Cloning and functional analysis of the ndk1 gene encoding nucleoside-diphosphate kinase in Schizosaccharomyces pombe.

We cloned the ndk1 gene encoding a subunit of nucleoside-diphosphate kinase (NDK) from Schizosaccharomyces pombe, by using polymerase chain reaction. The deduced ndk1 gene product has 151 amino acid residues and is approximately 60% identical with both Saccharomyces cerevisiae and mammalian NDKs. The gene product exhibited NDK activity and cross-reacted with antibodies raised against rat NDK. Disruption of ndk1 greatly reduced the cellular NDK activity but caused no obvious phenotype in cell growth and sexual development of the organism. However, a mutated allele of ndk1 could inhibit sexual development in a dominant-negative manner. This allele carried a point mutation in cysteine 116, which locates next to the putative active center histidine 117, and the mutant gene product showed no NDK activity. Gene expression inducible in response to mating pheromone signaling was decreased in cells carrying the dominant-negative allele. Cases have been reported in higher eukaryotes in which NDK appears to play a more sophisticated role than a simple catalyst in cell physiology, and the results of this study suggest that S. pombe NDK may also perform such a role in regulation of sexual development in the fission yeast.

We cloned the ndk1 gene encoding a subunit of nucleoside-diphosphate kinase (NDK) from Schizosaccharomyces pombe, by using polymerase chain reaction. The deduced ndk1 gene product has 151 amino acid residues and is ϳ60% identical with both Saccharomyces cerevisiae and mammalian NDKs. The gene product exhibited NDK activity and cross-reacted with antibodies raised against rat NDK. Disruption of ndk1 greatly reduced the cellular NDK activity but caused no obvious phenotype in cell growth and sexual development of the organism. However, a mutated allele of ndk1 could inhibit sexual development in a dominant-negative manner. This allele carried a point mutation in cysteine 116, which locates next to the putative active center histidine 117, and the mutant gene product showed no NDK activity. Gene expression inducible in response to mating pheromone signaling was decreased in cells carrying the dominantnegative allele. Cases have been reported in higher eukaryotes in which NDK appears to play a more sophisticated role than a simple catalyst in cell physiology, and the results of this study suggest that S. pombe NDK may also perform such a role in regulation of sexual development in the fission yeast.
Nucleoside-diphosphate kinase (NDK; EC 2.7.4.6) 1 is a ubiquitous enzyme that catalyzes transfer of the terminal phosphate group of ribo-and deoxyribonucleoside 5Ј-triphosphates ((d)NTPs) to nucleoside 5Ј-diphosphates ((d)NDPs) by a pingpong mechanism (1). This enzyme functions as a hexamer, and two kinds of subunits have been identified in higher eukaryotes including human (2). NDK has been believed to be involved in the production of all kinds of nucleoside triphosphates except ATP and in the maintenance of their levels in the cell.
Molecular cloning of cDNAs encoding NDK has been performed in various species (3)(4)(5). Interestingly, recent studies suggest that NDK may have a function other than as a simple enzyme. In Drosophila melanogaster, the abnormal wing discs gene (awd) encodes a NDK homolog (6). The null mutation in awd causes abnormalities in the development of larvae, which result in their death at the prepupal stage (7,8). The nm23 gene encodes NDK in mammals (9). It was originally cloned as a gene that was differentially expressed in two lines of murine K-1735 melanoma cells, which were distinct in their metastatic activities, and it was considered to be a candidate suppressor of tumor metastasis (10 -12). Furthermore, human PuF, a factor required for the transcriptional activation of c-myc in vitro, and mouse I-factor, which inhibits differentiation of murine leukemia M1 cells in the monocyte/macrophage pathway, are highly similar to the product of human nm23-H2 (13,14). NDK has been shown to be able to interact with a GTP-binding protein G s , microtubules, and succinyl-CoA synthetase, and it has been postulated that NDK may regulate them (15)(16)(17). It has been also shown that the cytoplasmic membrane of Dictyostelium possesses NDK activity that is stimulated by a cAMP receptor and that this activity in turn contributes to the activation of G proteins, although precise mechanisms for this regulation remain unknown (18).
