Deoxyribonucleic Acid Ligase ISOLATION AND PHYSICAL CHARACTERIZATION OF THE HOMOGENEOUS ENZYME FROM COLI”

SUMMARY DNA ligase of Escherichia coli has been purified to homogeneity as judged by polyacrylamide gel electrophoresis and by analytical ultracentrifugation. The molecular weights of the native, and the denatured and reduced forms of the enzyme are 74,000 f 3,000, hence, ligase consists of a single polypeptide chain. When the unadenylylated form of the enzyme is incubated with diphosphopyridine nucleotide, ap-proximately 1 mole of adenosine 5’-monophosphate becomes covalently linked per 74,000 g of protein. The DNA ligase of Escherichia coli catalyzes the synthesis of phosphodiester bonds at single strand interruptions in duplex DNA, coupled to the cleavage of the pyrophosphate bond of DPN (l-3). Although it has been demonstrated that enzyme-adenyl-ate and DNA-adenylate are formed in the DNA joining reaction (l-5), definitive proof that these two compounds are true kinetic intermediates requires the demonstration that their rates of reaction be as great or greater than the over-all DNA joining rate. We have therefore purified the E. coli DNA ligase to physical homogeneity and undertaken a kinetic analysis of the reaction which it catalyzes. paper we

The DNA ligase of Escherichia coli catalyzes the synthesis of phosphodiester bonds at single strand interruptions in duplex DNA, coupled to the cleavage of the pyrophosphate bond of DPN (l- 3).
Although it has been demonstrated that enzyme-adenylate and DNA-adenylate are formed in the DNA joining reaction (l-5), definitive proof that these two compounds are true kinetic intermediates requires the demonstration that their rates of reaction be as great or greater than the over-all DNA joining rate.
We have therefore purified the E. coli DNA ligase to physical homogeneity and undertaken a kinetic analysis of the reaction which it catalyzes.
In this paper we describe a purification procedure which leads to the isolation of physically homogeneous enzyme in good yield.
We also describe the physical properties of the pure DNA ligase. In the following paper we present a detailed kinetic analysis of the reaction catalyzed by the pure enzyme. A preliminary report of this work has appeared (6).

Materials
[32P]AMP (50 Ci per mmole) was prepared by the method of Symons (7). [32P]DPN was synthesized chemically from the * This is Paper XII in a series entitled "Enzymatic Joining of Polynucleotides." The previous paper is Reference 29. This investigation was supported in part by a grant from the United States Public Health Service (GM-06196).
$ Predoctoral Fellow of the National Science Foundation. Present address, Department of Biological Chemistry, Harvard Medical School, Boston, Mass.
$ Present address, Faculty of Pharmaceutical Sciences, University of Tokyo, Hongo, Tokyo, Japan.
[32P]AMP as described by Shuster et al. (8). Alternatively, the [32P]AMP was converted to [a-32P]ATP (9), which was then used to synthesize DPN enzymatically (10) The DPN, the first major 32P-containing peak eluting from the column, was concentrated by rotary evaporation. It was subjected to further purification by high voltage electrophoresis (40 volts per cm, 3 to 4 hours) on 589 Orange Ribbon paper (Schleicher and Schuell) in 0.05 M triethylammonium acetate (pH 5.5).
[32P]DPN prepared by either method had a typical radiochemical purity of 95% as judged by thin layer chromatography on polyethyleneimine cellulose (11). 3H-Labeled d(A-T)n copolymer was prepared as described previously (12). Unlabeled DPN (chromatopure grade) and NMN were purchased from P-L Biochemicals. Electrophoresis grade acrylamide, N , N'-methylenebisacrylamide, and N , N , -N' , N'-tetramethylethylenediamine were from Bio-Rad. This strain was grown in loo-liter cultures of yeast extract-phosphate-glucose medium (15) at 37" with aeration in a New Brunswick fermenter.
When the absorbance at 595 nm reached 2.5 to 2.7, the culture was chilled to 4" with the refrigeration unit of the fermenter. The cells were harvested with a refrigerated Sharples cent,rifuge and the cell paste was stored at -20" until ready for use.

