Systems Biology and Metabolic Engineering in Bacteria - Systems Biology of Metabolic and Signaling Networks - page 350

13.4.1 Systems that Require Dynamical Models
These phenomena can only be adequately represented by nonlinear differential
equations because the system behavior would simply be lost in a linear approxima-
tion. Models of such emergent phenomena are needed not only for understanding
system properties but also for constructing novel systems by methods of synthetic
biology. The first synthetic oscillatory circuit constructed and modeled was proba-
bly the “repressilator,” composed of three regulators that repress each other in a
cyclic arrangement (Elowitz and Leibler 1999 ). A more complex synthetic,
oscillatory circuit, the “metabolator” combines genetic and metabolic regulations.
The expression of the enzymes that interconvert acetyl coenzyme A and acetyl
phosphate is regulated by the concentration of the second metabolite, acetyl
phosphate (Fung et al. 2005 ). The nonlinear ODE model of the system correctly
describes the observed oscillatory behavior, predicting properties such as the
frequency of the oscillations as a function of the carbon flux through glycolysis.
The synthetic constructs can be further extended to incorporate cell-cell communi-
cation (Song et al. 2008 ). An artificial predator-prey construct combines two strains
of E. coli . Both strains produce a toxin molecule that kills the cell transcribing the
corresponding gene. The predator strain sends a diffusible molecule to the prey
strain, eliciting there the production of the toxin. The prey strain produces a
different diffusible molecule that activates the production of an antidote in the
predator strain. The predator thus depends on the prey for survival, but the prey will
be killed in the presence of many predators. The observed population dynamics are
oscillations of the number of predators and prey in anti-phase, as predicted by the
dynamical model. The key to success of all these projects was parameter optimiza-
tion of the system components based on a detailed dynamical model of the synthetic
system. The number of system components was still relatively modest, which
permitted the construction of a complete, nonlinear ODE model.
Intrinsically, nonlinear phenomena also govern major aspects of the “natural”
physiology of the cell. Metabolic oscillations have been observed over 50 years ago
and the first models describing these phenomena were developed soon thereafter
(Song et al. 2008 ). Such oscillations concern many microorganisms and
metabolites: for example, amidase activity in Pseudomonas aeruginosa , cAMP in
Dictyostelium discoideum , lactose metabolism in E. coli , and glycolytic oscillations
in yeast. This latter phenomenon has been studied in detail experimentally and
theoretically. Continuous cultures of yeast spontaneously synchronize their cell
physiology. During the 40 min of growth (corresponding to one mass doubling) in
constant environmental conditions, the population traverses a cycle comprising a
reductive and an oxidative phase (Klevecz et al. 2004 ). This metabolic oscillation
affects all major cellular activities, ATP:ADP ratio, transcription, replication, and
cell growth. A recent model proposes that the observed periodicity of cell physiol-
ogy is controlled by ATP-dependent nucleosome remodeling (Thiele et al. 2009 ).
The alternate anabolic and catabolic phases of the yeast metabolism could strongly
affect the efficiency of biotechnological applications in this organism. However,