- Email: [email protected]
North American clades
Vaccine 21 (2003) 1072–1081
The out of Africa model of varicella-zoster virus evolution: single nucleotide polymorphisms and private alleles distinguish Asian clades from European/North American clades Timothy R. Wagenaar a , Vincent T.K. Chow b , Chantanee Buranathai c , Pranee Thawatsupha c , Charles Grose a,∗ a
Departments of Microbiology & Pediatrics, University of Iowa College of Medicine, University of Iowa Hospital/2501 JCP, 200 Hawkins Drive, Iowa City, IA 52242, USA b Department of Microbiology, Faculty of Medicine, National University of Singapore, Singapore, Singapore c National Institute of Health, Department of Medical Sciences, Nonthaburi, Bangkok, Thailand Received 20 June 2002; received in revised form 4 October 2002; accepted 9 October 2002
Abstract Until 1998, varicella-zoster virus (VZV) was generally considered sufficiently stable to allow the use of a single sequenced virus (VZV-Dumas) as a consensual representation of the world VZV genotype. But recent investigations have uncovered a gE mutant virus called VZV-MSP with a second genotype and a distinguishable accelerated cell spread phenotype. A subsequent study suggested that single nucleotide polymorphisms (SNPs) could be applied toward the genetic analysis of the VZV genome. To further assess the scope of genetic variation in the VZV genome on a worldwide basis, we carried out an extensive SNP analysis of structural glycoprotein genes gB, gE, gH, gI, gL, as well as the IE62 regulatory gene in viruses collected from Western Europe, North America and Asia, including the VZV vaccine strain. The SNP data showed segregation of viral isolates of Asian origin from those of Western ancestry into distinct phylogenetic clades. Unexpectedly, however, VZV from Thailand segregated with VZV from Iceland and the United States, i.e. it was more Western than Asian in nature. Further, SNP analysis disclosed strikingly unusual genotypes, e.g. gH genes with up to five missense mutations and gL genes with insertions of an in-frame methionine codon. In summary, these VZV genomic analyses have shown that individual VZV strains, like closely related human beings, have distinctive SNP profiles containing private alleles within just five VZV genes (gB, gH, gE, gL and IE62) that provide a fingerprint to localize ancestry of the viral strain. © 2002 Elsevier Science Ltd. All rights reserved. Keywords: Varicella-zoster virus; Chickenpox; Bioinformatics
1. Introduction Based on data provided by the sequencing of several diverse herpesvirus genomes, varicella-zoster virus (VZV) is considered to have evolved from its primordial ancestor about 60–70 mya and has continued to cospeciate in close association with its primate hosts [1,2]. Based on this geologic time frame, VZV most likely arose in a primordial primate in Africa [3]. Until recently the genotype/phenotype of VZV was considered to be conserved throughout the world because of the presumed low VZV mutability. The initially sequenced VZV laboratory strain, called VZV-Dumas, represented the prototype [4]. In 1998, we reported the discovery of a mutant VZV that had lost a major B cell epitope ∗
Corresponding author. Tel.: +1-319-356-2270; fax: +1-319-356-4855. E-mail address: [email protected] (C. Grose).
in the gE ectodomain [5]. The mutant virus has been extensively characterized. Because the mutation occurred in gE, a glycoprotein largely associated with the property of cell-to-cell spread in all alphaherpesviruses, this property was assessed in VZV-MSP by a newly designed imaging technique based on confocal microscopy [6,7]. VZV-MSP showed an accelerated cell to cell spread, possibly representing a gain-of-function VZV phenotype. Following the discovery of VZV-MSP, further investigations were conducted among North American VZV strains to classify VZV by an approach similar to the Human Genome Project, namely, using single nucleotide polymorphisms (SNPs) found in five VZV glycoproteins genes gB, gE, gH, gI, gL and the major early transactivation gene IE62 [8]. Preliminary results showed that VZV strains collected mainly from North America could be distinguished on the basis of SNP analysis. SNPs are single base pair variations
0264-410X/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 2 6 4 - 4 1 0 X ( 0 2 ) 0 0 5 5 9 - 5
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
in genomic DNA which differ from the prototype but occur in at least 1% of the population [9,10]. In our VZV analyses, we required that a polymorphism occur in at least two strains to be called a SNP, in other words, even if 98 as yet unsequenced VZV strains lacked the polymorphism, it would still be a valid SNP [8]. Based on the out-of-Africa model of human evolution [11], we put forward the hypothesis that VZV in Asia would possess different SNP profiles that VZV from Europe and North America colonized mainly by Europeans over the past 250 years. To test this hypothesis, VZV strains were obtained from two Asian countries: Singapore and Thailand. Both countries had experienced varicella outbreaks in the late 1980s–early 1990s and therefore virologists had collected samples to look for VZV restriction fragment length polymorphisms. The subsequent genomic and bioinformatic analyses described herein not only show that SNP analysis is an excellent means by which to study evolution of VZV within different human populations, they also document unexpected evidence of selective pressure on certain VZV genes, e.g. the gH/gL genes in one Asian subpopulation, as well as hybrid strains containing both Asian and Western polymorphisms. 2. Materials and methods 2.1. Viral strains The viral isolates VZV-376, VZV-529, VZV-681, VZV-1140, VZV-1242, VZV-1419 were collected from children and adults in Singapore in the early 1990s and stored in low temperature freezers [12]. VZV-V8 and VZV-Z24 were collected in the late 1980s from children in a municipal hospital in Bangkok, Thailand, and stored in low temperature freezers [13]. No children of European parentage were included in the chickenpox studies carried out in Bangkok. Vaccine Oka virus was kindly provided by M. Takahashi in 1979 [14]. The VZV strains from North America and Europe were previously described by Faga et al. [8]. VZV-32 was isolated in Texas in the 1970s. VZV-80-2 was collected from Pennsylvania in 1982. VZV-Ellen was collected from Georgia in 1960s. VZV-LAX1 and VZV-LAX2 originated from Los Angeles in the 1990s. VZV-MSP was isolated in Minnesota in 1995. VZV-VIA was obtained in Iowa in 1990s. VZV-VSD originated from South Dakota in the 1980s. VZV-Iceland was a wild-type virus collected in the country of Iceland in the 1990s. VZV-Dumas was isolated in the Netherlands in 1970s [4]. GenBank Accession numbers: VZV-Dumas: NC 001348, VZV-MSP: AY005330, VZV-Oka: AY016462.
1073
plete medium supplemented with 10% fetal calf serum (FCS GibcoBRL). Monolayers were infected with the appropriate VZV strain and incubated at 32 ◦ C until 80–90% cytopathic effect (CPE) was visible. Thereafter, the monolayer was harvested and DNA was extracted using Qiagen DNeasy Tissue Kit according to manufacturer specifications. 2.3. Sequencing and genetic analysis The VZV genes gB (ORF 31), gE (ORF 68), gH (ORF 37), gI (ORF 67), gL (ORF 60), and IE62 (ORF 62) were amplified by polymerase chain reaction (PCR) with a Boehringer Mannheim Expand High Fidelity PCR system using corresponding oligonucleotides. PCR reaction products were analyzed for appropriate molecular weight product by 1% agarose gel electrophoresis followed by ethidium bromide staining. The amplified products were purified with a Microcon YM-50 filter according to modified manufacturer specifications. Fluorescence-tagged sequencing was preformed on an Applied Biosystems Model 373A stretch fluorescent automated sequencer (Perkin-Elmer) at the University of Iowa DNA facility. The sequences of each open reading frame from the corresponding viral isolates were analyzed by MacDNASIS version 3.7 (Hitachi) for maximum homology to the genomic sequence VZV-Dumas. Any mutations were designated numerically from the adenosine of the start codon with the Dumas base pair listed first, followed by the base pair number and the resulting base substitution. Viral gene sequences were aligned using Bioedit version 5.0.9 [15] and phylogenetic trees were constructed using MEGA version 2.1 [16]. 3. Results 3.1. The VZV gB gene (ORF 31) The gB gene is one of the most highly conserved herpesviral genes [2]. Polymorphisms were designated after sequencing of the individual viral genes from Asian strains. The SNP profiles for the gB gene showed three distinctive Asian subpopulations called Singapore subpopulation 1 (S1), and Singapore subpopulation 2 (S2), and Bangkok, Thailand (Fig. 1). The presence of one coding SNP T73P and two noncoding SNPs a294c and a390t characterized the S2 and Oka strains and separated them from the subpopulations called S1 and Bangkok. The nucleotide substitution c948t within the S1 strains distinguished them from the S2, Oka and Bangkok strains. The Bangkok population was unique among Asian strains by lacking any nucleotide substitution within gB; in this regard, gB Bangkok was identical to the prototype VZV-Dumas gB gene [4].