The fission yeast Schizosaccharomyces pombe provides a useful model system for the molecular genetic study of intracellular signal transduction. S. pombe haploid cells proliferate mitotically under rich nutritional conditions. They undergo sexual development, namely mating, meiosis, and sporulation, when starved for nitrogen (19,20). Many genes have been shown to be involved in this sexual process. They include the genes for two kinds of heterotrimeric G proteins. One of them mediates the nutritional signal and regulates the cAMP cascade (21). The other, together with the single ras homolog in fission yeast, controls mating pheromone signaling (22)(23)(24). In this report we describe cloning of a fission yeast gene encoding NDK. Molecular genetic analysis of its physiological function suggested that fission yeast NDK may have a regulatory role in the mating pheromone signaling.

EXPERIMENTAL PROCEDURES
Yeast Strains, Media, and Genetic Procedures-S. pombe strains used in this study are listed in Table I. General genetic procedures for S. pombe have been described (25). Yeast complete medium YPD (26) and three kinds of minimal medium, namely SD (26), SSA (27), and PM (28,29), were used. PM-N is a nitrogen-free derivative of PM. SD, PM, and PM-N used in this study contained 1% glucose instead of 2% described in the original recipes. A lithium method was used for transformation of S. pombe (30). A semiquantitative assay of the mating efficiency was done by staining colonies with iodine vapor, which stains spores dark brown (25).
Cloning of ndk1-Two sets of degenerated oligonucleotides, namely 5Ј-GCNAT(T/C/A)AA(G/A)CCNGA(T/C)GGNGT-3Ј and 5Ј-CC(T/C)TC-CCA(A/T)ACCAT(G/A/T)G(G/C)NACNACNGGNCC-3Ј, were prepared as the primers for polymerase chain reaction (PCR). N in the primers stands for a mixture of G/A/T/C. The former set corresponds to the amino acid sequence AIKPDGV and the latter, GPVV(L/A/P)MVWEG (antisense), both of which are highly conserved among NDKs of various species (see Fig. 2). PCR was carried out under the following conditions. A reaction mixture (30 l) contained approximately 50 ng of S. pombe genomic DNA as a template, 100 pmol of each primer, 50 mM KCl, 100 mM Tris-HCl (pH 8.3), 2 mM MgCl 2 , 0.2 mM each dNTP, 0.01% gelatin, 0.4% Triton X-100, and 1 unit of Taq polymerase (Perkin Elmer or Promega). This mixture was subjected to 45 cycles of 93°C 1 min, 45°C 1 min, and 72°C 1 min. After one round of amplification, an aliquot of the mixture was transferred as the template to a fresh reaction mixture devoid of S. pombe DNA, and another round of amplification was performed. The mixture was then electrophoresed in a 4% agarose gel (NuSieve) and was stained with ethidium bromide. DNA fragments were recovered from a band corresponding to the size of 0.2 kb, using Mermaid kit (BIO 101, Inc.). The fragments were cloned into the EcoRV site of pBluescript SK(ϩ) vector (Stratagene). The resulting clones were screened using a part of the human nm23-H1 gene, which we amplified from a HeLa cell cDNA library by PCR (data not shown), as the hybridization probe. We obtained altogether eight positive clones in this screen and sequenced their inserts. Four of them turned out to be identical except for the terminal regions, which appeared to reflect heterogeneity of the original PCR primers. Using one of these inserts as a probe, a 6.0-kb EcoRI fragment that covered the entire ndk1 gene was cloned from the S. pombe genome into pUC119.
