細菌致病因子的基因體學分析

Abstract

這個論文由三個針對細菌致病因子的研究構成。首先在第一個部份，是克雷白氏菌(Klebsiella pneumoniae)裡一個重要的致病性巨型質體(plasmid) pLVPK的核酸定序以及基因註解(annotation)工作。我們定序了這個全長21.9萬鹼基配對(base pair)的巨型質體並且在其中註解出251個開讀框架(ORF; open reading frame)。在這些可能的基因之中，我們發現了許多明顯可能與細菌致病性有關的毒性基因(virulence gene)，其中包括了控制莢膜多醣體(CPS)合成的rmpA2以及它的同源基因rmpA，各種的抓鐵系統基因如iucABCD-iutA，iroBCDN，fepBC，fecIRA等。此外，我們也在這個質體的核酸序列中發現了與其他細菌中負責產生對於銅、銀、鉛、碲等的抗性基因組(gene-cluster)相似的數個基因組。在質體上主要的基因組之間，我們總共發現了十三個由插入序列(insertion sequence, IS)組成的區域，這些插入序列使我們相信這個大質體極有可能是經由一系列水平轉移(horizontal-transferred)基因組的重組而形成。 在論文的第二部份，我們在綠膿桿菌(Pseudomonas aeruginosa)的全基因體序列中針對一類重要的毒性因子—雙分子調控系統基因(2CS; two-component system)—的演化進行分析。我們分析了這些基因的排列方式以及它們所編碼(encode)的蛋白質序列的功能區塊(domain)組成。再經由比對感受子(sensor)與反應子(regulator)這兩個雙分子調控系統的組成單元的演化樹，我們可以推論至少有一半以上的雙分子調控系統它所包含的兩個單元之間有很明顯的共演化(co-evolution)現象。我們也發現，以上的共演化特徵，在一群帶有與OmpR相似反應子的雙分子調控系統中顯得特別的明顯。相反的在一些其它的雙分子調控系統中，特別是那些帶有與NarL相似反應子的，這種特徵就比較不明顯。在分別針對感受子與反應子所進行的分組之間找到的關聯性，也支持了以上的結果。此外，證據也顯示六群雙分子調控系統極有可能分別演化來自六個共同的來源。從鄰近基因的功能看來，這些雙分子調控系統基因非但是經由基因組整個的重製(duplication)而產生，它們還很可能在重製之後仍然保有相同的功能。我們更進一步分析並比較它們開讀框架前端的未轉錄序列(untranslated sequence)時，發現細菌可能對這些基因組採取不同的基因轉錄調控來避免功能的重疊。 最後一個章節中，我們則利用HMMER從38個已經完成基因體序列定序的微生物中找出並且分析其中OmpR家族雙分子調控系統的演化。OmpR家族雙分子調控系統在這些物種之中的分布也支持了雙分子調控系統源自細菌而後傳播至其他物種的理論。一般而言，OmpR家族的雙分子調控系統，它們的基因排列都是呈現一個從反應子到感受子(RS)的排列方式。我們分別分析了屬於RS與SR(與RS恰好相反的基因排列順序)的兩群雙分子調控系統，發現它們感受子上，接受磷酸基修飾(phosphorylated)的組胺酸殘基(histidine residue)附近的序列都非常的相似。這意味著雙分子調控系統的兩個組成單元之間的蛋白質－蛋白質作用很可能約束(constrain)了蛋白質上特定功能區域的演化。這個現象也同時說明了感受子與反應子之間的作用應該發生於兩個單元在演化歷史中組合成為成RS或SR基因組之前。在這些基因組之間，還保留著相當相似的基因排列順序及蛋白質功能區域，這成為雙分子調控系統共演化的強力佐證。

The thesis covered three major approaches aimed to identify the property of bacterial virulence determinants. In the first part, we determined the entire DNA sequence of pLVPK, a 219-kb virulence plasmid harbored in Klebsiella pneumoniae. A total of 251 open reading frames were annotated. The obvious virulence-associated genes carried by the plasmid are the capsular polysaccharide synthesis regulator rmpA and its homolog rmpA2, and multiple iron-acquisition systems, including iucABCD-iutA and iroBCDN siderophore gene clusters, fepBC ABC-type transporter, and fecIRA that encodes iron uptake regulatory system. In addition, several gene clusters homologous with copper, silver, lead, and tellurite resistance genes of other bacteria were also identified. The presence of thirteen insertion sequences located mostly at the boundaries of the aforementioned gene clusters suggests that pLVPK was derived from a sequential assembly of various horizontally-acquired DNA fragments. In the second part, we analyzed the complete genome sequence of P. aeruginosa PAO1 to unravel the evolution of a group of important virulence factors, the two component systems. Gene organization and functional motif analyses of the 123 two component system (2CS) genes in Pseudomonas aeruginosa PAO1 were carried out. By comparing the phylogenetic trees built respectively for the two components, we showed that more than half of the sensor-regulator gene pairs, especially the 2CSs with OmpR-like regulators, are derivatives of a common ancestor, and have most likely co-evolved through gene pair duplication, while several of the 2CS pairs, especially those with NarL-like regulators, appeared to be relatively divergent. Correlation of the classification of sensor kinases and response regulators further provides support for these models. We have identified six congruent clades, which represent the group of the most recently duplicated 2CS gene pairs. Sequence comparison showed that certain paralogous 2CS pairs may carry a redundant function even after a gene duplication event. However, comparative analysis of the putative promoter regions of the paralogs suggested that functional redundancy could be prevented by a differential control. Finally in the third part, we analyzed 38 completed genomes using HMMER to identify the putative 2CS components and investigate the evolution of the 2CSs of OmpR-family. The distribution of OmpR-like response regulators among different genomes of different taxonomy groups also supported the hypothesis that 2CSs are originated in the last common ancestor of bacteria and subsequently passed to the other species. Mostly, the 2CS genes containing an OmpR-like regulator-encoding gene were found in the order of regulator-to-sensor (RS). The amino acid sequences around the phosphorylated histidine residue of the sensor kinases from either RS or SR (sensor-to-regulator) 2CSs were nearly identical. This suggested that the interaction of 2CS component may have constrained the sequences of the interacting domain between sensor kinase and the cognate response regulator while the ancestral components were brought together during evolution into a RS or SR gene cluster, where coordinated transcriptional control may be economically favored. The nearly invariant gene order and the conservation of catalytic domains of these 2CSs provide strong evidence for the co-evolution of 2CSs.