Protein kinase Ypk1 phosphorylates regulatory proteins Orm1 and Orm2 to control sphingolipid homeostasis in Saccharomyces cerevisiae

Abstract

The Orm family proteins are conserved integral membrane proteins of the endoplasmic reticulum that are key homeostatic regulators of sphingolipid biosynthesis. Orm proteins bind to and inhibit serine:palmitoyl-coenzyme A transferase, the first enzyme in sphingolipid biosynthesis. In Saccharomyces cerevisiae, Orm1 and Orm2 are inactivated by phosphorylation in response to compromised sphingolipid synthesis (e.g., upon addition of inhibitor myriocin), thereby restoring sphingolipid production. We show here that protein kinase Ypk1, one of an essential pair of protein kinases, is responsible for this regulatory modification. Myriocin-induced hyperphosphorylation of Orm1 and Orm2 does not occur in ypk1∆ cells, and immunopurified Ypk1 phosphorylates Orm1 and Orm2 robustly in vitro exclusively on three residues that are known myriocin-induced sites. Furthermore, the temperature-sensitive growth of ypk1^ts ypk2∆ cells is substantially ameliorated by deletion of ORM genes, confirming that a primary physiological role of Ypk1-mediated phosphorylation is to negatively regulate Orm function. Ypk1 immunoprecipitated from myriocin-treated cells displays a higher specific activity for Orm phosphorylation than Ypk1 from untreated cells. To identify the mechanism underlying Ypk1 activation, we systematically tested several candidate factors and found that the target of rapamycin complex 2 (TORC2) kinase plays a key role. In agreement with prior evidence that a TORC2-dependent site in Ypk1(T662) is necessary for cells to exhibit a wild-type level of myriocin resistance, a Ypk1(T662A) mutant displays only weak Orm phosphorylation in vivo and only weak activation in vitro in response to sphingolipid depletion. Additionally, sphingolipid depletion increases phosphorylation of Ypk1 at T662. Thus, Ypk1 is both a sensor and effector of sphingolipid level, and reduction in sphingolipids stimulates Ypk1, at least in part, via TORC2-dependent phosphorylation.

In eukaryotic cells, the lipid phase of the plasma membrane is highly regulated in terms of overall lipid content and the distinct lipid compositions of the inner and outer leaflets (1, 2). This organization is maintained despite continuous flux due to insertion of exocytic vesicles and removal of endocytic vesicles, as well as lipid biosynthetic and degradative processes. How a cell detects changes in membrane composition and dynamically regulates lipid biosynthetic output and transbilayer translocation of lipids to maintain membrane homeostasis are central questions in cell biology. Sphingolipids, in particular, are essential structural components of membranes and critical signaling molecules whose levels must be closely regulated.

We recently showed that sphingolipid homeostasis is dependent on a feedback pathway mediated by Orm family proteins. Specifically, in the budding yeast Saccharomyces cerevisiae, Orm1 and Orm2 are inhibitors of serine:palmitoyl-coenzyme A transferase (SPT), encoded by LCB1 and LCB2 (3), which catalyzes the first committed step in de novo sphingolipid synthesis. When sphingolipid biosynthesis is compromised, Orm proteins are inactivated by phosphorylation, which alleviates their inhibition of SPT and allows for a compensatory increase in production of sphingolipid precursors (4). However, the mechanism by which cells sensed sphingolipid levels and correspondingly adjusted Orm protein phosphorylation was not known.