Southern Hybridization-Essentially, a standard protocol (31) was followed. Hybridization using a part of the nm23-H1 gene as a probe was done at 42°C overnight in 5 ϫ SSC (0.75 M NaCl, 0.075 M sodium citrate) containing 0.1% N-lauroylsarcosinate, 0.02% SDS, 5% blocking reagent (Boehringer Mannheim), and 20% formamide. Washing was carried out in 2 ϫ SSC containing 0.1% SDS at room temperature for 5 min twice and then at 42°C for 30 min twice. In all other cases, the concentration of formamide in the hybridization solution was raised to 50%, and additional washing was carried out with 0.1 ϫ SSC containing 0.1% SDS at 42°C for 30 min after the same washing as above was done.
DNA Sequence Determination-Nucleotide sequences were determined by the dideoxy chain termination method (32). DNA fragments to be sequenced were cloned into either pUC119 or pBluescript. Subclones for stepwise sequencing were generated by progressive deletion with exonuclease III and S1 nuclease (Takara-shuzo) (33). Single-stranded DNA was prepared by using M13KO7 as a helper bacteriophage. The sequence shown in Fig. 1B was determined in both directions at least once. A primer 5Ј-ATTCATTCGTAAATCCAAGG-3Ј, which is complementary to the carboxyl-terminal region of ndk1, was used to determine the sequence displayed in Fig. 5B.
Disruption of the ndk1 Gene-One-step gene disruption of ndk1 was done essentially as described (34,35). A pUC-based plasmid was constructed by inserting the following three fragments into the multicloning sites of the vector in this order: a 2.0-kb SphI-EcoRV fragment carrying the region 5Ј upstream to the ndk1 ORF, a 1.8-kb S. pombe ura4 ϩ cassette (36), and a 3.0-kb PvuI-EcoRI fragment carrying the 3Ј-terminal region of ndk1 and its downstream vicinity, of which the PvuI end was blunted. The resultant plasmid had a disrupted ndk1 allele, which lacked a 0.6-kb EcoRV-PvuI fragment covering most of the ndk1 ORF and carried the ura4 ϩ cassette in place. An SphI fragment that encompassed this disrupted allele (see Fig. 1A) was introduced into a diploid strain constructed by crossing JY878 with JY879 (h 90 /h 90 ade6-M210/ade6-M216 leu1/leu1 ura4-D18/ura4-D18). Ura ϩ transformants were selected, and genomic DNA was prepared from each of them. DNA preparations thus obtained were digested with EcoRV and examined by Southern hybridization. Diploid strains in which one ndk1 allele was properly replaced were chosen and subjected to tetrad analysis.
ndk1 Plasmids-pREP-ndk1 was constructed by ligating a 1.2-kb SalI-HincII fragment that carried the ndk1 ORF to the vector pREP1 (41), which was digested with SmaI and SalI and treated with calf intestine alkaline phosphatase. The SalI site covered the first through third amino acid residues of Ndk1 (see Fig. 1B), and the exact aminoterminal sequence of Ndk1 was regenerated in pREP-ndk1. pREP-ndk1-dn was constructed by ligating a 0.7-kb SalI-ClaI fragment that carried the ndk1 variant allele to pREP1 by a similar procedure.
Enzymatic Assay of NDK-S. pombe cells (1-3 ϫ 10 8 ) were collected by centrifugation and disrupted in a lysis buffer (20 mM Tris-HCl (pH 7.4), 10 mM MgCl 2 , 1 mM EDTA, 5% glycerol, 1 mM dithiothreitol, 0.3 M (NH 4 ) 2 SO 4 ) containing 1 mM phenylmethanesulfonyl fluoride. Cellular debris was removed by centrifugation in a microcentrifuge. The protein concentration of each preparation was determined by the BCA Protein Assay System (Pierce). Measurements of NDK activity were carried out according to a previous report (42) with a slight modification, as described below. 10 l of each cell lysate diluted appropriately with TEA buffer (100 mM triethanolamine HCl (pH 8.0), 10 mM MgCl 2 , 2 mM ␤-mercaptoethanol) was added to 20 l of TEA buffer containing 12.5 mM [␥-32 P]ATP (2 mCi/mmol) and 10 mM dCDP. The mixture was incubated at 37°C for 15 min and then boiled for 5 min to inactivate the enzymatic activity. 3 l of each sample was spotted onto a polyethyleneimine TLC plate (Macherey-Nagel) and developed in 0.85 M KH 2 PO 4 (pH 3.4). Radioactivity in the resultant dCTP spots was either detected by exposing an x-ray film to them or quantitated by using a Bioimage Analyzer (BAS 2000, FUJIX).
