• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of molcellbPermissionsJournals.ASM.orgJournalMCB ArticleJournal InfoAuthorsReviewers
Mol Cell Biol. Sep 2003; 23(17): 6279–6290.
PMCID: PMC180942

Transcription Initiation of the Yeast IMD2 Gene Is Abolished in Response to Nutrient Limitation through a Sequence in Its Coding Region


The yeast IMD2 to IMD4 and GUA1 genes, involved in GMP synthesis, are highly expressed in exponentially growing cells but are shut off when cells cease to grow upon nutrient limitation. We show for the IMD2 gene that this effect is not specific to certain carbon sources or to growth rate. Strikingly, the cis elements responsible for this nutritional response are contained within a 23-nucleotide sequence in the coding region of the IMD2 gene. Despite its very unusual location, this regulatory sequence mediates the repression of transcription initiation. From our data, we conclude that GMP synthesis is downregulated upon nutrient limitation through an active mechanism. We show that this transcriptional shutoff abolishes any possibility of the induction of IMD2, even under drastic conditions of guanylic nucleotide limitation. Taken together, these results indicate that low levels of guanylic nucleotides could be required for proper entry into stationary phase.

GMP synthesis from IMP results from two successive enzymatic reactions catalyzed by IMP dehydrogenase (IMPDH) and GMP synthetase (Fig. (Fig.1A).1A). This metabolic pathway is highly conserved through evolution; sequencing of the Saccharomyces cerevisiae genome has revealed four homologs of IMPDH-encoding genes (IMD1 to IMD4), while the second step of the pathway requires the product of the GUA1 gene (6). Guanylic nucleotide level feedback regulates the transcription of the IMD genes (7). Indeed, the addition of guanine, a precursor of guanylic nucleotides, represses IMD2 transcription (7), while the addition of inhibitors of GMP synthesis, such as mycophenolic acid (MPA) or 6-azauracil, strongly induces IMD2 transcription (7, 25). A specific sequence, called the guanine response element (GRE), was identified in the promoter of IMD2 and found to be required for this regulation (7). Besides this specific regulation by the end product(s) of the pathway, IMD2 expression was also found to respond to growth phase. Indeed, microarray experiments revealed that the expression of IMD2 decreases when cells enter stationary phase (5, 9), and this result was confirmed by Northern blot analysis (26); however, the mechanism leading to this regulation has not been yet elucidated.

FIG. 1.
GMP synthesis is repressed upon nutrient limitation. (A) Metabolic pathway for GMP synthesis from IMP. Genes are shown in italic type and encode the following enzymatic activities: GUA1, guanosine-5′-monophosphate synthetase; and IMD2, IMD3, and ...

While the transcription of many genes decreases during stationary phase, understanding of this phenomenon is still incomplete. It is important to find out whether this regulation is the “passive” result of a general decrease in transcriptional efficiency or whether it indeed is an “active” regulation mechanism. It is clear that transcription is not simply abolished during stationary phase, since a subset of genes is induced during this phase (5, 9). Interestingly, mutations affecting the induction of YGP1, a gene induced by nutrient deprivation and entry into stationary phase, also affect the expression of ACT1, a gene repressed in stationary phase (1). Therefore, common mechanisms could regulate increased or decreased expression of specific genes in stationary phase in response to metabolic signals.

Because decreased GTP synthesis is critical for entry into stationary phase of Bacillus subtilis (21), we thought that the regulation of GMP synthesis genes in yeast cells could also be an important signaling step in this complex process. Importantly, in mammalian cells, the synthesis of IMPDH is tightly correlated to cellular proliferation. It is indeed well documented that quiescent or differentiated cells have low IMPDH levels, while dividing cells have high levels of IMPDH (see, for example, references 2, 14, 19, and 29 for reviews). Furthermore, the overexpression of IMPDH can bypass the antiproliferative effect of p53, indicating that correct control of IMPDH levels is essential for the antiproliferative role of p53 (16). We therefore studied the regulation of the yeast GMP synthesis pathway in response to growth phase.

In this report, we show that all of the genes involved in GMP synthesis are expressed less in stationary phase. We establish that the regulation of IMD2 by growth phase is indeed an active process requiring a regulatory sequence. We also show that this regulatory sequence, despite its very unusual location within the coding sequence, affects transcription initiation. Finally, we establish that this transcriptional downregulation in response to growth phase abolishes the induction of IMD2 transcription in the presence of MPA, therefore documenting the hierarchy between these two types of transcriptional regulation.


Strains and media.

SD medium is 2% glucose, 0.17% nitrogen base, and 0.5% ammonium sulfate. SDcasa medium is SD medium supplemented with 0.2% Casamino Acids (Difco). Uracil at a final concentration of 200 mg/liter was optionally added. SDcasaU medium is SDcasa medium supplemented with uracil.

Saccharomyces cerevisiae wild-type strain Y350 (MATa ura3-52 leu2-3,112 lys2Δ201) was used for all experiments except when otherwise indicated (see Fig. Fig.3B,3B, ,5C,5C, ,6C,6C, ,7A,7A, ,7D,7D, and and8B).8B). For some experiments (see Fig. Fig.3B),3B), wild-type strain LG812-2A (MATα ura3-52) was used. For other experiments (see Fig. Fig.5C,5C, ,7D,7D, and and8B),8B), we used strain MY3473, obtained by transforming into strain Y350 the BamHI digestion fragment of URA3 integration plasmid pMAC470B, constructed as described below. Strain MY3473 carries two tandem copies of the IMD2 gene, one complete copy with a wild-type sequence and one incomplete copy with a mutated nutrient-sensing element (NSE) (NSE1, mutated from TCTT to AATT at nucleotides 131 and 132 [mutations in boldface type], resulting in an L43K change in the protein sequence). Furthermore, an artificial HindIII site was introduced after nucleotide C155 downstream of the NSE, resulting in the insertion of a KL peptide in the protein sequence. For additional experiments (see Fig. Fig.6C),6C), wild-type strain BY4742 (MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0) and an isogenic gcr1-disrupted strain were used. This gcr1 haploid strain was obtained by sporulation of the GCR1/gcr1 diploid strain Y22753 (MATahis3Δ1/his3Δ1 leu2Δ0/leu2Δ0 lys2Δ0/LYS2 MET15/met15Δ0 ura3Δ0/ura3Δ0 YPL075w::kanMX4/YPL075w) and tetrad dissection in 2% lactate-1% glycerol medium. Both BY4742 and Y22753 were purchased from Euroscarf. For further experiments (see Fig. Fig.7A),7A), strain RY260 (MATa ura3-52 rpb1-1), created in the laboratory of Rick Young (20), was used.