Mutations in components of the target of rapamycin complex 2 (TORC2) kinase complex were isolated in a classic genetic screen for regulators of sphingolipid synthesis (5), and TORC2 activity promoted ceramide production (6). Despite a growing body of evidence from yeast and other organisms linking TORC2 to lipids (7–9), a precise link between TORC2 substrates and sphingolipid metabolism had not been established. However, we and others demonstrated that two paralogous protein kinases, Ypk1/Ykl126w and Ypk2/Ymr104c (mammalian orthologs are SGK1, SGK2, and SGK3) (10), now known to be TORC2 targets (11, 12), have essential roles in processes linked to membrane dynamics (13–15). In particular, we had previously implicated Ypk1 function in the regulation of sphingolipid biosynthesis in two ways. First, Ypk1 overexpression confers resistance to the SPT inhibitor myriocin, whereas loss of YPK1 causes profound hypersensitivity to this drug (16, 17). Second, Ypk1 is a pivotal component of regulatory circuitry that imposes a balance between the pool of complex sphingolipids and the amount of aminophospholipids in the outer leaflet (12). Specifically, among transposon insertions that improve the temperature-sensitive growth of ypk1^ts ypk2Δ cells (14), we found loss-of-function mutations in DNF3, one of five genes encoding inward-directed P-type ATPases (flippases) that transport aminophospholipids (18, 19), suggesting that one role of Ypk1 is to inhibit lipid flippase action, directly or indirectly. Indeed, Ypk1 does not act on the flippases directly, but does so indirectly by phosphorylating and thereby inhibiting the flippase protein kinase 1 (Fpk1; and its paralog Fpk2), which are known flippase activators (20) that require a complex sphingolipid, mannosylinositol phosphorylceramide (MIPC), for their optimal function (12).

Most significantly, in addition to transposon insertions in DNF3, a hit in ORF YLR350w, now designated the ORM2 gene, was also recovered (14), but nothing was known about its function at the time. Given the aforementioned role of Orm proteins in the control of sphingolipid metabolism, these observations established a more direct connection between sphingolipids and Ypk1 function. In particular, the fact that a presumptive loss-of-function mutation in ORM2 largely suppressed the temperature-sensitive growth of ypk1^ts ypk2Δ cells (14) made Ypk1 a prime candidate to be the protein kinase responsible for the observed sphingolipid depletion-induced phosphorylation of Orm2 (and Orm1). Here we show that both Orm proteins are direct physiological targets of Ypk1. Moreover, we have uncovered a primary mechanism by which Ypk1-mediated phosphorylation of the Orm proteins is stimulated by TORC2 in response to inhibition of sphingolipid synthesis. These findings reveal critical processes in the regulatory pathways that control sphingolipid homeostasis.

Results and Discussion

Orm Proteins Are Substrates of Ypk1.

We previously showed that Orm1 and Orm2 are inactivated by phosphorylation upon sphingolipid depletion, and mapped the primary phosphorylation sites involved to S51, S52, and S53 in Orm1 (and S46, S47, and S48 in Orm2) (4) (Fig. 1A). The protein kinase responsible for these modifications was not identified; however, several findings pinpointed Ypk1 as the most likely candidate. First, all six phosphorylated Ser residues lie in a sequence context that corresponds to the optimal consensus phosphoacceptor motif for Ypk1 [-R-x-R-x-x-S/T-(Hpo)-, where (Hpo) indicates a modest preference for a hydrophobic residue] (10) (Fig. 1B). Second, the phosphorylation state of Ypk1 itself is responsive to sphingolipid levels (12, 21). Third, Ypk1 activity mediates resistance to the SPT inhibitor myriocin (16). Fourth, the fact that inactivation of Orm2 via a transposon insertion significantly rescued the temperature-sensitive growth of ypk1^ts ypk2∆ cells (14) suggested that a primary function of Ypk1 is to prevent Orm-mediated SPT inhibition. Lack of Ypk1 would therefore impede growth by prohibiting an adequate rate of de novo sphingolipid synthesis. This scenario predicted that this deleterious effect of the lack of Ypk1 should be substantially ameliorated by lowering the amount of the Orm proteins. Indeed, we found that the ability of ypk1^ts ypk2Δ cells to grow at an otherwise restrictive temperature was markedly improved by introduction of an orm2∆ mutation, and further improved in cells lacking both Orm1 and Orm2 (Fig. 1C), corroborating the results obtained in the original transposon screen. Thus, absence of Orm1 and Orm2 compensates for Ypk deficiency, in full agreement with the conclusion that Ypk1 is the protein kinase responsible for phosphorylating and negatively regulating the capacity of the Orm proteins to inhibit SPT.

Fig. 1.

