Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2026 Jan 2;66(2):1203–1213. doi: 10.1021/acs.jcim.5c02787

Osimertinib’s Proton-Catalyzed, Pseudoconcerted EGFR Inhibition Guides Next-Generation Inhibitor Design

Akihiro Kondo †, Kazuhiro J Fujimoto †,‡,*, Shin-ichi Koda §,∥, Shinji Saito §,∥, Takeshi Yanai †,‡,*
PMCID: PMC12848964  PMID: 41481194

Abstract

Third-generation inhibitors such as osimertinib irreversibly inhibit the epidermal growth factor receptor tyrosine kinase (EGFR-TK) via Michael addition to Cys797, yet the mechanism of covalent bond formation remains unsettled. Here, density functional theory-level quantum mechanics/molecular mechanics (QM/MM) computations delineate a proton-catalyzed, pseudoconcerted mechanism for covalent inhibition of EGFR by osimertinib. The computed Gibbs energy profile features a single transition state (TS1) in which Cys797 deprotonation, nucleophilic attack at osimertinib’s Michael acceptor, and protonation of the acceptor’s carbonyl oxygen by osimertinib’s terminal aliphatic aminium occur simultaneously, in contrast to previous stepwise proposals. Natural bond orbital (NBO) analysis shows that carbonyl O-protonation attenuates carbonyl π bonding, increases β-carbon (Cβ) electrophilicity, and thereby facilitates formation of the Cys797–Cβ bond. The resulting path proves to be more favorable in both kinetics and thermodynamics than stepwise alternatives, consistent with experimental trends. The calculations also rationalize osimertinib’s preferential inhibition of the T790M mutant. Electrostatic potential (ESP) analysis demonstrates that T790M subtly redistributes charge within osimertinib’s pyrimidine–indole scaffold, enhancing the cationic character at the indole N-methyl substituent and strengthening its electrostatic interaction with Asp855 to stabilize the reactant complex. Together, these results provide a mechanistic framework for covalent bond formation underlying EGFR inhibitionhighlighting inhibitor-derived proton catalysis and Asp855 engagementand offer design principles for next-generation covalent inhibitors with improved potency and resistance-breaking potential.

graphic file with name ci5c02787_0008.jpg

graphic file with name ci5c02787_0006.jpg

1. Introduction

Targeted covalent inhibitors (TCIs) have emerged as a transformative class of anticancer agents, whose high target selectivity enables reduced off-target toxicities relative to conventional agents. Within this class, epidermal growth factor receptor tyrosine kinases (EGFR-TKs) constitute principal targets: EGFR signaling orchestrates cellular development, differentiation, and survival, and its aberrant activation drives diverse malignancies including carcinomas, sarcomas, non-small cell lung cancer (NSCLC), and malignant gliomas. Pharmacologic blockade of the intracellular catalytic domain by EGFR-TK inhibitors (EGFR-TKIs) prevents EGFR autophosphorylation and downstream signaling pathways, constraining cancer cell proliferation (G0–G1 phase arrest), suppressing VEGF-driven angiogenesis, and limiting invasive capacity. Across successive generations, design has converged on covalent engagement of Cys797: second-generation agents such as afatinib and dacomitinib and third-generation agents exemplified by osimertinib (Figure a) achieve durable target engagement through covalent modification of Cys797 within the EGFR-TK active site. Cocrystal structures for afatinib at Cys797 together with evidence from a mouse plasma-mimicking solution that osimertinib reacts readily with free cysteine are consistent with covalent inhibition via Michael addition, in which Cys797 acts as the nucleophile and the inhibitor’s acrylamide warhead serves as the Michael acceptor. Clinically, second-generation inhibitors retain activity against exon 19 deletions (Del19) and the L858R mutation, whereas third-generation agents additionally address the T790M gatekeeper mutation, supporting regulatory approvals and clinical use worldwide. Nevertheless, the elementary chemical steps of covalent bond formation remain incompletely resolved; furthermore, therapy with agents such as osimertinib is associated with the emergence of the on-target C797S substitution, which accounts for 10–26% of resistance mutations, thereby reinforcing the need for mechanism-guided inhibitor design. , The development of fourth-generation EGFR inhibitors that engage alternative nucleophiles (e.g., Lys745) is underway, , and a more precise delineation of osimertinib’s inhibitory mechanism could inform these efforts.

1.

1

Open in a new tab

(a) Chemical structures of osimertinib and Cys797 in the reactant state and as the Cys797–osimertinib thioether adduct (Michael addition); osimertinib is depicted with its terminal aliphatic amine in the protonated (aminium) form (red circle). X-ray cocrystal of wild-type EGFR bound to osimertinib (PDB ID: 4ZAU), highlighting the active-site region. (b) Synopsis of previously proposed cysteine-targeting mechanisms for EGFR and BTK covalent inhibitors.

A definitive mechanistic account of EGFR-TKI reactivity is essential for the rational design of next-generation covalent inhibitors. Yet conventional X-ray crystallography rarely resolves hydrogen positions and therefore cannot directly assign protonation states to key EGFR active-site residues or to the bound inhibitor. Given that protonation states govern acid–base catalysis, nucleophilicity/electrophilicity, and barrier heights within protein active sites, computational chemistry becomes indispensable for interrogating these variables and their roles in the reaction coordinate. Despite numerous studies of EGFR covalent inhibition, a consensus remains elusive. Two questions at the rate-determining step are particularly contested: (i) whether the inhibitor’s terminal aliphatic amine is protonated and, if so, whether it donates a proton as a general acid; and (ii) which species acts as the general base to deprotonate Cys797 (e.g., Asp800, an active-site water, or the inhibitor itself).

Previous quantum mechanics/molecular mechanics (QM/MM) investigations arrived at conflicting mechanistic conclusions (Figure b and Table S1). Capoferri et al. attributed Cys797 deprotonation to Asp800, on the basis of self-consistent-charge density functional tight-binding (SCC-DFTB)/AMBER99SB free energy simulations of N-(4-anilinoquinazolin-6-yl) acrylamidean inhibitor lacking a protonatable terminal amine. Callegari et al., modeling osimertinib with SCC-DFTB/AMBER99SB, assumed a protonated terminal amine but likewise assigned Asp800 as the general base. Ma et al., using pairwise distance directed Gaussian modified PM3 (PDDG-PM3)/CHARMM22 simulations of afatinib, assumed the terminal amine to be unprotonated and proposed intramolecular deprotonation of Cys797 by that amine. Kaewkham et al., using density functional theory (DFT)/AMBER99SB for rociletinib, also held the terminal amine unprotonated and assigned Cys797 deprotonation to either a water molecule or Asp800; notably, they further proposed that an active-site water donates a proton to the Michael acceptor’s carbonyl oxygen to generate an enol-type intermediate, introducing a new element into the EGFR-TKI literature. An analogous enol-type pathway had earlier been advanced for Bruton’s tyrosine kinase (BTK) inhibition by ibrutinib: Voice et al., with third-order self-consistent-charge density functional tight-binding (DFTB3)/MM umbrella sampling, showed that the Michael acceptor’s carbonyl oxygen can deprotonate the catalytic cysteine, thereby invoking an enol-type intermediate.

