Hi Kay,

I'd agree with the approach of manually editing and merging the two pdb files to create a new pdb file that can be used as a template for comparative modelling. You should be able to do this in chimerax or pymol.

You could then do the heteromer modelling with that and probably you would want to include the full sequences (which you could crop) to account for missing residues. I haven't tried this inside chimerax, but it is definitely something that works directly within stand-alone modeller and also within the scipion-chem-modeller plugin for scipion.

Best wishes James

Elaine Meng via ChimeraX-users chimerax-users@cgl.ucsf.edu escribió:

Hi Kay, I somewhat doubt that comparative modeling is the way to solve this problem. Consider:

(1) modeling a multimer requires a multimeric template with same stoichiometry: https://rbvi.ucsf.edu/chimerax/docs/user/tools/modeller.html Only one of your structures has both proteins, so only that structure could be used as a template. Alternatively you could try modeling only the monomer of the 2-domain protein using both structures as templates, but the process may change the relative orientations of the domains that you say were already in the correct relative orientation, as it will not take the second protein structure into account.

(2) modeling a heteromultimer requires a separate sequence window for each different chain: https://rbvi.ucsf.edu/chimerax/docs/user/tools/modeller.html It sounds like the structures you have are already for the same proteins you are interested in (identical sequences, not homologs from a different species or anything like that) so you wouldn't even need an alignment, just the single sequence itself. For heteromultimer modeling, could simply open the structure that contains both proteins, then show both of its sequences (menu: Tools... Sequence... Show Sequence Viewer), which would be in two different sequence windows.... but as mentioned above, heteromultimer modeling may not make sense in your situation. For monomer modeling, you could just open the single sequence of that one protein with two domains (same menu entry, just choose that one chain as the sequence to view), but I'm not sure that makes sense in your situation either.

Instead you might just try manually editing the PDB files (i.e. using a text editor) to substitute the helical loop conformation into the structure that already has both domains. If the two structures aren't already superimposed appropriately to do this, you would first superimpose them (e.g. using Matchmaker, menu: Tools... Structure Analysis... Matchmaker), then save PDB of one of the structures choosing the option to save "relative to" the other structure, then use this newly saved PDB for the text-editing part. You may need to do some renumbering after that, e.g. open your newly edited PDB, then use menu: Tools... Structure Editing... Renumber Residues https://rbvi.ucsf.edu/chimerax/docs/user/tools/renumber.html

Another possibility is to try to use AlphaFold to predict the multimer (menu: Tools... Structure Prediction... AlphaFold, see the prediction option) but this may not (a) be feasible depending on the sizes of these two proteins, or (b) give the expected/desired result. https://rbvi.ucsf.edu/chimerax/docs/user/tools/alphafold.html

I hope this helps, Elaine

---

Elaine C. Meng, Ph.D. UCSF Chimera(X) team Department of Pharmaceutical Chemistry University of California, San Francisco

...

On Apr 4, 2023, at 10:44 AM, Prof. Dr. Kay-Eberhard Gottschalk via ChimeraX-users chimerax-users@cgl.ucsf.edu wrote:

Hi There,

I am interested in a complex formed by a protein with two domains connected by a long loop in complex with a second Protein, that binds this loop, causing it to form a short helix. I have two pdb stuctures: One with the two domains in the correct relative domain orientation, connected by an unstructured loop, and one with the part of the loop forming a helix in complex with a second Protein. I am sure that comparative modelling can do the trick, but somehow I dont understand the workflow. Do I need do load both templates into chimeraX, oben sequence file conatining both Proteins and load that also, if so in which format? I went over the Manual, but somehow it is not clear to me.

Thanks a lot for your help! Best, Kay

---

ChimeraX-users mailing list ChimeraX-users@cgl.ucsf.edu Manage subscription: https://www.rbvi.ucsf.edu/mailman/listinfo/chimerax-users

Hi Kay, Sorry no, the ChimeraX interface to Modeller doesn't include the ability to specify restraints. We aimed to keep it simple and easy to use, so it does not allow for many of the functions that you could perform with Modeller directly.

However, one thing that might help is that in the Advanced options section in the ChimeraX tool, you can specify a "temporary folder location" to put associated files. https://rbvi.ucsf.edu/chimerax/docs/user/tools/modelloops.html

When you run the calculation that has the other aspects that you want (other than the restraints, I mean) and specify a temp folder location, it will generate the corresponding Modeller script file. Then you could take the script file as a starting point to run Modeller directly, e.g. after you edit it to specify using the restraints. Since I don't have experience with running Modeller directly, however, I don't know whether you can specify them directly in the script or whether the script has to refer to a separate file with the restraints.

