Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

How do conformational changes take place?

  1. Jul 10, 2017 #1
    Hi everyone,

    I don't understand how ATP binding, hydrolysis, and dissociation actually causes movement within a molecule, say a protein. I'd like to understand this in terms of seeing ATP and the protein themselves as molecular orbitals and changes in energy states of electrons.
    Take this example: http://www.nature.com/nm/journal/v18/n10/fig_tab/nm.2924_F5.html
    ATP binds myosin, causing a converter domain to rotate and a switch loop to close, forming a system of hydrogen bonds and some more rotations occur, leading to ATP hydrolysis. With ADP + Pi bound, myosin now binds actin, after which Pi dissociates, causing the lever arm to move, et cetera et cetera.
    I understand ATP hydrolysis provides the (thermal?) energy required for these changes to take place, but I have a gap in knowledge how. It just makes the electrons of the myosin vibrate more? And how are the physical movements taking place?

    Many thanks
     
  2. jcsd
  3. Jul 10, 2017 #2

    Ygggdrasil

    User Avatar
    Science Advisor

    From a thermodynamic point of view, ATP hydrolysis makes one step of the thermodynamic cycle irreversible so that the molecule will traverse the cycle only in one direction.

    From a mechanical point of view, ATP generally binds to motor proteins in a pocket in the interior of the protein. The different states associated with ATP hydrolysis (ATP --> ADP + Pi --> ADP) will have very different shapes that can cause changes to the shape of the pocket surrounding the nucleotide. For example, after ATP hydrolysis, you have the beta and gamma phosphate groups moving apart, which can induce changes to the surrounding protein, as can the reduction in size of the pocket that occurs when the phoshpate leaves. These changes to the shape of the nucleotide binding pocket can propagate throughout the protein, causing larger conformational changes throughout the protein. The nucleotide binding pockets of ATPases often occur at the interface between different subdomains of the protein, so changes to the shape of the nucleotide binding pocket will greatly change the quaternary structure of the protein.

    Aside from the chemical catalysis of the ATP hydrolysis reaction, I believe most of the changes are essentially mechanical and do not require understanding the electronic states of the atoms in the protein.
     
  4. Jul 13, 2017 #3
    When a ligand binds protein, say diacylglycerol binding PKA, then the molecular orbitals of the PKA must change, either in character or shape, correct?
    Because the shape of PKA changes, so thus its MO must be changing.
    Why is this? Is it simple electron (orbital) repulsion? When a ligand binds, hydrogen bonds (and van der waals and ionic bonds?) within the protein, in favour of new ones with the incoming ligand. So basically electrons are moving, from being shared internally to being shared with a ligand?
     
  5. Jul 13, 2017 #4
    One thing to keep in mind is that proteins are not static entities - the level of ordering and rigidity can vary dramatically, with some proteins being like rocks, while others can be unstructured except in certain conditions. While I'm not super familiar with that particular example with PKA, I do recall a nice study from a while back looking at PKA interacting with a nucleotide and a protein substrate, where nucleotide binding shifted the population of open and closed states of the PKA enzyme which facilitated substrate binding and catalysis.
     
  6. Jul 14, 2017 #5

    Ygggdrasil

    User Avatar
    Science Advisor

    A good way to think about protein conformational dynamics is through energy landscape theory. There is a multidimensional energy landscape describing the thermodynamic stability of the different protein conformations, and this energy landscape can change upon binding to the ligand. These energy landscapes account for electrostatic, van der Waal, and other intermolecular interactions between atoms in the protein and in the ligand.

    Here is a nice review article discussing how conformational change influences intermolecular interactions from the perspective of energy landscape theory: https://www.nature.com/nchembio/journal/v5/n11/full/nchembio.232.html
     
  7. Aug 23, 2017 #6
    It seems that what happens when a protein is changing shape is that an electron pair on the end of a C-O (for example) of the peptide backbone comes into proximity with the N-H of another region of the backbone, and because N-H lacks electron density while C-O is rich in it, a lone pair of the C-O (are lone pairs in p orbitals?) donates electron density towards the 1s orbital of the H, which is largely empty.
    So basically, the shape of the electron orbitals changes, there's a change in electron density, and this causes likewise changes across the protein, due to moving electron densities and changes in electron orbital distributions. This causes force somehow (due to their charges being forced together, right?), moving the protein. And hydrophobic and Van der Waals forces also contribute. This can all be represented on an energy landscape like Ygggdrasil pointed out.
    (again, so basically the answer is that electron density and the orbital shapes are changing, correct)?
     
  8. Aug 25, 2017 #7

    Ygggdrasil

    User Avatar
    Science Advisor

    This is better known as hydrogen bonding and can be modeled as a dipole-dipole interaction.

    Electron densities and orbital shapes of molecules will always change in any process. For example, the idea behind van der Waals interactions is that fluctuations in electron density will create temporary dipoles that will pull on the molecular orbitals of nearby molecules to create induced dipoles, resulting in attractive dipole-dipole interactions.

    The bigger question is whether you need to consider details about molecular orbitals in order to model the process or whether some simplifying assumptions that paper over some of the quantum-mechanical details (e.g. modeling the combined effects of van der Waals interactions and steric repulsion via a Lennard-Jones potential). For the most part, one does not need to consider how molecular orbitals are changing in order to understand protein conformational changes, though because most of the forces that hold a protein together are electrostatic, understanding how molecular orbitals and electron densities affect these interactions, in general, is useful for understanding protein mechanics.

    As an analogy, consider a ball flying through the air. Clearly interactions between the ball and the air molecules surrounding the ball are critical to creating the air resistance that will influence the flight of the ball. In these collisions and interactions, the molecular orbitals of the molecules in the air and the ball are surely changing as well. However, no one would model the flight of a ball by modeling every single molecular orbital of each air molecule in the flight path of the ball. Instead, one can simply use a phenomenological model of air resistance (resisting force proportional to velocity squared) rather than trying to model air resistance from first principles.
     
  9. Aug 25, 2017 #8
    One thing that I think should be emphasized is, while ligand binding, hydrolysis, etc. can make certain conformations more favorable, the force that ultimately causes the proteins to move is through collisions with neighboring molecules. Thermal interactions provide the kick the protein needs to navigate its energy landscape. I like to think that the various biochemical changes a protein goes through "terraform" the energy landscape while the actual point the protein is on the landscape is (locally) dictated by thermal fluctuations.
     
  10. Sep 1, 2017 #9
    Thanks for the replies so far... What I'm trying to do here is get to the root of the question of what really is happening when certain objects change shape, and proteins are just one example.
    The wavefunctions of each electron in the C=O and the N-H concerned are synonymous with the orbital of the hydrogen bond formed in between right?
    It's only that they're individual one moment and one wavefunction, the hydrogen bond orbital, the next when the bond is formed, right?
    And when an electron from one species moves to form a bond with another species, that's because the empty orbital that was filled was of the lowest energy available correct?
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted