Figures and data
Structures of the MEK1DDGRA-ERK2T185V complex.
A Diagram describing the open questions of selectivity and substrate recognition mechanism between MAP2Ks and their substrate MAPKs B Schematic representation of protein constructs. Numbers indicate domain limits and locations of mutants following Uniprot numbering. Kinase core domain for MEK1 highlighted in yellow, ERK2 kinase core domain shown in dark green. The introduced GRA24 KIM-motif on MEK1 is illustrated in pink, followed by the linker region (residues 12 to 67), represented with an orange line. C Representative 2D class averages from the dataset collected on HexAuFoil grids. D Cryo-EM reconstructions of MEK1DDGRA-ERK2T185V complexes coloured according to A (state 1: inactive; state 2: active; state 3: nucleotide free). E Models of MEK1DDGRA-ERK2T185V structures, with the flexible linker indicated by a dashed line. Side chains are displayed on residues of interest and labeled: The A-loop residue Y187 on ERK2 is the phosphate acceptor. D218 and D222 mutant sites render MEK1DDGRA constitutively active. F The LocScale sharpened Coulomb potential maps (grey mesh) showing the differences in the interface regions between the kinases with models. H100 on MEK1 and H61 on ERK2T185V were identified as a novel site of contact.
Interaction sites in the MEK1DDGRA-ERK2T185V complex.
A Beyond the KIM, MEK1 interacts with ERK2 at the hydrophobic pocket formed by the MAPK specific insert via the αG-helix (zoom on the left), a histidine triad coordinating a glutamate between the αG-helix of ERK2 and the loop between the αF-helix and MEK1 A-loop, and a permissive His-His interaction in the N-lobes (zoom on the right). Sharpened Coulomb potential from the State 1 reconstruction is displayed as a grey mesh. B The histidines are conserved in the MEK-ERK branch but not the other MAPK pathways and are conserved in metazoans (Fig S9). C State 2 and 3 aligned on the substrate ERK2. Nucleotide free MEK1 is rotated 5.6° and shifted up by 2 Å to slightly detach from its substrate. D The N-lobe of state 3 has opened 12.5° relative to the C-lobe of active MEK1, releasing nucleotide.
Dynamics of the MEK1-ERK2 complex.
A Composite model of the MEK1DDGRA-ERK2T185V complex (active, state 2) constructed from the cryo-EM model and the AF3 prediction to complete disordered regions (full prediction in Fig. S7). B Top view of the same model. C Interaction sites identified with HDX-MS on MEK1DD (top) and ERK2WT (bottom) plotted on the composite active model. Binding partners colored as previously are shown for orientation. Regions showing protection upon the addition of ERK2WT or MEK1DD, indicative of interaction, are highlighted in blue. Regions showing increased exposure upon binding, indicative of allosteric conformational change, are highlighted in red. D Plots for relevant peptide regions of MEK1DD, showing deuteration uptake over time for MEK1DD alone or in the presence of a 1.5 fold excess ERK2WT. Schematic of MEK1 sequence shown next to the plots for clarity. E From left to right: Dimensionless Kratky plot from SEC-SAXS experiments for MEK1DD-ERK2T185V(dark purple) and MEK1DDGRA-ERK2T185V(purple) complexes overlaid for comparison, the introduction of the GRA KIM yields a more globular particle with reduced flexibility. The second dimensionless Kratky plot shows overlay of MEK1DDGRA-ERK2T185V(purple) and MEK1DDGRA-ERK2T185V-ADP.AlF4-(light purple); the addition of nucleotide and transition state analogue further compacts the complex but increases local flexibility. The third panel shows P(r) distribution for MEK1DD-ERK2T185V (dark purple) and MEK1DDGRA-ERK2T185V(purple) which have similar dmax values. In the fourth panel, the P(r) distribution for MEK1DDGRA-ERK2T185V(purple) and MEK1DDGRA-ERK2T185V-ADP.AlF4-(light purple) are seen, where a lower Dmax for sample containing ADPAlF4- addition indicates complex compaction.
MD simulations show that ERK2 Y187 can reach MEK1 ATP and that the flexibility of the system is consistent with HDX-MS and cryo-EM data.
A Frame extracted from an MD simulation of the fully active state in which Y187 of ERK2WT approaches the γ-phosphate of MEK1DDGRA ATP at a catalytically compatible distance (∼3.8 Å). The inset shows K97 interacting with the ATP α-phosphate as well as with E114 to form the salt bridge that stabilizes the αC-in active state. In the distances over time plot, the frames where ERK2WT Y187 reaches MEK1DDGRA ATP at a catalytically compatible distance are highlighted in blue. B Three frames extracted from an MD simulation of the inactive state of the MEK1DD-ERK2WT at different times where the α-helices of the N-terminal linker of MEK1DD show large rearrangements. C Average root mean square fluctuation (RMSF) for all the MEK1DDGRA-ERK2WT MD simulations (24 replicas, 48 μs). D Distances between MEK1DDGRA H100 and ERK2WT H61 at the N-lobe interface and between the MEK1DDGRA F311 and ERK2WT Y263 at the MAPK-specific insert αG-helix interface. The insets show frames extracted from the corresponding time series.
Molecular details of MAPK pathway specificity.
A Schematic showing the specificity of MAP2Ks for MAPKs and the consequences of signalling, nodes with structural data on the interaction are coloured. B Diagram describing the selectivity checkpoints in the MAP2K-MAPKT185V interactions of the MAP kinase pathways. C Comparison of the MEK1DDGRA-ERK2T185V complex with the MKK6DDGRA-p38α complex. While the overall face-to-face conformation is maintained between pathways significant differences exist in the relative positioning of the kinases (a 36° rotation of MKK6 relative to ERK2 leads to a higher N-lobe interaction) and the residues that interact. From left to right the complexes are displayed as space filling surfaces, the substrate MAPK as a surface and the MAP2K as a ribbon, and the surface of the substrate MAPK with the interface residues highlighted.
