Synthesis and Anti-Acetylcholinesterase Properties of Novel β- and γ-Substituted Alkoxy Organophosphonates

Abstract

Activated organophosphate (OP) insecticides and chemical agents inhibit acetylcholinesterase (AChE) to form OP-AChE adducts. Whereas the structure of the OP correlates with the rate of inhibition, the structure of the OP-AChE adduct influences the rate at which post-inhibitory reactivation or aging phenomena occurs. In this report, we prepared a panel of β-substituted ethoxy and γ-substituted propoxy phosphonoesters of the type p-NO₂PhO-P(X)(R)[(O(CH₂)_nZ] (R = Me, Et; X = O, S; n = 2, 3; Z = halogen, OTs) and examined the inhibition of three AChEs by select structures in the panel. The β-fluoroethoxy methylphosphonate analog (R = Me, Z = F, n = 2) was the most potent anti-AChE compound comparable (k_i ~ 6 ×10⁶ M⁻¹min⁻¹) to paraoxon against EEAChE. Analogs with Z = Br, I, or OTs were weak inhibitors of the AChEs, and methyl phosphonates (R = Me) were more potent than the corresponding ethyl phosphonates (R =Et). As expected, analogs with a thionate linkage (P=S) were poor inhibitors of the AChEs.

Organophosphate (OP) compounds (e.g., diazinon, parathion; Fig. 1; P=S) have been widely employed as insecticides in which their mode of toxic action first requires conversion to the more reactive P=O (oxon) form.^{1, 2} Conversely, OP compounds that contain the oxon functional group and possess a leaving group that is readily stabilized upon release (e.g., F⁻, CN⁻) are readily reactive and define the properties present in nerve agents like VX, sarin (GB) and soman (GD) (Fig. 1).^{3, 4}

Figure 1.

Representative OP insecticides and nerve agents.

The mechanism of toxicity of most highly reactive OPs results from the inhibition of acetylcholinesterase (AChE), specifically phosphorylating an active site serine to produce an OP-AChE adduct.^{1, 2, 4} Formation of OP-AChE adducts diminish or halt the hydrolysis of the neurotransmitter, acetylcholine (ACh) causing a surplus of ACh in the neuronal synapse that triggers overexcitation of nicotinic and muscarinic receptors leading to neurotoxic events. Once the OP-AChE adduct is formed it can undergo reactivation (restoration of activity) or aging, which is defined as loss of a phosphoester group leading to an oxyanion that is intractable to reactivation (Scheme 1).^5–7

Scheme 1.

OP-inhibition of AChE and subsequent reactivation and aging reactions.

The reactivation of OP-AChE adducts occurs by nucleophilic displacement at phosphorus (by water, oximes), whereas aging is influenced by branching and/or inductive properties that favor cationic (proton-assisted) mechanisms. For example, dimethoxyphosphoryl-AChE (R = CH₃, Y = OCH₃) undergoes reactivation within 1–2 h (or minutes using oxime antidotes) and ages with a t_½ ~ 2–10 h. Conversely, OP-AChE adducts containing branched alkoxy groups like sarin (R = CH(CH₃)₂) or soman (R = CH(CH₃)(C(CH₃)₃) undergo aging to the methyl phosphonate anion (Y = CH₃) in 3-6 h and 2–5 min, respectively, affording little chance for the OP-adducts to reactivate.⁸

When the OP-ester component attached to AChE is a small, unbranched alkoxy group (R = CH₂CH₃, CH₂CH₂CH₃), these adducts are known to reactivate slowly and do not undergo appreciable aging (t_½ ~ 36 – 48 h) suggesting that these OP-AChE adducts are relatively stable.⁹ Herein, we report the synthesis and anti-AChE activity of OP compounds designed to form analogs of these stable OP-AChE adducts, for example, (ZCH₂CH₂O)(Y)P(O)-AChE and (ZCH₂CH₂CH₂O)(Y)P(O)-AChE (Y = CH₃, CH₂CH₃) in which Z is a halogen or tosylate substituent. It was thought that these customized inhibitors would form OP-AChE adducts in which the Z-substituent is positioned to influence the adduct stability and post-inhibitory reaction pathways. Since there are no systematic reports that prepare AChE-reactive, organophosphate compounds containing β- or γ-substituent phosphoester groups, this study was aimed to identify useful synthetic strategies to these structure supported by preliminary enzymatic evidence demonstrating their potential as enzyme inhibitors.

