Understanding the temperature response of biological systems: Part II -- Network-level mechanisms and emergent dynamics

TL;DR

The paper investigates how temperature shapes network-level dynamics in biology, bridging Arrhenius-type single-reaction kinetics with emergent system-wide responses. It develops both deterministic (ODE-based) and stochastic (Markov-chain) network models to show that non-Arrhenius scaling, thermal limits, and temperature compensation arise from network topology and dynamical organization, with mean-first-passage-time frameworks revealing triphasic temperature responses in large networks. Two canonical case studies are analyzed: (i) an embryonic Xenopus laevis cell-cycle oscillator where different activation energies of Cyclin synthesis and degradation generate curved period–temperature relations and potential thermal limits; and (ii) a Goodwin-type circadian oscillator where degradation-dominated timing together with adaptive modification cycles yields robust temperature compensation. The work highlights that universal features such as quadratic Arrhenius behavior near a reference temperature, thermal limits, and compensation mechanisms can be predicted from network architecture and energy-distribution statistics, informing strategies for forecasting responses to warming and for engineering temperature-resilient biological functions.

Abstract

Building on the phenomenological and microscopic models reviewed in Part I, this second part focuses on network-level mechanisms that generate emergent temperature response curves. We review deterministic models in which temperature modulates the kinetics of coupled biochemical reactions, as well as stochastic frameworks, such as Markov chains, that capture more complex multi-step processes. These approaches show how Arrhenius-like temperature dependence at the level of individual reactions is transformed into non-Arrhenius scaling, thermal limits, and temperature compensation at the system level. Together, network-level models provide a mechanistic bridge between empirical temperature response curves and the molecular organization of biological systems, giving us predictive insights into robustness, perturbations, and evolutionary constraints.

Figures (2)

• Figure 1: Temperature response of biological oscillators.A–D: Temperature scaling in the early embryonic cell cycle.(A) Core regulatory architecture of the Xenopus embryonic cell-cycle oscillator: Cyclin B synthesis ($k_s$), Cdk1-dependent degradation ($k_d$), and a bistable activation module governing the switch-like transition between interphase and mitosis. (B) Characteristic sawtooth-like Cyclin accumulation and sharp Cdk1 activation pulses reproduced by the minimal two-ODE model of Yang et al.Yang2013 (adapted from Rombouts et al.rombouts2025mechanistic). (C) Phase plane of the two-variable model showing Cyclin and Cdk1 nullclines, the unstable steady state (US), and the emergent relaxation-type limit cycle. (D) Temperature scaling arises from differing activation energies of synthesis ($k_s$) and degradation ($k_d$): Arrhenius plots for the two rates (left) and the corresponding nonlinear, non-Arrhenius dependence of the model-predicted oscillation period (right). E–H: Temperature compensation in a Goodwin-type circadian oscillator.(E) Minimal transcription–translation feedback architecture underlying temperature-compensated circadian rhythms franccois2012adaptivekidd2015temperature. (F) Example time series of X and Z generated by the Goodwin oscillator, showing robust 24-h rhythms. (G) Phase portrait of the Goodwin system, illustrating the X and Z nullclines, the unstable steady state (US), and the resulting limit cycle. (H) Temperature affects the synthesis rates ($k_X$, $k_Y$, $k_Z$) more strongly than degradation. Because the period is governed primarily by degradation, scaling only the production terms with temperature leaves the oscillation period nearly constant.
• Figure 2: Temperature response of stochastic network models.(A) Schematic representation of a generic stochastic biochemical network, where individual transitions follow Arrhenius temperature dependence. Mean first-passage times can be expressed in terms of spanning trees ($\mathcal{T}$) and spanning forests ($\mathcal{F}$), whose total activation energies determine the cumulants entering the Taylor expansion of $\ln r(T)$voits2025generic. (B) In large networks, the distributions of $E_{\mathcal{T}}$ and $E_{\mathcal{F}}$ become approximately Gaussian, causing all cumulants of order $n\geq 3$ to vanish and yielding a quadratic exponential dependence of the log-rate on inverse temperature. (C) A linear cascade of reversible reactions provides an analytically tractable example illustrating the emergence and breakdown of the quadratic exponential jacobs2025beyond. Near the reference temperature $T^*$, forward-biased transitions dominate and collectively generate the generic quadratic form. At extreme temperatures, however, individual critical cycles control the mean first-passage time. (D) Resulting triphasic temperature response. Around $T^*$, the rate follows the quadratic exponential predicted for large networks, while below $T_+$ and above $T_-$ the rate reverts to Arrhenius behavior with positive and negative effective activation energies, respectively.