|Home | About | Journals | Submit | Contact Us | Français|
The glycoprotein gO (UL74) of human cytomegalovirus (HCMV) forms a complex with gH/gL. Virus mutants with a deletion of gO show a defect in secondary envelopment with the consequence that virus spread is restricted to a cell-associated pathway. Here we report that the positional homolog of HCMV gO, m74 of mouse CMV (MCMV), codes for a glycosylated protein which also forms a complex with gH (M75). m74 knockout mutants of MCMV show the same spread phenotype as gO knockout mutants of HCMV, namely, a shift from supernatant-driven to cell-associated spread. We could show that this phenotype is due to a reduction of infectious virus particles in cell culture supernatants. m74 knockout mutants enter fibroblasts via an energy-dependent and pH-sensitive pathway, whereas in the presence of an intact m74 gene product, entry is neither energy dependent nor pH sensitive. This entry phenotype is shared by HCMV expressing or lacking gO. Our data indicate that the m74 and UL74 gene products both codetermine CMV spread and CMV entry into cells. We postulate that MCMV, like HCMV, expresses alternative gH/gL complexes which govern cell-to-cell spread of the virus.
Herpesviruses are complex viruses displaying many different glycoproteins in their envelopes. The exact role of individual glycoproteins in binding to target cells, activation of cellular signaling pathways, and fusion of the viral envelope with cellular membranes is still unclear. For human cytomegalovirus (HCMV), more than 10 different envelope proteins have been identified (53). Three glycoprotein complexes (designated gB, gM/N, and gH/gL complexes) have been characterized in more detail. The gB complex is formed by homodimers of the glycoprotein B (gB) (5) and is required for virus entry and cell-to-cell spread (19). The gM/gN complex is formed by the highly conserved glycoproteins gM (UL100) and gN (UL73) (26) and probably plays a role in initial attachment and tethering to heparan sulfate proteoglycans on the cell surface (8). For gH/gL, two alternative complexes have been described: a complex of gH (UL75), gL (UL115), and gO (UL74) (16, 24) and a complex of gH, gL, and pUL128, pUL130, and pUL131A (1, 56). HCMV gH/gL complexes are supposed to play a role in promoting fusion of the viral envelope and cellular membranes and probably act in concert with gB and potentially through binding to integrin receptors (11, 22, 36, 37, 52, 58). Recently, the role of the gH/gL/gO complex was analyzed in more detail. gO deletion mutants showed a severely impaired release of infectious virus into cell culture supernatants (20, 60), accompanied by accumulation of naked capsids in the cytoplasm (20). This implied a role for gO in virus secondary envelopment. It was shown that deletion of gO did not affect viral cell-to-cell spread or cell tropism of HCMV, indicating that gO is not essential for entry. In contrast, the gH/gL/pUL(128-131A) complex is essential for infection of endothelial, epithelial, and dendritic cells but is not needed for infection of fibroblasts and neuronal cells (12, 15, 55). Deletion of any of the UL128, UL130, and UL131A genes and in addition of gO resulted in a spread-deficient virus also in fibroblasts, indicating that for virus entry and spread in fibroblasts, either gO or pUL(128,130,131A) is needed (20). The biochemistry of the 464-amino-acid (aa)-long gO of HCMV has been extensively studied. It forms a complex with gH and gL of about 250 kDa. Under reducing conditions, the complex is disrupted and gO appears as multiple bands of about 115 to 125 kDa in sodium dodecyl sulfate (SDS)-polyacrylamide gels. gO is N and O glycosylated and contains an amino-terminal hydrophobic membrane anchor which is cleaved off during processing (17, 49, 50). Data on the contribution of gO to the fusogenic activity of the gH/gL complex are controversial (36, 52). It is also not exactly known whether gH/gL/gO complexes are consistently incorporated into HCMV particles or whether gO merely acts as a chaperone to promote incorporation of gH/gL complexes into virions (44, 60).
It is likely that the alternative gH/gL complexes of HCMV bind to different receptors and in concert with gB initiate fusion (43). It is a matter of current research whether fusion takes place at the plasma membrane or at membranes of endocytic vesicles after endocytosis, whether binding of different gH/gL complexes to different receptors initiates different entry pathways, and whether entry pathways vary between cell types (3, 9, 37, 45, 47, 57). The role of alternative gH/gL complexes for cell-type-specific infection in vivo is a matter of speculation and is difficult to study in humans.
Mouse CMV (MCMV) infection resembles the human infection with respect to organ and cell tropism, pathogenesis during acute infection, establishment of latency, and reactivation after immunosuppression (23, 31, 41). Yet, the homologous glycoproteins of MCMV have barely been investigated. MCMV gB and gH have been expressed as recombinant proteins and biochemically characterized (39, 40). Most of the homologous glycoproteins have been identified as virion-associated proteins (21). Here we analyzed the MCMV m74 gene product both with respect to its complex formation with MCMV gH and with respect to the phenotype of m74 knockout mutants. The m74 gene product is, like HCMV gO, a glycosylated protein associating with gH. Knockout of the m74 gene resembles UL74 knockout mutants in every respect: delayed reconstitution of bacterial artificial chromosome (BAC)-derived mutants, a shift from supernatant-driven to cell-associated virus spread, and impaired secondary envelopment. We could additionally show that in contrast to wild-type (wt) MCMV, infection of fibroblasts with m74 knockout mutants is energy dependent and pH sensitive, a phenotype that is shared by HCMV lacking gO.
