Elsevier

Biochemical Pharmacology

Volume 66, Issue 3, 1 August 2003, Pages 471-479
Biochemical Pharmacology

Cytosolic and mitochondrial deoxyribonucleotidases: activity with substrate analogs, inhibitors and implications for therapy

https://doi.org/10.1016/S0006-2952(03)00290-9Get rights and content

Abstract

Nucleoside analogs act as prodrugs that must be converted to 5′-phosphates by intracellular kinases to become active in the treatment of viral and oncological diseases. Activation may be reversed by dephosphorylation if the 5′-phosphates are substrates for 5′-nucleotidases. Dephosphorylation by cytosolic enzymes decreases the efficacy of the analogs, whereas dephosphorylation by mitochondrial enzymes may decrease mitochondrial toxicity. Both effects may influence the outcome of therapy. We investigated the dephosphorylation of the 5′-phosphates of commonly used nucleoside analogs by two cytosolic (cN-II and dNT-1) and one mitochondrial (dNT-2) nucleotidase. Most uracil/thymine nucleotide analogs were dephosphorylated by all three human enzymes but cytosine-containing nucleotide analogs were inactive. Only cN-II showed some activity with the monophosphates of the two purine analogs 2-chloro-2′-deoxyadenosine and 9-β-d-arabinosylguanine. We conclude that overproduction of any of the three 5′-nucleotidases cannot explain development of resistance against cytosine analogs but that overproduction of cN-II could lead to resistance against purine analogs. Of the tested analogs, only (E)-5-(2-bromovinyl)-2′-deoxyuridine was preferentially dephosphorylated by mitochondrial dNT-2. We propose that in future developments of analogs this aspect be considered in order to reduce mitochondrial toxicity. We tested inhibition of dNT-1 and dNT-2 by a large variety of synthetic metabolically stable nucleoside phosphonate analogs and found one (PMcP-U) that inhibited dNT-1 and dNT-2 competitively and a second (DPB-T) that inhibited dNT-2 by mixed inhibition. Both inhibitors are useful for specific 5′-nucleotidase assays and structural studies and may open up possibilities for therapy.

Introduction

5′-Nucleotidases are ubiquitous enzymes that dephosphorylate nucleoside monophosphates producing nucleosides and inorganic phosphate. The enzymes show varying specificities for the sugar and base moieties. The cDNAs of seven enzymes have been cloned [1], [2], [3], [4], [5], [6], [7]. We are particularly interested in two related 5′-nucleotidases that prefer deoxyribonucleotides as substrates, one located in the cytosol (=dNT-1) [4] and the other in mitochondria (=dNT-2) [5]. Both in humans and in the mouse the genes for dNT-1 and dNT-2 map to the same chromosome and have identical intron/exon organization. The amino acid sequences of all four proteins are approximately 50% identical, disregarding the leader sequences for the two mitochondrial enzymes [8]. The structures of human dNT-2 [9] and dNT-1,1 including the active site of the enzymes, were recently determined and a detailed reaction mechanism was proposed for their catalytic function.

We proposed that both dNT-1 and dNT-2 function in the homeostasis of deoxynucleoside triphosphate pools required for DNA synthesis [10]. dNT-2 has a narrow specificity for dUMP and dTMP among natural deoxyribonucleotides [5]. Also dNT-1 is very active with these two deoxyribonucleotides but, in addition, dephosphorylates other deoxyribonucleotides, notably dGMP and dIMP [4]. We found that in cultured cells the catabolic activity of mouse dNT-1 counteracted the anabolic activity of thymidine and deoxycytidine kinases, thereby preventing the accumulation of pyrimidine deoxyribonucleotides in the cytosol and providing evidence for a regulatory role of the enzyme [11]. dNT-2 probably has a similar intramitochondrial function with respect to thymidine and deoxyuridine nucleotides. In humans intramitochondrial accumulation of dTTP leads to a severe genetic disease [12], attesting to the importance of a strict control of this pool.

