The view of the role of trehalose in yeast has changed in the last few years. For a long time considered a reserve carbohydrate, it gained new importance when its function in the acquisition of thermotolerance was demonstrated. More recently the cellular processes in which the trehalose biosynthetic pathway has been implicated range from the control of glycolysis to sporulation and infectivity by certain fungal pathogens. There is now enough experimental evidence to conclude that trehalose 6-phosphate, an intermediate of trehalose biosynthesis, is an important metabolic regulator in such different organisms as yeasts or plants. Its inhibition of hexokinase plays a key role in the control of the glycolytic flux in Saccharomyces cerevisiae but other, likely important, sites of action are still unknown. We present examples of the phenotypes produced by mutations in the two steps of the trehalose biosynthetic pathway in different yeasts and fungi, and whenever possible examine the molecular explanations advanced to interpret them.
Trehalose is quite an oddity to most biochemists. This disaccharide, formed by two glucose units linked in an α,α-1,1-glycosidic linkage (Fig. 1), apparently absent from vertebrates, present in yeasts, fungi and plants but not located in a central metabolic pathway, is hardly considered in non-specialized biochemistry books. However, trehalose has a great importance in nature as well as in industry. In nature trehalose plays a variety of roles, from being an energy source to a stress protectant. In industry its importance as a determinant of baker's yeast stability has been known for a long time and its use as a stabilizer for a wide range of molecules and structures is gaining importance. Trehalose even reached the newspapers when it was shown that its accumulation in transgenic rice plants resulted in increased tolerance to different abiotic stresses . A brief historical appreciation of trehalose discovery and identification is provided in Singer and Lindquist  and an overview of the distribution and different roles of the disaccharide may be found in Elbein et al. .
Our purpose in this minireview is to bring to consideration the importance of the proper functioning of the biosynthetic pathway of trehalose for the life of yeasts and some fungi and to highlight the relevance of trehalose 6-phosphate (T6P), an intermediate of trehalose biosynthesis (Fig. 1), as a metabolic regulator in different organisms. We will not deal with the role of trehalose in the acquisition of thermotolerance, a process in which the disaccharide itself is implicated. To achieve our purpose we will consider the phenotypes produced by mutations in the different steps of the trehalose biosynthetic pathway and examine the molecular explanations advanced to interpret them.
2The biosynthesis of trehalose in yeasts
Different pathways of trehalose biosynthesis have been identified in various organisms (for a review see ). In yeasts this disaccharide is synthesized through a pathway elucidated by Cabib and Leloir almost 50 years ago . The synthesis starts with the transfer of a glucosyl residue from uridine-diphospho-glucose to glucose 6-P to provide T6P that is subsequently dephosphorylated to yield trehalose (Fig. 1). In Saccharomyces cerevisiae, the enzymes that catalyze the reactions of trehalose biosynthesis, T6P synthase (Tps1) and T6P phosphatase (Tps2), are part of a complex in which two other proteins, Tsl1 and Tps3, without catalytic activity, participate [5,6]. Curiously all these proteins share a high degree of sequence similarity. In vitro assays have shown that the activity of Tps1 from S. cerevisiae within the complex is increased by fructose 6-P and inhibited by phosphate. At 50°C these effects are much reduced but the enzyme is intrinsically more active .
Tsl1 and Tps3 stabilize the complex; individual deletion of the TSL1 or TPS3 genes does not markedly affect the trehalose content of the yeast but a double deletion tsl1 tps3 decreased it by more than 50%. Deletion of these subunits also caused a change in the effect of phosphate on T6P synthase activity: it became activated by phosphate instead of being inhibited . Tps1 may function independently of the other subunits of the complex [8,9] and it has been hypothesized that in this form it may be implicated in the rapid increase of T6P observed upon glucose addition to yeast cells growing on non-fermentable carbon sources .
In some culture conditions S. cerevisiae strains mutated in the TPS1 gene were found to accumulate trehalose . To explain this finding a biosynthetic pathway for trehalose using adenine-diphospho-glucose instead of uridine-diphospho-glucose was proposed . However, a gene encoding such activity has not yet been identified and the trehalose accumulation seems to be due to its uptake from complex media by a broad-specificity transporter encoded by the AGT1 gene .
