metal-ligand distances, and Optimize X-ray/stereochemistry weights
I'm refining a model at 2.9 A, where four chains are in the ASU, and each active site has an iron with a few ligands; the configuration of this active site is the topic here. Currently the iron-ligand distances are not ideal (according to CheckMyMetal and the ideal distances generated by phenix.metal_coordination), but they are within a range observed previously for structures of truncated versions of this enzyme; furthermore, the more the metal-ligand distances change towards ideal distances, the more positive and negative density I observe in the Fo-Fc, so it appears that this deviation from the 'ideal' is supported by the data. However, when I use the 'Optimize X-ray/stereochemistry weights' option during refinement, the metal-ligand distances change (the metal appears to do most of the moving), getting closer to ideal, and this generates more positive and negative density in the Fo-Fc map). However, clearly it's nice that optimizing the X-ray/stereochemistry weights reduces R-free (and keeps the R-f/R-work ratio fine as well). So what is the most appropriate approach here, in order to keep the metal-ligand distances that best support the data, but still reap the benefits of the optimized stereochemistry weights? Some details of the structure and refinement are these: The iron in each of the four chains in the ASU is coordinated with atoms from two histidines, a glutamate, and a water (so four ligands per Fe) and there remains un-modeled density in the iterative build comp OMIT map used for modeling (because resolution is 2.9A and adding more water (which is my best guess, given what truncated structures at higher resolution show coordinating the Fe) does nothing to the Rf). I am working with an .edits file generated by phenix.metal_coordination. After some experimentation, I've left the sigma values (for Fe-O and Fe-N) at default values, and added a restraint for Fe-O (water), using an ideal distance of 2.3 and a sigma value of 0.1. Emily.
Fe-Ligand bond lengths depend significantly on redox and spin state of the iron (which drives the conformational change in hemoglobin). Could it be the state of your iron is different from what metal_coordination assumed in making the .edits file? eab On 05/20/2015 07:03 PM, Emilia C. Arturo (Emily) wrote:
I'm refining a model at 2.9 A, where four chains are in the ASU, and each active site has an iron with a few ligands; the configuration of this active site is the topic here.
Currently the iron-ligand distances are not ideal (according to CheckMyMetal and the ideal distances generated by phenix.metal_coordination), but they are within a range observed previously for structures of truncated versions of this enzyme; furthermore, the more the metal-ligand distances change towards ideal distances, the more positive and negative density I observe in the Fo-Fc, so it appears that this deviation from the 'ideal' is supported by the data. However, when I use the 'Optimize X-ray/stereochemistry weights' option during refinement, the metal-ligand distances change (the metal appears to do most of the moving), getting closer to ideal, and this generates more positive and negative density in the Fo-Fc map). However, clearly it's nice that optimizing the X-ray/stereochemistry weights reduces R-free (and keeps the R-f/R-work ratio fine as well).
So what is the most appropriate approach here, in order to keep the metal-ligand distances that best support the data, but still reap the benefits of the optimized stereochemistry weights?
Some details of the structure and refinement are these: The iron in each of the four chains in the ASU is coordinated with atoms from two histidines, a glutamate, and a water (so four ligands per Fe) and there remains un-modeled density in the iterative build comp OMIT map used for modeling (because resolution is 2.9A and adding more water (which is my best guess, given what truncated structures at higher resolution show coordinating the Fe) does nothing to the Rf). I am working with an .edits file generated by phenix.metal_coordination. After some experimentation, I've left the sigma values (for Fe-O and Fe-N) at default values, and added a restraint for Fe-O (water), using an ideal distance of 2.3 and a sigma value of 0.1.
Emily. _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]
Emily
The ideal value in your .edits file may not be what is the correct value
for your situation. You should feel free to adjust the ideal value as you
see fit to improve the fit to the density. I have also included some other
metal fitting tips from Comput. Cryst. Newsl.
http://www.phenix-online.org/newsletter/CCN_2015_01.pdf#page=6
Tips for coordinated metal refinement
It is not uncommon to observe pronounced residual (difference) map features
around metal ions. These features may originate from a number of possible
reasons, such as: a) underrefined metal parameters, b) inoptimal metal
parametrization,
c) Fourier map artifacts, d) partial or/and shared occupancy, e) incorrect
metal identity.
Provided that the metal identity is correctly assigned, refinement hints
below may be helpful:
1) Ensure charge is in the model input file. In case of PDB file it is
defined in rightmost part of ATOM record, for example:
HETATM 3241 SN SN C 3 5.000 5.000 5.000 0.25
41.55 SN4+
2) Refine occupancy of the metal.
3) If it is a heavy metal (has substantially more electrons than typical
macromolecular atoms C, O and N), refine anisotropic ADP of metal.
4) If it is anomalous scatterer refine f’ and f’’.
