Structural and dynamic effects of paraoxon binding to human acetylcholinesterase by X-ray crystallography and inelastic neutron scattering

Abstract

Organophosphorus (OP) compounds, including nerve agents and some pesticides, covalently bind to the catalytic serine of human acetylcholinesterase (hAChE), thereby inhibiting acetylcholine hydrolysis necessary for efficient neurotransmission. Oxime antidotes can reactivate the OP-conjugated hAChE, but reactivation efficiency can be low for pesticides like paraoxon (POX). Understanding structural and dynamic determinants of OP inhibition and reactivation can provide insights to design improved reactivators. Here X-ray structures of hAChE with unaged POX, with POX and oximes MMB4 and RS170B, and with MMB4 are reported. A significant conformational distortion of the acyl loop was observed upon POX binding, being partially restored to the native conformation by oximes. Neutron vibrational spectroscopy combined with molecular dynamics simulations showed that picosecond vibrational dynamics of the acyl loop soften in the ~20–50 cm⁻¹ frequency range. The acyl loop structural perturbations may be correlated with its picosecond vibrational dynamics to yield more comprehensive template for structure-based reactivator design.

eTOC

Gerlits et al. used X-ray crystallography to visualize conformational plasticity of the acyl loop in human acetylcholinesterase upon paraoxon and subsequent oxime reactivator binding. Using inelastic neutron scattering the study visualized softening of the acyl loop vibrational dynamics in the paraoxon-conjugated enzyme providing insights for reactivator design.

Graphical Abstract

Introduction

Organophosphorus (OP) compounds continue to pose risk to millions of people worldwide. Although over the past twenty years isolated OP nerve agent attacks still occurred, albeit infrequently and resulting in a limited number of deaths, the number of people dying from incidental or deliberate use of OP pesticides amount to hundreds of thousands every year (Gunnell et al. 2007; Dawson et al. 2010; King & Aaron 2015; Mew et al. 2017). OP compounds act by covalently binding to the catalytic Ser amino acid residue of acetylcholinesterase (AChE), a vital enzyme for the nerve signal termination through hydrolysis of a neurotransmitter acetylcholine (ACh). Inhibition of human AChE (hAChE, Figure 1a) catalytic function by OPs causes increase in ACh concentrations at synapses to dangerous levels leading to overstimulation of the peripheral and central nervous systems. To counter the effects of OPs the standard therapeutic intervention options include administering atropine as a cholinergic receptor antagonist, a benzodiazepine drug as an anticonvulsant (e.g., diazepam), and an oxime reactivator (e.g., pralidoxime or 2-PAM, and alternatively HI-6 and obidoxime) (Johnson et al. 2000; Cannard 2006; Worek et al. 2016; Dhuguru et al. 2022).

Figure 1. Human acetylcholinesterase and the studied ligands.

(A) Structure of the hAChE homodimer in cartoon representation. The acyl pocket loops (acyl loops), the Ω loops, and the four-helix bundle dimer interface regions are colored orange, purple and red, respectively. (B) Chemical structures of POX, its covalent conjugate with the catalytic Ser203, and oximes methoxime (MMB4) and RS-170B.

Many OP pesticides are phosphate triesters that contain a good, yet relatively large, leaving group that makes them non-volatile. Small leaving groups render nerve agent OPs volatile and extremely hazardous. Furthermore, most OP pesticides are symmetric having two identical alkoxy chains that remain attached to the phosphorus after it conjugates with the side-chain hydroxyl of the catalytic Ser203 in hAChE. Interestingly, hAChE conjugated with the more potent and lethal, yet chiral methylphosphonyl nerve agent OPs is reactivated more efficiently by oximes (except for the phosphoramidate tabun) than the symmetric diethylphosphorylated enzyme inhibited by some pesticides such as paraoxon (POX) or diisopropylphosphorylated by DFP. POX contains two ethoxy groups (OEt) (Figure 1b) that bind one in the choline-binding site and the other in the acyl pocket of the AChE active site located at the bottom of a 20 Å-long narrow gorge. DFP has two isopropoxy (OiPr) groups. The choline-binding site easily accommodates an OEt (or OiPr) group as it has evolved to bind the sterically bulky choline substituent of the ACh substrate, whereas the acyl pocket is not sufficiently large to fit the other OEt of POX or OiPr of DFP. To avoid steric clashes, positions of Phe295 and Phe297 in the acyl pocket significantly shift and the structural changes propagate further to cause perturbations in the acyl pocket loop (referred to as acyl loop hereafter, residues 287–299) (Figure 1a). Similar in magnitude, yet somewhat distinct in shape acyl loop perturbations were observed in DFP inhibited AChEs from Torpedo californica (TcAChE), mouse (mAChE) and human AChE in both non-aged and aged forms (reviewed in Radić 2021; Figures S1 and S2). Moreover, acyl loop-stabilized OEt group of R_P isomer of VX in hAChE and ethylmethyl carbamylated TcAChEs (in complex with oximes) show significant disturbances in the acyl loop conformations (Radić 2021). An interesting exception is X-ray structure of Novichok A-234 conjugated hAChE (PDB ID 6NTL) where in spite of OEt group stabilization in the acyl pocket along with very large substituent stabilization in the choline binding site geometry of the acyl loop remains in the native conformation (reviewed in Luedtke et al. 2021) due to ~ 1.5 Å shift of the P atom covalently bound to Ser203 towards choline binding site. Some of pM reversible inhibitors of TcAChE disturb acyl loop conformation as well (Radić 2021). Such structural perturbations of the acyl loop geometry have been proposed as a mechanism for reduced affinity and/or access of oximes to AChE inhibited by POX or DFP and consequently inefficient enzyme reactivation by oximes (Ekström et al. 2009; Radić et al. 2010; Winter et al. 2016). Furthermore, analysis of all PDB-deposited X-ray structures of TcAChE, mAChE and hAChE where acyl loop was found in a disturbed conformation showed associated shifts in α-helices located between acyl loop and dimer interface (Radić 2021; Figure 1A). Accordingly, in SAXS (Small Angle X-ray Scattering) experiments with hAChE it was demonstrated (Blumenthal et al. 2021) that POX (or R_P-OP, but not S_P-OP) binding to hAChE promotes dissociation of the reversible physiological-like enzyme homodimer, possibly due to additional structural perturbations extending to the dimer interface. Relatively small acyl pocket bound substituents such as dimethylamino group of tabun-conjugated TcAChE and also ethyl methylcarbamylated TcAChE fit in the acyl pocket without disturbance, but soaking-in of uncharged oximes stabilizes acyl loop in a different conformation (Santoni et al., 2018). Thus, change in an ensemble of alternative conformations of the acyl loop in AChEs can be triggered by OP conjugation and then stabilized by a reversible ligand at a different, likely higher energetic level.

