Signalling in primary metabolism
Signals, sensing and plant primary metabolism: the 3rd International Symposium of the Collaborative Research Center SFB 429, Potsdam, Germany, April 2006
Primary metabolism forms a backbone of connected core metabolic pathways that generate energy and lead to the synthesis of the polymeric building blocks of the plant cell. In addition, primary metabolism provides the precursors from which secondary metabolic pathways radiate. Overall, metabolism forms a highly interconnected ‘small world’ network in which each metabolite is connected to every other metabolite (Wagner & Fell, 2001). Given the obvious regulatory challenge this imposes, it is remarkable that metabolic networks can be maintained in perfect balance such that a steady state of metabolite concentration is achieved. The mechanisms that govern this metabolic balancing act involve feedback of information on metabolite concentration and a responsive alteration of enzyme activity, generally through allosteric regulation. The emphasis on steady state and the textbook view of metabolic pathways as set-in-stone structures can easily lead one to forget that metabolism is a dynamic process. In particular, primary metabolism responds to the availability of core nutrients such as carbon and nitrogen as well as to altered environmental conditions. A change in these parameters will cause a shift in metabolism from one steady state to another, such that the metabolic output is tailored exactly to the new condition. While analysis of enzyme regulatory properties has been the mainstay of classical plant biochemistry, investigation of dynamic changes in plant metabolic networks has become the new focus of the postgenomic effort. The major challenges now facing plant biochemists are to understand how changes in nutrient availability and environment are sensed and signalled and to understand the nature of the subsequent metabolic reconfiguration. The 3rd International Symposium of the Collaborative Research Center SFB 429 (a consortium of five institutions with interests in plant metabolism, based in or around Berlin, Germany) brought together scientists from across the world to discuss these issues.
‘The idea that metabolites are signals of gene expression is not a new one, but considerable advances have been made in recent years in identifying new metabolic signals and in beginning to characterize the mechanisms by which these signals are sensed and transduced’
Flux change and transcript change: which is chicken and which is egg?
In the famous dilemma, deciding which came first is an issue of circular causality. You cannot have a chicken without an egg, but equally you cannot have an egg without a chicken. In metabolism, such strict causality between transcript and flux does not exist, but the sequence of events during metabolic change is not straightforward. The issue is an important one because it has a direct bearing on what is sensed to generate a signal. There are two possible scenarios. First, a change of environment is sensed by processes outside metabolism generating a remote signal leading to gene expression which in turn drives metabolic flux. Secondly, the change in environment directly affects flux (e.g. as a result of altered substrate availability or enzyme inactivation) which leads to a perturbation in the metabolic steady state. The altered metabolite concentrations themselves are sensed and lead to altered gene expression which either brings metabolism back to its original steady state (i.e. homeostatic) or drives flux to generate a new steady state. Which of these two mechanisms operates probably depends on the nature and rapidity of the environmental change. Many changes are regular (e.g. the day/night cycle) and can be anticipated at the gene expression level by linkage to the circadian clock. Nevertheless, even such regular changes may lead to a primary signal from metabolism. Environmental changes can also be irregular and generate a sudden insult that is likely to directly perturb metabolism. The idea that metabolites are signals of gene expression is not a new one, but considerable advances have been made in recent years in identifying new metabolic signals and in beginning to characterize the mechanisms by which these signals are sensed and transduced. There have been two main areas of progress: response to nutrient availability (via sugar and nitrogen signalling) and response to abiotic stress (via redox signalling).
