呼吸机
管路Current Protein and Peptide Science, 2005, 6, 413-4224131389-2037/05 $50.00+.00
© 2005 Bentham Science Publishers Ltd.
HIV-1 gp120 V3 Loop for Structure-Based Drug Design
Suzanne Sirois 1,2,4,*, Tobias Sing 3 and Kuo-Chen Chou 4
1
Université du Québec à Montréal (UQAM), Chemistry Department, C.P. 8888 Succursale Centre-Ville, Montréal,Québec, Canada, H3C 3P8, 2Immune Deficiency Treatment Centre (IDTC), Montreal General Hospital, McGill Univer-sity Health Centre, 1650 Cedar Avenue, Montréal, Québec, H3G 1A4, Canada, 3Max-Planck-Institut für Informatik,Stuhlsatzenhausweg 85, 66123 Saarbrücken, Germany, 4Gordon Life Science Institute, 13784 Torrey Del Mar Drive,San Diego, California 92130, USA
Abstract: HIV-1 cell entry is mediated by sequential interactions of the envelope protein gp120 with the receptor CD4and a coreceptor, usually CCR5 or CXCR4, depending on the individual virion. Considera
ble efforts on exploiting the HIV coreceptors as drug targets have led to the new class of coreceptor antagonists. While these antiretroviral drugs aim at preventing virus/coreceptor interaction by binding to host proteins, neutralizing antibodies directed against the core-ceptor-binding sites on gp120 have attracted attention as possible vaccine candidates. However, both approaches are com-plicated by the multiple protective mechanisms of gp120 which allow for rapid escape from selective pressures exerted by drugs or antibodies. Thus, advances in rational drug and vaccine design rely heavily on improved insights into the relation between genotype and phenotype, the evolution of coreceptor usage, and, ultimately the structural biology of coreceptor usage and inhibition. The third variable (V3) loop of gp120, crucially involved in all these aspects, will be a major focus of this review.
Keywords : HIV-1, gp120, V3 loop, coreceptor, CCR5 , CXCR4, structure-based drug design.1. INTRODUCTION
More than ten years after the discovery that HIV-1 could only infect cells with the CD4 receptor [1], it was found that,unlike other retroviruses, HIV was also dependent on a core-ceptor to enter a host cell [2-9] and that the chemokine re-ceptors CCR5 [10] and CXCR4 [11] were the preferred HIV coreceptors in vivo . First, Cocchi [12] observed that three chemokines, MIP1-α, MIP1-β and RANTES (nowadays also called CCL3 to CCL5) were potent repressors of strains with in vitro tropi
sm for macrophages. Next, Feng et al. [13], dis-covered that strains infecting T-cells in vitro needed a chemokine receptor which we call CXCR4 today, as a core-ceptor, in addition to CD4. Due to these two observations, it was widely hypothesized in 1996 that CCL3 to CCL5 were the natural ligands of a receptor which is needed by macro-phage-tropic strains as a coreceptor. Indeed, in the same year, this receptor, today called CCR5, was identified by several groups [2-4,6,14]. At the very moment of their dis-covery, the idea of inhibiting the HIV coreceptors to prevent viral entry was born. Moreover, the discovery of differential coreceptor usage also replaced previous phenotype classifi-cation systems based on cell tropism, replication rate in pe-ripheral blood mononuclear cells (PBMCs), or the cytopa-thology in MT-2 cells [15]. Based on these “old” classifica-tion schemes, which are highly correlated but not identical to coreceptor usage, the third variable region of gp120 had al-ready been identified as a major determinant of phenotype [16]. The rough region being identified, interest grew in
*Address correspondence to this author at the Université du Québec àMontréal (UQAM), Chemistry Department, C.P. 8888 Succursale Centre-Ville, Montréal, Québec, Canada, H3C 3P8;E-mail: suzanne.sirois@gmail
determining the relation between genotype and phenotype more exactly. Two pioneering papers app
eared in 1992, im-plicating a net V3 charge of at least 5 [17] or the presence of a basic residue at V3 positions 11 or 25 [18] with usage of the CXCR4 receptor. At this time it was already known that a phenotype switch was associated with progression to AIDS. Thus, a direct link between the sequence evolution of V3 and disease progression was established. Reflecting the growing interest in this link, the first article on V3 evolution appeared in 1992 [18-19]. Long before the V3 region was implicated as a phenotype determinant, it had already been recognized as a major target for neutralizing antibodies. In an early bioinformatics paper by Modrow et al. [20], a re-gion within V3 was predicted as an epitope, along with other regions. This prediction was confirmed in the following year [21], and in 1989, the V3 region was termed the “principal neutralizing determinant” (PND), because its deletion stopped the activity of neutralizing antibodies. Thus, in 1996,all questions were there which have occupied HIV corecep-tor research until today, and will continue to do so. Conse-quently, our short historical survey will stop here (good overviews of the history of HIV research can be found in Tang [22].
