% Artificial contamination of a well. Water samples collected on the day of the experiment prior to bio-aerosol exposure were coliform negative. A sub-sample of air taken from the Cubitainer contained 5.4CFU/L of the suspended Enterobacter agglomerans coliform. Water samples collected 5 days after the experimental well infection yielded Enterobacter agglomerans from the well. The API Profile, however, was slightly different. While the original bacteria produced a positive indole production test, and a negative sorbitol reaction, the indole test on the organism isolated from the well was negative and the sorbitol test was
positive. The researchers feel these anomalies may be insignificant as Enterobacter organisms recently cultured from the wild are often found to be genetically labile upon passage in artificial media. The researchers intend to repeat this experiment in the fall of 1999. CONCLUSION These results clearly demonstrate that bio-aerosols containing coliform bacteria commonly exist near well-heads. This fact is compounded with the finding that a well was 1.7 times as likely to test total coliform positive when total coliform bacteria are detected in bio-aerosols near the well-head. Additionally, recent precipitation and/or the presence of animals or vegetation in close proximity to the well-head increase the likelihood of the existence of coliforms in these bioaerosols. Finally, the experimental attempt to contaminate a well with a total coliform organism introduced as an aerosol was successful. This strongly implicates air-borne coliforms entering vented wells during pumping as a mechanism of well contamination with total coliform bacteria and provides a possible explanation for previously unexplained occurrences of coliform contamination in well water systems. Further study of engineering interventions to eliminate air from entering wells or to filter air entering wells may be warranted. LITERATURE CITED 1. LeChevallier, M. W., T. M. Babcock, and R. G. Lee. 1987. Examination and characterization of distribution system biofilms. Appl. Environ. Microbiol. 53:2714-2724. 2. LeChevallier, M. W., N. J. Welch, and D. B. Smith. 1996. Full-scale studies of factors related to coliform regrowth in drinking water. Appl. Environ. Microbiol. 62:2201-2211. 3. Camper, A. K., W. L. Jones, and J. T. Hayes. 1996. Effect of growth conditions and substratum composition on the persistence of coliforms in mixed-population biofilms. Appl. Environ. Microbiol. 62:4014-4018. 4. Olstadt, J. M., J. H. Standridge, and S. M. Kluender. 1998. A study of the seasonal occurrence of total coliform bacteria positivity in drinking water. In Proceedings of the Water Quality Technology Conference, American Water Works Association. 5. LeChevallier, M. W., C. D. Cawthon, and R. G. Lee. 1988. Factors promoting survival of bacteria in chlorinated water supplies. Appl. Environ. Microbiol. 54:649-654. 6. LeChevallier, M. W., C. D. Lowry, and R. G. Lee. 1990. Disinfecting biofilms in a model distribution system. Jour. AWWA. 82:87-99. 7. LeChevallier, M. W., W. Schulz, and R. G. Lee. 1991. Bacterial nutrients in drinking water. Appl. Environ. Microbiol. 57:857-862. 8. Wierenga, J. T. 1985. Recovery of coliforms in the presence of a free chlorine residual. J. Am. Water Works Assoc. 77(11):83-88.
9. Ridgway, H. F. and B. H. Olson. 1982. Chlorine resistance patterns of bacteria from two drinking water distribution systems. Appl. Environ. Microbiol. 44:972-987. 10. Herson, D. S., B. McGonigle, M. A. Payer, and K. H. Baker. 1987. Attachment as a factor in the protection of Enterobacter cloacae from chlorination. Appl. Environ. Microbiol. 53:1178-1180. 11. Center for Disease Control and Prevention. 1998. A Survey of the Quality of Water Drawn from Domestic Wells in Nine Midwestern States. Atlanta, GA. 12. Anderson, A. 1958. New Sampler for the collection, sizing and enumeration of viable airborne particles. Journal of Bacteriology. 76:471-484. 13. American Public Health Association. 1992. Standard methods for the examination of water and wastewater, 18 th ed. American Public Health Association, Washington, D.C. 14. Jensen, P. A., W. F. Todd, G. N. Davis, and P. V. Scarpino. 1992. Evaluation of eight bioaerosol samples challenged with aerosols of free bacteria. Am. Ind. Hyg. Assoc. Jour. 53:660-667. 15. Kenline, P. A. and P. V. Sarpino. 1972. Bacterial air pollution from sewage treatment plants. Am. Ind. Hyg. Assoc. 33:346-352. 16. Sawyer, B., G. Elenbogen, K. C. Rao, P. O Brien, D. R. Zenz, and C. Lue- Hing. 1993. Bacterial aerosol emission rates from municipal wastewater aeration tanks. Appl. Environ. Microbiol. 59:3183-3186.