Báo cáo khoa học: Amino acid biosynthesis and metabolic flux profiling of Pichia pastoris ppt

Amino acid biosynthesis and metabolic flux profiling of Pichia pastoris Aina Sola ` 1,2 , Hannu Maaheimo 2,3 , Katri Ylo¨ nen 3 , Pau Ferrer 1 and Thomas Szyperski 2 1 Department of Chemical Engineering, Escola Te ` cnica Superior d’Enginyeria (E.T.S.E), Universitat Auto ` noma de Barcelona, Bellaterra, Spain; 2 Department of Chemistry, University at Buffalo, The State University of New York at Buffalo, NY, USA; 3 NMR-laboratory and Structural Biology and Biophysics Program, VTT Biotechnology, Helsinki, Finland Amino acid biosynthesis and central carbon metabolism of Pichia pastoris were studied using biosynthetically directed fractional 13 C labeling. Cells were grown aerobically in a chemostat culture fed at two dilution rates (0.05 h )1 , 0.16 h )1 ) with glycerol as the sole carbon source. For investigation of amino acid biosynthesis and comparison with glycerol cultivations, cells were also grown at 0.16 h )1 on glucose. Our results show that, firstly, amino acids are synthesized as in Saccharomyces cerevisiae. Secondly, bio- synthesis of mitochondrial pyruvate via the malic enzyme is not registered for any of the three cultivations. Thirdly, transfer of oxaloacetate across the mitochondrial membrane appears bidirectional, with a smaller fraction of cytosolic oxaloacetate stemming from the mitochondrial pool at the higher dilution rate of 0.16 h )1 (for glucose or glycerol cul- tivation) when compared to the glycerol cultivation at 0.05 h )1 . Fourthly, the fraction of anaplerotic synthesis of oxaloacetate increases from 33% to 48% when increasing the dilution rate for glycerol supply, while 38% is detected when glucose is fed. Finally, the cultivation on glucose also allowed qualitative comparison with the flux ratio profile previously published for Pichia stipitis and S. cerevisiae grown on glucose in a chemostat culture at a dilution rate of 0.1 h )1 . This provided a first indication that regulation of central carbon metabolism in P. pastoris and S. cerevisiae might be more similar to each other than to P. stipitis. Keywords: 13 C NMR; central metabolism; flux profiling; metabolic engineering; Pichia pastoris. The methylotrophic yeast Pichia pastoris has emerged as an important host for heterologous protein expression in both biomedical research and industrial biotechnology [1,2]. As a eukaryotic organism, P. pastoris offers well known advan- tages for protein expression, such as facilitated proteolytic processing as well as efficient protein folding, disulfide bond formation and glycosylation. Very recently, the capability of P. pastoris to produce complex human glycoproteins has been demonstrated in a paradigmatic study [3], indicating that this organism might well become the premier choice for future expression of human proteins for medical applica- tions. Specific advantages of P. pastoris are due to (a) the availability of an unusually tightly regulated promoter from the methanol-regulated alcohol oxidase I gene (AOX1)(b) the system’s efficient protein secretion which, combined with the very low secretion levels of endogenous proteins, is a major advantage for their purification, and (c) P. pastoris’ preference to grow in a respiratory mode, which tends to reduce the excretion of fermentation byproducts such as ethanol or acetic acid and allows one to reach exceptionally high cell densities. Furthermore, P. pastoris expresses pro- teins at high levels when grown on a minimal medium, which also makes this yeast strain attractive for the production of stable isotope labeled proteins [4,5] for NMR-based struc- tural biology and structural genomics [6]. The level of protein expression in P. pastoris depends critically on the growth conditions, and the attainment of high cell densities has been shown to improve protein yields substantially [7]. Typically, aerobic growth is achieved in two phases. First, the cells are grown in a batch culture with glycerol. Subsequently, a mixture of glycerol and methanol Correspondence to P. Ferrer, Department of Chemical Engineering, Escola Te ` cnica Superior d’Enginyeria (E.T.S.E), Universitat Auto ` noma de Barcelona, 08193-Bellaterra, Spain. Fax: +34 935 812013, Tel.: +34 935 812141, E-mail: pau.ferrer@uab.es and Thomas Szyperski, Department of Chemistry, University at Buffalo, The State University of New York at Buffalo, NY 14260, USA. Fax: + 1 716 6457338, Tel.: + 1 716 6456800 ext. 2245, E-mail: szypersk@chem.buffalo.edu Abbreviations:[ 13 C, 1 H]-COSY, [ 13 C, 1 H] correlation NMR spectroscopy; BDF, biosynthetically directed fractional; METAFoR, metabolic flux ratio; ROL, Rhizopus oryzae lipase; D,dilutionrate;l max , maximum specific growth rate; cyt, cytosolic; mt, mitochondrial; TCA, tricarboxylic acid; PEP, phosphoenolpyruvate; PPP, pentose phosphate pathway; PYR, pyruvate; OAA, oxaloacetate; SHMT, serine hydroxymethyl- transferase; GCV, glycine cleavage pathway; ICL, isocitrate lyase; MS, malate synthase. Enzymes: alanine:glyoxylate amino transferase (EC 2.6.1.44), glycine cleavage system (EC 2.1.2.10), glycerol kinase (EC 2.7.1.30), FAD-dependent glycerol-3-phosphate dehydrogenase (EC 1.1.99.5), isocitrate lyase (EC 4.1.3.1), lipase [triacylglycerol acylhydrolase] (EC 3.1.1.3), malic enzyme (EC 1.1.1.39/1.1.40), malate synthase (EC 2.3.3.9), pyruvate carboxylase (EC 6.4.1.1), phosphoenolpyruvate carboxykinase (EC 4.1.1.49), serine hydroxymethyltransferase (EC 2.1.2.1), succinate dehydrogenase (EC 1.1.1.42), threonine aldo- lase (EC 4.1.2.5), transaldolase (EC 2.2.1.2), transketolase (EC 2.2.1.1). (Received 19 December 2003, revised 23 March 2004, accepted 20 April 2004) Eur. J. Biochem. 