The Cdc42p GTPase Is Involved in a G2/M Morphogenetic Checkpoint Regulating the Apical-Isotropic Switch and Nuclear Division in Yeast*

The Cdc42p GTPase is involved in the signal transduction cascades controlling bud emergence and polarized cell growth in S. cerevisiae. Cells expressing thecdc42 V44A effector domain mutant allele displayed morphological defects of highly elongated and multielongated budded cells indicative of a defect in the apical-isotropic switch in bud growth. In addition, these cells contained one, two, or multiple nuclei indicative of a G2/M delay in nuclear division and also a defect in cytokinesis and/or cell separation. Actin and chitin were delocalized, and septin ring structure was aberrant and partially delocalized to the tips of elongated cdc42 V44Acells; however, Cdc42V44Ap localization was normal. Two-hybrid protein analyses showed that the V44A mutation interfered with Cdc42p’s interactions with Cla4p, a p21(Cdc42/Rac)-activated kinase (PAK)-like kinase, and the novel effectors Gic1p and Gic2p, but not with the Ste20p or Skm1p PAK-like kinases, the Bni1p formin, or the Iqg1p IQGAP homolog. Furthermore, the cdc42 V44A morphological defects were suppressed by deletion of the Swe1p cyclin-dependent kinase inhibitory kinase and by overexpression of Cla4p, Ste20p, the Cdc12 septin protein, or the guanine nucleotide exchange factor Cdc24p. In sum, these results suggest that proper Cdc42p function is essential for timely progression through the apical-isotropic switch and G2/M transition and that Cdc42V44Ap differentially interacts with a number of effectors and regulators.

The Cdc42p GTPase is involved in the signal transduction cascades controlling bud emergence and polarized cell growth in S. cerevisiae. Cells expressing the cdc42 V44A effector domain mutant allele displayed morphological defects of highly elongated and multielongated budded cells indicative of a defect in the apicalisotropic switch in bud growth. In addition, these cells contained one, two, or multiple nuclei indicative of a G 2 /M delay in nuclear division and also a defect in cytokinesis and/or cell separation. Actin and chitin were delocalized, and septin ring structure was aberrant and partially delocalized to the tips of elongated cdc42 V44A cells; however, Cdc42 V44A p localization was normal. Two-hybrid protein analyses showed that the V44A mutation interfered with Cdc42p's interactions with Cla4p, a p21(Cdc42/Rac)-activated kinase (PAK)-like kinase, and the novel effectors Gic1p and Gic2p, but not with the Ste20p or Skm1p PAK-like kinases, the Bni1p formin, or the Iqg1p IQGAP homolog. Furthermore, the cdc42 V44A morphological defects were suppressed by deletion of the Swe1p cyclin-dependent kinase inhibitory kinase and by overexpression of Cla4p, Ste20p, the Cdc12 septin protein, or the guanine nucleotide exchange factor Cdc24p. In sum, these results suggest that proper Cdc42p function is essential for timely progression through the apical-isotropic switch and G 2 /M transition and that Cdc42 V44A p differentially interacts with a number of effectors and regulators.
Cdc42p is a member of the Rho/Rac family of GTPases, which play an essential role in the signal transduction pathways that lead to the establishment and maintenance of cell polarity and polarized cell growth in eukaryotic cells (1). In Saccharomyces cerevisiae, Cdc42p functions in selection of non-random bud sites, rearrangement of the actin cytoskeleton during bud emergence, and in directing actin-dependent secretion into an enlarging bud (2)(3)(4)(5). The characteristic loss-of-function cdc42 phenotype, typified by the cdc42-1 ts , cdc42 T17N , cdc42 T35A , cdc42 W97R , cdc42 D118A , and cdc42 C188S alleles (2, 4 -8), is lethality with cells becoming large, round, and unbudded with multiple nuclei and delocalized actin and chitin. These alleles affect different aspects of Cdc42p function, but all lead to defects in bud emergence, suggesting that Cdc42p primarily functions in this process. However, Cdc42p localizes to the tips and sides of the enlarging bud following bud emergence (7), suggesting that Cdc42p regulates polarized cell growth after bud emergence. The mechanisms by which Cdc42p, in conjunction with various effectors and regulators, selects the bud site and maintains polarized growth throughout the cell cycle are still unclear.
There are a number of S. cerevisiae proteins that function as regulators of Cdc42p activity, including the guanine nucleotide exchange factor Cdc24p (9 -13), and the Bem3p, Rga1p, and Rga2p potential GTPase-activating proteins (GAPs) 1 (14 -17). Cdc24p shows genetic as well as two-hybrid interactions with Cdc42p, exhibits in vitro guanine nucleotide exchange activity against Cdc42p, and is required for bud emergence. Bem3p, Rga1p, and Rga2p are potential Cdc42p GAPs, but only Bem3p has been shown to stimulate the intrinsic GTPase activity of Cdc42p in vitro. Characterization of bem3 rga1 mutants suggest that the GAPs do not have an essential role in bud emergence, but their elongated bud morphology (17) suggests that inactivation of Cdc42p to a GDP-bound state is necessary for cessation of apical polar bud growth. Analysis of these regulators suggests that the regulation of Cdc42p is critical for the establishment of bud emergence and may affect cellular processes later in the cell cycle.
Additionally, Cdc42p and Cla4p function in the localization of the S. cerevisiae septin proteins that are thought to be components of the 10-nm filament ring that forms at the bud site prior to budding and remains at the mother-bud neck region through cytokinesis (38).
