Rhesus Cytomegalovirus Particles Prevent Activation of Interferon Regulatory Factor 3^†

Cells, antibodies, and virus.

HCMV strain AD169 and RhCMV strain 68-1 were used in all experiments. Virus particles were purified by centrifugation through a 20% sorbitol cushion, resuspended in Dulbecco modified Eagle medium (DMEM), and stored at −80°C. Telomerized rhesus fibroblasts (TRFs) were cultured in DMEM supplemented with 10% fetal bovine serum (FBS), 1% glutamine, and 1% penicillin-streptomycin. Cells were grown to confluency in 35-mm plates and infected in 300 μl of serum-free DMEM at a multiplicity of infection of 3 PFU/cell. Infection in the presence of CHX and UV inactivation of virus were performed as described previously (8). Two separate antibodies to IRF3 were used. Rabbit anti-IRF3 used for immunofluorescence assays (IFAs) was a generous gift from Takashi Fujita (Tokyo Metropolitan Institute of Medical Science) (42). Rabbit anti-IRF3 used for Western blotting was obtained from Santa Cruz Biotechnology (catalog no. SC-9082). Rabbit anti-NF-κB used for both Western blotting and IFA was obtained from Santa Cruz Biotechnology (SC-372). Antibody to RhCMV IE1 was raised in the Vaccine and Gene Therapy Institute antibody core. Rabbit and mouse antibody to HCMV IE1 and rabbit anti-HCMV gB were obtained from Jay Nelson (Oregon Health and Science University).

Array hybridization.

At 8 and 24 h postinfection (hpi), inoculum was removed, cells were washed with phosphate-buffered saline (PBS), and total RNA isolated by using the Zymo Mini-RNA isolation kit according to the manufacturer's instructions (Zymo Research, Orange, CA). Gene expression was examined by using hybridization arrays displaying 8053 human cDNA clones manufactured by the OHSU Gene Microarray Shared Resource. Human genes were selected from a library of sequence verified IMAGE consortium clones (Invitrogen catalog no. 97001.V). Arrays were printed on aminosilane coated glass slides by using a Cartesian PixSys 5500 XL microarray printer (Genomic Solutions), and each cDNA was spotted twice to account for intrachip variation. Homology between rhesus macaque and human genes is high enough to allow cross species array hybridization to be an accurate technique for examining rhesus samples (9). mRNA was amplified to make cRNA from 1 μg of total RNA prior to probe synthesis according to the method of Phillips and Eberwine (32). Probes were synthesized from 1 μg of cRNA by randomly primed reverse transcription (RT) in the presence of biotin or fluorescein-conjugated dCTP. Probe-target hybridization of infected and mock-infected and cyanine-3 (Cy3) and cyanine-5 (Cy5) fluorophore deposition was performed by using the Micromax TSA labeling and detection kit according to the manufacturer's instructions (Perkin-Elmer, Wellesley, MA). Probe labeling and fluorophore deposition were reversed in each experiment to control for dye bias resulting in two arrays for each biological replicate. Arrays were scanned by using a confocal laser scanner (Perkin-Elmer) and intensities quantified by using ImaGene 5.0 (Biodiscovery).

Data analysis.

Subtraction of background signal was performed by using ImaGene 3.5.1. Intensity data were Loess normalized (smoothing parameter, 0.5; degree of fitness, 2) by using GeneSight (Biodiscovery) (41). Signals obtained from the same clone on an array were used to calculate the mean signal intensity. Differences in gene transcription between infected and uninfected samples were examined by using the ratio of signal intensity between fluorophores. Expression differences between infected and mock-infected samples were calculated by using the ratio of normalized intensities, and statistical significance was determined as described in reference 21. Genes were considered differentially regulated if they exhibited significant (P > 0.95) changes in the same direction for both Cy3/Cy5 orientations and both biological replicates and if the fold change was >1.5.

Quantitative real-time RT-PCR.

