Journal of Applied Bacteriology 1985, 58, 139-143 I528/2/84

A note on Yersinia enterocolitica in a swine farm watershed

P.J . W A L K E R * & D.J. GRIMES?, River Studies Center and Department of’Biolog.y, Uni- versity of Wisconsin-La Crosse, La Crosse, Wisconsin 54601, U S A

Received 22 February 1984 and accepted 13 June 1984

WALKER, P.J. & GRIMES, D.J. 1985. A note on Yersinia enterocolitica in a swine farm watershed. Journal of Applied Bacteriology 58, 139-143.

Swine faeces from three pig farms in the La Crosse River watershed near La Crosse, Wisconsin, were sampled for Yersinia enterocolitica; 19 presumptive isolates were recovered and biochemically confirmed as Y . enterocolitica. Simultaneously, during a 2.2 cm rainfall, the confluences of runoff water flowing from the swine holding pens and of nearby streams were also sampled; a single isolate was obtained from one holding pen runoff-stream confluence. Biochemical analysis showed that the water isolate was a biotype identical with that of a swine isolate from the adjacent farm. These results demonstrate one possible mechanism for the introduction of Y . enlrrocolilicu into water supplies; faecal material from swine, a suspected natural reservoir of the bacterium, is transported via runoff water to streams.

Swine and dogs have been implicated as natural reservoirs of Yersinia enterocolitica (Gutman et al. 1973; Toma & Deidrick 1975; Wilson et al. 1976; Schiemann & Fleming 1981), but they have not been linked to human infection (Schiemann & Fleming 1981). Drinking water contaminated with the organism has been suspected in several human infections (Lassen 1972; Keet 1974), and aquatic isolates are indeed frequent (Botzler et al. 1976; Harvey et al. 1976; Brenner et al. 1980; Ursing et ul. 1980; Bartley e f al. 1982). In the present study, the objective was to demonstrate that swine infected with Y . enterocolitica could contaminate adjac- ent surface water.

Materials and Methods

S A M P L I N G

Samples of swine faeces were collected twice from three pig farms. Selection of the farms was based on the proximity of drainage streams to

* Present address: 6475 Pulpit Rock Drive, Colorado Springs, Colorado 80907, USA. i Present address: Department of Microbiology,

University of Maryland, College Park, Maryland 20742 USA.

swine holding areas and on the presence of Y. enterocolitica in the herds. Each farm had a stream within 100 m of the main swine holding areas, and each stream was located in the drain- age basin of the La Crosse River, a tributary of the Mississippi River. Two farms, A and B, were located near Barre Mills, Wisconsin, on small creeks known as Bostwick Creek (A) and Tollefson Coulee (B). The third farm (C) was near West Salem, Wisconsin, on Gill Coulee.

The faeces samples were 24 h old or less. A specimen the size of a pea (approx. 3-5 gf was collected from the centre of the faeces, placed in a screw-top vial containing 10 ml of phosphate- buffered glycerol saline (Lennette et al. 1974) and emulsified by shaking. All specimens were incubated at 4°C; the time between sample col- lection and incubation never exceeded 2 h. Each pig farm was sampled twice, once to ascertain the presence of Y . enterocolitica in the herd and a second time in conjunction with sampling adjacent streams.

Water samples were taken from two locations at each of the farms during a 2.2 cm rainfall that occurred within a two hour period in October, 1977. One sample was taken upstream from the area where runoff from the swine holding area entered the stream. The other

P. J. Walker and D. J . Grimes

sample was collected at the furthest point down- stream where, based on the topography of the area, runoff from the holding area was observed to be entering the stream. Samples, totaling 7.5 1, were collected in sterile 3.8 1 poly- propylene milk containers. The container was submerged with the opening facing upstream, taking care not to disturb the bottom sediment. All water samples were processed within 18 h of collection by aseptic filtration through 0.22 pm membrane filters (GSWP 047 SO, Millipore Corp.).

