Complete Sequence and Comparative Analysis of the Genome of Herpes B Virus (Cercopithecine Herpesvirus 1) from a Rhesus Monkey

Abstract

The complete DNA sequence of herpes B virus (Cercopithecine herpesvirus 1) strain E2490, isolated from a rhesus macaque, was determined. The total genome length is 156,789 bp, with 74.5% G+C composition and overall genome organization characteristic of alphaherpesviruses. The first and last residues of the genome were defined by sequencing the cloned genomic termini. There were six origins of DNA replication in the genome due to tandem duplication of both oriL and oriS regions. Seventy-four genes were identified, and sequence homology to proteins known in herpes simplex viruses (HSVs) was observed in all cases but one. The degree of amino acid identity between B virus and HSV proteins ranged from 26.6% (US5) to 87.7% (US15). Unexpectedly, B virus lacked a homolog of the HSV γ₁34.5 gene, which encodes a neurovirulence factor. Absence of this gene was verified in two low-passage clinical isolates derived from a rhesus macaque and a zoonotically infected human. This finding suggests that B virus most likely utilizes mechanisms distinct from those of HSV to sustain efficient replication in neuronal cells. Despite the considerable differences in G+C content of the macaque and B virus genes (51% and 74.2%, respectively), codons used by B virus are optimal for the tRNA population of macaque cells. Complete sequence of the B virus genome will certainly facilitate identification of the genetic basis and possible molecular mechanisms of enhanced B virus neurovirulence in humans, which results in an 80% mortality rate following zoonotic infection.

Comparative genome analyses of closely related viruses offer insight into the protein coding capacities of viral genomes (18, 69), a means to biologically classify viruses (2, 19) and identify viral genes involved in virulence and pathogenicity (1, 37, 67). The number of completely sequenced viral genomes has been increasing rapidly and now exceeds 1,000, including 29 herpesvirus genomes (GenBank data, http://www.ncbi.nlm.nih.gov/PMGifs/Genomes/viruses.html).

B virus (Cercopithecine herpesvirus 1, monkey B virus) is a member of the subfamily Alphaherpesvirinae, which together with human herpes simplex virus types 1 and 2 (HSV-1 and HSV-2) constitutes the genus Simplexvirus. B virus generally causes only mild localized or asymptomatic infections in its natural hosts, Asian monkeys of the genus Macaca (33, 34, 74). In contrast, B virus infections in foreign hosts, humans or monkey species other than macaques, often result in encephalitis, encephalomyelitis, and death (53, 73, 74).

The genome organization of B virus is similar to that of HSV-1 and HSV-2: the unique long (U_L) and unique short (U_S) segments flanked by inverted long (R_L) and short (R_S) repeat sequences are covalently joined in four possible isomeric configurations (26). Sequence analysis of the partial U_S regions of B virus and simian agent 8 virus demonstrated that human and primate viruses are colinear in this genomic segment (51). However, only a limited number of B virus gene sequences from the U_L region have been published (3, 36, 62, 63), and nothing has been reported about the structure and the gene content of the repeat genomic elements.

The size of the B virus genome was estimated previously as 165 kb (26) and 162.5 kb (40), which is significantly larger than the HSV-1 and HSV-2 genomes (152 and 155 kb, respectively). Extra DNA might contain additional genes that are not present in human viruses and may provide insight about B virus pathogenicity in foreign hosts.

In this study, the complete genomic sequence of the B virus reference strain E2490, isolated from a rhesus macaque (Macaca mulatta), was determined, and viral DNA termini were located. The genomic structure and gene content were analyzed and compared to those of human herpesviruses HSV-1 and HSV-2.

MATERIALS AND METHODS

Viruses, cells, and media.

B virus laboratory strain E2490 and two B virus clinical isolates were obtained from the National B Virus Resource Laboratory (Atlanta, Ga.). Viruses were propagated in Vero cells maintained in minimal essential medium (MEM) supplemented with 2% fetal bovine serum (Gibco). Viral DNA was purified from the infected-cell lysate by ultracentrifugation in an NaI density gradient as described previously (70). Escherichia coli Top10 (Invitrogen) cells were used for propagation of plasmids.

DNA cloning and sequencing.

The BamHI, KpnI, SalI, PstI, and XhoI restriction fragments of B virus genomic DNA were cloned into the pUC19 vector digested with the same enzyme. The following three techniques were employed to generate full sequences of the clones: primer walking, unidirectional deletions, and random transposon insertions (Epicentre Technologies).

To obtain the sequence of the IR_L-IR_S regions, two shotgun libraries of random overlapping clones were prepared from the adjacent PstI fragments P13 and P1, covering this region (Fig. 1). DNA fragments of 13 kb (P13) and 4.8 kb (P1) were isolated after separation of PstI-digested genomic DNA on a 0.8% agarose gel. Each DNA fragment was then treated with DNase I from a DNaseI Shotgun kit (Novagen) according to the manufacturer's suggested protocol. The DNase I-treated fragments were fractionated on a 1.4% agarose gel, and then 300- to 1,000-bp fragments were eluted from the gel, blunt ended with T4 DNA polymerase, dA-tailed with Tth polymerase and a dA tailing kit (Novagen), and ligated into the pScreen T vector (Novagen). The ligation mixtures were transformed into E. coli NovaBlue cells (Novagen).

FIG. 1.

Molecular cloning of B virus genome. The first letter in the clone designations indicates the cloning site: B, BamHI; K, KpnI; P, PstI, S, SalI; X, XhoI. The number in the clone designations reflects the order of clone isolation. Double-sided arrows indicate PCR fragments used for verification of clone junctions. The map scale is in kilobases.

DNA sequencing was performed with BigDye Terminator (PE Applied Biosystems) sequencing chemistry on an automatic DNA sequencer ABI 377 (Applied Biosystems). Sequencing reactions were performed in the presence of Sequence Enhancer A (Gibco) to resolve sequence compressions due to the high G+C content of templates. A 3:1 mixture of regular and dGTP BigDye sequencing kits (Applied Biosystems) was used in order to overcome premature termination of sequencing reactions in G-rich regions. DNAStar SeqMan program (DNAStar Inc.) was used to assemble the complete genome sequence from the sequences of the overlapping plasmid clones. For adjacent fragments with no overlap, areas across the junction were amplified with B virus genomic DNA as the template, and the resulting PCR fragments were sequenced.

Cloning of genomic DNA termini.

B virus genomic DNA (≈0.5 μg) was blunt ended by treatment with 5 U of T4 DNA polymerase (Novagen) as suggested by the manufacturer. After purification on a Qiaquick spin column (Qiagen), the blunt-ended DNA was digested overnight with restriction enzyme SphI and ligated into SmaI- and SphI-digested vector pUC19. E. coli NovaBlue (Novagen) competent cells were transformed with the ligation mixture.

Southern blots.

SphI-digested genomic DNA was transferred from a 0.8% agarose gel onto a nylon membrane (Roche) by alkali blotting according to the instructions supplied with the membranes. Plasmid DNA was labeled with digoxigenin with a HighPrime labeling kit (Roche). Prehybridization, hybridization, and washing steps were carried out according to standard protocols (59). Chemiluminescence detection of the bound probe was performed with an ECL detection kit (Amersham Pharmacia Biotech).

DNA sequence analysis.

Identification of open reading frames (ORFs), repeats, and DNA regulatory sequences was performed with DNAStar suite of programs. GenBank database searches were carried out with BlastN, BlastP, and BlastX with default settings. Multiple alignments between B virus, HSV-1, and HSV-2 genes and proteins were performed with and analyzed by DNAStar MegAlign program (version 4.0.3) with the PAM250 amino acid substitution matrix and Joltun Hein method (30) and the following parameters: gap opening penalty = 11 and gap extension penalty = 3. Identity values were determined from the generated alignments.

The GeneMark and GeneMark.hmm gene-finding programs were used to refine the procedure of gene identification (4, 41). The parameters of the statistical models were defined by training on the set of B virus ORFs encoding protein products homologous to known HSV-1 and HSV-2 proteins. These statistical methods are able to identify frameshifts and unusual gene starts in addition to detecting genes not found by similarity search. To prove that a homolog of the HSV neurovirulence factor is absent in the B virus genome, BlastP and PSI-Blast searches with ICP34.5 protein as the query were used against the whole B virus genome translated in six frames. In addition, a Hidden Markov Model profile was created from the conservative domain of the ICP34.5 protein and used to scan the translated B virus genome.

In a search for unique genes, the PSI-Blast and RPS-Blast programs were used to search for possible conserved domains in proteins predicted by statistical methods. Predicted proteins shorter than 20 residues were excluded from the analysis. Analyses of protein structures were performed with the DNAStar Protean program (version 4.0.3), signal peptide prediction program SignalP version 2.0.b2 (http://www.cbs.dtu.dk/services/SignalP-2.0/), and transmembrane prediction programs TMHMM version 2.0 (http://www.cbs.dtu.dk/services/TMHMM/), DAS (http://www.sbc.su.se/≈miklos/DAS/), and TMpred (http://www.ch.embnet.org/software/TMPRED_form.html).

GenBank accession numbers.

The complete sequence of the B virus genome (GenBank accession number AF533768) and the L-terminal sequences of two clinical B virus isolates (GenBank accession numbers AY230747 and AY230748) have been deposited in the NCBI database.

RESULTS AND DISCUSSION

Cloning and sequencing of B virus genome.

The relative positions of all fragments used for B virus genome sequencing are shown in Fig. 1. The U_L and U_S genomic sequences were derived from the nucleotide sequences of the plasmid-cloned BamHI, KpnI, SalI, PstI, and XhoI overlapping genomic fragments. The sequences of the internal copies of the large and small repeat regions (IR_L and IR_S, respectively) were determined from the nucleotide sequences of short overlapping clones derived from shotgun libraries prepared from the adjacent P13 and P1 genomic fragments. Terminal copies of repeat elements (TR_L and TR_S) were assumed to be identical to the internal copies, as in the HSV-1 and HSV-2 genomes (18, 43, 44). An assembly of 1,389 single readings resulted in generation of the U_L-IR_L-IR_S-U_S genomic sequence comprising 139,269 bp, with an average redundancy of 3.9. Exact U_L-TR_L, U_L-IR_L, U_S-IR_S, and U_S-TR_S boundaries were determined by sequence comparisons of PCR fragments amplified across these junctions.

Location of B virus genomic termini.

To determine the first and last nucleotides in the B virus genome, genomic termini were defined. Initially, the locations of the internal a sequence and the joint between the B virus long and short internal repeat regions (L/S joint) were estimated after aligning the HSV-1 and B virus IR_L-IR_S sequences (not shown). Two SphI restriction sites were located approximately 1.8 kb (in the IR_L region) and 1.5 kb (in the IR_S region) from the predicted internal a sequence. Since the R_L and R_S internal and terminal copies in alphaherpesvirus genomes are identical, SphI restriction sites were expected to be located ≈1.8 kb and ≈1.5 kb from the B virus L and S genomic termini, respectively, and therefore could be used for cloning the B virus genomic ends.

