以RdRp酶为例通过蛋白结构比对研究RNA病毒进化历史的方法学探讨

Inferring the evolutionary history of RNA viruses using protein structure alignment:An RdRp perspective

目的以RdRp酶为例探讨病毒蛋白的保守性以及通过蛋白结构比对研究RNA病毒进化历史的可能性。方法从GenBank下载6个病毒科RdRp酶的序列,通过多维尺度分析筛选代表性的序列,使用在线预测工具SwissModel预测三维结构,分别用Dali和Matt软件进行结构两两比对,对比对结果进行聚类分析。筛选的序列同时进行系统发育分析。结果同一病毒科内RdRp酶结构比对的分值一般小于1.0,不同病毒科之间结构比对的分值一般大于3.0。对于同一病毒种,分离自不同年份的病毒蛋白结构相似,病毒蛋白结构的差异性与年份无明显相关性。聚类分析表明同一科病毒均聚在一起,但Dali和Matt得到的不同病毒科之间的进化关系并不完全相同,与系统发育树所揭示的进化关系也不完全相同。结论用现有的结构比对方法研究RNA病毒进化所得到的结果并不一致,与系统发育树的结论也不尽相同。因此,通过蛋白结构比对来研究病毒的进化历史,需要开发新的、更复杂的结构比对算法。

Objective RdRp sequences served as an example to explore whether structural alignment of virus proteins could be used to infer the evolutionary history of RNA viruses. Methods RdRp sequences of six virus families were downloaded from GenBank. Representative sequences were selected using multidimensional scaling and were then used to predict structures using the online prediction tool Swiss-Model. Pairwise structural alignment of the predicted structures was performed using Dali and Matt, respectively and the score matrices were analyzed using clustering analysis. In addition, phylogenetic analysis of the representative sequences was also performed. Results Structural alignment scores within a virus family were mostly smaller than 1.0, but those between different virus families were generally greater than 3.0. For a single virus species, protein structures isolated in different years were quite similar, and there was no clear correlation between structures and years in which they were isolated. Although viruses belonging to the same family were always clustered together in cluster analysis, the evolutionary relationships between different virus families obtained from Dali and Matt differed and they also disagreed with results from the phylogenetic analysis. Conclusion Results of the structural alignment techniques used did not agree, and they also differed from results of phylogenetic analysis. There fore, novel and more complex structural alignment algorithms should be developed to infer the evolutionary history of viruses using protein structural alignment.