Physiological and morphological effects of genetic alterations leading to a reduced synthesis of UDP-glucose in Saccharomyces cerevisiae


  • Jean Marc Daran,

    1. Centre de Bioingenierie Gilbert Durand, UMR-CNRS 5504, Laboratoire associé INRA, Complexe Scientifique de Rangueil, 31077 Toulouse Cedex 04, France
    Search for more papers by this author
    • 1

      Manchester Biotechnology Centre, UMIST, P.O. Box 88, Manchester, UK.

  • Walter Bell,

    1. Centre de Bioingenierie Gilbert Durand, UMR-CNRS 5504, Laboratoire associé INRA, Complexe Scientifique de Rangueil, 31077 Toulouse Cedex 04, France
    Search for more papers by this author
  • Jean François

    Corresponding author
    1. Centre de Bioingenierie Gilbert Durand, UMR-CNRS 5504, Laboratoire associé INRA, Complexe Scientifique de Rangueil, 31077 Toulouse Cedex 04, France
    Search for more papers by this author

Corresponding author. Département de Génie Biochimique et Alimentaire, Institut National des Sciences Appliquées, Complexe Scientifique de Rangueil, 31077 Toulouse Cedex 04, France. Tel.: +33 (5) 61559492; Fax: +33 (5) 61559400; E-mail:


Yeast cells lacking UDP-Glc pyrophosphorylase (UGPase) encoded by UGP1 are not viable. Two strategies were developed to drastically reduce the intracellular concentration of UDP-Glc in order to study the consequences of this metabolic engineering on physiology and morphology. Firstly, UGP1 was placed under the strongly regulatable THI4 promoter. This resulted in a 95% reduction of UGPase activity in the presence of thiamine. The phenotypic effects of this reduction were slightly stronger than those of glucose on the GAL10/CYC1-UGP1 gene fusion [Daran et al. (1995) Eur. J. Biochem. 230, 520–530]. A further reduction of flux towards UDP-Glc was achieved by deletion of the two phosphoglucomutase genes in the ugp1 conditional strain. The growth of this new mutant strain was hardly affected, while it was extremely sensitive to cell wall interfering drugs. Surprisingly, UDP-Glc levels were reduced only by 5-fold, causing a proportional decrease in both glycogen and β-glucans. Taken altogether, these results indicate that a few percent of enzymatic activities leading to the formation of UDP-Glc appears sufficient to provide the UDP-Glc demands required for cell viability, and that the loss of function of UGP1 is lethal mainly because of the inability of yeast cells to properly form the cell wall.


UDP-Glc plays a pivotal role in the intermediary metabolism of yeast since it serves as a glucosyl donor in numerous biosynthetic pathways including the biosynthesis of reserve carbohydrates[1], the formation of cell wall β-glucans[2] and glucomannoproteins[3], protein N-glycosylation[4] and galactose entry into glycolysis[5]. UDP-Glc is formed from UTP and Glc-1P in a reversible reaction catalyzed by uridine diphosphoglucose pyrophosphorylase (UGPase). This enzyme is encoded by UGP1, the deletion of which is lethal for the cells[6]. To determine which metabolic pathways are the most sensitive to changes in UDP-Glc levels, yeast strains in which UGPase could be either overproduced or depleted were generated. One strategy was to clone the coding region of UGP1 in frame with a glucose-repressible GAL10/CYC1 hybrid promoter and to use this construction to rescue a haploid ugp1 null mutant strain. This experiment led to a 90% reduction of UGPase activity with marginal effects on growth and on both UDP-Glc pools and cell wall β-glucans. On the other hand, a 50-fold overproduction of UGPase was deleterious only specifically for cells growing on galactose[6]. The reasons for this defect remain to be clarified.

The aim of this present work was to improve the reduction in the availability of UDP-Glc in two complementary ways. A first approach was to consider the use of a more tightly regulatable promoter which should be otherwise controlled independently of the carbon sources. This possibility was tested with the THI4 promoter characterized by Praekelt et al.[7] who reported a 5000-fold reduction in the activity of this promoter in the presence of thiamine. A second approach was to reduce the flux through Glc-1P by deletion of PGM1 and PGM2 encoding the two phosphoglucomutase isoenzymes in a conditional ugp1 mutant. This strategy appeared feasible because the growth on glucose of a pgm1/pgm2 double mutant is not impaired due to the capability of phosphomannomutase and N-acetylglucosamine-1-phosphate mutase to catalyze, albeit at only 1% efficacy with respect to phosphoglucomutase, the conversion of Glc-6P into Glc-1P[8, 9]. The results of these genetic alterations on the physiology, morphology and cell wall integrity of yeast are reported. They show that only a few percent of enzymatic activities leading to the formation of UDP-Glc is enough for growth and viability of yeast, while rendering it hypersensitive to cell wall interfering drugs.

