Heating up a Protein, One residue at a time

Protein structures are fascinating things. Irregular in shape, but sturdy in design, proteins can undergo enormous changes in structure – they bend, spin, expand and pince. As such, proteins are sometimes called nano-machines. If we are lucky, we can catch the protein in flagranto, as certain small reactive molecules can induce a protein to switch between different motions. Some motions are very subtle.

Over the last year, I've been working on simulation methods to probe motions that involve only very slight changes in the structure of a protein. The work I've been doing is an extension of earlier work of a previous postdoc, Nobo Ota, in the Agard lab. Nobu had originally come up with the brilliant idea of freezing a protein, and then heating the protein at one single residue. By following where the energy would flow, he hoped to catch a glimpse of the subtle transfer of energy, and thereby understand something about subtle motions in the protein. He developed a method that he called Anisotropic Thermal Diffusion.

Nobu studied the PDZ domain from the 3rd domain of PSD-95 [1BE9], a very well studied protein. He first froze the PDZ protein to T=10K. Then, he pumped energy into a particular protein (green outline) at room temperature:

He could see tiny fluctuations in energy, which could be measured with sidechain RMSD. The actual motion is tiny, and artificial position constraints needs to be applied to the backbone to accentuate the heating of the residues.

Over the last year, I've come up with a novel way of heating up a residue, which I've called Rotamerically Induced Perturbation, or RIP for short. In RIP, the energy is pumped in using only the rotamer degrees of freedom where the the rotamer velocity is explicitly scaled to that of for room temperature. All other motions are damped out.

Using RIP, the residue moves around a lot more, inducing lots of interesting motions in the rest of the protein. One of the really nice things about RIP is that the energy is cleanly pumped into the system. Even though there are no backbone constraints, residues on the surface don't hit anything and won't induce spurious energy transfer along the backbone:

Given that RIP perturbs the protein cleanly, we can use RIP to really smack the protein around and see if we shake off some loops that are not locked down tightly by the intra-molecular interactions. Here, we first heat the protein to room temperature, then we apply a RIP equivalent to T=3000K in the rotamers: