Paradoxical activation of the protein kinase-transcription factor ERK5 by ERK5 kinase inhibitors

Abstract

The dual protein kinase-transcription factor, ERK5, is an emerging drug target in cancer and inflammation, and small-molecule ERK5 kinase inhibitors have been developed. However, selective ERK5 kinase inhibitors fail to recapitulate ERK5 genetic ablation phenotypes, suggesting kinase-independent functions for ERK5. Here we show that ERK5 kinase inhibitors cause paradoxical activation of ERK5 transcriptional activity mediated through its unique C-terminal transcriptional activation domain (TAD). Using the ERK5 kinase inhibitor, Compound 26 (ERK5-IN-1), as a paradigm, we have developed kinase-active, drug-resistant mutants of ERK5. With these mutants, we show that induction of ERK5 transcriptional activity requires direct binding of the inhibitor to the kinase domain. This in turn promotes conformational changes in the kinase domain that result in nuclear translocation of ERK5 and stimulation of gene transcription. This shows that both the ERK5 kinase and TAD must be considered when assessing the role of ERK5 and the effectiveness of anti-ERK5 therapeutics.

Fig. 1. Schematic diagrams of the reagents used in this study.

a Schematic diagram of ERK5 (full length) and ERK5 ΔTAD, which lacks the C-terminal extension. Functional domains with amino acid positions are: cytosolic targeting domain (1–77), kinase domain (48–383), activation-loop TEY motif (219–221), proline rich domain (PR) 1 (434–485) and 2 (578–701), nuclear localisation signal (NLS) (505–539), minimal transactivation domain (TAD) (664–789), and the N-terminal interaction domain (740-816). b Chemical structures of ERK5i: compounds 25, 26 and AX15836. c Schematic representation of the ERK5:MEF2D luciferase assay. d Chemical structure of the MEK5i, BIX02189. Structures were drawn using ChemDraw v16.0.

Fig. 2. ERK5i induce transcriptional activity in the ERK5:MEF2D reporter system independently of kinase activity.

a, b, c and e HEK293 cells were transfected with GAL4-MEF2D, GAL4:LUC and CMV:Renilla, together with either wild-type HA-ERK5 (full length) or HA-ERK5 ΔTAD and EGFP-MEK5D or EGFP (control) as indicated. Four hours post transfection, cells were treated with either DMSO (control) or compound 26 (a), compound 25 (b), AX15836 (c) or BIX02189 (e) at concentrations indicated. Twenty-four post transfection, cells were lysed and firefly luciferase activity was measured and normalised to Renilla. The results are presented as the mean of three independent experiments ± SEM. Source data are provided as a Source Data file. d HEK293 cells were transfected as in (a). Twenty-four hours post transfection cells were lysed and processed as in a. f HEK293 cells were transfected with GAL4-MEF2D, GAL4:LUC and CMV:Renilla together with either wild-type HA-ERK5 (full length) or HA-ERK5 kinase dead and EGFP-MEK5D or EGFP (control) as indicated. Twenty-four hours post transfection cells were lysed and processed as in a. g and h HEK293 cells were transfected as in (f). Four hours post transfection, cells were treated with either DMSO (control) or compound 26 (g) or AX15836 (h) at concentrations indicated. Twenty-four hours post transfection cells were lysed and processed as in (a).

Fig. 3. Generation of compound 26-resistant kinase-active mutants of ERK5.

