Cell cycle deregulation by a poxvirus partial mimic of anaphase-promoting complex subunit 11

Abstract

The anaphase-promoting complex (APC), or cyclosome, is a ubiquitin ligase with major roles in cell cycle regulation. It is required for mitotic exit, but must be deactivated for the G₁/S phase transition to occur. APC consists of at least 12 subunits with the catalytic core formed by a scaffold protein, APC2, and a RING-H2 protein, APC11. APC11 facilitates ubiquitin chain formation by recruiting ubiquitin-charged conjugating enzymes through its RING-H2 domain. We report that a small number of poxviruses encode RING-H2 proteins with sequence similarities to APC11. We show that a representative of these viral proteins mimics APC11 in its interactions with APC, but unlike APC11, the viral protein fails to promote ubiquitin chain formation. This absence of ubiquitin ligase activity is linked to a distinctive sequence variation within its RING-H2 domain. Expression of the viral protein led to cell cycle deregulation and the accumulation of APC substrates in a manner consistent with impaired APC function. Our data characterize this protein as a regulator of APC activity, and consequently, we have called it PACR (poxvirus APC/cyclosome regulator). Deletion of the PACR gene substantially reduced viral replication. Here, we report a viral mimic of an APC component and reveal an intriguing mechanism by which viruses can manipulate cell cycle progression and, thereby, promote their own replication.

Cell cycle progression is controlled by a tightly orchestrated periodic oscillation of key regulators. A major player in the process is the anaphase-promoting complex (APC), a multisubunit ubiquitin ligase, which via ubiquitin chain formation and subsequent proteasomal degradation, controls levels of numerous cell cycle regulators (1). APC thereby regulates progression through mitosis and the maintenance of G₁ phase (2–4). APC is also active in differentiated cells and is required for maintenance of G₀ (5). Disruption of APC function can result in cell cycle arrest at metaphase, compromised G₁ maintenance and genome instability (2, 3, 6, 7). Vertebrate APC is composed of at least 12 subunits with its catalytic core formed by the RING-H2 protein, subunit 11 (APC11), and APC2 (8). APC11 does not determine substrate specificity, but it facilitates the ubiquitination of substrates by recruiting ubiquitin-charged ubiquitin conjugating enzymes.

Poxviruses such as the devastating human pathogen, Variola virus, are large DNA viruses, which replicate exclusively in the cytoplasm and encode many of their own factors for genome replication and expression (9). These characteristics provide an unusual degree of independence from the host cell nucleus, and in contrast to many other viruses, there is only limited evidence of autocrine modulation of the cell cycle by poxviruses (9–11). We detected a RING-H2 protein encoded by a subset of vertebrate poxviruses with substantial sequence similarities to APC11, raising the possibility that these viruses might use these proteins to target APC and manipulate cell cycle control in a manner not previously observed. Here, we report that one of these poxviral RING-H2 protein mimics APC11 in its association with APC, but unlike APC11, lacks ubiquitin ligase activity and, thereby, inhibits APC function. We characterized this protein as a regulator of APC (also called cyclosome) activity; therefore, we call it PACR (poxvirus APC/cyclosome regulator). Our analysis of PACR reveals a previously undescribed mechanism by which viruses manipulate cell cycle progression.

Results

Subset of Poxviruses Encode APC11 Homologs.

An analysis of available poxvirus genome sequences revealed that members of two genera of vertebrate poxviruses, the parapoxviruses and the molluscipoxviruses, as well as the unclassified Crocodile poxvirus and Squirrel poxvirus, each encode a RING-H2 protein with sequence similarities to APC11 (Fig. 1A). Like APC11, the viral proteins, which we have named PACR, contain a classic RING-H2 domain (Fig. 1A), and they share with APC11 the unusual feature of a third zinc binding site (8, 12). The RING-H2 domain of APC11 recruits ubiquitin-charged conjugating enzymes (E2s), whereas its N-terminal stalk is believed to bind the cullin homology domain of APC2, so as to position the ubiquitin for transfer to a substrate bound by other subunits of APC (Fig. 1B) (8). We identified considerable sequence similarities between the N-terminal domain of APC11 and the corresponding region of the viral proteins, suggesting that they too might interact with APC2 via this domain (Fig. 1A). We used the PACR protein of Orf virus (13), the type species of the parapoxvirus genus to explore this possibility.

