Solution structure of the HIV gp120 C5 Domain Laure Guilhaudis, Amy Jacobs and Michael Caffrey Department of Biochemistry and Molecular Biology, University of Illinois at Chicago, IL, USA In HIV the viral envelope protein is processed by a host cell protease to form gp120 and gp41. The C1 and C5 domains of gp120 are thought to directly interact with gp41 but are largely missing from the available X-ray structure. Bio- physical studies of the HIV gp120 C5 domain (residues 489– 511 of HIV-1 strain HXB2), which corresponds to the carboxy terminal region of gp120, have been undertaken. CD studies of the C5 domain suggest that it is unstructured in aqueous solutions but partially helical in trifluoroethanol/ aqueous and hexafluoroisopropanol/aqueous buffers. The solution structure of the C5 peptide in 40% trifluoroethanol/ aqueous buffer was determined by NMR spectroscopy. The resulting structure is a turn helix structural motif, consistent with the CD results. Fluorescence titration experiments suggest that HIV C5 forms a 1 : 1 complex with the HIV gp41 ectodomain in the presence of cosolvent with an apparent K d of  1.0 l M . The absence of complex forma- tion in the absence of cosolvent indicates that formation of the turn-helix structural motif of C5 is necessary for complex formation. Examination of the C5 structure provides insight into the interaction between gp120 and gp41 and provides a possible target site for future drug therapies designed to disrupt the gp120/gp41 complex. In addition, the C5 struc- ture lends insight into the site of HIV envelope protein maturation by the host enzymes furin and PC7, which provides other possible targets for drug therapies. Keywords:AIDS;gp41;gp120;HIV;NMR. Envelope proteins are known to play roles in directing viral particles to the appropriate target cell and initiating the fusion of the viral and cellular membranes in diverse viruses including retroviridae, herpesviridae, filoviridaie and para- myxoviridae [1]. In the retrovirus HIV, which is the causative agent of AIDS, the envelope proteins exist as a noncovalent complex of a surface subunit (gp120) and a transmembrane subunit (gp41), which are proteolytic products of the env gene. HIV infection, as well as SIV (simian immunodeficiency virus) infection, is thought to occur in three discreet steps [2]. During the initial step, the gp41/gp120 complex associates with the CD4 receptor. In the second step, gp120 associates with a chemokine receptor that is specific to the cell type and dissociates from gp41. In the final step, gp41 mediates fusion of the viral and the cellular membranes by a poorly understood mechanism, which is thought to involve the N-terminal hydrophobic region of gp41, termed the fusion peptide. gp120 is known to be well exposed on the HIV surface and play an initial and critical role in HIV infection. Thus, gp120 is an attractive target for vaccines and structure-based drugs designed for the prevention and treatment of AIDS. The search for new AIDS therapies is of critical importance in light of recent reports that current highly active antiretro- viral treatments are ineffective in 20–50% of treated patients [3]. Moreover, therapies designed to disrupt gp120 function are especially attractive because of their potential prophy- lactic properties, as well as being alternative and comple- mentary therapies to the protease and reverse transcriptase inhibitors currently utilized. In HIV, the precursor envelope protein, gp160, is proteo- lytically cleaved to form gp120 and gp41 by the host specific proteases furin and PC7 [4]. Cleavage of the precursor viral envelope protein at the carboxyl side of the amino acid sequence REKR is critical to HIV function, as shown by mutagenesis studies [5]. Moreover, inhibitors of furin and PC7 have been shown to block gp160 processing and HIV infection [6]. The mature gp120 consists of 5 conserved regions (C1-C5) and five variable regions (V1–V5). Recently, the structure of HIV gp120 in complex with regions of the CD4 receptor and a monoclonal antibody that blocks gp120 binding to the chemokine receptor has been determined by X-ray crystallography [7]. However, for technical reasons, large regions of the conserved and variable domains of gp120 are missing from the X-ray structure. Mutagenesis and deletion studies have suggested that the C1 and C5 domains of gp120, which are largely missing in the crystal structure but would be expected to be proximal to one another, are involved in the interaction with gp41 [8,9]. Furthermore, it has recently been shown that the introduction