Complete DNA Sequences of Two Oka Strain Varicella-Zoster Virus Genomes

Nucleic acid extraction.Total DNA was extracted from a single vial of recent production lots of Varilrix (lot VAV10118, produced in April 2002; GSK Biologicals, Rixensart, Belgium) and Varivax (lot 0895 M, purchased in 2003; Merck & Co., Inc., Whitehouse Station, NJ) vaccines using a High Pure viral nucleic acid kit from Roche (Basel, Switzerland). In brief, 100 μl of sample was lysed in a lysing-binding buffer in the presence of proteinase K. The lysis mixture was then applied to a glass fiber filter, which binds the nucleic acids in the presence of the lysis and binding buffer containing chaotropic salts. Bound nucleic acids were eluted in 50 μl of nuclease-free water by centrifugation and stored at −70°C.

PCR.Around 540 primers were designed using Primer D software (GSK in-house software) and the nucleotide sequence of the Dumas strain (GenBank accession no. X04370) (12). Overlapping primers were designed approximately 500 bases apart to cover the entire genomic sequence of VZV. Sequences of primers used for amplification and sequencing are available upon request. The reaction mixtures for PCR contained 15 μl of HotStarTaq Plus Master Mix solution (Qiagen, Valencia, CA), 0.3 μM of each primer, and 5 ng of template DNA. A Tetrad thermal cycler (MJ Research, Waltham, MA) was used for all amplifications. An initial hot-start PCR step of 96°C for 15 min was followed by 35 cycles of amplification (95°C for 20 s, 55°C for 30 s, and 72°C for 45 s) and a final elongation step at 72°C for 3 min. All amplified products were then purified using QIAquick PCR purification kit (Qiagen). Direct sequencing of both DNA strands was performed on the generated amplicons.

Sequencing.Direct sequencing of purified PCR products and plasmid DNA was performed with BigDye Terminator cycle sequencing kit and a 3730XL genetic analyzer (both from Applied Biosystems, Foster City, CA). The viral sequences were compiled and analyzed with Sequencher software (Gene Codes Corp, Ann Arbor, MI). The following GenBank sequences were used for comparison: for the European (The Netherlands) reference strain (Dumas), X04370 (12); for Oka-P, AB097933 (22); and for Oka-V, AB097932 (22), AF206304 (3), AY016450 (15), and the sequencing information provided by Schmidt et al. (60). Unless otherwise stated, all described nucleotide sequence positions in this paper correspond to the genome of Dumas strain, X04370 (12).

Cloning of PCR products.When direct sequencing did not generate information of sufficient quality or when particular single nucleotide polymorphisms (SNPs) could not be reliably confirmed, additional subcloning was performed, followed by sequencing of numerous generated clones to confirm the consensus sequence of the region. Direct sequencing of the PCR products derived from regions with highly complex secondary structure (flanking regions between internal repeat long and internal repeat short regions, and the R3 repeat region) was complemented by subcloning of amplicons and sequencing of plasmid clones. PCR products containing these regions were individually inserted into a pCR2.1 vector (Invitrogen, Carlsbad, CA) and then transformed into competent Escherichia coli by the TOPO TA cloning method (Invitrogen). The plasmid DNAs were purified from cultured bacteria with a QIAprep spin kit (Qiagen). DNA sequences of the cloned inserts were determined using vector-specific sequencing primers.

Sequencing of ends of the viral genomes.The direct sequencing data for viral genome ends were complemented by sequencing of overlapping amplicons generated after circularization using T4 DNA ligase (Roche). The PCR mixtures contained 500 μM of each deoxynucleoside triphosphate, 10 pmol of each primer, and 2.5 U high-fidelity Platinum Taq polymerase (Invitrogen). PCR products were inserted into a pCR4-TOPO vector and transformed into competent E. coli TOP10 bacteria by the TOPO TA cloning method (Invitrogen). The plasmid DNAs were purified with a QIAprep spin kit (Qiagen). The consensus sequence of the cloned amplicons was confirmed by sequencing and alignment of multiple E. coli plasmid clones.

