Efficient replication of rhesus cytomegalovirus variants in multiple rhesus and human cell types

UL128-UL131 Locus Promotes Efficient RhCMV Replication in Rhesus Epithelial Cells.

Virus recovered from the BAC clone of RhCMV strain 68-1, RhCMV-loxP(r) (24), infects rhesus epithelial cells with low efficiency and spreads slowly, but produces large amounts of extracellular virus over extended periods of time (10). Sequence analyses of the RhCMV strain 180.92 and other RhCMV isolates have shown that BAC-derived 68-1 and its parent, 68-1, are missing UL128 and the second exon of UL130 (8, 9, 23). However, strain 180.92 has a large deletion in the ULb′ region of the genome, and it has not been cloned as a BAC (9). To analyze the importance of the RhCMV UL128-UL131 locus for replication in epithelial cells, we repaired the locus in the 68-1 BAC clone using the genome of strain 180.92 as template. The resulting BAC, pBRh68-1.1, has a wild-type UL128-UL131 locus, albeit in the reverse orientation compared with strain 180.92 and other RhCMV isolates (9, 23). The sequence junction downstream of UL128 in the repaired genome is artificial, because the genomic sequences are different in the 68-1 and 180.92 viruses (8, 9). The DNA sequence of the repaired locus is identical to that in 180.92 with the exception of one missing A residue at the repair junction inside the UL130 intron. The RhCMV genomic sequence of pBRh68-1.1 can be generated by replacing nucleotides (nt) 167,056–167,079 from strain 68-1 (Accession No: AY186194 (8)) with the reverse complement of nt 162,629–164,443 from strain 180.92 (Accession No: DQ120516 (9)).

The second exon of the UL36 orthologue (Rh60 (8)) in strain 68-1 carries a 1-bp insertion that causes a premature truncation of the translated protein (25). We repaired this mutation in pBRh68-1.1, and named the BAC clone with the repaired UL36 and UL128-UL130 loci pBRh68-1.2.

As noted above, the HCMV UL128-UL131 locus is required for efficient viral entry into several cell types other than fibroblasts, including epithelial cells (13–20). To determine whether the repaired RhCMV isolates infect epithelial cells more efficiently than the parental viruses, we monitored the expression of the RhCMV IE1 orthologue (Rh156 (8)). Stocks of parental and repaired viruses were used to infect rhesus epithelial cells (26) or fibroblasts (27) at a multiplicity that generated 60–70% IE1-positive fibroblasts 24 h later. BAC-derived 68-1 and parental 68-1 virus infected epithelial cells at 0.2–0.4% of the frequency with which they infected fibroblasts (Fig. 1A), consistent with our previous observations (10). Strain 180.92 also infected epithelial cells at a reduced efficiency compared with fibroblasts (≈6%), presumably because it lacks a portion of the ULb′ region (8, 9, 23). The infectivity of 180.92 was not significantly higher than the infectivity of the 68-1 derivatives. In marked contrast to the parental viruses, the repaired viruses, BRh68-1.1 and BRh68-1.2, infected rhesus epithelial cells at 32 and 48%, respectively, of the efficiency with which they infected fibroblasts (Fig. 1A), which is significantly more efficient than the other viruses tested (P < 0.001 in all cases). The repaired viruses infected 100% of epithelial cells when used at a higher input multiplicity (data not shown).

Next, we performed growth assays to see whether the improved infectivity of BRh68-1.1 and BRh68-1.2 for rhesus epithelial cells translated to faster spread and accumulation of extracellular virus. The parental and repaired viruses produced >10⁷ TCID₅₀ of extracellular virus per culture at 12 days after infection of rhesus fibroblasts (Fig. 1B Left). BAC-cloned 68-1 virus grows similarly to parental 68-1 in fibroblasts (24). As anticipated, BAC-cloned 68-1 and its parent replicated slowly in rhesus epithelial cells, releasing ≈10³ TCID₅₀ per culture after 12 days (Fig. 1B Right). In contrast, 180.92, BRh68-1.1 and BRh68-1.2 replicated more efficiently in epithelial cells, generating yields of 10⁶-10⁷ TCID₅₀ per culture after 12 days (Fig. 1B Right), comparable to their yield in fibroblasts.

