Hi Ian, > All, I know Clemens Vonrhein raised a similar question about the HIS > restraint target values & SUs in the standard library some years ago. > The values currently in use appear to be the Engh & Huber (1991) ones > for the doubly-protonated (+1 charge) imidazole side-chain, so I have 4 > questions: > 1) Shouldn't we be using the E&H (1999) values by now? Obviously, we should. The makers of refinement programs should update their restraint libraries if they haven't done so already.
> 2) Is the doubly-protonated form the most appropriate, given that the > pKa is supposed to be ~ 6 (I'm aware that the physiological > extracellular pH of ~ 7.4 may not be relevant inside a protein)? Notwithstanding the biological question we want to answer, shouldn't we base this on the crystallisation conditions? You could of course argue that pH is a statistical property and does not necessarily apply to a local position in the protein. In that case the neutral form seems to me the most appropriate form. > 3) Do users have the option to use the parameters for the other forms > (assuming of course they can work out for each case which form is > present, and therein of course may lie the rub!)? E&H list values for 3 > forms of HIS: 'HISE' with ND1 unprotonated, NE2 protonated; 'HISD' with > ND1 protonated, NE2 unprotonated; and 'HISH' the doubly protonated form > which appears to be the default. WHAT_CHECK has had a check to find the proper form of HIS for ages. It considers both the geometry (which only works when the restraints ar not too tight) and the hydrogen bond network. > 4) Would a suitably weighted average of all 3 forms be more appropriate, > with suitable inflation of the SUs, given that the form actually present > is likely to be a mixture of tautomeric and/or > resonance-charged/uncharged forms anyway? This can vary locally. On many positions in the protein the protonation state of HIS won't change all that much because it would be energetically unfavourable. Cheers, Robbie
