Nucleic Acids Research

Volume 14 Number 20 1986 Nucleic Acids Research Chromatin 3'-phosphatase/5'-OH kinase cannot transfer phosphate from 3' to 5' across a strand nick in DNA Yvette Habraken and Walter G.Verly Biochimie, Facult6 des Sciences, Universit6 de Li6ge, Sart Tilnan B6, 4000 Li6ge I, Belgium Received 24 July 1986; Accepted 29 September 1986 Summary : Rat liver chromatin contains a 3'-phosphatase/5'-OH kinase which may be involved in the repair of DNA strand breaks limited by 3'-phosphate/5'-OH ends. In order to determine whether the phosphate group can be transferred directly from the 3' to the 5' position, a polynucleotide duplex was synthesized between poly(dA) and oligo(dT) segments which had 3'-[32P]phosphate and 5'-OH ends. The oligo(dT) segments were separated by simple nicks as shown by the ability of T4 DNA ligase to seal the nick after the 3'-phosphate was removed by a phosphatase and the 5' end was phosphorylated with a kinase. The chromatin 3'phosphatase/5'-OH kinase was unable to transfer phosphate directly from the 3' to the 5' end of the oligo(dT) segments in the original duplex; ATP was needed to phosphorylate the 5'-OH end. It is concluded that the chromatin 3'-phosphatase/5'-OH kinase is unable to convert a 3'-phosphate/5'-OH nick which cannot be repaired by DNA ligase directly into a 3'-OH/5'-phosphate nick which can be repaired by DNA ligase; the chromatin enzyme rather acts in two steps : hydrolysis of the 3'-phosphate followed by ATP-mediated phosphorylation of the 5'-OH end. INTRODUCTION We have found [1] and purified to near homogeneity an enzyme from rat-liver chromatin that specifically hydrolyzes 3'phosphates in DNA; the protein is also a polynucleotide 5'-OH kinase. Pheiffer and Zimmerman [2] found that a preparation of DNA kinase, purified from rat-liver nuclei, contained a 3' phosphatase activity. T4 polynucleotide 5'-OH kinase has also a 3' phosphatase activity [3]. The association of these two activities on the same protein in organisms as far apart as T4 phage and rat suggests the possibility that they may participate to a concerted repair mechanism. 3'-phosphate ends are found in irradiated DNA [4] where they are usually separated from 5'-phosphate ends by a one-nucleotide gap [5]. Simple nicks limited by 3'-phosphate and 5'- 0 I RL Press Umited, Oxford, England. 8103 Nucleic Acids Research OH ends are less common; they may exist in limited numbers in irradiated DNA or result from an aborted reaction catalyzed by an eukaryotic topoisomerase which cleaves the phosphodiester bonds in 3'-phosphates that normally attach to tyrosine residues of the enzyme, and 5'-OH [6]. The problem that we tried to solve in this work is whether 3'-phosphate/5'-OH nicks which DNA ligase cannot seal can be transformed in one step by the chromatin 3'-phosphatase/5'-OH kinase into 3'-OH/5'-phosphate nicks that DNA ligase can seal. This could occur if the enzyme could transfer the phosphate directly from 3' to the 5' end (figure 1). MATERIALS and METHODS Materials : Chromatin 3'-phosphatase/5'-OH kinase was prepared from rat liver [1 and unpublished]. Deoxynucleotidyl terminal transferase, T4 polynucleotide kinase, T4 DNA ligase, deoxynucleosides triphosphates and ATP were purchased from Boehringer; alkaline phosphatase from Sigma; (dT)8 and (dA)8 from PL-Biochemicals; Sephadex G-25 and Blue Sepharose from Pharmacia. The radioactive nucleotides came from Amersham. Buffer A : 50 mM Na succinate, 10 mM MgCl2, 1.3 mM 2-mercaptoethanol, bovine serum albumin (0.03 mg/ml), pH 5.5. Buffer B : 0.5 M sodium cacodylate, 25 mM MgCl2, 0.85 mM ZnCl2, 2 mM 2-mercaptoethanol, bovine serum albumin (5 mg/ml), pH 7.0. Acid-insoluble radioactivit Aliquots of the polymerization mixtures were brought to 1) 5'-OH + 3'-P +51-P + 31-oH .ligase 5,_-P-3, 2) 5'-OH + 3'-P - 5'-OH + 3'-OH AP 5.P + 3'-OH * 5,_p3, repair of 3'-phosphate/5'-OH nicks 1) phosphate transfer from 3' to 5' end of the nick; even if, when isolated, the first step might come to an equilibrium, Figure 1 : Two models for the it would go to completion when coupled with