Series 3 compounds targeting PKA

[holeung]

My docking experiments of OSM-S-106 vs all crystal structures of Plasmodium kinases supports PKA (protein kinase A) as the top target (PDB 5kbf).

OSM-S-106 was docked into the ligand binding site with the target side chains free to rotate. The docked structure was then energy minimized. 5kbf_docked_min.pdb.zip

Most of the molecule, especially the sulfonamide end, is tightly buried. I am proposing four compounds to explore this binding mode, incorporating modifications consistent with the modeled pocket.

ELN

[MFernflower]

You are the man!!!!!!!!!! If you have free time could you make a binding pocket diagram of the s4 drugs in the atpase? @holeung

Also it seems like we could replace that sulfur with an oxygen to save costs on reagents?

Might you be able to dock some of the S3 compounds on #546

[holeung]

The model predicts a significant penalty going from thiophene to furan.

[MFernflower]

@holeung I assume this is because of atom size? - A selenium ring system would be fascinating but a definite venture into uncharted territory - I know of less than 5 (actually only like 3) drugs that ever were tested that used selenium heterocycles

[mbhebhe]

Hi @holeung thank you for this. We appreciate your hard work. You are the man!

1) We were wondering if there was raw data relating to this, published somewhere . And how can we access it? ELN? 2) Is this in relation to your comments on #503? Because we did not get the notifications for that or your original post for this. Mat has invited you to the core contributing team, hopefully this will resolve the issue. In relation to the values posted on #503, how much of a better target is PKA-R than other kinases listed? And what is the confidence level? Because the values don’t mean much to non-modelling people. 3) The suggestions you made are new and we have not synthesised these before. Here is a list of the compounds we have made. These will be on the wiki soon.

4) 3/4 of the suggestions can be easily made. Please see the image below for more information. As for OSM-S-106-6, we cannot purchase the thiophene core but would have to look into cyclising from precursors.

5) In relation to the docking, did you only look at OSM-S-106, or were other molecules docked too? It would be interesting to see if predictions hold for other actives and non-active compounds List of actives and non actives and masterlist. Further to this, if you would be willing to, it would be interesting to see how the current series 3 targets fair (Tha's core heteroatom replacements, and Ryan's CLK1 targets)

6) Its great to see that changing from the thiophene to the furan improves predicted potency. We are currently working on the furan but this is on pause (#539). At the moment we are working on the pyrrole and the hypoxanthine analogues of OSM-S-106.

Heteroatom replacements

CLK1 targets

7) As a more open and general question; does anyone know if the target proteins are essential? Have gene-knockout studies been done? Not of vital importance right now, but something to bear in mind as we move forward

Thanks, Tha and @rjmacdonald

[holeung]

1, 2. Some of the docking work was previously described in #503. Not sure about the notifications. The docked pdb coordinates are available in my earlier post at the top of this thread. I will put up a more detailed description of the work on an ELN shortly. What is new here is that I've applied energy minimization after docking and then proposed and analyzed new compounds. Protein kinases can be tricky to model because conformational change and loop flexibility often play key roles in ligand binding. Even in the best situations, modeling is often not reliable and is better suited for hypothesis generation than making strong predictions. In the absence of major conformational changes, my docking experiments suggest PKA is the best target of all the Plasmodium kinases with crystal structures available.

5. I docked all the Series 3 compounds but not the recently proposed compounds. I will describe the work in my ELN post and will include predictions of the new compounds.

6. No, the furan is predicted to be less potent.

7. PKA as a plasmodium target is described in JBC (2016) 291: 25375.

[rjmacdonald]

That would be great, thank you again, if you could link the ELN in this issue when that's done it would be appreciated.

Of course, models only take us so far and with limited experimental data for S3 analogs thus far any predictions will help guide synthesis at this stage and only with experimental results (like the upcoming trials #546, hopefully) will we be able to evaluate any predicted potencies.

Perhaps when the results are in, a summary of the S3 SAR to date, with a combination of the modelling work from yourself and @spadavec, would be ideal?

Interesting paper, I was looking at this yesterday, in that paper they have a crystal structure of cAMP bound to PfPKA-R, would you happen to know if your model is predicting binding in the same region?

It would seem from the work in #503, that the compounds are binding in a similar region to ATP within CLK (comparable to the docking with ANP), it would be interesting to know if this is the case for cAMP and PKA-R.

Edit: Answered my own question after having time to review the paper in more detail, your model is showing binding in the same region observed with cAMP, and similar amino acid residues

[MFernflower]

Based on the post above - I hereby throw in my compound ideas:

@mbhebhe Would you be willing to draft a proposed scheme as you did with Ho's compounds?

[holeung]

For the most recent set of Series 3 compounds proposed in #546, the model predicts pyrrole (best) > thiophene > imidazole, but the differences are modest.

[mbhebhe]

The first two can be done via suzuki coupling, The 3-bromoisoxazole and 2-Bromo-1H-imidazole can be converted to boronic esters and then coupled to the thiophene core. The imizadole should/can be made by substituting a halogen for the imidazole (Cl should be better, but can try Br since we already have that) These schemes are just based on what we already know, I didn't do a scifinder search.

Hopefully these new scaffolds will be potent, they are't fun to make!

[mbhebhe]

Didn't mean to close the issue. Oops

[MFernflower]

I really appreciate the work of @mbhebhe as these compounds do not look like a pleasurable thing to synthesize as opposed to the s4 compounds - @holeung If you have time could you dock my compound ideas to the kinase?

[holeung]

New analog with highest predicted score so far, OSM-S-106-20. I found this just by manually building structures and aligning/scoring in Cresset Forge with my previously described model.

[MFernflower]

@holeung I think you are missing a double bond -- Is 7-(benzyloxy)benzothiophene easy to get/make? could treat with NBS and then couple to the standard bornic acid used and then deprotect + convert alcohol to amine? @mbhebhe / @mattodd

ref for the amine/alcohol conversion: http://www.tandfonline.com/doi/abs/10.1080/00397910008087402?journalCode=lsyc20

[mbhebhe]

Hi, 7-(benzyloxy)benzothiophene is hard to get. Making it also looks difficult, the starting material, benzo[b]thiophen-7-ol (also tricky to make) could also be used in the traditional route for S3 compounds (Chlorination, bromination, amination and then coupling to the benzenesulfonamide) instead of having to make 7-(benzyloxy)benzothiophene. I will look into that later, still trying to finish of my other target molecules! Hope that made sense

[MFernflower]

@mbhebhe Understood clear as a bell! Perhaps you could post us all an update as to where you are in the synthesis of your project molecules?

[drc007]

Since the sulphonamide appears to be interacting with Arg268 perhaps the simplest change to try is to replace the sulphonamide with a carboxylic acid?

A related analogue, https://www.ebi.ac.uk/chembl/compound/inspect/CHEMBL3759356 is known. https://doi.org/10.1016/j.ejmech.2015.11.012 may include useful chemistry.