Project Launch. 2. Which Fenarimol Analogs Should be Made Next?

[mattodd]

In the preprint that was posted at project launch, the resynthesis of some active compounds was described, along with the evaluation of a few new compounds. These are shown again below, and are in the Master List of molecules for the project.

Since the paper was submitted, an Honours student at Sydney Uni who carried out the synthetic work for the paper, Hung Duong (@fantasy121), completed another small set of compounds that have been shipped to Wendy van de Sande's lab for evaluation. We're awaiting potency data, but these compounds are in the Master List and shown below (note a couple of minor issues in the sheet for these compounds that need to be fixed). They interrogate the importance of the central heteroatom for the 3-aromatic ring subseries, and they examine the impact of varying the piperazine moiety.

We have a new student starting at the lab in Sydney today, Matthew O'Dowd (@mattodowd25). So, the question for him and for anyone else interested in becoming involved in the chemistry: What should be make next?

The aim (in response to a query below): to discover some additional potent compounds (alongside some inactives) in order to bolster the SAR.

(note that the original Epichem library #1 may contain some compounds of interest which therefore need not be made).

Chemdraws: Hung Duong Chemdraws.zip

[MFernflower]

Might I throw my hat into the ring??? 2-methylpyrimidine precursor for this really cheap http://www.matrixscientific.com/pyrimidines/067101.html

[drc007]

Before deciding on what to make, you need to decide on what question you are trying to address. Is increasing potency, improving physicochemical properties, pharmacokinetics, off-target activity etc..

[MFernflower]

@drc007 @mattodd The idea behind most of my compounds (barring the section on pyrimidine derivatives and the N-alkyl idea) was this: How far can we tinker with the ethyl piperazine‐1‐carboxylate fragment of BS0495 before we loose potency? Can we replace it with things like morpholine, azeitidne, etc to lower LogP because some of the piperzinyl class compounds are rather hydrophobic for having a hydrophillic ring system within them!

[MFernflower]

@holeung would you be able to dock some of these series 1 compounds to CYP51?

[mattodd]

@MFernflower maybe put some of these knowns into #1? Also, if you're asking someone to do something, like your request above to @holeung, then you should always explain the reasoning and context in detail so that they can judge the merits of the request. In this case: i) Why do you mention this target and ii) why do you think this target is relevant to this disease? @drc007 good question, which I'll edit the original post for. At this stage we don't have a lot of data, so to my mind it'd be helpful to obtain some potent compounds with some inactives just to bolster the SAR.

[MFernflower]

Paper detailing the mechanism of action of Fenarimol as being a CYP51 hitter @mattodd

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4254036/

[mattodd]

Great - that answers i). but ii)? Does M. mycetomatis have CYP51?

[MFernflower]

Fungi by default use ergosterol in there membranes thus they all posses CYP51 in-order to make the ergosterol @mattodd Blocking cyp51 also is the mechanism of action of most azole antifungals https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3810441/

[MFernflower]

Just came up with an interesting idea (it uses a fluconazole like motif instead of the fenarminol one) might be more potent at cyp51? Looks like a pain to try and make however!

[wwjvdsande]

Indeed, Madurella mycetomatis has a CYP51 gene in its genome. prof Sami Khalid build a model of it and is now verifying the model with azole classes of antifungal agents.

[wwjvdsande]

Problem with mycetoma is that the causative agent itself often is susceptible towards the antifungal agents, but when they form a grain in vivo than not. We need something which can penetrate trough the grain

Grain 100x magnificied: grain surrounding by neutrophils

Grain magnificied 400x: hyphae are embedded in brown cement material. This probably inhibits the penetration of many drugs

Problem we don't know what this grain is made of. We do know that chitin, melanin, proteins and betaglucan are part of it.

[mattodd]

Thanks @wwjvdsande (and welcome aboard!). The model being developed by Prof Khalid - are there any online resources for that, by which I mean publications? And you're talking about an in silico model, I assume, against which azoles are docked? Or do you mean an expressed protein?

[holeung]

Can you provide the protein sequence for Cyp51?

[MFernflower]

@wwjvdsande Would it be possible to screen miltefosine against this organism? It's shown ability to kill organisms with a thick chitin wall! https://www.cdc.gov/parasites/balamuthia/treatment-hcp.html

[drc007]

@MFernflower Milefosine was in the original Pathogen box described in the preprint.

[drc007]

The crystal structure human CYP51 complied with econazole an azole fungicide is known. https://www.rcsb.org/structure/3jus as is the TB CYP51 https://www.rcsb.org/structure/1e9x

@wwjvdsande has published Genome Sequence of Madurella mycetomatis https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4882942/

[MFernflower]

I think all docking studies should be done on human CYP51 as human and fungal CYP51 genes are really closely related @holeung

Also @drc007 what was the MIC for miltefosine? I cant find the word "miltefosine" anywhere in the preprint!!!!