Proteins are inherently dynamic entities, with their dynamics spanning at least fifteen orders of magnitude, from low femtoseconds (fs) corresponding to bond stretching to milliseconds signifying large conformational changes and induced fit of ligands (Henzler-Wildman et al. 2007). Within this timescale, the protein collective vibrational dynamics occurring in low picoseconds (< 10 ps) occupy a special place. At frequencies of 5–50 cm⁻¹, corresponding to ~0.7–7 ps or ~0.15–1.5 THz, there exist protein breathing vibrational motions encompassing a whole protein molecule, and the motions of the protein secondary structure elements, including collective vibrations of α-helices, β-sheets and loops (Chou 1988; Cusack et al. 1988). Neutrons are ideally suited to probe the protein dynamics employing elastic, quasi-elastic and inelastic neutron scattering (INS) techniques (Schiro, 2020). Several investigations have been carried out to study hAChE dynamics using neutrons. Elastic incoherent neutron scattering experiments under pressure identified a molten globule state of hAChE at 1750 bar which possessed increased flexibility compared to the enzyme state at atmospheric pressure (Marion et al. 2015). In additional studies, it was shown by quasi-elastic neutron scattering that binding of a non-covalent ligand huperzine A led to the softening of low frequency modes (i.e., shifting the frequencies to lower energies) (Saouessi et al. 2019), and a global effect of sucrose on quasi-elastic neutron scattering of mAChE was explored (Luschekina et al., 2020). Conversely, the complex of hAChE with the aged soman was found to be more rigid, with stiffer collective vibrational modes (i.e., shifting the frequencies to higher energies), than the apo enzyme, as revealed by the INS measurements (Peters et al. 2016). None of these studies, however, pinpointed specific regions of hAChE molecule where vibrational dynamics was affected by a bound ligand. We recently employed a combination of INS and molecular dynamics (MD) simulations to identify the secondary structure elements that underpin the changes in the collective vibrational dynamics of aspartate aminotransferase and HIV-1 protease upon a substrate analog and tight inhibitor binding, respectively (Dajnowicz et al. 2020; Kneller et al. 2022). Interestingly, our studies and those by various research groups indicate that ligand binding can either stiffen or soften protein vibrational dynamics (Dajnowicz et al. 2020; Kneller et al. 2022; Balog et al. 2004; Peters et al. 2016; Niessen et al. 2017). The observed dynamical changes are protein- and ligand-dependent, with no universal rule to predict the effect of a specific ligand on protein dynamics.

In this study, we present X-ray crystal structures of hAChE in complex with unaged POX: a binary complex POX-hAChE, and two ternary complexes with bound oxime reactivators MMB4 (POX-hAChE-MMB4) and RS170B (POX-hAChE-RS170B) at 2.6–2.8 Å resolutions. We also report a 2.2 Å crystal structure of unmodified hAChE in complex with MMB4 (hAChE-MMB4). MMB4 (Figure 1b) is a bispyridinium bis-oxime whose chemical structure is similar to those of obidoxime, TMB-4, and other bis-oximes (Kovarik et al. 2008; Musilek et al. 2011). MMB4 demonstrates superior reactivation properties compared to 2-PAM (Luo et al. 2010; Shih et al. 2012; Radić et al. 2013) and to RS170B (Figure 1b) which is an imidazole monopyridinium oxime reactivator (Sit et al. 2014; Kovarik et al. 2015). In the POX-hAChE structure, we detected a significant conformational distortion of the acyl loop upon binding of POX when compared to the unmodified hAChE structure (apo-hAChE, PDB ID 4EY4, Cheung et al. 2012, Figure 2a). However, the acyl loop geometry in POX-hAChE structure we determined is radically different from that observed in the earlier hAChE complex with POX (PDB ID 5HF5; Franklin et al. 2016), where the difference may be driven mainly by a flip in Arg296 position in the latter structure that kicks out Trp286 from the base of the peripheral anionic site (PAS), restructuring a stretch of residues from Val288 to Phe295 so that Pro290 is positioned at the site previously occupied by Arg296. Binding of MMB4 or RS170B to POX-modified hAChE appears to push the acyl loop conformation towards that found in the apo-enzyme but does not restore it completely to the native conformation. By combining INS spectra measured from the D₂O-hydrated powders of apo-hAChE and POX-hAChE with MD simulations we determined that the picosecond vibrational dynamics of the acyl loop soften in the ~20–50 cm⁻¹ frequency range. Our findings illustrate that the structural perturbations of the acyl loop found in the crystal structures may be correlated with the changes in its dynamic behavior in the sub-to-low THz frequency regime.

Figure 2. Covalent binding of POX to hAChE distorts the acyl loop conformation.

(A) Conformational distortion of the acyl loop geometry in POX-hAChE compared to the apo-hAChE structure (PDB ID 4EY4). (B) A close-up view of the active site in POX-hAChE complex. The 2F_O-F_C electron density map is contoured at 1.0 σ level. Hydrogen bonds are indicated by blue dashed lines and the distances are in Å. (C) Superposition of the current POX-hAChE structure (green carbon atoms and light teal cartoon) with the previously published POX-hAChE (PDB ID 5HF5; light pink) demonstrating significant differences in the acyl loop conformations (black arrows).

Results

POX inhibition of hAChE distorts the acyl loop geometry

The hAChE protein used in this study lacks glycosylation at Asn350 as described previously (Gerlits et al. 2019a), resulting in the enzyme’s inability to crystallize in the ligand-free or apo-form. Consequently, to obtain the crystal structures of hAChE with bound POX, we employed the strategy that was successfully applied to determine structures of hAChE in complex with VX (Gerlits et al. 2019b). Specifically, a crystal of the binary complex with reactivator RS194B (Gorecki et al. 2020) was soaked with POX for 1 hour resulting in OP-conjugation of Ser203 whereas RS194B molecule completely dissociated from the enzyme active site gorge. Thus, the resulting structure was of the binary POX-hAChE determined at 2.8 Å resolution. We focus on describing the complex in chain B as the electron density in the active site gorge and the acyl loop is of better quality than in chain A.

