RECQ1 Helicase Interacts with Human Mismatch Repair Factors That Regulate Genetic Recombination*[boxs]

Maintenance of the genome requires cells to duplicate, repair, and segregate their DNA precisely for proper transmission of genetic information to take place. A loss of function of gene products critical for maintaining genome integrity leads to genome instability, one of the hallmarks of cancer. A greater understanding of how cells maintain their genome has come from studies in both prokaryotic and eukaryotic organisms. These studies have revealed a variety of mechanisms involving DNA transactions performed by DNA replication, repair, and recombination proteins to preserve genomic integrity.

One class of enzymes that have been found to play a pivotal role in the genome preservation is the RecQ family of helicases (for review, see Refs. 1, 2). RecQ helicases play an important role in genome preservation through their roles in DNA repair, recombination, and replication. Of the five human RecQ helicases identified, three are associated with genetic disorders (Werner, Bloom, and Rothmund-Thomson syndromes) characterized by an elevated incidence of cancer and premature aging (3, 4). Two features commonly found in cells in which a RecQ helicase is mutated are abnormal DNA replication and elevated recombination (4). With the exception of RECQL4, which is reported to lack detectable helicase activity (5), all of the RecQ helicases characterized to date unwind duplex DNA, translocating in a 3′-5′ direction (for review, see Ref. 1). A great deal of knowledge about the biological function of RecQ helicases has been acquired through studies involving the identification of RecQ protein partners and their functional interactions. Consistent with their roles on a variety of DNA substrates that represent multiple intermediates in DNA processing, RecQ helicases interact with a number of proteins involved in DNA metabolism (for review, see Refs. 3, 4, 6). The network of RecQ protein interactions is likely to be important in pathways that recognize and process DNA structural perturbations that may otherwise lead to chromosomal rearrangements or mutations. A framework for understanding the roles of RecQ helicases in genome stability maintenance has been established by the identification and characterization of their protein interactors.

Of the five human RecQ helicases, RECQ1 (also referred to in the literature as RecQL and helicase Q1) is the smallest with a size of 649 residues (7, 8), only slightly larger in size than the founding member of the family, Escherichia coli RecQ. Although no disease has been linked to a mutation in RECQ1, the prominent roles of RecQ helicases in DNA metabolism and maintenance of genomic stability suggest that RECQ1 is likely to have an important biological function. RECQ1 was demonstrated to be a DNA-stimulated ATPase and helicase, capable of unwinding short duplex DNA substrates (9). In the presence of replication protein A (RPA),¹ the enzyme can unwind up to ∼500 bp (10). In addition to RPA, RECQ1 is reported to physically interact with topoisomerase IIIα (11) and importin-α homologs (12); however, the functional importance of RECQ1 interactions with topoisomerase IIIα or importin has not been demonstrated.

In this work, we sought to identify and characterize RECQ1-interacting partners to gain a better understanding of the physiological role of RECQ1. To accomplish this task, we utilized an antibody against a unique region of the protein to perform co-immunoprecipitation experiments. To explore the functional significance of these interactions, we examined the effects these partners display on the various enzymatic activities of the proteins. These interactions provide new insight into the role of RECQ1 in the maintenance of the genome.

MATERIALS AND METHODS

Proteins—Recombinant His-tagged RECQ1 protein was overexpressed using a baculovirus/Sf9 insect system and purified to near homogeneity as previously described (10). Recombinant full-length His-tagged EXO-1 and MSH2/6 were overexpressed using a baculovirus/insect system and purified as described elsewhere.² Recombinant MLH1/PMS2 was overexpressed using a baculovirus/insect system and purified as described elsewhere.³

Production of the RECQ1-K119A Mutant—The mutant K119A RECQ1 construct was generated with the QuikChange XL site-directed mutagenesis kit according to manufacturer's instructions (Stratagene). The pFastBac-RECQ1 vector containing the wild-type gene was used as the template DNA. The substitution (Lys→ Ala) was verified by dideoxy chain-termination sequencing of both DNA strands. The pFastBac-RECQ1-K119A construct was transformed next into DH10Bac E. coli for transposition of the gene into the baculovirus shuttle vector (bacmid) according to the Invitrogen protocol. After the transposition, the bacmid DNA was isolated and transfected into Sf9 insect cells with Cellfectin reagent (Invitrogen) to generate the recombinant baculoviral stock that was then used to infect cells for large-scale expression of the recombinant mutant RECQ1-K119A protein. The mutant RECQ1-K119A protein was purified to near homogeneity, applying the same protocol used for the wild-type protein.