2.2. DNA preparation
3.2. The VZV gE gene (ORF 68)
The human melanoma cell line designated as MeWo was grown at 37◦ in minimal essential medium with Eagle com-
The gE gene is present in the US segment of the VZV genome and is one of the genes which is considered to be
1074
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
Fig. 1. Alleles of VZV gB (ORF 31). In this figure and all subsequent figures synonymous SNPs are indicated by white bars and nonsynonymous SNPs are indicated by black bars. Amino acids are indicated by capital letters and nucleotides are indicated by lower case letters. Singapore subpopulation 2 resembled Oka. Bangkok VZV resembled the prototype VZV-Dumas.
a relatively recent evolutionary addition to the herpes viral genome, not being present in beta or gamma herpesviruses [1,17]. The individual Asian gE subpopulations also contained SNPs (Fig. 2). Moreover, the gE SNPs continued to segregate into three subpopulations. All gE genes within the S1, S2, Oka and Bangkok subpopulations contained the missense mutation T40I, which was absent from VZV-Dumas. The occurrence of a single silent mutation t1062g in the S1 strains distinguished them from the other strains. As with gB, the gE SNP profiles of the S2 and Oka subpopulations were identical. The Bangkok subpopulation showed a second coding SNP L536I which provided a unique gE SNP
profile. The VZV-MSP gE mutation D150N was not seen in any Asian strain. 3.3. The VZV gH gene (ORF37) The gH gene is another highly conserved glycoprotein found in alpha, beta and gamma herpesvirus; it houses important neutralization epitopes [2,18]. The SNP profiles of the Asian gH gene provided unexpected findings (Fig. 3). All viral subpopulations displayed in common two coding SNPs P269L and R700K. The S1 subpopulation had a noncoding SNP c2298t. In addition the S2 and Oka subpopulations
Fig. 2. Alleles of VZV gE gene (ORF 68). All Asian strains contained the coding SNP T40I, which was not present in VZV-Dumas. Singapore VZV subpopulation 2 resembled VZV Oka.
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
1075
Fig. 3. Alleles of VZV gH (ORF 37). Singapore VZV subpopulation 2 resembled VZV Oka. VZV Bangkok contained more polymorphisms than found in gH from any other geographic location. The R72K and S319A SNPs were not present in the three other Asian subpopulations.
contained a silent mutation g573t. In contrast, the Bangkok subpopulation was easily distinguished from S1, S2 and Oka subpopulations by finding one new noncoding SNP t39g and two more coding SNPs R72K and S319A. In short, both Bangkok strains possessed four coding SNPs. In addition, one of the two Bangkok strains had a fifth missense mutation T639A which was not included in Fig. 3. The large number of missense mutations in Bangkok gH was striking because they had not been seen previously in any other subpopulation [8]. 3.4. The VZV gI gene (ORF 67) The gI protein provides a chaperone function for the gE glycoprotein [17]. As seen in earlier studies, gI exhibited
few SNPs (Fig. 4). One noncoding SNP was observed in the S2 and Oka subpopulations and this SNP facilitated the separation of S2 and Oka subpopulations from the S1 and Bangkok subpopulations. But unlike other previously sequenced genes, S1 and Bangkok strains were not distinguishable due to no mutational differences in the gI sequence. 3.5. The VZV gL gene (ORF 60) The gL protein provides a chaperone function for the gH protein [19]. In addition to SNPs, some gL genes contained the in-frame insertion of an ATG methionine codon that may potentially serve as an alternative start site (Fig. 5). The S1 subpopulation contained a single noncoding polymorphism. All three other Asian subpopulations contained the in-frame
Fig. 4. Alleles of VZV gI gene (ORF 67). Singapore subpopulation 2 resembled VZV Oka. Both Singapore subpopulation 1 and Bangkok resembled VZV-Dumas.
1076
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
Fig. 5. Alleles of VZV gL (ORF 60). An additional in-frame ATG methionine codon (27atg28) was found in Singapore subpopulation 2, Bangkok and VZV Oka, but not in Singapore subpopulation 1 or VZV-Dumas. Singapore subpopulation 2 resembled VZV Oka.
insertion of the ATG methionine codon. The S2 and Oka subpopulations displayed a missense mutation A107T. The Bangkok strains were distinguishable from other strains by a noncoding SNP in addition to the insertion of the ATG methionine codon. The occurrence of the ATG methionine codon in gL was the only mutation seen in all sequenced genes that involved insertion of additional base pairs. The ATG insertion was never previously seen in Western VZV strains [4,8].
3.6. The VZV IE62 gene (ORF 62) The VZV IE62 gene is the homolog of the HSV ICP4 gene and is the predominant VZV regulatory protein [20]. The IE62 SNP patterns characteristic of the Singapore subpopulations contained more polymorphisms than those in the Bangkok strains (Fig. 6). All sequenced Asian isolates possessed in common one coding polymorphism g1969a > A657T and three noncoding polymorphisms g90c, c183t
Fig. 6. Alleles of VZV IE62 (ORF 62). Singapore subpopulation 2 resembled VZV Oka. VZV Thailand resembled VZV-Dumas except for one missense mutation. The same missense mutation was present in all Asian subpopulations.
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
1077
Fig. 7. Unrooted phylogenetic trees of VZV evolution constructed from the aligned genomic sequences of VZV genes gB, gE, gH, gI, gL and IE62. Viral isolates from North American/Europe and Singapore/Japan segregated into four distinct clades (top tree). Branches A and D represented North American/Europe strains, while branches B and C represented the Singapore strains as well as the Japanese Oka strain. The Bangkok isolates (V8 and Z24) segregated with the European/North American clade D (bottom tree).
1078
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
and t387c (latter three not shown). The S1, S2, and Oka subpopulations contained the SNP a1827g. The presence of SNP g2565a in the S1 strains separated them from the other three sub-populations. The S2 and Oka subpopulations possessed the a1419g, g1527t, and t3822c polymorphisms that facilitated differentiation from the S1 and Bangkok subpopulations. 3.7. The promoter region of IE62 The promoter region of the IE62 gene was sequenced in eight viruses starting approximately 600 bp prior to the IE62 start site. Four viruses had identical nucleotide regions when compared with the published Dumas strain, while four shared a common insertion of a guanine into a stretch of eight guanines between bp −113 and −120. There were no other consistent polymorphisms. Therefore, these data were not included in further analyses of polymorphisms. 3.8. Clades and phylogenetic trees After assessment of the SNP profiles shown in Figs. 1–6, an unrooted phylogenetic tree was constructed from nucleotide sequences of the five glycoproteins genes gB, gE, gH, gI, gL and the major early transactivator IE62. Analysis of a phylogenetic tree showed division into four distinct viral clades, arbitrarily assigned letters A–D (Fig. 7 top tree). Clades B and C contained viral isolates of Singapore/Japan origin. In contrast, viruses of Western European and North American origin segregated within clades A and D. Clade D revealed an anomaly however. When the two Thailand strains were inserted into the same phylogenetic analyses (bottom tree), they sorted with the European/North American strains in Clade D, and not with the other Asian strains in clades B and C. Close inspection of Figs. 1–6 showed that the Thailand strains contained some VZV genes which were more European in character and other genes which were more Asian in character. Furthermore, the validity of the segregation of VZV isolates into clade D was strongly supported by a high bootstrap value of 95. 3.9. Bioinformatic analysis of the VZV genome Computer technology has facilitated more detailed comparative genomics among the VZV strains [21]. A private allele is a marker within the human genome; the name was originally applied to alleles which segregated small groups of humans with rare blood types [22,23]. When a similar strategy was applied toward the VZV genomic analysis, we documented that private alleles were present in clades B,C and D (Fig. 7). Viral clade B was characterized by the private allele g390t in gB, g573t in gH, g546a in gI, and A107T in gL. Clade C included private alleles c948t in gB, t1062g in gE, g186t in gL and g2565a in IE62. Clade D contained the private allele L536I in gE, R72K and t39g in gH. Note again that the private alleles in clade D were
found in 2 Asian strains and 3 European strains. Clade A possessed no private alleles because Clade A contained the prototype Dumas strain.
