Skip Navigation

Nucleotide Sequences that Distinguish Oka Vaccine from Parental Oka and Other Varicella-Zoster Virus Isolates

  1. Takele Argaw1,
  2. Jeffrey I. Cohen2,
  3. Michael Klutch1,
  4. Kristen Lekstrom2,
  5. Tetsushi Yoshikawa3,
  6. Yoshizo Asano3 and
  7. Philip R. Krause1
  1. 1Division of Viral Products, Office of Vaccines Research and Review, Center for Biologics Evaluation and Research, Food and Drug Administration, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland
  2. 2Laboratory of Clinical Investigation, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland
  3. 3Department of Pediatrics, Fujita Health University, Aichi, Japan
  1. Reprints or correspondence: Philip R. Krause, 29A/1 C16, 29 Lincoln Dr., Bethesda, MD 20892-4555 (krause{at}cber.fda.gov).
 
Next Section

Abstract

The sequences of ∼34 kb from the 3′ end of the varicella-zoster virus (VZV) Oka vaccine strain and the previously sequenced Dumas strain were compared. Sequence differences were noted in the coding sequences of several VZV open reading frames (ORFs), including ORFs 48, 51, 52, 55, 56, 58, 59, 60, 62, 64, and 68. Tests based on differences in the ORF62 gene and in the ORF64 poly-A region successfully distinguished the Oka vaccine strain from its wild-type parent and from other Japanese and US clinical isolates. These changes remained stable after passage of the virus in humans.

Varicella-zoster virus (VZV) causes chickenpox, a common disease of children, which is characterized by a disseminated vesicular rash with fever [1]. A live, attenuated varicella vaccine, which uses the Oka strain of VZV, was licensed in the United States in 1995 for prevention of chickenpox [2]. This vaccine strain was derived by serial passage in cell culture of a viral isolate obtained in the early 1970s from a child with chickenpox [3].

Several rash-associated clinical syndromes have been described in vaccinees. Vaccinees may experience localized or disseminated rashes within a few weeks after vaccination. Some vaccinees are incompletely protected from chickenpox, and, when exposed to circulating wild-type VZV, they experience a milder form of the disease, termed “breakthrough” chickenpox. In rare cases, transmission of the vaccine strain to a secondary contact has been described [4–7]. In addition, the vaccine strain may reactivate, causing a zoster-like illness similar to that caused by reactivation of the wild-type virus. Distinguishing vaccine strains of VZV from circulating wild-type strains is important in determining which of these rash syndromes and other side effects are attributable to vaccine strains and which are attributable to the wild-type virus.

When virus can be cultured, strains may be differentiated on the basis of restriction-endonuclease digestion of whole virus DNA [8]. This method is time consuming and cannot be used in cases in which virus DNA is present but the virus cannot be recovered in tissue culture. Cell culture methods that distinguish vaccine and wild-type Oka strains on the basis of differences in temperature sensitivity and growth in guinea pig cells have been described [9], but these methods also are not practical for routine laboratory use.

Two genetic differences that lead to restriction-endonuclease polymorphisms have been used in rapid polymerase chain reaction (PCR)-based assays to distinguish the Oka vaccine strain from wild-type strains that typically circulate in the United States. These differences lead to the loss of a PstI site in open reading frame (ORF) 38 and to the gain of a BglI site in ORF54 of the Oka vaccine strain, relative to most US wild-type viruses [10–12]. In US studies, 100% of US wild-type viruses had the ORF38 PstI site, and 80% lacked the ORF54 BglI site; thus, 100% could be differentiated from the vaccine strain on the basis of a combination of these differences. In Japan, the ORF54 mutation does not distinguish the Oka vaccine strain from most circulating strains of wild-type VZV. Thus, the ORF38 (PstI) mutation is used together with a polymorphism in the R2 repeat region to differentiate between these strains [13, 14]. Although either of these Oka-characteristic sequences may be observed among typical Japanese wild-type strains, it is unusual for the strains to possess both. Japanese zoster cases in individuals who were exposed to chickenpox at the same time that the Oka parent virus was isolated appear to reflect Japanese wild-type viruses from the 1970s, which may not be distinguished from the Oka vaccine strain by these methods. Similarly, the wild-type Oka parental strain of VZV is not distinguished from the vaccine strain by these methods. Thus, although it is usually possible to rule out the presence of the Oka vaccine strain, some cases of wild-type viruses may be misclassified as the Oka vaccine strain, especially those from Japan.

