OUP user menu

Hexose and pentose transport in ascomycetous yeasts: an overview

Maria José Leandro, César Fonseca, Paula Gonçalves
DOI: http://dx.doi.org/10.1111/j.1567-1364.2009.00509.x 511-525 First published online: 1 June 2009


The biochemical characterization of sugar uptake in yeasts started five decades ago and led to the early production of abundant kinetic and mechanistic data. However, the first accurate overview of the underlying sugar transporter genes was obtained relatively late, due mainly to the genetic complexity of hexose uptake in the model yeast Saccharomyces cerevisiae. The genomic era generated in turn a massive amount of information, allowing the identification of a multitude of putative sugar transporter and sensor-encoding genes in yeast genomes, many of which are phylogenetically related. This review aims to briefly summarize our current knowledge on the biochemical and molecular features of the transporters of hexoses and pentoses in yeasts, when possible establishing links between previous kinetic studies and genomic data currently available. Emphasis is given to recent developments concerning the identification of d-xylose and l-arabinose transporter genes, which are thought to be key players in the optimization of S. cerevisiae strains for bioethanol production from lignocellulose hydrolysates.

  • yeast carbon metabolism
  • yeast transporter genes
  • xylose transport
  • arabinose transport
  • fructose transport
  • sugar proton symporter


Simple sugars are the favoured carbon and energy sources of ascomycetous yeasts, among which Saccharomyces cerevisiae and Schizosaccharomyces pombe stand out as well-established model organisms. Considerable research effort was devoted over the past decades to several aspects of sugar metabolism in yeasts, showing that the uptake into the cell has a high level of control over the metabolic flux (Elbing, 2004) and that it impacts catabolic repression and the relative fraction of the sugar that is fermented and respired (Goffrini, 2002; Barnett & Entian, 2005).

The first biochemical and physiological studies on sugar transport in yeasts date back to the 1960s (Cirillo, 1962; Kotyk, 1965), and the first monosaccharide transporter gene was isolated c. 20 years later (GAL2, encoding the galactose permease in S. cerevisiae) (Tschopp, 1986). Description of the GAL2 gene preceded, in turn, the identification of the complete set of the most physiologically relevant hexose transporter (HXT) genes by nearly a decade. This was mainly because of the genetic complexity of hexose transport in S. cerevisiae (c. 20 genes are involved), which hampered a straightforward identification of genes by functional complementation. The complete or partial sequences of the genomes of many ascomycetous yeasts are available and have been scrutinized for the presence of putative sugar transporters (Palma, 2007), considerably improving our knowledge concerning the evolution and phylogeny of this important group of proteins (Palma, 2009).

General characteristics of yeast sugar transporters

The sugar porter (SP) family is the largest within the major facilitator superfamily (MFS), and includes proteins from bacteria, archaea and eukarya, with very diverse sequences and functions (Maiden, 1987; Baldwin & Henderson, 1989; Henderson & Maiden, 1990). The Transporter Classification (TC) system is used to catalogue transmembrane transporter proteins, relying on both functional and phylogenetic data (Saier, 2006; http://www.tcdb.org/). Virtually all characterized yeast sugar transporters are classified in this system. Recently, 95 potential MSF proteins from the yeast Candida albicans were identified and assigned to 17 different TC families, using a comprehensive bioinformatics approach (Gaur, 2008). Another useful source of information is the Yeast Transport Protein Database (http://rsat.ulb.ac.be/ytpdb/), which provides access to annotations on yeast proteins classified as established or predicted membrane transporters.

Proteins belonging to the MSF family exhibit strong structural conservation although they may share little sequence similarity (Vardy, 2004). Generally, these permeases consist of a single integral membrane protein with two sets of six hydrophobic transmembrane-spanning (TMS) α-helices connected by a hydrophilic loop (Fig. 1). Many monosaccharide transporters in yeasts operate by facilitated diffusion, an energy-independent mechanism that catalyses the movement of a single solute molecule at a time down its concentration gradient, across a membrane. Other yeast monosaccharide transporters, which usually operate only when the amount of sugar in the medium is scarce, are energy-consuming systems that are able to transport a solute against its concentration gradient, coupled to the simultaneous movement of (a) proton(s). Finally, a few yeast members of the SP subfamily (e.g. Snf3p) are unable to transport sugars but rather seem to act as receptors, signalling the presence of sugars at the cell surface, thereby influencing the levels of gene expression (Lewis & Bisson, 1991; Ko, 1993; Coons, 1995; Liang & Gaber, 1996; Kruckeberg, 1998).


Topology of yeast sugar transporters. Generally, monosaccharide transporters in yeasts have 12 hydrophobic transmembrane domains (represented as cylinders in the cartoon). The position of the five conserved sequence motifs that have been recognized in sugar transporters is represented in the figure by the letters (A–E). Their respective sequences are as follows: A, RXGRR [the first and last R may be replaced by K; X is generally an amino acid with a large hydrophobic side chain (Henderson & Maiden, 1990)]. An extended version of the same motif can also be found: G-[RKPATY]-L-[GAS]-[DN]-[RK]-[FY]-G-R-[RK]-[RKP]-[LIVGST]-[LIM] (Pao, 1998). B, the ‘diffused’ motif R-X3-G-X3-G-X6-P-X-Y-X2-E-X6-R-G-X6-Q-X5-G through transmembrane domains four and five (Henderson, 1990; Griffith, 1992); the conserved glutamate (E) and a glycine (G) are thought to have, respectively, important catalytic and structural roles (Pao, 1998); C, PESPRXL (Henderson, 1990; Griffith, 1992); D, [LI]-Q-X2-Q-Q-X-[ST]-[GN]-X3-Y-Y-F (Horak, 1997); E, PETKGXXXE (Henderson, 1990; Griffith, 1992).

