Mode of action of an antiviral peptide from HIV-1. Inhibition at a post-lipid mixing stage.

DP178, a synthetic peptide corresponding to a segment of the transmembrane envelope glycoprotein (gp41) of human immunodeficiency virus, type 1 (HIV-1), is a potent inhibitor of viral infection and virus-mediated cell-cell fusion. Nevertheless, DP178 does not contain gp41 coiled-coil cavity binding residues postulated to be essential for inhibiting HIV-1 entry. We find that DP178 inhibits phospholipid redistribution mediated by the HIV-1 envelope glycoprotein at a concentration 8 times greater than that of solute redistribution (the IC(50) values are 43 and 335 nm, respectively). In contrast, C34, a synthetic peptide which overlaps with DP178 but contains the cavity binding residues, did not show this phenomenon (11 and 25 nm, respectively). The ability of DP178 to inhibit membrane fusion at a post-lipid mixing stage correlates with its ability to bind and oligomerize on the surface of membranes. Furthermore, our results are consistent with a model in which DP178 inhibits the formation of gp41 viral hairpin structure at low affinity, whereas C34 inhibits its formation at high affinity: the failure to form the viral hairpin prevents both lipid and solute from redistributing between cells. However, our data also suggest an additional membrane-bound inhibitory site for DP178 in the ectodomain of gp41 within a region immediately adjacent to the membrane-spanning domain. By binding to this higher affinity site, DP178 inhibits the recruitment of several gp41-membrane complexes, thus inhibiting fusion pore formation.

that mediate the fusion between the viral and the cellular membranes or between infected and healthy cells (12,13). Gallaher and co-workers (14,15) postulated a model of gp41, identifying a fusion peptide followed by a leucine/isoleucine zipper-like sequence (N-helix) and an amphipathic helical segment (C-helix) in the viral glycoprotein. The indispensability of the fusion peptide for viral infection was confirmed by sitedirected mutagenesis (16,17). Furthermore, gp41 was found to contain a protease-resistant core consisting of the postulated N-and C-helices (18). Specifically, peptides corresponding to these sequences co-crystallized as a six-helix bundle in which the N-and C-helices are arranged in a three-hairpin structure (19 -21). Three N peptides form a coiled-coil, and the C peptides are packed in an antiparallel manner into highly conserved, hydrophobic grooves on the surface of the coiled-coil. Recently, the solution and crystal structures of the ectodomain of the Simian immunodeficiency virus gp41, consisting of those two helices as well as the loop connecting them, confirmed the interplay of the N-and C-helices (22,23). Remarkably, the coiled coil is a common motif found in many diverse viral membrane fusion proteins (24), as well as in proteins involved in vesicular transport (25)(26)(27)(28)(29)(30)(31).
A synthetic peptide overlapping the C-terminal amphipathic helical segment of gp41 and its tryptophan-rich sequence (32) (DP178, Fig. 1) was reported to inhibit virus infection at extremely low concentration (33). Remarkably, DP178 blocks cell fusion and viral entry at a concentration of less than 2 ng/ml in vitro and was reported to be a promising drug for treating HIV-1-infected humans (34 -37). Since DP178 is a potent inhibitor of HIV-1-induced membrane fusion, elucidating its mode of action is of major importance. In essence, this means finding the stage at which DP178 inhibits the formation of the fusion-active conformation of gp41. It was suggested that the antiviral mode of action of DP178 involves interaction with the gp41 leucine/isoleucine zipper motif (38 -41). However, DP178 (residues 638 -673) does not contain the residues (Trp-628, Trp-631, and Ile-635) that were shown to be crucial for binding the prominent cavity in the coiled coil of gp41 (59). Weiss and colleagues (42) demonstrated that DP178 binds gp41 and inhibits envelope-mediated membrane fusion only after gp120 interacts with cellular receptors. It is currently accepted that DP178 binds to the gp41 leucine/isoleucine zipper sequence before the hairpin structure is formed (43)(44)(45), thus preventing the HIV envelope glycoprotein from adopting a fusogenic conformation. This model is supported by the results of Blumenthal and co-workers (6,46), suggesting that the pre-hairpin intermediate stage appears to be induced rapidly upon interaction of gp120 with CD4 and the co-receptor, but is then relatively stable for several minutes, allowing DP178 to interact with the exposed leucine/isoleucine zipper sequence.
