What do studies of transgenic plants reveal about the integration of metabolism
Within the past decade advances in genetics and molecular biology has facilitated brand new ways of looking at metabolic processes. Far from the traditional reductionalist approaches of the previous years, we are able undertake a more holistic approach towards understanding metabolic pathways and networks. The most important advance has been a move away from inferred models based on in vitro characteristics of enzymes to real-time studies in vivo of enzymes at work.
One of the best understood metabolic networks (and also among the best funded in plants) is the primary pathway for Carbon fixation, the Calvin cycle. In this essay I shall describe how the regulation of metabolic pathways was originally approached, how the use of genetics has changed this approach and describe in detail some experiments on enzymes in the Calvin cycle and how the results from these has caused a reappraisal of our understanding of how metabolism is controlled in plants.
Metabolic pathways consist of a series of chemical modifications to a compound which results in substrates being turned into products. At each step of the way enzymes are used in order to allow the reaction to occur at physiological temperatures and at a speed conducive to homeostasis. The regulation of the flux of a pathway has been one of the key questions in understanding metabolism; is flux regulated by a series of steps that act as a bottleneck to the system or by co-limitation by several enzymes.
Generally speaking enzyme activity can be modified via two different mechanisms. For short-term, “fine” changes, due to a change in the environment, enzymes can be modified by altering the existing enzymes kinetics, changing levels of substrate, inhibitors or activators and by post-translational modification (e.g. carbamylation of rubisco). Longer-term “coarse” changes, such as those incurred during development, require the amounts of enzymes to be altered, through modification of the transcription of the gene or by protein turnover.
Original thinking behind regulation of flux in a pathway has been that enzymes catalysing steps far from thermodynamic equilibrium (i.e. irreversible) are best suited for regulation as they provide a bottleneck for the pathway that relies on their catalytic abilities. Enzymes that catalyse steps that are at or near thermodynamic equilibrium (i.e. readily reversible) would be poor sites for regulation as they are likely to be present in high numbers and control will be unable to favour either the products or substrates in the reaction. Evidence to support this view comes in the form that irreversible enzymes tend to have regulatory properties, including allosteric regulation and/or post-translational regulation.
Also a good candidate for a regulatory enzyme would show some reciprocal relationship between its concentration of substrate and overall flux of the pathway. For example if the in vivo concentration of an enzymes substrate is reduced and subsequently the flux of the pathway it is involved in also goes down then it can be hypothesised that the enzyme catalysing that step is responsible in part for the overall flux. Given these hypothetical qualities Newsholme and Start in 1973 developed a theoretical analysis for the study of regulation. Stating that an enzyme is likely to regulate the flux of a pathway if it fulfils these criteria:
1. They catalyse an irreversible reaction.
2. They possess regulatory properties.
3. They show characteristic reciprocal relationship between the in vivo flux and their substrate concentration.
Using this framework enzymes were identified, purified and studied for their kinetic and regulative properties in vitro and by correlation with in vivo expression patterns, changes in fluxes and so on, it was then possible to frame a hypothesis about how the enzymes was regulated. This approach produced a large amount of information regarding individual properties of enzymes but there are many problems associated with this approach which made it very difficult to conclusively prove the models created. It is not to say that all study went on in vitro just that there was a large imbalance between the two due to the lack of technical ability to study in vivo systems properly.
The fundamental flaw in this analysis is that enzymes are integral to the pathway and once an enzyme is removed from its native surroundings it behaves differently. Any hypothesis being drawn from it can be misleading, much like observing an animal’s behaviour in the zoo and inferring how it reacts in the wild. By limiting analysis to only enzymes that facilitate irreversible reactions resulted in many potential (although thought unlikely) site for regulation were missed out. Also showing a correlation between flux and enzyme activity in vitro does not imply that the enzyme is primarily responsible for that change and merely because they have regulatory properties does not mean that they are used in vivo.
A New Approach
A step forward in thinking about how metabolic pathways should be analysed was the differentiation between enzymes “regulatability” and their “regulatory capacity” (Hoffmeyer and Cornish-Bowden1991). In order to address the problems described above, Hoffmeyer made the distinction between an enzymes potential to be regulated in vivo (regulatability), and the contribution that enzymes regulation makes to the overall flux of the system (regulatory capacity). The theoretical analysis traditionally applied to metabolism only identified enzymes with a high “regulatability” it does not provide us with a set of logical criteria with which to understand an enzymes “regulatory capacity” or help define exactly which enzymes are key control points in the system.