of non-native intermolecular disulfide bonds between the gp41 loop region and the C1 or C5 domains of gp120 stabilizes the gp41/gp120 complex and mimics the antigenic properties of the native envelope complex [10]. Taken together, the mutagenesis studies suggest that there is a direct interaction between the gp120 C1 and C5 domains and the gp41 loop region. In an effort to further characterize the gp41/gp120 interaction, the solution structure of the C5 domain of HIV gp120 has been determined by NMR spectroscopy. EXPERIMENTAL PROCEDURES The HIV gp120 C5 domain was prepared by solid phase peptide synthesis (Biological Resource Center, University of Illinois at Chicago). The peptide was modified by the Correspondence to M. Caffrey, Department of Biochemistry and Molecular Biology, University of Illinois at Chicago, 1819 W Polk St., Chicago, IL 60612, USA. Fax: + 1 312 413 0364, Tel.: + 1 312 996 4959. E-mail: caffrey@uic.edu. (Received 12 June 2002, revised 9 August 2002, accepted 19 August 2002) Eur. J. Biochem. 269, 4860–4867 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03187.x additions of acetyl and amido groups to the amino and carboxy termini, respectively. The identity of the peptide was verified by mass spectrometry (mass ¼ 2700.9). The wild-type HIV gp41 ectodomain was prepared as previously described [11]. CD spectra were recorded on a Jasco J-710 spectropolarimeter with a 0.1-cm path length at 23 °C. Fluorescence titrations were performed with a PTI Fluo- rescence System using a 0.1-cm path length at 25 °C. Before performing the titrations, the HIV gp41 ectodomain and gp120 C5 were equilibrated in 50 m M sodium formate (pH 3.0), 40% trifluoroethanol (v/v). Data were analyzed with KALEIDAGRAPH 3.08. For the NMR experiments, the conditions were 1 m M C5, 50 m M PO 4 (pH 6.0), 40% trifluoroethanol-d 3 (Cam- bridge Isotopes), 5% D 2 O. NMR experiments were per- formed on a Bruker DRX 600 equipped with a triple resonance probe. TOCSY and NOESY experiments were acquired at 285, 290, 295, 300 and 305 K. For the TOCSY experiments the isotropic mixing time was 72 ms; for the NOESY experiments, the mixing times were 75, 150, and 250 ms. 13 C-edited HSQC were acquired at 300 and 305 K. The HIV gp120 C5 structure was calculated with the program CNS [12]. The protocol consisted of four steps: (a) high temperature torsion angle molecular dynamics start- ing from an extended conformation; (b) slow-cooling torsion angle molecular dynamics; (c) slow-cooling Carte- sian dynamics; and (d) conjugate gradient minimization. For the first step, the temperature was set to 50 000 K for 2000 steps of 15 fs. In this step, the force constants for NOE, dihedral angles, and van der Waals forces were 150 kcalÆ mol )1 ÆA ˚ )2 , 100 kcalÆmol )1 Ærad )2 and 0.1 kcalÆmol )1 ÆA ˚ )4 , respectively. For the second step, the temperature was set to 50 000 K and cooled to 0 K over 1000 steps of 15 fs. In this step, the force constants for NOE, dihedral angles, and van der Waals forces were 150 kcalÆmol )1 ÆA ˚ )2 , 200 kcalÆ mol )1 Ærad )2 and 0.1–1.0 kcalÆmol )1 ÆA ˚ )4 , respectively. For the third step, the starting temperature was set to 2000 K and cooled to 0 K in 3000 steps of 5 fs. In this step, the force constants for NOE, dihedral angles, and van der Waals forces were 150 kcalÆmol )1 ÆA ˚ )2 , 200 kcalÆmol )1 Ærad )2 and 1.0–4.0 kcalÆmol )1 ÆA ˚ )4 , respectively. In the final step, 200 steps of conjugate gradient minimization were performed. In this step, the force constants for NOE, dihedral angles, and van der Waals forces were 75 kcalÆmol )1 ÆA ˚ )2 , 400 kcalÆ mol )1 Ærad )2 and 1.0 kcalÆmol )1 ÆA ˚ )4 , respectively. In all steps, an energy term for a conformational database was included with a force constant of 1.0 [13]. The / and u dihedral angles were set to the values derived from 1 Hand 13 C chemical shifts using the program TALOS [14]. For the structure calculation, errors of ± 30° were used for the dihedral restraints. NOEs were classified as strong (<2.7 A ˚ ), medium (<3.3 A ˚ or <3.5 A ˚ for amide protons), weak (<5.0 A ˚ )andveryweak(<6.0A ˚ ). A correction of 0.5 A ˚ was added to proton distances involving methyl groups. The electrostatic map and structural figures were generated by the molecular graphics program MOLMOL [15]. RESULTS To characterize the HIV gp120 C5 domain, a 23 residue peptide was synthesized, corresponding to residues 489–511 of HIV-1 strain HXB2 (Fig. 1A). The synthetic C5 peptide was found to be soluble under all experimental conditions that were examined. As shown in Fig. 1B, the CD