Abstract
Many biochemical and industrial applications involve complicated networks of simultaneously occurring chemical reactions. Under the assumption of mass action kinetics, the dynamics of these chemical reaction networks are governed by systems of polynomial ordinary differential equations. The steady states of these mass action systems have been analyzed via a variety of techniques, including stoichiometric network analysis, deficiency theory, and algebraic techniques (e.g., Gröbner bases). In this paper, we present a novel method for characterizing the steady states of mass action systems. Our method explicitly links a network’s capacity to permit a particular class of steady states, called toric steady states, to topological properties of a generalized network called a translated chemical reaction network. These networks share their reaction vectors with their source network but are permitted to have different complex stoichiometries and different network topologies. We apply the results to examples drawn from the biochemical literature.
Similar content being viewed by others
References
Angeli, D., & Sontag, E. (2008). Translation-invariant monotone systems, and a global convergence result for enzymatic futile cycles. Nonlinear Anal., Real World Appl., 9, 128–140.
Angeli, D., Leenheer, P., & Sontag, E. (2007). A Petri net approach to the study of persistence in chemical reaction networks. Math. Biosci., 210(2), 598–618.
Clarke, B. L. (1980). Stability of complex reaction networks. Adv. Chem. Phys., 43, 1–215.
Conradi, C., Flockerzi, D., & Raisch, J. (2008). Multistationarity in the activation of a MAPK: parametrizing the relevant region in parameter space. Math. Biosci., 211, 105–131.
Cox, D., Little, J., & O’Shea, D. (2007). Undergraduate texts in mathematics. Ideals, varieties and algorithms (3rd ed.). Berlin: Springer.
Craciun, G., & Feinberg, M. (2005). Multiple equilibria in complex chemical reaction networks: I. The injectivity property. SIAM J. Appl. Math., 65(5), 1526–1546.
Craciun, G., & Feinberg, M. (2006). Multiple equilibria in complex chemical reaction networks: II. The species-reaction graph. SIAM J. Appl. Math., 66(4), 1321–1338.
Craciun, G., Dickenstein, A., Shiu, A., & Sturmfels, B. (2009). Toric dynamical systems. J. Symb. Comput., 44(11), 1551–1565.
Deng, J., Feinberg, M., Jones, C., & Nachman, A. (2011). On the steady states of weakly reversible chemical reaction networks. Preprint available on the arXiv:1111.2386.
Dickenstein, A., & Pérez Millán, M. (2011). How far is complex balancing from detailed balancing? Bull. Math. Biol., 73, 811–828.
Érdi, P., & Tóth, J. (1989). Mathematical models of chemical reactions. Princeton: Princeton University Press.
Feinberg, M. (1972). Complex balancing in general kinetic systems. Arch. Ration. Mech. Anal., 49, 187–194.
Feinberg, M. (1979). Lectures on chemical reaction networks. Unpublished written versions of lectures given at the Mathematics Research Center, University of Wisconsin. Available at http://www.chbmeng.ohio-state.edu/~feinberg/LecturesOnReactionNetworks/.
Feinberg, M. (1987). Chemical reaction network structure and the stability of complex isothermal reactors: I. The deficiency zero and deficiency one theorems. Chem. Eng. Sci., 42(10), 2229–2268.
Feinberg, M. (1988). Chemical reaction network structure and the stability of complex isothermal reactors: II. Multiple steady states for networks of deficiency one. Chem. Eng. Sci., 43(1), 1–25.
Feinberg, M. (1989). Necessary and sufficient conditions for detailed balancing in mass action systems of arbitrary complexity. Chem. Eng. Sci., 44(9), 1819–1827.
Feinberg, M. (1995a). The existence and uniqueness of steady states for a class of chemical reaction networks. Arch. Ration. Mech. Anal., 132, 311–370.
Feinberg, M. (1995b). Multiple steady states for chemical reaction networks of deficiency one. Arch. Ration. Mech. Anal., 132, 371–406.
Flockerzi, D., & Conradi, C. (2008). Subnetwork analysis for multistationarity in mass-action kinetics. J. Phys. Conf. Ser., 138(1).
