Nak1, an Essential Germinal Center (GC) Kinase Regulates Cell Morphology and Growth in Schizosaccharomyces pombe *

We have identified and characterized Nak1, a 652-amino acid NH 2 -terminal kinase belonging to the group II germinal center kinase (GCK) family, in Schizosaccharomyces pombe . We found that nak1 is essential for cell proliferation. Furthermore, partial repression of nak1 , under regulation of an integrated nmt1 promoter, resulted in an aberrant round cellular morphology, actin and microtubule mislocalization, slow growth, and cell division defects. Overexpression of either a kinase-inactive mutant (Nak1 K39R ) or the non-catalytic domain resulted in similar phenotypes, suggesting dominant-negative effects. By deletion analysis, we mapped the region responsible for this dominant-negative effect to the COOH-terminal 99 residues. Furthermore, we found that deletion of the COOH-terminal 99 residues inhib-ited Nak1 autophosphorylation, and expression of a partially inactive (Nak1 T171A ) or truncated (Nak1 1–562 ) protein only weakly suppressed morphological and growth phenotypes, indicating that both kinase and COOH-ter-minal regions are important for Nak1 function. GFP-Nak1 localized uniformly throughout the cytoplasm, unlike many other proteins which influence cell polarity that preferentially localize to cell ends. Together, our results implicate Nak1 in the regulation of cell polarity, growth, and division and suggest that the COOH-termi-nal end plays an important role in the regulation of this kinase.

Schizosaccharomyces pombe are cylindrical shaped cells that elongate by polarized growth at the cell ends during interphase. The establishment of polarity and symmetrical directionality is essential in growth and developmental processes of most eukaryotic cell types (1)(2)(3)(4). The study of various model systems has identified many genetic elements involved in providing cells with positional information during growth and development. For example, studies of cell morphogenic processes in fission yeast have led to the identification of numerous proteins such as Cdc42 (5), Scd1 (6), casein kinase II (7), Tea1 (8), Orb6 (9,10), and Pom1 (11) that are involved in mediating polar cell growth. Through such studies it is evident that polarized cell growth involves the coordinated function of positional signals within the cell, regulation of signal transduction pathways, and cytoskeletal reorganization (3,12). However, the mechanisms by which such proteins regulate cell morphogenesis and the mode by which their dysfunction results in loss of cell polarity remain unclear.
The p21-activated kinases (PAKs) 1 have been implicated in the regulation of cell morphology and cytoskeletal dynamics (13), various signaling pathways (14 -19), and apoptotic responses (20,21). PAK-related kinases are grouped into two main families based on the arrangement of their respective functional domains (22,23). The true PAKs, originally characterized as primary downstream effectors for Rac/Cdc42 small molecular weight GTPases, have a COOH-terminal kinase domain and an NH 2 -terminal regulatory region. The NH 2 -terminal domain contains a conserved CRIB (Cdc42/Rac interactive binding) motif that mediates Cdc42/Rac binding to PAKs, resulting in their consequent activation (24,25). In fission yeast, Shk1/Pak1 is a critical effector for Cdc42, and it has been shown to play roles in the regulation of cell morphology, sexual differentiation, and mitosis (16, 26 -29). Genetic analyses suggest that the functions of the two fission yeast PAKs (Pak1/ Shk1 and Pak2/Shk2) are largely redundant (29,30).
A second PAK-related kinase family, the GCKs (germinal center kinases), comprise highly conserved NH 2 -terminal kinase domains and less conserved COOH-terminal regulatory domains. Unlike the true PAKs, GCKs do not have CRIB motifs and do not bind Rac/Cdc42 GTPases. GCKs can be subdivided into two groups based on their structural and functional properties. Group I GCKs are most similar to mammalian Gck1 and have homologous carboxyl termini containing at least two PEST motifs, two polyproline SH3 domain binding sites, and an additional ϳ350-amino acid highly conserved region (22). Various group I GCKs have been implicated in mediating stress response and cytoskeletal arrangement (22,23,31). The function and regulation of group II GCKs are less well characterized. Studies in fission yeast have shown that Sid1, a group II GCK, is essential in mediating cytokinesis, most likely by localizing and phosphorylating downstream targets at spindle pole bodies between anaphase and septation (32,33). Cdc14 was recently shown to positively regulate Sid1 by binding the COOH-terminal region of this kinase (34). Previous biochemical evidence suggested that the carboxyl-terminal region of some group II GCKs contains an autoinhibitory domain, al-though Sid1 requires the regulatory region for optimal activity (34 -36).