spectra of C5 in aqueous solution suggest the absence of regular secondary structure. The absence of secondary structure was found to be independent of temperature, pH and salt concentration (data not shown). In contrast, the CD spectra of C5 in trifluoroethanol/H 2 O or hexafluoroisopropanol/ H 2 O mixtures suggest the presence of helix. In trifluoro- ethanol/H 2 O mixtures, the maximal amount of helical structure is observed in 40% trifluoroethanol (Fig. 1B). A similar amount of helical structure is observed in 60% hexafluoroisopropanol (v/v, data not shown). In the presence of 40% trifluoroethanol, the molar ellipticity at 222 nm suggested that the helical content of C5 is  40% (i.e.  10 residues). Importantly, the use of trifluoroethanol or hexafluoroisopropanol poses no technical difficulties for high-resolution NMR studies due to the availability of deuterated compounds. We found that the NMR spectra of HIV C5 in 40% trifluoroethanol were superior to those of C5 in 60% hexafluoroisopropanol. Consequently, we chose to further characterize the 40% trifluoroethanol sample. The 1 H resonances of C5 in 40% trifluoroethanol were assigned by a series of TOCSY and NOESY experiments at different temperatures (285–305 K). The 13 C a and 13 C b were assigned by 13 C-edited HSQC experiments on a natural abundance sample. Due to the small degree of spectral dispersion, it was critical to make assignments at various temperatures. The NOESY spectra of C5 at lower temperatures are superior to those at higher temperatures, due to an increased correlation time (and thus an increased NOE). An example of the NOESY spectrum at 285 K is given in Fig. 2A. The Fig. 1. (A) Amino acid sequence and (B) CD spectra of the HIV gp120 C5 peptide. (A) Numbering corresponds to that of HIV-1 strain HXB2 [16]. (B) Dotted lines represent the spectrum in aqueous solution (25 l M C5, 100 m M Tris/HCl/pH 8.0); solid lines represent the spec- trum in the presence of an organic solvent (25 l M C5, 40% trifluoro- ethanol, 100 m M Tris/HCl/pH 8.0). Ó FEBS 2002 HIV Gp120 C5 solution structure (Eur. J. Biochem. 269) 4861 presence of sequential H N -H N crosspeaks is characteristic of helical regions. Indeed, the crosspeak pattern shown in Fig. 2A suggests the presence of a continuous helix between residues 499 and 509. The secondary structure was further characterized by 1 H a and 13 C a chemical shifts. As shown by Fig. 2B, the 1 H a and 13 C a chemical shifts of residues 499– 509 suggest the presence of a helix. The prolines at positions 493 and 498 are apparently in the trans conformation, based on the presence of strong a/d NOEs. Moreover, the absence of Ôpeak splittingÕ in the TOCSY spectra of C5 indicates that the trans isomer is the major isomer. Interestingly, the H N and H a line-widths of the region encompassing the prolines and preceding the helix (residues 489–498) are similar to those of the helix. Accordingly, the relaxation properties of this region indicate that it is not appreciably more mobile than the helix. The tertiary structure of HIV gp120 C5 was determined by an iterative method of torsion angle molecular dynam- ics followed by simulated annealing [18]. A summary of theNOEsusedinthecalculationispresentedinFig.3. From Fig. 3, it is apparent that there are numerous long- range contacts between residues in the N- and C-terminal regions of C5 (e.g. V489 to V505 and I491 to V506). An ensemble of 20 low energy structures is presented in Fig. 4A and the structural statistics of the ensemble are summarized in Table 1. The overall backbone RMSD of the ensemble to the mean is 0.53 A ˚ and the overall heavy atom RMSD (i.e. nonhydrogen) of the ensemble to the mean is 1.25 A ˚ . In Fig. 4B, the backbone RMSD of the C5 ensemble is plotted as a function of residue number to illustrate the relative uncertainties of the backbone. Figure 5A shows the ribbon diagram of the energy minimized mean structure of HIV gp120 C5. The structural motif of C5 is that of a turn-helix with the amino and carboxy termini in close proximity. As shown by Fig. 5A, the amino terminal end of the C5 domain would be attached to remaining domains of gp120 in the native structure. The carboxy terminal end of the C5 domain would be attached to the amino terminus of gp41 in the unprocessed form (termed gp160). Note that the last four residues of the C5 domain form the furin/PC7 recognition site (amino acid sequence ¼ REKR) and processing at this site by furin liberates the gp41 amino terminus, which contains the fusion peptide. In Fig. 5B, the electrostatic map of C5 is