New Genes from Old: Redeployment of dUTPase by Her
http://www.100md.com
病菌学杂志 2005年第20期
MRC Virology Unit, Institute of Virology, University of Glasgow, Church Street, Glasgow G11 5JR, United Kingdom
ABSTRACT
Published work (D. J. McGeoch, Nucleic Acids Res. 18:4105-4110, 1990; J. E. McGeehan, N. W. Depledge, and D. J. McGeoch, Curr. Protein Peptide Sci. 2:325-333, 2001) has indicated that evolution of dUTPase in the class of herpesviruses that infect mammals and birds involved capture of a host gene followed by a duplication event that resulted in a coding region comprising two fused dUTPase domains. Some of the conserved residues required for enzyme activity were then lost, resulting in a dUTPase containing a single active site with different elements contributed by each half of the protein. Further conserved residues were lost in one subfamily (the Betaherpesvirinae), yielding a protein that is related to herpesvirus dUTPases but has a different and as yet unrecognized function. Evidence from sequence similarities and structural predictions now indicates that several additional genes were derived from the herpesvirus dUTPase gene, probably by duplication. These are UL31, UL82, UL83, and UL84 in human cytomegalovirus (and counterparts in other members of the Betaherpesvirinae) and ORF10 and ORF11 in human herpesvirus 8 (and counterparts in other members of the Gammaherpesvirinae). The findings clarify the evolutionary history of these genes and provide novel insights for structural and functional studies.
INTRODUCTION
One means of generating proteins with new functions that has been used widely in evolution is that of gene duplication (70). This strategy allows the original function to be retained by one of the gene copies while permitting the development of a distinct function in the other. The occurrence of families of related genes indicates that gene duplication has also played a prominent role in the evolution of large DNA viruses (81). Repeated duplication has often resulted in families consisting of tandem head-to-tail gene arrays, but related genes may also be spread throughout the genome, presumably as a result of relocation during or after duplication. Moreover, viral gene families may be similar to host genes, suggesting that the parental gene was derived originally by capture from the cellular repertoire (20, 81).
The gene encoding 2'-deoxyuridine 5'-triphosphate pyrophosphatase (dUTPase; EC 3.6.1.23) is ubiquitous in all classes of organisms and has been incorporated into nonprimate lentiviruses and many families of large DNA viruses (5, 52, 54). This enzyme catalyzes the hydrolysis of dUTP and, in addition to supplying dUMP for synthesis of TMP by thymidylate synthase, reduces the dUTP pool and thus minimizes misincorporation of uracil residues into newly synthesized DNA. The dUTPases of Escherichia coli and Homo sapiens are identical in only 31% of their residues but exhibit similar crystallographic structures (12, 30, 61).
Figure 1a shows the amino acid sequence of the human protein and highlights regions of secondary structure and the five conserved sequence motifs identified by McGeoch (54). Mol et al. (61) described the domain fold of the human dUTPase monomer as a distorted eight-stranded ?-barrel consisting of a six-stranded antiparallel ?-barrel formed from ?2, ?7, ?4, ?5, ?6, and ?3, supplemented by the parallel strands ?1 and ?8', the latter contributed by an adjacent subunit (Fig. 1b). This structure is capped by a conical arrangement of five antiparallel ?-strands extending from the barrel (?2b, ?6b, ?5, ?6, and ?3), plus an -helix (1). The active enzyme is a trimer, each catalytic site involving residues from all three subunits (Fig. 1c and d). Thus, residues from one subunit primarily recognize uracil and deoxyribose (motif 3), residues from a second subunit recognize the phosphate groups (motifs 2 and 4), and residues from the C-terminal region of a third subunit close the substrate-binding pocket (motif 5). Specificity for deoxyuridine is provided by hydrogen bonding with the ?5-?6 hairpin and the exclusion of ribose by a tyrosine residue in the ?5-?6 turn.
Members of the family Herpesviridae fall into three very widely divergent phylogenetic classes: those infecting mammals, birds, or reptiles (henceforth called the mammalian herpesvirus class), those infecting fish or amphibians (henceforth called the fish herpesvirus class), and a single known virus infecting bivalves (20). Each class possesses a dUTPase gene that appears to have been captured from the host, perhaps on independent occasions. In the fish herpesvirus class, the dUTPases are similar in size to the cellular protein, have the same arrangement of motifs, and presumably are equivalent structurally and functionally to the cellular enzyme (18, 19). The bivalve virus also has a dUTPase that is similar to the cellular enzyme and in addition possesses two related genes of roughly the same size, which are located in other regions of the genome and are likely to have arisen by duplication. These lack active-site residues and probably fulfill other functions (21).
The mammalian herpesvirus class is divided into three phylogenetically distinct subfamilies, the Alpha-, Beta-, and Gammaherpesvirinae. In these viruses, the dUTPases are related to the cellular enzyme but are structurally distinct as the result of an unusual evolutionary history. The most intensively studied mammalian herpesvirus, herpes simplex virus type 1 (HSV-1), belongs to the Alphaherpesvirinae and has long been known to induce a novel dUTPase activity in infected cells (11). The dUTPase genes of other members of the Alphaherpesvirinae and also of the Gammaherpesvirinae have been shown to specify active enzymes (27, 45, 69, 83). The dUTPases of HSV-1 and murid herpesvirus 4 (MuHV-4; a member of the Gammaherpesvirinae) are not required for growth of the virus in cell culture but provide advantages in vivo (25, 71, 84).
In a comparative study, McGeoch (54) observed that the dUTPases of two members of the Alphaherpesvirinae (HSV-1 and varicella-zoster virus) and one member of the Gammaherpesvirinae (Epstein-Barr virus) are about twice the length of their cellular counterparts. Moreover, the conserved regions are arranged differently, with motifs 1, 2, 4, and 5 present in their usual order in the C-terminal half of the protein, and motif 3 in the N-terminal half (compare Fig. 2a and b). These observations prompted the conclusion that the dUTPase gene in an ancestral herpesvirus had been duplicated internally, giving rise initially to a protein containing two complete sets of motifs. Motifs 1, 2, 4, and 5 had then been lost from the N-terminal half and motif 3 from the C-terminal half. The herpesvirus enzyme was therefore envisaged as a monomer, as detected previously (10), with elements of the single active site contributed by each half of the protein. These insights into the contributions made to the active site by the conserved regions turned out to be entirely consistent with those defined subsequently in the trimeric cellular dUTPases.
Aided by the structure of Escherichia coli dUTPase (12) and additional sequence data, McGeehan et al. (53) concluded from secondary-structure predictions that the ?-strand arrangement is conserved in both halves of dUTPases from the Alpha- and Gammaherpesvirinae. This supported a molecular model comprising two structurally similar domains, in which motif 3 from the N-terminal domain is juxtaposed to motifs 1, 2, and 4 from the C-terminal domain, with motif 5 from the C terminus positioned at the active site by virtue of an extended region connecting it to the rest of the protein (Fig. 2c). The conserved region that occupies the position of motif 3 in the C-terminal domain has been termed motif 6, and it is possible that this herpesvirus-specific element contributes a novel function.
In contrast to the Alpha- and Gammaherpesvirinae, the dUTPase orthologs in the Betaherpesvirinae were found to lack motifs 1 to 5 but to have retained motif 6. This prompted the idea that these proteins may no longer function as dUTPases, but instead have other roles, perhaps requiring motif 6 (53, 55). This has recently found experimental support from the observation that the human cytomegalovirus (HCMV) protein lacks dUTPase activity (9). Like active dUTPases in the Alpha- and Gammaherpesvirinae, this protein is not required for growth in cell culture, and its function remains unknown (9, 24, 94).
Alternative functions thus appear to have been generated from the herpesvirus dUTPase in both the bivalve virus (involving gene duplications that permitted retention of the original function and the development of new ones) and in the subfamily Betaherpesvirinae of the mammalian herpesvirus class (involving replacement of the original function by a new one). We reasoned that additional herpesvirus functions could have been derived from dUTPase via gene duplication, but that these might be hidden by extensive divergence. Indeed, we were provoked by McGeehan et al.'s (53) mentioning marginal relationships between dUTPase and the proteins encoded by ORF10 and ORF11 in the Gammaherpesvirinae, although these authors assessed the significance as dubious and took the analysis no further. Having completed a comprehensive analysis of the mammalian herpesvirus class, for which sequence data are abundant, we conclude that several genes in the Beta- and Gammaherpesvirinae were indeed derived from the herpesvirus dUTPase. Our findings clarify the evolutionary history of these genes and provide novel leads for structural studies.
MATERIALS AND METHODS
Sources of sequence data. Sequences were derived from public databases or by personal communication, as detailed in Fig. 4. The tools available at http://www.ncbi.nlm.nih.gov and http://www.ebi.uniprot.org/uniprot-srv/uniProtPowerSearch.do were utilized for this purpose.
Computer programs. Searches for amino acid sequence similarity using BLASTP (3) and PSI-BLAST (2) were carried out with default parameters at http://www.ncbi.nlm.nih.gov. Amino acid sequences were aligned using CLUSTAL W (86) implemented locally, and in some instances the alignments were adjusted manually. Alignments were displayed using CHROMA (31) with default parameters. Secondary structures were predicted from amino acid sequences using PROFsec (74, 75; http://www.predictprotein.org), which provides a facility for processing precomputed alignments. SUPERFAMILY 1.65 was utilized for protein threading (33, 51; http://supfam.mrc-lmb.cam.ac.uk/SUPERFAMILY). This program relies on the SCOP database of all known protein folds (4; http://scop.mrc-lmb.cam.ac.uk/scop). PROFsec and SUPERFAMILY were used on-line with default parameters. Neighbor-joining phylogenetic trees were constructed using MEGA (46; http://www.megasoftware.net) with the PAM (Dayhoff) matrix model, removing gapped regions and assuming uniform substitution rates among sites. Bootstrap values were calculated from 100 replicates.
Gene nomenclature. Gene nomenclature varies from one herpesvirus to another, and as a result orthologous relationships are frequently not obvious. For clarity, HSV-1, HCMV, and human herpesvirus 8 (HHV-8) names were applied throughout for orthologs in the Alpha-, Beta-, and Gammaherpesvirinae, respectively, with alternative names given only in particular instances. Virus-specific names may be obtained via the accession numbers listed in Fig. 4.
The nomenclature problem is compounded for a family of tandemly arranged, related genes (UL82, UL83, and UL84) located in the middle of the HCMV genome and their counterparts in other members of the Betaherpesvirinae (14, 17). Depending on the virus, the number of genes in this family ranges from two to four, and the evolutionary history of the more distantly related members relative to one another is difficult to determine with confidence from phylogenetic analysis (see Table 1 and the Discussion). This confounds application of a common nomenclature that accurately represents orthology. In this study, the leftmost gene in the family was denoted as UL82 for every virus and the rightmost as UL84 where it is clearly orthologous to HCMV UL84. Any other genes were denoted UL83 (where there is a single gene) or UL83A and UL83B (where there are two genes).
RESULTS AND DISCUSSION
Pairwise-based sequence comparisons. Pairwise BLASTP and PSI-BLAST searches were conducted using various herpesvirus dUTPases as probes and the GenBank nonredundant set of viral proteins as the data set. Reciprocal searches were then carried out using proteins that emerged as potentially related to dUTPase. These investigations showed that dUTPases in the Alpha- and Gammaherpesvirinae (encoded by UL50 and ORF54, respectively) are related to the orthologous UL72 protein of the Betaherpesvirinae within a domain of approximately 100 amino acid residues in the C-terminal half of each protein. In addition, more distant relationships were revealed in the same domain to the proteins encoded by HCMV UL31, UL82, UL83, and UL84 (and their orthologs in other members of the Betaherpesvirinae) and HHV-8 ORF10 and ORF11 (and their orthologs in other members of the Gammaherpesvirinae). The locations of these genes in the respective genomes are illustrated in Fig. 3. The conserved dUTPase-related domain (DURD) corresponds closely to the region identified previously as the most highly conserved between UL82, UL83, and UL84, although the relationship of these genes to dUTPases was not recognized at the time (17).
Further PSI-BLAST searches were then conducted using representative DURDs in order to detect all dUTPase-related proteins (DURPs) encoded by the mammalian herpesvirus class. A comprehensive alignment of the DURD in these proteins is shown in Fig. 4. The DURD corresponds to approximately 100 residues containing herpesvirus dUTPase motifs 1, 2, 6, and 4. Conservation of motif 6, with its characteristic tryptophan residue, rather than the positionally equivalent motif 3 of nonherpesvirus dUTPases, with its tyrosine residue, indicates that the DURD was propagated from the herpesvirus dUTPase rather than from an external source. In every case, the DURD is located in the C-terminal half of DURPs (Fig. 5, green blocks).
Secondary structure analysis. Secondary structure predictions were carried out for the various groups of DURDs in order to assess their similarity to human dUTPase. All sequences in each group were analyzed precisely as aligned in Fig. 4, with the BLAST facility in PROFsec disabled so that no other sequences were incorporated. The results (Fig. 4, blue blocks) show that the predicted secondary structure of the DURD is conserved among the groups of DURPs and is strikingly similar to that of human dUTPase. The reliability of the predictions is indicated by the close correspondence of the known human dUTPase structure with the prediction (obtained by alignment with other dUTPase sequences detected using BLAST as implemented in PROFsec). ?-strands corresponding to ?2 to ?7 were predicted for each herpesvirus group, except for ?5a and ?5b in the UL82/UL83/UL84 group and ?5b in the UL50/ORF54 and "all herpes" groups. No -helix was predicted for any of the proteins, including human dUTPase. Taking into account the limitations of secondary structure prediction, each DURP is anticipated to contain in its C-terminal half the six-stranded antiparallel ?-barrel characteristic of dUTPases.
