Objective: Currently, dengue represents one of the most significant arboviral disease worldwide, for whicha vaccine is not yet available. Persistent challenges in live viral dengue vaccines have sparked a keeninterest in exploring non-replicating dengue vaccines. We have examined the feasibility of using themethylotrophic yeast Pichia pastoris to develop a chimeric vaccine candidate displaying the dengue virustype-2 (DENV-2) envelope domain III (EDIII), implicated in host receptor binding and in the inductionof virus-neutralizing antibodies, on the surface of non-infectious virus-like particles (VLP)-based on theHepatitis B virus core antigen (HBcAg).Methods: We designed a fusion antigen by inserting DENV-2 EDIII into c/e1 loop of HBcAg. A codon-optimized gene encoding this fusion antigen was integrated into the genome of P. pastoris, under thecontrol of the Alcohol Oxidase 1 promoter. The antigen was expressed by methanol induction and purifiedto near homogeneity by Ni2+ affinity chromatography. The purified antigen was characterized physicallyand functionally to evaluate its ability to assemble into VLPs, and elicit DENV-2-specific antibodies inmice.Results: This fusion antigen was expressed successfully to high yields and purified to near homogeneity.Electron microscopy and competitive ELISA analyses showed that it formed VLPs in which the EDIII moietywas accessible to different EDIII-specific antibodies. These VLPs were immunogenic in mice, stimulat-
ing the production of antibodies that could specifically recognize DENV-2 and neutralize its infectivity.However, virus-neutralizing antibody titers were modest.Conclusions: Our data show: (i) insertion of EDIII into the c/e1 loop of HBcAg does not compromise particleassembly; and (ii) the chimeric VLPs elicit a specific humoral response against DENV-2. The strategyof displaying dengue virus EDIII using a VLP platform will need further optimization before it may be
lterna
developed into a viable a
. Introduction
Dengue disease, which is prevalent in several countries of theorld, collectively representing roughly half the global population,
s caused by any of four closely related serotypes of dengue viruses,ENV-1, -2, -3 and -4. There is no drug to treat DENV infections and
vaccine is yet to be licensed. While the major thrust in the vaccineevelopment field is on live attenuated viral vaccines, efforts arenderway to explore non-replicating subunit vaccines as well, forheir obvious safety advantage [1]. In this context, the ∼100 aminocid (aa) residue DENV envelope domain III (EDIII), stabilized by a
ingle S-S bridge [2], has emerged as a potential subunit vaccineandidate [1,3]. Several key reasons underlie this emergence. Thisomain located at the C-terminus of the DENV envelope (E) protein
is exposed and accessible on the virion surface, implicated in virionbinding to the host cell surface receptor and contains multiple type-and sub-type specific neutralizing epitopes. Consequently, EDIII hasbeen the focus of many groups who have successfully expressed it,fused to a whole range of carriers, and demonstrated its ability toinduce virus-neutralizing antibodies [1,3].
In the context of recombinant subunit vaccines, initiatives seek-ing to utilize carriers with the capacity to self-assemble intovirus-like particles (VLPs) have come into focus with the recentsuccess of the VLP-based vaccine for human papilloma virus[4]. Several viral structural proteins, expressed in recombinantsystems, have an inherent ability to generate non-infectious, non-replicating VLPs and thus combine the advantages of whole virusvaccines and recombinant subunit vaccines [5]. In this regard, the183 aa residue hepatitis B virus core antigen (HBcAg) is well-
documented as a very promising VLP carrier [6]. RecombinantHBcAg, expressed in either prokaryotic or eukaryotic hosts, assem-bles into VLPs of two sizes, 30 nm and 34 nm, containing 180 and240 copies of the monomer, respectively. Importantly, it has been
ocumented that a variety of antigens, from bacterial, viral androtozoan pathogens, fused to HBcAg form chimeric VLPs. These
nsertions have been at three sites, the N and C termini and an inter-al region, known as the c/e1 loop, which projects from the VLPurface. While the N-terminal site can accommodate insertions ofp to 50 aa long, the C-terminal site can accept much larger inserts>100 aa long polypeptides). However, in contrast to the N-terminalnserts, the C-terminal inserts tend not to be well-displayed on theurface of the chimeric VLPs. Consequently, the C-terminal insertsre not as immunogenic as the N-terminal inserts. Interestingly, the/e1 loop, which represents the major immunodominant region,ot only accepts large antigen inserts, but can also display themn the surface of the chimeric VLPs [6]. A survey of the literatureocusing on >100 aa long antigen inserts in the c/e1 loop of HBcAgevealed several examples ranging in length from 108 to 256 aa7–15]. Most of these assemble into VLPs when expressed in E. coliSupplemental Table S1).
