HRSL3 directly binds PR65α, the regulatory subunit A of PP2A, and its N-terminal proline-rich region plays a major role in this interaction

To examine the mechanism by which HRSL3 suppresses cellular growth and induces apoptosis, we accomplished a search for its potential interacting proteins using a LexA-based yeast two-hybrid system. First, we generated a fusion construct containing the DNA-binding domain of the LexA protein and a truncated HRSL3 protein (HREV1dC), lacking 29 amino acids, which comprise its membrane-binding domain. The truncated form of HRSL3 was used because in a library screen with the full-length protein none of the clones survived, suggesting that the membrane-binding domain inhibits the transport of the recombinant protein into the nucleus. Using the truncated HREV1dC fusion construct, we screened a human kidney cDNA library, and successfully isolated several positive clones. One of the clones contained the PPP2R1A gene, encoding PR65α the regulatory subunit Aα of PP2A, which is a major serine/threonine phosphatase in eukaryotic cells (Millward et al., 1999). To confirm the interaction between the truncated HRSL3 (HREV1dC) and PR65α, we first performed co-immunoprecipitation of HA-fused HREV1dC (HREV1dCHA) and V5-fused PR65α (PR65αV5) in COS-7 cells, transiently transfected with the corresponding expression plasmids. Protein complexes were immunoprecipitated either with an anti-V5 or with an anti-HA antibody. Using the anti-V5 antibody for immunoprecipitation, we observed the HREV1dCHA protein only in the presence of PR65αV5 (Fig. 1A, left panel, top). As a control, western blotting with an anti-V5 antibody was used (Fig. 1A, left panel, bottom). The reverse experiment, precipitation with an anti-HA antibody, again confirmed the interaction of PR65αV5 with the HREV1dCHA (Fig. 1A, right panel). These results demonstrated that overexpressed HREV1dCHA and PR65αV5 form a complex in mammalian cells. In a GST pull-down assay we used a GST-HREV1dC recombinant construct containing the truncated HRSL3 protein fused to glutathione and coupled to glutathione-Sepharose. With this approach, we were able to recover PR65α protein from cell lysates thus confirming a direct interaction between the HRSL3 and PR65α proteins (Fig. 1B).

Next, we aimed to determine the domains of HRSL3 governing the interaction with PR65α. For that purpose we designed HRSL3 mutants defective in the conserved domains typical for H-REV107-like proteins (Fig. 1C). The mutant AWAIdCHA, carries a substitution of His23 to Ala within the HWAIY domain. We also modified the NCE domain and generated the NSEdCHA mutant, which involved a one-amino acid exchange of Cys112 to Ser. Additionally, we constructed the dNdCHA mutant, which carries a deletion of 10 N-terminal amino acids including the proline-rich region (Fig. 1C). The HREV1dCHA, AWAIdCHA, NSEdCHA and dNdCHA expression vectors were transiently cotransfected with PR65αV5 into COS-7 cells. Protein interactions were assessed by immunoprecipitation with an anti-V5 antibody. We detected similar amounts of the HREV1dCHA, the AWAIdCHA, and the NSEdCHA proteins in the PR65αV5 immunocomplex. However, the dNdCHA deletion mutant, lacking 10 N-terminal amino acids, did not interact with PR65αV5 (Fig. 1D, left panel, top). Re-probing of this immunoblot with an anti-V5 antibody demonstrated efficient immunoprecipitation of PR65αV5 in all samples (Fig. 1D, left panel, bottom). The reverse experiment, immunoprecipitation with an anti-HA antibody confirmed these results (Fig. 1D, right panel, top). Thus, deletion of the N-terminal region of HRSL3 abrogated its binding to PR65αV5, indicating that this region plays a major role in the interaction between HRSL3 and the regulatory subunit A of PP2A (PR65α).

The endogenous regulatory (PR65) but not the catalytic subunit (PR36) of PP2A is an interaction partner of HRSL3

The interaction between HRSL3 and PR65α suggested that HRSL3 might be a novel component of the PP2A complex. To prove this hypothesis, we aimed to examine the interaction between endogenous HRSL3 and PR65α and tested a number of cell lines for expression of the HRSL3 and the PR65α proteins (Fig. 2A). The regulatory subunit PR65 tightly associates with the catalytic subunit PR36 forming a “core dimer” (Millward et al., 1999), therefore, we also analyzed expression of PR36α. PR65α and PR36α were equally expressed in all cell lines analyzed, however, consistent with our earlier observations (Husmann et al., 1998; Sers et al., 2002), no HRSL3 protein was detected in several rat and human tumor cell lines and in COS-7 cells (Fig. 2A). The absence of the HRSL3 protein in cultured cells excluded investigation of an interaction between HRSL3 and PR65 under physiological conditions. Therefore, for further experiments we used OVCAR-3 cells, an ovarian carcinoma cell line, in which we had demonstrated previously a suppressive function of HRSL3 (Sers et al., 2002) upon transient overexpression of the HRSL3 protein.

We isolated the endogenous PR65α immunocomplex from OVCAR-3 cells using a PR65α-specific antibody bound to an AminoLink-activated coupling gel (Pierce Biotechnology Inc. IL). As an isotype control, we applied the same protein extracts to mouse IgG2a (Fig. 2B). Analysis of the recovered protein complexes revealed the presence of the endogenous regulatory subunit PR65α and the overexpressed HRSL3 protein (Fig. 2B), clearly demonstrating interaction between HRSL3 and endogenous PP2A.