Gatermann, K. (2001). Counting stable solutions of sparse polynomial systems in chemistry. In E. L. Green, S. Hosten, R. C. Laubenbacher, & V. A. Powers (Eds.), Contemporary math: Vol. 286. Symbolic computation: solving equations in algebra, geometry and engineering (pp. 53–69).
Gatermann, K., & Huber, B. (2002). A family of sparse polynomial systems arising in chemical reaction systems. J. Symb. Comput., 33(3), 275–305.
Gatermann, K., & Wolfrum, M. (2005). Bernstein’s second theorem and Viro’s method for sparse polynomial systems in chemistry. Adv. Appl. Math., 34(2), 252–294.
Gunawardena, J. (2005). Multisite protein phosphorylation makes a good threshold but can be a poor switch. Proc. Natl. Acad. Sci. USA, 102, 14617–14622.
Gunawardena, J. (2007). Distributivity and processivity in multisite phosphorylation can be distinguished through steady-state invariants. Biophys. J., 93, 3828–3834.
Holstein, K., Flockerzi, D., & Conradi, C. (2013). Multistationarity in sequentially distributed multisite phosphorylation networks. Bull. Math. Biol., 75, 2028–2058.
Horn, F. (1972). Necessary and sufficient conditions for complex balancing in chemical kinetics. Arch. Ration. Mech. Anal., 49, 172–186.
Horn, F., & Jackson, R. (1972). General mass action kinetics. Arch. Ration. Mech. Anal., 47, 187–194.
Johnston, M. D. (2011). Topics in chemical reaction network theory. PhD thesis, University of Waterloo.
Manrai, A., & Gunawardena, J. (2009). The geometry of multisite phosphorylation. Biophys. J., 95, 5533–5543.
Markevich, N. I., Hoek, J. B., & Kholodenko, B. N. (2004). Signaling switches and bistability arising from multisite phosphorylation in protein kinase cascades. J. Cell Biol., 164(3), 353–359.
Müller, S., & Regensburger, G. (2012). Generalized mass action systems: complex balancing equilibria and sign vectors of the stoichiometric and kinetic-order subspaces. SIAM J. Appl. Math., 72(6), 1926–1947.
Müller, S., Feliu, E., Regensburger, G., Conradi, C., Shiu, A., & Dickenstein, A. (2013). Sign conditions for injectivity of generalized polynomial maps with applications to chemical reaction networks and real algebraic geometry. Preprint available on the arXiv:1311.5492.
Pérez Millán, M., Dickenstein, A., Shiu, A., & Conradi, C. (2012). Chemical reaction systems with toric steady states. Bull. Math. Biol., 74(5), 1027–1065.
Savageau, M. A. (1969). Biochemical systems analysis II. The steady-state solutions for an n-pool system using a power-law approximation. J. Theor. Biol., 25, 370–379.
Shinar, G., & Feinberg, M. (2010). Structural sources of robustness in biochemical reaction networks. Science, 327(5971), 1389–1391.
Shinar, G., & Feinberg, M. (2012). Concordant chemical reaction networks. Math. Biosci., 240(2), 92–113.
Shiu, A. J. (2010). Algebraic methods for biochemical reaction network theory. PhD thesis, University of California, Berkeley.
Stanley, R. (1999). Enumerative combinatorics (Vol. 2). Cambridge: Cambridge University Press.
Vol’pert, A. I., & Hudjaev, S. I. (1985). Analysis in classes of discontinuous functions and equations of mathematical physics. Dordrecht: Martinus Nijhoff.
Wang, L., & Sontag, E. (2008). On the number of steady states in a multiple futile cycle. J. Math. Biol., 57(1), 25–52.
Wilhelm, T., & Heinrich, R. (1995). Smallest chemical reaction system with Hopf bifurcations. J. Math. Chem., 17(1), 1–14.
Wilhelm, T., & Heinrich, R. (1996). Mathematical analysis of the smallest chemical reaction system with Hopf bifurcation. J. Math. Chem., 19(2), 111–130.
Acknowledgements
The author is supported by NSF grant DMS-1009275 and NIH grant R01-GM086881. The author is also grateful for the numerous constructive conversations with Anne Shiu, Carsten Conradi, Casian Pantea, Stefan Müller, and others, over email and at the AIM workshop “Mathematical problems arising from biochemical reaction networks,” which pointed him toward the strong connection between toric steady states and complex balancing in generalized mass action systems. The author also thanks the two anonymous referees whose suggestions have significantly improved the paper.
