Photoaffinity Labeling of the recBCD Enzyme of Escherichia coli with 8-Azidoadenosine 5 ’-Triphosphate *

The recB and recD subunits of the recBCD enzyme (exonuclease V) from Escherichia coli were covalently photolabeled with the ATP photoaffinity analogue [as2P]8-azido-ATP. The labeling was specific for ATP binding sites by the following criteria. (i) Saturation occurs at high 8-azido-ATP concentrations with dissociation constants of 30 and 120 PM for the recD and recB subunits, respectively; (ii) ATP strongly inhibits the photolabeling; (iii) 8-azido-ATP is hydrolyzed by the recBCD enzyme and supports its double-stranded DNA exonuclease activity; and (iv) the label is largely confined to two peptides obtained by tryptic digestion of the photolabeled holoenzyme; one is derived from the recB subunit and the other from the recD subunit.

subunits can be labeled by this analogue. The results of titration and nucleotide inhibition experiments, the ability of 8-N3-ATP to serve as a substrate for the recBCD enzyme, and peptide mapping of the covalent modification sites all indicate that both sites are of high affinity and specificity and, therefore, likely to be of importance in the catalytic mechanism of the enzyme.

EXPERIMENTAL PROCEDURES
Materiak-Unlabeled 8-N3-ATP was obtained from Sigma. A stock solution was made up to 30 mM in 10 mM Tris.HC1 (pH 8.0), 1 mM EDTA, and stored at -80 "C. The 8-Na-ATP concentration was determined by its absorbance at 280 nm using an extinction coefficient of 13.1 mM" cm-I (Haley and Hoffman, 1974). [ c Y -~* P ]~-N~-ATP was purchased from ICN and ATP-yS from Boehringer-Mannheim. SephacryM300 and ATP-agarose (ATP attached through the N6-amino group) were from Sigma. Heparin-agarose, prepared according to Davison et al. (1979) was a gift from J. Kelly of this department.
pMOB45 plasmid DNA was a gift from Hisao Masai, DNAX Research Institute. Tritium-labeled supercoiled M13mp8 DNA (30,000-90,000 cpm/nmol) was prepared as described for Ml3Goril (Julin et al., 1986). Tritium-labeled E. coli DNA was prepared from strain K37. Cultures grown to saturation were lysed with lysozyme and Triton X-100 as in Davis et al. (1980), except that the lysis mixture was vortexed vigorously for 1 min to release the chromosomal DNA from the cell debris after the Triton X-100 treatment. The DNA was purified from the lysate by two CsCl/ethidium bromide equilibrium centrifugation steps as described in Davis et al. (1980). This procedure yielded 800 pg of DNA/liter of culture with a specific activity of 45,000 cpm/nmol (nucleotide). The product was sonicated for 30 s to reduce it to small fragments (average size 5 1000 nucleotides).
Aldolase (rabbit muscle, type IV) and ovalbumin were purchased from Sigma. The aldolase suspension in 2.5 M ammonium sulfate was dialyzed overnight uersus 1000 volumes of 10 mM Tris.HC1 (pH 8.0), 1 mM EDTA, and the protein concentration was determined by measuring the absorbance at 280 nm, using E% = 9.38 (Rose and O'Connell, 1977). Trypsin (type TRTPCK, 219 units/mg, tosylphenylalanyl chloromethyl ketone treated) was from Worthington. BamHI restriction endonuclease was from New England Biolabs.
Clones containing the desired plasmid were isolated by selecting for the ability to grow on plates containing chloramphenicol (5 pg/ml) but not tetracycline (5 pglml). The structures of these plasmids were determined by restriction endonuclease analysis to verify the presence of the 19.5-kilobase BamHI fragment. The ATP-dependent doublestranded exonuclease activity of the recBCD enzyme was overpro-

9044
Photoaffinity Labeling of recBCD Enzyme 9045 duced by as much as 80-fold upon temperature-shift induction of plasmid replication. The ATP-dependent double-stranded exonuclease activity of the recBCD enzyme was assayed during the purification by the trichloroacetic acid precipitation assay described by Eichler and Lehman (1977) (Maniatas et al., 1982) containing 0.5% cerelose and 5 pg/ml chloramphenicol at 30 "C to an absorbance at 595 nm of 0.42, at which time the temperature was shifted to 42 "C to induce plasmid replication. Growth was continued for an additional 4 h, and the cells were harvested and frozen in liquid nitrogen.
