Direct photoaffinity-labelling of human deoxycytidine kinase with the feedback inhibitor dCTP

Re-Discovery of Pyrimidine Salvage as Target in Cancer Therapy

Nucleotides are synthesized through two distinct pathways: de novo synthesis and nucleoside salvage. Whereas the de novo pathway synthesizes nucleotides from amino acids and glucose, the salvage pathway recovers nucleosides or bases formed during DNA or RNA degradation. In contrast to high proliferating non-malignant cells, which are highly dependent on the de novo synthesis, cancer cells can switch to the nucleoside salvage pathways to maintain efficient DNA replication. Pyrimidine de novo synthesis remains the target of interest in cancer therapy and several inhibitors showed promising results in cancer cells and in vivo models. In the 1980s and 1990s, poor responses were however observed in clinical trials with several of the currently existing pyrimidine synthesis inhibitors. To overcome the observed limitations in clinical trials, targeting pyrimidine salvage alone or in combination with pyrimidine de novo inhibitors was suggested. Even though this approach showed initially promising results, it received fresh attention only recently. Here we discuss the re-discovery of targeting pyrimidine salvage pathways for DNA replication alone or in combination with inhibitors of pyrimidine de novo synthesis to overcome limitations of commonly used antimetabolites in various preclinical cancer models and clinical trials. We also highlight newly emerged targets in pyrimidine synthesis as well as pyrimidine salvage as a promising target in immunotherapy.

siRNA knockdown of mitochondrial thymidine kinase 2 (TK2) sensitizes human tumor cells to gemcitabine

Nucleoside metabolism enzymes are determinants of chemotherapeutic drug activity. The nucleoside salvage enzyme deoxycytidine kinase (dCK) activates gemcitabine (2', 2'-difluoro-2'-deoxycytidine) and is negatively regulated by deoxycytidine triphosphate (dCTP). Reduction of dCTP in tumor cells could, therefore, enhance gemcitabine activity. Mitochondrial thymidine kinase 2 (TK2) phosphorylates deoxycytidine to generate dCTP. We hypothesized that: (1) TK2 modulates human tumor cell sensitivity to gemcitabine, and (2) antisense knockdown of TK2 would decrease dCTP and increase dCK activity and gemcitabine activation. siRNA downregulation of TK2 sensitized MCF7 and HeLa cells (high and moderate TK2) but not A549 cells (low TK2) to gemcitabine. Combined treatment with TK2 siRNA and gemcitabine increased dCK. We also hypothesized that TK2 siRNA-induced drug sensitization results in mitochondrial damage that enhances gemcitabine effectiveness. TK2 siRNA and gemcitabine decreased mitochondrial redox status, DNA content, and activity. This is the first demonstration of a direct role for TK2 in gemcitabine resistance, or any independent role in cancer drug resistance, and further distinguishes TK2 function from that of other dTMP-producing enzymes [cytosolic TK1 and thymidylate synthase (TS)]. siRNA knockdown of TK1 and/or TS did not sensitize cancer cells to gemcitabine indicating that, among the 3 enzymes, only TK2 is a candidate therapeutic target for combination with gemcitabine.

Phosphorylation of deoxycytidine kinase on Ser-74: impact on kinetic properties and nucleoside analog activation in cancer cells

Deoxycytidine kinase (dCK) (EC 2.7.1.74) is a key enzyme in the activation of several therapeutic nucleoside analogs (NA). Its activity can be increased in vivo by Ser-74 phosphorylation, a property that could be used for enhancing NA activation and clinical efficacy. In line with this, studies with recombinant dCK showed that mimicking Ser-74 phosphorylation by a S74E mutation increases its activity toward pyrimidine analogs. However, purine analogs had not been investigated. Here, we show that the S74E mutation increased the k(cat) for cladribine (CdA) by 8- or 3-fold, depending on whether the phosphoryl donor was ATP or UTP, for clofarabine (CAFdA) by about 2-fold with both ATP and UTP, and for fludarabine (F-Ara-A) by 2-fold, but only with UTP. However, the catalytic efficiencies (k(cat)/Km) were not, or slightly, increased. The S74E mutation also sensitized dCK to feed-back inhibition by dCTP, regardless of the phosphoryl donor. Importantly, we did not observe an increase of endogenous dCK activity toward purine analogs after in vivo-induced increase of Ser-74 phosphorylation. Accordingly, treatment of CLL cells with aphidicolin, which enhances dCK activity through Ser-74 phosphorylation, did not modify the conversion of CdA or F-Ara-A into their active triphosphate form. Nevertheless, the same treatment enhanced activation of gemcitabine (dFdC) into dFdCTP in CLL as well as in HCT-116 cells and produced synergistic cytotoxicity. We conclude that increasing phosphorylation of dCK on Ser-74 might constitute a valuable strategy to enhance the clinical efficacy of some NA, like dFdC, but not of CdA or F-Ara-A.

