Cloning, characterization and expression analysis of nucleotide metabolism-related genes of mycobacteriophage L5


  • Editor: Dieter Jahn

Correspondence: Sujoy K. Das Gupta, Department of Microbiology, Bose Institute, P1/12 C.I.T. Scheme VIIM, Kolkata 700054, India. Tel.: +91 33 23559416; fax: +91 33 2553886; e-mail:


The genomes of mycobacteriophages of the L5 family, which includes the lytic phage D29, contain several genes putatively linked to nucleotide-metabolizing functions. Two such genes, 48 and 50, encoding thymidylate synthase and ribonucleotide reductase (RNR), respectively, were overexpressed in Escherichia coli and the recombinant proteins were biochemically characterized. It was established that Gp50 was a class II RNR having properties similar to that of the corresponding enzyme from Lactobacillus leichmanni, whereas Gp48 was a flavin-dependent thymidylate synthase (ThyX) that resembled the Paramecium bursaria chlorella virus-1 ThyX enzyme in its properties. That both these proteins play a role in phage development was evident from the observation that they were detectable soon after the lytic phase of growth commenced. Gp48 and 50 were also found to coimmunoprecipitate, which indicates the possible existence of an L5 thymidylate synthase complex. Thymidylate synthase assays revealed that during the intracellular stage of phage growth, a significant decrease in the host thymidylate synthase (ThyA) activity occurred. It appears that synthesis of the viral enzyme (ThyX) is necessary to compensate for this loss in activity. In general, the results suggest that phage-encoded nucleotide metabolism-related functions play an important role in the lytic propagation of L5 and related mycobacteriophages.


Mycobacteriophages have been considered to be cornerstones of mycobacterial research (Hatfull & Jacobs, 1994). Among the mycobacteriophages that have been well studied are the closely related members of the L5 family, which includes the lysogenic phages L1 (Doke, 1960; Lee et al., 1991), L5 (Snapper et al., 1988; Hatfull & Sarkis, 1993) and BxB1 (Jain & Hatfull, 2000) and the lytic phage D29 (Ford et al., 1998). Although the L5 family of phages is considered to be lysogenic or potentially lysogenic (D29) (Ribeiro et al., 1997), they have features that strongly resemble lytic phages (Hatfull & Sarkis, 1993). One such feature is the ability of these phages to shut down host protein synthesis immediately following infection.

A large number of mycobacteriophage genome sequences are currently available in the databases (Pedulla et al., 2003). Comparative analysis of these sequences revealed that at least half of the mycobacteriophage genes have no database matches (NDMs), indicating thereby the possibility of the existence of novel genes, whose functions are unknown. Although several mycobacteriophage genome sequences are available, only a limited number of genes have been characterized, most of which belong to the L5 family. These are: gene 71, the phage repressor (Donnelly-Wu et al., 1993); gene 33, integrase (Lee & Hatfull, 1993); and gene 36, excisionase (Lewis & Hatfull, 2000). In addition, several other mycobacteriophage genes have been characterized, such as those encoding a tape measure protein from TM4 (Piuri & Hatfull, 2006), a polynucleotide kinase from Omega and Cjw1 (Zhu et al., 2004) and a lysis protein from Ms6 (Garcia et al., 2002).

In silico analysis (Hatfull & Sarkis, 1993; Ford et al., 1998) of L5 and D29 genome sequences revealed the presence of several genes that putatively code for enzymes involved in DNA synthesis. One such gene (44) encodes a viral DNA polymerase. Other DNA synthesis-related genes are 58, 59 and 65, which putatively encode primase, exonuclease and helicase, respectively. Apart from genes encoding enzymes related to DNA replication, there are several genes that appear to code for nucleotide metabolism-related functions. Two such genes are 48 (thymidylate synthase) and 50 [ribonucleotide reductase (RNR)]. In the case of D29, but apparently not in the case of L5, there may be a gene for dCMP deaminase also. All of these genes are clustered in the central region of the phage genome, indicating that their products have some related functions. If the entire scenario is considered in its totality, it seems that, as in case of the lytic phages such as T4, L5/D29 phage-encoded DNA polymerization and deoxy-ribonucleotide synthesis-related enzymes act in concert to meet the replicative demands of the viral DNA, particularly during the early stage of infection.

