C-terminal octylation rescues an inactive T20 mutant: implications for the mechanism of HIV/SIMIAN immunodeficiency virus-induced membrane fusion.

T20, a synthetic peptide corresponding to a C-terminal segment of the envelope glycoprotein (gp41) of human and simian immunodeficiency viruses, is a potent inhibitor of viral infection. We report here that C-terminal octylation of simian immunodeficiency virus gp41-derived T20 induces a significant increase in its inhibitory potency. Furthermore, when C-terminally octylated, an otherwise inactive mutant in which the C-terminal residues GNWF were replaced by ANAA has potency similar to that of the wild type T20. This effect cannot be explained by a trivial inhibitory effect of the octyl group added to the peptides, because the N-terminally octylated peptides have the same activity as the non-octylated parent peptides. The effects caused by octylation on the oligomerization, secondary structure, and membrane-interaction properties of the peptides were investigated. Our results shed light on the mechanism of inhibition by T20 and provide experimental support for the existence of a pre-hairpin intermediate.

T20, a synthetic peptide corresponding to a C-terminal segment of the envelope glycoprotein (gp41) of human and simian immunodeficiency viruses, is a potent inhibitor of viral infection. We report here that C-terminal octylation of simian immunodeficiency virus gp41-derived T20 induces a significant increase in its inhibitory potency. Furthermore, when C-terminally octylated, an otherwise inactive mutant in which the C-terminal residues GNWF were replaced by ANAA has potency similar to that of the wild type T20. This effect cannot be explained by a trivial inhibitory effect of the octyl group added to the peptides, because the N-terminally octylated peptides have the same activity as the non-octylated parent peptides. The effects caused by octylation on the oligomerization, secondary structure, and membrane-interaction properties of the peptides were investigated. Our results shed light on the mechanism of inhibition by T20 and provide experimental support for the existence of a pre-hairpin intermediate.
Primate lentiviruses, such as HIV 1 and SIV, enter into their host cells by membrane fusion. Gp120/gp41 are the viral envelope glycoproteins responsible for this process. Gp120, which is non-covalently attached to the integral membrane subunit gp41, mediates the binding to the target cell receptor CD4 (1) and other co-receptors (2,3). These interactions induce a conformational change in gp41, the first step in a process that will lead to membrane merging (4,5). The N terminus of gp41 contains a stretch of approximately 15 hydrophobic residues named the fusion peptide (6) that are believed to insert into and destabilize the membrane, thus facilitating fusion (7). Gp41 ectodomain contains an N-terminal heptad repeat region consecutive to the fusion peptide and a C-terminal heptad repeat region that precedes the transmembrane domain (8). The three-dimensional structures of a protease-resistant core of HIV-1 gp41 (9, 10) as well as a segment of SIV gp41 ectodomain that lacks the fusion peptide (11) show that three N-terminal heptad repeats fold into a trimeric coiled coil, against which three antiparallel helices formed by C-terminal heptad repeats are packed. This structure is thought to form at a late stage during the membrane fusion process (12).
Synthetic peptides that partially overlap the C-terminal heptad repeat segment and a consecutive Trp-rich region have been shown to be potent inhibitors of gp41-mediated membrane fusion. These peptides can be grouped in two distinctive classes: (i) C34 and its different analogs contain in their N termini the sequence WMEW, which is believed to fit into a deep cavity located at the opposite end of the N-terminal coiled coil (13,14), and (ii) T20 (formerly DP178) and its analogs, which are shifted toward the C terminus and therefore do not include these cavity-filling residues (15). The analogy between the structural organizations of gp41 and hemagglutinin, the fusion protein of influenza virus (16), and the inhibitory activity of gp41 Cpeptides lead to postulate a "pre-hairpin intermediate" in gp41induced membrane fusion in which the N-terminal coiled coil is formed but the C-terminal helices are not packed (10,13). At this stage, the C-peptide inhibitors can bind to the exposed coiled coil, thus preventing the subsequent refolding and blocking fusion (12). In addition, extensive biophysical studies done with constructs corresponding to part of SIV gp41 ectodomain suggested that gp41 monomers, which might be present in equilibrium with the trimers, could be also a possible target for C-peptides binding and subsequent inhibition (40). The inhibitory potency of HIV gp41-derived C34 and its analogs has been clearly correlated with their ability to interact with the cavity present in the coiled coil (14). Furthermore, increasing their helical content resulted also in more potent peptide inhibitors (17)(18)(19). HIV gp41-derived T20 is a promising anti-HIV candidate in clinical trials (20). However, the understanding of its detailed mechanism of inhibition has been hampered by the following. (i) T20 lacks the residues that bind to the coiled coil C-terminal