Dear jlliu liu, Also note that Mg2+ is significantly smaller than water. It fits in places where water cannot go. This doesn't look like a magnesium site on first glance. If you can (privately) give the PDBid of the previous publication, I can have a look in 3D. Cheers, Robbie > Date: Mon, 20 Dec 2010 21:31:58 +0000 > From: [log in to unmask] > Subject: Re: [ccp4bb] Mg2+ or water > To: [log in to unmask] > > Mg2+ is (almost) aways octahedrally coordinated, usually by oxygen atoms, with distances of ~2A. > Phil > > On 20 Dec 2010, at 21:16, jlliu liu wrote: > > > Hi All, > > > > I am refining a structure and encountered a problem of modeling a difference density as water or Mg2+, and would like to hear opinions from the community. It has the following coordinations (attached): the water/Mg2+ forms salt bridge/H-bonding interaction with a carboxylate group from the ligand, it also forms salt bridge/H-bonding interaction with a Glu residue from the protein, it is also within hydrogen bonding distance to the main chain N of another protein residue. In provious publication, it was modelled as a Mg2+ and the author reasoned the dual salt-bridge stabilizes the liganding binding, also the Mg2+ is present in the protein solution for crystallization. For my case, I have no Mg2+ present in the protein buffer, also modelling it with water refines perfectly with no indication of positive difference density even at 2.0 sigma cut off. Should I modelled this density as water or as Mg2+. Your opinions are appreciated. > > > > JL > > > > > > <test.png.odp>