Introduction

Riemerella anatipestifer infection, which causes disease primarily in domestic ducks, accounts for major economic losses to the duck industry worldwide as a result of high mortality, weight loss, and condemnation of carcasses (Ruiz and Sandhu 2013). To date, 21 R. anatipestifer serotypes have been identified (Pathanasophon et al. 2002). Of these, serotypes 1, 2, 6, and 10 have been responsible for most major outbreaks in China (Cheng et al. 2003; Hu et al. 2001). Once the bacterium infects a duck flock, it can become endemic, and eradication can be difficult, with repeated infectious episodes possible. However, little is known about the protective antigens of R. anatipestifer.

Outer membrane protein A (OmpA) is a main protein of a large array of Gram-negative bacteria. OmpA of Escherichia coli and other enterobacteria is reportedly a multifaceted protein that functions as an adhesin and invasin, acts as both an immune target and evasin, and serves as a receptor for several bacteriophages (Smith et al. 2007). In some bacteria, such as Haemophilus parasuis (Tian et al. 2011), Klebsiella pneumoniae (Kurupati et al. 2011), Salmonella enterica serovar Typhimurium (Lee et al. 2010), and Shigella flexneri 2a (Pore et al. 2011), OmpA can induce a protective immune response. In R. anatipestifer, the open reading frame (ORF) of the ompA gene is 1467-bp, so we called it ompA1467 in this study, while within this ORF, there is another 1164-bp ORF that shared the same C-terminal with ompA1467, and we called it ompA1164 (Fig. S1). Our previous results showed that OmpA of R. anatipestifer was an important virulence factor (Hu et al. 2011). In addition, OmpA is a predominant antigenic determinant of R. anatipestifer (Subramaniam et al. 2000). However, the recombinant OmpA (residues 101–488 aa) of serotype 15 R. anatipestifer strain 110/89 could not protect against subsequent challenge with the virulent serotype 15 strain, although specific antibodies were detected successfully in infected ducks (Huang et al. 2002). Nevertheless, the recombinant OmpA1467 (residues 33–466 aa) of serotype 2 R. anatipestifer strain Rf153 was reported to be able to induce protection against challenges with both homologous and heterologous strains (60% and 50% for Rf153 and serotype 1 strain Rf63, respectively) (Zhai et al. 2013). In addition, the homology of R. anatipestifer ompA1164 nucleotide sequence has been analyzed (Tsai et al. 2005; Yu et al. 2008), and the identity of the ompA1164 sequence of 15 Taiwan strains and eight reference strains and two other sequences retrieved from GenBank was 88.1–100.0% (Tsai et al. 2005). If OmpA1467 could induce protective immunity, partial cross immunity among different serotypes of R. anatipestifer strains may be found. However, very little or no cross-protection has been observed among different R. anatipestifer serotypes (Sandhu 1979). So, the results of the different studies seem to be confused and inconsistent as to whether OmpA in R. anatipestifer confers protective immunity. Therefore, the aim of the present study was to compare the protective immunity of rOmpA1467 and rOmpA1164 against virulent challenge and to evaluate the protective immunity of ompA deletion mutant.

Materials and methods

Bacterial strains and growth conditions

R. anatipestifer strains CH3 (serotype 1), WJ4 (serotype 1), Th4 (serotype 2), Yb2 (serotype 2), NJ-3 (serotype 2), HXb2 (serotype 10), YXb1 (serotype 10), and SX (serotype not determined) were isolated and identified by Qinghai Hu 10–20 years ago. R. anatipestifer type strain ATCC11845 was purchased from the American Type Culture Collection (ATCC; Manassas, VA, USA). The ompA deletion R. anatipestifer mutant Th4ΔompA was constructed previously in our lab, and 865-bp (+247 to +1111 nt) of ompA1467 ORF was deleted, and no OmpA protein was detected from Th4ΔompA by western blot (Hu et al. 2011). All R. anatipestifer strains were cultured in tryptic soy broth (TSB; Difco Laboratories, Franklin Lakes, NJ, USA) or on tryptic soy agar (TSA) at 37 °C under an atmosphere of 5% CO2. For selective growth of bacterial strains, kanamycin (50 μg/mL) or spectinomycin (60 μg/mL) was added as needed.

