Dealing with metal containing ligands - Garib Murshudov

Over the last two decades the quality of internal structures of ligands in the PDB has been improved substantially. The main reason for such improvements is the huge efforts put in by the wwPDB. Apart from the PDB, software tools associated with the refinement programs have also been developed and improved. Developments of such software tools as GRADE/GRADE21, phenix.elbow2 and AceDRG3 has helped in improving ligand chemistry as well as fit of ligands into the density. Despite these
developments metal-containing ligands have largely been ignored. Metal containing ligands pose special problems and they are not easy to deal with. The reasons are: 1) In general, metals can have different coordination geometry depending on their charge and environment they are in. 2) It is not always easy to abstract metal containing ligands from the macromolecule they are in. Classic examples are iron-sulphur clusters. Their full chemical identity could only be perceived if macromolecular atoms they are bonded to are accounted for. In extremely rare cases metal containing ligands can be considered and dealt with outside macromolecules.
Because of such difficulty it is not surprising that many metals containing ligands in the Chemical Components Dictionary (CCD) are with basic chemistry errors - wrong bond orders, missing hydrogens, wrong charges are examples of such problems. These and other problems should be
addressed urgently.
We decided to consider all metal-containing ligands from CCD, correct their chemistry if/when necessary. Currently, there are 857 metal-containing ligands. It seems that many of them can either be removed or replaced with single metal with their immediate environment. Many of these ligands are
present in one or two PDB entries, presumably these were a part of a scientific study or part of crystallographic structure solutions. In any case, our view is that all entries in the CCD as well as the monomer library distributed by the CCP4 should be as accurate as possible: in terms of chemistry and conformation.
To address the problems related with these monomers we decided to extract all metals with their environment from the Crystallography Open Database (COD), classify metal coordination geometries. A new program - metalCoord was written to improve bonds and angles around the metal. The program takes ligand name and a pdb/mmCIF file (either supplied by the user or taken from the PDB), takes the environment together with this metal, looks at the coordination classes, extracts bond lengths and angles
using coordination classes. Extracted geometric information is then written to the output cif file thus updating ligand parameters. It also writes a json file with all bonds and angles between the metal and atoms from this ligand as well as atoms from the macromolecule. The newly derived ligand description is used to replace existing ligands. The rest of the information is used as external distance/angle restraints in refinement programs.
In this talk application of the developed tools for all PDB entries containing metal containing ligands will be described. Special attention will be paid to heme like ligands and iron-sulphur clusters, as they form substantial portion of metal containing ligands.
1. Moriartyç N.W. et al (2009) Acta Cryst D65, 1074-80 (2009).
2. grade.globalph...
3. Long F et al. Acta Cryst (2017) D73 112-122