An acetylcholinesterase-inspired biomimetic toxicity sensor

Abstract

This work demonstrates the ability of an acetylcholinesterase-inspired biomimetic sensor to accurately predict the toxicity of acetylcholinesterase (AChE) inhibitors. In surface waters used for municipal drinking water supplies, numerous pesticides and other anthropogenic chemicals have been found that inhibit AChE; however, there is currently no portable toxicity assay capable of determining the potential neurotoxicity of water samples and complex mixtures. Biological assays have been developed to determine the toxicity of unknown samples, but the short shelf-life of cells and other biological materials often make them undesirable for use in portable assays. Chemical methods and structure–activity-relationships, on the other hand, require prior knowledge on the compounds of interest that is often unavailable when analyzing environmental samples. In the toxicity assay presented here, the acetylcholinesterase enzyme has been replaced with 1-phenyl-1,2,3-butanetrione 2-oxime (PBO) a biomimetic compound that is structurally similar to the AChE active site. Using a biomimetic compound in place of the native enzyme allows for a longer shelf-life while maintaining the selective and kinetic ability of the enzyme itself. Previous work has shown the success of oxime-based sensors in the selective detection of AChE inhibitors and this work highlights the ability of an AChE-inspired biomimetic sensor to accurately predict the toxicity (LD50 and LC50) for a range of AChE inhibitors. The biomimetic assay shows strong linear correlations to LD50 (oral, rat) and LC50 (fish) values. Using a test set of eight AChE inhibitors, the biomimetic assay accurately predicted the LC50 value for 75% of the inhibitors within one order of magnitude.

Introduction

Drinking water sources are continuously subjected to multiple sources of anthropogenic contamination. Wastewater treatment plant effluents and agricultural runoff, for example, have been identified as anthropogenic sources of surface water contamination across the United States (Kolpin et al., 2002). These contaminants can elicit toxic effects in parent form or as a transformation product formed during drinking water treatment (i.e., disinfection) (Kodama et al., 1997, Rule et al., 2005, Bedner and MacCrehan, 2006, Arnold et al., 2008, Duirk et al., 2009). Although the concentration of most anthropogenic chemicals may not elicit a toxicological response at concentrations found in water sources, the combination of pesticides, industrial solvents, etc. at the concentrations present may be additive or exacerbate a toxicological response (McCarty and Borgert, 2006, Rakotondravelo et al., 2006).

With current toxicological methodologies, it is difficult to rapidly assess total potential toxicity of water samples on-site. A novel method capable of rapidly monitoring the potential total toxicity of source and finished drinking water could help mitigate potential health concerns as well as increase the quality of water being delivered to consumers homes. Currently, in vitro biological methods used to assess the toxicity of unknown samples employ live cells, enzymes, or DNA (Cheol Gil et al., 2000, Choi and Gu, 2001, van der Schalie et al., 2001, Kim and Gu, 2003, Zhou et al., 2003, Mitchell and Gu, 2004, Lee and Gu, 2005, Lee et al., 2005, Yoo et al., 2007, Chapin et al., 2008, Edwards et al., 2008, Frampton et al., 2008, Girotti et al., 2008, Ahmad and Moore, 2009, Ma et al., 2009). These techniques have successfully demonstrated the prediction of toxicity and detection of toxic compounds; however, the short shelf-life of cells and other biological molecules make them poor candidates for use in portable devices.

Another challenge in adequately assessing potential health effects from contaminated drinking water is the methods used to quantify the presence of contaminants. Concentrations of either regulated or unregulated drinking water contaminants are generally very low in the aqueous phase (ppb level); therefore, complicated extraction and analytical methods are required to adequately quantify these contaminants. However, knowing the concentration of each analyte does not readily provide any indication of the total potential toxicity of the sample. The need exists to measure total toxicity of contaminant mixtures to alleviate public health concerns from exposure by either recreational use of surface waters or finished potable water.

