Purification and properties of T4 phage thymidylate synthetase produced by the cloned gene in an amplification vector.

We have introduced the T4 thymidylate synthetase gene, resident in a 2.7-kilobase EcoRI restriction fragment, into an amplification plasmid, pKC30. By regulating expression of this gene from the phage lambda pL promoter within pKC30 in a thyA host containing a temperature-sensitive lambda repressor, the T4 synthetase could be amplified about 200-fold over that after T4 infection. At this stage, a 20-fold purification was required to obtain homogeneous enzyme, mainly by an affinity column procedure. The purified plasmid-amplified T4 synthetase appeared to be identical with the T2 phage synthetase purified from phage-infected Escherichia coli in molecular weight, amino end group analysis, and immunochemical reactivity. The individual nature of the phage and host proteins was revealed by the fact that neither the T2 nor the T4 enzyme reacted with antibody to the E. coli synthetase, nor did antibody to the phage enzymes react with the E. coli synthetase. These differences were corroborated by DNA hybridization experiments, which revealed the absence of apparent homology between the T4 and E. coli synthetase genes. The techniques and genetic constructions described support the feasibility of employing similar amplification methods to prepare highly purified thymidylate synthetases from other sources.

impaired. This type of selectivity has been shown recently in the case of the synthetases from T2 phage and its host Escherichia coli with PteGlu6 (9).
Although most studies on thymidylate synthetase have been reported for the Lactobacillus casei enzyme, because of its abundance (10,11) and the availability of extensive data on its structural properties and primary sequence (12), it is difficult to extrapolate this information from the L. casei synthetase to that from other sources. To provide an understanding of the differences in responsiveness of the E . coli and T-even phage synthetases to folate derivatives, structural information is required which is not readily available because of the limited quantities of these enzymes. This problem is being resolved by manipulating the cloned thymidylate synthetase genes (13,14).
The T4 phage thymidylate synthetase is also of biological interest from the viewpoint of its regulation. The td gene, which codes for the enzyme, is part of a co-transcribed cluster of immediate early genes involved in nucleotide metabolism, including those for folate reductase and ribonucleoside diphosphate reductase (14,15). Moreover, it appears that coregulation of these enzymes as well as others involved in deox-yribonucleotide synthesis occurs post-translationally by formation of an active multienzyme complex (16). The regulation of these enzymes at several levels may account for discrepancies in their apparent order of expression in vivo and in vitro (17). Further interest in T4 thymidylate synthetase is generated by the fact that the enzyme is a structural component of the phage base-plate (18,19).
To increase our understanding of the regulation, mode of action, effect of mutation, and functional domains of thymidylate synthetases from various sources, we have amplified this enzyme by using DNA-cloning techniques. This paper describes the linkage of the T4 td gene, which has previously been isolated (14), to the strong regulatable promoter PI, of phage h in expression plasmid pKC30 (20). Purification and properties of the amplified gene product are also presented along with a preliminary comparison of the phage and host enzymes.

Materials
Strains, Vectors, a n d Recombinant Plasmids The bacterial and bacteriophage strains used in this study are listed in Table I. The recombinant plasmids which were constructed by using the plasmid vectors pBR322 (21) and the expression plasmid pKC30 (20) are described in Table I and Fig. 1. pKC3O comprises the pL-containing HindIII-BamHI fragment of phage h cloned into the HzndIII-BamHI interval of pBR322.

Preparation of Phage and Plasmid DNAs and DNA Fragment Purification
Phage A was prepared and purified and the DNA extracted in 50% formamide, as described by Davis et al. (24). Plasmid DNA to he used as a vector or for fragment isolations was purified by handing in CsC1ethidium bromide gradients (25). Rapid plasmid preparations were made by the procedure of Holmes and Quigley (26). Fragment purification was achieved by electrophoresis onto strips of DEAE-cellulose (Schleicher & Schuell, NA-45, 0.45 p~) and subsequent elution by adaptations of the procedure of Dretzen et al. (27).

