Yeast Surface Display of Capsid Protein VP7 of Grass Carp Reovirus: Fundamental Investigation for the Development of Vaccine Against Hemorrhagic Disease

Abstract

VP7, an outer capsid protein of grass carp reovirus (GCRV), was expressed and displayed on the surface of Saccharomyces cerevisiae for developing an efficient vaccine against hemorrhagic disease of grass carp. The result of flow cytometry analysis indicated that protein VP7 could be displayed on the surface of yeast cells after inducing with galactose. The expression of VP7 was confirmed by western blot analysis and further visualized with confocal microscopy. The specific antibodies against VP7 generated from mice were detectable from all immune groups except the control group, which was immunized with untransformed yeast cells. The displaying VP7 on glycosylation-deficient strain EBYΔMnn9 was detected to induce a relatively low level of specific antibody amongst the three strains. However, the antiserum of EBYΔM9-VP7 showed relative high capacity to neutralize GCRV. Further neutralization testing assays indicated that the neutralizing ability of antiserum of the EBYΔM9-VP7 group appeared concentration dependent, and could be up to 66.7% when the antiserum was diluted to 1:50. This result indicates that appropriate gene modification of glycosylation in a yeast strain has essential effect on the immunogenicity of a yeast-based vaccine.

Introduction

Grass carp reovirus (GCRV), a pathogen of grass carp (Ctenopharyngodon idella) hemorrhagic disease, can cause a mortality rate of up to 80% during an outbreak [18]. It also can infect black carp (Mylopharyngodon piceus) [10] and rare minnow (Gobiocypris rarus) [46]. GCRV was recognized to be the most virulent aquareovirus so far [25]. To prevent the spread of grass carp hemorrhagic disease, several kinds of vaccines have been developed, which could effectively prevent the spread of grass carp hemorrhagic disease to some extent [42,50]. However, development and manufacture of these vaccines are time consuming and expensive, and difficulties in the application of these vaccines to grass carp in the aquatic environment also meet with many limitations in the field. As GCRV has the potential to cause death in a wide variety of host fish, it is very important to find a simple, reliable, and effective method to produce large quantities of cheap and effective vaccines for the control of GCRV infection. Thus, it is necessary to find new methods to produce large quantities of cheap vaccines.

Saccharomyces cerevisiae can be an excellent vector to produce vaccine owing to its GRAS (generally regarded as safe) status, clear genetic background, easy cultivation, and cheap production [31,43]. The cell surface is a functional interface between the inside and the outside of the cell, where natively displayed molecules on the surface play important roles in various biological or physiological phenomena [35]. A special expression vector was developed to display the heterologous protein on the surface of yeast cells utilizing covalently linked GPI-anchored proteins, which is called yeast surface display technology [5] and has been widely used in bioconversion, vaccine, antibody development, and protein engineering [40]. Recombinant antigens were expressed as yeast surface protein by fusing to cell wall proteins and used as a whole-cell live vaccine. A yeast displaying cell live vaccine was proposed bySchreuder et al. [31], who displayed antigen of hepatitis B virus on the surface of S. cerevisiae and used it as antigen to immune mice, which resulted in a weak but very specific anti-HbsAg response. Yeast surface display antigens were proved to be efficient potential vaccines in the application of highly pathogenic avian influenza virus hemagglutinin, hemolysin from Vibrio harveyi and the 380R protein from red sea bream iridovirus [39,47,51]. More recently, an oral vaccine against candidiasis was also successfully developed based on yeast surface display technology [34].

