The human cytomegalovirus UL78 gene is highly

Journal of General Virology (2005), 86, 297–306

DOI 10.1099/vir.0.80436-0

The human cytomegalovirus UL78 gene is highly conserved among clinical isolates, but is dispensable for replication in fibroblasts and a renal artery organ-culture system Detlef Michel,1 Irena Milotic´,2 Markus Wagner,3 Bianca Vaida,1 Jens Holl,1 Ramona Ansorge1 and Thomas Mertens1 Correspondence

1

Abteilung Virologie, Universita¨tsklinikum, Albert-Einstein-Allee 11, 89081 Ulm, Germany

Thomas Mertens

2

thomas.mertens@medizin.

3

uni-ulm.de

Received 9 July 2004 Accepted 19 October 2004

Clinic for Infectious Diseases, Clinical Hospital Centre, Cambierieva 17, 51000 Rijeka, Croatia Harvard Medical School, NRB 836, Department of Pathology, 77 Avenue Louis Pasteur, Boston, MA 02115, USA

The human cytomegalovirus (HCMV) UL78 ORF is considered to encode a seven-transmembrane receptor. However, neither the gene nor the UL78 protein has been characterized so far. The objective of this study was to investigate the UL78 gene and to clarify whether it is essential for replication. UL78 transcription was activated early after infection, was inhibited by cycloheximide but not by phosphonoacetic acid, and resulted in a 1?7 kb mRNA. Later in the replication cycle, a second mRNA of 4 kb evolved, comprising the UL77 and UL78 ORFs. The 59 end of the UL78 mRNA initiated 48 bp upstream of the translation start and the polyadenylated tail started 268 bp downstream of the UL78 translation stop codon within the UL79 ORF. By using bacterial artificial chromosome technology, a recombinant HCMV lacking most of the UL78 coding region was constructed. Successful reconstitution of the UL78-deficient virus proved that the gene was not essential for virus replication in fibroblasts. The deletion also did not reduce virus replication in ex vivo-cultured sections of human renal arteries. Analysis of viral proteins at different stages of the replication cycle confirmed these results. Among clinical HCMV isolates, the predicted UL78 protein was highly conserved. However, an accumulation of different single mutations could be found in the N-terminal region and at the very end of the C terminus. Due to the absence of an in vivo HCMV model, the role of UL78 in the pathogenesis of HCMV infection in humans remains unclear.

INTRODUCTION In human cytomegalovirus (HCMV), the ORFs US27, US28 and UL33 encode putative G protein-coupled receptor (GPCR) homologues (Chee et al., 1990). A fourth ORF, UL78, is also thought to encode a GPCR. However, no experimental data exist concerning this function, so it is currently adequate to consider UL78 as a seventransmembrane (7TM) protein (Gompels et al., 1995). GPCRs play key roles in transduction of extracellular signals into the cell. Activation of these receptors most often requires binding of ligands, e.g. cytokines. The US28 protein (pUS28) has been shown to be a calciummobilizing receptor for several b-chemokines (Gao et al., 1993) and has been suggested to be responsible for b-chemokine sequestration in HCMV-infected fibroblasts in vitro (Bodaghi et al., 1998). It may also induce migration of infected smooth muscle cells (Streblow et al., 1999). It has been shown that pUS28 mediates constitutive, 0008-0436 G 2005 SGM

Printed in Great Britain

ligand-independent, pertussis toxin-insensitive activation of the phospholipase C Gq/11-dependent signalling pathway (Casarosa et al., 2001), and neither RANTES nor MCP-1 had any effect on this activity. We have shown that infection of fibroblasts with HCMV led to a vigorous, constitutively enhanced formation of inositol phosphates, which was insensitive to pertussis toxin (Minisini et al., 2003). The UL33 protein has been described as non-essential for growth in fibroblasts in vitro (Margulies et al., 1996). However, it has recently been shown that its homologue in rat cytomegalovirus (RCMV), the R33 protein (pR33), is important for viral pathogenesis. Unlike wild-type virus, an RCMV R33-deficient mutant could neither enter into, nor replicate in, the salivary-gland epithelial cells of infected rats (Beisser et al., 1998). A similar observation was made for mouse cytomegalovirus (MCMV) M33. Upon infection of mice with an MCMV M33-deletion mutant, virus could not be recovered from mouse salivary glands (DavisPoynter et al., 1997). It has also been shown that pM33 297

D. Michel and others

exhibits constitutive signalling activity (Waldhoer et al., 2002). We have reported that, like pR33, pUL33 activates multiple signalling pathways in a ligand-independent manner (Casarosa et al., 2003). Although both receptors constitutively activate phospholipase C via Gq/11-mediated, and partially Gi/o-mediated, pathways, they exhibit profound differences in the modulation of cAMP-responsive element (CRE) activation – pR33 constitutively inhibits, whereas pUL33 constitutively enhances CRE-mediated transcription. The genomes of the b-herpesviruses human herpesvirus (HHV)-6A, HHV-6B and HHV-7 contain homologues of two 7TM receptors, namely U12, which is homologous to HCMV UL33, and U51, which is homologous to HCMV UL78 (Gompels et al., 1995; Nicholas, 1996; Megaw et al., 1998). Interestingly, both the UL33and UL78-like genes exhibit conservation of position, as well as orientation, within the genomes of b-herpesviruses. Until now, six UL78-like genes had been identified: HCMV UL78, MCMV M78, RCMV R78 and the U51 ORFs of HHV-6A, HHV-6B and HHV-7 (Gompels et al., 1995; Nicholas, 1996; Rawlinson et al., 1996; Beisser et al., 1999). HHV-6 U51 is known to be expressed on the surface of infected T cells and has been found to bind RANTES/ CCL5 with nanomolar affinity, as well as to downregulate the transcription of this chemokine (Menotti et al., 1999; Milne et al., 2000). It has been reported that deletion of the R78 ORF in RCMV confers an attenuated syncytiuminducing phenotype (Beisser et al., 1999). MCMV pM78 led to an accumulation of viral mRNAs and is a constituent of the virion (Oliveira & Shenk, 2001). To our knowledge, this is the first study of the HCMV UL78 gene. We identified the transcription start point and the start of the polyadenylated tail. We found differences in the regions carrying polyadenylation signals for UL78-like mRNAs in rodent and human CMVs. Specifically, we generated and characterized a UL78-deficient mutant and showed that the absence of the UL78 ORF hampered neither virus replication in fibroblasts or an organ-culture system, nor the expression of different viral proteins in fibroblasts, although the predicted pUL78 was quite conserved among clinical HCMV isolates.

