Developing a generally applicable electrochemical sensor for detecting macrolides in water with thiophene-based molecularly imprinted polymers

Highlights

• The initiation potentials better meet: functional < linker< cross-linking monomer.

• Tuning initiation potentials by introducing electron-donating or withdrawing groups.

• Developing sensor for specific macrolide by merely replacing corresponding template.

Abstract

Our screening data revealed the threat macrolide antibiotics, especially azithromycin (AZN), posed to human health with its increasing occurrence in water environment. The electrochemical sensor based on molecularly imprinted polymer (MIP) is a promising platform that caters for the next generation of intelligent wastewater treatment plants (WWTPs) by virtue of its wide tolerance to water from all sources and in-situ monitoring. However, low initiation potentials of cross-linking monomers contributed by the electron-rich circumstance allowed them to usurp sites designed for functional monomers when electrically stimulated, leading to an unsatisfactory binding capacity. Another uncertainty is that multiple reaction sites of cross-linking monomers granted them complex polymerization routes and made it difficult to ensure the consistency of preparation. Serval monomers had been investigated with electrochemical tools and the performance of sensors constructed with these monomers were compared in this study. Based on the results, we proposed a protocol in which a novel functional monomer possessing a stronger electron-donating group, phenyl, was adopted to compete for the dominance in electropolymerization. Beyond that, the cross-linking monomer was modified with electron-withdrawing groups to raise its initiation potential. A monothiophene with a moderate initiation potential was also recruited as the linker to address the steric hindrance. In this way, polymerization proceeded in a specific order. It is worth mentioning that the Marangoni flow is an ideal tool to deal with the Coffee-ring deposition while drop-casting. The resulting sensor showed good performance with a limitation of detection (LOD) of 0.120 μM for AZN and a satisfactory selectivity, and the design can be applied to constructing sensors for a variety of macrolide antibiotics.

Introduction

The proliferation of new chemicals like pharmaceutical and personal care products (PPCPs) is challenging the alarm system of traditional WWTPs for ecological risks. Macrolide antibiotics (Figure S1), especially AZN, are emerging antibiotics for the treatment of respiratory infections (Gamerdinger and Deuerling, 2012; Hubble et al., 2019; Vazquez-Laslop and Mankin, 2018). Due to its excellent clinical performance, macrolides entered human society in large quantities, and some of them were excreted into nature. Consequently, macrolide antibiotics frequently appeared in the research reports of water pollutants around the world (Alygizakis et al., 2019; Oberoi et al., 2019; Senta et al., 2019). According to our own screening data in 2019 (Fig. 1), macrolide antibiotics like AZN and roxithromycin (ROX) exhibited worrying health risks. Worse still, the clinical usage of macrolides has risen sharply in the face of raging COVID-19, including the exploration of some new treatment methods (Liu et al., 2020; Nishiga et al., 2020; Sheridan, 2020; Takahashi et al., 2020). The situation forces us to be more alert to the rapid establishment of macrolide resistance. Two terrible outcomes are approaching, that is, a sharp increase in the mortality of drug-resistant diseases and the shortened lifespan of such antibiotics (Chakaya et al., 2021; Rossiter et al., 2017; Zeng et al., 2019), Only accurate concentration of macrolide antibiotics in the environment can properly guide the prescription and drug production, and make it possible to allocate the optimal resources for targeted removal. Thus, long-term monitoring of macrolide antibiotics should be on the agenda.

While the drawbacks of traditional high-accuracy equipment are burdens to township WWTPs when considering upgrading to the next generation of intelligent wastewater treatment plants. At present, the most widely used analysis method for macrolides is LC-MS (Patel et al., 2019; Senta et al., 2019; Terzic et al., 2018; Voigt et al., 2020). Choosing this strategy means to endure its costly equipment, massive manual labor, huge consumption of toxic reagents and extremely poor mobility (Patel et al., 2019). An alternative strategy is to resort to electrochemical methods. The electrochemical approaches do not require excessive reagents and manpower, and the equipment depreciation cost, power consumption and working period are much lower than large-scale analytical instruments, so they are always cost-effective (Clifford et al., 2021; Manjunatha, 2020; Tigari and Manjunatha, 2020). Equipped with these qualities, electrochemical sensor is a platform who meets the requirements of portability, sustainability and sensitivity (Manjunatha et al., 2012; Wongkaew et al., 2019). There have been some related reports, but most of these sensors were only used to determine the concentration of a certain macrolide without interfering substances (Vajdle et al., 2016, 2017, 2020). Also, traditional biosensors are heavily dependent on biological recognition elements that possessing shortcomings like fragile structure, poor long-term stability, reproducibility and high cost (Romanholo et al., 2021). In order to deal with the complex sewage, the selectivity and robustness of the sensor has to reach more stringent standards.

MIP is an ideal identification element as it selectively binds to the target through a specific spatial structure and matched covalent/ non-covalent bonds (BelBruno, 2019). Electropolymerization accurately produces homogeneous MIP on the electrode with adequate energy (Rebelo et al., 2021) in that the thickness of film depends on the charge. As for the unit used to seize macrolides, non-covalent interactions like hydrogen bonds are not supposed to be the only forces we rely on, since there are varieties of small polar molecules competing for binding sites in complex sewage (Kolkman et al., 2021). Taking all these factors into account, we chose thiophenes with boronic acid groups as functional monomers, which specifically form borate ester with the cis-diol moieties of macrolides (Aeridou et al., 2020). Although there have been reports on the preparation of MIP with boronic thiophene (Dabrowski et al., 2016; Sharma et al., 2016; Stoian et al., 2020), there has been no detailed description of considerations from selecting to configurating monomers, and finally producing reliable sensors that can be adjusted to detecting some other contaminants by merely replacing templates so far. In our laboratory, serval monomers had been investigated with electrochemical approaches and the performance of constructed sensors were compared. Based on the results, we discovered an adequate initiation sequence by which monomers could work together effectively. Interestingly, the high electron density of cross-linking monomer made it always oxidized first when the voltage increases, and their multiple oxidizable sites tend to attack each other when activated. Thus, these cross-linking monomers occupied most of the positions on the electrode surface. Specific monomers were selected to ensure the predominant role of functional monomers in polymerization. The monomers here performed their own duties like initiating others, passing through radicals or uniting clusters. After optimizing parameters, selectivity and linear range of reproducible sensors were studied. This work is dedicated to provide a comprehensive guidance on the selection and organization of monomers, and to explore a useful prototype for later workers who want to prepare MIP sensors with thiophene monomers.