In the POX-hAChE structure, there is unambiguous electron density at the catalytic Ser203 indicating conjugation with unaged POX (Figure 2b and S3). One OEt group of POX faces the choline-binding site but is rotated away from Trp86, because a glycerol molecule occupies the bulk of the site being stacked against the π-system of the Trp86 indole side chain. The presence of glycerol at this location is not unusual as it was observed in several low-temperature X-ray structures we have determined (Gerlits et al. 2019b : PDB IDs 6O5V, 6O66; Gorecki et al. 2020: PDB IDs 6U37). The other OEt group of POX is inserted into the mainly hydrophobic acyl pocket flanked by Trp236, Phe295, Phe297, Phe338, Val407, and the catalytic triad His447, with Phe295 and Phe297 belonging to the acyl loop. The unsubstituted phosphate’s oxygen is stabilized in the oxyanion hole, making two short hydrogen (H) bonds of 2.6 and 2.8 Å with the main-chain amide NH of Gly121 and Gly122, respectively, and one longer 3.1 Å H bond with the main-chain NH of Ala204. This is consistent with the coordination of POX in the previously published unaged POX-hAChE structure (Franklin et al. 2016). There is no electron density that would indicate presence of RS194B inside the active site gorge. Instead, we modeled a DMSO molecule clearly visible in the electron density and located about half-way between the Ser203-POX conjugate and Trp286 of the PAS (Figure 2b).

To understand the structural effect of POX binding to the catalytic Ser203 we superimposed POX-hAChE with the apo-hAChE structure determined previously (Cheung et al. 2012). It is apparent that to avoid a steric clash between OEt group positioned in the acyl pocket and Phe295, the latter moves away by ~1.8 Å. Phe297 is also pushed but by only ~0.8 Å (Figure 2a). As a consequence, the entire acyl loop geometry is reshaped, with Pro290 moving by as much as 5.9 Å and Ser293 by 4.4 Å towards Trp286 of the PAS so that Ser293 Cα is only 4 Å away from the Trp286 indole side chain in POX-hAChE. The observed acyl loop conformational change leads to the narrowing of the active site gorge entrance by ~ 4 Å, which would block oxime reactivators from reaching to the phosphorus atom of Ser203-POX conjugate. It is important to note that the kind of deviated acyl loop conformation found in the current POX-hAChE structure is somewhat unique. A markedly different organization of the loop (Figure 2c) was observed in the previously published POX-hAChE complex (Franklin et al. 2016), where Arg296 makes a dramatic flip relative to its location in the current POX-hAChE and apo-hAChE structures. The Cα atoms of Phe295 and Phe297 move accordingly with Arg296, but the shifts are smaller, 0.9 and 2.9 Å respectively relative to the current POX-hAChE. In the published POX-hAChE, the side chain of Arg296 invades the position occupied by the Trp286 side chain, forcing its indole to rotate ~180 ° to avoid the steric clash. Interestingly, the side chains of Phe295 and Phe297 are not greatly affected by the Arg296 flip, their phenyl groups occupying essentially the same positions as in the current POX-hAChE structure. The conformation of the 290–294 residue stretch is also very different in the two POX-hAChE structures. For example, the corresponding Cα atoms of Ser293 and Pro290 are 3.6 and 6.1 Å apart, respectively, when the two POX complexes are compared.

Acyl loop distortions were reported also for other AChEs (mAChE and TcAChE) and other covalent inhibitors (DFP, tabun and ethyl methyl carabamate). Distortion of the Phe295-Arg296-Phe297 acyl loop fragment, with large side-chain flip of the Arg, is also seen in DFP-AChE conjugates of both mAChE (PDB ID 2JGI -nonaged; PDB ID 2JGM -aged) and TcAChE (PDB ID 2DFP – aged) (Figures S1 and S2B and S2C). An alternative loop conformation, with no flip of Arg was reported for one of aged DFP-mAChE structures (PDB ID 5HCU). Both tabun-inhibited TcAChE and methyl ethylcarbamylated TcAChE have conformations identical to non-liganded TcAChE. However, reversible binding of oximes has capacity to stabilize a distorted acyl loop in both conjugate types (Figures S2D and S2E) without side chain flips of the intervening Arg (PDB IDs: 1GQR, 6FLD, 6FQN). Even formation of a non-covalent adduct with tight binding reversible inhibitor tacrine-benzofurane results in distortion of the acyl loop conformation without any additional disturbances of the TcAChE backbone (Figure S2F). The acyl loop therefore has a high degree of conformational flexibility and can adopt various conformations in crystals of hAChE, and similarly in mAChE and TcAChE.

Oxime reactivator partially restores the acyl loop geometry in POX-hAChE

To illustrate the modes of oxime reactivator binding to POX-inhibited hAChE and to find out whether unique distortion of the acyl loop conformation that we observed here can be restored due to presence of a reactivator in the active site gorge we determined the X-ray crystal structures of POX-hAChE-MMB4 and POX-hAChE-RS170B ternary complexes at 2.8 and 2.6 Å resolutions. There is strong electron density at the catalytic Ser203 in both complexes indicating full occupancy of the POX conjugate. For the description and analysis of the structures, we focus on chain A in POX-hAChE-MMB4 and chain B in POX-hAChE-RS170B because the electron densities for the reactivators are of better quality, and then compare these ternary complexes to our current POX-hAChE and published apo-hAChE structures (Cheung et al. 2012).

Similar to the current POX-hAChE, the POX conjugate in POX-hAChE-MMB4 makes three hydrogen bonds, two with Gly121 and Gly122 with the distances of 2.8 Å and one with His447 with the distance of 3.1 Å (Figure 3a and S3). Half of MMB4 is inserted into the active site gorge, with the pyridinium ring positioned in the PAS above Trp286 indole, making 3.6–4.0 Å π-π stacking interactions. The oxime group is directed towards the phosphorus atom of Ser203-POX conjugate, but is ~7 Å away unable to reach beyond the ‘choke point’ created largely by the Tyr124 phenolic side chain. The oxygen and nitrogen atoms of the oxime moiety are > 4 Å away from the Tyr124 phenolic oxygen. The other identical half of the reactivator remains outside and is disordered, as indicated by poorer electron density. MMB4 binds in an identical fashion to hAChE in the complex with a substrate analog 4K-TMA with similar distances to Trp286, Tyr124 and the central tetrahedral carbon (instead of phosphorus) of 4K-TMA conjugate (PDB ID 7RB5, Gerlits et al. 2021). The structure of 4K-TMA-hAChE-MMB4 complex was determined at room temperature, whereas X-ray diffraction data for POX-hAChE-MMB4 were collected at 100K. Thus, neither the temperature nor the binding of POX had a noticeable effect on the position of MMB4 in the hAChE active site gorge. Admittedly, in our MMB4 structures choline binding site is occupied, either by glycerol or by choline part of 4K-TMA, possibly preventing access of the oxime to the gorge base even if it were able to traverse the ‘choke point’. Widening of the ‘choke point’ (in Y124A mAChE mutants) or of the opening at the top of the gorge (in W286A mAChE mutants) reportedly enhanced reactivation rates by selected bis-quaternary oximes (Katalinić et al., 2018).