Sequence analysis of ompA1467 ORF from different serotypes of R. anatipestifer strains

The sequences of the ompA1467 ORF of R. anatipestifer strains Th4, YXb1, and SX were amplified by PCR using the primers ompA P1 plus ompA P2 for ompA1467 (ompA P1: 5′-GCGATTAAGGAGAGAGAAGCAA-3′; ompA P2: 5′-TTTTATCCCAACGAGCCATC-3′). The PCR products were sequenced by Shanghai Huagene Biotech Co., Ltd. (Shanghai, China). These nucleotide sequences of ompA1467 have been deposited in the GenBank database (https://www.ncbi.nlm.nih.gov/genbank/) under Accession No. MF458999-MF459001. The full genome of WJ4, HXb2, CH3, and Yb2 has been sequenced by our lab or other lab in our institute and released to GenBank database under accession No. CP041029, CP011859, CP006649, and CP007204.The ompA1467 sequences of other strains used in this study, such as ATCC11845, CH-1, CH-2, RA-GD etc., were retrieved from the GenBank database. The amino acid sequence homology of ompA1467 from R. anatipestifer strains with different serotypes was aligned and analyzed by the Clustal W method enclosed within MegAlign program of Lasergene 7.01 software (DNASTAR Inc., Madison, WI, USA).

Expression of OmpA1164 and OmpA1467 in Escherichia coli

The full ompA1467 gene of R. anatipestifer encodes 488 amino acids. The prediction results of the SignalP 4.1 server (http://www.cbs.dtu.dk/services/SignalP/) showed that the signal peptides of OmpA1467 and OmpA1164 were 21 and 23 amino acids in length, respectively. Two different lengths of ompA ORFs encoding for OmpA1467 (residues 22–488, OmpA22–488) and OmpA1164 (residues 125–488, OmpA125–488), without the signal peptides, were amplified from R. anatipestifer strain Th4 by PCR using the primer pairs ompA1467 P1 plus ompA P2′ and ompA1164 P1 plus ompA P2′ (ompA1467 P1, 5′-ATGGATCCCAGACTACTAGCAATCCTTGGTT-3′; ompA P2′, 5′-GCTACTCGAGATACTAATTATTTTCTTTTCTTTTTTACTACTTT-3′; ompA1164 P1, 5′-TAGGATCCAACGAAGATGCATGGTTTGAC-3′), respectively. Then, the sequenced ompA1164 and ompA1467 were linked to the expression vector pET30a (+) (Novagen, Inc., Madison, WI, USA) and then expressed in E. coli BL21 (DE3) cells (Novagen, Inc.). The bacteria were collected by centrifugation and lysed by sonication. The inclusion bodies were washed with washing buffer (50 mM Tris (pH 8.0), 300 mM NaCl, 1% Triton X-100, 2 mM EDTA, and 5 mM DTT), and dissolved with solubilization buffer (50 mM Tris (pH 8.0), 300 mM NaCl, 8 M Urea, 20 mM imidazole buffer), and then the recombinant proteins were purified by Ni-IDA affinity chromatography (DetaiBio, Nanjing, China), and the target proteins were eluted with different concentrations of imidazole. For rOmpA1467 refolding, the eluted proteins with high purity by SDS-PAGE analysis were dialyzed at 4 °C in refolding buffer 1 (1 × PBS (pH 7.4), 4 mM GSH, 0.4 mM GSSG, 2 mM EDTA, 0.4 M L-Arginine, 2 M Urea), while for rOmpA1164 refolding, refolding buffer 2 (50 mM Tris(pH 8.0), 150 mM NaCl, 4 mM GSH, 0.4 mM GSSG, 2 mM EDTA, 0.4 M L-Arginine, 2 M Urea) was used. After refolding, the rOmpA1467 and rOmpA1164 proteins were finally dialyzed into storage buffer 1 (1 × PBS, 10% Glycerol, pH 7.4) and storage buffer 2 (50 mM Tris (pH 8.0), 150 mM NaCl, 10% Glycerol), respectively. The purified proteins were detected by SDS-PAGE and western blotting using anti-His tag antibody (Beyotime Biotechnology, Shanghai, China).

