TEXT

Top Abstract Text References

Kaposi's sarcoma-associated herpesvirus (KSHV) is postulated to be the infectious cause of Kaposi's sarcoma (KS) (for reviews, see references 5 and 19). Due to the strong similarities in sequence and gene organization with herpesvirus saimiri (HVS), the prototype of the gamma-2 (Rhadinovirus) genus of the gammaherpesvirus subfamily, KSHV has been classified as a human rhadinovirus (6, 13). We have previously identified DNA sequences related to KSHV in retroperitoneal fibromatosis (RF) lesions from two macaque species, the pig-tailed macaque (Macaca nemestrina) and the rhesus macaque (Macaca mulatta) (17). RF is a vascular fibroproliferative neoplasm with similarities to KS which was prevalent in the macaque colony in the Washington Regional Primate Research Center (WaRPRC) during the late 1970s and early 1980s (9, 10). The macaque KSHV-like sequences were identified using a novel consensus-degenerate hybrid oligonucleotide primer (CODEHOP) PCR technique (16), which was employed to detect unknown herpesvirus DNA polymerase genes. Phylogenetic analysis of the available sequence data suggested that the DNA polymerase fragments were derived from macaque homologs of human KSHV, with a unique genotype present in each macaque species. These macaque homologs were designated RF-associated herpesvirus-M. nemestrina (RFHVMn) and -M. mulatta (RFHVMm). Subsequently, an additional simian homolog of KSHV was identified in an M. mulatta from the New England Regional Primate Research Center, and approximately 10 kb of the viral genome, including the DNA polymerase and flanking regions, was sequenced (8). Because of its sequence similarity to KSHV and HVS, this new homolog was designated rhesus rhadinovirus (RRV) isolate H26-95. Another isolate of RRV, RRV-17577, was identified in a simian immunodeficiency virus-infected rhesus macaque with a lymphoproliferative disorder at the Oregon Regional Primate Research Center, and its complete genomic sequence was determined (20). Due to the similarities in gene sequence and genomic structure, it was concluded that RRV was the macaque homolog of KSHV (8, 20). However, comparison of the partial DNA polymerase sequences of RFHVMn and RFHVMm with the corresponding region of RRV suggested that the putative RFHVMn and RFHVMm were more closely related to KSHV than was RRV (4). More recently, DNA fragments from two distinct KSHV-like herpesviruses have been identified in African green monkeys, with one virus, designated Chlorocebus rhadinovirus type 1 (ChRV-1), more closely related to KSHV and the other virus, ChRV-2, more closely related to RRV (11).

In order to further characterize the macaque viruses and establish their evolutionary relationship to KSHV and other rhadinoviruses, we have cloned and sequenced the entire DNA polymerase gene and the upstream glycoprotein B and downstream open reading frame 10 (ORF 10) flanking regions from the genomes of RFHVMn, RFHVMm, and KSHV. We have also cloned and characterized the majority of the DNA polymerase gene and the downstream flanking ORF 10 region of a close variant of RRV identified in M. nemestrina. Using these sequences, we have determined a phylogenetic relationship between KSHV and the macaque rhadinoviruses.

Identification of an RFHVMn isolate from an RF tumor of Mn442N. In our original studies of RFHVMn, only archived samples of paraffin-embedded, formalin-fixed samples of RF tissue were available (17). Subsequently, a simian retrovirus type 2-infected M. nemestrina no. 442N (Mne442N) from the National Institutes of Health, Bethesda, Maryland, was diagnosed with RF (21), and small frozen tissue specimens were collected after necropsy (kindly provided by Riri Shibata). DNA was extracted and amplified by PCR using the CODEHOP technique with the SLYP1A upstream primer and the GDTD1B downstream antisense primer (Fig. 1 and Table 1). This assay is similar to that described in the original identification of RFHVMn and RFHVMm (17), except that the DFASA primer was replaced with SLYP1A, which was designed with less degeneracy. A PCR amplification of 35 cycles (1 min at 94°C, 1 min at 60°C, and 1 min at 72°C) was performed in 0.067 M Tris-HCl (pH 8.8), containing 4 mM MgCl₂, 0.016 M (NH₄)₂SO₄, 0.01 M 2-mercaptoethanol, and 100 µg of bovine serum albumin per ml (17). A PCR product of the expected size was obtained. This DNA fragment was purified from excess primers with a Qiagen spin column, cloned into the pCR2.1 T-cloning vector (Invitrogen), and sequenced on an ABI model 377 automated sequencer using Big Dye terminator technology. The DNA sequence obtained was 99.4% (463 of 466 bp) identical to the original published RFHVMn fragment from M. nemestrina MneM78114 from the WaRPRC (17), suggesting that the NIH and WaRPRC macaques contained very similar but nonidentical isolates of RFHVMn (Table 2).

