Dynamic Boolean modelling reveals the influence of energy supply on bacterial efflux pump expression

Abstract

Antimicrobial resistance (AMR) is a global health issue. One key factor contributing to AMR is the ability of bacteria to export drugs through efflux pumps, which relies on the ATP-dependent expression and interaction of several controlling genes. Recent studies have shown that significant cell-to-cell ATP variability exists within clonal bacterial populations, but the contribution of intrinsic cell-to-cell ATP heterogeneity is generally overlooked in understanding efflux pumps. Here, we consider how ATP variability influences gene regulatory networks controlling expression of efflux pump genes in two bacterial species. We develop and apply a generalizable Boolean modelling framework, developed to incorporate the dependence of gene expression dynamics on available cellular energy supply. Theoretical results show that differences in energy availability can cause pronounced downstream heterogeneity in efflux gene expression. Cells with higher energy availability have a superior response to stressors. Furthermore, in the absence of stress, model bacteria develop heterogeneous pulses of efflux pump gene expression which contribute to a sustained sub-population of cells with increased efflux expression activity, potentially conferring a continuous pool of intrinsically resistant bacteria. This modelling approach thus reveals an important source of heterogeneity in cell responses to antimicrobials and sheds light on potentially targetable aspects of efflux pump-related antimicrobial resistance.

Efflux, Signaling and Warfare in a Polymicrobial World

“…Small proteins governing the expression of membrane channels (including components of efflux pumps) are also key here. More recently, a study using dynamic Boolean modeling has suggested that ATP variability and energy supply may control efflux pumps, with model bacteria developing heterogeneous pulses of efflux pump gene expression, which contribute to a sustained sub-population with increased efflux expression activity [ 71 ]. The regulation of these systems and their components is discussed in greater detail below and summarized Table 1 .…”

Unravelling microbial efflux through mathematical modelling

“…The previously mentioned Boolean model in [22] takes a gene regulation network and converts it into a mathematical network (sometimes referred to as a graph) where each node represents a regulator or a downstream target and is either 'on' or 'off ' . If an activator is on then it switches a downstream target on with an assigned probability.…”

“…It is frequently when models and data disagree that we learn the most. These pulses are also uncovered theoretically in [22] using a subtly different mathematical approach (Boolean modelling) where probabilities are assigned to the regulatory interactions of the gene regulation network. This approach and the transcriptional pulses are discussed further in the next section.…”

“…When aggregating the single-cell simulations in [22], the results show a homogeneous response (across the population) of efflux being switched on in the presence of a stressor (antibiotic). In the absence of stress, subpopulations of cells have efflux switched on, suggesting that a subset of the population are capable of responding quickly to stress.…”

Fine-grained modelling of ATP dependence of decision-making capacity in genetic regulatory networks

“…This mirrors the ATP dependence of other biomolecular pathways (Gawthrop and Crampin 2017; 2014; Qian and Beard 2006) – for example, the detailed role of ATP in shaping the dynamics of signalling cascades has recently been illustrated using quantitative modelling grounded in systems biology and thermodynamics (Forrest et al 2023). In GRNs, theoretical studies have shown that simplified GRN models (Karlebach and Shamir 2008), using coarse-grained descriptions of genes and gene expression processes, exhibit strong energy-dependent diversity in decision-making capacity (Johnston et al 2012; Kerr, Jabbari, and Johnston 2019; Kerr et al 2022). In such models of simple mutually repressing GRN motifs, increased ATP concentrations led to more attractor states (that is, distinct stable states of gene expression to which other states evolve over time), and hence to more decision-making capacity in cells.…”