METHODS Cells and viruses. HCMVs were propagated in human foreskin fibroblasts (HFFs) as described elsewhere (Michel et al., 1996, 1998). Eleven clinical isolates from nine patients from different clinics (one newborn, four children aged 1–5 years and four adults aged 38–77 years) were taken from our diagnostic virus laboratory from the second passage of HFFs after the appearance of the first cytopathic effect. The isolates derived from different specimens (urine, throat wash and leukocytes). Virus titration. Quantification of viral titres was performed by

standard plaque assays with end-point dilutions in triplicate. Virus yield assays (single-cycle replication) were performed on HFFs. Briefly, HFFs were grown in 48-well plates and infected 2 days later. Following adsorption, the monolayers were washed and fresh culture medium was added. At various days post-infection (p.i.), cell-culture 298

supernatants were harvested and titrated in triplicate on HFFs. Cells were fixed and stained with a mAb against the viral pp65 (pUL83). Viral titres were determined by counting the number of antigenpositive cells. Plasmid construction and bacterial artificial chromosome (BAC) mutagenesis. Plasmid cloning was performed by using

standard methods (Sambrook et al., 1989). Construction of the shuttle plasmid pSTdelUL78 was done by insertion of DNA fragments amplified with the Expand High Fidelity PCR system (Roche). For generation of the shuttle plasmid, two regions were cloned separately. Firstly, a fragment of 1?97 kbp containing 113 bp of the 59-UL78-coding region was amplified with the primer pair Hind-78 (59-CGGGTGCGCAAACGGaATtcGCGTCAG-39) and Xba78a (59-TAGCCGATGtcTAGATCCGAGGCGG-39) [nucleotides mutated to create the restriction sites (underlined) are in lower case], cut with HindIII and XbaI and inserted into pUC19, resulting in plasmid pUCUL78-59. Secondly, a 2?1 kbp fragment containing 159 bp of the 39-UL78-coding region was amplified with primers Xba78b (59-AACCACGTtctAGAAAGCCACGTTG-39) and Eco-78 (59-GACGGCGGgaATTCCGCGGGAAGA-39), cut with XbaI and EcoRI and subcloned into XbaI/EcoRI-digested pUCdel78-59. From the second plasmid, the entire insert was excised by EcoRI/HindIII digestion and inserted into the shuttle plasmid pST76K (Po´sfai et al., 1997), resulting in pSTdelUL78. Generation and reconstitution of the HCMV recombinant.

Mutagenesis of the HCMV BAC plasmid pHB5 with the shuttle plasmid pSTdelUL78 was performed in Escherichia coli strain CBTS as described previously (Borst et al., 1999; Wagner et al., 2000). Selection and identification of BAC plasmids lacking most of the UL78-coding region were performed as described previously (Wagner et al., 2000). Maxi-preparations of BAC plasmid DNA and reconstitution of infectious virus were performed as described elsewhere (Borst et al., 1999; Casarosa et al., 2003). As an important control, we used the virus RVHB5, obtained from transfection and reconstitution of BACmid pHB5. Cells were passaged 7 days after transfection and cultured until cytopathic effect became visible. Viral nucleic acid isolation and analysis. Fibroblasts were

infected at an m.o.i. of 1 and harvested 3 days p.i. Extraction of total DNA from infected cells and Southern blot analysis were performed as described previously (Wagner et al., 2000; Casarosa et al., 2003). The probe used represented nt 112520–112921 of the HCMV genome, comprising the 39-terminal region of the UL77 ORF and the 59 region of the UL78 ORF (AD169 strain, GenBank accession no. X17403). Primer extension. Primer extension was performed as described

previously (Michel et al., 1993). Briefly, HFFs were infected with HCMV strain AD169 at an m.o.i. of 1. Cells were harvested at 24 h p.i. and total RNA was extracted by using a Midi RNA Extraction kit (Qiagen) as recommended by the manufacturer. Total RNA (20 mg) was annealed for 5 h to 100 ng of a UL78-specific oligonucleotide (59-CCGACATAGAGTAGCCTTGC-39) spanning nt 113026–113007, labelled with [32P]ATP by using T4 kinase. The nucleic acid was precipitated and resuspended in 300 ml 50 mM Tris/HCl (pH 8?3), 50 mM KCl, 6 mM MgCl2, 100 mM each dNTP, 5 mM dithiothreitol (DTT), 30 mg actinomycin ml21, 50 mg BSA ml21, 1 U RNasin ml21 and incubated for 30 min at 42 uC with 67 U avian myeloblastosis virus (AMV) reverse transcriptase. The extension products were resuspended in 10 ml double-distilled water and separated on a 6 % polyacrylamide gel. Identification of the UL78 3§ end. For cDNA synthesis, total

RNA from HCMV-infected fibroblasts was hybridized with the primer mix 59-GTAAAACGACGGCCACTTTTTTTTTTTTTTTTTN39, where N stands for A, G or C. The 59 region of this primer is Journal of General Virology 86

