Chemical Genetics of AGC-kinases Reveals Shared Targets of Ypk1, Protein Kinase A and Sch9
2020; Elsevier BV; Volume: 19; Issue: 4 Linguagem: Inglês
10.1074/mcp.ra120.001955
ISSN1535-9484
AutoresMichael Plank, M. P. Perepelkina, Markus Müller, Stefania Vaga, Xiaoming Zou, Clélia Bourgoint, Marina Berti, Jacques Saarbach, Steven Haesendonckx, Nicolas Winssinger, Ruedi Aebersold, Robbie Loewith,
Tópico(s)Fungal and yeast genetics research
ResumoProtein phosphorylation cascades play a central role in the regulation of cell growth and protein kinases PKA, Sch9 and Ypk1 take center stage in regulating this process in S. cerevisiae. To understand how these kinases co-ordinately regulate cellular functions we compared the phospho-proteome of exponentially growing cells without and with acute chemical inhibition of PKA, Sch9 and Ypk1. Sites hypo-phosphorylated upon PKA and Sch9 inhibition were preferentially located in RRxS/T-motifs suggesting that many are directly phosphorylated by these enzymes. Interestingly, when inhibiting Ypk1 we not only detected several hypo-phosphorylated sites in the previously reported RxRxxS/T-, but also in an RRxS/T-motif. Validation experiments revealed that neutral trehalase Nth1, a known PKA target, is additionally phosphorylated and activated downstream of Ypk1. Signaling through Ypk1 is therefore more closely related to PKA- and Sch9-signaling than previously appreciated and may perform functions previously only attributed to the latter kinases. Protein phosphorylation cascades play a central role in the regulation of cell growth and protein kinases PKA, Sch9 and Ypk1 take center stage in regulating this process in S. cerevisiae. To understand how these kinases co-ordinately regulate cellular functions we compared the phospho-proteome of exponentially growing cells without and with acute chemical inhibition of PKA, Sch9 and Ypk1. Sites hypo-phosphorylated upon PKA and Sch9 inhibition were preferentially located in RRxS/T-motifs suggesting that many are directly phosphorylated by these enzymes. Interestingly, when inhibiting Ypk1 we not only detected several hypo-phosphorylated sites in the previously reported RxRxxS/T-, but also in an RRxS/T-motif. Validation experiments revealed that neutral trehalase Nth1, a known PKA target, is additionally phosphorylated and activated downstream of Ypk1. Signaling through Ypk1 is therefore more closely related to PKA- and Sch9-signaling than previously appreciated and may perform functions previously only attributed to the latter kinases. Cell growth is dynamic and highly regulated by signaling pathways that are conserved across evolution. To accomplish this regulation, eukaryotes have developed intricate means to assess growth conditions and to rapidly communicate this information to the processes controlling the accumulation of mass, the modification of cellular volume and of membrane surface area. Although many signal transduction pathways are involved in this regulation, those employing AGC-family kinases (named after Protein Kinases A, G and C) are prominent. The target of rapamycin complexes 1 and 2 (TORC1 and TORC2) are central sensors of environmental conditions and regulators of cell growth (1Loewith R. Hall M.N. Target of rapamycin (TOR) in nutrient signaling and growth control.Genetics. 2011; 189: 1177-1201Crossref PubMed Scopus (603) Google Scholar). Both complexes exert their functions by phosphorylating AGC-kinases as their main targets. In S. cerevisiae TORC1 primarily responds to changes in carbon and nitrogen availability and regulates ribosome biogenesis, cell cycle progression and stress responses via the AGC-kinase Sch9, like S6K downstream of mammalian TORC1 (1Loewith R. Hall M.N. Target of rapamycin (TOR) in nutrient signaling and growth control.Genetics. 