Another player joins the complex field of sugar-regulated gene expression in plants

All organisms need to be able to sense and respond to the availability of metabolites, such as carbohydrates. In many cases, these metabolic responses are mediated by alterations in the expression of specific genes. For instance, the phenomenon of carbon catabolite repression, wherein expression of a specific set of genes is repressed in the presence of glucose, has been studied extensively in bacteria (reviewed in ref. 1) and yeast (reviewed in ref. 2), and much is known about the underlying regulatory mechanisms in these systems. Specific genes have also been shown to be either induced or repressed in response to alterations in carbohydrate levels in plants (3–8). In recent years, several components of the plant metabolite sensing/response machinery have also been identified. Characterization of one of the most recently identified of these components, the PRL1 (pleiotropic regulatory locus) protein, is reported in this issue of the Proceedings by Bhalerao et al. (9).

The available evidence suggests that the molecular mechanisms by which gene expression is tied to the concentrations of certain metabolites, such as soluble sugars, are very complex. In organisms where metabolic regulation of gene expression has been relatively well characterized, evidence for several signal transduction pathways between soluble sugar levels and gene expression has been obtained (reviewed in ref. 10). For instance, studies on the yeast Saccharomyces cerevisiae have provided evidence for as many as four signal transduction pathways for sugar-regulated gene expression. Among the best-characterized of these pathways are the RAS-cAMP and the “main glucose repression/derepression” pathways. Both of these pathways are believed to be activated via glucose phosphorylation by a particular hexokinase (11, 12), although more recent data indicate that two unusual glucose transporters may also play a role in glucose sensing in yeast (13). Other proteins involved in yeast sugar sensing/response include the SNF1 protein kinase, which has been shown to play a key role in the main glucose repression/derepression pathway (14), and a specific G protein, Gpa2 (15). Other, less well characterized, pathways for sugar-regulated gene expression in yeast are the “fermentable growth medium induced pathway” and the “slow, proteolytic inactivation pathway” (reviewed in ref. 10). Adding to the complexity of mechanisms responsible for adapting cellular metabolism to changes in available sugars is the fact that the different sugar sensing/response pathways can affect gene expression at different levels. Examples of sugar-mediated gene regulation have been found at the transcriptional, RNA processing/stability, and post-translational levels (10, 16–18).

Although the field of metabolic regulation in plants is still in its infancy, evidence that plants contain more than one sugar sensing/response pathway, and that the plant pathways share some, but not all, of the components found in sugar sensing/response pathways from other organisms, has already been obtained. For instance, hexokinase, which has been postulated to play an important role in sugar sensing in other organisms (e.g., refs. 11 and 12), has also been implicated in the mechanism responsible for sensing sugars and causing the repression of a number of plant genes (19–21). However, the expression of other sugar-regulated plant genes has been shown to be hexokinase-independent, providing good evidence for the existence of multiple sugar sensing/response pathways in plants (22, 23). Besides hexokinase, another component of sugar sensing/response pathways that appears to have been well conserved between different organisms, including plants, is the SNF1 protein kinase (reviewed in refs. 24 and 25). However, while plant sugar sensing/response pathways do share some common features with analogous pathways from other organisms, the plant pathways also appear to have some unique features. For instance, calcium-dependent protein kinases have been suggested to play a role in sugar-regulated gene expression in plants but are not known to act in the sugar response pathways of other organisms (26, 27).

