The membrane-proximal external region of HIV-1 gp41: a vaccine target worth exploring
2005; Lippincott Williams & Wilkins; Volume: 19; Issue: 16 Linguagem: Inglês
10.1097/01.aids.0000189850.83322.41
ISSN1473-5571
Autores Tópico(s)Monoclonal and Polyclonal Antibodies Research
ResumoHIV-1 vaccine design and the membrane-proximal external region (MPER) of gp41 Neutralizing antibody titers are correlated to protection for most major vaccines [1–3], but eliciting robust neutralizing antibody responses against HIV remains a major stumbling block for vaccine development [4–7]. Even during HIV-1 infection, strong neutralizing antibody titers against heterologous primary isolates are rare, and quite slow to develop [8–10]. Nevertheless, passive transfer of neutralizing monoclonal antibodies can completely protect against lentiviral challenge, at least in animal models, demonstrating a potential for vaccine-elicited antibodies to prevent an established infection with HIV-1 [11–14]. As only antibodies against conserved epitopes on HIV-1 will neutralize a broad range of different viral isolates, eliciting such antibodies should be a priority for vaccine design. Fortunately, a few human monoclonal antibodies have been described that can neutralize primary HIV-1 with exceptional breadth and potency. These antibodies target conserved regions on the envelope glycoproteins, gp120 and gp41, which putatively assemble as a trimer of gp120–gp41 heterodimers on the surface of the virus. Among the broadly neutralizing antibodies, IgG1b12 [15,16] and 2G12 [17] cross react with monomeric gp120, and strategies to elicit anti-gp120 antibodies are discussed elsewhere [18–23]. The major focus of this review will be the target on the transmembrane protein, gp41, of the broadly neutralizing antibodies 2F5, 4E10 and Z13 [24–27]. The finding that these three antibodies recognize neighboring sites on the conserved membrane-proximal external region (MPER) on gp41, defined here roughly as residues 660 to 683 of HIV-1 gp160, has brought attention to this whole region for vaccine design [26]. Subsequently, interest in the MPER of gp41 as a vaccine target has risen, as numerous studies have broadened our understanding of the structure and function of the MPER of gp41, as well as of the antibodies that bind to it. It should be made clear, though, that a vaccine designed to elicit neutralizing antibodies against HIV will most likely integrate conserved targets on both gp120 and gp41. It should also be established that non-neutralizing antibodies appear to dominate in the antibody response against gp41, both during natural infection, and through immunization [28–30]. These antibodies typically target membrane-distal regions, which tend to be occluded by quaternary interactions within the HIV-1 envelope glycoprotein (Env) trimer [31–33]; the MPER appears to be an exception. This review will cover what is known about the structure and function of the MPER of gp41, compare it with other vaccine targets, and address some of the challenges of exploiting membrane-proximal epitopes on HIV-1 in vaccine design. In-vivo and in-vitro anti-HIV-1 activity of 2F5, 4E10 and Z13 The main justification for targeting the MPER of gp41 for vaccine design is the great breadth with which the human monoclonal antibodies 2F5 and 4E10 neutralize primary isolates of HIV-1 (see below). Importantly, however, these antibodies, often in various cocktails, can contribute some level of protection against infection in animal models, although their individual contributions are sometimes difficult to assess [11,13,14,34–36]. In a recent clinical study, a high-dose cocktail of 2F5, 4E10 and 2G12 could delay viral rebound after cessation of antiretroviral therapy (ART) in some individuals [37]. In that study, the average serum levels were higher for 2G12 than 2F5 and 4E10, and viral escape mutants to 2G12, but not 2F5 and 4E10 were observed. Many questions about the clinical efficacy of 2F5 and 4E10 have now been raised that will require future studies, perhaps using each antibody at higher doses. Note that neutralizing antibodies are not expected to have a profound effect once an infection with HIV-1 has been established, but may be more effective at lower levels in a vaccination setting, in which neutralization of free virus