Fig. 1.

Orf virus PACR is a RING-H2 protein, a homolog of APC11, and like APC11, interacts with C-terminal region of APC2. (A) Alignment of Orf virus PACR (ORFV gene 014) (13) with human APC11 (Hu) and PACR homologues encoded by Bovine papular stomatitis virus (BPSV gene 013) (36), Molluscum contagiosum virus (MOCV gene 026L (37), Squirrel poxvirus (SPV gene A11L) (38), and Crocodile poxvirus (CRV gene 047) (39). Residues forming the first, second, and third zinc binding sites are in pink, blue, and light blue, respectively. Residues of the canonical RING-H2 motif are numbered to assist with discussion of them in the text. Residues identical or similar to those of PACR are shaded in yellow and green, respectively. The red box marks Trp-63, a residue essential for the ubiquitin ligase activity of APC11, and corresponding residues of the viral homologs. The blue box indicates regions predicted to interact with the cullin homology domain (CHD) of APC2. (B) Diagram of APC (1) showing the recruitment of ubiquitin-charged E2 by APC11 and ubiquitin chain formation on a substrate. (C) APC2, APC11, and PACR constructs used in the immunoprecipitation (IP) assays (see SI Materials and Methods and Table S1 for details). (D–F) 293EBNA1 cells were transfected with the indicated (+) combinations of plasmids, harvested after 48 h and cell lysates immunoprecipitated with anti-FLAG or anti-HA agarose beads. The samples were analyzed by SDS/PAGE/Western blotting (WB) for the presence of coprecipitating proteins, using the indicated antibodies. (G) Transfected cell lysates were prepared and analyzed as for D–F except that IPs were performed by using anti-APC3 or anti-cyclin A and protein G-agarose. (H) 293T cells infected with the indicated recombinant virus at a multiplicity of infection (MOI) of 10 were harvested after 24 h and cell lysates prepared and analyzed as for D–F.

PACR Binds APC2 in a Manner that Mimics APC11.

PACR with a C-terminal FLAG fusion (PACR-FLAG) or APC11-FLAG was coexpressed in 293EBNA1 cells along with APC2 carrying an N-terminal HA fusion (HA-APC2) (Fig. 1C). Immunoprecipitates prepared with anti-FLAG-agarose were analyzed by SDS/PAGE and Western blotting. HA-APC2 coprecipitated with APC11-FLAG and with PACR-FLAG, but not with the empty FLAG vector (EV) negative control (Fig. 1D). Reciprocal immunoprecipitations confirmed these interactions (Fig. S1C). We further examined these interactions using two HA-tagged APC2 derivatives. One consisted of the C-terminal 319 amino acids (APC2 CT) containing the cullin homology domain, whereas the other was restricted to the N-terminal 503 amino acids (APC2 NT), and thus, excluded the cullin homology domain (Fig. 1C). It has been shown in vitro that APC2 CT binds to APC11 but APC2 NT does not (8). Both APC11 and PACR coprecipitated with HA-APC2 CT when the immunoprecipitation was conducted with either anti-HA (Fig. 1E) or with anti-FLAG-agarose (Fig. S1 A and B). In contrast, neither APC11 nor PACR coprecipitated with APC2 NT (Fig. 1E; Fig. S1C). These observations indicate that PACR, like APC11, binds to the cullin homology domain-containing C terminus of APC2.

PACR Associates with Endogenous APC.

