Identification and distribution of putative virulence genes in clinical strains of Yersinia enterocolitica biovar 1A by suppression subtractive hybridization



Jugsharan S. Virdi, Microbial Pathogenicity Laboratory, Department of Microbiology, University of Delhi South Campus, Benito Juarez Road, New Delhi 110021, India. E-mail:



To detect putative virulence genes in clinical strains of Yersinia enterocolitica biovar 1A by suppression subtractive hybridization between two closely related strains of clinical and nonclinical origin having the same serotype (O:6,30–6,31).

Methods and Results

Suppression Subtractive Hybridization (SSH) was used to identify genomic differences between clinical (serotype O:6,30–6,31, from diarrhoeic human stools) and nonclinical (serotype O:6,30–6,31, from wastewater) strains of Y. enterocolitica biovar 1A. Following genomic subtraction and DNA sequencing, nine DNA sequences that were present only in clinical biovar 1A strains were identified. The sequences identified using SSH showed similarity to conserved hypothetical proteins, proteins related to iron acquisition and haemin storage, type 1 secretion proteins, flagellar hook proteins, exported protein and ABC transport system. All these sequences showed high similarity with Y. enterocolitica 8081 (biovar 1B). The distribution of these genes was further analysed using PCR in 26 clinical strains of Y. enterocolitica biovar 1A. The results revealed that the distribution of these genes was not uniform.


Genes related to iron acquisition and storage, and flagellar proteins might be responsible for virulence of some of the clinical strains of Y. enterocolitica biovar 1A.

Significance and Impact of the Study

Genes identified in this study might be useful in understanding the pathogenic potential of clinical strains of Y. enterocolitica biovar 1A.


Yersinia enterocolitica, an important food- and water-borne enteropathogen, is known to cause a variety of gastrointestinal problems including acute diarrhoea, terminal ileitis and mesenteric lymphadenitis. Long-term sequelae following infection include reactive arthritis and erythema nodosum (Bottone 1999). Yersinia enterocolitica is currently represented by six biovars (1A, 1B, 2, 3, 4 and 5) and more than thirty serotypes. On the basis of clinical evidences, Y. enterocolitica have been divided into three groups viz. strains belonging to highly pathogenic group (biovar 1B), strains with moderate pathogenicity (biovars 2–5) and the so-called nonpathogenic (biovar 1A) strains (Bottone 1999). Pathogenicity of biovars 1B and 2–5 is because of the presence of virulence genes present on pYV (plasmid for Yersinia virulence) plasmid and some chromosomal genes. Plasmid-encoded virulence factors include Yersinia adhesin A (YadA) and Ysc–Yop type III secretion system, which is responsible for injection of Yop effector proteins into host cells (Cornelis et al. 1998). The Yop proteins are responsible for a large number of virulence traits, including evasion from macrophages (Grosdent et al. 2002), immunomodulation (Boland and Cornelis 1998; Carlos et al. 2004) and cytotoxicity (Andor et al. 2001). Virulence genes that are present on chromosome include ail (attachment and invasion locus), inv (invasin), yst (Yersinia stable toxin), myfA (mucoid Yersinia factor) (Revell and Miller 2001) and genes related to iron acquisition and utilization (fepA, fepD and fes) (Schubert et al. 1999). In addition to these factors, biovar 1B strains possess iron acquisition genes on a ‘high-pathogenicity island’ which also contribute to virulence (Carniel et al. 1996; Ratledge and Dover 2000).

