Elongation Factor Tu and Heat Shock Protein 70 Are Membrane-Associated Proteins from Mycoplasma ovipneumoniae Capable of Inducing Strong Immune Response in Mice

Chronic non-progressive pneumonia, a disease that has become a worldwide epidemic has caused considerable loss to sheep industry. Mycoplasma ovipneumoniae (M. ovipneumoniae) is the causative agent of interstitial pneumonia in sheep, goat and bighorn. We here have identified by immunogold and immunoblotting that elongation factor Tu (EF-Tu) and heat shock protein 70 (HSP 70) are membrane-associated proteins on M. ovipneumonaiea. We have evaluated the humoral and cellular immune responses in vivo by immunizing BALB/c mice with both purified recombinant proteins rEF-Tu and rHSP70. The sera of both rEF-Tu and rHSP70 treated BALB/c mice demonstrated increased levels of IgG, IFN-γ, TNF-α, IL-12(p70), IL-4, IL-5 and IL-6. In addition, ELISPOT assay showed significant increase in IFN-γ+ secreting lymphocytes in the rHSP70 group when compared to other groups. Collectively our study reveals that rHSP70 induces a significantly better cellular immune response in mice, and may act as a Th1 cytokine-like adjuvant in immune response induction. Finally, growth inhibition test (GIT) of M. ovipneumoniae strain Y98 showed that sera from rHSP70 or rEF-Tu-immunized mice inhibited in vitro growth of M. ovipneumoniae. Our data strongly suggest that EF-Tu and HSP70 of M. ovipneumoniae are membrane-associated proteins capable of inducing antibody production, and cytokine secretion. Therefore, these two proteins may be potential candidates for vaccine development against M. ovipneumoniae infection in sheep.


Introduction
Mycoplasma ovipneumoniae is the causative agent of chronic non-progressive pneumonia in sheep, goat, bighorn and wild small ruminants [1][2][3][4]. It has been a worldwide epidemic that has caused huge economic loss to the sheep industry for decades [1,3,5]. Progressive wasting, spasmodic cough, diarrhea and anemia are some of the characteristic symptoms of the disease [6]. Previous investigations led to conclusions that M. ovipneumoniae was primarily an epizootic pathogen, which could cause secondary infection in fatal respiratory diseases [7][8][9]. On the other hand, multiple strains of M. ovipneumoniae have been isolated from infected flock in different countries [2,6,10,11]. Gene polymorphism studies have shown differences in immunogenicity and toxicity among various M. ovipneumoniae strains. These disparities have only made the research harder, and has slowed down the progress of vaccine development [2].
M. ovipneumoniae is closely related to M. hyopneumoniae based on partial sequences of the 16s rRNA gene and HSP70 (DnaK) gene [12,13]. Genomic sequencing of M. ovipneumoniae strain SC01 and NM2010 has led to a better understanding of the role of virulent genes in mycoplasma's pathogenesis [14,15]. The data showed M. ovipneumoniae has some adhesionlike protein homologues that are closely related to M. hyopneumoniea, which include a M. hyopneumoniae P97-like protein. P97 from M. hyopneumoinae has been reported to be an outer membrane-associated protein and recombinant P97 protein has been shown to elicit immune response in pigs [16][17][18]. Shahzad et al. suggested that surface-exposed membrane proteins could suppress the activity of M. ovipneumoniae [19]. Lack of cell wall may further facilitate mycoplasma's membrane and cytoskeleton proteins to actively participate during infection of host cells. So here we postulate that the membrane-associated proteins of M. ovipneumoniae, especially the membrane adhesion-like protein, maybe a potential candidate to design a vaccine against M. ovipneumoinae.
Elongation factor-Tu (EF-Tu) is one of the most abundant and conserved bacterial proteins. It is involved in the protein synthesis and translation process in prokaryotic cells [20]. EF-Tu is a component of the bacterial membrane cytoskeleton involved in adhesion to the host cells in several pathogenic bacteria [21]. In M. pneumonia, EF-Tu is surface-localized and may mediate the binding of M. pneumoniae to fibronectin [22]. Recent studies have demonstrated that EF-Tu was also a bacterial virulence factor and could be an immunodominant protein [23][24][25][26].
Heat shock protein (HSP70), a molecular chaperone is one of the most conserved proteins [27]. Recent studies demonstrated that HSP70 could induce both innate and acquired immune responses, and also act as Th1 cytokine-like adjuvant in immune responses [28,29]. In M. pneumonia, HSP70 is suggested to play an important role in the organization of the terminal organelle, critical for adhesion to host cells [30].
In this study, we demonstrated that EF-Tu (44 kDa) and HSP70 (66.4 kDa) proteins were M. ovipneumoniae membrane-associated proteins. The immunogenicity of recombinant EF-Tu (rEF-Tu, a 44.8 kDa protein from a M. ovipneumoniae wild strain Mo-1) and recombinant HSP70 (rHSP70, a 67.1 kDa protein from a M. ovipneumoniae wild strain Mo-1) were evaluated in BALB/c mice. Moreover, growth inhibition test (GIT) results indicated that growth of M. ovipneumoniae was inhibited by the specific antisera in vitro.
Animals. All animal studies were performed in accordance with the China Agricultural University Institutional Animal Care and Use Committee guidelines (ID: SKLAB-B-2010-003) and approved by animal welfare committee of China Agricultural University.
Medium for M. ovipneumoniae was prepared with Tryptic Soy Broth (30 g/L), Lactose (10 g/L), Yeast extract (100 mL/L), Penicillin (200 U/mL), Porcine serum 20% (w/v) and 0.4% Phenol Red (w/v), and pH adjusted to 7.4-7.6 with 1 mM NaOH. This medium is a modified version of a medium described previously for culturing the M. ovipneumoinae [31]. The agar medium of M. ovipneumoniae contained 1.5% (w/v) agar. M. ovipneumoniae was cultivated at 37°C with 5% CO 2 , harvested at a mid-log phase by centrifugation at 12,000×g for 20 min at 4°C, washed three times with 0.01 M PBS and stored at -80°C.

