Heterologous mammalian Akt disrupts plasma membrane homeostasis by taking over TORC2 signaling in Saccharomyces cerevisiae

The Akt protein kinase is the main transducer of phosphatidylinositol-3,4,5-trisphosphate (PtdIns3,4,5P3) signaling in higher eukaryotes, controlling cell growth, motility, proliferation and survival. By co-expression of mammalian class I phosphatidylinositol 3-kinase (PI3K) and Akt in the Saccharomyces cerevisiae heterologous model, we previously described an inhibitory effect on yeast growth that relied on Akt kinase activity. Here we report that PI3K-Akt expression in yeast triggers the formation of large plasma membrane (PM) invaginations that were marked by actin patches, enriched in PtdIns4,5P2 and associated to abnormal intracellular cell wall deposits. These effects of Akt were mimicked by overproduction of the PtdIns4,5P2 effector Slm1, an adaptor of the Ypk1 and Ypk2 kinases in the TORC2 pathway. Although Slm1 was phosphorylated in vivo by Akt, TORC2-dependent Ypk1 activation did not occur. However, PI3K-activated Akt suppressed the lethality derived from inactivation of either TORC2 or Ypk protein kinases. Thus, heterologous co-expression of PI3K and Akt in yeast short-circuits PtdIns4,5P2- and TORC2-signaling at the level of the Slm-Ypk complex, overriding some of its functions. Our results underscore the importance of phosphoinositide-dependent kinases as key actors in the homeostasis and dynamics of the PM.


PI3K-activated Akt induces actin-supported membranous structures in yeast.
We previously reported that reconstitution of the PI3K-Akt pathway in yeast led to MAPK pathways activation and impaired growth. GFP-Akt1 was recruited to the yeast PM by PtdIns3,4,5P 3 generated in situ by co-expressed PI3K p110α catalytic subunit, a phenomenon dependent on Akt PH domain 37,38 . As shown in Fig. 1a, besides PM localization, cells expressing GFP-Akt1 from the galactose-inducible GAL1 promoter displayed large Akt1-enriched intracellular compartments. This was not observed in cells expressing kinase-dead GFP-Akt1 K179M , which just concentrated at small dots. We had previously demonstrated that the yeast PDK1 orthologues, Pkh1 and Pkh2, are responsible for phosphorylation and activation of heterologous Akt1 in its T-loop (Thr308) 37 . Immunofluorescence with specific anti-phospho-Akt1(Thr308) antibodies showed that it was the active form of GFP-Akt1 which accumulated at these large compartments (Fig. 1b, upper panel). Moreover, a non-activatable GFP-Akt1 T308A mutant version failed to produce these structures although it was efficiently recruited to the PM by PI3K-generated PtdIns3,4,5P 3 (Fig. 1b, lower panel). Thus, concomitant PI3K-and Pkh-dependent Akt1 activation is required for the development of intracellular Akt-enriched structures.
Actin staining with fluorochrome-conjugated phalloidine of yeast cells co-expressing mammalian PI3K and Akt1 revealed an abnormal actin distribution: restriction of actin patches to the growing bud was lost in cells expressing wild type -but not kinase-dead-Akt1 (Fig. 1c). Moreover, actin patches co-localized with GFP-Akt1-enriched structures (Fig. 1d). The proportion of cells containing such structures was significantly reduced by treatment with the actin-depolymerizing drug latrunculin A (Fig. 1e). Thus, the generation of Akt1-induced structures relies on actin dynamics.

Akt1-induced structures are PM invaginations.
To characterize the nature of Akt1-marked compartments, we used the lipophilic vital dye FM4-64, which binds PM lipids and is internalized by endocytosis, thus labeling compartments along the endocytic pathway until it reaches the vacuolar membrane 39 . As shown in Fig. 2a, in cells co-expressing p110α and GFP-Akt1, GFP fluorescence did not co-localize with FM4-64 at the vacuole after 30 min of incubation. Instead, GFP-Akt1-marked structures matched with intense smaller SCIEnTIFIC REPORTs | (2018) 8:7732 | DOI: 10.1038/s41598-018-25717-w FM4-64-labeled compartments. Blocking endocytosis by incubating cells on ice in the presence of azide as a metabolic inhibitor leads to persistent PM staining by FM4-64 39 . Interestingly, in these conditions, GFP and FM4-64 co-stained both the PM and active Akt-induced compartments (Fig. 2b), suggesting that such structures corresponded to intracellular PM extensions. In support of this idea, a specific probe for phosphatidylserine (PS) at the PM inner layer, GFP-C2(Lact) 40 , co-localized with mCherry-Akt3 in the presence of p110α, but not when co-expressed with kinase-dead p110α (Fig. 2c). These results indicate that the compartments in which activated Akt accumulated were large PM invaginations.

