Cdc1p Is an Endoplasmic Reticulum-Localized Putative Lipid Phosphatase That Affects Golgi Inheritance and Actin Polarization by Activating Ca²⁺ Signaling

Yeast strains and reagents. Saccharomyces cerevisiae strains were grown in rich glucose medium (yeast extract, peptone, and dextrose [YPD]) or minimal glucose medium (synthetic dextrose [SD]) (35). SD-URA is SD medium lacking uracil. The experiments shown in Fig. 1 were performed with the haploid strain BGY103, which has the genotype leu2-3,112ura3-52rme1trp1his4GAL⁺HMLaSEC7-EGFPbar1::URA3 (33). All other experiments were performed with derivatives of the haploid strain JK9-3d, which has the genotype leu2-3,112ura3-52rme1trp1his4 (21).

• Open in new tab
• Download powerpoint

FIG. 1.

Two cdc1 alleles show different sensitivities to Mn²⁺ but still exhibit activated Ca²⁺ signaling at the restrictive temperature. (a) Wild-type (WT), cdc1-310, and cdc1-314 mutant strains were grown to stationary phase and serially diluted (1:10) five times. The strains were replica plated on YPD medium at 25°C and 37°C, at 37°C with 2 mM MnCl₂, and at 25°C with 2 mM EGTA. (b) The activation of Ca²⁺ signaling was measured using a CDRE-lacZ reporter. Wild-type, cdc1-310, and cdc1-314 cells were grown in SD-URA medium to early log phase and shifted to 33.5°C for the indicated times. Reporter activity was measured every hour for 3 h.

The cdc1-314 allele contains a T377I mutation (ACT to ATT), and the cdc1-310 allele contains an L356F mutation (ACT to ATT). Our CDC1 wild-type allele has a serine codon (ATC) at position 441 instead of a proline codon (ACC), as indicated in the Saccharomyces Genome Database. To make a cdc1-314 derivative of a given strain, the full-length CDC1 gene with a T377I mutation was integrated at the CDC1 locus to create a gene duplication. Successful integrant colonies were identified by PCR. One copy of CDC1 was then popped out using 5-fluoroorotic acid (Toronto Research Chemicals, Toronto, Canada), and the resulting colonies were screened for temperature sensitivity.

To delete the CCH1 gene, a fragment was PCR amplified from a cch1Δ strain created by the Saccharomyces Genome Deletion Project (http://www-sequence.stanford.edu/group/yeast_deletion_project/deletions3.html ). This fragment, which included the kanMX gene flanked by 400 bp upstream of the CCH1 coding sequence and 500 bp downstream of the CCH1 coding sequence, was integrated into the genome by homologous recombination. A similar strategy was used to delete the MPK1 gene. The PER1 and GUP1 genes were deleted by replacement with LEU2 using PCR-based homologous recombination (3).

The inhibitor FK506 (LC Laboratories, Woburn, MA) was stored at −20°C in ethanol at 10 mg/ml. The final concentration of FK506 in cell cultures was 1 μg/ml.

Microscopy.Immunofluorescence microscopy was performed as previously described (33). Rabbit polyclonal anti-Kar2p antibody (a gift from Jack Rose) was used at a 1:5,000 dilution. Mouse monoclonal anti-myc antibody (clone 9E10; Roche, Indianapolis, IN) was used at 2 μg/ml.

Lipid analysis.Phospholipid radiolabeling, extraction, and separation by two-dimensional thin-layer chromatography (TLC) were performed as previously described (36). The lipid extracts were not dried with nitrogen because the deacetylation step was not needed. After extraction, the lipids were resuspended in solvent A (95% ethanol, water, diethyl ether, pyridine, concentrated ammonium hydroxide [15:15:5:1:0.018, by volume]) and resolved on TLC plates (Silica Gel 60-precoated plates for TLC; EMD, Germany). The results were quantified using a Molecular Dynamics Storm 860 PhosphorImager (GE Healthcare, Piscataway, NJ) with ImageQuant, version 1.2, software. For the lipid marked with the arrows in Fig. 6a, the signal was measured as a fraction of the total radioactivity present in all of the labeled species that migrated above the origin.

β-Galactosidase and Golgi inheritance assays.β-Galactosidase reporter assays were performed using the CDRE-lacZ (38) and MPK1-lacZ (18) reporter plasmids. The assays were performed using a yeast β-galactosidase assay kit (Pierce, Rockford, IL). Error bars in the figures represent standard errors of the means.

