Voltar
Artigo Acesso aberto Revisado por pares
BIBTEX RIS

Design of a Novel Peptide Inhibitor of HIV Fusion That Disrupts the Internal Trimeric Coiled-coil of gp41

2002; Elsevier BV; Volume: 277; Issue: 16 Linguagem: Inglês

10.1074/jbc.m201453200

ISSN

1083-351X

Autores

Carole A. Bewley, John M. Louis, Rodolfo Ghirlando, G. Marius Clore,

Tópico(s)

HIV/AIDS drug development and treatment

Resumo

The pre-hairpin intermediate of gp41 from the human immunodeficiency virus (HIV) is the target for two classes of fusion inhibitors that bind to the C-terminal region or the trimeric coiled-coil of N-terminal helices, thereby preventing formation of the fusogenic trimer of hairpins. Using rational design, two 36-residue peptides, N36Mut(e,g) and N36Mut(a,d), were derived from the parent N36 peptide comprising the N-terminal helix of the gp41 ectodomain (residues 546–581 of HIV-1 envelope), characterized by analytical ultracentrifugation and CD, and assessed for their ability to inhibit HIV fusion using a quantitative vaccinia virus-based fusion assay. N36Mut(e,g) contains nine amino acid substitutions designed to disrupt interactions with the C-terminal region of gp41 while preserving contacts governing the formation of the trimeric coiled-coil. N36Mut(a,d) contains nine substitutions designed to block formation of the trimeric coiled-coil but retains residues that interact with the C-terminal region of gp41. N36Mut(a,d) is monomeric, is largely random coil, does not interact with the C34 peptide derived from the C-terminal region of gp41 (residues 628–661), and does not inhibit fusion. The trimeric coiled-coil structure is therefore a prerequisite for interaction with the C-terminal region of gp41. N36Mut(e,g) forms a monodisperse, helical trimer in solution, does not interact with C34, and yet inhibits fusion about 50-fold more effectively than the parent N36 peptide (IC50∼ 308 nmversus ∼16 μm). These results indicate that N36Mut(e,g) acts by disrupting the homotrimeric coiled-coil of N-terminal helices in the pre-hairpin intermediate to form heterotrimers. Thus N36Mut(e,g)represents a novel third class of gp41-targeted HIV fusion inhibitor. A quantitative model describing the interaction of N36Mut(e,g) with the pre-hairpin intermediate is presented. The pre-hairpin intermediate of gp41 from the human immunodeficiency virus (HIV) is the target for two classes of fusion inhibitors that bind to the C-terminal region or the trimeric coiled-coil of N-terminal helices, thereby preventing formation of the fusogenic trimer of hairpins. Using rational design, two 36-residue peptides, N36Mut(e,g) and N36Mut(a,d), were derived from the parent N36 peptide comprising the N-terminal helix of the gp41 ectodomain (residues 546–581 of HIV-1 envelope), characterized by analytical ultracentrifugation and CD, and assessed for their ability to inhibit HIV fusion using a quantitative vaccinia virus-based fusion assay. N36Mut(e,g) contains nine amino acid substitutions designed to disrupt interactions with the C-terminal region of gp41 while preserving contacts governing the formation of the trimeric coiled-coil. N36Mut(a,d) contains nine substitutions designed to block formation of the trimeric coiled-coil but retains residues that interact with the C-terminal region of gp41. N36Mut(a,d) is monomeric, is largely random coil, does not interact with the C34 peptide derived from the C-terminal region of gp41 (residues 628–661), and does not inhibit fusion. The trimeric coiled-coil structure is therefore a prerequisite for interaction with the C-terminal region of gp41. N36Mut(e,g) forms a monodisperse, helical trimer in solution, does not interact with C34, and yet inhibits fusion about 50-fold more effectively than the parent N36 peptide (IC50∼ 308 nmversus ∼16 μm). These results indicate that N36Mut(e,g) acts by disrupting the homotrimeric coiled-coil of N-terminal helices in the pre-hairpin intermediate to form heterotrimers. Thus N36Mut(e,g)represents a novel third class of gp41-targeted HIV fusion inhibitor. A quantitative model describing the interaction of N36Mut(e,g) with the pre-hairpin intermediate is presented. Virus-cell and cell-cell fusion mediated by the viral envelope glycoproteins (Env) 1The abbreviations used are: Envviral envelope glycoprotein(s)HIVhuman immunodeficiency virusgp120surface envelope glycoprotein of HIVgp41transmembrane subunit of HIV envelopeN36 and C34peptides encompassing residues 546–581 and 628–661 of HIV-1 Env, respectivelyN36Mut(eg), peptide derived from N36 that contains nine substitutions at positionse and g of the helical wheel (defined in the context of the gp41 trimer of hairpins structure) corresponding to residues 549, 551, 556, 558, 563, 565, 570, 572, and 577 of HIV-1 EnvN36Mut(ad), peptide derived