To further elucidate Cdc42p cellular function during the cell cycle, the cdc42 V44A allele was characterized. The V44A mutation lies within the Cdc42p effector domain and was originally isolated due to its ability to block two-hybrid protein interactions between Cdc42p and Cdc24p (6). cdc42 V44A cells predominantly had elongated buds containing a single nucleus with a small percentage of cells conferring a multibudded, multinucleate phenotype indicative of a apical-isotropic switch defect, a G 2 /M delay, and a partial cytokinesis defect. The data presented here also show that localization of actin, chitin, and the septin ring was altered in elongated-budded cdc42 V44A cells while Cdc42 V44A p localization appeared to be normal. In conjunction with studies using a SWE1 CDK inhibitory kinase mutant, which regulates the G 2 /M morphogenetic checkpoint, these results indicated that the Cdc42 V44A mutant protein activated the G 2 /M checkpoint. Cdc42 V44A p is also defective in its interactions with a subset of CRIB (Cdc42/Rac interactive binding; Ref. 39) domain-containing downstream effectors, suggesting that Cdc42 V44A p differentially interacts with these proteins. Taken together, these data highlight Cdc42p differential interactions and show that Cdc42p function is required for timely cell cycle progression.

EXPERIMENTAL PROCEDURES
Reagents, Media, and Strains-Enzymes, polymerase chain reaction (PCR) kits, and other reagents were obtained from standard commercial sources and used as specified by the suppliers. Oligonucleotide primers for sequencing and PCR were obtained from Genosys (The Woodlands, TX). Growth media and maintenance of bacterial and yeast strains have been described previously (40,41). The S. cerevisiae strains used are listed in Table I. The ⌬cla4::TRP1 marker was disrupted by HIS3 in strain TRY3-H by transforming the trp1::HIS3 marker swap fragment (42), released from pTH4 by digestion with XhoI and EcoRI, into TRY1-6B (6). One of the stable His ϩ integrants was designated TRY3-H and used for further experiments. Yeast transformations were performed as described previously (41). Selection of transformants was on synthetic complete drop-out media lacking specified amino acid(s) and containing 2% glucose as a carbon source. To induce the GAL1 promoter, transformants were grown on solid or liquid media containing 2% raffinose ϩ 2% galactose.
PCR and Site-directed Mutagenesis-To create pEG202(cdc42 V44A,C188S ), cdc42 V44A was amplified from pAS1-CYH2(cdc42 V44A ) by PCR using the Expand Long Template PCR system (Roche Molecular Biochemicals). The 5Ј-primer sequence used in the PCR reaction was GGAATTCCATATGCA-AACGCTAAAGTGTG (underlined sequence is an EcoRI site; double underlined sequence is an NdeI site that contains the CDC42 start codon), and the 3Ј-primer sequence was CGCGGATCCGACTACAAAATTGTAGATTT-TTTACTTTTCTTGATAACAGG (underlined sequence contains a BamHI site followed by the CDC42 stop codon and the reverse complement of CDC42; Cys 188 to Ser; underlined G is C in the wild type sequence). The PCR cycling parameters were 30 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 1 min. The resulting 576-base pair PCR fragment was digested with EcoRI plus BamHI and inserted into EcoRI/BamHI-digested pEG202. The resulting construct was sequenced to verify the fusion and cdc42 V44A,C188S sequence.
pEG202(cdc42 G12V,V44A,C188S ) was created using the QuikChange TM sitedirected mutagenesis kit from Stratagene. pEG202(cdc42 V44A,C188S ) was the DNA template for the mutagenesis. The nucleotide sequence of the forward mutagenic primer was GTCGGTGATGTTGCTGTTGGGAAAACG (Gly 12 to Val; underlined T is G in the wild type sequence), and the reverse mutagenic primer was CGTTTTCCCAACAGCAACAT-CACCGAC (Gly 12 to Val; underlined A is C in the wild type sequence). The cycling parameters for the mutagenesis were 12 cycles of 95°C for 30 s, 55°C for 30 s, and 68°C for 21 min. pRS315(pGAL1-cdc42 G12V,V44A ) was also created using the QuikChange TM kit, pRS315(cdc42 V44A ) as the DNA template, and the mutagenic primers containing the G12V mutation and cycling parameters described above. The resulting plasmid pRS315(cdc42 G12V,V44A ) was cut with NdeI to release the cdc42 G12V,V44A fragment, which was inserted into pRS315(pGAL1-CDC42) cut with NdeI. Sequencing of both plasmids resulting from the mutagenesis confirmed the sequence of the cdc42 G12V,V44A,C188S and cdc42 G12V,V44A alleles.
p416MET(GFP S65T -A 8 -cdc42 V44A ) was also created using the QuikChange TM kit with p416MET(GFP S65T -A 8 -CDC42) as the DNA template. The nucleotide sequence of the forward mutagenic primer was CGA-TAACTATGCGGTGACTGCGATGATTGGTGATGAACC (Val 44 to Ala; underlined C is T in the wild type sequence) and the reverse mutagenic primer was GGTTCATCACCAATCATCGCAGTCACCGCATAGTTATCG (Val 44 to Ala; underlined G is A in the wild type sequence). The cycling parameters for the mutagenesis were as above.