Golgi inheritance was quantified as previously described (33). In brief, we measured the percentage of buds that contained late Golgi membranes as a function of bud size. Bud size classes III, IV, and V were combined into one class, termed III-V.

EndoH analysis.Total cellular protein was extracted by adjusting the culture to 10% trichloroacetic acid, incubating at 60°C for 5 min, chilling on ice for 5 min, centrifuging for 5 min at 3,000 × g, and resuspending in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer supplemented with 50 mM Na⁺-PIPES [piperazine-N,N′-bis(2-ethanesulfonic acid); pH 7.5]. The samples were split into endoglycosidase H (EndoH)-treated and -untreated samples. EndoH-treated samples were incubated in 10 μl of EndoH buffer (Prozyme, San Leandro, CA), boiled for 10 min, chilled on ice for 5 min, treated with 2.5 μl of EndoH (30 μU/μl) (Prozyme), incubated at 37°C overnight, boiled for 10 min, and examined using SDS-PAGE. Untreated samples were boiled for 10 min in SDS-PAGE buffer and examined using SDS-PAGE. The separated proteins were transferred to polyvinylidene difluoride membranes, and tagged Cdc1p-myc was detected with a SuperSignal West Femto kit (Pierce, Rockford, IL) using peroxidase-conjugated anti-c-myc antibody (clone 9E10; Roche, Indianapolis, IN) at 1:1,000.

Role of Mn²⁺ in Cdc1p function.We reasoned that if Cdc1p is a Mn²⁺-dependent protein that affects Ca²⁺ levels, then the Mn²⁺ and Ca²⁺ links could potentially be uncoupled in some cdc1 mutants. The previously described cdc1 strains had elevated intracellular Ca²⁺ at the restrictive temperature (25, 28). In addition, these strains were rescued by supplementing the medium with Mn²⁺ and were sensitive to the divalent cation chelator EGTA (28). Even though EGTA is not a specific Mn²⁺ chelator, the effect on cdc1 mutants was thought to be Mn²⁺ related because EGTA sensitivity could be rescued by Mn²⁺ addition or by overexpression of the Mn²⁺ transporter Smf1p (28). Thus, Mn²⁺ rescue and EGTA sensitivity are hallmarks of the link between Mn²⁺ and CDC1. We looked for new cdc1 alleles that did not exhibit Mn²⁺ rescue and EGTA sensitivity but still activated Ca²⁺ signaling. Fourteen different cdc1 mutants (33) were tested for their ability to grow in the presence of 2 mM MnCl₂ at 37°C or in the presence of 2 mM EGTA at 25°C. These mutants showed a wide range of sensitivities to EGTA, and some of them could be rescued by Mn²⁺. One mutant, cdc1-310, showed the best rescue response to Mn²⁺ and one of the more severe sensitivity responses to EGTA (Fig. 1a). Conversely, the cdc1-314 mutant was not rescued by Mn²⁺ and exhibited only a mild growth defect in the presence of EGTA (Fig. 1a). Despite having such different responses to Mn²⁺ and EGTA, the cdc1-314 and cdc1-310 strains had similarly elevated Ca²⁺ signaling pathway activation levels at the restrictive temperature as measured by the calcineurin-dependent response element-lacZ fusion reporter, CDRE-lacZ (38) (Fig. 1b). The assays were performed at 33.5°C to avoid any heat shock-related responses. The results indicate that the connections to Mn²⁺ and Ca²⁺ can be separated and favor the hypothesis that Cdc1p is a Mn²⁺-dependent protein that affects Ca²⁺ signaling.

This conclusion is also supported by sequence alignment data suggesting that Cdc1p has a metal ion-dependent phosphoesterase domain (34). We set out to test the importance of this domain for the function of Cdc1p. The residues Asp95, Asp144, Asn183, His184, His259, and His316 are predicted to chelate the divalent cation in the active site, so we mutated these residues individually to alanine (Fig. 2a). Whereas wild-type CDC1 on a CEN plasmid rescued the cdc1-314 mutant at the restrictive temperature, plasmids carrying the mutated versions of CDC1 failed to rescue (Fig. 2b). As a control, growth was rescued by a plasmid carrying a mutation of His287, which is not predicted to chelate the divalent cation. The wild-type and mutant Cdc1p proteins had similar localizations (see below) and expression levels (see Fig. S1 in the supplemental material). These results confirm that the putative divalent cation-chelating residues are important for Cdc1p function, suggesting that Cdc1p is indeed a phosphoesterase.

