In recent years, the biomedical applications of metal–organic frameworks (MOFs) for drug delivery have attracted increasing attention, largely due to their outstanding advantages, such as high surface area and porosity for high loading of therapeutic agents, along with facile structural modification. However, MOF drug carriers often suffer from uncontrollable burst release and inadequate spatiotemporal control, hindering their clinical application. To address this challenge, we report a novel nanoplatform with photoresponsive properties, ZIF-90-Azo, by postsynthetic modification (PSM) of ZIF-90 with 4-aminoazobenzene groups. This system leverages the reversible trans-to-cis photoisomerization of the azobenzene moieties to construct a sophisticated photogating mechanism for controlled drug release. The resulting ZIF-90-Azo efficiently encapsulates antimicrobial agents within its porous structure. Under ultraviolet (UV) light irradiation, the extended trans-Azo groups isomerize to the bent cis-Azo form, modulating the accessible pore fraction and activating drug release. Conversely, visible-light irradiation reverts the system to its closed state, preventing on-demand release. Kinetic studies reveal a pronounced, reversible divergence in release profiles between the trans and cis states, confirming that azobenzene photoisomerization dictates the diffusion flux. This mechanism enables on-demand drug release with exquisite spatiotemporal precision via external light-triggered control. Furthermore, the material demonstrates a notable synergistic antibacterial effect stemming from the combined action triggered by UV irradiation drug release and the mild intrinsic antimicrobial activity arising from Zn2+ release from the MOF framework. Upon UV light irradiation, the material achieves a 100% antibacterial rate against Staphylococcus aureus (106 CFU/mL). This ZIF-90-Azo system successfully demonstrates a highly efficient, on-demand drug delivery modality, offering a versatile strategy to achieve superior spatiotemporal control.
Wu et al. (2026) studied this question.