INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

G protein-coupled receptors (GCRs) play a key role in transduction of extracellular signals to the intracellular environment. They can be activated by a variety of stimuli, such as neurotransmitters, hormones, and photons (reviewed by Probst et al. [53]). Upon ligand binding, GCRs activate G proteins, which in turn activate effector enzymes and ion channels in a cascade-like fashion. Thousands of GCR variants are encoded by genes of both prokaryotes and eukaryotes. Additionally, some are encoded by virus genes. To date, 20 putative viral GCR genes have been discovered: 2 in poxvirus genomes (21, 44), 11 in betaherpesvirus genomes (9, 22, 33, 49, 54), and 7 in gammaherpesvirus genomes (4, 26, 48, 57, 63). The majority of these genes was found to be similar to genes encoding cellular chemokine receptors. Although the functions of most of the putative viral GCRs are unclear, several are capable of binding chemokines, hence invoking a classical signal transduction response (1, 4, 31, 38, 42, 47). The herpesvirus saimiri (HVS) ECRF3-encoded chemokine receptor is capable of transducing signals upon activation by  chemokines (1). The human cytomegalovirus (HCMV) US28-encoded chemokine receptor was shown to be a promiscuous calcium-mobilizing receptor for several  chemokines (31). Additionally, the US28 protein was suggested to be responsible for -chemokine sequestration in HCMV-infected fibroblasts (15). The GCR encoded by Kaposi's sarcoma-associated herpesvirus (KSHV) open reading frame (ORF) 74 not only binds both  and  chemokines but also is constitutively active, inducing second-messenger signalling in vitro (4). The KSHV chemokine receptor was also shown to stimulate cellular proliferation (4), transformation, and angiogenesis (7). A distinctive set of chemokine-like receptors is exclusively encoded by betaherpesviruses: HCMV UL33 (23), rat cytomegalovirus (RCMV) R33 (9), murine cytomegalovirus (MCMV) M33 (54), and human herpesvirus 6 and 7 (HHV-6 and -7) U12 (33, 49). Recently, we found that the RCMV R33 gene is essential for the pathogenesis of viral infection in vivo and that unlike wild-type (wt) virus, an RCMV R33 null mutant could neither enter nor replicate in salivary gland epithelial cells of infected rats (9). A similar observation was made for MCMV M33: upon infection of mice with an MCMV M33 deletion mutant strain, virus could not be recovered from mouse salivary glands (26). Interestingly, betaherpesviruses encode another distinctive set of yet uncharacterized viral GCRs: HCMV UL78 (22), MCMV M78 (54), and HHV-6 and -7 U51 (33, 49). Although the positions of these UL78-like genes within the betaherpesvirus genomes are conserved, their sequences are rather divergent. Despite the presence of distinct GCR characteristics, such as seven transmembrane domains, two conserved cysteine residues, and a G protein-coupling domain (53), these UL78-like gene products show little similarity with any of the thousands of GCRs known to date. Nevertheless, characterization of this unique family of GCRs may be crucial to the development of new antiviral therapeutics. In this report, we present the sequence and transcriptional analysis of the RCMV member of this family of GCR genes, which we termed R78. In addition, RCMV strains were generated in which the R78 (ORF) is either partially or completely deleted from the genome (RCMVR78a or RCMVR78c, respectively). We show that disruption of the R78 gene affects RCMV replication in permissive cell types in vitro and that the RCMV R78 deletion mutant strains induce syncytium formation in rat embryo fibroblasts (REF) in vitro, in contrast to wt RCMV. In addition, a dramatically lower mortality was observed in rats infected with either RCMVR78a or RCMVR78c than in wt RCMV-infected rats. We conclude that the RCMV R78 gene plays a vital role in both RCMV replication in vitro and the pathogenesis of viral infection in vivo.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cells and virus. Primary REF, rat fibroblast cell line Rat2 TK (ATCC CRL 1764), rat heart endothelium cell line 116 (RHEC), monocyte/macrophage cell line R2 (R2M), and rat aorta medial smooth muscle cells (RSMC) were cultured as described previously (18, 24, 50, 60). RCMV (Maastricht strain) was propagated in REF (18). Virus titers were determined by a plaque assay using standard procedures (19). RCMV DNA was isolated from culture medium as described by Vink et al. (61).

