A cocaine hydrolase engineered from human butyrylcholinesterase as a medical countermeasure for aldicarb poisoning

Abstract

Aldicarb is known to be the most toxic carbamate pesticide and could be used easily by terrorists to cause a mass casualty incident. It is highly desired to discover a medical countermeasure for aldicarb poisoning. Based on structural insights from molecular modeling and in vitro experimental validation, CocH3-Fc(M3), a potent cocaine hydrolase engineered from human butyrylcholinesterase (BChE), has a ~4-fold improved binding affinity with aldicarb compared to wild-type human acetylcholinesterase (AChE) and BChE and a 9–28-fold improved bimolecular rate constant for the carbamylation with aldicarb compared to AChE. CocH3-Fc(M3) also has significant catalytic activity for aldicarb hydrolysis and, hence, is identified as an aldicarb hydrolase. To the best of our knowledge, this is the first aldicarb hydrolase identified so far. Due to significant catalytic activity against aldicarb, CocH3-Fc(M3) effectively and dose-dependently rescued all mice that had been injected with a lethal dose of aldicarb, demonstrating the promise of CocH3-Fc(M3) as a potential medical countermeasure for aldicarb poisoning.

1. Introduction

Aldicarb is known to be the most toxic of the carbamate pesticides.¹ It is one of the Chemicals of Concern (CoC) identified by the US Department of Homeland Security, and it belongs to the Cholinergic Warfare and Pesticides category. This category consists of both organophosphorus (OP) compounds (OP chemical warfare or pesticides) and carbamates like aldicarb. As a potent, fast-acting inhibitor of acetylcholinesterase (AChE), aldicarb is readily absorbed from all routes of exposure, including oral and dermal exposure.¹ In all species tested, the acute toxicities of aldicarb are similar. For example, LD₅₀(aldicarb) = 0.3 mg/kg (IP) or 0.38–0.50 mg/kg (oral, with aldicarb in corn oil) in mice. LD₅₀(aldicarb) = 0.47 mg/kg (IV) or 0.57 mg/kg (IP) or 0.88–0.93 mg/kg (oral, with aldicarb in corn oil) in rats (EPA Report)². Due to its physical properties documented in literature, aldicarb could be used easily by terrorists to cause a mass casualty incident. Notably, when rats, mice, and guinea pigs were exposed to aldicarb, finely ground, mixed with talc, and dispersed in the air at a concentration of 200 mg/m³ for five minutes, all animals died; however, at a lower concentration (6.7 mg/m³), 15-minute exposure was not lethal, but five out of six animals died during a 30-minute exposure test.² According to a study with aldicarb aerosol, when rats were exposed to aldicarb aerosol for four hours, mortality was 1/10, 1/10, 5/10, 10/10, and 10/10 when the aldicarb concentration was 0.82, 2.0, 6.0, 8.7, and 46.3 mg/m³, respectively; the calculated 4-hour LC₅₀ was 3.9 mg/m³ (EPA Report).³

OP warfare/pesticides and carbamate pesticides have a common mode of action for their neurotoxicity as AChE inhibitors.⁴ However, they differ in the specific type of chemical reaction: an OP will first react with AChE via phosphorylation and then eventually go through the well-known aging process (without spontaneous dephosphorylation or reactivation observed/reported), whereas a carbamate like aldicarb will first react with AChE via carbamylation and then go through a decarbamylation process (see Figure 1).⁵ For this reason, in vivo AChE function can be recovered over a prolonged period of time (greater than 24 hours) due to spontaneous hydrolysis of carbamylated AChE.⁶ According to reported toxicokinetics/toxicodynamics (TK/TD) data in humans determined from case studies of severe (but non-fatal) aldicarb toxicity after overdose, aldicarb plasma concentration over time fits a two-compartment model with a distribution half-life of 0.4 h and an elimination half-life of 13 h.⁷ Aldicarb poisoning causes rapid-onset severe toxicity with muscarinic and nicotinic excess, seizures, and decreased consciousness.⁷ Plasma butyrylcholinesterase (BChE) activity rapidly recovers once aldicarb concentrations decrease. So, for the treatment of aldicarb poisoning, it is crucial to eliminate/degrade aldicarb in the body.

Figure 1.

Hydrolysis reactions of aldicarb and acetylcholine. (A) Aldicarb hydrolysis. (B) Acetylcholine hydrolysis. (C) Carbamylation stage of AChE/BChE-catalyzed aldicarb hydrolysis. (D) Decarbamylation stage of AChE/BChE-catalyzed aldicarb hydrolysis. The amino-acid residue numbers refer to human AChE/BChE. The amino-acid residue numbers refer to human AChE/BChE within the well-known catalytic triad (S-H-E or Ser-His-Glu) and oxyanion hole of the cholinesterase. The three residues within the oxyanion hole can potentially form hydrogen bonds with the carbonyl oxygen of the substrate.

