The nucleotide exchange factor Cdc24p may be regulated by auto-inhibition
2004; Springer Nature; Volume: 23; Issue: 5 Linguagem: Inglês
10.1038/sj.emboj.7600124
ISSN1460-2075
AutoresYukiko Shimada, Philippe Wiget, Marie‐Pierre Gulli, Efrei Bi, Matthias Peter,
Tópico(s)Genetics and Neurodevelopmental Disorders
ResumoArticle26 February 2004free access The nucleotide exchange factor Cdc24p may be regulated by auto-inhibition Yukiko Shimada Yukiko Shimada Swiss Institute for Experimental Cancer Research (ISREC), SwitzerlandPresent address: Department of Biochemistry, University of Geneva, 8093 Geneva, Switzerland Search for more papers by this author Philippe Wiget Philippe Wiget Swiss Institute for Experimental Cancer Research (ISREC), Switzerland Swiss Federal Institute of Technology Zurich (ETH), Institute of Biochemistry, Zurich, Switzerland Search for more papers by this author Marie-Pierre Gulli Marie-Pierre Gulli Swiss Institute for Experimental Cancer Research (ISREC), Switzerland Search for more papers by this author Efrei Bi Efrei Bi Department of Cell and Developmental Biology, University of Pennsylvania, Philadelphia, PA, USA Search for more papers by this author Matthias Peter Corresponding Author Matthias Peter Swiss Federal Institute of Technology Zurich (ETH), Institute of Biochemistry, Zurich, Switzerland Search for more papers by this author Yukiko Shimada Yukiko Shimada Swiss Institute for Experimental Cancer Research (ISREC), SwitzerlandPresent address: Department of Biochemistry, University of Geneva, 8093 Geneva, Switzerland Search for more papers by this author Philippe Wiget Philippe Wiget Swiss Institute for Experimental Cancer Research (ISREC), Switzerland Swiss Federal Institute of Technology Zurich (ETH), Institute of Biochemistry, Zurich, Switzerland Search for more papers by this author Marie-Pierre Gulli Marie-Pierre Gulli Swiss Institute for Experimental Cancer Research (ISREC), Switzerland Search for more papers by this author Efrei Bi Efrei Bi Department of Cell and Developmental Biology, University of Pennsylvania, Philadelphia, PA, USA Search for more papers by this author Matthias Peter Corresponding Author Matthias Peter Swiss Federal Institute of Technology Zurich (ETH), Institute of Biochemistry, Zurich, Switzerland Search for more papers by this author Author Information Yukiko Shimada1, Philippe Wiget1,2, Marie-Pierre Gulli1, Efrei Bi3 and Matthias Peter 2 1Swiss Institute for Experimental Cancer Research (ISREC), Switzerland 2Swiss Federal Institute of Technology Zurich (ETH), Institute of Biochemistry, Zurich, Switzerland 3Department of Cell and Developmental Biology, University of Pennsylvania, Philadelphia, PA, USA *Corresponding author. Swiss Federal Institute of Technology Zurich (ETH), Institute of Biochemistry, ETH Hoenggerberg HPM G 6.2, 8093 Zurich, Switzerland. Tel.: +41 1 632 3134; Fax: +41 1 632 1269; E-mail: [email protected] The EMBO Journal (2004)23:1051-1062https://doi.org/10.1038/sj.emboj.7600124 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Site-specific activation of the Rho-type GTPase Cdc42p by its guanine-nucleotide exchange factor (GEF) Cdc24p is critical for the establishment of cell polarity. Here we show that binding of Cdc24p to the small GTPase Rsr1p/Bud1p is required for its recruitment to the incipient bud site. Rsr1p/Bud1p binds to the CH-domain of Cdc24p, which is essential for its function in vivo. We have identified a cdc24-mutant allele, which is specifically defective for bud-site selection. Our results suggest that Cdc24p is auto-inhibited by an intramolecular interaction with its carboxy-terminal PB1-domain. Rsr1p/Bud1p appears to activate the GEF activity of Cdc24p in vivo, possibly by