Introduction

Motile aeromonads are gram-negative, facultative anaerobes that are ubiquitously present in aquatic environments. Bacteria of this genus have been reported to infect a variety of species including fish, amphibians, reptiles, and humans. Among the known species of genus Aeromonas, Aeromonas hydrophila, Aeromonas caviae, and Aeromonas veronii are considered to be the most pathogenic (Abbott et al. 2003), with A. hydrophila bacterial invasion fairly distributed in both vertebrates and invertebrates (Janda and Abbott 2010; Igbinosa et al. 2012). Over the past few years, A. hydrophila has caused fatal fishery infections all over the world, leading to huge economic losses (Janda and Abbott 2010). The infection is heightened in immunocompromised fishes or in those already facing stressful environmental conditions, such as changes in humidity and pH. Symptoms include skin infections such as fin paleness, reddening, ulcerations, skin lesions that escalate to hemorrhages, gastroenteritis, and exopthalmia. In humans, the symptoms of Aeromonas infection include osteomylitis, gastroenteritis (secretory) causing fever, pain, and dehydration, septicemia, skin and soft tissue infections, endocardium inflammation, pneumonia, and meningitis. Sources of infection include contaminated food or water, air transfer, and exposure of tissue wounds to nearby bacterial presence (Camus et al. 1998; Schlenker and Surawicz 2009; Behera et al. 2011).

Various approaches have been adopted to target these pathogenic bacteria. Terramycin and Remet-30 are the two registered antibiotics currently in use to treat this infection (Cipriano 2001). Vaccines are the alternative preventive measure for treatment of such bacterial infections. Rifampicin-resistant strain of bacteria as a live attenuated vaccine (patent no. US 7,988,977 B2; date of patent Aug. 2, 2011) has already been reported. However, the use of attenuated bacteria increases antigenic load in the host cell eliciting non-specific inflammatory responses, together with the risk of reversion to virulence and antibiotic resistance. This advocates the need for alternative vaccine strategies employing different A. hydrophila virulence factors as potential vaccine candidates.

Outer membrane proteins have previously been identified as suitable vaccine candidates and evaluated for their immunogenic and protective potential in different bacterial species (Pal et al. 1997; Wright et al. 2002; Khushiramani et al. 2012).

Exposed on the cell surface, these proteins are quickly recognized as extracellular foreign particles by the host cells, which then precipitate an immune response against the pathogen (Osman and Marouf 2014). The protective immune response against a pathogen is largely based on B cell-mediated humoral immune response (Wei et al. 2010). However, small synthetic peptides corresponding to these epitopes are poorly immunogenic and do not effectively stimulate T cell response (Purcell et al. 2007). This is overcome by conjugating the peptides to various adjuvants such as lipids, interleukin (IL)-12, and GM-CSF (Brown and Jackson 2005; Lee et al. 2001; Dakappagari et al. 2003; Markovic et al. 2006). Covalent coupling of the peptides to a carrier protein such as tetanus toxoid has also been used to augment immune response against the peptides (Herrington et al. 1987; Kurikka et al. 1995; Diethelm-Okita et al. 2000). Over the years, the use of subunit vaccine based on immunodominant regions of an antigen together with an adjuvant or a carrier protein, rather than the whole antigens, has gained popularity in vaccine design as it alleviates the adverse effect of the whole antigen, generally a virulent factor, and increases vaccine safety. Such vaccines provide many advantages over the conventional subunit vaccine approach including instigating a focused immune response, the facility to engineer epitopes in order to customize the type of immune response required, ease of synthetic preparation on a large scale, and the ability to add additional immunogenicity and stability-enhancing groups (carbohydrates, lipids, phosphate groups) (van der Burg et al. 2006; Purcell et al. 2007). B cell and T cell epitope-based vaccines have been successful against a variety of pathogens including viruses (Depla et al. 2008; Sun et al. 2013), bacteria (Zhu et al. 2010; Xu et al. 2011; Li et al. 2012), and parasites (Nardin et al. 2000).

The outer membrane proteins outer membrane protein F (OmpF) and OmpC are the two most common porins that make ~2 % of the total cellular protein at any time (Nikaido et al. 1996). The OmpF is upregulated during low osmolarity and poor nutritional conditions, favoring bacterial survival under adverse conditions (Ferrario et al. 1995). Absence of OmpK35, an OmpF homolog, has been correlated with reduced susceptibility to nalidixic acid highlighting the role of OmpF porin in antibiotic transport across membrane (Osman et al. 2012). OmpR, via the histidine kinase activity of envZ, regulates expression of both OmpF and OmpC porins, thus responding to sudden changes in the osmolarity of the environment (Cai and Inouye 2002). Role of OmpR in virulence was first established by Dorman et al. (1989) in Salmonella typhimurium strain. Mutation in OmpF and OmpC also led to attenuation of Salmonella virulence (Chatfield et al. 1991) demonstrating their role in bacterial virulence. Earlier studies from our laboratory have demonstrated the immunogenic and vaccine potential of the complete A. hydrophila OmpF. Immunization with the recombinant OmpF generated agglutination-positive antisera (Yadav et al. 2014). Immunization with a 14-mer synthetic peptide sequence representing an epitope of Psuedomonas aeruginosa OmpF conferred protection against P. aeruginosa bacterial challenge in murine model. The antisera raised against the same gave positive response in opsonophagocytic assay (Staczek et al. 2000), thereby highlighting the potential of OmpF derived epitopes to be used as promising vaccine candidates.

The present study envisages to develop an epitope-based vaccine against A. hydrophila using OmpF as an antigen. Employing an in silico approach that analyzes various physicochemical properties of amino acids, immunodominant surface epitopes of OmpF responsible for B cell-mediated immune response were identified. Since the identified peptides themselves do not generate sufficient immune response (Adar et al. 2009), the epitopes were genetically coupled to the heat labile enterotoxin B of Escherichia coli (LTB) as a carrier which has been reported to act as an adjuvant and elicit high magnitude epitope-specific immune response (Weltzin et al. 2000). The recombinant OmpF epitope-LTB fusion proteins were evaluated for their immunogenic and vaccine potential.