Identification, cloning, and sequence analysis of the RCMV R78 gene. Cloning of the 30-kb XbaI B fragment of the RCMV genome into vector pSP62-PL has previously been described (45). The XbaI B fragment was digested with various restriction endonucleases, and the resulting fragments were cloned into vector pUC119. Both strands of each clone were sequenced by using the Cy5 Autoread sequencing kit (Pharmacia Biotech, Roosendaal, The Netherlands) and ALFexpress automated DNA sequencer (Pharmacia Biotech). Sequence analysis was done with the program PC/Gene (version 2.11; IntelliGenetics). The sequences were checked for the presence of HCMV UL78-homologous regions by alignment with the GenBank nucleic acid database, using the BLASTN search algorithm (39). Thus, a 3.7-kb BamHI fragment was identified, which contains an ORF with considerable similarity to the HCMV UL78 ORF.

RCMVR78a recombination plasmid construction. Plasmid p115, which contains a large part of the R78 ORF on a 2.4-kb SalI fragment (Fig. 1), was digested with SalI and subsequently treated with deoxynucleotide triphosphates and DNA polymerase I Klenow fragment (Pharmacia Biotech). Subsequently, the DNA was digested with BglII, and the resulting 1.4-kb fragment was cloned into BamHI-SmaI-treated vector pUC119, generating plasmid pA. Plasmid p114, which contains the remaining part of the R78 ORF on a 1.7-kb SalI fragment (Fig. 1), was digested with SalI. The 1.7-kb SalI insert from p114 was ligated into the SalI site of pA, generating pB. A 1.5-kb BamHI-EcoRI fragment from Rc/CMV (Invitrogen, Leek, The Netherlands), containing the neomycin resistance (neo) gene, was treated with Klenow enzyme and ligated into a XbaI-digested and Klenow enzyme-treated vector pB. This final construct (p147 [see Fig. 4]) was linearized by digestion with Asp718I and HindIII and used for transfection, in order to generate mutant RCMVR78a.

View larger version (12K): [in this window] [in a new window]   FIG. 1.   Restriction map of the RCMV genome and relative position of the R78 gene, which encodes a putative GCR. An enlarged section of the map is shown below. Arrow boxes indicate the size and polarity of conserved RCMV ORFs.

RCMVR78c recombination plasmid construction. Plasmid p116, which contains the complete R78 ORF on a 3.7-kb BamHI fragment (Fig. 1), was digested with Asp718I and NcoI. The resulting 2.9-kb fragment was treated with Klenow enzyme and subsequently circularized with T4 DNA ligase, resulting in plasmid pC. One SalI site was removed from pC by digestion of the plasmid with XbaI and HindIII and subsequent treatment with Klenow enzyme. The resulting fragment was circularized, generating plasmid, pD. Plasmid pD was digested with SalI, treated with Klenow enzyme, and ligated to the blunt-ended DNA fragment containing the Rc/CMV neo gene (see above). The resulting plasmid, p120 (see Fig. 4), was linearized with Asp718I and XbaI prior to transfection, in order to generate mutant RCMVR78c.

Generation of RCMV R78 deletion mutants. Approximately 10⁷ Rat2 cells were trypsinized and subsequently centrifuged for 5 min at 500 × g. The cells were resuspended in 0.25 ml of culture medium, after which 10 µg of linearized plasmid of either p120 or p147 was added. The suspension was transferred to a 0.4-cm electroporation cuvette (Bio-Rad, Veenendaal, The Netherlands) and pulsed at 0.25 kV and 500 µF in a Bio-Rad Gene Pulser electroporator. The cells were subsequently seeded in 10-cm-diameter culture dishes. At 6 h after transfection, the cells were infected with RCMV at a multiplicity of infection (MOI) of 1. The culture medium was supplemented with 50 µg of G418 per ml at 16 h postinfection (p.i.). Recombinant viruses were plaque purified and cultured on REF monolayers as described earlier (9).

Southern blot hybridization. DNA was isolated from wt RCMV, RCMVR78a, and RCMVR78c and then digested with either BamHI, BamHI-BglII, BamHI-EcoRV, BamHI-NcoI, SalI, or SalI-EcoRV. Subsequently, the digested DNA was electrophoresed through a 1% agarose gel and blotted onto a Hybond N⁺ nylon membrane (Amersham, 's-Hertogenbosch, The Netherlands) as described previously (16). Both a 3.7-kb insert from p116 containing the intact R78 ORF as well as sequences from R77 and R79 (R78 probe [see Fig. 5A]), along with a 1.5-kb BamHI-EcoRI fragment containing the neo gene (neo probe) from Rc/CMV, were used as probes. Hybridization and detection experiments were performed with digoxigenin DNA labeling and chemiluminescence detection kits (Boehringer Mannheim, Almere, The Netherlands).

