Interaction Proteomics Suggests a New Role for the Tfs1 Protein in

Article pubs.acs.org/jpr

Interaction Proteomics Suggests a New Role for the Tfs1 Protein in Yeast Martine Beaufour, Fabienne Godin, Béatrice Vallée, Martine Cadene,* and Hélène Bénédetti Centre de Biophysique Moléculaire, Centre National de la Recherche Scientifique, UPR 4301, affiliated to the University of Orléans, rue Charles Sadron, 45071 Orléans cedex 2 S Supporting Information *

ABSTRACT: The PEBP (phosphatidylethanolamine-binding protein) family is a large group of proteins whose human member, hPEBP1, has been shown to play multiple functions, influencing intracellular signaling cascades, cell cycle regulation, neurodegenerative processes, and reproduction. It also acts, by an unknown mechanism, as a metastasis suppressor in a number of cancers. A more complete understanding of its biological role is thus necessary. As the yeast Saccharomyces cerevisiae is a powerful and easy to handle model organism, we focused on Tfs1p, the yeast ortholog of hPEBP1. In a previous study based on a two-hybrid approach, we showed that Tfs1p interacts and inhibits Ira2p, a GTPase Activating Protein (GAP) of the small GTPase Ras. To further characterize the molecular functions of Tfs1p, we undertook the identification of protein complexes formed around Tfs1p using a targeted proteomics approach. Complexed proteins were purified by tandemaffinity, cleaved with trypsin, and identified by nanoflow liquid chromatography coupled with tandem mass spectrometry. Overall, 14 new interactors were identified, including several proteins involved in intermediate metabolism. We confirmed by coimmunoprecipitation that Tfs1p interacts with Glo3p, a GAP for Arf GTPases belonging to the Ras superfamily of small GTPases, indicating that Tfs1p may be involved in the regulation of another GAP. We similarly confirmed the binding of Tfs1p with the metabolic enzymes Idp1p and Pro1p. Integration of these results with known functional partners of Tfs1p shows that two subnetworks meet through the Tfs1p node, suggesting that it may act as a bridge between cell signaling and intermediate metabolism in yeast. KEYWORDS: targeted proteomics, protein−protein interactions, protein function, cell signaling, intermediate metabolism, PEBP/RKIP homologues, Saccharomyces cerevisiae

■

(GAP) for Ras (Ras1p and Ras2p) GTPases.13 In normal growth conditions, the inhibition of Ira2p has been shown to be the predominant function.14 Ras is a small G-protein which oscillates between an inactive state, Ras-GDP, and an active state, Ras-GTP. Ras is tightly regulated by two antagonistic proteins: Cdc25p and the Ira proteins (Ira1p and Ira2p). The Ira proteins are GAPs which activate the weak intrinsic GTPase activity of Ras and favor the Ras-GDP form. Cdc25p, a guanine exchange factor (GEF), promotes the GDP/GTP exchange reaction on Ras, thereby activating it to its GTP-bound form.15 When activated, Ras stimulates the cAMP/PKA pathway, a cell signaling pathway which controls many important processes such as metabolism (glycolysis and glycogenesis), stress resistance, growth and meiosis.16 TFS1 overexpression is induced when a stress occurs.17 By inhibiting Ira2p, Tfs1p might therefore influence metabolic processes, and also induce a negative control of the general stress response. Indeed, the current hypothesis is that Tfs1p might exert a negative feedback

INTRODUCTION The PEBP (phosphatidylethanolamine-binding protein) family is a large group of conserved proteins with orthologs in a wide variety of organisms. PEBP proteins share a common structural scaffold as well as a surface binding pocket for anionic groups. One of the human family members, hPEBP1, also known as RKIP (Raf kinase inhibitor protein), is a multifunctional protein which has been shown to play a major role in various processes. Human PEBP1 (i) controls different signaling pathways by inhibiting different kinases;1−3 (ii) inhibits serine proteases,4 (iii) binds lipids;5 (iv) generates HCNP, a neurostimulatory peptide involved in the development of hippocampal and cardiac physiology;6,7 (v) affects heterotrimeric G protein-dependent signaling;8 and (vi) is involved in a growing number of cancers where it acts as a metastasis suppressor.9 Tfs1p is a PEBP ortholog in yeast. Two functions have been assigned to Tfs1p so far. On one hand, it inhibits the carboxypeptidase Y vacuolar serine proteinase (CPY, also known as Prc1p),10−12 and on the other hand, as we have previously shown, it inhibits Ira2p, a GTPase-activating protein © 2012 American Chemical Society

Received: December 19, 2011 Published: April 19, 2012 3211

dx.doi.org/10.1021/pr201239t | J. Proteome Res. 2012, 11, 3211−3218

Journal of Proteome Research

Article

loop on this response.13 By analogy with the human ortholog of Tfs1p, it is reasonable to assume that Tfs1p may carry other functions in addition to the two aforementioned ones. Moreover, as yeast is a powerful model organism, it could provide more clues about the role of PEBP proteins. Physical partners of a protein are potent indicators of its molecular and cellular functions. We decided to systematically look for partners of Tfs1p in order to gain insight into its biological role in the cell, and to shed light on new functions of its human homologue RKIP, which has been assigned multiple functions and is associated with an increasing number of diseases through its involvement with signal transduction pathways. The yeast two-hybrid screening we performed on Tfs1p allowed us to identify only one new partner of Tfs1p, Ira2p.13 High-throughput (HTP) affinity-MS studies, in addition to Prc1p,18 found only two other partners: Pet111p (a mitochondrial translational activator specific for the COX2 mRNA) and the isocitrate deshydrogenase Idp3p.19 In the approach described here, we decided to undertake a lowthroughput (LTP) identification of Tfs1p partners using TAPtag proteomics with careful manual validation, hoping to identify new partners compared with the two-hybrid and HTP approaches. In the work described here, 14 new Tfs1p partners were identified, including Glo3p, a GAP for Arf (ADP Ribosylation Factor) GTPases, as well as proteins involved in intermediate metabolism. The interaction of Tfs1p with Glo3p, and two enzymes of intermediate metabolism were confirmed using coimmunoprecipitation. A map combining Tfs1p physical and genetic interactions was drawn. This map sheds light on Tfs1p functions and suggests a new biological role of Tfs1p in the yeast cell.

