Vol. 32, No. 9JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 1994, p. 2128-21330095-1137/94/$04.00+0Copyright © 1994, American Society for Microbiology

Differentiation of Salmonella Serovar Infantis Isolates fromHuman and Animal Sources by Fingerprinting

IS200 and 16S rm LociSINIKKA PELKONEN,l.2* EEVA-LIISA ROMPPANEN,"3 ANJA SIITONEN,4 AND JUKKA PELKONEN2'3

Regional Laboratory of Kuopio, National Veterinary and Food Research Institute, FIN-70701 Kuopio,j Department ofClinical Microbiology, University of Kuopio, FIN-70211 Kuopio, and Laboratory of Enteric Pathogens,

National Public Health Institute, FIN-00300 Helsinki,4 Finland, and Max-Planck-Institutefor Immunobiology, D-79011 Freiburg, Germany2

Received 27 January 1994/Returned for modification 29 April 1994/Accepted 2 June 1994

We genotyped SalmoneUla serovar infantis (referred to as S. infantis), which is the most widespread serovar

among animals and the third most common cause of human salmonellosis in Finland. Molecular fingerprint-ing of the 16S rrn locus and the Salmonella-specific insertion sequence IS200 was used to type the 131 isolatesoriginating from the main sources of S. infantis infection. The number of IS200 elements in S. infantis variedfrom zero to seven; three or more copies were present in 97% of the isolates, and 71% had four copies. Therewere four conserved chromosomal positions of IS200, which allowed us to group the isolates into three majorclonal groups. We defined 11 unique IS200 profiles and five different ribotypes which, in combination,generated 15 genotypes highly restricted to the infection sources: 8 genotypes were typical of isolates frombroiler chickens and cattle and seven genotypes were typical of isolates from humans. The eight genotypes ofisolates from chickens represented two clonal groups which were differentially associated with chicken-producing companies. The typing scheme allows efficient discrimination between isolates from variousinfection sources and within sources and, therefore, provides a unique molecular tool for use in the study ofthe epidemiology of S. infantis infection.

Salmonella infections are regarded as classical examples ofzoonotic infections which spread from animals to humans.Human infections are most often acquired from contaminatedfood. The source of infection, however, is difficult to traceunless efficient typing methods exist. While the most importantserovars of Salmonella enterica subsp. enterica, like typhi,paratyphi B, typhimurium, and enteritidis, can be subgroupedinto different phage types (29), this is not the case for manyother serovars that cause gastroenteritis, like Salmonella sero-var infantis (referred to here as S. infantis). This is the mostwidespread Salmonella serovar among animals in Finland andthe third most important etiological agent of salmonellosis inhumans in Finland.

S. infantis has been the most common Salmonella serovar inpoultry in Finland since the first outbreak occurred in 1971 (16,20). The infection spread to the farms of all five broilerchicken-producing facilities. More recently, S. infantis has alsospread to cattle farms and is at the moment the most commonserovar that causes bovine salmonellosis (16). In the humanpopulation, S. infantis infection has been the third mostcommon type of salmonellosis after those caused by theserovars enteritidis and typhimurium since the infection hasprevailed in broiler chicken production facilities (15). Atpresent, however, roughly 80% of the human infections areassociated with a recent history of travel to a foreign country,while the remaining 20% are classified as true domesticallyacquired infections.There are only a few reports on the typing of S. infantis.

Phage typing schemes have been introduced by two groups (12,

* Corresponding author. Mailing address: Regional Laboratory ofKuopio, National Veterinary and Food Research Institute, P.O. Box92, FIN-70701 Kuopio, Finland. Phone: 358-71-201450. Fax: 358-71-201459.