In order to find out significant proteins of GCRV to produce subunit vaccines, several studies on the antigenicity of GCRV were performed. The antiserum generated from grass carps infected with GCRV could only recognize outer capsid protein VP7 of GCRV, which indicates that VP7 is an essential antigenic protein of GCRV [8]. In another report, both antisera generated from GCRV capsid proteins VP5 and VP7, respectively, showed neutralization ability in GCRV-infected CIK (Ctenopharyngodon idellus kidney) cells. Furthermore, the neutralizing capacity activity of VP7 antiserum was three times higher than that of VP5 antiserum, which also indicates that outer capsid protein VP7 might be the dominating antigen of the virus [33]. Therefore, in the present study, VP7 of GCRV was chosen to be expressed in S. cerevisiae and displayed on the surface of EBY100, and the glycosylation deletion mutant strains EBYΔMnn1, EBYΔMnn9, and EBYΔOch1. Intraperitoneal administration of these yeast cells in mice resulted in a high level (titer) of specific anti-VP7 antibodies. The neutralization abilities of the prepared antibodies to GCRV infection were also tested.

 

Materials and Methods

Plasmids, Strains, and Media

Escherichia coli DH5α and BL21 (DE3) were used for cloning and expression, respectively. Yeast strain EBY100 and plasmid pYD1 were purchased from Invitrogen (Carlsbad, CA, USA). LB medium and plates, YPD, minimal dextrose plates (with Amino acids), and YNB-CAA medium were prepared as recommended by Invitrogen. S. cerevisiae BY4741ΔMnn1 (MATa his3Δ0 leu2Δ0 met15Δ0 ura3Δ0 mnn1Δ::kanMX4) was kindly provided by Professor Xiao Wei (University of Saskatchewan, SK, Canada).

Construction of Hyperglycosylated Gene Deletion Strains

The hyperglycosylated gene deletion in S. cerevisiae was performed according to the strategy described by Baudin et al. [3]. The mnn1Δ::KanMX4 fragment containing the kanamycin cassette flanked with part of the MNN1 gene was amplified from genomic DNA of S. cerevisiae BY4741ΔMnn1 using the primers F-mnn1 and R-mnn1 (Table S1). S. cerevisiae BY4741ΔMnn1 is a MNN1 gene-deficient strain that was constructed by Professor Xiao’s laboratory. The primers F-mnn9, R-mnn9, F-och1, and R-och1 were designed to amplify mnn9Δ::KanMX4 and och1Δ::KanMX4 fragments, which contain the kanamycin resistance marker KanMX4 sequence flanked with the corresponding homologous arms. All the primers share parts of the sequence with their target genes at the 5’ end and complement with KanMX4 at the 3’ end. The polymerase chain reaction (PCR) procedure was as follows: initial denaturation at 94ºC for 10 min, followed by 30 cycles of 1 min at 94ºC, 30 sec at 55ºC, and 1 min at 72ºC, and 10 min at 72ºC for final extension. The final products, mnn1Δ::KanMX4, mnn9Δ::KanMX4, and och1Δ::KanMX4 fragments, contain the kanamycin resistance marker flanked at the 5’ and 3’ ends by 5’ and 3’ target gene fragment, respectively. The final products were transformed to S. cerevisiae strain EBY100 (Invitrogen) by the method of Gietz et al. [14] Transformants were selected on YPD containing 150 μg/ml G418, and single clones were picked up and cultured for preparing genomic DNA. Yeast genomic DNA was isolated according to the method described by Rose et al. [28]. PCR and DNA sequencing were applied to confirm whether the target genes were successfully deleted.

Construction of Vector and Transformation of Yeast

The sequence of GCRV-873 VP7 was amplified from pR/GCRVVP7 plasmid [49] using primers CCGGGATCCATGCCACTTCAC ATGATTCC and CCGCTCGAGATCGGATGGCTCCACATGCA containing restriction sites for BamHI and XholI (underlined). The amplified gene was cloned into vector pMD18-T (Takara, Dalian, China), analyzed by restriction enzymes digestion, and confirmed by DNA sequencing. The resulting amplified DNA fragment was separated by electrophoresis on 1.5% agarose gel, purified and cloned into BamHI/XholI predigested plasmid pYD1, analyzed by PCR, and confirmed by DNA sequencing. The recombinant pYD1/VP7 plasmid was transformed to S. cerevisiae strain EBY100 and three hyperglycosylated gene defect strains by the lithium acetate procedure. Transformants were selected on minimal plates. The single clones were picked out and cultured in YNBCAA medium with 2% glucose, and then the yeast cells were harvested and added to lysis buffer (10 mM Tris, 1 mM EDTA, 100 mM NaCl, 1% SDS, and 2% Triton X-100, pH 8.0). The recombinant plasmid was extracted from yeast cells according to the method that we utilized to isolate genomic DNA. PCR was used to detect the VP7 gene in the extraction of yeast cells.