Post-refolding analysis

Ultracentrifugation and circular dichroism were performed to analyze whether the rOmpA1467 and rOmpA1164 proteins were correctly refolded. For ultracentrifugation, the rOmpA1467 and rOmpA1164 protein solutions were centrifuged at 30,000 g for 3 h at 4 °C in an Optima L-100XP Beckman ultracentrifuge, and the concentrations of protein solution before ultracentrifugation (C1) and after ultracentrifugation (C2) were measured by BCA Protein Assay kit (Pierce, Thermo Scientific, IL, USA), and C2/C1 ratio was calculated. To examine the secondary structures of the refolded rOmpA1467 and rOmpA1164 proteins, circular dichroism analysis was performed on a Chirascan™ circular dichroism spectrometer (Applied Photophysics, UK) in a 0.5-mm path length quartz cuvette under N2 at room temperature. The UV absorption and CD spectra were recorded over the wavelength range 190–260 nm using a slit bandwidth of 1.0 nm and a scanning speed of 0.5 s per point. The CDNN CD spectra deconvolution software Version 2.1 was utilized to determine the secondary structure content of the rOmpA1467 and rOmpA1164 proteins.

Enzyme-linked immunosorbent assay (ELISA)

The antibodies from ducklings or hybridoma supernatants were measured by indirect ELISA using 96-well microplates as described previously (Prieto et al. 2003), with a coating antigen of whole R. anatipestifer cells at a concentration of 106 cells/well or soluble recombinant protein at a concentration of 1.2 μg/mL. Background noise was corrected by subtracting the absorbance of negative control wells. All samples were tested in triplicate, and the results are presented as the mean ± the standard error of the mean. In addition, the highest dilution of the serum (positive/negative ratio of ≥ 2.1) was recorded as the ELISA titer.

Immunization and challenge

One-day-old Cherry Valley ducklings, in which no serum antibodies against R. anatipestifer Th4 was detected by indirect-ELISA as described above, were purchased from Jinhu Duck Farm (Jiangyin, Jiangsu province, China).

To evaluate the protective immunity of the recombinant OmpA1467 (rOmpA1467) and rOmpA1164, 8-day-old Cherry Valley ducklings were divided into four groups (18–19 ducklings per group) and immunized once intramuscularly with 100 μg of rOmpA1164 or rOmpA1467, or 1 × 109 colony-forming units (CFU) of formaldehyde-inactivated Th4 cells, or phosphate-buffered saline (PBS), which were emulsified with ISA 70 V adjuvant (SEPPIC, Paris, France). Sera were collected 14 days after immunization, and the sera antibodies were detected using an enzyme-linked immunosorbent assay (ELISA) with rOmpA1467 or Th4 bacterial cells as the coating antigen. At post-immunization day 15, all ducklings were challenged with 3 × 109 CFU of Th4 cells. Ducklings that became moribund were killed humanely and counted as dead and then subjected to R. anatipestifer identification. The mortality of the ducklings was recorded daily for a period of 10 days after challenge.

To further evaluate the role of OmpA in the protective immunity of whole R. anatipestifer cells, 8-day-old ducklings were immunized with the inactivated ompA deletion mutant Th4ΔompA or wild-type Th4 cells and challenged with wild-type Th4. Briefly, Th4ΔompA or Th4 cells were washed with PBS, inactivated with 0.3% (v/v) formaldehyde. The ducklings were allocated to one of the three groups (16 ducklings per group) and immunized subcutaneously with 2 × 109 CFU of inactivated Th4 or Th4ΔompA cells, or 500 μL of PBS as negative control on days 5 and 18, respectively. On day 30, all ducklings were challenged with 3 × 109 CFU of the Th4 bacteria. Moribund ducklings were killed humanely. R. anatipestifer was isolated from the blood or liver tissues of all ducklings either after death or the termination of the experiment (10 days post-challenge). Duck sera before challenge were collected for the detection of the antibodies with ELISA.

