JOURNAL OF BACTERIOLOGY, June 1990, p. 2935-2939 Vol. 172, No. 6 0021-9193/90/062935-05$02.00/0 Copyright © 1990, American Society for Microbiology Bacteriophage T4 Deoxynucleotide Kinase: Cloning and Purificationt GEORGE S. BRUSH, SATISH K. BHATNAGAR, AND MAURICE J. BESSMAN* McCollum-Pratt Institute and Department ofBiology, The Johns Hopkins University, Baltimore, Maryland 21218 Received 9 November 1989/Accepted 2 March 1990

Gene 1 of bacteriophage T4 has been cloned into a A PL expression vector, resulting in the overproduction of deoxynucleotide kinase. A procedure that includes affinity chromatography on Cibacron Blue F3GA-agarose has been used to purify milligram quantities of from a 2-liter culture. The enzyme has been partially characterized in vitro and in vivo, and it appears to be identical to the deoxynucleotide kinase isolated from T4-infected Escherichia coli. These results prove the earlier contention that the of three dissimilar deoxynucleotides (5-hydroxymethyldeoxycytidylate, dTMP, and dGMP), to the exclusion of most others, is catalyzed by a single .

In Escherichia coli, Saccharomyces cerevisiae, and higher MATERIALS AND METHODS eucaryotes, four distinct and highly specific enzymes cata- lyze the phosphorylation of the canonical deoxynucleoside Chemicals and chromatographic media. All deoxynucle- monophosphates to the corresponding diphosphates accord- otides were purchased from P-L Biochemicals, Inc., and ing to the following equation: deoxynucleoside monophos- Sigma Chemical Co., except for dHMP, which was prepared phate + ATP = deoxynucleoside diphosphate + ADP. essentially as previously described (8, 9). ATP, phosphoe- However, after infection of E. coli with the T-even phages nolpyruvate, P-NADH, streptomycin sulfate, and ampicillin (T2, T4, and T6), enzymatic activities are induced (EC were also purchased from Sigma. The Geneclean kit for 2.7.4.12) that catalyze the phosphorylation of three deoxy- isolating DNA fragments from agarose gels was purchased monophosphates (5-hydroxymethyldeoxycytidy- from Bio-101. Sepharose 6B was obtained from Pharmacia late [dHMP], dTMP, and dGMP); these three activities Fine Chemicals, and Affi-Gel Blue was from Bio-Rad Labo- appear to reside in one protein (2, 3, 8). The enzyme is ratories. specific for these three structurally dissimilar (2, Enzymes. T4 DNA and restriction endonucleases 8), and it would be of considerable interest to discover how BamHI, HaeII, HpaII, and SmaI were purchased from the protein recognizes these nucleotides to the exclusion of Bethesda Research Laboratories, Inc. Restriction endonu- most others. cleases AvaIl, BsmI, and MnlI were from New England Since only small quantities of the kinase can be extracted BioLabs, Inc. Calf intestinal alkaline was from from large-scale cultures of phage-infected E. coli, the Boehringer Mannheim Biochemicals. Lactic dehydrogenase number and types of experiments that have been performed (type II) and (type II) were from Sigma. T4 to date have been limited. Sakiyama and Buchanan (21) DNA was from both Bethesda Research Labo- reported that the T4 protein exists as a homodimer with a ratories and Boehringer Mannheim. subunit molecular weight of 23,000. This value is comparable Bacteria, bacteriophages, and plasmids. E. coli B, CR63 to the molecular weight of 27,300 calculated from the DNA (22), DH5a, HB101 (4), and JM83 (24) were taken from sequence data of Broida and Abelson (5). Duckworth and stocks maintained in this laboratory at -80°C in 8.0% Bessman (8) provided supportive evidence that only one dimethyl sulfoxide. Bacteriophages T4D and T4amE957 (8) enzyme, encoded by gene 1 of T4 phage, is responsible for were from liquid stocks kept at 4°C. We are very thankful to the three activities. It has also been shown with T4 phage Donna Duckworth for providing us with the amber mutant. that dHMP, dTMP, and dGMP are competitive inhibitors The expression vector pHE6 (18) and the plasmid pTFR1501 of each other (2). This would imply the existence of only (J. Velten, Ph.D. thesis, Department of Chemistry, Univer- one for the three substrates, overlapping sites, sity of California, San Diego, 1981) were the generous gifts or -induced conformational changes in the protein. of Lawrence Grossman and John Abelson, respectively. The resolution of this question will require substantial Growth media. All media used were prepared as described quantities of pure enzyme. In this report we describe a by Miller (17). For antibiotic selection, ampicillin was added procedure for the molecular cloning of gene 1 into an to a final concentration of 50 ,ug/ml. expression vector, the overproduction of the enzyme, and Assays. The method of Lowry et al. (15) was used to the large-scale preparation of the pure protein. We also determine protein concentrations, with bovine serum albu- demonstrate unequivocally that one protein is responsible min as the standard. To determine enzymatic activity, a for all three activities. spectrophotometric assay was used in which the formation of nucleoside diphosphates from monophosphates is coupled to the oxidation of 1-NADH (8, 12, 14). The conditions were as previously described (8), with the following exceptions. (i) * Corresponding author. Purified pyruvate kinase and lactic dehydrogenase fractions t Publication no. 1452 of the McCollum-Pratt Institute. were used rather than a crude extract. As a result, no 2935 2936 BRUSH ET AL. J. BACTERIOL.

