Publications
* denotes equal contribution
Biochemically plausible models of habituation for single-cell learning.
Eckert L, Vidal-Saez MS, Zhao Z, Garcia-Ojalvo J, Martinez-Corral R*, Gunawardena J*.
Current Biology. 2024. 16;34(24):5646-5658.e3.
Here we developed plausible molecular models of habituation, one of the most basic forms of learning. In animals, it is known that there are 7 hallmarks that characterise habituation to a single stimulus, and we developed models that can account for all of them. Our models are based on differential equations, and represent interactions among proteins that activate and deactivate each other. Molecular species were arranged in motifs known to mediate adaptation, where molecules are linked in such a way that one molecule activates another molecule but this also gets inhibited, either in a feed-forward or a feedback manner. This work clarified some of the discrepancies in the habituation literature, and should enable scientists to experimentally characterise the basis for learning behaviours at the single cell level. This could have multiple applications in medicine, for example to understand why cells respond differently to multiple exposures to cytokines or drugs.
Emergence of activation or repression in transcriptional control under a fixed molecular context.
Martinez-Corral R*, Friedrich D, Frömel R, Velten L, Gunawardena J, DePace A H*.
bioRxiv. 2024
We investigate the basis for transcription factor duality - where a transcription factor behaves as an activator or a repressor - and non-monotonicity - where responses to a transcription factor first increase and then decrease. We show that for transcription factors that act incoherently, simultanoeusly affecting transcription positively and negatively, the affinity of the DNA binding site can tune the response between activation and repression. We also discuss under what conditions non-monotonicity can arise, and show experimental evidence for non-monotonicity and affinity-dependent activation or repression that we interpret in terms of incoherent regulation.
Frömel R, Rühle J, Bernal Martinez A, Szu-Tu C, Pacheco Pastor F, Martinez-Corral R, Velten L.
bioRxiv. 2024
We investigate the function of 38 TFs involved in hematopoiesis through a Massively Parallel Reporter Assay in primary hematopoietic differentiating cells. We reveal a large degree of context dependency, particularly highlighting the emergence of activating or repressive function depending on neighboring TFs, cell type or predicted occupancy of TFs on the regulatory sequence.
Bonnal S, Bajew S, Martinez-Corral R, Irimia M.
bioRxiv. 2024
We propose that different micro-exons have different sensitivies to the SRRM3/4 regulators due to variations in the sequences that determine the interaction with the core splicing regulatory sequences.
The Hill function is the universal Hopfield barrier for sharpness of input-output responses.
Martinez-Corral R, Nam Kee-Myoung, DePace AH, Gunawardena J.
PNAS 2024.121 (22) e2318329121
Biology is characterized by the expenditure of energy. In the processes of transcription and translation, energy can be used to reduce the error of the processes. If energy is not used, under conditions of thermodynamic equilibrium, there is a limit to how well errors can be corrected, which we call Hopfield barrier. In this paper, we investigate the limits imposed by thermodynamic equilibrium in another information processing task, which is to turn a graded input that binds to a target, like a transcription factor that binds to DNA, into a response in a very non-linear (sharp) manner, with a sharp transition between the no-response and response regimes. This kind of behaviour is usually modelled with a Hill function (x^n/1+x^n) , with n, the Hill coefficient, determining the sharpness. This function is convenient, but there is not a biophysical basis for it. In this paper we derive a universal model for ligand binding to n sites at equilibrium, which only depends on a very small number of core assumptions but is independent of the details of the model, therefore allowing to model a large class of potential mechanisms. We show that the sharpness of the input-output function is limited by that of the Hill function of Hill coefficient n, therefore providing a biophysical basis for this widely used mathematical function and a solid basis for ruling out equilibrium models.
Martinez-Corral R*, Park M*, Biette K*, Friedrich D, Scholes C, Khalil A S, Gunawardena J, DePace A H.
Cell Systems 2023.14(4):324-339.e7
Transcriptional synergy, whereby transcription is higher under a combination of transcription factors (TFs) than under each individually, is commonly understood to arise from physical interactions among the TFs and the transcriptional machinery. Here we combine experiments in a cell line-based synthetic platform, and a novel modelling approach, to show that synergy can also emerge if two TFs bind to the same, shared site, without physically interacting while bound. We interpret this as evidence for “kinetic synergy”, which emerges from the functional interactions among the TFs, and find that it depends nontrivially on the TFs’ DNA binding kinetics. These findings are broadly relevant for our understanding of transcriptional control by TF combinations in eukaryotic genomes. This work is a collaboration with experimentalists in the group of Prof. Angela DePace at Harvard Medical School and Prof. Ahmad (Mo) Khalil at Boston University. Watch a summary of this work (starting time 1:04:35)
Allosteric conformational ensembles have unlimited capacity for integrating information.