However, DP178 was reported to inhibit redistribution of lipid and aqueous dyes as a result of HIV-1 envelope glycoprotein-mediated fusion with significantly different values of IC 50 (46). Binding of DP178 to the leucine/isoleucine zipper sequence in the pre-hairpin intermediate should inhibit both lipid and aqueous redistribution, suggesting the existence of two target sites for inhibition by DP178. In an attempt to identify an additional target site for inhibition by DP178, we investigated potential interactions of DP178 with the membrane, and with the leucine/isoleucine zipper sequence within the membrane. We find that DP178 binds to the membrane and oligomerizes within, but does not interact with the leucine/ isoleucine zipper sequence within the membrane. We suggest that by binding to its corresponding segment in gp41, DP178 inhibits fusion pore formation.

EXPERIMENTAL PROCEDURES
Peptides Preparation and Fluorescent Labeling-DP178, N36, and C34 were synthesized by using the Boc chemistry, as described previously (47,48). Concentrations were measured by tryptophan and tyrosine absorbance (at 280 nm) in 8 M urea (49). Labeling of the N terminus of synthetic peptides was achieved as described previously (50,51). Briefly, resin-bound peptides, with their amino acid side chains fully protected, were treated with trifluoroacetic acid, to remove the BOC protecting group from their N-terminal amino groups, while keeping all the other reactive amine groups of the attached peptides still protected. The resin-bound peptides were then reacted with the desired fluorescent probe, cleaved from the resins by hydrogen fluoride, and finally precipitated using ether. This procedure yielded peptides selectively labeled with fluorescent probes at their N-terminal amino acids. The synthetic peptides were purified by reverse-phase high performance liquid chromatography on an analytical C 18 Vydac column 4.6 ϫ 250 mm (pore size of 300 Å). The column was eluted in 80 min, at a flow rate of 0.6 ml/min, using a linear gradient of 25-80% acetonitrile in water, in the presence of 0.05% trifluoroacetic acid (v/v).
Dye Transfer Fusion Assay-Peptide inhibition of cell-cell fusion was assayed by monitoring the redistribution of fluorescent probes, both water soluble and lipophilic, between target and effector cells upon their co-incubation with each other (46). The HIV-1 gp120-41 expressing TF228 cells (52) were labeled with either calcein or a green fluorescent fatty acid. Cells incubated with 1 M calcein for 60 min at 37°C were then washed and resuspended at 105 cells/Ml in RPMI. Cells labeled with the fatty acid (4,4-difluoro-5-methyl-4-bora-3a,4a-diaza-sindacene-3-dodecanoic acid (C 1 -BODIPY-C 12 )) were seeded at 105 cells/ml in RPMI medium containing 10 mg/ml of the C 1 -BODIPY-C 12 and grown for 3 days prior to experiment (Molecular Probes, Eugene, OR) (53). The fatty acid was eventually metabolized into phospholipid, primarily phosphatidylcholine (data not shown). The cells where then washed and resuspended at 105 cells/ml. Lipid (BODIPY) dye and aqueous (calcein) dye transfer were measured in separate experiments. The target cells, CD4ϩ CXCR4ϩ 3T3 mouse fibroblasts, were plated at 10 5 cells/ml the night before the experiment. The next day, they were labeled with 20 M CMTMR for 1 h at 37°C, washed several times and combined with the effector cells (1:3 target-effector cell ratio). Different concentrations of peptides, dissolved in phosphate-buffered saline or Tris buffer, were then added. The cells were co-cultured for 2 h at 37°C in 12-well plates (Costar, Cambridge, MA). Phase and fluorescent images were collected using an Olympus IX70 coupled to a CCD camera (Princeton instruments, Trenton, NJ) with a ϫ20 objective lens. An 82,000 optical filter cube (Chroma technology corp., Brattleboro, VT) was used for the excitation of calcein (494/517), BODIPY (500/510), and CMTMR (541/565). Three images per well were collected and then analyzed using metamorph software (Universal imaging, West Chester, PA) for dye transfer from the donor to the acceptor cell. The scoring of fusion events was conducted as previously (46).
