Portal de Periódicos da CAPES

Proteome Analysis and Morphological Studies Reveal Multiple Effects of the Immunosuppressive Drug Mycophenolic Acid Specifically Resulting from Guanylic Nucleotide Depletion

2001; Elsevier BV; Volume: 276; Issue: 49 Linguagem: Inglês

10.1074/jbc.m103416200

1083-351X

Mafalda Escobar‐Henriques, Axelle Balguerie, Chistelle Monribot, Hélian Boucherie, Bertrand Daignan‐Fornier,

Renal Transplantation Outcomes and Treatments

Mycophenolic acid (MPA), one of the most promising immunosuppressive drugs recently developed, is a potent inhibitor of IMP dehydrogenase, the first committed step toward GMP synthesis. We found that all the drug effects on yeast cells were prevented by bypassing GMP synthesis, thus confirming the high specificity of MPA. Although the primary target of MPA is clearly identified, we aimed to further understand how GTP depletion leads to growth arrest and developed a new approach based on proteome analysis combined with overexpression studies. Essential proteins down-expressed in the presence of MPA were identified by protein two-dimensional gel analysis and subsequently overexpressed in yeast. Two such proteins, Cdc37p and Sup45p, when overexpressed allowed partial relief of MPA toxicity, strongly suggesting that their lower amount after MPA treatment significantly contributed to the MPA effect. These conserved proteins involved in cell cycle progression and translation are therefore important secondary targets for MPA. Our data establish that MPA effects occur through inhibition of a unique primary target resulting in guanine nucleotides depletion, thereby affecting multiple cellular processes. Mycophenolic acid (MPA), one of the most promising immunosuppressive drugs recently developed, is a potent inhibitor of IMP dehydrogenase, the first committed step toward GMP synthesis. We found that all the drug effects on yeast cells were prevented by bypassing GMP synthesis, thus confirming the high specificity of MPA. Although the primary target of MPA is clearly identified, we aimed to further understand how GTP depletion leads to growth arrest and developed a new approach based on proteome analysis combined with overexpression studies. Essential proteins down-expressed in the presence of MPA were identified by protein two-dimensional gel analysis and subsequently overexpressed in yeast. Two such proteins, Cdc37p and Sup45p, when overexpressed allowed partial relief of MPA toxicity, strongly suggesting that their lower amount after MPA treatment significantly contributed to the MPA effect. These conserved proteins involved in cell cycle progression and translation are therefore important secondary targets for MPA. Our data establish that MPA effects occur through inhibition of a unique primary target resulting in guanine nucleotides depletion, thereby affecting multiple cellular processes. mycophenolic acid inosine monophosphate dehydrogenase synthetic dextrose SD+ casamino acids SD supplemented SD casa + uracil Mycophenolic acid (MPA)1in its morpholinoethyl ester pro-drug form, mycophenolate mofetil (Cellcept, Roche), is one of the most promising immunosuppressive drugs recently developed. It is now widely used to prevent allograft rejection and may be an important alternative to cyclosporin A, which was recently demonstrated to induce tumor development (1Hojo M. Morimoto T. Maluccio M. Asano T. Morimoto K. Lagman M. Shimbo T. Suthanthiran M. Nature. 1999; 397: 530-534Crossref PubMed Scopus (976) Google Scholar). MPA specifically inhibits inosine monophosphate dehydrogenase (IMPDH), the enzyme catalyzing the first committed step in GMP biosynthesis (Fig.1), and consequently depletes the GTP cellular pool severely (down to 10% of normal levels) (2Qiu Y. Fairbanks L.D. Ruckermann K. Hawrlowicz C.M. Richards D.F. Kirschbaum B. Simmonds H.A. Transplantation. 