THE JOURN~~ OF BIOLOGICAL CHEMI8TRY Vol. 247, No. 1, hue of January 10, pp. 232-240, 1972 Pn`ntcd in U.S.A. Deoxyribonucleic Acid Polymerase: Two Distinct Enzymes in One Polypeptide 11. A PROTEOLYTIC FRAGMENT CONTAINING THE 5` 3 3' EXONUCLEASE FUNCTION. RESTORA- TION OF INTACT ENZYME FUNCTIONS FROM THE TWO PROTEOLYTIC FRAGMENTS* (Fteceived for publication, July 16, 1971) PETER SETLOWS AND ARTWR KORNBERG From the Department of Biochemistry, Stanford University School of Medicine, Stanford, Cdijmia 9,4305 SUMMARY The small fragment (mol wt 36,000) produced by the lim- ited proteolytic cleavage of DNA polymerase (mol wt 109,000) retains only the 5' + 3' exonuclease activity. The small fragment resembles the 5' -+ 3' exonuclease of the intact enzyme in degrading DNA to mono- and oligonucleotides and in its capacity to excise mismatched regions such as thymine dimers. It differs from the intact enzyme in that deoxy- ribonucleoside triphosphates, which support polymerization, fail to stimulate the exonuclease or to increase the proportion of oligonucleotides among the products. However with a mixture of small fragment, large fragment (mol wt 76,000; polymerase and 3` .-* 5' exonuclease functions), nicked DNA, and suitable deoxyribonucleoside triphosphates, the same influences of polymerization on 5' -+ 3' exonucleases are seen as with the intact enzyme. Thus at the locus of a nick in DNA, the two fragments bind adjacent to one another to perform the coordinated polymerization-5' -+ 3' exonuclease functions that characterize the intact enzyme. Limited proteolysis of Escherichia coli DNS polymerase (mol wt 109,000) splits this single polypeptide chain into two frag- ments which retain the several catalytic activities of the intact enzyme (1-6). The large fragment (mol wt 76,000) which polym- erizes deoxyribonucleotides (polymerase) and hydrolyzes DNA in the 3' -+ 5' direction (3' + 5` exonuclease) (1,3) was described in the preceding paper (6). The smdl jragnent (mol wt 30,000) retains only the ability to hydrolyze double-stranded DNA in the 5' -+ 3' direction (5` -+ 3' exonuclease) (3, 7) and is the sub- ject of this report. Unlike the intact enzyme, the small fragment has no detectable * This study was supported in part by grants from the National Institutes of Health (United States Public Health Service) and the National Science Foundation. The previous paper in this series is Reference 6. $ National Science Foundation Postdoctoral Fellow. Present address, Department of Biochemistry, School of Medicine, Uni- versity of Connecticut Health Center, Hartford Plaza, Hartford, Connecticut 06105. polymerase activity, and no exonuclease activity on single- stranded DNA. However, the products of the action of the small fragment on double-stranded DNA are identical with those generated by the 5' -+ 3' exonuclease of the intact enzyme since both enzymes excise mono- and oligonucleotides and also thymine dimers (8, 9). Neither enzyme exhibits detectable endonuclease activity on a variety of substrates. The two dserent fragments can bind ne't to one another at a diester bond break (nick) in a DNA molecule, and this adjacent binding permits concomitant action of the polymerase and 5' + 3' exonuclease activities. Simultaneous action of the two frag- ments restores properties of the intact enzyme such as the stim- ulation of 5` 3` exonuclease by concurrent DNA synthesis, and the autocatalytically primed synthesis of poly[d(A-T)]? How- ever, the individual fragments show no direct affinity for one another in either the presence or absence of DNA. EXPERIMENTAL PROCEDURE Materials Nucleotides and P~lynucZeotides-[~~C]dATP and [14C]dCTP were purchased from SchwarB BioResearch; other deoxyri- bonucleotides were obtained as described