|Home | About | Journals | Submit | Contact Us | Français|
Rhesus cytomegalovirus (RhCMV) is an emerging model for human cytomegalovirus (HCMV) pathogenesis that facilitates experimental CMV infection of a natural primate host closely related to humans. We have generated a library of RhCMV mutants with lesions in genes whose HCMV orthologues have been characterized as nonessential for replication in human fibroblasts, and we characterized their replication in rhesus fibroblasts and epithelial cells. The RhCMV mutants grew well in fibroblasts, as predicted by earlier studies with HCMV. However, mutations in four genes caused replication defects in rhesus retinal pigment epithelial cells: Rh01 (an HCMV TRL1 orthologue), Rh159 (HCMV UL148), Rh160 (HCMV UL132), and Rh203 (HCMV US22). Growth of the Rh01-deficient mutant was examined in detail. After entry into epithelial cells, the mutant expressed representative viral proteins, accumulated viral DNA, and generated infectious virus, but it failed to spread efficiently. We conclude that Rh01 is a cell tropism determinant that has the potential to dramatically affect virus spread and pathogenesis.
Human cytomegalovirus (HCMV) is the prototypical β-herpesvirus and is carried as a latent infection by the majority of the worldwide population (26). Primary HCMV infection and reactivation from latency are generally asymptomatic but can lead to serious disease in patients with poorly developed or compromised immune systems such as neonates, AIDS patients, and organ transplant recipients. HCMV is strictly species specific, which has restricted studies aimed at understanding the pathogenesis of the virus (26). There are rodent models, most notably murine cytomegalovirus (MCMV), that have contributed importantly to our understanding of CMV-host interactions (22). Nevertheless, there are significant limitations in comparing rodent models to HCMV infection because the gene content of these two CMVs is markedly different (30). More recently, rhesus cytomegalovirus (RhCMV) infection of rhesus macaques has emerged as an attractive model system for HCMV pathogenesis. The RhCMV genome is substantially more closely related to HCMV than are the genomes of rodent CMVs (17, 31), and the epidemiology, pathogenesis, and immunological responses to RhCMV infection in macaques are similar to HCMV infection of humans (20, 25, 36). The genome of RhCMV strain 68-1 has been cloned as a bacterial artificial chromosome (BAC) to facilitate genetic manipulation of the virus (9) and, with the knowledge and experimental approaches made available with the recent completion of the rhesus macaque genomic sequence (14), the RhCMV model will certainly accelerate our understanding of HCMV pathogenesis.
CMVs infect a variety of different cell types in the host, including fibroblasts, epithelial cells, endothelial cells, and macrophages (28). Epithelial cells are among the first to be infected in a new host, and it is also believed that glandular epithelial cells are the source of virus that is shed during persistent stages of infection (23). Retinal pigment epithelial cells are a site of persistent HCMV infection in immunocompetent hosts, and HCMV-induced retinitis is one of the clinical manifestations of AIDS (34).
It has been shown that UL128, UL130, and UL131 are required for efficient entry of HCMV into epithelial cells, endothelial cells, and dendritic cells but are dispensable for replication in fibroblasts (1, 13, 16, 37, 38). Mutations in this locus are common in laboratory strains of HCMV, which have restricted cell tropism (2, 16). In addition to UL128-UL131, there have been few reports of additional tropism determinants in CMVs. A functional screen of mutants in the Towne strain of HCMV indicated that UL24 is important for replication in microvascular endothelial cells and that UL64 and US29 contribute to efficient replication in retinal pigment epithelial cells (11). RhCMV carries a cyclooxygenase-2 homologue that is required for efficient RhCMV replication in microvascular endothelial cells (32). The MCMV ribonucleotide reductase homologue M45 has been shown to be critical for MCMV replication in endothelial cells, although the HCMV orthologue, UL45, does not share this characteristic (7, 15).
We identified a rhesus epithelial cell (primary retinal pigment epithelial [RRPE] cells) that can be infected by wild-type RhCMV, generated a library of RhCMV mutants with a focus on genes that have HCMV orthologues that are nonessential for replication in cultured fibroblasts, and then screened the mutants for their ability to grow in rhesus fibroblasts compared to RRPE cells. We have identified four RhCMV genes (orthologous to HCMV TRL1, UL132, UL148, and US22) that are dispensable for replication in fibroblasts but required for normal RhCMV growth in epithelial cells. The TRL1-deficient virus was studied in detail and shown to produce virions in RRPE cells that are deficient for spread. The use of RhCMV provides the long-term potential to evaluate the role of these host-range genes in a well-established primate model.
Dermal rhesus macaque fibroblasts (telo-RFs) (21), which were life extended by the addition of a constitutively expressed human telomerase subunit (hTERT), were maintained in Dulbecco modified Eagle medium supplemented with 7.5% fetal calf serum. Rhesus macaque RRPE cells (5) were maintained in a 1:1 combination of Dulbecco modified Eagle medium and Ham F-12 nutrient mixture supplemented with 1 mM sodium pyruvate, nonessential amino acids (Gibco), and a final concentration of 5% fetal calf serum. RRPE cells were used for infection between passages 10 to 25. Cultured cells were incubated at 37°C in a 5% CO2 atmosphere.