Genetic epistasis indicates Ypk action negatively regulates Orm function. (A) Primary structure (Left) and predicted topology (Right) of Orm proteins. Presumptive transmembrane segments (green); candidate Ypk1 sites (red). (B) Alignment of the cytosolic N termini of Orm1 and Orm2. Identities (bold blue); conservative substitutions (bold black); three nested consensus Ypk1 phosphorylation sites (bold red). (C) Absence of Orm1 and Orm2 compensates for Ypk1 deficiency. Serial 10-fold dilutions of strains of the indicated genotypes were spotted on yeast extract peptone dextrose (YPD) plates and photographed after incubation for 2 d at the indicated temperatures.

To determine if Ypk1 is capable of phosphorylating Orm1 and Orm2, we tested the ability of active Ypk1, enriched from yeast cell extracts by immunoprecipitation, to phosphorylate the N-terminal cytosolic segments of Orm1 and Orm2, expressed as fusions to GST and purified from Escherichia coli. In such in vitro assays, both GST-Orm1(1–85) (Fig. 2A, Left) and GST-Orm2(1–80) (Fig. 2A, Right) were efficiently phosphorylated by Ypk1. GST-Orm1(1–85) was phosphorylated by an analog-sensitive mutant, Ypk1(L424A), or Ypk1-as, in the absence, but not in the presence, of the appropriate inhibitor (3-MOB-PP1) (22). By contrast, phosphorylation of the same substrate by wild-type Ypk1 was unaffected, verifying that the activity in the immunoprecipitates responsible for the observed phosphorylation was indeed Ypk1. All experiments were conducted with both GST-Orm1(1–85) and GST-Orm2(1–80) with virtually identical results; but, for clarity, we present the further in vitro analysis conducted just with Orm1. In other studies below, these two substrates were used interchangeably.

Fig. 2.

Ypk1 phosphorylates the N terminus of Orm1 exclusively at Ser51, Ser52, and Ser53. (A) GST-Orm1(1–85) or GST-Orm2(1–80), purified from E. coli, were incubated with [γ-³²P]ATP and either Ypk1 or an analog-sensitive (as) variant, Ypk1(L424A), purified from S. cerevisiae, in the absence or presence of 3-MOB-PP1, and the products resolved by SDS/PAGE and analyzed as described in Materials and Methods. (B) As in A, except that ATP was used and the products were analyzed by immunoblotting with anti–phospho-AKT substrate antibody (Cell Signaling Technology, Inc.) and with anti-GST antibody. (C) As in A, except that GST-Orm1(1–85) contained the indicated site-directed mutations.

We demonstrated that the observed Ypk1-dependent phosphorylation of GST-Orm1(1–85) occurs at its predicted Ypk1 sites using two independent methods. First, using a commercial phosphorylation site-specific antibody that recognizes the phosphorylated form of the Ypk1 motif, we found that this epitope was present in GST-Orm1(1–85) phosphorylated by active Ypk1 and Ypk1-as, but not when Ypk1-as was inhibited by 3-MOB-PP1 (Fig. 2B). Site-directed mutagenesis further demonstrated that incorporation of radiolabeled phosphate into GST-Orm1(1–85) was progressively reduced by successive mutagenesis of Ser51, Ser52, and Ser53, and totally lost when just those three residues were converted to Ala (Fig. 2C). Thus, these three Ser residues are the only sites phosphorylated by Ypk1 in the Orm1 N terminus, despite the presence of 17 other Ser and Thr residues in this region, some of which are detectably phosphorylated in vivo (4).

We took advantage of an electrophoretic mobility shift assay to confirm that Ypk1 mediates the sphingolipid depletion-evoked phosphorylation of the Orm proteins in vivo. After treatment of cells with the SPT inhibitor myriocin (23), (FLAG)₃-Orm1 undergoes a time-dependent conversion to multiple, slower-migrating species in immunoblots of the corresponding cell lysates resolved by phosphate affinity SDS/PAGE (24) (Fig. 3A, Left). As shown previously (4), these bands correspond to phosphorylated species of Orm1. Strikingly, the appearance of these slower-migrating bands was almost totally abrogated in a ypk1∆ mutant (Fig. 3A, Center). Absence of these bands cannot be attributed to reduced permeability to the drug, because ypk1∆ cells are significantly more sensitive to the growth-inhibitory effect of myriocin than otherwise isogenic wild-type cells (12, 16). The slight amount of slower-migrating Orm1 species in ypk1∆ cells presumably reflects residual action of the Ypk1 paralog, Ypk2/Ykr2 (14, 16). However, given that Orm1 underwent robust myriocin-induced phosphorylation in a ypk2∆ mutant (Fig. 3A, Right), it is clear that Ypk1 is the primary isoform responsible for Orm1 and Orm2 phosphorylation in vivo. Consistent with this conclusion, we showed previously that the three residues demonstrated here to be Ypk1 sites are necessary for the myriocin-induced mobility shift exhibited by the Orm proteins (4).

