论文精选
The selection pressure analysis of chicken anemia virus structural
The selection pressure analysis of chicken anemia virus structuralprotein gene VP1Dong Wang Wei Fan Guan-Zhu Han Cheng-Qiang HeReceived: 10 October 2008 / Accepted: 11 December 2008 Springer Science+Business Media, LLC 2009Abstract Chicken anemia virus (CAV) is the pathogen ofchicken infectious anemia. To clarify the driving force inCAV evolution, we have detected positive selection in thestructural protein gene VP1 by using maximum-likelihoodmodels. Strong evidence was found that VP1 proteins weresubject to the high rates of positive selection, and eight siteswere identified to be under positive selection using theBayes Empirical Bayesian method. Interestingly, fourselected sites (amino acids 75, 125, 141, and 144) might beresponsible for the attenuation exhibited. One selected site(amino acid 287) was connected with the virulence of CAV.This study provided some implication for the evolution ofCAV, development of vaccines, and investigation into thestructural and functional profiles of the VP1 protein.Keywords Chicken anemia virus VP1 Positive selection VirulenceChicken anemia virus (CAV) infection is an economicallyimportant clinical and subclinical disease in broiler chickenswith a worldwide distribution [1]. CAV, which belongs tothe family Circoviridae, Gyrovirus genus, is a non-envelopedvirus with a negative sense single-stranded circularDNA genome [2, 3]. It is known as a much conserved virusof one serotype with several genetic groups. The viralgenome consists of 2.3 kb with three partially or completelyoverlapping ORFs. ORF3 (1,347 bp) encodes the majorviral structural protein VP1 (52 kDa) and partially overlapswith ORF1 (648 bp), which encodes the 24 kDa proteinVP2 [4].The importance of positive Darwinian selection as aprocess shaping the evolution of protein coding genes hasbeen suggested by numerous recent studies [5, 6], includingin Circovirdae [7, 8]. One of the most stringent methods ofdetecting adaptive evolution is to compare the rate ofnonsynonymous substitutions (dN) with the rate of synonymoussubstitution (dS) [9]. Maximum-likelihood (ML)methods are most statistically satisfactory because theyemploy an explicit model of evolution [10], taking intoaccount the effects of unequal transition and transversionrates, unequal base and codon frequencies, and variable xvalues among lineages in a phylogeny [11].Amino acid substitutions, especially within the VP1gene, may play a role in the rate of virus growth and cytopathogenicity[12]. As the major viral structural protein,VP1 carries highly conformational epitopes responsible forimmune reaction, which poses it to highest selective pressure,and thus seems to present the highest variability in thethree ORFs [13]. Renshaw et al. [14] identified an hypervariableregion in VP1 and found that certain amino acidchanges in this region could influence the rate of virusreplication and spread of CAV. Previously, we reported thathomologous recombination can occur in VP1 of CAV [15].In this study, we investigated the nature of selection pressureacting on viral structural protein VP1 gene of CAV.All currently available 48 CAV complete VP1 genesequences were extracted from the GenBank database andaligned using CLUSTAL X program [16]. Putativerecombinant/mosaic sequences (SD24 and SD22) wereDong Wang and Wei Fan contributed equally to this work.D. Wang G.-Z. Han C.