Cell lysis and ammonium sulfate precipitation were carried out as described by Dykstra et al. (1984), except that phenylmethylsulfonyl fluoride and sodium bisulfite were not added. The redissolved ammonium sulfate pellet was chromatographed on DEAE-cellulose (Whatman) as in Eichler and Lehman (1977), followed by chromatography on Sephacryl-S300 and heparin-agarose as described by Dykstra et al. (1984). RecBCD enzyme-containing fractions from the heparin-agarose column were concentrated by ultrafiltration (YM-10 membrane, Amicon), dialyzed uersus buffer C (20 mM potassium phosphate, pH 6.9, 0.1 mM EDTA, 0.1 mM DTT, 50% glycerol (v/v)), and stored frozen at -80 "C. The enzyme was purified further by ATP-agarose chromatography. ATP-agarose (1.73 g) was washed twice in 50 ml of buffer D (10 mM potassium phosphate, pH. 6.7, 0.1 mM EDTA, 0.1 mM DTT, 10% (v/v) glycerol), poured into a column (24-ml bed volume), and washed with 250 ml of buffer D. A portion of the dialyzed heparin-agarose pool (1.2 ml, 8.2 mg of protein) in buffer C was diluted to 6 ml with 0.2 M potassium phosphate, pH 6.7, to convert the buffer conditions to those of buffer D. This material was applied to the column, which was then washed with 100 ml of buffer D. The protein content of the eluate was monitored by Na-DodS04-polyacrylamide gel electrophoresis. The gels were silver stained using Rapid-Ag-Stain (ICN) as described in the manufacturer's instructions. Contaminating protein was not retained by the column, and recBCD enzyme was eluted in the buffer D wash. Fractions containing recBCD enzyme were concentrated as above, dialyzed uersus buffer C, and stored at -80 "C. The yield was 23%. The enzyme was >95% pure, as judged from Coomassie Blue staining of 9 pg of protein electrophoresed through a NaDodS04-7.5% polyacrylamide gel.
Photolubeling-Standard buffer conditions for the 8-N3-ATP photolabeling reactions were 25 mM PIPES (pH 7) and 10 mM MgC12, with concentrations of 8-N3-ATP, proteins, and other additions as described in the relevant figure legends. [a-3ZP]8-N3-ATP was pipetted into a 1.5-ml microcentrifuge tube and the methanol solvent evaporated with a stream of air. Buffer and other reaction components were then added and the mixture incubated on ice for 30 s. UV irradiation was then performed using the short wavelength lamp of a hand-held Minerallight (Brinkmann Instruments, model UVSL-25, 254 nm, 200 microwatts/cm' at a distance of 6 inches) at a distance of 4 cm from the bottom of the reaction tube. Fluorescent room lights were kept off during the experiment. Photolabeling mixtures, or aliquots removed therefrom, were quenched by the addition of DTT (10 mM) and prepared for gel electrophoresis by the addition of 0.1 volume of 0.5 M Tris-HC1 (pH 6.8), 10% NaDodS04, 0.025% bromphenol blue, and glycerol.
Autoradiography of dried gels was done at -80 "C using XAR-5 xray film (Kodak) and a Cronex Lightning Plus intensifying screen (Du Pont). Films were preflashed before exposure when band intensities were to be quantitated. Densitometry of preflashed films was done with a Hoefer Scientific Instruments GS300 scanning densitometer. Peak areas were determined by cutting-and-weighing peaks from the chart recorder traces or by integration with a Hewlett-Packard 3390A integrator. In some experiments, protein bands were cut from the dried gels and radioactivity determined by scintillation counting using a toluene-based scintillation fluid.
Trypsin Digestion of Photolabeled recBCD Enzyme-RecBCD enzyme (0.088 mg in 100 p1 of buffer C (see above)) was diluted 10-fold with 25 mM PIPES (pH 7.0) and concentrated to 30-50 pl by centrifugation in a Centricon 30 filter (Amicon). The dilution and concentration were repeated and the enzyme recovered in a total volume of about 100 pl. PIPES (pH 7.0) and MgCl, were added to 25 and 10 mM, respectively, and [cx-~'P]~-N~-ATP (1350 pCi/pmol) was added at 36 p~. The solution was irradiated on ice for 1 min as described above, and DTT was added to 10 mM.
