Fission Yeast Homolog of Neuronal Calcium Sensor-1 (Ncs1p) Regulates Sporulation and Confers Calcium Tolerance*

The neuronal calcium sensor (NCS) proteins (e.g. recoverin, neurocalcins, and frequenin) are expressed at highest levels in excitable cells, and some of them regulate desensitization of G protein-coupled receptors. Here we present NMR analysis and genetic functional studies of an NCS homolog in fission yeast (Ncs1p). Ncs1p binds three Ca2+ ions at saturation with an apparent affinity of 2 μm and Hill coefficient of 1.9. Analysis of NMR and fluorescence spectra of Ncs1p revealed significant Ca2+-induced protein conformational changes indicative of a Ca2+-myristoyl switch. The amino-terminal myristoyl group is sequestered inside a hydrophobic cavity of the Ca2+-free protein and becomes solvent-exposed in the Ca2+-bound protein. Subcellular fractionation experiments showed that myristoylation and Ca2+ binding by Ncs1p are essential for its translocation from cytoplasm to membranes. The ncs1 deletion mutant (ncs1Δ) showed two distinct phenotypes: nutrition-insensitive sexual development and a growth defect at high levels of extracellular Ca2+ (0.1 m CaCl2). Analysis of Ncs1p mutants lacking myristoylation (Ncs1pG2A) or deficient in Ca2+ binding (Ncs1pE84Q/E120Q/E168Q) revealed that Ca2+ binding was essential for both phenotypes, while myristoylation was less critical. Exogenous cAMP, a key regulator for sexual development, suppressed conjugation and sporulation of ncs1Δ, suggesting involvement of Ncs1p in the adenylate cyclase pathway turned on by the glucose-sensing G protein-coupled receptor Git3p. Starvation-independent sexual development of ncs1Δ was also complemented by retinal recoverin, which controls Ca2+-regulated desensitization of rhodopsin. In contrast, the Ca2+ intolerance of ncs1Δ was not affected by cAMP or recoverin, suggesting that the two ncs1Δ phenotypes are mechanistically independent. We propose that Schizosaccharomyces pombe Ncs1p negatively regulates sporulation perhaps by controlling Ca2+-dependent desensitization of Git3p.

The neuronal calcium sensor (NCS) proteins (e.g. recoverin, neurocalcins, and frequenin) are expressed at highest levels in excitable cells, and some of them regulate desensitization of G protein-coupled receptors. Here we present NMR analysis and genetic functional studies of an NCS homolog in fission yeast (Ncs1p). Ncs1p binds three Ca 2؉ ions at saturation with an apparent affinity of 2 M and Hill coefficient of 1.9. Analysis of NMR and fluorescence spectra of Ncs1p revealed significant Ca 2؉ -induced protein conformational changes indicative of a Ca 2؉ -myristoyl switch. The amino-terminal myristoyl group is sequestered inside a hydrophobic cavity of the Ca 2؉ -free protein and becomes solvent-exposed in the Ca 2؉bound protein. Subcellular fractionation experiments showed that myristoylation and Ca 2؉ binding by Ncs1p are essential for its translocation from cytoplasm to membranes. The ncs1 deletion mutant (ncs1⌬) showed two distinct phenotypes: nutrition-insensitive sexual development and a growth defect at high levels of extracellular Ca 2؉ (0.1 M CaCl 2 ). Analysis of Ncs1p mutants lacking myristoylation (Ncs1p G2A ) or deficient in Ca 2؉ binding (Ncs1p E84Q/E120Q/E168Q ) revealed that Ca 2؉ binding was essential for both phenotypes, while myristoylation was less critical. Exogenous cAMP, a key regulator for sexual development, suppressed conjugation and sporulation of ncs1⌬, suggesting involvement of Ncs1p in the adenylate cyclase pathway turned on by the glucose-sensing G protein-coupled receptor Git3p. Starvation-independent sexual development of ncs1⌬ was also complemented by retinal recoverin, which controls Ca 2؉ -regulated desensitization of rhodopsin. In contrast, the Ca 2؉ intolerance of ncs1⌬ was not affected by cAMP or recoverin, suggesting that the two ncs1⌬ phenotypes are mechanistically independent. We propose that Schizosaccharomyces pombe Ncs1p negatively regulates sporulation perhaps by controlling Ca 2؉ -dependent desensitization of Git3p.
The primary sequence of S. pombe Ncs1p demonstrates its homology to the NCS branch of the EF-hand superfamily of Ca 2ϩ -binding proteins ( Fig. 1) (20,21). Recoverin, the most intensively studied NCS protein, serves as a Ca 2ϩ sensor in retinal rod and cone cells where it controls the desensitization of rhodopsin by inhibiting rhodopsin kinase only at high Ca 2ϩ levels (22)(23)(24)(25). The NCS family also includes neuronal Ca 2ϩ sensors such as neurocalcin (26), hippocalcin (27), and frequenin (28) as well as the budding yeast homolog Frq1 (29). All members of the NCS family are myristoylated and possess four EF-hands, although the first EF-hand motif (EF-1) contains substitutions that prevent Ca 2ϩ binding at this site ( Fig. 1). In recoverin, substitutions in EF-4 also prevent Ca 2ϩ binding, and hence only EF-2 and EF-3 are functional (30). Frq1, frequenin, and neurocalcin contain three occupied Ca 2ϩ -binding sites (EF-2, EF-3, and EF-4) in their atomic resolution structures (31)(32)(33). The S. pombe Ncs1p shares highest sequence identity with Frq1 and frequenin (60%) but is also highly similar in sequence to mammalian recoverin (46% identity). Interestingly structurally important amino acids in recoverin appear invariant in S. pombe Ncs1p but are not conserved in Saccharomyces cerevisiae Frq1 (see highlighted residues in Fig. 1). Moreover recoverin is significantly closer in sequence to Ncs1p (46% identity) than to Frq1 (35% identity). Hence recoverin appears to be structurally and evolutionarily more similar to S. pombe Ncs1p than it is to Frq1 of budding yeast.
The three-dimensional structures of myristoylated recoverin have been determined by NMR spectroscopy (30,34). A striking feature of these structures is a large Ca 2ϩ -induced conformational change. In the Ca 2ϩ -free state, the myristoyl group is buried inside the protein and not exposed to solvent. Binding of Ca 2ϩ to recoverin leads to extrusion of its myristoyl group and to a large rotation of the two domains of the protein such that many hydrophobic residues are exposed. The Ca 2ϩ -induced exposure of the myristoyl group, termed the calcium-myristoyl switch, enables recoverin to bind to membranes only at high Ca 2ϩ levels (35,36).