For further evaluation of HRSL3 binding to the endogenous PP2A complex in OVCAR-3 cells, we performed co-immunoprecipitation using transfected full-length HRSL3 (HREV1FL) and the interaction-deficient mutant (dNdC) and tested their binding to the regulatory subunit PR65α (Fig. 2C, left panel) and to the catalytic subunit PR36α of PP2A (Fig. 2C, right panel). Whereas binding of full-length HRSL3 to PR65α was confirmed (Fig. 2C, left panel, bottom), neither wild-type HRSL3 nor the interaction deficient mutant were detected in the PR36 protein complex (Fig. 2C, right panel, bottom). In addition, we examined whether the presence of the HRSL3 protein might influence the amount of the catalytic subunit PR36 and the regulatory subunit B56 precipitated within the PR65α immunocomplex. No difference in the relative amounts of the regulatory B subunit, B56 were observed in immunocomplexes containing either wild-type HRSL3 or its interaction-deficient mutant dNdC (Fig. 2C, left panel, middle). By contrast, a considerably lower amount of the catalytic subunit PR36 was detected in the presence of full-length HRSL3 in the PR65 immunocomplex as compared to the control (Fig. 2C, left panel, asterisk). These results indicate that HRSL3 binds specifically to the regulatory, but not to the catalytic subunit of PP2A and suggest that the interaction between PR65 and HRSL3 might displace PR36 but not B56 from the HRSL3-PR65α complex.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

The N-terminus of the HRSL3 protein (HREV1) is important for the binding of HRSL3 and PR65. (A) Left panel: proteins were immunoprecipitated with an anti-V5 antibody, and subsequently detected with anti-HA (top) and anti-V5 (bottom) antibodies. Lanes (from the left): 1, total cell lysate; 2, immunoprecipitation from cellular extracts harboring both proteins; 3, 4, immunoprecipitation from cellular extracts harboring PR65αV5 or HREV1dCHA, respectively. Right panel: proteins were immunoprecipitated with an anti-HA antibody and subsequently detected with anti-V5 (top) antibody. The efficiency of the immunoprecipitation was controlled using an anti-HA antibody (bottom). Lanes (from the left): 1, immunoprecipitation from cellular extracts harboring both proteins; 2, 3, immunoprecipitation from cellular extracts harboring PR65αV5 or HREV1dCHA, respectively. (B) To prove whether PR65 directly interacts with HRSL3, cleared COS-7 cell lysates were incubated with either GST or GST-HREV1dC proteins coupled to GST-Sepharose. Relative amounts of the proteins were analyzed using western blotting against PR65 and GST antibodies. (C) The HREV1FL expression construct carries wild-type full-length HRASL3 cDNA. Four domains of the HRSL3 protein are indicated: a proline-rich domain at the N terminus of the protein (PD), a HWAIY domain containing a conserved histidine (His23), an NCE domain containing a conserved cystein (Cys112), and a membrane binding domain (MBD). In the HREV1dCHA expression construct, 29 C-terminal amino acids are substituted by the HA-epitope. AWAIdCHA carries a mutation His23→Ala; NSEdCHA carries a mutation Cys112→Ser; dNdCHA harbors a deletion of the N-terminal proline-rich region. (D) COS-7 cells were cotransfected with the following expression plasmids: PR65αV5 and HREV1dCHA, PR65αV5 and HREV1-AWAIYdCHA, PR65αV5 and HREV1-NSEdCHA, and PR65αV5 and HREV1-dNdCHA. As negative controls, plasmids containing HA or V5 epitopes only were transfected (–). (Left panel) Top: proteins were immunoprecipitated with an anti-V5 antibody and subsequently analyzed with an anti-HA antibody. HREV1dCHA, HREV1-AWAIdCHA and HREV1-NSEdCHA associate with PR65αV5, whereas the HREV1-dNdCHA deletion mutant is unable to interact with PR65αV5. Bottom: incubation with an anti-V5 antibody demonstrates equal amounts of the immunoprecipitated PR65αV5 protein. (Right panel) Top: proteins were immunoprecipitated with an anti-HA antibody and subsequently analyzed with an anti-V5 antibody. The PR65αV5 protein was obtained in a complex with HREV1dCHA, HREV1-AWAIdCHA and HREV1-NSEdCHA. By contrast, no PR65αV5 was detected in the immunoprecipitated HREV1-dNdC complex. Bottom: immunoprecipitation was controlled using an anti-HA antibody and demonstrates equal amounts of the HA-tagged HRSL3 proteins.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

HRSL3 (HREV1) interacts with the endogenous regulatory, but not the catalytic, subunits of PP2A. (A) Protein extracts from several cell lines were analyzed by immunoblotting against an HRSL3-specific antibody described previously as H-REV107-1. As a control, COS-7 cells transiently transfected with an HRASL3 expression plasmid, were used. PR36α and PR65α were detected with appropriate antibodies. Actin was used as a loading control. (B) The PR65α immunocomplex was recovered from OVCAR-3 cells transiently transfected with an HRASL3 expression plasmid (HREV1FL). Immunoprecipitated proteins (IP), and protein extracts used for the immunoprecipitation (Input) were subjected to SDS-PAGE and western blot analysis. Presence of the regulatory PR65α (top) and the catalytic PR36α (middle) subunits was demonstrated by immunoblotting with appropriate antibodies. Application of an anti-HRSL3 specific antibody (bottom) revealed the presence of the HRSL3 protein in the PR65α protein complex (second lane) but not in the control immunoprecipitation (third lane), confirming the interaction between HRSL3 and endogenous PR65α. (C) OVCAR-3 cells were transiently transfected with HREV1FL, dNdC and pcDNA3 as a control. Forty-eight hours post-transfection immunoprecipitation was performed with a PR65α-specific antibody (left panel) and with a PR36α,β-specific antibody (right panel). Presence of the HRSL3 full-length and truncated proteins, PR65α,β, PR36α,β and B56 in the protein extracts used for immunoprecipitation (Input) and in the precipitated protein complexes was analyzed using western blotting with the appropriate antibodies. ^*Decreased amount of the PR36 proteins in the PR65 protein complex in the presence of HRSL3 was semiquantitatively evaluated (see bar chart below) using ImageJ freeware. The ratio between the amounts of immunoprecipitated PR36 and the light chain of the PR65 antibody is shown on top of the bars. (D) OVCAR-3 cells were transiently transfected with HREV1FL-V5 and incubated with PR65, and V5-specific antibodies. Then the cells were stained with DAPI to visualize nuclei (blue), with Alexa Fluor 488 and Alexa Fluor 594 secondary antibodies to visualize the distribution of the PR65α (red) and the HRSL3-V5 (green) proteins, respectively using confocal microscopy. (Merge) Yellow color indicates a colocalization of the HRSL3 and the PR65α proteins (overlay of red and green) in the cytoplasm. (E) OVCAR-3 cells were transiently transfected with HREV1FL and incubated with PR36, and HRSL3-specific antibodies. Then the cells were stained with DAPI to visualize nuclei (blue), with Alexa Fluor 488 and Alexa Fluor 594 secondary antibodies to visualize distribution of the PR36 (green) and the HRSL3 (red) proteins, respectively using confocal microscopy. (Merge) Absence of yellow color indicates no colocalization of the HRSL3 and the PR36 protein in the cells.