Akt-induced PM invaginations promote intracellular cell wall deposition.
Next, we performed transmission electron microscopy (TEM) on yeast cells co-expressing p110α and Akt1. As shown in Fig. 3, these cells displayed multiple membrane invaginations conforming cytoplasmic islands associated with the cell surface. These structures were frequently paired or concentrated at particular areas of the cell surface (Fig. 3a,b). The observation of invaginating membranes that do not form yet a closed surface-associated compartment (Fig. 3c) suggested that extended folds of membrane expand inwards until they find again the PM and fuse, thus originating cytoplasmic islands. Remarkably, electron-translucent cell wall (CW) polysaccharides were deposited filling some PM invaginations. None of these phenomena were observed in control cells co-expressing p110α and kinase-dead Akt1 (Fig. 3a).
We stained p110α and GFP-Akt1 co-expressing cells with calcofluor white (CFW) to visualize chitin-rich CW regions, to confirm that CW material was accumulated at invaginations. In agreement with TEM observations, some but not all GFP-Akt1 PM extensions were stained with CFW (Fig. 3d). CW growth is supported by actin-driven polarized secretion involving local activation of the Rho1 small GTPase 41 . The localization of the Rho1-activated protein kinase C Pkc1 serves as a spatial marker for sites of Rho1 activation 42,43 . Interestingly, abnormally located GFP-Pkc1 co-localized with some mCherry-Akt3 invaginations (Fig. 3e). important roles in endocytosis 44 . PM invaginations similar to those induced by Akt1 have been previously reported in cells lacking synaptojanin-like (Sjl) PtdIns 5-phosphatases, as a consequence of the up-regulation of this phosphoinositide 45, 46 . Consistently, we found intense staining of Akt-induced invaginations by the PtdIns4,5P 2 -specific fluorescent probe GFP-PH(PLCδ) 46 , which co-localized with mCherry-Akt3 (Fig. 4a). The presence of PtdIns4,5P 2 in the invaginations seems paradoxical because Akt localization and activation requires the conversion of PtdIns4,5P 2 into PtdIns3,4,5P 3 by co-expressed PI3K. Nevertheless, overexpression of either the Sjl2/Inp52 or Sjl3/Inp53 PtdIns 5-phosphatases partially counteracted the formation of Akt-induced PM invaginations (Fig. 4b), supporting that Akt1 activation by PtdIns3,4,5P 3 locally increases the levels of its precursor PtdIns4,5P 2 .  We also overproduced TORC2-signaling targets, like ceramide synthases Lac1 and Lag1 27 , and MCC/eisosome components Lsp1, Pil1 and Sur7 12 , because MCC/eisosomes are also related to PtdIns4,5P 2 regulation 11,48 . As shown in Table S1 and Fig. 5a, overexpression of Pkc1 and Slm1 was highly toxic for yeast cells. Among the other proteins overexpressed, only Lac1 and Lag1 faintly inhibited yeast growth. When co-expressed with p110α-Akt, only Pkh2 and Slm1 showed a negative genetic interaction, as their overexpression slightly enhanced Akt-induced growth inhibition (Fig. 5a). Remarkably, Slm1 overproduction led to large PM invaginations, reminiscent of those triggered by p110α-Akt (Fig. 5b), in around 60% of cells (Fig. 5c). Although overproduction of Pkh2, Ypk1 and TORC2 components also led to the appearance of PM invaginations (Table S1), they were smaller and less conspicuous than those induced by Slm1-or Akt-overexpressing cells (Fig. 5b). Eisosome disorganization by absence of Pil1 has also been reported to promote PM invaginations 12 . In our hands, around 60% of pil1Δ cells displayed at least one invagination (Fig. 5c), but not those lacking other eisosome components (lsp1Δ or sur7Δ) (Table S1). Interestingly, this phenotype was significantly enhanced when the SLM1 gene was overexpressed in a pil1Δ background (Fig. 5c). PM invaginations induced by SLM1 overexpression, like those triggered by Akt, were dependent on Pkh1/2 function, as their induction was significantly diminished in a pkh1-ts pkh2Δ strain at the restrictive temperature (Fig. 5d). Also, like Akt, overproduced Slm1 accumulated at these structures, as revealed by immunofluorescence on cells expressing poly-His-tagged Slm1 (Fig. 5e). Furthermore, Slm1-induced structures also contained PS and PtdIns4,5P 2 , as determined by using fluorescent GFP-C2(Lact)-and GFP-PH(PLCδ) probes, respectively (Fig. 5f). As expected, TEM analysis of SLM1-overexpressing cells revealed PM extensions into the cytoplasm, occasionally filled with CW material (Fig. S1a). We also found that the number and size of cytoplasmic CW inclusions upon SLM1 overexpression was enhanced in a pil1Δ background (Fig. S1b). Therefore, loss of eisosome function, overproduction of Pkh kinases, TORC2 components or, especially, Slm1 caused PM disturbances reminiscent of local Akt activation, indicating that this mammalian protein could be taking over the yeast TORC2 pathway.