from N36 that contains nine substitutions at positions a and d of the helical wheel (defined in the context of the gp41 trimer of hairpins structure) corresponding to residues 552, 555, 559, 562, 566, 569, 573, 576, and 580 of HIV-1 Env1The abbreviations used are: Envviral envelope glycoprotein(s)HIVhuman immunodeficiency virusgp120surface envelope glycoprotein of HIVgp41transmembrane subunit of HIV envelopeN36 and C34peptides encompassing residues 546–581 and 628–661 of HIV-1 Env, respectivelyN36Mut(eg), peptide derived from N36 that contains nine substitutions at positionse and g of the helical wheel (defined in the context of the gp41 trimer of hairpins structure) corresponding to residues 549, 551, 556, 558, 563, 565, 570, 572, and 577 of HIV-1 EnvN36Mut(ad), peptide derived from N36 that contains nine substitutions at positions a and d of the helical wheel (defined in the context of the gp41 trimer of hairpins structure) corresponding to residues 552, 555, 559, 562, 566, 569, 573, 576, and 580 of HIV-1 Env(1.Freed E.O. Martin M.A. J. Biol. Chem. 1995; 270: 23883-23886Abstract Full Text Full Text PDF PubMed Scopus (223) Google Scholar) gp41 and gp120 constitute the first step of infection and dissemination, respectively, of the human immunodeficiency virus (HIV) and hence represent a promising target for the development of antiviral therapeutics (2.Eckert D.M. Kim P.S. Annu. Rev. Biochem. 2001; 70: 777-810Crossref PubMed Scopus (1139) Google Scholar). Following binding of gp120 to CD4 and a chemokine receptor, a conformational change occurs in the gp120/gp41 oligomer that leads to insertion of the fusion peptide of gp41 into the target membrane and ultimately membrane fusion (2.Eckert D.M. Kim P.S. Annu. Rev. Biochem. 2001; 70: 777-810Crossref PubMed Scopus (1139) Google Scholar, 3.Moore J.P. Trkola A. Dragic T. Curr. Opin. Immunol. 1997; 9: 551-562Crossref PubMed Scopus (451) Google Scholar). The structure of the ectodomain of both HIV and simian immunodeficiency virus gp41 in its fusogenic/postfusogenic state has been solved by NMR (4.Caffrey M. Cai M. Kaufman J. Stahl S.J. Wingfield P.T. Covell D.G. Gronenborn A.M. Clore G.M. EMBO J. 1998; 17: 4572-4584Crossref PubMed Scopus (370) Google Scholar) and crystallography (5.Chan D.C. Fass D. Berger J.M. Kim P.S. Cell. 1997; 89: 263-273Abstract Full Text Full Text PDF PubMed Scopus (1833) Google Scholar, 6.Weissenhorn W. Dessen A. Harrison S.C. Skehel J.J. Wiley D.C. Nature. 1997; 387: 426-430Crossref PubMed Scopus (1459) Google Scholar, 7.Tan K.J. Liu J. Wang S. Shen S. Lu M. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12303-12308Crossref PubMed Scopus (518) Google Scholar, 8.Malashkevich V.N. Chan D.C. Chutkowski C.T. Kim P.S. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 9134-9139Crossref PubMed Scopus (190) Google Scholar) and shown to consist of a trimer of hairpins. Each subunit comprises long N- and C-terminal helices connected by a 25–30-residue loop. The N-helices form a parallel, trimeric coiled-coil in the interior of the complex surrounded by the C-terminal helices oriented antiparallel to the N-terminal helices (Fig. 1a, bottom). Peptides derived from the C- and N-helices inhibit Env-mediated fusion at nanomolar and micromolar concentrations, respectively (9.Wild C.T. Oas T. McDanal C.B. Bolognesi D. Matthews T.J. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 10537-10541Crossref PubMed Scopus (480) Google Scholar, 10.Wild C.T. Shugars D.C. Greenwell T.K. McDanal C.B. Matthews T.J. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 9770-9774Crossref PubMed Scopus (881) Google Scholar, 11.Chan D.C. Chutkowski C.T. Kim P.S. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 15613-15617Crossref PubMed Scopus (484) Google Scholar, 12.Eckert D.M. Kim P.S. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 11187-11192Crossref PubMed Scopus (248) Google Scholar). These peptides do not bind the fusion-active or postfusogenic state of gp41 as represented by the ectodomain of gp41 free in solution and are thought to interact with a pre-hairpin intermediate (2.Eckert D.M. Kim P.S. Annu. Rev. Biochem. 2001; 70: 777-810Crossref PubMed Scopus (1139) Google Scholar, 13.Furuta R.A. Wild C.T. Weng Y. Weiss C.D. Nat. Struct. Biol. 1998; 5: 276-279Crossref PubMed Scopus (466) Google Scholar, 14.Chan D.C. Kim P.S. Cell. 1998; 93: 681-684Abstract Full Text Full Text PDF PubMed Scopus (1110) Google Scholar) in which the N- and C-helices are not associated and the trimeric coiled-coil of N-helices is exposed (Fig. 1a, top left). Peptides derived from the C-terminal helix, such as C34 (residues 628–661 of HIV-1 Env) and T20 (residues 638–673 of HIV-1 Env; currently in phase III clinical trials (15.Kilby J.M. Hopkins S. Venetta T.M. DiMassimo B. Cloud G.A. Lee J.Y. Alldredge L. Hunter E. Lambert D. Bolognesi D. Matthews T. Johnson M.R. Nowak M.A. Shaw G.M. Saag M.S. Nat. Med. 1998; 4: 1302-1307Crossref PubMed Scopus (939) Google Scholar, 16.Pozniak A. J. HIV Res. 2001; 6: 92-94Google Scholar)) target the exposed face of the trimeric coiled-coil of N-helices (11.Chan D.C. Chutkowski C.T. Kim P.S. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 15613-15617Crossref PubMed Scopus (484) Google Scholar, 13.Furuta R.A. Wild C.T. Weng Y. Weiss C.D. Nat. Struct. Biol. 1998; 5: 276-279Crossref PubMed Scopus (466) Google Scholar, 14.Chan D.C. Kim P.S. Cell. 1998; 93: 681-684Abstract Full Text Full Text PDF PubMed Scopus (1110) Google Scholar, 17.Weissenhorn W. Dessen A. Calder L.J. Harrison S.C. Skehel J.J. Wiley D.C. Mol. Membr. Biol. 1999; 16: 3-9Crossref PubMed Scopus (320) Google Scholar, 18.Kliger Y. Shai Y. J. Mol. Biol. 2000; 295: 163-168Crossref PubMed Scopus (89) Google Scholar). Engineered constructs such as the chimeric protein NCCG-gp41 (19.Louis J.M. Bewley C.A. Clore G.M. J. Biol. Chem. 2001; 276: 29485-29489Abstract Full Text Full Text PDF PubMed Scopus (99) Google Scholar), which features an exposed, stable, disulfide-linked, trimeric coiled-coil of N-helices grafted onto the minimal, thermostable ectodomain of gp41; peptides in which the trimeric coiled-coil of N-helices is stabilized by fusion to the GCN4 trimeric coiled-coil (12.Eckert D.M. Kim P.S. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 11187-11192Crossref PubMed Scopus (248) Google Scholar); and the protein 5-helix (20.Root M.J. Kay M.S. Kim P.S. Science. 2001; 291: 884-888Crossref PubMed Scopus (379) Google Scholar), in which the internal trimeric coiled-coil of N-helices is surrounded by only two C-helices, specifically target the C-region in the pre-hairpin intermediate state (Fig. 1a,top left). In both instances, packing of the C-region onto the trimeric coiled-coil of N-helices is blocked, and hairpin formation is inhibited. Although the ectodomain of gp41 in free solution is highly thermostable (with a Tm in excess of 100 °C) (21.Wingfield P.T. Stahl S.J. Kaufman J. Zlotnick A. Hyde C.C. Gronenborn A.M. Clore G.M. Protein Sci. 1997; 6: 1653-1660Crossref PubMed Scopus (43) Google Scholar), it has been shown to exist as a monomer-trimer equilibrium (21.Wingfield P.T. Stahl S.J. Kaufman J. Zlotnick A. Hyde C.C. Gronenborn A.M. Clore G.M. Protein Sci. 1997; 6: 1653-1660Crossref PubMed Scopus (43) Google Scholar, 22.Caffrey M. Kaufman J. Stahl S.J. Wingfield P.T. Gronenborn A.M. Clore G.M. Protein Sci. 1999; 8: 1904-1907Crossref PubMed Scopus (44) Google Scholar). In the context of the fusion process, the trimeric coiled-coil of N-helices in the pre-hairpin intermediate state may also exist as a monomer-trimer equilibrium (4.Caffrey M. Cai M. Kaufman J. Stahl S.J. Wingfield P.T. Covell D.G. Gronenborn A.M. Clore G.M. EMBO J. 1998; 17: 4572-4584Crossref PubMed Scopus (370) Google Scholar, 22.Caffrey M. Kaufman J. Stahl S.J. Wingfield P.T. Gronenborn A.M. Clore G.M. Protein Sci. 1999; 8: 1904-1907Crossref PubMed Scopus (44) Google Scholar, 23.Peisajovich S.G. Shai Y. J. Mol. Biol. 2001; 311: 249-254Crossref PubMed Scopus (12) Google Scholar). If this is indeed the case, blocking the formation of the fusion-competent, homotrimeric coiled-coil of N-helices may provide another molecular target for inhibiting HIV cell fusion. In this article, we present the design and characterization of a peptide, derived from the N-helix of gp41, in which the sites of interaction with the C-helices have been mutated, but the sites of intermolecular contacts between the N-helices have been preserved. This peptide, which we term N36Mut(e,g), is about 50-fold more effective in inhibiting HIV Env-mediated cell fusion than the N36 peptide (residues 546–581 of HIV-1 Env) of gp41 from which it was derived. These data strongly suggest that the homotrimeric coiled-coil of N-helices in the pre-hairpin state can be disrupted. viral envelope glycoprotein(s) human immunodeficiency virus surface envelope glycoprotein of HIV transmembrane subunit of HIV envelope peptides encompassing residues 546–581 and 628–661 of HIV-1 Env, respectively g), peptide derived from N36 that contains nine substitutions at positionse and g of the helical wheel (defined in the context of the gp41 trimer of hairpins structure) corresponding to residues 549, 551, 556, 558, 563, 565, 570, 572, and 577 of HIV-1 Env d), peptide derived from N36 that contains nine substitutions at positions a and d of the helical wheel (defined in the context of the gp41 trimer of hairpins structure) corresponding to residues 552, 555, 559, 562, 566, 569, 573, 576, and 580 of HIV-1 Env viral envelope glycoprotein(s) human immunodeficiency virus surface envelope glycoprotein of HIV transmembrane subunit of HIV envelope peptides encompassing residues 546–581 and 628–661 of HIV-1 Env, respectively g), peptide derived from N36 that contains nine substitutions at positionse and g of the helical wheel (defined in the context of the gp41 trimer of hairpins structure) corresponding to residues 549, 551, 556, 558, 563, 565, 570, 572, and 577 of HIV-1 Env d), peptide derived from N36 that contains nine substitutions at positions a and d of the helical wheel (defined in the context of the gp41 trimer of hairpins structure) corresponding to residues 552, 555, 559, 562, 566, 569, 573, 576, and 580 of HIV-1 Env All peptides (Fig. 2b), purchased from Commonwealth Biotechnologies (Richmond, VA), were synthesized on a solid phase support, purified by reverse phase high pressure liquid chromatography, and verified for purity by mass spectrometry and amino acid composition. All peptides bear an acetyl group at the N terminus and an amide group at the C terminus. Concentrations of peptides were determined spectrophotometrically: the calculatedA280 values (1-cm path length) for a concentration of 1 mg/ml N36, N36Mut(e,g), N36Mut(a,d), and C34 are 1.35, 1.31, 1.34, and 2.90, respectively. The corresponding molecular masses are 4160, 4293, 4182, and 4286 Da, respectively. CD spectra of peptides (at a concentration corresponding to 0.7–0.8 A280) were recorded at 25 °C on a JASCO J-720 spectropolarimeter using a 0.05-cm path length cell. Quantitative evaluation of secondary structure from the CD spectrum was carried out using the program CDNN (www.bioinformatik.biochemtech.uni-halle.de/cd_spect/index.html; Ref. 24.Bohm G. Muhr R. Jaenicke R. Protein Eng. 1992; 5: 191-195Crossref PubMed Scopus (1010) Google Scholar). Sedimentation equilibrium experiments were conducted at 20.0 °C and three different rotor speeds (16,000, 20,000, and 24,000) on a Beckman Optima XL-A analytical ultracentrifuge. Peptide samples were prepared in 50 mmsodium formate buffer (pH = 3) and loaded into the ultracentrifuge cells at nominal loading concentrations of ∼0.2 and 0.7–0.8A280. Data were analyzed in terms of a single ideal solute to obtain the buoyant molecular mass, M(1 − vρ), using the Optima XL-A data analysis software (Beckman). The value for the experimental molecular mass Mwas determined using calculated values for the density ρ (determined at 20 °C using standard tables) and partial specific volumev (calculated on the basis of amino acid composition (25.Perkins S.J. Eur. J. Biochem. 1986; 157: 169-180Crossref PubMed Scopus (541) Google Scholar)). Inhibition of HIV Env-mediated cell fusion by peptides was carried out as described previously (19.Louis J.M. Bewley C.A. Clore G.M. J. Biol. Chem. 2001; 276: 29485-29489Abstract Full Text Full Text PDF PubMed Scopus (99) Google Scholar) using a modification (26.Salzwedel K. Smith E. Dey B. Berger E.J. J. Virol. 2000; 74: 326-333Crossref PubMed Scopus (134) Google Scholar) of the vaccinia virus-based reporter gene assay (using soluble CD4 at a final concentration of 200 nm). B-SC-1 cells were used for both target and effector cell populations. Target cells were co-infected with vCB21R-LacZ and vCBYF1-fusin (CXCR4), and effector cells were co-infected with vCB41 (Env) and vP11T7gene1 at a multiplicity of infection of 10. For inhibition studies, peptides were added to an appropriate volume of Dulbecco's modified Eagle's medium (2.5%) and phosphate-buffered saline to yield identical buffer compositions (100 μl) followed by addition of 1 × 105 effector cells (in 50 μl of medium) per well. After incubation for 15 min, 1 × 105 target cells (in 50 μl) and soluble CD4 were added to each well. Following 2.5 h of incubation, β-galactosidase activity of cell lysates was measured (A570; Molecular Devices 96-well spectrophotometer) upon addition of chlorophenol red-β-d-galactopyranoside (Roche Molecular Biochemicals). The curves for %fusion versus peptide inhibitor concentration were fit by nonlinear least-squares optimization using the program FACSIMILE (27.Chance E.M. Curtis A.R. Jones I.P. Kirby C.R. FACSIMILE, Atomic Energy Research Establishment Report R8775. Harwell, H. M. Stationary Office, London, UK1979Google Scholar, 28.Clore G.M. Geisow M.J. Barrett A.N. Computing in Biological Science. Elsevier/North-Holland, New York1983: 313-348Google Scholar). The helical wheel diagram in Fig. 2a illustrates the interactions between the N-helices and between the N- and C-helices as observed in both the NMR (4.Caffrey M. Cai M. Kaufman J. Stahl S.J. Wingfield P.T. Covell D.G. Gronenborn A.M. Clore G.M. EMBO J. 1998; 17: 4572-4584Crossref PubMed Scopus (370) Google Scholar) and x-ray (5.Chan D.C. Fass D. Berger J.M. Kim P.S. Cell. 1997; 89: 263-273Abstract Full Text Full Text PDF PubMed Scopus (1833) Google Scholar, 6.Weissenhorn W. Dessen A. Harrison S.C. Skehel J.J. Wiley D.C. Nature. 