About this Author
DBL%20Hendrix%20small.png College chemistry, 1983

Derek Lowe The 2002 Model

Dbl%20new%20portrait%20B%26W.png After 10 years of blogging. . .

Derek Lowe, an Arkansan by birth, got his BA from Hendrix College and his PhD in organic chemistry from Duke before spending time in Germany on a Humboldt Fellowship on his post-doc. He's worked for several major pharmaceutical companies since 1989 on drug discovery projects against schizophrenia, Alzheimer's, diabetes, osteoporosis and other diseases. To contact Derek email him directly: Twitter: Dereklowe

Chemistry and Drug Data: Drugbank
Chempedia Lab
Synthetic Pages
Organic Chemistry Portal
Not Voodoo

Chemistry and Pharma Blogs:
Org Prep Daily
The Haystack
A New Merck, Reviewed
Liberal Arts Chemistry
Electron Pusher
All Things Metathesis
C&E News Blogs
Chemiotics II
Chemical Space
Noel O'Blog
In Vivo Blog
Terra Sigilatta
BBSRC/Douglas Kell
Realizations in Biostatistics
ChemSpider Blog
Organic Chem - Education & Industry
Pharma Strategy Blog
No Name No Slogan
Practical Fragments
The Curious Wavefunction
Natural Product Man
Fragment Literature
Chemistry World Blog
Synthetic Nature
Chemistry Blog
Synthesizing Ideas
Eye on FDA
Chemical Forums
Symyx Blog
Sceptical Chymist
Lamentations on Chemistry
Computational Organic Chemistry
Mining Drugs
Henry Rzepa

Science Blogs and News:
Bad Science
The Loom
Uncertain Principles
Fierce Biotech
Blogs for Industry
Omics! Omics!
Young Female Scientist
Notional Slurry
Nobel Intent
SciTech Daily
Science Blog
Gene Expression (I)
Gene Expression (II)
Adventures in Ethics and Science
Transterrestrial Musings
Slashdot Science
Cosmic Variance
Biology News Net

Medical Blogs
DB's Medical Rants
Science-Based Medicine
Respectful Insolence
Diabetes Mine

Economics and Business
Marginal Revolution
The Volokh Conspiracy
Knowledge Problem

Politics / Current Events
Virginia Postrel
Belmont Club
Mickey Kaus

Belles Lettres
Uncouth Reflections
Arts and Letters Daily
In the Pipeline: Don't miss Derek Lowe's excellent commentary on drug discovery and the pharma industry in general at In the Pipeline

In the Pipeline

« Silicon In Drug Molecules: Not Quite There | Main | Biotech Takeover Targets »

November 8, 2012

Picosecond Protein Watching

Email This Entry

Posted by Derek

We're getting closer to real-time X-ray structures of protein function, and I think I speak for a lot of chemists and biologists when I say that this has been a longstanding dream. X-ray structures, when they work well, can give you atomic-level structural data, but they've been limited to static time scales. In the old, old days, structures of small molecules were a lot of work, and structure of a protein took years of hard labor and was obvious Nobel Prize material. As time went on, brighter X-ray sources and much better detectors sped things up (since a lot of the X-rays deflected from a large compound are of very low intensity), and computing power came along to crunch through the piles of data thus generated. These days, x-ray structures are generated for systems of huge complexity and importance. Working at that level is no stroll through the garden, but more tractable protein structures are generated almost routinely (although growing good protein crystals is still something of a dark art, and is accomplished through what can accurately be called enlightened brute force).

But even with synchrotron X-ray sources blasting your crystals, you're still getting a static picture. And proteins are not static objects; the whole point of them is how they move (and for enzymes, how they get other molecules to move in their active sites). I've heard Barry Sharpless quoted to the effect that understanding an enzyme by studying its X-ray structures is like trying to get to know a person by visiting their corpse. I haven't heard him say that (although it sounds like him!), but whoever said it was correct.