Quantitation of individual gene expression was performed by using quantitative real-time RT-PCR (qPCR) with the Applied Biosystems Sequence Detection System (Foster City, CA). This involves intercalation of SYBR green fluorescent molecules into DNA synthesized during PCR. Measurement of SYBR green occurs throughout the reaction and the cycle number at which fluorescence levels reach an arbitrary threshold during the exponential phase of amplification (cycle threshold) is recorded for each sample. According to the method of Livak and Schmittgen (27), this value is normalized within each sample to that of a housekeeping gene (GAPDH [glyceraldehyde-3-phosphate dehydrogenase]) and subsequently compared between mock-infected and infected samples which allows estimation of relative gene expression levels (i.e., fold change). Quantitation of expression of three ISGs was performed in this way: Viperin (nucleotide [nt] 765 to 879; GenBank accession no. NM_080657), Mx2 (nt 1811 to 1924; NM_002463), and IFI-15K (nt 327 to 446; NM_005101). PCR products were examined on a 1% agarose gel to ensure that specific amplification occurred.

Immunofluorescence.

Indirect immunofluorescence assay (IFA) was used to examine the subcellular localization of IRF3 and NF-κB in TRFs, as well as to monitor viral replication through examination of viral proteins. TRFs were grown on coverslips and infected or mock infected at subconfluency as described above. At room temperature, cells were washed twice with PBS, fixed for 30 min in 3.7% formaldehyde, washed, and quenched for 10 min with 50 mM NH₄Cl. Cells were then permeabilized with 0.1% Triton X-100 for 7 min and washed three times with PBS containing 2% bovine serum albumin (BSA). Cells were incubated with primary antibody (diluted 1:500 in PBS containing 2% BSA) at 37°C for 60 min, washed three times in PBS containing 2% BSA (10 min each wash), and incubated with fluorescence-labeled secondary antibody (goat anti-mouse 488 and/or goat anti-rabbit 594; Molecular Probes) diluted 1:1,000 in PBS containing 2% BSA for 60 min. Cells were washed twice in PBS containing 2% BSA (10 min each) and once in PBS. Coverslips were mounted on a slide with Vectashield mounting medium (Vector Laboratories, Burlingame, CA) containing DAPI (4′,6′-diamidino-2-phenylindole).

Western blotting.

Examination of IRF3 dimerization by native polyacrylamide gel electrophoresis and Western blotting was performed as described in reference 17. TRFs infected and mock infected in 100-mm dishes were treated with trypsin, collected by using centrifugation, and resuspended in lysis buffer (50 mM Tris-Cl [pH 8.0], 1% Ipegal, 150 mM NaCl) supplemented with 1× protease inhibitor cocktail (Invitrogen) and phosphatase inhibitor cocktail (Sigma). Whole-cell lysate protein was quantified by using a BCA assay (Pierce), and equal amounts were added to loading buffer (125 mM Tris-Cl [pH 6.8], 30% glycerol, 0.1% bromophenol blue). Samples were subjected to 7.5% non-sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to ImmunoBlot polyvinylidene difluoride membranes (Bio-Rad Laboratories). After transfer, membranes were incubated in a blocking buffer (PBS containing 10% milk and 0.1% Tween 20), followed by incubation of the primary antibody diluted in PBS containing 5% milk and 0.1% Tween 20. Blots were washed with PBS containing 0.1% Tween 20 and incubated with a horseradish peroxidase-conjugated secondary antibody (Amersham Biosciences, Piscataway, N.J.) diluted in PBS containing 5% milk and 0.1% Tween 20. Blots were washed and developed by using the PicoPure chemiluminescence reagent (Pierce) according to the manufacturer′s protocol.

Rhesus fibroblasts display an innate immune response to HCMV but not to RhCMV.

Expression profiling by DNA microarrays revealed previously that HCMV infection of human fibroblasts triggered massive changes in the transcriptional program of the infected cells mostly attributable to innate immune response genes such as ISGs and proinflammatory genes (1, 8, 37, 45). To determine whether infection with RhCMV would trigger a similar response, we compared the transcriptional response of rhesus fibroblasts (TRFs), telomerized as described previously by Bresnahan et al. (6), after infection with RhCMV or HCMV. To verify that HCMV was able to infect these nonhuman primate cells, we examined the expression of the immediate-early protein IE1 (UL123), as well as the late protein glycoprotein B (gB; UL55), at 24 hpi by using IFA. Figure 1 illustrates the expression in TRFs of RhCMV immediate-early open reading frame (ORF) Rh156 and HCMV proteins encoded by ORFs UL123 and UL55. Although the viral yield of HCMV from TRFs is lower than from human fibroblasts (not shown), our data are consistent with normal gene expression by HCMV in TRFs. To examine changes in host transcription, TRFs were infected with RhCMV or HCMV (MOI = 3) or mock infected, and total RNA was isolated at 8 and 24 hpi. After mRNA amplification, host cell gene expression was examined by using cDNA microarrays as described in Materials and Methods. Genes were considered differentially regulated between infected and mock-infected samples if statistically significant, similar direction fold changes of >1.5 were observed for both biological replicates and both dye label orientations.

