Studies on the Contribution of Human Cytomegalovirus UL21a and UL97 to Viral Growth and Inactivation of the Anaphase-Promoting Complex/Cyclosome (APC/C) E3 Ubiquitin Ligase Reveal a Unique Cellular Mechanism for Downmodulation of the APC/C Subunits APC1, APC4, and APC5

Cell culture and infection with HCMV.

Human foreskin fibroblasts (HFs) were cultured in HF growth medium (minimal essential medium [Gibco] supplemented with 10% fetal bovine serum, 1.5 μg/ml amphotericin B, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin). Cells were incubated at 37°C with 7% CO₂. To prepare for infections, cells were allowed to come to confluence and maintained for 3 days prior to infection to allow exit from the cell cycle. Virus was diluted in HF growth medium at one of the multiplicities of infection (MOI) indicated below and added to cells. After 6 h, virus was removed, cells were washed with phosphate-buffered saline (PBS), and fresh HF growth medium was added. For siRNA experiments, infection took place 96 h after siRNA transfection. For proteasome inhibition experiments, cells were treated with Cre and 100 nM salinosporamide A (SalA; gift from Bradley Moore, Scripps Institute of Oceanography, University of California, San Diego, CA) or vehicle and harvested 18 h later.

Lentiviral constructs.

The UL21a and UL97 open reading frames (ORFs) were amplified from the pHB5 bacterial artificial chromosome (BAC) and inserted downstream from the second loxP sequence in pLV-EF1α-loxp-Neo-loxp and pLV-EF1α-loxp-Puro-loxp, respectively (25). To accomplish this, UL21a was amplified using the forward and reverse primers 5′-GATCGAGTCGACATGGGAGGTAGCCCTGTTCC-3′ and 5′-GATCGGATCCTTAAAACTGGTCCCAATGTTCTTC-3′, placing a SalI site at the 5′ end. The amplicon was digested with SalI and then phosphorylated and ligated into the pLV-EF1α-loxp-Neo-loxp vector, which had been digested with unique SalI and SmaI sites. UL97 was amplified using forward and reverse primers 5′-GGGATCCACCGGCCCACTATGTCCTCCGCACTTCGGT-3′ and 5-CTTTACTTGTACCCCGCCTTTCCCCTCAGCAACCG-3′ and inserted into the SmaI-linearized pLV-EF1α-loxp-Puro-loxp backbone using the InFusion ligation-independent cloning kit (CloneTech).

Complementing cell lines.

Complementing cell lines were created by transducing HFs (passage 12-15) with lentivirus expressing the following bicistronic constructs. HF-21a cells (expressing UL21a) received pLV-EF1α-loxp-Neo-loxp-UL21a. HF-97 cells (expressing UL97) received pLV-EF1α-loxp-Puro-loxp-UL97. HF-21a/97 cells (expressing UL21a and UL97) first received pLV-EF1α-loxp-Neo-loxp-UL21a, followed by pLV-EF1α-loxp-Puro-loxp-UL97 several days later. The transductions were separated to avoid lentiviral recombination. Lentiviral particles were made by transfecting the lentiviral constructs into 293FT cells with ViraPower lentiviral packaging mix (Life Technologies), and transductions were performed as previously described (25). HF-97 cells were selected with 1 μg/ml puromycin, HF-21a cells were selected with 400 μg/ml G418, and HF-21a/97 cells were selected with both.