detectable nucleoside diphosphokinase was present in the column equilibrated with buffer A. The flow rate was ad- system. Ten units of each enzyme was added to each justed to 0.4 ml/min, and fractions containing approximately reaction mixture. (ii) Approximately 1.5 to 15 U of deoxy- 75% of the loaded activity were pooled. kinase was added to the assay mixtures. As (v) Affinity chromatography. A 2.5- by 16-cm Cibacron before, 1 U of deoxynucleotide kinase is the amount of Blue F3GA-agarose (Affi-Gel Blue) column was equilibrated enzyme that catalyzes the phosphorylation' of 1 nmol of with buffer A, 'and 10 ml of the previous fraction was applied dTMP per min under the standard assay conditions. to the gel bed. The sample was washed into the column with Cloning procedures. Most of the cloning procedures were 2 column volumes of buffer A, followed by 5 column carried out by the methods of Maniatis et al. (16), except for volumes of 50 mM Tris chloride (pH 7.4)-i mM dithiothre- blunt-end ligations (6) and transformations (19). When cells itol-150 mM NaCl (buffer B). Activity was then eluted with were being transformed with recombinant pHE6 (18) plas- buffer B containing 5 mM ATP and 10 mM MgC12, and the mids, they were incubated at 30'C rather than 37°C after the active fractions were combined. Apparently, Cibacron Blue 42°C pulse. In addition, the plated cells and small-scale binds to the same site of the enzyme as ATP, the phosphoryl cultures used for plasmid preparations were incubated at donor in the reaction (2, 8). 30°C. These modifications were included so that transcrip- (vi) Pressure filtration. The pooled Affi-Gel eluent was tion from the k PL promoter of pHE6 would be repressed by concentrated by filtration through an Amicon PM10 mem- the temperature-sensitive cI857 repressor during cell recov- brane and then dialyzed extensively versus buffer A. Glyc- ery and growth. erol was added to a final concentration of 20%, and the Complementation. These studies were done by modifying purified enzyme was stored at -80°C. the single-step bacteriophage growth curve procedure de- Polyacrylamide gel electrophoresis. Denaturing gel electro- scribed by Adams (1). Bacteria were grown in H broth phoresis was carried out by the method of Laemmli (13), and containing ampicillin at 30°C to an A6. of 0.3 and then native slab gel electrophoresis was a modification of the disc infected with phage at a low multiplicity (-0.02). The method of Ornstein (20) and Davis (7). temperature was shifted to 42°C to allow for analysis. An Applied Biosystems 470A gas- from recombinant plasmids. After a 5-min incubation period phase protein sequencer coupled to a PTH analyzer (model for phage adsorption, a 10-,ul sample, was filtered onto a 120A) was used to determine the N-terminal amino acid Micropore disc (0.45-,um pore size), and the retained cells sequence of'the purified enzyme. were washed with H broth. This filter was then placed in a tube containing 10 ml of H broth (tube 1) and incubated with RESULTS aeration at 42°C. At regular intervals, samples of this culture were plated on H-agar petri plates, using H soft agar with E. Subcloning of gene 1. The expression vector pHE6, which coli CR63 as the indicator. Plaques were scored after an contains both the strong A PL promoter upstream from a overnight incubation at 37°C. multiple cloning site and the gene for the temperature- Enzyme purification. (i) Crude extract. E. coli HB101 sensitive repressor cI857, was used to overproduce T4 containing pBK5 was