previously (6). The 1 The abbreviations used are: poly[d(A-T)], alternating co- polymer of deoxyadenylate and deoxythymidylate; d(T),, d(T)*, d(T)5, oligonucleotides (5',3') of deoxythymidylate containing 2, 3, and 5 residues, respectively, all with a 3'-hydroxyl and a 5'-phosphate terminus; d(C) 8, an oligonucleotide (5',3') of deoxy- cytidylate of 3 residues in length with a 3'-hydroxyl and a 5'- phosphate terminus; pppTpTpT, thymidylyl- (5',3`)-thymidylyl- (5',3')-thymidine 5'-triphosphate; d(A)tooo, polydeoxyadenylate of length about 4OOO residues; d(T)taoo, polydeoxythymidylate of length about 4000 residues; d(1) 1000, polydeoxyinosinic acid of length about 1000 residues; d(T)200, polydeoxythymidylate of length about 200 residues; poly d(A-T)looo, double-stranded, alternating copolymer of deoxyadenylate and deoxythymidylate of length about loo0 residues; (e), thymine dimer; d(C)no-d(T)rW, a block copolymer containing about 120 residues of deoxycytidyl- ate on the 5' side of the molecule, and about 160 residues of deoxy- thymidylate at the 3' side; d(C)tzo polydeoxycytidylate of length about 120 residues; BSA, bovine serum albumin; PMSF, phenyl- methylsulfonyl fluoride; $f5pTp(T)too, a 5'-triphosphate-termi- nated d(T)too with the 0 and y phosphates labeled with [s*P]; ppp- TpT, thymidylyl-(5', 3`)-thymidine 5'-triphosphate; SDS, sodium dodecyl sulfate. 232 Issue of January 10, 1972 P. Setlow and A. Kornberg 233 oligonucleotides d(T)p, d(T),, d(T)I, and d(C)r were prepared according to Weimann, Schaller, and Khorana (10) and [s, y"P]- pppTpTpT was a gift from Dr. N. R. Cozzarelli (11). d(A)400o and d(T)woo were prepared and purified according to Riley, Maling, and Chamberlin (12); [14C]d(A)a~~o and [aH]- d(T)4000 were prepared in an identical manner using [WIdATP and [aH]dTTP. d(I)1ooo was prepared and purified according to Chamberlin and Patterson (13). The synthesis of [aH]d(T)~oo and the addition of a 2',3'-dideoxythymidylate terminus to the 3' end of an oligo(dT) have been described previously (8, 14). Ls, -Y-"PIPPPT~T)zoo was prepared from W, T-~PIPPPTPTPT and [aH]dTTP using terminal transferase according to the method of Cozzarelli, Kelly, and Kornberg (11). Activated calf thymus DNA and activated poly[d(A-T)lo~o] were prepared by limited digestion with pancreatic deoxyribonuclease (15, 16). Concen- trations of oligo- and polynucleotides are expressed BS nucleotide residues. Enzymes-E. coli DNA polymerase (Fraction VII) was used and had a specific activity of 4500 units per mg with activated calf thymus DNA (4). The large fragment of DNA polymerase was prepared either as described in this communication or puri- fied from E. coli as described previously (6). No differences were observed between the two preparations (6). Pancreatic deoxyribonuclease was purchased from Worthington Biochemi- cals. Terminal deoxynucleotidyltransferase from calf thymus (17) was a gift of Dr. F. N. Hayes (Los Alamos Scientific Labora- tory, Los Alamos, New Mexico), and subtilisin (Carlsberg) (18) was a gift of Dr. T. Link. Methods Polymerase and Exonuclease Assays-The assay of DNA syn- thesis on activated calf thymus DNA has been described (6), and the specific assays for 3' 4 5' exonuclease and 5' + 3' exo- nuclease were carried out as described previously (6) except as noted in the text. Concurrent synthesis and hydrolysis of polydeoxythymidylate was measured under the conditions for assay of 5' -+ 3' exonuclease, but [aH]d(T)~~ was substituted for [aH]d(T)a~~, and [C~-~P]~TTP was present at 0.5 mM. Endonuclease AssaysEndonuclease assays were carried out in 67 mM potassium phosphate (pH 7.4) and 6.7 m~ MgCll using one of the following substrates; ['4Cld(A) 4000 9 [aHld(T) 4000; ["CI- d(A)tooo. [3Hld(T)4~~~ (7 % fi) ; and