OUP user menu

Molecular epidemiology of Yersinia enterocolitica infections

Maria Fredriksson-Ahomaa, Andreas Stolle, Hannu Korkeala
DOI: http://dx.doi.org/10.1111/j.1574-695X.2006.00095.x 315-329 First published online: 1 August 2006

Abstract

Yersinia enterocolitica is an important food-borne pathogen that can cause yersiniosis in humans and animals. The epidemiology of Y. enterocolitica infections is complex and remains poorly understood. Most cases of yersiniosis occur sporadically without an apparent source. The main sources of human infection are assumed to be pork and pork products, as pigs are a major reservoir of pathogenic Y. enterocolitica. However, no clear evidence shows that such a transmission route exists. Using PCR, the detection rate of pathogenic Y. enterocolitica in raw pork products is high, which reinforces the assumption that these products are a transmission link between pigs and humans. Several different DNA-based methods have been used to characterize Y. enterocolitica strains. However, the high genetic similarity between strains and the predominating genotypes within the bio- and serotype have limited the benefit of these methods in epidemiological studies. Similar DNA patterns have been obtained among human and pig strains of pathogenic Y. enterocolitica, corroborating the view that pigs are an important source of human yersiniosis. Indistinguishable genotypes have also been found between human strains and dog, cat, sheep and wild rodent strains, indicating that these animals are other possible infection sources for humans.

Keywords
  • Yersinia enterocolitica
  • molecular typing
  • epidemiology

Introduction

Yersinia enterocolitica is an important zoonotic pathogen that can cause yersiniosis in humans and animals. The predominant symptom in humans, particularly in young children, is diarrhoea (Bottone, 1999). Secondary immunologically induced sequelae, such as reactive arthritis, are not uncommon, especially in HLA-B27-positive individuals. Most cases of yersiniosis occur sporadically without an apparent source (Bottone, 1997, 1999). Yersinia enterocolitica is thought to be a significant food-borne pathogen, even though pathogenic strains have seldom been isolated from foods. In studying raw pork, higher detection rates have been obtained by PCR than by culture methods (Fredriksson-Ahomaa & Korkeala, 2003a). In some case-control studies, an increased risk of yersiniosis has been demonstrated when raw or undercooked pork has been consumed (Tauxe et al., 1987; Ostroff et al., 1994). The pig is the only reservoir from which pathogenic Y. enterocolitica strains have frequently been isolated (Bottone, 1999). Nevertheless, the epidemiology of Y. enterocolitica infections is complex and remains poorly understood. To identify reservoirs of infection, transmission vehicles and associations between clinical cases, several DNA-based methods have been used in molecular typing of Y. enterocolitica (Iteman et al., 1996; Wojciech et al., 2004; Fearnley et al., 2005; Virdi & Sachdeva, 2005).

Yersinia enterocolitica infection

Yersinia enterocolitica infections have been observed on all continents but appear to be most common in Europe (Bottone, 1999). Food-borne outbreaks are rare and most infections are sporadic. The prevalence of Y. enterocolitica O:3/O:9-specific antibodies in healthy blood donors has been shown by immunoblotting to be relatively high in Finland (31%) and in Germany (43%) (Mäki-Ikola et al., 1997). This may indicate a high rate of subclinical yersiniosis in the healthy population. Yersinia enterocolitica can cause gastrointestinal symptoms ranging from mild self-limiting diarrhoea to acute mesenteric lymphadenitis, which can lead to appendicitis (Bottone, 1997). The clinical manifestations of the infection depend to some extent on the age and physical state of the patient, the presence of any underlying medical conditions and the bioserotype of the organism. Gastroenteritis, caused by Y. enterocolitica, is the most frequent form of yersiniosis, typically affecting infants and young children. In older children and young adults, acute yersiniosis can be present as a pseudoappendicular syndrome, which is frequently confused with appendicitis. Sometimes long-term sequelae, including reactive arthritis, erythema nodosum, uveitis, glomerulonephritis and myocarditis, will occur. Postinfection manifestations are mainly seen in young adults. Sepsis is a rare complication of Y. enterocolitica infection, except in patients who have a predisposing underlying disease or are in an iron-overloaded state. Sepsis can also occur during blood transfusion (Bottone, 1999).

Yersinia enterocolitica infection is typically initiated by ingestion of contaminated food or water. Yersinia enterocolitica initially travels to the small intestine and invades the intestinal barrier via M cells, a specialized group of follicle-associated epithelial cells that function in antigen uptake (Bottone, 1999; Pujol & Bliska, 2005). Adherence to and invasion of epithelial layers require at least two chromosomal genes, inv (invasion) and ail (attachment invasion locus). After invasion of the intestinal epithelium, the bacteria replicate within lymphoid follicles known as Peyer's patches. The presence of an approximately 70 kb virulence plasmid, termed pYV (plasmid for Yersinia virulence) enables Y. enterocolitica to survive and multiply in lymphoid tissues. The bacteria can spread from the lymphoid follicles to mesenteric lymph nodes, resulting in mesenteric lymphadenitis.

Yersinia enterocolitica strains belonging to certain few bioserotypes can cause human disease. Most strains associated with yersiniosis belong to the following bioserotypes: 1B/O:8; 2/O:5,27; 2/O:9; 3/O:3; 4/O:3. These bioserotypes have been shown to have different geographical distributions. Strains largely responsible for human yersiniosis in Europe, Japan, Canada and the USA belong to bioserotype 4/O:3 (Bottone, 1999). Strains of five biotypes (1B, 2, 3, 4 and 5) can carry the pYV, which is required for full expression of virulence, and several chromosomally encoded virulence determinants. Strains of biotype 1A lack the virulence-associated markers of pYV-bearing strains and are considered to be nonpathogenic. However, growing clinical, epidemiological and experimental evidence suggests that some biotype 1A strains are virulent and can cause gastrointestinal disease (Tennant et al., 2003). Several studies have been conducted to investigate the distribution of different virulence genes (ail, inv, yst, yadA, virF and yopT) among Y. enterocolitica strains by PCR (Thoerner et al., 2003; Falcao et al., 2004; Lee et al., 2004; Gürtler et al., 2005). A correlation between biotypes and the presence of plasmid and chromosomal virulence genes has been found. However, plasmid-borne genes (yadA, virF and yopT) have been detected with variable efficiency owing to heterogeneity within the bacterial population for the presence of the virulence plasmid.

Molecular detection of Y. enterocolitica in natural samples

Considerable difficulties are associated with isolating Y. enterocolitica from clinical, food and environmental samples. Conventional culture-dependent methods have several limitations, such as long incubation steps taking up to 4 weeks, lack of identification between species, and lack of discrimination between pathogenic and nonpathogenic strains (Cocolin & Comi, 2005). In addition, the ability of bacteria, including Y. enterocolitica, to persist in samples in a viable but nonculturable state can be a problem (Alexandrino et al., 2004). Thus, investigations have been undertaken to develop rapid and reliable methods for detecting pathogenic Y. enterocolitica in natural samples. In the field of molecular diagnostics, PCR is the accepted method for detecting nucleic acids in a variety of samples. PCR is the most commonly used nucleic acid amplification technique for the diagnosis of infectious diseases, surpassing the probe and signal amplification methods. Using PCR, pathogenic Y. enterocolitica can be detected in samples rapidly and with high specificity and sensitivity (Fredriksson-Ahomaa & Korkeala, 2003).

