Inhibitory effect of agents altering the structure of DNA on the synthesis of pyrimidine deoxyribonucleotides in bacteriophage T4 DNA replication.

Previous studies from this laboratory (Flanegan, J. B. and Greenberg, G. R. (1977) J. Biol. Chem. 252, 3019-3027) have shown that bacteriophage T4-infected cultures of Escherichia coli synthesize thymine and 5-hydroxymethylcytosine deoxyribonucleotides via the de nova reductive pathway with rates in the proportion of 2:1, respectively, i.e. exactly the ratio of thymine to 5-hydroxymethylcytosine in T4 DNA. The ratio of synthesis of these deoxyribonucleotides is maintained at 2:l even after infection by Dnamutants until their concentrations reach about 100 times the initial levels. To examine the possible role of DNA in this regulation, several types of agents altering the structure of DNA were tested in vivo for their abilities to affect the de novo synthesis of dTMP and HmdCMP. Mitomycin C, an agent causing interstrand cross-linking between guanine residues, was administered to cultures infected by a phage T4xstrain after early protein synthesis was completed. This agent caused a prompt inhibition of HmdCMP synthesis with little effect on dTMP synthesis. Little or no inhibition was observed in phage T4x+-infected cells. Since the product of gene x repairs DNA damage caused by mitomycin C, these findings provide strong argument that the action of this agent is through its reaction with DNA and not by direct inactivation of enzymes in the synthetic pathway to deoxyribonucleotides. Addition of the intercalating agent proflavine at 10 min after infection, or exposure of T4xphage to ultraviolet irradiation prior to infection, caused a decrease in both dTMP and HmdCMP synthesis. Proflavine also was shown to inhibit pyrimidine deoxyribonucleotide synthesis in cells infected by a phage carrying an am mutation of gene 43 (T4 DNA polymerase). Exposure of phage T4xto methyl methanesulfonate, a reagent known to alkylate double-

Previous studies from this laboratory (Flanegan, J. B. and Greenberg, G. R. (1977) J. Biol. Chem. 252, 3019-3027) have shown that bacteriophage T4-infected cultures of Escherichia coli synthesize thymine and 5-hydroxymethylcytosine deoxyribonucleotides via the de nova reductive pathway with rates in the proportion of 2:1, respectively, i.e. exactly the ratio of thymine to 5-hydroxymethylcytosine in T4 DNA. The ratio of synthesis of these deoxyribonucleotides is maintained at 2:l even after infection by Dna-mutants until their concentrations reach about 100 times the initial levels. To examine the possible role of DNA in this regulation, several types of agents altering the structure of DNA were tested in vivo for their abilities to affect the de novo synthesis of dTMP and HmdCMP.
Mitomycin C, an agent causing interstrand cross-linking between guanine residues, was administered to cultures infected by a phage T4x-strain after early protein synthesis was completed. This agent caused a prompt inhibition of HmdCMP synthesis with little effect on dTMP synthesis. Little or no inhibition was observed in phage T4x+-infected cells. Since the product of gene x repairs DNA damage caused by mitomycin C, these findings provide strong argument that the action of this agent is through its reaction with DNA and not by direct inactivation of enzymes in the synthetic pathway to deoxyribonucleotides.
Addition of the intercalating agent proflavine at 10 min after infection, or exposure of T4x-phage to ultraviolet irradiation prior to infection, caused a decrease in both dTMP and HmdCMP synthesis. Proflavine also was shown to inhibit pyrimidine deoxyribonucleotide synthesis in cells infected by a phage carrying an am mutation of gene 43 (T4 DNA polymerase stranded DNA primarily at the guanine residues, inhibited dTMP synthesis slightly but HmdCMP synthesis somewhat greater. With ail the agents employed the effect on deoxyribonucleotide synthesis was not the result of blockage of DNA replication, or inhibition of early enzyme synthesis or by their direct inactivation.
We propose that T4 template DNA has an intimate role in the structure and function of the complex of enzymes synthesizing deoxyribonucleotides to be channeled into DNA replication and that the agents employed distort the DNA and modify the activities of the complex.
Bacteriophage T4-infected cells exhibit a remarkable control over the synthesis of pyrimidine deoxyribonucleotides in uiuo (l-8). In turn, the rate of deoxyribonucleotide synthesis is the limiting factor in DNA replication (5). The ratio of Thy/ HmCytl deoxyribonucleotides formed exactly mirrors the ratio of these nucleotides in T4 DNA, that is 2:l (1). This precise regulation is maintained even on infection by Dna-mutants until the accumulating deoxyribonucleotides reach exceedingly high levels. Therefore during T4 DNA replication in vivo, the regulation of deoxyribonucleotide biosynthesis does not appear to be primarily mediated by feedback control. Because of our earlier studies suggesting a direct role of the deoxyribonucleotide-synthesizing enzymes in DNA replication (6-81, comparable experiments from other laboratories (g-11), and new evidence of a channeled synthesis of deoxyribonucleotides (121, we have proposed that the regulation is intrinsic in the structure and activity of a complex of enzymes (1).