expression of the fusion Tfs1-TAP tag protein in the yeast cells was confirmed by Western blot using anti-Tfs1p antibodies.13 Heat Shock Resistance Tests

The tests were performed as previously described by Chautard et al.13 on BY4742, BY4742 TFS1-TAP, and BY4742ΔT (Δtfs1) strains. Tandem Affinity Purification of Tagged Complexes

Cells were pelleted by centrifugation at 7500g in a BR4i centrifuge and the pellets were washed with Tris buffer consisting of 10 mM Tris-HCl (pH 8), and 150 mM NaCl. Cell pellets were then suspended in a solution of Tris buffer, 0.1% NP-40, 1 mM PMSF, 2% (w/v) DNase, and a protease inhibitor cocktail tablet (Roche Diagnostics, Mannheim, Germany) to an absorbance of about 200 units/mL. Drops of the suspension (equivalent to 2.5 g of cells) were quickly frozen in liquid nitrogen. Cells were broken using a MM301 cell disrupter (Retsch, Haan, Germany) by performing five cycles of 3 min at 30 Hz. The supernatant was recovered after two centrifugations at 3200g followed by a 45 min centrifugation step at 14 000g to remove cell debris. The first affinity purification step was performed using IgG-sepharose beads (GE Healthcare, Uppsala, Sweden) in manually packed columns. For this purpose, cell extracts were mixed with IgG beads at a ratio of 10 mL of extract to 200 μL of beads for 2 h at 4 °C with rotary mixing. Five washing steps were performed to remove nonspecific interactions. Washing buffer consisted of 10 mL of 10 mM Tris-HCl, pH 8, with 150 mM NaCl. The bound material was released by proteolytic cleavage at the specific TEV site using TEV proteinase (Invitrogen, Carlsbad, CA) prepared at 100 units/mL in TEV buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% NP-40, 0.5 mM EDTA, 1 mM TCEP). Cleavage was performed by adding 1 mL of this solution to the beads and rotary mixing for 2 h at 16 °C. The eluate was recovered by gravity and the beads were washed with 200 μL of TEV buffer. The eluate was mixed in a 1 to 3 volume ratio with calmodulin buffer (10 mM Tris, pH 8.0, 150 mM NaCl, 0.1% NP-40, 1 mM Mg acetate, 1 mM imidazole, 2 mM CaCl2, and 5 mM TCEP) and incubated with 200 μL of calmodulin-coated Sepharose beads (Stratagene, Amsterdam, The Netherlands) for 1 h at 4 °C on a rotary mixer. The beads were washed 3 times with 5 mL of calmodulin buffer, and resuspended in ammonium bicarbonate buffer (10 mM NH4CO3, 1 mM TCEP, and 0.1% SDS). The material bound on calmodulin beads was released by boiling at 95 °C. The same protocol was applied to the control S. cerevisiae wild-type strain.

■

EXPERIMENTAL PROCEDURES All chemicals were reagent-grade unless otherwise noted. All solvents were HPLC-grade. Water was purified by reverse osmosis followed by ion-exchange (milli-Q system, Millipore, Milford, MA). Yeast Strain Construction

The TAP (Tandem Affinity Purification) tag approach20 was used in order to characterize complexes formed with the Saccharomyces cerevisiae Tfs1 protein. To maintain the expression of the TFS1 gene under its natural promoter, yeast homologous recombination was used to directly integrate the DNA sequence encoding the TAP tag in 3′ of TFS1 gene in the genome of the BY4742 strain (Euroscarf, Frankfurt, Germany), thereby coding for a C-terminally tagged fusion protein. For this purpose, a PCR-generated fragment containing the DNA sequence of the tag and the URA3 marker flanked by 57−43 bases homologous to the region located immediately upstream and downstream of the last codon of TFS1 was amplified from the pBS1539 plasmid with the oligonucleotides HB59 (5′-GCTAAGGAAAACAACCTGCAACTAGTTGCCTCCAATTTCTTCTATGCGGAAACGAAATCCATGGAAAAGAGAAG-3′) and HB60 (5′GTCAAGTGAAAAGCACTGAAATCTAAAAAATAAATACATATACTACGACTCACTATAGGG-3′), respectively. The S. cerevisiae BY4742 wild-type strain was used as a negative control.

1D SDS-PAGE

The purified protein complexes were loaded onto a 1 mm 4− 12% NuPAGE Bis-Tris mini-gel (Invitrogen), and electrophoresis was performed under reducing conditions using an Xcell SureLock cell (Invitrogen). Bovine serum albumin (1 mg/mL in water) was loaded as a positive control for protein identification. Dual Color Precision Plus protein standards (Bio-Rad, Hercules, CA) were used as molecular weight markers. The gel was stained using negative zinc staining for protein band visualization (Zinc staining kit, Bio-Rad). Gel bands were excised in a sterile laminar flow hood, transferred to 1.5 mL microtubes (Eppendorf AG, Hamburg, Germany), and cut into cubes of roughly 1 mm3. A gel band was also cut in a blank region as background control. At each of the following steps, the supernatant was discarded after table-top centrifuga-

Cell Growth

Yeast cells were grown in YEPD broth to an absorbance value of 2 at 600 nm in a culture volume of 2 L at 30 °C. The 3212



Article

tion. Gel cubes were destained for 45 min with 50 μL of the zinc staining kit Cu/Zn destain solution. Subsequently, gels pieces were shrunk with 100 μL of pure acetonitrile solvent, washed with 4 μL of 50 mM ammonium bicarbonate, and shrunk again with 100 μL of acetonitrile.

polysiloxanes contaminants was used. Dynamic exclusion was enabled with a repeat count of 2, and an exclusion duration of 30 s. Protein Search Methods