-Q. He (&)College of Life Science, Shandong Normal University, Jinan,Shandong 250014, Chinae-mail: hchqiang@yahoo.com.cnW. FanHospital of Qilu, Shandong University, Jinan 250012, China123Virus GenesDOI 10.1007/s11262-008-0316-zexcluded to avoid their influence on the statistical result[15, 17]. There are overlapping reading frames betweenCAV VP1 and VP2 gene, so we trimmed off the first 180nucleotides (overlapping region of VP1 and VP2) of VP1gene lying in the very region to discard the interference theof overlapping genetic codes, which corresponded to aminoacid positions 1039–2205 (389 amino acids) of the referencesequence M81223. Ambiguous regions of thealignment were removed [18] and the alignment was furtherrefined by visual inspection using the sequence editorBioEdit version [19] to generate the VP1 gene codingsequences alignment. ML trees were constructed usingPhyml online tool [20] (http://atgc.lirmm.fr/phyml/). Thesetrees were tested using bootstrap (500 replications). Theappropriate models of molecular evolution for use in thephylogenetic analyses were identified using the Kakusan2program [21] (Fig. 1).To analyze the possibility of positive selection acting onthe VP1 protein and to infer amino acid sites under positiveselection, the ML method was applied, which implementedin the codeml program from the PAML package [22, 23].Several site-specific models (M0, M1, M2, M3, M7, andM8) that allow for various dN/dS ratios among sites wereused to detect positive selection [24]. From these models,likelihood-ratio test (LRT) was conducted to determine themodel that best fit the data [25]. The test compares twonested model (M1 vs. M2; M0 vs. M3; M7 vs. M8): one thatallows for sites under positive selection (with x[1), andanother that does not (with 0\x\1), with the x2 distributionused for significance testing [25]. Positive selectioncan be inferred from this analysis when models M2, M3, orM8 indicate a group of codons with an x ratio greater thanone, the likelihood of the positive selection model is significantlyhigher than that of the nested null hypothesismodel (at P\0.05). The Bayes empirical Bayes (BEB)calculation of posterior probabilities for site classes wasused to calculate the probabilities of sites under positiveselection [26]. Codon sites with x values[1 and whereposterior probabilities summed to be[95% were identifiedas potentially being under positive selection. In addition, weanalyzed the dataset using a conservative method fordetection of positively selected codons, which is implementedas the fixed effects likelihood (FEL) method in theHyphy package available from DataMonkey [27].Interestingly, all three models (M2, M3, and M8) thatallow for selection were significantly favored over the othermodels (P\0.001) in all cases (Table 1). Six codon sites(sites 75, 139, 144, 287, 370, 447) were inferred to be underpositive selection, with P[90% under M8 (Table 2). Twomore sites (sites 125, 141) were found under M2. The FELmethod found six codon sites (sites 75, 139, 144, 287, 370,447) under positive selection in VP1 gene (data not shown).We concluded that positive selection might act on specificamino acids in VP1 gene of CAV.Based on mutagenesis and monoclonal antibody studiesof cloned infectious viruses, the coding regions of VP1Fig. 1 Maximum-likelihood(ML) tree of the 46 CAV VP1genes nucleotide sequencesalignment. Constructed usingPhyml (TN93 model, 500bootstrap replicates). Isolatesused: name/country/accessionnumberVirus