As is well-known, currently available options for treating OP and aldicarb poisoning, including administration of atropine or co-administration of atropine and pralidoxime (2-PAM, a well-known oxime reactivator developed to reactivate phosphorylated AChE), have limited efficacy. There have been extensive efforts in development of improved options for treatment of OP poisoning, such as novel types of more active oxime reactivators^{8, 9} that can cross the blood-brain barrier (BBB)¹⁰ to reactivate phosphorylated AChE in the brain and/or even reactivate the aged AChE (as demonstrated in recently reported studies^{11, 12}); stoichiometric bioscavengers using native or recombinant human BChE;^{13, 14} and catalytic OP bioscavengers¹⁵ using other engineered enzymes or BChE-polymer-oxime conjugates.¹⁶ In comparison, relatively fewer studies have been reported in literature in testing and developing treatments for aldicarb poisoning,^17–19 and there are no truly promising therapeutic strategies available in literature. Reported studies indicated that 2-PAM had neither positive nor negative effects on survival in animal studies on aldicarb intoxication.¹⁹ This outcome is understandable, as 2-PAM was developed to reactivate phosphorylated AChE (an intermediate formed from the reaction of AChE with an OP), but not carbamylated AChE (an intermediate formed from the reaction of AChE with a carbamate).¹⁹ It is highly desired to develop a new, effective treatment strategy for aldicarb poisoning.

Our research was focused on rational design and discovery of an engineered human enzyme as an aldicarb hydrolase capable of rapidly and efficiently detoxifying aldicarb as an effective aldicarb bioscavenger–medical countermeasure–for aldicarb poisoning. As the first necessary condition, the desired aldicarb bioscavenger must be able to react with aldicarb significantly faster than AChE reacting with aldicarb to protect AChE from further inhibition reaction (carbamylation) with aldicarb. This is because the exogenous aldicarb hydrolase and endogenous AChE will compete for reaction with aldicarb in the body. With the faster reaction of the exogenous protein (or the more potent the binding of the exogenous protein) with aldicarb, there will be less reaction of endogenous AChE with aldicarb.

Our aldicarb hydrolase discovery effort was motivated by our previous success in computational design, discovery, and development of protein drugs for treatment of cocaine use disorders (CUDs, including cocaine overdose and addiction)^20–25 that resulted in promising protein drug candidates, known as cocaine hydrolases (CocHs). One CocH is known as TV-1380 for treatment of cocaine addiction. TV-1380 is our computationally designed, discovered and patented first cocaine hydrolase (known as CocH1, i.e. the A199S/S287G/A328W/Y332G mutant of human BChE with a ~1,000-fold improved catalytic efficiency compared to the wild-type BChE against cocaine)^26–29 fused with human serum albumin (HSA)³⁰ to extend the biological half-life for clinical development to treat cocaine addiction. This fusion protein is also known as Albu-BChE, Albu-CocH, or Albu-CocH1 in literature.^30–35 A Phase II clinical trial³⁴ revealed that TV-1380 at a once-weekly dosing schedule with intramuscular (IM) injection of 150 or 300 mg/week significantly and dose-dependently decreased cocaine intake in the treatment groups compared to the placebo group, although its biological half-life (t_1/2 = 8 h in rats and 43–77 h in humans) was not long enough for the practical cocaine dependence treatment using TV-1380 with the desired weekly dosing schedule. Our more recently designed and discovered CocHs, such as CocH3 (the A199S/F227A/S287G/A328W/Y332G mutant of human BChE with a ~2,000-fold improved catalytic efficiency against cocaine compared to the wild-type BChE), have further improved the catalytic efficiency (k_cat/K_M) against cocaine compared to TV-1380.^{24, 25, 36, 37} Further, we have recently developed a new CocH entity, denoted as CocH-Fc for convenience, to improve the in vivo efficacy and also the yield of protein expression^{21–23, 38–40} through fusion of CocH (truncated after amino acid #529 to delete the tetramerization domain) with N-terminus of the Fc (wild-type or mutant) portion of human antibody immunoglobulin G1 (IgG1). We have demonstrated that CocH fusion with Fc can considerably improve the in vivo stability and prolong the biological half-life (as well as the yield of protein expression) without changing the protein folding or in vitro catalytic parameters of the enzyme (CocH) as shown in our previous studies.^{21, 23} So, the obtained CocH-Fc fusion protein has a desirable, excellent pharmacokinetic (PK) profile for a protein drug.^{21, 22, 38} As a result, a dose of 3 mg/kg CocH3-Fc completely blocked the discriminative stimulus and reinforcing effects of cocaine for 24/25 days and continued to significantly attenuate/decrease the cocaine effects for at least 29 days in rats.²²

In the present report, guided by molecular modeling, we carried out in vitro studies on a long-acting cocaine hydrolase, known as CocH3-Fc(M3)²³ engineered from human BChE, for its binding affinity with aldicarb and catalytic parameters for the aldicarb hydrolysis. Further, in vivo experiments were performed to examine whether CocH3-Fc(M3) can effectively rescue mice from a lethal dose of aldicarb.

