(W1030-02-10) Detection of Xanthines and Several Beta-Lactam Antibiotics Using Surface Modified Lipid/Polymer Membranes

Purpose: Taste sensors employing lipid/polymer membranes are widely used to assess the bitterness of oral pharmaceuticals, especially that of active pharmaceutical ingredients (APIs), which significantly impacts patient adherence [1]. These sensors offer an objective, safe method of evaluation and allow for the ethical testing of toxic, allergenic, or extremely bitter compounds. Among them, the commercially available BT0 sensor has shown strong correlation between its output (BT0 value) and the bitterness response of the human taste receptor hT2R14, as reported in BitterDB [2]. However, the BT0 sensor has limitations: it fails to detect the bitterness of caffeine, which possesses a xanthine skeleton lacking intramolecular charge, and also that of certain drugs, such as oral beta-lactam antibiotics like amoxicillin and cephalexin [3]. Recent insights into human taste perception, particularly regarding allosteric modulation of bitter taste receptors, have revealed that cholesterol is vital for receptor activation and that many bitter molecules act allosterically [4]. Building on these findings, this study aims to detect xanthines and certain oral beta-lactam antibiotics using surface-modified lipid/polymer membranes, incorporating allosteric interaction via hydrogen bonding

Methods: Four xanthine derivatives—caffeine, acefylline, doxofylline, and pentoxifylline—and ten drugs listed in the WHO Model List of Essential Medicines for Children (7th ed., 2019)—acyclovir, amoxicillin, cephalexin, isoniazid, linezolid, propylthiouracil, azathioprine, acetaminophen, carbamazepine, and folic acid—were selected due to their low BT0 sensitivity [3]. Additionally, two commonly used oral cephalosporins not included in the WHO list, cefaclor monohydrate and cefdinir, were tested. Quinine hydrochloride served as a positive control. All drugs were dissolved in a reference solution containing 30 mM KCl and 0.3 mM tartaric acid. Taste sensors with tetradecylammonium bromide (TDAB) membranes were developed to detect non-charged bitter compounds. A mixture of 10 mL of 3 mM TDAB in tetrahydrofuran (THF), 1.5 mL of dioctyl phenylphosphonate (DOPP; plasticizer), and 800 mg of polyvinyl chloride (PVC) was cast into a petri dish and dried at room temperature for three days. The dried membranes were mounted on electrodes and soaked in 0.03 wt% 3-bromo-2,6-dihydroxybenzoic acid (3-Br-2,6-DHBA) for 72 hours to modify the surface. Electrostatic and hydrophobic interactions between positively charged TDAB and the negatively charged carboxyl groups of 3-Br-2,6-DHBA enabled stable adsorption. For comparison, membranes were also modified with 2,6-dihydroxybenzoic acid (2,6-DHBA) and benzoic acid (BA). Measurements were conducted using the TS–5000Z taste sensing system (Intelligent Sensor Technology, Japan). TDAB sensors and AgCl-coated reference electrodes filled with 3.33 M KCl saturated with AgCl were cleaned and immersed in a saliva-mimicking reference solution. After immersion in the sample solution, potential (Vₛ) was measured, and the relative response was calculated as (Vₛ − Vᵣ). Membranes were cleaned between measurements using a solution of 10 mM KOH, 100 mM KCl, and 30% ethanol. Five intra-day replicates were conducted; data from the third to fifth replicates were used for analysis.

Results: As shown in Fig. 1a, the BT0 sensor responded strongly to 0.1 mM quinine hydrochloride but showed no response to caffeine or other non-charged xanthines. In contrast, the 3-Br-2,6-DHBA-modified sensor responded significantly (Fig. 1b): 100 mM caffeine (75 mV), 30 mM acefylline (31 mV), 30 mM doxofylline (25 mV), and 100 mM pentoxifylline (70 mV). Figure 2 shows that the BT0 sensor produced minimal responses to eleven WHO-listed drugs even at 10–30 mM. However, the 3-Br-2,6-DHBA-modified membrane yielded significantly higher responses. Caffeine (30 mM) showed the highest signal (~40 mV), while amoxicillin (10 mM) and cephalexin (30 mM) also produced responses above 20 mV. Other drugs showed little or no response. Further testing of amoxicillin, cephalexin, and acyclovir (control) confirmed that only the 3-Br-2,6-DHBA-modified membrane produced notable responses to the beta-lactams. Membranes modified with 2,6-DHBA or BA showed no response, nor did any respond to acyclovir (Fig. 3). These findings imply that hydroxyl groups in 3-Br-2,6-DHBA contribute to allosteric effects through hydrogen bonding. Additional tests with cefaclor monohydrate and cefdinir showed responses exceeding 30 mV, but only with the 3-Br-2,6-DHBA-modified membrane. This suggests that the modified membrane enables effective detection of xanthines and oral beta-lactam antibiotics.

Conclusion: We successfully detected bitterness signals for caffeine, acefylline, doxofylline, and pentoxifylline, as well as for amoxicillin, cephalexin, cefaclor, and cefdinir—compounds previously undetectable with the conventional BT0 membrane. Our findings indicate that hydrogen bonding between drugs and the hydroxyl groups on the aromatic carboxylic acid enhances binding by altering carboxyl group dissociation. This allosteric mechanism, wherein drug binding at one site promotes hydrogen bonding at another, appears central to improved detection. The increased responses observed with the 3-Br-2,6-DHBA-modified membranes are attributed to specific intra- and intermolecular hydrogen bonding interactions.

References: 1. Toko K., Research and Development of Taste Sensors as a Novel Analytical Tool. Proc Jpn Acad Ser B Phys Biol Sci., 2023. 99(6): p173-189.
2. Haraguchi T., Uchida T., et al The Utility of the Artificial Taste Sensor in Evaluating the Bitterness of Drugs: Correlation with Responses of Human TASTE2 Receptors (hTAS2Rs). Chem. Pharm. Bull. 2018, 66: p71–77.
3. Kojima H., Kurihara T., Uchida T.,.et al A New Bitterness Evaluation Index Obtained Using the Taste Sensor for 48 Active Pharmaceutical Ingredients of Pediatric Medicines. Chem. Pharm. Bull. 2021, 69: p537–547.
4. Kim, Y.; Gumpper, R.H.; Liu, Y., et al Bitter Taste Receptor Activation by Cholesterol and an Intracellular Tastant. Nature 2024, 628: p664–671.

Acknowledgements: The funding source was supported by JSPS KAKENHI Grant Number JP21H05006.


Figure1 Responses of the BT0 sensor (a) and the 3-Br-2,6-DHBA-treated sensor (b) to quinine hydrochloride and non-charged bitter substances. The standard deviations (SD) of the outputs from both types of electrodes, BT0 (n= 6) and the 3-Br-2,6-DHBA-treated sensor (n= 12), were calculated separately and are shown by error bars.


Figure 2a, 2b Responses of the BT0 sensor (a) and the 3-Br-2,6-DHBA-treated sensor (b) to caffeine
and ten drugs listed in the WHO Model List of Essential Medicines for Children. The standard deviations (S.D.s)
of the responses were calculated for both types of sensors: BT0 (n = 6) and the 3-Br-2,6-DHBA-treated sensor
(n = 12).



Figure 3 Responses to acyclovir (used as a reference), amoxicillin, and cefalexin measured by the membranes modified with 3-Br-2,6-DHBA, 2,6-DHBA, or BA (benzoic acid). The standard deviations (S.D.s) of the outputs were calculated for each type of electrode. The mean values and standard deviations (S.D.s) were obtained from 12 electrical response measurements
(n = 4 electrodes × 3 trials).