This persistent lack of consensus on the identities and timing of acid–base participants has impeded mechanism-guided inhibitor design and slowed the development of agents effective against emerging resistance mutations. Faithful modeling of covalent inhibitor–thiol reactions requires rigorous control of computational parameters, particularly selecting a DFT exchange-correlation functional (e.g., ωB97X-D or PBE0) that achieves qualitative accuracy sufficient to capture the mechanism at tractable cost. Accordingly, we integrate prior experimental insights with higher-accuracy QM/MM computations that employ DFT-level QM/MM geometry optimizations and Gibbs energy calculations to chart the covalent inhibition of wild-type EGFR-TK by the third-generation inhibitor osimertinib. Within this framework, we advance and test a focused hypothesis that the inhibitor’s terminal aliphatic amineprotonated and serving as the countercation to Asp800catalyzes the reaction via proton transfer. Anticipating our results, we find that (i) assuming a protonated terminal amine (aminium), proton transfer to the Michael-acceptor carbonyl activates the β carbon of the acrylamide and promotes Cys797 addition; (ii) the rate-determining transition state is pseudoconcerted, integrating Cys797/Asp800 proton transfer, nucleophilic attack, and carbonyl-oxygen protonation; and (iii) Asp855 engagement adjacent to the indole N1/N-methyl locus is a leverageable contact for design. Taken together, these findings reconcile prior mechanistic divergences and inform the structure-based optimization of next-generation covalent inhibitors.

2. Method

All computations employed the ONIOM (our own N-layered integrated molecular orbital and molecular mechanics) method. The wild-type EGFR-osimertinib cocrystal structure (PDB ID: 4ZAU) served as the starting model. Although the deposited coordinates span Ala698-Leu1017, the segment Glu985-Met1007 is largely unresolved; accordingly, we truncated the model at Asp984 and retained residues Ala698-Asp984. The truncation site lies ∼28 Å from the active site, measured to the Cβ of osimertinib, and is thus well outside the catalytic region. Protonation states were assigned with the Reduce module in UCSF Chimera, and any missing residues within the retained segment were rebuilt with MODELLER. The QM region comprised the side chains (from the α-carbons) of Lys745, Glu762, Thr/Met790, Cys797, Asp800, and Asp855, together with the entire osimertinib molecule. These residues were selected because they either lie within the active-site catalytic region or mediate key contacts with the inhibitor as suggested by prior studies. − ,, As the cocrystal structure (PDB ID: 4ZAU) shows no water molecules near the reaction center, the QM region was defined without explicit water molecules. Complementary pK a estimates (DelPhiPKa2.3 − and PROPKA3.1 , ) were used to assess titration behavior of QM-region residues and protonatable sites on osimertinib and to support the protonation-state assignments (Table S2). For geometry optimization, we first performed a constrained ONIOM optimization, relaxing only the osimertinib-Cys797 fragment while holding all other atoms fixed, followed by a full ONIOM optimization at the B3LYP-GD3BJ/6-31G­(d) level with all atoms relaxed. Vibrational frequency calculations at the same level provided zero-point energy and thermal corrections at 298.15 K, which were used to augment single-point electronic energies computed at ωB97X-D/6-311+G­(d) on the optimized structures to assemble Gibbs energies. A prior comparison of DFT methods for modeling thio-Michael additions identified ωB97X-D as one of the most reliable functionals for describing carbanion-intermediate stability. All remaining atoms formed the MM region and were treated with AMBER96 force fields. Electrostatic potential (ESP)-derived charges computed from a local optimization of the Cys797–osimertinib fragment were used as MM charges for these components in the subsequent full complex optimization. The reactant, transition state 1 (TS1), intermediate, transition state 2 (TS2), and product structures were fully optimized at the ONIOM level of theory. The transition states were verified by intrinsic reaction coordinate (IRC) calculations. All QM/MM calculations were performed with Gaussian16. Reaction path exploration was performed by additionally employing the direct MaxFlux method, seeded by an initial path built with the flat-bottom elastic network. No constraints were applied to the atomic positions during the ONIOM geometry optimizations or reaction path searches.

3. Results and Discussion

3.1. Experimental Rationale for Proton Catalysis

Three empirical observations motivate our hypothesis. First, the terminal aliphatic amine of osimertinib has a pK a ≈ 9.5, implying that it is predominantly protonated under physiological conditions. Second, across EGFR-directed covalent inhibitors, reported pK a values correlate with glutathione half-lives in glutathione reactivity assays, implicating proton participation in the covalent bond-forming step of inhibition. Computational pK a estimates for the EGFR–osimertinib complex adopted as our computational model (PDB ID: 4ZAU) give a value of 10.1 for the terminal amine, supporting its protonation under physiological conditions (Table S2). Third, Fenselau et al. reported acid-catalyzed hetero-Michael addition reactions (ACHMAR) under atmospheric conditions that yield secondary organic aerosol, and Wabnitz et al. presented evidence that, in Lewis acid-mediated systems, hetero-Michael additions proceed by proton catalysis. Together, these studies support the feasibility of enol-type intermediates in EGFR-TKI reactions, as suggested by Kaewkham et al. The combined evidence is consistent with a mechanism whereby protonation of the inhibitor’s Michael-acceptor carbonyl oxygen by the inhibitor’s terminal aliphatic aminium (i.e., the protonated amine) (Figure a) accelerates covalent bond formation.

Regarding intermediate protonation, we define the α-carbon (Cα) as the vinylic carbon adjacent to the carbonyl within osimertinib’s acrylamide Michael acceptor (with Cβ denoting the distal vinylic carbon) (Figure a). Although activation energies for keto–enol tautomerization are typically high without catalysis, , enzymatic precedents (e.g., triosephosphate isomerase and enoyl-CoA hydratase) show that appropriately positioned residues (e.g., glutamate) can catalyze the requisite substrate protonations. In the EGFR active site, the absence of nearby waters in the cocrystal structure, together with the proximity of Asp800 to Cys797, motivates a working model in which Asp800 mediates Cα protonation in the intermediate state. On this basis, we propose a reaction path catalyzed jointly by the inhibitor’s terminal aminium and Asp800.