Methods
Enzyme Assays-The DNA ligase was diluted into 0.05 M Tris-HCl (pH 8.0), I rnM EDTA, 3 mM MgC12, 0.01 M (NH&S04, 0.05 mg per ml of BSA. DNA joining activity of the E. co& ligase was assayed by d(A-T), circle formation as previously described (la), except that the enzyme diluent and reaction mixture were made 0.01 M in (NH&S04 (diluted from a 1 M stock solution neutralized to pH 7.5 with NH,OH). Instead of collection of the exonuclease III-resistant d(A-T)n circles by acid precipitation, 80-to loo-p1 aliquots of the exonuclease III digest were spotted on 1.5.cm squares of DEAE-paper (Whatman DE81).
The DE81 squares were washed with 0.3 M ammonium formate (pH 7.8) as described by Brutlag and Kornberg (16). Adenylylation of the ligase was carried out in a buffer consisting of 0.015 M Tris-HCl (pH 8.0), 0.01 M (NH&S04, 4 mM MgC12, 1 mM EDTA, and 100 pg per ml of BSA. [32P]DPN (0.5 to 1 X lo5 cpm per pmole) was present in at least a IO-fold excess over the ligase, which was present at concentrations ranging from 0.1 to 100 nM. Incubation was at 30" in siliconized (Siliclad Clay-Adams) glass tubes. The reaction was terminated by re-. moving samples of appropriate volume and making them 0.2 M in Tris base, 20 mM in EDTA, and 0.2 mg per ml in BSA. The 32P-labeled ligase-AMP was precipitated by adding cold trichloroacetic acid to a final concentration of 770. After 5 min at 0" the precipitate was collected on Millipore filters (HA, 0.45 p) which were washed four times with 10 ml of cold 1 N HCI.
When [32P]ligase-AMP was isolated on a preparative scale, the adenylylation reaction was terminated by adding EDTA to a final concentration of 10 mM. The reaction was applied to a siliconized column (30 cm x 0.5 cm2) of Sephadex G-50 (fine) equilibrated with 0.02 M Tris-HCl (pH 8.0), 1 mM EDTA, 0.01 M (NH&S04, 0.5 mM 2-mercaptoethanol, and 0.5 mg per ml of USA. The ligase-AMP, which was excluded by the gel, was collected into siliconized glass tubes. Polyacrylamide Gel Electrophoresis-Electrophoresis in gels containing 5% acrylamide and 0.1% sodium dodecyl sulfate was performed as described by Weber and Osborn (18). Gel length was usually 8.5 cm.
Sedimentation Analysis-Analytical centrifugation experiments were performed with an ANH-4 rotor in a Beckman model E ultracentrifuge equipped with schlieren and Rayleigh optics. Prior to sedimentation, the unadenylylated form of the ligase was dialyzed exhaustively at 4" against 0.02 M potassium phosphate (pH 6.50), 0.01 M NH&l (pH 6.50), 0.20 M KCl, 0.5 rniM dithiothreitol.
Velocity sedimentation of the enzyme was carried out in a filled Epon 12.mm double sector cell with sapphire windows. The sample sector was filled with 0.45 ml of the dialyzed ligase solution (2.7 mg per ml), and the reference sector contained 0.46 ml of the dialysis buffer. Sedimentation was at 56,050 rpm at 20.6". Schlieren photographs were taken at intervals of about 10 min on Kodak Spectroscopic IIG plates.
B phase angle of 60" was used. The sedimentation coefficient was corrected to standard conditions by assuming that ST/@0~) is invariant, where s is the sedimentation coefficient, 7 is solvent viscosity, ir is the protein partial specific volume, and p the solvent density.
Sedimentation equilibrium analysis was performed as described by Chervenka (20) in a 12-mm double-sector cell with a doublesector capillary synthetic boundary centerpiece.
The sample sector contained about 0.12 ml of dialyzed enzyme at 2.2 mg per ml (corresponding to a column height of 3.17 mm) which was raised off the cell bottom by 0.03 ml of fluorocarbon oil (FC-43, 3M Co.). The reference sector contained 0.18 ml of dialysis buffer.
The cell was centrifuged at 22,000 rpm for 2.5 hours and then at 9,000 rpm for 30 hours at 16.6". Rayleigh photographs were taken on Kodak Spectroscopic IIG plates at 25 hours and 30 hours after reduction of speed. Since there was no difference in the number of fringes on the two plates, the run was judged to be at equilibrium.