In this paper, we describe the identification and characterization of Nak1, a fission yeast group II GCK. Our results indicate that Nak1 is required for bipolar cell morphology and cell growth and suggest that the COOH-terminal region plays an important role in the regulation of this kinase.

EXPERIMENTAL PROCEDURES
Yeast Strains and Methods-Genotypes of S. pombe strains used in this study are listed in Table I. Methodology employed in culturing yeast strains, transformation, iodine staining, and transformations was undertaken according to procedures described previously (37).
Identification of nak1-A DNA fragment encoding an NH 2 -terminal Nak1 peptide was amplified from a S. pombe cDNA library, as described previously (38). The amplified nak1 DNA fragment was used as a probe to isolate several clones containing the nak1 gene from a S. pombe genomic DNA library constructed in the pWH5 vector (gift of D. Beach); a 7-kb XbaI DNA fragment encoding the entire nak1 gene was subcloned into pBluescript II SKϪ (Stratagene) to generate pNak1. The nak1 DNA sequence was determined and has been deposited in the GenBank TM database (accession number AF091345).
nak1 Gene and Promoter Replacement-The region encoding the Nak1 catalytic domain was replaced with the ura4 selectable marker in the S. pombe diploid strain SP826 by the gene replacement method as follows (43,44). pNak1⌬Ura was derived from pNak1 by replacing the 653-bp BsiWI-NdeI fragment with a 1.8-kb fragment containing ura4. SP826 was transformed with the 6.4-kb XbaI insert of pNak1⌬Ura, and Uraϩ transformants were selected on PMA ϩ Leu medium. Independent transformants were tested for stability of the Uraϩ phenotype, and Southern blot analysis was performed to confirm that they contained the proper disruption in one copy of the endogenous nak1 genes. h 90 /h ϩN revertants of these strains, which occur at a frequency of ϳ10 Ϫ3 , were detected by the iodine vapor staining test. Diploid strains were analyzed by tetrad analysis.
The endogenous nak1 promoter was replaced with the repressible nmt1 promoter in the haploid S. pombe strain RL143. pGEM-nmt1-nak1 was constructed as follows: (i) a 1016-bp fragment was amplified from S. pombe strain RL143 genomic DNA using primers XK3F and B3REV, appended with an additional 3Ј dA overhang, and cloned into pGEM-T (Promega) to produce pGEM-3RE; (ii) a 2951-bp ura4/nmt1 promoter fragment, amplified from pREP82x using a XF and KR primers, was cloned into XhoI/KpnI sites of pGEM-3RE to produce pGEM-3RE/UN; (iii) a 998-bp fragment, amplified from S. pombe (strain RL143) genomic DNA using a X5F and N5REV primers, was inserted into the XhoI/NheI sites of pGEM-3RE/UN. The 4927-bp NotI/SphI DNA fragment from pGEM-nmt1-Nak1 was used to replace the endogenous nak1 promoter with the pREP82X nmt1 promoter. S. pombe (strain RL143) was transformed with the nmt1-nak1 replacement cassette, and Ura ϩ transformants were passaged on non-selective minimal medium (PMA with Leu and Ura) to select for stable Ura ϩ integrants. Southern blot analysis was performed to confirm integration of the nmt1-nak1 regulatory cassette into the correct locus.
Cell Cycle Analysis-nmt1-nak1 cells were grown in the absence or presence of 100 g/ml thiamine and rapidly fixed in 70% ethanol. Cells were stained with propidium iodide by methods described previously (45), analyzed for cell cycle distribution by flow cytometry, and viewed under isothiocyanate optics for DNA staining. Fluorescence micrographs were compared with corresponding DIC images for the presence of dinucleated cells and the presence of a septum.