Photomicroscopy-Cells were grown in the appropriate liquid media at 23°C to mid-log phase. Cells were collected, sonicated, and examined morphologically. Methods for preparing and staining cells with 4Ј,6diamidino-2-phenylindole (DAPI), Calcofluor, and rhodamine-phalloidin have been described previously (48). Cells containing GFP-Cdc3p, GFP-Cdc12p, and GFP-A 8 -Cdc42p constructs were grown to mid-log phase, sonicated and observed. Cells containing GFP-A 8 -Cdc42p constructs were grown in SCϪUraϪMet media for expression from the methionine repressible promoter. Photomicroscopy using Hoffman modulation optics was performed on an Olympus BH-2 epifluorescence microscope. DAPI-stained cells were examined on a Nikon TE300 inverted microscope equipped with epifluorescence illumination and Nomarski optics. Video images were obtained using a VE1000SIT camera (Dage-MTI, Michigan City IN). The magnification of the image was adjusted with projection optics placed between the microscope and the video camera. Nikon filter cube UV-2A (excitation 330 -380 nm, emission Ͼ420 nm) was used for visualizing DAPI fluorescence. Photomicroscopy using phase contrast optics (Fig. 2, B and C) was also performed on an E400 Nikon microscope (Omega Optical, Brattleboro, VT). GFP-A 8 -Cdc42p expressing cells were visualized on this microscope using Omega optical filter cube XF100 and DAPI staining was visualized with Omega optical filter cube XF06 filter. All cell and plate images were digital images that were obtained using a Dage VE-VG-5 video frame grabber card (Scion Corp, Frederick, MD) and analyzed in Adobe Photoshop 5.0 (Adobe Systems, Inc., San Jose, CA) on a PowerTower 180 MacOS computer (PowerComputing, Round Rock, TX). Where indicated, cells from the same culture but different fields were manipulated into collages using Adobe Photoshop 5.0.

cdc42 V44A Mutant Cells Have an Apical-Isotropic Switch
Defect and a Partial Cytokinesis Defect-The V44A mutation lies within the Cdc42p effector domain and was identified in a screen for point mutations that interfered with two-hybrid interactions between the Cdc42 D118A, C188S mutant protein and Cdc24p (6). The V44A mutation intragenically suppressed the dominant negative growth and morphological phenotype associated with the cdc42 D118A allele at 23°C and the dominant lethality associated with the cdc42 G12V activated allele (data not shown), raising the possibility that the V44A mutation led to an inactive protein. However, the cdc42 V44A mutant allele could complement a ⌬cdc42 mutation in strain TRY5-3A, and these cells grew at 16°C, 23°C, and 30°C, but not at 37°C, indicating that cdc42 V44A encoded a functional protein at low but not high temperatures. At 23°C, ϳ80% of the cdc42 V44A cells had abnormally elongated buds and the cells appeared considerably larger than wild type cells (Fig. 1A, upper panel), suggesting that Cdc42 V44A p was partially functional at 23°C. Additionally, ϳ15% of the cells were multibudded (Fig. 1B). Staining of these cells with the fluorescent DNA stain DAPI (Fig. 1A, lower panel) indicated that 56% of the abnormally budded cells had 1 nucleus (Fig. 1A, arrowheads), 35% had two nuclei and 9% of the cells had more than two nuclei (Fig. 1A, arrows). The elongated-budded morphology along with the multinucleated cells indicated that Cdc42 V44A p conferred a delay in the apical-isotropic switch in bud growth and in the G 2 /M transition as well as a partial cytokinesis defect to cells at 23°C.
To determine whether cdc42 V44A had a cytokinesis or cell separation defect, cdc42 V44A cells were fixed, treated with the cell wall-digesting enzyme glusulase, and the morphological phenotype quantified pre-and post-glusulase. The percentage of budded and multibudded cells in the glusulase-treated cells decreased by 30%, correlating with a 30% increase in the unbudded cell population. These results suggested that cdc42 V44A cells have a cell separation defect. However, ϳ8% of the budded population post-glusulase were still multibudded, suggesting that cytokinesis may be defective or at least delayed in these cells.
The cdc42 V44A phenotype at 37°C was also characterized to determine the basis for the lethality. After 6 h at 37°C, the cdc42 V44A strain had a mixed population of cells with 62% being abnormally budded and 38% being large, round, and unbudded as compared with prototypic cdc42-1 ts cells, 69% of which were large, round, and unbudded (n ϭ 100 cells). This 37°C phenotype differed not only from the cdc42-1 ts phenotype, but also from the abnormal 20°C morphology of cdc42 V44A cells, in which no large, round, unbudded cells were observed.
The cdc42 V44A Mutant Triggers the Swe1p-dependent Morphogenetic Checkpoint to Delay the Apical-Isotropic Switch and Nuclear Division-The cdc42 V44A elongated-budded phenotype suggested that these cells have a delay in the apical-isotropic growth switch, resulting in hyperpolarized cortical actin. Indeed, cdc42 V44A cells showed an abnormal cortical actin localization pattern as compared with wild type cells ( Fig. 2A). Cortical actin was polarized to the tips of elongated buds, suggesting that the apical-isotropic growth switch that occurs during the G 2 /M transition was delayed in cdc42 V44A cells. This actin localization pattern was similar to the patterns seen in abnormal ⌬cla4 cells ( Fig. 2A) and was less severe but similar to the actin localization seen in cdc12 ts cells at non-permissive temperatures (data not shown; Ref. 49). These results also suggested that cdc42 V44A defects may trigger the Swe1p-dependent G 2 /M morphogenetic checkpoint, which monitors the apical-isotropic switch through the Clb2p/Cdc28p regulatory kinase (50 -53).
