I think there would be major issues getting folded protein to 5K. Plus proteins "live" in 300K and the oscillations are probably very important for the function (perhaps even more than a single arbitrary structure you would get from a picture like this).
Yes to all of the above. Not to mention that we've long been able to visualize protein molecular structure using X-rays and NMR. This probably isn't very useful to anything related to proteins.