3The role of TPS1 in S. cerevisiae
3.1The tps1 phenotype in S. cerevisiae
Mutants affected in the gene encoding T6P synthase have been around at least since the early 1970s but the data available at the time did not allow to ascribe their defect(s) to this function. Four mutants have been instrumental to establish the importance of the TPS1 gene in S. cerevisiae: fdp1, cif1, byp1 and glc6. We will consider each of these mutations in due course according to the role played in the development of our knowledge about the relationship of trehalose biosynthesis and other pathways.
When mutants affected in glucose metabolism started to be isolated and characterized, most of them had a phenotype that could be reasonably explained with the available biochemical knowledge (see the first review on the topic by Fraenkel ). However, some of them, like the fdp1 mutant , exhibited a phenotype difficult to understand. This mutant was unable to grow on glucose although all the glycolytic enzymes were active when assayed in vitro. The mutant grew on glycerol but when glucose was added to the culture hexose phosphates accumulated and a precipitous drop in the internal ATP concentration occurred. Catabolite inactivation of the gluconeogenic enzyme fructose-1,6-bisphosphatase (then called fructose-1,6-diphosphatase, hence the name of the mutation) did not occur in the mutant and this led to the proposal of the existence of a futile cycle phosphofructokinase–fructose-1,6-bisphosphatase during glucose utilization that could explain the drop in ATP and thereby the lack of growth on glucose . This hypothesis appeared to be reinforced by the lack of catabolite inactivation of another gluconeogenic enzyme, PEPcarboxykinase, that could cause a second futile cycle PEPcarboxykinase–pyruvate kinase . However, this idea was not supported by the results of Bañuelos and Fraenkel  who did not find experimental evidence for a futile cycle in the fdp1 mutant. In the late 1970s a mutant with similar characteristics was isolated and named cif1 for catabolite inactivation of fructose-1,6-diphosphatase . Later on it was shown that fdp1 and cif1 were allelic forms of the same gene . Another interpretation of the phenotype of the fdp1 mutant, suggested by the increase of hexose monophosphates and fructose-1,6-bisphosphate and the precipitous drop of ATP triggered by glucose [12,13], , could be the existence of an imbalance between the rates of the initial steps of glycolysis consuming ATP and those regenerating it after the glyceraldehyde 3-P dehydrogenase step. Such a metabolic imbalance between energy-consuming and energy-recovering steps has been documented in liver after a load of fructose as a consequence of the function of a highly active fructokinase and a low-capacity aldolase .
The cloning of CIF1 did not immediately provide an explanation of the phenotype of the mutant, but shortly afterwards it was found that the sequence of the gene encoding T6P synthase was identical to that of CIF1. This information did not by itself explain the phenotype of the cif1 or fdp1 mutants but was crucial to later interpretations. In retrospect it is interesting to note that the lack of T6P synthase activity in the fdp1 mutant was noticed earlier but it was not considered to be the cause of the phenotype observed .
Although different names were used when the various mutations in the gene encoding T6P synthase were isolated, from now on we will refer to the gene as TPS1.
3.2Extragenic suppressors of the tps1 phenotype
Extragenic suppressors have provided valuable information to determine the mechanism(s) of action of the tps1 mutation. The suppressors isolated fit into two main categories: those that decrease the glycolytic flux and those that appear to relieve the functional block at the level of glyceraldehyde 3-P dehydrogenase suggested by the accumulation of hexose phosphates (Fig. 2). The suppressors that decrease glycolytic flux in turn act either at the glucose transport step or on hexokinase. Overexpression of MIG1, the action of the DGT1-1 allele of MTH1[24,25] or mutations in CAT3 (SNF4)  cause a decrease in glucose transport. Deletion of the gene HXK2 encoding hexokinase II or a partial decrease in its activity also restored the capacity to grow on glucose to S. cerevisiae tps1 mutants and reduced the accumulation of glycolytic intermediates upon glucose addition [27,28].
The other group of suppressors cause an increased excretion of glycerol in the mutants that carry them. They comprise the overexpression of FPS1 encoding a glycerol channel , or of GPD1 encoding glycerol 3-P dehydrogenase , and mutations in QCR9 that encodes one subunit of the cytochrome bc1 complex .