To examine whether cdc42 V44A defects triggered the Swe1pdependent morphogenetic checkpoint, the cdc42 V44A mutant phenotype was examined in a ⌬swe1 background. A cdc42 V44A ⌬swe1 double mutant no longer exhibited the elongated-budded phenotype characteristic of cdc42 V44A cells at 23°C (Fig. 2B), whichsupportedanapical-isotropicswitchdelaythatwasSwe1pdependent. However, the morphology of these cells was not completely normal, with 55% of the cdc42 V44A ⌬swe1 budded cells displaying a multibudded or irregular morphology (Fig.  2B). DAPI staining revealed that 29% of all cdc42 V44A ⌬swe1 cells contained greater than two nuclei (Fig. 2C), indicating that the ⌬swe1 mutation did not suppress and may exacerbate the cdc42 V44A multinucleate phenotype. The multiple nuclei were also unevenly distributed in mother and daughter cells with some daughter cells having two nuclei (Fig. 2C, leftmost  panel). The increased number of nuclei per cell and the abnormal multibudded morphology of cdc42 V44A ⌬swe1 cells indicated that loss of Swe1p exacerbated Cdc42 V44A p effects, highlighting the cytokinesis and/or cell separation defect. The nuclear staining of cdc42 V44A cells (Fig. 1A) as compared with cdc42 V44A ⌬swe1 cells also suggested that cdc42 V44A cells had a nuclear division delay, which is characteristic of activation of the G 2 /M morphogenetic checkpoint.
To confirm whether these mutant cells truly had a cell cycle delay, cell synchrony experiments to follow the nuclear cycle were attempted. However, due to the mutant morphology, standard cell synchrony experiments, such as ␣-factor arrest, were unsuccessful. However, it was found that growth in 0.5 M NaCl remediated the abnormally budded morphological and multinucleate phenotypes at 23°C with 82% of the budded cdc42 V44A cells (strain TRY5-8C) having a normal budded morphology and none of the normally budded cells having more than 2 nuclei. Furthermore, cdc42 V44A cells grown in NaCl and then shifted to restrictive temperatures in fresh media without NaCl were found to be a nearly uniform population of large unbudded cells. These results suggested that NaCl remediated the 23°C mutant morphology but did not remediate the cdc42 V44A temperature sensitivity.
Using this NaCl remediation as a means to increase cell synchrony, cell cycle reciprocal shift experiments were performed to determine if there was a delay in the nuclear cycle of cdc42 V44A cells relative to cdc42 V44A ⌬swe1 cells. cdc42 V44A cells (strain TRY5-8C) and cdc42 V44A ⌬swe1 cells (strain TRY15-21A) were grown in 0.5 M NaCl media to late-log phase, A, the morphological phenotype and DNA content of the cdc42 V44A strain TRY5-3A was compared with wild type C276 -4A cells at 23°C using Nomarski optics (upper panels) and DAPI staining (lower panels). Cells were grown in YEPD liquid media to mid-log phase and sonicated briefly before observation. Scale bar, 10 m. B, multibudded cdc42 V44A cells at 23°C (strain TRY5-3A).
washed, and then shifted to 37°C in fresh media without NaCl for 1 h to obtain a population in which ϳ80% of the cells were unbudded and Ͼ80% of the cells were uninuclear. These cells were then shifted to permissive temperature (23°C) and observed over time with DAPI to determine when nuclear division occurred. Both cdc42 V44A and cdc42 V44A ⌬swe1 mutant cells began budding ϳ1 h after the shift to 23°C and the abnormally elongated budded phenotype in cdc42 V44A mutant cells became apparent ϳ30 min later with buds increasing in length over time. At 4 h after the shift to permissive temperatures, ϳ50% of the abnormally budded cells had two nuclei, suggesting that these cells eventually complete nuclear division. In contrast to cdc42 V44A cells, ϳ50% of the cdc42 V44A ⌬swe1 cells had two or more nuclei within 2 h after the temperature shift. These results indicated that cdc42 V44A cells had an ϳ2-h nuclear division delay as compared with cdc42 V44A ⌬swe1 cells and indicated further that the cdc42 V44A cell cycle delay is Swe1p-dependent.

cdc42 V44A Cells Have Abnormal Chitin Localization and Aberrant and Delocalized Septin Ring
Structures-To determine if chitin was delocalized as seen in other cdc42 mutants, cdc42 V44A cells were stained with the chitin stain Calcofluor. The haploid cdc42 V44A cells had a normal axial budding pattern, but ϳ50% of the budded cells did not form a defined chitin ring at the mother-bud neck region and had diffuse chitin localized around the entire periphery of the cell (Fig. 3). The cdc42 V44A mutant cells staining pattern was comparable to that seen in ⌬cla4 (Fig. 3) and cdc12 mutants (data not shown; Ref. 54). The mother-bud neck region of cdc42 V44A cells also appeared wider than wild type cells and was comparable to the enlarged neck region of ⌬cla4 cells. These results suggested that one of the mechanistic defects in cdc42 V44A cells was delocalized chitin deposition. Since chitin localization is de-pendent on the septins, these observations suggested that cdc42 V44A cells had defects in septin ring localization.
To further explore a possible link between Cdc42 V44A p and septins, we compared the localization of Cdc3p and Cdc12p tagged with GFP, which were expressed from their endogenous promoters in plasmid pRS315, in wild type and cdc42 V44A cells ( Fig. 3; only GFP-Cdc12p is shown; GFP-Cdc3p gave identical results). In wild type cells, both GFP-Cdc3p and GFP-Cdc12p localized to the mother-bud neck region as a bright ring structure and many of the cells showed a double ring structure at the neck region, as was seen in immunofluorescence localization studies with anti-Cdc3p and anti-Cdc12p antibodies ( Fig.  3; Refs. 55 and 56). It was also noted that the overexpression of either GFP-Cdc3p or GFP-Cdc12p did not alter the morphology of wild type cells. In contrast to wild type cells, expression of GFP-Cdc3p and GFP-Cdc12p did alter the morphology of cdc42 V44A cells with Ͼ50% of cells appearing normal. The elongated budded cells typically seen in the cdc42 V44A cell population made up only a small percentage of the cdc42 V44A cell population overexpressing either septin protein, suggesting that Cdc3p or Cdc12p suppressed the morphological defects seen in cdc42 V44A cells alone. Overexpression of Cdc12p alone also showed the same quantitative level of suppression, but it did not suppress the cdc42 V44A temperature sensitive defect. However, the morphology of the elongated budded cells overexpressing either GFP-Cdc3p, GFP-Cdc12p, or Cdc12p was altered from the typical cdc42 V44A cell morphology as evidenced by the presence of cells with normal buds but a slightly elongated neck region (Fig. 3, arrows; normally budded cells with the slightly elongated neck regions were considered abnormally budded cells in the quantitative analysis of GFP-Cdc3p, GFP-Cdc12p, or Cdc12p suppression).