4. Discussion The specificity of VZV for its human host during an entire lifetime suggests that co-evolution of VZV and its human host occurs [1,2]. VZV presumably evolved from its progenitor 60–70 mya. It is reasonable to assume therefore that VZV has evolved solely in primates, which also originated 60–70 mya (reviewed in [3]). Based on the same hypothesis, VZV coevolved with Australopithecus, Homo erectus and finally Homo sapiens in Africa. The current VZV genomic studies, when combined with prior studies, demonstrated that VZV can be analyzed in the same manner as for the Human Genome Project [8,24,25]. In other words, SNPs can be located in individual VZV genes and these SNPs can serve as a markers to discriminate VZV subpopulations. The ability to use SNP analysis depended on the apparent contradictory properties of the VZV genome: (1) the VZV genome was much more stable than most viral genomes (2) but the VZV genome, like the human genome, underwent both synonymous and nonsynonymous mutations at consistent locations within the genome which could be determined by SNP analysis [4,8]. As with the human genome, some VZV genes were more highly conserved than others. Through the process of investigation and elimination, we have discovered that SNP analyses of just 5 of the approximately 70 VZV genes were sufficient to categorize VZV genomes into phylogenetic trees and determine shared ancestry. 4.1. VZV gH mutations The SNP analyses of the gH gene were the most surprising. We have previously used both BLAST peptide and MaxHom searches to examine the VZV gH mutation at codon 269. We discovered that the substitution of leucine at codon 269 leads to greater identity with gH found in pseudorabies virus [8]. Again, this comparison is of interest because of the relatively close evolutionary distance between VZV and pseudorabies virus [2]. Of further interest, bioinformatics analyses have suggested that the herpesviral gH gene is a candidate for positive selection [26]. A high ratio of missense to silent mutations within a gene can be an indication of positive natural selection upon a gene [24,27]. Within the human genome, it is exceedingly unusual to find exons with SNPs containing potentially deleterious alterations in proteins. Similarly, the observance of multiple missense mutations within the coding region of VZV gene gH from Thailand was a very surprising finding. This high number of missense mutations was not seen in gH from any other clade. The gH gene is known to code for one of the major complement independent neutralization epitopes of the virus [18]. Although the exact
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
locations of the epitopes are not known, they lie within the ectodomain. Thus, the observance of an increased number of nonsynonymous mutations within the Thailand gH population may suggest an immunological selection of these gH mutant viruses. A similar situation has been suggested during outbreaks of pseudorabies virus [28]. 4.2. VZV gL and new ATG site The gH glycoprotein forms a complex with the gL protein, which serves to chaperone gH through its biosynthesis [19]. With the exception of S1 subpopulation, the Asian subpopulations studied herein can be defined by the presence of the 27atg28 insertion within gL. The second atg site may serve as the primary or a secondary initiation of translation. In addition, the observation of the atg insertion may correlate with the rapidly evolving gH of Bangkok, as these two proteins form a complex [19]. The gH/gL complex may be subject to evolutionary pressure within the heterogeneous Bangkok population. Furthermore, the 27atg28 insertion is not observed within European and American strains and additional characterization of African strains would provide an answer as to when in the evolution of VZV this mutation occurred. 4.3. VZV IE62 and attenuation The IE62 gene is the major VZV regulatory protein. Therefore, polymorphisms within IE62 have been consid-
1079
ered to be a mechanism of attenuation [29]. Thus, they were considered the end result of passage in cell culture. However, the VZV-Dumas strain, which was a laboratory strain with an unclear passage history, has an IE62 amino acid sequence identical to low passage VZV isolates, such as the Bangkok strain. The VZV-MSP strain also has a very similar sequence. For purposes of comparison, the IE62 sequences of our VZV strains were compared with other published sequences (Table 1) [8,30,31]. Based on studies in the SCID-hu model of VZV infection, VZV Ellen appears to be equally attenuated to the VZV Oka vaccine strain. If the polymorphisms shared by Ellen and Oka are compared, and those found in other stains are excluded, coding SNP R958G may indicate attenuation. Another possibility is that attenuation may require two linked SNPs in the IE62 protein. This comparative genomic analysis is another indicator for the value of SNP profiles. 4.4. Thailand VZV and recombination Analyses of the two Thailand strains were very informative. SNPs within the Thailand isolates revealed that the VZV genes gB and gI were identical to the prototypic Western isolates. However, the 27atg28 insertion in the gL gene was found only in Asian strains. Additionally, the Thai isolates contained SNPs private to the Western clade. A reasonable explanation for the Thailand SNPs is recombination. Evolution of DNA viruses usually occurs via 2
Table 1 VZV IE62 polymorphisms Polymorphism
VZV strains Thailand
t124g > S42A t296c > M99T a583g > N195D t1337c > L446P t1535c > V512A c1805 > A602V a1882g > S628G g1969a > A657T t2108c > V703A a2872g > R958G a3170g > Q1057R a3215g > Q1072R t3590c > V1197A a3602g > H1201R t3622g > S1208A a3644g > Q1215R t3683c > L1228P a3721g > S1241G a3728g > E1243G t3763g > S1255A a3778g > I1260V t3824c > L1275S a
V-Okaa
P-Okab
V-Okab
V-Okac
×
×
×
×
× ×
× ×
×
×
×
×
×
×
×
×
×
×
×
×
× ×
Iceland
VSD
Via
32
× ×
× × × ×
80-2
×
× × × ×
Ellen
×
×
Faga et al. [8]. Gomi et. al. [30]. c Argaw et. al. [31]. b
Singapore
× ×
× × × ×
×
× × × ×
× ×
×
×
MSP
1080
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
mechanisms: mutation and recombination. Recombination events have been thoroughly evaluated in another DNA virus–adenovirus. Experimental assays with adenoviruses revealed that coinfection of cells with serologically distinct types 2 and 5 adenoviruses generated up to 20% hybrid virus [32]. Furthermore, recombination was able to found in almost every region of the viral genome [32]. Supplementary evidence from isolated field strains yielded further evidence of interviral recombination between viral serotypes [32]. Similarly, recombination events have been cataloged in several herpesviruses. Most commonly, recombination has been considered a likely explanation for the emergence of the gamma herpesviral EBV strains 1 and 2; in turn, recombination events may also have occurred subsequently between EBV-1 and EBV-2 themselves [33,34]. Recombination occurs in cell cultures infected with different strains of pseudorabies virus, an alpha herpes virus closely related to VZV in the herpesvirus phylogenetic tree [35]. Recombination has also been documented in cultures infected with VZV [36]. Another recent study has documented different VZV variants based largely on a heteroduplex mobility assay [37]. These investigators discovered at least three major VZV genetic clades in samples collected from humans residing on four continents. Most geographic regions contained strains representative of only one of the three clades. However, in the country of Brazil they discovered a probable recombinant VZV virus. There is yet one more intriguing connection between variant strains of DNA viruses and Brazil. In a large worldwide survey of the evolution of human papillomavirus type 16, the most unusual so called bipartite variants were discovered in Brazil [38]. The viruses had genetic attributes of both European and African HPV-16 strains. The authors speculated that the bipartite HPV-16 strains were the result of commingling of European and African viruses during the colonization period beginning in the early 1500s. Our data also provide a timeline for VZV recombination. Based on the Thailand findings, we postulate that recombination has occurred within the last 500 years. The Europeans first arrived in Thailand in 1511 and quickly established a permanent colony in the capital city. Some of these European immigrants presumably developed herpes zoster and thereby introduced European VZV strains into the native Thai population. The recent Brazil VZV findings of Muir et al. [37] support a similar timeline of 500 years. 4.5. Coevolution and private alleles In summary, the VZV genomics data support the concept that VZV coevolves with the human genome. Under this hypothesis, Homo erectus dispersed out of Africa and established populations in Asia and Europe. At the same time, VZV traveled within the human host and diversified in Asian and European populations. As with human populations, the VZV genome is continuing to evolve by mutation, e.g. a second VZV gE mutant virus extremely similar to
the VZV-MSP strain has been isolated from an elderly man with shingles in Vancouver, BC, Canada [39]. Finally, our data demonstrate the likely presence of private alleles specific to VZV populations within humans residing in small geographic areas. The concept of private alleles further extends the genetic similarities between the human and VZV genomes [23,40].
