Figure 5. Phosphorylation of S98 and S33 has opposing effects on MRTF-A nuclear accumulation.

(A,B) Cells expressing MRTF-A derivatives, with or without activated MEK or RafER, were treated as indicated, and their subcellular localisation scored by immunofluorescence. (A) ERK-mediated S98 phosphorylation controls MRTF-A(2–204)PK localisation. Data are mean ± SEM, n = 3. Mutant S98D baseline higher than wildtype or S98A (p<0.01); MEK and TPA potentiate wildtype (p<0.001); 98A increased by MEK (p<0.01; possibly reflecting chronic ERK stimulation); others not significant. Two-way ANOVA with Bonferroni post-test. (B) Alanine substitutions or 'phosphomimetic' aspartate substitutions at S33 and S98 have opposing effects on MRTF-A(2–204)PK subcellular localisation. Data are mean ± half-range, n = 2. Mutant S98D baseline higher than wildtype or S98A (p<0.01); S98D and S33A cooperate to elevate baseline (p<0.001); 4OHT significantly induces wildtype and S33A (both p<0.05); two-way ANOVA with Bonferroni post-test. (C) N-terminal phosphosite mutations affect nucleocytoplasmic shuttling of intact MRTF-A. Data are mean ± half-range, n = 2. TPA treatment significantly increases nuclear localisation of wild-type MRTF-A (p<0.001), 33A (p<0.001) and 98D (p<0.05); one-way ANOVA, Bonferroni multiple comparison test. (D) Schematic representation of the MRTF-A / RevΔGFP derivatives. (E–G) analysis of MRTF-A NES activity by RevΔGFP assay. (E) MRTF-A(2–204) contains a Crm1-dependent NES. Data are mean ± half-range, n = 2. (F) MRTF-A residues 2–67, but not the RPEL domain, function as an NES. Data are mean ± SEM, n = 3. The 2–67 and 2–115 derivatives differ significantly from each other and from 67–115 and 67–204 in resting cells (all p<0.001 except 2–67 vs 67–204, p<0.01), but not following LMB treatment; two-way ANOVA with Bonferroni post-test. (G) Alanine or 'phosphomimetic' aspartate substitutions at S33, but not S98 affect MRTF-A N-terminal NES activity. Data are mean ± half-range, n = 2; absence of error bar shows a single datapoint (ie mean of two technical replicates).

DOI: http://dx.doi.org/10.7554/eLife.15460.013

Figure 5—figure supplement 1. Use of RevΔGFP to detect NES activity in MRTF-A.

(A) The RevΔ nuclear export assay. Left, representative fluorescenbce images of RevΔGFP with or without re-insertion of the Rev NES and treatment with 50 nM LMB for 120 min. Right, serum does not induce relocalisation of RevΔGFP. N, predominantly nuclear; N/C, pancellular; C, predominantly cytoplasmic. A representative experiment is shown. (B, C) NES activity in the MRTF N-terminal sequences. N, predominantly nuclear; N/C, pancellular; C, predominantly cytoplasmic. Data are from two independent experiments. Error bars are half-range; absence of error bar shows a single point (i.e. mean of two technical replicates). (B) G-actin binding is required for the NES activity of the MRTF-A N-terminal sequences, but the RPEL domain does not exhibit NES activity. Each MRTF derivative was tested with weak (B3A) or strong (B2A) NLS mutations (Pawłowski et al., 2010) and scored as in (A). The MRTF-A RPEL domain (67–204) has no detectable NES activity even when the NLS is inactivated. Inclusion of N-terminal sequences reveals NES activity which is abolished by disruption of the G-actin-binding sites (123-1A mutant: alanine substituted at each RPEL motif). See also Figure 5. (C) Effect of alanine or 'phosphomimetic' aspartate substitutions at S33 and S98 on NES activity of RevΔMRTF-A(2–204) B2A, analysed as in (A).

DOI: http://dx.doi.org/10.7554/eLife.15460.014