Gene polymorphism analysis of Yersinia enterocolitica outer membrane protein A and putative outer membrane protein A family protein

Background Yersinia enterocolitica outer membrane protein A (OmpA) is one of the major outer membrane proteins with high immunogenicity. We performed the polymorphism analysis for the outer membrane protein A and putative outer membrane protein A (p-ompA) family protein gene of 318 Y. enterocolitica strains. Results The data showed all the pathogenic strains and biotype 1A strains harboring ystB gene carried both ompA and p-ompA genes; parts of the biotype 1A strains not harboring ystB gene carried either ompA or p-ompA gene. In non-pathogenic strains (biotype 1A), distribution of the two genes and ystB were highly correlated, showing genetic polymorphism. The pathogenic and non-pathogenic, highly and weakly pathogenic strains were divided into different groups based on sequence analysis of two genes. Although the variations of the sequences, the translated proteins and predicted secondary or tertiary structures of OmpA and P-OmpA were similar. Conclusions OmpA and p-ompA gene were highly conserved for pathogenic Y. enterocolitica. The distributions of two genes were correlated with ystB for biotype 1A strains. The polymorphism analysis results of the two genes probably due to different bio-serotypes of the strains, and reflected the dissemination of different bio-serotype clones of Y. enterocolitica.

Background Y. enterocolitica is spread primarily through contaminated food or water [1] and causes a wide range of intestinal diseases, including enteritis, mesenteric lymphadenitis, and sepsis in some severe cases; and also cause some complications such as erythema nodosum and reactive arthritis [2]. Currently, Y. enterocolitica is divided into six biotypes (1A, 1B, and 2-5) and more than 50 serotypes [3]. The biotypes of Y. enterocolitica are divided into three groups according to the bacterial pathogenic properties: non-pathogenic biotype 1A, weakly pathogenic biotypes 2-5, and highly pathogenic biotype 1B [4]. At present, the virulent factors of Y. enterocolitica are mainly referred to type III secretion system (TTSS) encoded by 70-kb plasmid pYV [3,5], Yersinia adhesin A (YadA), the virulence genes involved inv (invasion gene), ail (attachment and invasion locus gene), yst (Yersinia stable toxin gene), myfA (mucoid Yersinia factor gene) and the pathogenic island (HPI), etc. [6]. Biotype 1A strains are traditionally considered non-pathogenic, however in recent studies have confirmed that a portion of them can cause clinical symptoms similar to pathogenic strains [7].
Bacterial outer membrane proteins primarily contain outer membrane protein A, porin C and F, being the major immunogenic proteins, and widely present in Gram-negative Enterobacteriaceae [8]. Recently, multifunction of outer membrane protein A of intestinal bacteria has been demonstrated [9][10][11][12], but seldom referred to Y. enterocolitica. In our previous study [13], OmpA was the major immunogenic protein of both highly and weakly pathogenic Y. enterocolitica incubated at different temperatures. To further identify the characteristics of ompA for Y. enterocolitica, we sequenced and analyzed the polymorphism of ompA (and p-ompA) genes of Y. enterocolitica.