Expression and purification of EF-Tu and HSP 70 proteins
For cloning experiments, EF-Tu and HSP70 specific primers were designed by Primer Premier 5.0 as per HSP70 (DnaK, GenBank No. HM047293.1) of M. ovipneumoniae strain Y98 and EF-Tu (Tuf, GenBank No. JQ990999.1) of M. ovipneumoniae strain GH3-3. Two restriction sites, Nco I and Xho I, were introduced at both ends of target genes in the flanking primers. These two genes of M. ovipneumoniae wild strain Mo-1 were amplified by PCR and the amplified products were sequenced. Sequence analysis showed both EF-Tu and HSP70 genes carried the UGA codon. UGA is a termination codon in Escherichia coli expression system but codes for tryptophan in Mycoplasma. Therefore, an overlap extension PCR was used to mutate the A of UGA into G in the gene sequence to avoid any truncated gene products. All primers used in this study are shown in the Table 1. The full-length PCR products of mutated target genes were digested by the restriction enzymes Nco I and Xho I. The fragments were then cloned into the PET-28a (+) prokaryotic expression vector to construct the pET-28a (+)-EF-Tu and pET-28a (+)-HSP70 plasmids. Rosetta (DE3) chemically competent cell (TaKaRa, Dalian, China) was used to express the recombinant protein. The competent cells carrying pET-28a (+)-EF-Tu or pET-28a (+)-HSP70 plasmids were grown in sterile Luria-Bertani (LB) media containing kanamycin (25 μg/ml). Mid-log phase growth was achieved before 0.5 mM isopropyl β-D-thiogalactoside (IPTG) induction at 37°C for 6 h. Finally, the harvested cells were disrupted by sonication, and recombinant proteins purified by Ni-NTA columns as per manufacturer's instruction (Qiagen, Germany). Amicon 1 Ultra-15 Centrifugal Filter Devices (Merck Millipore, USA, MW cutoff 30 kDa or 50 kDa) were used to remove the imidazole and concentrate the purified recombinant proteins. Then the lipopolysaccharide (LPS) in recombinant proteins was removed using EtEraser Endotoxin Removal Kit (Chinese Horseshoe Crab Reagent Manufactory, Xiamen, China). LPS contamination was determined to be less than 1 EU/ml using the Quantitative Chromogenic Tachypleus Amebocyte Lysate (TAL) for endotoxin detection kit (Chinese Horseshoe Crab Reagent Manufactory, Xiamen, China).