Blunt-ended B virus genomic DNA was then digested with SphI and ligated into SmaI- and SphI-digested vector pUC19. All internal genomic fragments had SphI sticky ends on both sides, and consequently only terminal genomic fragments, which had one blunt end and another SphI end, were compatible with the prepared vector and successfully cloned by this procedure. As predicted, two sets of recombinant clones were generated with inserts of the estimated sizes. To confirm that the clones isolated included viral DNA ends, plasmid DNA from clones containing 1.5-kb and 1.8-kb inserts was used to probe SphI-digested B virus genomic DNA on Southern blots. Theoretically, if these clones indeed contain genomic ends, they will hybridize to the two SphI genomic fragments of 1.5 kb or 1.8 kb (terminal) and 3.1 kb (internal from the L/S junction). These fragments and two additional 2.1-kb and 3.3-kb fragments of lower intensity were detected (Fig. 2A). The observed cross-hybridization patterns indicated that the two genomic ends had common sequences. The additional fragments are most likely L-terminal (2.1 kb) and junction (3.3 kb) fragments containing an extra copy of a sequence. The variable number of a sequence repeats has also been described at the L terminus and L/S junction in HSV-1 (57).

FIG. 2.

Identification of B virus genomic termini. (A) Southern blot of SphI-digested genomic DNA hybridized with digoxigenin-labeled plasmids containing 1.6-kb (lane 1) and 1.8-kb (lane 2) terminal fragments. An autoradiograph of the membranes is shown. The positions of DNA size markers are shown on the left. Arrows indicate terminal small (T_S), terminal large (T_L), and junction (J) fragments. (B) Cloned genomic fragments containing B virus termini. The structural organization of the B virus genome is shown, with the U_L and U_S regions represented by solid lines and the TR_L, IR_L, TR_S, and IR_S regions represented by open boxes. Terminal a sequences and the oppositely oriented internal a′ sequence are indicated. Below the genome diagram, a schematic alignment of the isolated terminal fragments is shown. Arrows denote the locations of a sequence copies and their orientations in the genome. The numbers in parentheses indicate the number of clones sequenced. (C) Alignments of the a sequences of B virus (BV), HSV-1, and HSV-2. Arrows indicate genomic ends. The conserved motifs of pac1 and pac2 signals are shown in bold.

Nucleotide sequences adjacent to the vector cloning site SmaI (genomic termini) were derived from the inserts of 10 selected clones: six clones originated from the large terminal repeat (L-terminal clones) and four originated from the small terminal repeat (S-terminal clones) (Fig. 2B). All contained identical sequences (223-bp long) present in direct orientation on both genomic ends and in reverse orientation at the L/S joint in either one (nine clones) or two (one clone) directly repeated copies. The B virus 223-bp terminal sequence is a homolog of the HSV-1 a sequence that is also present in other herpesvirus genomes. Two cis-acting elements, pac1 and pac2, that are required for cleavage and packaging of herpesvirus genomes exist in the a sequence (17, 68).

Although overall sequence homology between the a sequences of B virus, HSV-1, and HSV-2 was low, the sequence composition of the cleavage and packaging signals was very well conserved (Fig. 2C). An A-rich motif followed by three CGCGGCG motifs formed the B virus pac2 signal. Interestingly, most herpesviruses have only one copy of the CGCGGCG motif, which contributes significantly to the efficiency of genome cleavage and packaging (45). The B virus pac1 contained three conserved motifs, an A-rich region flanked by two stretches of seven C residues and 13 C+G residues. Both pac1 and pac2 are located at conserved distances from the genomic ends (17), 31 bp from the L terminus and 34 bp from the S terminus of B virus DNA. There were no DR2 repeats between the pac1 and pac2 signals in the B virus a sequence, which were thus noticeably shorter than the HSV-1 a sequence.

Overall genome organization.

The complete B virus genomic sequence was assembled according to the HSV-1 prototype genome structure (57). The genome of B virus is 156,789 bp in length, considerably shorter than earlier predictions based on the constructed B virus physical map (26) or direct measurements by electron microscopy (40). The 74.5% G+C base composition of the genome is very similar to earlier estimates of 75% determined by DNA buoyant density centrifugation (40). Table 1 presents a comparison of the length and G+C content of B virus, HSV-1, and HSV-2 by genomic region. The G+C composition was elevated in all B virus regions compared to that in HSV types 1 and 2, with unique regions demonstrating the highest G+C content among known herpesviruses. The sizes of genomic segments were comparable among the viruses with the exception of R_S, which in the B virus genome was ≈1.5 kb longer.

TABLE 1.

Comparison of B virus, HSV-1, and HSV-2 genomic regions

Region	Virus	Length (bp)	% G+C
UL	B virus	107,815	72.9
	HSV-1	107,947	66.9
	HSV-2	109,689	68.9
US	B virus	14,687	73.2
	HSV-1	12,980	64.3
	HSV-2	14,329	66.2
TRL	B virus	9,021	79.4
	HSV-1	9,912	71.6
	HSV-2	9,297	74.4
TRS	B virus	8,234	80.4
	HSV-1	6,677	79.5
	HSV-2	6,711	80.1
Whole genome	B virus	156,789	74.5
	HSV-1	152,261	68.3
	HSV-2	154,746	70.4

Ten sets of tandem reiterations were identified in the B virus genome. Genomic locations of sets, numbers of repeat units in each set, sizes, and sequences are given in Table 2. Most reiterations were present at locations compatible with the locations of HSV-1 reiterations. Six sets were present in two copies in the genome due to location in the terminal and internal copies of the R_S and R_L segments. Only two reiteration sets were located in protein-coding regions, the UL36 and US4 (gG) genes.