5) Run refinement until convergence. Usually it takes more than default 3
macro-cycles, about 5-10 macro-cycles.
If the residual map features are Fourier map artifacts then there isn’t
much one can do about it.
Cheers
Nigel
---
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
On Wed, May 20, 2015 at 4:38 PM, Edward A. Berry
Fe-Ligand bond lengths depend significantly on redox and spin state of the iron (which drives the conformational change in hemoglobin). Could it be the state of your iron is different from what metal_coordination assumed in making the .edits file? eab
On 05/20/2015 07:03 PM, Emilia C. Arturo (Emily) wrote:
I'm refining a model at 2.9 A, where four chains are in the ASU, and each active site has an iron with a few ligands; the configuration of this active site is the topic here.
Currently the iron-ligand distances are not ideal (according to CheckMyMetal and the ideal distances generated by phenix.metal_coordination), but they are within a range observed previously for structures of truncated versions of this enzyme; furthermore, the more the metal-ligand distances change towards ideal distances, the more positive and negative density I observe in the Fo-Fc, so it appears that this deviation from the 'ideal' is supported by the data. However, when I use the 'Optimize X-ray/stereochemistry weights' option during refinement, the metal-ligand distances change (the metal appears to do most of the moving), getting closer to ideal, and this generates more positive and negative density in the Fo-Fc map). However, clearly it's nice that optimizing the X-ray/stereochemistry weights reduces R-free (and keeps the R-f/R-work ratio fine as well).
So what is the most appropriate approach here, in order to keep the metal-ligand distances that best support the data, but still reap the benefits of the optimized stereochemistry weights?
Some details of the structure and refinement are these: The iron in each of the four chains in the ASU is coordinated with atoms from two histidines, a glutamate, and a water (so four ligands per Fe) and there remains un-modeled density in the iterative build comp OMIT map used for modeling (because resolution is 2.9A and adding more water (which is my best guess, given what truncated structures at higher resolution show coordinating the Fe) does nothing to the Rf). I am working with an .edits file generated by phenix.metal_coordination. After some experimentation, I've left the sigma values (for Fe-O and Fe-N) at default values, and added a restraint for Fe-O (water), using an ideal distance of 2.3 and a sigma value of 0.1.
Emily. _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]
_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]
Could it be the state of your iron is different from what metal_coordination assumed in making the .edits file?
Yes, you're absolutely right, and I hadn't found documentation for phenix.metal_coordination to help me gauge whether 'my' irons were like those used to inform how phenix generated the edits file. Do you happen to know on what the tool bases its ideal distances? Do you have any references you can share that discuss how Fe-ligand distances change with charge and spin state, whether its observable in crystal structures or solution? Or, maybe you can suggest search terms to help me out as I am quite naive about bio-inorganic chemistry? Emily.
eab
On 05/20/2015 07:03 PM, Emilia C. Arturo (Emily) wrote:
I'm refining a model at 2.9 A, where four chains are in the ASU, and each active site has an iron with a few ligands; the configuration of this active site is the topic here.
Currently the iron-ligand distances are not ideal (according to CheckMyMetal and the ideal distances generated by phenix.metal_coordination), but they are within a range observed previously for structures of truncated versions of this enzyme; furthermore, the more the metal-ligand distances change towards ideal distances, the more positive and negative density I observe in the Fo-Fc, so it appears that this deviation from the 'ideal' is supported by the data. However, when I use the 'Optimize X-ray/stereochemistry weights' option during refinement, the metal-ligand distances change (the metal appears to do most of the moving), getting closer to ideal, and this generates more positive and negative density in the Fo-Fc map). However, clearly it's nice that optimizing the X-ray/stereochemistry weights reduces R-free (and keeps the R-f/R-work ratio fine as well).
So what is the most appropriate approach here, in order to keep the metal-ligand distances that best support the data, but still reap the benefits of the optimized stereochemistry weights?
Some details of the structure and refinement are these: The iron in each of the four chains in the ASU is coordinated with atoms from two histidines, a glutamate, and a water (so four ligands per Fe) and there remains un-modeled density in the iterative build comp OMIT map used for modeling (because resolution is 2.9A and adding more water (which is my best guess, given what truncated structures at higher resolution show coordinating the Fe) does nothing to the Rf). I am working with an .edits file generated by phenix.metal_coordination. After some experimentation, I've left the sigma values (for Fe-O and Fe-N) at default values, and added a restraint for Fe-O (water), using an ideal distance of 2.3 and a sigma value of 0.1.
