The herpesviruses have been subdivided into three subfamilies based on molecular and biological properties (
27). Although members of the three herpesvirus subfamilies exhibit a wide range of biological properties and the range of pathogenic manifestations (
27), their genomes contain a significant number of conserved genes. These conserved genes include genes encoding glycoproteins (e.g., gB and gH), regulatory proteins (e.g., ICP27), and a large array of conserved enzymes involved in posttranslational modification of protein (e.g., protein kinases), DNA synthesis (e.g., DNA polymerase, helicase, and primase), and processing of proteins (e.g., protein kinase) (
28). This conservation suggests that these gene products play an essential role in the life cycle of herpesviruses.
In an earlier study, we showed that infection of cells with herpes simplex virus 1 (HSV-1), a prototype of alphaherpesvirus, causes extensive hyperphosphorylation of translation elongation factor 1δ (EF-1δ) (
11). In a subsequent study, we found that the gene product of HSV-1 responsible for EF-1δ hyperphosphorylation is a viral protein kinase encoded by the U
L13 gene (
13). This observation and the fact that the amino acid sequence that encodes U
L13 protein kinase is conserved in members of all subfamilies of the family
Herpesviridae led us to investigate whether EF-1δ is also posttranslationally modified in cells infected with representative members of all subfamilies of herpesviruses. In this report, we present evidence that this is in fact the case. In one instance, we have specifically related the modification of EF-1δ to the protein kinase encoded by U
L97, the human cytomegalovirus (HCMV) homologue of the U
L13 gene of HSV-1. Relevant to this report are the following observations.
(i) EF-1δ is a subunit of EF-1, a complex of proteins which mediate the elongation of polypeptide chains during translation of mRNA (
16,
18,
26,
33). EF-1α transports aminoacyl tRNA for binding to ribosomes concurrent with the hydrolysis of GTP. EF-1δ is a component of the EF-1β–EF-1γ–EF-1δ complex which is responsible for GDP-GTP exchange on EF-1α, and numerous studies have ascribed to EF-1δ regulatory functions (
16,
19,
26).
(ii) EF-1δ interacts with two viral proteins in cells infected with HSV-1. As reported elsewhere, EF-1δ interacts with ICP0, an α protein known primarily for its function as a promiscuous transactivator. Thus, ICP0 interacts physically in the yeast two-hybrid system with EF-1δ, and the domain of ICP0 interacting with EF-1δ affects translational efficiency in vitro (
11). ICP0 is a multifunctional protein which interacts with and modulates many key cellular functions, including transcription (
7,
10), translation (
11), cell cycle regulation (
12), and the protein degradation pathway (
6). EF-1δ also interacts with and is hyperphosphorylated by the protein kinase encoded by the U
L13 gene. The observation that two viral proteins interact with EF-1δ suggests that this protein plays an important role in the viral life cycle.
(iii) The U
L13 gene was originally reported to contain motifs conserved in eukaryotic protein kinase (
1,
30) and in subsequent studies was associated with the phosphorylation of itself, EF-1δ, and several viral proteins (e.g., ICP22, ICP0, and gE) (
2,
13,
20,
22,
25). Recent evidence based on experiments on purified U
L13 protein unambiguously demonstrated that it is a protein kinase (
3). The U
L13 protein is packaged into virion, and indeed, a protein kinase activity is associated with purified virions (
2,
23). Experiments with genetically engineered viruses lacking the U
L13 gene showed that the viral protein kinase affects the accumulation of ICP0 and a subset of γ proteins, suggesting that the kinase activity of U
L13 may play a regulatory function (
24).
(iv) The HCMV protein U
L97 has homology to U
L13 of HSV, so the protein was predicted to be a protein kinase (
1). Interestingly, the U
L97 protein phosphorylates ganciclovir, a nucleotide analog effective against HCMV infection, and therefore, the viral protein can act as a deoxynucleoside kinase (
14,
31). Studies on recombinant HSV-1 in which the U
L13 protein was replaced by its HCMV homologue showed that U
L97 can substitute for the U
L13 protein kinase (Fig.
