HIV-1 entry and how to block it

Introduction Our understanding of the process by which the HIV enters cells and initiates its replication has considerably evolved over the past few years. Among the most spectacular achievements are the identification of novel cellular factors behaving as viral receptors, the elucidation, at least partial, of the spatial structure of the viral envelope proteins, and the discovery of compounds efficiently blocking the cell entry process. Knowledge has been growing at an unprecedented pace, allowing questions left open since the HIV disease emerged some 20 years ago to meet their answer. In this brief article, we shall attempt to summarize the recent evolution of this field of HIV research, and discuss the promises of antiviral strategies based on the novel findings. But we also would like to show that a number of important questions remain unsolved. The issues that will be discussed represent a selection, perhaps arbitrary, within the research field of HIV entry, for which both recent and excellent reviews are available [1-3]. We focus on the type 1 virus (HIV-1), for which there is considerably more information than for the related human (HIV-2), simian or feline immunodeficiency viruses. HIV-1 entry: an overview Like all retroviruses, HIV enters cells by fusing its envelope with a cellular membrane, allowing the inner content of the viral particle (the nucleocapsid) to gain access to the cytoplasm [4]. The envelope is a phospholipid bilayer formed during the virus budding process, carrying proteins from both viral and cellular origins [5]. The former represent the cell entry machinery of the virus. In contrast, there is no consensus for a well-defined role of the proteins of cellular origin, even though they can represent a significant fraction of the total envelope proteins [5]. They have been proposed to contribute to cell attachment or to triggering some form of activation of target cells [5]. Noteworthy is their relative enrichment for particular cellular proteins (MHC class I and class II, ubiquitine, intracellular adhesion molecule-1) or, on the contrary, their exclusion (e.g., CD45) could reflect the budding of HIV-1 from particular areas of the plasma membrane, such as the glycolipid-rich domains termed 'rafts" [6]. This could also explain the relatively high cholesterol content of the HIV-1 envelope [7]. The HIV-1 envelope proteins (Env) consist of two types of glycoprotein named after their apparent molecular mass, gp120 for the surface subunit (SU) and gp41 for the transmembrane subunit (TM), both resulting from the proteolytic cleavage of a gp160 precursor [4]. The gp41 and gp120 subunits remain non-covalently associated and form higher order oligomers at the surface of cells or virions. Although dimers and tetramers of Env were visualized [8], several structural studies have shown that the TM of HIV-1 or other retroviruses seem to spontaneously organize as trimers [9]. The spatial structure of the HIV-1 envelope proteins has long remained beyond reach, and knowledge on the structure of the gp120 or gp41 was mainly deduced from indirect approaches, such as mutagenesis or epitope mapping. We shall see that these deficiencies have now largely been filled. For the moment, let us recall that the surface protein gp120 is characterized by an overall high degree of genetic variability between isolates, and even more between HIV-1 and HIV-2. The most variable domains, termed hypervariable loops, are exposed at the surface of the protein and also carry most of the sugar groups, while the more conserved regions are folded to form the core of the protein [10]. Consequently, a large part of the gp120 subunit is masked in the native virions, and so is the gp41 subunit (see Fig. 1).Fig. 1: Organization of gp41 and the model for HIV-1 membrane fusion. (A) Linear representation of gp41. The elements shown in the ectodomain are the the fusion peptide, the proximal (P) and distal (D) helixes, and the conserved dicysteine motif (S-S), as well as the region corresponding to the T20 peptide. (B) Schematic model for the initiation of membrane fusion between HIV and the target cell. (a) Before contact of the gp120 complex with cellular receptors, the native conformation of gp41 is that of a trimeric 'hairpin' structure; note that D helixes (brown) form its inner core, and that the fusion peptides should therefore be buried and pointing toward the virus membrane. (b) Two models are proposed to account for the unmasking of fusion peptides and their anchoring to the target membrane. Top, the dissociation of the hairpin and formation of a hypothetical transient 'pre-hairpin' structure, as proposed in [11]; bottom, the alternative model is far simpler in terms of topology, since it only postulates tilting of the gp41 axis [94]. (c) Return to the hairpin structure, or gp41 tilting, resulting in close apposition of membranes and initiation of lipid merging. (d) Fusion pore formation and dilatation. The T20 peptide could act by binding to the 'pre-hairpin' structure, if any, thus preventing the transition from (b) to (c), or it could interact with gp41 and block a later step.The role of the SU Env subunit is to establish nonspecific contact with target cells and then to engage interactions with particular proteins of their surface, termed receptors [4]. The latter interactions result in a change in the conformation of the SU/TM complex, unmasking domains of the TM subunit that are responsible for the merging of the viral and cellular membranes and for the formation of a pore between the inner content of the virion and cytoplasm [9,11]. The next steps of the