Logo of aemPermissionsJournals.ASM.orgJournalAEM ArticleJournal InfoAuthorsReviewers
Appl Environ Microbiol. Sep 2006; 72(9): 6194–6203.
PMCID: PMC1563674

Global Transcriptional and Physiological Responses of Saccharomyces cerevisiae to Ammonium, l-Alanine, or l-Glutamine Limitation


The yeast Saccharomyces cerevisiae encounters a range of nitrogen sources at various concentrations in its environment. The impact of these two parameters on transcription and metabolism was studied by growing S. cerevisiae in chemostat cultures with l-glutamine, l-alanine, or l-ammonium in limitation and by growing cells in an excess of ammonium. Cells grown in l-alanine-limited cultures had higher biomass yield per nitrogen mole (19%) than those from ammonium-limited cultures. Whole-genome transcript profiles were analyzed with a genome-scale metabolic model that suggested increased anabolic activity in l-alanine-limited cells. The changes in these cells were found to be focused around pyruvate, acetyl coenzyme A, glyoxylate, and α-ketoglutarate via increased levels of ALT1, DAL7, PYC1, GDH2, and ADH5 and decreased levels of GDH3, CIT2, and ACS1 transcripts. The transcript profiles were then clustered. Approximately 1,400 transcripts showed altered levels when amino acid-grown cells were compared to those from ammonium. Another 400 genes had low transcript levels when ammonium was in excess. Overrepresentation of the GATAAG element in their promoters suggests that nitrogen catabolite repression (NCR) may be responsible for this regulation. Ninety-one genes had transcript levels on both l-glutamine and ammonium that were decreased compared to those on l-alanine, independent of the concentration. The GATAAG element in these genes suggests two groups of NCR-responsive genes, those that respond to high levels of nitrogen and those that respond to levels below 30 μM. In conclusion, our results reveal that the nitrogen source has substantial influence on the transcriptome of yeasts and that transcriptional changes may be correlated to physiology via a metabolic model.

In its natural habitat, the yeast Saccharomyces cerevisiae is often growth limited by the availability of nitrogen (1, 37). To survive and adapt to various environmental conditions, S. cerevisiae has evolved complex regulatory networks to ensure an efficient regulation of metabolism, as illustrated by the systems for sensing and regulation of the nitrogen metabolism (7, 13, 20). An example of global response to amino acid depletion is the transcriptional induction of amino acid biosynthesis genes by the transcription factor Gcn4 (20). Additional specific systems are responsible for the sensing of amino acids or ammonium in the medium and the subsequent induction of corresponding transporter genes and pseudohyphal growth (13, 16, 22, 27, 30).

Central for the transcriptional regulation of nitrogen as well as carbon metabolism are the target of rapamycin (TOR) pathway and the protein kinase A pathway. These pathways control the metabolism by sequestering transcription factors such as Msn2, Msn4, Gln3, and Crf1 to the cytosol when nutrients are in excess (4, 31, 49). When glucose is depleted from the medium, Msn2 and Msn4 enter the nucleus, where they induce transcription of tricarboxylic acid (TCA) cycle genes, gluconeogenesis genes, and genes carrying the stress response element STRE (40). The same is the case for Gln3, which enters the nucleus upon inactivation of the TOR pathway or a change in the nitrogen availability, where it induces genes with the GATAAG promoter element (4). Moreover, Gln3 is part of a network involved in nitrogen catabolite repression (NCR), which ensures selective use of different available nitrogen sources (7). This selection is obtained by transcriptional repression and induction of genes involved in nitrogen uptake and metabolism, such as GLN1, GDH2, GLT1, and GAP1 (7, 8, 32, 43). NCR is dependent on at least four transcription factors that compete for the cis-acting element GATAAG (7). Two GATA transcription factors, Gln3 and Gat1, are transcriptional activators, while Dal80 and Gzf3 function as transcriptional repressors (10, 11, 32, 43, 46). The TOR pathway has been shown to trigger NCR upon increased levels of intracellular l-glutamine (28), though the link is probably not direct, as different nitrogen sources, such as ammonium and l-glutamine, are shown to repress different sets of NCR-sensitive genes (43). Part of this complexity is probably due to the mutual regulation that the GATA factors impose on each other (7) as well as to the different sensitivities of NCR transcription factors to the repressing nitrogen source.

Besides the impact on nitrogen metabolism, the presence of different nitrogen sources also affects the carbon metabolism, e.g., via the TCA cycle intermediate α-ketoglutarate, which serves as a precursor for the synthesis of l-glutamate and as an amino donor for the synthesis of other amino acids, amino sugars, and nucleotides. α-Ketoglutarate is aminated by the glutamate dehydrogenases Gdh1 and Gdh3 and by the glutamate synthase-glutamine synthase pathway encoded by GLT1 and GLN1 (2). The fluxes through the α-ketoglutarate node are therefore highly dependent on the nitrogen source, as in the case of ammonium, which will lead to a higher net use of α-ketoglutarate than will amino acids such as l-glutamine. The retrograde synthesis of α-ketoglutarate from oxaloacetate is regulated at the transcript level by Rtg2, which represses a repressor, Mks1, resulting in activation of the transcription factors Rtg1 and Rtg3 (26). A recent study shows that the intracellular ammonium concentration induces the activity of Rtg2 and thereby the net production of α-ketoglutarate (44).

However, it is still unclear how nitrogen and carbon metabolism is affected when cells are grown on amino acids such as l-alanine, where the carbon backbone feeds into the TCA cycle via acetyl coenzyme A (acetyl-CoA). The close links between the carbon and nitrogen biochemistry, at the metabolic level as well as the regulatory level, therefore makes it important to understand how cells will respond to growth on different nitrogen sources.