may have a greater impact [34]. In the most comprehensive study of cross-clade neutralization to date, 2F5 and 4E10 neutralized 67 and 100%, respectively, of the diverse primary isolates tested (n = 90) [38]. Similarly, 2F5 and 4E10 neutralized 80 and 100%, respectively, of viruses (n = 91, presumably clade B) that were cloned directly from newly infected patients that had a negative or indeterminate serology [39]. In both of these examples, a relatively sensitive pseudovirus assay was used. Some experiments involving a classical peripheral blood mononuclear cell (PBMC)-based assay have shown that these antibodies can be less potent in this format, although still remarkably broad in their activity. For example, at 50 μg/ml, 2F5 and 4E10 neutralized (IC90) 60 and 8%, respectively, of primary isolates (n = 25) in a PBMC assay [38]. In other studies using the PBMC assay, 4E10 neutralized 71% (n = 28), 82% (n = 22) or 98% (n = 58) of the primary isolates tested [26,27,40]. In the latter study, an increased susceptibility to 4E10 and 2F5 of viruses taken during acute, relative to chronic infection was observed [40]. The molecular bases of these observed differences, and of some of the discrepancies in neutralization between pseudovirus and PBMC-derived virus, are not yet clear but may be instructive for vaccine design. A clear understanding of the mechanism of neutralization of HIV-1 by antibodies against the MPER is essential for rational approaches to elicit such antibodies. In general, neutralization of HIV-1 by antibody has been attributed to the ability of the antibody to bind to the native Env trimer, largely irrespective of epitope specificity [32,41–45]. Accordingly, studies have suggested that this explanation also extends to 2F5 [46,47]. To complicate the matter slightly, however, 2F5 and 4E10 also appear to neutralize subsequent to receptor engagement, a mechanism of inhibition that has been well established for the peptide drug, T20 (also named DP178, fuzeon and enfuvirtide) [48,49] and for other polypeptides corresponding to the N-heptad repeat (NHR) and C-heptad repeat (CHR) regions of gp41 [50–52] (Fig. 1). For example, a mutant HIV-1 pseudovirion, in which infectivity is controlled by an inducible disulfide bond in Env, is neutralizable by 2F5 and 4E10 post-attachment, albeit at a diminished potency relative to the neutralization observed under standard conditions [53]. Other studies have suggested that the 2F5 epitope is present on both native and a fusion-intermediate form of gp41 [54–56]. Moreover, substitutions that appear to affect Env stability and/or fusion kinetics have been shown to increase the sensitivity of HIV-1 to 2F5 and 4E10 [57–59]. Thus, there are many potential factors, both structural and kinetic, that can affect the access of antibody to the MPER (Table 1). It is also noteworthy that 2F5 and 4E10 can neutralize with several-fold greater potency as whole IgGs as compared to Fabs [60,61]. These results suggest that increased bulk does not impair access of bivalent antibodies to the MPER, in contrast to what has been observed with whole IgGs against CD4-induced epitopes that overlap the coreceptor binding site on gp120 [62]. However, access to the MPER may still be limited, as an IgM version of 4E10 appears to have little neutralizing activity [63].Fig. 1: Model of HIV-1 envelope-mediated fusion and its inhibition by targeting gp41. (a) A cartoon showing that, prior to attachment of gp120 to CD4 and co-receptor (CoR), gp41 is mostly occluded within the trimeric Env spike, leaving the MPER of gp41 (pink) partially exposed. Antibody binding to the MPER on the trimer is inferred from the neutralizing activity of antibodies 2F5, Z13 and 4E10. (b) Following receptor activation, gp41 assumes a transient fusion-intermediate state in which the N-terminal fusion peptide (orange) inserts into the host cell membrane. Access to this 'pre-hairpin' structure is brief (several minutes) and is somewhat restricted by the close proximity of the viral and host membranes and by other host and viral proteins. (c) Membrane fusion occurs upon formation of a stable six-helix bundle structure. (d) Fusion may be blocked by 'class 1' (e.g. T20, C34) or 'class 2' (e.g. 5-helix, IZN17, N35-NCCG-N13) inhibitors, against NHR or CHR regions, respectively [33,51,52,145]; or by 2F5, Z13 or 4E10, which may neutralize by binding to the native Env trimer (as in Panel (a)), but may also have activity post-receptor engagement (Panel (d); see text).Table 1: Potential factors affecting exposure of gp41 MPER to antibodies.Fab Z13 is less well characterized than 2F5 and 4E10, and whereas the latter two antibodies were obtained using combined polyethylene glycol-electrofusion of PBMCs and heteromyeloma cells [25], Z13 was isolated from an antibody phage display library prepared from the bone marrow RNA of a donor whose sera had broad neutralizing antibody titers [26]. Z13 binds to a determinant that overlaps the 4E10 epitope, has cross-clade neutralizing activity, but apparently neutralizes primary isolates with less breadth and potency than 4E10 [26]. In addition, Fab Z13 appears to bind with poorer affinity than 4E10 to the MPER, and its activity is dependent on a somewhat variable residue on gp41, D674, by which 4E10 recognition is less affected [26]. Future structural studies with Z13, and comparison of its epitope with that of 4E10 may shed light on how best to present the most conserved elements of this region to the immune system. HIV-1 gp41 MPER structure The first biophysical insights into the MPER of gp41 came with the determination of the NMR structure of a 19-mer peptide, KWASLWNWFNITNWLWYIK, comprising residues 665 to 683 of HIV-1 gp160 (HxB2 numbering). This peptide adopts an alpha helical conformation in dodecylphosphocholine micelles in which the aromatic and polar residues are distributed around the helical axis [64] (Fig. 2, middle). Classical amphipathic helices, in contrast, have more clearly segregated polar and non-polar faces. The defined helical structure of this MPER peptide shares a commonality with particular regions of the gp41 ectodomain that had previously been shown to fold into tight helical bundles using X-ray crystallography. Thus, the NHR region, and the CHR segment (Fig. 2, top), the latter of which abuts the N-terminal end of the MPER of gp41, can fold anti-parallel to one another such that the NHR polypeptides form a trimeric helical core around which the CHR helices pack at a slightly oblique angle [65–67]. The resulting 'six-helix bundle' is extremely stable and the favorable energy of its formation is thought to drive the fusion of the virus and host cell membranes during infection [33]. Importantly, prior to membrane fusion, or even receptor engagement, gp41 is thought to be in a very different conformation, and much less is known of the structure of this native state. C-terminal to the MPER is the membrane-spanning domain, which is most likely helical, as single-pass membrane-spanning domains are most stable as alpha helices in membranes [68].Fig. 2: Various structures overlapping the membrane-proximal external region (MPER) of HIV-1 gp41. A schematic diagram of HIV-1 gp41 depicts the hydrophobic fusion peptide (FP), N-terminal and C-terminal heptad repeat regions (NHR and CHR), MPER (defined here as residues 660-683 of gp160), the transmembrane (TM) domain, and the internal cytoplasmic domain. The clade B consensus sequence of the MPER is shown, below which is an alignment of the residues that differ among clades A, C, D and Group O, as determined using the Los Alamos database (http://www.hiv.lanl.gov/content/hiv-db/CONSENSUS/CONSENSUS_TOOL/consensus.html). The residues crucial for 2F5 and 4E10 neutralization are marked by asterisk [58]. A tube representation of the CHR helical region (top, yellow), bears the N-terminal portion of the 2F5 core epitope, shown in pink. The structure of the peptide corresponding to residues 666-683 of gp160, as determined in dodecylphosphocholine micelles using NMR [64], is shown (middle, yellow) with the core residues of the 2F5 and 4E10 epitopes colored in pink and blue, respectively. These same crucial residues are colored similarly in the structures of the 2F5- and 4E10- bound peptides (bottom, left and right, respectively), as determined by X-ray crystallography [61,70]. The structural coordinates of 1JAV [64], 1ENV [66], 1TJI [61], and 1TZG [70] were obtained from the PDB database and rendered using PMV software [146].It would be best for vaccine design to know the structure of gp41 in the context of the virus. However, there