presented. Pertinent features include a relatively uncharged hydrophobic ÔelbowÕ in the turn (residues 493–499, amino acid sequence ¼ PLGVAPT), the presence of two negative charges (E492 and E509), and the presence of numerous positive charges along the exterior of the helix (e.g. K490, K500, R504, R508 and R511). Interestingly, the hydrophobic ÔelbowÕ is absolutely conserved among Fig. 2. (A) NOESY spectrum and (B) secondary chemical shifts of 1 H a and 13 C a of HIV gp120 C5 in 40% trifluoroethanol. (A) Experimental conditions were 1 m M C5, 40% trifluoroethanol, 50 m M PO 4 (pH 6.0), 5% D 2 O at 285 K with a mixing time of 150 ms. (B) Secondary chemical shifts of 1 H a (a) and 13 C a (b) of the HIV gp120 C5 in 40% trifluoroethanol. Experimental conditions were 1 m M C5, 40% trifluoroethanol, 50 m M PO 4 (pH 6.0) at 300 K. Random coil values of a protein in trifluoroethanol were taken from Merutka et al.[17]. Fig. 3. (A) Summary of the number and type of NOEs observed for each residue and (B) NOE contact map for HIV gp120 C5 in 40% tri- fluoroethanol. (A) The NOE classification is given in Table 1. 4862 L. Guilhaudis et al. (Eur. J. Biochem. 269) Ó FEBS 2002 HIV species [16]. Moreover, the charged residues at positions 490, 500, 504, 508, 509, 510 and 511 are highly conserved among HIV species [16]. The next logical step is to assay the ability of HIV C5 to bind to HIV gp41. As discussed in the introduction, the loop region of gp41 is the most likely binding site for C5. The gp41 loop region contains four highly conserved tryptophan residues (W85, W99, W103, and W112), which could possibly serve as fluorescent reporters of a gp41/C5 interaction. Importantly, the C5 domain employed in the present study does not possess any tryptophan residues. Therefore, we performed fluorescence titrations of the HIV gp41 ectodomain in the presence of trifluoroethanol as cosolvent. As shown by Fig. 6A, the tryptophan fluores- cence of HIV gp41 increases as HIV C5 is added. The increased tryptophan fluorescence at 340 nm suggests that at least some of the gp41 tryptophan sidechains are being shielded from solvent by the addition of HIV C5 (Fig. 6B). Based on the fluorescence changes shown by Fig. 6A, HIV C5 forms a 1 : 1 complex with the HIV gp41 ectodomain with an apparent K d of  1.0 l M . Interestingly, titrations performed in the absence of cosolvent did not exhibit measurable changes in tryptophan fluorescence, which suggests that the turn-helix structural motif present of C5 is critical to complex formation. DISCUSSION The CD results suggest that HIV gp120 C5 is in a random coil conformation in aqueous solutions, which is in agree- ment with previous studies of similar C5 peptides [20,21]. In contrast the presence of helix is clearly indicated by CD Fig. 4. (A) Ensemble of 20 low-energy struc- tures of HIV gp120 C5 in 40% trifluoroethanol and (B) plot of the backbone RMSD of the final C5 ensemble as a function of residue number. (A) For clarity the amino and carboxy termini have been denoted by their residue number. (B) For this calculation, the backbone was defined as the N, C a and C atoms. Table 1. Structural statistics for the final 20 simulated annealing structures (ÆSAæ) and the minimized mean structure (ÆSAæ r ). None of the structures exhibited distance violations greater than 0.5 A ˚ or dihedral violations greater than 6°. ÆSAæÆSAæ r RMS deviations from distance restraints all (190) 0.062 ± 0.003 0.075 intraresidue (16) 0.042 ± 0.012 0.032 sequential (| i-j | ¼ 1) (93) 0.081 ± 0.004 0.095 short range (1 < | i- j| < 5) (63) 0.030 ± 0.004 0.054 longrange (|i–j | > 5) (18) 0.030 ± 0.015 0.001 RMS deviations from dihedral restraints (°) (47) a 0.99 ± 0.21 1.46 Deviations from idealized covalent geometry bonds (A ˚ ) 0.006 ± 0.001 0.004 angles (°) 0.61 ± 0.02 2.52 impropers (°) 0.48 ± 0.04 2.74 Measures of structure quality b % residues in most favorable regions 84.2 ± 4.1 88.9 % residues in additionally allowed regions 15.8 ± 4.1 11.1 % residues in generously allowed regions 0.0 ± 0.0 0.0 % residues in disallowed regions 0.0 ± 0.0 0.0 number of bad contacts/100 residues 2.2 ± 3.3 8.7 Coordinate precision c backbone (A ˚ ) 0.53 ± 0.20 heavy (A ˚ ) 1.25 ± 0.72 helix backbone (A ˚ ) 0.39 ± 0.06 helix heavy (A ˚ ) 1.22 ± 0.65 a Torsion angle restraints consisted of 21 /,21w,3v 1 , and 2 v 2 . b The overall quality of the structure was assessed by the program PROCHECK 3.5 [19]. c Defined as average RMS difference between the final 20 simulated annealing structures and the mean coordinates. Ó FEBS 2002 HIV Gp120 C5 solution structure (Eur. J. Biochem. 