Glyoxylate biosynthesis in Saccharomyces cerevisiae is traditionally mainly ascribed to the reaction catalyzed by isocitrate lyase (Icl), which converts isocitrate to glyoxylate and succinate. However, Icl is generally reported to be repressed by glucose and yet glyoxylate is detected at high levels in S. cerevisiae extracts during cultivation on glucose. In bacteria there is an alternative pathway for glyoxylate biosynthesis that involves a direct oxidation of glycine. Therefore, we investigated the glycine metabolism in S. cerevisiae coupling metabolomics data and 13C-isotope-labeling analysis of two reference strains and a mutant with a deletion in a gene encoding an alanine:glyoxylate aminotransferase. The strains were cultivated on minimal medium containing glucose or galactose, and 13C-glycine as sole nitrogen source. Glyoxylate presented 13C-labeling in all cultivation conditions. Furthermore, glyoxylate seemed to be converted to 2-oxovalerate, an unusual metabolite in S. cerevisiae. 2-Oxovalerate can possibly be converted to 2-oxoisovalerate, a key precursor in the biosynthesis of branched-chain amino acids. Hence, we propose a new pathway for glycine catabolism and glyoxylate biosynthesis in S. cerevisiae that seems not to be repressed by glucose and is active under both aerobic and anaerobic conditions. This work demonstrates the great potential of coupling metabolomics data and isotope-labeling analysis for pathway reconstructions.
The glyoxylate pathway is the main and well-known pathway that leads to glyoxylate biosynthesis in Saccharomyces cerevisiae,. Isocitrate lyase (Icl) is the key enzyme of the glyoxylate pathway, which bypasses the two decarboxylation steps in the TCA cycle and leads to the synthesis of succinate (C4) and glyoxylate (C2). However, there are strong evidences in the literature about the repression of Icl by glucose –. Nonetheless, by developing very sensitive and low-discriminative analytical techniques for metabolome analysis of yeasts, glyoxylate has been detected at high levels intra- and even extracellularly in S. cerevisiae cultures growing on glucose ,.
By determining the metabolite profile of S. cerevisiae during very-high – gravity ethanol fermentations , intracellular glycine levels appeared to be inversely related to the glyoxylate levels, which could point to glycine being a possible precursor for glyoxylate in S. cerevisiae. Biosynthesis of glyoxylate from glycine has been described in several prokaryotes such as Bacillus subtilis and Nitrobacter agilis, among others. However, the most well-described catabolic reaction of glycine in yeasts is its decarboxylation with subsequent conversion to serine, catalyzed by the glycine decarboxylase multienzyme complex (Gdc) . The Gdc, also known as the glycine cleavage system or glycine synthase (EC 126.96.36.199), fills a critical metabolic position connecting the metabolism of one-, two-, and three-carbon compounds and is linked to many different metabolic reactions.
Although glycine is usually described as a poor source of nitrogen for yeasts, S. cerevisiae can grow on glycine as the sole nitrogen source . Sinclair and Dawes  have investigated yeast strains with mutations in single genes involved in glycine uptake and decarboxylation, and they found a solid indication of a second pathway for glycine assimilation in yeasts, as two of the mutants tested could not decarboxylate glycine but could use it as the sole nitrogen source.
The putative second pathway for glycine assimilation could be a reversible reaction catalyzed by alanine: glyoxylate aminotransferase (Agt). Agt (EC 188.8.131.52) is one of three different enzymes used for glycine synthesis in S. cerevisiae. Glyoxylate is transaminated to glycine by Agt with a concurrent conversion of alanine to pyruvate. However, this enzyme has been reported to be repressed by glucose, and a purified enzyme preparation demonstrated to be highly selective for using l-alanine and glyoxylate as substrate, hence there was strong evidence for irreversibility of the reaction .
In this work, the synthesis of glyoxylate from glycine was investigated by cultivating two different S. cerevisiae reference strains and a mutant with a deletion in the gene that encodes Agt. The strains were grown on glucose and galactose, with galactose representing a non-fermentable carbon source and, thus, imposing little carbon catabolite repression. 13C-labeled glycine was used as the sole nitrogen source, and its catabolism was followed by metabolome analysis coupled to 13C-labeling analysis of key metabolic intermediates using GC–MS.
2 Material and methods
2.1 Yeast strains
Three S. cerevisiae strains were used: CEN.PK113-7D (MATaMAL2-8cSUC2) and BY4741 (MATa;his3Δ1;leu2Δ0;met15Δ0;ura3Δ0) as reference strains, and the mutant with BY background YFL030w (MATa;his3Δ1;leu2Δ0;met15Δ0;ura3Δ0;YFL030w::kanMX4), obtained from the EUROSCARF collection.
