ABSTRACT
We previously reported that the presence of dideoxythymidine (ddT) in the growth medium selectively inhibits the ability of bacteriophage T7 to infect Escherichia coli by inhibiting phage DNA synthese (N. Q. Tran, L. F. Rezende, U. Qimron, C. C. Richardson, and S. Tabor, Proc. Natl. Acad. Sci. U. S. A. 105:9373–9378, 2008, doi:10.1073/pnas.0804164105). In the presence of T7 gene 1.7 protein, ddT is taken up into the E. coli cell and converted to ddTTP. ddTTP is incorporated into DNA as ddTMP by the T7 DNA polymerase, resulting in chain termination. We have identified the pathway by which exogenous ddT is converted to ddTTP. The pathway consists of ddT transport by host nucleoside permeases and phosphorylation to ddTMP by the host thymidine kinase. T7 gene 1.7 protein phosphorylates ddTMP and ddTDP, resulting in ddTTP. A 74-residue peptide of the gene 1.7 protein confers ddT sensitivity to the same extent as the 196-residue wild-type gene 1.7 protein. We also show that cleavage of thymidine to thymine and deoxyribose-1-phosphate by the host thymidine phosphorylase greatly increases the sensitivity of phage T7 to ddT. Finally, a mutation in T7 DNA polymerase that leads to discrimination against the incorporation of ddTMP eliminates ddT sensitivity.
INTRODUCTION
When bacteriophage T7 infects an Escherichia coli cell, there is a rapid increase in DNA synthesis, resulting in the production of over 100 T7 genomes in a 10-min period (1). The enzymes and mechanisms by which the T7 DNA is replicated have been studied in great detail. The T7 replisome consists of four proteins: T7 DNA polymerase, T7 helicase/primase, T7 single-stranded DNA (ssDNA) binding protein, and E. coli thioredoxin (2). Much less well understood are the enzymes and mechanisms responsible for the production of the immediate precursors of DNA synthesis, the deoxynucleoside 5′-triphosphates (dNTPs). T7 derives most of the nucleotides found in its DNA from the breakdown of host DNA (3, 4). The host DNA is degraded to deoxynucleoside 5′-monophosphates (dNMPs) by the joint action of the gene 3 endonuclease and gene 6 exonuclease (5, 6). How these dNMPs are eventually converted to dNTPs is not well understood. The genome of E. coli encodes four different dNMP kinases, each specific for one dNMP (7, 8). At least one of these kinases, CMP kinase (CMK), is essential for T7 growth (9). It has been assumed generally that either the nucleoside diphosphate kinase (NDK) or the adenylate kinase (AMK) of the host converts the dNDPs to dNTPs (10), but the question remains as to whether the activity of these kinases is sufficient to meet the demand of T7 DNA replication.
In recent years, we have reported a serendipitous finding that has led to new insight into nucleotide metabolism in T7-infected cells (11–13). Phage T7 growth and T7 DNA synthesis are inhibited by dideoxythymidine (ddT) at concentrations that are not toxic to E. coli. The inhibition of DNA synthesis suggests that the ddT is converted to ddTTP, and then the chain-terminating ddTMP is incorporated into T7 DNA by T7 DNA polymerase. T7 DNA polymerase incorporates ddNMPs with essentially the same efficiency as it does dNMPs (14). Therefore, we originally sought to screen for mutations in gene 5, the structural gene for T7 DNA polymerase (15), in order to identify those that would cause T7 DNA polymerase to discriminate against the incorporation of ddNMPs. Surprisingly, when we isolated phage T7 that could grow in the presence of ddT, nearly all of the mutations were in gene 1.7, a nonessential gene of unknown function. Extensive screening of mutants did identify a T7 phage that encodes an altered DNA polymerase in which tyrosine 526 has been replaced by phenylalanine. This mutation bypasses the function of gene 1.7, and T7 phage with this altered DNA polymerase are resistant to ddT (12). This single alteration in T7 DNA polymerase leads to discrimination against the incorporation of ddTMP by 8,000-fold compared to that with dTMP (14). These results strongly suggest that defects in gene 1.7 lead to a defect in the conversion of ddT to ddTTP. Furthermore, overproduction of T7 gene 1.7 protein (gp1.7) from a plasmid renders E. coli cells sensitive to ddT, indicating that no other T7 proteins are required for conferring sensitivity to ddT. The inhibition of T7 phage and that of E. coli overproducing gp1.7 requires the E. coli thymidine kinase (12), suggesting that gp1.7 exerts its role after the formation of ddTMP.
We have purified T7 gp1.7 and shown that it is indeed a nucleoside monophosphate kinase (11, 13). It has a number of remarkable properties that distinguish it from all other known nucleotide kinases. It phosphorylates dGMP and dTMP to dGDP and dTDP, respectively, using GTP, dGTP, or dTTP as the phosphate donor. It phosphorylates ddTMP with an efficiency equal to that dTMP, in contrast to the host thymidylate kinase (TMK), which discriminates over 500-fold against phosphorylation of ddTMP (11). This observation explains the phenotype of sensitivity to ddT resulting from the presence of gp1.7. T7 gp1.7 shares no sequence homology with any known protein, and there are no identifiable nucleotide binding motifs found in its protein sequence. A most unusual feature is its full activity as a kinase in the absence of any metal ion (11, 13).
