Trehalose-6-phosphate synthesis controls yeast gluconeogenesis downstream and independent of SNF1

Trehalose-6-P (T6P), an intermediate of trehalose biosynthesis, was identified as an important regulator of yeast sugar metabolism and signaling. tps1Δ mutants, deficient in T6P synthesis (TPS), are unable to grow on rapidly fermentable medium with uncontrolled influx in glycolysis, depletion of ATP and accumulation of sugar phosphates. However, the exact molecular mechanisms involved are not fully understood. We show that SNF1 deletion restores the tps1Δ growth defect on glucose, suggesting that lack of TPS hampers inactivation of SNF1 or SNF1-regulated processes. In addition to alternative, non-fermentable carbon metabolism, SNF1 controls two major processes: respiration and gluconeogenesis. The tps1Δ defect appears to be specifically associated with deficient inhibition of gluconeogenesis, indicating more downstream effects. Consistently, Snf1 dephosphorylation and inactivation on glucose medium are not affected, as confirmed with an in vivo Snf1 activity reporter. Detailed analysis shows that gluconeogenic Pck1 and Fbp1 expression, protein levels and activity are not repressed upon glucose addition to tps1Δ cells, suggesting a link between the metabolic defect and persistent gluconeogenesis. While SNF1 is essential for induction of gluconeogenesis, T6P/TPS is required for inactivation of gluconeogenesis in the presence of glucose, downstream and independent of SNF1 activity and the Cat8 and Sip4 transcription factors.

trehalose-6-P, TPS, SNF1, gluconeogenesis

INTRODUCTION

Cells of all living organisms have evolved complex signal transduction networks to ensure that a wide range of physiological and growth properties are properly adjusted to the changing environmental conditions. For microorganisms like yeast cells, fluctuating extracellular nutrient availability is a key determinant of growth and cell cycle progression and in line with this, several conserved nutrient-controlled eukaryotic pathways that regulate cell growth, proliferation, metabolism and stress tolerance have been first characterized in Saccharomyces cerevisiae (Smets et al. 2010; Conrad et al. 2014). Interestingly, trehalose-6-P (T6P) was identified as an important regulator of yeast sugar metabolism and signaling (Thevelein and Hohmann 1995; Hohmann et al. 1996). T6P is the metabolic intermediate of trehalose biosynthesis from glucose-6-phosphate and UDP-glucose, catalyzed by T6P synthase (TPS, Tps1) and T6P phosphatase (TPP, Tps2) at the crossroads of primary carbon metabolism, making an excellent readout of carbon and energy status. Remarkably, yeast tps1Δ mutants are unable to grow on rapidly fermentable glucose medium and this growth defect is rescued by mutation of Hxk2 (the most active yeast hexokinase isozyme), suggesting that deregulation of glycolysis and uncontrolled carbon influx is causing the growth defect (Thevelein and Hohmann 1995; Hohmann et al. 1996). Consistently, upstream glycolytic sugar phosphates accumulate to high levels and ATP and phosphate levels drop dramatically after addition of glucose to tps1Δ mutant cells. Moreover, T6P inhibits yeast Hxk2 in vitro (reviewed in Thevelein and Hohmann 1995). However, a T6P insensitive Schizosaccharomyces pombe Hxk was used to show that the growth defect and deregulation of glycolysis can be uncoupled and that inhibition of Hxk activity by T6P is not the major mechanism by which the Tps1 protein controls glycolytic flux and growth on glucose (Bonini, Van Dijck and Thevelein 2003). Importantly, glucose inhibits growth of a tps1Δ strain already at low mM concentrations, consistent with a regulatory rather than or in addition to a metabolic effect. While several hypotheses and (metabolic) models have been proposed, the exact regulatory functions of T6P and cause for the dramatic metabolic and growth defect of TPS-deficient cells in the presence of glucose remain elusive.

T6P is recently also emerging as a novel growth regulator in plants, where it is required for embryogenesis and normal vegetative development as well as the transition to flowering (van Dijken, Schluepmann and Smeekens 2004; Ramon et al. 2009; Wahl et al. 2013). But while T6P does not inhibit plant Hxk (Eastmond et al. 2002), it was identified as an allosteric inhibitor of Snf1-related kinase 1 (SnRK1) activity (Zhang et al. 2009). The plant SnRK1 kinases are part of the evolutionarily conserved serine/threonine kinase family with orthologs in all eukaryotic organisms, from yeast (SNF1) to mammals (AMPK) and plants (SnRK1). Activated by energy-depleting stress conditions, the AMPK/SNF1/SnRK1 kinase complexes play key roles in eukaryotic energy homeostasis (Ghillebert et al. 2011). In S. cerevisiae, SNF1 primarily controls the glucose repression pathway, which involves adaptation of carbon metabolism to a decrease in extracellular glucose availability. When glucose becomes limiting, activation of SNF1 results in the transcriptional induction of enzymes involved in gluconeogenesis, respiration and uptake and metabolism of alternative carbon sources, such as glycerol, ethanol and sucrose, and a selection of glucose transporters (Ozcan and Johnston 1999; Smets et al. 2010; Ghillebert et al. 2011).

One of the prime SNF1 targets is Mig1, the main transcriptional regulator of glucose-repressed genes (Treitel and Carlson 1995) (Fig. 1A). Under glucose-limited conditions, SNF1 directly phosphorylates Mig1, resulting in its nuclear exclusion and subsequent relieve of glucose repression (De Vit, Waddle and Johnston 1997; Treitel, Kuchin and Carlson 1998; Smith et al. 1999). Hxk2, next to its catalytic function in glycolysis, also exhibits regulatory functions in SNF1-Mig1 signaling. A small fraction of Hxk2 localizes in the nucleus and interacts with Mig1 and Snf1, thereby inhibiting Mig1 phosphorylation and nuclear exclusion during growth in the presence of glucose (Ahuatzi et al. 2004, 2007). In line with this, hxk2Δ mutants show derepression of glucose-repressed genes on glucose medium (Zimmermann and Scheel 1977; Entian 1980), due to nuclear exclusion of Mig1 (Ahuatzi et al. 2007). Additional effectors of the Snf1 kinase are the transcription factors (TFs) Adr1, Cat8, Sip4 and Rds2 (Fig. 1A). The latter three are involved in the activation of ‘carbon source response element’ (CSRE)-controlled genes, encoding key enzymes in gluconeogenesis, like fructose-1,6-bisphosphatase (Fbp1), phospo-enolpyruvate carboxykinase (Pck1) and isocitrate lyase (Icl1) (Scholer and Schuller 1994), and in metabolism of alternative carbon sources, like acetyl-CoA synthetase (Acs1) and pyruvate decarboxylase (Pdc1) (Kratzer and Schuller 1997) (Fig. S1, Supporting Information). While CAT8 and SIP4 transcription is also regulated by Mig1 (Hedges, Proft and Entian 1995; Lesage, Yang and Carlson 1996; Rahner et al. 1996), RDS2 and ADR1 do not seem to be glucose-repressed (Young, Kacherovsky and Van Riper 2002; Young et al. 2003). All four TFs are activated by SNF1-dependent phosphorylation (Hedges, Proft and Entian 1995; Lesage, Yang and Carlson 1996; Rahner et al. 1996; Randez-Gil et al. 1997; Vincent and Carlson 1998; Hiesinger et al. 2001; Soontorngun et al. 2007, 2012; Ratnakumar et al. 2009). However, only rds2Δ and cat8Δ strains exhibit deficient growth phenotypes similar to snf1Δ mutants in the presence of non-fermentable carbon sources (Hedges, Proft and Entian 1995; Lesage, Yang and Carlson 1996; Soontorngun et al. 2012).

