A Novel Link Between REC2, a DNA Recombinase, the Retinoblastoma Protein, and Apoptosis*

The REC2 recombinase is essential for recombinational repair following DNA damage as well as for successful meiosis and gene targeting in the corn smut Ustilago maydis. Here we report that overexpression of REC2 induced apoptotic cell death in human HuH-7, Hep G2, and Hep 3B hepatoma cells. Apoptosis was related to recombinase activity and was significantly increased by inhibition of retinoblastoma (Rb) expression with transforming growth factor-β1. REC2-induced apoptosis was associated with a significantly reduced percentage of cells in the G1phase of the cell cycle and a significant reduction in G2/M only in theRb( − / − )Hep 3B cells. Overexpression of REC2 resulted in increased abundance of the hyperphosphorylated form of Rb. However, by immunoprecipitation REC2 was associated primarily with hypophosphorylated Rb, suggesting that REC2 may be involved in modulating the phosphorylation state of Rb. The A and B pocket domains with the spacer amino acid sequence and the carboxyl-terminal region of Rb were required for maximal binding to REC2. Overexpression of Rb significantly inhibited REC2-induced apoptosis even in the presence of transforming growth factor-β1. Taken together, these data suggest a novel interaction of Rb with the recombinase REC2 and a role for this complex in bridging DNA recombination and apoptosis.

The REC2 recombinase is essential for recombinational repair following DNA damage as well as for successful meiosis and gene targeting in the corn smut Ustilago maydis. Here we report that overexpression of REC2 induced apoptotic cell death in human HuH-7, Hep G2, and Hep 3B hepatoma cells. Apoptosis was related to recombinase activity and was significantly increased by inhibition of retinoblastoma (Rb) expression with transforming growth factor-␤1. REC2-induced apoptosis was associated with a significantly reduced percentage of cells in the G 1 phase of the cell cycle and a significant reduction in G 2 /M only in the Rb( ؊ / ؊ ) Hep 3B cells. Overexpression of REC2 resulted in increased abundance of the hyperphosphorylated form of Rb. However, by immunoprecipitation REC2 was associated primarily with hypophosphorylated Rb, suggesting that REC2 may be involved in modulating the phosphorylation state of Rb. The A and B pocket domains with the spacer amino acid sequence and the carboxyl-terminal region of Rb were required for maximal binding to REC2. Overexpression of Rb significantly inhibited REC2-induced apoptosis even in the presence of transforming growth factor-␤1. Taken together, these data suggest a novel interaction of Rb with the recombinase REC2 and a role for this complex in bridging DNA recombination and apoptosis.
The retinoblastoma tumor suppressor gene product (Rb) is a nuclear phosphoprotein that is involved in regulating progression through the cell cycle (1). It is hypophosphorylated in the quiescent state and early G 1 where it blocks cell cycle progression. Phosphorylation of Rb is initiated during mid to late G 1 , and it becomes fully phosphorylated prior to the G 1 -S phase transition. Moreover, the phosphorylation status of Rb is also involved in modulating its ability to associate and/or bind with other cellular and viral factors. In fact, the number of partner proteins that interact with Rb has increased dramatically in recent years and includes certain viral proteins (2)(3)(4)(5), the transcription factors Myc and E2F, the proto-oncogene MDM2, members of the cyclin family, the nuclear matrix proteins lamin A and C, and proteins involved in nucleosome disruption (6). In addition to its central role in regulating the cell cycle, it has been shown that Rb is involved in the TGF-␤1-associated 1 apoptotic pathway in hepatocytes and hepatoma cells (7,8). The ability of TGF-␤1 to induce apoptosis in these cells was associated with decreased phosphorylation of Rb as well as reduced expression. Depletion of Rb by antisense oligonucleotides also resulted in cell death, whereas overexpression of the protein protected against TGF-␤1-induced apoptosis (8).
Recently, a eukaryotic recombinase REC2 was cloned from the fungus Ustilago maydis and shown to play an essential role in facilitating homologous recombination with high target specificity as well as recombinational repair following DNA damage (9 -11). Sequence analysis revealed that the REC2 gene product shares a significant region of homology with the Escherichia coli RecA protein, comprising an amino acid stretch of conserved residues that span an ATP binding domain (9). The region is essential for biological activity of RecA in recombination as well as in homologous pairing reactions performed in vitro. The requirement of functional REC2 for these biochemical pathways has facilitated the isolation of recombinase-deficient mutants, thus aiding in the characterization of the mechanisms involved in recombination and repair (9).