We have consequently examined the impact of ammonium, l-alanine, and l-glutamine on the physiology and transcriptome of the yeast S. cerevisiae. Ammonia and l-glutamine enter metabolism via the glutamine and glutamate dehydrogenases and are classified as preferable nitrogen sources that induce NCR (29). l-Alanine, on the other hand, enters metabolism via acetyl-CoA, and its impact on regulation is largely unknown. Cells were grown in well-controlled aerobic chemostat cultures at a growth rate μ of 0.2 h−1 with ammonia, l-glutamine, l-alanine, or glucose as the limiting nutrient. Chemostat cultivation allowed the identification of nitrogen source responses at constant nutrient levels and constant growth rates and thereby minimized the number of confounding parameters found in batch cultivations (12). Finally, we linked the physiological data to the transcriptome data by the use of an enzyme subnetwork analysis combined with cluster and promoter analysis.


Strain and maintenance.

This study was performed with the prototrophic haploid laboratory strain S. cerevisiae CEN.PK113-7D (MATα) (46), which was provided by P. Kötter from the EUROSCARF strain collection (Frankfurt am Main, Germany). A frozen stock culture was used for the preculture, which was grown in a two-baffle shake flask (500 ml) containing 100 ml of minimal medium (48) at 30°C for 24 h in an orbital shaker at 150 rpm and a pH of 6.5.

Chemostat cultivation.

Steady-state aerobic chemostat cultures were grown at 30°C by use of 2-liter bioreactors (Applikon) with a working volume of 1.0 liter at a dilution rate D of 0.2 (± 0.005) h−1. Cultures were fed with a modified minimal medium containing a nitrogen source at a concentration corresponding to 6 mM nitrogen and 250 mM glucose (calculated on nitrogen and carbon atom bases, respectively), which ensured that growth was limited by the nitrogen source and that glucose was in excess. Three different nitrogen sources were used in the three comparative experiments: ammonium sulfate (NH4)2SO4, l-glutamine, and l-alanine. All other substrates were in excess. The pH was measured online and kept constant at 5.0 by automatic titration with 4 M KOH by use of an Applikon (ADI1030) biocontroller. The stirring speed was set to 700 rpm, and the dry airflow rate was 1.0 liter/min. The exhaust gas from chemostat cultivation was led through a condenser, and the mole-% values for carbon dioxide and oxygen were measured online by use of a personal computer-controlled acoustic gas analyzer (Brüel & Kjær, Denmark). After batch cultivation, dry weight, metabolite concentrations, and gas profiles were monitored, and when they were constant over at least five residence times, steady state was assumed. Thereafter, samples for RNA extraction were taken. All samples were collected by no later than 10 to 15 residence times from the start of continuous operation to avoid any strain adaptation generally occurring over long-term cultivation (15).


The nitrogen-limited minimal medium composition was based on that described in the work of Klein et al. (24). The amounts per liter of the compounds were as follows: KH2PO4, 3 g; MgSO4 · 7H2O, 0.5 g; d-glucose, 7.5 g; antifoam 289 (A-5551, Sigma-Aldrich), 0.05 ml; EDTA (Titriplex III), 15.0 mg; ZnSO4 · 7H2O, 4.5 mg; MnCl2 · 2H2O, 0.82 mg; CoCl2 · 6H2O, 0.3 mg; CuSO4 · 5H2O, 0.3 mg; Na2MoO4 · 2H2O, 0.4 mg; CaCl2 · 2H2O, 4.5 mg; FeSO4 · 7H2O, 3.0 mg; H3BO3, 1.0 mg; KI, 0.1 mg; biotin, 0.05 mg; p-benzoic acid, 0.2 mg; nicotinic acid, 1.0 mg; Ca-pantothenate, 1.0 mg; pyridoxin-HCl, 1.0 mg; thiamine-HCl, 1.0 mg; and m-inositol, 25.0 mg. The nitrogen source in the medium was 6 nitrogen mmol per liter for all the nitrogen-limited chemostat experiments. This corresponded to the following: (NH4)2SO4, 0.4 g/liter; l-glutamine, 0.4425 g/liter; or l-alanine, 0.534 g/liter. Media for three glucose-limited chemostat cultures had the same composition as those described above, except that the nitrogen concentration was increased to 5.0 g (NH4)2SO4 per liter.

Metabolite and dry-weight analysis.

Culture supernatants were obtained after centrifugation of samples or by immediate filtering through 0.45-μm-pore-size acetate filters (CAMEO 25GAS 0.45; Osmonics) and stored at −20°C. The samples were used for analysis of the residual substrate concentration and extracellular metabolite concentrations. Glucose, ethanol, acetate, glycerol, pyruvate, and succinate were analyzed by high-performance liquid chromatography (HPLC) using an Aminex HPX-87H column (Bio-Rad) (50). Ammonium, l-glutamine, and l-alanine concentrations were analyzed by ion exchange HPLC using a Waters 474 scanning fluorescence detector (3, 46). The same analysis was performed on the feed medium samples. Culture dry weight was determined via filtration and drying (14).

Transcript analysis.

Samples for RNA isolation were taken from the chemostat cultivations by rapidly sampling 20 ml of culture into a tube with 20 to 30 ml of crushed ice. Thereby, the temperature was lowered to below 4°C in less than 10 s. Cells were quickly pelleted (4,500 rpm at 0°C for 2 min), instantly frozen in liquid nitrogen, and thereafter stored at −80°C. Total RNA was extracted by using a FastRNA kit, Red (BIO 101, Inc., Vista, CA), after the samples were thawed on ice. mRNA extraction, cDNA synthesis, cRNA synthesis, and labeling were performed as described in the Affymetrix GeneChip expression analysis manual. Ten micrograms of cRNA was thereafter hybridized to the oligonucleotide array YG_S98 (Affymetrix, CA) as described in the Affymetrix GeneChip expression analysis manual. The result for each growth condition was derived from three independent chemostat experiments followed by transcript analysis.

Data acquisition.

The YG_S98 arrays were scanned with the GeneArray scanner, and the Affymetrix microarray suite v5.0 was used to generate CEL image files of the arrays. The array images were then normalized with the software dChip v1.2, and the transcript levels of all 9,335 probe sets were calculated with the perfect-match model in dChip v1.2 (25). The transcript levels of 6,091 unique open reading frames (ORFs) (ORFs defined by the Saccharomyces Genome Database) were extracted from the probe sets and used for further analysis.