Do all these suppressors tell a coherent story? The answer is yes, even if we may lack some clue to fully understand their action. If we accept the previously mentioned idea of the imbalance between the rates of the ATP-consuming and -regenerating glycolytic steps, the suppressors indicate that a new metabolic balance can be achieved either by a decrease in the rate of the initial glycolytic steps or by an increase in the activity of an ancillary pathway that may generate substrates (NAD+ and free phosphate) to activate the reactions of the second part of glycolysis. The question becomes then, how does the wild-type yeast achieve the balance?
3.3A possible molecular explanation for the phenotype of tps1 mutants: T6P, a new regulator of glycolysis that inhibits hexokinase
A remarkable feature of the tps1 mutants is that they can grow on galactose although they do not grow on glucose. Since the metabolic steps of catabolism are identical for both sugars after glucose 6-P, the cause of the lack of growth on glucose is likely to be either at the transport level, at the hexokinase step, or at some system implicated in glucose sensing. The problem was to find a link between the primary genetic defect of the mutant (lack of T6P synthase) and those steps. A plausible idea could be that they were regulated either by trehalose or by T6P. Following this reasoning Blázquez et al.  found that T6P competitively inhibits the hexokinases from S. cerevisiae. Hexokinase II, the enzyme present during growth on glucose was the most sensitive with a Ki of 40 μM. These authors proposed that the loss of this inhibition in the tps1 mutant was the reason for the lack of growth on glucose since the non-inhibited hexokinase would cause an increased ATP expenditure, exceeding the capacity to regenerate it further down in glycolysis. They suggested that T6P “could play in yeast a role as feed-back inhibitor similar to that played by glucose 6-P in higher organisms”. In fact up to now T6P is the only reported physiological inhibitor of yeast hexokinase. A kinetic model has been developed showing not only the importance of the hexokinase inhibition by T6P but also its need for a proper functioning of glycolysis in yeast due to the turbo design of this pathway . (Turbo design refers to those pathways in which the initial steps are primed by one of its end products, ATP in the case of yeast glycolysis).
The internal concentration of T6P in S. cerevisiae during growth on glucose is about 0.18 mM ; addition of glucose to cells grown on non-fermentable carbon sources transiently increases this value up to 1.5 mM [33,34].
The relevance of the inhibition of hexokinase by T6P for the in vivo control of glycolysis was questioned based on two arguments . One of them was that T6P would be channeled within the protein complex synthesizing trehalose and therefore was unlikely to reach the cytoplasm in a sufficient concentration to inhibit hexokinase. The other was that the inhibition of hexokinase by T6P was competitive with glucose and that the intracellular glucose concentration was an order of magnitude higher than the T6P concentration measured. Two alternative proposals were therefore advanced . The first was that the trehalose biosynthetic pathway could serve the additional function of recovering phosphate that is required for the functioning of glycolysis at the level of glyceraldehyde 3-P dehydrogenase. Thus a mutation in tps1 that abolishes the production of T6P would eliminate the possibility of phosphate recovery in the T6P phosphatase step. This idea was based on results that showed that in tps1 mutants the internal phosphate concentration dropped to lower values than in wild-type yeasts after glucose addition . If this hypothesis is correct it would be expected that tps2 mutants in which the accumulated T6P acts as a phosphate trap would be affected in their growth on glucose. However, this is not the case [36–38]. Also it was found that a strain carrying the byp1-3 mutation, a TPS1 allele with low T6P synthase activity, grew better on glucose after deletion of the TPS2 gene while the phosphate-recovering hypothesis would have predicted the contrary . (The byp1 mutants were isolated by Zimmermann and coworkers  and were named byp (from bypass) because they were supposed to affect a hypothetical pathway that bypassed the phosphofructokinase reaction in glucose metabolism.) In addition François and Parrou  have calculated that the rate of glycolytic flux greatly exceeds the rate of trehalose formation in yeast, making it difficult for this pathway to supply the necessary phosphate for the functioning of glycolysis. Therefore it seems that the phosphate-recovering role – if any – of the trehalose biosynthetic pathway is not chiefly implicated in the tps1 phenotype.