Isolation of poly(A)⁺ RNA and Northern blot analysis. RCMV poly(A)⁺ RNA was isolated from REF at 2, 8, and 48 h p.i. and at immediate-early (IE), early (E), and late (L) times of infection with RCMV (MOI = 1). To obtain IE mRNA, REF were treated with 100 µg of cycloheximide per ml 1 h before, during, and 16 h after infection. During the 1-h infection period, the cells were exposed to RCMV. E mRNA was isolated after infection of REF with RCMV and treatment of cells with 100 µg of phosphonoacetic acid per ml from 3 h p.i. until the cells were harvested at 13 h p.i. L mRNA was isolated after infection of REF with either RCMV, RCMVR78a, or RCMVR78c and harvesting of cells at 72 h p.i. To obtain mRNA from mock-infected cells, a procedure similar to that described for the purification of L mRNA was used except that RCMV infection was omitted. Poly(A)⁺ RNA was purified with a QuickPrep Micro mRNA purification kit (Pharmacia Biotech). Aliquots (1 µg) of poly(A)⁺ RNA were electrophoresed through agarose under denaturing conditions as described by Brown and Mackey (17); then the RNA was transferred to positively charged nylon membranes (Boehringer Mannheim) as described previously (17). The 754-bp BamHI-SalI fragment from p115 and the 950-bp BglII-SalI, 206-bp BglII-NarI, and 1,088-bp BglII-SalI fragments from p116 (see Fig. 6A) were used to generate probes. These fragments contain R77-, R78-, R79-, and R79/R80-specific sequences, respectively. Hybridization and detection experiments were performed with digoxigenin DNA labeling and chemiluminescence detection kits (Boehringer Mannheim).

Replication of R78a and R78c in vitro. REF, RHEC, R2M, and RSMC were grown either in 96-well plates or on glass slides and infected with either RCMV, RCMVR78a, or RCMVR78c at an MOI of 0.1 or 1. Culture medium samples (three per virus) were taken at 1, 3, 5, and 7 days p.i. and subjected to plaque titer determination. The cells were fixed and stained with monoclonal antibodies (MAbs) against RCMV E proteins (MAb RCMV 8 [20]) as described previously (60). The degree of infection was determined by counting the number of antigen-positive cells relative to the total number of cells in three different wells (four microscopic fields per well at a magnification of ×400).

Dissemination of wt RCMV and RCMVR78c in vivo. Male specific-pathogen-free Lewis/N RT1 rats (Central Animal Facility, Maastricht University, Maastricht, The Netherlands), used for all in vivo experiments in this study, were kept under standard conditions (55). Rats were immunocompromised by 5 Gy of total-body Röntgen irradiation 1 day before infection as described by Stals et al. (55), and all virus stocks that were used for inoculation in vivo were derived from tissue culture medium of virus-infected REF. Two groups of rats (10 weeks old; body weight of 250 to 300 g) were infected with 5 × 10⁶ PFU of either RCMV or RCMVR78c. On days 4 and 21 p.i., five rats from each group were sacrificed, and their internal organs were collected. These organs were subjected to both plaque assay and immunohistochemistry (19). Tissue sections (4 µm) of the submaxillary salivary gland, spleen, kidney, liver, lung, heart, pancreas, thymus, aorta, and cervical lymph nodes were stained with MAb RCMV 8.

Survival of immunocompromised rats infected with either RCMV, RCMVR78a, or RCMVR78c. Four-week-old rats (100 to 120 g) were divided into three groups of five rats. Intraperitoneal infection was carried out with 10⁶ PFU of either RCMV, RCMVR78a, or RCMVR78c. The number of surviving rats was recorded daily until day 28 p.i.

Nucleotide sequence accession number. The nucleotide sequence of the 3.7-kb BamHI fragment containing the R78 gene (Fig. 1) and the predicted amino acid sequence derived from R78 have been deposited in the GenBank database under accession no. AF077758.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Identification, cloning, and sequence analysis of the RCMV R78 gene. Previously, it was shown that the majority of RCMV genes are collinear with genes of both HCMV and MCMV (9-11, 61, 62). We hypothesized that the position of a putative RCMV UL78 homolog would be analogous to that of both HCMV UL78 (22) and MCMV M78 (54). Accordingly, we focused on the 33-kb RCMV XbaI B fragment (Fig. 1) (45). This fragment was digested with various restriction endonucleases, subcloned, and sequenced. The GenBank database was subsequently screened for homology with the generated sequence. Thus, we identified (in plasmid p116) a 3.7-kb BamHI fragment showing considerable similarity to a region within the genomes of HCMV as well as MCMV, containing the UL78 and M78 genes, respectively. A 1,422-bp ORF was identified within the BamHI fragment (Fig. 1), which has the potential to encode a 474-amino-acid polypeptide with a predicted molecular mass of 50 kDa. This polypeptide shows 25 and 20% similarity with the amino acid sequences of M78 (54) and UL78 (22), respectively (Table 1). This low level of similarity is not uncommon to UL78-like sequences, since the amino acid sequences derived from UL78 and M78 share only 21% similarity (Table 1). In analogy to the nomenclature for the corresponding HCMV and MCMV genes, the 1,422-bp RCMV ORF was termed R78.