cavity; (ii) the known crystal and solution structures of fragments of HIV or SIV gp41 ectodomains do not fully include the regions corresponding either to T20 or to its putative binding site at the N terminus of the coiled coil; and (iii) in addition to a primary target site within the N-terminal heptad repeat (21), T20 has been postulated to bind to a second site in gp41, presumably a non-specified region at the C terminus of the ectodomain close to the viral membrane (22,23). Indeed, that T20 acts in close proximity to the membrane is supported by the studies of von Laer and co-workers (24). They have shown that when T20 is anchored to the target cell membrane by fusion to a transmembrane domain, the wild type and the ANAA mutant T20 have similar inhibitory activities. These observations suggest that either the original residues are im-portant for binding of T20 to its target site in gp41 or that their replacement affects T20 in an indirect way, perhaps by altering its structure or membrane partitioning. In either case, membrane partitioning by fusion to a transmembrane domain might have resulted in higher local concentrations, and perhaps the membrane proximity might have restored the secondary structure required for inhibition. To further study the role of membrane proximity in T20 inhibitory potency and to possibly discriminate between the different hypothesis, we synthesized and analyzed eight variants of SIV gp41-derived T20 (Fig. 1) in which different acyl groups were added either to the N or to the C terminus and compared the inhibitory activities, membrane partitioning, and secondary structural content of the different peptides. The rationale behind the design was that if T20membrane interaction was dependent on the aromatic C-terminal residues, C-terminal acylation could compensate the detrimental effects of the ANAA mutation. Furthermore, if a particular interaction between the C terminus of T20 and the membrane was necessary for inhibition, then C-terminal acylation of the wild type peptide could result in a better inhibitor. Furthermore, it is known that a minimum length of the acyl chain is necessary to significantly influence the way in which the peptide interacts with the membrane bilayer (36). On the other hand, long acyl chains are prone to form micelles, leading to peptide aggregation. Thus, although they are likely to affect peptide-membrane interaction (our purpose), they might alter also the "real" concentration of the peptides in the cell-cell fusion inhibition experiments, seriously compromising their interpretation. Thus, we chose to modify the N or C terminus of T20 and its ANAA mutant with an octyl group. In addition, we further explored the minimum length necessary to affect peptide-membrane interaction by specifically modifying the C terminus of the mutant T20 also with propyl and hexyl groups. The results are discussed in relation to the mechanism of T20 inhibition and gp41-mediated membrane fusion.
Peptide Synthesis and Modifications-The peptides were synthesized by a standard solid phase method on butyloxycarbonyl-methylbenzhydrylamine resin as described previously (25,26) with modifications outlined by Hackeng et al. (41) required for inclusion of the linker moiety (S-trityl-␤-mercaptopropionic acid). The fully protected resinbound WT T20 and mut-T20 peptides were N-acetylated with 1 equivalent of acetic anhydride (in two equivalents of N,N-diisopropylethylamine) for 15 min and then cleaved from the resin by HF treatment and purified by reverse phase-high performance liquid chromatography. C-terminal amidation of WT T20 and mut-T20 was achieved by incubating the pure peptides with a molar excess of NH 3 in acetonitrile/ water (1:1). Reverse phase-high performance liquid chromatography was used to follow the reaction and to purify the product to Ͼ98% homogeneity. Replacement of the linker and leucine groups at the C termini of the peptides as well as the removal of the formyl-protecting groups from tryptophans were evidenced by the disappearance of the shoulder at ϳ240 nm (characteristic of the thiol esther bond) and by the shift of the peak at ϳ300 nm (characteristic of the formyl-Trp) to 280 nm (as expected for Trp). The WT octyl-T20 and mut-octyl-T20 peptides were obtained by N-terminal coupling of octanoic acid to the fully protected resin-bound peptides by incubation with 7 equivalents of octanoic acid, 1 equivalent of PyBOP, and 1.1 equivalent of N,N-diisopropylethylamine for 2 h. Cleavage, purification, replacement of the C-terminal linker and Leu groups by amides, and Trp deprotection were done as described above. The WT T20-octyl, mut-T20-propyl, mut-T20hexyl, and mut-T20-octyl were obtained by N-acetylation, cleavage, and purification as described above followed by the replacement of the C-terminal linker and Leu groups with a propyl, hexyl, or octyl group (as required) by incubation with a molar excess of propyl, hexyl, or octyl amine (as required) in acetonitrile/water (1:1). Reverse phase-high performance liquid chromatography was used to follow the reaction and to purify the product to Ͼ98% homogeneity as described before. Note that incubation under these basic conditions successfully removed the formyl groups from protected Trp. The peptide compositions were determined by mass spectrometry.