To investigate the status of ?1 and ?8, which contribute to the eight-stranded ?-barrel in cellular dUTPases and are located partly or wholly outside the DURD, secondary structures were predicted for entire proteins (Fig. 5, blue and red blocks). In contrast to predictions for the DURDs, which were based on the alignments exactly as shown in Fig. 4, other regions of the proteins are generally too divergent to permit the construction of robust alignments incorporating all members of a particular group. Consequently, predictions were carried out by allowing alignments to be generated within PROFsec and using a cutoff facility to remove proteins (in part or in entirety) that are either too closely or too distantly related to the probe sequence. This ensured that predictions were based on sound representative alignments, but had the disadvantage that they were restricted to a subset of sequences in any protein group, and to different numbers of sequences in different parts of the protein. The reliability of the predictions therefore varied from protein to protein, and from one part of a protein to another. The results (Fig. 5) indicated that the N-terminal part of each protein is rich in ?-strand, but were not revealing about the presence of ?1 and ?8. The most that could be inferred was that a ?-strand in each protein is positioned close to the N-terminal end of the DURD and may correspond to ?1. More equivocally, each protein contains one or more ?-strands between the DURD and the C terminus that may correspond to ?8.
McGeehan et al. (53) noted from secondary structure predictions that an arrangement of ?-strands similar to that in the C-terminal half of herpesvirus dUTPases of the Alpha- and Gammaherpesvirinae is present in the N-terminal half, implying the presence of two ?-barrels in the protein. However, the N-terminal portions of herpesvirus dUTPases are more divergent than the C-terminal portions, and any N-terminal ?-barrel is not obvious from amino acid sequence comparisons with the C-terminal one. Nonetheless, the likelihood that dUTPases contain two DURDs raises the question of whether the same applies to other DURPs.
Profile-based sequence comparisons. Protein threading programs, which attempt to fit query sequences to models derived from known protein structures, are showing increasing promise as a way of identifying folds in such sequences, even in the absence of detectable sequence conservation. SUPERFAMILY is a threading program that offers a number of advantages. It searches a comprehensive collection of protein structures, identifies multiple folds within the same protein, and distinguishes between significant and marginal predictions among the output scores. Having analyzed a wide range of herpesvirus proteins using SUPERFAMILY, our evaluation was that this program is sufficiently conservative to inspire a substantial degree of confidence.
Nevertheless, given that it can be difficult to discriminate between meaningful and spurious findings among the output from threading programs, our analysis was undertaken cautiously, with the aim of confirming the case for a ?-barrel in the C-terminal part of DURPs and of examining the N-terminal part for the presence of a second ?-barrel. The full-length versions of all the DURPs represented in Fig. 4 were submitted to SUPERFAMILY, and the presence of dUTPase-like domains (?-barrels) was recorded (Table 2 and Fig. 5, orange blocks). As expected, a ?-barrel was scored in the C-terminal part of every dUTPase (UL50 and ORF54) from the Alpha- and Gammaherpesvirinae, and in the C-terminal part of every UL72 protein from the Betaherpesvirinae. Confident predictions of a domain in the N-terminal part were made for most dUTPases, and marginal predictions for the others, supporting the conclusion of McGeehan et al. (53) from secondary-structure predictions that herpesvirus dUTPases consist of two similar domains.
SUPERFAMILY also predicted that the N-terminal part of most UL72 proteins contains a ?-barrel. Similarly, the C-terminal part of most other DURPs (UL82, UL83, UL84, and UL31 in the Betaherpesvirinae and ORF10 and ORF11 in the Gammaherpesvirinae) was predicted to contain a ?-barrel. However, the situation with the N-terminal parts of these proteins was much less clear, as only a minority were predicted to contain a ?-barrel. Unusually, predictions were made with greater confidence for the N-terminal part of the UL82 proteins of human herpesvirus 6 (HHV-6) and human herpesvirus 7 (HHV-7) than for the C-terminal part.
Bearing in mind the limitations of threading programs, we predict that the C-terminal part of DURPs contains a ?-barrel, and that the N-terminal part of herpesvirus dUTPases and their UL72 orthologs in the Betaherpesvirinae contains a second ?-barrel. Overall, the C-terminal ?-barrel is better conserved than the N-terminal one, indicating that the latter is more responsive to evolutionary pressures. There are indications that a ?-barrel is also present in the N-terminal part of some other DURPs (from Table 2, specifically UL82, UL83B, ORF10, and ORF11).
An analysis of HCMV proteins was performed previously by Novotny et al. using the threading program ProCeryon (66). These authors listed the UL82, UL84, and UL31 proteins as containing ?-barrel domains, although the coordinates given for the latter two assignments are C-terminal to the DURD, and the UL82 prediction was considered marginal. The UL84 prediction included the sequence of the DURD and was listed as being of medium confidence. In none of these cases was a dUTPase fold described as being the best match to the HCMV protein.
Evolution of DURPs. Our identification of DURPs depends only upon sequence and structure relationships and is independent of phylogenetic analysis. Indeed, phylogenetic analysis of the DURD sequences was not particularly robust since their modest length and wide divergence resulted in substantial sensitivity to the parameters used and generally low bootstrap support. The example tree shown in Fig. 6 groups positionally equivalent genes together. However, in the deeper branches only the bootstrap values that establish the UL72 (79 of 100), UL31 (78 of 100) and ORF11 groups (81 of 100) might be considered convincing. Other inferences from this particular tree, such as the UL84 group seeming more closely related to the UL31 group than to UL82/UL83 proteins, are poorly supported (29 of 100). Members of the UL82/UL83/UL84 family of the Betaherpesvirinae grouped generally with each other, consistent with the proposition that they have arisen by duplications of an ancestral gene encoding the "matrix protein ancestor" (17).
Nonphylogenetic considerations also contribute to understanding the evolution of the DURPs. The substantial evidence for a DURD in the C-terminal part and the weaker evidence for a second DURD in the N-terminal part (Fig. 5 and Table 2) suggests that DURPs arose as a result of whole-gene duplication events rather than transfer of the domain to existing genes. However, it is difficult to determine the order of events involved. Given their tandem arrangement, one possibility is that ORF10 and ORF11 may have arisen from local duplication of an ancestral gene derived earlier from the active dUTPase gene (ORF54) in the Gammaherpesvirinae. In contrast, the ancestor of the UL31, UL82, UL83, and UL84 proteins of the Betaherpesvirinae may already have lost dUTPase function by making the transition to UL72.
Perhaps more likely, however, is that duplication and loss of function prior to divergence of the Beta- and Gammaherpesvirinae gave rise to a common ancestor of ORF10, ORF11, UL31, UL82, UL83, and UL84. This is supported by the observation that the ORF10/ORF11 and UL82/UL83/UL84 families are positionally equivalent with respect to the gene encoding the protease (ORF17 and UL80, respectively; denoted PR in Fig. 3, shaded green). The protease gene is conserved in the Alpha-, Beta-, and Gammaherpesvirinae and is located near the end of block IV, one of the blocks of genes that are arranged differently in the three subfamilies (60). The protease gene is adjacent to UL82 in the Betaherpesvirinae and separated from ORF11 in the Gammaherpesvirinae by a variable number of recently captured cellular genes; this number happens to be rather large in HHV-8 but is only one or two in many members of the Gammaherpesvirinae. UL31 of the Betaherpesvirinae probably represents an additional relocated copy derived from the matrix protein ancestor gene.
Roles of the DURD. The above analyses indicate the existence of an extensive family of herpesvirus proteins that have arisen from an ancestral viral dUTPase and comprises several distinct sets of orthologous proteins retaining a common structural domain. How might this be reflected in the functions of these proteins?
The presence of the DURD could reflect the original biochemical function or the fact that it provides an amenable architectural element. In contrast to the functional dUTPases encoded by the Alpha- and Gammaherpesvirinae, none of the other DURPs retains the residues required for dUTPase activity. Nevertheless, parts of the protein fold involved in sugar and phosphate binding are predicted to be present, and it is possible that all these proteins exploit such a property. The dUTPase superfamily is one of several superfamilies that share a structure consisting of six ?-strands termed the ?-clip fold (38). This fold is thought to have been derived by duplication of a three-strand unit and includes the dUTPase ?-barrel containing ?2 to ?7. Many ?-clip proteins are involved in binding to carbohydrate residues (38, 54), and in dUTPases this activity centers on motif 3, containing ?5 and ?6. However, despite these observations, no evidence has been reported for other DURPs binding to carbohydrates.
Herpesvirus-specific motif 6 occupies the position of motif 3 in the cellular dUTPase monomer and is retained not only in the UL72 proteins of the Betaherpesvirinae, which are orthologs of the functional dUTPases, but throughout the whole family of DURPs (Fig. 4). This motif contains the most highly conserved residue in the DURD (a tryptophan in the ?5-?6 loop). Inspection of the human dUTPase structure (Fig. 1) indicates that motif 6 is likely to be located on the opposite face of the molecule from the dUTP-binding site. Any function conferred by this motif has not been identified, and detailed site-directed mutagenesis may prove informative.
The structure of the human enzyme shows that the two complementary faces in a DURD could facilitate intermolecular interactions. Thus, DURDs may enable proteins to assemble into dimers or larger structures. In this regard, it is interesting that several DURPs are incorporated into the tegument (the layer of the virus particle between the capsid and envelope), some in relatively large quantities. Furthermore, DURPs might have evolved to enable interactions with other proteins. Unfortunately, the area of research is insufficiently developed to indicate whether DURDs are important for any of the interactions of DURPs documented to date (16, 35, 41, 78, 85).
An interesting parallel to the herpesvirus DURPs is available among the adenoviruses. Human adenovirus E4 ORF1 encodes a catalytically inactive derivative of the cellular dUTPase monomer, whose ancestor was presumably acquired from the cell (22, 38, 89). Indeed, in various nonhuman adenoviruses this protein has retained the catalytic motifs of an active enzyme. Figure 7 shows that the full-length human adenovirus E4 ORF1 protein corresponds to a C-terminally truncated dUTPase monomer, with all eight ?-strands contributing to a predicted ?-barrel. It appears that the functions of the E4 ORF1 protein in transformation (26, 47) must be attributed directly to the DURD, since the protein contains little or nothing by way of additional elements. These observations point to some general utility of the dUTPase ?-barrel (as a specific type of ?-clip fold) that has proved valuable during evolution of herpesviruses and adenoviruses.
The finding that herpesvirus DURPs exhibit a well-characterized protein fold contributes insights relevant to their study. For example, Hofmann et al. (35) identified a motif in the HCMV UL82 protein (residues 324 to 331) that is similar to a sequence known to be important in other proteins for interaction with the human Daxx protein (hDaxx). Although deletion of the 8-residue sequence abolished interaction of UL82 with hDaxx, substitutions within it had no effect. The authors speculated that the deletion might have had a major effect on the three dimensional structure of the protein. Our analysis indicates that this is likely to be the case, since the region deleted corresponds closely to ?5. Similarly, it was observed that mutants with C-terminal truncations to residue 508 of HCMV UL84 retained the ability to localize to the nucleus, whereas those truncated at residue 488 remained cytoplasmic (48), even though the intervening sequence is not part of a functional nuclear localization signal (48, 91). Again, it seems probable that removal of these 20 residues, which correspond to ?6b and ?7, would result in major structural changes to the protein.
Functions of DURPs. To date, only a small number of DURPs have been studied in any detail, and it is already apparent that they perform various functions. It is important to emphasize that DURPs exhibit a high degree of divergence, and it is probable that certain protein functions are not conserved even between orthologous proteins, as is the situation with the UL50 and ORF54 dUTPases in the Alpha- and Gammaherpesvirinae in comparison with their UL72 orthologs in the Betaherpesvirinae. Thus, functions inferred from sequence comparisons must be viewed with caution. Moreover, it should be appreciated that the DURD is relatively small, and even the presence of two copies would account in most instances for only a minority of the DURP (Fig. 5). Key aspects of function are therefore probably specified by other regions, which may have evolved by independent gene expansion events, including the acquisition of other protein domains by recombination. Thus, DURPs may perform a wide range of unrelated roles while retaining a common function provided by the DURD.
HCMV UL82 encodes the tegument phosphoprotein pp71. This protein functions as a transcriptional transactivator and is essential for virus growth at low multiplicity of infection in cell culture (7, 37, 49, 76). Similar properties have been reported for the guinea pig cytomegalovirus (GPCMV) UL82 protein (57). HCMV pp71 interacts with cellular proteins hDaxx and Rb and induces cell cycle progression (35, 40, 42). An interaction with another tegument protein, ppUL35, has been reported, and appears to be important for activation of the major immediate early promoter (78).
The HCMV UL83 protein is also a phosphorylated tegument protein (pp65) and the major constituent of virions and noninfectious dense bodies (76, 88). In contrast to pp71, pp65 is not essential for growth in cell culture (80). Similar properties have been reported for the UL83 proteins of GPCMV and murine cytomegalovirus (MCMV) (43, 57, 63, 79). HCMV pp65 dampens the interferon response to infection, but conflicting mechanisms for this effect on the host have been presented (1, 8). It also represents an immunodominant target for host cytotoxic T-cell responses (58, 90). In contrast, the corresponding MCMV M83 protein is not a dominant factor in the T-cell response (64), but M83-specific T cells may nevertheless be important in resolving acute infection (36).
The HCMV UL84 protein has been reported to be present in virions, but appears to be a minor constituent (88). However, it is essential for growth in cell culture (24, 92, 94). The protein is capable of self-oligomerization, and interacts with and modulates the activity of the transcriptional transactivator IE2 (16, 28, 85). Transient DNA replication assays and characterization of a null mutant virus have indicated that the UL84 protein probably performs an essential role during initiation of HCMV DNA synthesis (67, 77, 92, 93). Surprisingly, however, the protein was found to be dispensable for HCMV origin-dependent DNA synthesis when the HSV-1 DNA replication fork proteins were employed in a transient assay (72).
