Deoxyribonucleoside triphosphates (dNTPs) are the building blocks of DNA, and a constant supply is essential for the synthesis and maintenance of both the nuclear and mitochondrial genomes. Antiviral nucleoside analogues and inborn errors of nucleotide metabolism frequently cause dNTP pool imbalances, leading to depletion of mtDNA (mitochondrial DNA) in non-replicating tissues. mtDNA depletion, in turn, causes failure of the mitochondrial respiratory chain, resulting in cellular energy depletion and cell death. Accordingly, it is important to understand the origin and regulation of dNTPs in order to develop safe and effective treatments. In this issue of the Biochemical Journal, Morris et al. have pursued the origin of pyrimidines in perfused adult rat heart. They found no evident role for the nucleotide de novo synthesis pathway and also demonstrated that AZT (3′-azido-3′-deoxythymidine; also known as zidovudine) substantially decreased the TTP pool. Their results underscore the general importance of the mitochondrial deoxyribonucleoside salvage pathway in adult tissues, and particularly in AZT-mediated toxicity. Although the role of nucleoside salvaging versus de novo synthesis in humans remains unclear, the study of tissue cultures and animal models contribute to the understanding of the intricate network of biochemical pathways, maintaining the cellular dNTP supply.
- de novo synthesis pathway
- deoxyribonucleotide salvage pathway
- deoxyribonucleoside triphosphate pool
- mitochondrial DNA (mtDNA)
- mitochondrial respiratory chain
- nucleoside analogue
- nucleotide metabolism
Deoxyribonucleoside triphosphates (dNTPs) are the building blocks of DNA, and a balanced supply of all four dNTPs is crucial for the synthesis and repair of both the nuclear and mitochondrial genomes. The mammalian dNTP supply is maintained and regulated by an intricate network of enzymes; dysfunction of one component leads to disruption of the whole network, causing lack or surplus of one or more dNTP. Imbalances in the dNTP pool lead to arrest of DNA synthesis, mutations, and eventually to cell death.
dNTP pool imbalance is not only caused by inborn errors of nucleotide metabolism, but may also be acquired, secondary to treatment with antiviral and antineoplastic NAs (nucleoside analogues). As mitochondria are constantly replicating throughout the cell cycle, the mtDNA (mitochondrial DNA) is especially vulnerable to changes in dNTP concentration and the toxic side effects of NAs frequently present as severe mitochondrial dysfunction in differentiated tissues. Accordingly, understanding the origin and regulation of the cellular dNTP pool is of the utmost importance, and contributes to the development of safer and more effective antiviral and anticancer treatments.
THE INTRICATE ENZYMATIC NETWORK MAINTAINING THE dNTP POOLS
The mammalian dNTP pool is contained in two separate compartments, the cytosolic and the mitochondrial. Both pools are maintained and regulated by cytosolic and mitochondrial enzymes, some of them with overlapping substrate specificities, some counteracting each other and many subjected to allosteric regulation. Furthermore, nucleosides and nucleotides are constantly exchanged between the two compartments via a number of dedicated transporters, rendering the whole network extremely complex (reviewed in [1–3]) (Figure 1).
In replicating cells, the cytosolic dNTP pool is mainly replenished by de novo synthesis from pRpp (phosphoribosyl pyrophosphate) via rNDP (ribonucleotide diphosphate) reduction by the RNR, R1-R2 (ribonucleotide reductase, subunits R1 and R2), and to a minor extent by recycling via the deoxyribonucleoside salvage pathway. The salvage pathway consists of three consecutive phosphorylation steps; the first one catalysed by TK1 (thymidine kinase 1) specific for thymidine and deoxyuridine, and dCK (deoxycytidine kinase) specific for deoxycytidine and deoxypurines. The following phosphorylation steps are performed by NMPKs (deoxyribonucleoside monophosphate kinases) and NDPK (deoxyribonucleoside diphosphate kinase). The synthesis of TTP is somewhat more complicated, and also involves the activity of thymidylate synthase, which converts dUMP derived from the de novo or salvage pathway into TMP .
The mitochondrial dNTPs are supplied either by import from the cytosolic pool or by the mitochondrial deoxyribonucleoside salvage pathway. The mitochondrial salvage pathway is catalysed by mitochondrial enzymes with the deoxypyrimidine-specific TK2 (thymidine kinase 2) and the deoxypurine-specific dGK (deoxyguanosine kinase) performing the first phosphorylation step.
In contrast with replicating cells, in differentiated tissues, nuclear DNA synthesis is relatively low, reducing the requirement for DNA precursors [1,3]. Nevertheless, mtDNA synthesis and DNA repair is constantly ongoing, independently of the cell cycle, necessitating a constant (albeit lower) supply of dNTPs. Because two crucial cytoplasmic enzymes, TK1 and RNR, R1-R2 and, to a certain extent, also dCK, are S-phase-regulated, the cellular dNTP supply becomes highly dependent on the mitochondrial salvage pathway, and especially on TK2 and dGK, which are rate-limiting [2,4]. However, recent findings confirm that dNTP pools in resting cells are not only dependent on the mitochondrial enzymes, but also on the cytosolic p53-inducible R2 subunit of RNR (R1-p53R2) of the de novo pathway .
