INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Correct progression of the events of the eukaryotic cell cycle is essential for successful division. Cellular components, especially chromosomes, must be accurately replicated and faithfully segregated between the mother and daughter cells. To accomplish this partitioning with high fidelity, mechanisms have evolved to monitor the proper execution of these processes and to impose a delay in cell cycle progression to allow for repair of any damage or mistakes that may have occurred. These control mechanisms have been termed checkpoints (38). Checkpoints triggered by DNA damage (90), by errors in DNA replication (89), and by defects in spindle assembly and kinetochore attachment (33, 85) have been uncovered by genetic analysis in budding and fission yeasts (for reviews, see references 36 and 65). A checkpoint that delays cytokinesis in response to spindle misalignment also has been described (61). Additional control mechanisms must exist to impose the intricate choreography of cell cycle events.

In particular, some feedback control must assess the status of the incipient daughter and impose a delay until the recipient cell is big enough to accept its complement of chromosomes before the cell commits to mitosis. Indeed, in Saccharomyces cerevisiae, such a morphology checkpoint may be based on mechanisms that monitor the state of assembly of the actin cytoskeleton because cells exhibit a delay in G₂ when the actin cytoskeleton is perturbed (53, 56). Even if a daughter is of sufficient size, a cell also needs a mechanism that gauges the state of assembly of the cytokinesis apparatus and imposes a delay until the cleavage machinery is properly assembled and ready to operate before allowing chromosomes to be segregated by the mitotic spindle. Indeed, there seems to be a cytokinesis checkpoint in yeast that monitors formation of a distinct cytoskeletal structure, the septin rings (which mark the incipient site of cell division), because defects in septin assembly also cause a G₂ delay (5). Imposition of both of these checkpoints requires the function of the Swe1 protein kinase (5, 79), a homolog of Schizosaccharomyces pombe Wee1 (74). Swe1 is a negative regulator of the Cdc28 protein kinase (6).

Major stages of the S. cerevisiae cell cycle are initiated by action of the CDC28 gene product (reviewed in reference 57), a founding member of the conserved family of cyclin-dependent kinases (CDKs) (reviewed in reference 59). Cdc28 associates with different sets of cyclins to mediate different cell cycle transitions. Mitotic events are controlled by Cdc28 bound to a set of functionally redundant B-type cyclins, Clb1, Clb2, Clb3, and Clb4 (26, 62). Clb-bound Cdc28 is susceptible to inhibitory phosphorylation on a conserved residue (Tyr19) in its ATP-binding loop mediated by Swe1, and phosphorylation of Cdc28-Clb complexes on Tyr19 blocks entry into mitosis (6). The regulatory role of phosphorylation at Tyr19 in Cdc28 was not appreciated until recently because a Y19F mutation does not have catastrophic consequences (2, 82), unlike the equivalent mutation in the corresponding fission yeast CDK (Cdc2) (31). Phosphorylation at Tyr19 can be reversed by the phosphoprotein phosphatase Mih1 (72), a homolog of S. pombe Cdc25 (73). Thus, a potential cause of the delay in entry into mitosis provoked by the septin-based cytokinesis checkpoint is blockade of Cdc28 activity via phosphorylation at Tyr19 due to activation of Swe1 or inhibition of Mih1 (or both). Conversely, in a normal cell cycle, proper septin assembly must trigger events that assist in preventing phosphorylation of Cdc28 at Tyr19.

The septins are a family of proteins, originally identified in budding yeast (35, 54) but conserved in all eukaryotes (16). Loss-of-function septin mutations cause cytokinesis defects in yeast (8, 37), in Drosophila melanogaster (63), and in mice (48). In S. cerevisiae, septins involved in mitotic cell division are encoded by the CDC3, CDC10, CDC11, and CDC12 genes, assemble into 10-nm filaments (9), and colocalize at the bud neck (54). In vitro, these septins form a tight complex and can assemble into long filaments (27). As visualized by indirect immunofluorescence, the septins appear to form two stacked rings. The apparent plane of the rings is orthogonal to the mother-bud axis; one ring faces the mother cell, and the other faces the daughter. However, there is evidence that the rings may instead represent the prominent lips of a continuous cylindrical collar formed by lateral association of septin filaments parallel to the mother-bud axis (25a, 54a). Under restrictive conditions, temperative-sensitive septin mutants exhibit elongated buds and display biochemical hallmarks of cells delayed in cell cycle progression (11, 17). The precise contribution of septins to cell division is not entirely understood, but it has been suggested that septins form a landmark or scaffold necessary for the recruitment and/or spatial organization of other structures (e.g., actomyosin) required for cytokinesis (19).