Results
Distribution of ompA and p-ompA genes The ompA and p-ompA genes were detected in 318 Y. enterocolitica strains. The data showed 170 of all the pathogenic strains carried both ompA and p-ompA; 91 biotype 1A strains carried ompA, and 106 carried p-ompA ( Table 1). The ompA and p-ompA genes were both existed for biotype 1A strains carried ystB gene. However, the isolates only had one of the two genes for biotype 1A strains without ystB gene. The distribution of the two genes was associated with ystB, especially for ompA (Tables 2 and 3). The correlation analysis (P < 0.05) showed correlation coefficient (r) was 0.67 between ompA and ystB and 0.58 between p-ompA and ystB.
Polymorphism analysis of the two genes ompA: The Open Reading Frame (ORF) of ompA was 1,074 bp encoding 357 amino acids (Genbank: YP_00 1005874.1). The ompA gene of 261 strains formed 23 sequence types. The pathogenic group contained five types, Pattern A-E ( Figure 1A). 155 pathogenic strains (76 bio-serotype 3/O: 3, two 4/O: 3, 68 2/O: 9, one 4/O: 9 and eight 3/O: 9) were clustered into Pattern A, and accounted for 91.2% of all of the pathogenic strains. Therefore, pattern A was an absolute primary type in all pathogenic isolates. Other nine pathogenic strains were clustered into pattern B and C, seven (two bio-serotype 2/O: 3, four 4/O: 3 and one 3/O: 9) were clustered into Pattern C, two pathogenic bio-serotype 2/O: 5, 27 clustered into pattern B. Few nucleotide differences were found between pattern A and B ( Figure 2); while the nucleotide insertion was found compared with pattern A and C ( Figure 3). Six highly pathogenic bio-serotype 1B/O: 8 strains were clustered into Patterns D and E ( Figure 1A), and the nucleotide differences were shown with red bases compared with pattern A and B ( Figure 2).
The non-pathogenic group contained 91 biotype 1A strains, formed 18 sequence types, Pattern F-W; all the 83 strains carried the ystB gene were clustered into pattern F-T; Eight strains (except one) without ystB gene formed pattern U-W ( Figure 1A).
There were nucleotide fragment insertions and deletions for ompA gene sequences. The nucleotide fragment insertion was found for pattern C. The insertion site located at 659-664 nt of the ORF with the "-CTCCAG-" compared with pattern A ( Figure 3I), this made an alanine (A) and a proline (P) inserted at position 220 and 221 in the amino acid sequence of the OmpA. For biotype 1A strains, sequence type diversities were found compared with pathogenic strains, and the nucleotide fragment deletions were observed for pattern U to W. 15 nucleotides deletions located at 140-154 nt with "-TCTATGGCGACCAAG-", and 6 nucleotides deletions located at 409-414 nt with "-GACGCA-" of ompA were found ( Figure 3D). Eventually, this led to a change of the amino acid of the OmpA at sites 39-52, 94-98,133-146, included amino acid deletions or translation into other amino acids.
Although some predicted amino acid differences were found for ompA of Y. enterocolitica, the primary proteins and its functions were the same, even if the insertions of the pathogenic or deletions of the non-pathogenic strains occurred. For all the Y. enterocolitica possessed ompA in this study, the translated proteins were identical. The predicted secondary structure of OmpA for all the strains were almost the same, and the predicted tertiary structure of OmpA for all the isolates were similar as well. It was highly conserved for the structure and function of OmpA for Y. enterocolitica, no matter the pathogenic ability and other characteristics.
p-ompA: The Open Reading Frame (ORF) of p-ompA was 1,377 bp encoding 458 amino acids (GenBank:       Some nucleotide point mutants were found among pattern 1 to 7, as shown in Figure 4. Pattern 8-38 referred to non-pathogenic group, included 106 biotype 1A strains. All of the 84 isolates carried ystB were clustered into pattern 8 to 27; 22 strains without ystB were clustered into pattern 28 to 38 ( Figure 1B).
Although the diversity of cluster results was found for p-ompA, the primary proteins structure and function of different patterns were identical predicted by software. The predicted secondary and tertiary structure or functions of proteins for all patterns were also similar, shown the conservative characteristic.

Discussions
The surface structures of bacterial pathogens (including pilus, flagella, outer membrane proteins, and various secretion systems) are likely to interact with host tissue to regulate adhesion and invasion [14]. The outer membrane protein A belong to highly conserved protein in intestinal bacteria, and play a key role in bacterial integrity and virulence [15]. Currently, more evidence shows the pathogenicity of outer membrane protein A in a variety of pathogens [11,[16][17][18][19].
In our study, all the 170 pathogenic Y. enterocolitica strains carried the ompA and p-ompA genes. Few nucleotides changes were found for both ompA and p-ompA of the pathogenic strains. Therefore, the distributions of two genes were highly conserved for pathogenic Y. enterocolitica. The translated proteins or predicted structures of different patterns of the two genes were the same, certificated the conservative property of ompA and p-ompA for Y. enterocolitica. Several researches had been widely shown the distribution of outer membrane protein A among entero-pathogenic bacteria, and its important role in bacterial infection and immunogenicity [16][17][18][19]. However, seldom study referred to ompA or p-ompA of Y. enterocolitica, so it was the first time to perform this research. Our results showed the pathogenic strains and non-pathogenic strains were divided into different groups, and highly or weakly pathogenic strains were also distinguished based on sequence results of two genes, which reflected the different bio-serotype distributions of Y. enterocolitica. In China, serotype O: 3 and O: 9 strains were predominate pathogenic Y. enterocolitica, and most of these bacteria isolated from patients, swine and dogs. While, non-pathogenic strains referred to biotype 1A, and widely distributed among different hosts. Furthermore, no highly pathogenic 1B/O: 8 strain was isolated in China up to present, all the highly pathogenic 1B/O: 8 were foreign strains in our study. Additionally, the weakly pathogenic strains (biotype 2-4, serotype O: 3 or O: 9 strains) included wild strains from China and reference strains not from China showed no differences even if they were isolated from different origins in our study. Therefore, the cluster results for ompA or p-ompA explained the different bio-serotype distributions of Y. enterocolitica. Highly pathogenic biotype 1B strains have been shown to differ genetically from weakly pathogenic biotype 2-4 strains, and they belonged to different subtypes. The polymorphism analysis results of the two genes probably reflected the dissemination of different bio-serotype clones of Y. enterocolitica for a period of time.
Biotype 1A strains lack pYV plasmid and chromosomal virulence genes, and generally regarded as avirulent [1]. However, few studies have confirmed biotype 1A strains were related to outbreaks of nosocomial infections and foodborne diarrhea [20,21]; and some early studies found that biotype 1A strains could cause abortion in goats and cattle [22][23][24][25]. Grant et al. [26,27] showed biotype 1A strains invaded epithelial cells and resisted the killing effect of macrophage. Also biotype 1A strains were associated with the potential pathogenicity in humans [28]. Enterotoxin is an important pathogenic factor in most enteric pathogens, and the ystB gene coded for a class of thermo-resistant enterotoxin in biotype 1A Y. enterocolitica [29,30]. Virulence related gene ystB was a distinguishing marker of biotype 1A strains, presented in close to 100% of clinical isolates [31,32]. However, whether ystB gene as a virulent factor for biotype 1A of Y. enterocolitica has not been confirmed. Nakano et al. [33] found Salmonella enterotoxin (stn) regulated the OmpA membrane localization and functions, indicated the close relationship between them. In our research, ompA and p-ompA were correlated with ystB in biotype 1A strains, and formed the independent cluster patterns, the strains with ystB or without ystB for biotype 1A were also separated, which indicated the phenomenon that OmpA was linked with enterotoxins for Y. enterocolitica biotype 1A strains.

