Cells expressing either yellow or cyan fluorescent protein in response to pheromone treatment (a bright-field image is also shown). Images courtesy of Alejandro Colman-Lerner, The Molecular Sciences Institute, Berkeley, California, USA.

The activation of signalling cascades in genetically identical cells often leads to variable responses, raising the question about the origin of that variation. Colman-Lerner, Gordon, Brent and colleagues now report their studies of cell-to-cell variation in the output of a cell-fate decision system — the pheromone-response pathway in budding yeast.

The pheromone-response pathway, which includes a mitogen-activated protein kinase (MAPK) cascade, is induced by the α-factor pheromone. Pheromone induction triggers a cell-fate decision by switching normal, vegetative growth to the initiation of mating events, including the induction of gene expression and cell-cycle arrest. Colman-Lerner et al. measured the pheromone-induced expression of fluorescent-protein reporter genes as the readout of this cell-fate decision system — the output combines differences in the signal-transduction pathway and in gene expression from the reporters. To distinguish between these contributions, the authors generated a series of yeast strains containing the genes for yellow and cyan fluorescent protein (YFP and CFP, respectively). The results of experiments in which YFP and CFP were controlled by identical α-factor-responsive promoters were compared to those of experiments in which YFP was driven by an α-factor-responsive promoter and CFP by an α-factor-independent promoter.

This experimental set-up was accompanied by an analytical framework in which the system could be divided into two subsystems, 'pathway' and 'expression', and in each subsystem the authors distinguished between differences in 'capacity' and stochastic fluctuations in the workings of each subsystem ('noise'). This approach enabled the authors to discriminate among the contributions of four sources of variation — differences in pathway power or the ability to transmit a signal; differences in expression capacity or the ability to express proteins from genes; and noise in the two subsystems.

Colman-Lerner et al. showed that most of the cell-to-cell variation in the activation of gene expression in response to pheromone is caused by pre-existing differences among the cells and not noise, and that a main source of variation arises from differences in expression capacity. The cells also differed in the strength with which they carry the signal from the plasma membrane into the nucleus: at high pheromone concentrations most of the variation was due to cell-to-cell differences in their expression capacity; at low pheromone concentrations, the relative contribution of pathway strength was much higher — indicating that high levels of input might conceal pre-existing differences in pathway capacity.

These findings posed new questions about the regulation of cellular functions by signal-transduction cascades. For example, how can pheromone-induced gene expression be accurately regulated if cells vary so widely in their expression capacity? The authors show evidence suggesting that yeast cells adjust the strength of the mating cascade to compensate for the variability in expression capacity, allowing the amount of protein made to reflect the activity of the pathway more accurately.

Furthermore, Colman-Lerner et al. showed that the two MAPKs that transduce the signal, Fus3 and Kss1, function to regulate cell-to-cell variation in pathway capacity. Fus3 suppressed variation when cells are stimulated with high pheromone levels, whereas Kss1 enhanced variation at low pheromone levels.

In conclusion, the identification of the mechanisms that control the quantitative behaviour of a cell-fate decision system, and the first attempts to characterize the functional components that regulate cell-to-cell variation in its performance, form a crucial first step to deepen the understanding of biological systems.