triggering a conformational change that dissociates the PB1-domain from its intramolecular binding site. Genetic experiments suggest that Bem1p functions as a positive regulator of Cdc24p by binding to the PB1-domain of Cdc24p, thereby preventing its re-binding to the intramolecular inhibitory site. Taken together, our results support a two-step molecular mechanism for the site-specific activation of Cdc24p, which involves Rsr1p/Bud1p and the adaptor protein Bem1p. Introduction The establishment of asymmetry or cell polarity is critical for many different biological processes in both single- and multicellular organisms, including cell migration, differentiation, proliferation and morphogenesis (Drubin, 2000). The understanding of how cellular components can generate asymmetry represents one of the key steps toward understanding the molecular basis of cellular organization. The polarized assembly of the actin and microtubule cytoskeletons is regulated by site-specific activation of Rho-type GTPases, and many critical downstream targets have been identified (Aspenstrom, 1999; Johnson, 1999). However, much less is known about the spatial and temporal regulation of the Rho-type GTPases themselves. The yeast Saccharomyces cerevisiae has proven to be an excellent and genetically tractable model organism to study the biochemical pathways leading to the establishment of cell polarity (Pruyne and Bretscher, 2000). During vegetative growth, polarization of the actin cytoskeleton toward a single position on the cell cortex (incipient bud site) leads to bud formation, while during mating a pointed projection (shmoo) develops to allow contact and fusion with a cell of the opposite mating type. Central to the initiation of actin polarization during mating and budding is the local activation of the GTPase Cdc42p (Johnson, 1999). Regulation of Cdc42p involves its sole GEF Cdc24p (Bi et al, 2000) and several putative GAPs (Bem2p, Bem3p, Rga1p and Rga2p). In addition, Cdc24p is positively regulated by binding to the adaptor protein Bem1p, which is required to keep Cdc24p at sites of polarized growth (Butty et al, 2002). Finally, Cdc24p is sequestered in the nucleus during the G1-phase of the haploid cell cycle by binding to the adaptor Far1p (Toenjes et al, 1999; Nern and Arkowitz, 2000b; Shimada et al, 2000). Like most Rho-GEFs (Hoffman and Cerione, 2002), Cdc24p contains a domain with strong similarity to the Dbl-family of exchange factors (residues 284–452, Dbl-homology domain or DH) and a nearby pleckstrin-homology (PH) domain (residues 477–667; Toenjes et al, 1999; Zheng, 2001). Cdc24p also contains a calponin-homology (CH) domain in its amino-terminus (residues 154–226), which in some proteins has been implicated in binding to actin (Stradal et al, 1998). Finally, Cdc24p has a PB1-domain at its carboxy-terminus (residues 781–854; Ito et al, 2001; Butty et al, 2002). While the PH-domain is thought to serve as a membrane-targeting signal (Irvine, 1998), the functions of the CH- and PB1-motifs for the regulation of Cdc24p are not clear. We are interested in the spatial and temporal regulation of Cdc24p. During mating, cell polarity is determined by a gradient of mating pheromones secreted by the mating partner, which is interpreted with the help of the adaptor protein Far1p (Arkowitz, 1999; Gulli and Peter, 2001). Far1p interacts with Cdc24p and Gβγ and is thought to recruit Cdc24p to the site of receptor activation. Likewise, during vegetative growth, the position on the cell cortex where Cdc24p is activated is not random, but is regulated by several gene products in a cell-type-dependent manner (Chant, 1999). The transmembrane protein Bud10p/Axl2p is required to guide the budding