Conformation Selection by ATP-competitive Inhibitors and Allosteric Communication in ERK2

Abstract

Activation of the extracellular signal regulated kinase-2 (ERK2) by phosphorylation has been shown to involve changes in protein dynamics, as determined by hydrogen-deuterium exchange mass spectrometry (HDX-MS) and NMR relaxation dispersion measurements. These can be described by a global exchange between two conformational states of the active kinase, named "L" and "R", where R is associated with a catalytically productive ATP-binding mode. An ATP-competitive ERK1/2 inhibitor, Vertex-11e, has properties of conformation selection for the R-state, revealing movements of the activation loop that are allosterically coupled to the kinase active site. However, the features of inhibitors important for R-state selection are unknown. Here we survey a panel of ATP-competitive ERK inhibitors using HDX-MS and NMR and identify 14 new molecules with properties of R-state selection. They reveal effects propagated to distal regions in the P+1 and helix αF segments surrounding the activation loop, as well as helix αL16. Crystal structures of inhibitor complexes with ERK2 reveal systematic shifts in the Gly loop and helix αC, mediated by a Tyr-Tyr ring stacking interaction and the conserved Lys-Glu salt bridge. The findings suggest a model for the R-state involving small movements in the N-lobe that promote compactness within the kinase active site and alter mobility surrounding the activation loop. Such properties of conformation selection might be exploited to modulate the protein docking interface used by ERK substrates and effectors.

Keywords: ERK2; MAP kinase; NMR; cancer therapeutics; conformation selection; hydrogen-deuterium exchange; inhibitor.

Figure 1.. BVD523 and VTX11e stabilize the R-state in 2P-ERK2.

(A) Chemical structures of the ATP-competitive ERK1/2 inhibitors, BVD523, VTX11e and GDC0994. (B-D) 2D-HMQC spectra collected at 25°C and 5°C ([ERK2]:[inhibitor] = 1.0:1.2), showing [methyl-¹³C,¹H] peaks of residues (B) I72, (C) L220 and (D) L242, which report R and L conformers. Their locations in the ERK2 structure are shown in Suppl. Fig. S1. 2P-ERK2 complexed with BVD523 and VTX11e (shown in blue) shift to 100% R at all temperatures, while 2P-ERK2 complexed with GDC0994 (shown in grey) retains conformational exchange between R:L populations of 80:20, similar to the apoenzyme. Previous studies show that 0P-ERK2 complexed with all inhibitors retains the L conformer seen in the apoenzyme (Pegram et al., 2019). Full NMR spectra are shown in Suppl. Fig. S2. Titration of 2P-ERK2 with VTX11e and GDC0994 to demonstrate binding saturation is shown in Suppl. Fig. S3.

Figure 2.. Binding of R-state inhibitors reveals allosteric coupling between the active site and the activation loop.

(A-C) Summary of HDX experiments indicating regions that change in deuterium uptake upon binding of (A) BVD523, (B) VTX11e, and (C) GDC0994. (D-F) HDX time courses showing effects of inhibitors on deuterium uptake at the (D) DFG motif (peptide 161–168: LKICDFGL), (E) P+1 segment (peptide 191–198, YRAPEIML), and (F) helix αF (peptide 203–210: YTKSIDIW). Colored segments in panels A-C indicate regions where HDX decreases or increases upon binding each inhibitor at saturating concentration ([ERK2]:[inhibitor] = 1.0:1.2). Full peptide coverage and locations of segments that undergo changes in HDX with inhibitor binding are shown in Suppl. Figs. S4 and S5. Highlighted in light green in panels A-C are regions where a similar degree of HDX protection is seen with all inhibitors in both 0P-ERK2 and 2P-ERK2 (Gly loop, hinge, helices αC, αE and αL16). HDX protection is similar with all inhibitors in strands β7-β8 (dark green) in 2P-ERK2, but not 0P-ERK2. Time courses for these peptides are shown in Suppl. Fig. S6. Highlighted in light blue are regions where BVD523, VTX11e or GDC0994 lead to decreased HDX uptake around the DFG motif, in 0P-ERK2 or 2P-ERK2. Highlighted in dark blue are regions where BVD523 and VTX11e lead to increased HDX protection, compared to GDC0994. These occur only in 2P-ERK2, and include the DFG motif and adjacent strand β9, as well as the P+1 segment and helix αF. Time courses for strand β9 are shown in Suppl. Fig. S7. Full HDX datasets for all inhibitors are presented in Suppl. Dataset S1. Crystal structures shown in panels A-C are (A) PDBID: 6GDQ (left) and 2ERK (right); (B) PDBID: 4QTE (left) and 6OPK (right); (C) PDBID: 5K4I (left) and 6OPH (right).