As a bonus here's a open frame analogue of fenarminol I came up with:

Where X = C or Si and A = C or N and a possible way to make it: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1055775/

[drc007]

@MFernflower the paper only describes the actives, all compounds in the pathogen box and (in)activities are in the supplementary information (https://www.biorxiv.org/highwire/filestream/79406/field_highwire_adjunct_files/0/258905-1.xlsx) worth downloading and keeping as a reference.

[MFernflower]

Appears miltefosine is inactive against the fungus!! Will keep the spreadsheet @drc007

[wwjvdsande]

The CYP51 protein sequence can be found in the genome. It is in the NCBI database.

[wwjvdsande]

I'm not sure if chitin inhibition will work. When you disrupt the grain, it is usually lipid rich. It feels a bit oily when you work with it. You can disrupt it by crushing it through two glass slides

[MFernflower]

So a lipophillic cyp51 blocker would be required?

[wwjvdsande]

Yes that would be a possibility.

[drc007]

@MFernflower why do you want to dock against human CYP51? Would it not be better to use the fungal target ?

[MFernflower]

@drc007 @holeung Crystallographic structures of human cyp51 are known. We could use yeast cyp51 http://www.rcsb.org/structure/5EAB untill we get a good model of madura cyp51 https://user-images.githubusercontent.com/3164942/36170170-a2972fac-10cc-11e8-8145-a5762e94ef7e.png

@wwjvdsande could you do a blasp to compare mycetoma cyp51 and yeast cyp51? - I cant find the AA sequence for mycetoma cyp51

For what it's worth I think all docking studies should be done on yeast CYP51

[drc007]

@MFernflower Perhaps better to focus effort on getting a homology model of target protein, we can use ligand based-design until it becomes available. Are there any issues with inhibiting CYP51 in humans?

[MFernflower]

Not certain - fluconazole does it and it's pretty safe - only thing is possible liver tox @drc007

[MFernflower]

@mattodd have any new compounds been made since Hung Duong's lab work in 2017?

[holeung]

I believe this is the amino acid sequence of CYP51 from Madurella mycetomatis:

MSATKSIVGEALEYVNIGLSHFLALPLAQRISLIIIIPFIYNIVWQLLYSLRKDRPPLVF YWIPWVGSAVVYGMKPYEFFEECQKKYGDIFSFVLLGRVMTVYLGPKGHEFVFNAKLADV SAEAAYAHLTTPVFGKGVIYDCPNSRLMEQKKFVKGALTKEAFKSYVPLIAEEVYKYFRD SKNFRLNERTTGTIDVMVTQPEMTIFTASRSLLGKEMRAKLDTDFAYLYSDLDKGFTPIN FVFPNLPLEHYRKRDHAQKAISGTYMSLIKERRKNNDIQDRDLIDSLMKNSTYKDGVKMT DQEIANLLIGVLMGGQHTSAATSAWILLHLAERPDVQQELYEEQMRVLDGGKKELTYDLL QEMPLLNQTIKETLRMHHPLHSLFRKVMKDMHVPNTSYVIPAGYHVLVSPGYTHLRDEYF PNAHQFNIHRWNKDSASSYSVGEEVDYGFGAISKGVSSPYLPFGGGRHRCIGEHFAYCQL GVLMSIFIRTLKWHYPEGKTVPPPDFTSMVTLPTGPAKIIWEKRNPEQKI

From http://fungi.ensembl.org/Madurella_mycetomatis/Transcript/Sequence_Protein?db=core;g=MMYC01_202883;r=MMYC01_Sc2059:18815-20540;t=KXX80456

[MFernflower]

Dont know if that's the whole thing but that fragment shares 65% or so similarity to the known yeast cyp51 in the pdb @holeung

[holeung]

Working on the homology model. Will take some time.

In the meantime, the most closely related sequences in the PDB to M. mycetomatis CYP51 come from Aspergillus fumigatus. PDB codes 4uyl, 4uym, 5frb. Pymol session file attached.

overlay.pse.zip

[MFernflower]

Can you post a graphic detaling the compounds you docked? I dont have pymol

On Feb 15, 2018 2:33 PM, "Ho Leung Ng" notifications@github.com wrote:

Working on the homology model. Will take some time.

In the meantime, the most closely related sequences in the PDB to M. mycetomatis CYP51 come from Aspergillus fumigatus. PDB codes 4uyl, 4uym, 5frb. Pymol session file attached.

[image: overlay] https://user-images.githubusercontent.com/10951570/36276820-aaeb50e6-1254-11e8-938f-3b4a11308ede.png

overlay.pse.zip https://github.com/OpenSourceMycetoma/Series-1-Fenarimols/files/1729112/overlay.pse.zip

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/OpenSourceMycetoma/Series-1-Fenarimols/issues/2#issuecomment-366036777, or mute the thread https://github.com/notifications/unsubscribe-auth/ADBLDtfGyCGaEJ5UmPbOQOQoaUtHB5wpks5tVIZ2gaJpZM4R6O0Z .

