HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 48, 45811-45820, November 29, 2002

This Article Abstract Full Text (PDF) An addition or correction has been published All Versions of this Article: 277/48/45811    most recent M205862200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Joyce, J. G. Articles by Keller, P. M. Search for Related Content PubMed PubMed Citation Articles by Joyce, J. G. Articles by Keller, P. M. Social Bookmarking                      What's this?

Enhancement of -Helicity in the HIV-1 Inhibitory Peptide DP178 Leads to an Increased Affinity for Human Monoclonal Antibody 2F5 but Does Not Elicit Neutralizing Responses in Vitro

IMPLICATIONS FOR VACCINE DESIGN^*

Joseph G. Joyce^§, William M. Hurni, Michael J. Bogusky^¶, Victor M. Garsky^¶, Xiaoping Liang, Michael P. Citron, Renee C. Danzeisen, Michael D. Miller, John W. Shiver, and Paul M. Keller

From the  Departments of Virus and Cell Biology, ^¶ Medicinal Chemistry, and  Biological Chemistry, Merck Research Laboratories, West Point, Pennsylvania 19486

Received for publication, June 12, 2002, and in revised form, August 29, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The synthetic peptide DP178, derived from the carboxyl-terminal heptad repeat region of human immunodeficiency virus type 1 GP41 protein is a potent inhibitor of viral-mediated fusion and contains the sequence ELDKWA, which constitutes the recognition epitope for the broadly neutralizing human monoclonal antibody 2F5. Efforts at eliciting a 2F5-like immune response by immunization with peptides or fusion proteins containing this sequence have not met with success, possibly because of incorrect structural presentation of the epitope. Although the structure of the carboxyl-terminal heptad repeat on the virion is not known, several recent reports have suggested a propensity for -helical conformation. We have examined DP178 in the context of a model for optimized -helices and show that the native sequence conforms poorly to the model. Solution conformation of DP178 was studied by circular dichroism and NMR spectroscopy and found to be predominantly random, consistent with previous reports. NMR mapping was used to show that the low percentage of -helix present was localized to residues Glu⁶⁶² through Asn⁶⁷¹, a region encompassing the 2F5 epitope. Using NH₂-terminal extensions derived from either GP41 or the yeast GCN4 leucine zipper dimerization domain, we designed peptide analogs in which the average helicity is significantly increased compared with DP178 and show that these peptides exhibit both a modest increase in affinity for 2F5 using a novel competitive solution-based binding assay and an increased ability to inhibit viral entry in a single-cycle infectivity model. Selected peptides were conjugated to carrier protein and used for guinea pig immunizations. High peptide-specific titers were achieved using these immunogens, but the resulting sera were incapable of viral neutralization. We discuss these findings in terms of structural and immunological considerations as to the utility of a 2F5-like response.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The HIV-1¹ GP160 envelope glycoprotein is synthesized as a single precursor that is cleaved by a cellular endoprotease to generate two noncovalently associated subunits, GP120 and GP41 (1). GP120 is the receptor-interacting constituent, which mediates sequential binding to the cellular CD4 receptor and CXCR4 or CCR5 chemokine co-receptors. The GP41 transmembrane subunit mediates fusion between viral and cellular membranes (reviewed in Refs. 2 and 3). Both envelope segments are highly immunogenic but antibodies raised by vaccination with full-length or subunit versions of these proteins, while capable of neutralizing homologous virus, are generally not protective against heterologous challenge (4-6). In contrast, several broadly neutralizing human monoclonal antibodies (mAb) to diverse envelope epitopes have been identified in recent years. Among the most well characterized of these are 1b12 and 2G12, which bind to GP120 (7-9) and 2F5, which interacts with the COOH-terminal region of GP41 (10). MAb 2F5 has generated much interest because its epitope is well conserved across HIV clades and because of its ability to neutralize both laboratory-adapted and primary viral isolates (11).

The envelope protein undergoes a conformational change upon receptor binding, with significant perturbations occurring in both subunit domains. It is postulated that the structural rearrangement that takes place in GP41 provides the energy required to bring viral and cellular membranes into apposition and to allow proper positioning of the NH₂-terminal fusion peptide for bilayer interaction (12-15). This change from a prefusogenic to fusion-competent state has been intensively studied over the past several years, and the currently accepted model of the fusion-active conformation is that it is composed of a six-stranded helical bundle formed by association of NH₂- and COOH-terminal heptad repeat regions of the ectodomain (16-18). The interior of the bundle consists of a trimeric coiled coil formed by three NHRs comprising residues 546-581 (numbering based on the HXB2 GP160 variant as described in Ref. 17). This inner bundle is surrounded by three outer COOH-terminal helices that pack around the trimer in an antiparallel configuration. Proteolytic dissection studies identified a highly thermostable core structure as being composed of peptide domains N51, encompassing residues 540-590 and C43, encompassing residues 624-666, joined by a short linker sequence (19). Subsequent studies showed that shorter peptides such as N36/C34 (17) and N34/C28 (18) also associate to form stable core-like assemblies.

A number of isolated peptides whose sequences overlap either the NHR or CHR domains have been shown to be potent inhibitors of viral infectivity. These include N51 (19), C34 (20), and DP178 (21). Several lines of evidence suggest that these compounds act in a dominant-negative manner, inhibiting fusion by binding to a nascent prehairpin derivative formed during the receptor-induced conformational change. The 36-mer peptide DP178, encompassing residues 638-673, has been the most well studied of these compounds. DP178, also known as T20, comprises the final 24 residues of C34 and continues to Phe⁶⁷³, reaching approximately halfway into the tryptophan-rich region defined by residues 665-683, which is postulated to maintain important contacts with the viral membrane (22). Residues 662-667, ELDKWA, form the recognition epitope for mAb 2F5. Solution studies of DP178 show it to be largely unordered with a low content of -helix. However, when DP178 is mixed with a NHR-derived peptide such as N36 a large increase in helicity is observed (23, 24). The magnitude of this enhancement is higher than that predicted for theoretically noninteracting species, confirming that DP178 undergoes a conformational change upon interaction with N36 (23).

