Non-homologous recombination of deoxyribonucleoside kinases from human and Drosophila melanogaster yields human-like enzymes with novel activities

In antiviral and cancer therapy, deoxyribonucleoside kinases (dNKs) are often the rate-limiting step in activating nucleoside analog (NA) prodrugs into their cytotoxic, phosphorylated forms. We have constructed libraries of hybrid enzymes by non-homologous recombination of the pyrimidine-specific human thymidine kinase 2 and the broad-specificity dNK from Drosophila melanogaster; their low sequence identity has precluded engineering by conventional, homology-dependent shuffling techniques. From these libraries, we identified chimeras that phosphorylate nucleoside analogs with higher activity than either parental enzyme, and that possess new activity towards the anti-HIV prodrug 2',3'-didehydro-3'-deoxythymidine (d4T). These results demonstrate the potential of non-homologous recombination within the dNK family for creating enzymes with new and improved activities towards nucleoside analogs. In addition, our results exposed a previously unknown role for the C-terminal regions of these dNKs in determining substrate selectivity.     

A few amino acid substitutions can convert deoxyribonucleoside kinase specificity from pyrimidines to purines

In mammals, the four native deoxyribonucleosides are phosphorylated to the corresponding monophosphates by four deoxyribonucleoside kinases, which have specialized substrate specificities. These four enzymes are likely to originate from a common progenitor kinase. Insects appear to have only one multisubstrate deoxyribonucleoside kinase (dNK, EC 2.7.1.145), which prefers pyrimidine nucleosides, but can also phosphorylate purine substrates. When the structures of the human deoxyguanosine kinase (dGK, EC 2.7.1.113) and the dNK from Drosophila melanogaster were compared, a limited number of amino acid residues were identified and proposed to be responsible for the substrate specificity. Three of these key residues in Drosophila dNK were then mutagenized and the mutant enzymes were characterized regarding their ability to phosphorylate native deoxyribonucleosides and nucleoside analogs. The mutations converted the dNK substrate specificity from predominantly pyrimidine specific into purine specific. A similar scenario could have been followed during the evolution of kinases. Upon gene duplication of the progenitor kinase, only a limited number of single amino acid changes has taken place in each copy and resulted in substrate-specialized enzymes.     

Nucleoside 5'-triphosphate analogs as positive and negative effectors of mammalian ribonucleotide reductase

Ribonucleotide reductase activity is strongly regulated by nucleoside 5'-triphosphates acting as positive and negative effectors. With the use of dGTP analogs, araGTP and dITP, it was found that the structural requirements of dGTP to serve as a positive effector of ADP reductase were not the same as the requirements for dGTP to serve as a negative effector of CDP and ADP reductase activities. The dTTP analogs methylenedTTP and dideoxyTTP also gave different responses in terms of activating GDP reductase activity and inhibiting CDP and ADP reductase activities. Etheno-ATP and etheno-dATP were inactive as positive and negative effectors, respectively, of CDP reductase activity. DideoxyATP was less active than dATP as a negative effector. Formycin ATP was a very poor substitute for ATP as a positive effector of CDP reductase. These studies indicate that the effector sites are very specific in terms of binding nucleoside triphosphates as positive or negative modulators of ribonucleotide reductase activity.     

A quantitative model for allosteric control of purine reduction by murine ribonucleotide reductase

The reduction of purine nucleoside diphosphates by murine ribonucleotide reductase requires catalytic (R1) and free radical-containing (R2) enzyme subunits and deoxynucleoside triphosphate allosteric effectors. A quantitative 16 species model is presented, in which all pertinent equilibrium constants are evaluated, that accounts for the effects of the purine substrates ADP and GDP, the deoxynucleoside triphosphate allosteric effectors dGTP and dTTP, and the dimeric murine R2 subunit on both the quaternary structure of murine R1 subunit and the dependence of holoenzyme (R1(2)R2(2)) activity on substrate and effector concentrations. R1, monomeric in the absence of ligands, dimerizes in the presence of substrate, effectors, or R2(2) because each of these ligands binds R1(2) with higher affinity than R1 monomer. This leads to apparent positive heterotropic cooperativity between substrate and allosteric effector binding that is not observed when binding to the dimeric protein itself is evaluated. Allosteric activation results from an increase in k(cat) for substrate reduction upon binding of the correct effector, rather than from heterotropic cooperativity between effector and substrate. Neither the allosteric site nor the active site displays nucleotide base specificity: dissociation constants for dGTP and dTTP are nearly equivalent and K(m) and k(cat) values for both ADP and GDP are similar. R2(2) binding to R1(2) shows negative heterotropic cooperativity vis-à-vis effectors but positive heterotropic cooperativity vis-à-vis substrates. Binding of allosteric effectors to the holoenzyme shows homotropic cooperativity, suggestive of a conformational change induced by activator binding. This is consistent with kinetic results indicating full dimer activation upon binding a single equivalent of effector per R1(2)R2(2).     

