Report

Protein Design of an HIV-1 Entry Inhibitor

See allHide authors and affiliations

Science  02 Feb 2001:
Vol. 291, Issue 5505, pp. 884-888
DOI: 10.1126/science.1057453

Abstract

Human immunodeficiency virus type–1 (HIV-1) membrane fusion is promoted by the formation of a trimer-of-hairpins structure that brings the amino- and carboxyl-terminal regions of the gp41 envelope glycoprotein ectodomain into close proximity. Peptides derived from the carboxyl-terminal region (called C-peptides) potently inhibit HIV-1 entry by binding to the gp41 amino-terminal region. To test the converse of this inhibitory strategy, we designed a small protein, denoted 5-Helix, that binds the C-peptide region of gp41. The 5-Helix protein displays potent (nanomolar) inhibitory activity against diverse HIV-1 variants and may serve as the basis for a new class of antiviral agents. The inhibitory activity of 5-Helix also suggests a strategy for generating an HIV-1 neutralizing antibody response that targets the carboxyl-terminal region of the gp41 ectodomain.

Infection by HIV-1, the virus that causes AIDS, requires fusion of the viral and cellular membranes (1–3). This membrane-fusion process is mediated by the viral envelope glycoprotein complex (gp120/gp41) and receptors on the target cell. Binding of gp120/gp41 to cell-surface receptors (CD4 and a coreceptor, such as CCR5 or CXCR4) triggers a series of conformational changes in the gp120/gp41 oligomer that ultimately lead to formation of a trimer-of-hairpins structure in gp41 (Fig. 1A).

Figure 1

Targeting HIV-1 membrane fusion. (A) A schematic of HIV-1 membrane fusion depicting events that promote formation of the gp41 trimer-of-hairpins [adapted from (1)]. The NH2-terminal fusion peptide of gp41 (red), inaccessible in the native state, inserts into target cell membranes following gp120 interaction with CD4 and coreceptors. Formation of the prehairpin intermediate exposes the NH2-terminal coiled coil (gray), the target of C-peptide inhibition. This transient structure collapses into the trimer-of-hairpins state that brings the membranes into close apposition for fusion. (B) Lateral (left) and axial (right) views of a ribbon diagram representing the core of the gp41 trimer-of-hairpins. The ribbon diagram is derived from the crystal structure of a six-helix bundle formed by N36 (N-peptide, gray) and C34 (C-peptide, blue) (7). (C) A schematic model of the designed protein 5-Helix. Three N-peptide segments (N40, gray) and two C-peptide segments (C38, blue) are alternately linked (N-C-N-C-N) using short Gly/Ser peptide sequences (red loops) (21). The sequences of each segment in single-letter amino acid code are: N40, QLLSGIVQQQNNLLRAIEAQQHLLQLTVWGIKQLQARILA; C38, HTTWMEWDREINNYTSLIHSLIEESQNQQEKNEQELLE; N-to-C linker, GGSGG; and C-to-N linker, GSSGG.

The trimer-of-hairpins is a common structural element involved in the fusion process of many enveloped viruses, suggesting a critical role for this motif in promoting membrane fusion (3–5). In HIV-1 gp41, the core of the trimer-of-hairpins is a bundle of six α-helices (Fig. 1B): three α-helices (formed by the COOH-terminal regions of three gp41 ectodomains) pack in an antiparallel manner against a central, three-stranded coiled coil (formed by the NH2-terminal regions of the gp41 molecules) (6–9). The fusion peptide region, which inserts into the cellular membrane, is located at the extreme NH2-terminus of gp41, and the COOH-terminal region is adjacent to the transmembrane helix anchored in the viral membrane. Thus, the trimer-of-hairpins motif brings the two membranes together (Fig. 1A).

Agents that interfere with formation of the gp41 trimer-of-hairpins structure can inhibit HIV-1 infection. Peptides derived from the COOH-terminal region of the gp41 ectodomain, referred to as C-peptides (corresponding to the outer helices of the six-helix bundle), are effective inhibitors of HIV-1 infection (10–12). Studies from several groups support a mechanism of dominant-negative inhibition in which C-peptides bind to a transient gp41 species known as the prehairpin intermediate (Fig. 1A) (1, 6, 1215). In this prehairpin intermediate, the gp41 fusion peptide is embedded in the target-cell membrane, exposing the NH2-terminal three-stranded coiled coil [compare (16)]. Binding of C-peptides to the NH2-terminal region of the prehairpin structure prevents formation of the gp41 trimer-of-hairpins, ultimately leading to irreversible loss of membrane-fusion activity. C-peptides potently inhibit HIV-1 entry, with a mean inhibitory concentration (IC50) as low as 1 nM in vitro (11, 12). One such C-peptide is in clinical trials and shows antiviral activity in humans (2, 17). More recently, efforts to target a prominent pocket on the surface of the NH2-terminal coiled coil of the prehairpin intermediate have led to the discovery of small, cyclic D-peptides that inhibit HIV-1 infection, thereby validating the pocket as a potential target for development of small, orally bioavailable HIV-1 entry inhibitors (18).