2011; 189: 1177-1201Crossref PubMed Scopus (603) Google Scholar, 2Urban J. Soulard A. Huber A. Lippman S. Mukhopadhyay D. Deloche O. Wanke V. Anrather D. Ammerer G. Riezman H. Broach J.R. Virgilio C.D. Hall M.N. Loewith R. Sch9 Is a Major Target of TORC1 in Saccharomyces cerevisiae.Mol. Cell. 2007; 26: 663-674Abstract Full Text Full Text PDF PubMed Scopus (616) Google Scholar). Another AGC-kinase, protein kinase A (PKA) 1The abbreviations used are:PKAProtein kinase AAGC-kinasesProtein kinases of the family comprising PKA, PKG and PKCTORC1/TORC2TOR complex 1/21NM-PP1C3–1′-naphthyl-methyl PP1FDRfalse-discovery rateGOgene ontologyPSMpeptide spectrum match. 1The abbreviations used are:PKAProtein kinase AAGC-kinasesProtein kinases of the family comprising PKA, PKG and PKCTORC1/TORC2TOR complex 1/21NM-PP1C3–1′-naphthyl-methyl PP1FDRfalse-discovery rateGOgene ontologyPSMpeptide spectrum match., performs many, if not most of its functions in parallel to Sch9 by regulating an overlapping set of functions and potentially by cross-talk (3Broach J.R. Nutritional control of growth and development in yeast.Genetics. 2012; 192: 73-105Crossref PubMed Scopus (411) Google Scholar, 4Schmelzle T. Beck T. Martin D.E. Hall M.N. Activation of the RAS/cyclic AMP pathway suppresses a TOR deficiency in yeast.Mol. Cell Biol. 2004; 24: 338-351Crossref PubMed Scopus (211) Google Scholar, 5Zaman S. Lippman S.I. Schneper L. Slonim N. Broach J.R. Glucose regulates transcription in yeast through a network of signaling pathways.Mol. Syst. Biol. 2009; 5: 245Crossref PubMed Scopus (158) Google Scholar). Indeed, Sch9 was originally identified by virtue of its ability to suppress growth phenotypes associated with loss of PKA activity (6Toda T. Cameron S. Sass P. Wigler M. SCH9, a gene of Saccharomyces cerevisiae that encodes a protein distinct from, but functionally and structurally related to, cAMP-dependent protein kinase catalytic subunits.Genes Dev. 1988; 2: 517-527Crossref PubMed Scopus (146) Google Scholar). Reciprocally, hyper-activation of PKA signaling can suppress phenotypes linked to the loss of Sch9 activity (4Schmelzle T. Beck T. Martin D.E. Hall M.N. Activation of the RAS/cyclic AMP pathway suppresses a TOR deficiency in yeast.Mol. Cell Biol. 2004; 24: 338-351Crossref PubMed Scopus (211) Google Scholar). Protein kinase A Protein kinases of the family comprising PKA, PKG and PKC TOR complex 1/2 C3–1′-naphthyl-methyl PP1 false-discovery rate gene ontology peptide spectrum match. Protein kinase A Protein kinases of the family comprising PKA, PKG and PKC TOR complex 1/2 C3–1′-naphthyl-methyl PP1 false-discovery rate gene ontology peptide spectrum match. Tpk1, Tpk2, and Tpk3 are the partially redundant paralogs of the catalytic subunit of PKA. When cells are starved of carbon, cAMP levels are low and, therefore, PKA is kept inactive by its regulatory subunit Bcy1. Glucose addition induces activation of the adenylate cyclase Cyr1/Cdc35 via the small GTPase Ras1/2 and the G protein-coupled receptor Gpr1. The subsequent increase in cAMP levels triggers the dissociation of Bcy1 from the Tpks allowing them to phosphorylate their substrates (7Broach J.R. RAS genes in Saccharomyces cerevisiae: signal transduction in search of a pathway.Trends Genet TIG. 1991; 7: 28-33Abstract Full Text PDF PubMed Scopus (170) Google Scholar, 8Toda T. Cameron S. Sass P. Zoller M. Wigler M. Three different genes in S. cerevisiae encode the catalytic subunits of the cAMP-dependent protein kinase.Cell. 1987; 50: 277-287Abstract Full Text PDF PubMed Scopus (506) Google Scholar). Additionally, cAMP-independent activation of PKA has been reported (9Peeters T. Louwet W. Gelade R. Nauwelaers D. Thevelein J.M. Versele M. Kelch-repeat