Saccharomyces cerevisiae contains a large family of (putative) hexose transporters (Boles and Hollenberg, 1997; Kruckeberg, 1996). The hexose transporter family includes HXT1–HXT17, GAL2, SNF3 and RGT2. HXT1–HXT4 and HXT6–HXT7 encode the major proteins that transport the hexoses glucose, fructose and mannose over the plasma membrane (Reifenberger et al., 1995). SNF3 and RGT2 encode two proteins that are involved in the sensing of the available glucose (Özcan et al., 1996), and GAL2 encodes the galactose permease (Tschopp et al., 1986).
The kinetic parameters of the individual HXT gene products have been determined in an hxt null background (Reifenberger et al., 1997). It was shown that HXT1 and HXT3 encode low-affinity glucose transporters, HXT2 and HXT4 encode moderate-affinity glucose transporters, and HXT6 and HXT7 encode high-affinity glucose transporters. A strain in which HXT1–HXT4 and HXT6–HXT7 were deleted (the RE605 strain) has been reported to be not significantly different from a hxt1–hxt7 null strain (the RE700 strain), in that it was unable to grow on glucose (Reifenberger et al., 1995). In another yeast parental strain, all HXT genes have been shown to be able to restore growth on different hexoses (except HXT12) when overexpressed individually in an hxt1-hxt17 gal2 null strain (Wieczorke et al., 1999).
The hexose transport proteins are differentially regulated depending on the source and the amount of carbon (Boles and Hollenberg, 1997; Özcan and Johnston, 1995, 1999). In a previous study on the expression of the members of the hexose transporter family during various conditions in a wild-type strain, including aerobic batch cultivation on glucose and aerobic glucose limited chemostat cultivation, it was confirmed that the transcription of HXT1–HXT4 and HXT6–HXT7 correlated with the extracellular glucose concentration (Diderich et al., 1999). In this same study a consistent relationship was observed between the expression of the individual HXT genes and the glucose transport kinetics determined from the uptake of 14C-glucose. Surprisingly, transcription of HXT5, a functionally non-characterized member of the hexose transporter family, was abundant at low growth rates and/or conditions of low glucose. In addition, the presence of HXT5 transcript did seem to correlate with the expression of a high-affinity hexose transport component under several conditions.
The gene HXT5 is located on chromosome III upstream of HXT1 and HXT4, and its product encodes a protein of 592 amino acids with 12 (putative) membrane spanning helices, a property common to all members of the major facilitator superfamily. The protein encoded by HXT5 (YHR096c) is described in the Yeast Transport Protein Database as a protein with strong similarity to the hexose facilitators and an unknown function (http://alize.ulb.ac.be/YTPdb/), as a probable glucose transporter in the Swiss-Prot database (http://www.expasy.ch/sprot/) and in the Yeast Protein Database (http://www.proteome.com) as a probable integral plasma membrane protein involved in small molecule transport with high similarity to other hexose transporters from S. cerevisiae and other organisms from prokaryotes to mammals. Expression of HXT5 is induced after a shift to high osmolarity by either 0.7 M NaCl or 0.95 M sorbitol (Rep et al., 2000). Transcript levels of HXT5 increase strongly approaching and beyond the diauxic shift during growth on glucose (DeRisi et al., 1997; Diderich et al., 1999). A study in which the promoter of HXT5 was fused to lacZ implied the presence of Hxt5p during growth on glycerol and suggested that expression of HXT5 is induced by low glucose (Özcan and Johnston, 1999). One study (Klaassen and Raamsdonk, 1998) reported fermentation in a strain that contained only HXT5 and none of the major hexose transporters HXT1–HXT4 and HXT6–HXT7; however, no further details were given.
In the present study the kinetic characteristics of Hxt5p have been determined and an attempt was made to reveal the function of the protein encoded by HXT5.
Materials and methods
Strains, media, and growth conditions
The strains of S. cerevisiae used in this study are given in Table 1. Aerobic batch cultivations were performed at 30°C on a rotary shaker at 250 rpm (Gallenkamp) in Erlenmeyer flasks. Cells were grown in a rich medium which contained 2% (w/v) yeast extract, 1% (w/v) peptone (YEP) and a carbon source as specified in the text, or in a minimal medium that consisted of 0.67% (w/v) yeast nitrogen base (YNB) and 0.1 M potassium phthalate, pH 5.0, and a carbon source as specified in the text. To induce sporulation, cells were incubated on sporulation medium as described in Sherman (1991).
Table 1. S. cerevisiae strains
The KY92, RE605A and RE700A strains listed are derived from the MC996A strain and are all ura3-52 his3-11,15 leu2-3,112 MAL2 SUC2 GAL MEL (Reifenberger et al., 1995).
