Portal de Periódicos da CAPES

The Cdc42p effector Gic2p is targeted for ubiquitin-dependent degradation by the SCFGrr1 complex

1998; Springer Nature; Volume: 17; Issue: 18 Linguagem: Inglês

10.1093/emboj/17.18.5360

1460-2075

Malika Jaquenoud, Marie‐Pierre Gulli, Katrin Peter, Matthias Peter,

Cancer-related Molecular Pathways

Article15 September 1998free access The Cdc42p effector Gic2p is targeted for ubiquitin-dependent degradation by the SCFGrr1 complex Malika Jaquenoud Malika Jaquenoud Swiss Institute for Experimental Cancer Research (ISREC), Ch. des Boveresses 155, 1066 Epalinges/VD, Switzerland Search for more papers by this author Marie-Pierre Gulli Marie-Pierre Gulli Swiss Institute for Experimental Cancer Research (ISREC), Ch. des Boveresses 155, 1066 Epalinges/VD, Switzerland Search for more papers by this author Katrin Peter Katrin Peter Present address: Institute of Biochemistry, University of Lausanne, 1066 Epalinges/VD, Switzerland Search for more papers by this author Matthias Peter Corresponding Author Matthias Peter Swiss Institute for Experimental Cancer Research (ISREC), Ch. des Boveresses 155, 1066 Epalinges/VD, Switzerland Search for more papers by this author Malika Jaquenoud Malika Jaquenoud Swiss Institute for Experimental Cancer Research (ISREC), Ch. des Boveresses 155, 1066 Epalinges/VD, Switzerland Search for more papers by this author Marie-Pierre Gulli Marie-Pierre Gulli Swiss Institute for Experimental Cancer Research (ISREC), Ch. des Boveresses 155, 1066 Epalinges/VD, Switzerland Search for more papers by this author Katrin Peter Katrin Peter Present address: Institute of Biochemistry, University of Lausanne, 1066 Epalinges/VD, Switzerland Search for more papers by this author Matthias Peter Corresponding Author Matthias Peter Swiss Institute for Experimental Cancer Research (ISREC), Ch. des Boveresses 155, 1066 Epalinges/VD, Switzerland Search for more papers by this author Author Information Malika Jaquenoud1, Marie-Pierre Gulli1, Katrin Peter2 and Matthias Peter 1 1Swiss Institute for Experimental Cancer Research (ISREC), Ch. des Boveresses 155, 1066 Epalinges/VD, Switzerland 2Present address: Institute of Biochemistry, University of Lausanne, 1066 Epalinges/VD, Switzerland *Corresponding author. E-mail: [email protected] The EMBO Journal (1998)17:5360-5373https://doi.org/10.1093/emboj/17.18.5360 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info Cdc42p, a Rho-related GTP-binding protein, regulates cytoskeletal polarization and rearrangements in eukaryotic cells. In yeast, Gic1p and Gic2p are effectors of Cdc42p involved in actin polarization at bud emergence. Gic2p is expressed in a cell cycle-dependent manner and rapidly disappears shortly after bud emergence concomitant with the activation of the G1 cyclin-dependent kinase Cdc28p–Clnp. Here we have shown that the rapid disappearance of Gic2p results from ubiquitin-dependent proteolysis. Biochemical and genetic evidence demonstrates that degradation of Gic2p required the Skp1–cullin–F-box protein complex (SCF) components Cdc34p, Cdc53p, Skp1p and Grr1p, but not Cdc4p. Phosphorylation of several C-terminal sites of Gic2p served as part of the recognition signal for ubiquitination. In addition, binding of Gic2p to Cdc42p was a prerequisite for degradation, suggesting that specifically the active form of Gic2p is targeted for destruction. Finally, our data indicate that degradation of Gic2p may be part of a mechanism which restricts cytoskeletal polarization in the G1 phase of the cell cycle. Introduction The actin cytoskeleton mediates many biological functions in all eukaryotic cells, such as maintenance of cell shape and polarity, cell motility, cytokinesis and phagocytosis (Hall, 1998). Members of the Rho family of small GTPases including Rho, Rac and Cdc42 have emerged as key regulators that control the organization and dynamics of the actin cytoskeleton and the assembly of focal adhesion contacts (Burridge et al., 1997; Hall, 1998; Keely et al., 1998). In addition, these GTPases play important roles in the establishment of cell polarity (Drubin and Nelson, 1996) and may be directly involved in axon guidance (Gallo and Letourneau, 1998). The biological effects of GTPases are mediated through their interaction with