This work highlights a new development in toxicity assays: replacing biological molecules with stable biological mimics. Biological mimics are chemical molecules that imitate a biological active site (structural mimics) or the catalytic chemistry of enzymes (functional mimics) (Breslow, 1980, Breslow, 1990, Rodriguez, 2012). Structural mimics are an excellent choice for use in toxicity screening because they have the selectivity of a biological molecule paired with the stability of a chemical compound. In this paper, we describe the fabrication of an acetylcholinesterase (AChE)-inspired biomimetic sensor for the prediction of toxicity via AChE inhibition, one of the leading causes of acute chemical toxicity. AChE inhibitors encompass a wide-range of compounds including organophosphate and carbamate pesticides, phenanthrenes, and piper dines.

Previously developed biomimetic sensors for the detection of nitrosothiol compounds in the blood have demonstrated the enhancement in shelf-life from the replacement of biomolecules with biomimics. These biomimetic sensors employ an organoditelluride polymer as a glutathione peroxidase mimic and amperometrically detect the resulting NO produced (Cha and Meyerhoff, 2006, Hwang et al., 2008, Cha et al., 2009). The sensor was stable for over 30 d (at room temperature) and showed no decrease in sensitivity over time (Hwang et al., 2008), extending the shelf-life of an enzyme-based glutathione peroxidase sensor (held at 4 °C) by over 2 weeks (Musameh et al., 2006). Recently, the Monty Research Laboratory has shown that biomimics can effectively replace biological molecules in the prediction of acute toxicological effects (Rodriguez, 2012). Through that work, hepatotoxicity was predicted, within one order of magnitude, using a bio-inspired compound (metalloporphyrin) as a structural mimic for the liver enzyme cytochrome P450 (CYP450).

In order to imitate AChE inhibition, oxime molecules are employed as structural mimics of the AChE active site. Oxime molecules are a class of organic compounds with a strongly ionizable (NOH) group. It has been shown that oxime molecules have a high affinity for phosphorus and can be used to reactivate a phosphrylated AChE enzyme (Green and Saville, 1956). During inhibition of the AChE active site, inhibitors bind (reversibly or irreversibly) to the oxygen atom of the amino acid serine residue. This binding is stabilized by the nitrogen containing histidine residue. Oxime molecules are able to imitate this inhibition by initiating a nucleophilic substitution reaction between the hydrolytic oxygen of the oxime, in close proximity to a stabilizing nitrogen atom (Fig. 1), and the target acetylcholinesterase inhibitor. The oxygen molecule attacks the acetylcholinesterase inhibitor, in this case an organophosphate compound, forming an oxime-inhibitor complex. The oxime-inhibitor complex then decomposes to produce a cyanide ion. The production of cyanide ions can be monitored potentiometrically by the toxicity sensor. Previous work has shown that 1-phenyl-1,2,3,-butanetrione-2-oxime (PBO) has superior reaction kinetics and multi-phase micro-sensors-based on the oxime chemistry, depicted in Fig. 1 – have been successful in the selective detection of parts-per-trillion levels of vapor-phase organophosphates, as well as other related compounds in the presence of complex mixtures (Oh et al., 2008).

In this work a microchannel oxime sensor was tested in order to correlate sensor response to sample toxicity over a wide range of liquid-phase AChE inhibitors. A linear correlation was found for a training set of AChE inhibitors (Table 1) between toxicity (LD50 and LC50) and inverse sensor response, demonstrating that the magnitude of response is indicative of toxicity. The LC50 correlation was then validated by predicting the toxicity of a test set of AChE inhibitors. The predicted toxicity for the AChE inspired assay was within one order of magnitude of the measured LC50. The AChE-inspired biomimetic sensor provides a more accurate prediction of toxicity for the test set than the accepted web-based structure–activity relationship (SAR) Ecological Structure Activity Relationships (ECOSARs) Class Program model developed by the EPA, demonstrating that the sensor can be used as a viable real-world toxicity assay. In addition, the method presented in this work is selective only to AChE inhibitors and shows no response to a variety of common interferents known to give false positives in commercially available detectors.