Deoxynucleoside kinases encoded by the yaaG and yaaF genes of Bacillus subtilis. Substrate specificity and kinetic analysis of deoxyguanosine kinase with UTP as the preferred phosphate donor

The overlapping yaaG and yaaF genes from Bacillus subtilis were cloned and overexpressed in Escherichia coli. Purification of the gene products showed that yaaG encoded a homodimeric deoxyguanosine kinase (dGK) and that yaaF encoded a homodimeric deoxynucleoside kinase capable of phosphorylating both deoxyadenosine and deoxycytidine. The latter was identical to a previously characterized dAdo/dCyd kinase (Møllgaard, H. (1980) J. Biol. Chem. 255, 8216-8220). The purified recombinant dGK was highly specific toward 6-oxopurine 2'-deoxyribonucleosides as phosphate acceptors showing only marginal activities with Guo, dAdo, and 2',3'-dideoxyguanosine. UTP was the preferred phosphate donor with a Km value of 6 microm compared with 36 microm for ATP. In addition, the Km for dGuo was 0.6 microm with UTP but 6.5 microm with ATP as phosphate donor. The combination of these two effects makes UTP over 50 times more efficient than ATP. Initial velocity and product inhibition studies indicated that the reaction with dGuo and UTP as substrates followed an Ordered Bi Bi reaction mechanism with UTP as the leading substrate and UDP the last product to leave. dGTP was a potent competitive inhibitor with respect to UTP. Above 30 microm of dGuo, substrate inhibition was observed, but only with UTP as phosphate donor.

Differences in kinetic properties of pure recombinant human and mouse deoxycytidine kinase

Human and mouse deoxycytidine kinase (dCK) (EC 2.7.1.74) were cloned and expressed in Escherichia coli. Michaelis-Menten kinetics were determined for the purified enzymes with 2'-deoxycytidine (dCyd), 2'-deoxyadenosine (dAdo), 2-chloro-2'-deoxyadenosine (CdA), 2',3'-dideoxycytidine (ddCyd) and 9-beta-D-arabinofuranosylguanine (araG) as substrates and ATP and UTP as phosphate donors. Both human and mouse dCK showed highest affinity to dCyd with Km values of 0.05-0.2 microM. The anti-leukaemic compound CdA was the superior substrate of the nucleoside analogues tested. Both enzymes were able to efficiently utilize ATP and UTP as phosphate donors. However, the use of UTP instead of ATP as phosphate donor decreased Km values for all substrates investigated. The kinetic properties of mouse and human dCK differed in that the human enzyme showed higher affinity for the substrates dAdo, CdA, ddCyd and araG. The human enzyme also showed higher affinity for ATP and UTP. The ability to phosphorylate dCyd was, however, similar for both human and mouse dCK. At physiological concentration of the feedback inhibitor dCTP, mouse dCK showed lower activity than human dCK for all substrates investigated.

Mammalian deoxyribonucleoside kinases

The mammalian deoxyribonucleoside kinases are deoxycytidine kinase, thymidine kinase 1 and 2 and deoxyguanosine kinase. These enzymes phosphorylate deoxyribonucleosides and thereby provide an alternative to de novo synthesis of DNA precursors. Their activities are essential for the activation of several chemotherapeutically important nucleoside analogues. In recent years, these enzymes have been thoroughly characterised with regard to structure, substrate specificity and patterns of expression. In this review, these results are reviewed and furthermore, the physiologic metabolic role of the anabolic enzymes is discussed in relation to catabolic pathways. The significance of this information for the development of therapeutic protocols and choice of animal model systems is discussed. Finally, alternative pathways for nucleoside analogue phosphorylation are surveyed, such as the phosphotransfer capacity of 5'-nucleotidase.