Since the complex of enzymes carrying out polymerization reactions (13) and the complex-forming deoxyribonucleotides appear to be interacting systems (81, the DNA template itself might be an integral part of both. Because deoxyribonucleotide synthesis so accurately reflects the ratio of nucleotides found in the DNA template, the question may be posed whether or not the DNA itself has a role in this regulation. If so, altering its structure might affect not only DNA synthesis but the in uiuo activities of the enzymes forming deoxyribonucleotides. To test such a possibility, we have used various agents known to alter DNA, using the in uivo assay described 6032

Structural
Role of T4 DNA in Deoxyribonucleotide Synthesis previously (1,2,5). The present study shows that such agents do, indeed, have a profound effect on the regulation of deoxyribonucleotide biosynthesis, apparently without affecting the in vitro activities or the induction of the phage-induced early enzymes.

EXPERIMENTAL PROCEDURES
Materials-Strains of Escherichia coli and mutants of bacteriophage T4 used in this study were those described earlier (1,41 Warner and Lewis (16). Tritium Release Assay, ThylHmCyt Ratios, and DNA synthesis-The release of tritium from 5-labeled pyrimidine precursors, the ratio of Thy/HmCyt deoxyribonucleotides synthesized via the de nova pathway, and DNA synthesis were measured as previously described (1,5). The release of tritium from administered [5-3Hluridine occurs at the levels of dTMP synthetase and of dCMP hydroxymethylase ( Fig. 1) and measures all of the dTMP and HmdCMP derivatives formed via the reductive de novo pathway including those derivatives incorporated into the DNA and also excreted into the medium (1). Further definition of this measurement is presented in earlier papers (1,5,8 Fig. 2 show that proflavine added 10 min after infection by phage T4D inhibits not only DNA synthesis but also tritium release activity. At all concentrations, DNA synthesis was more sensitive than tritium release activity. At 20 pglml, the acridine dye gave almost total inhibition of both activities. The inhibition of tritium release activity resulting from the addition of proflavine at 10 min after infection cannot result from an effect on protein synthesis. Thus, chloramphenicol added at 10 min to inhibit protein synthesis had little effect on the rate of tritium release or of DNA synthesis (5,8,17). The inhibition of tritium release also is not caused by a block of DNA synthesis per se since DNA-negative mutants show no inhibitory effect on tritium release activity (1,5). Furthermore, the results in Fig. 3 show that proflavine still inhibits tritium release in a culture infected by an amber mutant of gene 43, the structural gene for T4 DNA polymerase. That is, DNA synthesis is not related to the inhibitory effect of proflavine on the synthesis of dTMP and HmdCMP.
It is conceivable that the inhibition of tritium release activity by proflavine could result from a direct interaction with the enzymes in the pyrimidine deoxyribonucleotide biosynthetic pathways. At 20 pg/ml, the agent had no effect on the activity of dCMP hydroxymethylase, thymidylate synthetase, or dihy- Escherichia coli Thywas infected with phage T4D at 30" for 12 min and plasmolyzed with 2 M sucrose (6, 12). DNA synthesis in the plasmolyzed cells was assayed using [6-3Hluridine as previously described (6, 12). Tritium release activity was assayed using the same reaction mixture but with [S"Hluridine (5). In the preincubation experiments, the plasmolyzed cells were mixed with actinomycin D and the other components of the reaction mixture except for the labeled compounds. After 10 min the labeled compounds were added and the system assayed as in the nonpreincubated preparation. The concentration of mitomycin C was 3 pg/ml. The procedure was as described in the legend to Fig. 2  Except where indicated the procedure was the same as given in Fig. 2. The temperature was 37". The rates of DNA synthesis and tritium release between 30 and 40 min are given as percentages of the control rates. Note that higher concentrations of mitomycin C were required in this experiment than in the experiment shown in Fig. 6, apparently because of a lower potency of the particular batch. Other samples of mitomycin C (all from Sigma with different lot numbers) produced the same level of inhibition at about one-seventh the concentration.  9 (left). Tritium release activity and DNA synthesis after infection with ultraviolet( UV)-irradiated phage T4c. The phage were irradiated to the indicated levels prior to infection. Escherichia coli Thy-cultures at 37" were infected and diluted into medium with either (A) 0.09 mM 15-3H]uridine (3.8 x IO3 cpm/nmol) or (B) 0.09 rnM 16JHluridine (4.1 x lo3 cpmlnmol). The cultures infected with irradiated phage were assayed in the dark. FIG. 10 (center). Effect of ultraviolet irradiation of phage T42-on release, DNA synthesis, and the activities of dTMP synthetase and dCMP hydroxymethylase. The measurements of tritium release and DNA synthesis were made at 30 min after infection and the enzymes were measured at 12 min. If the inactivation curves of Fig. 10 are plotted on a semilogarithmic scale, the slopes for tritium release, DNA synthesis, and viable phage are about -1, -2.2, and -12.6, respectively, indicating the relative sensitivities of these activities. From the phage inactivation curve, an average value of 0.019 hit/erg/mm? was obtained. Fig. 10 also confirms that at the levels of ultraviolet irradiation which greatly inhibit DNA synthesis, the levels of dTMP synthetase and dCMP hydroxymethylase in extracts were unaffected (20,21).