Dr P. Kötter (Frankfurt, Germany)
MATα MAL2-8c SUC2
Dr P. Kötter (Frankfurt, Germany)
MATα MAL2-8c SUC2 HXT5::GFP
Dr P. Kötter (Frankfurt, Germany)
MATaMAL2-8cSUC2 leu2-3,112 Δhxt5::LEU2
MATaMAL2 SUC2 GAL MEL ura3-52 his3-11,15 leu2-3,112
The MSY1 strain was created by disruption of HXT5 in strain CEN.PK111-32D with an hxt5::LEU2 disruption cassette from plasmid pD5 (Reifenberger et al., 1995). Leu+ isolates were checked for correct integration of LEU2 by Southern blot analysis with a KpnI restriction fragment of pD5 as a probe for HXT5 and with an EcoRV–ClaI fragment from pRS425 (Sikorski and Hieter, 1989) as a probe for LEU2. The probes were isolated from agarose gels using the Qiagen gel extraction kit. The isolated probes were then labelled with α-[32P]-dATP (Amersham), using the Prime-a-Gene labelling kit from Promega. Total yeast genomic DNA was isolated according to Hoffman and Winston (1987), cut with HindIII and transferred to nylon membranes using a vacuum blotter (BioRad). The membranes were washed for 1 h in 1× SSC/0.1% SDS, and UV cross-linked in a UV oven (Stratagene). The membranes were then washed again for 4 h at 68°C in a Hybaid oven in hybridization buffer (6× SSC, 5× Denhardt's solution, 1% SDS and 0.2 g/l boiled salmon sperm DNA). The membranes were hybridized with the probes overnight at 68°C in the Hybaid oven. The blots were washed in 7× SSC/0.1% SDS, 1× SSC/0.1% SDS and 0.1× SSC/0.1% SDS, respectively, and exposed to Kodak X-Omat AR film.
The KY98 strain was constructed by tagging HXT5 with the gene encoding enhanced green fluorescent protein (GFP) in the wild-type strain CEN.PK113-7D as follows. The plasmid pFA6a–GFPS65T–kanMX6 (Wach et al., 1994) was amplified by PCR using oligonucleotides AK41 (AATGATCCGAGACCATTTTATAAAAGGATGTTCACTAAAGAAAAAAGTAAAGGAGAAGAACTTTTC; the underlined nucleotides correspond to the DNA immediately 5′ of the HXT5 stop codon) and AK42 (AACATTGCAAGTATGCGAAAATAGTTGATCCTACACTACAAGAGAGGATGGCGGCGTTAGTATC; the underlined nucleotides correspond to the DNA immediately 3′ of, and including, the HXT5 stop codon). The resulting DNA was transformed into CEN.PK113-7D and the transformants were selected for geneticin resistance. Proper integration of the GFP cassette was confirmed by PCR with primers AK42 and AK43 (GCTACTAGAAATGATCCGAG; annealing at base 1722 in the HXT5 open reading frame).
Strain KY92 was constructed in an identical manner, resulting in HXT5 tagged with GFP in strain RE605, which is disrupted in HXT1–HXT4 and HXT6–HXT7 (Reifenberger et al., 1995).
Using standard genetic techniques (Sherman, 1991), the diploid strain KY132 (MATa/MATα HXT5/HXT5::GFP) was produced by crossing strain KY98 with strain CEN.PK113-1A. KY132 was sporulated and strain KY126 (MATα HXT5::GFP) was recovered. The diploid strain KY152 (MATa/MATα HXT5::GFP/HXT5::GFP) was produced by crossing strains KY98 and KY126.
Northern blot analysis
Total RNA was isolated from the yeast cells by acid–phenol extraction. RNA samples were separated by electrophoresis in 1% agarose formaldehyde gels (Sambrook et al., 1989). Transfer to nylon membranes, prehybridization, hybridization with an HXT5-specific oligonucleotide probe and washing were performed as described in (Diderich et al., 1999).
Zero trans-influx of glucose
Cells were harvested by centrifugation at 4°C (5 min at 4000×g), washed three times in ice-cold 0.1 M KH2PO4 buffer (pH 6.5), and then kept on ice in 0.1 M KH2PO4 buffer (pH 6.5) until further use. Zero trans-influx of glucose, fructose, mannose and galactose was determined according to Walsh et al. (1994) at 30°C in 0.1 M phosphate buffer (pH 6.5). The kinetic parameters of glucose transport were derived using Enzfitter software.
