**4. Functional significance of ABC-type ATPase properties of RecF**

#### **4.1 ATP-dependent dimerization is required for DNA binding**

The DNA binding properties of RecF and their role in recombination initiation remain poorly understood and controversial. Different publications presented contradicting results of DNA junction recognition by RecF (Hegde et al., 1996; Webb et al., 1999). RecR was shown to stabilize ATP-dependent interaction of RecF with DNA. However, RecR also stimulated ATP hydrolysis, which theoretically should lead to destabilizing of RecF complex with DNA (Webb et al., 1995). Therefore, multiple complimentary equilibrium binding techniques were utilized to comprehensively address the relationship between dimerization, DNA binding and ATP binding and hydrolysis (Makharashvili et al., 2009). Quantitative characterization of RecF dimerization was performed using Föster (or Fluorescence) Resonance Energy Transfer (FRET) technique with a mixture of Cy3- and FAM(fluorescein) labeled DrRecF (Fig. 4). The cysteine substitutions were introduced either at a topical part of domain II arm or at the C-terminal tail to crosslink DrRecF with fluorophores. The labeling of domain II interfered with DNA-binding (Makharashvili, 2009), indirectly confirming the dimer model presented in Fig. 3, where apical parts of domain II arms are situated close to each other in the dimer and the presence of bulky polar fluorophores may interfere with DNA binding. C-terminally labeled protein (A355C) was fully functional. Apparent dimerization constant of Ld = 0.15 ± 0.02 µM was calculated from the plot of FRET signal versus DrRecF concentration (Fig. 4C). Alternatively, multiple data sets (Fig. 4B) were globally fitted into a two-step reaction model consisting of the ATP-binding and dimerization processes resulting in a dimerization constant of Ld = 0.13 ± 0.02 µM and an ATP-binding constant of KdATP = 13 ± 2 µM.

Fig. 4. ATP-dependent dimerization of DrRecF. **A**) Location of cysteines (A355C) are shown by magenta spheres on the model of DrRecF dimer with one monomer is colored in yellow and the other color-coded accordingly to its secondary structure elements with α-helixes in red and β-strands in green. The DNA is shown in cyan. **B**) Titration of labeled DrRecF by ATP. Different isotherms represent different concentration of DrRecF in solution (values are shown on the right). The black isotherm corresponds to titration of 2 μM DrRecF by ADP. **C**) A plot of maximal FRET signal versus DrRecF concentrations.

The DNA binding was first assayed using short FAM-labeled oligonucleotides with the fluorescent polarization anisotropy method (Fig. 5). To address initial DNA binding rate,

to the presence of high molecular weight protein aggregates, and it was utilized to optimize

The DNA binding properties of RecF and their role in recombination initiation remain poorly understood and controversial. Different publications presented contradicting results of DNA junction recognition by RecF (Hegde et al., 1996; Webb et al., 1999). RecR was shown to stabilize ATP-dependent interaction of RecF with DNA. However, RecR also stimulated ATP hydrolysis, which theoretically should lead to destabilizing of RecF complex with DNA (Webb et al., 1995). Therefore, multiple complimentary equilibrium binding techniques were utilized to comprehensively address the relationship between dimerization, DNA binding and ATP binding and hydrolysis (Makharashvili et al., 2009). Quantitative characterization of RecF dimerization was performed using Föster (or Fluorescence) Resonance Energy Transfer (FRET) technique with a mixture of Cy3- and FAM(fluorescein) labeled DrRecF (Fig. 4). The cysteine substitutions were introduced either at a topical part of domain II arm or at the C-terminal tail to crosslink DrRecF with fluorophores. The labeling of domain II interfered with DNA-binding (Makharashvili, 2009), indirectly confirming the dimer model presented in Fig. 3, where apical parts of domain II arms are situated close to each other in the dimer and the presence of bulky polar fluorophores may interfere with DNA binding. C-terminally labeled protein (A355C) was fully functional. Apparent dimerization constant of Ld = 0.15 ± 0.02 µM was calculated from the plot of FRET signal versus DrRecF concentration (Fig. 4C). Alternatively, multiple data sets (Fig. 4B) were globally fitted into a two-step reaction model consisting of the ATP-binding and dimerization processes resulting in a dimerization constant of Ld = 0.13 ± 0.02 µM and an

