|
Advertisement | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
J. Biol. Chem., Vol. 279, Issue 6, 3956-3979, February 6, 2004
The Yeast Mitochondrial Proteome, a Study of Fermentative and Respiratory Growth*![]() From the Biocenter Oulu and Department of Biochemistry, P. O. Box 3000, University of Oulu, Oulu FIN-90014, Finland
Received for publication, September 12, 2003 , and in revised form, November 3, 2003.
Saccharomyces cerevisiae is able to switch from fermentation to respiration (diauxic shift) with major changes in metabolic activity. This phenomenon has been previously studied on the transcriptional level. Here we present a parallel analysis of the yeast mitochondrial proteome and the corresponding transcriptional activity in cells grown on glucose (fermentation) and glycerol (respiration). A two-dimensional reference gel for this organelle proteome was established (available at www.biochem.oulu.fi/proteomics/), which contains about 800 intense spots. From 459 spots 253 individual proteins were identified, among them low abundant and hydrophobic proteins, and 37 proteins previously deemed hypothetical, with partially unknown cellular localization. After the diauxic shift, mitochondrial levels of only 18 proteins were changed (17 increased, with 1 decreased), among them proteins involved in the tricarboxylic acid cycle (Sdh1p, Sdh2p, and Sdh4p) and the respiratory chain (Cox4p, Cyb2p, and Qcr7p), proteins contributing to other respiratory pathways (Ach1p, Adh2p, Ald4p, Cat2p, Icl2p, and Pdh1p), and two proteins with unknown function (Om45p and Ybr230p). Apart from an overall increase in mitochondrial protein mass, the mitochondrial proteome remains remarkably constant, even in a major metabolic adaptation. This seemingly disagrees with results of the DNA microarray analyses, where a rather heterogenous up- or down-regulation of genes encoding mitochondrial proteins implies large changes in the proteome. We propose that the discrepancy between proteome and transcriptional regulation, apart from different translation efficiency, indicates a changed turnover rate of proteins in different physiological conditions.
Mitochondria have been known for a long time as the organelles responsible for the energy metabolism of eukaryotic cells. More recent studies suggest an additional role of the mitochondrion in apoptosis and aging (16). Likewise, numerous diseases, e.g. Alzheimer's disease, Parkinson's disease, Friedreich ataxia, diabetes mellitus, and malignant tumors (4, 714), were connected to mutations in genes coding for mitochondrial proteins. These findings have motivated increasing efforts to study the putative role of mitochondrial proteins in pathological events (11, 15). The majority of mitochondrial proteins are nuclear encoded and targeted to the organelle by weakly conserved mitochondrial targeting sequences (MTS)1 (1618). The MTS has been used as a marker for the prediction of mitochondrial proteomes from genomic databases. Currently, the expected proteome size of the yeast mitochondrion varies between 400 and 786 proteins (1820). The mitochondrial proteome of higher eukaryotes is significantly larger, e.g. about 1000 proteins for the human mitochondrion (21). The reason for this increased proteome size of higher eukaryotes is not fully understood. A possible explanation offers the presence of isoforms for several mitochondrial enzymes and additional mitochondrial functions (22).
Elucidating the function and regulatory networks of proteins is one of the main challenges in the post-genomic era. Changes in the proteome under different environmental conditions, e.g. nutrient limitation or stress, can provide a starting point for the analysis of function of proteins. Thus far, studies of the mitochondrial proteome by two-dimensional gel electrophoresis have been described only for some higher eukaryotes (11, 2229), but not for yeast. Because its genome is fully sequenced and its cell physiology has been exhaustively investigated, Saccharomyces cerevisiae is one of the most suitable eukaryotic model organism. Due to the fact that most genes coding for mitochondrial proteins are highly conserved among eukaryotes (22, 29, 30), it can serve as a model for mitochondrial physiology at the protein level.