2.2 Media supplements
Histidine, leucine and methionine were obtained from Serva Electrophoresis GmbH (Heidelberg, Germany). Uracil and 13C1-glycine was purchased from Sigma (St. Louis, MO).
S. cerevisiae CEN.PK113-7D was cultivated aerobically and anaerobically using shake flasks containing glucose or galactose (20 g l−1), 13C1-glycine (5.0 g l−1), MgSO4· 7H2O (0.5 g l−1), KH2PO4 (3.0 g l−1), and vitamins and trace elements according to Verduyn et al. . Aerobic cultivations were performed using a rotary shaker at 30 °C and 200 rpm, in shake flasks containing 150 ml medium and cotton plugs. Anaerobic cultivations were carried out under moderate shaking (130 rpm) at 30 °C in shake flasks, containing 150 ml of medium, with tight rubber plugs. The flasks were flushed with nitrogen prior to cultivation and the medium was supplemented with ergosterol (10 mg l−1) according to Verduyn et al. .
S. cerevisiae BY4741 and YFL030w were cultivated aerobically using shake flasks containing glucose or galactose (20 g l−1), 13C1-glycine (5.0 g l−1), MgSO4· 7H2O (0.5 g l−1), KH2PO4 (3.0 g l−1), histidine (50 mg l−1), leucine (50 mg l−1), methionine (50 mg l−1), uracil (50 mg l−1), vitamins and trace elements according to Verduyn et al. . The cultivations were performed using a rotary shaker at 30 °C and 200 rpm, in shake flasks containing 150 ml medium and cotton plugs.
2.4 Extracellular sampling
Five milliliters of the culture medium were filtrated using Millipore membrane (0.45 μm) and then the filtrate was lyophilized under low temperature (−56 °C) using an ALPHA 1-4 freeze–dryer (Christ, München, Germany).
2.5 Intracellular sampling, quenching, and extraction
Samples from each flask were harvested during glucose or galactose consumption (upper exponential phase). The cellular metabolism was quenched and the intracellular metabolites were extracted according to the procedure described by Koning and van Dam . The metabolites were concentrated by lyophilization under low temperature (−56 °C) using an ALPHA 1-4 freeze–dryer.
2.6 Sample derivatization
After lyophilization, the solids from both intracellular and extracellular samples were dissolved in 200 μl of sodium hydroxide solution (1%) and the metabolites were derivatized according to the protocol described previously in Villas-Bôas et al. .
2.7 GC–MS analysis
We used a Hewlett-Packard system HP 6890 gas chromatograph coupled to a HP 5973 quadrupole mass selective detector (EI) operated at 70 eV. The column used for all analysis was a DB1701 (J&W Scientific, Folsom, CA, USA), 30 m × 250 μm i.d. × 0.15 μm film thickness. The MS was operated in scan mode (start after 5 min, mass range 38–550 a.m.u. at 2.88 s/scan) and in SIM mode for the selective detection of the ions 103 and 104 a.m.u. (typical ions of glyoxylate-MCF derivative). The detailed analysis parameters have been described in Villas-Bôas et al. .
2.8 Determination of labeling pattern
The labeling pattern of the detected metabolites from central carbon and amino acid metabolism were determined by estimating the increase of 13C-fraction in the selected ion clusters as shown in Table 1.
List of ion clusters used to determine the increase of 13C-labeling in the detected metabolites
Mass isotopomer relative abundance
3.1 Cultivations and metabolite profiles
S. cerevisiae CEN.PK grew poorly on glucose and galactose with glycine as the sole nitrogen source, presenting very low specific growth rates (Table 2). During cultivation on glucose the maximum OD600 reached was 2.8, while on galactose the cultures reached a maximum OD600 of 6.0. These results suggest that glucose repression is behind the limited growth of S. cerevisiae on glucose with glycine as sole nitrogen source. Glucose could have repressed either the glycine transportation across the cell membrane or its catabolic reactions intracellularly.
Fermentation parameters obtained during aerobic cultivation of three Saccharomyces cerevisiae strains on different media
MM + glucose
MM + galactose
MM + S + glucose
MM + S + galactose
MM: minimal media; S: supplemented with histidine, leucine, methionine and uracil; –: no growth observed.