Immunoblotting of recombinant EF-Tu protein (rEF-Tu) and recombinant HSP70 proteins (rHSP70)
Purified rEF-Tu and rHSP70 proteins were subjected to 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride (PVDF) membranes. The membranes were blocked in PBST with 5% nonfat milk for 1h and then incubated overnight at 4°C with specific sheep M. ovipneumoniae hyper-immune sera. Following washing with PBST three times, membranes were incubated with rabbit anti-goat IgG HRPconjugated antibody at 37°C for 1 h and exposed to SuperSignal 1 West Pico Chemiluminescent Substrate reagent for detection (Thermo Scientific, USA).

Identification of EF-Tu and HSP70 as membrane-associated protein
Immunogold. Fresh intact M. ovipneumoniae strain Y98 cells were harvested at mid-log phase and washed with 0.01 M PBS three times. The cell pellets were fixed in PBS containing 2% (w/v) paraformaldehyde and 2.5% (w/v) glutaraldehyde and processed as previously described [32]. After embedding in Lowicryl, sample sections were cut, mounted onto Formvar-coated nickel grid, blocked with PBS-Glycine buffer (PBS with 50 mM glycine) for 10 min, and washed with PBGT buffer (PBS with 0.1% gelatin, 1.0% BSA, 0.1% Tween 20) six times. Grids were incubated overnight at 4°C with anti-rEF-Tu or anti-rHSP70 sera (The sera were collected from mice immunized with either recombinant protein respectively 42 days after immunization, both 1:1,000 diluted). After washes with PGBT, grids were incubated at 25°C for 60 min with goat anti-mouse IgG (whole molecule)-Gold (5 nm) antibody (1:20 diluted, Sigma, USA). Samples were counterstained with aqueous uranyl acetate and examined by transmission electron microscope (TEM). TEM measurements were made on an H-7500 (Hitachi, Japan) at 80 kV. Sera from non-immunized mice were used as negative control.
Immunoboltting. Membrane proteins of 12 strains of M. ovipneumoniae, including strain Y98 and Mo-1, were extracted using 1.5%(v/v) Triton X-114. Briefly, 50 mL fresh intact M. ovipneumoniae cells were harvested at mid-log phase and washed with 0.01 M PBS three times. Then, the M. ovipneumoniae pellets were suspended in 2 mL 0.01 M PBS containing 1.5% (v/v) Triton X-114, incubated at 25°C for 10 min and centrifuged at 12,000×g for 30 min at 4°C. 2.5 mL cold ethanol was added to the supernatant, mixed well and precipitated overnight at -20°C, and then centrifuged at 12,000×g for 30 min at 4°C. The pellets were suspended in 5.0 mL 0.01 M PBS and centrifuged at 12,000×g for 30 min at 4°C. Finally, the pellets were suspended in 0.5 mL 0.01 M PBS. Immunoblotting was performed in the membrane proteins extracts of 12 strains of M. ovipneumoniae in order to identify whether EF-Tu and HSP70 were membraneassociated proteins. Mouse anti-rEF-Tu sera or anti-rHSP70 sera collected at 42 DAI (day after immunization), were used as primary antibodies at 1:2,000 dilution. Purified rEF-Tu and rHSP70 proteins were used as the positive controls in Western bolting.

Grouping and immunization of experimental animals
Seven-eight weeks old SPF female BALB/c mice were randomly divided into four groups. The first three groups were immunized with 1) recombinant protein rEF-Tu, 2) recombinant protein rHSP70 or 3) Mo extracts (the supernatant of Mo-1 whole cell ultrasonic lysates, LPS were less than 1 EU/ml) respectively. The fourth group served as a non-immunized control. Each mouse in each group was inoculated three times subcutaneously with 50 μg of antigen in 100 μL of either Freund's complete adjuvant (CFA) or Freund's incomplete adjuvant (IFA) at two week intervals. Antigens mixed and emulsified with the same volume of CFA were used for the first immunizations and antigens with IFA for the next two immunizations. Simultaneously, 100 μL of sterile PBS (pH 7.4-7.6) mixed with respective adjuvant as antigens were injected into each mouse of non-immunized control group.