Mouse embryonal fibroblasts (MEF) from BALB/c mice, MEF-BL/6-1 (ATCC), human foreskin fibroblasts (HFF), and human kidney epithelial cells (293) (13) were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum, 100 units penicillin, and 100 units streptomycin. NIH 3T3 (ATCC) cells were maintained in DMEM supplemented with 5% fetal calf serum, 100 units penicillin, and 100 units streptomycin.
w.t.-MCMV and w.t.-FRT-MCMV are BAC-derived viruses (pSM3fr and pSM3fr-FRT) cloned from MCMV strain Smith (7, 54). The latter contains an FLP recombination target (FRT) site inserted between genes m16 and m17.
Virus stocks were prepared from supernatants of infected fibroblasts. Supernatants were cleared of cellular debris by centrifugation for 15 min at 3,500 × g and stored at −80°C.
Virus titers of cleared supernatants of cells infected with MCMV or HCMV were determined either by a 50% tissue culture infective dose (TCID50) assay on 96-well plates on MEF or HFF, respectively, or by using a standard plaque assay on MEF.
To infect cells, medium was removed from 90% confluent cell monolayers and replaced by virus diluted in medium. Virus uptake was enhanced by a centrifugation step (30 min, 2,000 × g at room temperature) followed by an incubation at 37°C for 90 min.
To monitor cell death or damage, lactate dehydrogenase (LDH) levels were detected in cell culture supernatants using the LDH cytotoxicity detection kit (TaKaRa, Japan).
The m74 open reading frame (ORF) was amplified by PCR from pSM3fr BAC DNA (primers: m74-for, 5′-AGCCCTGCAGATGAACCCCTTATTACTCAT-3′; m74-rev, 5′-AGGGCGG CCGCTCAGACACGGCTAAAGGATA-3′; m74HA-rev, 5′-AGGGCGGCCGCCTAAGCGTAGTCTGGGACGTCGTATGGGTAGACACGGCTAAAGGATATTG-3′). The PCR fragments were either cloned in pCR3 (Invitrogen) via PstI and NotI restriction sites, sequenced, and expressed as hemagglutinin (HA)-tagged and untagged m74 proteins under the control of the CMV promoter (pCR3-m74+/−HA) or cloned in pori6-SVTentry (42) via an HpaI restriction site and expressed under the control of the simian virus 40 (SV40) promoter (pori6-SVTentry-m74+/−HA). For stable expression in NIH 3T3 cells, m74 and m74HA were cloned into a modified pEPi-luc vector (38), carrying the blasticidin resistance gene and an expression cassette derived from pGL3, containing the firefly luciferase gene under the control of a minimal SV40 promoter inserted after the scaffold/matrix attached region (S/MAR) site. The luciferase gene was replaced by blunt-end ligation of m74 and m74HA excised from the respective pCR3 vectors via Ecl136II into HindIII and XbaI sites, resulting in the vectors pEPi-m74 and pEPi-m74HA.
The MCMV strain Smith cloned as BACs with and without an FRT site inserted between genes m16 and m17 (pSM3fr-FRT and pSM3fr, respectively) (7, 54) was used for MCMV BAC mutagenesis. An m74 deletion mutant of pSM3fr-FRT (pΔm74) was generated as follows. A linear PCR fragment containing the kanamycin resistance gene flanked by sequences homologous to the N terminus of m74 was generated using the primers deltam74-for (5′-GCCTATTAAAACACAAAGAGAGCCGCGATTAATGTCCGCTGTATTCAACGCGGA GATCAGCCCTCCCGCGATTTATTCAACAAAGCCACG-3′) and deltam74-rev (5′-CGTATCGCATTTTTTAAAATATTTGGCGGTGATGTTACTTTTCGGGGTGATGAGGTCTCTCCGCCAGTGTTACAACCAATTAACC-3′) and pACYC177 as template. The resulting PCR fragment was inserted into pSM3fr-FRT by homologous recombination in Escherichia coli, thereby deleting the N-terminal 532 bp of the m74 reading frame and inserting a kanamycin resistance cassette derived from pACYC177.
To screen for recombination in genomes of vΔm74 progeny trans complemented in NIH 3T3-m74, PCR analysis using a primer located upstream of the m74 start codon, m74upstream (5′-GTATTCAACGCGGAGATCAG-3′), and a primer located within the deleted N-terminal sequence of m74, m74deletion (5′-CTGAGTTGACCGTGAATGAG-3′), was performed.
For cis complementation of m74, SV40-driven m74+/−HA tag was inserted into the FRT site of pΔm74 by using pori6-SVTentry-m74+/−HA tag and the temperature-sensitive FLP expression plasmid pCP20 (6). The resulting BACmids were called pΔm74-m74+/−HA.