Nucleotide sequence accession numbers.The complete nucleotide consensus sequences of the Oka-V_GSK (Varilrix) and Oka-V_Merck (Varivax) strains are available in GenBank under the accession numbers DQ008354 and DQ008355, respectively.

Oka-V_GSK and Oka-V_Merck genome organization.The full-length consensus sequence of Oka-V_GSK and Oka-V_Merck vaccine strains was essentially determined by bidirectional sequencing of overlapping PCR-amplified fragments. Occasionally, when the amplified region contained SNPs that could not be conclusively resolved, the amplified fragments were subcloned and a consensus sequence was derived from multiple plasmid clones. The obtained sequences were assembled and the complete genomes of the vaccines were annotated using the VZV sequence of the Dumas strain published by Davison and Scott as a template (12). The full annotations for Oka-V_GSK and Oka-V_Merck are presented in Tables 1 and 2, respectively.

The complete genomes of Oka-V_GSK and Oka-V_Merck strains are comprised of 124,821 and 124,815 bp, respectively. Like the wild-type Dumas strain and the parental Japanese Oka-V strain, the Oka-V_GSK and Oka-V_Merck genomes consist of a unique long region flanked by terminal repeat long and internal repeat long inverted repeat regions, as well as a unique short region flanked by internal repeat short (IRS) and terminal repeat short (TRS) inverted repeat regions. An origin of replication was found in both the IRS and TRS regions. Four unique reiteration regions (R1 to R4) were found along the genome, with R4 duplicated in the IRS and TRS regions.

All the open reading frames (ORFs) described for the Dumas VZV strain (12) and the Oka vaccine parental strain (22) were found in the two Oka-derived vaccine strains (Tables 1 and 2). The 72 ORFs predicted to encode proteins were evenly distributed on both DNA strands. Three genes were located within the repeat sequences and were therefore duplicated within the VZV genome, so that ORFs 69 to 71 in the IRS region correspond to ORFs 62 to 64 in the TRS region.

Comparison of Oka strain genomes to the Dumas strain genome.The obtained sequences of Oka-V_GSK and Oka-V_Merck were aligned with the full-length VZV genomes of Oka-P, Oka-V, and Dumas strains. All sequence differences between the four Oka strains and the Dumas strain are given in Table 3. A total of 326 nucleotide positions displaying differences relative to the genome of Dumas strain (X04370 [12]) were identified. Among these, 228 were common to the four Oka strains, and the remaining 98 were specific to one, two, or three of the Oka strains. Several deletions or insertions were found, but most mutations were substitutions of one nucleotide, i.e., SNPs. Frequently, the original nucleotide was nonetheless preserved, resulting in a mixture of two nucleotides present at the same position (Table 3). Because, to our knowledge, the vaccine strains were never cloned, this is consistent with the existence of multiple viral species that evolved during the attenuation process. Multiple SNPs were found to still contain the original Oka-P-specific nucleotide. This supports the cooperative effect of the overall pattern of nucleotide substitutions in the expression of the attenuation phenotype and, to a lesser extent, the contribution of individual SNPs.

The 98 differences between the Oka-V_GSK (124,821 bp), Oka-V_Merck (124,815 bp), Oka-P (125,125 bp), and Oka-V (125,078 bp) genomes were found in 25 ORFs (ORFs 1, 2, 6, 9A, 10, 11, 14, 18, 21, 22, 31, 35, 39, 45, 47, 48, 50, 51, 52, 54, 55, 62, 64, and 71), the R1 and R3 repeat regions (in ORFs 11 and 22, respectively), and one origin of replication (Table 3).

The total number of differences between the four Oka strains was determined (Table 4). Of the 98 differences identified, 69 were found between Oka-P and Oka-V, 51 between Oka-V and Oka-V_GSK, and 68 between Oka-V and Oka-V_Merck. Consequently, Oka-V_Merck contains 17 more differences that discriminate it from Oka-V compared with Oka-V_GSK.

Although the highest convergence was found for Oka-V_GSK and Oka-V_Merck, they still had 36 nucleotide differences (Table 4). For 12 of these positions, Oka-V_Merck had nucleotides matching the Oka-P strain, whereas the Oka-V_GSK strain had only a single position (119683) where the sequence was Oka-P-like. Overall, for the positions in which Oka-V_GSK differed from Oka-V_Merck, the Oka-V_GSK sequence was closer to Oka-V, whereas the Oka-V_Merck sequence was closer to Oka-P.

