Retroviruses integrate a DNA copy of their genome into host DNA as an obligatory step in their replication cycle. Our work focuses on the molecular mechanism of integration, and in particular on the structure and function of HIV integrase and other proteins involved in retroviral DNA integration. Although integrase carries out the key DNA cutting and joining steps of integration in vitro, integration in the cell is carried out by a large nucleoprotein complex, the preintegration complex (PIC), that is derived from the core of the infecting virion. The PIC contains the viral DNA, integrase and a number of other viral and cellular proteins. The major impediment to further structural, biophysical and further biochemical studies of retroviral DNA integration is that it has not yet been possible to reconstitute a nucleoprotein complex from purified integrase and DNA substrate that exhibits all the properties of PICs isolated from infected cells. We have recently made substantial progress towards this goal. Under most reaction conditions for integration in vitro, the majority of the reaction products are ?half-site? products that result from integration of only one viral DNA end into one strand of the target DNA. Pre-processed DNA substrates are more efficient substrates for half-site reactions than are blunt-ended substrates that require removal of two nucleotides prior to integration. In contrast, we find that blunt-ended DNA is a better substrate for the biologically relevant reaction of concerted integration of pairs of viral DNA ends. The reaction pathway is channeled to concerted integration and half-site integration products are reduced with blunt-ended DNA substrate that must first be processed by integrase. In addition, the terminal nucleotide requirements for concerted integration are more stringent than for the half-site reaction. Longer DNA is more efficient for the concerted reaction than is shorter DNA that is capable of efficient half-site integration. This suggests that non-specific interactions of integrase with viral DNA distant from the termini contribute to assembly of a complex that is competent for concerted integration. Finally, differential effects of mutation of a residue in the C-terminal domain of integrase on concerted versus half-site integration implicate protein-protein interactions involving this domain to be important for concerted integration. Specific complexes of integrase with DNA have not been physically detected except in the context of the PIC made in vivo. Integrase binds DNA non-specifically in simple DNA binding assays and the protein and DNA aggregate under reaction conditions; discrete stable complexes have not been physically detected. The above improvements in the reaction conditions for concerted integration in vitro allowed us to ask whether stable complexes of integrase and DNA participate in concerted integration. We find that for the concerted integration pathway, but not the half-site reaction pathway, integrase makes highly stable nucleoprotein complexes with the viral DNA that parallels the stable association of integrase with the viral DNA in the PIC. This will enable us, for the first time, to determine the nucleoprotein organization of the active complex that mediates retroviral DNA integration. These studies are in progress. Another focus of our work is on the role of cellular proteins in retroviral DNA integration. Elucidation of the role of host factors in retroviral DNA integration is necessary to gain a more complete understanding of the integration process, and to potentially identify targets other than integrase that allow viral replication to be blocked at the step of DNA integration. We have previously identified a cellular protein (BAF) that blocks self-destructive autointegration of retroviral DNA. We have focused on the mechanism by which BAF blocks autointegration and obtained evidence that supports our hypothesis that BAF blocks autointegration by compacting the viral DNA within the preintegration complex. By gel filtration and equilibrium ultracentrifugation analysis, we have demonstrated that a 7 bp DNA is sufficient to make a complex with BAF. The structure of the complex, solved in collaboration with Dr. Dyda?s group, revealed that it consists of a dimer of BAF with DNA bound to each side of the dimer. Thus the bridging of DNA by BAF can be explained by a dimer with one DNA binding site on each monomer. The structure also explains why the interaction of BAF with DNA is sequence non-specific. However, the association of BAF with DNA is less stable than the association of BAF with the PIC. We find that an interacting partner of BAF, lamina-associated-protein-2-alpha (LAP2alpha) is also present in the PIC and that these two proteins collaborate to make a stable complex with DNA. The details of the interactions between BAF, LAP2alpha and DNA are under investigation.
Showing the most recent 10 out of 15 publications