MS/MS spectra were searched against the S. cerevisiae subset of the nonredundant NCBI protein sequence database (Release 20070316) using Mascot (v.2.0, Matrix Science, London, U.K.) and PEAKS Studio (version 4.5, BSI, Waterloo, Canada) software. Raw ion trap data were submitted to the Mascot server after preprocessing with DataAnalysis version 3.1 (Bruker). Mass lists from mass spectral peaks were generated with this program using the following parameters: S/N threshold = 1, fwhm = 0.1, relative intensity threshold = 2, absolute intensity threshold = 100. Charge deconvolution was also performed for 1+ to 3+ charge states. The resulting peak lists were then converted to mgf file format for Mascot analysis. For PEAKS, raw data were first converted to the mzXML file format and then charge attributed (precursor ions, 1+ to 3+), filtered (quality factor = 0.65), and preprocessed (noise removal, centroiding, and deconvolution of MS/MS spectra) within the software. Data were merged within a 3 min window. The search parameters settings for both Mascot and PEAKS search engines were as follows: 2 maximum missed cleavages, 2 Da mass accuracy for precursor ions, and 0.5 Da tolerance for fragment ions. The variable modifications were methionine oxidation and cysteine propionamidation (from residual acrylamide). In the Mascot scoring algorithm, the protein score is related to the size of the database that is used. In the case of this study, a match with an ion score greater than 34 is considered to be identical or have extensive homology, and only proteins with scores above this threshold are reported, with a minimum peptide score of 20. For PEAKS, only proteins having peptide scores greater than 50% and a protein score greater than 90% were retained. With such thresholds, all proteins that were found with PEAKS had a minimum of two unique peptide sequences identified. As proteins found in the wild-type strain interact nonspecifically with the beads, proteins that were found in both wild-type and genomically TAP-tagged strain experiments were not listed in the results.

In-Gel Trypsin Proteolysis

Gel pieces were subjected to tryptic proteolysis. A 5 μL aliquot of 1 μg/μL modified trypsin, sequencing grade (Roche Diagnostics) in 10 mM hydrochloric acid was activated by adding 45 μL of mQ water and 50 μL of 100 mM ammonium bicarbonate. A 2 μL aliquot of the activated trypsin solution was added to each gel sample and the gel pieces were allowed to swell for 5 min. Addition of 15 μL of 50 mM ammonium bicarbonate was followed by a 4 h incubation at 37 °C. Proteolytic peptides were extracted from the samples using POROS beads (POROS 20 R2 resin, Applied Biosystems, Framingham, MA). For this purpose, a POROS beads suspension was prepared by diluting a 50% methanol POROS beads solution 20-fold into a 5% formic acid, 0.2% TFA solution. The peptide extraction was performed overnight at 4 °C using 10 μL of POROS beads suspension per microtube. Beads were pelleted down and stored at −80 °C before mass spectrometry analysis. Sample Preparation for Mass Spectrometry Analysis

The POROS-extracted samples were brought back to room temperature and spun down. The POROS beads were resuspended using gel-loading pipet tips (VWR, Strasbourg, France), and loaded on top of preconditioned C18 ZipTips (Millipore) for peptide recovery. All steps were performed by top-loading of solutions and table-top microfuge centrifugation. ZipTips were washed with 60 μL of 0.1% trifluoroacetic acid and the peptides were eluted with 10 μL of elution solvent consisting of 50% acetonitrile with 0.1% formic acid. The eluates were evaporated to dryness in a Speed-Vac (Fisher Scientific, Illkirch, France) and reconstituted in 10 μL of 2% acetonitrile solution containing 0.1% formic acid for subsequent injection in the chromatographic system. LC-NanoESI−MS/MS

Liquid chromatographic separations were performed on an UltiMate nano-HPLC system equipped with an autosampler (LC Packings, Amsterdam, The Netherlands). A 15 cm reversephase column (Pepmap C18, 75 μm inner diameter, 3 μm particle size diameter, 100 Å pore size, LC Packings) was used to separate the peptide mixtures. Peptides were injected in a 1 μL volume and separated using a linear gradient of 0−50% B, solvent A consisting of 5% acetonitrile with 0.1% formic acid and solvent B consisting of 80% acetonitrile with 0.1% formic acid. The total run time was 70 min and the flow rate was measured at 250 nL/min with mobile phase A. The column was connected to the nanospray source of the ion trap (Esquire HCT from Bruker, Bremen, Germany). A distal coated PicoTip emitter (360 μm outer diameter, 20 μm inner diameter, 10 μm inner tip diameter, New Objective, Woburn, MA) was used for spraying. Mass spectrometry parameters were set via EsquireControl software version 5.1 (Bruker). The source was operated at a spray voltage of 2000 V. The eluent from the LC was not subjected to MS/MS for the first 8 min. The ion trap scanned in MS mode between m/z 250 and 2000. The 5 highest intensity peaks were selected for MS/MS fragmentation with a 4 Da isolation window. An exclusion list of m/z values for common trypsin autolysis products, plasticizers, and

Co-immunoprecipitation Experiments

Four hit proteins identified in the TAP-MS experiment were selected for confirmation by co-immunoprecipitation. Commercial strains based on the BY4741 background (WT, Wild Type) and containing plasmids encoding the N-terminally tagged GST fusion proteins of the selected hits were purchased from Open Biosystems (Huntsville, AL). The negative control experiment consisted of the BY4741 TFS1 knockout cells (Δtfs1) which were transformed with the plasmids extracted from the commercially available GST-tagged cells. Cells were grown on galactose selective medium until the absorbance at 600 nm reached a value of 1.0. Cell lysis was done in the presence of glass beads in a lysis buffer containing PBS (4 mM phosphate buffer, 150 mM NaCl, pH 7,4), 0.1% NP-40, 1 mM PMSF, and complete protease inhibitor cocktail (Roche Diagnostics). Protein A coated magnetic beads from Invitrogen (1.8 mg in 60 μL) were washed with solution A (PBS, 0.1% NP-40) and incubated for 10 min with rabbit anti-Tfs1p antibodies prepared as described.13 Beads were washed 3 times with solution A and then incubated with 300 μL of cell extract for 15 h at 4 °C. After 3 washes with solution A and 2 washes with PBS, beads were eluted with electrophoresis sample buffer. 3213