are serious practical limitations in probing this structure that relate to the heterogeneity and instability of HIV-1 virion particles. The antibodies 2F5, Z13 and 4E10 are therefore useful as structural probes, since they bind to the functional Env trimer (as prerequisite of their neutralizing capacity). At first, the X-ray crystal structure was determined for a complex of 2F5 with a 7-mer peptide, ELDKWAS [69], then, later, with a 17-mer peptide, EKNEQELLELDKWASLW [61]. These studies reveal that this portion of the MPER is not in a helix, but an extended conformation with a distinct Type I β turn at the 'DKW' in the core of the peptide epitope [61,69] (Fig. 2, bottom left). The crystal structures also show that the residues 'DKW' are among the most buried in the interface, indicating that these residues are exposed on the Env trimer prior to fusion. More recently, a crystal structure was determined of a complex of 4E10 and a peptide bearing the sequence, WNWFDITNW; the peptide is shown bound to 4E10 in a helical conformation [70] (Fig. 2, bottom right). Unlike the corresponding region on the NMR structure of the longer MPER peptide in micelles, the 4E10-bound peptide has a characteristic 'loosening' of the alpha helix within, and just N-terminal to the residues W672 and F673, which are deeply buried in the paratope of 4E10 [70]. Thus, W672, which buries most deeply into 4E10, must be well exposed at some point prior to fusion. If we assume that most of the MPER is helical at some point in time on the Env trimer, this would position an exposed or 'neutralizing' face of the helix on the exterior of the trimer, and the opposite or 'non-neutralizing' face would be occluded within the Env trimer (or in the membrane interface as some have argued). Note that the unfavorable energy of exposing the aromatic side-chains to solvent prior to 4E10 recognition must be balanced by favorable interactions made elsewhere by Env. Nuclear magnetic resonance and other spectroscopic techniques have also been used to show the structural preferences of unconstrained 2F5-epitope peptides in solution, although the application of this knowledge to vaccine design has been problematic. For example, a 2F5-binding peptide has been shown to adopt a 3-10 helix in solution [71]. In addition, 2F5 preferentially binds to mutants of the peptide DP178 having the greatest helical character [72]. In contrast, a β-lactam constraint, engineered into a 2F5-epitope peptide to stabilize a β-turn, enhances recognition of the peptide by 2F5 [73]. The multispecific capacity of 2F5 may explain these findings, as affinity-optimized peptides may bind to 2F5 using slightly different binding mechanisms [74]. It was recently suggested that the 2F5-epitope peptide, ELLELDKWASLWN, has structural characteristics that dictate its corresponding structure (and function) within the HIV-1 Env spike, acquiring, at most, only minimal structure from the virus context [46]. In support of this, when the 2F5 epitope sequence was grafted into the MPER of the VSV G protein, the resulting VSV virions bearing the 2F5-epitope were shown to be neutralized by 2F5 [75]. It must be noted, however, that it can be misleading to relate the binding affinity of a neutralizing antibody for an artificial antigen to the success in mimicry between that antigen and the corresponding viral epitope. Indeed, alanine-scanning mutagenesis in the MPER of gp41 on infectious virus showed that the only changes to produce neutralization resistance to 2F5 occurred in residue D, K, or W of the core epitope; likewise, 4E10 resistance arose by replacing three residues, two (W and F) were in the core epitope, and one (W) was seven residues C-terminal to these two (NWFDISNWLW) [58]. Therefore, residues on the virus that are not crucial for the neutralizing activity of 2F5, can boost the affinity of cognate peptides, to 2F5 in the context in apparently irrelevant ways. Despite the above, one should not rule out other conformations of the MPER of gp41, to which as yet undiscovered neutralizing antibodies may bind. Does the short segment that intervenes between the 2F5 and 4E10 epitopes prefer an alpha-helical or extended conformation, and is it accessible to neutralizing antibodies? By its proximity to the membrane, the MPER of gp41 may acquire