Skip to Content

X-Ray Structures: Handle With Care

X-ray crystallography is wonderful stuff – I think you’ll get chemists to generally agree on that. There’s no other technique that can provide such certainty about the structure of a compound – and for medicinal chemists, it has the invaluable ability to show you a snapshot of your drug candidate bound to its protein target. Of course, not all proteins can be crystallized, and not all of them can be crystallized with drug ligands in them. But an X-ray structure is usually considered the last word, when you can get one – and thanks to automation, computing power, and to brighter X-ray sources, we get more of them than ever.
But there are a surprising number of ways that X-ray data can mislead you. For an excellent treatment of these, complete with plenty of references to the recent literature, see an excellent paper coming out in Drug Discovery Today from researchers at Astra-Zeneca (Andy Davis and Stephen St.-Gallay) and Uppsala University (Gerard Kleywegt). These folks all know their computational and structural biology, and they’re willing to tell you how much they don’t know, either.
For starters, a small (but significant) number of protein structures derived from X-ray data are just plain wrong. Medicinal chemists should always look first at the resolution of an X-ray structure, since the tighter the data, the better the chance there is of things being as they seem. The authors make the important point that there’s some subjective judgment involved on the part of a crystallographer interpreting raw electron-density maps, and the poorer the resolution, the more judgment calls there are to be made:

Nevertheless, most chemists who undertake structure-based design treat a protein crystal structure reverently as if it was determined at very high resolution, regardless of the resolution at which the structure was actually determined (admittedly, crystallographers themselves are not immune to this practice either). Also, the fact that the crystallographer is bound to have made certain assumptions, to have had certain biases and perhaps even to have made mistakes is usually ignored. Assumptions, biases, ambiguities and mistakes may manifest themselves (even in high-resolution structures) at the level of individual atoms, of residues (e.g. sidechain conformations) and beyond.

Then there’s the problem of interpreting how your drug candidate interacts with the protein. The ability to get an X-ray structure doesn’t always correlate well with the binding potency of a given compound, so it’s not like you can necessarily count on a lot of clear signals about why the compound is binding. Hydrogen bonds may be perfectly obvious, or they can be rather hard to interpret. Binding through (or through displacement of) water molecules is extremely important, too, and that can be hard to get a handle on as well.
And not least, there’s the assumption that your structure is going to do you good once you’ve got it nailed down:

It is usually tacitly assumed that the conditions under which the complex was crystallised are relevant, that the observed protein conformation is relevant for interaction with the ligand (i.e. no flexibility in the active-site residues) and that the structure actually contributes insights that will lead to the design of better compounds. While these assumptions seem perfectly reasonable at first sight, they are not all necessarily true. . .

That’s a key point, because that’s the sort of error that can really lead you into trouble. After all, everything looks good, and you can start to think that you really understand the system, that is until none of your wonderful X-ray-based analogs work out they way you thought they would. The authors make the point that when your X-ray data and your structure-activity data seem to diverge, it’s often a sign that you don’t understand some key points about the thermodynamics of binding. (An X-ray is a static picture, and says nothing about what energetic tradeoffs were made along the way). Instead of an irritating disconnect or distraction, it should be looked at as a chance to find out what’s really going on. . .

15 comments on “X-Ray Structures: Handle With Care”

  1. NH_chem says:

    13C and 15N labeling of compounds with high field NMR gives a better look since it is in solution but you all know that………….

  2. ET says:

    On my planet, where water is unknown and we all walk about in our crystalline states, xray data is extremely useful.

  3. xtal_dave says:

    As a protein crystallographer I am always sure to point out the many caveats associated with crystal structures to my colleagues. At resoluition above 2.8A you are pretty much using prior knowledge (bias) to fit a small molecule ligand, however the validity of the data is in the interpretation.
    No-one should claim that crystallography reflects binding affinity. A low potency, highly soluble compound may be bound but the hyper-potent compound that is brick-dust will not.
    Having done a bit of NMR it is fair to say that both techniques have their flaws, but they can be used powerfully ‘in tandem’

  4. Sili says:

    And I never did anything but small-molecule work. And I’m not good enough to get one of the few positions available in that (even if I had a ph.d.).

  5. Sili says:

    And I never did anything but small-molecule work. And I’m not good enough to get one of the few positions available in that (even if I had a ph.d.).