GFP-Akt1
Akt activation leads to oxidative stress in yeast cells. TORC2 signaling has been reported to influence both PM homeostasis and cytoskeletal regulation via reactive oxygen species (ROS) 49 . We analyzed ROS production by dihydroethidium (DHE) staining and flow cytometry in cells co-expressing p110α with either Akt1 or Akt1 K179M . Overexpression of the YBH3, a gene coding for S. cerevisiae BCL-2 Homology domain 3 (BH3)-containing protein 50 was used as a positive control for ROS generation. As shown in Fig. 5g, Akt1 co-expressed with p110α led to a higher percentage of ROS-producing cells than the kinase-dead Akt1 version, and did so more efficiently than overproduction of Ybh3.
We also performed a transcriptomic analysis to study Akt-induced transcriptional response. Global mRNA expression of cells expressing either p110α with Akt1 or kinase-dead Akt1 K179M were compared. Table S2 show  Term Finder analyses on up-regulated genes highlighted three significantly enriched functional categories: "pentose phosphate shunt" (p-value = 9×10 −15 by chi square test with Bonferroni correction), "oxidation-reduction" (p = 2×10 −4 ) and "cell wall organization" (p = 0.05). In down-regulated genes, two functional categories were underscored: "mitotic cell cycle" (p = 2×10 −4 ) and again "cell wall organization" (p = 0.01) (Fig. S2a). We then compared our dataset of Akt-induced genes to reported datasets related to oxidative stress (treatment with diamide), cell wall stress (treatment with Congo red) and inhibition of Tor2. Hits from our microarray significantly overlapped with these datasets (56.5%, 43.2% and 22.9%, respectively; Fig. S2b), in agreement with the previously reported activation of the yeast cell wall integrity (CWI) pathway by Akt 38 , as well as the above-described induction of ROS production and alteration of TORC2 signaling. The "pentose phosphate shunt"-related hits were Akt-specific, as it did not overlap with other datasets.   the phosphorylation state of Slm1 in the presence of Akt. As shown in Fig. 6a, overexpressed polyHis-tagged Slm1 migrated as a single band. However, when Slm1 was co-overproduced with p110α-Akt1, but not with p110α alone, an additional band of reduced electrophoretic mobility appeared. This mobility shift corresponded to phosphorylation, as it could be eliminated by phosphatase treatment in vitro (Fig. 6b). However, a mutant Slm1 S659A version, known to lack Pkh1/2-dependent phosphorylation upon stress 52 , displayed the same pattern as wild type Slm1 (Fig. 6c). Thus, this residue was not involved in the post-translational modification detected upon PI3K-Akt expression. To determine whether this phosphorylation occurred on endogenous Slm1, we integrated a SLM1-6×myc version by gene replacement in a slm2Δ mutant so that all cellular Slm function was dependent on tagged Slm1. The electrophoretic shift of Slm1-6×Myc was still observed (Fig. 6d). Moreover, immunoprecipitation with anti-myc antibodies and immunodetection with anti-phospho-Akt substrate [(R/K)X(R/K)XX(pT/pS)] antibodies showed that Slm1 was recognized by these antibodies only in the presence of Akt1 (Fig. 6e).