Colletti et al. recently reported that HCMV UL84 encodes a nucleoside triphosphatase (NTPase) activity with a preference for UTP as the substrate (15). It should be emphasized that UTPase activity (which converts UTP to UDP plus phosphate) is enzymatically unrelated to that of dUTPase (which converts dUTP to dUMP plus pyrophosphate). These authors also noted the presence of several sequence motifs within the UL84 protein and suggested it belongs to the DExD/H box family, many members of which are known to be NTP-utilizing helicases (32, 34, 82, 87). Although matches were found for DExD/H box motifs I, Ia, II, IV, and V, there appeared to be no identifiable counterparts of the two other conserved motifs, III and VI (15). The spacing of the identified motifs was also atypical for DExD/H box proteins, with motifs I and Ia separated by an unusually large distance and motifs II and IV by only 6 residues. In addition, our amino acid sequence alignment for HCMV UL84 and the homologous proteins of other primate cytomegaloviruses revealed that proposed motifs I and II were not conserved, but rather were located in regions of the HCMV protein characterized by significant insertions or deletions in the homologous proteins (data not shown). Interestingly, the suggested matches to motifs Ia, II, and part of motif IV were located within the region of the UL84 protein corresponding to the DURD.
The structures of several DExD/H box proteins are known and the regions containing the conserved motifs exhibit a high degree of conservation in their folding pattern, which is quite distinct from that of the DURD (73, 34, 87). When authentic DExD/H box proteins were submitted to SUPERFAMILY, they were identified as containing the P-loop nucleoside triphosphate hydrolase fold in the region corresponding to motifs I, Ia, II, and III, whereas this fold was not predicted for HCMV UL84 (data not shown). Taken together, these observations raise serious doubt as to whether the HCMV UL84 protein is a member of the DExD/H box family and suggest that site-directed mutagenesis experiments should be performed to confirm the assignment of a UTPase activity to this protein.
MCMV M84 has sometimes been viewed as the homolog of HCMV UL84. However, M84 is distantly related to UL84 (17), and, indeed, is more closely related in the DURD to the UL82/UL83 proteins (Fig. 6). The properties of the M84 and UL84 proteins are also distinctly different. The M84 protein is not essential for replication in cell culture (63) and is therefore unlikely to play a key role in DNA synthesis. Rather, it appears to function in the T-cell response (36, 64). However, despite its closer relationship to the UL82/UL83 proteins, M84 was not detected in virions (43). Given the high degree of conservation between other HCMV and MCMV DNA replication proteins, it is surprising that MCMV encodes no obvious homolog of UL84. This could reflect functional differences between MCMV and HCMV, for example, in the origin of DNA replication. Nonetheless, several interesting questions remain regarding the functions of the DURPs in this region of the genome of the nonprimate cytomegaloviruses, only some of which appear to have sequence counterparts of UL84, and the identities of the DNA synthesis initiator proteins for these viruses. Unfortunately, insufficient data are available to facilitate functional comparisons with the corresponding HHV-6 and HHV-7 genes.
Very little information is available on the DURPs other than those in the UL82/UL83/UL84 group. HCMV UL31 is not essential for growth in cell culture (24, 94) and, although the HCMV protein was not detected in virions, its MCMV ortholog was (43, 88). As far as expression kinetics are concerned, the DURPs that have been examined are early or late lytic cycle genes, compatible with their proposed connections to nucleotide metabolism, DNA synthesis, gene regulation, or virion structure (13, 17, 39, 44, 50, 79). HHV-7 U10 (the UL31 ortholog) is unusual in being transcribed under immediate-early conditions (59). EBV LF1 and LF2 (the ORF10 and ORF11 orthologs) are not required for growth in cell culture, since they lie in a region absent from the B95-8 strain (68). MuHV-4 ORF10 and ORF11 are also not essential, and the latter encodes a virion component (6, 62). The HHV-8 ORF11 protein has been reported to be present in virions (95).
In conclusion, despite our presently limited knowledge of the roles of members of the DURP family, it is clear that these proteins exhibit wide differences with regard to their functions, presence in the virion, kinetics of expression, requirement for growth in cell culture, immunological properties, and effects on the host cell. The discovery that they have evolved from an ancestral herpesvirus gene and share predicted structural properties should serve to focus further functional characterization of this large family of related proteins.
ACKNOWLEDGMENTS
This work was supported by the United Kingdom Medical Research Council.
We are grateful to Aidan Dolan, Mark Schleiss, and James Stewart for generously providing unpublished sequence data for SCMV, GPCMV, and OvHV-2, respectively. We thank Edgar Sevilla-Reyes, Chris Preston, and Duncan McGeoch for comments on the manuscript.
REFERENCES
Abate, D. A., S. Watanabe, and E. S. Mocarski. 2004. Major human cytomegalovirus structural protein pp65 (ppUL83) prevents interferon response factor 3 activation in the interferon response. J. Virol. 78:10995-11006.
Altschul, S. F., T. L. Madden, A. A. Sch?ffer, J. Zhang, Z. Zhang, W. Miller, and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25:3389-3402.
Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. Basic local alignment search tool. J. Mol. Biol. 215:403-410.
Andreeva, A., D. Howorth, S. E. Brenner, T. J. Hubbard, C. Chothia, and A. G. Murzin. 2004. SCOP database in 2004: refinements integrate structure and sequence family data. Nucleic Acids Res. 32:D226-229.
Baldo, A. M., and M. A. McClure. 1999. Evolution and horizontal transfer of dUTPase-encoding genes in viruses and their hosts. J. Virol. 73:7710-7721.
Boname, J. M., J. S. May, and P. G. Stevenson. 2005. Murine gammaherpesvirus 68 open reading frame 11 encodes a nonessential virion component. J. Virol. 79:3163-3168.
Bresnahan, W. A., and T. E. Shenk. 2000. UL82 virion protein activates expression of immediate early viral genes in human cytomegalovirus-infected cells. Proc. Natl. Acad. Sci. USA 97:14506-14511.
Browne, E. P., and T. Shenk. 2003. Human cytomegalovirus UL83-coded pp65 virion protein inhibits antiviral gene expression in infected cells. Proc. Natl. Acad. Sci. USA 100:11439-11444.
Caposio, P., L. Riera, G. Hahn, S. Landolfo, and G. Gribaudo. 2004. Evidence that the human cytomegalovirus 46-kDa UL72 protein is not an active dUTPase but a late protein dispensable for replication in fibroblasts. Virology 325:264-276.
Caradonna, S. J., and D. M. Adamkiewicz. 1984. Purification and properties of the deoxyuridine triphosphate nucleotidohydrolase enzyme derived from HeLa S3 cells. Comparison to a distinct dUTP nucleotidohydrolase induced in herpes simplex virus-infected HeLa S3 cells. J. Biol. Chem. 259:5459-5464.
Caradonna, S. J., and Y. C. Cheng. 1981. Induction of uracil-DNA glycosylase and dUTP nucleotidohydrolase activity in herpes simplex virus-infected human cells. J. Biol. Chem. 256:9834-9837.
Cedergren-Zeppezauer, E. S., G. Larsson, P. O. Nyman, Z. Dauter, and K. S. Wilson. 1992. Crystal structure of a dUTPase. Nature 355:740-743.
Chambers, J., A. Angulo, D. Amaratunga, H. Guo, Y. Jiang, J. S. Wan, A. Bittner, K. Frueh, M. R. Jackson, P. A. Peterson, M. G. Erlander, and P. Ghazal. 1999. DNA microarrays of the complex human cytomegalovirus genome: profiling kinetic class with drug sensitivity of viral gene expression. J. Virol. 73:5757-5766.
Chee, M. S., A. T. Bankier, S. Beck, R. Bohni, C. M. Brown, R. Cerny, T. Horsnell, C. A. Hutchison III, T. Kouzarides, J. A. Martignetti, E. Preddie, S. C. Satchwell, P. Tomlinson, K. M. Weston, and B. G. Barrell. 1990. Analysis of the protein-coding content of the sequence of human cytomegalovirus strain AD169. Curr. Top. Microbiol. Immunol. 154:125-169.
Colletti, K. S., Y. Xu, I. Yamboliev, and G. S. Pari. 2005. Human cytomegalovirus UL84 is a phosphoprotein that exhibits UTPase activity and is a putative member of the DExD/H box family of proteins. J. Biol. Chem. 280:11955-11960.
Colletti, K. S., Y. Xu, S. A. Cei, M. Tarrant, and G. S. Pari. 2004. Human cytomegalovirus UL84 oligomerization and heterodimerization domains act as transdominant inhibitors of oriLyt-dependent DNA replication: evidence that IE2-UL84 and UL84-UL84 interactions are required for lytic DNA replication. J. Virol. 78:9203-9214.
Cranmer, L. D., C. L. Clark, C. S. Morello, H. E. Farrell, W. D. Rawlinson, and D. H. Spector. 1996. Identification, analysis, and evolutionary relationships of the putative murine cytomegalovirus homologs of the human cytomegalovirus UL82 (pp71) and UL83 (pp65) matrix phosphoproteins. J. Virol. 70:7929-7939.
Davison, A. J. 1992. Channel catfish virus: a new type of herpesvirus. Virology 186:9-14.
Davison, A. J. 1998. The genome of salmonid herpesvirus 1. J. Virol. 72:1974-1982.
Davison, A. J. 2002. Evolution of the herpesviruses. Vet. Microbiol. 86:69-88.
Davison, A. J., B. L. Trus, N. Cheng, A. C. Steven, M. S. Watson, C. Cunningham, R. M. Le Deuff, and T. Renault. 2005. A novel class of herpesvirus with bivalve hosts. J. Gen. Virol. 86:41-53.
Davison, A. J., M. Benk?, and B. Harrach. 2003. Genetic content and evolution of adenoviruses. J. Gen. Virol. 84:2895-2908.
Dolan, A., C. Cunningham, R. D. Hector, A. F. Hassan-Walker, L. Lee, C. Addison, D. J. Dargan, D. J. McGeoch, D. Gatherer, V. C. Emery, P. D. Griffiths, C. Sinzger, B. P. McSharry, G. W. G. Wilkinson, and A. J. Davison. 2004. Genetic content of wild-type human cytomegalovirus. J. Gen. Virol. 85:1301-1312.
Dunn, W., C. Chou, H. Li, R. Hai, D. Patterson, V. Stolc, H. Zhu, and F. Liu. 2003. Functional profiling of a human cytomegalovirus genome. Proc. Natl. Acad. Sci. USA 100:14223-14228.
Fisher, F. B., and V. G. Preston. 1986. Isolation and characterisation of herpes simplex virus type 1 mutants which fail to induce dUTPase activity. Virology 148:190-197.
Frese, K. K., S. S. Lee, D. L. Thomas, I. J. Latorre, R. S. Weiss, B. A. Glaunsinger, and R. T. Javier. 2003. Selective PDZ protein-dependent stimulation of phosphatidylinositol 3-kinase by the adenovirus E4-ORF1 oncoprotein. Oncogene 22:710-721.
Fuchs, W., K. Ziemann, J. P. Teifke, O. Werner, and T. C. Mettenleiter. 2000. The non-essential UL50 gene of avian infectious laryngotracheitis virus encodes a functional dUTPase which is not a virulence factor. J. Gen. Virol. 81:627-638.
Gebert, S., S. Schmolke, G. Sorg, S. Floss, B. Plachter, and T. Stamminger. 1997. The UL84 protein of human cytomegalovirus acts as a transdominant inhibitor of immediate-early-mediated transactivation that is able to prevent viral replication. J. Virol. 71:7048-7060.
Glenn, M., L. Rainbow, F. Aurade, A. Davison, and T. F. Schulz. 1999. Identification of a spliced gene from Kaposi's sarcoma-associated herpesvirus encoding a protein with similarities to latent membrane proteins 1 and 2A of Epstein-Barr virus. J. Virol. 73:6953-6963.
González, A., G. Larsson, R. Persson, and E. Cedergren-Zeppezauer. 2001. Atomic resolution structure of Escherichia coli dUTPase determined ab initio. Acta Crystallogr. D Biol. Crystallogr. 57:767-774.
Goodstadt, L., and C. P. Ponting. 2001. CHROMA: consensus-based colouring of multiple alignments for publication. Bioinformatics 17:845-846.
Gorbalenya, A. E., and E. V. Koonin. 1993. Helicases: amino acid sequence comparisons and structure-function relationships. Curr. Opin. Struct. Biol. 3:419-429.
Gough, J., K. Karplus, R. Hughey, and C. Chothia. 2001. Assignment of homology to genome sequences using a library of hidden Markov models that represent all proteins of known structure. J. Mol. Biol. 313:903-919.
Hall, M. C., and S. W. Matson. 1999. Helicase motifs: the engine that powers DNA unwinding. Mol. Microbiol. 34:867-877.
Hofmann, H., H. Sindre, and T. Stamminger. 2002. Functional interaction between the pp71 protein of human cytomegalovirus and the PML-interacting protein human Daxx. J. Virol. 76:5769-5783.
Holtappels, R., J. Podlech, N. K. A. Grzimek, D. Thomas, M.-F. Pahl-Seibert, and M. J. Reddehase. 2001. Experimental preemptive immuonotherapy of murine cytomegalovirus disease with CD8 T-cell lines specific for ppM83 and pM84, the two homologs of human cytomegalovirus tegument protein ppUL83 (pp65). J. Virol. 75:6584-6600.
Homer, E. G., A. Rinaldi, M. J. Nicholl, and C. M. Preston. 1999. Activation of herpesvirus gene expression by the human cytomegalovirus protein pp71. J. Virol. 73:8512-8518.
Iyer, L. M., and L. Aravind. 2004. The emergence of catalytic and structural diversity within the beta-clip fold. Proteins 55:977-991.