To identify a mutation that is specific for the vaccine strain, we partially sequenced a cloned VZV MstIIA cosmid derived from the Japanese vaccine strain [15]. This sequence was compared with that of the Dumas laboratory strain, for which the entire sequence is available [16]. For cases in which sequence differences led to more dramatic (nonconservative) amino acid changes (e.g., changes in charge or hydrophilicity) in known ORFs, we also sequenced portions of the Oka parental strain (the original clinical isolate from which the Oka vaccine strain was derived [3]), to determine whether these changes could distinguish the Oka vaccine strain from its parent or from the Oka Merck strain marketed in the United States [2]. We reasoned that any such changes might play a role in virus attenuation and could be used as a basis for an assay that could differentiate between vaccine and nonvaccine strains.

Previous SectionNext Section

Materials and Methods

Viruses and tissue culture

All VZV nucleotide positions are given relative to the Dumas strain [16]. In this article, the term “Oka vaccine” refers to marketed vaccine obtained from the United States and Japan, manufactured by Merck and Co. (West Point, PA) and by Biken (Osaka, Japan), respectively. “Oka parent” refers to the original clinical isolate from which Oka vaccine was derived; Oka parent was provided by Dr. Michiaki Takahashi (Osaka University, Osaka, Japan). “Recombinant Oka” refers to the sequence of the cloned MstIIA cosmid derived from the Japanese vaccine strain, spanning nt 84,970–124,884 of the VZV genome [15]. Japanese clinical isolates were obtained from patients in Nagoya, Japan. All VZV isolates were grown on MRC-5 human diploid lung fibroblast cells (obtained from American Type Culture Collection, Rockville, MD), which were maintained in MEM containing 2% fetal bovine serum.

Cells were treated with trypsin, recovered, and centrifuged to extract viral DNA. Cell pellets were lysed, and viral capsids were separated by high-speed centrifugation through a discontinuous glycerol gradient. The viral DNA was extracted with phenol-chloroform before precipitation with ethanol. In other cases, DNA was purified from infected cells by use of the QIAamp tissue kit (Qiagen, Valencia, CA).

Sequencing and PCR

Automated DNA sequencing of recombinant Oka was performed on an ABI PRISM 377 DNA Sequencer (Perkin Elmer, Foster City, CA). Sequences of DNA derived from other VZV isolates were determined in selected regions by use of one or more of the following techniques: (1) direct sequencing of gel-purified PCR products, (2) sequencing of TA-cloned (Invitro-gen, Carlsbad, CA) PCR products, and (3) restriction-endonuclease digestion of PCR products. All sequences were verified on both strands, and all differences among Oka strains were confirmed in at least 2 experiments. PCR was done, in accordance with the manufacturer's instructions, by use of the HotStarTaq DNA Poly-merase kit (Qiagen, Valencia, CA). We used the following after-cycling conditions to amplify a fragment containing the polymorphism at nt 106,262 of ORF62: 95°C for 15 min, followed by 30 cycles of 94°C for 1 min, 45°C–50°C for 1 min, and 72°C for 1 min, followed by 72°C for 10 min.