How the state of septin assembly transduces a signal that modulates Tyr19 phosphorylation in Cdc28 is not well understood. Clues about the mechanism have come, first, from the demonstration that Hsl1, a protein kinase that is a homolog of S. pombe Nim1 (75), and Hsl7, a protein of previously unknown function, are negative regulators of Swe1 (55). Second, it has been shown recently that Hsl1 localizes to the septin ring and can physically associate with at least one septin (5). Moreover, autophosphorylation of Hsl1 requires an intact septin ring (5), suggesting that assembled septin structures stimulate the catalytic activity of Hsl1. Thus, the state of septin assembly may regulate Swe1 via Hsl1. However, the role of Hsl7 in this regulatory mechanism has been obscure. The primary structure of Hsl7, unlike that of Hsl1, does not possess sequence features diagnostic of any protein of known function. Nonetheless, readily recognizable homologs of Hsl7 are present in all eukaryotes that can be readily examined, including fission yeast (30), Caenorhabditis elegans (55), and humans (49). Skb1, the S. pombe Hsl7 homolog, reportedly interacts with members of the PAK (p21-activated protein kinase) family, as judged by the yeast two-hybrid assay (30), which raised the possibility that Hsl7 itself can serve as a linker or scaffold to assembly or localize a signaling complex. To address its function, we developed reagents to examine the subcellular distribution of Hsl7 and to determine the nature of the other proteins with which Hsl7 interacts. Here we describe our findings, which provide considerable insight as to how Hsl7 participates in the septin assembly checkpoint and assists in communicating a signal from Hsl1 to Swe1.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strains and growth conditions. Yeast strains used in this study are listed in Table 1. Two hsl7 null alleles were constructed. The hsl7-10 mutation was generated by using PCR to amplify the 3' end of the HSL7 gene and its immediately flanking DNA, using appropriate primers and genomic DNA of strain W303 as the template. The resulting product (with EcoRI ends) was ligated into EcoRI-digested pUC19 (93) carrying a 3.8-kb BamHI-BglII fragment containing an hisG::URA3::hisG cassette (1), yielding pCJ61. PCR was used to amplify the 5' end of the HSL7 gene and its immediately flanking DNA, and the resulting product (with flush ends) was ligated with XbaI-digested pCJ61 that had been converted to blunt ends by incubation with T4 DNA polymerase, yielding pCJ65. A 6.3-kb MscI-SalI fragment of pCJ65 carrying the resulting hsl7-10(::hisG::URA3::hisG) construct (in which codons 99 to 806 of HSL7 have been deleted) was used for DNA-mediated transformation of the desired ura3 yeast strains, selecting for Ura⁺ colonies. To generate the hsl7-20 mutation, an internal 2-kb segment of the HSL7 sequence (codons 83 to 740) was deleted from the Litmus-28-derived plasmid pMJS1 (see below) by complete digestion with XbaI and partial digestion with EcoRV and replaced with an XbaI-SmaI fragment carrying the HIS3 gene, excised from pJJ215 (44), simultaneously eliminating the EcoRV site in the HSL7 DNA, and yielding pMJS11. A 3.6-kb fragment carrying the resulting hsl7-20(::HIS3) construct (in which HIS3 is inserted in the same transcriptional orientation as HSL7) was excised from the polylinker in pMJS11 by digestion with StuI and EcoRV and used for DNA-mediated transformation of the desired his3 yeast strains, selecting for His⁺ colonies. Correct transplacement of the HSL7 locus on chromosome II was confirmed both by Southern hybridization analysis (83) and by PCR using appropriate primers and genomic DNA isolated from the transformants and the parental strain as templates. Standard rich (YP) and defined minimal (SC) media (78), containing either 2% glucose (Glc), 2% raffinose (Raf), or 2% galactose (Gal) as the carbon source and supplemented with the appropriate nutrients to maintain selection for plasmids, were used to culture yeast cells. For the two-hybrid screen, SCGlc-Leu-Trp-Ura plates were supplemented with 0.1 M potassium phosphate (pH 7.0) and 40 µg of 5-bromo-4-chloro-3-indolyl--D-galactopyranoside (X-Gal; American Biorganics, Inc.) per ml.

Plasmids and recombinant DNA methods. Plasmids were constructed by using standard procedures (76) in Escherichia coli DH5 (32), SURE (Stratagene, Inc.), or JM110 (93). Unless indicated otherwise, PfuI DNA polymerase (Stratagene) was used for all PCRs. The lack of errors in constructs prepared by PCR was verified by nucleotide sequence analysis using the dideoxynucleotide chain termination method (77). To isolate genomic DNA carrying the HSL7 gene, PCR was used to amplify the corresponding open reading frame from chromosome II (25) by using appropriate primers and genomic DNA of strain W303. The resulting fragment was radiolabeled by the random primer method (24) and used to screen a yeast genomic DNA library in the vector, pSB32 (gift of F. Spencer and P. Hieter, then at Johns Hopkins University School of Medicine, Baltimore, Md.). A 3.7-kb HindIII-HindIII fragment of genomic DNA carrying the entire HSL7 locus and its flanking regions was excised from one of the plasmids isolated and inserted into the polylinker of HindIII-digested vector, Litmus-28 (23), yielding pMJS1.

Two-hybrid screen. Strain YD116 (Table 1) harboring pAS1-HSL7 was transformed with a library of yeast cDNAs fused to the carboxyl terminus of the Gal4 transcriptional activation domain [Gal4(TAD)] driven from the ADH1 promoter on a LEU2-marked 2µm DNA plasmid (gift of S. Elledge, Baylor College of Medicine, Houston, Tex.). Total number of potential transformants screened was estimated by plating appropriate dilutions of a small fraction of the transformed cells on Trp Leu plates to select for cells containing both TRP1-marked pAS1-HSL7 and a LEU2-marked library plasmid. Positive clones that stimulated expression of the URA3 reporter were selected on the same medium lacking uracil and containing X-Gal, which permitted concomitant secondary screening for expression of the lacZ reporter. Candidate plasmids were recovered from these Ura⁺ LacZ⁺ transformants in E. coli (39), reintroduced into YD116 harboring pAS1-HSL7, and retested for activation of expression of the URA3 and lacZ reporters. Specificity was confirmed by failure of the candidates to stimulate reporter expression in the absence of pAS1-HSL7. Candidate clones that passed all of these criteria were tested again for the ability to interact with Gal4(DBD)-Hsl7 in an independent reporter strain, Y190 (3). Partial nucleotide sequences of the inserts in the final confirmed positives were determined and identified by comparison to the S. cerevisiae genome database (Stanford University).