Several PCR assays have been developed to detect pYV-positive Y. enterocolitica in clinical, food and environmental samples (Fredriksson-Ahomaa & Korkeala, 2003a). Many of these samples use primers targeting the yadA or virF gene located on the pYV. Because of possible plasmid loss, PCR methods targeting chromosomal virulence genes have also been created for natural samples. The ail, inv and yst genes, located in the chromosome of pathogenic Y. enterocolitica strains, are the most frequently used chromosomal targets. In addition, some PCR assays using the Yersinia-specific region of the 16S rRNA gene have been designed to detect Yersinia spp., especially in blood samples (Table 1).

View this table:
Table 1

Real-time PCR assays developed for detection of Yersinia enterocolitica in clinical, food and environmental samples

SampleGene regionDetection systemReference
Blood16S rRNATaqmanSen (2000); Sen & Asher (2001)
16S rRNASYBRGreenWolffs et al. (2004)
FaecesailTaqmanJourdan et al. (2000)
16S rRNASYBRGreenWolffs et al. (2004)
yadASYBRGreenFukushima et al. (2003)
FoodailTaqmanBoyapalle (2001); Jourdan (2000)
ystTaqmanVishnubhatla (2000); Wu (2004)
16S rRNASYBRGreenWolffs et al. (2004)
16S rRNA, yadASYBRGreenWolffs et al. (2005)
Water16S rRNA, yadASYBRGreenWolffs et al. (2005)

Real-time PCR is a powerful advancement of the basic PCR technique. Since real-time PCR thermal cyclers became commercially available in the late 1990s, the technology has been recognized as an outstanding tool in molecular diagnostics. Compared with conventional PCR, real-time PCR facilitates automation, computerization and quantification of nucleic acids. A significant improvement introduced by real-time PCR is the increased speed with which it can produce results. This is largely due to reduced cycle times, removal of separate post-PCR detection procedures and use of sensitive fluorescence detection equipment, allowing earlier amplicon detection. This helps to increase throughput and reduce carry-over contamination.

At present, the most popular real-time PCR assays are based on ‘Taqman’ and ‘SYBRGreen’ approaches. The Taqman system is a 5′-nuclease assay that utilizes specific hybridization of a dual-labelled Taqman probe to the PCR product. The SYBRGreen system is based on the binding of the fluorescent SYBRGreen dye to the PCR product. Although both assays are potentially rapid and sensitive, their principles of detection and optimization are different, as is the resulting price per assay. Recently, several real-time PCR assays for qualitative detection of Y. enterocolitica in natural samples have been developed (Table 1). Boyapalle (2001) reported that the Taqman assay was 1000–10 000 times more sensitive than the culture method or traditional PCR assay when pig samples were studied. Sensitive methods are particularly necessary to detect pathogenic Y. enterocolitica in asymptomatic carriers, when studying, for example, possible animal reservoirs for this pathogen. Rapid and sensitive methods are also needed to detect small numbers of Y. enterocolitica organisms in blood units used for transfusion.

Molecular identification of Y. enterocolitica strains

Identification systems for Y. enterocolitica based on biochemical reactions do not guarantee reliable identification at species level. Members of nonpathogenic Yersinia species can easily be misidentified as Y. enterocolitica. To identify presumptively typed Y. enterocolitica, Neubauer (2000b) described a PCR method targeting the 16S rRNA gene combined with sequencing. Based on different DNA– DNA hybridization values and 16S rRNA gene sequences, Neubauer (2000a) demonstrated that Y. enterocolitica should be divided into two subspecies: Y. enterocolitica ssp. enterocolitica and Y. enterocolitica ssp. palearctica. Yersinia enterocolitica ssp. enterocolitica consists of strains of biotype 1B, and Y. enterocolitica ssp. palearctica contains the remaining strains.

Plasmid-encoded targets, like virF and yadA genes, and chromosomally encoded targets, like ail, inv, yst and rfb genes, have also been used in PCR to identify Y. enterocolitica strains (Fredriksson-Ahomaa & Korkeala, 2003a; Bhaduri et al., 2005; Thisted Lambertz & Danielsson-Tham, 2005). Thisted Lambertz & Danielsson-Tham (2005) designed a multiplex PCR targeting virF, ail, yst and rfbC genes for identification of pathogenic Y. enterocolitica isolates. Using these four primer pairs, it was possible to differentiate Y. enterocolitica serotype O:3 from other pathogenic serotypes as well as from Yersinia pseudotuberculosis in the same reaction. Jacobsen (2005) have recently developed a real-time PCR assay based on the perosamine synthetase (per) gene to detect Y. enterocolitica O:9. This serotype and Brucella spp. cross-react in serological tests due to identical O-antigens in their lipopolysaccharides.

Molecular subtyping of Y. enterocolitica strains

Outbreaks of infectious disease often result from exposure to a common source of the aetiological agent. Generally, the aetiological agent causing an outbreak of bacterial infection is derived from a single cell whose progeny are genetically identical or closely related to the source organism. In epidemiological studies, subtyping is important in recognizing outbreaks of infection, in detecting the cross-transmission of pathogens and in determining the source of the infection. Subtyping has been accomplished by a number of different approaches. All organisms within a species must be typeable by the method used. Moreover, the subtyping methods should have high reproducibility and good discriminatory power.

Subtyping of Y. enterocolitica strains has previously largely relied on phenotypic characteristics such as biochemical properties (biotyping), O and H antigens (serotyping), antimicrobial susceptibility (antibiogram typing) and bacteriophage lysis patterns (phage typing). However, because of their low discriminatory power, these techniques have partly been replaced by DNA-based molecular methods. Genotyping of Y. enterocolitica has made great strides in the last decades, and several different DNA-based methods have been used to characterize Y. enterocolitica strains (Table 2). Ideally, the typing methods should have high typeability, reproducibility and discriminatory power. They should also be inexpensive, generally available and easy to use. Finally, the results should be easy to interpret. However, the high similarity between strains and the predominating genotypes within the bioserotype have limited the benefit of these methods in epidemiological studies.

View this table:
Table 2

Methods for molecular typing of Yersinia enterocolitica isolates

Typing methodTypeabilityReproducibilityDiscriminatory powerUseInterpretation
REAPVariableGoodPoorEasyEasy
REACExcellentModerateModerateEasyDifficult
RibotypingExcellentExcellentVariableModerateEasy
PFGEExcellentExcellentGoodModerateEasy
PCRExcellentModerateVariableEasyModerate
AFLPExcellentGoodGoodModerateModerate
DNA sequencingExcellentExcellentGoodDifficultModerate
  • Modified from Virdi et al. (2005).

  • REAP, restriction endonuclease analysis of plasmid; REAC, restriction endonuclease analysis of chromosome; PFGE, pulsed-field gel electrophoresis; AFLP, amplified fragment length polymorphism.

Restriction endonuclease analysis of plasmid and restriction endonuclease analysis of chromosome

Plasmid analysis, the first bacterial typing tool, has been used to differentiate bacterial strains, but pathogenic strains of Y. enterocolitica contain only one virulence plasmid (pYV) of about 70 kb. Thus, for subtyping of pYV-positive Y. enterocolitica strains, the isolated plasmid has been cut with different frequent-cutting restriction enzymes (Table 3). Restriction endonuclease analysis of the plasmid (REAP) has been applied in several studies to compare pYV-positive Y. enterocolitica strains. This method has the advantage of being rapid and easy to perform but cannot be used for plasmid-negative strains. In most studies, the profiles were serotype-specific (Nesbakken et al., 1987; Kapperud et al., 1990; Kwaga & Iversen, 1993; Iteman et al., 1996). Serotype O:8 strains display the highest degree of REAP polymorphism, whereas strains of serotypes O:3 and O:9 display low diversity. Fukushima (1993, 1997) showed a close correlation between geographical distribution and REAP patterns among strains of serotype O:5,27. However, because of the low degree of polymorphism within serotypes O:3 and O:9, the use of REAP for epidemiological tracing of these serotypes did not yield much information.

View this table:
Table 3

Restriction enzymes used for typing of Yersinia enterocolitica strains

Typing methodEnzymes usedReference
Restriction endonuclease analysis of plasmidEcoRIIwata et al. (2005)
EcoRI, BamHINesbakken (1987); Kapperud (1990); Fukushima (1993), (1997); Iteman (1996); Hayashidani (2003)
EcoRI, BamHI, HindIIIKwaga & Iversen et al. (1993)
EcoRI, BamHI, HindIII, XbaIPulkkinen et al. (1986)
Restriction endonuclease analysis of chromosomeHaeIIIKapperud (1990); Fukushima (1993)
Southern hybridization
RibotypingEcoRIHayashidani (2003); Iwata (2005)
NciIBlumberg et al. (1991)
PvuIIBurnens et al. (1996)
AvaI, NciIAndersen & Saunders et al. (1990)
EcoRI, EcoRVFukushima et al. (1998)
HindIII, BglILobato et al. (1998)
EcoRI, EcoRV, NciIIteman et al. (1996)
HindIII, BglI, SalI, NciI, EcoRIMendoza et al. (1996)
YeO:3RSNciI, BglIHallanvuo et al. (2002)
IS Yen2EcoRIGolubov et al. (2005)
Pulsed-field gel electrophoresisNotISaken (1994); Pilon (2000); Hayashidani (2003); Korte (2004); Iwata (2005); Thisted Lamberzt & Danielsson-Tham, 2005
XbaIFavier et al. (2005)
NotI, BlnIJones et al. (2003)
NotI, XbaIBuchrieser (1994); Asplund (1998); Fredriksson-Ahomaa (2000a,b)
XbaI, SpeIFiletici et al. (2000)
NotI, ApaI, XhoIFredriksson-Ahomaa, (1999, 2001a,b, 2003, 2004)
NotI, XbaI, SpeINajdenski et al. (1994)
NotI, XbaI, SpeI, SfiIShayegani et al. (1995)
Amplified fragment length polymorphismBamHI, BspDIFearnley et al. (2005)
  • YeO:3RS, probe for Y. enterocolitica O:3 repeated sequence.

  • IS Yen2, probe for Y. enterocolitica insertion sequence element.

Chromosomal DNA restriction analysis was the first of the chromosomal DNA-based typing schemes. A major limitation of this technique is the difficulty in interpreting complex profiles, which consist of hundreds of bands that may be unresolved and overlapping. Kapperud (1990) used restriction endonuclease analysis of chromosome (REAC) to study polymorphism in restriction fragment patterns among Y. enterocolitica strains belonging to different bioserotypes. A total of 22 distinct REAC patterns were distinguished among the 72 Yersinia strains examined, and each pattern was bioserotype-specific. Some DNA profiles also varied between strains belonging to the same bioserotype. REAC proved to have the greatest discriminatory power within bioserotype 1B/O:8. Compared with bioserotype 1B/O:8, bioserotypes 3/O:3, 4/O:3, 2/O:5,27 and 2/O:9 were homogeneous. Most (12/13) of the strains of bioserotype 4/O:3 isolated in Norway could not be distinguished from each other by REAC, showing an overall genetic homogeneity.

REAC and Southern blotting

To avoid problems associated with complex REAC patterns, probes that hybridize to specific DNA sequences are used. The separated DNA fragments are transferred from the agarose gel to a membrane by Southern blotting. The membrane-bound nucleic acid is then hybridized to one or more labelled probes. Ribotyping refers to the use of nucleic acid probes to detect polymorphisms associated with the ribosomal operons. Automated ribotyping is very useful in epidemiological surveys as the first step owing to its ability to analyse a large number of bacterial isolates in a very short time and with minimal human effort. Ribotyping has been applied to characterize Y. enterocolitica isolates in several studies, but interstudy comparisons are difficult because of the use of different restriction enzymes (Table 3). However, these studies demonstrate a good correlation between ribotypes and bioserotypes. Ribotyping has also allowed, to some extent, the subtyping of strains within a given bioserotype. Polymorphisms appear to be higher for nonpathogenic strains of biotype 1A found in the environment than for pathogenic strains of bioserotypes 4/O:3 and 2/O:9, which are adapted to an animal host. This technique has limited potential for outbreak investigations because of the low degree of polymorphism within the most common pathogenic bioserotypes. However, it could be used to trace the global dissemination of Y. enterocolitica clones.

Blumberg (1991) studied 53 Y. enterocolitica O:3 strains and distinguished four different ribopatterns. All but two of the strains belonged to ribotype patterns I and II. Ribotype I strains corresponded to phage type 9b, and ribotype II strains to phage type 8. Phage type 9b is known as the ‘Canadian phage type’ because it was initially thought to be present only in Canada. It was later identified in the USA, and now this clone appears to be widely disseminated. Import of a large number of Canadian pigs into the USA may explain how strains of ribotype pattern I were introduced. Phage type 8, originally known as ‘European and Japanese phage type’, also appears to be widely disseminated. This type was identified almost 10 years after the first documentation of phage 9b. Although many fewer pigs have been imported to the USA from Europe than from Canada, this practice may have led to the introduction of ribotype pattern II strains. Andersen & Saunders (1990) have also found some evidence of geographical variation in the distribution of ribotypes among the strains of bioserotype 4/O:3; Canadian strains were found to be different from Danish strains. Fukushima (1998) studied 103 strains of serotype O:9 with REAP and ribotyping. They demonstrated a close correlation between genotypes (ribotypes and REAP profiles) and geographical distribution of strains of serotype O:9. Ribotype 1 has been widely detected in isolates from humans, animals and food worldwide, ribotype 2 in West European countries and the USA, ribotype 3 in Russia and ribotype 4 in Japan.