At 8 hpi HCMV infection induced transcriptional changes in 212 genes relative to mock-infected cells (150 upregulated, 62 downregulated), and RhCMV induced changes in 32 genes (12 upregulated, 20 downregulated) (see Tables S1 and S2 in the supplemental material). At 24 hpi HCMV induced changes in 720 genes (436 upregulated, 284 downregulated) and RhCMV induced changes in 167 genes (85 upregulated, 82 downregulated). Figure 2A graphically illustrates the differences in quantity and the fold changes of differentially expressed genes between the two viruses. Clearly, the changes in mRNA levels induced by HCMV infection are by far more numerous and of greater magnitude than those induced by RhCMV. Fold changes relative to mock infection were greater overall in HCMV-infected cells compared to RhCMV-infected cells. Figures 2B and C illustrate the number and average degrees of change for significantly regulated host genes across experiments. Fold changes of differentially regulated genes in RhCMV-infected cells at 24 hpi averaged 3.3 (standard deviation [SD] = 1.2) for upregulated genes and −4.5 (SD = 7.3) for downregulated genes, whereas upregulated genes at 24 hpi in HCMV-infected cells averaged 23.7 (SD = 120.8) for upregulated genes and −5.8 (SD = 7.1) for downregulated genes. Thus, infection of rhesus fibroblasts by RhCMV triggered a remarkably low host transcriptional response.

At 8 hpi, 7 genes and at 24 hpi 131 genes were differentially expressed in both HCMV- and RhCMV-infected cells (see Table S1 in the supplemental material). The majority of these genes are involved in general cellular metabolic, physiological processes or in cell-cell communication. In contrast, many innate immune response genes were induced by HCMV, but not RhCMV, including ISGs and proinflammatory cytokines and chemokines, as well as genes involved in prostaglandin synthesis. The induction of these genes is consistent with previous observations for human cells, suggesting that rhesus fibroblasts reacted very similarly to infection with HCMV. These genes were not induced, only minimally induced, or even repressed in TRFs infected with RhCMV. The dramatic response to HCMV infection ruled out the possibility that this low transcriptional response to RhCMV was caused by a general defect of innate immune responses in TRFs. Thus, we conclude that, in contrast to HCMV, RhCMV does not trigger an innate response to viral infection in vitro.

To independently confirm both expression levels measured by using microarray hybridization and the general disparity in ISG expression, we used qPCR to examine mRNA levels of three ISGs (viperin, IFI-15K, and Mx2) in two independent experiments. Gene-specific primers were selected based on the corresponding human sequence and successful amplification of a DNA band corresponding to the expected molecular size. In all cases, qPCR confirmed that ISGs were induced by HCMV but not by RhCMV (Fig. 3). Similar results were obtained after HCMV and RhCMV infection of TRFs in the presence of CHX at 8 hpi (Fig. 3), although magnitude of upregulation is distinctly lower presumably due to the absence of IFN secretion and subsequent JAK/STAT pathway stimulation. The apparent downregulation of ISGs in RhCMV-infected cells is the result of experimental variability in measurements obtained from very low mRNA copy numbers. The difference in ISG expression between RhCMV-infected and mock-infected cells is thus not significant. In contrast, the same transcripts were present in high copy numbers in HCMV-infected cells resulting in a significant difference to mock levels. Thus, we conclude that, unlike HCMV, RhCMV does not induce ISGs in infected fibroblasts.

Differential activation of IRF3, but not NF-κB, in HCMV-infected compared to RhCMV-infected cells.