used for the enzyme purification. deoxynucleotide kinase. This plasmid also contains the These cells were grown in 2 liters of LB with ampicillin at 1-lactamase gene, allowing for ampicillin selection of trans- 30°C until an A6. of 0.3 was reached. At this point the formed cells. culture was shifted to 42°C to activate K PL and induce the Velten (Ph.D. thesis) cloned several fragments from the production of deoxynucleotide kinase, and the culture was T4 tRNA region into pBR322, some of which contained the incubated overnight. proximal gene 1. One of the recombinant plasmids, The cells were harvested by centrifugation at 4,200 x g pTFR1501, was the source of this gene for our subcloning (this and subsequent steps were carried out at 4°C) and then procedure. A detailed restriction map of the cloned region washed with 100 ml of 0.5% NaCl-0.5% KCl. After a second could be generated from the map of Fukada et al. (10) and centrifugation, the cells were suspended in approximately 3 the nucleotide sequence determined by Broida and Abelson volumes of 50 mM Tris chloride'(pH 7.4-1 mM dithiothre- (5). The cloning scheme is shown in Fig. 1. itol (buffer A) and broken by one passage through a French The insert of the constructed plasmid pBK3 includes T4 pressure cell. The cell debnrs was removed by centrifugation 1 and 57A and an upstream E. coli consensus promoter for 20 min at 10,000 x g, and the supernatant was diluted to sequence (P) (5). Since this regulatory region might be approximately 10 mg/ml with buffer A. recognized by E. coli RNA polymerase (5), the transcription (ii) Streptomycin fractionation. To remove nucleic acids of gene 1 may not be entirely under the control of the A PL and associated , a streptomycin precipitation step promoter in this recombinant plasmid. Furthermore, cells was included. The crude extract was stirred while a fresh 5% containing pBK3 overproduce not only deoxynucleotide (wt/vol) streptomycin sulfate solution (in buffer A) was kinase but also a smaller protein, which'is presumably the added dropwise (0.2 ml per ml of extract). The resulting product of gene 57A (data not shown). The final step in the mixture was stirred for 10 min, and the precipitate was cloning procedure removes the possible promoter region as discarded after a 15-min centrifugation at 10,000 x g. well as gene 57A. Cells transformed with the resulting (ui) Ammonium sulfate fractionation. The streptomycin plasmid, pBK5, have very high deoxynucleotide kinase fraction was stirred while solid ammonium sulfate was activity (see below). slowly added to 45% saturation (11). After 15 min of addi- Complementation of phage mutant in gene 1. Although the tional stirring, the sample was centrifuged at 20,000 x g for in vitro studies indicated that deoxynucleotide kinase activ- 15 min, and the supernatant was discarded. The precipitate ity was being overexpressed by cells containing pBK5, it was dissolved with buffer A to 4.6 ml. This fractionation was of interest to determine whether enzyme was being served both as a purification step and as a concentration step produced that would function in vivo. Bacteriophage in preparation for the gel filtration column. T4amE957, which is defective in gene 1, was used in this (iv) Sepharose 6B chromatography. The ammonium sulfate experiment, as it cannot grow in strains of E. coli lacking an fraction was passed through-a 2.5- by 88-cm Sepharose 6B amber suppressor. Two transformants of E. coli JM83, one VOL. 172, 1990 T4 DEOXYNUCLEOTIDE KINASE OVERPRODUCTION 2937

