The essential phosphatidylinositol kinase homologue TOR2 of Saccharomyces cerevisiae controls the actin cytoskeleton by activating a GTPase switch consisting of RHO1 (GTPase), ROM2 (GEF) and SAC7 (GAP). We have identified two mutations, rot1‐1 and rot2‐1, that suppress the loss of TOR2 and are synthetic‐lethal. The wild‐type ROT1 and ROT2 genes and a multicopy suppressor, BIG1, were isolated by their ability to rescue the rot1‐1 rot2‐1 double mutant. ROT2 encodes glucosidase II, and ROT1 and BIG1 encode novel proteins. We present evidence that cell wall defects activate RHO1. First, rot1, rot2, big1, cwh41, gas1 and fks1 mutations all confer cell wall defects and suppress tor2ts. Second, destabilizing the cell wall by supplementing the growth medium with 0.005% SDS also suppresses a tor2ts mutation. Third, disturbing the cell wall with SDS or a rot1, rot2, big1, cwh41, gas1 or fks1 mutation increases GDP/GTP exchange activity toward RHO1. These results suggest that cell wall defects suppress a tor2 mutation by activating RHO1 independently of TOR2, thereby inducing TOR2‐independent polarization of the actin cytoskeleton and cell wall synthesis. Activation of RHO1, a subunit of the cell wall synthesis enzyme glucan synthase, by a cell wall alteration would ensure that cell wall synthesis occurs only when and where needed. The mechanism of RHO1 activation by a cell wall alteration is via the exchange factor ROM2 and could be analogous to signalling by integrin receptors in mammalian cells.
The Saccharomyces cerevisiae proteins TOR1 and TOR2 were originally identified genetically as the targets of the immunophilin–immunosuppressive complex FKBP–rapamycin (Heitman et al., 1991; Kunz et al., 1993; Cafferkey et al., 1994; Helliwell et al., 1994). They contain a lipid kinase domain near their C‐terminus and are thus structurally related to PI 3‐ and PI 4‐kinases. The kinase domain is essential for TOR function, although no kinase activity has been demonstrated for either TOR1 or TOR2 (Zheng et al., 1995; Schmidt et al., 1996). The mammalian counterpart of TOR1 and TOR2, which is also the target of FKBP‐rapamycin, has protein kinase activity (Brown et al., 1995; Brunn et al., 1997).
TOR2 has two essential signalling functions (Kunz et al., 1993; Helliwell et al., 1994; Zheng et al., 1995; Hall, 1996). One function is shared with TOR1 and is required for signalling activation of translation initiation and early G1 progression in response to nutrients (Barbet et al., 1996; Di Como and Arndt, 1996). Translation initiation and G1 progression in mammalian cells are also controlled by TOR (mTOR/FRAP/RAFT), suggesting that this TOR signalling pathway is highly conserved (Beretta et al., 1996; von Manteuffel et al., 1996; Thomas and Hall, 1997). The second TOR2 function, which TOR1 is unable to perform, mediates the cell cycle‐dependent organization of the actin cytoskeleton (Schmidt et al., 1996, 1997). TOR2 signals to the actin cytoskeleton by activating a Rho‐type GTPase switch.
In yeast, the organization of the actin cytoskeleton is controlled so as to optimize secretion toward regions of polarized growth (Adams and Pringle, 1984; Kilmartin and Adams, 1984). As in mammalian cells, the yeast cytoskeleton is regulated by a family of small GTPases comprising CDC42 (Johnson and Pringle, 1990), RHO1, RHO2 (Madaule et al., 1987), RHO3 and RHO4 (Matsui and Toh‐e, 1992b). During the G1 phase of the cell cycle, growth occurs uniformly over the entire cell surface and cortical actin patches are distributed randomly. As cells progress into the cell cycle, CDC42 mediates bud emergence by controlling the concentration of actin cortical patches at the chosen bud site (Bender and Pringle, 1989; Adams et al., 1990). RHO1, and possibly its non‐essential homologue RHO2, are required for further bud growth (Yamochi et al., 1994). RHO1 controls three different proteins which are important for cell wall synthesis and actin organization. It binds and activates β‐1,3 glucan synthase which produces a major structural component of the cell wall (Drgonova et al., 1996; Qadota et al., 1996). It also binds and activates PKC1 (Nonaka et al., 1995; Kamada et al., 1996) which transcriptionally activates many cell wall biosynthetic enzymes, including β‐1,3 glucan synthase (Shimizu et al., 1994; Jiang et al., 1995; Igual et al., 1996; Madden et al., 1997). PKC1 may also play a role in organizing the actin cytoskeleton (Mazzoni et al., 1993). Lastly, RHO1 may control the actin cytoskeleton by binding BNI1 which binds profilin (Kohno et al., 1996; Evangelista et al., 1997). RHO3 and RHO4 are also important for bud growth; depleting them causes cells to lyse or arrest growth with small buds and delocalized actin (Matsui and Toh‐e, 1992a).