The structure of DP178 in the prefusogenic configuration on the native virion is unknown. Oligomerization studies of soluble, uncleaved GP140 constructs stabilized by trimeric GCN4 leucine zipper or T4 bacteriophage fibritin motifs offer compelling evidence that a trimeric structure is consistent with known antibody reactivities (25). Furthermore, these studies suggest that the trimeric forms more closely resemble the prefusogenic state as supported by their low affinity for mAb NC-1 (26) and peptide DP178. The association of the CHR with the trimeric NHR bundle in the fusion-competent configuration is not that of a coiled coil because the orientation is tilted with regard to the core helices (17). Nevertheless, predictive computational studies show a high propensity for the CHR region to exist as an amphipathic coiled coil (27). Rabenstein et al. (24) showed that the heptad repeat motif found in the CHR can be described as having "imperfect" amphipathic character and postulated that the entire sequence span from Val⁶⁰⁸ to Leu⁶⁸⁴ is likely helical in the viral fusion-competent state. Finally, Lawless et al. (23) had shown that self-association of DP178 occurred in a concentration-dependent manner, with tetramers forming at peptide concentrations greater than 20 µM. They suggest that the native prefusogenic state could involve an oligomeric T20 structure that re-arranges upon CD4 binding to form the final highly thermostable 6-helix bundle.

DP178 effectively inhibits both cell-free viral infectivity and cell-cell fusion at low nanomolar concentrations, and it is currently being evaluated as a therapeutic in human clinical trials (28). The potential value of this compound is further enhanced by observations that primary viral isolates are slow to develop resistance to it. However, a relatively long peptide such as DP178 suffers from several limitations including bioavailability, proteolytic sensitivity, and eventual mutation-induced resistance. Alternatively, the use of the peptide as a vaccine candidate is attractive because it contains the recognition epitope for 2F5. However, fusion proteins containing the epitope have uniformly failed to produce a broadly protective response (6, 29). One explanation for this may be that the 2F5 epitope was not presented in the correct structural context in these constructs. Sattentau et al. (30) showed that whereas 2F5 could bind to virally infected cells, binding was abrogated when these cells were first treated with soluble CD4. Similar results were observed in a recent study of binding to synthetic peptides in solution (31). These data suggest that 2F5 recognizes either the native prefusogenic conformation of ELDKWA or a transient intermediate structure. Studies such as those by Judice et al. (32) in which chemical constraints were introduced into DP178 to maintain helicity, and Root et al. (33) who prepared a linear protein that self-folded into a stable 5-helical bundle show that enhanced infectivity inhibition can be obtained with highly constrained structures. Finally, a recent crystallographic report of the peptide ELDKWAS bound to the Fab' fragment of 2F5 suggests that the epitope adopts a defined turn (34).

We were interested in evaluating whether increasing the helicity of DP178 constructs led to an enhanced interaction with mAb 2F5 and if these constructs were capable of eliciting a broadly neutralizing antibody response. We decided to optimize the amphipathic character of the peptide with the reasoning that if this provided a significantly better interaction with 2F5, it would support the hypothesis for interacting CHR subunits in the prefusogenic virion. Working from a model that described optimization of amphipathic helices in peptides (35) we prepared chemically unconstrained peptide constructs designed to present the ELDKWA epitope in the context of a highly -helical structure and used circular dichroism (CD) to confirm the predictions of the model. To determine whether presentation of the ELDKWA epitope in a helical format was a supportable hypothesis, we studied the structure of native DP178 by NMR and show, for the first time, that the carboxyl-terminal portion of DP178 accounts for the low -helical content reported in previous studies. To quantitatively rank the ability of these peptides to bind to mAb 2F5 we developed a true solution-based competitive binding assay that represents a significant improvement over previously described solid-phase enzyme-linked immunosorbent assay systems in terms of preserving peptide structure. Finally, to determine whether these structured peptides were capable of generating a 2F5-like immune response, modified DP178 peptides were conjugated to a carrier protein and tested in guinea pig vaccination experiments. We discuss the implications regarding the ability of these immunogens to elicit high titer, nonneutralizing responses with regard to efforts aimed at generating 2F5-like responses by peptide immunization.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Peptide Synthesis-- Synthetic peptides corresponding to DP178 and modified analogs were prepared by standard t-Boc solid phase synthesis in-house or purchased from Bio-Synthesis, Inc. (Lewisville, TX). Peptides were synthesized as the free amino and carboxyl forms. Peptides intended for conjugation to carrier proteins were synthesized with a cysteine residue on either the amino or carboxyl terminus and these constructs were made as the N-acylated and C-amidated derivatives. Peptides that were not directly soluble in water or neutral pH buffer were initially solubilized at 1-5 mg/ml in dilute NaOH and immediately neutralized by dilution into appropriate buffers. Solution concentrations were determined by quantitative amino acid analysis.

Circular Dichroism (CD) Spectroscopy-- CD measurements were made at 25 °C on a Jasco model J810 spectropolarimeter equipped with Peltier temperature controller and running Jasco version 1.52.01 Spectra Manager software. Spectra were acquired from 180 to 250 nm using a bandwidth of 1 nm and data pitch of 0.1 nm at a scan speed of 100 nm/min with 10 accumulations per sample. Spectra were corrected for solvent contribution and the CD signal was converted to molar ellipticity, [], by [] = _observed (MRW/10 × l × c), where MRW is the mean residue weight (molecular mass divided by number of peptide bonds), l is path length, and c is concentration in mg/ml. The fractional percentage of  helix (f) was estimated from []_{222 nm} by, 100% f = 40,000 [1  (2.5/n)], where n = number of peptide bonds. For helix-induction studies, samples were adjusted to a final concentration of 30% 2,2,2-trifluoroethanol (TFE).

NMR Analysis-- NMR spectra were recorded on a Varian Inova 600 MHz spectrometer at 25 °C. Samples for NMR measurements contained 0.5-1.5 mM peptide in either phosphate-buffered saline (PBS, 6.25 mM sodium phosphate, pH 7.2, 0.15 M NaCl) or 50% CD₃CN, 50% H₂O (v/v) or 50% CD₃CN, 50% D₂O (v/v) in a final sample volume of 0.650 ml. The water resonance in all experiments was suppressed by low power saturation using an attenuated transmitter pulse. All two-dimensional experiments were acquired in the hypercomplex mode for phase-sensitive presentation. ¹H chemical shifts were referenced to sodium 3-(trimethylsilyl)propionate-2,2,3,3-d₄ at 0.00 ppm. Sequence-specific chemical shift assignments were obtained with the combination of TOCSY (36, 37), NOESY (38), and ROESY (39) experiments using standard methodologies (40). Clean TOCSY (37, 41) spectra were recorded with 1 K complex points in t₂ and 512 points in t₁ consisting of 4-32 transients per increment. Spin locking was achieved with an MLEV16 + 60° mixing sequence for a duration of 50-75 ms preceded by a 2.0-ms trim pulse. Data sets were multiplied by a shifted Gaussian apodization function and zero filled to 2 K by 1 K complex points prior to Fourier transformation. ROESY spectra were acquired using a spin lock radiofrequency strength, H₂ of ~3.8 kHz with a 100-ms mixing period. Spin locking was achieved by the application of a series of 30° pulses to minimize TOCSY artifacts in the spectrum and two hard 90° pulses on either side of the spin lock to compensate for resonance offset effects (42). The data were acquired with 1 K complex points in t₂ and 512 points in t₁ consisting of 32-96 transients per increment. The data was processed as described above. NOESY spectra were acquired using mixing times of 100-500 ms. Solvent suppression was achieved by selective saturation of the water resonance during the recycle delay, t₁ period, and mixing period. Data sets were acquired with 1 K complex points in t₂ and 512 points in t₁ with 32-64 transients per increment. The data was processed as described above. The NOE intensities were analyzed by volume integration of the cross-peaks in well resolved regions of the spectrum.