The cdc42 V44A cells containing GFP-Cdc3p or GFP-Cdc12p FIG. 2. Actin localization and suppression of the cdc42 V44A morphology by ⌬swe1. A, C276 -4A (wild type), TRY5-3A (cdc42 V44A ), and TRY1-9D (⌬cla4) cells were grown in YEPD at 23°C to mid-log phase, fixed and stained with the actin fluorescent stain rhodaminephalloidin using standard fixation and staining procedures. All cells were sonicated briefly before observation. B, TRY15-23B (⌬swe1), TRY15-23A (cdc-42 V44A ), and TRY15-21A (cdc42 V44A ⌬swe1) were grown in YEPD liquid media at 23°C to mid-log phase, sonicated briefly, and observed. C, TRY15-23B (⌬swe1) and TRY15-21A (cdc42 V44A ⌬swe1) were grown in YEPD liquid media at 23°C to mid-log phase, stained with DAPI, and sonicated briefly before observation. Scale bars, 10 m. that appeared to have normal buds had intact, properly localized septin rings. In normally budded cells with a slightly extended neck region, both GFP-Cdc3p and GFP-Cdc12p containing septin rings appeared to be intact. However, in these abnormal cells, the septin ring was localized at the base of the rounded bud at the point of constriction as opposed to at the mother-bud neck region (arrows in Fig. 3 show GFP-Cdc12pcontaining cells which are representative of GFP-Cdc3p cells). Furthermore, of the elongated budded cells seen, many appeared to have a diffuse septin ring and GFP fluorescence could be seen at the tip and around the sides of the bud (Fig. 3,  arrowheads). This aberrant septin ring localization was seen in 82% of the abnormally elongated budded cells. These results suggested that the Cdc42 V44A p had an adverse effect on the localization of Cdc3p and Cdc12p and the structural formation of the septin ring.
Abnormal localization of GFP-Cdc3p and GFP-Cdc12p was also observed in ⌬cla4 cells, with the abnormally elongated budded cells having a diffuse staining pattern and bud tip staining pattern that was similar to the patterns seen in cdc42 V44A cells (Fig. 3, arrowheads). The ⌬cla4 cells also had intact septin rings in normal cells and intact rings that were wider or delocalized in elongated budded cells (Fig. 3, arrows). Interestingly, the percentage of abnormally budded cells in the ⌬cla4 cell population did not change upon overexpression of GFP-Cdc3p or GFP-Cdc12p, suggesting that overexpression of Cdc3p or Cdc12p could not suppress the abnormal morphology associated with ⌬cla4 cells (see "Discussion").
GFP-A 8 -Cdc42 V44A p Localizes Properly to Regions of Polarized Growth-The altered localization of actin, chitin, and septin proteins in cdc42 V44A cells raised the possibility that these morphological defects could be due to delocalization of Cdc42p. To explore Cdc42p localization patterns through the cell cycle, Cdc42p was fused in-frame to the C terminus of GFP containing a C-terminal eight alanine linker, and expressed from a methionine-repressible promoter in wild type strain TRY11-7D. The GFP-A 8 -Cdc42p fusion was able to complement the cdc42-1 ts mutant, indicating that GFP-A 8 -Cdc42p was functional. 3 As seen previously with endogenous Cdc42p and anti-Cdc42p antibodies (7), GFP-A 8 -Cdc42p localized to the presumptive bud site prior to bud emergence, and to the sides and tip of emerging buds (Fig. 4, upper panels). In most medium to large budded cells, GFP-A 8 -Cdc42p was localized to the tips and sides of buds (Fig. 4, upper panels), which was not clearly distinguished previously (7). In addition, in most large budded cells, GFP-A 8 -Cdc42p was localized at the mother-bud neck region (Fig. 4, upper panels), suggesting that Cdc42p may function later in the cell cycle around the time of cytokinesis and/or cell separation. A similar localization pattern was seen for a GFP-A 8 -Cdc42 V44A fusion protein (Fig. 4, upper panels), indicating that the cdc42 V44A morphological defects were not due to delocalized Cdc42 V44A p. GFP-A 8 -Cdc42p and GFP-A 8 -Cdc42 V44A p were also expressed in the ⌬cla4 strain TRY1-9D and both GFP-A 8 -Cdc42p constructs localized to the presumptive bud sites, to the tips of elongated buds and to the motherbud neck region in some cells (Fig. 4, lower panels). These results suggested that Cdc42p localization was not dependent on Cla4p. The localization of GFP-A 8 -Cdc42p to the elongated bud tip suggested that Cdc42p may regulate, or be dependent on, the apical-isotropic growth switch and that the mitotic delay characteristic of cla4 mutants may result in prolonged localization of Cdc42p to the bud tip.