14), but until the method is well established the results aredifficult to interpret (unpublished data). Plasmid profiling isnot a potential subtyping method, because only 12 to 18% ofthe isolates have plasmids (10) (unpublished data). Resistanceto antimicrobial drugs is also rare. For these reasons, wesought a means of genotyping S. infantis. The finding ofgenotypes typical of broiler chicken production facilities butdifferent from the genotypes prevalent abroad would allow usto trace sources of human infections and to study the epide-miology of S. infantis infection in the important animal reser-voirs, poultry and cattle.The Salmonella-specific insertion sequence IS200 can be

used for molecular fingerprinting, because the copy numberand chromosomal location of this mobile element may vary(13). Restriction fragment length polymorphism (RFLP) ofIS200 and its adjacent sequences has been studied in Salmo-nella serovars berta, bovismorbificans, dublin, enteritidis, hei-delberg, java, paratyphi B, and typhimurium (2, 4, 6, 7, 13, 23,26, 27). Ribotyping by RFLP analysis at the 16S ribosomal rrnlocus and determination of its location in the chromosome (8,28) are already well-established methods in molecular epide-miology and have been applied alone or in combination withother approaches for the typing of Salmonella species (1, 4-7,19, 23). This report describes a typing scheme for S. infantisthat is based on a combination of fingerprinting of IS200 and16S mnn loci. By using only one restriction enzyme, BanI, for thedigestion of chromosomal DNA, the scheme allowed us todivide S. infantis isolates into 15 genotypes and to differentiatebetween isolates of animal and human origin.

MATERIALS AND METHODS

Bacterial isolates. Salmonella enterica subsp. enterica serovarinfantis (S. infantis) isolates of animal origin were obtained

2128

on Novem

ber 6, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

GENOTYPING OF S. INFANTIS WITHIS200 AND 16S mn 2129

from the National Veterinary and Food Research Institute,Helsinki, Finland, and its Regional Laboratory in Seinajoki,and the strains of human origin were obtained from theNational Public Health Institute, Helsinki, Finland. The strainshad been isolated by or referred to these national salmonellareference centers, which had serologically confirmed the sero-var. The S. infantis isolates (n = 131) originated from thefollowing sources: various broiler chicken farms and broilerchickens (n = 59) of five broiler chicken-producing companiesduring the years 1986 to 1991, various cattle farms (n = 11)from one enzootic area during the years 1988 to 1991, ninedomestic human epidemics (n = 14) in the 1980s, and humanswith a recent history of travel to a foreign country (n = 47) in1989. The strains from broiler chickens were isolated by thenational salmonella control program for the control of salmo-nellosis in poultry. As control strains we used our own Salmo-nella isolates of serovars typhimurium, enteritidis, and dublin,Yersinia pseudotuberculosis, Pasteurella multocida, and Esche-richia coli. Salmonella stock cultures were maintained on eggagar slopes at 20°C. The purity of the cultures was checked onlactose bromthymol blue agar plates, from which one colonywas further cultured for the preparation of bacterial DNA.

Isolation of bacterial DNA. Bacterial DNA was isolated asdescribed previously (28), with minor modifications, from 10ml of a Luria broth culture. The DNA samples were treatedwith RNase and proteinase K; this was followed by phenolextraction. Standard molecular biological methods (21) wereused throughout the study when not otherwise stated.

Preparation ofDNA probes. The 16S rRNA gene probe wasgenerated by PCR amplification of a 1.3-kb DNA fragment(bases 1539 to 2877) of theE. coli 16S rnB operon (3) with theprimer pair 5'-TTCGAGCTCAGATTGAACGCTG-3' and5'-ATTGGATCCACGATTACTAGCG-3' by using E. coliDNA as a template. The PCR was performed by using theGeneAmp PCR core reagents kit (Perkin-Elmer Cetus, Nor-walk, Conn.) and by following the manufacturer's recommen-dations. A 25-cycle program (1 min at 94°C, 1 min at 56°C, and2 min at 72°C) was used. The 1.3-kb DNA PCR product wasisolated by low-melting-point agarose electrophoresis and wascloned into the BamHI and Sacl polylinker sites of the vectorpGEM 3Z f(-) (Promega, Madison, Wis.) after digestion withthe appropriate enzymes. DNA isolated from transformed E.coli NM522 bacteria was digested with BamHI and Sacl, andthe 1.3-kb 16S rm gene probe was isolated by using a DEAEmembrane (Schleicher & Schuell, Dassel, Germany). TheIS200 probe was generated by PCR with the primers describedrecently (2). As a template we used PstI-fragmented DNAfrom our own Salmonella serovar typhimurium isolate HE1 19/89. A 30-cycle program (1 min at 94°C, 0.5 min at 59°C, and 2min at 72°C) was used. The 692-bp PCR product was purifiedwith a QlAquick-spin PCR purification kit (Qiagen, Chats-worth, Calif.) and was cut with the HaeIII enzyme. Thisenzyme cleaves a 592-bp fragment of IS200 (2), which wasisolated with a DEAE membrane, and was used as a probe forIS200. The probes were labelled with [a-32P]dATP by usingrandom prime labelling to a specific activity of 108 cpm/jig andwere purified of free radionucleotides with a NICK column(Pharmacia, Piscataway, N.J.).