Anti-RECQ1 Antibody Production—Custom polyclonal anti-RECQ1 antibody was produced as previously described (10). A 20 amino acid peptide corresponding to residues 644-662 (GNFQKKAANMLQQSGSKNT) in the C-terminal region of RECQ1 with an N-terminal cysteine conjugated to Sepharose 4B for affinity purification (Alpha Diagnostic International, San Antonio, TX) was conjugated to keyhole limpet hemocyanin carrier protein and injected into rabbit. An IgG fraction from the antiserum was purified using the peptide coupled to Sepharose 4B via a cysteine group. Affinity-purified antibody titration was checked by ELISA using free peptide.

Co-immunoprecipitation Experiments—HeLa cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum at 37 °C in 5% CO₂. Nuclear extracts were prepared from exponentially growing HeLa cells as previously described (13). For co-immunoprecipitation experiments, precleared nuclear extract (1.36 mg protein) was incubated with rabbit polyclonal anti-RECQ1 antibody (1:20) in buffer D (50 mm HEPES (pH 7.5), 100 mm KCl, 10% glycerol) for 4 h at 4 °C. The mixture was subsequently incubated with 60 μl of protein G-agarose (Roche Applied Science) at 4 °C overnight. The beads were then washed three times with buffer D supplemented with 0.1% Tween 20. Proteins were eluted by boiling in SDS sample buffer, and the eluate was resolved on 10% polyacrylamide Tris-glycine SDS gels and transferred to polyvinylidene difluoride membranes (Amersham Biosciences). The membranes were blocked with 5% nonfat dry milk in phosphate-buffered saline containing 0.1% Tween 20 and probed for RECQ1 using rabbit anti-RECQ1 polyclonal antibody (1:250), for EXO-1 using rabbit polyclonal antibody against a recombinant EXO-1 nuclease domain fragment characterized previously (1:5000, courtesy of Dr. D. M. Wilson III (NIA, National Institutes of Health)) (14), for MSH2 using mouse anti-MSH2 monoclonal antibody (1:250, BD Biosciences), for MSH6 using mouse anti-MSH6 monoclonal antibody (1: 500, BD Biosciences), and for MLH1 using mouse anti-MLH1 monoclonal antibody (1:250, BD Biosciences). Donkey-anti-rabbit IgG (1: 5000, Santa Cruz Biotechnology) or horse-anti-mouse secondary (1: 2500, Vector Laboratories Inc.) antibodies conjugated to horseradish peroxidase were subsequently added, and proteins on immunoblots were detected using ECL Plus (Amersham Biosciences). The detected endogenous proteins (RECQ1, EXO-1, MSH2, MSH6, and MLH1) from co-immunoprecipitation experiments using HeLa nuclear extracts were confirmed by co-migration with the corresponding purified full-length recombinant proteins.

ELISA for Detection of Direct RECQ1 Protein Interactions—Purified recombinant RECQ1 was diluted to a concentration of 18 nm in carbonate buffer (0.016 m Na₂CO₃, 0.034 m NaHCO₃ (pH 9.6)) and then added to the appropriate wells of a 96-well microtiter plate (50 μl/well), which was incubated at 4 °C overnight. Bovine serum albumin (BSA) was used in the coating step for control reactions. The samples were aspirated, and the wells were blocked for 2 h at 37 °C with blocking buffer (phosphate-buffered saline, 0.5% Tween 20, and 3% BSA). The procedure was repeated. Following blockage, the wells were incubated with the indicated concentrations of purified recombinant EXO-1, MLH1-PMS2, or MSH2/6 complex for 1 h at 37 °C. For ethidium bromide (EtBr) treatment, 50 μg/ml EtBr was included in the incubation with EXO-1, MLH1-PMS2, or MSH2/6 during the binding step in the corresponding wells. The samples were aspirated, and the wells were washed three times before the addition of the appropriate primary antibody (the same as used for Western blot detection) against EXO-1, MLH1-PMS2, or MSH2/6 in blocking buffer. Plates were incubated at 37 °C for 1 h. Following three washings, horseradish peroxidase-conjugated anti-rabbit or anti-mouse secondary antibody (1:10,000) was added to the wells and the samples were incubated at 37 °C for 30 min. After washing five times, any protein bound to the immobilized RECQ1 was detected using o-phenylenediamine substrate (Sigma). The reactions were terminated after 3 min with 3 m H₂SO₄, and absorbance readings were taken at 490 nm. The values represent the mean of three independent experiments performed in duplicate with the mean ± S.D. indicated by error bars.

