Rhys, If crystals grow reproducibly and within a reasonable timeframe, I would always do co-crystallization, particularly if the HA is a good anomalous scatterer. We have had good success with this method, including a recent membrane protein structure. Even if you get crystals that are not isomorphous with the native, SIRAS is easy to do these days.
Also broaden your soaking screens of HAs (if you haven't already), to include mercury compounds, which don't always need a Cys to bind well and love hydrophobic nooks and crannies, and metal clusters (like tantalum bromide). As this is a beta-barrel membrane protein, iodine might also be a way to go. Good luck, Michael **************************************************************** R. Michael Garavito, Ph.D. Professor of Biochemistry & Molecular Biology 603 Wilson Rd., Rm. 513 Michigan State University East Lansing, MI 48824-1319 Office: (517) 355-9724 Lab: (517) 353-9125 FAX: (517) 353-9334 Email: [email protected] **************************************************************** On Jul 27, 2014, at 5:48 PM, RHYS GRINTER <[email protected]> wrote: > Hi All, > > I thought I might put a question to the community, with the hope of getting > some tips of the best way to proceed with my heavy atom phasing problem. > I'm working on solving the structure of an integral beta-barrel membrane > protein of approximately 100 kDa. I've crystallised protein, growing some > very flimsy needle like crystals, and collected datasets to around 3.1 A. > I then produced selenomet derivative protein and repeated crystallisation > trials in the same conditions and also repeated broad screens, however the > derivative protein failed to produce crystals that diffracted beyond 10 A (in > fact it barely crystallises at all). > So I've moved on to heavy atom soaks and have had some success with > tetrachloroplatinate and tetranitroplatinate compounds, in that the crystals > didn't dissolve (as they did with gold and samarium compounds) and diffracted > to some degree. I collected SAD data to around 6.5 A from these crystals and > there seems to be anomolous signal. However, while I get a good CC of 0.4 > from HYSS in phenix with this dataset and the phaser EP FOM is 0.56, the maps > before and after DM are uninterpretable. I'm guessing the quality and > resolution of the data I collected just aren't good enough (the data is > reasonably anisotropic). > I performed the metal soaking, by taking a small amount of the platinate salt > and adding it to the crystallisation drop as the crystals are extremely > fragile and don't stand up well to handling through a soaking or cryo > solution. Leaving the crystals to soak for 48 hours and then, freezing them > directly. The solution is on the border of cryoprotection (the conditions has > PEG2000MME and PVP and the precipitants), but with native crystals this > doesn't seem to be a parameter which affects diffraction. The crystals are > very variable in performance, so while I feel that the heavy atom soaking has > compromised their diffractability to a degree, inherent variation may play a > part. > > What I was wondering is if some one with more experience than me found > themselves in this position, how would they proceed? Questions which spring > to mind are, how much heavy atom compound do people add and how long do they > soak for? Is there anyway I can squeeze something out of the anomalous data I > have, given I have 'reasonable' native data, or will poor quality data give > spuriously positive statistics for heavy atom phasing? And are there any > tricks people have experienced to improve performance of crystals like these > (aside from the usual seeding, additives, different detergents etc which I > have spend a fair bit of time on optimization already). > > Thanks in advance, > > Rhys
