Nucleoside Diphosphate Kinase (Molecular Biology)

For the first 35 years after its discovery in the early 1950s, nucleoside diphosphate kinase (NDP kinase) was considered the quintessential "housekeeping enzyme," responsible only for the transfer of the g-phosphate from a variety of nucleoside triphosphates to any common nucleoside diphosphate . Initially, NDP kinase was discovered as a partner in the succinyl-CoA synthetase reaction, using GTP synthesized by that enzyme to drive the synthesis of ATP from ADP. Later, it was recognized that the enzyme can participate in the biosynthesis of all common ribo- and deoxyribonucleoside triphosphates from the respective diphosphates and a phosphate donor (usually ATP) (see Nucleotides, Nucleosides, And Nucleobases). The past decade, however, has revealed a bewildering array of regulatory and developmental roles ascribed to NDP kinase, as well as additional catalytic capabilities.

1. NDP Kinase Reaction

NDP kinase functions via a phosphoenzyme intermediate. The reaction begins with transfer of the g-phosphate from the nucleoside triphosphate substrate—NjTP in reaction (1) to a specific histidine residue of the enzyme. After dissociation of the resultant diphosphate (NjDP), the second substrate (N2DP, the phosphate acceptor) binds, and phosphate is transferred from the enzyme, yielding N2TP, the nucleoside triphosphate product. Thus, the enzyme displays classical ping-pong kinetics (see Kinetic Mechanisms, Enzyme):

The enzyme is quite efficient, with kcat values over 1000 sec 1. The equilibrium constant for the reaction is close to 1. Because of this, and because ATP is the most abundant nucleoside triphosphate in most cells and organelles (see Adenylate Charge), the primary role of the enzyme has been considered to be the synthesis of the other seven triphosphates from the respective diphosphates and ATP.

Steady-state kinetic analysis revealed that the enzyme is rather nonspecific with regard to both phosphate acceptor and donor. However, recent pre-steady-state kinetic analysis of a human NDP kinase has revealed more specificity than was previously thought. This analysis depended on, first, a change in intrinsic tryptophan fluorescence accompanying phosphorylation of the enzyme and, second, the stability of the phosphoenzyme intermediate, which permitted its isolation for analysis of enzyme dephosphorylation by N2DP. As shown in Table 1, the second-order rate constants for the formation of E ~P varied 17-fold among the five triphosphates tested, whereas the dephosphorylation rate constants varied 145-fold among the eight phosphate acceptors tested. In general, purines are better substrates than pyrimidines, and ribonucleotides are better substrates than deoxyribonucleotides.

Table 1. Rate Constants for Phosphorylation and Dephosphorylation of Human NDP Kinase

Base	Second-Order Rate Constant
	rNTP	dNTP	rNDP	dNDP
Adenine	7.9	—	17.9	4.5
Guanine	12.2	—	29.1	10.3
Uracil	2.0	—	3.8	1.3
Cytosine	0.7	—	1.3	0.2
Thymine	—	2.3	—	4.3a

^a Data for rNTPs and dNTPs refer to enzyme phosphorylation rates; data for rNDPs and dNDPs refer to dephosphorylation of the phosphoenzyme.Source of the data: ref. 2.

2. Structure of NDP Kinase

The NDP kinase protein has been rather highly conserved through evolution, and there is about 43% amino acid sequence identity between the human enzyme and that of Escherichia coli. Figure 1 summarizes relationships among some of the known NDP kinase sequences. Although the primary structure is highly conserved, the quaternary structure is not; the enzymes of eukaryotic origin are homohexamers whereas the prokaryotic enzymes have a homotetrameric structure. Subunit molecular weights range from 15 to 18 kDa. Interestingly, preliminary data suggest that at least one mitochondrial isoform of a eukaryotic NDP kinase might not have a hexameric structure.

Figure 1. A phylogenetic tree illustrating the evolutionary relationships among NDP kinase amino acid sequences. The numbers above the horizontal lines represent evolutionary time in arbitrary units. The vertical lines indicate the positions of common ancestors.