Preparation of Lipid Vesicles-Small unilamellar vesicles were prepared from PC, SPM, PE, and Cho (4.5:4.5:1:1) as follows. Dry mixed lipid films were suspended in PBS buffer by vortexing to produce large multilamellar vesicles. Small unilamellar vesicles were then obtained by sonication of the lipid suspension for ϳ1 h until it cleared.
Dye Transfer Fusion Assay-Peptide inhibition of cell-cell fusion was assayed by monitoring the redistribution of water-soluble fluorescent probes between target and effector cells upon their co-incubation with each other (27). The expression of SIVmac251 gp41/gp120 on the effector cells was achieved by overnight infection of HeLa cells with a Vaccinia recombinant virus V194 at 31°C in Dulbecco's modified Eagle's medium, 10% fetal bovine serum. Effector cells were labeled with the fluorescent dye CMTMR (20 M) for 1 h at 37°C followed by two washes with PBS and one with Dulbecco's modified Eagle's medium, 10% fetal bovine serum. The target cells, ϩCD4 ϩCXCR5 3T3 fibroblasts, were labeled with the fluorescent dye calcein AM (1 M) for 1 h at 37°C followed by two washes with PBS and one with Dulbecco's modified Eagle's medium, 10% fetal bovine serum, and 3 g/ml puromycin and combined with the effector cells at room temperature. Different concentrations of the peptides dissolved in PBS were then added (or equivalent amounts of PBS were added to the controls). The cells were then co-cultured at 31°C for 30 min in 24-well plates (Costar, Cambridge, MA). Phase and fluorescence images were collected using a Nikon 200TE inverted microscope coupled to a CCD camera (SPOT Camera, Diagnostic Instruments, Inc.) with a ϫ20 objective lens (Plan-Fluor 20x, ELWD, NA ϭ 0.45). Single beam-splitter cubes were used for the excitation of calcein (494/517, Nikon B-2E/C) and CMTMR (541/565, Nikon G-2E/C). 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 based on the cells positive for both dyes in ratio with the total potential fusion events (target cells with bound or fused effector cells) as described previously (27). An average of fusion efficiency from 3 to 4 images is plotted against peptide concentration ([peptide]) with error bars for each point representing the mean Ϯ S.D. The data were fit to a hyperbolic decay function: (IC 50 ) is reached for each peptide. An estimation of error for IC 50 was also derived from this fit. Mut-T20, mut-octyl-T20, mut-T20-propyl, and mut-T20-hexyl yielded poor fitting coefficients, and therefore, no estimation of their IC 50 s was made.