Fig. 4. ATP-dependent dimerization of DrRecF. **A**) Location of cysteines (A355C) are shown by magenta spheres on the model of DrRecF dimer with one monomer is colored in yellow and the other color-coded accordingly to its secondary structure elements with α-helixes in red and β-strands in green. The DNA is shown in cyan. **B**) Titration of labeled DrRecF by ATP. Different isotherms represent different concentration of DrRecF in solution (values are shown on the right). The black isotherm corresponds to titration of 2 μM DrRecF by ADP. **C**)

The DNA binding was first assayed using short FAM-labeled oligonucleotides with the fluorescent polarization anisotropy method (Fig. 5). To address initial DNA binding rate,

A plot of maximal FRET signal versus DrRecF concentrations.

**4. Functional significance of ABC-type ATPase properties of RecF** 

**4.1 ATP-dependent dimerization is required for DNA binding** 

RecF solution conditions for other experiments.

ATP-binding constant of KdATP = 13 ± 2 µM.

reactions were performed for a relatively short time (10-15 min) and with the excess of ATP, taking an advantage of RecF being a slow ATPase (Fig. 6C, below). Alternatively, the rate of ATP hydrolysis was measured over 1 or 2 hours time upon titration of RecF by different DNA oligonucleotides (Fig. 6B). The binding of all DNA substrates was relatively weak with the apparent dissociation constants greater than 15 µM (Fig. 5). Neither a wild type DrRecF in the presence of ADP nor a signature motif mutant S279R in the presence of ATP were able to bind DNA (Fig. 5), suggesting that the ATP-dependent dimerization is essential for RecF interaction with all DNA substrates.

Fig. 5. ATP-dependent binding of DrRecF to different DNA substrates (top) and DNAdependent ATP hydrolysis rates (bottom). DNA substrates are schematically represented above each plot with **A**) ssDNA, **B**) dsDNA and **C**) ds/ssDNA junction. Solid isotherms correspond to binding in the presence of ATP, dashed black – in the presence of ADP, dotted – to the binding of signature motif mutant S279R in the presence of ATP. Red isotherms correspond to DrRecF binding in the presence of ATP and 50 μM DrRecR. The maximum estimated ATP hydrolysis rates of DrRecF (Fig. 6A) are shown at the bottom with the top lane corresponding to reactions without DrRecR and the bottom – with RecR. DrRecF concentration is 10 μM, DNA- 20 nM, ATP – 2 mM.

#### **4.2 RecR-dependent DNA specificity of RecF**

DNA binding of DrRecF is drastically alters in the presence of DrRecR (red isotherms in Fig. 5). DrRecR significantly increases the affinity of DrRecF to dsDNA (Fig. 5B) with the estimated association binding constant at least two orders of magnitude stronger than without DrRecR. DrRecR does not alter DrRecF ssDNA binding according to the DNA binding assay. However, the ATPase assay clearly demonstrated interaction of DrRecR with DrRecF in the presence of ssDNA. ssDNA does not stimulate ATP hydrolysis by DrRecF, while the presence of both DrRecR and ssDNA results in strongest ATPase rate. This suggests that DrRecR stimulates the ATPase rate of DrRecF bound to ssDNA, potentially destabilizing dimerization and ssDNA binding. In case of dsDNA, maximum ATPase rates were similar with and without DrRecR. Therefore, DrRecR stabilizes DrRecF complex with dsDNA without increasing its ATPase rate. Due to this stabilization effect of RecR, we are able to measure DNA binding and dimerization of DrRecF in the presence of ATP analogs (Fig. 6B). Curiously, a weak dimerization is observed at highest DrRecF concentration even in the presence of ADP. Therefore, DrRecR selectively stimulates binding of DrRecF dimer to dsDNA, while potentially destabilizing DrRecF complex with ssDNA. Both dimerization and DNA binding reactions were also measured as a function of time to verify that under