Preparation of Lipid Vesicles-Dry phospholipids were hydrated in phosphate-buffered saline and dispersed by vortexing to produce large multilamellar vesicles. Small unilamellar vesicles (SUV) were then prepared by sonication.
DNS-tryptophan FRET-The binding of a peptide to the membrane was detected by using FRET from the tryptophan residues of the peptide to DNS chromophores incorporated into lipid vesicles, as described previously (54). Excitation was set at the tryptophan maximum absorbance peak (280 nm), and binding was detected by measuring the increase in the sensitized DNS fluorescence (518 nm), thus diminishing light scattering.
Membrane Partition of the Peptides-NBD fluorescence increases directly with the increase in the environmental hydrophobicity. The degree of peptide association with lipid membranes was measured by titration of the NBD-labeled peptide with lipid vesicles in phosphatebuffered saline. The fluorescence intensity was measured with excitation set at 467 nm, and emission set at 530 nm. The fluorescence values were later corrected by taking into account the dilution factor corresponding to the addition of microliter amounts of liposomes and by subtracting the corresponding blank (buffer with the same concentration of vesicles). The binding isotherms were analyzed as partition equilibria (55,56), as described in detail previously (50). The partition coefficient (K) is determined from the slope of the curve. Free energy (⌬G) was calculated according to ⌬G ϭ ϪRT ln([H 2 O]⅐K). The universal gas constant was taken as R ϭ 1.987 cal/(K mol), and H 2 O concentration is 55.6 M.
NBD-Rho FRET-The interaction of peptides within the membrane was evaluated using NBD-labeled peptides serving as donors and rhodamine-labeled peptides as energy acceptors (57,58). In a typical experiment, a donor peptide was added to a dispersion of SUV in phosphate-buffered saline, followed by adding the acceptor peptide in several sequential doses. Fluorescence spectra were obtained before and after adding the acceptor. The efficiency of energy transfer (E) was determined by measuring the decrease in the quantum yield of the donor as a result of the presence of the acceptor. E was determined experimentally from the ratio of the fluorescence intensities of the donor in the presence (I da ) and absence (I d ) of the acceptor, at the donor's maximum emission wavelength (530 nm and 524 nm for NBD-N36 and NBD-DP178, respectively), The correction for the contribution of acceptor emission as a result of direct excitation was made by subtracting the signal produced by the acceptor-labeled analogue alone. The contributions of the buffer and vesicles were also subtracted from all measurements.

