Development of Chemical Sensors for Rapid Identification of Amphetamine-Related New Psychoactive Substances

Abstract

A molecular receptor for mephedrone, an amphetamine-like NPS, was developed using host-guest chemistry and pharmacophoric design. The in-field detection of new psychoactive substances (NPS) is an area that has garnered considerable attention in the last few years. With the continuously expanding number of NPS on the market, traditional detection mechanisms lack the selectivity needed. In this project a new methodology has been developed for the design of host molecules for use in in-field detection, based on biomimetic design.

To understand what a sensory molecular needs to be selective against, GC-MS and HPLC analysis were employed to identify and quantify thirteen aminoindane internet samples. It was found that the composition of internet samples varies greatly in terms of concentration of active ingredient, with a range of 17-95 % w/w of active ingredient identified. It was also found that caffeine was the most common cutting agent with a range of 27.7-30.2 % w/w identified. This highlights the need for both selectivity and sensitivity in detection mechanisms.

Using the principles of biomimetic design, a methodology for the treatment of protein-ligand interactions was developed. Protein-ligand binding data collected from the Protein Databank was analysed for mephedrone related structures and common cutting agents, identified through aminoindane internet sample analysis and literature sources. From this work a three-point pharmacophoric model was developed, upon which two host molecules were considered, macrocyclic calixarenes and acyclic anthraquinones. Both contained the three binding interactions deduced from the pharmacophore design; two p-stacking interactions and one hydrogen bond acceptor.

The final host molecule taken forward for testing was 1,8-dibenzylthiourea anthracene (Probe 1). The binding affinity of Probe 1 to mephedrone was tested using 1H-NMR. An estimated association constant of 104 M-1 was calculated, with a 1:1 binding stoichiometry. Along with ESI-MS and DFT calculations, it was found that mephedrone binds to Probe 1 in a concerted fashion with a three-point binding geometry, with two hydrogen bonds and one p-stacking interaction. A modest optical response using fluorescence spectroscopy was also observed between mephedrone and Probe 1 at high molar concentrations. A more pronounced response was observed upon addition of high molar concentrations of flephedrone.

1H-NMR showed that Probe 1 selectively bound mephedrone over methamphetamine as well as the four most common cutting agents identified from literature: lidocaine, caffeine, paracetamol and benzocaine, which have been shown to cause false positives in previous studies. Probe 1 showed significant selectivity for the β-ketoamine arrangement. This is supported by the systematic analysis of mephedrone, methamphetamine, mephedrone precursor and flephedrone. This is the first time this has been achieved using host-guest chemistry. A protocol was developed to successfully detect mephedrone via Probe 1 using NMR spectroscopy in a simulated street sample containing two of the most common cutting agents, benzocaine and caffeine. To further aid future design of small host molecules a methodology for the in silico analysis of small molecule host-guest binding using metadynamics was explored. Solvent interactions with the host and guest molecules were observed, highlighting the importance of solvent choice in binding studies. Metadynamics shows potential to be used in further work for improving the approach in which host molecules are designed in future.