2/64 3 1 57A 57B ipI tRNA ~~~ H II _,_ ~ ~ ~~ Deletion EcoRI HpalI HaeII

1. Hpoll, Haell digestions 2. 1.2 kb fragment isolation 3. End blunting Mnll BsmI

1. SnaI digestion Ligtion 2.

1. Mnlt, Bsml digestions 1. SnaI digestion 2. 0.8 kb fagment isolation 2. Deplosplorylation 3. End blunting ~l-

(Mnlt) (Bsml)

pBKS AmpR

4.8 kb

FIG. 1. Subcloning of gene 1. The arrows under gene 1 indicate the direction of transcription. Only the pertinent HpaII, HaeII, MnlI, and BsmI sites are shown. Plasmid sizes are given in kilobases (kb).

containing pHE6 and the other containing pBK5, were of the N terminus was determined to confirm the identify of infected with either T4D (wild type) or amE957, and single- the enzyme and assess its homogeneity. A fraction V sample step growth (1) was followed. Both cell types could support was precipitated with ammonium sulfate at 80% saturation the growth of wild-type T4 (data not shown). However, (11), and the pellet was dissolved in a small volume of buffer results with amE957 show that cells containing vector alone could not support growth, whereas complementation did take place with the pBK5-transformed cells (Fig. 2). This 300 indicates that bacteria containing pBK5 produce biologically 250 functional T4 deoxynucleotide kinase. Enzyme purification. A summary of the purification pro- 200 0 cedure (see Materials and Methods) is presented in Table 1. 150 * - * .I I. Various samples were subjected to polyacrylamide gel elec- A. trophoresis (Fig. 3). The denaturing gel revealed an intense 100 band in the HB1O1(pBK5) crude extract that was not present 50 in the other extracts. The position of this band indicates a i 0 0000 - p p p molecular weight of approximately 28,000, close to the 10 20 30 40 50 60 70 molecular weight of 27,300 predicted from the nucleotide Tine (min) sequence (5). Also shown are samples of fraction VI elec- FIG. 2. Complementation of phage defective in gene 1 (T4am trophoresed under denaturing and native conditions. These E957). Single-step growth curves of amE957-infected JM83(pHE6) lanes indicate a highly purified protein preparation. (0) and JM83(pBK5) (0). Results are given as PFU/p,l of tube 1 N-terminal sequence. The sequence of the first 10 residues culture. 2938 BRUSH ET AL. J. BACTERIOL.