Hallanvuo (2002) used a probe containing the O-antigen-coding region, which is present in multiple copies in the genome of Y. enterocolitica. This repeated sequence, termed YeO:3RS, was found only in strains of the European pathogenic serotypes (O:1, O:2, O:3, O:5,27 and O:9) among the genus Yersinia. Using the YeO:3RS probe and the NciI and BglI restriction enzymes, all European pathogenic Y. enterocolitica isolates of bioserotype 4/O:3, 2/O:5,27 and 2/O:9 could efficiently be detected and subtyped when 112 Y. enterocolitica isolates from humans, animals and food were studied. Within bioserotype 4/O:3, YeO:3RS typing using NciI and BglI restriction enzymes was as discriminatory as pulsed-field gel electrophoresis (PFGE) typing using NotI and XbaI enzymes; the discrimination indices of the YeO:3RS and PFGE methods were 0.73 and 0.69, respectively. When these methods were combined, the two predominant PFGE types could be divided into six subtypes, thus increasing the discriminatory power to 0.85.

Golubov (2005) investigated a new Y. enterocolitica insertion sequence element, IS Yen2, which is present in multiple copies in Y. enterocolitica O:3 and O:9. The novel IS element is restricted to strains of biotypes 2–5 and is absent in the strains of biotype 1B, but also in the strains of nonpathogenic biotype 1A. Using the IS Yen2 probe and EcoRI restriction enzyme, all strains of biotypes 2–5 could be differentiated from strains of biotypes 1A and 1B. Furthermore, the number of IS Yen2 copies varied among the Y. enterocolitica O:3 strains, indicating that this method can serve as an additional tool for Y. enterocolitica differentiation and epidemiological studies.

PFGE

PFGE is a variation of agarose gel electrophoresis that permits analysis of large-sized DNA fragments of the whole bacterial genome. PFGE used with various rare-cutting restriction enzymes is considered the gold standard in subtyping of bacteria. PFGE provides a highly reproducible restriction profile that typically shows distinct, well-resolved fragments representing the entire bacterial DNA. The method has proved to be highly discriminatory and superior to other available techniques (Lukinmaa et al., 2004). With the aid of computerized gel scanning and analysis software, creating data banks of PFGE patterns is possible, which enables the development of reference databases to which any new strain can be compared in order to define its relationship to other strains. One of the factors that has limited the use of PFGE is the time involved in completing the analysis. Although the procedural steps are straightforward, the time needed to complete the procedure can be 2–3 days. This can reduce the laboratory's ability to analyse large numbers of samples.

Several studies have been conducted to characterize Y. enterocolitica with PFGE (Table 3). Iteman (1996) compared PFGE with ribotyping and REAP, and found PFGE to be the most suitable technique for epidemiological tracing of Y. enterocolitica. PFGE allows subtyping of strains belonging to the same bioserotype (Buchrieser et al., 1994; Najdenski et al., 1994; Saken et al., 1994). Najdenski (1994, 1995) showed that the pulsotype is more closely associated with the biotype than with the serotype. They also demonstrated that the degree of genomic instability is low among Y. enterocolitica strains. NotI enzyme is the most frequently used enzyme for PFGE typing of Y. enterocolitica isolates (Table 3). However, one drawback is that thus far no optimal enzyme and gel electrophoresis conditions have been found; all enzymes used in subtyping Y. enterocolitica isolates produce a large number of closely spaced restriction fragments, which can make analysis of patterns difficult.

The global homogeneity of the pulsotypes among strains of bioserotype 4/O:3 has been shown to be high (Najdenski et al., 1994; Saken et al., 1994; Iteman et al., 1996; Asplund et al., 1998). However, using NotI, ApaI and XhoI enzymes, this group can efficiently be divided into several genotypes with a discrimination index of 0.93 (Fredriksson-Ahomaa et al., 1999). Diversity of different genotypes of Y. enterocolitica 4/O:3 strains recovered from pig tonsils in southern Germany and Finland has been studied using these three restriction enzymes. All genotypes found in German strains differed from Finnish strains, indicating that genotypes of Y. enterocolitica 4/O:3 strains have a different geographical distribution (Fredriksson-Ahomaa et al., 2003).

PCR

Randomly amplified polymorphic DNA (RAPD) assay, also called arbitrary primed PCR, is a variation of the PCR technique employing a single short (typically 10 bp) primer that is not targeted to amplify any specific bacterial sequence. RAPD is a very simple and quick method, but its reproducibility can be low and standardization of the technique is difficult. However, a stable interlaboratory RAPD protocol may be obtained by careful optimization of technical procedures (Blixt et al., 2003). Some studies have been conducted to characterize Y. enterocolitica isolates with RAPD (Rasmussen et al., 1994; Odinot et al., 1995; Leal et al., 1999). This method allows discrimination between Y. enterocolitica isolates belonging to different bioserotypes and also, in some cases, between isolates belonging to the same bioserotype (Odinot et al., 1995; Leal et al., 1999). With RAPD, subtypes of pathogenic Y. enterocolitica strains correlated well with the geographical origin. Blixt (2003) divided all 27 Y. enterocolitica O:3 strains into two subclusters, representing the geographical origin of the strains. One of the subclusters contained 21 Y. enterocolitica O:3 strains originating from Sweden, Finland and Norway, while the other subcluster comprised six Danish and English O:3 strains together with O:9 and O:5,27 strains.

Repetitive element PCR (rep-PCR) uses primers corresponding to interspersed repetitive DNA elements present in various locations within the bacterial genome to generate specific genomic fingerprints. Rep-PCR with primers based on repetitive extragenic palindromic (REP) and enterobacterial repetitive intergenic consensus (ERIC) sequences has been successfully used to differentiate Y. enterocolitica strains (Sachdeva & Virdi, 2004; Wojciech et al., 2004). Sachdeva & Virdi (2004) characterized 81 Y. enterocolitica strains of biotype 1A isolated from India, Germany, France and the USA. Although both REP and ERIC fingerprinting gave comparable results, ERIC primers discriminated the strains better than REP primers; ERIC and REP primers produced 23 and 19 genotypes, respectively. The rep-PCR genotyping showed that strains having different serotypes produced identical rep-profiles, indicating a limited genetic diversity among strains of biotype 1A. Wojciech (2004) analysed 35 Y. enterocolitica strains belonging to different bioserotypes. The strains were isolated from humans, pigs and foxes in Poland. They also found that different serotypes produced identical rep-profiles. However, they demonstrated that strains belonging to the same serotype can represent different rep-profiles. REP-PCR proved to be more discriminatory than ERIC-PCR.

PCR-ribotyping (ITC-profiling) is a modification of conventional ribotyping, in which the PCR is used to amplify the spacer region between the 16S and 23S rRNA genes. This spacer region has shown a significant length and sequence polymorphism across species lines, in contrast to rRNA genes, which are highly conserved. To analyse the intergenic space, the relevant DNA is amplified by PCR with defined primers. The PCR product can further be digested with a restriction enzyme, and the resulting fragments resolved electrophoretically. Internal transcribed spacer (ITS)-profiling can also be used to subtype Y. enterocolitica strains (Wojciech et al., 2004). Its discriminatory power was shown to be similar to ERIC-PCR but lower than REP-PCR when 35 Y. enterocolitica strains of different bioserotypes were studied.