Five conserved motifs have been found in sugar transporters (Fig. 1), irrespective of their mechanism or substrate specificity (Maiden, 1987; Henderson & Maiden, 1990; Griffith, 1992; Horak, 1997; Pao, 1998), which considerably facilitated the identification of putative novel sugar transporter genes. The SP subfamily in yeasts is one of the most variable in the number of members, with 34 transporters in S. cerevisiae, 17 in Candida glabrata, 20 in Kluyveromyces lactis, 48 in Debaryomyces hansenii and 27 in Yarrowia lipolytica (De Hertogh, 2006). Genes involved in hexose transport and the most important characteristics of the proteins they encode are listed in Table 1, while the biochemical properties of the pentose-uptake systems identified in yeasts so far are listed in Table 2.

View this table:

Genes encoding functionally characterized hexose transporters and structurally related glucose sensors in yeasts

YeastTransporter (aa) [FD vs. SYMP]Uptake Substrate(s)Km (mM)SimilarityReferences
Candida albicansHgt1Glucose194% to Hgt2Varma (2000); Fan (2002)
(545)51% to KlHgt1
Hgt4Glucose sensorNA54% to Rgt2Fan (2002); Brown (2006); Arnaud (2007)
(748)46% to Snf3 and Rag4
Hgt12GlucoseND48% to Snf3Fan (2002); Luo (2007); Arnaud (2007)
(526)FructoseND46% to Hgt4 and Rgt2
[NC]MannoseND28–31% to Hxt1-7
Hansenula polymorphaHxs1Glucose sensorNA54% to Snf3Stasyk (2008)
(638)53% Rgt2
[NA]44% to Gcr1
Hxt1GlucoseND60% to Hxt3 andStasyk (2008)
[NC]37% to Gcr1
Kluyveromyces lactisFrt1Fructose0.1670% to Fsy1Diezemann & Boles (2003)
Hgt1Glucose125–30% to Hxt family and Rag1Billard (1996); Diezemann & Boles (2003); Dujon (2004); Baruffini (2006)
Kht1Glucose13Identical to Rag1 except for the last 2 residues of Kht1 C-terminusWeirich (1997); Milkowski (2001)
Kht2Glucose3.775% to Rag1Weirich (1997); Milkowski (2001)
(566)75% to Hxt6 and Hxt7
[FD]73% to Gal2
Kht3GalactoseND47% to Hxt14Wiedemuth & Breunig (2005)
(528)42% to Rag1
[ND]39% to Kht2
Rag1Glucose20–5078% to Hxt1Wésolowski-Louvel (1992); Goffrini (2002); Diezemann & Boles (2003)
(567)FructoseND73% to Hxt3
[FD]69% to Hxt2
67% to Gal2
Rag4Glucose sensorNA52–53% to Snf3 and Rgt2Betina (2001)
Pichia stipitisSut1Glucose1.565–67% to Hgt6, Hgt7 and Hgt8Weierstall (1999)
(553)Fructose3660% to Rag1 and Hxt5
[FD]Xylose14557% to Kht2
Sut2Glucose1.1 and 5577% to Sut1Weierstall (1999)
(550)59% to Rag1
[FD]Xylose4958% to Hxt11 and Hxt2
Sut3Glucose0.8 and 3199% to Sut2Weierstall (1999)
(550)Fructose4977% to Sut1
Saccharomyces cerevisiaeHxt1Glucose10086.4% to Hxt3Kruckeberg (1996); Reifenberger (1997); Liu (2004)
Arsenic trioxideND
Hxt2Glucose10; 1.5 and 6081% to Hxt10Kruckeberg & Bisson (1990); Reifenberger (1997);
(541)8369% to Gal2,
[FD]Hxt6 and Hxt7Kruckeberg (1999); Maier (2002)
Hxt3Glucose6086.4% to Hxt1Kruckeberg (1996); Reifenberger (1997); Liu (2004)
(567)Fructose12575% to Hxt6, Hxt7 and Hxt4
Arsenic trioxideND
Hxt4Glucose983% to Hxt6 and Hxt7Johnston (1994); Reifenberger (1997); Liu (2004)
Arsenic trioxideND
Hxt5Glucose1072% to Hxt1Kruckeberg (1996); Diderich (2001); Liu (2004)
Arsenic trioxideND
Hxt6 and Hxt7Glucose1–299.7% to each otherKruckeberg (1996); Reifenberger (1997); Boles & Hollenberg (1997); Liu (2004)
(570)Fructose2.6–4.683% to Hxt4
Arsenic trioxideND
Hxt8GlucoseND70% to Hxt6,Kruckeberg (1996); Wieczorke (1999)
(569)FructoseNDHxt7 and Hxt4
[FD]MannoseND67% to Gal2
Hxt9 and Hxt11GlucoseND97% to each otherKruckeberg (1996); Wieczorke (1999); Liu (2004)
(567)FructoseND71% to Hxt7 and Hxt6
Arsenic trioxideND
Hxt10GlucoseND79% to Hxt2Kruckeberg (1996); Boles & Hollenberg (1997); Wieczorke (1999)
Hxt13 and Hxt17GlucoseND90–99% among the fourKruckeberg (1996); Boles & Hollenberg (1997); Wieczorke (1999)
(564)FructoseND<58% to other Hxts
Hxt15 and Hxt16GlucoseND
Hxt14GalactoseND≤38% to other HxtsKruckeberg (1996); Boles & Hollenberg (1997); Wieczorke (1999)
Gal2Glucose272% to Hxt6 and Hxt7Nishizawa (1995); Kruckeberg (1996); Reifenberger (1997); Boles & Hollenberg (1997)
[FD]Galactose3 and >20
Rgt2Glucose sensorNA60% to Snf3Kruckeberg (1996); Özcan & Johnston (1999)
Snf3Glucose sensorNA26–30% to HxtsKruckeberg (1996); Özcan & Johnston (1999)
Saccharomyces pastorianusFsy1Fructose0.1670% to KlFrt1Gonçalves (2000)
(570)49% to BcFrt1
[SYMP]SorboseND30% to E. coli galactose/H+ symporter GalP
Schizosaccharomyces pombeGht1Glucose578% to Ght5Lichtenberg-Fraté (1997); Heiland (2000); Wood (2002)
(557)41% to Hxt2, Hxt8, Hxt10 and Rag1
Ght2Glucose271% to Ght1 and Ght5Heiland (2000); Wood (2002)
(531)41% to Hxt2, Hxt9 and Hxt11
Ght3Gluconate384% to Ght4Hoever (1992); Heiland (2000)
(555)54–58% to other Ghts
Ght5Glucose0.676% to Ght1Heiland (2000)
(546)62–72% to other Ghts
Ght6Glucose869% to Ght1 and Ght5Heiland (2000)
(535)58–68% to other Ghts
Torulaspora delbrueckiiLgt1Glucose3678% to Hxt9 and Hxt11Alves-Araújo (2005)
(567)71% to Kht2
[FD]Fructose5165% to Rag1
Zygosaccharomyces bailiiFfz1Fructose8028% to S. pombe probable transporter Yao5Pina (2004)
  • * Number of amino acids.