Comes now this paper in PNAS, a multinational effort with the latest on the attempts to change that situation. The team is looking at photoactive yellow protein (PYP), a blue-light receptor protein from a purple sulfur bacterium. Those guys vigorously swim away from blue light, which they find harmful, and this seems to be the receptor that alerts them to its presence. And the inner workings of the protein are known, to some extent. There's a p-courmaric acid in there, bound to a Cys residue, and when blue light hits it, the double bond switches from trans to cis. The resulting conformational change is the signaling event.

But while knowing things at that level is fine (and took no small amount of work), there are still a lot of questions left unanswered. The actual isomerization is a single-photon event and happens in a picosecond or two. But the protein changes that happen after that, well, those are a mess. A lot of work has gone into trying to unravel what moves where, and when, and how that translates into a cellular signal. And although this is a mere purple sulfur bacterium (What's so mere? They've been on this planet a lot longer than we have), these questions are exactly the ones that get asked about protein conformational signaling all through living systems. The rods and cones in your eyes are doing something very similar as you read this blog post, as are the neurotransmitter receptors in your optic nerves, and so on.
This technique, variations of which have been coming on for some years now, uses multiple wavelengths of X-rays simultaneously, and scans them across large protein crystals. Adjusting the timing of the X-ray pulse compared to the light pulse that sets off the protein motion gives you time-resolved spectra - that is, if you have extremely good equipment, world-class technique, and vast amounts of patience. (For one thing, this has to be done over and over again from many different angles).

And here's what's happening: first off, the cis structure is quite weird. The carbonyl is 90 degrees out of the plane, making (among other things) a very transient hydrogen bond with a backbone nitrogen. Several dihedral angles have to be distorted to accommodate this, and it's a testament to the weirdness of protein active sites that it exists at all. It then twangs back to a planar conformation, but at the cost of breaking another hydrogen bond back at the phenolate end of things. That leaves another kind of strain in the system, which is relieved by a shift to yet another intermediate structure through a dihedral rotation, and that one in turn goes through a truly messy transition to a blue-shifted intermediate. That involves four hydrogen bonds and a 180-degree rotation in a dihedral angle, and seems to be the weak link in the whole process - about half the transitions fail and flop back to the ground state at that point. That also lets a crucial water molecule into the mix, which sets up the transition to the actual signaling state of the protein.

If you want more details, the paper is open-access, and includes movie files of these transitions and much more detail on what's going on. What we're seeing is light energy being converted (and channeled) into structural strain energy. I find this sort of thing fascinating, and I hope that the technique can be extended in the way the authors describe:

The time-resolved methodol- ogy developed for this study of PYP is, in principle, applicable to any other crystallizable protein whose function can be directly or indirectly triggered with a pulse of light. Indeed, it may prove possible to extend this capability to the study of enzymes, and literally watch an enzyme as it functions in real time with near- atomic spatial resolution. By capturing the structure and temporal evolution of key reaction intermediates, picosecond time-resolved Laue crystallography can provide an unprecedented view into the relations between protein structure, dynamics, and function. Such detailed information is crucial to properly assess the validity of theoretical and computational approaches in biophysics. By com- bining incisive experiments and theory, we move closer to resolving reaction pathways that are at the heart of biological functions.

Speed the day. That's the sort of thing we chemists need to really understand what's going on at the molecular level, and to start making our own enzymes to do things that Nature never dreamed of.

Comments (13) + TrackBacks (0) | Category: Analytical Chemistry | Biological News | Chemical Biology | Chemical News


1. Chemjobber on November 8, 2012 10:50 AM writes...

Wasn't it Jeremy Knowles who said that “Taking a photograph of a horse does not necessarily tell you how fast it can run"?