An m74stop mutant of pSM3fr was generated by a markerless BAC mutagenesis as described previously (20, 51). Briefly, an I-SceI-aphAI cassette was amplified from the plasmid pEP-KAN-S by PCR using the primers m74stop-for (5′-GGAGGTTCGGTCGCATCGATTGTATCATAACCTCCGTCTTCATA ATCATCGGCTAGTTAACTAGCCAGGATGACGACGATAAGTAGGG-3′) and m74stop-rev (5′-AAAGTGTAGCATACAACCCGGCCGTTACCGGCTATATCGAGATGAGCGAAGGCTAGTTAACTAGCCGATGATTATGAAGACGGAGGCAACCAATTAACCAATTCTGATTAG-3′). The MCMV-specific primer extensions contain a 16-bp stop cassette (boldface) consisting of stop codons (underlined) in all three reading frames (10). This cassette was introduced at nucleotide position 120 of m74. In a first Red recombination, the PCR fragment was inserted in pSM3fr, resulting in a BAC carrying a kanamycin resistance cassette and an I-SceI restriction site. The kanamycin cassette was then removed from kanamycin-resistant clones by an I-SceI digest and a subsequent Red recombination, resulting in a markerless insertion of the stop cassette at position 120. The resulting BACmid was called pm74stop.
To screen for recombination in genomes of vm74stop progeny trans complemented in NIH 3T3-m74, PCR analysis using a primer at the insertion site of the stop cassette, stopsensitivem74 (5′-GAGATGAGCGAAGATGATTA-3′), and a primer binding downstream of the stop insertion site, m74/437-418 (5′-TTGATTGGGGTGTAAGTAGC-3′), was performed.
A TB40-BAC4-UL131Astop mutant was generated by markerless mutagenesis of TB40-BAC4 (48) using primers UL131Astop-for (5′-GCTTTCTTTCTCAGTCTGCAACATGCGGCTGTGTCGGGTGTAGCTGTCTGTTTGTCTGTGCGAGGATGACGACGATAAGTAGGG-3′) and UL131Astop-rev (5′-GCACTGACCCAGCACCACGGCGCACAGACAAACAGACAGCTACACCCGACACAGCCGCATGCAACCAATTAACCAATTCTGATTAG-3′). This strategy created a markerless nucleotide exchange at position 20 (G-A), resulting in a premature stop codon after 6 aa.
Deletions and insertions were controlled by restriction pattern analysis and subsequent sequencing.
BACmids were reconstituted to virus by transfection of BAC DNA into MEF (MCMV) or HFF (HCMV) using FugeneHD transfection reagent (Roche Diagnostics) according to the manufacturer's instructions. Supernatants from these transfections or trypsinized cells were used for further passages in fibroblast cell cultures. For infection experiments, virus passaged two to four times was used.
NIH 3T3 cells were transfected in 6-well plates with 2.5 μg of pEPi-m74+/−HA using TransIT-3T3 (Mirus) transfection reagent according to the manufacturer's protocol. One day after transfection, cells were selected for those carrying the pEPi constructs with 10 μg/ml blasticidin. When the first colonies grew, either cells were kept as a pool or clones were isolated by limiting dilution. m74-expressing NIH 3T3 cells were called NIH 3T3-m74, and vΔm74 and vm74stop trans complemented by growth in these cell lines were called vΔm74/m74trans and vm74stop/m74trans, respectively.
293 cells were transiently transfected in 6-well plates with 6 μg of plasmid DNA using FuGeneHD transfection reagent (Roche Diagnostics) according to the manufacturer's protocol. Twenty-four hours after transfection, cells were harvested and lysed in sample buffer for Western blot analysis.
Cells were lysed in RIPA buffer (50 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 0.1% SDS, 0.5% deoxycholate) containing a protease inhibitor cocktail (Roche Diagnostics). Comparable protein amounts were subjected to immunoprecipitation using either rat anti-HA antibody (3F10; Roche Diagnostics) and protein G-Sepharose (GE Healthcare) or mouse anti-MCMV gH (8D1.22A; kindly provided by Lambert Loh, University of Saskatchewan, Canada) (25) and protein A-Sepharose (GE Healthcare). Mouse IgG2a (Perkin-Elmer) served as an isotype control for the anti-gH antibody. For Western blot analysis, precipitates or pelleted cells were lysed either in reducing sample buffer (0.13 M Tris-HCl [pH 6.8], 6% SDS, 10% α-thioglycerol) or in nonreducing sample buffer without α-thioglycerol, separated on 10% SDS-polyacrylamide gels, and transferred onto Hybond ECL nitrocellulose (GE Healthcare). Membranes were blocked with 5% low-fat milk in Tris-buffered saline (TBS) and incubated with anti-HA antibody, and tagged proteins were detected by using peroxidase-coupled anti-rat antibody (Dianova) and enhanced chemiluminescence (ECL system; GE Healthcare).
N-linked carbohydrates were removed using endoglycosidase H (Endo H; Roche Diagnostics) on cells lysed in a Triton X-100 buffer (20 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1% Triton X-100, protease inhibitor cocktail) according to the manufacturer's protocol. The reaction was stopped by adding an equal volume of sample buffer.
Cells were fixed in 50% acetone-50% methanol and stained by indirect immunofluorescence using a monoclonal mouse anti-MCMV immediate-early 1 (ie1) antibody (Croma 101; kindly provided by Stipan Jonjic, University of Rijeka, Croatia) or a mouse anti-HCMV ie1 antibody (Perkin-Elmer) and Cy3-coupled goat anti-mouse antibody (Dianova). For counterstaining of cell nuclei, cells were incubated in phosphate-buffered saline (PBS) containing 5 μg/ml Hoechst 333258 (Invitrogen) for 1 min.