Characterization of HCMV UL78 complementary to the M13 universal primer. The reaction mix consisted of 3 mg RNA, 10 mM primer, 1/5 vol. first-strand buffer (Gibco-BRL), 200 mM each dNTP, 50 mM DTT, 100 U RNasin (Promega) and AMV reverse transcriptase. After incubation for 30 min at 37 uC, synthesis was stopped by heating to 94 uC. A 5 ml aliquot of the reaction mix was used for amplification with the M13 universal primer and the UL78-specific primer 1160 (59-TCTGGTCGAGAGATGTCAGC-39) or 1360 (59-TGTTGGTAACGACAACCACG-39), hybridizing 355 and 159 bp, upstream of the translation stop codon, respectively. Amplicons were sequenced and the obtained sequences were aligned with the sequence of strain AD169. Sequencing of UL78 regions from clinical isolates. Total

DNA was extracted from infected fibroblasts by proteinase K digestion as described previously (Michel et al., 2001). UL78 sequences were amplified by PCR. Primers were designed on the basis of data from strain AD169 (GenBank no. CAA35351.1). For amplification, the primers 59-GGGTATATTCGTTCGGCGAG-39 and 59-TATCTGCCACTTTTCTCCCCG-39 were used, both hybridizing outside the UL78-coding region. All amplifications were done in duplicate. PCR was performed as follows: 2 min at 94 uC, followed by 35 cycles of 15 s at 94 uC, 15 s at 58 uC and 120 s at 72 uC, using a GeneAmp 9700 PCR cycler (Perkin-Elmer) and the Expand High Fidelity PCR system (Roche). Amplification products were purified by using a GFX PCR DNA and Gel Band Purification kit (Amersham Biosciences). Direct sequencing was performed as described previously (Michel et al., 2001). Briefly, for sequencing, an ABImed Cycle Sequencing kit (Applied Biosystems) was used according to the manufacturer’s instructions (Roche Diagnostics) on the GeneAmp 9700 PCR cycler (Perkin-Elmer), performing the program as recommended by the manufacturer and using the sequencing primers UL78-675P (59-AGACGCGCGAAAACGCCGG-39), UL78-1195M (59-GAATGCCAAGACGCATGGTG-39), UL78-250M (59-CCCTTGGACAACATGGTGG-39), UL78-500 (59-ATAGGGATCTGACGACAGCCTAGC-39), UL78-outP (59-GGGTATATTCGTTCGGCGAG39), UL78-outM (59-TATCTGCCACTTTTCTCCCCG-39), 1160 (59TCTGGTCGAGAGATGTCAGC-39), 1360 (59-TGTTGGTAACGACAACCACG-39) and UL78-970 (59-CTACTACTGCTTCAGAGTCCT39). DNA sequences were separated in a 310 Genetic Analyser (Applied Biosystems). Western blot analysis. Western blot analysis was performed as

described previously (Michel et al., 1996, 1998) using mouse mAbs directed against the HCMV proteins IE1, pUL69, pp65 and pp28 (kindly provided by M. Mach, Universita¨t Erlangen, Germany) and a polyclonal antiserum against pUL97 (Michel et al., 1996). Briefly, fibroblasts were infected at an m.o.i. of 1 with the indicated virus strains and total cell lysates were extracted at different time points p.i. SDS-PAGE was performed according to standard protocols. Proteins were visualized with an enhanced chemiluminescence (ECL) reaction according to the manufacturer’s recommendations (Amersham Biosciences). Northern blot analysis. HFFs were infected at an m.o.i. of 1 and

harvested at the indicated time points. To arrest virus replication in the immediate-early or early stage, the metabolic inhibitors cycloheximide (CHX, 100 mg ml21) and phosphonoacetic acid (PAA, 250 mg ml21) were added to the infected fibroblasts as described previously (Michel et al., 1996). Total RNA was extracted by using a Midi RNA Extraction kit (Qiagen). Total RNA (10 mg) was separated as described by Sambrook et al. (1989). RNA was transferred on to a nylon membrane and immobilized by incubation for 2 h at 60 uC. For subsequent hybridization, DNA probes specific for UL78 and UL77 or for the cellular glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene were amplified by PCR. Membranes were stripped for 20 min with 0?16 SSC and 0?1 % SDS for rehybridization. Probes were labelled with [32P]dCTP by using a random priming kit (Pharmacia). Hybridization was carried out at 42 uC for http://vir.sgmjournals.org

18 h in the presence of 50 % formamide, 26 Denhardt’s solution, 56 SSPE, 0?05 % SDS and 200 mg salmon sperm DNA ml21. Filters were washed to a stringency of 0?26 SSC, 0?1 % SDS at 65 uC. Renal artery organ-culture system. Human renal arterial

segments were obtained after informed consent of the patients. Segments were cultivated according to Voisard et al. (1999) and Reinhardt et al. (2003) in a mixture of Ham’s F-12/L-glutamine medium and Waymouth medium (1 : 2) (BioWhittaker) supplemented with 15 % FCS, 1 % penicillin/streptomycin at 37 uC and 5 % CO2. Segments were inoculated with 56105 cell-free virus particles ml21 of either AD169 or the recombinant HCMV viruses, or were mock-infected. After 24 h, the arterial segments were washed three times with PBS and further cultivated as described above. Once a week up to day 56 p.i., aliquots of the culture supernatant were quantified for infectious virus. All of the tested supernatant aliquots from mock-infected segments were always negative for HCMV, irrespective of whether the donor was seropositive or seronegative (Reinhardt et al., 2003).

