Transcriptional Regulation in Constraints-based metabolic Models of E. coli Published by Markus Covert and Bernhard Palsson, 2002

Outline Background – Metabolic Network Modeling – What is FBA model? – Application and Challenge of FBA – How to integrate FBA with regulatory constrains Method Results

Metabolic network modeling(MNM)

Terminology I Flux: the rate of flow of metabolites along a metabolic pathway. Stoichiometric matrix

Terminology II Mass balance Constrains – Invariant constrains Physico-chemical – Variant constrains Environmental constraints Regulatory constrains

FBA model

Summary of FBA model

Resource and tools for FBA

Application of FBA model

Challenges of FBA Reconstruction problem – Rely fundamentally on the availability of genome sequences and annotations Incomplete annotation – Some path and enzyme may be missing Select objective function

rFBA Add regulatory constrains to FBA – Boolean formalism ： AND, OR, NOT – ON and OFF If OFF, flux 0 If ON, the flux is calculated by FBA model Trans= IF (G) AND NOT (B) Rxn= IF (A) AND (E)

Metabolic and Regulatory Network Reconstruction The metabolic network was reconstructed by identifying a set of biochemical reactions in the central E. coli metabolism, taken from the annotated genome sequence as well as from biochemical and physiological literature. The regulatory network was derived from the literature data and represented as a set of regulatory rules following established procedures. These rules were based on external conditions and/or internal conditions of the system. Regulatory constraints were described using a Boolean formalism in which gene products are either available (ON) or unavailable (OFF) to the cell.

Method Regulatory network: 149 genes  16 regulatory proteins and 73 enzymes  113 reactions

Transcriptional Regulation and the Calculation of Steady-state Metabolic Flux Distributions FBA was used to determine an optimal metabolic flux distribution for the given conditions For the purposes of these simulations, capacity constraints included maximum uptake rates of oxygen as well as substrates such as glucose, acetate, and lactose, as determined from growth experiments found in the literature. The production of growth precursors in certain ratios was used here as an approximation. LINDO was used to calculate the optimal flux distributions.

Method

Changing Environments and Time-dependent Cell Behavior The time constants that describe metabolic transients are fast (on the order of milliseconds to tens of seconds) as compared with the time constants associated with transcriptional regulation (generally on the order of a few minutes or slower) or cell growth (on the order of hours to days). Therefore, dynamic simulations may be performed by considering the behavior inside the cell to be in a quasi-steady state during short time intervals relative to the environment.

Time-dependent Cell behavior Beginning at T 0, all the condition as initial Generate regulatory rules based on current environmental and internal conditions Determined which genes are up-regulated Set reactions related to up-regulated genes as unconstrained, otherwise set to 0 Run FBA to calculate the flux distribution Terminated in 3sec, calculate environmental and internal conditions Time Delay rFBA

Combined regulatory/metabolic network for central metabolism in E. coli

Mutant Study

rFBAFBA Correct prediction 106/11697/116

Dynamic Growth Simulation Case 1: Aerobic growth of E. coli on acetate with glucose reutilization » When glucose is deleted from the environment, the acetate is then reutilized as a substrate Case 2: Anaerobic Growth on glucose Case 3:Aerobic growth on glucose and lactose

Dynamic Growth Simulation Case 1: Aerobic growth of E. coli on acetate with minimal glucose » When glucose is deleted from the environment, the acetate is then reutilized as a substrate Case 2: Anaerobic Growth on glucose Case 3:Aerobic growth on glucose and lactose

Aerobic growth on acetate with glucose reutilization

Dynamic Growth Simulation Case 1: Aerobic growth of E. coli on acetate with glucose reutilization » When glucose is deleted from the environment, the acetate is then reutilized as a substrate Case 2: Anaerobic Growth on glucose Case 3:Aerobic growth on glucose and lactose

Anaerobic Growth on glucose

Dynamic Growth Simulation Case 1: Aerobic growth of E. coli on acetate with glucose reutilization » When glucose is deleted from the environment, the acetate is then reutilized as a substrate Case 2: Anaerobic Growth on glucose Case 3:Aerobic growth on glucose and lactose

Aerobic growth on glucose and lactose

Reference Transcriptional Regulation in Constraints-based Metabolic Models of Escherichia coli, Markus Covert and Bernhard Palsson, doi: /jbc.M Markus W Covert, Christophe H. Schilling and Bernhard PalssonRegulation of Gene Expression in Flux Balance Models of Metabolism, J Theor Biol Nov 7;213(1): Flux balance analysis of biological systems: applications and challenges, karthik Raman and Nagasuma Chandra, Brief Bioinform (2009) 10 (4): doi: /bib/bbp011 Genome-scale metabolic networks, Marco Terzer Nathaniel D. Maynard Markus W. Covert and J org Stelling, DOI: /wsbm.037 Cellular Metabolic Network Modeling, Eivind Almaas, NetSci Conference