Skip to main content

Analytical Chemistry

Vibrating Proteins, Resolved

Here’s something that many of us don’t tend to think about when we think about enzymes: vibrational energy. But it’s long been thought that anisotropic vibrational energy transfer (VET) plays a role in both enzyme active sites and in things like coupling to allosteric sites. Getting a handle on that, though, has not been easy – how would you pick out one particular part of a protein and suddenly increase just its vibrational energy?

Well, this new paper has found an ingenious way of doing just that. The group (a multinational effort from Germany, Canada, and Denmark) used an unnatural amino acid, the azulene analog of phenylalanine. Azulene is beloved by those who know it for its striking blue color and odd spectroscopic behavior, and that’s what’s being taken advantage of here. When the molecule is excited by 600nm light, it undergoes an unusual conversion from the singlet excited state, shedding all that energy into vibrational modes. In effect, when you do this to an AzAla-containing protein, you are quickly and selectively heating up that particular side chain, which you have placed wherever you choose in the protein’s structure.

Of course, to realize that last part isn’t necessarily trivial, and in this case the team used a genetic code expansion, evolving a new aminoacyl-tRNA synthetase enzyme to handle the unnatural amino acid. They were able to use this system to put the AzAla residue into a model protein for allostery (PDZ3) and see what happens when vibrational energy gets suddenly injected. As a sensor for that, they also incorporated azido-homoalanine (Aha) residues, whose side chains can be monitored by IR spectroscopy, into the peptide ligand for the protein.

When the team irradiated systems of this kind with 613nm light, monitoring the azide group by IR, they did indeed see an instant VET signal. The azide absorption shifted to a lower wavenumber in two different experiments, one where the AzAla was relatively buried and one when it was on the surface. The latter showed less of an effect, though, because some of the vibrational energy got transferred out into the water instead. It appears that the vibrational energy transfer happens on the order of picoseconds, which is what you’d expect.

So now that this technique is feasible, the next step will be to incorporate this donor/sensor pair into a wider variety of positions and proteins. Eventually, this will allow a map of vibrational energy transfer for a given protein – we’ll be able to see how this propagates, through what parts of the protein structure, and at what speeds. A key part of protein behavior, one that we’ve been fairly blind to, is now opening up – this and new spectroscopic techniques in the area are going to tell us a lot that we’ve never known.

16 comments on “Vibrating Proteins, Resolved”

  1. Bodrell Spicer says:

    Derek, remember your SI units! You mean 600 nm light, not nM.

    1. Derek Lowe says:

      How right you are. I’m home with a cold today; we’ll put that one down to viral infection!

  2. Fred says:

    oh… I thought he meant “600 nano-molar light”… 🙂

    1. Fred says:

      …maybe “homeopathic light” 😉

    2. Kevin H says:

      It’s a measure of irradiation dose. “600 nM light” equals 600 nanomoles of photons per litre of solution.

  3. Grumpyoldgeek says:

    Engineer here, so please excuse my ignorance. Are you saying that these enzymes physically vibrate? At what frequency? And does a group of them vibrate in phase, so the vibrations are easily detected?

    1. Derek Lowe says:

      Oh, all molecules at room temperature are vibrating – lots of different modes involving bonds stretching and bending. In something as large as an enzyme, such vibrations are going to be rather different in different parts of the structure, and propagating through the framework in various ways.

    2. What Derek said. Normal modes. 3N – 6 frequencies (N=number of atoms).

    3. anon says:

      Vibrational frequencies lie in the IR region of the electromagnetic spectrum, ~10^13 Hz. Frequencies vary with the strength of the “springs” (bonds) and the masses of the atoms. And no, they’re not coherent.

  4. Barry says:

    Forty years ago, Dick Zare showed that when you pump the O-H bond of (t)Butylhydroperoxide with a laser, it’s still the O-O bond that ruptures. I.e. vibrational energy equilibrates among the (small) molecule’s bonds/modes faster than any bond can rupture.

    1. MrXYZ says:

      There is a literature on local breathing modes (low frequency, high amplitude) helping to promote enzyme catalysis (particularly involving hydrogen transfer). I wonder if this technology could be use to probe that phenomena or whether the vibrational energy is too high frequency (and equilibrates too fast) to be useful?

  5. electrochemist says:

    Curious how much perturbation to the tertiary and quaternary structure of the protein occurs due to insertion of these unnatural amino acids. Was just thinking about sickle cell disease and the changes that occur to hemoglobin due to a single AA substitution. Still as useful technique, I am sure, but I wonder about its generality….

    1. sciguy says:

      Proteins are big enough, and sufficiently tolerable to mutations (in a very general sense) that this isn’t usually a problem. Caveats: Probe positions must be chosen with care, and some analysis of structure & stability vs wild-type is always essential. (Circular Dichroism works, NMR is better, crystallography is best.)

      Specific to this particular study: Azulenylalanine is a reasonably good homolog of tryptophan; I’ve synthesized structured AzAla test peptides to confirm this. I have no experience with the azido homoalanine, but I figure it could pass for Met or maybe Phe if the surrounding residues are in a forgiving mood. (And this is only an issue if it’s buried in the core.)

  6. David Edwards says:

    I’ve been absent for a while (up to my neck in JavaScript) and I’ve been missing my doses of Derek’s reporting on the ingenuity of chemists. This is a welcome reminder of what I’ve been missing whilst coding!

    Though one tangential thought occurs to me whilst reading this. There are two representations of molecules that I’ve seen in circulation, the classic ball and stick representation (the sticks presumably representing bonds) and the space-filling representation, where the atoms are depicted as intersecting spheres. If the space-filling model is more physically realistic with respect to the actual manner in which the atoms occupy space, then where are the bonds in this model? I suspect I’ve alighted upon a phenomenon that has a horribly complex quantum description that will take even an expert a week to teach me, but I thought I’d ask anyway, in case someone can provide a simpler version, without the simplification degenerating into the outright misleading (I’m aware of that danger at least in this matter). Only the ball and stick model always struck me as being informative about bonds, even if it doesn’t represent space filling properly. It’s not that difficult to picture vibrational modes using the ball and stick model, not is it difficult to think of bonds as having a ‘length’ in that model, whereas in the space-filling model, these aspects of bonds become a lot harder to visualise.

    Though I can imagine the fun and games involved in tracking vibrational propagation in a protein, even with a ball and stick model, and the ingenuity required to do this in a real protein is manifestly considerable. Definitely worth a big hat tip to the researchers involved, and to Derek for bringing this to my attention!

  7. Larry says:

    The significance of coupled protein vibrational modes in enzymatic catalysis was recognized many years ago by the mechanistic cognoscenti, e.g., the late great Stephen Benkovic:
    and other papers cited therein

  8. loupgarous says:

    Could this be of help in studying phenylketonuria, or is a light-emitting analog of alanine not likely to be helpful in studying it?

Comments are closed.