FIG. 3.
The metabolic shutdown of IMD2 expression is a general response to any nutrient limitation. (A) Effect of carbon source on IMD2-lacZ expression. Wild-type strain Y350 transformed with the IMD2-lacZ plasmid P354 was grown in SDcasa medium with different ...
FIG. 5.FIG. 5.
Identification of cis NSEs in the coding sequence of the IMD2 gene. (A) Effect of clustered triple mutations in the region located at nucleotides 130 to 155 and required for IMD2 regulation by nutrient limitation. Wild-type strain Y350 was transformed ...
FIG. 6.
Assays of Abf1p and Gcr1p as trans-acting regulators of the IMD2 nutrient limitation response. (A) Schematic representation of the IMD2 sequence in the NSE region and comparison with the consensus binding elements of Gcr1p and Abf1p. pb, base pairs. (B) ...
FIG. 7.
The regulation of IMD2 expression in response to nutrient limitation is a transcription initiation event. (A) Analysis of IMD2 messenger stability in an rpb1-1 temperature-sensitive (rpb1-1ts) RNA polymerase II mutant strain. The rpb1-1 strain was grown ...
FIG. 8.
MPA cannot induce the expression of the IMD2 gene when it has been turned off in response to nutrient limitation. (A) Wild-type cells were grown from exponential to post-diauxic-shift phases and then were treated with the indicated amounts of MPA for ...


The following plasmids containing IMD2-lacZ fusions were described elsewhere (7): P354, P853, P855, and P890 (containing, respectively, 853, 442, 355, and 108 nucleotides upstream of the IMD2 ATG and 1,135 nucleotides of coding sequence fused to lacZ). All of the other IMD2-lacZ fusion-containing plasmids were constructed in the URA3 vectors described by Myers and coworkers (18) as follows. The desired IMD2 fragments were obtained by PCR amplification with the promoter-coding strand synthetic oligonucleotide IMP2-355 and the appropriate noncoding-strand synthetic oligonucleotide. The resulting PCR fragments were digested with BamHI-HindIII and cloned at the BamHI-HindIII sites of either YEp355 (for P1376 and the constructs with point mutations) or YEp356 (for P1204, P1611, P1335, and P1200). Plasmids containing 723, 155, 130, 95, and 3 bp of the IMD2 coding sequence and 355 bp upstream of the ATG were named P1204, P1376, P1611, P1335, and P1200, respectively, and were obtained with oligonucleotides Hind(+723), Hind(+155), Hind(+130), Hind(+95), and Hind(ATG), respectively. Plasmids containing triple AAA mutations in the +130 to +153 region of the IMD2 coding region, named P1807, P1950, P1803, P1801, P1949, P1799, P1798, and P1796, were obtained with oligonucleotides 155A151, 155A148, 155A145, 155A142, 155A139, 155A136, 155A133, and 155A130, respectively. Plasmid P2292 with the frameshifted sequence was obtained with oligonucleotide IMD2(−1).

Plasmid pMAC470B, used to create the endogenous tandem IMD2-mutated strain MY3473, was obtained by cloning into the EcoRI-HindIII sites of vector YIplac211 (URA3 integration vector) the EcoRI-BamHI-digested PCR product amplified with oligonucleotides 5′IMD2(−1730) and 3′IMD2(−312)Bam, the BamHI-HindIII fragment of plasmid P1796, and the HindIII-digested PCR product amplified with oligonucleotides 5′IMD2(+156) and 3′IMD2(1160).

The oligonucleotides used to PCR amplify the IMD2 fragments from the S288C genomic DNA template were as follows: coding strand—IMP2(−355), 5′ CGG GAT CCC GAA GAA AGC GGA AAA ATA A 3′; 5′IMD2(−1730), 5′ CCG GAA TTC AAC ATA TCC TTG CAA 3′; and 5′IMD2(+156), 5′ CCC AAG CTT GTC CTC TGA AGT TAG CC 3′; noncoding strand—Hind(+723), 5′ CCC AAG CTT TGG ACG CTA ATG GGT A 3′; Hind(155), 5′ CCC AAG CTT GCA AAA TCG ACT AAA CCT 3′; Hind(+130), 5′ CCC AAG CTT TTA AAA ATC GTT ATA AGT CA 3′; Hind(95), 5′ CCC AAG CTT TGG AGT CCA TCA ATT CC 3′; Hind(ATG), 5′ CCC AAG CTT CCA TTG CTT TTG CTA CT 3′; 155A151, 5′ CCC AAG CTT GCT TTA TCG ACT AAA CCT GG 3′; 155A148, 5′ CCC AAG CTT GCT AAT TTG ACT AAA CCT GG 3′; 155A145, 5′ CCC AAG CTT GCA AAA TCT TTT AAA CCT GGT AA 3′; 155A142, 5′ CCC AAG CTT GCA AAA TCG ACT TTA CCT GGT AAG AT 3′; 155A139, 5′ CCC AAG CTT GCA AAA TCG ACT AAT TTT GGT AAG ATT AA 3′; 155A136, 5′ CCC AAG CTT GCA AAA TCG ACT AAA CCT TTT AAG ATT AAA AA 3′; 155A133, 5′ CCC AAG CTT GCA AAA TCG ACT AAA CCT GGT AAG ATT AAA AAA TC 3′; 155A130, 5′ CCC AAG CTT GCA AAA TCG ACT AAA CCT GGT AAT TTT AAA AAA TCG TT 3′; IMD2(−1), 5′ CCC AAG CTT GCA AAT CGA CTA AAC CTG GTA AGA TTT AAA AAA TCG TT 3′; 3′IMD2(−312)Bam, 5′ CGG GAT CCT TTT TTT TTT TTA TTT TTT CGT TTT 3′; and 3′IMD2(1160), 5′ CAG TAG AAG AAC CAA GAG C 3′.