The OP targets (4 and 5) were designed and synthesized (Scheme 2) to include a number of key features. To produce OPs with useful inhibitory properties, a p-nitrophenol ester was installed as a leaving group. To examine the role, if any, of the alkyl group on the inhibition mechanism, both methyl and ethyl phosphonates were considered as targets. Both phosphonothionate (P=S) and phosphonate oxon (P=O) analogs were prepared. Phosphonothionates are expected to be less reactive toward nucleophilic displacement than oxons, and therefore likely to be poor inhibitors of AChE.¹⁰ However, the thionate P=S bond reduces reactivity at the phosphorus center allowing functional group transformations to more readily proceed elsewhere on the molecule. Substituted, ethoxy and n-propoxy phosphoester groups were considered to evaluate the ease of substituent installation and potentially examine the contribution of inductive effects on inhibition. Additionally, if certain alkoxy substituents (Z) are leaving groups, the alkylation of AChE residues residing in proximity to the active site serine will be examined in a future study. To this end, twenty novel OP structures were produced with the most notable variation occurring in the alkoxy β, γ-substituents as F, Br, I and OTs.

Scheme 2.

(i) p-NO₂PhOH, Et₃N, CH₂Cl₂, 0 °C to rt, 4 h; (ii) 0.5 M LiOH, CH₃CN, rt, 1 h; (iii) Z(CH₂)_nOH, DCC or BOP•PF₆, CH₂Cl₂, rt, 24 h; (iv) Lawesson’s reagent, toluene, 80 °C, 3h.

Compounds 4a–j and 5a–j were first synthesized (Scheme 2) starting from the corresponding phosphonyl chlorides 1a (R = CH₃) and 1b (R = CH₃CH₂). Reaction with p-nitrophenol (PNP) in base affords the bis(4-nitrophenoxy) alkylphosphonates (2a–b) in good yield 87–91%.¹¹ To convert 2a–b to the 4-nitrophenyl, hydrogen alkylphosphonic acids, LiOH, NaOH, KOH, Ba(OH)₂, Ca(OH)₂ were examined. Most conditions formed the hydrolyzed diacids or mixtures of hydrolysis products¹¹ except LiOH (0.5 M aq.) with CH₃CN that led to monoesters 3a–b in moderate yield 51–54%.¹²

Monoesters 3a–b were converted to the target compounds 4a–j by reaction of the corresponding haloalcohol or tosyl alcohol in the presence of a coupling reagent, dicyclohexylcarbodiimide (DCC) or benzotriazol-1-yl-oxytris-(dimethylamino)-phosphonium hexafluorphosphate (BOP•PF₆). Both coupling reactions were successful; however, the use of DCC afforded higher yields.¹³ Coupling of monoesters 3a–b to the alcohols was also attempted using Mitsunobu conditions or water soluble carbodiimides, however, neither worked.^{14, 15} Oxons 4a–j were treated with Lawesson’s reagent to form the corresponding phosphonothionates 5a–j in 60–65% yield.¹⁶