We next examined whether the viral protein might be incorporated into endogenous APC by transiently expressing PACR-FLAG in 293EBNA1 cells and testing anti-FLAG immunoprecipitates for the presence of endogenous APC3 or APC4 by Western blotting. APC3 and APC4 are located in the outer shell of APC whereas APC11 and APC2 form a separate subcomplex, and are not directly linked with either APC3 or APC4 (Fig. 1B) (1, 14). Endogenous APC3 was shown to coprecipitate with PACR-FLAG or with APC11-FLAG (Fig. 1F). The small amount of APC3 precipitating in the negative control (EV) represents nonspecific binding to the agarose beads as has been reported (15). Reverse immunoprecipitations using anti-APC3 antibody were performed and again showed coprecipitation of PACR-FLAG with APC3 (Fig. 1G). The same strategy was applied to cells transiently expressing HA-APC2 and coprecipitation with APC3 was observed as expected. An anti-cyclin A antibody was used to conduct immunoprecipitations from cells expressing PACR-FLAG, and this negative control revealed no evidence of precipitation of PACR (Fig. 1G). Also, we detected coprecipitation of endogenous APC4 with transiently expressed APC11-FLAG or PACR-FLAG (Fig. S1D). This coprecipitation was not seen with a truncated version of PACR limited to the RING domain (PACR CT-FLAG; see Fig. 3), indicating that, like APC11, interaction of PACR with APC requires the N-terminal domain of PACR. Together, these results suggest that ectopic PACR can incorporate into endogenous APC.

Fig. 3.

Domain swap mutants of PACR and APC11 retain the ability to bind to APC2. The diagrams represent constructs used in the IP assays. The mutated (MU) RING domains are described in Fig. 2. PACR CT is limited to the RING domain and lacks amino acids 2–17; 293EBNA1 cells were transfected with the indicated (+) plasmids. Lysates and IP were analyzed by WB with the indicated antibodies.

To confirm that these interactions were relevant in the context of poxvirus infection, we made use of a recombinant Vaccinia virus expressing a FLAG-tagged PACR (VV-PACR-FLAG). The 293T cells were infected with VV-PACR-FLAG or a control recombinant (VV-Lac). Western blotting of anti-FLAG immunoprecipiates prepared from infected cells revealed coprecipitation of FLAG-tagged PACR and APC4, indicating that interactions of PACR with endogenous APC components are also observed during virus infection (Fig. 1H).

PACR, Unlike APC11, Lacks Ubiquitin Ligase Activity.

Most RING-H2 proteins studied have ubiquitin ligase (E3) activity as manifested by their ability to stimulate polyubiquitin chain formation in an in vitro, substrate-independent ubiquitination assay, and this activity has been demonstrated with APC11 (8, 12, 16). Therefore, we examined PACR for the same activity. In vitro ubiquitination assays of APC11 and PACR were conducted using Ubc5b-His, a promiscuous and robust E2 enzyme (17). As expected, APC11 promoted the formation of polyubiquitin chains (Fig. 2A, lane 1). In contrast, we were surprised to discover that PACR failed to stimulate the formation of polyubiquitin chains (Fig. 2A, lane 2). These observations prompted us to reexamine the sequence similarities between PACR and other RING proteins shown to have ubiquitin ligase activity.

Fig. 2.

PACR lacks the ubiquitin ligase activity of APC11 and analysis of domain swap mutants identifies a region essential to this (in)activity. (A) In vitro ubiquitination assays mixtures contained the indicated (+) components. Reaction products were subjected to SDS/PAGE, and the presence of polyubiquitin chains was detected by antiubiquitin Western blotting. Molecular mass (kDa) are shown to the left. (B) (Upper) Schematic representation of wild-type and domain swap mutated (MU) proteins. The sequences and the locations of the exchanged domains are indicated. Residues of APC11 marked in bold are W63, which is essential for the ubiquitin ligase activity of APC11, and P74 that is conserved among RING-H2 ubiquitin ligases. The corresponding residues of PACR, T61 and F67, are also in bold. (Lower) Products of in vitro ubiquitination assays conducted as described in A except that the E2 used was Ubc5b-GST. The assay mixtures contained the indicated E3.

Crystallographic structural studies or NMR-monitored binding assays of E3-E2 pairs have identified residues of RING ubiquitin ligases that are required for their interaction with E2s (18, 19). These residues include a widely conserved Trp, 4-aa C-terminal of the 6th Cys/His of the RING-H2 motif (equivalent to Trp-63 of huAPC11, marked in Fig. 1A). Mutation of this residue abolishes or significantly reduces ubiquitin ligase activity (12, 19, 20). This Trp is located within a region shown, or predicted, to form an α-helix, which has also been shown to be an important component of the E3-E2 interaction. In the case of PACR, a corresponding Trp is not present, and an α-helix is not predicted to form in this region (for an expanded description of the composition and predicted structure of the PACR RING-H2 domain, see Fig. S2). To examine the functional significance of these features, two domain swap mutants were constructed.