Construction of Recombinant Plasmids
pBTd-EcoRI-treated hT4tdl (500 ng) was ligated to EcoRIcleaved pBR322 (250 ng) by incubation with 30 units of T4 DNA ligase in ligase buffer at 15 "C for 18 h. The ligation mixture was used to transform a CaC12-treated RuelO (Thy-) culture (28). While no Thy+ colonies were obtained when transformants were selected directly on minimal medium, about 10% of ampicillin-resistant transformants produced zones of weak but discernible growth when replica plated onto minimal medium plates. Several of these transformants  (14). comprises the 2.7-kb td-containing EcoRI fragment of phage T4 cloned into the EcoRI replacement interval of the Xirnm2l vector NM816 (23). The arrows show the direction of transcription from p~, which is located beside the immunity region (irnm), as well as the transcriptional orientation of the td gene. The solid circle marks the position of the X attachment site. The 2.7-kb T4 td fragment was subcloned into the EcoRI site of pBR322. These constructions are designated pBTd. B, cloning the T4 td fragment into expression plasmid pKC30. The 2.7-kb EcoRI restriction fragment was excised from a pBTd derivative, purified, and made blunt-ended by treatment with E. coli DNA polymerase I in the presence of the four deoxynucleotide triphosphates. This bluntended td fragment was then ligated into the HpaI sites of pKC30,321 nucleosides downstream from the hp,, promoter. Recombinants carrying the T4 td insert in the correct transcriptional orientation relative top,. are designated pKTd.
were purified and used for preparation of plasmid DNAs by rapid isolation procedures. Restriction analyses and Southern blotting experiments ( Fig. 6) confirmed the presence of the 2.7-kb EcoRI fragment of T4 cloned in either orientation into the EcoRI site of pBR322. These constructions, designated pBTd, were phenotypically Thywhen restreaked onto minimal medium plates lacking thymine and produced only very low levels of thymidylate synthetase ( Table 11, line 4).
pKTd-The 2.7-kb EcoRI T4 td fragment was purified from EcoRI-digested pBTd on an 0.8% agarose gel. The 5' extensions were filled in by incubating the fragment (10 ng) with DNA polymerase I of E. coli (29) in the presence of the four deoxynucleoside triphosphates (5 p~) in 20 mM Tris-HC1 (pH 7.5), 10 mM MgCL, 1 mM 2mercaptoethanol, and 50 pg of bovine serum albumin/ml at 15 "C for 2 h. The reaction mixture was heated to 65 "C for 10 min followed by blunt end ligation of the fragment to HpaI-cleaved pKC30 with T4 DNA ligase. This mixture was used to transform competent RuelO(Xc+) lysogens. Again no Thy' transformants appeared on minimal medium, but 2% of the ampicillin-resistant transformants produced patches of growth on minimal medium replica plates. Of the eight transformants purified, all contained the 2.7-kb T4 td fragment in the correct transcriptional orientation relative to PI.. These recombinants, designated pKTd, were used to transform cells containing a temperature-sensitive X repressor (RuelO(Xc1ts) or MBI47(hcItsdef)). All of the transformants overproduced T 4 thymidylate synthetase after the cultures were shifted from 32-42 "C. Recombinant pKTd2 was chosen for further study.

Hybridization and Autoradiography
Restriction endonuclease-digested DNAs were separated on an 0.8% agarose gel and transferred to a nitrocellulose filter (Schleicher & Schuell) as described by Thomas (30). Nick-translated "'P-labeled  Degree of enhancement of thymidylate synthetase activity Enzyme activity was measured as described under "Experimental Procedures" in crude cell-free extracts. Trimethoprim (0.3 mM) was added to the T4-infected extract to inhibit dihydrofolate reductase (line 2). All cells were grown in TBYET medium to midlog phase. Phage infections and temperature inductions were carried out for the times indicated, and the cultures were chilled to 2 "C in an ice-ethanol bath before harvesting. ' Cultures grown at 32 "C and shifted to 42 "C for the Cultures grown a t 37 "C.
__ times pBTd (0.1 pg) was used as a probe (5 X 10' cpm/pg). Hybridization was carried out at 42 "C for 16 h in the presence of 10% dextran sulfate and 50% formamide (31). Hybridization was detected by autoradiography of the filter at -80 "C with a DuPont-Cronex Lightning-Plus intensifying screen.