Southern hybridization. Bacterial DNA samples (0.5 ,ug)were digested with BanI or with EcoRI and PstI restrictionenzymes and were electrophoresed (1 V/cm) through a 0.8%agarose gel in lx TBE buffer (21). Denatured DNA wastransferred to a Biodyne A nylon membrane (Pall Bio Support,East Hills, N.Y.) in 20X SSC (lx SSC is 0.15 M NaCl plus0.015 M sodium citrate) and was fixed to the membrane with aUV crosslinker (Stratagene, La Jolla, Calif.). The membranes

2 1 43 1 5

W -as 414W

a'

- 6.0

- 4.8

-2.2

FIG. 1. Genomic Southern blot of selected S. infantis ribotypestrains. Genomic DNAs were digested to completion with restrictionendonuclease BanI, Southern blotted, and hybridized with the 1.3-kb16S rm probe which was labelled with 32p. Lanes 1 to 5, ribotypes 1 to5, respectively, obtained with Banl-fragmented DNA. The numbers onthe right of the panel represent the sizes (in kilobases) of the respectivehybridizing fragments.

were prehybridized (3X SSC, 1% sodium dodecyl sulfate[SDS], 5 mM EDTA, 1OX Denhardt's solution, 100 jig ofdenatured salmon sperm DNA per ml) for 1 h at 650C andwere then hybridized overnight at 65°C in a hybridizationmixture (prehybridization mixture plus 10% dextran suilfate[Pharmacia] and the denatured DNA probe [105 cprn/mll)After hybridization the membranes were washed twice in 3 xSSC-1% SDS for 30 min at 650C and once in 0.3x SSC--I %SDS for 15 min at 65°C. The hybridization probe was removedby washing the membranes in 0.4 N NaOH at 42°C for 30 minand then in 0.1x SSC-0.1% SDS-0.2 M Tris-HCl (pH 7.5) at42°C for 30 min.

Designation of genotypes. The genotypes were designated.for example, 1A, where 1 refers to the 16S rn profile (ribotype)and A refers to the IS200 profile of BanI-digested DNA.

RESULTS

Ribotypes obtained with 16S rrn probe. By using the 16S rnprobe we observed five different profiles with the BanI-digested DNA samples (Fig. 1). The use of EcoRI-PstI diges-tion provided one ribotype in addition to the five ribotypesdefined with BanI-digested samples alone, because ribotype 3could be divided further (data not shown). Combination of theresults from IS200 profiling and ribotyping revealed that thesixth ribotype, obtained for only one isolate, was not ofepidemiological interest; thus, the results obtained with BanIfragments alone accounted for the genotypic grouping. Ofribotypes 1 to 5, only ribotype 2 (six isolates) was specific toone infection source, broiler chickens, while ribotypes 4 and 5were represented by only one strain each (Table 1). The majorribotypes, ribotypes 1 (94 isolates) and 3 (29 isolates), wereequally common among human and animal sources. Thus,ribotyping alone was not efficient in differentiating betweenvarious infection sources. However, ribotyping considerablyimproved the efficacy of genotyping by dividing the IS200profiles B and F.

VOL. 32, 1994

on Novem

ber 6, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

2130 PELKONEN ET AL.

TABLE 1. Distribution of genotypes among S. infantis isolatesfrom various sources

No. of isolates from the following source:Genotypea Broiler . Human, Human, Total

chicken Bovine domestic foreign

1A 31 11 0 2 442B 5 0 0 0 53B 18 0 0 0 182C 1 0 0 0 11E 1 0 0 0 14D 1 0 0 0 13J 1 0 0 0 11L 1 0 0 0 11B 0 0 13 11 241F 0 0 1 21 223F 0 0 0 6 65D 0 0 0 1 11G 0 0 0 2 23H 0 0 0 2 23K 0 0 0 2 2

Total (n = 15) 59 11 14 47 131a Genotypes obtained with BanI-digested DNA. The number refers to the

ribotype, and the letter refers to the IS200 profile.