For energy depletion, cells were preincubated in energy depletion medium (glucose-free DMEM with 2% bovine serum albumin [BSA], 50 mM 2-deoxy-d-glucose, and 0.1% sodium azide) for 1 h followed by coincubation with virus for 90 min in the presence of energy depletion medium. All incubations were performed at 37°C. Virions that had not penetrated during coincubation were inactivated by washing the cells three times with PBS, pH 3.0. Cells were then incubated for 3 h (MCMV) or 24 h (HCMV) in DMEM, and virus entry was assessed by indirect immunofluorescence staining for ie1 protein.
For inhibition of pH-dependent endocytosis, cells were pretreated with medium containing NH4Cl or bafilomycin A1 (Sigma) for 1 h at 37°C. Infection (coincubation with virus for 90 min) and further incubations were all performed in the presence of the respective inhibitors. Three hours (MCMV) or 24 h (HCMV) after infection, virus entry was assessed as described above.
Viral DNA in infected cells and cell culture supernatants was determined by quantitative real-time PCR using an ABI Prism 7700 sequence detector (Applied Biosystems). DNA was extracted from infected cells or supernatants using the DNeasy blood and tissue kit (Qiagen) according to the manufacturer's instructions. Supernatants were pretreated with 2 U DNase I (Fermentas) for 60 min at 37°C. To quantify the relative ratio of viral DNA per cell, two independent PCRs were performed for each sample in triplicate: a PCR specific for MCMV M54 using the primers MCMV-for (5′-ATCATCCGTTGCATCTCGTTG-3′) and MCMV-rev (5′-CGCCATCTGTATCCGTCCAT-3′) and the specific probe M54-as (5′-6-carboxyfluorescein [FAM]-AACGTACATCGCTCTCTGCTGGCCG-6-carboxytetramethylrhodamine [TAMRA]-3′) and a PCR specific for the cellular lamin B receptor (LBR) gene using the primers LBR-for (5′-GGAAGTTTGTTGAGGGTGAAGTGGT-3′) and LBR-rev (5′-CCAGTTCGGTGCCATCTTTGTATTT-3′) and a specific probe for LBR (5′-FAM-CTGAGCCACGACAACAAATCCCAGCTCTAC-TAMRA-3′). Viral DNA copy numbers were calculated by comparing the amplification to standard curves using pSM3fr or pSM3fr-LBR BAC DNA. DNA amplification was performed using the TaqMan 1000 RXN PCR core reagent kit (50 cycles of 95°C for 15 s and 60°C for 1 min) (Applied Biosystems). The data are presented as viral genome copy numbers/ml or viral genome copy numbers relative to the copy numbers of LBR.
MEF-BL/6-1 or NIH 3T3 cells were infected with w.t.-FRT-MCMV or vΔm74/m74trans at a multiplicity of infection (MOI) of 0.5. Forty-eight hours after infection, trypsinized or proteinase K-mobilized cells were fixed and analyzed using two different methods: (i) conventional chemical fixation, dehydration, and staining of sections with uranyl acetate and lead citrate as described previously (20) and (ii) physical fixation by high-pressure freezing, followed by cryosectioning and cryo-electron microscopy of vitreous sections (CEMOVIS) (14). By omitting the replacement of cellular water with organic solvents, embedding medium, and contrast-endowing chemicals, CEMOVIS allows the study of biological structures in their natural hydrated state, three-dimensionally and, potentially, at macromolecular resolution (reference 14 and reviews cited therein). Projection images from vitreous sections were corrected for thickness-dependent compression along the cutting direction (knife marks) by rescaling and interpolation and were low pass filtered for an improved signal-to-noise ratio.
The m74 open reading frame of MCMV is the positional homolog of HCMV gO (UL74), but sequence analysis shows only a low level of identity to gO. To study the role of m74 during virus infection, we constructed two BAC-derived m74 mutants of MCMV: a deletion mutant lacking the 532 N-terminal bp of m74 (pΔm74), and an m74 stop mutant carrying a stop cassette at position 120 (pm74stop) (Fig. (Fig.1).1). The m74 deletion was reversed by inserting the complete m74 ORF with and without a C-terminal HA tag into a preexisting FRT site at position m16/m17 in pΔm74 (pΔm74-m74+/−HA) (Fig. (Fig.1).1). The m74 knockout mutants and the ectopic insertions were controlled by restriction enzyme pattern analysis and by sequencing (data not shown). During virus reconstitution in mouse embryonal fibroblasts (MEF), we observed a strong delay in reconstitution with five independent BAC clones (three m74 deletion and two m74 stop mutants). Whereas after transfection of the parental BAC DNA plaques appeared 5 days later, foci of cells infected with the m74 knockout mutants appeared no earlier than 2 weeks after transfection (data not shown). In contrast to reconstitutions of parental BAC DNA, supernatants from reconstitutions of the m74 knockout mutants exhibited only low titers of infectious virus even if 100% of the cells were infected (data not shown).