269) 4863 when C5 is in the presence of the cosolvents trifluoroethanol or hexafluoroisopropanol. The high-resolution solution structure of HIV gp120 C5 in 40% trifluoroethanol was determined by NMR spectroscopy. The presence of a carboxy terminal helix (residues 499–510) was consistent with the CD experiments, which predicted a helical content of  10 residues. We recognize that from a physiological standpoint the structural and dynamic characterization of C5 in the presence of a cosolvent is less desirable than under strictly aqueous conditions. However, it is increasingly apparent that trifluoroethanol only stabilizes helical struc- tures in protein domains with inherent propensity for helix formation [22–24]. Moreover, trifluoroethanol/aqueous mixtures are thought to mimic the hydrophobic environ- ment of protein interiors [22,25–28]. In the present case, C5 in isolation is missing the other domains of gp120 as well as its putative hydrophobic interaction site on gp41 [10,18,29]. Consequently, the trifluoroethanol/aqueous mixture may very well mimic the hydrophobic environment of the gp120 C5 domain in vivo. Interestingly, the presence of helix at residues 500–511 was predicted for HIV gp120 by various methods of secondary structure prediction [30] and thus the C5 propensity for helical structure is not surprising. Importantly, we have shown by fluorescence titration that HIV C5 directly interacts with the HIV gp41 ectodomain in the presence of cosolvent, suggesting that the C5 structure presented herein is physiologically relevant. The HIV gp120 C5 structure provides insight into the functional properties of the C5 domain. First, the C5 structure encompasses the site of envelope protein matur- ation. Specifically, the furin/PC7 recognition site REKR is present at the C terminal helix of the present C5 construct (Fig. 5). Mutations at this cleavage site inhibit fusion and hence HIV infection [5] and thus this site is a valid target for drug therapies against HIV. However, it is important to note that the presence of helix in this site is unknown in the precursor envelope protein and further study is clearly warranted. Second, the C5 structure provides insight into the structure and action of a peptide that displays antiviral activity against HIV-1 and HIV-2 [31]. This peptide, which is called CLIV, corresponds to four linked copies of residues 499–511. The peptide has been shown to disrupt the fusion process after binding of CD4, which indicates the CLIV peptide does not function by disrupting gp120 maturation [32]. Third, the C5 structure can be used to interpret the results of mutagenesis studies of the gp120–gp41 interaction. For example substitutions of C5 residues inhibit association of gp120 with gp41 [8]. Moreover, substitution of I491, P493, G495, A497, P498, T499 or A501 by cysteines resulted in intermolecular crosslinks to non-native cysteines intro- duced into the gp41 loop region [10]. Together the mutagenesis, structural studies, and fluorescence titration experiments [7,18,29,33] (and the present work) provide compelling evidence that the hydrophobic loop of gp41 interacts with the hydrophobic turn region of C5. The HIV gp120 C5 structure can be used to model the gp41/gp120 complex. In Fig. 7, we present a model of the gp41 and gp120 components, which were determined by Fig. 5. (A) Ribbon diagram of the minimized mean structure of HIV gp120 C5 in 40% trifluoroethanol and (B) electrostatic map of the minimized mean structure of HIV gp120 C5. (A) The location of the furin/PC7 site is shown. The locations of the missing gp120 domains and gp41 are denoted. (B) In this orientation, the conserved Ôhydro- phobic elbowÕ of C5 is located at the upper left corner of the figure. Fig. 6. (A) Fluorescence titration of HIV gp41 ectodomain by HIV gp120 C5 and (B) location of tryptophan residues in HIV gp41 ectodomain (coordinates taken from reference [33]). (A) Experimental conditions were 20 l M HIV gp41 ectodomain in 50 m M sodium formate (pH 3.0), 40% trifluoroethanol at 25 °C. Titrations of a buffer blank by HIV C5 exhibited no change in the 340 nm fluorescence. The solid line represents the least squaresfitofthedatawithK d ¼ 1.0 ± 0.8 l M and n ¼ 0.98 ± 0.13. (B) The exposed tryptophans found in the gp41 loop are labeled. The unlabeled tryptophans, which include W60, W117 and W120, are found in the protein interior and not exposed to solvent [18,33]. 4864 L. Guilhaudis et al. (Eur. J. Biochem. 