View larger version (11K): [in this window] [in a new window]   FIG. 1.   CODEHOPs PCR cloning strategy. The expected linear arrangement of the glycoprotein B (ORF 8), DNA polymerase (ORF 9), and ORF 10 homologs in the KSHV-like gammaherpesviruses is indicated. The positions and orientation of CODEHOPs derived from conserved sequence motifs are shown. The name of the CODEHOPs indicates conserved amino acids within the motif, and the terminal A or B indicates sense or antisense orientation, respectively. The positions of the conserved DNA polymerase regions are indicated for reference (see reference 22).

Cloning of the DNA polymerase gene and flanking regions of KSHV, RFHVMn, and RFHVMm. To further characterize and compare KSHV and the KSHV-like herpesviruses in macaques, we chose to clone and sequence the genomic regions containing the DNA polymerase genes using the CODEHOP PCR strategy (16). In gammaherpesviruses, such as HVS, the DNA polymerase gene (also known as ORF 9) is flanked upstream by the glycoprotein B gene (also known as ORF 8) and downstream by the homolog of the HVS ORF10 gene. We, therefore, designed CODEHOPs from blocks of conserved sequences in the DNA polymerase and flanking genes to use in PCR cloning, as shown in Fig. 1 and Table 1. The CODEHOPs were designed manually or with CODEHOP prediction software that we have written and made available at http://www.blocks.fhcrc.org/blocks/codehop.html as part of the BLOCKS database of the Fred Hutchinson Cancer Research Center, Seattle, Wash. The primers are named after the predominant sequence motif in the conserved block, where the terminal A or B designation indicates sense (A) or antisense (B) orientation.

Characterization of the DNA polymerase and flanking regions of RFHVMn and RFHVMm. The nucleotide and encoded amino acid sequences of the overlapping PCR products spanning the DNA polymerase gene and flanking sequences obtained from RFHVMn isolate Mne442N (3,554 bp) and RFHVMm isolate MmuYN-91 (3,540 bp) were aligned pairwise using GenePro software (Riverside Scientific, Bainbridge Island, Wash.) (data not shown). ORF analysis and BLAST similarity searches demonstrated that both sequences encode herpesvirus homologs of the C terminus of a glycoprotein B gene, an intact DNA polymerase gene, and the N terminus of a HVS ORF 10 gene homolog in the same linear order found in KSHV (18) and HVS (2). The 475-bp PCR fragment of the DNA polymerase gene of the RFHVMn isolate that we previously identified in the RF sample from MneM78114 (17) aligned with the sequences from RFHVMn Mne442N isolate between nucleotides 2087 and 2562 with only a three-nucleotide difference (99.4% identity), as described above. The original 454-bp fragment of RFHVMm obtained from MmuYN-91 (17) was present within the cloned genomic region of RFHVMm from the same source between nucleotides 2166 and 2619 (Fig. 2).

View larger version (636668K): [in this window] [in a new window]   FIG. 2.   Comparison of the nucleotide and encoded amino acid sequences of the 3' end of the glycoprotein B gene (ORF 8), the entire DNA polymerase gene (ORF 9), and the 5' end of the ORF 10 gene for RFHVMm (top line) and RRV (bottom line). Asterisks indicate the positions of identical nucleotide sequences. The positions of identical amino acids are indicated within the RRV sequence with a · , and gaps are indicated with a -. The position of a potential TATA box in the promoter region of the DNA polymerase genes is indicated in boldface and underlined.

View larger version (68K): [in this window] [in a new window]   FIG. 3.   Amino acid sequence comparison of the DNA polymerases of the KSHV-like gammaherpesviruses. The positions of gaps and cysteine residues are boxed, and the general nucleotide polymerase conserved regions are indicated (see reference 22). Positions with identities with the KSHV sequence are shown as a · , and the unidentified N-terminal region of MneRV-2 is indicated. The numbering refers to the KSHV sequence. The source of the sequences is given in Table 2.

View larger version (18K): [in this window] [in a new window]   FIG. 4.   Comparison of the amino acid sequences encoded within the 5' end of the cloned ORF 10 homologs of HVS and the KSHV-like gammaherpesviruses. Highly conserved amino acids are boxed, and the position of the GDWE sequence motif from which the ORF 10 CODEHOPs were derived is highlighted (shaded area), with known sequences in uppercase and presumed sequences in lowercase.

