Reengineering the nucleotide cofactor specificity of the RecA protein by mutation of aspartic acid 100.

We have recently obtained evidence for a direct linkage between the S0.5 (S0.5 is the substrate concentration required for half-maximal velocity) value of a nucleoside triphosphate and the conformational state of the RecA-ssDNA complex, with an S0.5 value of 125 μM or less required for stabilization of the strand exchange-active conformation. For example, although ATP and ITP are hydrolyzed by the RecA protein with the same turnover number (18 min−1), ATP (S0.5 = 45 μM) functions as a cofactor for the strand exchange reaction, whereas ITP (S0.5 = 500 μM) is inactive as a strand exchange cofactor. The RecA protein crystal structure suggests that cofactor specificity is determined by Asp100, which likely forms a hydrogen bond with the exocyclic 6-amino group of ATP; the higher S0.5 value for ITP is presumably due to unfavorable interactions between Asp100 and the 6-carbonyl group of the inosine ring. To test this hypothesis, we prepared a mutant RecA protein in which Asp100 was replaced by an asparagine residue. The S0.5(ITP) for the [D100N]RecA protein is 125 μM, indicating favorable interactions between the Asn100 side chain and the 6-carbonyl group of ITP. Correspondingly, ITP functions as a cofactor for the strand exchange activity of the [D100N]RecA protein. This result demonstrates the importance of the residue at position 100 in determining nucleotide cofactor specificity and underscores the importance of the S0.5 value in the RecA protein-promoted strand exchange reaction.

The RecA protein of Escherichia coli (M r 37,842, 352 amino acids) is essential for homologous genetic recombination and for the postreplicative repair of damaged DNA. The purified RecA protein will promote a variety of DNA pairing reactions that presumably reflect in vivo recombination functions. The most extensively investigated DNA pairing activity is the ATPdependent three-strand exchange reaction, in which a circular ssDNA 1 molecule and a homologous linear dsDNA molecule are recombined to yield a nicked circular dsDNA molecule and a linear ssDNA molecule. This reaction proceeds in three phases. In the first phase, the circular ssDNA substrate is coated with RecA protein to form a presynaptic complex; this complex will catalyze the hydrolysis of ATP to ADP and P i . In the second phase, the presynaptic complex interacts with a dsDNA molecule, the homologous sequences are brought into register, and pairing between the circular ssDNA and the complementary strand from the dsDNA is initiated. In the third phase, the complementary linear strand is completely transferred to the circular ssDNA by unidirectional branch migration to yield the nicked circular dsDNA and displaced linear ssDNA products (Roca and Cox, 1990;Kowalczykowski et al., 1994).
The presynaptic complex formed between RecA protein and ssDNA is the active recombinational entity in the strand exchange reaction. The RecA protein binds cooperatively to ssDNA, forming a right-handed helical protein filament with one RecA monomer per four nucleotides of ssDNA and six RecA monomers per turn of the filament. In the absence of nucleotide cofactor or in the presence of ADP, the helical filament adopts a "collapsed" or "closed" conformation (helical pitch: 65 Å) that is inactive in strand exchange. In the presence of ATP or the nonhydrolyzable ATP analog, ATP␥S, however, the filament assumes an "extended" or "open" conformation (helical pitch: 95 Å) that is active in strand exchange (Egelman, 1993).
We have been examining the mechanism of the nucleotide cofactor-mediated isomerization of the RecA-ssDNA complex and have identified a linkage between the S 0.5 value 2 of a nucleoside triphosphate and the conformational state of the RecA-ssDNA complex (Menge and Bryant, 1992;Meah and Bryant, 1993;Bryant, 1994, 1995). These studies have shown that a nucleoside triphosphate must have an S 0.5 value of 100 -120 M or lower in order to stabilize the strand exchange-active conformation of the RecA-ssDNA complex. For example, although ATP and ITP are hydrolyzed by the RecA protein with identical turnover numbers (18 min Ϫ1 ), ATP (S 0.5 ϭ 45 M) functions as a cofactor for the strand exchange reaction, whereas ITP (S 0.5 ϭ 500 M) is inactive as a strand exchange cofactor. The x-ray crystal structure of the RecA protein-ADP complex indicates that cofactor specificity is determined by Asp 100 , which forms a hydrogen bond with the exocyclic 6-amino group of adenosine base (Story et al., 1992); it seems likely that a similar contact is made with ATP in the RecA-ssDNA-ATP complex ( Fig. 1), although no structural information is available for this complex. The higher S 0.5 value for ITP, relative to that for ATP, is presumably due to unfavorable interactions between the negatively charged Asp 100 side chain and the 6-carbonyl group of the inosine ring. To test this hypothesis, we prepared a mutant RecA protein in which Asp 100 was replaced by an asparagine residue. The effect of this mutation on the nucleotide cofactor specificity of the RecA protein is described in this report.