Figure 3.. Novel ATP-competitive inhibitors of ERK1/2.

(A) A panel of 17 novel ERK1/2 inhibitors surveyed in this study. Estimates of K_i were measured using kinase assays for phosphorylation of Omnia peptide substrate (Invitrogen). (B) Variations among ERK1/2 inhibitors in their left-side, central scaffold, and a sampling of right-side substituents. (C) Contacts formed by VTX11e and GDC0994 with active site residues in 2P-ERK2, based on published X-ray structures (PDBID:6OPK, PDBID:6OPH).

Figure 4.. HDX assays survey conformation selection among compounds in the ERK inhibitor panel.

HDX measurements performed with representative inhibitors shown in Fig. 3, chosen for variations in their left-side, central scaffold, and right-side substituents. Time courses show deuterium uptake at the (A) DFG motif (peptide 161–168: LKICDFGL), (B) P+1 segment (peptide 191–198, YRAPEIML), and helix αF (peptide 203–210: YTKSIDIW). Enhanced HDX protection (strongly decreased uptake) in each segment by inhibitors #5, #6, #8 and #16 (blue) suggest properties of conformation selection for the R-state, while lower protection by inhibitors #4 and #15 (grey) suggest retention of conformational exchange. Inhibitor #1 (cyan) shows HDX properties intermediate to these two groups. HDX time courses for the full set of 17 inhibitors are shown in Suppl. Fig. S8A,B.

Figure 5.. ERK inhibitors differentially cluster with VTX11e/BVD523 or GDC0994.

Effects of inhibitors on HDX were quantified by a difference AUC measurement, which calculates dAUC = Σ_t(HDX_Apo- HDX_Inhibitor)_t over all time points. (A) Plot comparing dAUC for the DFG motif (peptide 161–168) vs P+1 segment (peptide 191–198) reveals 13 inhibitors clustered with VTX11e/BVD523, 3 inhibitors clustered with GDC0994, and one inhibitor with intermediate properties. Labels mark the representative set of inhibitors shown in Fig. 4. (B) dAUC values for each inhibitor.

Figure 6.. 2D-HMQC NMR confirms R-state selection by ERK inhibitors.

2D-HMQC NMR spectra were collected on 2P-ERK2 at 25°C and 5°C. Effects on [methyl ¹³C,¹H] peaks of residues (A) I72, (B) L220 and (C) L242 are shown for the representative set of inhibitors in Fig. 4. Inhibitors #5, #6, #8 and #16 (shown in blue) shifted the R⇌L equilibrium to 100% R at all temperatures, confirming R-state selection as suggested by their HDX behaviors in Figs. 4 and 5. Inhibitors #4 and #15 (in grey) retained R:L populations comparable to those with GDC0994. Inhibitor #1 (in cyan) showed partial selection for the R state but retained conformational exchange, as evidenced by an L state population present at 5°C. Full NMR spectra are shown in Suppl. Fig. S9.

Figure 7.. NMR chemical shift patterns reveal long distance perturbations by R-state inhibitors.

2D-HMQC NMR spectra of 2P-ERK2 at 25°C showing [methyl ¹³C,¹H] peaks with significant changes in chemical shift upon binding ERK inhibitors. NMR peaks colored blue correspond to inhibitors characterized as R-state selective (VTX11e, BVD523, #5, #6, #8, #16). Peaks colored grey correspond to inhibitors that allow conformational exchange (GDC0994, #4, #15). Peaks in cyan correspond to inhibitor #1, which shows partial R-state selection. Shown in red spheres on the X-ray structure (PDBID:6OPK) are residues (I196, I345) that separate the R-state selective inhibitors from those that allow exchange. Green spheres are residues (L26, L105, L155, L161) that separate inhibitors with left-side tetrahydropyran (#4, #6, #8, #15) from other substituents. Black spheres are residues that broaden upon binding all inhibitors (I82, L154), or all inhibitors except VTX11e (V12, V37, I101); These together with residues in orange (I138, L148) reflect first- and second- sphere regions of contact around the binding site. Chemical shifts and chemical shift perturbations (Δδ) for all assigned residues are shown in Suppl. Fig. S10 and Suppl. Dataset S2.