TABLE 2.

Sets of reiterated sequences in the B virus genome

Set	Genomic region	Location of set		Sequence of repeat unit	Unit size (bp)	No. of units	Comments
		First copy	Second copya
1	RL	824-1022	124836-125034	GGGGGTCCTGGGGGTCCGGGGTCGCC	26	8
2	RL	2244-2315	123543-123614	GGGGGTCTC	9	8	Located in ICP0 intron
3	RL	3658-3729	122129-122200	GCCCGGCGCCCAAGTCCC	18	4
4	RL	5164-5244	120614-120694	CCAGAAGCAGAGAGGGGCGGGGGCTCC	27	3
5	UL	43039-43123		GGGGGTGCGGGGGCGGT	17	5	Located between UL21 and UL22
6	UL	71559-71756		GCAGGGGCA	9	22	Encodes PAA repeat in UL36
7	UL	115984-116238		CCCCCCTCCCCTCCCCCGCG	20	13
8	RS	131851-132461	149963-150573	CCCTTCCCCCCTT	13	47	Flanks oriS
9	RS	133448-133785	148639-148976	CCCCCGCGCACCCCTCGCCCTCCCCT	26	13	Flanks oriS
10	US	139337-139444		CCCGCCCCCACCACCACC	18	6	Encodes PAPTTT repeat in gG

^aA location of the second copy of a set is given for reiteration sets from the R_S and R_L genomic regions.

Both origins of B virus replication, oriL and oriS, were tandemly duplicated and present in locations corresponding to HSV oriL and oriS locations. Thus, six origins of DNA replication exist: two copies of oriL (oriL1 and oriL2) in the U_L region and two copies each of oriS (oriS1 and oriS2) in the terminal and internal R_S regions. Duplications of either the oriL or oriS sequence were also found in HSV-1 strain ANG and HSV-2 strain HG52, respectively (18, 24), but concurrent duplication of both origins has not been described previously for herpes simplex viruses. All B virus origins share the core element of a 94-bp perfect palindrome (Fig. 3) containing two predicted binding sites for the origin-binding protein (OBP), box I, and box III (22, 27, 52). Furthermore, the nucleotide sequences of the B virus oriS and oriL core elements were extremely conserved and almost identical to HSV oriL but differed from HSV oriS (25, 42). Whether the existence of six nearly identical origins of replication in the B virus genome has any functional significance or is just an artifact due to repeated passage of the virus in cell culture currently remains unknown.

FIG. 3.

Comparison of B virus and HSV-1 origins of replication. The DNA sequences of the oriL and oriS core elements are shown. BV/L1, B virus oriL1; BV/L2, B virus oriL2; BV/S, B virus oriS1 and oriS2; HSV1/L, HSV-1 oriL; HSV1/S, HSV-1 oriS. The OBP-binding sites (box I, box II, and box III) are boxed. Residue substitutions in all B virus origins relative to HSV-1 oriL are shown in bold. Residue substitutions in B virus oriL2 relative to oriL1 are underlined.

Gene characterization.

Both major approaches to gene identification, extrinsic and intrinsic, were used to identify and characterize B virus genes. Extrinsic methods, such as Blast, identify protein-coding genes by detecting similarity of translated protein sequences to the primary structure of a known protein. The intrinsic approach, an ab initio statistical method such as GeneMark, identifies protein-coding regions by detecting specific frequency patterns in nucleotide order, including the codon usage pattern. These two types of methods have complementary strengths in terms of sensitivity and specificity. Extrinsic, similarity search methods have high specificity but may miss some genes, up to 30 to 40%, which could be recovered by the intrinsic, statistical method. The B virus genomic sequence can be efficiently analyzed by similarity search methods due to extensive knowledge of the closely related viral species HSV-1 and HSV-2. Still, unique virus-specific genes could be missed. Therefore, in addition to the similarity search, we used ab initio methods, GeneMark and GeneMark.hmm, trained on the set of genes confirmed by similarity search. These methods were modified to allow a noncanonical translation initiation codon, GTG (46), that could be used in this highly G+C-rich genome.

(i) HSV gene homologs.

All but one B virus gene were identified on the basis of sequence homology to HSV-1 and HSV-2 genes and named correspondingly. Table 3 summarizes their locations in the genome, predicted sizes, and percent identities with the corresponding HSV-1 and HSV-2 gene products. Seventy-two genes existed as a single copy within unique genomic regions, whereas two genes, ICP0 and ICP4, appeared twice due to duplication of the large and small repeat regions where they reside. Two ORFs (RL2 and UL15) are predicted to have introns matching ones in the corresponding HSV genes.

TABLE 3.