The biovar 1A strains have generally been regarded as nonpathogenic as these lack both pYV plasmid and the major chromosomal virulence genes (Tennant et al. 2003). However, some studies have implicated biovar 1A strains in gastroenteritis in children (Morris et al. 1991; Burnens et al. 1996). Other investigators have implicated these in nosocomial and food-borne outbreaks (Greenwood and Hooper 1990) of gastroenteritis. Recent studies showed that some biovar 1A strains, especially those of clinical and animal origin, invaded and survived in cultured epithelial cells and macrophages better than strains isolated from nonclinical samples (Grant et al. 1999; Singh and Virdi 2005; McNally et al. 2006). McNally et al. (2006) also observed that Y. enterocolitica biovar 1A strains were capable of colonizing tissues (in vitro organ culture) like other biovars of Y. enterocolitica. Previous studies in our laboratory demonstrated existence of two clonal groups among strains of biovar 1A using REP-ERIC DNA fingerprinting (Sachdeva and Virdi 2004). These studies also revealed that clinical serotype O:6,30–6,31 strains formed a discrete cluster and the aquatic serotype O:6,30–6,31 strains formed yet another tight cluster, even though these belonged to same serotype. The study suggested that clinical strains have some genetic differences that separate these from nonclinical strains. The objective of this study was to identify genomic differences between closely related clinical and nonclinical strains using suppression subtractive hybridization. SSH is a powerful tool to identify DNA sequences that are present in one strain (the tester) while absent in the other (the driver), especially in two closely related strains (Diatchenko et al. 1996; Agron et al. 2001; Tennant et al. 2005; Steele et al. 2009; Williams et al. 2010). Using SSH, genes related to virulence have been identified among the strains of different biovar of Y. enterocolitica (Iwobi et al. 2002; Golubov et al. 2003). This technique has also been used extensively to identify genomic differences between strains in different bacterial species (Calia et al. 1998; Zhang et al. 2000; Iwobi et al. 2003; Bernier and Sokol 2005; Qi et al. 2008; Dai et al. 2010).

Using this approach, we have identified genes that may be implicated to play role in the virulence of biovar 1A strains of clinical origin. Furthermore, distribution of these genes among a group of clinical isolates of Y. enterocolitica biovar 1A was also studied.

Materials and methods

Bacterial strains, plasmids and growth conditions

SSH was carried out between clinical (IP 27366, serotype O:6,30–6,31, tester strain) and nonclinical (IP26315, serotype O:6,30–6,31, driver strain) strains of Y. enterocolitica biovar 1A. The presence of the tester-specific sequences thus identified was further studied in a group of 26 clinical strains of Y. enterocolitica biovar 1A. The details of all the strains are summarized in Table 1. All clinical strains were isolated previously in our laboratory from human stools (Singh and Virdi 1999; Sinha et al. 2000; Singh et al. 2003). These strains have been authenticated by Yersinia National Reference Laboratory and WHO collaborating centre, Pasteur Institute, Paris (France). Details of other strains and plasmids used in the study are given in Table 2.

Table 1. Details of Yersinia enterocolitica biovar 1A strains used in the present study
S. no.Strain designationSerotypeCountry of originSourceClonal group
Laboratory no.Ref. laboratory no.*MTCC No.
  1. Yersinia National Reference Laboratory and WHO collaborating centre, Pasteur Institute, Paris.

  2. † Microbial Type Culture Collection (MTCC) and Gene Bank, Institute of Microbial Technology (IMTECH), Chandigarh (India).

  3. ‡ Clonal groups based on rep-PCR genotyping (Refer to Sachdeva and Virdi 2004); IP, Institut Pasteur; NAG, Nonagglutinable.

1C16IP273594861O:6,30–6,31IndiaHuman stoolsA
2C17IP273604867O:6,30–6,31IndiaHuman stoolsA
3C27IP273624869O:6,30–6,31IndiaHuman stoolsA
4C51IP273636101O:6,30–6,31IndiaHuman stoolsA
5C64IP273644862O:6,30–6,31IndiaHuman stoolsB
6C92IP273656102O:6,30–6,31IndiaHuman stoolsA
7C93IP273666103O:6,30–6,31IndiaHuman stoolsA
9C760IP274034865O:6,30IndiaHuman stoolsA
10C770IP274054844O:6,30IndiaHuman stoolsA
11C777IP274074842O:6,30IndiaHuman stoolsB
12C791IP274254847O:6,30IndiaHuman stoolsA
13C801IP274274849O:6,30IndiaHuman stoolsA
14C845IP274294851O:6,30IndiaHuman stoolsA
15C871IP274314853O:6,30IndiaHuman stoolsA
16C927IP274334854O:6,30IndiaHuman stoolsA
17C931IP274344855O:6,30IndiaHuman stoolsA
18C945IP274814856O:6,30IndiaHuman stoolsA
19C764IP274046109O:6,30IndiaHuman stoolsA
20C772IP274064845O:6,30IndiaHuman stoolsA
21C782IP274084846O:6,30IndiaHuman stoolsA
22C792IP274264848O:6,30IndiaHuman stoolsA
23C842IP274284850O:6,30IndiaHuman stoolsB
24C855IP274304852O:6,30IndiaHuman stoolsA
25C876IP274324843O:6,30IndiaHuman stoolsA
26C998IP274844859O:6,30IndiaHuman stoolsB
27C112IP273826104NAGIndiaHuman stoolsA
28C192IP273886108NAGIndiaHuman stoolsA
Table 2. Details of Escherichia coli and plasmids used in this study
Strain/plasmidRelevant characteristicsSource
Escherichia coli DH5αUsed as competent cells for transformationInvitrogen (USA)
pGEMT-easy vectorCloning vector AprPromega (USA)

