Acquired Immunodeficiency Syndrome (AIDS) is a modern day pandemic that afflicts millions of people worldwide with a constant and bewildering increase in the number of infected individuals. Human Immunodeficiency Virus type 1 (HIV-1) is a retrovirus and the definitive causative agent of AIDS. Despite the availability of an effective and specific anti-retroviral therapy, at the present time AIDS, is still considered an incurable disease (Barre-Sinoussi et al., 1983; Fauci, 2003; Gallo et al., 1984; Popovic et al., 1984; Vilmer et al., 1984). Given the devastating effects of the viral infection on the host immune system, and the present absence of therapy that would purge the host of the disease, AIDS is now viewed as one of the most threatening epidemics of the century (Fauci, 1999; Fauci, 2003). Therapeutic strategies designed to combat viral replication have been successful in achieving significant levels of reduction of the viral spread, but are not sufficient to eliminate the virus from the host, due to the emergence of resistant viruses and persistence of latently-infected viral reservoirs (Belmonte et al., 2003; Capiluppi et al., 2000; Nunnari et al., 2002; Pomerantz, 1999a; Pomerantz, 1999b; Pomerantz, 2001a; Pomerantz, 2001b). One of the major reasons for the emergence of resistant viruses is the apparent ease with which the low fidelity reverse transcriptase enzyme of HIV-1 generates mutations during viral replication. The persistence of latent viruses is due to the inability of the current drugs to inhibit integration, a process that mediates the insertion of reverse transcribed viral DNA into the host genome. Thus a pressing need exists to identify novel drug targets and new therapeutic strategies to combat AIDS (Anthony, 2004; Tarrago-Litvak et al., 2002).

Host-virus interactions are attractive to explore for the development of novel anti-HIV-1 strategies, as many steps of the viral replication involve an intricate interplay between the virus and the host machinery (Moore & Stevenson, 2000; Rowland-Jones et al., 2001). Given the apparent paradox between the minimalist nature of the retroviral genome and the complexity of the HIV-1 lifecycle, it is essential that the virus utilizes the host machinery both to promote its replication and, at the same time, to subvert and evade the antiviral responses of the cell (Freed, 2004; Ott, 2002; Popik & Pitha, 2000; Stevenson, 2003). In recent years, it has become apparent that virtually all the steps of viral replication rely heavily on host cell processes. Activities of cellular factors seem to be either hijacked by the virus to perform the pro-viral functions, or suppressed to minimize their potential anti-viral functions. Many aspects of the intricate interplay between the virus and pro-viral and anti-viral cellular proteins remain unclear, and many participants in this process have yet to be identified. Nevertheless, an array of cellular factors contributing both to propagation of the virus, and to cellular defense has been identified in the past couple of years, and paints a fascinating picture of the convoluted relationship that has evolved between the virus and the invaded cell. Clear understanding of cellular and viral contributors to this interplay will not only provide insights into the mechanism of various steps of retroviral replication, but will also likely lead to the development of novel therapeutic strategies to specifically disrupt virus-host protein interactions (Greene & Peterlin, 2002; Moore & Stevenson, 2000; Tang et al., 2002).

Viral replication takes place in several distinct steps, and can be divided into early and late events (Freed, 2001; Freed and Martin, 2001; Goff, 2001a). Early steps of viral replication initiate with the virus contacting the target cell, and proceed through integration of the viral genome into the host chromosome; late events take place after integration and lead to the release of newly formed virus particles from the infected cell (Figure 1). Viral entry into the target cell is mediated by the interaction between the envelope glycoprotein on the mature infectious virus particle with the appropriate receptor (CD4) and co-receptor (CCR5 or CXCR4) on the surface of the cell, followed by a fusion step. Upon entry into the cell, the viral nucleoprotein complex, sequestered within the capsid (CA) shell, undergoes uncoating, exposing the viral machinery to the cellular environment, and allows the initiation of reverse transcription. The viral RNA genome is then reverse transcribed by the virally encoded reverse transcriptase (RT) to produce a linear double-stranded cDNA molecule. The cDNA is assembled into a high molecular weight nucleoprotein complex termed the pre-integration complex (PIC) that contains a number of known, as well as many unidentified, cellular and viral proteins. The PIC is then translocated into the nucleus of the target cell through a nuclear pore complex (NPC). The viral cDNA carried into the nucleus within the PIC is subsequently integrated into the host chromosome, the process of which is catalyzed by the virally encoded Integrase (IN). The above events constitute the early stages of viral replication. During late events the integrated viral DNA, termed the provirus, is transcriptionally activated by the virally-encoded transactivator TAT and cellular transcription factors. Singlyspliced, multiply-spliced and unspliced RNA molecules that are formed within the nucleus are then exported into the cytoplasm. While singly- and multiply-spliced RNA molecules encode viral proteins, the unspliced message is the genomic viral RNA that subsequently gets assembled into the immature virus particle. The assembled virus particle then buds out of the producer cell and simultaneously undergoes maturation to form the infectious particle. Throughout these steps of viral replication, viral components interact with, confront and exploit the host cellular machinery (Figure 1).