M E D I A A N D C U L T U R E

The filters were placed in 50 ml of cold enrich- ment medium containing (g/l) NaCI, 8.5; manni- tol, 10; K, HPO,, 2.0. The pH was adjusted to 7.3 and the medium was autoclaved at 121°C for 15 min.; cultures were incubated at 40°C (E.H. Christenson & G.P. Jansen, State Labor- atory of Hygiene, Madison, Wisconsin, unpub- lished method for the isolation of Y . enterocolitica from stools). The faecal samples, suspended in phosphate-buffered glycerol saline (Lennette et al. 1974), were held at 4°C for 6 days. A sterile cotton swab was then saturated with the stool specimen and transfered asepti- cally to 10 ml of cold enrichment medium held at 4°C. At the same time, the sample was streaked for isolation on Drigalski-Norwegian (DN) and Deoxycholate-Norwegian (DCLN) agar plates (Christenson & Jansen, unpublished).

The composition of D N (g/l) was: peptone, 10; NaCl, 5; beef extract, 4; agar, 20. This D N base was adjusted to pH 7.5, autoclaved (121"C, 15 min), and then combined with a solution of the following composition: bromthymol blue, (0.2%), 40 ml; lactose, 7 g; Na,S,O,, 1 g; crystal violet (0.1%) 5 ml.

DCLN contained (g/l): sodium deoxycholate, 2.5; trisodium citrate, 10.5; lactose, 5; peptone, 7 ; beef extract, 2.4; Na,S,O3.5H,O, 5; agar, 15; and neutral red, 0.06. This medium was adjusted to pH 7.2 and was boiled for 2 min; it was not autoclaved and was used within 3 days of preparation.

The faeces and water enrichments were cul- tured on DN and DCLN immediately and at 6, 14, and 21 days. The cultures were incubated at room temperature for up to 6 days and exam- ined periodically. Typical colonies (blue on D N

and translucent pale pink on DCLN) were usually subcultured at 48 to 72 h, as prolonged incubation resulted in overgrown plates.

I D E N T I F I C A T I O N

Suspect Y . enterocolitica colonies were trans- fered to Triple Sugar Iron (TSI) agar (Difco) slants and incubated at 37°C for 48 h. Strains producing an acid slant and butt, without gas (or a very small amount of gas), and without hydrogen sulphide were transfered to Tryptic Soy Agar (Difco) plates to check culture purity. Pure cultures were re-evaluated in TSI and then stained by Gram's method and tested for urease, lysine and ornithine decarboxylase, arginine dihydrolase, and motility a t 22 and 37°C. Strains that produced the appropriate reactions were tentatively identified as Y . enterocolitica and subjected to further biochemical tests (Table 1).

Rabbit antisera were prepared to the follow- ing strains of Y . enterocolitica: serotypes 0:2, 3, 0:5, 0:6, 0:6, 30 (lactose-positive), 0:6 , 30 (lactose-negative), 0:8, and 0: 16. These serotypes were provided by G.P. Jansen, State Laboratory of Hygiene, Madison, WI. The method of Winblad et al. (1966) was used and modified in that the second, third and fourth injections were subcutaneous, using 1 ml antigen mixed with 1 ml Freund complete adju- vant (Difco). The tube dilution method was used to determine the titres of the resultant antisera to homologous strains. Slide agglutination was used to determine the serotype of swine and water isolates.

Results and Discussion

A total of 34 strains of Y . enterocolitica was isolated during the two sampling periods. Of these 34 strains, 33 were obtained from swine faeces and one strain was from runoff water. The number of confirmed strains isolated at the first sampling was distributed as follows: Farm A, six strains from 51 samples; Farm B, six strains from 21 samples; Farm C, two strains from 15 samples. Further consideration of these 14 strains is not given in this paper. At the second sampling, which occurred during the 2.2 cm rainfall, 25 samples were taken from each farm, yielding nine, three and seven con- firmed isolates, respectively. The biochemical

Yersinia enterocolitica in swine 141 Table 1. Biochemical characteristics for one water and 19 faecal strains of Yersinia enterocolitica isolated during

a 2.2 cm rainfall, October 1977

Farm A Farm B Farm C

Group 2 Group 1 Group 2 Group 3 Group 1 (1 faecal Group 3 Group 1 Group 2 ( 2 faecal (1 faecal (6 faecal (1 faecal & 1 water (1 faecal (4 faecal (3 faecal

