APPLIED AND ENVIRONMENTAL MICROBIOLOGY, OCt. 1984, p. 830-832 Vol. 48, No. 4 0099-2240/84/100830-03$02.00/0 Copyright © 1984, American Society for Microbiology Isolation of Legionella Species from Drinking Water S. C. HSU,* R. MARTIN, AND B. B. WENTWORTH Laboratory and Epidemiological Services Bureau, Lansing, Michigan 48909 Received 7 March 1984/Accepted 25 July 1984 Three different species of Legionella were recovered from samples of water taken from chlorinated public water supplies where no coliform bacteria were simultaneously detected. Five of 856 samples yielded Legionella isolates. Three isolates were identified as Legionella pneumophila serogroup 1, the fourth was identified as Legionella dumoffli, and the fifth was identified as Legionella jordanis. Studies to determine the survival of L. pneumophila Flint 1 serogroup 1 in tap water at various temperatures and in tap water with added sodium hypochlorite were done. These organisms were found to survive for 299 days in tap water at 24 and 5°C but not at 35°C. A concentration of at least 0.2 mg of residual chlorine per ml was required to eliminate at least 90% of L. pneumophila and Escherichia coli inocula in 2 h. Legionella species have been recognized as environmental gionella pneumophila serogroups 1 through 6, Legionella inhabitants, having been found not only in waters from bozemanii, Legionella micdadei, Legionella dumoffli, Le- cooling towers (5, 7, 8) and in plumbing systems of hospitals gionella gormanii, Legionellajordanis, and Legionella long- and hotels (3, 11) but also in natural water such as lakes, beachae serogroups 1 and 2. ponds, and streams (6, 10). If these organisms are wide- Survival testing. Samples of water from a laboratory tap spread in nature, they may be expected to occur in drinking were used to represent the municipal system serving the water. This study was done to determine whether Legionella laboratory. Portions of this tap water and distilled water species could be isolated from chlorinated drinking water from a central distiller were collected in screw-capped sources and whether Legionella species remained viable bottles. Portions that were autoclaved at 121°C for 20 min over time in tap water, distilled water, and water with added were designated as "sterilized." Legionella organisms to be sodium hypochlorite. tested were removed from a 7-day-old BCYE agar plate culture, washed twice with 10 ml of sterile distilled water by MATERIALS AND METHODS membrane filtration, and suspended in 10 ml of sterile Survey of drinking water supplies. Water samples for distilled water. Samples of sterilized water, unsterilized tap isolation of Legionella species were selected from among water, and distilled water were inoculated with ca. 2.0 x 105 those from chlorinated public water supplies submitted to CFU of L. pneumophila Flint 1 serogroup 1, and 0.1 ml of the laboratory for routine coliform testing. The only criteria each inoculated sample was spread on a BCYE agar plate to used for selection were that (i) the water sample came from a establish the CFU at zero time. Inoculated water samples public water supply listed as chlorinated, and (ii) the sample were incubated at three different temperatures, 5, 24, and was sent for a routine bacteriological analysis during the 35°C, with three replicates at each temperature. At various time period of the study. Each sample (100 ml) was divided times, samples (0.1 to 1.0 ml) of each type of water were equally into two portions; one 50-ml sample was tested for plated on BCYE agar plates. These plates were incubated at coliform contamination, and the other was tested for isola- 35°C in a 2.5% CO2 atmosphere for 10 days. Legionella tion of Legionella species. The portion for coliform analysis colonies were counted to estimate survival over time in the was tested by the standard membrane filter technique (1), three types of water. being filtered through a membrane filter (pore size, 0.45 ,um) Survival in water with added sodium hypochlorite. Three which was transferred to M-Endo agar LES (Difco Labora- series of dilutions of sodium hypochlorite (Miesel Bleach; tories, Detroit, Mich.) and incubated at 35°C. The second active ingredient, 5.25% sodium hypochlorite; 94.75% inert portion was filtered in the same way, and the filter was ingredient; Miesel Co., Detroit, Mich.) were made in steril- transferred to glycine-vancomycin-polymyxin B agar (12) ized tap water for examination at different times (see Table which was then incubated at 35°C in a 2.5% CO2 atmosphere 2). Both test organisms were washed and prepared in the for 6 days. same manner for inoculation. Approximately 106 CFU each Suspect colonies were picked and transferred to buffered of