For transposon insertion mutagenesis (40), the transposon delivery vector pYD-1721 was electroporated into E. coli DH10B cells containing pRhCMV/BAC-Cre. Transformed clones were selected for growth at 30°C on LB agar plates containing 100 μg of ampicillin/ml and 15 μg of chloramphenicol/ml. The resulting colonies were grown at 30°C in LB broth containing the same antibiotics. These cultures were then used to inoculate LB agar plates containing 15 μg of chloramphenicol/ml and 25 μg of kanamycin/ml, followed by incubation at 43°C. The resulting colonies were replica plated onto LB agar plates containing 100 μg of ampicillin/ml to confirm loss of the temperature-sensitive delivery vector and 15 μg of chloramphenicol and 25 μg of kanamycin/ml to maintain selection on the desired clones. Cells from the latter colonies were grown up and used for restriction enzyme digests and sequence analysis to confirm the integrity of the mutant BAC and to determine the position of transposon insertion.
Site-directed mutagenesis was done by linear recombination essentially as described previously (24). In brief, 77- to 79-nucleotide (nt) PCR primers were designed for each targeted gene: the 5′ 51 nt were complementary to the targeted gene and the 3′ 26 to 28 nt were complementary to the 1.8-kb kanamycin/lacZ cassette in pYD-C54 (41). Each primer pair was used to amplify the kanamycin/lacZ cassette, and the resulting products were introduced into heat-induced E. coli DY380 containing pRhCMV/BAC-Cre by electroporation. Recombinant clones were selected by incubating for 36 to 48 h at 30°C on YENB (7.5 g of yeast extract/liter, 8 g of nutrient broth/liter) agar plates containing 15 μg of chloramphenicol/ml, 25 μg of kanamycin/ml, IPTG (isopropyl-β-d-thiogalactopyranoside), and X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside). The resulting blue colonies were grown up and used for restriction enzyme digests and sequence analysis to verify the desired product.
The Rh01 ORF in RhCMV was FLAG tagged by using Escherichia coli strain SW105, which in addition to facilitating linear recombination also allows for arabinose-inducible expression of Flp recombinase (39). PCR primers (74 to 78 nt) were designed to substitute the stop codon of the TRL1 open reading frame (ORF) with a FLAG-FRT-kanamycin-FRT cassette that was amplified from pSLFRTKn (4). Linear recombination was performed as described above, with the following addition (24): after selection on YENB agar plates with chloramphenicol and kanamycin at 30°C, colonies were picked and grown in YENB liquid culture overnight at 30°C and then diluted 50-fold and grown again at 30°C until the optical density at 600 nm was 0.5. Flp expression was induced by the addition of l-arabinose (0.1% final concentration), followed by incubation at 30°C for 1 h. After a 10-fold dilution, the cultures were incubated at 30°C for another hour, plated on YENB agar plates with 15 μg of chloramphenicol/ml, and incubated for 36 to 48 h at 30°C. The integrity of the TRL1FLAG RhCMV BAC was confirmed by restriction enzyme analysis, and the desired mutation was verified by sequence analysis.
Mutant BACs were electroporated into telo-RFs to reconstitute the corresponding mutant viruses. Cell-free virus was harvested from infected cultures 1 day after all cells showed cytopathic effect. These primary stocks were expanded once by low-multiplicity infections of telo-RFs and harvested at 100% cytopathic effect.
Centrifugal enhancement was used to increase the efficiency of RRPE cell infection. After the addition of the virus inoculum, cell monolayers were subjected to centrifugation at 500 × g for 30 min at room temperature. Cells were then allowed to recover for 30 min at 37°C in a 5% CO2 atmosphere before being washed with phosphate-buffered saline (PBS) and the addition of fresh medium. To monitor viral growth kinetics, telo-RFs or RRPE cells in six-well dishes at ca. 90% confluence were infected at various multiplicities. Supernatants were harvested after various intervals of infection, cleared by centrifugation, and stored at −80°C. To harvest cell-associated virus, cells were scraped into a small volume of culture medium and pelleted. Washed pellets were resuspended in a small volume of medium and sonicated to release intracellular virus. Cellular debris was then pelleted by centrifugation and the resulting supernatants stored at −80°C. Infectivity was determined by 50% tissue culture infective dose (TCID50) assay on telo-RFs.
To generate monoclonal antibodies to RhCMV proteins, portions of the Rh156 (HCMV UL123 orthologue), Rh70 (HCMV UL44), and Rh112 (HCMV UL83) ORFs were amplified by PCR using the following primer pairs: (i) UL123fwd, 5′-AAGCTTTTACTTGTCAGTCTTGCTTCTGGT-3′; and UL123rev, 5′-TCTAGATATGGGCATAGATAAGATGAAGTTCA-3′; (ii) UL44fwd, 5′-AAGCTTCTATGTACATTTCTGCTTTTTGCTG-3′; and UL44rev, 5′-TCTAGATATGCGCGTTCACGTGCAGCTCAAGA-3′; and (iii) Rh112fwd, 5′-AAGCTTCTAACTACGGTGCTTTTTAGGAAC-3′; and Rh112rev, 5′-TCTAGATATGATTCTCACCAAGGGTACCGC-3′. The amplified DNAs were inserted into pGTK, and propagated in E. coli BL21 cells. The sequences of the cloned constructs were verified, and the chosen clones were used to produce glutathione S-transferase fusion proteins that were purified and used to immunize BALB/c mice. After seroconversion, spleens were harvested and used to generate hybridomas that were subsequently screened for reactivity and specificity to RhCMV-infected cell lysates, cloned by dilution, and amplified.
