INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Cercopithecine herpesvirus 1 (herpesvirus simiae, or monkey B virus [BV]) is an alphaherpesvirus indigenous to macaque monkeys (Macaca spp.) and is closely related to other simian herpesviruses and the herpes simplex viruses (HSV1 and HSV2) of humans (9, 10, 14). Like HSV in humans, BV causes inapparent or self-limiting localized infections of the mucous membranes (ocular, oral, and/or genital) in macaques (11, 17, 28). The virus becomes latent in sensory ganglia (21) and may cause recurrent infections or be shed asymptomatically following reactivation from the latent state (29, 30); only rarely does BV cause serious infections in macaques. In contrast, transmission of BV to other species including humans usually produces severe and often fatal infections involving the central nervous system (10, 22, 23, 28).

Of the documented human cases of BV infection, virtually all have been traced back to some form of exposure to macaque monkeys or macaque tissues such as primary cell cultures or anatomical specimens (2, 4-6, 16, 22, 28). In every case where a specific macaque species was identified, patients had exposure to rhesus monkeys (Macaca mulatta) and in some cases to cynomolgus monkeys (Macaca fascicularis) as well. No human BV cases reported in the literature have been traced back to exposure to other macaque species or to cynomolgus monkeys alone. Although no macaque species was specified for several human cases of BV occurring in the 1950s, rhesus monkeys were widely employed in biomedical research during this era. As a result, anecdotal wisdom has held for some time that BV from rhesus monkeys may be more lethal to humans than BV from other macaque species. However, no hard data exist to support this supposition or provide any rational basis for it.

One study has reported some variation in restriction fragment length polymorphism (RFLP) profiles among BV isolates and noted that two isolates from rhesus macaques had RFLP profiles different from those of a number of cynomolgus macaque BV isolates (26). We recently observed a lack of strain variation in RFLP profiles of PCR products generated from multiple strains of BV isolated from rhesus monkeys, just as is observed for human HSV and baboon herpesvirus papio 2 (HVP2) isolates (3). However, one isolate from a cynomolgus monkey did have a slightly altered RFLP profile. Similarly, Slomka et al. (25) observed that a PCR test developed by using primers based on the DNA sequence from a cynomolgus monkey BV isolate failed to detect BV isolates from rhesus monkeys. In comparing this sequence to a homologous genomic sequence from a rhesus macaque BV isolate (unpublished data), we noted significant differences between the two viruses in the sequence where these PCR primers were located, which probably accounts for the failure of this test to detect rhesus macaque BV isolates.

Taken together, these observations suggest that there may well be significant differences among BV isolates from rhesus and cynomolgus macaques. This study was undertaken to directly determine if substantial variation exists among BV isolates and, if so, whether such variation is at all related to the macaque species from which the virus was isolated. Were such differences to be found, they would provide a molecular basis for the possible existence of BV strains from different macaque species which may have different pathogenic properties in nonmacaque species.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cells, virus, and antigens. Monkey kidney cells (Vero) were used to propagate viruses and for all experimental procedures. BV isolates E2490, a standard laboratory strain originally isolated from a rhesus macaque (15); strains 12930, 16293, and 20620, isolated from rhesus macaques, and strain 1504-11, isolated from a pigtail macaque (Macaca nemistrina), were provided by J. Hilliard, Southwest Foundation for Biomedical Research. Strain E90-136, isolated from a cynomolgus macaque (24), was provided by K. Mansfield and D. Lee-Parritz, New England Regional Primate Research Center. Additional isolates provided by R. Heberling (VRL Inc., San Antonio, Tex.) included strain 9400371 isolated from a cynomolgus monkey kidney cell culture, strain Kumquat isolated from a pigtail macaque, and strains 7709609 and 7709642 isolated from Japanese macaques (Macaca fuscata).