X-ray crystallography structures have been described for the NDP kinases from Dictyostelium discoideum (4, 5), Drosophila melanogaster (6), and human tissue (7). The tertiary structures are similar. The Dictyostelium structure has been determined complexed with ADP and with dTDP. Figure 2 shows the active-site structure of the ADP-enzyme complex. The nucleotide is bound near His122, the residue that becomes phosphorylated during the reaction. The b-phosphate interacts with two arginine residues and a threonine that are conserved in all known NDP kinases, and the ribose 2′ and 3′ hydroxyls similarly have polar interactions with Lys19 and Asn119, which are also conserved. Of considerable interest, the purine base forms no polar contacts with the enzyme, nor does thymine in the dTDP-enzyme complex. The absence of specific contacts involving the purine or pyrimidine base accounts for the relatively low nucleotide specificity of the enzyme. However, the structure explains a distinct aspect of specificity that the enzyme does display. Bourdais et al (8) reported that the diphosphates of the anti-HIV drugs azidothymidine, dideoxyadenosine, and dideoxythymidine are very poor substrates. This finding points up the importance of the interaction between the enzyme and the sugar 3′ hydroxyl, which is missing in these analogues, a feature that may be significant in design of more effective antiviral nucleoside drugs.

Figure 2. Structure of the active site of Dictyostelium NDP kinase with bound ADP. ADP is shown with open bonds, with the purine base at the bottom of the figure. Amino acid side chains are shown in black; W is a water molecule.

The structure of the ADP-NDP kinase complex suggests a model for the transition state leading to the phosphoenzyme intermediate (4). Amino acid residues involved in both binding and catalysis, shown in Figure 3, are residues that are conserved in all NDP kinases analyzed to date.

Figure 3. Structure of the proposed NDP kinase transition state, based on the structure of Dictyostelium NDP kinase complexed with ADP.

3. Additional Functions of NDP Kinase

Several recent developments have graphically demonstrated that the NDP kinase protein does more than simply synthesize nucleoside triphosphates from diphosphates. Some of these additional roles are hinted at by the existence of specific protein-protein interactions involving NDP kinase. Other functions have been traced by genetic analysis to the structural genes for NDP kinase. It is evident that NDP kinase lies at the heart of a number of developmental, genetic, and metabolic control systems.

3.1. Human nm23 Genes and Metastatic Cancer

Analysis of a closely related series of murine tumor cell lines revealed the existence of a gene, named nm23 (nonmetastatic clone no. 23), whose action correlated with the potential of the cell lines to develop metastatic tumors after implantation in mice. Lines of low metastatic potential expressed the gene at high levels (9). Sequence analysis revealed the nm23 gene to be the structural gene for an NDP kinase. Human cells contain two major Nm23 proteins, Nm23-H1 and Nm23-H2 (NDP kinase isoforms A and B, respectively). Action of the A isoform is specifically associated with metastasis. The NDP kinase enzymatic activity of the protein is not essential for this activity because some mutant forms of the protein that lack kinase activity still retain the metastatic suppression activity. Since the original cloning and analysis of the two human nm23 genes, two additional members of the NDP kinase protein family, less closely related to each other and to Nm23-H1 and -H2, have been described in human tissues.

Important clues to the biochemical activity involved in tumor suppression are beginning to accumulate (10). NDP kinase has a protein kinase activity, being able to transfer phosphate from its catalytically important histidine residue to other proteins, including histidine residues on ATP citrate lyase and succinyl-CoA synthetase, and aspartate or glutamate residues on a number of human membrane proteins. Using a cell motility assay with breast cancer cells, Wagner et al (10) explored the activities of a series of nm23-H1 mutants. Mutations that abolished motility suppression activity also interfered with the phosphorylation of Asp and Glu residues in the membrane proteins, with minimal effects on the protein histidine kinase activity. Thus, the suppression of metastasis correlates closely with the transfer of phosphate to acidic amino acid residues. This activity is similar to that seen in two-component regulatory systems in bacteria; indeed, E. coli NDP kinase also can transfer phosphate from histidine to protein aspartate residues, and evidence strongly suggests regulatory roles for this activity (11).

3.2. Wing Development and the Prune Gene in Drosophila

In Drosophila, NDP kinase is encoded by the awd gene (abnormal wing development). Mutations in this gene are normally not lethal, but one particular mutation, a Pro-to-Ser substitution at residue 97, is lethal when in combination with a mutation in the prune gene. Prune controls eye color, by somehow regulating the activity of GTP cyclohydrolase, an enzyme involved in the synthesis of the pteridine eye pigments. However, the prune gene itself evidently encodes a small protein similar to the mammalian GTPase-activating proteins involved in Ras-mediated signal transduction pathways (12). In vitro experiments with one such protein (13) suggest that NDPK can use a protein-GDP complex as a substrate, with direct conversion of the bound GDP to GTP occurring in the absence of dissociation from the protein. Similar findings have been described for mammalian NDP kinases (14) but, to date, interactions of this type have not been unambiguously demonstrated in vivo.

