Retrovirus Assembly and Maturation

Division Of Basic Sciences - Nci

Investigators

Linked publications, trials & patents

Abstract

Understanding the molecular mechanisms underlying virus particle assembly and maturation is of critical importance for several reasons. First, it can serve as a model for the broader problem of complex macromolecular assembly, with the particular virtue that the relevant structures are foreign to the cell but will be assembled in the cellular context, with the help of cellular constituents. Second, it offers many potential targets for antiviral therapy, which have not been exploited to date because we lack the detailed knowledge needed for the development of therapeutic strategies. We have continued our analysis of the assembly and maturation of HIV-1 particles. Specifically, we characterized the properties of the HIV-1 Gag protein, the fundamental building block of the HIV-1 virus particle. Gag assembles in mammalian cells into the immature virus particle, containing several thousand Gag molecules. Similarly, recombinant HIV-1 Gag protein, purified from bacteria, can assemble into virus-like particles (VLPs) in a defined system in vitro; this shows that the protein will, upon interaction with a single-stranded nucleic acid, engage in rather regular protein-protein interactions, forming a roughly spherical structure with a rather uniform radius of curvature. The arrangement of Gag molecules is the same in particles assembled in vitro and in cells, and many mutants of Gag have analogous effects in the two systems. We study assembly of this protein both in vitro and in mammalian cells, where Gag is obviously far more dilute than in vitro and is surrounded by myriad cellular constituents. While the structures in the assembled particle have been defined to a considerable degree, little is known about the steps by which a Gag molecule is converted from a free, soluble protein to a molecule ready to assemble and then to a constituent of the final, assembled structure. We are addressing the following: What is the structure of an HIV-1 Gag molecule and how does this structure enable it to coassemble, with other Gag molecules and with nucleic acid, into immature retrovirus particles? If the structure of a Gag molecule in solution is different from that in an immature virion, what controls this difference? How is particle assembly triggered by binding to nucleic acid? How are Gag proteins and virions from different retroviral genera similar to each other and how do they differ? Do these differences affect the control of assembly? We have exploited the simplicity of the in vitro system to analyze the molecular changes accompanying HIV-1 Gag assembly. Using a wide variety of experimental approaches, including gel filtration, static and quasi-elastic light-scattering, sedimentation velocity analysis, and small-angle neutron scattering, as well as molecular modeling, we have characterized the conformation of the HIV-1 Gag protein. We found that when Gag molecules are brought into close proximity, they undergo a major conformational change that prepares them for assembly. All of the data indicate that HIV-1 Gag protein is folded over in solution, with its ends near each other in three-dimensional space, in striking contrast to its rod-like shape in immature particles. We have recently focused on conformational changes in the SP1 domain of Gag, a short "linker" between the capsid (CA) and nucleocapsid (NC) domains. Our data show that SP1 assumes an alpha-helical conformation when two or more molecules containing SP1 are brought into juxtaposition. This change occurs because SP1 can form an amphipathic helix; association of several such helices produces helical bundles in which hydrophobic residues face inward, shielded from the solvent. We proposed that this change in SP1 leads to changes in the CA domain and exposes new interfaces for the Gag-Gag interactions that are needed to convert Gag to an "assembly-ready" protein. The data also suggest that the associative interactions between SP1 domains participate, along with those of CA domains, in immature particle assembly. In collaboration with Dr. Eric Freed (HIV Dynamics and Replication Program), we are testing the possibility that immature virus particles, or small SP1-containing helical bundles, will bind maturation inhibitors such as bevirimat; if so, they would be extremely useful for in vitro screens for new inhibitors of HIV-1. Our data in this study also support the hypothesis that maturation inhibitors stabilize these bundles. We have found that IP6 also stabilizes the mature core of the gammaretrovirus Moloney murine leukemia virus (MLV). The role of this cofactor appears analogous to its functions in HIV-1; for example, it stabilizes the core in lysed virions, enabling the reverse transcriptase in virus lysates to copy the genomic RNA into double-stranded DNA. We have shown that MLV particles contain IP6. However, one striking difference between IP6 function in MLV and HIV-1 is that depletion of IP6 from target cells renders them less infectable by MLV, but not by HIV-1. This difference may reflect the fact that gammaretrovirus cores dissociate, at least partially, in the cytoplasm in newly infected cells, whereas it is now understood that HIV-1 cores remain intact until they enter the nucleus. Perhaps the gammaretrovirus cores lose some or all of their IP6 upon dissociation and are thus dependent upon a supply in the new host cell for completing reverse transcription. This contrast between MLV and HIV-1 highlights the specific role of IP6 in retroviral replication. We have also investigated the role of iP6 in immature HIV-1 virus particle assembly, using differential scanning fluorimetry (DSF) in in vitro assembly experiments. Briefly, recombinant HIV-1 Gag protein (lacking the myristate modificatrion at its N-terminus, the p6 domain at its C-terminus, and the large globular head region of the matrix (MA) domain) is soluble, but assembles into virus-like particles (VLPs) upon addition of either nucleic acid or IP6. The mechanism underlying this induction of assembly is not known, and could be kinetic, thermodynamic, or both in nature. As the conformation of Gag is stabilized by intermolecular contacts in a VLP, the assembly is amenable to quantitative analysis using DSF. Particles assembled upon addition of IP6 are denatured at a temperature 10 C higher than unassembled protein, reflecting the stabilization of Gag conformation due to Gag-Gag and Gag-IP6 contacts within VLPs. Studies with Gag mutants yield further information on the mechanism of IP6 action. At physiological ionic strength, the highly charged IP6 molecule binds to many sites on Gag (Datta et al., JMB 2007). This interaction is salt-sensitive, indicating that it is largely electrostatic in nature. We find that induction of assembly depends upon the basic character of the nucleocapsid (NC) domain, even though in assembled VLPs, IP6 is localized in the capsid domain. Cooperativity of the IP6 response is drastically enhanced if the NC domain is replaced with a trimerizing isoleucine-zipper domain. Taken together with other results, the results suggest that IP6-induced assembly is always initiated by the juxtaposition of Gag molecules at their C-termini, either by clustering of basic NC domains around IP6 or by zipper-mediated trimerization. We have also characterized the IP6 response of Gag mutants which lack the lysines in CA that are known to bind IP6 in wild-type VLPs. It is notable that some of these mutants retain some response to IP6 in the DSF assay; in other words, IP6 can still induce assembly of these proteins, even though the IP6 binding site in assembled VLPs is not present. This observation confirms the view that the role of IP6 in assembly is not limited to interaction with these lysines. In ongoing experiments, we are using DSF to compare IP6-induced VLPs with nucleic acid-induced VLPs with respect to stability.