Two-hybrid and Genetic Evidence Indicated Cdc42 V44A p Had Altered Interactions with Downstream Effectors Cla4p, Gic1p, and Gic2 but Not Ste20p, Skm1p, Bni1p, or Iqg1p-Since the V44A mutation is located in the effector domain, it seemed likely that the primary defect of Cdc42 V44A p was due to altered interactions with its regulators and effectors. Previous twohybrid experiments had indicated that the V44A mutation affected interactions between Cdc42p and Cdc24p (6)   tions, two-hybrid analysis was performed with known Cdc42p effectors. Two-hybrid interactions between Cla4p and the Cdc42 V44A,C188S mutant protein were significantly decreased, but not abolished, in comparison to interactions between Cla4p and the Cdc42 C188S p (Fig. 5; the C188S mutation was introduced to override the normal plasma membrane localization of Cdc42p). The Cdc42 G12V,V44A,C188S p also showed a reduced interaction with Cla4p as compared with Cdc42 G12V,C188S p interactions (data not shown). These results suggested that the V44A mutation partially inhibited interactions between Cdc42p and Cla4p. The V44A mutation also affected interactions with the novel effectors Gic1p and Gic2p with both proteins showing a reduced interactions with Cdc42p when the V44A mutation was present (Fig. 5) and also when in conjunction with the G12V mutation (data not shown).
In contrast, the V44A mutation did not affect interactions of either Cdc42 C188S p or Cdc42 G12V, C188S p with downstream effectors Ste20p, the third PAK homolog Skm1p, the Bni1p formin or the cytokinesis effector Iqg1p (Fig. 5). The interactions between Skm1p and Cdc42p were relatively weak (4.3 Ϯ 1.4 Miller units) when compared with the interactions between Cdc42 C188S p and Cla4p or Ste20p (Fig. 5), but they did increase when Cdc42p was locked in the activated GTP-bound state by the G12V mutation (43.3 Ϯ 1.4 Miller units with Cdc42 G12V, C188S p and 33.1 Ϯ 3.7 Miller units with Cdc42 G12V,V44A,C188S p), suggesting that Skm1p is a bona fide Cdc42p downstream effector. None of the proteins tested interacted with the presumably GDP-bound Cdc42 D118A p as was previously reported (data not shown; Refs. 17 and 28)), or with a Cdc42 deletion derivative missing amino acids 32-40 of the effector domain (data not shown), suggesting that all these proteins have some contact with the Cdc42 effector domain. Overall, these twohybrid results suggested that the V44A mutation differentially altered interactions with Cla4p, Gic1p, and Gic2p, but did not affect interactions with Ste20p, Skm1p, Bni1p, or Iqg1p.
To further define which Cdc42p regulators and effectors were affected by the V44A mutation, suppression of the cdc42 V44A temperature sensitivity or morphological defect by proteins known to interact with Cdc42p was examined. High copy plasmids containing CDC42, BEM3, RGA1, CDC24, STE20, CLA4, or SKM1 under their endogenous promoters, or GIC2 under the ADH promoter were transformed into cdc42 V44A strain TRY5-3A, and transformants were streaked at 23°C and 37°C. Plasmids containing CDC42 completely complemented the cdc42 V44A morphological defect at 23°C and could also complement the temperature sensitive lethality. However, it could only partially complement the morphological phenotype at 37°C (data not shown), suggesting that the V44A mutation was semi-dominant at 37°C. Overexpression of the other seven proteins could not suppress the cdc42 V44A temperature-sensitive lethality (data not shown), and overexpression of Bem3p, Rga1p, Skm1p, or Gic2p did not affect the morphology of cdc42 V44A cells at 23°C. In contrast, overexpression of Cla4p from a high copy plasmid showed efficient suppression with 81% of the budded cells displaying a normal budded morphology. Effective suppression of the cdc42 V44A morphology was also seen with the low copy number plasmid pRS315(CLA4), but the level of abnormally budded cells was consistently higher with YEp351(CLA4) versus pRS315(CLA4), suggesting that suppression by CLA4 was dosage dependent. When Cdc24p was overexpressed in cdc42 V44A cells, 43% of the budded cells had a normal budded phenotype while 55% of the budded cells had a normal budded phenotype when Ste20p was overexpressed, suggesting that expression of these proteins only partially suppressed the mutant phenotypes. Together with the two-hybrid data, these results suggest that Cdc42 V44A p altered interactions with Cla4p are the primary contributor to the cdc42 V44A morphological defect.
To further examine the functional relationship between Cdc42 V44A p and Cla4p, a cdc42 V44A ⌬cla4 double mutant was tested for synthetic lethality. The cdc42 V44A ⌬cla4::trp1::HIS3 diploid strain TRY3-HV was sporulated and, of 95 spores (from 50 tetrads), none were Ura ϩ Trp ϩ His ϩ , suggesting that the cdc42 V44A ⌬cla4 double mutant is inviable at 23°C. To confirm this apparent synthetic lethality, a LEU2-based CDC42 plasmid, pRS315(CDC42), was transformed into TRY3-HV, and upon sporulation and tetrad dissection, three Ura ϩ Trp ϩ His ϩ Leu ϩ spores were isolated, indicating that a cdc42 V44A ⌬cla4 double mutant could be rescued by CDC42 on a plasmid. All three segregants were stably Leu ϩ , indicating that these spores required plasmid-borne CDC42 for survival and confirming that cdc42 V44A ⌬cla4 double mutant displayed synthetic lethality. Similar synthetic lethality was observed with the cdc42-1 ts mutant allele (28). These results indicate that these mutant Cdc42 proteins cannot function as the sole copy of Cdc42p in the cell without a functional Cla4p. Similar cdc42 V44A constructs were made with ⌬ste20, ⌬bem3, ⌬rga1, cdc3-1 ts , and cdc12-5 ts mutants, but none of these double mutants displayed synthetic lethality. DISCUSSION S. cerevisiae Cdc42p has been implicated in a number of different processes from bud emergence to pheromone response, but mostly through indirect information obtained from characterization of known Cdc42p interacting regulators and effectors. Insight into Cdc42p direct functions and interactions have been limited to characterization of mutants that primarily have defects that disrupt the polarity of the actin cytoskeleton and prevent bud emergence or affect the nucleotide-binding state of Cdc42p. The cdc42 V44A allele gave rise to an elongated multibudded, multinuclear phenotype, a phenotype that differed from other previously characterized cdc42 mutant alleles. Analysis of this novel effector domain mutant cdc42 V44A showed that Cdc42p was required for triggering the apicalisotropic growth switch and for timely progression through the G 2 /M transition.