2000; 69: 890-897Crossref PubMed Scopus (42) Google Scholar). Although the primary target of MPA is clearly identified and human IMPDH has been crystallized in the presence of MPA (3Sintchak M.D. Fleming M.A. Futer O. Raybuck S.A. Chambers S.P. Caron P.R. Murcko M.A. Wilson K.P. Cell. 1996; 85: 921-930Abstract Full Text Full Text PDF PubMed Scopus (370) Google Scholar), it is not yet clear how MPA treatment leads to immunosuppression. Two major questions are still unresolved. First, are all of the MPA in vivo effects attributed to decreased synthesis of GMP, and second, what are the consequences on cell physiology of GMP starvation caused by MPA treatment, i.e.which proteins are the secondary targets of MPA? To address these questions, we have used a proteomic approach on yeast as a model eukaryotic system. SD, SD casa, and SD S are as described previously (4Boy-Marcotte E. Tadi D. Perrot M. Boucherie H. Jacquet M. Microbiology. 1996; 142: 459-467Crossref PubMed Scopus (63) Google Scholar, 5Escobar-Henriques M. Daignan-Fornier B. J. Biol. Chem. 2001; 276: 1523-1530Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar). Uracil, leucine, lysine, and guanine at final concentrations of 200 μg/ml, 600 μg/ml, 100 μg/ml, and 0.13 mm, respectively, were added optionally. The wild-type strain Saccharomyces cerevisiae Y350 (MATα;ura3–52;leu2–3,112;lys2Δ201) was used for all experiments except for those presented in Fig. 6, in which the following set of isogenic wild-type and disrupted strains purchased from Euroscarf was used: BY4742 (MATα;his3Δ1;leu2Δ0;lys2Δ0;ura3Δ0),cln1 (Y10785), cln2 (Y11036), cln3(Y10366), mih1 (Y10612), and swe1 (Y11238). Plasmids pC166 (CDC37,URA3, 2 μm), pSL2307 (PSA1,URA3,2 μm), pGS342 (SRP1,URA3,2 μm), pSP3545 (SUP35, SUP45,URA3, 2 μm), and pAFS125 (TUB1-GFP,URA3, int) were kind gifts from M. Winey, G. S. Sprague, G. Schlenstedt, J. P. Rousset, and A. Straight, respectively. Strains in exponential growth phase were resuspended in water to an A 600 = 1. 10-μl drops and four serial 1:10 dilutions of this suspension were spotted on SD casa medium supplemented or not supplemented with uracil and MPA and incubated for 4 days at 30 °C. Yeast wild-type strain Y350 was grown overnight in SD casa U to an A 600 = 0.1. The culture was then divided in halves, and MPA was added to one of them. Four hours later (or 48 h for calcofluor-staining experiments), cells were imaged by microscopy. Conventional epifluorescence microscopy was performed on a Leica DMRXA microscope using a × 100 immersion objective. DNA was visualized by fluorescence microscopy after staining with Hoechst 33342 dye (1 μg/ml, Molecular Probes, Inc.). Cells were harvested, washed with water, and recentrifuged. The pellet was then resuspended in 1 μg/ml Hoechst dye, and cells were incubated for 10 min in the dark before observation. Chitin was visualized by fluorescence microscopy after staining with Calcofluor white dye (1 mg/ml Fluorescent Brightener 28, Sigma) (6Pringle J.R. Adams A.E. Drubin D.G. Haarer B.K. Methods Enzymol. 1991; 194: 565-602Crossref PubMed Scopus (601) Google Scholar). For both DNA and chitin stainings, we used a standard UV filter cube (Leica A). For green fluorescent protein studies, fluorescence was directly visualized in vivo using a standard fluorescein isothiocyanate filter cube (Leica L4). Actin was visualized with rhodamin-phalloidin (7Balguerie A. Sivadon P. Bonneu M. Aigle M. J. Cell Sci. 1999; 112: 2529-2537Crossref PubMed Google Scholar) using a standard TX filter cube (Leica TX). Images were acquired with a cooled CCD camera MicroMax (Princeton Instruments) controlled by the Metamorph 3.0 software (Universal Imaging Corp.). Yeast wild-type strain Y350 was grown overnight in SD S to an A 600 = 0.1. Aliquots of the culture were withdrawn and treated with 0, 0.03, 1, 30, or 100 μg/ml MPA and with or without guanine. Four hours later, flow cytometry analysis (FACS) was performed. For results presented in Fig. 5 b, the same experiment was done in SD casa with the Y350 (wild-type) strain overexpressing CDC37 or the URA3 control plasmid. Cells were harvested, fixed in 70% ethanol, and stained with propidium iodide as described previously (8Paulovich A.G. Hartwell L.H. Cell. 1995; 82: 841-847Abstract Full Text PDF PubMed Scopus (527) Google Scholar). A DAKO flow cytometer was used for FACS analysis. Cell size was acquired with the forward angle light scatter signal. DNA content was acquired with the red fluorescence from the propidium iodide dye detected by a 633-nm filter directed to the FL3 channel. Data from 50,000 cells (acquired at a rate of ∼200 events/second) were analyzed and prepared for presentation using the Flow Mate application. Y350 wild-type strain was grown overnight in SD S to A 600 = 0.5. 