In higher eukaryotes, the role of DNA recombinases in cell growth and survival is largely uncharacterized. In this study we examined the effect of REC2 overexpression on the growth characteristics of several human hepatoma cell lines. Our results indicate that overexpression of REC2 is associated with an increase in apoptosis that is regulated in part by the level of recombinase activity. In addition, the response is modulated by Rb, which binds REC2 directly, thereby suggesting a role for this complex in bridging DNA repair and apoptosis.

EXPERIMENTAL PROCEDURES
Cell Culture, Transfections, and Immunohistochemistry-Cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, grown at 37°C, and transfected as described previously (8,12). Briefly, 5 g of each expression plasmid pCMV.CAT, pCMV.REC2, pCMV.REC2-10, and pCMV.REC2-5 (9), and pCMV.Rb (13) under control of the cytomegalovirus promoter was transfected into the cell lines using Lipofectin (Life Technologies, Inc.). After 36 h, 1 nM TGF-␤1 was added to the cultures, which were then maintained for an additional 30 h. Transfection efficiency was monitored by chloramphenicol acetyltransferase gene expression and fluorescence labeling of the transfected gene products. Morphological evaluation of apoptosis was carried out as described (8). A minimum of 300 -500 cells were counted for each culture, and the percentage of apoptotic cells was determined from at least three independent experiments.
For immunohistochemical analysis, the cells were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS), pH 7.4, at room temperature for 10 min. Following three washes with PBS, they were incubated with rabbit polyclonal antibody against U. maydis REC2 (2.5 * This work was supported in part by a grant from Kimeragen, Inc. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ʈ To whom correspondence should be addressed: g/ml) in PBS containing 0.3% Triton X-100 at 4°C overnight. After blocking with 5% normal goat serum, the cells were incubated with fluorescein isothiocyanate-conjugated goat anti-rabbit IgG in PBS for 1 h at room temperature. Cell nuclei were stained using Hoechst dye 33258 (1 g/ml) for 5 min. The images were acquired, and the nuclear morphology was evaluated as described previously (8). Levels of REC2 expression were quantitated by gray scale using the MetaMorph Imaging System (Universal Imaging Corp.). High and low level expressions of REC2 were arbitrarily determined based on two standard deviations of apoptotic and nonapoptotic cells. The mean gray values for high and low expression of REC2 were 9494.2 Ϯ 2831.7 and 750.9 Ϯ 244.2, respectively (p Ͻ 0.0001). REC2-5 expression by gray scale analysis was 8949.9 Ϯ 1129.8. For each determination, greater than 300 fields were examined. Statistical analysis was performed by the Mann-Whitney U test.
Stable REC2 transfectants of Hep G2 and Hep 3B cells were established using the pCMV.REC2 construct containing a neomycin resistance gene. The transfected cells were selected using G418 (800 g/ml), and the resulting clonal cells were maintained in G418 at 400 g/ml.
REC2 positive cells were identified by immunohistochemistry as outlined above, and the positive clonal lines from the Hep 3B or Hep G2 transfectants were pooled and expanded.
Immunoprecipitation and Western Blotting-Whole cell lysates were prepared and analyzed by Western blotting as described previously (7) using monoclonal anti-Rb XZ161 and polyclonal anti-REC2 antibodies. Co-immunoprecipitation using anti-Rb XZ161 monoclonal antibody or anti-REC2 polyclonal antibody was performed as outlined (14,15), and the resulting precipitates were subjected to immunoblot analysis using the same two antibodies. The proteins were detected using the ECL Chemiluminescence System (Amersham Life Science, Inc.).
In  terminal region of GST in pGEX-2T (Pharmacia Biotech Inc.) (17). The proteins were expressed in E. coli and purified by chromatography using glutathione (GSH)-Sepharose (Pharmacia). Recombinant REC2 protein was isolated from E. coli (11). The cell extracts from the REC2or control vector-transfected Hep 3B cells were freshly prepared as described previously (8). Binding assays were performed as follows: 20 g of REC2 in NETN buffer (18) was incubated with either the GSH-Sepharose beads alone, GST (20 g) ϩ GSH-Sepharose beads, or the GST-Rb proteins (ϳ20 g protein) ϩ GSH-Sepharose beads at 4°C overnight with rotation. The beads were then washed four times with NETN buffer, resuspended in denaturing buffer, and boiled for 4 min. After brief centrifugation, the supernatants were subjected to SDSpolyacrylamide gel electrophoresis. After transfer, the Western blots were incubated with polyclonal anti-REC2 antibody, and the proteins detected using the ECL Chemiluminescence System. GST-Rb (379 -928) beads were incubated with extracts derived from REC2-transfected Hep 3B cells and analyzed as described above.