BRh68-1.1, with a repaired UL128–130 locus, infects and replicates more efficiently in epithelial cells than strain 68-1 viruses. BRh68-1.2, in which the UL36 ORF also was repaired, appeared to replicate marginally better than BRh68-1.1 in epithelial cells (Fig. 1B Right), but the possible improvement was not statistically significant. To further investigate a role for pUL36 in the efficient replication of RhCMV, we constructed a derivative of BRh68-1.2 lacking the entire sequence encoding UL36x2 (BRh68-1.2-subUL36x2), and compared its replication to BRh68-1.2 and BRh68-1.1. BRh68-1.2 produced 2-fold more extracellular virus than BRh68-1.1 (P = 0.07) and 4-fold more than BRh68-1.2-subUL36x2 (P = 0.02) in rhesus fibroblasts at 12 days after infection (Fig. 1C Left), but the improvement was of marginal significance. The difference was more pronounced in rhesus epithelial cells, where BRh68-1.2 yielded 8-fold higher titers than BRh68-1.1 and 13-fold higher titers than BRh68-1.2-subUL36x2 (Fig. 1C Right). The difference in yield was most significant when pUL36-deficient BRh68-1.2-subUL36x2 was compared with pUL36-expressing BRh68-1.2 in epithelial cells at 8 days after infection (P < 0.0001). Although we did not consistently observe a significant effect of pUL36 (compare Fig. 1 B and C), it is possible that the gene product modestly enhances RhCMV replication in epithelial cells.

Expression of UL128-UL131 and UL36 in BRh68-1.2-Infected Cells.

We mapped the transcripts expressed by the repaired UL128-UL131 locus in BRh68-1.2. We amplified 5′ and 3′ RACE PCR products with overlapping sequences. The two reactions each produced only one product, suggesting that the BRh68-1.2 UL128-UL131 locus is carried on a single transcript (Fig. 2A). This is consistent with the absence of predicted poly(A)-signals within the locus. HCMV encodes a similar RNA, but it also produces a transcript that codes for only UL128 (14, 16). We did not detect an additional start site using 5′ RACE primers that should have detected a UL128-specific species, although it is possible that the analysis failed to detect a low abundance RNA.

The sequences of cloned UL128-UL131 RACE products precisely identified the predicted (9) splice junctions within UL128 and UL131 (Fig. 2A). As for HCMV (16), RhCMV UL128 has 2 introns (nucleotides 167,998–168,085 and 168,224–168,309) and RhCMV UL131 has 1 intron (nucleotides 166,383–166,464). In contrast to the unspliced human transcript, an intron was present in UL130 as recently reported (23). The 5′ end of the UL128-UL131 transcript mapped to nt 166,116, 10 nt upstream of the predicted start AUG of UL131 and 32 nt downstream of a predicted TATA-box. The 3′ end of the transcript mapped to nt 168,780, 16 nt downstream of the UL128 stop codon and the predicted poly(A)-signal.

We further characterized the pUL130 coding region in BRh68-1.2 by generating a variant, BRh68-1.2-inUL130F, with a FLAG epitope tag fused to the C-terminal end of UL130x2. This virus grows like BRh68-1.2 in rhesus fibroblasts (Fig. 2B Left), but displays delayed growth kinetics in epithelial cells (Fig. 2B Right), consistent with a deleterious effect of the tag on pUL130 function. Immunofluorescent analysis of rhesus fibroblasts and epithelial cells at 72 h after infection with BRh68-1.2-inUL130F showed that pUL130FLAG localizes to the cytoplasm of infected cells (Fig. 2C). A Western blot of lysates from fibroblasts infected with the tagged virus showed that pUL130FLAG accumulated to a detectable level between 24 and 48 h after infection (Fig. 2D), consistent with the late expression kinetics of the corresponding HCMV transcript (14). Because of the growth defect of the tagged virus (Fig. 2B Left), there was a delay in pUL130FLAG accumulation in epithelial cells, but a specific product was detected at 96 h after infection (Fig. 2D). The apparent size of pUL130FLAG is larger (≈40 kDa) than the predicted size (37 kDa) of the fusion protein, presumably because of glycosylation, as documented for HCMV pUL130 (17).

We also mapped the repaired UL36 transcript in BRh68-1.2 (Fig. 3A). 5′ RACE analysis located its start to nt 49,532, 76 nt upstream of the predicted start of UL36x1 and 31 nt downstream of a putative TATA box. 3′ RACE analysis located its end at nt 47,915, 73 nt downstream of the UL36 stop codon and 17 nt downstream of a poly(A)-signal. Sequence analysis of the RACE products verified the repaired sequence and mapped the UL36 intron to nt 49,179–49,254.

To confirm that the repaired UL36 ORF encoded a full-length protein, we fused a FLAG tag to the C-terminal end of the ORF in BRh68-1.2, generating BRh68-1.2-inUL36F. Replication in fibroblasts was not affected by the tag (Fig. 3B), and immunofluorescent analysis showed that pUL36FLAG is localized at or near the cytoplasmic viral assembly zone (28) at 48 h after infection (Fig. 3C). Immunoblot analysis of pUL36FLAG accumulation identified a single protein that was first detected 24 h after infection (Fig. 3D), consistent with the known delayed-early expression of its mRNA (25). pUL36FLAG migrates at ≈48 kDa, which is smaller than predicted for RhCMV UL36 (53 kDa). Sequence analysis of RACE products demonstrated that the mRNA contains the expected sequence, so we anticipate that the size discrepancy results from the use of an internal AUG, posttranslational processing or aberrant electrophoretic migration.