relatively high molecular mobility that is characteristic of lipids in membranes. Extreme rigidity, in this case, may not be a desirable feature of a designed immunogen based on the MPER of gp41, since antibodies may need to accommodate a more conformationally variable target. This may be particularly important in vivo if membrane composition is found to vary among different target cell types and influences MPER conformation. Notwithstanding, as a reasonable start for vaccine design, the MPER of gp41 should be presented in a way that mimics the structure that is recognized precisely when 2F5 and/or 4E10 bind (i.e. neutralize) the virus. Note that the virus itself may not be the best immunogen, even with safety considerations put aside, given the apparent paucity of 2F5 and 4E10-like antibodies that are elicited during natural infection, and that considerable heterogeneity and non-neutralizing epitopes are present in HIV-1 viral preparations [76]. Another caveat is that, in comparison to the neutralizing epitopes on gp120, those recognized by antibodies b12 and 2G12, the 2F5 and 4E10 epitopes appear to be less exposed on virus [26,77,78] and on infected cells [26,32]. HIV-1 gp41 MPER function Several studies have shown the importance of the MPER of gp41 in viral or cell to cell fusion. Simple deletion of residues 665-682 abrogates syncytium formation, for example [79,80]. However, Ala-substitution of the five conserved W residues produces a phenotype in which Env does not induce syncytia but does permit lipid–dye mixing between host and target cells, suggesting a failure in fusion-pore expansion with this mutant [80]. In a recent study to clarify the defect in the fusion mechanism with the deletion mutant Δ665-682, it was found that the mutant Env, displayed on the host cell surface, was fully competent to form six-helix bundle structures following activation by CD4-CXCR4 complexes [81]. The deletion in the MPER of gp41 was suggested to restrict the flexibility in the C-terminal portion of the ectodomain such that the fusion peptide is unable to insert into the target cell membrane, and that self-insertion of the fusion peptide leads to membrane destabilization and the dye leakage that was observed. In this model, the MPER of gp41 serves as a 'linker' which positions the fusion apparatus in a conformation compatible with fusion at the appropriate time. Whether additional features of the MPER of gp41 also contribute to the fusion and/or viral entry mechanisms is not entirely clear. The same MPER peptide that was shown to be an alpha-helix in DPC micelles using NMR can also bind and disrupt membranes. Thus, it has been suggested that the MPER interacts directly with the membrane during fusion [64,82]. The notion of an interaction between the MPER and lipid has been re-iterated in structural studies with the 2F5 epitope [46,61], and in studies showing that lipids enhance the affinity of 2F5 and 4E10 to recombinant gp140s that have been captured on proteoliposome beads [61,83]. For simian immunodeficiency virus (SIV), the stability of the bundle itself depends on sequences N-terminal to the N-HR domain and on the MPER [84]. Thus, for SIV at least, the MPER may directly contribute energy toward the fusion reaction via bundle formation. Finally, the MPER of gp41 may affect overall Env production or transport. For example, subtle changes to the MPER of gp41 (e.g. W672F) can decrease the incorporation of Env into replication-competent virions and diminish virus infectivity, although gp160 appears to be processed and expressed normally on the cell surface, and syncytium-formation appears to be unaffected with these mutants [79]. The above studies on the function of the MPER of gp41 help to explain its sequence conservation. However, many substitutions to the MPER of gp41 have little or no effect on virus infectivity, especially in envelope-complementation (pseudovirus) formats [58,85]. Furthermore, the natural sequence variation that occurs in the MPER of gp41 itself is an indication that some aspects of this region of gp41 are not directly required for infection. Recently, a replication-defective mutant of HxB2 with three Trp to Ala substitutions in the MPER was serially passaged in culture and a variant was selected that restored replication competence to