Yersinia enterocolitica and Y. enterocolitica-like ... - Springer Link

Eur J Clin Microbiol Infect Dis (2009) 28:757–765 DOI 10.1007/s10096-008-0696-y

ARTICLE

Yersinia enterocolitica and Y. enterocolitica-like species in clinical stool specimens of humans: identification and prevalence of bio/serotypes in Finland L. M. Sihvonen & K. Haukka & M. Kuusi & M. J. Virtanen & A. Siitonen & YE study group* Received: 17 September 2008 / Accepted: 22 December 2008 / Published online: 14 February 2009 # Springer-Verlag 2009

Abstract This study investigated the prevalence of Yersinia enterocolitica (YE) bio/serotypes and YE-like species in clinical stool specimens. The special aim was to find the best methods for accurate identification of YE species and, further, pathogenic strains among YE isolates. Of the 41,848 specimens cultured in ten laboratories during a 12-month period, 473 Yersinia strains were isolated from 462 patients. The strains were identified by 21 biochemical tests, serotyping, colony morphology, as well as by 16S rRNA and gyrB gene sequencing. The most prevalent Yersinia findings were YE biotype 1A (64% of the strains) and pathogenic bio/serotype *Finnish clinical microbiology laboratories’ YE study group: Martti Vaara, Eveliina Tarkka (Hospital District of Helsinki and Uusimaa), Antti Nissinen, Jaakko Uksila (Central Hospital of Central Finland, Microbiology Laboratory), Erkki Eerola (Turku University, Department of Medical Microbiology), Hannu Sarkkinen, Pauliina Kärpänoja (Päijät-Häme Social and Health Care Group, Clinical Microbiology), Päivi Strandén (Microbiological laboratory, Central Hospital of Seinäjoki), Pekka Ruuska (Joint Authority of Kainuu Region, Clinical Laboratory, Kainuu Central Hospital), Markku Koskela (Oulu University Hospital, Microbiology laboratory), Ulla Larinkari, Vesa Kirjavainen (Laboratory of Clinical Microbiology, Kymenlaakso Central Hospital, Kotka), Anja Kostiala-Thompson, Tiina Muuronen (VITA Terveyspalvelut Oy), Anna Muotiala (Medix Laboratories Ltd). L. M. Sihvonen : K. Haukka : M. Kuusi : M. J. Virtanen : A. Siitonen (*) National Institute for Health and Welfare, P.O. box 30, FI-00271 Helsinki, Finland e-mail: [email protected] L. M. Sihvonen e-mail: [email protected] K. Haukka e-mail: [email protected] M. Kuusi e-mail: [email protected] M. J. Virtanen e-mail: [email protected]

4/O:3 (16%). The cold-enrichment increased the number of all isolates, and 25% of the bio/serotype 4/O:3 and 2/O:9 strains were only found by cold-enrichment. In routine diagnostic laboratories, 50% of the YE-like species were identified as YE and in 26% the identification differed from that of the reference laboratory. The microscopic colony identification on CIN agar with positive CR-MOX test, combined with several biochemical tests, identified reliably the pathogenic YE bioserotypes and most YE BT 1A strains, but some strains of the YE-like species were so heterogenic that gene sequencing was the only way to identify them.

Introduction Yersinia enterocolitica (YE) is a zoonotic enteropathogen that encompasses strains of varying pathogenicity. Pathogenic YE strains usually cause acute gastroenteritis, which may result in severe postinfectious complications, such as reactive arthritis, myocarditis and erythema nodosum [3]. The high pathogenic YE belong to the subspecies (ssp.) enterocolitica biotype (BT) 1B [27]. The BT 1B strains possess a genomic island of high-pathogenicity [5] in addition to 70 kB Yersinia virulence plasmid (pYV). The majority of YE worldwide, however, belong to the ssp. palearctica [27]. YE ssp. palearctica includes five BTs (1A, 2, 3, 4 and 5) and non-biotypable (NBT) strains. All BTs are associated with numerous serotypes. BTs 2–5 have pYV and chromosomal virulence marker genes, such as attachment-invasion-locus (ail) and invasin (inv) [3]. The most frequently isolated pYV-positive YE ssp. palearctica strains in Europe belong to bio/serotypes 4/O:3 and 2/O:9 [9]. The BT 1A is commonly called non-pathogenic or environmental YE because these strains lack pYV and most of the chromosomal virulence markers. However, the BT 1A strains have also been reported to be linked to gastroenteritis or

758

potential pathogenecity [2, 4, 14, 18, 24, 25]. In addition to Y. enterocolitica, there are also Yersinia species that are often collectively called YE-like species including Y. aldovae, Y. bercovieri, Y. frederiksenii, Y. intermedia, Y. kristensenii, Y. mollaretii, Y. rohdei, Y. alecksiciae, and recently described Y. similiensis [31] and Y. massiliensis [26]. Some strains of these species may be able to cause disease in humans [32]. In Europe, the incidence of human yersiniosis caused commonly by YE has recently been reported to be 2.6 and 2.1 cases per 100,000 population in 2005 and 2006, respectively [9]. However, these data are believed to represent underreporting, largely because of the lack of systematic yersinia culturing of the stool samples in most countries and of difficulties with isolation techniques. On the other hand, problems in characterization of the yersinia isolates can lead to overreporting, when non-pathogenic yersinia findings are also reported. As a consequence, the data collected in Europe, as well as worldwide, are not comparable, leading to major uncertainties regarding the true incidence and burden of disease due to human yersiniosis [9]. In Finland, YE is the third most common bacterial enteric pathogen after Campylobacter and Salmonella with around 500 reported human infections annually [1]. YE infections confirmed by culturing or by detection of antibodies are reported to the Finnish National Infectious Diseases Register (NIDR). Most of the cases are sporadic but food-borne outbreaks have also occurred [1]. Several Finnish clinical microbiology laboratories use cold-enrichment to isolate YE but biotypes are determined only rarely. The differentiation of YE from YE-like species is challenging if the identification is based on phenotypic characteristics and biochemical reactions. For instance, the widely used Api 20E (BioMerieux, Marcy l’Etoile, France) incorrectly identifies Y. aldovae, Y. mollaretii, Y. bercovieri, Y. rohdei, or Y. alecksiciae as YE. Furthermore, if serotyping is used, similar O-antigens are found in strains of the pathogenic YE bio/serotypes (4/O:3; 3/O:5,27; 1B/O:8), in YE BT 1A strains (O:5, O:8), and in those of other Yersinia species, such as Y. aldovae (O:8), Y. mollaretii (O:3), Y. bercovieri (O:8), Y. kristensenii (O:8), and Y. frederiksenii (O:3) [30, 32]. This study investigated all human faecal Yersinia findings (excluding Y. pseudotuberculosis) in ten Finnish clinical microbiology diagnostic laboratories in 2006. The aim was to find the best methods for accurate recognizing the strains belonging to Y. enterocolitica and pathogenic strains among Y. enterocolitica isolates. For that, detailed phenotyping and 16S rRNA gene and DNA gyrase B subunit (gyrB) gene sequencing when necessary were applied to obtain precise identification and prevalence of different bio/serotypes of YE and YE-like strains. The results obtained were compared to the original data received from the clinical laboratories. Based on the results, an identification scheme has been

Eur J Clin Microbiol Infect Dis (2009) 28:757–765

suggested to improve Yersinia identification in routine laboratories in Finland and elsewhere.

Materials and methods Bacterial strains Ten clinical microbiology laboratories throughout Finland sent all their Yersinia strains isolated from 41,848 human stool samples in 2006 to the Enteric Bacteria Laboratory (EBL) of the National Public Health Institute (KTL) for further investigations. When isolating Yersinia spp. from stool samples, all ten laboratories used direct plating on selective cefsulodin-irgasan-novobiocin (CIN) agar and 1+ 1 days incubation varying from room temperature to 35°C. Seven of the ten laboratories used an additional 5–7 days’ cold-enrichment in peptone broth at 4–6°C prior to plating on CIN. In all, 462 Yersinia strains were isolated in the clinical laboratories and sent to EBL. In addition, EBL received 308 primary CIN plates of either direct culturing or of culturing after cold-enrichment to search for the possible coexistence of different Yersinia species in the stool sample. Several extra Yersinia colonies were marked on the primary plates by the forwarding laboratory and two of these colonies were always investigated at EBL in addition to the isolates received. This way, 11 additional strains that differed from the original strains were isolated. Thus, 473 Yersinia strains isolated from 462 subjects were included in the study. The laboratories reported to EBL if a strain had been isolated by cold-enrichment. Also, information on whether the patient had travelled abroad within a month prior to illness onset was reported on 208 patients by the laboratories. Phenotypic characterisation Initial identification of the strains was carried out in the clinical laboratories mainly using Api 20E (BioMerieux, Marcy l’Etoile, France). Some laboratories used biochemical tests for esculin, D-sucrose, L-rhamnose, D-xylose and pyrazinamidase (Roscolab ltd, London, UK). Most laboratories also used serotyping with commercially available YE antisera. In EBL, the colony morphology of the strains on CIN agar [16] was recorded by the Olympus SZX9 stereomicroscope (Olympus Europe, Hamburg, Germany). YE biotypes were determined as previously described [16, 36]. Reactions of the strains for 21 different substrates were determined after culturing for 48 h at 25°C (Table 1). The presence of pYV in the strains was tested on Congo-red magnesium oxalate (CR-MOX) agar, which demonstrates calcium dependency and Congo red absorption of a strain carrying pYV [29]. For cultivation on CRMOX, several colonies were streaked onto the plate to


759

Table 1 Biochemical reactions of the 473 Yersinia strains. The varied reactions of strains are indicated with percentages of positive reactions Species, bio/serotypea

Test

YE 4/O:3 (n=75)

YE 3/O:3 (n=2)

YE 2/O:9 (n=5)

YE BT 1A (n=302)

YE NBT (n=2)

Y. bercovieri (n=23)

Y. frederiksenii (n=29)

Y. intermedia (n=10)

Y. kristensenii (n=9)

Y. mollaretii (n=14)

Y. rohdei (n=2)

CR-MOX agarb Lipase Esculin D-Salisin D-Xylose D-Trehalose Pyrazinamidase Indole Voges-Proskauer L-Rhamnose D-Sucrose D-Melibiose L-Sorbose Mucate L-Fucose D-Raffinose D-Cellobiose Maltose D-Arabitol Glycerol α -Methyl-D-Glucoside

+ − − − − 99 − − 97 − + − 99 4 − − 63 91 1 95 −

+ − − − + + − − + − + − − − − − 50 + − + −

+ − − − + + − + + − + − + − − − + + − + −

(1) 98 + 99 + + 99 + 99 − + − + 1 87 4 + + 89 98 3

− + − − + + + − + − + − + − 50 50 + + + + −

(9) − − − + + 96 − 13 − + − 13 87 96 − + + 17 91 4

− 55 + + + + 83 + 66 93 + − + 90 97 4 + + + 97 4

(20) 60 + + + + + + 90 90 90 50 + + 90 50 + + 60 + 80

− 56 11 11 + + 78 11 11 11 11 − + 44 56 − + + 78 + −

− − 21 21 + + 86 − 7 7 + − + 36 − − 86 93 43 79 21

− − − − + + − − − − + − − − − + + + − + −

a b

“ + ” = 100% positive reactions, “ − “ = 100% negative reactions Values in parentheses represent atypical reaction with minor intake of Congo-red on CR-MOX agar differing from reaction of pYV+ strain

avoid the possibility of a single colony having lost its virulence plasmid. Serotyping was performed by slide agglutination of colonies incubated at 25°C for 48 h on Conradi-Drigalski agar using antisera against YE O:3, O:5, O:27, O:8, and O:9 (Sifin, Berlin, Germany). Since the reaction in the indole test is fundamental in the separation of YE BTs 2 and 3, all these strains, as well as Y. frederiksenii, Y. intermedia, and Y. kristensenii strains, were tested using the rapid spot indole test [10], Api 20E (24 h at 30°C), and the trypticase peptone (20g/L) broth (4 ml) tube test using 400 µl of Ehrlich reagent after incubation (48 h at 25°C). Sequencing of the 16S rRNA and gyrB genes DNA sequencing was applied to the strains with contradictory results in phenotypic tests or with untypical colony morphology. In that case, cells were grown at 30°C overnight on Drigalski agar and 5–10 colonies were collected for DNA extraction performed as described earlier [34]. The 16S rRNA genes were amplified with primers FD1mod (5′-AGAGTTTGATCYTGGYTYAG-3′) [22] and r533 (5′-TTACCGCGGCTGCTGGCAC -3′) [17]. The total reaction volume was 50 µl, containing 2 U of Dynazyme DNA polymerase (Finnzymes, Espoo, Finland), 5 µl of buffer

provided with the enzyme, 10 pmol of each primer, and 10–50 ng of template DNA. PCR was performed as previously described [22]. The DNA gyrase B subunit (gyrB) genes were amplified with primers gyrBf (5′CGGCGGTTTGCAYGGYGTRGG-3′) and gyrBr (5′CAGSGTRCGRGTCATYGCCG-3′) as reported earlier [21]. The partial 16S rRNA and gyrB genes were sequenced for both directions using Big Dye Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) with an ABI Prism 310 Genetic Analyzer (Applied Biosystems) according to the manufacturer’s instructions. Sequences were viewed and manually edited with Sequence scanner software v1.0 (Applied Biosystems) and aligned in Bioedit [15]. Neighbor joining (NJ) trees, presented, were constructed from selected 16S rRNA and gyrB gene sequences using PAUP* 4.0 using a F84 susbstitution model [33]. Bootstrap values for NJ trees were calculated from 1,000 reiterations. Statistical analysis of biochemical profiles The usefulness of the biochemical tests for Yersinia identification was analysed using Bayesian discriminant analysis. The goal was to get as accurate a prediction of the YE biotype or YE-like species identity as possible

760

from as small a group of 21 tests as possible. Each test was assumed to be independent of the others. The number of tests needed for prediction (the size of the model) was assumed a priori to be uniform for all possibilities from 0 to 21. Given the size, each test had an equal chance of being included in the model. The results were presented as posterior probabilities of being included in the model, averaging over all model sizes. The resulting model was evaluated using JAGS version 1.0.1. [28]. The analysis was first done separately using all 11 biochemical profiles shown in Table 1 and then by grouping the isolates into three groups (CR-MOX positive YE, BT 1A, and YE-like Yersinia species).

