Radiation damage can cause specific changes in your structure. Disulphide bonds break up and acidic side chains become decarboxylated. The unit cell volume increases, and the molecule might undergo small rotations and translations. Non-isomorphism is introduced, which can easily cause larged differences in your structure factors than the difference you are looking for when moving from the inflection point to the remote wavelength during MAD data collection.
    There seems to be a maximum amount of photons/cell volume that a crystal can handle before the crystalline diffraction is lost. Henderson's limit, 2x10^7 Gray, can be obtained in as short as 200-400s total exposure time on ID14-4!
    Try to underexpose your sample rather than overexpose it! For MAD, first aim at some phases at low resolution, even with low redundancy, and increase the redundancy and resolution only afterwards. It is better to collect several MAD data sets for the same project, starting with a very conservative data set first. We routinely use SAD as THE alternative for MAD.

 


 

The program RADDOSE
    A program called RADDOSE has been written to calculate the absorbed dose in Gray (J/kg) that a crystal received during an experiment, the theoretical lifetime of a crystal as well as a theoretical maximum heating using the "lumped model" approximation from Kuzay et al. 2001.

    The program will calculate absorption coefficient, relative contribution of different atom-types to the total absorption, crystal density, total absorbed dose, expected life time of the crystal in the X-ray beam as well as theoretical maximum heating of the crystal.

    The absorption coefficient is calculated based on the unit cell contents and the energy of the incoming beam. The unit cell content is calculated based the number of residues, to which non-C/H/N/O atoms could be added. The solvent area is uniformly filled up with water-molecules, although a part can be replaced by absorbing atoms that were used for the crystallisation/heavy-atom soaks. Thus, a 2M AS solution could significantly enhance the absorption power of a crystal, especially if it has a large solvent content. The calculation of the total unit cell content is rather rough, resulting in only an estimate of the absorption coefficient and crystal density. The author is interested receiving ideas on improvements on this.

    The incoming X-ray beam intensity should be given in #photons per second. A uniform beam profile is assumed, and the beam dimensions form an input to the program. The crystal is described by 3 dimensions, 2 being perpendicular to the beam. The calculations do not take into account any rotation of the crystal. For a crystal being much larger than the beam, the calculation would therefore only be valid for the part of the crystal that stays in the beam throughout the data collection. Using calculated absorption, crystal and beam size, photon flux and total exposure time, the program will calculate the number of photons absorbed in the crystal, which is converted to energy deposited in the crystal. This divided by the mass of the exposed part of the crystal gives the absorbed dose in Gray.

    Using Henderson's limit of 2x10^7 Gray, the program will calculate an expected life time of the crystal in the beam.

    Using the exposure time per frame and all parameters given or calculated above, the program will calculate the steady-state temperature increase and the system response time tsys according to the "lumped model" approximation from Kuzay et al. 2001. This model neglects thermal gradients within the crystal, which will not be a valid assumption for large crystals and a non-uniform beam-intensity profile. Nevertheless, it could help to reveal some interesting trends, such as sensitivity to changes in absorption coefficient due to high concentrations of heavy atoms.

    The program is written by and can be obtained from Raimond Ravelli.

 



References

  • The 'fingerprint' that X-rays can leave on structures
    Raimond B.G. Ravelli and Sean M. McSweeney
    Structure, Vol 8, 315-328, 2000

     

  • Specific chemical and structural damage to proteins produced by synchrotron radiation
    Martin Weik, Raimond B. G. Ravelli, Gitay Kryger, Sean McSweeney, Maria L. Raves, Michal Harel, Piet Gros, Israel Silman, Jan Kroon, and Joel L. Sussman
    PNAS Vol. 97, Issue 2, 623-628, 2000

     

  • Structural changes in a cryo-cooled protein crystal owing to radiation damage
    Wilhelm P. Burmeister
    Acta Cryst. D, Vol D56, 2000

     

  • Atomic resolution structures of trypsin provide insight into structural radiation damage
    H.K. Schrøder Leiros, S.M. McSweeney and A.O. Smalas
    Acta Cryst. D, Vol D57, 488-497, 2001

     

  • Primary radiation damage of protein crystals by an intense synchrotron X-ray beam
    T. Teng and K. Moffat
    J. Synchrotron Rad., Vol 7, 313-317, 2000

     

  • Specific protein dynamics near the solvent glass transition assayed by radiation-induced structural changes
    M. Weik, R.B.G. Ravelli, I. Silman, J.L. Sussman, P. Gros and J. Kroon
    Protein Science , Vol 10, 1953-1961, 2001

     

  • X-ray beam/biomaterial thermal interactions in third-generation synchrotron sources
    T. M. Kuzay, M. Kazmierczak and B. J. Hsieh
    Acta Cryst. D, Vol D57, 69-81, 2001