The initial protein concentration, Co, in units of fringes, was determined by a boundary forming run following the equilibrium run. The molecular weight was determined from the slope of a plot of In C versus r2 according to the equation M = 2RT/(l -tp)u2.d In C/d(r2). A Gaertner comparator was employed for measurements on schlieren and Rayleigh plates.
A partial specific volume, 0, of 0.740 at 25" was calculated on the basis of the amino acid composition (21,22). For use in calculations, this value was corrected for an assumed temperature dependence of 5 x lop4 rnl'g-I. degree-l (23).
Amino Acid Analysis-Amino acid analyses were performed on a Beckman model 121 Automatic Amino Bcid analyzer.
The enzyme was dialyzed exhaustively against 0.01 M potassium phosphate (pH 7.5). Portions of the dialyzed enzyme (about 100 pg) were taken to dryness on a rotary evaporator, a crystal of phenol was added (24), and the residue dissolved in 1.2 ml of 6 x HCl. Hydrolysis was performed at 110' for 24,48, and 73 hours.
The values for serine and threonine were obtained by extrapolating to zero time of hydrolysis.
The values cited for valine and leucine are averages of the 48. and 73.hour determinations. Half-cystine was determined as cysteic acid after performic acid oxidation (25). Tryptophan was estimated from the absorption spectrum of the native enzyme assuming a molar extinction coefficient at 278 nm of 5.55 x lo3 (26).
Other Methods-A Zeiss PM&II spectrophotometer was employed for optical measurements.
Radioactivity was determined in a Nuclear Chicago Unilux scintillation counter. Protein was determined by the method of Lowry et al. (27) or by the absorbance at 280 and 260 nm (28). The protein concentration of purified ligase preparations was determined by amino acid analysis.
Unless indicated otherwise, pH was determined at 25" at a buffer concentration of 0.05 M.

PuriJication of DNA Ligase
Unless indicated otherwise, all steps were performed at O-4", and centrifugation was at 15,000 x g for 20 to 30 min. The enzyme was routinely concentrated after each step by precipitation with (NH&SO4 (0.47 kg per liter, to 70% saturation). Solid (NH&S04 was added over a period of 20 to 30 min with stirring.
After stirring for an additional hour, the precipitate was collected by centrifugation.
A summary of the purification procedure is given in Table I.
The cells were disrupted by two passages through a Manton-Gaulin homogenzier at a pressure of 6500 to 7000 pounds.
The extract was clarified by centrifugation and buffer added to adjust the protein concentration to 20 mg per ml (Fraction I).
Streptomycin Fractionation-A freshly prepared 5% aqueous solution of streptomycin sulfate (620 ml) was added to 3.1 liters of extract over a 20.min period with stirring.
After stirring for an additional 20 min, the suspension was centrifuged. The supernatant fluid (3.4 liters) was diluted by the addition of 6.8 liters of cold H20, and then an additional 3.4 liters of 5% streptomycin sulfate were added.
The suspension was stirred for 20 min, and the insoluble material removed by centrifugation in lliter batches in a Sorvall Superspeed centrifuge.
The supernatant fluid (13.3 liters) was concentrated by (NH&S04 precipitation (Fraction II DEAE-cellulose Adsorption and E&ion-This step removes residual nucleic acid and is required for retention of the enzyme on phosphocellulose. Fraction III was dissolved in Buffer A and the volume adjusted to 215 ml to yield a protein concentration of 20 mg per ml. The solution was dialyzed against 4-liter portions (three changes) of 0.15 M potassium phosphate (pH 7.5), 0.01 M (NH&Sod, 2 mM EDTA, and 1 rnnf 2-mercaptoethanol for a total of 4 hours and then applied at a rate of 300 ml per hour to a DEAE-cellulose column (17.5 cm X 15.4 cm2) equilibrated with the same buffer. The column was washed with buffer (500 ml) and protein that did not adsorb (590 ml) collected (Fraction IV).