Profiles obtained with IS200 probe. Molecular fingerprint-ing of the Salmonella-specific insertion sequence IS200 and itsinsertion site in the chromosome was carried out by using thesame DNA membranes used for the 16S rn gene probe. BanIand Pstl do not cut IS200, while EcoRI has one site in IS200 ofSalmonella serovar typhimurium which is located in the middleof the probe, a 592-bp HaeIII fragment of IS200 (2). Hybrid-ization of BanI-fragmented DNA with the IS200 probe pro-duced sharp bands, and it was easy to estimate the copynumber of IS200. The IS200 probe did not hybridize with DNAfrom P. multocida, Y pseudotuberculosis, or E. coli. Thenumber of IS200 sequences observed for the control strainsSalmonella serovars typhimurium, dublin, and enteritidis afterBanI digestion were in accordance with previously publishedresults obtained by Hincll or PstI digestion (4, 13, 25, 27).

ABC E D. FG H KLa

a. a. so GM a. a.

rna.a.mw 4Wrn4W4w4w 4

- 8.5_ -7 0

The S. infantis strains contained up to seven insertionelements, as judged from hybridization of IS200 with BanI-digested DNA (Fig. 2a; Table 2). While 2 isolates had noIS200, 88 strains had four copies and 28 strains had threecopies. Eleven unique profiles for IS200 in the chromosomewere observed (Fig. 2a). One IS200-containing fragment withan estimated molecular size of 2.2 kb was present in nineprofiles. In addition, eight profiles contained bands of 1.6 and0.6 kb. The 1.6-kb band appeared to contain two copies ofIS200 in genotypes A, E, and L, as judged by the intensity ofthe band. The copies could be separated into two separatefragments when DNA was digested with EcoRI-PstI, since inall three genotypes the IS200 probe hybridized with a unique5.0-kb fragment which was missing from strains with the otherIS200 profiles (Fig. 2b). Strains of genotype F differed fromthose of genotype A only in the copy number of IS200 in the1.6-kb band and, consequently, did not show any 5.0-kb bandamong EcoRI-PstI fragments (Fig. 2). Eleven different bandingprofiles were also obtained by hybridization of the IS200 probewith the EcoRI-PstI DNA fragments (Fig. 2b). For each BanIgenotype we obtained a typical EcoRI-PstI genotype, and forthis reason, the discriminatory power of IS200 profiling was notenhanced by using two different digestions. Thus, the resultsobtained with BanI fragments alone accounted for the geno-typic grouping.The IS200 profiles were fairly well restricted to certain

infection sources (Table 1). Three major IS200 profiles wereobserved: profile A for 44 isolates almost solely of animalorigin, profile B for 47 isolates of both animal and humanorigin, and profile F for 28 isolates of human origin. Ribotyp-ing could further divide the IS200 profile B into three geno-types and profile F into two genotypes, thus allowing completedifferentiation between animal and human infection sources(Table 1).

Distribution of genotypes among infection sources. By com-bining the ribotypes and the IS200 profiles obtained by BanIdigestion, 15 unique genotypes were defined (Table 1). Thegenotypes were highly restricted to the source of infection; thegenotypes of two human isolates of type 1A from foreignsources were the only ones overlapping those of isolates fromanimal and human sources. Broiler chicken farms in Finland

b

_ -3 .5

__ Z- 2. 2

4a -1.6-1 3

_4W_r 4_W4_ _ --0.6

A B C E D J F G II K L

_ _4 * _0-5.0

4.

0

.p.iaa-hWPa

4db

am 2. "I

41_ l .8

--t i:.I8

a -0.4

FIG. 2. Genomic Southern blot of selected S. infantis IS200 profile strains. Genomic DNAs were digested to completion with restrictionendonuclease BanI (a) or EcoRI-PstI (b), Southern blotted, and hybridized with the 0.6-kb IS200 probe which was labelled with 32p. The samestrains were analyzed in blots a and b. (a) Lanes A to L, IS200 profiles A to L, respectively, obtained with BanI-fragmented DNA. (b) Lanes Ato L, IS200 profiles A to L, respectively, obtained with EcoRI-PstI-fragmented DNA. The numbers on the right of both panels represent the sizes(in kilobases) of the respective hybridizing fragments.