1997; 387: 426-430Crossref PubMed Scopus (1459) Google Scholar, 7.Tan K.J. Liu J. Wang S. Shen S. Lu M. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12303-12308Crossref PubMed Scopus (518) Google Scholar, 8.Malashkevich V.N. Chan D.C. Chutkowski C.T. Kim P.S. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 9134-9139Crossref PubMed Scopus (190) Google Scholar) structures of the fusogenic/postfusogenic state of the ectodomain of gp41. Internal contacts between the N-helices involve positions a and d of the helical wheel (5.Chan D.C. Fass D. Berger J.M. Kim P.S. Cell. 1997; 89: 263-273Abstract Full Text Full Text PDF PubMed Scopus (1833) Google Scholar). Each C-helix interacts with two N-helices (one intra- and the other intersubunit): these contacts involve positions e andg of the N-helices and positions a andd of the C-helix (5.Chan D.C. Fass D. Berger J.M. Kim P.S. Cell. 1997; 89: 263-273Abstract Full Text Full Text PDF PubMed Scopus (1833) Google Scholar). The first crystal structure of the HIV-1 gp41 ectodomain core consisted of a complex of N36 and C34 peptides comprising residues 546–581 and 628–661, respectively, of HIV-1 Env (5.Chan D.C. Fass D. Berger J.M. Kim P.S. Cell. 1997; 89: 263-273Abstract Full Text Full Text PDF PubMed Scopus (1833) Google Scholar). Using the N36 and C34 peptides as starting points, we designed two peptides: N36Mut(e,g), which can only undergo self-association but cannot interact with C34, and N36Mut(a,d), which can no longer self-associate but could potentially still interact with C34 (Fig. 2b). In the case of N36Mut(e,g), the residues at positions e andg of N36 have been replaced by residues at positionse and g of C34. Since the latter residues are located on the external surface of C34 in the context of the ectodomain gp41 core (4.Caffrey M. Cai M. Kaufman J. Stahl S.J. Wingfield P.T. Covell D.G. Gronenborn A.M. Clore G.M. EMBO J. 1998; 17: 4572-4584Crossref PubMed Scopus (370) Google Scholar, 5.Chan D.C. Fass D. Berger J.M. Kim P.S. Cell. 1997; 89: 263-273Abstract Full Text Full Text PDF PubMed Scopus (1833) Google Scholar, 6.Weissenhorn W. Dessen A. Harrison S.C. Skehel J.J. Wiley D.C. Nature. 1997; 387: 426-430Crossref PubMed Scopus (1459) Google Scholar, 7.Tan K.J. Liu J. Wang S. Shen S. Lu M. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12303-12308Crossref PubMed Scopus (518) Google Scholar, 8.Malashkevich V.N. Chan D.C. Chutkowski C.T. Kim P.S. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 9134-9139Crossref PubMed Scopus (190) Google Scholar) and since C34 on its own is monomeric (29.Lu M. Blacklow S.C. Kim P.S. Nat. Struct. Biol. 1995; 2: 1075-1082Crossref PubMed Scopus (668) Google Scholar), 2C. A. Bewley, J. M. Louis, R. Ghirlando, and G. M. Clore, unpublished data. this set of substitutions will prevent any interaction between N36Mut(e,g) and the C-region of gp41 in its pre-hairpin intermediate state while preserving the intermolecular contacts required to form the trimeric coiled-coil of N-helices. In the case of N36Mut(a,d), the residues at positions a andd of N36 have been substituted by residues at positionsf and c of C34, which are located on the solvent-exposed face of the ectodomain core of gp41 (4.Caffrey M. Cai M. Kaufman J. Stahl S.J. Wingfield P.T. Covell D.G. Gronenborn A.M. Clore G.M. EMBO J. 1998; 17: 4572-4584Crossref PubMed Scopus (370) Google Scholar, 5.Chan D.C. Fass D. Berger J.M. Kim P.S. Cell. 1997; 89: 263-273Abstract Full Text Full Text PDF PubMed Scopus (1833) Google Scholar, 6.Weissenhorn W. Dessen A. Harrison S.C. Skehel J.J. Wiley D.C. Nature. 1997; 387: 426-430Crossref PubMed Scopus (1459) Google Scholar, 7.Tan K.J. Liu J. Wang S. Shen S. Lu M. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 12303-12308Crossref PubMed Scopus (518) Google Scholar, 8.Malashkevich V.N. Chan D.C. Chutkowski C.T. Kim P.S. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 9134-9139Crossref PubMed Scopus (190) Google Scholar), thereby removing the intermolecular contacts required to form the trimeric coiled-coil of N-helices. The results of analytical ultracentrifugation on N36Mut(e,g) and N36Mut(a,d) are presented in Fig. 3a. N36Mut(e,g)behaves as a single monodisperse species at concentrations of ∼36 μm (in monomer; A280 ∼ 0.2) and ∼124 μm (in monomer; A280 ∼ 0.7) with a molecular mass of ∼12,000–12,500 Da, corresponding to a trimer. In this context it is worth noting that N36 on its own aggregates and does not form a well defined trimer (12.Eckert D.M. Kim P.S. Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 11187-11192Crossref PubMed Scopus (248) Google Scholar),2presumably due to further self-association involving the predominantly hydrophobic residues at positions e andg, which have been substituted by predominantly hydrophilic residues in N36Mut(e,g) (Fig. 2b). N36Mut(a,d) also behaves as a single monodisperse species at a concentration of ∼140 μm(A280 ∼ 0.8), but its molecular mass is only ∼3700 Da, corresponding to a monomer. CD spectra of N36Mut(e,g) and N36Mut(a,d) are shown in Fig. 3b. N36Mut(e,g) displays a double minimum at 208 and 222 nm, characteristic of an α-helix; and quantification of the CD data (24.Bohm G. Muhr R. Jaenicke R. Protein Eng. 1992; 5: 191-195Crossref PubMed Scopus (1010) Google Scholar) indicates a helical content of ∼80%. N36Mut(a,d), on the other hand, is largely random coil (characterized by a minimum around 200 nm) with a small amount of α-helix (∼20%). We were unable to detect any evidence of interaction between either N36Mut(e,g) or N36Mut(a,d) and C34 by either analytical ultracentrifugation or CD. The absence of interaction between N36Mut(e,g) and C34 is exactly as predicted from the design since the points of contact with C34 have been mutated (cf. Fig. 2). The absence of interaction between N36Mut(a,d) and C34 was initially somewhat surprising since the residues that contact C34 in the context of the fusogenic/postfusogenic state of the gp41 ectodomain were preserved. This result therefore indicates that C34 can only form a complex with a stable trimeric coiled-coil of N-helices. From a structural standpoint, this is readily understood since each C-helix contacts two N-helices of the trimeric coiled-coil (one intramolecular and the other intersubunit; cf. Fig. 2a), and the buried surface area for each of the two interactions is comparable. To exclude the remote possibility that N36Mut(e,g) could behave in a manner analogous to C34 and bind to the surface of the trimeric coiled-coil of N-helices in the pre-hairpin intermediate of gp41, we also examined the interaction of N36Mut(e,g) with the engineered protein NCCG-gp41. NCCG-gp41 is a chimeric protein that features an exposed trimeric coiled-coil of N-helices that is stabilized both by fusion to a minimal thermostable ectodomain of gp41 and by engineered intersubunit disulfide bonds (19.Louis J.M. Bewley C.A. Clore G.M. J. Biol. Chem. 2001; 276: 29485-29489Abstract Full Text Full Text PDF PubMed Scopus (99) Google Scholar). The exposed trimeric coiled-coil of N-helices in NCCG-gp41 mimics that of the pre-hairpin intermediate of gp41, but in contrast to native gp41, the N-helices cannot dissociate since they are covalently tethered by disulfide bonds. Analytical ultracentrifugation on various mixtures of N36Mut(e,g) and NCCG-gp41 in ratios of 4.5:1 and 11.7:1 (24 μm N36Mut(e,g) plus 5.3 μmNCCG-gp41 and 51 μm N36Mut(e,g)plus 4.4 μm NCCG-gp41, respectively, with concentrations expressed in trimer) provided no evidence of any interactions between these two molecules, and the data were readily accounted for by a mixture of two ideal species. The results of a quantitative vaccinia virus-based reporter gene assay (26.Salzwedel K. Smith E. Dey B. Berger E.J. J. Virol. 2000; 74: 326-333Crossref PubMed Scopus (134) Google Scholar) for HIV Env-mediated cell fusion are shown in Fig. 4. N36 inhibits fusion with an IC50 of 16 ± 2 μm in agreement with previous results (19.Louis J.M. Bewley C.A. Clore G.M. J. Biol. Chem. 2001; 276: 29485-29489Abstract Full Text Full Text PDF PubMed Scopus (99) Google Scholar). N36Mut(e,g) inhibits fusion with an IC50 308 ± 22 nm. Thus N36Mut(e,g) is ∼50-fold more active in inhibiting fusion than N36. N36Mut(a,d), on the other hand, fails to inhibit fusion even at concentrations as high as 0.1 mm. The lack of any fusion-inhibitory activity for N36Mut(a,d) is exactly as predicted from the biophysical data since N36Mut(a,d) does not self-associate and does not interact with C34. Since N36Mut(e,g) forms a well defined trimeric species that does not interact with either C34 or the chimeric protein NCCG-gp41 (in which the N-helices of the solvent-exposed trimeric coil-coil are covalently linked by interhelical disulfide bonds), it must target the N-region of the pre-hairpin intermediate by forming fusion-incompetent heterotrimers (Fig. 1b). Analytical ultracentrifugation on the ectodomain of gp41 indicates the presence of only monomer and trimer species in equilibrium (21.Wingfield P.T. Stahl S.J. Kaufman J. Zlotnick A. Hyde C.C. Gronenborn A.M. Clore G.M. Protein Sci. 1997; 6: 1653-1660Crossref PubMed Scopus (43) Google Scholar, 22.Caffrey M. Kaufman J. Stahl S.J. Wingfield P.T. Gronenborn A.M. Clore G.M. Protein Sci. 1999; 8: 1904-1907Crossref PubMed Scopus (44) Google Scholar); that is, assembly of the trimer is a highly cooperative process. The fusion-inhibitory activity of N36Mut(e,g) therefore indicates the presence of a dynamic equilibrium between monomeric and trimeric