machinery to the correct position in haploid cells, while Bud8p and Bud9p mark the site of polarization in diploid cells. The GTPase Rsr1p/Bud1p is regulated by its GEF Bud5p and its GAP Bud2p, both of which localize to the incipient bud site (Park et al, 1999). Bud5p co-localizes with the transmembrane protein Axl2p/Bud10p, a component of the budding landmark, and directly interacts with its cytoplasmic tail (Kang et al, 2001), suggesting that Bud5p may be recruited to the incipient bud site by binding to Axl2p/Bud10p. Several lines of evidence suggest a crucial role for the small GTPase Rsr1p/Bud1p in the local activation of Cdc24p at bud emergence. Rsr1p/Bud1p was originally isolated as a multicopy suppressor of cdc24-4-mutant cells (Bender and Pringle, 1989), and subsequently shown to bind directly to Cdc24p in a GTP-dependent manner (Park et al, 1997). Moreover, the recruitment of Cdc24p to the proper bud site depends on Rsr1p/Bud1p (Park et al, 2002). These results suggest that Rsr1p/Bud1p is an important regulator of Cdc24p, which may recruit Cdc24p to the incipient bud site. In this study, we investigated the molecular mechanisms of Cdc24p activation. Our results suggest that Rsr1p/Bud1p is not only responsible for the localization of Cdc24p but may also be involved in its activation in vivo. In addition, we propose that Cdc24p is auto-inhibited by binding of its carboxy-terminal PB1-motif to an intramolecular site. As Bem1p also interacts with this PB1-motif, we favor a model for Cdc24p activation where Bem1p may specifically stabilize the active conformation of Cdc24p by preventing intramolecular inhibition by its PB1-domain. Results The CH-domain of Cdc24p is required for its site-specific localization of Cdc24p Besides the DH- and PH-domains typically found in Rho-GEFs (Hoffman and Cerione, 2002), Cdc24p contains a CH-domain in its amino-terminal regulatory domain, and a PB1-domain at its carboxy-terminus (Figure 1A). Both Cdc24p-ΔCH and Cdc24p-ΔDH were unable to complement the growth defect of cdc24Δ cells (Figure 1B and C), implying that both domains are essential for their function in vivo. For control, we also analyzed a Cdc24p mutant harboring a 10 amino-acid deletion within its catalytic DH-domain (amino acids 406–415). As expected, Cdc24p-ΔCH-GFP did not accumulate in the cell nucleus of G1 cells (Toenjes et al, 1999; Nern and Arkowitz, 2000b; Shimada et al, 2000), but surprisingly was also unable to localize to the incipient bud site or tips of mating projections (Figure 1, panel D). In contrast, Cdc24p-ΔDH-GFP was found at cortical sites and the cell nucleus in a cell cycle-dependent manner (Figure 1, panel D), implying that GEF activity is not necessary for proper localization of Cdc24p. Taken together, these results suggest that the CH-domain is required for the recruitment of Cdc24p to the cell cortex. Figure 1.An intact CH- but not DH-domain is required for membrane localization of Cdc24p. (A) Schematic representation of the domain structure of Cdc24p. The numbers indicate amino acids starting from the start methionine. CH: calponin-homology domain; DH: Dbl-homology domain; PH: pleckstrin-homology domain, PB1: Phox-Bem1-homology domain. (B) cdc24Δ cells transformed with a URA3-marked plasmid carrying wild-type Cdc24p (YMG315) and, as indicated, either an empty plasmid (vector), wild-type Cdc24p, Cdc24p-ΔCH or Cdc24p-ΔDH. The cells were grown at 25°C until the mid-log phase, and five-fold serial dilutions were spotted on plates with (right panel) or without 5-FOA (left panel), a drug that selects against the URA-plasmid. (C) The expression of wild-type and the indicated Cdc24p mutants was analyzed in cdc24-5 cells, which express very low levels