2. THE ENVELOPE PROTEIN gp120 AND ITS V3LOOP
The envelope protein gp120 initiates the process of cell entry by interacting with the main receptor CD4 and one of the chemokine receptors CCR5 or CXCR4. It is derived from the polyprotein gp160,
which also contains the transmem-brane protein gp41. This polyprotein is encoded by the env gene, present in all retroviruses. To denote specific regions within the sequence coding for gp120, either a region name or a position number can be used. Inconsistent and inaccu-
414Current Protein and Peptide Science, 2005, Vol. 6, No. 5Sirois et al.机器人拉车
rate numbering, caused by frequent insertions and deletions, has been a serious problem in the literature on gp120. Trying to unify the language in the field Berger et al. [23] have pro-posed to number sequence positions relative to the reference strain HXB2 (GenBank accession K03455). The position of insertions relative to HXB2 is denoted with the aid of post-fixed letters. For example, two insertions between HXB2 positions 465 and 466 would be referred to as 465a and 465b, respectively. While it is strongly advised to stick to this scheme when denoting positions relative to the whole genome or to gp120, position numbering within the third variable region of gp120 is frequently relative to a subtype B consensus strain, which has one insertion and two deletions relative to HXB2. Following these established traditions, position numbering will be relative to HXB2 in this review, except for the V3 region, where the subtype B consensus will be used as a reference.
2.1. Coding Sequence of gp120
In HXB2, the envelope protein is encoded by 483 amino acids within an env gene consisting of 856 residues, pre-ceded by a signal peptide of length 28, and followed by the transmembrane protein gp41. Though the leader peptide se-quence is cleaved concurrently with translation, it is included in the proposed numbering scheme for gp120. Thus, in this scheme, the actual gp120 positions are referred to by the numbers 29 to 511. The gene has been divided into five vari-able (V1-V5) and five constant (C1-C5) regions, although the constant regions exhibit a substantial variety as well. Most of this review will focus on the third variable region, due to its crucial relevance for coreceptor usage. This region, occupying the gp120 positions 296 to 331 relative to HXB2, is typically 34 to 36 residues in length. Table 1 shows the V3 regions of HXB2 and of the subtype B consensus sequence provided by the HIV Sequence Database at Los Alamos, (v) which is frequently used as a V3 reference. An important exploratory step that should precede any more specific analysis is to investigate the overall se-quence variability and site covariation. The literature on V3 sequence variability is abundant, mostly specific for a certain geographic region or risk group. Since most of these studies have focused on finding sequence signals predictive of genotype, they will be discussed below.
Two classic analyses focused on site covariation within the V3 region. Korber et al. [24], determined
the mutual in-formation between all pairs of positions in an alignment of 308 V3 sequences, and reported the strongest covariation among the pairs 24/25, 13/25, 13/19, 13/24, 20/25, 11/25, and 11/13. These results were later confirmed and extended by Bickel et al. [25], on a larger data set and using alterna-tive measures of covariation. Though covariation analysis can yield insights into the dependencies between different sequence positions, additional analyses are needed to deter-mine their evolutionary or phenotypic relevance. For exam-ple, while the reported covariation between 11/25, with a particularly strong correlation between 11S and 25E/D is a highly phenotype-associated dual of the “charge rule” [17] other covariation patterns might simply reflect general func-tional constraints. Another descriptive measure of sequence variability is to analyze the pairwise distances of a set of sequences or the distances to a consensus strain.