The identification of T7 gp1.7 as a nucleotide kinase suggests that the E. coli nucleotide kinases are not sufficient to provide an adequate supply of dNTPs for the synthesis of T7 DNA. dTTP in particular is required in large amounts by T7, not only as a substrate for DNA synthesis but also as the energy supply for T7 DNA helicase where dTTP is hydrolyzed to dTDP (16). While the role of gp1.7 as a nucleoside monophosphate kinase has been shown clearly, there are other steps in the pathway from exogenous nucleosides in the media to the synthesis of dTTP precursors that are less understood. For example, in E. coli, thymidine is transported into the cell by the nup gene products that form a nonspecific nucleoside permease (17). Upon entry into the cell, thymidine is either degraded into thymine and deoxyribose-1-phosphate by thymidine phosphorylase (18, 19) or is phosphorylated to dTMP by thymidine kinase (Fig. 1). The sequestering of thymidine inside the cell requires its phosphorylation to dTMP, since only the phosphorylated compound is unable to diffuse back out of the cell. Once thymidine is phosphorylated to dTMP, it can enter into the thymidine salvage pathway to be converted to dTTP. In fact, the thymidine salvage pathway is the only salvage pathway for deoxyribonucleosides in E. coli (20). Since phage T7 is no more sensitive to ddT in hosts that lack the tdk gene (12), the conversion of ddT to ddTMP must be catalyzed by the E. coli thymidine kinase, although it is not known whether any other E. coli genes are required as well.
Thymidine salvage pathway in E. coli. deoA, thymidine phosphorylase (29); tdk, thymidine kinase (41); yjjG, dUMP phosphatase (40); tmk, thymidylate kinase (42); ndk, nucleoside diphosphate kinase (34); amk, adenylate kinase (10); thyA, thymidylate synthase (43); dT, thymidine; dU, deoxyuridine; dR-1-P, deoxyribose-1-phosphate; broken arrow, synthesis de novo.
The pathway by which ddTDP is phosphorylated to ddTTP in T7 phage-infected cells is not clear. It has been assumed generally that either the host NDK or the adenylate kinase of the host converts the dNDPs to dNTPs (10), but the question remains whether the activity of these kinases also converts ddTDP to ddTTP.
In this study, we use a genetic analysis to address these questions concerning the metabolism of thymidine and dideoxythymidine in E. coli infected with phage T7. A thorough understanding of this pathway will provide important details on how a bacteriophage, particularly phage T7, is selectively killed by ddT, as well as insight into the synthesis of the essential precursors used for both unwinding and synthesis of the DNA of replicating phage.
MATERIALS AND METHODS
Bacterial strain E. coli HMS89 (xth-1 thi argE mtl xyl rpsL ara his galK lacY proA leu thr tsx glnX44) was used for all experiments in which dideoxynucleosides were added to the media. E. coli HMS89(DE3) was obtained by lysogenizing E. coli HMS89 with the lambda phage DE3 using the λDE3 lysogenization kit (Novagen). This strain expresses T7 RNA polymerase upon induction by isopropyl-β-d-thiogalactopyranoside (IPTG). E. coli DH5α (Invitrogen) was used for all subcloning manipulations. E. coli BL21(DE3) (Novagen) was used for overexpression of gene 1.7. E. coli AGI, containing the pCA24N-6His-tdk/tmk/ndk vector from the ASKA library, was used for overexpression of the E. coli thymidine kinase, thymidylate kinase, and nucleoside diphosphate kinase as previously described (21). E. coli thymidine phosphorylase (catalog number T2807) was purchased from Sigma-Aldrich Co. T7 DNA polymerase/thioredoxin was purified and reconstituted as previously described (22). Radioactive materials were purchased from Moravek Biochemicals Inc.
Purification of gp1.7 and nucleotide kinase assays.T7 gene 1.7 was cloned into the vector pET28a for overexpression in E. coli BL21(DE3). The T7 gene 1.7 protein was purified as previously described (11). The standard nucleoside monophosphate kinase assay for gp1.7 measured the conversion of [3H]dTMP or [3H]ddTMP to [3H]dTDP or [3H]ddTDP. Reaction mixtures contained 100 mM Tris-HCl, pH 7.5, 5 mM dithiothreitol (DTT), 100 μM [H3]dTMP or [H3]ddTMP (∼10 cmp/pmol), 2.5 mM dTTP, and 50 nM gp1.7. Reactions were carried out at 37°C and were terminated by heating at 95°C for 3 min. Where indicated, the nucleoside diphosphate kinase activity of gp1.7 was measured by the conversion of [3H]dTDP or [3H]ddTDP to [3H]dTTP or [3H]ddTTP. Reaction products were analyzed by polyethylenimine (PEI) cellulose thin-layer chromatography (TLC) as previously described (11).