Author information
Authors and Affiliations
Corresponding author
Electronic Supplementary Material
Below is the link to the electronic supplementary material.
Appendices
Appendix A: Deficiency Result
Lemma 5
The deficiency δ=n−ℓ−s of a chemical reaction network \(\mathcal{N}\), where n is the number of stoichiometrically distinct complexes, ℓ is the number of linkage classes, and \(s =\operatorname{dim}(S)\), also satisfies \(\delta= \operatorname{dim}(\operatorname{ker}(Y) \cap \operatorname{Im}(I_{a}))\).
Proof
It follows from basic dimensional considerations that
From the rank–nullity theorem we have
The rank of I a corresponds to the number of complexes minus the number of linkage classes, so that \(\operatorname{dim}(\operatorname{Im}(I_{a})) = n- \ell\). It follows that
It follows that
and we are done. □
Appendix B: Kernel of A k
In this appendix, we present a more detailed characterization of ker(A k ) for a mass action system \(\mathcal{M} = (\mathcal{S},\mathcal {C},\mathcal{R},k)\).
Consider a weakly reversible chemical reaction network \(\mathcal{N} =(\mathcal{S},\mathcal{C},\mathcal{R})\) and let \(\mathcal{L}_{k}\), k=1,…,ℓ, denote the network’s linkage classes. Define a subgraph \(\mathcal{T} \subseteq\mathcal{R}\) to be a spanning i-tree if \(\mathcal{T}\) spans all of the complexes in some linkage class \(\mathcal{L}_{k}\), contains no cycles, and has the unique sink \(i \in\mathcal{C}\). Let \(\mathcal{T}_{i}\) denote the set of all spanning i-trees for i=1,…,n. We define the following network constants.
Definition 13
Consider a weakly reversible chemical reaction network \(\mathcal{N} = (\mathcal{S},\mathcal{C},\mathcal{R})\) with reaction weights k j , j=1,…,r. Then the tree constant for i=1,…,n is given by
Remark 12
To compute the tree constants K i , we restrict ourselves to the linkage class containing the complex \(i \in\mathcal {C}\). We then determine all of the spanning trees which contain this complex as the unique sink, multiply across all the weighted edges in each tree, and then sum over all such trees. The terms K i can also be computed by computing specific minors of the kinetic matrix A k restricted to the support of the linkage classes (Proposition 3, Craciun et al. 2009). Note that the term “tree constant” is our own.
The following result characterizes \(\operatorname{ker}(A_{k})\) in terms of the tree constants (31). This result appears in various forms within the chemical reaction network literature. A basic form, just concerned with the signs of the individual components, can be found in Feinberg (1979) (Proposition 4.1) and Gatermann and Huber (2002) (Theorem 3.1). A more specific result can be obtained by the Matrix-Tree Theorem (Stanley 1999). This form is explicitly connected with the reaction graph of a chemical reaction network in Craciun et al. (2009) (Corollary 4). A direct argument is also contained in Sect. 3.4 of Johnston (2011). We defer to these references for the proof.
Theorem 6
Let \(\mathcal{N} = (\mathcal{S},\mathcal{C},\mathcal{R})\) denote a weakly reversible chemical reaction network. Let K i denote the tree constants (31) corresponding to i=1,…,n. Then
where K j =([K j ]1,[K j ]2,…,[K j ] n ) has entries
Remark 13
This theorem may be extended to networks which are not weakly reversible by considering the terminal strongly linked components of a chemical reaction network. As all the relevant networks considered in this paper are weakly reversible, however, Theorem 6 will suffice for our purposes here.