We also examined the intracellular distribution of the regulatory (PR65α) and the catalytic (PR36) subunits of PP2A and investigated a potential colocalization with the overexpressed full-length HRSL3 protein (HREV1FL or HREV1FL-V5). In OVCAR-3 cells, PR65α and HRSL3 were preferably localized in the cytoplasm where both proteins are co-distributed (Fig. 2D, merge). Consistently with our previous observations (Sers et al., 2002), HRSL3 overexpression correlated with the development of an apoptotic morphology of the nuclei (Fig. 2D, DAPI). Interestingly, we could not detect a co-distribution of HRSL3 and PR36 (Fig. 2E, merge), further supporting our co-immunoprecipitation data, which suggest that the HRSL3-PR65α complex might not contain PR36.

The N-terminal domain of HRSL3 is required for inhibition of the PP2A catalytic activity

Since HRSL3 interacts with the PP2A regulatory subunit but not with the catalytic subunit, we tested if the catalytic activity of intracellular PP2A would change in the presence of HRSL3 (Fig. 3A). We overexpressed full-length HRSL3 and the interaction-deficient mutant dNdC in OVCAR-3 cells. As controls, we analyzed cells transfected with pcDNA3, and cells treated with the natural PP2A inhibitor okadaic acid (OA) (Holmes et al., 1990). To achieve effective inhibition of PP2A in cell culture, we used 10 nM OA (Fig. 4A) (Favre et al., 1997). At this concentration OA also inhibits several other members of the serine/threonine protein phosphatase family (Bastians and Ponstingl, 1996; Brewis et al., 1993; Cohen et al., 1996; Sasaki et al., 1994), therefore we measured phosphatase activity of the immunoprecipitated PR65 protein complexes (Fig. 3A). In parallel, we controlled the amounts of the precipitated phosphatase and HRSL3 (Fig. 3B). As shown in Fig. 3A, the presence of HRSL3 led to a 40% inhibition of PP2A activity, whereas the interaction-deficient mutant dNdC did not inhibit PP2A, suggesting that the interaction between HRSL3 and PR65α might be important for the diminished catalytic activity of PP2A.

The observed reduction of the PR36 catalytic subunit within the PP2A complex after HRSL3 transfection might be one reason for a lower PP2A activity. However, other mechanisms might contribute to an HRSL3-mediated suppression of PP2A activity. For instance, the intracellular localization of PP2A plays a role in the capacity of PP2A to dephosphorylate particular substrates (Janssens et al., 2005; Janssens and Goris, 2001; Zuluaga et al., 2007). To investigate this possibility, we examined the subcellular distribution of the PP2A regulatory and catalytic subunits in OVCAR-3 cells either treated with OA or transfected with wild-type HRSL3 and the interaction-deficient mutant dNdC. Only moderate inconsistent changes in the protein levels were observed in different subcellular fractions after treatment or HRSL3 overexpression (see Fig. S1 in supplementary material), suggesting that HRSL3 regulates PP2A activity by other mechanisms.

Suppression of PP2A activity by HRSL3 is sufficient to induce apoptosis in ovarian carcinoma cells

In our previous investigation, we had detected loss of HRSL3 from human ovarian carcinomas and demonstrated that overexpression of the HRSL3 protein induced apoptosis in ovarian cancer cells (Sers et al., 2002). Because only expression of wild-type HRSL3 was able to interfere with PP2A activity, we now addressed a potential role of the HRSL3-PR65α interaction in cell death. OVCAR-3 cells were transfected with wild-type HRSL3 and with the N-terminal deletion construct, then caspase cleavage and the DNA content were analyzed (Fig. 4A,B). Cleavage of caspase-3 was observed 48 hours after transfection with wild-type HRSL3 and could be completely inhibited in cells pretreated with the caspase inhibitor zVAD-fmk (Fig. 4A). The fraction of cells showing a subG1-characteristic DNA content was increased after transfection with HRSL3 by approximately 23%. The N-terminal deletion construct did not induce DNA fragmentation above background, as determined by the empty vector pcDNA3 (Fig. 4B). The pan-caspase inhibitor zVAD-fmk abrogated DNA fragmentation induced by HRSL3, confirming an involvement of caspases in this process (Fig. 4B, HREV1FL + zVAD). These observations indicated that inhibition of PP2A by the tumor suppressor HRSL3 might be the crucial event in the regulation of cell survival in ovarian carcinomas. To test whether inhibition of the phosphatase alone would induce apoptosis in ovarian carcinoma cells, we treated OVCAR-3 cells with 10 nM OA and measured the fraction of cells with a subG1 DNA content (Fig. 4B, right panel). In concert with published data showing a survival role of PP2A (Li et al., 2002; MacKeigan et al., 2005), after OA application more than 80% of the cells displayed a subG1 DNA content, which was reduced to 30% when the caspase-inhibitor zVAD-fmk was added (Fig. 4B).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