J. CLIN. MICROBIOL.

on Novem

ber 6, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

GENOTYPING OF S. INFANTIS WITH IS200 AND 16S rrn 2131

TABLE 2. Clonal groups of S. infantis on the basis of conserved IS200 insertion sites

No. of IS200 copies in bands of the following sizeClonal IS200 No. of Ribotype(S)a No. of IS200 hybridizing with the IS200 probeb No. of othergroup profile strains P copies bands >2.2 kb

2.2 kb 1.6 kb 1.3 kb 0.6 kb

A 44 1 4 1 2 0 1 0AEL E 1 1 5 1 2 0 1 1

L 1 1 3 1 2 0 0 0

BCJ B 47 1, 2, 3 4 1 1 1 1 0C 1 2 6 1 1 1 1 2J 1 3 6 1 1 1 1 2

FG F 28 1, 3 3 1 1 0 1 0G 2 1 7 1 1 0 1 4

K K 2 3 4 1 0 0 1 2

HD H 2 3 1 0 0 0 0 1D 2 4,5 0 0 0 0 0 0

a Ribotypes associated with the IS200 profiles; BanI digestion.b Number of IS200 copies in the band; BanI digestion.

are affiliated with a parent company that provides birds andsome management services. The S. infantis isolates from all ofthe broiler chicken production companies in the country couldbe divided into eight different genotypes (Table 1). For two ofthe companies, sufficient numbers of isolates obtained duringthe period from 1986 to 1989 (24 and 19, respectively) wereanalyzed to allow assignment of company-associated S. infantisgenotypes, 1A and 3B. All of the bovine isolates were ofgenotype 1A, which was also prevalent in the broiler chickenfarms in the same geographical part of the country (Table 1).Eight genotypes were found among 47 S. infantis isolates frompeople who, preceding the infection, had been to a foreigncountry. The distribution of the genotypes among isolates frompeople who visited 12 countries varied considerably, and thesmall number of isolates made it impossible to draw anyconclusions about country-associated genotypes. All of thestrains isolated from human domestic epidemics shared acommon genotype, 1B or 1F, with strains involved in infectionscaused by isolates from foreign sources (Table 1). Genotype1B could be differentiated from the broiler chicken isolateseither by ribotype (ribotypes 2B and 3B in broilers) or by IS200profile (profiles 1A, 1E, and 1L in broiler chickens).

Clonal groups on the basis of IS200 profiles. Changes ininsertion sequence numbers and positions as defined by IS200profiling can reflect evolutionary changes (22). S. infantisstrains seemed to have four conserved chromosomal positionsfor IS200; 124 of 131 of the isolates had IS200 in three of thesepositions, and 3 strains had IS200 in two of these positions(Table 2). On the basis of the differential presence of theconserved insertion sites, we could classify three major clonalgroups: AEL, consisting of 46 strains; BCJ, consisting of 49strains; and FG, consisting of 30 strains (Table 2). The fourthclonal group, group K, shared two common bands with theother groups. The remaining four strains with one or no IS200copy did not share any similarities with the other strains.Group AEL was the most homogeneous; all of the strains wereof ribotype 1, and all but two strains were from animals. GroupFG contained two ribotypes and was solely associated withhumans, whereas group BCJ showed the highest degree ofheterogeneity, with three different ribotypes, and was prevalentin both human and animal populations.

DISCUSSION

Molecular fingerprinting of the Salmonella-specific insertionsequence IS200 and 16S rm gene proved to be a most suitablemethod for the typing of S. infantis. By using BanI digestion ofchromosomal DNA, 15 different genotypes were obtained bycombining the ribotypes and the IS200 profiles, and isolatesfrom different sources could be differentiated. AlthoughEcoRI-PstI digestion provided only one additional ribotype ofminor epidemiological interest, it confirmed the differencebetween IS200 profiles A and F and the existence of conservedinsertion sites of IS200. Until now there has not been anywidely acceptable method for subtyping serovar infantis; thus,we cannot compare the efficacy of the typing scheme with thatof any other epidemiological typing approach. A comparisonof IS200 profiling with phage typing for the subgrouping ofSalmonella serovar typhimurium has shown that some phagetypes can even be subtyped with the IS200 probe (2), and inparticular, clinical phage type strains showed great diversity intheir IS200 profiles (23). The combination of ribotyping andIS200 profiling for the genotyping of S. infantis may efficientlysubstitute for the lack of any suitable typing method for thisserovar.