  • Transport mechanism: FD, facilitated diffusion; SYMP, proton symport; NC, not characterized.

  • Amino acid identity (in %) with the most similar transporters, as determined by blastp (Altschul, 1997).

  • NA, not applicable.

  • ND, not determined.

View this table:

Biochemical properties of pentose transport systems characterized in yeasts

YeastGrowth substrate[Sugar] (g L−1)Substrate uptakeCharacteristicsReferences
C. intermediaXylose20XyloseFD51.510.0Gárdonyi (2003)
C. shehataeXylose or glucose20XyloseFD100–15320–25Lucas & van Uden (1986)
Glucose (starved)0XyloseSYMP1.01.4
C. succiphilaXylose20–50XyloseSYMP3.80.9Stambuk (2003)
C. utilisXylose20XyloseSYMP1.9NDKilian (1993)
Glucose (starved)20, 0XyloseFD10.5ND
D. hanseniiXylose20XyloseFDND10.7Nobre (1999)
K. marxianusXylose20–50XyloseFD11011.4Stambuk (2003)
P. heidiiXylose20XyloseNC40–50NDDoes & Bisson (1989)
P. stipitisXylose20XyloseSYMP2.263.4Kilian & van Uden (1988)
PYCC 4374SYMP0.0790.8
P. stipitisXylose20XyloseNC380NDDoes & Bisson (1989)
NRRL 7124NC0.9ND
P. stipitisXylose20XyloseFD19–212Weierstall (1999)
CBS 5774FD0.2–2.80.4–1
P. stipitisGlucose20XyloseSYMP3.692.6Kilian & van Uden (1988)
PYCC 4374SYMP0.380.4
P. stipitisGlucose20XyloseFD43–801–6Weierstall (1999)
CBS 5774FD3.20.5
S. cerevisiaeGlucose20XyloseFD92–175NDKotyk (1967)
201204.8Heredia (1968)
20FD1906.0Kötter & Ciriacy (1993)
C. arabinofermentansArabinose5ArabinoseFD125205Fonseca (2007b)
XyloseSYMP0.61.3–1.6Fonseca et al., unpublished data
P. guilliermondiiArabinose5ArabinoseFD123574Fonseca (2007b)
XyloseSYMP0.04–0.051.1–1.6Fonseca et al., unpublished data
  • * Transport mechanism: FD, facilitated diffusion; SYMP, proton symport; NC, not characterized.

  • Km (mM); Vmax (mmol h−1g−1 dw).

  • Similar to lower l-arabinose concentration (5 g L−1); ND not determined.

Hexose transport

The S. cerevisiae Hxt family

General features

In S. cerevisiae, the uptake of hexoses occurs only through facilitated diffusion (Lagunas, 1993) mediated by several transporters, the Hxt proteins, with different kinetic properties and modes of regulation (Reifenberger, 1997). The Hxt family encompasses 20 different hexose transporter-related proteins, Hxt1p–Hxt17p, Gal2p, Snf3p and Rgt2p, which were identified using mutant strains [gal2 and snf3 in S. cerevisiae and rag1 in K. lactis (Lewis & Bisson, 1991; Prior, 1993; Reifenberger, 1995; Boles & Hollenberg, 1997)] and, after complete sequencing of the yeast genome, by sequence similarity (Kruckeberg, 1996). Alignment of the amino acid sequences of the 20 proteins brings to light a marked sequence conservation with identity values ranging from 99.7% (Hxt6p and Hxt7p) to 25% (Snf3p and Hxt16p or Hxt17p). The most conserved regions are the 12 putative transmembrane domains and the most divergent in length and sequence are the cytosolic amino- and carboxyl-terminal regions (Kruckeberg, 1996). Hxt1p–Hxt17p and Gal2p have similar sizes whereas Snf3p and Rgt2p have longer carboxyl termini (Kruckeberg, 1996). These C-terminal tails seem to be related to the fact that Snf3p and Rgt2p act as glucose sensors and not as hexose transporters (Özcan, 1998).