HRSL3 suppresses the PP2A catalytic activity. (A) OVCAR-3 cells were transiently transfected with HREV1FL and HREV1-dNdC, transfection with pcDNA3 was used as a negative control, and treatment with 10 nM okadaic acid as a positive control. Forty-two hours after transfection cells were lysed and immunoprecipitated with an anti-PR65α antibody. Phosphatase activity of the immunoprecipitates was measured as the release of free phosphate (pmol/minute; y axis). Six independent experiments were performed, and included into a statistical analysis. Significance of the obtained results was assessed using an F-test. The difference between HREV1FL and dNdC, and HREV1FL and pcDNA3 values was considered as significant (P=0.02). (B) Immunoprecipitates used for the measurement of the phosphatase activity in the presence of HRSL3 and its interaction-deficient mutant were controlled by western blotting using HRSL3- and PR65-specific antibodies.

In order to prove a direct role of the HRSL3 interaction partner PR65α in preventing apoptosis in OVCAR-3 cells, we suppressed the PR65α subunit via RNAi and subsequently investigated caspase-3 cleavage (Fig. 4C, upper panel). As a control, we also performed a knock-down of PR65β, the efficiency of which was measured by RT-PCR (Fig. 4C, bottom panel). Cleaved caspase-3 was detected 48 hours after transfection of PR65α-specific siRNAs, whereas siRNA-mediated inhibition of the PR65β isoform induced a weaker activation of caspase-3. The caspase inhibitor zVAD-fmk prevented caspase cleavage during the siRNA experiments (Fig. 4C, upper panel), indicating that inhibition of PP2A, whether mediated through HRSL3 or via siRNA is sufficient to activate the caspase cascade in OVCAR-3 cells. Further measurement of the DNA content in siRNA-treated cells demonstrated that the PR65α-specific RNAi induced DNA fragmentation in approximately 49% of the cells, 29% were affected by inhibition of PR65β (Fig. 4D, upper panel), whereas a strong induction of DNA fragmentation (83%) was achieved after siRNA-dependent downregulation of both isoforms. Transfection with scrambled control siRNA induced DNA fragmentation in ∼24% of the cells (Fig. 4D, upper panel). Concomitant application of the pan-caspase inhibitor zVAD in each case strongly reduced the fraction of cells exhibiting a subG1-characteristic DNA content (Fig. 4D, middle panel). These results further demonstrate that both individual downregulation of PR65α and HRSL3 overexpression are sufficient to induce caspase-dependent apoptosis in ovarian carcinoma cells. Taking into account the observed decrease of the PR36 protein in the PR65 immunocomplex in the presence of HRSL3, we would speculate that HRSL3 might disrupt the PR65-PR36 protein complex thereby inhibiting PP2A activity and inducing apoptosis.

Interaction of HRSL3 with PR65α does not influence stability of PP2A

It has previously been demonstrated that ablation of either the regulatory or the catalytic subunits induces degradation of the whole PP2A enzyme (Li et al., 2002; Silverstein et al., 2002; Strack et al., 2004). Assuming that HRSL3 acts via a disruption of the PR65-PR36 interaction, the level of intracellular PR36 should be decreased in the presence of HRSL3. To prove this hypothesis, we analyzed the amounts of the PR36 and PR65 proteins in OVCAR-3 cells after PR65- and PR36-specific RNAi and after transfection with HRSL3. As shown in Fig. 5A, PR65α,β-specific RNAi leads to a reduction in the PR36 protein level. In the control experiment, PR36-specific RNAi leads to a decrease in the relative amount of the PR65 protein. By contrast, HRSL3 has no effect on the protein level of either the catalytic or the regulatory subunit of PP2A (Fig. 5B), albeit we demonstrated that the amount of PR36 in the PR65 immunocomplex decreased in the presence of HRSL3. This suggests that HRSL3 might change the composition of the PP2A protein complex, yet does not suppress PP2A activity by inducing its degradation.

Atypical protein kinase C (PKC) is potentially involved in HRSL3 signaling in ovarian carcinoma cells

To decipher the apoptotic pathway induced by HRSL3-mediated suppression of PP2A in OVCAR-3 cells in more details, we analyzed the amounts and phosphorylation of anti-apoptotic Bcl and pro-apoptotic Bad, which are known to be PP2A targets (Garcia et al., 2003; Grethe and Porn-Ares, 2005). Western blot analysis revealed that neither HRSL3, nor administration of OA can influence phosphorylation and relative levels of these proteins in OVCAR-3 cells (Fig. 6A).