Due to the relative abundance of data and the relevance for determining the phenotype, most studies of sequence variability and covariation have focused on the V3 region. However, since other regions than V3 have been implicated with coreceptor interaction as well, significant covariation between residues inside and outside of V3 is to be expected. Indeed, Hoffman et al. [26] report that a cluster containing HXB2 positions 190 to 200 (C3 region), and position 440 (C4) is linked to changes in V3. In a recent study focusing not on covariation but on phenotype-associated changes p
art of these observations were confirmed, particularly a strong association between HXB2 position 440 and V3 positions 11 and 25 [27-28]. Clearly, the major obstacle for a full-length covariation analysis of gp120 is lack of data.
2.2. gp120 V3 Loop
艾叶油胶丸
The V3 region encodes a surface accessible loop formed by a disulfide bridge between two invariant cysteines at HXB2 positions 296 and 331 of gp120 (see the yellow resi-dues in Fig. 1 and Table 1 for numbering).Comparison of different isolates show that the N- and C-terminus are con-served, as well as the “crown” (or crest) of the loop, marked by a GPG motif. In contrast, the regions flanking the crown Fig. (1). 3D structure of the V3 loop obtained from NMR studies. The GPG (Gly15-Pro16-Gly17) sequence is shown at the crown of the loop in green, orange, and green, respectively. The disulfide bridge is represented by the two Cys residues in yellow (Cys1-Cys35). Position 25 is represented by both Arg (positively charged) and Glu (negatively charged) residues in red blue and red, respectively. Position 10 is represented by Lys (positively charged) in blue.
HIV-1 V3 Loop Current Protein and Peptide Science, 2005, Vol. 6, No. 5 415
show considerable variability. The GPG crest forms a beta
turn, with the flanking regions as the two strands of an anti-
parallel beta sheet [19,30,31,94,95]. V3 loop-derived pep-
tides have been found to be structurally similar to distinct
chemokines, the natural ligands of CCR5 and CXC4 [81].
增感
油墨This suggests that alternative V3 conformations are respon-
sible for selective interactions with the coreceptors. Other
混合交换structural analyses on V3 peptides have predicted a C-
terminal alpha helix [33]. As mentioned above, the majority
of difference between CCR5- and CXCR4-tropic strains lies
in the region flanking the central GPG segment, encom-
passsing residues 306-320 (V3 position 11-25).
For example, the LAI V3 crown contains an RQ insertion
between I309 and G310 and substitutions at Y316V, T3181,
E320K. In the LAI C-terminus the I322 is absent. The crown
and stem of V3 loop32 are two functionally distinct domains
[32]. The V3 stem alone mediates soluble gp120 binding to
the N-terminus of CCR5, and the V3 crown alone determines
coreceptor usage. Based on this observation, it has been hy-
pothesized that crown and stem interact with distinct CCR5
regions in order to mediate viral entry [32] and a C-terminal
α-helix [19, 30-31]. Residues 306-320 (11-25) were desig-nated V3 crown and residues 296 to 305 (1-10) as well as
321 to 330 (26-35) were designated N and C-terminal strands
of the V3 stem [32]. The majority of differences between R5
and X4-like virus lie in the crown (see Table 1). The LAI V3
crown contains an RQ insertion between I309 and G310 and卷帘门控制箱
substitutions at Y316V, T318I, E320K. In the LAI C-
terminus the I322 is absent. The crown and stem of V3
loop32 are two functionally distinct domains. The V3 stem
alone mediates soluble gp120 binding to CCR5 N-terminal
and V3 crown alone determines coreceptor usage. Based on
this consensus sequence a neural network approach predicted
an N-terminal β strand followed by a type II β-turn (GPRC),
a second β-strand, and a C-terminal helix [33]. Thus, the
GPG crown (crest) forms a β-turn and the variable regions
flanking the crest form the two strands of an anti-parallel β-
sheet[30] [94-95].