Membrane Partition of the Peptides-Surface Plasmon Resonance (SRP) experiments to determine membrane partition were carried out with a BIAcore 3000 analytical system (Biacore, Uppsala, Sweden) using an L1 sensor chip (Biacore). The L1 chip has a long chain alkanethiol that contains polar head groups, so when in contact with liposomes, a lipid bilayer is formed. The running buffer used for all of the experiments was PBS, pH 7.3. The washing solution was 40 mM N-Octyl ␤-D-glucopyranoside. The regenerating solution was 10 mM NaOH. All of the solutions were freshly prepared, degassed, and filtered through 0.22-m pores. The operating temperature was 25°C. After cleaning as indicated by the manufacturers, the BIAcore instrument was left running overnight using Milli-Q water as eluent to thoroughly wash all of the liquid handling parts of the instrument. The L1 chip was then installed, and the alkanethiol surface was cleaned by an injection of the non ionic detergent (40 mM N-octyl ␤-D-glucopyranoside (25 l)) at a flow rate of 5 l/min. PC/SPM/PE/Cho (4.5:4.5:1:1) small unilamellar vesicles (80 l, 0.5 mM) were then applied to the chip surface at a low flow rate (2 l/min). To remove any multilamellar structures from the lipid surface, NaOH (50 l, 10 mM) was injected at a flow rate of 50 l/min, which resulted in a stable base line corresponding to the lipid bilayer linked to the chip surface. The bilayer linked to the chip surface was then used as a model membrane surface to study peptide-membrane binding. Peptide solutions (35 l of PBS solution of 0.4 -40 M peptide) were injected on the lipid surface at a flow rate of 5 l/min. PBS alone then replaced the peptide solution for 15 min to allow peptide dissociation. Surface plasmon resonance detects changes in the reflective index of the surface layer of lipids in contact with the sensor chip. A sensogram is obtained by plotting the surface plasmon resonance angle against time. An analysis of the peptide-lipid binding event was performed from a series of sensograms collected at several different peptide concentrations. Our system reached binding equilibrium during sample injection, and therefore, the affinity constant could be calculated from the relationship between the equilibrium binding response and the peptide concentration using a steady state affinity model. The affinity constants were derived from the following equation (by nonlinear least squares fitting) as shown in Equation 1, where R.U. (resonance unit) i is the signal measured, [peptide]i is the concentration of a generic peptide "i", R.U. MAX is the maximal response unit (or equilibrium binding response), and K A is the affinity constant. Determination of the Secondary Structure of the Peptides by CD-The CD spectra of the different peptides (10 M) in PBS were determined in an Aviv 202 spectropolarimeter in a capped quartz optical cell with a 1-mm path length at 25°C in the range of 193-260 nm (1 nm steps, 10 s averaging). The spectrum of PBS was subtracted from the spectra of the peptides.
Determination of Changes in the Oligomeric State of the Peptides by Trp Fluorescence-The fluorescence of Trp is sensitive to the polarity of its environment. Changes in the oligomerization state of short peptides alter the Trp surroundings with a concomitant change in their fluorescence signals. The spectra of the peptides at increasing concentrations (ϳ0.5-10 M) were determined in PBS at 25°C with excitation at 280 nm (4 nm slits) and emission in the range of 300 -400 nm (1-nm steps, 4-nm slits). To ensure equilibrium between possible different oligomeric states, the peptides were preincubated at the desired concentrations for 24 h at 4°C. The spectrum of PBS was subtracted from the spectra of the peptides.

C-terminal Octylation of SIV Gp41-derived T20 Enhances Its
Inhibitory Activity-First, we determined the ability of the different peptides to inhibit gp41-mediated cell-cell fusion. Specifically, we used effector HeLa cells expressing SIVmac251 gp41/gp120 (by infection with the Vaccinia recombinant virus V194) labeled with the fluorescent dye CMTMR and target cells, ϩCD4 ϩCXCR5 3T3 fibroblasts, labeled with the fluorescent dye calcein AM. The two cell types were co-incubated in the presence of increasing concentrations of the different peptides (or equivalent amounts of PBS in the controls) for 30 min, and then the percentage of gp41-mediated cell-cell fusion was calculated. As shown in Fig. 2, the non-acylated wild type peptide (Wt-T20, filled squares) efficiently prevents cell-cell fusion in a dose-dependent manner with an IC 50 of 400 Ϯ 25 nM. On the contrary, the replacement of the four C-terminal residues GNWF in the SIV strain studied here by ANAA resulted in an inactive peptide (Mut-T20, empty squares). Strikingly, the addition of an octyl group to the C termini of the wild type and the mutant peptides caused significant improvements of their inhibitory potencies with IC 50 of 123 Ϯ 30 nM and 534 Ϯ 129 nM, respectively (Wt-T20-octyl, filled diamonds; Mut-T20octyl, empty diamonds). Interestingly, the addition of the octyl group to the N termini of the wild type and mutant peptides (Wt-octyl-T20, filled triangles; Mut-octyl-T20, empty triangles) did not alter their inhibitory potencies. The IC 50 of WT octyl-T20 was 360 Ϯ 57 nM, whereas the mut-octyl-T20 showed no inhibition, indicating that the increase observed for WT T20-octyl and mut-T20-octyl is not caused by an intrinsic inhibitory ability of the octyl groups. Furthermore, mut-T20-propyl (crosses) and mut-T20-hexyl (empty circles) are also inactive, indicating that a minimum acyl chain length is necessary to bring about inhibitory activity to the mutant peptide. Interestingly, excluding the inactive peptides, Student's t test analysis showed that only WT T20-octyl had a statistically significant difference in IC 50 value as compared with that of wild type (data not shown).