The irradiated solution was diluted 10-fold into 1% ammonium bicarbonate (pH 8.0) and concentrated by centrifugation in Centricon 30 filters. This was performed three times, and the protein was recovered in a total volume of about 100 pl. Two aliquots of trypsin (10 pg each) were added at 3-h intervals, and the mixture was incubated at 37 "C for a total of 6-8 h. NaDodSO, (1% w/v) and DTT (20 mM) were added, and the mixture was refrigerated until HPLC analysis (see below).
When the photolabeled recB and recD subunits were separated before tryptic digestion (Fig. 14), one-half of the mixture was treated as described above for the holoenzyme, while the remaining half was subjected to NaDodS0,-polyacrylamide gel electrophoresis. The recB and recD bands were cut from the destained gel, shaken for 5 min at 37 "C in 25% isopropyl alcohol, followed by 10% methanol and then 1% ammonium bicarbonate, pH 8.0 (4 X 1 ml each), and the slices were minced by forcing them through a fine steel mesh. Ammonium bicarbonate (pH 8.0) (0.8 ml of 1%) and trypsin (8 pl of 5 mg/ml) were added and the samples incubated at 37 'C. Two more additions of trypsin were made at 3-h intervals, and NaDodSO, was added to 0.1% (w/v) after a total digestion time of 11 h. The peptides were separated from the gel fragments by centrifugation through Centricon 30 filters. The fragments were washed twice on the filter with 0.6 ml of water. The combined filtrates were concentrated almost to dryness by centrifugation under vacuum in a Speed-vac concentrator and taken up again in 80 pl of water and 10 mM DTT.
HPLC Analysis of Tryptic Digests-Reversed phase chromatography was done using a Vydac C-18 column. Elution buffers were: A, 0.1% (v/v) trifluoroacetic acid; and B, 0.1% (v/v) trifluoroacetic acid and 70% (v/v) acetonitrile. Trypsin-digested [ c x -~' P ]~-N~-A T P -~~beled peptides were injected onto the column (equilibrated in 100% A) and eluted by (i) 5 min of 100% buffer A, (ii) a 90-min linear gradient of 0-60% buffer B, and (iii) 10 min of 40% A, 60% B, at a total flow rate of 1 ml/min. Peptides were detected in the eluate by continuous monitoring of the absorbance at 214 nm. Fractions were collected at 0.6-min intervals, and the radioactivity was determined by Cerenkov counting of the entire fraction.
Two-dimensional Thin Layer Chromatography-Electrophoresis of Peptides Purified by HPLC-Two-dimensional thin layer chromatography was carried out essentially as described by Gracy (1977). Peak fractions containing radioactivity from the HPLC runs were pooled, evaporated to dryness in the Speed-vac concentrator, and redissolved in 10 pl of TLC buffer I (acetic acid/formic acid/water, 15:5:80 (v/v/ v)) containing 2% methyl green dye. The samples were applied near one corner of a polygram CEL400 (Brinkmann Instruments) cellulose TLC plate (20 X 17.5 cm). Electrophoresis was performed in the first dimension in buffer I at 800 volts with the sample spot placed nearest the anode. The plates were then dried thoroughly and chromatographed in the second dimension in butanol/pyridine/acetic acid/ water, 32.5:25:5:20, v/v/v/v. Radioactive peptides were visualized by autoradiography.
Separation of r e d Subunit from recBC Compkx-The recD subunit was separated from the recBC complex as described by Amundsen et al. (1986). RecBCD enzyme (0.044 mg) and ovalbumin (0.1 mg) were treated with 4 M NaCl at 0 "C for 2.5 h and sedimented by centrifugation for 39 h at 49,000 rpm in a SW 50.1 rotor, 4 "C, through a 10-30% glycerol gradient containing 3 M NaCl and 0.1 mg/ml ovalbumin. Reconstitution of the double-stranded exonuclease activity was also performed as described by Amundsen et al. (1986).