PCR and DNA sequencing. Purified viral DNA was used in all PCRs. Sequences of primers used to sequence a 450-bp region of the UL27 (gB) gene have been published previously (3) (primers B4 and B9). Although multiple primers were used to amplify and sequence DNA from the unique short (U_S) regions of all BV isolates, two primer pairs provided initial amplification and sequence from all isolates: primers AS9 (5'-TC[A/T]CCCGGGCTAGACTT[T/C][A/C]TCTTCCTGCTCAG-3') and AS2 (5'-ATGGCGGCCAGGGTCAGCGCGCAGAGG-3') amplified approximately 650 bp extending from the 3' end of the US4 (gG) open reading frame (ORF) to near the middle of the US5 (gJ) ORF; primers AS8 (5'-CTCTGCGCGCTGACCCTGGCCGCCATGG-3') and AS7 (5'-CACGTCGGGGGG[G/A]TCCGTCTTCTGCTCC-3') amplified approximately 800 bp extending from near the middle of the US5 ORF into the US6 (gD) ORF. The GeneAmp XL PCR kit (Perkin-Elmer, Foster City, Calif.) was used for all PCR. Reaction mixtures were initially incubated for 3 min at 95°C, followed by the addition of Tth polymerase and 30 cycles of 1 min at 95°C, 1 min at 55°C, and 1 min at 72°C, with a final elongation step of 7 min at 72°C. All PCR products were gel purified before sequencing. The phylogenetic analysis of sequence data was performed by using the MEGA program package (18). Distances were estimated by the Tamura-Nei method, and phylogenetic trees were constructed by the neighbor-joining method.

Nucleotide sequence accession numbers. DNA sequences for the U_S regions of all BV strains have been deposited in the GenBank database and are available under accession no. AF082804 through AF082814 and AF083210.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Infected cell proteins and glycoproteins of 10 different isolates of BV were initially compared by SDS-PAGE. As shown in Fig. 1, there was little variation in the relative sizes or amounts of proteins or glycoproteins synthesized by the four rhesus and two Japanese macaque isolates. One of the two cynomolgus macaque isolates (9400371) was also very similar to these six isolates. The other cynomolgus macaque isolate (E90-136) showed some variation in proteins in the low-molecular-weight range. However, the two BV isolates from pigtail macaques were much more different from all the other BV isolates.

View larger version (90K): [in this window] [in a new window]   FIG. 1.   SDS-PAGE analysis of BV-infected cell proteins and glycoproteins. Vero cells were infected with BV strains and labeled from 4 to 24 h postinfection with [35S]methionine (A) or [14C]glucosamine (B). Polypeptides were separated on SDS-8% polyacrylamide gels. The monkey species from which isolates were made are identified above the lane numbers. Virus isolates shown in lanes 1 to 10 are E2490, 16293, 12930, 20620, 9400371, E90-136, 1504-11, Kumquat, 7709642, and 7709609, respectively. MCP, major capsid protein.

Based on this variation in viral proteins, several BV isolates were tested to determine if there were any detectable differences in their antigenic properties. ELISA antigens from four BV isolates, one from each macaque species, were prepared. The reactivities of sera from different species of macaques with each antigen were then compared. All macaque sera reacted well with each of the antigens (Fig. 2A), which was expected since all isolates used to prepare antigens were strains of BV. However, slight but consistent preferential reactivities of sera from cynomolgus and pigtail macaques with the homologous cynomolgus or pigtail macaque BV antigen were apparent. Rhesus monkey sera similarly exhibited slightly lower reactivities with pigtail and Japanese macaque BV isolate antigens, but were equally reactive with rhesus and cynomolgus BV antigens.

View larger version (42K): [in this window] [in a new window]   FIG. 2.   Antigenic reactivities of macaque sera and BV isolates. In a standard ELISA (A) sera from cynomolgus monkeys (352-84, 267-76, and 410-94), pigtail macaques (91099, 92209, and 90153), and rhesus monkeys (MM339-95, MM336-95, and MM329-95) were tested at a 1:4,000 dilution with antigens prepared from rhesus (E2490), cynomolgus (E90-136), pigtail (Kumquat), and Japanese (7709609) macaques. cELISA (B) was also performed with homologous serum-viral antigen combinations from rhesus (left), pigtail (center), and cynomolgus (right) monkeys. Soluble antigens used to compete these reactions were from rhesus (), cynomolgus (), and pigtail () macaque BV isolates.