Structural basis for feedback inhibition of the deoxyribonucleoside salvage pathway: studies of the Drosophila deoxyribonucleoside kinase

Deoxyribonucleoside kinases are feedback inhibited by the final products of the salvage pathway, the deoxyribonucleoside triphosphates. In the present study, the mechanism of feedback inhibition is presented based on the crystal structure of a complex between the fruit fly deoxyribonucleoside kinase and its feedback inhibitor deoxythymidine triphosphate. The inhibitor was found to be bound as a bisubstrate inhibitor with its nucleoside part in the nucleoside binding site and with its phosphate groups partially occupying the phosphate donor site. The overall structure of the enzyme--inhibitor complex is very similar to the enzyme--substrate complexes with deoxythymidine and deoxycytidine, except for a conformational change within a region otherwise directly involved in catalysis. This conformational change involves a magnesium ion, which is coordinated in the inhibitor complex to the phosphates and to the primary base, Glu52, that normally is positioned close to the 5'-OH of the substrate deoxyribose.     

Structural basis for substrate specificities of cellular deoxyribonucleoside kinases.

Deoxyribonucleoside kinases phosphorylate deoxyribonucleosides and activate a number of medically important nucleoside analogs. Here we report the structure of the Drosophila deoxyribonucleoside kinase with deoxycytidine bound at the nucleoside binding site and that of the human deoxyguanosine kinase with ATP at the nucleoside substrate binding site. Compared to the human kinase, the Drosophila kinase has a wider substrate cleft, which may be responsible for the broad substrate specificity of this enzyme. The human deoxyguanosine kinase is highly specific for purine substrates; this is apparently due to the presence of Arg 118, which provides favorable hydrogen bonding interactions with the substrate. The two new structures provide an explanation for the substrate specificity of cellular deoxyribonucleoside kinases.     

Interaction of DNA polymerase I of Escherichia coli with nucleotides. Antagonistic effects of single-stranded polynucleotide homopolymers

Binding of deoxyribonucleoside 5'-triphosphates to DNA polymerase I of Escherichia coli was measured by using a microscale nonequilibrium dialysis method. It allowed rapid and economic measurement of dissociation constants, with negligible interfering side reactions. A stoichiometry of 1 mol of nucleoside 5'-triphosphate/mol of DNA polymerase was measured, and the occurrence of a single binding site was established, for which the nucleotides competed in the binary complex with the polymerase. Binding affinities decreased in the order dGTP greater than or equal to dATP greater than dCTP congruent to dTTP. These results are in agreement with previous findings [Englund, P. T., Huberman, J. A., Jovin, T. M., & Kornberg, A. (1969) J. Biol. Chem. 244, 3038-3044] except that, in a few cases, values of dissociation constants were smaller by factors of 2-3. The cations Mg2+ and Mn2+, as well as spermine, slightly enhanced complex stability at low levels and decreased it at high concentrations, while NaCl and Hg2+ had only destabilizing effects. Recognition between nucleoside 5'-triphosphates and nucleotide templates was studied by titration of the polymerase-[3H]dGTP complex with polynucleotide homopolymers. Complementary poly(dC) did not affect binding of dGTP, and non-complementary templates caused rejection of the nucleotide. Rejection of dGTP followed a saturation dependence with an equivalence of 110 +/- 10 monomer units of polynucleotides bound per molecule of DNA polymerase. The results favor a model by which recognition arises chiefly from the stereogeometrical fit of complementary template and nucleoside 5'-triphosphate into a rigid binding site.     