ORFs and other features of the B virus genome

ORF or feature	Location		Strand	Length (codons)	Identity (%)		Characteristics and predicted function(s)a (reference)
	Start	Stop			HSV-1	HSV-2
a sequence	1	223					Terminal direct repeat
TRL	1	9021					Terminal copy of large repeat region
RL2			+	701	43.5	40.1	Immediate-early protein ICP0; multifunctional regulatory protein
Exon 1	2194	2241	+	16
Exon 2	2461	3003	+	181
Exon 3	3125	4636	+	504
UL	9022	116837					Unique large region
UL1	9072	9746	+	225	65.5	53.4	Virion membrane glycoprotein L; in complex with gH; membrane fusion
UL2	9814	10575	+	254	71.3	68.9	Uracil-DNA glycosylase; DNA repair
UL3	10711	11394	+	228	61.8	61.8	Colocolization with ICP22 and UL4 in small, dense nuclear bodies
UL4	12120	11509	−	204	56.1	56.5	Colocolization with ICP22 and UL3 in small, dense nuclear bodies
UL5	14820	12172	−	883	82.7	83.0	Component of helicase-primase complex
UL6	14819	16870	+	684	69.5	71.6	Capsid protein; DNA cleavage/packaging
UL7	16821	17711	+	297	65.2	65.2	Capsid protein; DNA cleavage/packaging
UL8	20171	17886	−	762	58.2	60.4	Component of helicase-primase complex
UL9	22847	20223	−	875	78.3	78.8	ori binding protein; helicase activity
UL10	22714	24132	+	473	65.2	65.1	Virion membrane glycoprotein M; proposed role in capsid envelopment
UL11	24733	24437	−	99	50.0	53.1	Myristylated tegument protein; capsid envelopment
UL12	26553	24679	−	625	67.8	69.1	DNase; endoexonuclease; processing of DNA replication intermediates
UL13	28097	26553	−	515	66.9	68.8	Virion protein kinase
UL14	28513	27869	−	215	65.6	66.5	Minor tegument protein
UL15			+	739	87.5	87.5	DNA cleavage/packaging; transiently associated with maturing capsids
Exon 1	28599	29624	+	342
Exon 2	33199	34389	+	397
UL16	30898	29795	−	368	64.2	66.7	Capsid associated; DNA cleavage/packaging; located in UL15 intron
UL17	33015	30919	−	699	66.8	69.5	Tegument protein; DNA cleavage/packaging
UL18	35574	34618	−	319	82.8	82.4	Capsid protein VP23; forms triplexes with VP19C that connect pentons and hexons in capsids
UL19	39918	35785	−	1,378	86.9	85.5	Major capsid protein VP5; forms pentons and hexons of capsid shell
UL20	40783	40112	−	224	64.6	66.4	Virion membrane protein; virion egress; syn5 locus
UL21	41384	42964	+	527	66.2	68.3	Nucleotidylylated phosphoprotein; interacts with microtubules and facilitates intracellular transport of the virus (65)
UL22	45736	43193	−	848	59.1	60.9	Virion membrane glycoprotein H; in complex with gL; membrane fusion, entry, cell-to-cell spread
UL23	47107	45995	−	371	59.8	60.7	Thymidine kinase
UL24	47042	47842	+	267	66.4	66.4	Nonglycosylated membrane-associated protein; syn5 locus
UL25	48036	49775	+	580	79.8	79.8	Minor capsid protein; DNA packaging; possible role in DNA anchoring (50)
UL26	49937	51763	+	609	65.7	65.4	Capsid maturation protease
UL26.5	50834	51763	+	310	58.8	57.0	Scaffolding protein
UL27	54815	52137	−	893	79.9	80.4	Virion membrane glycoprotein B; cell entry; contains syn3 locus
UL28	57192	54835	−	786	84.0	84.8	DNA cleavage/packaging; transiently associated with maturing capsids
UL29	61257	57667	−	1,197	82.7	82.2	Single-strand DNA-binding protein; key role in assembly of DNA replication proteins
oriL1	61592	61789					Center of replication origin oriL1
oriL2	61795	61992					Center of replication origin oriL2
UL30	62173	65916	+	1,248	79.6	80.0	DNA polymerase catalytic subunit; complexes with UL42
UL31	66766	65861	−	302	80.9	81.3	Nuclear phosphoprotein; interacts with UL34; capsid egress from nucleus
UL32	68531	66759	−	591	74.2	74.5	DNA packaging; not associated with capsids
UL33	68530	68928	+	133	72.1	72.9	DNA packaging; not associated with capsids
UL34	68988	69803	+	272	66.8	71.3	Type II nuclear membrane-associated phosphoprotein; interacts with UL31; capsid egress from nucleus
UL35	69939	70283	+	115	51.8	47.3	Basic phosphorylated capsid protein VP26
UL36	80356	70490	−	3,289	61.2	62.2	Very large tegument protein; interacts with UL19 and UL37
UL37	84145	80624	−	1,174	69.4	68.7	Minor tegument protein
UL38	84598	85974	+	459	70.6	69.1	Capsid protein VP19C, forms triplexes with VP23 that connect pentons and hexons in capsids
UL39	86392	89385	+	998	66.7	66.0	Large subunit of ribonucleotide reductase
UL40	89434	90450	+	339	79.7	80.4	Small subunit of ribonucleotide reductase
UL41	92073	90619	−	485	72.3	73.0	Tegument phosphoprotein; virion-associated host shutoff (vhs) protein
UL42	92575	94002	+	476	48.6	48.8	Double-stranded DNA-binding protein, DNA polymerase subunit
UL43	94131	95270	+	380	44.4	49.1	Predicted membrane-associated protein
UL44	95527	96930	+	468	49.9	51.5	Virion membrane glycoprotein C; cell attachment; blocking host immune response
UL45	97166	97690	+	175	63.6	59.0	Type II membrane protein; possible role in cell fusion
UL46	100140	97978	−	721	57.7	58.9	Tegument phosphoprotein VP11/12; modulates alpha trans-inducing factor activity
UL47	102327	100267	−	687	59.9	58.5	Tegument phosphoprotein VP13/14; O-glycosylated; modulate α-TIF activity; RNA binding (60)
UL48	104246	102783	−	488	69.3	69.2	Major tegument protein VP16 (α-TIF); trans-activator of α genes
UL49	105449	104580	−	290	45.9	45.0	Major tegument protein VP22; binds RNA; carrier of mRNA from infected to uninfected cells (60)
UL49A	106092	105853	−	80	40.5	43.0	Envelope protein
UL50	106107	107216	+	370	55.9	54.3	Deoxyuridine triphosphatase
UL51	108064	107381	−	228	67.5	69.3	Capsid/tegument-associated phosphoprotein (15)
UL52	108126	111302	+	1,059	73.0	73.1	Component of helicase-primase complex
UL53	111254	112267	+	338	66.0	68.6	Membrane glycoprotein K; virion egress; contains syn1 locus
UL53A			−	300			Hypothetical protein predicted by GeneMark and GeneMark.hmm
Exon 2	112186	112495	−	197
Exon 1	112755	113344	−	103
UL54	112644	114116	+	491	59.1	60.8	Immediate-early protein ICP27; regulates some early and all late gene expression
UL55	114429	115001	+	191	62.0	64.3	Nuclear matrix-associated protein
UL56	115834	115154	−	227	39.7	42.9	Type II membrane protein (38); involved in virus pathogenicity (35)
IRL	116837	125634					Internal copy of large repeat region
RL2			−	701	43.5	40.1	Immediate-early protein ICP0; multifunctional regulatory protein
Exon 3	122733	121222	−	504
Exon 2	123397	122855	−	181
Exon 1	123664	123617	−	16
a′ sequence	125635	125857					Inverted copy of a sequence
IRS	125858	133868					Internal copy of small repeat region
RS1	131284	127724	−	1,187	65.0	66.8	Immediate-early protein ICP4; regulator of gene expression
Oris1	132795	132796					Center of replication origin oriS1
Oris2	132998	132999					Center of replication origin oriS2
Us	133869	148554					Unique small region
US1	133900	135252	+	451	41.1	43.5	Immediate-early protein ICP22; required for optimal ICP0 expression
US2	136386	135478	−	303	56.1	54.0	Tegument protein
US3	136708	138084	+	459	60.8	61.1	Protein kinase; antiapoptotic activity
US4	138221	140242	+	674	29.2	39.2	Virion membrane glycoprotein G; entry into polarized cells
US5	140465	140830	+	122	34.0	26.6	Glycoprotein J; block apoptosis
US6	141296	142480	+	395	57.0	59.0	Virion membrane glycoprotein D; cell entry; interacts with cellular receptors
US7	142680	143885	+	402	46.1	51.0	Virion membrane glycoprotein I; in complex with gE; basolateral viral spread
US8	144253	145872	+	540	46.0	48.0	Virion membrane glycoprotein E; in complex with gI; basolateral viral spread
US8.5	145817	146185	+	123	42.3	45.9	Nucleolar phosphoprotein
US9	146309	146581	+	91	58.9	57.3	Tegument protein
US10	148139	147204	−	312	43.7	45.9	Tegument protein
US11	148290	147847	−	148	45.2	46.7	RNA-binding tegument protein; interacts with protein kinase R
US12	148555	148310	−	82	26.8	26.8	Immediate-early protein ICP47; inhibits antigen presentation
TRs	148556	156566					Terminal copy of small repeat region
oriS2	149425	149426					Center of replication origin oriS2
oriS1	149628	149629					Center of replication origin oriS1
RS1	151140	154700	+	1187	65.0	66.8	Immediate-early protein ICP4; regulator of gene expression
a sequence	156567	156789					Terminal direct repeat

