Introduction

Streptococcus suis (S. suis) can cause meningitis, endocarditis, septicemia, polyarthritis, polyserositis, pneumonia, and even acute death in swine. Among the 35 known serotypes of this pathogen, serotype 2 is the most pathogenic, prevalent, and prone to cause human infection.

Two recent large-scale outbreaks of human infection caused by S. suis 2 showed the new clinical manifestation of streptococcal toxic shock syndrome (STSS) in Jiangsu and Sichuan, China, and the mortality of patients with STSS reached 62.7% to 81.3% [38]. Obviously, the prevention and control of S. suis 2 infection have become an urgent task.

Vaccination generally is recognized as one of the most effective means for controlling and eradicating infectious diseases [32]. Most of traditional bacterium vaccines, whether live or killed bacterial vaccines, are based on whole cells [16, 31]. This type of vaccine has the disadvantage of causing many and complex responses when injected into the body. The presence of some complicated components in whole-cell vaccine probably induces a dominant but nonprotective response and sometimes causes serious side effects. Therefore, subunit vaccine has greatly attracted the interest of researchers.

Some known virulent factors of S. suis 2 have been studied as vaccine candidates. Suilysin, or muramidase-released protein, and extracellular protein factor have been shown to induce a protective response in pigs against S. suis 2 [18, 39], but vaccines based on these molecules are clinically unavailable because the molecules are not always expressed in some virulent isolates of S. suis 2 [11, 13, 28]. A vaccine based on capsule polysaccharide is unsatisfactory because of its poor immunogenicity [9, 21]. Therefore, identification of protective S. suis 2 antigens would lay a foundation for the development of a subunit vaccine.

With the progress of microbial genomics and bioinformatics, it became possible to scan and select protective antigens from the whole genome sequence. Pizza et al. [30] reported that 570 gene products were selected as vaccine candidates from the complete genome sequence of serogroup B Neisseria meningitidis (MenB) by means of bioinformatics software. Then 350 candidates successfully expressed in Escherichia coli were purified and used to immunize mice. Finally, 25 antigens were identified that could induce a bactericidal antibody response.

Such work has laid the foundation for the development of an effective vaccine against MenB. Currently, this research strategy of reverse vaccinology has already been applied successfully to the study of vaccines for Bacillus anthracis, Chlamydia pneumoniae, Streptococcus pneumoniae, and the like [1, 26, 40]. In this study, the strategy of whole genome sequence analysis was applied to identify protective antigens against S. suis 2 infection.

Materials and Methods

Bacterial Strains

The highly pathogenic strain 05ZYH33 of S. suis 2 was isolated from an infected Chinese patient and kept in our laboratory. Strain 05ZYH33 was grown in Todd-Hewitt broth (Difco Laboratories, Detroit, Mich) for use in challenge experiments and also for extraction of the genomic DNA for polymerase chain reaction (PCR) amplification. The genome sequence of strain 05ZYH33 is available at GenBank (accession no. NC_009442).

Bioinformatics Analysis

First, the subcellular locations of all 2,194 putative proteins of strain 05ZYH33 were predicted via Cell-Ploc package (http://chou.med.harvard.edu/bioinf/Cell-PLoc/) [6], and secreted proteins were selected. Second, the amino acid sequence of each protein was analyzed with BioEdit software, and proteins containing cell wall–sorting signal motifs (LPXTG, IPXTG, NPKTG, NXZTN, LPXAG, FPXTG, LPXTN, and LPXTS) [3, 19, 23, 24] were selected. The proteins possessing the lysin motif (LysM) or choline-binding motif [2] were selected via the online server of Search Pfam at the Web site of Sanger Institute (http://pfam.sanger.ac.uk/search). Third, the amino acid sequences of protective antigens for streptococcus were collected based on the literatures. Using the bl2seq software (http://www.ncbi.nlm.nih.gov/blast/Blast.cgi), protective antigens for S. suis 2 could be derived from the similarity alignment between the collected protective antigen sequences and the whole sequence of S. suis 05ZYH33. Finally, the proteins possessing sequence similarity with virulent factors or surface antigens in the database could be obtained through a sequence homology search (http://www.ncbi.nlm.nih.gov/blast/Blast.cgi).