Article

Table 1. Proteins Identified by NanoLC−MS/MS Analysis Following Purification with TAP-Tagged Tfs1p PEAKS Studio

a

Mascot

protein

gene

score (%)

number of unique peptides

score

number of unique peptides

molecular function (biological process)a

Rpg1p Yap1p Nrd1p Bfr1p Cpr6p Lap4p Glo3p Rio2p Kes1p Lpd1p Cit1p Idp1p Ade3p Pro1p

YBR079C YML007W YNL251C YOR198C YLR216C YKL103C YER122C YNL207W YPL145C YFL018C YNR001C YDL066W YGR204W YDR300C

99 98 91 95 99 92 94 99 98 93 99 92 94 91

2 3 2 2 2 2 3 3 2 2 5 3 3 2

130 147 43 98 210 71 166 171 152 128 240 144 145 136

3 4 1 3 5 2 3 4 3 3 6 4 5 3

Transcription initiation factor Transcription factor Termination of transcription regulator mRNP component (Translation) Prolyl cis−trans isomerase (Folding) Proteinase Arf GAP (Trafficking) Protein kinase (Ribosome synthesis) Lipid binding (Biosynthesis/Trafficking) Energy metabolism enzyme Energy metabolism enzyme Energy metabolism + Biosynthesis enzyme Biosynthesis enzyme Biosynthesis enzyme

According to the Saccharomyces Genome Database.

(Lap4p) is a proteinase, suggesting a role in proteinase inhibition as previously established for CPY.10−12 The remaining five proteins (Rpg1p, Yap1p, Nrd1p, Bfr1p, and Cpr6p) participate in transcription/translation/folding processes. The identification engines’ results seem to show a higher stringency for Peaks Studio identification compared to Mascot, which is probably due to the use of Peaks Studio’s additional filtering at the protein score level. The proteins identified and reported in Table 1 were compared to known physical interactors of Tfs1p, as determined either by high-throughput (HTP) MS-based approaches (Table 2) or by biochemistry/biology-based

Eluted samples were subjected to Western blot. Immunodetection was done using mouse anti-GST antibodies as primary antibodies (Invitrogen) and rabbit anti-mouse antibodies coupled to horseradish peroxidase (Invitrogen) as secondary antibodies. The latter were detected using the chemiluminescent SuperSignal West Dura substrate (Thermo Scientific, Rockford, IL).

■

RESULTS

TAP-Tag Results

We expressed the C-terminally TAP-tagged Tfs1p in a genomically modified yeast strain (BY4742 TFS1-TAP). Tfs1p has been shown to positively regulate the Ras/cAMP/ PKA pathway which controls yeast cells sensitivity to stress.13,21 Stress sensitivity can be used to ascertain that the gene is fully functional. BY4742 TFS1-TAP sensitivity to heat shock was tested and compared to that of a wild-type strain (BY4742) and a mutant strain deleted of TFS1 (BY4742ΔT). The BY4742 TFS1-TAP strain was shown to be as sensitive to heat shock as the wild-type strain, thereby demonstrating that the TAPtagged Tfs1p was functional (see Supporting Information). The wild-type strain was used as a control and the same protocol applied to both strains for the preparation of cell extracts and their subsequent tandem affinity purification. TAP eluates were separated by one dimension SDS-PAGE (see Supporting Information) followed by in-gel trypsin proteolysis of the gel bands. The final steps consisted in mass spectrometry analysis by LC−MS/MS and database searching to identify the interactors of Tfs1p. Two search engines using different algorithms, Mascot and PEAKS Studio, were used in order to increase confidence in the results. Proteins that were found in both strains were considered as nonspecifically interacting with the purification beads. Table 1 presents the 14 new interactors of Tfs1p identified in this work, classified by functional families. Seven proteins are involved in energy metabolism and/or biosynthesis: two of them in energy metabolism (Lpd1p and Cit1p), four in biosynthesis (Kes1p, Ade3p, Pro1p, and Rio2p), one (Idp1p) in both processes. Glo3p, which participates in cell trafficking,22 is a GTPase activating protein for Arf1p and Arf2p G-proteins23 belonging to the Ras superfamily.24 Ira2p, a known interactor of Tfs1p,13 is also a GAP for Ras.25,26 One identified partner

Table 2. Proteins Identified by LTP (This Work) versus HTP Proteomics-Based Experiments

protein

LTP experiment PEAKS score (%)

HTP experiments

Prc1p (CPY) Idp3p

34 not found

Collins et al.b, 18 Krogan et al.19

Pet111p

not found

Krogan et al.19

molecular function (biological process)a Proteinase Energy metabolism + Biosynthesis enzyme Translation regulator

a According to the Saccharomyces Genome Database. bIn this case, Prc1p was the bait and Tfs1p was the hit.

methods (Table 3). These known interactors were extracted from the BioGRID relational database (http://www.thebiogrid. org).27 In the course of HTP studies on the yeast interactome by Krogan et al.19 and Collins et al.,18 Pet111p and Idp3p were retrieved as preys19 and Prc1p was shown to interact as a bait with Tfs1p.18 However, these partners were not identified in the experiments of Gavin et al.28 Previously, the Ira2p,13,29 and Prc1p10−12 proteins were identified by non-MS-based methods as physical partners of Tfs1p, while CDC25,13,21 Nab2p,30 GPI1,31 PRP4,31 CDC7,31 VMA9,31 and COX1331 interact with TFS1 at the genetic level. None of these gene products were clearly identified by the low-throughput experiment presented here. Nevertheless, it is worthwhile to notice that Prp4p, Prc1p, and Ira2p were reported by the PEAKS search engine below the 3214



Article

Table 3. Physical and Genetic Interactions Found by Other Methods and Corresponding Score in This LTP Experiment protein/gene

LTP experiment PEAKS score (%)

Prc1p (CPY) Ira2p

Nab2p CDC25 GPI1 PRP4 CDC7 VMA9 COX13 a

34 3

not found not found not found 33 not found not found not found

molecular function (biological process)a

method 10−12

Biochemical activity Two hybrid13 Reconstituted complex13 Co-immunoprecipitation29 Affinity capture-RNA30 Dosage rescue13,21 Negative genetic31 Negative genetic31 Positive genetic31 Positive genetic31 Negative genetic31

Proteinase Ras GAP (Cell signaling)