Figure 8.. R-state inhibitors promote outward movements of N-lobe structural elements.

(A) Summary of contacts formed by inhibitors #8 and #16 with active site residues in co-crystal structures with 2P-ERK2 (PDBID: 8U8K and 8U8J). (B) The active site of 2P-ERK2 complexed with GDC0994 (PDBID:6OPH, grey) and VTX11e (PDBID:6OPK, blue) (Pegram et al, 2019). Left panel: Front view showing movement of the Gly loop, helix αC and helix αL16 in an “outward” direction (away from the inhibitor) upon binding of VTX11e (blue), relative to GDC0994 (grey). The movement can be attributed to the right-side 3-chlorobenzyl substituent in VTX11e which interacts with the π orbital of Y34 in the Gly loop (Cl-π distance, 3.5 Å). In turn, π–π stacking interactions between Y34 and Y62 couples movements of the Gly loop to helix αC. Right panel: Side view showing left-side hydrogen bond contacts with main chain atoms of hinge residue M106, as typical of ATP-competitive kinase inhibitors. (C,D) Active site of 2P-ERK2 complexed with GDC0994 (PDBID:6OPH, grey) and (C) inhibitor #8 (PDBID:8U8K, slate) or (D) inhibitor #16 (PDBID:8U8J, cyan). Like VTX11e, inhibitors #8 and #16 move the Gly loop, helix αC and helix αL16 outward, relative to GDC0994. Right panels B-D show that all inhibitor complexes share a bound water (W1) bridging the central scaffold to the gatekeeper residue in ERK2 (Q103). Left panels show that a bound water (W2) bridges active site residues K52 and D165 in complexes with VTX11e, inhibitor #8 and inhibitor #16, but not GDC0994. (E) Overlay of GDC0994 (grey), inhibitor #8 (slate) and inhibitor #16 (cyan). The relative N-lobe movements in panels C and D may be explained by differential hydrogen bonding of K52 and W1 to the triazolopyridine central scaffold of inhibitors #8 and #16, distinct from the pyridone scaffold of GDC0994. The position of the hydrogen bond of the triazole nitrogen with K52 relative to the pyridone oxygen moves the K52-E69 salt bridge in an outward direction in inhibitors #8 and #10 relative to GDC0994. Structures were superpositioned by aligning Cα atoms within the C-terminal domain (residues 109–141, 205–245, 272–310).

Figure 9.. ATG017 promotes Gly loop opening and inward movement of N-lobe elements.

(A) Summary of contacts formed by inhibitor ATG017 with active site residues in a co-crystal structure with 0P-ERK2 (PDBID:6SLG). (B) The active site of 0P-ERK2 complexed with ATG017 (PDBID:6SLG, green) and BVD523 (PDBID:6GDQ, light blue). Left panel: Front view showing movement of the Gly loop, helix αC and helix αL16 in an outward direction by BVD523, attributed to the close proximity between the right-side chlorobenzyl substituent and the π orbital of Y34 in the Gly loop (Cl-π distance, 3.3 Å). Right panel: View showing left-side hydrogen bonds with main chain atoms of hinge residue M106. (C) 2D-HMQC spectra of 2P-ERK2 complexed with ATG017 at 25°C and 5°C, showing [methyl ¹³C,¹H] peaks of residues I72, L220 and L242. Unlike 2P-ERK2 complexed with BVD523, the ATG017 complex retains R⇌L exchange resembling that of GDC0994 (Fig. 1). (D) HDX time courses with ATG017 measuring deuterium uptake at the DFG motif, P+1 segment, and helix αF. Time courses for strand β9 are shown in Suppl. Fig. S11. Enhanced HDX protection by ATG017 binding is observed at the DFG and adjacent β9 segments, but minimally at the P+1 and helix αF, similar to that seen with GDC0994. The results suggest that allosteric coupling between the ligand binding pocket and distal regions surrounding the activation loop, but not the DFG motif or β9, are characteristic of R-state inhibitors. Structures were superpositioned by aligning Cα atoms within the C-terminal domain (residues 109–141, 205–245, 272–310).