Receptors, Coreceptors and Lipid Rafts

One of the first observations of host factor involvement in HIV-1 replication following the discovery of HIV-1 was that the virus utilizes CD4, a member of the immunoglobulin superfamily, as a primary receptor for targeting the cells for infection (Dalgleish et al., 1984; Klatzmann et al., 1984). Shortly after this discovery, it became clear that CD4 alone was not sufficient to support efficient HIV-1 infection, leading to the discovery of two major coreceptors for the virus (Alkhatib et al., 1996; Choe et al., 1996; Deng et al., 1996; Dragic et al., 1996; Feng et al., 1996). These coreceptors, CCR5 and CXCR4, that belong to a family of Gprotein- coupled chemokine receptors, have been shown to confer susceptibility to macrophage-tropic and T cell-tropic strains of HIV-1, respectively. The coreceptor proteins appear to act in cooperation with CD4 to mediate efficient fusion of the virus driven by the viral envelope (Env) proteins. CCR5 is viewed as an attractive target for the development of antiviral drugs, as homozygous mutation in this co-receptor in certain individuals confers resistance to HIV-1 (Dean et al., 1996; Liu et al., 1996; Samson et al., 1996). Selective distribution of the co-receptors on different cell types, which has been shown to determine the tropism of HIV-1, has led to the classification of viruses based on the type of co-receptors they utilize. HIV-1 that utilizes CXCR4 is termed X4; virus utilizing CCR5 is termed R5; and virus capable of using both types of receptors (dual-tropic) is classified as R5X4 (Berger et al., 1999).

Cooperative binding of viral gp120 to host CD4 and a co-receptor promotes fusion of the virus with the cell membrane by utilizing biological lipid membrane microdomains called lipid rafts. Lipid rafts are rich in cholesterol and sphingolipids and facilitate viral fusion and entry purportedly in a pH-independent manner (Liao et al., 2001; Liao et al., 2003; Nguyen & Taub, 2002; Popik et al., 2002; Raulin, 2002; Viard et al., 2002).

In addition to direct utilization of the receptor- and coreceptor-expressing cells by HIV-1, other types of cells, such as antigen-presenting dendritic cells (DC), appear to be involved in HIV-1-mediated infection. The C-type lectin, DC-SIGN, present on the surface of DC, promotes their indirect infection by associating with the virus. DCSIGN binds to HIV-1 via gp120 Env and takes up and anchors the intact virus particle, but most DC cells do not undergo productive infection (Geijtenbeek et al., 2000; Geijtenbeek & van Kooyk, 2003; McDonald et al., 2003). However, HIV-1-harboring DC are believed to facilitate viral transmission by carrying HIV-1 to its target cells.

Host Restriction Factors

Upon entry of the virus, CA sequesters the viral nucleoprotein complex and encounters a multitude of host factors within the cytoplasm of the infected cell (Goff, 2001b). Removal of the CA and release of the viral nucleoprotein machinery into the cellular environment constitutes the process of uncoating, which in turn leads to the initiation of reverse transcription. For many years it was virtually unknown how virus/host encounters at the early post-entry stages affect subsequent steps of viral replication, although it was clear that the dynamic interplay between viral factors and cellular machinery occurring during these stages plays a critical role in determining whether the virus would be able to continue its propagation. Recent landmark studies have not only shed light on the specific processes of post-entry events, but have also imparted a deep appreciation for the evolutionary relationship that has developed between viruses and mammalian cells.

Viruses have coexisted with cells for tens of millions of years. The evidence of such co-existence is provided by the endogenous retroviruses that have been incorporated into the human genome. Throughout mammalian evolution, some viruses have progressed to cause harm to host cells, and replication of those viruses can result in disease (Bock & Stoye, 2000; Stoye, 2001; Sverdlov, 2000; Weiss et al., 1999). Perhaps because of this reason, mammalian cells have developed a variety of mechanisms to protect themselves from pathogenic viruses. Both innate immune responses, as well as numerous adaptive immune responses, have evolved to prevent enslaving of the organism by viruses. In addition to global protective mechanisms provided by the immune system, an array of cell-autonomous factors that bestow protection against specific viruses has evolved in mammalian cells.

In recent years it has become apparent that these factors act in a dominant manner to inhibit viral replication in certain cell types and play a critical role in determining the host range of viruses. Both viral nucleic acids and viral proteins seem to be targeted by host factors conferring anti-viral activities at the early post-entry steps of viral replication (Freed, 2004; Stevenson, 2003). For example, the anti-viral activity of the cellular protein APOBEC3G, which will be described in more detail below, has recently been demonstrated to act directly on the viral genome, but is counteracted by the viral protein viral infectivity factor (Vif) (Sheehy et al., 2002; Sheehy et al., 2003; Simon et al., 1998). Another class of factors that direct their action towards the viral CA protein to inhibit the post-entry pre-integration events has been termed “restriction factors” (Besnier et al., 2002; Bieniasz, 2003; Cowan et al., 2002; Munk et al., 2002; Towers et al., 2000).