Flow Cytometry-HuH-7 cells and Rb( Ϫ / Ϫ ) Hep 3B cells (19) were transfected with 5 g of either pCMV.REC2 or pCMV vector for 36 h. The cells were analyzed by flow cytometry as described previously (8), except the samples also included the detached cells.

REC2 Overexpression Induces Apoptosis in Human
Hepatoma Cells-To identify the effects of REC2 on cell growth and survival, Hep G2 human hepatoma cells were transfected with an expression construct containing the full-length REC2 gene from U. maydis under control of the cytomegalovirus promoter. Immunofluorescent labeling with anti-REC2 antibody revealed that overexpression of REC2 was associated with cell blebbing, nuclear fragmentation, and subsequent cell death by apoptosis (Fig. 1A, d-f). Apoptosis was also observed with overexpression of REC2-10 (T697A) (Fig. 1A, g-i), a recombinase-active missense mutant defective in the putative cdc2 kinase association motif (9). In contrast, similar overexpression of the recombi-nase-inactive missense mutant REC2-5 (D234A) did not induce apoptosis (Fig. 1A, j-l). The same results were observed when the wild type and the two missense REC2 mutants were overexpressed in rat primary hepatocytes (data not shown).
Previous studies indicated that the Rb gene product protected hepatocytes and human hepatoma cells from apoptosis induced by TGF-␤1 (8). We therefore investigated whether the high levels of endogenous Rb expressed in HuH-7 cells (7) could modulate REC2-induced apoptosis. We characterized cells expressing two significantly different levels of the transfected REC2 protein. HuH-7 cells with low level expression of REC2 displayed a morphologically normal nuclear appearance (Fig.  1B, a-c). In contrast, cells with higher level expression of REC2 exhibited nuclear blebbing and fragmentation similar to that observed in the transiently transfected Hep G2 cells (Fig. 1B, d-f).
Cell cycle analysis was performed on pCMV.REC2-transfected HuH-7 cells as well as on transfected Hep 3B cells, a Rb( Ϫ / Ϫ ) human hepatoma cell line (19). The results demonstrated that a "sub-G 1 " cell population of apoptotic cells (20,21) increased significantly (p Ͻ 0.001) in both REC2 transiently transfected cell lines (Table I). However, the percentage of Hep 3B cells in sub-G 1 increased 4-fold compared with a 2.7-fold increase observed in the transfected HuH-7 cells. Both cell lines exhibited significant decreases in G 1 cell populations, but the decrease was 2-fold greater in Hep 3B than HuH-7 cells. The S phase populations in both cell lines were unchanged by overexpression of REC2. Interestingly, only the Hep 3B cells showed a decrease in the G 2 /M phase cell population relative to control (p Ͻ 0.001) ( Table I). These data suggested that expression of Rb may affect the ability of REC2 to promote apoptosis as well as alter the phase of the cell cycle that exits into apoptosis.
REC2 Binds Hypophosphorylated Rb-The induction of apoptosis in HuH-7 cells has been linked to decreased Rb expression and/or inactivation (8). Thus, if overexpressed REC2 binds Rb during the cell cycle, the interaction could be a potential trigger for REC2-induced apoptosis. In fact, Western blot analyses demonstrated that stable expression of REC2 in Hep G2 cells resulted in an increase of the hyperphosphorylated inactive form of Rb from approximately 10 to 60% (Fig. 2, A and C). Meanwhile, those selected clones no longer exhibited extensive apoptosis, suggesting that compensatory changes had occurred. The up-regulation of Rb steady-state levels may, in fact, be a compensatory mechanism(s) required for survival with overex-pression of the recombinase. However, by immunoprecipitation with anti-REC2 polyclonal antibody, only the hypophosphorylated Rb species was associated with REC2 (Fig. 2B). Immunoprecipitation using the anti-Rb monoclonal antibody XZ161 also established the presence of REC2 in this complex (Fig. 2D). The results suggested that REC2 binds Rb either directly or indirectly through a bridge protein and appears to modulate its phosphorylation state promoting accumulation of the functionally inactive hyperphosphorylated Rb.
In vitro binding assays were performed (22) to confirm binding between the REC2 and Rb proteins and to identify the putative REC2 binding region in Rb. Recombinant GST-Rb fusion proteins containing different regions of Rb (Fig. 3B) were bound to GSH-Sepharose and used as an affinity reagent for recombinant REC2 protein purified from E. coli (11). The results established that REC2 bound to the GST-Rb fusion proteins (Fig. 3A) but not to the GST-Sepharose nor the beads alone. Binding required the Rb A and B pocket domains as well as the spacer sequence (lanes 1 and 3-6). The in vitro binding assay also confirmed that the REC2 present in the lysate from the transfected Hep 3B cells was capable of binding to the GST-Rb (379 -928) fusion protein containing the A and B pocket domains, spacer sequence, and carboxyl terminus (Fig. 3C).