Detection of antibody level and antibody subtypes
Mice sera were collected at 3, 7, 14, 21, 28, 35, 42, 49, 56, 63 and 70 DAI. Indirect ELISA was performed to detect IgG titers and subtypes in the immunized groups. Antibody titers were performed in all time point sera, whereas detection of antibody subtypes was performed only in sera at 42 DAI. ELISA plates were coated with purified rEF-Tu proteins or rHSP70 proteins or Mo extracts of M. ovipneumoniae wild strain Mo-1 at a concentration of 100 ng per well. For antibody titers, a 2-fold serial dilution of sera starting at 1:50 was performed. Goat-antimouse IgG (H+L) HRP diluted at 1:5,000 was used as secondary antibody. On the other hand, sera harvested at 42DAI were diluted at 1:1,000, HRP-goat-anti-mouse IgG1 or IgG2a (diluted at 1:500) were used for detection of antibody subtypes. All indirect ELISA was developed using TMB/ H 2 O 2 . Color development was stopped by adding 2M H 2 SO 4 and plates were read at 450 nm.

In vitro M. ovipneumoniae Growth inhibition test
In order to test the immunogenicity of the two proteins in vitro, GIT were performed in this study.
GIT on agar medium. Mycoplasma agar medium was melted and poured into culture dishes with a diameter of 60 mm. M. ovipneumoniae suspension of 100 μL (10 6 CCU/mL) was inoculated and distributed uniformly onto each dish after its solidification. After incubation at 37°C for 1 h, 2.00 mm diameter paper discs saturated with 10 μL of undiluted antisera (rEF-Tu antisera, rHSP70 antisera, Mo Extracts antisera or non-immune control (PBS) sera collected at 70 DAI) were placed on the agar. Dishes were incubated at 37°C with 5% CO 2 for 7 days and then the diameter of inhibition zones were measured.
GIT in liquid medium. A modified assay was performed as described previously [33]. Briefly, 100 μL (10 6 CCU/mL) M. ovipneumoinae strain Y98 were poured into 96-well microplates. Then, 100 μL of serial dilutions (1:20-1:2,560) of each group sera (collected at 70 DAI) were then added into wells in triplicate. Mycoplasma that cultured in the absence of serum severed as a positive control. As no antiserum control, only the culture medium was in the wells. Absorbance was recorded at 560 nm after incubation at 37°C with 5% CO 2 for 7 days. The percentage of growth inhibition was determined as follows: [absorbance of sample − absorbance of positive control]/[absorbance of no antiserum control − absorbance positive control]×100%.

Statistical analysis
Levels of antibodies and concentrations of cytokines between different groups were compared using One-way ANOVA (SPSS 17.0 and GraphPad Prism 6). P values below 0.05 were considered statistically significant.

Expression and identification of recombinant EF-Tu and HSP70
EF-Tu (Tuf) gene (1209bp) and HSP70 (DnaK) gene (1818bp) of M. ovipneumoniae wild strain Mo-1 were obtained by site-mutation (UGA to UGG) and overlap extension PCR assay (S1 Fig). The two expressed recombinant proteins were purified with Ni-NTA. The SDS-PAGE results showed that rHSP70 protein was about 67.1 kDa and rEF-Tu protein was about 44.8 kDa (Fig 1A). The two purified recombinant proteins could be recognized by sheep hyperimmune sera against M. ovipneumoniae strain Y98 in immunoblotting (Fig 1B).

Confirmation of HSP70 and EF-Tu as membrane-associated proteins
Immunogold staining and imaging by transmission electron microscope (TEM) were performed to locate EF-Tu and HSP70 to verify whether the two proteins were membrane-associated proteins. Based on the observations of EF-Tu and HSP70 in TEM micrographs, most M. ovipneumoniae presented circular morphology with immunogold staining mainly localized The observations thus indicated that the two proteins were indeed membrane-associated proteins (Fig 2). In order to locate EF-Tu and HSP70 on M. ovipneumoniae and to verify whether they were membrane associated proteins of M. ovipneumoniae, membrane proteins of 12 strains of M. ovipneumoniae, including Y98, Mo-1 and 10 wild-type strains were extracted using 1.5% (v/v) Triton X-114. Immunoblotting was performed with membrane proteins of 12 strains of M. ovipneumoniae, and the results confirmed presence of EF-Tu and HSP70 in extracted membrane proteins (Fig 3).