Transient expression assays

Plasmid pCMV62 [17], which contains the ORF62 gene sequence from the Scott strain of VZV under the control of the cytomegalovirus (CMV) immediate-early (IE) promoter, was mutagenized by replacing a PmlI-CpoI fragment (nt 106,232–106,897) obtained by PCR of the Oka vaccine and Oka parent strains; the altered plasmids were designated pCMV62-vac-cine and pCMV62-parent, respectively. The entire insert was sequenced for both clones and was shown to differ only at nt 106,262. We tested the ability of ORF62 protein to transactivate a VZV IE gene (ORF4) promoter and of ORF62 with ORF4 to synergistically transactivate a VZV late gene (gpC, formerly gpV) promoter. Vero cells (American Type Culture Collection, Rockville, MD) were transfected with p4CAT (containing the chloramphenicol acetyl transferase [CAT] gene under the control of the VZV ORF4 promoter [17]), together with pCMV (empty vector), pCMV62-vaccine, or pCMV62-parent. Alternatively, cells were transfected with the plasmid pgpVCAT (containing the CAT gene under the control of the VZV gpC promoter [17]) and pCMV4 (containing VZV ORF4 under the control of the CMV IE promoter [17]), together with plasmids pCMV, pCMV62-vaccine, or pCMV62-parent. Equal quantities (2–3 μg) of each plasmid were used within each experiment. After 3 days, CAT assays to measure the proportion of acetylated chloramphenicol were performed. Each experiment was performed in duplicate on at least 3 different days.

Previous SectionNext Section

Results

Oka vaccine VZV differs from its parental virus at several sites

Sequencing of ∼34 kb from the 3′ end of recombinant Oka revealed numerous changes relative to the Dumas strain. These included multiple nucleotide changes leading to nonconservative amino acid substitutions in coding sequences for virus gene products (table 1). The Oka parent strain was examined for the presence of some of these changes. Differences were observed between recombinant Oka and the Oka parent at 6 loci: at positions 105310, 105356, and 106262 in ORF62; in the polyadenylation signal for ORF64 at position 112130; at position 111650 in ORF64; and at position 101089 in ORF59.

Table 1
View larger version:
Table 1

Differences between recombinant Oka and Dumas strain, with comparison of recombinant Oka, Oka parent (P), and Oka Merck (M).

The most dramatic predicted difference in amino acid sequence in ORF62, which contained the largest number of differences between the recombinant Oka and Dumas sequences, was a substitution of C for T at nt 106,262, resulting in a substitution of glycine for arginine. This change results in the loss of a BstN1 and the gain of a SmaI restriction-endonuclease cleavage site in recombinant Oka virus DNA, relative to Dumas virus DNA. The presence or absence of these sites can be determined by the ability of these enzymes to cleave a PCR-amplified fragment of viral DNA that contains this polymorphism. PCR primers (AGGTTGGCAAACGCAGTC and ATTACTGTCGACCCGAGACC) that amplify ORF62 between nt 106,078 and nt 106,380 were used. In this assay, the band sizes after cutting amplified DNA with SmaI were 118 bp, 112 bp, and 72 bp for the vaccine strain or 230 bp and 72 bp for nonvaccine strains. The ORF62 mutation also distinguished 2 different lots of the Oka (Merck) vaccine from the Oka parent and 7 clinical isolates (2 from the United States and 5 from Japan; table 2). Seventeen other differences were also observed in ORF62 in the vaccine strain, relative to the Dumas strain. Eight of these led to amino acid differences (at least 2 of which also distinguished Oka vaccine from Oka parent strains), and 9 led to no difference in the protein-coding sequence (table 1).

Table 2
View larger version:
Table 2

Ability of tests based on nucleotide differences to correctly identify varicella isolates from the recombinant Oka vaccine strain.

The change at nt 112,130 involves the addition of 8 adenosine bases in recombinant Oka vaccine, relative to Dumas, and of 7 adenosine bases within the consensus polyadenylation signal for ORF64, relative to the Oka parent. Differences between the vaccine strain and 5 clinical isolates were also shown by direct sequencing of PCR-amplified viral nucleic acids. This difference also distinguished commercial US Oka vaccine from the Oka parent (although the difference between these strains was 3 bp). Because this does not alter any restriction-endonuclease cleavage sites, it is more difficult to screen isolates for this alteration than for the change at nt 106,262. This mutation appears unlikely to influence the actual polyadenylation of ORF64, because the core AAATAAA consensus sequence remains intact in both viruses.

Two sequence differences did not consistently distinguish the Oka parent strain from all vaccine isolates. By using a PCR-based test, we found that the change observed in ORF59 (nt 101,089) was not observed in the commercially available US Oka vaccine preparation. In addition, the US and Japanese commercial vaccines were both heterogeneous at position 111,650 in ORF64, whereas the clonal recombinant Oka represents only 1 of the 2 possibilities at this position.