Using an E. coli expression system, we recently showed thatENV-2 EDIII (EDIII-2, 104 aa long), inserted into the c/e1 loopf HBcAg, could also be displayed on the surface of chimericLPs [15]. Unexpectedly, the chimeric VLPs generated from the. coli-expressed HBcAg-EDIII-2 fusion antigen elicited low titersf DENV-2 neutralizing antibodies. As the structural and antigenicntegrity of EDIII depends on a single S-S linkage [2], it is likely thathe reducing environment in the E. coli host cell which is not con-ucive to efficient formation of S-S bonds may have compromisedhe ability of EDIII to elicit high titer virus-neutralizing antibodies.his study was undertaken to examine if using a eukaryotic expres-ion host instead of E. coli would circumvent this issue. To this end,e expressed the chimeric HBcAg-EDIII-2 antigen in the methy-
otrophic yeast, Pichia pastoris, because of its well-documentedigh expression potential [16], its utility in expressing S-S linked
iral antigens [17,18] and the observations that HBcAg expressedsing this yeast assembles into VLPs [19]. This work showed that. pastoris offers an alternate host for the creation of HBcAg-basedhimeric VLPs capable of displaying DENV-2 EDIII on its surface.
ig. 1. Design and expression of chimeric Pp-HBcAg-EDIII-2 antigen. (a) Schematic represe chimeric antigen (bottom). The numbers in black indicate the aa residue numbers of thehe polyhistidine tag for affinity purification. (b) Map of the expression vector integrated iene insert encoding the chimeric antigen, shown in panel ‘a’; T: transcription terminanalysis. Total lysates of cultures of a P. pastoris clone harboring the plasmid shown in ‘becombinant chimeric antigen expression in Western blots using mAbs specific to the DE). (d). Induction optimization. Logarithmically growing cultures of the P. pastoris clone w.5% (lane 1), 1% (lane 2), 1.5% (lane 3) and 2% (lane 4), or for varying durations of 0 h (lane2%). Induced cultures were lysed and analyzed in Western blots using mAb 24A12. In botizes (in kDa) are shown to the left of the panels. The arrow to the right of these two pane
1 (2013) 873– 878
Further, the chimeric VLPs generated using P. pastoris appeared tobe relatively more efficient, than their E. coli-expressed counter-parts, in stimulating DENV-2-specific neutralizing antibodies.