MAb 2F5 Reactivity Assay-- Eight-well MAXISORP^TM Immuno^TM Module strips (Nalge Nunc, Rochester, NY) were coated with mAb 2F5 (Polymun, Vienna, Austria) at 20 ng/well in PBS and overcoated with 1% bovine serum albumin in PBS containing 0.1% sodium azide. The reference peptide (biotin-DP178) was Ac-Cys-DP178-NH₂, which had been biotin-labeled by reaction with PEO-maleimide activated biotin (Pierce) as per the manufacturer's protocol and purified by gel filtration chromatography. The reference peptide was used at 8 × 10¹⁴ mol/well. Test samples were dissolved at a nominal concentration of 1 mg/ml (w/v) and diluted 500-fold in assay diluent (1% bovine serum albumin in PBS containing 0.1% Tween 20^TM and 0.1% sodium azide) as the top point for a 3-fold, 7-point dilution series. An equal volume of test sample and biotin-DP178 (125 µl of each) were mixed in a tube and 200 µl was added to a well precoated with mAb 2F5. After an overnight incubation at ambient temperature the wells were washed with diluent and detection of bound biotin-DP178 was accomplished with horseradish peroxidase-coupled streptavidin utilizing 3,3',5,5'-tetramethylbenzidine substrate.

Single-cycle HIV Infectivity Assay-- P4/R5 cells (HeLa cells with a stably integrated LTR-LacZ reporter gene cassette and expressing CD4 and CCR5) maintained in phenol red-free Dulbecco's modified Eagle's medium, 10% fetal bovine serum were seeded in 96-well plates at 2.5 × 10³ cells/well and infected the following day with the HXB2 strain of HIV-1 (Advanced Biotechnology Inc., Bethesda, MD) at a multiplicity of infection of ~0.01 in the presence of titrations of inhibitors or antisera. After incubating an additional 48 h, cells were lysed and -galactosidase was detected using GalScreen^TM chemiluminescent substrate (Tropix, Bedford, MA) according to the manufacturer's instructions. Data were obtained using a Dynex luminometer. IC₅₀ values were calculated by fitting data to the following 3-parameter logistic equation, y = m₁ + ((m₂  m₁)/(1 + (10^(Log(x)  Log(IC₅₀))))), where x = concentration of inhibitor, y = blue/green ratio, m₁ = assay background, m₂ = highest (100%) signal in assay.

Preparation of Peptide-Carrier Conjugates-- Maleimide-activated keyhole limpet hemocyanin (KLH) (Pierce) was used as an immune stimulatory carrier for peptide immunizations. Cysteine-containing peptides were solubilized in 10 mM NaOH, immediately adjusted to neutral pH, and thiol content was determined by the method of Ellman (43). Peptide and carrier were mixed to target a given mole percent of available KLH maleimide groups as determined by the manufacturer. For quantitative assessment of conjugation a maleimide-activated KLH control (no added peptide) was processed identically along with the conjugates. Conjugations were performed in 83 mM sodium phosphate, pH 7.2, 0.9 M NaCl at a final carrier concentration of 1-2 mg/ml for 3-5 h at 25 °C. Residual maleimide groups were quenched using a 10-fold molar excess of L-cysteine hydrochloride for 1 h. Residual free peptide was removed by exhaustive dialysis against conjugation buffer using 300-kDa molecular mass cutoff membranes. Protein content of dialyzed conjugates and control was determined by the bicinchoninic acid assay (Pierce). Quantitation of peptide incorporation into the conjugate was determined by amino acid analysis as previously described (44).

Animal Immunizations-- Female guinea pigs of 4-8 weeks old were obtained from Charles River Laboratories, Wilmington, MA, and maintained in the animal facilities of Merck Research Laboratories in accordance with institutional guidelines. Peptide-carrier conjugates were prepared in phosphate-buffered saline at a concentration of 250 µg/ml. Immediately prior to immunization the saponin-based adjuvant QS21 (45) was added to the conjugate to a final concentration of 100 µg/ml. For immunizations, guinea pigs were injected via the quadriceps with 400 µl of formulated conjugate. The immunization was given three times at 4-week intervals. Blood samples were collected 3 weeks post each injection, and stored at 2-8 °C until assayed.

Enzyme-linked Immunosorbent Assay-- Streptavidin-coated 96-well plates (Pierce) were coated overnight with 0.2 µg of biotinylated peptide in 50 µl of bicarbonate buffer, pH 9.6, per well at 4 °C. Plates were washed with PBS containing 0.05% Tween 20 (PBST) and blocked with 3% skim milk in PBST (milk-PBST). Serial 1:4 dilutions of testing samples were prepared in milk-PBST and 100 µl was added to each well of the antigen-coated plates. The plates were incubated at 25 °C for 1 h, washed three times with PBST, and then incubated with horseradish peroxidase-conjugated goat anti-guinea pig IgG (Zymed Laboratories Inc., San Francisco CA) at 25 °C for 1 h. The plates were washed three times and 100 µl of 1 mg/ml o-phenylenediamine dihydrochloride substrate (Sigma) was added per well. Plates were incubated 30 min and the absorbance at 490 nm was read. Antibody titers were calculated as reciprocals of the highest dilutions, which gave optical density values above two standard deviations of the mean of the secondary antibody conjugate control.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Helical Enhancement of DP178-- The model of Mant et al. (35) was used to design DP178-based peptides of increased -helical content. Fig. 1A presents the Mant optimized model sequence and shows how DP178 and the modified analogs conform to the model. Two strategies were employed for the analog constructions. The first used NH₂-terminal extensions derived from either (a) native GP41 sequence or (b) the GCN4 yeast transcription factor leucine zipper (LZ) dimerization domain (46) as a means of driving downstream helical induction without mutating the primary sequence. The second strategy was point mutation of residues along the native DP178 backbone.

