Biological membranes are characterised by having a complex composition of lipid species. This complexity arises from the diversity of lipid classess (e.g. phosphatidylcholine, phosphatidylethanolamine, diacylglycerol, phosphatidic acid etc.) and from the variety of acylation/alkylation patterns within each class.
Although the precise lipid composition of a membrane is determined by several factors, it is extremely unlikely that each of the species are regulated individually. However, it is clear that the lipid composition is under homeostatic control. Indeed the steady state concentrations of the lipid classes are generally insensitive to changes in the availability of dietary precursors. The lipid composition of a membrane is determined by a need to maintain its structural and functional properties. This is an example of functional robustness. In addition to species complexity, membrane lipids are connected through a network of biosynthetic transformations. In many cases multiple routes are available for the production of a particular lipid class. This connectivity constitutes a further level of complexity in lipid membranes. When lipid biosynthesis is viewed as a system, it becomes clear that the robustness of bioembranes is a consequence of the combination of species complexity and connectivity complexity. Our work on CTP:phosphocholine cytidylyltransferase (CCT) has shown the activity of membrane-associated enzymes can be modulated by feedback signals arising from the stored elastic energy of membranes. This type of physical feedback sharply increases the connectivity complexity of the system. We have implemented a simulation of a system that contains all the major membrane lipid classes and analysed its robustness with respect to changes in precursor concentrations and with respect to changes in the rates of the individual reactions. We subsequently incorporated physical feedback loops to different enzymes in the pathway to identify what effects these have on robustness. Our data show that this type of feedback at only 3 or 4 reactions leads to a dramatic increase in the robustness of the system. We are now extending the model to encompass the variety of acylation patterns for the different lipid classes. The resulting model will enable us to predict changes in the steady state composition of the various lipid species that result from inhibition of particular enzymes in the pathway. These predictions will be compared with experimental data that are being obtained from mass spectrometry studies by our colleagues in the Department of Child Health.