Western Blot Analysis-Crude extracts were prepared from S. pombe wild type cells, ndk1⌬ cells, and cells overexpressing ndk1, and their protein concentration was determined as above. Each sample containing 50 g of protein was separated by SDS-polyacrylamide (14%) gel electrophoresis (43) and then electrotransferred onto a piece of Immobilon-P membrane (Millipore) at 8 V/cm in a transfer buffer (24 mM Tris, 192 mM glycine (pH 8.3), 20% methanol, 0.1% SDS) at 0°C for at least 3 h. The membrane was incubated in TBST buffer (20 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.1% Tween 20) containing 3% skim milk for 1 h at room temperature and washed a few times with TBST buffer. The membrane was then incubated with anti-rat NDK polyclonal antibodies (44) in TBST buffer containing 3% skim milk for 1 h at room temperature. After washing, alkaline phosphatase-conjugated goat anti-rabbit IgG (Sigma), diluted with TBST buffer containing 3% skim milk, was applied as the secondary antibodies. The color reaction was performed in AP buffer (100 mM Tris-HCl (pH 9.5), 100 mM NaCl, 5 mM MgCl 2 ) using 5-bromo-4-chloro-3-indolyl phosphate and nitro blue tetrazolium (Sigma), essentially as described (45).
Hydroxylamine Mutagenesis-Hydroxylamine is a mutagen known to introduce point mutations. We essentially followed a standard protocol (46). 350 mg of hydroxylamine was resolved in 0.56 ml of a freshly prepared 4 N NaOH, which was then diluted with 5 ml of sterile water (Solution A). 0.5 ml of Solution A was mixed with 5 l of 0.2 M EDTA and 5 g of pREP-ndk1. The mixture was incubated at 37°C for 20 h. 10 l of 5 M NaCl, 50 l of 0.1% bovine serum albumin, and 1 ml of cold ethanol were then added, and DNA was recovered by centrifugation.

RESULTS
Cloning of the ndk1 Gene Encoding a Homolog of NDK-To clone a gene encoding a NDK homolog in fission yeast, we carried out the polymerase chain reaction using a pair of primers corresponding to highly conserved amino acid sequences in NDK, as detailed under "Experimental Procedures." We eventually obtained a 0.2-kb DNA fragment that could hybridize with a piece of human nm23-H1 cDNA. Sequence analysis suggested that this fragment carried part of the gene encoding a NDK homolog, which we call ndk1 hereafter. We then cloned a 6.0-kb EcoRI fragment that appeared to contain the entire ndk1 gene, using the 0.2-kb fragment as a probe. Fig. 1A shows a restriction map of the ndk1 locus. Further Southern blot analysis suggested that the ndk1 ORF is located on a 1.2-kb SalI-HincII fragment. We determined the nucleotide sequence of this fragment and its vicinity (Fig. 1B). The ndk1 gene appeared to have a coding capacity of 151 amino acids with no intron. The deduced Ndk1 protein was more than 50% identical with most of NDKs from other species. The highest identity (62.3%) was observed with S. cerevisiae Ynk and the lowest (41.7%), with one from Myxococcus xanthus (Fig. 2). Disruption of ndk1 Results in No Obvious Phenotype-S. pombe diploid strains in which one ndk1 allele was disrupted as schematically shown in Fig. 1A were constructed (see "Experimental Procedures"). Sporulation was induced in these strains, and resultant asci were subjected to tetrad analysis. Four viable spores were recovered in most asci, and Mendelian segregation of the disrupted and normal ndk1 alleles among progeny spores was confirmed by Southern blot analysis (data not shown). Haploid cells carrying the disrupted ndk1 allele appeared to be normal in vegetative cell growth and sexual development.