Retroviral restriction factors were originally described in murine cells that showed controlled susceptibility to the Friend’s murine leukemia virus (Lilly, 1967). The gene products controlling resistance of the murine cells to the virus were identified as Friend virus susceptibility factors (Fv). One of the Fv genes, Fv1, was particularly interesting due to its ability to exert restrictive activity in vitro (Gardner et al., 1980; Hartley et al., 1970; Lilly, 1967; Pincus et al., 1971; Rasheed & Gardner, 1983). Two alleles of the gene, Fv1b and Fv1n confer resistance to the B-tropic and N-tropic murine leukemia viruses, respectively (Hartley et al., 1970; Pincus et al., 1971). Intriguingly, the Fv1 protein itself seems to have a retroviral origin (Best et al., 1996). Its protein sequence has homology to the capsid-related domain of the ERV-L family of the endogenous retroviral Gag proteins. It is not clear how the product of the Fv1 gene exerts its inhibitory effect on viral replication, although it is apparent that the restriction occurs post-entry, after the initiation of reverse transcription but before the integration (DesGroseillers & Jolicoeur, 1983; Stoye, 1998) (Figure 2A). Since introduction of Fv1 into the non-murine cells is sufficient to confer resistance to MLV, it is likely that Fv1 does not require additional factors unique to the murine cells to exert its function. In addition, the number of Fv1 protein molecules within the cell seems to be a limiting factor in restraining infection. The levels of expression of the Fv1 protein in the murine cells are very low, and it has been demonstrated that even though the FV1-mediated restriction is relatively strong, it can be abrogated by saturation of the target cells with non-infectious virus particles derived from the susceptible cells (Boone et al., 1990). These observations, combined with the finding that the determinant of the virus to the Fv1 resistance maps to a single amino acid at position 110 of the capsid protein, suggest a model where restriction is mediated by direct association of the Fv1 protein with the incoming capsid (DesGroseillers & Jolicoeur, 1983; Goff, 1996; Kozak & Chakraborti, 1996). This model, although attractive, remains to be proven.

In recent years it became clear that the inhibitory activity of restriction factors is not limited to murine cells, but rather is widespread throughout the mammalian family (Towers et al., 2000). Restriction factors are now considered one of the major driving forces of species-specific infection by retroviruses (Bieniasz, 2003). In human cells, which do not express Fv1, but are still resistant to infection with N-MLV, a factor termed restriction factor 1 (Ref1) exerts inhibitory activity on the viral spread in a manner highly reminiscent of Fv1 restriction in murine cells (Towers et al., 2002) (Figure 2B). While the restrictive activity of Ref1 appears to take place at the early postentry stages of viral infection, at or before the initiation of reverse transcription, the Fv1 factor appears to mediate its effect after the initiation of the reverse transcription (Figure 2). Remarkably, the viral determinant that controls restrictive ability of Ref1 maps to the same amino acid within the capsid sequence as that of Fv1, suggesting a strong evolutionary connection between the two factors (Towers et al., 2000).

In primate cells, a similar mode of restriction is mediated by a group of factors collectively termed lentivirus susceptibility factor 1 (Lv1) (Hatziioannou et al., 2003; Towers et al., 2000) (Figure 2B). The presence of Lv1 factors in cells of Old World and New World monkeys confers resistance to infection by HIV-1, and similar to the cases of Fv1 and Ref1, restriction is determined by the viral capsid (Towers et al., 2000). Unlike Fv1, which has been shown to specifically restrict MLV, but not the lentiviruses, Lv1 and Ref1 seem to restrict infection by numerous divergent retroviruses (Bieniasz, 2003; Hofmann et al., 1999; Towers et al., 2000) (Figure 3). In the case of Lv1, a particularly interesting example of such diversified restriction is represented by primate cells derived from African Green Moneys (AGM) that are resistant to infection by HIV-1, HIV-2, N-MLV, SIVmac, and Equine infectious anemia virus (EIAV). In human cells, restriction by Ref1 also efficiently inhibits N-MLV and EIAV (Besnier et al., 2002; Cowan et al., 2002; Hatziioannou et al., 2003). Heterokaryon studies have demonstrated that these factors act in a dominant way to restrict infection (Cowan et al., 2002; Munk et al., 2002). In addition, despite the wide divergence of the retroviruses that are restricted in primate and human cells, a series of abrogation studies have demonstrated that the same restriction factor is dominantly mediating the restriction of various incoming viruses within the cell (Besnier et al., 2002; Cowan et al., 2002; Hatziioannou et al., 2003). These studies have elegantly shown that when the cells are saturated with virus-like particles derived from one type of retrovirus that is restricted in these cells, and then infected with another type of virus, restriction is abrogated. Saturation of the cells with non-restrictive virus particles does not diminish the restriction, supporting the idea of specific recognition of the viral capsid molecules by a single type of factor in these cells. This notion of a single factor that selectively recognizes a variety of capsids is quite remarkable. It also strongly suggests that the restriction factors present in mammalian cells are evolutionarily closely related to each other. A series of recent reports confirmed this notion when it was revealed that Ref1 and Lv1 belong to a single family of tripartite interaction motif 5α (TRIM5α)-containing proteins (Hatziioannou et al., 2004; Keckesova et al., 2004; Perron et al., 2004; Yap et al., 2004).