Jenner, R. G., M. M. Alba, C. Boshoff, and P. Kellam. 2001. Kaposi's sarcoma-associated herpesvirus latent and lytic gene expression as revealed by DNA arrays. J. Virol. 75:891-902.
Kalejta, R. F., and T. Shenk. 2003a. The human cytomegalovirus UL82 gene product (pp71) accelerates progression through the G1 phase of the cell cycle. J. Virol. 77:3451-3459.
Kalejta, R. F., and T. Shenk. 2003b. Proteasome-dependent, ubiquitin-independent degradation of the Rb family of tumor suppressors by the human cytomegalovirus pp71 protein. Proc. Natl. Acad. Sci. USA 100:3263-3268.
Kalejta, R. F., J. T. Bechtel, and T. Shenk. 2003. Human cytomegalovirus pp71 stimulates cell cycle progression by inducing the proteasome-dependent degradation of the retinoblastoma family of tumor suppressors. Mol. Cell. Biol. 23:1885-1895.
Kattenhorn, L. M., R. Mills, M. Wagner, A. Lomsadze, V. Makeev, M. Borodovsky, H. L. Ploegh, and B. M. Kessler. 2004. Identification of proteins associated with murine cytomegalovirus virions. J. Virol. 78:11187-11197.
Kieff, E., and A. B. Rickinson. 2001. Epstein-Barr virus and its replication, p. 2511-2573. In D. M. Knipe and P. M. Howley (ed.), Fields virology, 4th ed. Lippincott Williams and Wilkins, Philadelphia, Pa.
Kremmer, E., P. Sommer, D. Holzer, S. A. Galetsky, V. A. Molochkov, V. Gurtsevitch, C. Winkelmann, R. Lisner, G. Niedobitek, and F. A. Grasser. 1999. Kaposi's sarcoma-associated herpesvirus (human herpesvirus-8) ORF54 encodes a functional dUTPase expressed in the lytic replication cycle. J. Gen. Virol. 80:1305-1310.
Kumar, S., K. Tamura, I. B. Jakobsen, and M. Nei. 2001. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 17:1244-1245.
Leppard, K. N. 1997. E4 gene function in adenovirus, adenovirus vector and adeno-associated virus infections. J. Gen. Virol. 78:2131-2138.
Lischka, P., G. Sorg, M. Kann, M. Winkler, and T. Stamminger. 2003. A nonconventional nuclear localization signal within the UL84 protein of human cytomegalovirus mediates nuclear import via the importin /? pathway. J. Virol. 77:3734-3748.
Liu, B., and M. F. Stinski. 1992. Human cytomegalovirus contains a tegument protein that enhances transcription from promoters with upstream ATF and AP-1 cis-acting elements. J. Virol. 66:4434-4444.
Lu, M., J. Suen, C. Frias, R. Pfeiffer, M. H. Tsai, E. Chuang, and S. L. Zeichner. 2004. Dissection of the Kaposi's sarcoma-associated herpesvirus gene expression program by using the viral DNA replication inhibitor cidofovir. J. Virol. 78:13637-13652.
Madera, M., C. Vogel, S. K. Kummerfeld, C. Chothia, and J. Gough. 2004. The SUPERFAMILY database in 2004: additions and improvements. Nucleic Acids Res. 32:D235-D239.
McClure, M. A. 2001. Evolution of the DUT gene: horizontal transfer between host and pathogen in all three domains of life. Curr. Protein Peptide Sci. 2:313-324.
McGeehan, J. E., N. W. Depledge, and D. J. McGeoch. 2001. Evolution of the dUTPase gene of mammalian and avian herpesviruses. Curr. Protein Peptide Sci. 2:325-333.
McGeoch, D. J. 1990. Protein sequence comparisons show that the ‘pseudoproteases’ encoded by poxviruses and certain retroviruses belong to the deoxyuridine triphosphatase family. Nucleic Acids Res. 18:4105-4110.
McGeoch, D. J., and A. J. Davison. 1999. The molecular evolutionary history of the herpesviruses, p. 441-465. In E. Domingo, R. Webster, and J. Holland (ed.), Origin and evolution of viruses. Academic Press, London, England.
McGeoch, D. J., M. A. Dalrymple, A. J. Davison, A. Dolan, M. C. Frame, D. McNab, L. J. Perry, J. E. Scott, and P. Taylor. 1988. The complete DNA sequence of the long unique region in the genome of herpes simplex virus type 1. J. Gen. Virol. 69:1531-1574.
McGregor, A., F. Liu, and M. R. Schleiss. 2004. Molecular, biological, and in vivo characterization of the guinea pig cytomegalovirus (CMV) homologs of the human CMV matrix proteins pp71 (UL82) and pp65 (UL83). J. Virol. 78:9872-9889.
McLaughlin-Taylor, E., H. Pande, S. J. Forman, B. Tanamachi, C. R. Li, J. A. Zaia, P. D. Greenberg, and S. R. Riddell. 1994. Identification of the major late human cytomegalovirus matrix protein pp65 as a target antigen for CD8+ virus-specific cytotoxic T lymphocytes. J. Med. Virol. 43:103-110.
Menegazzi, P., M. Galvan, A. Rotola, T. Ravaioli, A. Gonelli, E. Cassai, and D. Di Luca. 1999. Temporal mapping of transcripts of human herpesvirus-7. J. Gen. Virol. 80:2705-2712.
Minson, A. C., A. Davison, R. Eberle, R. C. Desrosiers, B. Fleckenstein, D. J. McGeoch, P. E. Pellett, B. Roizman, and M. J. Studdert. 2000. Family Herpesviridae, p. 203-225. In M. H. V. van Regenmortel, C. M. Fauquet, D. H. L. Bishop, E. B. Carstens, M. K. Estes, S. M. Lemon, J. Maniloff, M. A. Mayo, D. J. McGeoch, C. R. Pringle, and R. B. Wickner (ed.), Virus taxonomy: seventh report of the International Committee on Taxonomy of Viruses. Academic Press, London, England.
Mol, C. D., J. M. Harris, E. M. McIntosh, and J. A. Tainer. 1996. Human dUTP pyrophosphatase: uracil recognition by a ? hairpin and active sites formed by three separate subunits. Structure 4:1077-1092.
Moorman, N. J., C. Y. Lin, and S. H. Speck. 2004. Identification of candidate gammaherpesvirus 68 genes required for virus replication by signature-tagged transposon mutagenesis. J. Virol. 78:10282-10290.
Morello, C. S., L. D. Cranmer, and D. H. Spector. 1999. In vivo replication, latency, and immunogenicity of murine cytomegalovirus mutants with deletions in the M83 and M84 genes, the putative homologs of human cytomegalovirus pp65 (UL83). J. Virol. 73:7678-7693.
Morello, C. S., L. D. Cranmer, and D. H. Spector. 2000. Suppression of murine cytomegalovirus (MCMV) replication with a DNA vaccine encoding MCMV M84 (a homolog of human cytomegalovirus pp65). J. Virol. 74:3696-3708.
Neipel, F., J. C. Albrecht, and B. Fleckenstein. 1997. Cell-homologous genes in the Kaposi's sarcoma-associated rhadinovirus human herpesvirus 8: determinants of its pathogenicity? J. Virol. 71:4187-4192.
Novotny, J., I. Rigoutsos, D. Coleman, and T. Shenk. 2001. In silico structural and functional analysis of the human cytomegalovirus (HHV5) genome. J. Mol. Biol. 310:1151-1166.
Pari, G. S., and D. G. Anders. 1993. Eleven loci encoding trans-acting factors are required for transient complementation of human cytomegalovirus oriLyt-dependent DNA replication. J. Virol. 67:6979-6988.
Parker, B. D., A. Bankier, S. Satchwell, B. Barrell, and P. J. Farrell. 1990. Sequence and transcription of Raji Epstein-Barr virus DNA spanning the B95-8 deletion region. Virology 179:339-346.
Preston, V. G., and F. B. Fisher. 1984. Identification of the herpes simplex virus type 1 gene encoding the dUTPase. Virology 138:58-68.
Prince, V. E., and F. B. Pickett. 2002. Splitting pairs: the diverging fates of duplicated genes. Nat. Rev. Genet. 3:827-837.
Pyles, R. B., N. M. Sawtell, and R. L. Thompson. 1992. Herpes simplex virus type 1 dUTPase mutants are attenuated for neurovirulence, neuroinvasiveness, and reactivation from latency. J. Virol. 66:6706-6713.
Reid, G. G., V. Ellsmore, and N. D. Stow. 2003. An analysis of the requirements for human cytomegalovirus oriLyt-dependent DNA synthesis in the presence of the herpes simplex virus type 1 replication fork proteins. Virology 308:303-316.
Rocak, S., and P. Linder. 2004. DEAD-box proteins: the driving forces behind RNA metabolism. Nat. Rev. Mol. Cell. Biol. 5:232-241.
Rost, B. 1996. PHD: predicting one-dimensional protein structure by profile-based neural networks. Methods Enzymol. 266:525-539.
Rost, B., and C. Sander. 1993. Improved prediction of protein secondary structure by use of sequence profiles and neural networks. Proc. Natl. Acad. Sci. USA 90:7558-7562.
Ruger, B., S. Klages, B. Walla, J. Albrecht, B. Fleckenstein, P. Tomlinson, and B. Barrell. 1987. Primary structure and transcription of the genes coding for the two virion phosphoproteins pp65 and pp71 of human cytomegalovirus. J. Virol. 61:446-453.
Sarisky, R. T., and G. S. Hayward. 1996. Evidence that the UL84 gene product of human cytomegalovirus is essential for promoting oriLyt-dependent DNA replication and formation of replication compartments in cotransfection assays. J. Virol. 70:7398-7413.
Schierling, K., T. Stamminger, T. Mertens, and M. Winkler. 2004. Human cytomegalovirus tegument proteins ppUL82 (pp71) and ppUL35 interact and cooperatively activate the major immediate-early enhancer. J. Virol. 78:9512-9523.
Schleiss, M. R., A. McGregor, N. J. Jensen, G. Erdem, and L. Aktan. 1999. Molecular characterization of the guinea pig cytomegalovirus UL83 (pp65) protein homolog. Virus Genes 19:205-221.
Schmolke, S., H. F. Kern, P. Drescher, G. Jahn, and B. Plachter. 1995. The dominant phosphoprotein pp65 (UL83) of human cytomegalovirus is dispensable for growth in cell culture. J. Virol. 69:5959-5968.
Shackelton, L. A., and E. C. Holmes. 2004. The evolution of large DNA viruses: combining genomic information of viruses and their hosts. Trends Microbiol. 12:458-465.
Silverman, E., G. Edwalds-Gilbert, and R.-J. Lin. 2003. DExD/H proteins and their partners: helping RNA helicases unwind. Gene 312:1-16.
Sommer, P., E. Kremmer, S. Bier, S. Konig, P. Zalud, M. Zeppezauer, J. F. Jones, N. Mueller-Lantzsch, and F. A. Grasser. 1996. Cloning and expression of the Epstein-Barr virus-encoded dUTPase: patients with acute, reactivated or chronic virus infection develop antibodies against the enzyme. J. Gen. Virol. 77:2795-2805.
Song, M. J., S. Hwang, W. H. Wong, T.-T. Wu, S. Lee, H.-I. Liao, and R. Sun. 2005. Identification of viral genes essential for replication of murine -herpesvirus 68 using signature-tagged mutagenesis. Proc. Natl. Acad. Sci. USA 102:3805-3810.
Spector, D. J., and M. J. Tevethia. 1994. Protein-protein interactions between human cytomegalovirus IE2-580aa and pUL84 in lytically infected cells. J. Virol. 68:7549-7553.
Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673-4680.
Tuteja, N., and R. Tuteja. 2004. Unraveling DNA helicases: motif, structure, mechanism and function. Eur. J. Biochem. 271:1849-1863.
Varnum, S. M., D. N. Streblow, M. E. Monroe, P. Smith, K. J. Auberry, L. Pasa-Tolic, D. Wang, D. G. Camp II, K. Rodland, S. Wiley, W. Britt, T. Shenk, R. D. Smith, and J. A. Nelson. 2004. Identification of proteins in human cytomegalovirus (HCMV) particles: the HCMV proteome. J. Virol. 78:10960-10966.
Weiss, R. S., S. S. Lee, B. V. Prasad, and R. T. Javier. 1997. Human adenovirus early region 4 open reading frame 1 genes encode growth-transforming proteins that may be distantly related to dUTP pyrophosphatase enzymes. J. Virol. 71:1857-1870.
Wills, M. R., A. J. Carmichael, K. Mynard, X. Jin, M. P. Weekes, B. Plachter, and J. G. P. Sissons. 1996. The human cytotoxic T-lymphocyte (CTL) response to cytomegalovirus is dominated by structural protein pp65: frequency, specificity, and T-cell receptor usage of pp65-specific CTL. J. Virol. 70:7569-7579.
Xu, Y., K. S. Colletti, and G. S. Pari. 2002. Human cytomegalovirus UL84 localizes to the cell nucleus via a nuclear localization signal and is a component of viral replication compartments. J. Virol. 76:8931-8938.
Xu, Y., S. A. Cei, A. Rodriguez Huete, and G. S. Pari. 2004a. Human cytomegalovirus UL84 insertion mutant defective for viral DNA synthesis and growth. J. Virol. 78:10360-10369.
Xu, Y., S. A. Cei, A. Rodriguez Huete, K. S. Colletti, and G. S. Pari. 2004b. Human cytomegalovirus DNA replication requires transcriptional activation via an IE2- and UL84-responsive bidirectional promoter element within oriLyt. J. Virol. 78:11664-11677.
Yu, D., M. C. Silva, and T. Shenk. 2003. Functional map of human cytomegalovirus AD169 defined by global mutational analysis. Proc. Natl. Acad. Sci. USA 100:12396-12401.