Materials and methods

Materials

All chemicals used in the study were of analytical grade and were procured from Sigma-Aldrich Chemical Co., USA, Sisco Research Laboratories, India, or SD Fine–Chemicals Limited, India, unless otherwise stated. Enzymes used for DNA modification were purchased from New England Biolabs (NEB, USA). Gel extraction and DNA purification kits were from Qiagen, Germany. Protein molecular weight markers were obtained from New England Biolabs (USA). Fermentas PageRuler Prestained Protein Ladder was procured from Thermo Scientific, USA (product code 11854544). Alkaline phosphatase (AP)-conjugated anti-His and anti-Fc anti-mouse antibodies were purchased from Sigma-Aldrich Chemical Co., USA. Bicinchoninic acid (BCA)-based protein estimation kit and Ni-NTA2+ slurry were obtained from G-Biosciences, USA. Oligonucleotides used for the experiments were synthesized by Sigma-Aldrich Chemical Co., USA.

Bacterial strains, plasmids, and animals

E. coli DH5α and BL21(λDE3) pLysS strains were obtained from GIBCO BRL, USA, and Novagen, USA, respectively. A. hydrophila (isolate EUS112) was kindly provided by Dr. I. Karunasagar, College of Fisheries, Mangalore, India, and has been deposited to the national repository for microorganisms, Microbial Tissue Culture Collection (MTCC no. 12301), Chandigarh, India. Other Aeromonas species namely Aeromonas culicicola (MTCC no. 3249), Aeromonas sobria (MTCC no. 1608), and Aeromonas liquefaciens (MTCC no. 2654) were procured from MTCC, Chandigarh. The plasmid pQELTB harboring LTB in pQE32 vector, used to create genetic fusion of OmpF epitopes and LTB, was a kind gift from Dr. Lalit C. Garg, National Institute of Immunology (NII), New Delhi, India (Kaushik et al. 2013). Plasmid pET22b + was used for recombinant protein expression purpose (Novagen, USA).

Female Balb/c mice (n = 5/group) were obtained from the Jawaharlal Nehru University (JNU) animal house facility. The animals were maintained on feed (Hindustan Lever Ltd., Mumbai, India) and water ad libitum. The Institutional Animal Ethics Committee, JNU, New Delhi, India, approved the animals’ stated use.

In silico prediction of B cell epitopes

Linear B cell epitopes of the OmpF of A. hydrophila (GenBank accession no. CCO02501.1) were predicted using IEDB analysis resource’s prediction tools. Epitope analysis was conducted using Emini surface accessibility (Emini et al. 1985), Chou and Fasman beta turn prediction (Chou and Fasman 1978), Karplus and Schulz flexibility prediction (Karplus and Schulz 1985), Kolaskar and Tongaonkar antigenecity (Kolaskar and Tongaonkar 1990), Parker hydrophilicity prediction (Parker et al. 1986), and BepiPred linear epitope prediction (Larsen et al. 2006). Peptides showing positive results for the majority or all of the methods used were selected as putative candidate epitopes. Non-peptide regions between the selected peptides and reshuffled regions were also analyzed by the programs as negative controls.

Construction of LTB-OmpF epitope fusion clones

Amino acid sequences of the predicted peptides were reverse-translated into their corresponding coding sequence. Complementary oligonucleotides corresponding to various epitopes were synthesized with overhanging KpnI and HindIII sites at 5′ and 3′ ends, respectively. These oligonucleotides were annealed and cloned into KpnI and HindIII digested pQELTB vector obtained from NII (Kaushik et al. 2013). This results in recombinant expression of LTB-epitope fusion protein under the control of T5 promoter in E. coli. However, due to negligible expression of the recombinant fusion protein from this construct in appropriate host (E. coli M15 cells), alternate strategy for cloning the fusion gene was adapted. The fusion gene fragment harboring the peptide sequence as a C-terminal fusion with the LTB was released from pQELTB-epitope construct using SacI and HindIII, ligated to pET22b + expression vector digested with the same enzymes. The ligation mix was transformed into competent E. coli DH5α cells, and transformants were selected on LB agar plates containing ampicillin (100 μg/ml). The putative recombinants were analyzed by digesting isolated plasmid DNA with SacI and HindIII and confirmed by automated DNA sequencing at the DNA sequencing facility, University of Delhi South Campus, New Delhi. The recombinant thus obtained was named as pETLTB.OmpF aa-aa, the “aa-aa” indicating the amino acid residue (aa) position of the epitope with respect to the OmpF protein.

Expression and purification of recombinant LTB-peptide fusion proteins

For expression of the fusion proteins, E. coli BL21(λDE3) pLysS cells were transformed with the recombinant plasmid and the transformants were selected using ampicillin (100 μg/ml). Expression analysis of the fusion proteins was carried out as described earlier (Yadav et al. 2014). An overnight (O/N) culture of the E. coli BL21(λDE3) pLysS cells harboring the pET22b + -fusion gene construct was used as an inoculum (1 %) for secondary cultures which were allowed to grow until the culture density reached 0.6 OD600. The cells were induced using 1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG), and allowed to grow for 6 h post-induction. Cells from both the uninduced and induced cultures were harvested by centrifugation at 6000 rpm (Sorvall, SS34 rotor), and cell lysates were prepared by suspending the cells in appropriate volume of lysis buffer (50 mM Tris–HCl, pH 6.8, 2 % sodium dodecyl sulfate (SDS)). The lysates were analyzed for expression of the fusion proteins by SDS polyacrylamide gel electrophoresis (PAGE) (15 %) following the method of Laemmli (1970).

To study the localization of the expressed protein, soluble and insoluble fractions from induced cell cultures were prepared essentially as described earlier (Yadav et al. 2014) and analyzed by 15 % SDS-PAGE, followed by Coomassie Brilliant Blue staining.