2.3. Net Charge
The consensus V3 loop sequence derived from 245 V3 loop sequences [33] corresponds to existing macro-phagetropic strains. V3 loop is composed approximately of 35 residues and has a global positive charge that can vary from +2 to +10 [32]. A single amino acid (AA) change in V3 loop can switch coreceptor usage from R5-like to X4-like virus. This switch is also associated with an increased posi-tive charge.2.4. GPG Motif
The amino acid sequence of the V3 loop is highly vari-able among different isolates, especially in the regions flanking the highly conserved GPG central part [33]. The GPG(R/K/Q) crown (or crest) is situated in the center of the neutralizing domain [34-36].Sequence changes close to the GPG motif can alter the stability of the β-sheet and/or alter the surface accessibility, thereby influencing coreceptor us-age.
2.5. Role of V3 Loop Residues in Co-Receptor Binding
Highly conserved residues are at positions alanine 328 (33), arginine at position 3 for HIV-2 and SIV as well as the two cysteine residues at postion 296 (1) and 331 (35). Site-directed mutagenesis studies revealed that V3 Arg-298 [3] has an important role in CCR5 utilization. Other residues that are critical are Lysine 10, Isoleucine 12, Arginine 18, and Phenylalanine 20. Mutational studies of CCR5 have re-ported several acidic and aromatic residues in the extracel-lular (EC) domain of CCR5 as critical for CCR5 utilization. Critical V3 residues include both basic and hydrophobic residues, compatible with the interpretation that the interac-tion between V3 and CCR5 could involve electrostatic as well as hydrophobic interactions. Residues adjacent to the as well as can have a role in CCR5 utilization [37]. Interest-ingly 12 is conserved among all known R5 viruses of type B but not among subtypes A, C, and E.
2.6. Binding Domain of V3 Loop with CCR5
The V3 loop binds to the cell surface in a conformation dependent manner and its N-terminal domain is responsible for the interaction [38]. It has been observed that V3 loop can enhance the entry of its own HIV strains. Pre-treatment of the target cells with V3 peptides followed by removal of the peptid
es also enhanced infectivity, indicating that the binding of the peptides to the target cells also plays a role in this enhancement. The V3 stem is responsible for gp120 binding to the CCR5 N-terminus. Both the V3 crown and stem are required for soluble gp120 binding to cell surface CCR5 [32]. The V3 crown interacts with residues in the EC of CCR5, most likely ECL2. The V3 crown alone is neces-sary and sufficient to direct exclusive usage of CCR5 or CXCR4. The V3 stem, despite being able to mediate specific binding to CCR5 Nt sulfopeptides, is not the main determi-nant of coreceptor usage.
Table 1.Alignment of HXB2 and Subtype B Consensus Strain (“B”). Position Numbering is Relative to the Consensus Strain
123456789101112131414a14b1516
B C T R P N N N T R K S I H I--G P HXB2C T R P N N N T R K R I R I Q R G P 17181920212223242526272829303132333435
G R A F Y T T G E I I G D I R Q A H C
G R A F V T I G K I-G N M R Q A H C
416Current Protein and Peptide Science, 2005, Vol. 6, No. 5Sirois et al.
3. gp120 V3 LOOP 3D STRUCTURES
The central GPG sequence of the V3 loop is recognized by most of HLA optimal epitopes and can serve as a basis for vaccine development. Especially, the GPG motif is found in the crown of gp120- V3 loop (see Fig. 1) which also de-termines HIV-1 coreceptor usage. HIV infects host cells via docking of its env gp120 to the CD4 receptor. Upon binding to CD4, gp120 undergoes conformational changes involving the V3 loop, which facilitate subsequent interactions with the coreceptors CCR5 and/or CXCR4. The V3 loop of gp120 interacts directly with the co-receptors CCR5 and CXCR4.
3.1. gp120 X-Ray Structure and no V3 Loop!