Preparation of monoclonal antibodies against rOmpA1164

To prepare monoclonal antibodies against rOmpA1164, 6-week-old BALB/c mice were intraperitoneally immunized with 100 μg of rOmpA1164 with an equal volume of Freund’s complete adjuvant (Sigma-Aldrich, MO, USA), followed by booster immunization with 100 μg of rOmpA1164 with Freund’s incomplete adjuvant twice at 2-week intervals. Three days after the final booster, spleen cells from the immunized mice were fused with mouse myeloma SP2/0-Ag14 cells (ATCC, VA, USA), and the hybridoma supernatants were screened with indirect ELISA using whole Th4 cells or rOmpA1164 as coating antigen. The serum from rOmpA1164-immunized mice and serum from naïve mice were used as positive and negative control. The positive cells were subcloned by limiting dilution and further characterized by western blot. Positive hybridoma cells were subcloned two to three times by limiting dilution. Six hybridoma cell lines were then established, and culture supernatant of these hybridoma cell lines was collected. The reactivity of these monoclonal antibodies to rOmpA1164 and rOmpA1467 was detected by western blot.

Expression of OmpA1467 and OmpA1164 in R. anatipestifer cells

R. anatipestifer strains ATCC11845, CH3, WJ4, NJ3, Yb2, Th4, HXb2, and YXb1, which belong to different serotypes, were grown on TSA agar for 24 h at 37 °C under 5% CO2. The cells were harvested into sterile PBS, washed and pelleted by centrifugation, and then resuspended in PBS and adjusted to an optical density at 600 nm (OD600) of 1.0. OmpA expression of these strains was measured by SDS-PAGE and western blot using monoclonal antibody 2A4 as the first detection antibody as described above. The ompA deletion mutant Th4ΔompA, rOmpA1164, and rOmpA1467 were used as control.

Statistical analysis

The statistical significance of the data was determined using survival analyses and survival curve program or one-way ANOVA and Tukey’s multiple comparison test within GraphPad Prism 5 software (GraphPad Software, Inc., La Jolla, CA, USA). A probability (p) value of < 0.05 was considered statistically significant.

Results

OmpA1467 from some R. anatipestifer strains with different serotypes shared the same amino acid sequence

In this study, our results showed that ompA1467 of R. anatipestifer strains with different serotypes shared the same amino acid sequence. OmpA1467 from strains Th4, CH3, CH-1, WJ4, YXb1, RA-GD, ATCC11845, NCTC11014, and RCAD0133, which belong to different serotypes, shared 100% amino acid identity, and so did strains SX, CH-2, Rf153 and 153, 17 and Yb2 (Fig. 1).

Fig. 1
figure 1

OmpA1467 from some R. anatipestifer strains with different serotypes shared the same amino acid sequence. Amino acid sequence of OmpA1467 ORF from R. anatipestifer strains with different serotypes, including some retrieved from the GenBank database, was aligned and analyzed by Clustal W method enclosed within MegAlign program of Lasergene 7.01 software. A phylogenetic tree was built up based on the amino acid sequence of OmpA1467. The bar below indicates the distance between sequences. Units indicate the number of substitution events

Full size image

Expression and purification of the recombinant OmpA1467 and OmpA1164

In this study, the OmpA1467 and OmpA1164 without signal peptide were expressed as inclusion bodies in E. coli. The recombinant fusion proteins His-OmpA1467 (rOmpA1467) and His-OmpA1164 (rOmpA1164) were obtained after washing, dissolution, purification by affinity chromatography, and refolding. Only one clear band of recombinant proteins rOmpA1467 and rOmpA1164 occurred by SDS-PAGE and western blot (Fig. S2).