Immunoprecipitation and Kinase Assays-Yeast cultures were grown in minimal medium with adenine (PMA) to saturation in the presence of thiamine (100 g/ml), washed once with PMA, diluted into thiamine-free medium, and grown to an OD 600 ϭ 0.8. Cells were subsequently collected by centrifugation and resuspended in yeast lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM EDTA, 0.1% Nonidet P-40, 10 g/ml leupeptin, 1 g/ml pepstatin A, 100 g/ml phenylmethylsulfonyl fluoride, 1 g/ml aprotinin). Yeast cells were vortexed with glass beads, and crude lysates were cleared by centrifugation at 7000 rpm for 1 min. Relative protein concentrations were subsequently determined by Bio-Rad protein assay.
Immunoprecipitation of HA-tagged proteins from yeast cell lysates were undertaken by preclearing equal quantities of yeast cell lysate with protein A-Sepharose slurry for 20 min prior to incubation with 12CA5 (anti-HA) antibody cross-linked to protein A-Sepharose for 2 h at 4°C. Immune complexes were centrifuged at 6000 rpm and washed three times with yeast lysis buffer (750 l per wash); samples were divided evenly upon the last wash. Half of the samples were resuspended in SDS-PAGE protein sample buffer, boiled, separated by SDS-

Nak1
Belongs to the Group II GCK Family-We previously identified kinases from a fission yeast cDNA library by PCR using degenerate oligonucleotides derived from conserved regions of kinase catalytic domains as primers (38). The PCR products were cloned, and their DNA sequences were determined. One clone we identified encoded a region of a novel kinase that we have named Nak1. We used this clone to generate a probe to screen a S. pombe genomic DNA library, and we identified and determined the DNA sequence of a 7-kb DNA insert containing the entire nak1 gene (see "Experimental Procedures"). The DNA sequence of nak1 revealed three exons encoding a 652-amino acid residue protein that exhibits a high degree of sequence identity with PAKs and is most closely related to the group II GCKs (Fig. 1). Like other GCKs, Nak1 has an NH 2 -terminal catalytic region containing the PAK GTPY/FWMAPE signature motif and other conserved domains typical of GCKs. In addition, Nak1 contains a 411-residue COOH-terminal non-catalytic domain that does not share significant sequence identity with other proteins. Similar to group II GCKs, Nak1 lacks a CRIB CDC42/RAC-binding motif characteristic of PAKs, or SH3-binding motifs unique to PAKs and group I GCKs.
Nak1 Is Essential for Viability, and nak1 Repression Results in Loss of Cell Polarity and Growth Inhibition-To investigate the function of Nak1, we constructed and examined the pheno-types of S. pombe strains in which nak1 is deleted. One copy of the nak1 open reading frame was replaced with the ura4 selectable marker in a diploid S. pombe strain (see "Experimental Procedures"). Sporulation and tetrad analysis of several independently derived Ura ϩ diploid strains consistently resulted in two viable Ura Ϫ spores, indicating that nak1 is essential for germination and/or cell growth. On closer examination, it was clear that the other two spores in each tetrad had germinated and given rise to microcolonies (2-8 cells) prior to growth arrest. The cells in these small colonies exhibited an abnormal round shape (data not shown).
To further investigate the function of Nak1, we generated a strain in which the endogenous nak1 promoter was replaced with the thiamine-repressible nmt1 promoter (see "Experimental Procedures"). We observed that 12 h after addition of thiamine to repress nak1 expression, cells were smaller and more oval-shaped, rather than exhibiting the normal rod-shaped morphology of fission yeast (Fig. 2A). This indicates that cells have begun to lose normal polarity, but suggests that they still retain sufficient levels of Nak1 to maintain partial cell polarity. (␣-tubulin) were grown at 30°C in PMA minimal medium in the presence or absence of 100 g/ml thiamine to repress nak1 expression for 12 or 18 h. Cells were either fixed and stained with TRITC-phalloidin to visualize actin localization or visualized for GFP-␣-tubulin to observe microtubule localization by fluorescence microscopy. Cells were also examined by DIC microscopy. B, TYH1 (nmt1-nak1) and SPU (wild type) strains were grown on PMA minimal medium plates in the presence or absence of 100 g/ml thiamine (ϩB1) and 1.2 M KCl (ϩKCl) at the indicated temperature (30°C or 35°C) for 3-5 days. C, TYH (nmt-nak1) cells were grown in the presence (ϩB1) or absence (ϪB1) of 100 g/ml thiamine for 20 h at 30°C and analyzed by flow cytometry.