Transmembrane domains 3, 5, 7, 8 and 11 were proposed to be involved in substrate binding or substrate translocation across the membrane (Kruckeberg, 1996). More recently, chimeric studies involving the Hxt1 and Hxt2 proteins led to the identification of residues crucial for high-affinity glucose transport (Kasahara, 2007).

Hexose transporters Hxt1p, Hxt3p and Gal2p can be converted into cation transporters, capable of mediating K+, Na+ and probably Ca2+ transport, by the introduction of single mutations in residues located within or immediately adjacent to transmembrane regions (Liang, 1998).

Comparison of HXT nucleotide sequences suggests that some of these genes resulted from gene duplication events and recombination between different chromosomes. This is further supported by the fact that several HXT genes are clustered in small regions of the same chromosome, often subtelomeric. In addition, some HXT genes are also identical in both the 5′ and the 3′ flanking regions (Bargues, 1996).

The expression of Hxt proteins is strongly regulated by the presence and concentration of glucose, among other factors, as described below. Functional interactions between different transporter proteins have been proposed (Reifenberger, 1997), but have not been demonstrated experimentally.

The most important glucose transporters in S. cerevisiae are the six proteins Hxt1–4, 6 and 7, because all of them are capable of supporting the growth of S. cerevisiae on glucose on their own. These proteins also accept fructose and mannose as substrates (Reifenberger, 1995; Wieczorke, 1999).

The construction of S. cerevisiae strains devoid of hexose transporters hxt17 (Reifenberger, 1995) and hxt1–17 (Wieczorke, 1999) allowed the individual characterization of the Hxt proteins (Reifenberger, 1995; Liang & Gaber, 1996; Wieczorke, 1999; Hamacher, 2002) and of several other fungal hexose transporters, for example, Saccharomyces pastorianus Fsy1 (Gonçalves, 2000), K. lactis Frt1 (Diezemann & Boles, 2003), Botrytis cinerea Frt1 (Doehlemann, 2005) and Candida intermedia Gxs1 and Gxf1 (Leandro, 2006).

Regulation and biochemical properties of individual Hxt family members

Hxt1p exhibits a low affinity for glucose and fructose (Reifenberger, 1997) and is accordingly induced about 300-fold by high levels of glucose, but not by galactose or raffinose (Özcan & Johnston, 1995). HXT1 was also found to be upregulated in cells exposed to osmotic stress (Hirayama, 1995; Greatrix & van Vuuren, 2006).

Hxt2p is a high-affinity glucose transporter (Kruckeberg & Bisson, 1990) whose expression is induced by low levels of glucose and repressed both by high glucose concentrations and in its absence (Özcan & Johnston, 1995). When cells carrying an Hxt2-green fluorescent protein fusion protein are exposed to high glucose concentrations, the protein is removed from the plasma membrane through internalization by endocytosis and degraded in the vacuole (Kruckeberg, 1999).

The HXT3 gene was initially isolated as a mutant allele able to restore K+ uptake in S. cerevisiae cells deleted for the genes encoding the K+ transporters (Trk1p and Trk2p), but the corresponding wild-type allele is unable to complement this phenotype (Ko, 1993). Overexpression of HXT3 suppresses the growth defect of the snf3 mutant on raffinose, indicating that it encodes a glucose transporter (Theodoris, 1994). Hxt3p is the most similar to Hxt1p, and is likewise a low-affinity hexose transporter that is expressed only in the presence of glucose (Özcan & Johnston, 1995). Recently, a different Hxt3p variant that seems to have improved affinity for fructose was characterized in a wine production strain (Guillaume, 2007).

The HXT4 gene was cloned by complementation of the rag1 mutation in a strain of K. lactis-defective in low-affinity glucose transport and has an intermediate affinity for glucose and a low-affinity for fructose (Reifenberger, 1995). It is induced by low levels of glucose, but is completely repressed at high glucose concentrations (Özcan & Johnston, 1995).

Hxt5p has a moderate affinity for glucose and a low affinity for fructose and for mannose (Diderich, 2001). It is apparently regulated by the growth rate of the cells rather than by the external glucose concentration (Verwaal, 2002). In addition, it is upregulated during growth on ethanol and on glycerol (Greatrix & van Vuuren, 2006) and upon nitrogen and carbon starvation (Buziol, 2002).

Hxt6p and Hxt7p are nearly identical, differing only in two amino acids (293:Val/Ile) and (556:Thr/Ala), located outside the 12 putative membrane-spanning domains. Both are high-affinity glucose transporters (Reifenberger, 1995). The two cognate genes are located in tandem on the right arm of chromosome IV, downstream of HXT3, and the high degree of similarity between them extends to 96 bp upstream of the start codon (Boles & Hollenberg, 1997). Both genes are repressed at high glucose concentrations, and, in derepressed cells, HXT7 is the most strongly expressed of the HXT genes (Liang & Gaber, 1996; Diderich, 2001). HXT6 is repressed by moderate and high concentrations of glucose, but is slightly induced by very low levels of glucose or by raffinose. HXT6 and HXT7 are also highly expressed on nonfermentable carbon sources such as glycerol and ethanol, and on maltose and galactose (Liang & Gaber, 1996). Both Hxt6p and Hxt7p are subject to catabolite inactivation by high glucose concentrations (Krampe, 1998).