Construction of BAC recombinants.

pHB5-IE-Cre-1.2-flp has been described previously (25) and was used as the starting BAC for the mutant lacking UL21a (Δ21a) and the Δ21a/97 mutant. To construct deletions, the UL21a open reading frame was replaced by AmpR, and the UL97 ORF was replaced by TetR. AmpR and the promoter sequence were amplified from pcDNA3 using the following primer sequences: 5′-AGCCATGCAGCGTGCGGCGCCTCTCTCATGGATCCACTGTCACCGTCGCGGTGCGCGGAACCCCTATTTG-3′ and 5′-ATACAGACTTTTATATGATCCTTGTACAGATGTAAATAAAATGTTTTTATGGGGTCTGACGCTCAGTGG-3′. TetR was amplified from the pACYC184 vector using the following primer sequences: 5′-GTCGGTGTGGTAGCTAGTGCAGCCTTAGGAACAGGGAAGACTGTCGCCACGCCCCATACGATATAAGTTG-3′ and 5′-GTACCTTCTCTGTTGCCTTTCCCCTCAGCAACCGTCACGTTCCGCGTCCCGGACAGGAGCACGATCATG-3′. Recombination of the insert into the BAC was carried out in SW105, a modified strain of DH10B Escherichia coli, via heat-inducible recombinase. Clones were selected and confirmed by restriction endonuclease digestion and field inversion gel electrophoresis (FIGE) and by PCR. To grow deletion viruses, complementing cell lines were replated at low density. One day later, they were harvested and electroporated with the BAC along with pcDNA3-pp71 as described previously (25). Viral supernatants were harvested periodically, and their titers determined on the HF-21a/97 complementing cell lines.

Preparation of cell-permeable Cre.

Cell-permeable Cre recombinase was prepared as described in reference 26. Briefly, His-Tat-nuclear localization signal (NLS)-Cre protein was purified on Ni-NTA (Ni-nitrilotriacetic acid) Superflow resin and dialyzed against HEPES-buffered saline and HEPES-buffered glycerol, both tissue culture grade. The amount of Cre required to induce recombination in nearly all cells was determined by treating HF-97 cells with dilutions of Cre, followed by anti-UL97 immunofluorescence.

Depletion of APC/C subunits.

Cells were seeded at low density in antibiotic-free growth medium. The next day, depletion of a specific APC/C subunit was performed with 75 pmol siGENOME Smartpool siRNA (Dharmacon/Thermo Fisher) per well in 6-well dishes using Lipofectamine RNAiMax (Life Technologies). The transfection mixture was removed 4 to 6 h after transfection, and HF growth medium was added. The oligonucleotides targeting the sequences were as follows. For APC1, GGACAAGUGUGCACAAUUG, GGUUACAAUCCACGAGAAA, GAGUGGUUCGGGUGGGAAA, and GGCAUUGGCAGUUUAUAAA. For APC3, CAUGCAAGCUGAAAGAAUA, CAACACAAGUACCUAAUCA, GGAGAUGGAUCCUAUUUAC, and GGAAAUAGCCGAGAGGUAA. For APC4, GGACAUAGAUGAUGAAUGG, CAACACAGCUGGCGAGGUU, GCUCAAAUCUUCGUCAUGU, and GAGGAUGAAUCAAAUCUUC. For APC5, GAAGGCGAGUUGAAGGAUA, UGUCAGAGCUCAUCGAUAU, UAGGAAAGCGGUUGUAUUA, and GAACAGGCCUUAAGUCUUC. For APC6, GUAGAUGGCUUGCAAGAGA, GCUACAAGCUUACUUCUGU, UGGAAGAGCCCAUCAAUAA, and CUAUGGACCUGCAUGGAUA. For APC8, GCAGUUGCCUAUCACAAUA, UGAAACAGUUGAUAGCUUA, GUAGAAACGUGCUGUGUAA, and GAAAUUAAAUCCUCGGUAU. Finally, for nontargeting pool 1, UAGCGACUAAACACAUCAA, UAAGGCUAUGAAGAGAUAC, AGUAUUGGCCUGUAUUAG, and AUGAACGUGAAUUGCUCAA.

Western blot analysis.