To further define the caspases involved in apoptosis induction, we performed western blot analysis of cleaved caspases after HRSL3 transfection and OA treatment. We did not detect cleavage of the initiator caspase-8 and caspase-10 after HRSL3 transfection, although caspase-8 was slightly activated by OA (data not shown). By contrast, caspase-9 was cleaved after transfection with the HRSL3 expression vector and after OA treatment (Fig. 6A, left panel). Activation of caspase-9 resulted in the cleavage of its effector caspases-7 and 3. Transfection with the empty vector or treatment with the drug vehicle, DMSO, did not affect any of the caspases examined (Fig. 6A, left panel). This result suggests that HRSL3 induces apoptosis via a caspase-9-dependent, but caspase-8- and 10-independent pathway.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Both HRSL3-dependent PP2A inhibition and siRNA-mediated suppression of PP2A regulatory subunit A, induce apoptosis via a caspase-dependent pathway. (A) Cleaved caspase-3 was detected by immunoblotting using appropriate antibodies. Total amount of caspase-3 was used as a loading control. In all experiments HRSL3 expression was controlled by immunoblotting with the HRSL3 specific antibody. Application of the caspase inhibitor zVAD-fmk leads to the abrogation of caspase cleavage. (B) OVCAR-3 cells were electroporated with the HREV1FL, dNdC expression plasmids. Twelve hours after transfection, cells were treated with 40 μM caspase inhibitor zVAD-fmk, if indicated. Alternatively, cells were treated with OA (10 nM), OA with zVAD-fmk (40 μM), or vehicle DMSO only. Seventy-two hours after transfection or treatment, nuclear DNA content was measured by flow cytometry. The relative number of cells displaying an apoptotic, sub-G1 content, is given between the marker bars. DNA histograms are representative of two independent experiments. (C) OVCAR-3 cells were transfected with PR65α- and PR65β-specific siRNAs individually, together, and treated with caspase inhibitor zVAD-fmk (40 μM), if indicated. The level of the PR65α protein and cleavage of caspase-3 were controlled 48 hours after transfection (upper panel). Efficiency of the PR65-specific RNA interference was verified using RT-PCR 12 and 24 hours after transfection. As a control, GAPDH mRNA was amplified (lower panel). (D) OVCAR-3 cells were transfected with PR65α- and PR65β-specific siRNAs individually or in combination. Apoptosis was analyzed at the single-cell level by measuring the nuclear DNA content. The relative number of cells displaying a sub-G1 DNA content is given between the marker bars. DNA histograms are representative of three independent experiments.

Next, we analyzed activation and phosphorylation of RAF-MAPK, PI 3-kinase, and Wnt pathways through HRSL3 and OA in OVCAR-3 cells, because PP2A is known to be involved in the regulation of these signaling cascades (Alessi et al., 1995; Chen et al., 2005; Gao et al., 2002; Gotz et al., 2000; Ikeda et al., 2000; Li et al., 2003; Li et al., 2001; Ratcliffe et al., 2000; Ricciarelli and Azzi, 1998; Seeling et al., 1999; Sim and Scott, 1999; Yamamoto et al., 2001; Yang et al., 2003; Zhou et al., 2002; Zuluaga et al., 2007). Treatment with OA led to the activation of ERK1,2, AKT (PKB), and β-catenin (Fig. 6A, right panel), whereas neither RAF no PDK seemed to be regulated by OA-dependent phosphatases. In contrast to OA, HRSL3 did not affect phosphorylation of ERK and AKT (PKB), which is consistent with our previous observation that HRSL3 only partially inhibits PP2A catalytic activity (Fig. 3A).

We asked if an HRSL3-independent partial inhibition of PP2A might be sufficient to induce apoptosis in ovarian carcinoma cells. Since the IC₅₀ of OA varies depending on the substrate employed and on the amount of the enzyme, both under cell-free conditions and in cellular systems (Boudreau and Hoskin, 2005; Favre et al., 1997; McCluskey et al., 2002; Schonthal, 1998), we empirically defined the OA concentrations required for apoptosis in OVCAR-3 cells. 0.5 nM OA was sufficient to induce caspase cleavage, with an efficiency similar to that of HRSL3 expression (Fig. 6B, right panel), confirming that partial inhibition of PP2A is sufficient to induce the apoptotic cascade in ovarian carcinoma cells. In addition, treatment of the cells with 1 nM OA did not significantly influence phosphorylation of ERK1/2 and MEK1/2 (Fig. 6B, left panel), suggesting that at this concentration, the cells maintain a pool of active PP2A which dephosphorylates the kinases analyzed.

Most interestingly, a clear increase in the phosphorylation of the atypical PKCζ was detected 48 hours post-transfection with full-length HRSL3, but not with the empty vector or with the vehicle (Fig. 6C). Thus PKCζ is a PP2A target, whose phosphorylation is regulated by both OA and HRSL3, suggesting a potential implication of PKC in HRSL3-dependent signaling. To prove whether PKCζ is involved in HRSL3-dependent apoptosis, we utilized a myristilated pseudosubstrate peptide, which has been shown to specifically inhibit PKCζ (Standaert et al., 1999). Cells, pre-treated with different concentrations of the peptide, were transfected with the HRSL3 expression plasmid, and then caspase-3 cleavage was analyzed. As shown in Fig. 6D, PKCζ inactivation abrogated HRSL3-induced apoptosis to the background level, suggesting that PKCζ activity might be critical for the pro-apoptotic HRSL3 function in ovarian cancer cells.

