Abstract
To investigate the occurrence of different Klebsiella spp. in aquatic environments, a total of 208 samples of natural surface waters was examined. From half (53%) of these samples, 123 Klebsiella strains were isolated, the most common species being Klebsiella pneumoniae. A comparison of these isolates to a group of 207 clinical K. pneumoniae isolates demonstrated that water isolates of K. pneumoniae, unlike those of K. oxytoca and K. planticola, are as capable as clinical isolates of expressing putative virulence factors such as serum resistance and capsular polysaccharides, pili, and siderophores.
Bacteria of the genus Klebsiella are a frequent cause of nosocomial infections (8). Klebsiella spp. are ubiquitous in nature. Their nonclinical habitats encompass the gastrointestinal tract of mammals as well as environmental sources such as soil, surface waters, and plants (1). Environmental isolates have been described as being indistinguishable from human clinical isolates with respect to their biochemical reactions and virulence (12). While the medical significance of Klebsiella obtained in the natural environment is far from clear, such habitats are thought to be potential reservoirs for the growth and spread of these bacteria which may colonize animals and humans (11).
At present the genus Klebsiella is subdivided into five species. Clinically, the most important species are Klebsiella pneumoniae and K. oxytoca, while K. ornithinolytica, K. terrigena, and K. planticola are rarely isolated from human clinical specimens (6, 18). K. planticola and K. terrigena are considered to be environmental species, as reflected in their species designations. In contrast to K. pneumoniae, neither species grows at elevated temperatures, such as at 44.5°C.
To date, however, studies on the frequency of Klebsiella in nonclinical habitats have focused on K. pneumoniae or have not identified isolates to the species level at all (2, 3, 4, 5, 10, 11, 12, 19). No data are currently available on the incidence of the various Klebsiella spp. in environmental habitats.
The purpose of the present study was to determine the occurrence and distribution of the five Klebsiella species in natural surface waters. Having done this, we investigated whether environmental isolates are capable of producing putative Klebsiella virulence factors, such as pili, serum resistance properties, siderophores, or particular capsular types. For comparison, a group of previously described K. pneumoniae human clinical isolates was used (15).
Collection of water samples.
From November 1997 to June 1998, 208 water samples were collected in sterile 250-ml glass bottles from 196 different sampling sites at streams, lakes, and the Baltic Sea in various geographic areas of Schleswig-Holstein, Germany. Samples were taken 30 cm below the water surface, stored on ice for transportation, and processed for bacteriological analysis within 4 h of collection. The samples were classified as freshwater (conductivity, <1,500 μS/cm), brackish water (1,500 to 15,000 μS/cm), or salt water (>15,000 μS/cm).
Isolation of strains.
Each 250-ml sample was filtered through a 0.45-μm-pore-size membrane (Sartorius, Göttingen, Germany). The membranes were transferred onto Simmons citrate agar with 1% (wt/vol) inositol and incubated for 48 h at 37°C. This medium is highly selective but not inhibitory for the recovery of Klebsiella (21). Presumptive Klebsiella colonies were isolated, followed by identification according to the biochemical tests given by Ørskov (13), which include fermentation of melezitose and l-sorbose, gas production from lactose at 44.5°C, growth at 10°C, pectate degradation, and utilization of m-hydroxybenzoate and hydroxy-l-proline. A group of 207 human clinical K. pneumoniae isolates previously obtained from human infections was used for comparison (15).
Capsule typing.
The isolates were serotyped by the capsular swelling method as described by Ullmann (20). Polyvalent rabbit anticapsule sera were used for screening, and monospecific sera were used for typing.
Hemagglutination assay.
Expression of type 1 pili (mannose-sensitive hemagglutination [MSHA]) and type 3 pili (mannose-resistant Klebsiella-like hemagglutination [MR/K-HA]) was examined as described previously (16), with MSHA being assessed on guinea pig erythrocytes and MR/K-HA on tanned ox red blood cells. Bacteria were grown statically at 48-h intervals. Fifty microliters of bacterial suspensions (approximately 1011 bacteria/ml) and 50 μl of erythrocytes (5 × 108/ml) were mixed in porcelain tiles, rocked, and observed for 3 min at room temperature. Agglutination was finally read after further incubation for 10 min at 4°C.
Determination of siderophore production.
For detection of enterobactin and aerobactin production the cross-feeding bioassay of Hantke (7) was performed as described elsewhere (14). Nutrient agar supplemented with 2,2′-dipyridyl (200 μM) served as iron-restricted agar medium. Escherichia coli strain H1887 (ColV− Aer− Iut+ FepA− Fiu− Cir− aroB) was used as the indicator strain for aerobactin production, and strain H1939 (FepA+ Fiu− Cir− FhuA− FhuB− aroB) was used to indicate enterobactin production. Aerobactin production was counterchecked using the E. coli strain H1886, which is the Iut− parent strain of H1887. Each isolate was tested twice. The indicator strains were kindly provided by K. Hantke, University of Tübingen, Tübingen, Germany.