Cells were harvested by trypsinization, washed with PBS, and pelleted. Pellets were lysed in Laemmli buffer with protease inhibitor cocktail (Sigma) and Halt phosphatase inhibitor cocktail (Thermo Scientific). Lysates were sonicated to shear DNA, and equal cell number equivalents of protein were run on Tris-glycine SDS gels and transferred to nitrocellulose. Membranes were probed with the following antibodies (Ab): rabbit (Rb) anti-geminin Ab (clone FL-209; Santa Cruz Biotechnology), mouse (Ms) anti-Cdc6 Ab (Neomarkers), Ms anti-Cdh1 Ab (Calbiochem), Ms anti-tubulin Ab (Sigma), Rb anti-APC1 Ab (A301-653A; Bethyl), Rb anti-APC3 Ab (Thermo Scientific), Rb anti-APC4 Ab (A301-176A; Bethyl), Rb anti-APC5 Ab (A301-026A; Bethyl), Rb anti-APC6 Ab (A301-165A; Bethyl), Rb anti-APC8 Ab (A301-181A Bethyl), and Rb anti-cyclin A Ab (Clone H-432, Santa Cruz Biotechnology). Antibodies to HCMV proteins included Ms anti-IE2 Ab (Chemicon), Ms anti-IE1/2 Ab (CH160; Virusys), Ms anti-pp28 Ab (Virusys), Ms anti UL44 Ab (Virusys), Rb anti-glutathione S-transferase (GST)-UL97 Ab (a kind gift from Don Coen), Rb anti-UL21a Ab (A kind gift from Dong Yu), Ms anti-UL57 Ab (Virusys), and Ms anti-pp65 Ab (Virusys).

Determination of viral titers.

Infected cell supernatants were centrifuged to remove cell debris, mixed with 1% dimethyl sulfoxide (DMSO), and stored at −80°C. Plaque assay was performed using the HF-21a/97 complementing cells to efficiently determine the titer of the UL21a/UL97 double mutant.

Immunofluorescence.

Cells grown on coverslips were fixed for 10 min at room temperature with 3.7% formaldehyde in PBS. Cells were permeabilized for 5 min with 0.2% Triton-X in PBS, washed twice, and blocked with 10% normal goat serum in PBS. Primary antibodies against UL21a or UL97 were diluted in 5% normal goat serum and incubated with the coverslips for 25 min at room temperature. After washing, the coverslips were incubated in Alexa Fluor 594 (Life Technologies)-labeled secondary antibodies, nuclei were stained with Hoechst in 5% normal goat serum, and cells were mounted on slides with Prolong gold antifade reagent (Life Technologies). Images were acquired on an epifluorescence microscope.

Real-time RT-PCR.

RNA was isolated from pellets using the DNA/RNA/protein extraction kit (Norgen) according to the manufacturer's instructions. Reverse transcription (RT) was performed on 1 μg of RNA per sample using ThermoScript reverse transcriptase (Life Technologies). APC1 (5′-TCATGGCTGGCTCAGGAAACCTAA-3′ and 5′-ACAGAGAAGAGCGGCAATGGAAGA-3′) and tubulin (sequences described previously [27]) were amplified using SYBR green (Life Technologies). APC5 (sequences described previously [10]) was amplified with TaqMan polymerase (Life Technologies). Real-time PCR was performed on an ABI Prism 7000 (Life Technologies), and relative quantifications were determined by comparison to standard curves for each cDNA. APC1 and APC5 transcript levels were normalized to the level of tubulin.

Creation of cell-permeable Cre-inducible fibroblast cell lines expressing UL21a and UL97 alone or together.

In order to study the individual and combined contributions of UL97 and UL21a to APC/C function both in the absence of other viral proteins and in the context of infection with a mutant lacking both genes, we used a Cre recombination-based system to conditionally express UL21a and UL97 in HF cells (Fig. 1A). HFs were transduced with bicistronic lentiviruses encoding a selectable marker flanked by loxP sites upstream from UL21a or UL97. Translation of this constitutively produced transcript yields the selectable marker gene product. In the presence of Cre recombinase, the selectable marker is recombined out, and translation of the viral protein commences. This allows the expression of each viral protein, alone or in combination, upon treatment with cell-permeable Cre (His-Tat-NLS-Cre) (26) or infection with the Cre-expressing wild-type (WT) or mutant HCMV. These cells are referred to as HF-21a, HF-97, and HF-21a/97 cells. The results in Fig. 1B show that treatment of HF-21a and HF-97 cells with cell-permeable Cre results in the expression of the corresponding viral proteins in a high percentage of the cells.