• Open in new tab
• Download powerpoint

FIG. 2.

The putative phosphoesterase domain of Cdc1p is functionally important and faces the lumen of the ER. (a) Sequence analysis revealed a putative phosphoesterase domain (red box) and three hydrophobic putative transmembrane domains (green boxes). The conserved amino acid residues labeled in red are predicted to chelate the divalent metal ion. (b) Cdc1p mutants with alanine substitutions of the conserved residues indicated in panel a were expressed from a CEN plasmid in the cdc1-314 strain at 25°C and 37°C. A mutant of His287, which is not predicted to chelate the divalent metal ion, was used as a control. Lack of growth at 37°C shows that the mutated residues are important for Cdc1p function. (c) Cells expressing myc-tagged Cdc1p were immunolabeled with anti-myc antibody (red) and anti-Kar2p antibody (green). Scale bar, 5 μm. (d) The five potential N-linked glycosylation sites in Cdc1p were mutated separately or in combination in a myc-tagged Cdc1p protein, which was expressed as a second copy in the wild-type (WT) strain. The cell extracts were treated with EndoH and examined using SDS-PAGE. The lack of a mobility shift after EndoH treatment in the N215Q N233Q double mutant indicates that these residues are N glycosylated and are therefore in the lumen of the secretory pathway. (e) A diagram showing the inferred topology of Cdc1p.

The known Cdc1p temperature-sensitive mutations are in the putative phosphoesterase domain (33) (see Materials and Methods). Presumably, these mutations destabilize Cdc1p, and high Mn²⁺ levels can partially overcome the destabilization for alleles such as cdc1-310 but not for alleles such as cdc1-314. To avoid complications due to Mn²⁺ sensitivity, we used the cdc1-314 allele for the remainder of this study.

Cdc1p localization and topology.What might be the substrate of Cdc1p? To address this question, we examined the localization of myc-tagged Cdc1p. Surprisingly, Cdc1p-myc colocalized with the ER marker Kar2p (32) (Fig. 2c). A hydrophobicity analysis revealed that Cdc1p has three predicted transmembrane domains, implying that Cdc1p is an integral membrane protein of the ER. The phosphoesterase domain is between the first and second hydrophobic domains (Fig. 2a). To determine whether the phosphoesterase domain faces the cytosol or the ER lumen, we took advantage of the fact that N-linked glycosylation occurs only in the ER lumen (1). If a protein is glycosylated, treating it with EndoH will cleave off the N-linked glycan (26) and reduce the molecular weight. To determine if a potential N-linked glycosylation site is, in fact, glycosylated, we can mutate that site and ask whether EndoH treatment of the resulting protein shifts the molecular weight. Cdc1p has five potential N-linked glycosylation sites. When we simultaneously mutated Asn5, Asn434, and Asn461 to Gln and then treated with EndoH, we saw the same shift in molecular weight that was seen with the wild-type protein (Fig. 2d). When we individually mutated Asn215 or Asn233, which are located in the phosphoesterase domain, we observed a smaller shift (Fig. 2d). When Asn215 and Asn233 were simultaneously mutated, the molecular weight shift disappeared completely, and the EndoH-treated and -untreated mutant proteins had the same mobility as the EndoH-treated wild-type protein (Fig. 2d). These results indicate that only the phosphoesterase domain residues Asn215 and Asn233 are glycosylated, implying that the phosphoesterase domain faces the lumen of the secretory pathway. Our topology assignment is supported by previous data showing that the C terminus of Cdc1p is lumenally oriented (20). Together, these findings demonstrate that the phosphoesterase domain of Cdc1p faces the ER lumen (Fig. 2e).

Mpk1p activation in cdc1 cells.Our data suggest that inhibiting the enzymatic function of Cdc1p activates Ca²⁺ signaling. One possibility is that inhibition of Cdc1p triggers a cell stress response. This idea fits with the actin depolarization seen in cdc1 mutants (33) because actin becomes depolarized in cells experiencing stress (23). We tested whether cdc1 mutants were experiencing cell stress by examining the activity of Mpk1p, a key component of the cell stress signal transduction pathway (18). These experiments employed an MPK1-lacZ reporter (18, 23). The results showed that Mpk1p is activated in a cdc1 mutant at the restrictive temperature, consistent with a cell stress response (Fig. 3a).