Zhu, F. X., J. M. Chong, L. Wu, and Y. Yuan. 2005. Virion proteins of Kaposi's sarcoma-associated herpesvirus. J. Virol. 79:800-811.(Andrew J. Davison and Nig)
ABSTRACT
Published work (D. J. McGeoch, Nucleic Acids Res. 18:4105-4110, 1990; J. E. McGeehan, N. W. Depledge, and D. J. McGeoch, Curr. Protein Peptide Sci. 2:325-333, 2001) has indicated that evolution of dUTPase in the class of herpesviruses that infect mammals and birds involved capture of a host gene followed by a duplication event that resulted in a coding region comprising two fused dUTPase domains. Some of the conserved residues required for enzyme activity were then lost, resulting in a dUTPase containing a single active site with different elements contributed by each half of the protein. Further conserved residues were lost in one subfamily (the Betaherpesvirinae), yielding a protein that is related to herpesvirus dUTPases but has a different and as yet unrecognized function. Evidence from sequence similarities and structural predictions now indicates that several additional genes were derived from the herpesvirus dUTPase gene, probably by duplication. These are UL31, UL82, UL83, and UL84 in human cytomegalovirus (and counterparts in other members of the Betaherpesvirinae) and ORF10 and ORF11 in human herpesvirus 8 (and counterparts in other members of the Gammaherpesvirinae). The findings clarify the evolutionary history of these genes and provide novel insights for structural and functional studies.
INTRODUCTION
One means of generating proteins with new functions that has been used widely in evolution is that of gene duplication (70). This strategy allows the original function to be retained by one of the gene copies while permitting the development of a distinct function in the other. The occurrence of families of related genes indicates that gene duplication has also played a prominent role in the evolution of large DNA viruses (81). Repeated duplication has often resulted in families consisting of tandem head-to-tail gene arrays, but related genes may also be spread throughout the genome, presumably as a result of relocation during or after duplication. Moreover, viral gene families may be similar to host genes, suggesting that the parental gene was derived originally by capture from the cellular repertoire (20, 81).
The gene encoding 2'-deoxyuridine 5'-triphosphate pyrophosphatase (dUTPase; EC 3.6.1.23) is ubiquitous in all classes of organisms and has been incorporated into nonprimate lentiviruses and many families of large DNA viruses (5, 52, 54). This enzyme catalyzes the hydrolysis of dUTP and, in addition to supplying dUMP for synthesis of TMP by thymidylate synthase, reduces the dUTP pool and thus minimizes misincorporation of uracil residues into newly synthesized DNA. The dUTPases of Escherichia coli and Homo sapiens are identical in only 31% of their residues but exhibit similar crystallographic structures (12, 30, 61).
Figure 1a shows the amino acid sequence of the human protein and highlights regions of secondary structure and the five conserved sequence motifs identified by McGeoch (54). Mol et al. (61) described the domain fold of the human dUTPase monomer as a distorted eight-stranded ?-barrel consisting of a six-stranded antiparallel ?-barrel formed from ?2, ?7, ?4, ?5, ?6, and ?3, supplemented by the parallel strands ?1 and ?8', the latter contributed by an adjacent subunit (Fig. 1b). This structure is capped by a conical arrangement of five antiparallel ?-strands extending from the barrel (?2b, ?6b, ?5, ?6, and ?3), plus an -helix (1). The active enzyme is a trimer, each catalytic site involving residues from all three subunits (Fig. 1c and d). Thus, residues from one subunit primarily recognize uracil and deoxyribose (motif 3), residues from a second subunit recognize the phosphate groups (motifs 2 and 4), and residues from the C-terminal region of a third subunit close the substrate-binding pocket (motif 5). Specificity for deoxyuridine is provided by hydrogen bonding with the ?5-?6 hairpin and the exclusion of ribose by a tyrosine residue in the ?5-?6 turn.
Members of the family Herpesviridae fall into three very widely divergent phylogenetic classes: those infecting mammals, birds, or reptiles (henceforth called the mammalian herpesvirus class), those infecting fish or amphibians (henceforth called the fish herpesvirus class), and a single known virus infecting bivalves (20). Each class possesses a dUTPase gene that appears to have been captured from the host, perhaps on independent occasions. In the fish herpesvirus class, the dUTPases are similar in size to the cellular protein, have the same arrangement of motifs, and presumably are equivalent structurally and functionally to the cellular enzyme (18, 19). The bivalve virus also has a dUTPase that is similar to the cellular enzyme and in addition possesses two related genes of roughly the same size, which are located in other regions of the genome and are likely to have arisen by duplication. These lack active-site residues and probably fulfill other functions (21).
The mammalian herpesvirus class is divided into three phylogenetically distinct subfamilies, the Alpha-, Beta-, and Gammaherpesvirinae. In these viruses, the dUTPases are related to the cellular enzyme but are structurally distinct as the result of an unusual evolutionary history. The most intensively studied mammalian herpesvirus, herpes simplex virus type 1 (HSV-1), belongs to the Alphaherpesvirinae and has long been known to induce a novel dUTPase activity in infected cells (11). The dUTPase genes of other members of the Alphaherpesvirinae and also of the Gammaherpesvirinae have been shown to specify active enzymes (27, 45, 69, 83). The dUTPases of HSV-1 and murid herpesvirus 4 (MuHV-4; a member of the Gammaherpesvirinae) are not required for growth of the virus in cell culture but provide advantages in vivo (25, 71, 84).
In a comparative study, McGeoch (54) observed that the dUTPases of two members of the Alphaherpesvirinae (HSV-1 and varicella-zoster virus) and one member of the Gammaherpesvirinae (Epstein-Barr virus) are about twice the length of their cellular counterparts. Moreover, the conserved regions are arranged differently, with motifs 1, 2, 4, and 5 present in their usual order in the C-terminal half of the protein, and motif 3 in the N-terminal half (compare Fig. 2a and b). These observations prompted the conclusion that the dUTPase gene in an ancestral herpesvirus had been duplicated internally, giving rise initially to a protein containing two complete sets of motifs. Motifs 1, 2, 4, and 5 had then been lost from the N-terminal half and motif 3 from the C-terminal half. The herpesvirus enzyme was therefore envisaged as a monomer, as detected previously (10), with elements of the single active site contributed by each half of the protein. These insights into the contributions made to the active site by the conserved regions turned out to be entirely consistent with those defined subsequently in the trimeric cellular dUTPases.
Aided by the structure of Escherichia coli dUTPase (12) and additional sequence data, McGeehan et al. (53) concluded from secondary-structure predictions that the ?-strand arrangement is conserved in both halves of dUTPases from the Alpha- and Gammaherpesvirinae. This supported a molecular model comprising two structurally similar domains, in which motif 3 from the N-terminal domain is juxtaposed to motifs 1, 2, and 4 from the C-terminal domain, with motif 5 from the C terminus positioned at the active site by virtue of an extended region connecting it to the rest of the protein (Fig. 2c). The conserved region that occupies the position of motif 3 in the C-terminal domain has been termed motif 6, and it is possible that this herpesvirus-specific element contributes a novel function.
In contrast to the Alpha- and Gammaherpesvirinae, the dUTPase orthologs in the Betaherpesvirinae were found to lack motifs 1 to 5 but to have retained motif 6. This prompted the idea that these proteins may no longer function as dUTPases, but instead have other roles, perhaps requiring motif 6 (53, 55). This has recently found experimental support from the observation that the human cytomegalovirus (HCMV) protein lacks dUTPase activity (9). Like active dUTPases in the Alpha- and Gammaherpesvirinae, this protein is not required for growth in cell culture, and its function remains unknown (9, 24, 94).
Alternative functions thus appear to have been generated from the herpesvirus dUTPase in both the bivalve virus (involving gene duplications that permitted retention of the original function and the development of new ones) and in the subfamily Betaherpesvirinae of the mammalian herpesvirus class (involving replacement of the original function by a new one). We reasoned that additional herpesvirus functions could have been derived from dUTPase via gene duplication, but that these might be hidden by extensive divergence. Indeed, we were provoked by McGeehan et al.'s (53) mentioning marginal relationships between dUTPase and the proteins encoded by ORF10 and ORF11 in the Gammaherpesvirinae, although these authors assessed the significance as dubious and took the analysis no further. Having completed a comprehensive analysis of the mammalian herpesvirus class, for which sequence data are abundant, we conclude that several genes in the Beta- and Gammaherpesvirinae were indeed derived from the herpesvirus dUTPase. Our findings clarify the evolutionary history of these genes and provide novel leads for structural studies.
MATERIALS AND METHODS
Sources of sequence data. Sequences were derived from public databases or by personal communication, as detailed in Fig. 4. The tools available at http://www.ncbi.nlm.nih.gov and http://www.ebi.uniprot.org/uniprot-srv/uniProtPowerSearch.do were utilized for this purpose.
Computer programs. Searches for amino acid sequence similarity using BLASTP (3) and PSI-BLAST (2) were carried out with default parameters at http://www.ncbi.nlm.nih.gov. Amino acid sequences were aligned using CLUSTAL W (86) implemented locally, and in some instances the alignments were adjusted manually. Alignments were displayed using CHROMA (31) with default parameters. Secondary structures were predicted from amino acid sequences using PROFsec (74, 75; http://www.predictprotein.org), which provides a facility for processing precomputed alignments. SUPERFAMILY 1.65 was utilized for protein threading (33, 51; http://supfam.mrc-lmb.cam.ac.uk/SUPERFAMILY). This program relies on the SCOP database of all known protein folds (4; http://scop.mrc-lmb.cam.ac.uk/scop). PROFsec and SUPERFAMILY were used on-line with default parameters. Neighbor-joining phylogenetic trees were constructed using MEGA (46; http://www.megasoftware.net) with the PAM (Dayhoff) matrix model, removing gapped regions and assuming uniform substitution rates among sites. Bootstrap values were calculated from 100 replicates.
Gene nomenclature. Gene nomenclature varies from one herpesvirus to another, and as a result orthologous relationships are frequently not obvious. For clarity, HSV-1, HCMV, and human herpesvirus 8 (HHV-8) names were applied throughout for orthologs in the Alpha-, Beta-, and Gammaherpesvirinae, respectively, with alternative names given only in particular instances. Virus-specific names may be obtained via the accession numbers listed in Fig. 4.
The nomenclature problem is compounded for a family of tandemly arranged, related genes (UL82, UL83, and UL84) located in the middle of the HCMV genome and their counterparts in other members of the Betaherpesvirinae (14, 17). Depending on the virus, the number of genes in this family ranges from two to four, and the evolutionary history of the more distantly related members relative to one another is difficult to determine with confidence from phylogenetic analysis (see Table 1 and the Discussion). This confounds application of a common nomenclature that accurately represents orthology. In this study, the leftmost gene in the family was denoted as UL82 for every virus and the rightmost as UL84 where it is clearly orthologous to HCMV UL84. Any other genes were denoted UL83 (where there is a single gene) or UL83A and UL83B (where there are two genes).
RESULTS AND DISCUSSION
Pairwise-based sequence comparisons. Pairwise BLASTP and PSI-BLAST searches were conducted using various herpesvirus dUTPases as probes and the GenBank nonredundant set of viral proteins as the data set. Reciprocal searches were then carried out using proteins that emerged as potentially related to dUTPase. These investigations showed that dUTPases in the Alpha- and Gammaherpesvirinae (encoded by UL50 and ORF54, respectively) are related to the orthologous UL72 protein of the Betaherpesvirinae within a domain of approximately 100 amino acid residues in the C-terminal half of each protein. In addition, more distant relationships were revealed in the same domain to the proteins encoded by HCMV UL31, UL82, UL83, and UL84 (and their orthologs in other members of the Betaherpesvirinae) and HHV-8 ORF10 and ORF11 (and their orthologs in other members of the Gammaherpesvirinae). The locations of these genes in the respective genomes are illustrated in Fig. 3. The conserved dUTPase-related domain (DURD) corresponds closely to the region identified previously as the most highly conserved between UL82, UL83, and UL84, although the relationship of these genes to dUTPases was not recognized at the time (17).
Further PSI-BLAST searches were then conducted using representative DURDs in order to detect all dUTPase-related proteins (DURPs) encoded by the mammalian herpesvirus class. A comprehensive alignment of the DURD in these proteins is shown in Fig. 4. The DURD corresponds to approximately 100 residues containing herpesvirus dUTPase motifs 1, 2, 6, and 4. Conservation of motif 6, with its characteristic tryptophan residue, rather than the positionally equivalent motif 3 of nonherpesvirus dUTPases, with its tyrosine residue, indicates that the DURD was propagated from the herpesvirus dUTPase rather than from an external source. In every case, the DURD is located in the C-terminal half of DURPs (Fig. 5, green blocks).
Secondary structure analysis. Secondary structure predictions were carried out for the various groups of DURDs in order to assess their similarity to human dUTPase. All sequences in each group were analyzed precisely as aligned in Fig. 4, with the BLAST facility in PROFsec disabled so that no other sequences were incorporated. The results (Fig. 4, blue blocks) show that the predicted secondary structure of the DURD is conserved among the groups of DURPs and is strikingly similar to that of human dUTPase. The reliability of the predictions is indicated by the close correspondence of the known human dUTPase structure with the prediction (obtained by alignment with other dUTPase sequences detected using BLAST as implemented in PROFsec). ?-strands corresponding to ?2 to ?7 were predicted for each herpesvirus group, except for ?5a and ?5b in the UL82/UL83/UL84 group and ?5b in the UL50/ORF54 and "all herpes" groups. No -helix was predicted for any of the proteins, including human dUTPase. Taking into account the limitations of secondary structure prediction, each DURP is anticipated to contain in its C-terminal half the six-stranded antiparallel ?-barrel characteristic of dUTPases.