View larger version (16K): [in this window] [in a new window]   FIG. 2.   Schematic representation of the gene products of RCMV R78, MCMV M78, HCMV UL78, and HHV-6 and -7 U51. The top drawing shows the two-dimensional orientation of GCRs (not to scale with diagrams of the other GCRs). Horizontal lines represent the one-dimensional structure of the R78-like gene products; designations of the genes that encode these receptors are shown at the left. Black boxes represent putative transmembrane  helices; the white box indicates an additional unique hydrophobic -helix domain predicted by the computer program TMpred (58); the hatched box indicates a hydrophobic stretch of amino acid residues collinear with predicted transmembrane domains of the other GCRs, although not predicted by TMpred. Conserved cysteine residues are indicated by C's. The (D/E)R(I/L) motifs indicate positions of the conserved putative G protein-coupling domains. The sequences enclosed in white boxes indicate positions of putative tyrosine kinase phosphorylation sites that are conserved among CMV genes relative to R78.

R78 transcription. Although the ORFs of R78-like betaherpesvirus genes are poorly conserved, they are all preceded by a remarkable double TATA signal motif. Moreover, the start codons of R78, M78, and UL78 conform to the Kozak consensus (41). The positions and sequences of these motifs are summarized in Table 2. We set out to identify transcripts derived from the RCMV R78 gene by Northern blot analysis. As shown in Fig. 3, R78-specific hybridization signals were detected in phosphonoacetic acid-treated (lane 3) and untreated (lane 4) cells but not in cycloheximide-treated (lane 2) or mock-infected (lane 1) cells. In untreated, RCMV-infected REF, R78-specific signals were detected at 8 and 48 h p.i. (lanes 7 and 8, respectively) but not at earlier time points (lane 5 and 6, respectively). These results classify R78 as an E gene that is also transcribed during the L phase of infection in vitro.

View larger version (77K): [in this window] [in a new window]   FIG. 3.   The RCMV R78 gene is transcribed at both E and L times of infection in REF. Lanes 1 and 5 of the autoradiograph of a Northern blot hybridized with an R78-specific probe represent mRNA from mock-infected (M) cells. Lanes 2 to 4 represent the IE, E, and L phases of infection, respectively; lanes 6 to 8 represent transcripts from REF at 2, 8, and 48 h p.i., respectively. The estimated lengths of the different transcripts are indicated at the right in kilobases.

Generation of RCMV strains with a deletion of the R78 gene. To investigate the role of R78 in the pathogenesis of RCMV disease, we constructed a mutant RCMV strain (RCMVR78a) in which a 1,030-bp BglII-SalI fragment from the R78 gene was deleted and replaced with a 1.5-kb neo expression cassette (Fig. 4). Another RCMV mutant strain (RCMVR78c) was constructed such that an 80-bp SalI fragment was replaced by the neo cassette (Fig. 4). Consequently, the part of R78 that encodes the putative intracellular C terminus was deleted and replaced by an irrelevant sequence of similar length, while the part that encodes the N terminus (including the seven predicted transmembrane helices) was preserved (Fig. 4). The deletion/insertion mutations were first introduced into plasmids containing the R78 gene (Fig. 4). The R78 gene within the RCMV genome was subsequently replaced by either of the mutated R78 genes via homologous recombination, after transfection of fibroblasts with the recombination plasmids and infection with RCMV. Selection for recombinant strains was established by supplementing the growth medium with G418. After plaque purification of the virus, the purity and integrity of the recombinant strains were checked by Southern blot analysis. Virion DNA from RCMV, RCMVR78a, and RCMVR78c was purified and digested with BamHI-NcoI (Fig. 5). We also digested purified virion DNA with either BamHI, BamHI-BglII, BamHI-EcoRV, NcoI, SalI, or SalI-EcoRV (data not shown). After agarose gel electrophoresis and transfer of the DNA to a filter, hybridization was done with either an R78-specific probe or a neo-specific probe. As shown in Fig. 5, the observed hybridization signals correspond exactly to the predicted restriction fragments. In addition, contaminating wt virus fragments were not detected in the RCMVR78a and RCMVR78c lanes (Fig. 5). These findings indicated that both recombinant strains are pure and contain the appropriate mutations.