In nature, RNR (Poole et al., 2002) accounts for the acquisition of three (dCTP, dGTP and dATP) of the four deoxyribonucleotides needed for DNA biosynthesis. The fourth nucleotide, dTTP, is produced by two well-known processes. Exogenous thymidine can be directly salvaged by thymidine kinase or dTMP can be synthesized de novo from dUMP, a reaction catalyzed by thymidylate synthase. Thymidylate synthases may be of two types (Myllykallio et al., 2002) – the classical thymidylate synthase designated as ThyA and the alternative class of thymidylate synthases known as ThyX. Both these enzymes catalyze the conversion of dUMP to dTMP, but the reductive mechanisms are different. In ThyA-catalyzed reactions, CH2H4folate serves both as a reductant and as a CH2– donor whereas in a ThyX-catalyzed reaction CH2H4folate acts only as a CH2– donor, reducing equivalents being donated by NADPH through the intermediate reduction of the enzyme-bound FAD. The protein encoded by L5 gene 48 contains the highly conserved ThyX motif RHRX7S (Myllykallio et al., 2002). ThyX proteins in general are found in many pathogenic bacteria and it has been proposed that this protein is an attractive target for developing drugs against such bacteria (Graziani et al., 2004). Interestingly, mycobacteria possess both the versions, ThyX as well as ThyA, but the significance of this is not clearly understood (Myllykallio et al., 2002; Sampathkumar et al., 2005).

Mycobacteriophage L5 gene 50 putatively codes for a class II RNR (RNR II). RNRs of the class II family typically use adenosylcobalamin as the cofactor for the generation of the free radicals (Poole et al., 2002). It is generally considered that Class II enzymes function under micro-aerophillic conditions, whereas class I and III enzymes function under aerobic and anaerobic conditions, respectively.

Enzymes catalyzing sequential metabolic reactions are often organized in the form of multienzyme complexes. Such multienzyme complexes also exist in case of dNTP biosynthesis. Although the dNTP-synthesizing complexes exist in several prokaryotic and eukaryotic systems, the best-characterized system is the complex encoded by the Escherichia coli lytic phage T4 (Tomich et al., 1974; Reddy et al., 1977; Prem veer Reddy & Pardee, 1980; Allen et al., 1983; Mathews, 1993). In case of T4, it has been demonstrated that several phage-encoded nucleotide metabolism-related proteins such as thymidylate synthase, dihydrofolate reductase, ribonucleotide reductase, deoxycytidylate deaminase, dC/UTPase and deoxycytidylate hydroxymethylase form a functional aggregate. It is believed that such an aggregate acts as a ‘metabolome’, which plays an important role in maintaining an adequate supply of deoxy-nucleotide precursors for the rapid synthesis of viral DNA (Mathews, 1993).

The presence of genes for ThyX and RNR II in the genome of L5 and related mycobacteriophages has been predicted purely on the basis of in silico analysis. No functional characterization of these proteins has been carried out so far. The present work was thus initiated with the dual objective of characterizing the translated products of genes 48 and 50 and analyzing their expression status during infection. The results, apart from confirming bio-chemically the in silico predictions, shed significant light on the nature of these enzymes, their expression status and their ability to associate with each other.

