The self-assembling protein nanoparticle (SAPN) is an antigen-presenting system that has been shown to be suitable for use as a vaccine platform. The SAPN scaffold is based on the principles of icosahedral symmetry, beginning from a monomeric chain that self-assembles into an ordered oligomeric state. The monomeric chain contains two covalently linked α-helical coiled-coil domains, an N-terminal de novo-designed pentameric tryptophan zipper and a C-terminal de novo-designed trimeric leucine zipper, which assemble along the internal symmetry axes of an icosahedron. In this study, we incorporated the membrane proximal external region (MPER) of HIV-1 gp41 from HXB2 into the N-terminal pentamer, referred to as MPER-SAPN, attempting to reproduce the α-helical state of the 4E10 epitope while maintaining a structurally less-constrained 2F5 epitope. Sprague–Dawley rats were immunized with MPER-SAPNs, and their sera were analyzed for induced humoral anti-HIV-1 responses. We show that high membrane proximal external region-specific titers can be raised via the repetitive antigen display of MPER on the SAPN without the need for adjuvant. However, none of the sera displayed a detectable neutralizing activity against HIV-1. Thus, 4E10- and 2F5-like neutralizing antibodies could not be elicited by MPER conformationally restrained in the SAPN context.
HIV vaccine design has varied in strategy and included viral vectors, DNA vaccines, and soluble protein vaccine formulations. None of these approaches has thus far resulted in protection in the human host (1–3).
Toward the development of a protective HIV-1 vaccine (3,4), a small number of human monoclonal antibodies (mAbs) with cross-neutralizing activity have been identified and studied in detail. These include the 2G12 neutralizing antibody (nAb), which recognizes a conformation of high mannose sugars on the glycosylated face of the virus and binds them to inhibit the attachment of the gp120 protein to the CD4 receptor molecule (4). Also, the b12 nAb recognizes the exposed CD4-binding region on gp120 and attaches to it irreversibly (3,4). Two other very important nAbs, 2F5 and 4E10, have been shown to bind and maintain the prefusion or fusion intermediate states of the virus when gp120 exposes gp41 for fusion, thus preventing gp41 from folding into the hexameric precursor bundle that is necessary for cell entry (5).
Membrane proximal external region
The membrane proximal external region (MPER) of HIV-1 gp41 has been recognized as a target for broadly neutralizing antibodies (2,6). The MPER extends from amino acid residues 660–683 of gp160 (HXB2 amino acid numbering) and is important for lipid membrane perturbation and viral entry into target cells (7,8). The conformation of MPER in native, prefusion gp41 is possibly an α-helical trimer and, additionally, at least two further conformational states are induced by the consecutive interactions between gp120 and cell surface receptors (CD4, CCR5, or CXCR4). These are first the fusion intermediate ‘prehairpin’ conformation and then the postfusion state in which the ectodomain of gp41 has formed a six helix bundle consisting of antiparallel coiled coils of the N-terminal and C-terminal heptad repeats (7). In these two latter states, the MPER itself is present in a currently unknown secondary structure (9–11).
For neutralization of virus infectivity to occur, it is thought that MPER antibodies target either the prefusion state or the fusion intermediate state (with potentially higher epitope exposure) of gp41 (10–12). 2F5 and 4E10 target adjacent epitopes in this site and broadly neutralize different HIV-1 isolates (6). Structural data of peptides bound to the respective mAbs indicate that the 4E10 epitope has an α-helical conformation while the 2F5 epitope exhibits a more extended, unrestrained structure with a small beta turn (2,4,13). However, MPER-specific antibodies elicited by several previous vaccines strategies have failed to be broadly neutralizing (2–4,14,15).
In this study, we present an adjuvant-free immunogen, MPER-self-assembling protein nanoparticle (SAPN), as a subunit vaccine candidate for the presentation of conformationally appropriate 4E10 and 2F5 epitopes. Outbred Sprague–Dawley rats were immunized with MPER-SAPN, and the elicited humoral immune responses were characterized in vitro. In case of promising results, HIV receptor-complex-transgenic rats derived from this strain would in principle allow us to expand these analyses into the context of an in vivo challenge (16,17).