ArrayExpress accession number.

The ArrayExpress accession number is E-MEXP-724.

Data analysis.

Pairwise significance analysis of microarrays (SAM) data analysis of all possible combinations of the four nutrient limitations was performed using the Microsoft Excel SAM add-in (v1.12) (45). Genes with a score greater than a threshold resulting in a false discovery rate of 1% were deemed potentially significant (45).

Cluster analysis was performed on the 2,466 transcripts that were found to alter transcript level under at least one of the four growth conditions (analysis of variance [ANOVA] on the original 6,091 transcripts, 5% significance level). The transcript profiles were clustered into groups of transcripts with similar profiles by use of the software ClusterLustre (19) (see the supplemental material and http://www2.cmb.dtu.dk/additional_material_for_publications/papers/6/index.html). Briefly, the Pearson's correlation coefficient was used as the metric, and the data set was clustered according to the formula k · r = 20 · 26 = 520 times; that is, k = 5… 30 clusters and 20 repetitions (r) for each cluster size (k), with the variational Bayes mixture of Gaussians. The likelihood of the Gaussian mixtures was used to estimate the number of clusters as 27. Clustering with Gaussian mixtures lead to different cluster solutions in each run, and therefore a consensus cluster solution was extracted from the most robust pattern in the initial 520 cluster runs. The 27 clusters were sorted and displayed in a dendrogram, and hence the similarities and differences between the different clusters could be studied. For instance, cluster 1 is more similar to cluster 2 than it is to cluster 3, etc.

Gene expression changes in response to nitrogen limitation were mapped on the genome-scale metabolic model of S. cerevisiae (17) in order to identify metabolic hot spots that significantly responded to the nitrogen limitation at the transcriptional level (35). The genome-scale model of yeast was first represented as a graph in which each metabolite is connected to all enzymes that catalyze a reaction involving that particular metabolite. Each enzyme involved in this graph was then scored based on the significance of the change in the expression level of the corresponding gene. This significance score was calculated by using a t test and transforming the resulting P value to a Z-score using the inverse normal cumulative distribution function. Each metabolite was assigned the average score of its k neighboring enzymes, and this score was then corrected for the background by subtracting the mean and dividing by the standard deviation of average scores of 10,000 enzyme groups of size k selected from the same data set. These corrected scores were then converted back to P values by using the normal cumulative distribution function. The 10 top-scoring metabolites were identified as reporter metabolites and selected for further analysis. Thus, the reporter metabolites are those around which transcriptional changes are significantly concentrated.

To uncover the global coordinated response of the genes spread across different metabolic pathways, the genome-scale model was converted to a graph of enzyme interactions. In this graph, only enzymes are represented as the nodes, and two enzymes sharing a common substrate in the corresponding reactions are connected to each other. Each enzyme in this graph was scored based on its altered transcription level as in the case of reporter metabolite identification. Subnetworks, defined as connected subgraphs of the enzyme interaction graph that showed maximum transcriptional response (quantified as the average background-corrected scores of all enzymes in that subnetwork), were then identified by using a simulated annealing algorithm (see the supplemental material or http://www2.cmb.dtu.dk/additional_material_for_publications/papers/6/index.html).

Promoter analyses were performed using the web-based software Regulatory Sequence Analysis Tools with default settings (http://rsat.ulb.ac.be/rsat/) (47). Promoters (from −800 bp to −1 bp) from genes in each cluster were analyzed for overrepresented hexanucleotides. Hexanucleotides sharing large common patterns were merged by alignment of elements with common patterns. Finally, the relative abundance of the consensus elements was determined by a query against genes in the relevant cluster and a query against genes in the entire genome (5,862 ORFs). Overrepresented gene ontology groups were defined with the Gene Ontology Term Finder (http://db.yeastgenome.org/cgi-bin/GO/goTermFinder) for the Saccharomyces Genome Database.


Physiological data.

To compare the physiological impacts of different nitrogen limitations, the prototropic S. cerevisiae strain CEN.PK113-7D was grown in a chemostat with ammonium, l-glutamine, or l-alanine as a sole and limiting nitrogen source. Chemostat experiments with the three different limitations were performed in triplicate, and HPLC analysis confirmed that the nitrogen source was limiting for growth in all the continuous cultures. The residual concentrations of l-alanine and l-glutamine were found to be below the detection level (<30 μM) in the cultures where they served as nitrogen sources. For the ammonia-limited chemostat, the residual ammonium concentration was found to be 50 μM, which is lower than the Km of the high-affinity ammonium permease Mep2 (42). The carbon source (glucose) was found to be in excess in all the chemostats (Table (Table11).

Physiological parameters of strain CEN.PK113-7D when cultivated in nitrogen-limited chemostatsa

Cells grown under amino acid limitations had biomass yields 17% and 19% higher per nitrogen mole than cells grown under ammonia limitation (Table (Table1).1). This indicated a more efficient use of the nitrogen source when cells were grown on amino acids and thereby a change in the nitrogen metabolism between ammonium- and amino acid-limited cultures. The excretion of by-products from the central carbon metabolism was also affected by the nitrogen source. More ethanol, succinate, and CO2 were produced by the amino acid-grown cells than by the ammonium-grown cells, while less glucose was consumed by the amino acid-grown cells (Table (Table11).

Consensus cluster analysis.

Aiming to gain insight into the events leading to the physiological changes, we next conducted a whole-genome transcript analysis of the nine nitrogen-limited cultures. To relate these results to conditions where nitrogen was in excess, we also included results from three glucose-limited cultures (μ = 0.2 h−1) where ammonium was in surplus.

An initial ANOVA of the transcript analysis revealed that 2,644 genes had altered transcript levels in at least one of the four conditions (P < 0.05) (see the supplemental material or http://www2.cmb.dtu.dk/additional_material_for_publications/papers/6/index.html). The cluster analysis revealed that the different nitrogen limitations affected as many transcripts (approximately 1,700; clusters 1 to 7, 11 to 21, and 24 [Fig. [Fig.1])1]) as the change to a nitrogen surplus (approximately 1,900; clusters 3 to 12 and 17 to 27 [Fig. [Fig.1]).1]). The largest differences between the nitrogen-limited cultures were found between amino acid limitation and ammonium limitation. In 14 of the clusters (clusters 2, 5 to 7, and 11 to 20) (Fig. (Fig.1),1), a total of 1,404 genes had expression profiles under amino acid limitations different from those of cells grown under ammonium limitation.