  6. Wavefunction says:

    Kleywegt is one of the wittiest speakers I have ever heard and his chronicling of the problems with crystal structures has been swashbuckling. An older and very readable Angew. Chem. review by him:
    “Application and Limitations of X-ray Crystallographic Data in Structure-Based Ligand and Drug Design”
    Angewandte Chemie International Edition
    Volume 42, Issue 24, Date: June 23, 2003, Pages: 2718-2736
    Andrew M. Davis, Simon J. Teague, Gerard J. Kleywegt”
    DOI: 10.1002/anie.200200539

  7. David Pearlman says:

    As a computational chemist for many years, I’ve been trying to preach the “there is no such thing a simple single rigid structure, xray or not” to chemists for a very long time. The trouble is, it’s a lot easier to live with that picture than it is to deal with the reality. And, to an extent, it’s like a lottery scratcher: Once a chemist has used an x-ray derived snapshot to help direct a change, and that change works, they’re sold on that method and you’ll never convince them that this isn’t going to be the ultimate road to riches.
    There’s no better proof than the fact that at most companies, NMR has been marginalized (or eliminated completely), despite the fact that the information NMR can provide is both more experimentally relevant (solution) and provides important information that one cannot obtain from X-ray about dynamics. It’s not that chemists don’t need that information, but they THINK they don’t need it.

  8. Chrispy says:

    Crystallography has become a crutch. The effort required to crystallize could be spent better making compounds.
    Most of the time, by the time a crystal structure is solved the chemical space is already well trodden by the chemists. At least at any real company with a serious target.
    Unfortunately, the crystallographers have fallen into the same trap as the computer modellers, coming in late with some pretty pictures and claiming some usefulness. Witness the embarrassing articles around HIV protease, where grand claims were made for structure-based design but the leads actually came from screening old renin libraries.
    I do like the pretty pictures, but you chemists need to keep stuff real. The compound is king!

  9. xtal_dave says:

    Crystallogarphy is only a crutch at lead op. The move is on to make structural biology impact at Hit ID and H2L. The current fragment methodologies are the way forward. In my opinion crystallography is pretty much useless at after H2L unless it helps explain some spurious results.
    as I said earlier, to maximise structyural input you have to use an armory of techniques, Xray, NMR, SPR etc. Used in the correct way its incredibly powerful.
    Each technique on its own has massive pitfalls… like everything
    Sometimes I wish I was one of you all-powerful chemists who rule the world, not a lowly ‘service’ to give you pretty pictures for talks.

  10. Morten says:

    I hate when people say that you should look at the resolution to judge whether the structure is good or not. 2.8A today is a lot better today than it was ten years ago. If you want to know anything then you should go look at the electron density maps in the area that you are interested in. B-factors are a poor substitute. And even those are ignored much too often.
    That’s why it’s such a huge problem when big journals (like PNAS) don’t require the deposition of structure factors (last I checked whether they did was Oct. 2007).
    The best crystallography in drug design story is probably the DPP-IV inhibitor from Novartis, vildagliptin. In that case the crystallographers were able to do the structure approx. as fast as the assays were performed and the information could be integrated in to the next round of synthesis. AFAIR it took them around 6 years from target validation to market. Don’t shoot me if I’m wrong. Other than that there’s Astex – fragment-based lead design.

  11. xtal_dave says:

    Structure factor deposition is now mandatory for most big journals (PNAS may be a special case as described by derek below)…
    The biggest crime is ‘over-interpretation’ of a crystal structure… 2.8A is generally good enough to distinguish a gross binding mode ….. not hugely exciting at LO but at hid ID and H2L it provides key information.
    I have also worked on projects where we could turn round a crystal structure in 16hours from receiving the compound to publishing the data. As quick as waiting for a batch of assay results.

  12. RTW says:

    I think everyone puts too much emphasis on a perturbed system (crystal stacked and soaked in many instances). Crystal structures I don’t believe to be truly representative of the bound structure of a small molecule to binding site on a target protein in a biological environment. I have to agree with #7. David Pearlman on this one. There are many many examples where crystal structures have taken medicinal chemists on a very merry chase….
    NMR structure studies in addition to X Ray is probably the way to go. But I am also reminded that it is turning out that the supposed cytoplasm of a cell is not all that much a clasical solution. I think I read an article recently (probably in Science) about the cell being much more like a gel than a solution in the interior. It was a very interesting study. Suggests things may not approach solution phase dynamics either.

  13. bitter pill says:

    I have to say in my experience working with crystallographers was much more productive and less disruptive than working with modelers (with a few exceptions). Modelers always wanted me to make ridculous compounds that no med chemist would ever make. NMR was an order of magnitude less useful than modeling.

  14. Anonymous BMS Researcher says:

    Sometimes the X-ray folks get a protein to crystallize by making a construct with the floppy loops cut off, since those floppy loops get in the way of crystallization.
    But all too frequently it seems the loops are ALSO where the biology we care about seems to happen!

  15. Ted says:

    I will jump on this paper bandwagon..
    Protein crystallography for non-crystallographers, or how to get the best (but not more) from published macromolecular structures

Comments are closed.