3.2. ONIOM Geometry Optimizations and Free Energy Calculations

ONIOM optimizations together with IRC paths from the located transition states delineate a two-step, proton-catalyzed Michael-addition mechanism for the covalent inhibition of EGFR by osimertinib (Figure a). Step 1 comprises three coupled events: Asp800 deprotonates Cys797, the nascent Cys797 thiolate executes a nucleophilic attack on the Michael acceptor, and osimertinib’s terminal aminium donates a proton to the Michael-acceptor carbonyl oxygen. Step 2 completes the reaction: the α-carbanion is protonated by Asp800, and the proton donated to the Michael acceptor in Step 1 is returned to the amine, regenerating the aminium.

2.

2

Open in a new tab

(a) Proposed proton-catalyzed, pseudoconcerted mechanism for EGFR tyrosine kinase inhibition by osimertinib. (b) ONIOM-optimized geometries of the stationary points along the reaction coordinate: reactant, TS1, intermediate, TS2, and product. (c) Gibbs energy profile at 298.15 K, referenced to the reactant. The activation energy for TS1 computed with ONIOM­(ωB97X-D/6-311+G­(d):AMBER96) is 15.2 kcal/mol, in good agreement with the higher-level ONIOM (scaled opposite-spin second-order Møller–Plesset perturbation theory/6-311+G­(2d,2p):AMBER96) value of 15.6 kcal/mol (Table S3). (d) IRC trajectories initiated from TS1 toward the reactant and the intermediate, with representative structures (IRC = 11.8 and 16.3 au) and relative energies (reactant = 0 kcal/mol).

Optimized stationary-point geometries are presented in Figure b. In the reactant complex, the distance between the terminal amine nitrogen and the Michael-acceptor carbonyl oxygen is 2.77 Å, consistent with a hydrogen bond that orients the active site for proton transfer to the carbonyl oxygen. The proposed reaction path involves transient proton transfer between Cys797 and Asp800. Computational pK a estimates derived from the crystal structure (PDB ID: 4ZAU) assign pK a values of 11.1 to both residues (Table S2), and the Cys797 estimate agrees with the value reported for apo-form EGFR Cys797 by the replica-exchange thermodynamic integration method, together supporting the mechanistic plausibility of this proton transfer.

Following this path, we computed the free energy (ΔG) profile for the formation of the Cys797–osimertinib thioether adduct (Michael addition) in wild-type EGFR using ONIOM (Figure c). Relative to the reactant, the intermediate and product lie at −11.7 and −21.3 kcal/mol, respectively; thus, both elementary steps are thermodynamically favorable (ΔG < 0). Free energy barriers (ΔG ‡) of 15.2 and 5.7 kcal/mol for TS1 and TS2, respectively, identify TS1corresponding to the nucleophilic attack by the Cys797 thiolate on the Michael acceptoras the rate-determining step (details in Table S4). Experimentally, the inactivation rate (k Inact) for osimertinib against wild-type EGFR is 0.036 s–1; by the Eyring equation at 298.15 K, this corresponds to ΔG ‡ ≈ 19.4 kcal/mol. The computed barrier for TS1 (15.2 kcal/mol) is in qualitative agreement with the experimental estimate, supporting the proposed mechanism. The remaining ∼4.2 kcal/mol discrepancy may reflect limitations of the rigid-rotor harmonic-oscillator approximation for entropic contributions, which cannot be further constrained in the absence of experimental kinetic isotope effect data. Vibrational frequency analyses of the optimized TS1 and TS2 geometries each yielded a single imaginary mode associated with the intended reaction coordinate, confirming their assignment as first-order saddle points (Table S5).

3.3. IRC Analysis of TS1: Evidence for a Pseudoconcerted Elementary Step

In contrast to most prior proposals of a stepwise sequence in which Cys797 deprotonation precedes Michael addition, our IRC analysis indicates that these events are encompassed within a single, pseudoconcerted transition state. IRC trajectories initiated from the optimized TS1 connect continuously to both the reactant and intermediate (Figure d and Table S6), which also present representative structures along the downhill path toward the intermediate at IRC = 11.8 and 16.3 au. The earlier structure (11.8 au) shows Cys797 deprotonated but prior to C–S bond formation, whereas the later structure (16.3 au) displays nucleophilic attack by the Cys797 thiolate on the Michael acceptor. These results support the interpretation that Cys797 deprotonation, nucleophilic attack, and carbonyl O-protonation proceed in a single elementary step (TS1). Plots of interatomic distances for the reacting atoms along the IRC further corroborate this mechanism (Figure S1), and the same reaction path is independently recovered from reaction path optimization by using the direct MaxFlux method (Figure S2). By concentrating the elementary events into a single rate-determining transition state, this pseudoconcerted mechanism streamlines the reaction coordinate and rationalizes the modest rate-determining free energy barrier relative to stepwise alternatives, offering a novel framework for understanding covalent bond formation underlying EGFR inhibition by TKIs. This pseudoconcerted behavior likely reflects the protein microenvironment as a tailored reaction fieldfavoring mechanisms that are inaccessible or disfavored in solution.

3.4. Molecular-Orbital Analysis: Protonated Terminal Amine Enhances β-Carbon Electrophilicity

To probe the functional role of proton catalysis, we generated two alternative protonation-state models at a common IRC point (11.8 au) along the IRC path from TS1 toward the intermediate: (i) a protonated model in which osimertinib’s terminal aliphatic amine is an aminium and (ii) a deprotonated model lacking that proton. We then performed natural bond orbital (NBO) analysis on both structures to compare how protonation modulates the π system of the Michael acceptor and the reactivity of Cys797. Figure displays the relevant NBOsthose representing the π-network of osimertinib’s Michael acceptor and the lone pair on the deprotonated Cys797 sulfur. In the protonated model, proximity of the proton renders the Michael acceptor’s carbonyl-oxygen π-electrons effectively lone-pair-like; consequently, π density is reinforced between the carbonyl carbon and Cα, yielding a strong π-bond along CCα. In contrast, in the deprotonated model, the carbonyl oxygen engages more strongly in π-bonding to the carbonyl carbon, and the π orbital instead localizes along Cα–Cβ, producing a π-bond on Cα–Cβ. These observations strongly suggest that the proton promotes electron localization on the carbonyl oxygen (lone-pair-like), weakening π overlap with the carbonyl carbon and reducing the contribution of Cβ to conjugation.

3.

3

Open in a new tab

Natural bond orbital (NBO) analysis at IRC = 11.8 au along the TS1-to-intermediate trajectory, comparing (1) osimertinib bearing a protonated terminal aliphatic amine (aminium) and (2) the deprotonated (neutral)-amine counterpart. Shown are the NBOs corresponding to the π orbital of osimertinib’s Michael acceptor and the lone pair on Cys797 sulfur.