UL21a but not UL97 is sufficient to promote APC/C substrate accumulation in the absence of infection.

To study the effects of UL97 and UL21a, alone or together, on APC/C and substrate accumulation, we added cell-permeable Cre to HF-21a, HF-97, and HF-21a/97 cells after they were confluence synchronized in G₀/G₁. Cell-permeable Cre was also added to control confluence-synchronized HFs that did not express either protein. After 12 h of treatment, cells were washed twice with PBS, and growth medium was added. Cells were harvested 72 h after the addition of Cre and processed for Western blotting. For comparison, we infected HF with WT HCMV. We noted that Cre induction of the viral proteins in these cells, as well as infection with WT virus, results in a mobility shift of Cdh1 in cells expressing UL97, consistent with its phosphorylation by this viral kinase (Fig. 2). Upon the expression of UL21a, the APC/C substrates geminin and Cdc6 accumulated. This is consistent with previous reports (20) that UL21a causes substrate accumulation. However, UL97 was not sufficient to induce this accumulation. It appears that UL97 alone, despite its ability to phosphorylate Cdh1, is not sufficient to induce APC/C substrate accumulation in confluence-synchronized HFs. Additionally, there was no synergistic effect on geminin or Cdc6 levels from the expression of both proteins. We also show that the extent of APC/C substrate accumulation corresponds with the level of UL21a expressed. In an HF-21a/97 cell line expressing lower levels of UL21a (Fig. 2, left), the accumulation of substrates was less than that in a second line of HF-21a/97 cells expressing more UL21a (Fig. 2, right).

UL21a is necessary and sufficient for the degradation of APC4, APC5, and APC1.

To confirm that the Cre-induced expression of UL21a in the absence of other viral proteins also led to the degradation of APC4 and APC5 as previously reported (19), we assayed the lysates described above by Western blotting with antibodies to APC/C subunits. The results in Fig. 2 show that, as expected, the levels of APC4 and APC5 were significantly reduced during the infection with WT virus. The levels of APC4 and APC5 were also lower when UL21a was expressed alone or in the presence of UL97, and the magnitude of the decrease was dependent on the levels of UL21a. Surprisingly, we also found that in the infected cells or cells expressing UL21a, there was a loss of APC1, in addition to the lower levels of APC4 and APC5. There was not a global degradation of APC/C subunits, as APC6 and APC8 levels did not change.

Since this result differed from that obtained previously, in which it was shown with a different antibody directed against APC1 (28, 29) that this subunit appeared to be stable and localized to the nucleus during infection (9), we confirmed the specificity of the antibody used here with siRNAs directed against the APC1 transcript and a nontargeting control siRNA. As shown by the results in Fig. 3A (right), the detected band disappears in the lane containing lysate from mock-infected cells in which APC1 was depleted by siRNA. For comparison, we show that in cells infected with WT HCMV, the band was not present when the cells were treated with either APC1 siRNA or nontargeting control siRNA. To document that the APC1 antibody used in this study detects a band of the correct molecular weight, we also included a lane that contained lysate from uninfected cells that had been immunoprecipitated with APC3 antibody (Fig. 3A, left, IP). We suspect that the previously used antibody, which was an affinity-purified rabbit antibody directed against a peptide of APC1 (aa 326 to 342) (28, 29), detects an alternative stable form, but we cannot confirm this as the antibody is no longer available. However, in support of the results shown here, a recent study showed by quantitative mass spectrometry that there is a 4- to 5-fold reduction in APC1 levels in infected cells (30). As further confirmation that UL21a is needed for the degradation of APC1, we infected HF cells with WT HCMV or a mutant virus lacking UL21a (Δ21a virus) and harvested the cells at 24 h postinfection (hpi) for Western blot analysis. The results in Fig. 3B show that, as was the case for APC4 and APC5, UL21a is necessary for degradation of APC1. To test whether APC1 degradation was proteasome dependent as has been shown for APC4 and APC5 (10), we induced UL21a expression in the HF-21a cells with Cre treatment in the presence or absence of the proteasome inhibitor salinosporamide A (SalA). We harvested the cells 18 h posttreatment for Western blot analysis. The results in Fig. 3C show that UL21a levels increased with proteasome inhibition, as expected, and that APC1 and APC5 degradation was prevented.

