[buster-discuss] Creating link dictionary for a ligand cross-linking two chains

Clemens Vonrhein vonrhein at globalphasing.com
Mon Mar 19 10:24:27 CET 2018


Dear Florian,

On Tue, Mar 13, 2018 at 10:36:17AM +0100, Florian Sauer wrote:
> Dear Buster users,
> 
> I'm refining a 3.5 A structure of a Ubiquitin bound protease. The
> Ubiquitin is covalently bound to the protease active site through a
> Propargylamide (AYE) which replaces Ub Gly76 and connects the Ub
> C-terminus and SG of the active site's Cys.

Ok.

So you have Ub-Gly75 --> AYE -> SG-Cys-Protease? Do you use different
chain identifiers for you ubiquitin part and for the protease part? Or
is this handled via an offset in residue number? And what chain
identifier did you give to the AYE residue?

It is important to know if the two connections are inter- or
intra-chain (where we are talking of a chain as a distinct entity
according to the PDB chain identifier).

> Following the Buster Wiki tutorial (https://www.globalphasing.com/buster/wiki/index.cgi?GradeCovalentTutorial)
> I have created a *.cif file with the grade webserver containing the
> ligand including both covalent bonds Gly->AYE->Cys.

That tutorial is still useable, but with the latest release (20171219)
we introduced a tool that tries to simplify the whole procedure:
aB_covalent_ligand [1]. However, since your situation includes a
compound that is not just connected to one residue, but bridges two
different ones, things might get a bit complicated.

If you have the two correct LINK records in your PDB file (from
ubiquitin Gly 76 to AYE and another one from protease Cys to AYE),
then running aB_covalent_ligand twice should give you the two relevant
link dictionaries and the two required linkage descriptors (so that
the MakeLINK tool knows - either when running manually or as run
during a normal 'refine' job - how to describe connectivity).

> I would now like to create a link dictionary using
> covalent_grade_create_linkdic.pl . How can I specify the two
> covalent bonds (1) Gly C - AYE N  and (2) AYE C2 -> Cys SG
> separately with this script?

We've switched basically to using the aB_covalent_ligand tool now:
could you try and see if that will work for you?

 ---------------------------------------------------------------

Do you think that entry 5ohk would be another example for this? What
seems for us here is the following (we will simplify handling of
multiple linkage types further in one of the next releases):

  # get PDB and MTZ:
    fetch_PDB 5OHK
    cd 5OHK

  # create restraints dictionary for AYE:
    grade_PDB_ligand AYE      

  # create two linkage descriptions:
    aB_covalent_ligand 5ohk.pdb  # select link no 1

    aB_covalent_ligand 5ohk.pdb  # select link no 6

  # combine those:
    cat *AYE*.dat > combined.dat

  # use in BUSTER refinement:
    refine -p 5ohk.pdb \
           -m 5ohk-unique.mtz \
           -l AYE.grade_PDB_ligand.cif \
           MakeLINK_LinkagesFile=`pwd`/combined.dat \
           -l CYS-SG_AYE-C2.dic \
           -l GLY-C_AYE-N1.dic \
           ExcludeResiduesFromFetching="CYA|GLA" \
           -d buster.01 | tee buster.01.lis

You see the 4 additional arguments: 2 additional linkage dictionaries
(*.dic), a file containing the linkage description (combined.dat) and
some (currently required) workaround for the fact that you have two
different covalent linkages in a single structure ... something
the released aB_covalent_ligand tool doesn't yet cover.

> Thanks in advance for your answers!

Cheers

Clemens & Andrew (for BUSTER developers)

[1] https://www.globalphasing.com/buster/ReleaseNotes/ReleaseNotes-BUSTER_snapshot_20171219.txt


More information about the buster-discuss mailing list