The cultures of Y. enterocolitica were grown overnight at 28°C in trypticase soya broth. Escherichia coli was cultivated in Luria-Bertani (LB) medium at 37°C with shaking in the presence of ampicillin or on LB agar plates (HiMedia, Mumbai, India). Hundred microgram Ampicillin per millilitre was used, when required.

DNA isolation and restriction digestions

Genomic and plasmid DNA were extracted using the Genomic DNA Kit (Qiagen, Hilden, Germany) and plasmid mini kit (Qiagen), respectively. Restriction endonuclease digestion was accomplished by standard methods (Sambrook et al. 1989).

Genomic suppression subtractive hybridization

Bacterial genomic subtraction was carried out in accordance with the user manual of the PCR-select™ bacterial genome subtraction kit (Clontech; Takara, CA, USA). Genomic DNA of a clinical strain of Y. enterocolitica biovar 1A (IP27366, serotype O:6,30–6,31) was used as tester, whereas the DNA of an environmental (wastewater) strain of Y. enterocolitica biovar 1A (IP26315, serotype O:6,30–6,31) was used as driver.

Briefly, genomic DNA of both the tester (clinical strain) and the driver (nonclinical) strains was digested with RsaI. After digestion, the tester DNA was subdivided into two portions, and each of these was ligated with a different adaptor provided in the subtraction kit. Subsequently, two hybridizations were performed. In the first hybridization, an excess of driver DNA was added to each adaptor-ligated tester sample. The samples were then heat-denatured and allowed to anneal. In the second hybridization, the two primary hybridization samples were mixed together without denaturing. The entire population of molecules was then subjected to PCR to amplify the tester-specific sequences. Finally, nested PCR was performed to enrich the tester-specific sequences. The nested PCR-amplified fragments were cloned into pGEM®-T Easy vector (Promega, WI, USA) and transformed into competent E. coli DH5α cells. Positive clones were screened on LB medium supplemented with X-Gal (Sigma, St Louis, MO, USA), IPTG (Sigma) and ampicillin. Further, plasmid DNA was isolated from these clones. The recombinant clones were screened by colony PCR and restriction digestion of plasmid DNA.

Identification of subtractive clones containing tester-specific genomic fragments

Dot blot hybridization (Guo et al. 2006) was performed to identify the subtractive clones that might contain tester (clinical strain)-specific genomic fragments. Briefly, 50 ng of colony PCR product was spotted onto positively charged nylon membrane (Pall Life Sciences, MI, USA). The blot was cross-linked by UV at 700 Hz for 2 min to immobilize DNA on the membrane using UV cross-linker (Amersham Biosciences, Buckinghamshire, UK). The RsaI digested genomic DNA from both the tester and driver strains was used as probes labelled with Biotin-11-dUTP using Biotin DecaLabel DNA labelling Kit (Fermentas Life Sciences, MD, Mumbai, USA). The blots were prehybridized in the hybridization oven (Neolab, Mumbai, India) at 42°C for 1 h and then hybridized at 42°C for 16 h with 40 ng ml−1 of the probe. After this, the membrane was washed thrice for 20 min each in low stringency washing solution [2X- saline sodium citrate (SSC), 0·5% sodium dodecyl sulphate (SDS)] and twice for 20 min each in high stringency washing solution (0·2X- SSC, 0·5% SDS), at 30 and 65°C, respectively. The colour was developed and detected using Biotin chromogenic Detection kit (Fermentas Life Sciences). The clones showing positive results with the tester probes were selected. All the tester-specific affirmative clones were amplified using colony PCR and sequenced.

