OUP user menu

Posttranscriptional regulation of FLO11 upon amino acid starvation in Saccharomyces cerevisiae

Claudia Fischer, Oliver Valerius, Heike Rupprecht, Marc Dumkow, Sven Krappmann, Gerhard H. Braus
DOI: http://dx.doi.org/10.1111/j.1567-1364.2007.00331.x 225-236 First published online: 1 March 2008


Various starvation conditions cause adhesive growth of haploid cells or pseudohyphae formation of diploid cells of Saccharomyces cerevisiae. For the genetic Σ1278b background, these morphological changes depend on the expression of the gene encoding the cell wall glycoprotein Flo11p, which is increased during nutritional limitations. Deletion of the genes encoding the transcriptional coactivators Rsc1p or Gcn5p impairs FLO11 transcription, which consequently leads to a loss of both haploid invasive growth and diploid pseudohyphae development upon glucose and nitrogen limitation, respectively. In contrast, amino acid starvation induces FLO11-dependent adhesive growth of the rsc1Δ and gcn5Δ strains although FLO11 transcription remains very low. The double deletion strain rsc1Δflo11Δ, however, does not grow adhesively, suggesting that the adhesion of the rsc1Δ strain at amino acid starvation is still FLO11-dependent. The FLO11prom-lacZ-encoded β-galactosidase activities of the rsc1Δ and gcn5Δ mutant strains increase manifold upon amino acid starvation. It is therefore concluded that low levels of FLO11 transcripts are essential and sufficient for derepression of FLO11 expression and adhesive growth during amino acid starvation. A posttranscriptional control is assumed to be behind this phenomenon that permits the increased FLO11 expression from low FLO11 transcript abundances.

  • amino acid starvation
  • adhesion
  • FLO11 expression
  • RSC1
  • HTZ1
  • translation


Starvation for nutrients triggers a variety of responses to keep cells alive. Nutrient sensing and adjustment of gene expression are essential to adapt to changing environmental conditions. Saccharomyces cerevisiae transmits e.g. information of utilizable glucose by two central signal transduction pathways, namely the Ras/Gpa2p/cyclic AMP (cAMP)/protein kinase A (PKA) and the Snf3p/Rgt2p/Yck pathway. Subsequent adjustment of cellular metabolism and gene expression are achieved by activating as well as repressing cellular processes (reviewed in Santangelo, 2006). During glucose limitation, the Snf1p/AMP-activated protein kinase family of protein kinases plays an essential role in the transcriptional and metabolic carbon catabolite repression and the adaptation to low carbon supply.

Several cellular systems regulate the use of nitrogen compounds in S. cerevisiae (Schneper et al., 2004). Ammonium as a nitrogen source is sensed by the high-affinity ammonium permease Mep2p localized at the plasma membrane. Amino acids, both for incorporation into protein and as sources of nitrogen in biosynthesis, are internally sensed by the ‘general control of amino acid biosynthesis’ (gc) (Hinnebusch, 2005). The nitrogen discrimination pathway is utilized in the absence of favourable ammonium sources such as glutamine or ammonium switching to sources like proline and allantoin. Therefore, mainly the transcription factor Gln3p acts as the master regulator for the respective transcriptome modulation.

Amino acid starvation is sensed in S. cerevisiae intracellularly by the kinase Gcn2p recognizing uncharged tRNAs and subsequently phosphorylating the α-subunit of the translation initiation factor eIF2 (Hinnebusch, 2005). Cap-dependent initiation of translation is thereupon reduced and only specific mRNAs become translationally active under such conditions. The GCN4 mRNA is a well-studied example for its derepression during amino acid starvation. It encodes a transcriptional key regulator modulating large parts of the transcriptome to adjust cells to periods of compromised protein biosynthesis. The target of rapamycin (TOR) pathway is another nutrient sensing system controlling cell growth and proliferation by responding to both cellular nitrogen and amino acid supplies (reviewed in Dann & Thomas, 2006; De Virgilio & Loewith, 2006; Wullschleger et al., 2006). TOR also regulates the gc by regulating the phosphorylation of a specific residue of Gcn2p that in turn determines eIF2α phosphorylation and translational activity.

Different nutrient sensing and signalling pathways have a strong impact on cellular differentiation of S. cerevisiae-like diploid pseudohyphae development or haploid adhesive growth (Pan et al., 2000; Cutler et al., 2001; Gagiano et al., 2002; Braus et al., 2003). Various limitations finally activate the transcription of the FLO11 (also referred to as MUC1) gene through their respective signalling pathways. FLO11 encodes a glycosylphosphatidylinositol-(GPI)-anchored cell wall adhesin that mediates tight cell–cell and cell–surface interactions (Lo & Dranginis, 1996). FLO11 expression is essential for adhesive as well as for pseudohyphal growth (Lo & Dranginis, 1998), which is reflected by one of S. cerevisiae's largest promoters and a complex transcriptional regulation (Gancedo, 2001).

Here, amino acid starvation-induced FLO11 expression is compared with glucose or nitrogen starvation-induced FLO11 expression and surprising differences were found. Saccharomyces cerevisiae strains lacking certain transcriptional coregulators as e.g. Rsc1p or Gcn5p are impaired in activating FLO11 transcription and, therefore, also fail in adhesive growth and pseudohyphae development at glucose and nitrogen limitation, respectively. In contrast, during amino acid starvation, the same strains still show very low FLO11 transcript levels but they nevertheless grow adhesively. The present data suggest an efficient posttranscriptional control for FLO11 expression.