Based on the altered two-hybrid interactions with Cla4p, the dosage-dependent suppression by CLA4 and the synthetic lethality of the cdc42 V44A ⌬cla4 double mutant, Cdc42 V44A p effects on morphology and the cell cycle were likely manifested through Cla4p. Data implicating Cdc42p in the regulation of FIG. 4. GFP-A 8 -Cdc42p localization. p416MET(GFP S65T -A 8 -CDC42) and p416MET(GFP S65T -A 8 -cdc42 V44A ) were transformed into TRY11-7D (wild type) and TRY1-9D (⌬cla4). Transformants were selected on SCϪUra plates, and then selected transformants were grown at 23°C to mid-log phase in SCϪUraϪMet liquid media for expression from the methionine-repressible promoter.
Cla4p kinase activity during the cell cycle have been reported previously. Cla4p kinase activity was found to be most active during mitosis, and when the Cla4p CRIB domain was deleted, kinase activity peaked closer to G 1 /START (29). Cla4p also has been shown to biochemically and genetically interact with the Gin4p and Nap1p kinases, which are thought to play a role in regulating the Clb2p/Cdc28p kinase-dependent G 2 /M transition (37,57,58). Therefore, Cdc42 V44A p altered interactions with Cla4p may be affecting the regulatory pathway that controls passage through the G 2 /M transition to cause a nuclear division delay that at least partially contributes to the mutant phenotype.
Recently, the septin proteins, among their many roles during the cell cycle, have also been implicated in regulating entry into mitosis (58,59), which suggests that the septin structural defects seen in cdc42 V44A may trigger the G 2 /M checkpoint. Septin ring formation occurs in the G 1 phase of the cell cycle ϳ15 min prior to bud emergence (56), and is required for proper chitin ring formation prior to bud emergence at the presumptive bud site, but is not required for bud emergence itself (38,54). Previously, cdc42 loss of function mutants were found to have delocalized chitin and were unable to localize or form a septin ring (38). The delocalized chitin deposition and lack of a chitin ring in 50% of cdc42 V44A cells suggested that Cdc42 V44A p is affecting the function or structure of the septin ring and the delocalization of GFP-Cdc3p and GFP-Cdc12p corroborated this hypothesis. The diffuse and delocalized GFP-Cdc3p and GFP-Cdc12p staining patterns in cdc42 V44A elongated-budded cells (Fig. 3) suggested that Cdc42 V44A p affected septin ring formation or maintenance but not localization of the septins to the bud. Therefore, along with its role in localizing the septins to the presumptive bud site, Cdc42p may play a role in the structural assembly and maintenance of the septin ring.
The correlation between the cdc42 V44A morphological defects and Cdc42p interactions with Cla4p and the aberrant ring structure seen in these mutant cells suggests that together Cdc42p and Cla4p regulate the septins. There are several lines of evidence that support a functional relationship between Cdc42p, Cla4p, and the septins. First, a ⌬cla4 mutant had defects in the maintenance of the septin ring to the mother-bud neck region as shown in Fig. 3 that were similar to the septin ring defects seen in cdc42 V44A cells. cla4 ts ste20 double mutant cells have also been shown to have similar defects in septin ring maintenance (28). Second, several genetic interactions between Cdc42p, Cla4p, and the septins have been established including synthetic lethality observed with ⌬cla4 cdc42 V44A and ⌬cla4 cdc42-1 double mutants as well as cla4 cdc12 double mutants (28). Further evidence to support this hypothesis is that Gin4p mitosis-specific phosphorylation is dependent on Cla4p whose mitosis-specific phosphorylation is dependent on Nap1p and Cdc42p (37). Gin4p kinase activity promotes septin organization and the septins interact with, and are dependent on Gin4p for proper ring structure (60). Therefore, Cdc42 V44A p reduced interaction with Cla4p may be affecting Cla4p functional association with Gin4p thereby interfering with septin ring structure. The ability of Cla4p and Cdc12p to suppress cdc42 V44A could be due to the overexpressed proteins localizing and functioning properly, allowing cells to progress through G 2 /M.
Interestingly, overexpression of GFP-Cdc12p did not suppress the ⌬cla4 elongated-bud phenotype, but was able to partially suppress the cdc42 V44A phenotype, suggesting that the cdc42 V44A mutant had a milder septin localization defect than the ⌬cla4 mutant. It should also be noted that the penetrance of the elongated-bud morphological defect was higher in cdc42 V44A cells (80% of budded cells) than in ⌬cla4 mutant cells (50% of budded cells). This difference in penetrance suggested that the cdc42 V44A phenotype was not solely the result of altered interactions with Cla4p. Two-hybrid analysis revealed that the V44A mutation, in addition to reducing interactions with Cla4p, affected interactions with Cdc24p (6) and effectors Gic1p and Gic2p (Fig. 5). The correlation between the reduced FIG. 5. Cdc42 V44A p two-hybrid protein interactions. ␤-Galactosidase liquid assays were performed in triplicate. These data are the averages of one triplicate assay for each interaction and are representative of at least four independent ␤-galactosidase liquid assays for each interaction.