Plasmids expressing the Abf1p size variants were kind gifts from D. Gallwitz (10).

S1 nuclease assays.

Wild-type yeast strain Y350 was either grown overnight in SDcasaU medium to an optical density at 600 nm of 0.8 or allowed to grow until the postdiauxic shift phase. For some experiments (see Fig. Fig.5C),5C), the IMD2 tandem strain MY3473 was grown as described above but in SDcasa medium. For other experiments (see Fig. Fig.7A),7A), the rpb1-1 strain RY260 was grown as described above but in yeast extract-peptone-dextrose complete medium and was then transferred from 24 to 37°C. Subsequently, aliquots were removed at various times. In each experiment, 50 μg of total cellular RNA was analyzed for the presence of specific transcripts as previously described (3), except that hybridization was performed at 63°C and that S1 nuclease digestion was done with 1.5 times more enzyme. Each gene probe was concomitantly labeled along with the DED1 gene probe, used as an internal control in all experiments. The oligonucleotide used for DED1 has already been described (3), and those used for the IMD, GUA1, and ACT1 genes were as follows: IMD1(S1), 5′ CCG ATA CCA TCC AAC AGA GCC ATA AAA GTG GCC ATT TCA GAT TCC GTC ACA GTG TCC ATT GGC CG 3′; IMD2(S1) (probe D), 5′ CCG ATA CCA CCC AAC AGA GCC ATA AAA GTG GCC ATT TCT GAC TCT GTC ACC GTG TCC ATT GGC CG 3′; IMD3(S1), 5′ CCG ATA CCA CCC AAC AAA GCC ATG AAG ATG GCC ATT TCT GAT TCT GTC ACG GTG TCC ATT GGC CG 3′; IMD4(S1), 5′ CCG CCA ATA CCA CCC AAT AAA GCC ATA TAA ATA GCC ATA TCA GCT TCA GTG ACT GTG TCC ATA GG 3′; GUA1(S1), 5′ CCA AAG TCT AAC ACT AAG ATA GTG TCA AAC ATG TTA GAA ACT TGT TCA CCG GCA GCC CG 3′; ACT1(S1), 5′ CGA GCA ATT GGG ACC GTG CAA TTC TTC TTA CAG TTA AAT GGG ATG GTG CAA GCG CGC C 3′; NSE wt S1 (probe A), 5′ GAG GAC GCA AAA TCG ACT AAA CCT GGT AAG ATT AAA AAA TCG TTA TAA GTG TTG 3′; NSE A130 S1 (probe B), 5′ AAG CTT GCA AAA TCG ACT AAA CCT GGT AAT TTT AAA AAA TCG TTA TAA GTG TTG 3′; and IMD2S1avantNSE (probe C), 5′ CCA TCC GGT CTT GGT AGG CTC TTG GTA AAG TCT AGT GCG GTC TTG TAG TCT CTA ATG GCC G 3′

β-Galactosidase assays.

Cells grown in medium selective for the reporter plasmids were collected at various times to measure expression from the IMD2-lacZ fusions as previously described (7). Units of β-galactosidase activity are optical density at 420 nm × 1,000/(optical density at 600 nm × time [in minutes] × volume [in milliliters]). For some experiments (see Fig. Fig.5B5B and and6C),6C), transformed cells were grown overnight in SDcasa medium or in SDcasa medium with 2% lactate and 1% glycerol, respectively, either to exponential phase or until the post-diauxic-shift phase.

Band shift assays.

Extracts of Escherichia coli strain DH5α transformed with plasmids expressing Abf1p or the Abf1Δ551-728p size variant were prepared and used to perform band shift assays as described previously (10). The control (ADE5,7) 81-bp probe containing the Abf1p DNA binding site was obtained by PCR with yeast genomic DNA in the presence of 5 μl of [α-32P]dATP (400 Ci/mmol) and with the following oligonucleotides: 125, 5′ CGC CCC GTC GGT AG 3′; and 126, 5′ AGT TCA AGC CCA TCG C 3′. The IMD2 106-bp probe containing the NSE was obtained by the same procedure with the following oligonucleotides: 448, 5′ CAA GAT CAG AAA AGG GTT GAC TT 3′; and 256, 5′ CCC AAG CTT TGG AGT CCA TCA ATT CC 3′.

ChIP analysis.

Chromatin immunoprecipitation (ChIP) assays were performed as described previously (4) with Y350 cultures at various growth phases. For cultures in exponential phase (see culture point 1 in Fig. Fig.7C7C and culture points 1 and 3 in Fig. Fig.7D),7D), 100 ml of the cell culture was taken; for cultures in the post-diauxic-shift phase (see culture points 2 and 3 in Fig. Fig.7C7C and culture points 2 and 4 in Fig. Fig.7D),7D), 30 ml of the cell culture was taken. One of the IMD2 real-time PCR amplifications (see Fig. Fig.7C)7C) was performed with oligonucleotides 5′GREbis (5′ CAA AAT TAT TGG TTT TCG TAA CC 3′) and 3′GREbis (5′ GCC ACC AAG ATT CTC CGG T 3′). Another IMD2 real-time PCR amplification (see Fig. Fig.7D)7D) was performed with oligonucleotides 5′GREtris (5′ AAA AAA AAG TAT AAA TAG TGA AGA CTT 3′) and 3′GRE (5′ CAG CAA AAA ACT TCC AGC CG 3′).