wild-type levels; the compensatory mutation was a single substitution just N-terminal of the MPER [86]. If limited sequence variation and targeted mutagenesis is well-tolerated in the MPER of gp41, then why are some residues still so highly conserved? Perhaps a major reason for the particular sequence conservation that is observed in the MPER of gp41 is that these sequences are the least immunogenic in context of the virus, and thus the most successful in the viral population. Thus, there are at least two competing selection pressures that shape the MPER of gp41: antibody avoidance as a viral defense mechanism, and necessity of function (viral offence), whether that be fulfilling the role of a specialized linker for the fusion apparatus, controlling the energetics of fusion via an extended coiled-coil, or by maintaining Env trimer stability (Table 2).Table 2: Putative functions of the HIV-1 gp41 membrane-proximal external region (MPER).Aromatic residues in membrane interfaces Many other viruses also have clusters of aromatic residues in the MPER region of their transmembrane glycoproteins (TMs), including SIV, feline immunodeficiency virus (FIV), equine infectious anemia virus (EIAV), 'South African ovine Maedi Visna virus' (SA-OMVV; or 'ovine lentivirus', or 'Visna virus'), Moloney murine leukemia virus (MoMuLV), Ebola virus, vesicular stomatitis virus (VSV), and SARS-CoV, coronavirus linked to severe acute respiratory syn [82,87,88]. For FIV, it has been shown that deletions within a W-rich portion of the MPER prevent productive viral infection without affecting virus attachment to cells [87]. Likewise, substitution for Ala of three conserved W-residues in the MPER of FIV TM abrogates viral fusion [87,89]. Interestingly, an octapeptide corresponding to this region on FIV can block viral replication in PBMCs, an activity that is inhibitable by a peptide from the N-terminal portion of the TM protein [90]. The MPERs of the F protein from paramyxovirus SV5 and human parainfluenza type 2 viruses are important for viral fusion as well [91,92]. For SV5 F, however, only extension, but not deletions or mutations within the 7 aa long MPER abrogates fusion; this apparent permissibility has been attributed to a compensatory effect of the unusually long and flexible intervening region between the NHR and CHR regions of the SV5 F protein [93]. Even the neuronal SNARE proteins, VAMP and syntaxin 1a, which are involved in intracellular and vesicular fusion, have W residues in their MPERs. The insertion of soluble linkers in the MPERs of these SNARE proteins can diminish fusion efficiency [94]. Notwithstanding, some viruses do not appear to require their membrane-proximal aromatic residues for viral fitness or entry into target cells. In the case of VSV-G, for example, mutagenesis of the highly conserved membrane-proximal W residues has no effect on viral replication and little effect on cell–cell fusion activity, although deletion of the entire MPER domain renders the virus completely non-infectious [95]. Studies on many transmembrane proteins of diverse architecture including the β-strand-rich porins, helix-bundle proteins like cytochrome c oxidase, as well as single-span membrane proteins have shown that Trp residues are typically found at the membrane interface, that is, between the polar head groups and the acyl chains of membrane lipids [96,97]. Trp-rich antimicrobial peptides appear to bury Trp side-chains in the membrane interface as well [98]. As mentioned above, peptides corresponding to the MPER of gp41 have been shown to bind and disrupt liposomes [82]. Moreover, gp41 peptides that overlap with the MPER can aggregate in an array on the membrane surface [99], or have lectin-like activity [100]. In addition, several studies have demonstrated the cholesterol-sequestering ability of the 5-mer motif, LWYIK, in the MPER of gp41 [101–103]. The latter studies have related the cholesterol-binding activity to the ability of HIV-1 Env to partition into cholesterol-rich lipid raft domains. Whether or not this activity has a direct effect on the antibody response to the MPER of gp41 is unknown. Despite the strong interfacial propensity of Trp residues, it is not inconceivable that Trp residues may transiently dip deeper into the membrane core, although this would