A more straightforward route to the target OPs 4a–j and 5a–j was undertaken by attempting to directly introduce the Z substituent by nucleophilic displacement of a leaving group on the ester side chain (Scheme 3). For example, the bromoethyl ester 4c was treated with silver tosylate to form 4f in 48% yield. We next attempted to prepare the 2-fluoro derivative 4a using tetrabutyl ammonium fluoride (TBAF; 1.0 M aq. THF; 10 min) and the yield was low (13%). We attributed this low product yield to competing nucleophilic substitution at phosphorus and formation of a P-F bond as confirmed by the appearance of a d_F-P in the ³¹P (J = 1074 Hz). To circumvent the P-F bond formation problem, which results in a toxic product, we examined the TBAF reaction using the thionate (P=S) 5f as starting material, which owing to stronger pπ-dπ overlap, would be less reactive toward nucleophiles at phosphorus. When 5f was treated with TBAF, fluoride ion preferably displaced the OTs leaving group rather than the p-nitrophenol ester moiety in 10 min to afford 5a in 33% yield. This yield was acceptable because there was no evidence of P(S)-F bond formation. The β–fluoroethoxy thionate 5a was treated with m-CPBA to obtain 4a in 58% yield.^17–19 In sum, nucleophilic displacement of a side chain leaving group (Z) works with phosphonothionates as substrates but poorly with phosphonates leading to the reaction occurring preferentially at phosphorus producing hazardous phosphonofluoridates. All compounds were characterized by ¹H, ¹³C, ³¹P, ¹⁹F NMRs and HRMS data.^20–23

Scheme 3.

Alternate route to target organophosphates: (i) TsO⁻Ag⁺, CH₃CN, 65 °C, 4 h; (ii) 1.0 M TBAF in THF, rt, 10 min; (iii) m-CPBA, CH₂Cl₂, rt, 30 min.

Many of the new compounds were screened for their inhibitory potency against EEAChE (Table 1), and five were assessed in detail for their anti-cholinesterase activity. Owing to variations in the inhibitor solubility and reactivity, the bimolecular inhibition constants (k_i) for select compounds were determined using two kinetic methods: (a) a set inhibitor concentration [I] incubated with AChE for various time points (time-dependent) and (b) variable [I] incubated with AChE and measured at a set time point (concentration-dependent²⁰) kinetics (Table 1). Compounds 4a and 4b differ only in the alkyl-P bond as methyl- and ethyl phosphonates, respectively, and were examined by concentration- and time-dependent methods. Compounds 4a, 4c and 4f are all methyl phosphonates that vary in the β-ethyl substituent Z and were examined by concentration or time-dependent kinetics. Paraoxon was also examined as a control.

Table 1.

Inhibition rate constants (k_i M⁻¹min⁻¹) for 4a, 4b, 4c, 4f and 5c against EEAChE, rHAChE and RBAChE.

Cmpd	Method	EEAChE	rHAChE	RBAChE
4a	CD	5.90±0.15×106	7.51±0.21×106	6.11±0.25×106
	TD	2.50±0.50×106	1.40±0.59×107	6.16±0.28×106
4b	CD	5.52±0.09×104	6.16±0.16×104	1.16±0.06×106
	TD	8.64±2.20×104	1.02±0.05×105	3.77±0.48×106
4c	CD	8.11±0.29×104	nd	nd
4f	TD	1.14±0.03×105	nd	nd
5c	TD	1.17±0.14×104	nd	nd
paraoxon	CD	1.96±0.05×106	nd	1.58±0.22×108
	TD			3.39±0.14×107

EEAChE: electric eel acetylcholinesterase; rHAChE: recombinant human acetylcholinesterase; RBAChE: rat brain acetylcholinesterase. All k_i values (mean ± SEM) were determined from 2–4 different experiments at 25 C. nd: not determined. CD: concentration-dependent. TD: time-dependent. All rate constants (k_i) were determined by Ellman assay.²¹

First, the difference between methyl- and ethyl phosphonates was examined. The fluoroethoxy methyl phosphonate 4a was found to be a 30 to 100-fold better inhibitor of EEAChE than the corresponding ethyl phosphonate 4b as determined by both time- and concentration-dependent kinetic methods. The inhibitory potency of 4a was k_i ~ 5 × 10⁶ M⁻¹ min⁻¹ toward EEAChE, which is comparable to the highly potent inhibitor, paraoxon.