Self-assembling protein nanoparticle
The geometric basis of the SAPN design is derived from the concept of monomeric protein chains capable of self-assembling into a larger nanoparticle architecture. The monomeric chains transit toward an oligomeric scaffold (Figure 1A,B) by self-assembling along the fivefold and threefold symmetry axes of higher-order polyhedra, such as dodecahedrons, icosahedrons, or quasi-equivalent species, as has been previously described by Raman et al. (18). This self-assembling scaffold provides an epitope presentation platform that generates highly immunogenic responses for a diverse group of pathogens, without the need for adjuvants (18–20). Conveniently, antigens can be fused either to the trimer and/or the pentamer leading to a repetitive display of B-cell epitopes on or near the surface of the SAPN (Figure 1). Therefore, SAPN provides a new type of immunogenic platform that we have previously shown to be a successful subunit vaccine, without the necessity for adjuvants, and containing the pliability to present multiple types of epitopes (18,19,21–23).
Methods and Materials
Synthesis of MPER-SAPN and CONT-SAPN
Oligonucleotides coding for the MPER sequence (HIV-1 gp41 HXB2 strain residues 662–683): ELDKWASLWNWFNITNWLWYIR were ordered from IdtDNA (Integrated DNA Technologies Inc., Coralville, IA, USA), annealed, and ligated into a pPEP-T vector coding for the de novo monomeric protein chain. The final plasmid vector encoding the protein sequence, MGHHHHHHASELDKWASLWNWFNITNWLWYIRSWQTWNAKWDQWSNDWNAWRSDWQAWKDDWARWRALWMGGRLLLRLEELERRLEELERRLEELERAINTVDLELAALRRRLEELARGGSGDPPPPNPNDPPPPNPND, was then transformed into an Eschericia Coli KRX strain (Promega, Madison, WI, USA) and incubated per company instructions. Briefly, the transformed E. Coli culture was grown in Terrific Broth (Luria broth with 0.89 m Potassium Phosphate) containing 100 μg/mL ampicillin and 34 μg/mL chloramphenicol at 37 °C with shaking at 190 rpm to an optical density (O.D.) of approximately 1.0. The temperature was then reduced to 25 °C and the culture allowed to grow to an O.D. of approximately 1.5, when protein expression was induced with 1 mm isopropyl β-d-thiogalactopyranoside and 0.1%l-rhamnose monohydrate.
The induced cell culture was grown overnight at 25 °C and then harvested via centrifugation at 4000 × g for 10 min. The pellet was frozen at −80 °C, thawed at 4 °C, and resuspended in 9 m urea, 100 mm NaH2PO4, 10 mm Tris at pH 8.0. The resuspended pellet was lysed via sonication, and the cell debris centrifuged at 30 600 × g for 45 min. The supernatant was incubated for 1 h with nickel beads (Qiagen, Madison, WI, USA) and applied to a Nickel-NTA column. This was washed with a pH gradient to remove any contaminants. The pH gradient consisted of the following buffers (all at 9 m urea): pH 6.3, pH 5.9, pH 5.1, and pH 4.5, in the presence of 100 mm NaH2PO4, 20 mm sodium citrate, and 10 mm imidazole. An imidazole gradient was then applied to elute bound protein. This consisted of 50, 150, 300, 500 mm, and 1 m imidazole, pH 8.0. The purities and molecular weights of the proteins in purified fractions were assessed via sodium dodecyl sulfate polyacrylamide gel electrophoresis and correlated with the expected 16985.8 Da size. Most of the purified protein eluted in the protein fractions from pH 5.1, pH 4.5, and 50 mm imidazole (pH 8.0).
The purified protein was then filtered through a 0.1-μm polyvinylidene fluoride membrane (Millipore #SLVV033 RS; Millipore, Billerica, MA, USA) and dialyzed into a buffer containing 8 m urea, 150 mm NaCl, 20 mm Hepes, 5% glycerol, set at pH 7.5, to begin a stepwise refolding in which the concentration of urea was lowered sequentially. Over a period of 4 h per dialysis step, and while maintaining the underlying buffer conditions, the concentration of urea was reduced stepwise from 8, 6, 4, 2, 1, to 0 m urea (twice dialyzed in 0 m urea, to remove underlying traces of denaturant). Room temperature was maintained up to 4 m urea, and then, the dialysis was transferred to 4 °C, until completion. Filtered protein (with a 0.1-μm Millipore filter) was frozen at −80 °C.