Recombinant fusion proteins were purified from solubilized inclusion bodies using immobilized-metal affinity chromatography following the protocol standardized for the rOmpF protein (Yadav et al. 2014). After washing the column thoroughly with wash buffer (8 M urea, 20 mM Tris–HCl, pH 8.0, 500 mM NaCl), non-specifically bound proteins were eluted with wash buffer-I containing 20 mM imidazole. The bound proteins were eluted with increasing concentrations of imidazole in wash buffer-I. Chromatographic fractions were analyzed for the presence of fusion proteins by 15 % SDS-PAGE. Fractions containing the fusion proteins were pooled and refolded using urea gradient dialysis method (Yadav et al. 2014) until urea concentration was reduced to 0.5 M; thereafter, the proteins were transferred to 1× PBS. Purified fusion proteins were stored in small aliquots at −80 °C until further use. Protein concentration in cell lysates and different fractions was estimated using BCA protein estimation kit.

Recombinant expression and purification of rOmpF and rLTB

The parent protein rOmpF of A. hydrophila was expressed and purified from E. coli BL21(λDE3) cells harboring pETAhompF (available in the lab) essentially as described earlier (Yadav et al. 2014).

Soluble rLTB purified by the method of Panda et al. (1995) was kindly provided by Dr. Lalit C. Garg, National Institute of Immunology, New Delhi.

Immunization of mice with recombinant fusion proteins

Post-collection of preimmune sera a mice were immunized (intraperitoneally (i.p.)) with 15 μg of recombinant fusion proteins in 1× PBS emulsified in complete Fruend’s adjuvant (CFA) in 1:1 ratio (five mice/group). Control group was immunized with 1× PBS emulsified in CFA. After an interval of 14 days, booster dose (15 μg of the respective protein or PBS, i.p.) emulsified in incomplete Fruend’s adjuvant (IFA) was given. Mice were bled a week after each immunization/booster dose. After incubation at 37 °C for 1 h, sera were collected by centrifugation at 5000 rpm at 4 °C for 10 min and stored at −20 °C for further use.

ELISA for end point titer determination and antibody isoytyping

To determine the efficacy of the antisera generated, fusion protein (500 ng/100 μl/well) was coated in a 96-well round-bottom plate (Nunc, USA) and incubated at 4 °C O/N. The unbound protein was removed by washing the wells three times with 1× 0.05 % Tween 20 in 1× PBS (PBST). Non-specific sites were blocked with 2 % BSA (in 1× PBST) at 37 °C for 1 h. The wells were washed three times with 1× PBST between each treatment. Different dilutions of antisera raised in mice (100 μl/well, diluted in 1× PBS) were added to the wells, followed by incubation at 37 °C for 1 h. After thoroughly washing with 1× PBST, alkaline phosphatase-conjugated anti-mouse IgG (Fc specific) antibody (1:5000 in 1× PBS) was added. The color was developed by the addition of p-nitrophenylphosphate (PNPP) substrate (1 mg/ml) made in AP buffer (1 mM MgCl2, 50 mM Na2CO3, pH 9.8), and the absorbance was measured at 405 nm using an enzyme-linked immunosorbent assay (ELISA) reader (Tecan, USA).

A similar ELISA was used to determine the type of immune response (Th1 or Th2) generated after immunization. Antisera generated in mouse were used as the primary antibody while anti-IgG1, anti-IgG2a, or anti-IgG2b antibodies conjugated with horse-radish peroxidase (1:5000 in 1× PBS) were used as secondary antibodies. The color was developed using 3,3′,5,5′-tetramethylbenzidine (TMB) substrate and absorbance was read at 450 nm with 570 nm as the reference wavelength.

Western blot analysis

Cell lysates from the uninduced or induced E. coli cell cultures harboring the fusion gene construct and purified proteins were separated on a 15 % SDS-PAGE and blotted onto nitrocellulose membrane (Advanced Microdevices, Ambala, India) for 1.5 h at 50 V using 1× electrode transfer buffer [25 mM Tris–HCl, pH 8.3, 192 mM glycine, 20 % (v/v) methanol]. The membrane was incubated in 2 % BSA at room temperature (RT) for 1 h at RT or at 4 °C O/N followed by three washes with 1× PBST for 10 min each (as done for ELISA earlier). The membrane was then incubated with primary antibody (either antisera generated against the fusion proteins or with anti-histidine antibody; 1:10,000 in 1× PBS) for 1 h at RT. The blot was again washed three times with 1× PBST. Secondary antibody (alkaline phosphatase-conjugated goat anti-mouse Fc region, 1:10,000 in 1× PBS) was then added to the blot and allowed to incubate at 37 °C for 1 h, followed by three 1× PBST washes. Immunoreactive bands were visualized by the addition of NBT/5-bromo-4-chloro-3-indolyl phosphate (BCIP) (Promega, USA). The color reaction was terminated by the addition of double-distilled water.

Lymphocyte proliferation assay

Mice were sacrificed within 7–10 days of receiving the last booster dose. Spleens were surgically removed, and the cells were suspended in chilled Dulbecco’s modified Eagle’s medium (DMEM, Biological Industries, Israel). After lysing RBCs with 0.9 % ammonium chloride, splenocytes were washed with DMEM and counted using trypan blue. Splenocytes were seeded at a density of 1 × 105 cells/100 μl/well in 96-well plates and stimulated in vitro with the respective fusion proteins (15 μg/ml). Splenocytes from immunized mice stimulated with the fusion protein were taken as the experimental group. Splenocytes from the immunized group stimulated with ConA (5 μg/ml) were used as the positive control while negative controls included unstimulated splenocytes from the mice immunized with the recombinant proteins, unstimulated and stimulated splenocytes from the mice immunized with PBS. The plates were incubated in a humidified CO2 incubator (5 % CO2, 37 °C). Cell proliferation was measured at different time intervals using an XTT kit (Biological Industries, Israel) as per the manufacturer’s protocol. XTT (50 μl/100 μl medium) was added to each well and incubated in a CO2 incubator for 4 h. The reaction was stopped using 1 M orthophosphoric acid (50 μl/well). Absorbance was measured at 450 nm with reference wavelength of 570 nm. All cell culture experiments were performed in triplicate.

Cytokine ELISA for IL-4 and IFN-γ levels

Culture supernatants from the plate used for lymphocyte proliferation assays were collected at 24 and 48 h and stored at −80 °C until further use. The levels of interferon-γ (IFN-γ) and IL-4 in the supernatant were measured using BD cytokine-ELISA kit (Becton Dickinson Pharmingen, USA) as per the manufacturer’s instructions.