Nak1, a GC Kinase Regulates Fission Yeast Morphology
Although the small cell size could reflect a reduced G 1 phase, this effect appears to be temporary, since many cells become large and completely round by 18 h after addition of thiamine. Also, at the 18-h time point, actin was no longer polarized and was redistributed around the cellular circumference. We also observed that actin still localized to medial sites of division after loss of cell polarity. Also, microtubules, which normally extend along the long axis of cells, are short and appear to be scattered in differing directions within the cell 12 h after nak1 repression and subsequently lengthen around the radial axis as cells became round and enlarged 18 h after repression. In addition, nak1-repressed cells exhibited a slow growth phenotype, which was more severe at high temperature and high salt concentration (Fig. 2B). We also examined cell cycle progression of nak1-repressed cells by flow cytometry. Actively growing fission yeast spend ϳ70% of their time in G 2 , with a 2C DNA content. Nuclear division occurs in M phase resulting in binucleated cells, which enter the G 1 phase of the cell cycle. However, cell division is not normally completed until S phase as DNA replication takes place. Thus, most cells contain 2C DNA content, except during S phase when the DNA content varies between 1C (cells that divide before DNA replication) and 4C (cells that remain undivided after DNA replication) (48,49). Interestingly, we found that a large population of cells with greater than 2C DNA content accumulated upon nak1 repression (Fig. 2C), suggesting a delay in cell division or cell separation until late S phase. This conclusion is further supported by our observation that there is a concomitant increase in the proportion of binucleated cells upon nak1 repression (Table II). Previous reports indicate that ϳ10% of wild-type cells are septated/binucleated in a growing culture (50, 51). We observed a similar level of nmt1-nak1 binucleated cells in the absence of thiamine, but this level increased significantly 12-18 h after the addition of thiamine to repress nak1 expression. A significant proportion of these binucleated cells were septated, implying a delay in cell separation following septa formation. Together, these results indicate that Nak1 is essential for bipolar cell morphology, cell proliferation, and normal cell cycle progression.
Overexpression of a Mutant Kinase-inactive Nak1 or the Noncatalytic Domain Results in Dominant-negative Phenotypes-To further examine the function and regulation of Nak1, we generated expression constructs containing a series of deletions and specific mutations within the Nak1 coding sequence (Fig. 3), including mutation of the critical ATP-binding region (HA-Nak1 K39R ) predicted to produce an inactive kinase. We also generated a construct predicted to encode a partially inactive kinase (HA-Nak1 T171A ) by site-directed mutation of a critical autophosphorylation/activation Thr residue (19). These mutant proteins were detectable in yeast extracts by Western blot analysis using anti-HA antibody, indicating that they were stably expressed (data not shown). We found that overexpression of HA-Nak1 in a normal haploid S. pombe background did not have a significant effect on growth or morphology (data not shown). However, overexpression of either the kinase-inactive mutant (HA-Nak1 K39R ) or the non-catalytic domain of Nak1 (HA-Nak1 262-652 ) produced a round morphology similar to that resulting from nak1 repression (Fig. 4A), suggesting that these mutant proteins act in a dominant-negative manner to block endogenous Nak1 function. Furthermore, overexpression of the partially inactive mutant HA-Nak1 T171A produced a similar dominant-negative morphological effect (data not shown). These morphological aberrations are also associated with slow growth at high temperature and high salt concentration (35°C, 1.2 M KCl) (Fig. 4B).
The causative region of Nak1 producing these dominantnegative morphological and proliferative phenotypes was mapped to the COOH-terminal end (residues 554 -652), which we refer to as the CTR (COOH-terminal region). Expression of the CTR (HA-Nak1 554 -652 ) alone was sufficient to produce a dominant-negative morphological phenotype, while cells expressing the Nak1 non-catalytic domain lacking the CTR (HA-Nak 262-562 ) appeared normal (Fig. 4A). Moreover, expression of kinase-inactive Nak1 (HA-Nak1 K39R ), the non-catalytic domain (HA-Nak1 262-652 ), or the CTR (HA-Nak1 481-652 ) resulted in severe growth inhibition at high temperature and high salt concentration (35°C, 1.2 M KCl), but cells expressing the kinase-inactive mutant lacking the CTR (HA-Nak1 1-562, K39R ) or the non-catalytic domain lacking the CTR (HA-Nak1 262-562 ) were able to grow under these conditions (Fig. 4B). In summary, these results suggest that overexpression of the CTR abrogates endogenous Nak1 activity resulting in dominantnegative phenotypes.