Sixty-nine nucleotide changes between the Oka-V and the Oka-P strains were identified (Table 5). Among these 69 differences, 56 positions in Oka-P were identical to the reference Dumas strain, whereas only 11 positions in Oka-V were identical to the Dumas strain. Identical nucleotides for many of these positions were also present in Oka-V_GSK and Oka-V_Merck.

To better characterize the observed differences, the substitution spectra were analyzed (Fig. 2). The large majority of mutations were SNPs and only partial, with two different nucleotides at the same position. Compared to Oka-P, transitions (i.e., mutations resulting in substitution of a purine for a purine [A↔ G] or a pyrimidine for a pyrimidine [C↔ T]) were more frequently (64% to 69%) observed for the Oka-V_GSK and Oka-V_Merck strains than transversions (i.e., mutations resulting in substitution of a purine for a pyrimidine and vice versa; 13% to 17%). Transversions were more common than insertions or deletions (≤10%). The majority of the identified mutations were silent mutations, either because they were located in intergenic regions or because of the degenerated genetic code. A significant proportion of mutations in intragenic regions (∼45%) caused single amino acid substitutions in both the Oka-V_GSK and Oka-V_Merck strains (Fig. 2). No stop or frameshift mutations were identified. All deletions and insertions either were located in intergenic regions or, when located within coding regions, were multiples of three bases.

• Open in new tab
• Download powerpoint

FIG. 2.

Type (A) and function (B) of the mutations between Oka-P and the Oka-V_GSK and Oka-V_Merck vaccine strains of VZV. The numbers indicate the number of events identified for each category of mutations. aa, amino acid.

Comparison of Oka-V_GSK and Oka-V_Merck genomes.Sequence differences observed between the Oka-V_GSK and Oka-V_Merck strains are described in Table 6. Only 36 differences were found throughout the complete genomes (i.e., ∼125 kb), three of which were repeated in ORF 62 and its duplicate, ORF 71. These 33 nucleotide unique position changes resulted in 14 amino acid changes, 1 each in ORFs 6, 9A, 10, 31, 39, and 52 and 2 each in ORFs 14, 55, and 62/71 and the R3 repeat region.

Among these 36 position differences between Oka-V_GSK and Oka-V_Merck, Oka-V_GSK had 23 nucleotide sequences identical to Oka-V but only 3 identical to Oka-P. In contrast, Oka-V_Merck had 18 positions identical to Oka-P but only 6 identical to Oka-V (Table 6 and Fig. 3).

• Open in new tab
• Download powerpoint

FIG. 3.

Sequence comparisons of Oka-V_GSK and Oka-V_Merck with Oka-P and Oka-V strains of VZV. The 36 nucleotide positions that are different in Oka-V_GSK and Oka-V_Merck vaccines were compared to the sequence of the original vaccine strain Oka-V and its parental virus, Oka-P.

Transactivation.ORF 62 encodes immediate early protein 62 (IE62), also known as a transcription regulator, which is the major component of the virion tegument and an important transactivating protein for all classes of VZV promoters (28, 49, 56). It is located in the short repeat sequences and has therefore a duplicate gene, ORF 71. These two duplicated genes cover 7% of the whole VZV genome. Recent studies suggested that ORF 62 could play a central role in the attenuated phenotype of the Oka vaccine strains (3, 20-22). Defined amino acid substitutions in ORF 62 that are associated with individual virus variants purified from the vaccine mixture have been linked to enhanced virus growth and spread in monolayer cell culture (22).