Photoaffinity
Labeling of recB and recD Subunits with 8-N3-ATP-The recBCD enzyme was photolabeled by UV irradiation at 0 "C in the presence of [cY-~'P]~-N,-ATP. As shown in Fig. 1, the recB and recD subunits were strongly labeled after irradiation times of up to 30 s, whereas much less label was incorporated into the recC subunit (lanes 2-5). Aldolase which was used to control for nonspecific labeling, since it does not interact with ATP or other nucleotides, showed only low levels of incorporation. No labeling occurred without UV irradiation ( l a n e 1) even after incubations as long as 5 min (not shown). As judged by densitometer scanning of the autoradiogram, labeling of the recB and recD subunits was 7-and 13-fold greater, respectively, than that of aldolase, IRRADIATION TIME (sed 0 5 10 20 30 on a weight basis, after 30 s of irradiation (more than 20 times more label appeared in recB and recD than in aldolase on a molar basis, calculated using molecular weights of 40,000 (aldolase), 134,000 (recB, Finch et al., 1986a), and 67,000 (recD, Finch et al., 198613) (Fig. 2). The extent of labeling was as great as 50%, depending on irradiation time, as determined by liquid scintillation counting of bands cut out from the dried gel (data not shown). Both the recB and recD subunits were also photolabeled at 25 "C, although the extent of incorporation was slightly lower than when the experiment was performed at 0 "C (data not shown). Labeling of the recD subunit was virtually abolished by the addition of NaDodSO, (0.1%) or by preheating the mixture for 2 min at 100 "C prior to irradiation (data not shown). These treatments reduced the labeling of the recB subunit only slightly and led to increased  (lanes 7-12) [a-"P]8-N3-ATP (218 pCi/pmol) and recBCD enzyme (0.022 mg/ml) in a total volume of 20 pl. Sonicated calf thymus DNA was added at 0.15 mg/ml (lanes 2 and 8), 0.36 mg/ml (lanes 3 and 9), and 0.73 mg/ml (lanes 5 and 11). Cytidine was added at 1.24 mM (lanes 4 and 10) or 2.49 mM (lanes 6  and 12). Irradiation was for 30 s on ice, as described under "Experimental Procedures." Effect of DNA on Photolubeling-The extent of photolabeling was reduced substantially by high concentrations of sonicated calf thymus DNA (Fig. 3). Cytidine at a concentration giving an absorbance at 280 nm identical to that of the DNA also inhibited photolabeling (Fig. 3) indicating that the inhibition by DNA is at least partially due to the reduced UV flux caused by the high optical density (A2BO = 16 cm") of the solution. Cytidine at these concentrations did not affect the DNA-dependent 8-N3-ATPase activity of the recBCD enzyme (not shown).

Photoaffinity Labeling of recBCD Enzyme
Enzymatic Activities of recBCD Enzyme with 8-N3-ATP us Substrate-The ability of 8-N3-ATP to serve as a substrate for the ATPase and as a co-substrate for the double-stranded exonuclease activity of the recBCD enzyme was measured to determine whether the photolabeling detects interactions between 8-N3-ATP and specific binding sites on the recB and recD subunits. Azido-ATP supported the double-stranded exonuclease activity of the recBCD enzyme, although not as effectively as ATP at low concentrations (Fig. 4), and was hydrolyzed in a DNA-dependent reaction (Fig. 5). (The ap-  , 1977).) The double-stranded exonuclease activity was reduced more than %fold when the recBCD enzyme and 8-N3-ATP (1 mM) were irradiated together and then diluted into the assay mixture, whereas the recBCD enzyme and 8-N3-ATP irradiated separately and then mixed to the same final concentration had the same activity as the unirradiated recBCD enzyme or enzyme that had been irradiated without 8-N3-ATP (Fig. 6). No inhibition of the doublestranded exonuclease was observed when the recBCD enzyme and 8-N3-ATP were mixed but not irradiated before being diluted into the double-stranded exonuclease assay mixture (Fig. 6). Titration of recBCD Enzyme with 8-N3-ATP-Photolabeling of the recBCD enzyme at increasing concentrations of 8-  10 (lanes 1 and 2), 20 (lanes 3 and 4). 40 (lanes 5 and 6),75 (lanes 7  and 8), 150 (lanes 9 and l o ) , 250 (lanes 11 and 12), and 500 p~ (lanes 13 and 14).