A domain of APC11 (amino acids 61–74) spanning the α-helix containing Trp-63 was incorporated in place of the corresponding region of PACR (amino acids 59–67). Ubiquitination assays conducted with the mutated PACR (PACR MU) indicated that it had acquired ubiquitin ligase activity (Fig. 2B, lane 5). However, an equivalent domain swap in APC11 (APC11 MU), in which amino acids 61–74 were replaced with amino acids 59–67 of PACR, resulted in loss of ubiquitin ligase activity of APC11 (Fig. 2B lane 2). These data indicate that the lack of ubiquitin ligase activity of PACR can be linked to the sequence differences identified in the region between the 6th and 8th Cys/His.

We next examined the ability of the two domain swap mutants to interact with APC2, and showed that they remained capable of coprecipitating APC2 (Fig. 3, lanes 1 and 5). We also showed that neither PACR CT nor APC11 CT was able to precipitate APC2 (Fig. 3, lanes 2 and 6). These data indicate that the N-terminal region of PACR is required for binding to APC2 and mutation of the RING had no effect on PACR or APC11 binding to APC2.

PACR Impairs APC Activity.

We had shown that PACR interacts with APC2 in a manner indistinguishable from APC11, but does not possess ubiquitin ligase activity. These observations raised the possibility that PACR might incorporate into APC and, thereby, disrupt function of APC. To explore this possibility, we constructed a set of cell lines stably expressing either full-length or truncated versions of PACR, APC11, or APC2. DNA content profiles of actively growing populations of these cells lines were then obtained by flow cytometry. The PACR cell line exhibited a distinctive DNA content profile consistent with impaired APC function, with fewer cells in G₁ phase, more in S, and an accumulation of cells in G₂/M (Fig. 4A). In contrast, APC11, APC2, and control (EV) cell lines exhibited normal cell cycle profiles, indicating that expression of functional units of APC had no impact on cell cycle progression. Similar profiles were also seen with cell lines expressing the RING-H2 domain of either protein (PACR CT and APC11 CT), indicating that this domain was not able, by itself, to target APC, a result consistent with our immunoprecipitation data (Fig. 3). In contrast, cells expressing only the N-terminal region of either protein induced cell cycle profiles similar to cells expressing PACR, consistent with the ability of these N-terminal domains to interact with APC2. Expression of the C-terminal domain of APC2 was predicted to interfere with the assembly of a functional APC by binding to and occluding available APC11. Indeed, the cell cycle profile displayed by cells expressing APC2 CT was similar to that seen with cells expressing PACR.

Fig. 4.

PACR expression causes accumulation of cells in G₂/M and an accumulation of the APC substrates, cyclin A, cyclin B, and TK. (A) Stable cell lines expressing the indicated proteins were stained with PI and DNA contents measured by flow cytometry, followed by Modfit software analysis. (Left) Representative dataset; (Right) percentages (means ± SD) of cells in G₁, S, and G₂/M phase of three independent clones of each cell type. (B) Stable cell lines expressing PACR, APC11, or EV were synchronized by treatment with nocodazole (+) or untreated (−); 16 h later, cyclin A and B levels were monitored by Western blotting. Actin was used as a loading control. (C) PACR cell lines, expressing high (PACR2) and low (PACR1) levels of PACR, as well as APC2 CT and EV cell lines were serum starved for 48 h. The levels of TK, PACR-FLAG, HA-APC2 CT, and actin were monitored by Western blotting.

APC promotes mitotic exit and maintains the duration of G₁ by targeting key cell cycle regulatory proteins such as cyclin A and cyclin B for degradation (2). To further test the inhibitory effects of PACR on APC function, cyclin levels were compared in the PACR, APC11, and EV cell lines. Elevated levels of cyclin B and cyclin A were observed in the PACR cell line, suggesting an early mitotic block specific to that cell line (Fig. 4B Left). To more directly link these observations to inhibition of APC, we treated the cell lines with nocodazole to activate the spindle check point. Under these conditions, degradation of cyclin B by APC is inhibited (15), and in line with our predictions, cyclin B levels were the same in all three cell lines (Fig. 4B Right). In contrast, cyclin A is degraded by APC independently of the spindle check point (21), and therefore, is unaffected by nocodazole treatment as was seen in the APC11 and EV cell lines. However, cyclin A levels remained high in the nocodazole-treated PACR cell line consistent with a direct inhibition of APC.