To investigate the status of ?1 and ?8, which contribute to the eight-stranded ?-barrel in cellular dUTPases and are located partly or wholly outside the DURD, secondary structures were predicted for entire proteins (Fig. 5, blue and red blocks). In contrast to predictions for the DURDs, which were based on the alignments exactly as shown in Fig. 4, other regions of the proteins are generally too divergent to permit the construction of robust alignments incorporating all members of a particular group. Consequently, predictions were carried out by allowing alignments to be generated within PROFsec and using a cutoff facility to remove proteins (in part or in entirety) that are either too closely or too distantly related to the probe sequence. This ensured that predictions were based on sound representative alignments, but had the disadvantage that they were restricted to a subset of sequences in any protein group, and to different numbers of sequences in different parts of the protein. The reliability of the predictions therefore varied from protein to protein, and from one part of a protein to another. The results (Fig. 5) indicated that the N-terminal part of each protein is rich in ?-strand, but were not revealing about the presence of ?1 and ?8. The most that could be inferred was that a ?-strand in each protein is positioned close to the N-terminal end of the DURD and may correspond to ?1. More equivocally, each protein contains one or more ?-strands between the DURD and the C terminus that may correspond to ?8.
McGeehan et al. (53) noted from secondary structure predictions that an arrangement of ?-strands similar to that in the C-terminal half of herpesvirus dUTPases of the Alpha- and Gammaherpesvirinae is present in the N-terminal half, implying the presence of two ?-barrels in the protein. However, the N-terminal portions of herpesvirus dUTPases are more divergent than the C-terminal portions, and any N-terminal ?-barrel is not obvious from amino acid sequence comparisons with the C-terminal one. Nonetheless, the likelihood that dUTPases contain two DURDs raises the question of whether the same applies to other DURPs.
Profile-based sequence comparisons. Protein threading programs, which attempt to fit query sequences to models derived from known protein structures, are showing increasing promise as a way of identifying folds in such sequences, even in the absence of detectable sequence conservation. SUPERFAMILY is a threading program that offers a number of advantages. It searches a comprehensive collection of protein structures, identifies multiple folds within the same protein, and distinguishes between significant and marginal predictions among the output scores. Having analyzed a wide range of herpesvirus proteins using SUPERFAMILY, our evaluation was that this program is sufficiently conservative to inspire a substantial degree of confidence.
Nevertheless, given that it can be difficult to discriminate between meaningful and spurious findings among the output from threading programs, our analysis was undertaken cautiously, with the aim of confirming the case for a ?-barrel in the C-terminal part of DURPs and of examining the N-terminal part for the presence of a second ?-barrel. The full-length versions of all the DURPs represented in Fig. 4 were submitted to SUPERFAMILY, and the presence of dUTPase-like domains (?-barrels) was recorded (Table 2 and Fig. 5, orange blocks). As expected, a ?-barrel was scored in the C-terminal part of every dUTPase (UL50 and ORF54) from the Alpha- and Gammaherpesvirinae, and in the C-terminal part of every UL72 protein from the Betaherpesvirinae. Confident predictions of a domain in the N-terminal part were made for most dUTPases, and marginal predictions for the others, supporting the conclusion of McGeehan et al. (53) from secondary-structure predictions that herpesvirus dUTPases consist of two similar domains.
SUPERFAMILY also predicted that the N-terminal part of most UL72 proteins contains a ?-barrel. Similarly, the C-terminal part of most other DURPs (UL82, UL83, UL84, and UL31 in the Betaherpesvirinae and ORF10 and ORF11 in the Gammaherpesvirinae) was predicted to contain a ?-barrel. However, the situation with the N-terminal parts of these proteins was much less clear, as only a minority were predicted to contain a ?-barrel. Unusually, predictions were made with greater confidence for the N-terminal part of the UL82 proteins of human herpesvirus 6 (HHV-6) and human herpesvirus 7 (HHV-7) than for the C-terminal part.
Bearing in mind the limitations of threading programs, we predict that the C-terminal part of DURPs contains a ?-barrel, and that the N-terminal part of herpesvirus dUTPases and their UL72 orthologs in the Betaherpesvirinae contains a second ?-barrel. Overall, the C-terminal ?-barrel is better conserved than the N-terminal one, indicating that the latter is more responsive to evolutionary pressures. There are indications that a ?-barrel is also present in the N-terminal part of some other DURPs (from Table 2, specifically UL82, UL83B, ORF10, and ORF11).
An analysis of HCMV proteins was performed previously by Novotny et al. using the threading program ProCeryon (66). These authors listed the UL82, UL84, and UL31 proteins as containing ?-barrel domains, although the coordinates given for the latter two assignments are C-terminal to the DURD, and the UL82 prediction was considered marginal. The UL84 prediction included the sequence of the DURD and was listed as being of medium confidence. In none of these cases was a dUTPase fold described as being the best match to the HCMV protein.
Evolution of DURPs. Our identification of DURPs depends only upon sequence and structure relationships and is independent of phylogenetic analysis. Indeed, phylogenetic analysis of the DURD sequences was not particularly robust since their modest length and wide divergence resulted in substantial sensitivity to the parameters used and generally low bootstrap support. The example tree shown in Fig. 6 groups positionally equivalent genes together. However, in the deeper branches only the bootstrap values that establish the UL72 (79 of 100), UL31 (78 of 100) and ORF11 groups (81 of 100) might be considered convincing. Other inferences from this particular tree, such as the UL84 group seeming more closely related to the UL31 group than to UL82/UL83 proteins, are poorly supported (29 of 100). Members of the UL82/UL83/UL84 family of the Betaherpesvirinae grouped generally with each other, consistent with the proposition that they have arisen by duplications of an ancestral gene encoding the "matrix protein ancestor" (17).
Nonphylogenetic considerations also contribute to understanding the evolution of the DURPs. The substantial evidence for a DURD in the C-terminal part and the weaker evidence for a second DURD in the N-terminal part (Fig. 5 and Table 2) suggests that DURPs arose as a result of whole-gene duplication events rather than transfer of the domain to existing genes. However, it is difficult to determine the order of events involved. Given their tandem arrangement, one possibility is that ORF10 and ORF11 may have arisen from local duplication of an ancestral gene derived earlier from the active dUTPase gene (ORF54) in the Gammaherpesvirinae. In contrast, the ancestor of the UL31, UL82, UL83, and UL84 proteins of the Betaherpesvirinae may already have lost dUTPase function by making the transition to UL72.
Perhaps more likely, however, is that duplication and loss of function prior to divergence of the Beta- and Gammaherpesvirinae gave rise to a common ancestor of ORF10, ORF11, UL31, UL82, UL83, and UL84. This is supported by the observation that the ORF10/ORF11 and UL82/UL83/UL84 families are positionally equivalent with respect to the gene encoding the protease (ORF17 and UL80, respectively; denoted PR in Fig. 3, shaded green). The protease gene is conserved in the Alpha-, Beta-, and Gammaherpesvirinae and is located near the end of block IV, one of the blocks of genes that are arranged differently in the three subfamilies (60). The protease gene is adjacent to UL82 in the Betaherpesvirinae and separated from ORF11 in the Gammaherpesvirinae by a variable number of recently captured cellular genes; this number happens to be rather large in HHV-8 but is only one or two in many members of the Gammaherpesvirinae. UL31 of the Betaherpesvirinae probably represents an additional relocated copy derived from the matrix protein ancestor gene.
Roles of the DURD. The above analyses indicate the existence of an extensive family of herpesvirus proteins that have arisen from an ancestral viral dUTPase and comprises several distinct sets of orthologous proteins retaining a common structural domain. How might this be reflected in the functions of these proteins?
The presence of the DURD could reflect the original biochemical function or the fact that it provides an amenable architectural element. In contrast to the functional dUTPases encoded by the Alpha- and Gammaherpesvirinae, none of the other DURPs retains the residues required for dUTPase activity. Nevertheless, parts of the protein fold involved in sugar and phosphate binding are predicted to be present, and it is possible that all these proteins exploit such a property. The dUTPase superfamily is one of several superfamilies that share a structure consisting of six ?-strands termed the ?-clip fold (38). This fold is thought to have been derived by duplication of a three-strand unit and includes the dUTPase ?-barrel containing ?2 to ?7. Many ?-clip proteins are involved in binding to carbohydrate residues (38, 54), and in dUTPases this activity centers on motif 3, containing ?5 and ?6. However, despite these observations, no evidence has been reported for other DURPs binding to carbohydrates.
Herpesvirus-specific motif 6 occupies the position of motif 3 in the cellular dUTPase monomer and is retained not only in the UL72 proteins of the Betaherpesvirinae, which are orthologs of the functional dUTPases, but throughout the whole family of DURPs (Fig. 4). This motif contains the most highly conserved residue in the DURD (a tryptophan in the ?5-?6 loop). Inspection of the human dUTPase structure (Fig. 1) indicates that motif 6 is likely to be located on the opposite face of the molecule from the dUTP-binding site. Any function conferred by this motif has not been identified, and detailed site-directed mutagenesis may prove informative.
The structure of the human enzyme shows that the two complementary faces in a DURD could facilitate intermolecular interactions. Thus, DURDs may enable proteins to assemble into dimers or larger structures. In this regard, it is interesting that several DURPs are incorporated into the tegument (the layer of the virus particle between the capsid and envelope), some in relatively large quantities. Furthermore, DURPs might have evolved to enable interactions with other proteins. Unfortunately, the area of research is insufficiently developed to indicate whether DURDs are important for any of the interactions of DURPs documented to date (16, 35, 41, 78, 85).
An interesting parallel to the herpesvirus DURPs is available among the adenoviruses. Human adenovirus E4 ORF1 encodes a catalytically inactive derivative of the cellular dUTPase monomer, whose ancestor was presumably acquired from the cell (22, 38, 89). Indeed, in various nonhuman adenoviruses this protein has retained the catalytic motifs of an active enzyme. Figure 7 shows that the full-length human adenovirus E4 ORF1 protein corresponds to a C-terminally truncated dUTPase monomer, with all eight ?-strands contributing to a predicted ?-barrel. It appears that the functions of the E4 ORF1 protein in transformation (26, 47) must be attributed directly to the DURD, since the protein contains little or nothing by way of additional elements. These observations point to some general utility of the dUTPase ?-barrel (as a specific type of ?-clip fold) that has proved valuable during evolution of herpesviruses and adenoviruses.
The finding that herpesvirus DURPs exhibit a well-characterized protein fold contributes insights relevant to their study. For example, Hofmann et al. (35) identified a motif in the HCMV UL82 protein (residues 324 to 331) that is similar to a sequence known to be important in other proteins for interaction with the human Daxx protein (hDaxx). Although deletion of the 8-residue sequence abolished interaction of UL82 with hDaxx, substitutions within it had no effect. The authors speculated that the deletion might have had a major effect on the three dimensional structure of the protein. Our analysis indicates that this is likely to be the case, since the region deleted corresponds closely to ?5. Similarly, it was observed that mutants with C-terminal truncations to residue 508 of HCMV UL84 retained the ability to localize to the nucleus, whereas those truncated at residue 488 remained cytoplasmic (48), even though the intervening sequence is not part of a functional nuclear localization signal (48, 91). Again, it seems probable that removal of these 20 residues, which correspond to ?6b and ?7, would result in major structural changes to the protein.
Functions of DURPs. To date, only a small number of DURPs have been studied in any detail, and it is already apparent that they perform various functions. It is important to emphasize that DURPs exhibit a high degree of divergence, and it is probable that certain protein functions are not conserved even between orthologous proteins, as is the situation with the UL50 and ORF54 dUTPases in the Alpha- and Gammaherpesvirinae in comparison with their UL72 orthologs in the Betaherpesvirinae. Thus, functions inferred from sequence comparisons must be viewed with caution. Moreover, it should be appreciated that the DURD is relatively small, and even the presence of two copies would account in most instances for only a minority of the DURP (Fig. 5). Key aspects of function are therefore probably specified by other regions, which may have evolved by independent gene expansion events, including the acquisition of other protein domains by recombination. Thus, DURPs may perform a wide range of unrelated roles while retaining a common function provided by the DURD.
HCMV UL82 encodes the tegument phosphoprotein pp71. This protein functions as a transcriptional transactivator and is essential for virus growth at low multiplicity of infection in cell culture (7, 37, 49, 76). Similar properties have been reported for the guinea pig cytomegalovirus (GPCMV) UL82 protein (57). HCMV pp71 interacts with cellular proteins hDaxx and Rb and induces cell cycle progression (35, 40, 42). An interaction with another tegument protein, ppUL35, has been reported, and appears to be important for activation of the major immediate early promoter (78).
The HCMV UL83 protein is also a phosphorylated tegument protein (pp65) and the major constituent of virions and noninfectious dense bodies (76, 88). In contrast to pp71, pp65 is not essential for growth in cell culture (80). Similar properties have been reported for the UL83 proteins of GPCMV and murine cytomegalovirus (MCMV) (43, 57, 63, 79). HCMV pp65 dampens the interferon response to infection, but conflicting mechanisms for this effect on the host have been presented (1, 8). It also represents an immunodominant target for host cytotoxic T-cell responses (58, 90). In contrast, the corresponding MCMV M83 protein is not a dominant factor in the T-cell response (64), but M83-specific T cells may nevertheless be important in resolving acute infection (36).
The HCMV UL84 protein has been reported to be present in virions, but appears to be a minor constituent (88). However, it is essential for growth in cell culture (24, 92, 94). The protein is capable of self-oligomerization, and interacts with and modulates the activity of the transcriptional transactivator IE2 (16, 28, 85). Transient DNA replication assays and characterization of a null mutant virus have indicated that the UL84 protein probably performs an essential role during initiation of HCMV DNA synthesis (67, 77, 92, 93). Surprisingly, however, the protein was found to be dispensable for HCMV origin-dependent DNA synthesis when the HSV-1 DNA replication fork proteins were employed in a transient assay (72).