The importance of trimer-of-hairpins formation for HIV-1 entry leads to the hypothesis that the COOH-terminal region on gp41 might also serve as a target for potential membrane-fusion inhibitors (Fig. 1A). If the COOH-terminal region is accessible (at least transiently) before formation of the trimer-of-hairpins, then agents that bind to this region of gp41 may prevent fusion. Consistent with this notion (6), peptides derived from the gp41 NH2-terminal region (referred to as N-peptides) are modest inhibitors of HIV-1 membrane fusion (micromolar IC50) (6, 19). The inhibitory mechanism of N-peptides, however, has not been ascertained, in part because these peptides have a strong tendency to aggregate. Indeed, a plausible alternative mechanism of action for the N-peptides is that they intercalate into the gp41 NH2-terminal coiled coil, thereby disrupting the trimeric interface (19, 20).

To directly test the hypothesis that the C-peptide region of gp41 is a potential target for the inhibition of HIV-1 entry, we designed a protein that binds tightly and specifically to this site. The design takes advantage of the binding properties of the N-peptide coiled coil while minimizing the tendency of the N-peptides to aggregate. In this designed protein, denoted 5-Helix, five of the six helices that make up the core of the gp41 trimer-of-hairpins structure are connected with short peptide linkers (Fig. 1C) (21). The 5-Helix protein lacks a third C-peptide helix, and this vacancy is expected to create a high-affinity binding site for the COOH-terminal region of gp41.

Under physiological conditions, 5-Helix is well folded, soluble, and extremely stable, with an α-helical content in close agreement with the value predicted from the design (Fig. 2, A and B). In affinity-interaction experiments, 5-Helix interacts strongly and specifically with epitope-tagged C-peptides (Fig. 2C). Moreover, this interaction induces a helical conformation in the bound C-peptide, as judged by the difference in circular dichroism (CD) before and after mixing (Fig. 2D). These properties are consistent with the intended design of 5-Helix.

Figure 2

Properties of 5-Helix (41). (A) CD spectrum of 5-Helix (10 μM) at 25°C. The spectrum indicates that the 5-Helix protein adopts >95% of the helical content expected from the design. (B) Thermal denaturation of 5-Helix monitored by ellipticity at 222 nm in TBS (filled squares) and in 3.7 M GuHCl/TBS (open circles). The denaturation observed in the GuHCl solution is >90% reversible. (C) Nickel-NTA precipitation of 5-Helix with a His-tagged C-peptide. Untagged 5-Helix and His-tagged C-peptide [denoted C37-H6 (39)] were mixed before Ni-NTA agarose was added in order to precipitate complexes containing C37-H6 (lanes 1 and 5; lanes numbered from left to right). Addition of excess untagged C-peptide (C34) shifts the 5-Helix molecules from the bound to the unbound fraction (lanes 2 and 6). (D) CD spectra of 5-Helix and C37-H6 before (filled squares) and after (open circles) mixing in a mixing cuvette (41). The increase in ellipticity at 222 nm upon mixing indicates an interaction between the two species that increases the total helical content (corresponding to an additional 28 helical residues per associated C-peptide).

The 5-Helix protein potently inhibits HIV-1 membrane fusion (nanomolar IC50), as measured by viral infectivity and cell-cell fusion assays (Fig. 3, A and B). In contrast, a control protein, denoted 6-Helix, in which the C-peptide binding site is occupied by an attached C-peptide (i.e., all six helices that constitute the gp41 trimer-of-hairpins have been linked into a single polypeptide) (21, 22), does not have appreciable inhibitory activity (Fig. 3A). Likewise, a 5-Helix variant, denoted 5-Helix(D4), in which the C-peptide binding site is disrupted by mutation of four interface residues (V549, L556, Q563, and V570) to Asp (23), does not block membrane fusion even at 1 μM (Fig. 3A). We conclude that C-peptide binding is the key determinant of antiviral activity in 5-Helix.

Figure 3

Inhibition of HIV-1 envelope-mediated membrane fusion by 5-Helix (42). (A) Titration of viral infectivity by 5-Helix (filled squares), 6-Helix (open triangles), and 5-Helix(D4) [open circles (23)]. The data represent the mean ± SEM of two or more separate experiments. (B) Antagonistic inhibitory activities of 5-Helix and C34. The number of syncytia were measured in a cell-cell fusion assay performed in the absence or presence of 5-Helix, C34, or mixtures of 5-Helix and C34 at the indicated concentrations. The IC50 values for 5-Helix and C34 in this assay are 13 ± 3 nM and 0.55 ± 0.03 nM (12), respectively. Data represent the mean and range of mean of duplicate measurements, except for the control (mean ± SEM of five measurements). (C) Shown is 5-Helix inhibition of pseudotyped virus containing different HIV-1 envelope glycoproteins. The reported IC50 values represent the mean ± SEM of three independent experiments.