Thymidine kinase assay.Thymidine kinase activity was measured by the conversion of dT and ddT to dTMP and ddTMP, respectively. The reaction mixture (200 μl) contained 200 μM [3H]dT or [3H]ddT (10 cpm/pmol), 5 mM ATP, 10 mM MgCl2, 100 mM Tris-HCl, pH 7.8, 20 μg bovine serum albumin (BSA), and 100 nM purified E. coli thymidine kinase. The reaction mixture was incubated at 37°C. At the indicated times, 20-μl aliquots were removed and the reactions were terminated by heating the mixtures at 85°C for 5 min. The samples were then applied to a PEI cellulose plate. The plate was wash thoroughly in distilled water and then air dried. Chromatography was carried out in distilled water. The spot corresponding to dTMP was cut out, and the radioactivity was measured using a liquid scintillation counter.
Phosphorylase assay.E. coli thymidine phosphorylase catalyzes the phosphorolysis of thymidine to thymine and deoxyribose-1-phosphate. Phosphorolysis activity was assayed according to the method of continuous spectrophotometric rate determination, described in detail elsewhere (23). The reaction mixture (1 ml) contained 200 mM potassium phosphate buffer, pH 7.4, 1 mM thymidine (or dideoxythymidine), and 10 U of thymidine phosphorylase. The decrease in A290 in a 1-cm-light-path cuvette was measured at room temperature using a diode array spectrophotometer (Hewlett Packard).
Construction of pDeoA and pTDK expression plasmids.The coding sequence of the deoA gene of E. coli was cloned into the vector pET28a. In the resulting plasmid, pDeoA, the expression of the cloned deoA gene was under the control of a T7 RNA polymerase promoter. pTDK was constructed analogously. Plasmids were transformed into the E. coli HMS89 strain for complementation studies.
Inhibition of phage T7 by ddT and complementation assays.Inhibition of phage T7 by ddT was carried out by mixing 300 μl of an overnight culture of E. coli HMS89 with 3 ml of premelted soft agar (1% tryptone, 0.5% yeast extract, 0.5% NaCl, 0.7% agar) at 45°C in the presence of 1 mM ddT. To the mixture, 50 μl of wild-type (wt) phage T7 (4 × 103/ml) was added. The mixture was overlaid onto petri plates. The plates were incubated in a 37°C incubator for 6 h or left at room temperature (∼25°C) overnight. The plates were then photographed. Plating efficiencies were determined by dividing the number of plaques observed in the presence of ddT by the number observed in its absence.
For mapping of the region of gp1.7 that is essential for ddT sensitivity, gene 1.7 mutants containing various deletions at either the 5′ or 3′ end were cloned into the expression vector pET28a. The resulting plasmids then were transformed into E. coli HMS89 for complementation analysis. In this system, the gene 1.7 deletions will be expressed upon infection of the E. coli cells by phage T7. The cells not infected by phage T7 will grow normally in the presence of ddT. Complementation assays were carried out to determine the ability of the various truncated gp1.7 mutants to confer sensitivity of phage T7Δ1.7 to the presence of ddT in the media. Phage T7Δ1.7 has the entire gene 1.7 deleted and, in contrast to wild-type T7, is resistant to the presence of ddT in the media (1).
Construction of gene knockouts in E. coli HMS89.E. coli HMS89 mutants were constructed in which either deoA, deoB, deoC, deoD, dcd, ndk, nup, tdk, or yjjG was deleted based on the protocol described in the Quick & Easy E. coli gene deletion kit (Gene Bridges GmbH) and as previously described (12). The requirement of the deleted gene for conferring sensitivity of phage T7 to ddT in the media was examined by measuring the plating efficiency and plaque size of wild-type phage T7, infecting each of the mutant E. coli strains in the presence of 1 mM ddT in the media as described above.
Preparation of E. coli crude lysate.E. coli HMS89 cells were grown in 10 ml LB at 37°C to an A600 of 1. Cells were harvested by centrifugation and then suspended in 500 μl of lysis buffer containing 20 mM Tris-HCl, pH 7.5, 10 mM β-mercaptoethanol, and 0.1 mM EDTA. Cells were disrupted using a Branson digital sonifier at a setting of 20% pulse for 10 pulses of 5 s each, followed by centrifugation at 14,000 rpm for 30 min in a microcentrifuge. The supernatant was collected. The concentration of crude lysate was determined by the Bradford method (24).
Measurement of dideoxythymidine uptake.E. coli HMS89 tdk (DE3) and HMS89 tdk (DE3)/pGP1.7 mutant strains were grown at 30°C to an A600 of 0.3 in LB medium. Gene 1.7 was induced by the addition of 1 mM IPTG for 20 min. Nalidixic acid was added to a final concentration of 200 μg/ml, and the cells were incubated for an additional 10 min to inhibit DNA synthesis. At time zero, [3H]ddT was added to a final concentration of 100 μM (50 μCi/ml). At the indicated times, 200 μl was removed, filtered by vacuum through nylon Millipore filters (pore size, 0.45 μm), washed twice with 2 ml of cold LB medium, and dried, and then the 3H label retained on the filters was measured using a liquid scintillation counter.