This difference is mirrored in the behavior of the Cys797 sulfur lone pair (Figure ): in the protonated model, it converts to a σ-bond with Cβ, whereas in the deprotonated model, it remains nonbonding. Quantitatively, this trend is supported by second-order perturbation energy (E (2)) analysis of donation from the Cys797 sulfur lone pair to the π­(Cα–Cβ)* orbital: the E (2) values are 59.0 and 55.2 kcal/mol in the protonated and deprotonated models, respectively, indicating a stronger NBO donor–acceptor interaction in the protonated case (Table S7). Together, the NBO results indicate that protonation at the Michael acceptor’s carbonyl oxygen by the terminal amine enhances the electrophilicity of Cβ and facilitates σ-bond formation to sulfur, thereby promoting the nucleophilic attack that defines the rate-determining step.

3.5. Key Mechanistic Specificities Relative to Prior Models

3.5.1. Fewer Elementary Steps via Proton-Coupled Activation

By explicitly invoking the proton on the inhibitor’s terminal aliphatic amine, our path achieves inhibition with fewer elementary steps than prior schemes. Most previous models either omitted participation of the terminal amine proton or invoked water-mediated proton donation to the Michael acceptor and typically treated Cys797 deprotonation and nucleophilic addition as separate, sequential steps. In our model, these events proceed pseudoconcertedly at TS1; proton transfer from the inhibitor’s terminal aliphatic aminium coordinates the events and drives the Michael addition, a role not previously ascribed to an aliphatic-amine proton in the EGFR-TKI literature.

3.5.2. Stabilization of the Intermediate and Enhanced Forward Progression

In many previously proposed mechanisms, the Cys797 addition intermediate lies above the reactant in free energy, providing little thermodynamic drive for the onward reaction. In contrast, our path yields an intermediate that is lower in free energy than the reactant, thereby facilitating the progression toward the product. The intermediate identified here is structurally distinct from any species proposed in earlier mechanisms summarized in Figure b (details in Table S1). While t-butyl cyanoacrylamide has been suggested to form a stable enolate intermediate with a cysteine residue in BTK, our work provides, to our knowledge, the first evidence for a stable enol-type intermediate formed by an EGFR inhibitor. Mechanistically, the IRC energy profile (Figure d) pinpoints proton transfer from the terminal aminium to the carbonyl oxygen as the principal source of this stabilization, biasing the reaction coordinate toward product formation.

Such behaviorwhere local interactions reshape proton affinities and modulate the free energy landscapeis characteristic of reactions in protein active sites. Our proposed path parallels ACHMAR, an acid-catalyzed hetero-Michael addition reported under atmospherically relevant acidity, which rationalizes reactivity under naturally generated conditions and offers a useful analogue for protein environments. Enzymatic precedents illustrate that active-site microenvironments can markedly alter proton affinities: in acetoacetate decarboxylase (AAD), the pK a of an active-site lysine is significantly lowered by steric proximity, enabling acid catalysis. By analogy, binding of osimertinib in the EGFR kinase domain colocalizes the terminal aliphatic amine and the Michael acceptor, promoting proton-assisted activation of the carbonyl and facilitating covalent bond formation. Thus, EGFR-targeted covalent inhibitors appear to leverage the geometric architecture of the kinase active site to drive inhibition via proton transfer.

3.6. Determinants of Enhanced Inhibition of T790M by Osimertinib

To rationalize osimertinib’s preferential inhibition of T790M mutant EGFR relative to the wild type, we analyzed the reactant and product free energies obtained from ONIOM geometry optimizations and vibrational frequency calculations (Figure a). The product free energy relative to the reactant is ΔG = −21.3 kcal/mol for the wild type and −22.0 kcal/mol for T790M (details in Table S8). At 298.15 K, the Boltzmann distribution ratio of product to reactant (ρP/ρR) is defined as

ρPρR=exp(−GP−GRRT) 1

where R is the gas constant, T is the temperature, and G P and G R are the free energies of the product and reactant, respectively. The corresponding ρP/ρR values are 3.7 × 1015 for wild type and 1.4 × 1016 for T790M, indicating that the inhibited state is populated 3.73-fold more in T790M than in wild type. Although these are thermodynamic estimates, they are directionally consistent with kinetic measurements from a stopped-flow double mixing assay, which reported a k Inact of 0.036 s–1 for wild type and 0.12 s–1 for the L858R/T790M mutant (3.33-fold higher), used here as a proxy in the absence of kinetic data for the T790M single mutant. Despite the difference in constructs (double mutant versus T790M alone), the magnitudes are comparable, lending qualitative support to the computed preference.

4.

4

Open in a new tab

(a) Free energy changes (ΔG) for osimertinib adduct formation with wild type and T790M, and the corresponding Boltzmann population ratios of product to reactant. (b) Pairwise interaction energies (E Int) between osimertinib and six active-site residues in the reactant complexes of wild type and T790M. (c) ESP maps and ESP-fitted atomic charges for selected atoms in the reactant complexes, illustrating mutation-dependent charge redistribution; red denotes negative, and blue denotes positive ESP-fitted charges (au).

To identify the source of this preference, we computed pairwise interaction energies (E Int) between osimertinib and the six active-site residues included in the QM regionLys745, Glu762, Thr/Met790, Cys797, Asp800, and Asp855for the reactant complexes of wild type and T790M (Figure b and Table S9). The Asp855–osimertinib interaction is notably stronger in T790M (−52.0 kcal/mol) than in the wild type (−46.8 kcal/mol), a stabilization of approximately 5 kcal/mol. Summed over the six residues, E Int is −132.3 kcal/mol for T790M versus −126.6 kcal/mol for the wild type, with the Asp855 term providing the dominant differential. Because Asp855 lies within the Asp855–Phe856–Gly857 (DFG) motif, and strengthened inhibitor–Asp855 contacts have been proposed as a design principle for potent EGFR inhibitors, , these results support the view that third-generation EGFR inhibitors, including osimertinib, benefit from enhanced Asp855 engagement in the T790M background.

3.7. Electrostatic Origin of Enhanced Asp855 Engagement in T790M

To investigate whether the T790M substitution enhances Asp855–osimertinib engagement through alterations in local electrostatics, we generated electrostatic potential (ESP) distribution maps and extracted ESP-fitted atomic charges for a cluster comprising osimertinib and the six QM-region residues (Figure c). Within the osimertinib scaffold, the N-methyl substituent at the indole N1 atom, which lies closest to Asp855, is more cationic by ESP-fitted charge in T790M: for wild type, the carbon and three hydrogens carry atomic ESP charges of −0.259, 0.130, 0.109, and 0.094 au (sum 0.074 au), whereas for T790M the corresponding values are −0.150, 0.082, 0.089, and 0.061 au (sum 0.082 au). By the same ESP-fitted charge metric, the nearest Asp855 oxygen is more anionic in T790M (−0.692 au) than in wild type (−0.685 au). Taken together, these differences yield greater polarization between the N-methyl group and Asp855 in T790M, thereby strengthening electrostatic attraction and accounting for the ∼5 kcal/mol stabilization in E Int noted above. The 790 substitution also perturbs charge distribution within the inhibitor’s π-system: the pyrimidine C5 atom bears −0.637 au in wild type versus −0.396 au in T790M, reflecting the greater ring polarization induced by Thr790’s hydroxyl group relative to Met790’s less polar thioether. Attenuation of this polarization in T790M redistributes electron density across the pyrimidine–indole framework, increasing the positive character at the N-methyl substituent and thereby strengthening Asp855 engagement in the reactant complex. Collectively, the ESP analysis supports the view that osimertinib’s conjugated scaffold leverages active-site electrostatic field: the T790M substitution rebalances charge within the inhibitor, reinforces inhibitor–Asp855 engagement, and thereby contributes to preferential inhibition of T790M.