Adding to the challenges associated with unraveling multiple sugar sensing/response pathways that affect gene expression at different levels is the fact that the roles of some of even the best-characterized components of these pathways remain controversial. For example, while hexokinase has been postulated to function as a sugar sensor in organisms ranging from yeast to plants, the exact role of hexokinase in sugar sensing is still being debated (reviewed in ref. 28). Early evidence suggesting that hexokinase functions as a sugar sensor came from studies using the yeast Saccharomyces cerevisiae, where partial loss-of-function mutations in the gene encoding the major hexokinase were found to result in the partially constitutive expression of glucose-repressed genes (29). However, so far, attempts to dissociate the role of this hexokinase in catalyzing phosphorylation of hexoses from its postulated role in sugar sensing have been unsuccessful (11, 12). Additional support for the hypothesis that hexokinase is a sugar sensor comes from experiments with glucose analogues. Supplying plants with glucose analogues, such as 2-deoxyglucose, that can be phosphorylated by hexokinase but not further metabolized at an appreciable rate results in alterations in the expression levels of sugar-regulated genes that are similar to the alterations seen when plants are supplied with glucose (30). In contrast, glucose analogues, such as 3-O-methylglucose, that are not phosphorylated by hexokinase do not affect the expression of most sugar-regulated genes. These results suggest that phosphorylation by hexokinase, but not further metabolism, is required for sugar sensing. However, interpretation of these results is complicated by the fact that feeding organism compounds, such as 2-deoxyglucose, that can be phosphorylated but not efficiently further metabolized, may have numerous effects on metabolism. For instance, the phosphorylation of such compounds may result in a significant reduction in ATP levels. In fact, feeding yeast 2-deoxyglucose at levels as low as 1 mM caused a 67% reduction in ATP levels, while 40 mM 2-deoxyglucose caused a 97% decline in ATP (31).

Experiments in which plants were supplied with glucose analogues have shown that only hexokinase substrates are active in controlling the expression of the majority of sugar-regulated genes (3, 19). However, these experiments are subject to the same limitations stated above for similar experiments performed on yeast. Namely, differentiating between a specific role for these compounds in sugar signaling, as opposed to more general metabolic effects, is problematic. Experiments in which different hexokinase-encoding genes were expressd in sense and antisense orientations in transgenic plants also suggest a role for hexokinase in sugar sensing (21). In these experiments, transgenic Arabidopsis seedlings that overexpressed either of two Arabidopsis hexokinase genes, AtHXK1 and AtHXK2, showed retarded growth, relative to wild-type seedlings, on 330 mM glucose. In contrast, seedlings that contained antisense constructs of either AtHXK1 or AtHXK2 showed much better growth than wild-type seedlings on 330 mM glucose. Intriguingly, transgenic Arabidopsis expressing a yeast hexokinase showed increased hexokinase catalytic activity in in vitro assays but decreased sensitivity to sugar, suggesting that it may be possible to separate the role of hexokinases in sugar sensing from their role in phosphorylating sugars (21).

The role of the SNF1 gene in sugar signaling has been less controversial than that of hexokinase. In Saccharomyces cerevisiae, missense and null mutations in the SNF1 gene have been shown to result in derepression of glucose-repressed genes (14). Genes with significant sequence similarity to the yeast SNF1 gene have been found in organisms ranging from mammals (reviewed in ref. 24) to plants (reviewed in ref. 25). In addition, putative SNF1 homologs from rye (32), tobacco (33), and Arabidopsis (9) have been shown to complement yeast snf1 mutants, demonstrating that the plant proteins are functionally similar to SNF1. Further evidence for a role for SNF1 homologs in plant sugar response comes from an experiment where antisense expression of a putative SNF1 homolog in potato resulted in loss of the sugar-inducible expression of sucrose synthase (34). Still, interpretation of the role of SNF1 and SNF1 homologs in sugar signaling is complicated by the fact that, at least in some organisms, SNF1 also plays a role in stress adaptation by responding to stress-induced alterations in AMP/ATP ratios (24, 35). This response involves the post-translational control, via phosphorylation, of a number of proteins. The affected proteins include key biosynthetic enzymes, thus allowing cellular metabolism to adapt to the imposed stress. This second role of SNF1 raises the possibility that some SNF1-mediated responses to sugar may result from its role in stress signaling, rather than in sugar signaling per se. For instance, some SNF1-mediated responses to high sugar concentrations could be a result of osmotic shock-induced stress, rather than a specific response to sugar.