DNA sequencing and confirmation of tester-specific sequences

The amplicons obtained after colony PCR were purified using HiYield™ Gel/PCR Extraction Kit (RBC Biosciences, New Taipei City, Tiawan) according to the manufacturer's instructions and sequenced directly using appropriate primers in either one or both directions. The commercial facilities provided by M/s Link Biotech (Ocimum Biosolutions, Delhi, India) and M/s Lab India (Gurgaon, India) were used for sequencing of the samples. The sequences obtained were analysed further for homology and identity using blastn and blastx programs available at NCBI (National Center for Biotechnology Information). To further confirm the tester-specific sequences, the DNA sequences obtained above were used to design and synthesize (Sigma) tester-specific primers. PCR was carried out for both tester and driver DNA using these primers. Amplifications were carried out in My Cycler (Thermal Cycler; Bio-Rad, CA, USA) for 30 consecutive cycles consisting of denaturation at 95°C (1 min), annealing at temperature as shown in Table 3 (35 s) and extension at 72°C (90 s). Following completion of 30 cycles, final extension was carried out at 72°C for 10 min. Primers that amplified the DNA from tester strains only are shown in Table 3. The tester-specific sequences have been submitted to GenBank under the accession numbers GU253386GU253393.

Table 3. Primers designed using tester-specific sequences and used to screen a group of 26 clinical strains of Yersinia enterocolitica biovar 1A
Primer forPrimer namePrimer seq. (5′–3′)Annealing Temp. (°C)Amplicon size (bp)
Clone no. 10C10 fpCGGGGGCTGGGAGTAGT54236
Clone no. 18C18 fpCGGCGGCGCAGATAAGA54265

Detection of tester-specific sequences in other clinical strains

Using the primers listed in Table 3, the presence of tester-specific sequences was also analysed in a group of 26 clinical strains of Y. enterocolitica biovar 1A belonging to different serotype and clonal groups by PCR.


Construction of tester-specific clones library

Suppression subtractive hybridization between a clinical (IP 27366, serotypes O:6,30–6,31, Human stools) and nonclinical (IP26315, serotypes O:6,30–6,31, wastewater) strain of Y. enterocolitica biovar 1A yielded 149 clones after one round of subtraction. These were confirmed using colony PCR, and the results of some representative clones are shown in Fig. 1(a,b). Restriction digestion by RsaI revealed presence of inserts ranging from 200 to 1500 bp. Of these, 24 clones were found to carry no insert. Thus, 125 clones were confirmed as positive.

Figure 1.

(a) Colony PCR to identify subtractive clones: Lane M, 100 bp DNA ladder; Lanes 1–12 showing the colony PCR products from the subtractive clones, (b) Lane M, 100 bp DNA ladder; Lanes 1–11 showing the colony PCR products from another group of subtractive clones.

Identification of tester-specific sequences

Further, the clones carrying tester-specific genomic fragments were identified by dot blot hybridization (Fig. 2) with genomic DNA from both tester and driver strains. After dot blot hybridization, only those clones that showed twofold or more intensity with the tester DNA as compared to that with the driver DNA were selected.

Figure 2.

Dot blot hybridization to identify tester-specific clones: The DNA (PCR product) from the subtractive clones was blotted on nylon membrane and hybridized with DNA probes prepared from driver and tester strains. The clones that showed twofold or more intensity with the tester DNA compared to driver DNA were considered as tester-specific. Spots in the boxes represent the clones harbouring tester-specific genomic fragments.

Confirmation of tester-specific sequences

Finally the tester-specific sequences were confirmed using PCR with tester and driver genomic DNA. Nine tester-specific (clinical) sequences were confirmed, which were not found in driver (nonclinical) strain of Y. enterocolitica biovar 1A (serotype O:6,30–6,31). The results are shown in Fig. 3. These sequences were analysed at both nucleotide and protein level using BLAST homology search. The results showed their homology to proteins involved in iron acquisition, haemin storage, outer membrane protein PgaA, flagellar hook proteins, secretion system, transport systems and others of unknown function. These are summarized in Table 4. All these sequences showed high homology to sequences with Y. enterocolitica 8081 biovar 1B, Yersinia pseudotuebculosis, Salmonella enterica, Aeromonas spp., and Serratia spp.