View larger version (22K): [in this window] [in a new window]   FIG. 4.   Construction of RCMV strains in which the R78 gene is disrupted. The RCMV genome, of which a segment representing the R78 region is shown in the middle, was modified by homologous recombination with different recombination plasmids (p147 [top] or p120 [bottom]), resulting in RCMVR78a or RCMVR78c, respectively. Approximately 85% of the R78 ORF has been deleted in the RCMVR78a genome (resulting in a gene encoding only the N-terminal part including half of the first predicted transmembrane domain, shown at the top right). In mutant RCMVR78c, the part of the R78 that encodes the putative intracellular C terminus has been replaced by a sequence encoding a stretch of 97 amino acids of irrelevant sequence (bottom right, indicated as a black arrow box). ORFs are shown as arrow boxes. Wild-type and mutated R78 ORFs are indicated with descending hatches. The neo genes that were inserted in the recombination plasmids are indicated with ascending hatches.

View larger version (34K): [in this window] [in a new window]   FIG. 5.   Southern blot analysis of recombinant viruses RCMVR78a and RCMVR78c. (A) DNA from RCMV, RCMVR78a, and RCMVR78c was digested with BamHI and NcoI, electrophoresed, blotted, and hybridized with either a probe derived from the p116 insert or a neo probe (both indicated with black boxes). ORFs are shown as arrow boxes. Wild-type and mutated R78 ORFs are indicated by descending hatches. The neo genes that were inserted in the recombination plasmids are indicated by ascending hatches. (B) Autoluminograph of a Southern blot containing wt (w) DNA, RCMVR78a (a) DNA, and RCMVR78c (c) DNA. The estimated lengths of BamHI-NcoI-digested DNA fragments are indicated at the sides in kilobases.

RCMVR78c transcripts. The deletion of the R78 ORF might invoke unforeseen events such as disruption of unknown promoter/enhancer regions of the neighboring gene R77, R79, or R80. To investigate the effect of modification of the R78 genes on transcription of these neighboring genes, a Northern blot hybridization experiment was performed with poly(A)⁺ RNA from REF infected with one of the two recombinant viruses (RCMVR78c) and wt RCMV-infected REF. The genes upstream and downstream of the R78 gene were found to have considerable sequence similarity to the MCMV M77, M79, and M80 genes (54), respectively (data not shown). These RCMV genes are therefore referred to as R77, R79, and R80, respectively. Probes were generated from R77-, R78-, R79-, R79/R80-, and neo-specific DNA fragments (Fig. 6A and B) and hybridized with poly(A)⁺ RNA extracted from either mock-, RCMV-, or RCMVR78c-infected fibroblasts at 48 h p.i. (Fig. 6C).

View larger version (37K): [in this window] [in a new window]   FIG. 6.   Transcription of R78 and neighboring genes. The wt RCMV (A) and RCMVR78c (B) genomes are represented by lines. ORFs are indicated by white, hatched, and black arrow boxes, probes are indicated by black boxes, and transcripts are indicated by black arrow boxes below the genomes. Lengths are shown at the sides in kilobases. (C) Autoluminographs from Northern blots that contain poly(A)+ RNA from virus-infected REF. The transcript lengths that correspond to the detected hybridization signals are indicated in kilobases. M, mock infected; W, wt RCMV infected; , RCMVR78c infected.

The R78 gene is important for efficient virus replication in various cell types in vitro. The replication characteristics of RCMV, RCMVR78a, and RCMVR78c were assessed in four different cell types: REF RHEC, RSMC, and R2M. Cells were infected with either RCMV, RCMVR78a, or RCMVR78c, and the proportion of infected cells relative to the total population of cells was determined at various time points. Additionally, the amount of excreted infectious virus was determined for each cell type. As shown in Fig. 7, the infected cell ratios did not differ significantly among REF and RSMC infected with either wt RCMV or each of the recombinant viruses during the observed time course. Also, infected cell ratios as well as virus titers in the culture medium of RHEC and R2M infected with either wt or recombinant virus did not differ significantly (data not shown). However, clear differences were observed in virus titers in the culture medium of virus-infected REF and RSMC. In particular, at days 5 and 7 p.i., virus titers were 10- to 100-fold lower in the culture medium of both RCMVR78a- and RCMVR78c-infected REF than in the culture medium of wt RCMV-infected REF (Fig. 7). Additionally, at day 5 p.i., virus titers were 10-fold lower in the culture medium of recombinant virus-infected RSMC than in the medium of wt virus-infected RSMC (Fig. 7). Taken together, these data suggest that (i) R78-deleted viruses enter cells and spread throughout the monolayers of cells at similar rates as wt virus and (ii) in REF and RSMC, the production of recombinant virus is less efficient than the production of wt virus.