Temperature-induced Enzyme Synthesis
Transformants of lysogens harboring a temperature-sensitive X repressor were grown in TBYET medium containing 50 pg of ampicillin/ml a t 32 "C. Small cultures (<250 ml) were shifted directly to a 42 "C water bath shaker and maintained a t this temperature for the duration of the induction period. Large scale cultures were brought to 42 "C in I-liter batches in a Litton 1520 microwave oven monitored with a temperature probe2 and were maintained a t 42 "C in a high capacity New Brunswick air shaker. Both the specific activity and total yield of thymidylate synthetase in microwave-heated cultures were at least as high as in cultures brought to 42 "C by conventional methods.

Extract Preparation
Cells grown in TBYET medium for at least 8 generations to an Asso of 0.2-0.3 under conditions specified in the figure and table legends were chilled in an ice-ethanol bath to 2 "C and harvested by centrifugation. Pellets were washed in ice-cold 10 m~ Tris-HC1 (pH 7.4), quick frozen in a dry ice-ethanol bath, and stored at -80 "C. Frozen pellets were resuspended in 3 volumes of Buffer A (20 mM Tris-HC1 (pH 7.4), 10 m~ MgCI,, 1 m~ EDTA, and 10 mM /3mercaptoethanol), and the cells were broken by sonication. Cell debris was removed by centrifugation. Crude extracts could be stored at -80 "C for several months and thawed a t least twice without appreciable loss of synthetase activity.

Assays of Enzyme Activity
Thymidylate synthetase was measured by the spectrophotometric assay of Wahba and Friedkin (32). In some cases, the tritium release assay of Roberts (33) was used with [5-,"H]dUMP as substrate. One unit of activity is defined as the amount of enzyme required to convert 1 pmol of dUMP to dTMP/min at 30 "C.

Enzyme Purification
The phage synthetases were purified to homogeneity from crude extracts by a modification of the quinazoline affinity column procedure of Rode et al. (34).

Ternary Complex Formation
T o form this complex, enzyme, FdUMP, and 5.10-CHzH4PteGlu were incubated for 10 min a t 37 "C in a solution containing 100 mM potassium phosphate (pH 7.1), 10 mM ascorbate, and 100 mM pmercaptoethanol. 5,lO-CHzH4PteGlu was present a t 0.6 mM and [6-3H]FdUMP at 0.6 p~ or [2-'4C]FdUMP at 25 pM. When it was important to ensure conversion of all of the enzyme present in a sample to the ternary complex (as in Fig. 2), the sample was incubated for an additional 10 min in the presence of unlabeled FdUMP (250 p~) .
After incubation, the samples were either lyophilized or added directly to electrophoresis loading buffer.

Gel Electrophoresis
All samples were heated at 95 "C for 3 min in loading buffer containing 1% SDS and 1% P-mercaptoethanol before being applied to the gels. Separations were performed on either a 20% acrylamide low cross-linked gel (bisacrylamide/acrylamide ratio of 1:300) or on a 12.5% acrylamide high cross-linked gel (ratio 8:300). The gel composition, electrophoresis conditions, and differences in separating properties of the two gel types have been described (35). T4 thymidylate synthetase is more clearly resolved from neighboring bands after Coomassie blue staining in the low cross-linked gels.

Carboxymethylation and Sequence Determination
Enzyme (5-6 mg) was reduced in 6 M guanidine-HC1, 1 M Tris-HC1 (pH 8.0), and 0.1 M 2-mercaptoethanol for 4 h at room temperature under a nitrogen atmosphere. S-Carboxymethylation was accomplished with iodoacetate (36). After 20 min, the reaction was terminated by adding 2-mercaptoethanol to a final concentration of 1%. The carboxymethylated protein was then dialyzed a t 4 "C against four changes of 5% acetic acid and lyophilized.
The carboxymethylated proteins (50 nmol) were sequenced by using 0.1 M Quadrol as buffer with a Beckman Model 890 automatic sequencer. The released thiazolinone derivative were converted to their respective phenylthiohydantoin-amino acids by treatment with 1 N HC1 for 10 min a t 80 "C. The latter were characterized by thin layer chromatography on silica gel (37) and quantitated by high performance liquid chromatography in a modified Waters Associates Model 440 system equipped with a Model 660 solvent programmer. The ClxpBondapak columns (Waters) were developed with a gradient formed from 0.01 M sodium acetate (pH 5.0) and 38% acetonitrile in 0.01 M sodium acetate (pH 5.0) (38).