CONT-SAPN with the sequence MGHHHHHHASWKWDGGLVPRGSWQTWNAKWDQWSNDWNAWRSDWQAWKDDWARWRALWMGGRLLLRLEELERRLEELAKFVAAWTLKAAAVDLELAALRRRLEELARGGSGDRAAGQPA GDRADGQPA was synthesized in a fashion similar to MPER-SAPN, with dialysis from 8 m urea to 0 m urea, into 20 mm TRIS, 50 mm NaCl, 5% glycerol, at a pH of 7.5.
Recombinant proteins, defined antibodies
Recombinant HXB2 gp120, generated in insect cells, was purchased from ImmunoDiagnostic, Woburn, MA, USA and HXB2-gp41-ectodomain (amino acids 541–682), generated in yeast cells, from Viral therapeutics, Ithaca, NY, USA. Membrane proximal external region peptide (ELLELDKWASLWNWFNITNWLWY) coupled to ovalbumin (MPER-OVA) was generated by Peptide Speciality Laboratories GmbH, Heidelberg, Germany. Cross-neutralizing human gp41 mAbs 2F5 (24) and 4E10 (25), the cross-neutralizing human mAb b12 (26), and mouse mAb H902 specific for the Env-V3 loop of HIV strain BH10 (27) were obtained from the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH, Bethesda, MD, USA.
Dynamic light scattering
The refolded protein (MPER-SAPN) was characterized via dynamic light scattering (DLS) obtained from a Malvern Zetasizer Nano S (Malvern, Worcestershire, UK), at 25 °C with a buffer containing 150 mm NaCl, 20 mm Hepes, 5% glycerol, set at pH 7.5. An average of five DLS readings was obtained over a period of 10 min per run.
Transmission electron microscopy
An FEI Tecnai T12 S/Transmission electron microscopy (TEM) microscope (FEI Worldwide, Hillsboro, OR, USA) was utilized at an 80 kV accelerating voltage to visually characterize the diameter and morphology of 0.03 mg/mL MPER-SAPN stained on a copper grid (Electron Microscopy Sciences, Hatfield, PA, USA) with 1% uranyl acetate.
Immunization of rats
For this immunization study, outbred Sprague–Dawley rats were obtained from Charles River (Sulzfeld, Germany) (all female, age: 8–13 weeks). Animal experiments were conducted according to the German animal welfare act and with authorization of the Regierungspräsidium Karlsruhe (35-9185.81/G-152/07). The experiments were supervised by animal welfare officers of Heidelberg University. A total of six rats were immunized with preparations of MPER-SAPN (rats 1–6, Table 1). Additionally, two animals were immunized with a mixture of recombinant HXB2-gp120 and HXB2-gp41-ectodomain (amino acids 541–682) (rats 7 and 8, Table 1). The amounts of immunogen injected, the application routes, and adjuvant employed were as described in Table 1. These rats and, additionally, two unimmunized animals (rats 9 and 10, Table 1) were killed at day 86 and sera prepared using standard procedures.
Table 1. CD4/CXCR4 transgenic rat immunization schedule
Animal number, route of administration, and adjuvant
Prime (day 0) Boost (days 30,51,72) Bleed (day 86)
MPER, membrane proximal external region; SAPN, self-assembling protein nanoparticle.
Analyses of antibody-binding titers of rat sera or control human mAbs, 2F5, or 4E10, to either MPER-SAPN, CONT-SAPN, recombinant gp41 proteins, or to MPER peptide (ELLELDKWASLWNWFNITNWLWY) coupled to ovalbumin (MPER-OVA) (all coated at 50 ng/well) were carried out as described previously (28).
To remove antibodies directed against the SAPN scaffold and thus common to both MPER-SAPN and CONT-SAPN, adsorptions of the sera were carried out where indicated. Aliquots of the sera (diluted 1:500) were incubated overnight at 4 °C either alone or with 20 μg/mL of MPER-SAPN or CONT-SAPN. Centrifugation at 4 °C at 21 000 × g for 1 h was then carried out to remove immune complexes and the resulting serum supernatants subjected to ELISA as above.