EXPERIMENTAL PROCEDURES
Materials-Wild-type RecA protein was prepared as described previously (Cotterill et al., 1982). ATP and ITP were from Sigma. [␣-32 P]ATP and [␥-32 P]ATP were from ICN. [␥-32 P]ITP was prepared from IDP using [␥-32 P]ATP and nucleoside diphosphate kinase (Sigma) as described previously (Menge and Bryant, 1992). E. coli SSB was from Pharmacia Biotech Inc.. Circular X ssDNA ((ϩ)-strand) and circular X dsDNA were from New England Biolabs; linear X dsDNA was prepared from circular X dsDNA as described (Cox and Lehman, 1981). Single-and double-stranded DNA concentrations were determined by absorbance at 260 nm using the conversion factors 36 and 50 g/ml/A 260 , respectively. All DNA concentrations are expressed as total nucleotides.
Preparation of [D100N]RecA Protein-The mutant [D100N]RecA gene was produced by the polymerase chain reaction (PCR)-based overlap extension method essentially as described (Ho et al., 1989). The template for RecA mutagenesis consisted of a pUC19 vector containing a 1300-base pair E. coli DNA fragment carrying the wild-type recA gene and promoter cloned into a BamHI/HindIII site; the mutagenesis primers were 5Ј-CACGCGCTGAACCCAATCTACG-3Ј and 5Ј-CGTAGATT-GGGTTCAGCGCGTG-3Ј (the codon for Asn 100 is underlined and the nucleotide mismatch is in bold). The resulting PCR fragment containing the mutant RecA gene and promoter region was cloned into pUC19 to yield the plasmid, pUCrecA (D100N). The entire [D100N]RecA gene and promoter region was sequenced to confirm that only the desired change had been introduced during the mutagenesis procedure. pUCrecA (D100N) was then introduced into the RecA deletion strain, BNN124, and the mutant [D100N]RecA protein was expressed and purified by methods that have been described previously (Bryant, 1988).

RESULTS AND DISCUSSION
Experimental Design-In an effort to reengineer the nucleotide cofactor specificity of the RecA protein, we prepared a new mutant RecA protein in which aspartic acid 100 was replaced by an asparagine residue. The expectation was that the D100N mutation would allow the protein to form hydrogen bonding interactions with the 6-carbonyl group of ITP that are not possible in the wild-type protein (Fig. 1). If the new interactions resulted in a significant decrease in the S 0.5 value for ITP, we predicted that the mutation would convert the RecA protein into an ITP-activated DNA recombinase. The purified [D100N]RecA protein is shown in Fig. 2.
ssDNA-dependent NTP Hydrolysis Activity of the [D100N]RecA Protein-The ssDNA-dependent hydrolysis of ATP and ITP by the wild-type and [D100N]RecA protein was analyzed under standard reaction conditions (pH 7.5, 37°C). The dependence of the rate of ssDNA-dependent NTP hydrolysis on NTP concentration is shown in Fig. 3, and the steadystate kinetic parameters for the hydrolysis of each NTP are presented in Table I.
The turnover number (V max /[E t ]) for ssDNA-dependent ATP hydrolysis by the [D100N]RecA protein was 18 min Ϫ1 , a value identical to that obtained for the wild-type RecA protein. The S 0.5 (ATP) for the [D100N]RecA protein was 85 M, approximately 2-fold higher than the value of 45 M for the wild-type RecA protein. The ATP saturation curves were sigmoidal and identical Hill coefficients (n H ) of 3 were obtained for both pro-teins, indicating that both proteins are subject to positive cooperativity with respect to ATP concentration. These results show that the D100N mutation has a minimal effect on the steady-state kinetics of ssDNA-dependent ATP hydrolysis by the RecA protein.
The turnover number for ssDNA-dependent ITP hydrolysis by the [D100N]RecA protein was also 18 min Ϫ1 , a value again equivalent to that obtained for the wild-type protein. Moreover, initiated by the addition of enzyme and were carried out at 37°C. ATP and ITP hydrolysis reactions were measured using a thin-layer chromatography method as described previously (Weinstock et al., 1979). The points represent the initial rates of ATP or ITP hydrolysis measured at the indicated concentrations of NTP. The solid lines represent fits of the data to the Hill equation.