2. Materials and methods
The ∼0.8 kilobase Pp-HBcAg-EDIII-2 gene (GenBank accessionnumber JQ723012), codon optimized for expression in P. pastoris,encoding ∼31 kilodalton (kDa) chimeric Pp-HBcAg-EDIII-2 protein(Fig. 1a) was obtained by chemical synthesis (Geneart AG, Regens-burg, Germany). This gene was cloned into P. pastoris plasmidpPICZa, integrated into the genome of P. pastoris host strain KM71Husing zeocin selection and purified by Ni-NTA affinity chromatog-raphy under denaturing conditions essentially as described before[17]. Samples containing Pp-HBcAg-EDIII-2 antigen were charac-terized by immunoblot analyses using in-house EDIII-specific mAb24A12 [17], anti-His mAb, or anti-HBcAg mAb ab8638, using a pro-tocol described earlier [15]. Protein estimation was done using BCAmethod with BSA as the standard [20]. The purified protein wasdialysed against 20 mM sodium bicarbonate buffer, pH 9.2 and itsability to assemble into VLPs was assessed by Electron Microscopy(EM) as described previously [15]. Groups (n = 6) of 4–6 weekold Balb/C mice were immunized intraperitoneally with differentantigens (Pp-HBcAg-EDIII-2 and its precursors HBcAg and EDIII-2)formulated in alum (20 �g antigen coated on 500 �g alum in 100 �l)on days 0, 30 and 90. Sera were obtained one week after the finalimmunization for analysis of antibody titers. Animal experimentswere performed in accordance with Government of India animalethics guidelines after approval by the Institutional Animal EthicsCommittee. Competitive ELISA to probe the surface-accessibilityof the EDIII moiety on the chimeric VLPs using mAb 24A12 [17],anti-EDIII-T antiserum [18] and anti-Pp-HBcAg-EDIII-2 antiserum,
indirect ELISA to assess EDIII-2- and DENV-2-specific antibodytiters in sera of immunized mice, indirect immunofluorescenceassay to evaluate the ability of anti-Pp-HBcAg-EDIII-2 polyclonalserum to recognize and bind to infectious DENV-2, and plaque
ntation of the C-terminally truncated HBcAg antigen (top) and the Pp-HBcAg-EDIII- two antigens. The empty box at the C-terminal end of the fusion antigen representsnto P. pastoris. Abbreviations are as follows. P: AOX1 promoter; FG: synthetic fusiontor; Z: zeocin selection marker; O: plasmid origin of replication. (c). Immunoblot’ were analyzed before (lane 1) and after methanol induction (lanes 2, 3 and 4) forNV EDIII moiety (lane 2), the HBcAg moiety (lane 3) and the polyhistidine tag (laneere either induced for a fixed duration (72 h) at varying methanol concentrations of
5), 24 h (lane 6), 48 h (lane 7), and 72 h (lane 8), at a fixed methanol concentrationh panels ‘c’ and ‘d’, pre-stained protein markers were run in lanes marked ‘M’. Theirls indicates the position of the Pp-HBcAg-EDIII-2 antigen.
cine 31 (2013) 873– 878 875
rtpm
3
3p
ltDtoFitaTscmHta
taeldahtaoio
3
eiumtpitof∼saf
3
tVpa
Fig. 2. Purification of Pp-HBcAg-EDIII-2 antigen. (a) Ni-NTA affinity chromato-graphic purification of Pp-HBcAg-EDIII-2 antigen. The solid curve depicts the elutionprofile monitored by absorbance at 280 nm. The dotted curve indicates the pH gra-
VLP as well, we carried out a competitive ELISA assay wherein theresidual antigen-binding activity of different anti-EDIII antibod-ies pre-incubated with HBcAg-EDIII-2 antigen was measured usingmonomeric EDIII-2 as the coating antigen [17]. The rationale was
Fig. 3. Characterization of Pp-HBcAg-EDIII-2 antigen. (a) Size exclusion chromato-graphic analysis of the purified fusion protein. The pooled peak material (shown
U. Arora et al. / Vac
eduction neutralization test (PRNT) to determine DENV-2 neu-ralizing antibody titers were carried out essentially as describedreviously [15]. Detailed methodology is presented in the Supple-ental file.
. Results
.1. Design of HBcAg-EDIII-2 antigen and its expression in P.astoris
A pentapeptide (LEDPI) spanning aa residues 77–81, in the c/e1oop of a deletion variant of the HBcAg antigen lacking the C-erminal 17 aa residues, was replaced with the 104 aa residueENV-2 EDIII. Further, this construct was provided with a C-
erminal poly-histidine tag to facilitate its purification. The designf the chimeric antigen, HBcAg-EDIII-2, is shown schematically inig. 1a. A P. pastoris expression vector, pPIC-HBcAg-EDIII-2, shownn Fig. 1b, capable of expressing this antigen under the control ofhe AOX1 promoter in response to methanol induction, was creatednd integrated into the genome of P. pastoris strain KM71H (MutS).his yeast clone was tested for Pp-HBcAg-EDIII-2 expression ashown in Fig. 1c. Aliquots of a methanol-induced culture of thislone were analyzed in immunoblots using an in-house anti-EDIIIonoclonal antibody (mAb 24A12), anti-HBcAg mAb and penta-is mAb (Fig. 1c, lanes 2–4). All three mAbs, identified a protein of
he predicted size, confirming successful expression of the fusionntigen.