The Increased Inhibitory Potency of the C-terminally Octylated Peptides Is Not the Result of Higher Peptide Concentrations in the Proximity of the Membrane-To determine whether the observed increase in the inhibitory activity of the C-terminally octylated peptides was simply the result of higher local concentrations caused by enhanced membrane binding, we calculated the membrane affinity constants of the different peptides by surface plasmon resonance using a model PC:SPM:PE: Cho (4.5:4.5:1:1) bilayer on a L1 sensor chip. An analysis of the peptide-lipid binding event was performed from a series of sensograms collected at several different peptide concentrations. The sensograms reached equilibrium during sample injection; therefore, the affinity constant was calculated from the

FIG. 2. Ability of the peptides to inhibit gp41-induced cell-cell fusion.
Effector HeLa cells expressing SIVmac251 gp41/gp120 (by infection with the Vaccinia recombinant virus V194) were labeled with the fluorescent dye CMTMR, and target cells, ϩCD4 ϩCXCR5 3T3 fibroblasts, were labeled with the fluorescent dye calcein AM. The two cell types were co-incubated in the presence of increasing concentrations of the different peptides (or equivalent amounts of PBS in the controls) for 30 min, and then the percentage of gp41-mediated cell-cell fusion was calculated. Symbols: f, WT (Wt) T20; OE, Wt-octyl-T20; ࡗ, Wt-T20-octyl; Ⅺ, Mut-T20; ‚, Mut-octyl-T20; छ, Mut-T20-octyl; crosses, Mut-T20-propyl; E, Mut-T20-hexyl. relationship between the equilibrium binding response and the peptide concentration (see Fig. 3) using a steady state affinity model. The derived affinity constants (K A ) and the r 2 of the fittings are listed in Table I. A comparison of the K A demonstrates that the non-acylated wild type and mutant peptides as well as the C-terminally octylated wild type and mutant peptides have similar membrane affinities. This suggests that the differences observed in their inhibitory potencies are not attributed to significantly different local concentrations. On the other hand, N-terminal octylation seems to have more complex effects on membrane partition.
Octylation Influences the Secondary Structure of T20 -Secondary structure has been shown to affect the inhibitory potency of both C34 and T20 analogs (17)(18)(19)28). Thus, we investigated whether octylation altered the secondary structure of the wild type and mutant peptides by means of CD. HIV-1 gp41-derived T20 has been reported to lack a well defined structure in aqueous environment (21). However, a 13-residue C-terminal fragment of HIV-1 gp41-derived T20 was recently shown to form a stable monomeric 3(10) helix by NMR (29). Indeed, the CD spectrum of this short fragment, which shows a minimum at ϳ205 nm typical of 3(10) helices (30 -34), is similar to the spectrum reported earlier for HIV-1 gp41-derived T20 (21,35), suggesting that the C-terminal region of T20 folds into a 3(10)-helical structure. Here, we observed a similar spectrum for SIV gp41-derived WT T20 (Fig. 4A, filled squares) in which a minimum at around 203-204 nm suggests that part of the peptide (or part of the population of the peptide) adopts a 3(10) structure and part is disordered (the minimum observed is actually a combination of the minima at ϳ200 and ϳ205 nm expected for random and 3(10) structures, respectively). The spectrum of mut-T20 (Fig. 4B, empty squares) shows a stronger minimum at around 201-202 nm, suggesting a higher proportion of disordered structure. Interestingly, C-terminal octylation of both the wild type and mutant peptides results in more stable structures. In the case of mut-T20-octyl (Fig. 4B, empty diamonds), the spectrum now is identical to that of WT T20, whereas in the case of WT T20-octyl (Fig. 4A, filled diamonds), the spectrum shows a double minimum at ϳ205 and 217 nm, suggesting the presence of both ␣-helical and unstructured regions (or populations). On the other hand, N-terminal octylation did not alter the structure of mut-octyl-T20 (Fig. 4B, empty triangles) and even destabilized that of WT octyl-T20 (Fig. 4A, filled triangles).