• Open in new tab
• Download powerpoint

FIG. 3.

Mpk1p is activated in the cdc1-314 mutant but is not the major cause of Ca²⁺ signaling activation. (a) An MPK1-lacZ reporter was used to measure Mpk1p activity in cdc1-314 mutant cells. Wild-type (WT) and mutant cells were grown in SD-URA medium to log phase and then shifted to 33.5°C for the indicated times. Reporter activity was measured every hour for 3 h. (b) A CDRE-lacZ reporter was used to measure the activation of Ca²⁺ signaling. Wild-type, cdc1-314, mpk1Δ, and mpk1Δ cdc1-314 cells were grown in SD-URA medium to early log phase and shifted to 33.5°C for the indicated times. Reporter activity was measured every hour for 3 h.

The results presented thus far suggested that loss of Cdc1p function might cause cell stress and activate Mpk1p, which would then activate Ca²⁺ signaling. This interpretation seemed reasonable because it has been proposed that under some stress conditions, Mpk1p activates the plasma membrane Ca²⁺ channel Cch1p (4). To test whether Mpk1p activates Cch1p in a cdc1 mutant, we deleted MPK1 and then monitored Ca²⁺ signaling activation using the CDRE-lacZ reporter. Surprisingly, deletion of MPK1 caused only a small decrease in cdc1-induced Ca²⁺ signaling activation (Fig. 3b), implying that in cdc1 cells, Cch1p is activated primarily by an Mpk1p-independent pathway. These results argue against the hypothesis that a generalized cell stress response can explain the Ca²⁺ signaling activation in cdc1 mutant cells.

Relationship of Ca²⁺ signaling activation to the cdc1 mutant phenotypes.Because cdc1 mutations activate both Ca²⁺ signaling and Mpk1p, we wanted to know which of these pathways mediates the additional cdc1 phenotypes of defective Golgi inheritance and depolarization of actin. For this purpose, we deleted the Ca²⁺ channel Cch1p and performed a Golgi inheritance assay. The results showed that loss of Cch1p completely suppressed both the Golgi inheritance and actin polarization phenotypes (Fig. 4a and b). By contrast, when we deleted MPK1 in a cdc1 mutant, actin was still depolarized (Fig. 4b). We were unable to test the effect of MPK1 deletion on Golgi inheritance because many of the small-budded cells in the mpk1Δ mutant had lysed. Nevertheless, these results clearly indicate that in cdc1 mutants, Cch1p-dependent activation of Ca²⁺ signaling leads to defective Golgi inheritance and depolarization of actin.

• Open in new tab
• Download powerpoint

FIG. 4.

Deletion of the Ca²⁺ channel Cch1p completely suppresses the Golgi inheritance and actin polarization phenotypes in the cdc1-314 mutant. (a) The inheritance of late Golgi cisternae was measured by assessing the presence of the late Golgi marker protein Sec7p-GFP in the buds of wild-type (WT), cdc1-314, cch1Δ cdc1-314, and cch1Δ cells after a shift to 33.5°C for 3 h. (b) Actin was visualized by staining with Alexa 594-phalloidin. Wild-type, cdc1-314, cch1Δ cdc1-314, cch1Δ, mpk1Δ, and mpk1Δ cdc1-314 cultures were grown at 25°C, shifted to 33.5°C for 3 h, fixed, and stained. Fluorescence images are shown. Most of the wild-type, cch1Δ cdc1-314, cch1Δ, and mpk1Δ cells contained polarized actin patches and cables, whereas most of the cdc1-314 and mpk1Δ cdc1-314 cells contained depolarized patches and few visible cables. Scale bar, 5 μm.

In a further test of the importance of Ca²⁺ signaling activation for the cdc1 mutant phenotypes, we introduced the per1Δ mutation, which strongly suppresses the growth defect of cdc1 strains (27). As expected, a per1Δ cdc1 double mutant showed no Ca²⁺ signaling activation at the restrictive temperature (see Fig. S2 in the supplemental material).