ELISA Data Analysis—The fraction of the immobilized RECQ1 bound to the microtiter well that was specifically bound by EXO-1, MLH1-PMS2, or MSH2/6 was determined from the ELISA. A Hill plot was used to analyze the data as described previously (15).

DNA Substrates—PAGE-purified oligonucleotides (Midland Certified Reagent Co.) (Table I) were used for preparation of duplex DNA substrates. DNA substrates (nicked duplex, 1-nt 5′-flap, 25-nt 3′-flap) were prepared as previously described (16).

EXO-1 Incision Assay—20-μl reactions contained 0.5 nm DNA substrate and the indicated concentrations of RECQ1 and/or EXO-1 in 30 mm HEPES (pH 7.6), 5% glycerol, 40 mm KCl, 0.1 mg/ml bovine serum albumin, and 8 mm MgCl₂. RECQ1 was mixed with the substrate on ice prior to the addition of EXO-1. Reactions were incubated at 37 °C for 15 min (unless indicated otherwise), terminated with the addition of 10 μl of formamide dye (80% formamide (v/v), 0.1% bromphenol blue, 0.1% xylene cyanol), and heated to 95 °C for 5 min. Products were resolved on 20% polyacrylamide, 7 m urea denaturing gels. A PhosphorImager was used for detection, and the ImageQuant software (Amersham Biosciences) was used for quantitation of the reaction products. The percent incision was calculated as described previously (17).

RECQ1 Helicase Assay—Helicase reaction mixtures (20 μl) contained 20 mm Tris-HCl (pH 7.5), 10 mm KCl, 5 mm MgCl₂, 5 mm ATP, 10% glycerol, 80 μg/ml BSA, 0.5 nm DNA substrate, 1 nm RECQ1, and the indicated concentrations of MSH2/6. Reactions were initiated by the addition of RECQ1, and the reaction mixtures were incubated for 15 min at 37 °C. Helicase reactions were terminated by the addition of 20 μl of Stop buffer (35 mm EDTA, 0.6% SDS, 25% glycerol, 0.04% bromphenol blue, 0.04% xylene cyanol), and subsequently incubated with 75 μg/ml proteinase K at 37 °C for 15 min. The products were resolved on nondenaturing 12% polyacrylamide gels. Radiolabeled DNA species in polyacrylamide gels were visualized using a PhosphorImager and quantitated using ImageQuant software. The percent helicase substrate unwound was calculated by the formula: percent unwinding = 100 × (P/(S + P)), where P is the product and S is the residual substrate. The values of P and S have been corrected after subtracting background values in controls having no enzyme and heat-denatured substrate, respectively. Helicase data represent the mean of at least three independent experiments with mean ± S.D. indicated by error bars.

RESULTS

Relatively little is known regarding the cellular pathways of DNA metabolism in which human RECQ1 helicase participates. To gain insight into the molecular functions of RECQ1, we investigated its in vivo protein interactions by co-immunoprecipitation studies and in vitro physical and functional interaction assays. Using these complementary approaches, we have identified novel RECQ1 protein interactions with DNA repair factors involved in mismatch correction and DNA recombination.