2. Methods

2.1. Materials

The CocH3-Fc(M3) protein was expressed using our previously reported stable CHO-S cell line with plasmid pCSC-CocH3-Fc(M3) using pCSC-SP-PW lentiviral vector (plasmid 12335) encoding CocH3-Fc(M3) as described in detail in our previous report.²³ The stable CHO-S cells were cultured using Gibco FreeStyle^™ CHO expression media + 100 U/mL penicillin + 100 μg/mL streptomycin + 8 mM L-Glutamine (as base media). After seed vial thawing, cells were pelleted and resuspended in complete media, then grown in adherent culture for 72 hours in base media + 10% fetal bovine serum (FBS), at 37 °C in a 5% CO₂ environment. After achieving confluence, adhered cells were suspended by addition of trypsin, pelleted, resuspended in base media, and introduced into suspension culture (base media + 3% FBS, orbital shaker at 125 RPM, 37 °C, and 5% CO₂) for 2–3 days for further cell expansion. Cells produced from this expansion culture were used to inoculate enzyme expression cultures (base media, orbital shaker at 125 RPM, 37 °C, and 5% CO₂) at an initial viable cell density (VCD) of 5 × 10⁵ cells/mL. The culture temperature was reduced to 32 °C on Day 2 and 30 °C on Day 4, and the culture medium was harvested when the cell viability ≤ 90%, and the supernatant was collected and clarified by vacuum-assisted filtration (pore size 0.2 μm). Secreted CocH3-Fc(M3) protein was purified from clarified supernatant by Protein A affinity chromatography. The protein purification process using Protein A affinity chromatography consists of protein binding step at pH 7.4, wash step 1 (0.05 M Tris, 0.15 M NaCl, pH 7.4), wash step 2 (0.05 M Tris, 1.0 M NaCl, pH 7.4), wash step 3 (0.05 M acetate, pH 5.5), and elution step (0.05 M acetate, 0.15 M L-Arginine, pH 4.0). Eluted enzyme was concentrated and exchanged into storage buffer (20% w/v D-sorbitol, 50 mM HEPES, 1.0 M L-glycine, pH 7.4) by centrifuge-assisted filtration using Amicon 50 KDa MWCO centrifuge filters, before sterile filtration and storage at −80 °C. The CocH3-Fc(M3) enzyme concentration in this final material was ~10 mg/mL with a purity >90%. The purified CocH3-Fc(M3) protein was stored at −80 °C before use.

Aldicarb (catalog # 33386), butyrylthiocholine Iodide (BTC, catalog # B3253–5G), and 5,5-Dithiobis (2-nitrobenzoic acid) (DTNB, catalog # D8130–5G) were ordered from Sigma-Aldrich (St. Louis, MO). Aldicarb was prepared in 0.1 M phosphate buffer (PB) at pH 7.4. (−)-Cocaine was provided by the National Institute on Drug Abuse (NIDA) Drug Supply Program (Bethesda, MD); and [³H](−)-Cocaine (50 Ci/mmol) was ordered from PerkinElmer (Waltham, Massachusetts).

All other general supplies for the experimental studies were obtained from Sigma-Aldrich (St. Louis, MO) or Thermo Fisher Scientific (Waltham, MA).

2.2. Animals

32 male CD-1 mice (body weights: 28–32 g) were ordered from Envigo (Indianapolis, IN). The in vivo experiments were carried out in our animal lab within the animal laboratories of the Division of Laboratory Animal Resources (DLAR) facility (PHS assurance number A3336–01; USDA number 61-R-0002; AAALAC, Intl. Unit # 13) at the University of Kentucky, with the DLAR staff providing veterinary care and animal husbandry services. All mice were housed in clean, adequately-sized, stainless-steel cages at 21–22 °C and were allowed ad libitum access to food and water. All in vivo experiments were performed in accordance with the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the National Institutes of Health (NIH), and were in fact also consistent with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines (https://arriveguidelines.org). The animal procedures used in this study had been approved by the University of Kentucky’s Institutional Animal Care and Use Committee (IACUC).

2.3. Molecular modeling

All computational modeling studies in this work started from X-ray crystal structures of human AChE (PDB ID: 4EY5)⁴¹ and BChE (PDB ID: 1P0M).⁴² The X-ray crystal structures (chain A) were cleaned using the Protein Preparation Wizard⁴³ of the Maestro-V13.0 within the Schrödinger software suite.⁴⁴ The cleaned AChE structure (4EY5) was used to model the AChE-aldicarb binding, and the cleaned BChE structure (1P0M) was used to model the structures of aldicarb binding with BChE and various BChE mutants (including CocH3 and others) through molecular docking using the Schrödinger software suite. The docked binding complexes with aldicarb were energy-minimized by using the Amber20 program (Sander module in the AmberTools)⁴⁵ using the AMBER ff19SB force field⁴⁶ for protein and the general AMBER force field (gaff2)⁴⁷ and the restricted electrostatic potential (RESP) charges for aldicarb. The RESP charges were determined by first quantum mechanically (QM) optimizing the aldicarb geometry and calculating the electrostatic potential at the B3LYP/6–31+G* level using the Gaussian16 program⁴⁸ and then the RESP fitting of the QM-calculated electrostatic potential.⁴⁹ For the energy minimization, the docked protein-aldicarb binding structure was soaked in an orthorhombic box of TIP3P water molecules⁵⁰ with a minimum solute wall distance of 12 Å. The solvated complex system was neutralized by adding counter ions. For each protein-aldicarb complex, the energy minimization was first performed by using the steepest descent method for 2,500 steps and then the conjugate gradient method until convergence.

2.4. Evaluation of aldicarb non-covalent inhibition of CocH3-Fc(M3)

The in vitro inhibitory activity of aldicarb against CocH3-Fc(M3) was determined first by using a sensitive radiometric assay of [³H](−)-cocaine hydrolysis as used in our previous studies.^{21, 24, 25} To measure the instantaneous inhibition of CocH3-Fc(M3) by aldicarb at 25 °C as a function of aldicarb concentration, CocH3-Fc(M3) was simultaneously and briefly (5 minutes) incubated with (−)-cocaine—including trace quantities of [³H](−)-cocaine—and aldicarb; [E] = 0.02 nM, [cocaine] = 1 μM, and aldicarb concentrations varying from 0.1 μM to 3.16 mM (log₁₀[aldicarb] = −7.0 to −2.5 when the unit of aldicarb concentration [aldicarb] is M), in 0.1 M sodium phosphate buffer, pH 7.4. Reactions were quenched after 5 minutes reaction time with one reaction volume of 0.1 N HCl, and hydrolyzed cocaine was measured by scintillation counting involving detection of [³H](−)-cocaine hydrolysis product as used in previous studies.^{21, 24, 25} All these reactions were run in triplicate, except for the aldicarb-free control (100% activity) and negative control reactions, in which twelve replicates were run (three replicates per set of parallel reactions, in which the complete assay involved four sets of reactions). Scintillation counting was measured in triplicate upon each reaction replicate.

Enzyme activity of CocH3-Fc(M3) incubated with aldicarb relative to the 100% activity controls was calculated from scintillation counting results per Eq. 1:

$$\mathit{Relative}\mathit{Activity}(\mathrm{\%})=\frac{C_{\mathit{Test}}-C_{NC}}{C_{IFC}-C_{NC}}\cdot 100\mathrm{\%}$$ (Eq. 1)

Where $C_{\mathit{Test}}$, $C_{NC}$, and $C_{IFC}$ are scintillation counting results from test (involving addition of aldicarb) reactions, negative control (enzyme/aldicarb-free) reactions, and inhibitor-free control reactions, respectively. The IC₅₀ of the resulting relative activity versus aldicarb concentration data set was calculated by fitting the data to Eq. 2 (log₁₀([inhibitor]) vs normalized response, variable slope) in

GraphPad Prism by nonlinear regression:

$$Y=\frac{100}{1+{10}^{a·\left({\mathit{log}}_{10}\left(I{C}_{50}\right)-X\right)}}$$ (Eq. 2)

where $Y=\mathrm{relative}\mathrm{activity}$, $X={\log }_{10}([\mathrm{aldicarb}])$, and the fitting parameter ‘$a$’ is the Hill coefficient.

The resulting IC₅₀ value was converted to the non-covalent inhibition constant $K_I$ via the Cheng-Prusoff equation (Eq. 3) in which the manner of non-covalent aldicarb/CocH3-Fc(M3) interaction was considered to be competitive inhibition. The parameters [S] (the concentration of substrate (−)-cocaine) and $K_{\mathrm{M}}$ are the inhibition assay concentration and Michaelis-Menten constant of the substrate (cocaine). In this case, [S] = 1 μM, and $K_{\mathrm{M}}=3.1\mu \mathrm{M}$ was used for CocH3-Fc(M3) against cocaine at 25 °C.²⁵

$$K_{\mathrm{i}}=\frac{I{C}_{50}}{1+\frac{\left[S\right]}{K_M}}$$ (Eq. 3)

This competitive inhibition constant, $K_{\mathrm{i}}$, is the dissociation equilibrium constant for reversible binding of aldicarb to the active site of the enzyme.

2.5. In-vitro evaluation of enzyme inactivation and reactivation kinetics versus aldicarb

To determine the in vitro activity of enzyme CocH3-Fc(M3) against aldicarb, CocH3-Fc(M3) was incubated with aldicarb at 25 °C; enzyme concentration [E] = 300 nM, initial aldicarb concentration [aldicarb]₀ = 1,000 nM, in 0.1 M Phosphate Buffer (PB), pH 7.4. Enzyme activity was determined at various time points by measurement of the initial hydrolysis rate ($V_0$) of butyrylthiocholine (BTC), in which 5,5-dithiobis-(2-nitrobenzoic acid) (DTNB) was used for detection of hydrolyzed BTC via kinetic measurement of A₄₅₀ in a well-plate format. Assay reaction conditions were total enzyme concentration [E]_rxn = 0.25 nM (50 μL 300X pre-diluted aldicarb/enzyme mixture per 200 μL final well volume), [BTC]_rxn = 20 mM, [DTNB]_rxn = 6.25 mM, in 0.1 M PB + 0.25 mg/mL BSA, pH 7.4, at 25 °C. This relatively high concentration of BTC was used to ensure that the BTC concentration was kept near constant during the reaction process. The DTNB concentration was chosen to provide enough DTNB for detection of hydrolyzed BTC product during measurement of the initial reaction rate. Spontaneous BTC hydrolysis was measured in negative control reactions ([E]_rxn = 0), and the enzymatic hydrolysis was calculated from negative-control-subtracted A₄₅₀ versus time profiles. Assay reactions were run in five replicates for each time point, and BTC hydrolysis $V_0$ measurements were compared to that of an aldicarb-free CocH3-Fc(M3) preparation for calculation of the active enzyme concentration over aldicarb incubation time. The obtained time-dependent enzyme activity data were analyzed according to Scheme 1 (as used previously⁵¹ for similar reactions) to obtain the carbamylation rate constant $k_2$ and decarbamylation rate constant $k_3$ when $K_{\mathrm{d}}$ was fixed as 13.1 μM (see below for the $K_{\mathrm{d}}$ value determined):

$$E+S\stackrel{k_1/k_{-1}}{⟷}E\cdot S\stackrel{k_2}{\to }ES\stackrel{k_3}{\to }E$$ (Scheme 1)

in which $S$ is aldicarb (free substrate), $E$ represents the free enzyme, $E·S$ refers to the enzyme-aldicarb complex, and $ES$ represents the carbamylated enzyme (intermediate). The enzyme-aldicarb binding affinity $K_{\mathrm{d}}=k_{-1}/k_1$.