In order to assess the “regulatory capacity” of an enzyme, small changes in the activity of an enzyme must be made in vivo and measuring the effect it has on the flux through the whole pathway. This can be visualised by plotting the pathway flux against the enzyme activity normalised to the wild type value.
If the enzyme is the sole control point, limiting the entire flux of the system, then there is a linear and strictly proportional relationship between enzyme activity and flux of the system. When the enzyme together with other enzymes, co-limits flux there will be a curvilinear response with a finite but non proportional slope in the range corresponding to wild type plants. Where the enzyme has no significant control in the system the slope will be zero in the wild type range. What is important to note is that at some point the flux will change according to enzyme activity although this only shows that the enzyme is essential or redundant not whether it plays a role in regulation of flux.
The degree to which an enzyme is said to control the flux in a pathway is referred to as the enzymes flux control coefficient. Derived by Kacser and colleagues from the above graph and formalised into this equation.
Where J is the flux and E is the enzyme activity normalised to wild type and C is the flux control coefficient. Since this relates to the gradient of said graph it is deducible that C will always take a value between 0 (non-limiting) and +1 (strictly limiting). For the entire pathway the sum of all enzymes flux control coefficient must equal 1 (in the wild type plant) otherwise there are factors that play a regulatory role.
The value of an enzyme’s flux control coefficient cannot be predicted from its properties, or the thermodynamic state of its reaction. The model developed by Kacser predicts that the values of flux control coefficients of the enzymes in the pathway emerge from an interaction between all the enzymes of the pathway. Due to this value depending on the interactions within the system it can only be determined by a holistic approach. Experimental measurements of flux control coefficients started with estimates from metabolic models, using inhibitors to titrate out the enzyme activity in vivo and careful control of substrate concentrations. The problem with these assumptions is that they require a simplification of the system based on prior knowledge of the enzymes and their kinetic properties and also that specific inhibitors and substrates have been identified and can be modulated in vivo.
Another approach is to search for mutants that express varying levels of the enzyme in question and assess its flux with comparison to a wild type. Problems arise here due to the fact that the enzyme in question must confer some sort of selective phenotype for screening processes and the majority of metabolic processes either have no easily identifiable phenotype or it is covered up by redundancy in the system. In order to accurately obtain the flux control coefficient a range of mutants must be identified each displaying different levels of enzyme activity, you cannot plot a graph with two points. With this in mind it appears that mutant screening for different enzyme expression levels is a hopeless task although some groups have managed to achieve it.
With the advent of molecular biology and an increased understanding of genetics, the last decade or so has provided a methodology that has revolutionised the way in which the study of flux control coefficients has been approached. Now it is possible to directly manipulate levels of individual enzyme activity by altering their expression patterns though either upregulation (as in ectopic expression) or downregulation (antisense, cosupression (is it really understood yet?) and targeting of foreign proteins (I don’t understand the point of them, it is in order to move orthologues to the right place as in Miyagawa)) of the genes that encode them. Another decisive advantage of transgenic plants it that they reflect the importance of the enzyme not only acting in the pathway but for whole plant development.
The most popular method of creating mutants with varying levels of expression is to use antisense technology and rely on random insertion and position effects to generate different enzyme activities (is this RNAi now?). Others have been used and have shown promising results especially when orthologues of native genes are ectopically expressed which do not undergo normal regulation (Miyagawa 2001). Ideally plants with an activity of enzyme relative to wild type ranging from 60-70% right down to 10% above these values the enzyme activity is hard to determine exactly due to compensatory effects and below the enzymes do not show regulation but necessity.
The Calvin Cycle
One of the fields to have benefited the most from this research is the primary carbon fixation cycle or Calvin cycle. Photosynthetic carbon metabolism in higher plants is thought to be one determining factor in plant growth and yield. The cycle combines a five carbon compound (ribulose-1,5-bisphosphate) with a molecule of CO2 and through a process of reduction creates two triose phosphates which can be siphoned off into other metabolic pathways or used to regenerate the ribulose 1,5-bisphosphate. Overall there is a net gain in carbon, hence the term carbon fixation cycle.