4. Conclusions

Our computations delineate a pseudoconcerted, proton-assisted path in which the rate-determining transition state (TS1) encompasses three coupled events: deprotonation of Cys797 by Asp800, nucleophilic Michael addition of the Cys797 thiolate to the inhibitor, and protonation of the inhibitor’s carbonyl oxygen by the terminal aminium. This pseudoconcerted coupling of steps contrasts with prior stepwise proposals and yields an intermediate that is lower in free energy than the reactant yet higher than the product, thereby facilitating forward progression and rationalizing the modest barrier we compute. Overall, these findings identify inhibitor-derived proton transfer as the decisive element that renders this path superior, both thermodynamically and kinetically, to previously proposed mechanisms that omit the catalytic proton (Figure ). Novel insights emerging from this work are as follows:

  • 1.

    Catalytic role of the terminal amine. The terminal aliphatic amine of osimertinib, in its protonated (aminium) form, donates a proton to the Michael-acceptor carbonyl, thereby activating Cβ and enabling more efficient nucleophilic attack by Cys797 than in previously proposed stepwise routes.

  • 2.

    Pseudoconcerted rate-determining step. The rate-determining event proceeds as a single elementary step that integrates proton transfer between Cys797 and Asp800, nucleophilic attack by Cys797, and proton transfer between the aminium and the Michael acceptor. Within this transition state, the aminium-derived proton promotes formation of the bond between Cys797’s sulfur and the Michael acceptor’s Cβ, lowering the free energy barrier.

  • 3.

    Design-relevant residue engagement. Interaction-energy analysis of the reactant complex identifies Asp855 as the dominant stabilizing contact, proximal to the indole N1/N-methyl region of the inhibitor. Strengthening interactions at this locus, therefore, constitutes a practical design principle for next-generation EGFR covalent inhibitors.

5.

5

Open in a new tab

Integrated schematic summarizing (a) the IRC trajectory of the pseudoconcerted step and its promotion of inhibition in both thermodynamic and kinetic terms, (b) an orbital/proton transfer depiction on osimertinib, and (c) key takeaways: aminium-mediated activation of the Michael acceptor; pseudoconcerted, rate-determining TS1; and enhanced inhibitor–Asp855 contact near indole N1/N-methyl underpinning T790M selectivity.

In summary, this mechanistic picture, which highlights the proton-catalytic function of the inhibitor’s terminal aliphatic amine and the design relevance of Asp855 engagement, provides actionable guidance for structure-based optimization of covalent inhibitors.

Using ONIOM geometry optimizations and free energy calculations, we have developed a novel mechanism for covalent inhibition of wild-type EGFR by osimertinib. The computed Gibbs energy profile agrees well with experimental measurements, lending credence to the proposed reaction path. A defining feature of this path is inhibitor-derived proton catalysis: the terminal aliphatic aminium donates a proton to the Michael-acceptor carbonyl, thereby activating Cβ and accelerating Cys797 addition. To our knowledge, this role for the inhibitor’s terminal proton has not been articulated in prior EGFR-TKI work and constitutes the principal conceptual advance of the study. As such, it offers actionable design guidance for next-generation covalent inhibitors. Future studies that probe Asp800 mutations may provide a more stringent test of the proposed proton catalysis mechanism and refine these design principles. Further analyses also clarify why osimertinib shows preferential inhibition of the T790M mutant over the wild type. Electrostatic potential mapping indicates that the mutation rebiases charge within the pyrimidine–indole scaffold, increasing the positive character of the indole N-methyl substituent and strengthening electrostatic engagement with Asp855. These insights suggest that tuning the charge distributionparticularly around the indole N1/N-methyl regionmay enhance activity not only against Thr790 but also against emergent resistance variants. In future work, a more statistically robust assessmentcombining extended MD simulations to probe ligand stability and hydrogen-bonding dynamics with testing of structurally diverse inhibitors to evaluate the generality of Asp855 engagementwould substantially strengthen the broader implications of these findings. We anticipate that the mechanistic and electrostatic principles delineated here will inform the structure-based optimization of molecular targeted drugs with improved selectivity and reduced off-target effects.

Supplementary Material

ci5c02787_si_001.pdf (484.2KB, pdf)
ci5c02787_si_002.zip (21KB, zip)

Acknowledgments

This study was supported by JSPS KAKENHI (Grant Nos. JP24H02268 to K.J.F.; JP25H00429, JP25H01253, JP25H01266, JP24H00449, and JP22K21346 to T.Y.) and JST CREST (Grant No. JPMJCR21O5 to K.J.F.). We are grateful for additional support provided by the MEXT Promotion of Development of a Joint Usage/Research System Project: Coalition of Universities for Research Excellence Program (CURE) (Grant No. JPMXP1323015482) and for the CPU time granted by the Research Center for Computational Science, Okazaki, Japan (25-IMS-C083 and 24-IMS-C082). The authors thank C. Ushiro for her helpful comments on the manuscript.

Data supporting this article have been included as part of the Supporting Information.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jcim.5c02787.

  • Key interatomic distances along the intrinsic reaction coordinate (Figure S1); electronic energy along the reaction path from reactant toward intermediate (Figure S2); summary of prior computational proposals (Table S1); estimated pK a values for amino acid residues (Table S2); free energies of the reactant and TS1 for wild-type EGFR (Table S3); free energies of stationary points (Table S4); imaginary vibrational frequencies of the transition states (Table S5); IRC points and corresponding ONIOM energies (Table S6); second-order perturbation energies (Table S7); free energies of the reactant and product in T790M EGFR (Table S8); and BSSE-corrected pairwise interaction energies (Table S9) (PDF)

  • Atomic coordinates of the optimized structures (ZIP)

K.J.F. conceived and designed the project. A.K. and S.K. performed the computations. A.K., K.J.F., and T.Y. analyzed and interpreted the data. T.Y. supervised the study. A.K. and K.J.F. drafted the manuscript. T.Y., S.S., and S.K. reviewed and edited the manuscript. All authors approved the final version of the manuscript.

The authors declare no competing financial interest.