Antibodies. To prepare antigen, pGEX-Hsl7 was expressed in E. coli BL21(DE3)[pLysS], and the resulting GST-Hsl7 fusion protein was purified according to standard procedures. To raise polyclonal antisera, the purified GST-Hsl7 antigen was injected into adult female New Zealand White rabbits (primary inoculation, 0.5 mg; secondary boosts, 0.2 mg) and into adult female mice (primary inoculation, 100 to 200 µg; secondary boost, 100 to 200 µg) using standard methods (34). The immunoglobulin fraction was enriched from rabbit anti-Hsl7 sera by ammonium sulfate fractionation (34). Antibodies were affinity purified from mouse anti-Hsl7 sera by adsorption and elution from purified GST-Hsl7 immobilized on nitrocellulose by using minor modifications of established procedures (81, 86). Other antibodies were obtained from the following sources: mouse MAb 9E10, from S. Grell, Cancer Research Laboratory, University of California, Berkeley); mouse MAb 12CA5, (from BAbCo, Inc., or Santa Cruz Biotechnology, Inc.); and rabbit polyclonal anti-Cdc11, anti-Clb2, and anti-Swe1, from D. Kellogg.

Microscopy. Indirect immunofluorescence was performed by minor modifications of methods described in detail elsewhere (70). Briefly, exponentially growing cells were fixed with 5% formaldehyde in 0.1 M potassium phosphate (pH 6.5) containing 0.5 mM MgCl₂ for 30 min at room temperature and washed in 0.1 M potassium phosphate (pH 6.5). Fixed cells were resuspended in 0.2 M Tris-HCl (pH 9.0) containing 20 mM EDTA (pH 8.0), 1 M NaCl, and 80 mM -mercaptoethanol, incubated at room temperature for 10 min, washed once with potassium phosphate-sodium citrate (pH 5.8) containing 1 M NaCl and twice with potassium phosphate-sodium citrate (pH 5.8), resuspended in 1 ml of solution A (1.2 M sorbitol, 0.1 M potassium phosphate [pH 6.5], 0.5 mM MgCl₂) containing 0.14 M -mercaptoethanol, and digested with 110 µl Glusulase (NEN) and 0.6 mg of Zymolyase 100T (Seikagaku Corp., Tokyo, Japan) per ml. The digested cells were washed twice with solution A, applied to the wells of multiwell microscope slides, and permeabilized by treatment at 20°C with, successively, methanol for 6 min and acetone for 30 s. Permeabilized cells were rehydrated in phosphate-buffered saline, pH 7.3 (PBS), blocked in PBS containing 1 mg of bovine serum albumin per ml, and incubated overnight at 4°C with primary antibodies: a 1:500 dilution of affinity-purified polyclonal mouse anti-Hsl7 antibodies and a 1:500 dilution of rabbit polyclonal anti-Cdc11 antibodies (11). After incubation, cells were washed several times with PBS containing 1 mg of bovine serum albumin per ml, and incubated for 2 h in the dark with secondary antibodies: a 1:500 dilution of indocarbocyanine (Cy3)-conjugated donkey anti-mouse immunoglobulin antibody and a 1:200 dilution of fluorescein isothiocyanate (FITC)-conjugated donkey anti-rabbit antibody (both from Jackson ImmunoResearch Laboratories, Inc.). After staining, cells were washed several times with PBS and examined in an epifluorescence microscope (Zeiss) with a 100× objective lens. Images were collected by using a charge-coupled device camera (Sony) and then processed with Phase 3 imaging software (Northern Exposure, Inc.) and Photoshop (Adobe Systems, Inc.). To view cells expressing GFP-Hsl7, cultures were grown under appropriate selective conditions, and the live cells were examined directly under the fluorescence microscope equipped with an FITC band-pass filter. Images were captured and processed as described.