S. cerevisiae can grow either with fermentative or respiratory metabolism. Glucose in abundance is catabolized mainly by fermentation (31). As glucose becomes exhausted, a transient growth arrest occurs. During this arrest, called the diauxic shift, the metabolism is changed from fermentation to respiration and adapted to alternative carbon sources, e.g. ethanol, glycerol, and oleic acid (32). This diauxic shift has been studied extensively at the mRNA level (3237), and changes in transcription of genes involved in the tricarboxylic acid cycle and respiratory chain (37, 38), the gluconeogenesis and glyoxylate cycle (34, 39), as well as peroxisome biogenesis and The aim of the present work was to achieve a more detailed characterization of the yeast mitochondrial proteome to study the diauxic shift at the proteomic level. This allows the investigation of how the transcriptional regulation, reported for this metabolic switch, is reflected at the protein level.
Preparation of MitochondriaFor the analysis of the fermentative growth the S. cerevisiae wild type strain BJ1991 (45) was cultivated on rich YPD (1% (w/v) yeast extract, 2% (w/v) peptone, and 2% (w/v) D-glucose) and collected at an optical density (600 nm) of 1.1. The glucose concentration of the culture media was monitored. For the study of the respiration a wild type culture was grown in YPD to an optical density (600 nm) of 12, pelleted, and transferred to YPG containing 3% (v/v) glycerol instead of D-glucose. After 16 h cells were collected, and mitochondria were prepared as described by Meisinger et al. (46). For two-dimensional gel electrophoresis the mitochondrial pellet was resuspended in urea buffer (8 M urea, 2 M thiourea, 1% (w/v) CHAPS, 20 mM 1,4-dithio-DL-threitol, 0.8% (v/v) carrier ampholytes 310, 100 mM Tris-HCl, pH 7.5, 1 mM EDTA, 14 µM phenylmethylsulfonyl fluoride). The protein concentration was determined with a commercially available kit (RotiNanoquant, C. Roth, Karlsruhe, Germany) according to the method of Bradford (47), and protein aliquots (100 and 500 µg) were stored at 20 °C. For transcriptional analyses the RNA was isolated from collected cells using the hot acidic phenol extraction method (84). The RNA was then further purified using the Qiagen RNeasy kit, following the RNA cleanup protocol provided by the manufacturer. The RNA was visually inspected on ethidium bromide-stained agarose gel to confirm that it was not degraded. Two-dimensional Gel Electrophoresis and Protein Identification The protein solution was adjusted with urea buffer to a final volume of 350 µl. Proteins were then transferred into IPG strips (pH 310, nonlinear, 18 cm, Amersham Biosciences, Uppsala, Sweden) overnight by in-gel rehydration. Isoelectric focusing was carried out at 20 °C with the Multiphor II system (Amersham Biosciences) under paraffin oil for 55 kVh (500 V 500 Vh, 500 V 2500 Vh, 3500 V 10 kVh, 3500 V 42 kVh) (48). Prior to separation in the second dimension the strips were equilibrated according to Görg et al. (49). The electrophoresis was done overnight in polyacrylamide gels (12.5%) with the Ettan DALT II system (Amersham Biosciences) at 12 W per gel and 12 °C. The gels were silver-stained (50) and analyzed with the two-dimensional PAGE image analysis software Melanie 3.0 (GeneBio, Geneva, Switzerland). The apparent isoelectric points (pI) and molecular masses of the proteins were calculated with Melanie 3.0 (GeneBio) using identified proteins with known parameters as a reference. The same software was used for the comparison of fermentative and respiratory protein pattern. An expression change was considered significant if the intensity of the corresponding spot differed reproducibly more than 3-fold. Changes in protein abundance were compared with the transcriptional data of the same samples. Excised spots were digested in-gel and identified from the peptide fingerprints as described elsewhere (85). The identification of a protein was accepted if the peptides (mass tolerance, 20 ppm) covered at least 30% of the complete sequence. A sequence coverage between 30 and 20% or a sequence coverage below 20% for protein fragments was only accepted if at least three main peaks of the mass spectrum matched with the sequence and the number of weak intensity peaks was clearly reduced. Monoclonal antibodies were obtained as a gift from Alexander Tzagoloff (anti-Kgd1p), Benedikt Westermann (anti-Aco1p), and Bernard D. Lemire (anti-Sdh2p).