virus entry process are considered to be the expansion of this pore and the translocation of the nucleocapsid [12-14]. Understanding these steps of the entry process can shed light on aspects of the viral tropism and pathogenesis. In addition, each step represents an actual or potential target for antiviral therapeutics. First contacts The adhesion of HIV-1 to cells seems principally due to non-specific interactions between the gp120 and charged groups of the cell surface, in particular the sugars of glycolipids (e.g., galactosyl ceramide) and glycoproteins, particularly heparan sulfate [15]. These interactions are not sufficient to trigger virus entry, but are probably important for the subsequent recognition of the receptors by the gp120, either by modification of its conformation or more simply by concentrating virus particles at the cell surface. Compounds blocking the interaction with sugars, for example heparin or anionic polymers (D2S, SPC3), can indeed inhibit HIV-1 entry [16]. Contact with cells can result in internalization of virions by an endocytosis process apparently independent of viral receptors [17]. Recently, endocytosis was proposed to be a route of efficient infection for certain HIV-1 strains [18], but the relevance of these observations to the in vivo situation is unknown. There is indeed a huge body of arguments to show that HIV-1 fusion occurs at the cell surface and at neutral pH [19], at the difference with influenza virus for which the fusiogenic protein (hemagglutinin) must be activated by low pH, hence the requirement for transit in a late endosome [20]. The interest in the step of cell adhesion was recently raised by the finding that dendritic cells, a class of antigen-presenting cells present in the skin and mucosae, could capture virions and retain them for extended periods of time (up to 4 days) in an infectious form. They then efficiently present these virions to permissive T cells, resulting in a dramatic enhancement of infection [21]. This process could be particularly important in the mucosal transmission of infection [2,22]. The uptake of HIV-1 seems mediated by the interaction of the gp120 with the mannose-binding C-type lectin domain of a type II membrane protein termed DC-SIGN present on some types of dendritic cells [21,23,24]. Some important issues about this mechanism of HIV-1 transmission still have to be clarified, such as the nature of the interaction of DC-SIGN with gp120, the reasons for the prolonged stability of virions at the surface or in an intracellular compartment of dendritic cells, the possible effects of their productive infection on the presentation of virions to T cells, or the possible influence of a DC-SIGN polymorphism on the individual susceptibility to HIV-1 [25]. If the infection of dendritic cells proves important for the establishment of HIV-1 infection, DC-SIGN could become a potential target for antiviral approaches. Receptors The issue of HIV receptors is probably the most spectacular example of an evolution, or revolution, of concepts on HIV entry. Soon after HIV-1 was discovered in 1983, the CD4 molecule was identified as its receptor [26]. Considerable efforts have been devoted to elucidate the interaction of gp120 with CD4 cells [27], and to develop antiviral agents that could block it [16,28]. However, ample evidence was gathered during the years 1984-1996 that the presence of CD4 cells was not always necessary for infection by HIV-1 or related retroviruses [29] and, more importantly, not sufficient for the infection of certain cell types, such as macrophages or mouse cells [30]. There was no satisfying explanation to these disturbing facts until the discovery in 1996 that another class of cell surface molecules, chemokine receptors, was involved in the cell entry process [31]. These proteins are members of the family of receptors with seven membrane-spanning domains (7TM) whose function is to transduce a cell signal, in particular through coupling to heterotrimeric G proteins, after contact with one or several given ligands [32]. Chemokines are themselves a family of some 45 small proteins (8-14 K) principally involved in the activation of leukocyte chemotaxis. Most chemokines are classified as CC or CXC based on the relative position of conserved cysteines in their amino-terminal region, and the same terminology (CCR or CXCR) is used for their receptors (for a review, see [33]). We shall not recall the steps that led to this major discovery, emerging from two seemingly distinct phenomena (restriction of HIV-1 entry to certain target cell types, and control of HIV-1 infection by soluble suppressor factors [34]), and shall only present a concise list of the principal results established over the past few years. First, the presence of a chemokine receptor at the cell surface is necessary for infection by HIV-1 and related retroviruses, while that of CD4 cells is not always required [1]. In the case of HIV-1, the only chemokine receptors found to play a role in vivo are CCR5 and CXCR4 [35,36], even though many others can permit experimental infection [1]. Second, CCR5 and CXCR4 can interact with gp120, usually in the presence of CD4, hence the terminology CD4-associated co-receptors [37-39]. Third, the ability of HIV-1 strains to use either CCR5 (viral strains termed R5) or CXCR4 (X4 strains), or both (R5X4 strains), determines biological properties such as tropism for macrophages (R5 viruses) or for immortalized cell lines (X4 strains) [40]. In vivo, X4 and R5X4 strains generally emerge at later stages of infection, while R5 viruses can be isolated throughout infection [1]. The X4 and R5X4 strains are