5.2.1
Metabolic control analysis
Metabolic control analysis (MCA) is a framework for quantifying the control properties of a steady-state metabolic system in terms of the responses of its fluxes and metabolite con- centrations towards perturbations in the rates of its reactions [8,9]. MCA also relates these sensitivities to the properties of the system components through the use of the connectivity and summation properties [8]. Below we briefly describe the fundamental coefficients of MCA and define their relationships.
Elasticity coefficients describe the sensitivities of isolated reactions towards changes in their parameters and the concentrations of their substrates, products, or direct effectors. For a reaction i affected by a metabolite X , the elasticity coefficient is defined as a ratio of relative change of the reaction rate vi to the relative change of the concentration x:
"vi
x =
∂ ln vi
∂ ln x (5.1)
Control coefficients describe the sensitivities of system variables of a metabolic pathway, such as flux or steady-state metabolite concentrations, towards changes in the local rates of
its reactions. These sensitivities, unlike elasticity coefficients, are a function of the complete system and depend on the system topology and component properties and are thus defined using total derivatives. Control coefficients are defined as a ratio of relative change of any system variable y to the relative change of a reaction rate vi:
Ciy = dln y
dln vi (5.2)
The summation property describes the distribution of control between the reactions within a metabolic pathway. This property applies to all metabolic systems and states that the sum of the control coefficients of all reactions on any particular flux is equal to 1. For the two step system with reactions 1 and 2:
S −*)− X1 −*)− P2 (5.3)
the flux summation property is defined as:
C1J+ C2J= 1 (5.4)
The connectivity theorem describes the relationship between control and elasticity co- efficients. The flux connectivity states that in a system where a metabolite affects multiple reactions, the sum of the products of each of the control coefficients of a particular flux to- wards these reactions multiplied by the corresponding elasticity coefficients of these reactions towards the metabolite is equal to 0. For our two step system (Equation5.3) where both re- actions are affected by x, it can be expressed as:
C1J"v1
x + C J
2"xv2= 0 (5.5)
Solving these simultaneous equations produces expressions for each of the two control coefficients CJ
1 and C2J in terms of the elasticity coefficients:
C1J = " v2 x "v2 x − " v1 x and CJ 2 = −" v1 x "v2 x − " v1 x (5.6) The relationship between control and elasticity coefficients expressed above can also be obtained through one of the various matrix-based formalisations of MCA [19–21,56,57,98–
110]. One such method, previously described in [57], combines the summation and connec- tivity properties into a generalised matrix form called the control matrix equation. Here a matrix of independent control coefficients, Ci, and a matrix expressing structural properties and elasticity coefficients, E, have the relationship:
5 4 2 6 S1 1 S2 S3 3 X4 X6 X0 S5 Backbone Multiplier 1 Multiplier 2
Figure 5.1: Control patterns for a 6 step branched pathway. The backbone and multiplier patterns for two control patterns of CJ1
1 are depicted. Green bubbles indicate the backbone pattern, while red and
blue bubbles each indicate a multiplier pattern. Each multiplier-backbone combination represents a single control pattern. The chain of local effects for the backbone originates from the perturbation of reaction 1 which affects the concentration of S1, in turn affecting reaction 2, and so forth, until
reaching reaction 4. An increase in the concentration of the enzyme catalysing reaction 1 will therefore cause a chain of positive effects, each playing a small role in determining the sensitivity of J1towards v1.
This means that Cican be calculated by the inversion of E. In the case of a symbolic inversion of E [23], expressions similar to those shown in Equation5.6are obtained.
These expressions of control also translate into physical concepts. Each term of the nu- merator of a control coefficient expression represents a chain of local effects that radiates from the modulated reaction as demonstrated in Fig.5.1[17]. These chains of local effects, otherwise known as control patterns, describe the different paths through which a perturba- tion can propagate towards the controlled system variable, with each unique path potentially carrying a different weight towards the entire control coefficient. The contribution of a con- trol pattern towards its control coefficient is quantified as its percentage absolute contribution towards the sum of the absolute values of all the control patterns of the control coefficient. This metric is used instead of a conventional percentage to account for cases where control patterns have different signs, which could lead to the contribution of a single pattern being more than 100%.
Control patterns can be further subdivided into smaller backbone and multiplier patterns [118]. In the case of flux control coefficients, a backbone pattern is defined as an uninter- rupted chain of reactions that links two terminal metabolites and passes through the flux being controlled, known as the reference flux. Multiplier patterns are chains of reactions that are not directly linked with a backbone pattern, and occur only in the case of branched pathways. Different multiplier patterns can therefore be combined with a single backbone to form different control patterns. Conversely, the same multipliers can also be found among patterns with different backbones.
5.2.2
Regulation by enzymes
One definition of regulation, as it pertains to metabolic systems, is the alteration of reaction
properties to augment or counteract the mass-action trend in a network of reactions[25]. One
way that the mass-action trend can be counteracted is through the activity of enzymes, with higher potential for regulation being achieved far from equilibrium. Therefore, in order to quantify the regulatory effect of enzymes in a system, it is necessary to be able to determine the distance from equilibrium of a reaction, and to be able to distinguish between the effect of mass action and enzyme binding [16].