[holeung]

Compounds are described in the PDB files and are azole drugs. These are from crystal structures, not docking.

[MFernflower]

Understood - can you docking software predict the drugs binding to the iron center of cyp51? If you have time could you try one of my proposed compounds docking to aspergillus cyp51?

On Feb 15, 2018 3:21 PM, "Ho Leung Ng" notifications@github.com wrote:

Compounds are described in the PDB files and are azole drugs. These are from crystal structures, not docking.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/OpenSourceMycetoma/Series-1-Fenarimols/issues/2#issuecomment-366049743, or mute the thread https://github.com/notifications/unsubscribe-auth/ADBLDuj6IrS5UQOVlJQ-cVOGLWUSgZNQks5tVJHSgaJpZM4R6O0Z .

[drc007]

@MFernflower Pymol is a free download https://pymol.org/2/

If we are going to be discussing docking it is pretty much essential.

[MFernflower]

@drc007 pymol complains about not being licensed? will I still be able to view PSE files indefinitely?

[drc007]

Should still be fine

[wwjvdsande]

Sami Khalid also made a model on the cyp51 gene. Might be good to join forces there. I will see if he can join us here.

[wwjvdsande]

He docked several of the adolescenten and fenarimols in his model. Het sent an abstract to the isham conference to be held in Amsterdam this summer

[wwjvdsande]

There will be a session on mycetoma there for those who are interested

[drc007]

This meeting https://www.isham2018.org/en/Home_10_6_12.html ?

[wwjvdsande]

https://www.isham2018.org

[wwjvdsande]

Yes. The globale mycetoma working group and the isham mycetoma working group will be there too. Great if every body can meet and participate

[mattodd]

Hi @wwjvdsande obviously it'd be very good if Prof Khalid could join us here, yes, so that we avoid any duplication of effort. I'd agree with @drc007 that we aim for the most relevant homology model, and in the meantime we look to Apergillus for guidance. I'd imagine we'd only want to look at human in order to identify compounds that do not hit that target?

If there is much more discussion on the development of a homology model could I suggest we move to a parallel Issue, since this one is becoming long. Clearly relevant, but becoming a large discussion of its own.

Pivoting back to the question of which compounds to make, in the absence of a model:

1) @MFernflower for suggestions (which are always welcome) please do check through the Epichem library to see if any related compounds already exist. And also a Pubchem search to see whether any identical/similar molecules are already known. 2) All - would a sheet in the Master List that contains all suggested molecules be helpful? i.e. one that contains strings as well as a URL to where the suggestion was made? 3) @wwjvdsande - the potency data on the compounds Hung submitted in October (shown further up the page) - do we have data on these yet by any chance? The activity (or not) of these molecules will influence which molecules @mattodowd25 goes after in the first instance, and we're at that planning stage.

[MFernflower]

@mattodd Do you think the compounds I proposed are worth making? The pyrimidine type compounds I proposed could be made in a one pot synthesis!

Also regarding the docking studies what I said before about human cyp51 was dead wrong - we are not gonna be able to make a compound that only hits fungal cyp51 - some human cyp51 is going to get bound up no matter what - We need to make compounds that bind stronger to fungal cyp51 than human cyp51 as this prevents liver damage https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5241696/

[mattodowd25]

Does anyone know if this tertiary amine scaffold (see below) has been explored yet? I noticed that this substructure was the most potent in the EpiChem library.

[MFernflower]

@mattodowd25 See: https://user-images.githubusercontent.com/4386101/35832528-bd331fa2-0b21-11e8-8cd5-7a52fc44ac07.jpeg They were not very good

All compounds that were effective had a carbon atom in that position

[mattodowd25]

@mfernflower I have about 500 mg of this compound that I synthesised yesterday. Any suggestions what I should make with it? Btw I am currently making the 1-Boc-piperazine intermediate.

[MFernflower]

Some ideas (I forgot to post smiles strings but will do soon) @mattodd

[drc007]

@mattodd @mattodowd25 @MFernflower With respect to the tertiary amine scaffold

I've had a browse through the Epichem results and can only find 3 analogues, all very weakly active but also not bearing the optimal substitution 1 example shown below. Perhaps I'm missing some results? One advantage of this chemotype is that it eliminates the chiral centre. An alternative might be to incorporate it using the amide linker shown below.

I've done searches of ChEMBL and PubChem and there is nothing similar.

SMILES = FC(F)(F)c1ccc(cc1)N1CCN(CC1)C(=O)N(c1ccc(Cl)cc1)c1cccnc1