Dear Dmitry, I only work with N-glycosylated proteins and here the people from coot and refmac have done a wonderful job in creating all necessary dictionaries. I would be very surprised if this would not be true for O-glycosidic bonds. Instead of reinventing the wheel myself, I would first try the dictionaries available in coot. So what I would do is:
Go to the Ser - Get monomer MAN (to get a monomer with the most recent atom names) - delete O4, or which oxygen gets replaced by the OG of the Ser - create Link (from the extenstions/modeling menu) and see what happens. Doing a real-space refinement on the sugar only usually produces a fit to the density which is sufficient for refinement with Refmac or Buster. By me Asn-Nag bond is shown as a dotted line, so once the Ser-Man bond is created, it should also be displayed. Alternatively, you could also manually add a LINK record to your pdb file with an editor. Best regards, Herman -----Ursprüngliche Nachricht----- Von: CCP4 bulletin board [mailto:CCP4BB@JISCMAIL.AC.UK] Im Auftrag von Dmitry Rodionov Gesendet: Mittwoch, 20. November 2013 19:49 An: CCP4BB@JISCMAIL.AC.UK Betreff: [ccp4bb] Dealnig with O-linked mannose Good day! I am refining what appears to be O-mannosylated protein structure. In my hands Coot (0.7.1) does not form the SER-MAN bond automatically. I made a SER-MAN.cif with jLigand, which takes care of the glycosidic bond. However, now the peptide bonds are not made to this custom residue (same chain, consecutive numbering). Am I doing something wrong? How can I fix this? Many thanks, Dmitry