Inducing VP7 to Express on the Surface of the Yeast

A single colony of EBY100-VP7, EBYΔM1-VP7, EBYΔM9-VP7, and EBYΔO1-VP7 was respectively inoculated into 10 ml of YNBCAA containing 2% glucose and grown overnight at 30ºC with shaking. The absorbances of the cell cultures were determined by a SP-723 spectrophotometer (Shanghai Spectrum Instruments Company, Shanghai, China) at 600 nm. When the OD6600nm reached between 2 and 5, the cell cultures were centrifuged at 3,000 ×g at room temperature. The cell pellets were resuspended in YNBCAA containing 2% galactose. The OD600nm of the cell cultures was changed to 0.5 to 1. The cell cultivation with 2 OD600nm units was removed and placed on ice immediately; this point was defined as time zero. The rest of the cells were cultured at 20ºC with shaking, and then 2 OD600 units of cell culture was removed every 12 h. At the end of the induction process, all these samples were analyzed by flow cytometry (FACS Aria; Becton Dickinson, NJ, USA) after staining with mouse anti-VP7 protein antibody and fluorescein dylight488-conjugated goat anti-mouse IgG to determine the optimal induction time for maximum display.

A volume of the yeast cells equivalent to 10.0 OD600nm units was harvested and washed twice with PBS buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, and 1.8 mM KH2PO4, pH 7.4) after EBY100/pYD1-VP7 and other strains were induced in YNB-CAA medium containing 2.0% galactose for 24 h. These cells were stored at -20ºC and prepared for SDS–polyacrylamide gels and western blot analysis.

SDS–Polyacrylamide Gels and Western Blot Analysis

The microcentrifuge tubes containing yeast cells were added to 100 μl of ultrapure water and 100 μl of 2× loading buffer (80 mM Tris, 2% SDS, 10% Glycerol, 0.0006% Bromophenol blue, and 0.1 M dithiothreitol, pH 6.8). After mixing the cells using a vortex mixer, the tubes were placed in a 100°C (boiling) water bath for 10 min. It is possible to release the Aga2p protein fusion from the cell wall by treatment with 0.1 M dithiothreitol (DTT) in buffer for 10 min treatment in boiling water. Proteins were separated by SDS–polyacrylamide gel electrophoresis (SDS–PAGE) according to the method of Laemmli [19] on 12% polyacrylamide gels. After being separated by SDS–PAGE, the proteins were transferred onto nitrocellulose membranes using a semi-dry transfer cell (Bio-Rad, CA, USA). Membranes were blocked by 4% PBSM (4% skim milk in PBS) and incubated at 30°C for 1 h with His tag antibody HRP conjugate. The membrane was washed three times with PBST (0.1% Tween-20 in PBS) and PBS, respectively. Subsequently, the antibody binding was visualized by the ECL method with FluorChem Imaging System Q (Alpha Innotech Corporation, San Leandro, USA).

Immunofluorescence Microscopy

A small volume of induced cells was collected and washed twice with PBS buffer. The cells were resuspended in 100 μl of 4% PBSM containing the mouse anti-VP7 protein antibody at a dilution of 1:500 and incubated for 1 h at room temperature. The cells were washed three times with PBS and resuspended in 50 μl of 4% PBSM containing fluorescein dylight488-conjugated goat anti-mouse IgG (Amyjet Scientific Inc., Wuhan, China) at 1:100 dilution, and incubated for 1 h at room temperature. After washing out the unreacted antibodies with PBS for three times, the cells were visualized with a Zeiss LSM 710 confocal microscope (Carl Zeiss AG, Oberkochen, Germany).