N3-ATP was carried out to determine whether the extent of photolabeling saturated a t high 8-N3-ATP concentrations. The extent of photolabeling increased linearly with time up to 1 min of irradiation a t 40 and 500 p~ 8-N3-ATP (Fig. 7), indicating that the amount of enzyme labeled was directly proportional to the recBCD enzyme-8-N3-ATP complex in the solution. Scintillation counting of bands cut from the gels showed that the recB and recD subunits were labeled to only about 5% after 30 s of irradiation in the presence of 500 p~ 8-N3-ATP. No change was observed in the UV absorbance spectrum of 90 p~ 8-N3-ATP in the 270-280-nm region after as much as 5 min of irradiation under photolabeling conditions (not shown). Thus, there was no significant change in either the enzyme or 8-N3-ATP concentration in the course of a brief UV irradiation. In addition, preincubation of the recBCD enzyme with 8-N3-ATP for 5 min a t 0 "C prior to irradiation had no effect on the extent of photolabeling, indicating that binding to both subunits was rapid over the time scale of the experiment. Taken together these findings show that the equilibrium between recBCD enzyme and 8-N3-ATP was not perturbed significantly by the UV irradiation.
The recBCD enzyme and aldolase were titrated with increasing concentrations of 8-N3-ATP in the presence and absence of 50 p~ ATP (Fig. 8). As shown in Fig. 9A, labeling of the recB and recD subunits was saturated at high 8-N3-ATP concentrations while aldolase was not saturated at even 500 p~ 8-N3-ATP. Analysis of Scatchard plots of the data (Fig. 9C) gave dissociation constants of 30 and 120 p~ for binding to the recD and recB subunits, respectively. Substan- Some inhibition by other nucleoside triphosphates was also observed (Fig. 10); however, the extent of inhibition was substantially less than that observed with ATP, dATP, or ATP-& showing that the sites are specific for adenine nucleotides. ADP was also inhibitory (Fig. l l ) , apparently binding with greater affinity to the recD than to the recB subunit.
Tryptic Peptide Mapping of recBCD Enzyme Photolabeled with [a-32PJ8-N3-ATP-The high affinity of both the recB and recD subunits for 8-N3-ATP and ATP suggests that both subunits have specific binding sites for ATP. Tryptic peptide mapping of enzyme photolabeled with [a-32P]8-N3-ATP was performed to determine whether labeling is confined to only one or a few peptides as would be expected for a specific binding interaction. The recBCD enzyme (1.5 nmol) was photolabeled a t 0 "C by irradiation for 1 min in the presence of [cI-~'P]~-N~-ATP (280 p~) .
The labeled protein was digested with trypsin and the peptides analyzed by HPLC. Three major radioactive peaks were observed (Fig. 12). The first corresponded to the column flow-through volume and eluted at the same position as [a-32P]8-N3-ATP. The other two radioactive peaks contained peptides as detected by their absorbance a t 214 nm (not shown). The 32P-containing fractions were pooled, lyophilized separately, and rerun over the same column using shallower elution gradients. Radioactivity was detected in each case in many (-20) fractions. To determine whether these broad peaks represent more than one labeled peptide, the radioactive fractions were again lyophilized and analyzed by two-dimensional electrophoresis-thin layer chromatography. One major radioactive species, along with a minor one, is visible in the autoradiogram of the chromatograms (Fig. 13). These results indicate that there are two peptides containing the bulk of the incorporated label, consistent with photolabeling at two specific sites in the recBCD enzyme.
If the recB and recD subunits each contain specific ATP binding sites, then only a single labeled peak would be expected to appear in tryptic digests of the isolated photolabeled subunits. Tryptic digestion and HPLC analysis were, therefore, performed on the isolated subunits as well as the intact enzyme. After photolabeling, one-half of the reaction mixture was subjected to NaDodS04-polyacrylamide gel electrophoresis, and the separated recB and recD subunits were cut out and digested with trypsin. The remainder of the mixture was treated with trypsin without separating the subunits. The HPLC elution profile of the digest of the intact enzyme (Fig.  14, top) showed three major peaks of radioactivity, as in Fig.  12. The HPLC profiles of the tryptic digests of separated photolabeled recB and recD subunits each contained a single peak of radioactivity (Fig. 14, middle and lower panels) 60 s (even-numbered lanes) of UV irradiation. Nucleotides added at 500 p~ were: none (hnes 1 and 2), ATP (lanes 3 and 4), GTP (lanes 5 and 6), CTP (lanes 7 and 8), UTP (lanes 9 and lo), dATP (lanes 11 and 12), dGTP (lanes 13 and  14), dCTP (lanes 15 and 16), dTTP (lanes 17 and 18), and ATPyS  (lanes 19 and 20).