Nucleotide analogs and molecular modeling studies reveal key interactions involved in substrate recognition by the yeast RNA triphosphatase

RNA triphosphatases (RTPases) are involved in the addition of the distinctive cap structure found at the 5′ ends of eukaryotic mRNAs. Fungi, protozoa and some DNA viruses possess an RTPase that belongs to the triphosphate tunnel metalloenzyme family of enzymes that can also hydrolyze nucleoside triphosphates. Previous crystallization studies revealed that the phosphohydrolase catalytic core is located in a hydrophilic tunnel composed of antiparallel β-strands. However, all past efforts to obtain structural information on the interaction between RTPases and their substrates were unsuccessful. In the present study, we used computational molecular docking to model the binding of a nucleotide substrate into the yeast RTPase active site. In order to confirm the docking model and to gain additional insights into the molecular determinants involved in substrate recognition, we also evaluated both the phosphohydrolysis and the inhibitory potential of an important number of nucleotide analogs. Our study highlights the importance of specific amino acids for the binding of the sugar, base and triphosphate moieties of the nucleotide substrate, and reveals both the structural flexibility and complexity of the active site. These data illustrate the functional features required for the interaction of an RTPase with a ligand and pave the way to the use of nucleotide analogs as potential inhibitors of RTPases of pathogenic importance.     

Nucleotide specificity in microtubule assembly in vitro

A procedure is described for removing most of the GDP bound at the exchangeable GTP binding site (E site) of tubulin. Microtubule protein containing substoichiometric amounts of GDP at the E site is found to polymerize in response to: (a) two nonhydrolyzable ATP analogues, adenylyl imidodiphosphate (AMP-PNP) and adenylyl beta, gamma-methylenediphosphonate (AMP-PCP); and (b) substoichiometric levels of GTP or dGTP. The results are interpreted as suggesting that: (1) when GDP is removed from tubulin, the E site shows broad specificity for nucleoside triphosphates: (2) microtubule assembly can be induced by the binding of substoichiometric amounts of nucleoside triphosphate to the E site.     

An AZT Analog with Strongly Pairing Ethynylpyridone Nucleobase and Its Antiviral Activity against HSV1

Challenges resulting from novel viruses or new strains of known viruses call for new antiviral agents. Nucleoside analogs that act as inhibitors of viral polymerases are an attractive class of antivirals. For nucleosides containing thymine, base pairing is weak, making it desirable to identify nucleobase analogs that pair more strongly with adenine, in order to compete successfully with the natural substrate. We have recently described a new class of strongly binding thymidine analogs that contain an ethynylmethylpyridone as base and a C-nucleosidic linkage to the deoxyribose. Here we report the synthesis of the 3′-azido-2′,3′-deoxyribose derivative of this compound, dubbed AZW, both as free nucleoside and as ProTide phosphoramidate. As a proof of principle, we studied the activity against Herpes simplex virus type 1 (HSV1). Whereas the ProTide phosphoramidate suffered from low solubility, the free nucleoside showed a stronger inhibitory effect than that of AZT in a plaque reduction assay. This suggests that strongly pairing C-nucleoside analogs of pyrimidines have the potential to become active pharmaceutical ingredients with antiviral activity.     

Transient state kinetic studies of the MutT-catalyzed nucleoside triphosphate pyrophosphohydrolase reaction