probably counter-selected when the immune defenses are fully efficient, but their emergence could represent an aggravating factor; in particular, by allowing HIV-1 to gain access to a larger spectrum of target cells, including resting T cells [41] or CD4-negative cells [42,43]. Finally, the chemokine ligands of CCR5 (MIP-1α, MIP-1β, RANTES) or CXCR4 (SDF-1) can block HIV-1 infection by different mechanisms (mainly steric hindrance and receptor endocytosis) [44,45], and this is also the case for synthetic ligands, which of course led to envision novel antiviral approaches [16]. This issue will be discussed in the final section of this article. To summarize the evolution of concepts, CD4 that was initially considered the key element determining the possibility for a cell to become infected by HIV-1 are now rather seen as an activator of the interaction of HIV-1 with chemokine receptors. This explains a number of observations that did not fit well with the CD4 paradigm, and the relative inefficiency of CD4-based antiviral strategies. But certain questions remain and others are raised regarding the precise role of chemokine receptors in the HIV-1 entry process. In particular, it can be wondered why only the relatively divergent receptors CCR5 and CXCR4 are used in vivo, while other chemokine receptors can only mediate HIV-1 entry experimental and most are not at determines the between CCR5 and CXCR4 or the ability to use both of Finally, are chemokine receptors the of the in terms of cellular factors required for HIV-1 the has been to that the ability of chemokine receptors to mediate HIV-1 entry only on their ability to interact with gp120 in a This is HIV-1 which is a but not the most as it a receptor role for Considerable efforts are therefore devoted to elucidate the interaction of gp120 with chemokine receptors at the molecular can and the interaction between gp120 and CXCR4 or CCR5 but they can only be with gp120, which from the native trimeric Env complex of virions. For of CCR5 to gp120 in were found to mediate HIV-1 entry Most studies have therefore used infection by HIV-1 or fusion with cells, to the interaction of chemokine receptors with the these the of gp120 binding and of other properties of chemokine receptors that could be required for a interaction to membrane fusion and virus entry. We shall only summarize the of these The of viral strains for CCR5 or CXCR4 seems largely on the of variable regions of gp120, in particular the and to a degree the region However, the that viruses with different can interact with CCR5 for the of a relatively conserved binding in the to conserved domains of gp120 were involved in the interaction with CCR5 on the spatial structure of the gp120 core as a complex with CD4 and the chemokine receptor binding was proposed to be a between the inner and regions of gp120 and the that them This is close to the of the and loops, and could be masked by them in the native gp120 conformation it has some form of usually by contact with CD4 This could explain the influence of these variable on the interaction with chemokine receptors and on the for CCR5 or model for the interactions of HIV envelope proteins (Env) with cell surface receptors from (a) Schematic representation of the Env and the CD4 cell and CCR5 or CXCR4 chemokine receptors. amino-terminal (b) to domain 1 of CD4 cells in gp120, unmasking a chemokine receptor binding a and a conserved structure in (c) with the chemokine receptor close apposition of the virus and cell membrane, and results in in the complex, gp41 allowing its fusion peptide and in the target cell regions of chemokine receptors apparently involved in the interaction with gp120 are domains, in particular the amino-terminal domain and the in The role of in the can be deduced from mutagenesis studies and from the of the of the domain These studies the of for the role of in the variable of gp120 that have long been to influence HIV-1 cell tropism However, not all found important for the of CCR5 or CXCR4 are charged and their can have effects In the case of of the of the or even within TM domains were found involved in the HIV-1 The gp120 could therefore become relatively close to the target membrane, subsequent of the gp41. of results with different viral strains or with different types of that the interaction of gp120 with chemokine receptors is relatively and not which is in apparent with the if not role of CCR5 or the presence of these receptors explain certain facts the HIV-1 cell such as the of most X4 strains to macrophages even though CXCR4 is In addition, there are that in gp120 or chemokine receptors, to their can have effects on viral or tropism can therefore be wondered the HIV-1 of CCR5 or CXCR4 only on their interaction with gp120, or some other properties could explain the of these receptors. Among properties of chemokine receptors that could have a or influence are the formation of with CD4 or in to gp120 even though of CCR5 or CXCR4 to trigger a cell can mediate infection experimental a could play a role in steps in certain cell types, such as macrophages that the properties of these chemokine receptors could be to a particular at the plasma membrane, for example in that are in transmembrane cholesterol from the cell membrane seems to the formation of infectious HIV-1 virions there is not evidence for the of domains or in HIV-1 entry, for other viruses and The interaction of gp120 with chemokine receptors conformation in the Env complex that are by the unmasking of gp120 and gp41 or even by the of gp120 from virions in the case of cell X4 HIV-1 strains their to compounds such as soluble CD4 or certain We have