Structural biology of host factors affecting retroviral integration

1. 逆转录病毒整合的一个奇怪的方面是，为什么首选的目标是宿主细胞DNA而不是逆转录病毒DNA本身，从整合酶的角度来看，逆转录病毒DNA似乎是最容易接近的DNA，并且具有最高的局部浓度。答案似乎与宿主细胞蛋白有关，即自动整合障碍因子BAF (Lee & Craigie, 1998)，它被整合到预整合复合体中。BAF是后生动物中一种高度保守的必需蛋白。它以与序列无关的方式结合和桥接双链DNA，并且BAF似乎与病毒DNA结合，并交联并将其压缩成一种屏蔽整合的形式。虽然已经确定了BAF的结构，但没有结构信息表明它如何与DNA相互作用，并且提出了几种相互排斥的模型。我们决定确定BAF与DNA的结构，因为我们希望它能够深入了解预整合复合体中一个重要的相互作用，并可能提出削弱BAF保护病毒DNA免受自杀性自我整合能力的方法。

1. One curious aspect of retroviral integration is why the preferred target is host cell DNA rather than the retroviral DNA itself, which would seem to be the DNA that is the most accessible and with the highest local concentration from the perspective of the integrase enzyme. The answer appears to lie with a host cell protein, the barrier-to-autointegration factor, BAF (Lee & Craigie, 1998), that is incorporated into the preintegration complex. BAF is a highly conserved and essential protein among metazoans. It binds and bridges double-stranded DNA in a sequence-independent manner, and it appears that BAF binds to the viral DNA, and crosslinks and compacts it into a form that is shielded from integration. Although the structure of BAF had been determined, there was no structural information indicating how it interacts with DNA, and several mutually exclusive models had been proposed. We undertook to determine the structure of BAF with DNA as we hoped it would provide insight into one of the important interactions within the preintegration complex, and might suggest ways to weaken BAF's ability to protect viral DNA against suicidal self-integration. Non-specific DNA binding proteins can be difficult to crystallize as they tend to form heterogeneous complexes in solution. We circumvented this problem by systematically characterizing complexes of human BAF and short DNA oligonucleotides using a range of biophysical approaches. We established that BAF forms a discrete complex with a 7-mer in which a BAF dimer binds two DNA molecules. The structure of this complex reveals that the DNA duplexes are bound at opposite ends of the dimer and are approximately perpendicular to each other. BAF uses only four protein side chains to contact DNA through interactions that involve only the phosphate backbone of DNA, thus ensuring sequence independence. The interactions are mediated by an helix-turn-helix (HtH) motif, a pseudo-HtH motif, assisted by residues located on the first alpha-helix. The observed packing of the BAF-7-mer crystal suggests a model for DNA compaction in which BAF bridges potentially distant regions of DNA. 2. The nuclear lamina is a thin web of proteins assembled from intermediate filaments comprised of A- and B-type lamins that supports the inner membrane of the nucleus. Lamins serve as crucial architectural components and organizers of nuclear structure. Mutations in the LMNA gene that encodes the vertebrate A/C-type lamins give rise to a variety of human diseases known as laminopathies. These range from tissue-specific diseases such as muscular dystrophies to systemic diseases such as Hutchinson-Gilford progeria syndrome which causes premature aging in children. One specific lamin-interacting protein, LAP2alpha, has been identified as a component of retroviral preintegration complexes (Suzuki et al., 2004) and shown to bind to BAF (Furukawa, 1999; Shumaker et al., 2001). Although the role of LAP2alpha in the retroviral lifecycle is not yet clear, it has been reported that the replication of murine leukemia virus is inhibited in LAP2alpha knockdown cells. We initiated our structural studies of domains of LAP2alpha to provide insight into its roles in nuclear structure and dynamics. Mammalian cells have six isoforms of LAP2, but only the alpha isoform has been identified as a component of retroviral preintegration complexes. The alpha isoform is distinguished from the other isoforms by a unique C-terminal domain which binds to the nuclear lamina protein, lamin A/C. We therefore focused our efforts on the LAP2alpha C-terminal domain with the hope that its structure would shed light on its role in nuclear function, and how that role might be important for preintegration complexes. We recently determined the three-dimensional structure of the C-terminal domain of LAP2alpha (CTD), and showed that it forms an elongated dimer in which a pair of antiparallel alpha-helices from each monomer pack together to form a left-handed, four-stranded, antiparallel coiled coil. Although the evidence is clear that LAP2alpha is multimeric, it has been most recently proposed to be a trimer (Snyers et al., 2007). Our structure does not support this, and we suggest that the data indicating a trimeric form could reasonably be interpreted to support a dimeric assembly. The coiled-coil region of LAP2alpha is packed against smaller helices and loops that flank it. We used a solid-phase overlay assay to show that the CTD binds to lamin A/C. We then assessed the roles of various CTD residues in lamin A/C binding by generating 159 single point mutants that were purified in a high-throughput 96-well format and assayed for lamin A/C binding. The vast majority of CTD mutations had no effect, suggesting that the lamin A/C binding surface on LAP2alpha is extensive and involves multiple residues. Only six point mutants had significantly altered lamin A/C binding, and these are in flexible loops that might be expected to form protein-protein interfaces. 3. It has recently become clear that an important host cell protein involved in retroviral integration is lens-epithelium-derived growth factor (LEDGF/p75), a transcriptional coactivator that has been shown to influence target site selection by HIV integrase (Ciuffi et al., 2005; Llano et al., 2006). LEDGF interacts directly with HIV integrase, and has been proposed to act as a tether that links pre-integration complexes to chromatin. Recently, the co-crystal structure of the integrase-binding region of LEDGF and the catalytic domain of HIV integrase has been determined (Cherepanov et al., 2005). However, the region of LEDGF that interacts with DNA and is presumed to direct integrase to its site of action has not been structurally characterized. To this end, we have expressed and purified several constructs encoding single and multiple domains of LEDGF. We are currently attempting to crystallize the DNA binding domain(s) of LEDGF alone and in complex with various DNA substrates. Cherepanov, P., Ambrosio, A.L.B., Rahman, S., Ellenberger, T., and Engelman, A. (2005) Proc. Natl. Acad. Sci. USA 102, 17308-17313. Ciuffi, A., Llano, M., Poeschla, E., Marshall, H., Hoffman, C., Leipzig, J., Shinn, P., Ecker, J., and Bushman, F.D. (2005) Nature Med. 11, 1287-1289. Furukawa, K. (1999) J. Cell Sci. 112, 2485-2492. Lee, M.S. and Craigie, R. (1998) Proc. Natl. Acad. Sci. USA 95, 1528-1533. Llano, M., Saenz, D.T., Meehan, A., Wongthida, P., Peretz, M., Walker, W.H., Teo, W., and Poeschla, E.M. (2006) Science 314, 461-464. Shumaker, D.K., Lee, K.K., Tanhehco, Y.C., Craigie, R., and Wilson, K.L. (2001) EMBO J. 20, 1754-1764. Snyers, L., Vlcek, S., Dechat, T., Skegro, D., Korbei, B., Gajewski, A., Mayans, O., Schofer, C., and Foisner, R. (2007) J. Biol. Chem. 282, 6308-6315. Suzuki, Y., Yang, H., and Craigie, R. (2004). EMBO J. 23, 4670-4678.