The APC/C degradation phenotype induced by HCMV can be mimicked by disruption of the APC8-APC1-APC4-APC5 interaction via siRNA-mediated depletion of individual APC/C subunits.

The specific APC/C subunit(s) that UL21a interacts with to mediate the degradation of the platform subunits is unknown. To further investigate the mechanism of APC/C subunit degradation by UL21a, we proceeded to test the hypothesis that UL21a might first target one of the platform subunits and that depletion of this subunit might prevent the degradation of the other two subunits. To address this, we first transfected HFs with siRNA against APC1 or APC5. Unexpectedly, we found that knocking down either of these subunits was sufficient to cause the degradation of APC1, APC4, and APC5 in uninfected cells. These experiments were repeated multiple times, and representative results are shown in Fig. 4. There was not global degradation of subunits, as the levels of two other APC/C subunits in the TPR subcomplex, APC6 and APC8, were not affected by the depletion of APC1 or APC5. To rule out off-target effects of the siRNA, we performed real-time RT-PCR on these samples and found that the siRNAs against APC1 and APC5 did not affect the RNA levels of the other subunit (Fig. 4B).

We extended this experiment to include knockdown of the other APC/C subunit in the platform subcomplex, APC4, and three APC/C subunits in the TPR subcomplex, APC3, APC6, and APC8 (Fig. 4C). We were surprised to find that knockdown of APC8 is sufficient to cause a reduction of APC1, APC4, and APC5, and knockdown of any one of the group of APC1, APC4, and APC5 causes reduction of the other three but not of APC8. This was of interest since the recently solved structure of the human APC/C places APC1 next to APC8 (15). The levels of APC6 were not affected by depletion of these subunits, and the depletion of APC6, albeit incomplete, did not significantly affect the levels of the other subunits. In the case of APC3, we sometimes observed that its depletion modestly affected the levels of some of the other subunits.

This unexpected result may underlie the cellular mechanism by which the APC/C subunits are degraded during infection and demonstrates that UL21a may need only to disrupt the complex in order to effect APC1, APC4, and APC5 degradation.

A double mutant, the Δ21a/97 virus, is growth compromised compared to Δ21a and WT virus.

Previously, it was shown that a mutant virus that contained point mutations in UL21a that still allowed degradation of APC4 and APC5 seemed to have no growth defect but was severely growth impaired if a second mutation that eliminated the expression of UL97 was introduced (19). A second UL21a mutant virus that contained point mutations in UL21a that abrogated its ability to degrade APC4 and APC5 also replicated like the WT but was even more severely growth impaired than the first mutant when the UL97 mutation was introduced. One difficulty in interpreting these experiments was that the mutant viruses were not propagated on complementing cell lines and replication of the mutant viruses was analyzed in a multistep growth assay over 21 days. We therefore proceeded to adapt the method we previously used for complementation of IE2 mutant virus growth (25). We constructed a mutant virus that contained deletions of both UL21a and UL97, denoted as Δ21a/97 virus (Fig. 5A), and propagated it on the Cre-inducible HF-21a/97 cell line to facilitate growth and avoid compensatory mutations. BAC recombinant clones were monitored by PCR, restriction digest, and FIGE (Fig. 5B). In comparison to the results for the WT, the appearance of a 1.4-kbp fragment was diagnostic for deletion of UL21a, and the disappearance of the 13.5-kbp fragment indicated deletion of UL97 (Fig. 5B). Since the WT, Δ21a, and double mutant viruses express Cre under the direction of the IE promoter, there is induction of both UL21a and UL97 in the HF-21a/97 cells upon infection. The virus stock and viral supernatants produced in experiments also had their titers determined on HF-21a/97 cells in order to accurately determine the amounts of infectious virus produced.