Coimmunoprecipitation. Protease-deficient strain BJ2168 (Table 1) carrying plasmids expressing the gene products of interest under control of the GAL1 promoter were pregrown under appropriate selective conditions in SCRaf medium to an A₆₀₀ of 0.6, induced by addition of galactose (2% final concentration), and incubated for 2 h. Cells were harvested, washed with PBS, and lysed by vigorous vortex mixing with glass beads in ice-cold lysis buffer (20 mM Tris-HCl [pH 7.2], 12.5 mM potassium acetate, 4 mM MgCl₂, 0.5 mM EDTA, 5 mM sodium bisulfite, 0.1% Tween 20, 12.5% glycerol) containing 1 mM dithiothreitol and protease inhibitors (2 µg of leupeptin per ml, 2 µg of pepstatin A per ml, 1 mM benzamidine, 2 µg of aprotinin per ml, and 1 mM phenylmethylsulfonyl fluoride). The resulting crude extracts were clarified first by centrifugation in a microcentrifuge for 10 min at 4°C and then by sedimentation at 30,000 × g in a table-top ultracentrifuge for 30 min at 4°C. In some experiments, two proteins of interest were coexpressed in the same cell, and then samples (1 mg of total protein) of the extracts were diluted into lysis buffer (200-µl final volume) and mixed with 20 µl of a suspension of protein A-protein G (A/G)-agarose beads (Calbiochem, Inc.). In other experiments, proteins of interest were expressed in separate cell cultures, and then samples (500 µg of total protein) of two different extracts were mixed and diluted into lysis buffer (200-µl final volume) prior to addition of the A/G-agarose beads. For preclearing, these mixtures were incubated for 1 h at 4°C on a roller drum and then subjected to centrifugation in a microcentrifuge for 5 min at 4°C. The resulting supernatant solution was transferred to a fresh tube containing another aliquot (20 µl) of A/G-agarose beads and either 1 µl of mouse ascites fluid containing anti-c-Myc MAb 9E10 or 2 µl of affinity-purified anti-HA MAb. After incubation on the roller drum for 2 h at 4°C, the bead-bound immune complexes were collected by brief centrifugation, washed three times (1 ml each) with ice-cold lysis buffer, resuspended in sample buffer for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and solubilized by incubation in a boiling water bath for 10 min. After removal of any residual particulates by centrifugation for 10 min at room temperature, samples of the resulting supernatant fraction were resolved by SDS-PAGE, transferred to a membrane filter (Immobilon-P; Millipore) by using a semidry transfer apparatus (Bio-Rad, Inc.), analyzed by immunoblotting with appropriate primary antibodies followed by appropriate horseradish peroxidase-conjugated secondary antibodies, and visualized by using a commercial chemiluminescence detection system (Renaissance; New England Nuclear, Inc.).

Protein binding to immobilized GST fusions. Analysis of protein-protein association via interaction of soluble proteins with immobilized GST fusion proteins was performed similarly to immunoprecipitations, with the following modifications. After clarification of extracts by sedimentation in the table-top ultracentrifuge, the resulting supernatant fractions were dialyzed rapidly in a micro-concentration device (Microcon-30; Amicon, Inc.) to remove endogenous glutathione. Samples of an extract to be tested and an extract containing the GST fusion (500 µg of total protein each) were combined, diluted in lysis buffer (200-µl final volume), and mixed with a slurry of glutathione-Sepharose beads (Pharmacia). After incubation for 1 h at 4°C on a roller drum, the beads were collected and washed, and bound proteins were eluted and analyzed, all as described above for immunoprecipitations.

Assay for Swe1 modification. Cell cycle synchronization, analysis of cell cycle progression, and assessment of Swe1 hyperphosphorylation by electrophoretic mobility shift were carried out by procedures described in detail elsewhere (84), with the following modifications. To enhance sensitivity of MATa haploids to -factor-imposed G₁ arrest, all strains used carried the sst1-3 mutation (12) and were treated with 50 ng of -factor per ml for 3 h at 37°C. Culture samples (1.6 ml) were taken at the indicated time points and collected by brief centrifugation, and the resulting pellets were snap frozen in liquid N₂. Frozen pellets were resuspended in 100 µl of analysis buffer (2% SDS, 50 mM Tris-HCl [pH 6.8], 100 mM dithiothreitol, 0.1% bromophenol blue, 10% glycerol) containing protease inhibitors (2 µg of leupeptin per ml, 2 µg of pepstatin A per ml, 1 mM benzamidine, 2 µg of aprotinin per ml, and 1 mM phenylsulfonyl fluoride) and phosphatase inhibitors (10 mM sodium pyrophosphate, 10 mM sodium fluoride, 5 mM sodium metavanadate, 5 mM sodium orthovanadate, 1 mM -glycerol phosphate, and 2 µg of phosvitin per ml) and lysed by vigorous vortex mixing for 1 min with an equal volume of glass beads, followed, successively, by incubation in a boiling water bath for 1 min, vortex mixing for 1 min, and boiling for 5 min. After removal of glass beads and particulates from the solubilized material by brief centrifugation, proteins were resolved by SDS-PAGE on a 8% gel for 2 h at 200 V and analyzed by immunoblotting using either anti-Swe1 antibodies (84) or anti-Clb2 antibodies (47).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Hsl7 acts in a pathway that modulates tyrosine phosphorylation of Cdc28. Loss of HSL1 or HSL7 function causes cells to grow more slowly and to display altered morphology (55). Using either of two different deletion alleles, hsl7-10 (data not shown) and hsl7-20, carrying different selectable markers (URA3 and HIS3, respectively), we confirmed that absence of Hsl7 results in aberrant morphology in strains derived from two distinct lineages (W303 and S288C). Although the mutant cells are slightly larger than otherwise isogenic HSL7⁺ cells (Fig. 1A), the most striking aspect of the altered morphology is the dramatically elongated shape of the bud (Fig. 1B). Elongated buds can arise as the result of defects in genes involved directly in elaboration of the cell wall and other aspects of the actual mechanics of bud formation (reviewed in reference 14). However, elongated buds are also the hallmark of cells that have not undergone the switch from polarized to isotropic growth of the actin-based cytoskeleton (reviewed in reference 52). The transition from polarized to isotropic growth requires the function of Cdc28 complexed to B-type cyclins (Clb1 and Clb2) (51). It has already been demonstrated that the elongated bud morphology in hsl1 mutants is prevented if a swe1 mutation or a CDC28(Y19F) variant is introduced into the same cell, indicating that absence of Hsl1 causes elongated buds because Cdc28-Clb complexes are inhibited by Swe1-dependent phosphorylation at Tyr19 (55). The fact that the morphological phenotype of hsl7 cells is also suppressed by a swe1 mutation (55), and the observation that an hsl1 hsl7 double mutant does not display a phenotype that is more severe than that of either single mutant (data not shown), suggested that Hsl7 functions in the same pathway as Hsl1. We found that the elongated bud morphology and the larger cell size of an hsl7 mutant were completely suppressed by the CDC28(Y19F) allele (Fig. 1C), confirming unequivocally that the morphological effects caused by absence of Hsl7 are mediated solely via tyrosine phosphorylation of Cdc28.