To analyze proteins by Western blot, cells were infected at a multiplicity of 3 TCID50/cell for 1 h (telo-RFs) or 3 TCID50/cell with centrifugal enhancement (RRPE cells). At the indicated times postinfection, the cells were scraped from the culture plate into PBS and pelleted, washed in PBS, and stored at −80°C until all samples had been collected. Each cell pellet was resuspended in radioimmunoprecipitation assay buffer: 50 mM Tris (pH 7.4), 150 mM sodium chloride, 1% Nonidet P-40, 0.25% sodium deoxycholate, 1 mM EDTA, and Complete Mini, EDTA-free protease inhibitor cocktail (Roche). The lysate was sonicated and centrifuged to pellet insoluble debris. Polypeptides were separated by electrophoresis in sodium dodecyl sulfate-containing polyacrylamide gels and transferred to nitrocellulose membranes. The membranes were blocked in 0.2% Tween 20 in PBS (PBST) with 10% nonfat dry milk for ≥3 h, probed with primary antibodies for 1 h in PBST with 2% bovine serum albumin (BSA), washed with PBST, probed with an anti-mouse secondary antibody conjugated to horseradish peroxidase for 1 h in PBST with 2% BSA, washed in PBST, treated with ECL Western blotting reagent (Amersham), and exposed to film.
For analysis of intracellular localization by fluorescence microscopy, telo-RFs or RRPE cells cultured on glass coverslips in six-well dishes were infected at a multiplicity of 0.5 or 3 TCID50/cell, respectively. At various times after infection, cells were washed with PBS, fixed for 15 min at 37°C with 2% paraformaldehyde, washed, permeabilized with 0.1% Triton X-100 in PBS, and then washed with PBST. Next, the cells were incubated for ≥2 h with blocking buffer made up of PBST with 2% BSA, reacted for 1 h at room temperature with a primary antibody in blocking buffer, washed with PBST, incubated for 1 h in blocking buffer with goat anti-mouse Alexa 488-conjugated secondary antibody (Molecular Probes) and DAPI (4′,6′-diamidino-2-phenylindole; Molecular Probes), washed with PBST, mounted on glass slides with Slow Fade solution (Molecular Probes), and viewed on a Zeiss LSM 510 confocal microscope.
For counting Rh156-positive nuclei, cells were cultured in six-well dishes and infected at a multiplicity of 1 TCID50/cell (telo-RFs) or 3 TCID50/cell with centrifugal enhancement (RRPE cells). At 24 h postinfection, the infected cells were fixed for 15 min at 37°C with 2% paraformaldehyde, permeabilized, reacted with antibodies, and viewed on a Nikon Eclipse TE200 fluorescence microscope. The percent Rh156-positive nuclei corresponds to the ratio of the number of nuclei seen in the green channel (Alexa 488) to the number of nuclei seen in the blue channel (DAPI). The number given is the mean percentages of four randomly chosen fields.
To monitor intracellular RhCMV DNA accumulation by quantitative real-time PCR (qPCR), cells were infected at a multiplicity of 0.1 TCID50/cell (telo-RFs) or 3 TCID50/cell with centrifugal enhancement (RRPE cells). Infected cells were harvested by trypsin treatment, pelleted by centrifugation, and stored at −80°C until all samples had been collected. To prepare DNA, each pellet was resuspended in 500 μl of qPCR lysis buffer (10 mM Tris [pH 7.5], 400 mM NaCl, 10 mM EDTA, 40 μg of proteinase K/ml, 0.2% SDS), agitated on a vortex mixer until resuspended, and incubated at 37°C overnight. The samples were phenol-chloroform extracted, treated with 40 μg of RNase A/ml for 1 h at 37°C, phenol-chloroform extracted, ethanol precipitated, and resuspended. To quantify genome-containing RhCMV particles in culture medium, samples were taken from infected cells, centrifuged to pellet cellular debris, and the supernatant was stored at −80°C until all samples had been collected. To prepare viral DNA, 100 μl of sample was mixed with 400 μl of qPCR lysis buffer and treated as described above. qPCR analysis was performed by using SYBR Green PCR Master Mix (Applied Biosystems) according to the manufacturer's recommendations. Viral DNA was quantified using primers specific to RhCMV UL123 (5′-TCTGCATATGGTGCTTGCTC-3′ [qRhUL123fwd] and 5′-GGAAGAGGAAGGTGCTGGAC-3′ [qRhUL123rev]). Cellular DNA was quantified using primers specific to the rhesus tubulin gene (10). Absolute quantities of RhCMV genomes from virus particles were obtained by using a standard curve made from known concentrations of pRhCMV/BAC-Cre. Relative quantities of intracellular RhCMV genomes and tubulin were obtained from a standard curve made from 10-fold serial dilutions of a sample from the final time point of the wild-type RhCMV infection.
RhCMV replication is generally studied in fibroblasts, and we wanted to identify and study viral genes that influenced tropism in epithelial cells. Consequently, it was necessary to identify a rhesus epithelial cell that is permissive for replication of wild-type RhCMV before we could initiate a screen for host range mutants. We chose to examine primary retinal pigment epithelial cells since HCMV establishes a persistent infection in this cell type (34), and persistent infection likely requires a different portion of the viral proteome than basic, lytic replication.