Using this immunoblot assay, we next sought to optimize induc-ion conditions for obtaining maximal expression of the fusionntigen, as shown in Fig. 1d. Cultures were induced with differ-nt methanol concentrations, for a fixed time duration (Fig. 1danes 1–4), or at a single methanol concentration for different timeurations (Fig. 1d, lanes 5–8). Methanol concentrations beyond 2%nd induction times longer than 3 days were not tested as weave observed a reduction in recombinant product formation underhese conditions, which may be associated with methanol toxicitynd manifestation of proteolytic activity, respectively [17]. Basedn the data obtained in this experiment, we used 2% methanolnduction for 3 days to obtain maximal expression for the purposef purification (see below).
.2. Purification of Pp-HBcAg-EDIII-2 antigen
A preliminary localization experiment revealed that thexpressed fusion antigen is almost exclusively associated with thensoluble fraction (data not shown). As a result, we used the insol-ble fraction obtained from the induced biomass as the startingaterial and performed extraction and Ni-NTA affinity purifica-
ion under denaturing conditions. After washing away unboundrotein, the bound proteins were eluted by decreasing the pH. It
s evident from the chromatographic profile shown in Fig. 2a thathe bound material eluted as a single sharp peak at pH 4.3. Basedn the data in Fig. 2b, which represents an SDS-PAGE analysis ofractions across this peak, we estimated the purity attained to be95%. This is in agreement with the His-Sorb ELISA data which
howed that the pooled peak material was ∼93% pure. Starting from one-liter culture, we obtained 26 mg of purified Pp-HBcAg-EDIII-2usion antigen, representing a 37% yield (Supplemental Table S2).
.3. Pp-HBcAg-EDIII-2 assembles into VLPs
As recombinant HBcAg and several of its in-frame fusion deriva-
ives with several antigenic partners are known to assemble intoLPs [6], we analyzed if the purified Pp-HBcAg-EDIII-2 antigenossesses this ability. An initial size exclusion chromatographicnalysis of the purified protein showed that it emerged in the void
dient employed during chromatography. (b) SDS-PAGE analysis of the wash (lane1) and fractions across the major peak (lanes 2–7). Low molecular weight proteinmarkers were run in lane M; their sizes (in kDa) are indicated to the left.
volume of a Sephacryl S-300 column suggesting that it is present asa multi-molecular entity (Fig. 3a). It has been shown that orderedmulti-subunit particles appear as distinct slow-migrating bands ascompared to non-associated monomeric molecules which appearas fast-migrating and more diffuse bands upon native agarose gelelectrophoresis. Both HBV capsid and its GFP-derivative whichform VLPs manifest this electrophoretic behavior [7]. Interestingly,the Pp-HBcAg-EDIII-2 protein that emerged in the void volume ofSephacryl S-300 appeared as a distinct slow-migrating band uponnative agarose gel electrophoresis (data not shown), suggestingthat it is indeed capable of forming higher order structures. Thiswas confirmed by electron microscopic (EM) analysis after nega-tive staining with uranyl acetate as shown in Fig. 3b. It is evidentthat purified EDIII-2 containing fusion antigen exists as distinctVLPs predominantly in the size range of ∼40–60 nm. Further, wealso observed that these VLPs retained their structural integrityfollowing storage at 37 ◦C for 2 weeks (data not shown).
3.4. EDIII-2 component of the VLPs is accessible to anti-EDIII-2antibodies
It is well documented that antigens inserted into the c/e1 loopof HBcAg tend to be displayed on the surface of the chimeric VLPs[6]. To assess if this may be true for the Pp-HBcAg-EDIII-2 chimeric
in Fig. 2a) was dialyzed and analyzed on a Sephacryl S-300 column. The columneffluent was monitored for absorbance at 280 nm. The downward arrow (above thex-axis) indicates the void volume of the column. (b) The protein in the major peak(shown in panel ‘a’) was negatively stained with uranyl acetate and visualized byelectron microscopy.