Figs. 12 and 13, show that the photolabeling by 8-N3-ATP occurs largely at a single site in the recB and recD subunits, indicating that each subunit interacts specifically with the ATP analog. The broadness of the radioactive peaks in the HPLC elution profiles could have several explanations. Covalent labeling might occur at several amino acid side chains within a single peptide, giving rise to modified peptides with slightly different mobilities. There may also be some nonspecific labeling, although this would not necessarily give rise to a single broad peak but rather to multiple peaks with different mobilities.
Interaction of Separated recBC and recD Subunits with ATP and 8-N3-ATP-Direct evidence for binding and hydrolysis of ATP by both recB and recD subunits was sought by separating the recD subunit from the recBC complex by sedimentation through glycerol gradients containing 3 M NaCl (Amundsen et al., 1986). The separation was quite efficient, although some cross-contamination remained, particularly of the recD subunit by the recBC complex, as judged by silver staining of polyacrylamide gels. No double-stranded exonuclease activity was detected in the separated subunits, although the activity could be reconstituted by incubation under conditions described by Amundsen et al. (1986). A low level of DNA-dependent ATPase was found in the fractions containing the recBC complex and a somewhat lower level in fractions containing mainly the recD subunit, which could be due to contamination with the recBC complex. Neither subunit could be photolabeled under the usual conditions. Thus, the subunits interact much more weakly with ATP when separated (i.e. recD from recBC) than in the holoenzyme.

DISCUSSION
We have shown that the recB and recD subunits of the recBCD enzyme are photolabeled by UV irradiation in the presence of 8-N3-ATP. Both polypeptides are labeled to a greater extent than the recC subunit or aldolase, a protein known to lack an ATP binding site. The photolabeling of recB and recD, therefore, represents the interaction of 8-N3-ATP with specific binding sites on the two polypeptides.
Further experiments were carried out to prove the specificity of the interaction. Titration of the recBCD enzyme with 8-N3-ATP, using irradiation times short enough that the equilibrium could be sampled without being significantly perturbed, showed that the recB and recD subunits are saturated with dissociation constants of 120 and 30 @, respectively, while aldolase was not saturated even at 500 PM 8-N3-ATP. The saturation with relatively high affinity is to be expected for interaction of a ligand with an enzyme binding site.
The inhibition of photolabeling of both subunits by low concentrations of ATP and ADP further indicates that the sites have high affinity for ATP and other adenosine nucleotides, as well as for 8-N3-ATP. These compounds all inhibit labeling of the recD subunit to a greater extent than the recB subunit, consistent with the tighter binding of 8-N3-ATP to recD than to recB. Particularly noteworthy is the substantially greater inhibition of labeling of the recD than the recB subunit at low ADP concentrations. This fact, along with the different affinities observed for 8-N3-ATP binding, provides a strong argument against the possibility that there is only a single binding site situated at the interface of the recB and recD subunits, since in that case labeling of both subunits should be inhibited equally at a given ADP concentration. Inhibition of labeling by nonadenine-containing ribo-and deoxynucleoside triphosphates was somewhat weaker, again RecBCD enzyme (4.4 mg/ml) was photolabeled under standard conditions (see "Experimental Procedures") in a total volume of 0.1 ml by irradiation for 1 min in the presence of 280 p~ [a-"P]8-N3-ATP (200 pCi/pmol). Unincorporated label was removed by three successive 20-fold dilutions followed by centrifugation through Centricon 30 filters. The recovered protein was digested with trypsin, lyophilized, and the digest dissolved in 100 pl of 10 mM Tris-HC1 (pH 8.0), 1 mM EDTA, 1 mM DTT, and 1% NaDodS04. The redissolved digest was injected onto a C18 reverse-phase column (Vydac) and eluted with buffer A (0.1% (v/v) trifluoroacetic acid) for 5 min followed by a linear gradient over 90 min of 0-60% buffer B (0.1% trifluoroacetic acid (v/v) and 70% acetonitrile (v/v)) in buffer A. Fractions were collected a t 0.6-min intervals and the radioactivity determined by Cerenkov counting of the entire fraction. Fractions containing 32P from the HPLC run shown in Fig. 12 were pooled, concentrated, and dissolved in TLC buffer I (acetic acidformic acidH20, 15:5:80 (v/v/v) containing 2% methyl green). The redissolved material was spotted on Polygram CEL400 TLC plates (20 X 17.5 cm). Electrophoresis was carried out in the first dimension in buffer I at 800 V (40 V/cm). The plates were dried thoroughly, and ascending chromatography was carried out in the second dimension, eluting with butano1:pyridine:acetic acidH20, 32.5:25:5:20, v/v/v/v. The plates were then dried and autoradiographed. A , fractions 33-39 (peak 2). B, fractions 71-75 (peak 3). emphasizing the specificity of the binding sites for adenine. No dramatic difference in affinity of one site over the other was observed for any of these nucleotides.