proteins interacting with the G protein Gpa2 bypass adenylate cyclase for direct regulation of protein kinase A in yeast.Proc. Natl. Acad. Sci. 2006; 103: 13034-13039Crossref PubMed Scopus (79) Google Scholar). In addition to cell growth, Tpk effectors influence many other processes including carbohydrate metabolism, cell cycle progression, sporulation, pseudohyphal development and longevity by controlling the activities of metabolic enzymes, transcription and autophagy factors (10Smith A. Ward M.P. Garrett S. Yeast PKA represses Msn2p/Msn4p-dependent gene expression to regulate growth, stress response and glycogen accumulation.EMBO J. 1998; 17: 3556-3564Crossref PubMed Scopus (283) Google Scholar, 11Shenhar G. Kassir Y. A positive regulator of mitosis, Sok2, functions as a negative regulator of meiosis in Saccharomyces cerevisiae.Mol. Cell Biol. 2001; 21: 1603-1612Crossref PubMed Scopus (63) Google Scholar, 12Longo V.D. The Ras and Sch9 pathways regulate stress resistance and longevity.Exp. Gerontol. 2003; 38: 807-811Crossref PubMed Scopus (100) Google Scholar, 13Van de Velde S. Thevelein J.M. Cyclic AMP-protein kinase A and Snf1 signaling mechanisms underlie the superior potency of sucrose for induction of filamentation in Saccharomyces cerevisiae.Eukaryot. Cell. 2008; 7: 286-293Crossref PubMed Scopus (35) Google Scholar). Similarly to TORC1, TORC2 phosphorylates and activates AGC-kinases, including Ypk1 and its redundant paralog Ypk2, in their hydrophobic motif (14Kamada Y. Fujioka Y. Suzuki N.N. Inagaki F. Wullschleger S. Loewith R. Hall M.N. Ohsumi Y. Tor2 directly phosphorylates the AGC kinase Ypk2 to regulate actin polarization.Mol. Cell Biol. 2005; 25: 7239-7248Crossref PubMed Scopus (166) Google Scholar, 15Niles B.J. Mogri H. Hill A. Vlahakis A. Powers T. Plasma membrane recruitment and activation of the AGC kinase Ypk1 is mediated by target of rapamycin complex 2 (TORC2) and its effector proteins Slm1 and Slm2.Proc. Natl. Acad. Sci. 2012; 109: 1536-1541Crossref PubMed Scopus (99) Google Scholar). Deletion of YPK2 produces no obvious phenotype suggesting that it only plays a minor role in cell growth control (16Chen P. Lee K.S. Levin D.E. A pair of putative protein kinase genes (YPK1 and YPK2) is required for cell growth in Saccharomyces cerevisiae.Mol. Gen. Genet. 1993; 236: 443-447Crossref PubMed Scopus (52) Google Scholar). Recent work has demonstrated that TORC2 is regulated downstream of membrane tension (17Berchtold D. Piccolis M. Chiaruttini N. Riezman I. Riezman H. Roux A. Walther T.C. Loewith R. Plasma membrane stress induces relocalization of Slm proteins and activation of TORC2 to promote sphingolipid synthesis.Nat. Cell Biol. 2012; 14: 542-547Crossref PubMed Scopus (234) Google Scholar), oxidative stress (18Niles B.J. Powers T. TOR complex 2–Ypk1 signaling regulates actin polarization via reactive oxygen species.Mol. Biol. Cell. 2014; 25: 3962-3972Crossref PubMed Scopus (41) Google Scholar) and carbon cues (19Hatano T. Morigasaki S. Tatebe H. Ikeda K. Shiozaki K. Fission yeast Ryh1 GTPase activates TOR Complex 2 in response to glucose.Cell Cycle Georget. Tex. 2015; 14: 848-856Crossref PubMed Scopus (36) Google Scholar). In turn, Ypk1, which is homologous to the mTORC2 substrate SGK in humans (20Eltschinger S. Loewith R. TOR complexes and the maintenance of cellular homeostasis.Trends Cell Biol. 2016; 26: 148-159Abstract Full Text Full Text PDF PubMed Scopus (126) Google Scholar, 21Casamayor A. Torrance P.D. Kobayashi T. Thorner J. Alessi D.R. Functional counterparts of mammalian protein kinases PDK1 and SGK in budding yeast.Curr. Biol. 1999; 9: 186-197Abstract Full Text Full Text PDF PubMed Scopus (203) Google Scholar), couples TORC2 signals to the regulation of membrane lipid biosynthesis