The second proposal suggested that the Tps1 protein itself played a role in the inhibition of the glycolytic flux as part of a hypothetical general glucose-sensing complex formed by the association of Tps1, hexokinase and the glucose carrier. During a certain time the TPS1 gene was named in several articles GGS1 (glucose general sensor) to stress this hypothetical role. In such a complex T6P would immediately reach hexokinase and act with greater efficacy. Attractive as this hypothesis may be, it has not yet been substantiated by experimental work. As far as we are aware no description of interactions between Tps1 and hexokinase or sugar transport proteins has been reported. Therefore up to now the only experimentally demonstrated connection found between trehalose biosynthesis and glycolysis is the inhibition of hexokinase caused by T6P . But does the inhibition of hexokinase by T6P completely explain the lack of growth on glucose caused by the tps1 mutation? Are we not missing something important?
The hypothesis of the imbalance between the rates of the initial and later glycolytic steps explains the accumulation of large amounts of hexose monophosphates and fructose-1,6-bisphosphate and the depletion of ATP in tps1 mutants after glucose addition. However, there is a feature of these mutants that is not easily accounted for by that idea, namely their lack of glucose fermentation in spite of the accumulation of hexose phosphates. An idea to explain this behavior could be that T6P, Tps1 itself or both could activate the second part of glycolysis. Experimental evidence consistent with this hypothesis comes from results on the fermentation of fructose by a phosphoglucose isomerase mutant. This mutant does not ferment glucose but in principle ought to ferment fructose that enters glycolysis after the metabolic block; however, fructose is not fermented and glycolytic metabolites accumulate in a pattern similar to that of a tps1 mutant, suggesting a block at the step catalyzed by glyceraldehyde 3-P dehydrogenase. If small amounts of glucose are added simultaneously with fructose, normal fermentative metabolism of fructose takes place . This restoration of fermentation may be due to its activation by some metabolite derived from glucose, possibly T6P. However, up to now no experimental in vitro evidence supports this hypothesis .
Another possible cause of the arrest of glucose fermentation in a tps1 mutant might be the lack of ADP, or phosphate, whose levels also decrease after glucose addition [13,20]. The stimulating effects of antimycin A or of the qcr9 mutation on glucose fermentation  are compatible with this thought.
3.4The relationship between T6P synthase, sporulation, and glycogen biosynthesis
Another characteristic of tps1 mutants is their decreased sporulation capacity. De Silva-Udawatta and Cannon  have provided a possible explanation for this trait. They found that under sporulation conditions the levels of IME1 mRNA were reduced in diploid tps1 strains with respect to those of a wild-type strain. IME1 encodes a transcriptional activator of meiotic gene expression that requires the presence of the protein kinase Mck1 for its optimal expression . In tps1 diploids the abundance of MCK1 mRNA is decreased  and this could explain the lower sporulation frequency as a consequence of a decrease in IME1 expression.
The same authors also advanced an explanation for the phenotype of the glc6-1 mutant (now tps1-H223Y). This mutant was isolated as one that did not accumulate glycogen. It turned out to be allelic with tps1 but it grew on glucose plates . It has now been found that, in contrast with the T6P synthase activity of a wild-type, the activity of the mutated protein is not inhibited by phosphate and is higher when assayed at 50°C. Moreover, the mutated allele behaved as dominant with respect to wild-type . The increase in T6P synthase activity in the mutant could explain the diminished glycogen levels through two effects: (a) reduced availability of uridine-diphospho-glucose and glucose 6-P for glycogen synthesis, (b) lowered glycogen synthase activity due to a decrease in the concentration of its allosteric activator (glucose 6-P).
4Phenotype of tps1 mutants in other yeasts and fungi
The striking phenotypic effects observed in the tps1 mutant of S. cerevisiae led to the question about the generality of these effects in other species. As will be seen in the following lines, although mutations in the corresponding gene show in general important effects, there is not a common picture for the different organisms.
In Kluyveromyces lactis ggs1 mutants (GGS1 is the name in this yeast for the gene encoding T6P synthase) showed a phenotype similar to that of tps1 mutants from S. cerevisiae. Suppressors were isolated in which glucose uptake was decreased, pointing also to an uncontrolled glycolytic flux as the origin of the phenotype observed .