In vivo dynamics during glucose consumption
Cultured and washed yeast cells were preincubated for 3 min in a stirred thermostated vessel at 30°C; then glucose in 0.1 M phosphate buffer (pH 6.5) was added. Samples were taken at different times by adding a volume of the glucose-metabolizing cells to an equal volume of 10% (w/v) perchloric acid, which was kept on ice. Samples were neutralized within 1 h after extraction with 2 M K2CO3 and stored at −20°C. Before analysis, samples were centrifuged for 1 min at 16 000×g. Samples were analysed for extracellular and intracellular metabolites by NAD(P)H-coupled enzymatic reactions on a COBAS-FARA automatic analyser (Roche).
Cells were harvested from cultivation and embedded in molten 1% low-melt agarose, then chilled briefly. They were immediately examined with a Leitz Aristoplan epifluorescence microscope. Micrographs were recorded using an Apogee CCD camera and processed for display using Image-Pro Plus and Adobe Photoshop.
Expression of HXT5
The expression of a gene under a particular set of conditions often hints to a function under those conditions. To analyse the function of the protein product of HXT5, transcript levels and abundance of the Hxt5 protein tagged with GFP were measured under different conditions.
The abundance of Hxt5p–GFP was low during exponential growth on fermentable carbon sources (Table 2). However, upon approaching depletion of the carbon source, the HXT5::GFP fusion was expressed, as apparent from an increase in plasma membrane-localized GFP (see also Figure 1). In contrast, during growth on galactose, which confers a lower growth rate than glucose, Hxt5p–GFP was already present and membrane localized during exponential growth. Similarly, it was observed that during growth on non-fermentable carbon sources, i.e. ethanol, glycerol and a mixture of ethanol and glycerol, when growth was slow or almost absent, Hxt5p–GFP was present and plasma membrane localized. Interestingly, HXT5 expression was most abundant (through visual inspection) with 2% glycerol as a carbon source, when growth was almost absent.
Table 2. Expression of Hxt5p during growth on various carbon sources
Growth rate (h–1)
Abundance of Hxt5p–GFP during exponential growth on carbon source
Abundance of Hxt5p–GFP after depletion of carbon source
The HXT5::GFP fusion strain (KY98) was grown in minimal media containing YNB in 100 mM phthalic acid (pH 5.0) with 2% of various carbon sources. The abundance of Hxt5p–GFP was followed during growth by fluorescence microscopy by visual inspection. +, membrane-associated GFP expression; –, no membrane-associated GFP expression.
nd, not determined.
2% EtOH+2% Glycerol
The expression of HXT5 was studied in more detail during growth on 2% glucose (Figure 1). Hxt5p–GFP expression in KY98 was followed during growth by fluorescence microscopy (Figure 1B) and compared with phase contrast microscopy (Figure 1C), HXT5 transcript levels (Figure 1A), and the concentration of residual glucose (Figure 1).
During exponential growth on glucose, transcription of HXT5 was absent; only upon approaching glucose depletion were HXT5 transcript levels increased (Figure 1A). After glucose depletion, the transcription of HXT5 increased even further, and the expression of the fusion protein Hxt5p–GFP became visible and was mostly plasma membrane-localized (Figure 1B, C), although some cells showed GFP localized in the cell interior (Figure 1B, at t=8 h).
HXT5 is highly expressed during sporulation (Figure 2). In heterozygous and homozygous diploid cells of HXT5::GFP, plasma membrane-localized Hxt5p–GFP was abundant during sporulation. Spore formation was confirmed by short-wave fluorescence microscopy, which visualizes dityrosine in the ascospore wall. In the homozygous cells (Figure 2A; KY152; HXT5::GFP/HXT5::GFP), Hxt5p–GFP was present in all four spores. In the heterozygous cells (Figure 2B; KY132; HXT5/HXT5::GFP), Hxt5p–GFP was only present in the two spores that contained the HXT5::GFP fusion, which implies that HXT5 is expressed after spore formation.
Kinetic parameters of HXT5
The kinetic parameters of the protein encoded by HXT5 were determined in RE605, a strain that does not possess HXT1–HXT4 and HXT6–HXT7 (Reifenberger et al., 1995). Growth on glucose of the RE605 strain has been reported to be not significantly different from RE700, a strain lacking HXT1–HXT7 and unable to grow on glucose (Reifenberger et al., 1995). In our hands, RE605 showed slow growth on YEP/2% glucose plates. In contrast, RE700 produced a large number of revertants on a background of non-growing cells.
