Characterization of mayor capsid protein VP7 of African horsesickness virus as an antigen display system

Abstract

Due to the poor stability and immunogenicity of small antigenic peptides, it is necessary that they be presented to the immune system in conjugation with larger carrier proteins or on the surface of chimeric viral or virus-like particles in order to elicit an immune response. When the major capsid protein VP7 of African horsesickness virus (AHSV) is expressed in Spodoptera frugiperda (Sf9) insect cells by means of a recombinant baculovirus strategy, the VP7 trimers aggregate and are assembled into unique, distinctive crystal-like particles. Immunization of mice with AHSV-9 VP7 crystal-like particles has been found to protect 80-100% of the animals against virulent, heterologous AHSV-7 challenge (Wade-Evans et al., 1997). These results suggested the possibility of using such particles for the immune display of foreign antigenic peptides. Natural hydrophilic loops within the AHSV-9 VP7 proteins’ top domain were targeted for the display of foreign peptides by inserting multiple cloning sites into three different positions within the gene encoding the VP7 protein. In this study we explored the feasibility of using the AHSV major capsid protein VP7 as an antigen display system by investigating how peptide insertions into the VP7 top domain affect VP7 particle formation, solubility, trimer formation and immune display. To generate the VP7 vectors with multiple cloning sites, small 6 amino acid peptides were inserted downstream of amino acids 144, 177 and 200 within the VP7 top domain in previous studies. These VP7 vectors were subsequently used for the insertion of a range of peptides that varied in size from 11 to 150 amino acids. These peptides were derived from outer capsid protein VP1 of foot-and-mouth-disease virus (FMDV), the major neutralization-specific antigen of AHSV (VP2) and well as the envelope protein gp41 of HIV-1. The largest sequence to be inserted into the VP7 top domain was the 239 amino acid enhanced green fluorescent protein (eGFP). To address the question of the effect of foreign peptide insertion on AHSV-9 VP7 particle formation and protein solubility, the expressed chimeric VP7 proteins were analyzed by sucrose density gradient sedimentation analysis. The effect of these peptide insertions on the ability of the AHSV-9 VP7 protein to form trimers was investigated by either a low density sucrose sedimentation strategy or a modified SDS-polyacrylamide gel electrophoresis strategy. The ability of an inserted peptide to elicit an immune response, when presented on the surface of the VP7 trimer, was determined by inoculating the chimeric VP7 fusion protein into guinea pigs and analyzing the humeral immune response elicited by means of insert-specific ELISA and neutralization assays. Since these insertions affected VP7 protein solubility to a significant degree, an important question that was addressed was whether the chimeric VP7 particles or the soluble chimeric VP7 trimers presented the best display strategy for eliciting a strong insert-specific humeral immune response. The effect of various solubilization strategies, such as protein sample dilution and treatments with the solubilization agents L-arginine and sarkosyl, on chimeric VP7 protein solubility were also investigated as well as the effect of these treatments on the immunogenicity of the inserted peptide. The insertion of different 6 amino acid peptides downstream of amino acids 144, 177 and 200 in the AHSV-9 VP7 top domain during vector construction affected VP7 protein solubility. The inserted peptide and insertion site were both found to play a role. The insertion of a 6 amino acid peptide downstream of VP7 amino acid 144 converted the hydrophobic VP7 protein to a largely soluble protein whereas the insertion of such a peptide downstream of amino acid 200 had no apparent effect on solubility. Although the ability of the VP7 protein to form trimers was not affected, the stability of these trimers under strong denaturing conditions, such as in the presence of SDS at room temperature, was dependent on the site of insertion. Trimers with peptides inserted downstream of VP7 amino acid 177 were the most stable whereas the insertion of peptides downstream of amino acid 200 resulted in the least stable trimer molecules. The insertion of larger foreign peptides into the VP7 vector proteins also influenced protein solubility. Insertion of a 36 amino acid peptide from the FMDV VP1 protein (aa129-164) into site 177 yielded the largely soluble chimeric VP7-177(P1)-FMDVepi protein. A marked reduction in solubility was observed when the same peptide was inserted into a different site. VP7 solubility and particle formation was also strongly affected by the length and amino acid composition of the inserted peptide. This demonstrated that both the VP7 top domain insertion site and the nature of the inserted peptide influence the overall equilibrium between the soluble VP7 chimeric trimers and VP7 particles. It is, as yet, not possible to predict chimeric VP7 protein solubility and the solubility of each construct needs to be individually assessed. The insertion of these larger peptides did not affect the ability of the VP7 fusion proteins to form trimers but trimer stability was again affected as described in the case of