TRIM5α-α: Unified Host Restriction Factor?

A breakthrough study that has allowed recognition of these factors identified the cytoplasmic body component TRIM5α as an HIV-1 restriction factor in rhesus monkey cells (Stremlau et al., 2004). The rhesus monkey TRIM5α protein has been shown to confer resistance to HIV-1 much more potently than human TRIM5α, and the determinant of the restriction by TRIM5αrh has been mapped to the HIV-1 capsid. The mode of action by TRIM5αrh to confer restriction, as well as the specificity of its effects have been recognized as highly similar to that of Lv1 factors, and, in a short time, have led to the identification of species-specific variants of the TRIM5α proteins that have bona fide Ref1- and Lv1- type restriction activity. The human TRIM5α protein has been definitively shown to account for Ref1 activity in human cells. When TRIM5αh is expressed in permissive cells, the cells become resistant to infection by N-MLV, but not B-MLV. At the same time, when the expression of the TRIM5αh is abrogated in human cells, the cells are sensitized to infection by N-MLV. Species-specific variants of primate TRIM5α protein have also been identified and have been demonstrated to confer restriction to an array of retroviruses in the manner of Lv1. AGM TRIM5α has been shown to protect cells from infection by HIV-1 and SIVmac, as well as the nonprimate lentivirus EIAV, and at the same time allowing replication of AGM SIV. Consistent with its function as the Lv1 restriction factor, siRNA-induced abrogation of AGM TRIM5a made the AGM cells susceptible to infection by HIV-1. Several studies have thus concluded that the species-specific isoforms of the single polymorphic factor TRIM5α perform the function of Lv1 in Old World monkeys.

The fact that variations of the single factor are able to specifically confer restriction through recognition of the viral capsid is unprecedented (Owens et al., 2004; Owens et al., 2003). The capsid proteins of retroviruses, although conserved, vary significantly from one virus to another. At the same time, N-MLV and B-MLV capsids differ by only a few amino acids, and this difference is sufficient to confer resistance. Specific conformational changes of capsid proteins are therefore likely to contribute to the specificity of recognition by the restriction factors. It is also possible that the ability of capsids to interact with other species-specific host factors contributes to the specificity of the restriction. In one case of the New World primates, this notion has been recently confirmed. Among the New World primates, the owl monkey is the only species that is restrictive to infection by HIV-1 (Hofmann et al., 1999). Interestingly, restriction by Lv1 in these primates requires interaction of the viral CA with the cellular factor Cyclophilin A (CypA) (Towers et al., 2003). Studies have demonstrated that the disruption of the interaction between the HIV-1 capsid and CypA by means of mutating the interaction domain of CA, or using drugs that abrogate the interaction by sequestering CypA, relieved restriction to the HIV-1 infection in these cells. The fact that the CA-CypA interaction is required for sustained restriction in the owl monkey cells is paradoxical since, in human cells, the same type of interaction seems to promote replication of HIV-1 (Braaten et al., 1996; Braaten & Luban, 2001). An interesting discovery has recently been made that might explain the connection between the TRIM5α-mediated restriction in these cells and CypA. An overview of this discovery, as well as its implications will be discussed below in the description of CypA.

Cyclophilin A

Cyclophilins are a part of a large family of cellular factors that were originally shown to interact with the allograft rejection immunosuppressive drug cyclosporine (Thali, 1995; Wiederrecht et al., 1993). One of the members of this family of about 15 proteins, CyPA, has been shown to directly interact with the N-terminal domain of CA (Luban et al., 1993). The interaction of CyPA with the proline-rich loop in the CA domain of GAG allows incorporation of CyPA into HIV-1 particles in a stoichiometric ratio of CA to CyPA of roughly 10:1 (Franke et al., 1994; Thali et al., 1994). Although the mechanism of CyPA action in the viral lifecycle remains elusive, it appears to be important for efficient HIV-1 replication (Braaten et al., 1996; Braaten & Luban, The Einstein Journal of Biology and Medicine 17 !SCIENTIFIC REVIEW HIV-1 Dynamics in the Host Cell 2001; Franke & Luban, 1995; Saphire et al., 2000). The cellular function of CyPA as a cis-trans prolyl isomerase had been originally speculated to play a role in viral assembly, but later studies have shown that blocking CyPA incorporation does not affect virus particle assembly and release. At the same time it has been demonstrated that CyPA is required for normal HIV-1 replication kinetics, and abrogation of CyPA expression by specific knockdown results in reduced infectivity of HIV-1 particles (Braaten & Luban, 2001). Disruption of the interaction between the CypA and CA pharmacologically with Cyclosporin A, which binds competitively to the CA-interacting sites of CyPA, or by mutagenesis of the CyPA-binding residues of CA, results in abrogation of CyPA encapsidation into virions. Such virions are able to undergo normal assembly and release, but their infectivity is reduced for the next round of infection. The virions depleted of CyPA are blocked at early stages of reverse transcription, suggesting the requirement of CyPA for the early post-entry steps of replication (Towers et al., 2003).