Five ribotypes were obtained for S. infantis, which is in thesame range as those reported for many other Salmonellaserovars (5-7, 19, 23). Only in the case of Salmonella serovartyphi does ribotyping allow for the efficient differentiationbetween strains (1, 18). It is noteworthy that in some casesRFLP at the 16S rn locus can discriminate among isolateswithin a particular IS200 profile; isolates of IS200 profile B (47isolates) were divided into three genotypes, and isolates ofprofile F (28 isolates) were divided into two genotypes. Thisdifferentiation is critical for the complete discrimination be-tween human and animal infection sources.

Molecular fingerprinting of IS200 is a very sensitive methodfor the typing of Salmonella serovars which harbor severalcopies of the insertion sequence, as has recently been shownfor Salmonella serovars typhimurium, heidelberg, paratyphi B,and java (7, 23, 24). Most of the S. infantis strains (97%) hadat least three copies of IS200, while 71% of the isolates carriedfour copies (Table 2). It has been suggested that Salmonella

VOL. 32, 1994

on Novem

ber 6, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

2132 PELKONEN ET AL.

ba,aeria prefer to harbor a certain number of IS200 elements,and these numbers may differ from one serovar to another (13,24, 25). In the case of S. infantis, this copy number would befour.

T'he IS200 profiles are based on RFLP at insertion sites andDNA rearrangements following transposition (13). For each ofthe 11 profiles obtained with BanI-fragmented DNA, weobserved a corresponding IS200 profile with EcoRI-PstI-di-gested DNA. This finding suggests that the RFLP differencesdid not result from small nucleotide changes in the cleavagesite of BanI but, rather, resulted from the different insertionsites of IS200. However, the use of other restriction enzymesmight allow further discrimination between some strains. Wecould differentiate within the IS200 profile F with PstI diges-tion one strain that was dissimilar to the others (data notshown). Insertion sequences can integrate both into the chro-mosomal and into the plasmid DNA (13, 27). Studies withother Salmonella serovars have shown that IS200 is most often,except in some cases of Salmonella serovar enteritidis (23),inserted into the chromosome even in those serovars in whichplasmids occur frequently (23, 24). We have found plasmidsonly in 18% of S. infantis strains (unpublished data). In thepresent study we observed no correlation between a particularIS200 profile and plasmids. Thus, the IS200 profiles most likelyreflect chromosomal genotypic differences and can, therefore,be applied to the study of clonal relationships among isolateswithin this serovar.The isolates from two infection sources examined in the

present study, broiler chickens and human infections acquiredabroad, were supposed to be epideniiologically far apart fromeach other. Among broiler chickens the infection has prevailedirn Finland since the first outbreak in 1971 (17). Another largeepidemic occurred in 1975 and 1976, but the relation of theisolates to those involved in the first epidemic remainedunclear (20). Later, the surveillance studies of the nationalsalmonella control program showed that new Salmonella sero-vars have only rarely been introduced into chicken productionfacilities (16). Therefore, the strains causing the prevalent S.nJantis infections might have originated from the epidemicsthat occurred during the 1970s. In the present study weregarded the chicken strains as domestic and as having theleast possible genetic change from that in the past. The isolatesfrom human infections acquired in 1 of 12 foreign countries, onthe other hand, were thought to represent strains with thehighest possible genetic variation. The finding that particulargenotypes are typical of isolates from broiler chicken produc-tion facilities and differ from the genotypes prevalent abroadsupports the hypothesis of different genetic backgrounds.However, genetic variability among the broiler chicken strains,surprisingly, was as high as that among isolates acquiredabroad; eight genotypes, which were almost equally distrib-uted, could be found among isolates from both infectionsources. Still, a clear conservation of genotypes could bedetected among isolates from within the same broiler chickenfacilities. Genotypes 1A and 3B, which were associated withtwo of the facilities, differed from each other both in theirribotypes and in their IS200 profiles, which firmly proves thatthey represent different clonal lineages. The prevalent S.infantis infection in broiler chickens evidently originates fromat least two infection sources, which tempts us to speculate thatthe two sources would represent the major epidemics of the1970s.