The signalling of gene expression by sugar (sucrose and glucose) concentration has been intensively studied in the last decade, and at least some of the molecular sensors have been identified, including hexokinase and SNF1 (Moore et al., 2003; Rolland et al., 2006). In addition, new signalling molecules such as trehalose 6-P have begun to take centre stage (Halford & Paul, 2003; Satoh-Nagasawa et al., 2006). Nevertheless, we have only begun to scratch the surface of the sugar response and there are many, many unanswered questions. Do different sugars elicit a different response? Do different signalling mechanisms target a different set of genes (Tiessen et al., 2003)? What are the downstream components of the signal transduction cascade? Recently, a transcriptomic approach was taken by Mark Stitt's group to analyse the interaction of circadian clock regulation and sugar signalling (Blasing et al., 2005). The study concluded that both the circadian clock and sugars are major factors in diurnal gene expression patterns. A key part of the experimental strategy was an investigation of diurnal changes in the starchless pgm mutant in which sucrose is very high during the day and very low at night (in contrast to the wild type, in which sugar concentrations show much less diurnal variation). The idea was to provide genetic evidence that sugar contributes to diurnal gene expression changes. While the mutant study did indeed serve this purpose, it also threw up an extremely interesting observation: most of the transcript changes in the mutant were triggered by low, not high, sugar. Stitt calls this the ‘bank manager response’– your bank manager only cares when you have too little money, not too much! The reason that this is particularly interesting is that most of the genetic screens that have been performed to date to isolate sugar-signalling mutants are based on the sensitivity of plants to large amounts of sugar added to the growth medium. Not only are such screens potentially complicated by osmotic effects (Rook & Bevan, 2003), but they may not actually be testing the conditions under which the strongest sugar signalling response occurs.
Abiotic as well as biotic stress conditions lead to an enhanced production of reactive oxygen species (ROS) in different subcellular compartments, depending on the type of stress. Extensive molecular studies of the response to imposition of stress, to direct exogenous application of ROS and to mutation in key antioxidant enzymes have demonstrated that a shift in redox state is perceived within the cell and that a signal is transduced to the nucleus, leading to widespread changes in gene expression (Laloi et al., 2004). Redox signals originating from the photosynthetic electron transport chain have been the most intensively studied, and three different types of redox signals can be defined: (i) those that derive from the redox state of photosynthetic electron transport chain components, (ii) those that derive from changes in the redox state of thiol-containing proteins such as thioredoxins, and (iii) those that derive from generation of ROS, for example in the Mehler reaction of photosystem I (Dietz, 2003).
In the last 30 years, the critical importance of redox-active cysteine residues in the regulation of plant primary metabolism has become an established dogma and is now recognized to represent a central element in the regulation of both gene expression and enzyme activity. Much of our knowledge of redox regulation centres on thioredoxin, a cysteine-containing protein that catalyses thiol/disulfide exchanges. The role of thioredoxin in the activation of the Calvin cycle is one of the classic examples of redox regulation. Recently, proteomic experiments have considerably extended the list of possible proteins that are regulated by thioredoxin, a list that now encompasses every major primary metabolic pathway (Buchanan & Balmer, 2005). Moreover, the thioredoxin family of proteins continues to grow, with at least 19 isoforms in the Arabidopsis genome from the f-, h-, m-, o-, x- and y-types. To further complicate the picture, there are many other thioredoxin-fold containing proteins, including some 31 glutaredoxins and several other thioredoxin-related proteins in Arabidopsis, although the extent to which these proteins play a regulatory role in metabolism remains to be established (Mora-Garcia et al., 2006).
Redox signalling is not exclusively a local phenomenon. The redox state of the chloroplast, for example, is the starting point of a retrograde signalling pathway that communicates between organelle and nucleus, resulting in altered nuclear gene expression (Fey et al., 2005a). Genes regulated by this retrograde pathway were identified in a microarray experiment (Fey et al., 2005b). One major group of regulated transcripts encoded enzymes of amino acid, nucleotide and energy metabolism. It is increasingly clear that the metabolic state of the organelles is a major factor that influences transcriptional regulation of metabolism across the cell.
While it is generally accepted that different parts of the metabolic network must ‘talk’ to each other, we are only just beginning to unravel the mechanisms by which this communication occurs. It is clear that metabolites themselves are important signals, but that other metabolic outputs such as the poise of redox couples can also be sensed and signalled. The challenge that faces us is not only to identify new signalling components but also to understand how different signalling pathways are integrated and how multiple signals are interpreted. Given the complexity of such interactions, computer models are likely to play an important role in developing this understanding (Thum et al., 2003).
LJS is funded by the Biotechnology and Biological Sciences Research Council and IF is a Feodor Lynen fellow of the Alexander von Humboldt Foundation.