be energetically costly. Interestingly, Trp residues in the MPER of the SNARE protein, VAMP-2, may provide an anchor for the core-region that is suggested to be angled obliquely to the membrane such that basic residues can interact with the negatively charged lipid heads. The MPER of VAMP-2 serves then as a regulatory module for subsequent membrane merger that is driven by coiled-coil formation in the core-region [104,105]. Such studies have highlighted that, even if the force generated by bundle formation is enough to bring the membranes together, this energy may be dissipated if the MPER is not appropriately configured. Immunogenicity considerations for membrane-proximal epitopes In addition to the synthetic peptide immunogens mentioned above, a variety of recombinant immunogens have been engineered to display the 2F5 epitope using such proteins as the hepatitis B surface antigen [106], influenza virus haemagglutinin [107], maltose binding protein [108], gp120 [109,110], potato virus X coat protein [111], and anti-HLA-DR immunoglobin molecule [112]. Despite eliciting high antibody titers against the epitope sequence, these immunogens have not elicited neutralizing antibody against primary HIV-1. More recently, some neutralizing antibodies have been detected with immunizations using bovine papilloma-HIV-1 gp41 chimeric virus-like particles [113], and with rhinovirus display [114]. However, these approaches will require further and independent evaluation. Overall, the above studies indicate how difficult it is to elicit 2F5-like antibodies, and how alternate conformations or surfaces that overlap the 'neutralizing' face of the 2F5-epitope sequence on these immunogens elicit predominately non-neutralizing antibodies. It may be that some non-neutralizing antibodies to irrelevant conformations of the MPER are elicited during natural infection, although there is no direct evidence so far to support this. Recently, a fluorescence-based binding assay was used to demonstrate the presence of high titers of antibodies against the fusion peptide, NHR and CHR regions, and the MPER of gp41 among several seropositive individuals, suggesting that no region of the gp41 ectodomain is immunologically silent [115]. The same study showed no correlation between neutralizing antibody titer and the antibody titer against any of the gp41 antigens tested. In another study, a panel of HIV-1 patient sera was screened using overlapping peptides and an immunodominant site was identified in which the residues WNWFDI were most critical for Ab recognition [116]. However, other serum panel screens using overlapping peptides do not usually identify the MPER as an immunodominant site, at least in comparison with some other determinants on Env such as V3 on gp120 or the immunodominant loop on gp41 [117–119]. Note that antibodies against conformational epitopes may be overlooked using peptide screens. Whatever their prevalence (or lack thereof), non-neutralizing antibodies in serum recognizing the WNWFDI motif or irrelevant conformations of the MPER of gp41 may confound attempts to elicit neutralizing antibodies to the MPER, so their existence must be considered. Since antibodies are 'elicited' as selectable B-cell receptors (BCRs), prior to their production as soluble immunoglobins by committed plasma cells, steric restrictions imposed by the viral membrane could conceivably limit the immunogenicity, or at least the angle of approach by the BCR to the MPER of gp41. The immunogenicity of haptens that are linked to lipid heads on a liposome is optimal, for example, at intermediate linker length, but low at very short and very long linker lengths; this has been attributed to the tendency of the membrane to mask the hapten when it is close to the membrane [120,121]. Clearly, an antibody cannot approach the MPER from beneath the membrane in which it resides. An antibody against the membrane-proximal end of the Env spike (haemagglutinin) of influenza virus appears to illustrate this point. The murine monoclonal antibody, LMBH6 was elicited with bromelain-cleaved haemagglutinin (BHA) and, in contrast to other antibodies to HA, binds much better to BHA than to virus or infected cells, ostensibly because the approach to its epitope on the virus is hindered by the membrane [122].
Referência(s)