Nuclear mRNA export Ras GEF (Cell signaling) Biosynthesis enzyme Splicing factor (Translation) Protein kinase (Replication) H+ ATPase (ion transport) Regulatory subunit of energy metabolism enzyme

According to the Saccharomyces Genome Database.

threshold score. The relevance of these differences in identification will be discussed in the following section. Two homologous proteins, that is, Idp1p in the present study and Idp3p in Krogan’s study,19 were retrieved as Tfs1p interactors. Idp1p and Idp3p have a 70% sequence identity. In the case of Idp1p, the corresponding MS/MS spectra were validated by manual interpretation with both search engines (see Supporting Information). Since detailed sequence coverage information for Idp3p identification in the Krogan et al. experiment is not available, it cannot be determined whether the proteins identified in these two studies were indeed different homologues. Co-immunoprecipitations

Figure 1. Western blot detection of proteins co-immunoprecipitated with Tfs1p. WT or Δtfs1 strains were transformed with a plasmid carrying GST-tagged hit proteins (either Glo3p, or Idp1p or Pro1p). After cell lysis, Tfs1p was immunoprecipitated with anti-Tfs1p antibodies incubated with Protein A coated beads. Beads were then eluted with sample buffer. Eluates were subjected to Western blot analysis using anti-GST antibodies. WCL, whole cell lysate; SN, supernatant. Expected molecular masses for GST-tagged hit proteins: Glo3p, 81 kDa; Idp1p, 74 kDa; Pro1p, 73 kDa.

To validate some of the results obtained by MS-based proteomics, co-immunoprecipitation experiments were conducted on four newly identified partners of Tfs1p: Glo3p, Lpd1p, Idp1p, and Pro1p. We chose Glo3p, a GAP of the small GTPase Arf proteins, because it was previously shown that Tfs1p regulates Ira2p, a GAP of the small GTPase Ras.13 We also chose Lpd1p, Idp1p, and Pro1p as all three are involved in intermediate metabolism. Moreover, the GST-tagged version of these four proteins is commercially available. Wild-type or Δtfs1 (negative control) strains were transformed with a plasmid carrying GST-tagged version of our four genes of interest. The native complexes were immunoprecipitated from the different cell lysates using an anti-Tfs1p antibody incubated with Protein A coated beads. Eluates were then analyzed by Western blot using anti-GST antibody. With this method, protein−protein interactions could be recovered in the case of Glo3p, Idp1p, and Pro1p (Figure 1). Interaction with Lpd1p could not be confirmed, possibly due to a competition between Lpd1p and anti-Tfs1p antibody for Tfs1p, or due to steric hindrance caused by the GST tag. Results for Glo3p show two bands in the 75 kDa region for the whole cell lysates lanes. This can be explained by proteolytic cleavage. The upper band corresponds to the GST-tagged intact protein while the lower band probably corresponds to the degradation product, which apparently does not contain a functional Tfs1p binding domain.

newly discovered as well as the previously known partners of Tfs1p were represented as linked to Tfs1p. This link indicates that the identified protein belongs to a complex containing Tfs1p. Second, physical or genetic interactors of the new Tfs1p partners were added to the map on the basis of BioGRID data. However, proteins were left on the map only if they could satisfy both of the following criteria: (i) they are in a complex with Tfs1p or interact with a Tfs1p partner, (ii) the protein is part of a subnetwork with at least one identified Tfs1p partner. Thus, the insertion of Ira2p, a well-established direct physical partner of Tfs1p, leads to the addition of Ras2p, a direct partner of Ira2p, underscoring the involvement of Tfs1p in the Ras signaling pathway. Idh1p and Idh2p were added as genetic interactors as they are known to participate in the tricarboxylic acid (TCA) cycle along with Cit1p, Idp1p, and Ldp1p. On the other hand, Ade3p was not indicated because none of its known genetic or physical interactors could be related to Tfs1p partners. Additional proteins shown by BioGRID data to form subnetworks with partners of Tfs1p such as Lte1p, Kdg1p, and Arf2p were included in the scheme. In summary, the approach used to draw this scheme leads to the building of a network that includes one previously known partner (Ira2p), 8 new Tfs1p partners (one of which, Idp1p, is homologous to a previously identified interactor), and 7 other interacting proteins. The biological significance of these subnetworks is discussed below.

Construction of a Synthetic Map Including Known and Found Tfs1p Partners

To get insight into the biological role of Tfs1p in the cell, we drew a map representing the functional subnetworks drawn from the present experiment combined with a yeast interaction database provided by BioGRID (Figure 2). Subnetworks can shed light on or emphasize possible functions of Tfs1p. The map was constructed using the following rationales. First, the 3215



Article

Figure 2. Tfs1 protein functional network map based on this work and BioGRID data. Proteins co-purified with the TAP-tagged Tfs1p as well as proteins shown by BioGRID (http://www.thebiogrid.org) to physically and genetically interact with them, are included in the map (for inclusion criteria, see Results section). AC-MS, Affinity Capture-MS (TAP Tag or Flag-tag -MS); AC-W, Affinity Capture-Western; BA, Biochemical activity; Co-P, Co-purification; GI, Genetic interaction; PCA, Protein-fragment Complementation assay; RC, Reconstituted complex; Y2H, Two Hybrid.