FIG. 1.
The transcript profiles of 2,644 genes with altered transcript profiles under glucose, ammonium, l-glutamine, or l-alanine limitation. Genes with similar transcript profiles were divided into 27 solid clusters according to the consensus-clustering algorithm ...

The most significant feature of transcripts with higher levels under amino acid limitation (Fig. (Fig.1,1, clusters 2 and 11 to 17) was the number of genes involved in chronological aging, cytoskeleton organization (65 transcripts), morphogenesis (39 transcripts), and regulation of the physiology (104 transcripts [20%]). These last-mentioned transcripts were particularly involved in the regulation of the M phase of the mitotic cell cycle, which includes chromosome segregation and cytoskeleton organization. Another fraction consisted of transcripts encoding specific transcription factors (56 transcripts), including MSN2, MSN4, MIG2, and GCR1, which are involved in the transcriptional induction of the TCA cycle and glycolytic genes (39), as well as GLN3, which encodes one of the two positive GATA transcription factors (7). Though increased levels of transcription factor mRNAs could potentially impose transcriptional effects on other genes, GATA factor targets or Msn2/Msn4 targets were not found to be enriched in these clusters. This observation was also supported by promoter analysis, which showed no overrepresentation of the NCR upstream activating sequence GATAAG or the Msn2/Msn4 upstream activating sequence AGGGG (Table (Table22).

Overrepresented sequences retrieved from the promoters of coregulated genes

Furthermore, several hundred genes were found to have lower transcript levels under amino acid limitation (clusters 5 to 7 and 18 to 20). This group was characterized by a weak overrepresentation of 4 genes involved in protein peroxisome targeting, of 25 in structural components of ribosomes, and of 38 involved in intracellular transport. These biological functions could all be important for a more efficient nitrogen utilization of cells grown on amino acids (Table (Table11).

Hence, the consensus cluster analysis together with promoter and gene ontology analysis revealed a large and complex transcript response to amino acid limitation but gave little information about the link to the physiological data (Table (Table1).1). We therefore used a new model-driven approach (35) to get a better understanding of the molecular mechanisms controlling the differential response to ammonium and amino acids.

Reporter metabolite analysis.

The bioinformatics tool used here is based on the hypothesis that expression of genes encoding enzymes is coordinated in order to satisfy the metabolic demands of the cell. This assumption implies that metabolic changes can be predicted through the transcriptional changes of genes encoding enzymes in the same metabolic pathway. The tool, reporter metabolite analysis, therefore identifies these groups of genes (subnetworks) rather than individual genes. Metabolites serving as substrate or product for the corresponding enzymes are called “reporter metabolites” and are identified as follows.

First, for each pair-wise comparison of studied conditions, the changes in transcript levels of all genes encoding enzymes are determined by using the classical t test. Metabolites are then scored based on the average P value (calculated from the t test) of their k neighboring enzymes. Thus, an enzyme contributes towards the scores of all metabolites that participate in the corresponding reactions. In order to test the statistical significance of the metabolite scores, we corrected the individual metabolite scores by using the mean and the standard deviations of the average scores of 10,000 randomly selected enzyme groups of size k. These corrected scores were then converted to P values by using the normal cumulative distribution. A high score is consequently equivalent to a low P value. The 10 metabolites with the lowest P values were identified as reporters. In order to explore a very large proportion of all known enzymes and metabolites from S. cerevisiae, we used a genome-scale metabolic model (17) with 1,175 reactions (708 gene products) and 584 metabolites.

The reporter metabolites indicate specific parts of metabolism where significant transcriptional regulation is exerted either to maintain homeostasis or to adjust the metabolite levels to altered demands or both. Similarly, enzyme subnetworks denote which groups of enzymes are functionally connected through metabolic networks and also respond significantly at the transcription level. These subnetworks usually span several metabolic pathways and thus offer a global view of the transcriptional response in a metabolic network (see the supplemental material or http://www2.cmb.dtu.dk/additional_material_for_publications/papers/6/index.html). The subnetworks, together with the reporter metabolites, thereby projected potential metabolic changes and hot spots in response to different nitrogen sources (Fig. (Fig.22 and Tables Tables33 and and44).

FIG. 2.
Yeast metabolic subnetworks affected by ammonium-, l-glutamine-, or l-alanine-limited growth conditions. Reporter metabolites and enzymes were identified by comparing transcript levels of 6,091 unique ORFs from l-alanine-limited cells versus ammonium-limited ...
Reporter metabolites identified in comparisons of two nitrogen limitationsa
Genesa encoding enzymes that synthesize the reporter metabolites listed in

Reporter metabolite analysis of the ammonium- versus l-glutamine- and l-alanine-limited cultures suggested that the main changes occur around the TCA cycle and the glyoxylate cycle (Fig. 2A and C). This was especially the case for l-alanine versus ammonium, where pyruvate, acetyl-CoA, glyoxylate, and α-ketoglutarate were identified as reporter metabolites (Fig. (Fig.2A2A and Table Table3).3). The flux from l-alanine to acetyl-CoA was potentially increased by the upregulation of the putative l-alanine-transaminase gene ALT1 (Fig. (Fig.2A).2A). Acetyl-CoA and glyoxylate could further be converted into malate via an increased level of the malate synthase, Dal7. The main function of the malate synthase is to increase anabolic activity and to make possible the synthesis of four and six carbons from the two-carbon level of acetyl-CoA.