2Materials and methods

2.1Yeast strains, media and culture conditions

The relevant genotypes of yeast strains used in this work are listed in Table 1. The haploid JF291 or the isogenic diploid strains were used as hosts for transformations. Yeasts were grown at 30°C either on a rich medium (1% yeast extract and 1% bactopeptone) or on Wickerham's minimal medium (MIN)[10] with or without 5 μM thiamine-HCl (Sigma). The carbon sources were added to the media at a final concentration of 2%, and the MIN medium was complemented with 100 μg/ml of the auxotrophic requirements. Heat shock was carried out as described by Neves and François[11], except that the temperature was set at 42°C instead of 40°C.

Table 1.  Strains used in this work
  1. aAll strains are isogenic to JF291.

JF291MATαleu2,3-112 his3–1 ura3-52 PGM1PGM2UGP1[6]
JF291, pWA2JF291 p(URA3 THI4::UGP1 ARS1 CEN4)this study
JMD663MATa/αleu2,3-112/leu2,3-112 ura3-52/ura3-52 his3-1/his3–1 UGP1/ugp1::HIS3this study
JF647MATαleu2,3-112 his3–1 ura3-52 pgm2Δ::HIS3this study
JF649MATαleu2,3-112 his3–1 ura3-52 pgm1Δ::KanRthis study
JF645MATαleu2,3-112 his3–1 ura3-52pgm2Δ::HIS3 pgm1Δ::KanRthis study
JMD105MATαleu2,3-112 his3–1 ura3-52 ugp1::LEU2 pYJMGalcen (URA3 GAL10::UGP1)[6]
JF650MATαleu2,3-112 his3–1 ura3-52 pgm2Δ::HIS3 ugp1::LEU2 pYJMGalCen (URA3 GAL10/CYC1::UGP1)this study
JF591MATαleu2,3-112 his3–1 ura3-52 pgm1Δ::KanRugp1::LEU2 pYJMGalcen (URA3 GAL10/CYC1::UGP1)this study
JF652MATαleu2,3-112 his3–1 ura3-52pgm2::HIS3 pgm1Δ::KanRugp1::LEU2 pYJMGalcen (URA3 GAL10/CYC1::UGP1this study

2.2Construction of the THI4-UGP1 gene fusion

The fusion of UGP1 to the THI4 promoter was constructed by PCR. Firstly, the 1.9-kb Bam HI/Hin dIII fragment of pYJMGalcen[6] bearing UGP1 was subcloned into the centromeric vector pRS316 to yield pWA1. Then, two primers, P1 (5′-ccggatccTTTGATAGTTAGTTGATTTTTTTGG-3′) carrying a 5′ extension (lower case letters) including a Bam HI site (underlined) and P2 (5′-ggccgcggactagtAGATTACTTTTTTAATTTTCATACT-3′) carrying a 5′ extension containing Spe I and Sac I restriction sites, were used to amplify the THI4 promoter from pUP39a[7]. Amplification was carried out with VentR DNA polymerase (New England Biolabs) following standard protocols. The PCR product and pWA1 were digested with Spe I and Bam HI, ligated to each other to yield pWA2.

