Life Sciences Division

More efficient control of energy and material resource use – inspired by plants

CEA

Living organisms have developed their own systems for regulating the metabolic pathways so crucial to maintaining their survival. Why not draw inspiration from these mechanisms to improve how we control our own cycles of material and energy use? This is the perspective that prompted CNRS, INRA and CEA researchers to mathematically model the biosynthetic pathways of amino acids in the Arabidopsis plant. This model has made it possible to faithfully reproduce the measurements performed on whole plants, thus demonstrating its accuracy. Running simulations with the models thus provides key qualitative and quantitative insight into how this kind of regulatory control system works.

Imagine a car factory that, starting out with the same stock of spare parts, can simultaneously produce several different car models at an output rate of 20,000 cars per second. Imagine that the factory has to adapt production rates to a more or less efficient input supply of spare parts and to highly volatile consumer demand. This would make the factory capable of manufacturing more than a billion cars in just one day. At this kind of output rate, it is clear that the slightest lag in any process step could result in absolute chaos.

The processes driving the synthesis of small molecules (metabolism) in living systems have, over the course of evolution, developed control systems capable of handling flows like these, and in some cases – like photosynthesis in plants –at rates 1000-fold more intensive. To illustrate, sunlight – used by plants as their source of energy – can increase sharply when a cloud has passed, and a squall of wind can cause at ten-degree drop in temperature (halving the work rates in the cell). Control mechanisms step in as essential, preventing stock shortages or indeed excess stock feed that would slow plant growth or possibly even kills the cells.

Quantitative control systems research in biochemical systems could therefore yield key insights that could, for example, be usefully recycled into our own process flow control measures governing energy or material distribution networks.

This was the idea that set the ball rolling a decade ago, when scientists from the Institute of Life Sciences Research and Technologies (CEA Grenoble) set themselves the goal of modelling a complex metabolic system (Fig. 1) comprising numerous branches and a highly dense network of regulatory steps (synergy, feedback, inhibitions, activation, dual controls). Although metabolic system models had already been published at the time and other models have since been released, none of them tackled with the same level of complexity as that illustrated in Figure 1.

Building on plant research led since the 1950s into the biosynthetic pathways of essential amino acids (those that animals are unable to synthesize) and exploiting the completed DNA sequencing of Arabidopsis, the team was able to perform the following operations:

- separation of each individual enzyme (worker) from the other bacterial proteins. This was achieved by inserting plant enzyme genes into bacteria which were then used to produce huge numbers of enzymes.

- gauging the rate at which these workers transform or assemble parts (metabolites).

- gauging how certain worker-groups (allosteric enzymes) are sensitive to different signals (allosteric enzymes work faster or slower depending on the available levels of certain stocks). These ‘long-distance’ signals are highlighted as the arrows in Figure 1.

- counting the numbers of workers in the factory (the chloroplast) that are enrolled in each step, using antibodies specifically recognizing these enzymes (antibodies produced by rabbits immunized with the purified proteins).

Each individual information building-block alone does not carry any essential meaning. Nevertheless, when the full dataset was assembled into a mathematical model, it became possible to simulate system behaviour (Fig.2) and reproduce measurements performed on whole plants.

The model was also able to quantitatively reproduce the behaviour of a number of mutant plants. With these results validating the mode, simulations were run, providing key qualitative and quantitative insights into how this kind of regulatory control system works.

Regulatory control mechanisms acting immediately upstream of each branch ensure what in engineering terms is understood as a feedback function. However, there are other mechanisms acting further upstream whose function is to control the potential consequences of changes in the demand on one branch on fluxes in the other branches. With this type of control mechanism in place, fluxes are allowed to vary significantly without any great impact on the levels of intermediate or final levels of stock.