The remaining genes, HXT8HXT17, appear to be unable to support growth on glucose of the hxt1–7 null strain (Reifenberger, 1995) seemingly because they are transcribed at very low levels (Özcan & Johnston, 1999). However, when individually overexpressed in an hxt1–17 gal2 null strain, all except HXT12 are able to restore growth on at least one of the hexoses, glucose, fructose, mannose or galactose, confirming that they are functional hexose transporters (Wieczorke, 1999).

Expression of Hxt8p is induced by low and repressed by high levels of glucose (Özcan & Johnston, 1999). The HXT9, HXT11 and HXT12 genes encode very similar proteins (Kruckeberg, 1996). The HXT12 sequence is interrupted by a 2-bp insertion (the two ORFs YIL170w and YIL171w can be combined upon deletion of two nucleotides to yield the predicted Hxt12p ORF sequence encoding 567 amino acids). This agrees with the inability of HXT12 to restore growth on hexoses of an hxt1–17 gal2 null strain (Wieczorke, 1999). HXT9 and HXT11 are under control of transcriptional regulators Pdr1 and Pdr3. These regulators also control the expression of ATP-binding cassette (ABC) transporters, which, when overexpressed, confer resistance to several unrelated drugs by acting as drug efflux pumps. Deletion of HXT9 and/or HXT11 confers resistance to the drugs cycloheximide (a protein synthesis inhibitor), sulphomethuron methyl (an acetolactate synthase inhibitor) and 4-nitroquinoline-N-oxide (a mutagen), and overexpression of HXT11 increases sensitivity to those drugs, suggesting that these permeases interact with ABC transporters or can themselves accept drugs as substrates (Nourani, 1997). Although Hxt10 is very similar to Hxt2p (79% identity), it seems to be unable to transport glucose in significant amounts (Reifenberger, 1997) and it is repressed by glucose (Özcan & Johnston, 1999).

Hxt13p, Hxt15p, Hxt16p and Hxt17p form another subgroup of very similar proteins (Kruckeberg, 1996) whereas Hxt14p is more distantly related to the other Hxt proteins. Deletion of each of these five genes or of all of them simultaneously did not cause any obvious growth defect phenotypes on media with different carbon sources at different temperatures (Boles & Hollenberg, 1997). HXT13, HXT14 and HXT15 are induced by low levels and repressed by high levels of glucose (Özcan & Johnston, 1999). HXT13 and, to a lesser extent, HXT15, are upregulated during growth on ethanol and glycerol, and HXT17 promoter activity is induced by a shift from pH 4.7 to 7.7 at low glucose concentrations (Greatrix & van Vuuren, 2006).

GAL2 was isolated by functional complementation of a gal2 mutant strain (Tschopp, 1986). Gal2p is the main galactose transporter in S. cerevisiae and is also a high-affinity glucose facilitator (Reifenberger, 1997). Galactose uptake kinetics by Gal2p in galactose-grown wild-type cells or hxt1–7 null cells show biphasic uptake with high- and low-affinity components, suggesting that galactose uptake by the Gal2 transporter may be modulated or influenced by other proteins (Ramos, 1989; Nishizawa, 1995). GAL2 expression is induced by galactose and repressed by high glucose concentrations (Özcan & Johnston, 1999). The substrate recognition domain of the Gal2p galactose/glucose transporter is located in transmembrane domain 10, as determined by functional analysis of chimeras between Gal2 and Hxt2 (Kasahara, 1996). Tyr446 and Trp455 were shown to be important amino acids for galactose recognition by Gal2p. Also, Tyr352, located in the extracellular boundary of putative transmembrane domain 7, and Phe504, located in the middle of transmembrane domain 12, are critical for Gal2p-mediated galactose transport (Kasahara & Kasahara, 2000).

The S. cerevisiae Snf3p and Rgt2p have only limited sequence similarity to the other members of the hexose transporter subfamily, have long cytoplasmic C-terminal segments, are transcribed at very low levels and are unable to transport glucose even when overexpressed (Özcan & Johnston, 1999). The sequences of their C-terminal tails are similar to each other only over a stretch of 25 amino acids that are thought to be important for sugar sensing. Rgt2 possesses only one of these sequences while Snf3 contains two of them (Özcan, 1998). Both proteins have well-established functions in the regulation of HXT gene expression, functioning as sensors of high (Rgt2) and low (Snf3) levels of glucose (Özcan, 1996).

Specific fructose transporters

The large majority of hexose transporters can accept both glucose and fructose (and often also mannose) as substrates, but some exceptions have been found. A specific, high-affinity fructose/H+ symporter was first detected in S. pastorianus and Saccharomyces bayanus in the late 1980s (Cason, 1986). The cognate FSY1 gene was isolated from S. pastorianus (Gonçalves, 2000). It also accepts sorbose as a substrate, but not glucose, and it seems to be phylogenetically unrelated to the Hxt proteins (Gonçalves, 2000). Expression of the FSY1 gene occurs only at low sugar concentrations, its transcription being repressed at high fructose and glucose concentrations (Rodrigues de Sousa, 2004). Fsy1p is very similar to another high-affinity specific fructose/H+ symporter (Frt1p) found in the grey mould fungus B. cinerea (Doehlemann, 2005). A close relative is also present in K. lactis (Diezemann & Boles, 2003) and in other yeasts such as Lanchancea (Saccharomyces) waaltii, Pichia stipitis, D. hansenii, C. albicans and Clavispora lusitanea, in addition to various ascomycetous filamentous fungi (P. Gonçalves, unpublished data). Hence, the FSY1 gene seems to have appeared early in the evolution of the Ascomycetes, but presumably after the separation of the Taphrinomycotina, the ancient lineage to which Schizosaccharomyces pombe belongs. The gene seems to have been subsequently selectively lost in some lineages of both yeasts and moulds.