BRh68-1.1 and BRh68-1.2 Replicate Efficiently in Human Epithelial and Endothelial Cells.

RhCMV was cultured in human fibroblasts after its initial isolation (29), but we are not aware of reports that it can also replicate in other human cell types. We were not able to grow BAC-cloned 68-1 virus in human epithelial and endothelial cells, but this was expected, given the role of UL128-UL131 in infection of rhesus epithelial cells (Fig. 1). We next infected rhesus fibroblasts, human lung fibroblasts (MRC5), and human retinal pigment epithelial cells (ARPE-19) with the parental or repaired viruses at a multiplicity that resulted in 60–70% IE-1-positive rhesus fibroblasts, and compared the percentage of IE-1 positive cells for each virus (Fig. 4A). The 68-1 viruses infected human epithelial cells poorly, at 0.6 and 0.2% of the efficiency with which they infected rhesus fibroblasts, whereas the repaired BRh68-1.1 and BRh68-1.2 viruses infected human epithelial cells at 19 and 35% of the efficiency with which they infect rhesus fibroblasts. As for rhesus epithelial cells, the 180.92 strain exhibited an intermediate level of infectivity for human epithelial cells, at 4% of the efficiency with which it infects rhesus fibroblasts. All RhCMV variants infected rhesus and human fibroblasts with similar efficiency, showing that the rhesus UL128-UL131 locus is specifically required for infection of human epithelial cells.

We also compared the growth kinetics of the different RhCMV strains in several human cell types. All viruses tested generated robust yields in human fibroblasts (Fig. 4B), although strain 180.92 produced an ≈50-fold reduced yield compared with the other viruses. Strains 180.92, BRh68-1.1, and BRh68-1.2 also produced substantial yields in human epithelial cells (Fig. 4C). BRh68-1.2 grew most efficiently, producing a yield of 8 × 10⁷ TCID₅₀ per culture at 12 days after infection, somewhat higher than its yield in human fibroblasts and ≈10-fold higher than its yield in rhesus epithelial cells (compare Fig. 1B Right with Fig. 4C). As expected, the 68-1 viruses did not produce progeny in these cells (Fig. 4C), even though they replicated in human fibroblasts (Fig. 4B). We expanded our analysis to primary human endothelial cells (HUVEC) (Fig. 4D). BRh68-1.1 and BRh68-1.2 produced 2–3 × 10⁶ TCID₅₀ at 12 days after infection. RhCMV 180.92, which lacks a portion of the ULb′ region, replicated much more poorly in human endothelial than epithelial cells compared with BRh68-1.1 and BRh68-1.2, yielding ≈10³ TCID₅₀ at 12 days after infection. Not surprisingly, the 68-1 viruses replicated poorly (≈10³ TCID₅₀ for BAC-derived 68-1) or not at all in human endothelial cells. We also assayed growth in human papilloma virus E6/E7-immortalized endothelial cells (iHUVEC) (30) and, again, BRh68-1.1 and BRh68-1.2 generated much higher yields than the other viruses (Fig. 4E).

Although BRh68-1.1 and BRh68-1.2 replicated efficiently in a variety of human cells, none of the variants generated a significant yield after infection of mouse embryo fibroblasts (data not shown).

In summary, RhCMV variants that contain an intact UL128-UL131 region produce robust yields in human epithelial or endothelial cells.

Cells and Viruses.

Telomerase life-extended, rhesus fibroblasts (telo-RFs) (27) were cultured in DMEM with 8% FBS; human MRC5 lung fibroblasts (45) were grown in DMEM with 8% FBS and 1 mM sodium pyruvate; rhesus retinal pigment epithelial (RRPE) cells (26) and human ARPE-19 retinal pigment epithelial cells (46) were propagated in a 1:1 combination of DMEM and Ham's F12 nutrient mixture with 5% FBS, 1 mM sodium pyruvate, and nonessential amino acids; and primary human umbilical vein endothelial cells (HUVECs) and E6/E7-immortalized HUVECs (30) were maintained in EGM-2 MV (Lonza Walkersville).