Colletti et al. recently reported that HCMV UL84 encodes a nucleoside triphosphatase (NTPase) activity with a preference for UTP as the substrate (15). It should be emphasized that UTPase activity (which converts UTP to UDP plus phosphate) is enzymatically unrelated to that of dUTPase (which converts dUTP to dUMP plus pyrophosphate). These authors also noted the presence of several sequence motifs within the UL84 protein and suggested it belongs to the DExD/H box family, many members of which are known to be NTP-utilizing helicases (32, 34, 82, 87). Although matches were found for DExD/H box motifs I, Ia, II, IV, and V, there appeared to be no identifiable counterparts of the two other conserved motifs, III and VI (15). The spacing of the identified motifs was also atypical for DExD/H box proteins, with motifs I and Ia separated by an unusually large distance and motifs II and IV by only 6 residues. In addition, our amino acid sequence alignment for HCMV UL84 and the homologous proteins of other primate cytomegaloviruses revealed that proposed motifs I and II were not conserved, but rather were located in regions of the HCMV protein characterized by significant insertions or deletions in the homologous proteins (data not shown). Interestingly, the suggested matches to motifs Ia, II, and part of motif IV were located within the region of the UL84 protein corresponding to the DURD.
The structures of several DExD/H box proteins are known and the regions containing the conserved motifs exhibit a high degree of conservation in their folding pattern, which is quite distinct from that of the DURD (73, 34, 87). When authentic DExD/H box proteins were submitted to SUPERFAMILY, they were identified as containing the P-loop nucleoside triphosphate hydrolase fold in the region corresponding to motifs I, Ia, II, and III, whereas this fold was not predicted for HCMV UL84 (data not shown). Taken together, these observations raise serious doubt as to whether the HCMV UL84 protein is a member of the DExD/H box family and suggest that site-directed mutagenesis experiments should be performed to confirm the assignment of a UTPase activity to this protein.
MCMV M84 has sometimes been viewed as the homolog of HCMV UL84. However, M84 is distantly related to UL84 (17), and, indeed, is more closely related in the DURD to the UL82/UL83 proteins (Fig. 6). The properties of the M84 and UL84 proteins are also distinctly different. The M84 protein is not essential for replication in cell culture (63) and is therefore unlikely to play a key role in DNA synthesis. Rather, it appears to function in the T-cell response (36, 64). However, despite its closer relationship to the UL82/UL83 proteins, M84 was not detected in virions (43). Given the high degree of conservation between other HCMV and MCMV DNA replication proteins, it is surprising that MCMV encodes no obvious homolog of UL84. This could reflect functional differences between MCMV and HCMV, for example, in the origin of DNA replication. Nonetheless, several interesting questions remain regarding the functions of the DURPs in this region of the genome of the nonprimate cytomegaloviruses, only some of which appear to have sequence counterparts of UL84, and the identities of the DNA synthesis initiator proteins for these viruses. Unfortunately, insufficient data are available to facilitate functional comparisons with the corresponding HHV-6 and HHV-7 genes.
Very little information is available on the DURPs other than those in the UL82/UL83/UL84 group. HCMV UL31 is not essential for growth in cell culture (24, 94) and, although the HCMV protein was not detected in virions, its MCMV ortholog was (43, 88). As far as expression kinetics are concerned, the DURPs that have been examined are early or late lytic cycle genes, compatible with their proposed connections to nucleotide metabolism, DNA synthesis, gene regulation, or virion structure (13, 17, 39, 44, 50, 79). HHV-7 U10 (the UL31 ortholog) is unusual in being transcribed under immediate-early conditions (59). EBV LF1 and LF2 (the ORF10 and ORF11 orthologs) are not required for growth in cell culture, since they lie in a region absent from the B95-8 strain (68). MuHV-4 ORF10 and ORF11 are also not essential, and the latter encodes a virion component (6, 62). The HHV-8 ORF11 protein has been reported to be present in virions (95).
In conclusion, despite our presently limited knowledge of the roles of members of the DURP family, it is clear that these proteins exhibit wide differences with regard to their functions, presence in the virion, kinetics of expression, requirement for growth in cell culture, immunological properties, and effects on the host cell. The discovery that they have evolved from an ancestral herpesvirus gene and share predicted structural properties should serve to focus further functional characterization of this large family of related proteins.
ACKNOWLEDGMENTS
This work was supported by the United Kingdom Medical Research Council.
We are grateful to Aidan Dolan, Mark Schleiss, and James Stewart for generously providing unpublished sequence data for SCMV, GPCMV, and OvHV-2, respectively. We thank Edgar Sevilla-Reyes, Chris Preston, and Duncan McGeoch for comments on the manuscript.
REFERENCES
Abate, D. A., S. Watanabe, and E. S. Mocarski. 2004. Major human cytomegalovirus structural protein pp65 (ppUL83) prevents interferon response factor 3 activation in the interferon response. J. Virol. 78:10995-11006.
Altschul, S. F., T. L. Madden, A. A. Sch?ffer, J. Zhang, Z. Zhang, W. Miller, and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25:3389-3402.
Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. Basic local alignment search tool. J. Mol. Biol. 215:403-410.
Andreeva, A., D. Howorth, S. E. Brenner, T. J. Hubbard, C. Chothia, and A. G. Murzin. 2004. SCOP database in 2004: refinements integrate structure and sequence family data. Nucleic Acids Res. 32:D226-229.
Baldo, A. M., and M. A. McClure. 1999. Evolution and horizontal transfer of dUTPase-encoding genes in viruses and their hosts. J. Virol. 73:7710-7721.
Boname, J. M., J. S. May, and P. G. Stevenson. 2005. Murine gammaherpesvirus 68 open reading frame 11 encodes a nonessential virion component. J. Virol. 79:3163-3168.
Bresnahan, W. A., and T. E. Shenk. 2000. UL82 virion protein activates expression of immediate early viral genes in human cytomegalovirus-infected cells. Proc. Natl. Acad. Sci. USA 97:14506-14511.
Browne, E. P., and T. Shenk. 2003. Human cytomegalovirus UL83-coded pp65 virion protein inhibits antiviral gene expression in infected cells. Proc. Natl. Acad. Sci. USA 100:11439-11444.
Caposio, P., L. Riera, G. Hahn, S. Landolfo, and G. Gribaudo. 2004. Evidence that the human cytomegalovirus 46-kDa UL72 protein is not an active dUTPase but a late protein dispensable for replication in fibroblasts. Virology 325:264-276.
Caradonna, S. J., and D. M. Adamkiewicz. 1984. Purification and properties of the deoxyuridine triphosphate nucleotidohydrolase enzyme derived from HeLa S3 cells. Comparison to a distinct dUTP nucleotidohydrolase induced in herpes simplex virus-infected HeLa S3 cells. J. Biol. Chem. 259:5459-5464.
Caradonna, S. J., and Y. C. Cheng. 1981. Induction of uracil-DNA glycosylase and dUTP nucleotidohydrolase activity in herpes simplex virus-infected human cells. J. Biol. Chem. 256:9834-9837.
Cedergren-Zeppezauer, E. S., G. Larsson, P. O. Nyman, Z. Dauter, and K. S. Wilson. 1992. Crystal structure of a dUTPase. Nature 355:740-743.
Chambers, J., A. Angulo, D. Amaratunga, H. Guo, Y. Jiang, J. S. Wan, A. Bittner, K. Frueh, M. R. Jackson, P. A. Peterson, M. G. Erlander, and P. Ghazal. 1999. DNA microarrays of the complex human cytomegalovirus genome: profiling kinetic class with drug sensitivity of viral gene expression. J. Virol. 73:5757-5766.
Chee, M. S., A. T. Bankier, S. Beck, R. Bohni, C. M. Brown, R. Cerny, T. Horsnell, C. A. Hutchison III, T. Kouzarides, J. A. Martignetti, E. Preddie, S. C. Satchwell, P. Tomlinson, K. M. Weston, and B. G. Barrell. 1990. Analysis of the protein-coding content of the sequence of human cytomegalovirus strain AD169. Curr. Top. Microbiol. Immunol. 154:125-169.
Colletti, K. S., Y. Xu, I. Yamboliev, and G. S. Pari. 2005. Human cytomegalovirus UL84 is a phosphoprotein that exhibits UTPase activity and is a putative member of the DExD/H box family of proteins. J. Biol. Chem. 280:11955-11960.
Colletti, K. S., Y. Xu, S. A. Cei, M. Tarrant, and G. S. Pari. 2004. Human cytomegalovirus UL84 oligomerization and heterodimerization domains act as transdominant inhibitors of oriLyt-dependent DNA replication: evidence that IE2-UL84 and UL84-UL84 interactions are required for lytic DNA replication. J. Virol. 78:9203-9214.
Cranmer, L. D., C. L. Clark, C. S. Morello, H. E. Farrell, W. D. Rawlinson, and D. H. Spector. 1996. Identification, analysis, and evolutionary relationships of the putative murine cytomegalovirus homologs of the human cytomegalovirus UL82 (pp71) and UL83 (pp65) matrix phosphoproteins. J. Virol. 70:7929-7939.
Davison, A. J. 1992. Channel catfish virus: a new type of herpesvirus. Virology 186:9-14.
Davison, A. J. 1998. The genome of salmonid herpesvirus 1. J. Virol. 72:1974-1982.
Davison, A. J. 2002. Evolution of the herpesviruses. Vet. Microbiol. 86:69-88.
Davison, A. J., B. L. Trus, N. Cheng, A. C. Steven, M. S. Watson, C. Cunningham, R. M. Le Deuff, and T. Renault. 2005. A novel class of herpesvirus with bivalve hosts. J. Gen. Virol. 86:41-53.
Davison, A. J., M. Benk?, and B. Harrach. 2003. Genetic content and evolution of adenoviruses. J. Gen. Virol. 84:2895-2908.
Dolan, A., C. Cunningham, R. D. Hector, A. F. Hassan-Walker, L. Lee, C. Addison, D. J. Dargan, D. J. McGeoch, D. Gatherer, V. C. Emery, P. D. Griffiths, C. Sinzger, B. P. McSharry, G. W. G. Wilkinson, and A. J. Davison. 2004. Genetic content of wild-type human cytomegalovirus. J. Gen. Virol. 85:1301-1312.
Dunn, W., C. Chou, H. Li, R. Hai, D. Patterson, V. Stolc, H. Zhu, and F. Liu. 2003. Functional profiling of a human cytomegalovirus genome. Proc. Natl. Acad. Sci. USA 100:14223-14228.
Fisher, F. B., and V. G. Preston. 1986. Isolation and characterisation of herpes simplex virus type 1 mutants which fail to induce dUTPase activity. Virology 148:190-197.
Frese, K. K., S. S. Lee, D. L. Thomas, I. J. Latorre, R. S. Weiss, B. A. Glaunsinger, and R. T. Javier. 2003. Selective PDZ protein-dependent stimulation of phosphatidylinositol 3-kinase by the adenovirus E4-ORF1 oncoprotein. Oncogene 22:710-721.
Fuchs, W., K. Ziemann, J. P. Teifke, O. Werner, and T. C. Mettenleiter. 2000. The non-essential UL50 gene of avian infectious laryngotracheitis virus encodes a functional dUTPase which is not a virulence factor. J. Gen. Virol. 81:627-638.
Gebert, S., S. Schmolke, G. Sorg, S. Floss, B. Plachter, and T. Stamminger. 1997. The UL84 protein of human cytomegalovirus acts as a transdominant inhibitor of immediate-early-mediated transactivation that is able to prevent viral replication. J. Virol. 71:7048-7060.
Glenn, M., L. Rainbow, F. Aurade, A. Davison, and T. F. Schulz. 1999. Identification of a spliced gene from Kaposi's sarcoma-associated herpesvirus encoding a protein with similarities to latent membrane proteins 1 and 2A of Epstein-Barr virus. J. Virol. 73:6953-6963.
González, A., G. Larsson, R. Persson, and E. Cedergren-Zeppezauer. 2001. Atomic resolution structure of Escherichia coli dUTPase determined ab initio. Acta Crystallogr. D Biol. Crystallogr. 57:767-774.
Goodstadt, L., and C. P. Ponting. 2001. CHROMA: consensus-based colouring of multiple alignments for publication. Bioinformatics 17:845-846.
Gorbalenya, A. E., and E. V. Koonin. 1993. Helicases: amino acid sequence comparisons and structure-function relationships. Curr. Opin. Struct. Biol. 3:419-429.
Gough, J., K. Karplus, R. Hughey, and C. Chothia. 2001. Assignment of homology to genome sequences using a library of hidden Markov models that represent all proteins of known structure. J. Mol. Biol. 313:903-919.
Hall, M. C., and S. W. Matson. 1999. Helicase motifs: the engine that powers DNA unwinding. Mol. Microbiol. 34:867-877.
Hofmann, H., H. Sindre, and T. Stamminger. 2002. Functional interaction between the pp71 protein of human cytomegalovirus and the PML-interacting protein human Daxx. J. Virol. 76:5769-5783.
Holtappels, R., J. Podlech, N. K. A. Grzimek, D. Thomas, M.-F. Pahl-Seibert, and M. J. Reddehase. 2001. Experimental preemptive immuonotherapy of murine cytomegalovirus disease with CD8 T-cell lines specific for ppM83 and pM84, the two homologs of human cytomegalovirus tegument protein ppUL83 (pp65). J. Virol. 75:6584-6600.
Homer, E. G., A. Rinaldi, M. J. Nicholl, and C. M. Preston. 1999. Activation of herpesvirus gene expression by the human cytomegalovirus protein pp71. J. Virol. 73:8512-8518.
Iyer, L. M., and L. Aravind. 2004. The emergence of catalytic and structural diversity within the beta-clip fold. Proteins 55:977-991.
Jenner, R. G., M. M. Alba, C. Boshoff, and P. Kellam. 2001. Kaposi's sarcoma-associated herpesvirus latent and lytic gene expression as revealed by DNA arrays. J. Virol. 75:891-902.