Cytokine array analysis

Culture supernatant (collected 72 h post-stimulation) from the splenocytes of immunized mice, which gave maximum proliferation upon stimulation with the respective fusion protein, were analyzed for a global cytokine response using Proteome Profiler antibody array (R&D Systems, USA) as per the manufacturer’s protocol. Culture supernatant from the splenocytes isolated from PBS-immunized mice and stimulated with the same protein constituted the negative control. The ready-to-use analyte-specific capture antibodies precoated membranes were blocked with array buffer 6 at RT for 1 h. Blocking buffer was then removed, and the culture supernatants (1 ml of supernatant added to 0.5 ml of array buffer 4) mixed with 15 μl of reconstituted detection antibody cocktail were added to the membrane and incubated at 4 °C O/N on a rocking platform shaker. After washing with 1× wash buffer (20 ml) for 10 min, streptavidin-HRP solution (1:2000 in array buffer 6) was added to the membrane, followed by incubation at RT for 1 h with shaking. The membrane was again washed with 1× wash buffer, and immunoreactive spots were detected by the addition of 1 ml of Chemi Reagent mix (provided in the kit) for 1 min. The images were captured by a ChemiDoc, Biospectrum-500 (UVP, USA), and densitometric analysis of the immunoreactive spots was performed using VisionWorks® LS analysis software version 6.8 (UVP Laboratory Products, USA).

Agglutination assay

To assess whether the antisera generated could interact with live A. hydrophila cells, an agglutination assay was performed as described by Yadav et al. (2014) with minor modifications. Cultures of different A. hydrophila strains were as grown at 37 °C overnight with shaking at 200 rpm. A secondary culture was inoculated from the O/N culture (1 %) and allowed to grow till mid-log phase. For the experimental group, 1 × 108 colony-forming units (cfu) were incubated with antisera (1:320 in 1× PBS) generated against different fusion proteins. For the control group, the same population of bacteria was incubated with preimmune serum (same dilution). After incubation at 37 °C for 1 h, the cells were pelleted down at 5000 rpm at RT for 10 min and washed with 1× PBS to remove the unbound antibodies. The cells were then suspended in 1× PBS and smeared uniformly on a glass slide. The bacteria were then heat-fixed, slides air-dried, and stained with methylene blue. Slides were washed with double-distilled water to remove excess stain and visualized under a bright-field microscope (Model Eclipse TE2000S, Nikon, USA).

Flow cytometry

Surface availability of the outer membrane protein to the antibodies generated against the fusion proteins was confirmed by fluorescence-activated cell sorting (FACS) analysis. An O/N culture of A. hydrophila was used to inoculate a secondary culture which was harvested during the mid-log phase. Bacterial cells (1 × 106 cfu) were washed with chilled 1× PBS and incubated with antisera raised in mice against the fusion protein or the preimmune sera (1:200 in 1× PBS). Unbound antibodies were washed away with 1× PBS, and the cells were further incubated with FITC-conjugated goat anti-mouse IgG. Fluorescence intensity was measured using FACS Calibur (Beckon Dickinson Immunocytometry System, San Jose, CA) and analyzed using FCS express 4 flow cytometry analysis software.

Statistical analysis

Data represent the mean and standard deviation of three independent experiments, each performed in triplicate. Student’s two-tailed t test was used to determine the p value. A p value ≤0.05 was considered statistically significant.

Gene sequence accession numbers

The OmpF of A. hydrophila (EUS112): GenBank accession no. CCO02501.1

Heat labile enterotoxin chain B of E. coli (LTB): GenBank accession no. M17874.1

Results

B cell epitope prediction and cloning in expression vector

B cell epitopes were predicted using different online software, and the peptide sequences with maximum hits among all software used were identified (Fig. 1a–d). Putative B cell epitopes on the OmpF of A. hydrophila were thus selected. These spanned amino acid residues 66–80, 201–214, 235–250, 270–285, and 290–309 (Fig. 1e). Region 290–309 was predicted by only one of the prediction programs and was included in the study as negative control. Nucleotide sequences corresponding to these peptides with respective amino acid positions are listed in Table 1.

Fig. 1
figure 1

In silico analysis of linear immunodominant B cell epitopes of rOmpF protein. Different algorithms were used for predicting surface-exposed regions of rOmpF. a Parker hydrophilicity epitope prediction. b BepiPred linear epitope prediction. c Emini surface accessibility prediction. d Chou and Fasman beta turn prediction. e Schematic representation of identified epitopes of OmpF A. of hydrophila (not to scale). Numbers represent amino acid residue positions of the epitopes

Table 1 Predicted nucleotide and amino acid sequences of B cell epitopes in the OmpF of A. hydrophila

Cloning of the oligonucleotide corresponding to the epitope as a C-terminal fusion with LTB in plasmid pQE.LTB, as described earlier in the “Materials and methods” section, and its subsequent expression in E. coli M15 cells resulted in negligible expression of the fusion protein. Hence, the LTB.OmpF epitope fusion inserts released from pQE constructs by SacI and HindIII digestion were cloned in pET22b + expression vector digested with the same enzyme. Cloning of the fusion gene was confirmed by digesting the putative recombinants with SacI and HindIII, releasing inserts of expected size (data not shown). DNA sequencing of the recombinant plasmid DNA further confirmed in frame cloning of the OmpF epitope(s) with the LTB. Recombinant clones were designated as pETLTB.ompF 66–80 , pETLTB.ompF 201–214 , pETLTB.ompF 235–250 , pETLTB.ompF 270–285 , and pETLTB.ompF 290–309 , the subscript numbers indicating the amino acid residues in OmpF constituting the epitopes.