A comparison of in vitro growth curves of vm74stop with w.t.-MCMV (Fig. (Fig.2A,2A, upper panel) and vΔm74 with w.t.-FRT-MCMV (Fig. (Fig.2B,2B, upper panel) revealed a 2- to 3-log-reduced release of infectious virus into the supernatant by the m74 mutants. To ensure that the observed differences reflect a difference in spread or replication of the m74 mutants and not in cell entry, cells were stained for MCMV immediate-early protein 1 (ie1) 24 h after infection. The percentage of initially infected cells was found to be comparable for wild type and m74 knockout mutants (data not shown). The attenuated phenotype of vm74stop and vΔm74 also became obvious when spread of the virus was followed in cell culture by staining for infected cells (Fig. 2A and B, lower panels). Infections with parental viruses readily spread throughout the culture, whereas infections with m74 knockout mutants stayed focal. Cells between foci were rarely infected, which strengthens the finding that infectious virus was not readily released into the supernatant.
To show that the observed phenotype is due to deletion of m74 and not to effects on neighboring genes, we cis or trans complemented the m74 knockout mutants. The attenuated growth phenotype of the m74 mutant viruses could be rescued by both complementation approaches. cis complementation was achieved by insertion of the complete m74 ORF into the m16/m17 FRT site within the pΔm74 genome as either a C-terminally HA-tagged version or an untagged version. Virus could be readily reconstituted (data not shown). Expression of the HA-tagged protein could be detected by Western blot analysis (see below). Virus production into the supernatant was rescued by both HA-tagged and untagged m74 (Fig. (Fig.3A).3A). Protein trans complementation of vΔm74 and vm74stop was achieved by infecting NIH 3T3 cells stably transfected with pEPI-m74 expressing the complete m74 ORF (NIH 3T3-m74). In NIH 3T3-m74 cells, release of infectious virus by wild-type parental viruses and that by m74 mutants were comparable (Fig. 3B and C). Expression of m74 in trans-complementing cells was controlled by analysis of m74 transcript levels in NIH 3T3-m74 clones and in parallel in NIH 3T3-m74HA clones (data not shown). For the NIH 3T3-m74HA clones, it could be shown that high transcription levels resulted in high levels of protein expression and of supernatant virus after infection with m74 mutant viruses (data not shown).
Recombination of mutated ORFs with intact ORFs expressed in trans-complementing cell lines has been observed. Due to the deletion of the 5′ end of m74 in the Δm74 mutant and the insertion of a kanamycin cassette, recombination in addition to trans complementation is highly unlikely. Virus progeny with wild-type properties was never observed when trans-complemented virus was passaged (>5 passages) in noncomplementing cells (data not shown). In addition, PCR amplification of DNA from mutant and trans-complemented virus progeny using primers which bind within the deleted N terminus of m74 was always negative and sequencing of the region always confirmed the introduced deletion (data not shown). For the vm74stop mutant, recombination in trans-complementing cells remains a possibility. Yet, also for vm74stop and trans-complemented vm74stop, virus progeny with wild-type properties was not observed. Loss of the stop cassette could be detected neither by PCR amplification using primers which discriminate between sequences carrying or lacking the stop cassette nor by sequencing m74 of trans-complemented virus progeny (data not shown).
Virus progeny of m74 knockout mutants grown in trans-complementing cells is complemented at the protein level. Infection of noncomplementing cells with the trans-complemented virus should thus result in a progeny revealing the knockout phenotype during subsequent infection of noncomplementing cells. This was observed when growth properties of m74 mutant viruses trans complemented by passage in NIH 3T3-m74 cells (vm74stop/m74trans and vΔm74/m74trans) and of noncomplemented m74 mutants were compared in cell culture. trans-complemented and noncomplemented m74 mutants showed the same reduction in release of infectious virus (Fig. (Fig.4).4). Comparable initial infection rates were controlled by staining cells 24 h after infection for MCMV ie1 protein (data not shown).
Recently, it has been shown that deletion of HCMV UL74 (gO) results in impaired release of infectious virus into the supernatant (20, 60). Electron microscopy studies of HCMV ΔUL74 mutants showed cytoplasmic accumulation of nonenveloped capsids, which might explain the observed phenotype (20). To find out whether the same is true for m74 deletion mutants of MCMV, we performed similar experiments. We used vΔm74 trans complemented in NIH 3T3-m74 to compare virus infection of w.t.-FRT-MCMV with that of vΔm74/m74trans in electron microscopy studies. Cells were infected at an MOI of 0.5. In infections with w.t.-FRT-MCMV, enveloped viral particles and particles in larger vesicles could readily be found in the cytoplasm (Fig. 5A and B) (59), whereas in cells infected with vΔm74 almost only naked capsids could be detected in the cytoplasm. Often, accumulations of naked capsids embedded in an amorphous structure were surrounded by vesicles (Fig. 5C and D), but virus particles were only rarely detected within these vesicles (data not shown). Most often, these vesicles were electron-optically empty, showing no signs of budding even if they were embedded in cytoplasmic capsid aggregates (Fig. 5C and D).
Comparing our conventional and cryo-electron microscopy approaches, we noticed a complementary character of the results. Whereas, for instance, the embedding amorphous mass of the cytoplasmic naked capsid aggregates was better visualized in stained sections (Fig. 5A and C), better preservation of the membrane and capsid structures was achieved with CEMOVIS (Fig. 5B and D).