A similar effect was found for the l-glutamine-limited cells compared to cells grown under ammonium limitation. Here, the glyoxylate cycle intermediates glyoxylate and malate were identified as reporter metabolites when l-glutamine served as the nitrogen source (Fig. (Fig.2C).2C). The latter analysis also suggested that changes in the nitrogen source influenced the metabolism of complex lipids, since intermediates in ergosterol biosynthesis (methylzymosterol and zymosterol) and in phospholipid metabolism (phosphatidylethanolamine) were identified as reporter metabolites (Table (Table33).

The model-driven data analysis also revealed large changes between the l-alanine-limited and the l-glutamine-limited growth conditions. In this case, the reporter metabolites suggested changes primarily in amino acid metabolism (Fig. (Fig.2B),2B), with little effect on the glyoxylate cycle. Most of the genes with higher transcript levels in the l-alanine-grown cells, as identified by enzyme subnetwork analysis, were involved in transaminase and deaminase reactions, a pattern that suggests a higher ammonium turnover in the l-alanine-grown cells. The reporter genes were primarily found in cluster 1 from the clustering analysis (ALT1, ASN1, BAT2, CPA2, DAL7, GDH2, PYC1, and YFR055W) (Fig. (Fig.3)3) where transcript levels were higher on l-alanine than on l-glutamine or ammonium.

FIG. 3.
Transcript profiles of clusters with NCR-responsive genes. The relative transcript levels of each gene (rows) under the four different growth conditions (columns), where green represents transcript levels between −1 and 0 and red represents transcript ...

Promoter analysis.

All in all, three clusters (1, 4, and 24 [Fig. [Fig.1])1]) identified 192 genes with changed transcript levels for l-alanine-grown cells. Cluster 1 had an overrepresentation of genes (49%) with the GATAAG NCR cis-acting element (Table (Table2),2), indicating that ammonium and l-glutamine could impose NCR, despite the low levels of these compounds in the medium. This result would also explain a majority of the changes in the reporter metabolite analysis between l-alanine-limited and l-glutamine-limited cells (Fig. (Fig.2A).2A). Genes in this cluster (91 genes) were involved in nitrogen and amino acid metabolism (ALT1, ASN1, BAT2, CHA1, CPA1, CPA2, DAL1, DCG1, ECM17, FOL1, GDH2, HOM3, LEU4, MCT1, PDA1, UGA1, and YFR055W), transport (DAL4, DIP5, MEP3, OAC1, SXM1, VMA7, and YHM1), and nucleotide metabolism (DAL7, DAL1, IMD1, IMD2, IMD3, and SDT1). Many of these genes have previously been described as targets for NCR.

However, it was surprising to find the classical NCR-sensitive genes in cluster 1, since the shape of this cluster indicated that l-glutamine and ammonium induced repression even though they are limiting for growth. NCR-sensitive genes have previously been found to be repressed in media where ammonium is in excess (7, 44). The current data revealed a second group of NCR-sensitive genes in clusters 8 to 10 (Fig. (Fig.1).1). These clusters represented genes with lower transcript levels in the glucose-limited cultures where ammonium was in excess. The current data therefore indicate that NCR-sensitive genes can be divided into two groups, one that responds to nitrogen in concentrations below 30 μM and another group that responds to nitrogen in the millimolar range.

The promoter analysis also revealed that the Mig1 core consensus element, ACCCC, the Ume6 consensus element, WGCCGCCGA, and the Cat8 consensus element, CCNNYNRNCCG, were overrepresented in clusters 25 to 27. A higher transcript level characterized these clusters when cells were grown under glucose-limiting conditions where glucose was in excess. This observation correlates with previously published results (6), where cells were found to be under Mig1 and Ume6 regulation when grown under nitrogen-limited conditions.


In the current paper, we have shown that cells grown with amino acids as the limiting nitrogen source have altered metabolism with higher biomass yields per nitrogen mole than cells grown under ammonium-limiting conditions. To understand these metabolic changes, we measured the global transcript levels and analyzed the results with two novel expression analysis tools developed in our laboratory (19, 35). The cluster analysis revealed 18 clear clusters of genes that had changed transcript levels in response to one of the nitrogen sources, containing a total of approximately 1,700 transcripts (Fig. (Fig.1,1, clusters 1 to 7 and 11 to 21). Another set of 21 clusters covered genes that had changed expression in response to glucose limitation (approximately 1,900 genes in clusters 3 to 12 and 17 to 27). Whole-genome transcript profiles have previously been compared for S. cerevisiae grown in glucose-limited culture and that grown in ammonium-limited culture (μ = 0.1 h−1) (6, 44). In this case, transcript profiles were sorted with SAM, which led to a total of 301 genes with different transcript levels for the two conditions. These genes were also identified in the current work (76% overlap of clusters 2 to 12 and 21 to 27). As the reader may note, these clusters contained a much larger number of transcripts than the ones identified by the SAM analysis. To investigate if the difference in the number of identified transcripts was due to the choice of analysis tool or the quality of the data in the two data sets, we conducted a similar SAM analysis (P < 0.01) with the current data. The SAM analysis identified 560 genes with altered transcript levels. The current consensus cluster analysis combined with ANOVA thus allowed us to extract more information from the transcript data set.

The consensus cluster analysis gave information about the coregulated transcripts (Fig. (Fig.11 and and3),3), while promoter analysis of the different gene clusters indicated possible means of regulation (Table (Table2).2). However, this analysis provided limited insight into the transcriptional changes causing altered metabolism under the different nitrogen limitations. This shortcoming is primary, due to the equal weight that transcript changes are given in cluster analysis, ignoring the fact that the gene products act in different metabolic pathways and different cellular compartments. One way to include this information is to integrate knowledge about metabolic pathways into the analysis of the transcript data. Such integration enabled us to systematically identify coregulated enzymatic subnetworks as well as the metabolic changes that could follow from a given change.

Altered anabolic activity in l-alanine-grown cells.