Figure 3. Effect of oxime binding to POX-conjugated hAChE on the acyl loop geometry.

(A) A close-up view of the active site in POX-hAChE-MMB4 complex and (B) in POX-hAChE-RS170B. The 2F_O-F_C electron density map is contoured at 1.0 σ level. Hydrogen bonds are indicated by blue dashed lines and the distances are in Å. Val294 side chain is omitted for clarity. (C) Superimposition of POX-hAChE-MMB4 (green carbons and cartoon) and POX-hAChE-RS170B (light purple carbons and cartoon) structures with our POX-hAChE structure (colored orange) and with the published apo-hAChE structure (PDB ID 4EY4; colored violet).

In POX-hAChE-RS170B, the geometry of the Ser203-POX conjugate and the interactions it makes with the catalytic His447 and the oxyanion hole are preserved relative to the other complexes containing conjugated POX (Figure 3b and S3). A glycerol molecule is again seen occupying the choline-binding site. RS170B binds mainly in the PAS, with the pyridinium ring and the alkyl spacer located above Trp286 side chain indole resulting in the close 3.6–3.9 Å π-π stacking and 3.4–3.6 Å C-H-π interactions. The imidazole group of RS170B pushes against the phenolic oxygen of Tyr124 and is rotated so that the oxime moiety is turned away from the phosphorus of Ser203-POX towards the acyl loop to form an H bond with the main chain amide NH of Phe295. Interestingly, the RS170B mode of binding is essentially identical to that found in the room-temperature structure of the binary hAChE-RS170B complex (Gerlits et al. 2019b). Therefore, RS170B is observed in the non-productive orientation in POX-hAChE-RS170B but was found to reorient through a ~ 180° flip of the imidazole ring to bring the oxime moiety closer to the phosphorus atom in VX-hAChE-RS170B complex (Gerlits et al. 2019), from ~ 10 Å (in POX-hAChE) to ~ 6.4 Å (in VX-hAChE). The non-productive orientation of RS170B in the complex with POX may partially be the reason for its order of magnitude lower reactivation potency towards POX than VX (Sit et al. 2014).

It is important to analyze whether oxime reactivator binding to POX-modified hAChE restores the acyl loop geometry back to its original conformation found in apo-hAChE. Hence, we superimposed POX-hAChE-MMB4 and POX-hAChE-RS170B structures with the published apo-hAChE structure (Cheung et al. 2012). As seen in Figure 3c, there is a tendency for the acyl loop to return to the original conformation before POX bound to Ser203; however, the complete restoration of the apo-hAChE -like geometry is not achieved. Residues 292–299 are closest to their original positions in both ternary complexes, but in the POX-hAChE-MMB4 Ser293 has shifted back by 3.1 Å, with >1 Å left to go. The rest of the acyl loop residues (residues 287–291), although moved significantly from their locations in the current POX-hAChE structure upon oxime binding, are nonetheless far away from their original positions. For example, the oxime presence pushes Gln291 by 2.5 Å from its position in the current POX-hAChE, but this residue is still 3.8–3.9 away from where it was in apo-AChE. Surprisingly, Phe295 and Phe297 also shift towards their positions in apo-hAChE when oximes bind, even in the presence of Ser203-POX conjugate. Although Phe297 moves very slightly and perhaps insignificantly by 0.2–0.4 Å considering the structures were obtained at moderate resolutions, Phe295’s shift is significant, 0.7–1.1 Å, with the larger shift corresponding to the binding of RS170B.

MMB4 binding does not perturb the acyl loop

We have also sought to establish the binding mode of MMB4 to the uninhibited hAChE and to confirm that MMB4 does not perturb the acyl loop conformation in the absence of POX modification at Ser203. Thus, we determined a 2.2 Å resolution X-ray structure of the binary hAChE-MMB4 complex. There is clear electron density for the bispyridinium methylene moiety and the oxime group inserted into the active site gorge in both chain A and B; the other oxime group facing away from the protein surface is poorly defined in the electron density. MMB4 binds identically to both hAChE chains and the intermolecular interactions are very similar. Henceforth, we focus our structural analysis on chain A (Figure 4a and S3).

Figure 4. Binding of MMB4 to hAChE.

(A) A close-up view of the active site in hAChE-MMB4 complex. The 2F_O-F_C electron density map is contoured at 1.0 σ level. Hydrogen bonds are indicated by blue dashed lines and the distances are in Å. (B) Superimposition of hAChE-MMB4 (magenta carbons and light blue cartoon) with our POX-hAChE (colored orange) structure and the published apo-hAChE structure (PDB ID 4EY4; colored violet).

MMB4 mostly occupies the PAS with one of the pyridinium moieties stacked against Trp286 indole side chain. The central CH₂ group makes C-H…π contacts as close as 3.3 Å and the pyridinium group interacts with the indole by means of π-π contacts with the distances of 3.6–3.8 Å. The MMB4 positions are virtually identical in both hAChE-MMB4 and POX-hAChE-MMB4 complexes. Thus, POX conjugation does not alter the oxime’s binding mode. In hAChE-MMB4, the active site gorge is hydrated by harboring six water molecules (Figure 4a). Of these, two water molecules create a tight hydration shell for the Ser203 hydroxyl, with the hydrogen bond distances of 2.4 and 2.5 Å. Two other waters hydrate the MMB4 oxime group, with the hydrogen bond distances of 2.9 and 3.3 Å. In addition, the oxime group has long-range contacts with Ser203 hydroxyl mediated by a chain of three water molecules. A glycerol molecule occupies the choline-binding site similar to the other crystal structures. hAChE-MMB4 superimposes on the apo-hAChE complex with the RMSD of 0.6 Å. The acyl loops have identical conformations in both complexes, indicating that the oxime binding does not perturb the acyl loop geometry (Figure 4b).