RESULTS Transcription of HCMV UL78 mRNA in infected fibroblasts Northern blot analysis was performed to determine the presence of UL78 mRNA. Fibroblasts were infected and total RNA was extracted at 16 and 36 h p.i. As shown in Fig. 1(a), an RNA band of approximately 1?7 kb was detectable at 16 h p.i. and at the later time point, the UL78specific probe clearly identified an additional mRNA species migrating at approximately 4 kb. To further elucidate the temporal kinetics of UL78 transcription, infected cells were treated with CHX or PAA. As shown in Fig. 1(b), the appearance of the UL78 1?7 kb mRNA could be detected by 4 h p.i. and was inhibited completely by CHX, but not by PAA, treatment in fibroblasts. On the other hand, the larger transcript of approximately 4 kb was clearly inhibited by PAA. These observations are in line with the recent results of Wang et al. (2004). In conclusion, these results classified UL78 as an early gene that is also transcribed during the late stage of infection. The larger mRNA is a bicistronic transcription unit consisting of UL77 and UL78 coding regions (see Fig. 4). Transcription start point and start of the polyadenylated tail To identify the UL78 transcription start, primer-extension analysis was performed. A UL78-specific oligonucleotide was annealed to total RNA prepared from HCMV-infected fibroblasts and extended by reverse transcriptase. As depicted in Fig. 2(b), the start of the UL78 mRNA was located 48 bp upstream of the translation start codon of the UL78 ORF at nt 112876 (according to the AD169 sequence; GenBank accession no. X17403). For identification of the 39 end of the UL78 mRNA, RT-PCR was performed. An anchor primer was used with a sequence complementary to polyadenylated tails of mRNA. Next, two different UL78specific primers were used in order to amplify the 39 end of the UL78 mRNA. A fragment of approximately 360 bp was 299


Fig. 1. Northern blot analysis of HCMV UL78 transcription kinetics in infected fibroblasts. HFFs were infected with strain AD169. (a) Total RNA was extracted at 16 and 36 h p.i. The nylon membrane was hybridized with a UL78-specific probe. (b) Cells were infected and treated with CHX or PAA and harvested at the indicated time points. The membrane was subsequently hybridized with probes specific for ORF UL78, ORF UL77 or GAPDH.

expected when using primer 1160 and a fragment of approximately 160 bp with primer 1360. Independent experiments were done with target RNA from AD169infected fibroblasts and cells infected with a clinical HCMV isolate. PCR results revealed unexpectedly that both amplified fragments were approximately 250 bp larger than predicted. These results suggested that the site within the UL78 mRNA where the polyadenylated tail is attached comprised sequences of the UL79 ORF. Sequencing of the amplicons revealed that the polyadenylated tail started 268 bp downstream of the UL78 translation stop codon within the UL79 ORF (see Fig. 2b). As shown in Fig. 2, these results suggested that a putative polyadenylation signal (AATAAA) is located 18 bp upstream of the polyadenylation site. A similar configuration can be found in the chimpanzee CMV (ChCMV) genome. As summarized in Fig. 3, to date, several members of a UL78-like gene family have been identified in b-herpesviruses: MCMV M78

(Rawlinson et al., 1996), RCMV R78 (Vink et al., 2000), ChCMV UL78 (Davison et al., 2003), HHV-6 U51 (Gompels et al., 1995), HHV-7 U51 (Nicholas, 1996) and tupaia herpesvirus T78 (Bahr & Darai, 2001). The position and orientation of these genes within the respective herpesvirus genomes are well-conserved. Comparison of the 39 regions of the UL78-like genes revealed significant differences between CMVs of rodents and primates (see Table 1). Putative polyadenylation signals of RCMV R78 and MCMV M78 were located in regions upstream of the coding regions of R79 and M79, respectively. Generation of a recombinant HCMV deficient for UL78 To investigate the role of the UL78 gene product during replication, we used the AD169 genome cloned as a BAC in E. coli and constructed a mutant strain (HCMV-delUL78)

Fig. 2. 59 and 39 ends of UL78 mRNA. (a) Schematic drawing of ORF UL78 and the surrounding region. Arrow boxes represent ORFs. Putative RNAs transcribed from the UL77 and UL78 ORFs are indicated as hatched boxes below. Putative translation start and stop codons are depicted. (b) Schematic drawing of the UL78 mRNA. Nucleotides at the 59 end, as well as at the polyadenylation site, are indicated, as determined from primer-extension analysis and amplification of the 39 end. 300

Journal of General Virology 86

Characterization of HCMV UL78

Fig. 3. Unrooted bootstrap tree of UL78-like proteins. A phylogenetic tree was constructed by using CLUSTAL X alignments of full amino acid sequences. The viral proteins were compared with the prototype CC chemokine receptor, as well as the human GPCR45.

in which 1030 bp of the coding region was deleted (Fig. 4a). The BAC DNA was transfected into HFFs and viable virus was isolated, showing that UL78 is not essential for the replication of HCMV in fibroblasts. The deletion was confirmed with DNA extracted from fibroblasts infected with the reconstituted virus and by Southern blot analysis, as shown in Fig. 4(b). Northern blot analysis was performed to investigate the effect of the deletion on UL78/UL77 transcription. Total RNA was extracted at 36 h p.i. from cells infected with either AD169 or the deletion mutant. Membranes with the separated RNAs were hybridized sequentially with probes

specific for the UL78 ORF, UL77 ORF and the cellular GAPDH gene. As shown in Fig. 4(c), different effects on UL78 transcription could be observed. As mentioned above, hybridization of mRNA from AD169-infected fibroblasts showed clearly that the 1?7 kb RNA contained the UL78coding region, whereas the later-appearing longer RNA consisted of both UL77 and UL78. No signal was detected in RNA samples derived from cells infected with the HCMV UL78-deficient mutant after hybridization with a UL78specific probe. In the RNA sample from cells infected with HCMV-delUL78, a weak signal at approximately 3 kb was detected after hybridization with the UL77-specific probe, suggesting that the UL77 gene might still be active to some extent. The UL77 mRNA level may have been affected by the absence of the UL78 region, which could result in increased instability of the mRNA molecules.

Absence of UL78 does not modify expression of viral proteins representing the different stages of virus replication In order to answer the question of whether the absence of the UL78 gene influences viral protein expression at different stages of virus replication, immunoblotting experiments were performed. HFFs were infected with AD169 or HCMV-delUL78. At different time points p.i., cells were harvested. Crude protein extracts were separated by SDSPAGE and incubated with mAbs directed against IE1, pUL69, pp65 or pp28, or with a polyclonal pUL97 antiserum. These proteins represent different stages of the virus replication cycle. However, as shown in Fig. 5, no significant differences in protein expression could be detected. In individual experiments, there were some differences in band intensities (e.g. pUL69 at 6 h or pp28 at 48 h p.i.), but these were not reproducible in repeated experiments.