Two Overlapping Anti-HIV-1 Peptides Inhibit Cell-Cell Fusion via a Different
Mechanism-Previously, we reported that DP178 inhibits redistribution of lipid and aqueous dyes as a result of HIV-1 envelope glycoprotein-mediated fusion with significantly different values of IC 50 (46). To investigate DP178 mode of action, we compared its inhibition ability with that of C34, another antiviral peptide which overlaps with DP178 in 24 out of its 34 residues (59) (Fig. 1). We performed a series of new experiments to investigate inhibition of HIV-1 envelope glycoprotein-mediated fusion by DP178 and C34. The experi- DP178, Mechanism of Inhibition ments with the two inhibitors were done on the same cells labeled the same way. In our previous study (46) we used lipophilic carbocyanine membrane probes which are incorporated into the plasma membrane from the outside and have the potential of nonspecifically exchanging between cells. To avoid potential nonspecific dye transfer we used in this study phospholipid analogs that were produced in the cells by metabolic incorporation of BODIPY-labeled fatty acids (53). Fig. 2 shows the difference between inhibition by the two peptides. The data was plotted and then fit to a hyperbolic decay function. The concentration at which 50% inhibition is reached (IC 50 ) can be extracted from each curve. The IC 50 for the DP178 peptide was ϳ43 Ϯ 10 nM (R 2 ϭ 0.96) and 335 Ϯ 125 nM (R 2 ϭ 0.89) for solute and lipid mixing, respectively, while C34 was estimated at 11 Ϯ 3 nM (R 2 ϭ 0.97) and 25 Ϯ 4 nM (R 2 ϭ 0.98) for solute and lipid mixing, respectively. The difference in IC 50 between solute and aqueous redistribution of dye for DP178 is 8-fold, whereas that for C34 is only about 2-fold. In addition, statistical t tests were conducted on each point in both these curves to compare contents versus lipid mixing. There is no statistical difference for any of the concentrations tested for C34, but 100, 200, 400, and 800 nM concentrations of DP178 show a statistical difference (see Fig. 2).
A reasonable explanation for this difference is that DP178 has two target sites (46). One is the leucine/isoleucine zipper sequence of gp41 before it binds the membrane (43-45). We further checked whether interactions within the membrane may reveal another binding site.
DP178 Binds to the Membrane-We detected the binding of peptides to the membrane by using fluorescence resonance energy transfer (FRET) from the polypeptide tryptophan residues to DNS chromophores incorporated into the lipid vesicles. Excitation was set at the tryptophan maximum absorbance (280 nm), and binding was detected by measuring the changes in the sensitized DNS fluorescence (518 nm). When DP178 (0.45 M) was added to 68 M PS/PC/cholesterol/DNS-PE (8:8: 2:1, w/w) SUV (Fig. 3a), we observed a substantial increase in the DNS fluorescence, indicating that DP178 binds to the membrane. Since the DNS groups are attached to the head groups of the phospholipids, DP178 must be located near the surface of the membrane. In contrast, when C34 (0.45 M) was added to the vesicles (Fig. 3b), there was no change in the DNS fluorescence, indicating that the tryptophans of C34 do not interact with the DNS groups, or that the interaction is too weak to be detected by this assay (see also Fig. 4). Interestingly, while treatment of HIV-1-infected cells with soluble CD4 increases binding of a monoclonal antibody (98-6) to an epitope that spans residues 644 -663 (overlaps with C34 and DP178, see Fig. 1), binding of a monoclonal antibody (2F5) to an epitope that spans residues 662-667 (overlaps with DP178, but not with C34) was reduced (60). This is consistent with a model in which the conformational change of gp41 results in binding the DP178 region to the membrane, thus hiding the corresponding epitope.
The biological relevance of this membrane-binding feature is supported by the location of DP178 just upstream of the transmembrane domain. Experiments performed with either negatively charged PS/PC/cholesterol/DNS-PE (8:8:2:1, w/w) vesicles or zwitterionic PC/cholesterol/DNS-PE (16:2:1, w/w) vesicles qualitatively yielded similar results (data not shown), indicating that nonelectrostatic forces are involved in the peptide-membrane interactions.
