Fig. 5.

Intermittent ERK oscillations are less prevalent in more differentiated cells. (A) Schematic of experimental approach to compare dynamical signatures of ERK pulsing in ESC and EpiSCs growing in the same medium. (B,B′) Smoothened single cell traces of KTR signal in EpiSCs (brown, B) and ESCs (green, B′) growing in FAX medium. Pulses are indicated by the maxima (blue dots) and corresponding minima (black dots) that define them. (C,C′) Left: fraction of time that individual EpiSCs (C) or ESCs (C′) spent pulsing (brown/green) or non-pulsing (gray) in FAX alone (top) or upon addition of MEKi (bottom). Right: average time that cells were pulsing (brown/green) or non-pulsing (gray) in the cell population. Error bar indicates s.e.m. (D) Pulse duration distribution for EpiSCs (top, brown, n=402 pulses) and ESCs (bottom, green, n=588 pulses). (E) Interpulse interval (IPI) distribution for EpiSCs (top, brown, n=351 pairs of pulses) and ESCs (bottom, green, n=540 pairs of pulses). Pulse recognition resolution limit (yellow bar) and quartiles (Q) 25, 50 and 75 are indicated in D,E. (F) Joint pulse duration versus silence intervals for successive pairs of pulses in EpiSCs (top, brown) and ESCs (bottom, green). The slope 2 dashed line classifies pairs of pulses into consecutive (above) and non-consecutive (below). The axes range was adjusted to better resolve individual data points, leaving off the scale 45 out of 351 data points for EpiSCs, and 36 out of 540 data points for ESCs. (G) Number of pulse trains as a function of the number of consecutive pulses in the train. Counts have been normalized by the number of traces. Top: EpiSCs data (brown dots), ESCs data (green dots) and heterogeneous population model for EpiSCs data (brown triangles). Bottom: ESCs data (green dots, same as top) and heterogeneous population model for ESCs data (green triangles). Shaded area is the s.d. from 200 independent realizations of the stochastic models. Data in D-G from n=52 cells in FAX alone for each cell type.