2-ml aliquots of the culture were treated with 0, 0.03, 1, 30, or 100 μg/ml MPA for 60 min followed by in vivo protein labeling for 5 min by adding 30 μCi of [35S]methionine (60 Ci/mmol, ICN). Protein sample preparation was carried out by a trichloroacetic acid (5%) precipitation, and total radioactive incorporation was measured. The intensity of each sample was compared with the control culture without MPA, corresponding to 100% translation. Y350 wild-type strain was grown overnight in SD S to A 600 = 0.5. 2-ml aliquots of the culture were treated with 0, 0.03, or 1 μg/ml MPA and with or without guanine. Thirty minutes after the addition of MPA (and guanine), proteins were labeled in vivo for 30 min by adding 100 μCi of [35S]methionine (1000 Ci/mmol, ICN). Protein sample preparation and two-dimensional gel electrophoresis were carried out as described previously (9Boucherie H. Dujardin G. Kermorgant M. Monribot C. Slonimski P. Perrot M. Yeast. 1995; 11: 601-613Crossref PubMed Scopus (74) Google Scholar). Each two-dimensional pattern allows the visualization of 1000 spots corresponding to approximately one-sixth of the yeast proteome. The radioactive gels were exposed to PhosphorImager plates, and images were scanned in a PhosphorImager (Molecular Dynamics). Quantification of spots and comparative analysis were performed with the ImageMaster two-dimensional elite software (Amersham Pharmacia Biotech). The intensity of each spot was normalized to 30 internal standards of different intensities, such as Act1p. Normalization to the actin spot or to the global ratio of all matched spots on the gel gave similar results. At least three independent two-dimensional patterns realized with different35S-labeled protein extracts were analyzed for each condition tested. Only proteins displaying at least 2-fold intensity variations in response to MPA in all three experiments were considered. Proteins affected by the addition of MPA were all identified by matching two-dimensional maps with a reference gel containing more than 400 proteins identified previously (10Perrot M. Sagliocco F. Mini T. Monribot C. Schneider U. Shevchenko A. Mann M. Jeno P. Boucherie H. Electrophoresis. 1999; 20: 2280-2298Crossref PubMed Scopus (118) Google Scholar). Yeast strain Y350 (wild-type) was grown overnight in SD casa U to an A 600 = 0.5, and the culture was then supplemented or not supplemented with 0.03 μg/ml MPA or with guanine. Aliquots were withdrawn at the times indicated in Fig.3. For results presented in Fig. 5, after reaching anA 600 = 0.5, the culture was supplemented or not supplemented with 30 μg/ml MPA and grown for one more hour. RNA was isolated using the Tri-Reagent RNA/DNA/PROTEIN isolation reagent (EUROMEDEX) (5Escobar-Henriques M. Daignan-Fornier B. J. Biol. Chem. 2001; 276: 1523-1530Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar). RNA blots were prepared and probed with labeled polymerase chain reaction fragments (11Denis V. Boucherie H. Monribot C. Daignan-Fornier B. Mol. Microbiol. 1998; 30: 557-566Crossref PubMed Scopus (62) Google Scholar) amplified from the S288C genomic DNA template with the following oligonucleotides: IMP2coli, 5′-TTTTTGCATGCCGCCATTAGAGACTAC-3′; IMP22, 5′- TCCCCCGGGCCCAAATACCGTA-3′; CDC37Bam, 5′-CGCGGATCCATGGCCATTGATTACTC-3′; and CDC37NSI, 5′-CCAATGCAT- CTAGTCAACAGTGTCGG-3′. Growth of a wild-type (Y350) yeast strain was monitored at four concentrations of MPA. Doubling time was 120 min in the absence of MPA and 120, 135, 260, and 310 min at 0.03, 1, 30, and 100 μg/ml of MPA, respectively. The addition of guanine, a GMP precursor, which allows overcoming the lack of IMPDH activity (see Fig.1), completely prevented the growth defect at all concentrations of MPA (Fig.2 a). Although MPA inhibited cell growth, which resulted in large cells with abnormal DNA content (Fig. 2 b), MPA did not severely affect cell viability (data not shown), nor did it arrest cells in a particular state, thus