Strain-release alkylation of Asp12 enables mutant selective targeting of K-Ras-G12D

Abstract

K-Ras is the most commonly mutated oncogene in human cancer. The recently approved non-small cell lung cancer drugs sotorasib and adagrasib covalently capture an acquired cysteine in K-Ras-G12C mutation and lock it in a signaling-incompetent state. However, covalent inhibition of G12D, the most frequent K-Ras mutation particularly prevalent in pancreatic ductal adenocarcinoma, has remained elusive due to the lack of aspartate-targeting chemistry. Here we present a set of malolactone-based electrophiles that exploit ring strain to crosslink K-Ras-G12D at the mutant aspartate to form stable covalent complexes. Structural insights from X-ray crystallography and exploitation of the stereoelectronic requirements for attack of the electrophile allowed development of a substituted malolactone that resisted attack by aqueous buffer but rapidly crosslinked with the aspartate-12 of K-Ras in both GDP and GTP state. The GTP-state targeting allowed effective suppression of downstream signaling, and selective inhibition of K-Ras-G12D-driven cancer cell proliferation in vitro and xenograft growth in mice.

Fig. 1. Malolactone (RS)-G12Di-1 (1) is a selective, covalent inhibitor of K-Ras-G12D.

a, Chemical structure of racemic (RS)-G12Di-1. b, Deconvoluted protein mass spectra of Cyslight K-Ras-G12D•GDP in the presence or absence of (RS)-G12Di-1. Mass spectra are representative of two independent experiments. c, Compound (RS)-G12Di-1 selectively labeled K-Ras-G12D not WT or other mutants. All data points represent individual biological replicates. Data are presented as mean ± standard deviation (n = 3). d, Kinetics of K-Ras-G12D (200 nM) labeling with (RS)-G12Di-1 (10 µM) (n = 3, replicates are plotted as individual data points). e, Time-resolved fluorescence energy transfer dose–response of (RS)-G12Di-1 induced Ras-Raf•RBD binding disruption. All data points represent individual biological replicates. Data are presented as mean ± standard deviation (n = 3). f, Biolayer interferometry dose–response of GST-Raf•RBD binding with K-Ras-G12D•GppNHp without or with covalently labeled (RS)-G12Di-1. g, Co-crystal structure of K-Ras-G12D•GDP•(RS)-G12Di-1. h, 2F_o–F_c map for the covalently bound ligand (RS)-G12Di-1 and Asp12 is depicted in blue mesh (1.0σ). i, Chemical structure of covalently bound ligand with the configuration of β-carbon assigned as S.

Fig. 2. Sterically hindered β-lactones are stable and potent inhibitors of K-Ras-G12D.

a, Mechanistic analysis of the Asp12-alkylating (red) and hydrolytic (teal) processes reveals the higher sensitivity of the latter to C3 substituents. b, Chemical structures of C3-substituted malolactones (R)-G12Di-2 (2), (2R,3S)-G12Di-3 (3), and (2R,3S)-G12Di-4 (4). c, Recombinant K-Ras-G12D (200 nM) labeling kinetics with 3-substituted malolactones (10 µM) (n = 3, replicates are plotted as individual data points). d, Stability of 3-substituted β-lactones in PBS (pH 7.4) by LC–MS (n = 3, replicates are plotted as individual data points).

Fig. 3. Rapid covalent modification of K-Ras-G12D•GTP is essential for in-cell target engagement and oncogenic signaling suppression.

a, Chemical structures of (2R,3S)-G12Di-4 (4), (2R,3S)-G12Di-5 (5) and (2R,3S)-G12Di-6 (6). b, Recombinant K-Ras-G12D (200 nM) labeling kinetics with 3-isopropyl malolactones (10 µM) (n = 3, replicates are plotted as individual data points). c, Western blot time course of cellular K-Ras-G12D covalent engagement and downstream signaling inhibition. Data are representative of two independent experiments. d, Covalent K-Ras-G12D labeling kinetics in both nucleotides by (2R,3S)-G12Di-5, (2R,3S)-G12Di-6 and (2R,3S)-G12Di-5a–5f (7–12). Substitutions X, Y and Z vary between compounds 7–12. e, Analysis of the determining substitution on covalent labeling kinetics by an unpaired t-test comparing X = H (n = 5) or F (n = 3), Y = H (n = 4) or OH (n = 4), and Z = H (n = 4) or F (n = 4). All data points represent individual chemical compound. Data are presented as mean ± standard deviation. Source data

Fig. 4. Malolactone (R)-G12Di-7 (13) covalently and mutant-selectively modified recombinant and endogenous K-Ras-G12D oncoprotein in cancer cell lines.