Tumor suppressing function of the H-REV107-like proteins

HRSL3 belongs to the H-REV107-like family of tumor suppressors (Hughes and Stanway, 2000). Although the biochemical properties of these proteins are still elusive, there is ample evidence that H-REV107-like proteins are involved in the regulation of cell proliferation, differentiation and apoptosis. HRSL3 was found to suppress growth of rodent and human carcinoma cells in vitro and in vivo and is downregulated in breast, ovary and kidney carcinomas (Husmann et al., 1998; Sers et al., 1997; Sers et al., 2002). The RARRES3 protein is suppressed in many human tumors such as ovarian (Lotz et al., 2005), gastric (Huang et al., 2000), colorectal (Huang et al., 2000; Shyu et al., 2003) and head and neck tumors (Higuchi et al., 2003), in psoriatic lesions and in basal cell carcinomas (Duvic et al., 2000; Duvic et al., 2003). It can activate type I transglutaminase by an unknown mechanism thereby inducing terminal differentiation (Sturniolo et al., 2003). In cervical cancer cells, RARRES3 suppresses the activation of RAS proteins (Tsai et al., 2006) and negatively regulates the activity of Jun and p38 MAPKs (Huang et al., 2002). Interestingly, both HRASL3 and RARRES3 were identified as downstream targets of IFNγ (Sers et al., 2002) and are suggested to regulate ovarian carcinoma cell growth by a concerted action (Lotz et al., 2005). However, whereas RARRES3 has been identified as a suppressor of RAS and MAPK activation (Tsai et al., 2006), HRSL3 negatively affects PP2A, thereby specifically targeting PKC, independent of MAPK and PI 3-kinase activation. This indicates that the concerted loss of HRSL3 and RARRES3 from advanced ovarian carcinomas (Lotz et al., 2005; Sers et al., 2002) might allow a sustained activation of several distinct branches of the intracellular oncogenic signaling network.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 6.

Expression of the full-length HRSL3 protein (HREVFL) was controlled by incubation with the HRSL3-specific antibody (H-REV107-1). HRSL3 regulates PKCζ, which is required for HRSL3-induced apoptosis. (A) Phosphorylation of the indicated signaling proteins was analyzed by immunoblotting using antibodies specific for their phosphorylated forms. Immunoblotting against total amount of the proteins was used as a loading control. Expression of the HRSL3 protein, full-length and truncated form, was controlled by incubation with the HRSL3-specific antibody. (B) OVCAR-3 cells were either treated with 1, 10 and 20 nM of OA and with an inhibitor of the PI 3-kinase (LY294002; LY), or transfected with HRSL3. To determine the concentration of OA that inhibits PP2A activity only partially, phosphorylation of MEK1/2 and ERK1/2 kinases was examined using the corresponding antibodies (left panel). Then caspase-9 cleavage was analyzed (right panel). Actin- and caspase-9-specific antibodies were used as loading controls. (C) Phosphorylation of PKCζ is induced by treatment with both OA and HRSL3. Cells were treated with 10 nM OA, with DMSO or transfected with HRSL3 or the empty vector. Phospho-PKCζ or total PKCζ were measured by western blotting using the appropriate antibodies. (D) Twelve hours prior to transfection, OVCAR-3 cells were treated with the myristoylated PKCζ-specific pseudosubstrate. Then the cells were transfected with HRASL3 expression plasmid and caspase cleavage was analyzed 48 hours after transfection.

In a recent publication, Anantharaman and Aravind proposed that H-REV107-like proteins belong to the NlpC/P60 protein superfamily, consisting largely of prokaryotic members, most of which act as peptidases. Lecithin retinol acyltransferase (LRAT), an enzyme responsible for the esterification of retinol to retinyl ester (Ruiz et al., 1999), was suggested to be the founder of the eukaryotic members of this superfamily (Anantharaman and Aravind, 2003). Similar to H-REV107-like proteins, LRAT expression is reduced in several human tumors and in corresponding cell lines (Andreola et al., 2000; Boorjian et al., 2004; Guo et al., 2000; Guo et al., 2001; Guo et al., 2002; Guo and Gudas, 1998; Jurukovski and Simon, 1999; Simmons et al., 2002; Zhan et al., 2003). In contrast to the H-REV107 proteins and most of the prokaryotic related proteins, LRAT has a defined biochemical function and the active motifs within the protein have been characterized (Mondal et al., 2000; Mondal et al., 2001; Xue and Rando, 2004). Mutation of the critical residue Cys112, which is conserved between LRAT and HRSL3, had no influence on binding and suppression of PP2A by HRSL3. This suggests that the HRSL3 protein is not likely to exert an acyl-transferase function, and also leaves open to what extend the conservation of distinct domains between these proteins reflects a functional similarity. Most importantly, LRAT has been found recently also to harbor a palmitoylation function, indicating that this protein family might have a number of different biochemical properties (Xue et al., 2006). Since PP2A activity is controlled by post-translational modifications such as phosphorylation and methylation, it is an attractive hypothesis that HRSL3 might exert a protein modifying function (Bryant et al., 1999; Janssens and Goris, 2001; Turowski et al., 1995; Wu et al., 2000), however, currently, no data supporting a possible enzymatic activity of H-REV107 proteins are available.

A role for HRSL3 in controlling the PR65 protein complex

We detected the ectopically expressed HRSL3 protein in a complex with the endogenous PR65α protein in OVCAR-3 cells and could prove in vitro interaction between HRSL3 and PP2A. More specifically, we showed that the amount of the catalytic subunit PR36α recovered from the PR65 protein complex is considerably reduced in the presence of HRSL3. Additionally, PR36 and HRSL3, despite their predominant cytoplasmic distribution, were not colocalized within the cell. This suggests, but does not prove, that HRSL3 might somehow substitute for PR36 in a distinct pool of PP2A holoenzymes. Additionally, the absence of PR36 protein degradation following HRSL3 overexpression suggests that upon depletion of PR36 from the PR65 protein complex, PR36 binds to other interaction partners such as alpha-4 or cyclin G2 known to directly interact with PR36 thereby protecting it from degradation (Bennin et al., 2002; Chen et al., 1998; Chung et al., 1999; Prickett and Brautigan, 2004).