Amplified fragment length polymorphism (AFLP)

Amplified fragment length polymorphism (AFLP) is a recently adopted PCR-based typing method. The AFLP approach is based on selective PCR amplification of restriction fragments from total digest of genomic DNA. The technique involves four steps: restriction of DNA, ligation of oligonucleotide adapters, selective amplification of sets of restriction fragments and gel analysis of the amplified products. Selective amplification is achieved by the use of primers that extend beyond the restriction site into the restriction fragments, amplifying only the subset of fragments in which the primer extension matches the nucleotide flanking the restriction sites. Conventionally, AFLP is performed using a combination of a 6-base-specific and a 4-base-specific restriction enzyme. The primers are labelled with a fluorescent label on the 5′ end such that the amplified DNA fragments are labelled in the PCR reaction. The number of fragments that can be analysed simultaneously is dependent on the resolution of the detection system. Typically, 50–100 restriction fragments are amplified and detected on denaturing polyacrylamide gels. This protocol yields AFLP fragments that are 100–500 bp long. Gel electrophoresis of the fragments is carried out on a DNA sequencing machine, and interpretation of the results is performed using gene scan analysis software. The AFLP technique provides a very powerful means of DNA fingerprinting for DNA of bacterial origin.

Fearnley (2005) have developed an AFLP method to genotype Y. enterocolitica. The combination of restriction enzymes BamHI (four base specific) and BspDI (two base specific) resulted in bands ranging from 35 to 500 bp. In total, 70 strains belonging to different bioserotypes isolated from humans, pigs, sheep and cattle from the UK were studied. AFLP primarily distinguished Y. enterocolitica strains according to their biotype, with strains dividing into two distinct clusters: cluster A included biotypes 2–5 and cluster B biotypes 1A and 1B. Within these two clusters, subclusters were formed, largely on the basis of serotype. AFLP profiles also allowed differentiation of strains within these serotype-related subclusters, indicating a high discriminatory power of the technique for Y. enterocolitica. Strains of bioserotype 4/O:3 were the most clonal, with 93.2% identity found among the strains tested. Strains of bioserotype 3/O:5,27 and 3/O:9 had 92.2% and 91.8% identity, respectively.

Sequencing

Full-genome sequences have given us new opportunities to study bacteria and diseases. So far, the genome of more than 160 bacteria, including food-borne bacteria, has been partly or completely sequenced (ftp://ftp.ncbi.nih.gov/genbank/genomes/bacteria). Sequencing allows the cataloguing of all genetic variables, providing knowledge about bacterial pathogenicity and helping us to understand the origin and spread of microbial disease better.

Based on the sequencing of specific housekeeping genes, a multilocus sequence typing (MLST) method has been developed for strain typing. MLST is a relatively new approach that has been used to determine the genetic relatedness among strains of various bacterial species (Kotetishvili et al., 2005). Genetic variation is studied by nucleotide sequencing of chromosomal genes encoding enzymes and proteins of several functional types. DNA sequences of multiple housekeeping genes reveal the existence of clonal grouping even within bacterial species, such as Neisseria meningitis, Streptococcus pneumoniae and Helicobacter pylori, which have a high level of recombination (Achtman et al., 1999). Using five housekeeping genes and a gene involved in lipopolysaccharide biosynthesis, Achtman (1999) showed that some sequence polymorphisms occurred among strains between bioserotypes and within the same bioserotype when 13 Y. enterocolitica strains were studied. However, Y. enterocolitica proved to be relatively uniform, especially the bioserotypes of 2/O:5, 2/O:9 and 4/O:3. Kotetishvili (2005) identified two groups of Y. enterocolitica, one containing biotypes 1A and 2, and the other composed of biotype 1B, when four housekeeping genes were sequenced.

Short nucleotide sequences repeated in tandem, called tandem repeats or short sequence repeats, constituting microsatellite elements, are very frequent in eukaryotic organisms as well as in bacteria. In some cases, these bacterial repetitive loci are polymorphic owing to differences in the number of repeated units that they contain. These are known as variable number tandem repeats (VNTRs). This method has been used to characterize bacterial species that are genetically homogeneous and difficult to type with other methods. The VNTR method is easy to use, fast and reproducible (Lukinmaa et al., 2004). The banding patterns can also be easily analysed. Preliminary results from de Benito (2004) suggest that the analysis of VNTR in the orf528 locus may provide a new method to discriminate Y. enterocolitica 4/O:3 isolates found to be identical with other epidemiological tools.

Epidemiology of Y. enterocolitica

Indirect evidence suggests that food, particularly pork, is an important link between the pig reservoir and human infections. In case-control studies, a correlation has been demonstrated between the consumption of raw or undercooked pork and the prevalence of yersiniosis (Tauxe et al., 1987; Ostroff et al., 1994). To identify reservoirs of infections, transmission vehicles and associations between clinical cases, several DNA-based methods have been used to subtype Y. enterocolitica strains (Table 2). However, the high genetic similarity between Y. enterocolitica strains and the predominating genotypes among the strains have limited the benefit of these methods in epidemiological studies. Thus, many factors related to the epidemiology of Y. enterocolitica, such as sources and transmission routes of yersiniosis, remain obscure.

Reservoirs

Animals have long been suspected of being reservoirs for Y. enterocolitica and, hence, sources of human infections (Bottone, 1997). Numerous studies have been carried out to isolate Y. enterocolitica strains from a variety of animals. However, most of the strains isolated from animal sources differ both biochemically and serologically from strains isolated from humans with yersiniosis. Y. enterocolitica strains that belong to bioserotypes associated with human disease have frequently been isolated from tonsils and faecal samples of slaughtered pigs. In several countries, Y. enterocolitica of bioserotype 4/O:3 has been shown to be the predominant bioserotype in asymptomatic pigs. Occasionally, pathogenic Y. enterocolitica strains, mostly of bioserotype 4/O:3, have been isolated from dogs and cats (Fredriksson-Ahomaa et al., 2001b). Dogs excrete this organism in faeces for several weeks after infection. Thus, pets may be one source of human infections because of their close contact with people, especially young children. Y. enterocolitica strains of biotypes 2 and 3 and serotypes O:5,27 and O:9 have sporadically been isolated from slaughter pigs, cows, sheep and goats; however, the reservoir of these bioserotypes is not clearly established (Fukushima et al., 1993; Wojciech et al., 2004; Fearnley et al., 2005). Wild rodents and pigs have been shown to be reservoirs for Y. enterocolitica O:8 strains in Japan (Hayashidani et al., 2003). Strains of very rare bioserotypes, such as bioserotype 5/O:2,3, have been isolated from sheep, hares and goats, and bioserotype 3/O:1,2a,3 from chinchillas.

The PCR method has been applied in only a few studies investigating the prevalence of pathogenic Y. enterocolitica in pigs (Fredriksson-Ahomaa et al., 2000a; Korte et al., 2004; Bhaduri et al., 2005). The detection rate is clearly higher with PCR, especially when using real-time PCR (Boyapalle et al., 2001; Bhaduri et al., 2005). The prevalence of ail-positive Y. enterocolitica in mesenteric lymph nodes of pigs was surprisingly high (40%) when the Taqman assay was used (Boyapalle et al., 2001) (Table 4).