ATP-Binding Cassette Properties of Recombination Mediator Protein RecF 441

ATPase rates had an average between ss- and dsDNA substrates values. Although all data were obtained with *D. radiodurans* proteins, RecF and RecR are highly homologous proteins. Moreover, *E. coli* RecR stimulates DNA binding of *D. radiodurans* RecF similarly to that of *D. radiodurans* RecR suggesting that DrRecF binds both Dr- and *E. coli* RecR proteins with similar affinities (Makharashvili, 2009). Therefore, the described above properties of *D. radiodurans* proteins are likely to be conserved for *E. coli* homologs. While DNA binding and ATPase assays did not reveal specificity of RecF towards DNA junction, functional studies clearly evidence the role of RecF at ss/dsDNA junction (Chow and Courcelle, 2004; Handa et al., 2009; McInerney and O'Donnell, 2007; Morimatsu and Kowalczykowski, 2003; Webb et al., 1997). The potential specificity of RecF to ds/ssDNA junction is likely to require additional protein partners of recombination initiation reaction including SSB, RecO and RecA. For example, RecR can be recruited to SSB-bound ssDNA while in complex with RecO (Ryzhikov et al., 2011). The increased local concentration of RecR on SSB-coated ssDNA may subsequently stimulate RecF interaction with the adjacent dsDNA region.

Fig. 7. Schematic representation of RecF interaction with ATP, ATP-dependent dimerization, DNA binding, and the effect of RecR on DNA binding and ATP hydrolysis. The complex formed on dsDNA in the presence of RecR (red box) is the most stable intermediate. In case

The role of RecF SMC motifs *in vivo* was initially addressed with *E. coli RecF* mutant cells transformed with RecF-containing vector (Koroleva et al., 2007). Only wild type RecF complemented the UV sensitivity of a *recF* cells. Mutations of Walker A, -B and signature motifs did not restore the UV resistance. Since the overexpression of RecF can potentially affect its function, similar mutants of *RecF* were constructed in chromosome (Michel-Marks et al., 2010). Importantly, different steps of RecFOR function were tested with each mutant. Those include the rate of DNA synthesis, degradation of nascent DNA, the presence of DNA intermediates, and cell survival upon UV irradiation. Mutants included Walker A motif K36M, deficient in ATP binding, a Walker A motif K36R and a Walker B D303N, which both retain ATP binding but are deficient in ATP hydrolysis, and two signature motif mutants

Following the UV-induced arrest of replication, the nascent DNA is partially degraded at the replication fork by RecQ helicase and RecJ nuclease and RecF limits such degradation (Courcelle and Hanawalt, 1999). The degree of nascent DNA degradation was measured with pulse labeling of growing cell culture with [14C]thymine and [3H]thymidine. Similarly to a null mutant(Courcelle and Hanawalt, 1999), approximately 50% of nascent DNA was degraded with all mutants with the exception of D303N, where degradation was less severe.

of *D. radiodurans* homologs, RecF dimer interacts with RecR tetramer.

S270R and Q273A, which prevent an ATP-dependent dimerization.

**4.4 In vivo function of RecF conserved motifs** 

these conditions ATP hydrolysis does not significantly alter either interaction within first 10 minutes (Fig. 6C).

DrRecR is characterized by a weak DNA binding affinity in a millimolar range, while binding of *E. coli* RecR to DNA was not detected. DrRecR forms a tetrameric DNA clamplike structure (Lee et al., 2004). This conformation is likely to be conserved for other RecR homologs since *E. coli* RecR is either a dimer or tetramer in solution (Umezu et al., 1993), and *H. influenzae* RecR also was crystallized in a similar tetrameric conformation (Koroleva, O., Baranova, E., Korolev, S. unpublished data). One way to explain the DNA-dependent interaction of RecR with RecF is through the binding of both proteins to a shared DNA substrate, as beads on a string. Moreover, since dimer to tetramer transition was proposed as a clamp loading mechanism (although not confirmed), the ATP-dependent dimerization of RecF may stimulate such loading of RecR clamp on DNA. To test the hypothesis of shared DNA substrate requirement for RecF interaction with RecR, the RecR-stimulated DNA binding of RecF and the ATPase rate were tested in the presence of different length dsDNA substrates. Surprisingly, 10 bp short oligonucleotide stimulates DrRecF interaction with DrRecR. Structural modeling suggests that RecF dimer can bind 12-15 bp long DNA, while RecR clamp may cover up to 8-12 bp. These results rule out the beads-on-a-string model of RecFR binding to dsDNA. Alternatively, RecR may interact with the domain II arms encircling RecF bound DNA in a model similar to that of Rad50/Mre11 complex (Hopfner et al., 2001; Lammens et al., 2011; Williams et al., 2011).

