Feedback control of ERBB1 and PKC signalling to ERK: does distributive activation cause temporal gating?

Rebecca Perrett 1 Robert Fowkes 2 Christopher Caunt 3 Krasimira Tsaneva-Atanasova 4 Craig McArdle 1
1Department of Clinical Sciences, University of Bristol, Bristol
2Endocrine Signalling Group, Royal Veterinary College, London
3Department of Biology and Biochemistry, University of Bath, Bath
4Department of Engineering Mathematics, University of Bristol, Bristol

Many extracellular signals act via the Raf-MEK-ERK cascade in which signal amplitude, kinetics, compartmentalisation and cell-cell variability can all influence cell fate. The system is subject to negative regulation, notably by ERK-mediated feedback. We have used automated fluorescence microscopy to explore consequences of ERK-mediated feedback in HeLa cells acutely stimulated with EGF (ErbB1 activation) or phorbol 12,13 dibutyrate (PDBu, PKC activation). Using siRNA to knock down endogenous ERK and recombinant adenovirus to add back either wild-type (WT) ERK2 (as a GFP fusion) or catalytically inactive K52R ERK2, we found that ERK-mediated feedback reduces both average ppERK levels and cell-cell heterogeneity in ppERK levels in un-stimulated cells. Both stimuli caused concentration- and time-dependent increases in ERK phosphorylation and nuclear translocation, and ERK-driven transcription. The phosphorylation responses were transient, and frequency distribution plots revealed graded (rather than all-or-nothing) ppERK responses, with indistinguishable Hill coefficients for EGF- and PDBu-stimulated ERK phosphorylation at 5 min and 4 hr. Thus, we found little evidence for the anticipated feedback effects of ERK on response amplitude, kinetics, variability or input-output relationships in stimulated cells. We also binned cells according to ERK2-GFP expression and observed slower ppERK responses at higher ERK2-GFP levels. Remarkably, ERK2-ppERK2 input-output relationships were bell-shaped at early time points (2-5 min) with maximal ERK activation occurring at submaximal ERK2-GFP levels. Mathematical modelling predicted this as consequence of distributive activation rather than (pseudo)processive activation. It also predicted occurrence of the switch without negative feedback, which we confirmed experimentally. Thus, in this model ERK-mediated negative feedback plays a major role in shaping system parameters (amplitude and variance) under equilibrium conditions but has less effect in the non-equilibrium condition of acute stimulation. Under these conditions there is a rapid switch in ERK-ppERK input-output behaviour that is indicative of distributive activation and could provide a novel temporal gate on ERK activation.









Powered by Eventact EMS