The strains with BY background presented much higher specific growth rates as well as higher biomass production (Table 2). S. cerevisiae with BY background is auxotrophic to three amino acids and uracil, requiring medium supplementation with minimal amounts of these compounds (50 mg l−1). Cultivating S. cerevisiae CEN.PK background in a medium supplemented with these amino acids and uracil also resulted in a considerably higher specific growth rate. The specific growth rate increased 181% during glucose cultivation and 45% during galactose cultivation (Table 2). Similarly, the biomass production increased 74% and 11.7% during cultivation on glucose and galactose, respectively.
It is quite clear, therefore, that the medium supplementation with low amounts of histidine, leucine, methionine and uracil improved the ability of S. cerevisiae to grow in minimal media containing glycine as the main nitrogen source. The presence of other nitrogenated compounds in the media could have facilitated the transport of glycine across the membrane. The transport of glycine in S. cerevisiae is not well elucidated yet, but the general amino acid permease (Gap1) is believed to be the key transporter for glycine . Generally, the nitrogenated compounds are transported across the plasma membrane against a concentration gradient by Gap1, which presents high capacity and low affinity. According to Hofman-Bang , the transcriptional activity of Gap1 in S. cerevisiae is repressed in a minimal medium containing ammonium, asparagine, or glutamine, and it is induced by poor nitrogen sources, such as leucine and methionine that were all added to the cultivation medium.
The mutant strain presented exactly the same specific growth rate and biomass production of its reference strain during cultivation on galactose (Table 2), but during the cultivation on glucose it presented a lower specific growth rate than the reference strain (34% lower), and a slightly lower biomass production (8% lower). Nevertheless, the mutant grew comparatively well on minimal medium with glycine as the main nitrogen source even through the alanine:glyoxylate aminotransferase-encoding gene was deleted. These results, therefore, confirm that it is unlikely that the catabolism of glycine involves reversibility of the alanine:glyoxylate aminotransferase reaction, corroborating the results obtained by Takada and Noguchi  and Sinclair and Dawes .
3.2 13C-labeling pattern
Figs. 2 and 3 present the relative abundances of m+ 1 ions of the principal ion clusters selected for each metabolite described in Table 1. A few metabolites were detected only in the extracellular media (Supplementary Tables 1 and 2). In all cultivations glyoxylate clearly presented a drastic increase in the abundance of its m+ 1 ion, suggesting that it is a direct product/intermediate from 13C-glycine metabolism. Increase in the abundance of m+ 1 ion from 2-oxovalerate was also detected in most of the cultivations (except for the BY strains cultivated on galactose).
Relative abundance of m+ 1 ions in different metabolites detected during aerobic cultivation of two S. cerevisiae strains (BY4741 and YFL030w) on glucose and galactose with glycine as main nitrogen source. (a) Reference strain cultivated on glucose; (b) mutant strain cultivated on glucose; (c) reference strain cultivated on galactose; (d) mutant strain cultivated on galactose. White bars, reference values determined using non-labeled standards; black bars: abundances observed at higher levels compared to the respective references.
Glycine metabolism in Saccharomyces cerevisiae. We prove that there are at least two pathways for glycine catabolism in S. cerevisiae: (1) via Gdc and (2) via a de novo described Gda. We postulate that 2-oxovalerate is synthesized from glyoxylate by an unknown reaction/enzyme with its subsequent conversion to 2-oxoisovalerate by Dhad. Gdc: glycine decarboxylase multienzyme complex; Sda: serine deaminase; Agt: alanine: glyoxylate aminotransferase; Gda: glycine deaminase; Dhad: dihydroxy acid dehydratase; Ipms: isopropylmalate synthase; Icl: isocitrate lyase; Tb: transaminase B. Full arrows indicate confirmed pathways and dashed arrows indicate speculative pathways. The numbers specify the number of reaction steps in the pathway.
Decarboxylation of glycine to CO2 and Decarboxylation of NH4+ by Gdc yields the activated one-carbon unit for the formation of serine via 5,10-methylene-tetrahydrofolate. However, serine was not detected in the samples from any of the cultivations. Nonetheless, serine is metabolized in S. cerevisiae by serine deaminase (EC 184.108.40.206) to pyruvate. Pyruvate is transported to the mitochondria or it can be converted to alanine, valine and leucine via 2-oxoisovalerate and isopropylmalate, or isoleucine via 2-oxobutanoate. We expected that the labeling pattern in pyruvate and posterior intermediates would suffer a huge dilution since the main carbon source (glucose/galactose) was not labeled and, thus, the 13C incorporated from glycine would consist of a fairly small fraction, possibly below the detection limit of the instrument. Indeed, pyruvate did not appear labeled in any samples analyzed, but 2-oxoisovalerate and isopropylmalate appeared labeled in several samples (Figs. 2 and 3), and isoleucine presented a huge increase in m+ 1 ion during aerobic cultivation on galactose (Fig. 1(c)). Labeled isoleucine was also detected at a considerable level in samples from the mutant strain cultivated aerobically on glucose (Fig. 2(b)). However, no significant 13C-label was found in isoleucine during anaerobic cultivations (Fig. 1(b) and (d)).