It is striking that none of the isolates from human domesticepidemics had any genotype typical of isolates from broilerc;hicken production facilities. It is of utmost importance toconfi-m this finding by more comprehensive epidemiological

studies before drawing any conclusions about the sources ofdomestic human infections. Traditionally, broiler chickenshave been regarded as the main source of domestic humaninfections, as also indicated by the occurrence of human S.infantis infections after the first outbreak in poultry in the1970s (17). We did not include in our study isolates fromsporadic cases of salmonellosis acquired at home. Thosestrains may have still another type of genotypic distribution;however, we would at least expect to see some genotypes thesame as those from isolates from broiler chickens. Two largerdomestic S. infantis epidemics were recently traced back to aninfection carrier working in a kitchen (9), but the originalsource of infection often remains obscure. Because of thewidespread use of a competitive exclusion method as a pre-ventive measure, the current prevalence of Salmonella infec-tions has declined to less than 5% among broiler chicken flocks(11, 17). Only 5 to 11% of broiler chicken carcasses arecontaminated, and these only with a minute amount of Salmo-nella bacteria (11). Therefore, it is evident that other potentialsources of human infection must be considered as well, espe-cially as foreign travel has increased. We believe that the typingscheme presented here will help us to determine the infectionsources of epidemics and understand the epidemiology of S.infantis infections in human and animal populations.

ACKNOWLEDGMENTS

We thank Tuija Saranpaa for providing the isolates from cattleinfections; Matti Jahkola, Nancy Noben, and Esko Nurmi for criticallyreading the manuscript; and John Stanley for sharing in-press results.The study has been supported in part by Finnish Veterinary Science

Foundation and the I. Smolander Foundation.

REFERENCES1. Altwegg, M., F. W. Hickman-Brenner, and J. J. Farmer. 1989.

Ribosomal RNA gene restriction patterns provide increased sen-sitivity for typing Salmonella typhi strains. J. Infect. Dis. 160:145-149.

2. Baquar, N., E. J. Threlfall, B. Rowe, and J. Stanley. 1993.Molecular subtyping within a single Salmonella typhimurium phagetype, DT204c, with a PCR-generated probe for IS200. FEMSMicrobiol. Lett. 112:217-222.

3. Brosius, J., T. J. Dull, D. D. Sleeter, and H. F. Noller. 1981. Geneorganization and primary structure of a ribosomal RNA operonfrom Escherichia coli. J. Mol. Biol. 148:107-127.

4. Chowdry, N., E. J. Therfall, B. Rowe, and J. Stanley. 1993.Genotype analysis of faecal and blood isolates of Salmonelladublin from humans in England and Wales. Epidemiol. Infect.110:217-225.

5. Esteban, E., K. Snipes, D. Hird, R. Kasten, and H. Kinde. 1993.Use of ribotyping for characterization of Salmonella serotypes. J.Clin. Microbiol. 31:233-237.

6. Ezquerra, E., A. P. Burnens, K. Frith, M. Costa, and J. Stanley.1993. Molecular genotype analysis of Salmonella bovismorbificans.Mol. Cell. Probes 7:45-51.

7. Ezquerra, E., A. P. Burnens, C. Jones, and J. Stanley. 1993.Genotypic typing and phylogenetic analysis of Salmonella para-typhi B and S. java with IS200. J. Gen. Microbiol. 139:2409-2414.

8. Grimont, F., and P. A. D. Grimont. 1986. Ribosomal ribonucleicacid gene restriction patterns as potential taxonomic tools. Ann.Inst. Pasteur/Microbiol. (Paris) 137B:165-175.

9. Hatakka, M. 1992. Salmonella outbreak among railway and airlinepassengers. Acta Vet. Scand. 33:253-260.

10. Helmuth, R, R Stephan, C. Bunge, B. Hoog, A. Steinbeck, and E.Bulling. 1985. Epidemiology of virulence-associated plasmids andouter membrane protein patterns within seven common Salmo-nella serotypes. Infect. Immun. 48:175-182.

