For this particular project, we mounted two native crystals from the crystal drop before adding the Ta6Br12(2+). The Ta6Br12(2+) was added as powder and left o/n. Since various amounts of Ta6Br12(2+) was added this way we never knew what the concentration was, many crystals were ruined his way, I suspect that the final concentration of Ta6Br12(2+) in the drops with useful crystals had a [Ta6Br12(2+)] below 1mM. The structure was initially solved by SIRAS using a native and derivative dataset collected on a crystal originating from the same drop. This derivative was collected at the peak. Most of the phasing power came from the isomorphous signal. However later as Poul mentioned, the data we got the best result from, was a dataset collected at an energy above the absorption edge of Ta and used both the anomalous and the isomorphous signal to get the best phase information. Preben
On 04/10/2010, at 21.48, Jacob Keller wrote: > ----- Original Message ----- From: "Poul Nissen" <[email protected]> > To: <[email protected]> > Sent: Monday, October 04, 2010 2:21 PM > Subject: Re: [ccp4bb] Radiation damage with crystals containing metal centers > (TaBr people chime in?) > > > [1] the signal from Ta6Br12 is enormous and one will typically focus on low > resolution (below 7 Å) so radiation sensitivity can be handled by a fairly > low dose data collection > We collected several data sets with Ta6Br12(2+) on the Na+,K+-ATPase (Morth > JP et al. 2007) and found that although we got the strongest anom. diff. > Fourier peaks from a data set collected on the Ta peak, we got far better SAD > phases from a data set collected on the high-energy remote wavelength. This I > think is also often observed for SeMet. > > > > > > Interesting phenomenon--has it been documented, I wonder? I wonder which > datasets were collected first? If the peak was collected first, as usual I > think, and one assumes an exponential decay of the resonant signal as a > function of radiation dose, it makes sense that the resonant signal would be > more constant after the first data set, where the decay curve would have > flattened out a bit. This would also be true for two consecutive data sets > collected at high energy. Also, I think the decay function itself is steeper > at the peak wavelength, leading to a less-internally-consistent data set at > the peak. Does this argument hold water? > > Jacob Keller J. Preben Morth, Ph.D Group Leader Membrane Transport Group Nordic EMBL Partnership Centre for Molecular Medicine Norway (NCMM) University of Oslo P.O.Box 1137 Blindern 0318 Oslo, Norway Email: [email protected], [email protected] Tel: +47 2284 0600 http://www.ncmm.uio.no/research/ncmm-embl-group-leaders/