876 U. Arora et al. / Vaccine 31 (2013) 873– 878
Fig. 4. Competitive ELISA. Varying concentrations of purified HBcAg (dashed curves) and Pp-HBcAg-EDIII-2 (solid curves) VLPs were pre-incubated with anti-EDIII mAb2 serumc -HRPp ations
tttaiEbmsaEtrstcltba
3D
c
Fpv(D
4A12 (a), murine anti-EDIII-T polyclonal serum (b) or anti-Pp-HBcAg-EDIII-2 antioated with purified P. pastoris-expressed EDIII-2 and revealed with anti-mouse IgGrotein in the pre-incubation step was taken to represent 100%. The regression equ
hat if the EDIII-2 moiety of the chimeric VLP was indeed accessibleo anti-EDIII antibody, it would render antigen-combining sites onhe latter unavailable for binding to the monomeric EDIII-2 coatingntigen in the microtiter wells. Thus, a reduction in ELISA reactivityn this assay would be a reflection of the relative efficiency of theDIII-2 moiety (of the fusion antigen), to bind to anti-EDIII anti-ody. This notion is borne out by data presented in Fig. 4. WhenAb 24A12 (EDIII-specific in-house mAb) was tested in this assay,
hown in Fig. 4a, its reactivity to the monomeric EDIII-2 coatingntigen was inversely proportional to concentration of Pp-HBcAg-DIII-2 VLP it was pre-treated with. However, pre-treatment ofhis mAb with HBcAg VLPs (lacking EDIII-2) did not affect its ELISAeactivity toward the EDIII-2 coating antigen (Fig. 4a). Essentiallyimilar results were obtained, shown in Fig. 4b, when we performedhis experiment with a murine polyclonal serum raised against ahimeric antigen containing the EDIIIs of the four DENV serotypesinked in tandem [18]. This suggests that the EDIII-2 moiety onhe chimeric VLP is accessible on the surface and therefore capa-le of recognizing and binding to the monoclonal and polyclonalnti-EDIII antibodies.
.5. Anti-Pp-HBcAg-EDIII-2 antibodies bind to and neutralize
ENV-2
If the EDIII-2 moiety is indeed displayed on the surface of thehimeric VLPs as the above data suggest, then it must be possible to
ig. 5. Evaluation of murine anti-Pp-HBcAg-EDIII-2 antisera. (a) Serial dilutions of murinrecursors (HBcAg: circles; EDIII-2: squares) were analyzed using recombinant EDIII-2; (birus; in panels ‘a’ and ‘b’, the dashed curves represent ELISA titration profiles of sera fromupper left), or immunized with HBcAg (upper right), EDIII-2 (lower left) or Pp-HBcAg-EDENV-2 in infected BHK-21 cells in an indirect immunofluorescence assay in conjunction
(c). Residual antibodies in the pre-incubation mix were captured on ELISA platesO in conjunction with TMB substrate. ELISA reactivity in the absence of any added
for the solid curves are shown just above the horizontal axes.
elicit antibodies specific to this moiety using these chimeric VLPs asimmunogens. To this end, mice were immunized with the chimericVLPs and the resultant antisera used in the competitive ELISAdescribed above. Consistent with expectation, we found that theELISA reactivity of this polyclonal serum once again manifested thesame behavior as in the previous experiments, as evident from thedata presented in Fig. 4c. These data strengthen the conclusion thatanti-Pp-HBcAg-EDIII-2 antibodies do recognize monomeric EDIII-2antigen. This was corroborated by directly assaying the reactivityof the Pp-HBcAg-EDIII-2-induced antibodies on monomeric EDIII-2in a standard indirect ELISA format, shown in Fig. 5a. For com-parison, we also analyzed antibodies elicited by the monovalentprecursors, EDIII-2 and HBcAg. This experiment revealed that theEDIII-2-specific titer of the anti-Pp-HBcAg-EDIII-2 antiserum wasquite high (ELISA titer ∼18,500). However, this was markedly lesscompared to the EDIII-2-specific antibody titer of anti-EDIII-2 anti-serum (ELISA titer »64,000). As expected, anti-HBcAg antiserumdid not manifest any significant EDIII-2-specific titers. When thisexperiment was repeated using DENV-2 as the coating antigeninstead of recombinant EDIII-2 protein, we observed a similar pat-tern with anti-EDIII-2 serum manifesting higher DENV-2-specifictiters than anti-Pp-HBcAg-EDIII-2 serum, as shown in Fig. 5b. How-
ever, the notable difference was that titers were more than twoorders of magnitude lower compared to those observed againstEDIII-2 protein (Fig. 5b). Anti-DENV-2 ELISA titers elicited byEDIII-2 and Pp-HBcAg-EDIII-2 were ∼200 and <100, respectively.