Appendix C: Proof of Lemma 4
Proof
Consider an improper translation \(\tilde{\mathcal{N}} = (\tilde{\mathcal {S}},\tilde{\mathcal{C}},\tilde{\mathcal{C}}_{K},\tilde{\mathcal{R}})\) of a chemical reaction network \(\mathcal{N} = (\mathcal{S},\mathcal {C},\mathcal{R})\) which is resolvable. Let \(\mathcal{M} = (\mathcal {S},\mathcal{C},\mathcal{R},k)\) be a mass action system corresponding to \(\mathcal{N}\) and \(\tilde{\mathcal{M}} = (\tilde{\mathcal{S}},\tilde {\mathcal{C}},\tilde{\mathcal{C}}_{K},\tilde{\mathcal{R}},\tilde{k})\) be an improperly translated mass action system corresponding to \(\tilde {\mathcal{N}}\) and defined by Definition 11. We will write the steady state condition for \(\mathcal{M}\) as
and the steady state condition for \(\tilde{\mathcal{M}}\) as
Since \(\tilde{\mathcal{N}}\) is improper, the vector \(\tilde{\varPsi }_{K}(\mathbf{x})\) contains a subset of the monomials in Ψ(x) by property 3 of Definition 6. Consequently, to relate (32) and (33), we need to remove explicit dependence on the monomials in Ψ(x) corresponding to complexes not in \(\tilde{\mathcal{C}}_{K}\). We will accomplish this by rewriting the unused monomials in Ψ(x) in terms of the monomial in \(\tilde{\varPsi}(\mathbf{x})\) corresponding to the kinetically relevant complex, and absorbing the required adjustment factor into the matrix I k . (This will produce a state-dependent matrix \(\tilde {I}_{\tilde{k}(\mathbf{x})}\). We will resolve the state-dependency at a later stage.)
Since \(\tilde{\mathcal{N}}\) is resolvable, it follows that it is weakly reversible and \(\tilde{S}_{I} \subseteq\tilde{S}_{K}\). Consequently, by Lemma 3, it follows that, for every \(i \in\mathcal{R}_{I}\), there are constants c ij and pairs \(p_{j}, q_{j} \in\tilde{\mathcal {CR}}\), \(j =1, \ldots, \tilde{s}\), such that
We now introduce
and define
for \(j=1, \ldots, \tilde{n}\). This gives rise to the state dependent kinetic matrix \(\tilde{I}_{\tilde{k}(\mathbf{x})} \in\mathbb {R}_{\geq0}^{\tilde{r} \times\tilde{n}}\) with entries \([\tilde {I}_{\tilde{k}(\mathbf{x})}]_{ij} = \tilde{k}_{i}(\mathbf{x})\) if \(\tilde {\rho}(i)=j\) and \([\tilde{I}_{\tilde{k}(\mathbf{x})}]_{ij} = 0\) otherwise.
We now will prove that \(Y \; I_{a} \; I_{k} \; \varPsi(\mathbf{x}) = \tilde{Y} \; \tilde{I}_{a} \; \tilde{I}_{\tilde{k}(\mathbf{x})} \; \tilde{\varPsi }_{K}(\mathbf{x})\) by showing that, for all \(j=1, \ldots, \tilde{r}\),
where Γ:=Y I a , \(\tilde{\varGamma} := \tilde{Y} \; \tilde {I}_{a}\), R(x):=I k Ψ(x), and \(\tilde {R}(\mathbf{x}) := \tilde{I}_{\tilde{k}(\mathbf{x})} \; \tilde{\varPsi }_{K}(\mathbf{x})\). We first make several observations, listed in the order they will be used:
-
By definition, [Γ]⋅,i =y ρ′(i)−y ρ(i) and \([\tilde{\varGamma}]_{\cdot, j} = \tilde{y}_{\tilde{\rho}'(j)} - \tilde {y}_{\tilde{\rho}(j)}\).
-
For h 1(i)=j, we have \(\tilde{\rho}(j) = h_{3}(\tilde{\rho}(h_{1}(i))) = h_{3}(h_{2}(\rho(i)))=\rho(i)_{K}\) (since \(\tilde{\rho}(h_{1}(i)) = h_{2}(\rho (i))\) by property 2 of Definition 6) so that \([\tilde {\varPsi}_{K}(\mathbf{x})]_{\tilde{\rho}(j)} = \mathbf{x}^{y_{\tilde{\rho }(j)}} = \mathbf{x}^{y_{\rho(i)_{K}}}\).
-
By the constructions (34) and (35), we have \(k'_{i}(\mathbf {x})\:\mathbf{x}^{y_{\rho(i)_{K}}} = k_{i}\:\mathbf{x}^{y_{\rho(i)}} = R_{i}(\mathbf{x})\) for all i=1,…,r.