Figure 1.

Deletion of SNF1 rescues the growth defect of a tps1Δ mutant on glucose-containing medium. (A) Schematic representation of the SNF1-signaling network. The SNF1 complex (with the Snf1 catalytic α subunit) activates a number of partially redundant TFs involved in the induction of alternative carbon source metabolism via direct TF phosphorylation and induction of expression of a subset of these TFs through Mig1 phosphorylation and inhibition. Hxk2 interferes with SNF1 signaling by controlling Mig1 phosphorylation and localization. (B) Growth assay showing the growth defect of the tps1Δ mutant on glucose-containing medium and its Mig1-dependent suppression by additional deletion of SNF1. Dilution series of cells are spotted on solid rich growth medium containing 2% galactose (YPGal) or 2% glucose (YPGlc). Growth is analyzed after 2–3 days at 30°C. (C, D) Growth analysis of WT, tps1Δ and tps1Δsnf1Δ mutants growing to stationary phase in liquid rich (YP) glucose- (C) and galactose- (D) containing medium, showing that snf1Δ effectively suppresses the tps1Δ growth defect on glucose-containing medium. Shown are mean values with standard error bars (n = 6). (E) Growth assay on solid synthetic media confirming that overexpression of MIG1 can partially rescue the tps1Δ growth defect in the presence of glucose. (F) Growth assay on solid synthetic media showing that hxk2Δ suppression of the tps1Δ growth defect in the presence of glucose is not associated with the regulatory nuclear function of Hxk2. (G) Growth assay on solid synthetic media confirming that T6P production by the bacterial TPS protein OtsA can complement the tps1Δ mutant phenotype in the presence of glucose.

SNF1, like AMPK and SnRK1, functions as a heterotrimeric kinase complex consisting of a catalytic α subunit and regulatory β and γ subunits required for complex stability, localization, kinase activity and substrate specificity (Ghillebert et al. 2011). Saccharomyces cerevisiae encodes single α (Snf1) and γ (Snf4) subunits and three partially redundant β-subunits (Gal83, Sip1 and Sip2), yielding three possible complexes. Activity of the SNF1 kinase complex is regulated by multiple mechanisms. First, when glucose becomes limiting, Snf4 interacts with Snf1, thereby relieving the auto-inhibitory interaction of the C- and N-terminal domain of the catalytic subunit (Jiang and Carlson 1996). Furthermore, phosphorylation of Thr210 in the activation (T) loop of the catalytic subunit, mediated by three partially redundant upstream kinases (Sak1, Elm1 or Tos3), is essential for kinase activity (Estruch et al. 1992; Wilson, Hawley and Hardie 1996; McCartney and Schmidt 2001; Sutherland et al. 2003). T-loop dephosphorylation on high glucose is mediated by the Reg1-Glc7 (PP1), Sit4 (PP2A-related) and Ptc1(PP2C-type) phosphatases (Tu and Carlson 1995; Ruiz, Xu and Carlson 2011, 2013). Finally, SNF1 complex function is also regulated by protein localization in response to glucose availability and this depends on the β subunits (Vincent et al. 2001). Recently, an additional regulatory mechanism was identified in which glucose induces SUMOylation of the catalytic Snf1 subunit, causing its inactivation (despite T-loop phosphorylation) and ubiquitin-dependent degradation (Simpson-Lavy and Johnston 2013). SNF1 signaling also cross-talks with other pathways. cAMP-dependent protein kinase (PKA) e.g. acts at the core of the major nutrient signaling pathway activated in the presence of rapidly fermentable sugars and consistently is found to inhibit SNF1 function at different levels, including complex localization (Hedbacker, Townley and Carlson 2004), Sak1 phosphorylation (Barrett et al. 2012), PP1 phosphatase activation (Castermans et al. 2012) and SUMO E3 ligase activation (Simpson-Lavy et al. 2015). Conversely, SNF1 signaling also affects PKA activity, e.g. through phosphorylation of adenylate cyclase, leading to reduced cAMP levels (Nicastro et al. 2015).

Despite the structural and functional conservation of the AMPK/SNF1/SnRK1 complexes, the actual metabolic signals triggering their activation appear to have diverged according to the organisms’ different lifestyles (Ramon et al. 2013; Emanuelle et al. 2015). In animals, cellular AMP/ATP ratios determine the activity of AMPK, but in plants and yeast cells, this does not seem to be the major regulatory factor (Davies et al. 1995; Suter et al. 2006; Sanders et al. 2007). Interestingly, in photo-autotrophic plants, sugar phosphates such as glucose-6-P and T6P were found to allosterically inhibit SnRK1 kinase activity in vitro (Toroser, Plaut and Huber 2000; Zhang et al. 2009). Inverse correlations between SnRK1-controlled gene expression and T6P levels (a sensitive readout of plant sucrose status) appear to confirm T6P as an important SnRK1 regulator (Zhang et al. 2009), although relations might be more complex (Lunn et al. 2006; Schluepmann, Berke and Sanchez-Perez 2012; Nunes et al. 2013b). We now report that the growth defect of a yeast tps1Δ mutant on glucose is restored by additional deletion of SNF1 and that T6P synthesis regulates yeast SNF1 targets. This would be consistent with problematic persistent SNF1 activity on glucose medium in the absence of T6P synthesis. However, while lack of T6P (or the TPS enzyme) results in an inability to inhibit gluconeogenesis in the presence of glucose, possibly contributing to the metabolic and/or growth defect, suppression of the growth defect is not associated with deficient repression of respiration. In line with this, normal glucose-induced dephosphorylation and subsequent inactivation of the SNF1 kinase complex is observed in the tps1Δ mutant, indicating that T6P synthesis negatively regulates gluconeogenesis, but downstream and independent of SNF1.

MATERIALS AND METHODS

Strains and plasmids

Saccharomyces cerevisiae strains used in this study are listed in Table 1. For abrogation of the mtDNA (rho⁻), cells were grown on YPD containing 25 μg ml⁻¹ ethidium bromide. Cells were transformed using the Gietz method (Gietz et al. 1995), however, tps1 mutants were heat-shocked for 2 h at 37°C instead of 20–25 min at 42°C. Double mutants were obtained by crossing of the single mutants. Sporulation was performed by spotting diploid cells on plates containing 1% K-acetate, 0.1% KHCO₃, pH 6.0 for 3 to 6 days at 24°C. Tetrads were dissected using a micromanipulator (Singer Instruments, Roadwater, UK). Deletions with auxotrophic markers were confirmed on selective media and/or by PCR.

Table 1.

Strains used in this study.

