Expression variation and covariation impair analog and enable binary signaling control

Due to noise in the synthesis and degradation of proteins, the concentrations of individual vertebrate signaling proteins were estimated to vary with a coefficient of variation (CV) of approximately 25% between cells. This high variation enables population-level regulation of cell functions but abolishes accurate single-cell signal transmission. Here we measure cell-to-cell variability of relative protein abundance using quantitative proteomics of individual Xenopus laevis eggs and cultured human cells and show that variation is typically much lower, in the range of 5-15%, compatible with accurate single-cell transmission. Furthermore, we show that MEK and ERK expression covary which improves controllability of the fraction of cells that activate bimodal ERK signaling, arguing that covariation has a role in facilitating population-level control of binary cell-fate decisions. Together, our experimental and model data argues for a control principle whereby low covariation limits signaling noise for accurate control analog single-cell signaling. In contrast, increased covariation widens the stimulus-range over which external inputs can regulate binary cell activation, thereby enabling accurate control of the fraction of activated cells at the population level.