Preparation of Antibody to Thymidylate Synthetase
Antibody to T2-induced thymidylate synthetase was prepared in rabbits (39), and the globulin fraction was partially purified by ammonium sulfate fractionation as described previously (40).
The yglobulin fraction was dialyzed against a 5 mM potassium phosphate (pH 7.1), 0.85% NaCl solution and stored a t -10 "C. Antibody to the protein antigen was determined by immunodiffusion on a microscope slide (41). Precipitin lines were visible after several hours.

Amino Acid Analysis
Carboxymethylated protein samples were hydrolyzed in sealed evacuated tubes containing 1.0 ml of constant boiling HCl for 24 h a t 110 "C. Single column amino acid analysis was performed on a Model 119CL Beckman analyzer in triplicate.

Expression of the T4 td Gene in Plasmid Vectors
Subcloning the T4 td Gene-Phage XT4td1, which contains the td gene of T4, was selected by its ability to complement thymidylate synthetase-negative (thyA) mutants of E. coli (14). The 2.7-kb EcoRI fragment of T 4 , on which the td gene resides, is carried in this A phage in the orientation shown in Fig. 1. This arrangement permits gene expression from the distant pI, promoter of phage A during a lytic infection. Lysogens of AT4td 1 made in a thyA host have been reported to be

Cloned T4 Phage Thymidylate Synthetase
Thy+, suggesting that the T4 insert includes a sequence that can function as a promoter for the td gene since the main promoters are repressed in the lysogenic state (14). To obtain this 2.7-kb td fragment from XT4td 1 for further study, it was subcloned into the EcoRI site of pBR322. ThyA cells transformed by these constructions, designated pBTd, could not be selected on Thy-medium and contained only trace amounts of thymidylate synthetase, with the insert in either orientation (Table 11, line 4). This observation implies that if there is a promoter-like sequence within the insert, it functions very inefficiently. The presence of the insert was verified by restriction analysis (data not shown) and Southern hybridization (see below).
To amplify expression of the td gene, the 2.7-kb td fragment was placed immediately proximal to the A p l . promoter contained within expression plasmid pKC30 (20). E. coli DNA polymerase I was used to make the 5"EcoRI extensions double-stranded. Subsequently, this blunt-ended fragment was ligated into the HpaI site of pKC30, 321 base pairs from the PI, transcriptional start site. Because we were concerned about potentially detrimental effects of T4 thymidylate synthetase overproduction in E. coli and aware of the fact that dTMP excess can be both mutagenic (42) and lethal (43) in yeast, transformants were selected in a thyA-lysogen, RuelO(Xc+). The X repressor protein (cI) in these cells prevents transcription frompL and thereby permits the cloning of lethal functions which would be expressed if CI were not present (20). Ampicillin-resistant transformants which appeared on replica plates to be weakly Thy' were screened by size and restriction analysis. All eight of the transformants checked contained the 2.7-kb fragment in the correct transcriptional orientation relative to PI.. This bias may be accounted for by low level transcription originating from p1. in the presence of X repressor (Table 11, line 5), leading to a weak Thy' phenotype, an effect which influenced our choice of recombinants in favor of those with the td gene oriented appropriately relative to p~. Low level transcription from p~ may also account for the fact that XT4td 1 lysogens are Thy' (14).
To induce high level expression from PI,, pKTd2 was transferred into thyA lysogens harboring prophages with a temperature-sensitive repressor gene (cIts). These cells were maintained at 32 "C and were shifted to 42 "C when expression of the td gene was desired. The mutant repressor was inactivated at 42 "C, enabling transcription to proceed from PI..
High Level Production of T4 Phage Thymidylate Synthetase-Lysogens containing the multicopy pKTd2 plasmid (Table 11, lines 6 and 7) produced about 4-5-fold more enzyme before temperature shift than was obtained after T4 phage infection (line 2). This was roughly the same degree of amplification (4.5-fold) as was effected by infection with hT4td1, where the td gene is under remote control of PI, in a lytic infection (line 3). However, 30 min after the temperature was raised to 42 "C, the synthetase was elevated 115-fold in RuelO(XcIts)/pKTdP (line 6) and 155-fold in the defective lysogen MB147(hcItsdef)/pKTd2 (line 7). Since cell lysis occurs beyond 30 min in the former lysogen, later time points were feasible only with the strain harboring a defective prophage. In this experiment, a 240-fold amplification was achieved in 60 min (line 7), corresponding to 480 milliunits/ mg of protein. Specific activities as high as 700 milliunits/mg, which represent 5-7% of the total soluble protein, have been obtained.
Identification of Thymidylate Synthetase in Crude Extracts-Since T4 phage thymidylate synthetase comprises such a high percentage of the total cellular protein after derepression ofp1, (5-7%), the protein profiie in crude extracts from MB147(XcItsdef)/pKTd2 raised to 42 "C was compared with extracts from cells maintained at 32 "C. In the Coomassie-stained SDS gel in Fig. 2