To determine the kinetic parameters of hexose transport by Hxt5p, the RE605 strain was grown on YEP 2% ethanol–glycerol, in which HXT5 is highly expressed, yet growth is present (Table 2). In a parallel experiment, RE700 and KY92 (the RE605 strain with a GFP-tag on HXT5) were grown. Again, in all strains growth was slow on ethanol–glycerol, a feature also shown by wild-type yeast strains (results not shown and Table 2). The kinetic parameters of hexose transport were determined by zero trans-influx of glucose assays with 14C-hexose (Walsh et al., 1994). In the RE700 strain glucose (Figure 3), fructose, galactose and mannose uptake was virtually undetectable (during 5 and 10 s incubation). In contrast, the RE605 strain showed a moderate affinity of 10±1 mM for glucose (Figure 3) and a low-affinity of 40±4 mM for fructose, with a maximal velocity (Vmax) of 150±20 nmol/min/(mg protein) and 150±10 nmol/min/(mg protein), respectively. Although some affinity for mannose was present, it was too low to obtain an accurate measure of the Km (>100 mM). Affinity for galactose was absent in the RE605 strain. The parallel experiment with the KY92 strain confirmed the presence of Hxt5p under the conditions assayed. The GFP tag on the Hxt5 protein did not change the kinetic parameters of Hxt5p to a great extent, and an affinity constant for glucose of 11±1 mM was determined (Figure 3).
In vivo flux through HXT5
As mentioned before, in a study on the individual hexose transporters it has been reported that the RE605 strain does not grow on glucose (Reifenberger et al., 1995). However, as described above, Hxt5p is able to transport glucose, fructose and mannose over the plasma membrane. Furthermore, it was described that the RE605 strain will ferment supplemented glucose to ethanol and carbon dioxide after growth on maltose (Klaassen and Raamsdonk, 1998). Here we determined the flux through glycolysis under a defined condition in the RE605 strain and compared this with the RE700 strain that lacks any of the functional hexose transporters.
The RE605 and RE700 strains were harvested during exponential growth on YEP containing 2% ethanol/glycerol and resuspended in potassium phosphate buffer at pH 6.5. Flux through glycolysis was determined by measuring glucose consumption and ethanol production (Figure 4A, B). In addition, the intracellular concentrations of glucose-6-phosphate, fructose-6-phosphate and ATP were followed in time to determine intracellular responses in the presence of Hxt5p as sole hexose transporter or in the absence of any functional hexose transporter (Figure 4C, D).
The RE605 strain showed a glucose consumption rate of approximately 90 nmol/min/(mg protein) (Figure 4A). This approximately equals the zero trans-influx of glucose at that glucose concentration, as determined from the kinetic parameters of Hxt5p (Figure 3). Only a small fraction of the glucose was converted into ethanol during glucose consumption (Figure 4A). In the RE700 strain, glucose consumption was absent (Figure 4B). In the RE605 strain, the concentration of the intracellular metabolites glucose-6-phosphate and fructose-6-phosphate increased rapidly after the addition of glucose and remained at that same level during glucose consumption (Figure 4C), while the concentration of ATP was not affected by the pulse of glucose. Remarkably, in the RE700 strain the intracellular ATP concentration first increased after the addition of glucose, and thereafter steadily decreased to the level before addition of glucose (Figure 4D). This suggests metabolic activity in the RE700 strain as a consequence of the addition of glucose, despite the absence of measurable glucose consumption.
Functional analysis of HXT5
As was shown above, transcription of HXT5 and presence of Hxt5p tagged with GFP was prominent under specific conditions. However, from these data a function for Hxt5p was not evident. To obtain a better insight into the function of the protein encoded by HXT5, the physiological characteristics of a wild-type strain deleted in HXT5 were determined.
The growth rates of the hxt5 deletion strain, MSY1, and the wild-type CEN.PK113-7D strain were virtually identical on minimal medium with 2% glucose as a carbon source. After glucose was depleted, the cells were grown to stationary phase (i.e. left for 24 h after glucose depletion) and then shifted to fresh minimal medium containing 2% glucose. Both in the wild-type strain and in the hxt5 deletion strain growth resumed after a lag period of approximately 3 h; however, growth, according to measurements of the optical density, was reproducibly slower in the hxt5 deletion strain (Figure 5).