271, 2462–2470 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04176.x is fed to the culture in a fed-batch mode. During this second growth phase, the production of the recombinant protein is induced by methanol, which activates the AOX1 promoter controlling the heterologous gene. Notably, adaptation to growth on methanol leads to the induction of several key enzymes, e.g. alcohol oxidase, catalase, formaldehyde dehydrogenase and dihydroxyacetone synthase, as well as peroxisome biosynthesis [8]. Although methanol assimil- ation is rather strongly repressed by multicarbon sources such as glucose and glycerol, coassimilation of a multicar- bon source and methanol can be triggered at certain growth conditions [9]. In turn, this allows one to generate a significant fraction of the heterologous protein from the cheap C1 source methanol. Although the central metabolic bioreaction network is quite similar for all yeast strains, important variations exist with respect to its regulation [10,11]. It is, for example, well documented that during aerobic growth of Saccharomyces cerevisiae, catabolism of glucose and related sugars causes a strong impairment in respiratory capacity (the so called ÔCrabtreeÕ-effect). In contrast, most non-Saccharomyces yeasts grow under aerobic conditions in a respiratory mode, that is, reduction equivalents are used to reduce oxygen to water. In fact, variations in the regulation of central carbon metabolism in non-S. cerevisiae genera are essentially unexplored. Moreover, even more comprehensive investi- gations of amino acid metabolism have so far been pursued only for S. cerevisiae (for example [12]). In view of the outstanding role of P. pastoris for biotechnology research, this organism represents an obvious target for studies of its metabolism. Stable isotope labeling experiments employed in con- junction with NMR spectroscopy and/or mass spectrometry [13] are a powerful tool for metabolic studies. In particular, biosynthetically directed fractional (BDF) 13 C labeling of proteinogenic amino acids has been developed into a cost- effective approach to assess the topology of active bioreac- tions (i.e. the active pathways) and to quantify metabolic flux ratios [14]. BDF labeling has been applied to study central carbon metabolism of eubacteria [14–16] as well as eukaryotic yeast cells [17,18]. Moreover, when feeding glycerol as the sole carbon source, such labeling enabled one to explore amino acid biosynthesis pathways in the halo- philic archaeon Haloarcula hispanica [19]. In this publication, we employ BDF 13 C-labeling to elucidate the central carbon metabolism of P. pastoris cells growing under different dilution rates in chemostat cultures. Specifically, this study focuses on characterizing P. pastoris cells growing on glycerol. For comprehensive investigation of amino acid biosynthesis and comparison with flux ratios obtained for growth on glycerol, P. pastoris was also grown with glucose as the sole carbon source. Materials and methods Strain and media A prototrophic P. pastoris strain expressing a heterologous protein – a Rhizopus oryzae lipase (ROL) – under the transcriptional control of the AOX1 promoter has been chosen for metabolic flux ratio profiling. Pichia pastoris x-33/pPICZaA-ROL [20] is the wild type x-33 strain (Invitrogen, Carlsbad, CA, USA) with the pPICZaA-derived expression vector (Invitrogen) containing the ROL gene, pPICZaA-ROL, integrated in its AOX1 locus. Chemostat cultures were fed with a defined minimal medium containing per litre of deionized water: yeast nitrogen base (YNB; Difco, Detroit, MI, USA), 0.17 g; (NH 4 ) 2 SO 4 , 5 g; glycerol or glucose, 10 g; Antifoam Mazu DF7960 (Mazer Chem- icals, PPG Industries, Gurnee, IL, USA), 0.1 mL. The YNB components were sterilized separately by microfiltration and then added to the bioreactor. The medium used for starter cultures was YPD medium containing 1% (w/v) yeast extract, 2% (w/v) peptone, 2% (w/v) dextrose. Chemostat cultivations Continuous cultivations were carried out at a working volume of 0.8 L in a 1.5 L bench-top bioreactor (BiofloIII, New Brunswick, NJ, USA) at 30 °C, and a minimum of 30% dissolved oxygen tension. Cultivations using glycerol as the sole carbon source were performed at two different dilution rates, D (defined as volumetric flow rate/working volume) of 0.05 h )1 and 0.16 h )1 . The cultivation using glucose as sole carbon source was performed at D ¼ 0.16 h )1 . The maximum specific growth rates, l max , of P. pastoris on excess glycerol or glucose are virtually identical (0.17 and 0.18 h )1 , respectively). Hence, at D ¼ 0.16 h )1 cells are growing at around 90% of l max , for both the glycerol and the glucose cultivation. This ensures comparability of metabolic flux ratios. Medium feeding was controlled by a Masterflex pump (Cole-Parmer, Vernon Hills, IL, USA). The working volume was kept constant by removal of effluent from the center of the culture volume by use of a peristaltic pump (B. Braun Biotech Int., Melsungen, Germany). The pH of the culture was maintained at 5.5 by addition of 1 M KOH, and the airflow was maintained at 1 LÆmin )1 with filter-sterilized air using a mass flow controller (Brooks Instruments B.V., Veenendaal, the Netherlands). The agitation speed was set to 500 r.p.m. Starter cultures (100 mL) were grown in 1 L baffled shake flasks at 200 r.p.m., 30 °C for 24 h. Cells were harvested by centrifugation (4000 g,10mins)andresus- pended in fresh medium prior to the inoculation of the bioreactor. Analytical procedures Cell biomass was monitored by measuring the attenuance at 600 nm. For cellular dry weight, a known volume of cultivation broth was filtered using preweighted filters; these were washed with two volumes of distilled water and dried to constant weight. Samples for extracellular metabolite analyses were centrifuged at 12 000 g for 2 min in a microcentrifuge to remove the cells. Glycerol, glucose, organic acids and ethanol were analyzed by HPLC (Hewlett Packard 1050, Wilmington, DE, USA) analysis using an ionic exchange column, Aminex HPX-87H (Bio-Rad, Hercules, CA, USA). The mobile phase was 15 m M sulfuric acid. The injection volume was 20 lL. Data was quantified by the MILLENIUM 2.15.10 software (Waters, Milford, MA, USA). The exhaust gas of the bioreactor was cooled in a condenser at 2–4 °C (Frigomix R; B. Braun Biotech Int., Melsungen, Germany) and dried through a silica gel Ó FEBS 2004 Metabolic flux profiling of Pichia pastoris (Eur. J. Biochem. 