Expression and purification of recombinant fusion proteins

Transformation of the confirmed recombinant plasmids harboring the LTB-epitope fusion gene into E. coli BL21(λDE3) pLysS cells and induction with IPTG resulted in expression of the LTB-peptide fusion proteins (Fig. 2a). These are rLTB.OmpF66–80 (rEpiF1, lane 2), rLTB.OmpF201–214 (rEpiF2, lane 4), rLTB.OmpF235–250 (rEpiF3, lane 6), rLTB.OmpF270–285 (rEpiF4, lane 8), and rLTB.OmpF290–309 (rEpiF5, lane 10) (indicated by arrow). No band at the expected position of the respective fusion proteins in the uninduced culture of the respective clones (lanes 1, 3, 5, 7, and 9) suggested tight control of the recombinant protein expression (Fig. 2a). Immunoblot analysis using anti-6× His tag antibody further confirmed the expressed products in the induced cell cultures (lanes 2, 4, 6, 7, and 9, respectively) to be the recombinant proteins (Fig. 2b). All the five recombinant LTB-epitope fusion proteins could be purified to near 98 % homogeneity (established by densitometric analysis of the purified proteins resolved on SDS-PAGE) using Ni2+-NTA affinity chromatography (Fig. 2c). Approximately 60 mg of refolded purified proteins could be obtained at shake flask level.

Fig. 2
figure 2

a SDS-PAGE analysis of E. coli BL21(λDE3) pLysS cells harboring pET.LTB.ompF epitopes for expression analysis. Lanes 1, 3, 5, 7, and 9 show cell lysates prepared from the uninduced E. coli BL21(λDE3) pLysS cells while lanes 2, 4, 6, 8, and 10 show cell lysates prepared from induced E. coli cells harboring rEpiF1 (rLTB.OmpF66–80), rEpiF2 (rLTB.OmpF201–214), rEpiF3 (rLTB.OmpF235–250), rEpiF4 (rLTB.OmpF270–285), and rEpiF5 (rLTB.OmpF290–309), respectively. The arrow points to the recombinant proteins only in the induced cell lysates. b Western blot analysis of the rEpiF1, rEpiF2, rEpiF3, rEpiF4, and rEpiF15. Authenticity of the expressed proteins was confirmed by blotting using anti-His antibody. Lanes 1, 3, 5, 8, and 10 denote uninduced cell lysates whereas lanes 2, 4, 6, 7, and 9 denote induced cell lysates of E. coli BL21(λDE3) pLysS cells expressing rEpiF1, rEpiF2, rEpiF3, rEpiF4d, and rEpiF5, respectively. Immunoreactive bands at ~13 kDa were seen only in the induced cell lysates. c Purification of recombinant LTB-epitope fusion proteins using immobilized Ni2+-NTA affinity chromatography. Lanes 1–5 show purified rEpiF1, rEpiF2, rEpiF3, rEpiF4, and rEpiF5, respectively, eluted with 300 mM imidazole. Lane M indicates protein molecular weight migration (kDa)

Antigenicity of the recombinant fusion proteins and cross-reactivity analysis of the antifusion protein antisera with fusion partners

Antisera generated against the five recombinant LTB.OmpF fusion proteins were analyzed for antibody titer by coating the respective recombinant protein in a 96-well ELISA plate.

All recombinant proteins except the rEpiF5 (harboring the epitope spanning 290–309 aa of the OmpF), which generated minimal antigenic response, were found to be highly immunogenic with end point titers >1:80,000. In order to assess whether the antisera generated against these four LTB.OmpF epitope fusion proteins were able to recognize the parent OmpF protein, ELISA using recombinant OmpF as the target antigen was performed (Fig. 3a). Relative absorbance observed on incubation with the antisera generated against the four recombinant LTB.OmpF epitope fusion proteins on day 35 post-immunization clearly shows that the anti-rEpiF1 antisera gave maximum absorbance in the assay when compared to the other three fusion proteins. The absorbance with the anti-rEpiF1 antisera was ~50 % of that obtained with the anti-rOmpF antisera, suggesting that the epitope region 66–80 of the OmpF is highly immunogenic and is able to generate antibodies that could cross-react with the full-length rOmpF efficiently. The anti-rEpiF1 antisera was also analyzed for its cross-reactivity with the other fusion partner LTB, rOmpF, and the fusion protein by Western blot analysis (Fig. 3b). Detection of strong immunoreactive band in all the three lanes shows that the anti-rEpiF1 antisera could recognize both the rLTB (lane 3) and rOmpF (lane 4) as efficiently as it recognized the fusion protein rEpiF1 (lane 2), suggesting that the antigenicity of both the LTB and the epitope is preserved in the fusion protein. BSA (lane 1) included as negative control was not recognized by the anti-rEpiF1 antisera.

Fig. 3
figure 3

Specificity determination of the antibody generated in mice. a Cross-reactivity analysis of the fusion proteins. Ability of the anti-rEpiF1, anti-rEpiF2, anti-rEpiF3, and anti-rEpiF4 antisera to recognize native rOmpF, i.e., the parent protein was demonstrated using ELISA. Absorbance at 450 nm indicates relative cross-reactivity of the antisera against different fusion proteins with the rOmpF. Anti-rOmpF antisera was included as positive control. b Specificity of the antisera generated by immunoblot analysis. Lanes 1–4 represent BSA, induced cell lysate from E. coli cells expressing rEpiF1, rLTB, and rOmpF, respectively. Intense immunoreactive bands in lanes 2, 3, and 4 show the antisera to be highly specific for rEpiF1, rLTB, and rOmpF. Lane M indicates protein molecular weight migration (kDa). c Sandwich ELISA to determine GM1 ganglioside receptor binding activity of the rEpiF1 using anti-LTB antibodies. Positive and negative controls included LTB and BSA, respectively

Since LTB was used as a fusion partner because of its immunostimulatory property, it was imperative to establish if the LTB in the recombinant fusion protein retained its receptor binding activity. ELISA-based GM1 ganglioside receptor binding assay shows that the rEpiF1 could bind to the receptor whereas no binding was observed with BSA, included as negative control (Fig. 3c). The binding efficiency of the fusion protein was ~61 % when compared to the native LTB. Since GM1 binding activity of the LTB is required for its immunogenicity (de Haan et al. 1998; Bhatia et al. 2014), the rEpiF1 fusion protein is likely to be significantly immunogenic, a primary requirement for a vaccine candidate.