Analysis of antibody level and subtypes in sera of immunized mice
IgG titer in sera of BALB/c mice were detected by indirect ELISA using rEF-Tu, rHSP70 or Mo extracts as the coating antigens (Fig 4A). Results showed that mice immunized with rEF-Tu, rHSP70 or Mo extracts proteins elicited strong humoral responses with the rHSP70-immunized mice showing higher levels of antibody than rEF-Tu or Mo extract groups at 7 DAI. However, the differences among the three groups were not significant (p>0.05). At 14 DAI, the IgG levels in sera of three immunized groups were significantly increased especially from the rHSP70 group. The rHSP70 and rEF-Tu groups showed higher antibody titers compared to Mo extracts group, and the significant differences (p<0.01) lasted till 70 DAI. The difference between rHSP70 and rEF-Tu, however, were not significant (p>0.05). With the exception of the PBS negative control group, all the immunized groups could generate IgG1 and IgG2a simultaneously. There was no significant difference in the amount of IgG1 and IgG2a elicited by either rHSP70 or rEF-Tu, however they elicited a significantly higher IgG2a production when compared with Mo extracts group (Fig 4B).
While TNF-α, IL-12(p70) and IL-6 were chosen for their early immune response participation, IFN-γ, IL-4 and IL-5 levels were analyzed to assess the progression of the immune response induced by the recombinant proteins 14 days after the first immunization (Fig 5).
At 3 DAI, all three immunized groups showed higher levels of all six cytokines compared to the negative control group. The concentrations of IFN-γ and TNF-α in rHSP70 and rEF-Tu groups were significantly higher than the negative control group (p<0.05 or p<0.01). rHSP70 protein immunized group also showed significantly higher levels of IL-12(p70) than the negative control group. The level of all six cytokines in the Mo extracts group reached their peak concentration at 7 DAI, while the cytokines in rHsp70 and rEF-Tu groups were either slightly decreased or showed no change when compared with 3 DAI. At 14 DAI, all cytokines of rHsp70 group were higher than the other groups with the exception of IL-5.
Seven days after third immunization (35 DAI), all Th1 and Th2 type cytokine concentrations of rHSP70 and rEF-Tu groups were increased and at the highest level compared to 3 DAI, 7 DAI and 14 DAI. rHsp70 induced the higher expression levels of cytokines among four groups (Fig 5). There were significant differences (p<0.05, p<0.01 or p<0.001) between rHsp70 and Mo extracts in all cytokines except IL-6 ( Fig 5A-5F). The rHsp70 also induced a  higher expression level of IFN-γ compared to rEF-Tu group at 35 DAI (Fig 5B), while the concentration of TNF-α in rEF-Tu group was significantly higher with Mo extracts (p<0.05) ( Fig  5A). Besides, although the cytokine levels of Mo extracts group were higher than the negative control, only TNF-α value showed significant difference compared to the negative control (p<0.05) (Fig 5A).

Detection of lymphocytes secreting IFN-γ by ELISPOT
The cellular immune response was assessed by the amount of IFN-γ + secreting lymphocytes using ELISPOT assay. Mice spleen lymphocytes were isolated at 35 DAI and stimulated by rHSP70 or rEF-Tu or Mo extracts for 48h. Each spot represented one IFN-γ + secreting lymphocyte ( Fig 6A) and the mean spot of PHA positive control group was 313 in every 100,000 lymphocytes. Fig 6A and the calculated results (Fig 6B) demonstrated that rHSP70 (339/ 100,000 lymphocytes) and EF-Tu (252/100,000 lymphocytes) stimulation significantly induced IFN-γ secretion of lymphocytes compared to Mo extracts (31/100,000 lymphocytes) or PBS negative control (1/100,000 lymphocytes). Furthermore, the amount of IFN-γ + secreting lymphocytes in rHSP70 group was significantly different from rEF-Tu group, which were in concordance with the results of cytokines test (Figs 5B and 6B).