Connecting Restriction to CypA

A link between restriction and CyPA was revealed when it was shown that HIV-1 virions, harboring CA mutants defective for CyPA binding, were poorly infectious in human cells expressing Ref1 (Towers et al., 2003). Surprisingly, the restriction was eliminated by pretreating cells with virus-like particles containing the CyPA-binding CA mutant, suggesting that Ref1 was saturated with the mutant CA. This observation clearly established the connection between restriction and CyPA (Towers et al., 2003). Moreover, the region of CA that confers the sensitivity to the post-entry block was mapped near the CyPA-binding loop (Kootstra et al., 2003; Owens et al., 2004). The above results, combined with several other observations, led to a current model which proposes that the interaction between the CyPA and CA in human cells protects HIV-1 Gag from Ref1- mediated restriction (Towers et al., 2003). The paradox observed in owl monkey cells does not fit into such a model since, in these primate cells which are restricted to HIV-1 infection, disruption of the CA-CyPA interaction alleviates restriction. This suggests that, unlike in human cells, CyPA is required for Lv-1-mediated activity (Towers et al., 2003).

One recent study has demonstrated that abrogation of CyPA expression by RNAi knockdown in owl monkey kidney (OMK) cells relieved the restriction to HIV-1 infection, and confirmed the requirement of CyPA to mediate restriction by Lv1 (now referred to as TRIM5α) (Sayah et al., 2004). Surprisingly, re-introduction of CyPA back into the OMK cells did not restore restriction to HIV-1 infection. An unexpected explanation of this phenomenon was found when it was realized that a previously unidentified mRNA, homologous to CyPA, was targeted by siRNA in the above experiments. The protein encoded by this mRNA turned out to contain the sequences of both TRIM5α and CyPA. The presence of this novel chimeric protein that contains 299 Nterminal amino acids of TRIM5α, linked to the complete CyPA sequence at the C-terminus, appears to be essential for mediating the restriction in the owl monkeys, as reconstitution of TRIM5α-CyPA in the cells knocked down for CyPA restores restriction to HIV-1. Since other New World monkeys lack the restriction to HIV-1, the TRIM5α-CyPA protein is likely to be unique to the owl monkey species (Sayah et al., 2004). Discovery of this novel protein is likely to be followed in the near future by the multitude of other species-specific factors mediating restriction. The demonstration of the highly specific restrictive action by these newly found but uncharacterized proteins marks the beginning of the unraveling of the highly complex system of antiviral defense built by mammalian cells throughout evolution. Elucidating the mechanism of the restrictive factors’ actions and understanding the evolutionary processes that stand behind formation of these factors are clearly of critical importance in both understanding the roots of retroviral epidemics and the design of antiviral therapeutic strategies.

APOBEC3G

In addition to the activities of the restriction factors that target the CA of the incoming virus, an activity against the viral nucleic acids has been shown to contribute to early post-entry antiviral cell defense (Sheehy et al., 2002). Recent studies have demonstrated that this anti-viral activity, at least in part, is attributed to the cellular factor APOBEC3G (Apolipoprotein B mRNA-editing enzyme, catalytic polypeptide-like 3G; also known as CEM15) (Harris et al., 2003; Lecossier et al., 2003; Mangeat et al., 2003; Mariani et al., 2003; Sheehy et al., 2003; Stopak et al., 2003; Zhang et al., 2003). This protein, expressed in human T lymphocytes, exerts its function through the interaction with one of the viral auxiliary proteins, Vif. Vif has long been recognized as a factor that enhances the infectivity of HIV-1, but, until recently, its mode of action was not clear. Interestingly, despite the fact that Vif acts during early stages of viral replication, its presence is required in the producer cells rather than in the target cells. The observation that infectivity of Vif-defective virions is predetermined exclusively by the type of cells used to produce the virus, suggested that the producer cells contain activity that is somehow counteracted by Vif. APOBEC3G has been shown to account for such an activity, and its presence in the producer cells renders Vif-negative virions non-infectious. Several studies have demonstrated that when Vif is present in the producer cells, it binds to APOBEC3G and promotes its degradation. In the absence of Vif, APOBEC3G is incorporated into the virus particles, and upon the entry of the APOBEC3G-carrying virus into the newly infected cell, it exerts its intrinsic cytidine deaminase activity on the minus strand DNA product of viral reverse transcription. Conversion of cytosines to uracils catalyzed by APOBEC3G on the minus strand viral DNA triggers the response by the host uracil-DNA glycosidases and repair enzymes, leading to degradation of the viral DNA, and thus to limiting the quantity of the viral cDNA available for subsequent replication steps. The cDNA that escapes the degradation undergoes G-to-A hypermutation, which accounts for limitation in the quality of the viral cDNA that is produced. As a result, the virus is severely impaired in its ability to proceed to steps following reverse transcription.