Phosphocellulose Chromatography of Adenylylated Enxyme-Fraction IV was dialyzed against 11 liters of Buffer A for 2 hours. Dialyzed enzyme was made 5 mM in MgC12 and 50 PM in DPN and placed in a 30" bath for 10 min, at which time the temperature of the solution was 23". After an additional 30-min incubation at O", EDTA was added to a concentration of 9 mM. The adenylylated enzyme was dialyzed against ll-liter portions (two changes) of 0.2 M potassium phosphate (pH 6.5), 2 mM EDTA, 1 mM 2-mercaptoethanol, and 2 PM DPN for 2.5 hours, and immediately applied at a rate of 90 ml per hour to a phosphocellulose column (82 cm x 9.2 cmz) equilibrated with the DPNsupplemented buffer.
The column was washed with 1.5 liters of this buffer and then eluted with a 36liter linear gradient of potassium phosphate (pH 6.5, 0.02 to 0.2 M) containing 2 mM EDTA, 1 mM 2-mercaptoethanol, and 2 PM DPN. Most of the DNA ligase activity was recovered from the column at the end of the 0.02 M potassium phosphate wash, well separated from the unadsorbed protein.
However, a second peak of activity eluting at 0.07 M potassium phosphate along with a major peak of contaminating protein was often observed. Usually (six experiments) the early eluting fraction contained the majority of recovered activity; however, in two experiments activity eluting at 0.07 M potassium phosphate was much more significant, comprising 40% or 90% of recovered activity (see below).
Early eluting fractions with a specific activity greater than 300 units per mg were pooled (1,200 ml) and taken to 707, saturation with ultrapure (NH&S04. The (NH&S04 suspension was allowed to stand overnight at O", and the precipitate collected by centrifugation at 15,000 x g for 3 hours (Fraction V).

DEAE-Xephadex
Chromatography-Fraction V was dissolved in 28 ml of 0.02 M Tris-HCl (pH X.0), 2 mM MgCl,, 1 mM 2-mercaptoethanol, and 0.2 M NaCl, and then dialyzed against two lliter portions of this buffer for a total of 4 hours. After removal of a small amount of insoluble material by centrifugation, the dialyzed enzyme was applied to a column of DEAE-Sephadex A-50 (28 cm X 2.9 cm2) equilibrated with the dialysis buffer. The column was washed with 160 ml of starting buffer, and then eluted with a 560.ml linear gradient of NaCl (0.2 to 0.5 M) containing 0.02 M Tris-HCl (pH 8.0), 2 mM XIgCl, and 1 mM 2mercaptoethanol.
The flow rate was 20 ml per hour and lo-ml fractions were collected.
Ligase activity eluted as a single peak at an NaCl concentration of 0.25 M. Fractions with a specific 7498 activity greater than 2000 units per mg were pooled (47 ml) and precipitated with ultrapure (NH&SO4 as in Fraction V (Fraction VI). Phosphocellulose Chromatography o.f Unadenylylated Enzyme-Fraction VI was dissolved in 5 ml of Buffer A containing 5y0 glycerol.
It was dialyzed against 1 liter of 0.05 M Tris-HCl (pH 7.5), 2 rnM EDTA, 1 mM 2-mercaptoethanol, and 5% glycerol for 2 hours. The dialyzed enzyme was made 0.28 mM in NMN and 5 mM in MgClz and incubated at 30" for 3 min. Aft,er further incubation at 0" for 60 min, EDTA was added to a final concentration of 7 mM, and the solution dialyzed against 250.ml portions of Buffer A (two changes) for a total of 2.5 hours. The dialyzed enzyme was applied at a flow rate of 10 ml per hour to a column of phosphocellulose (23 cm x 1.1 cm2) equilibrated with Buffer A. After washing with 50 ml of Buffer A, the column was eluted with a 200-ml linear gradient of potassium phosphate (pH 6.5,0.02 to 0.2 M) containing 2 mM EDTA-1 mM 2-mercaptoethanol, and 6-ml fractions were collected. The ligase was eluted from the column at a potassium phosphate concentration of 0.06 to 0.07 M. The peak fractions representing about 75% of the recovered activity were pooled (21 ml) and the protein precipitated with ultrapure (NH&Sod. The pellet was dissolved in 0.7 ml of 0.15 M potassium phosphate (pH 6.5), 0.01 M (NHI)SOI, 2 mM EDTA, and 1 mM 2-mercaptoethanol, and then dialyzed against loo-ml portions of this buffer (two changes) for 3 hours. The dialyzed enzyme was made 50% in glycerol and stored at -20" (Fraction VII).