Acknowledgements We thank M.B. Coulthart and G.A. Tipples, National Microbiology Laboratories, Health Canada, Winnipeg, Canada, for their comments. T.R. Wagenaar was supported by a fellowship from the Howard Hughes program at the University of Iowa. Research by C. Grose was supported by NIH grants AI22795 and AI36884. References [1] McGeoch DJ, Cook S. Molecular phylogeny of the alphaherpesvirinae subfamily and a proposed evolutionary timescale. J Mol Biol 1994;238:9–22. [2] McGeoch JJ, Davison AJ. The molecular evolutionary history of herpesviruses. In: Domingo E, Webster R, Holland J, editors. Origin and evolution of viruses. San Diego: Academic Press; 1999. p. 441–465. [3] Grose C. Varicella-zoster virus: less immutable than once thought. Pediatrics 1999;103:1027–8. [4] Davison AJ, Scott JE. The complete DNA sequence of varicellazoster virus. J Gen Virol 1986;67:1759–816. [5] Santos RA, Padilla JA, Hatfield C, Grose C. Antigenic variation of varicella-zoster virus Fc receptor gE: loss of a major B cell epitope in the ectodomain. Virology 1998;249:21–31. [6] Santos RA, Hatfield CC, Cole NL, et al. Varicella-zoster virus gE escape mutant VZV-MSP exhibits an accelerated cell-to-cell spread phenotype in both infected cell cultures and SCID-hu mice. Virology 2000;275:306–17. [7] Dingwell KS, Johnson DC. The herpes simplex virus gE–gI complex facilitates cell-to-cell spread and binds to components of cell junctions. J Virol 1998;72:8933–42. [8] Faga B, Maury W, Bruckner DA, Grose C. Identification and mapping of single nucleotide polymorphisms in the varicella-zoster virus genome. Virology 2001;280:1–6. [9] Wang DG, Fan JB, Siao CJ, et al. Large-scale identification, mapping, and genotyping of single-nucleotide polymorphisms in the human genome. Science 1998;280:1077–82. [10] Brookes AJ. The essence of SNPs. Gene 1999;234:177–86. [11] Templeton AR. Out of Africa? What do genes tell us? Curr Opin Genet Dev 1997;7:841–7. [12] Chow VT, Lim KP. Amplification and sequencing of varicella-zoster virus (VZV) gene 4: point mutation in a VZV strain causing chickenpox during pregnancy. Acta Virol 1997;41:277–83. [13] Thawaranantha D, Balachandra K, Jongtrakulsiri S, Yamkunthong W, Chimabutra K, Bhumiswasdi J. Genome differences among varicella-zoster viruses isolated in Thailand. Southeast Asian J Trop Med Public Health 1995;26:677–83. [14] Takahashi M, Okuno Y, Otsuka T, Osame J, Takamizawa A. Development of a live attenuated varicella vaccine. Biken J 1975;18:25–33. [15] Hall TA. BioEdit: A user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl Acids Symp Ser 1999;41:95–8.
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081 [16] Kumar S, Tamura K, Jakobsen IB, Nei M. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 2001;17:1244–5. [17] Grose C. The predominant varicella-zoster virus gE and gI glycoprotein complex. In: Bogner E, Holzenburg A, editors. Structure function relationships of human pathogenic viruses. London: Kluwer Academic Press; 2002. [18] Grose C. Glycoproteins encoded by varicella-zoster virus: biosynthesis, phosphorylation, and intracellular trafficking. Annu Rev Microbiol 1990;44:59–80. [19] Duus KM, Grose C. Multiple regulatory effects of varicella-zoster virus (VZV) gL on trafficking patterns and fusogenic properties of VZV gH. J Virol 1996;70:8961–71. [20] Kinchington PR, Hougland J, Arvin A, Ruyechan W, Hay J. The varicella-zoster virus immediate-early protein IE62 is a major component of virus particles. J. Virol. 1992;66. [21] Roos DS. Computational biology. Bioinformatics—trying to swim in a sea of data. Science 2001;291:1260–1. [22] Slatkin M. Rare alleles as indicators of gene flow. Evolution 1985;39:53–65. [23] Neel JV. “Private” genetic variants and the frequency of mutation among South American Indians. Proc Natl Acad Sci USA 1973;70:3311–5. [24] Lander ES, Linton LM, Birren B, et al. Initial sequencing and analysis of the human genome. Nature 2001;409:860–921. [25] Venter JC, Adams MD, Myers EW, et al. The sequence of the human genome. Science 2001;291:1304–51. [26] Endo T, Ikeo K, Gojobori T. Large-scale search for genes on which positive selection may operate. Mol Biol Evol 1996;13:685–90. [27] Messier W, Stewart CB. Episodic adaptive evolution of primate lysozymes. Nature 1997;385:151–4. [28] Goldberg TL, Weigel RM, Hahn EC, Scherba G. Comparative utility of restriction fragment length polymorphism analysis and gene sequencing to the molecular epidemiological investigation of a viral outbreak. Epidemiol Infect 2001;126:415–24. [29] Loparev VN, Argaw T, Krause PR, Takayama M, Schmid DS. Improved identification and differentiation of varicella-zoster virus
[30]
[31]
[32] [33]
[34] [35]
[36]
[37]
[38]
[39]
[40]
1081
(VZV) wild-type strains and an attenuated varicella vaccine strain using a VZV open reading frame 62-based PCR. J Clin Microbiol 2000;38:3156–60. Gomi Y, Imagawa T, Takahashi M, Yamanishi K. Oka varicella vaccine is distinguishable from its parental virus in DNA sequence of open reading frame 62 and its transactivation activity. J Med Virol 2000;61:497–503. Argaw T, Cohen JI, Klutch M, Lekstrom K, Yoshikawa T, Asano Y, et al. Nucleotide sequences that distinguish Oka vaccine from parental Oka and other varicella-zoster virus isolates. J Infect Dis 2000;181:1153–7. Sambrook J, Sleigh M, Engler JA, Broker TR. The evolution of the adenoviral genome. Ann N Y Acad Sci 1980;354:426–52. Midgley RS, Blake NW, Yao QY, et al. Novel intertypic recombinants of Epstein-Barr virus in the Chinese population. J Virol 2000;74:1544–8. McGeoch DJ. Molecular evolution of the gamma-herpesvirinae. Philos Trans R Soc Lond B: Biol Sci 2001;356:421–35. Dangler CA, Deaver RE, Kolodziej CM. Genetic recombination between two strains of Aujeszky’s disease virus at reduced multiplicity of infection. J Gen Virol 1994;75:295–9. Dohner DE, Adams SG, Gelb LD. Recombination in tissue culture between varicella-zoster virus strains. J Med Virol 1988;24:329– 41. Muir WB, Nichols R, Breuer J. Phylogenetic analysis of varicellazoster virus: evidence of intercontinental spread of genotypes and recombination. J Virol 2002;76:1971–9. Chan SY, Ho L, Ong CK, et al. Molecular variants of human papillomavirus type 16 from four continents suggest ancient pandemic spread of the virus and its coevolution with humankind. J Virol 1992;66:2057–66. Tipples GA, Stephens GM, Sherlock C, Bowler M, Cook D, Grose C. A new variant of varicella-zoster virus. Emerg Infect Dis 2002;8:1424–7. Stoneking M. Single nucleotide polymorphisms. From the evolutionary past. Nature 2001;409:821–2.
