Probing Passive Permeation of Tetracycline: Are Simulations Ready for beyond-Rule-of-Five Drug Permeability Calculation?

Abstract

Passive permeation across lipid membranes is a key determinant of drug bioavailability and efficacy. Accurate computational estimation of drug permeability is essential for rational drug design, yet remains challenging, particularly for ionizable and beyond-Rule-of-Five (bRo5) compounds. In this study, we employ advanced molecular simulations and the inhomogeneous solubility-diffusion model to calculate the effective permeability and elucidate the membrane permeation mechanism of the antibiotic tetracycline (TC), the six hydrogen-bond donors of which violates one of Lipinski’s Rule-of-Five. By integrating the pH-partitioning and Boltzmann-weighted average potential schemes and accounting for both its neutral (TCN) and zwitterionic (TCZ) tautomers, we show that the dominant contribution to the effective permeability of TC arises from TCN, despite its low abundance. This result is attributed to the relatively small permeation barrier of TCN, explaining why the antibiotic exhibits moderate effective permeability even though it predominantly exists in the zwitterionic form at neutral pH. A further systematic investigation reveals that membrane patch size significantly impacts permeability estimates for TC, in contrast to the relative insensitivity observed for three other permeants. This unique sensitivity can be attributed to the hydrogen-bond network formed between TC and its lipid environment, with the smallest 32-POPC patch artificially raising the drug molecule’s permeation barrier and the largest 256-POPC patch exhibiting significant hysteresis that compromises the quality of the one-dimensional free-energy calculation. Overall, our results suggest that while probing the passive permeation of bRo5 drugs by molecular simulations appears increasingly feasible, the protonation and tautomeric states of the permeants, the uncertainty in their microscopic acid dissociation constants, as well as the potential impact of membrane patch-size effects need to be fully considered in permeability predictions for these complex molecules.