Serum bactericidal assay.
The susceptibility of bacteria to human serum was determined by the method of Hughes et al. (9) as slightly modified previously (17). Twenty-five microliters of bacterial suspensions (2 × 106 cells/ml) and 75 μl of normal human serum were put into microtiter trays and incubated at 37°C. Viability was determined immediately and after 1, 2, and 3 h of incubation by plating on brain heart infusion agar for colony counts. Responses were graded as highly sensitive, intermediately sensitive, or serum resistant according to the system of Hughes and colleagues (9). Each strain was tested three times.
Statistical analysis.
The significance of differences between groups of bacteria was evaluated by Yates' corrected chi square test for 2 by 2 contingency tables. Medians were compared using the nonparametric analysis of variance test of Kruskal-Wallis and by the Mann-Whitney test. All tests were performed using GraphPad InStat, version 3.00 (GraphPad Software, San Diego, Calif.).
Over an 8-month period, 208 samples of natural surface water were taken from 196 different sampling sites. One hundred ten of the water samples (52.9%) were found to contain 123 Klebsiella strains. Thirteen water samples each contained two different Klebsiella species, and two other samples contained two different strains of the same species. In our experience, the frequency of Klebsiella isolation from surface water depends to a large extent on the volume investigated. In a preliminary experiment using 1-ml samples, we were not able to isolate this bacterium from any of the 47 water samples examined. We therefore decided to investigate 250-ml samples. Even then, the number of Klebsiella colonies per sample was low (usually 1 to 5 CFU/250 ml). The incidence of positive water samples was conspicuously independent of source and season. We found no significant differences in the frequency of isolation between freshwater, salt water, and brackish water (Table 1). Likewise, no seasonal effects could be observed (data not shown). K. pneumoniae was most common (n = 62; 52%), followed by K. oxytoca (n = 34; 27%) and K. planticola (n = 27; 22%). Neither K. ornithinolytica nor K. terrigena was detected. With respect to the latter species, this stands in striking contrast to the view that K. terrigena is an environmental species that can be isolated from surface waters (1). The lack of K. terrigena in our surface water samples, however, was not due to a possible inhibiting effect of the Klebsiella-selective agar used since in preliminary experiments we confirmed that all Klebsiella species can grow on this medium. It is conceivable, though, that there are geographic differences in the occurrence of K. terrigena.
TABLE 1.
Incidence of Klebsiella spp. in samples of different types of surface water
| Organism | No. (%) of organisms in:
|
||
|---|---|---|---|
| Freshwater (n = 133) | Brackish water (n = 12) | Salt water (n = 63) | |
| K. pneumoniae | 42 (31.6) | 3 (25)a | 17 (25.4)a |
| K. oxytoca | 18 (12.8)a | 2 (16.7) | 14 (22.2) |
| K. planticola | 17 (12.8) | 1 (8.3) | 9 (14.3) |
| Total | 77 (51.9)a | 6 (41.7)a | 40 (57.1)a |
Detection of two different species or two different isolates of the same species in a single water sample.
Serotyping revealed K types K33 and K69 as the most common capsular types among K. pneumoniae and K. oxytoca isolates (13 to 15%) (Table 2). Both serotypes were rarely observed (<2%) in the group of 207 clinical K. pneumoniae isolates investigated for comparison purposes. In contrast, clinical K. pneumoniae strains predominantly expressed the K antigen K2 (14%), which is considered to be a main determinant of Klebsiella virulence (18). Surprisingly, this serotype was also the most common K type among K. planticola isolates (11%).
TABLE 2.
Distribution of capsule types in Klebsiella spp. from natural surface waters and from clinical specimens
| K type | No. (%) of isolates with each K type
|
|||
|---|---|---|---|---|
| K. pneumoniae(n = 62) | K. oxytoca(n = 34) | K. planticola(n = 27) | Clinical K. pneumoniaea(n = 207) | |
| K33 | 6 (9.7) | 5 (14.7) | 2 (7.4) | 3 (1.4) |
| K69 | 8 (12.9) | 2 (5.9) | 0 (0) | 1 (0.5) |
| K2 | 0 (0) | 2 (5.9) | 3 (11.1) | 28 (13.5) |
| K70 | 0 (0) | 2 (5.9) | 2 (7.4) | 0 (0) |
| K56 | 1 (1.6) | 0 (0) | 2 (7.4) | 1 (0.5) |
| K13 | 4 (6.5) | 0 (0) | 0 (0) | 2 (1.0) |
| K35 | 4 (6.5) | 2 (5.9) | 1 (3.7) | 3 (1.4) |
| K23 | 1 (1.6) | 2 (5.9) | 0 (0) | 2 (1.0) |
| K26 | 0 (0) | 2 (5.9) | 0 (0) | 1 (0.5) |
| Other K types | <5% Each | <5% Each | <5% Each | <5% Each |
| Untypeable | 6 (9.7) | 1 (2.9) | 0 (0) | 16 (7.7) |
Comparison group of clinical K. pneumoniae isolates. Data are from reference 15.