Materials and methods

Yeast strains and growth conditions

All the yeast strains used in this study and their genotypes are listed in Table 1. Strains are derivatives of the S. cerevisiaeΣ1278b genetic strain background (Gimeno et al., 1992). Deletion mutants were generated using the Euroscarf deletion collection's BY strain (Brachmann et al., 1998) for PCR amplification of the respective deletion cassettes. Transformations were carried out using the lithium-acetate yeast transformation method (Ito et al., 1983). Integrations were confirmed by Southern analysis (Southern, 1975) and/or analytical PCR. The rsc1Δ/flo11Δ double-mutant strain was constructed by mating RH3184 (rsc1Δ∷kanMX) and RH3275 (flo11Δ∷NAT) and following tetrad dissection.

View this table:

Yeast strains used in this study

RH2817MATα, ura3-52, trp1hisGThis study
RH2656MATa/α,ura3-52/ura3-52, trp1hisG/TRP1Braus (2003)
RH2661MATa/α, ura3-52/ura3-52, trp1hisG/TRP1, Δflo11kanRflo11kanRBraus (2003)
RH2662MATa, ura3-52, Δflo11kanRBraus (2003)
RH3184MATα, ura3-52, trp1hisG, Δrsc1kanRThis study
RH3186MATa/α,ura3-52/ura3-52, trp1hisG/TRP1, leu2hisG/LEU2, Δrsc1kanRrsc1kanRThis study
RH3187MATα, ura3-52, trp1hisG, Δisw2kanRThis study
RH3189MATa/α,ura3-52/ura3-52, trp1hisG/TRP1, leu2hisG/LEU2, Δisw2kanRisw2kanRThis study
RH3190MATα, ura3-52, trp1hisG, Δgcn5kanRThis study
RH3192MATa/α,ura3-52/ura3-52, trp1hisG/TRP1, leu2hisG/LEU2, Δgcn5kanRgcn5kanRThis study
RH3214MATα, ura3-52, trp1hisG, Δhtz1kanRThis study
RH3216MATa/α,ura3-52/ura3-52, trp1hisG/TRP1, leu2hisG/LEU2, Δhtz1kanRhtz1kanRThis study
RH3217MATα, ura3-52, trp1hisG, Δssn6kanRThis study
RH3219MATa/α,ura3-52/ura3-52, trp1hisG/TRP1, leu2hisG/LEU2, Δssn6kanRssn6kanRThis study
RH3220MATα, ura3-52, trp1hisG, Δsfl1kanRThis study
RH3222MATa/α,ura3-52/ura3-52, trp1hisG/TRP1, leu2hisG/LEU2, Δsfl1kanRsfl1kanRThis study
RH3265MATα, ura3-52, trp1::hisG, Δrsc1kanR, Δflo11NATThis study
RH3275MATa, ura3-52, Δflo11NATThis study

Yeast strains were generally cultivated in YPD complete medium (1% yeast extract, 2% bacto peptone, 2% glucose) at 30 °C. For selection of gene disruption, 200 μg mL−1 geneticin/G418 was added. For wash tests, cells were grown on solid YPD medium (2% agar), and alternatively on synthetic minimal medium [0.15% yeast nitrogen base (YNB), 50 mM (NH4)2SO4, 200 mM myo-inositol, 2% glucose] supplemented with the appropriate amino acids. 3-aminotriazol was added to a final concentration of 10 mM to induce histidine starvation. Low-ammonium medium (SLAD; YNB medium containing 50 μM ammonium sulphate) was used to monitor pseudohyphal formation.


The plasmids used in this study are listed in Table 2. Deletion cassettes were amplified from the respective Euroscarf deletion strains (Brachmann et al., 1998), and inserted via TA-cloning (Zhou & Gomez-Sanchez, 2000) into pBlueskript II SK (Stratagene) using the restriction site EcoRV. An FLO11prom-lacZ reporter with mutated upstream ORF (uORF) was generated by first transferring the EcoRI–HindIII 3-kb fragment of plasmid pME2063 comprising the wild-type FLO11 promoter into pBlueskript II KS (+), yielding pME3008. Site-directed mutagenesis by PCR with oligonucleotides oHR3 (5′ ccgttctcttctgatgaggtaacctttac 3′) and oHR4 (5′ ggatttttgaggcctacaaaagtatattttaattagaaccacaacaagacgaggg 3′) resulted in a 420-bp product that changes the uORF start codon ATG to TTG. Using the BstEII and StuI restriction sites, this PCR product was cloned into pME3008, yielding pME3009. The mutated FLO11 promoter was retransferred into the pME2063 backbone using EcoRI and HindIII restriction, yielding the FLO11prom-lacZ reporter with mutated uORF in plasmid pME3010.