Physical Interaction between RECQ1 and EXO-1—To identify the protein partner candidates of human RECQ1 that exist in vivo, co-immunoprecipitation experiments were performed using HeLa nuclear extracts and an antibody raised against a 20-amino acid peptide of a unique sequence in the C-terminal region of RECQ1 (10). The anti-RECQ1 antibody precipitated RECQ1 (Fig. 1A, bottom panel, lane 2) and co-immunoprecipitated EXO-1 (Fig. 1A, top panel, lane 2) from normal HeLa nuclear extract. Approximately 12% EXO-1 from the HeLa nuclear extract input (lane 1) was co-immunoprecipitated with RECQ1 using the anti-RECQ1 antibody. Control experiments using normal rabbit IgG incubated in the HeLa nuclear extracts confirmed the specificity of the anti-RECQ1 antibody (Fig. 1A, lane 4). Analogous experiments carried out in the presence of EtBr (50 μg/ml) demonstrated that this DNA-intercalating drug does not affect the co-immunoprecipitation of EXO-1 with RECQ1 (Fig. 1A, lane 3). These results indicate that RECQ1 and EXO-1 are associated with each other in vivo and that the interaction is not mediated by DNA.

To confirm a direct interaction between RECQ1 and EXO-1, we performed ELISA experiments with the purified recombinant proteins. Increasing amounts of EXO-1 were incubated in wells that had been previously coated with RECQ1 (18 nm) (Fig. 1B). After multiple washings, the RECQ1-EXO-1 complex was detected with a polyclonal rabbit antibody raised against EXO-1 and a colorimetric assay. The results were used to build binding curves that showed binding saturation of RECQ1-EXO-1 at EXO-1 concentrations of ≥10 nm. The specificity of this interaction was demonstrated by the absence of color in wells that had been precoated with BSA rather than RECQ1 (data not shown). To determine the affinity of the RECQ1-EXO-1 interaction, ELISA data from specific binding of EXO-1 to RECQ1-coated wells were analyzed according to the Scatchard binding theory using Hill plots as described under “Materials and Methods.” The transformed data were linear, indicating a single site on EXO-1 for binding to RECQ1. The apparent dissociation constant (K_d) was determined to be 1.37 nm. Analogous experiments performed in the presence of EtBr yielded almost identical binding curves (data not shown), indicating that the RECQ1-EXO-1 interaction is not mediated by DNA.

RECQ1 Stimulates EXO-1 Incision Activity on a 1-nt 5′-Flap Substrate—The physical interaction between RECQ1 and EXO-1, as well as the ability of a related human RECQ helicase, WRN (Werner syndrome protein), to stimulate EXO-1 cleavage by a direct protein interaction (18) suggested that RECQ1 might stimulate the incision activities of human EXO-1. To address this possibility, we first examined the ability of RECQ1 to stimulate EXO-1 cleavage of a 1-nt 5′-flap substrate, a common replication intermediate. We examined EXO-1 incision activity at various concentrations of EXO-1, either in the absence or presence of RECQ1 (5 nm). RECQ1 was observed to stimulate the incision activity of EXO-1 at each EXO-1 concentration tested (Fig. 2A). RECQ1 alone did not cleave the substrate (Fig. 2A, lane 2). Quantitation of the data revealed an ∼2-fold stimulation of EXO-1 by RECQ1 at all three concentrations of EXO-1 (Fig. 2B).

To determine whether higher concentrations of RECQ1 might increase the stimulation of EXO-1 cleavage, we tested the effect of increasing RECQ1 concentrations on EXO-1 incision of the 1-nt 5′-flap substrate. A concentration of EXO-1 (1.25 nm) that was determined to yield a small level of incision was used (Fig. 2A, lane 5). With increasing concentrations of RECQ1, a significant stimulation of EXO-1 incision activity was observed (Fig. 2, C and D). Quantitation of the 1-nt cleavage product revealed that RECQ1 stimulated EXO-1 cleavage in a concentration-dependent manner. Stimulation of EXO-1 by RECQ1 reached ∼5-fold at the highest RECQ1 concentration (20 nm) tested (Fig. 2D).

