MoVE identifies metabolic valves to switch between phenotypic states

Abstract

Metabolism is highly regulated, allowing for robust and complex behavior. This behavior can often be achieved by controlling a small number of important metabolic reactions, or metabolic valves. Here, we present a method to identify the location of such valves: the metabolic valve enumerator (MoVE). MoVE uses a metabolic model to identify genetic intervention strategies which decouple two desired phenotypes. We apply this method to identify valves which can decouple growth and production to systematically improve the rate and yield of biochemical production processes. We apply this algorithm to the production of diverse compounds and obtained solutions for over 70% of our targets, identifying a small number of highly represented valves to achieve near maximal growth and production. MoVE offers a systematic approach to identify metabolic valves using metabolic models, providing insight into the architecture of metabolic networks and accelerating the widespread implementation of dynamic flux redirection in diverse systems.

Essentiality of local topology and regulation in kinetic metabolic modeling

“…A topological orthogonality principles was successfully used to designed the bioengineering strains with minimal interaction between desired product-associated pathways and metabolic components related to biomass synthesis (Pandit et al, 2017). In addition, several topological constraints, such as removing the thermodynamically infeasible loops, decoupling two desired phenotypes, have been applied to the CBM methods to improve their metabolic production in recent studies (Venayak et al, 2018; Wang et al, 2018). All those observed metabolic topological properties can be concluded as: the network function is mainly determined by its topology (Ahnert & Fink, 2016), the fundamental principle of complex network theory.…”

“…These observations reflect the fact that current CBMs have inadequate constraints in modeling in vivo metabolism and cannot fully utilize the omics information defining cellular state in question. In particular, recent studies revealed the coordinated regulation of topologically highly-coupled reactions in central metabolism influences empirical flux distribution (Pandit et al, 2017; Venayak et al, 2018), and some metabolites also play important roles in the metabolic flux regulation (Hackett et al, 2016; Kochanowski et al, 2017; Park et al, 2016). However, neither of these constraints has been exploited in current CBM methods.…”

“…We previously found the highly-coupled reactions of central metabolism (the core of bowtie) involve a large number of precursors and cofactors, with fast interconversion among corresponding metabolites and complex interactions between their enzymes and regulators (Li et al, 2018). These characteristics all contribute to buffering the concentration of essential precursors, quick responses to diverse intra/extra-cellular perturbations, so as to maintain global metabolic optimum under various growth conditions (Braakman & Smith, 2012; Hackett et al, 2016; Kochanowski et al, 2017; Li et al, 2018; McCloskey et al, 2018; Okada & Mochizuki, 2017; Park et al, 2016; Venayak et al, 2018). For example, the dynamic trade-off between TCA cycle and glycolysis is regulated by both the ADP/ATP ratio and the concentrations of pyruvate, the starting metabolite of energy metabolism (Berg, 2007a; Berg, 2007b); the reaction fluxes in pentose phosphate pathway (PPP) are primarily determined by the ratio of NAPD/NADPH and the ribose 5-phosphate, the main precursor for the synthesis of nucleotides (Berg JM, 2002).…”

Novel two-stage processes for optimal chemical production in microbes

“…Recent advances in CRISPR 44 , transcriptional switches 45 , riboswitches 46 , and other gene regulatory elements present an exciting outlook for the experimental implementation of such intervention strategies. There has also been interest in computational algorithms that predict dynamic control strategies which begin with high growth and switch over to growth-coupled production as required by the best TS production strategies predicted in this study 47 .…”

Two-stage Dynamic Deregulation of Metabolism Improves Process Robustness & Scalability in EngineeredE. coli

“…[18][19][20][21][22][23][24][25][26][27] Two-stage dynamic control, which decouples growth from production, offers additional potential benefits in large scale bioprocesses, enabling metabolic states to be pushed beyond the boundaries required for growth. 8,21,22,24,28 We have recently implemented two-stage dynamic control in the commonly used and well characterized microbe, E. coli , by utilizing synthetic metabolic valves, which rely on the combination of proteolysis and gene silencing to dynamically reduce levels of key enzymes in the context of a standardized two-stage phosphate depleted process (as illustrated in Figure 1) 17,21,22,29 Proteolysis is accomplished by appending C-terminal degron (DAS+4) tags to a given gene and silencing via expression of the native E. coli CRISPR CASCADE system as well as silencing gRNAs (from pCASCADE plasmids). 30,31 Importantly, in these studies where we produced the organic acid citramalate 22 and the sugar alcohol xylitol, 21 we identified that dynamic deregulation of metabolism was a fundamental mechanism to improve stationary phase metabolic fluxes.…”