To confirm these observations, cELISAs using rhesus, cynomolgus and pigtail macaque BV antigens and sera were performed. The results shown in Fig. 2B are representative of those obtained and confirm the results obtained by standard ELISA. The reactivities of rhesus macaque sera with rhesus macaque BV antigen were competed almost as well by cynomolgus and pigtail macaque antigens as by rhesus macaque BV antigen (Fig. 2B, left graph). However, rhesus macaque BV antigen consistently competed the reaction slightly better than the other two BV antigens. This preferential competition was more readily apparent in reactions in which both the antibody and antigen were from pigtail or cynomolgus macaques (Fig. 2B, middle and right graphs, respectively). In each case the homologous BV antigen competed the reaction more efficiently than either of the heterologous BV antigens. As expected, heterologous antibody-antigen reactions (e.g., rhesus macaque antibody with pigtail macaque antigen) were competed equally well by all three competing antigens (data not shown). Thus, although various BV isolates from different macaque species are antigenically very similar, there do appear to be some virus-specific determinants which are related to the host species from which the viruses were isolated.

Differences among BV strains were also assessed by RFLP analysis of genomic DNA. As shown in Fig. 3, differences between BV isolates from different macaque species were readily apparent. With all restriction enzymes tested, the two isolates from pigtail macaques had similar RFLP profiles but these profiles were distinct from those of all other BV isolates. Similarly, the two isolates from Japanese macaques and all four rhesus macaque isolates had nearly identical RFLP profiles with all enzymes tested. However, the two cynomolgus macaque isolates had RFLP profiles quite different from each other. One (9400371) had an RFLP profile identical to those of rhesus and Japanese macaque isolates, while the other (E90-136) had a characteristically distinct RFLP profile with all restriction enzymes tested.

View larger version (57K): [in this window] [in a new window]   FIG. 3.   RFLP profiles of BV isolates. Purified viral DNA from BV isolates was digested with BamHI, PstI, or SalI, and fragments were separated on 0.8% agarose gels. Isolates from rhesus macaques are in lanes 2 to 5 (strains E2490, 16293, 12930, and 20620, respectively); cynomolgus macaque isolates are in lanes 6 and 7 (strains 9400371 and E90-136, respectively); pigtail macaque isolates are in lanes 8 and 9 (1504-11 and Kumquat, respectively); and Japanese macaque isolates are in lanes 10 and 11 (7709642 and 7709609, respectively). HVP2 strain OU1-76 is shown in lane 1 for comparison. Phage  DNA cut with EcoRI and EcoRI plus HindIII was used as a size marker. Fragment sizes in kilobase pairs are indicated.

Genomic variation among BV isolates was further examined by sequencing two regions of the genome. A small region of the UL27 gene, which encodes the gB glycoprotein (19), was initially sequenced. This region spanned the D2 region of the gB gene which is highly divergent among different primate herpesviruses but is conserved among strains of any single virus (3). There was very little variation in either nucleotide or predicted amino acid sequences among BV isolates (data not shown). The two Japanese macaque BV isolates had one common silent nucleotide change that distinguished them from rhesus macaque BV isolates. The two pigtail macaque BV isolates had eight nucleotide changes that distinguished them from other isolates, five of which were silent and three of which resulted in amino acid substitutions. One of the silent nucleotide substitutions resulted in elimination of a HaeIII site, which altered the RFLP profile indicative of BV, as determined by a diagnostic PCR previously described (3). The two cynomolgus macaque isolates had no common distinguishing substitutions, although, relative to rhesus macaque BV isolates, there were five nucleotide substitutions between them (all in isolate E90-136), which resulted in two amino acid substitutions at different sites; the sequence from isolate 9400371 was identical to those of the rhesus macaque isolates. These relatively minor differences in the UL27 D2 region were consistent with the conclusion that all the isolates tested were strains of BV.

Genomic variation was further assessed by PCR amplification and sequencing of about 1.3 kbp from the U_S region of the genome (20) encompassing the 3' terminus of the US4 gene through the 5' end of the US6 gene. Unlike the D2 region of the gB gene, differences among BV isolates were readily detected in this region of the genome (Fig. 4). Within the last 19 codons of the US4 (gG) ORF, there were no differences between BV isolates from Japanese and rhesus macaques. The two pigtail macaque isolates, however, had 10 identical nucleotide substitutions relative to the rhesus macaque isolates, 5 of which resulted in amino acid substitutions. The cynomolgus macaque isolates again were different from one another. The sequence of isolate 940037 was identical to those of rhesus isolates, while isolate E90-136 had five nucleotide substitutions resulting in three amino acid differences from the rhesus macaque isolates. The published sequence for another cynomolgus macaque isolate (26) was identical to that of isolate E90-136.