Growth inhibition test
Mycoplasma growth inhibition test was performed to verify immunogenicity of rEF-Tu and rHSP70 with sera from four groups of immunized mice. When compared to the non-immune control (PBS) group, inhibition zones on agar medium appeared in all the three other groups (Fig 7A). The average diameter of inhibition zone in rEF-Tu, rHSP70 and Mo extracts groups was 5.17±0.41 mm,7.00±0.32 mm and 5.5±0.32mm respectively. As shown in Fig 7B, sera from rEF-Tu, rHSP70 and Mo extracts groups immunized groups also inhibited the growth of M. ovipneumoniae in liquid medium. As expected, the anti-rHSP70 sera showed a higher growth inhibitory activity than two other immunized groups. This result also provides further evidence that the antisera of two recombinant proteins could inhibit the growth of M. ovipneumoniae in vitro.

Discussion
The epidemic of M. ovipneumoniae around the world has caused considerable economic loss to sheep industry. Despite extensive research on M. ovipneumoniae [2], a potential vaccine is yet to be discovered due to insufficient database on M. ovipneumoniae pathogenesis. Thus, it is necessary to characterize novel protective antigens for developing M. ovipneumoinae vaccines. In this study, we confirmed that both EF-Tu and HSP70 were membrane-associated proteins of M. ovipneumoniae and evaluated the immunogenicity of the two proteins.
Mycoplasma is the smallest and simplest self-replicating organism that contains the minimal complement of life-enabling genes and lacks cell wall which may further facilitate its membrane and cytoskeleton proteins to actively participate during infection of host cells [34]. As an intracellular pathogen, adhesion mediated by membrane-associated proteins is believed to be a significant step in mycoplasma pathogenesis [35,36]. Therefore, it may be possible to reduce infection of M. ovipneumoniae by immunization with membrane-associated proteins.
It has been recently reported that EF-Tu was a membrane-associated protein of M. pneumonia and mediated the fibronectin binding of M. pneumonia during pathogenesis [22,23]. HSP70 protein also plays an important role in the organization of the terminal organelle and is critical for adhesion to host cells in M. pneumonia [30], however, the location of HSP70 on mycoplasma is not clear. Here, we characterized EF-Tu and HSP70 were membrane-associated proteins by immunogold staining and immunoblotting. In addition to the rounded morphology of M. ovipneumoniae TEM data also indicated that these membrane-associated proteins were not exclusive to the surface alone, but also were intracellular as evidenced by cytoplasmic staining of the proteins. Using extracted membrane proteins of 12 strains of M. ovipneumoniae, we demonstrated the presence of EF-Tu and HSP70 by Western blotting, which further confirmed both EF-Tu and HSP70 are membrane-associated proteins of M. ovipneumoniae (Fig 3).
In previous studies, the membrane-associated proteins like P97, P42 and MnuA of M. hyopneumoniae, P30 of M. pneumoinae, MGC3 of M. gallisepticum were determined as the potential antigen candidates for vaccine development [18,[37][38][39][40]. EF-Tu of M. hyopneumoniae and M. bovis were reported to be potential antigens [24,41]. Moreover, HSP70 has been reported as an immunogenic biomolecule of mycoplasma that can induce and enhance humoral and cellular immunity [42]. Chen et al. also demonstrated that heat shock protein P42 of M. hyopneumoniae can be recognized by multiple B-cell and T-cell clones [37]. Taken all together, these studies lend the support to our observations of immunogenic EF-Tu and HSP70 in M. ovipneumoinae. Here, we demonstrated that EF-Tu and HSP70 of M. ovipneumoniae induce both humoral and cellular immune responses in mice.
The analysis of the antibody response revealed that Mo extracts, recombinant proteins rEF-Tu and rHSP70 were able to induce specific IgG antibodies, suggesting that these recombinant proteins could elicited a strong humoral immune response in mice. It is known that Th1 cells induce B cells to proliferate and produce antibodies such as IgG1 while Th2 cells induce B cells to produce IgG2a in mice [43]. We determined that both rEF-Tu and rHSP70 had the ability to produce more IgG1 and IgG2a antibodies than Mo extracts group (Fig 4B). Furthermore, the results revealed that both IgG1 and IgG2a in rHSP70 and rEF-Tu groups were induced at similar levels, which indicates that the recombinant proteins may induce a mixed Th1 and Th2 response in the mice.