NDK Activity of the ndk1 Gene Product-A plasmid named pREP-ndk1, which overexpresses ndk1 from the nmt1 promoter (41), was constructed as described under "Experimental Procedures." Crude extracts were prepared from S. pombe strains JY450 (ndk1 ϩ ), JX23 (ndk1⌬), and JX23 transformed with pREP-ndk1. The NDK activity in each extract was as-sayed, and the results are summarized in Fig. 3A. The activity in JX23 was only 20% of that in JY450, whereas the activity in JX23 transformed with pREP-ndk1 was more than 10 times higher than that in JY450. This strongly suggests that the product of ndk1 is responsible for the NDK activity. Furthermore, Western blot analysis of crude extracts prepared from JY450, JX23, and JY450 transformed with pREP-ndk1 demonstrated that the ndk1 gene product cross-reacts with anti-rat NDK antibodies previously described (44) (Fig. 3B). We thus conclude that the ndk1 gene encodes NDK in fission yeast.
Expression of the ndk1 Gene-Expression of ndk1 was examined in S. pombe heterothallic haploid (h ϩ and h Ϫ ), homothallic haploid (h 90 ), and diploid (h ϩ /h Ϫ ) cells by Northern blot analysis. The ndk1 gene was transcribed into a 0.7-kb-long transcript irrespective of the mating type of the cell (Fig. 4). The level of this transcript was increased upon nitrogen starvation, suggesting that ndk1 may have a function required under the starved conditions.
The ndk1-dn Allele Interferes with Sexual Development in the Fission Yeast-We postulated that the ndk1-disrupted strains showed no obvious phenotype because the remaining NDK activity, although very small, was sufficient for cell growth and sexual development. Postulating further that NDK subunits other than Ndk1, which can complex with Ndk1 to form a hexamer, are responsible for this residual activity, we suspected that it might be possible to obtain a dominant-negative (dn) allele of ndk1 that completely diminishes cellular NDK activity. Alternatively, if Ndk1 has any regulatory function in S. pombe cells, it may also be possible to obtain a dn allele of ndk1 that blocks the function of its target molecule. We thus set out to screen for such an ndk1-dn allele. pREP-ndk1, which carried the ndk1 gene, was mutagenized with hydroxylamine, and a homothallic haploid strain JY450 was transformed with the mutagenized plasmid library. As it was suggested that ndk1 may play a significant role under the starved conditions (Fig. 4), we examined 5,000 transformants by iodine staining for their ability to mate and sporulate. Three transformants were found to be poor in this ability. We could recover a plasmid from only one of them. The ndk1 ORF on this plasmid was excised and recloned into pREP1 that was not treated with hydroxylamine. The resultant plasmid (pREP-ndk1-dn) could inhibit mating and sporulation upon transformation (Fig. 5A), confirming that a mutation(s) in the ndk1 gene is responsible for the sterility of the transformant. The mating efficiency of JY450 cells carrying pREP-ndk1-dn was about 20%, while that of the cells carrying the vector was about 70%. Similarly, the sporulation efficiency of JY362 cells carrying pREP-ndk1-dn was about 15%, while that of the control was 45%.
Characterization of the ndk1-dn Mutation and the Mutant Gene Product-We determined the nucleotide sequence of the ndk1-dn allele. One point mutation, which replaces the cysteine residue at position 116 with tyrosine, was found (Fig. 5B). This mutation site was next to the putative active center for the enzymatic activity (His 117 ). To examine NDK activity of the mutant gene product, the ndk1-dn allele was overexpressed from the nmt1 promoter in ndk1⌬ cells. However, the overexpression caused no increase in the NDK activity, clearly indicating that the ndk1-dn gene product does not have the enzymatic activity (Fig. 5C). We tried to determine the level of the residual NDK activity in the ndk1-disrupted cells transformed with pREP-ndk1-dn. Despite repeated trials, we were unable to reach any firm conclusion concerning whether the cells transformed with ndk1-dn have a lower level of NDK activity than the ndk1⌬ cells, because the level was very low in both types of cells and the detected difference was within the range of experimental fluctuation.