PCR can rapidly distinguish Oka vaccine from Oka parent and other VZV isolates

Table 2 compares the ability of the ORF62 and ORF64 poly-A changes to distinguish Oka vaccine, parent, and clinical isolate strains with that of previously used polymorphisms in ORF54 and ORF38. Neither of the previously used polymorphisms distinguished Oka vaccine from Oka parent. The ORF54 (BglI) polymorphism failed to distinguish 3 of 4 wild-type varicella strains in the United States and Japan from the Oka vaccine, identifying only 1 of 2 US strains as unique. The ORF38 (PstI) mutation successfully distinguished both US isolates from the vaccine strain but distinguished only 1 of the 2 Japanese varicella isolates.

In contrast, all the viruses cultured from children with natural chickenpox, from the United States and Japan, were found to have a nonvaccine genotype, on the basis of both the ORF62 (nt 102,262) and ORF64 poly-A (nt 112,130) polymorphisms. In addition, all 13 of the tested viruses from Japanese patients with herpes zoster (representing patients with a greater likelihood of having virus strains that were circulating in Japan at or before the time of vaccine development) were also identified as wild type on the basis of the ORF62 polymorphism.

Nonconservative amino acid substitution at nt 106,262 in ORF62 does not reduce the transactivating activity of vaccine strain ORF62

To study whether the sequence difference between vaccine and parental virus at nt 106,262 of ORF62 might alter the ability of this regulatory gene product to influence expression of other viral genes, we constructed 2 plasmids expressing ORF62, differing only at nt 106,262, under the control of the CMV IE promoter.

The ORF62 construct with the Oka vaccine sequence at nt 106,262 had a slightly greater transactivating activity for the ORF4 promoter (1.8-fold) than the Oka parent ORF62 construct. However, this difference did not reach statistical significance by Student's t test. When transfected together with pCMV4, the ORF62 Oka vaccine construct had greater transactivating activity for the VZV gpC promoter (mean, 2.5-fold; P = .03 by Student's t test), when compared with the ORF62 Oka parent construct. These results imply that the change at nt 106,262 may modestly increase the ability of vaccine ORF62 protein to transactivate late gene products, relative to the parent ORF62 protein.

Two nucleotide changes identified in Oka vaccine viruses are stable after passage in humans

Tests based on the ORF62 (nt 106,262) and ORF64 poly-A (nt 112,130) mutations correctly identified the Oka vaccine strain obtained from vaccinees who developed varicella-like rashes after immunization (table 2). These viruses were also identified as the Oka vaccine strain by the ORF54 and ORF38 mutations. These results indicate that these specific nucleotide sequences in the vaccine strain remained unchanged after passage in humans. These results are consistent with clinical studies showing that vaccine virus transmitted from viremic recipients to their siblings appears to maintain an attenuated phenotype [7].

Previous SectionNext Section

Discussion

This study reports the first known sequence differences between the Oka vaccine strain and the wild-type Japanese isolate from which it was derived and identifies a restriction-fragment polymorphism and a set of nucleotide insertions that appear to distinguish vaccine from wild-type strains. Thes differences appear stable in humans in a limited test, but further study would be required to be certain that they could not revert after multiple passages in humans. Study of more isolates will also be useful in further assessing the diagnostic value of these sequence differences. Other VZV sequence polymorphisms identified in this study may also be of value in epidemiological studies of VZV strains.

The homogeneity of wild-type viruses at nt 106,262 of ORF62, which has a different sequence in the vaccine isolate, implies that there may be some selective pressure against mutation of this base among wild-type isolates. Of the genes se-quenced, the greatest number of mutations relative to Dumas was observed in ORF62. It is thus tempting to speculate that this or other ORF62 mutations could play a role in the relative attenuation of the Oka vaccine strain. In a previously reported study, mutation of specific amino acid residues (nt 107,490– 107,507) of ORF62 impaired the DNA recognition and regulatory functions of the gene [18]. In the transient expression assays reported in the current study, the change at nt 106,262 modestly increased the ability of ORF62 to transactivate gpC in conjunction with ORF4.