1), and only recently it has been shown to be a serine/threonine kinase that phosphorylates itself (
8,
21). The results supported the hypothesis that U
L97 is a protein kinase with the unusual property of being able to phosphorylate deoxynucleosides.
In this report we demonstrate that EF-1δ is commonly modified in cells infected with representative members of the alpha-, beta-, and gammaherpesvirus subfamilies and that this phenomenon is not unique to HSV-1-infected cells. We also present evidence that HCMV UL97 can mediate the modification of EF-1δ.
Vero and human lung fibroblast (HLF) cells were obtained from the American Type Culture Collection and Aviron (Mountain View, Calif.), respectively. Madin-Darby bovine kidney (MDBK) and cloned porcine kidney (CPK) cells were provided by H. Ohtsuka (The University of Tokyo). Crandell feline kidney (CRFK) cells were a gift from T. Mikami (The University of Tokyo). Fetal horse kidney (FHK) cells were isolated as described elsewhere (
15) and used within seven passages of primary culture. The cell lines were grown in Dulbecco’s modified Eagle’s medium supplemented with either 10% fetal calf serum (HLF, MDBK, CPK, or CRFK cells) or 5% fetal calf serum (Vero cells).
The properties of the virus strains HSV-1(F), HSV-2(G), HCMV(Towne) and the genetically engineered recombinant viruses listed in Table
1 have been described elsewhere (
5,
9,
21,
22,
24,
25). Feline herpesvirus 1 [FHV-1(C7301)] (
17) was provided by T. Mikami. Bovine herpesvirus 1 [BHV-1(LA)] and pseudorabies virus [PRV(Indiana)] were supplied by H. Ohtsuka. Japanese field isolates of equine herpesvirus 2 [EHV-2(87C26) and EHV-2(87C33)] were isolated with FHK cells as described elsewhere (
15). gB regions of the genomes of the EHV-2 strains were amplified, cloned into pGEM-T vector (Promega), and sequenced. The sequences were completely identical to the published sequence of the EHV-2 gB (
32).
The rabbit polyclonal antibody to EF-1δ was generated as described elsewhere (
11). The electrophoretically separated proteins transferred to nitrocellulose sheets were reacted with the antibody to EF-1δ as described previously (
11,
12).
The objective of the first series of experiments was to determine whether alphaherpesviruses other than HSV-1 also cause hyperphosphorylation of EF-1δ in infected cells. As reported previously, EF-1δ consists of two predominant forms, a hypophosphorylated form (apparent
M r of 38,000) and a hyperphosphorylated form (apparent
M r of 40,000) (
11,
13,
19,
29). The polyclonal antibody used in this study can readily detect both forms of EF-1δ, and the pattern of bands of EF-1δ radiolabeled by
32Pi is exactly the same as that of EF-1δ detected by Western blotting (
11,
13). To monitor changes in EF-1δ, Vero, CRFK, CPK and MDBK cells were harvested 24 h after mock infection or infection with HSV-2(G), FHV-1(C7301), PRV(Indiana), and BHV-1(LA), respectively, solubilized, electrophoretically separated in denaturing gels, electrophoretically transferred to nitrocellulose sheets, and reacted with the rabbit polyclonal antibody to EF-1δ. The results were as follows.
(i) In mock-infected Vero and CRFK cells, the two predominant bands migrating with
M rs of 38,000 and 40,000, respectively, were detected (Fig.
2) as described previously (
11,
13). The result that the polyclonal antibody raised against human EF-1δ-glutathione
S-transferase fusion protein detected EF-1δ in feline cells (Fig.
2) indicates that the epitopes to which the antibody is directed are conserved in nonhuman EF-1δ. In cells infected with HSV-2(G) or FHV-1(C7301), the amount of EF-1δ protein in slower-migrating bands was significantly increased (Fig.
2).
(ii) In mock-infected CPK and MDBK cells, the slow-migrating bands of EF-1δ protein observed in Vero and CRFK cells were barely detectable, whereas the slow-migrating bands of EF-1δ protein were prominent in cells infected with PRV(Indiana) or BHV-1(LA) (Fig.