The present analysis of vaccine strains confirmed that a high number of mutations could be detected within ORF 62 (20-22). As previously discussed, these SNPs in ORF 62/71 may be important for attenuation of VZV (50, 69). The current analysis identified a nucleotide transition (position 105356 in ORF 62, corresponding to 124541 in ORF 71) that altered an Ile of the IE62 protein to a Val only partially in Oka-V_Merck and completely in Oka-V_GSK and Oka-V. Because Oka-P encodes only Ile at this position, it is likely that the Oka-V_Merck passaging history has selected for minor Oka-P-related species that might be present in the Oka-V vaccine. The second substitution (position 107599 in ORF 62 and 122298 in ORF 71) partially changed a Val to Ala in Oka-V_GSK and Oka-V; for Oka-V_Merck, this position is identical to the one found in Oka-P and encodes Val only. In both cases, the amino acids involved are small hydrophobic residues. Gomi et al. demonstrated that five amino acid substitutions, including the 105356 mutation, in the carboxyl terminus of IE62 directly reduced transactivational activity (22). Experiments with recombinant VZV will be required to determine how these mutations in ORF 62 modulate VZV gene expression and which amino acid substitutions are responsible for the differences in viral spreading.

The product of ORF 10, the virion-associated transactivator, is a tegument protein that regulates the IE62 promoter (28, 46). A nucleotide substitution (C→C/T) at position 12779 results in a conversion of an Ala in the Dumas, Oka-P, and Oka-V_Merck strains to a mixture of Ala and Val in the Oka-V and Oka-V_GSK strains. Similarly, minor subspecies from Oka-V that were originally present in Oka-P may have been selected in Oka-V_Merck. This position, which corresponds to a location in the middle of the protein, in Oka-V and Oka-V_GSK encodes two small hydrophobic amino acids (Val and Ala) that could have similar functions. Indeed, no statistically significant differences in transactivational activity of the ORF 10 gene product could be detected between the wild type and the mutant form, suggesting that this alternative form of ORF 10 has a minimal effect on viral attenuation through modulation of the expression level of IE62 (22). Furthermore, in vitro studies have shown that ORF 10 product was dispensable for VZV replication in vitro (10).

The helicase-primase complex.The helicase-primase complex consists of three proteins encoded by ORF 6 (primase) and ORFs 52 and 55 (helicase). Interestingly, we found four amino acid substitutions in these proteins, three of which were described previously (22).

The first amino acid substitution was located near the C terminus of ORF 6 (position 5745, A→G), which was a Ser in Dumas and Oka-P, a Pro in Oka-V and Oka-V_GSK, and a mixture of both in Oka-V_Merck. Pro is a rigid residue that could induce substantial changes in the protein conformation. This nucleotide substitution was also comprised in an AluI restriction site. Interestingly, Quinlivan et al. found no differences between Oka-V_GSK and Oka-V_Merck by AluI restriction analysis: both Oka-V_GSK and Oka-V_Merck were A/G (±AluI), whereas Oka-V was G (+AluI) (52).

The second substitution occurred in ORF 52 (position 90535, A→A/G). In this case, the amino acid residue in Oka-P and Oka-V_Merck was Ile, which was partially changed to Val in Oka-V and Oka-V_GSK.

The last two substitutions were located in ORF 55. Val at position 97479 was partially replaced by Ala in Oka-V_Merck, and Cys at position 97796 was partially replaced by Arg in Oka-V and Oka-V_GSK.

Gomi et al. demonstrated that pathogenicity and spreading of VZV were affected by mutations in ORFs 6, 10, and 62, whereas ORFs 52 and 55 did not seem to be important for efficient VZV spreading (22). Because these substitutions were only partial, they could result in the coexistence of different helicase-primase activities resulting from different isomeric complexes, as shown by restriction fragment length polymorphism analysis (22). For most of these positions Oka-V_GSK was similar to Oka-V, whereas Oka-V_Merck was similar to Oka-P.

Envelope glycoproteins.VZV produces at least seven glycoproteins, gK, gC, gB, gH, gL, gI, and gE, which are the products of ORFs 5, 14, 31, 37, 60, 67, and 68, respectively (11). Two putative additional glycoproteins were recently described, gN (ORF 9A) and gM (ORF 50) (55, 77). It is known that VZV glycoproteins induce a strong humoral immune response following either natural infection or vaccination with the Oka strain. The SNPs in the nine VZV glycoproteins were reviewed in a recent comparative analysis (64). Some of them are specific to the vaccine strains and thus could be involved in VZV attenuation.