View larger version (28K): [in this window] [in a new window]   FIG. 7.   RCMV is attenuated upon deletion of the R78 gene. REF were infected at an MOI of 0.01 with either wt RCMV (), RCMVR78a (), or RCMVR78c (). RSMC were infected with wt and mutant virus at an MOI of 1 (relative to REF infection). The upper graphs show the infected cell/total cell ratios at various time points p.i.; the lower graphs show virus titers determined in culture medium up to 7 days p.i. Standard deviations are indicated by vertical bars.

RCMVR78a and RCMVR78c induce syncytium formation in REF. In addition to the differences in virus replication between wt and mutant viruses, the morphology of REF infected with either RCMVR78a or RCMVR78c clearly contrasted with that of wt virus-infected REF. Within the monolayers of mutant RCMV-infected REF (Fig. 8C and D) but not in those of infected RSMC, RHEC or R2M, we observed syncytium-like cellular cultures that were larger than those typically seen in wt RCMV-infected REF monolayers (Fig. 8B). These structures appear 3 to 4 days p.i. To investigate the nature of these structures more closely, they were subjected to immunofluorescence staining and confocal laser scan microscopy. To this purpose, virus-infected REF were stained with MAb RCMV 8, which detects E-phase-expressed antigens in the nuclei (20). Additionally, these cells were counterstained with phalloidin, which binds to cytoskeletal F-actin fibers localized in the cytoplasm. MAb RCMV 8 and phalloidin were labeled with fluorescent markers fluorescein isothiocyanate and rhodamine, respectively. Surprisingly, the large structures that were seen in either RCMVR78a- or RCMVR78c-infected REF appeared to consist of a single cytoplasmic entity encapsulating 2 to 20 distinct nuclei. A typical example of such a syncytium-like structure is shown in Fig. 8E. The significance of these polykaryotic cells, or syncytia, and their relation to the function of the R78 gene is discussed below.

View larger version (277K): [in this window] [in a new window]   FIG. 8.   RCMV R78 deletion mutants induce syncytium formation in vitro. Immunofluorescence micrographs (magnification of ×400) show uninfected REF (A) and REF infected with either wt RCMV (B), RCMVR78a (C), or RCMVR78c (D). (E) Confocal laser scan micrograph taken from a syncytium structure in a monolayer of RCMVR78c-infected REF. The scale bar indicates 20 µm. The left-hand frame shows intracellular F-actin fibers that were stained with phalloidin-rhodamin (Eugene, Leiden, The Netherlands); the right-hand frame, of the same microscopic view, shows the nucleic distribution of RCMV E antigens detected by MAb RCMV 8 plus anti-mouse-fluorescein isothiocyanate.

R78 has a critical function in the pathogenesis of RCMV infection in vivo. The role of R78 in the pathogenesis of RCMV disease was investigated by infection of rats with either wt or mutant virus. To compare virus dissemination in wt virus-infected and mutant virus-infected rats, two groups of immunocompromised rats were infected with 5 × 10⁶ PFU of either RCMV or RCMVR78c. At days 4 and 21 p.i., the presence of virus in internal organs (aorta, heart, kidney, liver, lung, lymph nodes, pancreas, salivary glands, spleen, and thymus) of the infected rats was determined. At day 4 p.i., virus could be detected in 18% of the aforementioned organs in wt RCMV-infected rats and in 9% of the organs of RCMVR78c-infected rats, as determined by either plaque assay or immunohistochemistry. At this time point, the highest virus titers were found in the spleens of wt virus- or mutant virus-infected rats. In contrast to the marked differences in replication efficiency between wt and deletion mutant virus in vitro, the differences in the amount of virus recovered from either spleen (Table 3) or other organs (not shown) of infected rats were less dramatic. Neither wt nor mutant virus could be detected in the liver, pancreas, and salivary glands at day 4 p.i. At day 21 p.i., virus could be recovered from salivary glands of both wt RCMV- and RCMVR78c-infected rats but not from any of the other organs analyzed. Surprisingly, virus titers in salivary glands did not differ significantly between RCMV- and RCMVR78c-infected rats (Table 3).