and the regulation of cell surface area. Despite their central role in the regulation of fundamental cellular processes and intense efforts in understanding the functions controlled by these kinases, a systematic assessment of their targets has been lacking to date. A previous study aimed at systematically defining changes in the phospho-proteome associated with the absence of protein kinases, using kinase deletion strains (22Bodenmiller B. Aebersold R. Quantitative analysis of protein phosphorylation on a system-wide scale by mass spectrometry-based proteomics.Methods Enzymol. 2010; 470: 317-334Crossref PubMed Scopus (38) Google Scholar). However, this approach is limited to the study of non-essential kinases and allows cells to adapt to the absence of a kinase. To overcome these limitations, we employed yeast strains expressing analog-sensitive (23Bishop A.C. Buzko O. Shokat K.M. Magic bullets for protein kinases.Trends Cell Biol. 2001; 11: 167-172Abstract Full Text Full Text PDF PubMed Scopus (202) Google Scholar) variants of PKA, Sch9 and Ypk1. Mutation of the gatekeeper residue of these kinases allows binding of the bulky ATP-analog C3–1′-naphthyl-methyl PP1 (1NM-PP1), thus preventing ATP-binding and rendering the enzymes inactive. Using this selective and acute way of kinase inhibition, we explored the phospho-protein targets downstream of each of these major AGC-kinases by means of quantitative mass spectrometry. In these phospho-proteomics data sets we identified both known and potentially new targets of each tested kinase. As expected, we found extensive substrate overlap between PKA and Sch9. Unexpected was our finding that several substrates were shared between PKA and/or Sch9 and Ypk1 and that many sites hypo-phosphorylated upon Ypk1 inhibition resided in an RRxS/T-motif, which has previously been associated with PKA and Sch9, rather than Ypk1. Among the numerous potentially new kinase-substrate relationships discovered in this study, we chose neutral trehalase Nth1 as a candidate for follow-up experiments. Nth1 has been employed as a model PKA-substrate in multiple previous studies (24van der Plaat J.B. Cyclic 3′,5′-adenosine monophosphate stimulates trehalose degradation in baker's yeast.Biochem. Biophys. Res. Commun. 1974; 56: 580-587Crossref PubMed Scopus (148) Google Scholar, 25Schepers W. Van Zeebroeck G. Pinkse M. Verhaert P. Thevelein J.M. In vivo phosphorylation of Ser21 and Ser83 during nutrient-induced activation of the yeast protein kinase A (PKA) target trehalase.J. Biol. Chem. 2012; 287: 44130-44142Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar). Trehalose is a disaccharide that functions as a stress-protectant and reserve carbohydrate under adverse conditions (26Panek A. Synthesis of trehalose by baker's yeast (Saccharomyces cerevisiae).Arch. Biochem. Biophys. 1962; 98: 349-355Crossref PubMed Scopus (28) Google Scholar, 27Singer M.A. Lindquist S. Multiple effects of trehalose on protein folding in vitro and in vivo.Mol. Cell. 1998; 1: 639-648Abstract Full Text Full Text PDF PubMed Scopus (547) Google Scholar). Upon return to favorable conditions, trehalose is converted into two molecules of glucose by trehalases, including the neutral trehalase Nth1 (25Schepers W. Van Zeebroeck G. Pinkse M. Verhaert P. Thevelein J.M. In vivo phosphorylation of Ser21 and Ser83 during nutrient-induced activation of the yeast protein kinase A (PKA) target trehalase.J. Biol. Chem. 2012; 287: 44130-44142Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar). The regulation of Nth1 by PKA has long been recognized as important for cell survival (24van der Plaat J.B. Cyclic 3′,5′-adenosine monophosphate stimulates trehalose degradation in baker's yeast.Biochem. Biophys. Res. Commun. 