The RhCMV 68-1 strain (29), BAC-derived RhCMV 68-1, i.e., RhCMV-loxP(r) (24), and RhCMV strain 180.92 (9) were propagated in telo-RFs. pBRh68-1.1 was generated by BAC “recombineering” (47, 48). Briefly, a galK cassette was amplified by PCR using 5′-GTTTTGTATCATAACGTATAATATTCATAATACAGATTATACCAATACTTTCCTGTTGACAATTAATCATCGGCA-3′ and 5′-GCCATTTTCGAAAACTTTTGGAACACGGGCGAGTGTAGTGCACTACAATGATCAGCACTGTCCTGCTCCTT-3′ primers. The product was electroporated into E. coli SW105 containing pRhCMV/BAC-Cre (24), galK-expressing clones were selected and electroporated with a UL128-UL130 DNA fragment generated by PCR amplification of strain 180.92 genomic viral DNA, using 5′-GTTTTGTATCATAACGTATAATATTCATAATACAGATTATACCAATACTTTTGGG-3′ and 5′-GCCATTTTCGAAAACTTTTGGAACACGGGCGAGTGTAGTGCACTACAATGA GCCGATTTTACCCGAGTCTCAAACC-3′ primers, and the desired variants were obtained by galK counterselection. pBRh68-1.2 was generated from pBRh68-1.1, using a similar approach. The galK cassette was PCR amplified using 5′-ATTGTACCGAACTTGTCAACTAGTACATAGAGTCTGACTAGGAACTCATTT CCTGTTGACAATTAATCATCGGCA-3′ and 5′-TAGTGTGTGTTAATCAATTGCTCGGGGTGCTAGGTTGCTTCCGTAAAGAAA TCAGCACTGTCCTGCTCCTT-3′ primers and the UL36 mutation was corrected using an oligonucleotide comprised of 5′-AACTTGTCAACTAGTACATAGAGTCTGACTAGGAACTCATTTTTTTCTTTACGGAAGCAACCTAGCACCCCGAGCAATTGATTAA-3′ annealed to its complementary strand. The pBRh68-1.2-inUL36F and pBRh68-1.2-inUL130F BAC clones were made by using linear recombination and the arabinose-inducible Flp recombinase in SW105 bacteria (47), as described in ref. 10. Clones were analyzed by restriction enzyme digestion to monitor integrity of the BAC and sequence analysis to confirm the desired modification. Virus stocks were made by electroporating BAC DNAs into telo-RFs. Extracellular virus was harvested from the transfected cells 1 day after all cells showed cytopathic effect (CPE) and expanded once by low multiplicity infection of telo-RFs. The infectivity of virus stocks was determined by TCID₅₀ assays on telo-RFs.

Virus replication assays were performed by infecting cells at a multiplicity of 0.01 TCID₅₀ per cell and collecting the supernatant after various time intervals. Growth curves report the mean yield of extracellular virus from 2 infections, each of which was titered in duplicate by TCID₅₀ assay on telo-RFs. Virus infection, i.e., entry, was quantified after infection at a multiplicity that resulted in IE1 (Rh156) expression in 60–70% of telo-RFs at 24 h after infection, using an IE1-specific monoclonal antibody (10) and a fluorescently labeled secondary antibody. Infectivity was calculated from the mean fraction of fluorescent nuclei in 5 randomly chosen fields.

RNA and Protein Analysis.

To map UL128-UL131 and UL36 transcripts, telo-RFs were infected with BRh68-1.2 at a multiplicity of 1 TCID₅₀ per cell. Total cell RNA was harvested at 24, 48, and 72 h after infection, using TRIzol (Invitrogen). The three samples were pooled, treated with DNA-free (Ambion), and used as starting material for RACE analysis, using a FirstChoice RLM-RACE kit (Ambion). The nested 5′ RACE analysis of the UL128-UL131 transcript used 5′-CCCAAGTGATGACACGCATCTG-3′ (ogs5 in Fig. 2A) and 5′-GTCGAGGAGATGGATACATGGGAC-3′ (igs5 in Fig. 2A) as the UL128-UL131-specific outer and inner primers, respectively. Similarly, 5′-CTCTAGCACGACCCATTCCGC-3′ (ogs5 in Fig. 3A) and 5′-GGATCACTGGGTTCTTCACCG-3′ (igs5 in Fig. 3A) were used as the UL36-specific outer and inner primers. The UL128-UL131-specific 3′ RACE primer was 5′-CCCCATCTTACTCGTAGGATTGG-3′ (gs3 in Fig. 2A) and the UL36-specific 3′ RACE primer was 5′-GTTGCGAAGAACGGTTACCGC-3′ (gs3 in Fig. 3A). The PCR amplified RACE products were cloned in pGEM-T Easy (Promega) and sequenced.

Accumulation and intracellular localization of FLAG epitope-tagged proteins was determined by immunoblot and immunofluorescence assays (10). Cells were infected at a multiplicity of 3 TCID₅₀/cell (immunoblots) or 0.5 TCID₅₀/cell (immunofluorescence), and the primary murine monoclonal antibody for both assays was anti-FLAG M2 (Sigma-Aldrich).