of endogenous Cdc24p (Butty et al, 2002). Actin was used as a loading control. (D) The localization of wild-type Cdc24p-GFP, Cdc24p-ΔCH-GFP or Cdc24p-ΔDH-GFP was analyzed in YMG258 (cln1Δ cln2Δ cln3Δ MET-CLN2) cells arrested in G1 by depletion of the G1-cyclins (−Cln2p) or 20 min after release from the arrest by induction of Cln2p (+Cln2p). Where indicated, the cells were treated with α-factor for 2 h. Note that, in contrast to Cdc24p-ΔDH-GFP, Cdc24p-ΔCH-GFP fails to localize to sites of polarized growth (arrows) during budding or mating. Download figure Download PowerPoint Binding of Cdc24p to Rsr1p/Bud1p is required for its recruitment to the axial bud site Cdc24p was previously shown to function as an effector of Rsr1p/Bud1p (Zheng et al, 1995; Park et al, 1997). To determine the functional significance of this interaction, we searched for cdc24 mutants, which are unable to interact with Rsr1p/Bud1p. Interestingly, cdc24-4 cells exhibited a random budding pattern when grown at the permissive temperature (Figure 2A), and their growth defect at restrictive temperature (36°C) was suppressed by increased levels of Rsr1p/Bud1p (Figure 3A). In contrast, cdc24-4 cells did not exhibit a mating defect even when assayed toward orientation-defective far1-c cells (Figure 2B; Chenevert et al, 1994), implying that they are able to efficiently interpret a natural pheromone gradient (Valtz et al, 1995). Immunoblotting with Cdc24p antibodies revealed that the Cdc24p-4-mutant protein was not grossly truncated and present in normal amounts (Figure 3B). However, the phosphorylation pattern of Cdc24p (upper blot) and the specific accumulation of the Cdc42p-effector Gic2p (middle blot) indicated that the Cdc24p-4 protein is inactive when the cells are shifted to the restrictive temperature. We recovered the cdc24-4-mutant allele by PCR, and identified a single point mutation at position 503, which changed glycine 168 to a glutamic acid residue (G168D; Figure 2C). Site-directed mutagenesis confirmed that this mutation is solely responsible for the observed defects (Figure 2D). Figure 2.Cells expressing Cdc24p-G168D are specifically defective for bud-site selection. (A) The budding pattern of wild-type (YEF241) and cdc24-4 (YEF313) cells grown at 25°C until the mid-log phase was examined by staining the bud scars with calcofluor white. The defect was quantified and presented as the percentage (%) of cells with an axial budding pattern. (B) Wild-type (YEF241), cdc24-4 (YEF313) and cdc24-m1 (YPW81) cells were tested at 25°C for their ability to mate with orientation-defective far1-c (YMP325) mating testers. (C, D) The positions of the identified mutation in cdc24-4 (G168D) and cdc24-m3 (S189P, Nern and Arkowitz, 1998) are marked on the schematic domain structure of Cdc24p (C). Site-directed mutagenesis was used to confirm that this single amino-acid change is sufficient to confer the temperature-sensitive phenotype (D). The assay was performed as described in the legend of Figure 1B. (E) The localization of Cdc24p-GFP or Cdc24p-G168D-GFP expressed from the CYC1-promoter was analyzed by time-lapse microscopy in cdc24Δ (YMP315) or bud1Δ cells (YACB191). The arrowheads mark the position of Cdc24p-GFP or Cdc24p-G168D-GFP. The numbers indicate the time (min) after the first image was taken (time 0). Note that Cdc24p-G168D-GFP is recruited to random positions, analogous to wild-type Cdc24p-GFP in bud1Δ cells. Download figure Download PowerPoint Figure 3.Interaction of Rsr1p/Bud1p with the CH-domain of Cdc24p is required for correct bud-site selection. (A) Wild-type (YEF241) and cdc24-4 (YEF313) cells were transformed with an empty control plasmid (vector) or a multi-copy plasmid expressing