Figure 3.

Confirmation of tester-specific sequences by PCR using genomic DNA of tester (T) and driver (D) strains: Lane M – 100 bp DNA ladder; Lane 1, PCR-amplified product of clone no 8 using tester strain; Lane 3, clone no. 10; Lane 5, clone no. 18; Lane 7, clone no. 51; Lane 9, clone no. 57; Lane 11, clone no. 66; Lane 13, clone no. 70; Lane 15, clone no. 87; Lane 17, clone no. 124; Lane – 2, 4, 6, 8, 10, 12, 14, 16, 18 – the corresponding amplicons were not detected when genomic DNA of driver strain (D) was used for PCR-amplification.

Table 4. Homologies of the tester-specific sequences of Yersinia enterocolitica biovar 1A identified by SSH
Clone no.Insert sizePredicted encoded proteinE-valueSimilarity (% amino acid)Genbank accession no.
Clone no. 8276Conserved hypothetical protein2e-3981/84 (96%) GU253386
Clone no. 10236Hemophore A (HasA)2e-1442/52 (80%) GU253387
Clone no. 18265Type 1 secretion protein2e-48110/123 (89%) GU253388
Clone no. 51643Outer membrane protein (PgaA)/haemin storage system (hms) protein8e-104212/218 (97%) GU253389
Clone no. 57688Flagellar hook protein1e-71206/216 (95%) GU253390
Clone no. 66402Putative 5′ nucleotidase5e-94110/113 (97%)
Clone no. 70521Restriction modification system7e-90127/140 (96%) GU253391
Clone no. 87492Putative exported protein2e-34133/166 (80%) GU253392
Clone no. 124430ABC-type multidrug transport system2e-68122/139 (87%) GU253393

Distribution of tester-specific sequences in clinical strains of Yersinia enterocolitica biovar 1A

The distribution of the tester-specific sequences was studied in a group of 26 clinical strains of Y. enterocolitica biovar 1A by PCR. The details of distribution of these genes in individual strains are given in Table 5. The sequence related to conserved hypothetical protein was present in 69% of the strains, Hemophore A (HasA) in 38%, type 1 secretion protein in 42%, Outer membrane protein (PgaA) in 35%, Flagellar hook protein in 62%, Putative 5′ nucleotidase in 38%, Restriction modification system in 58%, Putative exported protein in 23%, and ABC-type multidrug transport system was present in 35% of clinical strains of Y. enterocolitica biovar 1A. The results revealed that these sequences were distributed widely in clinical strains.

Table 5. Details of the distribution of tester-specific sequences and the genes thereof, as identified by SSH in other clinical strains (N = 26) of Yersinia enterocolitica biovar 1A
S. no.StrainsClonal groupRef no.*Conserved hypothetical protein (69%)HasA (38%)Type 1 secretion protein (42%)PgaAα (35%)Flagellar hook protein (62%)Putative 5′nucleotidase (38%)Restriction modification system (58%)Putative exported protein (23%)ABC-type multidrug transport system (35%)
  1. Yersinia National Reference Laboratory and WHO collaborating centre; Pasteur Institute, Paris; +/− presence/absence of gene.

  2. † Hemophore A, α Outer membrane protein.



Serological heterogeneity is the hallmark of Y. enterocolitica strains belonging to biovar 1A (Bottone 1999; Bhagat and Virdi 2011). Genotyping of biovar 1A strains using repetitive (REP/ERIC) genomic elements (Sachdeva and Virdi 2004), and rrn and gyrB loci (Gulati and Virdi 2007), however, indicated their limited genetic heterogeneity. These studies also revealed that clinical serotype O:6,30–6,31 strains clustered into one group, whereas the nonclinical serotype O:6,30–6,31 formed a yet another discrete cluster (Sachdeva and Virdi 2004). In the present study, suppression subtractive hybridization was used to discern genetic difference between two such closely related strains of Y. enterocolitica biovar 1A to understand the bases of the supposedly more pathogenic potential of clinical strains of Y. enterocolitica biovar 1A as compared to the nonclinical strains (Grant et al. 1999; Singh and Virdi 2005; McNally et al. 2006).