I hope this helps, Elaine ----- Elaine C. Meng, Ph.D. UCSF Chimera(X) team Department of Pharmaceutical Chemistry University of California, San Francisco

On Apr 5, 2023, at 1:52 AM, Kay-E. Gottschalk via ChimeraX-users chimerax-users@cgl.ucsf.edu wrote:

Makes sense, thanks a lot!

A related question/approach: is it possible to restrain secondary structure during loop modeling? Then I can restrain the respective residues as helix and build the complex using matchmaker. I know the original modeller there was something like:

# Residues 20 through 30 should be an alpha helix: rsr.add(secondary_structure.alpha(self.residue_range(’20:’, ’30:’)))

but I didn't find an implementation in ChimeraX. Thanks again for your quick help!! Best,

Kay

Am 04.04.2023 um 21:36 schrieb Elaine Meng:

...

Hi Kay, I somewhat doubt that comparative modeling is the way to solve this problem. Consider:

(1) modeling a multimer requires a multimeric template with same stoichiometry: https://rbvi.ucsf.edu/chimerax/docs/user/tools/modeller.html Only one of your structures has both proteins, so only that structure could be used as a template. Alternatively you could try modeling only the monomer of the 2-domain protein using both structures as templates, but the process may change the relative orientations of the domains that you say were already in the correct relative orientation, as it will not take the second protein structure into account.

(2) modeling a heteromultimer requires a separate sequence window for each different chain: https://rbvi.ucsf.edu/chimerax/docs/user/tools/modeller.html It sounds like the structures you have are already for the same proteins you are interested in (identical sequences, not homologs from a different species or anything like that) so you wouldn't even need an alignment, just the single sequence itself. For heteromultimer modeling, could simply open the structure that contains both proteins, then show both of its sequences (menu: Tools... Sequence... Show Sequence Viewer), which would be in two different sequence windows.... but as mentioned above, heteromultimer modeling may not make sense in your situation. For monomer modeling, you could just open the single sequence of that one protein with two domains (same menu entry, just choose that one chain as the sequence to view), but I'm not sure that makes sense in your situation either.

Instead you might just try manually editing the PDB files (i.e. using a text editor) to substitute the helical loop conformation into the structure that already has both domains. If the two structures aren't already superimposed appropriately to do this, you would first superimpose them (e.g. using Matchmaker, menu: Tools... Structure Analysis... Matchmaker), then save PDB of one of the structures choosing the option to save "relative to" the other structure, then use this newly saved PDB for the text-editing part. You may need to do some renumbering after that, e.g. open your newly edited PDB, then use menu: Tools... Structure Editing... Renumber Residues https://rbvi.ucsf.edu/chimerax/docs/user/tools/renumber.html

Another possibility is to try to use AlphaFold to predict the multimer (menu: Tools... Structure Prediction... AlphaFold, see the prediction option) but this may not (a) be feasible depending on the sizes of these two proteins, or (b) give the expected/desired result. https://rbvi.ucsf.edu/chimerax/docs/user/tools/alphafold.html

I hope this helps, Elaine

---

Elaine C. Meng, Ph.D. UCSF Chimera(X) team Department of Pharmaceutical Chemistry University of California, San Francisco

...

On Apr 4, 2023, at 10:44 AM, Prof. Dr. Kay-Eberhard Gottschalk via ChimeraX-users chimerax-users@cgl.ucsf.edu wrote:

Hi There, I am interested in a complex formed by a protein with two domains connected by a long loop in complex with a second Protein, that binds this loop, causing it to form a short helix. I have two pdb stuctures: One with the two domains in the correct relative domain orientation, connected by an unstructured loop, and one with the part of the loop forming a helix in complex with a second Protein. I am sure that comparative modelling can do the trick, but somehow I dont understand the workflow. Do I need do load both templates into chimeraX, oben sequence file conatining both Proteins and load that also, if so in which format? I went over the Manual, but somehow it is not clear to me. Thanks a lot for your help! Best, Kay

---

ChimeraX-users mailing list ChimeraX-users@cgl.ucsf.edu Manage subscription: https://www.rbvi.ucsf.edu/mailman/listinfo/chimerax-users

I tested it, it works like a charm! Great advice, thanks a lot!!

Am 05.04.2023 um 17:31 schrieb Elaine Meng:

Hi Kay, Sorry no, the ChimeraX interface to Modeller doesn't include the ability to specify restraints. We aimed to keep it simple and easy to use, so it does not allow for many of the functions that you could perform with Modeller directly.