The enzyme subnetwork analysis suggests that l-alanine-grown cells have increased anabolic activity compared to l-glutamine- and ammonium-grown cells via the pyruvate carboxylase gene PYC1 and the malate synthase gene DAL7 (Fig. 2A and B), which may explain why l-alanine-grown cells have biomass yields higher than those of ammonium-grown cells. The higher biomass yields may also be supported by increased levels of pyruvate due to higher transcript levels of YFR055W and ALT1 in the l-alanine-grown cells. The enzyme subnetwork analysis also revealed that l-alanine-grown cells could have increased levels of the l-glutamate-degrading l-glutamate dehydrogenase transcript GDH2 and decreased levels of the citrate synthase transcript CIT2 (Fig. (Fig.2A).2A). Glutamate dehydrogenase and citrate synthase are key enzymes in the fixation of ammonium to amino acids, and altered levels of these enzymes may therefore also contribute to the higher biomass production in l-alanine-grown cells. CIT2 is induced by the Rtg2/Mks1/Rtg1/3 pathway, which ensures retrograde production of α-ketoglutarate (26), while GDH2 is repressed by NCR via GATA factors (5). Tate and Cooper have recently suggested a model in which ammonium regulates the level of l-glutamate by induction of CIT2 via Rtg2 and inhibition of GDH2 by NCR (44), a hypothesis that is in good agreement with the current enzyme subnetwork analysis (Fig. (Fig.2A).2A). Somewhat unexpectedly, both enzyme subnetwork analysis and cluster analysis showed that GDH2 and other NCR-sensitive genes have lower transcript levels under l-glutamine- and ammonium-limited conditions (Fig. (Fig.3,3, cluster 1), suggesting that concentrations below 50 μM are sufficient to induce NCR. This phenomenon is similar to those observed when cells are exposed to both high and low levels of glucose. In the first case, genes carrying the glucose-repressible Mig1 consensus element have transcript levels in glucose-limited cultures that are higher than those in cultures where glucose is in excess (6) (Fig. (Fig.1,1, clusters 21 to 27). A second group of Mig1-repressible genes appears to be derepressed only when glucose is completely abolished from the media, as observed when glucose-limited cultures are compared to cultures limited for the nonrepressible carbon sources acetate and ethanol (12).

Changes in the transport properties of nitrogen-limited cells.

This apparent ability to sustain repression at very low l-glutamine and ammonium levels in the media could be due to higher intracellular concentrations. Ammonium uptake and amino acid uptake are driven by the electrochemical gradient across the plasma membrane, and this mechanism can lead to differences between the intracellular concentration and that in the medium of several hundredfold (21, 36). The concentrations of intracellular ammonium and l-glutamine could thereby be held at a repressing level, provided that the corresponding high-affinity transporters are available in the plasma membrane.

The transcript analysis does indeed indicate that cells adapted to nitrogen limitations have changed the uptake of nitrogenous compounds. Previous studies have shown that to maximize the uptake of a limited nitrogen source, the cells have two options: one is to increase the expression of genes encoding high-affinity permeases (30), and the other is to change the transmembrane potential together with a change in the Km of the transporter (30, 33). The current study shows that the gene encoding the high-affinity ammonium permease, MEP2, has higher transcript levels under all three nitrogen limitations than it does under glucose limitation (Fig. (Fig.3,3, cluster 9). This is an observation that supports the first strategy and points to Mep2 being important in the uptake of ammonium under nitrogen-limited conditions (30).

The transcript levels of genes encoding the main l-alanine permeases GAP1 and PUT4 were found to be high for all nitrogen limitations (cluster 9), whereas the transcript level of a third l-alanine-permease gene, DIP5, was increased only under l-alanine-limiting conditions (cluster 1). This regulation is probably mediated by NCR, as both GAP1 and PUT4 are known to be under this regulation (9, 23), while DIP5 can be repressed by an NCR-inducing nitrogen source (38). Besides the genes encoding specific nitrogen transporters, the plasma membrane H+-ATPase gene, PMA1, also had a higher transcript level in the nitrogen-limited cultures (Fig. (Fig.3,3, cluster 10). Pma1 is the main H+-ATPase in the plasma membrane and is essential for the maintenance of the electrochemical gradient across the plasma membrane (41). A potential effect of the higher mRNA is therefore an increase in the membrane potential, suggesting that nitrogen-limited cells also apply the second strategy. According to the Nernst equation for electric potential, this opens the possibility of secondary active transport against a steeper gradient and thereby the uptake of lower concentrations of nitrogen through the nitrogen transporters. Another way to increase the plasma membrane potential is to lower the cytosolic H+ concentration through acidification of the vacuole. A number of transcripts encoding subunits of the vacuolar H+-ATPase (TFP1, STV1, VPH1, VMA5, VMA13, and VMA2) were increased in the nitrogen-limited cultures. However, it is questionable whether the higher transcript levels have a physiological effect, as the V1 domain and the V0 domain of the vacuolar H+-ATPase are known to dissociate and inactivate the ATPase at high glucose concentrations (34).

Supplementary Material

[Supplemental material]


This work was supported by grant 26-02-0098 from the Danish Technical Research Councils.

We thank Christoffer Bro for transcriptional data from carbon-limited chemostat cultivations, Lene Christiansen for technical assistance, and Mats Åkesson for assistance with the analysis of the DNA microarray data.


Supplemental material for this article may be found at http://aem.asm.org/.