Genes123were thought to contribute to virus attenuation and reducedpathogenicity in chickens, especially the N-terminalregion. Most of the positively selected sites (75, 125, 139,141, and 144) fall within an about 70 aa segment of theN-terminal of VP1, suggesting that the positively selectedsites might lie in the same domain that interacts with host.Todd et al. [28] find that four amino acid changes at residues75I, 125L, 141L, and 144E combining with amino acidposition 89 change in the VP1 would attenuate. Interestingly,Yamaguchi et al. [29] identified one amino acid (394aa), which was a major determinant of pathogenicity, butthis site was not under positive selection in our study. Genemutation would induce the variation of the conformation ofVP1 protein, leading more efficient binding to the receptorpresent on host cells, and mutation might evolve undercontinuous positive selection. This is in accordance withtrade-off model [30], which has been used to explain awide range of phenomena related to virulence [31].Furthermore Todd et al. [32] found that an attenuatedCAV isolate was shown previously to recover pathogenicityfollowing 10 passages in young chicks. Theconsensus nucleotide sequence of the ‘revertant’ (Rev)virus changed amino acid 287 of the capsid protein from Ato D. As one of the 10 selective sites (site 287), its ability tomutate and undergo change in response to virus: hostinteraction is connecting with the virulence of CAV. Thissupported the opinion that the change of virulence of virusis under a positive selection in host.In conclusion, this study provides evidence for that thestructural protein genes VP1 has undergone positive selectionand identifies eight positively selected sites. Someselected sites may be responsible for the attenuation exhibitionand connect with the virulence. Further studies will berequired to assess the contribution of the positively selectedsites to virus virulence and pathogenesis. Understanding thefunctional importance of these positively selected aminoacid positions could help to predict possible changes in virulenceas well as be used to design live vaccines.Acknowledgement This work was supported by Innovative PostdoctoralProject Foundation of Shandong Province in China(200702013).References1. B.M. Adair, F. McNeilly, C.D. McConnell, D. Todd, R.T. Nelson,M.S. McNulty, Avian. Dis. 35, 783–792 (1991). doi:10.2307/1591611Table 1 Likelihood-ratio test (LRT) to detect adaptive evolutionModels 2Dl v2 Value df P-valueM1 versus M2 2[(-6472.421567)–(-6507.940805)] 71.38467 2 P\\0.001M0 versus M3 2[(-6438.088381)–(-6932.751737)] 989.326712 4 P\\0.001M7 versus M8 2[(-6446.667423)–(-6580.270305)] 267.205764 2 P\\0.001Note: P\\0.001(v20:001;2 13:82; v20:001;4 18:47); df Degrees of freedom between nested models; Dl Likelihood-ratio statisticTable 2 Parameter estimates under six models of variable x’s among sitesModel P Parameters l dN/dS Positively selected sitesM0: one ratio 1 x = 0.08236 -6932.751737 0.0824 Not allowedM1: neutral 1 p0 = 0.94965, x = 0.01253p1 = 0.05035, x = 1.00000-6507.940805 0.0623 Not allowedM2: selection 3 p0 = 0.94405, x = 0.00861p1 = 0.05077, x = 1.00000p2 = 0.00518, x = 10.9286-6472.421567 0.1207 75*, 125*, 139*, 141,144*, 287*, 370*, 447*M3: discrete 5 p0 = 0.92917, x = 0.00959p1 = 0.05016, x = 0.49520p2 = 0.02067, x = 3.63641-6438.088381 0.1089 75*, 139*, 144*, 287*,370*, 447*M7: beta 2 p = 0.01316, q = 0.02697 -6580.270305 0.3071 Not allowedM8: beta & x 4 p0 = 0.99479, p = 0.01526q = 0.10747, p1 = 0.00521x = 9.52289-6446.667423 0.1599 75, 139*, 144*,287*, 370*, 447*Note: Codons with a posterior probability greater than 0.95 of belonging to the positively selected class (x[1) are shown in normal typeface;* Codons with a posterior probability greater than 0.99; l Log-likelihood scoreVirus Genes1232. C.R. Pringle, Arch. Virol. 