e antisera raised against the fusion antigen Pp-HBcAg-EDIII-2 (diamonds) and its) similar indirect ELISA as in panel ‘a’, except that the coating antigen was DENV-2
mock-immunized mice; (c) antisera from mice that were either mock-immunizedIII-2 (lower right) antigens were used as the source of primary antibodies to detect
with secondary anti-murine IgG-FITC conjugate.
cine 3
NbaEpeoitV(
4
rvadpeeVaot
otscabhnawha
eaaeoetirscgdbppTccapvtDt
[
U. Arora et al. / Vac
ext, we analyzed the ability of these antisera to recognize andind to infectious DENV-2 using an indirect immunofluorescencessay (Fig. 5c). Interestingly, we observed that the anti-Pp-HBcAg-DIII-2 antiserum worked better than anti-EDIII-2 antiserum inicking up DENV-2 in this assay, suggesting a qualitative differ-nce in the interaction of these antibodies with DENV-2 dependingn whether it is coated on a microtiter plate or is present innfected cells. To see if the anti-Pp-HBcAg-EDIII-2 antiserum con-ained any virus-neutralizing antibodies, we performed PRNT usingero cells. This showed modest levels of DENV-2-neutralizing titers
PRNT50 ∼40).
. Discussion
In the field of dengue vaccine development research, there hasecently been an increasing awareness to explore non-replicatingaccine candidates, in addition to the live virus vaccines [1]. Whilen insect cell-expressed recombinant DENV E based vaccine can-idate has reached phase I trials, EDIII has emerged recently as aromising vaccine immunogen for a variety of reasons enumeratedarlier [1,3]. In this regard, the unique ability of HBcAg to accept for-ign epitopes into its structure and yet retain its potential to formLPs has resulted in its being widely explored as carrier for vaccinentigens [6]. Two vaccine candidates, one for malaria [21], and thether, influenza [22], have undergone safety and immunogenicityrials with encouraging results.
The size of DENV-2 EDIII and our intention to try and display itn the surface of the chimeric VLPs dictated that we insert it intohe c/e1 loop of HBcAg. Recently, using an E. coli expression host, wehowed that an HBcAg chimera with DENV-2 EDIII inserted into its/e1 loop assembles into VLPs in which the EDIII-2 moiety, whichppeared to be displayed on the surface elicited virus-specific anti-odies [15]. However, the neutralizing antibody titers were notigh, leading us to suspect that the E. coli system presumably didot permit efficient S-S bond formation within the EDIII moiety. Toddress this possibility, we undertook the present study in whiche replaced the E. coli with methylotrophic yeast P. pastoris as theost. To ensure efficient expression we codon-optimized the fusionntigen for the yeast host.