Figure Figure1A1A shows RRPE cells (5) infected with RhTn106, an RhCMV variant that expresses green fluorescent protein (GFP) from a transposon inserted into the intergenic region between the Rh142 (HCMV UL105 orthologue) and Rh143 (HCMV UL111A) ORFs and grows with wild-type kinetics in telomerase life-extended fibroblasts (telo-RFs) (data not shown). Small GFP-positive foci and individual GFP-positive cells are present at 1 week postinfection of RRPE cells (Fig. (Fig.1A,1A, top panels) and, at 5 weeks, large infected foci are evident by bright-field and fluorescence imaging as clusters of GFP-positive cells displaying a typical CMV cytopathic effect (Fig. (Fig.1A,1A, bottom panels). The foci are different from plaques formed on fibroblasts in that there is little or no disruption of the underlying cell monolayer (Fig. (Fig.1A,1A, bottom left panel). If the infections are allowed to progress further, cell-free areas in the RRPE cell monolayers eventually appear (data not shown). These observations are consistent with a slow, smoldering lytic infection, an infection that is clearly different than that observed in telo-RFs.
RhCMV successfully infected the epithelial cells much less efficiently than fibroblasts. When the infectivity of a wild-type virus stock was determined by TCID50 assay on the two cell types, the measured titer was ~1,000-fold less on RRPE cells compared to telo-RFs (Fig. (Fig.1B1B).
We analyzed the accumulation of cell-free and cell-associated virus in RRPE cells compared to telo-RFs after infection with wild-type RhCMV. When telo-RFs were infected at a multiplicity of 0.01 TCID50/cell, RhCMV generated 8 × 107 TCID50 of extracellular virus/ml and 2 × 106 TCID50 of intracellular virus/ml by 12 days after infection with complete cytopathic effect. RRPE cells were infected at a much higher multiplicity (3 TCID50/cell) with centrifugal enhancement. A substantial amount of virus was detected in infected RRPE cultures for the entire 6-week period assayed (Fig. (Fig.1C,1C, right panel), but an eclipse period was clearly evident when viral DNA accumulation was assayed at 1 day after infection (see Fig. Fig.4B).4B). The culture medium was changed each week, and virus was present in each weekly sample, demonstrating that a high level of infectious virus was produced on a continuing basis. Similar amounts of cell-free and cell-associated virus were produced in RRPE cultures (~106 TCID50/ml on week 6), and the amount of infectivity generated in RRPE cells at 6 weeks postinfection was only 10 to 100-fold lower than the maximum achieved in telo-RFs (Fig. (Fig.1C,1C, compare left and right panels). The extended period of virus production presumably results from a combination of slow virus spread and continued growth of RRPE cells within the infected cultures. Indeed, this high level of virus production can be maintained for at least 20 weeks (data not shown).
We conclude that cultured primary RRPE cells are permissive for RhCMV infection and produce substantial titers of extracellular virus over an extended time frame. However, relative to growth in fibroblasts, the initial infection is inefficient and spread is slow.
Our next goal was to identify RhCMV mutants that grew normally on telo-RFs but were deficient for replication in RRPE cells. To generate a library of mutants in pRhCMV/BAC-Cre (9), we initially used a random transposon insertion strategy (40). We generated 244 transposon insertion mutants, but, unfortunately, two hot spots for transposon insertion limited the utility of this approach. Approximately 50% of the transposon insertions targeted the Rh122-134 region (orthologous to HCMV UL87-98), and another 10% were inserted into the Rh80-83 region (HCMV UL49-52) (data not shown). Nevertheless, this procedure targeted 39 distinct ORFs with known HCMV orthologues (multiple transposon insertions were obtained for many of these ORFs) and four ORFs unique to RhCMV.
Due to the transposon insertion hot spots, we changed our approach and used site-directed mutagenesis mediated by linear recombination (24) to specifically target ORFs in pRhCMV/BAC-Cre. Since our desire was to study the effects of single-gene mutations on cell host range, we targeted RhCMV orthologues of HCMV ORFs that are nonessential for replication in cultured fibroblasts. Substituting all or part of each targeted RhCMV ORF for a kanamycin/lacZ cassette in individual BACs, we knocked out 36 HCMV orthologues and one RhCMV-specific ORF that were not targeted by transposon insertion, and we made additional mutations in several transposon-targeted ORFs (Table (Table1).1). All but two of the site-directed mutations disrupted a single gene. Rh214, Rh215, Rh216, Rh218, and Rh220 (multiple ORFs orthologous to HCMV US28) were targeted in one mutation, as were Rh221, Rh223, Rh225, Rh226, and rh228 (orthologous to HCMV US29-US32, except for rh228, which is unique to RhCMV).
In sum, the library of 72 RhCMV mutants, generated by a combination of random transposon insertion and site-directed mutation, targets 80 RhCMV ORFs (Table (Table1).1). The mutations disrupt 75 RhCMV ORFs orthologous to HCMV ORFs, corresponding to ~50% of the known HCMV sequence orthologues in the rhesus virus (17, 31). Five of the mutations target ORFs unique to RhCMV.
We initially monitored the growth of the RhCMV mutants in rhesus fibroblasts (telo-RFs). The BAC DNAs were delivered into telo-RFs by electroporation, and mutants that failed to produce progeny virus after two electroporation attempts were considered likely to contain lesions within ORFs essential for lytic RhCMV replication in fibroblasts. A total of 27 essential RhCMV ORFs were identified by this criterion (Table (Table1),1), and, with the exception of Rh103 and Rh132, the corresponding HCMV genes have been described as essential in at least one viral strain (11, 40). The HCMV orthologues of Rh103 and Rh132 are UL74 and UL97, respectively, and both ORFs have been defined as augmenting, albeit with severe growth defects (18, 29).