To determine whether reduced release of infectious virus by m74 knockout mutants is mirrored by a reduced release of viral genome copies into the supernatants of infected cells, a quantitative PCR analysis was performed. NIH 3T3 cells were infected with w.t.-FRT-MCMV or vΔm74/m74trans at comparable initial infection rates (Fig. (Fig.6A).6A). Real-time PCR analysis of DNA from infected cells showed a comparable replication of the viral genome (Fig. (Fig.6B).6B). In supernatants harvested at several time points after infection, a comparable release of viral DNA was observed for wild-type (wt) and mutant virus (Fig. (Fig.6C),6C), although only w.t.-FRT-MCMV produced detectable amounts of infectious virus at the time points shown. To exclude the possibility that the viral DNA measured in supernatants represents free DNA released from damaged cells, supernatants were pretreated with DNase I. Additionally, gross cell damage by wt or mutant virus resulting in, for example, release of cytoplasmic capsids was excluded by measuring release of lactate dehydrogenase (LDH). For the 48-h period analyzed, LDH levels in supernatants from cells infected with wild-type or mutant virus were comparable to those for uninfected controls (data not shown). These results indicate that the low infectivity of supernatants of cells infected with m74 mutants is due rather to differences in secondary envelopment, resulting in virus which is less infectious, than to a complete block of particle release.
The m74 open reading frame predicts a protein of 438 amino acids containing seven potential glycosylation sites (16). For initial biochemical characterization of the protein product of m74, we transiently transfected the m74 open reading frame with a C-terminal HA tag under the control of the human CMV promoter into 293 cells (pCR3-m74HA). The HA-tagged protein was also cloned in pΔm74 under the control of an SV40 promoter (pΔm74-m74HA) and could thus be expressed in cis in the context of an MCMV infection with vΔm74-m74HA. In the Western blot the m74 protein product showed three major protein bands migrating at about 70, 55, and 45 kDa in reducing gels (Fig. (Fig.7A).7A). Treatment of total cell extracts of MEF cells infected with vΔm74-m74HA with endoglycosidase H (Endo H) resulted in a shift of the 70-kDa band to about 55 kDa, indicating that the m74 protein product is N glycosylated. In extracts of cells infected with vΔm74-m74HA, a band of about 250 kDa could be detected under nonreducing conditions. This band very likely represents a protein complex with the m74 protein (Fig. (Fig.7B).7B). As gO of HCMV forms a complex with gH/gL of a comparable size (24), we speculated that the m74 homolog also forms a complex with MCMV gH/gL. Accordingly, the 250-kDa complex could be precipitated with an anti-HA antibody and a monoclonal anti-gH antibody but not with an isotype control (Fig. (Fig.7B).7B). Under reducing conditions, the anti-gH antibody precipitated the 70-kDa band of the m74 protein but, in contrast to the anti-HA antibody, neither the 55-kDa band corresponding to the Endo H deglycosylated product nor smaller proteins. Two different exposures of the Western blot are depicted to show that coprecipitation is restricted to the 70-kDa band (Fig. (Fig.7C).7C). This strongly indicates that gH forms a specific complex with the mature glycosylated m74 protein.
The virus pairs vΔm74-vΔm74/m74trans and vm74stop-vm74stop/m74trans are genetically identical and differ only with respect to m74 protein complementation. We used these pairs to study the potential role of the m74 protein with regard to virus entry. As means of virus entry into cells, both fusion and endocytosis have been observed for most of the human herpesviruses tested. To differentiate entry pathways, we applied conditions which can discriminate endocytosis and fusion. Endocytosis is an energy-dependent process and can be blocked by inhibitors of ATP synthesis (28). Preincubation with energy depletion medium containing sodium azide and 2-deoxy-d-glucose had no effect on the infection of fibroblasts with w.t.-FRT-MCMV but significantly inhibited virus entry of the m74 mutant viruses (Fig. (Fig.8A).8A). trans complementation with m74 protein abolished the inhibitory effect (Fig. (Fig.8A).8A). Following endocytosis, some viruses need an acidic environment for penetration. Ammonium chloride (NH4Cl) buffers the pH of acidic cellular compartments and inhibits entry of viruses that require a low pH for release from endocytotic compartments. Bafilomycin A1 inhibits the vacuolar type H+-ATPases and thus also prevents acidification (61). To determine the role of pH for MCMV entry, infections with vΔm74, vΔm74/m74trans, and w.t.-FRT-MCMV were performed in the presence of increasing concentrations of NH4Cl and bafilomycin A1 (Fig. 8B and C). NH4Cl and bafilomycin A1 significantly inhibited entry of vΔm74 (P < 0.001, two-way analysis of variance [ANOVA]) but had no or only minor effects on w.t.-FRT-MCMV and vΔm74/m74trans entry. trans complementation with m74 protein completely abolished pH sensitivity of vΔm74 virus progeny (Fig. (Fig.8).8). The in vitro growth phenotype of vΔm74 resembles the growth phenotype described for HCMV gO knockout mutants. Therefore, we tested whether HCMV ΔgO mutants show a comparable phenotype regarding energy dependence and pH sensitivity of entry. Indeed, infection of human foreskin fibroblasts (HFF) with TB40-BAC4-derived vΔgO was significantly inhibited in the presence of energy depletion medium compared to vTB40-BAC4 wild-type virus (Fig. (Fig.9A).9A). The entry of a UL131A stop mutant of TB40-BAC4, exhibiting a disruption of the alternative gH/gL complex, was not sensitive to energy depletion (Fig. (Fig.9A).9A). Infections with vΔgO, but not with vTB40-BAC4 and vUL131Astop, were also significantly inhibited in the presence of NH4Cl and bafilomycin A1 (P < 0.001, two-way ANOVA) (Fig. 9B and C). For control, inhibitor pretreatment of cells or virus alone was tested. Cells and virus stocks were preincubated for 30 min with the highest concentration of inhibitor, and then either the inhibitor was washed away or the virus stocks were diluted and infection was compared to mock-treated cells. Inhibitor pretreatment did not affect virus entry (data not shown).