Neutralization assays with 1:10 dilutions of individual immunized rat sera were carried out as described previously (28). Cross-neutralizing human mAb b12, mouse mAb H902, specific for the Env-V3 loop of HIV strain BH10, and rabbit anti-gp120 (HIV strain BH10) were employed as positive controls.
In this study, we have self-assembled MPER-SAPN, a protein nanoparticle incorporating the MPER amino acid sequence at the N-terminal end of the pentameric unit of the SAPN (Figure 1B). The MPER epitope was added in a manner that maintained the native α-helical presentation of the 4E10 epitope but left the conformationally more flexible 2F5 epitope in a less-constrained conformation. This was achieved by aligning the 4E10 epitope into the heptad repeat of the de novo tryptophan zipper oligomerization domain (Figure 1A), creating a coiled-coil 4E10 epitope pentameric bundle. The 2F5 epitope sequence was not incorporated into the coiled-coil heptad repeat of MPER-SAPN’s pentameric domain, and therefore, it was left unrestrained, and conformationally flexible. Upon assembly of the nanoparticle, the MPER epitope will be displayed on its surface (Figure 1C). It is expected that immunization with MPER-SAPN should lead to the formation of conformation-specific, and thus potentially cross-neutralizing, MPER antibodies. In this study, we immunized outbred Sprague–Dawley rats by different routes in the presence and absence of adjuvants and analyzed the humoral responses elicited.
The amino acid sequence of the MPER-SAPN protein (Figure 1C) and of a control SAPN (CONT-SAPN, Figure 1D) is shown. An identical core sequence without MPER yielded aggregating protein samples so that the sequence of the CONT-SAPN protein differed at two additional sites (total 29 amino acids) within the C-terminal part of the protein sequence (Figure 1D).
SAPN refolding and analysis
Removal of urea and refolding of MPER-SAPN and CONT-SAPN was achieved by sequential dialysis steps. The results of dialysis at two different pHs (7.5 and 8.5) were compared. As shown by DLS (Figure 2A), MPER-SAPN had a more uniform distribution at pH 7.5. The peak position points to MPER-SAPN having a hydrodynamic diameter of approximately 31 nm, a width of approximately 14 nm, contained within an overall percent volume of 96%. The pH 8.5 species of MPER-SAPN, while appearing soluble during dialysis, and being filterable through a 0.1-μm filter, showed an aggregation propensity. Two DLS peaks were detected, representing two species of nanoparticle, one with a hydrodynamic diameter approximately 7 nm, and the other at approximately 21 nm. CONT-SAPN also refolded stably at pH 7.5 with the DLS showing the main species at approximately 35 nm diameter (Figure 2A). Transmission electron microscopy confirmed these distributions, showing a discrete set of MPER-SAPN with a diameter ranging from about 25 to 40 nm (Figure 2B). Thus, the MPER-SAPN species has aligned according to the principles of icosahedral self-assembly, with monomeric chains (MW: 16985.8 Da) refolding into a nanoparticle that displayed helicity at 222 nm on CD spectra.
Reactivity of MPER-SAPN with human MPER mAbs 2F5 and 4E10
To confirm that the gp41 MPER epitopes on the MPER-SAPNs were displayed in a relevant conformation, binding of the cross-neutralizing human mAbs 2F5 and 4E10 to MPER-SAPN or CONT-SAPN as well as to ectodomain HXB2-gp41 and MPER-OVA was assessed by ELISA. As shown in Figure 3, both mAbs bound efficiently to MPER-SAPN (Figure 3A) but, as expected, failed to bind to CONT-SAPN (Figure 3A). This confirms the presence and accessibility to mAb binding of the MPER epitopes on the MPER-SAPNs. Both 2F5 and 4E10 mAbs bound to recombinant gp41 (Figure 3B) to similar levels as to MPER-SAPN. In the case of MPER-OVA, binding of 4E10 was weaker indicating that in the peptide context, its conformation was not compatible with strong binding.