Given the similarities of Ncs1p to the NCS family, it was of interest to assess whether this ancestral homolog displays a Ca 2ϩ -dependent myristoyl switch and to characterize its physiological role in fission yeast. Here we combine NMR spectroscopy, biochemical analysis, and yeast genetics to determine the Ca 2ϩ and membrane binding properties, Ca 2ϩ -induced protein structural changes, and biological function of Ncs1p.

EXPERIMENTAL PROCEDURES
Strains and Media-S. pombe strains constructed in this study were derived from either wild type KGY425, KGY461, KGY553, KGY554, or SP870 (Table I). Standard culture media and genetic manipulation techniques were used (37). For vegetative growth of S. pombe, rich medium (YES) and synthetic minimum medium (EMM) were used. Malt extract solid medium was used for conjugation and sporulation. Nutritional supplements were added as needed. To maintain strains bearing S. pombe plasmid with LEU2 marker and nmt1 promoter, leucine was omitted from the supplements in EMM (EMM-L) and 2 M thiamine was added. To see starvation-independent spore formation, EMM containing a final concentration of 3% glucose, a concentration equivalent to that in rich medium, was used. For the calcium sensitivity assay, CaCl 2 was added to YES at a final concentration of 0.1 M before autoclaving. The deletion mutant strain, bearing a plasmid based on S. pombe expression vector pREP1 (LEU2 ϩ nmt1 promoter), was grown on EMM-L minus thiamine to induce the nmt1 promoter before being streaked on YES containing 0.1 M CaCl 2 .
Ploidy was confirmed by the color of colonies on an agar medium containing 0.0005% Phloxine B (yeast extract with Phloxine B), by the color of colonies on agar medium containing a low concentration of adenine (YES with no adenine added), and by the cell size and presence of azygotic sporulation. Routine microscopic observation was performed using an Olympus BX51 fluorescence microscope equipped with a charge-coupled device camera (SPOT-RT Slider, Diagnostic Instruments). Live cells were stained with 4Ј,6-diamidino-2-phenylindole by mixing equal volumes of cell suspension and 50 g/ml 4Ј,6-diamidino-2-phenylindole solution. Live cells were also examined by confocal microscopy using a TCS-SP scanner with a DMRE fluorescence microscope (Leica, Heidelberg, Germany) equipped with an Ar ϩ laser source. Excitation for green fluorescent protein (GFP) was 488 nm and emission was 500 -575 nm. Twenty to 30 slices of optical sectioning were made by serial scanning, and images were stored for further three-dimensional reconstruction. Three-dimensional animation movies were generated from these data (see Supplemental Material).
Cloning of ncs1 and Plasmid Construction-Standard protocols were used (38), and Escherichia coli DH5␣ was used as a host for all plasmid manipulations. Nucleotide sequences were determined after gene cloning and each mutagenesis by Dye Terminator Cycle Sequencing using an ABI 373 or 3100 sequencer (PerkinElmer Life Sciences).
The ncs1 gene was cloned by reverse transcription-PCR. First, total RNA was extracted from S. pombe canonical wild type strain 972 by TRIzol reagent (Invitrogen) followed by reverse transcription. The resultant cDNA mixture served as a template in the PCR, and upstream and downstream primers 5Ј-AAGGTACCGT CGACATATGG GAAAAT-CACA ATCAAAATTG TCTCAAGATC AACTCC-3Ј and 5Ј-AATCTA-GAGG ATCCTTCATT ATATTAAAAA ATGTAATAAA AGGCAAAATG TAC-3Ј, respectively, were used. An ϳ0.8-kb PCR product was first cloned into pBlueScript KSII(ϩ) (Stratagene) at KpnI/XbaI sites yielding pBSII-ncs1. The nucleotide sequence of the ncs1 open reading frame was confirmed, and this plasmid was used as a template for all the mutagenesis. Triple EQ mutations (3EQ; E84Q/E120Q/E168Q) that abolished Ca 2ϩ binding at EF-2, EF-3, and EF-4 were created by QuikChange site-directed mutagenesis kit (Stratagene). The G2A mutation was performed by standard PCR using Pfu polymerase (Stratagene).
The multicopy S. pombe expression vector pREP1 2 was used to express ncs1, ncs1 G2A , ncs1 3EQ , bovine recoverin cDNA (Reco), or S. cerevisiae FRQ1 gene. To construct these plasmids, ncs1 and its mutant genes as well as Reco and FRQ1 in pTIS1DT (39) were all subcloned into pREP1 at either the NdeI/XmaI site or NdeI/BamHI site.
To construct fusion proteins of Ncs1p and its unmyristoylated mutant with enhanced green fluorescence protein (EGFP) (F64L/S65T; mut1 of Cormack et al. (40)), the ncs1 and ncs1 G2A genes were modified at their stop codons and ligated into pREP42EGFP-C 2 at the NdeI/ XmaI site. For protein expression in E. coli, the ncs1 open reading frame from pBSII-ncs1 was subcloned into pPET11a (Novagen) at the NdeI/ BamHI sites.
Gene Disruption of ncs1 Allele-Deletion mutant allele ncs1⌬::his3 was produced as shown in Fig. 6 by one-step gene disruption. The coding region of ncs1 was replaced with a his3 gene cassette through homologous recombination in wild type diploid NHSP001D. The his3 gene cassette was made by PCR using pAF1 2 as a template and 5Ј-G-GTTTTTTAA AGACTGCCCT TCTGGCCATT TGAATAAGTC TGAAT-TTCAG AAAATCTATA AACAATTTTT CCCATTCGGT GATCCCTTTC AACGTTTTC TTTACT-3Ј and 5Ј-AACCAATCCA TCGTATAAAG AAA-GGGCTGA CACGATAGTC GGATCACGTT TGGACCCTTC GCAGAA-TTCC TCAAGTGTCA GTTGCTCTAT GCAAAGCTAA CGAAT-3Ј as upstream and downstream primers, respectively. The His ϩ haploid NHP013 was obtained from resultant heterozygous diploid NHP009D by random sporulation (41). The isogenic strains NHP57-60 were ob-2 Plasmid pREP42EGFP-C was a kind gift from Dr. I. M. Hagan. Plasmid vector pREP1 was a gift from Dr. K. Maundrell, pCS2pkSu was a gift from Dr. J. R. McIntosh, and pAF1 was a gift from Dr. K. L. Gould.  (Table I). Homothallic deletion mutant NHP033 was obtained using the same cassette, and SP870 was used as a host strain.
The chromosomal ncs1-GFP-tagged allele (ncs1 ϩ -GFP::ura4 ϩ ) was constructed by the same method using GFP::ura4 ϩ cassette delivered from pCS2pkSu 2 and KGY425 as a host strain. Resultant NHP002 was then back-crossed with KGY554 to obtain NHP078 (Table I). Ncs1p-GFP was confirmed to be as functional in nutrition-dependent sporulation as wild type.
Preparation and Purification of Recombinant Myristoylated Ncs1p-To prepare recombinant myristoylated Ncs1p protein uniformly labeled with nitrogen-15 and/or [ 13 C 14 ]myristate, the protein was expressed in E. coli strain BL21(DE3) (Novagen) carrying the pET11a vector harboring the ncs1 coding sequence and co-expressing a yeast N-myristoyltransferase (42) grown in M9 minimal medium containing [ 15 N]NH 4 Cl according to well established procedures (43,44). Unlabeled or [ 13 C 14 ]myristic acid (5 mg/liter) was added to the medium 0.5 h before induction of protein expression. Labeled myristoylated Ncs1p protein was purified from the soluble fraction of a bacterial cell lysate using hydrophobic interaction chromatography as described previously for recoverin and frequenin (35,45). Peak fractions were concentrated to 5 ml and subjected to size exclusion chromatography (Sephacryl S-100, Amersham Biosciences) in buffer B (1 mM dithiothreitol, 2 mM CaCl 2 , 50 mM HEPES, pH 7.4). Final purity was greater than 98% as judged by SDS-PAGE (see Supplemental Data). 3 Electrosprayionization mass spectrometry indicated that at least 85% of the recombinant protein was properly myristoylated.
Binding of 45 Ca 2ϩ -45 Ca 2ϩ radioactive isotope (calcium-45, calcium chloride in aqueous solution, specific activity ϭ 850 mCi/ml, Amersham Biosciences) was used to quantitate the binding of Ca 2ϩ to Ncs1p. 45 Ca 2ϩ binding to Ncs1p was measured as the protein-bound radioactivity retained after ultrafiltration using a procedure described previously (47) based on the original method of Paulus (48). The buffer used in the Ca 2ϩ titration (25 mM HEPES, 0.1 M KCl, 1 mM dithiothreitol, pH 7.4) and protein samples were decalcified by treatment with Chelex resin (Bio-Rad) using the batch method. A Centricon-10 concentrator (10-kDa cutoff, 2-ml sample compartment, Millipore Corp.) used in the titration was pretreated to remove contaminating Ca 2ϩ . The lower chamber was rinsed with 0.1 M HCl followed by several rinses with decalcified buffer. The concentrator membrane was decalcified by rinsing with 5% NaHCO 3 followed by several rinses with decalcified buffer. A decalcified protein sample (1.5 ml, 70 M) was placed into the sample compartment, and 12 l of 0.25 mM 45 Ca 2ϩ solution (2.6 Ci) was added. The sample was carefully mixed and centrifuged (2300 rpm, 2 min) using a tabletop centrifuge (Beckman Model TJ-6), forming 25 l of filtrate. The filtrate was returned to the sample chamber, mixed, and centrifuged a second time to minimize any dead volume. The radioactivity of 10 l of the filtrate (free Ca 2ϩ ) and an equal volume of the protein sample (total Ca 2ϩ ) were determined by liquid scintillation counting (Liquid Scintillation Analyzer, Packard Instrument Co.). Aliquots of nonradioactive Ca 2ϩ were added serially to the protein sample to adjust the total Ca 2ϩ concentration throughout the titration (5.5, 15.7, 28.7, 52.5, 78.7, 131, 180, 240, 300, 400, 500, and 600 M), and the above centrifugation procedure was repeated in triplicate for each point in the titration. The free and bound Ca 2ϩ concentrations at each point in the titration were calculated from the measured radioactivity as described previously (29), and fractional saturation was plotted as a function of free Ca 2ϩ concentration (see Table II and Fig. 2).
Subcellular Fractionation and Western Blot Analyses-S. pombe NHP057 (ncs1⌬) bearing pREP42-ncs1-EGFP or pREP42-ncs1 G2A -EGFP were grown for 20 h after induction of nmt1 promoter (by decreasing the thiamine concentration in the media from 2 M down to 0.05 M) to midlog phase and rinsed in EB buffer (15 mM KCl, 10 mM HEPES-KOH, pH 7.8, 3 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride) containing either 1 mM CaCl 2 or 10 mM each of EDTA and EGTA. Cells were then lysed by glass beads in the same buffer containing either 0.2 mM CaCl 2 or chelating reagents (5 mM each of EDTA and EGTA), and the cell-free extract (lysate) was obtained after centrifugation at ϳ500 ϫ g for 5 min at 4°C. Protein concentration of the lysate was determined, and each lysate was adjusted with its respective buffer to the same final concentration. Part of these cell-free extracts, 100 l each, was subjected to centrifugation for 30 min at ϳ16,000 ϫ g resulting in the pellet and supernatant fractions. The pellet was resuspended with an equal volume of the appropriate lysis buffer. The lysate, supernatant fractions, and resuspended pellets were subjected to SDS-PAGE (15% acrylamide) followed by Western blot analysis with anti-GFP antibody (mouse polyclonal antibody, Clontech). Signals were detected using chemiluminescent substrate (Super Signal West Dura, Pierce). In the case of experiments in Fig. 9C (right panel), cells were lysed in EB buffer without any additional CaCl 2 or chelating reagents, and whole lysate was loaded on the gel.
Assay for Mating, Sporulation, and Germination Efficiency-To score the "color test" for diploid stability shown in Fig. 7A, pairs of fresh haploid strains from NHP057-60 were mixed on a mating/sporulation plate (malt extract), and diploid cells from each pair were isolated on EMM lacking adenine (EMM-A) containing 3% glucose. More than four colonies from EMM-A were periodically streaked on a YES ϩ low adenine (no adenine added) plate. After 2 or more days, white colonies (representing diploid cells) and pink colonies (representing haploid cells) appeared. Stability of diploid cells was quantified as the percentage of white colonies counted as a function of time. Two independent experiments were performed. Experiment 1 quantified the percentage of white colonies from cells grown on a low adenine plate for 2 days, and Experiment 2 quantified from cells grown for 4 days.
