Structural Studies Of DNA Recombination, Repair, and Rep

基因组DNA必须在细胞分裂的每个周期之前复制。虽然复制型DNA聚合酶有一种内置的校对机制，可以最大限度地减少复制过程中的错误，但偶尔也会发生由于复制错误而导致的不匹配。错配修复系统防止复制错误导致的突变存在于大多数生物体中。大肠杆菌有一个甲基导向的错配修复系统，包括MutS、MutL和Muth蛋白。在人类中也发现了MutS和MutL蛋白的同源物。在90%的遗传性非息肉病性结直肠癌中发现了MutS或MutL同源物的突变。到2001年10月，本课题组确定了错配修复蛋白Muth的晶体结构，MutL的一个保守的40kD的ATPase片段及其与核苷酸的复合体，人MutL同系物PMS2的ATPase结构域，单独与ATP和不可水解性的ATP类似物络合，以及190kD的Taq MutS单独与DNA络合，并作为DNA和ADP/Mg2+的三元络合物。去年(2001-2002)，(1)基于晶体结构，我们构建了47个MutS、MutL和Muth大肠杆菌突变体，进行了体外和体内(与加州大学洛杉矶分校的J.Miller合作)的生化研究，发现每个未能修复错配的突变体在阻止同源重组方面都存在缺陷。我们的研究还表明，MutL的非特异性DNA结合特性是MutS切割DNA后进行错配修复所必需的，并且MutS中的结构域协同工作，实现了异源双链与同源双链的最大差异结合，而不是最大DNA结合。(2)MutL-ATPase活性位点与DNA拓扑异构酶、Hsp90以及细菌和线粒体蛋白激酶具有保守的序列基序。这些蛋白质中有许多是潜在的药物靶点。我们发现，尽管这些酶之间有相似的ATP结合口袋，对镁离子的需求也是相同的，但它们在ATP结合口袋中的一价离子偏好上有所不同。在对MutL和一个大鼠线粒体蛋白激酶进行结构和序列比较的基础上，我们对MutL进行了点突变，并将MutL从任何一种单价离子特异性地转化为Na+。我们还测定了野生型和突变型MutL蛋白的晶体结构，并观察到K+交换的Na+。我们认为单价离子的专一性可能会被用于未来的药物设计。 我们还确定了与DNA结合的半甲基化GATC结合蛋白E.ColiSEQA的晶体结构。这个项目源于我们对Muth如何识别半甲基化的GATC序列并特异性地针对子链的错配修复的研究。SEQA-DNA复合体的结构表明，甲基化腺嘌呤的识别是由蛋白质主链原子通过紧密的van der Waals接触实现的。最有趣的是，这种结构表明了一种机制，即SEQA可以隔离DNA复制起点ORIC，使其不被过早、重复或异步使用。 最近发现了一个新的DNA聚合酶家族--Y-家族。它们与已知的DNA聚合酶在初级序列、病变旁路能力和容易出错的DNA合成方面不同。2001年，我们与NICHD的罗杰·伍德盖特博士合作，确定了来自Solfotaricus的Y家族DNA聚合酶Dpo4的晶体结构，该DNA聚合酶与未损坏的DNA和输入的核苷酸形成了复合体。这些晶体结构提供了Y-家族聚合酶作用的第一个原子视图，并揭示了低保真DNA合成和绕过修饰的DNA碱基的分子机制。最近，我们已经结晶了DPO4与受损或错配的DNA的络合物。新的结构表明，DPO4与传入的核苷酸结合，而不依赖于正确的碱基与模板链的配对，并在催化核苷酸转移反应之前对模板链进行改组，以找到传入核苷酸的合作伙伴。我们认为，这种“自由加载”的核苷酸服务于Dpo4，以绕过DNA损伤，并进行低保真DNA合成。

Genomic DNA has to be replicated before every cycle of cell division. Although replicative DNA polymerases have a built-in proofreading mechanism to minimize errors during replication, occasionally mismatch due to replication-error occurs. Mismatch repair systems to prevent mutations from replicative errors exist in most organisms. E. coli has a methyl-directed mismatch repair system comprising MutS, MutL and MutH proteins. Homologues of MutS and MutL proteins are also found in humans. Mutations in MutS or MutL homologs have been identified in 90% of the hereditary nonpolyposis colorectal cancers. By Oct. 2001, our group determined the crystal structures of mismatch repair proteins MutH, a conserved 40KD ATPase fragment of MutL and its complexes with nucleotides, the ATPase domain of a human MutL homologue, PMS2, alone and complexed with ATP and nonhydrolyzeable ATP analog, and the 190 Kd Taq MutS alone, complexed with DNA, and as a ternary complex with DNA and ADP/Mg2+. Last year (2001-2002), (1) based on the crystal structures, we constructed 47 MutS, MutL and MutH E. coli mutants, carried out in vitro and in vivo (in collaboration with J. Miller at UCLA) biochemical studies, and found that every mutant that fails to repair mismatch is deficient in preventing homeologous recombination. Our studies also revealed that the non-specific DNA binding property of MutL is required for mismatch repair after the DNA incision by MutH and that the structural domains in MutS work cooperatively to achieve maximal differential binding of heteroduplex versus homoduplex rather than maximal DNA binding. (2) The MutL ATPase active site shares conserved sequence motifs with DNA topoisomerases, Hsp90 and bacterial and mitochondrial kinases. Many of these proteins are potential drug targets. We found that despite the similar ATP binding pocket among these enzymes and a shared requirment of Mg ion, each of them differs in monovalent ion preference in the ATP binding pocket. Based on structure and sequence comparison between MutL and a rat mitochondrial protein kinase, we made a point mutation in MutL and converted MutL from using any monomalent ion to Na+ specifically. We have also determined the crystal structures of the wildtype and mutant MutL proteins and observed a Na+ for K+ exchange. We propose that the monovalent ion specificity may be exploited for future drug design. We have also determined the crystal structure of a hemimethylated GATC binding proten, E. coli SeqA, bound to the DNA. This project was derived from our research on how MutH recognizes hemimethylated GATC sequence and targets mismatch repair to the daughter strand specifically. The structure of SeqA-DNA complex reveals that the recognition of the methylated adenine is achieved by protein mainchain atoms via close van der Waals contacts. Most interestingly, the structure suggests a mechanism for SeqA to sequestrate DNA replication origin, oriC, from being used prematurely, repeatedly or asynchronously. A new family of DNA polymerases, the Y-family, has recently been identified. They differ from the previously known DNA polymerases in the primary sequence and in the ability of lesion-bypass and error-prone DNA synthesis. In collaboration with Dr. Roger Woodgate of NICHD, we determined the crystal structure of a Y-family DNA polymerase, Dpo4 from S. Solfotaricus, in complex with undamaged DNA and an incoming nucleotide in 2001. These crystal structures provide the first atomic view of the Y-family polymerase in action and reveal a molecular mechanism for the low fidelity DNA synthesis and bypassing modified DNA bases. Recently, we have crystallized Dpo4 in complex with damaged or mispaired DNA. The new structures suggest that Dpo4 binds an incoming nucleotide independent of correct base pairing with the template strand and shuffles the template strand to find a partner for the incoming nucleotide before catalyzing the nucleotidyl-transfer reaction. We propose tht this "free-loading" of nucleotide serves Dpo4 to bypass DNA lesions as well as to make low fidelity DNA synthesis.