Microarray Hybridization and Data AnalysisExperimental procedures for GeneChip were performed according to the Affymetrix Gene-Chip Expression Analysis Technical Manual. In essence, using 10 µg of total RNA as template double-stranded DNA was synthesized by means of the Superscript Choice System (Invitrogen, Rockville, MD) and T7-(dT)24 primer. The DNA was purified using GeneChip Sample Cleanup Module (Qiagen, Valencia, CA). In vitro transcription was performed to produce biotin-labeled cRNA using a BioArray HighYield RNA Transcription Labeling Kit (Enzo Diagnostics, Farmingdale, NY) according to the manufacturer's instructions. Biotinylated cRNA was cleaned with an GeneChip Sample Cleanup Module (Qiagen), fragmented to 35200 nucleotides, and hybridized to Affymetrix Yeast genome S98 array, which contains Analysis of Protein Function and SequenceInformation about function, subcellular localization, and specific parameters of the identified proteins were obtained from the databases SWISS-PROT (www.expasy.org/sprot/), MIPS (mips.gsf.de/proj/yeast/CYGD/db/index.html), and SGD (genome-www.stanford.edu/Saccharomyces/). Duplication of genes was analyzed with the Yeast Gene Duplications data base (acer.gen.tcd.ie/~khwolfe/yeast/nova/index.html). The subcellular localization and the putative MTS cleavage site of mitochondrial proteins were predicted by TargetP (www.cbs.dtu.dk/services/TargetP/), PSORT II (psort.nibb.ac.jp/form2.html), and MITO-PROT (www.mips.biochem.mpg.de/cgi-bin/proj/medgen/mitofilter). The expected position (pI and molecular mass) of the protein spot within the two-dimensional gel was calculated with Compute pI/Mw (us.expasy.org/tools/pi_tool.html). Grand average of hydropathicity index (GRAVY) scores were predicted with SOSUI (sosui.proteome.bio.tuat.ac.jp/sosuiframe0.html).
Mitochondria were purified from yeast cultures grown on a sufficient glucose concentration (at the beginning of the culture time 2% glucose and 1.2% glucose at the time of cell collection) to maintain glucose repression (Refs. 33 and 46, see below) and the protein fraction separated by two-dimensional gel electrophoresis. This resulted in about 800 intense spots detectable in a silver-stained two-dimensional gel. A two-dimensional pattern (reference gel) characteristic for fermentative growth, obtained from five mitochondrial preparations, is shown in Fig. 1. The corresponding two-dimensional gels specific for respiratory metabolism were obtained from seven mitochondrial preparations after growth on glycerol. Analyses of the detected proteins by tryptic digestion and mass spectrometry analyses (see "Experimental Procedures" for details) allowed the identification of 253 proteins from 459 spots (Table I).