We characterized the growth kinetics of Δ21a/97 virus by infecting confluence-synchronized HFs with Δ21a/97, Δ21a, or WT virus at a high MOI (MOI of 4) and collected cell pellets and cell supernatants at several times postinfection. Viral titers were determined by plaque assay on HF-21a/97 cells. We found the double Δ21a/97 mutant to be severely growth defective, releasing 2 to 3 log less virus than the WT and 10- to 50-fold less than Δ21a virus (Fig. 6A). The titer of the single mutant Δ21a virus was approximately 5-fold lower than that of the WT. Western blot analysis of viral proteins demonstrated that the double mutant was deficient in the production of late proteins (Fig. 6B) but not of early or IE proteins. While there was little difference in the levels of IE2 86 protein and early protein UL57, there were significant decreases in the levels of pp28 and the late IE2 protein p40 in the double mutant-infected cells and only modest decreases in the late protein levels in the Δ21 mutant-infected cells.

UL21a and UL97 contribute to but are not essential for the accumulation of APC/C substrates in infected cells.

We next tested the ability of Δ21a/97 virus to promote the accumulation of APC/C substrates. Cells were infected at high MOI with Δ21a/97, Δ21a, or WT virus and harvested at 24, 48, and 72 hpi for Western blot analysis. Lysates were probed for Cdh1 and several APC/C substrates. We found that at 24 hpi, there was a significant defect in the accumulation of APC/C substrates in cells infected with either mutant. At later times, however, APC/C substrates accumulated in cells infected with both mutants but with reduced efficiency in cells infected with the Δ21a/97 double mutant compared to their levels in cells infected with either Δ21a or WT virus (Fig. 7). These data indicate that other viral proteins or cellular proteins induced during the infection can contribute to blocking APC/C function in the absence of UL21a and UL97. As expected, in cells infected with the double mutant, Cdh1 failed to exhibit the change in migration that results from phosphorylation by UL97.

The relative contributions of exogenously expressed UL97 and UL21a to complementation of the growth of double mutant Δ21a/97 virus are MOI dependent.

Having compared the influence of UL21a and UL97 on APC/C substrate levels in the context of the infection, we sought to characterize the relative contribution of each protein in the growth of the double mutant Δ21a/97 virus during a single-cycle HCMV infection at high and at low MOI. This was accomplished by comparing the results for Δ21a/97 mutant infection on HF cells to those on HF-21a, HF-97, and HF-21a/97 complementing cells, where UL21a and/or UL97 was provided exogenously upon the expression of Cre by the mutant virus. All cells were confluence synchronized in G₀/G₁ and then infected at a low MOI of 0.1 or a high MOI of 3. At 96 hpi, cell pellets were harvested for Western blot analysis. Supernatants were collected at various times postinfection to determine the titers of infectious virus produced.