Counteracting deoxyribonucleotide synthesis are the catabolic enzymes, cytoplasmic (cN-II, dNT-1) and mitochondrial (dNT-2) nucleotides, and the nucleoside-degrading enzymes TP (thymidine phosphorylase) and PNP (purine nucleoside phosphorylase). The final dNTP concentration is thus a result of synthesis and degradation pathways [4,5] (Figure 1). This complex regulation allows the cellular dNTP supply to remain adequate and balanced during both replication and quiescence.
To which extent each pathway, and possibly also a mitochondrial de novo pathway, contributes to the dNTP pool in different tissues is currently a topic of debate, and several research groups have invested considerable effort in trying to clarify this issue [1–5].
ANTICANCER AND ANTIVIRAL TREATMENTS
The cell membrane does not permit diffusion of highly charged molecules such as nucleotides. Accordingly, NAs used for anticancer and antiviral treatment are supplied as inactive pro-drugs which are dependent on intracellular phosphorylation converting them into their active form. NA activation is catalysed by the cytoplasmic and mitochondrial salvage pathways. Subsequently, the NA phosphates are incorporated into the replicating DNA of malignant cells or viruses acting as chain terminators. On the other hand, nucleotide phosphatases degrade the activated NAs. and their overexpression renders resistance towards anticancer treatment [4,5]. Consequently, there is an ongoing effort to synthesize and characterize ‘tailor-made’ NAs that are activated by the salvage pathway enzymes, yet without being degraded.
MITOCHONDRIAL TOXICITY OF ANTIVIRAL NUCLEOSIDE ANALOGUES
Apart from inhibiting the replication of malignant cells and viruses, NAs frequently cause toxic side effects. A recognized unwanted effect is mitochondrial toxicity due to HAART (highly active antiretroviral viral therapy) in the treatment of AIDS. Specifically, the NRTIs (nucleoside reverse transcriptase inhibitors) AZT (3′-azido-3′-deoxythymidine; also known as zidovudine) and fialuridine may cause mitochondrial dysfunction with hepatic failure, skeletal myopathy, cardiac dysfunction and lactic acidosis. As HAART therapy has changed AIDS from a lethal disease to a chronic one, treatment is now long term, and elucidation of the mechanisms leading to mitochondrial toxicity is important for the development of safer treatment regimes. The ‘pol-γ hypothesis’ postulates that the toxicity is a result of the NRTI triphosphates inhibiting the mitochondrial DNA polymerase (pol-γ) causing reduction of mtDNA copy number (i.e. mtDNA depletion), which in turn leads to respiratory chain dysfunction, energy depletion and increased free radical production . However, lately the pol-γ hypothesis has been questioned and the focus has shifted towards the AZT-activating deoxyribonucleoside kinases; cycling cells mainly phosphorylate AZT via TK1, whereas resting cells and tissues phosphorylate AZT via mitochondrial TK2. Notably, TK2 is also the sole enzyme performing the first step of deoxypyrimidine phosphorylation in non-replicating cells. Accordingly, there is a distinct possibility that mtDNA depletion due to AZT is, in fact, caused by interference with normal TK2 function [4,7]. To corroborate the ‘AZT–TK2’ theory, a comprehensive understanding of pyrimidine and AZT metabolism is required. The mechanism of AZT toxicity and TTP metabolism has previously been investigated in tissue culture models using replicating and quiescent cells of various origins [1,3,4]. In this issue of the Biochemical Journal , Edward McKee and co-workers have taken these studies a step forward, and addressed this topic by using a whole-organ model. They studied isolated adult rat hearts perfused with labelled nucleotide precursors, and tracked the formation of deoxypyrimidines. Although they examined meticulously the fate of orotate (a pyrimidine nucleotide precursor), they did not detect any formation of pyrimidines or deoxypyrimidines. They concluded that there is no evident role for the de novo pyrimidine synthesis pathway in adult heart. Moreover, they demonstrated the importance of the salvage pathway by showing that TTP and dCTP were solely synthesized by phosphorylation of thymidine and deoxycytidine, and that perfusion with AZT clearly decreased the TTP pool . The present results are in accordance with their previous findings in adult rat heart and liver mitochondria that AZT is a competitive inhibitor of thymidine phosphorylation, and that AZT monophosphate accumulates and is not phosphorylated to AZT triphosphate. Taken together, these results contribute to the understanding of tissue pyrimidine metabolism in general, and the mechanism of AZT toxicity in particular. The findings also highlight the need to develop NRTIs that do not interfere with mitochondrial deoxyribonucleotide metabolism.