View larger version (73K): [in this window] [in a new window]   View larger version (37K): [in this window] [in a new window]   FIG. 4.   Comparison of sequence from US4 (gG) through US6 (gD). DNA sequences were amplified by PCR, and products were sequenced directly. Numbering above the aligned sequences represents the positions in the alignment rather than nucleotide numbers of the E2490 sequence. All sequences are shown referenced to the E2490 strain sequence, with identical residues indicated by dots. Termination codons of US4 and US5 and start codons of US5 and US6 in the E2490 sequence are shaded; mRNA poly(A) and transcriptional termination motifs in the US5-to-US6 intergenic region are underlined. The US4 ORF runs from positions 1 to 60 of the aligned sequences; the US5 ORF runs from positions 293 to 688; and the US6 ORF runs from positions 1186 to 1318 of the alignment. The sequence for isolate SMHV is taken from Bennett et al. (1) and Slomka et al. (26).

Similar results were obtained for the sequence of the first 44 codons of the US6 (gD) ORF. DNA sequences of Japanese and rhesus macaque BV isolates were identical. There were 22 nucleotide substitutions characteristic of the two pigtail macaque BV isolates, which resulted in 11 amino acid changes. The cynomolgus macaque isolate 9400371 sequence was again identical to those of rhesus macaque isolates, while the sequence of isolate E90-136 was identical to the sequence published for another cynomolgus macaque BV strain (1). In these two viruses, there were four characteristic nucleotide substitutions resulting in two amino acid differences from the rhesus macaque isolates.

The complete coding sequence of the US5 (gJ) gene was determined for all BV isolates, and again the viruses fell into three distinct groups. Isolates from Japanese and rhesus macaques did not have any consistent sequence differences in the 122 codons of this gene which distinguished them from each other. BV isolates from pigtail macaques had 34 amino acid substitutions and 10 codon deletions in the US5 ORF relative to rhesus macaque isolates. The two isolates from cynomolgus macaques again gave disparate results. Isolate 9400371 was identical to the rhesus macaque group of isolates with the exception of a triple repeat of five codons in the extracellular domain of the gJ glycoprotein, while the sequence of isolate E90-136 was again nearly identical to that previously reported for another cynomolgus macaque isolate (1). These two cynomolgus macaque BV sequences had 16 characteristic amino acid substitutions and five codon deletions relative to the rhesus macaque BV group sequence. There were also a few additional amino acid substitutions between members of each BV group (one between the two pigtail macaque isolate sequences, three among rhesus macaque isolate sequences, and two between the two similar cynomolgus macaque sequences).

An analysis of the pattern of nucleotide substitutions within the coding sequences of US4, US5, and US6 also provides evidence supporting the division of the BV isolates into three distinct groups, or genotypes. As shown in Table 1, tabulation of nucleotide substitutions occurring among rhesus and Japanese macaque isolates, between two cynomolgus macaque isolates (9400371 was excluded from this analysis), and between the two pigtail macaque isolates ranged from 2 to 6 substitutions (0.4 to 1.1%). In contrast, nucleotide differences between genotypes occurring at positions which were invariant within a genotype ranged from 50 to 97 substitutions (9.2 to 18.8%). Of these differences, greater than 65% were transversions, and over half of these were GC transversions. Both of these characteristics suggest that these differences have been positively selected for and fixed within the lineages rather than representing an accumulation of neutral substitutions.