The ndk1-dn Allele Decreases the Level of Transcripts Induced by Mating Pheromone Signaling-Effects of ndk1-dn on the expression of genes involved in the regulation of sexual development in fission yeast were examined. Characteristics of the genes we examined are summarized briefly below. The ste11 gene encodes a transcription factor that induces a number of genes required for sexual development, which include mei2, mat1-P, mat1-M, and ste6 (38). Ste11 carries an high mobility group motif and binds to the TR box (TTCTTTGTTY) commonly found in the 5Ј-upstream of these genes. The mei2 gene encodes an RNA-binding protein, which is essential for the initiation of both premeiotic DNA synthesis and meiosis I, and forms a complex with meiRNA to promote meiosis I (47). The mat1-P gene has two transcription units, termed mat1-Pc and mat1-Pi (m), the former of which is required for mating whereas the latter is required for meiosis (39). The sxa2 gene encodes a protease that is supposed to degrade mating pheromone P-factor (40). As shown in Fig. 6, transcripts of mat1-Pi and sxa2 were less abundant in cells transformed with pREP-ndk1-dn than in cells transformed with the vector, while transcripts of ste11 and mei2 were equally abundant. The expression of mat1-Pi and sxa2 calls for both mating pheromone signaling and nutritional signaling (23,40,48), while that of ste11 and mei2 depends only on nutritional signaling (29,38). Thus, our results suggest the possibility that the ndk1-dn allele interferes with the mating pheromone signaling pathway.
Other Point Mutations in ndk1-NDK mutations that affect FIG. 3. Biochemical analysis of the ndk1 gene product. A, ␥-phosphate transfer activity of the ndk1 gene product. Crude extracts were prepared from JY450 (denoted wild type, closed bar), JX23 (disruptant, open bar), and JX23 carrying pREP-ndk1 (disruptant ϩ pREP-ndk1, hatched bar). Given amounts of each sample were assayed for NDK activity. B, Western blot analysis of cell extracts with anti-rat NDK antibodies. Crude extracts were prepared from JY450 (wild type), JX23 (disruptant), and JY450 carrying pREP-ndk1 (wild type ϩ pREP-ndk1). 50 g of protein of each sample was separated in SDS-polyacrylamide gel electrophoresis, transferred to a membrane, and analyzed using anti-rat NDK antibodies. The arrow indicates the ndk1 gene product cross-reacting with the antibodies. The numbers on the left-hand side represent molecular mass markers (kDa).  1, 3,  5, and 7). The remainder was transferred to PM-N medium, incubated for 4 h, and collected (lanes 2, 4, 6, and 8). Total RNA was prepared from these samples. 10 g of each RNA preparation was separated in formaldehyde-agarose gel electrophoresis, blotted onto a membrane, and analyzed with a hybridization probe carrying the ndk1 ORF. rRNAs stained with ethidium bromide are shown in the bottom panel to confirm that nearly the same amount of RNA was loaded in each lane. biological processes have been identified in higher organisms. The Killer of prune mutant of D. melanogaster has a point mutation in the awd gene (49). In human nm23-H1, a deletion and a point mutation were discovered (50,51). We carried out site-directed mutagenesis of ndk1 to obtain analogous mutations. The proline residue at position 95 was substituted with serine, mimicking the mutation in awd K-pn , and the serine residue at position 119 was substituted with glycine, mimicking the mutation in nm23-H1 in neuroblastoma. However, overexpression of these ndk1 derivatives did not cause any specific phenotype (data not shown). Unlike the dominantnegative Ndk1, the mutant products of the fly and human genes have been reported to retain the NDK activity (49,51), which may suggest that the ndk1-dn mutation is either qualitatively different from or quantitatively more effective than these mutations. DISCUSSION In this study we identified the ndk1 gene encoding a NDK homolog in S. pombe. The ndk1 gene product cross-reacted with anti-rat NDK antibodies and apparently possessed the enzymatic activity as NDK. The deduced amino acid sequence of Ndk1 was highly similar to those of NDKs from other species (Fig. 2). Ndk1 showed the highest similarity to S. cerevisiae Ynk, and it was also similar to mammalian NDK nearly as well. The histidine residue known to be phosphorylated when the enzyme assumes a high energy intermediate form was conserved in Ndk1 as His 117 .