One effector of Cch1p-dependent Ca²⁺ influx is the Ca²⁺/calmodulin-dependent phosphatase calcineurin (9, 29). By activating the transcription factor Crz1p, calcineurin turns on many genes (9). We tested whether the growth and actin polarization phenotypes of cdc1 cells are caused by calcineurin activation. This experiment employed the potent calcineurin inhibitor FK506 (9). When cdc1 cells were incubated at 33.5°C, the addition of FK506 did not rescue the growth defect, nor did it prevent actin depolarization (Fig. 5b and c). As a control, FK506 was added to cdc1 cells, where it completely suppressed the elevation of CDRE-lacZ activity (Fig. 5a). Thus, the growth and actin polarization phenotypes seen in cdc1 mutants are not due to the activation of calcineurin but must be mediated by a separate Ca²⁺-dependent pathway.

• Open in new tab
• Download powerpoint

FIG. 5.

Treatment of cdc1-314 cells with the calcineurin inhibitor FK506 does not suppress the growth and actin polarization phenotypes. (a) Activation of Ca²⁺ signaling was measured using the CDRE-lacZ reporter. Wild-type (WT) and cdc1-314 cells were grown in 10 ml of SD-URA medium to early log phase and shifted to 33.5°C for the indicated times. Immediately before the shift, 10 μl of dimethyl sulfoxide or 1 mg/ml of FK506 in dimethyl sulfoxide was added. The final concentration of FK506 in cultures was 1 μg/ml. Reporter activity was measured every hour for 3 h. (b) Cell growth was measured by taking the absorbance at 600 nm. Cultures were treated as described in panel a. Readings were taken every hour for 3 h. (c) Actin was visualized by staining with Alexa 594-phalloidin. Cultures were treated as described in panel a. After 3 h, the cells were fixed and stained. Fluorescence images are shown. Most of the treated and untreated wild-type cells contained polarized actin patches and cables, whereas most of the treated and untreated cdc1-314 cells contained depolarized patches. Scale bar, 5 μm.

Accumulation of a phospholipid in cdc1 cells.To understand how Cdc1p affects Ca²⁺ signaling, it is important to clarify the enzymatic function of Cdc1p. The topology assignment constrains the possible substrates for the phosphoesterase domain. Proteins are unlikely substrates because, to our knowledge, there are no protein kinases or phosphatases in the ER lumen. However, several lipid phosphatases are present in this compartment (11, 19). We therefore hypothesized that a phosphorylated lipid might be the substrate for Cdc1p. To test this idea, we radiolabeled wild-type and cdc1 cells with ³²P_i, extracted the radioactive lipids, and performed two-dimensional TLC analysis. A phospholipid species reproducibly accumulated in the cdc1 mutant at the restrictive temperature (Fig. 6a). Quantitation indicated that after 3 h at 33.5°C, this unidentified lipid was approximately twice as concentrated in cdc1 cells as in wild-type cells (Fig. 6b).

• Open in new tab
• Download powerpoint

FIG. 6.

A phospholipid species accumulates in the cdc1-314 mutant at the restrictive temperature. (a) Wild-type (WT) and mutant strains were grown overnight in YPD medium supplemented with ³²P_i. Upon reaching log phase, the cells were shifted to 33.5°C for 3 h. Lipids were extracted and run in two dimensions on TLC plates. These images from representative plates show the area where the major radioactive lipid species migrated. Arrows mark an unidentified phosphorylated lipid that was consistently present at higher levels in cdc1 cells than in wild-type cells. (b) The level of the unidentified radioactive lipid was measured in each of the indicated strains as a fraction of the total radioactive lipid, and the results were normalized by defining the signal from wild-type cells as 1.0.

We postulate that in cdc1 mutants, abnormal accumulation of a lipid leads to hyperactivated Cch1p-dependent Ca²⁺ signaling, which in turn causes toxicity. This model predicts that a cdc1 strain carrying a cch1Δ suppressor mutation should still accumulate the unidentified lipid. Indeed, in a cch1Δ cdc1 double mutant, the unidentified lipid was about twice as concentrated as in the cch1Δ single mutant (Fig. 6b). Interestingly, in a per1Δ strain the unidentified lipid was about twice as concentrated as in wild-type cells, and this concentration was not enhanced in a per1Δ cdc1 double mutant (Fig. 6b). It therefore seems likely that the per1Δ and cch1Δ mutations suppress the cdc1 phenotypes by different mechanisms.