Dynamics And Functions Of Gene Regulatory Network Motifs

Cells are complex systems composed of thousands of different types of molecules with complicated regulatory interactions.In the view of systems biology,such a system can be characterized by a biochemical network,i.e.,interactions among molecules can be characterized by the connectivity of the network components.The research on the behaviors of biochemical networks can help us understand many cellular processes at the system level.Network motifs are recurring circuits of regulatory interactions from which the networks are built.They are thought of as the building blocks and functional elements of the whole network.Thus,studies of network motifs are fundamental for understanding the whole system.Dynamics of biological network motifs can reveal the behaviors and functions of those motifs,explain the essential mechanisms for the cellular processes they underlie.In this dissertation,I focus on two common types of dynamics of coupled feedback loops—multistability and oscillation.Multistability is a crucial theme in cell fate decision,and is often used to explain the mechanism for cell differentiation.Qualitatively,it is attributed to the presence of positive feedback loop(PFL)in coupled feedback loops,but the general condition and essential mechanism for realizing multistability remain unclear.In our first work,we build a generic circuit model comprising two transcription factors and a microRNA,representing a kind of core architecture in gene regulatory networks.The circuit can be decomposed into two positive feedback loops(PFLs)or one PFL and one negative feedback loop(NFL),which are multiplicatively coupled.We formulate the gener-al requirement for tristability in terms of logarithmic gain of the circuit.Bifurcation analyses of the model reveal that the circuit can achieve tristability through four kinds of bifurcation scenarios when parameter values are varied in a wide range.Coupling two PFLs with bistability or one NFL with a bistable PFL is most likely to generate tristability,but the underlying mechanisms are largely different.The parameter ranges for tristability and possible transition routes among steady states are determined by the combination of gain features of individual feedback loops,which provides the theoreti-cal fundamentals for controllable differentiation and reprogramming.We also interpret published results and make testable predictions.This work sheds new light on inter-linking feedback loops to realize tristability.The proposed theoretical framework can be of wide applicability.Oscillations are common in cellular processes,and observed in circadian rhyth-micity,cardiac rhythmicity and cell cycle.It has been found that biological oscillators often engage interlinked PFL and NFL to increase the robustness of oscillation and the accuracy of timing.Since time delays in feedback loops are often inevitable,it is necessary to understand the roles of time delay in the dynamics of such systems.In our second work,we build a simplified model for a synthetic genetic circuit,with the delayed PFL and NFL multiplicatively coupled,and explore various effects of time de-lays on the system dynamics.The stability of steady states and the routes to oscillation are analyzed in detail.We show that large-amplitude oscillations can be induced when the PFL has a longer delay than the NFL and unravel the underlying mechanisms.We also develop a stochastic algorithm for simulating a single reaction with two time de-lays and display that robust oscillations can be maintained by the PFL with long delay.This work provides an effective method for constructing robust large-amplitude oscil-lators and interprets why similar circuit structures are involved in timing keeping such as circadian rhythmicity.To summarize,this dissertation studies the dynamics and functions of one of the important gene regulatory network motifs—coupled feedback loops.In our first work,we explore the multistability in coupled feedback loops in terms of logarithmical gain-s,which greatly facilitates the quantitative analysis of the conditions for multistability.Especially,not only coupling two bistable PFLs can produce multistability,but cou-pling one bistable PFL and one NFL can.These findings expand the knowledge of multistability in coupled feedback loops.Besides,the features of individual feedback loops and the multistability of coupled systems are clarified,which may offer a pos-sible theoretical fundamental for the control and design of multistable circuits.In our second work,we systematically investigate the modulation of oscillation by the PFL in the coupled delayed PFL and NFL.We find the functions of the PFL depend on its time delay relative to the NFL's.Importantly,we emphasize that the PFL with a long delay may promote oscillation to a larger degree and generate larger robustness,compared with the non-delayed PFL.We explain this phenomenon analytically in an extreme case.These findings not only update the traditional notion on the function of PFLs,but also help people understand the roles of PFLs with long delay in circadian clock circuits.Finally,the stochastic simulation algorithm we developed for multi-delayed chemical reaction can be of wide applicability.The dissertation is organized as follows:In Chapter ?,I briefly introduce the back-ground of systems biology,methods of network dynamics,network motifs and my re-search topics.In Chapter ?,I present the work on how to realize the multistability in genetic regulatory networks.In Chapter ?,I present the work on how a delayed PFL modulates the oscillation of the coupled PFL and NFL.In Chapter ?,I draw the conclusions of this dissertation and give suggestions for future work.