MRSA/VRE Assay and Potency Results Oct 2020

[edwintse]

The following 38 compounds have been screened against MRSA by Paul Stapleton here at UCL. In addition to this, all compounds will be screened against VRE (vancomycin-resistant Enterococcus faecalis) to see if our compounds are active against other gram +ve bacteria. The MIC values (ug/mL) are shown below for each bacteria. Colour coding: 2 and 4 (green); 8 and 16 (orange); 32 and >32 (red).

Top compounds synthesised by @edwintse and @danaklug including benzofurans and +ve and -ve controls (822 and 820, respectively). Bottom compounds donated from DNDi (Thanks @bendndi!)

Quick observations

• There is good correlation of activity between the gram +ve bacteria (MRSA and VRE)
• The positive control (822) had slightly decreased activity again (compared to the original of 2 ug/mL and the previous run of 4 ug/mL)
• The 5,6-core is tolerated better with the benzothiophene (871) => @danaklug currently pursuing the benzofuran derivative with this core
• Remake of the benzofuran (821) matches the original results => the 2-pyridyl is important as the phenyl, 3- and 4-pyridyl (869, 861 and 862 respectively) are inactive => need to evaluate this set for tox (Human Cytotoxicity Assay #3)
• The p-tolyl compound (865) is also active
• All other UCL compounds are inactive
• The majority of the DNDi compounds (unsaturated 5,6-core with amine spacer) are inactive with a few exceptions that have weak activity => but tricky to make meaningful connections between these weak actives

---

OSA_000XXX	SMILES
837	CS(C(C=C1)=CC=C1C2=C(N=C3N2CCC3)C4=NC=CC=C4)(=O)=O
838	CS(C(C=C1)=CC=C1C2=C(N=C3N2CCC3)C4=NC=CC=C4)=O
839	COC1=CC2=NC(C3=NC=CC=C3)=C(NC4=CC=C(F)C=N4)N2C=C1
840	CN(C)C(C=C1)=CC=C1C2=C(NC3CCCC3)N4C(C=C(C)C=C4)=N2
841	CC1=CN2C(C=C1)=NC(C3=NC=CC=C3)=C2NC4=CC=C(OCCO5)C5=C4
842	CC1=CN2C(C=C1)=NC(C3=NC=CC=C3)=C2NCC4=CC=CC=C4
843	COC1=CC2=NC(C3=CSC=N3)=C(NC4=CC=CC(F)=C4)N2C=C1
844	FC1=CC(NC2=C(N=C3N2C=C(CN(C)C)C=C3)C4=NC=CC=C4)=CC=C1
845	C1(C2=C(NC3CCCC3)N4C(C=CC=C4)=N2)=NC=CS1
846	CN(C)C(C=C1)=CC=C1C2=C(NC3=CC=C(F)C=C3)N4C(C=C(C)C=C4)=N2
847	FC(C=C1)=CC=C1NC2=C(N=C3N2C=CC=C3)C4=NC=CN4C
848	COC1=CC2=NC(C3=CN(C)C=N3)=C(NC4=CC=CC(F)=C4)N2C=C1
849	CC1=CN2C(C=C1)=NC(C3=CC(OC)=C(OC)C=C3)=C2NCC4=CC=CO4
850	C1(C2=C(N3CCOCC3)N4C(C=CC=C4)=N2)=NC=CC=C1
851	COC1=CC2=NC(C3=NC=CC=C3)=C(NC4=CC=CC=C4F)N2C=C1
852	FC1=CC(NC2=C(N=C3N2C=C(COC)C=C3)C4=NC=CC=C4)=CC=C1
853	COC1=CC2=NC(C3=NC=CC=C3)=C(NCC4=CN(C)C=N4)N2C=C1
854	COC1=CC2=NC(C3=NC=CC=C3)=C(NC4=CC=NC(C)=C4)N2C=C1
855	COC1=CC2=NC(C3=CN(C)N=C3)=C(NC4=CC=CC=C4F)N2C=C1
856	COC1=CC2=NC(C3=NC=CC=C3)=C([H])N2C=C1
857	COC1=CC2=NC(C3=NC=CC=C3)=C(C4=CC=CC(Cl)=N4)N2C=C1
858	COC1=CC2=NC(C3=NC=CC=C3)=C(C4=CC=C(F)C=C4)N2C=C1
859	FC(C=C1)=CC=C1C2=C(N=C3N2C=CC=C3)C4=NC=CC=C4
860	COC1=CC2=NC(C3=NC=CC=C3)=C(C4=CC=CC=C4)N2C=C1
821	C1(C2=C(C3=CC=C(OC=C4)C4=C3)N5C(CCC5)=N2)=NC=CC=C1
861	C1(C2=C(C3=CC=C(OC=C4)C4=C3)N5C(CCC5)=N2)=CN=CC=C1
862	C1(C2=C(C3=CC=C(OC=C4)C4=C3)N5C(CCC5)=N2)=CC=NC=C1
864	N#CC1=CC=CC(C2=C(N=C3N2CCC3)C4=NC=CC=C4)=C1
865	CC(C=C1)=CC=C1C2=C(N=C3N2CCC3)C4=NC=CC=C4
866	COC1=CC2=NC(C3=CNC=N3)=C(NC4=CC=C(OCCO5)C5=C4)N2C=C1
867	COC1=CN2C(C=C1)=NC(C3=CNC=N3)=C2NC4=CC=C(OCCO5)C5=C4
868	FC(OC(C=C1)=CC=C1C2=C(N=C3N2CCC3)C4=NC=CC=C4)(F)F
869	C1(C2=C(C3=CC=C(OC=C4)C4=C3)N5C(CCC5)=N2)=CC=CC=C1
870	C1(C2=C(C3=CC=CC=C3)N4C(CCC4)=N2)=NC=CC=C1
871	C1(C2=C(C3=CC=C(SC=C4)C4=C3)N5C(C=CC=C5)=N2)=NC=CC=C1
872	FC1=CC(C2=C(N=C3N2CCC3)C4=NC=CC=C4)=CC=C1
822	C1(C2=C(C3=CC=C(SC=C4)C4=C3)N5C(CCC5)=N2)=NC=CC=C1
820	COC(C(OC)=C1)=CC=C1C2=C(N=C3N2CCC3)C4=NC=CC=C4

[edwintse]

UPDATE: The original post has now been updated to include the potency results and some quick observations. A table of SMILES strings have also been added.