Materials and methods

Phages, bacterial strains and plasmids

Escherichia coli XL1Blue was used for routine purposes. The expression vector pQE 30 (Qiagen) was used for overexpression of the phage genes. For phage induction experiments, the thermoinducible lysogenic strain Mycobacterium smegmatis L1c1ts (obtained from Dr N.C. Mandal) was used (Chaudhuri et al., 1993; Chatterjee et al., 2000). This strain harbors a thermo-inducible lysogenic mutant of mycobacteriophage L1 as a prophage. Although the complete genome sequence of L1 phage is not available, restriction maps (Lee et al., 1991) and DNA sequences of several L1 genes (Datta & Mandal, 1998; Chattopadhyay et al., 2003; Sau et al., 2004; Ganguly et al., 2007) indicate that L1 is only a minor variant of L5. Hence, the use of an L1 lysogen to study the L5 developmental pathways is justified. Cloning and expression of mycobacteriophage genes 48 and 50 were performed using genomic DNA derived from mycobacteiophage L5 (Hatfull & Sarkis, 1993). In case of Gp50, both enzymatic and immunological experiments were conducted using the L5 protein. In case of Gp48, immunological experiments were conducted using antisera raised against the L5 protein, but for enzymological studies Gp48 of D29 (Ford et al., 1998) was used, as the L5 protein could not be recovered in the soluble fraction. D29 is a lytic phage belonging to the L5 family. Many of the polypeptides encoded by D29 share extensive homology with those encoded by L5. In case of Gp48, the homology is >70%.

Cloning of mycobacteriophage genes in expression vectors

Mycobacteriophage DNA were isolated from L5 or D29 as per standard methods using CsCl purified phages (Chatterjee et al., 2000). The phage DNA was then used for PCR amplification of desired genes. For L5 gene 50, the primers used were P1, forward primer (5′-GGGGTACCTTGACTGACGAAATCC-3′) and P2, reverse primer (5′-CCCAAGCTTCACTTAATGGGGCATGC-3′), (the KpnI and HindIII sites are underlined). The PCR product was digested with KpnI and HindIII for cloning in the expression vector pQE 30 (Qiagen). The L5 gene 48 was amplified using primers P3, forward primer (5′-GAAGATCTCATATGAAAGCCAAACTGATC-3′), and P4, reverse primer (5′-CCCAAGCTTCATATGTCAGTAGCTGTAG-3′) (the BglII and HindIII sites are underlined). The PCR product was digested with BglII and HindIII for cloning in the expression vector pQE 30. D29 gene 48 was amplified using the forward primer (5′-CGGGATCCATGAAAGTCCAACTGATC-3′) and the reverse primer (5′-CCCAAGCTTTCAGCCTCCGTAGCTG-3′) (the BamHI and HindIII sites are underlined). The amplicon was cloned at the BamHI–HindIII site of pQE-30.

Expression and purification of recombinant proteins

Hexa-histidine-tagged recombinant protein purification was performed using Ni-nitrilotriacetic acid (NTA) agarose chromatography either under denaturing conditions or under native conditions as per standard protocols (Qiagen). Cells were harvested by centrifugation and sonicated in buffer A (50 mM sodium phosphate pH 8.0, 300 mM NaCl, 20 mM imidazole). After centrifugation at 14 000 g, the clear supernatant was loaded onto a 2-mL Ni2+-NTA agarose column, pre-equilibrated with buffer A. The column was washed with 10 column volumes of buffer B (50 mM sodium phosphate pH 8.0, 300 mM NaCl, 50 mM imidazole) and the bound protein was eluted with buffer C (50 mM sodium phosphate pH 8.0, 300 mM NaCl, 250 mM imidazole) and analyzed by 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Fractions containing pure protein were pooled and dialyzed overnight at 4 °C in the dark against 50 mM Tris-HCl, pH 7.4, containing 10% glycerol. For isolating proteins under denaturing conditions, pellet was dissolved in lysis buffer D (10 mM Tris, 50 mM sodium phosphate, 300 mM NaCl, 8 M urea, pH 8.0), sonicated, centrifuged and loaded onto a Ni-NTA column. After washing with 10 column volume of buffer E (10 mM Tris, 50 mM sodium phosphate, 300 mM NaCl, 8 M urea, pH 6.3), protein was eluted in buffer F (10 mM Tris, 50 mM sodium phosphate, 300 mM NaCl, 8 M urea, pH 4.5).

Induction of lysogen

The thermoinducible lysogen, M. smegmatis L1c1ts, was grown to an OD of 0.8 at 28 °C in Middlebrook 7H9 (MB7H9) broth as described earlier (Chaudhuri et al., 1993) and induced at 42 °C for 30 min, followed by growth at 37 °C for 3 h. Aliquots were removed at definite intervals (15, 30 and 60 min) after the flasks were immersed in the 42 °C water bath and processed for immunological and enzymatic assays.


Recombinant mycobacteriophage proteins were purified to near homogeneity by affinity chromatography on Ni-NTA agarose under denaturing conditions. The purified protein was further separated from any contaminant proteins by SDS-PAGE. The bands were excised and crushed before injecting into rabbits according to previously standardized protocols (Basu et al., 2002). Both immune and preimmune sera were collected. The specificities of the derived antisera were confirmed by Western blotting (1 : 1000 dilution) using purified proteins. For immunodetection of phage-derived proteins, cell-free extracts of thermo-induced M. smegmatis L1c1ts lysogens were subjected to Western blotting using specific antibodies (1 : 1000 dilution) according to standard protocols (Sambrook et al., 1989).


Mycobacterium smegmatis L1c1ts lysogenic cells were grown at 28 °C and then the temperature was shifted to 42 °C. After 15 min of induction, the proteins expressed were pulse labeled with 35S methionine (specific activity 3000 Ci mmol−1), obtained from BRIT, Mumbai, India), for 3 min. Extracts were prepared by sonicating the labeled cells in immunoprecipitation buffer (50 mM Tris pH 7.4, 500 mM NaCl, 1 mM EDTA, 0.1% Nonidet P-40) and then incubated with immune or preimmune sera (1 : 1000 dilution), overnight at 4 °C in the same buffer. Antibody-target complexes were then pulled down using 25 μL of Protein-A Sepharose beads (Amersham GE) (pre-equilibrated with the immunoprecipitation buffer containing 1% bovine serum albumin to block nonspecific interactions) by centrifugation at 1000 g. After washing five times with the immunoprecipitation buffer, the beads were suspended in 1 × SDS sample buffer, boiled for 5 min and resolved by 13.5% SDS-PAGE. Bands were visualized by autoradiography.

Determination of ThyX activity

The NADPH oxidase activity of ThyX was measured in a total volume of 1 mL. The basal reaction components were 50 mM Tris-HCl, pH 7.4, 1 mM MgCl2, 10% glycerol. In addition, dUMP, NADPH and FAD were added either alone or in specific combinations at desired concentrations. Activity was monitored by net decrease of A340 nm using a CARY 50 spectrophotometer (Varian). An extinction coefficient of 6400 cm−1 at 340 nm (ɛ340) was used to quantify absorption changes.

Thymidylate-synthesizing activity of ThyX was determined by monitoring the amount of tritium transferred to water using [5-3H] dUMP (Movarek Biochemicals) as a substrate as described previously (Griffin et al., 2005). The standard reaction mixture (100 μL) contained 50 mM Tris-HCl, pH 7.4, 200 μM 5, 10-CH2H4folate, 500 μM NADPH, 60 μM FAD, 1 mM MgCl2 and 10% glycerol. The specific activity of the [5-3H]dUMP stock used in the experiments was 14.3 Ci mmol−1. The reaction was started with the addition of 9 μg of purified ThyX. After 3 min, the reaction was terminated by the addition of 300 μL of a 100 mg mL−1 activated charcoal suspension containing 2% tricarboxylic acid, to remove the unused radiolabeled substrate. Samples were mixed at room temperature for 1 h and centrifuged at 12 000 g for 10 min to pellet the charcoal. Radioactivity in the supernatant was determined by liquid scintillation counting. Enzymatic activity of the protein was expressed as nmol of dTMP formed min−1 mg−1.

