|Home | About | Journals | Submit | Contact Us | Français|
The molecular diversity of the gene encoding the outer membrane protein A (OmpA) of Haemophilus parasuis has been unclear. In this study, the structural characteristics, sequence types, and genetic diversity of ompA were investigated in 15 H. parasuis reference strains of different serovars and 20 field isolates. Three nucleotide lengths of the complete open reading frame (ORF) of ompA were found: 1098 base pairs (bp), 1104 bp, and 1110 bp. The OmpA contained 4 hypervariable domains, mainly encoding the 4 putative surface-exposed loops, which makes it a potential molecular marker for genotyping. Western blot analysis showed that the recombinant OmpAs of serovars 4 and 5 could cross-react with antiserum to all 15 serovars. Hence, although ompA of H. parasuis exhibited high variation among serovars, this variation did not seem to affect the strong antigenic characteristics of OmpA.
La diversité moléculaire des gènes codant la protéine A de la membrane externe (OmpA) d’Haemophilus parasuis demeure confuse. Dans la présente étude, les caractéristiques structurales, les types de séquence, et la diversité génétique d’ompA ont été étudiés pour 15 souches de référence d’H. parasuis de sérovars différents et 20 isolats de champs. Trois longueurs de nucléotides du cadre de lecture ouvert (ORF) complet d’ompA ont été trouvées : 1098 paires de bases (bp), 1104 bp et 1110 bp. L’OmpA possédait 4 domaines hypervariables encodant les 4 boucles exposées en surface putatives, ce qui en fait un marqueur de virulence potentiel pour le génotypage. L’analyse par immunobuvardage a montré que les OmpAs recombinantes des sérovars 4 et 5 pouvaient avoir des réactions croisées avec des antisérums produits contre les 15 sérovars. Ainsi, bien qu’une grande variation soit démontrée pour ompA chez les différents sérovars d’H. parasuis, cette variation n’a pas semblé affectée les caractéristiques antigéniques d’OmpA.
(Traduit par Docteur Serge Messier)
Haemophilus parasuis, a member of the family Pasteurellacea, is the causative agent of Glässer’s disease in swine, which is characterized by fibrinous to fibrinopurulent polyserositis and polyarthritis (1). The disease has produced large losses in swine populations worldwide. However, the strains of H. parasuis are genetically diverse and phenotypically variable (2). Because of the serovar diversity and the lack of cross-reaction among the serovars, it has been difficult to develop an effective serodiagnostic method (3). The heterogeneity of strains and serovar diversity have highlighted the need for practical molecular typing methods and for seeking a common antigenic component among the serovars as the basis of a universal serodiagnostic tool.
Outer membrane protein A (OmpA) is an integral component of the outer membrane of gram-negative bacteria and is highly conserved (4). Functions that have been attributed to OmpA include participating in biofilm formation, acting as both an immune target and in immune evasion, serving as a receptor for several bacteriophages (5), and playing a part in bacterial adherence to host tissues (6–8). In addition, ompA, the gene encoding OmpA, is considered an important genetic marker in bacteria (4,9,10). The OmpA of H. parasuis was identified by Tadjine et al (11). In our previous study, the ompA genes of H. parasuis serovars 4 and 5 were cloned on the basis of the reported N-terminal sequence (12), which was named ompP5 in the H. parasuis genome (GenBank accession no. ABKM01000058). The recombinant OmpAs of serovars 4 and 5 cross-reacted with each other’s antiserum. The molecular diversity of the ompA of other H. parasuis serovars has been unclear.
In this study the structural characteristics, sequence types (STs), and genetic diversity of the ompA gene of 35 H. parasuis strains — 15 reference strains representing the 15 serovars and 20 field isolates — were investigated. The immunoreactivity of recombinant OmpA from serovars 4 and 5 was further investigated by western blotting with antiserum against the reference strains.
The 15 reference strains of H. parasuis were kindly supplied by the Queensland Animal Research Institute of Australia. The 20 field isolates were recovered from swine in China with Glässer’s disease and meningitis, arthritis, or pneumonia; they were of serovars 5 (n = 9), 13 (n = 4), and 14 (n = 3) or were nontypable (n = 4) according to the indirect hemagglutination test (13). The bacteria were cultivated on trypticase soy agar (Tianhe Microorganism Reagent Company, Hangzhou, China) supplemented with 0.02% nicotinamide adenine dinucleotide (Sigma Chemical Company, St. Louis, Missouri, USA) and 5% inactivated bovine serum at 37°C in a 5% CO2-enriched atmosphere, and DNA was prepared from them previously described (2).