Results Of the 473 strains isolated from 462 subjects, 82 (17%) belonged to BTs 2, 3, or 4 and serotypes (STs) O:3 or O:9 of YE, and 304 (64%) were of YE BT 1A (302 strains) or non-biotypable (2 strains) (Table 1). Of the 302 BT 1A strains, 56 (19%) and 42 (15%), agglutinated with antisera O:8 and O:5, respectively. The remaining 87 strains (18%) were assigned to Y. frederiksenii (29 strains), Y. bercovieri (23 strains), Y. mollaretii (14 strains), Y. intermedia (10 strains), Y. kristensenii (9 strains) and Y. rohdei (2 strains) species based on the biochemical tests or 16S rRNA and gyrB gene sequencing (Table 1). All of the additional 11 strains isolated in EBL from 11 subjects belonged to another Yersinia species or to a different YE biotype than the strain originally isolated in a clinical microbiology laboratory from the same subject. In ten of these patients, YE BT 1A co-existed with a YE-like species and in one case Y. kristensenii and YE 4/O:3 were found from the patient. Of the 208 patients whose travel information was known, 97 (47%) had travelled abroad prior to falling ill. The reference laboratory (EBL) investigated the colony morphology of the 473 strains on CIN agar by the stereomicroscope. All 82 YE ST O:3 and four of the five O:9 strains were identified correctly, as well as 97% of the 304 YE BT 1A and NBT strains. The colonies of O:3 and O:9 strains were approximately 500–700 µm in diameter after overnight incubation with dark red centre in the colony (“bulls-eye”). The BT 1A colonies were bright red with a diameter of over 1,000 µM. The colony morphology of other Yersinia species, especially Y. frederiksenii, Y. kristensenii, and Y. intermedia varied among the strains and some Y. frederiksenii (24%), Y. intermedia (22%), and Y. kristensenii (17%) had colony morphology resembling YE BT 1A on CIN agar. All of the 82 YE BT 4, 3 and 2 strains (all serotype O:3 or O:9 strains), mentioned above, were clearly positive in CR-MOX test and, thus, presumable harboured the pYV plasmid. Some YE BT 1A, Y. bercovieri, and Y. intermedia strains showed intake of


Congo-red on CR-MOX agar, but the colour of the colonies was lighter and their size was much larger than the size of the colonies of the bio/serotypes 3–4/O:3 and 2/O:9. The biochemical reactions of 75 YE bio/serotype 4/O:3 strains showed uniform reactions in tests used for identification and further biotyping (Table 1). Two strains of serotype O:3 differed from BT 4 by fermenting xylose instead of sorbose and were therefore assigned to BT 3. Some of the YE BT 1A strains showed variation with lipase and pyrazinamidase tests. The indole test of Api 20E at 30° C was reliable for recognising indole positive strains (e.g. YE 2/O:9 and Y. frederiksenii). The statistical analysis used was not able to predict suitable biochemical tests for separation of YE bio/serotypes and different YE-like species or the three groups (CR-MOX positive YE, BT 1A, and YElike Yersinia species). The most likely candidates for removal from the biochemical test panel as unnecessary were D-trehalose (posterior probability of inclusion 0%/1.1%), glyserol (0%/5.4%), maltose (0%/33.2%), L-rhamnose (88%/32%), D-melibiose (100%/98%) and D-sucrose (97%/95%). Seven laboratories with 364 Yersinia findings used cold-enrichment. This increased the number of all Yersinia findings: 25% (14/55) of the 4/O:3 and 2/O:9 strains, 72% (171/236) of the YE BT 1A strains, and 85% (57/67) of the YE-like strains were found only after coldenrichment. For 344 of the 462 strains (74%) the identification results in the ten clinical microbiology routine laboratories were the same to species level as in EBL (Fig. 1). However, biotypes of the YE strains were not reported by 9/10 laboratories. The YE 3–4/O:3 and 2/O:9 strains were identified with 94% (77/82), and YE BT 1A or NBT strains with 86% (257/300) accuracy as YE . Of the strains identified as YE-like species 50% (40/80) were identified as YE in routine laboratories. The percentage of correct identification of Y. frederiksenii and Y. kristensenii was 21% and 67%, respectively. Y. intermedia, Y. bercovieri, Y. mollaretii, and Y. rohdei were not identified to species level. Because of the contradictory biochemical identification results, partial 16S rRNA gene sequences of 43 strains and gyrB gene sequences of 34 strains were obtained. They clustered the tested YE, as well as Y. bercovieri, Y. mollaretii and Y. rohdei strains. Phylogenetic analysis of gyrB genes separated YE 1A from YE 4/O:3 and 2/O:9 strains, which all clustered in one group with 16S rRNA gene sequences (Fig. 2). Sequencing of the gyrB gene was also useful for separating Y. kristensenii and Y. intermedia. Y. frederiksenii grouped in two clusters in the 16S rRNA gene tree and in three clusters in the gyrB tree (Fig. 2). The GenBank/EMBL/DDBJ accession numbers for the sequences are FM958188FM958264.