Stimulation of EXO-1 by RECQ1 on a Nicked Duplex DNA Substrate—It was recently observed that EXO-1 is required in a reconstituted mismatch repair reaction of human nuclear extracts with a DNA substrate where mismatch-provoked incision is directed by a strand break located either 5′ or 3′ to the mismatch (19). Although in vitro data suggest that replication factor C and PCNA can direct EXO-1 nick-directed repair in the reconstituted mismatch repair system (20), factors that modulate the efficiency of EXO-1 incision in mismatch repair have not been identified. Therefore, we set out to test the effect of RECQ1 on EXO-1 cleavage activity of a proposed DNA mismatch repair intermediate, a nicked duplex DNA substrate. The nicked duplex substrate that was tested contained a 5′-³²P label on the downstream primer of the nicked duplex substrate. To verify its integrity, the nicked duplex substrate, incubated with T4 DNA ligase under conditions suitable for ligation, was found to have ∼95% of its upstream and downstream primers joined (data not shown), providing evidence that the substrate does indeed contain a ligatable nick.

We first examined EXO-1 cleavage of the nicked duplex DNA substrate in the presence or absence of a fixed RECQ1 concentration of 5 nm (Fig. 3A). Using a limiting concentration of EXO-1 (1.25 nm), EXO-1 5′-3′ incision was stimulated 4-fold by 5 nm RECQ1, resulting in 13% incision compared with 3% by EXO-1 alone (Fig. 3B). 2.5 nm EXO-1 incised 10% nicked duplex DNA substrate. The addition of RECQ1 increased the amount of substrate incised by EXO-1 to 40%, a 4-fold stimulation. We also tested a nicked duplex with a 3′-end label on the upstream primer and found that EXO-1 in the absence or presence of RECQ1 did not incise the 3′-label under conditions where the 5′-end label of the downstream primer was catalytically removed (data not shown).

We next studied EXO-1 cleavage of the nicked duplex DNA substrate as a function of RECQ1 concentration. Using a limited amount of EXO-1 (2.5 nm), RECQ1 stimulated EXO-1 incision of the nicked duplex to yield the 1-nt product in a RECQ1 concentration-dependent manner (Fig. 3C). In the control reactions, RECQ1 (20 nm) did not cleave the substrate (Fig. 3C, lane 2). Quantitation of the EXO-1 cleavage in the presence of RECQ1 is shown in Fig. 3D. At 5 nm RECQ1, EXO-1 incision was increased 4-fold. Further increases to 7- and 10-fold were observed at RECQ1 concentrations of 10 and 20 nm RECQ1, respectively.

Helicase Activity of RECQ1 Is Not Required for Stimulation of EXO-1 Incision—The helicase activity of RECQ1 may play a role in the functional interaction with EXO-1. To address this possibility, we tested the effects of full-length recombinant RECQ1 protein with a site-directed mutation in the conserved ATP-binding site on EXO-1 incision of the 1-nt flap and nicked duplex DNA substrates. The RECQ1-K119A mutant protein, devoid of helicase activity (Supplemental Fig. 1), was capable of stimulating the EXO-1 incision reaction similar to wild-type RECQ1 on both the 1-nt flap and nicked duplex DNA substrates (Fig. 4). In control reactions, RECQ1-K119A alone did not yield the products (Fig. 4). Thus, DNA unwinding of the substrate is dispensable for the RECQ1-EXO-1 functional interaction. This result is consistent with our observation that RECQ1 stimulates EXO-1 incision of either the 1-nt flap or nicked duplex substrates in the absence of ATP (Figs. 2 and 3).

Physical Interaction between RECQ1 and Other DNA Repair Factors—The interaction of RECQ1 with EXO-1 and RPA (10) suggested to us that RECQ1 might interact either sequentially or in a complex with other DNA repair factors. The robust stimulation by RECQ1 of EXO-1 incision of a nicked duplex substrate, a key intermediate in mismatch repair and related processes, suggested that proteins implicated in the mismatch repair pathway might also interact with RECQ1. To address this issue, co-immunoprecipitation experiments from HeLa nuclear extracts were performed using the anti-RECQ1 peptide antibody and the immunoprecipitated proteins were probed with antibodies against selected mismatch repair proteins by Western blot analysis. A mismatch repair protein that was consistently co-immunoprecipitated with RECQ1 was MLH1 (Fig. 5A). Approximately 8% MLH1 from the HeLa nuclear extract input (Fig. 5A, lane 3) was co-immunoprecipitated with RECQ1 using the anti-RECQ1 antibody (lane 1). Control experiments using normal rabbit IgG incubated in the HeLa nuclear extracts confirmed the specificity of the anti-RECQ1 antibody (Fig. 5A, lane 4). Experiments carried out in the presence of EtBr demonstrated that the co-immunoprecipitation of RECQ1 and MLH1 is not mediated by nucleic acids (Fig. 5A, lane 2).

To explore the nature and specificity of the putative RECQ1-MLH1 interaction, we conducted ELISAs using the purified recombinant proteins. MutLα, a heterodimer of MLH1 and PMS2, was added in increasing concentration to wells that had prebound with RECQ1 (18 nm application). After washing, the RECQ1-MutLα complex was detected with a mouse polyclonal antibody raised against MLH1 and a colorimetric assay was used to build binding curves that reached saturation at MutLα concentrations of ∼10 nm (Fig. 5B). Experiments performed in the presence of EtBr yielded similar binding curves (data not shown), indicating that the RECQ1-MLH1 interaction is not DNA-mediated. The specificity of the RECQ1-MLH1 interaction was demonstrated by the absence of color in wells that had been precoated with BSA rather than RECQ1. The transformed data were linear, indicating a single site on the MutLα complex for binding to RECQ1. Specific binding of MutLα to RECQ1-coated wells was analyzed according to the Scatchard binding theory. The apparent dissociation constant (K_d) was determined to be 3.2 nm when the data were analyzed by Hill plots as described under “Materials and Methods.”

Co-immunoprecipitation experiments from HeLa nuclear extracts also revealed that the MSH2 and MSH6 proteins were consistently co-immunoprecipitated with RECQ1 (Fig. 6, A and B, lane 1). Approximately 4% MSH2 and MSH6 were immunoprecipitated with RECQ1 from the nuclear extract with the RECQ1 antibody. A control experiment with normal rabbit IgG incubated with HeLa nuclear extract confirmed the specificity of the anti-RECQ1 antibody (Fig. 6, A and B, lane 4). The presence of EtBr did not negate the ability of MSH2 or MSH6 to be co-immunoprecipitated with RECQ1 (Fig. 6, A and B, lane 2), suggesting that a DNA bridge was not responsible for their association.

To determine whether a direct interaction between RECQ1 and MSH2/6 existed, we performed ELISAs with the purified recombinant proteins. Increasing concentrations of MSH2/6 complex were incubated in wells that had been previously coated with RECQ1 (18 nm). After washing, bound MSH2 complex was detected with a mouse monoclonal antibody raised against MSH2 and a colorimetric assay was used to build binding curves that reached saturation at MSH2/6 concentrations of ∼36 nm (Fig. 6C). The presence of EtBr in the ELISAs with RECQ1 and MSH2/6 did not interfere with the colorimetric signal, indicating that a DNA bridge was not responsible for the protein interaction. The specificity of this interaction was demonstrated by the absence of color in wells that had been precoated with BSA rather than RECQ1. Binding data analyses yielded an apparent dissociation constant (K_d) for the RECQ1-MSH2/6 interaction of 8.0 nm.

The MSH2/6 Mismatch Recognition Complex Stimulates RECQ1 Helicase Activity—In vivo data suggest that RecQ helicases may function with the mismatch repair machinery to regulate recombination mediated by strand invasion (21). One mechanism for their functional interaction might be for the mismatch recognition complex to modulate DNA unwinding of a strand invasion intermediate. Because RECQ1 preferentially unwinds a 3′-flap substrate,⁴ we tested the effect of MSH2/6 on this activity. As shown in Fig. 7, MSH2/6 complex stimulated RECQ1 unwinding of the substrate in a dose-dependent manner. Maximum stimulation of RECQ1 helicase activity was observed at the highest MSH2/6 concentration tested (25 nm), resulting in a 3-fold stimulation of RECQ1 helicase activity on the 3′-flap substrate. The stimulatory effect of MSH2/6 on RECQ1 helicase activity was specific, because the presence of MutLα complex in RECQ1 helicase assays with the 3′-flap substrate or other DNA substrates did not have a stimulatory or inhibitory effect on DNA unwinding (data not shown).