269) Ó FEBS 2002 NMR spectroscopy and X-ray crystallography [7,18,33] and the present work. As discussed above, knowledge of the gp41/gp120 structure is limited by the absence of structural information for the C1 and C5 domains of gp120, which are indicated to interact directly with gp41. Accordingly, we have modeled the complex by first placing C5 in the context of the gp120 core structure determined by X-ray crystallography [7]. The first three residues of the C5 domain used in the present study (amino acids VKI) are observed in the gp120 X-ray structure; thus, the C5 domain can be overlaid with the C terminus of the gp120 X-ray structure (the RMSD of the NMR VKI backbone to the X-ray VKI backbone is 0.4 A ˚ ). The extended conformation of these residues in solution is consistent with being part of a b sheet observed in the gp120 X-ray structure (Fig. 7). The location of C5 with respect to the gp120 core is also consistent with immunological studies that suggest that the C5 domain interacts directly with the C1 and C2 domains [34]. In the second step, we have used mutagenesis studies of gp41 and gp120 to reveal residues that are in close contact in the gp41/gp120 complex. As discussed above, a number of non-native cysteines placed within the gp41 loop form intermolecular crosslinks to non-native cysteines placed in the C5 domain [10]. Interestingly, the most reactive non- native cysteines occur at position 94 of gp41 (denoted by the blue sphere in Fig. 7) and at position 501 of gp120 (denoted by the red sphere in Fig. 7), which can be placed in close proximity in the present model. The C5/gp41 interactiondepictedinFig.7wouldbeexpectedto partially shield W85, W99 and W103 of gp41 from solvent (cf. Figure 6B), which is consistent with the fluorescence titration experiments discussed above. Finally, the placement of the viral and target cell membranes is facilitated by the structures of gp41 and CD4 receptor. Specifically, gp41 contains a transmembrane domain that is C-terminal to the ectodomain and thus located at the ÔleftÕ of Fig. 7. The CD4 receptor contains a C-terminal membrane anchor and thus is located at the ÔrightÕ of Fig. 7. In addition, the location of the target cell chemokine receptor can be inferred by the binding site of the anitbody 17b, which blocks binding of gp120 to the chemokine receptor [7]. Note that the model is consistent with immunological studies that indicate that residues 501–511 are well exposed and immunodominant in the native complex [35], consistent with its being charged and thus hydrophilic (Fig. 5B). Furthermore, antibody acces- sibility studies have suggested that the amino terminal region of the gp120 C5 domain (residues 489–500) is not exposed in the gp120/gp41 complex but exposed in the monomeric form when gp120 dissociates from gp41 [36], consistent with the notion that the hydrophobic ÔelbowÕ of C5 (Fig. 5B) interacts with the hydrophobic loop region of gp41 [18,33]. In total, the resulting model is consistent with a large number of biochemical and structural studies. It is important to note that the gp41 helical regions are generally thought to undergo large-scale structural chan- ges during the binding of gp120 to CD4 and the subsequent fusion event (cf. [37] but see [18,38] for an alternative viewpoint). Accordingly, the known structures for gp41 [18,33] may not represent the gp41 conformation bound to gp120. However, we are unaware of any evidence for structural changes in the gp41 loop connect- ing the helical regions. In conclusion, we feel that the HIV gp120 C5 structure is relevanttobiomedicalresearchonAIDSingeneral and vaccine development in particular. Firstly, a full understanding of gp120 function requires high-resolution structural and dynamic information for all regions of gp120. The present work presents the first high-resolution struc- tural information for C5, albeit in 40% trifluoroethanol. Secondly, based on the immunological studies [35,36] and the model of the gp41/gp120 complex presented in Fig. 7, the C5 domain is expected to be partially accessible to solvent and thus accessible to the immune system. Conse- quently, the gp120 C5 domain, which is highly conserved among HIV and SIV isolates, may be an attractive target for vaccine development. Finally, as discussed above, the C5 domain represents the most likely site for gp41 binding. Disruption of this binding is expected to inhibit HIV infection and thus the C5 domain is an attractive target for structure-based drug therapies. Fig. 7. Model for the gp41–gp120 complex upon association with the target cell receptors. The HIV gp41 ectodomain is shown in blue and is a taken from the model structure [33], which is based on the NMR structure of the SIV gp41 ectodomain [18]. The gp120 C5 structureisshowninredandistakenfromthe present work. The structure of the core gp120- CD4 receptor-antibody 17b complex is taken from Kwong et al. [7]. The gp120 core is shown in green; the CD4 receptor is shown in cyan, and the antibody 17b is shown in violet. The sites of the most reactive non-native cysteines in gp41 (at position 94) and gp120 (at position 501) are shown as spheres and are taken from Binley et al.