Emily. _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected]
On Thu, May 21, 2015 at 11:10 AM, Emilia C. Arturo (Emily) wrote: Yes, you're absolutely right, and I hadn't found documentation for
phenix.metal_coordination to help me gauge whether 'my' irons were like
those used to inform how phenix generated the edits file. Do you happen to
know on what the tool bases its ideal distances? There are taken from a literature survey some time ago. Do you have any references you can share that discuss how Fe-ligand
distances change with charge and spin state, whether its observable in
crystal structures or solution? Or, maybe you can suggest search terms to
help me out as I am quite naive about bio-inorganic chemistry? A more recent survey could possibly be of help to you but you have data
that may be telling you what you need to do. You can also add a slack
option to the edit which makes the potential well flat bottomed. If you
made it say 0.2 then the metal coordinate would be very flexible and you
wouldn't have the nonbonded term pushing them apart.
Cheers
Nigel
---
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
A more recent survey could possibly be of help to you
Yes, I'd been looking at Zheng et al's 2008 paper http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2872550/ reporting the NEIGHBORHOOD database, and the Fe-ligand distances that I observe in my structure are significantly outside of this range (by up to ~35%). I'm going to further into this and maybe pose the question to a broader audience at some point.
but you have data that may be telling you what you need to do. You can also add a slack option to the edit which makes the potential well flat bottomed. If you made it say 0.2 then the metal coordinate would be very flexible and you wouldn't have the nonbonded term pushing them apart.
Ah! I hadn't known about the slack parameter; now I know (this is now bookmarked: http://www.phenix-online.org/download/documentation/cci_apps/structure_refin...). Thanks to you both! Emily.
Cheers
Nigel
--- 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
Well - I'm not an expert on the programs or on coordination chemistry - I just know the problem exists. Perhaps others can help out here. I would try setting a rather large sigma value and perhaps a few pm of slack for these bond lengths in the .edits file. That down-weights or unweights the restraints for these bonds where the ideal values are uncertain. Then re-optimize overall weights and see if you get the same improvement without disturbing the iron coordination. anisotropic ADP refinement sounds like a good idea. His Ne2 - Fe distances in cytochromes tend to be right around 2.0 A. there are structures of cyt c that purport to be in oxidized and reduced states, but this may have been before the reducing effect of radiation was recognized -maybe both are reduced. His - Fe in Oxyhemoglobin is around 2.0 A (1.94 in 1hho), increasing by about 0.1 A in deoxy to as much as 2.15 A. For Fe - carboxylate you could look at feroxidase site in bacterioferritins, don't know if there are any really high resolution structures. eab On 05/21/2015 02:10 PM, Emilia C. Arturo (Emily) wrote:
Could it be the state of your iron is different from what metal_coordination assumed in making the .edits file?
Yes, you're absolutely right, and I hadn't found documentation for phenix.metal_coordination to help me gauge whether 'my' irons were like those used to inform how phenix generated the edits file. Do you happen to know on what the tool bases its ideal distances?
Do you have any references you can share that discuss how Fe-ligand distances change with charge and spin state, whether its observable in crystal structures or solution? Or, maybe you can suggest search terms to help me out as I am quite naive about bio-inorganic chemistry? Emily.
eab
On 05/20/2015 07:03 PM, Emilia C. Arturo (Emily) wrote:
I'm refining a model at 2.9 A, where four chains are in the ASU, and each active site has an iron with a few ligands; the configuration of this active site is the topic here.
Currently the iron-ligand distances are not ideal (according to CheckMyMetal and the ideal distances generated by phenix.metal_coordination), but they are within a range observed previously for structures of truncated versions of this enzyme; furthermore, the more the metal-ligand distances change towards ideal distances, the more positive and negative density I observe in the Fo-Fc, so it appears that this deviation from the 'ideal' is supported by the data. However, when I use the 'Optimize X-ray/stereochemistry weights' option during refinement, the metal-ligand distances change (the metal appears to do most of the moving), getting closer to ideal, and this generates more positive and negative density in the Fo-Fc map). However, clearly it's nice that optimizing the X-ray/stereochemistry weights reduces R-free (and keeps the R-f/R-work ratio fine as well).
So what is the most appropriate approach here, in order to keep the metal-ligand distances that best support the data, but still reap the benefits of the optimized stereochemistry weights?
Some details of the structure and refinement are these: The iron in each of the four chains in the ASU is coordinated with atoms from two histidines, a glutamate, and a water (so four ligands per Fe) and there remains un-modeled density in the iterative build comp OMIT map used for modeling (because resolution is 2.9A and adding more water (which is my best guess, given what truncated structures at higher resolution show coordinating the Fe) does nothing to the Rf). I am working with an .edits file generated by phenix.metal_coordination. After some experimentation, I've left the sigma values (for Fe-O and Fe-N) at default values, and added a restraint for Fe-O (water), using an ideal distance of 2.3 and a sigma value of 0.1.
Emily. _______________________________________________ phenixbb mailing list [email protected] mailto:[email protected] http://phenix-online.org/mailman/listinfo/phenixbb Unsubscribe: [email protected] mailto:[email protected]
participants (3)
-
Edward A. Berry -
Emilia C. Arturo (Emily) -
Nigel Moriarty