2.3Construction of deletion strains and genetic methods

The construction of a ugp1 null mutant strain bearing UGP1 under the GAL10/CYC1 hybrid promoter was performed as described previously[6], except that a 1.1-kb Bam HI HIS3 fragment from YDp-H[12] was inserted at the place of the 0.55-kb Bgl II fragment internal to the UGP1 coding sequence. The PGM1 and PGM2 loci were disrupted in the diploid strain using the 2.8-kb Kpn I-Sal I fragment of pPGM1Δ and the 2.2-kb Sac I-Nco I fragment of pPGM2Δ, respectively. These disruption plasmids were generated from pEB-PGM1 and pEB-PGM2 (kindly provided by E. Boles and F.K. Zimmermann) as follows. A 1.5-kb Sna BI-Sna BI fragment of pEB-PGM1 which removed 90% of the PGM1 coding sequence was replaced by a 1.1-kb Sma I-Eco RV of the pFA6aKanMX4 module[13] containing KanR, resulting in pPGM1Δ. To get pPGM2Δ, a 2.0-kb Hpa I-Xho I fragment of pEB-PGM2 containing the region of PGM2 from −28 bp to 317 bp behind the stop codon was replaced by a 1.1-kb Sma I-Sal I HIS3 fragment of YDp-H. DNA fragments were used to transform the diploid strain according to[14]. Transformants were isolated on the appropriate selective media and some of them were verified by Southern blot analysis (data not shown). Standard genetic procedures for diploid construction, sporulation and tetrad dissection were followed[15] to obtain the set of the isogenic deletion strains reported in Table 1.

2.4Analytical procedures

Extracts and assay of UGPase activity was as described previously[6]. Protein concentrations were determined by the method of Bradford[16] using bovine serum albumin as a standard. Cell sampling and metabolite extraction were performed in boiling ethanol[17]. Metabolites were measured by conventional spectrofluorometric techniques[18]. Determination of glycogen and trehalose in whole yeast cells was performed as in[19]. The isolation of cell walls were carried out as described by Ram et al.[20]. The cell wall polysaccharides were measured by the phenol-sulfuric acid method[21] using a mixture of Man/Glc at a ratio of 1:1 as a standard.

3Results and discussion

3.1Strategy for generating strains with reduced UGPase and UDP-Glc synthesis

Fusion of UGP1 to the regulatable THI4 promoter was chosen because studies using a THI4-lacZ gene fusion showed that the transcript disappeared in less than 1 h, and the activity of β-galactosidase decreased to an undetectable level 6 h after the addition of 1 μM thiamine to the thiamine-free medium[7]. Therefore, pWA2 bearing the UGP1 coding sequence under the control of the THI4 promoter was constructed to transform the diploid strain JMD663 which is heterozygous for the ugp1::HIS3 mutation. After selection of transformants on Ura plates, sporulation and tetrad dissection, we found to our surprise that all four spores from 10 complete tetrads grew well on MIN plates supplemented with 5 μM thiamine, and, as expected, only two of them had a His+ phenotype. A typical tetrad was analyzed for the effect of thiamine on UGP1 expression and UGPase activity. When the two segregants lacking the chromosomal copy of UGP1 were cultivated in the presence of thiamine, the transcript of UGP1 was undetectable (not shown), and the UGPase activity was reduced by 95%, as compared to the two wild-type segregants (Fig. 1). The same results were obtained using a glucose-rich medium. In both media, the growth rate was reduced by 25% (not illustrated). In contrast, growth in the absence of thiamine lead to a 50–80-fold increase of UGPase activity in two wild-type spores, and to a 500–630-fold increase in two other ugp1 spores (Fig. 1). This very high activity of UGPase under this condition suggests either that the copy number of the pWA2 is unexpectedly high or that the THI4 promoter is much stronger than the native promoter of UGP1. Support for this latter suggestion comes from the demonstration by Praekelt et al.[7] that THI4 protein is one of the most abundantly synthesized proteins under non-repressing conditions. Since the repressible effect of thiamine on the THI4-UGP1 chimeric gene was only slightly stronger than that exerted by glucose on the GAL10/CYC1-UGP1 construction[6], one could suggest the existence of a UAS-like element within the UGP1 coding sequence. However, this hypothesis is difficult to reconcile with the fact that the UGP1 transcripts were undetectable in cells cultivated in the presence of either glucose[6] or thiamine. Therefore, we favor the hypothesis that the residual UGPase activity could result from an extremely low, glucose- or thiamine-independent, activity of the promoter, combined with a very high stability of the enzyme in the cell.

Figure 1.

Activity of UGPase in haploid segregants from a relevant tetrad. The heterozygous diploid JMD663 (aUGP1/ugp1::HIS3 leu2/leu2 ura3/ura3 his3/his3) was transformed with pWA2, subjected to sporulation and tetrads were dissected on YPD agar plates. UGPase activity was measured in the four segregants from a tetrad that were grown to mid-exponential phase on MIN medium in the absence (open bars) or in the presence (filled bars) of 5 μM thiamine.