View this table:
Table 4

Detection of pathogenic Yersinia enterocolitica in natural samples with PCR and culture methods

SampleNo. of samplesNo. of culture-positive samples (%)No. of PCR-positive samples (%)Reference
Clinical
Pig tonsils18548(26)58(31)Fredriksson-Ahomaa (2000a)
252090(36)Boyapalle et al. (2001)
2415(63)18(75)Nesbakken et al. (2002)
Pig faeces255080(31)Boyapalle et al. (2001)
243(13)3(13)Nesbakken et al. (2002)
2793114(4)345(12)Bhaduri et al. (2005)
Mesenteric lymph nodes2570103(40)Boyapalle et al. (2001)
241(4)2(8)Nesbakken et al. (2002)
Submaxillary lymph nodes241(4)3(13)
Food
Pig tongues157(47)10(67)Vishnubhatla et al. (2001)
9979(80)82(83)Fredriksson-Ahomaa & Korkeala (2003b)
Pig offal11038(35)77(70)
Chitterlings3508(2)278(79)Boyapalle et al. (2001)
Ground pork3500133(38)
10032(32)47(47)Vishnubhatla et al. (2001)
Ground beef10023(23)31(31)
Minced pork2554(2)63(25)Fredriksson-Ahomaa & Korkeala (2003b)
Pork3006(2)50(17)Johannessen et al. (2000)
916(7)9(10)Thisted Lambertz & Danielsson-Tham et al. (2005)
Chicken4300Fredriksson-Ahomaa & Korkeala (2003b)
Fish20000
Lettuce10103(3)
Tofu5006(12)Vishnubhatla et al. (2001)
Vegetables271(4)4(15)Cocolin & Comi et al. (2005)
Environment
Water1051(1)11(10)Sandery et al. (1996)
Slaughterhouse895(6)12(13)Fredriksson-Ahomaa (2000b)
  • Pathogenicity of isolates confirmed.

  • All meat samples are raw.

  • Liver, heart, kidney.

  • Except pig tongues and offal.

Korte et al. (2005) determined the prevalence of yadA-positive Y. enterocolitica in pig tonsils in Finland using a conventional PCR method. The detection rate in fattening pigs (56%) was high compared with sows (14%), indicating that fattening pigs are an important reservoir of pathogenic Y. enterocolitica. The authors further characterized the Y. enterocolitica 4/O:3 strains isolated from sows and fattening pigs using PFGE. A total of nine and 34 NotI profiles were obtained for 26 sow strains and 122 fattening pig strains, respectively. Only three sow strains differed genotypically from pig strains, suggesting that Y. enterocolitica-positive sows could be an important reservoir in pig houses. Sows can transmit pathogenic Yersinia directly to weaning pigs or indirectly to other pigs via a contaminated environment. Similar genotypes were found in 1995 and 1999, demonstrating that the genomes of Y. enterocolitica 4/O:3 strains are very stable. Pilon (2000) studied the distribution of different genotypes of pathogenic Y. enterocolitica in pig herds in eastern Canada. Most (143/153) of the isolates belonged to serotype O:3, which is the predominant serotype in eastern provinces. The genotypic profiles obtained with PFGE using NotI enzyme were similar within a given farm over all visits. This indicates that in each herd certain strains are present, which persist on the farm for a long time. All environmental isolates, except one, has a NotI profile identical to that of an isolate recovered in pig faeces from the same farm. This suggests that the environment represents a source of contamination of pigs by Y. enterocolitica. However, because the prevalence of pathogenic Y. enterocolitica in the environment was clearly lower than in pigs, the pigs probably are the main source of pathogenic isolates on the farms.

Several studies using different typing methods have been conducted to compare human strains with animal, mostly pig, strains. Most of the reports support the hypothesis that pigs are the main source of human Y. enterocolitica infections (Andersen & Saunders, 1990; Kapperud et al., 1990; Lee et al., 1990; Fukushima et al., 1993; Fredriksson-Ahomaa et al., 2001a; Wojciech et al., 2004; Fearnley et al., 2005). Studies of REAC patterns have shown that strains of serotypes O:3 and O:9, which are generally implicated in human disease in Europe, had patterns identical to those of strains isolated from pigs (Kapperud et al., 1990). Andersen and Saunders (1990) detected no differences in the ribotypes of Y. enterocolitica strains from pigs and human patients in Denmark. Similarly, Lee (1990) found identical ribotype patterns of Y. enterocolitica O:3 in chitterlings and human specimens from the Atlanta outbreak. With three PCR-based techniques (ITC-profiling, ERIC-PCR and REP-PCR), Wojciech (2004) demonstrated that many of the human Y. enterocolitica strains of bioserotypes 4/O:3 and 2/O:9 share common genotypes with pig strains. However, the Y. enterocolitica strains recovered from humans and pigs exhibited limited heterogeneity, and two prevailing genotypes were observed. Fredriksson et al. (2001a) compared 212 Y. enterocolitica 4/O:3 strains isolated from humans with sporadic yersiniosis with 334 Y. enterocolitica 4/O:3 strains isolated from porcine origin using PFGE with NotI, ApaI and XhoI enzymes. They found 80% (170/212) of human strains to be indistinguishable from 247 strains of porcine origin in Finland. Two common genotypes obtained among human strains were also found among dog and cat strains, indicating that pet animals are also a possible source of human infection. Investigation of the relationship between strain and host AFLP profiles showed that pigs and sheep may be a potential source of human Y. enterocolitica 4/O:3 and 3/O:9 infections in Great Britain (Fearnley et al., 2005). Fukushima (1993) examined 123 strains of serotype O:5,27 using REAP and obtained three distinct restriction patterns. All human strains, most of the pig strains and some of the dog strains showed similar REAP patterns, suggesting the possibility that pigs and dogs are sources of serotype O:5,27 infections in humans in Japan.

Using PFGE, ribotyping and REAP, wild rodents and pigs have been shown to be reservoirs for Y. enterocolitica O:8 strains in Japan (Hayashidani et al., 2003). Indistinguishable genotypes were found among strains isolated from humans, wild rodents and pigs, indicating that wild rodents and pigs are possible infection sources for human Y. enterocolitica O:8 infections. Iwata (2005) recently reported the first fatal cases of Y. enterocolitica O:8 infection in monkeys. Using PFGE, ribotyping and REAP, they demonstrated that the genotypes of almost all outbreak isolates were undistinguishable from each other, indicating a common infection source. Furthermore, the strains from black rats had the same genotype as the strains from monkeys, which further supports the hypothesis that wild rodents are the infection source and an important reservoir for serotype O:8. Hosaka (1997) characterized a Y. enterocolitica O:8 strain isolated from a human with septicaemia using PFGE with XbaI enzyme and compared this strain with an earlier Y. enterocolitica O:8 strain isolated from humans with gastroenteritis. The XbaI patterns between the strains were very similar. No specific food could be determined as the source of infection for the septicaemia patient. The patient had not been exposed to small rodents, which are considered to be carriers of this pathogen in Japan. Furthermore, the patient had not travelled to the area, where serotype O:8 had been sporadically isolated. The authors therefore suggested that the bacterium persists latently in healthy humans throughout Japan; however, the true infection source remains obscure.