Thymidine kinase (TK) is another protein ubiquitinated by APC and whose levels are negatively regulated by APC during G₁ phase (22). To investigate the effect of PACR expression on TK levels, PACR, APC2 CT, or EV cell lines were starved of serum to synchronize them at G₁, when APC is active (7). Only a small amount of TK was detected in the EV cell line after serum starvation, as expected (Fig. 4C). In contrast, substantially greater amounts were seen in APC2 CT-expressing cells, which were predicted to have impaired APC function. A similar effect was seen with PACR-expressing cells. Also, the TK levels correlated with the differing PACR expression levels of two PACR clones (PACR1 and PACR2; Fig. 4C). These results suggest that PACR expression interferes with APC activity in G₁ phase, potentially undermining G₁ maintenance.

It should be noted that with continued passaging of these cell lines expression of PACR (and the other inhibitory constructs) was eventually lost, consistent with some inhibitory effect of PACR on cell proliferation. Also, significant numbers of cells with subG₁ DNA contents, indicative of apoptotic DNA fragmentation, were observed in these same cell lines and these numbers, along with the size of the G₂/M population correlated with the extent of PACR expression (Fig. S3 A and B). These observations suggested that PACR expression could result in a block in the cell cycle at G₂/M and apoptotic cell death. Indeed, inhibition of APC has been shown to result in metaphase-specific cell death (23, 24). We examined this possibility further and confirmed that transfection of cells with high concentrations of the expression plasmid so as to give rise to transient hyperexpression of the viral protein resulted in substantial changes in cellular DNA profiles. Most dramatic was the appearance of a major proportion of the cells in a subG₁ population (Fig. S3C). Also, elevated caspase activity, an indication of apoptosis, was observed in PACR-transfected cells (Fig. S3D). Also, expression of the domain swap form of APC11 (APC11 MU), which we had shown lacked ubiquitin ligase activity, generated a cell DNA content profile similar to that of PACR with a substantial subG₁“shoulder” (Fig. S3C). In contrast, expression of PACR MU, which had gained ubiquitin ligase activity, resulted in a profile indistinguishable from that seen with APC11. These observations are consistent with PACR having the ability to inhibit the function of APC.

PACR Promotes Viral Growth.

To examine the effect of PACR on Orf virus replication, we constructed a recombinant virus in which the PACR gene was deleted (OV-PACR-KO). The successful recovery of this virus demonstrated that PACR is not essential to Orf virus replication in cultured cells. However, the recombinant virus demonstrated markedly reduced plaque size (Fig. 5). We conducted single step and multistep growth curves, which revealed that growth of OV-PACR-KO was substantially impaired, with yields of ≈5% of that seen with wild-type virus. Restoration of the PACR gene to the knockout virus, generating OV-PACR-RE, restored plaque size, and growth characteristics to that seen with wild-type virus (Fig. 5). These observations demonstrate that PACR has a significant role in Orf virus replication.

Fig. 5.

Deletion of the PACR gene restricts viral growth. LT cells were infected with wild type Orf virus (WT), OV-PACR-KO (KO), or OV-PACR-RE (RE). Single plaques (Upper) were photographed (Olympus 1 × 71) at 3 days postinfection after staining with crystal violet. (Scale bar, 0.2 mm.) For multistep growth analyses cells were infected at MOI 0.02 and titers of pfu determined in triplicate (mean + SD). The results are representative of three independent experiments. For single step growth cells were infected at MOI 4. Samples were taken at the indicated times, and viral titres determined in duplicate (mean shown). The results are representative of two independent experiments.