Alternative gH/gL glycoprotein complexes in herpesvirus envelopes have been described for a number of human herpesviruses (18, 32, 56). The core gH/gL complex promotes fusion of the viral envelope with host membranes during the viral entry process. Additional proteins complexed with gH/gL define the specificity for certain cell types, cellular receptors, and also entry pathways (15, 18, 33). Although different gH/gL complexes direct the virus to different cell types, it is not easily conceivable how the usage of one or the other complex is regulated. For Epstein-Barr virus (EBV), it has been shown that virus particles released from B cells lack the gH/gL/gp42 complex which is used for B-cell infection and are thus directed to epithelial cells and vice versa (4). Whether a similar mechanism applies for HCMV is currently not known. HCMV is one of the few herpesviruses for which a good mouse model exists. Thus, the role of gH/gL complexes of CMV might be studied by infection of mice with MCMV.
The MCMV genome has an extensive sequence homology to HCMV. HCMV proteins can even substitute for MCMV proteins in infection experiments (46). Yet, the positional homologs of the HCMV glycoproteins show only low levels of identity (16). This includes HCMV UL74 (gO) and MCMV m74 (20.2% identity and 34% similarity between HCMV TB40-BAC4 and MCMV Smith strain). There are no apparent conserved domains (data not shown). The two ORFs have about the same length (464 aa for UL74 and 438 aa for m74), and both are predicted to be highly N glycosylated (18 sites for UL74 versus 7 sites for m74). The UL74 sequence analysis shows a hydrophobic N-terminal membrane anchor and predicts an N-terminal signal peptide, which is cleaved off during maturation (49). Analyses of the m74 sequence gave heterogeneous results with respect to a potential signal peptide and showed no potential hydrophobic transmembrane domains (data not shown). The m74 protein product expressed in transfected cells and in cells infected with cis-complemented vΔm74 exhibited three major protein bands of about 70, 55, and 45 kDa when run in SDS-polyacrylamide gels under reducing conditions. Treatment with Endo H shifted the 70-kDa band to a band of about 55 kDa, showing that the m74 protein is like gO of HCMV N glycosylated. Whether the observed shift reflects complete or partial deglycosylation and whether glycosylation becomes Endo H resistant during protein maturation were not addressed. This would require extensive analyses of glycosylation, including pulse-chase experiments as performed for HCMV gO (17, 49, 50). Thus, we cannot conclude on comparable protein modifications of m74 and gO. Under nonreducing conditions, a complex of about 250 kDa is formed which is precipitated by a monoclonal antibody specific for gH of MCMV. Under reducing conditions, the anti-gH antibody exclusively coprecipitated the 70-kDa m74 protein, indicating that only the fully modified m74 protein forms a complex with gH.
More important, the functional homology of the UL74 and m74 proteins became evident when the m74 ORF was knocked out, either by deletion of 532 N-terminal base pairs or by insertion of a stop cassette at position 120 of the m74 DNA sequence. Deletion of the complete m74 reading frame was not possible, as the C-terminal end of the m74 ORF overlaps with the C-terminal end of the M73 ORF (gN). Deletion of the M73 3′ end very likely would result in a replication-deficient virus as described for HCMV UL73 mutants (27). The stop was inserted at position 120 to exclude the usage of alternative internal start codons between positions 1 and 120. Whereas the deletion of the m74 N-terminal 532 bp and the insertion of a kanamycin resistance cassette may also influence neighboring genes, a markerless insertion of a stop cassette usually does not disturb the genomic locus. The two m74 knockout mutants showed identical phenotypes which resembled the UL74 deletion mutant of HCMV in all respects (20). (i) Virus propagation in fibroblast culture was severely attenuated due to impaired release of infectious virus. (ii) Virus spread was mainly cell associated. It is currently not clear whether the low levels of infectious virus detected in supernatants of cell cultures infected with the knockout mutants represent free virus. They could also represent virus particles sticking to cellular debris which escapes centrifugal clearance of supernatants. Cells in the center of infected foci of knockout mutants regularly detach and are likely degraded (Fig. (Fig.2),2), a finding which might support the latter explanation. (iii) Electron microscopy revealed a defect in secondary envelopment for MCMV Δm74 as seen for HCMV ΔgO mutants (20). Accumulations of capsids in the cytosol were characterized by a lack of detection of enveloped particles. Yet, when we analyzed supernatants from infected cells, comparable numbers of viral genomes were found to be released from infections with wt and m74 knockout virus. DNase I treatment to remove free DNA and the absence of obvious cell damage which might result in release of naked capsids excluded two potential explanations for the release of noninfectious viral genomes. Hence, the data strongly suggest that the viral DNA released from cells infected with m74 knockout mutants might be packaged in enveloped particles which are not infectious or are much less infectious than wild-type particles. The differences in appearance of secondary envelopment of wild-type and m74 knockout virus thus do not necessarily represent different envelopment efficiencies. This is supported by two recent publications which showed that a ΔgO mutant of the HCMV strain TR released wild-type levels of extracellular virus particles which are impaired in infecting fibroblasts (44, 60).