To obtain the correct and complete open reading frame (ORF) of ompA, we designed a pair of primers (forward, 5′-GCATTCTTGCCTCGTTCTTT-3′; reverse, 5′-CCGGTGAAGAAATAGATGGG-3′) according to fragment 297550 of H. parasuis (GenBank accession no. ABKM 01000058) to amplify a fragment of about 1200 base pairs (bp) by polymerase chain reaction (PCR). The PCR products were purified and then cloned into the pMD-18T vector (TaKaRa Biotechnology Company, Dalian, China) for sequencing. The complete ORF and its structural characteristics were further analyzed.
The 15 reference strains and 20 field isolates yielded ORFs of 3 lengths: 1098 bp, 1104 bp, and 1110 bp. The GenBank accession numbers were FJ667983 to FJ668015 for the ompA sequence of 13 reference strains and the 20 field isolates and EU846096 and EU846097 for the ompA sequence of reference serovars 4 and 5, respectively. Reference serovars 1, 2, 7, 10, 11, 12, 13, and 15, as well as 3 of the field isolates, had ORFs of 1098 nucleotides; serovars 3, 6, 8, and 14, as well as 17 of the field isolates, had ORFs of 1104 nucleotides; and serovar 9 had an ORF of 1110 nucleotides. The ompA gene of serovars 4 and 5 had an ORF of 1104 nucleotides, as reported previously (12). The predicted proteins encoded by ompA were 365, 367, and 369 amino acids in length, respectively, and varied from 39.12 to 39.55 kDa in molecular mass. Since these proteins contained a putative signal sequence of 21 amino acids, the predicted molecular mass of the putative mature proteins varied from 37.12 to 37.51 kDa.
The ompA genes were aligned by means of MegAlign (DNAStar, Madison, Wisconsin, USA). The amino acids deduced from the ompA genes were aligned, and the numbers of sites containing variable amino acids were calculated with MEGA 4.1 software (www.megasoftware.net). Gene alignment showed 86.3% to 100% similarity among the 35 H. parasuis strains, with 138 polymorphic nucleotide sites (12.6%), whereas the predicted amino acid alignment suggested that there would be 48 variable sites (13.2%).
Secondary structure prediction was performed by the Psipred secondary structure prediction method (http://bioinf.cs.ucl.ac.uk/psipred/psiform.html). The location of the polymorphic nucleotide sites and the variable inferred amino acid positions were strikingly nonrandom in distribution. Most of the polymorphic sites occurred within 4 hypervariable regions, named HV1 to HV4, which encode the 4 predicted surface-exposed loops of OmpA (Table I). Within the hypervariable domains, 93 (29%) of the 318 sites were polymorphic and 37 (35%) of the 106 inferred amino acid positions were variable. In contrast, the remainder of ompA was relatively conserved: only 45 (6%) of the 770 sites were polymorphic and 11 (4%) of the 255 inferred amino acid positions were variable. Domains HV3 and HV4 were characterized by amino acid deletions or insertions, or both, which accounted for the variation in molecular mass of the putative mature protein. The hypervariable domains corresponded to the surface-exposed loops of OmpA of Bibersteinia trehalosi, Pasteurella multocida, and Haemophilus influenzae, which also belong to the family Pasteurellacea (4,8,9). However, the variation was greater for B. trehalosi than for H. parasuis, whereas ompA of H. parasuis and B. trehalosi were highly homologous (12).
The genes encoding surface-exposed bacterial structural proteins or virulent factors may experience high selection pressure from interactions with host immune systems or environmental conditions and undergo more frequent genetic variation than do housekeeping genes. Therefore, the surface protein-encoding genes are useful genetic elements for differentiating isolates in epidemiologic investigations, which ultimately benefits clinical diagnosis and disease control (10). In our study, the levels of variation in the ompA sequence of H. parasuis differed among the strains, with unbalanced rates of variation between the putative surface-exposed loops and the relatively conserved regions. During a study of 16S rRNA genes in 35 strains we noticed that the overall variation rate for H. parasuis ompA was higher than the average overall variation rate for 16S rRNA genes: 12.6% and 5.6%, respectively (unpublished data). Hence, ompA would provide more resolution at the genetic level then the 16S rRNA gene sequences.