View this table:

Plasmids used in this study

pBIISK(+)Commercial cloning vectorStratagene, La Jolla
pME2632pBIISK(+) with integrated deletion cassette of RSC1 ORF (kanR)This study
pME2626pBIISK(+) with integrated deletion cassette of ISW2 ORF (kanR)This study
pME2628pBIISK(+) with integrated deletion cassette of GCN5 ORF (kanR)This study
pME2631pBIISK(+) with integrated deletion cassette of HTZ1 ORF (kanR)This study
pME2648pBIISK(+) with integrated deletion cassette of SSN6 ORF (kanR)This study
pME2653pBIISK(+) with integrated deletion cassette of SFL1 ORF (kanR)This study
B3782 (pME2063)3 kb FLO11 promoter fragment in YEp355Rupp (1999)
pME30083 kb EcoRI–HindIII flo11prom in pBIISK(+)This study
pME3009As pME3008 but FLO11 uORF-ATG mutated to TTGThis study
pME3010As B3782 but FLO11 uORF-ATG mutated to TTGThis study

Genetic crosses, sporulation and tetrad dissection

Diploid homozygous S. cerevisiae strains were obtained by mating of haploid yeast strains that had been verified by Southern hybridization experiments. Mating, sporulation and tetrad dissection were performed according to Sherman (1991).

Adhesive growth and pseudohyphal formation assays

Adhesive growth of haploids and diploids was performed as described previously (Roberts & Fink, 1994). Cells were patched on the respective medium and incubated for 1–5 days at 30 °C. Plates were photographed to show total growth and washed under a stream of water. The remaining adhesive cells were documented by photographing the washed plate.

Pseudohyphal growth after growing 5 days on SLAD medium at 30 °C was viewed using an Axiovert microscope (Carl Zeiss, Jena, Germany), and photographed with a Kappa DX30 digital camera and the Kappa Image Base software (Kappa Opta-Electronics, Gleichen, Germany).

RNA analysis

Total RNAs were isolated from exponentially growing yeast cells according to the method of Cross & Tinkelenberg (1991). For Northern hybridizations, 20 μg of total RNA was separated on 1.4% agarose gels containing 3% formaldehyde and subsequently transferred onto nylon membranes by capillary blotting. FLO11 and ACT1 transcripts were detected using specific 32P-radiolabelled DNA fragments with the Prime-It labelling kit from Stratagene (La Jolla). Signal intensities were visualized and quantified using a BAS-1500 phospho-imaging scanner (Fuji, Tokyo, Japan).

Quantification real-time PCR

Total RNA was isolated from yeast cells that were grown in YNB in the presence and absence of 3-AT. The quality of the RNA was checked using the Bioanalyzer 2100 from Agilent. Then, RNA samples were treated with DNAseI (Roche) at 37 °C for 20 min and, after checking the RNA quality again, cDNA synthesis was carried out using 1 μg RNA and the BioRad iScript cDNA Synthesis Kit according to the manufacturer's instructions. Additionally, the cDNA was purified via Qiagen gel extraction columns. Twenty nanograms of cDNA was used as a template for real-time PCR. The FLO11 gene was correlated to two reference genes ACT1 and CDC28 for normalization. The primers were designed using the Primer3 program (available at http://primer3.sourceforge.net/). The following primers were used: 5′ CTGGTCCAAAAGATACCGTCCAAC 3′ and 5′ ATGCATATTCAGCGGCACTACCTT 3′ for FLO11; 5′ GAATTGAGAGTTGCCCCAGAAGAA 3′ and 5′ AGAAACCAGCGTAAATTGGAACGA 3′ for ACT1; and 5′ CTTTTGGTGTTCCGTTGAGAGCTT 3′ and 5′ TCGATCTCACTATCGCCACTGAAG 3′ for CDC28.

The 16 μL real-time PCRs were performed on the iQ5 Cycler (BioRad) under the iQ SYBR Green Supermix Kit (BioRad). RT Q-PCR amplification was performed in triplicate using the following conditions: initial denaturation for 30 s at 98 °C, followed by 40 × cycle 2: denaturation (1 s at 94 °C), annealing (15 s at 64 °C) and extension (10 s at 72 °C). For melt curve data collection, these two cycles were followed by cycle 3 (15 s at 94 °C), cycle 4 (30 s at 64 °C) and cycle 5 (10 s at 59 °C; 67 × with increasing temperature after each cycle by 0.5 °C).

The specificity of the amplicons was confirmed by generating a melt curve profile of all amplified products. Gene expression was quantified on the basis of the threshold cycle values (Ct) at which a statistically significant increase in fluorescence intensity was first detected. The Ct-value for FLO11 was normalized to both reference genes within each sample, and the normalized mean values were compared with the sample containing wild-type mRNA without 3AT treatment, generating ΔΔCt values. The normalized level of target mRNA in the sample relative to the reference (wt-3AT) was expressed as 1+FΔΔCt, with E being the real PCR efficiency, which has been found to vary between 80% and 100%. The amplification efficiency was calculated using the slope of a standard curve using serial dilutions.