Strain	Genetic marker	Reference
W303-1A WT	MATa leu2-3/112 ura3-1 trp1-1 his3-11, 15 ade2-1 can1-100	Thomas and Rothstein (1989)
tps1Δ	W303-1A tps1::TRP1	Hohmann et al. (1993)
snf1Δ	W303-1A snf1::HIS3	Ostling, Carlberg and Ronne (1996)
mig1Δ	W303-1A mig1::LEU2	Nehlin and Ronne (1990)
hxk2Δ	W303-1A hxk2::LEU2	Thomas and Rothstein (1989)
tps1Δ snf1Δ	W303-1A tps1::TRP1 snf1::HIS3	This study
tps1Δ hxk2Δ	W303-1A tps1::TRP1 hxk2::LEU2	Thomas and Rothstein (1989)
tps1Δ snf1Δ mig1Δ	W303-1A tps1::TRP1 snf1::HIS3 mig1::LEU2	This study
cat8Δ	W303-1A cat8::TRP1	Vincent and Carlson (1998)
tps1Δ cat8Δ	W303-1A tps1::TRP1 cat8::TRP1	This study
sip4Δ	W303-1A sip4::LEU2	Vincent and Carlson (1998)
tps1Δ sip4Δ	W303-1A tps1::TRP1 sip4::LEU2	This study

All plasmids used in this study are listed in Table 2. New plasmids were created using the specific cloning primers, listed in Table 3.

Table 2.

Plasmids used in this study.

Name	Backbone	Ori	Promoter	Insert	Marker	Reference
pYX212	pYX212	2μ	HXT7	–	URA3	Vandesteene et al. (2010)
pYX242	pYX242	2μ	TPI	–	LEU2	Ingenous Inc.
pYEp352-HXK2	pYEp352	2μ	lac	HXK2	URA3	Pelaez et al. (2012)
pYEp352-HXK2-nls1	pYEp352	2μ	lac	HXK2K6,7,12A	URA3	Pelaez et al. (2012)
pJS465	pYEplac181	2μ	CAT8	CAT8	LEU2	Rahner, Hiesinger and Schuller (1999)
pJS466	pYEplac181	2μ	met25	CAT8	LEU2	Rahner, Hiesinger and Schuller (1999)
pAR33	pYEplac181	2μ	met25	CAT8-INO2TAD	LEU2	Rahner, Hiesinger and Schuller (1999)
pYX212-SNF1	pYX212	2μ	HXT7	SNF1-HA	URA3	This study
pYEplac195-FBP1	pYEplac195	2μ	FBP1	FBP1-HA	URA3	This study
pYEplac195-PCK1	pYEplac195	2μ	PCK1	PCK1-HA	URA3	This study
pYX242-ACC1	pYX242	2μ	TPI	ACC1tr-GFP-HA	LEU2	This study
pYX242-ACC1-S79A	pYX242	2μ	TPI	ACC1tr-S79A-GFP-HA	LEU2	This study
pYX212-otsA	pYX212	2μ	HXT7	otsA-HA	URA3	This study
pYX242-otsA	pYX242	2μ	TPI	otsA-HA	LEU2	This study
pBM3315-MIG1	pBluescript	CEN	MIG1	MIG1-GFP	URA3	De Vit, Waddle and Johnston (1997)

Table 3.

Cloning primers used in this study.

Gene	Forward primera	Reverse primera	RE Fwdb	RE Revb
SNF1/YDR477W	CG-GGATCC-ATGAGCAGTAACAACAACACAAAC	A-AGGCCT-ATTGCTTTGACTGTTAACGGCTA	BamHI	StuI
FBP1/YLR377C	GA-AGATCT-GCCAAGGAAGGTGGGTTTAC	AA-CTGCAG-CTGTGACTTGCCAATATGGTCT	BglII	PstI
PCK1/YKR097W	G-GAATTC-GGACACATGTCGACGAGTTTG	ACAT-GCATGC-CTCGAATTGAGGACCAGCGG	EcoRI	SphI
otsA/ECK1895	G-GAATTC-ATGAGTCGTTTAGTCGTAGTATC	A-AGGCCT-CGCAAGCTTTGGAAAGGTAGC	EcoRI	StuI
MIG1/YGL035C	CG-GGATCC-ATGCAAAGCCCATATCCAATGA	A-AGGCCT-GTCCATGTGTGGGAAGGGCA	BamHI	StuI

a

The restriction sites in cloning primers are indicated in italic.

b

The corresponding restriction enzymes (RE) for the forward (fwd) and reverse (rev) primer are specified.

Growth conditions

Yeast cells were grown in standard rich medium containing 2% (w/v) bacteriological peptone, 1% (w/v) yeast extract (YP) and 2% (w/v) galactose. Cells containing plasmids were pre-grown in minimal medium containing 0.19% (w/v) yeast nitrogen base w/o amino acids, supplemented with synthetic drop-out amino acid/nucleotide mixtures as required, 0.5% (w/v) (NH₄)₂SO₄ and 2% (w/v) galactose. Solid medium contained an additional 1.5% (w/v) agar. Cells were incubated at 30°C, liquid cultures under continuous shaking. For qPCR analysis, enzyme activity measurements, western blot analyses (expression, phosphorylation, protein stability) and localization studies by confocal microscopy, cells were grown overnight in SGal-U-L at 30°C to OD₆₀₀ 1, washed three times, resuspended in S-U-L medium containing 3% glycerol and 2% ethanol and incubated for another 3 h. Glucose was then added to a final concentration of 2% (w/v). Samples were taken at the indicated time points. A total of 1% glucose was added to snf1Δ pre-cultures on galactose to enable sufficient growth.

Growth analysis

For growth complementation spot tests, cells were grown overnight in YPGal or selective medium to OD₆₀₀ 1. Next, 10-fold serial dilutions (starting from OD₆₀₀ 0.1) were spotted solid media. Pictures were taken after 2–3 days at 30°C. Representative results are shown. For growth curves, cells were grown in YPGal medium until stationary phase and subsequently diluted to OD₆₀₀ 0.01 in a 96-well plate. Growth of the cells was measured at 600 nm every 2 h for 60 h using a Multiskan GO (Thermo Scientific, Thermo Scientific, Waltham, MA, USA).

qPCR analysis

Yeast liquid culture samples were taken at the indicated time points and rapidly cooled by addition to ice-cold water. Cells were pelleted and washed once with ice-cold water and stored at −80°C. Isolation of total RNA was performed as described previously (Giots, Donaton and Thevelein 2003). For each sample, 1 μg of total RNA was used for reverse transcription (RT) with the Reverse Transcription System (Promega A3500, Madison, WI, USA). cDNA was diluted to 5 ng μL⁻¹. The 10 μL PCR reaction was composed of 5 μL GoTaq^® qPCR Master Mix 2× (Promega, Madison, WI, USA), 0.1 μL of CXR reference dye and 0.2 μL of each primer (100 nM final concentration). For the PCR, 2 μL of cDNA (10 ng) was added to each reaction mix. The PCR program comprised an initial denaturation step of 2 min at 95°C and amplification by 40 cycles of 3 s at 95°C, 30 s at 60°C in an ABI Prism 7000 cycler. The ALG9 and TAF10 genes were used as controls for normalization (Teste et al. 2009). Primer sequences used for real-time PCR analysis are provided in Table 4. For each strain, the relative expression is calculated using the ΔΔCt method with values from cells grown on galactose as reference.

Table 4.

qRT-PCR primers used in this study.

