> Diffraction measurements at a single temperature do not distinguish > between these two components. In principle a series of diffraction > measurements from the same crystal at different temperatures would > allow partitioning the observed vibrational into the two components. > [Burgi (2000) Rev. Phys. Chem. 51:275] So far as I know this has > been confirmed for small molecule crystals but is too difficult > experimentally to be worth the trouble for protein crystals > (and I've tried :-)
In an early effort to investigate the role of protein dynamics in enzyme funciton, I studied this question as part of my PhD thesis. We found that, for a well-behaved protein that already crystallized well, we were able to nicely distinguish thermal disorder (only present at the higher temp, 300 K) from static disorder (ie differences protein packing in the crystal or more local vibrations that get frozen in). We published it in Rader and Agard, Protein Science, 1997 (https://www.ncbi.nlm.nih.gov/pubmed/9232638). Note that the key for us was to do multiple conformation refinement, so we could actually observe the clustering of states in the disordered regions at low temp. This required high resolution data to have sufficient refinement constraints. Stephen Rader Dept of Chemistry U. of Northern BC On 2017-01-19, at 1:16 PM, Ethan A Merritt <merr...@u.washington.edu> wrote: > On Thursday, 19 January, 2017 20:35:14 you wrote: >> A PhD student asked me what causes diffraction anisotropy. Quoting from the >> Diffraction Anisotropy Server webpage that it is caused by whole-body >> anisotropic vibration of unit cells. He asked whether a colder cyrostream >> could improve anisotropy. My answer would be yes, as colder temperatures >> would lower the vibrations. >> >> My two questions are; (1) am I right? and (2) if so, has it ever been done >> before in practice? > > I do not know if there is past work and literature that answers your > question with specific regard to whole-body anisotropic vibration of > unit cells. > > However with regard to anisotropy in general you must consider two > components. > > (1) Vibration that is still present in the crystal, so that > atoms or larger groups are moving while the diffraction is measured. > The vibrational amplitude will be temperature dependent, but > the anisotropy may remain the same since it depends on the ratio of > vibration amplitude in different directions rather than the > magnitude in any one direction. > > (2) Vibrational displacement of a group in one unit cell relative > to copies of the same group in other unit cells that was "locked in" > when the crystal was frozen. The frozen crystal captures a > sampling of states that were present at room temperature. > The diffraction experiment sees a positional average over space > that is equivalent to a single-copy average over time. > This component is not temperature dependent so long as the > crystal stays frozen. > > Diffraction measurements at a single temperature do not distinguish > between these two components. In principle a series of diffraction > measurements from the same crystal at different temperatures would > allow partitioning the observed vibrational into the two components. > [Burgi (2000) Rev. Phys. Chem. 51:275] So far as I know this has > been confirmed for small molecule crystals but is too difficult > experimentally to be worth the trouble for protein crystals > (and I've tried :-) > > This equivalence of states sampled from a single copy over time > to multiple copies in a frozen crystal is the basis for TLSMD > analysis. In the special case of a single molecule per unit cell > I suppose a one-group TLS treatment reduces to what you originally > asked about - vibration of whole unit cells - but in general > it does not. > > cheers, > > Ethan > > > -- > Ethan A Merritt > Biomolecular Structure Center, K-428 Health Sciences Bldg > MS 357742, University of Washington, Seattle 98195-7742