[10]. Ó FEBS 2002 HIV Gp120 C5 solution structure (Eur. J. Biochem. 269) 4865 ACKNOWLEDGEMENTS The coordinates of the minimized mean structure of the HIV gp120 C5 have been deposited in the Protein Data Bank (ID 1MEQ). This work was supported by RO1 grant AI47674. REFERENCES 1. White, J.M. (1992) Membrane fusion. Science 258, 917–924. 2. Dimitrov, D.S. (1997) How do viruses enter cells? The HIV coreceptors teach us a lesson in complexity. Cell 91, 721–730. 3. Major, E., Rausch, D., Marrak, C. & Clifford, D. (2000) HIV- associated dementia. Science 288, 440–441. 4. Hallenberger, S., Moulard, M., Klenk, H. & Garten, W. (1997) The role of eukaryotic subtilisin-like endoproteases for the acti- vation of human immunodeficiency virus glycoproteins in natural host cells. J. Virol. 71, 1036–1045. 5. Bosh, V. & Pawlita, M. (1990) Mutational analysis of the human immunodeficiency virus type 1 env gene product proteolytic cleavage site. J. Virol. 64, 2337–2344. 6. Hallenberger, S., Bosch, V., Angliker, H., Shaw, H., Klenk, H. & Garten, W. (1992) Inhibition of furin-mediated cleavage activation of HIV-1 glycoprotein-gp160. Nature 360, 358–361. 7. Kwong,P.D.,Wyatt,R.,Robinson,T.,Sweet,R.W.,Sodroski,J. & Hendrickson, W. (1998) Structure of an HIV gp120 envelope glycoprotein in complex with the CD4 receptor and a neutralizing antibody. Nature 393, 648–659. 8. Helseth, E., Olshevsky, U., Gabuzda, D., Ardman, B., Haseltine, W. & Sodroski, J. (1990) Changes in the transmembrane region of the human immunodeficiency virus type -1 gp41 envelope glyco- protein affect membrane fusion. J. Virol. 64, 6314–6318. 9. Ivey-Hoyle, M., Clark, R.K. & Rosenberg, M. (1991) The NH2- terminal 31 amino acids of human immunodeficiency virus type-1 envelope protein gp120 contain a potential gp41 contact site. J. Virol. 65, 2682–2685. 10. Binley,J.,Sanders,R.,Clas,B.,Schuelke,N.,Master,A.,Guo, Y., Kajumo, T., Anselma, D., Maddon, P., Olson, W. & Moore, J. (2000) A recombinant human immunodeficiency virus type 1 envelope glycoprotein complex stabilized by an intermolecular disulfide bond between the gp120 and gp41 subunits is an anti- genic mimic of the trimeric virion-associated structure. J. Virol. 74, 627–643. 11. Wingfield, P., Stahl, S., Kaufman, J., Zlotnick, A., Hyde, C., Gronenborn, A. & Clore, G. (1997) The extracellular domain of immunodeficiency virus gp41 protein: expression in Escherichia coli, purification, and crystallization. Prot. Sci. 6, 1653–1660. 12. Brunger, A., Adams, P., Clore, G., DeLano, W., Gros, P., Grosse-Kunstleve, R., Jiang, J S., Kuszewski, J., Nilges, M., Pannu, N., Read, R., Rice, L., Simonson, T. & Warren, G. (1998) Crystallography and NMR System: a new software suite for macromolecular structure determination. Acta Cryst. D54, 905–921. 13. Kuszewski, J., Gronenborn, A. & Clore, G. (1997) Improvements and extensions in the conformational database potential of NMR and x-ray structures of proteins and nucleic acids. J. Magn. Reson. 125, 171–177. 14. Cornilescu, G., Delaglio, F. & Bax, A. (1999) Protein backbone angle restraints from searching a database for chemical shift and sequence homology. J. Biomol. NMR 13, 289–302. 15. Koradi,R.,Billeter,M.&Wuthrich,K.(1996)MOLMOL:a program for display and analysis of macromolecular structures. J. Mol. Graphics 14, 52–55. 16. Douglas, N.W., Munro, G.H. & Daniels, R.S. (1997) HIV/SIV glycoproteins: structure-function relationships. J. Mol. Biol. 273, 122–149. 17. Merutka, G., Dyson, H. & Wright, P. (1995) Random coil H-1 chemical-shifts obtained as a function of temperature and trifluoroethanol concentration for the peptide series GGXGG. J. Biomolec. NMR 5, 14–24. 18. Caffrey, M., Cai, M., Kaufman, J., Stahl, S.J., Wingfield, P.T., Gronenborn, A.M. & Clore, G.M. (1998) Solution structure of the 44 kDa ectodomain of SIV gp41. EMBO J. 17, 4572–4584. 19. Laskowski,R.,MacArthur,M.,Moss,D.&Thorton,J.(1993) PROCHECK – a program to check the stereochemical quality of protein structures. J. Appl. Cryst. 26, 283–291. 