Schizosaccharomyces pombe presents an interesting situation since its two hexokinases are not significantly affected by T6P . It was found that disruption of the corresponding tps1+ gene did not have a noticeable effect on growth on glucose , a result consistent with the idea that the inhibition of hexokinase by T6P was important for the tps1 phenotype in S. cerevisiae. Spores carrying a tps1+ disruption did not germinate  but it was not clarified if this lack of germination was due to lack of trehalose or T6P.
In the respiratory yeasts Hansenula polymorpha and Yarrowia lipolytica disruption of the respective genes encoding Tps1 did not produce a phenotype of lack of growth on glucose  (our unpublished results). This result is particularly surprising in the case of Y. lipolytica, which presents the most sensitive hexokinase to inhibition by T6P yet found . These results indicate that in yeasts with a respiratory metabolism the effects of the TPS1 disruption are less severe than in a yeast like S. cerevisiae with a higher demand on the glycolytic flux.
Mutants affected in the first step of trehalose biosynthesis have been found in Aspergillus niger and in A. nidulans. While in A. niger two genes encoding T6P synthase have been found, tpsA and tpsB, there is only one in A. nidulans, tpsA. In both cases disruption of tpsA did not markedly influence growth on glucose, although in A. nidulans increased levels of sugar phosphates were detected in the mutant strain . There is no information on the phenotype of a double tpsA tpsB disruption in A. niger. Interestingly, Aspergillus hexokinases are also sensitive to inhibition by T6P (Ki for T6P 0.01 mM) .
The trehalose biosynthetic pathway is an attractive potential specific target for the development of antifungal drugs due to its absence in vertebrates; therefore the effects of the tps1 mutation (and that of tps2, see below) have been studied in some pathogenic yeasts. In Candida albicans a tps1/tps1 mutant grew at 30°C as the wild-type; however, at 42°C it ceased growth on glucose while it was able to grow on galactose or glycerol. Incubation with glucose at the restrictive temperature produced a rapid ATP depletion and accumulation of hexose phosphates similar to those found in the corresponding S. cerevisiae or K. lactis mutants. Hypha formation and infectivity were impaired in the mutant .
Although there are no biochemical data available on Cryptococcus neoformans tps1 mutants, it has been reported that they are cleared from the host within 3–7 days after inoculation, a result consistent with an implication of the trehalose biosynthetic pathway in infectivity . More evidence for the implication of the trehalose biosynthetic pathway in infectivity comes from results with Magnaporthe grisea, a fungus responsible for the rice blast disease. M. grisea tps1 mutants sporulated poorly and were unable to invade the host plant due to lack of functional apressoria, a structure needed for invasion . The molecular bases of these phenotypes are not clear.
5The tps2 phenotype in different yeasts and fungi
Much more work has been devoted to the tps1 mutations than to those affecting TPS2. However, the phenotype produced by mutations in TPS2 is conspicuous and likely points to unknown regulatory phenomena that would be worth studying. Common to all yeasts and fungi in which a mutation abolishes T6P phosphatase activity is a thermosensitive growth phenotype.
Piper and Lockheart  isolated a S. cerevisiae mutant that did not grow on any carbon source at temperatures above 34°C during a screen for mutants defective in RNA synthesis. The mutant accumulated a phosphoric ester that turned out to be T6P and the authors concluded that they had isolated a mutant affected in T6P phosphatase. Later De Virgilio et al.  cloned and disrupted the TPS2 gene in S. cerevisiae and found a phenotype similar to the one previously reported. An explanation for the growth thermosensitivity of the tps2 mutant could be the lack of trehalose, a well-known protector against heat stress. However, the fact that tps1 mutants that also lack trehalose are not thermosensitive argues against this possibility. The results of Elliot et al.  point to the accumulation of T6P as the main cause of the phenotype. These authors found that a tps1 mutant and a tps1 tps2 double mutant had heat shock resistance similar to that of the wild-type. However, a tps2 single mutant was extremely sensitive to that treatment and accumulated a large amount of T6P (14.7 nmol mg−1 wet weight vs. undetectable in the wild-type). Moreover, a suppressor that did not restore trehalose synthesis but decreased the accumulation of T6P to 0.3 nmol mg−1 wet weight suppressed the thermosensitive phenotype. It is worth mentioning that they also found a gene, PMU1 (YKL128c), that when expressed in a multicopy vector decreased the level of T6P and suppressed the tps2 phenotype. The physiological role of this gene is unknown.