multiple target proteins, which ensure coordinated control of the actin cytoskeleton with other cellular activities such as gene transcription and adhesion. In yeast, activation of Cdc42p by its exchange factor Cdc24p is required to organize the actin cytoskeleton towards the incipient bud site during vegetative growth and towards the pheromone-secreting partner during mating (Chant, 1996). Cytoskeletal polarization also requires Bem1p, a protein with two SH3 domains which is thought to function as an adaptor for Cdc42p and Cdc24p (Chenevert et al., 1990; Leeuw et al., 1995). In the absence of Cdc42p function, cells fail to grow in a polarized manner and instead increase in size isotropically (Adams et al., 1990; Johnson and Pringle, 1990). Several effectors of Cdc42p recently have been identified, including the PAK kinases Ste20p, Cla4p and Smk1p (Cvrckova et al., 1995; Peter et al., 1996; Leberer et al., 1997a; Martin et al., 1997; Ebi et al., 1998), the formins Bni1p and Bnr1p (Evangelista et al., 1997; Imamura et al., 1997), and Gic1p and Gic2p (Brown et al., 1997; Chen et al., 1997). None of these proteins alone can account for the effect of Cdc42p on polarized growth, suggesting that Cdc42p might affect the cytoskeleton by interacting with multiple targets. The actin cytoskeleton is highly dynamic during all stages of the yeast life cycle and is regulated by both internal and external signals (Lew and Reed, 1995; Leberer et al., 1997b). During mating, pheromones activate a signal transduction cascade leading to polarization and projection formation. In contrast, activation of the cyclin-dependent kinase (CDK) Cdc28p–Clnp during the G1 phase of the cell cycle triggers polarization of the cytoskeleton resulting in bud emergence. In G2, activation of the Cdc28p–Clbp kinase inactivates polarized growth, leading to uniform expansion of the bud. Finally, in mitosis, actin is redistributed to form a cleavage ring in the middle of the two separating cells. How these different cytoskeletal rearrangements are regulated in response to extracellular signals or by the cell cycle machinery is poorly understood. Regulation may alter the activity of the Cdc42p GTPase module and/or may be mediated through one or several downstream effectors. Ubiquitin-dependent degradation of key components has emerged as an important mechanism controlling multiple cell cycle transitions. For example, anaphase is initiated by the destruction of Pds1p, an inhibitor of sister chromosome separation (Cohen-Fix et al., 1996; Michaelis et al., 1997). Degradation of the cyclin-dependent kinase inhibitor (CKI) Sic1p activates Cdc28p–Clbp kinase which in turn initiates DNA replication (Schwob et al., 1994). At least two distinct ubiquitination systems trigger these events: Pds1p and mitotic cyclins are degraded by the anaphase-promoting complex (APC; King et al., 1996), whereas Sic1p is ubiquitinated by a complex termed SCF (Skp1–cullin–F-box protein complex; Deshaies, 1997; Feldman et al., 1997; Skowyra et al., 1997; Patton et al., 1998a). Homologues of several APC and SCF components have been identified in a variety of systems and, like in yeast, they have been shown to function in ubiquitin-dependent degradation of proteins involved in cell cycle control and other biological processes (Krek, 1998). SCF complexes are ubiquitin ligases which interact with both a ubiquitin-conjugating enzyme and substrates (Hershko, 1997; Patton et al., 1998a). Several substrates need to be phosphorylated for subsequent ubiquitination, and it is thought that phosphorylation of the substrate constitutes at least part of the recognition signal which mediates binding to the SCF complex (Willems et al., 1996; Feldman et al., 1997; Henchoz et al., 1997; Skowyra et al., 1997; Verma et al., 1997). SCFCdc4 is composed of Cdc53p, Cdc34p, Skp1p and Cdc4p, and is necessary and sufficient for the ubiquitination of Sic1p (Feldman et al., 1997; Skowyra et al., 1997). It is also involved in the degradation of Far1p (Henchoz et al., 1997), Cdc6p (Drury et al., 1997), Ctf13p (Kaplan et al., 1997) and possibly Gcn4p (Kornitzer et al., 1994). CDC53 encodes a cullin (Willems et al., 1996; Mathias et al., 1996) whereas Cdc34p shows homology to ubiquitin-conjugating enzymes (Goebl et al., 