Small GTPases exist in two states, a GTP‐bound active state and a GDP‐bound inactive state (Hall, 1990; Bourne et al., 1991). GDP/GTP exchange factors (GEFs) catalyse the exchange of GDP for GTP and are thus G‐protein activators. GTPase‐activating proteins (GAPs) accelerate the rate of GTP hydrolysis and are thus negative regulators. The mechanisms by which GEFs and GAPs are controlled are poorly understood.
The yeast cell wall is a dynamic structure that makes up 15–30% of the dry weight of the cell (Klis, 1994; Orlean, 1997). It is composed of three major components, mannoproteins, β‐1,6‐glucan and β‐1,3‐glucan, all cross‐linked to each other and to chitin, an N‐acetylglucosamine polymer (Kapteyn et al., 1997; Kollar et al., 1997). The cell wall determines cell shape, provides rigidity that counteracts the outward turgor pressure on the plasma membrane, and acts as a diffusion barrier delimiting the periplasmic space. Cell wall synthesis is a poorly understood, highly regulated process that is subject to cell growth and environmental cues.
We have shown recently that TOR2 activates RHO1 and RHO2 via their GEF ROM2 (Schmidt et al., 1997). Here we show that cell wall damage also activates RHO1 via ROM2, independently of TOR2. Our results suggest that yeast cells have a cell wall monitoring system that activates cell wall synthesis. Such a monitoring system signals in response to cell wall alterations caused by duress and possibly also by normal cell wall remodelling.
Mutations in ROT1 and ROT2 suppress the loss of TOR2
TOR2 has at least two essential functions. One is redundant with TOR1 and controls translation initiation; the other is unique to TOR2 and controls the actin cytoskeleton (Kunz et al., 1993; Helliwell et al., 1994; Barbet et al., 1996; Schmidt et al., 1996). To gain insight into the TOR2‐unique function, we selected second‐site suppressors of tor2ts mutations in a TOR1 background. The TOR1 tor2ts mutants were defective only in the TOR2‐unique function, and temperature sensitive for growth. Approximately 300 spontaneous mutants able to grow at 37°C on rich medium were isolated (see Materials and methods). Eight mutants containing suppressors that were able to suppress a tor2‐null allele were chosen for further study (see Materials and methods). The eight mutants were back‐crossed to the parental tor2ts strain and diploids were sporulated and dissected. The suppressor mutations were all recessive and segregated 2:2, indicating that a single nuclear locus was affected in each case. The eight mutants were then crossed to each other and diploids were sporulated and dissected. The suppressor mutations were linked to one of two loci which were named ROT1 and ROT2 for reversal of TOR2 (rot1 was isolated six times and rot2 twice). We chose one mutant from each segregation group (rot1‐1, MB136‐2d; rot2‐1, MB137‐1d) for further study. The rot1‐1 and rot2‐1 mutations suppressed both the growth defect and the actin organization defect of the tor2ts mutant (Figure 1 and data not shown). Interestingly, a tor2ts diploid heterozygous for rot1‐1 and rot2‐1 (MB66) was viable at 37°C, indicating non‐allelic non‐complementation (data not shown).
To determine whether the rot mutations could suppress a loss of the TOR‐common function, rot1‐1 tor2ts and rot2‐1 tor2ts strains were crossed to a tor1 tor2ts mutant (strain SH221). None of the tor1 tor2ts segregants was viable at 37°C, indicating that the rot mutations could suppress the loss of only the TOR2‐unique function (data not shown).
Isolation of ROT1, ROT2, BIG1 and SAC7 by suppression of rot1‐1 rot2‐1 synthetic lethality
The rot1‐1 and the rot2‐1 mutations were introduced into our wild‐type strain background to generate strains MB134‐1a and MB135‐4b, respectively. As expected from the wild‐type‐like growth of the rot1‐1 tor2ts and the rot2‐1 tor2ts mutants at all temperatures, the rot1‐1 and rot2‐1 mutations conferred no growth phenotype in a wild‐type background at 24, 30 or 37°C (see Materials and methods for scoring of the rot mutations). However, a combination of rot1‐1 and rot2‐1 in an otherwise wild‐type background (MB145‐3b) conferred a recessive, temperature‐sensitive growth defect (see Materials and methods). We took advantage of this recessive synthetic lethal phenotype to clone ROT1 and ROT2. Both a high‐copy‐number (2μ) and a single‐copy‐number (CEN) plasmid library were screened for rescue of the temperature‐sensitive growth phenotype of the rot1‐1 rot2‐1 mutant. Four open reading frames and a tRNA gene which conferred growth at 37°C were identified (see Materials and methods): YMR200w; YBR229c; YHR101c; SAC7; and a Gln tRNA [tQ(CUG)M]. When present in a single copy, only YMR200w and YBR229c were able to suppress completely the synthetic lethality (Figure 2). YMR200w and YBR229c were subsequently shown, by genetic linkage, to encode ROT1 and ROT2, respectively (see Materials and methods). This assignment was confirmed by transforming the single copy YMR200w and YBR229c plasmids into the rot1‐1 tor2ts and rot2‐1 tor2ts mutants. YMR200w was able to complement only rot1‐1, and YBR229c was able to complement only rot2‐1 (data not shown).