Refolded rOmpA1467 and rOmpA1164 are soluble and include secondary structural elements

After refolding, the rOmpA1467 and rOmpA1164 protein solutions were centrifuged, and the concentration ratios of rOmpA1467 and rOmpA1164 proteins after and before ultracentrifugation were 83.3% and 84.7%, respectively. The results indicated that 83.3% rOmpA1467 and 84.7% rOmpA1164 folded proteins were soluble, and over 80% rOmpA1467 and rOmpA1164 proteins were folded correctly. The second structures of rOmpA1467 and rOmpA1164 were further analyzed by circular dichroism. The CD of rOmpA1467 revealed 6.2% α-helices, 54.1% β-sheets, 16.3% β-turn, and 28.6% random coil, while rOmpA1164 appeared to contain 15.8% α-helices, 36.4% β-sheets, 20.6% β-turn, and 30.3% random coil (Fig. S3). The results indicated that compared to the secondary structures of rOmpA1467, rOmpA1164 contained more α-helices and β-turn but less β-sheets.

rOmpA1164, but not rOmpA1467, provided partial protective immunity against virulent Th4 challenge

As shown in Fig. 2a, 16 (84.2%) of 19 ducklings immunized with inactivated Th4 whole cells and 8 (44.4%) of 18 ducklings immunized with the soluble recombinant OmpA1164 were protected against virulent Th4 challenge. However, only 4 (21.1%) of 19 ducklings immunized with the soluble recombinant OmpA1467 survived after challenge, which was the same as group PBS. There was significant difference in protective immunity between groups rOmpA1467 and rOmpA1164 (p < 0.05). The results suggested that rOmpA1164, but not rOmpA1467, provided partial protective immunity against virulent Th4 challenge. Serum samples were collected from ducklings from each group, and the antibody titers were measured using rOmpA1467, rOmpA1164, or Th4 whole cells coated ELISA plates. As shown in Fig. 2b, when rOmpA1467 or rOmpA1164 proteins were used as the coating antigen, the serum antibodies titer from rOmpA1467-immunized ducklings was significantly higher than those from rOmpA1164 or Th4 cells-immunized ducklings (p < 0.05). When Th4 cells were used as the coating antigen, there was no significant difference in the serum antibodies titer from rOmpA1467- and rOmpA1164-immunized ducklings (p > 0.05), but the serum antibodies titer from Th4 cells-immunized ducklings was significantly higher than those from rOmpA1467- and rOmpA1164-immunized ducklings (p < 0.0001). It suggested that high antibody titer of OmpA1467 in the serum of rOmpA1467-immunized ducklings could not protect immunized ducklings from virulent challenge. The immunization-challenge experiment was performed twice, and the data from one representative experiment was shown.

Fig. 2
figure 2

The rOmpA1164, but not rOmpA1467, induced protective immunity against challenge. a Ducklings were immunized with 100 μg of rOmpA1164 or rOmpA1467, or 1 × 109 CFU of formaldehyde-inactivated Th4 cells, or PBS, which were emulsified with ISA 70 V adjuvant. Two weeks later, the ducklings were challenged with virulent Th4 cells. The mortality of the ducklings was recorded daily for a period of 10 days after challenge. Survival curves were drawn using GraphPad Prism 5 software. Asterisks indicate statistically significant differences between two groups (*p < 0.05). b The titers of sera from ducklings immunized with rOmpA1467 and rOmpA1164 was measured using ELISA coated with rOmpA1467, rOmpA1164, or whole Th4 cells, respectively

Full size image

Protection of the Th4△ompA-immunized ducklings from challenge with virulent isolates