Table 1. Polyadenylation signals of UL78-like genes Numbers in parentheses indicate the first nucleotide positions of the putative polyadenylation signals according to the respective published sequences. Virus

HCMV ChCMV RCMV MCMV

Gene

Translation stop codon

Putative polyadenylation signal

TGA TAG TGA TGA

AATAAA AATAAA AATAAA AATAAA AATAAA

UL78 UL78 R78 M78

Position UL78* +247 bp +225 bp +95 bp +1 bp +167 bp

(114466)d (115633)§ (100615)|| (112405) (112571)

UL79D +190 +151 2114 2236 269

bp bp bp bp bp

*With respect to the stop codon of the UL78-like gene. DWith respect to the translation start codon of the UL79-like gene. dGenBank accession no. X17403. §GenBank accession no. AF480884. ||GenBank accession no. NC_002512. GenBank accession no. U68299. http://vir.sgmjournals.org

301


Fig. 4. Construction of UL78-deficient HCMV and Northern blot analysis. (a) Schematic representation of HCMV ORFs UL76–UL79. The shaded and open boxes represent the sizes of the ORFs. The top shows the restriction map of AD169 and the bottom shows the organization of the region after disruption of the UL78 ORF. Filled boxes represent the DNA probes used for Southern and Northern blot experiments. Restriction sites: P, PstI; B, BamHI; BX, BstXII; A, AsuII. (b) Southern blot analysis of HCMV DNA extracted from HFFs infected with AD169 or the HCMV UL78-deficient mutant (del 78). DNA was separated on a 0?5 % agarose gel, transferred to a nylon membrane and hybridized with the BamHI–AsuII DNA probe and the entire UL78 coding region indicated in (a). (c) HFFs were infected with the indicated HCMV variants and total RNA was extracted at 48 h p.i. Total RNA was separated and transferred to nitrocellulose membranes. Membranes were hybridized with probes specific for UL78 or UL77. Signals representing the 1?7 kb UL78 and the putative 4 kb UL77/UL78 transcripts are indicated. The presumed UL77 mRNA deficient for the UL78 coding region is indicated by a black arrow.

The UL78 gene is not important for efficient virus replication Virus yield assays were performed to analyse whether the UL78 ORF is important for efficient infection and replication in fibroblasts. As shown in Fig. 6(a), no significant differences in virus yield kinetics were observed between supernatants of fibroblasts infected with AD169 and HCMVdelUL78. Infections were performed at an m.o.i. of 0?1. However, the same results were obtained at an even lower m.o.i. of 0?01 (data not shown). This showed that the UL78 gene product is dispensable for replication in fibroblasts in vitro and does not influence the efficiency of replication. UL78-deficient HCMV can replicate in a renal artery organ-culture system As no in vivo model for HCMV infection is available, we used an in vitro organ-culture model based on human renal 302

arteries that allowed us to monitor HCMV infection over a long period of time (Reinhardt et al., 2003; Voisard et al., 1999). In order to investigate the replication characteristics of the recombinant viruses in this more complex system, human renal arteries were inoculated in vitro with HCMV wild-type strain AD169, as well as with recombinant HCMV strains. Virus yields were determined from the medium at different times after inoculation. Due to the small size of the artery segments removed from patients, not all viruses could be tested at the same time and on the same artery. Therefore, data were collected from five different independent experiments, as shown in Fig. 6(b–f). Each experiment represents the comparison of the virus yield from artery slices from the same patient infected with different viruses. The recombinant HCMV strain RVHB5 used for generation of the HCMV UL78-deletion mutant virus served as a control. RVHB5, derived from AD169, is the recombinant virus obtained by transfection and reconstitution of BACmid Journal of General Virology 86


Fig. 5. Expression of viral proteins in HFFs infected with the UL78-deficient mutant. HFFs were infected with the indicated HCMV variants and crude cell extracts were harvested at the indicated time points p.i. Samples were separated by SDSPAGE. Membranes were incubated with mAbs directed against the immediate-early protein IE1, the early/late protein pp65, the true late protein pp28, and pUL69, which is expressed throughout the entire replication cycle. Additionally, a polyclonal antiserum specific for the early/late protein pUL97 was used. Proteins were visualized with secondary antibodies conjugated to peroxidase and ECL.

viruses were comparable, although the amount of virus produced by the recombinants was slightly lower than that of wild-type AD169 during the first weeks after infection (Fig. 6b, d and f). However, approaching day 17 p.i., virus yields were fairly similar. Comparing the UL78-deficient strain and the parental RVHB5 recombinant (Fig. 6c and e), no effect on overall replication in the arteries could be observed. HCMV reactivations were never observed in the segments of seropositive donors. Thus, these results further strengthened the assumption that the HCMV UL78 gene is not essential for virus replication. The cell types in the arteries supporting virus replication were not determined in this study. pHB5 (Borst et al., 1999), which is lacking the region between ORFs US2 and US6. This region is substituted by the BAC cassette. Virus titres in the culture medium of the arteries were determined up to a maximum of 56 days p.i. As shown in Fig. 6, the growth kinetics of the different

UL78 is conserved among clinical isolates In order to elucidate whether sequence variations in UL78 existed, as is observed in the US28 protein, 11 low-passaged clinical isolates from nine patients were investigated. The

Fig. 6. Virus yield kinetics in HFFs and in a renal artery organ-culture system. (a) HFFs were infected at an m.o.i. of 0?1 with either the laboratory strain AD169 or the HCMV UL78-deficient strain. The graphs show the virus titres determined in the culture medium. #, UL78-deficient mutant; &, AD169. (b–f) Segments of renal arteries were removed during routine nephrectomies from different patients. After inoculation of the renal-artery segments with the indicated viruses, a long-lasting permissive HCMV infection was induced. Production of virus in the medium supernatant was quantified at the indicated time points p.i. RVHB5 is the reconstituted HCMV control strain lacking the US2–US6 region, which is substituted by the BAC cassette as described elsewhere (Borst et al., 1999). http://vir.sgmjournals.org