Mammalian thymidine kinase 2. Direct photoaffinity labeling with [32P]dTTP of the enzyme from spleen, liver, heart and brain

Thymidine kinase 2 (TK2), also called mitochondrial thymidine kinase, is a pyrimidine deoxyribonucleoside kinase expressed in all cells and tissues. It was recently purified to apparent homogeneity from human leukemic spleen and the active enzyme was shown to be a monomer of a 29-kDa polypeptide. The enzyme is feedback-inhibited by both end products, dCTP and dTTP. Here we show that TK2 purified from several different sources, including purified beef heart mitochondria, could be directly photoaffinity labeled with radioactive dTTP (approximately 18% of all TK2 molecules were cross-linked to dTTP after 20 min of ultraviolet irradiation) or to a lower extent with dCTP. Photo-incorporation was inhibited by the presence of the other effector but also the phosphate donor ATP blocked photolabeling, with dTTP. Addition of nucleoside substrates gave only a marginal inhibition of photo-incorporation. There were no detectable difference in the molecular size of photolabeled TK2 isolated from human spleen, brain or placenta, monkey liver, beef heart and beef heart mitochondria. Nor was there any significant differences in the enzyme kinetic properties of these enzymes. Cleavage of labeled TK2 with cyanogen bromide showed that dTTP was incorporated into a single 3-kDa peptide. TK2 was the only pyrimidine deoxynucleoside kinase expressed in liver, heart and brain. A detailed characterization of the subunit structure and substrate specificity of this enzyme is of importance for the design of new antiviral and cytostatic therapies based on nucleoside analogs.

CDP-choline: 1,2-diacylglycerol cholinephosphotransferase from rat liver microsomes. II. Photoaffinity labeling by radioactive CDP-choline analogs

Photoaffinity labeling of cholinephosphotransferase from rat liver microsomes directly by its substrate, [32P]CDP-choline or by a synthetic photoreactive CDP-choline analog, 3'(2')-O-(4-benzoyl)benzoyl [32P]CDP-choline (BB-[32P]CDP-choline), was examined for the possible identification of its molecular form on subsequent SDS-PAGE followed by 32P-autoradiography. When the partially purified cholinephosphotransferase was photoirradiated in the presence of [32P]CDP-choline, a considerable amount of 32P-radioactivity was incorporated into the TCA-insoluble component. This incorporation was dependent on irradiation time, Mg2+ or Mn(2+)-requiring and inhibited strongly by the presence of Ca2+. Either CDP-choline or CDP-ethanolamine inhibited the ultraviolet irradiation-dependent incorporation of 32P-radioactivity into the TCA-insoluble component in a dose-dependent manner, whereas neither phosphocholine or 5'-CDP had any effect on this process. These results strongly suggested that the observed 32P-incorporation from [32P]CDP-choline into the protein component could be a consequence of the covalent interaction between cholinephosphotransferase and its substrate, [32P]CDP-choline. Two polypeptides, 25 kDa and 18 kDa, with high 32P-radioactivity were clearly identified on a SDS gel after the direct photoaffinity labeling with [32P]CDP-choline for more than 5 min of ultraviolet irradiation. On the other hand, when BB-[32P]CDP-choline was used as a photoaffinity ligand, a single polypeptide with apparent molecular size of 55 kDa could be rapidly photolabeled within 2.5 min, then this band gradually lost its 32P-radioactivity with increasing time of ultraviolet irradiation. Thus, the overall results strongly indicated that cholinephosphotransferase in rat liver microsomes exists most likely as a 55 kDa polypeptide (or subunit) and that 25 kDa and 18 kDa peptides identified after the direct photoaffinity labeling with [32P]CDP-choline were probably the photo-cleavage products of cholinephosphotransferase during the prolonged ultraviolet irradiation, both of which could contain the catalytic domain of the original enzyme protein(s).

Viral thymidine kinases and their relatives

Thymidine kinases were described for cellular life long before it was shown that they could also be encoded by viruses, but the viral thymidine kinase genes were the first to be sequenced. These enzymes have been extraordinarily useful to the researcher, serving first to help label DNA, then to get thymidine analogs incorporated into DNA for therapeutic and other purposes and more recently to move genes from one genome to another. Knowledge of the nucleotide and amino acid sequences of these enzymes has allowed some deductions about their possible three-dimensional structure, as well as the location on the polypeptide of various functions; it has also allowed their classification into two main groups: the herpesviral thymidine/eukaryotic deoxycytidine kinases and the poxviral and cellular thymidine kinases; the relationships of the mitochondrial enzyme are still not clear.