Fig. 2. ERK5 and ERK1/2 act differently to disrupt the actin cytoskeleton. Activation of ERK5 does not block the ability of constitutively activated RhoA to stimulate the formation of actin stress fibres. (A) NIH3T3 cells were microinjected, in the presence of serum, with expression vectors for constitutively activated RhoA (RhoAV14) alone or in combination with activated MEK1 (MEK1EE) + ERK2 or activated MEK5 (MEK5DD) + ERK5. 3 hours post-injection cells were serum starved and after 24 hours were fixed and stained with ERK2 and ERK5 antibodies. RhoAV14 expression was recognised with either a mouse or rabbit RhoA antibody and polymerised actin was detected using Texas Red-phalloidin. Arrows indicate injected cells; (B) treatment with 1 µM PD184352 restores the actin cytoskeleton in Ras-transformed NIH3T3 cells (clone 149169). 149169 cells were treated for 24 hours with 1 µM PD184352 or vehicle, permeabilised, fixed and stained for polymerised actin as in A and for vinculin, as a marker of focal adhesions, with a mouse monoclonal antibody followed by a anti-mouse FITC-coupled antibody. Bar, 20 µm.