The Pro-to-Ser mutation that makes awd lethal in combination with prune is called the kpn mutation (killer of prune). The proline residue involved lies in a loop that participates in subunit contacts. When the same mutation was engineered into the Dictyostelium NDP kinase , the mutant enzyme was found to undergo ready dissociation from its hexameric structure to folded monomers that retained some enzymatic activity. Lascu et al. (15) speculated that comparable dissociation occurs in vivo in kpn mutant cells and that the conditional lethality of the mutation derives from a deleterious effect of these monomers, still unknown.

Similar conclusions have been drawn from analysis of a natural mutation in human NDP kinase A (Nm23-H1), which is found in several aggressive neuroblastoma tumors (16). This mutation involves a Ser-to-Gly substitution at position 120. In vitro, the mutant enzyme was found to be quite unstable, and when it renatured, an intermediate accumulated that had the properties of a molten globule. Whether the accumulation of such a folding intermediate at abnormal levels in vivo might somehow contribute toward the tumor phenotype is an exceedingly interesting, unanswered question.

3.3. Control of c-myc Gene Transcription

Postel et al (17) proposed that human NDP kinase B also serves as a transcription factor for the c-myc proto-oncogene. This role was identified by screening a cervical cell carcinoma cDNA library with DNA containing binding sites for PuF, the purine-binding transcription factor. Sequence analysis of a positive clone revealed the PuF gene to be identical to nm23-H2, the structural gene for NDP kinase B. The enzyme has been shown to activate c- myc transcription in vitro and to bind preferentially to pyrimidine-rich single-stranded DNA (18). This is a specific activity of the B isoform, whereas the metastatic tumor suppression described earlier is associated with the A isoform. Nevertheless, it is intriguing to consider whether there is a common element in the actions of these two proteins with respect to tumorigenesis and tumor suppression.

4. Protein-Protein Interactions Involving NDP Kinase

As befits a protein involved in a range of biological processes, NDP kinase has been shown to interact with numerous proteins. In Pseudomonas aeruginosa, NDP kinase copurifies with succinyl-CoA synthetase (19), which is consistent with the idea that one role of NDP kinase is to generate ATP from the substrate-level phosphorylation step of the citric acid cycle. The possible connection with substrate-level phosphorylation is seen also in frog heart preparations, where NDP kinase was found to copurify with five proteins (20), including glyceraldehyde 3-phosphate dehydrogenase, creatine kinase, pyruvate kinase, and vimentin.

Relatively little attention has been paid to the role of NDP kinase in DNA precursor biosynthesis. One system that has been explored is bacteriophage T4-infected E. coli, in which all the reactions in dNTP and DNA synthesis are catalyzed by phage-coded enzymes, except for the synthesis of dNTPs from the respective NTPs. In this system, the NDP kinase of the host bacterium, encoded by the ndk gene, forms part of a multienzyme complex for dNTP synthesis (21). Affinity chromatography with immobilized E. coli NDP kinase as the affinity ligand reveals specific interactions with several T4 phage proteins, including dihydrofolate reductase, ribonucleotide reductase, and single-strand DNA-binding protein (22). However, NDP kinase is dispensable, because an E. coli strain bearing a targeted deletion of the ndk gene is viable, and it supports phage T4 growth (23). Interestingly, these mutant cells have abnormal nucleoside triphosphate pools—in particular, a twentyfold elevation of the dCTP level—and, probably as a result, an elevated spontaneous mutation rate. Lu and Inouye (24) searched in the ndk deletion strain for an enzyme capable of synthesizing nucleoside triphosphates from the respective diphosphates, and they made the surprising observation that adenylate kinase, the product of the adk gene, possesses this capacity. Because adenylate kinase was previously known only to catalyze the ATP-dependent phosphorylation of AMP or dAMP to the respective diphosphate, this is a novel activity of the adk protein, representing another important but incompletely understood aspect of the biochemistry of nucleoside diphosphate kinase.