Inhibition of DNA synthesis by dideoxythymidine in vitro.Kinase reaction mixtures (500 μl) contained 50 mM Tris-HCl, pH 7.5, 50 mM NaCl, 10 mM MgCl2, 5 mM DTT, 0.5 mM ddT, 2.5 mM ATP, and 100 nM E. coli thymidine kinase. Reaction mixtures were preincubated at 37°C for 30 min. DNA polymerase activity was then assayed as previously described (13). Briefly, DNA synthesis was initiated by the addition of the following components to the kinase mixture: 100 μM each dATP, dCTP, dGTP, and [3H]dTTP (∼10 cpm/pmol); 20 nM single-stranded M13 DNA primed with a 24-nucleotide primer; and 5 nM T7 DNA polymerase/thioredoxin. Where indicated, the reaction mixtures also contained 50 nM either T7 gp1.7 or E. coli thymidylate kinase. DNA synthesis was carried out at 37°C. Aliquots were removed at the indicated times and were spotted on DE81 ion exchange filters. Unincorporated nucleotides were removed by washing the filters three times (5 min each) in 300 mM ammonium formate. The 3H label retained on the filters was determined by liquid scintillation counting.
RESULTS
Effect of mutations in nucleoside uptake genes on the ability of dideoxythymidine to inhibit T7 phage growth.The uptake of nucleosides from the media into E. coli cells is believed to involve four nucleoside permeases, the products of the nupA (25), nupC (26), nupG (27), and nupX (28) genes. We tested whether any of these four gene products are solely responsible for the uptake of ddT into the cell by constructing different E. coli strains with a deletion of each of these genes. When T7 phage infected each of these strains, in each case the phage were inhibited to the same extent as in wild-type E. coli (Fig. 2). Furthermore, we constructed strains containing deletions in combinations of two of the nup genes; T7 phage had the same sensitivity to ddT in these strains as well. Unfortunately, we were unsuccessful in our attempt to simultaneously knock out three or four nup genes. The inability to do so could be due to technical difficulties, since at least three marker genes plus one from the plasmid must be inserted for selection. The results suggest that transport of ddT does not depend solely on a single or combination of two nup gene products.
Role of E. coli nup genes in sensitivity of phage T7 to ddT in the media. For each pair of plates shown, T7 phage was plated on a different E. coli strain either in the absence of ddT (left) or the presence of 1 mM ddT (right). From top to bottom: the E. coli HMS89 wt strain, E. coli HMS89 nupA strain, E. coli HMS89 nupC strain, E. coli HMS89 nupG strain, and E. coli HMS89 nupX strain. Plates were incubated at 37°C for 4 h and then photographed.
Effect of gene 1.7 on dideoxythymidine uptake.We previously showed that when T7 gp1.7 was overproduced in E. coli, it dramatically increased the intracellular pools of dideoxythymidine nucleotides (ddTMP, ddTDP, and ddTTP) when the cells were grown in the presence of ddT in the media (11). However, this experiment could not distinguish between an effect of gp1.7 on the transport of ddT into the cell directly or an indirect effect of sequestering the phosphorylated derivatives inside the cell as a result of their phosphorylated derivatives being unable to be transported out of the cell. In order to address this issue more directly, we examined the effect of gp1.7 on ddT transport using E. coli tdk. Since this strain lacks thymidine kinase activity, which is required for the initial phosphorylation of ddT, this experiment examines the role of T7 gp1.7 on ddT transport under conditions where the ddT will not be chemically altered when it enters the cell. Interestingly, gp1.7 significantly increases the intracellular concentration of ddT in this strain, even though the nucleoside is not a substrate for gp1.7 (Fig. 3). The mechanism by which gp1.7 is facilitating this sequestration is not known.
Effect of gene 1.7 protein on dideoxythymidine uptake. The E. coli HMS89 (DE3) tdk (▲) or HMS89(DE3) tdk/pGP1.7 (△) strain was grown at 37°C to an A600 of 0.4 in LB medium. Cells were induced with 1 mM IPTG for 30 min, followed by addition of 250 μg/ml nalidixic acid to inhibit DNA synthesis. [3H]ddT then was added to a final concentration of 100 μM (50 μCi/ml). Aliquots of 200 μl were removed at the indicated times, and the [3H]ddT taken up by the cells was determined as described in Materials and Methods.
E. coli gene tdk is essential for conversion of ddT to ddTMP.The tdk gene of E. coli encodes thymidine kinase (TDK; EC 2.7.1.21). Thymidine kinase plays an essential role in the thymidine salvage pathway by converting thymidine to thymidylate (dTMP) (Fig. 1). We have shown that phage T7 was able to grow on the E. coli HMS89 tdk strain, lacking the entire gene tdk, in the presence of ddT up to 1.5 mM, with essentially no effect on the plating efficiency; the plaque size was comparable to those observed in the absence of ddT (12). In this study, we purified thymidine kinase of E. coli and directly compared its ability to phosphorylate dT and ddT. E. coli thymidine kinase phosphorylates ddT to yield ddTMP, albeit at a rate ∼20% that of dT to dTMP. Taken together, these results show that E. coli thymidine kinase is the enzyme responsible for the conversion of ddT to ddTMP when T7 phage infects E. coli in the presence of ddT.