Gene	Forward	Reverse	Reference
ALG9/YNL219C	TGTCACGGATAGTGGCTTTG	CATTCACTACCGGTGCCTTC	This study
TAF10/YDR167W	CGTGCAGCAGATTTCACAAC	CAACAGCGCTACTGAGATCG	This study
PCK1/YKR097W	CTTCCTAGCCTTGCACCCTA	CATTGGCTAACGAACCATCA	This study
FBP1/YLR377C	AATGGTAGCCGCTTGCTATG	AATTCGCCCAAGTTTGTGTC	This study
ICL1/YER065C	ACTTTCGGTGCCCTAGATCC	TGAGGTGGAAGCAGTTGATG	This study
MLS1/YNL117W	TTGGATATCGGGTTGCAGTAC	CCGTGTTTCACCCATTGATAC	Infante et al. (2011)
ACS1/YAL054C	ATCAAGAAGCATTTGGTCTTTACTG	TGCCAGGGTTTGACAATGTAG	Infante et al. (2011)
CIT1/YNR001C	TGGTTTAGCTGGCCCATTAC	TCCCTGCGTTCAAAGTATCC	This study

Statistical analysis

For all statistical analysis, two-way ANOVA with Bonferroni post tests were performed using GraphPad Prism version 5.01 for Windows (GraphPad Software, San Diego, CA, USA, www.graphpad.com). Statistical significance is indicated with *P < 0.05; **P < 0.01, ***P < 0.001 or not ns (not significant).

Western blot analysis

Cells were harvested and resuspended in 500 μL ice-cold lysis buffer [50 mM Tris-HCl pH 7.4, 1% Triton, 10% glycerol, 2.5 mM MgCl₂, 200 mM NaCl] containing protease inhibitors (Complete EDTA-free, Roche) and phosphatase inhibitors (PhosSTOP Roche, Basel, Switzerland). After extraction with glass beads (FastPrep, MP Biomedicals, Santa Ana, CA, USA), lysates were clarified by centrifugation for 10 min at 13 000 rpm at 4°C. Protein concentrations in extract were determined using Lowry's method as described (Lowry et al. 1951). Sample buffer was added to the supernatant fractions, followed by boiling for 5 min. After brief centrifugation, equal amounts of solubilized protein were separated by SDS–PAGE (Sodium dodecyl sulphate-polyacrylamide gel electrophoresis) and transferred to PVDF membrane (Immobilon, Millipore, Billerica, MA, USA) by semi-dry blotting. HA-tagged proteins were detected using HRP-coupled anti-HA high affinity rat antibodies (Roche) and enhanced chemiluminescence (SuperSignal, Thermo Scientific, Waltham, MA, USA). For detection of Snf1 phosphorylation, polyclonal Phospho-AMPKα (Thr172) antibody (Cell signaling Technology, Danvers, MA, USA) was used in combination with a secondary goat-anti-rabbit IgG-HPR antibody (Santa Cruz Biotechnology, Dallas, TX, USA).

in vivo SNF1 activity

A synthetic SNF1 activity reporter based on an AMPK-phosphorylated mammalian (rat) acetyl-CoA carboxylase (ACC1) peptide was expressed using the pYX242 expression vector. Samples were taken and proteins were purified as described above. SNF1 kinase activity was determined using the HPR-coupled anti-HA high affinity rat antibodies (Roche) for detection of construct expression and a specific Phospho-Acetyl-CoA Carboxylase (Ser79) Antibody (Cell Signaling Technology, Danvers, MA, USA) for detection of the phosphorylated fraction.

Confocal microscopy

For Mig1-GFP localization, cells were grown as described above and samples were taken after 3 h of starvation and 30 min after re-addition of glucose for analysis with fluorescence microscopy (FluoView FV1000 confocal system, Olympus, Tokyo, Japan).

Enzyme activities

Cells were grown in liquid culture as described above. A total of 350 mL of culture was sampled (55–65 mg of biomass) and harvested by centrifugation (5 min, 4750 rpm), washed twice with 10 mM potassium phosphate buffer (pH 7.5) containing 2 mM EDTA, resuspended in 4 mL and stored at −20°C. Samples were thawed, washed and resuspended in 2 mL 100 mM Tris-HCl buffer (pH 7.5) containing 2 mM MgCl₂, and 1 mM DTT (Sigma-Aldrich, USA). After extraction (FastPrep, MP Biomedical, four bursts at speed six alternated with cooling on ice) with acid-washed (425–600 μm) glass beads (Sigma-Aldrich, USA), phosphoenolpyruvate carboxykinase (PCK; EC 4.1.1.32) and fructose-1,6-bisophosphatase (FBP; EC 3.1.3.11) activity was determined as described (de Jong-Gubbels et al. 1995). Protein content was determined using the Lowry method (Lowry et al. 1951). Bovine serum albumin (essentially fatty acid free) was used as a standard.

RESULTS

Deletion of SNF1 restores growth of a tps1Δ mutant in the presence of glucose

The yeast tps1Δ mutant growth defect in the presence of rapidly fermentable sugars like glucose can be suppressed by additional deletion of HXK2, encoding the most active yeast hexokinase isozyme, that is inhibited by T6P in vitro (Thevelein and Hohmann 1995; Hohmann et al. 1996). However, inhibition of Hxk2 is not the only mechanism by which T6P controls glycolytic flux and growth (Bonini, Van Dijck and Thevelein 2003). An important cellular response for optimal growth in glucose medium is the repression of SNF1 activity and its downstream regulatory network (Fig. 1A). Interestingly, in plants, T6P was recently identified as a specific inhibitor of SnRK1 kinase activity when carbon and energy supplies are abundant (Zhang et al. 2009; Schluepmann, Berke and Sanchez-Perez 2012; Nunes et al. 2013b). We therefore explored the possibility of evolutionary conservation of such a mechanism and hence whether the metabolic and growth defect of the yeast tps1Δ mutant is partly due to an inability to inactivate the SNF1 kinase and/or SNF1-regulated processes in the presence of glucose. Consistent with such a scenario, deletion of SNF1 restores growth of tps1Δ mutant on glucose medium (Fig. 1B). As a control, the different strains were grown on galactose-containing medium. Growth of the snf1Δ mutant is significantly affected on galactose, as reported earlier, which surprisingly can be partially suppressed by tps1Δ. Analysis of growth in liquid medium confirms efficient suppression of the tps1Δ growth defect (Fig. 1C and D). Moreover, this growth suppression appears to be dependent on the Mig1 transcriptional repressor, as additional disruption of MIG1 in a tps1Δ snf1Δ strain completely abolishes growth under these conditions. In line with this, overexpression of MIG1 partially also restores growth of a tps1Δ mutant in the presence of glucose (Fig. 1E), as observed before (Hohmann et al. 1992).

In addition to its catalytic function as the major hexokinase on glucose medium, Hxk2 also exhibits regulatory functions, directly interfering with nuclear SNF1-Mig1-signaling (Ahuatzi et al. 2007). To discriminate between the catalytic and signaling function of Hxk2 in suppression of the tps1Δ growth defect, we expressed a wild-type Hxk2 and a mutant allele (K6A, K7A, K12A) unable to localize to the nucleus (Pelaez et al. 2012), in the tps1Δ hxk2Δ strain. Both constructs restored the growth defect (Fig. 1F), suggesting that the nuclear regulatory function of Hxk2 is not involved in the suppression of the tps1Δ growth phenotype.

Finally, we confirmed that introduction of bacterial TPS activity (Escherichia coli OtsA, with moderate homology to the yeast Tps1 protein) restores growth of the tps1Δ mutant, pointing at a role for T6P (Fig. 1F). However, complementation with OtsA is not absolute, possibly due to only partial restoration of T6P levels. In addition, a regulatory role for the Tps1 protein itself cannot be excluded (Bonini et al. 2000; Petitjean et al. 2015). Together, these growth complementation analyses are consistent with T6P/TPS inhibiting SNF1 activity and/or target processes.