a knowledge of the that to activation of the fusiogenic of gp41. with the influenza virus fusion probably of several gp41 trimers There is also evidence that interactions with receptors are required to the change in Env least CD4 proteins and to chemokine receptors in with the of the Env function and with the of receptors in the presence of Env fusion There has been important in understanding of the structure of the gp41 and other transmembrane proteins in but we still have a knowledge of the to the merging of the membranes and then to the formation of a pore allowing translocation of the The structure of the part of or results from the of two domains termed proximal (P) and distal relative to the of the protein The helixes from gp41 through interaction to form a type structure which are D helixes in in In this structure, the amino-terminal part of each gp41 as the fusion peptide, must be relatively buried and toward the viral membrane, while the region of some found between the and D helixes a that is in contact with the gp120 This is also characterized by the presence of a dicysteine motif with a proposed to contribute to the with gp120, and to an epitope of the gp41 The of the in gp41 membrane fusion postulates the of a transition termed the in which the is so that the fusion peptide can toward the target membrane Return to a hairpin structure is considered to the merging of the viral and cell allowing for lipid and fusion pore This model is considered to some with in the fusiogenic protein of the influenza virus (hemagglutinin) [9]. However, it far more important in particular, the dissociation of the gp41 since the helixes are its inner The major arguments in of the of a 'pre-hairpin' structure are the unmasking of in the and D helixes during the fusion process and, the antiviral of peptides from these gp41 domains But the unmasking of could simply result from dissociation of the complex, while the of peptides could result from interaction with gp41 in its native was indeed that a peptide from the distal was on a step to the initiation of lipid that it on the native gp41 and not on a 'pre-hairpin' model was therefore proposed to account for the of membrane in which gp41 in the The only structural change required would be the tilting of the gp41 axis so that the fusion peptide can access the target membrane and its role This could result from the or dissociation of the gp120 after interaction with chemokine receptors. If this model is to be the initiation of lipid between the viral and cell membrane has to be as one of the of which must also be involved in the next steps to formation of a fusion pore and its dilatation. a major for this model is the of of peptides from the distal of such as T20 and others or of the a apparently blocking a late step of HIV-1 entry and the region of the gp41 of these compounds could be to the of gp41 which is probably required for membrane fusion as proposed for or for the elements intracellular fusion approaches The cell entry of HIV-1 therefore many to block we shall focus on compounds chemokine receptors or blocking the fusiogenic of the other viral and cellular as well as approaches charged but not related compounds such as or the were found to to the charged surface of in particular its and to block its interaction with gp120, and therefore replication of X4 HIV-1 strains (for a review, see is a small molecule that was found to be in vivo in a mouse model potential could be the CXCR4 blocking mediated by the chemokine The indeed seems to be involved in biological the and HIV-1 receptor of CXCR4 can be and and it should be possible to develop blocking In the case of no such is given the apparent of of the of CCR5 through different types of genetic chemokine ligands of CCR5 and their can efficiently block HIV-1 entry in seems to be an antiviral given the of of these proteins and their are The was the but the and compounds seem more given their antiviral with antiviral approaches CCR5 could be the of X4 and should be given to also both the CCR5 and CXCR4 receptors. This could be in since it is by a chemokine to gp41 we have peptides from the gp41 helixes, in particular from the distal such as T20 which is with a relatively high antiviral and is apparently efficient in a [16]. of gp41 are to be for HIV-1, and was between T20 and or Finally, the structure of the gp41 core has a that could be a target for and, more a based on the use of a form of gp41 has been proposed understanding of the of gp41 to the membrane fusion process will one to and to available receptors play a role in the cell entry process of HIV-1 and related other retroviruses seem to use proteins with membrane-spanning domains as their receptors [4]. The for the of such proteins during the evolution of these viruses is probably to their particular They indeed form a in the lipid and not its surface CD4, for The latter probably the merging of the viral and cellular membranes after the adhesion while the former could be important for the formation of the fusion by the exclusion of from the of fusion. the of such a cellular receptor can be the why CCR5 and CXCR4 were by HIV-1 is We that the interaction of gp120 with these two receptors was not a fully also why X4 strains are most until late stages of infection, or why the R5X4 strains can use either CCR5 or which have different domains, while they use or which are more related to each of to these questions and others regarding the role of chemokine receptors in the HIV-1 entry process will probably the on aspects of HIV-1 entry rather than on the between molecular even if the latter is probably the key to in terms of antiviral