Effect of host thymidine phosphorylase activity on ddT sensitivity of phage T7.Gene deoA of E. coli encodes a thymidine phosphorylase (TP; EC 2.4.2.4) that catalyzes the reversible reaction: thymidine + phosphate ⇆ thymine + deoxyribose-1-phosphate (dR-1-P) (Fig. 1) (29, 30). Once thymidine is degraded to thymine and deoxyribose-1-phosphate, the thymidine no longer is incorporated into DNA, since the conversion of thymine back to thymidine is inefficient except in thymine-requiring mutant strains. Therefore, an E. coli mutant that lacks thymidine phosphorylase effectively incorporates thymidine into DNA (31) and provides a useful tool for the measurement of DNA synthesis using labeled thymidine (31–33). We constructed an E. coli deoA strain containing a deletion of the entire deoA gene. We expected that T7 phage would be more sensitive to ddT in this strain, since the ddT would not be broken down to thymine and dideoxyribose-1-phosphate. Surprisingly, we found that phage T7 was considerably less sensitive to ddT in the E. coli deoA strain than in wild-type E. coli (Fig. 4A). Although phage T7 produced smaller plaques when phage T7 infects the E. coli deoA strain in the presence of ddT than those produced in the absence of ddT, the efficiency of plating is essentially identical under both conditions (Fig. 4A, compare the two upper plates). This effect is due to the absence of the deoA gene, since, when phage T7 infects the E. coli deoA/pDeoA strain, which contains a plasmid that expresses the deoA gene upon infection of phage T7, it has the same high sensitivity to ddT as when the phage infected wild-type E. coli (Fig. 4A, lower right).
Requirement of host thymidine phosphorylase (TP) for maximum inhibition of phage T7 by ddT. (A) Effect of E. coli thymidine phosphorylase on ddT sensitivity of phage T7. (Top) E. coli deoA mutant strain infected with phage T7 in the absence (left) and in the presence (right) of 1.5 mM ddT. (Bottom) E. coli deoA strain carrying the pDeoA plasmid infected with phage T7 in the absence of ddT (left) and in the presence of 1.5 mM ddT. pDeoA induces the expression of the deoA gene upon infection of phage T7. The number of phage-infected HMS89 deoA and HMS89 deoA/pDeoA strains in the presence of ddT are 10- and 100-fold higher, respectively, than in the absence of ddT. Plates were left at room temperature overnight and then photographed. Black circles indicate plaques of phage T7 formed on the lawn of E. coli cells. (B) Comparison of the ability of E. coli thymidine phosphorylase to use thymidine and dideoxythymidine. Reaction mixtures contained either thymidine (left) or dideoxythymidine (right) and 10 U of enzyme thymidine phosphorylase in 100 mM KH2PO4, pH 7.4. The decrease in A290 in a 1-cm light path at room temperature was monitored using a diode array spectrophotometer (Hewlett Packard), as described in Materials and Methods and elsewhere (23). Open circles indicate control reactions in the absence of enzyme, and filled circles indicate reaction mixtures containing 10 U of thymidine phosphorylase.
These results suggest that thymidine phosphorylase is required in order to obtain the maximum inhibition of phage T7 by ddT. This finding is contrary to what would be expected if thymidine phosphorylase cleaved ddT to thymine and ddR-1-P, comparable to the reaction it carries out with dT.
To address this question, we compared the activities of purified E. coli thymidine phosphorylase on dT and ddT substrates to determine if a difference in activity could explain the effect of the presence of thymidine phosphorylase on the ability of ddT to inhibit T7 phage. We measured thymidine phosphorylase activity using a spectrophotometric assay based on the observation that dT and ddT have higher molar extinction coefficients than does thymine (see Materials and Methods). Cleavage of dT or ddT to produce thymine and dR-1-P or ddR-1-P will result in a decrease in the A290. When dT is used as the substrate for thymidine phosphorylase, there is a linear decrease of ∼10% in A290 over 5 min (Fig. 4B, left). In contrast, when ddT is the substrate for thymidine phosphorylase, there is no detectable change in the A290 (Fig. 4B, right). We conclude that ddT is not a substrate of E. coli thymidine phosphorylase. This result explains why inhibition of phage T7 growth by ddT is significantly increased in the presence of thymidine phosphorylase. Thymidine phosphorylase will selectively degrade dT over ddT, increasing the ratio of ddT to dT in the cell. A higher ratio of ddT to dT as a substrate for thymidine kinase will result in a larger amount of ddTMP produced relative to dTMP.
T7 gp1.7, and not host enzymes, is responsible for the conversion of ddTDP to ddTTP.In E. coli, nucleoside diphosphate kinase (NDK; EC 2.7.4.6) encoded by the ndk gene is a ubiquitous enzyme responsible for the conversion of nucleoside diphosphates to nucleoside triphosphates (34). The function of nucleoside diphosphate kinase was believed to be the maintenance of balanced nucleoside triphosphate pools in the cells (35, 36). However, the deletion of the ndk gene does not affect the viability of E. coli. Although we have previously shown (13) that gp1.7 can catalyze the synthesis of dTTP or ddTTP from dTDP or ddTDP, respectively, we were curious as to the ability of nucleoside diphosphate kinase to convert ddTDP to ddTTP. We engineered an E. coli ndk strain and used this host to examine whether nucleoside diphosphate kinase was required to confer the sensitivity of phage T7 to ddT (Fig. 5). Serving as the controls, top plates of Fig. 5 show the plaques produced by T7 phage on wild-type E. coli in the absence and presence of 1 mM ddT. Surprisingly, we obtained the same results when the E. coli strain is missing the ndk gene (Fig. 5, lower plates). Phage T7 produces plaques of normal size and frequency in the absence of ddT (Fig. 5, compare the plates on the left) but is inhibited to the same extent in the presence of ddT (Fig. 5, compare plates on right). This finding suggests that the conversion of ddTDP to ddTTP is not dependent on nucleoside diphosphate kinase; thus, another enzyme must be responsible for this step.