The above results led us to hypothesize that the effect of Akt was due to hyperactivation of the Slm-dependent TORC2-Ypk pathway. In that case, the lack of Slm proteins would attenuate Akt-mediated toxicity in yeast. To test this, we used a triple slm1∆ slm2∆ sac7∆ mutant strain, because the absence of the Rho1 GTPase-activating protein Sac7 counteracts the lethality of a slm1∆ slm2∆ double mutant 23 . However, although growth of this strain in galactose was deficient, Akt1 still had an inhibitory effect as compared to kinase-dead Akt1 (Fig. 6f), suggesting that Slm function was not determinant for Akt effects in yeast.
Ypk kinases are activated when the TORC2 pathway is stimulated by defects in sphingosine biosynthesis, so treatment with myriocin, a serine palmitoyltransferase inhibitor, is commonly used to activate this pathway 25,53 . To test whether Akt activation led to Ypk1 phosphorylation by TORC2, we analyzed cell lysates expressing Ypk1-HA by immunoblotting with anti-phospho-Ypk1(Thr662) 47 . As a control we used cells expressing the non-phosphorylatable version Ypk1 S644A,T662A -HA 22 . As shown in Fig. 6g, whereas myriocin led to Ypk1-HA Thr662 phosphorylation, co-expression of p110α and Akt1 in the absence of this compound did not. Thus, in spite of inducing Slm1 phosphorylation, Akt is not promoting Ypk1 phosphorylation.
We also tested whether Akt was able to phosphorylate known Ypk1 substrates, such as Orm1/2 proteins. In response to low sphingolipid levels, these negative regulators of sphingolipid biosynthesis are inactivated by TORC2-activated Ypk1 26 . Treatment with myriocin triggers a electrophoretic mobility shift of 3×FLAG-Orm1, while co-expression of p110α and Akt1 did not (Fig. 6h). This result was consistent with the lack of Ypk1 Thr662 phosphorylation observed in Akt-expressing cells. We also checked whether the production of Akt-induced PM invaginations was dependent on Ypk activity. As shown in Fig. 6i, p110α-Akt1 co-expression still led to the appearance of invaginations upon reduction of Ypk activity in a ypk1-ts ypk2Δ mutant at semi-permissive temperature. The fact that Akt-induced phenotypes mimic TORC2 signaling without involving Slm-Ypk activation suggested that Akt is overriding the pathway downstream TORC2 by taking over, at least partially, Ypk function.

Akt complements loss of Ypk and TORC2 function.
To test whether Akt is mimicking the essential function of Slm1-Ypk1 downstream TORC2, we performed growth assays with a ypk1-ts ypk2Δ mutant expressing p110α with either wild type or kinase-dead Akt1. At the permissive temperature (24 °C), this mutant grew like the wild type and Akt activation negatively affected growth, as expected. However, at the restrictive temperature in galactose-based medium (34 °C), the growth defect of the Ypk-deficient strain was fully relieved by Akt (Fig. 7a). This indicates that Akt can complement a dysfunction of the Slm-Ypk module in the TORC2 pathway. To further prove this point, we tested whether Akt could complement loss of TORC2 function. We used a strain in which TORC2 can be specifically inhibited by rapamycin 54 . Co-expression p110α and Akt, but not the kinase-dead version, indeed rescued the lethality observed upon rapamycin-dependent TORC2 inhibition in galactose-based medium and even partially in glucose-based medium, in which Akt expression should be very low (Fig. 7b).