suggesting that there is no checkpoint for guanylic nucleotide pools in yeast. Importantly, these MPA effects were prevented by the addition of guanine (Fig. 2 b). Microscopic observation of the MPA-treated cells confirmed the observed increased cell size and revealed a multinucleate or puffy DNA content (Fig. 2 c, IV). Other obvious defects were noticed such as mislocalization of the nucleus and accompanying microtubule structures that either laid entirely within the emerging bud (Fig. 2 c, VI) or that otherwise laid in the mother cell at the opposite of the bud (Fig.2 c, V). The actin cytoskeleton was also clearly affected. In the presence of MPA, 24% (25/105) of small budded cells showed depolarized actin patches (randomly distributed in the mother cell cortex) and exhibited either misoriented or no actin cables compared with 2% (2/102) in the untreated cells (Fig. 2 d). Finally, MPA-treated cells also displayed a defect in bud-site selection, showing a random bud-site selection instead of the normal axial budding haploid pattern (Fig. 2 e). All these effects were prevented by the addition of guanine (data not shown). Therefore, guanine nucleotides clearly play important roles in cell polarity. To gain insight into the mechanism of action of MPA at the molecular level, we used a global approach based on the study of the yeast proteome. Because GTP plays an important role on translation, we first monitored the effect of MPA on this process.In vivo translation rates at 0.03, 1, 30, and 100 μg/ml of MPA were 104, 78, 49, and 0% of the untreated MPA culture translation rate, respectively. We also performed a kinetic study of MPA effect on the IMD2 gene expression encoding yeast IMPDH (5Escobar-Henriques M. Daignan-Fornier B. J. Biol. Chem. 2001; 276: 1523-1530Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar, 12Shaw R.J. Reines D. Mol. Cell. Biol. 2000; 20: 7427-7437Crossref PubMed Scopus (113) Google Scholar). The effect of MPA and guanine on the IMD2 gene was maximal after a 60-min treatment (Fig. 3). Consequently, the yeast proteome was examined after 60 min of MPA treatment by labeling proteins in vivo with [35S]methionine. Protein patterns without MPA were compared with patterns obtained after 0.03 or 1 μg/ml MPA treatment. Higher MPA concentrations having a drastic effect on translation (as discussed above) could not be used. Three protein spots were more intense after growth in the presence of 1 μg/ml MPA (Table I A). The most strikingly induced protein was Imd3p, one of the three IMPDH homologs in yeast (Fig. 4 a). The two other induced spots corresponded to proteins induced under various stress conditions, such as the heat shock protein Ssa4p (Fig.4 c). The up-regulation of the other two IMPDH homologs, Imd2p and Imd4p, could not have been observed, because they are too basic and migrate outside of the two-dimensional gels.Table IEffect of MPA on the yeast proteomeProteinRatio of spot intensity0.03 μg ml −1 MPA/MPA untreated1-bl and r for left and right spots on the 2D gel, respectively.1 μg ml −1 MPA/MPA untreated1-aFor each protein spot, values are the ratio to the MPA untreated level.Induced Imd3p2581 Ssa4p140 Ynl134cp12Repressed Essential proteins Cct5p1.00.4 Cct8p0.90.5 Cdc37p0.80.4 Psa1p1.10.1 Srp1p0.90.5 Sup45p0.80.3 Non-essential proteins Purine metabolism Ade1p0.80.3 Ade2p1.00.2 Ade4p0.70.5 Ade13p1.60.3 Ade17pl1-bl and r for left and right spots on the 2D gel, respectively.1.00.3 Ade17pr1-bl and r for left and right spots on the 2D gel, respectively.0.70.3 Amino acid metabolism Arg1pl1-bl and r for left and right spots on the 2D gel, respectively.0.80.2 Arg1pr1-bl and r for left and right spots on the 2D gel, respectively.0.80.6 Asn1p1.10.2 Car1p0.70.2 Cys3p1.10.4 Gln1p1.00.2 Gly1pl1-bl and r for left and right spots on the 2D gel, respectively.0.70.3 Gly1pd1-bl and r for left and right spots on the 2D gel, respectively.0.70.3 His1p0.80.3 Sam1pl1-bl and r for left and right spots on the 2D gel, respectively.1.00.4 Sam1pr1-bl and r for left and right spots on the 2D gel, respectively.0.80.5 Sam2p0.90.3 Carbohydrate metabolism Acs2p1.00.4 Adh1pl1-bl and r for left and right spots on the 2D gel, respectively.1.20.3 