Discussion

This work provides a demonstration of a viral protein able to disrupt the function of APC by partial mimicry of APC11. PACR is also the first RING-H2 protein shown to lack ubiquitin ligase activity, at least with the promiscuous E2, Ubc5. The analysis of domain swap mutants confirmed our bioinformatic prediction that the inability of PACR to support the formation of polyubiquitin chains in an in vitro assay is associated with the sequence between the 6th and 8th Cys/His. We postulate that this lack of activity relates to the absence of residues essential for E2-E3 interaction. Key differences are the lack of an α-helix, the absence of a residue equivalent to Trp 63 of APC11, and, to some extent, the lack of highly conserved Pro within the 7th and 8th Cys/His (Fig. 1A). Also, each of the PACR homologs encoded by different poxviruses shares these distinctive features that distinguish PACR from APC11 (Fig. 1A), suggesting that each of these proteins lacks ubiquitin ligase activity. Furthermore, the similarities between their N-terminal domains allow us to propose that each of these proteins also interacts with APC2 as we have demonstrated for Orf virus PACR.

Might the incorporation of PACR direct APC-PACR to function with an E2 other than Ubc10 or Ubc5? We have not examined this question directly; however, ubiquitin chain formation promoted by most classes of E2s, including Ubc5, leads to the ubiquitinated protein's proteasomal degradation. However, ubiquitination involving Ubc9 and Ubc13 promotes other fates for target proteins such as changes in subcellular localization (25, 26). Therefore, even if APC-PACR is capable of working with Ubc9 and 13, the fate of APC substrates would be changed by the presence of PACR.

One of the hallmark functions of APC is its ability to maintain cells in the quiescent stage of the cell cycle and inhibition of APC leads to cell cycle reentry (5). Orf virus replicates only in epidermal skin cells, and the virus is detected in zones containing differentiated G₀ cells (27). These features suggest that Orf virus is likely to require means of manipulating cells into a state supportive of viral genome replication. Our results suggest that Orf virus infection might stimulate differentiated epidermal skin cells to enter an S phase-like state as a result of the functional disruption of APC by PACR. This state could provide cellular factors to assist viral DNA replication. In support of this model, we show that deletion of the PACR gene substantially reduced viral replication in cultured cells. There is an intriguing correlation that further supports this hypothesis. Most vertebrate poxviruses encode their own TK and ribonucleotide reductase (RR), two enzymes that provide nucleotides for DNA replication and the cellular versions of which are ubiquitinated by APC (22, 28). These viruses uniformly do not encode a PACR-like factor. However, the small number of poxviruses that encode homologs of PACR uniformly do not encode their own TK or RR (Table S2), raising the possibility that these viruses use inhibition of APC as an alternative method of obtaining specific enzymatic activity that can support their replication. Inhibition of APC as well as promoting cell cycle reentry also leads to a subsequent blockage at G₂/M. We have not examined whether such a block occurs in Orf virus-infected cells, and the role that a potential M phase block induced by PACR might have in infection by these viruses remains to be determined, but there is growing evidence of an association between viral infection and G₂/M arrest (29).

This report of the manipulation of a major ubiquitin ligase adds to a growing body of recent evidence of poxviral manipulation of the ubiquitin–proteasome system (30, 31). However, this report of a viral mimic of an APC component reveals a previously undescribed mechanism by which viruses manipulate the ubiquitin system and, specifically, the cell cycle. As a potential competitive inhibitor of APC11, PACR would have the ability to block activities of APC during both G₁ and M phases. This range of inhibitory activity contrasts with identified cellular inhibitors of APC, which tend to be focused on substrate recognition (1). Two reported instances of viral inhibition of APC also focus on APC coactivators (Cdc20 and Cdh1) with roles in substrate recognition. Human cytomegalovirus encodes an undefined factor that disrupts APC function by modifying the phosphorylation status of Cdh1 (32), and the E2 protein of some human papillomaviruses inhibits Cdh1 and Cdc20 association with APC (23). Also, the chicken anemia virus protein, apoptin, binds APC1 causing APC destabilization (24). The mechanism of action of PACR is distinct from each of these and unique among cellular or viral inhibitors of APC.

The regulation of APC is of particular interest in understanding mechanisms of cell hyperproliferation and cancer development. Further investigation of the viral APC inhibitor described here will provide insights into such mechanisms and potentially new therapeutic opportunities.