Rescue of tps1Δ mutant growth by snf1Δ is lost by Cat8 activation

The SNF1 kinase complex is required for the induction (derepression) of two major processes when glucose becomes limiting, i.e. respiration and gluconeogenesis and the inability to induce (one of) these processes might contribute to the tps1Δ snf1Δ strain efficiently growing in the presence of glucose. We therefore analyzed in more detail how inhibition or induction of respiration and gluconeogenesis affects growth of tps1Δ and tps1Δ snf1Δ mutant strains.

It was already reported by others that blocking respiration at the level of the cytochrome bc₁ complex via addition of antimycin or disruption of QCR9 partially restores growth of a tps1Δ mutant in the presence of glucose (Blazquez and Gancedo 1995). To explore this further, we investigated whether blocking respiration on different levels restores growth of the tps1Δ and the tps1Δ snf1Δ mig1Δ strains under these conditions. We found that, while blocking the bc₁ complex partially restores growth of both strains, inhibition of cytochrome oxidase using NaN₃ has no effect (Fig. 2A). Moreover, anaerobic growth and specific disruption of mitochondrial activity (ρ⁻) does also not suppress the growth defects in the presence of glucose (Fig. 2A). Hence, blocking the cytochrome bc₁ complex appears to have additional specific effects independent of loss of respiration, allowing yeast cells to grow on glucose in the absence of Tps1, although tps1Δ snf1Δ growth appears to be slightly affected. One such putative effect is the leakage of O₂^−., increasing oxidative stress in the cell. However, induction of oxidative stress using addition of H₂O₂ or Fe²⁺ does not restore growth of the tps1Δ or tps1Δ snf1Δ mig1Δ strain in the presence of glucose (Fig. 2A). In any case, the results indicate that the growth defect of the tps1Δ mutant in the presence of glucose does not involve deficient regulation of respiration.

Figure 2.

Rescue of the tps1Δ mutant growth phenotype by snf1Δ is associated with inactivation of gluconeogenesis rather than respiration. (A) Growth assay assessing the involvement of respiration in the tps1Δ mutant growth phenotype. Dilution series of WT and mutant cells are spotted on solid rich growth medium containing 2% galactose (YPGal) or 2% glucose (YPGlc) and compounds or conditions affecting respiration. Growth is analyzed after 2–3 days at 30°C. While antimycin (1 μg ml⁻¹) seems to suppress the growth defect, NaN3 (0.15 mM), anaerobic growth, H₂O₂ (0.5 mM), FeSO₄ (10 mM) and the ρ0 phenotype do not resort the same effect. (B) Growth assay exploring the involvement of gluconeogenesis in the tps1Δ mutant growth phenotype. Expression of CAT8 from its own promoter (WT CAT8) or from a constitutively active promotor (Met25-CAT8) does not restore the growth defect in a tps1Δsnf1Δ mutant, but a constitutively active Cat8 (Met25-CAT8-INO2_TAD) does.

We then investigated whether disruption of SNF1 suppresses the tps1Δ growth defect in the presence of glucose by loss of induction of gluconeogenesis. Activation of gluconeogenesis occurs when cells are starved for glucose and requires a functional SNF1 (Schuller 2003; Turcotte et al. 2010). As multiple partially redundant downstream factors have been implicated in the induction of gluconeogenesis (Fig. 1A), complicating its disruption (Turcotte et al. 2010), we activated gluconeogenesis in a tps1Δ snf1Δ strain using Cat8 (Rahner, Hiesinger and Schuller 1999). As described above, transcription of CAT8 is repressed by Mig1 and Cat8 is activated by SNF1 phosphorylation. We therefore used CAT8pr-CAT8, met25pr-CAT8 and met25pr-CAT8-INO2_TAD constructs, the latter containing an auto-activation domain, resulting in constitutive activation of Cat8 (Rahner, Hiesinger and Schuller 1999). Interestingly, the met25-CAT8-INO2_TAD construct completely abolishes growth of the tps1Δ snf1Δ strain in the presence of glucose, while the others did not have any effect under these conditions (Fig. 2B). These data indicate that suppression of the tps1Δ growth defect by disruption of SNF1 might be linked to a deficient induction of Cat8 target processes like gluconeogenesis and that a lack of T6P/TPS-mediated repression of gluconeogenesis in the presence of glucose (and the associated futile cycling) possibly contributes to the metabolic and growth defect of a tps1Δ strain.

T6P/TPS acts downstream and independent of SNF1 kinase activity

The above analyses suggest a role for T6P in the regulation of specific processes downstream of SNF1, rather than the kinase itself. We therefore investigated the possible effect of TPS deficiency on SNF1 kinase activity in more detail. Phosphorylation of the conserved Thr210 residue in the activation loop of Snf1 is essential for complex kinase activity and addition of glucose to glucose-starved cells induces rapid dephosphorylation (Ruiz, Xu and Carlson 2011). To assess if SNF1 remains active in a tps1Δ strain, the dephosphorylation of Snf1^Thr210 was analyzed in wild type (WT) and tps1Δ strains after addition of glucose to glucose-starved cells (Fig. 3A). Loss of T6P did not affect dephosphorylation of Snf1 under these conditions, indicating that SNF1 is effectively inactivated on glucose in a tps1Δ strain. To corroborate this, and to exclude any phosphorylation-independent activity, we decided to directly analyze SNF1 kinase activity in vivo using a new synthetic reporter based on a rat ACC1 peptide that is specifically phosphorylated by AMPK/SNF1 (Fig. 3B). Analysis of the phosphorylation status of the reporter (with commercially available phospho-specific P-ACC antibodies) indicated that both in a WT and a tps1Δ strain, SNF1 activity is repressed after glucose addition to derepressed cells (Fig. 3B).

Figure 3.

Yeast SNF1 activity is not regulated by T6P/TPS. (A) Efficient dephosphorylation of Snf1 in wild type and tps1Δ cells. Cells expressing Snf1-HA (pYX212-SNF1) were pre-grown on 2% galactose medium and transferred to 3% glycerol/2% ethanol medium for 3 h. Protein samples were taken 0^″, 60^″ and 120^″ after supplementation of 2% glucose (Glc). Equal amounts of protein (20 μg) were loaded. Western blot analysis was repeated several times, representative results are shown. (B) SNF1 kinase activity was evaluated in vivo in WT, tps1Δ and bacterial TPS (pYX212-otsA) expressing tps1Δ mutants using a novel Acetyl-CoA Carboxylase peptide-based reporter (pYX242-ACC and pYX242-ACC-S79A). Total reporter expression (HA-antibody) and phosphorylation (specific Ser79 Phospho-ACC antibody) were analyzed in cells after 3 h on 3% glycerol/2% ethanol medium (-) and 30^′ after addition of glucose (+). For detection of the reporter construct 5 μg of protein was loaded, for detection of phosphorylation 100 μg of protein was loaded. Controls for specificity include: a. the ACC-reporter in WT cells, b. the ACC-S79A mutant reporter (with mutated SNF1 phosphorylation site) in WT cells and c. the ACC-reporter in the snf1Δ mutant. (C) Regulation of Mig1 localization is not affected in the tps1Δ mutant. Confocal fluorescence microscopy of Mig1-GFP (pBM3315) expressing WT and mutant cells after 3 h on 3% glycerol/2% ethanol (Gly/Eth) medium and 30^′ after addition of glucose (Glc). Under the conditions used, the majority of cells showed a predominantly nuclear or cytoplasmic localization of Mig1-GFP. Representative pictures are shown. Cells with a snf1Δ and hxk2Δ background, respectively, showed a constitutively nuclear and cytoplasmic localization of Mig1-GFP.