Rsr1p/Bud1p, and tested for their ability to grow at 36°C (A). (B) Extracts prepared from wild-type (YEF241) or cdc24-4 (YEF313) cells grown to the mid-log phase at 25°C and, where indicated, shifted to 37°C for 2 h were immunoblotted with antibodies against Cdc24p (upper panel), Gic2p (middle panel) or for control actin (bottom panel). (C) The interaction of Bud1p-G12V, Far1p and Bem1p with full-length or the indicated mutants of Cdc24p was analyzed by two-hybrid analysis using EGY48 cells grown at 25°C. The numbers indicate Miller units with standard deviations. An empty vector was included as control. A schematic representation of the various constructs is shown on the left. Download figure Download PowerPoint We next used two-hybrid assays and time-lapse microscopy to test whether Cdc24p-G168D is defective for its interaction with activated Rsr1p/Bud1p (Bud1p-G12V; Figures 2E and 3C). Deletion analysis of Cdc24p revealed that Rsr1p/Bud1p and Far1p interact with the amino-terminal domain, while Bem1p binds to its C-terminal PB1-domain (Ito et al, 2001; Butty et al, 2002). Importantly, the interaction between Cdc24p-G168D and Bud1p-G12V was strongly reduced, while its interaction with Far1p and Bem1p remained largely intact. Conversely, Cdc24p-S189P was defective for its binding to Far1p (Figure 3C; Nern and Arkowitz, 1998), while it was still interacting with Bud1p-G12V. Finally, time-lapse experiments confirmed that Cdc24p-G168D-GFP expressed in cdc24Δcells was recruited to random positions on the cell cortex with respect to the position of the previous bud (Figure 2E). In contrast, wild-type Cdc24p-GFP accumulated at the expected axial position in wild-type cells, and at random positions when expressed in bud1Δ, bud2Δ and bud5Δ cells (Figure 2E and data not shown; Park et al, 2002). Taken together, these results demonstrate that binding of Rsr1p/Bud1p to Cdc24p is specifically required for its correct recruitment to the axial position on the cell cortex. Rsr1p/Bud1p is sufficient to recruit Cdc24p to the cell cortex and may also be involved in its activation in vivo We observed that strong overexpression of an activated mutant form of Rsr1p/Bud1p (Bud1p-G12V) was toxic for cells (Figure 4A), and the cells arrest predominantly with a large unbudded morphology and an unpolarized cytoskeleton (data not shown). This toxicity required an intact effector domain of Rsr1p/Bud1p, as overexpression of Bud1p-G12V/T35A had no effect. We speculated that Rsr1p/Bud1p may recruit Cdc24p to the membrane leading to uniform activation of Cdc42p, which in turn may interfere with polarized growth toward a single bud site. To test whether Bud1p-G12V was able to recruit Cdc24p to the plasma membrane (Figure 4B), we arrested YMG258 cells expressing Cdc24p-GFP (upper panels) or for control Cdc24p-G168D-GFP (lower panels) in G1 by depletion of the G1-cyclin Cln2p (Gulli et al, 2000). The expression of Bud1p-GTP (Bud1p-G12V) or for control no protein (vector) was then induced by the addition of 2% galactose, and the localization of Cdc24p-GFP was examined by GFP microscopy after 2 h. As expected, Cdc24p and Cdc24p-G168D-GFP were predominantly nuclear with some cytoplasmic staining under these conditions (Figure 4B). However, Cdc24p-GFP, but not Cdc24p-G168D-GFP, was uniformly recruited to the plasma membrane after expression of Bud1p-G12V from the inducible GAL1-10-promoter (Mumberg et al, 1995). These results suggest that binding of Rsr1p/Bud1p to Cdc24p is required for the recruitment of Cdc24p to the cell cortex. Figure 4.Overexpression of an activated mutant form of Rsr1p/Bud1p is toxic and may recruit and uniformly activate Cdc24p at the plasma membrane in vivo. (A) Wild-type (K699) cells were