271) 2463 column. Concentrations of oxygen and carbon dioxide in the exhaust gas of bioreactor cultivations were determined on line with a mass spectrometer (Omnistar; Balzers Instruments, Liechtenstein). Biosynthetically directed fractional (BDF) 13 C-labeling P. pastoris cells were fed with a minimal medium containing either 10 gÆL )1 glycerol or glucose for five volume changes to reach a metabolic steady state, as is indicated by a constant cell density and constant oxygen and carbon dioxide concentrations in the bioreactor exhaust gas. BDF 13 C- labeling was achieved as described [16,18], that is, by feeding the medium containing about 10% (w/w) of uniformly 13 C- labeled and 90% (w/w) unlabeled substrate for one volume change. Uniformly 13 C-labeled glycerol and glucose (isotopic enrichment of > 98%) were purchased from Martek Biosciences (Columbia, MD, USA) and Spectra Stable Isotopes (Columbia, MD, USA), respectively. Cells were then harvested by centrifugation at 4000 g for 10 min, resuspended in 20 m M Tris/HCl (pH 7.6) and centrifuged again. Finally, the washed cell pellets were lyophilized (Benchtop 5L Virtis Sentry, Virtis Co., Gardiner, NY, USA), of which 200 mg were resuspended in 3 mL of 20 m M Tris/HCl (pH 7.6). After addition of 6 mL of 6 M HCl, the biomass was hydrolyzed in sealed glass tubes at 110 °Cfor 24 h, the solutions were filtered using 0.2 lm filters (Millex- GP, Millipore, Bedford, MA, USA) and lyophilized. NMR spectroscopy and data analysis The lyophilized hydrolysates were dissolved in 0.1 M DCl in D 2 O, and 2D [ 13 C, 1 H]-COSY spectra were acquired for both aliphatic and aromatic resonances as described [14] at 40 °C on a Varian Inova spectrometer (Varian, Inc., Palo Alto, CA, USA) operating at a 1 H resonance frequency of 600 MHz. The spectra were processed using the program PROSA [21] or standard Varian spectrometer software VNMR (version 6.1, C). The program FCAL 2.3.1 [22] was used for the integration of 13 C- 13 C scalar fine structures in 2D [ 13 C, 1 H]-COSY, for the calculation of relative abundances (f-values) of intact carbon fragments arising from a single carbon source molecule [14], and for the calculation of the resulting flux ratios through several key pathways in central metabolism [14,17]. As described previously [13–19], the calculation of the flux ratios when using fractional 13 C-labeling of amino acids is based on assuming both a metabolic (see above) and an isotopomeric steady state. To establish an affordable protocol for 13 C-labeling, it has been proposed to feed a chemostat, which is operating in metabolic steady state, for the duration of one volume change with the medium containing the 13 C-labeled substrate [16,18] before harvest- ing the biomass. Then, the fraction of unlabeled biomass produced prior to the start of the supply with 13 C-labeled medium can be calculated following simple wash-out kinetics [16,18,23–25]. When chemostats are fed in a Ôcarbon limitedÕ manner, one usually finds that (a) the steady-state concentration of labeled carbon source in the bioreactor is small (or even zero), and (b) only small metabolic byprod- ucts, which could be potentially re-imported into the cell after some time, are synthesized. Both conditions are met for the chemostat cultivations of the present study. Hence, after the supply was switched from unlabeled to labeled substrate, only smaller intracellular pools of metabolites had to reach isotopic steady state, which is usually attained for yeast within about half an hour or less [25]. (Notably, bacterial cells reach isotopic steady state within a few minutes or even faster [26], except when very large intracellular pools are present in high-yield overproducing strains [24,27–29].) At D ¼ 0.16 h )1 (0.05 h )1 ) one volume change requires 6.25 (20) hours, and we thus find that only a small fraction of the labeled biomass is generated while intracellular metabolism is far from isotopic equilibrium. In fact, van Winden et al. [25] determined in their experimental setup that the slowest Ôwash-inÕ rate in yeast cells growing in achemostatwas0.63· 10 )3 s )1 .EvenatD ¼ 0.16 h )1 this is much faster than the turnover rate of biomass (4.4 · 10 )5 h )1 ). Hence, the errors of flux ratios resulting from biomass that was fractionally labeled under isotopic nonsteady state conditions were neglected. In agreement with this proposition, we do not observe inconsistencies for labeling patterns that serve to validate the bioreaction network (Fig. 1). Such inconsistencies are expected if isotopomeric steady state was not reached before the majority of fractional 13 C labeled biomass was generated, as the pools of different metabolite approach isotopic steady state at different rates [25]. A special technical comment relates to the fact that the formalism developed for calculating flux ratios for cells that were grown on glucose is also readily applicable for experiments where glycerol is fed. This is because the catabolic breakdown of uniformly labeled [ 13 C]glycerol does not generate isotopomers which differ from those created when glucose is fed. A more detailed discussion can be found in a recent article describing a study in which fractional labeling of biomass was achieved with glycerol as the sole carbon source [19]. Biochemical reaction network model for P. pastoris Because the central carbon metabolism of P. pastoris has so far not been comprehensively characterized, the biochemical reaction network model taken for data interpretation was the one recently identified for S. cerevisiae [17,18], which has also been shown to be suitable for Pichia stipitis [18]. Following consideration of published data [10,30], only the pathway for glycerol metabolism was added (Fig. 1). This involves glycerol phosphorylation by a cytosolic glycerol kinase to 3-phosphoglycerol which is subsequently oxidized by a mitochondrial (membrane) FAD-dependent glycerol phosphate ubiquitone oxidoreductase in order to yield dihydroxyacetone phosphate. The thus generated dihyd- roxyacetone phosphate serves for both pyruvate synthesis and gluconeogenesis [10]. In principle, the glyoxylate cycle also had to be included (and is thus indicated in grey in Fig. 1). The two characteristic reactions of this cycle are catalyzed by malate synthase (MS) and isocitrate lyase (ICL), which are subject to catabolite repression in S. cere- visiae [31]. Notably, repression occurs to a lesser extent when S. cerevisiae grows on a medium containing a nonfermentable carbon source such as glycerol as the sole carbon source [32,33]. Although ICL and MS are most probably cytosolic in S. cerevisiae [10], they are assumed to 2464 A. Sola ` et al.