Zygosaccharomyces bailii is a fructophilic yeast. So far, only one of the two transport systems thought to contribute to its fructophilic behaviour has been characterized at the molecular level (Ffz1p; Pina, 2004). It appears to be a low-affinity facilitator specific for fructose, but displays a low degree of similarity to previously characterized yeast sugar transporters, and seems to belong to the drug H+ antiporter-1 family (TCDB, family 2.A.1.2, Saier, 2006).

Hexose transporters in non-Saccharomyces yeasts

Kluyveromyces lactis

The transport of hexoses has been fully characterized at the molecular level in K. lactis. The RAG1 gene encodes a low-affinity glucose transporter (Wésolowski-Louvel, 1992), inducible by a high concentration of mono or oligosaccharides (fructose, mannose, sucrose, raffinose and galactose). Some K. lactis strains, however, do not possess the RAG1 gene, but harbour, instead, two tandemly arranged genes, KHT1 and KHT2, at the RAG1 locus. The KHT1 gene is identical to the RAG1 gene, except for the last two codons, which are identical to the 3′ end of KHT2. This suggests that RAG1 is a chimeric gene, resulting from recombination between KHT1 and KHT2, with loss of most of the KHT2 sequence and of the intergenic region between KHT1 and KHT2 (Weirich, 1997). Kht1p, like Rag1p, is a low-affinity glucose transporter, and the respective gene is similarly induced by glucose (Milkowski, 2001). Uptake of hexoses by Kht1p or Rag1p promotes their fermentation (Diezemann & Boles, 2003). Kht2p has an intermediate affinity for glucose. Hence, KHT2 transcription, which is high even in the absence of glucose, is further induced by low glucose concentrations, and repressed by high levels of glucose (Milkowski, 2001). The HGT1 gene encodes a constitutive high-affinity glucose transporter (Billard, 1996), which plays a prominent role in galactose transport (Baruffini, 2006). The FRT1 gene encodes a specific high-affinity fructose/H+ symporter similar to FSY1 of S. pastorianus. Like RAG1 and KHT1, FRT1 is induced by glucose, fructose and, to a lesser extent, by galactose (Diezemann & Boles, 2003). Rag4p is a K. lactis glucose sensor protein that possesses the long characteristic C-terminal tail of the S. cerevisiae sensors, also containing one ‘25-amino acid motif’. This protein seems to be the only K. lactis glucose sensor (Betina, 2001).

Kluyveromyces lactis is a predominantly aerobic yeast (Chen, 1992) that is unable to grow anaerobically on certain sugars such as galactose, raffinose and maltose – a phenomenon known as the Kluyver effect. Notably, the Kluyver effect in K. lactis seems to be closely related to the capacity of sugar transport, and at least in certain cases, it can be overcome by the expression of suitable transporters (Goffrini, 2002). In contrast to S. cerevisiae, K. lactis is a Crabtree-negative yeast capable of adjusting its glycolytic flux to the requirements of respiration by regulating glucose uptake [regulation of glucose uptake is apparently coupled to oxygen availability and not to the external glucose concentration as in Crabtree-positive yeasts (Bar, 2003)].

Schizosaccharomyces pombe

GTH1 was the first glucose transporter gene isolated from the Crabtree-positive fission yeast S. pombe. It encodes a high-affinity glucose, fructose/H+ symporter that is able to restore growth on glucose and fructose of an hxt1–7 null S. cerevisiae strain (Lichtenberg-Fraté, 1997; Heiland, 2000). Low-stringency hybridization using the GTH1 gene as a probe led to the identification of GTH2 and GTH5 and similarity searches revealed three additional genes: GTH3, GTH4 and GTH6. Heterologous expression of Gth proteins in an hxt1–7 null S. cerevisiae strain showed that only Ght1, Ght2, Ght5 and Ght6 can restore glucose uptake and growth on glucose or fructose. Glucose is the preferred substrate for Ght1p, Ght2p and Ght5p, while Ght6p displays a slightly higher affinity for fructose. Ght5p seems to be the predominant glucose transporter in wild-type S. pombe because the Km measured for Ght5p-mediated glucose uptake is similar to that measured in wild-type S. pombe (Heiland, 2000). Schizosaccharomyces pombe cells cultivated in medium containing high glucose concentrations express GHT2, GHT5 and GHT6, while GHT1, GHT3 and GHT4 are repressed under these conditions. At low glucose concentrations, GHT5 is further induced and GHT3 and GHT4 are derepressed. GHT3 and GHT4 are the most highly expressed transporters in cells cultivated on gluconate, which is in line with the proposed role of Ght3p as a specific gluconate/H+ symporter (Hoever, 1992; Flores, 2000). Curiously, despite the high amino acid identity between Ght3p and Ght4p, Ght4p does not seem to be a gluconate transporter (Heiland, 2000). Expression of GHT1 was detected only in S. pombe cells cultivated on gluconate and maltose, and it was suggested that it could function as a signalling membrane protein, although it complements growth on glucose and fructose in hexose transport-deficient S. cerevisiae strains (Heiland, 2000).

