Follow up for the fragment hits for MurD Ligase

[drc007]

Four fragments have been identified, full details are on the wiki.

The next step is compile a list of suggestion for follow up testing. These can be commercial fragments (better) or suggestions for molecules to be made (which will take longer). So feel free to make suggestions, it would be much, much better if you could include the reason behind your suggestion.

[drc007]

From the crystal structure it appears the weakly basic nitrogen (pKa 7.0) of the piperazine interacts with Glu132 Details on wiki. It may be possible to enhance this interaction by increasing the basicity of the nitrogen by moving to the corresponding piperidine (pKa 8.9). The analogue is commercially available CSC048045984.

[drc007]

Looking at the interaction map for 373 it is clear that methylpiperazine is pointing towards solvent and whilst this might be a useful handle for manipulating physicochemical properties it is likely to have a limited impact on binding. In contrast the fluorophenyl ring points into the pocket and might be a better starting point for looking for additional interactions.

This might be a good instance to explore simple halogen scan, looking forward the bromine also makes a good starting point for further modification.

All 3 molecules are commercially available

[bendndi]

@drc007 How much work would it be to run similarity searches for the four identified fragments via the Enamines REALSPACE database? I think your structure based suggestions are a good option but could also be useful to try out the new virtually accessible analoging pioneered by REALSPACE...

https://www.biosolveit.de/realspacenavigator/

[drc007]

@bendndi Yes, I was thinking of that also. Gives me a chance to try it out for real.

On 30 Apr 2019, at 18:37, bendndi notifications@github.com wrote:

@drc007 https://github.com/drc007 How much work would it be to run similarity searches for the four identified fragments via the Enamines REALSPACE database? I think your structure based suggestions are a good option but could also be useful to try out the new virtually accessible analoging pioneered by REALSPACE...

https://www.biosolveit.de/realspacenavigator/ https://www.biosolveit.de/realspacenavigator/ — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/opensourceantibiotics/murligase/issues/1#issuecomment-488044996, or mute the thread https://github.com/notifications/unsubscribe-auth/ABWAURA5VGMLRPZYPZS75ITPTB7VBANCNFSM4G3JSN3Q.

[drc007]

@bendndi Ran all four ligands in "Most Similar" mode using default setting in RealSpaceNavigator. Exported results as sdf file Combined all files.

cat 349similar.sdf 373similar.sdf 374similar.sdf 378similar.sdf > allsimilar.sdf

Imported into Vortex and looked for duplicates, there were none. Ran K-mean clustered to give 50 clusters, chose centroid of each cluster to yield 50 fragments.

Checked against Diamond Fragment library and none of these fragments were in the original screen.

SelectedClustered.sdf.zip

[drc007]

I've created an interactive viewer for the individual fragments bound to the protein using the excellent NGLviewer. The views can be found here.

I've also recorded a very quick intro movie here

[fidiris]

Hi, my name is Fahima and I’m a postdoc working with Professor Mat Todd at UCL. I’m hoping to start making molecules for the Mur Ligase projects soon. I have started to note down some ideas of compounds I could possibly synthesise following discussions with @drc007 and @mattodd last week.

MurD

Fragment 349

Having used the NGL viewer (with help from Chris) we can see that the fragments (349, 373, 374, 378) bind close to the ADP binding site. The aryl ring points towards the ADP site, hence introducing substitution here to grow the fragment may allow access to the binding site.

In terms of piperazine fragment 349, a good starting would be investigating cross-coupling chemistry to quickly access a range of analogues. 1-(4-Bromophenyl)piperazine is commercially available and can be methylated to give methyl piperazine A via a reductive amination following a literature procedure. It is envisaged that aryl bromide A will serve as a key synthetic intermediate to access coupled compounds C-E. Sonogashira, Buchwald-Hartwig and Suzuki-Miyaura cross coupling reactions would likely be the methods of choice to diversify A. Mild reaction conditions and broad substrate scope make these an ideal starting point.

Reference for the synthesis of A and B

Additionally, aryl bromide A can be converted into aryl boronic acid B via a lithiation and subsequent borylation, following a literature protocol. Boronic acid B can be coupled with commercially aryl halides to access analogues which may be difficult to synthesise directly from intermediate A. I’ve noted down some potential scaffolds we can consider making using the different routes described.