Transcriptional expression of IMD2 responds to nutritional state but not to growth rate.

The expression of IMD2 was previously reported to decrease at the diauxic transition, both by microarray analyses and by Northern blot experiments (5, 9, 26). To further examine this phenomenon, we first assayed the expression of individual IMD genes in exponential or post-diauxic-shift cultures. We used S1 nuclease protection analyses, which allowed us to discriminate among the transcripts originating from the four IMD genes that are highly homologous (Fig. (Fig.1B).1B). The DED1 transcript (metabolic and technical control) was clearly regulated by growth phase, as previously reported (15), while the ACT1 transcript (loading control) was not, as expected, since its expression is affected only after several days of growth (1). IMD1 expression could not be detected, establishing that it is most likely a pseudogene, as previously suspected from lacZ fusion analyses (7). The four other genes involved in GMP synthesis from IMP (Fig. (Fig.1A)1A) were clearly expressed much more during exponential growth than after the diauxic shift, establishing that all of the genes required for GMP synthesis strongly respond to growth phase.

Similar results were obtained with an IMD2-lacZ fusion, showing that the level of IMD2-lacZ expression is low under conditions of slow growth, i.e., post-diauxic-shift and lag phases (Fig. (Fig.1C).1C). Since cells in post-diauxic-shift phase divide slowly due to nutrient limitation, it was critical to elucidate whether the low level of IMD2 expression was due to decreased growth rate (increased doubling time) or to nutrient limitation. This question was examined by modulating doubling time under conditions where nutrients were not limiting. When yeast cells were grown in SDcasa medium at 22 and 30°C, IMD2-lacZ activities were exactly the same during exponential growth, despite very different doubling times (253 versus 124 min at 22 and 30°C, respectively) (Fig. (Fig.2A).2A). In both cultures, the decrease in IMD2-lacZ expression coincided with the bending of the growth curve and was not correlated with the growth rate. We next studied yeast cells grown in SD medium (not rich) and SDcasa medium (rich). Under these conditions, growth rates were clearly different (doubling times were 243 and 124 min, respectively), but growth ceased simultaneously, although at very different cell densities (Fig. (Fig.2B).2B). In these media, IMD2-lacZ expression followed exactly the same patterns of expression (Fig. (Fig.2B).2B). We conclude that the repression of IMD2-lacZ expression is a response not to growth rate but to nutrient availability. This conclusion might explain why cells in lag phase have low IMD2-lacZ activity. Lag phase results from dilution of stationary-phase cells from a preculture into new growth medium; such cells are in the process of metabolic adaptation to their new growth conditions and could still behave as if they were nutritionally limited cells. Indeed, when experiments were performed with precultured cells that had not reached stationary phase, no lag was observed and IMD2-lacZ expression was high (Fig. (Fig.3A3A).

FIG. 2.
The metabolic shutdown of IMD2 expression is a response to nutrient limitation and not to doubling time. Wild-type strain Y350 transformed with the IMD2-lacZ plasmid P354 was grown in either SD medium or SDcasa medium at 30 or 22°C, conditions ...

The response of IMD2 to nutritional state is not specific to carbon limitation.

A major alteration occurring in the bending part of the growth curve in glucose medium is the shift from fermenting to respiratory metabolism. A simple explanation for the decreased IMD2 expression in this part of the growth curve could be that IMD2 expression is high under fermentative growth conditions and low when yeast cells are grown under respiratory conditions. Our results obtained with yeast cells grown in the presence of either glucose, galactose, ethanol, or glycerol clearly showed that the expression of IMD2-lacZ was high under exponential-growth conditions whatever the carbon source and decreased at entry into stationary phase (Fig. (Fig.3A).3A). Therefore, the signal for IMD2 regulation is not specifically the exhaustion of glucose or the increased concentration of ethanol in the medium but rather appears to be a nutrient limitation-sensing mechanism. We next examined the specificity of the nutrient limitation signal leading to the decreased expression of IMD2-lacZ. Limitation of any of the three components of SD medium (glucose, ammonium sulfate, or yeast nitrogen base) led to early entry into stationary phase and to concomitant decreased expression of IMD2-lacZ (Fig. (Fig.3B).3B). Therefore, decreased IMD2 expression is not a specific response to carbon limitation but rather is a general response to nutrient limitation.

We next wondered whether previously described signaling pathways involved in nutrient sensing (see references 12, 23, and 28 for reviews) participated in IMD2 regulation. Several mutants making the Ras-protein kinase A pathway either constitutively active (RAS2Val19, ira2, pde1, and pde2) or inactive (ras1, ras2, tpk1, tpk2, and tpk3) were tested, and no effect of these mutations on the repression of IMD2-lacZ expression at the post-diauxic shift could be found (data not shown). Mutations affecting the TOR or GCN4 pathways did not have any effect either (data not shown). We conclude that these previously described general pathways are not fully responsible for the nutrient-sensing regulation of IMD2.

Sequences required for the nutritional regulation of IMD2 are within the coding region.

While it is clear that the expression of many yeast genes (such as GMP synthesis genes) decreases after the diauxic shift, it is not yet known whether this decreased expression is the result of a general decrease in the activity of the transcription apparatus or is the result of an active regulatory process. Should such an actively controlled process exist, sequences in cis and trans factors would be required for this regulation. We therefore sought to identify cis elements in the IMD2 gene that are important for the nutritional limitation response.