Fig. 3.

Ypk1 is required for Orm phosphorylation and is activated by sphingolipid depletion. (A) Wild-type (YDB146), ypk1∆ (YDB344), and ypk2∆ (YDB340) cells, each expressing (FLAG)₃-Orm1 from the ORM1 promoter at the ORM1 locus, were grown to midexponential phase in YPD and then treated with myriocin (0.4 μM). At the indicated times, samples were withdrawn, lysates prepared, and the resulting extracts resolved and analyzed with anti-FLAG antibodies as described previously (4). (B) Wild-type (BY4741) cells expressing from the GAL1 promoter either Ypk1-myc or a catalytically inactive (KD) mutant, Ypk1(K376A)-myc, were grown to midexponential phase, and a portion of each culture was then treated with (+) myriocin in methanol (1.25 μM final concentration) or (−) an equal volume of the same solvent only. After 2 h, all of the samples were lysed and Ypk1 was recovered by immunoprecipitation with mouse ascites fluid containing anti–c-Myc mAb 9E10. The resulting immunoprecipitates were then incubated with [γ-³²P]ATP and GST-Orm2(1–80) as described in Materials and Methods.

Inhibition of Sphingolipid Biosynthesis Activates Ypk1.

Sphingolipid depletion-evoked Orm phosphorylation by Ypk1 could occur by a variety of mechanisms. Potentially, the reduction in the sphingolipid level could affect Orm accessibility, Ypk1 localization, and/or Ypk1 activity. To differentiate among these possibilities, we examined whether there was any change in the catalytic potency of Ypk1 after inhibiting sphingolipid biosynthesis. Using GST-Orm2 as a substrate, we found that the specific activity of wild-type Ypk1, but not of a catalytically inactive (kinase-dead or KD) mutant, Ypk1(K376A), immunoprecipitated from myriocin-treated cells, was markedly and reproducibly higher than the equivalent amount of enzyme immunoprecipitated from untreated control cells (Fig. 3B). This finding indicates that a major (and perhaps sole) cause of the observed increase in Orm protein phosphorylation in response to inhibition of sphingolipid production is an increase in Ypk1 kinase activity. Hence, we sought to determine the molecular basis for Ypk1 activation.

We demonstrated previously that truncation of the N-terminal regulatory domain of Ypk1 increases the activity of its kinase domain (14), and it has been reported that Ypk1 undergoes proteolysis under other conditions (such as nitrogen limitation) that inhibit growth (25). Because we immunoprecipitated C-terminally tagged Ypk1-myc from control and myriocin-treated cells, it was theoretically possible that the major species present had undergone a cleavage that removed its N-terminal domain. However, immunoblotting of both total cell extracts and the immunoprecipitated enzyme demonstrated that there was no change after myriocin treatment.

Ypk1 is inactive unless phosphorylated on a conserved Thr residue (T504) in its activation loop by the upstream kinase Pkb-activating kinase homolog 1 (Pkh1; or its paralog Pkh2) (10, 16). Hence, it was possible that the fraction of Ypk1 molecules phosphorylated at this site was increased after inhibition of sphingolipid synthesis with myriocin. However, we demonstrated previously that neither drastic reduction nor drastic increase of phytosphingosine had any discernible effect on Pkh-mediated phosphorylation of Ypk1 at T504 in vivo (12).

Also, it has been claimed recently that Ypk1 binds to and is activated by ergosterol (26). Hence, it was hypothetically possible that sphingolipid limitation enhanced the encounter between Ypk1 and this putative essential activator. However, we found that myriocin-induced activation of Ypk1 occurred robustly in an erg4∆ mutant (incapable of ergosterol production) and is, in fact, even greater than in wild-type cells (SI Appendix, Fig. S1A, Upper and Lower). Hence, ergosterol does not have an essential role in Ypk1 activation upon sphingolipid depletion.