Mice Immunization and Sample Preparations of Antiserum

The induced yeast cells were washed with sterile PBS buffer by spin. The cells were resuspended in PBS buffer and adjusted to a series of different cell concentrations of 5 × 109, 5 × 108, 5 × 107, and 5 × 106 cells/ml. Six-week-old SPF BALB/c mice were intraperitoneally injected with 0.2 ml of yeast cell suspension. Untransformed EBY100 cells at the same concentration gradient were injected as a negative control. In each group, three mice were used. The 2nd and 3rd immunizations were performed after 2 and 4 weeks with the same amount of yeast cells. Ten days after the last immunization, the blood was collected from the tail vein of each mouse and tested for antibodies against VP7.

Both transformed and untransformed standard EBY100 and glycosylation-deficient strains were administered intraperitoneally to mice at a concentration of 5 × 108 cells/ml. The immunization procedure was performed the same way as we did in the former immunization. After the last immunization, the mice blood was collected and also tested for antibodies against VP7.

ELISA

Wells of ELISA plates were coated with VP7 expressed in E. coli at a concentration of 1 μg/ml, adding 100 μl to each well and incubating at 4ºC overnight. After all of the coated wells were blocked with 4% PBSM, 100 μl of 4% PBSM containing different dilutions of the sera were added to the wells and incubated for 1 h at 37ºC. The plate was washed with PBST and PBS for three times, respectively . Goat anti-mouse I gG w ith HRP c onjugate (Boster, Wuhan, China) was added to the wells and incubated for 1h at 37ºC. After washing out the unbound antibody with PBST and PBS, the substrate 3,3’,5,5’-tetramethylbenzidine (Serva, Heidelberg, Germany) was added to the wells. The reaction was stopped by adding 2 M H2SO4. The absorbance was determined at 450 nm using a spectrophotometer (BioTek, Seattle, WA, USA).

Neutralization Assay of Prepared Antibodies

The infectivity of antibody-treated virus was determined by plaque assays. GCRV-873 virus samples were subjected to a series of 10-fold serial dilutions in MEM (mininum essential medium) for titer determination. Plaque assays were done by determining the plaque forming unit (PFU) value per milliliter according to the method indicated before [13], and the relative infectivity of different particles was also measured using endpoint experiments by calculating the 50% tissue culture infective dose (TCID50) as described previously [12,25]. Four original antisera samples induced by EBY100, EBY100-VP7, EBYΔM1-VP7, and EBYΔM9-VP7 were filtered, respectively. In addition, the antisera of EBYΔM1-VP7 and EBY100 were serially double diluted from 1:50 to 1:400 with MEM and incubated with GCRV. Ctenopharyngodon idellus kidney cells were infected with 100 TCID50 GCRV virions, as the positive control, and mock-infected cells served as the negative control. For the neutralization assay, 100 μl of diluted antiserum and 100 μl of virus (containing 100 TCID50 virions) were added to each well of 24-well microtiter plates and mixed gently, followed by incubating the mixtures at 28°C for 60 min. The mixtures were added to each well of 24-well microtiter plates containing CIK cells and incubated at room temperature for 30 min, followed by washing cells with MEM for two to three times. The detailed neutralization test was then performed as described elsewhere [26,45]. The neutralizing capacity of the polyclonal antibodies was evaluated as the percentage of plaque reduction. All the data obtained in this assay were the means obtained from triple determinations for each sample, with three repeats.