The out of Africa model of varicella-zoster virus evolution: single nucleotide polymorphisms and private alleles distinguish Asian clades from European/North American clades Timothy R. Wagenaar a , Vincent T.K. Chow b , Chantanee Buranathai c , Pranee Thawatsupha c , Charles Grose a,∗ a
Departments of Microbiology & Pediatrics, University of Iowa College of Medicine, University of Iowa Hospital/2501 JCP, 200 Hawkins Drive, Iowa City, IA 52242, USA b Department of Microbiology, Faculty of Medicine, National University of Singapore, Singapore, Singapore c National Institute of Health, Department of Medical Sciences, Nonthaburi, Bangkok, Thailand Received 20 June 2002; received in revised form 4 October 2002; accepted 9 October 2002
Abstract Until 1998, varicella-zoster virus (VZV) was generally considered sufficiently stable to allow the use of a single sequenced virus (VZV-Dumas) as a consensual representation of the world VZV genotype. But recent investigations have uncovered a gE mutant virus called VZV-MSP with a second genotype and a distinguishable accelerated cell spread phenotype. A subsequent study suggested that single nucleotide polymorphisms (SNPs) could be applied toward the genetic analysis of the VZV genome. To further assess the scope of genetic variation in the VZV genome on a worldwide basis, we carried out an extensive SNP analysis of structural glycoprotein genes gB, gE, gH, gI, gL, as well as the IE62 regulatory gene in viruses collected from Western Europe, North America and Asia, including the VZV vaccine strain. The SNP data showed segregation of viral isolates of Asian origin from those of Western ancestry into distinct phylogenetic clades. Unexpectedly, however, VZV from Thailand segregated with VZV from Iceland and the United States, i.e. it was more Western than Asian in nature. Further, SNP analysis disclosed strikingly unusual genotypes, e.g. gH genes with up to five missense mutations and gL genes with insertions of an in-frame methionine codon. In summary, these VZV genomic analyses have shown that individual VZV strains, like closely related human beings, have distinctive SNP profiles containing private alleles within just five VZV genes (gB, gH, gE, gL and IE62) that provide a fingerprint to localize ancestry of the viral strain. © 2002 Elsevier Science Ltd. All rights reserved. Keywords: Varicella-zoster virus; Chickenpox; Bioinformatics
1. Introduction Based on data provided by the sequencing of several diverse herpesvirus genomes, varicella-zoster virus (VZV) is considered to have evolved from its primordial ancestor about 60–70 mya and has continued to cospeciate in close association with its primate hosts [1,2]. Based on this geologic time frame, VZV most likely arose in a primordial primate in Africa [3]. Until recently the genotype/phenotype of VZV was considered to be conserved throughout the world because of the presumed low VZV mutability. The initially sequenced VZV laboratory strain, called VZV-Dumas, represented the prototype [4]. In 1998, we reported the discovery of a mutant VZV that had lost a major B cell epitope ∗
Corresponding author. Tel.: +1-319-356-2270; fax: +1-319-356-4855. E-mail address: [email protected] (C. Grose).
in the gE ectodomain [5]. The mutant virus has been extensively characterized. Because the mutation occurred in gE, a glycoprotein largely associated with the property of cell-to-cell spread in all alphaherpesviruses, this property was assessed in VZV-MSP by a newly designed imaging technique based on confocal microscopy [6,7]. VZV-MSP showed an accelerated cell to cell spread, possibly representing a gain-of-function VZV phenotype. Following the discovery of VZV-MSP, further investigations were conducted among North American VZV strains to classify VZV by an approach similar to the Human Genome Project, namely, using single nucleotide polymorphisms (SNPs) found in five VZV glycoproteins genes gB, gE, gH, gI, gL and the major early transactivation gene IE62 [8]. Preliminary results showed that VZV strains collected mainly from North America could be distinguished on the basis of SNP analysis. SNPs are single base pair variations
0264-410X/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 2 6 4 - 4 1 0 X ( 0 2 ) 0 0 5 5 9 - 5
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
in genomic DNA which differ from the prototype but occur in at least 1% of the population [9,10]. In our VZV analyses, we required that a polymorphism occur in at least two strains to be called a SNP, in other words, even if 98 as yet unsequenced VZV strains lacked the polymorphism, it would still be a valid SNP [8]. Based on the out-of-Africa model of human evolution [11], we put forward the hypothesis that VZV in Asia would possess different SNP profiles that VZV from Europe and North America colonized mainly by Europeans over the past 250 years. To test this hypothesis, VZV strains were obtained from two Asian countries: Singapore and Thailand. Both countries had experienced varicella outbreaks in the late 1980s–early 1990s and therefore virologists had collected samples to look for VZV restriction fragment length polymorphisms. The subsequent genomic and bioinformatic analyses described herein not only show that SNP analysis is an excellent means by which to study evolution of VZV within different human populations, they also document unexpected evidence of selective pressure on certain VZV genes, e.g. the gH/gL genes in one Asian subpopulation, as well as hybrid strains containing both Asian and Western polymorphisms. 2. Materials and methods 2.1. Viral strains The viral isolates VZV-376, VZV-529, VZV-681, VZV-1140, VZV-1242, VZV-1419 were collected from children and adults in Singapore in the early 1990s and stored in low temperature freezers [12]. VZV-V8 and VZV-Z24 were collected in the late 1980s from children in a municipal hospital in Bangkok, Thailand, and stored in low temperature freezers [13]. No children of European parentage were included in the chickenpox studies carried out in Bangkok. Vaccine Oka virus was kindly provided by M. Takahashi in 1979 [14]. The VZV strains from North America and Europe were previously described by Faga et al. [8]. VZV-32 was isolated in Texas in the 1970s. VZV-80-2 was collected from Pennsylvania in 1982. VZV-Ellen was collected from Georgia in 1960s. VZV-LAX1 and VZV-LAX2 originated from Los Angeles in the 1990s. VZV-MSP was isolated in Minnesota in 1995. VZV-VIA was obtained in Iowa in 1990s. VZV-VSD originated from South Dakota in the 1980s. VZV-Iceland was a wild-type virus collected in the country of Iceland in the 1990s. VZV-Dumas was isolated in the Netherlands in 1970s [4]. GenBank Accession numbers: VZV-Dumas: NC 001348, VZV-MSP: AY005330, VZV-Oka: AY016462.
1073
plete medium supplemented with 10% fetal calf serum (FCS GibcoBRL). Monolayers were infected with the appropriate VZV strain and incubated at 32 ◦ C until 80–90% cytopathic effect (CPE) was visible. Thereafter, the monolayer was harvested and DNA was extracted using Qiagen DNeasy Tissue Kit according to manufacturer specifications. 2.3. Sequencing and genetic analysis The VZV genes gB (ORF 31), gE (ORF 68), gH (ORF 37), gI (ORF 67), gL (ORF 60), and IE62 (ORF 62) were amplified by polymerase chain reaction (PCR) with a Boehringer Mannheim Expand High Fidelity PCR system using corresponding oligonucleotides. PCR reaction products were analyzed for appropriate molecular weight product by 1% agarose gel electrophoresis followed by ethidium bromide staining. The amplified products were purified with a Microcon YM-50 filter according to modified manufacturer specifications. Fluorescence-tagged sequencing was preformed on an Applied Biosystems Model 373A stretch fluorescent automated sequencer (Perkin-Elmer) at the University of Iowa DNA facility. The sequences of each open reading frame from the corresponding viral isolates were analyzed by MacDNASIS version 3.7 (Hitachi) for maximum homology to the genomic sequence VZV-Dumas. Any mutations were designated numerically from the adenosine of the start codon with the Dumas base pair listed first, followed by the base pair number and the resulting base substitution. Viral gene sequences were aligned using Bioedit version 5.0.9 [15] and phylogenetic trees were constructed using MEGA version 2.1 [16]. 3. Results 3.1. The VZV gB gene (ORF 31) The gB gene is one of the most highly conserved herpesviral genes [2]. Polymorphisms were designated after sequencing of the individual viral genes from Asian strains. The SNP profiles for the gB gene showed three distinctive Asian subpopulations called Singapore subpopulation 1 (S1), and Singapore subpopulation 2 (S2), and Bangkok, Thailand (Fig. 1). The presence of one coding SNP T73P and two noncoding SNPs a294c and a390t characterized the S2 and Oka strains and separated them from the subpopulations called S1 and Bangkok. The nucleotide substitution c948t within the S1 strains distinguished them from the S2, Oka and Bangkok strains. The Bangkok population was unique among Asian strains by lacking any nucleotide substitution within gB; in this regard, gB Bangkok was identical to the prototype VZV-Dumas gB gene [4].