Fig. 6. **A**) ATP hydrolysis by DrRecF over 120 min was measured upon titration by different DNA substrates, with circles corresponding to ssDNA, triangles to dsDNA, and squares to ds/ssDNA. Red symbols correspond to titrations in the presence of 50 μM RecR. Concentration of RecF is 10 μM, and ATP 2 mM. **B**) dsDNA binding by RecFR in the presence of ATP analogs measured with the fluorescence polarization assay performed similarly to that in Fig. 5 with the following nucleotides: ATP (red), ATPγS (green), AMPPNP (blue), and ADP (cyan). **C**) Time dependence of RecF dimerization upon titration with ATP as measured by FRET of labeled RecF. Isotherms of different colors correspond to the FRET value at different time points shown on the right.

#### **4.3 The lack of ss/dsDNA junction specificity**

The steps of RecF interaction with DNA and RecR are schematically represented in Fig. 7. ATP binding stimulates RecF dimerization, essential for binding of all DNA substrates. The DNA-bound RecF dimer interacts with RecR, which either stabilizes the complex with dsDNA or destabilizes with ssDNA. Importantly, neither of the performed assays revealed any specificity of RecF and RecFR complex for ss/dsDNA junction. Both DNA binding and

these conditions ATP hydrolysis does not significantly alter either interaction within first 10

DrRecR is characterized by a weak DNA binding affinity in a millimolar range, while binding of *E. coli* RecR to DNA was not detected. DrRecR forms a tetrameric DNA clamplike structure (Lee et al., 2004). This conformation is likely to be conserved for other RecR homologs since *E. coli* RecR is either a dimer or tetramer in solution (Umezu et al., 1993), and *H. influenzae* RecR also was crystallized in a similar tetrameric conformation (Koroleva, O., Baranova, E., Korolev, S. unpublished data). One way to explain the DNA-dependent interaction of RecR with RecF is through the binding of both proteins to a shared DNA substrate, as beads on a string. Moreover, since dimer to tetramer transition was proposed as a clamp loading mechanism (although not confirmed), the ATP-dependent dimerization of RecF may stimulate such loading of RecR clamp on DNA. To test the hypothesis of shared DNA substrate requirement for RecF interaction with RecR, the RecR-stimulated DNA binding of RecF and the ATPase rate were tested in the presence of different length dsDNA substrates. Surprisingly, 10 bp short oligonucleotide stimulates DrRecF interaction with DrRecR. Structural modeling suggests that RecF dimer can bind 12-15 bp long DNA, while RecR clamp may cover up to 8-12 bp. These results rule out the beads-on-a-string model of RecFR binding to dsDNA. Alternatively, RecR may interact with the domain II arms encircling RecF bound DNA in a model similar to that of Rad50/Mre11 complex (Hopfner et

Fig. 6. **A**) ATP hydrolysis by DrRecF over 120 min was measured upon titration by different DNA substrates, with circles corresponding to ssDNA, triangles to dsDNA, and squares to

AMPPNP (blue), and ADP (cyan). **C**) Time dependence of RecF dimerization upon titration with ATP as measured by FRET of labeled RecF. Isotherms of different colors correspond to

The steps of RecF interaction with DNA and RecR are schematically represented in Fig. 7. ATP binding stimulates RecF dimerization, essential for binding of all DNA substrates. The DNA-bound RecF dimer interacts with RecR, which either stabilizes the complex with dsDNA or destabilizes with ssDNA. Importantly, neither of the performed assays revealed any specificity of RecF and RecFR complex for ss/dsDNA junction. Both DNA binding and

ds/ssDNA. Red symbols correspond to titrations in the presence of 50 μM RecR. Concentration of RecF is 10 μM, and ATP 2 mM. **B**) dsDNA binding by RecFR in the presence of ATP analogs measured with the fluorescence polarization assay performed similarly to that in Fig. 5 with the following nucleotides: ATP (red), ATPγS (green),

the FRET value at different time points shown on the right.