1974; 56: 580-587Crossref PubMed Scopus (148) Google Scholar, 28Wera S. De Schrijver E. Geyskens I. Nwaka S. Thevelein J.M. Opposite roles of trehalase activity in heat-shock recovery and heat-shock survival in Saccharomyces cerevisiae.Biochem. J. 1999; 343: 621-626Crossref PubMed Scopus (61) Google Scholar). Here, we site-specifically validated that Ypk1-, as well as PKA-inhibition reduced its phosphorylation of an RRxS-site and showed that this is associated with reduced trehalase activity. These findings highlight the need for revisiting the Ypk1 consensus motif and prompt further investigation of the relationship of PKA and TORC1-Sch9 signaling on one hand and TORC2-Ypk1 on the other hand. Saccharomyces cerevisiae strains used in this study are listed in Table I. Strains were constructed using standard yeast genetic manipulation (29Amberg D.C. Burke D. Strathern J.N. Burke D. Methods in yeast genetics: a Cold Spring Harbor Laboratory course manual. 2005 ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y2005: 230Google Scholar).Table IS. cerevisiae strains used in this studyReferred to asGenotypeNamepkaas+sch9asTB50a tpk1-M164G tpk2-M147G tpk3-M165G SCH9::sch9-T492G(pRS304) ypk1::Kan ypk2::HphNT pRS416-YPK1wtMPP88pkaasTB50a tpk1-as tpk2-as tpk3-as SCH9::SCH9wt(pRS304) ypk1::Kan ypk2::HphNT pRS416-YPK1wtMPP86sch9asTB50a Sch9::sch9-T492G(pRS304) ypk1::Kan ypk2::HphNT pRS416-YPK1-wt(791)MPP84ypk1asTB50a SCH9::SCH9wt(pRS304) ypk1::Kan ypk2::HphNT pRS416-ypk1-L424GMPP91pkaas+ypk1asTB50a tpk1-M164G tpk2-M147G tpk3-M165G SCH9::SCH9wt(pRS304) ypk1::Kan ypk2::HphNT pRS416-ypk1-L424GMPP87wtas NTH1–3xFLAGTB50a SCH9::SCH9wt(pRS304) ypk1::Kan ypk2::HphNT pRS416-YPK1wt NTH1–3xFLAG::HIS3yMJP012pkaas NTH1–3xFLAGTB50a tpk1-M164G tpk2-M147G tpk3-M165G SCH9::SCH9wt(pRS304) ypk1::Kan ypk2::HphNT pRS416-YPK1wt NTH1–3xFLAG::HIS3yMJP010ypk1as NTH1–3xFLAGTB50a SCH9::SCH9wt(pRS304) ypk1::Kan ypk2::HphNT pRS416-ypk1-L424G NTH1–3xFLAG::HIS3yMJP009pkaas+ypk1as NTH1–3xFLAGTB50a tpk1-M164G tpk2-M147G tpk3-M165G SCH9::SCH9wt(pRS304) ypk1::Kan ypk2::HphNT pRS416-ypk1-L424G NTH1–3xFLAG::HIS3yMJP040wtas nth1–5A-3xFLAGTB50a SCH9::SCH9wt(pRS304) ypk1::Kan ypk2::HphNT pRS416-YPK1wt nth1-S20A,S21A,T58A,S60A,S83A-3xFLAG::HIS3yMJP033nth1ΔTB50a SCH9::SCH9wt(pRS304) ypk1::Kan ypk2::HphNT pRS416-YPK1wt nth1Δ::NatNT2yMJP013 Open table in a new tab Three replicates each of DMSO- and 1NM-PP1- (Merck) treated sch9as, pkaas, pkaas+sch9as and ypk1as cultures were prepared for label-free phospho-proteomics as previously described (30Huber A. Bodenmiller B. Uotila A. Stahl M. Wanka S. Gerrits B. Aebersold R. Loewith R. Characterization of the rapamycin-sensitive phosphoproteome reveals that Sch9 is a central coordinator of protein synthesis.Genes Dev. 2009; 23: 1929-1943Crossref PubMed Scopus (261) Google Scholar). In brief, all cultures were diluted to OD 0.2 and grown to exponential phase (OD 0.7–0.8). Each strain was treated with 500 nm 1NM-PP1 or DMSO for 15 min after which ice cold 100% TCA was added to a final concentration of 6%. The samples were incubated on ice for 30 min. After centrifugation the cells were washed twice with acetone and kept at −70 °C. Cell pellets were suspended in a buffer consisting of 8 m urea, 50 mm ammonium bicarbonate, and 5 mm EDTA, then lysed by bead beating. For each biological replicate, 3 mg protein were reduced using 5 mm TCEP, alkylated in 10 mm iodoacetamide, and digested overnight with trypsin (1:125 w/w). Reverse phase chromatography was used to purify samples before phospho-peptide enrichment, which was performed with titanium dioxide resin (1.25 mg resin for each sample) (22Bodenmiller B. Aebersold R. Quantitative analysis of protein phosphorylation on a system-wide scale by mass spectrometry-based proteomics.Methods Enzymol. 