View this table: [in this window] [in a new window]   TABLE 3.   Virus titers and immunohistochemical detection of RCMV and RCMVR78c in salivary glands and spleen tissue

View larger version (12K): [in this window] [in a new window]   FIG. 9.   The R78 gene plays a vital role in RCMV pathogenesis in vivo. The graph indicates survival of three groups of immunocompromised rats after intraperitoneal inoculation with 106 PFU of either wt RCMV (), RCMVR78a (), or RCMVR78c (). Survival was recorded up to day 28 p.i.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Herpesviruses are known for the large size and complexity of their genomes, which encompass a vast number of genes, many of which are homologous to genes of the host organism. Beta- and gammaherpesvirus genomes contain genes homologous to GCR genes of the host. Eighteen beta- and gammaherpesvirus-encoded GCR genes have been recognized to date (4, 9, 21-23, 25, 33, 44, 49, 48, 54, 57, 63). These genes can be arranged into four groups based on sequence homology, genome location, and function of the respective gene product. One group (the US28 family) consists of two GCR genes: US27 and US28 (23). Both are exclusively present within an unconserved region of the HCMV genome (22). These GCRs are highly similar to mammalian chemokine receptors (15, 30). Although little is known about the function of the GCR encoded by US27, the HCMV US28-encoded GCR is a well-studied example of a virus-encoded GCR with an immunomodulatory function. This receptor was shown to bind  chemokines MIP1, MIP1, RANTES, MCP-1, and MCP-3 (14, 31, 47). Additionally, Bodaghi et al. reported that the US28 gene product modifies the chemokine environment of HCMV-infected cells through sequestering of  chemokines by continuous internalization (15). A second group (the UL33 family) contains five GCR genes: HCMV UL33 (23), RCMV R33 (9), MCMV M33 (26), and HHV-6 and -7 U12 (33, 49). Both sequence and position of these genes within their respective genomes are conserved. Similar to the US28 family, GCRs of the UL33 family were shown to be related to chemokine receptors. Previously, we reported that the predicted amino acid sequences of UL33-like receptors share more similarity with mammalian chemokine receptors than with nonchemokine receptors (9). Moreover, Isegawa et al. (38) demonstrated that one member of the UL33 family of GCRs, HHV-6 U12, is a functional -chemokine receptor capable of binding MIP1, MIP1, RANTES, and MCP-1 (38). Although not much is known about the function of this chemokine receptor-like family, Margulies et al. detected the UL33 protein in both the membranes of HCMV-infected cells and the envelopes of HCMV virions (43a). Additionally, both RCMV R33 and MCMV M33 genes were shown to be essential for RCMV and MCMV replication in salivary glands of infected rats and mice, respectively (9, 26). A third group (the gammaherpesvirus GCR family) consists of seven gammaherpesvirus-encoded chemokine receptor-like genes: Epstein-Barr virus BILF1 (5, 25), HVS ECRF3 (48), KSHV ORF 74 (4), murine gammaherpesvirus 68 ORF 74 (63), and equine herpesvirus E1, E6, and E8 (57). In contrast to GCR genes of the US28 and UL33 family, sequences of the gammaherpesvirus GCR family are less well conserved. However, like GCR genes of the US28 and UL33 family, they show significant similarity with mammalian chemokine receptors (4, 26, 48, 57, 63). The role of any of these GCRs in replication or persistence of gammaherpesviruses is unknown. To date, two members of this family have been studied in detail. The HVS ECRF3-encoded GCR was shown to be capable of binding  chemokines interleukin-8, GRO/MGSA, and NAP-2 (1). Another gammaherpesvirus-encoded GCR, the KSHV ORF 74 gene product, was demonstrated to bind both  (interleukin-8, MGSA, NAP-2, and PF-4) and  (I-309 and RANTES) chemokines. In addition, the ORF 74 protein was found to be constitutively active (4), having both oncogenic and angiogenic potential (7). A fourth, novel group of GCR genes (the UL78 family) comprises five betaherpesvirus ORFs that were recently recognized as putative GCR genes: HCMV UL78, RCMV R78, MCMV M78, and HHV-6 and -7 U51 (references 33, 49, and 54 and this report). Similar to the positions of GCR genes of the UL33 family, the positions of the UL78-like genes are conserved within betaherpesvirus genomes. However, in contrast to the sequences of the UL33-like genes, the sequences of the UL78 family are poorly conserved. The predicted amino acid sequences derived from the UL78-like genes significantly resemble neither chemokine receptors nor any other of the thousands of GCRs currently known. The assumption that UL78-like genes are GCRs is based on three characteristics: (i) the presence of seven hydrophobic regions within the predicted amino acid sequences derived from all UL78-like genes (53); (ii) the presence of two cysteine residues within these amino acid sequences, which might be required for correct folding of the GCR polypeptide (53); and (iii) a stretch of amino acids within these amino acid sequences which bears similarity to a domain known to be required for G protein coupling (53). Consequently, the UL78 family is a novel class of orphan GCRs encoded by betaherpesviruses.