Determination of the Membrane Binding Energy of the Peptides-The high sensitivity of NBD fluorescence to the polarity of its environment can be used to detect and quantify membrane binding (50,61). When SUV (PS/PC, 1:1) were added to NBD-labeled DP178 or C34, an increase concomitant with a blue shift of the NBD fluorescence was observed, suggesting that both peptides bind to the membrane. To evaluate the biological relevance of their membrane binding ability, we measured the energy of peptide-membrane interactions. To this end, NBD-labeled C34 (62) or DP178 (0.5 M) were titrated with SUV. Plotting the resulting increase in the fluorescence intensities of NBD-labeled peptides as a function of lipid/peptide molar ratios yielded conventional binding curves (Fig. 4a). The binding isotherms for the peptides can be obtained by plotting X b * (the molar ratio of bound peptide per lipid in the

DP178, Mechanism of Inhibition
outer leaflets) versus the equilibrium concentration of free peptide in solution. The surface partition coefficient (K) is calculated from the slope of the curve. The curve for the binding of DP178 to phospholipid vesicles has two phases, suggesting positive cooperativity for this binding (Fig. 4b, filled squares). The initial slope reveals K of 4 ϫ 10 4 M Ϫ1 (⌬G ϭ Ϫ8.7 Kcal/mol). Remarkably, C34 binds Ͼ10-fold weaker and with no cooperativity (Fig. 4b, open circles). Its surface partition coefficient was 3 ϫ 10 3 M Ϫ1 (⌬G ϭ Ϫ7.1 Kcal/mol).
The inhibitory activity of DP178 can be partially explained by its ability to bind to the gp41 leucine/isoleucine zipper sequence before binding to the membrane (43)(44)(45). Here, we examined whether the same sequence may also be the target for inhibition within the membrane.
The Leucine/Isoleucine Zipper Sequence Is Not a Target for Inhibition by DP178 within the Membrane-DP178 interacts with the leucine/isoleucine zipper sequence (represented by N36, see Fig. 1) in aqueous solution (45). Contrary, NBD-Rho FRET experiments that used NBD-DP178 and Rho-N36 (Fig. 5,  open squares) or NBD-N36 and Rho-DP178 (Fig. 5, open circles) revealed efficiencies of energy transfer that were close to the calculated values assuming a random distribution of donors and acceptors (Fig. 5, dashed line) (58). These data suggests that membrane-bound DP178 does not specifically interact with membrane-bound N36, therefore the leucine/isoleucine zipper sequence is not the target for inhibition by DP178 within the membrane.
Membrane-bound DP178 Forms Homo-oligomers-The ability of DP178 to oligomerize within the membrane was also measured by NBD-Rho FRET. When Rho-DP178 (a final concentration of 0.035-0.105 M) was added to a mixture of NBD-DP178 (0.11 M) and SUV (303 M), a dose-dependent quenching of the donor's emission, which is consistent with energy transfer, was observed (Fig. 5, filled squares). Note that the acceptor peptide was added only after the donor peptide was already bound to the membrane, thus preventing any association in solution. The lipid/peptide ratio in these experiments was kept high to create low surface density of donors and acceptors to reduce the energy transfer between unassociated peptide monomers. To confirm that the observed energy transfer was due to peptide aggregation, we compared the transfer efficiencies observed in the experiments with the energy transfer expected for randomly distributed membrane-bound donors and acceptors (Fig. 5, dashed line). The results reveal that membrane-bound DP178 forms homo-oligomers. DISCUSSION The striking similarities between structural motifs in various viral envelope glycoproteins led to the notion that the native conformation of gp41 is metastable and it is stabilized by gp120 (63). Upon binding of gp120 to its receptors, gp41 is free to form the more energetically favorable hairpin structure (Fig.  6, a-c). The gp41 ectodomain core is a six-helix bundle composed of three helical hairpins, each consisting of an N-helix paired with an antiparallel C helix. The N-helices form an interior, trimeric coiled coil with three conserved, hydrophobic grooves; a C-helix packs into each of these grooves. This structure is believed to correspond to the core of the fusion-active state of gp41 and shows similarity to the proposed fusogenic structures of envelope fusion proteins from influenza (64), Moloney murine leukemia virus (65,66), simian parainfluenza virus 5 (67), Ebola virus (68), and simian immunodeficiency virus (22,23,69). Synthetic C peptides (peptides corresponding to the C-helix), such as DP178 and C34, potently inhibit membrane fusion by both laboratory adapted strains and