The high‐copy‐number suppressor of the rot1‐1 rot2‐1 synthetic lethality YHR101c, a weak suppressor, was named BIG1 (bad in glucose, see below). The high‐copy‐number suppressor SAC7 showed strong homology to Rho‐type GTPase‐activating proteins (GAPs). SAC7 was shown indeed to encode a GAP, for RHO1, as described elsewhere (Schmidt et al., 1997). The remaining high‐copy‐number suppressor encoded a Gln tRNA and was not pursued further.
ROT1 and BIG1 were novel proteins with one putative transmembrane domain each and no homologue in the DDBJ/EMBL/GenBank database. Upon resequencing the 3′‐end of ROT1, we observed an additional thymidine at position 758 compared with the sequence in the database. Our sequence changed the last five predicted amino acids and extended the predicted protein by 58 amino acids. The extended ROT1 protein consisted of 316 amino acids, with a calculated molecular weight of 34 kDa. The predicted BIG1 protein, taking into account an 87‐nucleotide intron in BIG1, was 335 amino acids with a molecular weight of 37 kDa. ROT2 (predicted 954 amino acids, 110 kDa) showed strong homology to glucosidases, and Trombetta et al. (1996) have reported recently that YBR229w (GLS2/ROT2) encodes glucosidase II. Glucosidase II is an ER protein that removes the two α‐1,3 linked glucose residues of N‐linked oligosaccharides after glucosidase I has removed the terminal α‐1,2 linked glucose residue (Kilker et al., 1981; Saunier et al., 1982; Herscovics and Orlean, 1993). Mammalian glucosidase II has been shown to be important, along with calnexin, calreticulin and the UDP:Glc glycoprotein glucosyltransferase, for monitoring the maturation of glycoproteins in the ER (quality control) (Bergeron et al., 1994). A similar quality control mechanism may also exist in yeast (Parlati et al., 1995).
The ROT1, ROT2 and BIG1 open reading frames were disrupted with PCR‐generated cassettes (see Materials and methods). ROT1 was essential for growth; rot1‐null (rot1::kanMX2) spores germinated and arrested growth within one cell cycle as medium‐ or large‐budded cells. A strain containing ROT1 under control of the GAL promoter (MB121‐1b) (see Materials and methods) also arrested growth, when shifted from galactose to glucose medium (Figure 7A). Thus, the original recessive rot1‐1 allele, which did not confer a growth defect, was not a null allele. ROT2 was not essential, as reported previously (Trombetta et al., 1996). A ROT2 deletion (rot2:: HIS3MX6; strain MB86‐1b) conferred no obvious growth defect and suppressed a tor2ts mutation (MB104‐2a), as observed for the original rot2‐1 mutation (Figure 3A). BIG1 was important but not essential for growth. Disruption of BIG1 (big1::HIS3MX6) caused a severe growth defect; big1 cells (PA5‐7b) formed microcolonies after 8 days on YPD medium (Figure 7A). The big1 cells were very large and often unbudded and multinucleate. The poor growth of big1 cells was partly suppressed on media containing galactose as a carbon source or 1 M sorbitol (data not shown). The BIG1 disruption did not suppress a tor2ts mutation (data not shown). However, a tor2ts mutant containing BIG1 under control of the GAL promoter (PA31‐7b) was viable at 37°C on glucose (data not shown). Thus, although not originally identified as a suppressor of tor2ts, loss of function of BIG1 could also suppress tor2ts. The difference in suppression by the two big1 alleles is presumably due to the fact that the BIG1 disruption caused a severe growth defect on its own, whereas the GAL‐BIG1 allele did not (see Materials and methods).
Loss of ER glucosidase activity suppresses tor2ts
ROT2 was identified by Trombetta et al. (1996) as glucosidase II. To examine the link between suppression of tor2ts and glucosidase activity in the ER, we searched for a yeast homologue of mammalian glucosidase I. Lack of glucosidase I activity should also abolish trimming by glucosidase II as the enzymes act sequentially. The nearest homologue to mammalian glucosidase I in yeast is CWH41, which has been reported to be an ER protein involved in β‐1,6 glucan synthesis (Jiang et al., 1996). We constructed a cwh41 tor2ts strain (MB119) and tested for growth at 37°C. We found that tor2ts was suppressed by the loss of CWH41 (Figure 3A). To determine whether CWH41 is indeed glucosidase I, we constructed cwh41 sec18ts (PA37‐5c) and rot2 sec18ts (PA36‐2d) strains; the sec18ts mutation was necessary to block glycoproteins in the ER and thereby prevent heterogeneous outer chain carbohydrate addition that obscures small changes in electrophoretic mobility (Esmon et al., 1984). These strains were shifted to 37°C to impose the ER block, labelled with [35S]Met and lysed, and the glycoprotein carboxypeptidase Y (CPY) was immunoprecipitated. CPY from the cwh41 sec18ts and rot2 sec18ts mutants showed reduced motility compared with CPY from the sec18ts strain (PA33‐6a). The differences in mobility were due to differences in N‐linked glycosylation as endo H‐treated CPY from all three strains showed the same mobility (Figure 3B). These results suggest that CWH41 is a glucosidase, and confirm the glucosidase activity of ROT2. In agreement with this, Romero et al. (1997) and Simons et al. (1998) have found, while this study was in progress, that CWH41 is glucosidase I. Thus, loss of glucosidase activity in the ER correlates with suppression of tor2ts.
