When I discuss medical applications in my talks, I often mention the advantage of targeting cells with active agents, such as toxins, directed by adding and thresholding signals from sensors of different kinds. This is not so necessary in targeting bacteria and viruses, since they look so different from mammalian cells, but could help greatly in recognizing the more subtle differences between normal and malignant cells. This is a natural application for atomically precise self-assembled biomolecular structures of the kind discussed in the Roadmap project.
I was reminded of this by an article in Science that discusses the structure and function of stressosomes, complex mechanisms found in a wide range of bacterial cells. Bacteria respond to several kinds of stress. Since the effects of stress are additive and some of the protective responses are generic, the cells need to respond to the combined effects of multiple stressors. Stressosomes accomplish this, combining inputs from distinct sensors that respond (for example) to ethanol, to salt, and to ultraviolet light. When the aggregate noxiousness surpasses a threshold, the stressosome sounds the alert.
A stressosome differs greatly from a multi-sensor drug-targeting mechanism, but the basic task is similar. A useful multi-sensor targeting mechanism, though, can be smaller and simpler than a stressosome, exploiting the natural additivity of binding energies in multivalent interactions.