P4C1d(A)4ooo. PHI~(T)zoo. The dA and dT concentrations were 30 p~ and 26 p~, respectively. After 30 min at 37", the assay mixture was made 10 mM in EDTA to stop the reaction, and 100 pl were layered on a 4-ml alkaline sucrose gradient (5 to 20 % sucrose in 0.1 M NaOH, 0.9 M NaC1). The gradients were centrifuged at 50,000 rpm for 6 hours at 4" in a Spinco L-2-65B centrifuge with an SW56 rotor. Fractions of 10 drops (0.1 ml) were collected and counted in Triton-toluene scintillation fluid (19). Endonucleolytic cleavage was estimated by comparing the sedimentation profile of the polynucleotides in the assay to those of untreated polynucleo- tides sedimented in parallel gradients. Since the labeled polynucleotides are somewhat heterogeneous, the assay is not very sensitive, but one endonucleolytic break in 50% of the polynucleotide molecules could have been detected. Synthesis of [14Cld(C)lIzcr[aH]d( T) IsrThe block copolymer d(C) 1t0-d(T)160 was prepared by synthesizing d(C)lto first and then extending this chain with deoxythymidylate residues using terminal transferase. [14C]d(C)lz~ was synthesized in 1 ml of 100 mM cacodylate, 100 mM potassium phosphate (pH 6.8) contain- ing 0.25 mM CoCb, 0.67 m~ [WIdCTP, 16 pad d(C),, and 54 pg of terminal transferase. The polymerization was complete (>W% utilization of the [WIdCTP) after 8 hours at 37", and the reaction wm stopped by addition of 25 pl of 45% KOH. After 10 min at 4" the mixture was neutralized with 20 pl of 85% &PO4 and dialyzed for 36 hours against 1 M NaCl, 10 mM po- tassium phosphate (pH 7.4), and l mM EDTA. This deoxy- cytidylate polymer (calculated to be d(C)uo) was extended with deoxythymidylate residues in a reaction mixture identical with that described above but containing 0.5 mM ['HIdTTP, 0.38 mM d(C)120, and 36 pg of terminal transferase. Polymerization was complete in 6 hours with >95% utilization of the dTTP; the reaction was stopped and the polymer purged as described above. The theoretical composition of the block copolymer is d(C)Imd(T)1so, and this value agrees with the ratio of ["C] to Further purification of the d(C)lro-d(T)lso was effected by annealing the polymer to an excess of d(A)rooo and sedimenting the duplex in a neutral sucrose gradient (5 to 20% sucrose in 0.1 M NaC1,lO mM potassium phosphate bder (pH 7.4), and 1 mar EDTA) as described above. Of the dC residues, 7% did not sediment with the d(A)4000, but remained near the top of the gradient in a region which had a mole ratio of dC to dT of 15; these fractions were discarded. The peak fractions further down the gradient (dC to dT mole ratio = 0.75) were pooled and concentrated. The d(C) 1zo-d(T)Iso was separated from the d(A)tooo by sedimentation in alkaline sucrose as described above. This is the preparation of d(C)Izo-d(T)160 used in the nuclease assay. The copolymer was also tested for the presence of thymidylate strands which contained only a few cytidylate residues at the 5' end by annealing to d(1) 1000 (20) and sedimenting the mixture in a neutral sucrose gradient. Of the dT residues, 6% remained at the top of the tube in a region with a molar ratio of dC to dT of 0.1. However the copolymer was not purified using this method due to our inability to separate completely all of the dI strands from the copolymer; introduction of this small amount of oligo(d1) would have complicated assays using the block copolymer. Proteolytic Cleavage of DNA Polymerase and Separation of Fragments-DNA polymerase was split into two fragments by proteolytic cleavage as described by Klenow and Overgaard- Hansen (3). Of intact DNA polymerase, 6.0 mg were incubated in 20 ml containing 67 my potassium phosphate (pH 6.5), 25 mM EDTA, 30 mM (NH& sod, 10% glycerol, 3 mg of BSA, 6.25 mg of activated