Last, we previously showed that the MIPC-dependent protein kinases Fpk1 and Fpk2/Kin82 phosphorylate Ypk1 at S51 and S71, and these modifications seem to down-regulate Ypk1 function (12). Thus, relief of Fpk-mediated inhibition could provide another potential mechanism for the observed activation of Ypk1 upon inhibition of sphingolipid synthesis. However, (FLAG)₃-Orm1 phosphorylation was not constitutively higher in a fpk1∆ fpk2∆ mutant (SI Appendix, Fig. S2A) or in cells expressing an active Ypk1 mutant, Ypk1(S51A S71A)—that is, “immune” to Fpk-mediated modification (SI Appendix, Fig. S2B). Additionally, the response to myriocin was not altered in either strain (although a somewhat higher level of myriocin was used on the fpk1∆ fpk2∆ cells because they have a reduced permeability to this compound, presumably due to the absence of activation of the aminophospholipid flippases in cells lacking Fpk1 and Fpk2) (20). Moreover, as judged by capacity to phosphorylate GST-Orm2(1–80), immunoprecipitated Ypk1(S51A S71A) showed the same increase in activity as wild-type Ypk1 upon myriocin treatment (SI Appendix, Fig. S2C). Thus, relief of Fpk-mediated phosphorylation of Ypk1 plays little or no role in the activation of Ypk1 that occurs upon inhibition of sphingolipid biosynthesis.

Phosphorylation of Ypk1 by Target of Rapamycin Protein Kinase 2 Is Enhanced upon Sphingolipid Depletion.

In the Ypk1 paralog, Ypk2, phosphorylation of S641 (in the turn motif), and T659 (in the hydrophobic motif) depends on target of rapamycin protein kinase 2 (Tor2) (11), specifically in the TORC2 complex (27, 28), and absence of these modifications prevents optimal ceramide biosynthesis (6). Similarly, we (16) and others (17) found previously that a Ypk1(T662A) mutant grows normally on rich medium, but is unable to grow on the same medium containing myriocin, suggesting that phosphorylation of Ypk1 at T662 in the hydrophobic motif is critical for survival when sphingolipids become limiting. By contrast, the same T662A alteration had no effect on sensitivity of cells to a known TORC1-specific inhibitor, rapamycin (16). These observations suggested that enhanced TORC2-mediated phosphorylation at T662 might play an important role in Ypk1 activation in response to inhibition of sphingolipid synthesis.

A combination of mass spectrometry and mutational analysis allowed us to determine Ser and Thr residues in Ypk1 that, when phosphorylated, lead to slower mobility species upon phosphate-affinity SDS/PAGE. To simplify the pattern and to serve as a specific reporter of phosphorylation at T662, we used a derivative of Ypk1 in which 11 phosphosites (including its Fpk and Pkh sites) were mutated to Ala. We demonstrated that the slowest-mobility species observed with this construct corresponded to phosphorylation at T662, because mutation of this Thr to Ala abolished the appearance of this band (Fig. 4A, Left). Moreover, as expected on the basis of our prior genetic findings, phosphorylation of T662 depends on Tor2 activity, because this modification is severely diminished in a tor2^ts strain, even at permissive temperature (Fig. 4A, Right), and abrogated by treatment of cells with a TOR-directed inhibitor NVP-BEZ235 (29) (Fig. 4B).

Fig. 4.