transformed with an empty control plasmid (vector) or plasmids expressing Bud1p-G12V or Bud1p-G12V/T35A from the inducible GAL1-10-promoter. The cells were grown at 30°C until mid-log phase, and five-fold serial dilutions were spotted on plates containing 2% glucose (left panel) or 2% galactose (right panel). Plates were photographed after 3 days at 30°C. (B) YMG258 (cln1Δ cln2Δ cln3Δ MET-CLN2) cells expressing from the GAL1-10-promoter either no protein (vector, left panels) or activated Bud1p-G12V (BUD1-G12V, right panels) were grown in media containing 2% raffinose (GAL-promoter off) and arrested in G1 by depletion of the G1-cyclins. The localization of Cdc24p-GFP (upper panels) or Cdc24p-G168D-GFP (lower panels) was analyzed by GFP microscopy 2 h after the addition of 2% galactose. The percentage (%) of cells with Cdc24p-GFP or Cdc24p-G168D-GFP, which were uniformly located at the plasma membrane, is indicated. (C) The phosphorylation state of Cdc24p was analyzed by immunoblotting in YMG258 (cln1Δ cln2Δ cln3Δ MET-CLN2) cells arrested in G1 by depletion of the G1-cyclins (left panel), or wild-type (DLY1) and cdc42-27 (MOSY0124) cells 2 h after expression of either no protein ('−') or activated Bud1p-G12V ('+') from the GAL1-10-promoter. (D) YMG258 (cln1Δ cln2Δ cln3Δ MET-CLN2) cells expressing from the GAL1-10-promoter either no protein (vector) or activated Bud1p-G12V (BUD1-G12V) were grown in media containing 2% raffinose (GAL-promoter off) and arrested in G1 by depletion of the G1-cyclins. The recruitment of the Cdc42p-target Gic2p-GFP to the plasma membrane was analyzed 2 h after addition of 2% galactose. (E) Extracts prepared from wild-type (K699) cells transformed with an empty control plasmid (vector) or a plasmid expressing Bud1p-G12V from the inducible GAL1-10-promoter were incubated with glutathione–sepharose beads coated, as indicated, with GST-CRIB, or for control a nonfunctional GST-CRIB-mutant protein (GST-CRIB*) purified from E. coli. Bound Cdc42p (upper panel) and a fraction of the total Cdc42p present in the extract (lower panel) were detected by immunoblotting. Download figure Download PowerPoint We used three assays to determine whether activated Rsr1p/Bud1p is involved in the activation of Cdc24p at the cell cortex. First, Cdc24p accumulated in its hyperphosphorylated form in YMG258 cells expressing Bud1p-G12V (Figure 4C), and this hyperphosphorylation was largely dependent on functional Cdc42p. Second, the Cdc42p-target Gic2p-GFP was uniformly localized at the plasma membrane when Bud1p-GTP was overexpressed in G1-arrested YMG258 cells (Figure 4D; Jaquenoud and Peter, 2000). Finally, pull-down assays with GST-CRIB beads revealed that Cdc42p-GTP levels were increased in YMG258 cells expressing Bud1p-G12V (Figure 4E). Taken together, these data imply that membrane recruitment of Cdc24p by Rsr1p/Bud1p is sufficient for uniform activation of Cdc42p at the plasma membrane. Two possibilities could account for the observed increase in Cdc42p-GTP in vivo. First, membrane localization of Cdc24p may be sufficient for its activation. Alternatively, binding of Rsr1p/Bud1p to Cdc24p may activate Cdc24p. To distinguish between these possibilities, we fused the amino-terminus of Cdc24p to the myristoylation signal of c-Src (Figure 5A; myr-Cdc24p), and overexpressed the fusion protein from the weak, but inducible GALL-promoter (Mumberg et al, 1995). For control, we used a mutated myristoylation signal, where the myristoylated glycine residue was replaced with a non-myristoylatable alanine residue (myrG2A-Cdc24p). As expected, myr-Cdc24p-GFP, but not myrG2A-Cdc24p-GFP, was uniformly recruited to the plasma membrane and some internal membranes (Figure 