Candida albicans

CaHGT1 was the first gene encoding a glucose transporter isolated from the opportunistic pathogenic yeast C. albicans. Expression of CaHGT1 in an hxt1–7 null S. cerevisiae strain restores growth on glucose and established CaHgt1p as a high-affinity glucose transporter (Varma, 2000).

Screening of the C. albicans genome for sequences similar to CaHGT1 yielded 19 putative glucose transporter genes that were designated HGT2–HGT20. Hgt5p and Hgt19p have extended N-terminal domains whereas Hgt3p and Hgt4p have long C-terminal tails. HGT9, HGT10, HGT12 and HGT17 are strongly expressed at low glucose concentrations and HGT10 and HGT12 are weakly expressed at high-glucose concentrations (Fan, 2002), suggesting that Hgt10p and Hgt12p are high-affinity hexose transporters (Brown, 2006). Hgt12p accepts glucose, fructose and mannose as substrates (Luo, 2007). Expression of HGT20 is poorly detectable at high glucose concentrations. The other 15 HGT genes, HGT1HGT8, HGT11, HGT13–HGT16 and HGT18–HGT19, are expressed at both high and low glucose concentrations (Fan, 2002). Expression of HGT7 is further induced by high concentrations of glucose, fructose and galactose. HGT18 and HGT20 show a high degree of sequence similarity to the Glut human transporters, suggested to be the possible result of functional convergence (Fan, 2002). Hgt4p is a glucose sensor similar to Snf3p and Rgt2p in S. cerevisiae (Brown, 2006).

Pichia stipitis

SUT1 is a glucose transporter gene isolated from P. stipitis CBS 5747 by functional complementation in S. cerevisiae (Weierstall, 1999). The SUT2 and SUT3 genes were subsequently identified by their sequence similarity to SUT1 (Weierstall, 1999). Both genes complement growth on glucose of an S. cerevisiae hxt1–17 gal2 null strain, but not as efficiently as SUT1 (Weierstall, 1999). Sut1p is a high-capacity, low-affinity glucose transporter whose transcription is strongly induced by glucose independent of the oxygen supply, whereas Sut2p and Sut3p are low-capacity transporters, which are regulated by the oxygen availability rather than by the carbon source. The Sut proteins are also able to transport other monosaccharides including d-xylose (Table 2), but with a lower affinity (Weierstall, 1999). Recently, complete sequencing of the P. stipitis genome revealed several additional sugar transporter genes, including a fourth putative SUT gene (Jeffries, 2007).

Other yeasts

The Torulaspora delbrueckii LGT1 gene encodes a low-affinity glucose and fructose facilitator similar to the K. lactis low-affinity transporters Kht1/Rag1, both in its sequence and in its mode of regulation (Alves-Araújo, 2005).

Other transporters have been studied in some detail from the biochemical point of view, but the respective genes were not isolated so far. In the Crabtree-negative yeast Candida utilis, a glucose/H+ symporter was characterized using membrane vesicles (van den Broek, 1997), while the methylotrophic yeast Hansenula polymorpha (formerly Pichia angusta) was found to harbour a high-affinity glucose/H+ symporter (Km=0.05–0.06 mM) and a lower affinity transporter (Km=1.75 mM) (Karp & Alamae, 1998). The H. polymorpha Gcr1 protein is similar to the Amanita muscaria monosaccharide transporter AmMst1 and to the putative glucose sensors Snf3p, Rgt2p and Rag4p, but whether it functions as a sensor or as a transporter is still unclear (Stasyk, 2004). Searches in the H. polymorpha genome for sequences similar to Gcr1p led to the identification of Hxs1p, which functions as a sensor, although it does not seem to have the ‘glucose-sensing’ domains of Snf3p, Rgt2p and Rag4p in its long C-terminal sequence. In addition, Hxt1 (also identified by sequence similarity) functions as a low-affinity glucose and fructose transporter (Stasyk, 2008).

Although facilitated diffusion seems to be the prevailing system for hexose transport in yeasts, a wide comparative study involving 205 yeast species led to the conclusion that glucose/H+ symporters are present in at least one-third of the species and that it is usually under control of catabolite repression (Loureiro-Dias, 1988). The reciprocal regulation of the facilitated diffusion of glucose vs. glucose/H+ symport was studied in detail in Candida wickerhamii (Spencer-Martins & van Uden, 1985) and Pichia ohmeri (Verma, 1987). The later study presents evidence for the occurrence of protein synthesis-independent recovery of the facilitated diffusion system in repressed cells, an observation that may be worth revisiting in light of our current knowledge of the molecular mechanisms regulating the turnover and cellular localization of sugar transporters in yeasts.

Pentose transport

Biochemical properties and regulation of d-xylose and l-arabinose transporters in yeasts

Xylose and arabinose are two abundant sugars in lignocellulose hydrolysates, a potential substrate for the production of fuel ethanol that can be obtained from agricultural and forestry residues. As a result, the poorly explored field of pentose transport in fungi is considered particularly important because of its potential impact on the improvement of S. cerevisiae strains engineered to ferment these substrates (Hahn-Hägerdal, 2007). Table 2 summarizes data concerning the characterization of xylose and arabinose uptake and regulation in different ascomycetous yeasts. Efficient xylose-utilizing yeasts usually have a high-affinity and a low-affinity xylose-uptake system. In general, the low-affinity system transports xylose (and in most cases also glucose) by a facilitated diffusion mechanism whereas the high-affinity system is a xylose/H+ symporter (Gárdonyi, 2003). With very few exceptions, symporters are expressed at low-sugar concentrations while the lower affinity facilitators are expressed when sugar is abundant.