The differences observed at nt 101,089 in ORF59 and at nt 105,010 between the Japanese vaccine-derived recombinant Oka and both its parent and the US vaccine suggest that there might be a difference between US and Japanese vaccines at these sites, that a minority population of viruses in the vaccine contains these polymorphisms, or that these changes are related to the cloning and propagation of recombinant Oka in bacteria. The heterogeneity of commercial vaccine (but not the Oka parent or wild-type strains) at nt 111,650 in ORF64 suggests that this mutation may be useful for identifying the Oka vaccine strain. However, the absence of this mutation in an isolate would not rule out the presence of the vaccine strain. This polymorphism may be of use in studying the effect on the polyclonal nature of the vaccine of differences in passage levels or in manufacturing techniques.

Although it is possible that attenuation of the vaccine strain could be attributable to increased transactivation activity of ORF62, it is difficult to postulate a mechanism by which up-regulation of a factor such as gpC would decrease viral virulence. Thus, this in vitro study suggests that additional differences between the vaccine virus and its parent (possibly elsewhere in ORF62) are likely to be responsible for the attenuation of the vaccine strain, relative to its parent and other wild-type VZVs. Because both viruses have been passaged repeatedly in tissue culture, comparison of the Oka vaccine strain with the Dumas strain may not identify all differences between the Oka vaccine strain and its parent. The sequences presented here represent a conservative estimate of these differences and may provide a basis for additional direct comparisons between the Oka vaccine strain and its parent.

Previous SectionNext Section

Acknowledgment

We thank Phil Brunell for supplying the clinical isolates of varicella-zoster virus and Koichi Yamanishi for helpful discussions.

Previous SectionNext Section

Footnotes

  • Grant support: National Vaccine Program via an appointment to the postgraduate research participation program at the Center for Biologics Evaluation and Research that was administered by the Oak Ridge Institute for Science and Education through the interagency agreement between the US Department of Energy and the Food and Drug Administration (FDA).

  • The opinions expressed in this article are those of the authors. No endorsement by the FDA is implied or should be inferred.

  • Received August 11, 1999.
  • Revision received November 23, 1999.
Previous Section
 