Expression and Purification of Interest Proteins

Target genes were amplified from the genomic DNA of 05ZYH33 by PCR with specific oligonucleotide primers (Table 1). The PCR products were cloned respectively into the BamHI/EcoRI and XhoI sites of the prokaryotic expression vector pET-30b(+) (Novagen, Madison, WI) containing two 6-histidine tags. The inserted genes were verified by DNA sequencing. Recombinant plasmids were transformed into E. coli Rosetta (Novagen) for expression of proteins of interest. The culture was incubated with agitation until the optical density at 600 nm was approximately 0.8. Then 0.5 mmol/l isopropylthiogalactoside (IPTG) was added to induce production of the fusion proteins. After 5 h of induction, five (RfeA, glutamate dehydrogenase [GDH], cell wall–associated serine proteinase [CWSP], epidermal surface antigen [ESA], surface immunogenic protein [SIP]) of eight fusion proteins were found in the bacterial periplam, and the remainder (IBP, SLY, fibrinogen-binding protein [FBP]) were located largely in the cytoplam in the form of inclusion body.

Table 1 Oligonucleotide primers used for polymerase chain reaction (PCR) amplification of candidate genesa
Full size table

The expressed proteins were purified by Ni-nitrilotriacetic acid affinity chromatography following the manufacturer’s instruction (Bio-Rad, Hercules, CA). For the extraction of soluble proteins, the E. coli cell pellets were suspended in affinity column-binding buffer (50 mmol/l NaH2PO4, 300 mmol/l NaCl, 10 mmol/l iminazole, pH 8.0), and the cells were broken by sonication. The supernatants of the E. coli lysates were loaded onto an affinity column (1 × 20 cm; Bio-Rad) equilibrated with affinity column-binding buffer. The soluble fusion proteins were eluted with 250 mmol/l iminazole in binding buffer. The purified proteins were dialyzed against 10 mmol/l Tris (pH 8.0) and sterile water to remove the iminazole. The inclusion body was extracted and dissolved as described by Sambrook and Russell [33]. The proteins were eluted with Buffer D (8 mol/l Urea, 100 mmol/l NaH2PO4·2H2O, 10 mmol/l Tris, pH 4.5) from the Ni-nitrilotriacetic acid affinity column. The eluted proteins were dialyzed to promote refolding as described in The Qiaexpressionist: A Handbook for High-Level Expression and Purification of 6xHis-Tagged Proteins (Qiagen, 1999).

The purity of all proteins was evaluated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The purified proteins were concentrated by freeze-drying and resuspended in 10 mmol/l phosphate-buffered saline (PBS). The concentration of the proteins was determined by the Bradford [4] assay.

Mouse Immunization and Challenge

The sera of mice used in the experiment were collected before immunization. Indirect enzyme-linked immunosorbent assay (ELISA) was performed to examine the existence of antigen-specific antibodies in sera. Female BALB/c mice, about 4 weeks of age, specific pathogen-free (SPF) grade (SLAC, Shanghai, China), were immunized subcutaneously in groups of 10 with 25 μg of various proteins of interest formulated in complete Freund’s adjuvant. The mice, 14 and 21 days later, were respectively given two booster immunizations in the same way with proteins formulated in incomplete Freund’s adjuvant (the mice in the control group did not receive any immunization). Then 3 days after the last booster, blood samples of the mice were drawn by vena caudalis and used to determine specific antibody titer and IgG isotypes by indirect ELISA. After 4 days, 1 ml of the highly pathogenic S. suis 2 strain 05ZYH33 diluted to fivefold 50% lethal doses (LD50 4 × 107 colony-forming units [CFU]s) in sterile Todd-Hewitt broth was injected intraperitoneally into the mice. The mice were monitored 14 days for mortality. All the animal experiments were conducted in a biosafety level 3 facility and approved by the local ethics committee.

Indirect ELISA

An indirect ELISA was used to detect total IgG antibody titer. Initially, 96-well microplates were coated overnight at 4ºC with purified recombinant proteins diluted to 1 μg/ml in carbonate buffer (15 mmol/l Na2CO3, 35 mmol/l NaHCO3 [pH 9.6]). After three washes with PBS containing 0.05% Tween-20 (PBST), the plates were blocked with PBST containing 5% nonfat dry milk for 1 h at 37ºC. Mice sera from the control and immunized groups diluted serially (twofold dilutions ranging from 1:400 to 1:3276800) in PBST were added to appropriate wells at 100 μl per well, and the plates were incubated for 1 h at 37ºC.