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Amphipathic modeling of DP178 and modified DP178 analog peptides. A, synthetic extended and substituted DP178 analogs were compared with a model peptide described by Mant et al. (35). Italicized lowercase lettering represents the position of the (abcdefg)n heptad repeat. Underlined residues indicate optimization of leucine or -branched aliphatic in positions d and a, respectively. Bars above residues indicate optimized potential i, i + 3, or i, i + 4 salt bridges. Italicized uppercase lettering represents residues derived from GCN4 LZ extensions; in the text the numbering for these residues is italicized. Boxed residues are the 2F5 mAb recognition epitope. Numbering of sequence residues is based on the HIV-1 HXB2 variant. B, prediction of coiled coil propensities using PAIRCOIL program (47) (www version at nightingale.lcs.mit.edu/cgi-bin/score). Data are plotted as peptide residue number versus probability. A value of 0.5 or above is considered positive for formation of an amphipathic coiled coil. , Mant model; , DP178; , C34DP178; , (LZ)10DP178; ×, (LZ)15DP178.

Structural extensions were designed so as to maintain correct phasing of the (abcdefg)_n heptad repeat motif predicted by Weissenhorn et al. (16) and were made only on the NH₂ terminus in the interest of minimizing disruption to any localized secondary structure involving the 2F5 epitope. For C34DP178, 10 GP41 residues immediately NH₂-terminal to the start of DP178 were added to coincide with the start of the outer CHR determined by crystallography (16). The extensions in peptides (LZ)₁₀DP178 and (LZ)₁₅DP178 were based on the dimerization domain (residues 245-281) of GCN4. The Tyr⁶³⁸ residue of DP178 is in position d based on crystallography data. Because Arg²⁸¹ of GCN4 was also a d residue, the GCN4 sequences used to construct the 10- and 15-residue extensions were Val²⁷¹ to Glu²⁸⁰ and His²⁶⁶ to Glu²⁸⁰, respectively. To maintain clarity, residues derived from GCN4 are italicized in the text although their numbering is based on GP41 HXB2 sequence position. The substitution mutant, DP178_mut, was constructed to meet the optimal constraints of the model as discussed below without alteration of the 2F5 epitope or subsequent downstream sequence.

The Mant model predicted -branched aliphatic residues (Ile or Val) in position a and leucine at position d to be highly favorable for helix formation. Charged residues of opposite polarity in the b and e positions further contribute to stabilization through the formation of i, i + 3 and i, i + 4 intrachain electrostatic interactions (35). In native DP178 the majority of a and d positions are not occupied by optimal residues. Although only Glu⁶⁵⁹ constitutes a charged substitution, Ser⁶⁴⁹, Gln⁶⁵², and Asn⁶⁵⁶ are polar, and thus nonconservative with regard to aliphatic residues. Because the model was optimized for amphipathic helices, sequences were input into the program PAIRCOIL (47) to estimate the propensity of individual peptides to form this structure. The output in the form of probability versus peptide residue number is plotted in Fig. 1B. The program also predicts a register position for each residue in the sequence. For all peptides, Tyr⁶³⁸ of DP178 was predicted to be in a d position, which was in agreement with our design strategy (data not shown). Consistent with lack of conformity of DP178 with the model, the probability for coiled coil formation is 0 for this peptide. Extension of the amino terminus by 10 residues optimizes four a or d registers in both C34DP178 and (LZ)₁₀DP178 and introduces an additional potential i, i + 4 salt bridge in each. Correspondingly, the propensity scores rise, although both still fall below the 0.5 cut-off predicted for coiled coil formation. In (LZ)₁₅DP178 the addition of Leu⁶²⁴ optimizes a d register and Glu⁶²⁷ forms another potential i, i + 3 salt bridge with Arg⁶³⁰. The PAIRCOIL program predicts a coiled coil structure for residues Leu⁶²⁴ through Lys⁶⁵⁵ with a probability score of 0.61. The highest propensity score was achieved by DP178_mut. Substitutions Y638L, S649I, Q652L, N656I, and E659L optimized a and d registers for residues 1-26, whereas H643E, I646K, Q650E, Q653K, and L660K maximized electrostatic interactions along the helix backbone. The propensity score for this peptide was 1.0 for residues 1 through 32 and >0.5 for residues 33 through 34. The last 2 residues fell below the cut-off value.

CD Conformation of Peptides-- To determine whether the native solution structure of our DP178 analogs correlated with predictions based on the Mant model, we acquired CD spectra at near neutral pH in aqueous solution at 25 °C. Fig. 2A shows that GP41 peptides DP178 and C34DP178 are characterized by a single broad negative ellipticity centered at ~202 nm, indicative of an unordered structure. In contrast, the LZ peptides and DP178_mut exhibit a double minima at ~208 and 222 nm along with a strong positive ellipticity at ~195 nm, features that are typical of -helices. The absolute magnitude of these peaks increases in the order (LZ)₁₀DP178 < (LZ)₁₅DP178 < DP178_mut. Table I gives the percent of -helix as estimated from the []₂₂₂ value. The values obtained for DP178 and C34DP178 are consistent with previous reports of CHR peptides in aqueous solution (23). The addition of TFE to a final concentration of 30% (Fig. 2B) markedly enhanced the helical content of the less structured forms, with an ~3-fold increase observed for DP178 and C34DP178. The effect, as expected, was less dramatic for those peptides already containing a significant amount of helical character and may represent coiled-coil association rather than increased helicity in the monomeric chain. Fig. 2C shows the spectra of those constructs used for conjugation. Whereas the helicity enhancement trend is the same as that seen in panel A, the average helical content appears higher for all constructs except DP178, which appeared to decrease by about 50%. Most strikingly, the helicity of (LZ)₁₅DP178 was increased from 54 to 103% by the addition of an NH₂-terminal cysteine and terminal capping.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   CD spectra of DP178 and analogs. Spectra were measured at 25 °C at 0.1-nm intervals and plotted at 1.0-nm intervals as described under "Experimental Procedures." A, peptides under native conditions. Samples were solubilized at 1 mg/ml in 1 mM NaOH and diluted to 50 µM in 10 mM sodium phosphate, pH 7.3. , DP178; +, C34DP178; , (LZ)10DP178; , (LZ)15DP178; , DP178mut. B, effect of TFE was measured by adjusting samples used for A to 30% TFE. Final peptide concentration was 35 µM. Symbols as in A. C, cysteine-containing peptides under native conditions. Samples were solubilized at 1 mg/ml in 20 mM Tris-HCl, 10 mM EDTA, pH 9.2 and diluted to 50 µM in 100 mM sodium phosphate, 10 mM EDTA, pH 7.3. Symbols as in A.