(Eur. J. Biochem. 271) Ó FEBS 2004 be located in peroxisomes in methylotrophic yeasts such as P. pastoris [34]. However, the 13 C-labeling pattern arising from the action of the glyoxylate cycle and the efflux of oxaloacetate (OAA) from the mitochondria cannot be distinguished [18]. Hence, the exchange of OAA between the cytosol and mitochondria was likewise considered to be bidirectional, as discussed for yeast cells growing in glucose-limited chemostat cultivations [18]. (Note, that Fig. 1. Network of active biochemical pathways constructed for P. pastoris cells grown with either glycerol or glucose as the sole carbon source. The network is based on the networks recently identified for S. cerevisiae [17] and P. stipitis [18] and on the literature on P. pastoris metabolism [10,30] (see text). The central carbon metabolism of P. pastoris is dissected into cytosolic and mitochondrial subnetworks. In addition, the glyoxylate cycle reactions are supposed to reside in peroxisomes in methylotrophic yeast like P. pastoris. Because the reactions of the glyoxylate cannot be identified with the currently employed 13 C-labeling strategy (see text), its reactions are depicted in grey. The amino acids and the carbon fragments originating from a single intermediate of the central carbon metabolism are represented in rectangular boxes. Thin lines between the amino acid carbon atoms denote carbon bonds that are formed between fragments originating from different precursor molecules, while thick lines indicate the intact carbon connectivities in the fragments arising from a single precursor molecule. The carbon skeletons of the intermediates of the glycolysis, TCA cycle and pentose phosphate pathway are represented by circles, squares and triangles, respectively. The numbering of the carbon atoms refers to the corresponding atoms in the precursor molecule. Abbreviations: AcCoA, Acetyl-Coenzyme A; DHAP, dihydroxyacetone phosphate; E4P, erythrose 4-phosphate; F6P, fructose 6-phosphate; Fum, fumarate; G6P, glucose 6-phosphate; Glc, glucose; Glyox, glyoxylate; G3P, glyceralde- hyde 3-phosphate; 3PG, 3-phosphoglycerate; Mae, malic enzyme; Mal, malate; OAA, oxaloacetate; 2Og, 2-oxoglutarate; PYR, pyruvate; PEP, phosphoenolpyruvate; S7P, seduheptulose-7-phosphate; Ser, serine; Succ, succinate. For AcCoA, Fum, OAA, PYR and Succ cytosolic (cyt) and mitochondrial (mt) pools are indicated separately. Ó FEBS 2004 Metabolic flux profiling of Pichia pastoris (Eur. J. Biochem. 271) 2465 exchange of C4 intermediates in cytosol and mitochondria may occur via either shuttle transport mechanisms of the tricarboxylic acid (TCA) cycle intermediates, e.g. succinate- fumarate shuttle [35,36], or mitochondrial redox shuttles, e.g. malate-oxaloacetate shuttle, malate-aspartate shuttle and malate-pyruvate shuttle [37]). Results and discussion P. pastoris cultivations in aerobic chemostats using glycerol as the sole carbon source were performed at dilution rates of D ¼ 0.16 h )1 , which is slightly below the maximum specific growth rate of the organism previously observed in a batch culture on glycerol (0.17 h )1 ,A.Sola ` , unpublished results), and D ¼ 0.05 h )1 , where the glycerol supply is growth- limiting (Table 1). Consistently, some residual glycerol (3.0 gÆL )1 ) was detected in the chemostat operating at D ¼ 0.16 h )1 , while no glycerol was found at D ¼ 0.05 h )1 . Biomass yield per gram of glycerol was 0.63 g (cell dry weight) in both cases. An aerobic chemostat cultivation using glucose as sole carbon source was performed at D ¼ 0.16 h )1 (Table 1). In this case, the biomass yield was 0.57 g (cell dry weight) per gram of glucose, slightly lower than for cells grown on glycerol. Moreover, cells growing on glucose exhibited higher CO 2 production rates compared with cells growing on glycerol at the same growth rate. The residual glucose concentration in the corresponding culti- vation was 0.5 gÆL )1 , indicating that the cells were likewise growing close to the maximum specific growth rate (which was 0.18 h )1 for P. pastoris grown in a batch culture with excess glucose). Notably, ethanol, acetate, succinate and pyruvate were not detected by HPLC in any of the cultivations, and carbon balances closed within 5%. Hence, P. pastoris used both glycerol and glucose entirely to generate biomass and CO 2 . This supports the notion that P. pastoris cells grow exclusively in a respiratory manner and are thus efficient biomass and protein producers. The metabolic flux ratio analyses were performed with hydrolyzed biomass samples that were harvested from these chemostat cultures in physiological steady-state. 2D [ 13 C, 1 H]-COSY data were analyzed as described [17], yielding the desired relative abundances (f-values) of intact carbon fragments arising from a single source molecule of glycerol or glucose (Table 2). Biosynthesis of proteinogenic amino acids and C1 metabolism in P. pastoris As expected, the f-values obtained for the glucose and glycerol cultivations (Table 2) show that the proteinogenic amino acids are primarily synthesized in P. pastoris accord- ing to the pathways documented for S. cerevisiae [12,17,38,39]. In particular, the data confirm that (a) Lys synthesis occurs primarily via the a-aminoadipate pathway, (b) Ser is (primarily) synthesized from 3-phosphoglycerate, and (c) the pool of Ser molecules is affected by reversible cleavage by serine hydroxymethyltransferase (SHMT; about 40% were cleaved in all three cultivations). For Gly synthesis, yeasts can cleave either Ser (via SHMT) or Thr (via threonine aldolase). Due to near degeneracy of f-values, however, it is not possible to accurately determine the relative contribution of the two pathways, or to distinguish between cytosolic and mitochondrial SHMT activity. Nonetheless, the data prove that the SHMT pathway is active. This is consistent with