The measurement of starvation-independent conjugation/spore formation was also performed by observation of morphological change using the homothallic strain NHP033 (ncs1⌬) bearing various expression plasmids: empty vector pREP1, pREP1-ncs1, pREP1-ncs1 G2A , pREP1-ncs1 E84Q/E120Q/E168Q , pREP1-recoverin, or pREP1-FRQ1 (Table  III). Homothallic (h 90 ) strains change their mating type periodically and start a series of sexual development upon change of nutritional condition. In this case, sexual development was triggered by temperature shift because ncs1 mutant was less sensitive to nutritional starvation. Cells from a single colony isolated from a minimal medium plate (EMM) lacking leucine and containing 3% glucose and thiamine were grown in liquid media (EMM-L) at 37°C, to avoid sporulation, until A 600 reached between 1-2. Cells were resuspended in the same media giving an A 600 of 1.0 and incubated at 25°C without shaking. Alternatively a drop of cell suspension was placed on an EMM-L plate and incubated at 25°C. Morphology was observed periodically using a microscope. Individual cells and asci were identified and counted. Typically 600 cells were counted.

RESULTS
Cloning of ncs1-Primers for reverse transcription-PCR of S. pombe gene (SPAC18B11.04) were designed according to the sequence of chromosome I cosmid c18B11 (GenBank TM accession number Z50728), and a cDNA (819 base pairs) was cloned from canonical wild type fission yeast (the same as a strain used in genome sequencing) as described under "Experimental Procedures." The open reading frame consists of four exons and encodes a protein of 190 amino acids that is highly homologous to NCS proteins (17). Hence this gene has been named ncs1 (the nucleotide sequence of the ncs1 cDNA has been deposited into GenBank TM under accession number AY225216).
Equilibrium Ca 2ϩ Binding Measurements-Ncs1p contains four EF-hand Ca 2ϩ -binding motifs (Fig. 1). The first EF-hand (EF-1) of Ncs1p contains substitutions (Lys 36 , Cys 38 , and Pro 39 ) that would be expected to disrupt the structure of this binding loop and prevent Ca 2ϩ binding to EF-1 as seen in the crystal structures of recoverin, frequenin, and neurocalcin (32,33,49). In recoverin, the disabling of Ca 2ϩ binding to EF-1 seems to be important for its calcium-myristoyl switch mechanism because residues of EF-1 in Ca 2ϩ -free recoverin make physical contact with the myristoyl group (30). Likewise the presence of residues that potentially disable Ca 2ϩ binding to EF-1 suggests that Ncs1p may possess a similar calcium-myristoyl switch mechanism. In contrast to recoverin, in which EF-4 is also disabled by substitutions incompatible with Ca 2ϩ binding, the remaining motifs in Ncs1p (EF-2, EF-3, and EF-4) are good matches to the consensus and are expected to bind Ca 2ϩ . To quantitate the number of ions that bind at saturation, direct measurements of Ca 2ϩ binding were performed on samples of recombinant, myristoylated Ncs1p purified from bacteria.
Equilibrium measurements of 45 Ca 2ϩ binding were conducted on Ncs1p ( Fig. 2 and Table II). At saturation, three Ca 2ϩ bind to the protein, consistent with the view that EF-1 is disabled, whereas EF-2, EF-3, and EF-4 are functional. The fractional saturation (Y), which can be obtained from the same data, can be represented by the Hill equation, where [Ca 2ϩ ] is the free Ca 2ϩ concentration, K d is the apparent dissociation constant, and a denotes the Hill coefficient. The Ca 2ϩ binding isotherm for Ncs1p (Fig. 2) is best fit by the Hill equation using the parameters, K d ϭ 2.1 M and a ϭ 1.9. A Hill coefficient of nearly 2 suggests that two or more Ca 2ϩ bind cooperatively to the myristoylated protein. The positive cooperativity of Ca 2ϩ binding suggests that Ncs1p undergoes a concerted protein conformational change like the Ca 2ϩ -myristoyl switch of recoverin (50).
Structural Studies Using NMR-The above analysis of Ca 2ϩ binding to Ncs1p suggests two allosteric states of the protein.
To test whether Ncs1p actually undergoes Ca 2ϩ -induced structural transitions, we collected two-dimensional NMR spectra ( 1 H-15 N HSQC) of uniformly 15 N-labeled Ca 2ϩ -free and Ca 2ϩbound forms of myristoylated Ncs1p (Fig. 3). Peaks in each spectrum represent main chain and side chain amide protons that serve as fingerprints of overall conformation. The NMR spectrum of Ca 2ϩ -free Ncs1p exhibited many sharp and highly resolved peaks (Fig. 3A). The number of observed peaks (215) was, as expected, very close to the total number of amide protons in the protein (190 main chain ϩ 32 side chain ϭ 222 amide protons). The sharpness of the peaks and uniform peak intensities indicated that Ca 2ϩ -free, myristoylated Ncs1p adopts a stable three-dimensional fold and is monomeric in solution, similar to the NMR spectrum and structure of Ca 2ϩfree recoverin that is highly stabilized by a sequestered myristoyl group (34). Our preliminary structural analysis of Ncs1p suggests that the amino-terminal myristoyl group is also sequestered inside a hydrophobic cavity located primarily in the amino-terminal domain of the protein (see below).
In contrast to Ca 2ϩ -free Ncs1p, the NMR spectrum of Ca 2ϩbound Ncs1p exhibits very broad and poorly resolved peaks (Fig. 3B). The striking Ca 2ϩ -induced spectral differences are consistent with large Ca 2ϩ -induced structural changes in the protein. The intensity of peaks characteristic of the Ca 2ϩbound form saturated upon the addition of 3 molar eq of Ca 2ϩ to the sample, in good agreement with the stoichiometry of Ca 2ϩ binding determined from the equilibrium Ca 2ϩ binding experiments (Fig. 2). The number of observed peaks in the  were not detected apparently because they were severely broad and weak. The variable range of peak intensities suggested that some of the peaks might be broadened due to self-association of Ncs1p molecules in concentrated solution (1 mM protein concentration) required for NMR. Indeed dynamic light scattering measurements performed on the solutions used for NMR confirmed that the Ca 2ϩ -bound Ncs1p samples contained a broad distribution of multimeric species with an average molecular mass of ϳ72 kDa (data not shown). The Ca 2ϩ -induced protein aggregation observed here for Ncs1p might be caused by a solvent-exposed myristoyl group like that of Ca 2ϩ -bound recoverin (30,46). NMR Analysis of [ 13 C]Myristate-labeled Ncs1p-NMR analysis of fatty acyl chain resonances provides a sensitive probe of the disposition of the covalently attached myristoyl group and its interaction with the protein. Previously we developed and performed three-dimensional ( 13 C-filtered NOESY-HMQC) NMR experiments on samples of recoverin that contained a 13 C-labeled myristoyl group to selectively probe the chemical environment around the amino-terminal myristoyl group (46,47,51). These studies revealed that the covalently attached fatty acyl chain in recoverin is sequestered in a hydrophobic cavity in the Ca 2ϩ -free protein and that binding of Ca 2ϩ leads to conformational changes that extrude the myristoyl group into solvent. In contrast, similar NMR experiments performed on Frq1 containing a 13 C-labeled myristoyl group revealed that the fatty acyl chain is solvent-exposed in both the Ca 2ϩ -free and Ca 2ϩ -bound forms (31). This difference in the disposition of the myristoyl group between Frq1 and Ca 2ϩ -free recoverin might be due to changes in critical residues required for the Ca 2ϩ -myristoyl switch (Fig. 1).