Analysis of immune response generated against the rEpiF1

Antibody titers against the IgG1, IgG2a, and IgG2b in the antisera generated against the rEpiF1 were used as tools to assess the type of immune response generated against the fusion protein (Fig. 4). An increase in all the three isotypes was observed in the anti-rEpiF1 antisera in comparison to the preimmune sera. IgG1/IgG2a and IgG1/IgG2b ratios in the antisera were found to be >1 on 35 days post-immunization (dpi), suggesting that the fusion protein rEpiF1 generated strong Th2-biased immune response.

Fig. 4
figure 4

Antibody isotyping of anti-rEpiF1 antisera raised in mice. Anti-rEpiF1 antisera collected at 35 days post-immunization was analyzed for the presence of different isotypes, namely IgG1, IgG2a, and IgG2b using isotype-specific secondary antibodies

In vitro stimulation of splenocytes isolated from rEpiF1-immunized mice with the same protein resulted in marked increase in T cell proliferation with a stimulation index of ~1.62 in comparison to that of control splenocytes (~1.1) (Fig. 5a). This increased proliferation can be directly related to heightened T cell activation in response to antigenic stimulation and that immunization with the rEpiF1was capable of generating T cell memory.

Fig. 5
figure 5

a In vitro lymphocyte proliferation assay. The spleen was surgically removed from Balb/c mice immunized with 15 μg rEpiF1 7 days after the last booster dose. Splenocytes were stimulated in vitro with 15 μg/ml rEpiF1. 1× PBS was used for stimulating splenocytes from PBS-immunized control mice. Cells were cultured for 24, 48, and 72 h under 5 % CO2 humified environment at 37 °C. Stimulation index was calculated for both control and protein immunized spleenocytes. b, c IL-4 and IFN-γ, respectively, in the culture supernatants collected at 24 and 48 h. *p ≤ 0.05; **p ≤ 0.02; ***p ≤ 0.01

Culture supernatants of the splenocytes were also analyzed for IFN-γ and IL-4 levels using cytokine specific ELISA kits to evaluate T-helper cell polarization in vitro. As evident from the figure, a significant increase in both IL-4 (Fig. 5b) and IFN-γ (Fig. 5c) levels were observed in sensitized splenocytes as early as 24 h of stimulation when compared to unstimulated splenocytes. The levels of both the cytokines continued to increase till 48 h (experimental duration), suggesting that the rEpiF1 has been able to generate mixed immune response (cell mediated as well as humoral). No detectable increase in the levels of either of the cytokines was observed in the splenocytes isolated from unimmunized mice.

In order to have an insight into global changes in the cytokine profile of the culture supernatant obtained from the stimulated splenocytes (collected 72 h post-stimulation) of rEpiF1-immunized mice in comparison to control splenocytes (PBS-stimulated), cytokine array analysis (R&D Systems) was performed (Fig. S1, supplementary material). An increase in diverse cytokines related majorly with the Th1-type immune response was detected in the culture supernatant of the protein-stimulated splenocytes in comparison to the control (Fig. 6). Significant increase in the levels of CCL5 at p ≤ 0.005; CCL3/MIP-1α, CXCL2, IL16, CD54, and IFN-γ at p ≤ 0.02; IL-1 F2 also termed as IL-1β and IL-1 F3 at p ≤ 0.05 was observed in the culture supernatant of protein-stimulated splenocytes when compared to that of control.

Fig. 6
figure 6

T cell immune response analysis by cytokine profiling. Culture supernatants from the control and sensitized splenocytes (as described in Fig. 5) were collected 72 h post-stimulation were subjected to cytokine analysis using mouse cytokine array kit from R&D Systems. Cytokines showing significantly increased levels are shown. *p ≤ 0.05; **p ≤ 0.02; ***p ≤ 0.005

Evaluation of vaccine potential of rEpiF1 immunization

Since the fusion construct was generated with an aim to develop an epitope-based vaccine against A. hydrophila using an epitope of the OmpF of A. hydrophila in fusion with the LTB (as carrier), it was essential to determine if the antisera generated against the fusion protein was able to bind to the OmpF protein on A. hydrophila cells.

Interaction of the anti-rEpiF1 antisera raised in mice with the bacterial membrane of A. hydrophila was visualized using fluorescent activated cell sorting. Positive interaction with anti-rEpiF1 and anti-rOmpF (Figs. 7, S2b, and S2c in the supplementary material) was directly linked to an increase in the fluorescence intensity, while minimal fluorescence was observed when the bacteria were incubated with preimmune sera (Figs. 7 and S2a, supplementary material). Fluorescent cell population (%) obtained upon incubation with the antisera raised against the parent protein rOmpF (35.26 ± 1.43) was not significantly different when compared to that observed with the anti-rEpiF1 antisera (30.4 ± 1.15) (Figs. 7 and S2), further confirming that the antigenicity of the OmpF epitope spanning 66–80 amino acid residues was retained in the fusion protein and that the antisera generated against the fusion protein could specifically interact with live A. hydrophila cell membrane.

Fig. 7
figure 7

Fluorescent cell population (%) of A. hydrophila cells incubated with preimmune (PI) sera, anti-rOmpF, and anti-rEpiF1 antisera. Data represent mean ± SD of three independent experiments, shown in Fig. S1 (supplementary material)

Vaccine potential of the rEpiF1 was confirmed by the ability of antisera generated against this protein to agglutinate live A. hydrophila cells (Fig. 8g, h). No agglutination was observed when the cells were incubated with preimmune sera of these mice (Fig. 8e, f). The anti-rEpiF1 antisera exhibited broad cross-reactivity as it was able to agglutinate other species of Aeromonas namely A. culicicola (Fig. 8k, l), A. sobria (Fig. 8o, p), and A. liquefaciens (Fig. 8s, t). As observed with A. hydrophila cells, no agglutination was observed when these species were incubated with preimmune sera (A. culicicola, Fig. 8i, j; A. sobria, Fig. 8m, n; A. liquefaciens, Fig. 8q, r). Antisera generated against the recombinant fusion protein rEpiF2 (harboring the OmpF epitope spanning 201–214 residues), which did not give significant cross-reactivity with the rOmpF in ELISA (Fig. 8c, d), included as a negative control was also not able to agglutinate A. hydrophila cells, and the bacterial cells remained isolated as observed with the preimmune sera of the respective mice (Fig. 8a, b). These data further confirm the specificity of interaction of anti-rEpiF1 antisera with the OmpF of A. hydrophila cells resulting in agglutination.

