Dear James and all, just to throw into the pot an idea I never came across anywhere, and I could never test because I never had such data/problem:
1. many NCS copies in the asymmetric unit; 2. either isomorphous differences and/or anomalous differences from heavy atoms 3. more than one heavy atom bound per molecule; 4. clear self rotation function indicating the directions of the NCS axes and the point group of the NCS I always thought in this case one should try and average under the NCS point symmetry the isomorphous and difference Pattersons around the origin to boost the intramolecular heavy-atom Patterson vectors. Has anyone ever had such data and tried that strategy? Is it implemented in any of the currently available difference Patterson-solvers/heavy atom finders? Ciao! Pietro Sent from my Desktop Dr. Pietro Roversi Oxford University Biochemistry Department - Glycobiology Division South Parks Road Oxford OX1 3QU England - UK Tel. 0044 1865 275339 ________________________________________ From: CCP4 bulletin board [[email protected]] on behalf of James Holton [[email protected]] Sent: 22 January 2014 02:24 To: [email protected] Subject: Re: [ccp4bb] Phasing with Many Monomers/AU The problem of many monomers in the "ASU" is not restricted to macromolecules. An interesting recent small molecule example is the structure of L-tryptophan (http://dx.doi.org/10.1107/S0108768112033484) which, amazingly, was not published until 2012. This is perhaps in part due to difficulty in accepting 16 monomers in the ASU (they call this Z=16), which was unprecedented. As a beamline scientist, I have seen "high Z" macromolecular crystals on many occasions, but they almost never get solved. Yes, they don't diffract well, but neither does anything else in the early stages of a project. The reason for not solving them seems more psychological than anything else. The prospect of amplifying the building and refinement headache by a factor of "Z" when Z > 10 is perhaps too much for an early term graduate student to bear. On the other hand, automated building and refinement has come a long way, and 24-fold NCS is a great restraint if you can get it! In fact, for virus structures, it has been shown that you can phase the structure starting with nothing but a crude spherical envelope and lots of density modification (http://dx.doi.org/10.1107/S0108767391013211). but your initial problems are going to be phasing. Ideally what you'd want is a way of folding back NCS information into the heavy atom finding and phase refinement process, but I know of no programs that actually do that. In fact, both molecular replacement and heavy-atom finding are hindered by this "pseodo-translation" rather than helped by it. Personally, I blame the fact that methods developers seldom get their hands on "interesting" datasets like yours. And if you look in the PDB there are very few examples of "high Z'" structures. Ahem. Best advice I can give is to try the "usual" approach, but look very seriously for NCS as early as you can. Then apply building/phasing packages like shelx{cde}, phenix.autobuild, or the newly-released Crank2. -James Holton MAD Scientist On 1/18/2014 11:18 PM, Felix Frolow wrote: > Francis, It can happened > We have (not yet published) P1 with 24 molecules. When we cut His-tag we get > P1 with 32 molecules. > In our case we believe it is dictated by very strong interaction between two > monomers, and strong interaction between dimers with build a flattish > tetramer. Probably such formations > is more difficult to oaks than globular oligomers. > In this moment I do not recall what we see in solution, I have to check. > Relating to structure solution, P1 is very convenient space group. > I would go for determination this structure by SAD (SHELXC/D/E pipe, PHENIX > or SHARP). For the native - molecular replacement. > In our time after tremendous developments in Refmac and Phenix and > development o DM refinement is 3-3.4 Ang. Is not very difficult. > I would use in addition to NCS restraints in refinement also multi crystal > averaging. Roumors say it is the most strongest phasing method (attributed to > Eleanor Dodson, myself never used it). > > > FF > > Dr Felix Frolow > Professor of Structural Biology and Biotechnology, Department of Molecular > Microbiology and Biotechnology > Tel Aviv University 69978, Israel > > Acta Crystallographica F, co-editor > > e-mail: [email protected] > Tel: ++972-3640-8723 > Fax: ++972-3640-9407 > Cellular: 0547 459 608 > > On Jan 19, 2014, at 08:48 , Francis Reyes <[email protected]> wrote: > >> You sure about this space group? 24 monomers in P1 is unusual (at least to >> me) >> >> F >> >>> On Jan 18, 2014, at 9:14 AM, Chris Fage <[email protected]> wrote: >>> >>> Hello Everyone, >>> >>> I am currently trying to phase a structure with an asymmetric unit >>> predicted to contain 20-24 monomers (space group P1). The native crystals, >>> while beautiful in appearance (see attached), only diffract to ~3.4-3.0 >>> angstroms at best, and SeMet-derived crystals grow with poor morphology >>> (small needles). Also, based a fluorescence scan, I know that mercury does >>> not bind appreciably. Other than screening for a new space group, what >>> options might I have for phasing this many monomers at lower resolution? Is >>> there any real chance of solving the structure in this space group? >>> >>> Thank you in advance for any suggestions! >>> >>> Regards, >>> Chris >>> <Crystals.jpg>