Acyl-loop vibrational dynamics softens upon POX binding

To link the structural results to specific dynamics changes, we subjected both wild-type and POX-conjugated hAChEs to INS measurements to characterize how POX binding affects the collective dynamics in the hAChE molecule. Neutron scattering is ideally suited to studying changes in protein structure and dynamics due to ligand binding, complementing crystallographic studies (Dajnowicz et al. 2020; Kneller et al. 2022). Using INS, the vibrational density of states (VDOS, g(ω)), for hAChE can be determined, allowing for thermodynamic quantities associated with the vibrational change to be derived (Kneller et al. 2022). The integrated dynamic structure factor S(q, ω) shows a distinct difference in the inelastic peak (~40 cm⁻¹) (Figure 5) between apo-hAChE and POX-hAChE. To quantify the difference in the inelastic peak heights, we computed the vibrational density of states g(ω). The g(ω) are very similar at the low wavenumbers (< 30 cm⁻¹) which correspond to collective protein breathing dynamics but differ at higher wavenumbers (30–50 cm⁻¹) indicating some differences to local collective vibrational dynamics of the secondary structure elements. The vibrational density of states suggest shift in the frequencies of some local motions to lower energies upon POX-conjugation.

Figure 5. Inelastic neutron spectra and vibrational density of states for apo-hAChE and POX-hAChE.

Experimental dynamic structure factor versus frequency S(q, ω) for apo-hAChE (blue curve) and paraoxon-conjugated hAChE (black curve) at 120 K (left). Data from all scattering angles are summed. Vibrational density of states, g(ω), for apo-hAChE (blue curve) and paraoxon-conjugated hAChE (black curve) (right). Data shown represent mean of the inelastic neutron scattering signal integrated over 0.31 to 5.19 Å scattering angle q, and error represents SD of the signal within this q range at each specific frequency.

To analyze the vibrational changes in hAChE upon POX-conjugation, we performed a series of in silico impulse simulations. Three local motions were selected for testing based on their functional importance and structural differences between the apo and POX-bound hAChE: the acyl loop, the four-helix bundle at the dimer interface and the omega loop (residues 69–96) (Figure 1). An impulse along the linear interpolation between the two crystal structure conformations was used as an approximation of the motion between these two conformations. The same magnitude of these impulses was applied to the model powder systems for both apo- and POX-hAChEs, thus allowing direct comparison of the power spectra. The acyl loop impulse shows an increase in vibrational modes upon POX conjugation (Figure 6) that corresponds with the frequency range (~20–50 cm⁻¹) observed in g(ω) similar to the observed changes in the INS experiment (Figure 5). Conversely, the four-helix bundle impulse shows a decrease in the lower frequency vibrational modes upon POX-conjugation in the frequency range of ~10–20 cm⁻¹. In addition, the omega loop impulse is unchanged upon POX-conjugation. These results indicate that the collective vibrational motions of the acyl loop in the frequency range of ~20–50 cm⁻¹ soften, whereas those of the four-helix bundle stiffen and those of the omega loop remain unchanged, upon catalytic Ser203 modification by POX in the hAChE active site. We emphasize here that other secondary structure elements of hAChE, not considered in our simulations, may also contribute to the increase in the vibrational modes observed experimentally by INS. The vibrational dynamics of all secondary structure elements sum up to the changes seen in the INS spectra upon POX binding.

Figure 6. Vibrational density of states from MD simulations.

Fluctuation power spectra of the apo-hAChE and the POX-hAChE calculated from MD simulations with impulse forces applied to (A) the acyl-loop (black) and the four-helix bundle (red), and (B) the omega-loop (blue). Data shown represent mean of five independent MD simulations of 8 hAChE monomers each resulting in 40 power spectra, and error represents SD n = 40.

Discussion

By combining static and dynamic biophysical approaches in the analysis of the structure of human AChE we reveal unique properties of the acyl loop region( residues 287–299) within the enzyme’s catalytic subunit. Backbones of the catalytic subunits from hundreds of PDB-deposited X-ray structures of AChEs reveal only single, largely “closed” protein conformation (Radić and Taylor, 2006; Siman and Sussman, 2008; Xu et al.,2008). In apparent contradiction, the resulting narrow 20 Å long active center gorge channel has to be traversed by a tight-fitting substrate molecule in a microsecond catalytic timeframe. A distinct “open” conformation, observed for example upon large Ω loop lift in X-ray structures of similarly folded α/β-hydrolase, Candida rugosa lipase (Grochulski et al. 1993), was not seen in AChEs. The acyl loop has been the only AChE region found in multiple, distinct conformations and with capacity to affect the size of the active center gorge opening. The X-ray structures that we contribute here add to only 14 PDB-deposited structures with acyl loop conformations diverging from the ones in apo-AChEs (PDB IDs: 5HF5, 5HF8, 5HF6, 6CQX, 2JGI, 2JGM, 5HCU, 2DFP, 6G4O, 6G4P, 1GQR, 6FLD, 6FQN, 4W63). In all but 4W63, this structural diversity was associated with covalent OP inhibition where either ethoxy- or isopropoxy- substituents on conjugated phosphorus of paraoxon, DFP and R_P isomer of VX, point to the acyl pocket. In addition, acyl loops of tabun and ethylmethyl carbamate-inhibited AChEs diverge from apo-AChE conformations as well, but only in the presence of reversibly bound oxime ligands (Santoni et al. 2018). The conformational flexibility of the acyl loop, regulating the size of the active center gorge opening, thus seems critical for maintaining fast catalytic turnover of acetylcholine, but could also affect single-cycle transphosphorylation in the nucleophilic antidote reactivation of the OP-conjugated AChE.

A distinct, previously unobserved deviation of the acyl loop conformation that we report here in the X-ray structure of POX-conjugated hAChE reverts, but only in part, to the apo-hAChE conformation in the X-ray structures of oximes RS-170B or MMB4 in individual complexes with POX-hAChE. Reversals of the distorted acyl loop reported previously in complexes of POX-hAChE with 2PAM and HI6 were, however, practically complete to match the loop conformation in Apo-hAChE (Franklin et al. 2016). In our structures, the most displaced residue of the acyl loop was not Arg296 that exhibited a drastic sidechain flip in all published structures of nonaged and aged conjugates of POX- and DFP-inhibited hAChE, TcAChE and mAChE as well as in R_PVX-inhibited hAChE. In our POX-hAChE structure, Arg 296 kept its native position, while Pro290 and Ser293 of the outer rim of the loop shifted most. This, at least, illustrates diversity of conformational ensemble that acyl loop has capacity to assume, and was similarly observed in TcAChE, mAChE and hAChE. Compensatory effects of reversibly bound oximes in covalently inhibited AChEs that revert acyl loop conformation, indicate that “ligand stabilization” rather than “induction by ligand” is the likely cause of observed conformational diversity. Thus both, reversal of already distorted acyl loop conformations of OP-AChEs back to its native conformation upon addition of oximes (RS-170B and MMB4, observed here, for example) as well as, appearance of a new acyl loop conformation upon reversible binding of selected oximes to tabun or ethylmethyl carbamyl inhibited TcAChE (Santoni et al. 2018), may indicate stabilization of already existing acyl loop conformations, rather than generating new ligand-induced ones. Since X-ray structures represent only static snapshots of dynamic proteins, we turned to a biophysical technique capable of detecting vibrational dynamic properties of a macromolecular backbone on a low-picosecond timescale. Assuming that AChE domains with documented conformational diversity, such as acyl loop, should upon OP inhibition also exhibit altered vibrational characteristics (i.e., the vibrational density of states, VDOS, g(ω)), we compared vibrational spectra from the apo-hAChE and from POX-inhibited hAChE collected using the INS technique in the frequency range of 5–50 cm⁻¹.