Three-dimensional 13 C(F 1 )-edited and 13 C(F 3 )-filtered NOESY experiments (46,52) were performed on unlabeled Ncs1p protein containing a 13 C-labeled myristate (Fig. 4). These spectra selectively probed atoms of residues in the protein that lie within 5 Å of the labeled C-14 methyl group of the myristoyl chain. If the myristoyl group is sequestered within the protein, then the terminal methyl group of the fatty acyl chain is expected to be very close to atoms of the protein, resulting in strong off-diagonal cross-peaks in the filtered NOESY spectrum. The filtered NOESY spectrum of Ca 2ϩ -free Ncs1p (Fig. 4A) exhibits many dipolar interactions between the myristate C-14 methyl group (F 2 ϭ 16.9 ppm, F 1 ϭ 0.9 ppm) and the protein. Strong off-diagonal cross-peaks near 1.0 and 7.0 ppm represent close contacts with aliphatic groups and aromatic side chains, respectively. Therefore, the methyl group of myristate interacts intimately with the Ca 2ϩ -free Ncs1p protein and appears buried in a hydrophobic cavity akin to that of Ca 2ϩ -free recoverin (34) and in contrast to the solventexposed myristoyl group of Ca 2ϩ -free Frq1 (31).
On the contrary, no off-diagonal cross-peaks were observed in the filtered NOESY spectrum of Ca 2ϩ -bound Ncs1p (Fig.  4B), suggesting that the C-14 methyl group of the fatty acyl chain does not interact closely with the Ca 2ϩ -bound Ncs1p protein. Hence the myristoyl group of Ca 2ϩ -bound Ncs1p appears solvent-exposed and undergoes a calcium-induced change in environment as expected for a Ca 2ϩ -myristoyl switch. The Ca 2ϩ -induced exposure of the myristoyl group might serve to anchor the Ncs1p protein to cell membranes only at high Ca 2ϩ levels.
Subcellular Fractionation of Ncs1p-To assess whether Ncs1p associates with cell membranes, we examined the distribution of myristoylated and unmyristoylated Ncs1p in the soluble and particulate fractions of whole-cell extracts prepared in the presence and absence of Ca 2ϩ (Fig. 5). For this purpose, either myristoylated Ncs1p or an unmyristoylated mutant (Ncs1p G2A ) was expressed as a GFP fusion protein from plasmids in fission yeast that lacked endogenous protein (ncs1⌬ mutant, see below). The resulting extracts were separated into an insoluble fraction (pellet) consisting mainly of plasma membranes and soluble fraction (supernatant) containing most of the cytosolic components (53). When the cell lysate was treated with saturating calcium, about half of the myristoylated Ncs1p was observed in the pellet (Fig. 5A, lane 2), indicating that a substantial portion of the Ca 2ϩ -bound, myristoylated protein was associated with membrane components. In sharp contrast, when the lysate was treated with calciumchelating agents (EGTA and EDTA), nearly all of the Ca 2ϩfree, myristoylated protein appeared in the soluble fraction, whereas almost none of it could be found in the pellet fraction (Fig. 5A, lane 5). The same results were obtained when Ncs1p was expressed under control of its endogenous promoter in chromosome (Fig. 5D). Ca 2ϩ -induced membrane binding by the Ncs1p-GFP fusion protein was not due to the attached GFP because GFP itself expressed under the same condition was observed only in the supernatant fraction regardless of calcium level (Fig. 5C, lanes 2 and 3 and lanes 5 and 6). The subcellular fractionation of the unmyristoylated Ncs1p G2A mutant is presented in Fig. 5B. The unmyristoylated protein was observed in both the soluble and insoluble fractions. The cellular distribution of unmyristoylated Ncs1p was found to be independent of Ca 2ϩ level (Fig. 5B, lanes 2 and 3  and lanes 5 and 6) in contrast to the Ca 2ϩ -induced membrane binding by myristoylated Ncs1p (Fig. 5A). Taken together, these results indicate that the myristoyl group is essential for Ca 2ϩ -induced membrane translocation of Ncs1p, and the Ca 2ϩinduced exposure of the myristoyl group (Fig. 4) may help to anchor the protein to membranes only at high Ca 2ϩ levels.
Deletion Mutant of ncs1 Exhibits Starvation-independent Conjugation and Sporulation-The ncs1 open reading frame was disrupted by inserting the his3 marker gene (1.9 kb), deleting the entire EF-2 and EF-3 coding region (Fig. 6). The deletion allele (ncs1⌬::his3) was confirmed by Southern blot analysis. A tetrad dissection of heterozygous diploid cells (ncs1 ϩ /ncs1⌬::his3) produced haploid deletion mutants that were viable with a normal vegetative growth rate. This is a clear difference from the lethal phenotype of deletion mutant of S. cerevisiae homolog, Frq1. The haploid ncs1⌬ strains did not show any temperature sensitivity at 15°C or 37°C and exhibited normal morphology.
During our attempt to isolate ncs1⌬ haploid cells by conventional tetrad dissection, we found that the heterozygous diploid cells (ncs1 ϩ /ncs1⌬::his3) were very unstable and sporulated and germinated in nutrient-rich medium. More than half of the ncs1⌬::his3 diploid cells grown on EMM-A containing 3% glucose spontaneously converted into haploid cells within a few days, whereas wild type diploid cells remained much more stable during the same period under the same conditions. The relative instability of ncs1⌬ diploid cells was quantified using a simple color test as illustrated in Fig. 7A and described under "Experimental Procedures." The percentage of white colonies (representing diploid cells) was much greater for wild type (54%) compared with that of heterozygous (8 and 9%) and homozygous (11%) strains of ncs1⌬::his3 (Fig. 7A). This phenotype was seen in both heterozygous and homozygous diploid cells and was therefore haploid-insufficient. Later we noticed that freshly prepared heterozygous diploid cells could be isolated temporarily but exhibited diploid instability after only a few days, hereafter referred to as starvation-independent sporulation.
Since sexual development in S. pombe involves germination that subsequently follows conjugation and sporulation, the observed diploid instability in ncs1⌬ could be due to abnormal germination. The germination efficiency of ncs1⌬ mutant measured by random sporulation method was shown to be similar to that of homozygous wild type (data not shown). Hence the ncs1 deletion does not affect germination efficiency.