The CTR Is Required for Proper Nak1 Function-Expression of HA-Nak1 results in full reversion of morphological and slow growth phenotypes associated with nak1 repression in nmt1-nak1 strains (Fig. 5, A and B). However, expression of the partially inactive mutant (HA-Nak1 T171A ) or Nak1 lacking the CTR (HA-Nak1 1-562 ) only partially rescued the morphological and slow growth phenotypes. These results suggest that both kinase activity and the CTR are important for Nak1 function.
To further examine the role of the CTR, we expressed HA-Nak1 and HA-Nak1 1-562 in a wild-type S. pombe strain, immunoprecipitated the HA-tagged proteins from cell extracts, and performed in vitro kinase assays. We found that immunoprecipitated HA-Nak1 exhibited a detectable level of autophosphorylation and phosphorylation of myelin basic protein, whereas deletion of the CTR resulted in a significant decrease in kinase activity, indicating that the CTR is critical for Nak1 kinase activity in vitro (Fig. 5C). Together, our evidence implicates the CTR as an important regulatory sequence and suggests that Nak1 may be regulated by the association of key factors with the CTR.
FIG. 3. Schematic diagram of mutant Nak1 expression constructs. The Nak1 NH 2 -terminal kinase domain (residues 1-262), COOH-terminal non-catalytic domain (residues 262-652), and the CTR region (562-652) are indicated (above). The numbers at the left and the bars at the right indicate the regions of Nak1 encoded by the various deletion constructs.

Nak1, a GC Kinase Regulates Fission Yeast Morphology
tagged Nak1 from the adh1 promoter. Expression of GFP-Nak1 rescued growth and morphological defects in nak1-repressed strains, indicating that it is functional (data not shown). Visualization of various GFP-Nak1 mutants expressed in wild-type cells revealed that GFP-Nak1, GFP-Nak1 1-562 (lacking the CTR), GFP-Nak1 481-652 (the CTR alone), and the GFP control localized throughout the cytoplasm (Fig. 6). Although other kinases, which regulate cell polarity in fission yeast, localize to cell ends in the presence of high salt (52), GFP-Nak1 localization was observed to be unaffected in the presence of 1.2 M KCl (data not shown). Interestingly, we found that GFP-Nak1 and GFP-Nak1 1-562 were excluded from the nucleus, whereas the GFP control and GFP-Nak1 481-652 constructs localized throughout the cell including the nucleus. However, GFP-Nak1 481-652 was consistently more concentrated around the nucleus. Although the significance of this observation is unclear, it may reflect active cellular mechanisms that exclude Nak1 from the nucleus and localize the Nak1 CTR to regions around the nucleus. DISCUSSION Nak1 Is a GC Kinase That Mediates Cell Growth and Morphology in S. pombe-In this study, we report the identification and characterization of the Nak1, a protein kinase in S. pombe.
The lack of proline-rich SH3-binding motifs within the COOHterminal non-catalytic domain of Nak1 indicates that Nak1 belongs to the group II subclass of GCKs. In addition to Nak1,FIG. 4. Overexpression of the Nak1 CTR results in dominant-negative growth and morphological phenotypes. A, wild-type (RL143) cells expressing the indicated mutant Nak1 proteins under the control of the thiamine-repressible nmt1 promoter were grown on thiamine-free medium for 3 days at 30°C and examined by DIC microscopy. B, RL143 cells expressing the indicated mutant Nak1 proteins were also grown on thiamine-free medium at 35°C for 7 days in the presence of 1.2 M KCl. Expression plasmids used for A and B were pREP3XHA (control), pREP3XHA-Nak1 K39R , pREP3XHA-Nak1 1-562, K39R , pREP3XGFPHA-Nak1 262-652 , pREP3XG-FPHA-Nak1 262-562 , and pREP3XGFPHA-Nak1 554 -652 .