The vibrational characteristics of the apo- and POX-bound hAChEs were analyzed here in a region-specific manner. Simulation of the vibrational spectra with molecular dynamics (MD) allowed us to identify the region(s) in hAChE that contributed to the differences in the INS spectra and provided connection to our X-ray structural data. Impulse simulations of the hAChE acyl loop domain, unlike those of the Ω-loop or the four-helix bundle regions, showed an increase in VDOS upon POX conjugation consistent with the VDOS change calculated using the INS measurements, indicating softening of the acyl loop collective vibrational dynamics. We therefore conclude that the acyl loop remains as a most likely source of functionally important conformational flexibility in hAChE. This conclusion is consistent with the multitude of X-ray structural data indicating the acyl loop in AChEs is capable of assuming the largest diversity of conformations in different structures. Consequently, we suggest that, from a pre-existing ensemble of acyl loop conformations of the apo-hAChE, either a covalent ligand binding in the acyl pocket or possibly also reversible binding on the AChE surface (adjacent to the peripheral site) could lead to stabilization of non-apo acyl loop conformations with altered VDOS. Contribution by the Ω-loop to the breathing motions of the active center gorge opening suggested in early MD simulations of AChEs (Wlodek et al., 1997; Shi et al.,2003; Xu et al.,2008) seems less likely. Our impulse simulations of Ω-loop and observed changes in INS spectra upon covalent binding of POX do not support that role of the Ω-loop. The acyl loop is thus emerging as a dominant structural region for both dynamic and static control of AChE catalysis and its specificity (Vellom et al. 1993; Radić et al.,1993). Its stabilization in a closer-to-apo conformation by surface ligand binding (of the kind observed in PDB ID structures 6G4O, 6G4P) could promote and accelerate nucleophilic oxime reactivation of POX-conjugated hAChE. The effect of reversibly bound oximes (MMB4, RS170B, HI6 and 2PAM) to largely revert the acyl loop conformation of POX-hAChE to the one found in apo-hAChE, is however initiated from within the AChE gorge. The approach of an oxime nucleophile to conjugated phosphorus atom and progress towards nucleophilic transition state in reactivation reaction thus appears already compromised. Finding a suitable ligand that would revert acyl loop conformation from the outside the gorge would have more beneficial effect on reactivation process by allowing unobstructed access to the conjugated phosphorus directly through the narrow active center gorge opening. In this manner chance of finding a broad range, more universal reactivator of structurally diverse OP-hAChE conjugates could be enhanced. Similarity in sites of binding and orientation for the same oxime antidote (MMB4 or RS170B) to both apo- and OP-conjugated hAChEs at distances non-productive for nucleophilic attack to phosphorus reflect their sub-optimal interaction concerning reactivation. Our structural and vibrational analysis of hAChE contributes comprehensive evidence for an improved template in reactivator antidote design. It is suggesting that avenues of improvement of structure/activity properties for nucleophilic antidotes towards accelerated reactivation remain wide open.

STAR Methods

Resource Availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Andrey Kovalevsky (kovalevskyay@ornl.gov).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• X-ray crystallographic data have been deposited at PDB and are publicly available as of the date of publication. Accession numbers are listed in the key resources table.

• In-house Python code to calculate vibrational density of states from MS simulations created in this study is available from the authors.

• Any additional information required to reanalyze the data reported in this paper is available from the lead author upon request.

Key resources table.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Chemicals, peptides, and recombinant proteins
Paraoxon	Millipore-Sigma	Cat#36186
MMB4	Millipore-Sigma	Cat#SML1059
RS170B	Sit et al., 2014	N/A
Deposited data
POX-hAChE	This work	PDB: 8DT2
POX-hAChE-MMB4	This work	PDB: 8DT4
POX-hAChE-RS170B	This work	PDB: 8DT5
hAChE-MMB4	This work	PDB: 8DT7
Experimental models: Cell lines
Stably transfected Gnt1- HEK293 cells	Gerlits et al., 2019	N/A
Software and algorithms
HKL3000	Minor et al., 2006	https://hkl-xray.com/hkl-3000
CCP4	CCP4, 1994	https://www.ccp4.ac.uk/
Phenix	Liebschner et al., 2019	https://phenix-online.org/download/
NAMD2.9	Phillips et al., 2005	https://www.ks.uiuc.edu/Research/namd/
CGenFF	Vanommeslaeghe et al., 2010	https://cgenff.umaryland.edu/
In-house Python code for MD simulations analysis	Available from the authors	N/A

Experimental model and subject details

Recombinant hAChE was purified from the expression in the Gnt1- HEK293 mammalian cell culture.

Method details

Chemicals and supplies

Imidazole-based aldoxime RS-170B was synthesized as described earlier (Sit et al. 2014). MMB4 and POX were purchased from Millipore-Sigma (St. Louis, MO, USA). Each compound was dissolved in DMSO at 100 mM concentration. Protein purification supplies were purchased from GE Healthcare (Piscataway, New Jersey, USA). Crystallization reagents were purchased from Hampton Research (Aliso Viejo, California, USA). Crystallographic supplies were purchased from MiTeGen (Ithaca, New York, USA) and Molecular Dimensions (Maumee, OH).

Expression and purification

The recombinant hAChE containing the N-terminal FLAG tag was expressed in the Gnt1- HEK293 mammalian cell culture deficient in complex N-glycans as previously described (Cochran et al., 2011). In short, stably transfected Gnt1⁻ HEK293 mammalian cell culture deficient in complex N-glycans was used for expression. Cells were grown at 37 °C and 10% CO₂ in Dulbecco’s modified Eagle’s medium, containing 10% fetal bovine serum. The enzyme was eluted from the anti-FLAG affinity column by specific human rhinovirus (HRV) 3C proteolysis (Sinobiological.com), cleaving the FLAG tag off at the engineered PreScission protease recognition site. Resulting N-terminal sequence of the pure eluted hAChE was G-P-L-E-G-R-… where amino acid sequence of the mature hAChE protein starts at E-G-R-… and ends at the truncated C-terminus with sequence …-S-A-T-D-T-L-D⁵⁴⁷. Physiologically relevant hAChE dimers spontaneously and reversibly associate in solution from the expressed monomers. The expressed hAChE lacks glycosylation at Asn350.