Eur J Clin Microbiol Infect Dis (2009) 28:757–765 Fig. 1 (a) Primary identification of the strains by clinical laboratories. (b) Identity of the 462 Yersinia isolates of the study determined in EBL

761

a)

Suspected Yersinia sp. (n=65) 14 %

YE ST O:3 (n=45) 10 %

Y. fred / Y. kris (n=1) 0%

YE ST O:9 (n=8) 2% YE ST O:5 (n=19)) 4%

Y. fred / Y. int (n=8) 2%

YE ST O:8 (n=25) 5%

Y. kristensenii (n=6) 1 % Y. frederiksenii (n=8) 2%

YE BT 1A (n=26) 6%

YE, no BT/ST determined (n=251) 54 %

b)

Y. mollaretii (n=12) 3% Y. kristensenii (n=6) 1% Y. rohdei (n=2) 0% Y. intermedia (n=9) 2% YE 3-4/O:3 (n=77) 17%

Y. frederiksenii (n=28) 6 % Y. bercovieri (n=23) 5%

YE 2/O:9 (n=5) 1%

YE NBT (n=2) 0%

YE BT 1A (n=298) 65 %

Discussion The prevalence of different YE bio/serotypes and YE-like species in almost 42,000 stool samples collected for diagnosis of intestinal bacterial pathogens were investigated. During a 12-month period, YE and YE-like species were isolated from the stools of 462 patients in ten laboratories.

YE BT 1A was the most prevalent Yersinia species constituting almost two-thirds of all Yersinia isolates. In Finland, about 500 Yersinia infections are reported annually by all 27 clinical microbiology laboratories [1]. Previously, however, no reliable data on the prevalence of YE bio/ serotypes and YE-like species in the stool specimens of Finns was available.

762


Fig. 2 Phylogenetic analysis of partial 16S rRNA genes (533 bp) (a) and gyrB genes (477 bp) (b) of selected strains

Our study showed that 50% of the YE-like species were identified as YE in routine diagnostic laboratories and in 26% the identification of the species level differed from that of the reference laboratory. This demonstrates that the identification tests used in these laboratories were not sufficient for separating YE from YE-like strains. A previous study showed that colony microscopy on CIN agar distinguished the Y. bercovieri and Y. mollaretii strains from clinically significant YE strains [16]. Also in the current study, the investigation of colony morphology on CIN agar was effective as a preliminary identification in separating pathogenic biotypes from the YE BT 1A and YE-like species, as well as most YE-like species from YE BT 1A. A CR-MOX test was positive for all the YE strains that had pathogenic biotype. However, it is important to use a reference strain as a positive control with CR-MOX since some YE-like strains can give an atypical colour reaction. The serotyping results showed that in Finland O:3 was the most common serotype (94%) among pYV positive (CRMOX positive) strains, O:9 strains were rarely (6%) seen, while O:5,27 strains were absent among the studied strains. If serotyping without biotyping is used for identification, it is noteworthy that 15% of the BT 1A strains agglutinated with antisera against O:5. Therefore, it is important to also use O:27 antiserum in serotyping to distinguish the

pathogenic bio/serotype 2–3/O:5,27 strains from the potentially non-pathogenic YE BT 1A O:5 strains. If, for example, Api 20E is used in a laboratory as a preliminary test, some additional biochemical tests or PCR-based methods [12] can be adopted. We present a simple identification scheme for identification of YE strains in routine laboratories (Fig. 3). It exploits the results of microscopy of the colonies on CIN agar, growth on CR-MOX, reactions of lipase, D-xylose, pyrazinamidase, esculin, D-trehalose and indole, as well as serotyping. Small colonies with dark red centres, positive CR-MOX together with negative pyrazinamidase, and esculin reactions were characteristics that, in our study, distinguished common pathogenic YE biotypes from BT 1A and YE-like species. Characteristic reactions for distinguishing pathogenic YE biotypes from each other are negative D-xylose (BT4 and BT 5), positive indole test (BT 1B, BT 2), and negative D-trehalose (BT 5). The use of the lipase test is necessary for separating BT 1B from other pathogenic biotypes. If a BT 1B is suspected, 16S rRNA sequencing may be wise to use for verifying this rare biotype of YE ssp. enterocolitica. In addition, serotyping with antisera against O:3, O:9, O:5, and O:27 combined with the tests above, indicated the identity of the most common pathogenic strains. In our study, YE 4/O:3 strains were always D-xylose negative, which differentiated