FIG. 6. Nak1 localizes to the cytoplasm. Wild-type (SPU) strains were transformed with constructs expressing GFP alone, GFP-Nak1, GFP-Nak1 1-562 (lacking the CTR), and GFP-Nak1 481-652 (the CTR alone). Transformants were grown in liquid medium to mid-log phase at 30°C and visualized by fluorescence microscopy for GFP. Expression constructs used include pAALNeGFP, pAALNGFPHANak1, pAALNGFPHANak1 1-562 , and pAALNGFPHANak1 481-652 . the fission yeast genome encodes two related GCK family kinases, which include Sid1 and an uncharacterized protein kinase (accession number CAB16374) lacking a non-catalytic domain. Although the three GCK family members in S. pombe reveal extensive homology within their catalytic regions, it was unknown whether these kinases are functionally related. Sid1 has been previously shown to be involved in the regulation of septation and cytokinesis during the later stages of anaphase. In contrast, our results indicate that Nak1 is required for cell polarity and growth, suggesting that Nak1 and Sid1 have distinct functions. In support of this conclusion, we found that overexpression of Sid1, or the related budding yeast GC kinases Nrk1 or Sps1, failed to rescue morphology and growth defects in nak1-repressed cells, indicating that Nak1 function is not conserved between related family members (data not shown). However, our results also suggest that nak1 repression results in a delay in cell division and cell separation, which indicate that proper cell division requires both Nak1 and Sid1.
Although the round morphology due to nak1 repression is somewhat similar to cells defective for the Wee1 kinase, nak1repressed cells are more spherical and larger than wee1 mutant cells. Also, we found that nak1-repressed cells lose specific localization of actin to the cell ends, whereas Wee1-deficient cells retain bipolar actin localization (10). Furthermore, Wee1 deficiency results in a greater proportion of cells with 1C DNA content due to a longer G 1 phase, which allows cells to divide prior to S phase (48,49). We did not observe a similar increase in the proportion of cells with 1C DNA content upon nak1 repression. Thus the abnormal round morphology of nak1-repressed cells appears to be due to dysregulation of cell polarity rather than a wee1-like defect. Interestingly, the round morphological phenotype associated with nak1 repression bears strong resemblance to the morphology of cells deficient for Pak1 (29), suggesting that Nak1 and Pak1 may have related functions in the regulation of bipolar cell morphology. However, we found that overexpression of Pak1/Shk1, Orb6, or Wee1 did not suppress growth and morphology defects associated with nak1 repression (data not shown).
Several proteins have been identified that appear to be involved in regulating polarized cell growth in fission yeast. Many of these proteins, such as Tea1, Bud6, SspI, and Pom1, localize to the growing cell ends (8,11,(52)(53)(54)(55). Although our genetic studies suggest that Nak1 is required for polarized cell morphology, we found that GFP-Nak1 did not exhibit specific localization to the cell ends. However, GFP-Nak1 overexpression could mask low levels specifically localized to the cell ends. Nevertheless, it is possible that Nak1 influences the localization of other cell polarity components to specific sites, such as sites of growth and division. Also, since Nak1 localization overlaps with sites of cell growth and division, it may regulate the activities of cell polarity components already present at these sites. Therefore, Nak1 may regulate cell polarity and growth by influencing the localization and activities of other proteins involved in these processes.
The Nak1 CTR Is an Important Regulatory Sequence-Although the COOH-terminal non-catalytic domains of GCKs have been implicated to both inhibit and mediate GCK function, the mechanism by which these domains regulate kinase activity remains unclear (34 -36). Our results demonstrate that a region (the CTR) within the non-kinase domain of Nak1 is important for Nak1 function as measured by in vitro kinase assays and functional complementation of nak1repressed growth and morphological defects. Furthermore, overexpression of the CTR produced dominant-negative morphological and growth inhibitory phenotypes. This dominant-negative effect may result from the ineffectual binding of positive regulators of Nak1 to the CTR.
Together, our results indicate that Nak1 is essential for cell growth, polarity, and normal cell division, and the CTR is important for Nak1 function. Further studies of the interactions between Nak1 and associated proteins may provide insight into the regulation of Nak1-dependent morphogenic mechanisms.