Conclusions
We showed the ompA and p-ompA genes of Y. enterocolitica were highly conserved in pathogenic strains; specially, the two genes showed a high correlation with ystB in biotype 1A strains. The pathogenic and non-pathogenic strains, highly and weakly pathogenic strains were divided into different groups based on sequence polymorphism analysis of the two genes, which reflected the different bio-serotype distributions of Y. enterocolitica.

Bacterial strains and identification of biotype and serotype
The bacterial strains used in this study were screened from the Chinese Yersinia enterocolitica library which contains nearly 4,000 strains gathered by our laboratory and derived from diarrhea patients, food, animals, and the environment. Strains were selected to cover different isolation dates, different hosts, and separated locations. We chose 150 pathogenic and 148 biotype 1A Y. enterocolitica strains isolated from China; 16 pathogenic reference strains from Europe, United States, and Japan; and four pathogenic complete-genome-sequenced strains ( Table 4). The serotypes of these strains were determined as previously described [1,[34][35][36], and the biotypes of strains were identified using the scheme reviewed by Bottone [37]. The pathogenic strains were positive for all genes (ail + , ystA + , virF + , and yadA + ); however, some pathogenic strains lost the plasmid virulence genes for virF and yadA, but still had ail and ystA genes located on the chromosome, the non-pathogenic strain was negative for all these genes.
The sample collection and detection protocols were approved by the Ethics Review Committee from the National Institute for Communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention.

Primer design
Two genes for OmpA of Y. enterocolitica reference strain 8081 (NC_008800.1) were shown when we searched the NCBI web, one was ompA, another was p-ompA. Therefore, we designed the primers of the two genes by using CloneManager software 4.0, and the primers were showed in Table 5. Primers were synthesized by Shanghai Sangon Biological Engineering & Technology and Service Co., Ltd, China.

PCR, DNA sequencing and sequence analysis
Bacteria were cultured as previously described [35]. The bacterial DNA was extracted using a Blood & Tissue Kit (QIAGEN, USA). PCR was performed in a 20 μl volume containing 10 μl PCR premix (TaKaRa, Japan), 8 μl ultrapure water, 0.5 μl of each forward and reverse primer  (25 μmol/l), and 10 ng DNA template. Thermal cycling was performed in a MJ PTC200 (Bio-Rad, USA) and the conditions were: denaturation at 94°C for 5 min, followed by 25 cycles of melting at 94°C for 25 s, annealing for 30 s at various temperatures depending on the primers used (Table 5), elongation at 72°C for 30 s, and a final extension at 72°C for 10 min. The specific PCR products were purified using a Gel Extraction Kit (QIAGEN, USA) and sequenced at TaKaRa Biotechnology (Dalian) Co., Ltd. Nucleotide sequence alignments and cluster tree construction were performed using MEGA (Version 4). The statistical tests were performed using statistical analysis software SAS version 9.2 (Statistics Analysis System). The different sequences of two genes were translated to predict amino acid by MEGA 4.0 software, and the second structure of the proteins were predicted by PredictProtein (http://www.predictprotein.org); the tertiary structure of proteins were predicted and analyzed by SWISS-MODEL (http://swissmodel.expasy.org/workspace).