Subcellular Localization of the Identified ProteinsFirst we investigated whether the identified proteins are known as mitochondrial proteins or belong to different subcellular compartments (Table I). From the literature we found experimental evidence for a mitochondrial localization of 151 identified proteins. We also searched the databases MIPS, SGD, and SWISS-PROT, which do not strictly distinguish between experimental determined or predicted localization. In that way we confirmed the mitochondrial localization for additional eight proteins (Arg8p, Gcv2p, Lys4p, Mef1p, Mrs1p, Qcr10p, Ygl068p, and Ypr004p). For several proteins with no data available we used the bioinformatic tools PSORT II, MITOPROT, and TargetP to predict their localization. To estimate the confidence of these predictions we first processed sequences of known mitochondrial proteins. We found the highest agreement with 76.4% for TargetP (110 of 144 proteins), followed by MITOPROT and PSORT II with 70.6% and 63.9%. TargetP predicted a putative mitochondrial localization for 17 additional proteins (Ape2p, Grx5p, Rhr2p, Thr4p, Yer004p, Yer078p, Yer080p, Yer182p, Ygr086p, Yhr198p, Yil157p, Yjl200p, Ykr065p, Ylr201p, Yor108p, Ypl004p, and Ypl063p), increasing the number of mitochondrial proteins to 176. However, for eight proteins (Hom6p, Pdh1p, Sam2p, Ydr071p, Yjr003p, Yjr085p, Ymr226p, and Ypl222p), no information about their localization was available, neither from databases nor by prediction with TargetP. Our data provide here a preliminary base to address them also as mitochondrial proteins. A number of proteins detected in our mitochondrial fraction have been described so far for other subcellular compartments. Similar results were obtained also for mitochondrial preparations of some higher eukaryotes (27, 29, 53). The majority (46 proteins) of the non-mitochondrial proteins identified here are cytoplasmic. According to data base information many of them are highly abundant and thus likely contaminants (Table I). Interestingly, five of them (Ach1p, Eno1p, Hxk2p, Pgk1p, and Sti1p) have been localized earlier within purified mitochondrial protein complexes (53). Among the nonmitochondrial proteins we also detected two vacuolar proteins (Tfp1p and Vma5p), three peroxisomal proteins (Fat2p, Fox1p, and Fox2p, the latter two only in the glycerol culture), a nuclear protein (Rvb1p), which was found also in other mitochondrial extracts (53), and Faa1p, which had been reported so far to be localized in lipid particles (54). Altogether, we found 54 putative contaminants in the mitochondrial extract. However, a number of the non-mitochondrial proteins identified in the two-dimensional gel are less abundant. This argues against a contamination due to incomplete mitochondrial purification. Instead, they may indicate a functional association of the mitochondrion with other cellular compartments. This conclusion is supported by earlier studies showing specific interaction between mitochondria and ribosomes (1618), endoplasmic reticulum (55) and cytoskeleton (29, 56). The identification of 14 endoplasmic reticulum proteins (Ayr1p, Cdc48p, Cpr5p, Cwh41p, Emp24p, Erg6p, Erg26p, Eug1p, Fpr2p, Gtt1p, Kar2p, Pdi1p, Pst2p, and Rot2p) agreed also with earlier observations (27). This lends support to the proposed interaction between endoplasmic reticulum and mitochondrion, which has been suggested to mediate the phospholipid translocation between these organelles (55). The only cytoskeleton protein detected so far in our two-dimensional gels is actin (Act1p). This could indicate a contamination due to the high abundance of actin, or, alternatively, it might originate from the interaction between cytoskeleton and mitochondrion, which has a function in positioning the organelle within the cell or mediating the transfer of mitochondria into the budding cell (29, 56). Interestingly, several of the proteins identified here have been also found previously within different compartments (Table I). It has been shown that such "multiple localization" can be dependent on different cellular conditions, e.g. cell cycle (57) or carbon source (58). This could explain the presence of these proteins in the mitochondrial fraction and indicates a functional association with this organelle. The subcellular localization of all proteins identified here is summarized in Fig. 2. It presents the classification into mitochondrial, "mitochondrion-associated" (endoplasmic reticulum and cytoskeleton), probable contaminants (cytoplasm, peroxisome, vacuole, nucleus, and lipid particle), and proteins of so far unknown localization.