763

Fig. 3 Simple identification scheme for Yersinia enterocolitica. Abbreviations: lip lipase, xyl D-xylose, pyz pyrazinamidas e, esc esculin, tre D-trehalose, ind indole test. * Few exceptions may have a negative reaction

Yersinia enterocolitica? Colony microscopy on CIN-agar after o/n incubation

Is it pathogenic? 1. CR-MOX agar test using several colonies or 2. PCR test based on pYV virulence genes

Pathogenic YE biotypes?

YE BT 1B lip + xyl + pyz esc tre + ind +

Serotyping?

e.g. O:8 ssp. enterocolitica

YE BT 2 lip xyl + pyz esc tre + ind +

Serotyping?

Biotyping

YE BT 3 lip xyl + pyz esc tre + ind -

Serotyping?

O:9

O:5,27 ssp. palearctica

them from other YE BTs and YE-like strains. In addition, the esculin hydrolysis test worked better in distinguishing pathogenic YE biotypes from BT 1A than the pyrazinamidase test. In other words, some YE BT 1A strains, as well as some Y. frederiksenii, Y. kristensenii, Y. rohdei, Y. mollaretii, and Y. bercovieri also had a negative pyrazinamidase reaction like pathogenic YE biotypes. The biochemical heterogeneity of the YE-like strains studied, especially those of species Y. frederiksenii, Y. intermedia, and Y. kristensenii, was high. Some Y. bercovieri and Y. mollaretii strains may also have similar biochemical profiles. Sequencing of the partial 16S rRNA gene separated YE well from the other species. Sequencing of the gyrB gene has been widely used in phylogenetics of Enterobacteriaceae [6], and therefore there are many reference gene sequences available in public genomic databanks. With gyrB it was also possible to separate YE BT 1A from 4/O:3 and 2/O:9 strains. In addition, Y. kristensenii and Y. intermedia were better separated with the gyrB than the 16S rRNA gene. Y. bercovieri, Y. mollaretii, and Y. rohdei clustered well into their own groups with both genes. The Finnish Y. frederiksenii strains studied clustered in two groups by the 16S rRNA gene tree and in three groups in the gyrB tree. The high genetic diversity of Y. frederiksenii has been well recognised in earlier studies [7, 21]. The most common pathogenic YE bio/serotype in the present study was 4/O:3. Notably, 25% of the pathogenic bio/ serotypes were only detected after cold-enrichment showing that it was a valuable method for isolating these types. However, this finding might be affected, for example, by the

YE BT 4 lip xyl pyz esc tre + ind -

Serotyping?

O:3

YE BT 5 lip xyl pyz esc tre ind -

Serotyping?

YE BT 1A or YE-like species?

YE BT 1A lip +/(-)* xyl + pyz +/(-)* esc + tre + ind +

YE-like species Y. aldovae Y. alecksiciae Y. bercovieri Y. frederiksenii Y. kristensenii Y. intermedia Y. mollaretii Y. rohdei Y. similis Y. massiliensis

YE-like species are biochemically very heterogeneous, and can often be identified by gene sequencing only.

phase of the disease, namely, acute diarrhea vs. chronic gastroenteritis, or even the convalescent phase. Otherwise, our results were congruent with the previous observations that cold-enrichment increases even more the number of YE BT 1A and other Yersinia spp. strains [19, 35]. Pigs have been the only source in which YE 4/O:3 have frequently been isolated [13] and prevalence of YE in Finnish fattening pigs has been reported to be 56% [20]. Therefore, it can be assumed that most of the Yersinia infections in Finland are domestic and originate from pork. Nevertheless, 47% of the patients whose travel history was known had returned from a trip abroad before getting ill. This suggests that not all the infections were of domestic origin. For instance, the two strains in our study with rare bio/serotype 3/O:3 were isolated from patients who had been to Thailand. This may indicate an Asiatic origin for these strains, since bio/serotype 3/O:3 have previously been reported mainly from Asia [1]. Only a few YE serotype O:9 strains were detected; the strains of this are seemingly also uncommon in Finland, although they probably are domestic pathogens [19]. YE biotype 1B was absent among our Yersinia findings which was not surprising since this biotype has never been reported from Finland and is rare in Europe [9]. The high prevalence of YE BT 1A (64%) in our study agrees with a study by McNally et al. [23]. They found that the majority of Yersinia isolates (53%) from humans belonged to YE BT 1A in Great Britain in 1999– 2000. In many countries Yersinia infections may be underdiagnosed [37]. Inefficient detection methods have been