recent 13 C-labeling experi- ments with S. cerevisiae cells growing on glucose batch cultures, where the threonine aldolase pathway accounts for only about one-third of the Gly biosynthesis, leaving the SHMT pathway as the major route to Gly in these cells [40]. In contrast to the SHMT pathway, Thr cleavage reaction via threonine aldolase is, if present, irreversible. This can be readily inferred from the fact that nearly identical f-values were obtained from Thr and Asp. Gly may also be synthesized from a C1 unit and CO 2 via the mitochondrial glycine cleavage (GCV) pathway. In contrast to the previous study with S. cerevisiae [17] and the present glucose cultivation of P. pastoris, we find in the glycerol cultivations no evidence for the efflux into cytosol of Gly which has been reversibly cleaved by GCV. Hence, it may either be that the mitochondrial GCV pathway is operating irreversibly, or that Gly is not exported into the cytosol when cells are grown on glycerol. In principle, yeasts can also synthesize Gly from TCA cycle intermediates via ICL and the alanine, glyoxylate aminotransferase [41]. However, our data suggest that the activity of the glyoxylate cycle is low (see below), so that this route for Gly synthesis is, if active at all, likely to be of minor importance. Central carbon metabolism of P. pastoris growing on glycerol in chemostats The use of the C3 source glycerol for BDF 13 C-labeling of proteinogenic amino acids enabled the determination of the flux ratios for reactions associated with the TCA cycle, while those related to glycolysis and the pentose phosphate pathway (PPP) cannot be assessed (Table 3). This is because labeled glycerol being metabolized through gluconeogenesis and oxidative PPP does not produce labeling patterns that are sufficiently distinct from those generated when glycerol is channeled through the nonoxidative PPP. In fact, the only information that can be derived with respect to the operation of the PPP is obtained from the f-values of His Table 1. Growth parameters at steady state chemostat cultivations of P. pastoris. Y s/x represents the biomass yield, q glyc ,q glc and q O2 are specific utilization rates; q CO2 specific production rates, where glyc and glc indicate glycerol and glucose, respectively. ND, not determined. Carbon source and dilution rate (D,[h )1 ]) Residual substrate concentration (gÆL )1 ) Y s/x (g (dry wt)Æg )1 ) q glyc ,q gluc (mmolÆg )1 Æh )1 ) q CO2 (mmolÆg )1 Æh )1 ) q O2 (mmolÆg )1 Æh )1 ) Glycerol (0.05) 0.0 0.63 0.86 0.9 ND Glycerol (0.16) 3.0 0.63 2.72 2.55 1.29 Glucose (0.16) 0.5 0.57 1.55 3.18 2.73 2466 A. Sola ` et al.(Eur. J. Biochem. 271) Ó FEBS 2004 Cb (Table 2). These reveal the reversible activity of the transketolase and transaldolase reactions when P. pastoris is grown on glycerol. On the other hand, the wealth of information obtained for pathways associated with the TCA intermediates can be summarized as follows (Fig. 2): (a) Gluconeogenesis from cyt-OAA via phosphoenolpyru- vate (PEP) carboxykinase is not registered. (b) Synthesis of mitochondrial pyruvate (mt-PYR) from malate via malic Table 2. Relative abundances of intact C2 and C3 fragments in proteinogenic amino acids. The first column indicates the carbon position for which the 13 C fine structure was observed. The f-values were calculated as described [14], and are given for the glycerol chemostat cultivation at D ¼ 0.05 h )1 in columns 2–5, for the glycerol chemostat cultivation at D ¼ 0.16 h )1 in columns 6–9 and for the glucose chemostat cultivation at D ¼ 0.16 h )1 in columns 10–13. Note, first, that for terminal carbons f (2 * ) and f (3) are not defined, second, that in cases where f (2 * ) is not given for a mid-chain carbon, the carbon-carbon scalar coupling constants are similar and the two doublets cannot be distinguished, and third, that for Tyr the two carbons d 1 and d 2 ,ande 1 and e 2 , respectively, give rise to only one 13 C fine structure each [14]. ND, not determined. Carbon atom Relative abundance of intact carbon fragments Glycerol (0.05 h )1 ) Glycerol (0.16 h )1 ) Glucose (0.16 h )1 ) f (1) f (2) f (2 * ) f (3) f (1) f (2) f (2 * ) f (3) f (1) f (2) f (2 * ) f (3) Ala-a 0.02 0.06 0 0.92 0.01 0.07 0 0.92 0 0.14 0.01 0.85 Ala-b 0.05 0.95 – – 0.02 0.98 – – 0.01 0.99 – – Arg-b 0.71 0.29 – 0 0.56 0.43 – 0.01 0.60 0.39 – 0.01 Arg-d 0.08 0.92 – – 0.10 0.90 – – 0.089 0.91 – – Asp-a 0.26 0.09 0.42 0.23 0.14 0.04 0.18 0.64 0.12 0.12 0.16 0.60 Asp-b 0.26 0.23 0.42 0.09 0.14 0.65 0.17 0.04 0.13 0.71 0.14 0.02 Glu-a 0.24 0.21 0.45 0.10 0.15 0.32 0.39 0.14 0.22 0.27 0.41 0.10 Glu-b 0.68 0.32 – 0 0.50 0.50 – 0 0.61 0.39 – 0 Glu-c 0 0 1 0 0 0 1 0 0 0.02 0.96 0.02 Gly-a 0.06 0.94 – – 0.08 0.92 – – 0.17 0.83 – – His-a 0.04 0 0 0.96 0.06 0 0 0.94 0.06 0 0.04 0.9 His-b 0.06 0.92 0.02 0 0.06 0.94 0 0 0.13 0.53 0 0.34 His-d 2 0.55 0.45 – – 0.55 0.45 – – 0.23 0.77 – – Ile-a 0.38 0 0.62 0 0.21 0 0.79 0 0.23 0 0.77 0 Ile-c 1 0.46 0.51 – 0.03 0.76 0.21 – 0.03 0.74 0.19 – 0.07 Ile-c 2 0.07 0.93 – – 0.06 0.94 – – 0.06 0.94 – – Ile-d 0.52 0.48 – – 0.79 0.21 – – 0.84 0.16 – – Leu-a 0.06 0 0.94 0 0.05 0 0.95 0 0.06 0 0.94 0 Leu-b 0.93 0.02 – 0.05 0.95 0 – 0.05 0.94 0.06 – 0 Leu-d 1 0.07 0.93 – – 0.07 0.93 – – 0.71 0.29 Leu-d 2 0.99 0.01 – – 1 0 – – 1 0 – – Lys-a 0.03 0.07 0.87 0.03 0.03 0.1 0.81 0.06 0.06 0 0.92 0.02 Lys-b 0.76 0.24 – 0 0.56 0.42 – 0.02 0.63 0.37 – 0 Lys-c 0.68 0.29 – 0.03 0.56 0.44 – 0 0.66 0.34 – 0 Lys-d 0.05 0.95 – 0 0.07 0.93 – 0 0.06 0.94 – 0 Lys-e 0.02 0.98 – – 0.04 0.96 – – 0.1 0.9 – – Met-a 0.23 0.13 0.39 0.25 0.12 0.08 0.22 0.58 0.14 0 0.21 0.65 Phe-a 0.04 0 0 0.96 0.03 0 0 0.97 0.03 0.1 0.01 0.86 Phe-b 0.03 0.97 0 0 0.03 0.97 0 0 0.06 0.80 0.14 0 Pro-a 0.29 0.15 0.42 0.14 0.27 0.29 0.27 0.17 0.29 0.28 0.37 0.06 Pro-b 0.72 0.28 – 0 0.57 0.43 – 0 0.65 0.35 – 0 Pro-c 0.06 0.90 – 0.04 0.13 0.85 – 0.02 0.12 0.86 – 0.02 Pro-d 0.15 0.85 – – 0.15 0.85 – – 0.11 0.89 – – Ser-a 0.01 0 0.42 0.57 0.03 0 0.40 0.57 0.13 0.04 0.36 0.47 Ser-b 0.46 0.54 0.45 0.55 – – 0.47 0.53 – – Thr-a 0.27 0.09 0.40 0.24 0.13 0.05 0.17 0.65 0.16 0.09 0.13 0.62 Thr-b 0.24 0.66 0.10 0.12 0.84 – 0.04 0.14 0.85 – 0.01 Thr-c 2 0.49 0.51 – – 0.76 0.24 – – 0.82 0.18 – – Tyr-a 0.06 0 0.01 0.93 0.06 0 0.01 0.93 0.03 0.11 0 0.86 Tyr-b 0.02 0.98 0 0 0.04 0.96 0 0 0.05 0.95 0 0 Tyr-d x 0.03 0.97 – 0 0.05 0.95 – 0 ND ND ND ND Tyr-e x 0.56 0 – 0.44 0.54 0 – 0.46 0.33 0.21 – 0.46 Val-a 0.04 0 0.96 0 0.05 0 0.95 0 0.13 0 0.87 0 Val-c 1 0.04 0.96 – – 0.04 0.96 – – 0.05 0.95 – – Val-c 2 0.96 0.04 – – 0.99 0.01 – – 0.97 0.03 – – Ó FEBS 2004 Metabolic flux profiling of Pichia pastoris (Eur. J. Biochem. 