The rates of enzyme-catalyzed aldicarb hydrolysis (producing the carbamylated enzyme intermediate) and enzyme reactivation (decarbamylation), termed $r_{\mathit{inact}}$ and $r_{\mathit{act}}$, are defined in Eq. 4 and Eq. 5:

$$r_{\mathit{inact}}=k_2[E]\left(\frac{\left[S\right]}{K_d+\left[S\right]}\right)$$ (Eq. 4)

$$r_{\mathit{act}}=k_3\left[ES\right]=k_3\left([E{]}_0-[E]\right)$$ (Eq. 5)

These equations for the rates of enzyme-catalyzed aldicarb hydrolysis and enzyme decarbamylation can be applied to describe the instantaneous rates of change in active enzyme and in aldicarb ($d\left[E\right]/dt$ and $d\left[S\right]/dt$, respectively), as defined by Eq. 6 and Eq. 7:

$$\frac{d\left[E\right]}{dt}=-r_{\mathit{inact}}+r_{act}=-k_2[E]\left(\frac{\left[S\right]}{K_d+\left[S\right]}\right)+k_3\left([E{]}_0-[E]\right)$$ (Eq. 6)

$$\frac{d\left[S\right]}{dt}=-r_{\mathit{inact}}=-k_2[E]\left(\frac{\left[S\right]}{K_d+\left[S\right]}\right)$$ (Eq. 7)

A custom program was created within the MATLAB software package to fit kinetic constants $k_2$ and $k_3$, which involved systematic simulation of active enzyme concentration [E] versus time curves—by numerically solving the nonlinear system of ordinary differential equations (ODEs) corresponding to Scheme 1 for instantaneous rates of the change in concentration of active enzyme and of aldicarb (Eq. 6 and Eq. 7)—for combinations of values of the kinetic constants $K_{\mathrm{d}}$, $k_2$, and $k_3$. Initial conditions for all simulations were [E] = 300 nM (= [E]₀), and [S] = 1,000 nM. Here, [S] refers to the concentration of aldicarb as a substrate. The input value for $K_d$ was fixed across simulations for the enzyme. Solutions of [E]-vs-time were produced over input ranges of $k_2$ and $k_3$, and the sum of squared error versus the experimental time course data was calculated for each solution. Simulations continued (with adjustment of input $k_2$ and $k_3$ evaluation ranges) until a stable error-minimized solution was obtained with kinetic constants precision of three significant figures.

2.6. Rescue experiment

All mice (n = 8 per group) were injected intraperitoneally (IP) with a lethal dose (0.75 mg/kg ≈ 2 × LD₅₀) of aldicarb (LD₅₀ = 0.3 mg/kg for IP route²). The injection volume was 10 μL/g body weight, and aldicarb concentration injected was 0.075 mg/mL. At 1–3 minutes after the aldicarb injection, the mice in the treatment group were injected intravenously (IV) with 100 or 50 or 25 mg/kg CocH3-Fc(M3), followed by visually monitoring the survival or death at various time points after aldicarb injection. The survival data was analyzed by the Chi-square test using the GraphPad Prism 10 software (GraphPad Software, La Jolla, CA). A difference was considered significant when p < 0.05 (*, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001).

3. Results

3.1. Insights from molecular modeling

Considering the success in the development of CocHs for treatment of CUDs, we computationally screened the CocHs for their interactions with aldicarb and identified that CocH3 can also favorably interact with aldicarb (see Figure 2). As shown in the modeled binding structures depicted in Figure 2, aldicarb can bind with both AChE (Figure 2A) and BChE (Figure 2B) in a similar binding mode, including two hydrogen bonds between the carbonyl oxygen of aldicarb and the oxyanion hole of the enzyme (AChE or BChE). The obtained similar binding structures of aldicarb with AChE and BChE are consistent with the reported experimental observation showing that aldicarb had IC₅₀ = 67 μM against AChE¹⁹ and that aldicarb had an essentially identical inhibitory potency against both AChE and BChE (based on the observed essentially identical dose-dependent inhibition curves against both AChE and BChE activities in the published report).¹⁹ Interestingly, aldicarb also binds with CocH3 in a similar binding mode, but with an additional hydrogen bond between the carbonyl oxygen of aldicarb and the hydroxyl group on the side chain of the mutated S199 of CocH3 (Figure 2C). Our computational modeling predicted that CocH3 could have a markedly improved binding affinity with aldicarb compared to human AChE and BChE wild-type enzymes.

Figure 2.

Modeled binding structures of aldicarb with (A) AChE, (B) BChE, and (C) CocH3. The binding structures were determined by molecular docking, followed by energy minimization.

Further, based on the similarity between the serine esterase-catalyzed hydrolyses of a carboxylate ester such as acetylcholine (Figure 1A) and a carbamate such as aldicarb (Figure 1A), the binding mode shown in Figure 2C will allow CocH3 to convert aldicarb to its hydrolysis products. According to the binding mode depicted in Figure 2C, the hydroxyl oxygen of S198 is close to the carbonyl carbon of aldicarb, ready for the desired nucleophilic attack. While the C^…O distance becomes closer and closer, the hydrogen bonds between the carbonyl oxygen of aldicarb and the oxyanion hole of CocH3 will become stronger and stronger due to the increasing negative charge on the carbonyl oxygen of aldicarb during the reaction process. This means that the transition state of the first reaction step (which is usually the rate-determining step for esterase-catalyzed hydrolysis⁵²) will be stabilized by the hydrogen bonds between the carbonyl oxygen of aldicarb and the oxyanion hole.⁵² So, the stronger hydrogen bonds between the carbonyl oxygen of aldicarb and the oxyanion hole in the transition state will help to lower the energy barrier for the enzymatic hydrolysis.