Trait isolates) isolate) isolates) isolate) isolate) isolate) isolates) isolates) - - Indole + + + + +

Aesculin + + + + + + Salicin + + + + + + ONPG -

Cellobiose + + + + + + V-Pat 22°C - - Rhamnose -

-

- -

- -

- - - + + + + - -

+ + + + + + - + - + - - -&

All strains positive for: motility at 22"C, arabinose, glucose, sucrose, xylose, mannose, maltose, urease, orni- thine decarboxylase, catalase, methyl red, mannitol and nitrate reduction. All negative for: motility at 37T, hydrogen sulphide, arginine dihydrolase, lysine decarboxylase, citrate (Simmons), lactose (in phenol red broth), phenylalanine deaminase, malonate, oxidase, Voges-Proskauer (V-P) at 37°C and lactose (O/F).

characteristics of these 19 isolates are given in Table 1.

Water samples were collected immediately after the 25 faecal samples were obtained at each farm. The control water samples, taken above the runoff areas, did not yield Y . entero- colitica. Samples collected below the area of surface runoff from the holding pens were also negative for Y . enterocolitica, with the exception of one sample, collected from Tollefson Coulee, downstream of Farm B. The fact that the control sample collected above Farm B was negative strongly suggests that the strain down- stream came from swine in the Farm B holding pen. The biochemical characteristics of the water isolate (BW-1) were identical with those of an isolate made from pig faeces (BS-2) in the holding pen of Farm B (Table 1).

The majority of isolates (23 of 34) were reco- vered only after 21 days incubation in cold enrichment medium at 4"C, in agreement with the results of Toma & Deidrick (1975) and Weissfeld & Sonnenwirth (1980). Plating the cold enrichment medium at the end of 6 and 14 days yielded only 1 and 8 isolates of Y . entero- colitica respectively. The water sample was posi- tive at the end of a 21 day period of cold enrichment. Faecal samples were considered negative if no strains were isolated after 21 days of enrichment, since additional time does not produce any appreciable increase in the number of isolates recovered (Nilehn 1973; Eiss 1975; Toma & Deidrick 1975).

Biochemical tests (Table 1) were performed in

duplicate at 22°C and 37°C. Forty percent of the strains gave negative indole reactions; 30Y0 fermented rhamnose, 85% gave a positive Voges-Proskauer test at 22"C, and 50% were positive for 8-galactosidase (ONPG) (Table 1). Strains of Y . enterocolitica isolated from the natural environment have tended to be indole variable and biochemically atypical (Lassen 1972; Toma 1973; Harvey et a/. 1976; Kap- perud 1977; Saari & Jansen 1979; Brenner et a/. 1980; Ursing et al. 1980). In addition, variation in growth characteristics at different tem- peratures is well known (Bercovier et a/. 1980).

Biotyping of the isolates revealed two major biotypes. Strains from Farms A and B were Nilehn (1969) and Wauters (1973) biotype 1 with the atypical reactions of lactose negative in O/F medium and ONPG negative. One strain, AS-3, was indole negative; the remaining iso- lates from Farm A and B were indole positive. These strains belong to Group 3, atypical Y . enterocolitica in the classification schema pro- posed by Knapp & Thal (1973). They also appear to belong to biotype 1 or 2, as defined by Bercovier et ai. (1980), and biovar 1 or 3, as described by Kaneko & Hashimoto (1982). Certain key tests (lipase, melibiose, trehalose, rafhose, cc-methyl-D-glucoside, lecithinase), however, that would have precisely placed our strains in these recent classifications were not performed. Strains from Farm C were biotype 4 in the Nilehn (1969) and Wauters (1973) bio- types, and Group 1, typical Y . enterocolitica, as defined by Knapp & Thal (1973). The water

142 P. J . Walker and D. J . Grimes

strain, BW-1, was identical biochemically, for all tests performed, to strain BS-2, obtained from a Farm B faecal specimen on the same sampling day. These strains were both rhamnose positive and hence could be referred to as ‘Y . enterocolitica-like’ (Knapp & Thal 1973). Ursing rf a/. (1980) and Brenner et al. (1980) proposed the names Y . frederiksenii and Y . intermedia to describe rhamnose-positive strains that differ in melibiose, raffinose and z-methyl-D-glucoside reactions. Recent numerical taxonomic studies by Kaneko & Hashimoto (1982) and by Kap- perud et al. (1981), however, d o not agree with this classification, and prefer to retain rhamnose-positive strains in Y . enterocolitica.