2).
These results indicate that EF-1δ is commonly modified in cells infected with representative human and nonhuman alphaherpesviruses.
To address the question of whether beta- and gammaherpesviruses cause the modification in the migration of EF-1δ associated with hyperphosphorylation of the protein, we examined the electrophoretic mobility of EF-1δ accumulating in HLF or FHK cells infected with a human betaherpesvirus, HCMV, and an equine gammaherpesvirus, equine herpesvirus 2 (EHV-2), respectively. As shown in Fig.
3, there was a dramatic increase in the abundance of the highly modified form of EF-1δ in cells infected with HCMV or EHV-2. In Fig.
3B, the amount of the hypophosphorylated form of EF-1δ also increased slightly after EHV-2 infection. However, the increase in hyperphosphorylated EF-1δ after EHV-2 infection is not due to a simple increase in the EF-1δ translation products but a consequence of modification of EF-1δ induced by EHV-2 infection since the slower-migrating band of EF-1δ was barely detectable when lysate from mock-infected cells containing a larger amount of hypophosphorylated form of EF-1δ than that from EHV-2-infected cells was subjected to Western blotting (data not shown). We infer from these results that the modification of EF-1δ is a conserved function in members of all subfamilies of
Herpesviridae. The evidence that EF-1δ was modified in nonprimate cell lines such as feline, porcine, bovine, and equine cells (Fig.
2 and
3) indicates that the modification is not restricted to primate cells infected with human herpesviruses. Rather, it is conserved phenomena in a variety of species of mammalian cells infected with various animal herpesviruses.
The observation that herpesviruses of all subfamilies modify EF-1δ raised the possibility that the conserved U
L13 homologues of other herpesviruses are involved in the modification of the protein. To test this hypothesis, we used the genetically engineered HSV-1 recombinants R4969 and R4970 (Fig.
1), in which the coding domains of U
L13 were replaced by a chimeric gene capable of expressing HCMV U
L97 (
21). As reported elsewhere (
13), the amounts of hypo- and hyperphosphorylated EF-1δ in cells infected with mutant viruses lacking the U
L13 gene could not be differentiated from those of mock-infected cells. Therefore, if HCMV U
L97 protein were involved in the modification of EF-1δ, the amount of the modified form of EF-1δ would increase after infection with R4969 and R4970. In this series of experiments, HLF cells were harvested 24 h after mock infection or infection with wild-type virus or mutant viruses, solubilized, electrophoretically separated in a denaturing gel, transferred to a nitrocellulose sheet, and reacted with the antibody to EF-1δ. The results were as follows.
(i) As reported previously (
13), in mock-infected cells, the hypophosphorylated form of EF-1δ is dominant, whereas the amount of the hyperphosphorylated form of the protein increased after infection with wild-type viruses (Fig.
4, lanes 2 and 5). The electrophoretic pattern of EF-1δ extracted from mock-infected cells (lane 1) could not be differentiated from those observed for the protein extracted from cells infected with U
L13 deletion viruses (lanes 3 and 6). Furthermore, the wild-type virus phenotype was restored in cells infected with the recombinant R7358, in which the U
L13 sequence had been repaired (lane 8).
(ii) In cells infected with U
L13
−/U
L97
+ viruses (Fig.
4, lanes 4 and 7), the electrophoretic pattern of EF-1δ forms was clearly differentiated from those of cells infected with U
L13 deletion mutants and mock-infected cells. The amount of the hyperphosphorylated form of EF-1δ significantly increased in cells infected with U
L13
−/U
L97
+ viruses compared to that in mock-infected cells or cells infected with U
L13 deletion mutant viruses, indicating that U
L97 conferred on the recipient U
L13
− virus the capacity to modify EF-1δ. The pattern of EF-1δ in cells infected with U
L13
−/U
L97
+ viruses was quite similar to that observed in cells infected with wild-type HCMV (Fig.
3A). These results indicate that the posttranslational modification of EF-1δ is mediated by HCMV U
L97.