The ST method has been widely used for epidemiologic investigation in many bacterial species (10,14). By taking every different sequence as an ST, even if just 1 nucleotide was different, as with the hsp60 ST and 16S rRNA ST of H. parasuis (14), we identified 17 different ompA STs among the 35 H. parasuis strains (Figure 1: consecutive lettering from A to Q). Interestingly, 1 of the 9 serovar-5 field isolates and 1 of the 3 serovar-14 field isolates had STs different from the other field isolates of the same serovar, which indicates that determining STs may provide more information for classification than serotyping of H. parasuis isolates. In addition, the serotyping results seemed to be associated with the results of ompA genotyping in the isolates with certain differences. For example, the 9 serovar-5 field isolates contained ST A (8 isolates) and ST D (1 isolate), the 4 serovar-13 field isolates contained ST E, and the 3 serovar-14 field isolates contained ST B (2 isolate s) and ST C (1 isolate). Moreover, 8 serovar-5 field isolates and the 4 nontypable field isolates contained ST A, the predominant ST in this study. However, these results could be biased owing to insufficient sample size. More isolates from different geographic areas will be required to investigate whether the ompA ST and the serovar of H. parasuis are associated.
From the neighbor-joining dendrogram of ompA for the 35 H. parasuis strains, 4 monophyletic clusters were defined (Figure 1). No obvious association with virulence was observed. For example, cluster 4 included 1 virulent reference strain (serovar 1), 1 strain of intermediate virulence (serovar 2), and 2 strains generally believed to be avirulent (serovars 7 and 11). Similarly, no obvious correlation was observed between serovars or STs among the 15 reference strains or the 20 field isolates. For example, cluster 1A contained the serovar-5 reference strain Nagasaki (ST A), 9 serovar-5 field isolates (ST A, ST D), and 4 nontypable field isolates (ST A). The 4 serovar-13 field isolates (ST E), designated NO634, NO635, NO648, and NO649, were included in cluster 1B. In contrast with previous cases, they showed a close genetic relationship with the serovar 3 and 4 reference strains (both ST H). Similarly, in cluster 4 the 3 serovar-14 field isolates, designated NO125 (ST B), NO127 (ST B), and NO132 (ST D), were considered to have a close genetic relationship with the reference strains of serovar 1 (ST F) and 11 (ST M).
Recombinant OmpAs of serovars 4 and 5 had been produced in our laboratory, as described previously (12), and we had found that they could cross-react with each other’s antiserum. To test their immunoreactivity against the other serovars of H. parasuis, we separated the proteins from whole cells by sodium dodecyl sulfate polyacrylamide gel electrophoresis, transferred them to a polyvinylidene fluoride membrane (Millipore Corporation, Bedford, Massachusetts, USA), and incubated them with antiserum prepared from whole bacteria belonging to serovars 1 to 3 and 6 to 15 of H. parasuis. Western blotting showed that the recombinant OmpAs cross-reacted with antiserum against the other reference strains (Figure 2), which suggests that the recombinant proteins had good immunoreactivity.
Owing to serovar diversity and lack of cross-reaction among the serovars, it has been very difficult to develop a universal serodiagnostic tool for H. parasuis infection (3). Our results revealed that the same antigenic epitope exists in the OmpAs from strains of different serovars. The cross-reactivity of recombinant H. parasuis OmpAs with other bacterial species should be investigated to decide whether recombinant OmpA is a good antigen candidate for a universal tool to diagnose infection by all serovars of H. parasuis.
In conclusion, the H. parasuis ompA gene presents high variation and a specific molecular structure. It contains 4 hypervariable domains, which encode mainly the 4 putative surface-exposed loops of OmpA. The ompA ST is a potential tool for genotyping H. parasuis. Although ompA exhibited high variation among different serovars, this variation did not seem to affect the strong antigenic characteristics of OmpA.
This work was funded through 2 programs of the Ministry of Science & Technology of China: the National 11th Five-year Plan Scientific and Technical Supporting Program (contract 2006BAD06A11) and Standardization and Sharing of Veterinary Microorganism Resources (contract 2005DKA21205-11).