2.2. DNA preparation
3.2. The VZV gE gene (ORF 68)
The human melanoma cell line designated as MeWo was grown at 37◦ in minimal essential medium with Eagle com-
The gE gene is present in the US segment of the VZV genome and is one of the genes which is considered to be
1074
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
Fig. 1. Alleles of VZV gB (ORF 31). In this figure and all subsequent figures synonymous SNPs are indicated by white bars and nonsynonymous SNPs are indicated by black bars. Amino acids are indicated by capital letters and nucleotides are indicated by lower case letters. Singapore subpopulation 2 resembled Oka. Bangkok VZV resembled the prototype VZV-Dumas.
a relatively recent evolutionary addition to the herpes viral genome, not being present in beta or gamma herpesviruses [1,17]. The individual Asian gE subpopulations also contained SNPs (Fig. 2). Moreover, the gE SNPs continued to segregate into three subpopulations. All gE genes within the S1, S2, Oka and Bangkok subpopulations contained the missense mutation T40I, which was absent from VZV-Dumas. The occurrence of a single silent mutation t1062g in the S1 strains distinguished them from the other strains. As with gB, the gE SNP profiles of the S2 and Oka subpopulations were identical. The Bangkok subpopulation showed a second coding SNP L536I which provided a unique gE SNP
profile. The VZV-MSP gE mutation D150N was not seen in any Asian strain. 3.3. The VZV gH gene (ORF37) The gH gene is another highly conserved glycoprotein found in alpha, beta and gamma herpesvirus; it houses important neutralization epitopes [2,18]. The SNP profiles of the Asian gH gene provided unexpected findings (Fig. 3). All viral subpopulations displayed in common two coding SNPs P269L and R700K. The S1 subpopulation had a noncoding SNP c2298t. In addition the S2 and Oka subpopulations
Fig. 2. Alleles of VZV gE gene (ORF 68). All Asian strains contained the coding SNP T40I, which was not present in VZV-Dumas. Singapore VZV subpopulation 2 resembled VZV Oka.
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
1075
Fig. 3. Alleles of VZV gH (ORF 37). Singapore VZV subpopulation 2 resembled VZV Oka. VZV Bangkok contained more polymorphisms than found in gH from any other geographic location. The R72K and S319A SNPs were not present in the three other Asian subpopulations.
contained a silent mutation g573t. In contrast, the Bangkok subpopulation was easily distinguished from S1, S2 and Oka subpopulations by finding one new noncoding SNP t39g and two more coding SNPs R72K and S319A. In short, both Bangkok strains possessed four coding SNPs. In addition, one of the two Bangkok strains had a fifth missense mutation T639A which was not included in Fig. 3. The large number of missense mutations in Bangkok gH was striking because they had not been seen previously in any other subpopulation [8]. 3.4. The VZV gI gene (ORF 67) The gI protein provides a chaperone function for the gE glycoprotein [17]. As seen in earlier studies, gI exhibited
few SNPs (Fig. 4). One noncoding SNP was observed in the S2 and Oka subpopulations and this SNP facilitated the separation of S2 and Oka subpopulations from the S1 and Bangkok subpopulations. But unlike other previously sequenced genes, S1 and Bangkok strains were not distinguishable due to no mutational differences in the gI sequence. 3.5. The VZV gL gene (ORF 60) The gL protein provides a chaperone function for the gH protein [19]. In addition to SNPs, some gL genes contained the in-frame insertion of an ATG methionine codon that may potentially serve as an alternative start site (Fig. 5). The S1 subpopulation contained a single noncoding polymorphism. All three other Asian subpopulations contained the in-frame
Fig. 4. Alleles of VZV gI gene (ORF 67). Singapore subpopulation 2 resembled VZV Oka. Both Singapore subpopulation 1 and Bangkok resembled VZV-Dumas.
1076
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
Fig. 5. Alleles of VZV gL (ORF 60). An additional in-frame ATG methionine codon (27atg28) was found in Singapore subpopulation 2, Bangkok and VZV Oka, but not in Singapore subpopulation 1 or VZV-Dumas. Singapore subpopulation 2 resembled VZV Oka.
insertion of the ATG methionine codon. The S2 and Oka subpopulations displayed a missense mutation A107T. The Bangkok strains were distinguishable from other strains by a noncoding SNP in addition to the insertion of the ATG methionine codon. The occurrence of the ATG methionine codon in gL was the only mutation seen in all sequenced genes that involved insertion of additional base pairs. The ATG insertion was never previously seen in Western VZV strains [4,8].
3.6. The VZV IE62 gene (ORF 62) The VZV IE62 gene is the homolog of the HSV ICP4 gene and is the predominant VZV regulatory protein [20]. The IE62 SNP patterns characteristic of the Singapore subpopulations contained more polymorphisms than those in the Bangkok strains (Fig. 6). All sequenced Asian isolates possessed in common one coding polymorphism g1969a > A657T and three noncoding polymorphisms g90c, c183t
Fig. 6. Alleles of VZV IE62 (ORF 62). Singapore subpopulation 2 resembled VZV Oka. VZV Thailand resembled VZV-Dumas except for one missense mutation. The same missense mutation was present in all Asian subpopulations.
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
1077
Fig. 7. Unrooted phylogenetic trees of VZV evolution constructed from the aligned genomic sequences of VZV genes gB, gE, gH, gI, gL and IE62. Viral isolates from North American/Europe and Singapore/Japan segregated into four distinct clades (top tree). Branches A and D represented North American/Europe strains, while branches B and C represented the Singapore strains as well as the Japanese Oka strain. The Bangkok isolates (V8 and Z24) segregated with the European/North American clade D (bottom tree).
1078
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
and t387c (latter three not shown). The S1, S2, and Oka subpopulations contained the SNP a1827g. The presence of SNP g2565a in the S1 strains separated them from the other three sub-populations. The S2 and Oka subpopulations possessed the a1419g, g1527t, and t3822c polymorphisms that facilitated differentiation from the S1 and Bangkok subpopulations. 3.7. The promoter region of IE62 The promoter region of the IE62 gene was sequenced in eight viruses starting approximately 600 bp prior to the IE62 start site. Four viruses had identical nucleotide regions when compared with the published Dumas strain, while four shared a common insertion of a guanine into a stretch of eight guanines between bp −113 and −120. There were no other consistent polymorphisms. Therefore, these data were not included in further analyses of polymorphisms. 3.8. Clades and phylogenetic trees After assessment of the SNP profiles shown in Figs. 1–6, an unrooted phylogenetic tree was constructed from nucleotide sequences of the five glycoproteins genes gB, gE, gH, gI, gL and the major early transactivator IE62. Analysis of a phylogenetic tree showed division into four distinct viral clades, arbitrarily assigned letters A–D (Fig. 7 top tree). Clades B and C contained viral isolates of Singapore/Japan origin. In contrast, viruses of Western European and North American origin segregated within clades A and D. Clade D revealed an anomaly however. When the two Thailand strains were inserted into the same phylogenetic analyses (bottom tree), they sorted with the European/North American strains in Clade D, and not with the other Asian strains in clades B and C. Close inspection of Figs. 1–6 showed that the Thailand strains contained some VZV genes which were more European in character and other genes which were more Asian in character. Furthermore, the validity of the segregation of VZV isolates into clade D was strongly supported by a high bootstrap value of 95. 3.9. Bioinformatic analysis of the VZV genome Computer technology has facilitated more detailed comparative genomics among the VZV strains [21]. A private allele is a marker within the human genome; the name was originally applied to alleles which segregated small groups of humans with rare blood types [22,23]. When a similar strategy was applied toward the VZV genomic analysis, we documented that private alleles were present in clades B,C and D (Fig. 7). Viral clade B was characterized by the private allele g390t in gB, g573t in gH, g546a in gI, and A107T in gL. Clade C included private alleles c948t in gB, t1062g in gE, g186t in gL and g2565a in IE62. Clade D contained the private allele L536I in gE, R72K and t39g in gH. Note again that the private alleles in clade D were
found in 2 Asian strains and 3 European strains. Clade A possessed no private alleles because Clade A contained the prototype Dumas strain.