 

Results

Construction of Hyperglycosylated Gene Deletion Mutant

In order to reduce the influence of hyperglycosylation of foreign proteins caused by yeast, three hyperglycosylated genes MNN1, MNN9, and OCH1 were deleted from yeast EBY100 strains and the mutants were named EBYΔM1, EBYΔM9, and EBYΔO1, respectively. As shown in Fig. 1A, the kanamycin resistance cassette is flanked by a 5’ and 3’ part of the target gene, allowing its integration at the locus of the target gene by homologous recombination (Fig. 1A). Transformants collected from YPD containing G418 were analyzed by the PCR method. Primer F-M1 was designed according to the upstream sequence of gene MNN1, whereas the primer KanMX4 could complement to gene KanMX4. If the KanMX4 gene was integrated to the locus of gene MNN1, part of gene KanMX4 from yeast genomic DNA would be amplified. The band on lane 2 in Fig. 1B confirmed the deletion of MNN1 in strain EBYΔM1. The primers F-M9, R-M9, F-O1, and R-O1 were designed according to the sequences of genes MNN9 and OCH1, and therefore they could be applied to amplify the whole genes of MNN9 and OCH1. Genes MNN9 (about 1.2 kb) and OCH1 (about 1.5 kb) were detected when wild-type EBY100 genomic DNA was used as template for replication, whereas no band appeared in the mutant strains (Fig. 1B). The result confirmed the successful deletion of MNN9 and OCH1 genes in EBYΔM9 and EBYΔO1, respectively. Subsequent DNA sequencing further confirmed the successful construction of these three hyperglycosylated defective gene strains.

Fig. 1.Construction of hyperglycosylated gene deletion strains.

Time-Course Induction of Yeast Displayed Protein and Its Stability at Different Storage Temperatures

In order to examine the time-course induction of yeast-displayed VP7 protein, different inducing time points were chosen for the expression of Aga2-VP7 fusion protein on the surface of the standard EBY100 containing the recombinant vector pYD1/VP7, by flow cytometry. Results showed that VP7 protein could be detected from 12 to 48 h after inducing with YNB-CAA containing 2% galactose. The peak value of mean fluorescence was 256 at the inducing time point of 24 h (Fig. 2A). Thus the time point of 24 h was probably the best time for inducing protein VP7 expressed on the surface of yeast cells. In our subsequent experiments, yeast cells were induced with 2% galactose for maximal expression of VP7.

Fig. 2.The optimal induction time and storage temperature of strain EBY100-VP7.

In the present study, because we tried to use yeast surface displayed VP7 as live vaccine, it is necessary to find out the time point at which VP7 protein starts to degrade. The EBY100-VP7 yeast cells were placed at different storage temperatures to detect the stability of the fusion protein Aga2-VP7. Western blot analysis was used to test whether the heterologous proteins still displayed on the surface of yeast cells (Fig. 2B). We observed that heterologous proteins degraded obviously by the 8th day at the temperature of 30ºC and by 2 months at 4ºC. VP7 still displayed on the surface of yeast cells after placing for 3 months at temperature of -20ºC and -80ºC. Therefore the induced yeast cells were suggested to be stored at a lower temperature for protecting VP7 from proteolytic degradation.

VP7 Fusion Protein Could Be Displayed on the Surface of Yeast Cells

The expressed Aga2-VP7 fusion protein on the surface of the yeast was also detected by western blot assay, using an antibody specific against the 6His tag. As expected, 6His tag fused with the C terminal of VP7 protein was detected in the extracts of cell walls from all the constructed strains induced with 2.0% galactose. Meanwhile, no signals were detected in the extracts of cell walls from the non-induced strains (Fig. 3A). Western blot results showed that the molecular masses of the fused VP7 proteins in all constructed yeast strains were larger than theoretical molecular mass of Aga2-VP7 protein. When glycosylation-related genes MNN1, MNN9, and OCH1 were deleted, the VP7 protein bands almost had the same molecular mass. Therefore, we deduced that the protein VP7 was not hyperglycosylated in yeast.