We first took advantage of a set of deletions in the promoter region of IMD2 upstream of the 5′ transcription start previously mapped at position −106 upstream of the start codon (7). The results presented in Fig. Fig.4A4A show that deletion of the entire promoter region of the IMD2 gene (P890) did not affect the regulation of the expression of the IMD2-lacZ fusion constructs in response to nutritional status. Therefore, the metabolic regulation of the IMD2 gene is not dependent upon the promoter region. We therefore investigated whether regulatory sequences could lie in the coding region of IMD2. Indeed, a fusion of lacZ just downstream of the IMD2 initiation codon resulted in the deregulated expression of IMD2-lacZ (compare P1200 and P1204 in Fig. Fig.4B).4B). Finally, the regulatory region could be delimitated to 25 nucleotides between +130 and +155 in the coding region (compare P1611 and P1376 in Fig. Fig.4C).4C). To investigate more precisely this region, clustered mutations were introduced by site-directed mutagenesis and assayed for their effects on IMD2-lacZ expression. The results show that the 25-nucleotide region can be separated into three subregions. Mutations in the region from +145 to +153 led to partial deregulation (Fig. (Fig.5A,5A, top panel), while mutations in the central region (+136 to +144) did not affect regulation (Fig. (Fig.5A,5A, middle panel). Finally, mutations affecting nucleotides +131 to +135 strongly affected the regulation process (Fig. (Fig.5A,5A, bottom panel), while further mutations between nucleotides +100 and +131 had no major effect (data not shown). These two critical regions were designated the NSE.

FIG. 4.
The cis region responsible for the metabolic shutdown of IMD2 expression is located in the coding region. (A) Effect of deletions in the IMD2 promoter on nutrient limitation regulation. Wild-type strain Y350 was transformed with the indicated IMD2-lacZ ...

Mutations in the NSE affected both the nucleotide sequence and the translated fusion. Thus, it was not clear from this experiment whether the regulatory sequence was required as a DNA element for transcriptional regulation or could be involved as a specific peptide sequence in destabilizing the fusion protein. We therefore designed a fusion in which the regulatory sequence was still present but in a frame different from that in the original construct, leading to a totally different translation product (Fig. (Fig.5B).5B). It was clear that this new fusion was regulated by nutritional state (Fig. (Fig.5B),5B), establishing that the NSE sequences operate at the nucleic acid and not the protein level.

Finally, the role of the NSE was confirmed by S1 nuclease protection experiments with a yeast strain carrying two tandem copies of IMD2, one copy with the wild-type NSE in its coding sequence and the other with a mutated form of the NSE (Fig. (Fig.5C).5C). Due to the mutation in the NSE region, the expression of both copies could be monitored simultanously by using specific oligonucleotides. It was clear that only the wild-type NSE could allow correct post-diauxic-shift regulation, confirming that the NSE sequence in the coding region of IMD2 was responsible for the decreased expression of IMD2 mRNA in response to nutritional limitation (Fig. (Fig.5C).5C). Furthermore, the mutated NSE copy led to increased levels of basal mutant IMD2 expression compared to wild-type IMD2 expression (Fig. (Fig.5C),5C), an observation also made for the equivalent lacZ fusions (compare P1376 and P1796 in Fig. Fig.5A5A).

From these results, we conclude that a critical regulatory cis element for IMD2 expression resides within the coding sequence. A search for homology between the NSE sequences and known regulatory sequences revealed that the NSE sequences are close to the consensus sequences for two DNA binding proteins, Abf1p and Gcr1p (Fig. (Fig.6A).6A). We therefore tested whether these proteins could be involved in the regulation of IMD2 expression. Since ABF1 is an essential gene, we could not assay directly the effect of the lack of Abf1p on IMD2 expression in vivo. We thus examined the binding of Abf1p to IMD2 in vitro. The binding of Abf1p to an IMD2 probe could not be detected (Fig. (Fig.6B),6B), while the same protein could bind to the control ADE5,7 probe (24). Finally, IMD2 expression was not affected in a gcr1 mutant strain (Fig. (Fig.6C).6C). We conclude that ABF1 and GCR1 are probably not major players in IMD2 regulation. Band shift experiments were then performed with a labeled NSE and total cell extracts prepared before and after the diauxic shift. Although these experiments revealed the presence of a protein(s) that could associate with the NSE specifically in extracts from post-diauxic-shift phase cells (data not shown), the affinity of this factor(s) for the NSE was low and could easily be competed for with large excesses of nonspecific DNA, so we did not pursue this finding further.

Nutritional limitation regulation of IMD2 occurs at the transcription initiation level.

Because of the unusual location of the regulatory NSE within the coding sequence, we sought to determine whether this sequence could affect late steps of IMD2 transcriptional expression rather than transcription initiation. An obvious reason for decreased IMD2 expression in the post-diauxic-shift phase could be that the NSE would somehow trigger IMD2 transcript instability under nutritional limitation conditions. We used an RNA polymerase II temperature-sensitive mutant to stop transcription either in exponential growth phase or in post-diauxic-shift phase. S1 nuclease protection assays carried out with RNA extracted at various times after the temperature shift revealed that, while the IMD2 transcript was less abundant in post-diauxic-shift phase, it was not less stable (Fig. (Fig.7A7A).

Since IMD2 expression is affected in several transcription elongation mutants in the presence of drugs affecting GMP metabolism (25), we suspected that the NSE sequence could affect transcription elongation, for example, by inducing pausing during the elongation process. Should this hypothesis be true, pausing of the transcriptional apparatus at the NSE might result in a population of shorter RNAs containing only sequences 5′ to the NSE. We therefore carried out S1 nuclease protection assays with two different oligonucleotide probes, one hybridizing upstream of the NSE (probe C) and the other one hybridizing downstream (probe D; same as that used in the experiment shown in Fig. Fig.1B)1B) (Fig. (Fig.7B).7B). No major difference could be found between signals with probes C and D, either in exponential-phase or in post-diauxic-shift phase cells (Fig. (Fig.7B).7B). These results suggest that the nutritional regulation of IMD2 does not lead to the accumulation of incompletely elongated transcripts, although the possibility that incompletely elongated transcripts might not be stable enough to be detected cannot be ruled out.