Additionally, the results of cytokines production test showed that all proteins could active an innate immune response in mice ( Fig 5). As an important inflammatory cytokine, TNF-α is the earliest inflammatory cytokine to activate neutrophils and lymphocytes, which can increase the permeability of vascular endothelial cells, adjust tissue metabolism and promote synthesis and release of cytokines [44,45]. We here observed that rEF-Tu and rHSP70 elicited a significant early increase (3 DAI) in TNF-α levels which doubled at 35 DAI (Fig 5A). This increased TNF-α levels were also observed with Mo extracts when compared with non-immunized mice, but was not as robust as that produced by rEF-Tu or rHSP70. The same early and late cytokine production response was observed for both proteins when measuring IL-12(p70) and IL-6 ( Fig  5C and 5F). Based on our results, we suggest that the recombinant proteins could induce a stronger innate immune response than Mo extracts.
In general, recombinant protein vaccines are known to preferentially induce Th2 responses with weak or no Th1 responses. However, it has been reported that some recombinant protein of M. hyopneumoniae can induce Th1 response in mice [38,46,47]. The analysis of the cellular immune responses revealed that rEF-Tu and rHSP70 induced high levels of IFN-γ than Mo extracts group from 3 DAI to 35 DAI (Fig 5B). The ELISPOT assay also demonstrated the ability of both recombinant proteins to induce immune response by secretion of IFN-γ at 35 DAI (Fig 6). As expected, the ELISPOT results corresponded to the cytokine production test. Additionally, since the concentration of IL-4 and IL-5 induced by rEF-Tu and rHSP70 were significantly higher than non-immunized group (Fig 5D and 5E), we suggest that the two recombinant proteins could elicit Th1-and Th2-based immune responses. Furthermore, rHSP70 could re-activated lymphocytes quickly and induce specific cellular immune response according to the higher level of IFN-γ in HSP70 group.
In previous studies, HSP70 has been reported to induce antigen-specific immune responses thereby transducing various signals into dendritic cells (DCs) to enhance immune responses [48]. The C-terminal peptide-binding domain of HSP70 might act as a microbial adjuvant that upregulates the expression of cytokines such as IL-12 and TNF-α. IL-12 (p70) known to be a very potent cytokine for Th1 polarization, and it was the most induced from 3 DAI to 35 DAI in rHSP70 group in our study. Thus we suggest that rHSP70 might also act as a Th1 cytokinelike adjuvant for immune response induction in M. ovipneumoniae.
Since there is still no applicable laboratory animal model for experimental infection with M. ovipnemoniae with the exception of the sheep model, which is a tedious model to work with, we here performed M. ovipneumoniae GIT to verify the immunogenicity of the two protein specific antibodies (Fig 7). Herein, the experiments in vitro proved that antisera of mice immunized with rEF-Tu or rHSP70 could inhibit growth of M. ovipneumonia strain Y98 and the rHSP70 antiserum was more potent at inhibiting the growth of M. ovipneumoniae than the other ones. These results may provide us a critical marker to evaluate potential vaccine antigens of M. ovipneumoinae.
In summary, the results of this study showed that both HSP70 and EF-Tu were membraneassociated proteins of M. ovipneumoniae, and the recombinant proteins were able to induce Th1 and Th2 based immune responses elicited to produce high levels of IgG1 and IgG2a with increased INF-γ expression. Additionally, compared with the three immunized groups, rHSP70 protein could induce the highest level of IFN-γ expression and anti-HSP70 protein antibody showed a better ability inhibiting M. ovipneumoniae growth. The results demonstrated that both rHSP70 and rEF-Tu antigens might be good candidates for the development of M. ovipneumoniae subunit vaccine. However, it should be note that the immune responses of mice may not be extrapolated to other species. Therefore, further studies in sheep are required to evaluate the ability of the two proteins to control chronic non-progressive pneumonia.