The MutT pyrophosphohydrolase, in the presence of Mg2+, catalyzes the hydrolysis of nucleoside triphosphates by nucleophilic substitution at Pbeta, to yield the nucleotide and PP(i). The best substrate for MutT is the mutagenic 8-oxo-dGTP, on the basis of its Km being 540-fold lower than that of dGTP. Product inhibition studies have led to a proposed uni-bi-iso kinetic mechanism, in which PP(i) dissociates first from the enzyme-product complex (k3), followed by NMP (k4), leaving a product-binding form of the enzyme (F) which converts to the substrate-binding form (E) in a partially rate-limiting step (k5) [Saraswat, V., et al. (2002) Biochemistry 41, 15566-15577]. Single- and multiple-turnover kinetic studies of the hydrolysis of dGTP and 8-oxo-dGTP and global fitting of the data to this mechanism have yielded all of the nine rate constants. Consistent with an "iso" mechanism, single-turnover studies with dGTP and 8-oxo-dGTP hydrolysis showed slow apparent second-order rate constants for substrate binding similar to their kcat/Km values, but well below the diffusion limit (approximately 10(9) M(-1) s(-1)): k(on)app = 7.2 x 10(4) M(-1) s(-1) for dGTP and k(on)app = 2.8 x 10(7) M(-1) s(-1) for 8-oxo-dGTP. These low k(on)app values are fitted by assuming a slow iso step (k5 = 12.1 s(-1)) followed by fast rate constants for substrate binding: k1 = 1.9 x 10(6) M(-1) s(-1) for dGTP and k1 = 0.75 x 10(9) M(-1) s(-1) for 8-oxo-dGTP (the latter near the diffusion limit). With dGTP as the substrate, replacing Mg2+ with Mn2+ does not change k1, consistent with the formation of a second-sphere MutT-M2+-(H2O)-dGTP complex, but slows the iso step (k5) 5.8-fold, and its reverse (k(-5)) 25-fold, suggesting that the iso step involves a change in metal coordination, likely the dissociation of Glu-53 from the enzyme-bound metal so that it can function as the general base. Multiple-turnover studies with dGTP and 8-oxo-dGTP show bursts of product formation, indicating partially rate-limiting steps following the chemical step (k2). With dGTP, the slow steps are the chemical step (k2 = 10.7 s(-1)) and the iso step (k5 = 12.1 s(-1)). With 8-oxo-dGTP, the slow steps are the release of the 8-oxo-dGMP product (k4 = 3.9 s(-1)) and the iso step (k5 = 12.1 s(-1)), while the chemical step is fast (k2 = 32.3 s(-1)). The transient kinetic studies are generally consistent with the steady state kcat and Km values. Comparison of rate constants and free energy diagrams indicate that 8-oxo-dGTP, at low concentrations, is a better substrate than dGTP because it binds to MutT 395-fold faster, dissociates 46-fold slower, and has a 3.0-fold faster chemical step. The true dissociation constants (KD) of the substrates from the E-form of MutT, which can now be obtained from k(-1)/k1, are 3.5 nM for 8-oxo-dGTP and 62 microM for dGTP, indicating that 8-oxo-dGTP binds 1.8 x 10(4)-fold tighter than dGTP, corresponding to a 5.8 kcal/mol lower free energy of binding.     

ANALOGS OF GUANINE NUCLEOSIDE TRIPHOSPHATES FOR SEQUENCING APPLICATIONS

We have synthesized more than 30 different deoxyribonucleosides and triphosphates with modifications either in the base or the phosphate moiety as analogs of 2′-dGTP for DNA sequencing applications. All the modified nucleoside triphosphates were tested as substrates for DNA polymerases, including Sequenase? T7 DNA polymerase or Thermo Sequenase? DNA polymerase. Two of the analogs, 7-ethyl-7-deaza-dGTP and 7-hydroxymethyl-7-deaza- dGTP meet our requirements as better sequencing reagents.     

The gene 1.2 protein of bacteriophage T7 interacts with the Escherichia coli dGTP triphosphohydrolase to form a GTP-binding protein

Escherichia coli encodes a dGTP triphosphohydrolase (dGTPase) that cleaves dGTP to deoxyguanosine and tripolyphosphate. dGTP is hydrolyzed with a Michaelis constant (Km) of 5 microM and a maximal velocity (Vmax) of 1.8 mumols/min/mg. The ribonucleotide GTP is a poor substrate with a much lower affinity. It is hydrolyzed with a Km of 150 microM and Vmax of 0.07 mumols/min/mg. Bacteriophage T7 encodes a specific inhibitor of dGTPase, the gene 1.2 protein, that forms a tight complex with the enzyme. The enzyme-inhibitor complex binds dGTP with a dissociation constant (KD) of 1.5 microM, but the bound dGTP is not hydrolyzed. It remains stably bound to the complex with a half-life of approximately 5 min. In contrast, dGTP is unable to bind to gene 1.2 protein alone, and dGTP bound to dGTPase alone is quickly hydrolyzed and released. Surprisingly, the dGTPase-gene 1.2 protein complex has a higher affinity for GTP than for dGTP. GTP is stably bound to the dGTPase-gene 1.2 protein complex with a half-life greater than 30 min and KD of 0.8 microM; GTP is not stably bound to either dGTPase or gene 1.2 protein alone. Both GTP and dGTP bind to and stabilize the dGTPase-gene 1.2 protein complex, inhibiting its dissociation. Although the presence of dGTP induces conformation changes in dGTPase so that it is unable to associate with the gene 1.2 protein, saturating concentrations of GTP have no such effect. The enzyme efficiently associates with its inhibitor in the presence of GTP. These results indicate that E. coli dGTPase and gene 1.2 protein interact to form a high affinity GTP-binding site. dGTP is most effective in preventing the association of the enzyme with the inhibitor whereas GTP is most effective in preventing the dissociation of the enzyme-inhibitor complex.     