We further compared the protective immunity of the ompA deletion mutant Th4△ompA with the wild-type Th4. As shown in Fig. 3a, all 16 (100%) ducklings immunized with the Th4△ompA bacterin were protected from challenge with virulent Th4, and most showed no clinical signs of R. anatipestifer infection, with the exception of one that displayed a very brief episode of transient depression. In the group immunized with the Th4 bacterin, 14 (87.5%) of 16 ducklings were protected from the challenge with no clinical signs of disease, while the other two showed signs of mild depression and died at 24 and 36 h, respectively, after challenge. No bacteria were isolated from the blood or liver tissues of live-immunized ducklings at 10 days post-challenge. Of the 16 ducklings in the PBS control group, 15 died from challenge by day 5. The remaining duck showed severe depression after challenge but had recovered at day 5 post-challenge. There was no significant difference in conferred protection from immunization with Th4△ompA and Th4 bacterin (p > 0.05), but the protective rate of Th4△ompA bacterin was higher than that of Th4 bacterin. Serum samples were collected from eight ducklings from each group, and the antibody titers were measured using ELISA plates coated with either whole cells of Th4△ompA or the Th4 bacterin. When Th4 cells were used as the coating antigen, there was no significant difference in the serum antibodies titer between the Th4 and Th4△ompA groups (p > 0.05, Fig. 3b). Likewise, when Th4△ompA cells were used as the coating antigen, there was no significant difference in the mean ELISA titers between the Th4 and Th4△ompA groups (p > 0.05, Fig. 3b). Interestingly, sera from the Th4△ompA group had a higher antibody tendency than that from Th4 group with the coating antigen as Th4△ompA or Th4. This was in accord with the results of immunization-challenge experiments in ducklings. It suggested that, in both Th4 and Th4△ompA-immunized ducklings, sera antibodies to proteins other than OmpA are the main components of antibodies against R. anatipestifer, and these antibodies contain neutral antibodies against R. anatipestifer infection. The immunization-challenge experiments were performed twice, which yielded similar results.

Fig. 3
figure 3

There was no significant difference in protective immunity between inactivated ompA deletion mutant Th4ΔompA or wild-type Th4 cells. a Ducklings were immunized with inactivated Th4ΔompA or wild-type Th4 cells and then challenged with wild-type Th4 2 weeks later. The mortality of the ducklings was recorded daily for a period of 10 days after challenge. Survival curves were drawn using GraphPad Prism 5 software. b The titers of sera from ducklings immunized with the Th4△ompA mutant and wild-type Th4 were measured using ELISA coated with whole Th4 or Th4△ompA cells

Full size image

Monoclonal antibodies against rOmpA1164

We generated mAbs against rOmpA1164 by traditional hybridoma method and obtained six mAbs against rOmpA1164. In this study, we focused on the property of two monoclonal antibodies 1C5 and 2A4. 1C5 and 2A4 are IgG1 subtype. The specificities of mAbs 1C5 and 2A4 were identified with rOmpA1164, rOmpA1467, total cell lysate of wild-type Th4, and ompA deletion mutant Th4△ompA by western blot using the mAbs as primary antibodies. The results showed that 1C5 and 2A4 could react with rOmpA1164, rOmpA1467, and Th4 cells, but not Th4△ompA cells. It also suggested that 1C5 and 2A4 were specific to OmpA and recognized the common epitope on OmpA1164 and OmpA1467. In addition, mAb 2A4 was better to detect OmpA1164 and OmpA1467 than 1C5 due to clean signal and less smear below the specific band in the western blot (Fig. S4). Therefore, 2A4 was chosen to detect OmpA expression in R. anatipestifer.

OmpA1467 was the main existing form of OmpA protein in R. anatipestifer cells

To determine in what form (OmpA1467 or OmpA1164) does OmpA protein exist in R. anatipestifer cells, the OmpA1467 and OmpA1164 proteins were detected in different serotypes of R. anatipestifer strains which were cultured on TSA agar by western blot analysis. As shown in Fig. 4, the results showed that all the tested R. anatipestifer with different serotypes, except ompA deletion mutant Th4△ompA, expressed OmpA1467, and OmpA1467 in these strains gave very strong band compared with that of other bands in western blot using mAb 2A4 as the primary antibodies, while other bands were very weak except CH3. It suggested OmpA1467 was the prominent form of OmpA in R. anatipestifer strains.