, an additional band is evident in the 42 "C cells (lane 4 versus lane 2). This band corresponds in size to purified T2 phage thymidylate synthetase (lane 6).
In gels which resolve smaller peptides, an additional band (about 11,000 daltons) was also seen in heat-treated extracts (data not shown). This band probably corresponds to polypeptide y of unassigned function (11,300 daltons) previously identified by Mileham et al. (14). To verify that the amplified band in Fig. 2 is indeed T4 phage thymidylate synthetase, crude enzyme preparations were treated with [14C]FdUMP and 5,lO-CH2H4PteGlu to determine whether the characteristic thymidylate synthetase ternary complex would be formed (7). While no change was noted in the 32 "C banding pattern (lane 3), the putative T4 phage synthetase subunit in lane 4 migrated more slowly after ternary complex formation (lane 5 ) . A similar increase in subunit molecular weight was noted with purified T2 phage synthetase (lanes I and 7). In the accompanying autoradiogram, the radiolabeled bands correspond to T4 phage synthetase ternary complex in the heatinduced crude extract (lane 5 ) and to the T2 phage synthetase ternary complex (lunes I and 7), confirming the identity of the heat-induced band in lane 4 as belonging to T4 phage thymidylate synthetase, the product of the cloned td gene.
The molecular weight estimate derived from our gels (Figs. 2 and 3) for the T2 and T4 synthetase subunits was about 32,000. This is in exact agreement with our previous estimates for the T2 phage enzyme, based on SDS-gel electrophoresis and sedimentation equilibrium measurements (39). It should be noted, however, that values in the range of 25,000 daltons (44) to 29,000 daltons (14,45) have been obtained for the T4 phage enzyme. Discrepancies in molecular weight estimates from polyacrylamide gels are common (35) and our molecular weight assignments must be considered tentative. Nevertheless, it is clear from Fig. 3 that the T2 synthetase ternary complex (lane 2), that of the T4 phage enzyme from extracts of heat-treated pKTd2-containing lysogens (lane 4 ) , and that formed by T4 synthetase purified from this source (lane 3 ) appear identical in size. Furthermore, these are indistinguishable in size from the derivatized subunit encoded by wild type T4 phage (lane 5) and by the XT4tdl-transducing phage (lane 6). This method clearly distinguishes the T-even enzymes from those of E. coli (lanes 7 and 8 ) , Bacillus su6tili.s (lanes 9 and IO), and H-35 hepatoma cells (data not shown).

Comparative Analysis of T-Even a n d E. coli Thymidylate Synthetases
The plasmid-amplified T 4 td gene product was purified to homogeneity by affinity chromatography (Fig. 4), as were the enzymes produced after T2 and T4 infection. These products were then sequenced from their NH2-terminal ends, and for a t least 20 cycles, the amino acid sequences of the T2-induced, T4-induced, and TCcloned enzymes appeared to be identi-