These results indicated that REC2 binds to the regions of Rb that have been identified for binding of the viral proteins SV40 T antigen, E1A, HPV E7, and Rep A (2)(3)(4)(5), as well as the E2F proteins (13,23). However, only the E2F transcription factors require both the A and B pocket domains as well as the carboxyl-terminal region of Rb for high affinity binding. In addition, protein sequence alignment analysis demonstrated that neither the REC2 nor the E2F proteins contain the Rb-binding motif LXCXE previously identified in the viral E1A, large T, E7, and Rep A transforming protein products (2)(3)(4)(5). Therefore, REC2 and the E2F transcription factors require similar structural domains of the Rb protein for maximal binding.
TGF-␤1 Modulates Apoptosis Associated with REC2 Overexpression in HuH-7 Cells-We next determined whether the interaction between REC2 and Rb was involved in the Rbmediated apoptotic pathway. High level overexpression of REC2 in HuH-7 cells was associated with apoptosis in 24.0% of cells compared with less than 0.4% in the chloramphenicol acetyltransferase plasmid-transfected controls (Figs. 4 and 5B). In addition, overexpression of the recombinase-active REC2-10 mutant allele was also associated with significantly increased apoptosis (p Ͻ 0.001). In contrast, the recombinaseinactive REC2-5 mutant allele did not induce apoptosis. Interestingly, it also exhibited no detectable binding to Rb in Hep G2 cells (Fig. 6) despite high level overexpression of the protein.
When the REC2-overexpressing HuH-7 cells were incubated with 1 nM TGF-␤1, which inhibits Rb expression and phosphorylation (8), REC2-induced apoptosis was significantly increased (p Ͻ 0.001) in both wild type and REC2-10-transfected cells (Fig. 4). However, in the wild type REC2-transfected cells the increase in apoptosis was significantly greater (ϳ20%) than the sum of the individual inducers ( Figs. 4 and 5, A and B). These data suggested that Rb expression modulates a TGF-␤1/ REC2 apoptotic pathway(s) and that the putative association of REC2 with CDK may be involved in the induction of apoptosis. Because loss of functional Rb results in apoptosis in these cells (8), we determined whether alterations in the Rb levels would modulate the increased apoptosis observed in TGF-␤1-treated REC2-transfected cells.
Overexpression of Rb Inhibits REC2-induced Apoptosis-When co-overexpressed in HuH-7 cells, Rb inhibited TGF-␤1 as well as REC2-induced apoptosis by an average of 74.0% (p Ͻ 0.001) (Fig. 5, A and B). Interestingly, an 82% decrease (p Ͻ 0.001) in the percentage of apoptosis was observed in the cells overexpressing both REC2 and Rb that were also treated with TGF-␤1 (Fig. 5, A and B). These data coupled with that from the REC2-10 mutant imply that the interactive effect of TGF-␤1 and REC2 is likely mediated by Rb and that the puta-tive association of CDK with REC2 may be involved in modulating the phosphorylation state of Rb. The inability of the REC2-5 recombinase-deficient mutant to bind Rb and induce apoptosis in either HuH-7 cells or primary rat hepatocytes suggests that Rb potentially monitors the level of recombinase activity during the cell cycle. Cells with an abnormal increase in REC2 activity may then, in part, be regulated via apoptosis.
Our results support a novel link between DNA recombination and apoptosis in which Rb appears to be involved. Interestingly, it has recently been reported that the tumor suppressor p53 associates with Rad51, a human homolog of REC2, and suppresses spontaneous homologous recombination (24,25). It remains to be elucidated how these tumor suppressor genes balance both DNA repair and apoptosis in the control of cell growth.
FIG. 6. Lack of association of the recombinase-inactive REC2-5 mutant with Rb. The expression plasmids REC2 and REC2-5 were transfected into Hep G2 cells for 36 h as described under "Experimental Procedures." Cells were harvested, and whole cell lysates were prepared for immunoprecipitation with either anti-Rb monoclonal antibody XZ161 or polyclonal anti-REC2 antibody. The resulting precipitates were separated by SDS-polyacrylamide gel electrophoresis and subjected to Western blot analysis using anti-REC2 antibody and the ECL Chemiluminescence System.