References

  1. Schaefer D., Cheng X.. Recent Advances in Covalent Drug Discovery. Pharmaceuticals. 2023;16(5):663. doi: 10.3390/ph16050663. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Lemmon M. A., Schlessinger J.. Cell Signaling by Receptor Tyrosine Kinases. Cell. 2010;141(7):1117–1134. doi: 10.1016/j.cell.2010.06.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Normanno N., Maiello M. R., De Luca A.. Epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs): simple drugs with a complex mechanism of action? J. Cell Physiol. 2003;194(1):13–19. doi: 10.1002/jcp.10194. [DOI] [PubMed] [Google Scholar]
  4. Wee P., Wang Z.. Epidermal Growth Factor Receptor Cell Proliferation Signaling Pathways. Cancers. 2017;9:52. doi: 10.3390/cancers9050052. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Ciardiello F., Tortora G.. EGFR Antagonists in Cancer Treatment. N. Engl. J. Med. 2008;358(11):1160–1174. doi: 10.1056/NEJMra0707704. [DOI] [PubMed] [Google Scholar]
  6. Sha C., Lee P. C.. EGFR-Targeted Therapies: A Literature Review. J. Clin. Med. 2024;13(21):6391. doi: 10.3390/jcm13216391. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Miller V. A., Hirsh V., Cadranel J., Chen Y.-M., Park K., Kim S.-W., Zhou C., Su W.-C., Wang M., Sun Y.. et al. Afatinib versus placebo for patients with advanced, metastatic non-small-cell lung cancer after failure of erlotinib, gefitinib, or both, and one or two lines of chemotherapy (LUX-Lung 1): a phase 2b/3 randomised trial. Lancet Oncol. 2012;13(5):528–538. doi: 10.1016/S1470-2045(12)70087-6. [DOI] [PubMed] [Google Scholar]
  8. Engelman J. A., Zejnullahu K., Gale C.-M., Lifshits E., Gonzales A. J., Shimamura T., Zhao F., Vincent P. W., Naumov G. N., Bradner J. E.. et al. PF00299804, an Irreversible Pan-ERBB Inhibitor, Is Effective in Lung Cancer Models with EGFR and ERBB2Mutations that Are Resistant to Gefitinib. Cancer Res. 2007;67(24):11924–11932. doi: 10.1158/0008-5472.CAN-07-1885. [DOI] [PubMed] [Google Scholar]
  9. Cross D. A., Ashton S. E., Ghiorghiu S., Eberlein C., Nebhan C. A., Spitzler P. J., Orme J. P., Finlay M. R. V., Ward R. A., Mellor M. J.. et al. AZD9291, an irreversible EGFR TKI, overcomes T790M-mediated resistance to EGFR inhibitors in lung cancer. Cancer Discovery. 2014;4(9):1046–1061. doi: 10.1158/2159-8290.CD-14-0337. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Ayati A., Moghimi S., Salarinejad S., Safavi M., Pouramiri B., Foroumadi A.. A review on progression of epidermal growth factor receptor (EGFR) inhibitors as an efficient approach in cancer targeted therapy. Bioorg. Chem. 2020;99:103811. doi: 10.1016/j.bioorg.2020.103811. [DOI] [PubMed] [Google Scholar]
  11. Solca F., Dahl G., Zoephel A., Bader G., Sanderson M., Klein C., Kraemer O., Himmelsbach F., Haaksma E., Adolf G. R.. Target Binding Properties and Cellular Activity of Afatinib (BIBW 2992), an Irreversible ErbB Family Blocker. J. Pharmacol. Exp. Ther. 2012;343(2):342–350. doi: 10.1124/jpet.112.197756. [DOI] [PubMed] [Google Scholar]
  12. Yuan Z., Yu X., Wu S., Wu X., Wang Q., Cheng W., Hu W., Kang C., Yang W., Li Y., Zhou X. Y.. Instability Mechanism of Osimertinib in Plasma and a Solving Strategy in the Pharmacokinetics Study. Front Pharmacol. 2022;13:928983. doi: 10.3389/fphar.2022.928983. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Liao B.-C., Lin C.-C., Yang J. C.-H.. Second and third-generation epidermal growth factor receptor tyrosine kinase inhibitors in advanced nonsmall cell lung cancer. Curr. Opin Oncol. 2015;27(2):164. doi: 10.1097/CCO.0000000000000164. [DOI] [PubMed] [Google Scholar]
  14. Thress K. S., Paweletz C. P., Felip E., Cho B. C., Stetson D., Dougherty B., Lai Z., Markovets A., Vivancos A., Kuang Y.. et al. Acquired EGFR C797S mutation mediates resistance to AZD9291 in non-small cell lung cancer harboring EGFR T790M. Nat. Med. 2015;21(6):560–562. doi: 10.1038/nm.3854. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Leonetti A., Sharma S., Minari R., Perego P., Giovannetti E., Tiseo M.. Resistance mechanisms to osimertinib in EGFR-mutated non-small cell lung cancer. Br. J. Cancer. 2019;121(9):725–737. doi: 10.1038/s41416-019-0573-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Arafet K., Scalvini L., Galvani F., Martí S., Moliner V., Mor M., Lodola A.. Mechanistic Modeling of Lys745 Sulfonylation in EGFR C797S Reveals Chemical Determinants for Inhibitor Activity and Discriminates Reversible from Irreversible Agents. J. Chem. Inf. Model. 2023;63(4):1301–1312. doi: 10.1021/acs.jcim.2c01586. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Heppner D. E., Günther M., Wittlinger F., Laufer S. A., Eck M. J.. Structural Basis for EGFR Mutant Inhibition by Trisubstituted Imidazole Inhibitors. J. Med. Chem. 2020;63(8):4293–4305. doi: 10.1021/acs.jmedchem.0c00200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Yosaatmadja Y., Silva S., Dickson J. M., Patterson A. V., Smaill J. B., Flanagan J. U., McKeage M. J., Squire C. J.. Binding mode of the breakthrough inhibitor AZD9291 to epidermal growth factor receptor revealed. J. Struct. Biol. 2015;192(3):539–544. doi: 10.1016/j.jsb.2015.10.018. [DOI] [PubMed] [Google Scholar]
  19. Kovalevsky A. Y., Liu F., Leshchenko S., Ghosh A. K., Louis J. M., Harrison R. W., Weber I. T.. Ultra-high Resolution Crystal Structure of HIV-1 Protease Mutant Reveals Two Binding Sites for Clinical Inhibitor TMC114. J. Mol. Biol. 2006;363(1):161–173. doi: 10.1016/j.jmb.2006.08.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Pajak M., Fichna J., Woźniczka M.. Protonation constants of endo- and exogenous L-amino acids and their derivatives in aqueous and mixed solution: Unraveling molecular secrets. Q. Rev. Biophys. 2024;57:e10. doi: 10.1017/S0033583524000118. [DOI] [PubMed] [Google Scholar]
