One of the features that Global Phasing's routine runs of deposited PDB structures often pick up is very close contacts between the SG of cysteine residues and the CAB and CAC atoms of the propenyl groups on HEM ligands, not described by LINK cards in the header of the deposited structure. There is generally a CXXC motif in the protein which provides two cysteines to hold a haem in place.
Am I right that, in general, a haem bound to cysteines should be modelled as the molecular entity called by the PDB HEC, with the CAB and CAC atoms essentially tetrahedral, the CAB-CBB and CAC-CBC bond lengths the same as a carbon-carbon single bond, and a link from SG to CAB with angle and bond lengths around the SG as in methionine? There are a number of high-resolution structures deposited recently containing haems near cysteines, and in most of them a re-refinement gives substantial positive difference density about the CM atoms; there is even occasionally a sign of some kind of longer tail coming out from the CMB position. It seems to be purely positive density, rather than the dipole that tends to be diagnostic of anisotropy. My current thought is that even a 1.4A structure of a haem-containing protein (I'm looking at an autoBUSTER re-refinement of the 3fo3 deposition from EMBL Hamburg) may well come from a crystal sufficiently well-ordered that we're seeing hydrogens - one of these structures has at least one isoleucine with green blobs at positions which would be reasonable for every hydrogen on the side-chain - but I would be interested to know other peoples' experience and interpretation. Tom Womack (Global Phasing)
