CCP4 web logo CCP4i: Graphical User Interface
Feasibility of MAD experiment

On the CCP4 Bulletin Board the following was presented:


   Copper is a competitive inhibitor of my enzyme. I have crystals
 which have turned blue on soaking with copper salt. Assuming the worst
 scenario where there is only one copper bound to one enzyme (350 amino acids)
 would it be worth it to do a MAD experiment ? Thanks


There a some rules of thumb to check whether a MAD experiment is feasible or not :

  • The average anomalous and isomorphous differences are given by
    Dano ~ 1/sqrt(2) sqrt(Na/Np) 2f"/<fp>
    Diso ~ 1/sqrt(2) sqrt(Na/Np) Df'/<fp>
    Nanumber of anomalous scatterers (1 in your worst case)
    Npnumber of protein atoms (approx 350x11= 3850)
    f"f" of your anomalous scatterer (best is 4 electrons with Cu)
    Df'difference in f' between 2 lambda (max 8 electrons with Cu)
    <fp>average scattering factor of your protein atoms (6.7)
  • You should have anomalous and isomorphous differences of at least 2.5 - 3 in order to have some hope.
  • In the case described in the question, with 1 Cu for 3850 protein atoms:
    Dano around 1.3%, same for Diso...
    Better hope that you have more than 1 Cu / protein..., in which case:
    2 Cu would give you 1.9 %, and 3 Cu would give 2.3 % etc., so something like 5 - 6 Cu would be needed.

However, these are not hard rules....