Multiple Spots, Protein Modification, and MaturationA large number of proteins were detected within multiple spots. This phenomenon has also been described before (25, 27) and indicates post-translational modification or processing. From the 459 spots identified here we found 175 single spots, whereas 195 spots appeared as "train of spots," differing only by their pI. Interestingly, 7 of all identified spots were detected at a higher molecular mass than expected, whereas 82 spots showed a decreased molecular mass. A comparison of expected and detected spot positions within the two-dimensional gel offers the possibility to analyze modification and maturation of the protein. Hence, 18 identified proteins with known N-terminal sequences were selected as marker spots (Abf2p, Atp7p, Atp18p, Cat5p, Cox4p, Cox6p, Cox12p, Cys3p, Hem14p, Kar2p, Mas1p, Mas2p, Mrpl3p, Mrpl17p, Nde1p, Pdb1p, Pim1p, and Rim1p) to determine pI and molecular mass of all identified proteins within the two-dimensional gel (Table I). In addition, we calculated their theoretical positions based on the protein sequence with Compute pI/Mw (Table I). A comparison between theoretical and experimental data allowed new conclusions about post-translational modification, processing, or fragmentation. A pI shift, which was found for a large number of identified spots, might indicate protein modification, e.g. acetylation and phosphorylation. N-terminal acetylation, which lowers the pI, can be excluded for mitochondrial proteins with a cleavable targeting sequence (59). Indeed, the non-mitochondrial proteins in a shifted position might be acetylated as described earlier (59). However, a more detailed characterization of post-translational modifications has to remain subject of future studies. A lowered pI together with a slightly decreased molecular mass indicates cleavage of the mitochondrial targeting sequence (MTS) (23). All mitochondrial proteins identified here are nuclear-encoded. Thus, these proteins are commonly targeted by a specific N-terminal MTS into the organelle (1618). To estimate whether the proteins identified here are processed upon mitochondrial import, we first calculated the expected cleavage site by using the bioinformatic tools TargetP, MITOPROT, and PSORT II. The confidence of such prediction can be estimated by processing known mature protein sequences. TargetP has been reported to have the lowest error rate (23). Our analyses revealed a correct prediction for 26 of 41 known cleavage sites. Thus, we predicted the cleavage sites for all mitochondrial proteins with TargetP and calculated the positions of unprocessed and mature proteins within the two-dimensional gel (Table I). The comparison of predicted and experimental data revealed 87 mitochondrial proteins detectable at the position of their cleaved form, including 47 mature sequences calculated by TargetP. Mcr1p was identified within two spots and might present both the unprocessed and the mature protein. Our data also suggest that 18 mitochondrial proteins are unprocessed. A significant fraction of these are well characterized proteins of the outer membrane (Fis1p, Om45p, Por1p, Por2p, Tom20p, Tom40p, and Tom70p), which could indicate that also the other unprocessed proteins, e.g. Mpm1p, Ykl027p, and Yor251p, are components of the outer membrane. At least Mpm1p has been found in mitochondrial membrane fractions before (60). For 26 gene products the corresponding pI and molecular mass shift is too small to allow discrimination to be made between uncleaved and cleaved protein in our two-dimensional gel. Interestingly, 19 proteins showed a shift indicating cleavage, but not to the position predicted for MTS processing. Moreover, for 12 proteins there is a significant shift of the spot suggesting removal of the MTS, although TargetP does not predict MTS cleavage. For 13 proteins the difference between expected and experimental position cannot be explained by MTS cleavage and might be due to additional protein modifications. Proteins with significantly decreased molecular masses in our two-dimensional reference gel indicate protein fragmentation as described for other mitochondrial preparations (25, 27). However, in contrast to rat liver mitochondria (27) with up to 649 spots for the carbamoyl-phosphate synthase we found a maximum of nine fragments for the ketol-acid reductoisomerase Ilv5p in the yeast mitochondrial proteome. An example for a significantly decreased molecular mass is presented by Arg5p,6p, involved in arginine biosynthesis. ARG5,6 codes for a 94.9-kDa protein, but in the two-dimensional gel it appears in multiple spots concentrated at 37 and 52 kDa. Mass spectrometry analyses of the 52-kDa spots yielded peptides spanning amino acids 90444, whereas the 37-kDa spots corresponded to protein fragments covering amino acids 555859. This confirms earlier studies on Arg5,6p, which is supposed to be transported as precursor into the mitochondrion and processed into Arg5p and Arg6p (51, 61). We suggest, that the 37-kDa spots present Arg6p and the 52-kDa spots Arg5p (Table I). Abundance and HydrophobicityThe Codon Bias Index (CBI) is indicative of the protein abundance, a CBI