We also examined whether ER glucosidase activity was affected upon depletion of ROT1 or BIG1. GAL‐ROT1 sec18ts (PA38‐2a) and GAL‐BIG1 sec18ts (PA35‐19d) strains were passaged in glucose medium at non‐permissive temperature (see Materials and methods), and assayed for ER glucosidase activity as described above. In neither case was an effect on glucosidase activity observed (Figure 3B). Thus, because depletion of BIG1 suppresses tor2ts but does not affect glucosidase activity, defects in ER glucosidases appear to be sufficient but not necessary for suppression of tor2ts.
ROT1, ROT2 and BIG1 mutants have cell wall defects
Mutants defective in CWH41, like many other mutants altered in N‐linked glycosylation, also have a cell wall defect (Klis, 1994; Jiang et al., 1996). The cell wall is composed of three major components, mannoproteins, β‐1,6‐glucan and β‐1,3‐glucan, all cross‐linked to each other and to chitin (Kapteyn et al., 1997; Kollar et al., 1997). CWH41 mutants have been reported to have low levels of β‐1,6‐glucan (Jiang et al., 1996). As a measure of cell wall composition, we examined the distribution and levels of chitin in rot1‐1, rot2‐1 and big1 (GAL‐BIG1) mutants (strains MB134‐1a, MB135‐4b, and PA11‐6a). First, rot1‐1, rot2‐1 and big1 cells were grown at 37°C in YPD (conditions under which tor2ts is suppressed) and stained with calcofluor white to visualize chitin. The rot1‐1, rot2‐1 and big1 strains exhibited an altered distribution of chitin; unlike in wild‐type cells, chitin staining was observed occasionally in bud tips and at elevated levels in the lateral wall of mother cells. Furthermore, the rot1‐1, rot2‐1 and big1 strains showed a random distribution of bud scars reflecting a random budding pattern. The bud scars of wild‐type cells were aligned in rows reflecting a normal axial budding pattern (Figure 4A). Second, the chitin content in the different strains was measured by a colorimetric assay (Bulawa et al., 1986). The rot1‐1, rot2‐1 and big1 cells had increased chitin levels (188, 176 and 173% of that of the wild type, respectively) in YPD at 37°C (Figure 4B). Thus, we conclude that the cell walls of ROT1, ROT2 and BIG1 mutants are abnormal. While this study was in progress, Simons et al. (1998) found that a ROT2 (GLS2) mutant, like a CWH41 mutant, has low levels of β‐1,6‐glucan, providing further evidence that ROT2 mutants have a cell wall defect.
Cell wall defects suppress tor2ts
Because rot1, rot2, big1 and cwh41 mutations all confer cell wall defects and suppress the loss of TOR2, we examined whether other cell wall defects could also suppress tor2ts. First, we examined if destabilizing the cell wall by supplementing the growth medium with the detergent SDS (Igual et al., 1996) suppressed the growth defect of a tor2ts mutant. A tor2ts mutant (MB133‐2b) incubated at non‐permissive temperature on rich medium supplemented with 0.005% SDS grew almost like wild type (Figure 5A). Second, we disrupted individually, in a tor2ts mutant, three genes directly involved in specific aspects of cell wall assembly or synthesis: GAS1, FKS1 and CHS3 (see Materials and methods). GAS1 encodes a GPI‐anchored plasma membrane protein required for proper cell wall assembly (Popolo et al., 1993, 1997; Ram et al., 1995). FKS1 encodes a component of the plasma membrane β‐1,3 glucan synthase complex (Douglas et al., 1994; Eng et al., 1994; Castro et al., 1995; Mazur et al., 1995). CHS3 encodes the putative catalytic subunit of the chitin synthase that produces 90% of cellular chitin (Shaw et al., 1991; Cabib et al., 1993). The gas1 tor2ts, fks1 tor2ts and chs3 tor2ts double mutants (strains MB120, MB140 and MB147) were incubated at 37°C on rich medium. Deletions of GAS1 and FKS1 suppressed the lethality of tor2ts. However, the suppression by gas1 was weak, possibly because gas1 itself caused a severe growth defect. Deletion of CHS3 had no effect (Figure 5B). Thus, an SDS‐induced defect or a specific defect in any one of the three main components of the cell wall, mannoproteins (gas1), β‐1,3 glucan (fks1) and β‐1,6 glucan (cwh41), suppressed the growth defect of a tor2ts mutant. Taken together, these results suggest that disturbing the cell wall is sufficient to compensate for the loss of TOR2. A mutation in CHS3 does not suppress tor2ts, possibly because an increase in chitin levels, as observed in rot, big1, gas1 and fks1 mutants, may be necessary for suppression. Cell wall mutants in general appear to have an increase in chitin content that is essential to compensate for the primary cell wall defect (Kapteyn et al., 1997; Popolo et al., 1997).