Growth rescue of the tps1Δ mutant by snf1Δ was dependent on Mig1, a key downstream target of SNF1, responsible for glucose-repression. SNF1 regulates the activity of Mig1 via phosphorylation at four different residues and its subsequent translocation out of the nucleus (Treitel, Kuchin and Carlson 1998; Smith et al. 1999). Conversely, Hxk2 regulates Mig1 localization and activity by inhibiting phosphorylation of Ser311 during high glucose conditions (Ahuatzi et al. 2007). Since TPS1 deletion does not directly affect SNF1 activity in yeast, we assayed the downstream regulation of Mig1. Confocal fluorescence microscopy with a Mig1-GFP construct indicated that Mig1 translocation is also not affected in the tps1Δ in the presence of glucose (Fig. 3C). As reported before, absence of SNF1 results in constitutive nuclear Mig1 localization, while deletion of HXK2 confers a constitutive cytosolic localization. We can conclude that in yeast T6P is not a direct regulator of SNF1, since neither phosphorylation, kinase activity or Mig1 localization are affected by lack of T6P, but that it acts downstream and independent of SNF1 and Mig1.

Lack of T6P is associated with lack of repression of gluconeogenesis in the presence of glucose

To confirm that T6P indeed acts downstream of SNF1 and Mig1, regulating Cat8 targets, we more specifically investigated gluconeogenesis in WT and the tps1Δ mutant after adding glucose to glucose-starved cells. The two key gluconeogenic genes PCK1 and FBP1 encode phospo-enolpyruvate carboxykinase and fructose-1,6-bisphosphatase, respectively (Fig. S1, Supporting Information). As their activity is only required in the absence of glucose, both enzymes are tightly regulated by glucose availability both at the transcriptional and post-translational level (Horak 2004; Hung et al. 2004; Turcotte et al. 2010).

The expression of both FBP1 and PCK1 is induced in the absence of rapidly fermentable carbon sources and repressed when glucose is available, a process that is mediated by multiple TFs (Fig. 1A) (Turcotte et al. 2010). Shifting cells from galactose to glycerol/ethanol triggers strong induction of gluconeogenic gene expression in WT and tps1Δ mutant cells, indicating that T6P and Tps1 are not involved in the induction of gluconeogenesis (Fig. 4A). Interestingly, transcriptional repression of FBP1 and PCK1 upon addition of glucose to these derepressed cells is completely abolished in the tps1Δ mutant. Restoration of T6P production by expression of E. coli OtsA results in strong but incomplete repression, confirming that T6P (or the TPS protein) is indeed required for transcriptional repression of PCK1 and FBP1 (Fig. 4A). As expected, a snf1Δ mutant did not show induction of FBP1 and PCK1 upon glucose starvation (data not shown). Consistent with gene expression, (genomic) HA-tagged Pck1 and Fbp1 protein levels were maintained after glucose addition to derepressed tps1Δ cells. (Fig. 4B).

Figure 4.

Deficient repression of gluconeogenic Fbp1 and Pck1 expression and activity in response to glucose in the absence of T6P. (A) Relative FBP1 and PCK1 expression assessed by RT-qPCR in WT, tps1Δ and bacterial TPS (pYX242-otsA) expressing tps1Δ mutants. Samples were taken of cells growing on 2% galactose (Gal) medium, after 3 h on 3% glycerol/2% ethanol medium (Gly/Eth) and 30^′, 60^′ and 90^′ after supplementation with 2% glucose (Glc). Shown are mean values with standard error (two-way ANOVA; n = 6; statistical significance is indicated with **P < 0.01, ***P < 0.001 or not significant (ns)). (B) Protein expression of (genomic) HA-tagged Fbp1 (pHAC195-FBP1) and Pck1 (pHAC195-PCK1). Samples were taken of cells growing on 2% galactose (Gal) medium, after 3 h on 3% glycerol/2% ethanol medium (Gly/Eth) and 60^′, 120^′ and 240^′ after supplementation with 2% glucose (Glc). Equal amounts of protein were loaded (20 and 5 μg, respectively, for Fbp1 and Pck1). Equal loading was verified by Coomassie staining of the membrane. Western blot analysis was repeated several times, representative results are shown. (C) FBP and PCK specific enzyme activities in cell free protein extracts. Cells were grown on 2% galactose, samples were taken after 3 h on 3% glycerol/2% ethanol medium (Gly/Eth) and 240^′ after supplementation with 2% glucose (Glc). Shown are mean values with standard error (two-way ANOVA; n = 5; statistical significance is indicated with ***P < 0.001 or ns (not significant)). FBP enzyme activity could not be detected (nd) on glucose medium.

Finally, we corroborated these findings by showing that also FBP and PCK enzymatic activity is maintained in the presence of glucose in the tps1Δ strain (4C). Complementation with OtsA rescues repression of both protein levels and activity (Fig. 4B and C). In a snf1Δ strain, no enzyme activities could be measured (data not shown), consistent with the lack of transcriptional induction and protein expression.

As Cat8 and related TFs control more genes involved in gluconeogenesis, the glyoxylate and TCA-cycles, and alternative carbon source metabolism (Fig. S1, Supporting Information), more expression profiles were evaluated. In addition, FBP1 and PCK1 expression, MLS11 (encoding malate synthase), ICL1 (isocitrate lyase), ACS1 (acetyl-coA synthetase) and CIT1 (citrate synthase) expression regulation was also deficient in the tps1Δ mutant, with sustained high expression levels in presence of glucose (Fig. 5A). These genes all share a CSRE in their promoter region that is controlled by Cat8 (Tachibana et al. 2005; Turcotte et al. 2010), identifying this TF as a possible direct T6P/TPS target. Interestingly, while rescue of the tps1Δ growth defect by snf1Δ is depending on Mig1 and while MIG1 overexpression also consistently partially rescues growth (Fig. 1A), Mig1 does not appear to be necessary for repression of gluconeogenic gene expression (Fig. 5B), consistent with more downstream T6P effects and post-translational regulation of Cat8. However, the low FBP1 and PCK1 expression levels in the tps1Δ snf1Δ mig1Δ mutant (Fig. 5B) suggest that the growth defect in the presence of glucose is uncoupled from persistent gluconeogenesis.

Figure 5.

(A) Alternative carbon source metabolism is affected in the tps1Δ mutant. Relative expression of MLS1, ICL1, CIT1 and ACS1 genes involved in alternative carbon metabolism, assessed by RT-qPCR. Shown are mean values with standard error (two-way ANOVA; n = 4). (B) Role of Mig1 in the regulation of gluconeogenic gene expression. Relative expression of FBP1 and PCK1 in different mig1Δ mutants, assessed by RT-qPCR. Samples were taken of cells growing on 2% galactose (Gal) medium, after 3 h on 3% glycerol/2% ethanol (Gly/Eth) medium and after supplementation with 2% glucose (Glc). Shown are mean values with standard error (two-way ANOVA; n = 4; statistical significance is indicated with **P < 0.01, ***P < 0.001 or ns (not significant)).