View larger version (147K): [in this window] [in a new window]   FIG. 1.   Elongated bud morphology of the hsl7 mutant arises from G2/M delay imposed by phosphorylation of Cdc28 at Tyr19. Three congenic strains, MJY112 (HLS7+) (A), MJY102 (hsl7) (B), and MJY124 [hsl7 CDC28(Y19F)] (C), were grown to mid-exponential phase in YPGlc at 30°C, and samples of each culture were photographed by using phase-contrast microscopy. Three independent fields are shown for each strain.

Hsl7 colocalizes with the daughter-side septin ring. It has been shown recently that Hsl1 (also known as Nik1 [87]) localizes to the neck between a mother cell and its bud and specifically decorates the bud-side septin ring (5). Likewise, two other Hsl1-related protein kinases, Gin4 (66) and Kcc4 (YCL024w) (5), also localize to the neck in a septin-dependent manner. Given the evidence that Hsl7 functions in the same pathway as Hsl1, it was important to also determine the subcellular localization of Hsl7. To address this question, we first constructed a GFP-Hsl7 fusion, which was fully functional, as judged by its ability to complement completely the morphological defect of hsl7 cells, even when expressed from the authentic HSL7 promoter on a CEN plasmid (data not shown). Examination of cells expressing the GFP-Hsl7 chimera at near-normal levels from the HSL7 promoter (Fig. 2A) or at a much higher level from the GAL1 promoter (Fig. 2B) on CEN plasmids showed that the fusion protein localized as a single ring at the mother-bud junction and seemed to be more closely apposed to the bud side of the neck. This neck-specific deposition of GFP-Hsl7 was observed in cells with small, intermediate, and large buds (Fig. 2A and B); in contrast, neither a ring nor any other obvious localization of the fluorescence were detectable in unbudded cells (data not shown).

View larger version (115K): [in this window] [in a new window]   FIG. 2.   Both a GFP-Hsl7 chimera and native Hsl7 localize to the daughter-side septin ring. Strain MJY112 cells transformed with CEN plasmids expressing GFP-Hsl7 from either the authentic HSL7 promoter (A) or the GAL1 promoter (B) were grown to mid-exponential phase at 30°C in SCGlc-Trp or SCRaf-Leu (followed by induction with 2% galactose for 2 h), respectively, and samples of each culture were viewed directly in a fluorescence microscope, as described in Materials and Methods. Strain MJY112 was grown to mid-exponential phase in YPGlc at 30°C, fixed, permeabilized, costained with affinity-purified mouse anti-Hsl7 antibodies (detected with a Cy3-labeled secondary antibody) and with purified rabbit antibodies directed against the septin, Cdc11 (detected with a FITC-labeled secondary antibody), and viewed in the fluorescence microscope, as described in Materials and Methods (C). Three different individual cells are shown, and the images were merged by using appropriate computer software, as described in Materials and Methods.

Identification of proteins that interact with Hsl7. Although genetic evidence suggests that Hsl7 acts in the same pathway as Hsl1 to negatively regulate Swe1 (55), it was unclear whether Hsl7 acts directly on Hsl1 or on Swe1, or indirectly via interaction with a septin or another protein (such as the Mih1 phosphatase), to alleviate the Swe1-dependent phosphorylation of Cdc28 at Tyr19. To gain further insight about the role of Hsl7 in these processes in an unbiased manner, we used the two-hybrid method to screen a total yeast cDNA library for gene products capable of interacting with Hsl7. For this purpose, we constructed a plasmid that constitutively expressed a Gal4(DBD)-Hsl7 fusion. Stable expression of the fusion was confirmed by immunoblotting with anti-Hsl7 antibodies (data not shown). When the Gal4(DBD)-Hsl7 fusion was expressed in an hsl7 mutant, normal morphology was restored to the vast majority of (but not all) cells (data not shown), suggesting that the chimera is able to supply HLS7 function, but may do so less efficiently than native Hsl7 because some fraction of the hybrid protein is presumably directed to the nucleus via the Gal4(DBD). To identify interacting proteins, cells expressing Gal4(DBD)-Hsl7 were transformed with the library of cDNAs fused to the Gal4(TAD) in a reporter strain (YD116) in which restoration of GAL promoter-dependent transcriptional transactivation drives both URA3 and lacZ expression. From approximately 7 × 10⁶ total transformants screened, 200 Ura⁺ LacZ⁺ colonies were obtained initially. Library plasmids were rescued from half of these original positives. Of these 100 candidates, 30 were able to reproducibly stimulate reporter gene expression in a Gal4(DBD)-Hsl7-dependent manner in both YD116 and an independent tester strain (Y190 [Table 1]). The identity of the gene products in these inserts was deduced by nucleotide sequence analysis and querying of the Saccharomyces genome database by computer, using standard search algorithms.