The screen identified 45 RhCMV mutants (43 disrupting single ORFs and two disrupting multiple ORFs) that produced progeny after transfection, and their multistep growth kinetics in telo-RFs were monitored by assaying the amount of infectious virus that accumulated in the supernatant of infected cultures over time. Mutants whose extracellular accumulation of infectious particles were at least a factor of 10 lower than parental, wild-type RhCMV at any time point assayed defined ORFs that are augmenting for RhCMV replication (Table (Table1).1). Four mutants disrupting an augmenting RhCMV ORF were identified, a finding consistent with earlier studies showing that the corresponding HCMV ORFs (UL35, UL72, UL82, and UL88) are augmenting (6, 8, 11, 33, 40). Mutants that grew with normal kinetics, i.e., producing amounts of extracellular infectivity within a factor of 10 of wild-type virus at all time points assayed, were defined as genes that are nonessential for RhCMV replication in cultured telo-RFs (simply referred to as nonessential from this point). A total of 41 mutants disrupting 49 RhCMV ORFs fall into this category (Table (Table1).1). Of the 49 nonessential RhCMV ORFs, 45 have HCMV orthologues, all of which have been shown to be nonessential in at least one strain of HCMV (11, 40).
We next screened the 45 mutants with HCMV orthologues, which were able to grow in fibroblasts, for growth in RRPE cells. Virus stocks produced and titered in telo-RFs were used to infect RRPE cells at a multiplicity of 3 TCID50/cell, and the accumulation of extracellular virus was monitored over time. Four mutants that grew with wild-type kinetics in telo-RFs proved to exhibit a ≥10-fold growth defect in RRPE cells (Fig. (Fig.2).2). Rhsub01 (HCMV TRL1) grew indistinguishably from wild-type virus in telo-RFs (Fig. (Fig.2B,2B, left panel) but was severely defective in RRPE cells, producing >1,000-fold less infectious extracellular virus than wild-type RhCMV (Fig. (Fig.2A,2A, right panel). Similarly, Rhsub159 (HCMV UL148) and Rhsub160 (HCMV UL132) exhibited ~100-fold growth defects, and Rhsub203 (HCMV US22) grew to an ~1,000-fold reduced yield in RRPE cells (Fig. 2B to D, right panels), while growing normally in telo-RFs (Fig. 2B to D, left panels). Consistent with the virus yields, the foci of infected RRPE cells produced by these four mutants were smaller than those observed for wild-type virus at all times assayed (Fig. (Fig.66 and data not shown). The majority of the screened mutants grew with yields that were similar to the wild-type virus in both telo-RFs and RRPE cells, as shown for a mutant targeting the five HCMV US28 orthologues in RhCMV (Fig. (Fig.2E2E).
To verify that the growth defects exhibited by these mutants resulted from inactivation of the targeted genes, we generated a second set of identical, but independently derived, mutants. In each case, the independently made mutant exhibited a growth defect similar to that documented for the original mutant, as shown in Fig. Fig.2F2F for a second Rhsub01 mutant. We conclude that the Rh01 (HCMV TRL1), Rh159 (HCMV UL148), Rh160 (HCMV 132), and Rh203 (HCMV US22) ORFs are important for RhCMV replication in RRPE cells, in spite of being fully dispensable for replication in cultured telo-RFs.
We decided to examine the growth defect in Rhsub01 in greater detail. The RhCMV Rh01 ORF is predicted to code for a 508-amino-acid polypeptide that is ~37% identical to its considerably smaller HCMV TRL1 orthologue (17). Curiously, the TRL1 ORF in AD169, an HCMV laboratory strain with restricted host cell tropism, is truncated compared to TRL1 genes in other HCMV strains (27). Since the expression of a TRL1-coded protein has not been demonstrated for any CMV to date, we used an epitope tag to test for pRh01 expression in RhCMV. Using linear recombination, a short DNA segment encoding a FLAG tag was fused to the C terminus of the Rh01 ORF in pRhCMV/BAC-Cre, and a virus expressing epitope-tagged pRh01 (Rhin01-FLAG) was recovered. The procedure left a residual 34-bp FRT site and an additional 29-bp noncoding sequence downstream of the FLAG tag.
After establishing that Rhin01-FLAG grew with normal kinetics in both fibroblasts and RRPE cells (Fig. (Fig.3A),3A), we used it to assay for the accumulation of FLAG-tagged pRh01 by performing Western blots on lysates of infected telo-RFs. A band specific to the FLAG antibody was evident at 24 h postinfection, and it increased in amount by 48 and 72 h postinfection (Fig. (Fig.3B,3B, lanes 4 to 6). This expression pattern is similar to that observed for pRh70 (see Fig. Fig.5A),5A), an orthologue of the HCMV pUL44 DNA polymerase accessory factor, which is expressed with early kinetics. The size of the FLAG-specific band is ~60 kDa, corresponding well with the predicted molecular mass of ~57 kDa (56 kDa for pRh01 plus 1.0 kDa for the tag). We repeated this experiment in RRPE cells and identified a FLAG-specific band of the expected size that appears beginning at 4 weeks postinfection (Fig. (Fig.3C).3C). Presumably, the RRPE cultures contained too few infected cells for the protein to be detected at earlier times. We also used Rhin01-FLAG to examine the localization of TRL1 in infected telo-RFs and RRPE cells by indirect immunofluorescence assay. In infected telo-RFs, the FLAG antibody showed a relatively punctate cytoplasmic distribution, a finding consistent with the localization of pRh01 to vesicles or endosomes, with no apparent nuclear staining (Fig. (Fig.3D,3D, left panel). In infected RRPE cells, FLAG staining was also seemingly exclusively cytoplasmic but exhibited a more diffuse distribution compared to telo-RFs (Fig. (Fig.3D,3D, right panel).