calf thymus DNA, and 20 I.cg of subtilisin (Carlsberg). After 40 min at 37", the mixture was chilled in ice and then 0.2 ml of 33 m~ PMSF was added to inactivate the subtilisin (21). After 1 hour at 4", the mixture was adjusted to 0.2 M in potassium phosphate (pH 6.8), 1 mM in @-mer- captoethanol, and DNA was removed by passage of the enzyme through a 5-ml column of DEAE-cellulose equilibrated in the same buffer. The fraction passing through the column was dialyzed for 24 hours against two changes of 20 mM potassium phosphate (pH 6.8), 1 m~ j%mercaptoethanol, and then adsorbed to a 35-ml phosphocellulose column equilibrated in the dialysis buffer. The two fragments were separated by a hear gradient of 0.04 M + 0.16 M potassium phosphate (pH &S), 1 mM @-mer- captoethanol (100 ml of each buffer). Appropriate column in the final copolymer. 234 DNA Polymerase: Two Distinct Enzymes in One Polypeptide. II Vol. 247, Xo. 1 148 16 27 260" fractions were pooled and concentrated by vacuum dialysis. Total recoveries of both fragments were 70% (see Fig. 1). Chronmtography-Degradation products of [aH]d(T)p~~ and [*H]~(T)zoo (7% T^T) were separated by descending chromatog- raphy on DE&-paper by sequential elution with 0.25 M ammo- nium bicarbonate and 0.3 M ammonium formate (9). Degra- dation products of [fit y-a2P]pppTp(T)~~~ (;&Tp(T)zoo) were chromatographed for 40 hours on Schleicher and Schuell orange ribbon paper with markers of dTTP, pppTpT, and pppTpTpT. The solvent system was isobutyric acid-1 M NHrOH-0.1 M EDTA (100:60:1.6) (11). Strips of 5 mm were cut out and counted. Miscellaneous-Thymine dimers (TT) were formed in oligo- and polynucleotides of deoxythymidylate by irradiation with a low pressure mercury lamp. The absorbance at 267 nm was followed with time, and the percentage of thymine residues con- h <3 <0.2 37 37b Fracti on^ num ber FIG. 1. Separation of the active fragments of DNA polymerase produced by proteolytic cleavage of intact DNA polymerase. Details of the cleavage and phosphocellulose chromatography are under "Methods." The gradient was started at fraction number 4, and 3-ml fractions were collected. The 5` -.) 3` exonu- clease unit is 10 nmoles of nucleotide excised per hour. TABLE I Catalytic activities of small fragment compared with intact enzyme DNA synthesis was measured with calf thymus DNA, and 5' 4 3' exonuclease with concomitant DNA synthesis was as- sayed on d(A)~oo~~[*H]d(T)z~~i with 0.5 mM dTTP. Enzyme con- centrations were determined from the absorbance at 278 nm. DNA synthesis.. ................... 3' -+ 5' Nuclease. .................. 5' 4 3' Nuclease. .................. 5' -+ 3' Nuclease in the presence of deoxynucleoside triphosphates. ... 0 5` 4 3' Nuclease concomitant with DNA synthesis. a No DNA synthesis. verted to dimers was calculated from the data of Deering and Setlow (22). SDS-acrylamide gels were run as previously described (1) and protein molecular weights were determined on these gels by the method of Shapiro et al. (23) with the modifications of Weber and Osborn (24). DNA polymerase, bovine serum albumin, the dimer of bovine serum albumin, ovalbumin, and chy- motrypsinogen were used as molecular weight markers. RESULTS Isolation of Small Fragment Two laboratories have demonstrated that proteolysis of E. coli DNA polymerase yields an active fragment (the large fragment) of molecular weight 76,000 which retains the polymerase and 3' -+ 5' exonuclease activities of the intact enzyme (1, 2). Klenow and Overgaard-Hansen have further demonstrated that if the proteolysis is carried out in the presence of DNA, the po- lymerase is cleaved and one obtains not only the large fragment, but also a small fragment (mol wt 36,000) which contains only the 5' + 3' exonuclease activity (3). We have confirmed this result and have separated the two fragments by phosphocellulose chro- matography (Fig. 1). Each of the purified fragments gave a single band on SDS-acrylamide gel electrophoresis and was stable for at least 6 months when stored in liquid nitrogen. How- ever, the small fragment was quite labile at 37", and lost over 50% of its activity in 25 min under assay conditions; in contrast, both the large fragment and the intact enzyme are stable under these conditions? Catalytic Activities of Small Fragment The small fragment had full 5' 4 3' exonuclease activity on a double-stranded DNA substrate, with even a slightly higher turnover number than that of the intact enzyme (Table 1). However, the small fragment had no detectable nuclease activity on single-stranded DNA and no detectable polymerase activity (Table I). The 5` + 3' exonuclease of the intact enzyme was dramatically stimulated by concomitant DNA synthesis (Table I) (8). This effect was, as expected, not seen with the small fragment (Table 1). Specijieity of Small Fragment Degradation Products of d(A)4000'd(T)200 and d(A)4000-d(T)~,o (7% TT)-The 5' + 3' nuclease of the intact enzyme generated a characteristic mixture of mono- and oligonucleotides upon degradation of d(A)dooo .d(T)zoo. The distribution of products was unaffected by the extent of hydrolysis of the substrate (Table 11) (8). The small fragment also produced the same distribution of mono- and oligonucleotides and exhibited the same specificity as did the intact enzyme (Table 11). The products were also similar when the 5` + 3' nuclease action was on a d(A)4000. d(T)200 in which 7 % of the thymine residues were linked thymine dimers. Both the intact enzyme and the small fragment de- graded this substrate to about the same extent, and excised the same amount of thymine dimers as oligonucleotides of 4 to 8 residues in length (Table 11) (9). Larger oligonucleotides do not migrate from the origin in the chromatographic system used and were therefore not detected. A 2 P. Setlow and A. Kornberg, unpublished results. Issue of January 10, 1972 P. Setlow and A. Kornberg 235 130 TABLE I1 -.- Degradations were carried out at 37" in 0.4 ml of 67 mM potas- sium phosphate (pH 7.4), 6.7 m~ MgCL, and 67 fig per ml of BSA containing 32 nmoles of d(A)mo and 28 nmoles of either [aH]- d(T)eoo or [aH]d(T)~o~ in which 7% of the residues were thymine dimers. Intact polymerase (40 pmoles) or the small fragment (48 pmoles) was added to start the reaction; loO-pl samples were removed and hydrolysis was halted by addition of 3 pl of 0.5 M EDTA. The digest was chromatographed on DEAE-paper with appropriate markers. The percentage of degradation of the d(T)loo was determined by absorption of a small sample (10 pl) of the digest to DEAE-paper, and elution of mononucleotides and smdl 01igonucleot.ides with 0.3 M ammonium formate (pH 8.5). I __-- d (T) I d (TI : d(T)a tl (T)<:8 ~____ % % % % 79 79-81 72 71 17-18 16-17 16 13 3-4 3 <1 <1 <1 (1 13' 16" I Srnoll Fragincnt ~Is~/lOa J wactwn 2300" <0.3b 480" + 0 (3 u ln 1 .- .- ? 200 E n - 0 Q IO0 0 2.50 0.50 0.16 I , Sedi rncntation R 0.80 355 0.80 1966 0.30 123 116 0.12 70 61 Fraction number exonuclease eventually degrades all the product. This was not the case with the mixture of the two fragments presumably because the lability of the small fragment at 37" led to denaturation of the 5' -+ 3' exonuclease activity early in the reaction? Lack of Direct Binding between Small and Large Fragments The preceding experiments demonstrated that a mixture of the small and large fragments of DNA polymerase in the presence of DNA restores the properties of the intact enzyme. This result might be caused by reassociation of the two fragments on the DNA or independently