View larger version (46K): [in this window] [in a new window]   FIG. 3.   A two-hybrid screen identifies both Hsl1 and Swe1 as Hsl7-interacting proteins. (A) A reporter strain (YD116) in which URA3 expression is GAL1 promoter dependent was cotransformed with a plasmid expressing a Gal4(DBD)-Hsl7 chimera and plasmids expressing either Gal4(TAD) alone (top row), Gal4(TAD)-Hsl1(987-1518) (second row), Gal4(TAD)-Swe1(295-819) (third row), or, as an additional negative control, Gal4(TAD)-Ste5 (fourth row). The same strain was cotransformed with a plasmid expressing Gal4(DBD)-Ste7 and plasmids expressing, as a positive control, either Gal4(TAD)-Ste5 (fifth row), Gal4(TAD)-Hsl1(987-1518) (sixth row), or Gal4(TAD)-Swe1(295-819) (last row) to demonstrate specificity. (B) Schematic diagrams of the primary structure of the Hsl1 and Swe1 protein kinases. Catalytic domain (solid box), noncatalytic regions (open boxes), and the minimal Hsl7-binding domain (overlined) in each protein, as delineated by the smallest common segment shared by the corresponding clones isolated in the two-hybrid screen, and their relative sequence positions (numbers below) are indicated.

Hsl7 associates with Hsl1 and Swe1 in cell extracts. To confirm the results of the two-hybrid screen by an independent biochemical method and to determine whether native Hsl7 can interact with full-length Hsl1 and full-length Swe1, we examined whether these proteins bind tightly enough to each other to be coimmunoprecipitated. First, in the process of investigating the interaction between Hsl7 and Swe1, we found that we were unable to cooverexpress HSL7 and SWE1 efficiently in the same cell. Consequently, we prepared separate cell extracts, one from a protease-deficient strain (BJ2168) overexpressing HSL7 and another from the same strain overexpressing a fully functional derivative of SWE1 tagged at its N terminus with a c-Myc epitope (MycSwe1), after each was induced briefly (2 h) from the GAL1 promoter on a CEN plasmid. Equivalent amounts of these extracts, either alone or mixed with each other, were subjected to immunoprecipitation with the anti-c-Myc MAb 9E10. The resulting immunoprecipitates were resolved by SDS-PAGE and transferred to a filter, and the proteins present were analyzed by immunoblotting with appropriate antibodies. As expected, MycSwe1 was efficiently immunoprecipitated by the anti-c-Myc MAb in either the presence or absence of the Hsl7-containing extract (Fig. 4A, left). Conversely, Hsl7 was not nonspecifically absorbed by the anti-c-Myc MAb. Strikingly, however, Hsl7 was efficiently coimmunoprecipitated in the sample that contained both MycSwe1 and Hsl7 (Fig. 4A, right), demonstrating the presence of MycSwe1-Hsl7 complexes. To confirm the tight association of Hsl7 and Swe1 in the reverse manner, we also examined the ability of GST-Hsl7 to bind MycSwe1 in solution. In this experiment, extract containing MycSwe1 was mixed with extracts from BJ2168 that expressed either GST alone, GST-Hsl7, or underivatized Hsl7. After incubation, protein complexes were captured on glutathione-agarose beads. Bound proteins were resolved by SDS-PAGE, transferred to a filter, and analyzed by immunoblotting with appropriate antibodies. We found that GST alone and GST-Hsl7 were efficiently retained by the beads, but not free Hsl7, as expected (Fig. 4B, right panel). Reassuringly, MycSwe1 was bound specifically only to the beads that were exposed to GST-Hsl7 and not to the beads exposed to GST alone or to underivatized Hsl7 (Fig. 4B, left). Thus, full-length Hsl7 and full-length Swe1 are able to physically associate.

View larger version (36K): [in this window] [in a new window]   FIG. 4.   Hsl7 physically associates with Swe1 in cell extracts. (A) Extracts of protease-deficient cells expressing either MycSwe1 or Hsl7 were prepared. Equivalent amounts of total protein from these extracts were incubated either alone or together, as indicated, and then subjected to immunoprecipitation using anti-c-Myc MAb 9E10. The resulting immunoprecipitates were resolved by SDS-PAGE and analyzed by immunoblotting either with MAb 9E10 to detect MycSwe1 (left) or with purified rabbit polyclonal anti-Hsl7 antibodies (right). Samples representing ~1% of the extracts added to each incubation (Input) were examined on the same gels. The asterisk denotes a species present in mouse ascites fluid that adsorbs to the A/G-agarose beads used for immunoprecipitation and nonspecifically cross-reacts with the horseradish peroxidase-linked goat anti-mouse immunoglobulin used for detection of the primary antibody (MAb 9E10). (B) Extracts of protease-deficient cells expressing GST, GST-Hsl, or Hsl7 were prepared. Equivalent amounts of total protein from these extracts were mixed with identical amounts of an extract containing MycSwe1 and then incubated with glutathione-agarose beads. After washing the beads, bound proteins were resolved by SDS-PAGE and analyzed by immunoblotting with mouse anti-Hsl7 antibodies (which were raised against an GST-Hsl7 fusion) to demonstrate specific binding of GST and GST-Hsl7 to the beads (right) and with MAb 9E10 to detect the presence of MycSwe1 (middle). A sample representing ~1% of the MycSwe1-containing extract added to each incubation (Input) was examined on the same gel (left).