Asn-Gly-Tyr-Glu-Thr-Asp-Asp-.
This information, coupled with the molecular weight data in Figs. 3 and 4, suggests that T2-and T4-induced synthetases are very closely related, if not identical. The immunochemical similarity of these proteins is also evident from the fact that the T2-induced (Fig. 5A, wells 1 and 4 ) and T4-cloned synthetases (wells 2 and 5) both cross-reacted with antibody to the former enzyme. No cross-reaction was obtained with the E. coli synthetase. Similarly, antibody to the latter reacted only with the E. coli synthetase (Fig. 5B, wells 3 and 6) and not with the phage enzymes.
These differences between the T4 and E. coli enzymes are corroborated by apparent differences in their coding sequences. The Southern hybridization experiment, in which "'P-labeled pBTd was used as probe (Fig. 6), showed the expected homology between the probe and the 2.7-kb EcoRI fragment in both XT4tdl (lane b ) and pBTd (lane c). In the latter case, the 4.4-kb band corresponds to the vector portion of pBTd. No hybridization was evident, however, either with X vector alone (lane a) or with a 7.8-kb E. coli thyA fragment cloned into X (13) (lane d). This absence of apparent homology between the T4 and E. coli thymidylate synthetase fragments under these hybridization conditions is somewhat surprising in view of certain homologies in amino acid sequence between the phage and host enzymes.'  "'P-labeled pBTd DNA, which comprises the 2.7-kb EcoRI T4 td fragment cloned into the EcoRI site of pBR322, was used as probe. Lane a, 10 ng of EcoKIdigested XNM816 vector DNA; lane b, 10 ng of EcoKI-digested XT4tdl containing the 2.7-kb td fragment; lane c, 1 ng of EcoRIdigested pBTd plasmid lane d, 10 ng of HindIII-digested XthyA DNA containing the 7.8-kb Hind111 fragment encoding thymidylate synthetase of E . coli. DISCUSSION T o obtain adequate amounts of T-even phage thymidylate synthetases for comparison with each other, as well as with the E. coli enzyme and with those from other sources, it was found advantageous to exploit the expression plasmid pKC30 described by Shimatake and Rosenberg (20). By positioning a 2.7-kb fragment containing the td gene of T 4 phage downstream from the phage h PI, promoter residing within this plasmid (Fig. l), the expression of the td gene was amplified more than 200-fold above that produced by phage infection (Table 11). This was achieved in lysogenic cells containing a thermosensitive repressor gene (cI), which could be inactivated by raising the temperature of the cultures to 42 "C. Transcription could then proceed fromp,, into the adjacent td gene to an extent that the gene product represented as much as 5-7% of total cellular protein in crude extracts prepared from such cultures. Similar results have been obtained by using pKC30 to amplify the expression of the cII gene of phage X (20) and the uurA gene of E. coli (46). The use of thyA cells lacking host synthetase aided in the selection of plasmids expressing the td gene and in the ensuing enzyme purification since extracts were devoid of the E. coli synthetase. Purification of the amplified T4 synthetase was greatly simplified with an affinity column procedure (34).
The amplified T4 td gene product is easily visible in gel patterns of crude cell extracts prepared after inactivation of the temperature-sensitive CI repressor (Fig. 2, lane 4 ) . The identity of the amplified band was verified either directly on the Coomassie-stained gel or on autoradiograms by converting this band to its corresponding FdUMP-5,10,CH2H4PteGluternary complex.
Comparison of the amplified T4 enzyme with the previously purified T2 synthetase (39) revealed them to be identical in subunit molecular weight (32,000), amino end group analysis through 20 amino acids, and immunochemical cross-reactivity (Fig. 5). These data and the apparent sequence homology in the td region of the T2 and T4 genetic maps (47) suggest that the T2 and T4 synthetases may be identical.
In contrast to the similarities between the T-even phage synthetases, striking differences have been found between the phage and host enzymes. We have shown here that differences exist in the electrophoretic mobility (Fig. 3) and immunological specificity (Fig. 5) of the synthetases. These findings were corroborated by DNA hybridization experiments in which no homology was apparent between the td gene of the T4 phage and the thyA gene of E. coli K12 (Fig. 6). These differences are somewhat surprising in view of the similarities in amino acid content and peptide patterns previously reported by us for the T2 and E. coli B synthetases (48). On the other hand, their differential response to inhibition by the folylpolyglutamates (9) implicates distinctive differences in their folatebinding regions. A comparative sequence analysis of the Teven phage and E. coli synthetases and the genes encoding them is in progress. These results should not only help to explain the molecular basis for the differential behavior of these proteins, but may also give us some insights into the evolutionary relationships between duplicated phage and host functions.