Permalink to Comment

2. Canageek on November 8, 2012 11:07 AM writes...

The big problem I see with this is that you are still looking at a crystal structure, not the solution-state conformation. Now, I may be biased by having done my undergrad thesis with a group doing protein NMR chemistry, but I think a much more promising area is the wealth of NMR techniques that are appearing: We were implementing several new techniques from other groups when I was there, and had just published one of our own (chemical shift covariance analysis [CHESCA]) they were quite proud of and was getting some good attention. They have a follow up technique in the works they are even more exited about.

XRD is cool and all, and you get lovely images, but what is the point of the timescale resolution if you are still in a weird crystal-state?

Also, why does everyone use X-ray diffraction on proteins, when neutron diffraction works so much better on light elements? Sure, you need a bigger crystal, and there aren't a lot of places you can do it, but there aren't many places you can do synchrotron work either.

Permalink to Comment

3. couldageek on November 8, 2012 12:31 PM writes...

This method surely brings dynamics to crystallography, great work.

@Canageek: Neutron diffraction works so much better on light elements in theory. Though you need crystals of ~1mm^3 - and a suitable beam source, from i.e. a nuclear reactor. Now for the majority of proteins readily giving these crystals sizes, we already know their structure quite well. Challenging are the ones, which do not bring on such nice and homogeneous crystals, let alone them not crystallizing at all.

Now NMR and especially Solid-State-NMR are all pretty much evolving towards solving structures of large proteins and protein complexes. But look at all the labeling effort you have to put in, as soon as there are similar subunits getting into the picture.

Then there's cryo-EM, an actual single molecule technique, and it also offers insights into molecule dynamics in solvate environments. (There are some really great publications on spliceosomes and ribosomes out there.)

And SAXS...and even others...

I don't see any big problems. What I do see is the option to use all these techniques complimentary. It may finally lead to an answer to the question, how much impact our investigation methods actually have on conformation and (maybe) structural dynamics. These different techniques do not replace each other and they surely are not superior to well-established ones, until we know better, why we are seeing, what we are seeing there.

Permalink to Comment

4. leftscienceahwileago on November 8, 2012 12:54 PM writes...

Canageek, remember that a lot of X-Ray structures are solved 'at home' (on a rotating anode source). Growing crystals can be pretty easy and a relatively efficient use of material (a few mgs to run a screen, with a chance you will get a complete high res structure).

I wonder what happened to the "Compact Light Source" that was supposed to be a synchrotron source meant to provide tunable X-Rays for phasing at home....I think the beam intensities never got there...

Permalink to Comment

5. MacDeezy on November 8, 2012 5:44 PM writes...

I have to agree with Canageek with regards to the limitations of crystal although I too am biased by being presented with solution state protein NMR techniques.

Also Deuterium exchange Mass Spectroscopy can be a complementary technique for observing "in situ" dynamics that was overlooked in this thread so far...

Permalink to Comment

6. MJ on November 9, 2012 12:52 AM writes...

While the *in vivo* NMR crowd can't be accused of this, there's plenty of solution NMR where they find the friendliest possible buffer and gleefully crank down the salt concentration and/or pH in order to get pretty spectra. Of course, the *in vivo* stuff is very neat - see, for example Marc Baldus & coworkers earlier this year in PNAS on looking at membrane proteins in whole cells by solid state NMR.

Neutron scattering (whether it's for diffraction, SANS, or spectroscopy) is isotope-dependent, last I checked. This is in contrast to x-ray scattering, where the scattering relates to the electron density, which varies with atomic number. I wouldn't say it uniformly works "better" for lighter elements - going back to my earlier sentence about its isotope dependence, boron-10 has a rather high neutron absorption cross section, while boron-11 has a very modest one. I'm sure I can dig up other illuminating examples once I can find my ILL data booklet. I think the issue is that since neutrons scatter off the nuclei and not the electrons, you're going to be able to detect light elements that don't scatter x-rays well, especially in compounds where there are heavy elements present.

