The fact that you can tack on an extra protein domain to each protein subunit is really neat because it lets you assemble little macromolecules containing several custom functionalities. Presumably these extra domains face outwards, coating the surface of the cage. Hopefully these extra domains don't interfere with the cage assembly process.<p>For example, say you created fusions between the cage mononmer and antibody Fv chains that bind two different proteins (A and B) and then you created fusions between the cage monomer and two different fluorescent proteins (GFP and RFP).<p>In bacteria strain 1, you express fusion A and fusion GFP so that you assemble cages that will show a green signal where-ever A is found (e.g., for fluorescent microscopy). Likewise, in bacterial strain 2, you expression fusions B and RFP so that you get a red signal where ever B is. You could use these as detectors in a western blot, for example. (Obviously there are better, established ways to detect proteins of interest in a western blot, but that's just a simple idea).<p>Another idea that I remember reading about was creating glucuronidase fusions to antibody fragments designed to bind cancer cells. The patient would then be administered the glucuronide of a cytotoxic drug, which is not toxic in its glucuronide form. The idea is that the antibody fusion protein would activate the drug so that it would be concentrated highly near the cancer cell and nowhere else.