Docking Enamine Compounds

[lindapatio]

Hi everyone!

I'm currently trying to come up with new compounds for my master's project with prof @mattodd. I've done some docking experiments for the following four enamine compounds posted on issue 10 by @edwintse and you can check out the screenshot posted below.

Can anyone confirm if the mean minimised affinity values obtained are sensible or makes sense? This would really help me understand what sort of values our current compounds have, especially these enamine compounds since they are considered to be the best ones so far! @kipUNC would you be able to help please?

I also made sure that all poses bound to the expected binding site on nsp13. Here's a snapshot of one of the conformations; pose 118 of Compound 59479544 is shown in pink and silver ligand is 5RM3 (from Hebba's @H-agha fragments). I used the 5RM3 PDB file for my docking experiments by the way!

[lindapatio]

Hi Prof @mattodd! Here are the mean minimised affinity for all the enamine compounds. Most compounds range 4-5. Best two compounds are 59479544 and 138540345

[mattodd]

Well, these are comparable to the numbers you were getting for the UCL synthesised compounds. I think @kipUNC was getting values for the Enamine compounds that were signifcantly better. It would be useful if you could compare results for one or two compounds to see if the poses/docking match? Or even the rank order of the minimised affinities?

[lindapatio]

Hi Prof @mattodd, I managed to compare the interactions of different docking results. I picked two poses (least and best minimised affinity scores) for compound 44 and 45. I think the interactions explain why some poses are binding more tightly than others, but you can check my analysis below to confirm. I've also contacted @kipUNC to see if he can help me with this, just waiting for his reply:D

[lindapatio]

Hi @jemimahaque! I've been docking the five UCL compounds: TK 1-1, JH-A2, JH A1, TK 2-1 and EGT 563 and analysing their affinities. Your compound (JH-A2) had the highest affinity Mean = -4.87 Range = -3.85 to -5.66

After this, I tried to optimise your fragment to see if I can improve the binding affinity. I accidentally put a carbonyl instead of an amide when i was trying to dock your compound but it actually improved the affinity!

Would it be okay if you could check these compounds and see if we'd be able to synthesise them please? Also to check if the chemistry is correct, because my knowledge is quite limited...

*note: amide to carbonyl: position was changed as well so that two carbons are between instead of one

cc: @mattodd

[jemimahaque]

Hi @jemimahaque! I've been docking the five UCL compounds: TK 1-1, JH-A2, JH A1, TK 2-1 and EGT 563 and analysing their affinities. Your compound (JH-A2) had the highest affinity Mean = -4.87 Range = -3.85 to -5.66

After this, I tried to optimise your fragment to see if I can improve the binding affinity. I accidentally put a carbonyl instead of an amide when i was trying to dock your compound but it actually improved the affinity!

Would it be okay if you could check these compounds and see if we'd be able to synthesise them please? Also to check if the chemistry is correct, because my knowledge is quite limited...

*note: amide to carbonyl: position was changed as well so that two carbons are between instead of one

cc: @mattodd

Hi Linda - that looks great, thank you! Those compounds can easily be synthesised I think. In the ketones, is it important that the CF3 group is para to the other substituent? Also, in JH-A2.3-2 - what if the amide substituent is attached to C-2 of the thiophene, or the carbonyl on the amide is directly attached to the thiophene? I think it would also be useful to see the results with a pyrrole instead of a thiophene - as it is more closely related to Heba's original pharmacophore structure?

[lindapatio]

hi @jemimahaque here's what i found:

• ortho position has a better minimised affinity vs para position for JH-A2.3-2 (see first table)
• attaching the carbonyl directly to thiophene while keeping it attached to C3 of thiophene gave better binding (second table)
• swapping thiophene with pyrrole resulted in weaker binding. (so would we keep thiophene?) amide position also played a role, MA was better when amide is at C2. (table 3)
• points 2 and 3 were also for JH-A2.3-2, i chose to focus on this compound since it gave the best minimised affinity! please check table 4 for a summary of the key findings for JH-A2.3-2 and let me know what you think :)

[TomkUCL]

@H-agha @mattodd interesting points from @lindapatio about thiophene over pyrrole, and ketone over amide?

[jemimahaque]

hi @jemimahaque here's what i found:

• ortho position has a better minimised affinity vs para position for JH-A2.3-2 (see first table)
• attaching the carbonyl directly to thiophene while keeping it attached to C3 of thiophene gave better binding (second table)
• swapping thiophene with pyrrole resulted in weaker binding. (so would we keep thiophene?) amide position also played a role, MA was better when amide is at C2. (table 3)
• points 2 and 3 were also for JH-A2.3-2, i chose to focus on this compound since it gave the best minimised affinity! please check table 4 for a summary of the key findings for JH-A2.3-2 and let me know what you think :)

thanks Linda, I will look into how we could synthesise some of those compounds. Does the carbonyl interact on the c-3 of thiophene interact with any amino acid residues?

[TomkUCL]

Nice work @lindapatio . Would it be possible to get 2D interaction maps for those top four scoring compounds?

[lindapatio]

interact @jemimahaque Thank you! Yes it does according to PyMOL, it makes 2 hydrogen bonds with Arg502's guanidino group. The file below contains all data from this docking experiment: Alldata for amide subbing copy.sdf.zip

[lindapatio]

@TomkUCL Thank you! Sure, I'll try to get it done and posted here by the end of the day

[lindapatio]

@TomkUCL Here are the 2D interaction diagrams for the top 4 compounds. I'm not really sure how to show interactions on maestro so i included what I found on PyMOL. Interactions for top 4 JH compounds.pdf

@H-agha Would you be able to confirm the interactions for these compounds please? A few of them are showing up as poor contacts for example pose 80 para to ortho looks like this (see below) but when I try to show the ligand-interaction diagram they don't appear.

SDF files for docking experiments:

1. Original 3 Alldata for ucl compounds.sdf.zip

2. Para to ortho ja-a2.3-2 Alldata for ortho and meta.sdf.zip

3. carbonyl directly attached to thiophene Alldata for amide subbing copy.sdf.zip

4. Thiophene to pyrrole Alldata for thiophene to pyrrole.sdf.zip

[lindapatio]

Hello @drc007, it's Linda from UCL!

I was wondering if you would be able to help me again please? I've found a potential ligand that docks to our target protein quite tightly but I need to produce a 2D interaction diagram. When I get my docking results from the jupyter notebook and open the SDF file on pymol it only shows the ligand pose and not with the protein (as shown in the image below for example), so I have to open a separate file containing the protein to see where it binds. How do I combine these two "objects" on PyMol and view the interactions on Maestro?

I've tried combining the two objects by saving the molecule as one? But I don't think it works as the interactions do not show up on Maestro

Hope you can help!

[drc007]

@lindapatio Open the pdb file you used for the docking experiment first then open the sdf containing the docked structures. Can you then see both protein and small molecule?

[drc007]

@lindapatio Hi, I just wondered if you had been given the link to the online lab manual I created on GitHub? https://github.com/UCL/Open_Docking_Lab_Handbook/wiki

[lindapatio]

@drc007 thanks for the prompt reply! I've been following that lab book for my project, but section 10.4 uses a software that is not available on mac. That is why I tried to use maestro.

I only want to show interactions between one pose and the protein but there are more than 30 states in PyMol, would i need to extract the specific pose I want? And how should I save the PDB file so that it contains the docked ligand with the protein?

[drc007]

@lindapatio In pymol if you have both protein and you have identified the pose you want to evaluate in view from all the docked structures. Select the ligand, and it should a become a selection object. Then click on the Action (A) menu for that selection object, one of the menu items should be extract object. You can then delete the original sdf file that contained all the poses. Then under the File Menu choose Export molecule (Under Selection, choose All). This will give both protein and ligand in the same file. There are various options in the Export molecule dialog box (PDB Options) these may need to be included for Maestro. I don't have a copy of Maestro so I can't try I out. If you want a 2D interaction map have you tried https://www.ebi.ac.uk/thornton-srv/software/LigPlus/. Includes instructions for installation on Mac.

[lindapatio]

Okay I'll try to follow this and see if Maestro will show the interactions. I've also heard of Ligplot but have never used it before so I'll give it a go as well! Thank you very much for your help @drc007