■

DISCUSSION The goal of the present work is to make a thorough inventory of the functions of Tfs1p in order to help understand its global biological role in the yeast cell. This can potentially also start bringing elements about the role of its orthologs in higher eukaryotic cells, especially in human cells. A review of literature data and databases on Tfs1p partners showed very few physical interactions identified to date. Of the two interactors previously found by molecular biology and biochemistry methods, only Prc1p was found in HTP TAP-MS studies.18 On the other hand, interactions with Idp3p and Pet111p identified in HTP studies19 seemed uncorrelated to any other biological data. This is what prompted us to perform a low-throughput identification of partners with careful, detailed manual validation. This validation led to the identification of 14 new interactors of Tfs1p. Affinity-MS methodologies and identification algorithms have been refined over time in an effort to decrease the false discovery rate, which has a direct impact on the reliability of the result. However, much less attention is usually given to false negatives, although they may include partners that are essential for function. This problem can be compounded in the case of high-throughput analyses where little time can be devoted to the manual validation of results. As stated in the Results section, the identification scores were below the threshold score with the PEAKS search engine for the two known physical interactors of Tfs1p, namely, Ira2p and Prc1p (Table 3). While TAP tag proteomics is a powerful tool for the purification of protein complexes in close to native conditions, the particular properties of some of the interactors can impair identification. The tag itself can generate problems by interfering with binding

to other proteins. In the particular case of Ira2p, the fact that it is a very low-abundance, high molecular mass (350 kDa) membrane-bound protein32 makes identification by affinity-MS even more challenging. Proteins with low and high molecular mass can be difficult to separate in gel electrophoresis and to recover thereafter. Furthermore, both the TAP purification and the gel electrophoresis steps can be unfavorable for the recovery of membrane proteins due to solubility issues during these steps. The Prc1 protein is a proteinase; therefore, the complexes in which this protein is involved could possibly undergo degradation. Here, the lack of recovery of this protein is probably due to its biological function. It is possible that some interesting interactors remained unidentified in this study, and that these could be unearthed after further methodology development. Nevertheless, careful manual validation and integration of the identified partners with known biological data can lead to insights into the biological role of the target protein. As represented in Figure 2, Tfs1p was found to be in physical complexes with Cit1p, Lpd1p, Pro1p, and Idp1p. Cit1p and Lpd1p are energy metabolism enzymes participating in the TCA cycle, Pro1p is involved in biosynthetic metabolism (via glutamate-based proline synthesis), and Idp1p is involved in both intermediate metabolism processes. A secondary network could be built around these proteins with Idh1p, Idh2p, and Kgd1p, which participate in the TCA cycle as well. Tfs1p/ Idp1p and Tfs1p/Pro1p interactions have been confirmed by co-immunoprecipitation, thereby suggesting that Tfs1p is involved in the control of intermediate metabolism. The Glo3p protein has been identified as a Tfs1p partner and the interaction has been confirmed by co-immunoprecipitation. In fact, Glo3p is an Arf GAP protein and it physically interacts 3216



Article

with Arf2p (Figure 2) and Arf1p.23 Arf1p and Arf2p are GTPbinding proteins thought to control the regulation of coated vesicles formation in intracellular trafficking within the Golgi. Thus, Tfs1p is capable of interacting with a new GAP protein in addition to Ira2p, and consequently Tfs1p may be involved in another G-protein-regulated pathway. Since the two GAP proteins do not share any sequence homology, the mode of interaction between the partners might be different. Taken together with known physical interactions, as shown in Figure 2, these observations suggest that Tfs1p is a connecting node between metabolism and cell signaling subnetworks. TFS1 expression is under the control of Msn2p and Msn4p transcription factors (specific of the general stress response)17 which are activated after a stress and are negatively controlled by the Ras/cAMP/PKA pathway.33 We have previously demonstrated that Tfs1p inhibits Ira2p13 and thereby activates the Ras/cAMP/PKA pathway through negative control of Ras by Ira2p. Tfs1p might therefore exert a negative feedback loop on the general stress response. The Ras/cAMP/PKA pathway also regulates glycolysis and glycogen synthesis. By activating it, Tfs1p therefore plays an indirect role on cell metabolism. We have shown in this work that it is also directly involved in other metabolic pathways such as the glutamate-based proline synthetic pathway, as well as the TCA cycle. This work therefore allows us to have a broader view of the multiple roles played by a protein overproduced by the activation of the general stress response, which not only participates in the downregulation of the stress response, but also helps the cell to adapt to stress by acting directly or indirectly on its metabolism. The observation in this LTP study of an interaction between Tfs1p and Yap1p, an oxidative stress-response transcription factor, also points to a role for Tfs1p in adaptation to stress. The Lap4 protein, another new Tfs1p interactor, is an aspartyl aminopeptidase zinc metalloproteinase and belongs to the peptidase M18 family (MH/MC/MF clan). Interestingly, Tfs1p is already known to be an inhibitor of the serine carboxypeptidase Prc1p (S10 peptidase family, SC clan). However, since the two exopeptidases share no sequence homology, Tfs1p might simply recognize convergent homology domains. The murin ortholog of Tfs1p was also found to inhibit thrombin.4 More importantly, it is interesting to note that tumor invasion requires proteolysis through the extracellular matrix, and that hPEBP1, the human homologue of Tfs1p, is a natural metastasis suppressor.9 In the present work, after a detailed, low-throughput study, the yeast protein Tfs1p is shown to potentially act directly on metabolism, and to connect cell signaling and metabolism. Further work will be needed to uncover the molecular mechanism of this activity, and to determine how compelling and general this role may be for proteins of the PEBP family.

■

the blastp program. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: +33-238-255624. Fax: +33-238-63-1517. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by Region Centre, the Ligue contre le Cancer, and by an ATIP grant from the CNRS. We are grateful to Hélène Chautard for the conception of the genomically tagged yeast strain and initiation of this study. We thank Dr. Emmanuelle Bouveret for helpful discussions and the kind gift of the pBS1539 plasmid. We thank Corinne Buré for technical help with preliminary experiments.