After three washes, the secondary antibody, peroxidase-conjugated goat antimouse IgG (Santa Cruz Biotechnology, Santa Cruz, CA), at a 1:5000 dilution, was added at 100-μl per well, and the plates were incubated for 30 min at 37ºC. Finally, the plates were incubated with O-phenylenediamine as horseradish peroxidase (HRP) substrate for visualization of the reaction. End point titers were defined as the highest dilution of serum at which the optical density at 490 nm (OD490) was 2.1-fold higher than the mean OD490 of preimmune serum. For IgG1 and IgG2a detection, mice sera from immunized groups diluted 1:2,000 were added at 100 μl per well. Peroxidase-conjugated goat antimouse IgG1 and IgG2a (Santa Cruz Biotechnology) were used as the secondary antibodies. The results were expressed as means ± standard deviations.

Statistical Analysis

The statistical significance of the differences in detection of the IgG isotypes was determined by Student’s t-test. Differences between the survival rates for the mice in the immunized and the control groups were analyzed using Fisher’s exact one-tailed test. All p values less than 0.05 were considered as significant.

Results and Discussion

In Silico Predication of Vaccine Candidates from the S. suis 2 Genome Sequence

Generally, protective antigens should be the bacterial components easily reached and recognized by immunocompetent cells [20]. Therefore, secreted proteins and surface proteins of the pathogens should be within the scope of our consideration when protective candidate molecules are predicated from the genome sequence. In addition, many previous studies have confirmed that the virulent factors of the bacteria usually are regarded as potential protective antigens [7, 18, 39].

In view of these findings, screening from the whole genome sequence focused first on the genes that encode the surface proteins, secreted proteins, and virulent-related factors. As a result, a total of 153 different molecules were selected from 2,194 predicated proteins of S. suis 2 strain 05ZYH33 (Table 2).

Table 2 List of vaccine candidates selected from the whole-genome sequence [5] of S. suis 2
Full size table

Expression of all 153 genes obviously was beyond the capability of our laboratory. We had to focus further on more important targets. Therefore, 10 candidates finally were selected from the 153 genes based on experimental evidences arising from the study of related bacteria such as Streptococcus pneumoniae, group B streptococcus, S. suis, and so on (Table 3).

Table 3 List of 10 candidate molecules selected finally
Full size table

Preparation of Recombinant Antigens

Ten selected genes encoding candidate molecules were amplified from genomic DNA of strain 05ZYH33 by PCR and subsequently ligated into the prokaryotic expression vector pET30b(+), respectively. Recombinant plasmids were transformed into Escherichia coli Rosetta for expression of target proteins. Of the 10 genes, 8 were successfully expressed in E. coli Rosetta (Fig. 1a). The remaining 2 genes, vaA and serP, were not expressed for some unknown reason.

Fig. 1
figure 1

Expression and purification of the candidate proteins. The samples were analyzed on 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) followed by Coomassie brilliant blue staining. (a) Expression of recombinant proteins in Escherichia coli Rosetta. Each lane contains total proteins from E. coli cells equivalent to 100 μl of cultures after induction with isopropylthiogalactoside (IPTG), except lane 1, in which the 3-μg protein marker was loaded. Lanes 1 to 10, respectively, contain protein marker, negative control condition, epidermal surface antigen (ESA), cell wall–associated serine proteinase (CWSP), glutamate dehydrogenase (GDH), IgG-binding protein (IBP), RTX family exoprotein A (RfeA), fibrinogen-binding protein (FBP), suilysin (SLY), and surface immunogenic protein (SIP). (b) The purified recombinant proteins. Each lane was loaded with 10 μl of the purified proteins. Lanes 1 to 9, respectively, contain protein marker (3 μg), ESA (5 μg), CWSP (10 μg), GDH (45 μg), IBP (12 μg), RfeA (0.5 μg), SLY (14 μg), FBP (10 μg), and SIP (5 μg)

Full size image

The apparent molecular weights of the recombinant proteins (ESA 66 kDa, CWSP 55 kDa, GDH 55 kDa, IBP 46 kDa, RfeA 32 kDa, FBP 72 kDa, SLY 66 kDa, SIP 23 kDa) were estimated from SDS-PAGE (Fig. 1a). The apparent molecular weights of six recombinant proteins, except for IBP and CWSP, were similar to those of the predicted ones. But the predicted molecular weights of IBP and CWSP were, respectively 36 kDa and 38 kDa, which were less than those estimated by SDS-PAGE (46 kDa and 55 kDa). The reason for the difference is unknown. However, some other proteins also have exhibited a deviation in apparent size from their theoretical molecular masses including the E. coli FtsY protein (92 vs. 54 kDa) [12] and the S. suis 2 Sao protein (110 vs. 74.8 kDa) [21]. Eight target proteins were purified by Ni-nitrilotriacetic acid affinity chromatography. The purity of prepared recombinant proteins exceeded 90% on SDS-PAGE gel after staining with Coomassie brilliant blue R250 (Fig. 1b).