The product of ORF 68, gE, is the most abundant glycoprotein expressed during infection. A single amino acid substitution in this protein was shown to induce the accelerated replication phenotype of the VZV-MSP mutant strain (57). Recently, Grose at al. reported the sequences of two VZV isolates harboring a D150N mutation within ORF 68 (25). That study identified only one mutation, common to all four Oka strains (position 115926, C→T), which induced the replacement of a Thr by Ile.

The product of ORF 31, gB, is the second most abundant and immunogenic envelope glycoprotein of VZV after gE. Along with gH and gC, it seems to play a role in the attachment and penetration of viral particles. It was also shown to have important fusogenic properties in the presence of gE and to be associated with cell-to-cell infection (39, 45). An A→G transition at position 58595 induced a conversion of Ile to Val in Oka-V_Merck and a mixture of Ile and Val in both Oka-V_GSK and Oka-V. Both amino acids are small and hydrophobic, suggesting that this substitution would probably not affect the properties of this glycoprotein.

ORF 14 exhibited one silent replacement and two amino acid substitutions in its product, gC. At position 20787, a coexistence of C and A induced the partial replacement of a Lys by Asn in Oka-V_GSK, whereas the original C residue was completely replaced by A in Oka-V_Merck, entirely replacing Lys with Asn. Both residues are hydrophilic; however, Lys is basic and Asn is polar, with an uncharged side chain. Because Asn was not detected in either Oka-P or Oka-V, it is likely that this amino acid substitution evolved as a result of additional vaccine passages and might reflect additional cell culture adaptation. The second modification (position 20879, A→T) was found only in the Oka-V_GSK strain. It partially changed a Ser to Thr. Both residues are hydrophilic and polar with an uncharged side chain. Grinfeld et al. showed that products of ORF 14 and ORF 67 were dispensable for the establishment of latency in a rat model (24). However, experiments with SCID-hu mice showed that gC is important for viral tropism in skin cells and that a decrease in gC plays a critical role in attenuation (44). It was also previously shown that expression of gC is dependent on the strain of VZV, with the gC level of Oka-V being much lower than that of wild-type viruses (29, 37).

A nucleotide substitution at position 71252 induced the replacement of a Met by a Thr in the product of ORF 39, one of the two multiply inserted membrane proteins of VZV. This change was complete in Oka-V_GSK and Oka-V but only partial in Oka-V_Merck. Although both amino acids are neutral, Thr is polar and Met is hydrophobic. This difference may have some effect on the properties of the protein, although these are unclear at present (23).

We also found a mutation in the gN envelope glycoprotein. This glycoprotein is the product of ORF 9A, a newly identified gene positioned closely upstream of the ORF 9 initiation codon (55). Glycoprotein gN is an 87-amino-acid protein whose amino-terminal extremity overlaps with the first nine amino acids of the ORF 8 product, coded on the complementary strand. The observed T-to-T/C shift at position 10900 involves a Trp-to-Arg switch in the very last amino acid of the protein in Oka-V and Oka-V_GSK. This change leads to replacement of a hydrophobic amino acid with a large aromatic side chain at the carboxyl terminus of the protein by a hydrophilic basic amino acid, which could alter the membrane topology of the protein or its stability.

R3 repeat region.The R3 reiteration region is located in ORF 22, the longest ORF in VZV. The product of ORF 22 is homologous to the UL36 virion tegument phosphoprotein of herpes simplex virus type 1 (41, 42). R3 is a highly variable region consisting of repeated elements that can vary in number and combination (12, 22), but the impact of this region on the function of ORF 22 phosphoprotein remains mostly unknown because the function of the phosphoprotein itself is poorly understood.

Despite the high variability of the R3 region, no frameshift mutations were detected in ORF 22, because the repeated elements are present as multiples of 3 bp. Nevertheless, two mutations (positions 41485 and 41494, C→T) in the Oka-V_GSK strain were identified, both of which convert Ala into Val, another amino acid with similar properties that probably does not affect the function of the ORF 22 protein.

