Final Report Summary - DEPICT (Design principles and controllability of protein circuits)
The DEPICT project uncovered unifying principles that can help us make sense of why complex biological circuits in the cell are shaped the way they are. It combined of theory and experiment on systems ranging from bacteria to cancer cells.
Paradoxical enzymes - called paradoxical because they do two opposite reactions - have an important role: they make biological circuits work precisely. Biological circuits are made of interacting proteins.
Circuits face a challenge: the concentration of each protein varies from cell to cell due to random processes inherent to biological material. It is thus of interest to understand mechanisms that can make the circuits work precisely despite such variations. In other words, to see how circuits can be made robust. In this project, we showed how seemingly paradoxical components can provide the needed robustness. For example, an enzyme that both adds and removes a chemical group that activates other proteins. We showed experimentally and theoretically how such a paradoxical enzyme makes a crucial E. coli system robust - namely the system that takes up nitrogen from the environment. We found an analogous mechanism in plants, where a paradoxical enzyme provides robustness to the system that takes up carbon from the atmosphere.
In a surprising twist, we realized that the same idea can be carried from the level of protein circuits inside a cell, to the larger-scale level of circuits made of cells that interact with each other. Such circuits are of central importance in the immune system; immune cells communicate by means of secreted signal molecules called cytokines. Some cytokines are paradoxical, because they tell cells to both grow and die- two opposing effects. We showed experimentally, with Nir Friedman, that this paradoxical design provides robustness to T-cell numbers in the face of natural variations that are unavoidable in the body.
Rules for making gene expression respond precisely to input signals.
In biological circuits, some genes are controlled by activators which bind DNA to increase gene expression. Others are controlled by repressors, which inhibit expression. The choice of whether to use an activator or repressor seems arbitrary. We tested a long-standing hypothesis by Savageau, that this choice is not arbitrary but rather reflects the demand for the gene in the natural environment: genes rarely needed are controlled by repressors, and those commonly needed are regulated by activators. We experimentally found that the Savageau rule makes genes optimally insulated from changes in environment.
Again, this project led us to unexpected directions, asking about evolutionary tradeoffs between tasks such as growth and survival in bacteria. We used concepts from economics about multi-objective Pareto optimality, to show that such tradeoffs can lead to unexpectedly simple patterns in gene-expression space – in which data falls on low dimensional polyhedra whose vertices are optimal gene expression profiles for certain key biological tasks. This explains complex gene expression data from E. coli, from breast tumors and even data on animal morphology.
Controllability of human protein dynamics with drug combinations.
Drug cocktails are expected to be very important in treating cancer; finding such cocktails is hard, because the number of combinations to be tested grows exponentially with the number of drugs. We are developing mathematical approaches to deduce the effects of multi-drug combinations on cancer cells, using only a small number of measurements (single drugs and drug pairs, for example). We test this with our unique system which allows following 1000 proteins in individual cancer cells in space and time. We aim to provide a basis for controlling cell protein dynamics in a rational way using drug combinations.
Paradoxical enzymes - called paradoxical because they do two opposite reactions - have an important role: they make biological circuits work precisely. Biological circuits are made of interacting proteins.
Circuits face a challenge: the concentration of each protein varies from cell to cell due to random processes inherent to biological material. It is thus of interest to understand mechanisms that can make the circuits work precisely despite such variations. In other words, to see how circuits can be made robust. In this project, we showed how seemingly paradoxical components can provide the needed robustness. For example, an enzyme that both adds and removes a chemical group that activates other proteins. We showed experimentally and theoretically how such a paradoxical enzyme makes a crucial E. coli system robust - namely the system that takes up nitrogen from the environment. We found an analogous mechanism in plants, where a paradoxical enzyme provides robustness to the system that takes up carbon from the atmosphere.
In a surprising twist, we realized that the same idea can be carried from the level of protein circuits inside a cell, to the larger-scale level of circuits made of cells that interact with each other. Such circuits are of central importance in the immune system; immune cells communicate by means of secreted signal molecules called cytokines. Some cytokines are paradoxical, because they tell cells to both grow and die- two opposing effects. We showed experimentally, with Nir Friedman, that this paradoxical design provides robustness to T-cell numbers in the face of natural variations that are unavoidable in the body.
Rules for making gene expression respond precisely to input signals.
In biological circuits, some genes are controlled by activators which bind DNA to increase gene expression. Others are controlled by repressors, which inhibit expression. The choice of whether to use an activator or repressor seems arbitrary. We tested a long-standing hypothesis by Savageau, that this choice is not arbitrary but rather reflects the demand for the gene in the natural environment: genes rarely needed are controlled by repressors, and those commonly needed are regulated by activators. We experimentally found that the Savageau rule makes genes optimally insulated from changes in environment.
Again, this project led us to unexpected directions, asking about evolutionary tradeoffs between tasks such as growth and survival in bacteria. We used concepts from economics about multi-objective Pareto optimality, to show that such tradeoffs can lead to unexpectedly simple patterns in gene-expression space – in which data falls on low dimensional polyhedra whose vertices are optimal gene expression profiles for certain key biological tasks. This explains complex gene expression data from E. coli, from breast tumors and even data on animal morphology.
Controllability of human protein dynamics with drug combinations.
Drug cocktails are expected to be very important in treating cancer; finding such cocktails is hard, because the number of combinations to be tested grows exponentially with the number of drugs. We are developing mathematical approaches to deduce the effects of multi-drug combinations on cancer cells, using only a small number of measurements (single drugs and drug pairs, for example). We test this with our unique system which allows following 1000 proteins in individual cancer cells in space and time. We aim to provide a basis for controlling cell protein dynamics in a rational way using drug combinations.