Sphingolipid depletion induces Tor2-dependent phosphorylation of Ypk1. (A) Wild-type cells (BY4741) or an otherwise isogenic tor2^ts mutant expressing either Ypk1^11A-myc (662T) or Ypk1^11A(T662A)-myc (662A) were grown at 26 °C and then treated with (+) myriocin (1.25 μM) or (−) solvent alone. After 2 h, the cells were lysed and the resulting extracts were resolved by phosphate-affinity SDS/PAGE and analyzed by immunoblotting with anti–c-myc mAb 9E10. (B) As in A, except that the cells expressing Ypk1^11A-myc were strain JRY8012 (pdr5∆ snq2∆ yor1∆) to reduce ATP-binding cassette transporter-mediated drug efflux. Cells were grown at 30 °C and treated with Tor inhibitor NVP-BEZ235 in DMSO (2 μM final concentration) or with an equal volume of the same solvent alone (0). When both drugs were present, myriocin and NVP-BEZ235 were added at the same time. (C) Serial 10-fold dilutions of tor2^ts or wild-type (BY4741) cells carrying pRS316 (empty vector) or expressing from the same vector a hyperactive Ypk1 variant, Ypk1(D242A) (pFR273), were spotted on plates lacking (Left) or containing myriocin (0.6 μM) (Right). After incubation for 2 d at the indicated temperatures, the plates were photographed. (D) Wild-type (BY4741) cells carrying pRS316 (vector) or expressing Ypk1(D242A) from the same vector were plated as lawns on plates lacking (0) or containing the indicated final concentration (0.3 or 0.6 μM) of myriocin and then overlaid with sterile filter discs on which was spotted in the same final volume (10 μL) either solvent alone (0) or 1- or 10-μL samples of a stock (10 mM) of NVP-BEZ235 in DMSO. After incubation at 30 °C for 3 d, the plates were photographed.

Most importantly, we observed that the level of the species corresponding to Ypk1 phosphorylated at T662 was elevated after cells were treated with myriocin (Fig. 4 A and B, Left), indicating that TORC2-dependent modification of Ypk1 at that site is negatively regulated by sphingolipids. These results are in agreement with our other findings, and support the model that a reduction in sphingolipid synthesis increases, in a TORC2-dependent manner, the amount of Ypk1 phosphorylated at T662. This, in turn, increases the efficiency of Ypk1-mediated phosphorylation of the Orm proteins, thereby alleviating their inhibition of SPT and allowing for a compensatory increase in flux through the sphingolipid pathway. In this regard, it has been suggested that a plasma membrane-localized tetraspanin, Nce102, is a critical component for sensing sphingolipid sufficiency (30); however, in our study, a nce102∆ mutant displayed myriocin-induced Ypk1-dependent Orm phosphorylation in a manner indistinguishable from otherwise isogenic wild-type cells (SI Appendix, Fig. S3).

Homeostatic Phosphoregulation of Orm Proteins by Ypk1 Requires Tor2.

The model we described herein predicts that, similar to cells expressing mutant Orm1 and Orm2, in which the Ypk1-dependent sites have been mutated to Ala (4), a Tor2-deficient strain should be hypersensitive to the growth-inhibitory effect of myriocin because it cannot activate Ypk1. Moreover, if the problem is lack of Ypk1 activation, then a constitutively hyperactive Ypk1 variant should ameliorate the myriocin sensitivity of Tor2-deficient cells. Consistent with both of these predictions, we found, first, that a tor2^ts strain is indeed hypersensitive to myriocin both at restrictive and permissive temperatures (Fig. 4C). Second, this sensitivity was substantially rescued, even at restrictive temperature, by expression of Ypk1(D242A) (Fig. 4C). Ypk1(D242A) is a gain-of-function allele of Ypk1 that we constructed on the basis of the prior demonstration that the corresponding Ypk2 mutant, Ypk2(D239A) (11), was constitutively hyperactive and able to rescue the phenotypes of cells defective in TORC2 function due to an alteration in AVO3, which encodes another essential subunit of the TORC2 complex (6). In further agreement with our model, the loss of viability caused by increasing concentrations of myriocin is much more severe in the presence of the TOR-directed inhibitor NVP-BEZ235 and, most dramatically, that this toxicity is substantially reduced by expression of Ypk1(D242A) (Fig. 4D).