NMR of DP178-- NMR spectroscopy was used to investigate the structure of native DP178 in more detail. In particular, it was desired to determine whether the low percentage of helicity observable by CD was confined to a specific region of the molecule or whether it represented an average across the entire peptide backbone. DP178 aggregates at concentrations above 10 µM exhibiting a complex monomer-tetramer equilibrium (23). Initial attempts to obtain NMR spectra of DP178 in PBS were unsuccessful because of extremely broad resonances that precluded the use of two-dimensional NMR for structural analysis. Earlier studies in our laboratory on related hydrophobic peptides had determined that organic:aqueous mixtures, in particular, acetonitrile-d₃:H₂O (1:1, v/v) are often effective in reducing or completely eliminating intermolecular aggregation. DP178 was completely soluble in a mixture of acetonitrile-d₃:H₂O (1:1, v/v) at concentrations of ~1 mM. The proton spectrum displayed narrow resonance line widths consistent with a monomeric species. There were no appreciable changes in line widths over the concentration range 0.060-1.1 mM suggesting that the peptide remains monomeric under these conditions. Fig. 3A shows CD spectra of DP178 in PBS versus acetonitrile-d₃:H₂O. The average percent helicity as estimated by []_{222 nm} increased from 15 to 25% with the addition of organic solvent. Qualitative secondary structure characterization was accomplished primarily by two-dimensional NMR. Sequential resonance assignments were obtained by standard methodologies using TOCSY and NOESY spectra. Sequential assignments were made for ~85% of the backbone and side chain resonances. The relative intensities of sequential and nonsequential NOE cross-peaks as well as the magnitude of backbone vicinal coupling constants (³J_NH,H) can be used to differentiate between helical and extended secondary structures in peptides. The observation of intense sequential NH to NH NOEs and small ³J_NH,H coupling constants (<6 Hz) are indicative of a helical conformation. An expansion of the NOESY spectrum of DP178 recorded in acetonitrile-d₃:H₂O is shown in Fig. 3B illustrating the pathway of sequential NH to NH NOEs. There is a 10-residue contiguous stretch of moderately intense NH to NH NOEs from Glu⁶⁶² to Asn⁶⁷¹. The identification of a helical segment is indicated by the presence of intense sequential NH-NH NOE cross-peaks, the presence of i to i + 3 long range NOEs, and nonsequential i to i + 2 NH-NH NOEs. Additionally the ³J_NH-H coupling constants for residues Glu⁶⁶² through Leu⁶⁶⁹ are all <6.0 Hz consistent with a helical secondary structure. Fig. 4 summarizes the NMR parameters used to define the helical region of DP178. Following dissolution in a acetonitrile-d₃:D₂O mixture, seven amide protons (Leu⁶⁶², Trp⁶⁶⁶-Asn⁶⁷¹) are slow to exchange with solvent deuterons; many are visible hours later, consistent with NH_i to O_i4 hydrogen bonding. These data indicate that a helical segment is present in the COOH-terminal region of DP178 comprising residues Glu⁶⁶² to Asn⁶⁷¹, which includes the 2F5 epitope. There is a single CH_i to NH_i+4 NOE observed between Ala⁶⁶⁷ and Asn⁶⁷¹ indicating that the Glu⁶⁶² to Asn⁶⁷¹ segment adopts an -helix as opposed to a 3₁₀-helix. This finding is also consistent with the CD study of the peptide that gave an estimated fractional helicity of ~25%. Beyond Leu⁶⁶⁹, the absence of long range NOEs and ³J_NH-H coupling constants exceeding 6 Hz indicate disorder or fraying of the helix terminus. The remainder of the peptide backbone, residues 638-661 is largely unstructured. The NOESY data for residues 638-661 are largely characterized by moderately intense CH_i to NH_i+1 NOEs and a paucity of long range interactions that are consistent with an extended structure. Similar results were obtained on a truncated DP178 analog comprising residues 559-671 demonstrating partial helical character (data not shown).

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Assessment of helicity in native DP178. CD and NMR spectroscopy were used to determine which residues of native DP178 were helical in solution. A, CD spectra were obtained at 0.1-nm intervals and plotted at 1.0-nm intervals. Spectra were recorded in PBS () or 50:50 (+) acetonitrile:water. B, an expansion of the 600 MHz two-dimensional NOESY spectrum of a sample of DP178 in acetonitrile-d3:H2O (1:1, v/v) recorded with a 100-ms mixing time at 25 °C. The amide proton region is shown illustrating sequential NH to NH NOEs. A contiguous stretch of intense NH-NH NOEs indicates the presence of a helix between residues 662 and 671 in the COOH-terminal region of DP178.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Summary of NMR parameters used to define the secondary structure of the COOH-terminal region of DP178 comprising residues 662-671. Sequential and nonsequential NOEs observed between two residues are indicated by bars. The thickness of the bar indicates the relative intensity of the observed NOE. Gaps in the dN (i, i + 1) NOE connectivities reflect ambiguities because of resonance degeneracy. Upward and downward pointed arrows indicate 3JNH-H vicinal coupling constants of >7.0 or <6.0 Hz, respectively. Asterisks indicate slow rates of amide proton exchange with solvent deuterons. The presence of intense sequential NH to NH NOEs, small values of 3JNH-H, and medium/long range NOEs indicate the presence of a helix between residues Glu662-Asn671.

Monoclonal Antibody Binding and Infectivity Inhibition by Peptides-- To assess and rank the ability of a peptide to bind to the broadly neutralizing mAb 2F5, a competitive solution binding assay was designed in which the peptide of interest competes with a fixed and limiting amount of a labeled reference peptide for a limiting amount of mAb 2F5 bound to plastic. The 50% maximum bound control consisted of biotin-DP178 mixed with diluent in place of inhibitor and run in triplicate. A negative control consisting of diluent only was also run in triplicate to monitor assay background. To normalize values obtained for samples on different test days so that they could be compared a 7-point, 3-fold dilution of unlabeled DP178 was run in each test. We determined the dilution of sample that provides for a 50% reduction in the maximum bound value by linear interpolation between two of the points in the sample dilution series. From this the number of moles of sample required to reduce the maximum bound response by 50% of 1 mol of biotin-DP178 was calculated. This value was then normalized to the standard unlabeled DP178 and reported as the reactivity value (IC₅₀).