Fig. 8
figure 8

Bacterial agglutination assay using the anti-rEpiF1 antisera. Live log phase A. hydrophila cells (1 × 108 cfu) were incubated with preimmune sera and anti-rEpiF1 antisera (1:320) in 1× PBS at 37 °C for 1 h. The left and right panels show cells of different Aeromonas species incubated with preimmune and antisera, respectively, raised against either rEpiF2 (included as negative control) or anti-rEpiF1 antisera (test protein). Agglutination patterns of A. hydrophila (ah), A. culicicola (il), A. sobria (mp), and A. liquefaciens (qt), respectively, incubated with preimmune and antisera against the respective proteins

Discussion

In spite of several attempts in progress, development of a fully protective vaccine against the fish pathogen A. hydrophila still remains a challenge. Outer membrane proteins have proven to be excellent targets for subunit vaccine development (Achouak et al. 2001). They have not only shown remarkable immunogenic potential, anti-rOmp48 antisera of A. hydrophila have been shown to cross-react with whole cell proteins of other bacteria including A. veroni, Vibrio parahaemolyticus, Edwardsiella tarda, and E. coli (Khushiramani et al. 2012). Previous studies from our laboratory have reported the immunogenic potential of the rOmpF protein of A. hydrophila. The antisera was agglutination positive and could cross-react with cell lysates of different Aeromonas isolates. As bacteria of Aeromonas spp. are known to have biochemical and antigenic heterogeneity, ability of an antigen to produce antibodies that recognize different strains of a bacterial pathogen increases its efficacy as a potential vaccine candidate (Yadav et al. 2014).

Epitope-based vaccines have for long stood out as promising vaccine candidates against diverse pathogens (Christy et al. 2012; Hurtgen et al. 2012; Pan et al. 2012; Kunthalert et al. 2013; Rojas-Caraballo et al. 2014) and have forayed their way in human clinical trials as well (Iinuma et al. 2014). Other than helping to focus on and thereby narrowing down the specificity of immune response, they also leave room for chimera generation including multiple epitopes from same as well as different and slightly more immunogenic antigens. Though epitopes can be linear or conformational, Zhu et al. (2010) have demonstrated the immunodominance of linear epitopes in the past. Therefore, in the present study, we have attempted to predict linear B cell epitopes and analyze their ability to evoke focused immune response. Identification of immunodominant B cell epitopes using an in silico approach over the experimental one is advantageous as it is rapid and economical, thus accelerating vaccine design process. We used multiple bioinformatic algorithms that assess different physicochemical properties such as hydrophilicity, flexibility, β turn structure, and surface accessibility to predict B cell epitopes of the OmpF of A. hydrophila. Use of multiple algorithms ensures that the predicted amino acid residues would enjoy benefits of having at least more than one B cell epitope characteristic. Five peptide sequences of OmpF of A. hydrophila were thus identified as putative epitopes.

Since the epitopes themselves are too small to generate an immune response, these need to be conjugated or covalently coupled to a carrier (Kurikka et al. 1995; Diethelm-Okita et al. 2000; Lee et al. 2001; Brown and Jackson 2005; Markovic et al. 2006; Ghosh et al. 2013). Earlier studies on refolded LT-toxoid fusion proteins have been reported to enhance the antigenicity of non-immunogenic antigens (Liu et al. 2011). C-terminal fusion of LTB with antigen has earlier been proven to enhance the potential immunogenicity, permitting the incorporation of adjuvant in the vaccine formula itself (Waheed et al. 2011). Even co-expression of LTB with the candidate antigen protein produced a four times higher antibody titer when compared with individual antigen strain immunization or a mixture of the two strains (Maggi et al. 2002). Therefore, we cloned the nucleotide sequences encoding the predicted epitopes in translational fusion with LTB. A pentaglycine linker was also incorporated in between the fusion partners to permit minimal stearic hindrance for B cell recognition (Ali et al. 2013). Presence of C-terminal 6× His tag facilitated purification of the fusion proteins from insoluble fraction, which were refolded using urea gradient dialysis. Immunization resulted in high end point titers except for rEpiF5 (>1:80,000) and were comparable to that of the parent protein, rOmpF (Yadav et al. 2014). Poor immunogenic response by the rEpiF5 is expected as this region was not predicted by all the software used for analysis and validates the accuracy of antibody epitope prediction by different software. The antisera against the fusion protein rEpiF1 gave maximum absorbance in ELISA against the targeted antigen, indicating greater concentration of OmpF-specific antibodies in antisera in comparison to antisera against other fusion proteins, and it was thus voted as the most immunodominant epitope. Ability of the antisera generated against the rEpiF1to cross-react with both the rOmpF and rLTB (Western blot analysis) indicated that the animals responded to both the antigens, producing LTB- and OmpF-specific antibodies.

Further, LTB, used as a fusion partner of the OmpF epitope, is known for its immunostimulatory property, which is mediated by ganglioside GM1 receptor binding (Nashar et al. 1996, 1997, 1998; Fingerut et al. 2006; Ran et al. 2008). The GM1 binding has been reported to be essential for its adjuvant activity (Nashar et al. 1998, 2001; Kaushik et al. 2013). Analysis of the receptor binding ability of rEpiF1 showed that it could effectively recognize the GM1 receptor and thus retained the native property of LTB in the fusion protein. Oral immunization with LTB accounts for both Th1- and Th2-related cytokine expressions (Nakagawa et al. 1996) as well as substantial serum antibody titer generation (de Aizpurua and Russell-Jones 1988). In earlier reports, co-administration of LTB (Todoroff et al. 2013) or fusion of LTB with the respective antigen (Qiao et al. 2009; Zhang et al. 2009) resulted in increased mucosal immune response. Oral administration of a vaccine is always cost-effective and easy to administer. LTB in oral administration of LTB-linked plant-derived antigen has been established to help in adhesion and uptake of the antigenic peptide, thereby provoking humoral immune response generation (Companjen et al. 2006), reconfirming its significance as a mucosal adjuvant. In the present study, the OmpF epitope in the rEpiF1 is only 15 amino acid residues long and not likely to generate efficient immune response. Generation of antisera against the fusion protein, with high OmpF-specific titer (~65 % of the antisera generated against the parent OmpF, Fig. 3a) clearly indicates that LTB in the fusion protein has been able to potentiate the antibody response against the epitope. This confirms the well-established role of LTB as an adjuvant and immunomodulator (Weltzin et al. 2000).