This is a diagram of the Calvin cycle showing the pathway of carbon as it is fixated. Of the 13 reactions involved in the cycle 11 of them are catalysed by enzymes.
Irreversible reactions are catalysed by the following 4 enzymes.
1. ribulose-1,5-bisphosphate carboxylase-oxygenase (Rubisco)
2. sedoheptulose-1,7-bisphosphatase (SBPase)
3. fructose-1,6-bisphosphatase (FBPase)
4. ribulose-5-phosphate kinase (PRK)
Although there are many flaws with the traditional approach these enzymes had already been well characterised and their encoding sequences were well known, making them perfect candidates for investigating the contribution they make to flux in the Calvin cycle.
One question to be asked is why regulate the Calvin cycle at all? The surrounding environment is not static and the system must be able to make changes in different light intensities and CO2 concentrations. Since the Calvin cycle relies upon the products of the light dependant reactions in photosynthesis ATP/NADPH, it must be able to switch off during the dark otherwise it uses carbohydrates in order to create the ATP/NADPH which in turn is used to make new carbohydrates. Such changes in the relative concentrations of intermediates will effect other pathways which share them, such as glycolysis and oxidative PPP.
I shall now outline some experiments performed on these enzymes and in turn describe what effects the results have had on the traditional ideas regarding regulation of the Calvin cycle.
Ribulose-1,5-bisphosphate carboxylase-oxygenase (Rubisco)
Rubisco is the most abundant enzyme on the planet. It is responsible for the fixation of CO2 in the Calvin cycle and was generally thought to regulate (limit) the rate of photosynthesis. It comprises of approximately 40% of the total protein in a leaf and represents a key site in the carbon and nitrogen economy of the plant. Rubisco was well known to be a highly regulated enzyme, undergoing carbamylation in order to become active which then allows a host of activators and phosphorylated intermediates to interact with it. What was not known was its “regulation capacity”. Stitt and Schulze (1994) used a set of “antisense” tobacco plants in order to determine the flux control coefficient. Under moderate light, ambient photosynthesis was only slightly inhibited when rubisco was reduced to about 60% of the wild type variants and a control coefficient of 0.05-0.15 was estimated.
This had massive implications for the traditional thinking. Here was an enzyme that was highly regulated yet appeared to contribute very little to the overall flux of the system. It was thought that since the rate of photosynthesis stayed nearly the same in the mutant 60% plants they must be using rubisco in a more efficient manner than their wild type counterparts. Studies showed that although there was a slight increase in substrates to compensate for the loss that would only account for 15% of the “missing” protein. What had occurred was that the level of the carbamylated (active) enzyme was increased in order to compensate. It was clear that redundant mechanisms were present in order to adjust the system to perturbations, revealing a new level of control and feedback within metabolism.
Perhaps one of the most interesting results to emerge from the study was that the flux control coefficient varied according to the conditions photosynthesis was measured in. It was already known that in moderate light C 0.05-0.15 but when plants were grown in low light then suddenly plunged into bright light (altering the rate of photosynthesis) there was a near proportional relationship between the amount of Rubisco and the rate of photosynthesis (C ;0.9!). This result was echoed when repeated with low CO2 growth levels and yet the converse was obtained when measurements were taken in 5% CO2. This clearly shows that an enzymes regulatory capacity depends on the short term conditions under which the flux was measured. And also due to the long term conditions the plant was subjected to before the change in conditions and flux was measured. (There doesn’t seem to be any difference between 4.1 and 4.2 in terms of experiments in Stitts paper the first will be the last)
In summary the analysis of Rubisco mutant revealed that it plays a much smaller part in regulation of the Calvin cycle, thus control must be shared with other enzymes. The only case where rubisco appears to be the major limitation in the Calvin cycle is in conditions of high irradiance and temperature. There exists a certain amount of redundancy within the system which can compensate for changes in the levels of enzyme expression. Also the enzymes regulatory capacity depends on the conditions it is measured in and the conditions it was produced in (short term and long term effects).