primary isolates of HIV-1 (33). The structural features of the gp41 core suggest that an intermediate, in which the leucine/isoleucine zipper sequence is not associated with the C-heptad repeat, exists before the formation of the hairpin structure, and is the target for inhibition by C-peptides (43)(44)(45). However, no direct evidence for the existence of this pre-hairpin intermediate exists, and an alternative for the mechanism of inhibition exerted by the C-peptides, based on the observed monomer-trimer equilibrium of the simian immunodeficiency virus gp41 has been postulated (22,70). According to the alternative model, gp41 exists in equilibrium between monomer and trimer. In the presence of excess inhibitory peptide, the equilibrium is driven from homotrimeric gp41 to a heterotrimer of gp41 and C-peptide. Since the peptides are only effective upon gp120 dissoci- ation (42), it follows that the absence of fusogenic activity displayed by the gp41-peptide heterotrimers is due to the fact that the heterotrimers can no longer present a sufficient number of fusion peptides to the target membrane for effective fusion to take place. Another explanation could be that the homotrimer structure is required for further oligomerization. The data also suggest that the homotrimeric state is stabilized by gp120 such that heterotrimers (which would be expected to bind less tightly to gp120 since they do not possess a trimeric loop structure) cannot be formed in the presence of bound gp120 (42). Anyway, the peptides act through a dominantnegative mechanism, in which exogenous C peptides bind to the leucine/isoleucine zipper sequence (N-helix) of gp41 and block further conformational changes needed for fusion.
Within each coiled-coil interface is a deep cavity, formed by a cluster of residues in the N-helix coiled coil that has been proposed to be an attractive target for the development of antiviral compounds. Chan et al. (59) showed that the inhibitory activity of the C-peptide C34 depends on its ability to bind to this coiled coil cavity. Moreover, examining a series of C34 peptide variants with modified cavity-binding residues, they observed a linear relationship between the logarithm of the inhibitory potency and the stability of the corresponding helical-hairpin complexes. A single mutation in the C34 peptide from Trp-631 3 Gly increased the IC 50 (43,44): a, native state of HIV-1 envelope glycoprotein; b, upon binding gp120 (green) to cellular receptors (not included in the figure), a conformational change in gp41 is induced, allowing an extension of the N-terminal fusion peptide (black) toward the cellular membrane. Subsequently, the leucine/isoleucine zipper (light blue) and the C-terminal amphipathic sequences (blue) form the threehairpin conformation (c). Binding of DP178 to the leucine/isoleucine zipper sequence can block this step (43)(44)(45). The resulting proximity of the viral and the cellular membranes within a limited area may allow lipid redistribution. Subsequent steps, not all of them defined, result in fusion between the viral and the cellular membranes. Among these steps, several membrane-bound gp41 oligomeric units interact to form a fusion pore. This step is blocked when DP178 peptide binds to the membrane-bound state of the DP178 region in gp41. The lower panel is based on the existence of a monomer-trimer equilibrium of gp41 ectodomain as suggested by Clore and colleagues (22,70): I, native state of HIV-1 envelope glycoprotein. Upon binding of gp120 (green) to cellular receptors (not included in the figure), shedding of gp120 and a conformational change in gp41 is induced. It is not known whether the result is the three-hairpin conformation (IIa), or a monomeric hairpin conformation (IIb) that will then trimerize. The resulting proximity of the viral and the cellular membranes within a limited area may allow lipid redistribution. Binding of DP178 to the monomeric form shifts this equilibrium toward the monomer. Subsequent steps, not all of them defined, result in fusion between the viral and the cellular membranes. Among these steps, several membrane-bound gp41 oligomeric units interact to form a fusion pore. This step is blocked when DP178 binds to the membrane-bound state of the gp41 segment that has a sequence identical to DP178. C34 appear to be quite similar (we find that the IC 50 for solute mixing is about 4-fold higher for DP178 as compared with C34). Since DP178 lacks all three residues from the C helix (Trp-628, Trp-631, and Ile-635) which insert into the cavity, we would expect the IC 50 to be at least 2 orders of magnitude higher for DP178.