Transcriptional start site (TSS) determination

To determine the TSS of the FLO11 gene, the method of rapid amplification of cDNA ends (5′ RACE) was applied. Therefore, the GeneRacer kit of Invitrogen (CA) was used comprising SuperScript III reverse transcriptase and the TOPO TA cloning® kit for sequencing. Before 5′ RACE, total RNA was treated with RNAse-free DNAse to prevent amplification from any genomic DNA. 5′ RACE experiments were performed according to the supplier's manual using oHR5 (5′ cgttgtaaccgtatagttggacggtacc 3′) and oHR6 (5′ gccaattgaagtctaagttgggacagcc 3′) as the FLO11-specific and as the nested FLO11-specific reverse primer, respectively. For each strain and condition used, at least six independent clones were sequenced for cDNA end determination to account for the possible selection of alternative TSSs.

β-Galactosidase assay

Strains carrying the FLO11prom-lacZ reporter (B3782) or the uORF-mut-flo11prom-lacZ reporter (pME3010) were grown in selective media to the exponential growth phase. Extracts were prepared and assayed for specific β-galactosidase activity as described previously (Rose & Botstein, 1983) and normalized to the total protein, resulting in specific enzyme activities (Bradford, 1976).


Rsc1p and H2A.Z/Htz1p are chromatin-associated coregulators that affect FLO11 transcription

To study the impact of transcriptional coregulators on starvation-induced adhesive growth on solid surfaces, yeast strains with different gene deletions were used to monitor the expression of the major adhesin-encoding gene in haploid Σ1278b cells, the FLO11 gene. Among the deletion strains tested, the rsc1Δ, gcn5Δ, isw2Δ and ssn6Δ strains showed significantly reduced FLO11 mRNA levels as determined by Northern hybridizations from total RNA preparations of liquid YPD-cultures (Fig. 1). The compromised FLO11 transcription is especially pronounced in the rsc1Δ strain with only one-tenth of FLO11 transcript abundance in comparison with the wild-type strain, normalized against the internal ACT1 mRNA level. FLO11 mRNA abundance is approximately halved in the isw2Δ and ssn6Δ strains, and reduced by one-third in the gcn5Δ strain. In contrast, the deletion of the HTZ1 gene, encoding the histone variant H2A.Z/Htz1p, causes a twofold increase in FLO11 mRNA levels (Fig. 1). This is similar to the threefold FLO11 induction in the sfl1Δ strain lacking the FLO11 repressor protein Sfl1p, which serves as a control here (Robertson & Fink, 1998). The gcn5 deletion strain also serves as a control and has already been described as an FLO11 compromising background (Laprade et al., 2002). Deletion of SSN6 in DBY747 cells, however, has earlier been shown to induce FLO11 transcription (Conlan & Tzamarias, 2001). DBY747-derivative strains are of the mating type a, whereas the present data derive from haploid Σ1278b strains of mating type α. It could be confirmed that a deletion of SSN6 in Σ1278b cells of the a mating type also causes increased FLO11 transcript levels by factor 1.5 (data not shown). This suggests that the effect of SSN6 on FLO11 transcription is mating type dependent.


(a) Northern hybridizations with total RNAs of haploid yeast strains with deletions in coregulator-encoding genes grown in liquid YPD. mRNA levels of isw2Δ (RH3187), rsc1Δ (RH3184), gcn5Δ (RH3190), ssn6Δ (RH3217), sfl1Δ (RH3220) and htz1Δ (RH3214) were compared. Twenty micrograms of total RNAs from each sample were used. 32P-labelled FLO11 and ACT1 riboprobes were used for hybridization. Numbers given indicate the relative expression levels of FLO11 mRNA normalized to the ACT1 mRNA, averaged from six individual experiments with total RNA derived from three independent cultures. (b) Visualization of quantities by graph.

In summary, these observations reveal two genes not yet described as players in FLO11 expression: RSC1 as an activator and HTZ1 as a repressor gene for FLO11 transcription.

RSC1 is essential for glucose limitation-induced adhesive growth and for nitrogen limitation-induced pseudohyphae development

The set of deletion strains with altered FLO11 expression was then studied for adhesive growth after 5 days of growth on YPD agar medium by wash tests (Fig. 2a). Prolonged growth on rich medium results in the local decline of glucose concentrations, which triggers FLO11 transcription that causes subsequent adhesive growth (Cullen et al., 2000). In contrast to the wild-type strain, the deletion strains rsc1Δ, gcn5Δisw2Δ and ssn6Δ are no longer able to grow adhesively on agar upon glucose limitation (Fig. 2a, flo11Δ strain as negative control). This finding correlates with the reduced FLO11 mRNA levels observed from liquid cultures as described before. The requirement of GCN5 for FLO11-dependent adhesive growth also confirms the results published by Laprade (2002). In contrast to the isw2Δ adhesion deficiency presented here, Trachtulcova (2004) have shown earlier that the deletion of ISW2 can induce adhesion. However, they also showed that their induction was independent of FLO11, which might explain the different results. For cells of the Σ1278b background, the nonadhesiveness of isw2Δ cells correlated to lowered FLO11 mRNA abundance.