Distance from equilibrium is given by the disequilibrium ratio Γ /Keq (or ρ), with kinetic control in the forward direction indicated by ρ ≤ 0.1, thermodynamic control in the forward direction indicated by ρ ≥ 0.9, and a combination of kinetic and thermodynamic control indicated by 0.1 < ρ < 0.9 [16].
Determining the effects of binding and mass action requires a rate equation to be cast in a form which separates these two functions and partially differentiating its logarithmic form with respect to the logarithm of the substrates or products [16,25]. This process yields two elasticity coefficients, one which quantifies the effect of the substrate or product binding on the reaction rate, and the other the effect of mass action. As an example, a Michaelis- Menten rate equation with substrate S and product P is expressed in a way that makes the rate contributions due to rate capacity (vcap), binding (Θ), and mass action (vma) explicit as factors arranged from left to right:
v= Vf Ks |{z} vcap × 1 1 + s Ks + p Kp | {z } Θ × s − p Keq | {z } vma (5.8)
Partial differential of the natural log form of Equation5.8with respect to ln s yields binding ("Θ
s ) and mass-action (" vma
s ) elasticity expressions, which add up to the total elasticity ("sv) [16]: "v s = 0 + − s Ks 1 + s Ks + p Kp + 1 1 − ρ = "Θ s + " vma s (5.9)
Similar expressions can be defined for the product. With these separate elasticity coeffi- cients we can quantify, and differentiate between, the effect of either a substrate or a product on the reaction rate though enzyme binding and the mass-action effect.
5.2.3
Software
Model simulations were performed using the Python Simulator for Cellular Systems (
PySCeS
) [91] within aJupyter
notebook [173]. Symbolic inversion of the E matrix (Equation5.7) and subsequent identification and quantification of control patterns was performed by theSymCa
[23] module of thePySCeSToolbox
[31] add on package forPySCeS
. Control pat- terns for each control coefficient were automatically numbered bySymCa
starting from 001, and we used these assigned numbers as-is in the presented results. Importantly,SymCa
was set to automatically replace zero-value elasticity coefficients with zeros. In other words, cer- tain elasticity coefficients are never found within any control coefficient expressions as their value will always be zero (such as in the case of irreversible reactions).The automatic identification of the Γ /Keq terms of the reactions, subsequent automatic separation of rate equations into the binding and mass-action terms, and calculation of elastic- ity coefficients for these terms was performed by the
ThermoKin
module ofPySCeSToolbox
. Additional manipulation of symbolic expressions, data analysis, and visualisations were per- formed using theSymPy
[175],NumPy
[179] andMatplotlib
[174] libraries for Python [157].5.2.4
Model
As mentioned in the Introduction to this chapter, we revisited results obtained in a previous study of a model of pyruvate branch metabolism in Lactococcus lactis [31] (Chapter4). This model, shown in Fig.5.2, was originally constructed by Hoefnagel et al. [86], and we retrieved it from the JWS online model database [160] in the PySCeS model descriptor language (see
http://pysces.sourceforge.net/docs/userguide.html
).Members of the of ATP/ADP, acetyl-CoA/CoA and NADH/NAD+ moiety-conserved cycles
were treated as ratios, respectively represented by the symbols φA, φC, and φN, in order to perform parameter scans of these conserved moieties without breaking moiety conservation. In this chapter, only the effect of a change in the value of φN is considered. The value of φN was thus fixed and varied between 0.0002 and 1.77 in order to generate Figs. 5.3,5.4,5.8,
5.9and5.10.
The value of φN was varied directly rather than modulating its demand by changing the activity of NADH oxidase for two reasons. Firstly, we wanted follow the same methodology used in our original study [31] (Chapter4), so that we may explain the results obtained there. Secondly, we wanted to simplify the system for control pattern analysis. These considerations are discussed further in Section5.3.6, where we perform a similar control pattern analysis for a range of φN values by varying the Vma x of NADH oxidase in a version of the model where
φN is not fixed. Pyr EtOH Lac Acp Ac
Aclac Acet Acet But 1 12 5 4 Glc 8 13 3 7 11 10 9 14 2 2 2 2 2 (outside) 6 N Acal C A 1. Glycolysis 3. Pyruvate dehydrogenase 2. Lactate dehydrogenase 4. Phosphotransacetylase 5. Acetate kinase 7. Alcohol dehydrogenase 10. Acetoin efflux 8. Acetolactate synthase 6. Acetaldehyde dehydrogenase 9. Acetolactate decarboxylase 12. ATPase 14. Non-enzymatic acetolactate decarboxylation 11. Acetoin dehydrogenase 13. NADH oxidase
Figure 5.2: The pyruvate branch pathway as defined by Hoefnagel et al. [86]. Reactions are num- bered according to the key. The stoichiometry of each reaction is 1 to 1, except for reaction 1 where Glc + 2ADP + 2NAD+→ 2Pyr + 2ATP + 2NADH and reaction 8 where 2Pyr Aclac. Intermediates
are abbreviated as follows: Ac: acetate; Acal: acetaldehyde; Acet: acetoin; Aclac: acetolactate; Acp: acetyl phosphate; Glc: glucose; Lac: lactate; But: 2,3-butanediol; Pyr: pyruvate; EtOH: ethanol.