MITOCHONDRIAL DISEASES CAUSED BY INBORN ERRORS OF DEOXYRIBONUCLEOTIDE METABOLISM
A distinct group of mitochondrial disorders are the MDSs (mtDNA-depletion syndromes), which are characterized by a profound reduction of mtDNA copy number in affected tissues. mtDNA depletion leads to decreased synthesis of respiratory chain complexes (I, III, IV and V) containing mtDNA-encoded subunits, while complex II remains relatively normal. It is not surprising that deficiency of pol-γ, which is directly involved in mtDNA synthesis, is associated with the largest group of patients affected with MSD or that Twinkle, the mitochondrial helicase, is mutated in other forms of mtDNA maintenance defects.
However, during the last decade, when mutations in enzymes participating in nucleotide metabolism were identified in conjunction with MDS, the study of deoxyribonucleotide pools became relevant also for the elucidation of the pathomechanism in these disorders (reviewed in [2,8]).
Mutations in both mitochondrial deoxyribonucleoside kinases (dGK and TK2) were identified in two forms of MDS manifesting in different tissues. Mutated dGK is associated with the hepatocerebral form of mtDNA depletion, whereas mutated TK2 mainly affects muscle. The different phenotypes most probably are derived from the lack of enzymes with overlapping substrates in the affected tissues or inadequate residual activity relative to demand [4,9]. The tissue specificity of SUCLA (succinyl-CoA synthase A) deficiency, another milder form of MDS, could be explained by the existence of a compensating subunit in unaffected tissues, and a link to nucleotide metabolism is provided by SCS interaction with mitochondrial NDPK. A connection between the mitochondrial salvage pathway and mtDNA depletion seemed logical; thus it was somewhat surprising that mutated cytosolic RNR subunit p53-R2 was recently shown to cause yet another form of severe MDS. This discovery led to a re-examination of the importance of the cytosolic de novo pathway in resting cells. This finding also led to a re-drawing of previous illustrations of mitochondrial nucleotide metabolism. (Remarkably, an additional revision is the ‘mitochondrial deoxyribonucleotide carrier’ DNC that was redefined as a thiamine pyrophosphate carrier.) Another cytosolic enzyme, TP (thymidine phosphorylase), was in fact the first identified nucleotide disorder associated with mtDNA aberrations, causing MNGIE (mitochondrial neurogastrointestinal encephalomyopathy). TP is a catabolic enzyme and deficiency leads to the accumulation of thymidine and deoxyuridine (MNGIE was finally corrected by bone marrow transplantation, after less successful treatment attempts to remove serum deoxypyrimidines by dialysis or platelet infusion). In addition to TP, PNP has also been linked with mtDNA aberrations.
From all data taken together, it can now be concluded that a lack, as well as a surplus, of nucleosides and nucleotides may cause defective mtDNA synthesis and maintenance.
“OF MICE (RATS) AND MEN”
Before the identification of the molecular causes of MDS, physicians and scientists involved in NA and NRTI research and treatment were working separately from those studying and treating MDS. The need for elucidation of the mechanisms of NA toxicity and MDS pathogenicity, as well as the practical challenges of research in the field of nucleotides and mitochondria, brought these two groups together. The collaboration yielded many interesting new data from different model systems (cell cultures, rat, mouse and human tissues and lately transgenic mouse models). Still, some fundamental questions remain open, exemplified by a few discrepancies in the quest to understand the origin of dTTP: (1) heart failure is a relatively common side-effect of NA treatments, so why is cardiomyopathy not the hallmark of TK2 deficiency? [6–8]; (2) why does one study in normal mouse tissues confirm the low TK2 activity detected in human muscle relative to liver, whereas another study shows the opposite? [2,9]; (3) why is the mitochondrial dNTP pool asymmetric with the dGTP content increased relative to TTP? [1,10]; (4) where does the enigmatic mitochondrial RNR hide? ; (5) and finally, if the mitochondrial salvage pathway is so efficient as demonstrated in rat tissues, why is human RNR-p52R2 deficiency so devastating? [7,8].
Clearly not all issues are resolved and although transgenic mice are promising, not all data obtained from animal and tissue culture models can accurately be translated to humans . Nonetheless, the pursuit towards better understanding of the biochemical pathways controlling the dNTP pool is ongoing, while the ultimate goal is to restore mitochondrial function by conscientiously intervening with nucleotide metabolism .
I am thankful for the support of the United Mitochondrial Disease Foundation.
Abbreviations: AZT, 3′-azido-3′-deoxythymidine (zidovudine); dCK, deoxycytidine kinase; dGK, deoxyguanosine kinase; dNTP, deoxyribonucleotide triphosphate; HAART, highly active antiretroviral viral therapy; MDS, mitochondrial DNA-depletion syndrome; MNGIE, mitochondrial neurogastrointestinal encephalomyopathy; mtDNA, mitochondrial DNA; NA, nucleoside analogue; NDPK, deoxyribonucleoside diphosphate kinase; NRTI, nucleoside reverse transcriptase inhibitor; pol-γ, mitochondrial DNA polymerase; RNR, ribonucleotide reductase; TK1/2, thymidine kinase 1/2; TP, thymidine phosphorylase
- © The Authors Journal compilation © 2009 Biochemical Society