Analysis of coregulator deletions for FLO11-dependent growth phenotypes. (a) Strains were analysed for adhesive growth and patched on YPD medium, incubated and photographed after 5 days of growth (total growth), washed and rephotographed to document adhesive growth. Nonadhesive yeast cells were washed away from the agar surface. The following strains were used: wt (RH2817) and flo11Δ (RH2662) as controls, isw2Δ (RH3187), rsc1Δ (RH3184), gcn5Δ (RH3190), ssn6Δ (RH3217), sfl1Δ (RH3220) and htz1Δ (RH3214). (b) Haploid Saccharomyces cerevisiae strains with increased FLO11 mRNA levels were tested for early manifestation of adhesion. Haploid cells of sfl1Δ (RH3220), htz1Δ (RH3214) and the wt (RH2817) and flo11Δ (RH2662) were plated on YPD and cultivated for 1 day. Nonadhesive yeast cells were washed away. (c) Diploid homozygous S. cerevisiae strains were streaked out on nitrogen starvation (SLAD) plates to induce pseudohyphal growth and were incubated for 5 days at 30°C. Colony phenotypes were microscopically examined. wt/wt (RH2656) as control, isw2Δ/isw2Δ (RH3189), rsc1Δ/rsc1Δ (RH3186), gcn5Δ/gcn5Δ (RH3192), ssn6Δ/ssn6Δ (RH3219) and sfl1Δ/sfl1Δ (RH3222), respectively, were tested. The htz1Δ/htz1Δ (RH3216) strain was not included in the assay due to its strong growth defect on SLAD medium.

Severe de-repression of FLO11 transcription is known to cause adhesive growth after short growth periods (1 day) on rich solid agar medium. Wild-type cells are nonadhesive after only 1 day of growth, whereas FLO11-derepressed cells of the sfl1Δ background adhere significantly after short growth periods (Robertson & Fink, 1998). It was found here that the htz1Δ strain is also characterized by hyperadhesive growth (Fig. 2b), which corresponds to the increased FLO11 mRNA levels in this strain (Fig. 1).

Pseudohyphae formation of diploid S. cerevisiae cells during nitrogen limitation was then investigated on SLAD medium. Pseudohyphae formation was observable for isw2Δ/isw2Δ, ssn6Δ/ssn6Δ and sfl1Δ/sfl1Δ strains but not for the rsc1Δ/rsc1Δ and the gcn5Δ/gcn5Δ strains (Fig. 2c). Therefore, RSC1 seems to play a similar important role for FLO11-dependent yeast morphogenesis as it is known for Gcn5p (Laprade et al., 2002). The different impact of ISW2 and SSN6 on haploid adhesion and diploid pseudohyphal growth might arise either from the different signal of nutrient limitation, finally activating FLO11 (glucose vs. nitrogen) or from the difference in ploidity/gene copy number (haploid vs. diploid).

Although adhesive growth upon glucose limitation is not so strong for diploid cells compared with haploid cells, it can also be addressed by wash tests. The diploid set of strains revealed wild-type-like adhesive growth for the isw2Δ/isw2Δ and ssn6Δ/ssn6Δ strains but, again, prevented adhesion for the rsc1Δ/rsc1Δ and gcn5Δ/gcn5Δ strains (data not shown). This is in accordance to the observations described before for diploid pseudohyphal growth and corroborates the importance of RSC1 and GCN5 for FLO11 expression.

Activated FLO11 transcription during amino acid starvation depends on RSC1

Beyond glucose and nitrogen limitation, also amino acid starvation caused by the addition of histidine or tryptophan analogues is a strong signal for the induction of FLO11 expression and adhesive growth (Braus et al., 2003). The set of deletion strains for FLO11 transcript abundance at amino acid starvation was then analysed. Cells were therefore cultivated in liquid YNB medium in the presence of 10 mM 3-amino triazole (3-AT), a histidine analogon imposing false feedback inhibition, thereby causing histidine starvation. A fivefold increase in FLO11 mRNA abundance of wild-type cells observed upon amino acid starvation (Fig. 3, wt+3-AT) confirms earlier data and serves as a control for induction of starvation (Braus et al., 2003). Even during amino acid starvation, the rsc1Δ, isw2Δ and ssn6Δ strains express low FLO11 mRNA levels that are clearly below those of nonstarving wild-type cells (wt−3-AT). Again, the rsc1Δ strain shows the severest decline in FLO11 transcript abundance, giving rise to only very faint signals. The gcn5Δ and htz1Δ strains display slightly higher FLO11 mRNA levels than the nonstarving wild-type cells. From the Northern hybridization data, the induction ratios between growth in the presence and absence of amino acid starvation for each strain were calculated from the relative ACT1-normalized FLO11 intensities (relating to the wild-type strain at nonstarvation) and are given in Table3. The wild-type strain and the gcn5Δ strain increase the FLO11 mRNA abundance five- and threefold, respectively. However, measurable FLO11-induction upon amino acid starvation is prevented in the rsc1Δ, isw2Δ or ssn6Δ strains.


(a) FLO11 mRNA abundance during growth at amino acid starvation. Total RNAs were prepared from haploid yeast strains cultivated in liquid YNB supplemented with 10 mM 3-AT. Strains isw2Δ (RH3187), rsc1Δ (RH3184), gcn5Δ (RH3190), ssn6Δ (RH3217), sfl1Δ (RH3220), htz1Δ (RH3214) and the wt strain (RH2817) as control were analysed. Numbers given indicate the relative expression levels of FLO11 mRNA normalized to the ACT1 mRNA, averaged from six individual experiments with total RNA derived from three independent cultures. (b) Visualization of quantities by graph.