A comparison of the DNA sequences of the two noncoding, intergenic regions separating the US4 and US5 ORFs and the US5 and US6 ORFs gave similar results. The same three groups or genotypes (rhesus and Japanese, pigtail, and cynomolgus macaque isolates) were apparent both by differences in sequence and particularly by insertions and deletions in the aligned sequences. While there was some variation between isolates of the same group or genotype, the differences were primarily single-nucleotide substitutions. This intragenotype variation was low, being about 2% among members of the rhesus and Japanese macaque isolate group (including cynomolgus macaque isolate 9400371) over both intergenic regions. This level of sequence variation in noncoding regions is similar to that occurring among strains of other alphaherpesviruses. Furthermore, canonical mRNA polyadenylation and termination signal sequences 3' of the US5 ORF were conserved in position and largely in sequence in all BV isolates (and HVP2). Similarly, AT-rich and GC-rich nucleotide tracks 5' of the US5 and US6 ORFs (positions 990 to 1005 and 1050 to 1065 in Fig. 4), possibly representing mRNA regulatory sequences, were also recognizably similar in all isolates.

Sequence data were used to assess the phylogenetic relationships among the various BV isolates. Shown in Fig. 5A is a phylogenetic tree derived from the complete 1,318-nucleotide U_S region sequence alignment shown in Fig. 4, with two HVP2 strains serving as an outgroup to root the BV tree. BV strains formed three separate lineages, which were composed of rhesus and Japanese macaque isolates, pigtail macaque isolates, and cynomolgus macaque isolates. Again, the sole exception was isolate 9400371, which, although identified as originating from a cynomolgus monkey, was grouped with isolates from rhesus and Japanese macaques. Since the alignment of sequence data from the noncoding intergenic regions was not as robust as that for coding sequences, phylogenetic analyses were also performed with coding sequences from the gB gene (438 nucleotides) and the US4, US5, and US6 ORFs (589 nucleotides total). For these two data sets, the same tree topology was again obtained. In each case, the three distinct lineages of BV isolates were evident, and isolate 9400371 was grouped with rhesus and Japanese macaque isolates. Bootstrap confidence levels for branches defining the three BV lineages were high, and the same tree topology was consistently derived using other distance and tree construction methods, further supporting the existence of the three BV genotypes.

View larger version (16K): [in this window] [in a new window]   FIG. 5.   Phylogenetic relationships of BV isolates. Three data sets were used to produce the trees shown. (A) The total aligned sequence shown in Fig. 4; (B) the combined coding sequences of the US4, US5, and US6 ORFs shown in Fig. 4; (C) 437 bp of coding sequence spanning the D2 region of the gB gene (UL27 ORF) (3). Tamura-Nei distances were used to construct trees by the neighbor-joining method. Numbers along branches represent bootstrap confidence levels generated with 500 replications.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

This study was undertaken to investigate the possibility that differences exist among BV isolates from different species of macaques which, if true, would provide a rational basis for the possibility of BV indigenous to different macaque species having different pathogenicities for humans. The results presented here suggest that there are at least three genotypes of BV, represented by isolates from rhesus and Japanese macaques (genotype 1), cynomolgus macaques (genotype 2), and pigtail macaques (genotype 3). There are, however, two points regarding this conclusion which need to be addressed.

First, the designation of a BV genotype carried by cynomolgus macaque isolates must be reconciled with the widely disparate results obtained for the two cynomolgus macaque isolates analyzed in this study (isolates 9400371 and E90-136). Isolate E90-136 was obtained from a cynomolgus macaque with a disseminated BV infection at the New England Regional Primate Research Center (24), while isolate 940037 was isolated from a monkey kidney cell culture submitted to VRL, Inc., by a biological supply company, which identified the origin of the cells as being from a cynomolgus macaque. While we cannot be certain that either of these isolates were not transmitted to cynomolgus monkeys from another macaque species, we believe that isolate E90-136 represents a true cynomolgus monkey virus. This is based on two observations. First, the 1.3 kbp of the E90-136 DNA sequence from the 3' end of the US4 ORF through the 5' third of the US6 ORF were nearly identical to published sequence data of other investigators for this region of DNA from a cynomolgus monkey BV isolate in Europe (1, 26). Second, Wall et al. (27) found that eight European BV isolates from three different colonies of cynomolgus macaques had characteristic RFLP profiles which differed from those of the two rhesus macaque BV isolates they examined. Their published BamHI restriction profiles for the cynomolgus macaque isolate DNAs were virtually the same as the one that we obtained for isolate E90-136 (Fig. 3), while the profiles of their two rhesus macaque BV isolates were indistinguishable from those of the rhesus macaque isolates and isolate 9400371 in this study. Given the similarity of isolate 9400371 to rhesus macaque isolates from the United States and Europe and the similarity of U.S. isolate E90-136 to other European cynomolgus monkey BV isolates, we conclude that isolate 9400371 is not a true cynomolgus monkey isolate but rather may have either been mistakenly identified as such or was possibly transmitted to a cynomolgus monkey from a rhesus monkey.