the ligation step. 2) hydrolysis of 3'-phosphate followed by ATP-mediated phosphorylation of 5'-OH. 8104 Nucleic Acids Research 100 p4 with water, and then 100 p4 14 mM Na2P207-bovine serum albumin (0.5 mg/ml) and 200 p4 10 % HC0l were added. After 10 min at 0°C, the tubes were centrifuged for 10 min at 12,000 x The supernatants were discarded and the pellets were dissolved in 100 pl 0.1 M NaOH and reprecipitated with 400 p4 5 % HC104. Finally, these precipitates, collected by centrifugation, were redissolved in 1 ml 2 M NH4OH for radioactivity determination. Inorganic phosphate radioactivity To the tubes containing the reaction solution (100 p1 maximum), cooled on ice, were added : 100 p4 25 mM Na4P207, 20 mM KH2PO4, bovine serum albumin (0.5 mg/ml); 400 p4 10 % HC104; and 200 p1 activated charcoal (20 % in water). The mixtures were shaken a few minutes and, after 10 more minutes at 0°C, were centrifuged for 10 min at 12,000 x j. The radioactivity of the supernatants, which contained the inorganic phosphate, was measured. Treatment with T4 polynucleotide kinase or with chromatin 3'phosphatase/5'-OH kinase : Reaction mixture consisted of : 50 p4 poly(dA)/poly(dT) (about 0.2 pg) in 10 mM Tris.HCl, pH 7.5; 12 p4 100 pM ATP (32p_ labelled in y position [6 pCi] in some experiments) in water; 10 p4 T4 polynucleotide kinase (4.5 units) or chromatlin 3'-phosphatase/5'-OH kinase in buffer A; and 37.5 p4 of 3 x concentrated buffer A. After 1 h at 370C, another 10-p4 enzyme solution was added and the incubation was carried out for another 1 h. Treatment with T4 DNA ligase 32 p4 210 mM Tris.HCl, 30 mM MgCl2, 1 mM EDTA, 1 mM ATP, pH 7.60, 60 p4 poly(dA)/poly(dT) (about 0.25 pg) in 10 mM Tris.HC1, pH 7.5, 1 p4 T4 DNA ligase (1 unit) solution, were incubated 16 h at 160C. EXPERIMENTS and RESULTS 1) Preparation of a double-stranded Dolydeoxynucleotide substrate containing 3'-phosphate/5'-OH nicks : 1.1 - Synthesis of poly(dA) : A mixture of 5.05 nmol (dA)8, 1125 nmol [3H]dATP (28 pCi), 400 units of deoxynucleotidyl terminal transferase, 600 p1 buffer B, comipleted to 2 ml with water, was incubated at 370C for 6 h. An aliquot showed that 85 % of 8105 Nucleic Acids Research the radioactivity had become acid-insoluble. The proteins were eliminated by filtration through a Blue Sepharose column equilibrated with 10 mM NH4HCO3. The mononucleotides were eliminated by filtration through a Sephadex G-25 column; the fractions corresponding to the poly(dA) peak, detected by absorbance at 260 nm and radioactivity measurements, were pooled. The average length of the [ H]poly(dA) molecules was 1125x085 + 8 = 197 nucleotides. A mixture of 5.5 nmol 1.2 - Synthesis of poly(dT,dA) (1 pCi), (40 pCi), 40 rnmol [a(dT)8, 2400 nmol 250 units deoxynucleotidyl terminal transferase, 600 pl 0.2 M potassium cacodylate, 4 mM CoCl2, 0.4 mM ZnCl2, bovine serum albumin (3 mg/ml), pH 7.4, 0.2 mM S-mercaptoethanol (added just prior utilization), completed to 1.5 ml with water, was incubated at 37°C for 16 h; 56 % of [ 3H] and 92 % of [ 32p] radioactivities had become acid-insoluble. Proteins and mononucleotides were eliminated as described Dreviously (1.1) except that elution was carried out with 10 mM Tris.HCl, pH 7.5. The average elongation of (dT)8 is (2400 xO.56) + (40x0.92) 2400 x 0.56 =_. rai 36. = 251 nucleotides, and the T/A ratio 40 x 0.92 1.3 - Depurination of poly(dT,dA) and fragmentation by 6 elimination : The poly(dT,dA) solution was made 10 mM with HC1 and heated 5 min at 1000C to hydrolyze away the A's (figure 2); it was then neutralized with NaOH. The extent of depurination was measured on an aliquot after addition of NaOH up to 0.2 M and a 30-min heating at 650C (to fragment the polynucleotide and induce S6 elimination of the apurinic sites), HC1 neutralization, exhaustive treatment with alkaline phosphatase and determination of the inorganic phosphate radioactivity. In a pilot preparation, the depurination was 100 %, but, when the scale was increased 10 times, the depurination was only 59 %. The solution of polynucleotides in which apurinic sites had replaced 