Discussion
Eukaryotic cells sense environmental stimuli through PM-located receptors that trigger complex molecular events to reprogram gene expression and protein synthesis. Phosphoinositides are key signal transducing landmarks at cellular membranes. In spite of recent advances, we are only beginning to understand how AGC superfamily protein kinase-mediated pathways respond to lipid second messengers to regulate PM homeostasis upon environmental challenges. In higher cells, one of the most relevant signaling pathways at the PM involves PI3K, which converts PtdIns4,5P 2 into PtdIns3,4,5P 2 that locally recruits Akt to phosphorylate multiple substrates. The yeast S. cerevisiae is an outstanding model for studies on signaling. Heterologous expression of mammalian PI3K p110α catalytic subunit efficiently produces PtdIns3,4,5P 3 , otherwise absent in yeast, leading to recruitment of Akt to the PM where it is fully activated by endogenous conserved PDK1-like (Pkh1/Pkh2) and presumably TORC2 kinases 37,38,47 . We show here that the most outstanding effect of Akt activity in yeast is the development of large and ubiquitous PM invaginations. A similar phenomenon was reported to occur in cells lacking the PtdIns4,5P 2 5-phosphatases Sjl1/Inp51 and Sjl2/Inp52 45, 46 . During endocytosis, an actin-driven process, elimination of PtdIns4,5P 2 by these phosphatases from the forming endocytic vesicle is essential for its scission from the PM 44,55 . The reduction of Akt-induced invaginations by treatment with the actin-depolymerizing drug latrunculin and quantitative analyses of PM invaginations in wild type (YPH499) or YPT40 (ypk1-1ts ypk2∆) at permissive (24 °C) or restrictive (34 °C) temperatures, co-transformed with YCpLG-p110α and pYES2-GFP-Akt1, after 24 h incubation in SG medium. Data are the average from 3 different fields (n > 30 cells/field). Error bars correspond to standard deviation. Images in (a-e,g,h) correspond to cropped blots for conciseness. Full-length blots are presented in Supplementary Fig. S3.
SCIEnTIFIC REPORTs | (2018) 8:7732 | DOI:10.1038/s41598-018-25717-w suggests that they derive of growing endocytic-like membranes unable to excise from a PtdIns4,5P 2 -rich PM. It seemed paradoxical that a behavior related to high PtdIns4,5P 2 levels was triggered by Akt, as its activation is subordinated to local removal of PtdIns4,5P 2 via its conversion to PtdIns3,4,5P 3 by co-expressed PI3K. However, we found that Akt-induced invaginated membranes, similar to those in sjl1Δ sjl2Δ mutants, were heavily marked by the PtdIns4,5P 2 PH(PLCδ) reporter, and their number decreased when PtdIns4,5P 2 levels were reduced by SJL2/SJL3 overexpression. Thus, local Akt kinase activity must be triggering endogenous pathways that upregulate PtdIns4,5P 2 synthesis, either by activating the Mss4 PtdIns4P 5-kinase or by inhibiting the PtdIns4,5P 2 5-phosphatases. Some PM invaginations induced by Akt were accompanied by abnormal intracellular CW deposition. Both recruitment of Pkc1 to invaginated membranes and transcriptomic data showing a significant differential expression of genes related to CW biogenesis are consistent with local hyperactivation of the Rho1-Pkc1 pathway at these sites to facilitate actin-based polarized secretion and CW deposition. Actually, we had previously reported Akt-dependent activation of the Pkc1-dependent CWI MAPK pathway 38 . Although our static EM analyses cannot establish the order of events at the PM following Akt activation, our data are consistent with the idea that large PM invaginations form prior to CW material deposition on the extracellular side, ultimately leading to their expansion. PM and CW growth into the cell is reminiscent of the sequence of events required for septation, when the PM is invaginated at the bud neck by mechanisms divergent to those governing endocytosis, likely maintaining high levels of PtdIns4,5P 2 in order to prevent membrane scission 56 .
Remarkably, analogous PM invaginations were observed by growing yeast cells in the presence of 1-O-hexadecyl-2-acetyl-sn-glycerophosphocholine, also known as C16:0 Platelet Activating Factor (PAF), a neurotoxic lipid that accumulates in neurons in Alzheimer disease 57 . In contrast to those induced by Akt, such structures were formed independently of the actin cytoskeleton and were not reported to be associated with internal cell wall deposition 57 . However, in spite of these differences, PAF-induced structures were also enriched in PtdIns4,5P 2 suggesting that both phenomena are caused by remodeling of PM PtdIns4,5P 2 distribution. Kennedy et al. concluded that the accumulation of sphingolipids in PAF-treated cells contributes to the re-localization of PtdIns4P 5-kinase Mss4, leading to an increase in PtdIns4,5P 2 levels 57 . Thus, Akt activation in yeast could be related to modulation of sphingolipid metabolism as well. Consistently, we observed that treatment with myriocin, an inhibitor of the sphingolipid synthesis pathway, did cause a 30-fold decrease in the formation of Akt-induced PM invaginations (data not shown). How alteration of complex sphingolipids might lead to a concomitant increase in PtdIns4,5P 2 is unclear, but mounting evidence suggests links between sphingolipid regulation and the TORC2-binding proteins Slm1 and Slm2, which are PtdIns4,5P 2 effectors 47,51,52 . Consistently, here we show that overexpression of TORC2 complex components, especially Slm1, led to PtdIns4,5P 2 -enriched invaginations reminiscent of those induced by PAF, Sjl phosphatases elimination or activated Akt.