-
For every i=1,…,r, we have \(\tilde{y}_{\tilde{\rho }'(h_{1}(i))} - \tilde{y}_{\tilde{\rho}(h_{1}(i))} = y_{\rho'(i)} - y_{\rho (i)}\) by property 1 of Definition 6.
It follows that, for every \(j=1, \ldots, \tilde{r}\), we have
It follows that we have
where \(\tilde{A}_{\tilde{k}(\mathbf{x})} := \tilde{I}_{a} \; \tilde{I}_{\tilde{k}(\mathbf{x})} \in\mathbb{R}_{>0}^{\tilde{n} \times \tilde{n}}\) is a state dependent kinetic matrix with positive off-diagonal entries corresponding to the structure of the translation \(\tilde{\mathcal{N}}\) and rates given by (35) and (36).
Consider the reaction graph of the network \((\tilde{\mathcal{S}},\tilde {\mathcal{C}},\tilde{\mathcal{R}}, \tilde{k}(\mathbf{x}))\) with state dependent edge weights \(\tilde{k}(\mathbf{x}) \in\mathbb{R}_{\geq 0}^{\tilde{r}}\) given by (35). In order to remove the state dependence in \(\tilde{A}_{\tilde{k}(\mathbf{x})}\), we consider the system at steady state. Since \(\tilde{\delta} = \operatorname{dim} ( \operatorname{ker}(\tilde{Y}) \cap\operatorname{Im}(\tilde{I}_{a})) = 0\), it follows that
Now let \(\tilde{K}_{j}(\mathbf{x})\), \(j=1, \ldots, \tilde{n}\), denote the state dependent tree constants (31) of the reaction graph of \((\tilde{\mathcal{S}},\tilde{\mathcal{C}},\tilde{\mathcal {R}},\tilde{k}(\mathbf{x}))\). Since \(\tilde{\mathcal{N}}\) is weakly reversible, by Theorem 6, we have that
where \(\tilde{\mathbf{K}}_{j}(\mathbf{x}) = ([\tilde{K}_{j}(\mathbf {x})]_{1},[\tilde{K}_{j}(\mathbf{x})]_{2},\ldots,[\tilde{K}_{j}(\mathbf {x})]_{\tilde{n}})\) has entries
It follows that, if \(i, j \in\tilde{\mathcal{L}}_{k}\) for some \(k=1, \ldots, \tilde{\ell}\), we have
Now consider \(i \in\mathcal{R}_{I}\) and let c ij and p j ,q j , \(j=1, \ldots, \tilde{s}\), denote the values guaranteed by Lemma 3. By (39), for every \(i \in\mathcal{R}_{I}\), we have
It follows by (35) and the assumption that \(\tilde{\mathcal{N}}\) is resolvable that (40) only depends on the state-independent rates \(k'_{i}(\mathbf{x})=k_{i}\), \(i \notin\mathcal{R}_{I}\). It follows that (40) may be written
where the tree constants \(\tilde{K}_{i}\) are determined with respect to the reaction graph of \((\tilde{\mathcal{S}},\tilde{\mathcal{C}},\tilde {\mathcal{R}},\tilde{k})\) with the rate constants given by (35) and (36) for \(k'_{i} = k_{i}\) for \(i \notin\mathcal{R}_{I}\) and \(k'_{i}\) arbitrary for \(i \in\mathcal{R}_{I}\) (since the product (40) does not depend on these rates). We may now substitute (41) into (35) to get
and
Notice that, while the tree constant pairs \(\tilde{K}_{p_{i}}\) and \(\tilde {K}_{q_{i}}\) depend upon a choice for the rate constants \(k'_{i}\) for \(i \in\mathcal{R}_{I}\), their ratios do not so that (42) has been defined consistently.