144, 2065–2070 (1999). doi:10.1007/s0070500507283. H. Gelderblom, S. Kling, R. Lurz, I. Tischer, V. von Bulow,Arch. Virol. 109, 115–120 (1989). doi:10.1007/BF013105224. B.M. Meehan, D. Todd, J.L. Creelan, J.A. Earle, E.M. Hoey, M.S.McNulty, Arch. Virol. 124, 301–319 (1992). doi:10.1007/BF013098115. S. Jagadeeshan, R.S. Singh, Mol. Biol. Evol. 24, 929–938 (2007).doi:10.1093/molbev/msm0096. C. Seoighe, F. Ketwaroo, V. Pillay, K. Scheffler, N. Wood, R.Duffet, M. Zvelebil, N. Martinson, J. McIntyre, L. Morris, W. Hide,Mol. Biol. Evol. 24, 1025–1031 (2007). doi:10.1093/molbev/msm0217. L. Heath, D.P. Martin, L. Warburton, M. Perrin, W. Horsfield, C.Kingsley, E.P. Rybicki, A.L. Williamson, J. Virol. 78, 9277–9284(2004). doi:10.1128/JVI.78.17.9277-9284.20048. A. Olvera, M. Cortey, J. Segales, Virology 357, 175–185 (2007).doi:10.1016/j.virol.2006.07.0479. M. Kimura, Sci. Am. 241, 98–100, 102, 108 passim (1979)10. Z. Yang, Mol. Biol. Evol. 15, 568–573 (1998)11. Z. Yang, R. Nielsen, J. Mol. Evol. 46, 409–418 (1998). doi:10.1007/PL0000632012. D. Todd, K.A. Mawhinney, M.S. McNulty, J. Clin. Microbiol. 30,1661–1666 (1992)13. A.N. Scott, M.S. McNulty, D. Todd, Arch. Virol. 146, 713–728(2001). doi:10.1007/s00705017014114. R.W. Renshaw, C. Soine, T. Weinkle, P.H. O’Connell, K.Ohashi, S. Watson, B. Lucio, S. Harrington, K.A. Schat, J. Virol.70, 8872–8878 (1996)15. C.Q. He, N.Z. Ding, W. Fan, Y.H. Wu, J.P. Li, Y.L. Li, Virology366, 1–7 (2007). doi:10.1016/j.virol.2007.06.00716. J.D. Thompson, D.G. Higgins, T.J. Gibson, Nucl. Acids Res. 22,4673–4680 (1994). doi:10.1093/nar/22.22.467317. M. Anisimova, R. Nielsen, Z. Yang, Genetics 164, 1229–1236(2003)18. M. Anisimova, J.P. Bielawski, Z. Yang, Mol. Biol. Evol. 19,950–958 (2002)19. H.F. Tippmann, Brief. Bioinform. 5, 82–87 (2004). doi:10.1093/bib/5.1.8220. S.Guindon,O.Gascuel, Syst.Biol. 52, 696–704 (2003). doi:10.1080/1063515039023552021. A.S. Tanabe, Mol. Ecol. Notes 7, 962–964 (2007). doi:10.1111/j.1471-8286.2007.01807.x22. Z. Yang, Mol. Biol. Evol. 24, 1586–1591 (2007). doi:10.1093/molbev/msm08823. R. Nielsen, Z. Yang, Genetics 148, 929–936 (1998)24. Z. Yang, R. Nielsen, Mol. Biol. Evol. 19, 908–917 (2002)25. M. Anisimova, J.P. Bielawski, Z. Yang, Mol. Biol. Evol. 18,1585–1592 (2001)26. Z. Yang, W.S. Wong, R. Nielsen, Mol. Biol. Evol. 22, 1107–1118(2005). doi:10.1093/molbev/msi09727. S.L.K. Pond, S.D.W. Frost, Mol. Biol. Evol. 22(5), 1208–1222(2005). doi:10.1093/molbev/msi10528. D. Todd, A.N. Scott, N.W. Ball, B.J. Borghmans, B.M. Adair, J.Virol. 76, 8472–8474 (2002). doi:10.1128/JVI.76.16.8472-8474.200229. S. Yamaguchi, T. Imada, N. Kaji, M. Mase, K. Tsukamoto, N.Tanimura, N. Yuasa, J. Gen. Virol. 82, 1233–1238 (2001)30. B. Soubeyrand, S.A. Plotkin, Nature 417, 609–610 (2002). doi:10.1038/417609b31. S. Gandon, M.J. Mackinnon, S. Nee, A.F. Read, Nature 414,751–756 (2001). doi:10.1038/414751a32. D. Todd, J.L. Creelan, T.J. Connor, N.W. Ball, A.N. Scott, B.M.Meehan, G.F. McKenna, M.S. McNulty, Avian. Pathol. 32,375–382 (2003). doi:10.1080/037945031000121121Virus Genes123