59 % of the A's, was brought to pH 11.5 by addition of 1 M glycine.NaOH, pH 12, then incubated at 370C. Fragmentation of the polynucleotide molecules at apurinic sites by a-elimination (figure 2) was followed by determination of the radioactivity soluble in 5 % HC104 at 0°C. When it was at a maximum [3H]dTTP 8106 32P]dATP Nucleic Acids Research 32 PPdATP P]depurination Figure2 L F()pTVTpTP 6-elimination PT'pTPTnpTp* H phosphatase TV pSVPTP* H- 6-elimination TpVpTp4 Synthesis of oligo(dT) with 5'-OH ends. 3'-[32Pphosphate and and that, consequently, the fragmentation was complete, the 5'phosphate-terminated pieces had an average length of : (37/0.59) - 1 = 62 nucleotides plus an apurinic site at the 3' end. 1.4 - Hydrolysis of the 5'-phosphate and 6-elimination of the apurinic site : The solution (1.23 ml) of depurinated and fragmented polynucleotide (19 pg) was brought to pH 9 and, after addition of 2.5 units of alkaline phosphatase, incubated at 37°C for 30 min (figure 2). NaOH was subsequently added up to 0.2 M before heating at 65°C for 30 min to eliminate the apurinic sites (figure 2). After cooling, the solution was neutralized with HC1. The last solution contained polynucleotides of 62-nucleotide average length with 5'-OH and 3'-[32P]phosphate ends. This was checked on an aliquot : exhaustive alkaline phosphatase treatment released 44 % of the 32P as inorganic phosphate; since we had only a 59 % depurination, this result means that all later steps had a yield near maximum. 1.5 - Hybridization : The solutions of poly(dA)197 (30 nmol A) and poly(dT)62 (33 nxmol T) were mixed and warmed 7 h at 22°C. The slight excess of T's relative to A's, was believed to oblige the 3'-[32P]phosphate end of a poly(dT) molecule to lie as close as possible to the 5'-OH end of the next molecule leavinq a simple nick between them in the T strand of the duplex. 2) Checking the substrate In this experiment, we used the poly(dT.dA) of the pilot preparation where the depurination was 100 % (see 1.3). A S6 8107 Nucleic Acids Research elimination performed in one step, followed by an exhaustive treatment with alkaline phosphatase, gave poly(dT) molecules, limited by 3'-OH and 5'-OH ends (the 32P was completely lost), having an average length of 36 nucleotides. Poly(dA)197 and poly(dT)36 were mixed in a T/A ratio of 1.1 and annealed. The hybrid (0.21 pg) was treated with T4 5'-OH kinase (4.5 units) and [y- 32P]ATP, then the solution was filtered through a Sephadex G-25 column to eliminate the radioactive ATP and inorganic phosphate. In the hybrid, the fragments of the T strand then had 3'-OH and 5'-[32P]phosphate ends. Treatment of an aliquot of this preparation by alkaline phosphatase released 96 % of the 32P as inorganic phosphate. The hybrid (16 ng) of poly(dA) and poly(dT) with 5,_[32PI phosphate ends, was incubated with T4 DNA ligase and ATP. Subsequent exhaustive treatment with alkaline phosphatase released only 42 % of the 32P as inorganic phosphate. The DNA ligase 42 =56 % of the phostreatment prevented the release of 96 96 phate, meaning that 56 % of the 5'-phosphates of the poly(dT) fragments had been included in phosphodiester bonds. We may conclude that many fragments of the T strand were contiguous; if a gap had remained between them, no phosphodiester linkage could have been formed by the DNA ligase. 3) Tentative transfer of phosphate from 3' to 5' ends across a nick using the 3'-phosphatase/5'-OH kinase of rat-liver chromatin In this experiment, we used the hybrid prepared in (1.5) where the T strand fragments were limited by 5'-OH and 3t_[32PI phosphate ends containing 44 % of the substrate 32p. When this hybrid (1.2 pg) was incubated with the chromatin was released as 3'-phosphatase/5'-OH kinase, 37 % of the inorganic phosphate whether or not 10 pM ATP was present. After filtration through a Sephadex G-25 column to eliminate the radioactive inorganic phosphate, a subsequent treatment with alkaline phosphatase did not release any more inorganic [ 32P1phosphate (Table I); an incubation with T4 DNA ligase interposed between the chromatin