2010; 470: 317-334Crossref PubMed Scopus (38) Google Scholar). Isolated phospho-peptides were analyzed on an LTQ-Obritrap XL mass spectrometer (Thermo Scientific). A 90-min gradient, starting with 3% and ending with 23% acetonitrile, was used for liquid chromatography elution. The 4 most intense ions detected in each MS1 measurement were selected for MS2 fragmentation. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (31Perez-Riverol Y. Csordas A. Bai J. Bernal-Llinares M. Hewapathirana S. Kundu D.J. Inuganti A. Griss J. Mayer G. Eisenacher M. Pérez E. Uszkoreit J. Pfeuffer J. Sachsenberg T. Yilmaz S. Tiwary S. Cox J. Audain E. Walzer M. Jarnuczak A.F. Ternent T. Brazma A. Vizcaíno J.A. The PRIDE database and related tools and resources in 2019: improving support for quantification data.Nucleic Acids Res. 2019; 47: D442-D450Crossref PubMed Scopus (4293) Google Scholar) partner repository with the data set identifier PXD015668 and 10.6019/PXD015668. Mass spectrometry data were processed with the MaxQuant software (v. 1.6.0.16), which uses the Andomeda MS/MS spectrum search engine (32Cox J. Mann M. MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification.Nat. Biotechnol. 2008; 26: 1367-1372Crossref PubMed Scopus (9154) Google Scholar, 33Cox J. Neuhauser N. Michalski A. Scheltema R.A. Olsen J.V. Mann M. Andromeda: a peptide search engine integrated into the MaxQuant environment.J Proteome Res. 2011; 10: 1794-1805Crossref PubMed Scopus (3450) Google Scholar). The following search settings were used: enzyme was Trypsin/P with 2 allowed missed cleavages. For precursor ions the first search tolerance was set to 20 ppm and the main search tolerance to 4.5 ppm and for fragment ions the tolerance was set to 0.5 Th. Phosphorylation of serine, threonine and tyrosine residues, oxidation of methionine, acetylation of protein N termini and deamidation of asparagine and glutamine were used as variable modifications (maximal 5 modifications per peptide). Carbamidomethylation of cysteine was used as fixed modification. The SGD yeast protein database (https://downloads.yeastgenome.org/sequence/S288C_reference/orf_protein/orf_trans_all.fasta.gz; 6713 entries) from 2015–01-13 was used to match the spectra. Finally, a peptide and modification site FDR of 0.01 was applied to filter peptide spectrum matches (PSMs). Contaminants and decoys were subsequently removed from the results. The MaxQuant result tables (msms.txt for PSMs and positioning information, modificationSpecificPeptides.txt for intensity values; provided as supplemental Table S1 and S2 respectively) were imported into the R statistical environment to perform missing value imputation and intensity normalization. Only peptides containing at least one phosphorylation were retained, together with all possible positions of the phosphorylation and their respective scores. All intensity values were log2 transformed (supplemental Table S3). Supplemental Fig. S1A shows the occurrence of missing values across the 3000+ phospho-peptides revealing that certain peptides and certain MS-runs are more prone to missing values. An example of a peptide with missing values can be seen in supplemental Fig. S1B, where all Sch9/PKA-1NM-PP1 values of the peptide HSS(ph)PDPYGINDKFFDLEK and one Sch9 and one PKA value are missing. In the sibling peptide (see below) HSS(ph)PDPYGINDK with clearly higher MS1 intensities all values are present (supplemental Fig. S1C). This shows that even if all 3 replicates values are missing, one cannot conclude absence of the corresponding peptide, but these missing values are most likely artifacts inherent to MS1 based quantitation. Phospho-peptides with no missing values in 3 replicates have stronger average MS1 signal intensities compared with peptides with missing values (supplemental Fig. S1D). If only one value out of 3 replicate values is missing, the 2 remaining positive replicate values can be used to estimate or impute the missing values (see below). However, if 2 or 3 values are missing, imputation becomes more error prone and can easily lead to false positives, when searching for peptides of differential abundance. Therefore, we only retained peptides with maximally one missing value per 3 replicates. 