Thymidylate synthase activity in cell extracts of induced M. smegmatis L1c1ts cells was performed by inducing 200 mL of M. smegmatis L1c1ts culture at 42 °C. Fifty milliliters of aliquots were removed at specified time points and the harvested cells were sonicated. After removal of debris by centrifugation, the extracts (60 μL) were used in the assays. The protein concentrations in the extracts were nearly the same. If necessary, the activity data were normalized with respect to the protein concentration. An assay for ThyX was performed as described above. For ThyA activity, the same buffer was used, except that NADPH and FAD were omitted. The results were expressed as relative activity taking the activity of the uninduced extract as unity.

Spectrophotometric detection of protein-bound FAD

Detection of FAD bound to protein was performed as described earlier (Myllykallio et al., 2002) by incubating the protein at 95 °C for 10 min in the dark. Precipitated protein were pelleted by centrifugation at 10 000 g for 10 min. The absorption spectra of the released flavin present in the supernatant was determined spectrophotometrically from 250 to 750 nm in a 1-cm quartz cuvette.

RNR assay

RNR activity was determined by the diphenylamine procedure of Blakley (1966) using ATP as the substrate. A complete reaction mixture (0.5 mL) contained 0.2 mM ATP, 20 mM dithiothreitol, 10 μM 5′-deoxyadenosylcobalamine (coenzyme B12), 1 mM EDTA, 50 mM potassium phosphate buffer, pH 7.3, and an appropriate amount of enzyme. The reaction mixture was incubated for 1 h at 37 °C and the reaction was terminated by placing the tubes in ice. The reaction mixtures were treated with chloroacetamide and subsequently with diphenylamine. Conversion of ATP to dATP was monitored by recording the A595 nm. In some cases, deoxyribonucleotides were incorporated in the assay at desired concentrations to study stimulatory/inhibitory effects.


Immunodetection experiments

To test whether Gp48 and 50 were at all produced during phage growth, the total protein extracted from M. smegmatis L1c1ts cells induced for varying lengths of time was subjected to Western blotting experiments using respective antisera. The results (Fig. 1a and b) showed that Gp48 or 50 are present at detectable levels in the induced, but not uninduced, extracts.

Figure 1.

 (a and b) Immunodetection of polypeptides synthesized from genes 48 and 50 (a and b, respectively) using the corresponding antisera. (c) Immunoprecipitation using antisera against Gp48, 50 and 49 as indicated. P and I are abbreviated forms of preimmune and immune, respectively. The identities of the immuno-precipitated proteins are indicated by solid arrows. Bands representing unidentified proteins are indicated by broken arrows. L5 Gp49 was expressed and processed for the antibody essentially as described in the case of Gp48 and Gp50. In each blot, the purified protein was incorporated as a control (Ctrl). The time periods (mins) for which the inductions were performed are shown on the top. The positions of molecular weight standards are shown.

Following the demonstration of the presence of these proteins in extracts of induced lysogenic cells by Western blotting, immunoprecipitation of 35S-methionine pulse-labeled proteins was attempted using the respective antibodies. The rationale for this was twofold – first, a positive immunoprecipitation would indicate that these proteins are synthesized actively during the induced phase of phage growth, and second, such an experiment might lead to important associated proteins. The pulse-labeled proteins obtained from cells induced for 15 min were thus subjected to immunoprecipitation using either immune or preimmune sera against Gp48 and 50, respectively (Fig. 1c). As a control experiment, immunoprecipitation of a third protein Gp49 expressed from gene 49 was performed using the corresponding antisera. The results showed that antisera against Gp48 immunoprecipitated the corresponding 28 kDa polypeptide (Fig. 1c marked by an arrow). No immunoprecipitation of this polypeptide, or for that matter any other polypeptide, was observed using the preimmune sera. In case of anti-Gp50, similar immunoprecipitation of the specific target was observed. Interestingly, anti-Gp48 and 50 cross-immunoprecipitated each other's targets and also a couple of other proteins (Fig. 1c, marked by broken arrows). This result indicates that in vivo, Gp48, 50 and several other unidentified proteins associate to form a complex. The fact that none of these bands were precipitated in case of Gp49 gives additional support to the claim that the precipitation is not artifactual. Moreover, it is unlikely that coimmunoprecipitation is due to epitope sharing because in the Western blots, performed at the same dilution of antisera (Fig. 1a and b) no evidence of aberrant reactivity was found.

