Subunit Interface Residues F129 and H294 of Human RAD51 Are Essential for Recombinase Function

RAD51 mediated homologous recombinational repair (HRR) of DNA double-strand breaks (DSBs) is essential to maintain genomic integrity. RAD51 forms a nucleoprotein filament (NPF) that catalyzes the fundamental homologous pairing and strand exchange reaction (recombinase) required for HRR. Based on structural and functional homology with archaeal and yeast RAD51, we have identified the human RAD51 (HsRAD51) subunit interface residues HsRad51(F129) in the Walker A box and HsRad51(H294) in the L2 ssDNA binding region as potentially important participants in salt-induced conformational transitions essential for recombinase activity. We demonstrate that the HsRad51(F129V) and HsRad51(H294V) substitution mutations reduce DNA dependent ATPase activity and are largely defective in the formation of a functional NPF, which ultimately eliminates recombinase catalytic functions. Our data are consistent with the conclusion that the HsRAD51(F129) and HsRAD51(H294) residues are important participants in the cation-induced allosteric activation of HsRAD51.


Introduction
Failure to repair DNA double strand breaks (DSBs) leads to tumorigenesis and genomic instability. Homologous recombination (HR) is an evolutionary conserved repair pathway utilized to restore DSBs. HR mediated DSB repair is initiated by resection of the 59-end of the break to leave a 39-single-stranded DNA (ssDNA) overhang [1,2]. In eukaryotes, RAD51 forms a nucleoprotein filament (NPF) on the newly formed ssDNA region aided by other recombination mediators such as RAD52 in yeast and BRCA2 in vertebrates [3,4,5,6,7,8]. The key function of the RAD51 NPF is to catalyze the homology search and initiate strand exchange. Deletion of Rad51 in mice results in embryonic lethality, while RAD51 knock down in chicken DT40 cell lines results in increased chromosomal instability [9,10]. Even though, no mutations of RAD51 have been found in cancers, its expression is elevated in many cancer cell lines; perhaps to provide an advantage to rapidly dividing cells by repairing DSBs that would lead to replication fork collapse [11,12,13].
Despite the functional conservation with bacterial RecA, human RAD51 (HsRAD51) possesses an essential cation salt requirement for efficient strand exchange in vitro. The most effective cation is ammonium (NH 4 + ), which is unlikely to be physiologically significant. However, potassium (K + ) also enhances HsRAD51 recombinase functions to a lesser extent [14,15] and has been shown to induce single stranded DNA-HsRAD51 complexes that structurally mimic active RecA NPFs [14]. The effect of salt has been attributed to an induced preferential single stranded DNA (ssDNA) binding by RAD51 over double stranded DNA (dsDNA) [15]. However, crystallographic analysis of the methanococcus voltae RadA (MvRAD51) has revealed important K + -induced amino acid and structural rearrangements within the intersubunit region of the NPF ( Figure 1A and 1B). For example, the MvRAD51(F107) residue in the highly conserved Walker A box rotates away from the ATP binding interface to accommodate K + cations (Figure 1b; [16]. This rotation of MvRAD51(F107) induces ordering of the L2 ssDNA binding domain and a conformational transition of MvRAD51(H280) that results in the formation of a direct hydrogen bond with the c-phosphate group of the ATP analogue; which is suggested to be involved in polarizing the water molecule involved in ATP hydrolysis. These two residues are conserved in RAD51 homologs from archaebacteria to human ( Figure 1C).
Here, we have examined substitution mutations of the analogous HsRAD51 residues to MvRAD51(F107) and MvRAD51(H280), HsRAD51(F129) and HsRAD51(H294). We find that HsRAD 51(F129V) and HsRAD51(H294V) affect ATP hydrolysis (ATPase) activity but not adenosine nucleotide binding or ADPRATP exchange. Moreover, they alter DNA binding properties in the presence of ATP and salt cations (K + or NH 4 + ) that ultimately results in defective recombinase functions. These data further delineate the importance of salt-induced allosteric changes at the subunit interface of HsRAD51 that promotes the formation of a functional NPF.
Reduced k cat could be due to several factors: an ATP binding deficiency, an ADP release deficiency or a deficiency in the catalytic step. To rule out the first two possibilities, we examined ATP binding, ADP binding and ADP-ATP exchange of the mutant proteins. We found no significant difference in ATP binding (Fig. 2C [15]. Only RAD51(F129V) displayed a similar salt (150 mM KCl) induced catalytic rate enhancement in the presence of ssDNA (Fig. 2E). Together these observations are consistent with the conclusion that HsRAD51(H294) residue and to a significantly lesser extent the HsRAD51(F129) residue are required for appropriate ATP catalysis.