Here are some links for reaction conditions for the cross-coupling reactions: • Ligand-free aryl amination using photoredox catalysis reported by Macmillan and Buchwald in 2016. • General set of reaction conditions for the Suzuki coupling of aryl and vinyl halides with aryl boronic acids by Fu and co-workers from 1999. Broad substrate scope and easy to source ligands/additives and catalysts. • Sonogashira cross-coupling condition.

These are just a few procedures to start with. As some of these reactions can be ligand dependent, optimisation may be required.

I can also investigate the ortho and meta-bromo derivatives of A and diversify those. This might give us a better idea of how best to grow the fragments.

Fragment 373

Due to the interaction of the acyl piperazine 373 with a Glu132, Chris has suggested investigating other amines to evaluate the influence of pKa, ring size and geometry. This will likely be more challenging as the obvious route to access these compounds is through an amide coupling between the corresponding carboxylic acid and aniline. I’ve investigated which carboxylic acids I can buy to make a few compounds to start off this part of the project (See below). Several substituted anilines can be coupled. In particular, it would be interesting to make aryl bromide derivatives which can be further functionalised.

I’ve listed some of the acids we can think about buying initially. The quinuclidine derived carboxylic acids tend to be expensive and scarce. However, I can make the 4-carboxylic acid derivative from the corresponding nitrile in a single step following a literature protocol.

Reference for hydrolysis of nitrile.

Apologies for the lengthy post. Any suggestions, advice would be appreciated!

[mattodd]

This looks like a really good analysis and starting point @fidiris. I like that C and D, and the bromo isomers, give us access to a range of compound geometries to try. Does this align with your thoughts @drc007?

[paulking1995]

Might be worth looking at 10.1039/c5sc04751j for a comparison of several reaction conditions for the Suzuki–Miyaura reaction and Buchwald–Hartwig amination (more the latter). Supposedly, the MacMillan aryl amination approach has a broader scope based on the table (but some with complementary-ish reactivity).

MacMillan often makes modifications to get better yields based on SI experimental. I suspect a short optimisation using the bromide will be useful.

[drc007]

@fidiris This is a fabulously comprehensive analysis. This nicely illustrates that with relatively simple chemistry it is possible to scope out potential SAR very quickly. Added to which they offer vectors for additional exploration. I look forward to seeing the first new results.

[drc007]

@mattodd Taken from here https://github.com/opensourceantibiotics/murligase/issues/8

Here is an analysis carried out right at the start of the project by the SGC's Paul Brennan (one of the people behind the science that started this project).

[drc007]

2-methyl-1,3,4,9-tetrahydropyrido[3,4-b]indoles are known in the literature, for example https://pubchem.ncbi.nlm.nih.gov/compound/122504662 which could be a useful intermediate.

[fidiris]

Thanks for posting this here @drc007. I will try to incorporate these structures into my synthesis plan.

[fidiris]

I’ve got some thoughts on potential synthetic routes to the merged fragments suggested by Paul Brennan (#8 ). We can consider a Pictet-Spengler mediated route for the synthesis of fragment 1 illustrated below containing a tryptoline core. Simple bond disconnections would leave us with tryptamine derivative 4 as a key building block. Unfortunately, this is a novel compound, hence a synthetic route to this intermediate would need to be developed. Another issue we may run into is the Pictet-Spengler cyclisation with this substrate. Aromatic compounds containing electron donating groups react more readily (i.e. methoxy) versus those containing deactivating, electron-withdrawing groups.

However, to explore the chemical space around this fragment, we could look at starting with tryptamine and making the core structure of fragment 1. We can then vary the substituents on both the piperidine and indole nitrogens. This would provide a good starting point for making analogues. There is then the option to see which tryptamine derivatives we can buy and functionalise quickly.

Synthetic Procedures: Pictet-Spengler Methylation via reductive amination

[fidiris]

Below are the structures of the compounds that myself and @danaklug have included in the 13/03/2020 shipment to Warwick, grouped by design rationale. We have included the parent fragment hits in the shipment too. These structures will shortly be posted on the wiki!

[danaklug]

Closing this issue - a link has been added to the relevant wiki page

[danaklug]

Screening results have been posted on the wiki

[drc007]

@danaklug Will you titrate using fresh sample?