Virion production was monitored by titration of supernatants on HF-21a/97 cells to provide complementing function. At both high and low MOI (Fig. 8A and B), infection of HF-21a cells resulted in higher titers of virus than infection of HF cells, with a greater increase in the titer when mutant virus was grown on HF-97 cells at high MOI. At low MOI, there was a delay in the production of virus on HF-21a cells, but at the last time point when supernatants were collected, the amount of virus produced on HF-21a cells was comparable to that produced on HF-97 cells. Infection of cells expressing both proteins (HF-21a/97 cells) with the double mutant at the low MOI resulted in still higher titers than infection of HF-97 cells, but the difference was less at high MOI. Cell pellets from these experiments harvested at 96 hpi were processed for Western blotting and probed for pp28, IE2 86, IE2 60, and IE2 40. Representative blots are shown in Fig. 8C and D. We observed that at high MOI, infection of HF-97 cells with the double mutant led to higher levels of the late proteins pp28, IE2 60, and IE2 40 than did infection of HF-21a cells, consistent with the titration results. (Fig. 8C). Likewise, the results of the Western blot analysis were consistent with the titers at low MOI, with infection on UL21a-expressing and UL97-expressing cells showing similar increases in late protein levels (Fig. 8D). At both high and low MOI, infection of HF-21a/97 cells led to the greatest increase in late protein expression compared to the results for double mutant infection of HFs.

Exogenous inactivation of the APC/C during double mutant infection results in increases in late protein expression and viral titers.

As both UL97 and UL21a perform other functions during the infection apart from phosphorylating Cdh1 and mediating the degradation of APC/C subunits, respectively, we were interested in determining the degree of rescue of the double mutant's growth deficiency that results from inactivation of the APC/C. To decrease the activity of the APC/C, we used siRNAs to deplete APC5 and APC8. A nontargeting siRNA was used as a control. The growth of the double mutant on the various siRNA-treated cells was assessed by measuring late viral protein accumulation and viral titers. Efficient knockdown of both APC5 and APC8 took several days, resulting in the cells coming to confluence (and thus a G₀/G₁ state) prior to infection. Since one function of UL21a is to degrade cyclin A2, we wanted to ensure that this would not be a confounding variable in the experiment. To document that the siRNA-treated cells are not cycling and do not express significant levels of cyclin A2, we harvested the cells at 96 h posttransfection and probed for cyclin A2 along with APC/C subunits 5 and 8 (Fig. 9A). We show that under these conditions, confluent siRNA-treated cells have only low levels of cyclin A2, in contrast with the very high levels found in asynchronous cycling nontreated cells and in siRNA-treated cells that were reseeded at low density to allow the cells to enter G₁ and then harvested 24 h later when they entered S phase of the cell cycle.

In the experiment whose results are shown in Fig. 9B and C, cells were mock infected (Fig. 9B) or infected (Fig. 9C) with Δ21a/97 or WT virus at an MOI of 0.5 4 days after transfection with siRNA. In the uninfected cells depleted of the APC/C subunits (Fig. 9B), there was accumulation of substrates, with slightly greater accumulation in cells treated with siRNA to APC8, probably due to the fact that depletion of APC8 also resulted in degradation of APC1, APC4, and APC5. It should be noted that, in accord with the results described above, depletion of APC8 also resulted in depletion of APC5. Supernatant was collected at various times postinfection, and the cells were harvested at 96 hpi for Western blot analysis. Membranes were probed for APC5 and APC8, to ensure maintenance of protein knockdown late in the infection, and for viral late proteins (Fig. 9C). Western blot analysis showed that the knockdown of both APC/C proteins was maintained until at least 96 hpi and that APC/C subunit knockdown resulted in a greater accumulation of APC/C substrates in cells infected with the double mutant. Knockdown of APC5 and, to a greater extent, knockdown of APC8 also resulted in increased production of late proteins. Supernatants were assayed for viral production by titration on the complementing HF-21a/97 cells. In accord with the levels of viral late proteins observed, viral titers were significantly increased by depletion of APC5, with a greater increase observed in cells depleted of APC8 (Fig. 9D). A comparison of the results in Fig. 8B with those in Fig. 9D shows that the increase in viral titer when mutant virus was grown on cells expressing UL21a relative to the titer when virus was grown on HF cells was comparable to the increase when mutant virus was grown on HF cells that were depleted of APC5 or APC8. While the mutant still showed debilitated growth relative to the growth of the WT, depletion of APC/C subunits was able to increase viral growth in a single-cycle infection. The growth of WT virus was not affected by knockdown of either protein.