interactions between Cdc42 V44A p and Gic1p or Gic2p is difficult to assess since the functions of Gic1p and Gic2p are not well defined and Gic2p, unlike Cla4p, was unable to suppress the cdc42 V44A morphological defect. However, gic1 gic2 double mutants appear to have defects in nuclear migration (27), suggesting that the reduced two-hybrid interactions seen between Cdc42 V44A p and Gic1p and Gic2p were contributing to the nuclear division delay seen in cdc42 V44A cells. Furthermore, genetic interactions between Gic1p, Gic2p, and Cla4p have been observed (27), suggesting that these proteins may have overlapping roles as Cdc42p effectors. Therefore, the reduced interactions between Cdc42 V44A p and Cla4p, Gic1p, and Gic2p are all likely contributing to the cdc42 V44A phenotype and may explain why the cdc42 V44A phenotype is more severe than the cla4 mutant phenotype. Taken together, these results suggested that the severity of the cdc42 V44A G 2 /M delay was not due to a specific alteration of Cdc42p function and/or protein interactions, but was the result of a combination of altered functions and interactions with several regulators and effectors including Cdc24p, Cla4p, Gic1p, and Gic2p.
The proper localization of the GFP-Cdc42 V44A fusion protein suggested further that the effects of the V44A mutation are most likely the result of altered protein interactions as opposed to a localization defect (Fig. 4). It is interesting to note that wild type GFP-A 8 -Cdc42p localized not only to the pre-bud site and the tips and sides of elongating buds but also to the mother-bud neck region in some larger budded cells, suggesting a function in cytokinesis and/or cell separation. The partial cytokinesis and cell separation defect in cdc42 V44A cells also supports a role for Cdc42p at later stages in the cell cycle although what that function is remains unclear. GFP-A 8 -Cdc42p also localized properly regardless of the presence of Cla4p, suggesting that Cdc42p localization is not dependent Cla4p.
The cdc42 V44A ⌬swe1 phenotype (Fig. 2) is consistent with Cdc42 V44A p activating the G 2 /M morphogenetic checkpoint that controls the apical-isotropic growth switch and entry into mitosis. Swe1p, the S. cerevisiae homolog of the Schizosaccharomyces pombe CDK inhibitory kinase Wee1p, phosphorylates the Cdc28p cyclin-dependent kinase on residue Y19 (61, 62), thereby inhibiting Cdc28p kinase activity and resulting in a G 2 phase cell-cycle delay. Swe1p is required for the morphogenetic checkpoint, and monitors the actin cytoskeleton and possibly other structural components required during the budding process (51,53,63), and ⌬swe1 mutants no longer delay at G 2 phase. The aberrant phenotype associated with the cdc42 V44A ⌬swe1 double mutant suggested that ⌬swe1 suppressed the elongated-budded phenotype characteristic of a G 2 /M delay but did not alleviate the aberrant multibudded and multinucleate defects associated with Cdc42 V44A p, suggesting that the G 2 /M delay is necessary to repair or compensate for the effects of Cdc42 V44A p in order to allow the cell to proceed through the checkpoint. However, since some cdc42 V44A cells are multibudded and multinucleate even post-glusulase treatment, which indicated that a percentage of the cells do have a cytokinesis defect, the G 2 /M delay may not always be sufficient for alleviating Cdc42 V44A p effects.
Actin depolarization prior to relocalization to the mother-bud neck region for cytokinesis is characteristic of the apical-isotropic growth switch that occurs during the G 2 /M transition and requires activation of the Clb2p/Cdc28p kinase complex (51,52). However, the cdc42 V44A G 2 /M delay did not seem to be the result of a depolarized cytoskeleton or a loss of bud emergence. Rather, proper actin polarization and bud emergence in cdc42 V44A cells suggested that Cdc42 V44A p exerts some other cellular effect prior to the G 2 /M transition that is initiating the G 2 /M morphogenetic checkpoint. Since one of the mechanistic defects observed in cdc42 V44A cells is in septin ring structure, passage through the G 2 /M checkpoint in these cells may be regulated through the septins. If this is the case, then cdc42 V44A cells that were able to form a stable septin ring during the mitotic delay may have progressed through the checkpoint allowing for viability, while those cells that never formed a stable ring did not recover from the checkpoint resulting in the cytokinesis defect seen in some cells. The septin proteins have recently been implicated in regulating the Swe1p morphogenetic checkpoint pathway through their interactions with Hsl1p, a Nim1p-like protein that negatively regulates Swe1p, and Hsl1p homologs Kcc1p and Gin4p (59). ⌬swe1 also suppressed the cdc12 elongated-bud phenotype and the ⌬swe1 cdc12 double mutant had similar morphological and nuclear phenotypes to those seen in cdc42 V44A ⌬swe1 cells (59). Therefore, Cdc42p-dependent regulation of septin organization may be required for progressing through the G 2 /M morphogenetic checkpoint.
The data presented here strongly suggest that Cdc42p is required for timely cell cycle progression and is dependent on differential interactions with its regulators and downstream effectors for function. The differential interactions seen between Cdc42 V44A p and various regulators and effectors most likely contributed to the severe morphological defects associated with cdc42 V44A . Surprisingly, this mutant is still viable, suggesting that Cdc42p functioning with other proteins in the cell can compensate for these altered interactions. The characterization of cdc42 V44A also suggested that Cdc42p has a direct role in septin ring structural maintenance that is mediated by Cla4p. What still remains unclear is how Cdc42p signals are passaged through its various interactions with regulators and downstream effectors to structural cellular components like actin and the septins. Future characterization of cdc42 mutants that mechanistically disrupt interactions with various regulators, effectors and structural components of the cell polarity pathway should be useful in discerning the Cdc42p-dependent signals required for bud emergence, the G 2 /M transition and cytokinesis.