Humoral responses induced in immunized rats
Two groups of rats were immunized with MPER-SAPNs either intraperitoneal (i.p.) with incomplete Freund′s adjuvant (IFA) or intradermally (i.d.) without adjuvant. Additionally, as positive controls for the elicitation of type-specific nAbs, two animals were immunized with a mixture of recombinant HXB2-gp120 and HXB2-gp41-ectodomain (amino acids 541–682) (Table 1). Two animals remained unimmunized. Sera from primed/boosted animals were first tested for binding to MPER-SAPN and CONT-SAPN. As shown in Figure 4A, all of the sera from animals immunized with MPER-SAPN contained antibodies that bound strongly to MPER-SAPN. The binding profiles were not affected by the presence or lack of adjuvant or by the route of administration of vaccine, be it i.p. or i.d. (Figure 4). The sera from the rats immunized with recombinant gp41 (and gp120) did not bind to MPER-SAPN despite the MPER epitope being present in the gp41 protein employed. All of the sera from the MPER-SAPN immunized rats also exhibited binding to CONT-SAPN (Figure 4B). This reflects the presence of antibodies induced against SAPN scaffold domains, common to both MPER-SAPN and CONT-SAPN. However in these cases, the binding titers were markedly lower (50% binding titers at least 10-fold lower than against MPER-SAPN).
To more directly demonstrate responses against the unique epitopes in MPER-SAPN, that is the MPER epitope itself as well as the further unique sequences in the C-terminus of the MPER protein, the sera were subjected to adsorption. This was performed either with MPER-SAPN or CONT-SAPN and subsequently specific antibody titers sera were (re-)evaluated by ELISA. For serum from rat 4 (immunized i.d. without adjuvant), for example, the binding activity to MPER-SAPN was very significantly reduced (by over 90%) by adsorption with homologous MPER-SAPN, but was not affected by adsorption to CONT-SAPN (Figure 5). This demonstrates that virtually all antibodies induced by MPER-SAPN are directed against the epitopes unique to MPER-SAPN. On the other hand, testing of these adsorbed sera on CONT-SAPN demonstrated that the adsorption procedure had successfully removed the antibodies against the common scaffold sequence. This result was replicated in all of the sera from rats immunized with MPER-SAPN.
The sera from the vaccinated rats were then tested against the ectodomain HXB2-gp41 from HIV-1. Three of the six rats showed a specific, but only moderate titer against ectodomain gp41 (Figure 6A). For comparison, both rats immunized with recombinant HXB2-gp41 (and HXB2-gp120) displayed high specific titers (Figure 6A). This disparity in the strengths of these responses may indicate that the MPER antibodies generated by immunization with MPER-SAPN recognize a different MPER conformation than those induced by the recombinant gp41 protein. On the other hand, testing of the sera against linear MPER-OVA (Figure 6B) showed that again a portion (4 of 6) of the MPER-SAPN immunized rats displayed moderate specific titers. However, in this case and in contrast to binding to gp41 protein, sera induced against recombinant HXB2-gp41 (and HXB2-gp120) exhibited only very low titers.
The results described above indicated that antibodies in sera induced by MPER-SAPN bound strongly to the MPER epitope only in its complex conformation as present in the MPER-SAPN. It was thus of great interest to establish whether these antibodies were able to neutralize HIV-1. To this end, homologous HXB2-HIV or heterologous YU2-HIV pseudotypes were incubated with dilutions of the individual rat sera (starting at 1:10), and HIV infectivity was quantified on TZM-bl indicator cells, as previously described (28). Type-specific neutralizing mouse H902 mAb and cross-neutralizing human mAb b12 were used as positive controls. Only conditions containing sera from rats immunized with recombinant gp120/gp41 or the H902 reference mAb exhibited strong type-specific neutralization (i.e., neutralized only HXB2-HIV). Human b12 mAb neutralized both strains of HIV-1. Importantly, the sera from the animals immunized with MPER-SAPN (rats 1–6) all failed to exhibit significant neutralization (data not shown).
