[buster-discuss] Grade2 restraints for cis-trans stereoisomers

[Oliver Smart]

Dear Wei-Chun,

Thank you for your report. Understanding Grade2 restraints for compounds with cis-trans stereoisomers is a complicated matter, so we have just written a new section in the Grade2 documentation. The section is entitled "How Grade2 handles cis-trans stereoisomers". Please see:

https://gphl.gitlab.io/grade2_docs/planes.html#cis-trans

For your ligand, the chem_comp_tor.source for the plane:

Mogul_plane_rms_out_of_plane_torsion3.7_degs_282_hits_trans_98.6%_within_10degs

Indicates that 282 CSD hits found for the torsion had a root-mean-squared out-of-plane torsion of 3.7º. This low value will result in a tight plane. 98.6% of the hits are with ±10º of the trans conformation. So, there are either no or very few examples of a CSD hit to the cis stereoisomer. So, data mining the CSD for cis-specific restraints is likely to be impossible in your case.

For your ligand, Grade2 restraints should be able well to represent both the E- and Z- stereoisomers. You may need to flip the stereoisomer when fitting. The documentation includes a subsection, "Flipping a stereoisomer in Coot", with a YouTube video showing exactly how to do this.

It is unlikely that you need a restraint file restricted to one stereoisomer. Still, in the unlikely circumstance that you do, the documentation now also describes how to do this.

Thanks for bringing the problem with butene to our attention. Unfortunately, there are few (in fact, three) CSD structures containing butene. Consequently, Grade2 cannot use Mogul information to set the plane on the double bond. The fallback source of information for restraints, if Mogul does not have enough information, is currently the MMFF94s forcefield (as implemented in RDKit). However, this tends incorrectly to apply planar restraints between aromatic rings. It was decided recently to modify Grade2 not to use MMFF94s as a source of planar restraints unless a hydrogen atom is involved (https://gphl.gitlab.io/grade2_docs/changelog.html#bug-fix-636). In most cases, this is an improvement, but for butene an obvious plane is missed. We are currently thinking about what best to do about this.

Please let us know if you have any further questions or if anything is unclear.

Best regards,

Oliver (for Buster Developers).



> On 30 Jul 2024, at 16:23, Wei-Chun Kao via buster-discuss <buster-discuss at globalphasing.com> wrote:
> 
> Dear BUSTER developers,
> 
> I am using grade2 version 1.6.0 with CCDC 2024.1. I am trying to generate restraints from a compound which undergoes E-Z conversion, therefore I created the two stereoisomers respectively in mol2 format and gave them to grade2. 
> 
> However, the output restraints as well as coordinates of the Z- and E-compound are the E-form, _chem_comp_plane_atom.plane_id and _chem_comp_tor.id <http://chem_comp_tor.id/> are trans-32 (I suppose that the one in the Z-form should be cis-32 ? as two substituents of the C=C of interest are hydrogen atoms), _chem_comp_plane_atom.source as well as _chem_comp_tor.source of four carbon atoms around the C=C of interest show 
> 
> Mogul_plane_rms_out_of_plane_torsion3.7_degs_282_hits_trans_98.6%_within_10degs
> 
> I wondered if the input coordinates in mol2 contain mistakes, hence I tried to use SMILES and explicitly assigned the stereochemistry (X/C=C/Y and X/C=C\Y). Interestingly, grade2 can write the Z-form coordinates correctly, but the restraints still contain the aforementioned out_of_plane information and _chem_comp_plane_atom.plane_id and _chem_comp_tor.id <http://chem_comp_tor.id/> are trans-32.
> 
> Another exercise I made was with cis-but-2-ene (C/C=C\C) and trans-but-2-ene (C/C=C/C). Grade2 wrote the coordinates of the respective stereoisomers correctly, and the _chem_comp_tor.source  in the restraints are both “unrestrained_default”. 
> 
> Could you please advise how to interpret these results? Thanks.
> 
> Best regards,
> 
> Wei-Chun Kao
> 
> —
> 
> Dr. rer. nat. Wei-Chun Kao
> wei-chun.kao at biochemie.uni-freiburg.de
> TEL: +49-761-203-5277 <tel:%2B49-761-203-5277>
> orcid.org/0000-0001-8687-6334						
> 
> Institute for Biochemistry and Molecular Biology
> AG Carola Hunte
> Albert-Ludwigs-Universitaet Freiburg
> Stefan-Meier-Str. 17
> D-79104 Freiburg im Breisgau
> Germany
> 
> _______________________________________________
> buster-discuss mailing list
> buster-discuss at globalphasing.com
> https://www.globalphasing.com/mailman/listinfo/buster-discuss

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.globalphasing.com/pipermail/buster-discuss/attachments/20240808/a99e0b91/attachment.html>