303


isolates were derived from different specimens (urine, throat wash and leukocytes). As summarized in Fig. 7, multiple sequence alignments of the predicted amino acid sequences revealed that the UL78 protein was highly conserved among the clinical isolates. However, several variations could be found in the N-terminal region and at the very end of the C-terminal region. Overall, when compared with AD169, the identity was 98?5–99?7 % at the nucleotide level and 98?2–100 % at the amino acid level. On the basis of the secondary structure as predicted by the SWISS-PROT program on the ExPASy Proteomics Server (last release 1999) (Bairoch & Boeckmann, 1992), the first 41 aa of UL78 might constitute the first extracellular part of the protein (Fig. 7). Isolates from the same patient, but isolated from different specimens, exhibited no sequence variations.

proteins of animal CMVs, only one conserved cysteine residue (aa C109) could be identified. The second cysteine residue in the second intracellular loop, which would be necessary to form a disulfide bridge, might be located at aa C187. The conserved motif DRL (aa 133–135) required for G-protein binding can be found in the second extracellular loop. It is unclear why most of the amino acid substitutions in different virus isolates are located in the N-terminal region. The higher variation of an extracellularly exposed N-terminal region indicates that this part of the molecule is not involved in binding of a specific ligand and may be helpful in immune evasion. This observation was also made in US28 proteins from different patient isolates but, until now, no functional relevance has been linked to this phenomenon (R. Minisini, B. Mo¨pps, T. Mertens & P. Gierschick, unpublished data).

DISCUSSION

Here, we showed that the UL78 mRNA has an early/late transcription pattern similar to those of a number of CMV GPCR homologues, including US28 (Bodaghi et al., 1998; Zipeto et al., 1999), RCMV R78 (Beisser et al., 1999) and MCMV M78 (Oliveira & Shenk, 2001), and is not suppressed by PAA. All of these mRNAs are transcribed early in the replication cycle and the transcripts are suppressed by CHX, but not by PAA. These findings are in accordance with the recently reported results of Wang et al. (2004) concerning the function of HCMV UL76. These authors also found that the UL78 transcript was not detected under immediate-early conditions and was not inhibited by PAA. In early/late and late stages, these ORFs were part of longer transcripts consisting of US27/US28, R77/R78 or M77/ M78, respectively. Our Northern blot analyses showed clearly that the longer mRNA, which is expressed preferentially during the late stage of HCMV replication, included UL77/UL78-coding sequences.

b-Herpesviruses of different hosts contain genes that are

considered to be homologous in function. In this study, we characterized the UL78 gene, which is thought to be a 7TM protein. In contrast to ORFs US27, US28 and UL33, the predicted amino acid sequence of pUL78 does not exhibit stringent homology to GPCRs. This assumption results from the comparison with HHV-6 U51, which is related closely to UL78 (Gompels et al., 1995). However, in contrast to U51, which can bind RANTES/CCL5 with nanomolar affinity, no pharmacological data exist for pUL78. The protein is only suspected to be a homologue of GPCRs because of structural properties, which include seven putative hydrophobic regions, two cysteine residues that are predicted to be required for correct folding of GPCRs and a region that exhibits similarity to a domain known to be required for G-protein coupling (Gompels et al., 1995; Vink et al., 2001). When compared with other UL78-like

Fig. 7. The predicted secondary structure of the putative UL78 protein (according to the SWISS-PROT program on the ExPASy Proteomics Server; Bairoch & Boeckmann, 1992) is shown. Amino acids located in the membrane are depicted as open circles and those that are located extra- or intracellularly are indicated in grey. Conserved amino acids are shown as dotted circles. Mutations that exhibited differences among the various UL78 proteins are represented by black circles. 304



Detection of the polyadenylated tail of the UL78 mRNA and sequence alignments revealed significant differences among the CMVs of different species. Although the functional relevance of the putative polyadenylation signals in the CMVs of humans, chimpanzee, rat and mouse is unknown, this observation might reflect evolution in different hosts. In the T78–T79 region of the tupaia herpesvirus, for example, no polyadenylation signal could be found, even though T78 has 25 and 26 % similarity to ChCMV UL78 and HCMV UL78, respectively. It should be noted that the region between T78 and T79 encodes a putative protein with homology to MCMV M59. Such a configuration has not been observed in the other CMV genomes.

or pM33 were found to replicate in vitro with efficiencies similar to those of the corresponding wild-type viruses (Beisser et al., 1998; Davis-Poynter et al., 1997). However, the replication of both pR33 and pM33 deletion mutants was found to be impaired in the salivary glands of infected animals. Functional differences among the homologous proteins R33 and UL33 have also been found (Casarosa et al., 2003). Although both receptors constitutively activate phospholipase C via Gq/11, and partially via Gi/o-mediated pathways, they exhibited profound differences in the modulation of CRE activation – pR33 constitutively inhibited, whereas pUL33 constitutively enhanced CREmediated transcription.