a, b Compound 25 (carbon atoms in green) bound to the ATP-binding pocket of ERK5 (PDB 4B99). Side-chains of residues proposed by Elkins et al. as key to the selectivity and binding affinity of 25 for ERK5 are shown in ball-and-stick and coloured blue (Ile115), purple (Leu189), gold (Gly199) and coral (Asp143), respectively. Asp200 from the ‘DFG’ motif at the start of the activation-loop (red), and Met140 in the hinge region (grey) are also shown. A transparent molecular surface is drawn over compound 25. Hydrogen bonds between 25 and ERK5 are shown as black dashed lines. The view in (b) is rotated ~90° about a vertical axis compared to the view in (a). Figure prepared using CCP4MG. c Comparison of sequences across the DCLK, ERK and PLK families highlights conservation of Ile115, Leu189 and Gly199 (ERK5 numbering), respectively, as key to inhibition by compounds 25 and 26. Sequences were retrieved from the UniProt database, aligned using Clustal Omega, and the alignment coloured by sequence identity (where darker blue indicates higher conservation) and annotated using Jalview. d Calculated binding energies for compound 26 to wild-type ERK5, and the mutant variants I115V, D143G, L189F and G199C, as derived from molecular dynamics simulations and MM/PBSA calculation (unadjusted for entropic contribution); the reduction in binding energy is statistically significant by one-way ANOVA for binding of compound 26 to mutants I115V (p = 0.0282) and L189F (p = 0.0149) when compared to binding to wild-type ERK5. Source data are provided as a Source Data file. e Estimated log-fold change in Kd derived from calculated binding energies demonstrating notably weaker affinity of compound 26 for all mutant variants of ERK5 compared to wild-type. Source data are provided as a Source Data file. f–i Comparison of compound 26 binding to ERK5 wild-type (grey) and I115V (f, blue), D143G (g, coral), L189F (h, purple) or G199C (i, gold) in silico, derived from molecular dynamics simulations with wild-type and mutated residue shown as sticks to denote mutation site. The D143G mutation causes compound 26 to shift toward the P-loop. The I115V, L189F and G199C mutations cause compound 26 to shift out into a shallower binding position in the active site, resulting in the weaker binding energies observed. Figures were prepared using UCSF Chimera. j HEK293 cells were transfected with FLAG-MEF2D and either wild-type HA-ERK5, kinase dead HA-ERK5, HA-I115V ERK5, HA-D143G ERK5, HA-L189F ERK5 or HA-G199C ERK5, and either EGFP-MEK5D or EGFP. Twenty-four hours post transfection cells were lysed, subjected to SDS-PAGE and immuno-blotted with the antibodies shown. The experiment was repeated three times and a representative image is shown. k HEK293 cells were transfected with GAL4-MEF2D, GAL4:LUC and CMV:Renilla, together with either wild-type HA-ERK5, HA-ERK5 ΔTAD, HA-I115V ERK5, HA-D143G ERK5, HA-L189F ERK5 or HA-G199C ERK5 and either EGFP-MEK5D or EGFP. Twenty-four hours post transfection, cells were lysed and firefly luciferase activity was measured and normalised to Renilla. The results are presented as the mean of at least three independent experiments ± SEM. Source data are provided as a Source Data file.

Fig. 4. Binding of compound 26 and AX15836 to the ERK5 kinase domain is required for paradoxical ERK5:MEF2D transcriptional activation.

a–d and h–k HEK293 cells were transfected with GAL4-MEF2D, GAL4:LUC and CMV:Renilla together with wild-type HA-ERK5 (full length) and either HA-L189F ERK5 (a, h), HA-G199C ERK5 (b, i), HA-I115V ERK5 (c, j) or HA-D143G ERK5 (d, k) and EGFP then 4 h post transfection treated with compound 26 (a–d) or AX15836 (h–k) at the concentrations indicated. Twenty-four hours post transfection, cells were lysed and firefly luciferase activity was measured and normalised to Renilla. The results are presented as the mean of three experiments ± SEM. Source data are provided as a Source Data file. e–g HEK293 cells were transfected with GAL4-MEF2D, GAL4:LUC and CMV:Renilla together with wild-type HA-ERK5 (full length) and either HA-L189F ERK5 (e), HA-G199C ERK5 (f) or HA-I115V ERK5 (g) and EGFP-MEK5D then 4 h post transfection treated with compound 26 at the concentrations indicated. Twenty-four hours post transfection, cells were lysed and processed as in (a–d and h–k).

Fig. 5. Compound 26 and AX15836 promote ERK5 nuclear localisation.