Phosphorylation of proteins by protein kinase is a major strategy employed by eukaryotic cells to regulate cellular function, including transcription, translation, protein degradation, etc. (
4). Viruses use a similar strategy, both to regulate their own replicative processes and to modify cellular proteins to suite their needs. Our knowledge regarding the requirements of herpesviruses with respect to host biosynthetic processes depends in part on the identification of cellular proteins modified by viral protein kinases and on elucidation of the functions altered by the phosphorylation. We should note that the original observation that HSV-1 protein kinase phosphorylates EF-1δ should, in retrospect, have come as no surprise, since protein synthesis is the key to viral replication and yet the translation of mRNA seems immune to viral interference.
The hypothesis we have attempted to test stems from the observation that the cellular translation factor, EF-1δ is hyperphosphorylated by the HSV-1 U
L13 protein kinase and the evidence that homologues of the U
L13 gene are present in the genomes of members of all three subfamilies of herpesviruses (
1,
30). Since protein kinases in general recognize specific substrates, the question arose whether other subfamilies of herpesviruses possess the ability to induce hyperphosphorylation of EF-1δ and whether homologues of U
L13 mediate the modification of the translation factor. The salient features of our results are as follows.
(i) Alpha-, beta-, and gammaherpesviruses commonly induce the posttranslational modification of EF-1δ associated with hyperphosphorylation of the protein. Infection of various mammalian cells with alphaherpesviruses (HSV-2, FHV-1, PRV, and BHV-1), betaherpesvirus (HCMV), and gammaherpesvirus (EHV-2) induced accumulation of the slow-migrating form of EF-1δ in various mammalian cells susceptible to these viruses. These results indicate that the modification of EF-1δ in various species of mammalian cells is a conserved function expressed by all herpesviruses and by extension, that EF-1δ plays a crucial role in replication of the viruses.
(ii) The U
L97 protein of HCMV mediates the posttranslational modification of EF-1δ. The electrophoretic profiles of EF-1δ from cells infected with the U
L13
−/U
L97
+ viruses were different from those of cells infected with the parent U
L13
− viruses. Although the U
L97 protein has long been thought to be a protein kinase because it has the motifs associated with protein kinases, the only available information of U
L97 acting as a protein kinase is the evidence that it mediated the phosphorylation HSV-1 proteins when inserted into a HSV-1 genome (
21) and that it phosphorylates itself (
8). This report shows the first evidence that U
L97 can also mediate the posttranslational modification that is involved in the phosphorylation of natural proteins except for U
L97 itself. Although we have not demonstrated that U
L97 directly phosphorylated EF-1δ, this is highly likely in light of the homology of U
L97 and U
L13 proteins. By extension, U
L13 homologues of all herpesvirus subfamilies may be responsible for the posttranslational modification of EF-1δ.
The significant new contribution of this report is the observation that the modification of EF-1δ is a specific, conserved objective of herpesviruses belonging to all three subfamilies of herpesvirus family and not merely that of HSV-1. This observation eliminates the slim possibility that UL13 by chance retained the motif necessary to bind and phosphorylate EF-1δ. Our results indicate that EF-1δ is of universal importance in herpesvirus infection. Although the physiological role of the posttranslational modification of EF-1δ by herpesvirus infection remains to be determined, it is conceivable that the modification is beneficial to viral replication and is therefore associated with activation of higher level of protein synthesis. Further experiments to unveil the direct significance of the EF-1δ hyperphosphorylation are of interest and are currently under way.
Acknowledgments
We thank T. Mikami for providing CRFK cells and FHV-1(C7301) strain, and H. Ohtsuka for providing CPK and MDBK cells and PRV(Indiana) and BHV-1(LA) strains.
The work at Tokyo Medical and Dental University was supported in part by grants from the Ministry of Education, Science, Culture and Sport of Japan and the Japan Health Science Foundation. Y.K. was supported by a grant from the Ichiro Kanehara Foundation. The work at the University of Chicago was aided by grants from the National Cancer Institute (CA47451, CA71933, and CA78766) and the U.S. Public Health Service.