Eisosome complexes are integrated by the BAR domain-containing proteins Pil1 and Lsp1 that assemble at PM MCC microdomains 13 . These compartments are regulated by Pkh kinases 31,58 , contain Slm proteins 30,53 , and play a role in PtdIns4,5P 2 -signaling towards actin-mediated endocytosis 24,52 . Eisosomes participate in the regulation of sphingolipid metabolism 59 and endosomal scission 60 , as well as in the maintenance of PtdIns4,5P 2 homeostasis by recruitment of the Sjl1 phosphatase 48 . Accordingly, defects in eisosome organization achieved by elimination of Pil1 are associated with an increase in PtdIns4,5P 2 48 . Therefore, PM invaginations observed here and reported by other authors in pil1∆ mutant cells 12,61 should also be related to high levels of this phosphoinositide. Consistently, we found that the pil1∆ phenotype and the one caused by Slm1 overexpression were additive. Slm proteins have been proposed to shift from MCC/eisosomes to MCT in response to PM stress, thus activating TORC2-Ypk1 signaling to readapt membrane composition to face such situation 26,28,47,53 . Moreover, Slm1 overexpression has been shown to trigger TORC2-Ypk1 activation even in the absence of stress 53 . In view of these data, it is plausible that Slm1 overexpression phenotypically resembles Akt activation by enhancing its presence at the MCT and promoting a local increase in PtdIns4,5P 2 , as happens when MCC/eisosomes are impaired. The fact that Slm1 becomes hyperphosphorylated upon Akt activation suggests their co-existence in particular PM domains. The recognition of Slm1 by antibodies specific to Akt-phosphorylated substrates supports the idea that it is a  Figure 7. Akt complements the lack of TORC2-Ypk1 function in yeast. (a) Growth assays of ten-fold serial dilutions of YPH499 (wild type; WT) and YPT40 (ypk1-1ts ypk2∆) cells transformed with YCpLG-p110α and pYES2-GFP-Akt1 K179M or pYES2-GFP-Akt1 cultured in SD (Glucose) and SG (Galactose), incubated at 24 °C or 34 °C. (b) Growth assays of ten-fold serial dilutions TB50 strain (WT) and mPR8 mutant cells (TOR1-1 avo3 ∆CT , otherwise isogenic to TB50) transformed with YCpLG-p110α and pYES2-GFP-Akt1 K179M or pYES2-GFP-Akt1 cultured in SD (Glucose) and SG (Galactose) in the presence or absence of 200 nM rapamycin, as indicated.
SCIEnTIFIC REPORTs | (2018) 8:7732 | DOI:10.1038/s41598-018-25717-w direct target of heterologous Akt. However, Slm1 sequence lacks a typical RxRxxS/T motif and these antibodies may also recognize phosphorylated minimum RxxS/T motifs that do not necessarily fit the Akt consensus, so we cannot discard that phosphorylation of Slm1 may be indirect through another kinase as a consequence of Akt interference with the TORC2-Ypk pathway.