TABLE 1. Purification of T4 deoxynucleotide kinase from 10 induced E. coli HB101 containing pBK5 8 Vol Protein % FractionFraction ((ml) (mg/ml) (kU/mg)aActivity Recovery I. Crude extract 37 9.9 96 100 II. Streptomycin 43 7.8 130 120 III. Ammonium sulfate 4.6 55 140 100 IV. Sepharose 56 3.3 150 74 V. Affi-Gel 1l0b 0.35 180 31 2 VI. Amicon 32 1.8 220 34 a 0 - is One kU is equivalent to 1 p.mol of dTMP phosphorylated to dTDP per 0 10 20 30 40 50 60 70 min. Protein b Only 10 ml of the Sepharose fraction was loaded onto the affinity column. (iig) Therefore, the experimental volumes of the Affi-Gel and Amicon fractions were multiplied by 5.6 to take this into account. The recoveries of these two FIG. 4. Substrate specificity. Reaction velocity versus protein fractions are based on these calculated values. (fraction VI) concentration determined by using 0.25 mM dHMP (0), 1.8 mM dTMP (0), 1.0 mM dGMP (x), 1.8 mM dAMP (0), and 1.8 mM dCMP (A). A. This concentrated solution was dialyzed extensively versus water, lyophilized, and sequenced (see Materials The crude extract of such a culture has a specific activity and Methods). The following unambiguous sequence was approximately 500 times greater than that of T4-infected found: Met-Lys-Leu-Ile-Phe-Leu-Ser-Gly-Val-Lys. This E. coli. am'ino acid sequence is colinear with that predicted from the Although a purification procedure for deoxynucleotide nucleotide sequence ATG AAA CTA ATC TTT TTA AGC kinase has been reported (8, 21), we chose a different course GGT GTA AAG (5). for two reasons. First, since the purified enzyme will be used Specificity. The canonical deoxynucleotides and dHMP for structural studies, we felt it was necessary to avoid were tested for their abilities to act as substrates of the fractionations that could denature even a small percentage of purified enzyme. Enzyme activity was observed with the preparation. The original procedure includes such a step, dHMP, dTMP, or dGMP, and this activity was proportional in which the sample is heated at 50°C for 10 min. Second, the to the enzyme concentration (Fig. 4). In the presence of the extremely high activity in the crude extract allowed us to other two deoxynucleotides (dAMP and dCMP), the prepa- devise a simpler procedure. This new scheme includes ration showed no appreciable activity (<0.5% of the dTMP chromatography with Cibacron Blue F3GA-agarose, a ma- value). This specificity is the same as that reported for the T4 trix that binds to many nucleotide-binding proteins (23). For deoxynucleotide kinase isolated from phage-infected cells our purposes, the affinity step provides an added advantage (8). in that only molecules capable of binding are recovered, ensuring that the final preparation consists of native enzyme. Nearly 60 mg of enzyme can be purified from 370 mg of DISCUSSION starting protein (Table 1). The power of the A PL expression The subcloning of gene 1 into pHE6 has resulted in the system is emphasized when this is compared with earlier dramatic overproduction of deoxynucleotide kinase. This work in which 0.48 mg of kinase was purified from 950 mg of enzyme constitutes between 40 and 45% of the protein in crude extract from T4-infected E. coli (8). The purification maximally induced E. coli HB101 cells containing pBK5. procedure has resulted in a preparation that is at least 95% pure asjudged by polyacrylamide gel electrophoresis. This is supported by the sequence analysis, in which no other proteins were detected. Duckworth and Bessman used biochemical and genetic B methods to show indirectly that T4 deoxynucleotide kinase ~>k is a single, trifunctional enzyme (8). The experiments pre- Fma sented here prove this contention, since high levels of these three deoxynucleotide kinase activities are seen to be asso- ciated with the cloning of a single gene. The availability of "a 11. large amounts of homogeneous protein resulting from the procedure described in this report will make possible a thorough investigation of the structure and catalytic function of the enzyme, perhaps leading to an understanding of the molecular basis for its unusual specificity. FIG. 3. Polyacrylamide gel electrophoresis. Gels were stained with Coomassie brilliant blue R. (A) Sodium dodecyl sulfate-20%o ACKNOWLEDGMENTS polyacrylamide gel of T4D-infected E. coli crude extract (lane 1), HB1O1(pHE6) crude extract (lane 2), HB1O1(pBK6) crude extract This work was supported by Public Health Service grant GM- (pBK6 contains the same insert as pBK5, but in the opposite 18649 from the National Institute of General Medical Sciences. orientation) (lane 3), fraction I (lane 4), fraction VI (lane 5), and the We thank P. Shenbagamurthi of The Johns Hopkins Protein/ following reference proteins (Sigma): egg albumin (45 kil6daltons), Peptide Facility for the N-terminal sequence analysis and Stephen carbonic anhydrase (29 kilodaltons), and trypsin inhibitor (20.1 Quirk'and Linda Bullions for their critical discussions. kilodaltons) (lane 6). The plasmid-containing strains were grown under identical conditions. Lanes 1 through 5 each contained 3.8 ,ug LITERATURE CITED of protein, whereas 11.4 ,ug was loaded in lane 6 (3.8 ,g of each 1. Adams, M. H. 1959. Bacteriophages, p. 473-485. Interscience standard). (B) Native gel (10%o polyacrylamide) offraction VI (4 ,ug). Publishers, Inc., New York. VOL. 172, 1990 T4 DEOXYNUCLEOTIDE KINASE OVERPRODUCTION 2939