Due to the fact that UL78-like genes are present in all known b-herpesviruses, it has been speculated that these genes may play a crucial role in the pathogenesis of infection. Despite the relatively low level of sequence similarity to known chemokine receptors, HHV-6A U51 was reported to bind several chemokines, e.g. CCL2 and CCL5, as well as vMIP-II, a chemokine encoded by HHV-8 (Milne et al., 2000). Concerning HHV-6 U12, it has been reported that RANTES/CCL5 elicits calcium release (Isegawa et al., 1998). However, to date, no ligand binding or signalling has been shown for HCMV pUL78. Some interesting differences have been observed following infection with mutant viruses deficient for UL78-like genes of CMVs. It has been shown that RCMV pR78 and MCMV pM78 serve important functions during in vivo infection (Beisser et al., 1999; Oliveira & Shenk, 2001). R78-deficient mutants are associated with a lower mortality rate in immunocompromised rats than in animals infected with wild-type virus, although no significant differences in virus yield could be found between strains with deletions and wild-type virus in vivo (Beisser et al., 1999). Furthermore, pR78 of RCMV seems to influence virus replication in fibroblasts and smooth muscle cells, where the R78-deficient mutants were found to replicate 10- to 100-fold less efficiently than wildtype RCMV. The M78-deficient mutant of MCMV only exhibited slightly reduced growth kinetics in vitro, but the authors suggested that pM78 facilitated accumulation of immediate-early viral mRNA (Oliveira & Shenk, 2001). The deletion of R78 led to an attenuated, syncytium-inducing phenotype, whereas no phenotype has been described after infection with an M78-deficient mutant (Oliveira & Shenk, 2001). As shown in the present study, the UL78-deficient HCMV mutant also exhibited no phenotype either in fibroblasts or in the in vitro renal-artery model system. Again, these results lead to the assumption of functional differences of the UL78-like proteins in different CMV species. Our results showed that the UL78 gene product is not as important for virus replication as pR78 for RCMV or pM78 for MCMV, at least in vitro. Recently, we showed that, although MCMV pM97 and HCMV pUL97 shared strong homologies in their amino acid sequences, they exhibited different functional properties (Wagner et al., 2000). Differences between in vitro and in vivo results were also obtained with other genes. Viruses deficient for pR33

The lack of an HCMV in vivo model still represents a great obstacle for investigating the influence of particular genes on HCMV pathogenesis. Thus, it is still questionable whether pUL78 can be considered as a potential target for future development of novel antiviral strategies as suggested (Beisser et al., 1999).

http://vir.sgmjournals.org

ACKNOWLEDGEMENTS This study was supported by the ‘Sonderforschungsbereich 451’, projects A2 and A3. I. M. was supported by a grant of the Ministerium fu¨r Wissenschaft, Forschung und Kunst Baden-Wu¨rttemberg in collaboration with the Croatian Ministry of Science and Technology. We thank Marlies Just and Birgit Jo¨rg for culturing the clinical HCMV isolates.

REFERENCES Bahr, U. & Darai, G. (2001). Analysis and characterization of the

complete genome of tupaia (tree shrew) herpesvirus. J Virol 75, 4854–4870. Bairoch, A. & Boeckmann, B. (1992). The SWISS-PROT protein

sequence data bank. Nucleic Acids Res 20, 2019–2022. Beisser, P. S., Vink, C., Van Dam, J. G., Grauls, G., Vanherle, S. J. V. & Bruggeman, C. A. (1998). The R33 G protein-coupled receptor

gene of rat cytomegalovirus plays an essential role in the pathogenesis of viral infection. J Virol 72, 2352–2363. Beisser, P. S., Grauls, G., Bruggeman, C. A. & Vink, C. (1999).

Deletion of the R78 G protein-coupled receptor gene from rat cytomegalovirus results in an attenuated, syncytium-inducing mutant strain. J Virol 73, 7218–7230. Bodaghi, B., Jones, T. R., Zipeto, D., Vita, C., Sun, L., Laurent, L., Arenzana-Seisdedos, F., Virelizier, J.-L. & Michelson, S. (1998).

Chemokine sequestration by viral chemoreceptors as a novel viral escape strategy: withdrawal of chemokines from the environment of cytomegalovirus-infected cells. J Exp Med 188, 855–866. Borst, E.-M., Hahn, G., Koszinowski, U. H. & Messerle, M. (1999).

Cloning of the human cytomegalovirus (HCMV) genome as an infectious bacterial artificial chromosome in Escherichia coli: a new approach for construction of HCMV mutants. J Virol 73, 8320–8329. Casarosa, P., Bakker, R. A., Verzijl, D., Navis, M., Timmerman, H., Leurs, R. & Smit, M. J. (2001). Constitutive signaling of the human

cytomegalovirus-encoded chemokine receptor US28. J Biol Chem 276, 1133–1137. Casarosa, P., Gruijthuijsen, Y. K., Michel, D. & 9 other authors (2003). Constitutive signaling of the human cytomegalovirus-encoded

305

D. Michel and others receptor UL33 differs from that of its rat cytomegalovirus homolog R33 by promiscuous activation of G proteins of the Gq, Gi, and Gs classes. J Biol Chem 278, 50010–50023. Chee, M. S., Satchwell, S. C., Preddie, E., Weston, K. M. & Barrell, B. G. (1990). Human cytomegalovirus encodes three G protein-

coupled receptor homologues. Nature 344, 774–777. Davison, A. J., Dolan, A., Akter, P., Addison, C., Dargan, D. J., Alcendor, D. J., McGeoch, D. J. & Hayward, G. S. (2003). The

human cytomegalovirus genome revisited: comparison with the chimpanzee cytomegalovirus genome. J Gen Virol 84, 17–28. Davis-Poynter, N. J., Lynch, D. M., Vally, H., Shellam, G. R., Rawlinson, W. D., Barrell, B. G. & Farrell, H. E. (1997). Identification

and characterization of a G protein-coupled receptor homolog encoded by murine cytomegalovirus. J Virol 71, 1521–1529. Gao, J.-L., Kuhns, D. B., Tiffany, H. L., McDermott, D., Li, X., Franke, U. & Murphy, P. M. (1993). Structure and functional expression of the human macrophage inflammatory protein 1a/

RANTES receptor. J Exp Med 177, 1421–1427.