The role of Slm1 in the TORC2 pathway is the recruitment of Ypk kinases in proximity of its activators: the TORC2 complex and yeast PDK1 orthologs Pkh1/2 47,53 (Fig. 8a). Although yeast Ypk kinases have been proposed to be presumptive ortologues to mammalian SGK and/or Akt 45, 62 , it has been demonstrated that Ypk kinases can be replaced by SGK but not Akt 20 . However, we show here that, when activated by PI3K, Akt is indeed capable of covering the essential functions of Ypk1/2 and TORC2 kinases in yeast. In these conditions, the presence of  Figure 8. A model for the interference of Akt with PtdIns4,5P 2 and TORC2 signaling. In physiological conditions (a), regulatory complexes at the MCC/eisosomes and MCT/TORC2 microdomains regulate PM PtdIns4,5P 2 and sphingolipid levels, respectively, in a coordinated way. PM Stress or lipid misbalance cause Slm1 to shift between compartments and bring Ypk1 in proximity to its activating kinases, turning on the TORC2 pathway. When mammalian p110α and Akt1 are co-expressed in yeast (b), PM pools of PtdIns4,5P 2 are converted into PtdIns3,4,5P 3 by PI3K activity, bringing Akt in proximity with PDK-like Pkh kinases and TORC2. Thus Akt takes over the role of Ypk short-circuiting TORC2 signaling and leading to enhanced PtdIns4,5P 2 and phosphoinositide-dependent signaling for membrane growth inwards and actin-supported cell wall deposition. Artificial overproduction of Slm1 (c) leads to similar effects, probably by enhancing its presence at the MCT and biasing Ypk signaling towards a PtdIns4,5P 2 -dependent response uncoupled from physiological TORC2 modulation.
SCIEnTIFIC REPORTs | (2018) 8:7732 | DOI:10.1038/s41598-018-25717-w a PH domain in Akt, which is absent in Ypks, avoids a requirement for Slm1 to interact with the yeast PM. By this means, PI3K-activated Akt takes over the function of the Slm-Ypk complex on selective targets related to PtdIns4,5P 2 -dependent signaling (Fig. 8b), uncoupling PtdIns4,5P 2 from sphingolipid-dependent signaling, an effect that can be mimicked by Slm1 overexpression (Fig. 8c). This situation seems to interfere with physiological TORC2-Ypk function on sensing PM stress, as Akt did neither stimulate TORC-dependent phosphorylation of Ypk1 nor activated known TORC2-Ypk1 substrates, such as Orm1 26 . Moreover, the yeast transcriptomic profile of Akt-expressing cells partially overlapped that of a tor2-Ts mutant shifted to restrictive temperature. In addition, it has been reported that TORC2-Ypk1 must be activated in order to suppress ROS accumulation 49 , which could explain the significant increase in cellular ROS levels observed upon Akt activation in yeast. Also, loss of TORC function has been related to inefficient endocytic scission, mislocalization of the phosphatase Sjl2 and a consequent increase in PM PtdIns4,5P 2 levels 63 , phenotypes reminiscent of Akt activation. Finally, the aforementioned PtdIns4,5P 2 -enriched invaginations induced by PAF have also been associated with inhibition of TORC2 57 . Therefore, although Akt is able to complement TORC2-Ypk loss of function, it seems to negatively interfere with the physiological activation of this pathway at the PM.
In sum, we report here that Akt activation in yeast short-circuits endogenous signaling pathways by taking over the function of Ypk kinases in the TORC2 pathway, underscoring the importance of phosphoinositides and downstream kinases in the control of PM homeostasis in all eukaryotes. Therefore, heterologous expression of Akt provides a means to studying its activity in a simple model while it offers a tool for studying phosphoinositideand TORC2-dependent signaling in yeast.  65 ; and YPT40 (YPH499 ypk1-1ts::HIS3 ypk2-∆1::TRP1) 20 . Strains SEY6210 and AAY1663 are described by Robinson et al. 66 and Audhya et al. 23 , respectively. Strains TB50 (MATa leu2 ura3 rme1 his3Δ) and isogenic mPR8 [MATα tor1-1 avo3Δ1274-1430::hphMX6] was described in Gaubitz et al. 54 . Strain YDB146 (BY4741, 3XFLAG-ORM1) is described by Breslow et al. 67 .
pESC-TRP-GFP-2xPH(PLCδ) was constructed by PCR amplification of a GFP N-terminal fusion containing two tandem copies of the PH domain of PLCδ, using plasmid pRS426-GFP-2 × PH(PLCδ) as a template 46  Flow cytometry assays. For the analysis of ROS, yeast transformants were grown in SR medium lacking uracil at 30 °C overnight. Then, the cultures were induced with galactose for 16 h and dihydroethidium (2.5 μg/mL) was added for 5 min. Three thousand cells per second were analysed on a FACScan flow cytometer (Becton Dickinson) on the FL2 log scale. WinMDI 2.7 software was used to analyze the graphics obtained.