Materials and Methods

Cell Culture.

Human embryonic kidney 293 cells expressing Epstein-Barr virus-encoded nuclear antigen (293EBNA1) or the SV40 large T antigen (293T) and HeLa cells were grown in DMEM (Gibco 12800-017) supplemented with 10% FCS, 2 mM glutamine, and antibiotics; 143B cells and primary lamb testis (LT) cells were grown in similarly supplemented Eagle MEM (Sigma M0769 or M0643, respectively). The 293EBNA1 cell lines with episomally maintained expression vectors were constructed by transfection with purified plasmid DNA using FuGENE6 (Roche), selection with hygromycin-B and clonal expansion.

Viruses.

Orf virus strain NZ2 was propagated in LT cells as previously described (13). A recombinant Orf virus (OV-PACR-KO) in which the PACR coding region was deleted and replaced with the Escherichia coli β-glucuronidase reporter gene under the control of a poxvirus promoter (PH5) was constructed by homologous recombination in LT cells according to standard procedures (33). A second recombinant Orf virus (OV-PACR-RE) was then constructed by replacing the reporter cassette of OV-PACR-KO with the PACR gene, under the control of its natural promoter, along with the β-galacotsidase coding region under control of a strong Orf virus late promoter, PF1. VV-PACR-FLAG, a recombinant Vaccinia virus strain Lister expressing a C-terminal FLAG-tagged PACR under control of the poxvirus promoter P7.5 from the TK locus was constructed according to standard procedures (34). A control recombinant expressing β-galactosidase was constructed in the same manner (VV-Lac). Details of each construct are provided in SI Materials and Methods.

Plasmid Constructs.

Coding regions amplified by PCR were cloned into expressions vectors based on pAPEX3 for mammalian expression (35) or pET-21D for expression in E. coli. The cloning strategies resulted in expression of proteins with N-terminal HA, C-terminal FLAG, or His tags. Details of each construct are provided in SI Materials and Methods.

Expression and Purification of His-Tagged Proteins.

E. coli BL21(DE3) was grown in LB/Sorbitol medium containing carbenicillin (50 μg/mL), chloramphenicol (34 μg/mL), betaine (2.5 mM), and ZnSO₄ (100 μM) at 37 °C to mid log phase and induced by addition of IPTG (0.3 mM). After incubation at 25 °C overnight, cells were harvested, and lysed by three cycles of freeze/thaw in the presence of 1% Triton X-100. Cleared lysates were mixed with Ni-NTA resin (Qiagen) (4 °C overnight). After extensive washing (20 mM Tris·HCl, pH 7.4/500 mM NaCl/10% glycerol/0.2% Nonidet P-40/2 mM β-mercaptoethanol), bound proteins were released by washing in the same buffer containing 250 mM imidazole. Purified proteins were concentrated using a Centricon YM-10 (Amicon) and dialyzed against ubiquitination buffer (50 mM Tris·HCl, pH 7.4/2.5 mM MgCl₂/1 mM DTT/50 mM NaCl).

Ubiquitination Assays.

In vitro ubiquitination assays were conducted at 37 °C for 1 h in a 10 μL of volume containing 75.7 nM E1 (human; Sigma), 606 nM E2 (human Ubc5b-GST; Sigma or Ubc5b-6xHis tagged), 4 μM RING protein, 52 μM ubiquitin (bovine erythrocytes; Sigma), 2 mM ATP, and reaction buffer. The reactions were stopped by the addition of 10 μL of SDS loading dye and boiled for 5 min before analysis by SDS/PAGE and Western blotting with anti-ubiquitin antibody.

Antibodies.

Antibodies used were anti-FLAG M2 HRP (Sigma, 1:2,500), anti-HA HRP (3F10; Roche, 1:1,000), anti-TK (3B3.E11; Abcam, 1:500), anti-Cyclin A (Ab-2; Calbiochem, 1:1,000), anti-Cyclin B (CNS1; Santa Cruz, 1:1,000), anti-ubiquitin (P4D1; Santa Cruz, 1:1,000), anti-actin (C-11; Santa Cruz, 1:1,000), anti-APC3 (Sigma, 1:1,000), and anti-APC4 (Abcam, 1:200). Secondary antibodies used were HRP conjugated anti-mouse (DakoCytomation, 1:1,000), anti-goat (Sigma, 1:10,000), and anti-rabbit (DakoCytomation, 1:2,500).