The observation that the two isolates obtained from Japanese macaques were nearly identical to isolates from rhesus monkeys also begs for explanation. Japanese macaques are the only species of the four examined whose natural geographic range does not overlap with that of any other macaque species. As such, one might expect that if any distinct BV genotype exists, it should be represented by isolates from this macaque species and that it would be the most divergent from any other BV genotypes. However, phylogenetic analyses of macaque mitochondrial DNA indicate that rhesus and Japanese macaques are in fact more closely related to each other than to either cynomolgus or pigtail macaques (12, 13). Given the close association between alphaherpesviruses and their hosts and their probable cospeciation, the fact that the tree topologies for macaques and BV isolates parallel each other is not surprising. Nonetheless, the Japanese macaque population from which the two isolates analyzed in this study were derived is a free-ranging troop located in southern Texas. These animals were imported to the United States several decades ago, and their history prior to arrival and release in Texas is not well documented. Thus, it is not known if these monkeys had ever been housed near rhesus macaques at any time where they could have contracted a rhesus strain of BV. Although it is known that wild Japanese macaques are seropositive for BV (21a) and that M. fuscata sera appear to differ from rhesus macaque sera in their reactivity with rhesus macaque BV antigen (21b), BV isolates from Japanese M. fuscata are not available. An analysis of BV isolates from Japanese macaques in Japan will be necessary to conclusively determine whether or not this species' BV truly has the same genotype as rhesus isolates of BV.

Of the three genotypes of BV identified in this study, the pigtail macaque group is more divergent from the rhesus and cynomolgus macaque genotypes than these two are from each other. This was apparent from an SDS-PAGE comparison of infected cell polypeptides, cELISA analysis of antigenicity, nucleotide sequence data, and phylogenetic analyses. However, ELISA results indicate that this variation is not sufficient to produce problems for BV serodiagnosis using other BV genotypes for the antigen. Similarly, Wall et al. (27) did not observe any difference in neutralization between rhesus macaque BV isolates and cynomolgus macaque BV isolates by using antiserum to a cynomolgus macaque BV isolate. Although one nucleotide substitution in the gB region used in a published PCR/RFLP test for detection and identification of BV results in the loss of a HaeIII restriction site used to discriminate BV isolates from other primate herpesviruses (3), the overall RFLP profiles of pigtail macaque isolates remain visibly typical of BV. In contrast, the use of more variable regions of the genome for location of PCR primers can result in failure to detect certain BV genotypes. This can be seen by a comparison of the sequences of the regions where the PCR primers and probe used by Slomka et al. (25) are located; the sequence variation between cynomolgus and rhesus macaque isolates is apparently sufficient to prevent these primers and/or the probe from identifying rhesus macaque isolates (a 1-bp deletion in one 21-mer primer and a 1-bp deletion plus five substitutions in the other 21-mer primer; residues 946 to 966 and 1121 to 1141 in Fig. 4). The even greater degree of DNA sequence variation in pigtail macaque isolates would make detection of these BV strains still more problematic. Thus, care must be taken in the design of primers for diagnostic use lest positive samples be identified as negative due to variation among BV genotypes in the sequences where primers are sited.

In summary, there appear to be at least three genotypes of BV which are to some degree related to the macaque species from which the virus is isolated. This conclusion is supported by all parameters examined in this study: antigenicity, overall genomic sequence similarity (RFLP), and nucleotide sequences of both coding sequences and noncoding intergenic regions. Thus, there is a molecular basis for the possible existence of BV strains having different degrees of pathogenicity in nonmacaque species. However, confirmation of variation in the lethality of different BV genotypes for species other than rhesus macaques will require testing and comparison of the pathogenic properties of each BV genotype.