The homothallic deletion mutant NHP033 (h 90 ncs1⌬) enabled synchronous observation of sporulation/conjugation events and was used to more quantitatively evaluate conjugation and sporulation efficiencies. The percentage of homothallic diploid cells that underwent sporulation in nutrient-rich medium after 24 h was significantly greater for ncs1⌬::his3 (13% zygotic asci) compared with that of wild type (Ͻ1%) (Fig. 7B). Similarly the percentage of conjugation in nutrient-rich medium after 24 h was also greater for ncs1⌬::his3 (9% zygotes) compared with that of wild type (4%) (Fig. 7B). Hence the ncs1 deletion mutant exhibits a phenotype similar to starvationindependent sporulation seen previously in the deletion mutants of git3 (glucose receptor) and gpa2 (␣-subunit of G protein) (54). In S. pombe, sexual development is regulated negatively by a nutritional signal (e.g. glucose or nitrogen) and positively by pheromone. Under normal growth conditions (non-starvation), the nutritional signal represses sexual development through sequential activation of the glucose receptor (git3), G protein ␣-subunit (gpa2), adenylate cyclase (cyr1), and protein kinase A (pka1) in a cascade known as the "adenylate cyclase pathway." Under nutrition starvation, the adenylate cyclase pathway turns off, causing a drop in the cellular cAMP level, which then triggers conjugation/spore formation (55,56). Sporulation is important for the preservation of species because spores are more durable than vegetative cells and more resistant to environmental stresses.
Starvation-independent conjugation and sporulation of the ncs1⌬ mutant was effectively blocked by the addition of exogenous cAMP, suggesting possible involvement of the adenylate cyclase pathway. When cells were incubated in EMM with 10 mM cAMP for more than 24 h, both wild type and ncs1⌬ mutant cells did not conjugate or sporulate: no zygote formation was FIG. 4. Ca 2؉ -induced extrusion of the amino-terminal myristoyl group. Three-dimensional 13 C(F 1 )-edited, 13 C(F 3 )-filtered NOESY-HMQC NMR spectra of Ca 2ϩ -free (A) and Ca 2ϩ -bound (B) Ncs1p. The attached myristoyl group was labeled with carbon-13, and the protein was unlabeled. Selected slices at F 2 ϭ 16.9 ppm are shown to specifically probe the C-14 methyl group of the fatty acyl chain. seen in 600 cells observed (Fig. 7C). This effect of cAMP suggests that Ncs1p might regulate a G protein cascade upstream of adenylate cyclase. Hence Ncs1p, like recoverin, may serve to couple calcium cascades and G protein cascades.
Retinal recoverin, S. cerevisiae Frq1, and Ncs1p G2A each complemented the starvation-independent sporulation phenotype of ncs1⌬ (see Table III). The ncs1⌬ homozygous diploid strain transformed with a multicopy plasmid (pREP1) carrying ncs1 exhibited a markedly lower percentage of cells that sporulate in rich media (3%, as defined in Experiment 1, Table III) compared with that of a negative control (16.7% for deletion mutant transformed with empty pREP1 vector). The ncs1⌬ mutant strain was also transformed with the plasmid overexpressing either Ncs1p G2A (unmyristoylated mutant), Ncs1p E84Q/E120Q/E168Q (Ca 2ϩ binding-deficient mutant), bovine recoverin, or S. cerevisiae Frq1. The expression of Ncs1p G2A in the ncs1⌬ mutant suppressed spore formation in nutrientrich media (5.7%), but the Ca 2ϩ binding-defective mutant (Ncs1p E84Q/E120Q/E168Q ) had less of an effect (11.7%). Hence Ca 2ϩ binding by Ncs1p is essential for suppressing the onset of sporulation in nutrient-rich media. Recoverin or Frq1 overexpressed in the ncs1⌬ mutant also suppressed sporulation in rich media, suggesting that both recoverin and Frq1 can functionally substitute for Ncs1p in fission yeast.
Calcium Sensitivity of ncs1⌬ Mutant-The ncs1 deletion mutant in the vegetative haploid state exhibited markedly lower tolerance for high concentrations of extracellular Ca 2ϩ (Fig. 8). When grown on agar plates, wild type cells grew at very high concentrations of extracellular Ca 2ϩ in excess of 0.5 M. In contrast, the growth of ncs1⌬ mutants was halted in the presence of 0.1 M CaCl 2 (Fig. 8A), but ncs1⌬ mutants could grow at 0.01 M CaCl 2 or lower concentrations of extracellular Ca 2ϩ in conventional media (ϳ1 mM). The growth defect at 0.1 M Ca 2ϩ was specific for Ca 2ϩ because ncs1⌬ mutants grew in the presence of 0.1 M Mg 2ϩ or 1 M K ϩ . Also the growth defect at 0.1 M CaCl 2 was not rescued by 1.2 M sorbitol (osmotic stabilizer), suggesting that the growth defect was not due to an osmotic impairment. The ncs1⌬ cells that stopped growing by Ca 2ϩ treatment for more than 48 h were able to start dividing again when placed on nutrient-rich media that lacked any added Ca 2ϩ . The Ca 2ϩ sensitivity test using homozygous and heterozygous diploid strains suggested that this phenotype is recessive; i.e. only homozygotic mutant showed sensitivity against 0.1 M Ca 2ϩ (data not shown).
The tolerance for high extracellular Ca 2ϩ was restored by overexpressing Ncs1p or Ncs1p G2A in the deletion mutant ( Fig.  8B) but not by the Ca 2ϩ -binding defective triple mutant (Ncs1p E84Q/E120Q/E168Q ). This observation indicates that Ca 2ϩ binding to Ncs1p plays a physiological role in fission yeast and Ca 2ϩ -bound Ncs1p prevents growth arrest that otherwise would occur at high concentrations of extracellular calcium. The Ca 2ϩ intolerance of the ncs1⌬ mutant was not rescued by overexpressing either Frq1 or recoverin. The expression levels of Ncs1p, Frq1, and recoverin in the ncs1⌬ cells were all confirmed to be same (data not shown), and therefore the lack of any complementation was not due to a lack of protein expression. The lack of complementation by Frq1 or recoverin suggests that the Ca 2ϩ tolerance may not be due to simple Ca 2ϩ buffering because both Frq1 and recoverin bind Ca 2ϩ with similar affinity (K d ϭ 10 Ϫ6 M) as Ncs1p. These data show that Ca 2ϩ -bound Ncs1p has a specific role in protecting the cell from effects at high Ca 2ϩ levels.