The only available gp120 three dimensional coordinate structures obtained from X-Ray crystallography (PDB 1G9M [39]) (see Fig. 4) have deglycosylated regions. The envelope proteins gp120’s were obtained from HXB2 and YU2 iso-lates where the V1, V2 and V3 loops have been deleted [39-40]. The study of the various type of conformations of V3 loop that are recognized by neutralizing antibodies alone with the understanding of sequence variabilities may bring forth a bette
r understanding of the molecular virological properties of the virus and their associated immune re-sponses.In 1998, a 2.5 Å structure of a gp120 core of the laboratory-adapted HXBc2 strain in complex with a frag-ment of CD4 and the antigen binding fragment of an anti-body was reported [40]. This model was later refined to 2.2Å and additionally, the gp120 core of a clinical isolate, YU2, in the same complex, was determined at 2.9 Å [39]. To date, these are the only available structures of gp120, and an es-sential basis for all structural studies of gp120 and its inter-action with the receptors. Envelope proteins are not isolated on the surface of a virion; rather, they are organized into trimeric “spikes” [41]. The gp120 structure has an overall heart-shaped appearance. Two major domains are connected by an antiparallel, four-stranded “bridging sheet”. The “in-ner” domain (with respect to the termini) consists mainly of a two-helix, two-strand bundle, while the “outer” domain forms a stacked double barrel. The V1-V4 regions form sur-face-exposed loops which are held together by disulphide bonds at their bases. The whole protein is heavily glycosy-lated; this glycan shield is thought to yield additional protec-tion from the immune system. When using the published gp120 structures for molecular modeling, one has to be aware of certain limitations. First, the structure is in complex with CD4 and is thought to adopt different conformations alone or in additional complex with a coreceptor. Second, the protein was heavily deglycosylated for crystallization. Third, and most importantly, gp120 was not only heavily truncated at the N- and C-ter
minal ends, but also its V1, V2, and V3 loops [39,40]. However, a lot of structural information on the V3 loop is available from other sources. LaRosa [33] pre-dicted the secondary structure of a V3 consensus sequence as β strand-type II β turn-β strand-α helix. For 7 out of 20 iso-lates, no helical region was predicted. Vranken [42] have analyzed the consensus sequence constructed by LaRosa [33] using 2D-NMR spectroscopy in water and in a 20% trifluo-rethanol/water solution, and confirmed the predicted secon-dary structure. Furthermore, several crystal structures of V3 peptides in complex with the antigen-binding fragments of antibodies have been solved. Still, all molecular modeling studies related to coreceptor usage will have to provide a model for the V3 loop.
3.2. V3 Loop Structures from NMR Studies
Many laboratories have carried out structural studies of V3 peptides from different HIV-1 isolates. For instance, PDB (Protein Data Bank) structure 1CE4 [42] is a confor-mational model of the consensus macrophagetropic V3 loop of the gp120 that was examined by proton 2D-NMR spec-troscopy in water and in a 20% trifluoroethanol/water solu-tion.In water, NOE data support a β-turn conformation for the central conservative GPG region and suggest partial for-mation of a helix in the C-terminal part. Upon addition of trifluoroethanol, a C-terminal helix is formed. The C-terminal helix is amphipathic and also occurs in other ex-amined strains. It could therefore be an important feature
for the functioning of the V3 loop.
3.3. V3 Loop Structures from X-Ray Studies in Complex with Neutralizing Antibodies
Crystal structures for V3 peptides in complex with neu-tralizing antibodies indicate that they recognize different types of conformations at the crown of the V3 loop. The type II β-turn predicted [33] for the GPG sequence GPG motif was observed in the 59.1 complex [43]. The unexpected one-residue shift and the type VI cis (trans) proline RGPG could possibly be attributed to the insertion of QR between posi-tions I309 and G310 (I14, G15), preceding the GPGR se-quence in the P1053 sequence (see Table 2). This insertion is a characteristic feature of the HIV-1IIIB strain and it is en-countered in approximately 10% of all HIV-1 isolates. Also, the half-turn in the QR insertion of P1053 may facilitate the observed conformational shift. In the PDB 1F58 structure, although the peptide conformations are very similar for the linear and cyclic forms, they differ from the ones seen for identical peptides bound to the Fab 59.1 neutralizing and also for a similar peptide bound to the MN-specific Fab 50.1 (see Table 2). The conformational difference in the peptide is localized around residues GPGR, which are highly con-served in different HIV-1 isolates and are predicted to adopt a type II beta turn (see Fig. 2). Type 2 β-turns are commonly Fig. (2). Nomenclature for residues in hairpin beta-turns. For the definition of beta turns and their types, the readers are referred to
[105].