Cell wall defects activate RHO1 via ROM2
Several lines of evidence suggested that mutations in ROT1, ROT2 and BIG1 activate RHO1. First, mutations in all three genes suppress tor2ts, which is also suppressed by activation of RHO1 (Schmidt et al., 1997). Second, the synthetic lethality of a rot1‐1 rot2‐1 double mutation and the toxicity of a big1‐null mutation are suppressed by overexpression of SAC7, which encodes a RHO1 GAP (Figure 2 and data not shown). Third, the growth defect conferred by disruption of ROT1 or BIG1 is partly suppressed by deletion of ROM2, which encodes one of two RHO1 GEFs (Figure 7A). The latter two findings suggest also that the growth defect of rot1‐1 rot2‐1, rot1‐null, or big1‐null strains is due to hyperactivation of RHO1. Hyperactivation of RHO1 is toxic (Madaule et al., 1987). To determine if mutations in ROT1, ROT2 and BIG1 indeed activate RHO1, we assayed extracts from rot1‐1, rot2‐1 and big1 (GAL‐BIG1) mutants (strains MB134‐1a, MB135‐4b and PA11‐6a) for GEF activity toward RHO1 (see Materials and methods). GEF activity on RHO1 from rot1‐1, rot2‐1 and big1 strains was increased to 185, 183 and 135% of that of the wild type, respectively (Figure 6). Because mutations in ROT1, ROT2 and BIG1 all cause cell wall defects, we examined if other conditions that cause cell wall defects also activate RHO1. We assayed GEF activity in extracts from cwh41, gas1 and fks1 mutants and in an extract from an SDS‐treated wild‐type strain (strains MB132‐2c, MB146‐3d, MB144‐2b and JK9‐3da) (see Materials and methods). GEF activity toward RHO1 in extracts from cwh41, gas1, fks1 mutants and SDS‐treated wild‐type cells was 200, 178, 143 and 185% of the activity from SDS‐untreated wild‐type cells, respectively (Figure 6). Thus, cell wall defects activate RHO1 via a GEF activity.
We have shown previously that TOR2 activates RHO1 via the GEF ROM2 (Schmidt et al., 1997). The findings that cell wall defects suppress a tor2ts mutation and stimulate GEF activity toward RHO1, and that the growth defect conferred by disruption of ROT1 or BIG1 is suppressed, albeit weakly, by deletion of ROM2 (Figure 7A) suggested that cell wall defects also activate RHO1 via ROM2. To investigate this further, we examined whether a rom2 mutation prevents the suppression of tor2ts by SDS. A rom2 tor2ts double mutant (AS153‐1a) grew very poorly on SDS‐containing medium compared with rom2 (AS152‐1c) and tor2ts (MB133‐2b) single mutants (Figure 7B), indicating that ROM2 is required for the suppression by SDS. The residual growth of the rom2 tor2ts double mutant on SDS could be due to stimulation of the other RHO1 GEF ROM1 (Ozaki et al., 1996). Furthermore, we found that rom2 also prevented the suppression of a tor2 mutation by rot2 (data not shown). Taken together, these results suggest that cell wall defects, like TOR2, also activate RHO1 via ROM2.
Our experiments do not directly address the issue of whether cell wall defects activate RHO1 via the GEF ROM1, or whether cell wall defects activate GTPases other than RHO1. However, cell wall defects may also activate RHO2, as TOR2 activates both RHO1 and RHO2 (Schmidt et al., 1997) and ROM2 is a GEF for both RHO1 and RHO2 (Ozaki et al., 1996).
We have shown that conditions which alter the cell wall suppress the growth defect of a TOR2 mutant and activate RHO1. These conditions include 0.005% SDS in the growth medium or loss‐of‐function mutations in ROT1, ROT2, BIG1, CWH41, GAS1 or FKS1. ROT1 and BIG1 are novel genes of unknown function. ROT2 and CWH41 encode the ER glucosidases II and I, respectively. It is not known why defects in ER glucosidases cause a cell wall defect. GAS1 encodes a GPI‐anchored, plasma membrane protein required for cell wall assembly. FKS1 encodes the catalytic subunit of β‐1,3 glucan synthase. Furthermore, we have shown that cell wall defects activate RHO1 via the GEF ROM2.
Why do cell wall defects activate RHO1? As RHO1 controls cell wall synthesis (see Introduction), activation of RHO1 by a cell wall alteration would ensure that cell wall synthesis occurs as needed. The cell wall alterations that activate RHO1 are alterations caused by duress, as shown here, but may also include alterations caused during normal cell growth. The yeast cell wall is remodelled during vegetative growth to allow bud emergence and cell fission, and during mating to allow shmoo formation and cell fusion.