The growth arrest at high calcium was not rescued by adding exogenous cAMP (data not shown). The lack of any complementation by cAMP, recoverin, or Frq1 suggests that the Ca 2ϩ sensitivity ncs1⌬ phenotype is mechanistically distinct from the starvation-independent sporulation phenotype of ncs1⌬. Hence the two ncs1⌬ phenotypes represent two different and perhaps unrelated biological functions for Ncs1p.
Cellular Localization of Ncs1p-GFP-The in vivo cellular localization of Ncs1p-GFP fusion protein was monitored using fluorescence microscopy. Fluorescence from Ncs1p-GFP in hap- loid cells grown in rich media (YES) was seen mostly in the cytoplasm and was excluded from the nucleus (Fig. 9A). It is noteworthy that Ncs1p-GFP was concentrated into punctate dots or short string-like patterns in the cytoplasm predominantly along the plasma membrane. The membrane association of Ncs1p-GFP was shown more clearly by confocal micros-copy (see Supplement Data). 4 A typical three-dimensional image was made from xy scanning plane of a single cell. Careful analysis of the three-dimensional images revealed that the punctate localization of Ncs1p-GFP was observed near the surface of the plasma membrane with few dots located inside the bulk of the cytoplasm. Significant translocation of Ncs1p-GFP was not observed during conjugation/sporulation. Also changes in glucose concentration in the media (2 and 8%) did not seem to affect the Ncs1p level in the cells (data not shown). Cellular localization of Ncs1p-GFP along the plasma membrane is intriguing and should provide a clue in the future to help identify potential target protein(s) of Ncs1p.
Fluorescence from Ncs1p-GFP in cells grown in liquid minimum media (EMM) was initially very weak and hard to detect. However, when extracellular Ca 2ϩ (Ͼ0.01 M CaCl 2 ) was added to EMM, the fluorescence signal from these cells increased significantly (Fig. 9B). In addition, the intracellular level of Ncs1p-GFP increased quite dramatically in cells grown in the FIG. 7. Mutants exhibit unstable diploidy and starvation-independent conjugation/sporulation. A, color test of homozygous wild type, heterozygous, or homozygous mutant diploid cells streaked on low adenine plates (YES with no adenine added). Shown from the left are NHP058(ncs1 ϩ ) ϫ NHP057(ncs1⌬), NHP057(ncs1⌬) ϫ NHP059(ncs1⌬), NHP058(ncs1 ϩ ) ϫ NHP060(ncs1 ϩ ), and NHP060(ncs1 ϩ ) ϫ NHP059(ncs1⌬). White colonies represent diploid cells, and pink colonies represent haploid cells. Results of counting individual white and pink colonies are summarized below the pictures. B, cell morphology of wild type h 90 strain and its ncs1⌬ mutant incubated in EMM containing 3% glucose (46 h) was observed microscopically. The percentage of zygotes (arrowheads) formed was 4% (ncs1 ϩ ) and 9% (ncs1⌬), and the percentage of zygotic asci (arrows) formed was Ͻ0.4% (ncs1 ϩ ) and 13% (ncs1⌬). Arrowheads and arrows indicate zygotes and zygotic asci, respectively. C, same as in B except carried out in the presence and absence of 10 mM cAMP. Pictures in both B and C are taken as Nomarski images (differential interference contrast (DIC)) (ϫ400). Black bars indicate 10 m.  presence of 0.01 M CaCl 2 compared with that of cells grown with no added CaCl 2 (Fig. 9C). These data suggest that intracellular level of Ncs1p is controlled in a highly calcium-dependent manner consistent with its role as a Ca 2ϩ sensor. DISCUSSION Our NMR structural analysis reveals that S. pombe Ncs1p is structurally similar to mammalian recoverin and exhibits a Ca 2ϩ -myristoyl switch (Figs. 3 and 4). Ncs1p is 46% identical in sequence to bovine recoverin, and several conserved residues are functionally important (Fig. 1). Ncs1p binds functionally to three Ca 2ϩ and exhibits Ca 2ϩ -induced and myristoylation-dependent binding to cell membranes (Fig. 5). The amino-terminal myristoyl group is sequestered in a hydrophobic cavity of Ca 2ϩ -free Ncs1p as evidenced by many NMR dipolar interactions of the myristoyl methyl group with hydrophobic side chains within the protein (Fig. 4A). The binding of three Ca 2ϩ to the protein leads to the extrusion of the myristoyl group out into the solvent (no detectable protein-myristate interactions, Fig. 4B), making the attached fatty acyl group available to interact with lipid bilayers or other hydrophobic targets. The Ca 2ϩ -induced exposure of the myristoyl group permits Ncs1p to associate with membrane-bound targets only at high Ca 2ϩ levels.
The physiological role of the Ca 2ϩ -myristoyl switch was investigated using the ncs1⌬ mutant in fission yeast that exhibited two distinct phenotypes: starvation-independent conjugation and sporulation as well as arrested cell growth at high levels of extracellular Ca 2ϩ (Ͼ0.1 M CaCl 2 ). In ncs1⌬ mutants, mammalian recoverin or S. cerevisiae Frq1 each complemented the starvation-independent sporulation but failed to restore tolerance at high Ca 2ϩ , suggesting that the two ncs1⌬ phenotypes are mechanistically distinct. The Ca 2ϩ binding-deficient triple mutant (Ncs1p E84Q/E120Q/E168Q ) did not influence any of the ncs1⌬ phenotypes, demonstrating that Ca 2ϩ -bound Ncs1p (and a Ca 2ϩ -myristoyl switch) is physiologically active and essential in fission yeast. However, the unmyristoylated Ncs1p G2A mutant partially complemented all the ncs1⌬ phenotypes, suggesting that the amino-terminal myristoyl group may not be essential for the recognition of signaling targets. Instead the fatty acyl group may serve as a dynamic anchor to cell membranes during signaling. Consistent with this interpretation, Ca 2ϩ -bound unmyristoylated recoverin is able to bind and regulate its target (rhodopsin kinase) (57) but is unable to bind to rod outer segment disc membranes (35).