We conclude that pRh01 is a cytoplasmic protein with a molecular mass of ~56 kDa. pRh01 appears to be expressed with early kinetics, although we did not perform transcription analysis with the appropriate inhibitors to show this conclusively.
We investigated the role of pRh01 in RRPE cells by comparing the replication cycle of wild-type RhCMV to that of Rhsub01. To monitor viral entry and the initiation of immediate-early gene expression, telo-RFs and RRPE cells were infected at a multiplicity of 1 or 3 TCID50/cell, respectively, with the wild-type or mutant virus, and pRh156-positive nuclei were assayed by immunofluorescence 24 h later. pRh156 is orthologous to the HCMV UL123-coded IE1 protein. Wild-type virus and Rhsub01 expressed pUL156 in ca. 30 and 40% of telo-RFs, respectively (Fig. (Fig.4A).4A). As expected, infection of RRPE cells was much less efficient; only 5% (wild-type virus) and 4% (Rhsub01) of these cells expressed pRh01 at 24 h after infection at a multiplicity of 3 TCID50/cell with centrifugal enhancement (Fig. (Fig.4A).4A). This low efficiency of pRh156 expression does not result from delayed expression of the immediate-early gene because the number of expressing cells was not different at 48 h postinfection (data not shown).
We used qPCR to assay viral DNA accumulation in wild-type-virus-infected compared to mutant-virus-infected telo-RFs and RRPE cells. We normalized the number of RhCMV genomes to the number of tubulin gene copies in each sample to produce a relative measure of RhCMV genomes per cell. In telo-RFs, wild-type and Rhsub01 accumulate viral genomes at a similar rate throughout infection (Fig. (Fig.4B,4B, left panel). Similar amounts of wild-type and mutant DNA had become cell associated at 1.5 h after infection at a multiplicity of 0.1 TCID50/cell, and, after a dip in DNA levels at 24 h postinfection, the levels of viral DNA in both infections increased steadily through the end of the 120-h period monitored. The patterns of wild-type and mutant DNA accumulation were also similar after infection of RRPE cells at a multiplicity of 3 TCID50/cell with centrifugal enhancement (Fig. (Fig.4B,4B, right panel). However, the mutant reproducibly accumulated somewhat more DNA than the wild-type virus at 48 h postinfection. By 72 h postinfection, the two viruses had produced similar amounts of DNA, and neither virus accumulated as many progeny genomes in RRPE cells as in telo-RFs, presumably because the viruses are rapidly spreading in fibroblast but not in epithelial cell cultures.
We next examined the accumulation and localization of a representative set of RhCMV proteins: pRh156, pRh70, and pRh112. As mentioned above, pRh156 and pRh70 are orthologues of the HCMV immediate-early protein, pUL123 (IE1), and the early protein, pUL44, respectively. pRh112 is one of two RhCMV orthologues of the HCMV late protein, pUL83 (17, 42). During the 72-h period tested, the three viral proteins accumulated with similar kinetics in telo-RFs after infection with wild-type or mutant virus (Fig. (Fig.5A).5A). pRh156, pRh70, and pRh112 were first detected at 8, 24, and 48 h postinfection, respectively, a finding consistent with their tentative classification as immediate-early, early, and late proteins by analogy to HCMV. Two pRh156 isoforms with apparent molecular masses of ~60 and ~80 kDa were detected; the pRh70-specific antibody detected a single band of ~45 kDa, and the pRh112 antibody detected an ~60-kDa species. The only difference evident between the two viruses was the presence of pRh112 at 1.5 h postinfection with the mutant but not with the wild type. The accumulation of the same set of RhCMV proteins was examined in RRPE cells (Fig. (Fig.5B)5B) and, again, the principal difference between the two viruses was the presence of a greater quantity pRh112 in Rhsub01-virus-infected compared to wild-type-virus-infected cells. We also observed a subtle delay in the accumulation of newly synthesized pRh70 (early) and pRh112 (late) proteins in the mutant compared to wild-type-virus-infected cells.
Immunofluorescence assays were performed to examine the localization of the RhCMV proteins in mutant compared to wild-type-virus-infected cells (Fig. 5C and D). pRh156, pRh70, and pRh112 were assayed at 24, 48, and 72 h postinfection, respectively. pRh156 and pRh70 were localized to the nucleus, and pRh112 localized predominantly to the cytoplasm with some protein at the intranuclear periphery. These localizations were independent of the cell type, and they were not influenced by the presence of pRh01.
When introduced to RRPE cultures at a relatively high input multiplicity with centrifugal enhancement, the mutant successfully enters, replicates its genome, and expresses a set of three representative viral proteins that become properly localized.
To monitor virus spread, we infected RRPE cells at a multiplicity of 3 TCID50/cell with centrifugal enhancement and assayed for pRh156-positive cell foci after various time intervals (Fig. (Fig.6A).6A). Individual cells or very small groups of cells expressed pRh156 at 1 week after infection with wild-type virus. These foci grow in size over time and by 5 weeks after infection the majority of cells in the culture expressed pRh156 (Fig. (Fig.6A,6A, left panels). In contrast, after infection with Rhsub01, pRh156 expression was restricted to individual cells and small foci throughout the 5-week period of observation (Fig. (Fig.6A,6A, right panels). Consistent with a failure to spread, Rhsub01 failed to generate a significant accumulation of intracellular or extracellular infectious virus during the 5-week period (Fig. (Fig.6B6B).