During aerobic batch growth on glucose the affinity for glucose of the transport step changes (Walsh et al., 1994). Low-affinity hexose transporters are present when the glucose concentration is relatively high, while high-affinity hexose transporters are present when the glucose concentration is relatively low. S. cerevisiae seems optimally adjusted to be able to grow on a broad range of glucose concentrations and to respond to small changes in the availability of glucose through subtle differences in the regulation of various hexose transporters (Boles and Hollenberg, 1997; Özcan and Johnston, 1995, 1999). The multitude of hexose transporter genes seems to be a consequence of ongoing duplications and evolution and to have resulted in optimal growth or survival under a variety of conditions. As an example, glucose-limited continuous cultures that were maintained for more than 250 generations showed cells that had evolved metabolism optimized for growth with a high biomass yield (Ferea et al., 1999). Multiple duplications in the genes encoding hexose transporters were shown to occur under these conditions (Brown et al., 1998). Also, the presence of Hxt5p can be seen as a consequence of duplication of a hexose transporter, evolved for a function under particular conditions. Since Hxt5p resembles the proteins that constitute the major hexose transporters during growth on glucose (Hxt1p–Hxt4p and Hxt6p–Hxt7p), a role for Hxt5p in hexose transport seems probable. The deletion of HXT5 is not lethal and does not result in a clear phenotype under the conditions tested. Therefore, it cannot be excluded that the existence of HXT5 is a consequence of a non-advantageous duplication, which resulted in a redundant or non-functional gene. However, the fact that HXT5 is transcribed and subsequently translated into a functional membrane protein with moderate capacity to transport hexoses under particular conditions (suggesting regulation of transcription), properties prone to mutations in time and evolution, strongly suggests that HXT5 encodes a functional protein and is advantageous to the cell.
As was shown by Wieczorke et al. (1999), none of the proteins encoded by the HXT genes is essential for viability. Instead, overexpression of any of the hexose transporters (except Hxt12p) in a hxt1-hxt17 gal2 deletion strain could restore growth on one or more hexoses. Here we confirmed that the deletion of HXT5 was not lethal. Only under conditions of starvation might the absence of Hxt5p result in a longer lag phase in growth when shifted to fresh glucose medium.
Diploid cells produce haploid cells in response to starvation through the process of sporulation (Chu et al., 1998; Mitchell, 1994). Sporulation in yeast involves meiosis and spore morphogenesis, processes which can be characterized by the sequential transcription of at least four sets of genes, expressed early, middle, mid-late and late during sporulation (Chu et al., 1998; Mitchell, 1994). HXT5 seems to be expressed in the later stages of sporulation (i.e. maturation), since in the heterozygous HXT5–GFP strain (Figure 2) Hxt5p–GFP was only present in two out of four spores, suggesting HXT5 expression after spore formation. Earlier studies on the transcription of genes during the course of meiosis and spore formation confirm that HXT5 transcription is induced in the later stages of sporulation (Chu et al., 1998).
The addition of glucose to cells grown on non-fermentable carbon sources (e.g. ethanol) induces a rapid cAMP signal (Beullens et al., 1988), a trigger for the RAS–adenylate cyclase pathway (Tadi et al., 1999; Thevelein, 1991, 1992), which is responsible for transducing the availability of extracellular nutrients, and governs the progression of the cell cycle. The abundance of HXT5 during nutrient limitation and sporulation might suggest that HXT5 is involved in the generation of an intracellular signal of nutrient availability (e.g. intracellular glucose). However, the fact that the RE700 strain showed some changes in intracellular metabolites after the addition of glucose, despite a lack of glucose consumption, implies that also in the absence of glucose uptake extracellular signals are transduced into the cell.
Alternatively, the fact that glucose seems to be transported out of the cell in ethanol-limited continuous cultures, concurrently with the abundance of HXT5 transcript under these conditions (Diderich et al., 1999), suggests a role for Hxt5p in the efflux of glucose. At first sight this seems futile; however, glucose efflux under conditions when glucose repression is unfavourable may involve a mechanism in which intracellular signals that would induce glucose repression (e.g. intracellular glucose) are suppressed by the secretion of intracellular glucose by a hexose transporter.
In conclusion, in the present study it is shown that Hxt5p is a functional hexose transporter with sufficient capacity to sustain a flux through glycolysis when present as the sole hexose transporter. The presence of a glucose transporter when glucose is absent or low suggests that HXT5 transcription is repressed by glucose. We suggest a role for Hxt5p in the initial uptake of glucose, when glucose is absent and becomes available again. Alternatively, we propose a function of Hxt5p in protecting the cell against switching on its glucose-repressing mechanism under inappropriate conditions.
We thank André Boorsma for technical assistance and Barbara Bakker and Joost Teixeira de Mattos for critical reading of the manuscript. This work was financially supported by the Council for Chemical Sciences of The Netherlands Organization for Scientific Research (CW-NWO), The Netherlands Association for Biotechnological Research Centres (ABON), and the European Union through Grant No. BIO4-CT95-0107 of the BIOTECH program. A.L.K. was supported by EU Grant No. BIO4-CT98-0562, DG12-SSM1.