8-N3-ATP is hydrolyzed by recBCD enzyme in a DNAdependent reaction. The rate observed for 8-N3-ATP hydrolysis is about one-tenth that seen at the same concentration of ATP. 8-N3-ATP also supports the double-stranded exonuclease activity of the recBCD enzyme and presumably the associated helicase activity (Muskavitch and Linn, 1982). These observations indicate that the interaction of 8-N3-ATP with at least one binding site on the enzyme is catalytically relevant. Inhibition of the double-stranded exonuclease activity when the enzyme and 8-N3-ATP are irradiated together is consistent with this conclusion. RecBCD enzyme (0.88 mg/ml) was photolabeled under standard conditions in the presence of 36 p~ [w3'P]8-N3-ATP (1350 pCi/ pmol) in a total volume of 100 p1 by irradiation for 1 min. Top panel, one-half of the mixture was digested with trypsin and run on the HPLC as in Fig. 12, except that NaDodSOI (1% w/v) and DTT (20 mM) were added directly to the digestion mixture without lyophilizing. Middle and lower panels, the remainder of the photolabeling mixture was loaded and run on a 7.5% NaDodS04-polyacrylamide gel, stained with Coomassie Blue, and destained. The recB and recD bands were cut from the gel, soaked in 25% isopropyl alcohol (4 X 1 ml), 10% methanol (4 X 1 ml), and 1% NH4HC03, pH 8.0 (4 X 1 ml) a t 37 "C. The washed gel slices were then minced by forcing them through fine steel mesh, digested with trypsin, and the digested protein removed from the gel fragments by centrifugation through Centricon 30 filters in 0.1% NaDodS04. The gel fragments were washed twice with 0.6 ml of H20, filtered again, and the combined filtrates concentrated to a small volume. The nearly dry mixtures were redissolved in 80 pl of HzO, 10 mM DTT, and run on the HPLC as in Fig. 12. Middle panel, 3ZP-labeled recD subunit. Lower panel, 32P-labeled recB subunit.
Finally, the tryptic peptide mapping experiments shown in Figs. 12-14 support the conclusion that photolabeling occurs at a single major site within each of the recB and recD subunits.
Separation of the recD subunit from the recBC complex was carried out in an effort to determine whether interaction of the isolated subunits with ATP or 8-N3-ATP can occur. The presence of ATPase activity in the recBC complex but not recD is in agreement with results reported with the isolated subunits by Lieberman and Oishi (1974) and with the isolated recB protein (Hickson et al., 1985). The isolated recD subunit could lose its ATPase activity upon separation from the other two subunits. Alternatively, binding of ATP to the recD subunit could have a purely regulatory or effector role. The inability to photolabel the isolated subunits suggests that they have reduced affinity for ATP compared to the holoenzyme. It should be noted that amino acid sequences homologous to those found in a number of ATPases are also found in the recB (Finch et al., 1986a) and recD (Finch et al., 1986b) subunits.
The standard conditions used in the photolabeling experiments (0 "C, no DNA) are not necessarily the most relevant to those in which the recBCD enzyme acts catalytically. As noted above, photolabeling does not depend on low incubation temperature but can take place at room temperature with only slightly lower efficiency. The effect of DNA is difficult