Mur Ligase Online Research Meeting November 2021

[mattodd]

Meeting time: Tue Nov 9th 4pm London Location: https://ucl.zoom.us/j/93738560190 Recording: https://youtu.be/67nIvJfADxY Meeting follows on from #51

Actions from last time

1) Laura XChem screen of Atomwise compounds vs EcMurE.

• [ ] @LauraDS1: SPR binding assays with MurC, MurD in presence and absence of ADP/AMPPNP (#55)
• [ ] @Rebecca-Steventon: Activity assays.
• [ ] @mattodd: Contact Atomwise: are these binding to the same site that Atomwise designed for? Atomwise original proposal location: ??

2) MurD/MurE dual inhibitors.

• [ ] @Rebecca-Steventon dose response experiments in progress.
• [ ] @danaklug to post structures of the hits and the non-hits. These structures to be forwarded to Jan for inclusion in the modelling.
• [ ] @LauraDS1 purifying E coli MurD to use in further experiments.

3) Yuhang compounds.

• [ ] @Yuhang-CADD compound structure (AZ cmpd 4) to be overlayed with AZ5595 (PDB 6X9N) and AZ8074 (PDB 6X9F).
• [ ] Peter progressing with Acinetobacter and/or Pseudomonas MurD (nano DSF + crystallization).
• [ ] @LauraDS1 to share her SPR optimization when in hand.
• [ ] @Yuhang-CADD to ship AZ5595 to SGC
• [ ] @Yuhang-CADD to synthesize amine derivative of AZ5595.
• [ ] @Yuhang-CADD: Give his structures an OSA compound code.
• [ ] @chrisdowson1 to work on compound transfer from SGC to Warwick and update the group.

4) Jan Jensen docking (more discussion at #46).

• [ ] @jhjensen2 to get docking poses on top scorers.

Older actions:

• [ ] @Rebecca-Steventon has posted data of @LizbeK fragment screen (https://github.com/opensourceantibiotics/murligase/wiki/Pocket-binding-site - but @drc007 would like structures to be added).
• [ ] @LauraDS1 to continue crystallisation trials of MurD and MurE in Warwick, with the plan to take crystals to Diamond. Streptococcus agalactiae MurD is 1-2 months, and being difficult so this is being parked. E. coli MurE is faster and will be taken to Diamond. @chrisdowson1 says the MurC biochemical assay is not yet on the road, but is certainly coming out of the garage.
• [ ] @chrisdowson1 to liaise with others to establish protein needs, and what's needed from SSGCID <-- this is getting mothbally. Needed?
• [ ] @LauraDS1 to liaise with Peter and Lizbe to see if a soak of the Joe Enamine compounds is possible at Diamond. <-- this is getting mothbally. Needed?
• [ ] @loriferrins pursuing inputs from Cyclica - prediction of actives. Not based on activity (where we do not have many data) but based on crystal structures (as for Atomwise). @eyermanncj did not think there were enough lit data (on binders at the ATP pocket) to build a good machine learning model. The Becca dose response curves may give us that. In the meantime, Peter has structures that may be use to Cyclica. Lori to invite Cyclica to next meeting and put them in touch with Peter Horanyi.
• [ ] @chrisdowson1 to maintain interaction with ELF following their interest in the assay/screen we proposed.
• [ ] Peter Horanyi to post deck that he presented in Aug. meeting.

[Yuhang-CADD]

[jhjensen2]

I probably won't make the meeting today, but @cstein has already posted the poses for the molecules here.

The next step for us is looking for primary amines with good docking scores using genetic algorithms.

[Yuhang-CADD]

Dear all,

@mattodd We have several questions before the meeting starts and wish to get some detailed help from you guys (experts). We would really appreciate it if anyone could help! And please bear my questions as some of them might sound stupid.

1. @Rebecca-Steventon Is that possible that we conduct enzyme inhibition assay against Pae MurC and another ligase, e.g. MurD (preferably from the same species) using WYH9-2-P which we sent to Warwick and Seattle during August. The sample is currently with Laura @LauraDS1?

2. PDB: 7SIR is an apo structure of MurD ligase from Acinetobacter baumannii AB5075-UW. How can we utilise this structure? As far as I know, we do have PDB: 6CAU MurC from Acinetobacter baumannii (strain AB307-0294) available online. Not being able to find a record of inhibitors especially for this Acinetobacter baumannii species of MurC. @eyermanncj

3. @mattodd According to @jhjensen2 @cstein latest results, we found that the following selections have great structural diversity. They (green) were occupying the space where AZ5595 (yellow) from PDB: 6X9N and AZ8074 (cyan) PDB: 6X9F sit.