To further improve the reduction of flux towards UDP-Glc formation, the two PGM genes encoding the phosphoglucomutase isoforms were deleted in a ugp1 null mutant rescued with UGP1 under the GAL10/CYC1 promoter (strain JMD105). It is shown in Table 2 that on a YPD medium, the growth rate of a pgm1/pgm2 double mutant was the same as the wild-type, and that the growth rate of a ugp1 conditional mutant, which was actually 20% lower than the wild-type, was not further decreased by disruption of both PGM1 and PGM2. However, on a synthetic minimal medium, the growth rates of the wild-type and the ugp1 conditional mutant were significantly affected by disruption of both PGM1 and PGM2. A likely explanation for this observation is that the synthetic medium contains a much higher salt concentration to which these mutants are very sensitive (see Table 3).

Table 2.  Growth rates, intracellular concentrations of UDP-Glc and levels of glycogen in the mutant strains
  1. aYeast cells were grown on YP+2% glucose (YPD) or on Wickerham's medium (MIN) supplemented with the auxotrophic requirements. Measurements of UDP-Glc and glycogen were made on late exponential phase cells (at about 5×107 cells/ml) cultivated on MIN.

  2. Data reported are from three independent experiments±S.E.

StrainRelevant genotypeGrowth rate (h−1)aUDP-Glc (μmol/g dry weight)Glycogen (mg/g dry weight)
JF645UGP1pgm1pgm20.43±0.020.38±0.030.84±0.12 9.9±1.6
JMD105ugp1PGM1PGM20.36±0.030.34±0.031.10±0.18 8.9±1.5
JF591ugp1pgm1PGM20.35±0.030.34±0.020.90±0.12 9.5±2.0
JF650ugp1PGM1pgm20.35+0.020.34+0.030.86±0.15 9.4±1.4
JF652ugp1pgm1pgm20.33±0.030.28±0.020.38±0.12 4.4±0.8
Table 3.  Test of cell wall integrity and composition of the cell wall expressed as Man/Glc ratio and glucosamine (% in cell wall) in the mutants
  1. aMean±S.E. of two independent experiments.

  2. Growth was scored after 3 days of incubation at 30°C on YP+2% glucose agar plates containing the different compounds at the indicated concentrations. ++, normal growth; +, slow growth; +/−, very slow growth; −, no growth. ND, not determined.

StrainRelevant genotypeCaffeineCalcofluorSDSNaClMan/GlcaGlucosaminea
  8 mM0.25 mg/ml0.5 mg/ml0.05%0.5 M % of total sugar in cell wall

3.2Metabolic characterization of the mutants

The metabolic consequences of these genetic alterations were quantitatively assessed by measuring the intracellular concentration of UDP-Glc and glycogen (Table 2) in the late exponential phase of growth of the various mutants on glucose. The concentration of UDP-Glc was not altered by deletion of either one of the two PGM genes, but it was decreased by 3-fold in a pgm1/pgm2 double mutant. This demonstrates that the enzymatic activity associated with the minor form of phosphoglucomutase encoded by PGM1 is sufficient to mediate UDP-Glc synthesis. In the conditional ugp1 mutant (JMD105), the concentration of the sugar nucleotide was 2-fold lower than that in the wild-type, and it was further decreased by 2.5-fold after deleting the two PGM genes. In all situations, the lowering of UDP-Glc was accompanied by a proportional reduction in the content of glycogen (Table 2). These data corroborate those previously reported by Corominas et al.[22] who showed a direct relationship between UDP-Glc and glycogen levels. Furthermore, they agree very well with the affinity of glycogen synthase for UDP-Glc[23]. The effect of UDP-Glc depletion on trehalose biosynthesis was also studied in exponentially growing cells subjected to a thermal shift from 30 to 42°C. It was also found that accumulation of trehalose in this condition was closely related to the UDP-Glc availability, as a 5-fold reduction of UDP-Glc in the ugp1 conditional mutant bearing pgm1/pgm2 mutations was accompanied by a 5-fold reduction in the rate and the extent of trehalose accumulation (Fig. 2). Taking these results together, it can be concluded that the biosynthesis of glycogen and trehalose is primarily regulated by the availability of UDP-Glc under these specific conditions. However, this conclusion does not hold for the synthesis of trehalose during growth on glucose, since the same levels of this disaccharide were found in all the different mutants. This seems to confirm the previous suggestion that the rate of trehalose under this condition is mainly controlled by Glc-6P, the other substrate of trehalose-6-phosphate synthase[19].