Ald6pl1-bl and r for left and right spots on the 2D gel, respectively.0.90.3 Ald6pr1-bl and r for left and right spots on the 2D gel, respectively.1.10.3 Gpp1p1.00.5 Hxk2pl1-bl and r for left and right spots on the 2D gel, respectively.0.60.6 Hxk2pr1-bl and r for left and right spots on the 2D gel, respectively.0.70.3 Viral protein CAPSID0.20.3 Others Yel071wp1.30.31-a For each protein spot, values are the ratio to the MPA untreated level.1-b l and r for left and right spots on the 2D gel, respectively. Open table in a new tab A total of 27 identified proteins was found less abundant after MPA treatment (Table I B). Among them are several enzymes of the IMP biosynthesis pathway encoded by the ADE genes (Fig. 4,a and b). We interpret this result as follows: MPA treatment decreases GMP synthesis from IMP and consequently reroutes nonmetabolized IMP to AMP synthesis (see Fig. 1), which in turn results in transcriptional down-regulation of the ADEgenes (13Guetsova M.L. Lecoq K. Daignan-Fornier B. Genetics. 1997; 147: 383-397Crossref PubMed Google Scholar). The other spots corresponded mainly to carbon and amino acids metabolism enzymes (Table I B). All the observed MPA effects were completely prevented by concomitant guanine addition, with the exception of Imd3p and some Ade proteins for which prevention was partial and probably reflected a competition between MPA and guanine during the labeling time lapse. Among the 27 down-regulated proteins, only six are encoded by essential genes, namely Psa1p, Cct5p, Cct8p, Cdc37p, Srp1p, and Sup45p (Fig. 4, c and d). Interestingly, Psa1p is an essential enzyme responsible for GDP-mannose synthesis, which utilizes GTP as a substrate (14Hashimoto H. Sakakibara A. Yamasaki M. Yoda K. J. Biol. Chem. 1997; 272: 16308-16314Abstract Full Text Full Text PDF PubMed Scopus (80) Google Scholar). Cct5p and Cct8p are components of the multiprotein chaperonin-containing T-complex (15Lin P. Sherman F. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 10780-10785Crossref PubMed Scopus (65) Google Scholar). The three other essential proteins Cdc37p, Srp1p, and Sup45p participate to cell cycle progression, nuclear import, and translation termination, respectively (16Stansfield I. Jones K.M. Kushnirov V.V. Dagkesamanskaya A.R. Poznyakovski A.I. Paushkin S.V. Nierras C.R. Cox B.S. Ter-Avanesyan M.D. Tuite M.F. EMBO J. 1995; 14: 4365-4373Crossref PubMed Scopus (427) Google Scholar, 17Loeb J.D. Schlenstedt G. Pellman D. Kornitzer D. Silver P.A. Fink G.R. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7647-7651Crossref PubMed Scopus (140) Google Scholar, 18Gerber M.R. Farrell A. Deshaies R.J. Herskowitz I. Morgan D.O. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 4651-4655Crossref PubMed Scopus (129) Google Scholar). We reasoned that in the presence of MPA down expression of these essential proteins might become limiting for cell growth and lead to the observed growth defect. Therefore, each of the following essential genes, CDC37, PSA1,SRP1, and SUP45 were expressed in yeast on multicopy plasmids. CCT5 and CCT8 were not tested, because they are part of a large complex, and their individual overexpression was therefore less likely to have an effect. For the same reason, SUP45 encoding one of the two subunits of the translation release factor was coexpressed with its interacting partner Sup35p. Overexpression of CDC37 or SUP45+SUP35clearly resulted in a significantly higher resistance to MPA (Fig.5 a), whereas the overexpression of the two other genes had no detectable effect on resistance. Cdc37p and Sup45p proteins are therefore important secondary targets of MPA and clearly contribute to the MPA effect. Because Cdc37p and Sup45p are essential for cell cycle and translation, two very active functions in rapidly proliferating cells, it is tempting to speculate that the observed MPA effect on these conserved proteins could be a general feature among eukaryotes. NeitherCDC37 nor SUP45 overexpression was sufficient to make yeast cells fully resistant to MPA, thus showing that other cellular components are most probably still limiting or unable to function properly. Sup45p (eRF1) together with the GTPase Sup35p (eRF3) forms the translational release factor complex (16Stansfield I. Jones K.M. Kushnirov V.V. Dagkesamanskaya A.R. Poznyakovski A.I. Paushkin S.V. Nierras C.R. Cox B.S. Ter-Avanesyan M.D. Tuite M.F. EMBO J. 1995; 14: 4365-4373Crossref PubMed Scopus (427) Google Scholar, 17Loeb J.D. Schlenstedt G. Pellman D. Kornitzer D. Silver P.A. Fink G.R. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 7647-7651Crossref PubMed Scopus (140) Google Scholar, 18Gerber M.R. Farrell A. Deshaies R.J. Herskowitz I. Morgan D.O. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 4651-4655Crossref PubMed Scopus (129) Google Scholar). In the presence of MPA, the GTPase activity of Sup35p could be affected, and translation termination could be abnormal and affect cell growth. Therefore, our results suggest that correct protein termination, an essential process for optimal growth, should be highly sensitive to GTP depletion. Nevertheless, overexpression of SUP45+SUP35 had no effect on the flow cytometry profiles (data not shown), clearly indicating that several independent processes are affected by MPA. Cdc37p is part of a complex with the Hsp90 chaperone and interacts with several kinases such as Mps1p, necessary for correct spindle duplication, or Cdc28p, necessary for progression through the cell cycle (18Gerber M.R. Farrell A. Deshaies R.J. Herskowitz I. Morgan D.O. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 4651-4655Crossref PubMed Scopus (129) Google Scholar, 19Schutz A.R. Giddings T.H.J. Steiner E. Winey M. J. Cell Biol. 1997; 136: 969-982Crossref PubMed Scopus (57) Google Scholar, 20Kimura Y. Rutherford S.L. Miyata Y. Yahara I. Freeman B.C. Yue L. Morimoto R.I. Lindquist S. Genes Dev. 1997; 11: 1775-1785Crossref PubMed Scopus (179) Google Scholar). For Cdc37p, no linkage is known between its function and guanine nucleotide pools, and its cellular role is not yet fully understood. Cdc37p may play a general role in protein folding or assembly of active kinase complexes and thus act in coordination of events responsible for correct cell cycle progression and spindle duplication. Flow cytometry analysis revealed that overexpression ofCDC37 significantly decreased both DNA content and cell size induced by MPA treatment (Fig. 5 b). Moreover, the cellular defects of the cdc37 mutants correlate quite well with some morphological defects because of MPA treatment (18Gerber M.R. Farrell A. Deshaies R.J. Herskowitz I. Morgan D.O. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 4651-4655Crossref PubMed Scopus (129) Google Scholar, 19Schutz A.R. Giddings T.H.J. Steiner E. Winey M. J. Cell Biol. 1997; 136: 969-982Crossref PubMed Scopus (57) Google Scholar, 20Kimura Y. Rutherford S.L. Miyata Y. Yahara I. Freeman B.C. Yue L. Morimoto R.I. Lindquist S. Genes Dev. 1997; 11: 1775-1785Crossref PubMed Scopus (179) Google Scholar). It is noteworthy that in mammals Cdc37 is highly expressed in proliferative tissues and functions as an oncogene in mice (21Stepanova L. Finegold M. DeMayo F. Schmidt E.V. Harper J.W. Mol. Cell. Biol. 2000; 20: 4462-4473Crossref PubMed Scopus (78) Google Scholar), whereas MPA has a clear anti-proliferative effect on rapidly proliferating cells, such as lymphocytes or tumor cells (22Allison A.C. Eugui E.M. Immunopharmacology. 2000; 47: 85-118Crossref PubMed Scopus (1117) Google Scholar). Because Cdc37p is required for proper assembly of the Cln-Cdc28p cyclin-kinase complexes (18Gerber M.R. Farrell A. Deshaies R.J. Herskowitz I. Morgan D.O. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 4651-4655Crossref PubMed Scopus (129) Google Scholar), knock-out mutants of the three yeastCLN genes were tested for MPA sensitivity. Mutantscln3 and cln1, although to a lesser extent, were more sensitive than the wild-type strain to MPA treatment (Fig.6). This result suggested that indeed by diminishing the guanine nucleotide pool, MPA somehow affected the proper function of the Cln-Cdc28p complexes. Growth of yeast strains mutated in the 6-yeast G2 cyclin (CLB) genes was not affected by MPA (data not shown). We also examined whether the down expression of Cdc37p in the presence of MPA was a consequence of inhibition of transcription. Results