Contamination of food and environment

Food has often been suggested to be the main source of Y. enterocolitica infection, although pathogenic isolates have seldom been recovered from food samples. Raw pork products have been widely investigated because of the association between Y. enterocolitica and pigs. However, the isolation rates of pathogenic bioserotypes of Y. enterocolitica have been low in raw pork, except for in edible pig offal, with the most common type isolated being bioserotype 4/O:3. The low isolation rates of pathogenic Y. enterocolitica in food samples may be due to limited sensitivity of culture methods (Fredriksson-Ahomaa & Korkeala, 2003a). The occurrence of pathogenic Y. enterocolitica in some foods has been estimated by both culture and PCR methods (Table 4). In all of these studies, the prevalence was higher by PCR than by culturing.

Prevalence of yadA-positive Y. enterocolitica in food has been studied in Finland using the PCR method (Fredriksson-Ahomaa & Korkeala, 2003b). The highest detection rate was obtained from pig offal, including pig tongues (83%), livers (73%), hearts (71%) and kidneys (67%). The detection rate was clearly higher in minced meat with the PCR method than with the culture method (Table 4). Using PCR, Thisted Lambertz (2005) detected ail-positive Y. enterocolitica in 10% (9/91) of raw pork samples (loin, fillet, chop, ham and minced meat) and in one of 27 ready-to-eat pork products. Surprisingly, Vishnuhatla et al. (2001) found a high occurrence of yst-positive Y. enterocolitica in ground beef. Pathogenic strains have only sporadically been isolated from beef. In the same study, yst-positive Y. enterocolitica was also detected in tofu by real-time PCR. These PCR results indicate that the true rate of contamination of pathogenic Y. enterocolitica in pork and other processed meats and foods is underestimated using culture methods. No pathogenic Y. enterocolitica has been detected in fish and chicken samples in Finland; however, three (3%) lettuce samples were positive with PCR (Table 4). In Korea, Lee (2004) isolated one ail-positive Y. enterocolitica strain of bioserotype 3/O:3 from 673 ready-to-eat vegetables, which supports the PCR results and shows that vegetables can be a source of human infection. Furthermore, Sakai (2005) reported a food-borne outbreak of Y. enterocolitica O:8 in Japan where the same PFGE pattern was obtained from all patient and salad isolates. Recently, Y. enterocolitica 2/O:9 has been isolated from chicken eggshell surfaces in Argentina (Favier et al., 2005). Using PFGE, XbaI patterns revealed a genomic heterogeneity among the strains, which suggests different contamination sources. Contamination of the egg surface might have occurred from contact with other Y. enterocolitica-contaminated animal products, such as pork, during collection on farms or during transportation or handling in retail shops.

In a case-control study, untreated drinking water has been reported to be a risk factor for sporadic Y. enterocolitica infections in Norway (Ostroff et al., 1994). Drinking water has been relatively widely investigated and revealed to be a significant reservoir for nonpathogenic Y. enterocolitica. However, Sandery (1996) detected pathogenic Y. enterocolitica in 10% of environmental water by PCR. Falcao (2004) recently tested 67 Y. enterocolitica strains isolated in Brazil from untreated water for the presence of virulence genes by PCR. They found that all 38 strains of serotype O:5,27 possessed inv, ail and yst genes, suggesting that water may be responsible for human infection with Y. enterocolitica. In Japan, Y. enterocolitica O:8 strains have been isolated from stream water (Hayashidani et al., 2003; Iwata et al., 2005). Indistinguishable genotypes using PFGE, ribotyping and REAP were found among human and stream water strains, indicating that stream water is a possible infection source for human Y. enterocolitica O:8 infections.

The combination of ribopatterns obtained with five different restriction enzymes allowed differentiation of Y. enterocolitica O:3 strains, including 48 human strains and 24 food strains from Spain, into 11 genotypes (Mendoza et al., 1996). Raw pig minced meat and raw sausages were identified as possible sources of human yersiniosis because all Y. enterocolitica O:3 strains from the 24 food samples were grouped into the three most frequent genotypes together with human strains. Thisted Lambertz and Danielsson-Tham (2005) characterized Y. enterocolitica 4/O:3 strains recovered from patients and pork products in Sweden with PFGE using the NotI enzyme. Ten NotI profiles were revealed for 44 human strains and four profiles for six pork strains. These four pork profiles were also present among isolates recovered from the yersiniosis patients. Furthermore, Fredriksson-Ahomaa (2001a) recovered indistinguishable genotypes among 140 human and 66 retailed pork strains of Y. enterocolitica 4/O:3 using NotI, ApaI and XhoI enzymes, which further reinforces the assumption that Y. enterocolitica infections are of food origin.

Pathogenic Y. enterocolitica has frequently been detected on pig carcasses and especially on pig offal in slaughterhouses in Finland (Fredriksson-Ahomaa & Korkeala, 2003b). The major contamination source of edible offal may be the tonsils, which are frequently Yersinia-positive. The tonsils are removed along with the pluck set (tongue, oesophagus, trachea, lungs, heart, diaphragm, liver and kidneys) and then hung on a hook or placed on a conveyer belt. During this process the spread of pathogenic Yersinia from the tonsils to the pluck set is unavoidable. The most common PFGE types found in pig tonsils in Finland (Fredriksson-Ahomaa et al., 2000a) were also found on pig carcasses and offal (ear, livers, hearts and kidneys) in the slaughterhouse (Fredriksson-Ahomaa et al., 2000b), supporting the hypothesis that tonsils contaminate the offal. In another study by Fredriksson-Ahomaa (2001a), a major contamination source of human Y. enterocolitica 4/O:3 infections was revealed to be edible pig offal: 151 (71%) of the human strains were indistinguishable from the strains isolated from tongues, livers, kidneys and hearts. The same genotypes have been found at the retail level on edible pig offal and pork. These results reveal that contaminated pig offal is an important vehicle in transmission of Y. enterocolitica bioserotype 4/O:3 from the slaughterhouse to humans.

yadA-positive Y. enterocolitica has been detected by PCR in a variety of environmental sources in Finnish slaughterhouses (Fredriksson-Ahomaa et al., 2000b). The PCR method yielded more positive samples than did the traditional culture methods (Table 4), indicating that virulent Yersinia may occur at higher frequencies in the slaughterhouse environment than previously assumed. Pathogenic Y. enterocolitica was detected on the brisket saw, the hook from which the pluck set hangs, the knife used for evisceration, the aprons used by trimming workers, the computer keyboard used in the meat inspection area and the handle of the coffee maker used by slaughterhouse workers. PCR-positive samples were also obtained from the floor in the eviscerating and weighing areas and from the table in the meat-cutting area, revealing that the pig slaughterhouse environment is contaminated with pathogenic Y. enterocolitica. The Y. enterocolitica 4/O:3 isolates recovered from the brisket saw, hook and knife showed the same PFGE pattern that had dominated among carcass and offal isolates, indicating that tools and machines may spread Yersinia in the slaughterhouse. Some of the PFGE types found on the computer were also observed on livers and kidneys, which reinforces the assumption that meat inspectors spread pathogenic Y. enterocolitica from offal with their hands.