Cell wall defects suppress the growth defect conferred by a tor2 mutation probably via activation of RHO1. First, TOR2 mutants are non‐viable because they are unable to activate RHO1 (Schmidt et al., 1997). Second, the suppression of a tor2 mutation by cell wall defects is reversed by overexpression of SAC7, the GAP for RHO1 (data not shown). Thus, cell wall defects most likely suppress by inducing a TOR2‐independent pathway for activating RHO1. The activation of RHO1 by cell wall defects is independent of TOR2 (although still dependent on ROM2) because at least rot1‐1 and rot2‐1 suppress a tor2‐null allele. We therefore propose a cell‐wall‐sensing pathway that signals to and activates RHO1 when cell integrity is compromised (Figure 8). The observation that a cell wall defect induces a compensating change in cell wall composition (Kapteyn et al., 1997; Popolo et al., 1997) suggests that a cell‐wall‐sensing and signalling mechanism exists.
How are cell wall defects sensed? In mammalian cells, integrin receptors interact with the extracellular matrix and, in response to mechanical stress, activate RhoA, which in turn induces formation of focal adhesions and actin stress fibers (Clark and Brugge, 1995; Nobes and Hall, 1995). As RhoA is a structural and functional homologue of RHO1 (Qadota et al., 1994; Yamochi et al., 1994), yeast cells may have an integrin‐like receptor that interacts with the cell wall and activates RHO1 in response to a cell wall alteration. The nature of the cell wall alteration sensed by such a receptor would have to be very general, as several different types of cell wall defects activate RHO1. An alternative model is that cell wall defects are sensed indirectly. Kamada et al. (1995) have suggested that PKC1, a downstream effector of RHO1, is activated by mechanosensors sensing an outward stretching of the plasma membrane. A cell wall defect could result in a turgor pressure‐induced outward stretching of the plasma membrane which would then activate RHO1.
Materials and methods
Strains, plasmids and media
The S.cerevisiae strains used in this study are listed in Table I. All strains were isogenic derivatives of JK9‐3d. Plasmids used in this study were as follows. pMB2 (YCplac33::ROT1) is YCplac33 (URA3 CEN) (Gietz and Sugino, 1988) containing a 1.4 kb SacI–EcoRI fragment bearing ROT1. pMB4 (YCplac111::GAL1‐HA‐ROT1) expresses 2×HA‐tagged ROT1 under control of the GAL1 promoter, in pAS24 (YCplac111::GAL1‐HA). pAS24 is YCplac111 (LEU2 CEN) (Gietz and Sugino, 1988) expressing the 2×HA tag under control of the GAL1 promoter. pMB6 (YCplac33::ROT2) is YCplac33 containing a 4.4 kb SacI–XbaI fragment bearing ROT2. pMB7 (YEplac195::SAC7) is YEplac195 (URA3 2μ) (Gietz and Sugino, 1988) containing a 3.5 kb EcoRI–HindIII fragment bearing SAC7. pMB8 (YCplac111::GAL1‐ROT1) expresses ROT1 under control of the GAL1 promoter, in pAS23. pAS23 (YCplac111::GAL1) is YCplac111 (LEU2 CEN) (Gietz and Sugino, 1988) containing the GAL1 promoter. pPAD24 (YEplac195:: BIG1) is YEplac195 containing a 1.6 kb FspI–ScaI fragment bearing BIG1. pPAD25 (pTB227::BIG1‐cmyc) is pTB227 (URA3 CEN GALp CYC1 terminator) expressing C‐terminally cmyc‐tagged BIG1 under control of the GAL promoter. pAS53 (pGST‐RHO1), expressing a GST‐RHO1 fusion under control of the GAL promoter, is the yeast GST‐fusion vector pEG‐KT (URA3 leu2‐d 2μ) (Mitchell et al., 1993) containing a 1 kb BamHI–HindIII fragment bearing the entire RHO1 open reading frame from pGAL‐HA‐RHO1 (Schmidt et al., 1997). YCplac111::tor2‐21 contains the temperature‐sensitive allele tor2‐21 (Helliwell et al., 1998). Rich media (YPD, YPGal) and synthetic minimal media (SD, SGal) complemented with the appropriate nutrients for plasmid maintainance were as described (Guthrie and Fink, 1991).
Construction of GAL‐ROT1 and GAL‐BIG1
The ROT1 and BIG1 open reading frames were placed under the control of the GAL1 promoter. For ROT1, a SalI site was introduced 5′ of the initiation codon by PCR with the following four oligonucleotides: 5′‐CACTTTGTCGACATGTGGTCGAAAAAG‐3′, 5′‐CGACCAGCG‐ CGCGCCCATAGTAATAAAGTGCAAC‐3′, and the #1211 and #1233 sequencing primers (Pharmacia). The 0.9 kb SalI fragment generated by PCR was ligated into pAS23 and pAS24 to yield pMB8 and pMB4, respectively. pMB4 and pMB8 were transformed into a rot1::kanMX2/ROT1 diploid, and transformants were sporulated and dissected on galactose‐containing medium. The HA‐ROT1 construct encoded by pMB4 was functional as determined by the ability of pMB4 to complement the rot1::kanMX2 disruption. BIG1 was amplified by PCR with the oligonucleotides 5′‐GGAGTATCTAGACAAGATGCAAAC‐3′ and the cmyc epitope‐containing oligonucleotide 5′‐CCATCGATTTAGTTCAAGTCTTCTTCAGAAATCAACTTTTGTT CAGATACTTGCTGGGATGG‐3′. A PCR‐generated 1.1 kb XbaI–ClaI fragment was cloned into pTB227 (URA3 GAL1p CEN CYC1 terminator) to yield pPAD25. pPAD25 was transformed into a big1::HIS3MX6/BIG1 diploid, yielding PA11, which was sporulated and dissected on YPGal. The cmyc‐tagged BIG1 was functional as HIS3 URA3 segregants (PA11‐6a) were obtained which grew like the wild type. Surprisingly, on glucose PA11‐6a cells grew almost like wild‐type cells, suggesting that very low levels of BIG1 were sufficient to confer near wild‐type growth.