View this table:

Factor of transcriptional FLO11 activation upon amino acid starvation according to the Northern hybridization data

NonstarvationFLO11 (norm.)0.20.1<
Amino acid starvation10.1<
FactorFLO11 aa-st. FLO11 n-st.51ND30.810.8
  • aa-st., amino acid starvation; n-st., nonstarvation; norm., normalized against ACT1 as an internal standard; ND, not determined because of too low intensities.

FLO11-dependent adhesive growth at amino acid starvation is independent from RSC1

So far, the presented FLO11 mRNA abundance data obtained from Northern hybridizations correlate with the adhesiveness or the nonadhesiveness of the respective deletion strain. If this generally holds true, one would expect that the isw2Δ and the ssn6Δ strains, but especially the rsc1Δ strain, are also nonadhesive at amino acid starvation, as for these strains the FLO11 mRNA levels remain very low even in the presence of 3-AT (Fig. 3). Surprisingly, the rsc1Δ, isw2Δ and ssn6Δ strains all showed strong adhesive growth on agar plates with 3-AT after only 1 day of growth (Fig. 4a). It was then tested whether this effect was still FLO11 dependent. As an rsc1Δflo11Δ double-deletion strain is not adhesive anymore at amino acid starvation, at least low FLO11 mRNA levels are required for adhesive growth in the rsc1Δ strain (Fig. 4b). Similarly, adhesion of diploid cells upon amino acid starvation was very pronounced for the rsc1Δ/rsc1Δ or gcn5Δ/gcn5Δ strains (data not shown).


Adhesive growth of the deletion strains at amino acid starvation. (a) Haploid yeast mutant strains were grown on YNB medium in the presence of 10 mM 3-AT to induce amino acid starvation. The plates were incubated at 30°C for 1 day and photographed before (total growth) and after (adhesive growth) washing off the nonadhesive cells. As control, a wt (RH2817) and an flo11 deletion strain (RH2662) were used. The tested deletion strains were isw2Δ (RH3187), rsc1Δ (RH3184), gcn5Δ (RH3190), ssn6Δ (RH3217), sfl1Δ (RH3220) and htz1Δ (RH3214). (b) Haploid yeast strains RH2817 (wt), as well as RH2662 (flo11Δ) as controls, RH3184 (rsc1Δ) and RH3265 (rsc1Δflo11Δ) were patched on YNB medium containing 10 mM 3-AT. The plates were incubated at 30°C for 1 day and photographed before (total growth) and after (adhesive growth) washing off the nonadhesive cells.

In order to actually rule out a significant transcription-based FLO11 induction, the low FLO11 mRNA levels were double-checked in the rsc1Δ and gcn5Δ strains at amino acid starvation by quantitative PCR (qPCR). The resulting FLO11 mRNA abundance data were normalized against ACT1 and CDC28, and the numbers were then related to the uninduced wild-type strain (Fig. 5a). The severe FLO11 mRNA decrease and its poor activation upon amino acid starvation for the rsc1Δ and gcn5Δ strains as shown by the Northern hybridizations are well confirmed by the qPCR experiments performed with total RNA isolations from new independent cultures. The results suggest that very low levels of FLO11 mRNA are sufficient but essential to trigger adhesive growth during amino acid starvation, suggesting an efficient posttranscriptional control of FLO11 expression.


FLO11 expression upon 3-AT induction. (a) Relative FLO11 mRNA abundances as determined by quantitative PCR were measured in haploid yeast strains RH2817 (wt), RH3184 (rsc1Δ) and RH3190 (gcn5Δ) and normalized against ACT1 and CDC28. Experiments were performed from three independent cultures for each strain and condition. SD are indicated as error bars. ΔΔCt is the ratio of the normalized threshold cycle values of the samples to be compared. (b) β-galactosidase activities encoded from the FLO11-lacZ reporter allele. The expression of FLO11-lacZ reporter gene was measured in haploid yeast strains RH2817 (wt), RH3184 (rsc1Δ) and RH3190 (gcn5Δ) carrying the reporter plasmid B3782. Specific β-galactosidase activities were measured in the absence or in the presence of 10 mM 3-AT. The mean values of six independent cultures are presented, and the SD are indicated as error bars. SD did not exceed 20%.

Amino acid starvation activates FLO11prom-lacZ expression even in rsc1Δ and gcn5Δ strains

As a rigid cell wall protein with poor solubility in cell extracts, reliable quantitative assays for the accurate measurement of cellular Flo11 protein levels have, to the authors' knowledge, so far not been established. Alternatively, the β-galactosidase activities of an FLO11prom-lacZ reporter allele were therefore determined. By virtue of an authentic FLO11 promoter including its 5′-untranslated region (5′ UTR), the β-galactosidase readout mainly reflects the combined consequences of transcriptional and translational impact on FLO11-lacZ expression. The reporter fusion consists of a 3-kb FLO11 promoter including its 5′-untranslated region (5′ UTR) together with the first three amino acid-encoding codons of the Flo11 protein, followed, in frame, by the lacZ encoding sequence. As described earlier (Braus et al., 2003), the thus encoded β-galactosidase activities increase significantly upon amino acid starvation in Σ1278b cells by a factor of c. 34 (Fig. 5b). Expressing this reporter allele in the rsc1Δ strain at nonstarvation results in less than 10%β-galactosidase activity in comparison with the basal activity in wild-type cells. This was expected from the low FLO11 mRNA present in Northern hybridizations and the qPCR experiment. However, starting from this low basal FLO11prom-lacZ expression rate in rsc1Δ cells, amino acid starvation causes an c. 15-fold increase in β-galactosidase activity, and an c. 17-fold increase in gcn5Δ cells (Fig. 5b). This β-galactosidase activity increase is about 10-fold higher than the respective FLO11 mRNA inductions measured by qPCR (Fig. 5). The absolute β-galactosidase activity of rsc1Δ cells at amino acid starvation is about threefold higher than that of wild-type cells under nonstarvation conditions, and it is about sevenfold higher in amino acid-starved gcn5Δ cells.