The yeast SNF1 gene encodes a protein serine/threonine kinase (14) that acts in a “SNF1 complex” together with the product of the SNF4 gene, as well as the product of the SIP1, SIP2, or GAL83 genes (36). Recently, a potential component of a plant “SNF1-like complex” was identified (37). This protein, the PRL1 protein, interacts with the yeast SNF1 and plant SNF1-like proteins in the yeast two-hybrid system and in vitro (9). The PRL1 gene was first identified by screening mutagenized Arabidopsis for the ability to grow on 175 mM sucrose or glucose. Unlike wild-type plants, plants carrying a null mutation in the PRL1 gene show arrested growth at this sugar concentration. Subsequent characterization of the prl1 mutant revealed defects in sugar-, light-, and stress-regulated gene expression, as well as in starch and sugar accumulation and in response to several phytohormones (37, 38). The PRL1 gene has been isolated and found to encode a protein that contains repetitive sequences characteristic of a family of regulatory proteins known as WD-40 repeat proteins (37). The PRL1 protein has also been shown to interact with an α-importin nuclear import receptor (37) and, as reported in this issue of the Proceedings by Bhalerao et al. with SNF1 (9). As the PRL1 protein interacts with SNF1, and mutations in the PRL1 gene confer a sugar-hypersensitive phenotype, the PRL1 protein has been postulated to function in plant sugar-regulated gene expression by acting as a negative regulator of SNF1 homologs (9).

At this time, the role of PRL1 in phytohormone and light response remains to be delineated. One possibility is that sugar, phytohormone, and light signal transduction pathways in plants interact, and may even share some components. In fact, the plant ethylene and sugar signaling pathways were recently proposed to share a common branch (39). In support of this hypothesis are experiments showing that exogenous ethylene can phenocopy a mutation, the gin1 (glucose insensitive) mutation, that allows plants to germinate, undergo cotyledon expansion, and develop true leaves on high concentrations of glucose that would otherwise inhibit these processes (39). Support for close interactions between sugar and light signaling pathways has also been provided by studies on the sun (sucrose uncoupled) mutants of Arabidopsis, which show defects in both sugar-related gene expression and light responses (40, 41). Alternatively, at least some of the connections between sugar-, phytohormone-, and light-induced responses might be indirect, possibly representing general effects on metabolism. Therefore, when a mutant that is disrupted in response to a particular phytohormone also shows an altered response to exogenous sugars, this does not necessarily mean there is crosstalk between two separate signaling pathways.

In fact, several possible indirect effects of sugars on plant gene expression, metabolism, and development remain to be fully explored. For instance, supplying plants with exogenous sugars, particularly at high concentrations, is likely to cause diverse alterations in general metabolism. Potential alterations include, but are not limited to, changes in AMP/ATP ratios, pH, and calcium levels. Any or all of these changes could then have effects on multiple signal transduction pathways. Besides affecting general metabolism, supplying plants with high concentrations of exogenous sugars also causes an osmotic shock that could trigger a general stress response. Distinguishing the effects of exogenous sugars on sugar sensing as opposed to osmotic shock response pathways is likely to be complicated by the fact that, as mentioned above, some factors, such as SNF1, have been implicated in both sugar and stress signaling pathways (24, 35). In addition, some of the chemicals, such as sorbitol and mannitol, most commonly used as osmotic controls are osmoprotectants (e.g., see refs. 42 and 43). Therefore, experiments demonstrating that particular mutations affect sugar sensitivity, but do not affect sensitivity to equimolar concentrations of sorbitol or mannitol, may not be sufficient to rule out an effect of the mutations on osmotolerance. Finally, experiments on a variety of organisms have demonstrated that glycation, the nonenzymatic glucosylation of proteins and lipids (44), results in the addition of glucose moieties to a wide variety of proteins (reviewed in ref. 45). Examples of proteins shown to glycated in different organisms include a glucose transporter (46), glucokinase (47), a calcium pump (48), and calmodulin (49). In at least some cases, glycation has also been shown to result in inactivation of the affected protein. While glycation has not been studied extensively in plants, the high levels of exogenous sugars used in many plant sugar response studies seem likely to result in increased formation of glycated proteins. The effects of increased glycation in plants represent an interesting area for further inquiry.

Interest in plant metabolic regulation has been increasing rapidly in recent years. Although the existence of multiple metabolite signal transduction pathways, potential interactions with other signaling pathways, and possible indirect effects of sugars (and other metabolites) on plants complicate this field of research, recent advances are providing useful tools for further investigations. One such advancement, the demonstration that the PRL1 protein interacts with SNF1 in a sugar-dependent manner, is reported in this issue of the Proceedings by Bhalerao et al. (9).