^aCharacteristics and functions of B virus proteins were deduced from the properties of known HSV-1 and HSV-2 homologs (31, 47, 57).

It was inferred that the B virus UL1 and UL2 genes use the GTG codon as a start codon at genomic sequence positions 9072 and 9814, respectively. This prediction was made statistically and was corroborated by the presence of conserved regions upstream of the canonical ATG. Both predicted translational initiation GTG codons aligned with the ATGs of the HSV UL1 and UL2 genes. Conservation of the UL1 GTG start codon was confirmed by sequence analysis of the 5′ region of the UL1 gene from two B virus clinical isolates (data not shown).

The extent of amino acid identity between B virus and HSV polypeptides varied from 26.6% (US5) to 87.7% (UL15). The conservation of specific protein domains observed in HSV-1 and HSV-2 was mirrored when HSV-1 and B virus were compared. The following proteins were significantly conserved in B virus: DNA cleavage and packaging proteins, i.e., UL15, UL28, UL32, and UL33; capsid proteins, i.e., UL18, UL19, and UL38; proteins involved in DNA replication, i.e., UL2, UL5, UL29, and UL30; and glycoprotein B. The three least-conserved proteins in B virus were US4, US5, and US12. Similar levels of conservation have been described for homologous proteins in many other mammalian herpesviruses (58).