Host nucleoside diphosphate kinase is not required for gene 1.7-dependent ddT sensitivity of phage T7. (Upper) Wild-type E. coli HMS89 infected with phage T7 in the absence (left) and presence (right) of 1 mM ddT. (Lower) E. coli HMS89 ndk strain infected with phage T7 in the absence (left) and presence (right) of 1 mM ddT. Plates were incubated at 37°C for 5 h then photographed. Black circles indicate plaques of phage T7 formed on the lawn of E. coli cells.
Lu and Inouye (10) have demonstrated that adenylate kinase (AMK; EC 2.7.4.3), an essential enzyme for the generation of ADP from AMP, also has nucleoside diphosphate kinase activity, and that this is the enzyme responsible for catalyzing this essential activity in E. coli deleted for the ndk gene. In order to address the enzyme(s) responsible for phosphorylation of ddTDP to ddTTP, we examined a reaction mixture containing either T7 gp1.7 alone (Fig. 6, solid lines) or in the presence of a crude extract prepared from wild-type E. coli (Fig. 6, broken lines). We monitored the radioactive products produced starting with either [H3]dTMP (Fig. 6A) or [H3]ddTMP (Fig. 6B) as the substrate. Under these conditions, T7 gp1.7 alone efficiently phosphorylates these substrates to [H3]dTDP (Fig. 6A, top solid line) and [H3]ddTDP (Fig. 6B, top solid line). There is also a small (∼10%) amount of [H3]dTTP (Fig. 6A, bottom solid line) or [H3]ddTTP (Fig. 6B, bottom solid line). When the reaction mixtures contain the E. coli extract in addition to T7 gp1.7, the amount of each product formed is dramatically different depending on whether the substrate is [H3]dTMP or [H3]ddTMP. When the substrate is [H3]dTMP, the major product formed is [H3]dTTP (Fig. 6A, top broken line). However, when the substrate is [H3]ddTMP, the major product formed is [H3]ddTDP (Fig. 6B, top broken line). In fact, the addition of the E. coli extract does not alter the amount of [H3]ddTTP formed compared to the small (∼10%) amount formed in the presence of T7 gp1.7 alone (Fig. 6B, compare the bottom broken line to the bottom solid line). This demonstrates that while the host enzymes efficiently phosphorylate dTDP to dTTP, it appears that there are no enzymes in an E. coli lysate that efficiently phosphorylate ddTDP to ddTTP.
Roles of host enzymes in the conversion of ddTDP to ddTTP. (A) Reaction mixtures (100 μl) contained 100 μM [H3]dTMP, 2.5 mM each dTTP and ATP, and 10 mM Mg2+. Reactions were initiated by the addition of 50 ng of gp1.7 (solid line) or 50 ng of gp1.7 plus 2.5 μg of E. coli crude lysate (broken line). (B) An experiment identical to that shown in panel A was performed, except that [H3]dTMP was replaced by [H3]ddTMP. At the indicated times, the formation of [3H]dTDP (▲) and [3H]dTTP (●) (left) or [3H]ddTDP (▲) and [3H]ddTTP (●) (right) was determined by TLC as previously described (11).
How does the gp1.7 phosphorylate both ddTMP and ddTDP? We have previously shown that the thymidine kinase reaction catalyzed by T7 gp1.7 is reversible: ddTMP + dTTP ⇆ dTDP + ddTDP (13). The reaction provides a pathway for the formation of ddTTP via the reverse reaction involving two ddTDP substrates or dTDP as a phosphate donor.
Inhibition of DNA synthesis by ddT in vitro.Thus far, genetic and biochemical data have implicated the host tdk, T7 gene 1.7, and T7 gene 5 as essential genes for conferring sensitivity of phage T7 to ddT in vivo. We purified these enzymes and reconstituted in vitro the inhibition of DNA synthesis by ddT (Fig. 7). DNA synthesis reactions were carried out coupled to the preincubated ddT-kinase reactions that contain E. coli thymidine kinase, ddT, and ATP as described in Materials and Methods. DNA synthesis is initiated by adding T7 DNA polymerase, primed M13 ssDNA, and dNTPs. As shown in Fig. 7, the addition of thymidylate kinase to the T7 DNA polymerase reaction has no effect on the amount of DNA synthesized; under the conditions shown in Fig. 7, DNA synthesis is linear for 90 min. However, when T7 gp1.7 was added to the reaction mixture, DNA synthesis was strongly inhibited after approximately 30 min. These results show that these three enzymes alone are sufficient to catalyze the synthesis of ddTTP from dT, which then is incorporated by T7 DNA polymerase to terminate DNA synthesis.