Finally, transcription initiation was investigated by measuring the in vivo association of the TATA box binding protein (TBP) with the IMD2 promoter under various growth conditions. We clearly found a tight correlation between IMD2 expression, as measured by S1 nuclease protection, and TBP association with the IMD2 promoter, as estimated by ChIP (Fig. (Fig.7C).7C). Furthermore, the importance of the NSE sequence for regulated TBP binding in response to growth phase was further established by showing significant increased occupancy of the IMD2 promoter by TBP in an NSE mutant strain in stationary phase (Fig. (Fig.7D).7D). Therefore, we conclude that the low level of expression of IMD2 after the post-diauxic shift is the result of an inhibitory effect on one of the steps leading to transcription initiation.

Nutritional limitation regulation through the NSE abolishes the guanylic nucleotide-specific response.

It was previously shown that the IMD2 gene was transcriptionally regulated in response to guanylic nucleotide availability (7), i.e., repressed in the presence of extracellular guanine and induced in the presence of MPA, a drug that decreases the guanylic nucleotide pool. Since nutritional limitation and MPA have antagonistic effects on IMD2 expression, we challenged yeast cells with both signals and found that, in a wild-type strain, the nutritional limitation regulation clearly prevented induction by MPA (Fig. (Fig.8A).8A). The same experiment carried out with a strain with two tandem copies of IMD2, one with a wild-type NSE and the other with a mutated NSE, revealed that the two copies behaved very differently. The wild-type copy was not induced by MPA after the post-diauxic shift (Fig. (Fig.8B).8B). In contrast, the IMD2 gene copy that was mutated in the NSE sequence was only poorly repressed at the diauxic shift. Some residual repression could still be observed in post-diauxic-shift phase and was most likely due to the NSE point mutation used in the construct which, unlike the total NSE deletion, does not lead to complete derepression (compare P1796 and P1611 in Fig. Fig.5).5). Most importantly, the mutant copy of IMD2 was still inducible by MPA (Fig. (Fig.8B).8B). Thus, IMD2 repression by metabolic downregulation precludes induction by MPA.


Nutrient limitation results in the decreased expression of many yeast genes. While it seems intuitively obvious that nondividing cells should require smaller amounts of newly synthesized proteins than actively dividing cells, the mechanisms leading to this regulation are not yet understood. Whatever the mechanism, it seems clear that decreased expression of these genes is not a passive process resulting from lower efficiency of the transcriptional apparatus. Indeed, while the expression of many genes decreases during stationary phase, the expression of some genes is enhanced, demonstrating that the transcriptional apparatus is still operational. Furthermore, loss-of-function mutations in genes, such as SRB10 or SRB11, encoding RNA polymerase II-associated factors, lead to increased expression of the ACT1 gene in stationary phase (1), and we show here that cis mutations in the IMD2 gene lead to a high level of expression of this gene under nutrient limitation conditions. Therefore, at least for some genes, cis sequences and trans factors are required for active negative regulation in response to nutrient limitation.

We have clearly established that IMD2 downregulation in post-diauxic-shift phase cells was not the result of the limitation of a specific nutrient and was not affected by mutations in previously characterized pathways involved in general nutritional responses. Moreover, it was not correlated to the cell doubling time, in good agreement with previously published results showing that yeast cells grown in rich or poor media have similar IMPDH activities (22). Therefore, what is the signal for the downregulation of IMD2 at the post-diauxic-shift stage? We cannot rule out the possibility that the IMD2 regulation that we have studied, rather than being a specific response to nutrient limitation, is a more general response to growth versus growth arrest. Indeed, challenging yeast cells with drugs impairing growth, such as the DNA-damaging agent methyl methanesulfonate or rapamycin, led to downregulation of the IMD2 mRNA (11, 13); sudden changes in temperature had a similar effect (9). It would be interesting to evaluate whether all of these forms of regulation result from a common mechanism.

Strikingly, a sequence required for proper regulation by nutrient limitation (the NSE) was found in the coding sequence of the IMD2 gene. Internal regulatory sequences have been described for higher eukaryotes (for a review, see reference 17) but are rather unusual in yeast genes. To our knowledge, there are only two documented cases in yeast genes: the internal regulatory sequences correspond to Rap1p and Abf1p binding sites in the SRP1 and LPD1 genes, respectively (8, 27). Strikingly, these positively acting transcription factors regulate the ribosomal genes which are transcriptionally repressed by the metabolic shutdown (5, 9). However, in the case of IMD2, our results do not suggest a major role for Abf1p and Gcr1p in the regulation, and a potential Rap1p binding site upstream of the NSE was shown not to affect regulation when mutated (data not shown). There may be factors that bind to the NSE specifically after the diauxic shift, as suggested by our band shift experiments, but the affinity of these factors for the NSE under the experimental conditions that we used was quite low. Thus, it is still unclear whether this finding is relevant for the function of the NSE in vivo.