Fig. 4
figure 4

OmpA1467 was the main existing form of OmpA protein in different R. anatipestifer strains. R. anatipestifer with different serotypes were grown on TSA, and OmpA expression was measured by SDS-PAGE and western blot using monoclonal antibody 2A4 as the first detection antibody. rOmpA1164, rOmpA1467, and the ompA deletion mutant Th4△ompA were used as control. M. Prestained protein Marker. Lane 1. ATCC11845; Lane 2. CH3; Lane 3. WJ4; Lane 4. NJ-3; Lane 5. Yb2; Lane 6. Th4; Lane 7. HXb2; Lane 8. YXb1; Lane 9. rOmpA1164; Lane 10. rOmpA1467; Lane 11. Th4△ompA

Full size image

Discussion

OmpA is a predominant antigenic determinant of R. anatipestifer (Subramaniam et al. 2000); thus some researchers have attempted to develop a subunit vaccine against R. anatipestifer based on OmpA (Chu et al. 2015; Gao et al. 2014). However, whether or not OmpA is a protective protein in R. anatipestifer remains unclear because of the conflicting results in previous reports.

The results of the present study showed that OmpA1164, but not OmpA1467, induced protective immunity, which may be due to the differences in the three-dimensional (3-D) structures of these two proteins. The N-terminal domain of OmpA1467, but not that of OmpA1164, could form an intact eight-stranded β-barrel by SWISS-MODEL (https://swissmodel.expasy.org/, Fig. S5). It suggested that some epitopes which could induce protective immunity may be exposed to the surface of OmpA1164. In previous report, rOmpA102–488 failed to induce protective immunity, and the recombinant OmpA102–488 proteins dissolved in 6 M urea were used to immunize ducklings (Huang et al. 2002), which may lead to improper protein refolding and/or incorrect exposure of the protective epitopes. In another report, OmpA33–466 could induce protective immunity, but not OmpA1467 (OmpA22–488) in our study. This may also due to the difference of the protein structure, OmpA33–466 cannot form an intact eight-stranded β-barrel as OmpA22–488 by SWISS-MODEL.

In this study, our results showed that OmpA1467 was the prominent form of OmpA in R. anatipestifer strains, and the amount of OmpA1164 was very low or none, while OmpA1164, but not OmpA1467, could induce protective immunity. Therefore, even ompA1467 ORF or ompA1164 ORF from many different serotypes of R. anatipestifer strains shared the same amino acid sequence, only very poor or none cross-protection has also been found between different serotypes of R. anatipestifer strains (Sandhu 1979). Moreover, when Th4 cells were used as the coating antigen, there was no significant difference in the serum antibodies titer from rOmpA1467- and rOmpA1164-immunized ducklings (p > 0.05), but the serum antibodies titer from Th4 cells-immunized ducklings was significantly higher than those from rOmpA1467- and rOmpA1164-immunized ducklings (p < 0.0001). It suggested that, in both Th4 and Th4△ompA -immunized ducklings, sera antibodies to proteins other than OmpA were the main components of antibodies against R. anatipestifer, and these antibodies contained neutral antibodies against R. anatipestifer infection.

In addition, the role of OmpA on protective immunity in whole R. anatipestifer cells was also evaluated using an ompA deletion mutant. The results showed that there was no significant difference in protection between Th4△ of the Th4△ompA bacterin was higher than that of the Th4 bacterin. The following reasons may account for this phenomenon: (i) OmpA1467 is the main existing form of OmpA in R. anatipestifer strains, but it could not induce protective immunity; (ii) there were some unknown protective antigens which play important roles in protective immunity of R. anatipestifer infection; (iii) OmpA deletion leads to immune refocusing and hidden protective epitopes may be exposed (Tobin et al. 2008); and (iv) other unknown reasons. Otherwise, these results suggest that deletion of the ompA gene in a vaccine candidate may be beneficial to its protective function. In fact, this is not an isolated phenomenon. The outer membrane proteins (OMPs) of Pasteurella multocida are widely recognized as important immunogens that contribute to disease pathogenesis and protection against challenge (Basagoudanavar et al. 2006). However, the protection afforded by vaccination with P. multocida OMPs alone was adversely affected by the addition of the recombinant PmOmpA to the vaccine preparation. Nonetheless, targeted inactivation of the ompA gene in P. multocida 232 represents a potential step toward the development of an effective vaccine candidate (Dabo et al. 2008).

In summary, rOmpA1164, but not rOmpA1467, induced protective immunity. The findings of this study will help to uncover the biological characteristics of OmpA and to develop novel vaccines against R. anatipestifer infection.