764

mentioned as a reason for underdiagnosis and the registered low prevalence of YE [11]. The relatively high annual incidence of Yersinia in Finland, 11 per 100,000 population, compared to other European countries [8], is partly explained by the fact that in Finland an examination for YE by selective culturing is part of the routine stool culturing done by the clinical microbiology laboratories. Our results also suggested that the incidence of yersiniosis appears high in Finland partly because the YE findings in NIDR have often been reported without mentioning bio/ serotypes. The lack of bio/serotype or virulence factor information of the Yersinia isolates has been recognized as an obstacle to proper assessment of relevance of Yersinia in causing human disease also at the European level [9]. Therefore, it is important to develop and implement suitable methods for the correct and precise identification of the Yersinia species. Acknowledgments We thank Heini Flinck, Tarja Heiskanen, Sari Jaakola, and Joonas Iivonen for skilful technical assistance. Elisa Huovinen is acknowledged for her useful comments. This work was funded by grant 4850/501/2004 from the Finnish Ministry of Agriculture and Forestry.


12.

13.

14.

15.

16.

17.

18.

References

19.

1. Anonymous (2005) Infectious diseases in Finland 1995–2004. Publications of National Public Health Institute (KTL) Series B 13/2005 2. Bhagat N, Virdi JS (2007) Distribution of virulence-associated genes in Yersinia enterocolitica biovar 1A correlates with clonal groups and not the source of isolation. FEMS Microbiol Lett 266:177–260 3. Bottone EJ (1999) Yersinia enterocolitica: overview and epidemiologic correlates. Microbes Infect 1:323–333 doi:10.1016/ S1286-4579(99)80028-8 4. Burnens AP, Frey A, Nicolet J (1996) Association between clinical presentation, biogroups and virulence attributes of Yersinia enterocolitica strains in human diarrhoeal disease. Epidemiol Infect 116:27–34 5. Carniel E (1999) The Yersinia high-pathogenicity island. Int Microbiol 2:161–167 6. Dauga C (2002) Evolution of the gyrB gene and the molecular phylogeny of Enterobacteriaceae: a model molecule for molecular systematic studies. Int J Syst Evol Microbiol 52:531–547 7. Dolina M, Peduzzi R (1993) Population genetics of human, animal, and environmental Yersinia strains. Appl Environ Microbiol 59:442–450 8. EFSA (2006) The community summary report on trends and sources of zoonoses, zoonotic agents, antimicrobial resistance and foodborne outbreaks in the European Union in 2005. EFSA J 94 9. EFSA (2007) Scientific opinion of the panel on BIOHAZ on a request from EFSA on monitoring and identification of human enteropathogenic Yersinia spp. EFSA J 595:1–30 10. Eisenberg HD (1992) Clinical microbiology procedures handbook. American Society for Microbiology, Washington DC 11. Fredriksson-Ahomaa M, Korkeala H (2003) Low occurrence of pathogenic Yersinia enterocolitica in clinical, food, and environ-

20.

21.

22.

23.

24.

25.

26.

27.