In addition to the decrease in tritium release activity, there was an increase in the ratio of Thy/HmCyt deoxyribonucleotides synthesized in cultures infected with ultraviolet-irradiated phage. The synthesis of dTMP and HmdCMP derivatives in cultures infected with irradiated (300 ergs/mm') and unirradiated phage are presented in Fig. 11. It is clear that the rates of synthesis of both were inhibited. Between 25 and 45 min the rate of dTMP synthesis was reduced by about 33% and the rate of HmdCMP synthesis was reduced by about 60%, resulting in a ratio of dTMP/HmdCMP synthesis of about 3.6 in contrast to the 2:l ratio exhibited by untreated phage. viability, DNA synthesis, tritium release activity, and the induction of dTMP synthetase and dCMP hydroxymethylase. All experimental conditions are described in the legend to Fig. 9. FIG. 11 (right). Synthesis of dTMP and HmdCMP derivatives after infection with ultraviolet irradiated T4x-phage. The phage were exposed to a dose of 300 ergs/mm2. Thy/HmCyt ratios and tritium release data obtained as described in Table I were used to calculate the amounts of deoxyribonucleotide synthesized. mechanisms inhibit not only phage T4 DNA replication but also the de novo synthesis of the pyrimidine deoxyribonucleotides. Two types of effects have been observed. In the first (proflavine treatment or ultraviolet irradiation) both dTMP and HmdCMP synthesis were inhibited. In the second (mitomycin C!) HmdCMP synthesis was immediately and specifically inhibited.
Since each of the agents tested inhibited DNA synthesis more effectively than pyrimidine deoxyribonucleotide synthesis, the question might be raised as to whether or not the resulting accumulation of the pyrimidine deoxyribonucleoside triphosphates causes a secondary feedback inhibition of the de novo pathway. However, as shown earlier, blockage of DNA replication by Dna-mutants has no effect on the combined rates of synthesis of dTMP and HmdCMP (1, 5; see the introduction). Finally, the inhibition of pyrimidine deoxyribonucleotide synthesis by proflavine occurred even with cells infected by a Dna-mutant (Fig. 3).
A number of findings provide evidence that these agents act by altering the input T4 DNA and not by a direct action on the enzymes forming the deoxyribonucleotides.
(a) None of the agents, as employed, decreased the induction or activities of dCMP hydroxymethylase and dTMP synthetase. (b) Addition of mitomycin C, proflavine, ethidium bromide, or actinomycin D after all of the proteins required for DNA synthesis had been formed, still inhibited tritium release from administered 15-"Hluridine.
(c) That mitomycin C brings about its inhibitory action on tritium release specifically by interaction with DNA was shown by the susceptibility of T4x-but not T4x+ phage to this agent (Figs. 6 and 81, inasmuch as mitomycin C damage of DNA is repaired by the action of the product of gene x (21, 22). Cd) The wide range of the DNA agents employed and their application by treatment of the infected cells or by modification of the phage itself argues against an interaction on systems other than DNA.
