Sequence Variation of Epstein-Barr Virus: Viral Types, Geography, Codon Usage, and Diseases

Abstract

One hundred thirty-eight new Epstein-Barr virus (EBV) genome sequences have been determined. One hundred twenty-five of these and 116 from previous reports were combined to produce a multiple-sequence alignment of 241 EBV genomes, which we have used to analyze variation within the viral genome. The type 1/type 2 classification of EBV remains the major form of variation and is defined mostly by EBNA2 and EBNA3, but the type 2 single-nucleotide polymorphisms (SNPs) at the EBNA3 locus extend into the adjacent gp350 and gp42 genes, whose products mediate infection of B cells by EBV. A small insertion within the BART microRNA region of the genome was present in 21 EBV strains. EBV from saliva of U.S. patients with chronic active EBV infection aligned with the wild-type EBV genome with no evidence of WZhet rearrangements. The V3 polymorphism in the Zp promoter for BZLF1 was found to be frequent in nasopharyngeal carcinoma cases from both Hong Kong and Indonesia. Codon usage was found to differ between latent and lytic cycle EBV genes, and the main forms of variation of the EBNA1 protein have been identified.IMPORTANCE Epstein-Barr virus causes most cases of infectious mononucleosis and posttransplant lymphoproliferative disease. It contributes to several types of cancer, including Hodgkin's lymphoma, Burkitt's lymphoma, diffuse large B cell lymphoma, nasopharyngeal carcinoma, and gastric carcinoma. EBV genome variation is important because some of the diseases associated with EBV have very different incidences in different populations and geographic regions, and differences in the EBV genome might contribute to these diseases. Some specific EBV genome alterations that appear to be significant in EBV-associated cancers are already known, and current efforts to make an EBV vaccine and antiviral drugs should also take account of sequence differences in the proteins used as targets.

Keywords: Epstein-Barr virus.

FIG 1

(A) PCA of 241 EBV genomes from the multiple-sequence alignment shown in a three-dimensional plot of PCA1, PCA2, and PCA3. Each genome is represented by a colored dot (geographic color codes are shown in panel B). The cluster of type 2 strains is circled. (B) Plot of PCA1 against PCA2 for the 217 type 1 strains, with each genome sequence represented by a colored dot. The Asian and Indonesian strains cluster away from African and European strains. (C) SNPs of a type 2 consensus genome relative to the type 1 consensus are plotted as number of SNPs in a 500-nt window along the EBV genome (numbered as in the MSA). The positions of EBNA2, BPLF1, gp350, EBNA3, gp42, and LMP1 coding sequences are marked. The peak marked by the asterisk is in the same location as the insertion shown in Fig. 1D. (D) Position of the insertion of 71 to 73 nt in 21 strains in the MSA of 241 EBV genome sequences on a scale numbered as in the NC_007605 reference sequence. The positions of Mir BART 21 and 18 are shown.

FIG 2

Linkage of gp42 protein sequence variation at the indicated amino acids relative to type 1 or type 2 EBNA3, comparing 24 type 2 EBNA3 protein sequences with 212 type 1 EBNA3 protein sequences. Detailed values and the phylogenetic tree are shown in Fig. S2. Significant differences were determined with an analysis of variance (ANOVA) test. **, P < 0.005.

FIG 3

(A) Schematic illustration of locations of the main EBNA1 variant amino acids under a scale of amino acid numbers of B95-8 EBNA1. The Gly-Ala region is marked, and the regions of nt 14 to 87 and 445 to 614, which were concatenated to make the phylogenetic tree (Fig. S3), are shown. (B) Numbers of strains (out of 260) with either the QEA, T20S, E24D+G2S, or T585I variant plotted in relation to amino acid 487 V, A, L, or T. In these EBNA1 sequences, 487 L or T is always found with the QCIGP haplotype. Significant differences were determined with an ANOVA test. **, P < 0.005. Full details are shown in Fig. S3.

FIG 4

(A) Percentage of G55391A and EBER2 alleles in EBV sequences from NPC or normal samples from China/Hong Kong and Indonesia. Details of strains are shown in Fig. S4, sheets 1 and 2. Significant differences were determined with an ANOVA test. **, P < 0.005. (B) Percentage of Zp P and V3 alleles in EBV sequences from NPC or normal samples from Asia, Indonesia, White British (WB)/Australia, and Kenya. Details of strains are shown in Fig. S4, sheet 3. Significant differences are indicated. **, P < 0.005. The Indonesian samples (NPC and normal samples) were collected in and around Jakarta and have been described previously ( and references therein).

FIG 5

Relative synonymous codon usage for combined open reading frames of NC_007605, AG876, and Akata EBV. (Upper) Latent and lytic cycle genes. (Lower) Lytic genes separated into early and late groups. Detailed values and results for each strain analyzed separately are shown in Fig. S5.

FIG 6

(A and B) Dot matrix plots of similarity between NKTLY97.1 and NC_007605 EBV sequences (window size, 30; minimum score, 85%; hash value, 8; MacVector). Templated BWA assembly of NKTLY97.1 (A) and de novo assembly of NKTLY97.1 (B) were used. (C) Agarose gel electrophoresis of products from PCR across the novel sequence boundary marked in panel B is shown with the 97.1 primers. PCR with WT primers (see Materials and Methods) is shown on the left.