It follows from (18) that (42) and (43) correspond to the choice of rate constants for the improperly translated mass action system \(\tilde{\mathcal{M}} = (\tilde{\mathcal{S}},\tilde{\mathcal{C}},\tilde{\mathcal{C}}_{K},\tilde {\mathcal{R}},\tilde{k})\) defined by Definition 11. Consequently, from (37) we have that
so that the steady states of the system (3) governing \(\mathcal {M}\) and the steady states of the system (6) governing \(\tilde {\mathcal{M}}\) defined by Definition 11 coincide, and we are done. □
Appendix D: Proof of Theorem 5
Proof
Let \(\mathcal{N} = (\mathcal{S},\mathcal{C},\mathcal{R})\) be a chemical reaction network and \(\tilde{\mathcal{N}} = (\tilde{\mathcal{S}},\tilde{\mathcal{C}},\tilde{\mathcal{C}}_{K},\tilde {\mathcal{R}})\) be a weakly reversible translation of \(\mathcal{N}\) which is either proper or improper and resolvable. Suppose \(\mathcal{M} = (\mathcal{S},\mathcal{C},\mathcal{R},k)\) is a mass action system corresponding to \(\mathcal{N}\). We define the translated mass action system \(\tilde{\mathcal{M}} = (\tilde{\mathcal{S}},\tilde{\mathcal{C}},\tilde {\mathcal{C}}_{K},\tilde{\mathcal{R}},\tilde{k})\) according to Definition 8 if \(\tilde{\mathcal{N}}\) is proper and by Definition 11 if \(\tilde{\mathcal{N}}\) is resolvable and improper.
From either Lemma 2 and Lemma 4 we have that the steady state set of \(\tilde{\mathcal{M}}\) corresponds to the steady state set of \(\mathcal{M}\). Correspondingly, by either (11) or (44) we have that
where \(\tilde{A}_{\tilde{k}}:= \tilde{I}_{a} \; \tilde{I}_{\tilde{k}}\) and \(\tilde{\varPsi}_{K}(\mathbf{x})\) has entries \([\tilde{\varPsi}_{K}(\mathbf {x})]_{j} = \mathbf{x}^{y_{h_{3}(j)}}\) for \(j \in\tilde{\mathcal{CR}}\).
Since \(\tilde{\delta}_{K} = 0\), we may conclude by Proposition 2.20 of (Müller and Regensburger 2012) that the translated mass action system \(\tilde{\mathcal{M}}\) has a complex balanced steady state. That is to say, there is a point \(\mathbf{a} \in\mathbb{R}_{>0}^{m}\) which satisfies
Furthermore, since \(\tilde{\delta} = \operatorname{dim}(\operatorname{ker}(\tilde{Y}) \cap \operatorname{Im}(\tilde{I}_{a}))=0\), we have from (38) that all steady states are complex balanced steady states. It follows from Proposition 2.21 of Müller and Regensburger (2012) that the set of such steady states may be parametrized by
where \(\tilde{S}_{K}\) is given by (10) of Lemma 1. This is sufficient to prove claim 3.
Now consider claims 1 and 2. Since \(\tilde{\mathcal{N}}\) is weakly reversible it follows by Theorem 6 that
where \(\tilde{\mathbf{K}}_{j} = ([\tilde{K}_{j}]_{1},[\tilde{K}_{j}]_{2},\ldots ,[\tilde{K}_{j}]_{\tilde{n}})\) has entries
where \(\tilde{K}_{i}\), \(i =1, \ldots, \tilde{n}\), are the tree constants corresponding to the reaction graph \((\tilde{\mathcal{S}},\tilde {\mathcal{C}},\tilde{\mathcal{R}}, \tilde{k})\).
It follows from (45), (46), and (47) that, for every \(i, j \in\tilde{\mathcal{L}}_{k}\) for some \(k=1, \ldots, \tilde {\ell}\), the steady states \(\mathbf{x} \in\mathbb{R}_{>0}^{m}\) satisfy
Since this set corresponds to the steady states of \(\mathcal{M}\) by either Lemma 2 or Lemma 4, we have shown that \(\mathcal{M}\) has toric steady states generated by binomials of the form required by claim 2. Since the choice of rate constants in the definition of \(\mathcal{M}\) was arbitrary, claim 1 follows and we are done. □
Rights and permissions
About this article
Cite this article
Johnston, M.D. Translated Chemical Reaction Networks. Bull Math Biol 76, 1081–1116 (2014). https://doi.org/10.1007/s11538-014-9947-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11538-014-9947-5