  21. Capoferri L., Lodola A., Rivara S., Mor M.. Quantum Mechanics/Molecular Mechanics Modeling of Covalent Addition between EGFR-Cysteine 797 and N-(4-Anilinoquinazolin-6-yl) Acrylamide. J. Chem. Inf. Model. 2015;55(3):589–599. doi: 10.1021/ci500720e. [DOI] [PubMed] [Google Scholar]
  22. Callegari D., Ranaghan K. E., Woods C. J., Minari R., Tiseo M., Mor M., Mulholland A. J., Lodola A.. L718Q mutant EGFR escapes covalent inhibition by stabilizing a non-reactive conformation of the lung cancer drug osimertinib. Chem. Sci. 2018;9(10):2740–2749. doi: 10.1039/C7SC04761D. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Ma S., Patel H., Peeples C. A., Shen J.. QM/MM Simulations of Afatinib-EGFR Addition: The Role of β-Dimethylaminomethyl Substitution. J. Chem. Theory Comput. 2024;20(13):5528–5538. doi: 10.1021/acs.jctc.4c00290. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kaewkham O., Gleeson D., Fukasem P., Santatiwongchai J., Jones D. J. L., Britton R. G., Gleeson M. P.. Probing the Effect of Protein and Inhibitor Conformational Flexibility on the Reaction of Rocelitinib-Like Covalent Inhibitors of Epidermal Growth Factor Receptor. A Quantum Mechanics/Molecular Mechanics Study. J. Chem. Inf. Model. 2025;65(7):3555–3567. doi: 10.1021/acs.jcim.4c01985. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Voice A. T., Tresadern G., Twidale R. M., van Vlijmen H., Mulholland A. J.. Mechanism of covalent binding of ibrutinib to Bruton’s tyrosine kinase revealed by QM/MM calculations. Chem. Sci. 2021;12(15):5511–5516. doi: 10.1039/D0SC06122K. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Awoonor-Williams E., Isley Iii W. C., Dale S. G., Johnson E. R., Yu H., Becke A. D., Roux B., Rowley C. N.. Quantum Chemical Methods for Modeling Covalent Modification of Biological Thiols. J. Comput. Chem. 2020;41(5):427–438. doi: 10.1002/jcc.26064. [DOI] [PubMed] [Google Scholar]
  27. do Amaral D. N., Lategahn J., Fokoue H. H., da Silva E. M. B., Sant’Anna C. M. R., Rauh D., Barreiro E. J., Laufer S., Lima L. M.. A novel scaffold for EGFR inhibition: Introducing N-(3-(3-phenylureido)­quinoxalin-6-yl) acrylamide derivatives. Sci. Rep. 2019;9(1):14. doi: 10.1038/s41598-018-36846-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Chung L. W., Sameera W. M. C., Ramozzi R., Page A. J., Hatanaka M., Petrova G. P., Harris T. V., Li X., Ke Z., Liu F.. et al. The ONIOM Method and Its Applications. Chem. Rev. 2015;115(12):5678–5796. doi: 10.1021/cr5004419. [DOI] [PubMed] [Google Scholar]
  29. Pettersen E. F., Goddard T. D., Huang C. C., Couch G. S., Greenblatt D. M., Meng E. C., Ferrin T. E.. UCSF ChimeraA visualization system for exploratory research and analysis. J. Comput. Chem. 2004;25(13):1605–1612. doi: 10.1002/jcc.20084. [DOI] [PubMed] [Google Scholar]
  30. Webb B., Sali A.. Comparative Protein Structure Modeling Using MODELLER. Curr. Protoc. Bioinformatics. 2016;54(1):5.6.1–5.6.37. doi: 10.1002/cpbi.3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Arter C., Trask L., Ward S., Yeoh S., Bayliss R.. Structural features of the protein kinase domain and targeted binding by small-molecule inhibitors. J. Biol. Chem. 2022;298(8):102247. doi: 10.1016/j.jbc.2022.102247. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Sutto L., Gervasio F. L.. Effects of oncogenic mutations on the conformational free-energy landscape of EGFR kinase. Proc. Natl. Acad. Sci. U.S.A. 2013;110(26):10616–10621. doi: 10.1073/pnas.1221953110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Wang L., Li L., Alexov E.. pKa predictions for proteins, RNAs, and DNAs with the Gaussian dielectric function using DelPhi pKa. Proteins. 2015;83(12):2186–2197. doi: 10.1002/prot.24935. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Wang L., Zhang M., Alexov E.. DelPhiPKa web server: predicting pKa of proteins, RNAs and DNAs. Bioinformatics. 2016;32(4):614–615. doi: 10.1093/bioinformatics/btv607. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Pahari S., Sun L., Basu S., Alexov E.. DelPhiPKa: Including salt in the calculations and enabling polar residues to titrate. Proteins. 2018;86(12):1277–1283. doi: 10.1002/prot.25608. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Søndergaard C. R., Olsson M. H. M., Rostkowski M., Jensen J. H.. Improved Treatment of Ligands and Coupling Effects in Empirical Calculation and Rationalization of pKa Values. J. Chem. Theory Comput. 2011;7(7):2284–2295. doi: 10.1021/ct200133y. [DOI] [PubMed] [Google Scholar]
  37. Olsson M. H. M., Søndergaard C. R., Rostkowski M., Jensen J. H.. PROPKA3: Consistent Treatment of Internal and Surface Residues in Empirical pKa Predictions. J. Chem. Theory Comput. 2011;7(2):525–537. doi: 10.1021/ct100578z. [DOI] [PubMed] [Google Scholar]
  38. Lee C., Yang W., Parr R. G.. Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density. Phys. Rev. B. 1988;37(2):785–789. doi: 10.1103/PhysRevB.37.785. [DOI] [PubMed] [Google Scholar]
  39. Grimme S., Ehrlich S., Goerigk L.. Effect of the damping function in dispersion corrected density functional theory. J. Comput. Chem. 2011;32(7):1456–1465. doi: 10.1002/jcc.21759. [DOI] [PubMed] [Google Scholar]
  40. Chai J.-D., Head-Gordon M.. Long-range corrected hybrid density functionals with damped atom–atom dispersion corrections. Phys. Chem. Chem. Phys. 2008;10(44):6615–6620. doi: 10.1039/b810189b. [DOI] [PubMed] [Google Scholar]
  41. Smith J. M., Jami Alahmadi Y., Rowley C. N.. Range-Separated DFT Functionals are Necessary to Model Thio-Michael Additions. J. Chem. Theory Comput. 2013;9(11):4860–4865. doi: 10.1021/ct400773k. [DOI] [PubMed] [Google Scholar]