Figure 2.

Trehalose accumulation during heat shock treatment at 42°C. Exponential-phase cells (about 1.5×107 cells/ml) cultivated on YP+2% glucose (YPD) at 30°C were analyzed for trehalose content at different times after shifting at 42°C. A: Strains bearing wild-type UGP1 and wild-type PGM genes (◯), or deleted for either PGM1 (•), PGM2 (█) or both genes (▴). B: Conditional ugp1 strains bearing wild-type UGP1 under the GAL10/CYC1 promoter and wild-type PGM genes (◯) or deleted for either PGM1 (•), PGM2 (█) or both genes (▴).

3.3Effects on morphology and cell wall integrity

In a previous report[6], we commented on the fact that yeast cells containing only 10% of the wild-type UGPase activity (e.g. strain JMD105) exhibited an abnormal morphology characterized by multi-budding and elongated forms. We found that deletion of the two PGM genes in this strain exacerbated this morphology and occasionally led to the formation of strings or pseudofilaments composed with 5–8 cells (not illustrated). A possible explanation for this observation is that any disturbance in cell wall synthesis and assembly will affect proper cell division as the two events are closely interlinked by mechanisms which remain to be clarified[2].

Since one of the key issues of UDP-Glc is in the biosynthesis of β-glucans, the cell wall integrity from the different mutants generated was tested by scoring their growth on various plates containing drugs or high salt concentrations (Table 3) which are known to affect or to destabilize the cell wall architecture[20]. Interestingly enough, the deletion of either one of the two PGM genes rendered the yeast cells more sensitive to 0.5 mg/ml calcofluor white than the wild-type. The cell wall structure appeared even more disturbed in a pgm2 deletant (JF647) because the growth of this mutant was already affected with 0.25 mg/ml calcofluor, 8 mM caffeine or 0.05% SDS. This unexpected phenotype could be related to the recent finding that the phosphoglucomutase encoded by PGM2 is a cytoplasmic glycoprotein and an acceptor for a UDP-Glc-dependent Glc-phosphotransferase reaction[24]. The significance of the post-translational modification on Pgm2p is not yet understood.

The deletion of the two PGM genes gave rise to a pronounced slow growth phenotype, which was roughly comparable to that of the conditional ugp1 mutant strains (JMD105 or JMD663-1C). When this latter strain was additionally disrupted for PGM1 and PGM2, growth on any drug-containing plate was totally inhibited. Such a behavior agreed fairly well with the low content of β-glucans as indicated by the high Man/Glc ratio (Table 3). Assuming that the mannose content of the cell wall was not modified, the β-glucans in the ugp1 conditional mutant bearing the pgm1/pgm2 deletion were reduced by about 4 times as compared to the wild-type. It is also shown in Table 3 that the higher the Man/Glc was, the more there was glucosamine, indicating a greater chitin content in the cell wall. These data are a direct demonstration of a previous suggestion that the reduction in β-glucans could be accompanied by an increase in chitin content[25], and support the physiological significance of the interlinkage between chitin and β-glucans as a mean to strengthen the rigidity of the cell wall, mainly when the amount of either glucan or mannan components in the cell wall is reduced [26, 27].

In conclusion, we have provided compelling evidence that the loss of function of UGP1 is lethal for the cells because they cannot properly form cell wall β-glucans. Indeed, the lower the UDP-Glc content, the lower the β-glucan content and the stronger the sensitivity of cells to cell wall interfering drugs. Levels of glycogen and trehalose which require UDP-Glc as a substrate were equally affected, but they are both dispensable for vegetative growth[1]. The most striking result was however that the enzymatic activities leading to the formation of UDP-Glc, the precursors of several macromolecules, can be reduced by more than 95% without affecting cell growth and viability.


We thank D. Gardner and V. Paquet for critical reading of the manuscript. Nathalie Dallies and Carine Tchilinguirian are acknowledged for having carried out the cell wall measurements. This work was supported in part by Grant 9407652 from the Region Midi-Pyrenées. J.M.D. was the recipient of a fellowship from the French MENESR.