enzyme and alkaline phosphatase treatments, obviously did not change the result. The conclusion is that the 3'-[32P]phosphate ends of the substrate were completely hydrolyzed by the chromatin 3'-phosphatase/5'-OH kinase and that no 32P 8108 Nucleic Acids Research Table I Evidence that the conversion of a 3'-phosphate/5'-OH into a 3'-OH/5'-phosphate nick in DNA by the chromatin 3'-phosphatase/5'-OH kinase is a two-step reaction. [y- 32 P]ATP DNA ligase ,. released by P04 phosphatase ~~~~~~~~(%) 0 + 0 + 54 23 + + The hybrid constituted by poly(dA) and poly(dT) fragments with 3'-[3 P]phosphate and 5'-OH ends (see 1.5) was incubated with the chromatin 3'-phosphatase/5'-OH kinase, with or without [y-32P]ATP. After filtration through a Sephadex G-25 column to eliminate the small radioactive molecules, the modified hybrid was incubated or not with T4 DNA ligase, before being submitted to alkaline phosphatase; the amount of inorganic [32p]phosphate released by this latter treatment is expressed as percents of the 32P radioactivity of the modified hybrid substrate after the filtration step. [32P]phosphate was transfered from 3' to 5' ends. The hybrid treatment with the chromatin 3'-phosphatase/5'OH kinase was also carried out in the presence of 10 WM [y- 32P] ATP, before filtration through a Sephadex G-25 column to eliminate the radioactive ATP and inorganic phosphate. The 32P radioactivity of the hybrid is now that left after hydrolysis of the 3'-[32P]phosphate ends by the 3' phosphatase activity of the chromatin enzyme and the [32 P]phosphate from [y- 32P]ATP added to the 5' ends by the 5'-OH kinase activity of the chromatin enzyme. A subsequent treatment with alkaline phosphatase released 54 % of the substrate 32P as inorganic phosphate; a preliminary treatment with T4 DNA ligase reduced this amount to 23 % (TableI). This control shows that, in the T strand of the hybrid, fragments were close enough so that, when their extremities had been changed to 3'-OH and 5'-phosphate, T4 DNA ligase could seal the interruption. DISCUSSION It is intriguing to find 3'-phosphatase and 5'-OH polynucleotide kinase activities within the same protein as well in the rat as in T4 phage-infected Escherichia coli. These two 8109 Nucleic Acids Research activities are capable of converting a 3'-phosphate/5'-OH strand nick, unreparable by DNA ligase, into a reparable 3'-OH/5'-phosphate nick in double-stranded DNA. The question was whether the conversion occurred in one or two steps (figure 1). In the onestep model, the phosphate would be moved from 3' to 5' by a concerted mechanism mediated by the two activities so that, if ATP was needed, it was only as an allosteric effector. In the twostep model, the 3' phosphatase first hydrolyzes the 3'-phosphate, the phosphate being recovered as inorganic phosphate, and ATP supplies its y-phosphate to phosphorylate the 5'-OH end. The experiments presented here do not show any hint of the possibility for the phosphate to be transferred from 3' to 5' across a nick in a polydeoxynucleotide duplex. It thus seems that conversion of 3'-phosphate/5'-OH into 3'-OH/5'-phosphate nick by the chromatin 3'-phosphatase/5'-OH kinase is a two step process; the simultaneous presence of the two activities within the same protein might have kinetical advantages, otherwise it remains a complete mystery. Acknowledgements : We thank Dr. David B. Ludlum of the Albany College of Medicine for the correction of our manuscript and many useful suggestions. This work was supported by grants from the Fonds Cancerologique de la C.G.E.R. and the Fonds de la Recherche Scientifique M6dicale. References 1. Habraken,Y. and Verly,W.G. (1983) FEBS lett., 160, 46-50. 2. Pheiffer,B.H. and Zimmerman,S.B. (1982) Biochem.Biophys.Res. Comm., 109, 1297-1302. 3. Cameron,V. and Uhlenbeck,O.C. (1977) Biochemistry, 16, 51205126. 4. von Sonntag,C., Hagen,U., Schon-Bopp,A. and Schulte-Frohlinde, (1981) Adv.Radiat.Biol., 9, 109-142. 5. Henner,W.D., Grundberg,S.M. and Haseltine,W.A. (1982) J.Biol. Chem., 257, 11750-11754. 6. Champoux,J. (1981) J.Biol.Chem., 256, 4805-4809. 8110