The inhibitory activities of 5-Helix and C-peptides are expected to be antagonistic: when 5-Helix binds C-peptide, the amino acid residues thought to be responsible for the antiviral activities of each inhibitor are buried in the binding interface. Indeed, mixtures of 5-Helix and C34 [a well-characterized and potent C-peptide inhibitor with an IC50 ≈ 1 nM (12)] display a dose-dependent antagonistic effect (Fig. 3B). In the presence of 5-Helix, high-potency inhibition by C34 is only observed when the peptide is in stoichiometric excess (Fig. 3B).

The 5-Helix protein inhibits infection by viruses pseudotyped with a variety of HIV-1 envelope proteins (from clades A, B, and D) with similar potency (Fig. 3C). This broad-spectrum inhibition likely reflects the highly conserved interface between the NH2- and COOH-terminal regions within the gp41 trimer-of-hairpins structure (Fig. 4). The residues in the C-peptide region of gp41 that are expected to make contact with 5-Helix are highly conserved in HIV-1, HIV-2, and simian immunodeficiency virus (SIV) (Fig. 4).

Figure 4

A helical wheel diagram depicting the interaction of 5-Helix with the C-peptide region of gp41. The (a) through (g) positions in each helix represent sequential positions in the 4,3 hydrophobic heptad repeat in each sequence. The (a) and (d) positions in the gp41 C-peptide region interact with the exposed (e) and (g) positions on the N40 coiled coil of 5-Helix. Residues are boxed according to their degree of conservation as determined from the alignment of 247 sequences from HIV-1, HIV-2, and SIV isolates (HIV-1 sequence database, August 2000, Los Alamos National Laboratory): black rectangle, >90% identical; gray rectangle, >90% conservative substitution (43)]; dotted rectangle, 70 to 90% conserved; no box, <70% conserved. Note the high degree of conservation in the (a) and (d) positions of the C-peptide region of gp41, a property markedly lacking in other positions [particularly (c) and (g)] of the C-peptide region not directly involved in binding 5-Helix.

As a potent, broad-spectrum inhibitor of viral entry, 5-Helix may serve as the basis for development of a new class of therapeutic agents against HIV-1 (24). Moreover, 5-Helix offers flexibility in the design of variants with better therapeutic characteristics. In principle, 5-Helix can be modified extensively to alter its immunogenic, antigenic, bioavailability, or inhibitory properties (25). For example, the C-peptide binding site might be lengthened, shortened, or shifted in the gp41 sequence in order to optimize inhibitory potency by targeting different regions of the gp41 ectodomain.

It would be desirable to generate neutralizing antibodies that mimic the binding properties of 5-Helix. Unstructured C-peptide immunogens may not elicit broadly neutralizing antibodies, because the linear sequence of the gp41 C-peptide region is variable among different HIV-1 strains. The potent and broad-spectrum inhibitory properties of 5-Helix suggest that an HIV-1 neutralizing antibody response might be generated using C-peptide analogs constrained in a helical conformation (i.e., as in the C-peptide region when it binds to 5-Helix).

Interestingly, the epitope for 2F5, a human monoclonal antibody directed against gp41 with broad neutralizing activity, is located immediately COOH-terminal to the C-peptide region targeted by 5-Helix (26–28). It is unknown if 2F5 inhibits infection by interfering with trimer-of-hairpins formation. The conformation of the 2F5-bound epitope has recently been shown to exist in a hairpin turn (29). Although antibodies elicited with fragments of gp41 containing this sequence do not possess significant virus-neutralizing activity (30, 31), it is possible that constrained analogs (perhaps with an adjacent helical C-peptide region) will lead to useful immunogens in efforts to develop an AIDS vaccine.

Alternatively, 5-Helix itself is a potential vaccine candidate. The possibility of eliciting an antibody response against transiently exposed conformations of proteins involved in HIV-1 fusion has been suggested (32). One possible well-defined target is the NH2-terminal coiled coil that is exposed in the prehairpin intermediate (18). Speculatively, a 5-Helix–like intermediate may be exposed during the fusion process (33) and, in this case, antibodies directed against 5-Helix may inhibit viral entry.

Finally, structural (4) and computational (5) methods predict similar trimer-of-hairpins motifs for viruses in diverse families, including orthomyxoviridae, paramyxoviridae, filoviridae, and retroviridae. In some of these cases, inhibition of viral entry by peptides analogous to the C-peptides of gp41 has been demonstrated (34–36). Thus, the 5-Helix design approach may offer a widely applicable strategy for inhibiting viral infections (37).

  • * To whom correspondence should be addressed. E-mail: kimadmin{at}wi.mit.edu

  • Present address: Merck Research Laboratories, 770 Sumneytown Pike, West Point, PA 19486, USA.

REFERENCES AND NOTES

View Abstract

Navigate This Article