Changing from 2-fluoroethoxy 4a to the 2-bromoethoxy methylphosphonate 4c led to a 70-fold decrease in inhibitory potency. This large difference in inhibition is somewhat surprising in light of the relatively small inductive and steric differences between F and Br that would be expected. However, additional or unforeseen steric considerations that affect the inhibitory potency other than those indirectly influencing the phosphorus atom cannot be excluded.

Replacing halogen substituents with a tosylate 4f afforded a 50-fold weaker inhibitor of EEAChE than the corresponding fluoro analog 4a but comparable to the bromo analog 4c. This result suggests that the larger tosylate group did not adversely affect the inhibition of EEAChE, whereas altering the phosphonate from methyl to ethyl played as significant role in reducing the reactivity at phosphorus.

None of the phosphonothionates (5: P=S) were inhibitors of EEAChE (k_i < 10² M⁻¹min⁻¹) except 5c that blocked enzyme activity at only seven-fold less than that of the corresponding oxon (P=O) (5: P=S). Typically, oxons are 100-fold more potent anticholinesterases than the corresponding thionates.¹⁰

Since the methyl β-fluoroethoxy phosphonate 4a results in an OP-AChE adduct that closely resembles inhibition by the nerve agent VX, the concentration- and time-dependent inhibition of two mammalian acetylcholinesterases, recombinant human and rat brain, was also examined. For comparison, the ethylphosphonate analog 4b and paraoxon were also studied. In general, the methyl β-fluoroethoxyphosphonate 4a was a marginally more potent an inhibitor of the mammalian enzymes than that of EEAChE. As found for EEAChE, compound 4a was a 100-fold more potent inhibitor of rHAChE than the ethyl analog 4b. To our surprise, the methylphosphonate 4a and ethyl phosphonate analog 4b showed comparable inhibition values (~ 10⁶ M⁻¹min⁻¹) against RBAChE suggesting that this enzyme is less sensitive to the added occlusion and inductively electron-donating properties of the ethyl group at phosphorus. Compounds 4a and 4b were 25 to 100-fold less potent inhibitors of RBAChE than paraoxon by their respective concentration-dependent analyses.

As noted, the (CH₃)(CH₃CH₂O)P(O)-serine adduct neither reactivates nor ages rapidly, which may be due to a combination of inductive and steric effects that reduce reactivity at the phosphorus atom and also reduce the propensity to form cation-like intermediates thought to underscore the aging mechanism. Since the phosphylation of AChE inhibition by compounds 4a, 4b, 4c, 4f, and 5c occur with concomitant loss of the p-nitrophenoxy group, the resulting OP-AChE adducts closely resemble that formed from VX, namely, the (CH₃)(CH₃CH₂O)P(O)-serine adduct, which differs only in the retained β-substituent.²² For example, reaction of AChE with 4a results in phosphylation at ser-200 to produce a (Me)(FCH₂CH₂O)P(O)-serine adduct.

In summary, a series of β- and γ-substituted alkoxy methylphosphonates were successfully prepared. The synthetic pathway is highlighted by a new approach that allows selective nucleophilic displacement at a phosphoester side chain rather than at a reactive phosphorus center. The enzymatic analyses demonstrated that the β-fluoroethoxy analog 4a was the best inhibitor of acetylcholinesterases of those tested, and comparable in strength to paraoxon. The kinetic analyses also revealed two noteworthy results. First, the size of the alkoxy group substituent does not adversely affect the inhibitor strength. Second, the phosphonothionate analog 5c was shown to possess anti-cholinesterase activity whereas P=S structures are normally inactive. The data produced in this study now makes possible advanced studies to further elucidate the mechanism of reactivation and aging by alkylphosphonates bearing substituted esters.