This procedure has also been used to purify the DNA ligase from the streptomycin supernatant fluid obtained as a side fraction of the DNA polymerase I purification (13) from E. coli B. Cold water, 0.56 volume, and a fresh 5% solution of streptomycin sulfate, 0.52 volume, were added to the streptomycin supernatant fraction.
After removal of the precipitate by centrifugation, the supernatant fluid was concentrated with (NH&S04 as for Fraction II. The remaining steps in the purification were carried out as described above. The specific activity of Fraction VII prepared from E. coli B was identical to the same fraction obtained from LC81 that overproduces ligase.
Except for phosphocellulose chromatography of adenylylated enzyme, all steps of the purification were easily reproducible. As noted above, a significant fraction of the enzyme was occasionally found to elute at 0.07 M potassium phosphate instead of at the end of the 0.02 M wash. In one instance 90% of the recovered activity eluted at 0.07 M potassium phosphate.
Despite the fact that the specific activity of the enzyme recovered in this case was only 30% of that expected for this step, it was possible to obtain a ligase preparation of 95% purity by continuing with the purification, and subjecting the enzyme recovered after the final step to rechromatography on DEAE-Sephadex.
The following procedure was used. Fraction VII (220,000 units in 11 ml, 7,500 units per mg, protein determined by absorbance at 280 and 260 nm (as)), was dialyzed against 1,000 ml of 0.025 M Tris-HCl (pH 7.5), 1 mM EDTA, 1 mM 2-mercaptoethanol, and 5% glycerol for 2 hours. The dialyzed enzyme was made 50 PM in DPN, 4 mM in MgC12 and incubated at 30" for 5 min. EDTA was then added to a final concentration of 5 mM, and the solution dialyzed against l,OOO-ml portions (three changes) of 0.02 M Tris-HCl (pH 8.0), 2 mM MgCl?, 1 mM 2-mercaptoethanol, and 0.2 M NaCl for a total of 3 hours. The dialyzed enzyme was then applied to a column of DEAE-Sephadex A-50 (22.5 cm x 2.9 cm2) and chromatographed as described above. Fractions with a specific activity of 10,000 units per mg (protein determined by absorbance at 280 and 260 nm) were pooled (142,000 units) and concentrated with ultrapure (NH&SOI. The enzyme was dialyzed and stored as described for Fraction VII. DNA ligase purified by this procedure was stable for at least 1 year (< 10% loss in DNA joining activity).
Furthermore, the purified enzyme was free of detectable endonuclease activity on double stranded DNA in the absence of AMP (29) and contained no detectable exonuclease activity on d&T), copolymer (< 1 nmole/30 min per mg). Fraction VII of the DNA ligase was used for all the studies to be described.
Fractions II and III were stable for at least 3 months, and Fraction V was stable for at least 2 weeks when stored as anmonium sulfate pellet at -20".
Fractions IV and VI were stable overnight at 0", while Fraction I lost 10 to 20% of its activity when stored under these conditions. Electrophoretic Analysis DNA ligase was subjected to electrophoretic analysis under both native and denaturing conditions. Results of typical gels are presented in Fig. 1. When reduced and denatured enzyme was analyzed by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate, a single protein band was detected.
Furthermore, the identical mobilities of 32P-labeled enzyme-AMP and the single protein band on sodium dodecyl sulfate acrylamide gels indicated that DNA ligase activity was associated with this major protein species (Fig. 2). A single prot,ein zone was also observed after electrophoresis under native conditions at pH 8.0 or 8.3. Since 25 to 40 pg of protein were applied to each gel, these results indicate a purity of greater than 98 '$& The mobility of the ligase on sodium dodecyl sulfate gels relative to the mobility of bromphenol blue was 0.49 (S.D. = 0.01, seven experiments).
This value was compared with the mobilities of several proteins of known molecular weight including BSA dimer, DNA polymerase I, USA, aldolase, and lysozyme (18). The apparent molecular weight for the denatured and reduced form of the DNA ligase calculated from this standard curve was 74,000 =I= 3,000. The mobilities of the enzymes obtained from E. coli B and from the overproducing K12 strain LC81 were indistinguishable.