This induction reflects the FLO11-dependent adhesiveness of the rsc1Δ and gcn5Δ strains during growth at amino acid starvation despite low FLO11 mRNA levels. Interestingly, also the opposite effect, which means high FLO11 mRNA abundance but abolished adhesive growth, has been observed for a gcn2Δ strain earlier. Despite highly activated FLO11 transcription in gcn2Δ cells at amino acid starvation, FLO11prom-lacZ reporter cells show low β-galactosidase activity, and the gcn2Δ strain does not grow adhesively (Braus et al., 2003). These earlier observations, together with the new data, support a posttranscriptional regulation of FLO11 at amino acid starvation that might function at the level of the mRNA translation.

Mutation of a putative uORF in the FLO11-5′ UTR does not significantly change the FLO11prom-lacZ reporter activity

cis-Elements in 5′ and 3′ UTRs are often involved in the posttranscriptional regulation of mRNAs. For the determination of potential regulatory elements in the 5′ UTR of the FLO11 mRNA, its TSSs were determined. Using of the rapid amplification of cDNA 5′ends approach (5′ RACE), FLO11 cDNA was generated using a gene-specific reverse primer. PCR amplification of the FLO11 cDNA and subsequent cloning of the PCR products, followed by DNA sequencing indicate FLO11 cDNA 5′ ends at positions −7, −11, −13, −17, −28, −30 and −36 for RNA derived from wild-type cells, and at positions −13, −28 and −30 for RNA derived from rsc1Δ cells, both strains cultured in both YNB with and without 3-AT (Fig. 6a). Amino acid starvation does not cause a differential use of TSSs. Using a forward primer that anneals upstream of a putative uORF and a gene-specific reverse primer, also a PCR product arises from cDNA, for which subsequent DNA sequencing confirmed the specificity of the FLO11 5′ UTR (Fig. 6b, lanes 10–13) in comparison with the controls (lanes 6–9). This hints at a rather small mRNA subpopulation including the putative uORF, which was not found among the 5′ end RACE experiments of more than 24 independent clones. To exclude a possible impact of the putative uORF in FLO11 expression at amino acid starvation, an FLO11prom-lacZ reporter plasmid with a mutated uORF-start codon was analysed. β-Galactosidase activities encoded from the uORF-mutated reporter are almost similar to those encoded from the wild-type uORF-FLO11prom-lacZ reporter in the absence or in the presence of starvation (Fig. 6c). Also, the expression of the uORF-mutated lacZ-reporter in the rsc1Δ strain is similar to that of the uORF-wild-type reporter. Therefore, the uORF itself does not seem to provide this posttranscriptional FLO11 mRNA control.


Transcriptional start site (TSS) determination of the FLO11 gene and analysis of the putative upstream ORF (uORF). (a) Arrows indicate TSSs as mapped by 5′ RACE relative to the actual ORF translational start site (+1). Bold arrows indicate cDNA 5′ ends found for at least six individual sequenced clones, and lean arrows represent 5′ ends found for only two individual sequenced clones. Oligonucleotide OLHR7 annealed immediately upstream of the putative uORF and was used as a forward primer for cDNA analysis shown in: (b) Each PCR was performed and loaded to the gel in replicate with a reverse primer annealing in the FLO11 ORF. Lanes 2–5: Total mRNA as a template for testing the RNA preparation for genomic DNA contamination. Lanes 6–9: DNAse-treated total mRNA as a template to ensure the absence of any genomic DNA contamination. Lanes 10–13: cDNA of DNAse-treated total mRNA (used as PCR template in 6–9) as a template. Lanes 14 and 15: positive control with genomic DNA, lane 16: negative control without a template. (c) β-galactosidase activities encoded from a mutated FLO11-lacZ reporter allele. Activity was measured for the haploid yeast strains RH2817 (wt), as well as for strain RH3184 (rsc1Δ) carrying either the wild-type FLO11-lacZ reporter plasmid B3782 (wt-uORF), the plain reporter plasmid YEp355 (empty) or the uORF-mutated reporter plasmid pME3010 (mut. uORF). Specific β-galactosidase activities were measured in the absence or in the presence of 10 mM 3-AT. The mean values of six independent cultures are presented, and the SD are indicated as error bars. SD did not exceed 20%.