Relative abundance of m+ 1 ions in different metabolites detected during cultivation of S. cerevisiae (CEN.PK113.7D) on glucose and galactose with glycine as sole nitrogen source. (a) Aerobic cultivation on glucose; (b) anaerobic cultivation on glucose; (c) aerobic cultivation on galactose; (d) anaerobic cultivation on galactose. White bars, reference values determined using non-labeled standards; black bars: abundances observed at higher levels compared to the respective references.
Labeled valine was detected in many samples, with exception of the samples from the BY strains cultivated on galactose, and oxaloacetate appeared labeled in the samples from the BY strains cultivated on glucose (Figs. 2 and 3). Several other metabolites, including some intermediates of the TCA cycle, such as fumarate, malate, isocitrate and citrate presented labeling in different samples from different cultivations.
Based on the results presented in this work it is clear that glycine can be directly oxidized to glyoxylate as demonstrated to occur in other microorganisms ,[9, and others].
The catabolic reaction of glycine via Gdc is believed to be repressed by glucose ,, and we could not directly determine the activity of this pathway by using 13C-glycine, due to the lack of serine detection in the metabolite pool. On the other hand, the label in isoleucine, detected in several samples from aerobic cultivations, probably coming from the glycine-serine pathway, can be taken as evidence of Gdc activity during growth on both carbon sources, although the route used for biosynthesis of the labeled isoleucine seems to be less active under anaerobic conditions.
On the other hand, the direct deamination of glycine to glyoxylate did not seem to be repressed by glucose since 13C-labelling was observed in glyoxylate in all cultivation conditions tested and at both aerobic and anaerobic growth conditions. However, the contribution of this pathway to the global catabolism of glycine by S. cerevisiae and its influence on the yeast ability to use glycine as nitrogen source still need to be elucidated by further studies (i.e. investigating the operation of this new pathway in gcv1-, gcv2-, gcv3- and lpd1-deleted mutants).
In addition our results support the view that the oxidation of glycine to glyoxylate is not via a reversible Agt reaction, as the mutant with the Agt-encoding gene deleted grew comparatively well on glucose with glycine as the main nitrogen source and this also resulted in glyoxylate with increased 13C-labelling (Fig. 3). Therefore, our results prove the presence of a yet undescribed pathway for glycine catabolism and glyoxylate biosynthesis in S. cerevisiae, a pathway the existence of which has earlier been indicated by Sinclair and Dawes  and by our previous works on yeast metabolomics ,.
However, it is still unclear why valine and isopropylmalate appeared labeled in several samples, while leucine did not (Figs. 2 and 3). A possible answer could be connected to the finding that 2-oxovalerate was labeled in all samples where it was detected. 2-Oxovalerate is an unusual metabolite that was first detected and identified in S. cerevisiae extracts through global metabolome analysis of central carbon and amino acids metabolism carried out in our previous work . We have speculated on the role of 2-oxovalerate and Fig. 3 shows our suggestion for the global pathways of glycine metabolism in S. cerevisiae and the putative biosynthetic reaction of 2-oxovalerate and its possible metabolic pathways. Based on the labeling pattern of 2-oxovalerate we postulate that it is possibly synthesized from glyoxylate. Once synthesized, 2-oxovalerate could be putatively converted to 2-oxoisovalerate, the main precursor of valine, by the dihydroxy-acid dehydratase (EC 220.127.116.11), which has been considered a low-specific enzyme .
In conclusion, besides confirming the presence of a so far undescribed metabolic pathway for glyoxylate biosynthesis and speculating on a few other unknown pathways in S. cerevisiae, we have shown how data from global metabolome analysis with simultaneous metabolite identification as presented here and in our previous works ,, coupled to the data from isotope labeling analysis, can be used to elucidate new metabolic pathways.
We thank Mrs. Kianoush K. Hansen for helpful technical support. This work has been supported by the Danish Biotechnological Instrument Center.
↵3 EUROSCARF collection: European S. cerevisiae Archive for Functional Analysis – Institute of Microbiology, Johann Wolfgang Goethe-University Frankfurt, Marie-Curie-Strasse 9; Building N250. D-60439 Frankfurt, Germany. Fax: +49 69 79829527, http://web.uni-frankfurt.de/fb15/mikro/euroscarf.
(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.