We then further explored whether this regulation by T6P/TPS is mediated by Cat8 or Sip4, two of the TFs involved in gluconeogenesis regulation. We generated tps1Δ cat8Δ and tps1Δ sip4Δ double mutants and evaluated their growth (Fig. 6A). Additional deletion of CAT8 or SIP4 does not rescue growth of the tps1Δ mutant on glucose, consistent with the observation that persistent gluconeogenesis per se is not the cause of the tps1Δ growth defect. We then analyzed the gene-expression profiles of FBP1 and PCK1 in these double mutants to verify whether T6P/TPS regulation of gluconeogenic gene-expression is mediated by Cat8 or Sip4 (Fig. 6B). The cat8Δ mutant is unable to efficiently induce FBP1 and PCK1 on non-fermentable medium (consistent with a major role for this TF), however, 1 h after glucose addition expression of both genes is still significantly repressed when compared with expression on glycerol/ethanol medium. In the tps1Δ cat8Δ double mutant this repression is lost; FBP1 and PCK1 expression even increases after addition of glucose. These results are consistent with our previous observations that T6P/TPS is required for repression of FBP1 and PCK1 expression on glucose-containing medium, but also indicate that this effect is not mediated by Cat8. In the sip4Δ and tps1Δ sip4Δ double mutant induction of FBP1 and PCK1 is not affected, but in the tps1Δ sip4Δ double mutant glucose repression is again lost. We conclude that T6P/TPS does not regulate gluconeogenic gene expression through (direct or indirect) inhibition of Cat8 or Sip4 activity.

Figure 6.

The transcription factors Cat8 and Sip4 are not regulated by T6P. (A) Growth analysis assessing the involvement of Cat8 and Sip4 in the growth phenotype of the tps1Δ mutant on glucose-containing medium. Dilution series of WT and mutant cells are spotted on solid rich growth medium containing 2% galactose (YPGal) or 2% glucose (YPGlc). Growth is analyzed after 2–3 days at 30°C. (B) Relative FBP1 and PCK1 expression assessed by RT-qPCR in WT, tps1Δ, cat8Δ, tps1Δ cat8Δ, sip4Δ and tps1Δ sip4Δ mutants. Samples were taken of cells growing on 2% galactose (Gal) medium, after 3 h on 3% glycerol/2% ethanol medium (Gly/Eth) and 30^′, 60^′ and 90^′ after supplementation with 2% glucose (Glc). Shown are mean values with standard error (two-way ANOVA; n = 3; statistical significance is indicated with *P < 0.05, ***P < 0.001 or ns (not significant)).

DISCUSSION

Since the early 1970s, a number of S. cerevisiae mutants with growth defects in the presence of glucose [fdp1 (deficient in the inactivation of fructose-1,6-bisphosphatase) (van de Poll, Kerkenaar and Schamhart 1974), cif1 (deficient in catabolite inactivation of fructose-1,6-bisphosphatase) (Navon et al. 1979; Gonzalez et al. 1992), byp1 (bypass of glycolysis) (Breitenbachschmitt et al. 1984), glc6 (deficient in glycogen accumulation) (Cannon et al. 1994)] have been isolated and later identified as mutants in TPS1, encoding the yeast T6P synthase (Gancedo and Flores 2004). Next to their inability to grow on rapidly fermentable carbon sources, these mutants share pleiotropic defects including accumulation of hexose monophosphates and fructose-1,6-bisphosphate (F-1,6-bP), a dramatic drop in ATP levels, a lack of catabolite inactivation (e.g. of fructose-1,6-bisphosphatase activity) and deficient cAMP-PKA signaling upon addition of glucose (Van Aelst et al. 1993). These diverse phenotypes identify the TPS enzyme and/or its product T6P as important regulators of sugar metabolism and signaling. The accumulation of sugar phosphates indicates a deregulation of the glycolytic pathway, possibly due to excessive influx in glycolysis or futile cycling with gluconeogenesis (Holzer 1976). The identification of T6P as a competitive inhibitor of yeast hexokinase in vitro (Blazquez et al. 1993) and suppression of the tps1 growth defect by Hxk2 mutation (Hohmann et al. 1993; Blazquez and Gancedo 1994) are in line with a key role for this metabolite as gatekeeper of glycolysis. However, the phenotype is more complex and alternative (downstream) effects and mechanisms and a role for the Tps1 protein itself have also been proposed (Blazquez and Gancedo 1995; Gancedo and Flores 2004; Turcotte et al. 2010). Consistently, multiple other extragenic suppressors were identified that restore mutant growth under these conditions. These are basically classified in two categories: those decreasing influx in (upstream) glycolysis and ATP consumption, like hxk2, and those relieving the apparent block in glycolysis at the level of glyceraldehyde-3-P dehydrogenase (GAPDH) e.g. by generating the necessary substrates (like NAD⁺ and P_i) and stimulating ATP producing downstream reactions (Gancedo and Flores 2004). Kinetic modeling recently provided evidence for an imbalanced state of glycolysis in the tps1 mutants due to reduced P_i release from trehalose production (van Heerden et al. 2014). This is consistent with the previously observed suppression of the tps1 growth phenotype by enhanced glycerol production, associated with P_i release (Luyten et al. 1995). But while several hypotheses and metabolic models have been proposed, the exact regulatory functions of T6P and cause for the metabolic and growth defects of T6P deficient cells in the presence of glucose remain elusive.

Interestingly, T6P more recently was also identified as an important regulator of carbon use, stress tolerance and development in plants (Wahl et al. 2013; Nunes et al. 2013a). While T6P does not inhibit plant HXK (Eastmond et al. 2002), it appears to act as an inhibitor of SnRK1 activity (Zhang et al. 2009). We now find that deletion of SNF1 effectively suppresses the growth defect of a yeast tps1Δ mutant (Fig. 1), suggesting that T6P regulates SNF1 signaling in yeast as well and offering the possibility to use yeast as a convenient model to investigate the molecular mechanisms in more detail. Consistent with our findings, different previously identified suppressors are also linked to SNF1 signaling. For example, deletion of SNF4, encoding the activating γ subunit of the SNF1 complex, and overexpression of MIG1 were reported to (partially) restore growth of a tps1Δ mutant in the presence of glucose, although both observations seemed unrelated at the time (Hohmann et al. 1992; Blazquez and Gancedo 1995). First, MIG1 was identified as a multicopy suppressor of the byp1 mutant growth defect in the presence of glucose and it was suggested that overexpression of MIG1 decreases glycolytic flux, thereby restoring growth under these conditions (Hohmann et al. 1992). However, overexpression of MIG1 also partially bypasses its inhibition by SNF1 phosphorylation, thereby reducing activation of alternative carbon metabolism and glucose import and hence partially restoring growth when glucose levels are repleted. Second, the cat3 mutant (deficient in catabolite repression, allelic to SNF4) was identified as an extragenic suppressor of the cif1 mutation (Blazquez and Gancedo 1994, 1995). Surprisingly, the effect on tps1Δ mutant growth seemed background-dependent and Mig1-independent (Blazquez and Gancedo 1995). Moreover, overexpression of the HXT1 low affinity glucose transporter in the tps1Δ snf4Δ mutant completely abolished growth on glucose, leading to the conclusion that deletion of SNF4 results in decreased glucose uptake and glycolytic flux, thereby allowing growth of a tps1Δ mutant in the presence of glucose. Interestingly, Hxk2 also exhibits regulatory functions in SNF1 signaling as a small fraction of Hxk2 localizes in the nucleus to interact with Mig1, thereby inhibiting Mig1 phosphorylation and inactivation during growth in the presence of glucose (Ahuatzi et al. 2004, 2007). However, using Hxk2 mutants deficient in nuclear translocation (Hxk2^K6,7,12A), we show that the loss of its catalytic function in glycolysis is likely responsible for the suppression of the tps1Δ phenotype. Hxk2^K6,7,12A is still able to abolish the growth of a tps1Δ hxk2Δ mutant on glucose (Fig. 1F), indicating that the restored growth of a tps1Δ hxk2Δ mutant is not related to its effects on SNF1 signaling. The unaffected localization of the Mig1 TF further supports this evidence (Fig. 3D).