However, one thing that might help is that in the Advanced options section in the ChimeraX tool, you can specify a "temporary folder location" to put associated files. https://rbvi.ucsf.edu/chimerax/docs/user/tools/modelloops.html

When you run the calculation that has the other aspects that you want (other than the restraints, I mean) and specify a temp folder location, it will generate the corresponding Modeller script file. Then you could take the script file as a starting point to run Modeller directly, e.g. after you edit it to specify using the restraints. Since I don't have experience with running Modeller directly, however, I don't know whether you can specify them directly in the script or whether the script has to refer to a separate file with the restraints.

I hope this helps, Elaine

---

Elaine C. Meng, Ph.D. UCSF Chimera(X) team Department of Pharmaceutical Chemistry University of California, San Francisco

...

On Apr 5, 2023, at 1:52 AM, Kay-E. Gottschalk via ChimeraX-users chimerax-users@cgl.ucsf.edu wrote:

Makes sense, thanks a lot!

A related question/approach: is it possible to restrain secondary structure during loop modeling? Then I can restrain the respective residues as helix and build the complex using matchmaker. I know the original modeller there was something like:

# Residues 20 through 30 should be an alpha helix: rsr.add(secondary_structure.alpha(self.residue_range(’20:’, ’30:’)))

but I didn't find an implementation in ChimeraX. Thanks again for your quick help!! Best,

Kay

Am 04.04.2023 um 21:36 schrieb Elaine Meng:

...

Hi Kay, I somewhat doubt that comparative modeling is the way to solve this problem. Consider:

(1) modeling a multimer requires a multimeric template with same stoichiometry: https://rbvi.ucsf.edu/chimerax/docs/user/tools/modeller.html Only one of your structures has both proteins, so only that structure could be used as a template. Alternatively you could try modeling only the monomer of the 2-domain protein using both structures as templates, but the process may change the relative orientations of the domains that you say were already in the correct relative orientation, as it will not take the second protein structure into account.

(2) modeling a heteromultimer requires a separate sequence window for each different chain: https://rbvi.ucsf.edu/chimerax/docs/user/tools/modeller.html It sounds like the structures you have are already for the same proteins you are interested in (identical sequences, not homologs from a different species or anything like that) so you wouldn't even need an alignment, just the single sequence itself. For heteromultimer modeling, could simply open the structure that contains both proteins, then show both of its sequences (menu: Tools... Sequence... Show Sequence Viewer), which would be in two different sequence windows.... but as mentioned above, heteromultimer modeling may not make sense in your situation. For monomer modeling, you could just open the single sequence of that one protein with two domains (same menu entry, just choose that one chain as the sequence to view), but I'm not sure that makes sense in your situation either.

Instead you might just try manually editing the PDB files (i.e. using a text editor) to substitute the helical loop conformation into the structure that already has both domains. If the two structures aren't already superimposed appropriately to do this, you would first superimpose them (e.g. using Matchmaker, menu: Tools... Structure Analysis... Matchmaker), then save PDB of one of the structures choosing the option to save "relative to" the other structure, then use this newly saved PDB for the text-editing part. You may need to do some renumbering after that, e.g. open your newly edited PDB, then use menu: Tools... Structure Editing... Renumber Residues https://rbvi.ucsf.edu/chimerax/docs/user/tools/renumber.html

Another possibility is to try to use AlphaFold to predict the multimer (menu: Tools... Structure Prediction... AlphaFold, see the prediction option) but this may not (a) be feasible depending on the sizes of these two proteins, or (b) give the expected/desired result. https://rbvi.ucsf.edu/chimerax/docs/user/tools/alphafold.html

I hope this helps, Elaine

---

Elaine C. Meng, Ph.D. UCSF Chimera(X) team Department of Pharmaceutical Chemistry University of California, San Francisco

...

On Apr 4, 2023, at 10:44 AM, Prof. Dr. Kay-Eberhard Gottschalk via ChimeraX-users chimerax-users@cgl.ucsf.edu wrote:

Hi There, I am interested in a complex formed by a protein with two domains connected by a long loop in complex with a second Protein, that binds this loop, causing it to form a short helix. I have two pdb stuctures: One with the two domains in the correct relative domain orientation, connected by an unstructured loop, and one with the part of the loop forming a helix in complex with a second Protein. I am sure that comparative modelling can do the trick, but somehow I dont understand the workflow. Do I need do load both templates into chimeraX, oben sequence file conatining both Proteins and load that also, if so in which format? I went over the Manual, but somehow it is not clear to me. Thanks a lot for your help! Best, Kay

---

ChimeraX-users mailing list ChimeraX-users@cgl.ucsf.edu Manage subscription: https://www.rbvi.ucsf.edu/mailman/listinfo/chimerax-users