Isolation and characterization of a mutant of L1210 murine leukemia deficient in nitrobenzylthioinosine-insensitive nucleoside transport

L1210 mouse leukemia cells exhibit two distinct types of nucleoside transport activity that have similar kinetic properties and substrate specificity, but differ markedly in their sensitivity to the inhibitor nitrobenzylthioinosine (NBMPR) (Belt, J. A. (1983) Mol. Pharmacol. 24, 479-484). It is not known whether these two transport activities are mediated by a single protein or by separate and distinct nucleoside transport proteins. We have isolated a mutant from the L1210 cell line that has lost the NBMPR-insensitive component of nucleoside transport, but retains NBMPR-sensitive transport. In the parental cell line 20-40% of the nucleoside transport activity is insensitive to 1 microM NBMPR. In the mutant, however, uridine and thymidine transport are almost completely inhibited by NBMPR. Consistent with the loss of NBMPR-insensitive transport, the mutant cells can be protected from the toxic effects of several nucleoside analogs by NBMPR. In contrast, the toxicity of the same analogs in the wild type cells is not significantly affected by NBMPR, presumably due to uptake of the nucleosides via the NBMPR-insensitive transporter. On the other hand, NBMPR-sensitive transport in the mutant appears to be unaltered. The mutant is not resistant to cytotoxic nucleosides in the absence of NBMPR and the cells retain the wild type complement of high affinity binding sites for NBMPR. Furthermore, the affinity of the binding site for the inhibitor is similar to that of parental L1210 cells. These results suggest that NBMPR-sensitive and NBMPR-insensitive nucleoside transport in L1210 cells are mediated by genetically distinct proteins. To our knowledge this is the first report of a mutant deficient in NBMPR-insensitive nucleoside transport.     

An axial binding site in the Tetrahymena precursor RNA.

Previous studies allow the construction of three distinct models of the binding of G and arginine within the active site of the Tetrahymena self-splicing preribosomal precursor RNA. These models (base triple, axial I and axial II) are now distinguished by measurements on the specificity of RNAs with nucleotide substitutions at positions spanning the site. Because the semi-conserved unpaired nucleotide 263 has no effect on substrate or inhibitor selection by the Tetrahymena RNA we conclude that the axial I model is improbable. In contrast, data with substituted RNAs and nucleoside analogs suggest that nucleotide 265 makes a hydrogen bond with the substrate. Accordingly the active site appears axial because substrate contacts exist at more than one nucleotide on the 5' side of the P7 helix. The effects of this hydrogen bond are observable in cases where the donor or acceptor is on the RNA, whether nucleotide 265 is a purine or pyrimidine, or whether nucleotide 265 is mispaired, wobble paired or normally paired. This pattern is consistent with the axial II model. Molecular dynamics and energy minimization calculations lead to the same conclusions as these site-directed substitutions; the base triple and axial I models are unstable dynamically. Under thermal agitation, the third model site (axial II) is transformed to a related, but more stable structure, axial III. The axial III active site is characterized by the extrusion of the conserved bulged base 263 from the P7 helix, a semi-pocket for G base formed by stacking of nucleotide 262, and formation of all bonds to the G base originally proposed for both the base triple and axial II sites. Because of these hydrogen bonds the axial III site is also consistent with data on enzymatic specificity. The axial III model indicates an unforeseen capacity for pocket formation within the groove of an RNA helix, suggests that the site may be unusually flexible, and bears on a hypothesis concerning the origin of the genetic code.     

Investigation of the Substrate Recognition of Drosophila melanogaster Nucleoside Kinase by Site Directed Mutagenesis

The deoxyribonucleoside kinase of Drosophila melanogaster (Dm‐dNK) has a broad substrate specificity and a higher catalytic rate than other known deoxyribonucleoside kinases. Therefore it is a natural candidate for possible use as a suicide gene in combined gene/chemotherapy of cancer. We have performed site directed mutagenesis and tested different truncated forms of the enzyme in order to increase the affinity for ganciclovir.     