Reconstitution of the inhibition of DNA synthesis by ddT in vitro. ddT kinase reaction mixtures containing 50 mM Tris-HCl, pH 7.5, 50 mM NaCl, 10 mM MgCl2, 5 mM DTT, 0.5 mM ddT, 2.5 mM ATP, and 100 nM E. coli thymidine kinase were incubated at 37°C for 30 min. The DNA synthesis assay was initiated by the addition of the following components to the kinase mixture: 100 μM each dATP, dCTP, dGTP, and [3H]dTTP; 20 nM M13 ssDNA primed with a 24-nucleotide primer; and 5 nM T7 DNA polymerase (●). Where indicated, the addition also contained 50 nM either T7 gp1.7 (○) or E. coli thymidylate kinase (▲). DNA synthesis was carried out at 37°C. [H3]dTMP incorporated into DNA was measured as previously described (13).
Mapping the region of gp1.7 that is essential for ddT sensitivity.We previously analyzed a series of phage T7 deletion mutants to determine the region responsible for conferring sensitivity of phage T7 to ddT (12). These results narrowed the critical region for this activity to the C-terminal 113 residues of gp1.7. We have more precisely defined the region essential for this activity via a series of additional deletions in gene 1.7 (Fig. 8). Specific deletions were constructed in the plasmid pET28a and then examined for their ability to complement gp1.7 for the ability to confer ddT sensitivity to phage T7Δ1.7 (Fig. 8A; also see Materials and Methods). When 121 amino acid residues were deleted from the N terminus of the T7 gp1.7 (gp1.7Δ121N), the truncated protein was able to confer nearly full sensitivity of the T7 phage to ddT (Fig. 8B). However, when one additional amino acid residue was deleted from the N terminus (gp1.7Δ122N), it was unable to complement gp1.7 and the T7Δ1.7 phage was completely resistant to ddT.
Deletion mapping of the region of gp1.7 essential for conferring ddT sensitivity to T7 phage growth. (A) Schematic of deletion map. Horizontal bars Δ125N and Δ126N, gp1.7 deletion mutants missing the first 125 and 126 residues, respectively, at the N terminus; bars Δ1C and Δ2C, gp1.7 deletion mutants missing the terminal 1 and 2 residues, respectively, at the C terminus. DNA sequences encoding truncated gp1.7 were cloned into expression plasmid pET28a. The resulting plasmids were transformed into E. coli HMS89 for examining the ability of the truncated gp1.7 to confer T7Δ1.7 sensitivity to ddT as described for panel B. The column at the right summarizes the effect of each deletion on the ability of the altered gp1.7 to confer sensitivity of T7Δ1.7 growth to ddT: +, confers inhibition; −, confers no inhibition. (B) Ability of truncated gp1.7 to confer T7Δ1.7 sensitivity to ddT. T7Δ1.7 (lacking the entire gene 1.7) was used to infect E. coli HMS89 containing pGP1.7 mutant plasmids that express the truncated forms of gp1.7 described for panel A. Complementation was performed as described in Materials and Methods.
We carried out a comparable analysis of deletions at the C terminus of gp1.7. Deletion of the terminal residue at the C terminus (gp1.7Δ1C) did not affect the ability of gp1.7 to confer sensitivity of T7 phage to ddT (Fig. 8B). However, deletion of two residues (gp1.7Δ2C) resulted in complete loss of the ability of gp1.7 to confer sensitivity to ddT. These results define the essential region of gp1.7 necessary for this activity to 74 residues at the C terminus of gp1.7 (Fig. 8A). It is remarkable that this truncated gp1.7, consisting of just 38% of the wild-type gp1.7, is able to confer complete sensitivity of T7 phage to ddT.
DISCUSSION
Phage T7 growth is inhibited by ddT when it is added to the growth media at concentrations that do not inhibit its host, E. coli (12). The inhibition is the consequence of ddT being converted to ddTTP, which then is incorporated into T7 DNA as ddTMP to terminate further DNA synthesis (37). ddT must be taken up and phosphorylated to ddTMP prior to being converted to ddTDP by gp1.7. Genetic and biochemical data show that gp1.7 does not play a role in this step. In E. coli, thymidine is phosphorylated to thymidylate (dTMP) by thymidine kinase (38). We previously showed that when T7 infects an E. coli host lacking the tdk gene, it is no longer inhibited by the presence of ddT (12). This observation indicates that T7 phage does not encode a thymidine kinase and that the host thymidine kinase is responsible for the conversion of ddT to ddTMP. In this study, we demonstrate that E. coli thymidine kinase indeed phosphorylates ddT to ddTMP at a rate approximately 20% that of the rate it converts dT to dTMP.
In this study, we also examined the role of different E. coli genes in the uptake of ddT into the cells. Four genes, nupA (25), nupC (26), nupG (27), and nupX (28), have been shown to be responsible for transporting nucleosides into the cells. We show that deletion of any of these four genes has no effect on the inhibition of T7 phage by ddT, suggesting that none of these genes is solely responsible for uptake of ddT. T7 phage also are inhibited normally by ddT in an E. coli mutant lacking two of the nup genes. However, we could not rule out the possibility that each individual nup gene is able to transport ddT into the cells. On the other hand, we have shown here that gp1.7 may play a role in ddT uptake; when gp1.7 is produced in the E. coli tdk mutant strain, there is a significant increase in the uptake of ddT. When E. coli cells lack the tdk gene, nucleosides taken up into the cell are unable to be phosphorylated; thus, they are free to pass back out of the cell. We do not know whether gp1.7 plays a direct role in the transport of ddT or an indirect role by stimulating a host gene product.