1. Attfield, P. V. 1997. Stress tolerance: the key to effective strains of industrial baker's yeast. Nat. Biotechnol. 15:1351-1357. [PubMed]
2. Avendano, A., A. Deluna, H. Olivera, L. Valenzuela, and A. Gonzalez. 1997. GDH3 encodes a glutamate dehydrogenase isozyme, a previously unrecognized route for glutamate biosynthesis in Saccharomyces cerevisiae. J. Bacteriol. 179:5594-5597. [PMC free article] [PubMed]
3. Barkholt, V., and A. L. Jensen. 1989. Amino acid analysis: determination of cysteine plus half-cystine in proteins after hydrochloric acid hydrolysis with a disulfide compound as additive. Anal. Biochem. 177:318-322. [PubMed]
4. Beck, T., and M. N. Hall. 1999. The TOR signalling pathway controls nuclear localization of nutrient-regulated transcription factors. Nature 402:689-692. [PubMed]
5. Blinder, D., and B. Magasanik. 1995. Recognition of nitrogen-responsive upstream activation sequences of Saccharomyces cerevisiae by the product of the GLN3 gene. J. Bacteriol. 177:4190-4193. [PMC free article] [PubMed]
6. Boer, V. M., J. H. de Winde, J. T. Pronk, and M. D. Piper. 2003. The genome-wide transcriptional responses of Saccharomyces cerevisiae grown on glucose in aerobic chemostat cultures limited for carbon, nitrogen, phosphorus, or sulfur. J. Biol. Chem. 278:3265-3274. [PubMed]
7. Cooper, T. G. 2002. Transmitting the signal of excess nitrogen in Saccharomyces cerevisiae from the Tor proteins to the GATA factors: connecting the dots. FEMS Microbiol. Rev. 26:223-238. [PubMed]
8. Coschigano, P. W., S. M. Miller, and B. Magasanik. 1991. Physiological and genetic analysis of the carbon regulation of the NAD-dependent glutamate dehydrogenase of Saccharomyces cerevisiae. Mol. Cell. Biol. 11:4455-4465. [PMC free article] [PubMed]
9. Courchesne, W. E., and B. Magasanik. 1983. Ammonia regulation of amino acid permeases in Saccharomyces cerevisiae. Mol. Cell. Biol. 3:672-683. [PMC free article] [PubMed]
10. Crespo, J. L., T. Powers, B. Fowler, and M. N. Hall. 2002. The TOR-controlled transcription activators GLN3, RTG1, and RTG3 are regulated in response to intracellular levels of glutamine. Proc. Natl. Acad. Sci. USA 99:6784-6789. [PMC free article] [PubMed]
11. Cunningham, T. S., and T. G. Cooper. 1993. The Saccharomyces cerevisiae DAL80 repressor protein binds to multiple copies of GATAA-containing sequences (URSGATA). J. Bacteriol. 175:5851-5861. [PMC free article] [PubMed]
12. Daran-Lapujade, P., M. L. Jansen, J. M. Daran, W. van Gulik, J. H. de Winde, and J. T. Pronk. 2004. Role of transcriptional regulation in controlling fluxes in central carbon metabolism of Saccharomyces cerevisiae. A chemostat culture study. J. Biol. Chem. 279:9125-9138. [PubMed]
13. Didion, T., B. Regenberg, M. U. Jorgensen, M. C. Kielland-Brandt, and H. A. Andersen. 1998. The permease homologue Ssy1p controls the expression of amino acid and peptide transporter genes in Saccharomyces cerevisiae. Mol. Microbiol. 27:643-650. [PubMed]
14. Dynesen, J., H. P. Smits, L. Olsson, and J. Nielsen. 1998. Carbon catabolite repression of invertase during batch cultivations of Saccharomyces cerevisiae: the role of glucose, fructose, and mannose. Appl. Microbiol. Biotechnol. 50:579-582. [PubMed]
15. Ferea, T. L., D. Botstein, P. O. Brown, and R. F. Rosenzweig. 1999. Systematic changes in gene expression patterns following adaptive evolution in yeast. Proc. Natl. Acad. Sci. USA 96:9721-9726. [PMC free article] [PubMed]
16. Forsberg, H., and P. O. Ljungdahl. 2001. Sensors of extracellular nutrients in Saccharomyces cerevisiae. Curr. Genet. 40:91-109. [PubMed]
17. Forster, J., I. Famili, P. Fu, B. O. Palsson, and J. Nielsen. 2003. Genome-scale reconstruction of the Saccharomyces cerevisiae metabolic network. Genome Res. 13:244-253. [PMC free article] [PubMed]
18. Gasch, A. P., A. M. Moses, D. Y. Chiang, H. B. Fraser, M. Berardini, and M. B. Eisen. 2004. Conservation and evolution of cis-regulatory systems in ascomycete fungi. PLoS Biol. 2:e398. [PMC free article] [PubMed]
19. Grotkjaer, T., O. Winther, B. Regenberg, J. Nielsen, and L. K. Hansen. 2006. Robust multi-scale clustering of large DNA microarray datasets with the consensus algorithm. Bioinformatics 22:58-67. [PubMed]
20. Hinnebusch, A. G. 2005. Translational regulation of GCN4 and the general amino acid control of yeast. Annu. Rev. Microbiol. 59:407-450. [PubMed]
21. Horak, J. 1986. Possible role of histidine in the l-proline transport system of Saccharomyces cerevisiae. Biochim. Biophys. Acta 862:407-412. [PubMed]
22. Iraqui, I., S. Vissers, F. Bernard, J. O. De Craene, E. Boles, A. Urrestarazu, and B. Andre. 1999. Amino acid signaling in Saccharomyces cerevisiae: a permease-like sensor of external amino acids and F-box protein Grr1p are required for transcriptional induction of the AGP1 gene, which encodes a broad-specificity amino acid permease. Mol. Cell. Biol. 19:989-1001. [PMC free article] [PubMed]
23. Jauniaux, J. C., and M. Grenson. 1990. GAP1, the general amino acid permease gene of Saccharomyces cerevisiae. Nucleotide sequence, protein similarity with the other bakers yeast amino acid permeases, and nitrogen catabolite repression. Eur. J. Biochem. 190:39-44. [PubMed]
24. Klein, C. J. L., L. Olsson, and J. Nielsen. 1998. Nitrogen-limited continuous cultivations as a tool to quantify glucose control in Saccharomyces cerevisiae. Enzyme Microb. Technol. 23:91-100.
25. Li, C., and W. H. Wong. 2001. Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection. Proc. Natl. Acad. Sci. USA 98:31-36. [PMC free article] [PubMed]