HsRAD51(F129V) and HsRAD51(H294V) are deficient in D-loop formation and strand exchange
We examined the recombinase activity of the HsRAD51(F129V) and HsRAD51(H294V) substitution mutations. HsRAD51 catalyzes D-loop formation in vitro between a 32 P-labeled 90-mer and a homologous supercoiled plasmid (Fig. 3A). Efficient D-loop formation occurs when ATP is present in a slow-hydrolysable state (e.g. in the presence of calcium) or when non-hydrolysable ATP analogues such as AMP-PNP are substituted for ATP [22,23]. Importantly, the Walker A box mutant HsRAD51(K133R) that binds ATP normally but is defective in ATP hydrolysis, catalyzes efficient D-loop formation with ATP and Mg 2+ [23]. These results have suggested that Dloop catalysis requires HsRAD51 to form and maintain an active ATP-bound NPF [22,23,24,25].
In the presence of ATP, HsRAD51 catalyzed D-loop formation with Ca 2+ but not with Mg 2+ (Fig. 3B). Previous work has demonstrated that Ca 2+ induces the formation of a stable ATPbound NPF while Mg 2+ allows the hydrolysis of ATP that results in a mixed (ADP/ATP) NPF that is largely inactive [22]. HsRAD51(F129V) and HsRAD51(H294V) did not catalyze Dloop formation in either Mg 2+ or Ca 2+ (Fig. 3B). Thus, even though HsRAD51(H294V) is ATPase deficient, and both HsRAD51(F129V) and HsRAD51(H294V) display ATP binding that is comparable to the wild type, these mutant proteins are incapable of catalyzing D-loop formation. Collectively these results indicate that suppression of ATP hydrolysis alone is not sufficient to confer enhanced D-loop recombinase activity.
Catalysis of strand exchange between a duplex DNA substrate and a homologous single stranded circular DNA substrate is a hallmark of RecA/RAD51 proteins [18,24,26,27,28,29]. Unlike D-loop formation that occurs in low salt, strand exchange requires specific cations (either NH 4 + or K + ) to activate HsRAD51 activity [14,15,30]. Using wX174 virion DNA and ApaL1 linearized wX174 dsDNA we compared the strand exchange activity of the HsRAD51(F129V) and HsRAD51(H294V) mutant proteins to wild type HsRAD51. We also examined the stimulatory effect of the human single-stranded binding complex RPA (HsRPA). Neither HsRAD51(F129V) nor HsRAD51(H294V) were able to form joint molecules in the presence of salt and/or RPA (Fig. 3C). We can conclude that the HsRAD51(F129) and HsRAD51(H294) residues are critical for cation-induced HsRAD51 recombinase functions.

HsRAD51(F129) and HsRAD51(H294) are essential for ATP-dependent DNA binding
We examined real-time HsRAD51 ssDNA and dsDNA binding by surface plasmon resonance (SPR, Biacore). Biotinylated dT 50  ssDNA and 50 bp dsDNA were immobilized on a streptavidincoated flow-cell surface. SPR measures the change in the refractive index that reflects protein binding and/or dissociation from DNA. We analyzed ssDNA and dsDNA binding in salt conditions similar to those used in for strand exchange. We titrated the protein from 100 nM to 1.6 mM in the presence of ADP, ATP and in the absence of adenosine nucleotide. For clarity the binding and dissociation curves shown correspond to 800 nM of protein (Fig. 4). After protein injection nucleoprotein filament formation was analyzed for 150 sec. For steady-state protein disassembly, protein free buffer was injected into the flow cells for 300 sec. Because the ssDNA and dsDNA were introduced into separate channels we could simultaneously examine protein binding and dissociation to the two DNA substrates. In the absence of a adenosine nucleotide only the wild type HsRAD51 and HsRAD51(H294V) bound to ssDNA (Fig. 4A, left panel, Table 2). In the presence of ADP, the wild type and mutant variants bound to ssDNA, but dissociated rapidly (Fig. 4A, middle panel, Table 2). These results are consistent with previous studies that have suggested protein turnover associated with ADP-bound RAD51 [22,31]. Interestingly, in the presence of ATP, wild type HsRAD51 bound ssDNA while both HsRAD51(F129V) and HsRAD51(H294V) were largely defective in ssDNA binding (Fig. 4A, right panel, Table 2). Only wild type HsRAD51 displayed dsDNA binding in the presence of ATP as well as in the absence of a nucleotide (Fig. 4B, Table 2). These results suggest that the HsRAD51(F129) and HsRAD51(H294) residues are important for the ATP induced DNA binding that ultimately results in an active NPF required for recombinase function.
The abnormal DNA binding did not appear to fully account for the altered steady-state ATPase activity of HsRAD51(F129V) and HsRAD51(H294V) (compare Fig. 2E, Table 1 with Fig. 4, Table 2). This is particularly true for HsRAD51(H294V) which showed reduced but not absent binding to ssDNA, yet near background ATPase activity. The dissociation of ssDNA binding activity from ATPase activity suggests that some other function(s) of these proteins are compromised. Based on the location of these residues in the crystal structures we postulate that cation induced conformational transitions are defective in HsRAD51(F129V) and HsRAD51(H294V) (see Figs. 1A and 1B). These conformational transitions do not influence ADP/ATP binding or ADPRATP exchange, but do affect the ability of the mutant proteins to interact appropriately with DNA and to properly catalyze ATP hydrolysis.

Discussion
Structural studies of the MvRAD51 and S.cerevisiae RAD51 (ScRAD51) have detailed significant conformational transitions associated with the NPF. The ScRAD51 residues analogous to HsRAD51(F129) and HsRAD51(H294), ScRAD51(F187) and ScRAD51(H352), were found in two alternate conformations within the NPF [32,33]. In one conformation ScRAD51(H352) was positioned above the ATP binding site while in the other ScRAD51(F187) excluded ScRAD51(H352) by moving it out of the active site. Importantly, K + cations induce similar conformational transitions of MvRAD51 of the analogous residues to HsRAD51(F129) and HsRAD51(H294), MvRAD51(F107) and MvRAD51(H280). In the presence of K + the MvRAD51(F107) rotates away from the ATP catalytic interface and MvRAD51(H280) rotates in to form a hydrogen bond interaction with the c-phosphate of ATP, while at the same time ordering the L2 ssDNA binding domain [16].
Our studies support the importance of these conformational transitions. While MvRAD51(F107V) can bind and hydrolyze ATP, it displays significantly reduced ssDNA binding in the presence of ATP. These results suggest an impaired ability to incite the conformational transitions necessary to the form an active NPF on ssDNA. We speculate that this impairment involves an inability of the smaller Val residue to incite ordering of the L2 ssDNAbinding domain that is normally provoked by cation-induced rotation of the Phe residue away from the ATP catalytic interface. In contrast, HsRAD51(H294V) binds but does not hydrolyze ATP, binds ssDNA in the absence of adenosine nucleotide, yet displays significantly reduced ssDNA binding in the presence of ATP. We speculate that these impairments are the result of an inability of the Val residue to form an appropriate hydrogen bond interaction with the c-phosphate of ATP in spite of the fact that L2 ssDNA-binding domain ordering has been provoked by the MvRAD51(F107) residue [34]. Modeling of the HsRAD51 structure strongly support such an allosteric communication between ATPase site and L2 ssDNA binding region [35,36]. A recent mutational analysis also revealed that the subunit interface residue ScRAD51(H352) is critical for functional NPF formation and strand exchange activity [37]. Ultimately, the defects associated with both HsRAD51(F129V) and HsRAD51(H294V) result in defective D-loop and strand exchange functions. Because both proteins bind ADP and ATP similar to the wild type HsRAD51, the functional defects are unlikely to be a result of improper folding and/or aggregation.
For RecA/RAD51 proteins ATP binding but not necessarily hydrolysis, is sufficient to catalyze strand exchange [23,26,27,28,38,39]. It has been suggested that RAD51 might utilize binding of ATP and the resulting release of the hydrolysis product as a conformational switch for regulating recombinase function similar to other members of the AAA+ superfamily [40]. Our data is consistent with the hypothesis that the HsRAD51(F129) and HsRAD51(H294) residues along with salt cations may play a significant role in this conformational switch during HRR.

HsRAD51 Protein Expression and Purification
The mutant RAD51 genes F129V and H294V were constructed using wild type HsRAD51 gene in the pET24d vector system (Novagen) using conventional PCR. All mutations were confirmed by DNA sequencing. Wild type and mutant hRAD51 proteins were expressed and purified following previously published protocols [17,41]. Briefly, hRAD51 was expressed in E. coli BLR strain and precipitated using spermidine-HCl. Resuspended pellet was purified using Reactive-Blue-4-agarose, Heparin sepharose, hydroxyapatite and Mono Q column chromatography. Purity of the fractions was verified by SDS-PAGE analysis. HsRPA was expressed in BL21(AI) cells using pET11d-tRPA purified as described [42], except for the resuspension of cells where HI buffer was supplemented with 100 mM KCl.

DNA Substrates
wX174 single-stranded (ss) virion DNA and replicative form I (RFI) were purchased from NEB. wX174 RFIII was obtained by linearizing RFI with ApaLI restriction enzyme and gel purifying with Qiaquick Gel Extraction kit (Qiagen). For surface plasmon resonance (SPR) analysis, a 59 biotinylated oligo dT 50 was used as ssDNA and for dsDNA 59 biotinylated 50-mer 59-TCG AGA GGG TAA ACC ACA-ATT ATT GAT ATA AAA TAG TTT TGG GTA GGC GA was annealed with its complement. D-loop assay substrates were prepared as described [43].

ATPase assay
ATP hydrolysis was measured as previously described [17]. Reactions were performed in 10 mL volumes in Buffer A containing 20 mM HEPES (pH 7.5), 10% glycerol, 100 mg/mL BSA, 1 mM DTT and 1 mM MgCl 2 and when indicated, 150 mM KCl. Each reaction mixture contained RAD51 (1 mM) and 6 mM (nt or bp) of wX174 ssDNA or dsDNA. Reactions were initiated by the addition of protein and incubation at 37uC. After 1 hr 400 mL of 10% activated charcoal supplemented with 10 mM EDTA was added to terminate the reaction and incubated on ice for another 2 hrs. After centrifuging for 10 min 50 mL duplicate aliquots were taken for counting [ 32 P] free phosphate by Cerenkov method. Kinetic parameters were obtained by fitting data into Michaelis-Menten equation using the software Kaleidagraph (Synergy software).

ATPcS/ADP binding assay
Experimental procedure was as previously described [17]. Reactions were performed in Buffer A and 150 mM KCl, when indicated. RAD51 (1 mM) and 6 mM (nt) wX174 ssDNA were used with the indicated amount of [c- 35 S]ATPcS or [ 3 H]ADP. After a 30 min incubation at 37uC, reactions were kept on ice until filtered. Reaction was added into 4 mL of ice-cold reaction buffer and filtered through HAWP nitrocellulose filters (Millipore) presoaked in the same buffer. Another 4 mL of buffer was used to wash the membrane and the filter was dried for 2 hrs. Radioactivity was counted after an overnight incubation in liquid scintillation fluid.

ADP-ATP exchange assay
Reactions were performed in Buffer A with 150 mM KCl. RAD51(1 mM) and 6 mM (nt) wX174 ssDNA was incubated with 3 mM of [ 3 H]ADP in a 60 mL reaction volume at 37uC for 10 mins. ADP to ATP exchange was initiated by addition of cold ATP to a 5 mM final concentration. 10 mL aliquots were withdrawn at indicated time points, filtered and analyzed as in ADP binding.

SPR analysis
Biotinylated DNA was immobilized on a streptavidin-coated chip (GE) for the analysis. Protein binding and dissociation were analyzed at 25uC with a 5 mL/min flow rate on a Biacore 3000 (GE). Reactions were performed in Buffer containing 20 mM HEPES (pH 7.5), 10% glycerol, 1 mM DTT and 1 mM MgCl 2 , 0.005% surfactant P-20 (GE), 2.5 mM of the indicated nucleotide and 150 mM KCl when indicated. For experiments with Ca 2+ , 1 mM CaCl 2 was used instead of MgCl 2 . DNA binding was

D-loop assay
2.4 mM (nt) labeled 90-mer was incubated with HsRAD51 (0.8 mM) in buffer A with 1 mM ATP supplemented with the indicated amounts of MgCl 2 or CaCl 2 at 37uC for 10 min. Reaction was initiated by adding 35 mM (bp) supercoiled pBS-SK(-) plasmid and incubated further for 15 min. Samples were deproteinized by addition of 1% SDS and 1 mg/mL proteinase K to a final concentration and incubated for 15 min at 37uC, mixed with 1/5 volume of gel loading dye and analyzed on 0.9% agarose gel in TAE buffer, run at 4 V/cm at 25uC. D-loops were quantified after drying and exposure to PhosphoImager screens.

DNA Strand Exchange Assay
HsRAD51 (5 mM) and 30 mM (nt) wX174 circular ssDNA were pre-incubated in buffer containing 20 mM HEPES (pH 7.5), 10% Glycerol, 1 mM DTT, 1 mM MgCl 2 supplemented with 2.5 mM of ATP at 37uC for 5 min before addition of 150 mM of the indicated salt and 15 mM (bp) linear wX174 dsDNA. After another 5 min incubation 2 mM of HsRPA was added and the incubation was continued. After 3 hrs samples were deproteinized by addition of 3 mL of stop buffer containing 2% SDS and 10 mg/mL proteinase K, and analyzed on 0.9% agarose gel in TAE buffer. Electrophoresis was carried out at 4 V/cm at 25uC with 0.1 mg/ mL ethidium bromide.