of it. Binding of one fragment to the DNA may or may not favor binding of the other fragment adja- cent to it. Inasmuch as the large and small fragments are readily sepa- rated by Sephadex (3) or phosphocellulose chromatography (Fig. l), afKnity between them cannot be strong. This was demonstrated more directly by preparing a mixture of the frag- ments and then sediienting them on a sucrose gradient (Fig. 5). The small and large fragments sedimented independently of one another; their positions in the gradient were identical when they were sedimented alone or together. Similarly, the separate fragments exhibited no physical affinity for one another on DNA (Table VII). The turnover number of the small fragment was measured in the presence of deox- yribonucleoside triphosphates and with several ratios of large and small fragments to 3' and 5' ends on the DNA. At a high level of saturation of the 3' and 5' ends, the two fragments bound adjacent to one another at the same nick, and the result was similar to that seen in Table V. The turnover number of the small fragment was increased by the concomitant DNA synthesis of an adjacent large fragment. However, the turnover number TABLE VI1 Lack of afinity between small and large fragments on DNA Assays were carried out aa described in Table V with different amounts of both fragments as noted and with 0.5 mM dTTP. Turnover number gf 5' - 3' exonuclease activity" Ratio of fragments to 3' or 5' ends FIG. 5. Sedimentation of a mixture of the large and small fragments in a sucrose gradient. The two fragments were mixed in 67 m~ potassium phosphate (pH 7.4), 0.2 mM p-mercaptoethanol, and incubated at 37' for 10 min and then at 4' for 50 min. One hundred microliters of this solution containing 300 pmoles of the large fragment and 100 pmoles of the small fragment were layered on a 4-ml sucrose gradient (5 to 20% sucrose, 67 mM potaesium phosphate (pH 7.4), 0.2 m~ 8-mercaptoethanol) and centrifuged at 4" in a SDinco model L2 66B centrifuge with a SW 56 rotor at 56,OOO rpm.- After 12 hours, 20-drop fractions were collected and assayed for polymerase and 5' + 3' exonuclease activities. In- tact enzyme (300 pmoles), the large fragment, and the small fragment were also sedimented as markers in three parallel gradi- ents. during DNA synthesis. Nanomoles of nucleotide excised per mole of small fragment per min. * These values are calculated assuming that the small and large fragments bind both very tightly to 5' and 3' ends, respectively, and that the binding is uninfluenced by the presence of another fragment. A value of 196 waa found for asmall fragment adjacent to a large fragment and 35 for a small fragment alone. Thus with a 0.50 ratio of large fragment, one-half of the small fragments would have a neighboring large fragment (196/2 = 98) and one- half would not (35/2 = 18). 0 Specific activity of small fragment alone. d Specific activity of small fragment adjacent to large fragment of the small fragment decreased as the ratio of fragments to 3' and 5' ends decreased; the experimental values were almost identical with values calculated assuming random binding of the fragments to DNA. If the fragments had had a strong affinity for one another on DNA, a decrease in the ratio of fragments to 3' or 5' ends would not have resulted in a corresponding decrease in the turnover number of the small fragment. DIBCUSSION Efects oj DNA Synthesis an Excision FunctionStudies of the action of the separate fragments of DNA polymerase suggest a mechanism whereby concomitant DNA synthesis could both stimulate the rate and alter the specificity of the 5' -+ 3' exonu- clease. The large fragment is fuUy active in catalyzing DNA synthesis on a nicked double-stranded DNA and probably dis- places