Hi Nigel, One thing I found out is that the slack keyword cannot be added into elbow.edits. After reading elbow.edits, phenix.refine will strip them off and replace them with slack=None. I think this must be a bug. -- Jianghai On Nov 23, 2011, at 11:32 AM, Nigel Moriarty wrote:

Tim

I did a survey of the literature some time ago and used Harding, Acta Cryst. D62 (2006), 678-682 for the data used in phenix.metal_coordination. I recently added the data from Zheng et al. Journal of Inorganic Biochemistry 102 (2008) 1765-1776 and plan to include it soon.

You are quite correct that the coordination number affects the bond length. This is currently ignored in PHENIX but it may well be useful. I'll read Shannon and Prewitt with a view to improving the coordination code.

One thing you can do is add the slack option to the edits giving the bond potential a flat bottom. This allow complete freedom for the metal to use the electron density for fitting within the slack distance of the ideal.

refinement.geometry_restraints.edits { bond { action = *add atom_selection_1 = chain F and resid 1 and name FE1 atom_selection_2 = chain A and resid 69 and name SG distance_ideal = 2.35 sigma = 0.1 slack = 0.1 } }

I'm very interested in discussing your situation and making it better. If you send me your model (off line) we can investigate a better solution.

Nigel

NB. Any files sent to me will be held in strictest confidence.

On Tue, Nov 22, 2011 at 2:48 PM, Timothy Springer wrote:

...

Hi, We are refining structures with Ca binding sites that are replaced by Mn. Using the coordination restraints in Phenix (distance only) obtained with phenix.metal_coordination, we wind up with the ligands coming somewhat too close, so there is positive density on the opposite side of the ligating residue from the metal. We may have a series of problems, including incomplete replacement of Ca by Mn. However, the bigger issue I want to address is coordination number. In the work by Shannon and Prewitt, such as Acta Cryst. B25, 925, metal oxides are examined, and the point is made that the Mn-O distance with a coordination number of 6 is shorter than for 7. Ca commonly has either 6 or 7 ligands, and with 7 ligands (most common) the O cannot squeeze as tightly together and are farther away from Ca. Mn more commonly has 6 ligands, but can have 7. What are the distance restraints in Phenix based on? They seem similar to those in Harding MM, Acta Cryst D62, 678 (2006). This paper lumps measurements for different coordination numbers together for Mn-O and Ca-O distances, but not for Co, Cu, and Zn. It does seem to help to add metal ion restraints in Phenix. Things don't seem to come out as well without them. Would it be helpful to treat 6 and 7-coordinate metals differently? Or should we just loosen the restraints? How does Refmac treat metals? One does not need to add restraints in its case. With my best regards, Tim Timothy A. Springer, Ph.D. Latham Family Professor of Biological Chemistry and Molecular Pharmacology

Harvard Medical School http://idi.harvard.edu/springer Immune Disease Institute [email protected] Program in Cellular and Molecular Medicine, Dept. Medicine Div. Hematology Children's Hospital Boston 3 Blackfan Circle, Rm 3103 phone: 617-713-8200 Boston MA 02115 fax: 617-713-8232

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

-- Nigel W. Moriarty Building 64R0246B, Physical Biosciences Division Lawrence Berkeley National Laboratory Berkeley, CA 94720-8235 Phone : 510-486-5709 Email : [email protected] Fax : 510-486-5909 Web : CCI.LBL.gov _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Tim The glycosidic bonds in PHENIX are taken from the Monomer Lib so I think that should be OK. O1 can be in the axial or equatorial position depending on the ring pucker but the O1 should always be on the opposite side of the ring (if you imagine the ring is flattened) from the C6. It is the chirality around the C1 atom that specifies the alpha/beta linkage. Please send me the PDB file (off-line) so I can double-check that it is correct. Just the three carbohydrate residues would be fine. Cheers Nigel On Tue, Nov 29, 2011 at 3:13 PM, Timothy Springer wrote:

Hi, I have just added alpha1-3 and alpha1-6 fucoses to NAG. The Fucose in the monomer library has O1 in the alpha position, which is axial. However, after refinement, the fucose moves so it is linked beta in the equatorial position.  I believe the alpha and beta linkages for Fucose may be reversed in Phenix.  Please advise. The links are below.

   apply_cif_link {      data_link = "ALPHA1-3"      residue_selection_1 = "chain A and resname NAG and resid 3821"      residue_selection_2 = "chain A and resname FUC and resid 3824"    }    apply_cif_link {      data_link = "ALPHA1-6"      residue_selection_1 = "chain A and resname NAG and resid 3821"      residue_selection_2 = "chain A and resname FUC and resid 3823"    }

With my best regards, Tim Timothy A. Springer, Ph.D. Latham Family Professor of Biological Chemistry and Molecular Pharmacology

Harvard Medical School                   http://idi.harvard.edu/springer Immune Disease Institute                 [email protected] Program in Cellular and Molecular Medicine, Dept. Medicine Div. Hematology                                 Children's Hospital Boston 3 Blackfan Circle, Rm 3103              phone:  617-713-8200 Boston MA 02115                               fax:        617-713-8232

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

-- Nigel W. Moriarty Building 64R0246B, Physical Biosciences Division Lawrence Berkeley National Laboratory Berkeley, CA 94720-8235 Phone : 510-486-5709     Email : [email protected] Fax   : 510-486-5909       Web  : CCI.LBL.gov