The SAPN HIV-immunogen described here represents an attempt at creating a subunit vaccine to induce neutralizing antibodies against the MPER region of the gp41 surface glycoprotein on HIV-1. Previous attempts at presenting 2F5 and 4E10 epitopes on immunogens have not been successful in eliciting neutralizing antibodies in vivo (2,13). This failure might be due to the conformational nature of the epitopes or their accessibility on the surface of the HIV membrane (4). Therefore, we attempted to utilize a vaccine platform that allowed us to maintain the conformational context of the helical 4E10 epitope, while also presenting an extended 2F5 epitope. While the 2F5 epitope has previously been presented on a conformationally appropriate scaffold, it has failed to elicit neutralizing antibodies (14,15). Our approach differs from this previous 2F5 epitope scaffold attempt in a couple of ways. The 2F5 epitope is presented in concert with the 4E10 epitope, to elicit two sources of neutralizing antibodies. Also, the presentation platform is a repetitive antigen array provided by SAPN, which has been shown to be highly immunogenic (19,22), while the previous scaffolds did not emphasize such an approach (14,15).
The in vitro mAb-binding assay indicated similar binding profiles of the 2F5 and 4E10 mAbs on ectodomain gp41 and on our nanoparticle vaccine, MPER-SAPN (Figure 3). Therefore, we performed immunization trials in immunocompetent Sprague–Dawley rats to test whether this conformational in vitro specificity was adequate for generating an in vivo MPER-specific antibody titer that was capable of neutralizing homologous HIV-1 HXB2. In principle, the HIV-susceptible transgenic rat model, which is build on this strain of rats (16,17), would allow in vivo-follow-up studies including the intravenous challenge of vaccinated rats with X4 HIV-1 strains. Although some specific anti-MPER titers were raised, a neutralizing response could unfortunately not be detected in vitro.
The question is why neutralization by the induced MPER-specific antibodies was not witnessed? It has been proposed that secondary antibody interactions with the virus membrane may be required to maintain antibody access to the neutralization site (9,29,30). It is possible that within the infectious virion, the MPER epitope’s proximity to the HIV membrane might keep the site buried and inaccessible for effective neutralization (3,31). On both 2F5 and 4E10 mAbs, this may be facilitated by the unusually long CDR H3 loop.
In fact, this membrane proximal location could be defining the conditions for the type of antibody required. Our platform, MPER-SAPN, might only induce antibodies that engage the primary binding site on the peptide sequence, but fail to account for the secondary interactions with the membrane (29,32). These results are similar to those from a study that utilized helical rods for MPER antigen presentation and also failed to elicit nAbs (13). 2F5 and 4E10 mAbs have extended CDR H3 loops, which are potential membrane interaction domains, and this may be a prerequisite for neutralization. In line with this notion, the ablation of this complementarity determining region has been shown to reduce the effectiveness of neutralization of the 2F5 mAb (33). However, an antibody with such a long CDR H3 loop may be perceived by the host immune system as generating an auto-antigenic response. Hence, although such long CDR H3 loops may be effective in neutralizing virus, the host immune system may limit production of such antibodies because they may also target self-antigens such as cardiolipin and generate B-cell tolerance (32–35).
While the SAPN immunogen did not generate a nAb response, these nanoparticles could raise an MPER-specific titer without the need for adjuvants. This is an important step for designing any new vaccine because most adjuvants have not (yet) been approved by the FDA, and they introduce additional risk factors into any vaccine formulation (36). The SAPN’s repetitive antigen display system and its helical platform have been implicated in T-cell-dependent B-cell maturation pathways, which have generated high avidity nAbs antibodies against a malaria epitope (19) and in particular raised conformation-specific antibodies against the heptad repeat region of the surface spike protein of SARS (37), which shares many features with the surface protein gp160 of HIV. This means various HIV epitopes, other than those for 2F5 and 4E10, could be presented on SAPN, potentially generating site-specific nAbs without adjuvant. Some epitope examples could include the V3 loop, PG9-, and PG16-binding sites, and also helical domains present on different sites of the HIV surface proteins (38,39).
The work was financially supported by the University of Connecticut Research Foundation (P.B.), the Deutsches Krebsforschungszentrum (V.B.) and by the German Centre for Infection Research, University of Heidelberg (O.T.K.). P.B. hold shares in the company Alpha-O Peptides the owner of the relevant IP.