Reverse Transcription and Uncoating

Assuming that the virus successfully evades the antiviral responses mounted by the cell upon entry, the viral nucleoprotein complex is released into the cytoplasm and proceeds to reverse transcribe the viral RNA genome. The process of reverse transcription is catalyzed by the virally encoded enzyme reverse transcriptase (RT). The DNA polymerase activity of RT that is capable of using both DNA and RNA as templates converts the viral RNA genome into the doublestranded DNA molecule (Telesnitsky and Goff, 1997). Although the mechanistic details of reverse transcription are fairly well understood, the timeline of the process with respect to the uncoating, as well as the extent of the involvement of the cellular environment are not very clear and undergoes continuous reassessment (Goff, 2001a; Goff, 2001b). The initial hypothesis, originating from the studies of the yeast Ty retrotransposition elements, suggested that reverse transcription takes place in the shell-like enclosed environment containing RT, IN and CA proteins, and is permeable to dNTPs (Bowerman et al., 1989; Brown et al., 1987; Brown et al., 1989; Eichinger & Boeke, 1988; Garfinkel et al., 1985). The notion of reverse transcription taking place within a sequestered viral core, protected from the cellular environment, was supported by the discovery of the MLV capsid protein within preintegration complexes (Bowerman et al., 1989; Fassati & Goff, 2001). However, subsequent studies demonstrated that early post-entry steps of HIV-1 differed from that of MLV, as the reverse transcription complexes (RTC) isolated from HIV-1-infected cells contained only trace amounts of CA (Fassati & Goff, 2001; Karageorgos et al., 1993; Miller et al., 1997). A series of elegant studies demonstrated that the in vitro reverse transcription of purified HIV-1 virions was greatly enhanced in the presence of physiological fluids (Zhang et al., 2000; Zhang et al., 1996; Zhang et al., 1998). In this so called endogenous reverse transcription assay, the virions were shown to undergo profound morphologic changes of the viral core in the absence of detergents. Together, the above results suggested that reverse transcription is likely to coincide with the process of uncoating and to involve cellular proteins (Nermut & Fassati, 2003). Results of the recently published EM studies of HIV-1 reverse transcription complexes support this hypothesis (McDonald et al., 2002; Nermut & Fassati, 2003). These studies revealed that the RTCs were large nucleoprotein complexes containing numerous proteins, and that they resembled fibers of histone H1-depleted chromatin.

It is clear that the involvement of cellular factors in reverse transcription is just beginning to be unraveled. Since reverse transcription and uncoating seem to coincide in time within the same cellular compartment, it is possible that host factors involved in the uncoating may also contribute to reverse transcription and vice versa. It is also possible that reverse transcription is functionally linked to other steps of replication through common use of cellular factors.

Formation and Nuclear Translocation of Pre-Integration Complexes

The process that follows reverse transcription involves assembly of the newly synthesized cDNA molecules into preintegration complexes (PICs). HIV-1 PICs have been characterized as large nucleoprotein complexes that, in addition to the viral nucleic acids, contain IN, RT, Vpr, MA and several known and unknown cellular components (Fouchier & Malim, 1999; Sherman & Greene, 2002). The host cell factors that have been shown to be part of PICs include HMG(I)Y, BAF, and LEDGF.

The ability of HIV-1 PICs to actively penetrate the nuclear membrane to gain access to the nucleus of the cell defines the uniqueness of HIV-1 and other lentiviruses, since they can infect non-dividing cells (Lewis & Emerman, 1994). Unlike lentiviruses, oncoretroviruses require disintegration of the nucleus during mitosis to gain access to the host genome. Lentiviruses have developed a means of actively transporting PICs through the nuclear membrane and, although the mechanism of such transport remains elusive and highly controversial, it is clear that it relies heavily on the cellular machinery. The simple fact that the size of the PIC is over twice the size of the nuclear pore complex (NPC) channel argues against the passive passage of the PIC into the nucleus. Indeed, studies have demonstrated that the nuclear translocation of PICs involves an active ATP-dependent mechanism (Bukrinsky et al., 1992). Several viral proteins have been identified to potentially mediate nuclear import of PICs. IN, MA and Vpr all contain nuclear localization sequences that have been implicated in the nuclear localization of PICs (Bukrinsky & Haffar, 1999; Bukrinsky et al., 1993). The question of requirement of the nuclear localization signal of these factors in this process remains highly debated and controversial (Bouyac-Bertoia et al., 2001; Dvorin et al., 2002; Fouchier et al., 1997; Gallay et al., 1997; Sherman & Greene, 2002).

The functional components of the NPC such as importins probably play an active role in the translocation of the PICs into the nucleus of the target cell (Gallay et al., 1997; Gallay et al., 1996). Importins are known to specifically recognize the nuclear localization signal of cytoplasmic proteins and can mediate their transport through the NPC by utilizing the gradient of RanGTP to RanGDP that forms across the nuclear membrane. It remains to be determined whether this classical import mechanism is utilized by HIV-1 PICs.

Transcription of the integrated HIV-1 provirus is a crucial step in the viral lifecycle that involves a highly regulated interplay between the virus and the host apparatus (Kingsman & Kingsman, 1996; Pereira et al., 2000). Various cell type-specific transcriptional activators as well as suppressors of transcriptional inhibitors are utilized to induce maximal viral transcription. Such an ability by the virus to adapt to the cell-specific environment allows for the rapid viral replication in a variety of cells, and at the same time ensures maintenance of latently-infected reservoirs that are activated for transcription only when the need arises.