The strains of Y . enterocolitica could not be serotyped using slide agglutination with the antisera prepared to serotypes 0: 2,3, 0:5, 0:6,30 lactose-positive, 0:6,30 lactose-negative, 0:8, and 0:16. The titre of the prepared antisera varied between 1/640 and 1/1280 to its homolo- gous strain, using the tube dilution method. It cannot be assumed that the strains were untypable. Definitive conclusions regarding serotype could have been made only if isolates had been typed using antisera to all known serotypes of Y . enterocolitica (Bercovier et al. 1980).

Because of the method of sampling, no con- clusions concerning the proportion of swine har- boring the organism can be drawn from these results. Random specimens were collected from large holding pens containing 30 to 50 head of swine. Therefore, it could not be ascertained if the faeces sampled were from the same pigs or from pigs not previously sampled. It can be ten- tatively assumed, however, that Y . enterocolitica was enzootic in area swine populations, because all three farms yielded isolates of the organism, a premise that has been advanced by other studies (Esseveld & Goudzwaard 1973; Winblad 1973 ; Schiemann & Fleming 198 1).

This study reinforces the need for additional research on the natural reservoir, or reservoirs, and transmission of Y . enterocolitica, particular- ly serotypes 0:3, 0:5,27 0:8 and 0:9, the virulent serotypes most frequently associated with human disease. The isolates reported in this study were of unknown virulence, and, unfor- tunately, are no longer available for testing. This should not detract from the epidemiological implication of this study, however, as potentially pathogenic bacteria were observed to move

from a reservoir into a waterway. The water could then serve as a vehicle for infecting humans, as well as other organisms that could act as incidental vectors.

The authors are grateful to G.P. Jansen for pro- viding the reference strains and to the La Crosse Veterinary Clinic for assistance in locating the study sites. We also thank the three farms for their cooperation.

References BARTLEY, T.D., QUAN, T.J., COLLINS, M.T. & MORRI-

SON, S.M. 1982 Membrane filter technique for the isolation of Yersinia enterocolitica. Applied and Environmental Microbiology 43,829-834.

BERCOVIER, H., BRENNER, D.J., URSING, J., STEIGEK- WALT, A.G., FANNING, G.R., ALONSO, J.M., CARTER, G.P. & MOLLARET, H.H. 1980 Characterization of Yersinia enterocolitica sensu stricto, Current Micro- biology 4, 201-206.

BOTZLER, R.G., WETZLER, F.T., COWAN, A.B. & QUAN, T.J. 1976 Yersinieae in pond water and snails. Journal of Wildlife Disease 12, 492496.

BRENNER, D.J., BERCOVIER, H., URSING, J., ALONSO, J.M., STEIGERWALT, A.G., FANNING, G.R., CARTER, G.P. & MOLLARET, H.H. 1980 Yersinia intermedia: A new species of Enterobacteriaceae composed of rhamnose-positive, melibiose-positive, rafinose- positive strains (formerly called Yersinia entero- colitica or Yersinia enterocolitica-like). Current Microbiology 4,2077212.

EM, J. 1975 Selective culturing of Yersinia entero- colitica at a low temperature. Scandanauian Journal of Infectious Diseases 7,249-251.

ESSEVELD, H., & GOUDZWAARD, C. 1973 On the epide- miology of Yersinia enterocolitica infection: pigs as the source of infection in man. Contributions to Microbiology and Immunology 2,99-101.

GUTMAN, L.T., OTTESEN, E.A., QUAN, T.J., NOCE, P.S. & KATZ, S.L. 1973 An inter-familial outbreak of Yersinia enterocolitica enteritis. New England Journal of Medicine 288, 1372-1377.