Gp48-a mycobacteriophage flavin-dependent thymidylate synthase (FDTS)

In order to characterize Gp48 biochemically, attempts were first made to express the L5 gene in E. coli. Although a high expression level was obtained, the protein could not be recovered in the soluble state. In order to overcome this difficulty, the D29 gene 48 that shares extensive homology with the L5 counterpart was expressed in E. coli (Fig. 2a). Suspecting that the purified protein is an FDTS, the possibility of FAD being tightly bound to the enzyme was investigated spectrophotometrically. The enzyme preparation was first heated to release the bound FAD if any and then the absorption spectra of the supernatant were recorded. The resulting spectra revealed discrete peaks that were characteristic of oxidized FAD, (Fig. 2b). This proved that the enzyme contained bound FAD confirming that it is a flavoprotein. ThyX enzymes are known to function through two different mechanisms: ping-pong and sequential. In the ping-pong mechanism, the enzyme reduces FAD in the presence of NADPH in a dUMP-independent manner. In the sequential mechanism, however, both dUMP and NADPH must bind to the enzyme for the reduction of FAD. In other words, the enzyme is expected to have dUMP-dependent NADPH oxidase activity. In order to obtain an insight into which of these mechanisms operates in case of D29 ThyX-catalyzed reactions, the NADPH oxidase activity of the protein was assessed in the absence or in the presence of increasing concentrations of dUMP, maintaining NADPH at a saturating concentration of 200 μM. The results show that the enzyme is a dUMP-dependent NADPH oxidase (Fig. 2c). From the corresponding double-reciprocal plot, the Km could be determined to be about 1 μM for dUMP. The ability of a ThyX to act as a dUMP-dependent NADPH oxidase has been reported in case of the Paramecium bursaria chlorella virus (PBCV)-I enzyme. In the above, assay CH2H4folate was absent. If the same NADPH oxidation assay was performed in the presence of CH2H4folate, a significant dose-dependent inhibition of the NADPH oxidase activity was observed (Fig. 2d). This indicates that as in the case of the PBCV-1 enzyme, the CH2H4folate-binding site of D29 ThyX overlaps with that of the NADPH-binding site. The ability of the enzyme to convert dUMP to dTMP in the presence of NADPH and CH2H4folate (200 μM each) was then assayed by the tritium release assay. From the resulting curve (as shown in Fig. 2e), the Km for dUMP was estimated to be about 3 μM, which is comparable to the 1 μM value obtained using the NADPH assay. The results, taken together, indicate that Gp48 acts as an FDTS by the sequential mechanism as proposed in case of the PBCV-1 enzyme (Graziani et al., 2004).

Figure 2.

 Characterization of D29 Gp48 ThyX activity: (a) Expression and purification of Hexa Histidine-tagged D29 Gp48 protein, – (sup) represents the soluble supernatant, while (E) denotes the protein eluted under native conditions. (b) Spectroscopic analysis of the purified protein following denaturation (solid line). The dotted line represents the typical spectra for FAD. The arrows represent characteristic oxidized flavin peaks at 260, 375 and 450 nm, respectively. (c) NADPH oxidation assay in the presence of varying concentrations of dUMP. The experiment was performed using 2 μM ThyX, in the presence of 60 μM FAD and 200 μM NADPH. The initial velocities were plotted against the concentration of dUMP. (d) Effect of CH2H4folate, the co-substrate for ThyX. The experiment was performed as in (c), except that the dUMP concentration was fixed at 100 μM (e) Conversion of dUMP to dTMP by Gp48. Enzyme activity was measured by monitoring the release of tritium from the radiolabeled substrate [5-3H]dUMP. The activities of the enzyme at different substrate concentrations were plotted according to Michaelis–Menten. Km and Vmax were determined from the corresponding double-reciprocal plot (not shown). All the enzymatic assays (c, d and e) were run in triplicate and the mean values±SE are shown. In some cases, error bars are not visible as they are shorter than the height of the symbol.