The reactivity of DP178 and its analogs in the competitive 2F5 binding and viral infectivity inhibition assays is summarized in Table II. Increasing the helicity of DP178 by extension apparently improved the binding to 2F5 by a modest 2-4-fold for both native and termini-capped cysteine-containing peptides. A quantitatively similar effect was observed regarding the ability of these constructs to inhibit infectivity. In both assays, cysteine-containing peptides gave overall lower IC₅₀ values, possibly as a result of increased local concentration resulting from dimer formation. In neither assay was a clear correlation between absolute percentage of -helix and reactivity observed. In fact, the peptide showing the highest helicity, DP178_mut, was ~10- and 90-fold less effective than DP178 in the 2F5 and infectivity assays, respectively. This was most likely attributable to mutation of important functional residues rather than conformational change. DP178_mut had 11 of its 36 residues (30%) substituted to optimize correlation with the model. Some of these substitutions, notably Q650E, Q652L, Q653K, N656I, and L660K, were of residues that are greater than 97% conserved in over 800 isolates (48). Because most of these are implicated in stabilizing the interaction between NHR and CHR helices in the six-stranded bundle it is not surprising that this peptide had low activity in infectivity inhibition. However, given the fact that the NMR data supported a helical structure for the region encompassing the ELDKWA epitope in native DP178 and because no residues of this epitope or downstream of it were substituted in DP178_mut it was strongly suggestive that residues outside the core epitope may be important for 2F5 binding.

To test this possibility we performed an alanine scan in which the 10 residues immediately NH₂- and COOH-terminal to the 2F5 epitope were systematically substituted as was the epitope itself. For the NH₂-terminal substitutions we used native DP178. To accomplish the COOH-terminal scan, DP178 was extended by 10 residues to produce GP41_638-683 and this construct was used for the epitope substitutions as well. The peptide design and 2F5 inhibition results are presented in Fig. 5. Consistent with previous reports (11, 49) core epitope residues Asp⁶⁶⁴, Lys⁶⁶⁵, and Trp⁶⁶⁶ are invariable and substitution reduced activity by >20-fold. Two additional substitutions, E659A and L661A, were ~7-fold less effective than native DP178. This observation was reproducible in replicate assays (data not shown). Interestingly, no substitution in the 10 residues COOH-terminal to the epitope produced any change in activity although the extended peptide itself was somewhat better of a competitor than native DP178. We performed CD analysis on E659A and L661A peptides to determine whether their conformation was altered relative to DP178. The estimated percent helicities were 12 and 13%, respectively, as compared with 15% for DP178. Likewise, constructs containing D664A and K665A gave 25 and 23% helicity, respectively, as compared with unsubstituted GP41_638-683, which gave a helical content of 30% (data not shown).

View larger version (65K): [in this window] [in a new window]   Fig. 5.   Alanine scan of DP178. Synthetic peptides were prepared to accomplish systematic alanine substitution of the 10 residues NH2- and COOH-terminal to, and inclusive of, the ELDKWA 2F5 recognition epitope. Peptides were solubilized at a nominal concentration of 1 mg/ml (w/v) in 20 mM sodium phosphate, pH 10.5, and immediately adjusted to pH 8.2 by titration with 1 N HCl. Each peptide was tested in the 2F5 competition assay to generate a reactivity (IC50) value. Underlined residues represent the XnA substitution in each peptide. Residue numbering as in Fig. 1A.

Conjugation and Immunogenicity-- Immunization of guinea pigs with peptide-KLH conjugates resulted in high peptide-specific enzyme-linked immunosorbent assay titers as shown in Table III for the post-dose three sera. The amount of peptide incorporated into the conjugate on a molar basis was determined using a multiple linear least squares regression analysis in which the amino acid composition of conjugate and a KLH-only control were compared (44). The variability between conjugates was less than 32%. This degree of variation did not appear to influence the level of titer achieved, and no correlation between titer and cysteine location was observed. None of the sera were able to inhibit viral infectivity in the in vitro assay at a 10-fold dilution, which was the highest testable level because serum components interfered at lower dilutions. The addition of a -GG- spacer to C34DP178 was done with the thought that it might enable the covalently coupled peptide to more readily preserve its structural integrity. This strategy did not produce an enhanced immune response nor did it result in attainment of neutralization. The geometric mean titers achieved with the (LZ)₁₅DP178 construct are quite high for a peptide-carrier antigen. These titers were peptide specific as evidenced by cross-reactivity of serum with a related peptide such as DP178 and lack of response with nonhomologous nonDP178 peptide controls (data not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The inability to develop an HIV-1 envelope immunogen capable of eliciting a broad neutralizing response has severely impeded vaccine development for this disease. Whereas recent strategies designed to elicit protective T-cell-mediated responses have been encouraging (50), some researchers believe that a humoral component may also be required to produce high efficacy in an HIV vaccine. Many of the characterized monoclonal antibodies that are broadly neutralizing against both wild type and laboratory-adapted viruses recognize specific envelope components, although the defined structure of these epitopes on the viral surface is largely unknown. Furthermore, for several of these, binding affinity is either enhanced or abrogated following the structural change that accompanies receptor binding. In the current study we examined the utility of increasing the helical content of DP178 as a means of inducing a 2F5-like response.

A number of studies support the theory that for mAb 2F5 the neutralization target sequence is presented in the prefusogenic conformation of GP41 but is perturbed or rendered inaccessible following transition to the prehairpin intermediate (31, 51). Although the fusion-active six-stranded helical GP41 core has been characterized in rigorous detail, the structure of the CHR region that encompasses 2F5 in the prefusogenic state is not known with certainty. Some recent reports have suggested possible structures for this domain prior to formation of the six-stranded bundle. The tryptophan-rich region comprising residues 665-683 contains the KWAS portion of the epitope as its amino terminus. Peptides overlapping this region show a strong propensity to interact with membranes and are critical for effective envelope-mediated fusion (52, 53). In a recent NMR study of the peptide in association with dodecylphosphocholine micelles, Schibli et al. (22) showed that it adopts a predominantly -helical structure. According to their model, the highly conserved Trp⁶⁶⁶, Trp⁶⁷², and Trp⁶⁷⁸ residues interact at the water-lipid membrane interface so that this region of GP41 would lie close to the bilayer surface with the 2F5 epitope bending outward in a loop to join the CHR region NH₂-terminal to it. DP178 also overlaps a portion of the tryptophan-rich region (residues 665-673) and has been recently implicated in membrane interactions (54). Although we and others have shown that the isolated peptide is largely disordered in solution, it is entirely plausible that it exists as a more highly helical structure in the membrane-bound state of native GP41.