This conundrum has led us to postulate that there are two targets for inhibition by DP178: (i) DP178 inhibits the formation of the heterotrimeric coiled coil by interacting with the leucine/isoleucine zipper sequence in the aqueous solution; (ii) fusion pore formation is inhibited by the interaction of DP178 with the membrane-bound state of its corresponding sequence in gp41 which will enhance melting of the coiled coil soluble form. Accordingly, a mutagenesis study, which revealed that the tryptophan-rich region (overlaps with DP178, see Fig. 1) must function in proximity to the membrane (71, 32) support the notion that this region binds to the surface of the membrane. For clarity, it is worth noting that three segments in the cytoplasmic tail of gp41 may associate with the membrane (48,(72)(73)(74)(75)(76)(77)(78)(79)(80), whereas both Rabenstein and Shin (81) and our results suggest that segments in the ectodomain bind to the membrane. These studies suggest that segments both upstream and downstream of the transmembrane domain of HIV-1 gp41 associate with the membrane.
Our results are consistent with the notion that C34 inhibits HIV-1-induced membrane fusion by blocking heterotrimeric coiled coil formation only. The relative low affinity of C34 to the membrane (Figs. 3 and 4) does not allow it to interfere with steps that occur within the membrane. DP178 interacts with this leucine/isoleucine zipper sequence when in aqueous environment (38,45), but this interaction cannot take place in the membrane environment (Fig. 5, open squares and open circles). On the other hand, DP178, which was found to be monomeric in aqueous solution when its concentration was less than 10 M (38), forms oligomers even at 100 times lower concentration in the membrane (Fig. 5, filled squares). However, we cannot rule out the possibility that an increase in its membrane local concentration might shift gp41 closed heterotrimeric (soluble)-gp41 open bundle (membrane-bound) equilibrium toward the closed bundle soluble form. In other words, DP178 can prevent melting of the heterotrimeric hairpin complex. The fact that DP107 (a shifted version of N36) inhibited (although at higher concentrations) HIV-1-mediated cell fusion with no differences between lipid and aqueous dye redistribution (46) may be correlated with its inability to interact with the membrane-bound state of the segment corresponding to DP178.
In the case of influenza hemagglutinin-mediated fusion, Chernomordik and co-workers (82) have shown that at reduced envelope glycoprotein surface density there is more hemi-fusion than full fusion. In the presence of different concentrations of peptide, the effective surface density of envelope glycoprotein is reduced, giving rise to different extents of hemi-fusion as compared with full fusion. This explains the 2-fold difference between inhibition of lipid and aqueous redistribution by C34 (Fig. 2a). However, the striking difference between inhibition of lipid versus content mixing in the case of DP178 (Fig. 2b) leads us to propose the following mode of inhibition by DP178 (see Fig. 6).
The binding of DP178 to the leucine/isoleucine zipper sequence (light blue) at an intermediate stage (Fig. 6, upper  panel, b; lower panel, IIb) inhibits the formation of the heterotrimeric coiled coil. How the binding of DP178 to the membrane-bound state of its corresponding segment in gp41 actually inhibits pore formation remains an open question. We can speculate that by binding to its corresponding segment in the membrane, DP178 prevents, by a dominant negative manner, further oligomerization of gp41. This possibility implicates the presence of another intermediate in the fusion reaction that takes place after the formation of the heterotrimeric coiled coil, but before the formation of the fusion pore. Indeed, it is believed that the recruitment of several oligomeric units of the fusion proteins is needed for complete fusion to occur (83)(84)(85)(86)(87)(88). Accordingly, by binding to its corresponding segment in the membrane, DP178 inhibits the formation of fusion pores (Fig.  6, the transition from c to d). Further studies are needed to understand the mechanism by which DP178 inhibits this process.