These investigations were performed with the D29 enzyme. However, a soluble glutatione-S-transferase-tagged version of the L5 protein was studied. The properties of this fusion protein were similar to that of the D29 protein.

Gp50- a Class II ribonucleotide reductase

In order to characterize Gp50, biochemically the L5 gene was overexpressed in E. coli and purified (Fig. 3a). To test whether this enzyme is capable of reducing ribonucleotides to their deoxycounterparts in a coenzyme B12-dependent manner, assays were performed in the presence of this cofactor, using ATP as the substrate. The results show that this enzyme is capable of reducing ATP to dATP. The activity was absolutely dependent on coenzyme B12 for its activity. Omitting this coenzyme resulted in complete loss of activity (data not shown). This confirms that the reductase belongs to the class II type. The enzyme was also found to reduce only triphosphates and not diphosphates. The Km for the enzymatic reduction of ATP by Gp50 was found to be 0.57 mM (Fig. 3b). Ribonucleotide reductases are known to be allosterically modulated by the activity of various nucleotides. The effect of various deoxy-nucleotide substrates on ATP reduction was studied. Only dGTP was found to stimulate the reduction of ATP (Fig. 3c). The remaining nucleotides had very little effect. The results suggest that the protein has properties similar to that of the reductase from Lactobacillus leichmanii (Booker & Stubbe, 1993).

Figure 3.

 Characterization of the RNR activity of Gp50: (a) Expression and purification of Gp50; – (sup) represents the soluble supernatant and (E) represents the eluted protein. (b) Lineweaver–Burk plot for the enzymatic reduction of ATP to dATP. Assays were carried out using 2 μM enzyme in the presence of 20 mM dithiothreitol, 10 μM Coenzyme B12 and varying concentrations of ATP from 0 to 2 mM. Reduction to dATP was determined by monitoring the absorbance at 595 nm following treatment with diphenylamine. Initial velocities (Δ OD595 nm min−1) were measured at different concentrations of the substrate ATP. (c) Effect of various deoxyribonucleotides. RNR assays were performed at different ATP concentrations in the absence (○) or in the presence of nucleotides – dATP (•), dCTP (□), dGTP (▪) and dTTP (Δ). The nucleotides were used at a concentration of 0.1 mM.

Thymidylate synthase activity in thermally induced M. smegmatis L1c1ts lysogens

In order to investigate the level of thymidylate synthase activity (ThyA and/or ThyX) during induction of phage, cell-free extracts of induced M. smegmatis L1c1ts lysogenic cells were assayed separately for both. To determine ThyA, activity assays were performed in the absence of FAD and NADPH. The results (Fig. 4) showed that ThyA activity in induced extracts decreased steadily relative to the uninduced extracts (line marked by open circles) whereas the ThyX activity increased moderately (squares). Both ThyX and ThyA assays share the same cosubstrate CH2H4folate; therefore, in the ThyX assay, ThyA is also expected to contribute proportionately. In other words, the ThyX assay should give the total thymidylate synthase activity. Now, because ThyA activity reduces it may be inferred that the contribution from ThyX must have increased, as a result of which the overall activity not only did not decrease but actually increased to a certain extent. The results indicate that gene 48 expression may be critical for the maintenance of an adequate supply of thymidylate in a situation where the host enzyme loses activity.

Figure 4.