View larger version (13K): [in this window] [in a new window]   FIG. 5.   Hsl7 physically associates with Hsl1 in cell extracts. (A) Extracts of protease-deficient cells expressing Hsl7 alone () or Hsl7 coexpressed with either full-length Hsl1 tagged with a triple-HA epitope [Hsl1(HA)3] or a C-terminal fragment, Hsl1(1021-1518), tagged with the same epitope [Hsl1N(HA)3], as indicated, were prepared. Equivalent amounts of protein from these extracts were subjected to immunoprecipitation with mouse anti-HA MAb 12CA5. The immunoprecipitates were resolved by SDS-PAGE and analyzed by immunoblotting with MAb 12CA5 to demonstrate specific recovery of the tagged proteins (right) and with rabbit anti-Hsl7 antibodies to detect the presence of Hsl7 (left). As a marker, a sample representing ~1% of the Hsl7-containing extract (Input) was analyzed on the same gel. (B) Extracts of protease-deficient cells expressing either Hsl1-HA3 or Hsl1N-HA3, as indicated, coexpressed with either Myc-tagged Hsl7(+) or an empty vector (), were prepared. Equivalent amounts of protein from these extracts were subjected to immunoprecipitation with anti-Myc MAb 9E10. The immunoprecipitates were resolved by SDS-PAGE and analyzed by immunoblotting with mouse anti-Hsl7 antibodies to confirm expression of Hsl7 (bottom) and with the anti-HA MAb to detect the presence of the HA-tagged Hsl1 derivatives (top). Samples representing ~1% of the Hsl1-HA3- and Hsl1N-HA3-containing extracts (Input) were also analyzed as markers.

Localization of Hsl7 is dependent on Hsl1. The findings that (i) Hsl7 directly interacts with Hsl1, as judged both by the two-hybrid method in vivo and by coimmunoprecipitation in vitro, and (ii) both proteins localize to a septin ring raised the issue of the order in which the septin-Hsl1-Hsl7 complexes are assembled. It has already been demonstrated that Hsl1 localization is dependent on proper septin assembly (5). Likewise, we found that GFP-Hsl7 was completely delocalized when a temperature-sensitive septin cdc10 mutant was shifted to the restrictive temperature (data not shown). Thus, both proteins are targeted to the bud neck in a septin-dependent manner. In this regard, it has been shown by others that Hsl1 coimmunoprecipitates with the septin, Cdc3 (5). We have confirmed this observation and, moreover, found that full-length Hsl1 coimmunoprecipitates with other septins, specifically, Cdc10 and Cdc11 (data not shown). In contrast, the HA₃-tagged C-terminal fragment of Hsl1 (residues 1021 to 1518), despite its ability to interact efficiently with Hsl7, does not target to the bud neck, as judged by indirect immunofluorescence with anti-HA antibodies (data not shown), suggesting that the septin-binding domain of Hsl1 lies elsewhere in the molecule. Although interaction of Hsl1 with septins can be demonstrated by coimmunoprecipitation, we have been unable under any condition tested to coimmunoprecipitate Hsl7 with any of three septins (Cdc3, Cdc10, and Cdc11) (data not shown). Taken together, these results suggested that the normal order of events involves septin assembly, followed by docking of Hsl1 and then recruitment of Hsl7. Indeed, in agreement with this scenario, it has been reported that Hsl1 localizes to the bud neck in an hsl7 mutant (unpublished results cited in reference 5).

View larger version (22K): [in this window] [in a new window]   FIG. 6.   Localization of Hsl7 at the bud neck requires Hsl1. An HSL1+ strain (MJY112; A) and an otherwise isogenic hsl1 derivative (MJY141; B) were grown to mid-exponential phase in YPGlc at 30°C, then costained for Hsl7 and for the septin, Cdc11, and viewed by indirect immunofluorescence, as described in the legend to Fig. 2. To confirm that stable expression of Hsl7 does not require Hsl1, samples (40 µg of total protein) of cell extracts from the same strains and from an otherwise isogenic hsl7 strain (MJY102) were prepared, resolved by SDS-PAGE, and analyzed by immunoblotting with mouse anti-Hsl7 antibodies (C).

Hsl1 and Hsl7 are required for efficient phosphorylation and degradation of Swe1. During the G₂-M transition, Swe1 becomes hyperphosphorylated and then degraded by ubiquitin-mediated proteolysis (45, 80). Ubiquitinylation of Swe1 is mediated by a so-called SCF complex that depends on a linker protein (Skp1), a specific ubiquitin-conjugating enzyme (Cdc34), and a targeting subunit (Met30) that contains an F-box motif (45). As observed for the destruction of other substrates by related SCF complexes, phosphorylation of the target to be destroyed is a prelude to its recognition by the cognate SCF (reviewed in references 50 and 69). Hyperphosphorylated forms of Swe1 can be detected readily as slower-migrating species by SDS-PAGE (80, 84). The Hsl1 homolog in S. pombe, Nim1/Cdr1, has been shown to phosphorylate directly and inhibit the activity of the S. pombe Swe1 homolog, Wee1 (15). The fact that Hsl7 interacts by the two-hybrid screen and by coimmunoprecipitation with both Swe1 and Hsl1 suggested that its function may be to deliver and/or present Swe1 to Hsl1 to permit its efficient Hsl1-dependent phosphorylation. To test this hypothesis, we examined the state of modification of Swe1 in hsl1 and hsl7 single mutants, in an hsl1 hsl7 double mutant, and in the otherwise isogenic HSL1⁺ HSL7⁺ parental strain. We analyzed the distribution of Swe1-related species both in asynchronous cultures at steady state and in cultures that were synchronized by mating pheromone-imposed G₁ arrest and then released from that block (see Materials and Methods).