The residual NDK activity in ndk1-disrupted cells was less than 20% of the wild type level. However, cells defective in ndk1 showed no obvious phenotype. This indicates that S. pombe cells naturally have NDK activity in large excess, most of which is dispensable for the usual life cycle. The same situation was observed when S. cerevisiae YNK was disrupted (52). What is responsible for the remaining NDK activity, then? It is suggested that (d)NTPs can be produced by the function of pyruvate kinase (53). However, it may be more likely that S. pombe has another gene encoding a NDK homolog, as is the case in mammals, although this has not been proven yet.
As described in the introduction to the text, cases are accumulating in which NDK appears to function not as a simple enzyme. This study presents evidence that NDK may play a role in promotion of sexual development in S. pombe. Introduction of a dominant-negative allele of ndk1, termed ndk1-dn, into wild type cells blocked their sexual development. The   FIG. 6. Expression of genes required for sexual development in ndk1-dn-transformed cells. JY450 cells transformed with either pREP1 or pREP-ndk1-dn were grown to midlog phase in PM medium. A portion of each culture was harvested (0 h), and the remainder was transferred to PM-N medium. Sampling was done at 4, 6, and 8 h after the transfer. Total RNA was extracted from each sample, electrophoresed in formaldehyde-agarose gel, blotted onto a membrane, and analyzed with appropriate hybridization probes for ste11, mei2, matP, and sxa2. rRNAs stained with ethidium bromide are shown in the bottom panel to confirm that nearly the same amount of RNA was loaded in each lane.

FIG. 5. Characterization of the dominant-negative allele of ndk1.
A, JY450 cells transformed with either pREP1, pREP-ndk1, or pREP-ndk1-dn were streaked on SSA medium and incubated for 4 days. Cells were then exposed to iodine vapor, which stains sporulated cells dark brown. B, the ndk1-dn allele was cloned in pBluescript, and the nucleotide sequence was determined. The left panel shows a sequence ladder for the wild type ndk1 gene, and the right panel shows the counterpart for the ndk1-dn allele. The nucleotide and deduced amino acid sequences were shown at the side of each panel. The arrow indicates the position where a mutation was found. C, NDK activity of the ndk1-dn gene product. Crude extracts were prepared from JX23 cells carrying either pREP-ndk1 (ndk1 ϩ ) or pREP-ndk1-dn (ndk1-dn), and NDK activity in each sample was assayed.
induction of mat1-Pi and sxa2 expression, but not that of ste11 and mei2 expression, was inhibited in these cells. This suggests that ndk1-dn may interfere with the mating pheromone pathway, because the former two genes are induced in response to mating pheromone signaling (23,39,40,48), as discussed in more detail under "Results." Although ndk1⌬ strains showed no obvious phenotype, the ndk1-dn allele could inhibit sexual development. One obvious explanation will be that the dominant-negative Ndk1 depletes the residual NDK activity provided by the NDK subunits other than Ndk1, by binding to them. However, we had difficulties in proving this experimentally, and there remains an alternative possibility that Ndk1-dn may sequester its target(s). This target will be essential for sexual development and will have a direct physical contact with the enzyme. In particular, an interesting possibility will be that S. pombe NDK regulates the activity of either Ras1 or Gpa1, both of which are GTP-binding proteins and are essential for the mating pheromone signaling in S. pombe (22,23,35). Further analyses of Ndk1 and related proteins are required to clarify how NDK controls mating pheromone signal transduction in S. pombe.