When dT is taken up by E. coli, it can either be phosphorylated to dTMP by thymidine kinase or it can be broken down by thymidine phosphorylase in the following reaction: dT + Pi ⇆ T + dR-1-P (30). While this reaction is reversible, in E. coli the breakdown of dT is favored, because thymidine phosphorylase is induced by thymidine and thymine is not readily converted to dTTP for DNA synthesis. Breakdown of thymidine is also favored by the metabolism of the resulting dR-1-P. At the outset of this study, we assumed that inactivation of the host thymidine phosphorylase would render T7 phage more sensitive to ddT, since it would prevent its breakdown to deoxythymine and dR-1-P. To our surprise, this mutation had the opposite effect, rendering phage T7 less sensitive to ddT. This apparent paradoxical result was explained by our biochemical analysis of the substrate requirements for the E. coli thymidine phosphorylase. While this enzyme readily cleaves thymidine to thymine and dR-1-P, it was inactive with ddT. The consequence of this differential activity results in increasing the relative intracellular concentration of ddT versus dT by selectively degrading dT. Thus, phage T7 is most sensitive to ddT when thymidine phosphorylase is present to reduce the concentration of dT.
In E. coli, there are two pathways for the conversion of dNDPs to dNTPs, catalyzed by nucleoside diphosphate kinase (34) and adenylate kinase (AMK) (10). In this study, we show that neither of these pathways is responsible for the conversion of ddTDP to ddTTP in T7-infected cells. Nucleoside diphosphate kinase is a nonessential enzyme. Phage T7 is fully inhibited by ddT when it infects E. coli mutants lacking this enzyme. Furthermore, whereas wild-type E. coli extracts efficiently convert dTDP to dTTP, they are unable to phosphorylate ddTDP to ddTTP. We have previous shown that T7 gp1.7 is a reversible kinase, catalyzing the reaction dTDP or ddTDP + ddTDP ⇆ dTMP or ddTMP + ddTTP (13). When labeled ddTMP is used as the substrate, ∼10% of the ddTDP generated is consistently converted to ddTTP. The amount of ddTTP generated is limited in these reactions, since the phosphate donor dTTP is always present in large excess, favoring the generation of dTDP/ddTDP. In T7-infected E. coli cells, the dTTP concentration is much lower, since it is a substrate for both T7 DNA polymerase and T7 helicase, being hydrolyzed to dTDP to drive unwinding activity. These conditions would favor the production of ddTTP. Consistent with this scenario, we have shown previously that in vivo, gp1.7 dramatically increases the ddTTP pool in E. coli cells growing in the presence of ddT (11). We conclude that ddTDP is not a substrate for either of the two host pathways for the conversion to nucleoside triphosphate, and that in T7-infected cells gp1.7 is the enzyme responsible for both the conversion of ddTMP to ddTDP and its subsequent phosphorylation to ddTTP. It remains to be determined whether this is an important pathway for the conversion of dTDP to dTTP as well or if the host nucleoside diphosphate kinase or AMP kinase pathways dominate with this substrate.
We believe that we have identified all host and phage genes involved in ddT metabolism. We did not see any difference in the sensitivity of phage T7 to ddT when infecting E. coli mutants lacking the deoB, deoC, deoD, yjjG, or dcd gene, mutations that have been shown to increase the ability of E. coli to use exogenous thymidine (19, 39, 40). Moreover, we could reconstitute in vitro the ddT inhibition of DNA synthesis catalyzed by T7 DNA polymerase by the addition of the two nucleotide kinases, E. coli thymidylate kinase and T7 gp1.7.
In addition to its ability to phosphorylate dideoxynucleotides, gp1.7 has a number of other novel properties not found in other known kinases. For example, it does not share any sequence homology with any other kinase, including any of the identifiable motifs found in nucleotide binding domains (12). Furthermore, it lacks any requirement for a divalent cation for catalytic activity (11). We had previously identified the region of gp1.7 essential for conferring ddT sensitivity to lie in the 113 carboxyl-terminal residues of the protein. In this study, we more precisely mapped the residues essential for activity. The ability to confer sensitivity to ddT was retained after deleting 121 residues at the amino terminus and just a single residue from the carboxyl terminus, leaving a minimal fragment of 74 residues that retained kinase activity. The role of the amino-terminal 121 residues is not known. This region has nine cysteine residues arranged in two putative zinc-binding motifs. One possibility is that this region plays a role in physical interactions between gp1.7 and other T7 DNA replication proteins.
ACKNOWLEDGMENTS
This work was funded by U.S. Public Health Service grant GM54397 and Department of Energy grant DE-FG02-96ER 62251.
FOOTNOTES
- Received 1 April 2014.
- Accepted 17 May 2014.
- Accepted manuscript posted online 23 May 2014.
- Address correspondence to Charles C. Richardson, ccr{at}hms.harvard.edu.
REFERENCES
- Copyright © 2014, American Society for Microbiology. All Rights Reserved.