between 0.6 and 1 corresponds to highly abundant and a CBI near 0 to low abundant proteins (40, 62). Here, we detected 48 proteins with a CBI under 0.1. Among them are 10 proteins with a CBI below 0.01 (Mia1p, Mrp1p, Mrpl17p, Pdh1p, Rsm26p, Yhr198p, Yjr003p, Yjr100p, Ynl026p, and Yor131p) with the lowest CBI (0.085) found for Yjr100p (Table I). Six of these low abundant proteins are mitochondrial, two are cytoplasmic, and for three the localization is still unknown. The GRAVY (grand average of hydropathicity index) score indicates the solubility of the proteins (63). Hydrophobic proteins with a positive GRAVY value are difficult to solubilize and usually not susceptible to two-dimensional gel electrophoresis (64). Based on this value, our proteome analyses comprise 13 hydrophobic proteins (Adh1p, Atp16p, Fpr2p, Hom6p, Hsp10p, Mir1p, Oac1p, Pad1p, Pet9p, Rpp0p, Sdh4p, Yjr085p, and Ypr004p) with positive GRAVY scores up to +0.358 (Table I). These data indicate that this experimental approach does not exclude low abundant and hydrophobic proteins. Protein FunctionAn overview about the functional classification of the 176 identified mitochondrial proteins according to the data base MIPS is shown in Fig. 3. A large number of proteins detected are involved in energy metabolism (13 in tricarboxylic acid cycle, 30 in respiratory chain) but also in metabolism of amino acids (24 proteins), and C-compound and carbohydrate metabolism (45 proteins). Interestingly, we detected here all known compounds of the mitochondrial pyruvate dehydrogenase complex (Lat1p, Ldp1p, Pda1p, Pdb1p, and Pdx1p). Eight identified proteins have functions in the mitochondrial protein import (Hsp60p, Mdj1p, Mge1p, Tim23p, Tim44p, Tom20p, Tom40p, and Tom70p). For 17 identified proteins no function is known at present (Ybr230p, Yer004p, Yer080p, Yer182p, Yfr011p, Ygr086p, Yhl021p, Yhr198p, Yil157p, Yjr100p, Ykl195p, Ykr065p, Ylr201p, Ynl026p, Yor215p, Ypl004p, and Ypl063p).
The Yeast Mitochondrial Proteome after the Diauxic Shift Yeast can switch from fermentative to respiratory growth after glucose exhaustion (diauxic shift), with large changes in metabolic regulation (86) and transcriptional activity (6567). So far little information is available how this adaptation appears at the proteomic level. We forced yeast cells to pass the diauxic shift by transferring the cells from glucose to glycerol (see "Experimental Procedures" for details) and analyzed the mitochondrial proteome of seven different extracts. The protein pattern were compared with the fermentative proteome and proteins with significantly changed abundance were identified (Figs. 4 and 5 and Table II). As expected we found changes for proteins involved in the respiratory metabolism and for proteins with so far unknown function. However, the overall change in the proteome appears surprisingly small when taking into account that the diauxic shift is a dramatic change in the metabolic state and that the mitochondrion is the organelle where the adaptation to the new growth conditions mainly occurs.
To verify that our data were indeed from samples of cells that were glucose-repressed or had undergone the diauxic shift, we also investigated transcriptional changes parallel to changes in the proteome, using the Affymetrix system. Our results indicated that more than 4000 genes showed more than 2-fold changes in expression when comparing the transcriptional data of glucose-versus glycerol-grown cells. The transcriptional data we obtained correlated well with previously published results (33, 35), with only minor deviations, and clearly indicated that the glucose-grown cells were glucose-repressed and that the glycerol-grown samples had undergone the diauxic shift (see data below). We focused on the subset of this data that corresponded to the 253 proteins that we were able to identify on our two-dimensional gels and 89 of them revealed more than 2-fold change (the complete data set is available at www.biochem.oulu.fi/proteomics/). At the proteomic level only 18 proteins had a significantly changed abundance (Table II), whereas the majority of proteins regulated at the transcriptional level did not show a corresponding change in the proteome. Selected marker proteins were also studied by Western blotting of the mitochondrial fraction. Por1p and Aco1p (1.3- and +1.7-fold change in transcription after respiratory growth on glycerol) were selected as controls with constant expression. KGD1 (+3.2-fold) and SDH2 (+9.8-fold) were chosen as significantly induced genes without and with detected change at the proteomic level. In agreement with our earlier observations, only Sdh2p was detected in increased amounts in the mitochondrial proteome (Fig. 6A). It is known that mitochondria have a different morphology and enlarge after the diauxic shift (87). To estimate how much the overall mitochondrial protein mass is increased upon metabolic derepression, we analyzed the amount of Por1p and Aco1p in whole cell extracts (Fig. 6B). The result clearly shows a higher proportion of these mitochondrial markers among cellular proteins, indicating an increase in mitochondrial proteins upon metabolic derepression.