Although arabinose catabolism closely resembles that of xylose, the two sugars often make use of distinct uptake systems (Lucas & van Uden, 1986; Spencer-Martins, 1994; Fonseca, 2007b). This is the case in Pichia guilliermondii and Candida arabinofermentans, where recently two low-affinity and high-capacity arabinose transport systems were biochemically characterized (Fonseca, 2007b). Both transporters seem to be highly specific for arabinose and the fact that they do not seem to transport glucose and xylose at all is of particular interest. Besides this low-affinity-facilitated diffusion, arabinose-utilizing yeasts also have high affinity but less-specific arabinose/H+ symporters (Fonseca, 2007b). The regulation of these transporters resembles that of glucose/xylose transporters. Both pentoses, d-xylose and l-arabinose, even though not metabolized by wild strains, are taken up by S. cerevisiae through the Hxt transporters. From the 18 known hexose transporters, the high- or intermediate-affinity transporters Hxt4p, Hxt5p, Hxt7p and Gal2p are the most relevant for d-xylose uptake (Hamacher, 2002; Lee, 2002; Sedlak & Ho, 2004), although d-xylose is a much weaker substrate compared with glucose –c. 100-fold lower affinity, Table 2 (Kotyk, 1967; Kötter & Ciriacy, 1993; Lee, 2002; Saloheimo, 2007). Both d-xylose and l-arabinose are poor substrates of the galactose transporter Gal2p (Cirillo, 1968).

Pentose transporter genes isolated from yeasts

Although substantial research effort has been made in attempting to isolate fungal xylose transporters, genetic approaches based on functional complementation in S. cerevisiae have so far yielded only the lower-affinity systems (Weierstall, 1999; Leandro, 2006). For example, an attempt to isolate xylose transporters from P. stipitis CBS 5774 by functional complementation yielded the previously mentioned low-affinity glucose/xylose transporters Sut1-3 (Table 2), but it did not result in the isolation of the gene encoding the high-affinity xylose symporter (Weierstall, 1999). A similar result was obtained while cloning the xylose transporter genes from C. intermedia, another efficient xylose-utilizing yeast. Two glucose/xylose transporter genes from this yeast were recently cloned and characterized: GXF1, encoding a glucose/xylose facilitator (Km∼50 mM), and GXS1, encoding a glucose/xylose symporter (Km∼0.4 mM). Although GXF1 was isolated by functional complementation of an hxt-null S. cerevisiae strain, isolation of the GXS1 cDNA required partial purification and microsequencing of the transporter (Leandro, 2006). Gxs1p was the first fungal glucose/xylose-H+ symporter to be characterized at the molecular level. While Gxf1p seems to be fully functional in S. cerevisiae, the symporter Gxs1p exhibits very low glucose/xylose transport activity, which could not be ascribed to insufficient production of the protein or incorrect subcellular localization. In addition, coexpression of glucose/xylose facilitators with Gxs1p under certain conditions strongly reduced GXS1 mRNA levels, an effect that seems to occur at the level of mRNA stability (Leandro, 2008). The expression of Gxf1p in recombinant S. cerevisiae was recently shown to improve xylose fermentation (Runquist, 2009).

The partial sequence of a D. hansenii putative xylose permease Xylhp (EMBL AAR06925, A. Nobre & C. Lucas) is available in the GenBank, but this protein has not been functionally characterized yet.

Although functional expression of arabinose transporters in S. cerevisiae has not been reported so far, some entries can be found in public databases concerning arabinose transport in fungi, mostly related to patent applications: Ambrosiozyma monospora LAT1 and LAT2 (EMBL AY923868 and AY923869, respectively, R. Verho et al.), C. arabinofermentans ART1 (Fonseca, 2007a), K. marxianus KmLAT1 and P. guilliermondii PgLAT2 (Knoshaug, 2007) and P. stipitis araT (Boles & Keller, 2008).


In the course of the last decade, our general knowledge of the molecular aspects of hexose and pentose transport in yeasts has increased considerably, but a number of important questions remain to be elucidated. For example, hexose transporter gene families vary dramatically in size among ascomycetous yeasts, but the physiological basis for the apparent redundancy of sugar transporters with similar properties found in some yeasts, such as S. cerevisiae, has not been fully elucidated yet. Another open issue, concerning not only yeast proteins but also all MFS transporters, is the identification of functional regions. Hopefully, the possibility to study an increasing number of transporter genes in a phylogenetic framework will help to shed some light on the most important structural open questions, such as the signature features of a symporter vs. a facilitator or the specificity for a substrate. So far, the lessons learned from the best-studied MFS protein, the lactose permease from Escherichia coli, suggest that it will be difficult to find general patterns in structure–function relationships, let alone to generate guide lines for the modification of the biochemical properties of transporters (Zhou, 2008). Because this is clearly a desiderate in the context of various metabolic engineering projects involving sugar metabolism in yeasts, it may be worthy to continue uncovering the existing natural diversity to find proteins with novel characteristics. In this respect, it should be emphasized that several lineages of yeast sugar transporters consist entirely of uncharacterized proteins whose function remains to be established (Palma, 2007).


We dedicate this review to the memory of Prof. Isabel Spencer-Martins, as a tribute to her long-lasting commitment to research in the field of sugar transport in yeasts.


  • Editor: André Goffeau


View Abstract