DISCUSSION

Dissecting the roles of RecQ helicases in the maintenance of genomic stability has posed a formidable challenge to researchers. Genetic studies suggest an important role of E. coli RecQ in the repression of homologous recombination (22, 23). In vitro studies with purified E. coli RecQ and recombination DNA intermediates suggested that RecQ helicase activity might serve to disrupt recombination intermediates; however, in a reconstituted system with the single-stranded DNA exchange protein RecA and the E. coli single-stranded DNA-binding protein, RecQ was shown to both initiate recombination as well as unwind RecQ-mediated joint molecules (23). Conceivably, E. coli RecQ may serve to prevent or promote recombination, depending on its subcellular interactions with DNA and/or recombination/repair protein factors.

Studies of the sole RecQ homolog in Saccharomyces cerevisiae have suggested that Sgs1 functions to prevent homologous recombination (24, 25) or act as a sensor for DNA damage (6). The Kolodner laboratory examined the nature of the genomic instability in sgs1 mutants and determined that mutations in sgs1 increased the rate of accumulating gross chromosomal rearrangements, including translocations and deletions containing extended regions of imperfect homology at their break-points (24). sgs1 mutations also triggered a 6-fold increase in the rate of recombination among DNA sequences that had 91% sequence homology. Epistasis analysis showed that Sgs1 is redundant with MSH2 for suppressing gross chromosomal rearrangements and for suppressing recombination among divergent DNA sequences. Recently, results from a single strand-annealing assay in yeast provided evidence that an sgs1 mutation inhibited repression of recombination among homologous sequences (25). Genetic and physical analyses of MSH2, MSH6, and SGS1 alleles harboring mutations that disrupt the biochemical activities of MSH2, MSH6, and Sgs1 support a model in which the MSH2/6 proteins interact with Sgs1 to unwind DNA recombination intermediates that contain mismatches (26). Further support that Sgs1 collaborates with mismatch repair proteins in the rejection of homologous recombination is provided by the physical interactions of Sgs1 with MLH1 (27) and Rad51 (28). Together, these data implicate Sgs1 as a regulator of homologous recombination.

Similar to RecQ and Sgs1, human RecQ helicases are suggested to be involved in recombination. A role for WRN in the promotion of homologous recombination by facilitating the resolution of recombination intermediates was evidenced by in vivo data (29, 30). In vitro studies suggested a role of BLM helicase with human topoisomerase IIIα to suppress crossing over during homologous recombination (31). Physical interactions of human RecQ proteins with recombination/repair factors suggest a cross-talk between recombination machinery and the human RecQ helicases as well. WRN interacts with the recombination mediator protein Rad52 (32). Rad51D physically interacts with BLM and stimulates BLM-catalyzed Holliday Junction branch migration activity (33). Physical interactions among BLM and Rad51 (28), MSH6 (34), and MLH1 (27, 35) have been shown; however, extracts of Bloom syndrome cells are proficient in mismatch repair (27, 35). Purified recombinant MSH2/6 complex stimulates branch migration activity of BLM on Holiday Junction structures (36). The early identification of BLM in BRCA1-associated complex-containing proteins such as MSH2/6 and MLH1 that are involved in the recognition and repair of aberrant DNA structures (37) further supports the idea of cooperativity among human RecQ helicases, such as BLM and genome surveillance proteins, that regulate recombination.

Unlike WRN or BLM helicases, the protein interactions of human RECQ1 have not been extensively investigated. To identify candidate RECQ1 protein partners, we performed co-immunoprecipitation experiments with HeLa nuclear extracts using a custom polyclonal antibody raised against a 20-amino acid peptide of unique sequence in the C-terminal region of RECQ1. This antibody was previously employed by us to demonstrate an interaction between RECQ1 and RPA, which was verified with purified recombinant proteins (10). Here, we present the first line of evidence that RECQ1 interacts with factors of the mismatch repair machinery that regulate genetic recombination. We have demonstrated a tight physical interaction of RECQ1 with EXO-1 and that RECQ1 dramatically stimulates EXO-1 nuclease activity. A conserved RecQ protein interaction is suggested by the fact that RECQ1 has a conserved noncatalytic domain found in WRN protein that mediates the physical and functional interaction with EXO-1 (18). EXO-1 has been implicated in mismatch repair (19, 20, 38), but the enzyme also has roles in recombination (38-42), telomere maintenance (43-46), and possibly replication (38, 47-49). The prominent role of RecQ helicases in pathways related to recombination raises the possibility that RECQ1 and EXO-1 function together when the nucleolytic digestion by EXO-1 is involved.