5B). Interestingly, cells overexpressing myr-Cdc24p were viable (Figure 5C). In contrast, co-expression of myr-Cdc24p with low levels of Bud1p-G12V was toxic (Figure 5, panel C), and these cells accumulated with a large unbudded morphology and an unpolarized actin cytoskeleton (Figure 5, panel D). This lethality required an intact DH-domain of Cdc24p (Figure 5, panel E), and was specific for activated Rsr1p/Bud1p, as neither wild-type Rsr1p/Bud1p nor the GDP-locked Bud1p-K16N or the effector-mutant Bud1p-G12V/T35A was able to interfere with polarization in the presence of myr-Cdc24p (Figure 5, panels C and D). Finally, co-expression of Bud1p-G12V was able to induce lethality with myr-Cdc24p and myr-Cdc24p-S189P, but not with myr-Cdc24p-G168D, which is unable to interact with Rsr1p/Bud1p (see Figure 3). Together, these results support the notion that cells co-expressing myr-Cdc24p and Bud1p-G12V are unable to polarize, because Cdc42p may be activated uniformly over the cell cortex and thus interfere with polarized growth toward a single bud site. Indeed, GST-CRIB pull-down assays confirmed that these cells express increased levels of Cdc42p-GTP (Figure 5F). Based on these results, we conclude that membrane localization is not sufficient to activate Cdc24p in vivo, and that the interaction between Rsr1p/Bud1p and Cdc24p may be required for this process. Figure 5.Binding of Rsr1p/Bud1p to Cdc24p may be required to activate Cdc24p in vivo. (A) Schematic representation of Cdc24p fused at its amino-terminus to the wild-type (myr-Cdc24p) or mutant (myrG2A-Cdc24p) myristoylation signal of c-Src. (B) Fusion of a wild-type (myr) but not a mutant (myrG2A) myristoylation signal is able to uniformly target Cdc24p-GFP to the plasma membrane. The proteins were expressed in wild-type (K699) cells from the weak, but inducible GALL-promoter, and analyzed after 3 h of induction by confocal microscopy. (C) Wild-type (K699) cells expressing either myr-Cdc24p or myrG2A-Cdc24p from the GALL-promoter were transformed with an empty control plasmid (vector) or multicopy plasmids carrying the indicated wild-type or mutant BUD1 alleles expressed from their endogenous promoter. Five-fold serial dilutions of cultures in mid-log phase were spotted on media containing glucose or galactose. (D) The morphology and actin polarization of the cells shown in (C) were examined after staining with rhodamine-phalloidin. The percentage (%) of large unbudded, unpolarized cells was quantified. (E) Wild-type (K699) cells co-expressing wild-type or the indicated myr-Cdc24p mutants from the inducible GALL-promoter, together with multicopy plasmids carrying wild-type or the indicated mutant forms of BUD1, were analyzed as described in (C). (F) Extracts prepared from wild-type (K699) cells co-expressing wild-type or the indicated myr-Cdc24p mutants from the inducible GALL-promoter, together with multicopy plasmids carrying BUD1-G12V or BUD1-K16N, were analyzed for the presence of Cdc42p-GTP by GST-CRIB pull-down assays. Bound Cdc42p (upper panel) and a fraction of the total Cdc42p present in the extract (lower panel) were detected by immunoblotting. Download figure Download PowerPoint Isolation of constitutively activated alleles of Cdc24p As membrane-targeted Cdc24p required binding of Rsr1p/Bud1p for its activation, we reasoned that Cdc24p may be present in an auto-inhibited, inactive conformation. If this model were correct, we should be able to isolate activated alleles of myr-Cdc24p. We thus randomly mutagenized myr-Cdc24p expressed from the inducible GALL-promoter by error-prone PCR, and screened for mutants that failed to grow in the presence of galactose (Figure 6A). Interestingly, these cells arrested with an unbudded