3.2. In vitro activity

With the above encouraging computational insights and predictions, we experimentally tested the in vitro activities of CocH3-Fc(M3) with aldicarb. First, we determined the inhibitory activity of aldicarb against the CocH3-Fc(M3)-catalyzed hydrolysis of 1 μM (−)-cocaine (see Figure 3A), demonstrating that IC₅₀ = 17.3 ± 1.3 μM or K_i = 13.1 ± 0.9 μM (according to Eq. 3). Notably, both the functional parameters IC₅₀ and $K_{\mathrm{i}}$ for an enzyme-aldicarb system reflect the enzyme-aldicarb binding affinity ($K_{\mathrm{d}}$).

Figure 3.

In vitro activity of CocH3-Fc(M3) with aldicarb. (A) Instantaneous inhibition of CocH3-Fc(M3) by aldicarb at 25 °C. Enzyme ([E] = 0.02 nM) was simultaneously and briefly (5 minutes) incubated with aldicarb and substrate cocaine over log₁₀([aldicarb]) = −7.0 to −2.5, and cocaine hydrolysis was compared to an aldicarb-free control preparation for determination of relative activity versus aldicarb concentration. The inhibition data was fit to Eq.2, giving a (Hill coefficient) = −1.134 ± 0.015. (B) Plot of the active enzyme CocH3-Fc(M3) concentration versus time for reaction of 300 nM CocH3-Fc(M3) with 1 μM aldicarb, with best-fit [E] versus time solution curve for input K_d = 13.1 μM (from instantaneous inhibition assays) and best-fit kinetic constants $k_2$ and $k_3$ (error minimization of simulated [E] versus time solutions against experimental data). The measured concentrations are present as the mean ± 95% Confidence Interval (CI). For each parameter value shown in the figure, the uncertainty (± standard error of the mean) was calculated from the parameter values determined by the individual repeats of measurements (n=3 for panel A and n=5 for panel B).

Further, we experimentally examined the in vitro activity of CocH3-Fc(M3) against aldicarb (see Figure 3B for the time-dependent enzyme activity after 300 nM CocH3-Fc(M3) was incubated with 1 μM aldicarb). The obtained time-dependent enzyme activity data were analyzed according to Scheme 1. It should be noted that the functional parameters $K_{\mathrm{i}}$ and $K_{\mathrm{d}}$ for an enzyme-aldicarb system reflect the same enzyme-aldicarb binding affinity. Hence, it is expected that K_d = K_i = 13.1 ± 0.9 μM.

Based on Scheme 1 with K_d = 13.1 μM, by fitting the time-course data in Figure 3B, we obtained k₂ = (2.90 ± 0.05) × 10⁻¹ min⁻¹, and k₃ = (1.03 ± 0.02) × 10⁻² min⁻¹ for CocH3-Fc(M3) against aldicarb. So, CocH3-Fc(M3) can be expected to first bind with aldicarb potently and then catalyze the aldicarb hydrolysis to effectively block aldicarb from interacting with endogenous AChE and BChE. As a result, 1 μM (or 1,000 nM) aldicarb can only partially inhibit 300 nM CocH3-Fc(M3), and the inhibited CocH3-Fc(M3) can fully recover, as shown in Figure 3B, after all the aldicarb molecules in the reaction system are hydrolyzed. Based on these kinetic data, the bimolecular rate constant (k) for the reaction of aldicarb with CocH3-Fc(M3) (from E + S to ES) can be estimated as k = k₂/K_d = 2.21 × 10⁴ M⁻¹ min⁻¹.

Using the obtained kinetic parameters, one may conveniently simulate the hydrolysis of aldicarb at a typically high concentration produced by a lethal dose of aldicarb in the presence of CocH3-Fc(M3). We note that the aldicarb hydrolysis consists of carbamylation step (Figure 1C) and decarbamylation step (Figure 1D) and that the main product (aldicarb oxime) of aldicarb hydrolysis is formed in the carbamylation step. In a study reported by Mohamed et al.,⁵³ a lethal dose of aldicarb (0.93 mg/kg) was administered to rats, followed by measurements until death, demonstrating that the aldicarb concentration in the blood reached as high as ~5.48 mg/L or ~28.8 μM. So, the initial free aldicarb concentration used in our kinetic simulation was 28.8 μM. Concerning the initial free CocH3-Fc(M3) concentration, our previous studies^{21–23, 38} have revealed that CocH3-Fc(M3) and the similar CocH-Fc fusion proteins all follow a linear pharmacokinetic (PK) profile in rodents (rats or mice) and that the Fc fusion enzyme at a dose of 0.075 mg/kg resulted in an average plasma fusion enzyme concentration of 1.85 mg/L or 0.0168 μM. Because the linear PK profile means that the CocH3-Fc(M3) concentration in blood is linearly proportional to the dose, the plasma concentration of CocH3-Fc(M3) (or another CocH-Fc protein) in rodents may be estimated to be ~22.4, ~11.2, and ~5.61 μM (without significant changes within the first 30 minutes according to our previous studies^{21–23, 38}) associated with the doses 100, 50, and 25 mg/kg, respectively. Hence, we simulated the kinetics of 28.8 μM aldicarb in the presence of 22.4, 11.2, or 5.61 μM CocH3-Fc(M3). The simulated results within the first 30 minutes are depicted in Figure 4. As shown in Figure 4A, 22.4 μM CocH3-Fc(M3) may rapidly decrease the aldicarb concentration from ~28.8 μM to ~10.0 μM within 5 minutes or ~2.85 μM within 30 minutes. According to Figure 4B, 11.2 μM CocH3-Fc(M3) may also rapidly decrease the aldicarb concentration from ~28.8 μM to ~18.4 μM within 5 minutes or ~14.8 μM within 30 minutes. Concerning the data shown in Figure 4C, 5.6 μM CocH3-Fc(M3) may decrease the aldicarb concentration from ~28.8 μM to ~23.4 μM within 5 minutes or ~21.7 μM within 30 minutes.