■

REFERENCES

(1) Yeung, K.; Seitz, T.; Li, S.; Janosch, P.; McFerran, B.; Kaiser, C.; Fee, F.; Katsanakis, K. D.; Rose, D. W.; Mischak, H.; Sedivy, J. M.; Kolch, W. Suppression of Raf-1 kinase activity and MAP kinase signalling by RKIP. Nature 1999, 401 (6749), 173−177. (2) Yeung, K. C.; Rose, D. W.; Dhillon, A. S.; Yaros, D.; Gustafsson, M.; Chatterjee, D.; McFerran, B.; Wyche, J.; Kolch, W.; Sedivy, J. M. Raf kinase inhibitor protein interacts with NF-kappaB-inducing kinase and TAK1 and inhibits NF-kappaB activation. Mol. Cell. Biol. 2001, 21 (21), 7207−7217. (3) Lorenz, K.; Lohse, M. J.; Quitterer, U. Protein kinase C switches the Raf kinase inhibitor from Raf-1 to GRK-2. Nature 2003, 426 (6966), 574−579. (4) Hengst, U.; Albrecht, H.; Hess, D.; Monard, D. The phosphatidylethanolamine-binding protein is the prototype of a novel family of serine protease inhibitors. J. Biol. Chem. 2001, 276 (1), 535−540. (5) Schoentgen, F.; Jolles, P. From structure to function: possible biological roles of a new widespread protein family binding hydrophobic ligands and displaying a nucleotide binding site. FEBS Lett. 1995, 369 (1), 22−26. (6) Tohdoh, N.; Tojo, S.; Agui, H.; Ojika, K. Sequence homology of rat and human HCNP precursor proteins, bovine phosphatidylethanolamine-binding protein and rat 23-kDa protein associated with the opioid-binding protein. Brain Res. Mol. Brain Res. 1995, 30 (2), 381− 384. (7) Goumon, Y.; Angelone, T.; Schoentgen, F.; Chasserot-Golaz, S.; Almas, B.; Fukami, M. M.; Langley, K.; Welters, I. D.; Tota, B.; Aunis, D.; Metz-Boutigue, M. H. The hippocampal cholinergic neurostimulating peptide, the N-terminal fragment of the secreted phosphatidylethanolamine-binding protein, possesses a new biological activity on cardiac physiology. J. Biol. Chem. 2004, 279 (13), 13054− 13064. (8) Kroslak, T.; Koch, T.; Kahl, E.; Hollt, V. Human phosphatidylethanolamine-binding protein facilitates heterotrimeric G protein-dependent signaling. J. Biol. Chem. 2001, 276 (43), 39772− 39778. (9) Granovsky, A. E.; Rosner, M. R. Raf kinase inhibitory protein: a signal transduction modulator and metastasis suppressor. Cell Res. 2008, 18 (4), 452−457. (10) Bruun, A. W.; Svendsen, I.; Sorensen, S. O.; Kielland-Brandt, M. C.; Winther, J. R. A high-affinity inhibitor of yeast carboxypeptidase Y is encoded by TFS1 and shows homology to a family of lipid binding proteins. Biochemistry 1998, 37 (10), 3351−3357. (11) Mima, J.; Suzuki, H.; Takahashi, M.; Hayashi, R. Overexpression and functional characterization of a serine carboxypeptidase inhibitor

ASSOCIATED CONTENT

S Supporting Information *

Supplementary Table S1, complete list of peptides identified by LC-MS/MS. Supplementary Figure S1, heat shock sensitivity test; Supplementary Figure S2, one-dimensional SDS-PAGE of TAP eluates; Supplementary Figure S3, annotated MS/MS spectra and corresponding ion tables of unique peptides of Idp1p which are not in common with Idp3p; Supplementary Figure S4, sequence alignment between Idp1p and Idp3p using 3217



Article

(I(C)) from Saccharomyces cerevisiae. J. Biochem. 2002, 132 (6), 967−973. (12) Mima, J.; Hayashida, M.; Fujii, T.; Narita, Y.; Hayashi, R.; Ueda, M.; Hata, Y. Structure of the carboxypeptidase Y inhibitor IC in complex with the cognate proteinase reveals a novel mode of the proteinase-protein inhibitor interaction. J. Mol. Biol. 2005, 346 (5), 1323−1334. (13) Chautard, H.; Jacquet, M.; Schoentgen, F.; Bureaud, N.; Bénédetti, H. Tfs1p, a member of the PEBP family, inhibits the Ira2p but not the Ira1p Ras GTPase-activating protein in Saccharomyces cerevisiae. Eukaryotic Cell 2004, 3 (2), 459−470. (14) Gombault, A.; Warringer, J.; Caesar, R.; Godin, F.; Vallée, B.; Doudeau, M.; Chautard, H.; Blomberg, A.; Bénédetti, H. A phenotypic study of TFS1 mutants differentially altered in the inhibition of Ira2p or CPY. FEMS Yeast Res. 2009, 9 (6), 867−874. (15) Créchet, J. B.; Poullet, P.; Camonis, J.; Jacquet, M.; Parmeggiani, A. Different kinetic properties of the two mutants, RAS2Ile152 and RAS2Val19, that suppress the CDC25 requirement in RAS/adenylate cyclase pathway in Saccharomyces cerevisiae. J. Biol. Chem. 1990, 265 (3), 1563−1568. (16) Thevelein, J. M.; de Winde, J. H. Novel sensing mechanisms and targets for the cAMP-protein kinase A pathway in the yeast Saccharomyces cerevisiae. Mol. Microbiol. 1999, 33 (5), 904−918. (17) Hasan, R.; Leroy, C.; Isnard, A. D.; Labarre, J.; Boy-Marcotte, E.; Toledano, M. B. The control of the yeast H2O2 response by the Msn2/4 transcription factors. Mol. Microbiol. 2002, 45 (1), 233−241. (18) Collins, S. R.; Kemmeren, P.; Zhao, X. C.; Greenblatt, J. F.; Spencer, F.; Holstege, F. C.; Weissman, J. S.; Krogan, N. J. Toward a comprehensive atlas of the physical interactome of Saccharomyces cerevisiae. Mol. Cell. Proteomics 2007, 6 (3), 439−450. (19) Krogan, N. J.; Cagney, G.; Yu, H.; Zhong, G.; Guo, X.; Ignatchenko, A.; Li, J.; Pu, S.; Datta, N.; Tikuisis, A. P.; Punna, T.; Peregrín-Alvarez, J. M.; Shales, M.; Zhang, X.; Davey, M.; Robinson, M. D.; Paccanaro, A.; Bray, J. E.; Sheung, A.; Beattie, B.; Richards, D. P.; Canadien, V.; Lalev, A.; Mena, F.; Wong, P.; Starostine, A.; Canete, M. M.; Vlasblom, J.; Wu, S.; Orsi, C.; Collins, S. R.; Chandran, S.; Haw, R.; Rilstone, J. J.; Gandi, K; Thompson, N. J.; Musso, G.; St Onge, P.; Ghanny, S.; Lam, M. H.; Butland, G.; Altaf-Ul, A. M.; Kanaya, S.; Shilatifard, A.; O’Shea, E.; Weissman, J. S.; Ingles, C. J.; Hughes, T. R.; Parkinson, J.; Gerstein, M.; Wodak, S. J.; Emili, A.; Greenblatt, J. F. Global landscape of protein complexes in the yeast Saccharomyces cerevisiae. Nature 2006, 440 (7084), 637−643. (20) Rigaut, G.; Shevchenko, A.; Rutz, B.; Wilm, M.; Mann, M.; Seraphin, B. A. A generic protein purification method for protein complex characterization and proteome exploration. Nat. Biotechnol. 1999, 17 (10), 1030−1032. (21) Robinson, L. C.; Tatchell, K. TFS1: a suppressor of cdc25 mutations in Saccharomyces cerevisiae. Mol. Gen. Genet. 1991, 230 (1− 2), 241−250. (22) Dogic, D.; de Chassey, B.; Pick, E.; Cassel, D.; Lefkir, Y.; Hennecke, S.; Cosson, P.; Letourneur, F. The ADP-ribosylation factor GTPase-activating protein Glo3p is involved in ER retrieval. Eur. J. Cell Biol. 1999, 78 (5), 305−310. (23) Poon, P. P.; Cassel, D.; Spang, A.; Rotman, M.; Pick, E.; Singer, R. A.; Johnston, G. C. Retrograde transport from the yeast Golgi is mediated by two ARF GAP proteins with overlapping function. EMBO J. 1999, 18 (3), 555−564. (24) Zhang, C. J.; Bowzard, J. B.; Anido, A.; Kahn, R. A. Four ARF GAPs in Saccharomyces cerevisiae have both overlapping and distinct functions. Yeast 2003, 20 (4), 315−330. (25) Tanaka, K.; Nakafuku, M.; Satoh, T.; Marshall, M. S.; Gibbs, J. B.; Matsumoto, K.; Kaziro, Y.; Toh-e, A. S. cerevisiae genes IRA1 and IRA2 encode proteins that may be functionally equivalent to mammalian ras GTPase activating protein. Cell 1990, 60 (5), 803− 807. (26) Tanaka, K.; Nakafuku, M.; Tamanoi, F.; Kaziro, Y.; Matsumoto, K.; Toh-e, A. IRA2, a second gene of Saccharomyces cerevisiae that encodes a protein with a domain homologous to mammalian ras GTPase-activating protein. Mol. Cell. Biol. 1990, 10 (8), 4303−4313.