The B virus proteins could be divided into three groups based on the degree of similarity to HSV-1 and HSV-2 proteins. The largest group (46 proteins) showed greater similarity to HSV-2 proteins, e.g., DNA cleavage and packaging proteins. The second group (20 proteins), with capsid proteins among others, showed higher levels of similarity to HSV-1 proteins. In seven proteins, differences in similarity to HSV-1 or HSV-2 homologs were marginal, and these proteins formed the third group. The proteins that enable HSV-1 to replicate and reactivate more efficiently at orofacial sites and HSV-2 at genital sites are unknown, but the fact that B virus replicates and is reactivated with similar efficiencies at both sites may be explained by the similarities of selected B virus proteins with either HSV type 1 or 2 proteins. These hybrid properties in B virus raise many questions about alphaherpesvirus evolution.

(ii) Gene homologs absent.

In recent years, the existence of additional HSV-1-specific genes has been proposed and substantiated by experimental data. UL20.5 is located between UL20 and UL21, ORF P and ORF O map antisense to the 34.5 gene, and UL43.5 and UL27.5 map antisense to UL43 and UL27, respectively (8, 55, 56, 71, 72). Like HSV-2 (18), B virus has no equivalents to these proposed genes.

One well-established gene, the HSV γ₁34.5 (RL1) homolog, was observed to be absent in B virus. This conclusion was reached after repeated searches with state-of-the-art computational genome analysis tools (see Materials and Methods). To determine that this observation was not limited to the laboratory strain of B virus, we cloned and sequenced L-terminal SphI restriction fragments from two low-passage clinical isolates, one of which was derived from a rhesus macaque (54) and the other post mortem from a zoonotically infected human patient (16). Sequence comparison of these fragments with the corresponding fragment of the laboratory strain did not reveal any significant differences: only single nucleotide substitutions and variations in the number of copies of short reiterations were found in the clinical isolates relative to the laboratory strain. Since the absence of the γ₁34.5 gene homolog was verified in three independent B virus strains, we concluded that it was a genuine feature of the B virus genome.

The protein product of the HSV-1 γ₁34.5 gene, infected-cell protein 34.5 (ICP34.5), is a neurovirulence factor with at least two known, distinct functional activities (9, 10, 12-14, 64, 66). One function, encoded by the carboxyl-terminal domain, negates the antiviral effect of induced protein kinase R by redirecting the host protein phosphatase 1α to dephosphorylate translation initiation factor eIF2α, preventing protein synthesis shutoff in infected cells (9, 11, 12, 29, 39). Another function, mapped to both amino-terminal and carboxyl-terminal domains, somehow enables the virus to replicate in the peripheral and central nervous systems of experimentally infected animals (10, 64). Deletion of γ₁34.5 leads to complete neuroattenuation of highly neurovirulent HSV-1 strains. In the absence of γ₁34.5, other genes may supply these functions, given the striking similarities between the replication characteristics of B virus and the human simplex viruses.

We predict that B virus currently uses compensatory strategies to block host responses to infection, similar to those described for HSV-1. For example, the HSV-1 US11 protein can inhibit protein kinase R activation and compensate for the absence of the ICP34.5 function in deletion mutants if expressed early in infection (5, 6, 28, 48). It was proposed that the US11 gene encodes an alternative mechanism to preclude the shutoff of protein synthesis that is currently inactive in HSV-1 (5). In addition, B virus might have evolved unique mechanisms to prevent termination of protein synthesis and elude the DNA replication blocks imposed by neuronal cells. To examine this possibility, a search for B virus-specific genes was performed.

(iii) Unique genes in B virus genome.

Two major sequence differences between the B virus and HSV-1 and HSV-2 genomes were detected. The B virus R_S region contains an additional ≈1.5 kb of sequence between the S terminus and the ICP4 gene homolog, while the R_L region is shorter in B virus than in the human viruses, with no sequence homology to the HSV ICP0 flanking regions. However, no potential genes were identified in these regions despite rigorous analyses.

A putative two-exon gene (UL53A) was identified by statistical analysis in the B virus U_L region on a complementary strand (positions 112186 to 112495 and 112755 to 113344). This putative gene had a codon usage pattern compatible with codons present in established B virus genes. The UL53A protein is a hydrophilic basic protein (pI 11.9) with an amino acid composition biased to Pro and Arg residues. The first 21 residues were predicted to be a signal peptide by the SignalP program, and residues 222 to 240 were predicted to be a transmembrane domain by two out of three programs applied (TMpred and DAS), but probability scores were not high enough to classify this protein as a membrane protein. PSI-Blast analysis detected sequence similarity to domains in three neuronal proteins encoded in mammalian genomes: neural cell adhesion molecule NCAM-180 (GenBank accession number P13595, E value of 0.16), calcineurin inhibitor cabin 1 (GenBank accession number AAD40846, E value of 0.014), and brain calcium channel α 1A subunit (GenBank accession number AAB64179, E value of 0.007). However, the similarity was insufficient to confidently characterize the gene at this time, and pending experimental evidence will be critical to determine whether this gene, which is missing in HSV-1 and HSV-2, provides B virus with a unique mechanism to attack neural cells.

B virus genome high G+C content and optimization of codon usage.

The G+C content of the protein-coding regions in B virus, 74.4%, is only slightly lower than the G+C content of the noncoding regions, 75.4%. These data indicate a strong mutation pressure toward G and C nucleotides in B virus DNA evolution. Obviously, the genetic drift toward the abundance of G's and C's has been compensated for by positive selection for conservative A's and T's in the first (26% A and T) and second (44% A and T) positions of a codon in protein-coding regions and by positive selection for A's and T's in a few evolutionarily conserved regions (promoters, repeats, and other regulatory sites) in noncoding DNA. The G+C content of the third position of codons in B virus genes is extremely high, 93.1%. This bias, created by mutation pressure, seems to be a driving force in the formation of the codon usage pattern in B virus.