Considering the complexity of multiple PP2A holoenzymes coexisting within a cell (Janssens and Goris, 2001; Sontag, 2001), HRSL3 is likely to interact with and to inhibit the activity of only a specific subset of PP2A holoenzymes. It is well known that different PP2A holoenzymes simultaneously interact with distinct targets and regulatory proteins. The composition of the PP2A heterotrimers determines their localization and their functional specificity. For instance, the PP2AB56C phosphatase interacts with the β-catenin degradation complex and inhibits Wnt-signaling in Drosophila (Li et al., 2001), whereas PP2AB56ϵC positively regulates the Wnt pathway during early Xenopus embryogenesis (Yang et al., 2003). When we addressed the activation status of several distinct signaling cascades downstream of the HRSL3 protein, we could show that inhibition of PP2A activity via HRSL3 activates apoptosis in a caspase-9-dependent manner. PP2A inhibition via OA or PP2A-specific RNAi also led to cell death, pointing to a role of the phosphatase in survival of ovarian carcinomas. This observation is supported by many reports showing that PP2A acts as one of the key regulators of cell proliferation and malignant transformation in other cell systems (Abraham et al., 2000; Chen et al., 2004; Chen et al., 2005; Jaumot and Hancock, 2001; Li et al., 2002).

Analysis of the PP2A substrates revealed PKCζ, shown earlier to be activated by OA in rat adipocytes, (Standaert et al., 1999), as a potential HRSL3 target in ovarian carcinoma cells. Importantly, treatment of cells with the PKCζ-specific pseudosubstrate abolished HRSL3-dependent caspase cleavage, supporting the possibility that PKCζ activity might be of functional relevance for apoptosis induction. The first evidence of a cross-talk between PP2A and PKCζ was provided by Sontag et al., who demonstrated that during an SV40 infection, inhibition of PP2A is essential for the activation of PKCζ and NF-κB signaling pathways in monkey kidney CV-1 cells (Sontag et al., 1997). Furthermore, PP2A was shown to regulate atypical PKCs during tight junction formation in epithelial cells (Nunbhakdi-Craig et al., 2002). Recently, the atypical PKCζ and the novel PKCθ were identified to be upregulated in ovarian carcinoma; PKCζ was suggested as a potential biomarker for aggressive disease (Zhang et al., 2006). Our current data suggest an additional role for PKCζ-dependent signal transduction in the progression of ovarian carcinoma and implicate HRSL3 as a signaling regulatory molecule functionally involved in the complex regulatory network mediating growth and survival of ovarian cancer.

Yeast two-hybrid analysis

A LexA-based yeast two-hybrid screening of a human kidney cDNA library was performed according to the recommendations of the manufacturer (BD Biosciences, Clontech, CA). The EGY48 yeast strain (OriGene Technologies, Inc., MD) was consecutively transformed with a p8op-lacZ reporter plasmid (BD Biosciences, Clontech, CA), a pEG202 plasmid (Pharmacia Biotech Inc., CA) carrying the HRSL3 cDNA, and a pJG4-5 plasmid containing a library insert (OriGene Technologies Inc, MD). Positive clones were selected on the SD, Gal, Raf, –His, –Leu, –Trp, –Ura, X-gal induction medium. Plasmid DNA was isolated from all positive clones, transferred into the KC8 bacterial strain and sequenced.

Expression constructs and siRNAs

HREV1FL was described previously (Hajnal et al., 1994). HREV1dCHA was generated by PCR amplification. It harbors the first 406 bp of the HRSL3 coding region, and 28 additional basepairs encoding an artificial HA-epitope (YPYDADYA). HREV1-AWAIdCHA (His23→Ala23) and HREV1-NSEdCHA (Cys112→Ser112) were generated using the QuikChange Mutagenesis Kit (Stratagene, CA). The dNdCHA construct harbors a deletion of 30 bp at the 5′-end of the HRSL3 cDNA coding region. An expression vector containing human PR65α fused to a V5 epitope (PR65αV5) was purchased from GeneStorm Collection (Invitrogen, CA). The pcDNA3 (Invitrogen, CA) plasmid was used as a control. PR65α,β-, PR36α,β-specific siRNAs and control duplexes were purchased from the Validated Stealth™ RNAi Collection (Invitrogen, CA).

RT-PCR

RT-PCR reactions were carried out using Access RT-PCR System according to the recommendations of the supplier (Promega, WI). Primers used for the amplification of 211 bp of the PPP2R1A gene (NM_014225) were 5′-CGAACTCCGCAATGAGGACG-3′ (fw), and 5′-AGGCAGTGCACGTACTCTGG-3′ (rv), and for the amplification of 260 bp the PPP2R1B gene (NM 002716) were 5′-CGACGAGCTCCGCAATGAAG-3′ (fw) and 5′-CAACAGTCTCTTCCACAGTTGCC-3′ (rv). As a control a G3PDH-specific fragment was amplified using primers 5′-GAACGGGAAGCTTGTCATTCA-3′ (fw) and 5′-GTAGCCAAATTCGTTGTCATAC-3′ (rv).

Cell culture and transfection procedures and treatment

Human epithelial ovarian carcinoma cell lines SKOV-3, OVCAR-3, CAOV-3, A27/80, the teratocarcinoma cell line PA1, the cervix adenocarcinoma cell line HeLa, the epidermoid carcinoma cell line A431, the African green monkey kidney fibroblasts COS-7, and the H-Ras transformed rat fibroblasts FE-8 (Griegel et al., 1986) were maintained in Dulbecco's modified Eagle's medium (DMEM; BioWhittaker, MD) supplemented with 10% fetal calf serum and 2 mM glutamine. DNA transfection was performed using Fugene reagent (Roche Diagnostics GmbH, Mannheim, Germany); siRNA transfection was done using Lipofectamine 2000 (Invitrogen, CA). For the analysis of the endogenous PR65α protein complex and DNA fragmentation, HRSL3 expression plasmids (HREV1FL and dNdC) were transfected into OVCAR-3 cells using an Amaxa device for electroporation (Amaxa Biosystems, Cologne, Germany) in solution R and the T16-program. For cell treatment, okadaic acid (Upstate, CA) and a myristoylated PKCzeta-specific pseudosubstrate inhibitor Myr-SIYRRGARRWRKL-OH, synthesized at the Peptide-synthesis Service Facility (Institute of Biochemistry, Charité, Berlin, Germany) were used.