The experimental observations presented in the current study appear to support this possibility. First, native DP178 has amphipathic character, although it fails to meet many of the optimal criteria found in the Mant model. The addition of low amounts of TFE to DP178 significantly increased its -helicity indicating that the propensity to exist as a helix is present. The ability of TFE to induce helical structures is partly a function of its effect on lowering the dielectric content of the aqueous medium, and this is believed to mimic the microenvironment of membrane surfaces (55). This observation agrees well with the proposed membrane-interacting properties of the GP41 CHR region. Second, NMR results unequivocally demonstrate that the COOH-terminal region comprising residues Glu⁶⁶² to Asn⁶⁷¹ adopts a helical structure in an acetonitrile-d₃:H₂O mixture. This finding is consistent with the 25% fractional helicity observed by CD in the same medium. An intriguing result was the behavior of DP178 in aqueous versus organic:water mixtures. The NMR spectra of the peptide in PBS displayed extremely broad resonances characteristic of an aggregated species. By CD the helical content was measured to be 15%. However, in 50:50 acetonitrile:H₂O, the NMR spectrum became much more highly resolved with narrow resonances, whereas the helicity by CD increased to 25%. Once again, it is likely that the organic:water mixture may act as a membrane mimetic, inducing increased structure in the sequence NH₂-terminal to the tryptophan-rich region.

Modification of DP178 by extension or substitution mutations resulted in conformational changes as measured by CD that were consistent with predictions based on both the Mant model and the PAIRCOIL program. These were directly attributable to introduction of optimized residues in critical register positions. The peptide that displayed the highest degree of helicity, DP178_mut, was specifically designed to contain optimized residues in all positions NH₂-terminal of the ELDKWA epitope and to maximize potential intrachain salt bridges. Of interest was the significant increase in helicity for (LZ)₁₀DP178 relative to C34DP178. Although both were extended by 10 residues and both extensions introduced one additional potential electrostatic interaction, the former peptide was 88% higher in helical content. This is likely explained by the fact that (LZ)₁₀DP178 contained three optimized a or d positions compared with one in C34DP178. Furthermore, the helical content of (LZ)₁₀DP178 was greater than that expected if only the 10-residue addition was a helix, implying that the LZ extension was driving downstream helicity in the remainder of the molecule. All peptides, with the exception of DP178, showed increased structure when prepared as termini-capped cysteine derivatives. The use of capping for induction and stabilization of -helices is a well known phenomenon, and although these peptides did not contain specific sequences designed for helix induction aside from the fusion extensions, terminal capping can markedly enhance structure in peptides with the propensity to form helices. The moderate increases in helicity observed for C34DP178 and (LZ)₁₀DP178 may reflect the fact that the uncapped native peptides would show a greater tendency to fray at the ends. The difference observed for (LZ)₁₅DP178 was more dramatic. Whereas capping may contribute some of the enhancement, it is possible that the extended GCN4 sequence might act to induce dimerization of individual helices and thus make disulfide bond formation more thermodynamically favorable. Houston et al. (56) had shown using model coiled coil peptides stabilized by lactam bridge formation that the helicity could be increased from 59 to 94% by introduction of an N-terminal disulfide bond.

The solution-based competitive 2F5 binding assay proved useful for ranking the ability of the peptide to interact specifically with the monoclonal as well as for identifying residues that have a potentially important role in mediating peptide-antibody affinity. This assay was superior to plate-based assays in which antigen is usually bound to plastic and binding is monitored either directly using a labeled antibody or indirectly by a secondary reporter antibody. The problem with this type of format is that structural antigens can be altered during the adsorption to plastic. In our assay, both reference peptide and competitor exist in solution and so any conformational dependent affinity for 2F5 would be maintained. Extended analogs showed a 2-4-fold increase in binding relative to native DP178, and whereas it might be argued that this improvement is modest at best, the lack of a dramatic enhancement in activity with increasing structure supports the NMR observation that this region of the molecule is already in a helical state even in the mostly disordered native DP178. Interestingly, the COOH-terminal extended peptide used for alanine scans, GP41_638-683, was also more reactive, and CD confirmed that its helical content was approximately twice that of DP178. The NMR data also supports the postulate that native DP178 in solution is flexible enough to allow an induced fit upon binding of antibody. This does not, however, account for the greatly reduced affinity displayed by DP178_mut. Instead, it suggests that residues outside of the core epitope are important for interaction with antibody. Efforts at defining the critical residues by alanine scanning were unsuccessful in that no single amino acid replacement (except those in the core epitope) produced as dramatic a decrease in reactivity as seen for DP178_mut. Of the two substitutions, E659A and L661A, which did affect binding, only E659L was present in DP178_mut. The likely explanation is that multiple substitutions along the face of DP178_mut acted synergistically to lower its 2F5 binding ability. Also, several substitutions in DP178_mut, notably E659L, L660K, and N656I, were of a nonconservative nature and were all proximal (within six residues) to ELDKWA.