**4.3 The lack of ss/dsDNA junction specificity** 

minutes (Fig. 6C).

al., 2001; Lammens et al., 2011; Williams et al., 2011).

ATPase rates had an average between ss- and dsDNA substrates values. Although all data were obtained with *D. radiodurans* proteins, RecF and RecR are highly homologous proteins. Moreover, *E. coli* RecR stimulates DNA binding of *D. radiodurans* RecF similarly to that of *D. radiodurans* RecR suggesting that DrRecF binds both Dr- and *E. coli* RecR proteins with similar affinities (Makharashvili, 2009). Therefore, the described above properties of *D. radiodurans* proteins are likely to be conserved for *E. coli* homologs. While DNA binding and ATPase assays did not reveal specificity of RecF towards DNA junction, functional studies clearly evidence the role of RecF at ss/dsDNA junction (Chow and Courcelle, 2004; Handa et al., 2009; McInerney and O'Donnell, 2007; Morimatsu and Kowalczykowski, 2003; Webb et al., 1997). The potential specificity of RecF to ds/ssDNA junction is likely to require additional protein partners of recombination initiation reaction including SSB, RecO and RecA. For example, RecR can be recruited to SSB-bound ssDNA while in complex with RecO (Ryzhikov et al., 2011). The increased local concentration of RecR on SSB-coated ssDNA may subsequently stimulate RecF interaction with the adjacent dsDNA region.

Fig. 7. Schematic representation of RecF interaction with ATP, ATP-dependent dimerization, DNA binding, and the effect of RecR on DNA binding and ATP hydrolysis. The complex formed on dsDNA in the presence of RecR (red box) is the most stable intermediate. In case of *D. radiodurans* homologs, RecF dimer interacts with RecR tetramer.

#### **4.4 In vivo function of RecF conserved motifs**

The role of RecF SMC motifs *in vivo* was initially addressed with *E. coli RecF* mutant cells transformed with RecF-containing vector (Koroleva et al., 2007). Only wild type RecF complemented the UV sensitivity of a *recF* cells. Mutations of Walker A, -B and signature motifs did not restore the UV resistance. Since the overexpression of RecF can potentially affect its function, similar mutants of *RecF* were constructed in chromosome (Michel-Marks et al., 2010). Importantly, different steps of RecFOR function were tested with each mutant. Those include the rate of DNA synthesis, degradation of nascent DNA, the presence of DNA intermediates, and cell survival upon UV irradiation. Mutants included Walker A motif K36M, deficient in ATP binding, a Walker A motif K36R and a Walker B D303N, which both retain ATP binding but are deficient in ATP hydrolysis, and two signature motif mutants S270R and Q273A, which prevent an ATP-dependent dimerization.

Following the UV-induced arrest of replication, the nascent DNA is partially degraded at the replication fork by RecQ helicase and RecJ nuclease and RecF limits such degradation (Courcelle and Hanawalt, 1999). The degree of nascent DNA degradation was measured with pulse labeling of growing cell culture with [14C]thymine and [3H]thymidine. Similarly to a null mutant(Courcelle and Hanawalt, 1999), approximately 50% of nascent DNA was degraded with all mutants with the exception of D303N, where degradation was less severe.

ATP-Binding Cassette Properties of Recombination Mediator Protein RecF 443

protein partners. For example, the ability of short DNA fragments to promote RecR binding suggests that the DNA-dependent conformational changes of RecF are important for

It is important to note that RecF does not represent the exact analog of Rad50. It is a much smaller protein without long coiled-coil structures. RecF does not support DNA unwinding or resection, as well as additional adenylate kinase activity of Rad50 and SMC proteins (Bhaskara et al., 2007; Lammens and Hopfner, 2010). Instead, it is involved in the initiation of the presynaptic complex formation, the function performed by BRCA2 or Rad52 in eukaryotes (Moynahan et al., 2001; New et al., 1998; Shinohara and Ogawa, 1998; Sung, 1997; Yang et al., 2005). While Rad52 is rather unique protein (Singleton et al., 2002), structural and functional motifs of BRCA2 resemble that of RecFOR system (Yang et al., 2002). BRCA2 interacts with ssDNA through OB-fold domain, similarly to RecO, and has a putative dsDNA-binding domain. The latter function is likely to be performed by RecF, even

RecF regulates RecQJ-dependent resection of nascent DNA at stalled replication fork (Courcelle and Hanawalt, 1999). This step occurs prior to RecA loading and initiation of SOS response. How RecF recognizes stalled replication remains unknown. It is tempting to speculate that RecF is a part of replisome (Kogoma, 1997) based on co-translation of RecF with replication initiation protein DnaA and polymerase subunit DnaN and on its early involvement in detection of replication arrest. However, no interactions of RecF with replication proteins have been identified so far. RecF may represent an alternative to PriA pathway of replication restart in case of arrested replication or postreplication repair (Sandler, 1996). Thus, it is important to find additional RecF-binding proteins. The detection of novel interactions is problematic due to low copy number of RecF in cells and poor solubility of purified RecF. The potential requirement of ATP- and DNA-dependent dimerization for RecF

interaction with other proteins further complicates the search for interacting proteins.

The relationship of specific steps of ATP-dependent reactions with the DNA damage recognition and processing by RecF and Rad50 remains elusive. Since RecF is the smallest known DNA-biding ABC ATPase composed of the head domain only, it represents an excellent model system to address the role of allosteric regulations, governing function of this class of proteins. Importantly, both ATP binding and hydrolysis are likely to play an important mechanistic role in most of reactions (Fig. 7). For example, the first step of limiting degradation of nascent lagging DNA by RecQJ and loading of RecA may only require formation of a stable RecF dimer at DNA junction, while ATP hydrolysis and dimer dissociation may be important for the following steps. However, the involvement of all the conserved motifs to prevent degradation of nascent DNA suggests that both ATP binding and hydrolysis are important even for this initial step. Therefore, all steps of RecF function in DNA repair are likely to depend on dynamic interactions of RecF with ATP, DNA and DNA repair proteins. Delineating molecular basis and principles of these interactions is essential for understanding fundamental mechanisms of DNA repair, recombination and replication.

The research was supported by National Institutes of Health grant GM073837. The author is grateful for help during manuscript preparation to Mikhail Ryzhikov and Ekaterina Los. Significant portion of RecF studies were performed by graduate student Nodar

protein-protein recognition rather than simple binding to the shared DNA substrate.

with the lack of structural similarity.

**6. Acknowledgment** 

Makharashivili.

Therefore, all steps of the dynamic interactions of RecF with ATP and DNA are important for the very first step of RecFOR function in replication repair. The weak functionality of D303N can be explained by a potential residual ATPase activity of this mutant, as shown for other SMC proteins (Lammens et al., 2004). Experiments with ATP analogs (Fig. 6B) demonstrated that even minor conformational changes significantly affect RecF properties. Therefore, an alternative explanation may be that D303N mutant introduces the least conformational distortion at the ATP-binding site and may retain conformation of a wild type wild type dimer and DNA-binding activities better than K36R mutant.

The rate of DNA synthesis is reduced by approximately 90% immediately after UV irradiation, but is recovered to nearly initial rate within 100 min in wild-type cells. The overall accumulation of DNA is increased at that time approaching the level of unirradiated cells. In *recF* cells the initial reduction of DNA synthesis rate is similar, but there is no recovery. Like in the previous assay, all mutants with exception of D303N were similar to the null mutant. D303N mutant did support slight recovery of DNA replication rate, yet it was significantly weaker than that of a wild type. RecF is associated with appearance of specific replication intermediates during DNA damage, as visualized on two-dimensional agarose gel (Courcelle et al., 2003). In this assay, all mutants were equally deficient in accumulation of such intermediates similarly to the null mutant, although the detection level of this assay may not be sufficient to reveal weak activity of D303N mutant. Finally, the survival rate of cell culture after UV irradiation was assayed. D303N mutant was partially resistant, while all other mutants were as hypersensitive to UV irradiation as deletion of *recF*. These studies demonstrate that all steps of ATP binding, dimerization and hydrolysis by RecF are important to maintain stalled replication and to restart cell growth after DNA damage.