Clinical K. pneumoniae strains and isolates from surface water were very similar with respect to the incidence of type 1 and type 3 pili. Both groups were significantly more often fimbriated than K. oxytoca or K. planticola strains (P < 0.005) (Table 3). However, the statistical significance of the differences between groups in the production of type 3 pili was only at a P value of 0.07 in some cases. MSHA expression in K. oxytoca was very rare (15%).
TABLE 3.
Distribution of fimbriae, serum resistance properties, and siderophores among Klebsiella spp. from natural surface waters and from clinical specimens
| Virulence factor | No. (%) of positive isolates
|
|||
|---|---|---|---|---|
| K. pneumoniae(n = 62) | K. oxytoca(n = 34) | K. planticola(n = 27) | Clinical K. pneumoniaea(n = 207) | |
| MSHA (type 1 fimbriae) | 54 (87) | 5 (15) | 16 (59) | 177 (86) |
| MR/K-HA (type 3 fimbriae) | 45 (73) | 18 (53) | 8 (30) | 145 (70) |
| Serum resistance | 7 (11) | 18 (53) | 1 (3.7) | 52 (25) |
| Production of: | ||||
| Enterobactin | 60 (97) | 33 (97) | 27 (100) | 205 (99) |
| Aerobactin | 0 (0) | 0 (0) | 0 (0) | 11 (5.3) |
| Mean cumulative no. of virulence factors per isolate | 2.7 | 2.2 | 1.9 | 2.9 |
Comparison group of clinical K. pneumoniae isolates. Data are from reference 15.
The incidences of serum resistance properties differed considerably between the groups of environmental isolates (Table 3). About half (53%) of the K. oxytoca strains proved to be serum resistant, whereas only 11% of the K. pneumoniae and 4% of the K. planticola isolates were resistant (P < 0.0001). Clinical K. pneumoniae strains exhibited serum resistance properties twice as often (25%) as environmental K. pneumoniae strains (P < 0.05).
Siderophore production by the bacterial groups was very similar. Except for five strains, all isolates investigated were able to synthesize the catechol-type siderophore enterobactin (Table 3). In contrast, production of the hydroxamate-type siderophore aerobactin was lacking (environmental isolates) or very rare (clinical K. pneumoniae strains).
To determine whether environmental species differ with respect to the expression of virulence factors and whether they differ in this respect from clinical K. pneumoniae isolates, the factors detected in each strain (MSHA, MR/K-HA, serum resistance, enterobactin synthesis, and aerobactin production) were added up to get the cumulative number of virulence factors expressed per isolate (Table 3). The mean number of factors expressed by environmental K. oxytoca (2.2) and K. planticola (1.9) strains was significantly lower than that expressed by environmental (2.7) or clinical (2.9) K. pneumoniae isolates (P < 0.01).
To sum up, a high percentage (53%) of surface water samples proved to be positive for Klebsiella spp., the most common species being K. pneumoniae. Furthermore, our data show that surface water isolates of K. pneumoniae resemble clinical strains in the expression of virulence factors, whereas water isolates of K. oxytoca and K. planticola differ from clinical strains in this respect. With respect to the factors examined, we found no evidence that environmental K. pneumoniae strains are less virulent than clinical strains. Whether this finding has any relevance to public health is at present unclear and should be evaluated by further studies.