Tricarboxylic Acid Cycle13 of the 15 known, nuclear-encoded proteins of the tricarboxylic acid cycle were identified in our two-dimensional reference gel (Aco1p, Cit1p, Idh1p, Idh2p, Kgd1p, Kgd2p, Lpd1p, Lsc1p, Lsc2p, Mdh1p, Sdh1p, Sdh2p, and Sdh4p). It is also suspected, that Yjl200p, with strong similarity to aconitate hydratases, might also have a function as Aco2p within the tricarboxylic acid cycle (68). At the transcriptional level eight tricarboxylic acid cycle compounds are more than 3-fold up- or down-regulated after glucose exhaustion (YJL200c, KGD1, KGD2, LSC2, SDH1, SDH2, SDH3, and SDH4). The protein level revealed on at least a 3-fold induction after the diauxic shift only for the compounds of the succinate dehydrogenase complex (Sdh1p, Sdh2p, and Sdh4p) (Figs. 5 and 7). Succinate dehydrogenase has a close interaction with the respiratory chain by coupling the oxidation of succinate with the transfer of electrons to ubiquinone (53). This key role for respiration might explain the strong induction after the diauxic shift, when the cellular ATP is mainly produced via the respiratory chain. The fact that the other enzymes of the tricarboxylic acid cycle are not switched off during fermentative growth (Fig. 7) can be explained by their involvement in amino acid biosynthesis, gluconeogenesis, heme biosynthesis or de novo fatty acid synthesis (6870).
Two additional proteins (Adh2p and Ald4p) were also found induced after the diauxic shift (Fig. 5). Adh2p and Ald4p catalyze a two-step conversion of ethanol to acetate under formation of NADH ("acetaldehyde pathway") (53). The induction of Adh2p and Ald4p after the diauxic shift under the current experimental conditions is not unexpected since ADH2 has been shown to be repressed by glucose (71) and indicates a derepression after glucose exhaustion. Ethanol, which is produced by fermentation of glucose and other fermentable sugars, becomes the major carbon source after the supply of these sugars has been exhausted. Thus, Adh2p and Ald4p present an alternative source of NADH, otherwise produced by the tricarboxylic acid cycle. Respiratory ChainThirty proteins identified in our two-dimensional reference gel are known parts of the respiratory chain, contribute to its function (Cyb2p, Mba1p, Mcr1p, and Sco1p) or play a yet unknown role within the respiratory chain (Yor356p and Ypr004p) (Table I). Our studies revealed an at least 3-fold mRNA induction for 13 corresponding genes (ATP3, ATP5, ATP7, COR1, COX4, COX14, CYB2, MCR1, QCR2, SDH1, SDH2, SDH3, and SDH4) (see Fig. 8). As for the tricarboxylic acid cycle proteins, most changes at the transcriptional level were not reflected in the protein pattern (Fig. 8). Only Qcr7p (complex III), Cox4p (complex IV), and Cyb2p were significantly induced during respiration (Figs. 5 and 8). Remarkably, Qcr7p and Cox4p are parts of different enzyme complexes, and their up-regulation is not followed by the other members of the corresponding complexes. Cyb2p (L-lactate cytochrome c oxidoreductase) is responsible for the conversion of lactate to pyruvate in the intermembrane space. Thus, it contributes to the function of the respiratory chain by transferring electrons to cytochrome c (73), which could explain its induction.