Kalejta, R. F., and T. Shenk. 2003a. The human cytomegalovirus UL82 gene product (pp71) accelerates progression through the G1 phase of the cell cycle. J. Virol. 77:3451-3459.
Kalejta, R. F., and T. Shenk. 2003b. Proteasome-dependent, ubiquitin-independent degradation of the Rb family of tumor suppressors by the human cytomegalovirus pp71 protein. Proc. Natl. Acad. Sci. USA 100:3263-3268.
Kalejta, R. F., J. T. Bechtel, and T. Shenk. 2003. Human cytomegalovirus pp71 stimulates cell cycle progression by inducing the proteasome-dependent degradation of the retinoblastoma family of tumor suppressors. Mol. Cell. Biol. 23:1885-1895.
Kattenhorn, L. M., R. Mills, M. Wagner, A. Lomsadze, V. Makeev, M. Borodovsky, H. L. Ploegh, and B. M. Kessler. 2004. Identification of proteins associated with murine cytomegalovirus virions. J. Virol. 78:11187-11197.
Kieff, E., and A. B. Rickinson. 2001. Epstein-Barr virus and its replication, p. 2511-2573. In D. M. Knipe and P. M. Howley (ed.), Fields virology, 4th ed. Lippincott Williams and Wilkins, Philadelphia, Pa.
Kremmer, E., P. Sommer, D. Holzer, S. A. Galetsky, V. A. Molochkov, V. Gurtsevitch, C. Winkelmann, R. Lisner, G. Niedobitek, and F. A. Grasser. 1999. Kaposi's sarcoma-associated herpesvirus (human herpesvirus-8) ORF54 encodes a functional dUTPase expressed in the lytic replication cycle. J. Gen. Virol. 80:1305-1310.
Kumar, S., K. Tamura, I. B. Jakobsen, and M. Nei. 2001. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 17:1244-1245.
Leppard, K. N. 1997. E4 gene function in adenovirus, adenovirus vector and adeno-associated virus infections. J. Gen. Virol. 78:2131-2138.
Lischka, P., G. Sorg, M. Kann, M. Winkler, and T. Stamminger. 2003. A nonconventional nuclear localization signal within the UL84 protein of human cytomegalovirus mediates nuclear import via the importin /? pathway. J. Virol. 77:3734-3748.
Liu, B., and M. F. Stinski. 1992. Human cytomegalovirus contains a tegument protein that enhances transcription from promoters with upstream ATF and AP-1 cis-acting elements. J. Virol. 66:4434-4444.
Lu, M., J. Suen, C. Frias, R. Pfeiffer, M. H. Tsai, E. Chuang, and S. L. Zeichner. 2004. Dissection of the Kaposi's sarcoma-associated herpesvirus gene expression program by using the viral DNA replication inhibitor cidofovir. J. Virol. 78:13637-13652.
Madera, M., C. Vogel, S. K. Kummerfeld, C. Chothia, and J. Gough. 2004. The SUPERFAMILY database in 2004: additions and improvements. Nucleic Acids Res. 32:D235-D239.
McClure, M. A. 2001. Evolution of the DUT gene: horizontal transfer between host and pathogen in all three domains of life. Curr. Protein Peptide Sci. 2:313-324.
McGeehan, J. E., N. W. Depledge, and D. J. McGeoch. 2001. Evolution of the dUTPase gene of mammalian and avian herpesviruses. Curr. Protein Peptide Sci. 2:325-333.
McGeoch, D. J. 1990. Protein sequence comparisons show that the ‘pseudoproteases’ encoded by poxviruses and certain retroviruses belong to the deoxyuridine triphosphatase family. Nucleic Acids Res. 18:4105-4110.
McGeoch, D. J., and A. J. Davison. 1999. The molecular evolutionary history of the herpesviruses, p. 441-465. In E. Domingo, R. Webster, and J. Holland (ed.), Origin and evolution of viruses. Academic Press, London, England.
McGeoch, D. J., M. A. Dalrymple, A. J. Davison, A. Dolan, M. C. Frame, D. McNab, L. J. Perry, J. E. Scott, and P. Taylor. 1988. The complete DNA sequence of the long unique region in the genome of herpes simplex virus type 1. J. Gen. Virol. 69:1531-1574.
McGregor, A., F. Liu, and M. R. Schleiss. 2004. Molecular, biological, and in vivo characterization of the guinea pig cytomegalovirus (CMV) homologs of the human CMV matrix proteins pp71 (UL82) and pp65 (UL83). J. Virol. 78:9872-9889.
McLaughlin-Taylor, E., H. Pande, S. J. Forman, B. Tanamachi, C. R. Li, J. A. Zaia, P. D. Greenberg, and S. R. Riddell. 1994. Identification of the major late human cytomegalovirus matrix protein pp65 as a target antigen for CD8+ virus-specific cytotoxic T lymphocytes. J. Med. Virol. 43:103-110.
Menegazzi, P., M. Galvan, A. Rotola, T. Ravaioli, A. Gonelli, E. Cassai, and D. Di Luca. 1999. Temporal mapping of transcripts of human herpesvirus-7. J. Gen. Virol. 80:2705-2712.
Minson, A. C., A. Davison, R. Eberle, R. C. Desrosiers, B. Fleckenstein, D. J. McGeoch, P. E. Pellett, B. Roizman, and M. J. Studdert. 2000. Family Herpesviridae, p. 203-225. In M. H. V. van Regenmortel, C. M. Fauquet, D. H. L. Bishop, E. B. Carstens, M. K. Estes, S. M. Lemon, J. Maniloff, M. A. Mayo, D. J. McGeoch, C. R. Pringle, and R. B. Wickner (ed.), Virus taxonomy: seventh report of the International Committee on Taxonomy of Viruses. Academic Press, London, England.
Mol, C. D., J. M. Harris, E. M. McIntosh, and J. A. Tainer. 1996. Human dUTP pyrophosphatase: uracil recognition by a ? hairpin and active sites formed by three separate subunits. Structure 4:1077-1092.
Moorman, N. J., C. Y. Lin, and S. H. Speck. 2004. Identification of candidate gammaherpesvirus 68 genes required for virus replication by signature-tagged transposon mutagenesis. J. Virol. 78:10282-10290.
Morello, C. S., L. D. Cranmer, and D. H. Spector. 1999. In vivo replication, latency, and immunogenicity of murine cytomegalovirus mutants with deletions in the M83 and M84 genes, the putative homologs of human cytomegalovirus pp65 (UL83). J. Virol. 73:7678-7693.
Morello, C. S., L. D. Cranmer, and D. H. Spector. 2000. Suppression of murine cytomegalovirus (MCMV) replication with a DNA vaccine encoding MCMV M84 (a homolog of human cytomegalovirus pp65). J. Virol. 74:3696-3708.
Neipel, F., J. C. Albrecht, and B. Fleckenstein. 1997. Cell-homologous genes in the Kaposi's sarcoma-associated rhadinovirus human herpesvirus 8: determinants of its pathogenicity? J. Virol. 71:4187-4192.
Novotny, J., I. Rigoutsos, D. Coleman, and T. Shenk. 2001. In silico structural and functional analysis of the human cytomegalovirus (HHV5) genome. J. Mol. Biol. 310:1151-1166.
Pari, G. S., and D. G. Anders. 1993. Eleven loci encoding trans-acting factors are required for transient complementation of human cytomegalovirus oriLyt-dependent DNA replication. J. Virol. 67:6979-6988.
Parker, B. D., A. Bankier, S. Satchwell, B. Barrell, and P. J. Farrell. 1990. Sequence and transcription of Raji Epstein-Barr virus DNA spanning the B95-8 deletion region. Virology 179:339-346.
Preston, V. G., and F. B. Fisher. 1984. Identification of the herpes simplex virus type 1 gene encoding the dUTPase. Virology 138:58-68.
Prince, V. E., and F. B. Pickett. 2002. Splitting pairs: the diverging fates of duplicated genes. Nat. Rev. Genet. 3:827-837.
Pyles, R. B., N. M. Sawtell, and R. L. Thompson. 1992. Herpes simplex virus type 1 dUTPase mutants are attenuated for neurovirulence, neuroinvasiveness, and reactivation from latency. J. Virol. 66:6706-6713.
Reid, G. G., V. Ellsmore, and N. D. Stow. 2003. An analysis of the requirements for human cytomegalovirus oriLyt-dependent DNA synthesis in the presence of the herpes simplex virus type 1 replication fork proteins. Virology 308:303-316.
Rocak, S., and P. Linder. 2004. DEAD-box proteins: the driving forces behind RNA metabolism. Nat. Rev. Mol. Cell. Biol. 5:232-241.
Rost, B. 1996. PHD: predicting one-dimensional protein structure by profile-based neural networks. Methods Enzymol. 266:525-539.
Rost, B., and C. Sander. 1993. Improved prediction of protein secondary structure by use of sequence profiles and neural networks. Proc. Natl. Acad. Sci. USA 90:7558-7562.
Ruger, B., S. Klages, B. Walla, J. Albrecht, B. Fleckenstein, P. Tomlinson, and B. Barrell. 1987. Primary structure and transcription of the genes coding for the two virion phosphoproteins pp65 and pp71 of human cytomegalovirus. J. Virol. 61:446-453.
Sarisky, R. T., and G. S. Hayward. 1996. Evidence that the UL84 gene product of human cytomegalovirus is essential for promoting oriLyt-dependent DNA replication and formation of replication compartments in cotransfection assays. J. Virol. 70:7398-7413.
Schierling, K., T. Stamminger, T. Mertens, and M. Winkler. 2004. Human cytomegalovirus tegument proteins ppUL82 (pp71) and ppUL35 interact and cooperatively activate the major immediate-early enhancer. J. Virol. 78:9512-9523.
Schleiss, M. R., A. McGregor, N. J. Jensen, G. Erdem, and L. Aktan. 1999. Molecular characterization of the guinea pig cytomegalovirus UL83 (pp65) protein homolog. Virus Genes 19:205-221.
Schmolke, S., H. F. Kern, P. Drescher, G. Jahn, and B. Plachter. 1995. The dominant phosphoprotein pp65 (UL83) of human cytomegalovirus is dispensable for growth in cell culture. J. Virol. 69:5959-5968.
Shackelton, L. A., and E. C. Holmes. 2004. The evolution of large DNA viruses: combining genomic information of viruses and their hosts. Trends Microbiol. 12:458-465.
Silverman, E., G. Edwalds-Gilbert, and R.-J. Lin. 2003. DExD/H proteins and their partners: helping RNA helicases unwind. Gene 312:1-16.
Sommer, P., E. Kremmer, S. Bier, S. Konig, P. Zalud, M. Zeppezauer, J. F. Jones, N. Mueller-Lantzsch, and F. A. Grasser. 1996. Cloning and expression of the Epstein-Barr virus-encoded dUTPase: patients with acute, reactivated or chronic virus infection develop antibodies against the enzyme. J. Gen. Virol. 77:2795-2805.
Song, M. J., S. Hwang, W. H. Wong, T.-T. Wu, S. Lee, H.-I. Liao, and R. Sun. 2005. Identification of viral genes essential for replication of murine -herpesvirus 68 using signature-tagged mutagenesis. Proc. Natl. Acad. Sci. USA 102:3805-3810.
Spector, D. J., and M. J. Tevethia. 1994. Protein-protein interactions between human cytomegalovirus IE2-580aa and pUL84 in lytically infected cells. J. Virol. 68:7549-7553.
Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673-4680.
Tuteja, N., and R. Tuteja. 2004. Unraveling DNA helicases: motif, structure, mechanism and function. Eur. J. Biochem. 271:1849-1863.
Varnum, S. M., D. N. Streblow, M. E. Monroe, P. Smith, K. J. Auberry, L. Pasa-Tolic, D. Wang, D. G. Camp II, K. Rodland, S. Wiley, W. Britt, T. Shenk, R. D. Smith, and J. A. Nelson. 2004. Identification of proteins in human cytomegalovirus (HCMV) particles: the HCMV proteome. J. Virol. 78:10960-10966.
Weiss, R. S., S. S. Lee, B. V. Prasad, and R. T. Javier. 1997. Human adenovirus early region 4 open reading frame 1 genes encode growth-transforming proteins that may be distantly related to dUTP pyrophosphatase enzymes. J. Virol. 71:1857-1870.
Wills, M. R., A. J. Carmichael, K. Mynard, X. Jin, M. P. Weekes, B. Plachter, and J. G. P. Sissons. 1996. The human cytotoxic T-lymphocyte (CTL) response to cytomegalovirus is dominated by structural protein pp65: frequency, specificity, and T-cell receptor usage of pp65-specific CTL. J. Virol. 70:7569-7579.
Xu, Y., K. S. Colletti, and G. S. Pari. 2002. Human cytomegalovirus UL84 localizes to the cell nucleus via a nuclear localization signal and is a component of viral replication compartments. J. Virol. 76:8931-8938.
Xu, Y., S. A. Cei, A. Rodriguez Huete, and G. S. Pari. 2004a. Human cytomegalovirus UL84 insertion mutant defective for viral DNA synthesis and growth. J. Virol. 78:10360-10369.
Xu, Y., S. A. Cei, A. Rodriguez Huete, K. S. Colletti, and G. S. Pari. 2004b. Human cytomegalovirus DNA replication requires transcriptional activation via an IE2- and UL84-responsive bidirectional promoter element within oriLyt. J. Virol. 78:11664-11677.
Yu, D., M. C. Silva, and T. Shenk. 2003. Functional map of human cytomegalovirus AD169 defined by global mutational analysis. Proc. Natl. Acad. Sci. USA 100:12396-12401.
Zhu, F. X., J. M. Chong, L. Wu, and Y. Yuan. 2005. Virion proteins of Kaposi's sarcoma-associated herpesvirus. J. Virol. 79:800-811.(Andrew J. Davison and Nig)