REFERENCES
- 1.Bagley S T. Habitat association of Klebsiella species. Infect Control. 1985;6:52–58. doi: 10.1017/s0195941700062603. [DOI] [PubMed] [Google Scholar]
- 2.Bagley S T, Seidler R J, Talbot H W, Jr, Morrow J E. Isolation of Klebsielleae from within living wood. Appl Environ Microbiol. 1978;36:178–185. doi: 10.1128/aem.36.1.178-185.1978. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Brown C, Seidler R J. Potential pathogens in the environment: Klebsiella pneumoniae, a taxonomic and ecological enigma. Appl Microbiol. 1973;25:900–904. doi: 10.1128/am.25.6.900-904.1973. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Campbell L M, Michaels G, Klein R D, Roth I L. Isolation of Klebsiella pneumoniae from lake water. Can J Microbiol. 1976;22:1762–1767. doi: 10.1139/m76-260. [DOI] [PubMed] [Google Scholar]
- 5.Edberg S C, Piscitelli V, Cartter M. Phenotypic characteristics of coliform and noncoliform bacteria from a public water supply compared with regional and national clinical species. Appl Environ Microbiol. 1986;52:474–478. doi: 10.1128/aem.52.3.474-478.1986. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Farmer J J, III, Davis B R, Hickman-Brenner F W, McWhorter A, Huntley-Carter G P, Asbury M A, Riddle C, Wathen-Grady H G, Elias C, Fanning G R, Steigerwalt A G, O'Hara C M, Morris G K, Smith P B, Brenner D J. Biochemical identification of new species and biogroups of Enterobacteriaceae isolated from clinical specimens. J Clin Microbiol. 1985;21:46–76. doi: 10.1128/jcm.21.1.46-76.1985. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hantke K. Dihydroxybenzoylserine—a siderophore for E. coli. FEMS Microbiol Lett. 1990;55:5–8. doi: 10.1016/0378-1097(90)90158-m. [DOI] [PubMed] [Google Scholar]
- 8.Horan T, Culver D, Jarvis W, Emori G, Banerjee S, Martone W, Thornsberry C. Pathogens causing nosocomial infections. Antimicrob Newsl. 1988;5:65–67. [Google Scholar]
- 9.Hughes C, Phillips R, Roberts A P. Serum resistance among Escherichia coli strains causing urinary tract infection in relation to O type and the carriage of hemolysin, colicin, and antibiotic resistance determinants. Infect Immun. 1982;35:270–275. doi: 10.1128/iai.35.1.270-275.1982. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Knittel M D. Occurrence of Klebsiella pneumoniae in surface waters. Appl Microbiol. 1975;29:595–597. doi: 10.1128/am.29.5.595-597.1975. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Knittel M D, Seidler R J, Eby C, Cabe L M. Colonization of the botanical environment by Klebsiella isolates of pathogenic origin. Appl Environ Microbiol. 1977;34:557–563. doi: 10.1128/aem.34.5.557-563.1977. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Matsen J M, Spindler J A, Blosser R O. Characterization of Klebsiella isolates from natural receiving waters and comparison with human isolates. Appl Microbiol. 1974;28:672–678. doi: 10.1128/am.28.4.672-678.1974. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ørskov I. Genus Klebsiella. In: Krieg N R, Holt J G, editors. Bergey's manual of systematic bacteriology. Baltimore, Md: The Williams & Wilkins Co.; 1984. pp. 461–465. [Google Scholar]
- 14.Podschun R, Fischer A, Ullmann U. Siderophore production of Klebsiella species isolated from different sources. Zentbl Bakteriol. 1992;276:481–486. doi: 10.1016/s0934-8840(11)80673-0. [DOI] [PubMed] [Google Scholar]
- 15.Podschun R, Fischer A, Ullmann U. Expression of putative virulence factors by clinical isolates of Klebsiella planticola. J Med Microbiol. 2000;49:115–119. doi: 10.1099/0022-1317-49-2-115. [DOI] [PubMed] [Google Scholar]
- 16.Podschun R, Sahly H. Hemagglutinins of Klebsiella pneumoniae and K. oxytoca isolated from different sources. Zentbl Hyg. 1991;191:46–52. [PubMed] [Google Scholar]
- 17.Podschun R, Teske E, Ullmann U. Serum resistance properties of Klebsiella pneumoniae and K. oxytoca isolated from different sources. Zentbl Hyg. 1991;192:279–285. [PubMed] [Google Scholar]
- 18.Podschun R, Ullmann U. Klebsiella spp. as nosocomial pathogens: epidemiology, taxonomy, typing methods, and pathogenicity factors. Clin Microbiol Rev. 1998;11:589–603. doi: 10.1128/cmr.11.4.589. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Seidler R J, Knittel M D, Brown C. Potential pathogens in the environment: cultural reactions and nucleic acid studies on Klebsiella pneumoniae from clinical and environmental sources. Appl Microbiol. 1975;29:819–825. doi: 10.1128/am.29.6.819-825.1975. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Ullmann U. The distribution of Klebsiella pneumoniae serotypes from different sources and their sensitivity to cephalosporins. Infection. 1983;11:S28–S31. doi: 10.1007/BF01641102. [DOI] [PubMed] [Google Scholar]
- 21.Van Kregten E, Westerdaal N A C, Willers J M N. New, simple medium for selective recovery of Klebsiella pneumoniae and Klebsiella oxytoca from human feces. J Clin Microbiol. 1984;20:936–941. doi: 10.1128/jcm.20.5.936-941.1984. [DOI] [PMC free article] [PubMed] [Google Scholar]