1988). Cdc4p is a member of the growing family of proteins containing an F-box, a small sequence motif which has been shown to mediate the interaction with Skp1p (Bai et al., 1996; Lisztwan et al., 1998). In addition to Cdc4p, yeast cells express several other F-box proteins including Grr1p and Met30p (Bai et al., 1996; Patton et al., 1998a). Cells lacking Grr1p exhibit severe morphological abnormalities and defects in glucose repression (Flick and Johnston, 1991), whereas cells deleted for MET30 are unable to grow in the absence of methionine (Thomas et al., 1995). Both Grr1p and Met30p are able to form complexes with Skp1p, Cdc34p and Cdc53p (Li and Johnston, 1997; Skowyra et al., 1997, Patton et al., 1998b). However, whereas Grr1p is required for the degradation of the G1 cyclins Cln1p and Cln2p, it is dispensable for the degradation of Sic1p (Barral et al., 1995; Patton et al., 1998b), suggesting that the F-box proteins provide substrate specificity to the SCF complex. To understand the function of individual F-box proteins, it is crucial to identify the substrates which they target for destruction. Here we show that Gic2p degradation at bud emergence is dependent on the SCFGrr1 complex, and is regulated by specific phosphorylation which requires active, GTP-bound Cdc42p. Cells overexpressing an undegradable Gic2 protein exhibit severe morphological defects in the G1 phase of the cell cycle, suggesting that after the Gic proteins have polarized the actin cytoskeleton their subsequent destruction is needed for efficient bud emergence. Taken together, our results suggest that ubiquitin-dependent degradation of the Gic proteins coordinates polarization of the actin cytoskeleton in the G1 phase of the yeast cell cycle. Results Cell cycle-dependent expression of Gic2p requires ubiquitin-dependent degradation by the SCFGrr1 complex Gic2p is expressed in a cell cycle-dependent manner and accumulates throughout G1, reaching maximal levels just before bud emergence (Brown et al., 1997). Gic2p abruptly disappears shortly after cells have polarized their cytoskeleton and is virtually undetectable during S and G2 phases (Figure 1A; Brown et al., 1997). The expression pattern of Gic2p is regulated in part at the level of transcription: expression of GIC2 mRNA is maximal at bud emergence but is absent at later stages of the cell cycle (data not shown; R.Tabtiang, unpublished results). To test whether regulated transcription is responsible for the abrupt disappearance of Gic2p, we placed Gic2p under the control of the constitutive CYC1 promoter. No obvious growth or morphological differences were apparent between Δgic2 cells expressing Gic2p from the constitutive CYC1 promoter or from the endogenous promoter, suggesting that transcriptional regulation of GIC2 was not essential for cell cycle progression and morphogenesis (data not shown). Moreover, both strains initiated bud formation ∼30 min after release from an α-factor block (Figure 1B). However, even when GIC2 was expressed constitutively, Gic2 protein levels still decreased shortly after bud emergence (Figure 1A, lower panel), demonstrating that post-transcriptional mechanism(s) downregulate Gic2 protein levels. In contrast, Gic2p was present later in the cell cycle only when expressed from constitutive promoters, suggesting that a transcriptional mechanism may regulate the levels of Gic2p during S and G2 (Figure 1A). Figure 1.Gic2p protein levels are regulated by ubiquitin-dependent degradation which requires components of the SCFGrr1 complex. (A and B) Δgic2 cells expressing Gic2p either from the endogenous (upper panel) or the constitutive CYC1 promoter (lower panel) were released from α-factor arrest and Gic2p levels were determined by immunoblotting at the times indicated (minutes after washing out α-factor). The asterisk marks the position of a protein which cross-reacts with the Gic2p antiserum. Cell cycle entry was monitored microscopically by analysing the percentage of unbudded cells (B; ○, endogenous promoter; ▪, constitutive CYC1 promoter). (C) Gic2p protein levels (upper panel) were determined by immunoblotting in strains harbouring defects in ubiquitin-dependent degradation. Lane 1, skp1-11; lane 2, skp1-12; lane 3, cdc4-1; lane 4, cdc34-2; lane 5, cdc53-1; lane 6, cim3-1; lane 7, Δgrr1; lane 8, wild-type. Immunoblotting against actin was used as a loading control (lower panel). The arrow marks the position of Gic2p; the arrowhead points to the slower migrating form of Gic2p which accumulates in some mutants. (D) Immunoblot demonstrating that Gic2p accumulates as a ladder of high molecular forms (bracket) in mutants defective in proteasome function. The following strains were analysed: lane 1, cim3-1; lane 2, cim5-1; lane 3, sen3-1; lane 4, wild-type; lane 5, Δgic2. (E and F) Accumulation of Gic2p is independent of the cell cycle arrest point. Gic2p levels in cdc4-1, cdc34-2 and cdc53-1 mutants (lanes 3, 5 and 7) were compared with accumulation of Gic2p in cdc4-1, cdc34-2 and cdc53-1 mutants lacking SIC1 (lanes 4, 6 and 8). FACS analysis confirming that cdc53-1 mutant cells arrest with a G1 DNA content (left panel) whereas cdc53-1 sic1Δ cells replicate their DNA and arrest in G2 (right panel). Download figure Download PowerPoint To examine how Gic2p is down-regulated at bud emergence, we first determined Gic2p levels in cells lacking components involved in ubiquitin-dependent degradation (Figure 1C). Interestingly, we found that Gic2p levels were dramatically increased in cdc34-2, cdc53-1 cells and cells deleted for GRR1. Moreover, a characteristic slower migrating Gic2p species became apparent in all of these mutants. Gic2p accumulated in skp1 cells in an allele-dependent manner: Gic2p levels were high in skp1-12 cells which arrest in G2, but were low in cells harbouring a G1-specific allele of SKP1 (skp1-11). In contrast, low levels and no slower migrating forms of Gic2p were present in cdc4-1 mutants which arrest at the same stage of the cell cycle as cdc34-2 and cdc53-1 cells (Figure 1C). Thus, we conclude that decreasing the levels of Gic2p at the time of bud emergence requires the SCFGrr1 but not the SCFCdc4 complex. To exclude further the possibility that cell cycle arrest rather than a direct involvement of these proteins accounted for the accumulation of Gic2p, we compared Gic2p levels in cdc53-1 and cdc34-2 cells, which arrest in the G1-phase of the cell cycle, and cdc53-1 Δsic1 and cdc34-2 Δsic1 strains which replicate their DNA and arrest in G2 (Figure 1F; data not shown; Schwob et al., 1994). Clearly, Gic2p accumulated in both strains (Figure 1E), demonstrating that Cdc53p and Cdc34p are directly involved in controlling Gic2p levels around bud emergence. Finally, elevated levels of Gic2p were also apparent in cim3, cim5 and sen3 cells which arrest at the G2–M boundary because of a defect in the 26S proteasome (Figure 1D; Ghislain et al., 1993; Hilt and Wolf, 1995). We conclude from these results that Gic2p degradation requires a functional proteasome. Importantly, Gic2p accumulated in these proteasome mutants as a ladder of high molecular weight species distinct from the characteristic slower migrating form (Figure 1D). We believe that these high molecular species represent ubiquitinated forms of Gic2p, although at present we are unable to demonstrate this directly. To determine whether the SCFGrr1 complex was required to degrade Gic2p, we compared the stability of Gic2p in cdc4-1, Δgrr1 rgt1 and cdc34-2 mutants (Figure 2A). We used the Δgrr1 rgt1 strain because the rgt1 mutation restores efficient glucose repression in cells lacking Grr1p (Barral et al., 1995). Gic2p was expressed from the GAL1 promoter, at which time synthesis was turned off by addition of glucose. Interestingly, Gic2p was stable throughout the time course and accumulated in the slower migrating form in Δgrr1 rgt1 and cdc34-2 cells (half-life >60 min). In contrast, rapid degradation of Gic2p was observed in cdc4-1 cells (half-life ∼10 min), demonstrating that Grr1p and Cdc34p but not Cdc4p are required for Gic2p degradation. Consistent with these results, we found that overexpression of Gic2p is lethal in Δgrr1 rgt1 or cdc34-2 cells grown at the semi-permissive temperature, whereas cdc4-1 cells overexpressing Gic2p grew normally (Figure 2B). Moreover, Gic2p tagged with the green fluorescent protein (GFP) was detected readily as a crescent at bud tips of cdc34-2 cells arrested at 37°C (Figure 2C, bottom panel), whereas Gic2p–GFP was degraded rapidly and was virtually undetectable in cdc4-1 