Gallery of Conformational Changes

I've been developing force-generation methods for MD simulations, which generates large conformational changes in picoseconds of simulations. These simulations can be used to access regimes of motion that would not be possible otherwise.

For experienced MD simulators: the background fluctuations may seem rather large! This is because implicit solvent models experience no viscosity.

The Ligand-Binding Loop of Triosephosphate Isomerase

Triosephosphate Isomerase (TIM) is a classic β-barrel domain. This fold provides a sturdy scaffold for catalytic reactions as the middle of the barrel forms a great catalytic site where sidechains have easy access to the ligand.

From analysis of X-ray structures, TIM has a ligand-binding loop that moves 6 to 7 Ångstromss. Here I induce the loop to move by applying a self-adaptive rotational force on the loop (residues 166-176). The loop rotates until it hits the body of the protein. This simulation is run in AMBER in GB/SA implicit solvent, where the loop experiences no viscosity, and surface residues experience a lot of thermal noise.

I repeated the calculation using explicit solvent in GROMACS. Notice that there is much less stochastic motion, and the loop motion is much smoother.

The Ligand-Binding Loop of GAD67

With the TIM example above, we knew what the conformations were. A more interesting case is to apply the loop motion technique to a new structure. Here I generate the loop-opening of the substrate binding loop in GAD67, a GABA synthase.

Allosteric Helix-12 of the Ligand-Binding Domain of the Estrogen Receptor

The Estrogen Receptor triggers large-scale nuclear activity upon binding of Estrogen in the ligand-binding domain. Binding of ligand involves the closing of Helix-12 over the ligand. In vivo, it appears that HSP90 is needed to keep Helix-12 from closing onto the ligand-binding site when the site is empty. In this simulation starting from the closed conformation of the protein, applying RIP to a highly conserved Trp knocks out the entire Helix-12, which is otherwise well-bound to the protein.

The Histidine Flip in the DNA-Dinding-Domain of the Glucocortocoid Receptor

The Glucocortocoid Receptor is another nuclear receptor like the Estrogen Receptor. Here we have the DNA-binding domain (DBD), which, surprisingly, can induce different conformations by simply binding to slightly different DNA sequences. A key histidine in the lever arm (near the middle of the movie) flips out depending on the DNA sequence. Here, I apply a domain pushing force, which simulates the effect of a longer DNA sequence. This flips the histidine in the lever arm.

The Mechanical Unfolding of Titin

The mechanical unfolding of the titin I27 domain has been studied extensively in Atomic Force Microscopy experiments. The I27 domain is mechanically pulled on the N-terminus and C-terminus in opposite directions.

If we pull with a slow enough velocity, we can see the protein trapped transiently in the metastable state before the hydrogen bonds are ripped apart and the protein unfolds.