TABLE I
Kinetic parameters for wild-type and [D100N]RecA protein-catalyzed NTP hydrolysis (pH 7.5) The steady state kinetic parameters were derived from the data presented in Fig. 3. Wild type  ATP  18  45  3  ITP  19  500  3   D100N  ATP  18  85  3  ITP  18 125 3 [D100N]RecA Protein the ITP saturation curves were sigmoidal and identical Hill coefficients of 3 were obtained for both proteins, indicating that both proteins are subject to positive cooperativity with respect to ITP concentration. However, the S 0.5 (ITP) for the [D100N]RecA protein was 125 M compared with a value of 500 M for the wild-type protein. Thus, the D100N mutation results in a 4-fold decrease in the S 0.5 (ITP), presumably due to more favorable interactions of the asparagine side chain with the 6-carbonyl group of the inosine ring. Three-strand Exchange Activity of the [D100N]RecA Protein-The three-strand exchange activity of the [D100N]RecA protein was evaluated under standard reaction conditions (pH 7.5, 37°C). In the three-strand exchange assay, a circular X ssDNA molecule and a linear X dsDNA molecule are recombined to form a nicked circular dsDNA molecule and a linear ssDNA molecule; the substrates and products of this reaction are readily monitored by agarose gel electrophoresis (Cox and Lehman, 1981).
As shown in Fig. 4, the wild-type RecA protein has full strand exchange activity in the presence of ATP (3 mM), but has no detectable activity in the presence of ITP (3 mM), consistent with our previous results (Menge and Bryant, 1992). In contrast, the [D100N]RecA protein exhibited full strand exchange activity in the presence of either ATP or ITP (3 mM). Thus, the D100N mutation converts the RecA protein into an ITP-activated DNA recombinase.

SUMMARY AND CONCLUSIONS
Our results show that it is possible to alter the nucleoside triphosphate cofactor specificity of the RecA protein by mutating the aspartic acid residue at position 100 of the RecA polypeptide. This finding confirms that the hydrogen bonding interaction between Asp 100 and the 6-amino group of the adenosine ring that is apparent in the crystal structure of the RecA-ADP complex (which presumably represents the strand exchange-inactive closed conformation of the protein) also contributes to nucleotide cofactor binding in the strand exchangeactive open conformation of the RecA-ssDNA complex.
Furthermore, our results show that the D100N mutation changes the S 0.5 value for nucleoside triphosphates, without affecting the turnover number for NTP hydrolysis or the cooperativity of binding of NTP to the RecA-ssDNA complex. This indicates that the residue at position 100 contributes only to cofactor recognition and not to the catalytic reaction and provides evidence to relate the S 0.5 value determined from the steady-state kinetics of ssDNA-dependent NTP hydrolysis with the binding affinity of the nucleotide cofactor for the NTP active site of the protein. The reduction of the S 0.5 (ITP) from 500 to 125 M upon mutation of Asp 100 to asparagine indicates that the 6-carbonyl group of ITP interacts more favorably with the Asn 100 side chain in the mutant protein than with the negatively charged Asp 100 side chain in the wild-type protein.
The small increase in the S 0.5 (ATP) that results from the mutation indicates that the 6-amino group of the adenosine ring can interact favorably with either an aspartic acid or asparagine at position 100, with the hydrogen bond between the negatively charged aspartic acid side chain being somewhat stronger than that with the neutral asparagine side chain.
The finding that the [D100N]RecA protein is able to use ITP as a cofactor for the three-strand exchange reaction also supports our proposal that it is not sufficient for an NTP cofactor to simply bind to and be hydrolyzed by the RecA protein, but that it also must bind with certain minimal affinity (S 0.5 of 125 M or lower) in order to support isomerization of the RecA-ssDNA complex to the strand exchange-active open conformation.