The dramatic difference in gene expression profiles between HCMV- andRhCMV-infected cells, especially with respect to ISGs and proinflammatory genes, suggests that important step(s) in the activation of these genes are either absent or inhibited in RhCMV-infected cells. Many of the transcripts upregulated by HCMV but not RhCMV are known to be induced by the transcription factor NF-κB, IRF3, or both. Transcriptional activity of both factors upon HCMV entry has been demonstrated (4, 7, 22, 26, 42, 44). Therefore, we examined whether these pivotal transcription factors were differentially accumulated in fibroblast nuclei upon HCMV or RhCMV infection. Both transcription factors are constitutively present in fibroblasts.

We examined the subcellular localization of NF-κB subunit p65 and IRF3 in TRFs infected with equivalent MOIs of either RhCMV or HCMV at 0.5, 1, 2, 4, 6, 8, and 24 hpi. We observed that neither HCMV nor RhCMV induced widespread redistribution of NF-κB from predominantly cytoplasmic to predominantly nuclear localization at any time point compared to mock-infected cells (Fig. 4). In contrast, cells transfected with poly(I-C) demonstrated a rapid nuclear accumulation of NF-κB. The lack of NF-κB accumulation in HCMV-infected cells is consistent with previous immunofluorescence observations for HCMV infection of human fibroblasts (7) but differ from other studies that demonstrated NF-κB activation by HCMV (22, 43, 44). It should be noted, however, that although the vast majority of infected cells primarily exhibited cytoplasmic NF-κB, cells were occasionally observed that contained weak cytoplasmic and slightly greater nuclear NF-κB staining at 2 to 24 hpi, a condition not observed in uninfected cells. Thus, our data do not rule out a low level of NF-κB activation that might be measurable in reporter gene or gel shift assays used in most studies (22, 43, 44). However, since we did not observe a substantial difference between HCMV and RhCMV in their ability to activate NF-κB at any time point, it seems highly unlikely that differential activation of NF-κB is responsible for the observed differential gene expression in cells infected with HCMV versus RhCMV.

In contrast, a dramatic difference in the activation of IRF3 was observed between RhCMV- and HCMV-infected cells (Fig. 5). IRF3 accumulated in the nuclei of TRFs infected with HCMV beginning at 1 hpi and was maximal by 3 hpi, whereas diffuse subcellular localization of IRF3 was observed inRhCMV-infected cells similar to the cytoplasmic staining of IRF3 in mock-infected cells (Fig. 5). This differential activation of IRF3 correlated perfectly with the differential induction of ISGs and proinflammatory genes described above.

RhCMV does not induce nuclear accumulation of IRF3 in the absence of host or viral gene expression.

Recently, Collins et al. (10) showed that the activation of IRF3 is a default reaction of mammalian cells to infection by enveloped DNA viruses unless the virus encodes inhibitors of this innate immune response pathway. Thus, diverse DNA and RNA viruses induced the phosphorylation of IRF3 upon inactivation of viral gene expression by either CHX or UV radiation. Therefore, we hypothesized that RhCMV would activate IRF3 and induce ISG expression upon UV-inactivation or CHX treatment.

RhCMV and HCMV virus particles were UV inactivated, resulting in the absence of immediate-early gene expression in infected cells (Fig. 6). Similarly, infected cells treated with CHX did not demonstrate viral immediate-early gene expression (Fig. 7). As observed for the live virus, UV-inactivated HCMV (UV-HCMV) induced nuclear accumulation of IRF3 (Fig. 7). This was also observed upon treatment with CHX (Fig. 7). In contrast, neither UV inactivation of RhCMV nor RhCMV infection in the presence of CHX resulted in nuclear IRF3 accumulation (Fig. 6 and 7). Thus, RhCMV did not activate IRF3 even in the absence of viral or host gene expression.

IRF3 does not dimerize after RhCMV infection.

Nuclear accumulation of IRF3 first involves phosphorylation of C-terminal serine and threonine residues by IKKɛ or TBK1, followed by homodimerization and association with nuclear transcription factors CBP/p300. To determine whether IRF3 did not accumulate in the nucleus because it did not form homodimers due to lack of phosphorylation in RhCMV-infected cells, we studied the dimerization state of IRF3 in virus-infected cells. Infections were conducted in the presence of CHX to eliminate the effects of viral gene expression. As shown in Fig. 8, more slowly migrating dimerized IRF3 is detected in TRFs transfected with poly(I-C) or infected with HCMV. In contrast, the dimerization status of IRF3 in RhCMV-infected cells was similar to mock-infected cells. We therefore conclude that the lack of IRF3 nuclear accumulation results from the absence of IRF3 phosphorylation-mediated dimerization in RhCMV-infected cells.