Octylation Does Not Alter the Oligomeric State of T20 -Oligomerization might affect the inhibitory potency of the peptides. HIV-1 gp41-derived T20 has been shown to be monomeric up to 10 M concentration (21). However, octylation could alter the ability of the peptides to self-associate. Thus, we analyzed whether the addition of octyl groups to the N or C terminus of wild type and mutant T20 affected their intrinsic oligomeric state in the range of concentrations relevant for the experiments on cell-cell fusion inhibition. Changes in the environment of tryptophan residues as a consequence of oligomerization result in changes in Trp fluorescence (36). Therefore, we studied possible changes in the oligomeric state of the peptides in PBS by Trp fluorescence. The effect of increasing peptide concentration on the position of the maximum of Trp emission as well as its intensity was used as an indicator of changes of  Table I. Symbols: f, WT (Wt) T20; OE, Wt-octyl-T20; ࡗ, Wt-T20-octyl; Ⅺ, Mut-T20; ‚, Mut-octyl-T20; छ, Mut-T20-octyl. The continuous lines are the fitted curves for the sets of wild type peptides, and the dotted lines represent the sets of mutant peptides.  the oligomeric state of the peptides. Increasing concentrations of the peptides from ϳ0.6 to ϳ10 M were added to PBS, and the Trp fluorescence was monitored. As a control, the fluorescence of free Trp at similar concentrations was also measured. As shown in Fig. 5, the intensity of the fluorescence of the six peptides, as well as that of the free Trp, increase linearly with the concentration of the peptides, suggesting that the oligomeric state of the peptides does not change within the range of concentrations analyzed. Furthermore, the position of the maximum of fluorescence for each peptide was constant for all of the tested concentrations (data not shown), confirming that their oligomeric states remain unchanged. Moreover, in particular, in the case of the wild type peptides, their Trp fluorescence varies within very similar values, suggesting that WT T20, WT octyl-T20, and WT T20-octyl have similar oligomeric states. DISCUSSION We have shown here that C-terminal octylation of SIV gp41derived T20 induces a significant increase in its inhibitory potency. Indeed, when C-terminally octylated, the otherwise inactive ANAA mutant has potency similar to that of the wild type T20 and the C-terminally octylated wild type peptide has a stronger inhibitory potency with an IC 50 statistically significantly lower than that of the non-octylated wild type peptide. This effect cannot be explained by a trivial inhibitory effect of the octyl group added to the peptides, because the N-terminally octylated peptides have the same activity of the non-octylated parent peptides. In the following paragraphs, we discuss possible hypotheses to explain these observations.
The Role of Secondary Structure on the Inhibitory Potencies of T20 and Its Analogs-Secondary structural content has been shown to influence the inhibitory potency of the two classes of inhibitory C peptides. C34 is unstructured in aqueous solution (18). On the contrary, it is fully helical when bound to the coiled coil (9 -11). This suggested the hypothesis that C34 inhibitory potency could be enhanced by simply helping C34 to adopt a helical structure in the unbound form. Indeed, adding short helix-capping sequences to both the N and C termini of a short peptide corresponding to the 19 N-terminal residues of C34 (17), replacing some of its residues by amino acids with stronger helical propensity (17) or by charged residues that would form matching ion pairs in a helical conformation (18), resulted in C34 analogs with higher helical content and stronger inhibitory abilities. Furthermore, the stabilization of the helical structure of a 14-residue peptide that partially overlaps the N terminus of C34 by chemical cross-linking and substitutions with unnatural helix-favoring amino acids was shown to significantly increase the inhibitory potency of the peptide (19). Similarly, Judice et al. (28) found a correlation between the helical content of short analogs of HIV-1 gp41-derived T20 designed to adopt constrained ␣-helical structures and their inhibitory potencies. These studies strongly indicate that helical structure is crucial for the inhibitory activity of the Cpeptides. However, as the modifications done to alter the structure of the peptides were located all along the sequences of the peptides, no conclusions can be drawn as to which region of the peptides is specifically required to be helical.