hsl1 hsl7

hsl1 hsl7

hsl1 hsl7

hsl1 hsl7

View larger version (51K): [in this window] [in a new window]   FIG. 7.   Modification of Swe1 requires Hsl1 and Hsl7. A MATa HSL1+ HSL7+ strain (MJY155; A) and its otherwise isogenic hsl1 (MJY157; B), hsl7 (MJY156; C), and hsl1 hsl7 (MJY158; D) derivatives were grown to mid-exponential phase in YPGlc at 30°C, and samples of these asynchronous cultures (lane A in each panel) were withdrawn prior to treatment with pheromone, resolved by SDS-PAGE, and analyzed by immunoblotting with anti-Swe1 and anti-Clb2 antibodies. Portions of the same cultures were then synchronized by arrest with -factor, as described in Materials and Methods. After release from the pheromone-imposed block, samples were taken at the time intervals indicated and the status of Swe1 and Clb2 was analyzed in the same fashion.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This study addressed how Hsl7 functions, at the mechanistic level, as a negative regulator of the Swe1 protein kinase. We found that Hsl7 is localized specifically to the bud neck, congruent with the septin ring closest to the daughter cell. We also found, using both genetic and biochemical methods, that Hsl7 physically associates with the Hsl1 protein kinase, another known negative regulator of Swe1 (55), as well as with Swe1 itself. In agreement with these findings, we showed that targeting of Hsl7 to the bud neck requires Hsl1. Moreover, we demonstrated that Hsl7 is required for the Hsl1- and cell cycle-dependent modification and destruction of Swe1. We feel that our results, and the previous observations of others, are best explained by the following model (Fig. 8).

View larger version (17K): [in this window] [in a new window]   FIG. 8.   Model for the role of Hsl1 and Hsl7 in linking septin ring assembly to activation of Cdc28-Clb complexes. See Discussion for details. The catalytic domain (KIN), Hsl7-binding domain (HDB), and postulated septin-binding domain (SBD) of Hsl1 are indicated.

A mechanism for coupling septin ring assembly to inactivation of Swe1. Both Hsl1 (170 kDa) and Hsl7 (95 kDa) are large molecules. In the absence of either Hsl1 or Hsl7, cells are viable and the septin rings seem to form normally at the bud neck, whereas defective septin function is lethal to the cell, suggesting that both Hsl1 and Hsl7 act downstream of septin assembly. As shown recently by Barral et al. (5), the Hsl1 protein kinase localizes to the bud-side septin ring, but in the absence of functional septins, Hsl1 is not present at the bud neck. These workers also demonstrated, by coimmunoprecipitation, that Hsl1 physically associates with at least one septin, Cdc3. Hence, it seems likely that septin assembly recruits Hsl1 to the ring. Moreover, these investigators found that functional septins are required to activate the catalytic activity of Hsl1, suggesting that interaction of this enzyme with the septins causes a conformational change that exposes its N-terminal kinase domain. Correspondingly, it seems that Hsl7 associates only weakly with full-length Hsl1 but binds strongly to the C-terminal 498 residues of Hsl1, as might be expected if the proposed septin-induced conformational change in Hsl1 also exposes its C terminus, permitting the docking of Hsl7.

hsl1 hsl7

Hsl1 and Hsl7 constitute a septin assembly checkpoint pathway. A physiological role for modulation of Cdc28 activity by phosphorylation at Tyr19 was obscure until recently because of the previous demonstration that a Y19F mutation in Cdc28 had no obvious effect on the delay in entering mitosis elicited in response to DNA damage or unreplicated DNA (2, 82). However, Swe1-dependent inhibitory phosphorylation at Tyr19 occurs during the G₂ delay evoked by inhibiting bud formation (79), and this regulation of Cdc28-Clb complexes appears to be the basis of a morphogenesis checkpoint that monitors the state of assembly of the actin cytoskeleton (53, 56). Likewise, Swe1-dependent phosphorylation of Cdc28-Clb appears to be responsible for a G₂ delay caused by defective septin ring assembly (5). Thus, Swe1 is the target of at least two pathways that monitor different aspects of cell morphology.

Hsl7 may regulate Swe1 only in the bud. Septin ring-dependent and Hsl1- and Hsl7-mediated inactivation of Swe1 also promotes another event that requires active Cdc28-Clb complex. Prior to mitosis, the bud grows in a polarized fashion at its tip; however, when the cell enters mitosis, bud growth switches from this anisotropic mode to an isotropic pattern in which the bud expands over its entire surface, thus forming a rounded structure (reviewed in reference 20). Experimental evidence indicates that this morphological transition requires active Cdc28-Clb complexes (51). Our findings suggest that septin ring assembly, and the ensuing Hsl1- and Hsl7-mediated inactivation of Swe1, must precede initiation of the switch from polarized to isotropic growth. Consistent with this conclusion, cells lacking Hsl1 or Hsl7 display marked elongated buds, as shown here and by others (55). This phenotype is also observed in septin mutants, in agreement with the suggestion that septin ring formation is a prelude to isotropic growth (11).

Conservation of this mechanism. In addition to Swe1 in S. cerevisiae, all other eukaryotes examined to date, including Drosophila melanogaster (10