The new version of the fusion antigen, Pp-HBcAg-EDIII-2, wasxpressed by methanol-induction, and purified using Ni2+-NTAffinity chromatography. In induced yeast cell lysates, it was mostlyssociated with the insoluble fraction. This is consistent with ourarlier observation regarding the expression of DENV-2 EDIII with-ut a fusion partner [17], and is presumably a reflection of efficientxpression arising from the use of a codon-optimized gene whichends to favor the association of the recombinant protein with thensoluble phase. Interestingly, and consistent with this notion, aecent report describes the purification of HBcAg VLPs from theoluble fraction of P. pastoris lysate, using a wild type gene withoutodon optimization [19]. As the chimeric Pp-HBcAg-EDIII-2 anti-en was insoluble, we purified it from the induced biomass underenaturing conditions followed by removal of the denaturing agenty dialysis. Preliminary analysis by Sephacryl S-300 chromatogra-hy indicated that following removal of the denaturing agent, theurified chimeric antigen assembled into higher order structures.his was confirmed by EM analysis which revealed the presence ofharacteristic VLPs in the purified preparation. Importantly, usingompetitive ELISA we found that the EDIII moiety of the chimericntigen was accessible to anti-EDIII antibodies indicating that it wasresumably surface-exposed akin to that in the wild type DENV-2
irion. This was corroborated by the ability of the chimeric VLPso elicit EDIII-specific antibodies which could recognize and bindENV-2 in an immunofluorescence assay and neutralize its infec-
ivity in PRNT.
[
1 (2013) 873– 878 877
In all the attributes described above, the Pp-HBcAg-EDIII-2chimeric VLPs were essentially similar to their counterparts pro-duced using E. coli host [15]. Notably, both were capable ofstimulating the production of a specific humoral response toDENV-2. Surprisingly, virus-neutralizing antibody titers elicited byPp-HBcAg-EDIII-2 VLPs were also modest, being comparable tothose elicited by the corresponding E. coli-based VLPs [15]. Thissuggests that though the EDIII-2 moiety may be accessible onthe chimeric VLP surface, when expressed using either host, itsmolecular microenvironment may not be identical to that on thevirion surface. In conclusion, retention of VLP-forming ability andsurface-display alone may not necessarily be adequate to ensureimmunogenicity of EDIII inserted into the c/e1 loop of HBcAg. Asthe efficient induction of neutralizing antibodies is considered to beimportant for protection against dengue, the VLP display strategyneeds further optimization/modification to evaluate its potentialas a viable alternate option for the development of non-replicatingdengue vaccines.
Acknowledgments
The authors are grateful to the Malaria group at ICGEB for theuse of the electron microscope. They acknowledge Nisha Dhar forhelp with cloning and Rajendra Raut for doing the virus neutral-ization assay. UA is the recipient of a senior research fellowshipfrom the Department of Biotechnology, Government of India. SSand NK gratefully acknowledge the funding support received fromthe Department of Biotechnology, Government of India.
Conflict of interest statement: The authors declare that they haveno conflict of interest.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.vaccine.2012.12.016.
References
[1] Swaminathan S, Batra G, Khanna N. Dengue vaccines: state of the art. ExpertOpin Ther Patents 2010;20:819–35.
[2] Roehrig JT, Volpe KE, Squires J, Hunt AR, Davis BS, Chang GJJ. Contribution ofdisulfide bridging to epitope expression of the dengue type 2 virus envelopeglycoprotein. J Virol 2004;78:2648–52.
[3] Guzman MG, Hermida L, Bernardo L, Ramirez R, Guillén G. Domain IIIof the envelope protein as a dengue vaccine target. Expert Rev Vaccines2010;9:137–47.
[4] Mao C, Koutsky LA, Ault KKA, Wheeler CM, Brown DR, Wiley DJ, Alvarez FB,Bautista OM, Jansen KU, Barr E. Efficacy of human papillomavirus-16 vaccine toprevent cervical intraepithelial neoplasia-a randomized controlled trial. ObstetGynecol 2006;107:18–27.
[5] Chackerian B. Virus-like particles: flexible platforms for vaccine development.Expert Rev Vaccines 2007;6:381–90.
[6] Pumpens P, Grens E. HBV core particles as a carrier for B cell/T cell epitopes.Intervirology 2001;44:98–114.
[7] Kratz PA, Böttcher B, Nassal M. Native display of complete foreign proteindomains on the surface of hepatitis B virus capsids. Proc Natl Acad Sci USA1999;96:1915–20.
[8] Wang XJ, Gu K, Xiong QY, Shen L, Cao RY, Li MH, Li TM, Wu J, Liu JJ. A novel virus-like particle-based on hepatitis B core antigen and substrate-binding domainof bacterial molecular chaperone DnaK. Vaccine 2009;27:7377–84.