Crystallization

Due to lack of Asn350 glycosylation, we are unable to grow apo-hAChE crystals (Gerlits et al. 2019a). For crystallization experiments a sample of hAChE was dialyzed in 10 mM HEPES, pH 7, 10 mM NaCl, and concentrated to 6–10 mg/mL. To generate POX-hAChE complex, we first grew crystals of hAChE-RS194B using previously published methodology (Gorecki et al. 2020). Briefly, the hAChE solution was combined with a 100 mM DMSO stock solution of RS194B at a molar ratio of 1:10 and the crystals were grown by vapor diffusion at 10 °C in sitting drops using Hampton Research microbridges. The well solution contained 100 mM HEPES, pH 7.5, 10 – 20 mM sodium citrate, and 7–8.5 % PEG6000. Crystals of hAChE complex with RS170B were obtained as previously published (Gerlits et al. 2019b), whereas crystals of hAChE-MMB4 were grown in a similar fashion using 100 mM HEPES, pH 7.5, 100 mM KNO₃ and 9 % PEG3350 as the well solution. Crystals of the hAChE-RS194B, hAChE-RS170B and hAChE-MMB4 complexes were then used for soaking experiments with POX.

Soaking experiments

For the soaking experiments 3 μL of 100 mM POX solution in ethanol were diluted with 400 μL of the solution from a chosen crystallization well. Crystals of hAChE-RS194B and hAChE-RS170B were soaked in a POX diluted with 100 mM HEPES, pH 7, 10 mM sodium citrate and 7 % PEG6000 well solution. For soaking of hAChE-MMB4 crystals well solution containing 100 mM HEPES, pH 7.5, 100 mM KNO₃ and 9 % PEG3350 was used. The soaking drop (10 μL of the corresponding POX solution) was supplemented with either 0.5 μL of RS170B or MMB4 stock solution. All crystals were soaked for 4 min at RT (~22 °C).

X-ray diffraction data collection and refinement

For all complexes reported here, X-ray crystallographic data were collected from frozen crystals at 100 K. Prior to data collection crystals were subjected to two very brief consecutive soaks in the cryoprotectant solutions, first in 12.5% glycerol followed by 25% glycerol, and then flash cooled by plunging into liquid nitrogen. Diffraction data were collected from a single crystal for each complex on the ID19 beamline at SBC-CAT using a Pilatus3 X 6M detector at the Advanced Photon Source (APS). X-ray diffraction data were integrated and scaled using the HKL3000 software suite (Minor et al. 2006). The structures were solved by molecular replacement using the CCP4 suite (CCP4 1994). The structure of the hAChE-RS194B complex (PDB ID 6U34) (Gorecki et al. 2020) was used as a starting model with all waters. Refinement was performed using the phenix.refine program in the PHENIX (Liebschner et al., 2019) suite and the resulting structure analyzed with molprobity (Chen et al., 2010). The structures were built and manipulated with program Coot (Emsley et al., 2010; Casañal et al., 2020). Figures were generated using the PyMol molecular graphics software (v.2.3.2; Schrödinger LLC). A summary of the crystallographic data and refinement is given in Table 1. Crystallographic data have been deposited to the PDB with the following codes: 8DT2 for POX-hAChE, 8DT4 for POX-hAChE:MMB4, 8DT5 for POX-hAChE-RS-170B, and 8DT7 for hAChE-MMB4 structures.

Table 1.

X-ray crystallographic data collection statistics and structure refinement.

	POX-hAChE	POX-hAChE-MMB4	POX-hAChE-RS170B	hAChE-MMB4
	PDB ID 8DT2	PDB ID 8DT4	PDB ID 8DT5	PDB ID 8DT7
Data collection:
Beamline/Facility	ID19, APS, ANL	ID19, APS, ANL	ID19, APS, ANL	ID19, APS, ANL
Space group	P31	P31	P31	P31
Cell dimensions:
a, b, c (Å)	124.7, 124.7, 128.7	124.5, 124.5, 129.1	124.7, 124.7, 129.2	124.7, 124.7, 130.1
α, β, γ (°)	90, 90, 120	90, 90, 120	90, 90, 120	90, 90, 120
Resolution (Å)	40.00–2.80 (2.90–2.80)*	40.00–2.80 (2.90–2.80)	40.00–2.60 (2.69–2.60)	40.00–2.20 (2.28–2.20)
No. reflections measured	135413	139169	164199	284949
No. reflections unique	53603 (5390)	53935 (5473)	66518 (6738)	111206 (11119)
R merge	0.058 (0.735)	0.063 (0.668)	0.068 (0.425)	0.062 (0.742)
CC1/2	0.974 (0.480)	0.986 (0.551)	0.971 (0.475)	0.985 (0.494)
I / σI	17.2 (2.4)	17.87 (2.7)	11.7 (2.0)	19.2 (2.4))
Completeness (%)	97.9 (98.8)	97.7 (98.8)	95.4 (96.2)	97.9 (97.6)
Redundancy	2.6 (2.6)	2.7 (2.7)	2.6 (2.4)	2.6 (2.6)
Refinement:
Resolution	38.9 – 2.80	38.9 – 2.8	38.9 – 2.60	38.9 – 2.20
No. reflections	52122	53195	64293	108822
Rwork / Rfree	0.188 / 0.220	0.192 / 0.220	0.188 / 0.219	0.192 / 0.211
No. atoms
Protein	8376	8376	8386	8413
Ligands	48	90	110	38
Water	162	155	254	798
B-factors
Protein	47.6	45.3	44.4	36.8
POX	44.1	41.2	45.2	N/A
Oxime	N/A	85.6	88.2	79.0
Glycerol / DMSO NO3	57.8	55.3	61.8	47.2
Water	36.8	35.3	40.0	43.5
R.M.S. deviations
Bond lengths (Å)	0.003	0.002	0.004	0.002
Bond angles (°)	0.584	0.566	0.694	0.564
Ramachandran statistics
Favored	95.82	96.38	95.82	96.11
Allowed	4.18	3.62	4.18	3.79
Outliers	0	0	0	0.09
Clash score	3.25	3.18	4.54	2.10
Diffraction precision index	0.37	0.38	0.28	0.24

^*- Values in parentheses are for highest resolution shell.