Antibody isotyping data suggested a strong Th2-biased immune response. The antibody isotype levels were in the order IgG1 > IgG2b > IgG2a (Fig. 4a) which are in accordance with the isotype type reported by Xu et al. (2011) in B cell epitope-based vaccine evaluation against Sterptococcus agalactiae. The opsonic activity of these isotypes for both human and murine phagocytes has been estimated as IgG2a > IgG1 > IgG2b. Though IgG1 has been reported to provide an excellent opsonic activity for neutrophils (Schlageter and Kozel 1990; Maurer and von Stebut 2004), vaccination against influenza showed that while IgG1 was involved in effective neutralization of viral particles, IgG2a assisted in clearance of the virus, indicating distinct role for both the antibody isotypes (Huber et al. 2006).

Increased proliferation of the sensitized splenocytes upon stimulation with the rEpiF1 indicates activation of T cell memory. IFN-γ and IL-4 levels were significantly higher in the culture supernatants of antigen-stimulated splenocytes. The increase in serum IFN-γ levels (681 pg/ml) and IL-4 levels (143 pg/ml) after rEpiF1immunization is comparable to that observed upon immunization with YscF antigen of Yersinia pestis (Ali et al. 2013). Significant increase (p ≤ 0.01) in both the IFN-γ and IL-4, hallmark cytokines of Th1 and Th2 immune response, respectively, in culture supernatant of stimulated splenocytes from rEpiF1-immunized mice (when compared to that of non-stimulated one) suggest mixed immune response. These findings were further confirmed by carrying out cytokine analysis of the culture supernatant using a mouse cytokine antibody arrays which allow detection of 40 different cytokines and chemokines simultaneously. Increased levels of CCL3/MIP-1α, CXCL2, G-CSF, IL-16, CD54/ICAM, CCL5/RANTES, IL-1 F2, and IL-1 F3 apart from IFN-γ in the culture supernatants of stimulated splenocytes collectively further confirm mixed immune response. While CCL3/MIP-1α and CXCL2/MIP-2 modulate inflammatory responses and promote macrophage-mediated degradation (Umemura et al. 2007), G-CSF, an IL-17 F-mediated factor, plays an important role in inflammation as it is involved in the secretion of neutrophil-specific cytokines (Manz and Boettcher 2014). An increase in IL-16 in the stimulated culture supernatant is of significance in generating type 1 immune response as it plays a pivotal role in orchestrating chemotaxis (including monocytes, eosinophils, and dendritic cells) as well as inducing recruitment and activation of CD4+ Th1 cells. The IL-16 also acts as an immunomodulator as it checks excess cytotoxicity by T cell damage (Wilson et al. 2004). An increase in CCL5/RANTES was noted which co-stimulates IL-2 (a Th1 cytokine) but also regulates Th2 differentiation (Liao et al. 2011). Immunization with the rEpiF1was able to generate T cell memory as established by splenocyte proliferation assay, and hence, an increase in the levels of ICAM/CD-54’s known to be involved in T cell memory maintenance (Cox et al. 2013) is not unexpected. Apart from these cytokines, some well-known chemokines were also detected by the array blot. The other molecules which showed an increase included IL-1 F2/lymphocyte activating factor (LAF) and IL-1 F3. While the former is an important inflammatory mediator belonging to the IL-1 cytokine family, the IL-1 F3/IL-1ra is known to block IL-1 receptor binding (Barksby et al. 2007). Since both the levels increased to a similar extent, it is suggested that they act antagonistically to each other thereby keeping the inflammatory damage in control. This detailed cytokine analysis clearly indicated the involvement of Th1 branch of cell-mediated immune response, while the antibody isotyping indicated Th2-biased immune response. Thus, it can be said that the rEpiF1 fusion antigen generated a mixed immune response. This has an advantage for a vaccine candidate as it has been reported that the antigens that generated mixed immune response conferred better protection in comparison to those which induced only a Th2 response (Cheers et al. 1999).

Having established an effective immune response generated by the rEpiF1 fusion protein, neutralizing ability of the anti-rEpiF1antisera was tested by an in vitro agglutination assay. The anti-rEpiF1 antisera could agglutinate live A. hydrophila cells, while no agglutination was observed with the preimmune and anti-rEpiF2 sera. It is of interest to note that agglutination of live A. hydrophila cells is directly related with the cross-reactivity of the antisera with rOmpF in ELISA. The antisera raised against the rEpiF2 showed minimal interaction with the rOmpF in ELISA, suggesting that the epitope 201–214 is not able to generate OmpF-specific antibodies and hence failed to agglutinate A. hydrophila cells. On the other hand, the anti-rEpiF2 sera which gave significant cross-reactivity with the rOmpF could agglutinate the bacteria. FACS analysis further confirmed that the agglutination results from specific interaction of the antisera with the protein on A. hydrophila cell membrane wherein surface-exposed OmpF was specifically recognized by antibodies present in anti-rEpiF1 antisera. Ability of the anti-rEpiF1 antisera to agglutinate live cells of different Aeromonas spp. is in line with our earlier reports, where the antisera raised against the complete OmpF could cross-react with a number of Aeromonas strains. The rEpiF1 can thus be evaluated as a potential vaccine against other Aeromonas species as well.

The present study reports potential of a recombinant outer membrane protein OmpF epitope (OmpF66–80) in translational fusion with LTB as a vaccine candidate against A. hydrophila and provides a detailed analysis of immune response generated against the same. Immunization with the fusion protein having LTB as a fusion partner resulted in immune sera of high end point titer which was agglutination positive. The LTB in the fusion protein conferred adjuvant activity and the ability of the fusion protein rEpiF1 to bind to GM1 receptor suggests its potential as an oral vaccine.