the strand ahead of the 3'-OH growing point (6). It is therefore probable that DNA synthesis by the intact enzyme may also cause some fraying at the 5' end of the strand ahead of a growing point. This frayed strand might be more susceptible to degradative attack by the same enzyme molecule (8) resulting in an increased rate of 5' 4 3' exonuclease. The fraying of the 5' end will also explain the excision of oligonucleotides (Table VI) since both the intact enzyme and the small fragment are known to excise large oligonucleotides from a DNA duplex containing frayed 5' ends produced by thymine dimer formation (Table 11) or the introduction of mismatched bases (9). The reason for the apparent upper limit on the size of the oligonucleotides produced by 5' -+ 3' exonuclease action is, however, not clear. Concomitant Action of Small and Large Fraction-The stimu- lation in the exonucleolytic rate of the small fragment by con- current DNA synthesis catalyzed by the large fragment can best be explained by concomitant action of two different fragments which have bound adjacent to one another at a single nick in a DNA molecule. The alternative explanation that DNA syn- thesis by the large fragment causes displacement of the 5' end which is then attacked by an unbound small frapent Seems less likely since the small fragment shows an absolute specificity for a double-stranded substrate, although the 5' end at which the small fragment acts may have some single stranded character, i.e. may be frayed due to a thymine dimer or some other cause. Previous work of Kelly et at. indicated that the stimulation in 5' + 3' exonuclease by DNA synthesis in the intact polymerase also involved concomitant action of the polymerase and 5' -+ 3' exonuclease activities of the same molecule (8). Lack of Afinity between Small and Large Fragment-The exper- iment demonstrating the lack of aEnity between the two frag- ments in the absence of DNA is quite conclusive (Fig. 5), but demonstration of their lack of &nity in the presence of DNA (Table VII) is based on two assumptions: (a) that the two frag- ments bind tightly to DNA, and (b) that the binding of the fragments to DNA is reversible and allows formation of thermodynamically favored states. Sedimentation studies have demonstrated that intact DNA polymerase binds tightly to both DNA and poly[d(A-T)] oligomers (29), and we have similar data for the binding of the large fragment to d(A)4000-d(T)200.2 The small fragment also binds tightly to DNA, since the rate of 5' -+ 3' exonuclease on d(A)cooo.d(T)200 was the same when the ratio of small fragment to 5' ends was reduced from 3 to 1 to 1 to 1. There are no data on the dissociation rates of the individual fragments from DNA, but the binding of the intact enzyme to Issue of January 10, 1972 P. Setlow and A. KomEerg 239 16. SCHACHMAN, H. K., ADLER, J., RADDING, C. M., LEHMAN, DNA is reversible (29). There are no data on the velocity of I. R., AND KORNBERG, A. (1960) J. Biol. Chem., 2!J6,3242. this dissociation, but it takes place at 4O, while our experiments were performed at 37" where the dissociation would be expected to be more rapid. Structure of DNA Polymerase-The cleavage of DNA polymerase into two active fragments which show no affinity for one another, and the similarity of their catalytic reactions to the analogous reaction catalyzed by the intact enzyme indicates that within the single polypeptide chain of DNA polymerase are two distinct enzymes, one enzyme containing the polymerase and 3' --+ 5' exonuclease function and the other enzyme containing the 5' -+ 3' excision function. The two enzymes are held to: gether by a polypeptide link (susceptible to proteases), which ensures that both the polymerase and excision function act si- multaneously