  42. Kollman P. A.. Advances and Continuing Challenges in Achieving Realistic and Predictive Simulations of the Properties of Organic and Biological Molecules. Acc. Chem. Res. 1996;29(10):461–469. doi: 10.1021/ar9500675. [DOI] [Google Scholar]
  43. Fukui K.. The path of chemical reactions - the IRC approach. Acc. Chem. Res. 1981;14(12):363–368. doi: 10.1021/ar00072a001. [DOI] [Google Scholar]
  44. Frisch, M. J. ; Trucks, G. W. ; Schlegel, H. B. ; Scuseria, G. E. ; Robb, M. A. ; Cheeseman, J. R. ; Scalmani, G. ; Barone, V. ; Petersson, G. A. ; Nakatsuji, H. ; et al. Gaussian 16, revision C.01; Gaussian Inc.: Wallingford, CT, 2016. [Google Scholar]
  45. Koda S.-i., Saito S.. Locating Transition States by Variational Reaction Path Optimization with an Energy-Derivative-Free Objective Function. J. Chem. Theory Comput. 2024;20(7):2798–2811. doi: 10.1021/acs.jctc.3c01246. [DOI] [PubMed] [Google Scholar]
  46. Koda S.-i., Saito S.. Flat-Bottom Elastic Network Model for Generating Improved Plausible Reaction Paths. J. Chem. Theory Comput. 2024;20(16):7176–7187. doi: 10.1021/acs.jctc.4c00792. [DOI] [PubMed] [Google Scholar]
  47. Administration, T. G. Attachment: Product Information: Osimertinib Mesilate; Australian Government Department of Health: Canberra, Australia, 2019. https://www.tga.gov.au/sites/default/files/auspar-osimertinib-mesilate-190822-pi.pdf. [Google Scholar]
  48. Birkholz A., Kopecky D. J., Volak L. P., Bartberger M. D., Chen Y., Tegley C. M., Arvedson T., McCarter J. D., Fotsch C., Cee V. J.. Systematic Study of the Glutathione Reactivity of N-Phenylacrylamides: 2. Effects of Acrylamide Substitution. J. Med. Chem. 2020;63(20):11602–11614. doi: 10.1021/acs.jmedchem.0c00749. [DOI] [PubMed] [Google Scholar]
  49. Fenselau R. Z., Alotbi A. R., Lee C. B., Cronin J. S., Hill D. R., Belitsky J. M., Elrod M. J.. A New Potential Atmospheric Accretion Mechanism: Acid-Catalyzed Hetero-Michael Addition Reactions. ACS Earth Space Chem. 2025;9(1):191–200. doi: 10.1021/acsearthspacechem.4c00342. [DOI] [Google Scholar]
  50. Wabnitz T. C., Yu J.-Q., Spencer J. B.. Evidence That Protons Can Be the Active Catalysts in Lewis Acid Mediated Hetero-Michael Addition Reactions. Chem. - Eur. J. 2004;10(2):484–493. doi: 10.1002/chem.200305407. [DOI] [PubMed] [Google Scholar]
  51. Moradi R., Jameh-Bozorghi S., Kadivar R., Mahdiani A., Soleymanabadi H.. Study of Mechanism Keto-Enol Tautomerism (isomeric reaction) Structure Cyclohexanone by Using Ab initio Molecular Orbital and Density Functional Theory (DFT) Method with NBO Analysis. APCBEE Proc. 2012;3:70–74. doi: 10.1016/j.apcbee.2012.06.048. [DOI] [Google Scholar]
  52. Harris T. K., Abeygunawardana C., Mildvan A. S.. NMR Studies of the Role of Hydrogen Bonding in the Mechanism of Triosephosphate Isomerase. Biochemistry. 1997;36(48):14661–14675. doi: 10.1021/bi972039v. [DOI] [PubMed] [Google Scholar]
  53. Jung Y., Lochan R. C., Dutoi A. D., Head-Gordon M.. Scaled opposite-spin second order Møller–Plesset correlation energy: An economical electronic structure method. J. Chem. Phys. 2004;121(20):9793–9802. doi: 10.1063/1.1809602. [DOI] [PubMed] [Google Scholar]
  54. Awoonor-Williams E., Rowley C. N.. How Reactive are Druggable Cysteines in Protein Kinases? J. Chem. Inf. Model. 2018;58(9):1935–1946. doi: 10.1021/acs.jcim.8b00454. [DOI] [PubMed] [Google Scholar]
  55. Zhai X., Ward R. A., Doig P., Argyrou A.. Insight into the Therapeutic Selectivity of the Irreversible EGFR Tyrosine Kinase Inhibitor Osimertinib through Enzyme Kinetic Studies. Biochemistry. 2020;59(14):1428–1441. doi: 10.1021/acs.biochem.0c00104. [DOI] [PubMed] [Google Scholar]
  56. Eyring H.. The Activated Complex in Chemical Reactions. J. Chem. Phys. 1935;3(2):107–115. doi: 10.1063/1.1749604. [DOI] [Google Scholar]
  57. Karandashev K., Xu Z.-H., Meuwly M., Vaníček J., Richardson J. O.. Kinetic isotope effects and how to describe them. Struct. Dyn. 2017;4(6):061501. doi: 10.1063/1.4996339. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Weinhold F.. Natural bond orbital analysis: A critical overview of relationships to alternative bonding perspectives. J. Comput. Chem. 2012;33(30):2363–2379. doi: 10.1002/jcc.23060. [DOI] [PubMed] [Google Scholar]
  59. Awoonor-Williams E., Rowley C. N.. Modeling the Binding and Conformational Energetics of a Targeted Covalent Inhibitor to Bruton’s Tyrosine Kinase. J. Chem. Inf. Model. 2021;61(10):5234–5242. doi: 10.1021/acs.jcim.1c00897. [DOI] [PubMed] [Google Scholar]
  60. Highbarger L. A., Gerlt J. A., Kenyon G. L.. Mechanism of the Reaction Catalyzed by Acetoacetate Decarboxylase. Importance of Lysine 116 in Determining the pKa of Active-Site Lysine 115. Biochemistry. 1996;35(1):41–46. doi: 10.1021/bi9518306. [DOI] [PubMed] [Google Scholar]
  61. Maity S., Pai K. S. R., Nayak Y.. Advances in targeting EGFR allosteric site as anti-NSCLC therapy to overcome the drug resistance. Pharmacol. Rep. 2020;72(4):799–813. doi: 10.1007/s43440-020-00131-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Tinivella A., Rastelli G.. Investigating the Selectivity of Allosteric Inhibitors for Mutant T790M EGFR over Wild Type Using Molecular Dynamics and Binding Free Energy Calculations. ACS Omega. 2018;3(12):16556–16562. doi: 10.1021/acsomega.8b03256. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ci5c02787_si_001.pdf (484.2KB, pdf)
ci5c02787_si_002.zip (21KB, zip)

Data Availability Statement

Data supporting this article have been included as part of the Supporting Information.

Articles from Journal of Chemical Information and Modeling are provided here courtesy of American Chemical Society

RESOURCES