5′-Nucleotidases contribute to the outcome of the treatment of cancer and viral diseases with nucleoside analogs. To interfere with DNA synthesis the analogs must first be phosphorylated and activated by cytosolic kinases. The efficiency of this process may be compromised if the resulting 5′-monophosphate is a good substrate for a cytosolic nucleotidase. Within a group of related nucleosides, an analog whose phosphorylated form is a poor substrate for dNT-1 would be more effective than a related analog that is rapidly dephosphorylated. A further concern is that sensitivity of cells to therapy by a nucleoside analog and the development of resistance against treatment might be related to the activity of cytosolic 5′-nucleotidases [13]. Experiments from different laboratories have demonstrated that cells kept in culture in the presence of increasing concentrations of an analog and developing resistance against the analog may increase their content of a soluble 5′-nucleotidase [14], [15], [16]. In clinical studies with patients suffering from leukemia the success of treatment was positively related to low levels of a cytosolic nucleotidase (high KM 5′-nucleotidase=cN-II) in the patient’s cells [17]. Also another cytosolic 5′-nucleotidase (=cN-I) may play a role as cell lines overproducing cN-I showed increased resistance towards CdA, dFdC and ddC [18]. It was suggested that nucleotidase inhibitors might increase the efficacy of some analogs [13].

Such considerations apply to cytosolic 5′-nucleotidases, including dNT-1. A different scenario can be envisaged for the mitochondrial dNT-2. Mitochondrial toxicity is a major problem during treatment with some nucleoside analogs, in particular during long-term treatment of HIV and hepatitis B virus [19]. The rationale of the treatment is to use analogs that selectively interfere with viral replication without affecting cellular nuclear DNA polymerases. In several cases the mitochondrial DNA polymerase is, however, affected and the ensuing deficit in mitochondrial DNA replication gives severe side-effects [20]. dNT-2 operates inside mitochondria and can there lower the concentration of the phosphorylated analogs provided that they are substrates for the enzyme. The monophosphate of an ideal nucleoside analog used in therapy thus should be a poor substrate for dNT-1 and other cytosolic nucleotidases but a good substrate for dNT-2.

Here we investigate the ability of the 5′-monophosphates of various nucleoside analogs used in cancer and virus therapy to act as substrates for dNT-2 and two cytosolic nucleotidases, dNT-1 and cN-II. We also describe experiments with two nucleoside phosphonoalkyl derivatives that are inhibitors of 5′-nucleotidases.

Section snippets

Materials

The 5′-phosphates of the nucleoside analogs were prepared from commercially available nucleosides by chemical synthesis (AZT, d4T, araC, dFdC, ddC, BVdU, FdU) [21] or by enzymatic phosphorylation with deoxycytidine kinase (araG, CdA) or with Drosophila melanogaster deoxynucleoside kinase (araT). The sources of the kinases and their use for the phosphorylation of the nucleoside analogs were described earlier [22]. The purity of all nucleoside monophosphates was >99% as based on NMR data (31P and

Specificity of human dNT-1, dNT-2 and cN-II for substrate analogs

The results in Table 1 demonstrate the ability of the three enzymes to dephosphorylate the 5′-phosphates of various nucleoside analogs used in viral and cancer therapy. Activities were measured at a single substrate concentration (2 mM) and are recorded in percent of the dephosphorylation of dUMP (for dNT-1 and dNT-2) or IMP (for cN-II).

Among the nucleotides-containing uracil or thymine as the base, all except araTMP were dephosphorylated. For both human dNT-1 and dNT-2 dTMP was the next best

Discussion

The results of Table 1 are of interest in connection with the therapeutic use of nucleoside analogs. AZT, d4T, ddC and 3TC are all used during multidrug treatment of HIV infection [19]. Fluorouracil, dFdC and araC are used to treat both leukemias and some solid tumors. HIV infections require intensive long-time chemotherapy and under those circumstances several of the analogs give rise to toxic effects that can be ascribed to their interference with mitochondrial DNA replication [19]. Because

Acknowledgements

We thank Dr. C. Pasti, Pasteur Institute, Paris, for the gift of 3TCMP, Dr. S. Sugano, Institute of Medical Science, University of Tokyo, for sending us the EST clone of human dNT-1, and Prof. J. Rozenski, Rega Institute, Leuven, for performing the structural determinations of the deoxyribonucleotidase inhibitors. This work was supported by the European Commission (Grant No. QLRT-CT-2000-01004). Financial support to V.B. by AIRC, Italian Association for Cancer Research, and Telethon (Grant No.

References (34)

Cited by (0)

View full text