cells (top panel). Taken together, these results strongly suggest that the stability of Gic2p is regulated by the SCFGrr1 complex. Figure 2.Gic2p is stabilized in Δgrr1 rgt1 and cdc34-2, but not in cdc4-1 mutant cells. (A) Cells carrying a plasmid coding for Gic2p from the inducible GAL promoter were grown in raffinose and expression was induced by the addition of galactose for 3 h (see Materials and methods). Glucose was then added to shut off the GAL promoter, and samples were taken every 10 min as indicated and immunoblotted for the presence of Gic2p. Equal loading was confirmed by immunoblotting the samples with antibodies against actin (lower panels). The following strains were analysed: top panel, cdc4-1 Δgic2; middle panel, Δgrr1 rgt1Δgic2; bottom panel, cdc34-2 Δgic2. The rgt1 mutation suppresses the glucose repression defect of Δgrr1 strains. (B) Overexpression of wild-type Gic2p from the GAL promoter is lethal in Δgrr1 rgt1 and cdc34-2 cells at the semipermissive temperature (30°C), whereas cdc4-1 cells tolerate overexpression of Gic2p. All strains grew normally when plated on medium containing glucose where expression of Gic2p was not induced (data not shown). (C) Overexpressed wild-type Gic2p fused to GFP is visualized readily at bud tips in arrested cdc34-2 cells (bottom panel), but is rapidly degraded and virtually undetectable in cdc4-1 cells (top panel). Photographs show GFP fluorescence superimposed with phase contrast. Download figure Download PowerPoint Phosphorylation of Gic2p is required for ubiquitin-mediated degradation Ubiquitin-dependent degradation of Sic1p, Far1p and the G1 cyclins requires phosphorylation of the substrate (Lanker et al., 1996; Schneider et al., 1996; Feldman et al., 1997; Henchoz et al., 1997; Skowyra et al., 1997; Verma et al., 1997). As noted earlier, Gic2p accumulated as slower migrating forms in cells lacking any component of the SCFGrr1 complex. To determine whether these slower migrating forms represented phosphorylated Gic2p, we incubated Gic2p isolated from cdc34-2 cells with alkaline phosphatase (calf intestine phosphatase; CIP). As shown in Figure 3A, the slower migrating forms of Gic2p were converted quantitatively to the faster migrating form (lane 4), demonstrating that Gic2p was indeed phosphorylated. No downshift was observed if CIP was inactivated by boiling (lane 5) or by the addition of phosphatase inhibitors (data not shown). Figure 3.Cdc28p–Clnp-dependent phosphorylation of Gic2p is required for degradation. (A) Gic2p accumulates in a phosphorylated form in cdc34-2 cells. Total cell extracts (T.E.) prepared from cdc34-2 cells (lanes 2–5) were incubated with purified calf intestine phosphatase (CIP; lane 4), buffer (lane 3), or CIP which was inactivated prior to addition by boiling for 5 min (lane 5). Note that CIP quantitatively converts the slower migrating forms of Gic2p (arrowhead) to the unphosphorylated form present in cdc28-4 cells (lane 1; arrow). (B) Phosphorylation and accumulation of Gic2p depends on the presence of Cdc28p–Clnp kinase. Gic2p levels were analysed by immunoblotting in cells lacking Cdc28p function (lanes 3–5) or various cyclins (lanes 6–8). The following strains were analysed: lane 1, wild-type (K699); lane 2, cdc34-2 (YMT670); lane 3, cdc28-1 (MJ254); lane 4, cdc28-4 (YMP599); lane 5, cdc28-1N (YML50), lane 6, ΔΔΔcln1,2,3 (EY569); lane 7, ΔΔclb5,6 (MJ202), lane 8, ΔΔΔΔΔΔclb1,2,3,4,5,6 (MJ207). The arrowhead marks the position of phosphorylated Gic2p; the arrow points towards the position of unphosphorylated Gic2p. Equal loading of the lanes was confirmed by immunoblotting actin (lower panel). (C) MJ270 were arrested at 25°C by depletion of the G1 cyclins CLN1, 2 and 3, and released by addition of galactose to induce expression of Cln3p. The culture simultaneously was shifted to 37°C to inactivate Cdc34p. Samples were collected at the times indicated and immunoblotted with antibodies against Gic2p (lanes 2–8). After 90 min at 37°C, glucose was added to half of the culture to repress Cln3p expression. Aliquots were removed at the times indicated (0′ represents addition of glucose) and analysed by immunoblotting with Gic2p antibodies (lanes 9–15). Note that maintenance of Gic2p phosphorylation depends on the presence of