The major viral regulator of proviral transcription is the Tat protein, which potently enhances gene expression by binding directly to the TAR region of the HIV-1 promoter, the long terminal repeat (LTR) (Gatignol et al., 1996; Gatignol & Jeang, 2000; Gaynor, 1995). The ability of Tat to activate transcription is driven by its association with a number of cellular transcription factors. One crucial cellular protein required for Tat transactivation activity has been identified as CyclinT1, which directly associates with Tat. This association, in turn, recruits a cyclin-dependent kinase (CDK9) to the TAR sequence, and leads to the formation of transcription elongation complex, known as positive transcription elongation factor b (P-TEF-b) (Garber et al., 1998; Wei et al., 1998). Formation of this complex results in subsequent hyper-phosphorylation of the C-terminus of RNA polymerase II, and leads to the efficient elongation of nascent RNA molecules.

In addition to Cyclin T1, Tat directly interacts with the cellular transcription factor Sp1 (Kamine & Chinnadurai, 1992; Kamine et al., 1991). Interaction between the Tat and Sp1 proteins seems to be required for the transactivation ability of Tat in all types of cells. Moreover, Sp1 itself exerts both cell-specific, and differential state-specific transcriptional activity on the proviral transcription, as it acts as an antagonist to another Sp-family factor, Sp3, which is known to repress LTR-driven transcription in certain cells. Interestingly, Sp1 also bypasses the requirement for Tat/Tar to drive proviral transcription by the direct recruitment of Cyclin T1 to the LTR (Yedavalli et al., 2003). A similar activity has been ascribed to another key cellular factor demonstrated to transactivate the LTR, nuclear factor κB (NF-κB). The NF-κB transcription factor is involved in multi-level regulation of proviral transactivation. During HIV-1 infection, a cell-specific activation of the NF-κB pathway by the virus leads to the liberation of its two subunits, p50 and p65, from the cytoplasmic inhibitor IκB, and the translocation of NF-κB to the nucleus, where it binds to the specific target sequences of the LTR and stimulates transcription (Rabson & Lin, 2000). Induction of LTR transcription by NF-κB has been linked to acetylation of NF-κB by the cellular coactivator p300 (Furia et al., 2002). Coactivator proteins such as CBP/p300 and pCAF are thought to mediate the Tat-induced transcriptional activation of the LTR promoter through the remodeling of HIV-1 containing chromatin structure. These factors also seem to mediate acetylation of the Tat itself and thus promote its activity.

Among the growing list of other cellular factors contributing to the establishment of equilibrium in the transcription of the proviral genome are such transcription factors as nuclear factor of activated T-cells (NFAT), which bind to an enhancer element in the LTR in T cells (Cron et al., 2000). The differential effect by two types of NFAT proteins, NFAT-1 and NFAT-2/NFAT-c allows either synergistic action with NF-κB and Tat to stimulate transactivation in cells such as activated T lymphocytes or inhibition of the NF-κB-mediated transactivation in resting lymphocytes (Mouzaki et al., 2000).

Inhibition of LTR transactivation involves a number of cellular factors and can be viewed as both a way for the cell to mount an anti-viral response and as a potent mechanism to establish latency. Several key transcription factors that have been identified to counteract Tat-mediated transactivation include tumor suppressor p53, which has been shown to associate with Tat, transcription factor Ying Yang1 (YY1) and lysofylline (LSF) (Coull et al., 2000; Duan et al., 1994; Li et al., 1995). The latter two cellular factors inhibit HIV-1 expression by the recruitment of the histone deacetylase activity to the LTR, which counteracts TAT-driven LTR transactivation (Coull et al., 2000).

Assembly and Budding

Upon successful export of viral RNA molecules into the cytoplasm, viral proteins are synthesized from the mRNAs in the form of large polyprotein precursor molecules. These polyprotein precursors are then assembled into the nascent viral particles, along with the viral genomic RNA and cellular proteins, (Gottlinger, 2001; Kaplan, 2002). Although it is clear that the assembly involves cellular factors, the identities of many of these factors, as well as the extent of their involvement, remain widely unknown. The major viral determinant in the assembly is considered to be Gag polyprotein, since the expression of the retroviral Gag protein itself is sufficient for the formation and release of virion-like particles (VLPs) (Gheysen et al., 1989). The process of assembly is arbitrarily divided into the following steps, some of which are likely to occur concurrently: 1) multimerization of Gag molecules; 2) binding of Gag complexes to viral genomic RNA; 3) formation of Gag- Gag/Pol complexes mediated by Gag; 4) formation of high order complexes containing viral proteins and host cell proteins required for assembly and budding; 5) cytoskeleton-directed transport of preassembled complexes to the inner plasma membrane of the cells.

Cellular factors that influence viral assembly include HP68 and Tumor Susceptibility gene 101 (Tsg101). HP68, an RNAse L inhibitor that was purified from the in vitro reconstituted assembly system, is suggested to promote the assembly of Gag into immature virus particles. It associates with Gag, Gag-Pol and Vif, and is selectively incorporated into the preassembled viral complexes (Zimmerman et al., 2002).

