Mur Ligase Soak-In Wednesday June 8th 2022

[mattodd]

Meeting will be at 2pm UK time on Wed June 8th at https://ucl.zoom.us/j/4808072370

Aim: To gather together what is known of the mur ligases from crystallography, and therefore what might be needed.

Host: @LauraDS1

Present: @mattodd @loriferrins @eyermanncj @chrisdowson1 Jan Abendroth @bartrum @Yuhang-CADD @LizbeK

Recording: https://youtu.be/pa3yXsd1MrI

Agenda/minutes:

1. Summary of SSGCID efforts

Slides were presented by Jan:

A summary of MurC proteins at SSGCID. Structures are at PDB 6CAU (AcbaC), 5VVW, 6X9F, 6X9N (PsaeA). Structures overlap well. Could get plenty of protein, but difficult to crystallise. Apo and ligand bound in same crystal form. A summary of crystallisation conditions was presented. Could get a MurC AcbaC with an AZ compound?

Also a summary of MurD proteins at SSGCID. Structures are at 7SIR (apo), 7TI7 (ADP) (AcbaC); 7SY9 (apo) and 7U35 (ADP) (PsaeA)). Latest news: MurD + UMA (both AcbaC (important based on pathogenicity/clinical need!) and PsaeA) has given crystals quickly, structures forthcoming (great news!). Inclusion of UMA appears to help. There are significant conformational changes upon ADP binding ("Pacman movement") - does that change where/how the frags bind?. @eyermanncj asked whether the crystals could be soaked with AZ compounds. But there were no shifts in nanoDSF for MurD and the AZ compounds. Only got that with MurC.

Summary of nanoDSF results with AZ compounds. Two batches of compounds tried. Only saw stabilisation for some compounds with MurC, not MurD. So get a MurC AcbaC with an AZ compound. @eyermanncj was most interested in AcbaC MurC structure with the AZ compounds because we don't have biochemistry for that. (the AZ compounds have given nothing with MurD). Jan can request crystallisation set-ups for that immediately. @LauraDS1 mentioned that there is one AcbaB MurC structure with APMP bound.

MurE at SSGCID: Interesting since one of the AZ hits hits MurC and MurE but not MurD (enzymatically). (Is the reason for this the modified lysine?). Jan says SSGCID does not have MurE. @bartrum said that they do - two entries in the database. Need target numbers - Bart to send to Jan, who will investigate.

An explanation was given of the meaning of the tags B1, Bn, K1, K11, K12, K13.

Everyone was very appreciative of all the work that SSGCID had been doing.

2. Summary of XChem/CMD/Warwick work

Slides were presented by @LauraDS1.
EcMurE: Isopropanol was causing problems. Apo structure has not been possible. Can we add in UMA? Yes, next step is with ligands.
Crystallisation with added small molecules (e.g. Atomwise compounds) generates the new pocket, which is otherwise not there. The dual inhibitors (synthesised at UCL) gave no structures with compounds bound, but protein movement could be seen. Needs further investigation.
(Follow-up: some compounds were mentioned that had been purchased by Tobias and Lizbe - what were these?)
@eyermanncj mentioned that there was also one fragment bound in the ATP pocket from the Xchem frag screen (which we're not pursuing).

EcMurD:
Apo needs his tag to crystallise.
Soaking with the crystals: a good pocket is created but as an artefact. Has been searching for different conditions, but they all generate the same pocket, unfortunately. One dual inhibitor compound bound, but at the crystal contact.

Generally apo has been with a low hit rate. A summary was given of what has been tried to generate the more closed protein conformation. Structures of EcMurD have been easiest, with addition of ADP, AMPPCP and UMA (slide - see below) but not yet with the dual inhibitors. In the latter (UMA) there is no magnesium bound. The others (ADP and AMPPCP) there are no structures like this in the PDB, so Laura will submit.
@eyermanncj mentioned that there are M.tb. structures (MurE + ADP + something else) in the PDB that might help find good conditions - @LauraDS1 is aware of such conditions. Might the magnesium be held in place by a cofactor?

Structural comparison on EcMurD ATP site (slide below): big shift between apo and ADP bound. In contrast, structures of ADP and UMA bound are very similar. @LauraDS1 is following up on the ADP-bound structures for soaking since easier to remove ADP than AMPPCP (and cheaper).

An action plan was presented by @LauraDS1 (below):
Run XChem with EcMurD with both ADP and UMA
Co-crystallisation with the new hits obtained from enzymatic assays (even though still micromolar).
Soaking of new hits:

• Apo EcMurD and EcMurE.
• EcMurD:ADP
• EcMurD:UMA
  Prepare more screens to search for EcMurE closed conformation structures
  EcMurC or PaMurC crystallisation work including apo and closed conformations
  Meet up again in a few months, e.g. in Oct/Nov? @mattodd to set up.

3. Next steps

• Some specific steps are listed above
• Publication should be pursued (as opposed to just PDB depositions). Do we need one consistent set of structures to help the story? Structures need enhancing with some biochemical data? What do the structures tell us about the protein dynamics? What lessons have been learned in attempts at apo vs co-cryst and the importance of the close conformations? - Jan's involvement will (officially) wind down after 3 months, since UCB not part of SSGCID going forward. So any papers should be soon. @bartrum proposed a data meeting - what do we have now with no extra experiments? So what's the title, i.e. aim of the paper? "Driving towards multi-targeted inhibitors..." - combination of crystal structures with compound bound plus enzyme inhibition data.
• Next steps involve targeted soaking of small number of hits. Can be informed by E coli biochemical data that is coming from Warwick screens.
• Large scale X-chem screen, what needed? Can happen when @LauraDS1 gets enough crystals. Highly reproducible protocol needed.
• As usual, we need to move towards funding a major body of work.

@chrisdowson1 mentioned that project students have recently characterised ATP-dependence of the hits identified so far vs MurD and MurE. The new hits are going to be discussed in #76.

Organise a SSGCID handover to Lizbe/Laura?

(Can share slides here (on this Issue) or in this folder.)

[mattodd]

Slides from Jan Abendroth

[mattodd]

Slides from Laura