#### **5. Conclusions**

RecFOR proteins regulate RecA binding to ssDNA under DNA damage conditions. This reaction initiates a variety of DNA repair pathways including maintenance and restart of stalled replication. Correspondingly, recombinational repair is tightly regulated in cell. While the exact role and mechanism of RecF in these pathways remain controversial, the majority of known data suggest a regulatory function of RecF during initiation and subsequent steps of recombinational DNA repair. Intricate properties of the ATP-dependent interaction of RecF with DNA and of the DNA-dependent ATP hydrolysis as well as the dependence of these interactions on RecR strongly supports this hypothesis.

Regulatory function is further reinforced by the sequence and structural homology with the head domain of Rad50 and SMC proteins. Rad50 is involved in multiple steps of DNA damage response including initial detection of DSBs, triggering of cell signaling cascades, and in resection of dsDNA to create 3' ssDNA tail for recombinase binding (Nicolette et al., 2010). In bacteria, RecF is likely to be involved in multiple steps of replication restart as well, including initial detection of replication arrest. Neither Rad50 nor RecF specifically recognizes functionally relevant DNA substrates, blunt-end DNA and ss/dsDNA junction, correspondingly (de Jager et al., 2002). Rad50 functions in complex with other DNA binding proteins, including Mre11 nuclease, and protein-protein interactions regulate DNA binding and ATPase activities (Lammens et al., 2011; Lim et al., 2011; Williams et al., 2011). By analogy, we can speculate that ATP binding and hydrolysis may not simply control DNA binding and dissociation of RecF, but also regulate binding of RecF dimer to different

Therefore, all steps of the dynamic interactions of RecF with ATP and DNA are important for the very first step of RecFOR function in replication repair. The weak functionality of D303N can be explained by a potential residual ATPase activity of this mutant, as shown for other SMC proteins (Lammens et al., 2004). Experiments with ATP analogs (Fig. 6B) demonstrated that even minor conformational changes significantly affect RecF properties. Therefore, an alternative explanation may be that D303N mutant introduces the least conformational distortion at the ATP-binding site and may retain conformation of a wild

The rate of DNA synthesis is reduced by approximately 90% immediately after UV irradiation, but is recovered to nearly initial rate within 100 min in wild-type cells. The overall accumulation of DNA is increased at that time approaching the level of unirradiated cells. In *recF* cells the initial reduction of DNA synthesis rate is similar, but there is no recovery. Like in the previous assay, all mutants with exception of D303N were similar to the null mutant. D303N mutant did support slight recovery of DNA replication rate, yet it was significantly weaker than that of a wild type. RecF is associated with appearance of specific replication intermediates during DNA damage, as visualized on two-dimensional agarose gel (Courcelle et al., 2003). In this assay, all mutants were equally deficient in accumulation of such intermediates similarly to the null mutant, although the detection level of this assay may not be sufficient to reveal weak activity of D303N mutant. Finally, the survival rate of cell culture after UV irradiation was assayed. D303N mutant was partially resistant, while all other mutants were as hypersensitive to UV irradiation as deletion of *recF*. These studies demonstrate that all steps of ATP binding, dimerization and hydrolysis by RecF are important to maintain stalled replication and to restart cell growth

RecFOR proteins regulate RecA binding to ssDNA under DNA damage conditions. This reaction initiates a variety of DNA repair pathways including maintenance and restart of stalled replication. Correspondingly, recombinational repair is tightly regulated in cell. While the exact role and mechanism of RecF in these pathways remain controversial, the majority of known data suggest a regulatory function of RecF during initiation and subsequent steps of recombinational DNA repair. Intricate properties of the ATP-dependent interaction of RecF with DNA and of the DNA-dependent ATP hydrolysis as well as the

Regulatory function is further reinforced by the sequence and structural homology with the head domain of Rad50 and SMC proteins. Rad50 is involved in multiple steps of DNA damage response including initial detection of DSBs, triggering of cell signaling cascades, and in resection of dsDNA to create 3' ssDNA tail for recombinase binding (Nicolette et al., 2010). In bacteria, RecF is likely to be involved in multiple steps of replication restart as well, including initial detection of replication arrest. Neither Rad50 nor RecF specifically recognizes functionally relevant DNA substrates, blunt-end DNA and ss/dsDNA junction, correspondingly (de Jager et al., 2002). Rad50 functions in complex with other DNA binding proteins, including Mre11 nuclease, and protein-protein interactions regulate DNA binding and ATPase activities (Lammens et al., 2011; Lim et al., 2011; Williams et al., 2011). By analogy, we can speculate that ATP binding and hydrolysis may not simply control DNA binding and dissociation of RecF, but also regulate binding of RecF dimer to different

dependence of these interactions on RecR strongly supports this hypothesis.

type wild type dimer and DNA-binding activities better than K36R mutant.

after DNA damage.

**5. Conclusions** 

protein partners. For example, the ability of short DNA fragments to promote RecR binding suggests that the DNA-dependent conformational changes of RecF are important for protein-protein recognition rather than simple binding to the shared DNA substrate.

It is important to note that RecF does not represent the exact analog of Rad50. It is a much smaller protein without long coiled-coil structures. RecF does not support DNA unwinding or resection, as well as additional adenylate kinase activity of Rad50 and SMC proteins (Bhaskara et al., 2007; Lammens and Hopfner, 2010). Instead, it is involved in the initiation of the presynaptic complex formation, the function performed by BRCA2 or Rad52 in eukaryotes (Moynahan et al., 2001; New et al., 1998; Shinohara and Ogawa, 1998; Sung, 1997; Yang et al., 2005). While Rad52 is rather unique protein (Singleton et al., 2002), structural and functional motifs of BRCA2 resemble that of RecFOR system (Yang et al., 2002). BRCA2 interacts with ssDNA through OB-fold domain, similarly to RecO, and has a putative dsDNA-binding domain. The latter function is likely to be performed by RecF, even with the lack of structural similarity.

RecF regulates RecQJ-dependent resection of nascent DNA at stalled replication fork (Courcelle and Hanawalt, 1999). This step occurs prior to RecA loading and initiation of SOS response. How RecF recognizes stalled replication remains unknown. It is tempting to speculate that RecF is a part of replisome (Kogoma, 1997) based on co-translation of RecF with replication initiation protein DnaA and polymerase subunit DnaN and on its early involvement in detection of replication arrest. However, no interactions of RecF with replication proteins have been identified so far. RecF may represent an alternative to PriA pathway of replication restart in case of arrested replication or postreplication repair (Sandler, 1996). Thus, it is important to find additional RecF-binding proteins. The detection of novel interactions is problematic due to low copy number of RecF in cells and poor solubility of purified RecF. The potential requirement of ATP- and DNA-dependent dimerization for RecF interaction with other proteins further complicates the search for interacting proteins.

The relationship of specific steps of ATP-dependent reactions with the DNA damage recognition and processing by RecF and Rad50 remains elusive. Since RecF is the smallest known DNA-biding ABC ATPase composed of the head domain only, it represents an excellent model system to address the role of allosteric regulations, governing function of this class of proteins. Importantly, both ATP binding and hydrolysis are likely to play an important mechanistic role in most of reactions (Fig. 7). For example, the first step of limiting degradation of nascent lagging DNA by RecQJ and loading of RecA may only require formation of a stable RecF dimer at DNA junction, while ATP hydrolysis and dimer dissociation may be important for the following steps. However, the involvement of all the conserved motifs to prevent degradation of nascent DNA suggests that both ATP binding and hydrolysis are important even for this initial step. Therefore, all steps of RecF function in DNA repair are likely to depend on dynamic interactions of RecF with ATP, DNA and DNA repair proteins. Delineating molecular basis and principles of these interactions is essential for understanding fundamental mechanisms of DNA repair, recombination and replication.