In Sch. pombe disruption of tpp1+ (designation in this yeast for the T6P phosphatase-encoding gene) also resulted in accumulation of T6P after heat shock (10.5 vs. 1.1 mg g−1 wet weight in the wild-type) and prevented cell growth above 37°C .
The same phenotype of thermosensitive growth and accumulation of T6P was also found in a homozygous C. albicans tps2/tps2 mutant [54,55]. The mutant flocculates in stationary phase and liberates proteins to the medium; both phenomena can be prevented by the addition of osmoprotectants to the culture medium . It has been suggested that T6P could regulate some step of cell wall organization or even participate in a cell integrity pathway, but direct proof of this implication is lacking . Infectivity of C. albicans tps2/tps2 mutants was greatly diminished, again showing the importance of the pathway for this process [54,55].
A mutation in the gene orlA of A. nidulans that encodes a T6P phosphatase also renders the organism thermosensitive and prevents viable germination of conidia due to their bursting when germinated at 42°C . The mutation causes a deficiency in chitin and is remediated by osmotic protectants. The observation that incubation at 32°C of extracts of the orlA disruptant caused a decrease in the activity of glutamine:fructose 6-P amidotransferase, the first enzyme of the pathway of aminosugar biosynthesis, led to the hypothesis that the accumulation of T6P in the mutant reduced the thermal stability of this enzyme, resulting in a defective cell wall . Concerning the sequence of the protein it has been shown that the protein is larger than previously assumed and contains an amino-terminal domain that is homologous to that of the A. nidulans T6P synthase .
6T6P is also an important molecule in plants
To widen the perspective on the roles of T6P we will briefly comment on the role ascribed to this molecule in flowering plants. The discovery of the regulatory role of T6P in yeast  and the identification of trehalose biosynthetic genes in Arabidopsis thaliana[57,58] have stimulated research on the role of the disaccharide in plants. TPS1 and TPS2 homologues have been detected in taxonomically unrelated plant species and at least 11 of these are found in the A. thaliana genome . AtTPS1 is essential for A. thaliana; its disruption leads to arrest of embryo development . It has also been shown that in this plant T6P controls carbohydrate utilization and growth ; plants with increased T6P levels had small green leaves while those with decreased T6P concentration produced large pale leaves. In addition, A. thaliana plants expressing a heterologous gene encoding T6P synthase grew better than wild-type ones .
The pleiotropic effects caused by mutations in TPS1 or TPS2, the two genes encoding the enzymes of the trehalose biosynthetic pathway in different yeasts and fungi described in the previous sections, show the importance of this pathway in the life of these organisms. Evidence from different studies has demonstrated the role of T6P that has been revealed as an important regulatory molecule in different organisms. Its effect on hexokinase places the glucose phosphorylation step among the important regulated steps of yeast glycolysis. It is curious that the role of T6P in the regulation of such a central metabolic pathway as glycolysis in yeast was not discovered before , however it should be remembered that fructose 2,6-bisphosphate, a key metabolite in the regulation of the same pathway in different organisms, was discovered only about 10 years before the role of T6P was proposed . Two late discoveries if one considers the amount of work devoted to the regulation of glycolysis over the last 50 years.
It should be stressed that some questions remain unexplained about the molecular relationships of the trehalose biosynthetic pathway and other pathways. One of them is the tantalizing question of the possible activation of the second part of glycolysis by T6P or Tps1. Another question that awaits further investigation is the molecular basis for the fragility and thermosensitivity of tps2 mutants in different organisms, a characteristic that suggests interactions of T6P with pathways implicated in cell wall synthesis. A deeper understanding of these questions will be important both to basic and to applied science.
We thank Juana M. Gancedo for stimulating discussions during our work, advice on the organization of the manuscript and critical reading of it, M.A. Blázquez (IBMCP, CSIC-UPV, Valencia, Spain) O. Zaragoza (Albert Einstein College of Medicine, New York, NY, USA) and J.M. Siverio (Universidad de la Laguna, Spain) for critical reading of the manuscript. Work in our laboratory over the last few years has been supported by different Spanish and EU grants and is currently funded by Grant BMC-2001-1690-CO2-01 from the Spanish Ministry for Science and Technology.