Cellular morphological switches and differentiation programmes depend on the sensing and transduction of important nutritional signals, and subsequent reorganization of gene expression. The corresponding control mechanisms include a wide spectrum of means ranging from the regulation of transcription, of mRNA half-life and export, the regulation of mRNA translation, to posttranslational protein modifications and localizations. The expression of the FLO11 gene is known to be essential for both haploid adhesive growth and diploid pseudohyphae formation in Σ1278b cells of S. cerevisiae during glucose and nitrogen limitation, respectively (Lo & Dranginis, 1998). So far, in the literature, these morphological changes always correlate with an increase in FLO11 transcription upon those nutritional limitations (Gancedo, 2001; Verstrepen et al., 2003). Also, the epigenetic heritage of an active FLO11 expression status results in a higher level of gene transcription, and was shown to be dependent on the histone deacetylase activity of Hda1p (Halme et al., 2004). The impact of internal repeats within the FLO11 ORF and the variability of their number can influence the transcriptional activity of the gene and finally determine the level of FLO11 mRNA (Voynov et al., 2006).

Amino acid starvation induced by the addition of amino acid analogues strongly induces FLO11 transcription, FLO11-dependent adhesion and FLO11prom-lacZ-encoded β-galactosidase activities in wild-type Σ1278b cells. In this study, it is shown, that strains with low and almost uninducible FLO11 mRNA levels, such as the rsc1Δ and the gcn5Δ strains, still grow adhesively during amino acid starvation. The FLO11 dependency of this adhesion and the increased β-galactosidase activity encoded from an FLO11prom-lacZ reporter allele support a posttranscriptional induction of FLO11 expression upon amino acid starvation. Such a posttranscriptional regulation of FLO11 could take place at the level of initiation of mRNA translation. The paradigm for regulated initiation of translation in yeast is the derepression of the GCN4 mRNA during amino acid starvation as part of the general control of amino acid biosyntheses (gc) (Hinnebusch, 2005). The gc system senses uncharged tRNAs of the limited amino acid(s) through the sensor kinase Gcn2p, which subsequently phosphorylates the α-subunit of the elongation initiation factor eIF2, thereby changing the translational settings within starving cells by reducing the amount of ternary complex (TC) required for the cap-dependent initiation of translation. Unlike the bulk of repressed mRNAs during starvation, the GCN4 mRNA is by that means translationally de-repressed, which is specifically conferred by four uORFs in the GCN4 mRNA 5′ leader sequence. Although the present experiments argue for the existence of a subfraction of FLO11 mRNAs including the putative uORF, the major transcription start sites exclude this uORF from the FLO11 transcripts. Also, mutation of this uORF of an FLO11prom-lacZ reporter construct gave no evidence for an exclusive role of the uORF in FLO11 expression during amino acid starvation. Remarkably, a gcn2Δ strain, characterized by an abolished amino acid starvation response, shows neither FLO11-dependent adhesion nor significant FLO11prom-lacZ expression despite very high FLO11 mRNA levels at amino acid starvation (Braus et al., 2003). This suggests that effective FLO11 mRNA translation is favoured under conditions of generally discriminated cap-dependent mRNA translation that e.g. is amino acid starvation.

The expression of the mammalian cationic amino acid transporter Cat-1 is also elevated upon amino acid starvation. This is the consequence of both increased transcription and translation of the corresponding mRNA (Fernandez et al., 2003). The 5′ UTR-uORF-containing Cat-1 mRNA is suggested to be translationally activated by a mechanism that stalls ribosomes at the uORF, which in turn leads to efficient IRES-mediated translation (Fernandez et al., 2005). It has to be investigated whether mechanisms like ribosome stalling or IRES apply with the posttranscriptional FLO11 regulation and what regulatory elements in the mRNA 5′/3′ UTR and/or whether mRNA-binding proteins are involved. The recent publication of Gilbert et al. demonstrates that cap-independent translation of mRNAs of genes required for invasive growth (such as GPR1, MSN1 or FLO8) is essential for starvation-induced differentiation in yeast (Gilbert et al., 2007). It is shown there that IRES elements in the 5′ UTR of these genes account for the cap-independent translation. The investigations of Gilbert et al. focus on upstream regulators of FLO11 expression, whereas this study concentrates on the expression of the adhesin-encoding gene FLO11 itself. Therefore, it is proposed that not only the upstream regulator genes of FLO11 but also the FLO11 mRNA can be subject to a posttranscriptional (and may be translational) regulation. It will be interesting to investigate whether riboswitches, which are an emerging field in the control of mRNA translation from bacteria to eukaryotes (Brantl, 2004; Eisenstein, 2005), might be involved in FLO11 regulation. Considering cell–tissue adhesion or filamentous growth as initial steps in the pathogenicity of fungi as Candida albicans or Candida glabrata, such alternative regulation mechanisms might provide further clues for developing strategies to fight fungal infections.


The authors thank Verena Pretz and Gaby Heinrich for excellent technical assistance during the course of this study. The authors are grateful to Maria Meyer and Lars Fichtner for the help with tetrad dissections, and the Göttingen Genomics Laboratory for DNA sequencing. Many thanks are due to Hans-Ulrich Mösch, Malte Kleinschmidt, Katrin Bömeke and Krystyna Nahlik for critical discussions and helpful comments. This work was supported by grants from the Volkswagen-Stiftung, Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie.


  • Editor: Isak Pretorius


View Abstract