20. Moulard, M., Challoin, L., Canarelli, S., Mabroluk, K. & Darbon, H. (1998) Retroviral envelope glycoprotein processing: Structural investigation of the cleavage site. Biochemistry 37, 4510–4517. 21. Ferrer, M., Sullivan, B., Godbout, K., Burke, E., Stump, H., Godoy, J., Golden, A., Profy, A. & van Schravendijk, M. (1999) Structural and functional characterization of an epitope in the conserved C-terminal region of HIV-1 gp120. J. Peptide Res. 54, 32–42. 22. Segawa, S., Fukuno, T., Fujiwara, K. & Noda, Y. (1991) Local structures in unfolded lysozyme and correlation with secondary structures in the native conformation – helix-forming or helix- breaking propensity of peptide segments. Biopolymers 31,497– 509. 23. Sonnichsen, F., Van Eyk, J., Hodges, R. & Sykes, B. (1992) Effect of trifluoroethanol on protein secondary structure – an NMR and CD study using a synthetic actin peptide. Biochemistry 31, 8790– 8798. 24. Hamada,D.,Kuroda,Y.,Tanaka,T.&Goto,Y.(1995)High helical propensity of the peptide-fragments derived from beta- lactoglobulin, a predominantly beta-sheet protein. J. Mol. Biol. 254, 737–746. 25. Dyson, H., Sayre, J., Merutka, G., Shin, H., Lerner, R. & Wright, P. (1992) Folding of peptide-fragments comprising the complete sequence of proteins – models for initiation of protein folding. 2. Plastocyanin. J. Mol. Biol. 226, 819–835. 26. Jasanoff, A. & Fersht, A. (1994) Quantitative-determination of helical propensities from trifluoroethanol titration curves. Biochemistry 33, 2129–2135. 27. Waterhous, D. & Johnson, W. (1994) Importance of environment in determining secondary structure in proteins. Biochemistry 33, 2121–2128. 28. Hirota, N., Mizuno, K. & Goto, Y. (1998) Group additive con- tributions to the alcohol-induced alpha-helix formation of melittin: Implication for the mechanism of the alcohol effects on proteins. J. Mol. Biol. 275, 365–378. 29. Caffrey, M., Cai, M., Kaufman, J., Stahl, S.J., Wingfield, P.T., Gronenborn, A.M. & Clore, G.M. (1997) Determination of the secondary structure and global topology of the 44 kDa ecto- domain of gp41 of the simian immunodeficiency virus by multi- dimensional nuclear magnetic resonance spectroscopy. J. Mol. Biol. 271, 819–826. 30. Hansen, J., Lund, O., Nielsen, J., Brunak, S. & Hansen, J. (1996) Prediction of the secondary structure of HIV-1 gp120. Proteins 25, 1–11. 31. Sabatier, J., Moulard, M., Rochat, H., Van Rietschoten, J. & Fenouillet, E. (1996) Anti-HIV activity of multibranched peptide constructs derived either from the cleavage sequence or from the transmembrane domain (gp41) of the human immunodeficiency virus type 1 envelope. Virology 223, 406–408. 32. Barbouche, R., Decroly, E., Kieny, M. & Fenouillet, E. (2000) An anti-human immunodeficiency virus multiple antigen peptide encompassing the cleavage region of the env precursor interferes with membrane fusion at a post-CD4 binding step. Virology 273, 169–177. 33. Caffrey, M. (2001) Model for the structure of the HIV gp41 ectodomain: insight into the intermolecular interactions of the gp41 loop. Biochim. Biophys. Acta 1536, 116–122. 4866 L. Guilhaudis et al. (Eur. J. Biochem. 269) Ó FEBS 2002 34. Moore, J., Willey, R., Lewis, G., Robinson, J. & Sodroski, J. (1994) Immunological evidence for interactions between the first, second and fifth conserved domains of the gp120 surface glyco- protein of human immunodeficiency virus type 1. J. Virol. 68, 6836–6847. 35. Palker,T.,Matthews,T.,Clark,M.,Cianciolo,G.,Randall,R., Langlois, A., White, G., Safai, B., Snyderman, R., Bolognesi, D. & Haynes, B. (1987) A conserved region at the cooh terminus of human-immunodeficiency-virus gp120 envelope protein contains an immunodominant epitope. Proc. Natl. Acad. Sci. USA 84, 2479–2483. 36. Moore, J., Sattentau, Q., Wyatt, R. & Sodroski, J. (1994) Probing the structure of the human immunodeficiency virus surface gly- coprotein gp120 with a panel of monoclonal antibodies. J. Virol. 68, 469–484. 37. Chan, D. & Kim, P. (1998) HIV entry and its inhibition. Cell 93, 681–684. 38. Caffrey, M., Kaufman, J., Stahl, S., Wingfield, P., Gronenborn, A. & Clore, G. (1999) Monomer-trimer equilibrium of the ecto- domain of SIV gp41: insight into the mechanism of peptide inhibition of HIV infection. Prot. Sci. 8, 1904–1907. Ó FEBS 2002 HIV Gp120 C5 solution structure (Eur. J. Biochem. 269) 4867 . interaction, the solution structure of the C5 domain of HIV gp120 has been determined by NMR spectroscopy. EXPERIMENTAL PROCEDURES The HIV gp120 C5 domain was. in the gp120 X-ray structure; thus, the C5 domain can be overlaid with the C terminus of the gp120 X-ray structure (the RMSD of the NMR VKI backbone to the