GST pull-down assay

GST-HREVdC fusion protein and GST control proteins were produced in B21 strain of E. coli and purified using gluthatione-Sepharose. Subsequently, OVCAR-3 cells were lysed in NP40-lysis buffer [50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% Nonidet P40, 10% glycerol, EDTA-free protease inhibitor complete tabs (Roche Diagnostics GmbH, Mannheim, Germany)]. Clarified protein extracts were incubated for 1 hour with GST-HREVdC and GST alone coupled to glutathione-Sepharose, washed four times with lysis buffer and analyzed using SDS-PAGE for presence of the PR65 protein.

Immunoprecipitation and western blot analysis

Transfected COS-7 cells were lysed in NP40-lysis buffer. Clarified lysates were used for immunoprecipitation with either anti-V5 or anti-HA antibody (Roche Diagnostics GmbH, Mannheim, Germany). Precipitated protein complexes were washed three times with lysis buffer and analyzed by SDS-PAGE. To immunoprecipitate the endogenous PR65 and PR36 protein complex, OVCAR-3 cells were lysed in Triton lysis buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% Triton X-100, 0.5 mM PMSF, EDTA-free protease inhibitor complete tabs); clarified protein lysates were precipitated either with a PR65α-specific antibody (Covance Research Products Inc, CA) or with PR36αβ-specific antibody (Upstate Biotechnology, NY). Precipitates were washed and subjected to SDS-PAGE. Alternatively, for the isolation of the PR65α immunocomplex (Fig. 3A), the ProFound Mammalian Co-immunoprecipitation Kit was used (Pierce Biotechnology Inc., IL). The experimental procedure was performed according to the recommendations of the supplier.

To detect cleaved caspases, cells were harvested in cold 1× PBS, and lysed in Chaps buffer (50 mM Pipes-HCl pH 6.5, 2 mM EDTA, 0.1% Chaps, 20 μg/ml leupeptin, 10 μg/ml pepstatin A, μg/ml aprotinin, 5 mM DTT). To disrupt the cells, three freeze-thaw cycles were used.

For western blot analyses, proteins were separated by SDS-PAGE, and transferred to a PVDF membrane (Amersham Pharmacia, Buckinghamshire, UK). Membranes were blocked and incubated with the following primary antibodies: anti-V5 (Invitrogen, CA), anti-HA (Sigma, MS), anti-PR65α (6F9; Covance Research Products Inc, CA), anti-PR65 and anti-PR36α (BD Biosciences, Heidelberg, Germany), anti-pan-actin (Chemicon, CA) and anti-HRSL3 (Sers et al., 2002). All other antibodies were purchased from Cell Signaling Technology Inc. (MA). Then the membranes were incubated with a peroxidase-conjugated goat anti-rabbit, goat anti-rat or goat anti-mouse antisera (Dianova, Hamburg, Germany) and developed using the ECL System (Amersham Pharmacia, Buckinghamshire, UK).

Immunofluorescence analysis and confocal microscopy

For immunofluorescence analysis cells were grown on glass coverslips. After fixation with 3% paraformaldehyde, cells were permeabilized using 0.2% Triton X-100 for 1 min, then washed with PBS and incubated with the primary antibodies against HRSL3, PR65 or PR36 (BD Biosciences, Heidelberg, Germany), in 1% BSA-PBS, washed and incubated with secondary antibodies, anti-rabbit Alexa Fluor 594, Alexa Fluor 546, anti-mouse Alexa Fluor 488 (Molecular Probes, OR), or anti-rat FITC (The Jackson Laboratory, MA). For nuclear staining, cells were incubated with diamidinophenylindole (DAPI). Colocalization of HREV1FL with PR65 and PR36 was analyzed using a Leica confocal microscope TCS SL (×63 oil objective, 2× zoom).

Phosphatase assay

Phosphatase activity was measured using the Ser/Thr Phosphatase Assay Kit (Upstate Biotechnology, NY). Cells were lysed in lysis buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% Nonidet P40, 10% glycerol, EDTA-free protease inhibitor complete tabs) and equal amount of clarified lysates were precipitated with an anti-PR65α antibody and protein G-agarose (Roche, Mannheim, Germany). For each reaction, 250 μM phosphopeptide K-R-pT-I-R-R was used as a substrate. The amount of free phosphate was determined in comparison with the standard curve and calculated as pmol/minute. The results of six independent experiments were used for statistical analysis using F-tests (MS Office Excel).

Measurement of apoptotic cell death

Apoptosis was determined at the single-cell level by measuring the DNA content of individual cells. The protocol was adapted from that of Hemmati et al. (Hemmati et al., 2002). Briefly, OVCAR-3 cells were trypsinized, collected and fixed in 2% paraformaldehyde. Cells were incubated in 70% ethanol and resuspended in PBS containing 40 mg/ml DNase-free RNase A (Roche Molecular Biochemicals, Mannheim, Germany). After incubation for 30 minutes cells were pelleted and resuspended in PBS containing 50 mg/ml propidium iodide (Sigma, Deisenhofen, Germany). Cellular DNA content was measured with a logarithmic amplification in the FL-3 channel of a FACScan flow cytometer (BD, Heidelberg, Germany) equipped with the CELL-Quest software. Data are expressed in percent hypoploidy (the percentage of cells with a sub-G1 DNA content), which reflects the percentage of apoptotic cells with fragmented genomic DNA.