67.3% of all phospho-peptides fall in this category (supplemental Fig. S1E). We imputed a missing log2-transformed intensity value IP,S,i of peptide P in replicate i of sample S by IP,S,i=αP,S,i+N(μ,σ)(1) where αP,S,i is the mean of the 2 non-missing values of peptide P in sample S and N(μ, σ) is a Gaussian error term with mean μ and standard deviation σ. In order to estimate this error term, we first need the notion of sibling peptides. Two peptides are called siblings if the sequence of one peptide is a N/C-terminal extension of the sequence of the other one (i.e. they differ by a missed cleavage) and if they carry the same modifications. It is unlikely that the sample condition has a direct influence on the trypsin cleavage, and it can be expected that two sibling peptides have similar expression profiles. The imputation parameters μ and σ are then estimated using the sibling peptides present in the MaxQuant results: for each peptide P with a missing value in replicate i in sample S, we searched for sibling peptides P' of P with no missing values in the 3 replicates of sample S. Then we calculated the difference between the IP',S,i and αP',S,i and used all these differences to estimate μ and σ. The result shown in supplemental Fig. S1F reveals that the missing values are on average slightly, but significantly lower than the average of the 2 non-missing values (t test p value of 0.003), a small bias that we took into account in the imputation Eq. 1). After missing value imputation, the R package limma (34Smyth G.K. Linear models and empirical bayes methods for assessing differential expression in microarray experiments.Stat. Appl. Genet. Mol. Biol. 2004; 3 (Article3)Crossref PubMed Scopus (9158) Google Scholar) was used for intensity normalization. First the normalizeBetweenArrays R-function was used to equalize the log2-intensity distributions in all 30 LC-MS runs. The lmFit function was used to remove eventual batch effects present in the 10 samples. The empirical Bayes (eBayes) method was the applied to compare the 1NM-PP1 treated to the control (DMSO) samples for each strain and to calculate fold-changes and p values adjusted for multiple testing (FDR adjustment) (supplemental Table S3). This method performs a correction of the standard deviation of each peptide by the overall empirical standard deviation. For each phospho-peptide that exhibited a change upon treatment with an adjusted p value of less than 0.05 the sequence from −5 to +5 amino acids from each of the phospho-sites localized on the peptide was extracted. In case of ambiguity between isoforms with different phospho-localizations, the peptide with the best localization score was chosen. If a phospho-site was represented by several peptides, the sequence surrounding the site was listed only once. This was performed individually for each strain for hyper- and hypo-phosphorylated peptides and the union of both. Sequence logos were generated with pLogo using all quantified phospho-peptides as a background (35O'Shea J.P. Chou M.F. Quader S.A. Ryan J.K. Church G.M. Schwartz D. pLogo: a probabilistic approach to visualizing sequence motifs.Nat. Methods. 