Of the various agents tested, mitomycin C appeared to cause the most specific effect on deoxyribonucleotide synthesis. Mitomycin C causes an immediate decrease and an eventual cessation in the synthesis of HmdCMP i n cells infected by T4z- (Fig. 81. The synthesis of dTMP derivatives is not decreased nor elevated. If the block were at dCMP hydroxymethylase, the accumulating dCMP would be diverted to dUMP by the action of dCMP deaminase and dTMP would be increased by 6036

Structural
Role of T4 DNA in Deoxyribonucleotide Synthesis about 33%. This analysis is based on the conclusion made in a yribonucleotide-synthesizing enzymes in L' ~L'O (5,8). Structural previous paper that on infection by T4 Dna-mutants product changes in the DNA template could conceivably affect coninhibition of dCMP hydroxymethylase occurs with time and formation of a bound polymerase which would in turn affect the resulting dCMP is siphoned into dUMP synthesis by the the activity of the other enzymes in the complex or might predeaminase so that dTMP synthesis increases by 33% (11.' vent normal complex formation. Therefore, we suggest that mitomycin C through its action on The precise agreement between the ratio of dTMP/ DNA causes a block prior to the formation of dCMP. The most HmdCMP biosynthesis and the 2:l Thy/HmCyt ratio in T4 reasonable site in the complex of enzymes synthesizing deoxy-DNA may suggest that the DNA sequence controls the procribonucleotide is at T4 ribonucleoside diphosphate reductase ess. However, this idea would seem to be untenable since since this enzyme is involved in the synthesis of all four of the regulation of the ratio still occurs in the absence of DNA deoxyribonucleotides.
synthesis (1) and presumably in the absence of movement of Treatment of infected cultures with uroflavine or exnosure the DNA template in relation to the replication complex. of phage before infection to ultraviolet irradiation produces Conceivably DNA could have an allosteric action on the ensimilar effects on the synthesis of thymine and HmCyt deoxy-zyme complex, the synthesis of dTMP and HmdCMP (and the ribonucleotides. That is, both agents show a prompt inhibition purine derivatives; see Ref. 1) being biased by the base seof tritium release from administered [ 5-"Hluridine.
Both treat-quences at the growing points, and an average value would be ments ultimately bring about a greater inhibition of HmdCMP seen in a mixed population of infected cells. synthesis than of dTMP synthesis. Ultraviolet irradiation An alternative role for DNA involving the integral memcauses the Thy/HmCyt ratio to rise to about 3 by 25 min (Fig. brane proteins formed by genes 39 and 52 (31, 32) also may be 111, and proflavine produces an increase by 35 min. While the considered. In mutants (DNA delay) lacking these proteins T4 ratios increase in these cases, the overall rates of dTMP syn-DNA synthesis is greatly decreased (33, 34). Huang and Buthesis actually decrease as compared to an untreated, infected chanan have shown that these proteins are bound by DNAculture. Nevertheless, the increased ratio with proflavine ap-cellulose columns (35), and the suggestion has been made (34, pears to mimic the process occurring on infection by Dna-351 that these products may link the replicating complex with mutants as outlined in the last paragraph. Again, these the membrane (36) reported that the dna gene 41 product catalyzes a GTP -+ GDP Phage DNA might interact with the deoxyribonucleotide-+ P, reaction which is dependent on single-stranded DNA (13). synthesizing enzymes in several ways. A direct interaction The present study provides evidence that DNA plays an intebetween DNA and the enzymes themselves may occur. Since gral role in a system which on the surface would not appear to DNA agents inhibit the synthesis of dTMP and HmdCMP require SO complex a molecule to maintain its structure or even with Dna-mutants, i.e. in the absence of replication, and regulation. since direct modification of the phage also causes inhibition, it The precise regulation of deoxyribonucleotide biosynthesis is clear that this interaction is with the input template. Evi-may be an important factor in insuring the fidelity of DNA dence has been presented earlier that the enzymes synthesiz-replication (1). If so, then the biochemical basis for the mutaing deoxyribonucleotides for DNA replication function in vivo genic effect of certain agents which modify the structure of the as components of a multi-enzyme complex (1,5,8), the prod-DNA template should be explained in part by then effect on ucts being channeled directly into the replication apparatus deoxyribonucleotide biosynthesis. (12). The present study suggesting that the DNA template itself is a necessary component of this complex is in keeping with these earlier proposals. Alternatively, the DNA template Acknowledgment-We are very grateful to Therese Ruetmay directly interact only with T4 DNA polymerase. Phage T4 tinger for carrying out some of the later experiments.
DNA polymerase is required for maximal activity of the deox-Note Added in Proof-Goodman and co-workers (42, 43) ' On infection by a cd mutant (dCMP deaminase-negative) the ratio of ThylHmCyt synthesis is lowered to about 0.6, indicating that have described the inhibitory effects of DNA intercalating _ _ approximately 40% of the dTMP normally derives from dCMPy Pro-agents on the activities of T4 DNA polymerases. With wild flavine treatment of cultures infected by this mutant decreased the type, mutator, and antimutator polymerases, these agents rate of tritium release but did not alter the ratio from 0.6. Unpub-had little effect on the misincorporation frequencies of 2-ami-. . . . 1. . . usned expcrnnents, this laboratory. nopurine deoxyribonucleotide, the analogue of dAMP, and on