HIV-1 V3 Loop Current Protein and Peptide Science, 2005, Vol. 6, No. 5 417 Table 2.X-Ray Crystal Structures and NMR Studies of Antibodies Targeting gp120 V3 Loop
Antibody/Peptide Epitope V3 pep-
tide Turn#AA Structure
ID
Method
447-52D/ V3MN type-II16X-Ray[79]
447-52D/ V3MN KRKRIHI GPGRAF YTTKN18NMR[80] 447-52D/ V3MN KRIHI -- GPGRAF YTT Inverse γNMR[81] V3MN[82,83] 0.5β/ V3IIIB KSIRI QR GPGRAF VTI P1053[84]type VI β 181B03NMR[31,45]
0.5β/ V3IIIB YNKRKKRIHI GPGRAF YTTKNI
IGC
RP13524NMR[85] 0.5β/ V3IIIB NMR[86] 0.5β/ V3IIIB RKSI-RI QR GPGRAF VT RP135a NMR[84] 0.5β/ V3IIIB NMR [87] 0.5β/ V3IIIB type VI βNMR [88] 50.1/ V3IIIB KRIHI-- GPG type II β
59.1/ V3IIIB KRIHI -- GPGRAF YT Aib142
RP142
type II-type I β[89]
58.2/ V3MN type II β1F58X-Ray[90]
I GPGRAF AFYTTKN TypeI-typeII -
typeI
X-Ray[91]
found to link two strands of anti-parallel β-sheet, forming a β-hairpin, and such secondary structure elements have been postulated as possible nucleation sites for the protein folding pathway [44]. A basic nomenclature for hairpin turns is shown in Fig. 2 where the loop residues are labelled L1, L2 and so on, and residues in the N-terminal strand are labelled -B1, -B2, etc. from the turn, and the C-terminal residues +B1, +B2, etc.
3.4. Cis-Trans Isomerisation of V3 Loop Gly-Pro Peptide Bond
Env gp120 adopts several conformational states [45] where the conformational changes include variation in V3 shape or exposure as shown by changes in V3 reactivity with conformation, dependant antibodies. Molecular interactions between V3 loop and its coreceptor induce further confor-mational changes in the envelope protein gp120, exposing the fusion domain of gp41, which ultimately mediates fusion of the cellular and viral membrane. Endrich and Gehring [46-47] have hypothesized that binding of cyclophilin A (CypA) to the V3 loop might catalyze, due to CypA confor-mase activity, the conformational changes in gp120, which in turn might enhance infectivity. The two
glycine residues on each side of the proline residue (GPG, see Fig. 1) provide conformational flexibility and enable different HIV-1 strains to adopt different conformations. This conformational vari-ability can be explained by the cis-trans isomerisation (see Fig. 3) of the Gly-Pro and Pro-Gly peptide bonds during an early step in the HIV-1 infection process. Such isomerisation has already been suggested by computer modeling studies which demonstrated that the enhanced susceptibility of the V3 loop to proteolytic cleavage following CD4 binding can be explained by a type VI xGPG turn, not by a type II GPGR turn [48]. Further support for the CypA isomerase activity to V3 loop comes from more recent studies showing that V3 peptides have a high affinity to cyclophilins and FKBP [47] and that Cyp A is involved in an early stage of the infection process [49]. Hence, one property of HIV-1 that makes it unique among all retroviruses is the incorporation during its life cycle of the host protein cyclophilin A (CypA) into its virion [50-52]. There is a lot of evidence that points to the fact that CypA is required for HIV-1 replication [50-55]. It is also the first known cellular protein other than the cell sur-face receptors CD4, and coreceptors CCR5 and CXCR4 to be involved in viral replication during an early step and be-fore the initiation of reverse transcription [56-59]. These immunophilins are prolyl-peptidyl isomerases and could catalyse the postulated cis-trans isomerization. Alternatively, immunophilins could bind to the V3 loop, preventing inter-action of the V3 loop with co-receptors. It was also hypothe-sized that the CypA-V3 loop interaction during primar
y in-fection shields the immune system to recognize the principal neutralizing domain GPG of HIV-1. Further during infection, antibodies production against the principal neutralizing do-main share close similarity with amino acid sequences and structures of RANTES, MIP-1α and MIP-1β. The interac-tions of antibodies with chemokines that structurally resem-ble the V3 loop results in HIV-1 infection becoming an autoimmune
豫公网安备41160202000603
站长QQ:729038198
关于我们
投诉建议