L. pneumophila Flint 1 serogroup 1 and Escherichia coli B charcoal-yeast extract (BCYE) agar (9) and blood agar (ATCC E11303) were inoculated into each dilution. Inoculat- (Trypticase soy agar base [BBL Microbiology Systems, ed samples were incubated at 24°C and examined at 2, 24, Cockeysville, Md.], 40 g; sterile defibrinated sheep blood, 50 and 96 h. The test waters were filtered, and the filters were ml; and distilled water, 1,000 ml). Colonies that grew on cultured as described for the survival test of Legionella. For BCYE agar but not on blood agar were then screened with recovery of E. coli, the filters were transferred to tryptone- polyvalent Legionella antisera by direct immunofluores- glucose-yeast extract agar and incubated at 35°C for up to 10 cence staining (2). Reactive isolates were further examined days. Colony counts were done to estimate the percent for Gram stain reaction, carbohydrate utilization, nitrate survival. After 2 and 96 h of incubation, the chlorine reduction, fatty acid profiles, and production of catalase, residuals of the filtrates were determined by the ortholidine gelatinase (film method), oxidase, and urease (4). Finally, method (1). isolates were classified by species and typed by direct immunofluorescence with anti-Legionella conjugates for Le- RESULTS Five of 856 water samples from public drinking water * Corresponding author. supplies yielded Legionella isolates. Coliform bacteria were 830 VOL. 48, 1984 ISOLATION OF LEGIONELLA SPECIES FROM DRINKING WATER 831 TABLE 1. Survival of L. pneumophila Flint 1 serogroup 1 in The ability of L. pneumophila to survive in various different types of water at different temperatures concentrations of sodium hypochlorite was compared with Survival" that of E. coli (Table 2). Neither organism could be recov- ered after 2 h in concentrations equal to or greater than 2.0 Days Nonsterilized tap Sterilized tap water Sterilized distilled water water mg of residual chlorine per liter in tested water. About 10% of the original inocula of both organisms survived in 0.2 mg 5°C 24°C 35°C 5°C 24°C 25°C 5°C 24°C 35°C of residual chlorine per liter for 2 h. There was little or no 4 100 100 95 100 100 100 90 80 40 effect on the population of either microorganism at concen- 7 100 100 100 100 100 100 95 80 10 trations below 0.1 mg of residual chlorine per liter. At a 14 100 100 100 100 100 100 95 80 1 residual chlorine concentration of 0.1 mg/liter the popula- 100 100 100 95 60 50 1 21 100 95 tions of both bacteria were reduced after day 1, but both 42 90 45 30 100 100 20 15 20 0 56 80 40 5 100 100 10 10 8 0 populations recovered after 96 h. There appeared to be little 70 50 20 0 100 90 0 0.5 0 0 difference between the survival of L. pneumophila and that 91 40 3 0 90 70 0 0.5 0 0 of E. coli in tap water with these various concentrations of 144 40 0.4 0 95 40 0 1 0 0 sodium hypochlorite. The same experiment has been repeat- 186 25 0.3 0 90 30 0 0.5 0 0 ed with similar results. 228 20 0.4 0 80 20 0 0.5 0 0 299 20 0.3 0 80 10 0 0.2 0 0 DISCUSSION a Percentage of population at zero time. The isolation of Legionella species from what are estab- lished chlorinated drinking water supplies and the survival of detected in 41 of the samples. None of the coliform-positive such organisms in tap water with and without additional samples yielded isolates of Legionella; all Legionella iso- sodium hypochlorite suggest that such waters could be a lates were obtained from coliform-negative waters. There source of Legionella infection. Isolation of Legionella spe- was, therefore, no correlation between coliform contamina- cies has been made primarily, if not exclusively, from the hot tion and the presence of Legionella isolates in these waters. water supplies of hospitals, where it is assumed that multipli- Three of the Legionella isolates were identified as L. pneu- cation occurs. The waters tested here would be similar to the mophila serogroup 1, one was identified as L. dumoffli, and source water for such supplies, and isolation from these one was identified as L. jordanis. waters suggests that public water supplies contain Legion- Viable L. pneumophila could be recovered from both tap ella species at levels detectable by the isolation technique and distilled water held at either 24 or 5°C for up to 299 days used here. Legionella species in water supplies appear to be (Table 1). Survival was best at 5°C in sterilized tap water. of environmental origin and not associated with fecal con- There was no indication that any bacterial growth occurred tamination detected by coliform testing, so that the tradition- at any temperature in these samples during the period of al methods for determining "unsafe contaminated" water incubation.