Although Rhsub01 failed to spread and generate normal levels of infectious virus in infected RRPE cultures, it remained possible that it produced substantial amounts of noninfectious particles. To test this possibility, we compared the accumulation of infectivity and DNA-containing particles in the medium of telo-RFs compared to RRPE cells. In fibroblasts infected at a multiplicity of 0.1 TCID50/cell, the wild-type and mutant viruses generated similar amounts of infectivity (Fig. (Fig.7A)7A) and DNA-containing particles (Fig. (Fig.7C),7C), and the ratio of genome-containing particles to infectious particles was similar for the two viruses at each time tested (Fig. (Fig.7E).7E). For infection of RRPE cells, we used a higher input multiplicity, 3 TCID50/cell, with centrifugal enhancement. Both Rhsub01 and wild-type underwent an eclipse period on days 1 and 2 and then generated similar yields of extracellular infectious virus on days 3 to 7 (Fig. (Fig.7B,7B, left panel). During this period, the yield of Rhsub01 was reduced by no more than a factor of 3 compared to the wild-type virus. However, during weeks 2 to 6, the yield of the two viruses diverged markedly (Fig. (Fig.7B,7B, right panel). Rhsub01-infected cultures produced less infectious virus than during the first week, whereas wild-type virus-infected RRPE cells generated increasing yields. By week 6, the yield of Rhsub01 was reduced by a factor of 500 in comparison to the wild type. The production of DNA-containing particles followed a similar pattern (Fig. (Fig.7D).7D). Rhsub01 generated nearly wild-type levels of particles for the first week and then showed no further increase. Calculation of the particle to TCID50 ratio revealed that the mutant produced particles with nearly normal infectivity during the first week of infection, but less infectious particles accumulated from 2 to 6 weeks after infection of RRPE cells (Fig. (Fig.7F7F).
We conclude that Rhsub01 generates a nearly normal yield of extracellular infectious progeny when it successfully enters RRPE cells (achieved by combining a high input multiplicity and centrifugal enhancement), but the mutant fails to initiate secondary infections and spread.
We generated a library of mutants in pRhCMV/BAC-Cre by using random transposon insertion and linear recombination (24, 40). In total, the library covers 75 RhCMV ORFs that correspond to annotated HCMV genes and five large, nonoverlapped genes that are unique to RhCMV (Table (Table1).1). With our current understanding of the coding capacity of the RhCMV genome, this represents ca. 50% of the known HCMV orthologues in RhCMV (17, 31).
Twenty-seven mutants carry lesions within RhCMV ORFs that are essential for replication in fibroblasts, and, with the exception of UL74 and UL97, the HCMV orthologues are essential for HCMV replication in at least one strain (11, 40). As noted above, UL74 and UL97 have been defined as strongly augmenting ORFs, i.e., viruses lacking these functions exhibit severe growth defects (18, 29). Consequently, it is not surprising that they have scored as essential in our RhCMV screen. Four mutations reside in ORFs that augment RhCMV replication, and 49 identify ORFs that are nonessential for replication in cultured telo-RFs. The augmenting and nonessential RhCMV ORFs exhibit growth phenotypes analogous to those described for the corresponding HCMV ORFs (11, 40), further validating RhCMV as a model for HCMV.
We used primary RRPE cells (5) to screen for RhCMV genes that are nonessential for replication in fibroblasts but required for replication in epithelial cells. RhCMV infects these cells inefficiently, exhibiting a block prior to immediate-early gene expression (Fig. (Fig.1).1). This phenotype might result from the absence of a UL128 orthologue in pRhCMV/BAC-Cre and its parental strain 68-1 (17). pUL128 is essential for the efficient infection of epithelial cells by HCMV (37, 38). An orthologue of this ORF is present in the 180.92 strain of RhCMV (31). In spite of the inefficient infection, the BAC-derived virus used in the present study produces substantial yields of infectious virus over an extended period of time in RRPE cells. Indeed, we have harvested virus weekly from infected RRPEs for a period of 20 weeks, with titers in the medium >105 TCID50/ml at each time point (data not shown).
Four of the 45 RhCMV ORFs that are nonessential for replication in cultured telo-RFs proved to be required for efficient replication in RRPE cells. Mutations in Rh01 (an HCMV TRL1 orthologue), Rh159 (UL148), Rh160 (UL132), and Rh203 (US22) caused moderate to severe growth defects in RRPE cells (Fig. (Fig.2).2). To the best of our knowledge, this is the first demonstration that these four genes influence the tissue tropism of a CMV. It is worth noting that mutations in the orthologues of UL130 and UL131A did not affect the replication of RhCMV in RRPE cells (data not shown). UL128, UL130, and UL131A in HCMV are believed to form a virion protein complex that is required for entry into epithelial and endothelial cells (1, 38). The fact that the corresponding RhCMV mutants showed no phenotype probably results from the fact that the complex is already disrupted in RhCMV BAC-derived virus due to the missing UL128 ORF (17, 31).
The defect in Rhsub01 was characterized. At an input multiplicity of 3 TCID50/cell with centrifugal enhancement, the mutant virus entered RRPE cells, expressed representative viral proteins and accumulated viral DNA as efficiently as wild-type virus (Fig. (Fig.44 to to7).7). Further, the pRh01-deficient mutant produced an amount of infectious extracellular virus similar to that of the wild-type virus during the first week after infection of RRPE cultures (Fig. (Fig.7B,7B, left panel). Importantly, however, the mutant virus failed to produce increasing amounts of progeny during weeks 2 to 6 after infection as was the case for the wild-type virus (Fig. (Fig.7B,7B, right panel). Further, although mutant and wild-type virus particles produced during the first week after infection were equally infectious (Fig. (Fig.7F,7F, left panel), mutant particles produced during weeks 2 to 6 were less infectious (Fig. (Fig.7F,7F, right panel). This might reflect the production of defective particles in cells that have been infected by the mutant for an extended period of time.