[9] Skamel C, Ploss M, Böttcher B, Stehle T, Wallich R, Simon MM, Nassal M. Hepati-tis B virus capsid-like particles can display the complete, dimeric outer surfaceprotein C and stimulate production of protective antibody response againstBorrelia burgdorferi infection. J Biol Chem 2006;281:17474–81.
10] Nassal M, Skamel C, Kratz PA, Wallich R, Stehle T, Simon MM. A fusion productof the complete Borrelia burgdorferi outer surface protein A (OspA) and thehepatitis B virus capsid protein is highly immunogenic and induces protectiveimmunity similar to that seen with an effective lipidated OspA vaccine formula.
Eur J Immunol 2005;35:655–65.
11] Bandurska K, Brodzik R, Spitsin S, Kohl T, Portocarrero C, Smirnov Y, Pogreb-nyak N, Sirko A, Koprowski H, Golovkin M. Plant-produced hepatitis B coreprotein chimera carrying anthrax protective antigen domain-4. Hybridoma2008;27:241–7.
8 cine 3
[
[
[
[
[
[
[
[
[
[
78 U. Arora et al. / Vac
12] Boulter NR, Glass EJ, Knight PA, Bell-Sakyi L, Brown CGD, Hall R. Theileria annu-lata sporozoite antigen fused to hepatitis B core antigen used in a vaccinationtrial. Vaccine 1995;13:1152–60.
13] Koletzki D, Biel SS, Meisel H, Nugel E, Gelderblom HR, Krüger DH, Ulrich R. HBVcore particles allow the insertion and surface exposure of the entire poten-tially protective region of Puumala hantavirus nucleocapsid protein. Biol Chem1999;380:325–33.
14] Geldmacher A, Skrastina D, Petrovskis I, Borisova G, Berriman JA, RosemanAM, Crowther RA, Fischer J, Musema S, Gelderblom HR, Lundkvist A, Ren-hofa R, Ose V, Krüger DH, Pumpens P, Ulrich R. An amino-terminal segment ofhantavirus nucleocapsid protein presented on hepatitis B virus core particlesinduces a strong and highly cross-reactive antibody response in mice. Virology2004;323:108–19.
15] Arora U, Tyagi P, Swaminathan S, Khanna N. Chimeric hepatitis B core antigen
virus-like particles displaying the envelope domain III of dengue virus type 2.J Nanobiotechnol 2012;10:30.
16] Macauley-Patrick S, Fazenda ML, McNeil B, Harvey LM. Heterologous pro-tein production using Pichia pastoris expression system. Yeast 2005;22:249–70.
[
1 (2013) 873– 878
17] Batra G, Raut R, Dahiya S, Kamran N, Swaminathan S, Khanna N. Pichia pastoris-expressed dengue virus type 2 envelope domain III elicits virus-neutralizingantibodies. J Virol Methods 2010;167:10–6.
18] Etemad B, Batra G, Raut R, Dahiya S, Khanam S, Swaminathan S, Khanna N. Anenvelope domain III-based chimeric antigen produced in Pichia pastoris elicitsneutralizing antibodies against all four dengue virus serotypes. Am J Trop MedHyg 2008;79:353–63.
19] Freivalds J, Dislers A, Ose V, Pumpens P, Tars K, Kazaks A. Highly efficient pro-duction of phosphorylated hepatitis B core particles in yeast Pichia pastoris.Protein Expr Purif 2011;75:218–24.
20] Smith PK, Krohn RI, Hermanson G, Mallia AK, Gartner FH, Provenzano MD,Fujimoto EK, Goeke NM, Olson BJ, Klenk DC. Measurement of protein usingbicinchoninic acid. Anal Biochem 1985;150:76–85.
R. Phase I trial of an alhydrogel adjuvanted hepatitis B core virus-like particlecontaining epitopes of Plasmodium falciparum circumsporozoite protein. PLoSOne 2008;3:e1556.
22] Fiers W, de Filette M, Bakkouri KE, Schepens B, Roose K, Schotsaert M, Birkett A,Saelens X. M2e-based universal influenza A vaccine. Vaccine 2009;27:6280–3.