 Thymidylate synthase (TS) activity monitored in cell-free extracts of Mycobacterium smegmatis L1c1ts cells. Aliquots from extracts of M. smegmatis L1c1ts induced or uninduced cells having the same protein concentration were used in the dUMP to dTMP conversion assay under conditions that yield either ThyA (FAD and NADPH absent) or ThyX activities (all cosubstrates present). The fold increase (or decrease) in activity in the case of induced cells relative to the uninduced was plotted against the induction time. Each data point represents the mean±SE of at least four independent experiments.


Despite being lysogenic by nature, mycobacteriophage L5 is known to inactivate its host in a manner reminiscent of lytic phages (Hatfull & Sarkis, 1993). Although the mechanisms are not known, it follows that following host inactivation the phage takes control of the cellular metabolism for its successful propagation. The central region of the genomes of mycobacteriophage L5 and the related phage D29 contains several genes that putatively encode enzymes required for DNA synthesis as well as nucleotide metabolism (Ford et al., 1998). It is not known as yet whether these proteins do indeed possess the predicted activities and whether they are at all expressed. This investigation was initiated to address these issues in the context of the two genes 48 and 50. It is for the first time that these genes have been expressed in E. coli and their enzymatic activities have been confirmed. It has also been demonstrated that they were detectable only following induction of the lysogen. This indicates that their expression is closely linked to the onset of the lytic cycle.

A novel observation presented here is that these proteins coimmunoprecipitate and are therefore possibly associated in vivo. Such associations involving thymidylate synthase and RNR have been reported in case of phage T4 (Allen et al., 1980), but this is the first time that a similar phenomenon has been indicated to occur in the context of mycobacteriophages. Moreover, it may be noted that in case of T4, the corresponding enzymes are a ThyA-type thymidylate synthase and an aerobic RNR, whereas in this case the thymidylate synthase is of ThyX type and the RNR is of the microaerophillic variety. Although the polypeptides Gp48 and 50 coimmunoprecipitated from cell-free extracts, a similar phenomenon could not be demonstrated using the purified proteins. This indicates that perhaps these proteins do not interact directly within the multiprotein complex, but only through the involvement of other proteins (Kim et al., 2005).

The ThyX activity reported for the L5/D29 Gp48 protein appears to be mechanistically distinct from that reported in case of several other bacterial systems including M. tuberculosis (Sampathkumar et al., 2005). The M. tuberculosis enzyme follows a ping-pong mechanism whereas L5/D29 ThyX, like the PBCV-1 enzyme, functions by the sequential mechanism. To the best of the authors' knowledge; these are the only two examples of ThyX operating through the sequential mechanism. The fact that both are viruses may be coincidental; alternatively, there could be some specific reason why these unrelated viruses use this type of ThyX.

The timing of the synthesis of the respective proteins strongly suggests that they have a role to play in the early stages of the lytic cycle. Evidence that supports such a possibility has also been presented in the form of thymidylate synthase assays using cell-free extracts of induced lysogens. One of the conclusions of these assays is that ThyA activity reduces following the onset of the lytic pathway. This in itself is a novel observation. Understanding how this occurs could in the long run give clues regarding how mycobacteriophages inactivate their host. In the ThyX assays that followed, a marginal increase in activity following induction was detectable. However, as ThyA is also likely to make a contribution in the ThyX assay and because ThyA activity decreases with time, it follows therefore that ThyX activity increases far more than is evident. The results strongly indicate that a shift from a ThyA to a ThyX mode of synthesis occurs as phage growth commences. Whatever is true for ThyX is also probably true for the phage-encoded RNR II, although this has not been verified. The results of this investigation indicate that like other well-known lytic phages, L5 (and related phages) manipulate host nucleotide metabolism in such a way as to benefit itself at the cost of the host.


The authors thank N.C. Mandal, A.K. Tyagi and R. McNerney for the phages L1, L5 and D29, respectively. The project was funded by a grant from DBT, Government of India. B.B. and N.G. are grateful to CSIR and UGC, Government of India for their fellowships. The authors thank P. Halder for technical assistance.