11. Hirn, J., E. Nurmi, T. Johansson, and L. Nuotio. 1992. Long-termexperience with competitive exclusion and salmonellosis in Fin-land. Int. J. Food Microbiol. 15:281-285.

J. CLIN. MICROBIOL.

on Novem

ber 6, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from

GENOTYPING OF S. INFANTIS WITH IS200 AND 16S rrn 2133

12. Kasatiya, S., T. Caprioli, and S. Champoux. 1979. Bacteriophagetyping scheme for Salmonella infantis. J. Clin. Microbiol. 10:637-640.

13. Lam, S., and J. R. Roth. 1983. IS200: a Salmonella-specificinsertion sequence. Cell 34:951-960.

14. Laszlo, V. G., K. Csak, and E. S. Csorian. 1988. A phage typingsystem for Salmonella infantis. Acta Microbiol. Hung. 35:55-69.

15. National Public Health Institute of Finland, Laboratory of EntericPathogens. 1992. Annual report. National Public Health Instituteof Finland, Helsinki.

16. National Veterinary and Food Research Institute of Finland,Department of Bacteriology. 1992. Annual report. National Vet-erinary and Food Research Institute of Finland, Helsinki.

17. Nurmi, E., and M. Rantala. 1973. New aspects of Salmonellainfection in broiler production. Nature (London) 241:210-211.

18. Pang, T., M. Altwegg, G. Martinetti, C. H. Koh, and S. Puthuc-heary. 1992. Genetic variation among Malaysian isolates of Sal-monella typhi by ribosomal RNA gene restriction patterns. Micro-biol. Immunol. 5:539-543.

19. Pignato, S. P., G. Giammanco, F. Grimont, and P. A. D. Grimont.1992. Molecular typing of Salmonella enterica subsp. entencaserovar Wien by rRNA gene restriction patterns. Res. Microbiol.143:703-709.

20. Raevuori, M., E. Seuna, and E. Nurmi. 1978. An epidemic ofSalmonella infantis infection in Finnish broiler chickens in 1975-76. Acta Vet. Scand. 19:317-330.

21. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecularcloning: a laboratory manual, 2nd ed. Cold Spring Harbor Labo-

ratory Press, Cold Spring Harbor, N.Y.22. Sawyer, S. A., D. E. Dykhuizen, R. F. Dubose, L. Green, T.

Mutangadura-Mhlanga, D. F. Wolczyk, and D. L. Hartl. 1987.Distribution and abundance of insertion sequences among naturalisolates of Eschenichia coli. Genetics 115:51-63.

23. Stanley, J., N. Baquar, and J. E. Threlfall. 1993. Genotypes andphylogenetic relationships of Salmonella typhimurium are definedby molecular fingerprinting of IS200 and 16S rrn loci. J. Gen.Microbiol. 139:1133-1140.

24. Stanley, J., A. Burnens, N. Powell, N. Chowdry, and C. Jones.1992. The insertion sequence IS200 fingerprints chromosomalgenotypes and epidemiological relationships in Salmonella heidel-berg. J. Gen. Microbiol. 138:2329-2336.

25. Stanley, J., N. Chowdry, N. Powell, and E. J. Threlfall. 1992.Chromosomal genotypes (evolutionary lines) of Salmonella berta.FEMS Microbiol. Lett. 95:247-252.

26. Stanley, J., M. Goldsworthy, and E. J. Threlfall. 1992. Molecularphylogenetic typing of pandemic isolates of Salmonella entenitidis.FEMS Microbiol. Lett. 90:153-160.

27. Stanley, J., C. S. Jones, and E. J. Threlfall. 1991. Evolutionarylines among Salmonella enteritidis phage types are identified byinsertion sequence IS200. FEMS Microbiol. Lett. 82:83-90.

28. Stull, T. L., J. J. LiPuna, and T. D. Edlind. 1988. A broad-spectrum probe for molecular epidemiology of bacteria: ribosomalRNA. J. Infect. Dis. 157:280-286.

29. Threlfall, E. J., and J. A. Frost. 1990. The identification, typing andfingerprinting of Salmonella: laboratory aspects and epidemiolog-ical applications. J. Appl. Bacteriol. 68:5-16.

VOL. 32, 1994

on Novem

ber 6, 2020 by guesthttp://jcm

.asm.org/

Dow

nloaded from