Figure 4.

Kinetic simulation of 5.48 mg/L (~28.8 μM) aldicarb hydrolysis in the presence of CocH3-Fc(M3) at a concentration of 22.4 μM estimated for the dose of 100 mg/kg CocH3-Fc(M3) (A), 11.2 μM estimated for the dose of 50 mg/kg CocH3-Fc(M3) (B), or 5.61 μM estimated for the dose of 25 mg/kg CocH3-Fc(M3) (C) using the determined kinetic parameters. The time step used for the simulation was 0.1 min. So, there are a total of 301 time points from time 0 to 30 min for each panel. Each panel shows the time-dependent concentrations of remaining aldicarb, aldicarb oxime (the main product of aldicarb hydrolysis shown in Figure 1), carbamylated enzyme CocH3-Fc(M3), and CH₃NHCOOH (the second product of aldicarb hydrolysis shown in Figure 1).

Additionally, the obtained kinetic parameters ($K_{\mathrm{d}}$, $k_2$, and $k_3$) may also be used to estimate the usually used parameters $k_{\mathrm{cat}}$, $K_{\mathrm{M}}$, and $k_{\mathrm{cat}}/K_{\mathrm{M}}$ using Eq. 8 to Eq. 10:⁵⁴

$$k_{\mathrm{cat}}=\frac{k_2{k}_3}{k_2+k_3}$$ (Eq. 8)

when $k_{-1}\gg k_2$,

$$K_{\mathrm{M}}=\frac{K_{\mathrm{d}}{k}_3}{k_2+k_3}$$ (Eq. 9)

$$\frac{k_{\mathrm{cat}}}{K_{\mathrm{M}}}=\frac{k_2}{K_{\mathrm{d}}}$$ (Eq. 10)

Using Eq. 8 to Eq. 10, we have k_cat = ~0.995 × 10⁻² min⁻¹, K_M = ~0.45 μM, and k_cat/K_M = ~2.21 × 10⁴ M⁻¹ min⁻¹. Notably, the estimated catalytic efficiency $\left(k_{\mathrm{cat}}/K_{\mathrm{M}}\right)$ is equal to the bimolecular rate constant $k$ when $k_{-1}\gg k_2$ according to Eq. 10.

3.3. In vivo activity against aldicarb in mice

Based on the promising in vitro reactivity data, we further tested CocH3-Fc(M3) at a dose of 100 mg/kg against a lethal dose (0.75 mg/kg ≈ 2 × LD₅₀) of aldicarb in male mice (n = 8 per group). As seen in Figure 5, all mice in the control group died at an average time of 8.2 min after the aldicarb injection. Prior to death, all mice in the control group exhibited convulsion-like movements. In comparison, post-exposure treatment using CocH3-Fc(M3) provided complete protection from aldicarb mortality at doses of 50 and 100 mg/kg and provided significant protection from mortality at a lower dose of 25 mg/kg (5/8 mice survived; p = 0.0035, Chi-square test), as also shown in Figure 5. No convulsion-like behavior was detected throughout the duration of visible toxicity symptoms in any of the mice in the 50 and 100 mg/kg CocH3-Fc(M3) treatment groups, nor in 3/5 rescued mice in the 25 mg/kg treatment group. So, the minimum effective dose of CocH3-Fc(M3) for providing full protection (no death) against the lethality of 0.75 mg/kg aldicarb was 50 mg/kg.

Figure 5.

Rescue effects of CocH3-Fc(M3) against a lethal dose (0.75 mg/kg) of aldicarb. Mice (n=8/group) were injected (IP) with 0.75 mg/kg aldicarb, and rescue with 100, 50, or 25 mg/kg CocH3-Fc(M3) (IV) began at 1–3 min after the aldicarb injection. Statistical significance (Chi-square test): ** p < 0.01; **** p < 0.0001.

4. Discussion

Based on the prediction from molecular modeling and in vitro experimental validation, CocH3-Fc(M3), a potent cocaine hydrolase engineered from human butyrylcholinesterase (BChE), has a ~4-fold improved binding affinity with aldicarb (K_i = 13.1 ± 0.9 μM) compared to previously reported binding affinity (IC₅₀ = ~67 μM) of aldicarb with both AChE and BChE.¹⁹ The improved binding affinity is consistent with our computational prediction.

Notably, Mulder and Bakker⁵⁵ estimated the bimolecular rate constant (k) for the reaction of aldicarb with AChE as 0.8–2.4 × 10³ M⁻¹ min⁻¹. In comparison, the bimolecular rate constant (k = k₂/K_d = ~2.21 × 10⁴ M⁻¹ min⁻¹) estimated using our kinetic parameters for the reaction of aldicarb with CocH3-Fc(M3) (from E + S to ES) is about 9–28 times faster than that estimated by Mulder and Bakker⁵⁵ for the reaction of aldicarb with AChE. So, CocH3-Fc(M3) may react with aldicarb about 9–28 times faster than AChE in terms of the carbamylation.