at the same nick in a DNA molecule. Possible Importance of Two Enzymes-Single Polypeptide Structure-The coupling of the polymerase and excision func- tions in a single enzyme may be of significant advantage in recom- binational events and in the in Vivo repair of lesions in DNA such as thymine dimers. As a thymine dimer is excised by the excision function, concurrently the polymerase function fills in the gap caused by the excision. Therefore excision repair by DNA polymerase at no time leaves a single strand gap which might be attacked at the 3' end by exonucleases such as exonu- clease 111, but rather leaves only a nick with a 3'-OH and a 5'-P from which the polymerase can be displaced by DNA ligase. Possibly the extensive DNA degradation observed after ultraviolet irradiation of E. coli mutants which lack DNA polymerase (pol A mutants (30)) is the result of the production of single strand gaps during the excision of thymine dimers by secondary repair systems (31, 32). The lifetime of these gaps may be long enough to enable enzymes such as exonuclease I11 to attack these single strand gaps and thereby cause extensive degradation of the DNA. REFERENCES 1. BRUTLAQ, D. L., ATKINSON, M. R., SETLOW, P., AND KORN- 2. KLENOW, H., AND HENNINGSEN, I. (1970) Proc. Nut. Acad. Sci. 3. KLENOW, H., AND OVERQAARD-HANSEN, K. (1970) Fed. Eur. 4. JOVIN, T. M., ENGLUND, P. T., AND BERTSCH, L. L. (1969) 5. KORNBERG, A. (1969) Science, 163, 1410. 6. SETLOW, P., BRUTLAG, D. L., AND KORNBERG, A. (1972) J. Biol. Chem., 247, 224. 7. DEUTSCHER, M. P., AND KORNBERG, A. (1969) J. Biol. Chem., 244, 3029. 8. KELLY, R. B., COZZARELLI, N. R., DEUTSCHER, M. P., LEH- MAN, I. R., AND KORNBERG, A. (1970) J. Biol. Chem., 246, 39. 9. KELLY, R. B., ATKINSON, M. R., HUBERMAN, J. A4., AND KORN- BERG, A. (1969) Nature, H4, 495. J. Amer. Chem. Soc., 86, 3835. J. Mol. Biol., 46, 513. Mol. Biol., 20, 359. Biol., 12, 410. A. F., AND MOFFATT, J. G. (1969) Biochemistry, 8,4897. 237, 519. BERG, A. (1969) Biochem. Biophys. Res. Commun., 37,982. U. S. A., 66, 168. Biochem. SOC. Lett., 6, 25. J. Biol. Chem., 244, 2996. 10. WEIMANN, G., SCHALLER, H., AND XHORANA, H. G. (1963) 11. COZZARELLI, N. R., KELLY, R. B., AND KORNBERG, A. (1969) 12. RILEY, M., MALING, B., AND CHAMBERLIN, h.I. J. (1966) J. 13. CHAMBERLIN, M. J., AND PATTERSON, D. L. (1965) J. Mol. 14. ATKINSON, M. R., DEUTSCHER, M. P., KORNBERG,A., RUSSELL, 15. APOSHIAN, H. V., AND KORNBERQ, A. (1962) J. Biol. Chem., 240 DNA Polymerase: Two Distinct Enzymes in One Polypeptide. II Vol. 247, No. 1 17. KATO, K., GONCALVES, J. M., HOUTS, G. E., AND BOLLUM, 18. GUNTELBERG, A. V., AND OTTESEN, M. (1952) Nature, 170, 802. 19. PATTERSON, M. S., AND GREENE, R. C. (1965) Anal. Chem., F. J. (1967) J. Biol. Chem., 242, 2780. 37, 854. 20. INMAN, R. B., IND BALDWIN, R. L. (1964) J. Mol. Biol., 8, 452. 21. BOYER, H. W., AND CARLTON, B. C. (1968) Arch. Biochem. 22. DEERING, R. A,, AND SETLOW, R. B. (1963) Biochem. Biophys. 23. SHAPIRO, A. L., VIN~ELA, E., AND MAIZEL, J. V., JR. (1967) Biophys., 128, 442. Acta, 68, 526. Biochem. Biophys. Res. Commun., 28. 815. 24. WEBER, K., AND OSBORN, M. (1969) J. Bid. Chm., 244, 4406. 25. ENGLUND, P. T., HUBERMAN, J. A., JOVIN, T. M., AND KORN. 26. PEARSON, M., AND JOHNS, H. E. (1966) J. Mol. Biol., 20, 215. 27. GUILD, W. R. (1968) Cold Spring Harbor Symp. Quani. Biol., 28. HOWARD-FLANDERS, P. (1968) Annu. Rev. Biochem., 37, 175. 29. ENGLUND, P. T., KELLY, R. B., AND KORNBERG, A. (1969) 30. DELUCIA, P., AND CAIRNS, J. (1969) Nature, 224, 1164. 31. BOYLE, J. M., PATERSON, M. C., AND SETLOW, R. B. (1970) 32. KANNER, L., AND HANAWALT, P. (1970) Biochem. Biophys. BERG, A. (1969) J. Biol. Chem., 244, 3038. 33,142. J. Biol. Chem., 244, 3045. Nature, 226, 708. Res. Commun., 39, 149.