Distribution of genotypes of Y. enterocolitica 4/O:3 strains in butcher shops in Munich has been studied with PFGE using NotI, ApaI and XhoI enzymes (Fredriksson-Ahomaa et al., 2004). Twelve genotypes were obtained among 33 isolates from 14 pork and two environmental samples, demonstrating that several different strains were distributed in butcher shops. The genotypes differed among butcher shops, possibly because raw material was purchased from different sources. In most shops, more than one genotype was found, indicating that the raw material was contaminated with different strains. These results show that pathogenic Y. enterocolitica can easily be transmitted from slaughterhouses via contaminated raw material to the retail level.

Possible transmission routes

The most common transmission route of pathogenic Y. enterocolitica is thought to be faecal–oral via contaminated food. Direct person-to-person contact has not been demonstrated, but Lee (1990) reported Y. enterocolitica O:3 infections in infants who were probably exposed to infection by their carers. This may happen when basic hygiene and hand-washing habits are inadequate. Indirect person-to-person transmission has apparently occurred in several instances by transfusion of contaminated blood products (Bottone, 1999). One transmission link may be direct contact with pigs, a common risk for pig farmers and slaughterhouse workers. However, transmission of pathogenic Y. enterocolitica from pigs to humans has not yet been proven.

The main sources of human infection are assumed to be pork and pork products. Pathogenic Y. enterocolitica can be transmitted from slaughterhouses to meat processing plants and then to retail level via contaminated pig carcasses and offal (Fredriksson-Ahomaa et al., 2001a, 2004). Contaminated pork and offal are important transmission vehicles from retail shops to humans (Fredriksson-Ahomaa et al., 2001a). Cross-contamination of offal and pork will occur directly or indirectly via equipment, air and food handlers in slaughterhouses (Fredriksson-Ahomaa et al., 2000b), retail shops (Fredriksson-Ahomaa et al., 2004) and residential kitchens. The detection rate of pathogenic Y. enterocolitica in raw pork products has been shown to be high using PCR. However, consumption of raw pork would play only a limited role in the development of yersiniosis as this is not a common habit in most developed countries. Nevertheless, in Germany, raw minced pork with pepper and onion is a delicacy that can be purchased in ready-to-eat form from butcher shops. Transmission probably more often occurs via cooked pork and other food products that have been undercooked or improperly handled.

Pet animals have also been suspected as being sources of human yersiniosis because of their close contact with humans, especially young children. However, transmission from pets to humans has not yet been proven. Pathogenic Y. enterocolitica may be transmitted to humans indirectly from pork and offal via dogs and cats (Fredriksson-Ahomaa et al., 2001b). Transmission of Y. enterocolitica 4/O:3 to pets via contaminated pork has been studied using PFGE with NotI, ApaI and XhoI enzymes. A total of 132 isolates, of which 16 were from cat and dog faeces and 116 from raw pork samples, were studied in Finland. The predominant genotype recovered from pig heart, liver, kidney, tongue and ear samples was also found in the cat, whose diet consisted mostly of raw pig hearts and kidneys. The dog, which was fed raw minced pork, excreted the same genotype found in the minced meat. These results show that raw pork should not be given to pets because pathogenic isolates can easily be transmitted from highly contaminated raw pork to pets. Dogs and cats may be an important transmission link of pathogenic Y. enterocolitica between pigs and young children.

Concluding remarks

Yersinia enterocolitica is an important zoonotic pathogen that can cause yersiniosis in humans and animals. Pigs are assumed to be the main source of human yersiniosis, even though a definite connection between pathogenic Y. enterocolitica strains isolated from pigs and human infections has not been established. A close genetic relationship between pig and human strains of Y. enterocolitica has been demonstrated by several DNA-based methods. However, the high similarity between strains and the predominating genotypes within the bioserotype have limited the benefit of these methods in epidemiological studies. PFGE typing is highly effective in molecular epidemiological studies of Y. enterocolitica and is superior to other methods in discriminating between Y. enterocolitica isolates of the same bioserotype. The AFLP technique is a recently adopted typing method that utilizes fingerprinting technique to subtype Y. enterocolitica strains. This method could provide a means of discriminating Y. enterocolitica strains found to be identical with other epidemiological tools.

There are considerable difficulties associated with isolating Y. enterocolitica from clinical, food and environmental samples. Conventional culture-dependent methods have several limitations, such as low sensitivity, long incubation time, lack of identification between species, and lack of discrimination between pathogenic and nonpathogenic strains. Using PCR, pathogenic Y. enterocolitica can be detected in natural samples rapidly and with high specificity and sensitivity. Recently, several real-time PCR assays for qualitative detection of Y. enterocolitica in clinical, food and environmental samples have been developed. However, to date, the PCR method has been used in only a few studies. Prevalence of pathogenic Y. enterocolitica in pigs has been determined by PCR in some countries; however, epidemiological data about other possible animal reservoirs and from many countries are still missing.

Food has often been suggested to be the main source of yersiniosis, although pathogenic strains have seldom been isolated from food samples. Raw pork products have been widely investigated because of the association between Y. enterocolitica and pigs. However, the isolation rates of pathogenic Y. enterocolitica have been low, which may be due to the limited sensitivity of the culture methods. The prevalence of pathogenic Y. enterocolitica in pork products has been shown to be higher with PCR than with culturing, indicating that the true contamination rate of pathogenic Y. enterocolitica in pork has been underestimated. Occasionally, pathogenic Y. enterocolitica has been detected by PCR in vegetables and environmental water; thus, vegetables and untreated water are also potential sources of human yersiniosis. To identify other possible transmission vehicles, different food items should be studied more extensively by PCR.

Using genotyping, only a few animal reservoirs of Y. enterocolitica infections have been identified. The primary source of pathogenic Y. enterocolitica is fattening pigs. A close genetic relationship between pig and human strains of Y. enterocolitica has been demonstrated by several DNA-based methods. Human pathogenic Y. enterocolitica strains share common genotypes with dog strains, indicating that dogs are a possible source of human yersiniosis. In Great Britain, sheep are suspected of being a potential reservoir of human yersiniosis. Similar AFLP patterns between human and sheep strains reinforce this assumption. Wild rodents have been shown to be an important reservoir of Y. enterocolitica O:8 strains in Japan. Indistinguishable genotypes have been found among strains isolated from humans and wild rodents.

Tonsils of fattening pigs are an important contamination source in slaughterhouses. Yersinia-positive tonsils will easily contaminate the carcass, the offal and the environment during the slaughtering process. Using PFGE, Yersinia-contaminated pork and edible pig offal have proven to be important transmission vehicles of pathogenic Y. enterocolitica from the slaughterhouse to the retail level and further to humans. Indirect transmission of pathogenic Y. enterocolitica from pets to humans may occur via contaminated pork and offal. Indistinguishable genotypes have been found among strains isolated from humans and environmental water, indicating that untreated water is a possible infection source for human yersiniosis. However, many factors related to the epidemiology of Y. enterocolitica, such as sources and transmission routes, remain obscure because of the low sensitivity of culture methods and the predominating genotypes among Y. enterocolitica strains.

Footnotes

  • Editor: Willem van Leeuwen

References

View Abstract