Do we just buy and test them all? In my opinion it's going to be hard to further narrow the list down. Shall we have any guidance from @eyermanncj as well?

Many thanks, Yuhang

[jhjensen2]

I don't think that it makes sense to narrow further based on the docking score. If you can buy and test them all, great!

But if you need to narrow down further:

• Note that Z57909504, Z2038227779, and Z57907808 have computed logP values >3.5.
• Also, Z57909504 and Z57907808 are very similar.
• Think about synthetic accessibility, in case you want to optimise further and need to make derivatives

[jhjensen2]

PS.Also, look at whether the ligands make the same kinds of interactions (e.g. H-bonds) with the protein as AZ5595 and AZ8074 do.

[mattodd]

Hi @jhjensen2 - sure, but as @Yuhang-CADD 's picture above shows, they make different interactions vs the AZ compound. Close, maybe overlapping, but certainly different. Is that bad? Yes, the repeat of Z57909504 and Z57907808 was the only one I could see.

[jhjensen2]

I think @eyermanncj would argue that the closer the interactions are to those of known binders, the higher the chances are that the compound will actually bind.

[Rebecca-Steventon]

1. @Rebecca-Steventon Is that possible that we conduct enzyme inhibition assay against Pae MurC and another ligase, e.g. MurD (preferably from the same species) using WYH9-2-P which we sent to Warwick and Seattle during August. The sample is currently with Laura @LauraDS1?

Depending on how the sample is stored and how much is left, we should be able to run an initial inhibitory assay test to determine if the fragments are able to inhibit specific Mur ligases.

[Rebecca-Steventon]

I am currently waiting on another meeting which may overrun and so I may be late attending today’s meeting. In regards to the activity assays, we are currently working on the dose response curves for Dana’s compounds against MurD from S.agalactiae, and these are our top priority right now.

[eyermanncj]

A couple of points. Most of these compounds docking scores are 'good' because of a salt-bridge between the ligand and Asp222. Docking scores can be very biased by electrostatic contributions so soem caution is warranted. Also note that pyridine is not protonated at physiological pH - so a protonated pyridine is not the best represented for docking. Similarly the protonation state for the piperazine in the dockign study is not correct. Only the terminal NH is protonated, not both N's of the piperazine.

[Yuhang-CADD]

A couple of points. Most of these compounds docking scores are 'good' because of a salt-bridge between the ligand and Asp222. Docking scores can be very biased by electrostatic contributions so soem caution is warranted. Also note that pyridine is not protonated at physiological pH - so a protonated pyridine is not the best represented for docking. Similarly the protonation state for the piperazine in the dockign study is not correct. Only the terminal NH is protonated, not both N's of the piperazine.

Yeah, I was thinking about this pH and salt bridge issue as well. So, is there any protocol you could suggest to be used for preparing the ligands in the prevention of these over-protonated situations (maybe over-scored)? @eyermanncj Or should we just follow the default Ligprep settings (pH=7.0+/-2.0) for HTVS, then select some of the best scorers whose charging states need to be modified manually and dock them back?

[Yuhang-CADD]

I am currently waiting on another meeting which may overrun and so I may be late attending today’s meeting. In regards to the activity assays, we are currently working on the dose response curves for Dana’s compounds against MurD from S.agalactiae, and these are our top priority right now.

Thanks Rebecca for letting us know! Really appreciate your efforts! So, shall we get WYH9-2-P (the following structure) to the back of the priority list? As we get it from the AZ paper, and we are definitely eager to test it once Dana's compounds are finished! Thank you so much!

[Yuhang-CADD]

A couple of points. Most of these compounds docking scores are 'good' because of a salt-bridge between the ligand and Asp222. Docking scores can be very biased by electrostatic contributions so soem caution is warranted. Also note that pyridine is not protonated at physiological pH - so a protonated pyridine is not the best represented for docking. Similarly the protonation state for the piperazine in the dockign study is not correct. Only the terminal NH is protonated, not both N's of the piperazine.

Yeah, I was thinking about this pH and salt bridge issue as well. So, is there any protocol you could suggest to be used for preparing the ligands in the prevention of these over-protonated situations (maybe over-scored)? @eyermanncj Or should we just follow the default Ligprep settings (pH=7.0+/-2.0) for HTVS, then select some of the best scorers whose charging states need to be modified manually and dock them back?

Thanks, Jan!

Note here: Jan has done the VS for all the structures with all possible protonation states, thus if the final list contains a good scorer with an unrealistic protonation status, we can just remove it. @mattodd

[mattodd]

@LauraDS1 you walked us through a 11-slide deck in the November meeting. If you were able to dump the file somewhere useful (e.g. here, below), that'd be awesome.