第一型登革病毒分子流行病學-親緣性演化與群體變化分析

Molecular epidemiology of dengue virus type I - the phylogenetic analysis and population dynamic

簡麗蓉(Li-Jung Chien)

指導教授 ： 何美鄉

共同指導教授 ： 金傳春

國立臺灣大學/公共衛生學院/流行病學研究所/博士(2006年)

https://doi.org/10.6342/NTU.2006.03026

摘要

背景
登革病毒共有四種血清型，其下又可以細分為不同的基因型，每一種血清型別都可能造成登革出血熱/登革休克症候群等嚴重疾病。登革熱是目前熱帶與亞熱帶地區最重要的蚊媒傳染病之一，全球有超過100個國家，估計25億人口居住在登革熱的流行範圍內，每年約有5千萬人受到感染。台灣地區自1987年發生光復後第一次本島性的流行以來，幾乎每年都有本土性的流行發生，主要流行地區是在台灣南部。
方法
本研究以反轉錄酶-聚合酶鏈鎖反應(RT-PCR)與直接定序(direct sequencing)的方式，建立台灣地區1987-2004年收集共計88株本土流行與境外移入第一型登革病毒株自5’端非編碼區與整個結構基因的核酸序列資料庫。並將這些資料與美國疾病管制中心提供的40株全球病毒株、由NCBI基因庫搜尋所得的95株病毒株、以及澳洲實驗室提供的6株病毒株的外套膜基因(envelope gene)序列共同進行後續的親緣性演化與群體變化分析。序列完全相同的病毒株都先被移除，僅保留一個代表株進入分析。親緣性演化分析採用最大似然法(maximum likelihood, ML)與貝氏方法(Bayesian method)兩種邏輯進行演算，使用的程式分別是PHYLIP及MrBayes 3.1。本研究並運用BEAST程式的Bayesian skyline plot方法推估第一型登革病毒在全球主要流行區的群體變化歷程，並同時估算基因的演化速率(substitution rate)。以PAML程式分析第一型登革病毒在演化的過程中是否有正向選擇(positive selection)的介入，造成胺基酸的改變。同時以CodonW程式分析不同時期或不同基因型別的病毒株在密碼子(codon)的選取上是否有顯著差異。
結果
根據親緣性分析的結果顯示，目前世界上流行的第一型登革病毒主要可區分為三個基因型，每一基因型別各有其特殊的地理分布，且隨著時間的演變，可以觀察到支系絕種(linage extinction)的情形發生。臺灣地區的本土第一型登革病毒主要來自基因型I和II，且具同一流行季的病毒株相似度高(p-distance<0.3%)，但發生年代相連的不同流行間的基因序列差異度大(p-distance>1%)的特性，顯示第一型登革病毒尚未在台灣本土化。同時，藉由基因資料庫的建立，配合基因序列可呈現病毒株的地理與時間分布的特性，本研究發現2004年在台中發生的單一本土病例係因實驗室感染所致。由185株第一型登革病毒株的外套膜基因資料庫分析，計算出病毒演化速率約為每年每個位置平均發生0.743x10-3 個變異(substitution/site/year)，95%最高事後密度區間(highest probability density, HPD)介於0.663-0.824 x10-3。現有病毒株的最近共同祖先(most recent common ancestor, MRCA)發生在距離西元2004年約98年前(95%HPD介於88.1~106.7年前)。族群變化分析發現，第一型登革病毒的三種基因型別在過去40年間都曾經歷過明顯的族群擴增，最近一次大約發生在西元2000年。但在外套膜基因的分析中，並未發現特定支系或胺基酸位置有正向選擇的情形發生，歸納其中所有可區分基因型I與II之主要或次要分支(clade and subclade)的胺基酸變異，也都是不會影響蛋白質特性的保守性的改變(conservative changes)。然而在不同基因型別間，卻可觀察到密碼子選用(codon usage)程度上的差異。在密碼子第3位置的GC含量上(G+C3s)，以基因型I為最低(0.404±0.006)，其次是基因型II (0.432±0.007)，基因型III的含量最高 (0.448±0.009)。而在密碼子選用普遍度的指標上(Nc)，則是以基因型II (52.57±0.65)顯著高於基因型I (49.02±0.63)，基因型III介於其中(50.22±0.53)。由對應分析(correspondence analysis)發現，不同基因型的外套膜基因在CGU、CGG、CCG三項密碼子的使用頻率，呈現顯著的差異。
結論
基於上述結果，可結論：台灣地區第一型登革病毒的本土流行都是因引入新的境外移入病毒株所致；截至目前為止，並沒有病毒已在台灣本土立基的證據。而建立完整的基因資料庫，配合親緣性分析，是作為追溯病毒感染來源或探討演化議題的重要工具。在本研究中利用Bayesian skyline plot建構出的第一型登革病毒族群變化情形和目前可得的流行病學資料時序吻合，顯示這項新方法的正確性與其重要性。至於不同基因型間密碼子選用度與組成的差異究竟是因為隨機漂浮(random drift)或是某種尚未明瞭的機轉選擇所致，還有待進一步探討。