In this work, we have found that the addition of an octyl group to the C termini, but not to the N termini, of the wild type and the ANAA mutant SIV gp41-derived T20 increases the helical content of the peptides as determined by CD spectroscopy in PBS. The different structural consequences of adding the octyl group to the N or to the C terminus of the peptides suggest that these two regions have different structural roles in the free peptides. Interestingly, in the NMR structure of a 13-residue C-terminal fragment of HIV gp41-derived T20 (corresponding to HIV-1 gp41 residues 659 -671, ELLELDK-WASLWN) (29), the segment ELDKWASLW was shown to fold into a monomeric 3(10) helical structure. Biron et al. (29) show FIG. 6. Model for the enhanced inhibitory potency of the Cterminally octylated peptides. Upon interaction of gp120 (shown in orange) with target cell receptors (for clarity not shown), gp41 undergoes a conformational change that leads to the pre-hairpin intermediate structure in which the N-terminal coiled coil is exposed (shown here as red cylinders) and the fusion peptide (shown here as pink arrows) is inserted into the target membrane. Note that gp41 is shown here as a trimer, although monomers in equilibrium with trimers have been also suggested to be present (40). Subsequent refolding into the trimeric helical hairpin leads to membrane fusion (not shown). The C-terminally octylated T20 (shown as a short green cylinder with the octyl group represented by a blue line) interacts with the membrane preferentially via its hydrophobic C terminus. In this way, it is in the most favorable orientation to interact in an antiparallel manner with its target site in the N-terminal coiled coil, thus blocking fusion. that the helix is stabilized by i, i ϩ 3 side chain-side chain interactions on the hydrophobic face, specifically by the packing of W666 between the side chains of L663 and L669. In the SIV gp41-derived sequence analyzed here, it is possible to hypothesize that the side chain of the equivalent W677 would be packed between the side chains of L674 and F680. Moreover, in the peptide analyzed here, the 3(10) helix is very likely to continue up to the W683 whose side chain would be interacting in a similar way with the aromatic ring of F680. On the other hand, the replacement of Trp-683 by Ala in the ANAA mutant eliminates this interaction. This might destabilize the helical structure at the C terminus of the peptide, thus decreasing its inhibitory potency. We would like to speculate here that the octyl group linked to the C terminus might fold back and create new interactions with the side chains of the C-terminal residues on the hydrophobic face of the helix, thus further stabilizing the wild type peptide and partially restoring the helical character of the ANAA mutant. Similarly, it has been reported that palmitoylation of the short antimicrobial peptide Magainin stabilizes its helical structure by providing hydrophobic surface with which the hydrophobic face of the peptide could interact (36).
The Role of Peptide-Membrane Interactions on the Inhibitory Potencies of T20 and Its Analogs-A number of observations suggest a possible role for peptide-membrane interactions in the inhibitory mechanism of T20. First, the region corresponding to T20 in the gp41 ectodomain is located close to the transmembrane domain (37); thus, it would not be surprising that the T20 region actually interacts with the membrane surface. Second, Nieva and co-workers (38,39) have shown that HIV gp41-derived peptides corresponding to a Trp-rich region that partially overlaps the C terminus of T20 destabilize model membranes, suggesting that the equivalent region in gp41 might participate in the actual process of membrane fusion. Third, HIV gp41-derived T20 has been shown to bind to phospholipid membranes (23). Fourth, as already mentioned, membrane anchoring by fusion to a transmembrane domain restores the inhibitory potency of the ANAA mutant of HIV gp41-derived T20 (24).
The addition of an octyl group, in principle, could alter the membrane interaction properties of short peptides. Indeed, we found here that N-terminal octylation increased membrane binding of WT octyl-T20 but decreased that of mut-octyl-T20, highlighting the inherent difficulty for predicting the actual outcome of the octylation (note that N-terminal octylation also reduced the helical content of WT octyl-T20, presumably compensating its higher local concentration and, therefore, not altering significantly its IC 50 ). On the contrary, C-terminal octylation did not significantly alter membrane affinity (WT T20/WT T20-octyl and mut-T20/mut-T20-octyl have similar K A ). These observations suggest that if C-terminal octylation enhanced the inhibitory potency of WT T20 and mut-T20 by altering peptide-membrane interactions, it did not do so by simply augmenting membrane binding. Indeed, we have found here that the non-octylated wild type and ANAA mutant T20 have similar membrane affinities but have different inhibitory potencies. Thus, the replacement of the C-terminal residues GNWF by ANAA could alter peptide-membrane interactions in a subtle way, perhaps by changing the degree of penetration, the orientation, or the structure of the membrane-bound peptide. Indeed, we would like to postulate the following working model (illustrated in Fig. 6) where C-terminal octylation might alter peptide-membrane interaction by favoring one preferential orientation in which T20 interacts with the target cell membrane by its C terminus. In this way, the N terminus of the peptide might extend away from the membrane. Because in the hypothetical pre-hairpin intermediate the N-terminal fusion peptide is believed to be inserted into the target cell membrane (10,13), C-terminal membrane interaction places T20 in the most favorable orientation to interact in an antiparallel manner with its target site in the N-terminal coiled coil.