presented in Fig. 5 c clearly established that MPA treatment did not affect the amount of CDC37 transcript, whereasIMD2 expression was strongly affected as described previously (5Escobar-Henriques M. Daignan-Fornier B. J. Biol. Chem. 2001; 276: 1523-1530Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar, 12Shaw R.J. Reines D. Mol. Cell. Biol. 2000; 20: 7427-7437Crossref PubMed Scopus (113) Google Scholar). Therefore, MPA affected Cdc37p expression at a posttranscriptional level and would not have been detected by a transcriptome analysis, thus illustrating the complementation of transcription profiling and proteomic analysis. Such a transcriptome approach has been used by others to examine the effect of immunosuppressants cyclosporin A or FK506 on yeast (23Marton M.J. DeRisi J.L. Bennett H.A. Iyer V.R. Meyer M.R. Roberts C.J. Stoughton R. Burchard J. Slade D. Dai H. Bassett D.E.J. Hartwell L.H. Brown P.O. Friend S.H. Nat. Med. 1998; 4: 1293-1301Crossref PubMed Scopus (593) Google Scholar). In comparison to their study (23Marton M.J. DeRisi J.L. Bennett H.A. Iyer V.R. Meyer M.R. Roberts C.J. Stoughton R. Burchard J. Slade D. Dai H. Bassett D.E.J. Hartwell L.H. Brown P.O. Friend S.H. Nat. Med. 1998; 4: 1293-1301Crossref PubMed Scopus (593) Google Scholar), our method not only uses global expression patterns but also combines it with overexpression studies, further validating potential secondary targets. We show for the first time that cell cycle regulation is affected by GTP depletion, since we observed that MPA induced increased DNA content and caused cytoskeleton defects as well as random budding. Several G-proteins are involved in cellular polarization and bud-site selection and could be the primary sensors of GTP depletion. Moreover, the morphological defects caused by MPA are similar to those observed in mutants such as cdc42, affected in bud-site assembly and cell polarity (24Pruyne D. Bretscher A. J. Cell Sci. 2000; 113: 365-375Crossref PubMed Google Scholar), although overexpression ofCDC42 did not result in increased MPA resistance (data not shown). Alternatively, tubulin polymerization proceeds through an active GTP-tubulin form and may therefore be directly affected by the MPA-induced GTP depletion. Because MPA treatment was found to cause actin patches depolarization (Fig. 2 d), the effect of mutations in theSWE1 and MIH1 genes affecting the "morphogenesis checkpoint" was evaluated. Under conditions of transient depolarization of the actin cytoskeleton, the protein kinase Swe1p phosphorylates Cdc28p, inhibiting cell cycle progression through G2. This action of Swe1p is opposed by the protein phosphatase Mih1p. The mih1 mutant was found more resistant to MPA than the wild-type strain, whereas on the opposite, the swe1 mutant showed increased sensitivity (Fig. 6). Therefore, the mitotic G2/M delay controlled by the morphogenesis checkpoint appears as an important step to recover from the effects of MPA. We establish that in yeast all MPA effects are consequences of IMPDH inhibition. In the case of decreased synthesis of IMP biosynthesis enzymes (ADE genes products), the MPA effect is most probably attributed to the accumulation of IMP, the IMPDH substrate. In all other cases the MPA effect is clearly because of decreased synthesis of guanine nucleotides, since it can be bypassed by extracellular guanine. Therefore, the critical effects of MPA must occur downstream of IMPDH (primary target) through the effect of guanine nucleotide deficiency on downstream targets (secondary targets). The proteome approach combined with overexpression studies allowed us to identify two essential proteins, Cdc37p and Sup45p, the amount of which is limiting in MPA-treated cells. The identification of such secondary targets, which are the cellular Achilles heels for a specific drug, is an important task to discriminate the therapeutic effects from the undesirable side effects. Proteome analysis combined with genetics is clearly a powerful tool to identify such important targets. We are grateful to Drs. Winey, Sprague, Schlenstedt, Rousset, and Straight for sending plasmids. We thank Drs. Pellman, Javerzat, and Breton for suggestions on the manuscript.