Pan-pathway based interaction profiling of FDA-approved nucleoside and nucleobase analogs with enzymes of the human nucleotide metabolism

To identify interactions a nucleoside analog library (NAL) consisting of 45 FDA-approved nucleoside analogs was screened against 23 enzymes of the human nucleotide metabolism using a thermal shift assay. The method was validated with deoxycytidine kinase; eight interactions known from the literature were detected and five additional interactions were revealed after the addition of ATP, the second substrate. The NAL screening gave relatively few significant hits, supporting a low rate of "off target effects." However, unexpected ligands were identified for two catabolic enzymes guanine deaminase (GDA) and uridine phosphorylase 1 (UPP1). An acyclic guanosine prodrug analog, valaciclovir, was shown to stabilize GDA to the same degree as the natural substrate, guanine, with a ΔT(agg) around 7°C. Aciclovir, penciclovir, ganciclovir, thioguanine and mercaptopurine were also identified as ligands for GDA. The crystal structure of GDA with valaciclovir bound in the active site was determined, revealing the binding of the long unbranched chain of valaciclovir in the active site of the enzyme. Several ligands were identified for UPP1: vidarabine, an antiviral nucleoside analog, as well as trifluridine, idoxuridine, floxuridine, zidovudine, telbivudine, fluorouracil and thioguanine caused concentration-dependent stabilization of UPP1. A kinetic study of UPP1 with vidarabine revealed that vidarabine was a mixed-type competitive inhibitor with the natural substrate uridine. The unexpected ligands identified for UPP1 and GDA imply further metabolic consequences for these nucleoside analogs, which could also serve as a starting point for future drug design.     

Exploring the active site of herpes simplex virus type-1 thymidine kinase by X-ray crystallography of complexes with aciclovir and other ligands

Antiherpes therapies are principally targeted at viral thymidine kinases and utilize nucleoside analogs, the triphosphates of which are inhibitors of viral DNA polymerase or result in toxic effects when incorporated into DNA. The most frequently used drug, aciclovir (Zovirax), is a relatively poor substrate for thymidine kinase and high-resolution structural information on drugs and other molecules binding to the target is therefore important for the design of novel and more potent chemotherapy, both in antiherpes treatment and in gene therapy systems where thymidine kinase is expressed. Here, we report for the first time the binary complexes of HSV-1 thymidine kinase (TK) with the drug molecules aciclovir and penciclovir, determined by X-ray crystallography at 2.37 Å resolution. Moreover, from new data at 2.14 Å resolution, the refined structure of the complex of TK with its substrate deoxythymidine (R = 0.209 for 96% of all data) now reveals much detail concerning substrate and solvent interactions with the enzyme. Structures of the complexes of TK with four halogen-containing substrate analogs have also been solved, to resolutions better than 2.4 Å. The various TK inhibitors broadly fall into three groups which together probe the space of the enzyme active site in a manner that no one molecule does alone, so giving a composite picture of active site interactions that can be exploited in the design of novel compounds. Proteins 32:350–361, 1998. © 1998 Wiley-Liss, Inc.     

Structural basis for the changed substrate specificity of Drosophila melanogaster deoxyribonucleoside kinase mutant N64D

The Drosophila melanogaster deoxyribonucleoside kinase (Dm-dNK) double mutant N45D/N64D was identified during a previous directed evolution study. This mutant enzyme had a decreased activity towards the natural substrates and decreased feedback inhibition with dTTP, whereas the activity with 3'-modified nucleoside analogs like 3'-azidothymidine (AZT) was nearly unchanged. Here, we identify the mutation N64D as being responsible for these changes. Furthermore, we crystallized the mutant enzyme in the presence of one of its substrates, thymidine, and the feedback inhibitor, dTTP. The introduction of the charged Asp residue appears to destabilize the LID region (residues 167-176) of the enzyme by electrostatic repulsion and no hydrogen bond to the 3'-OH is made in the substrate complex by Glu172 of the LID region. This provides a binding space for more bulky 3'-substituents like the azido group in AZT but influences negatively the interactions between Dm-dNK, substrates and feedback inhibitors based on deoxyribose. The detailed picture of the structure-function relationship provides an improved background for future development of novel mutant suicide genes for Dm-dNK-mediated gene therapy.