4. Discussion The specificity of VZV for its human host during an entire lifetime suggests that co-evolution of VZV and its human host occurs [1,2]. VZV presumably evolved from its progenitor 60–70 mya. It is reasonable to assume therefore that VZV has evolved solely in primates, which also originated 60–70 mya (reviewed in [3]). Based on the same hypothesis, VZV coevolved with Australopithecus, Homo erectus and finally Homo sapiens in Africa. The current VZV genomic studies, when combined with prior studies, demonstrated that VZV can be analyzed in the same manner as for the Human Genome Project [8,24,25]. In other words, SNPs can be located in individual VZV genes and these SNPs can serve as a markers to discriminate VZV subpopulations. The ability to use SNP analysis depended on the apparent contradictory properties of the VZV genome: (1) the VZV genome was much more stable than most viral genomes (2) but the VZV genome, like the human genome, underwent both synonymous and nonsynonymous mutations at consistent locations within the genome which could be determined by SNP analysis [4,8]. As with the human genome, some VZV genes were more highly conserved than others. Through the process of investigation and elimination, we have discovered that SNP analyses of just 5 of the approximately 70 VZV genes were sufficient to categorize VZV genomes into phylogenetic trees and determine shared ancestry. 4.1. VZV gH mutations The SNP analyses of the gH gene were the most surprising. We have previously used both BLAST peptide and MaxHom searches to examine the VZV gH mutation at codon 269. We discovered that the substitution of leucine at codon 269 leads to greater identity with gH found in pseudorabies virus [8]. Again, this comparison is of interest because of the relatively close evolutionary distance between VZV and pseudorabies virus [2]. Of further interest, bioinformatics analyses have suggested that the herpesviral gH gene is a candidate for positive selection [26]. A high ratio of missense to silent mutations within a gene can be an indication of positive natural selection upon a gene [24,27]. Within the human genome, it is exceedingly unusual to find exons with SNPs containing potentially deleterious alterations in proteins. Similarly, the observance of multiple missense mutations within the coding region of VZV gene gH from Thailand was a very surprising finding. This high number of missense mutations was not seen in gH from any other clade. The gH gene is known to code for one of the major complement independent neutralization epitopes of the virus [18]. Although the exact
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
locations of the epitopes are not known, they lie within the ectodomain. Thus, the observance of an increased number of nonsynonymous mutations within the Thailand gH population may suggest an immunological selection of these gH mutant viruses. A similar situation has been suggested during outbreaks of pseudorabies virus [28]. 4.2. VZV gL and new ATG site The gH glycoprotein forms a complex with the gL protein, which serves to chaperone gH through its biosynthesis [19]. With the exception of S1 subpopulation, the Asian subpopulations studied herein can be defined by the presence of the 27atg28 insertion within gL. The second atg site may serve as the primary or a secondary initiation of translation. In addition, the observation of the atg insertion may correlate with the rapidly evolving gH of Bangkok, as these two proteins form a complex [19]. The gH/gL complex may be subject to evolutionary pressure within the heterogeneous Bangkok population. Furthermore, the 27atg28 insertion is not observed within European and American strains and additional characterization of African strains would provide an answer as to when in the evolution of VZV this mutation occurred. 4.3. VZV IE62 and attenuation The IE62 gene is the major VZV regulatory protein. Therefore, polymorphisms within IE62 have been consid-
1079
ered to be a mechanism of attenuation [29]. Thus, they were considered the end result of passage in cell culture. However, the VZV-Dumas strain, which was a laboratory strain with an unclear passage history, has an IE62 amino acid sequence identical to low passage VZV isolates, such as the Bangkok strain. The VZV-MSP strain also has a very similar sequence. For purposes of comparison, the IE62 sequences of our VZV strains were compared with other published sequences (Table 1) [8,30,31]. Based on studies in the SCID-hu model of VZV infection, VZV Ellen appears to be equally attenuated to the VZV Oka vaccine strain. If the polymorphisms shared by Ellen and Oka are compared, and those found in other stains are excluded, coding SNP R958G may indicate attenuation. Another possibility is that attenuation may require two linked SNPs in the IE62 protein. This comparative genomic analysis is another indicator for the value of SNP profiles. 4.4. Thailand VZV and recombination Analyses of the two Thailand strains were very informative. SNPs within the Thailand isolates revealed that the VZV genes gB and gI were identical to the prototypic Western isolates. However, the 27atg28 insertion in the gL gene was found only in Asian strains. Additionally, the Thai isolates contained SNPs private to the Western clade. A reasonable explanation for the Thailand SNPs is recombination. Evolution of DNA viruses usually occurs via 2
Table 1 VZV IE62 polymorphisms Polymorphism
VZV strains Thailand
t124g > S42A t296c > M99T a583g > N195D t1337c > L446P t1535c > V512A c1805 > A602V a1882g > S628G g1969a > A657T t2108c > V703A a2872g > R958G a3170g > Q1057R a3215g > Q1072R t3590c > V1197A a3602g > H1201R t3622g > S1208A a3644g > Q1215R t3683c > L1228P a3721g > S1241G a3728g > E1243G t3763g > S1255A a3778g > I1260V t3824c > L1275S a
V-Okaa
P-Okab
V-Okab
V-Okac
×
×
×
×
× ×
× ×
×
×
×
×
×
×
×
×
×
×
×
×
× ×
Iceland
VSD
Via
32
× ×
× × × ×
80-2
×
× × × ×
Ellen
×
×
Faga et al. [8]. Gomi et. al. [30]. c Argaw et. al. [31]. b
Singapore
× ×
× × × ×
×
× × × ×
× ×
×
×
MSP
1080
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081
mechanisms: mutation and recombination. Recombination events have been thoroughly evaluated in another DNA virus–adenovirus. Experimental assays with adenoviruses revealed that coinfection of cells with serologically distinct types 2 and 5 adenoviruses generated up to 20% hybrid virus [32]. Furthermore, recombination was able to found in almost every region of the viral genome [32]. Supplementary evidence from isolated field strains yielded further evidence of interviral recombination between viral serotypes [32]. Similarly, recombination events have been cataloged in several herpesviruses. Most commonly, recombination has been considered a likely explanation for the emergence of the gamma herpesviral EBV strains 1 and 2; in turn, recombination events may also have occurred subsequently between EBV-1 and EBV-2 themselves [33,34]. Recombination occurs in cell cultures infected with different strains of pseudorabies virus, an alpha herpes virus closely related to VZV in the herpesvirus phylogenetic tree [35]. Recombination has also been documented in cultures infected with VZV [36]. Another recent study has documented different VZV variants based largely on a heteroduplex mobility assay [37]. These investigators discovered at least three major VZV genetic clades in samples collected from humans residing on four continents. Most geographic regions contained strains representative of only one of the three clades. However, in the country of Brazil they discovered a probable recombinant VZV virus. There is yet one more intriguing connection between variant strains of DNA viruses and Brazil. In a large worldwide survey of the evolution of human papillomavirus type 16, the most unusual so called bipartite variants were discovered in Brazil [38]. The viruses had genetic attributes of both European and African HPV-16 strains. The authors speculated that the bipartite HPV-16 strains were the result of commingling of European and African viruses during the colonization period beginning in the early 1500s. Our data also provide a timeline for VZV recombination. Based on the Thailand findings, we postulate that recombination has occurred within the last 500 years. The Europeans first arrived in Thailand in 1511 and quickly established a permanent colony in the capital city. Some of these European immigrants presumably developed herpes zoster and thereby introduced European VZV strains into the native Thai population. The recent Brazil VZV findings of Muir et al. [37] support a similar timeline of 500 years. 4.5. Coevolution and private alleles In summary, the VZV genomics data support the concept that VZV coevolves with the human genome. Under this hypothesis, Homo erectus dispersed out of Africa and established populations in Asia and Europe. At the same time, VZV traveled within the human host and diversified in Asian and European populations. As with human populations, the VZV genome is continuing to evolve by mutation, e.g. a second VZV gE mutant virus extremely similar to
the VZV-MSP strain has been isolated from an elderly man with shingles in Vancouver, BC, Canada [39]. Finally, our data demonstrate the likely presence of private alleles specific to VZV populations within humans residing in small geographic areas. The concept of private alleles further extends the genetic similarities between the human and VZV genomes [23,40].