Rescue of the m74 knockout mutants by cis and trans complementation with m74 proved that the loss of the m74 function and not indirect local genomic effects caused the observed phenotype. cis complementation was not complete, which might have several reasons. First, insertion of the m74 expression cassette results in a slight overlength of the genome, which may have an attenuating effect. Second, m74 was regulated by the SV40 promoter, which resulted in an overexpression of the m74 gene product compared to the expression of an HA-tagged m74 in its authentic locus (data not shown). This might affect cell viability or virus particle envelopment.
Altogether, the complex formation of the m74 protein with gH and the phenotype of the m74 deletion mutant show that the m74 protein is a true functional homolog of HCMV gO.
trans complementation with m74 rescued release of infectious virus by both m74 knockout mutants. Yet, the trans-complemented virus progeny retained the phenotype of impaired virus production and cell-associated virus spread in cell culture (Fig. (Fig.44 and data not shown). The trans-complemented viruses vΔm74/m74trans and vm74stop/m74trans differ from the m74 knockout mutants only with respect to the presence of m74 protein during virus particle maturation and very likely in the virion (21). We used the trans-complemented and the noncomplemented knockout mutants to study entry of genetically identical viruses differing only with respect to m74 protein complementation. Under energy depletion conditions, entry of m74 knockout mutants into fibroblasts was severely impaired whereas entry of parental and trans-complemented viruses was not affected. The latter viruses not only highlighted the importance of m74 for virus entry but also served as a control showing that energy depletion does not affect cell viability and thus indirectly entry success. The energy dependence of entry of m74 knockout mutants could be extended to a pH sensitivity of entry by using NH4Cl and bafilomycin A1 as inhibitors. Knockout of m74 thus directed virus entry into fibroblasts to an endocytotic pathway, whereas entry of viruses expressing m74 probably happens by fusion of the virions with the plasma membrane. Hence, trans-complemented m74 knockout mutants retained the m74 knockout phenotype with respect to impaired virus release and cell-associated virus spread but differed in the pathway used to enter fibroblasts. This indicates that m74 plays two roles in the MCMV life cycle, one in determining spread via free virus and one in determining the entry pathway into fibroblasts.
Knockout of m74 showed the same phenotype regarding virus spread as did knockout of UL74 of HCMV. Similarly, knockout of UL74 also rendered HCMV entry energy and pH dependent. When studying UL74-dependent HCMV entry into fibroblasts, we included a UL131A knockout mutant, which lacks the alternative gH/gL/pUL(128-131A) complex. This mutant expresses UL74 and consequently showed the same entry phenotype as did the parental TB40-BAC4-derived virus. The UL131A mutant confirmed that formation of a gH/gL/gO complex either during envelopment or in the particle envelope is sufficient for energy-independent entry into fibroblasts.
Different entry pathways into different cell types depending on envelope glycoprotein complexes have been described for several human herpesviruses (2, 29, 30, 34, 35). For HCMV, it has been reported that different virus strains use different entry pathways into endothelial or epithelial cells (37, 47, 57). To our knowledge, this is the first report that MCMV and HCMV also use different entry pathways into fibroblasts, a cell type for which fusion at the plasma membrane was long considered to be the only entry pathway (9).
The different entry pathways observed for w.t.-MCMV and Δm74 virus might be interpreted as that either the mutant mainly produces less-infectious particles carrying alternative gH/gL complexes which enter fibroblasts by an endocytotic pathway or it produces particles lacking gH/gL/gO complexes which are less infectious and enter if at all via endocytosis. For HCMV, the first scenario is more likely. We have observed that infection of fibroblasts with virus derived from ΔgO mutants of TB40-BAC4 is efficiently blocked by antibodies directed against the alternative gH/gL/pUL(128-131A) complex (B. Adler, unpublished observation). Thus, enveloped particles released by ΔgO mutants seem to infect fibroblasts using the alternative gH/gL/pUL(128-131A) complex. As shown here, ΔgO virus progeny enters through an endocytotic pathway, which very likely is mediated through binding of gH/gL/pUL(128-131A).
Future studies will have to show whether MCMV, like HCMV, uses an alternative gH/gL complex comparable to the gH/gL/pUL(128-131A) complex of HCMV. Virus in which one or the other gH/gL complex is deleted might then provide important insights into the role of distinct gH/gL complexes in cytomegalovirus infection and spread in vivo.
We thank Heike Hofmann and Adrian Prager for excellent technical assistance and Heiko Adler for critically reading the manuscript. The pEPi vector with blasticidin resistance was kindly provided by Martin Lipps (University of Witten Herdecke, Germany) and Rudolf Haase (LMU München, Germany). The conventional electron microscopy was performed in the laboratory of York Stierhof (Zentrum für Molekularbiologie der Pflanzen, University of Tuebingen, Germany).
This work was supported by the DFG through an individual grant to Barbara Adler (AD131/3-1), through the SPP1130 “Infections of the endothelium” (AD 131/2-3 and SI779/3-3), and through the DFG priority program 1775 “Dynamics of cellular membranes and their exploitation by viruses” (grant GR1990/2-1).
Published ahead of print on 24 February 2010.