Fig. 3.Detection of VP7 displayed on the surface of yeast cells.

In this study, protein VP7 was fused with Aga2. With the help of the signal peptide of VP7, the fused protein located on the surface of yeast cells. To further verify that the Aga2-VP7 fusion protein was associated with the yeast cell walls, immunofluorescence microscopy was used to visualize the location of the fusion protein (Fig. 3B) by using mice antibody specific against VP7 protein. No signal was detected on the surface of non-induced yeast cells. The detectable fluorescence on the surface of induced yeast cells confirmed that VP7 had been displayed successfully on all of the four yeast constructs. Within them, the EBYΔM9 and EBYΔO1 strains presented a challenge for complete separating after budding. These cells grew as a population of clumps of up to dozens of cells each. Since EBYΔM9 and EBYΔO1 almost have the same cell morphology and the same molecular mass of VP7 expressed in those cells, only EBY100, EBYΔM1, and EBYΔM9 were chosen for subsequent experiments.

Immunogenicity Assays of VP7 Expressed in Yeast Cells

For preparative experiment to test the immunogenicity of the yeast-displayed VP7, several groups of three female mice were injected intraperitoneally with whole cells of standard EBY100 expressing VP7 adjusted to a series of concentrations. Control group was inoculated with untransformed strain EBY100. After completing the immunization protocol, mice were bled, and the specific antibodies against the VP7 expressed in E. coli were tested by ELISA. Specific antibodies against VP7 from the serum of mice injected with whole yeast cells at the concentration of 5 × 109 cells/ml were detected after the first immunization (Fig. 4A). However, this concentration of yeast cells was too high to be tolerated by the injected mice. All of the three injected mice died after the second immunization. The mice inoculated with 5 × 108 cells/ml produced specific antibody after the second immunization, and with higher level of specific antibody after the 3rd immunization. Meanwhile, no specific antibody was detected from the serum of mice immunized with yeast cells of less than 5 × 108 cells/ml. These results indicated that the amount of inoculated yeast-displaying vaccine was critical for production of specific antibody against the target antigen. In the present case, 5 × 108 cells/ml of yeast cells is the most appropriate concentration for injection to produce specific antibody against VP7 protein in mice.

Fig. 4.lmmunogenicity assay of VP7 protein expressed in yeast cells.

Both transformed and untransformed standard EBY100 and glycosylation-deficient strains were also administered intraperitoneally to mice at a concentration of 5 × 108 cells/ml, and the serum was collected after the 3rd immunization. The specific antibodies against VP7 from injected mice were tested by ELISA (Fig. 4B). EBY100-VP7 and EBYΔM1-VP7 cells induced stronger responses compared with a very weak response of EBYΔM9-VP7 cells. No positive reaction was detected in mice inoculated with untransformed EBY100 cells.

Neutralization Assay

As indicated above, all of the mice injected with whole yeast cells displaying VP7 produced specific antibodies against protein VP7. To evaluate the neutralization ability of the prepared antibodies to GCRV infection, neutralization assays were performed with EBYΔM9-VP7 and the three other mice antisera. As shown in Fig. 5A, it appeared that the number of viral plaques could be severely reduced by the antiserum collected from mice injected with the EBYΔM9-VP7 strain, but there were no significant changes with the three other mice antisera of EBY100, EBY100-VP7, and EBYΔM1-VP7, which suggests that only the antiserum of strain EBYΔM9-VP7 can strongly inhibit the infection of GCRV into its host cells.

Fig. 5.Neutralization test of polyclonal antisera in CIK cells.