forms of membrane-bound gp41 that allows subunit exchange to take place in the pre-hairpin intermediate state. The rate of exchange between these species must be sufficiently fast to permit efficient heterotrimer formation within the lifetime (∼15 min) of the pre-hairpin intermediate (13.Furuta R.A. Wild C.T. Weng Y. Weiss C.D. Nat. Struct. Biol. 1998; 5: 276-279Crossref PubMed Scopus (466) Google Scholar, 30.Jones P.L. Korte T. Blumenthal R. J. Biol. Chem. 1998; 273: 404-409Abstract Full Text Full Text PDF PubMed Scopus (177) Google Scholar, 31.Muñoz-Barroso I. Durell S. Sakaguchi K. Appella E. Blumenthal R. J. Cell Biol. 1998; 140: 315-323Crossref PubMed Scopus (269) Google Scholar). The inhibition curve for N36Mut(e,g) is well fit by a simple Langmuir isotherm given by %fusion = 100/(1 + [N36Mut(e,g)]/IC50) (Fig. 4). Yet, mechanistically, the interaction of N36Mut(e,g) with the pre-hairpin intermediate of gp41 is far more complex, involving multiple species in different homo- and hetero-oligomerization states. The simplest scheme describing the situation is presented in Fig. 5a. L and M represent N36Mut(e,g) and the pre-hairpin intermediate of gp41 in their monomeric forms, respectively; LL and MM are homodimers; ML is a heterodimer; LLL and MMM are homotrimers; and MML and MLL are heterodimers. We assume that only the homotrimer MMM is fusion-active, and the fraction fusion activity is given by the ratio of [MMM]LT/[MMM]LT = 0. The interactions between ligand (in its various oligomerization states) and membrane-bound protein (in its various homo- and hetero-oligomeric states) are described by their respective bulk solution concentrations. The interactions involving only membrane-bound species, however, are dependent on their concentrations in the two-dimensional membrane (i.e. number of molecules per unit area) that are much higher than their concentrations in bulk solvent. In terms of thermodynamics, all equilibria in Fig. 5a can be related to the species concentrations in bulk solvent by multiplying the relevant equilibrium constants by a factor λ to yield appropriate apparent equilibrium constants (Fig. 5a, middle panel). The measured equilibrium association constant Kreftrimerfor the ectodomain of HIV-1 gp41 in free solution (i.e. the trimer of hairpins) is 4.8 × 1011m−2 and is given by the product of the equilibrium association constantsK1 (monomer-dimer equilibrium) andK2 (dimer-trimer equilibrium) (Fig. 5a, bottom panel). Since trimerization of the gp41 ectodomain is highly cooperative (21.Wingfield P.T. Stahl S.J. Kaufman J. Zlotnick A. Hyde C.C. Gronenborn A.M. Clore G.M. Protein Sci. 1997; 6: 1653-1660Crossref PubMed Scopus (43) Google Scholar, 22.Caffrey M. Kaufman J. Stahl S.J. Wingfield P.T. Gronenborn A.M. Clore G.M. Protein Sci. 1999; 8: 1904-1907Crossref PubMed Scopus (44) Google Scholar),K2 ≫ K1. Taking Kreftrimeras a reference point, the overall equilibrium association constant between monomeric and homotrimeric species of L is given by α Kreftrimer, between monomeric and homotrimeric species of M by λ Kreftrimer(note that λ also subsumes any difference in the energetics of trimerization between the pre-hairpin intermediate in the membrane and the ectodomain of gp41 in free solution), and between monomeric species of L and M and heterotrimeric species of M and L by 3β Kreftrimer(where the factor 3 is a statistical factor). The scheme in Fig. 5a has three unknowns: α, β, and λMT, where MT is the total protein concentration (in monomer units). The data, however, are insufficient to determine all three parameters independently. Nonlinear least-squares fitting to the experimental data, optimizing the values of α and β, was carried out for values of λMT ranging from 1.5 × 10−7 to 1.5 × 10−3m (which corresponds to values of λ of 1.5 × 104–1.5 × 108 for MT = 10 pm, the probable concentration of protein in bulk solution, estimated on the basis of a concentration of 5 × 103cells/μl and ∼5000 gp41 trimers/cell). (Note that the concentrations of the various species in the scheme shown in Fig. 5a as a function of total ligand concentration, LT, were calculated numerically by integration of the differential equations describing the reactions to essentially infinite time.) The optimized values of α and β depend on the product λMT, and the results are therefore equally valid for a wide range of MT concentrations. The data can be equally well fitted for values of λMT ranging from 10−7 to 10−3m with α varying from ∼10 to 0.1 and β varying from 1 to 10 (Fig. 5b). Best fits to the experimental fusion inhibition data for λMT = 1.5 × 10−5 and 1.5 × 10−4m are shown in Fig. 5c; the optimized values of α are 1.07 and 0.34 (with error estimates of ∼40%), respectively, and of β are 2.97 and 5.76 (with error estimates of 