morphology and an unpolarized actin cytoskeleton (Figure 6D and data not shown). This defect was not caused by increased levels of myr-Cdc24p, as the mutant alleles were expressed at levels comparable to myr-Cdc24p (Figure 6C). Sequencing revealed that all mutations mapped to the carboxy-terminal part of Cdc24p after the DH-motif (Figure 6B), and we identified several point mutations within the PH- and the carboxy-terminal domain. Surprisingly, one mutation introduced a stop codon already at position 457, which results in a truncated Cdc24 protein lacking its PH-domain. We wanted to exclude the possibility that some of the recovered alleles may cause a dominant-negative phenotype by interfering with the function of wild-type Cdc24p. First, expression of myr-Cdc24-G644C, but not myr-Cdc24-G644C-ΔDH, was toxic, implying that either Cdc42p binding or its ability to produce Cdc42p-GTP was required (Figure 6E). Second, the growth defect of cells overexpressing myr-Cdc24-G644C was suppressed by simultaneous overexpression of the Cdc42p GAP Bem2p (Figure 6F). Third, GST-CRIB pull-down assays revealed that cells overexpressing myr-Cdc24-G644C, but not myr-Cdc24-G644C-ΔDH, possess increased levels of Cdc42p-GTP (Figure 6G). Finally, we tested the functionality of the Cdc24p mutants by expressing them without the myristoylation signal from the endogenous promoter in cdc24Δ cells. Indeed, all tested Cdc24p mutants were able to support growth in this context (Figure 6H and data not shown), demonstrating that the mutant alleles are functional. Based on these results, we conclude that the identified mutations are constitutively activating Cdc24p in vivo. Figure 6.Identification and characterization of constitutively active alleles of Cdc24p. (A–C) Specific mutations within myr-Cdc24p that confer toxicity when expressed in wild-type cells (K699) from the GALL-promoter were screened as described in Material and methods. Five-fold serial dilutions were spotted on plates containing glucose or galactose, and grown for 3 days at 30°C (A). The responsible amino-acid changes were determined and are shown in (B). Arrowheads show single amino-acid changes, while stifled arrows point to mutations that result in a truncated protein by introducing a stop codon. The expression level of myr-Cdc24p and the indicated myr-Cdc24p mutants in wild-type cells (K699) was determined by immunoblotting with Cdc24p-antibodies (C). (D) The morphology and actin polarization of wild-type (K699) cells expressing for 3 h either no protein (vector) or myr-Cdc24p-G644C from the GALL-promoter were examined after staining with rhodamine-phalloidin. (E, F) Wild-type (K699) cells expressing from the GALL-promoter, as indicated, either no protein (vector), myr-Cdc24p, myr-Cdc24p-G644C or myr-Cdc24p-G644C/ΔDH were grown for 3 days at 30°C on plates containing galactose. The cells shown in (F) were transformed with a plasmid expressing the GAP BEM2 from the Gal1-10 promotor. (G) The levels of Cdc42p-GTP were compared by GST-CRIB pull-down assays in extracts prepared from wild-type (K699) cells expressing myr-Cdc24p, myr-Cdc24-G644C or myr-Cdc24p-G644C/ΔDH from the inducible GALL-promoter. Bound Cdc42p (upper panel) and a fraction of the total Cdc42p present in the extract (lower panel) were detected by immunoblotting. (H) cdc24Δ cells transformed with a URA3-marked plasmid carrying wild-type Cdc24p (YMG315) and, as indicated, either an empty plasmid (vector), wild-type Cdc24p or Cdc24p-G644C were grown at 25°C until mid-log phase, and serial dilutions were spotted on plates with (lower panel) or without 5-FOA (upper panel), as described in the legend to Figure 1B. Download figure Download PowerPoint The C-term
Referência(s)