a HEK293 cells were transfected with GST-ERK5 C-term, HA-ERK5 ΔTAD, GST, HA, EGFP-MEK5D or GFP as indicated. Four hours post transfection, cells were treated with compound 26, AX15836 (AX), SCH772984 (SCH) or DMSO as control (0) at the concentrations indicated for 24 h then the cells were lysed. Lysates were incubated with glutathione sepharose to precipitate GST-ERK5 C-term and measure co-pulldown of HA-ERK5 ΔTAD. Pulldowns (PD) were immuno-blotted (IB) for GST-ERK5 C-term using ERK5 (C-terminal) antibodies and HA-ERK5 ΔTAD using HA antibodies. To measure expression, whole-cell extracts (WCE) were blotted using ERK5 (C-terminal), HA and MEK5. Phosphorylation of ERK5 by MEK5D was measured using phospho-ERK5 TEY antibodies. Source data are provided as a Source Data file. b HEK293 cells were transfected with either myc-tag or myc-BRAF^V600E as indicated. Four hours post transfection, cells were treated with 0.1 µM SCH772984 (SCH) or DMSO as control (0). Cells were harvested and immuno-blotted for myc (for expression of myc-BRAF^V600E), phospho-T359 RSK and total RSK. A representative image of three independent experiments is shown. Source data are provided as a Source Data file. c HeLa cells were transfected with wild-type HA-ERK5 and either EGFP or EGFP-MEK5D. Twenty-four hours post transfection, cells were treated with compound 26, AX15836 or DMSO as indicated. Thirty minutes after treatment cells were fixed, permeabilised, blocked and stained with anti-HA then donkey-anti-mouse 568 Alexafluor and DAPI. Images were captured by high-content fluorescence microscopy. d and e HeLa cells were transfected with wild-type HA-ERK5 and either EGFP or EGFP-MEK5D. Four hours post transfection, cells were treated with compound 26 (d), AX15836 (e) or DMSO as indicated. Twenty-four hours post transfection, cells were fixed, permeabilised, blocked and stained with anti-HA then donkey-anti-mouse 568 Alexafluor and DAPI. High-content microscopy was used to determine the levels of nuclear and cytosolic HA-ERK5 staining. Results are presented as the % of transfected cells with nuclear≥cytosolic HA staining of three independent experiments ± SEM. Source data are provided as a Source Data file. f HeLa cells were transfected as in (d) and (e) then treated with compound 26 or AX15836 for the times indicated. Cells were then treated and processed as in (d and e). g HeLa cells were treated with 0.1 µM compound 26, AX15836 or DMSO for 30 min. Cells were harvested and the cytosolic and nuclear fractions were isolated and immuno-blotted for ERK5, MEK1 (cytosolic marker) and lamin A/C (nuclear marker). A representative image of three independent experiments is shown. Source data are provided as a Source Data file. h HeLa cells were transfected with wild-type HA-ERK5, HA-I115V ERK5, HA-D143G ERK5, HA-L189F ERK5 or HA-G199C ERK5 and EGFP. Twenty-four hours post transfection, cells were treated and processed as in (d). i HeLa cells were transfected with wild-type HA-ERK5, HA-I115V ERK5, HA-D143G ERK5, HA-L189F ERK5 or HA-G199C ERK5 and EGFP. Four hours post transfection, cells were treated with compound 26 or DMSO as indicated. Twenty-four post transfection, cells were treated and processed as in (d).

Fig. 6. ERK5i induce expression from the KLF2 promoter and is dependent on ERK5 TAD.

a HEK293 cells were transfected with KLF2:LUC and CMV:Renilla together with either FLAG or FLAG-MEF2D and either HA or HA-ERK5 (full length), and either EGFP (control) or EGFP-MEK5D as indicated. Twenty-four hours post transfection, cells were lysed and firefly luciferase activity was measured and normalised to Renilla. The results are presented as the mean of three independent experiments ± SEM. Source data are provided as a Source Data file. b HEK293 cells were transfected with KLF2:LUC and CMV:Renilla together with FLAG and HA, with either EGFP or EGFP-MEK5D (to activate endogenous ERK5). Four hours post transfection, cells transfected with KLF2:LUC and CMV:Renilla together with FLAG, HA and EGFP were treated with either compound 26, AX15836 or DMSO (control) at the concentrations indicated. Cells transfected with KLF2:LUC and CMV:Renilla together with FLAG, HA and EGFP-MEK5D were treated with DMSO. Twenty-four hours post transfection, cells were lysed and firefly luciferase activity was measured and normalised to Renilla. The results are presented as % KLF2 promoter activity where MEK5D-driven (through endogenous ERK5) KLF2 promoter activity is 100% (mean of three independent experiments ± SEM). Source data are provided as a Source Data file. c HEK293 cells were transfected with KLF2:LUC and CMV:Renilla, together with FLAG-MEF2D and either HA-ERK5 (full length) or HA-ERK5 ΔTAD, and either EGFP (control) or EGFP-MEK5D. Four hours post transfection cells were treated with either 300 nM AX15836 or DMSO (control). Twenty-four hours post transfection cells were lysed and processed as in b.