26. Liu, Z., and R. A. Butow. 1999. A transcriptional switch in the expression of yeast tricarboxylic acid cycle genes in response to a reduction or loss of respiratory function. Mol. Cell. Biol. 19:6720-6728. [PMC free article] [PubMed]
27. Lorenz, M. C., and J. Heitman. 1998. The MEP2 ammonium permease regulates pseudohyphal differentiation in Saccharomyces cerevisiae. EMBO J. 17:1236-1247. [PMC free article] [PubMed]
28. Magasanik, B. 1992. Regulation of nitrogen utilization, p. 283-317. In E. W. Jones, J. R. Pringle, and J. R. Broach (ed.), The molecular and cellular biology of the yeast Saccharomyces. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.
29. Magasanik, B. 2005. The transduction of the nitrogen regulation signal in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 102:16537-16538. [PMC free article] [PubMed]
30. Marini, A. M., S. Soussi-Boudekou, S. Vissers, and B. Andre. 1997. A family of ammonium transporters in Saccharomyces cerevisiae. Mol. Cell. Biol. 17:4282-4293. [PMC free article] [PubMed]
31. Martin, D. E., A. Soulard, and M. N. Hall. 2004. TOR regulates ribosomal protein gene expression via PKA and the Forkhead transcription factor FHL1. Cell 119:969-979. [PubMed]
32. Mitchell, A. P., and B. Magasanik. 1984. Three regulatory systems control production of glutamine synthetase in Saccharomyces cerevisiae. Mol. Cell. Biol. 4:2767-2773. [PMC free article] [PubMed]
33. Opekarova, M., T. Caspari, B. Pinson, D. Brethes, and W. Tanner. 1998. Post-translational fate of CAN1 permease of Saccharomyces cerevisiae. Yeast 14:215-224. [PubMed]
34. Parra, K. J., and P. M. Kane. 1998. Reversible association between the V1 and V0 domains of yeast vacuolar H+-ATPase is an unconventional glucose-induced effect. Mol. Cell. Biol. 18:7064-7074. [PMC free article] [PubMed]
35. Patil, K. R., and J. Nielsen. 2005. Uncovering transcriptional regulation of metabolism by using metabolic network topology. Proc. Natl. Acad. Sci. USA 102:2685-2689. [PMC free article] [PubMed]
36. Pena, A., J. P. Pardo, and J. Ramirez. 1987. Early metabolic effects and mechanism of ammonium transport in yeast. Arch. Biochem. Biophys. 253:431-438. [PubMed]
37. Pretorius, I. S. 2000. Tailoring wine yeast for the new millennium: novel approaches to the ancient art of winemaking. Yeast 16:675-729. [PubMed]
38. Regenberg, B., L. During-Olsen, M. C. Kielland-Brandt, and S. Holmberg. 1999. Substrate specificity and gene expression of the amino-acid permeases in Saccharomyces cerevisiae. Curr. Genet. 36:317-328. [PubMed]
39. Rolland, F., J. Winderickx, and J. M. Thevelein. 2002. Glucose-sensing and -signalling mechanisms in yeast. FEMS Yeast Res. 2:183-201. [PubMed]
40. Schmitt, A. P., and K. McEntee. 1996. Msn2p, a zinc finger DNA-binding protein, is the transcriptional activator of the multistress response in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 93:5777-5782. [PMC free article] [PubMed]
41. Serrano, R., M. C. Kielland-Brandt, and G. R. Fink. 1986. Yeast plasma membrane ATPase is essential for growth and has homology with (Na+ + K+), K+- and Ca2+-ATPases. Nature 319:689-693. [PubMed]
42. Soupene, E., R. M. Ramirez, and S. Kustu. 2001. Evidence that fungal MEP proteins mediate diffusion of the uncharged species NH3 across the cytoplasmic membrane. Mol. Cell. Biol. 21:5733-5741. [PMC free article] [PubMed]
43. Stanbrough, M., D. W. Rowen, and B. Magasanik. 1995. Role of the GATA factors Gln3p and Nil1p of Saccharomyces cerevisiae in the expression of nitrogen-regulated genes. Proc. Natl. Acad. Sci. USA 92:9450-9454. [PMC free article] [PubMed]
44. Tate, J. J., and T. G. Cooper. 2003. Tor1/2 regulation of retrograde gene expression in Saccharomyces cerevisiae derives indirectly as a consequence of alterations in ammonia metabolism. J. Biol. Chem. 278:36924-36933. [PubMed]
45. Tusher, V. G., R. Tibshirani, and G. Chu. 2001. Significance analysis of microarrays applied to the ionizing radiation response. Proc. Natl. Acad. Sci. USA 98:5116-5121. [PMC free article] [PubMed]
46. Van Dijken, J. P., J. Bauer, L. Brambilla, P. Duboc, J. M. Francois, C. Gancedo, M. L. Giuseppin, J. J. Heijnen, M. Hoare, H. C. Lange, E. A. Madden, P. Niederberger, J. Nielsen, J. L. Parrou, T. Petit, D. Porro, M. Reuss, N. van Riel, M. Rizzi, H. Y. Steensma, C. T. Verrips, J. Vindelov, and J. T. Pronk. 2000. An interlaboratory comparison of physiological and genetic properties of four Saccharomyces cerevisiae strains. Enzyme Microb. Technol. 26:706-714. [PubMed]
47. van Helden, J., B. Andre, and J. Collado-Vides. 2000. A web site for the computational analysis of yeast regulatory sequences. Yeast 16:177-187. [PubMed]
48. Verduyn, C., E. Postma, W. A. Scheffers, and J. P. Van Dijken. 1992. Effect of benzoic acid on metabolic fluxes in yeasts: a continuous-culture study on the regulation of respiration and alcoholic fermentation. Yeast 8:501-517. [PubMed]
49. Wade, J. T., D. B. Hall, and K. Struhl. 2004. The transcription factor Ifh1 is a key regulator of yeast ribosomal protein genes. Nature 432:1054-1058. [PubMed]
50. Zaldivar, J., A. Borges, B. Johansson, H. P. Smits, S. G. Villas-Boas, J. Nielsen, and L. Olsson. 2002. Fermentation performance and intracellular metabolite patterns in laboratory and industrial xylose-fermenting Saccharomyces cerevisiae. Appl. Microbiol. Biotechnol. 59:436-442. [PubMed]

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Compound
    PubChem Compound links
  • MedGen
    Related information in MedGen
  • Protein
    Published protein sequences
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem Substance links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...