[MFernflower]

@edwintse @danaklug @mattodd @drc007

A few ideas I thought up after seeing the new data:

[drc007]

The fact that the 2-pyridyl is essential underlines the importance of getting the toxicity data on the same compounds.

[danaklug]

A few more observations about the DNDi set:

• Methoxy- and methyl-substituted imidazopyridine cores are tolerated (858, 860, 839, 851, 855, and 841) and may improve MRSA activity (859 vs 858).

• Amine insertions are tolerated in many cases (839, 851, 855, 841).

• Replacement of 2-pyridyl with other heterocycles is not tolerated (no exact matched pairs, but none of these compounds show any activity).

• Compounds without aromatic substituents at the northwest position are not active, and substituents on phenyl rings at this position seem to be most beneficial at the para-position.

[MFernflower]

@danaklug @edwintse Might not be a bad idea to make a version of 839 where the fluoropyridine motif is replaced with the para-methylphenyl group of 865

https://molview.org/?smiles=COC1C=CN2C(=NC(=C2N([H])C2C=CC(C)=CC=2)C2C=CC=CN=2)C=1
https://molview.org/?q=COC1=CC2=NC(=C(N(C)C3=CC=C(C)C=C3)N2C=C1)C1=NC=CC=C1

[mattodd]

Activity vs VRE (actually Enterococcus faecalis, vanB-positive; Paul says he will try and obtain a vancomycin-resistant Enterococcus faecium strain for future reference) is roughly tracking with MRSA indicating the activity is likely not something that is MRSA-specific.

[MFernflower]

@mattodd I almost wonder if instead of sourcing another Enterococcus strain it would be better if we tried another pathogen on the list - Streptococcus pneumoniae is a promising one

[mattodd]

Quick thoughts on the data, since I'll be mobile on Friday and may not be able to dial in well (@danaklug has agreed to host the meeting). @bendndi @drc007 @loriferrins it'd be awesome if you were also able to join and give your gut feelings on the above data.

1. Compound 865 (the tolyl) is interesting, and this does indeed mean that exploration of that ring is a good idea (and synthetically simple) as @danaklug and @MFernflower have already suggested. We've retained activity and removed what was a significant metabolic liability (though we might have created a new one).

2. I'm also interested by 871, that the DNDi-style aromatic core works. The chimera with 865 would be interesting.

3. We seem to need the 2-pyridyl. We need to await the tox data (compounds being sent in Human Cytotoxicity Assay #3 and data next week), but the fact that so many compounds contain the 2-pyridyl and yet are inactive does suggest we're not talking about a generic toxicity effect arising from that motif. There could still be a common target between mammalian and gram+ve, which would be problematic, but let's see.

[drc007]

Can everyone include calculated LogP with their compound suggestions.

[drc007]

Here is a sdf file with all 39 compounds together with a range of calculated physicochemical properties, calculated using ChemAxon tools.

This can be viewed using the open source DataWarrior (http://www.openmolecules.org/datawarrior/download.html).

29Oct2020alldata.sdf.zip

[MFernflower]

@mattodd I can attend tommrow's meeting

[drc007]

A few thoughts. Why is the pyridyl nitrogen important? One possibility is that it takes part in a bidentate interaction as shown below in orange (there is no direct evidence for this), the low energy minima is shown in blue where the pyridyl nitrogen points away from the imidazole nitrogen. It might be interesting to do some NMR experiments to see if addition of zinc changes the preferred conformation?

Along similar lines perhaps look at eliminating one of the imidazo nitrogens. An alternative might be to explore alternatives to the pyridine that maintain a nitrogen that could form a bidentate interaction.

Whilst the benzofuran is the most active to date, the benzodioxan of OSA_000841 might be an interesting lower logP alternative (however the dimethoxy was inactive OSA_000820), perhaps the nitrogen linker is important.

[MFernflower]

@drc007 What about lopping off the core nitrogen that does not participate in the metal binding?

[mattodd]

@drc007 Yes, @bendndi suggested playing with that pyridyl (yet keeping a coordinating atom, like your thiazole) back here. Might make sense. Removing the imidazo N makes sense from a logical/academic standpoint to assess the metal binding.
I love NMR titration experiments. Nice idea, though I would be v surprised if it did not chelate in exactly this way.

[drc007]

@MFernflower I think that will exist as the pyrrole tautomer (hydrogen on the nitrogen).

[MFernflower]

@drc007 if I am not mistaken - the Zinc II ion would still be able to access the nitrogen's lone pair even if it assumes the pyrrole form in water - it's been a while since I studied this so forgive me!

[danaklug]

I've received materials to make these four compounds (brominated compounds will be used for Suzuki couplings). The chloro compounds will also allow for Topliss analysis - which I'm hoping will generate ideas for further analogs.

[danaklug]

Closing - these compounds have been added to the SAR wiki.