Immunoprecipition.

The 293EBNA1 cells were transfected with specific plasmids using FuGENE6. After 48 h, 1–3 × 10⁷ cells were lysed in 1 mL of lysis buffer (500 mM NaCl/0.5% Triton X-100/20 mM Tris·HCl, pH 7.5), and cell debris removed by centrifugation. Supernatants were incubated with either anti-FLAG M2-agarose (Sigma A2220) or anti-HA-agarose (HA-7; Sigma A2095). Beads were washed four times with lysis buffer when examining interactions between APC2 and APC11 or PACR, or with RIPA buffer (50 mM Tris·HCl, pH 7.4/150 mM NaCl/1 mM EDTA/1% Triton X-100/1% sodium deoxycholate/0.1% SDS) when examining interactions between APC11 or PACR, and APC2 NT or CT. When testing for an association of PACR with endogenous APC, lysates were precleared with 10 μL of EZview protein G affinity gel (Sigma, E3403) that had been prewashed with the immunoprecipitation buffer (20 mM Tris·HCl, pH 7.5/150 mM NaCl/0.5% Nonidet P-40/20 mM β-glycerophosphate/1 mM NaF/1 mM DTT/5 mM MgCl2). APC3 or cyclin A (20 μg) antibody was then added to the lysates. After 2–4 h incubation at 4 °C, 20 μL of prewashed EZview protein G affinity gel was added to the lysates followed by incubation at 4 °C overnight. Alternatively, FLAG-tagged proteins were precipitated with anti-FLAG M2-agarose. Beads were washed two times with the lysis buffer, four times with immunoprecipitation buffer, and analyzed by SDS/PAGE and immunoblotting.

Immunoblotting.

Samples were separated by SDS/PAGE and transferred to nitrocellulose membranes using standard methods. FLAG-tagged and HA-tagged proteins were detected with HRP-conjugated anti-FLAG M2 antibody and anti-HA antibody, respectively. Other primary antibodies were detected with secondary anti-mouse or anti-goat antibody conjugated to HRP. Bands were visualized with SuperSignal West Pico chemiluminesce substrate (Pierce 34080).

Cell Synchronization and Flow Cytometric Analysis of Cell Cycle.

Two to three × 10⁶ cells were washed twice in 5 mL of cold PBS with 0.5% BSA and fixed in 70% cold ethanol overnight at 4 °C. Cells were then washed twice with 5 mL cold PBS and incubated in 1 mL PBS containing 50 μg/mL propidium iodide (PI) and 0.5 μg/mL RNase for 30 min. Cellular DNA content was determined by fluorescence-activated cell sorting (BD FACSCalibur). For each sample, 10,000 events (within gated area) were analyzed using either the MODFIT software package (Verity Software) or Cell Quest (BD). Cells were arrested at M phase by treatment with nocodazole (330 nM) overnight at 37 °C. To synchronize cells in G₁ phase, cells (2–3 × 10⁶) were incubated in 37 °C for 48 h in DMEM growth medium with no FCS.

Caspase Activation Assay.

HeLa cells (1 × 10⁷) were transfected with 15 μg of a specific plasmid. Forty-eight hours later, cells were collected, resuspended in 1 mL of 1× lysis buffer (50 mM Hepes, pH 7.4/5 mM CHAPS/5 mM DTT), and incubated on ice for 30 min. Control cells were treated with 1 μg/mL of staurosporine for 6 h before harvesting. Reactions mixtures (650 μL) containing 300 μL of cell lysate, 20 μM caspase substrate (Ac-DEVD-AMC; Calbiochem) and 65 μL of 10× reaction buffer (200 mM Hepes, pH 7.4/1% CHAPS/50 mM DTT/20 mM EDTA) were incubated at 37 °C for 90 min. The release of AMC was measured in triplicate samples using a fluorescence-plate reader (InfiniteM20; Tecan), with an excitation wavelength of 360 nM and an emission wavelength of 460 nM.