It was suggested that yeast EXO-1 may substitute for the Mre11-Rad50-Xrs2 complex in the resection of double strand breaks by genetic studies in which the EXO-1 complementation of methyl methane sulfonate and ionizing radiation sensitivity of mre11, rad50, and xrs2 mutants was dependent on the nuclease activity of EXO-1 (40-42). EXO-1 may participate in telomere maintenance when telomerase or telomere-binding proteins are absent by creating long telomeric single-stranded DNA tails that promote recombination-driven lengthening of telomeres (43, 46) or activation of a DNA damage checkpoint (45). RECQ1, similar to other RecQ helicases, may facilitate genome maintenance by its role in recombination and through its functional interactions with structure-specific nucleases such as EXO-1. Proposed roles of EXO-1 in a mismatch repair-independent mutation avoidance pathway (50, 51) or repair/restart of stalled replication forks (52) may also involve RecQ helicases such as RECQ1 or WRN (18).

The cross-talk of RECQ1 with a nuclease that is implicated in mismatch repair and recombination suggested to us that RECQ1 might also interact with mismatch repair factors that regulate genetic recombination. Our results demonstrate that RECQ1 physically interacts with the DNA repair complexes MLH1/PMS2 and MSH2/6. In addition, MSH2/6, but not MLH1/PMS2, stimulated RECQ1 helicase activity. These results indicate that RECQ1, similar to BLM, interacts with mismatch repair proteins that serve to regulate genetic recombination.

Functional redundancy between RECQ1 and BLM in vivo was suggested by recent genetic studies. Simultaneous deletion of RECQ1 and BLM in chicken DT40 cells resulted in slow growth, increased cell death, and a higher incidence of sister chromatid exchange upon mitomycin C treatment (53). These results would suggest that BLM and RECQ1 may have partially redundant roles to maintain genome integrity, but their roles only overlap under certain circumstances in vivo. In the case of RECQ1 and BLM, the two proteins may substitute for each other when DNA damage blocks replication fork progression; however, RECQ1 does not perfectly substitute for BLM because BLM^-/- cells still display elevated sister chromatid exchange. Thus the requirement for BLM is more stringent because RECQ1^-/- cells did not display a mutant phenotype.

Recently, a model for RecQ helicase function with mismatch repair proteins was suggested that is relevant to the work presented here (21). During homologous recombination, the mismatch repair recognition complex will stall strand exchange in the presence of mismatches, yielding stabilized branched intermediates. At the DNA molecular level, the flap substrate is a substructure of the d-loop intermediate of homologous recombination. The unwinding of the annealed strands by the RecQ helicase is facilitated by interactions with bound MSH 2/6 complex. This permits a subsequent homology search to occur. Consistent with a potential role of RECQ1 on homologous recombination intermediates, RECQ1 preferentially unwinds the invading strand of d-loop structures and catalyzes branch migration activity on synthetic Holliday Junction structures in vitro.⁴

It is conceivable that RECQ1 shares with BLM and perhaps other RecQ helicases a specialized function in a subpathway of homologous recombination such as synthesis-dependent strand annealing. Recently, a genetic role for BLM in synthesis-dependent strand annealing was evidenced (54, 55). In the model presented, BLM is predicted to function downstream of strand invasion to unwind a d-loop intermediate to free the newly synthesized strand. If base pairing during synthesis-dependent strand annealing is imperfect, it seems likely that a RecQ helicase (e.g. BLM, RECQ1) would be signaled by the MSH complex to disrupt the heteroduplex joint molecule to avoid recombination with the imperfect locus. Future studies will address the importance of mismatches in defined DNA intermediates to the catalytic functions of human RecQ helicases in the proposed role of heteroduplex rejection.