To further prove the neutralization ability of EBYΔM9-VP7 antiserum, serial double-diluted antisera from 1:50 to 1:400 were mixed with GCRV, respectively, followed by adding to the CIK cells. The result (Fig. 5B) showed that there was obvious plaque reduction with EBYΔM9-VP7 group antiserum, especially with the dilution fold of 1:50, but no changes of plaque reduction could be observed with the EBY100 control serum. Further relative neutralization capacity was evaluated by calculating all the plaque amount in the assays. It is shown in Fig. 5C that the neutralizing ability of the antiserum was concentration dependent, and could be up to 66.7% with EBYΔM9-VP7 in its neutralization capacity when the antiserum was diluted 1:50 (Fig. 5C). Taken together, the result suggests that EBYΔM9-VP7 antibody has an ability to neutralize intact GCRV particles.

 

Discussion

Live vaccines have received increasing attention owing to their convenience and effectiveness [2,6,24]. In the case of developing a virus vaccine, a yeast-based vaccine is a good choice. Comparing with bacteria, yeast is more appropriate to express viral proteins owing to its glycosylation of proteins. This is very important because glycan moieties are an important factor in generation of neutralizing antibody [15,21,27]. The recombinant proteins produced in yeast have higher immunogenicity when administered with the yeast cell wall components [38]. Furthermore, β-glucans from S. cerevisiae can be used as adjuvants for intradermal and oral immunizations [4]. Proteins expressed on the cell surface are accessible to the immune system; even very small peptides can be immunogenic when displayed on a larger surface [32].

However, the glycosylation system in yeast is a double-edged sword for the development of yeast-displayed vaccine. On one side, the immunogenicity of glycosylated antigen was maintained for generating antibodies effectively. On the other side, the hyperglycosylation of yeast cell wall proteins may block the displayed antigen to be recognized exactly by the immune system of the host animal. Many proteins in S. cerevisiae are modified by adding O-linked mannose glycans to serine or threonine, and N-linked saccharides to asparagine. O-Linked glycosylation has minor influence because O-glycosidically linked linear glycans only consist of 1-5 mannose residues attached to serine or threonine of some proteins in yeast [17]. However, some N-linked oligosaccharides of hypermannosylated glycoproteins consist of up to 200 mannose residues [9]. N-Linked oligosaccharides consist of a core structure and an outer chain of variable size, which consists of a backbone of α-1,6-linked mannose residues and branches of α-1,2 and α-1,3-mannose-containing side chains. This process is initiated in Golgi α-1,6-mannosyltransferase (Och1p), adding a first α-1,6-linked mannose to the core oligosaccharide Man8GlcNAc2 [23]. After the addition of α-1,6-linked mannose by Och1p, the linear backbone of the outer oligosaccharide chain is extended by additional α-1,6-mannosyl transferases encoded by gene MNN9 and further branched by the addition of α-1,3-linked mannoses encoded by gene MNN1 [36]. The terminal α-1,3-linked mannose units that are added to the outer oligosaccharide chain are particularly immunogenic, and the serum collected from rabbits injected with whole yeast cells contains a large amount of antibodies that recognize this α-1,3-linked mannose epitope [9]. N-Linked glycosylation can only constitute a small fraction of the molecular mass for a given protein, while it significantly changes the hydrodynamic volume of the protein and therefore its pharmacodynamic behavior [48]. With the understanding of the glycosylation process in S. cerevisiae, we disrupted three hyperglycosylation-related genes by homologous recombination and constructed three hyperglycosylation gene deletion strains for development of yeast display vaccines. No morphological or physiological difference was observed in stain EBYΔM1 comparing with wild-type EBY100, whereas the EBYΔM9 and EBYΔO1 strains presented slower growth rates, increased clumpy growth, and aggregation. The aggregations of these yeast cells might be caused by the hyperglycosylation defects of cell wall proteins in both strains. The EBYΔO1 strain can only add core oligosaccharide Man8GlcNAc2 to the asparagine-linked glycosylation sites, whereas the EBYΔM9 strain can add one more mannose to the core oligosaccharide. Without the hyperglycosylation of cell wall proteins, the yeast cells tend to clump after budding [11].