Fluorescence microscopy and immunofluorescence. For in vivo fluorescence microscopy (GFP and mCherry observation), cells from exponentially growing SR cultures were induced with 2% galactose for 4 h harvested by centrifugation 10,000 rpm 1 min, washed once with phosphate saline buffer and viewed directly. Cells were examined with an Eclipse TE2000U microscope (Nikon) using the appropriate sets of filters. Digital images were acquired with Orca C4742-95-12ER charge-coupled device camera (Hamamatsu) and were processed with the HCImage software (Hamamatsu, Japan). For statistics on cell populations >100 cells were counted for each experiment. Observation of actin in yeast cells with rhodamine-conjugated phalloidin (Sigma, St. Louis, MO, USA) was performed as previously described 43  Secondary antibodies were anti-rabbit IgG Alexa Fluor ® 488 and anti-mouse IgG Alexa Fluor ® 568, respectively, both diluted to 1:500.
Electron microscopy. After galactose induction (6 h), yeast cells were fixed by adding to the culture an equal volume of 6% paraformaldehyde and 4% glutaraldehyde in 0.2 M potassium phosphate buffer (pH 6.5) and three times in water, and then treated with 1% KMnO 4 for 2 h on ice, followed by three times of rinses with water. The samples were subsequently dehydrated and then embedded in Spurr's low viscosity media (EM Science) as described by the manufacturer. Ultrathin sections were cut and examined under a Zeiss EM902 electron microscope.

Preparation of cell lysates, immunoprecipitation and Western blotting. Overnight cultures
of cells carrying GAL1-driven expression plasmids growing in SR media were refreshed to an OD 600 of 0.3 in SR-Galactose 2% in order to achieve GAL1-promoter induction. After 5-6 hours of incubation at 30 °C cells, yeast extracts were obtained as previously described 37 . For Ypk1-phosphorylation analysis, the same general procedure was followed but control cells were treated, after 3 h of GAL1-induction, with 1.5 µM myriocin (Cayman Chemical, Ann Arbor, MI, USA), for two additional hours as described by Niles et al. 47 . For the Orm1 electrophoretic shift assay, control cells were treated, after 4 h of GAL1-induction, with 0.4 µM myriocin for 90 additional minutes, as described in Roelants et al. 26 . In this particular case, yeast extracts were obtained following TCA precipitation. To this end, TCA was added to the cultures at a final concentration of 2% and they were kept on ice for 20 min. Afterwards they were centrifuged, washed with 10 mM sodium azide, resuspended in 500 µL of pre-chilled TCA buffer (10 mM Tris pH 8.0, 10% TCA, 25 mM NH 4 OAc, 1 mM Na 2 EDTA) and glass beads added.
Cells were broken in a FastPrep ® -2 4 at 5.5 rpm, during 30 sec for 3 times, chilling on ice in between. Samples were centrifuged and precipitated proteins were resuspended in 75 µL of resuspension buffer (100 mM Tris, 3% SDS, pH 11.0), and boiled at 95 °C for 5 minutes. Samples were centrifuged again to remove cellular debris and clarified lysates were collected to new tubes.
For endogenous Slm1-6xMyc immunoprecipitation experiments, 300 µL of yeast extracts were incubated with beads (Dynabeads ® Protein G, Thermofisher Scientific), which were previously treated overnight with the antibody anti-cMyc 9E10 (1:100, Santa Cruz). After washing three times, beads were resuspended in SDS-PAGE loading buffer and subjected to immunoblot. Western-blotting analyses were performed following the general procedure described previously 37 . Immunodetection was carried out with anti-phospho-Ypk1 (T662) (1: Secondary antibodies used for Western-blotting analyses were either horseradish peroxidase (HRP)-conjugated anti-mouse secondary antibodies (for blots in Fig. 6a,b), or anti-mouse IgG-Alexa FluorR 680, anti-rabbit IgG-IRDyeR 800 CW and anti-rabbit IgG-IRDyeR 680; all from LI-COR (Lincoln, NE, USA) at 1:5000 dilution (for the rest of the immunoblots). A chemiluminiscence detection system (ECL ™ ; Amersham Biosciences, UK) or an Oddissey infrared imaging system (LI-COR; Lincoln, NE, USA) system were used for developing the Western blots, respectively. Accession code. The accession code for the microarray data deposited in Gene Expression Omnibus Database is GSE107482.