Inelastic neutron scattering

The INS measurements were performed at 120K on the hydrated powder samples, which guarantees that the protein exhibits only harmonic vibrational dynamics. 100 mg of apo-hAChE and POX-hAChE were lyophilized and re-hydrated with D₂O to 33% and 41% solvation by mass, respectively, to minimize incoherent scattering from H atoms of water molecules. These samples were then wrapped in aluminum foil and placed within the time-of-flight instrument Cold Neutron Chopper Spectrometer (CNCS) at the Oak Ridge National Laboratory (ORNL).

The signal intensity, uncertainty and number of events were recorded in the range of −10 to +10 meV energy transfer ω and 0.31 to 5.19 Å⁻¹ scattering angle q. A vanadium sample was used to normalize data from each protein sample. The signal intensity and uncertainty at each q, ω was also normalized to the number of events at that grid point, which was not uniform over the samples. The integrated dynamic structure factor S(q, ω) was determined at both low-Q (0.76 to 1.26 Å⁻¹) and high-Q (3.3 to 4.1 Å⁻¹). A linear extrapolation of the operand in Equation 1 was used to determine the intercept at q = 0. The absolute density of states g(ω) followed the same procedure as in Balog et al. (2004) with the protein molecular mass calculated based on the structure.

$$g_{exp}(\omega )=\underset{q\to 0}{\lim }\frac{6\omega }{\hslash q^2}\left[e^{\hslash \omega /kT}-1\right]S(q,\omega )$$ Equation 1

MD simulations

MD simulations were performed following the experimental conditions under which the neutron scattering measurements were conducted to obtain molecular details on the vibrational modes altered by POX binding. Two powder model systems were constructed: apo-hAChE and POX-hAChE. Because we cannot obtain crystals of apo-hAChE using our hAChE construct due to the lack of Asn350 glycosylation, we took the VX-hAChE structure (PDB 6O66, Gerlits et al. 2019b) as the basis for apo-hAChE with the VX removed in order to accurately model the unmodified enzyme. The monomer structure of VX-hAChE is relatively similar to the glycosylated apo-hAChE (PDB 4EY4), with the backbone RMSDs of 0.32 and 0.33 Å respectively to each monomer in the apo-hAChE dimer. A total of 8 monomers of unmodified hAChE were randomly oriented and brought together to simulate a powder sample. The entire system was fully solvated and then waters were systematically removed until the hydration matched the experimental hydration by mass of 33%. The resulting powder model was heated up to 300 K over the course of 50 ps, and then equilibrated for an additional 120 ns. The final structures were then used as the starting structures for the impulse calculations described below. The same procedure was followed for the POX-hAChE powder model but using POX-hAChE crystal structure reported here and the hydration level was 41% by mass.

All MD simulations were performed using NAMD (v.2.9; Phillips et al. 2005) with the CHARMM c36 force field for the protein (Huang et al. 2013). The force field parameters for POX-conjugated Ser203 were obtained from CGenFF (v.1.0.0; Vanommeslaeghe et al. 2010; Yu et al. 2012; Vanommeslaeghe et al. 2012a, Vanommeslaeghe et al., 2012b). All protein residues were assigned to their canonical protonation states at neutral pH. Chloride and sodium ions were added to neutralize the system and reach a concentration of 100 mM. Production MD simulations were performed in the NPT ensemble. The temperature was maintained using the Langevin dynamics with a damping coefficient of 2 ps⁻¹, and the pressure at 1 atm using the Langevin piston method. All covalent bonds involving a hydrogen atom were constrained using the SHAKE algorithm. A cutoff of 12 Å was used for the short-range electrostatic interactions, while the long-range interactions were treated with the Particle-Mesh Ewald (PME) method.

Impulse velocity autocorrelation

Differences in the apo-hAChE and POX-hAChE crystal structures indicated which local regions of the protein were affected by POX-conjugation. These differences were roughly translated into vibrational modes via a linear extrapolation from one structure $({\overrightarrow{x}}_{i,A})$ to the other ${\overrightarrow{(x}}_{i,B})$ in localized regions (Equation 2). A scaled displacement vector $({\overrightarrow{v}}_P^{\prime })$ pointing from state A towards state B in a localized region P was added to the velocities of the atoms within P to simulate the effect of an instantaneous impulse along the vibrational mode with s=2.5 (Equation 3). Molecular dynamics was performed both in the presence and absence of this impulse, saving the velocity every 2 fs during the 10 ps simulations. The velocity of the atoms within the localized region P were then correlated with the initial impulse velocity (Equation 4). A Fourier transform of the velocity correlation function yielded a power spectrum, showing the vibrational frequencies exhibited by the localized region P. For each local region, 5 different MD trajectories of 8-monomer powder models were analyzed, and the average and standard deviation of the 40 resulting power spectra were used in the figures. The effects of the three impulse options (acyl-loop impulse, helix-bundle impulse or omega-loop impulse) on the protein vibrational spectra were computed.

$${\overrightarrow{d}}_P=\sum _{i\in P}{\overrightarrow{x}}_{i,B}-{\overrightarrow{x}}_{i,A}$$ Equation 2

$${\overrightarrow{v}}_P^{\prime }=\overrightarrow{v}+s\cdot {\overrightarrow{d}}_P$$ Equation 3

$$C_P(\tau )=\sum _i^N{\overrightarrow{v}}_P(\tau )\cdot {\overrightarrow{v}}_P^{\prime }\left({\tau }_0\right)$$ Equation 4

Quantification and statistical analysis

The crystal structures of hAChE in complex with POX, MMB4, and RS170B were determined using materials and software listed in the Key resources table. Statistics generated from X-ray crystallography data processing, refinement, and structure validation are displayed in Table 1. Data shown in Figure 5 represent mean of the inelastic neutron scattering signal integrated over 0.31 to 5.19 Å scattering angle q, and error represents SD of the signal within this q range at each specific frequency. Data shown in Figure 6 represent mean of five independent MD simulations of 8 hAChE monomers each resulting in 40 power spectra, and error represents SD n = 40.

Highlights.

• Structures of hAChE with POX and oximes show unseen acyl loop (AL) conformations

• Marked distortions of the AL in POX-hAChE reverse only in part upon oxime binding

• Inelastic neutron scattering from POX-hAChE is stronger than that from native hAChE

• MD shows vibrational dynamics softening of the AL, not of Ω loop or dimer interface