In this paper, we have shown that protein VP7 of GCRV873 could be expressed as a fusion protein on the surface of S. cerevisiae. The result of flow cytometry analysis indicated that VP7 displayed on the surface of yeast cells after inducing with 2% galactose. The peak value of mean fluorescence intensity appeared 24 h after induction. Only part of S. cerevisiae cells successfully express VP7 on the cell surface. Unlabeled cells represent the non-displaying fraction of yeast clones due to plasmid loss [7]. Growth conditions in galactose have an effect on the efficiency of the display, or the sequence of the protein or peptide fused to the Aga2p protein is another factor that influences the efficiency of the display [1].

Proteolytic degradation has been a problem since yeast was employed to express heterologous protein [44]. The yeast vacuole, which is homologous to the lysosome of higher eukaryotic cells, is filled with various proteases [30]. The heterologous proteins expressed in yeast cells are easily degraded by protease. We found that the heterologous proteins were degraded obviously on the 8th day at the temperature of 30ºC. In order to produce a potential live vaccine, some measures should be taken to slow down the process of degradation.

The fusion proteins were also confirmed by western blot analysis and microscopic observation. In the result of western blot analysis, we found that the VP7 expressed in EBY100 and three glycosylation-deficient strains almost had the same molecular mass, which was larger than the theoretical molecular mass of Aga2-VP7. When analyzed with the NetOGlyc 4.0 Server [37] and NetNGlyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/), four O-glycosylation sites and one potential N-glycosylation site were found in the amino acid sequence of VP7. It seems that the single N-glycosylation site was not hyperglycosylated, since the major bands of VP7 expressed in the EBY100 and three N-glycosylation-deficient strains have the same molecular masses. Some studies showed that not all the heterologous proteins with N-glycosylation sites needed to be hyperglycosylated [16,20,22]. Therefore we speculated that the enhancement of VP7 molecular mass in yeast was caused by O-glycosylation, theoretically. Interestingly, despite of the lowest specific antibody titer to VP7 because of cell aggregation, the serum from EBYΔM9-VP7 was the only serum sample with neutralizing ability amongst the three constructed yeast vaccines-induced serum. This indicated that modification of the hyperglycosylation-related gene is meaningful for the development of an efficient yeast-displayed vaccine.

Since the glycosylation of VP7 is not the major cause for the production of neutralizing antibodies in EBYΔM9, the hyperglycosylation of yeast cell wall proteins was deduced as the major reason for the different neutralizing ability within the serum induced by EBY100-VP7, EBYΔM1-VP7, and EBYΔM9-VP7. Plenty of long glycan chains of cell wall proteins may block the neutralization sites of VP7 displaying on the yeast surface in EBY100-VP7 and EBYΔM1-VP7. Because of the deletion of gene MNN9, hyperglycosylation of any protein in yeast was not possible. Therefore, all active antigenic sites, including the neutralizing sites, were exposed for immune response, except the sites that were blocked by cell aggregation. Hence, the serum induced by EBYΔM9-VP7 contains the neutralizing antibodies that inhibited the infection of GCRV in CIK cells. To expose the VP7 better on the surface of the cell wall, longer carboxyterminal fragments derived from FLO1 [41] or AGA1 [29] might be used as cell wall anchors. The probable reason of low level of antibody against VP7 induced by EBYΔM9-VP7 cells might be the cell aggregation. It is difficult for the mouse immune system to approach VP7 displaying on the surface of aggregated yeast cells.

In this study, using a modified glycosylation yeast strain, VP7 was successfully expressed and displayed on the surface of S. cerevisiae. The immunogenicity and neutralizing capacity of yeast-displayed VP7 vaccine EBYΔM9-VP7 has also been proved. Therefore, it has potential to be developed as an efficient and convenient live vaccine against hemorrhagic disease of grass carp. Subsequently, the immunoprotection ability of this novel vaccine will be evaluated in grass carp by injection and oral feeding of the transformed EBYΔM9-VP7 yeast cells.