Identification and characterization of an M. nemestrina variant of RRV. During the process of amplifying the Mne442N DNA above with the SLYP1A and GDTD1B DNA polymerase CODEHOPs, a second PCR clone was identified which was 65% (306 of 466 bp) identical to the RFHVMn Mne442N sequence. A BLAST search of the GenBank DNA database with the new DNA sequence as a probe revealed a close similarity to the DNA polymerase gene of the RRV isolate H26-95 (accession no. AF029302) having 424 of 466 (91%) identical nucleotides. This suggested that we had identified a DNA polymerase fragment of an M. nemestrina homolog of RRV, which we have provisionally named "M. nemestrina rhadinovirus type 2 (MneRV-2)," since it is the second rhadinovirus identified in M. nemestrina after RFHVMn. (After this manuscript was submitted for publication, two reports describing the existence of a virus similar to MneRV-2 were published. The virus was alternatively designated pig-tailed rhadinovirus (12) and pig-tailed monkey rhadinovirus (3).

Sequence comparison of RFHVMm-RFHVMn, RRV-MneRV-2, and KSHV. The alignment of the sequences of the two gammaherpesviruses, RFHVMm and RRV, both identified in M. mulatta (rhesus), is shown in Fig. 2. The alignment of the M. nemestrina herpesviruses, RFHVMn and MneRV-2, is not shown due to space considerations. Both pairs of viruses from each host macaque species showed significant similarities and distinct differences in nucleotide and encoded amino acid sequences throughout the 3' glycoprotein B fragment, the entire DNA polymerase gene, and the 5' ORF 10 gene fragment. The nucleotide sequences of RFHVMm and RRV were only 65% identical within the DNA polymerase gene, 62% identical within the fragment of the glycoprotein B gene, and 46% identical within the downstream ORF 10 gene fragment (Fig. 2 and Table 3). Comparison of the sequences of RFHVMn and MneRV-2 showed almost identical similarities. The encoded amino acid sequences of RFHVMm and RRV were 66, 53, and 22% identical within the DNA polymerase gene, the partial glycoprotein B gene fragment, and the ORF 10 gene fragments, respectively. This level of sequence similarity is greater than that seen between either of these viruses and HVS (Table 3). While the lengths of the gene and intergenic regions of RFHVMn and RFHVMm were quite similar, heterogeneity of length was detected between RFHVMn and MneRV-2 and between RFHVMm and RRV.

Phylogenetic analysis of the DNA polymerases. Phylogenetic analyses were performed on the amino acid sequences of the DNA polymerases shown in Fig. 3 with the addition of other gammaherpesvirus polymerases, as indicated in Table 2. Multiple alignments were performed using ClustalW, and gapped positions were removed. Figure 5 shows a maximum parsimony analysis (Phylip package; University of Washington, Seattle) of these sequence alignments with the tree file displayed with TreeView (15). Bootstrap analysis, performed with the programs Seqboot, Protpars, and Consense from the Phylip package, strongly supported all the branching patterns as indicated. Neighbor-joining analysis gave an equivalent topology. Recently, portions of the DNA polymerases of additional KSHV-like gammaherpesviruses have been identified in African green monkeys with a CODEHOP strategy similar to that previously described (17). Sequence analysis of the derived PCR products demonstrated that African green monkeys have a homolog of RFHVMm-RFHVMn, designated ChRV-1, and a homolog of RRV-MneRV-2, designated ChRV-2 (11). The available amino acid sequences from the African green monkey DNA fragments were used in parsimony analyses with the corresponding sequences from the other gammaherpesviruses, and the relative branch points of the ChRV-1 and ChRV-2 sequences were added to the phylogenetic tree obtained from the entire DNA polymerase sequences (Fig. 5). From these analyses, it is evident that macaques and African green monkeys contain two distinct lineages of KSHV-like rhadinoviruses. The first lineage, which we have designated rhadinovirus genogroup type 1 (RV-1), consists of RFHVMn, RFHVMm, ChRV-1, and human KSHV, which cluster together on one branch of the tree. The second lineage, which we have designated RV-2, consists of RRV, MneRV-2, and ChRV-2, which cluster together on a different branch more closely connected to the branch containing HVS, the prototype gamma-2 herpesvirus, and herpesvirus ateles type 3 (1), both rhadinoviruses of New World monkeys. These branching patterns demonstrate a distinct differentiation between the two groups of KSHV-like viruses and the prototypes of the gamma-1 Lymphocryptovirus genus, EBV, and the gamma-2 Rhadinovirus genus, HVS. Phylogenetic analysis of the available sequences derived from the glycoprotein B amino acid sequences (Table 2) confirmed this topology (data not shown).

View larger version (16K): [in this window] [in a new window]   FIG. 5.   Phylogenetic tree of the gammaherpesviruses. Phylogenetic trees were generated by maximum-parsimony analysis using the amino acid sequences of the DNA polymerases shown in Fig. 3 and Table 2. The branch points of the ChRV-1 and ChRV-2 sequences were derived from analyses of smaller sequence sets due to limited sequence data and are indicated with dotted lines. Bootstrap values derived from 100 replicates are shown for each branch point. The EBV (gamma-1) and HVS (gamma-2) prototypes (1 and 2) and the proposed rhadinovirus genogroups RV-1 and RV-2 are indicated.