References

  1. 1.↵
    1. Cohen JI,
    2. Brunell PA,
    3. Straus SE,
    4. Krause PR
    . NIH conference: recent advances in varicella-zoster virus infection. Ann Intern Med 1999;130:922-32.
    Abstract/FREE Full Text
  2. 2.↵
    1. Krause PR,
    2. Klinman DM
    . Efficacy, immunogenicity, safety, and use of live attenuated chickenpox vaccine. J Pediatr 1995;127:518-25.
    CrossRefMedlineWeb of Science
  3. 3.↵
    1. Takahashi M,
    2. Otsuka T,
    3. Okuno Y
    . Live vaccine used to prevent the spread of varicella in children in hospital. Lancet 1974;2:1288-90.
    MedlineWeb of Science
  4. 4.↵
    1. Hughes P,
    2. LaRussa P,
    3. Pearce JM,
    4. LePow M,
    5. Steinberg S,
    6. Gershon A
    . Transmission of varicella-zoster virus from a vaccinee with leukemia, demonstrated by polymerase chain reaction. J Pediatr 1994;124:932-5.
    CrossRefMedlineWeb of Science
  5. 5.
    1. Salzman MB,
    2. Sharrar RG,
    3. Steinberg S,
    4. LaRussa P
    . Transmission of varicella-vaccine virus from a healthy 12-month old child to his pregnant mother. J Pediatr 1997;131:151-4.
    CrossRefMedlineWeb of Science
  6. 6.
    1. Tsolia M,
    2. Gershon AA,
    3. Steinberg SP,
    4. Gelb L
    . Live attenuated varicella vaccine: evidence that the virus is attenuated and the importance of skin lesions in transmission of the varicella-zoster virus. J Pediatr 1990;116:184-9.
    CrossRefMedlineWeb of Science
  7. 7.↵
    1. LaRussa P,
    2. Steinberg S,
    3. Meurice F,
    4. Gershon A
    . Transmission of vaccine strain varicella-zoster virus from a healthy adult with vaccine-associated rash to susceptible household contacts. J Infect Dis 1997;176:1072-5.
    Abstract/FREE Full Text
  8. 8.↵
    1. Brunell PA,
    2. Geiser CF,
    3. Novelli V,
    4. Lipton S,
    5. Narkewicz S
    . Varicella-like illness caused by live varicella vaccine in children with acute lymphocytic leukemia. Pediatrics 1987;79:922-7.
    Abstract/FREE Full Text
  9. 9.↵
    1. Hayakawa Y,
    2. Torigoe S,
    3. Shiraki K,
    4. Yamanishi K,
    5. Takahashi M
    . Biologic and biophysical markers of a live varicella vaccine strain (Oka): identification of clinical isolates from vaccine recipients. J Infect Dis 1984;149:956-63.
    Abstract/FREE Full Text
  10. 10.↵
    1. Hawrami K,
    2. Breuer J
    . Analysis of United Kingdom wild-type strains of varicella-zoster virus: differentiation from the Oka vaccine strain. J Med Virol 1997;53:60-2.
    CrossRefMedlineWeb of Science
  11. 11.
    1. LaRussa P,
    2. Lungu O,
    3. Hardy I,
    4. Gershon A,
    5. Steinberg S,
    6. Silverstein S
    . Restriction fragment length polymorphism of polymerase chain reaction products from vaccine and wild-type varicella-zoster virus isolates. J Virol 1992;66:1016-20.
    Abstract/FREE Full Text
  12. 12.↵
    1. LaRussa P,
    2. Steinberg S,
    3. Arvin A,
    4. et al
    . Polymerase chain reaction and restriction fragment length polymorphism analysis of varicella-zoster virus isolates from the United States and other parts of the world. J Infect Dis 1998;178:S64-6.
    CrossRefMedline
  13. 13.↵
    1. Mori C,
    2. Takahara R,
    3. Toriyama T,
    4. Nagai T,
    5. Takahashi M,
    6. Yamanishi K
    . Identification of the Oka strain of the live attenuated varicella vaccine from other clinical isolates by molecular epidemiologic analysis. J Infect Dis 1998;178:35-8.
    Abstract/FREE Full Text
  14. 14.↵
    1. Takada M,
    2. Suzutani T,
    3. Yoshida I,
    4. Matoba M,
    5. Azuma M
    . Identification of varicella-zoster virus strains by PCR analysis of three repeat elements and a PstI-site-less region. J Clin Microbiol 1995;33:658-60.
    Abstract/FREE Full Text
  15. 15.↵
    1. Cohen JI,
    2. Seidel KE
    . Generation of varicella-zoster virus (VZV) and viral mutants from cosmid DNAs: VZV thymidylate synthetase is not essential for replication in vitro. Proc Natl Acad Sci USA 1993;90:7376-80.
    Abstract/FREE Full Text
  16. 16.↵
    1. Davison AJ,
    2. Scott JE
    . The complete DNA sequence of varicella-zoster virus. J Gen Virol 1986;67:1759-816.
    Abstract/FREE Full Text
  17. 17.↵
    1. Perera LP,
    2. Mosca JD,
    3. Ruyechan WT,
    4. Hay J
    . Regulation of varicella-zoster virus gene expression in human T lymphocytes. J Virol 1992;66:5298-304.
    Abstract/FREE Full Text
  18. 18.↵
    1. Tyler JK,
    2. Allen KE,
    3. Everett RD
    . Mutation of a single lysine residue severely impairs the DNA recognition and regulatory functions of the VZV gene 62 transactivator protein. Nucleic Acids Res 1994;22:270-8.
    Abstract/FREE Full Text
| Table of Contents

This Article

  1. J Infect Dis. (2000) 181 (3): 1153-1157. doi: 10.1086/315335
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 205 (5)
  1. Journal of Infectious Diseases
  1. Alert me to new issues

Most