HARVEY, S., GREENWOOD, J.R., PICKETT, M.J. & MAH, R.A. 1976 Recovery of Yersinia enterocolitica from streams and lakes of California. Applied and Environmental Microbiology 32, 352-354.

KANEKO, K.-I. & HASHIMOTO, N. 1982 Five biovars of Yersinia enterocolitica delineated by numerical tax- onomy. International Journal of Systematic Bacte- riology 32, 275-287.

KAPPERUD, G. 1977 Yersinia enterocolitica and Yersinia-like microbes isolated from mammals and water in Norway and Denmark. Acta pathologica microbiologica scandinavica Sect. B. 85, 129-1 35.

KAPPERUD, G., BERGAN, T. & LASSEN, J. 1981 Numeri- cal taxonomy of Yersinia enrerocolitica and Yersin- ia enterocolitica-like bacteria. International Journal of’Systematic Bacteriology 31,401-419.

KEET, E.E. 1974 Yersinia enterocolitica septicemia. New York State Journal of Medicine 74, 22262230.

Yersinia enterocolitica in swine 143 KNAPP, W. & THAL, E. 1973 Differentiation of Yersin-

ia enterocolirica by biochemical reactions. Contribu- tions to Microbiology and Immunology 2, 1&16.

LASSEN, J. 1972 Yersiniu enterocolitica in drinking water. Scandinavian Journal of Injectious Diseases 4,

LENNETTE, E.H., SPAULDING, E.H. & TRUANT, J.P. (ed.). 1974 Manual of Clinical Microbiology, 2nd ed. Washington, D.C.: American Society for Micro- biology.

NILEHN, B. 1973 The relationship of incubation iem- perature to serum bactericidal effect, pathogenicity and in vivo survival of Yersinia enterocolitica. Con- tributions to Microbiology and Immunology 2, 85-92.

NILEHN, B. 1969 Studies on Yersiniu enterocolitica with special reference to bacterial diagnosis and occurrence in human acute enteric disease. Acta pathologica rnicrobioloyica scandinauica Suppl. 206, 1-48,

SAARI, T.N. & JANSEN, G.P. 1979 Waterborne Yersin- ia enterocolitica in the Midwestern United States. Contributions to Microbiology and Immunology 5,

SCHIEMANN, D.A. & FLEMING, C.A. 1981 Yersinia enterocolitica isolated from throats of swine in eastern and western Canada. Canadian Journal of' Microbiology 27, 13261333.

TOMA, S. 1973 Survey on the incidence of Yersinia enterocolitica in the Province of Ontario. Canadian Journal of Public Health 64,477487.

125-127.

183-196.

TOMA, s. & DEIDRICK, V.R. 1975 Isolation of Yersinia enrerocolitica from swine. Journal of Clinicui Micro- biology 2, 478481.

URSING, J., BRENNER, D.J., BERCOVIBR, H., FANNING, G.R., STEIGERWALT, A.G., BRAULT, J. & MOLLARET, H.H. 1980 Yersinia frederiksenii: A new species of Enterobacteriaceae composed of rhamnose-positive strains (formerly called atypical Yersiniu entero- colitica or Yersinia enterocolitica-like). Current Microbiology 4,213-217.

WAUTERS, G. 1973 Correlation between ecology, bio- chemical behavior and antigenic properties of Yer- sinia enteracoliticu. Contributions to Microbiology and Immunology 2,3841.

WBISSFELD, A.S. & SONNENWIRTH, A.C. 1980 Ycrsinio enterocolitica in adults with gastrointestinal dis- turbances: need for cold enrichment. Journal of' CIinical Microbiology 11, 196197.

WILSON, H.D., MCCORMICK, J.B. & FEELEY, J.C. 1976 Yzrsinia enterocolitica infection in a 4 month old infant associated with infection in household dogs. Journal of Pediatrics 89,767-769.

WINBLAD, S. 1973 The clinical panorama of human Yzrsinia enterocolitica. Contributions to Micro- biology and Immunology 2, 129-132.

WINBLAD, S., NILEHN, B. & STERNBY, N.H. 1966 Yer- sinia enterocolitica (Pasteurella X ) in human enteric infections. British Medical Journal ii, 1363-1 366.