Biddle B*, Martinez-Corral R*, Wong F, Gunawardena J.
eLife, 2021. 10, e65498.
In this work we theoretically show that allostery can be a biophysical mechanism to generate very flexible higher-order cooperativities (HOCs), which quantify the extent by which ligand binding at one site in a macromolecule is affected by bound ligand at multiple other sites. We used the linear framework to develop a general model for allostery, with arbitrary numbers of conformations and sites, therefore subsuming and generalising previous allostery models. We developed a new method of coarse-graining, which revealed that HOCs can emerge from allostery in a very flexible way. Given the increased awareness of conformational flexibility in cellular regulatory processes, particularly in gene regulation, we suggest that allostery-mediated HOCs might underlie complex input-output behaviour, like the steep sigmoidal response of the classical Drosophila gene hunchback to the transcription factor bicoid. Watch a summary of this work (starting time 54:40).
Metabolic basis of brain-like electrical signalling in bacterial communities.
Martinez-Corral R, Liu J, Prindle A, Süel G M, Garcia-Ojalvo J.
Philosophical Transactions of the Royal Society B. 2019. 374(1774):20180382.
Previous research from the Süel group at University of California San Diego had discovered and characterised population-level oscillations in electrical signalling and metabolism in bacterial communities (biofilms) of Bacillus subtilis grown in microfluidic devices. Here we developed a spatially extended model of a single biofilm, where we investigated the coupling between glutamate metabolism and electrical signalling underlying the generation of the oscillations. Under this model, metabolic stress caused by lack of glutamate in the biofilm center is transmitted to the periphery by potassium-mediated electrical signalling. Potassium and glutamate are two central ions in the nervous system, and this biofilm phenomenon is reminiscent of cortical spreading depression in the brain. These findings have implications for our evolutionary understanding of the nervous system, and more broadly for the role of electrical signalling in non-excitable cell types.
Self-amplifying pulsatile protein dynamics without positive feedback.
Martinez-Corral R, Raimundez E, Lin Y, Elowitz M B, Garcia-Ojalvo J.
Cell Systems. 2018. 453-462.e1.
In this work we theoretically show that stochastic pulses can emerge in circuits of interacting molecules as a result of negative feedback and ultransensitivity. Therefore, a positive feedback interaction is not required, contrary to previous models in the field. This work was motivated by the stochastic pulses in nuclear localisation observed in the yeast transcription factor Msn2, whose regulation does not seem to involve any positive feedback.
Bistable emergence of oscillations in growing Bacillus subtilis biofilms.
Martinez-Corral R, Liu J, Süel G M, Garcia-Ojalvo J.
PNAS. 2018. 115(36): E8333-E8340.
We developed a delay-differential equation model of B. subtilis biofilm oscillations that, in combination with experiments, revealed that the oscillations emerge through what is known as a supercritical Hopf bifurcation. In other words, as the biofilm grows, its growth rate and membrane potential transitions from being constant to being oscillatory in an abrupt manner, and over a range of sizes, the two behaviours can occur (bistability). As a consequence, oscillations can be triggered by a perturbation, like transiently stopping the media flow in the microfluidics system. Moreover, this analysis revealed that the minimum size at which oscillations emerge is not fixed, as initially thought, but is reduced with higher basal stress, explaining the effect of lower glutamate concentrations experimentally observed.
Diversity of fate outcomes in cell pairs under lateral inhibition.
Guisoni N*, Martinez-Corral R*, Garcia-Ojalvo J, de Navascues J.
Development. 2017. 144(7):1177-1186.
In this work we proposed that the outcome of Notch/Delta signalling between pairs of cells can be modulated by the cell-cell contact area. In the Drosophila intestine, stem cells are found interspersed in the tissue. When they divide, each cell in the pair can either differentiate or remain as a stem cell in a Notch-dependent manner. Using a mathematical model, we found that cell pair fates could either be symmetric or asymmetric depending on a parameter that could be related to the contact area. In agreement with this, we found a correlation between cell pair fate and contact area experimentally, suggesting that stem cell fate can be modulated by mechanical and geometrical tissue properties. The work was a collaboration with the experimentalist Dr Joaquín de Navascués (currently at University of Essex).
Coupling between distant biofilms and emergence of nutrient time-sharing.
Liu J, Martinez-Corral R*, Prindle A*, Lee D Y D, Larkin J, Gabalda-Sagarra M, Garcia-Ojalvo J, Süel G M.
Science. 2017. 356(6338):638-642.
We combined modelling and experiments to show that oscillations of neighboring biofilms in the same chamber can become coupled, with oscillation peaks coinciding (in-phase) or alternating (anti-phase) depending on the glutamate concentration of the media. Biofilms switched from in-phase to anti-phase dynamics as the nutrient was reduced, but had counterintuitively higher growth in low than in high glutamate, which we interpreted in terms of a nutrient time-sharing strategy.