More importantly, CocH3-Fc(M3) has significant catalytic activity for aldicarb hydrolysis, with k₂ = (2.90 ± 0.05) × 10⁻¹ min⁻¹ and k₃ = (1.03 ± 0.02) × 10⁻² min⁻¹. So, CocH3-Fc(M3) is identified as an aldicarb hydrolase. To the best of our knowledge, this is the first aldicarb hydrolase identified so far, with a higher binding affinity and a higher reactivity with aldicarb compared to wild-type AChE/BChE and with a catalytic activity against aldicarb. Due to this significant catalytic activity, CocH3-Fc(M3) effectively and dose-dependently rescued all mice that had been injected with a lethal dose (0.75 mg/kg ≈ 2 × LD₅₀) of aldicarb. Based on the in vivo data, CocH3-Fc(M3) may also be used as a potential medical countermeasure for aldicarb poisoning, in addition to its previously known effectiveness in treatment of CUDs.

Concerning why CocH3-Fc(M3) at a dose of 100 mg/kg or 50 mg/kg can effectively rescue all mice from a lethal dose (0.75 mg/kg) of aldicarb, we note that the molar ratio of 0.75 mg/kg aldicarb to 100 mg/kg CocH3-Fc(M3) is ~4.3:1. On the other hand, it is known that aldicarb is readily absorbed from all routes of exposure, including oral and dermal exposure,¹ but is also rapidly metabolized and excreted in the urine.⁵⁶ As a result, only a fraction of aldicarb administered to mice contributed to the blood aldicarb level and needs to be converted to its hydrolysis product aldicarb oxime for the enzyme treatment to be effective. In our simplified kinetic simulation shown in Figure 4, within 30 minutes, CocH3-Fc(M3) at estimated concentrations associated with 100, 50, and 25 mg/kg CocH3-Fc(M3) may decrease the free aldicarb concentration from ~28.8 μM to ~2.85, ~14.8, and ~21.7 μM, respectively. The maximum tolerated concentrations of aldicarb in rodents or any other species are unknown according to the best of our knowledge. In a reported case of nonfatal aldicarb poisoning, the blood aldicarb level reached 3.22 mg/L (~16.9 μM).⁶ In another case report of fatal aldicarb poisoning, the blood aldicarb level reached 6.2 mg/L (~32.6 μM).⁵⁷ According to the simplified simulation, CocH3-Fc(M3) at the estimated concentrations associated with the enzyme doses of 100 and 50 mg/kg all can decrease the aldicarb concentration to be below ~16.9 μM in 30 minutes, which is consistent with the observed rescue efficacy. In comparison, CocH3-Fc(M3) at the estimated concentration associated with the enzyme dose of 25 mg/kg can only decrease the aldicarb concentration to ~21.7 μM in 30 minutes, higher than ~16.9 μM, which is consistent with the observed partial rescue efficacy.

There was a limitation in our in vivo experiment, as we only monitored the lethality (and whether there were convulsion-like movements) as we used to do in our previous studies^{24, 25} on the cocaine use disorder. Our observation of these alone did not show difference between 50 and 100 mg/kg enzyme treatments because 50 mg/kg CocH3-Fc(M3) is sufficient for the treatment in terms of the lethality. The higher dose (100 mg/kg) is expected to help mice to recover from any possible symptoms of aldicarb poisoning more rapidly. Future studies should also observe any other less-serious symptoms in rescue studies in addition to the lethality and convulsion-like movements.

Further, for the perspective use of CocH3-Fc(M3) in humans, the minimum effective dose of CocH3-Fc(M3) against 0.75 mg/kg aldicarb in mice is 50 mg/kg in terms of the full protection for the lethality. If the same dose level (50 mg/kg) is used for humans, a person with a body weight of 50 kg will require administration of 2.5 g CocH3-Fc(M3). Although the protein drug dose level of 2.5 g is practical in clinical application, it is highly desirable to further improve the catalytic activity of the enzyme against aldicarb. The higher the catalytic activity of the enzyme against aldicarb, the lower the minimum effective dose of the enzyme in practical clinical applications to the aldicarb poisoning treatment.

The discovery of CocH3-Fc(M3) as the first aldicarb hydrolase supports the computationally identified structural factors, i.e. the hydrogen bonds between the carbonyl oxygen of the carbamate and the oxyanion hole, affecting the esterase-catalyzed hydrolysis of aldicarb. The stronger the overall hydrogen bonds with carbonyl oxygen become, the faster the hydrolysis occurs. The similar structure-activity relationship could be useful for esterase-catalyzed hydrolysis of a carbamate in general, as well as other carbonyl-containing functional groups like carboxylate ester, carbonate, and amide. Because these functional groups (carboxylate ester, carbonate, carbamate, and amide) are all popularly used in the design and discovery of prodrugs, the structural insights obtained in this study might also be useful in future rational design of prodrugs to fine tune the hydrolysis rates for controlled drug release from their prodrug forms.

Research Highlights.

• BChE mutants effectively hydrolyze aldicarb when used in poisoning treatment.

• CocH3-Fc(M3) has a 9–28-fold improved bimolecular rate constant with aldicarb.

• CocH3-Fc(M3) has a significant catalytic activity for aldicarb hydrolysis.

• CocH3-Fc(M3) rescued all mice that had received a lethal dose of aldicarb.

Data availability

Data will be made available on reasonable request.