As judged by this criterion, as well as by their specific activities, the two enzymes appear to be identical.
The remainder of this section will be devoted to properties of the ligase obtained from LC81.

Sedimentation Analysis
The weight average molecular weight of the purified native ligase determined by sedimentation equilibrium ultracentrifugation was 77,000 (Fig. 3). The slight deviation from linearity of the In C versus r2 plot may reflect the presence of a high molecular weight impurity (approximately 5y0). Alternatively, it could represent a small amount of aggregation.
The value of 77,000 is close to that obtained for the denatured enzyme by sodium dodecyl sulfate gel electrophoresis (74,000)) indicating that the E. coli DNA ligase is composed of a single polypeptide chain.
The ~20,~ of the purified enzyme determined by analytical sedimentation at a protein concentration of 2.1 mg per ml (corrected for radial dilution) was 3.91 S. This value agrees reasonably well with the value of 4.2 S obtained for DNA ligase activity by sucrose gradient centrifugation (6). Except for the presence of the minor fast sedimenting contaminant mentioned above (see Fig. 4 Polyacrylamide gel electrophoresis was carried out as described under "Methods." The two gels on the left were run under native conditions, the center gel in System A, and the gel on the left in System B. The gel on the right was run under denaturing conditions in the presence of sodium dodecyl sulfate. From left to right the protein added was 30 pg, 40 pg, and 25 pg, respectively.
during the entire sedimentation velocity experiment, providing additional evidence that the ligase is homogeneous. The sedimentation coefficient of 3.9 S is lower than expected for a spherical protein of molecular weight 74,000, and suggests that the enzyme may have an asymmetric shape. *2P-Labeled ligase-AMP, isolated by filtration through Sephadex G-50, was subjected to electrophoresis in the presence of sodium dodecyl sulfate as described under "Experimental Procedure," except that a 7-cm gel was used. The gel was sliced into 2mm sections, and a2P was determined in a Nuclear Chicago gas flow counter.
A parallel gel was run with 15 pg of unlabeled ligase-AMP and the protein located by staining with Coomassie brilliant blue. Mobility is relative to the tracking dye bromphenol blue. tubes. However, residual adsorption may account for a stoichiometry of less than one.
As shown in Fig. 5  reaction. The stoichiometry determined by this method is 1.2 to 1.3 moles of AMP/74,000 g. It therefore seems likely that 1 mole of AMP is covalently bound per 74,000 g of protein. DISCUSSION The purification procedure described here for the E. coli DNA ligase leads to isolation in good yield of homogeneous enzyme as judged by several criteria. The complexity of the ligase reaction might suggest an oligomeric enzyme.
In fact, Zimmerman and Oshinsky (30) found that prolonged dialysis of a partially purified preparation in dilute buffer resulted in the appearance of a low molecular weight form of the enzyme that could still generate enzyme-AMP but did not catalyze the joining reaction, and on the basis of these observations, proposed a subunit structure for the ligase. Our physical studies of the homogeneous enzyme are inconsistent with an oligomeric structure and indicate that it is composed of a single polypeptide chain of molecular weight 74,000. The stoichiometry of the adenylylation reaction is also consistent with a functional unit of this molecular weight. Moreover, the pure enzyme has retained more than 90% of its DNA joining activity during storage for a year without any indication of an altered enzymatic species. Possibly the partially active, low molecular weight form of the ligase observed by Zimmerman and Oshinsky was generated by proteolytic cleavage of the native protein during prolonged dialysis. Mild trypsin treatment of the purified ligase does in fact yield fragments that can still be adenylylated by DPN but which are inactive in the DNA joining reaction (6). 7501 Given the molecular weight of the protein, it is possible to estimate the number of DNA ligase molecules per bacterium by comparing the specific activity of the pure enzyme with that observed in crude extracts.
Since crude extracts do not contain inhibitors of DNA joining activity as measured by the d(A-T)* copolymer assay, such a calculation is probably valid. Based on such an estimate, wild type E. coli growing in rich medium contain about 200 to 400 molecules of DNA ligase per cell.