Attenuation and reactivation.Comparison of the four Oka strain sequences of VZV indicated that Oka-V_GSK is genetically closer to Oka-V than Oka-V_Merck is but that Oka-V_Merck is closer to Oka-P than Oka-V_GSK is (reference 50 and the present study). In agreement with these findings, Sauerbrei et al. demonstrated that Oka-V_Merck is genetically closer to Oka-P than Oka-V_GSK is (59). Nevertheless, it is well established that both vaccine strains are attenuated but remain strongly immunogenic (43, 73). Recently, Quinlivan et al. suggested an association of particular SNPs in the VZV genome with frequency of vaccine-induced rash (53). Four SNP positions were suggested to contain nucleotides specific to Oka-P in most of the viruses isolated from vaccine rashes. The first one is a silent nucleotide change within ORF 51 (A→G at position 89734). The nucleotide at this position is A in Oka-V_Merck and Oka-P but G in Oka-V_GSK and Oka-V. The second SNP, at position 105169, contains mixed A/G nucleotides for Oka-V_Merck, Oka-V_GSK, and Oka-V, whereas the parental Oka-P contains only A. The third SNP, at position 105356, is located within ORF 62. The change from T to C is responsible for an amino acid switch (Ile→Val) for both Oka-V_GSK and Oka-V. At the same position, the parental Oka-P contains T (Ile), and Oka-V_Merck contains a mixture of T and C (Ile/Val). The last position (nucleotide 107797) was not identified as a SNP in our sequencing data, which agrees with a previous study (22).

Although the genetic basis of Oka-V attenuation has not been determined, Oka-V and Oka-P genomes have nucleotide differences predicted to change amino acids in every class of viral proteins (3, 21). VZV attenuation is a multifactorial phenomenon whose mechanism remains unclear (80), but it is conceivable that mutations of the vaccine genome, in particular, mutations resulting in amino acid modifications, could affect virulence or latency of the vaccine strain. Recently, Peters and coworkers sequenced 11 VZV genomes from different clades, bringing the current number of available full-length VZV sequences to 18 (50, 69). To assess variations that can occur during serial passage in cell culture, these studies included the four Oka strains (Oka-P and the three Oka vaccine strains) and a VZV strain sequenced at passages 5, 22, and 72. As discussed by Tyler et al. (69), the SNPs in ORF 62/71 found in the three Oka vaccine strains and in the VZV strain at high passage level (S628G, R958G, and I1260V in IE62) could be involved in the attenuation of VZV. In addition, it was suggested that other mutations could play a role, particularly those in regions containing ORFs 30 to 55 (69, 78, 80). However, we found that numerous SNPs contain the original Oka-P nucleotide, supporting the idea that the attenuated phenotype is the result of a cooperative effect between several SNPs rather than the result of selected mutations. Therefore, our analysis of SNP importance is aligned with the conclusions of Tyler et al. regarding VZV attenuation (69).

VZV remains latent in sensory-nerve ganglia and can reactivate later, causing herpes zoster. It was suggested that herpes zoster is less common after vaccination, because initial access of Oka-V to neural cells is reduced by limited skin replication or because Oka-V reactivation and secondary viral infection of skin are less efficient, rather than because of an intrinsic attenuation of Oka-V neurotropism (5). Indeed, even though the Oka-V strain of VZV can cause herpes zoster, it seems to reactivate less often than the wild-type VZV even in immunocompromised children (18, 27, 68), and increased incidences of reactivation after vaccination have not been demonstrated in either clinical studies or postmarketing surveillance (33a, 43, 62, 66, 70, 73). In addition, the Oka vaccines have been shown to elicit a strong and protective immune response against varicella (7, 43, 73).

Conclusion.Overall, this study shows that, throughout the entire VZV genome, only 36 nucleotide positions differ between the Oka-V_GSK and Oka-V_Merck vaccine strains. Analysis of the complete genome of VZV also shows that, genetically, Oka-V_GSK is closer to Oka-V and that Oka-V_Merck is closer to Oka-P. Although Oka-V_GSK and Oka-V_Merck exhibit differences, there is a high degree of conservation between these strains at both the nucleotide and amino acid levels. This result supports the clinical data showing that both vaccines are well tolerated and elicit strong immune responses against varicella.