Importantly, this work documents the role of a cis-acting element in the coding region of a yeast gene on an active regulatory process. We show that an “intracoding” sequence affects TBP binding to the promoter in vivo and most probably regulates transcriptional initiation. Strikingly, the NSE was still functional in a promoterless construct probably expressed from a cryptic promoter encoded by vector sequences (Fig. (Fig.4A);4A); thus, its action on transcription initiation might not have been highly specific to the IMD2 promoter. However, the NSE sequence inserted in the coding sequence of an ADE17-lacZ construct did not result in post-diauxic-shift regulation of the fusion (data not shown), showing that the NSE sequence is necessary but not sufficient for this regulation. Interestingly, we found that the NSE could be moved from +130 to +95 in the IMD2 coding region without affecting regulation, while placing it just downstream of the ATG (+3) totally abolished the post-diauxic-shift regulation (data not shown). The effect of the long distance between the NSE and the promoter of the IMD2 gene suggests that this growth arrest regulation could occur through an alteration of chromatin structure. However, it is also possible that it functions through the binding of a specific transcription factor; this point remains to be clarified. An attractive model is that by entry into stationary phase, the IMD2 gene adapts a “closed” chromatin structure, which would reduce TBP binding and shut down its expression, even precluding its induction by MPA. It will be interesting to determine whether the shutdown of regulation by nutrient limitation through intracoding region sequences is a rule or an IMD2 exception. A search for regions similar to the NSE in IMD3, IMD4, and GUA1 that could account for post-diauxic-shift regulation of these genes revealed that they all contain the NSE1 sequence in their coding regions. However, no NSE2 sequences could be found. Interestingly, a genome-wide search for NSE sequences, performed with the PatMatch program of the SGD database, revealed only two other yeast genes (GCD11 and YER039C-A) carrying NSE-like sequences, with only one mismatch, and the expression of GCD11 was reported to be lower (2.5-fold) after a diauxic shift (5). Further experiments will be needed to determine whether the regulation of these genes is similar to that of IMD2.

Why should IMD2 expression be tightly regulated under conditions of nutritional limitation? An appealing explanation would be that the synthesis of guanylic nucleotides produced from IMP by IMPDH and GMP reductase should decrease to allow correct entry into stationary phase. Indeed, GTP is a major factor in transcription, translation, and signal transduction; consequently, nutrient deprivation would result in the need for the cell to reduce its energy output in all of these metabolic areas. Such an important role for GTP has been documented for B. subtilis. Indeed, in this bacterium, low GTP levels are required for the correct regulation of several genes involved in stationary-phase establishment (21). For yeast cells, there is not yet any such evidence, but the expression of all of the genes required for GMP synthesis is clearly transcriptionally downregulated during nutrient limitation (Fig. (Fig.1B).1B). Moreover, in mammalian cells, there is a tight relationship between IMPDH activity and the replicative state of the cell, and the p53 antiproliferative role depends on a correct control of IMPDH levels (16). We believe that the nutrient response of the GMP biosynthesis genes could well be an important signal for stationary-phase establishment rather than just a consequence of growth arrest. Indeed, multiple attempts to integrate a full-length copy of an NSE-mutated IMD2 gene failed, strongly suggesting that it could be deleterious for the cell (data not shown). Finally, the strongest argument for an important role of this regulation comes from the fact that it prevails over the specific regulation of IMD2 by guanylic nucleotides through the GRE in the promoter. Indeed, the GRE, by modulating IMD2 expression in response to guanylic nucleotide supply, should be sufficient to adapt GTP synthesis in response to growth requirements. The fact that another form of regulation is present and furthermore prevails strongly suggests that the decreased synthesis of guanylic nucleotides is required for adaptation to nutrient limitation. Further manipulation of the metabolic flux through the guanylic nucleotide synthesis pathway should allow the investigation of this attractive hypothesis.


We thank D. Gallwitz and W. Hörz for yeast strains and plasmids.

This work was supported by grants from the CNRS (UMR5095) and ARC (4749) to B.D.-F., by a fellowship to M.E.-H. from the Portuguese Government (FCT SFRH/BPD/5725/2001), and by a grant (3100-059199.99) from the Swiss National Science Foundation to M.A.C.