2. Bello, L. J., and M. J. Bessman. 1963. The enzymology of assembly of the head of bacteriophage T4. Nature (London) virus-infected bacteria. IV. Purification and properties of the 227:680-685. deoxynucleotide kinase induced by bacteriophage T2. J. Biol. 14. Lieberman, I., A. Kornberg, and E. S. Simms. 1955. Enzymatic Chem. 238:1777-1787. synthesis of nucleoside diphosphates and triphosphates. J. Biol. 3. Bessman, M. J. 1959. Deoxyribonucleotide kinase in normal and Chem. 215:429-440. virus-infected Escherichia coli. J. Biol. Chem. 234:2735-2740. 15. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 4. Boyer, H. W., and D. Roulland-Dussoix. 1969. A complementa- 1951. Protein measurement with the Folin phenol reagent. J. tion analysis of the restriction and modification of DNA in Biol. Chem. 193:265-275. Escherichia coli. J. Mol. Biol. 41:459-472. 16. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular 5. Broida, J., and J. Abelson. 1985. Sequence organization and cloning: a laboratory manual. Cold Spring Harbor Laboratory, control of transcription in the bacteriophage T4 tRNA region. J. Cold Spring Harbor, N.Y. Mol. Biol. 185:545-583. 17. Miller, J. H. 1972. Experiments in molecular genetics. Cold 6. Cobianchi, F., and S. H. Wilson. 1987. Enzymes for modifying Spring Harbor Laboratory, Cold Spring Harbor, N.Y. and labeling DNA and RNA. Methods Enzymol. 152:94-110. 18. Milman, G. 1987. Expression plasmid containing the XPL pro- 7. Davis, B. J. 1964. Disc electrophoresis-II: method and applica- moter and cI857 repressor. Methods Enzymol. 153:482-491. tion to human serum proteins. Ann. N.Y. Acad. Sci. 121: 19. Morrison, D. A. 1979. Transformation and preservation of 404-427. competent bacterial cells by freezing. Methods Enzymol. 68: 8. Duckworth, D. H., and M. J. Bessman. 1967. The enzymology of 326-331. virus-infected bacteria. X. A biochemical-genetic study of the 20. Ornstein, L. 1964. Disc electrophoresis-I: background and the- deoxynucleotide kinase induced by wild type and amber mu- ory. Ann. N.Y. Acad. Sci. 121:321-349. tants of phage T4. J. Biol. Chem. 242:2877-2885. 21. Sakiyama, S., and J. M. Buchanan. 1973. Relationship between 9. Flaks, J. G., and S. S. Cohen. 1959. Virus-induced acquisition of molecular weight of T4 phage-induced deoxynucleotide kinase metabolic function. I. Enzymatic formation of 5-hydroxymeth- and the size of its messenger ribonucleic acid. J. Biol. Chem. yldeoxycytidylate. J. Biol. Chem. 234:1501-1506. 248:3150-3154. 10. Fukada, K., L. Gossens, and J. Abelson. 1980. The cloning of a 22. Stretton, A. 0. W., and S. Brenner. 1965. Molecular conse- T4 transfer RNA gene cluster. J. Mol. Biol. 137:213-234. quences ofthe amber mutation and its suppression. J. Mol. Biol. 11. Green, A. A., and W. L. Hughes. 1955. Protein fractionation on 12:456-465. the basis of solubility in aqueous solutions of salt and organic 23. Thompson, S. T., K. H. Cass, and E. Stellwagen. 1975. Blue solvents. Methods Enzymol. 1:67-90. dextran-sepharose: an affinity column for the dinucleotide fold 12. Kornberg, A., and W. E. Pricer, Jr. 1951. Enzymatic phosphor- in proteins. Proc. Natl. Acad. Sci. USA 72:669-672. ylation of adenosine and 2,6-diaminopurine riboside. J. Biol. 24. Vieira, J., and J. Messing. 1982. The pUC plasmids, an Chem. 193:481-495. M13mp7-derived system for insertion mutagenesis and sequenc- 13. Laemmli, U. K. 1970. Cleavage of structural proteins during the ing with synthetic universal primers. Gene 19:259-268.