關鍵字

登革病毒 ； 親緣性分析 ； 溯祖理論

並列摘要

Background: Dengue is among the most important mosquito-borne viral diseases in tropical and subtropical regions. All four serotypes of dengue viruses (DENV-1 through DENV-4) can cause severe disease and can be further divided into several genotypes. Dengue, being present in over 100 countries, is estimated to affect ~2.5 billion residing in dengue endemic areas with up to 50 million infections annually. The disease burden in Southeast Asia may be as high as 900 disability adjusted life years/million residents/year. In Taiwan, we had indigenous dengue outbreaks almost every year since 1987, with most of them occurred in southern Taiwan.

Methods: Genetic database for the 5’UTR and structural genes of 88 DENV-1 strains collected in Taiwan during 1987 to 2004 were established by RT-PCR and direct sequencing. These data were analyzed along with the E gene sequence of 40 representative global strains from the US-CDC, the sequence data of 95 DENV-1 strains available from GeneBack, and 6 E gene sequences (by J. Askove). All duplicated identical sequences were removed before analysis. Phylogenetic analyses using Maximum likelihood (ML) and Bayesian methods were constructed by PHYLIP and MrBayes 3.1, respectively. Application of Bayesian skyline plot by BEAST allows to infer the population dynamics as well as to co-estimate the substitution rates. PAML is used to test for positive selection pressure at amino acid sites and CodonW is used to exam for codon usage bias.

Results: Based on phylogenetic analysis, three major genotypes of DENV-1 are currently circulating in the world; each genotype exhibits unique geographical distributions, and linage extinction occurred through time. In Taiwan, DENV-1 isolated from different season (or cluster) of indigenous transmission fell into two genotypes (I and II); and indigenous strains from sequential years usually showed >1% p-distances, both findings were not supporting endemic transmission. The phylogenetic characteristics of an unusual indigenous dengue case in relation to the genetic database have assisted to trace the possible origin of dengue infection to laboratory-acquired in nature. The mean substitution rate estimated from 185 E gene sequences dataset was 0.743x10-3 substitution/site/year (95% highest probability densities (HPD), 0.663-0.824 x10-3). DENV-1 strains, was estimated to evolve ~96.8 years before 2004 (88.1-106.7 years). All three genotypes experienced significant population expansions in recent 40 years with the most recent event occurred around the year 2000. Selection analysis identified no positive selection on any site or specific lineage of DENV-1. The amino acid substitutions that characterize strains from different clades or subclades in genotype I and II are all conservative changes. Analysis of the codon usage on E gene indicated that genotype I had the lowest G+C3s (0.404±0.006), followed by genotype II (0.432±0.007) and then genotype III (0.448±0.009), while the Nc was significant higher in genotype II (52.57±0.65) than in genotype I (49.02±0.63); and genotype III was intermediate (50.22±0.53). Correspondence analysis showed exclusiveness in the usage of codon CGU (Arg), CGG (Arg), and CCG (Pro) is different among genotype.

Conclusions: Based on these results, we concluded that DENV-1 is reintroduced to Taiwan and resulted in subsequent indigenous transmission yearly without establishing endemicity in the most part. Establishment of comprehensive genetic database together with phylogenetic analysis is proven useful in tracking the source of transmission and in investigation of evolutionary parameters. The timeline of virus population expansion derived from Bayesian skyline plot was supported by the available epidemiological evidences, demonstrating the value in reconstructing the transition history of population growth of DENV-1. Whether the codon usage bias and composition difference observed among genotypes is a result of random drift or selection under some novel mechanism, needs to be further elucidated.

並列關鍵字

dengue virus ； phylogenetic analysis ； coalescent theory

參考文獻

Anonymous (2003). Dengue control in Singapore. Epidemiological News Bulletin 29, 29-34.

A-Nuegoonpipat, A., Berlioz-Arthaud, A., Chow, V., Endy, T., Lowry, K., Mai le, Q., Ninh, T. U., Pyke, A., Reid, M., Reynes, J. M., Su Yun, S. T., Thu, H. M., Wong, S. S., Holmes, E. C., and Aaskov, J. (2004). Sustained transmission of dengue virus type 1 in the Pacific due to repeated introductions of different Asian strains. Virology 329(2), 505-12.

Anonymous (1980). Laboratory safety for arboviruses and certain other viruses of vertebrates: the Subcommittee on Arbovirus Laboratory Safety of the American Committee on Arthropod-Borne Viruses. Am J Trop Med Hyg 291, 359-81.

Anonymous (2002). Surveillance for dengue fever/dengue haemorrhagic fever in Singapore. . Epidemiological News Bulletin 28, 25-28.

Auewarakul, P. (2005). Composition bias and genome polarity of RNA viruses. Virus Res 109, 33-37.