We want to better understand the molecular basis and physiological relevance of Ca 2ϩ intolerance in the ncs1 deletion mutant (see Fig. 8). Interestingly a very similar Ca 2ϩ sensitivity phenotype was reported recently for the deletion mutant of an ncs1 homolog in the plant fungus Magnaporthe grisea (58) and for prz1⌬ in fission yeast (59). The Ca 2ϩ intolerance in ncs1⌬ was not due to any defect in the osmotic stress response (60) because the intolerance was highly selective for Ca 2ϩ . Therefore, Ncs1p appears to protect the cell against lethal effects at elevated calcium levels. Indeed the cellular level of Ncs1p increased markedly when cells were treated with 0.01 M CaCl 2 (Fig. 9), consistent with its role as a Ca 2ϩ sensor. How-ncs1 ϩ -GFP::ura4 ϩ ) was grown on rich media plates (YES) for 48 h. GFP, 4Ј,6-diamidino-2-phenylindole fluorescence (DAPI), and Nomarski image (differential interference contrast (DIC)) are shown. When GFP alone was overexpressed, GFP signal was seen uniformly in the cytoplasm (data not shown). B, NHP078 was grown in minimum liquid media (EMM) in the presence or absence of additional CaCl 2 (0.01 or 0. FIG. 9. Ncs1p is localized in cytoplasm, and its cellular level is regulated by calcium. In A and B, the S. pombe strain that expresses Ncs1p-GFP under endogenous ncs1 promoter was observed by fluorescence microscopy (ϫ1000). Black bars indicate 10 m. A, NHP078 (Ncs1p-GFP expressed under endogenous ncs1 promoter, h Ϫ ever, the Ca 2ϩ tolerance is not due to simple Ca 2ϩ buffering by Ncs1p because recoverin and/or Frq1 (both of which bind Ca 2ϩ with the same capacity as Ncs1p) could not rescue the Ca 2ϩ intolerance of the deletion mutant (Fig. 8B). Instead Ca 2ϩbound Ncs1p might specifically control Ca 2ϩ influx in fission yeast by regulating ion channels. Indeed mammalian NCS-1 negatively regulates voltage-gated Ca 2ϩ channels (61), and the related K ϩ channel-interacting proteins regulate the gating kinetics of A-type K ϩ channels (62). Alternatively Ncs1p may serve as a calcium sensor that activates Ca 2ϩ -induced expression of stress response genes. A number of genes in S. pombe have been postulated to be up-regulated in response to a transient increase in intracellular calcium (3,5,6,59). Furthermore a recently discovered NCS protein called DREAM (downstream regulatory element antagonist modulator) has been shown to serve as a transcriptional repressor that controls Ca 2ϩ -regulated expression of c-fos and prodynorphin genes in the brain (63).
What is the molecular basis of the nutrition-insensitive sexual development in the ncs1 deletion mutant? An important clue is that exogenously added cAMP suppresses this phenotype, suggesting that Ncs1p may regulate the cAMP pathway in sporulation turned on by the glucose-sensing G proteincoupled receptor Git3p (54). Another clue is that this phenotype is complemented by the expression of retinal recoverin, which promotes Ca 2ϩ -regulated desensitization of the G protein-coupled receptor rhodopsin (23)(24)(25). An intriguing hypothesis is that S. pombe Ncs1p may regulate sporulation perhaps by controlling Ca 2ϩ -dependent desensitization of Git3p, analogous to the action of Ca 2ϩ -bound recoverin on rhodopsin.
Desensitization of G protein-coupled receptors is generally promoted by phosphorylation of the cytosolic carboxyl-terminal tail catalyzed by cognate G protein-coupled receptor kinases (GRKs) (64). A variety of Ca 2ϩ -binding proteins (such as recoverin, visinin-like protein, NCS-1, and calmodulin) selectively regulate the activity of GRK subtypes (65,66). Therefore, Ncs1p might regulate a GRK in fission yeast that specifically phosphorylates the glucose-bound form of Git3p, analogous to the action of retinal recoverin in regulating rhodopsin kinase activity (23,24). Fission yeast homologs of familiar GRKs like rhodopsin kinase or ␤-adrenergic receptor kinase have not been identified yet. However, the type I casein kinase genes (YCK1 and YCK2) in budding yeast are known to phosphorylate the carboxyl-terminal region of the ␣-factor pheromone receptor (67) and therefore may serve as a specialized class of GRK genes in yeast. Indeed fission yeast contains three homologs of YCK1 (cki1, cki2, and cki3 (68)) that might serve as cognate GRKs for git3 or pheromone receptors mam2 and map3. We are now in the process of identifying physiological target proteins of S. pombe that bind to Ncs1p in a Ca 2ϩ -regulated fashion to more rigorously understand the regulatory mechanism of conjugation and spore formation mediated by Ncs1p in fission yeast.
Another candidate target protein of Ncs1p is the fission yeast homolog of S. cerevisiae phosphatidylinositol 4-kinase isoform Pik1 (target of Frq1). Frq1 binds and activates Pik1, which is essential for vegetative growth of budding yeast (29). Correspondingly S. pombe contains a highly conserved homolog of Pik1 (SPAC22E12.16c) that most likely is activated by Ncs1p in an analogous fashion. Indeed we see that S. pombe Ncs1p binds tightly to residues 110 -192 of S. cerevisiae Pik1 (data not shown) like we have seen previously for Frq1 (69), suggesting that Ncs1p binds and possibly activates Pik1. Surprisingly the ncs1⌬ mutant of fission yeast was viable and exhibited normal vegetative growth in sharp contrast to the lethal phenotypes of frq1⌬ and pik1⌬ mutants in budding yeast (29). Therefore, if Ncs1p does activate a Pik1 homolog in fission yeast, this function is not essential for supporting vegetative growth as it is in the budding yeast. Conversely the observation that Frq1 complements the starvation-independent sporulation phenotype of ncs1⌬ fission yeast strain suggests that Frq1 might also control nutrition-regulated signaling pathways in the budding yeast.
Pheromone stimulation of budding yeast leads to a transient rise in the intracellular Ca 2ϩ concentration due to the opening of stretch-activated calcium channels encoded by the MID1 gene (70,71). The fission yeast homolog of Mid1 (Yam8p) also has been suggested to participate in the generation of Ca 2ϩ signals during pheromone stimulation of Mam2p and Map3p (5,6). It remains to be elucidated what roles, if any, Ncs1p might play in transmitting pheromone-induced Ca 2ϩ signals in fission yeast.
In summary, S. pombe Ncs1p binds functionally to intracellular Ca 2ϩ and possesses a Ca 2ϩ -myristoyl switch. The ncs1 gene is not essential for vegetative growth of fission yeast in contrast to the lethal null phenotype of the budding yeast homolog FRQ1 (29). The ncs1⌬ mutant exhibits nutrition-insensitive sexual development and a growth defect at high levels of extracellular Ca 2ϩ . The starvation-independent conjugation and spore formation of ncs1 null mutants was suppressed by exogenous cAMP and by the expression of mammalian recoverin or S. cerevisiae Frq1. We propose that Ncs1p may regulate conjugation and sporulation in fission yeast by modulating Ca 2ϩ -dependent desensitization of the G protein-coupled glucose receptor Git3p.