TSG101 protein, coupled with other cellular factors of the endosomal sorting machinery, such as Vsp28 and Vsp4, has been shown to bind to the PTAP motif of the Gag p6 to promote budding of the virus particle. Studies preceding identification of TSG101 demonstrated that the L domain of Gag that contains the conserved PTAP motif was responsible for the efficient budding of the nascent virions from the host cell. Mutations in this region caused blockage of the virions at the stage of budding and caused accumulation of Gag into VLPs that do not detach from the plasma membrane (Dorfman et al., 1994; Gheysen et al., 1989; Huang et al., 1995). The L domain was shown to facilitate a late budding process at the stage of membrane fusion required for the release of the assembled particles. Several studies, such as functional swapping of the L domains of various retroviruses, and placement of the p6 region into various locations within Gag suggested that the mode of action by which the L domain promotes budding is through its protein-protein interactions with cellular factors. The L domain was further shown to associate with the cellular ubiquitination and endosomal sorting machinery, and one factor from this pathway, TSG101, was identified to directly bind to p6 (Garrus et al., 2001; VerPlank et al., 2001). The cellular functions of TSG101 and associated Vsp proteins in the endosomal sorting pathway have been proposed to be utilized by the virus to promote budding. The mechanism by which HIV-1 employs the cellular endosomal sorting machinery, and specifically TSG101, has not been completely elucidated. The current favored hypothesis is that the L domains recruit the cellular endocytosis machinery to the site of budding to promote the specific sorting of the ubiquitinated cargo into the forming viral particles. Tsg101 has structural and sequence similarity to the ubiquitin-conjugating enzymes, and the efficient release of virus particles requires the ability of TSG101 to bind both PTAP and ubiquitin (Goff et al., 2003). Furthermore, the affinity of Tsg101 for Gag is enhanced when Gag is ubiquitinated (Garrus et al., 2001; Martin-Serrano et al., 2001; Pornillos et al., 2002). Results of the mutational analyses of Tsg101 are also in support of the above hypothesis, as Tsg101 truncation mutants inhibit budding of the virus by binding to the L domain and/or by disrupting the endosomal sorting pathway of the host cell (Goila-Gaur et al., 2003).

INI1/hSNF5

Late events of HIV-1 replication also seem to involve another cellular factor, INI1/hSNF5, which was originally identified as a host protein that specifically binds to HIV- 1 IN (Kalpana et al., 1994). INI1 is homologous to yeast SNF5 and is a component of the chromatin remodeling hSWI/SNF complex (Kalpana et al., 1994; Wang et al., 1996). SWI/SNF and related complexes are evolutionarily conserved, multi-subunit, high molecular weight (>2MDa), ATP-dependent complexes that regulate transcription by chromatin remodeling in eukaryotic cells (Carlson & Laurent, 1994; Peterson, 1998; Peterson & Tamkun, 1995; Wolffe, 2001). The function of INI1/hSNF5 within the context of SWI/SNF complex and its exact role in chromatin remodeling is presently unknown.

Creating a transdominant mutant can genetically test the role for a host protein in viral replication. A study has demonstrated that a truncation mutant of INI1 (termed S6) that harbors the minimal IN-interaction domain had a dramatic effect on the HIV-1 particle production, exhibiting up to 100,000-fold inhibition in 293T cells (Yung et al., 2001). Furthermore, it was found that expression of this mutant in T-cells protected these cells from the spread of the virus. These studies indicate that the inhibition is mediated by the direct protein-protein interaction between S6 and IN. Furthermore, it was found that HIV-1 particle production is reduced in INI1-deficient cells and that INI1/hSNF5 complements the defect (Yung et al., 2001). The above results suggested that INI1/hSNF is involved in the late stages of the viral lifecycle, although its precise role, as well as the mechanism of its action remains unclear. In addition, during the course of above studies, it was found that INI1/hSNF5 is incorporated into HIV-1 virions. Furthermore, incorporation of INI1/hSNF5 was specific to HIV-1 and was dependent on its ability to specifically interact with HIV-1 IN, indicating that HIV-1 may have evolved to utilize this host factor during its replication (Yung et al., 2004).

Until recently, the responses of the humoral and cellular immune systems were considered to be the major driving forces of mammalian host defence against viral pathogens including HIV-1. However, recent studies on the intracellular host-virus interactions illustrate the fact that host restriction factors strongly oppose HIV-1 replication within the cell and implicate the intricate dynamics in the battle between the host and HIV-1. These studies also illustrate the ingenuity of HIV-1 in its ability to hijack host cellular factors to assist in its propagation. It is clear that we do not yet fully understand all the players and the mechanisms involving host cell defence systems, and it is likely that there are many more host factors that mediate the intracellular immunity against infection by HIV-1. Further studies to elucidate the mechanism of intracellular host defence will not only provide new insights into this ancient battle between the host and the virus, but may also lead to the development of novel strategies to combat the onslaught by the virus. These novel strategies could include an array of interventions that could either mimic the natural defence mechanisms of the host or disrupt the adaptive mechanism of the virus, providing us with a powerful arsenal for the war against the currently devastating AIDS pandemic.