Protective Evaluation of Molecules of Interest in the Mouse Infection Model

The mice used in the experiments before immunization were confirmed by indirect ELISA to have no specific antibody for the target antigens in their sera. Antibody titers of mice immunized with different immunogens are shown in Table 4, which documents that four antigens (RfeA, ESA, IBP, and CWSP) induced much higher levels of antibody response than the other four (SIP, SLY, GDH, and FBP).

Table 4 Antibody titersa of mice immunized with different immunogens
Full size table

The BALB/c mice immunized with these purified proteins were challenged with a 5LD50-dose of the highly pathogenic S.suis 2 strain 05ZYH33 1 week after the final injection. The mortality of the challenged mice was observed for 14 days. The protection of the candidate molecules against S. suis 2 was evaluated with the survival rate of the mice. As shown in Fig. 2, 9 of the 10 mice without immunization protection in the control group died within 24 h after being challenged with strain 05ZYH33, whereas the RfeA-immunized group showed the best result, with 9 mice surviving the 14th day of the observation. Also, 7 of 10 mice survived in the ESA-immunized group, and the survival rates for 6 of 10 mice in the IBP- or SLY-immunized groups were significantly higher than in the control group (p < 0.029). The remaining four groups immunized, respectively, with GDH, CWSP, SIP, and FBP showed no significant differences from the control group in terms of survival rates (p > 0.05).

Fig. 2
figure 2

Survival time of actively immunized mice after lethal challenge with the highly pathogenic strain 05ZYH33 of S. suis 2. Each group of vaccinated mice (n = 10) was challenged respectively with 5LD50 of strain 05ZYH33. Mortality was recorded daily until the 14th day. One point represents a mouse

Full size image

Assessments of IgG Isotypes

The combined results of Fig. 2 and Table 4 make it clear that the protection of immunogens RfeA, ESA, and IBP is consistent with the high antibody level of each vaccinated group. However, the SLY group shows the good protection in Fig. 2 but shows a relative low antibody level in Table 4. This suggests that cellular immunity may play an important role in SLY-induced protection. To confirm this hypothesis, IgG isotypes of RfeA-, ESA-, IBP-, and SLY-induced antibody were determined because an antigen-induced IgG subclass can reflect the pattern of immune responses [10, 27, 37].

Some studies have shown that both IgG1 and IgG2a usually were produced in sera after animals had received the stimulation of immunogen, and that if the IgG1 level was significantly higher than the IgG2a level, the Th2-like immune response or the humoral immunity would dominate, whereas if the IgG1 level was significantly lower than the IgG2a level, the Th1-like immune response or the cellular immunity would dominate [10, 15, 27, 37]. However, if IgG1 and IgG2a in sera were comparable, a combined pattern of both the Th1 and Th2 responses would dominate [35].

The determined results for our four immunogens are shown in Fig. 3, from which it can be observed that RfeA, ESA, and IBP, respectively, induced a much higher level of IgG1 than IgG2a. This suggests that these three antigens induced humoral-mediated immunity in our immunization protocol.

Fig. 3
figure 3

Analysis of serum IgG isotypes. Serum IgG isotypes were determined by enzyme-linked immunosorbent assay (ELISA). Preimmune sera and specific antisera from each group were diluted at 1:2,000. Results were expressed as means of absorbance values ± standard deviations for 10 mice

Full size image

In contrast to these three antigens, SLY stimulated IgG1 and IgG2a at a comparable level, so it should induce a combined Th1 and Th2 response. That is, cellular-mediated immunity also should play a role in anti-infection in addition to humoral immunity. This can explain why SLY provided vaccinated mice with good protection under the condition of a relatively low antibody level.

In conclusion, candidate molecules RfeA, ESA, IBP, and SLY can induce a protective response of vaccinated animals against S. suis 2, suggesting that RfeA, ESA, IBP, and SLY are potential candidates for vaccine development. Immunoprotection of SLY against S. suis 2 had been reported previously [18], but RfeA, ESA, and IBP were explored first in this investigation. Identification of new protective antigens will provide a basis for develo** a subunit vaccine of S. suis 2.