the VP7 vector proteins. The only exception was the VP7-200(P1)-eGFP protein, the majority of which occurred in a monomeric rather than in a trimeric form. As both the VP7-144(P2)-eGFP and VP7-177(P1)-eGFP fusion proteins formed trimers, this indicated that the AHSV VP7 trimer is a stable molecule capable of accommodating even large insertions within its top domain without losing its ability to form trimers. A comparison of the insert-specific humeral immune response elicited against the FMDV VP1 peptide (aa129-164) presented either on VP7-FMDVepi particles or soluble VP7-FMDVepi trimers indicated that, whereas the particles elicited a weak FMDV-specific, neutralizing humeral immune response in guinea pigs, the soluble trimers induced a much stronger immune response. This result focused much of our investigation away from the use of insoluble AHSV VP7 particles as an antigen display system to the use of soluble AHSV VP7 trimer molecules. The large quantity (6.5mg) of soluble VP7-177(P1)-FMDVepi trimers harvested from 1x108 baculovirus-infected insect cells made this a feasible strategy. In order to increase the soluble component of less soluble VP7 proteins, various strategies were investigated. One such strategy involved the substitution of an uncharged leucine residue in position 345 within the bottom domain of the AHSV VP7 protein, with a positively charged arginine residue, the bluetongue virus (BTV) VP7 equivalent. Not only did this mutation significantly increase the solubility of the modified VP7 proteins, their ability to form stable trimers was not affected. This modification may thus be incorporated into future vector proteins to increase VP7 protein solubility. Treatments aimed at reversing VP7 trimer-trimer aggregation were also investigated as a means of increasing VP7 protein solubility. Dilution of the protein sample and/or treatment with the amino acid L-arginine (Tsumoto et al., 2004) was found to reverse the loose aggregation of correctly folded VP7 fusion trimer molecules while treatment with the ionic detergent Na+N-Lauroylsarcosine (Sarkosyl) solubilized both correctly folded and incorrectly folded trimer aggregates. Inoculation of the chimeric VP7 antigen in the presence of L-arginine or sarkosyl elicited a more consistent, strong immune response than untreated, soluble VP7-177(P1)-FMDVepi trimers. This indicated the importance of presenting soluble, non-aggregated antigen to the humeral immune system and that these treatments did not negatively affect antigen immunogenicity. Although both the L-arginine and sarkosyl treatments elicited equivalent humeral immune responses against the VP7-177(P1)-FMDVepi trimers, L-arginine is not required to be removed from the immunogen solution prior to inoculation and, as this amino acid is capable of solubilizing aggregates of correctly folded trimers, it is currently the preferred solubilization protocol. The eGFP protein folds equally well into its functioning, fluorescing conformation when inserted into any of the three top domain sites. Thus peptides inserted into these sites would most likely have the opportunity to fold correctly into their native conformations. It was, however, observed that the chimeric VP7-144-eGFP fusion protein, although apparently correctly folded and fluorescing, had a tendency to form loosely assembled aggregates that were readily solubilized by simple dilution or treatment with L-arginine. This was in contrast to the VP7 fusions with eGFP inserted at sites 177 and 200, which both had smaller soluble fluorescing fractions, and were not significantly affected by these solubilization treatments. The tendency of correctly folded VP7-144-eGFP fusion trimers to assemble into loosely aggregated particles is also observed with other inserted peptides, such as the AHSV VP2 110 amino acid peptide, and has formed the basis of a low speed centrifugation strategy designed to harvest large quantities of correctly folded, soluble VP7 fusion trimers. The VP7-HIV fusion proteins were found to be largely insoluble when expressed in insect cells via a recombinant baculovirus strategy and were subsequently explored with respect to their potential to detect HIV antibodies in HIV patient sera. The different binding profiles observed on immune blots against a range of VP7-HIVgp41 fusion proteins could be used to characterize HIV-1 antisera as either subtype B or C but this strategy needs to be further explored. In one of the sera an antibody response against the small ELDKWA epitope on the gp41 protein of HIV-1 subtype B was detected. The results of this study indicate that foreign peptides may be displayed by AHSV major capsid protein VP7 either on the surface of soluble VP7 trimers as a vaccine display system or by insoluble VP7 particles, in a denatured form on immunoblots, as an antiserum characterization strategy. As it is possible to produce more soluble AHSV VP7 trimers by modifying the VP7 protein, it may also be possible to increase the yield of core-like particles (CLPs) when this protein is co-expressed with AHSV VP3 in recombinant baculovirus-infected cells. Although the size of the presented peptide may be limited by a CLP strategy, these larger particles may induce cellular immunity. Although not yet investigated, it is likely that the soluble VP7 trimers may only elicit a humeral and not a cellular immune response.