To investigate the role of the R78 gene in virus replication, RCMV strains in which the R78 gene is disrupted were constructed. These R78 deletion mutant strains were found to replicate 10- to 100-fold less efficiently than wt RCMV in fibroblasts and smooth muscle cells in vitro. By contrast, RCMV and MCMV strains that carry a deletion of the R33 and M33 gene, respectively, were previously found to replicate in vitro with efficiencies similar to those of the corresponding wt viruses (9, 26). A difference between these mutants and the R78 deletion mutants was also seen in vivo: both R33 and M33 deletion mutants were found to be impaired in replication in salivary glands of infected animals, whereas R78 deletion mutants were demonstrated to replicate as efficiently as wt RCMV in salivary glands. It is therefore likely that although both R33 and R78 putatively encode GCRs, these genes exert unrelated functions.

A lower efficiency of virus replication in vitro was observed not only after deletion of the complete R78 ORF from the RCMV genome but also after deletion of the region that encodes the putative R78 GCR C terminus. This putative intracellular part of the protein is therefore likely to be essential for the function of the R78 protein. The C terminus, like the DRL motif in the second intracellular domain of the R78 protein, might be essential for G protein coupling and signal transduction, similar to what was found for other GCRs (53).

The RCMV R78 gene plays an important part in the pathogenesis of virus infection in vivo. This was inferred from the lower mortality seen among immunocompromised rats infected with either RCMVR78a or RCMVR78c than among animals infected with wt virus. Although we did not find significant differences in virus replication between recombinant and wt viruses in vivo, it is possible that the decrease in virulence that is seen after disruption of the R78 gene of RCMV is correlated with the observation that both RCMVR78a and RCMVR78c replicate less efficiently in fibroblasts and smooth muscle cells in vitro than wt virus. Similar to the R78 gene, the R33 GCR-like gene was reported to have a vital function in the pathogenesis of RCMV infection (9). It is likely that the HCMV counterparts of these genes (UL78 and UL33, respectively) have similar, important functions in the pathogenesis of HCMV infections in humans. The gene products of both UL78 and UL33 can therefore be considered as potential targets for future development of novel antiviral strategies.

The two recombinant RCMV strains described in this report induce syncytium formation in infected fibroblasts in vitro. Previously, syncytium formation has been observed in many different cell types infected with a variety of herpesvirus species: HCMV-infected human amnion cells (28), varicella-zoster virus-infected human melanoma cells (34), Epstein-Barr virus-superinfected Raji cells (8), HHV-6-infected human primary fetal astrocytes (35), HHV-7-infected T lymphocytes (56), and RCMV-infected Rat2 cells (11). In these cases, syncytium formation appears to be associated with the cell type or MOI, since the same virus strains fail to produce these syncytium (syn) phenotypes upon infection of other permissive cell lines (8, 28, 34, 35) or after infection at lower titers (8, 56). Well-defined syn loci were found within genomes of a limited number of herpesvirus species, in particular within the genomes of herpes simplex type 1 (HSV-1) strains: UL20 (6), UL24 (40), UL27-gB (29), and UL53-gK (37). Some HCMV genes were likewise associated with syncytium formation. Syncytia are produced in human glioblastoma cells that constitutively express HCMV UL55-gB (59). Additionally, the chemokine receptor encoded by HCMV US28 was shown to enhance cell-cell fusion in cells constitutively expressing retroviral envelope glycoproteins (52). The recombinant RCMV strains RCMVR78a and RCMVR78c are the first syn mutant CMVs reported. We postulate that functional R78-encoded GCRs transduce signals to the cellular interior, thereby creating an intracellular environment essential for the formation of protein complexes to establish intercellular junctions. In HCMV- and HSV-1-infected cells, these cell-to-cell junctions are basically composed of viral glycoproteins (27, 36, 43). Since many syn mutations are localized in HSV glycoprotein genes (29, 37), syncytia could be generated as a result of destabilized cell-to-cell contacts. The putative GCR encoded by R78 may mediate stabilization of glycoprotein complexes, directly by association with these glycoprotein complexes, or indirectly, through signal transduction to establish cell-to-cell contacts. These possible mechanisms for the function of the R78 gene product will have to be addressed in future studies.