[buster-discuss] Problem with hexacoordinated Mg2+

Fri Dec 16 17:34:55 CET 2022

Dear Clemens,

I was a bit sidetracked the last days, but finally today I could test your suggestions. It worked as a charm, both chains. Terrific!

Thank you very much!

Please accept my wishes of a Merry Christmas, and a fantastic 2023 - to you and to all the team.

Best,

---------------------------------------------------
Carlos KIKUTI, PhD
UMR144 - CNRS - Institut Curie
Pavillon Trouillet Rossignol
26 Rue d’Ulm - 75005 Paris, France
+33 6 82 87 62 76
carlos.kikuti at curie.fr<mailto:carlos.kikuti at curie.fr>

Le 12 déc. 2022 à 17:45, Clemens Vonrhein <vonrhein at globalphasing.com<mailto:vonrhein at globalphasing.com>> a écrit :

Dear Carlos,

On Wed, Dec 07, 2022 at 09:02:23PM +0000, Kikuti Carlos wrote:
I’m refining a structure at 2.5 angstroms (2.2 with Staraniso)
containing nucleotide + hexacoordinated Mg2+, but some of the side
chains in the active site seem to be pushed away from the ion,

Yes, if we don't have particular distance/bond restraints between two
atoms, a normal non-bonded interaction kicks in (avoiding atoms
getting too close). One can avoid the latter by using so-called
EXCLUDE cards, e.g.

 EXCLUDE H|604 H|723

(involving all atoms from residues 604 and 723 in the H
chain). However, a proper restraint is the better idea, yes.

and I am not sure the geometry of the phosphate moiety-Mg2+ bonds is
correct . In this structure, Mg2+ is surrounded by 2 water
molecules, 2 phosphate moieties from the nucleotide, two different
residues from the myosin. Bond lengths are between 1.8 and 2.3
angstroms.

Sounds sensible: the angles are probably close to the canonical 90/180
degree as well?

This is a very common scenario in the lab, so your answer
will potentially help improving many other future structures,
especially the ones in lower resolution, and I’ve got a bunch of
those.

What I’ve tried already, using the Nov/22 version of Buster:

LINK records -> well… I thought CCP4/PDB were recommending this as
the way to deal with all kinds of bonds (am I right?), but
Buster/TNT/Gelly prefer to use LINK only for covalent bonds. My LINK
records are filtered out from the beginning. Buster tells me to «
manually describe coordination » , which is the main reason of this
e-mail.

Ah LINK records and the different usages of those ;-)

According to the PDB documentation at

 http://www.wwpdb.org/documentation/file-format-content/format33/sect6.html#LINK

we have

 The LINK records specify connectivity between residues that is not
 implied by the primary structure. Connectivity is expressed in terms
 of the atom names. They also include the distance associated with
 the each linkage following the symmetry operations at the end of
 each record.

 Details

   * The atoms involved in bonds between HET groups or between a HET
     group and standard residue are listed.

So these are intended mostly as descriptions of existing bonds not
implicit in standard residue ordering (furthermore, the distance given
is not a restraint distance but the distance observed on the actual
ATOM/HETATM records).

At some point, REFMAC introduced non-standard LINKR records with a
connectivity type specification at the end (which then pointed into an
actual restraint dictionary). But being non-standard, there was no way
of relying on this really.

(This doesn’t look like the classic formatting/spacing issue as the
error message would be different; and I also tried to invert the
order of the ligands in the records, no success)

Yes, BUSTER doesn't use these LINK records to establish some kind of
connectivity - also because in the old days (still?) there was a lot
of random LINK record generation happening at deposition time without
manual confirmation and therefore very unreliable.

"good”(?) and old CONECT records -> nah… Buster just ignores them,
the output doesn’t have any. I can’t restrain the bond lengths with
this anyways.

Correct.

"-Gelly file.dat" with NOTE_BUSTER_DISTANCE… -> from the log it
seems this won’t work for HETATM, only for ATOM lines. Maybe there
is something I should do in the .seq file?

That's the way to get started, yes. You /could/ add it to the TNT
sequence file, but a cleaner way would be a separate text file with
those so-called Gelly commands ("utility restraints"):

 https://www.globalphasing.com/buster/manual/gelly/manual/gelly5.html

What you want is e.g. (7DY1 as a random example, MG D604)

 NOTE six coordinating atoms around MG    MG D 604
 NOTE BUSTER_DISTANCE =2.12 0.05  D|604:MG D|750:O
 NOTE BUSTER_UTILANGLE 90.00 1.60  D|750:O D|604:MG D|819:O
 NOTE BUSTER_UTILANGLE 90.00 1.60  D|750:O D|604:MG D|295:OG1
 NOTE BUSTER_UTILANGLE 90.00 1.60  D|750:O D|604:MG D|602:O3G
 NOTE BUSTER_UTILANGLE 90.00 1.60  D|750:O D|604:MG D|602:O1B
 NOTE BUSTER_DISTANCE =2.12 0.05  D|604:MG D|754:O
 NOTE BUSTER_UTILANGLE 90.00 1.60  D|754:O D|604:MG D|819:O
 NOTE BUSTER_UTILANGLE 90.00 1.60  D|754:O D|604:MG D|295:OG1
 NOTE BUSTER_UTILANGLE 90.00 1.60  D|754:O D|604:MG D|602:O3G
 NOTE BUSTER_UTILANGLE 90.00 1.60  D|754:O D|604:MG D|602:O1B
 NOTE BUSTER_DISTANCE =2.12 0.05  D|604:MG D|819:O
 NOTE BUSTER_UTILANGLE 90.00 1.60  D|819:O D|604:MG D|295:OG1
 NOTE BUSTER_UTILANGLE 90.00 1.60  D|819:O D|604:MG D|602:O3G
 NOTE BUSTER_DISTANCE =2.12 0.05  D|604:MG D|295:OG1
 NOTE BUSTER_UTILANGLE 90.00 1.60  D|295:OG1 D|604:MG D|602:O1B
 NOTE BUSTER_DISTANCE =2.12 0.05  D|604:MG D|602:O3G
 NOTE BUSTER_UTILANGLE 90.00 1.60  D|602:O3G D|604:MG D|602:O1B
 NOTE BUSTER_DISTANCE =2.12 0.05  D|604:MG D|602:O1B

This defines those 6 distance (here: 2.12A with esd of 0.05) and all
90 degree angles (esd = 1.6).

(I’ve found an old presentation from Clemens that shows a
hexacoordinated Mg2+. That made me salivate, but the protocol is not
described)

We always wanted to push the jiffy/script used at the time into a
normal BUSTER distribution, since the generation of the above
restraint example is tedious for anything larger. Maybe it is time to
get back to this and finally make it available ...

In the meantime and if you have only a few Mg ions to handle: define
the six distances (you might want to play a bit with value/esd) and
all 12 right angles as above. Afterwards, add this text file to your
command-line as

 refine -Gelly Mg-restraints.txt ...

or similar.

Cheers

Clemens, Gerard & Andrew

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.globalphasing.com/pipermail/buster-discuss/attachments/20221216/52eee6d1/attachment.html>