Acknowledgements We thank M.B. Coulthart and G.A. Tipples, National Microbiology Laboratories, Health Canada, Winnipeg, Canada, for their comments. T.R. Wagenaar was supported by a fellowship from the Howard Hughes program at the University of Iowa. Research by C. Grose was supported by NIH grants AI22795 and AI36884. References [1] McGeoch DJ, Cook S. Molecular phylogeny of the alphaherpesvirinae subfamily and a proposed evolutionary timescale. J Mol Biol 1994;238:9–22. [2] McGeoch JJ, Davison AJ. The molecular evolutionary history of herpesviruses. In: Domingo E, Webster R, Holland J, editors. Origin and evolution of viruses. San Diego: Academic Press; 1999. p. 441–465. [3] Grose C. Varicella-zoster virus: less immutable than once thought. Pediatrics 1999;103:1027–8. [4] Davison AJ, Scott JE. The complete DNA sequence of varicellazoster virus. J Gen Virol 1986;67:1759–816. [5] Santos RA, Padilla JA, Hatfield C, Grose C. Antigenic variation of varicella-zoster virus Fc receptor gE: loss of a major B cell epitope in the ectodomain. Virology 1998;249:21–31. [6] Santos RA, Hatfield CC, Cole NL, et al. Varicella-zoster virus gE escape mutant VZV-MSP exhibits an accelerated cell-to-cell spread phenotype in both infected cell cultures and SCID-hu mice. Virology 2000;275:306–17. [7] Dingwell KS, Johnson DC. The herpes simplex virus gE–gI complex facilitates cell-to-cell spread and binds to components of cell junctions. J Virol 1998;72:8933–42. [8] Faga B, Maury W, Bruckner DA, Grose C. Identification and mapping of single nucleotide polymorphisms in the varicella-zoster virus genome. Virology 2001;280:1–6. [9] Wang DG, Fan JB, Siao CJ, et al. Large-scale identification, mapping, and genotyping of single-nucleotide polymorphisms in the human genome. Science 1998;280:1077–82. [10] Brookes AJ. The essence of SNPs. Gene 1999;234:177–86. [11] Templeton AR. Out of Africa? What do genes tell us? Curr Opin Genet Dev 1997;7:841–7. [12] Chow VT, Lim KP. Amplification and sequencing of varicella-zoster virus (VZV) gene 4: point mutation in a VZV strain causing chickenpox during pregnancy. Acta Virol 1997;41:277–83. [13] Thawaranantha D, Balachandra K, Jongtrakulsiri S, Yamkunthong W, Chimabutra K, Bhumiswasdi J. Genome differences among varicella-zoster viruses isolated in Thailand. Southeast Asian J Trop Med Public Health 1995;26:677–83. [14] Takahashi M, Okuno Y, Otsuka T, Osame J, Takamizawa A. Development of a live attenuated varicella vaccine. Biken J 1975;18:25–33. [15] Hall TA. BioEdit: A user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl Acids Symp Ser 1999;41:95–8.
T.R. Wagenaar et al. / Vaccine 21 (2003) 1072–1081 [16] Kumar S, Tamura K, Jakobsen IB, Nei M. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 2001;17:1244–5. [17] Grose C. The predominant varicella-zoster virus gE and gI glycoprotein complex. In: Bogner E, Holzenburg A, editors. Structure function relationships of human pathogenic viruses. London: Kluwer Academic Press; 2002. [18] Grose C. Glycoproteins encoded by varicella-zoster virus: biosynthesis, phosphorylation, and intracellular trafficking. Annu Rev Microbiol 1990;44:59–80. [19] Duus KM, Grose C. Multiple regulatory effects of varicella-zoster virus (VZV) gL on trafficking patterns and fusogenic properties of VZV gH. J Virol 1996;70:8961–71. [20] Kinchington PR, Hougland J, Arvin A, Ruyechan W, Hay J. The varicella-zoster virus immediate-early protein IE62 is a major component of virus particles. J. Virol. 1992;66. [21] Roos DS. Computational biology. Bioinformatics—trying to swim in a sea of data. Science 2001;291:1260–1. [22] Slatkin M. Rare alleles as indicators of gene flow. Evolution 1985;39:53–65. [23] Neel JV. “Private” genetic variants and the frequency of mutation among South American Indians. Proc Natl Acad Sci USA 1973;70:3311–5. [24] Lander ES, Linton LM, Birren B, et al. Initial sequencing and analysis of the human genome. Nature 2001;409:860–921. [25] Venter JC, Adams MD, Myers EW, et al. The sequence of the human genome. Science 2001;291:1304–51. [26] Endo T, Ikeo K, Gojobori T. Large-scale search for genes on which positive selection may operate. Mol Biol Evol 1996;13:685–90. [27] Messier W, Stewart CB. Episodic adaptive evolution of primate lysozymes. Nature 1997;385:151–4. [28] Goldberg TL, Weigel RM, Hahn EC, Scherba G. Comparative utility of restriction fragment length polymorphism analysis and gene sequencing to the molecular epidemiological investigation of a viral outbreak. Epidemiol Infect 2001;126:415–24. [29] Loparev VN, Argaw T, Krause PR, Takayama M, Schmid DS. Improved identification and differentiation of varicella-zoster virus
[30]
[31]
[32] [33]
[34] [35]
[36]
[37]
[38]
[39]
[40]
1081
(VZV) wild-type strains and an attenuated varicella vaccine strain using a VZV open reading frame 62-based PCR. J Clin Microbiol 2000;38:3156–60. Gomi Y, Imagawa T, Takahashi M, Yamanishi K. Oka varicella vaccine is distinguishable from its parental virus in DNA sequence of open reading frame 62 and its transactivation activity. J Med Virol 2000;61:497–503. Argaw T, Cohen JI, Klutch M, Lekstrom K, Yoshikawa T, Asano Y, et al. Nucleotide sequences that distinguish Oka vaccine from parental Oka and other varicella-zoster virus isolates. J Infect Dis 2000;181:1153–7. Sambrook J, Sleigh M, Engler JA, Broker TR. The evolution of the adenoviral genome. Ann N Y Acad Sci 1980;354:426–52. Midgley RS, Blake NW, Yao QY, et al. Novel intertypic recombinants of Epstein-Barr virus in the Chinese population. J Virol 2000;74:1544–8. McGeoch DJ. Molecular evolution of the gamma-herpesvirinae. Philos Trans R Soc Lond B: Biol Sci 2001;356:421–35. Dangler CA, Deaver RE, Kolodziej CM. Genetic recombination between two strains of Aujeszky’s disease virus at reduced multiplicity of infection. J Gen Virol 1994;75:295–9. Dohner DE, Adams SG, Gelb LD. Recombination in tissue culture between varicella-zoster virus strains. J Med Virol 1988;24:329– 41. Muir WB, Nichols R, Breuer J. Phylogenetic analysis of varicellazoster virus: evidence of intercontinental spread of genotypes and recombination. J Virol 2002;76:1971–9. Chan SY, Ho L, Ong CK, et al. Molecular variants of human papillomavirus type 16 from four continents suggest ancient pandemic spread of the virus and its coevolution with humankind. J Virol 1992;66:2057–66. Tipples GA, Stephens GM, Sherlock C, Bowler M, Cook D, Grose C. A new variant of varicella-zoster virus. Emerg Infect Dis 2002;8:1424–7. Stoneking M. Single nucleotide polymorphisms. From the evolutionary past. Nature 2001;409:821–2.