Cryo-EM is a single molecule technique? I think you're referring to single particle analysis, which I'm not sure I'd put in quite the same category. I thought the idea was to examine lots of individual (noisy) particles, align as necessary, and reconstruct a (less noisy) particle structure. This is in contrast to "single molecule spectroscopy" as most chemists/physicists would likely recognize, where the goal is usually to elucidate the heterogenity in a sample of interest that is obscured by bulk/ensemble measurements. I'm willing to be persuaded, though.

Otherwise, it's always nice to see crystallographers actually using a wild-type protein and not doing stuff at 100 K. Heh. One of the authors, Phil Anfinrud, has been doing time-resolved studies like this for quite a while now that I can recall.

Permalink to Comment

7. Claire on November 9, 2012 5:24 AM writes...

How often are solution structures significantly different to crystal structures?

The 'solution structures are more realistic' comment seems to crop quite often and seems fairly logical, but I've never seen it backed up with data. Mind you, I haven't gone looking for that data - the only example I've come across of is barnase and in that example, the two structures seem to overlay pretty well.

Permalink to Comment

8. Derek Lowe on November 9, 2012 9:05 AM writes...

Here you go, Claire:

The short answer: there are differences, some of which are probably real, and some of which are probably artifacts of the measurement technique. Overall, I prefer solution structures, too, but I think that the relative convenience of X-ray is a strong point in its favor. I've certainly worked with a lot more X-ray structures than I ever have NMR structures.

Permalink to Comment

9. Anonymous on November 9, 2012 2:23 PM writes...

One of the big problems with using crystals to study protein dynamics is that the conformational changes occurring in the protein have to be small enough that they don't destroy your crystals in the process.

Permalink to Comment

10. yonemoto on November 9, 2012 4:54 PM writes...

Always beware of solvent accessible residues (especially the possibility of salt bridges). The crystal structure of insulin misses an intramolecular salt bridge in the b-chain that is not there because the lysine is making a salt bridge with the next unit element.

It's actually a conserved lysine (which you wouldn't know why) - except in murine insulin 2. Ironically, murine insulin 2 is used as the de facto substitute for human insulin because it has more identity (statistics of small numbers, IIRC it's 2 changes vs 3 changes) to human insulin.

So also don't trust bioinformatics.

Permalink to Comment

11. dave w on November 9, 2012 5:21 PM writes...

Re: the "Compact [Synchrotron] Light Source"... found this link:

Permalink to Comment

12. leftscienceawhileago on November 10, 2012 1:14 AM writes...

dave w,
Yes, I saw that press release 6 years ago as well. I am not aware if a CLS is actually running anywhere...

Permalink to Comment

13. Canageek on November 18, 2012 1:16 PM writes...

couldageek, MJ: Thanks for your explanations. As a note, you don't need a nuclear reactor to get a neutron beam, there are neutron generators you can purchase that produce neutron streams by accelerating deuterium into tritium. I haven't heard of anyone hooking one of these up to a scattering detector though. That could be because they are rather hard to get approval for; A lab I was in hired an undergrad for the summer whose only job was working on getting CNSC approval to purchase on of those.

Also, why is isotope dependence a problem? Hydrogen scatters neutrons better then it does x-rays, what with only 1 electron and all, and you can even see hydrides with it (Not a problem in proteins, I know). Deuterium and tritium have even better neutron cross sections as I recall.

The only other ones proteins worry about are CNO, right? I thought C-12 and N-14 were pretty good in neutron diffraction, is that wrong? Would you have to enrich them? I know that is expensive, but my NMR group did it a fair bit. Heck, they even grew deuterium enriched proteins for some experiments, which was rather expensive.

Permalink to Comment


Remember Me?


Email this entry to:

Your email address:

Message (optional):

The Last Post
The GSK Layoffs Continue, By Proxy
The Move is Nigh
Another Alzheimer's IPO
Cutbacks at C&E News
Sanofi Pays to Get Back Into Oncology
An Irresponsible Statement About Curing Cancer
Oliver Sacks on Turning Back to Chemistry