271) 2467 enzyme is likewise not observed. (c) The fraction of mt-OAA reversibly interconverted to fumarate is about the same at high and low dilution rates. (d) The fraction of cytosolic-OAA that stems from the mitochondrial pool of C4 intermediates, e.g. via malate-Asp and/or malate-OAA shuttles [37] (or has possibly been synthesized via the glyoxylate cycle) is about twice as high at the lower dilution rate. This indicates that at close to maximal growth rates, a largely unidirectional flux of OAA from the cytosol to the mitochondria occurs. (e) Significant variations were also identified for the anaplerotic supply of the TCA cycle – a higher fraction of mt-OAA arising from PEP at a higher growth rate reflects the increased demand for biosynthetic building blocks. Comparative profiling of P. pastoris grown on glycerol and glucose At D ¼ 0.16 h )1 , P. pastoris cells grow at about 90% of the maximum specific growth rate with both glycerol and glucose provided as the sole carbon source. Hence, the comparison of flux ratios at this dilution rate allows one to assess the impact of the different chemical nature of the two molecules (Fig. 1) on metabolic regulation when the same Ôbiomass production objectiveÕ is reached. Overall, flux ratios turn out to be rather similar (Table 3), as could be expected due to the very similar oxidation state of glycerol and glucose. The only notable difference is detected for the anaplerotic supply of the TCA cycle: for cells grown on glycerol, the fraction of mt-OAA arising from PEP is about 10% higher than on glucose. The ratio [biomass/(bio- mass + CO 2 )] formation changes accordingly, that is, cells growing on glucose produce CO 2 at higher rates with slightly less biomass being formed when compared to cells grown on glycerol. Among the invariant flux ratios, we find that mt-PYR synthesis from malate is negligible in all chemostat cultures. Consistently, malic enzyme activities were found only at low basal levels in P. pastoris cultures grown on either glycerol or glucose. This finding is in contrast to the important role of the malic enzyme for mt- PYR metabolism in respiro-fermentative glucose batch cultures of S. cerevisiae [17]. Due to degeneracy of the labeling patterns, the glyoxylate cycle activity cannot be Table 3. Origins of metabolic intermediates during aerobic growth of P. pastoris in glycerol chemostat cultures at D = 0.05 h -1 and D = 0.16 h -1 ,and in glucose chemostat culture at D = 0.16 h -1 . For comparison, corresponding data previously reported for P. stipitis and S. cerevisiae glucose- limited chemostat aerobic cultures at D ¼ 0.10 h )1 [18] are given in the two right-most columns. Metabolites Fraction of total pool [%] (mean ± SD) P. pastoris P. stipitis S. cerevisiae Glycerol 0.05 h )1 Glycerol 0.16 h )1 Glucose 0.16 h )1 Glucose a 0.10 h )1 Glucose a 0.10 h )1 Cytosol PEP derived from PPP through at least one transketolase reaction (upper bound) ND ND 41 ± 9 61 ± 11 40 ± 8 P5P from glucose (lower bound) – – 34 ± 2 28 ± 2 41 ± 2 R5P from G3P and S7P (TK reaction) ND ND 66 ± 2 72 ± 2 59 ± 2 R5P from E4P (TK and TA reactions) ND ND 23 ± 2 43 ± 2 33 ± 2 E4P from F6P (lower bound) ND ND 6 ± 5 27 ± 5 6 ± 6 PEP from cyt-OAA (PEP carboxykinase reaction) 0–3 0–6 0–6 0–3 0–10 cyt-OAA from cyt-PYR b 32 ± 2 68 ± 4 63 ± 4 24 ± 3 62 ± 4 cyt-OAA reversibly converted to fumarate at least once (cytosolic or intercompartmental exchange) 56 ± 13 12 ± 6 6 ± 5 47 ± 16 0–8 Mitochondria mt-PYR from malate 0–4 0–4 0–7 0–7 0–13 mt-OAA from PEP (anaplerotic supply of TCA cycle) 33 ± 2 48 ± 2 38 ± 2 32 ± 2 31 ± 2 mt-OAA reversibly converted to fumarate at least once 65 ± 14 61 ± 14 52 ± 14 58 ± 14 56 ± 14 a Data taken from [18]. b Values assuming absence of cytosolic-OAA from fumarate conversion. Fig. 2. Summary of flux information involving the pools of TCA inter- mediates when P. pastoris cells are grown in a chemostat. The top, middle and bottom values in the boxes correspond, respectively, to the glycerol cultivation at D ¼ 0.05 h )1 , the glycerol cultivation at D ¼ 0.16 h )1 and the glucose cultivation at D ¼ 0.16 h )1 .Notethat values associated with arrows pointing at the same metabolite pool add up to 100%. For abbreviations see the legend of Fig. 1. 2468 A. Sola ` et al.(Eur. J. Biochem. 271) Ó FEBS 2004 reliably identified from the NMR data. However, to confirm the previously established view that the glyoxylate cycle is low when yeast cells grow aerobically on glucose [17,18], we have measured ICL activities in the P. pastoris cultures. Indeed, about the same low ICL activities (0.018 UÆmg )1 of protein and 0.019 UÆmg )1 of protein, respectively) were detected in the glucose (D ¼ 0.16 h )1 ) and glycerol (D ¼ 0.05 h )1 ) limited cultures. For cells grownonglycerolatD ¼ 0.16 h )1 , the ICL activities were further reduced (5.07 · 10 )3 UÆmg )1 of protein). Glucose is known to repress the glyoxylate pathway in S. cerevisiae [42], and our enzyme assays show that a similar degree of repression is induced by glycerol in P. pastoris. Comparison of P. pastoris with P. stipitis and S. cerevisiae The central carbon metabolism of S. cerevisiae and P. stip- itis cells grown on glucose in chemostat cultures at D ¼ 0.1 h )1 has been previously characterized [18]. Here we compare (Table 3) these earlier studies with the new data for P. pastoris growing in glucose chemostat cultures at D ¼ 0.16 h )1 .TheS. cerevisiae and P. stipitis glucose limited continuous cultivations were carried out at about one-third of the l max , whereas for the present P. pastoris cultivation about 90% of the l max was reached. Moreover, the medium composition was somewhat different in the earlier studies [18]. Nonetheless, in all cultivations we find that cells exhibit a respiratory metabolism operating to support comparably fast cell growth with very little or no byproduct formation. Moreover, the same approach for flux ratio profiling was used [17], which facilitates comparison. Remarkably, we find that, overall, flux ratios of P. pas- toris at D ¼ 0.16 h )1 are more similar to those of S. cere- visiae than P. stipitis (both at D ¼ 0.1 h )1 ). For example, the fraction of cyt-OAA stemming from cyt-PYR is the same for P. pastoris and S. cerevisiae, but about 2.5 times larger in P. stipitis. This might indicate that in P. pastoris and S. cerevisiae the cytosolic and mitochondrial PYR pools are less well equilibrated in the aerobic growth regime. Similarly, the fraction of cyt-OAA reversibly converted to fumarate is very low for P. pastoris and S. cerevisiae, while about half of the cyt-OAA molecules in P. stipitis have undergone such an interconversion at least once. Consistent with the faster growth, a somewhat increased anaplerotic supply is detected for P. pastoris. Although any generalized conclusion is hampered by the fact that the dilution rate and the growth media were not the same for all cultivations, our findings might possibly provide a first indication that regulation of central carbon metabolism in P. pastoris and S. cerevisiae are more similar to each other than to P. stipitis. Conclusions This is the first comprehensive study of amino acid biosynthesis and central carbon metabolism of the yeast P. pastoris. In the framework of this study, we have established the BDF 13 C-labeling approach of proteinogenic amino acids as an analytical tool to study intermediary metabolism of yeast cells grown on glycerol. This approach allows one to accurately map the metabolic state of the TCA cycle and associated pathways, thus being an important methodological expansion for investigating the metabo- lism of eukaryotic cells grown with carbon sources other than glucose. Specifically, we have shown that (a) the common amino acids are synthesized in P. pastoris as previously described for S. cerevisiae, and that (b) growth on glucose and glycerol results in rather similar flux ratio profiles. Our investigation can be expected to become a valuable Ôbaseline studyÕ for future experiments that are geared towards profiling P. pastoris cells growing on mixtures of glycerol and methanol. Such studies will probably also support the optimization of the larger-scale production of glycosylated human proteins for biomedical applications. Acknowledgements This work was supported by the University at Buffalo, The State University of New York, the Spanish Ministry of Science and Technology (CICYT project PPQ2001-1908), and the Academy of Finland (projects 52311 and 202409). The authors thank J. M. Cregg and C. Gancedo for useful comments, and O. Cos for technical assistance with cultivation off-gas analyses. References 1. Lin Cereghino, J. & Cregg, J.M. (2000) Heterologous protein expression in the methylotrophic yeast Pichia pastoris. FEMS Microbiol. Rev. 24, 45–66. 2. Lin Cereghino, G.P., Lin Cereghino, J., Ilgen, C. & Cregg, J.M. (2002) Production of recombinant proteins in fermenter cultures of the yeast Pichia pastoris. Curr. Opin. Biotechnol. 13, 329–332. 3. Hamilton, S.R., Bobrowicz, P., Bobrowicz, B., Davidson, R.C., Li,H.,Mitchell,T.,Nett,J.H.,Rausch,S.,Stadheim,T.A., Wischnewski, H., Wildt, S. & Gerngross, T.U. (2003) Production of complex human glycoproteins in yeast. Science 301, 1244–1246. 4. Wood, M.J. & Komives, E.A. (1999) Production of large quan- tities of isotopically labeled protein in Pichia pastoris by fermen- tation. J. Biomol. NMR 13, 149–159. 5. de Lamotte, F., Boze, H., Blanchard, C., Klein, C., Moulin, G., Gautier, M.F. & Delsuc, M.A. (2001) NMR monitoring of accumulation and folding of 15 N-labeled protein overexpressed in Pichia pastoris. Prot. Express. Purif. 22, 318–324. 6. Montelione, G.T., Zheng, D., Huang, Y., Gunsalus, K.C. & Szyperski, T. (2000) Protein NMR spectroscopy for structural genomics. Nat. Struct. Biol. 7, 982–984. 7. Stratton, J., Chiruvolu, V. & Meagher, M. (1998) High cell-density fermentation. Methods Mol. Biol. 103, 107–120. 8. Harder, W. & Veenhuis, M. (1989) Metabolism of one-carbon compounds. In The Yeasts, Metabolism and Physiology of Yeasts (Rose, A.H. & Harrison, J.S., eds), Vol. 3, pp. 289–316. Academic Press, London. 9. Egli, T., Ka ¨ ppeli, O. & Fiechter, A. (1982) Mixed substrate growth of methylotrophic yeasts in chemostat culture: Influence of the dilution rate on the utilization of a mixture of glucose and me- thanol. Arch. Microbiol. 131, 8–13. 10. Gancedo, C. & Serrano, R. (1989) Energy-yielding metabolism. In The Yeasts, Metabolism and Physiology of Yeasts (Rose, A.H. & Harrison, J.S., eds), Vol. 3, pp. 205–259. Academic Press, London. 11. Flores, C L., Rodrı ´ guez, C., Petit, T. & Gancedo, C. (2000) Carbohydrate and energy-yielding metabolism in non-conven- tional yeasts. FEMS Microbiol. Rev. 24, 507–529. 12. Jones, E.W. & Fink, G.R. (1982) Regulation of amino acid and nucleotide biosynthesis in yeast. In The Molecular Biology of the Yeast Saccharomyces, Metabolism and Gene Expression Ó FEBS 2004 Metabolic flux profiling of Pichia pastoris (Eur. J. Biochem. 271) 2469 (Strathern, J.N., Jones, E.W. & Broach, J.R., eds), pp. 181–299. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 13. Szyperski, T. (1998) 13 C-NMR, MS and metabolic flux balancing in biotechnology research. Q. Rev. Biophys. 31, 41–106. 14. Szyperski, T. (1995) Biosynthetically directed fractional 13 C-labe- ling of proteinogenic amino acids. An efficient analytical tool to investigate intermediary metabolism. Eur. J. Biochem. 232, 433– 448. 15. Sauer, U., Lasko, D.R., Fiaux, J., Hochuli, M., Glaser, R., Szyperski, T., Wu ¨ thrich, K. & Bailey, J.E. (1999) Metabolic flux ratio analysis of genetic and environmental modulations of Escherichia coli central carbon metabolism. J. Bacteriol. 181, 6679–6688. 16. Sauer, U., Hatzimanikatis, V., Bailey, J.E., Hochuli, M., Szyper- ski, T. & Wu ¨ thrich, K. (1997) Metabolic fluxes in riboflavin-pro- ducing Bacillus subtilis. Nat. Biotechnol. 15, 448–452. 17. Maaheimo,H.,Fiaux,J.,C¸ akar, Z.P., Bailey, J.E., Sauer, U. & Szyperski, T. (2001) Central carbon metabolism of Saccharomyces cerevisiae explored by biosynthetic fractional 13 C labeling of common amino acids. Eur. J. Biochem. 268, 2464–2479. 18. Fiaux, J., C¸ akar, Z.P., Sonderegger, M., Wu ¨ thrich, K., Szyperski, T. & Sauer, U. (2003) Metabolic flux profiling of the yeasts Saccharomyces cerevisiae and Pichia stipitis. Eukaryotic Cell 2, 170–180. 19. Hochuli, M., Patzelt, H., Oesterhelt, D., Wu ¨ thrich,K.& Szyperski, T. (1999) Amino acid biosynthesis in the halophilic archaeon Haloarcula hispanica. J.Bacteriol. 181, 3226–3237. 20. Minning, S., Serrano, A., Ferrer, P., Sola ` ,C.,Schmid,R.D.& Valero, F. (2001) Optimisation of the high-level production of Rhyzopus oryzae lipase in Pichia pastoris. J. Biotechnol. 86, 59–70. 21. Gu ¨ ntert, P., Do ¨ tsch, V., Wider, G. & Wu ¨ thrich, K. (1992) Pro- cessing of multi-dimensional NMR data with the new software PROSA. J. Biomol. NMR 2, 619–629. 22. Szyperski, T., Glaser, R.W., Hochuli, M., Fiaux, J., Sauer, U., Bailey, J. & Wu ¨ thrich, K. (1999) Bioreaction network topology and metabolic flux ratio analysis by biosynthetic fractional 13 C- labeling and two dimensional NMR spectroscopy. Metab. Eng. 1, 189–197. 23. Gombert, A.K., Moreira dos Santos, M., Christensen, B. & Nielsen, J. (2001) Network identification and flux quantification in the central metabolism of Saccharomyces cerevisiae under different conditions of glucose repression. J. Bacteriol. 183, 1441–1451. 24. Marx, A., de Graaf, A.A., Wiechert, W., Eggeling, L. & Sahn, H. (1996) Determination of the fluxes in the central carbon metabo- lism of Corynebacterium glutamicum by nuclear magnetic reso- nance spectroscopy combined with metabolite balancing. Biotechnol. Bioeng. 49, 111–129. 25. van Winden, W., Schipper, D., Verheijen, P. & Heijnen, J. (2001) Innovations in generation and analysis of 2D [ 13 C, 1 H] COSY NMR spectra for metabolic flux analysis purposes. Metab. Eng. 3, 322–343. 26. Walsh, K. & Koshland, D.E. (1984) Determination of flux through the branch point of two metabolic cycles. The tri- carboxylic acid cycle and the glyoxylate shunt. J. Biol. Chem. 259, 9646–9654. 27. Sonntag, K., Eggeling, L., de Graaf, A.A. & Sahm, H. (1993) Flux partitioning in the split pathway of lysine synthesis in Coryne- bacterium glutamicum: Quantification by 13 C- and 1 H-NMR spectroscopy. Eur. J. Biochem. 213, 1325–1331. 28. Wiechert. W., Siefke, C., de Graaf, A.A. & Marx, A. (1997) Bidirectional reaction steps in metabolic networks. II. Flux esti- mation and statistical analysis. Biotechnol. Bioeng. 55, 118–135. 29. Klapa, M.I., Aon, J C. & Stephanopoulos, G. (2003) Systematic quantification of complex metabolic flux networks using stable isotopes and mass spectrometry. Eur. J. Biochem. 270, 3525–3542. 30. Gancedo, C., Gancedo, J.M. & Sols, A. (1968) Glycerol meta- bolism in yeasts. Pathways of utilization and production. Eur. J. Biochem. 5, 165–172. 31. Gancedo, J.M. & Gancedo, C. (1997) Gluconeogenesis and catabolite inactivation. In Yeast Sugar Metabolism: Biochemistry, Genetics, Biotechnology and Applications (Zimmermann, F.K. & Entian, K D., eds), pp. 359–377. Technomic Publishing, Lancaster, PA. 32. Ferna ´ ndez, E., Ferna ´ ndez,M.,Moreno,F.&Rodicio,R.(1993) Transcriptional regulation of the isocitrate lyase encoding gene in Saccharomyces cerevisiae. FEBS. Lett. 333, 238–242. 33. Scho ¨ ler, A. & Schu ¨ ller, H J. (1994) A carbon-source-responsive promoter element necessary for activation of the isocitrate lyase gene ICL1 is common to genes of the gluconeogenic pathway in the yeast Saccharomyces cerevisiae. Mol. Cell. Biol. 14, 3613–3622. 34. Bruinenberg, P.G., Blaauw, M., Veenhuis, M. & Ab, G. (1989) Purification and some properties of malate synthase from the methylotrophic yeast Hansenula polymorpha. FEMS Microbiol. Lett. 61, 11–16. 35. Palmieri, L., Vozza, A., Agrimi, G., De Marco, V., Runswick, M.J., Palmieri, F. & Walker, J.E. (1999) Identification of the yeast mitochondrial transporter for oxalacetate and sulfate. J. Biol. Chem. 274, 22184–22190. 36. Palmieri, L., Lasorsa, F.M., De Palma, A., Palmieri, F., Runs- wick, M.J. & Walker, J.E. (1997) Identification of the yeast ACR1 gene product as a succinate-fumarate transporter essential for growth on ethanol or acetate. FEBS Lett. 417, 114–118. 37. Bakker, B.M., Overkamp, K.M., van Maris, A.J.A., Ko ¨ tter, P., Luttik, M.A.H., van Dijken, J.P. & Pronk, J.T. (2001) Stoichio- metry and compartmentation of NADH metabolism in Sacchar- omyces cerevisiae. FEMS Microbiol. Rev. 25, 15–37. 38. Voet, D. & Voet, J.G. (1995) Biochemistry. John Wiley & Sons, New York. 39. Michal, G. (1998) Biochemical Pathways: an Atlas of Biochemistry and Molecular Biology. John Wiley & Sons, New York. 40. Woldman, Y. & Appling, D.R. (2002) A general method for determining the contribution of split pathways in metabolite production in the yeast Saccharomyces cerevisiae. Metab. Eng. 4, 170–181. 41. Takada, Y. & Noguchi, T. (1985) Characteristics of alanine: glyoxylate aminotransferase from Saccharomyces cerevisiae,a regulatory enzyme in the glyoxylate pathway of glycine and serine biosynthesis from tricarboxylic acid-cycle intermediates. Biochem. J. 231, 157–163. 42. Gancedo, J.M. (1998) Yeast carbon catabolite repression. Microbiol. Mol. Biol. Rev. 62, 334–361. Supplementary material The following material is available from http://blackwell publishing.com/products/journals/suppmat/EJB/EJB4176/ EJB4176sm.htm Fig. S1. 13 C Scalar coupling fine structures of BDF 13 C-labeled amino acids from the P. pastoris cells grown in glycerol chemostat cultivations at the dilution rates 0.05 h )1 and 0.16 h )1 . Fig. S2. 13 C Scalar coupling fine structures of BDF 13 C-labeled amino acids from the P. pastoris cells grown in glucose chemostat cultivation at dilution rate of 0.16 h )1 . 2470 A. Sola ` et al.(Eur. J. Biochem. 271) Ó FEBS 2004 . Amino acid biosynthesis and metabolic flux profiling of Pichia pastoris Aina Sola ` 1,2 , Hannu Maaheimo 2,3 , Katri Ylo¨ nen 3 , Pau Ferrer 1 and Thomas. Germany) and dried through a silica gel Ó FEBS 2004 Metabolic flux profiling of Pichia pastoris (Eur. J. Biochem. 271) 2463 column. Concentrations of oxygen and