Native and extended DP178 constructs were fully active in the single cycle infectivity assay with the more structured variants showing a slight increase in potency. Although the helical content of C34DP178 was unchanged relative to DP178, its increased activity in the assay can be explained by the inclusion of Trp⁶²⁸, Trp⁶³¹, and Ile⁶³⁵ as part of the NH₂-terminal extension. These residues, which lie along the same face of the fusion-active CHR, form critical interactions with a hydrophobic pocket lining the NHR trimer as described by Chan et al. (17). Interestingly, these residues are not present in the LZ-substituted peptides, but the corresponding replacements, W628V, W631L, and I635V are relatively conservative substitutions, with I635V also found in native SIV (57). This result was somewhat surprising because it was previously reported that a single point W631L mutation in C34 reduced activity by ~6-fold in cell-fusion based assays (20). It might be postulated that the higher helical content of the LZ-substituted peptides resulted in a stronger interaction with the NHR region because they were essentially pre-formed and did not have to undergo a binding-induced attainment of helical structure. As was observed in the 2F5 competition assay, DP178_mut was 90-fold less active than the native sequence and almost 500-fold less active than (LZ)₁₅DP178 in its ability to inhibit viral entry. The substitutions to enhance model conformity effected nonconservative mutations at a or d positions Q652L, N656I, and E659L and at e positions I646K, Q653K, and L660K. Crystallography studies of the protease-resistant core have shown that these a and d positions make critical contacts with e and g residues on the NHR trimer (17). The function of the CHR e residues is not known but they are also highly conserved between HIV-1 and SIV. Because both DP178 and DP178_mut lacked the upstream Trp⁶²⁸, Trp⁶³¹, and Ile⁶³⁵ residues, they must either (a) interact with the NHR trimer at residues COOH-terminal to this pocket-binding sequence, or (b) inhibit fusion through some stage distinct from the prehairpin intermediate. Peptide mixing studies support the ability of DP178 to interact with N-peptides to form thermostable core-like structures. Nevertheless, although dominant-negative inhibition of six-stranded core formation has been the mechanism most often postulated for DP178, several recent studies offer evidence for DP178 acting by alternate means. Interestingly for our findings, Kliger et al. (54, 58) proposed an inhibitory role for DP178 in the context of its membrane-bound form, a structure that has been shown to be highly -helical (22). Finally, DP178_mut contains a Q652L substitution, which had been shown to increase thermostability of an N34(L6)C28 core model peptide and increase its inhibitory activity by 10-fold (59). The fact that DP178_mut showed loss rather than enhancement of activity again supports the notion that multiple residues are important for the activity of DP178 as an inhibitor.

The ability of a peptide to serve as an effective inhibitor of fusion does not necessarily indicate its usefulness as an immunogen to raise a broadly neutralizing response, as is evident from previous DP178-based studies. Having shown that the ELDKWA epitope was partially helical in the native peptide, we undertook immunization studies using the more structured analogs. Our results showed that conjugated DP178 and structured analogs were uniformly incapable of eliciting a neutralizing response although all conjugates raised high peptide-specific titers, and some also were reactive against heterologous peptide constructs containing partial sequence homology to the immunizing construct. One possible explanation is that the helical nature of the peptides was not maintained following conjugation. Although the maleimide-thiol coupling strategy employed mild reaction conditions, helical disruption as a result of peptide-carrier interaction cannot be discounted. Furthermore, these molecules were not locked in a given conformation by any type of chemical constraint. Indeed, the NMR data suggests a high degree of freedom for interconversion. It may be that analogs stabilized by the use of i, i + 7 lactam bridges (60) or some other means would be more effective immunogens.

It would seem logical that an anti-HIV immune response recognizing the transition state of GP41 might be broadly neutralizing given the highly conserved nature of this portion of the protein and especially because evidence exists that the prehairpin intermediate can have a lifetime of several minutes (2). The identification of several human mAbs that neutralize primary isolates suggests that it should be possible to design an immunogen that would induce similar protection. Our work and that of others has been based on this premise, yet success has been elusive. Why do immunogens containing the 2F5 recognition sequence fail to induce a 2F5-like response? One possibility has already been discussed in the context of structural considerations. MAb 2F5 is broadly neutralizing only prior to CD4 binding, arguing that it either recognizes the transitional intermediate or the prefusion structure of DP178. Lacking clearly defined crystallographic data on these structures, any conformational design of peptides thought to mimic native structures is empirical. We examined helical structures in the context of our NMR findings and because earlier reports offered indirect evidence that such conformations of DP178 might exist on the prefusogenic virion. A recent description of the ELDKWAS peptide co-crystallized with the Fab' 2F5 fragment showed the epitope to adopt a type I -turn (34). However, the isolated epitope sequence is likely not physiologically relevant, and in our assay ELDKWA poorly competed with DP178 for binding to 2F5 (data not shown).

An alternative explanation focuses on the fact that the isolation of a 2F5-like antibody is a rare event and the observation that large amounts of antibody are required for effective neutralization. It takes between 10 and 100 µg/ml 2F5 depending on the primary isolate to neutralize HIV (11, 31), which suggests that the monoclonal has a low affinity for its epitope and thus the need for high concentrations to be effective. We have confirmed this hypothesis by a liquid phase affinity determination of 2F5 for DP178 using our competitive binding assay that resulted in a k_A = 2.0 × 10⁸ (data not shown). If 2F5 has a low variable affinity for most of the functional variants of HIV 1 this could account for its breadth of neutralization and suggests that it functions only at the high concentrations made possible by recombinant technology. These characteristics are not found in natural immune responses that generally rely on high affinity at low concentration to effect neutralization. This argues that 2F5 is most likely an affinity maturation intermediate rather than a mature response to a unique epitope and may explain why no equivalent to it has ever been isolated. If this hypothesis is correct, it is understandable why no 2F5-like response has been elicited by immunogens containing its recognition epitope and suggests that its broad neutralizing capability is an artifact of biotechnology.

In conclusion, we have shown that peptide analogs of DP178 with increased -helical character bind to mAb 2F5 as well as or better than native DP178 and also act as effective inhibitors of viral entry. Conjugate vaccines prepared from the peptides elicit high specific titers in guinea pigs, but the sera fail to neutralize in an in vitro infectivity model. Whereas the highly conserved NHR and CHR regions of GP41 remain viable candidates for elicitation of a broadly neutralizing protective vaccine, further work will be required to determine whether the ELDKWA epitope is the key to success.

	ACKNOWLEDGEMENTS

We thank Dr. Ned Landau for the P4/R5 cell line, Dr. Patrick Kanda for amino acid analysis, and Dr. Antonello Pessi for critical manuscript review and commentary.

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ To whom correspondence should be addressed: Dept. of Virus and Cell Biology, Merck Research Laboratories, West Point, PA 19486. Tel.: 215-652-5617; Fax: 215-652-2142; E-mail: joseph_joyce@merck.com.

Published, JBC Papers in Press, September 16, 2002, DOI 10.1074/jbc.M205862200

	ABBREVIATIONS

The abbreviations used are: HIV-1, human immunodeficiency virus type 1; GP, glycoprotein; mAb, monoclonal antibody; N(C)HR, amino (carboxyl)-terminal heptad repeat; SIV, simian immunodeficiency virus; CD, circular dichroism; TFE, 2,2,2-trifluoroethanol; KLH, keyhole limpet hemocyanin; LZ, leucine zipper; PBS, phosphate-buffered saline; italic numbers indicate residues derived from GCN4 LZ extensions..

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES