We have established conditions for in vitro assembly of stable synaptic complexes of a pair of viral DNA ends with HIV-1 integrase. These nucleoprotein complexes are intermediates in the integration of HIV DNA into a target DNA. Furthermore, the association of integrase with viral DNA in these complexes mimics all the properties of the association of integrase with viral DNA in preintegration complexes (PICs) isolated from virus infected cells. The synaptic complexes contain a tetramer of integrase tightly bounds to a pair of viral DNA ends. Footprinting of the viral DNA ends within the complex reveals that less than 20 base pairs of terminal viral DNA sequence are protected by integrase. within the SSC, a conclusion that is also in agreement with atomic force microscope images of the stable nucleoprotein complexes. 20bp of terminal viral DNA end sequence are efficient substrates for half-site integration in vitro, and are the only protected region observed in footprinting experiments. However, several hundred base pairs of non-specific flanking DNA sequence are required for efficient SSC assembly, stability and concerted integration. We are probing the function of this non-specific DNA sequence in SSC assembly and stability and propose that non-specific interactions between IN and DNA (distinct from the stable association of a tetramer of IN with the viral DNA ends) are involved. Having established methodologies to assemble synaptic complexes in vitro with purified integrase and HIV-1 DNA substrate, we are attempting both low and high-resolution structural studies combined with biochemical approaches to understand the detailed mechanism of DNA integration. In collaboration with Emilios Dimidriadis and Svetlana Kotova we have confirmed by atomic force microscopy (AFM) that a tetramer of integrase bridges the pair of viral DNA ends in the SSC and obtained evidence that assembly of an integrase tetramer on one viral DNA end is an intermediate step in assembly of the SSC. AFM also reveals that the viral DNA ends are arranged in antiparallel orientation in the SSC. This arrangement of the viral DNA ends is also supported by fluorescence resonance energy transfer (FRET) studies. Cy3 and Cy5 fluorophores have been incorporated at various positions along the DNA substrate and the FRET signal measured upon assembly into the SSC. The results show a significant FRET signal in the SSC when fluorophores are positioned close to the ends of the viral DNA. FRET efficiency was reduced when the fluorophores were located away from the end. AFM has also enabled us to detect a complex of a tetramer of integrase assembled on a single viral DNA end. These Single End Complexes (SECs) complexes are presumably less stable than the intasome and therefore not detected by gel electrophoresis. SECs predominate at early time points of assembly suggesting that they are intermediates in the assembly of intasomes. We propose that a tetramer of integrase first assembles on a single viral DNA and this is an intermediate in the pathway leading to intasome assembly. The potential role of cellular proteins in SSC assembly is under investigation. One cellular protein that has been implicated in playing an important role in HIV-1 DNA integration is Lens Epithelial Derived Growth Factor (LEDGF). We find that LEDGF does not stimulate assembly of the SSC and in fact inhibits complex assembly. LEDGF must therefore be acquired by the preintegration complex after the two viral DNA ends are engaged by integrase to form the SSC. Our goal is directed towards X-ray crystallographic studies of HIV-1 intasomes. The two major obstacles we need to overcome are the propensity of the intasomes to self-associate and the requirement of several hundred base pairs of non-specific flanking DNA sequence for their assembly. The role of the flanking DNA is not understood. Mixing experiments with long viral DNA ends, and short viral DNA ends that do not assemble intasomes alone, show that the short DNA ends are efficiently incorporated into intasomes if the partner is long. The length requirement of several hundred base pairs is similar to the length required for DNA to be easily able to bend back on itself. Our current hypothesis is that is that the flanking DNA transiently associates with the intasome, perhaps occupying the target DNA binding site. Cherepanov and colleagues have recently solved the structure of the prototype foamy virus (PFV) intasome. PFV belongs to the Spumaretrovirus family. Although it shares only 13% identify with HIV-1 integrase and has an extra domain not found in the HIV-1 protein, the arrangement of domains in the HIV-1 intasome is unlikely to be radically different. However, PFV integrase is sufficiently different that structures of the HIV-1 intasome are required to understand the detailed mechanisms of resistance to inhibitors such as Raltegravir. Four of the domains in the in the PFV intasome structure are disordered. We have tested whether these domains are necessary for function of HIV-1 integrase by constructing stable heterodimers lacking the corresponding domains. We find that these heterodimers are inactive for catalysis, indicating that the missing domains are required for function of HIV-1 integrase. The heterodimers are not only inactive for catalysis, but fail to assemble intasomes. The requirement of several hundred base pairs of internal DNA for HIV-1 intasome assembly and stability does not allow us approach structural studies in the way that has been successful with PFV intasomes. Our first approach was to assemble intasomes with long viral DNA substrate and cut off the internal DNA segment after assembly, but we found that the intasomes dissociate after cleavage of this flanking DNA. However, once target DNA is captured and stand transfer has taken place the internal DNA can be cleaved without loss of stability. A major effort is therefore directed at assembling HIV-1 STCs for structural studies. We are using the approach of assembling STCs on branched DNAs corresponding to the product of integration, a strategy we have successfully applied to the PFV system. Our second strategy is to modify the HIV-1 integrase based on the PFV structure to make its biophysical properties more amenable to structural studies. We have succeeded in making a mutant of HIV-1 integrase that is hyperactive for concerted DNA integration by fusing a non-specific DNA binding domain to the N-terminus. The rationale was guided by the presence such a domain in PFV integrase. This hyperactive integrase performs concerted DNA integration of short oligonucleotide substrate with a similar efficiency to that of PFV integrase. Importantly, the hyperactive HIV-1 integrase efficiently assembles intasomes with oligonucleotide DNA substrate, bypassing the requirement for long internal DNA that that been a major impediment for our structural studies of HIV-1 intasomes. The intasomes can be separated from unreacted DNA by gel filtration and are our most promising candidates to date for structural studies of HIV-1 intasomes.
Showing the most recent 10 out of 18 publications