a, Chemical structure of (R)-G12Di-7. b, Pseudo-first-order K-Ras-G12D labeling kinetics of (R)-G12Di-7. Conditions: K-Ras-G12D (200 nM), (R)-G12Di-7 (10 µM), room temperature. c, Second-order K-Ras-G12D labeling kinetics of (R)-G12Di-7. d, Covalent labeling selectivity against K-Ras WT and mutants. All data points represent individual biological replicates. Data are presented as mean ± standard deviation (n = 3). e, Immunoblot of Ba/F3:K-Ras-G12D, SW1990 AsPC-1, AGS, HCT116, A549, A375 and H1299 cells treated with DMSO or 10 µM (R)-G12Di-7 for 4 h. Data are representative of two independent experiments. f, Stability of covalent complex G12D•(R)-G12Di-7 in cells. Data are representative of two independent experiments. g, Relative growth of Ba/F3:K-Ras-G12D cells (with or without 10 ng ml⁻¹ IL-3) after treatment with (R)-G12Di-7 for 72 h. Data are presented as mean ± standard deviation (n = 3) and are representative of two independent experiments. Source data

Fig. 5. Covalent K-Ras-G12D inhibitor (R)-G12Di-7 selectively inhibits cell growth in cancer cell lines harboring KRAS^G12D mutation and tumor growth in mice bearing SW1990 xenograft.

a, Relative growth of cancer cell lines with (black, red or teal) or without (gray) KRAS^G12D mutation after treatment with (R)-G12Di-7 for 72 h (2D) or 120 h (3D). Data are presented as mean ± standard deviation (n = 3) and are representative of two independent experiments. b, Tumor volumes and body weight of mice bearing SW1990 xenografts and treated with vehicle (10% captisol in 1× PBS, n = 8 biologically independent mice), (R)-G12Di-7 (10 mg kg⁻¹, n = 9 biologically independent mice), or (R)-G12Di-7 (50 mg kg⁻¹, n = 9 biologically independent mice). BID, twice a day. All data points represent individual biological replicates. Data are presented as mean ± standard error of the mean. Tumor volumes for the (R)-G12Di-7 (50 mg kg⁻¹) treatment groups are statistically significant versus vehicle (P < 0.00001) by two-tailed Student’s t-test.

Extended Data Fig. 1. Covalent modification of K-Ras-G12D by strain-release electrophiles.

a, Covalent labeling kinetics. Conditions: 200 nM K-Ras-G12D•GDP, 10 µM compound, 20 mM HEPES pH 7.5, 150 mM MgCl₂, 23 °C. b, Chemical structures of strain-release electrophiles.

Extended Data Fig. 2. β-Propiolactone covalently labeled Asp12 of recombinant K-Ras-G12D.

a, Alkylation of Asp12 by β-propiolactone and identified fragment ions of aa. 6-16 peptide including covalently modified Asp12. b, Extent of covalent modification is dose dependent. All data points represent individual biological replicates. Data are presented as mean ± SD (n = 2).

Extended Data Fig. 3. Covalent K-Ras-G12D•1 adduct is stable in pH 4.5–7.5 buffers and to thiol (DTT) or amine (hydrazine) electrophiles.

Mass spectra are representative of two independent experiments.

Extended Data Fig. 4. Mechanistic rationale of the enantioselectivity of the covalent modification reaction.

a, Two possible reaction pathways to form (S)-adduct with distinct stereochemical events at α-carbon. b, K-Ras-G12D favors an (R)-enantiomer of the malolactone electrophile.

Extended Data Fig. 5. Cell growth inhibition of advanced β-lactone G12Di against a panel of G12D or non-G12D cell lines.

a, Cell growth inhibition curves. All data points represent individual biological replicates. Data are presented as mean ± SD (n = 3). b, Half-maximal inhibitory concentrations (IC₅₀s) of each biological replicate are plotted.

Extended Data Fig. 6. Malolactone compounds 4–6 inhibited 20 S proteasome while (R)-7 did not.

a, Chemical structures of Compound 6 and Belactosin C with β-lactone warhead highlighted in red. b, Immunoblot of poly-ubiquitinated proteins in HEK293 cells All data points represent individual biological replicates. Data are presented as mean ± SD (n = 3). Source data

Extended Data Fig. 7. Analysis of X-ray co-crystal structures containing covalent malolactone electrophile.

a, α-Carbon of malolactone G12Di could tolerate geminal disubstitution. b, Belactosin C could not tolerate a second α-substitution. c, Additional methyl group (circled in yellow) in belactosin analog caused significant loss of activity.

Extended Data Fig. 8. Intrinsic reactivity and stability of G12Di-6 and 7.

a, Stability in the presence of excess compounds (5 mM) bearing reactive functional groups. All data points represent individual biological replicates. Data are presented as mean ± SD (n = 2). b, Stability in 1X PBS at various pH. All data points represent individual biological replicates. Data are presented as mean ± SD (n = 2).

Extended Data Fig. 9. Melting point change of covalently crosslinked K-Ras-G12D.

a, Summary of thermal stabilization of K-Ras-G12D by G12Di-1–7. b, DSF melting curves of K-Ras-G12D•GDP or GppNHp covalently bound to (R)-G12Di-7. All data points represent individual biological replicates. Data are presented as mean ± SD (n = 3).

Extended Data Fig. 10. Covalency of malolactone is important for the G12D cell growth inhibition.

a, Chemical structures of covalent K-Ras-G12D inhibitor and non-covalent analogs. b, Non-covalent analogs inhibited K-Ras-G12D cell lines with significantly lower potency. All data points represent individual biological replicates. Data are presented as mean ± SD (n = 3).