10%), respectively. The resulting curves are essentially indistinguishable from each other as well as from that obtained with a Langmuir isotherm. The occupancy of the various species relative to MT and LT are shown in Fig. 5, d ande, respectively. For this set of parameters, the fraction M in the trimeric state in the absence of ligand is ∼86% for λMT = 1.5 × 10−5m and ∼97% for λMT = 1.5 × 10−4m; the value of LT at which 50% of L is monomeric is ∼2 × 10−6 and 5 × 10−6m, respectively. The occupancy of homodimeric ligand is less than 1% of LT; likewise the occupancy of homodimeric (MM) and heterodimeric (LM) protein is less than 1% of MT for all values of LT. Both MML and MLL heterotrimers are formed with the MML heterotrimer peaking at concentrations of LT slightly less than that at which 50% of the ligand is monomeric. The above calculations reveal two important findings. First, despite the complexities introduced by multiple homo- and hetero-oligomerization states, which might lead one to predict a complex relationship between fusion and total ligand concentration, a scheme such as that depicted in Fig. 5a can still yield rather simple inhibition data that is readily characterized by a Langmuir isotherm. Second, the values for the various equilibrium constants for trimerization required to best fit the experimental fusion data are entirely compatible with the experimentally measured value for the equilibrium constant for trimerization of the ectodomain of HIV-1 gp41 in solution. In the best fit calculations described above and depicted in Fig. 5, only the homotrimeric form of the pre-hairpin intermediate of gp41, MMM, is considered to be fusion-active. If the calculations are repeated assuming that the heterotrimer, MML, containing only one molecule of N36Mut(e,g), is also fusion-active, the resulting theoretical curves do not reproduce the experimental data. One can therefore conclude that the energetics of formation of a five-helix bundle comprising a heterotrimeric internal coiled-coil consisting of two N-helices of gp41 and one N36Mut(e,g)helix surrounded by two C-helices of gp41 is not sufficiently favorable to bring the target and viral membranes into sufficiently close proximity for fusion to take place. Using rational design, we have engineered two peptides derived from the N-helix of the ectodomain of gp41. The parent peptide, N36, corresponds to residues 546–581 of HIV-1 Env and encompasses the N-terminal helix of gp41. The N36Mut(a,d)peptide was designed to remove interactions leading to self-association and the formation of a trimeric coiled-coil of N-helices while preserving those residues that interact with the C-helix of the ectodomain of gp41. The absence of any fusion-inhibitory activity of N36Mut(a,d) leads us to conclude that the C-region of gp41 can only interact with a trimeric coiled-coil of N-helices. The N36Mut(e,g) peptide was designed to preserve the interactions leading to self-association while replacing those residues that interact with the C-region. N36Mut(e,g) forms a monodisperse trimer in solution that does not interact with the C-region of gp41 and yet still inhibits fusion about 50-fold more effectively than the native gp41 sequence (i.e. N36) from which it was derived. These results can only be explained by the existence of a dynamic equilibrium between monomeric and trimeric coiled-coil forms of the N-region of gp41 in the pre-hairpin intermediate on a time scale sufficiently fast to permit subunit exchange and the consequent formation of heterotrimers of the N-helices of gp41 and N36Mut(e,g). Thus, N36Mut(e,g)disrupts the homotrimeric coiled-coil of N-helices in the pre-hairpin intermediate state of gp41 and represents a novel third class of gp41-targeted fusion inhibitor. The other two classes of inhibitors bind to either the homotrimeric coiled-coil of N-helices (e.g. C34 and T20) or to the exposed C-region (e.g. NCCG-gp41 and 5-helix) of gp41 in the pre-hairpin intermediate state. Since C34 (and presumably T20) also binds to NCCG-gp41 and 5-helix (19.Louis J.M. Bewley C.A. Clore G.M. J. Biol. Chem. 2001; 276: 29485-29489Abstract Full Text Full Text PDF PubMed Scopus (99) Google Scholar, 20.Root M.J. Kay M.S. Kim P.S. Science. 2001; 291: 884-888Crossref PubMed Scopus (379) Google Scholar), these two classes of inhibitors antagonize each other. In contrast, one would predict that the N36Mut(e,g) class of inhibitors should act either additively or synergistically with either of the other two classes. Therefore, N36Mut(e,g) may represent a promising lead for the design of clinically effective, novel fusion inhibitors. We thank A. Szabo for stimulating discussions, I. Nesheiwat and L.C. Chang for technical assistance, E. Berger for soluble CD4, and L. Pannell for mass spectrometry.

Referência(s)