1. Chang, Y. W., S. C. Howard, Y. V. Budovskaya, J. Rine, and P. K. Herman. 2001. The rye mutants identify a role for Ssn/Srb proteins of the RNA polymerase II holoenzyme during stationary phase entry in Saccharomyces cerevisiae. Genetics 157:17-26. [PMC free article] [PubMed]
2. Collart, F. R., C. B. Chubb, B. L. Mirkin, and E. Huberman. 1992. Increased inosine-5′-monophosphate dehydrogenase gene expression in solid tumor tissues and tumor cell lines. Cancer Res. 52:5826-5828. [PubMed]
3. Collart, M. A., and K. Struhl. 1993. CDC39, an essential nuclear protein that negatively regulates transcription and differentially affects the constitutive and inducible HIS3 promoters. EMBO J. 12:177-186. [PMC free article] [PubMed]
4. Deluen, C., N. James, L. Maillet, M. Molinete, G. Theiler, M. Lemaire, N. Paquet, and M. A. Collart. 2002. The Ccr4-Not complex and yTAF1 (yTAFII130p/yTAFII145p) show physical and functional interactions. Mol. Cell. Biol. 22:6735-6749. [PMC free article] [PubMed]
5. DeRisi, J. L., V. R. Iyer, and P. O. Brown. 1997. Exploring the metabolic and genetic control of gene expression on a genomic scale. Science 278:680-686. [PubMed]
6. Dujardin, G., M. Kermorgant, P. P. Slonimski, and H. Boucherie. 1994. Cloning and sequencing of the GMP synthetase-encoding gene of Saccharomyces cerevisiae. Gene 139:127-132. [PubMed]
7. Escobar-Henriques, M., and B. Daignan-Fornier. 2001. Transcriptional regulation of the yeast GMP synthesis pathway by its end products. J. Biol. Chem. 276:1523-1530. [PubMed]
8. Fantino, E., D. Marguet, and G. J. Lauquin. 1992. Downstream activating sequence within the coding region of a yeast gene: specific binding in vitro of RAP1 protein. Mol. Gen. Genet. 236:65-75. [PubMed]
9. Gasch, A. P., P. T. Spellman, C. M. Kao, O. Carmel-Harel, M. B. Eisen, G. Storz, D. Botstein, and P. O. Brown. 2000. Genomic expression programs in the response of yeast cells to environmental changes. Mol. Biol. Cell 11:4241-4257. [PMC free article] [PubMed]
10. Halfter, H., B. Kavety, J. Vandekerckhove, F. Kiefer, and D. Gallwitz. 1989. Sequence, expression and mutational analysis of BAF1, a transcriptional activator and ARS1-binding protein of the yeast Saccharomyces cerevisiae. EMBO J. 8:4265-4272. [PMC free article] [PubMed]
11. Hardwick, J. S., F. G. Kuruvilla, J. K. Tong, A. F. Shamji, and S. L. Schreiber. 1999. Rapamycin-modulated transcription defines the subset of nutrient-sensitive signaling pathways directly controlled by the Tor proteins. Proc. Natl. Acad. Sci. USA 96:14866-14870. [PMC free article] [PubMed]
12. Hinnebusch, A. G. 1997. Translational regulation of yeast GCN4. A window on factors that control initiator-tRNA binding to the ribosome. J. Biol. Chem. 272:21661-21664. [PubMed]
13. Jelinsky, S. A., and L. D. Samson. 1999. Global response of Saccharomyces cerevisiae to an alkylating agent. Proc. Natl. Acad. Sci. USA 96:1486-1491. [PMC free article] [PubMed]
14. Konno, Y., Y. Natsumeda, M. Nagai, Y. Yamaji, S. Ohno, K. Susuki, and G. Weber. 1991. Expression of human IMP dehydrogenase types I and II in Escherichia coli and distribution in human normal and leukemic cell lines. J. Biol. Chem. 266:506-509. [PubMed]
15. Lemaire, M., J. Xie, M. Meisterernst, and M. A. Collart. 2000. The NC2 repressor is dispensable in yeast mutated for the Sin4p component of the holoenzyme and plays roles similar to Mot1p in vivo. Mol. Microbiol. 36:163-173. [PubMed]
16. Liu, Y., S. A. Bohn, and J. L. Sherley. 1998. Inosine-5′-monophosphate dehydrogenase is a rate determining factor for p53-dependent growth regulation. Mol. Biol. Cell 9:15-28. [PMC free article] [PubMed]
17. Müller, M. M., T. Gerster, and W. Schaffner. 1988. Enhancer sequences and the regulation of gene transcription. Eur. J. Biochem. 176:485-495. [PubMed]
18. Myers, A. M., A. Tzagoloff, D. M. Kinney, and C. J. Lusty. 1986. Yeast shuttle and integrative vectors with multiple cloning sites suitable for construction of lacZ fusions. Gene 45:299-310. [PubMed]
19. Nagai, M., Y. Natsumeda, and G. Weber. 1992. Proliferation-linked regulation of type II IMP dehydrogenase gene in human normal lymphocytes and HL-60 leukemic cells. Cancer Res. 52:258-261. [PubMed]
20. Nonet, M., C. Scafe, J. Sexton, and R. Young. 1987. Eucaryotic RNA polymerase conditional mutant that rapidly ceases mRNA synthesis. Mol. Cell. Biol. 7:1602-1611. [PMC free article] [PubMed]
21. Ratnayake-Lecamwasam, M., P. Serror, K. W. Wong, and A. L. Sonenshein. 2001. Bacillus subtilis codY represses early-stationary-phase genes by sensing GTP levels. Genes Dev. 15:1093-1103. [PMC free article] [PubMed]
22. Reichert, U., and M. Winter. 1975. Gene dosage effects in polyploid strains of Saccharomyces cerevisiae containing gua-1 wild-type and mutant alleles. J. Bacteriol. 124:1041-1045. [PMC free article] [PubMed]
23. Rohde, J., J. Heitman, and M. E. Cardenas. 2001. The TOR kinases link nutrient sensing to cell growth. J. Biol. Chem. 276:9583-9586. [PubMed]
24. Rolfes, R. J., F. Zhang, and A. G. Hinnebusch. 1997. The transcriptional activators BAS1, BAS2, and ABF1 bind positive regulatory sites as the critical elements for adenine regulation of ADE5,7. J. Biol. Chem. 272:13343-13354. [PubMed]
25. Shaw, R. J., and D. Reines. 2000. Saccharomyces cerevisiae transcription elongation mutants are defective in PUR5 induction in response to nucleotide depletion. Mol. Cell. Biol. 20:7427-7437. [PMC free article] [PubMed]
26. Shaw, R. J., J. L. Wilson, K. T. Smith, and D. Reines. 2001. Regulation of an IMP dehydrogenase gene and its overexpression in drug-sensitive transcription elongation mutants of yeast. J. Biol. Chem. 276:32905-32916. [PMC free article] [PubMed]
27. Sinclair, D. A., G. D. Kornfeld, and I. W. Dawes. 1994. Yeast intragenic transcriptional control: activation and repression sites within the coding region of the Saccharomyces cerevisiae LPD1 gene. Mol. Cell. Biol. 14:214-225. [PMC free article] [PubMed]
28. Werner-Washburne, M., E. Braun, G. C. Johnston, and R. A. Singer. 1993. Stationary phase in the yeast Saccharomyces cerevisiae. Microbiol. Rev. 57:383-401. [PMC free article] [PubMed]
29. Zimmermann, A. G., J. J. Gu, J. Laliberte, and B. S. Mitchell. 1998. Inosine-5′-monophosphate dehydrogenase: regulation of expression and role in cellular proliferation and T lymphocyte activation. Prog. Nucleic Acid Res. Mol. Biol. 61:181-209. [PubMed]

Articles from Molecular and Cellular Biology 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...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...