Selection and scoring of tor2 suppressors
The tor2ts mutants used to select the rot suppressors contained a chromosomal TOR2 disruption and a YEplac181‐borne (2μ LEU2) tor2ts allele. The tor2ts strains were streaked on rich medium and incubated at 37°C. Approximately 300 colonies were picked. Eight mutants, isolated from three different tor2ts alleles (rot1 alleles: four times from tor2‐21 and twice from tor2‐31; rot2 alleles: twice from tor2‐64) were noted to form two sizes of colonies at 37°C on YPD. The cells from the smaller colonies were leu2 (had lost the LEU2 tor2ts plasmid) and were cold sensitive for growth at 30°C and below, indicating that the rot mutation in these strains suppressed a tor2‐null allele, but only at 37°C. The cells from the larger colonies were LEU2 (still carried the tor2ts gene) and grew at all temperatures. The ability of rot1 and rot2 alleles to suppress a tor2‐null allele at 37°C was confirmed by crossing cells from the small colonies to a wild‐type strain and dissecting the resulting diploids at 37°C; rot1 tor2‐null and rot2 tor2‐null segregants grew at 37°C, but were unable to grow at 30, 24 and 15°C. Thus, the rot1 and rot2 mutations, from the eight mutants, suppressed a tor2‐null but only at 37°C. The rot1‐1 and rot2‐1 mutants selected for further study were isolated from tor2‐21 and tor2‐64 mutants, respectively.
The rot1‐1 and rot2‐1 mutations in strains with an otherwise wild‐type background were scored by back‐crossing to a tor2ts mutant. Diploids from these crosses were sporulated and dissected at 37°C. Strains that yielded segregants that contained the tor2ts allele and were viable at 37°C were inferred to carry a rot mutation.
Identification of YMR200w, YBR229c, YHR101c and SAC7
YMR200w, YBR229c, YHR101c and SAC7 were isolated by their ability to rescue a rot1‐1 rot2‐1 double mutant at 37°C. High‐ (LEU2‐based, provided by K.Nasmyth) and low‐copy‐number libraries (URA3‐based, provided by A.Koller) were transformed into the double mutant. Transformants were selected at permissive temperature and replica‐plated to selective medium at non‐permissive temperature. YMR200w was isolated 11 times on three independent plasmids from the low‐copy‐number library. Deletion analysis revealed that a 1.4 kb EcoRI–SacI fragment from the insert was both necessary and sufficient for rescue of the rot1‐1 rot2‐1 double mutant. YBR229c was isolated once from the LEU2‐based library. Deletion analysis revealed that a 3.9 kb SacI–XbaI fragment from the insert was both necessary and sufficient for rescue. YHR101c was isolated three times on three independent plasmids from the LEU2‐based library. Deletion analysis revealed that a 1.6 kb FspI–ScaI fragment was both necessary and sufficient for rescue. SAC7 was isolated four times on the same plasmid from the LEU2‐based library. Deletion analysis revealed that a 3.6 kb HindIII EcoRI fragment was both necessary and sufficient for rescue.
YMR200w and YBR229c were shown to encode ROT1 and ROT2, respectively, by genetic linkage. Plasmids encoding YMR200w and YBR229c were integrated at the genomic sites corresponding to the cloned open reading frames and found to segregate opposite rot1‐1 and rot2‐1. This assignment was further confirmed by transforming the single‐copy plasmids containing the open reading frames into the rot1‐1 tor2ts and rot2‐1 tor2ts double mutants. Only YMR200w was able to complement rot1‐1 and only YBR229c was able to complement rot2‐1 (data not shown).
Staining and quantification of chitin
Cells were grown to early logarithmic phase, fixed in formaldehyde and stained with calcofluor white to visualize chitin, as described (Pringle et al., 1989). Chitin was quantified as described (Bulawa et al., 1986). All reagents were purchased from Sigma.
Analysis of ER glycoproteins
ER glycosylation of carboxypeptidase Y (CPY) was analysed as described (Stevens et al., 1982) with the following minor modifications. Cells were labelled with [35S]methionine (Amersham) and were lysed with NaOH/β‐mercaptoethanol (Volland et al., 1994). The GAL‐ROT1 mutant PA38‐2a was grown in YPraffinose and shifted to YPD for 4 h (before shifting to 37°C) to turn off expression of ROT1. The GAL‐BIG1 strain PA35‐19d was grown in YPD.