The failure of the mutant to produce increasing amounts of infectious progeny following the first week after infection of RRPE cultures (Fig. (Fig.7B)7B) is consistent with the observation that the number of pRh156-expressing cells increases to a very limited extent over time after infection with Rhsub01 (Fig. (Fig.6A),6A), and it leads to the conclusion that the mutant fails to spread after producing an initial burst of virus from the first set of cells infected in RRPE cultures.
It is clearly possible to infect RRPE cells with wild-type and mutant viruses at the same efficiency by using virus stocks produced in telo-RFs at a multiplicity of 3 TCID50/cell with centrifugal enhancement (Fig. (Fig.44 to to7).7). It also is evident that extracellular, infectious mutant virus is produced after the initial infection (Fig. (Fig.7B,7B, left panel), but this virus fails to initiate secondary infections (Fig. (Fig.6A6A and and7B,7B, right panel). Perhaps Rhsub01 is inherently deficient for entry into RRPE cells, and a combination of high input multiplicity and centrifugal enhancement compensates for the deficiency. Low-speed centrifugation at the time of infection has been shown to enhance infection by a variety of different viruses, including CMV (19), but the mechanism remains obscure. Centrifugal enhancement might alter the physiology of RRPE cells to bypass the need for pRh01 during the entry process. Alternatively, the composition of Rhsub01 virions might be different when produced in telo-RFs compared to RRPE cells. A change in the mutant virion might then preclude subsequent infection of RRPE cells, even though these virions can infect telo-RFs in assays for the production of infectious progeny. It is noteworthy that pRh01 accumulates in the cytoplasm (Fig. (Fig.3D),3D), where it could potentially influence tegumentation and the final assembly of virions, but it does not appear to be a constituent of virus particles. We were unable to detect pRh01-FLAG in virions by immunoblot assay (data not shown), and its HCMV orthologue, pTRL1, has not been found in HCMV virions (35).
If it is not a constituent of virions, how could pRh01 influence the infectivity of virus particles produced in RRPE cells? In polarized epithelial cells, there are distinct sorting pathways for cargo destined for the apical and basolateral membranes (12). Although the RRPE cells used in the present study were not polarized, it is tempting to speculate that their sorting pathways present are nevertheless distinct from those in fibroblasts. pRh01 may facilitate the transport of viral proteins and possibly virions to the proper compartments in RRPE cells. It is possible that pRh01 serves a similar role in telo-RFs. RRPE cells may simply be very stringent filters that, since they are infected at relatively low effective multiplicities of infection, magnify the effect of minute, possibly multiplicity-dependent replication defects that are inconsequential in the fully permissive telo-RFs. Finally, it is conceivable that pRh01 plays no role in virion assembly but regulates the expression or secretion of viral or cellular proteins that normally modulate the innate immune response of neighboring RRPE cells to facilitate spread of the virus.
Thus far, it has not been possible to investigate the composition of Rhsub01 virions produced in RRPE cultures due to the very low yields of particles. We have, however, noted a difference in wild-type versus Rhsub01 virus stocks produced in telo-RFs. When cells are infected with the mutant, they contain much more pRh112 at 1.5 h after infection than when they are infected with wild-type virus (Fig. 5A and B). pRh112 accumulates with late kinetics (Fig. 5A and B), and its HCMV orthologue, UL83-coded pp65, is an abundant virion protein. Thus, the pRh112 detected at 1.5 h after infection is almost certainly delivered to the cells by infecting particles. Both virus stocks were used at a multiplicity of 3 TCID50/cell, and the TCID50/genome ratios of wild-type and mutant virions are similar, so the same number of genome-containing virions were delivered in the mutant and wild-type virus infections. Consequently, the greater amount of pRh112 suggests that mutant genome-containing virions contain more pRh112 than wild-type virions or that the mutant stock contains more noninfectious particles lacking genomes, i.e., so-called noninfectious enveloped particles or dense bodies (26), than does the wild-type virus stock. Alternatively, pRh112 in mutant virions could be modified so that it is more stable after infection, or mutant virions could disassemble more slowly (although there is no apparent delay in pRh156 expression), and this could stabilize the virion protein. No matter what the cause of the differential level of pRh112 at 1.5 h after infection of telo-RFs, it does not lead to an observable difference in the kinetics of mutant compared to wild-type virus replication in this cell type (Fig. 2A and E).
In conclusion, we have generated a library of RhCMV mutants and screened it to identify four viral genes that influence replication in epithelial cells but not fibroblasts. These four genes have the potential to markedly influence the spread and pathogenesis of the virus within its primate host.
We are grateful to J.-P. Cong and P. Robinson for generating monoclonal RhCMV antibodies, to D. Yu (Washington University) for reagents and helpful discussion, and to S. Wong (Oregon Health Science University) for sharing RhCMV sequence data prior to publication.
A.E.L. was a postdoctoral fellow of the New Jersey Commission on Cancer Research. The isolation and characterization of RRPE cells was supported in part by the Intramural Research Program of the NEI, National Institutes of Health (S.P.B.), and the virology in this report was supported by NIH grants AI-54430 and CA-082396 to T.E.S.
Published ahead of print on 19 December 2007.