Cln3p. (D) Degradation of Gic2p depends on the presence of Clnp. Cells were arrested as in (C) and released from the cdc34-2 block by shifting the temperature to 25°C either in the presence (Cln3p expressed, upper panel) or absence (Cln3p repressed, lower panel) of galactose. Samples were removed at the times indicated and immunoblotted with Gic2p antibodies. Time 0 indicates release of the culture from the cdc34-2 block. The immunoblots were quantified by phosphoimager and plotted against time in minutes after release. ▪, +galactose (Cln3p expressed); ♦, −galactose (Cln3p repressed). Download figure Download PowerPoint Interestingly, Gic2p accumulated in an unphosphorylated form in cells lacking a functional Cdc28p–Clnp kinase, suggesting that either the Cdc28p–Clnp kinase itself or a kinase dependent on Cdc28p–Clnp was responsible for Gic2p phosphorylation (Figure 3A, lane 1, and B). No accumulation of Gic2p was observed in strains lacking G2 cyclins (Figure 3B, lanes 7 and 8) or in strains harbouring the cdc28-1N allele (which arrests cells in the G2 phase of the cell cycle; Figure 3B, lane 5), indicating that Gic2p was phosphorylated efficiently. To corroborate these results further, we used a cdc34 strain deleted for CLN1, CLN2 and CLN3 harbouring CLN3 under the control of the inducible GAL promoter (Schneider et al., 1996). As expected, Gic2p accumulated in the unphosphorylated form when these cells were arrested by the addition of glucose to turn off expression of Cln3p (Figure 3C, lane 2), even when Cdc34p was inactivated simultaneously by shifting the culture to the restrictive temperature. Induction of Cln3p by the addition of galactose resulted in rapid phosphorylation of Gic2p (Figure 3C, lanes 3–8), consistent with the observation that the G1 cyclins are required for Gic2p phosphorylation. Subsequent repression of Cln3p by addition of glucose for 3 h led to dephosphorylation of Gic2p and its accumulation in the unphosphorylated form (Figure 3C, lanes 9–15), demonstrating that the continuous presence of the G1 cyclins was needed to keep Gic2p in the phosphorylated state. These findings enabled us to ask whether phosphorylation was required for degradation of Gic2p in vivo (Figure 3D). Cells grown in galactose medium at 25°C were arrested at the cdc34 arrest point by shifting the culture to the restrictive temperature. The culture was then divided and glucose was added to one half to repress Cln3p expression. After 2 h, Cdc34p was activated by shifting the temperature to 25°C and degradation of Gic2p was monitored by immunoblotting. Interestingly, Gic2p was degraded rapidly when Cln3p was present, but Gic2p was stable when Clnp activity was depleted. Thus, these results suggest that Clnp-dependent phosphorylation of Gic2p is required for subsequent ubiquitin-dependent degradation. To substantiate further the link between phosphorylation and degradation, we attempted to identify the phosphorylation sites on Gic2p. We found that a truncated Gic2p lacking the C-terminal domain (Gic2p-N1–208) was not fully phosphorylated in cdc34-2 mutant cells (Figure 4A, lane 4). Importantly, half-life measurements of Gic2pN1–208 in cdc4-1 mutant cells confirmed that the protein was stable (half-life ∼60 min; Figure 4B), suggesting that the critical phosphorylation sites involved in Gic2p degradation are located in the C-terminal domain of the protein. Sequence alignment of Gic1p and Gic2p revealed two closely spaced, conserved serine residues in this domain (serines 254 and 258 of Gic2p); both residues are followed by a proline, and hence conform to the minimal consensus site for both CDKs and MAP kinases (Nigg, 1995). We therefore mutated the two putative phosphorylation sites to non-phosphorylatable alanine residues and examined the phosphorylation state of the Gic2-S254A/S258A protein in cdc34-2 mutant cells. As shown in Figure 4A, the fully phosphorylated form of Gic2p was absent, suggesting that S254 and/or S258 are phosphorylated in vivo. However, S254 and S258 are not the only phosphorylation sites, because mutant Gic2p still migrated more slowly than unphosphorylated Gic2p. Nevertheless, Gic2p-S254A/S258A was significantly stabilized in cdc4-1 mutants with a half-life of 25 min compared with <10 min for wild-type Gic2p (Figure 4B). Consistent with these findings, we found that expression of either Gic2p-S254A/S258A or Gic2p-N1–208 was lethal in wild-type (data not shown) or cdc4-1 cells (Figure 4C); however, we cannot rigorously exclude that Gic2p-N1–208 may exhibit some additional gain-of-function effect unlinked to degradation which could contribute to the lethal phenotype. Taken together, these results suggest that phosphorylation of Ser254 and/or Ser258 is involved in degradation of Gic2p, supporting the notion that phosphorylation of Gic2p triggers its ubiquitin-dependent degradation. Figure 4.Phosphorylation of C-terminal phosphorylation sites triggers ubiquitination of Gic2p in vivo. (A) Phosphorylation of wild-type and mutant forms of Gic2p expressed in cdc34-2 (lanes 1–4) or for control cdc4-1 cells (lane 5) was analysed by Western blotting with Gic2-specific antibodies. The following Gic2p mutants were analysed: lane 1, wild-type; lane 2, Gic2p-S254A/S258A; lane 3, Gic2p-crib−; lanes 4 and 5, Gic2p-N1–208. (B) The half-life of wild-type and Gic2p mutant proteins was analysed in cdc4-1 cells as described in the legend to Figure 2 and quantified by phosphoimager. ▪, Gic2p-N1–208; ▴, Gic2p-S254A/S258A; ♦, wild-type. (C) Overexpression of either phosphorylation mutant of Gic2p (S254A S258A or N1–208) from the GAL promoter is lethal in cdc4-1 cells incubated at 30°C. In contrast, cells harbouring a control plasmid (vect.) or overexpressing wild-type Gic2p (WT) grew normally. All strains grew normally when plated on medium containing glucose (data not shown). (D) Phosphorylation of Gic2p is required for ubiquitination. Cim3-1 Δgic2 cells (MJ416) were transformed with plasmids expressing various Gic2 mutant proteins and grown at 37°C for 3 h. Extracts were prepared after 3 h and immunoblotted with polyclonal Gic2p antibodies. The following Gic2 proteins were analysed: lane 1, wild-type; lane 2, Gic2p-S254A/S258A; lane 3, Gic2p-crib−; lane 4, Gic2p-N1–208; lane 5, vector. Note that only wild-type Gic2p accumulates as a characteristic ladder of high molecular forms indicative of ubiquitination (bracket). Download figure Download PowerPoint Phosphorylation is thought to promote binding of the substrate to the SCF complex and, therefore, phosphorylation should precede ubiquitination. To test this prediction, we determined whether Gic2p mutants defective in phosphorylation would still be ubiquitinated in vivo (Figure 4D). To facilitate detection of ubiquitin conjugates, we took advantage of the proteasome-deficient cim3 cells which accumulated slower migrating, most likely ubiquitin-conjugated forms of Gic2p (Figure 1D). Interestingly, no slower migrating forms of Gic2p-N1–208 could be detected after expression in cim3 cells (Figure 4D, lane 4) and, similarly, ubiquitination of Gic2p-S254A/S258A was significantly reduced (lane 2). Thus, the defects of these Gic2 mutant proteins correlated with their defects in phosphorylation and degradation. We conclude from these results that phosphorylation of Gic2p is required for ubiquitination in vivo. Phosphorylation and degradation of Gic2p depends on its ability to bind Cdc42p Previous studies have suggested that the Gic proteins function in a complex with Cdc42p in its active, GTP-bound form (Brown et al., 1997; Chen et al., 1997). To determine whether Cdc42p is involved in the degradation of Gic2p, we measured Gic2p levels in mutants lacking Cdc42p. Interestingly, we found that Gic2p accumulated in cdc42-1 cells (Figure 5A, lane 3), indicating that Cdc42p is necessary for degradation. In contrast, low levels of Gic2p were detected in cells lacking Bem1p (Figure 5A, lane 1) or the Rho1p GTPase (Figure 5A, lane 4). Increased levels of Gic2p were also observed in cdc24-1 mutant cells (Figure 5A, lane 2), suggesting that specifically the GTP-bound form of Cdc42p is involved in the degradation of Gic2p. Gic2p accumulated in an unphosphorylated form in both cdc42 and cdc24 mutants, indicating that Cdc42p-GTP is required for phosphorylation of Gic2p. Taken together, these results indicate that the GTP-bound form of Cdc42p is required for phosphorylation and degradation of Gic2p. Figure 5.Binding of Cdc42p is required for phosphorylation and degradation of Gic2p. (A) Gic2p accumulates in an unphosph