mental samples: a methodological problem. Clin Microbiol Rev 16:220–229 doi:10.1128/CMR.16.2.220-229.2003 Fredriksson-Ahomaa M, Stolle A, Korkeala H (2006) Molecular epidemiology of Yersinia enterocolitica infections. FEMS Immunol Med Microbiol 47:315–329 doi:10.1111/j.1574-695X.2006. 00095.x Fredriksson-Ahomaa M, Stolle A, Siitonen A, Korkeala H (2006) Sporadic human Yersinia enterocolitica infections caused by bioserotype 4/O:3 originate mainly from pigs. J Med Microbiol 55:747–749 doi:10.1099/jmm.0.46523-0 Grant T, Bennett-Wood V, Robins-Browne RM (1999) Characterization of the interaction between Yersinia enterocolitica biotype 1A and phagocytes and epithelial cells in vitro. Infect Immun 67:4367–4375 Hall TA (1999) BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp Ser 41:95–98 Hallanvuo S, Peltola J, Heiskanen T, Siitonen A (2006) Simplified phenotypic scheme evaluated by 16S rRNA sequencing for differentiation between Yersinia enterocolitica and Y. enterocolitica-like species. J Clin Microbiol 44:1077–1080 doi:10.1128/ JCM.44.3.1077-1080.2006 Jalava J, Kotilainen P, Nikkari S, Skurnik M, Vanttinen E, Lehtonen OP, Eerola E, Toivanen P (1995) Use of the polymerase chain reaction and DNA sequencing for detection of Bartonella quintana in the aortic valve of a patient with culture-negative infective endocarditis. Clin Infect Dis 21:891–896 Kaneko S, Maruyama T (1987) Pathogenicity of Yersinia enterocolitica serotype O3 biotype 3 strains. J Clin Microbiol 25:454– 455 Kontiainen S, Sivonen A, Renkonen OV (1994) Increased yields of pathogenic Yersinia enterocolitica strains by cold enrichment. Scand J Infect Dis 26:685–691 doi:10.3109/00365549409008636 Korte T, Fredriksson-Ahomaa M, Niskanen T, Korkeala H (2004) Low prevalence of yadA-positive Yersinia enterocolitica in sows. Foodborne Pathog Dis 1:45–52 doi:10.1089/1535314047729 14455 Kotetishvili M, Kreger A, Wauters G, Morris JG Jr, Sulakvelidze A, Stine OC (2005) Multilocus sequence typing for studying genetic relationships among Yersinia species. J Clin Microbiol 43:2674–2684 doi:10.1128/JCM.43.6.2674-2684.2005 Kotilainen P, Jalava J, Meurman O, Lehtonen OP, Rintala E, Seppala OP, Eerola E, Nikkari S (1998) Diagnosis of meningococcal meningitis by broad-range bacterial PCR with cerebrospinal fluid. J Clin Microbiol 36:2205–2209 McNally A, Cheasty T, Fearnley C, Dalziel RW, Paiba GA, Manning G, Newell DG (2004) Comparison of the biotypes of Yersinia enterocolitica isolated from pigs, cattle and sheep at slaughter and from humans with yersiniosis in Great Britain during 1999–2000. Lett Appl Microbiol 39:103–108 doi:10.1111/ j.1472-765X.2004.01548.x McNally A, Dalton T, Ragione RML, Stapleton K, Manning G, Newell DG (2006) Yersinia enterocolitica isolates of differing biotypes from humans and animals are adherent, invasive and persist in macrophages, but differ in cytokine secretion profiles in vitro. J Med Microbiol 55:1725–1734 doi:10.1099/jmm.0.46726-0 McNally A, La Ragione RM, Best A, Manning G, Newell DG (2007) An aflagellate mutant Yersinia enterocolitica biotype 1A strain displays altered invasion of epithelial cells, persistence in macrophages, and cytokine secretion profiles in vitro. Microbiology 153:1339–1388 doi:10.1099/mic.0.2006/000919-0 Merhej V, Adekambi T, Pagnier I, Raoult D, Drancourt M (2008) Yersinia massiliensis sp. nov., isolated from fresh water. Int J Syst Evol Microbiol 58:779–784 doi:10.1099/ijs.0.65219-0 Neubauer H, Aleksic S, Hensel A, Finke EJ, Meyer H (2000) Yersinia enterocolitica 16S rRNA gene types belong to the same


28.

29.

30.

31.

32.

genospecies but form three homology groups. Int J Med Microbiol 290:61–64 Plummer M (2007) JAGS: Just Another Gibbs Sampler. http:// www-fis.iarc.fr/∼martyn/software/jags/. Accessed 18 January 2009 Riley G, Toma S (1989) Detection of pathogenic Yersinia enterocolitica by using congo red-magnesium oxalate agar medium. J Clin Microbiol 27:213–214 Robbins-Browne R (2007) Yersinia enterocolitica. In Doyle M, Beuchat LR (eds) Food microbiology: fundamentals and frontiers. ASM Press, Washington DC, pp 293–322 Sprague LD, Scholz HC, Amann S, Busse H-J, Neubauer H (2008) Yersinia similis sp. nov. Int J Syst Evol Microbiol 58:952– 958 doi:10.1099/ijs.0.65417-0 Sulakvelidze A (2000) Yersiniae other than Y. enterocolitica, Y. pseudotuberculosis, and Y. pestis: the ignored species. Microbes Infect 2:497–513 doi:10.1016/S1286-4579(00)00311-7

765 33. Swofford DL (2003) PAUP*. Phylogenetic analysis using parsimony (*and other methods). Sinauer Associates, Sunderland, MA 34. Thisted Lambertz S, Danielsson-Tham ML (2005) Identification and characterization of pathogenic Yersinia enterocolitica isolates by PCR and pulsed-field gel electrophoresis. Appl Environ Microbiol 71:3674–3681 doi:10.1128/AEM.71.7.3674-3681.2005 35. Van Noyen R, Vandepitte J, Wauters G, Selderslaghs R (1981) Yersinia enterocolitica: its isolation by cold enrichment from patients and healthy subjects. J Clin Pathol 34:1052–1056 doi:10.1136/jcp.34.9.1052 36. Wauters G, Kandolo K, Janssens M (1987) Revised biogrouping scheme of Yersinia enterocolitica. Contrib Microbiol Immunol 9:14–21 37. Zheng H, Sun Y, Lin S, Mao Z, Jiang B (2008) Yersinia enterocolitica infection in diarrheal patients. Eur J Clin Microbiol Infect Dis 27:741–752 doi:10.1007/s10096-008-0562-y