Disruption of ROT1, ROT2, BIG1, CWH41, GAS1, FKS1 and CHS3
The entire open reading frames of ROT1, ROT2, BIG1, CWH41, GAS1 and CHS3 were replaced by PCR‐generated kanMX2 or HIS3MX6 cassettes (Wach et al., 1994). The oligonucleotide pairs used for generation of the disruption cassettes were as follows, with the oligonucleotide corresponding to the 5′ end of the disrupted gene listed first. ROT1: 5′‐GGATCGGTATTATAACCATTATTATTTGTTGCACTTTATTACGA TATCAAGCTTGCCTCGTCC‐3′ and 5′‐GGGCCTCTTTCCCCCATTGGTTGAATAAGTATCCCTGTAGTGC GCGGATGGCGGCG‐ TTAGTATCG‐3′. ROT2: 5′‐CGGATATTTTTACGTTTACTTTGTAAGAAAGAGCTAAAAAGAT GTAAAGATCTGTTTAGCTTGCCTCGTCC‐3′ and 5′‐GATTGCTCTGCGTATCTTAAAATAGCGGTCT‐ CGAATCAACCGTATCGAGCTCGTTTAAACTGGATGGCG‐3′. BIG1: 5′‐GTTTTGTAGTTAGGAAACGTAGGCGCATACTAATTAGGAGTAG CAAGGCGCCAGATCTGTTTAGCTTGCC‐3′ and 5′‐GCGGTAGGTCAAATTTGCTCGAGCAATATATGTAGCTAAGAAACGT‐ TTTAATAGGTACTCTCTGAAGCG‐3′. CWH41: 5′‐CTGTTAATAAGGAAGTAGTGGATAATAACGGTTCAGGTTGCTTC ACAGATATCAAGCTTGCCTCGTCC‐3′ and 5′‐CTGGTTACATATCTGTTTATAT‐ GACTGAAAATATCTACATAGATGGATGGCGGCGTTAGTATCG‐3′. The disruption cassette for GAS1 was provided by M.Schönbächler. The disruption construct of FKS1 was provided by F.Foor (Mazur et al., 1995). All disruptions were confirmed by either Southern analysis or PCR.
GEF activity assay
GEF activity was assayed by measuring the binding of [35S]GTPγS to RHO1 as described (Yamamoto et al., 1990), with slight modifications. All procedures were at 4°C, unless stated otherwise. All strains were grown at 37°C (the temperature at which the cell wall defects are known to suppress tor2ts) in YPD to an OD600 of 0.5–0.8. SDS‐treated cells were grown in YPD plus SDS (0.005%) for 1.5 h, at 37 or 30°C, before harvesting. Whole‐cell extracts from strains to be assayed for GEF activity were prepared by resuspending cells in extraction buffer II (20 mM Tris–HCl, pH 7.5, 10 mM MgCl2, 2.5 mM EDTA, 1 mM DTT, 1 mM PMSF), lysing with glass beads in a bead‐beater and removing cell debris by centrifugation at 500 g for 10 min. GST‐RHO1 was prepared as follows. GST‐RHO1‐expressing yeast cells (JK9‐3dα/pAS53) were resuspended in extraction buffer I (50 mM Tris–HCl, pH 7.5, 100 mM NaCl, 1 mM EDTA, 0.5% NP‐40, 1 mM PMSF) and lysed with glass beads, and cell debris was removed by centrifugation at 500 g for 10 min. GST‐RHO1 was isolated by incubating the cell extract with GST–Sepharose beads (Pharmacia) for 2 h. Beads were washed with buffer III (50 mM Tris–HCl, pH 7.5, 500 mM NaCl, 1 mM EDTA, 0.5% NP‐40, 0.5% Triton X‐100, 1 mM PMSF). The GEF assay was as follows. GST‐RHO1 was washed with and then resuspended in reaction buffer (extraction buffer II), and incubated with 1 mM [35S]GTPγS (Amersham) in the presence of 200 μg of a whole cell extract to be assayed and 0.75 mM l‐α dimyristoylphosphatidylcholine for 5 min at 25°C. The reaction was stopped by adding 1 ml of ice‐cold stop buffer (20 mM Tris–HCl, pH 7.5, 25 mM MgCl2, 100 mM NaCl). The diluted mixture was centrifuged and washed twice with stop buffer. The radioactivity was quantified by scintillation counting. The GEF activity toward RHO1 in a mutant extract was expressed as a percentage of the GEF activity toward RHO1 in a wild‐type extract.
The EMBL and GenBank databases were searched for ROT1, ROT2, BIG1 and glucosidase I homologues using the BLAST program (Devereux et al., 1984).
We thank Sylvane Desrivières, Thomas Beck and Tobias Schmelzle for helpful discussions and comments on the manuscript. This work was supported by grants from the Swiss National Science Foundation and the Canton of Basel to M.N.H.
- Copyright © 1998 European Molecular Biology Organization