RhCMV inhibits HCMV-induced but not dsRNA-induced activation of IRF3.

Lack of IRF3 nuclear localization after RhCMV infection in the absence of viral or host protein synthesis indicates that RhCMV either lacks the ability to trigger the protein's activation or blocks activation in a manner that does not require virus gene expression, for example, via components of the virion. However, given the observation that diverse enveloped virus families (including HCMV) activate IRF3 upon cellular entry (10), it seems much more likely that RhCMV blocks IRF3 activation by a molecular mechanism that does not require viral or host gene expression. To examine whether RhCMV can inhibit IRF3 activation by a nonviral stimulus, we monitored IRF3 nuclear accumulation as induced by poly(I-C) after infection with RhCMV. TRFs were transfected with poly(I-C) for 2 to 3 h and subsequently infected with RhCMV (MOI = 3) for two to three more hours. The reverse sequence [RhCMV infection, followed by poly(I-C) transfection] was also performed. Significant IRF3 nuclear accumulation was detected under both conditions (Fig. 9), indicating that RhCMV neither blocks the signal transduction pathway activated by poly(I-C) nor reverses IRF3 nuclear accumulation induced by poly(I-C).

Induction of IRF3 by poly(I-C) is known to take place via TLR3 (2). In contrast, HCMV induction of inflammatory cytokines via NF-κB has been shown to occur via TLR2 and CD14 (11). To determine whether RhCMV can inhibit IRF3 activation by HCMV, we examined IRF3 nuclear accumulation after treatment with both RhCMV and HCMV. Two experimental designs were used. (i) Cells were preinfected with HCMV (MOI = 3) for 3 h and then infected with RhCMV (MOI = 3) for 3 h, or (ii) cells were infected first with RhCMV followed by HCMV. Positive control cells were also treated with HCMV alone (MOI = 3) for 3 h.

Coinfection of individual cells with live virus was verified by using antibodies that were specific for HCMV or RhCMV IE1 (Fig. 10). Preinfection with HCMV or UV-HCMV decreased the strength of RhCMV IE1 expression (Fig. 10). TRF cultures preinfected with HCMV or UV-HCMV and subsequently infected with RhCMV also seemed to exhibit additional cytopathic effects and increased syncytium formation (Fig. 10). However, the lower expression of RhCMV IE1 is presumably due to the action of the HCMV-induced innate antiviral response (4). Thus, HCMV probably induces an antiviral state that is antagonistic to RhCMV infection. In contrast, when TRFs were preinfected with RhCMV the vast majority of cells (>90%) expressed both RhCMV and HCMV specific IE1. Moreover, preinfection of cells with RhCMV inhibited or greatly diminished IRF3 induction by HCMV (Fig. 10). Since cells expressed immediate-early genes of both RhCMV and HCMV, we conclude that RhCMV actively prevented the activation of IRF3 by an HCMV-mediated trigger. UV-inactivated HCMV is known to be an even more potent inducer of ISG transcription than live virus (4, 7). Therefore, we also examined whether RhCMV could inhibit IRF3 activation by UV-inactivated HCMV. TRFs were (i) preinfected with RhCMV (MOI = 3) for 3 h, followed by treatment with UV-HCMV (MOI = 3), or (ii) pretreated with UV-HCMV for 3 h, followed by infection with RhCMV for 3 h. Positive control cells were treated only with UV-HCMV for 3 h. Cell cultures pretreated with UV-HCMV and subsequently infected with RhCMV generally displayed high IRF3 nuclear localization similar to cells treated with UV-HCMV only. In contrast, TRF cultures preinfected with RhCMV and subsequently treated with UV-HCMV exhibited greatly diminished IRF3 nuclear localization compared to cells treated with UV-HCMV only (Fig. 11). These data suggest that RhCMV is able to interfere with IRF3 activation and ultimately ISG induction triggered by HCMV entry and fusion events.