Here, we explored the possibility of SNF1 being subject to regulation by T6P, similar to the situation in plants. However, in the absence of T6P, Snf1 is still dephosphorylated upon re-addition of glucose to derepressed cells. To confirm SNF1 inactivation, we constructed a convenient in vivo reporter for SNF1 activity, which will also greatly facilitate future research into the molecular mechanisms of its regulation. Furthermore, the cytoplasmic-nuclear translocation of the key target TF Mig1 is unaffected, confirming that T6P is not a direct regulator of SNF1 activity (Fig. 3). Then how does deletion of SNF1 rescue the growth defect of tps1Δ mutant cells on glucose? SNF1 signaling is essential for the induction of genes required for growth on non-fermentable carbon sources, including genes involved in respiration and gluconeogenesis. However, different respiratory inhibitors are not able to rescue the growth defect of tps1Δ, while overexpression of activated Cat8 restored the tps1Δ snf1Δ growth defect. This initially pointed at active gluconeogenesis as a possible target responsible for the growth and metabolic defect on glucose. More specifically, we observed an elevated expression of the key gluconeogenic genes FBP1 and PCK1, both at the mRNA and protein level, leading to persistent, high FBP and PCK activity on glucose medium (Fig. 4). This supports the idea of an ATP-depleting futile cycle caused by simultaneous activation of glycolysis and gluconeogenesis. However, strains overexpressing FBP1 and PCK1 did not cause a growth defect on glucose-containing medium, even though gluconeogenic activity was significantly higher (Navas, Cerdan and Gancedo 1993). Possibly, this is only problematic in combination with an overactive glycolysis, e.g. due to uncontrolled Hxk activity.

In the recently proposed kinetic model of glycolysis mentioned above (van Heerden et al. 2014), trehalose metabolism provides the necessary P_i to stimulate downstream glycolytic flux to keep up with upstream sugar-phosphate production, an intrinsically risky design that results in an imbalanced state in individual cells due to (non-genetic) metabolic variability. The tps1 mutant, lacking this P_i regeneration step, predominantly exhibits the imbalanced-state phenotype with (experimentally confirmed) small isogenic subpopulations of cells with a global steady state. In such a model, however, increased Fbp1 activity due to snf1Δ would rather alleviate the tps1 transition problem. Another possibility is that SNF1-mediated expression of (high affinity) glucose transporters on non-fermentable growth medium contributes to the imbalanced state and transition problem in tps1 mutants upon glucose addition.

Remarkably, the glucose growth defect and the deregulation of glycolysis can also be uncoupled using a T6P-insensitive Hxk (Bonini, Van Dijck and Thevelein 2003). Consistently, we find that, while rescue of the tps1Δ growth defect by snf1Δ is depending on Mig1 and while MIG1 overexpression partially rescues growth (Fig. 1E), Mig1 is not required for repression of gluconeogenic gene expression (Fig. 5B).

Although it is not yet clear to what extend its misregulation contributes to the tps1 growth defect on rapidly fermentable growth medium, our analyses show that while SNF1 is essential for the induction of gluconeogenesis, T6P/TPS is required for inactivation of gluconeogenesis on glucose medium, downstream and independent of SNF1 activity. In line with this, the reg1Δ mutant, unable to dephosphorylate Snf1 in the presence of glucose, was reported to still exhibit repression of gluconeogenesis, while loss of glucose repression of genes involved in the utilization of alternative carbon sources was observed (Schuller 2003). This also provides additional evidence that T6P does not directly regulate SNF1 activity in yeast, but rather affects specific downstream targets. We focused on the gluconeogenic enzymes Fbp1 and Pck1 but found that associated pathways like the glyoxylate shunt (with ICL1 and MLS1 encoding two key enzymes in that process) and alternative carbon metabolism (with ADH2 and GUT1 encoding enzymes involved in alcohol and glycerol metabolization, respectively) and cytosolic malate dehydrogenase (encoded by MDH2, linking mitochondrial and peroxisomal metabolism to gluconeogenesis) are similarly regulated. The expression of all these genes is tightly regulated by a set of TFs with a predominant role for the CRSE-motif interacting zinc cluster protein Cat8 (Hedges, Proft and Entian 1995; Tachibana et al. 2005). FBP1 and PCK1 are weakly or not induced/activated on glycerol/ethanol medium in a cat8Δ mutant (Hedges, Proft and Entian 1995; Haurie et al. 2001). This TF is itself transcriptionally regulated by Mig1 and activated by Snf1 through phosphorylation. However, more TFs have been implicated in the derepression of non-fermentable carbon metabolism, including Sip4, Adr1, Rds2, Etr1, Gsm1 and Znf1, all acting downstream of SNF1 and each with a slightly different in target specificity (Lesage, Yang and Carlson 1996; Vincent and Carlson 1998; Tachibana et al. 2005; Soontorngun et al. 2007, 2012). The Sip4 TF is also dependent on SNF1 activity and regulates the expression of CRSE-containing promotors in collaboration and partly redundant with Cat8. However, Sip4 is not essential for derepression, since no specific knock-out phenotype is observed, while overexpression can complement the loss of Cat8 (Lesage, Yang and Carlson 1996; Vincent and Carlson 1998). Our analyses appear to disprove the possibility that the Cat8 and Sip4 TFs are targets of T6P regulation. Although Cat8 appears to be the prime TF, T6P/TPS might affect alternative carbon metabolism by inhibition of other TF involved in the regulation of alternative carbon metabolism. However, since a broad variety of genes, regulated by at least partially redundant TFs, is affected, T6P regulation of the more basal regulatory machinery, such as components of the general repressor (e.g. Tup1 and Ssn6/Cyc8) or mediator complexes, should not be excluded. Regulation of gluconeogenic mRNA stability (Mercado et al. 1994; Yin, Smith and Brown 1996; Yin, Hatton and Brown 2000; Braun et al. 2011; Braun, Dombek and Young 2015) and post-translational regulation of enzyme activity add an additional level of complexity that might also involve T6P/TPS control.

SUPPLEMENTARY DATA

Supplementary Data.

The authors thank Johan Thevelein, Stefan Hohmann, Marian Carlson, Mark Johnston, Hans-Joachim Schüller and Fernando Moreno for strains and plasmids, Tuong Thi Vi Dang (Vi) for expert help with western blot analysis, Catherina Coun for double mutant isolation, Pascale Daran-Lapujade and Marijke Luttik (TU Delft) for their hospitality and expert help with measurement of enzyme activities, and Bas Teusink and co-workers for helpful suggestions.

FUNDING

Work in the labs of Molecular Plant Biology and Functional Biology was supported by FWO, IWT and KU Leuven (OT).

Conflict of interest. None declared.