2-Methylcitrate CycleRespiratory growth on glycerol causes an induction of Pdh1p and Icl2p (Fig. 5 and Table II). At the transcriptional level both proteins were previously found to be induced by propionate and understood as a part of the 2-methylcitrate pathway (74). Because S. cerevisiae cannot grow on propionate as the sole carbon source, the physiological function of this cycle is not entirely clear (75). Apart from propionate metabolism, it may be involved in the degradation of certain amino acids via propionyl-CoA and also connected to the -oxidation of odd-chain fatty acids (76, 77). The Escherichia coli PrpD, homolog to Pdh1p, has a 2-methylcitrate dehydratase activity and can also use citrate and isocitrate as substrates (78). Thus, Pdh1p might take part in the conversion of 2-methylcitrate to 2-methylisocitrate. ICL2 encodes a 2-methylisocitrate lyase, which catalyzes the consecutive cleavage of 2-methylisocitrate to pyruvate and succinate (76). Both proteins might be responsible for the generation of succinate and pyruvate by metabolizing threonine as an additional source of succinate. This assumption is supported by the fact that the corresponding enzyme within the tricarboxylic acid cycle (succinate dehydrogenase) is also up-regulated (see above).
Peroxisomal
The end products of this Ybr230p and Om45pOm45p and Ybr230p are also induced by growth on glycerol (Fig. 5). The function of both proteins is unknown so far, but the transcriptional induction of the corresponding genes after glucose limitation (Table II) suggests a role in the respiratory metabolism. Om45p has been earlier identified in a purified mitochondrial enzyme complex together with components of the tricarboxylic acid cycle and respiratory chain (53). Likewise Ybr230p has been suggested to be co-regulated with proteins from tricarboxylic acid cycle and respiratory chain (81). On that background our results indicate a connection of both enzymes to tricarboxylic acid cycle and respiratory chain. Om45p, a mitochondrial outer membrane protein, has been suggested to function in protein import (82). Based on our results and the known localization, we propose a role for Om45p in the import of metabolic intermediates or proteins required for respiratory growth. The proteomic study presented here reveals that the yeast mitochondrial proteome is kept in a steady-state even upon large metabolic changes. Thus, the majority of proteins involved in respiratory metabolism is present also during fermentative growth, probably with lower activity or in different functional context. Tricarboxylic acid cycle enzymes for instance are also involved in different synthetic pathways (83). This could explain why the corresponding proteins are present under both conditions. The fact, that only some components of the respiratory chain are significantly up-regulated may indicate key functions in complex assembly or electron transfer under these conditions. Our analyses and previous reports on transcriptional regulation (33, 3537), however, revealed transcriptional changes for a much larger number of genes than found here at the proteome level. Partially, the transcriptional up-regulation can be accounted for by an overall increase in mitochondria upon the diauxic shift. However, the changes in transcriptional activity are heterogenous and do not reflect the situation found for the proteome. First, the discrepancy between transcriptome and proteome can simply reflect differences in translational efficiency for individual proteins. In addition, we assume that this discrepancy reflects differences in protein turnover for individual proteins. The turnover rate is likely to change under different physiological conditions, and the shift from fermentation to respiratory metabolism has probably pronounced effect on the "aging" of individual proteins.
* This work was supported by grants from the Academy of Finland and Sigrid Jusélius Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 The abbreviations used are: MTS, mitochondrial targeting sequence; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate; CBI, Codon bias index; GRAVY, grand average of hydropathicity index.
We thank Eeva-Liisa Stefanius and Marika Kamps for excellent technical assistance and Jussi Vuoristo for the Affymetrix services. We are grateful to Alexander Tzagoloff, Benedikt Westermann, and Bernard D. Lemire for the gift of antibodies.
This article has been cited by other articles:
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Advertisement | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||