novel human herpesvirus-6 b chemokine receptor. J Immunol 164, 2396–2404. Minisini, R., Tulone, C., Lu¨ske, A., Michel, D., Mertens, T., Gierschik, P. & Moepps, B. (2003). Constitutive inositol phosphate

formation in cytomegalovirus-infected human fibroblasts is due to expression of the chemokine receptor homologue pUS28. J Virol 77, 4489–4501. Nicholas, J. (1996). Determination and analysis of the complete nucleotide sequence of human herpesvirus 7. J Virol 70, 5975–5989. Oliveira, S. A. & Shenk, T. E. (2001). Murine cytomegalovirus M78

protein, a G protein-coupled receptor homologue, is a constituent of the virion and facilitates accumulation of immediate-early viral mRNA. Proc Natl Acad Sci U S A 98, 3237–3242. Po´sfai, G., Koob, M. D., Kirkpatrick, H. A. & Blattner, F. R. (1997). Versatile insertion plasmids for targeted genome manipula-

tions in bacteria: isolation, deletion, and rescue of the pathogenicity island LEE of the Escherichia coli O157 : H7 genome. J Bacteriol 179, 4426–4428. Rawlinson, W. D., Farrell, H. E. & Barrell, B. G. (1996). Analysis of

Gompels, U. A., Nicholas, J., Lawrence, G., Jones, M., Thomson, B. J., Martin, M. E., Efstathiou, S., Craxton, M. & Macaulay, H. A. (1995). The DNA sequence of human herpesvirus-6: structure,

the complete DNA sequence of murine cytomegalovirus. J Virol 70, 8833–8849.

coding content, and genome evolution. Virology 209, 29–51.

Reinhardt, B., Vaida, B., Voisard, R. & 7 other authors (2003).

Isegawa, Y., Ping, Z., Nakano, K., Sugimoto, N. & Yamanishi, K. (1998). Human herpesvirus 6 open reading frame U12 encodes a functional b-chemokine receptor. J Virol 72, 6104–6112.

Human cytomegalovirus infection in human renal arteries in vitro. J Virol Methods 109, 1–9.

Margulies, B. J., Browne, H. & Gibson, W. (1996). Identification of

a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

the human cytomegalovirus G protein-coupled receptor homologue encoded by UL33 in infected cells and enveloped virus particles. Virology 225, 111–125.

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning:

Streblow, D. N., Soderberg-Naucler, C., Vieira, P., Smith, J., Wakabayashi, E., Ruchti, F., Mattison, K., Altschuler, Y. & Nelson, J. A. (1999). The human cytomegalovirus chemokine receptor US28

Megaw, A. G., Rapaport, D., Avidor, B., Frenkel, N. & Davison, A. J. (1998). The DNA sequence of the RK strain of human herpesvirus 7.

mediates vascular smooth muscle cell migration. Cell 99, 511–520.

Virology 244, 119–132.

Vink, C., Beuken, E. & Bruggeman, C. A. (2000). Complete DNA

Menotti, L., Mirandola, P., Locati, M. & Campadelli-Fiume, G. (1999). Trafficking to the plasma membrane of the seven-

Vink, C., Smit, M. J., Leurs, R. & Bruggeman, C. A. (2001). The role

transmembrane protein encoded by human herpesvirus 6 U51 gene involves a cell-specific function present in T lymphocytes. J Virol 73, 325–333.

of cytomegalovirus-encoded homologs of G protein-coupled receptors and chemokines in manipulation of and evasion from the immune system. J Clin Virol 23, 43–55.

Michel, D., Salamini, F., Bartels, D., Dale, P., Baga, M. & Szalay, A. (1993). Analysis of a desiccation and ABA-responsive promoter

Voisard, R., von Eicken, J., Baur, R., Gschwend, J. E., Wenderoth, U., Kleinschmidt, K., Hombach, V. & Ho¨her, M. (1999). A human

isolated from the resurrection plant Craterostigma plantagineum. Plant J 4, 29–40.

arterial organ culture model of postangioplasty restenosis: results up to 56 days after ballooning. Atherosclerosis 144, 123–134.

Michel, D., Pavic´, I., Zimmermann, A., Haupt, E., Wunderlich, K., Heuschmid, M. & Mertens, T. (1996). The UL97 gene product of

Wagner, M., Michel, D., Schaarschmidt, P., Vaida, B., Jonjic, S., Messerle, M., Mertens, T. & Koszinowski, U. (2000). Comparison

human cytomegalovirus is an early-late protein with a nuclear localization but is not a nucleoside kinase. J Virol 70, 6340–6346.

between human cytomegalovirus pUL97 and murine cytomegalovirus (MCMV) pM97 expressed by MCMV and vaccinia virus: pM97 does not confer ganciclovir sensitivity. J Virol 74, 10729–10736.

Michel, D., Schaarschmidt, P., Wunderlich, K., Heuschmid, M., Simoncini, L., Mu¨hlberger, D., Zimmermann, A., Pavic´, I. & Mertens, T. (1998). Functional regions of the human cytomegalovirus protein

sequence of the rat cytomegalovirus genome. J Virol 74, 7656–7665.

Waldhoer, M., Kledal, T. N., Farrell, H. & Schwartz, T. W. (2002).

pUL97 involved in nuclear localization and phosphorylation of ganciclovir and pUL97 itself. J Gen Virol 79, 2105–2112.

Murine cytomegalovirus (CMV) M33 and human CMV US28 receptors exhibit similar constitutive signaling activities. J Virol 76, 8161–8168.

Michel, D., Ho¨hn, S., Haller, T., Jun, D. & Mertens, T. (2001).

Wang, S.-K., Duh, C.-Y. & Wu, C.-W. (2004). Human cytomegalovirus

Aciclovir selects for ganciclovir-cross-resistance of human cytomegalovirus in vitro that is only in part explained by known mutations in the UL97 protein. J Med Virol 65, 70–76.

UL76 encodes a novel virion-associated protein that is able to inhibit viral replication. J Virol 78, 9750–9762.

Milne, R. S. B., Mattick, C., Nicholson, L., Devaraj, P., Alcami, A. & Gompels, U. A. (2000). RANTES binding and down-regulation by a

306

Zipeto, D., Bodaghi, B., Laurent, L., Virelizier, J.-L. & Michelson, S. (1999). Kinetics of transcription of human cytomegalovirus chemo-

kine receptor US28 in different cell types. J Gen Virol 80, 543–547.