Finally, we found that the TORC2-dependent T662 site in Ypk1 is indeed physiologically important for acute Orm phosphorylation in response to sphingolipid depletion. Compared with cells expressing wild-type Ypk1 (Fig. 5A, Left), cells expressing Ypk1(T662A) exhibited markedly delayed and less-efficient Orm phosphorylation in vivo after addition of myriocin (Fig. 5A, Right). A similar, but less dramatic, reduction in the efficiency of myriocin-induced Orm phosphorylation was caused by the TOR inhibitor NVP-BZ235 (SI Appendix, Fig. S4A). Correspondingly, compared with wild-type Ypk1 immunoprecipitated from myriocin-treated cells, Ypk1(T662A) obtained in the same fashion was substantially reduced in its ability to phosphorylate GST-Orm1(1–85) in vitro (Fig. 5B, Left), even though the amount of the mutant enzyme recovered was detectably higher than the wild type, as judged by Coomassie Blue dye staining (Fig. 5B, Right). We presume that the residual ability of Ypk1(T662A) to phosphorylate Orm proteins both in vivo and in vitro is due to the fact that Tor2 also phosphorylates Ypk1 at S644 in its turn motif (and perhaps other sites). Indeed, as judged by in vivo complementation tests, mutation of the equivalent residue (S641) in Ypk2 abrogates its function (11). Treatment with the TOR inhibitor NVP-BZ235, which presumably should prevent all TORC2-dependent modifications, also caused a reproducible reduction in the ability of immunoprecipitated Ypk1 to phosphorylate GST-Orm2(1–80) (SI Appendix, Fig. S4B), but this effect was more modest than a T662A mutation. It seems, therefore, that this compound, developed for use in mammalian cells (29), is not fully potent in yeast.

Fig. 5.

Tor2-dependent phosphorylation of Ypk1 is necessary for optimal Orm phosphorylation. (A) As in Fig. 3A, except that the ypk1∆ cells expressing (FLAG)₃-Orm1 (YDB344) cells also carried plasmids expressing either wild-type Ypk1 (WT) (pAM20) or a Ypk1(T662A) mutant (pFR221). (B) As in Fig. 3B, except the cells (BY4741) expressed from the GAL1 promoter were either Ypk1-myc (pAM54) or Ypk1(T662A)-myc (pFR119). (C) Sphingolipid deficiency stimulates Ypk1-mediated phosphorylation of Orm proteins via TORC2. Apparent sites of action of phosphoprotein phosphatases are also indicated.

Taken in their entirety, our results show that Ypk1 is responsible for the phosphorylation of Orm proteins that occurs when sphingolipid synthesis is compromised, and strongly support the conclusion that the increase in Ypk1 activity that occurs upon sphingolipid depletion is due, at least in part, to an increase in Tor2/TORC2-dependent phosphorylation of Ypk1. Of course, the phosphorylation state of both the Orm proteins and Ypk1 is also influenced by the rate at which they are dephosphorylated (Fig. 5C). By screening the yeast deletion collection (31), we found (SI Appendix, Figs. S1 A and B and S5) that absence of two distinct classes of phosphoprotein phosphatases affect these processes (SI Appendix and Fig. 5C). Further investigation of how inhibition of sphingolipid synthesis promotes enhanced TORC2 action on Ypk1 should reveal additional features of the regulatory mechanisms that enable cells to achieve sphingolipid homeostasis.

Our findings also have implications for understanding the action of other cAMP-dependent, cGMP-dependent and protein kinase C (AGC) family kinases. In addition to mandatory phosphorylation of the T-loop in their kinase fold, many enzymes in this family are also regulated by phosphorylation at two conserved sites situated C terminal to the catalytic domain: the turn and hydrophobic motifs. The upstream kinases that phosphorylate these C-terminal motifs are different for different targets. For example, for yeast Sch9, like mammalian p70^S6K, TORC1 is responsible, whereas for Ypk1 (and Ypk2), like mammalian AKT and SGK, TORC2 is responsible (27, 32). It is noteworthy that mutation of the hydrophobic motif in Ypk1 does not diminish its “intrinsic” catalytic activity toward a peptide substrate (16); yet, as we showed here, Ypk1(T662A) is clearly impaired in its activity toward Orm1 and Orm2 both in vitro and in vivo, consistent with the observation that, unlike cells expressing wild-type Ypk1, this mutant is unable to grow when challenged with myriocin (16, 17). A similar finding has been made with mammalian AKT. In mice lacking an essential TORC2 component (Sin1) to block phosphorylation of AKT by TORC2, AKT is not inactivated, but rather impaired in its ability to phosphorylate certain of its substrates (33). This mode of regulation appears to be unique to those AGC kinases under TORC2 control because, by contrast, mammalian p70^S6K and yeast Sch9 absolutely require modification by TORC1 for their activity (27, 32). Likewise, activity and function of yeast Cbk1 (an AGC kinase of the Ndr/LATS subfamily) is completely dependent on phosphorylation of its hydrophobophobic motif by a distinct complex containing the kinase Kic1 (34, 35).

Materials and Methods

Strains and Growth Conditions.

Yeast strains used in this study are listed in SI Appendix, Table S1 and were grown routinely at 30 °C. Standard rich (YP) and defined minimal (SC) media (36) (containing 2% glucose; 2% raffinose and 0.2% sucrose; or 2% galactose as the carbon source and supplemented with appropriate nutrients to maintain selection for plasmids) were used for yeast cultivation. Conditions for gene induction by galactose and for treatment with drugs are described in detail in SI Appendix. Standard yeast genetic techniques were performed according to Sherman et al. (36).

Plasmids and Recombinant DNA Methods.

Plasmids used in this study are listed in SI Appendix, Table S2. Plasmids were constructed using standard procedures (37) in E. coli strain DH5α. Fidelity of all constructs was verified by nucleotide sequence analysis. All PCR reactions were performed using Phusion DNA polymerase (Finnzymes, Inc.). Site-directed mutagenesis using appropriate mismatch oligonucleotide primers was conducted using the QuikChange method and Pfu Turbo DNA polymerase (Stratagene, a division of Agilent Technologies, Inc.).

Cell Extracts and Immunoblotting.

To examine the in vivo phosphorylation state of (FLAG)₃-Orm1 or -Orm2, cells lysates were prepared and immunoblotting was performed as before (4). For analysis of the in vivo phosphorylation state of Ypk1^11A-myc, protein was extracted, and Phos-tag SDS/PAGE (Wako Chemicals USA, Inc.) was conducted as described in detail in SI Appendix.

Protein Kinase Assays.

Ypk1-myc (expressed in yeast from pAM54) or Ypk1(K376A)-myc (expressed in yeast from pAM49) were recovered from cell lysates by immunoprecipitation and their activity analyzed in the resulting immune complexes, using GST-Orm1(1–85) or GST-Orm2(1–80) (0.5 μg), prepared as described below, as the substrates, using procedures described previously (12). In some experiments, Ypk1 (0.4 μg) or an analog-sensitive variant, Ypk1(L424A) (1 μg), prepared by expression in and purification from S. cerevisiae (provided by Yidi Sun, Drubin Laboratory, University of California, Berkeley, CA), were resuspended in 20 μL of kinase assay buffer containing 100 μM [γ-³²P]ATP (∼5 × 10⁵ cpm/nmol) and incubated for 30 min with GST-Orm1(1–85) (0.5 μg) in the presence or absence of 3-MOB-PP1 (10 μM; provided by Chao Zhang, Shokat Laboratory, University of California, San Francisco). Reactions were terminated by addition of SDS/PAGE sample buffer containing 6% SDS followed by boiling for 5 min. Labeled proteins resolved by SDS/PAGE were analyzed by autoradiography using a PhosphorImager (Molecular Dynamics, a division of Amersham Pharmacia Biotech, Inc.). The same assay was repeated with 100 μM nonradiolabeled ATP, and the proteins were resolved by SDS/PAGE and analyzed by immunoblotting with rabbit polyclonal anti–phospho-Akt substrate (RXXXphosphoS/T; Cell Signaling Technology, Inc.; 1:1,000) diluted in 5% wt/vol BSA, 1× TBS, and 0.1% Tween-20, as recommended by the manufacturer, or anti-GST antibodies diluted in Odyssey blocking buffer that had been mixed 1:1 with PBS.

Purification of GST Fusion Proteins.

Freshly transformed BL21(DE3) cells carrying a plasmid expressing GST-Orm1(1–85) (pFR203) or GST-Orm2(1–80) (pFR215) were grown to A_{600 nm} = 0.6, and expression was induced by addition of isopropyl-β-d-thiogalactopyranoside (0.6 mM final concentration). After vigorous aeration for 8 h at 30 °C, cells were harvested and the GST-fusion protein was purified by column chromatography on glutathione-agarose using standard procedures.