Of the nine genes identified by SSH, some might be implicated in virulence. These are as follows: hemophore A (hasA), genes related to hemin storage system (hms), type 1 secretion system and flagellar hook proteins. It is conceivable that pathogenicity of the clinical strains of biovar 1A might be related to some of these genes. Iron acquisition has been shown to be critical for the multiplication of pathogenic bacteria during the course of infection (Lillard et al. 1997; Rossi et al. 2001). HasA (hemophore) is a small extracellular heme binding protein that delivers heme to a dedicated outer membrane receptor, HasR. It has already been demonstrated that hemophore-dependent heme acquisition system plays an important role in virulence of Serratia marcescens (Letoffe et al. 1994), Pseudomonas aeruginosa (Letoffe et al. 1998), Yersinia pestis (Rossi et al. 2001) and Y. enterocolitica biovar 1B (Bracken et al. 1999). For heme or iron acquisition, bacterial hemophore is transported to the extracellular medium, via type I secretion pathway (TISS) (Letoffe et al. 1994). The TISS is one of the six major secretion systems identified in Gram-negative bacteria that transport widely diverse proteins across the cell envelope without periplasmic intermediates (Delepelaire 2004). Thus, the role of hemophore A in pathogenic potential of biovar 1A strains of Y. enterocolitica needs to be investigated.

The flagellar hook proteins have been reported to play role in motility and virulence. Mutations in genes that encode structural components of the flagella viz. flgE and fliI resulted in loss of flagella (Kim et al. 2008). Recently, in vitro studies using isogenic mutants for flagellin genes have shown that biovar 1A strains require flagella for invasion of HEp-2 cells, and persistence within human macrophages (McNally et al. 2007). An aflagellate mutant of Y. enterocolitica biovar 1A was found to be attenuated in host cell invasion (McNally et al. 2007).

Flagella are involved not only in adherence (Giron et al. 2002), but also in development of biofilms as reported in organism like Pseudomonas aeruginosa, Listeria monocytogenes, E. coli and Aeromonas spp (O'Toole and Kolter 1998; Pratt and Kolter 1998; Reisner et al. 2003; Lemon et al. 2007). Kim et al. (2008) observed that Y. enterocolitica mutants, which lacked flagella or had paralysed flagella, showed defect in biofilm formation also. It would be worthwhile to further investigate the role of flagellar hook proteins in the pathogenicity of Y. enterocolitica biovar 1A.

Using PCR, we have also detected the hms genes, that is, hmsH, hmsF, hmsR and hmsS in Y. enterocolitica biovar 1A strains (data not shown). These genes not only showed identity with haemin storage system proteins but also 90–96% similarity with biofilm PGA synthesis N-glycosyltransferase pgaABCD gene cluster of other species of Yersinia viz. Ypestis, Yintermedia, Yfrederiksenii, Ykristensenii and Ymollaretii. Recently, Itoh et al. (2008) demonstrated that pgaABCD operon is necessary for the formation of biofilms in E. coli.

We have also analysed the distribution of nine tester-specific sequences in other clinical strains of Y. enterocolitica biovar 1A. However, the presence of these sequences varies from strain to strain, even though the strains were similar serotypically. This might be due to the acquisition of the virulence-associated factors by horizontal gene transfer that is mediated by the combination of plasmids, transposons and other mobile genetic elements. Adaptive mutations and environmental factors are also involved in the modification or loss of the pre-existing genetic elements. Distribution of nine tester-specific sequences in clinical biovar 1A strains was also correlated with ERIC types (Sachdeva and Virdi 2004), which showed that these sequences were associated mainly with the ERIC-type 1 (EI).

From the foregoing discussion, it is clear that it would be worthwhile to further explore the iron acquisition and storage systems, role of flagellar hook proteins and the ability to develop biofilms to understand the potential pathogenicity of clinical strains of Y. enterocolitica biovar 1A.


This work was supported by a research grant to J.S.V. from Indian Council of Medical Research (ICMR), New Delhi and Senior Research Fellowship to P.K. from Council of Scientific and Industrial Research (CSIR), New Delhi, India.