(27) Stark, C.; Breitkreutz, B. J.; Reguly, T.; Boucher, L.; Breitkreutz, A.; Tyers, M. BioGRID: a general repository for interaction datasets. Nucleic Acids Res. 2006, 34, D535−539. (28) Gavin, A. C.; Aloy, P.; Grandi, P.; Krause, R.; Boesche, M.; Marzioch, M.; Rau, C.; Jensen, L. J.; Bastuck, S.; Dümpelfeld, B.; Edelmann, A.; Heurtier, M. A.; Hoffman, V.; Hoefert, C.; Klein, K.; Hudak, M.; Michon, A. M.; Schelder, M.; Schirle, M.; Remor, M.; Rudi, T.; Hooper, S.; Bauer, A.; Bouwmeester, T.; Casari, G.; Drewes, G.; Neubauer, G.; Rick, J. M.; Kuster, B.; Bork, P.; Russell, R. B.; Superti-Furga, G. Proteome survey reveals modularity of the yeast cell machinery. Nature 2006, 440 (7084), 631−636. (29) Gombault, A.; Godin, F.; Sy, D.; Legrand, B.; Chautard, H.; Vallée, B.; Vovelle, F.; Benedetti, H. Molecular basis of the Tfs1/Ira2 interaction: a combined protein engineering and molecular modelling study. J. Mol. Biol. 2007, 374 (3), 604−617. (30) Batisse, J.; Batisse, C.; Budd, A.; Böttcher, B.; Hurt, E. Purification of nuclear poly(A)-binding protein Nab2 reveals association with the yeast transcriptome and a messenger ribonucleoprotein core structure. J. Biol. Chem. 2009, 284 (50), 34911−34917. (31) Costanzo, M.; Baryshnikova, A.; Bellay, J.; Kim, Y.; Spear, E. D.; Sevier, C. S.; Ding, H.; Koh, J. L.; Toufighi, K.; Mostafavi, S.; Prinz, J., St; Onge, R. P.; VanderSluis, B.; Makhnevych, T.; Vizeacoumar, F. J.; Alizadeh, S.; Bahr, S.; Brost, R. L.; Chen, Y.; Cokol, M.; Deshpande, R.; Li, Z.; Lin, Z. Y.; Liang, W.; Marback, M.; Paw, J.; San Luis, B. J.; Shuteriqi, E.; Tong, A. H.; van Dyk, N.; Wallace, I. M.; Whitney, J. A.; Weirauch, M. T.; Zhong, G.; Zhu, H.; Houry, W. A.; Brudno, M.; Ragibizadeh, S.; Papp, B.; Pál, C.; Roth, F. P.; Giaever, G.; Nislow, C.; Troyanskaya, O. G.; Bussey, H.; Bader, G. D.; Gingras, A. C.; Morris, Q. D.; Kim, P. M.; Kaiser, C. A.; Myers, C. L.; Andrews, B. J.; Boone, C. The genetic landscape of a cell. Science 2010, 327 (5964), 425−431. (32) Mitts, M. R.; Bradshaw-Rouse, J.; Heideman, W. Interactions between adenylate cyclase and the yeast GTPase-activating protein IRA1. Mol. Cell. Biol. 1991, 11 (9), 4591−4598. (33) Görner, W.; Durchschlag, E.; Martinez-Pastor, M. T.; Estruch, F.; Ammerer, G.; Hamilton, B.; Ruis, H.; Schüller, C. Nuclear localization of the C2H2 zinc finger protein Msn2p is regulated by stress and protein kinase A activity. Genes Dev. 1998, 12 (4), 586−597.

3218


Interaction Proteomics Suggests a New Role for the Tfs1 Protein in

Recommend Documents