Genomes of primates tend to have low G+C content, and the logical question arises whether the highly G+C-biased codon usage in B virus matches the proportions of the isoacceptor tRNA in host cells. Unfortunately, no experimental data are currently available regarding the tRNA pool in macaque cells. Still, relative amounts of the isoacceptor tRNAs could be predicted based on the frequencies of codon use in rhesus macaque genes because a strong direct correlation between these parameters has been demonstrated in a number of organisms (23, 32, 49).

A codon usage catalog of rhesus macaque genes was generated based on 288 protein-coding sequences from the GenBank database and compared with the codon usage in B virus genes (Table 4). As expected, the overall codon usage patterns of B virus and the macaque hosts differed substantially. However, remarkably, the favored codons of the B virus genes mirrored the most frequently used codons in the rhesus macaque genes (except Ser and Arg, both encoded by six codons). Moreover, these most frequent codons in B virus were almost the only codons used for each amino acid. This observation correlates with intriguing data from a number of unicellular and multicellular organisms that highly expressed genes utilize preferable codons in each synonymous group, the optimal codons (20, 21, 23, 61). Thus, the combined effects of directional mutation pressure and selection for preferential codons put the B virus genome in a favorable state to efficiently employ the host cell translation machinery for expression of viral proteins.

TABLE 4.

Codon usage in rhesus macaque and B virus genes^a

Amino acid	Codon	Macaca mulatta		B virus
		No. of occurrences	% of occurrences	No. of occurrences	% of occurrences
His	CAU	972	43	31	3
	CAC	1,284	57	870	97
Tyr	UAU	1,229	43	60	6
	UAC	1,602	57	887	94
Gln	CAA	1,161	30	55	6
	CAG	2,756	70	929	94
Asn	AAU	1,625	46	28	4
	AAC	1,902	54	692	96
Lys	AAA	2,094	45	50	10
	AAG	2,561	55	455	90
Asp	GAU	1,633	44	144	7
	GAC	2,050	56	1,978	93
Glu	GAA	2,141	42	157	7
	GAG	3,018	58	1,986	93
Phe	UUU	1,696	44	268	21
	UUC	2,187	56	1,028	79
Cys	UGU	1,141	47	83	12
	UGC	1,284	53	601	88
Ile	AUU	1,475	32	52	6
	AUC	2,303	50	808	89
	AUA	809	18	47	5
Val	GUU	1,056	19	141	5
	GUC	1,440	25	1,139	41
	GUA	516	09	46	2
	GUG	2,686	47	1,426	52
Pro	CCU	1,366	28	155	4
	CCC	1,476	31	2,014	53
	CCA	1,457	30	120	3
	CCG	518	11	1,527	40
Thr	ACU	1,300	24	31	2
	ACC	2,076	37	992	49
	ACA	1,528	28	58	3
	ACG	577	11	935	46
Gly	GGU	926	17	127	4
	GGC	1,666	30	1,724	54
	GGA	1,641	29	209	6
	GGG	1,352	24	1,180	36
Ala	GCU	1,548	27	144	2
	GCC	2,192	39	3,471	55
	GCA	1,298	23	127	2
	GCG	635	11	2,599	41
Arg	CGU	398	9	82	2
	CGC*	727	17	2,109	56
	CGA	368	8	208	5
	CGG	749	17	1,200	31
	AGA*	1,077	26	47	1
	AGG	1,014	23	180	5
Leu	UUG	1,109	12	80	2
	UUA	665	7	14	0
	CUU	1,261	14	55	1
	CUC	2,930	21	1,373	36
	CUA	562	6	58	2
	CUG	3,465	40	2,256	59
Ser	UCU	1,428	20	93	4
	UCC*	1,731	23	648	29
	UCA	1,122	16	37	2
	UCG*	326	5	778	35
	AGU	1,080	15	41	2
	AGC	1,505	21	620	28
Overall G+C			51.03		74.36
First position G or C			52.55		73.73
Second position G or C			42.12		56.22
Third position G or C			58.42		93.11

^aThe sequence of the M. mulatta genes were obtained from the GenBank database. The most frequently observed codons are in boldface. Codons for which discrepant results were obtained are marked with an asterisk. Nondegenerate amino acids (Met and Trp) are not included.

Conclusions.

The complete genomic sequence of the B virus E2490 laboratory strain demonstrates that the genome organization and gene repertoire are similar but not identical to those of the HSV-1 and HSV-2 genomes. Interestingly, B virus lacks the gene homolog encoding the neurovirulence factor ICP34.5. We postulate that B virus utilizes different mechanisms to block host antivirus responses and facilitate replication in neurons. The enhanced neurovirulence of B virus in human hosts may reflect the inability of human cells to mount efficient antiviral cellular responses against these divergent strategies. The other implication of our findings is that because B virus does not possess the γ₁34.5 gene homolog, acquisition of this gene by human viruses is an even more recent evolutionary event than was suggested previously (6, 7).

Comparative analysis of the B virus and HSV genomes is essential to understand the mechanisms of B virus pathogenesis in humans. The complete sequence of the B virus reference strain also supplies much-needed information to determine the genetic basis of phenotypic and pathogenic differences among B virus isolates derived from different macaque subspecies. With this information, new antiviral and vaccine strategies can be designed to target critical viral components in order to control this deadly zoonotic agent.