2013; 10: 1211-1212Crossref PubMed Scopus (244) Google Scholar). Sequence motifs of sites hypo-phosphorylated upon TORC2-inhibition in Rispal et al., 2015 were based on sites in supplemental Table S2 of that study and generated in an analogous manner (36Rispal D. Eltschinger S. Stahl M. Vaga S. Bodenmiller B. Abraham Y. Filipuzzi I. Movva N.R. Aebersold R. Helliwell S.B. Loewith R. Target of rapamycin complex 2 regulates actin polarization and endocytosis via multiple pathways.J. Biol. Chem. 2015; 290: 14963-14978Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar). For GO-analysis the systematic gene names of proteins harboring phospho-sites significantly changing upon kinase inhibition with an adjusted p value below 0.05 were used as a foreground and harboring any quantified phospho-site as a background. The analysis was performed using the "GO Term Finder" tool at the Saccharomyces Genome Database at a p value of 0.05 with FDR calculation enabled (37Cherry J.M. Hong E.L. Amundsen C. Balakrishnan R. Binkley G. Chan E.T. Christie K.R. Costanzo M.C. Dwight S.S. Engel S.R. Fisk D.G. Hirschman J.E. Hitz B.C. Karra K. Krieger C.J. Miyasato S.R. Nash R.S. Park J. Skrzypek M.S. Simison M. Weng S. Wong E.D. Saccharomyces Genome Database: the genomics resource of budding yeast.Nucleic Acids Res. 2012; 40: D700-D705Crossref PubMed Scopus (1144) Google Scholar). The network of proteins associated with GO-term "Endocytosis" in the Ypk1-data set was displayed in the STRING-database tool at default parameters (38Snel B. Lehmann G. Bork P. Huynen M.A. STRING: a web-server to retrieve and display the repeatedly occurring neighbourhood of a gene.Nucleic Acids Res. 2000; 28: 3442-3444Crossref PubMed Scopus (725) Google Scholar). In the label-free phospho-proteomics, targeted proteomics and the trehalase activity assay, the stated number of replicate cultures of analog-sensitive strains or wt-strain (wtas) were treated with ATP-analog (1NM-PP1) or mock (DMSO). Three replicates per strain were used for shotgun and targeted proteomics and six replicates for the trehalase assay. Independent starter cultures were used for each strain to generate replicates. The number of replicates used is the highest number deemed justified considering cost and the ability to process samples in parallel. No replication was used in Western blotting as the results were verified by Parallel Reaction Monitoring (PRM) as an alternative approach. A strain which was not expected to be affected by 1NM-PP1 addition (wtas) was employed in all experiments. For Western blotting, a strain not carrying a FLAG-tag and another in which five serines of Nth1 had been mutated to alanine were used as additional controls. For label-free phospho-proteomics, after data processing as described above, the empirical Bayes (eBayes) method of the limma R package (34Smyth G.K. Linear models and empirical bayes methods for assessing differential expression in microarray experiments.Stat. Appl. Genet. Mol. Biol. 2004; 3 (Article3)Crossref PubMed Scopus (9158) Google Scholar) was then applied to compare the 1NM-PP1 treated to the mock-treated samples for each strain and to calculate fold-changes and p values adjusted for multiple testing (FDR adjustment). Using this approach an observed false positive rate of 0.4% was calculated based on the results from the wtas strain. For targeted proteomics and the trehalase activity assay, 1NM-PP1-treated samples were compared with mock-treated samples using unpaired, two-sided Student's t-tests for each strain. When analyzing sets of data points, rather than individual data points (e.g. phospho-sites), a trade-off exists between the quality with which each point is measured and the number of points representing each set. To achieve meaningful set-sizes we did not apply further localization score thresholds when analyzing ensembles of phospho-sites
Referência(s)