Research Focus: HIV-1 Reverse Transcriptase and Integrase
We have a long-standing interest in HIV-1 reverse transcriptase (RT) and, more recently, have begun work on another essential viral enzyme, integrase (IN). Although much of our work is focused on understanding the mechanisms that underlie drug resistance, and using that information to develop more effective anti-HIV drugs, we are, in both the RT and IN projects, doing experiments that are intended to better understand the roles that these enzymes play in viral replication. We are also using redirected HIV integration as a tool to investigate chromatin structure and function and a tool to investigate the nature and origin of the HIV reservoir in patients on antiretroviral therapy.
Background. HIV-1 is the causative agent of AIDS. The three viral enzymes — RT, IN, and protease (PR) — have essential roles in the replication of HIV-1 and are the targets for all of the most potent anti-HIV drugs. Although considerable progress has been made in treating HIV-infected patients with three- and four-drug regimens, there is an immediate need for the development of effective ways to prevent new infections. A potent preventive vaccine would be ideal; however, despite a huge effort, the goal of developing an effective vaccine remains elusive. In the absence of an effective vaccine, reducing the transmission of HIV-1 must rely on barrier methods and/or drug treatments. There are two ways that anti-HIV drugs can be used to reduce viral transmission: (1) effective therapy in infected patients can reduce the viral load, making it less likely that an infected individual will transmit the virus to a partner; and (2) treating the uninfected partner with an anti-HIV drug can block transmission. Because most new infections are caused by a single virus, blocking transmission is an attractive option and there is now good evidence that giving an anti-HIV drug to the uninfected partner can significantly reduce viral transmission if the uninfected partner is compliant. Because of the problem of drug resistance, it would be better to use drugs with nonoverlapping resistance profiles for treatment and prophylaxis. Treatment would have to be long term and, for this reason, drug toxicity is an important consideration, which argues against the use of nucleoside RT inhibitors (NRTIs). It would also be better to block infection before the viral DNA is integrated, which argues against the use of PR inhibitors. Therefore, the two remaining options among the major classes of anti-HIV drugs are nonnucleoside RT inhibitors (NNRTIs) and IN strand-transfer inhibitors (INSTIs). We are using a combined approach that involves structural analysis, biochemistry, virology, modeling, toxicity testing, and chemistry to design, synthesize, and evaluate new NNRTIs and INSTIs. We have made good progress in developing new compounds that are effective against the wild-type (WT) and common drug-resistant viruses and that have good therapeutic indexes in tests done in cultured cells.
HIV-1 RT. There are two clinically important classes of inhibitors of HIV-1 RT: NRTIs and NNRTIs. In the past, a major focus of our work has been on the mechanism(s) of RT inhibitor resistance. There is, at this point, a reasonably good understanding of the mechanism(s) of NNRTI resistance, and considerable progress has been made in understanding NRTI resistance, although some important issues remain. We are continuing to investigate the mechanisms of resistance to nucleoside analogs (NRTIs), and we are also interested in elucidating the mechanism of action of compounds that do not cause an immediate or complete block to DNA synthesis. However, in terms of developing new anti-HIV compounds, we have shifted our focus to NNRTIs. Our chemistry collaborators, Drs. Joel Schneider, Gary Pauly, and Craig Thomas, have developed several promising compounds that are able to inhibit, in a one-round replication assay, WT and several common NNRTI-resistant viruses with IC50s below 5 nM. In cultured cells, these compounds have CC50s that are more than 4 logs higher than their IC50s. Additional compounds are being designed, synthesized, and tested. Recently, our collaborator, Dr. Zandrea Ambrose, selected a novel resistance mutation with one of our new NNRTIs. This mutation has some interesting properties (for example, it appears to make the RT of some, but not all, strains of RT temperature sensitive), and we are attempting to understand the mechanisms that underlie the effects of the new mutation.
In addition to the experiments designed to understand resistance to anti-RT drugs and to develop new RT inhibitors, we are studying the effects of RT mutations on the stability of RT in virions and on the fidelity of HIV-1 replication. We showed that a large percentage of the mutations we tested in the thumb subdomain make RT susceptible to PR cleavage; in some cases, this susceptibility creates a temperature-sensitive phenotype. We proposed that the mutations that lead to PR susceptibility partially unfold RT, exposing sites where PR can cleave. The data show that mutations that affect the stability of RT can have a significant impact on the ability of the virus to replicate, and by extension, on viral fitness. We recently showed that mutations that affect the stability of RT are found only very rarely in the Stanford database; this shows that these mutations not only affect the fitness of the virus in cultured cells, but also affect the fitness of the virus in patients.
We are using a LacZalpha complementation assay similar to the assays used by the Pathak and Mansky labs to measure the fidelity of HIV-1 replication in cultured cells. We improved the efficiency of the assay, which allows us not only to measure the mutation rate, but also to determine the positions in LacZalpha where mutations frequently arise (hotspots). The data show that (1) all of the published in vitro assays using purified RT overestimate the in vivo error rate and fail to correctly identify the mutational hotspots; (2) HIV-1 replication is not more error prone than the replication of other retroviruses; (3) which strand of LacZalpha is in the RNA genome does not affect the overall error rate, or the types of errors made, but it does affect the hotspots; (4) mutations in RT (including mutations that confer resistance to NRTIs) affect which sites are mutational hotspots.
We also showed that some of the compounds that Dr. Terrence Burke prepared that were intended to inhibit HIV-1 IN can inhibit the polymerase and/or the RNase H activities of HIV-1 RT. We are studying the effects of the compounds using purified HIV-1 RT, and HIV-1-based viral vectors, and our collaborator, Dr. Arnold, is preparing co-crystals of RT and three of the compounds.
HIV-1 IN. Like RT, HIV-1 IN is an important drug target; however, as is the case for all anti-HIV drugs, treatment with INSTIs leads to resistance. We are making good progress on two fronts: understanding how mutations in IN confer resistance to the currently available compounds, and developing new INSTIs that are effective against the common drug-resistance mutations. Dr. Terrence Burke is synthesizing new anti-IN compounds; Dr. Yves Pommier is using biochemical assays to test Dr. Burke's anti-IN compounds in vitro (using purified recombinant IN); and we are testing how the new inhibitors affect viral replication and measuring their toxicity in cultured cells. Dr. Peter Cherepanov has joined our collaborative effort and has solved the structures of prototype foamy virus (FV) IN in complex with some of the more promising compounds developed by Dr. Burke. The active site of FV IN is similar, but not identical, to the active site of HIV-1 IN, and we used Dr. Cherepanov's data to develop models of HIV-1 IN (both WT and mutant). These models have helped us understand how resistance arises and have been useful in the design of more effective compounds. Dr. Burke has recently synthesized several novel compounds that have IC50s in the low nanomolar range in a one-round replication assay, and that effectively inhibit both WT IN and most of the common drug-resistant variants. Based on tests done in cultured cells, these compounds have excellent therapeutic indexes (the CC50s are more than 3 logs higher than the IC50s). With help from the NCI, some of the most promising of the new compounds are being tested for uptake, stability, and toxicity in animals.
We also have three projects that involve studies of HIV-1 integration. In the first project, we are working with Drs. Alan Engelman and Vineet KewalRamani to define the host factors that are involved in transporting the preintegration complex (PIC) and to determine the exact roles they play in this process. In the second project, we are taking advantage of the fact that it is possible to redirect where HIV-1 DNA preferentially integrates. Redirecting HIV-1 DNA integration has the potential to make gene therapy safer because it may help solve the problems associated with the insertional activation of oncogenes; in addition, the technology can be used to determine where in the genome proteins/domains bind to chromatin. Different retroviruses have different integration-site preferences. There are good reasons to believe that such preferences are based on which host factor(s) the PIC interacts with; we now have information about the host factors that are targeted by HIV-1 IN and murine leukemia virus (MLV) IN. HIV-1 IN is known to bind to lens epithelium-derived growth factor (LEDGF); the distribution of LEDGF on chromatin is the major factor that determines the local sites where HIV-1 DNA integrates. However, working with Dr. KewalRamani, we showed that the host factor CPSF6, which binds to the viral CA protein, helps direct the PIC to nuclear speckles, which are associated with regions of the genome that are enriched in highly expressed genes. Thus, the interaction of CA and CPSF6 directs the PIC to broad regions of the genome, and the interaction of LEDGF and IN determines the exact sites where the HIV DNA is integrated.
We, and others, showed that replacing the N-terminus of LEDGF with chromatin-binding domains (CBDs) from other proteins changes the specificity of HIV-1 DNA integration. The initial experiments were done either with single CBDs or, in one case, with two linked domains taken from a larger protein. These analyses showed that the binding sites for CBD can be accurately determined by mapping redirected HIV-1 integration sites and that the distance between the CBD binding site and the integration site(s) is relatively small. We also showed that the binding sites for multiple-domain modules reflect the combinatorial interactions of the individual domains and chromatin and that the structural relationship of the domains helps define binding specificity. We have moved from an analysis of isolated CBDs to intact proteins. Although not every protein we have tried has worked, we have been able to get excellent data with a number of relatively large proteins. Last year we published a paper that explored how TAF3 interacts with its binding sites on chromatin using mutants with altered binding specificities (the TAF3 experiments are part of a collaboration with Dr. Robert Roeder). This year we have published two additional papers in which we mapped the genome-wide distribution of the binding sites for the proteins in the integrator complex (with Drs. Jeffrey Skaar and Michele Pagano) and WT and mutant DNMTs (with Drs. Kyung-Min Noh and C. David Allis).
In the third project, we are working with Drs. Xiaolin Wu, Frank Maldarelli, John Coffin, John Mellors, and Mary Kearney to determine the distribution of integration sites in HIV-infected patients. Last year we showed that there is extensive clonal expansion of HIV-infected cells in patients, and that, in some cases, integration of HIV DNA in specific oncogenes (MKL2 and BACH2) can contribute to this clonal expansion. However, there was a question of whether any of the clonally expanded cells carried an infectious provirus. This year, we have been able to show not only that a highly expanded clone carries an infectious provirus, but also that cells of this clone released detectable amounts of virus into the blood of the patient. We are currently asking whether all of the cells in the clone are making a small amount of virus, or whether (as we suspect) only a fraction of the cells are making virus at any one time. We are analyzing additional infected clones to see what fraction carry replication-competent proviruses. We are also looking at the distribution of integration sites in acutely infected patients, and have begun a collaboration with Dr. Jeffrey Lifson and his colleagues (Leidos, Frederick) to develop a monkey model for integration site analysis.
Last modified: 4 February 2019