Topic automatically created for discussing the designs at:
https://covid.postera.ai/covid/submissions/EDJ-MED-edeb0d3a
Hi Ed @edgriffen, comparison of potencies of MAT-POS-916a2c5a-2 and BEN-DND-7e92b6ca-2 suggests that N-methylation of amide is likely to lead to significant reduction in potency. Inspection of bound conformation in x2910 may provide insights into this SAR feature. You might want to check the 400 nM potency for ALP-POS-05819dc4-1
Interesting data for MAT-POS-3b536971-1 just received. Thanks for the heads up Pete, I’ll have a look.
Hi Ed, let’s discuss more when the data for MAT-POS-3b536971-1 are on the system and I’ll also mention @mc-robinson. I suspect an interaction between methoxy O and amide carbonyl C and this may be relevant to SAR. The quinolones and pyridines/isoquinolines present their amide carbonyl Os to different HB donors in the binding site. I believe that the HB donor exploited by the quinolones is more productive from the potency perspective as mentioned in this discussion.
So Pete, if you have a look at this: https://fragalysis.diamond.ac.uk/viewer/react/projects/409/306, the crystal structures of MAT-POS-916a2c5a-2 and BEN-DND-7e92b6ca-2, the NH and NMe respectively, the N-Me is as you suggest messing with the conformation. So… given we want to loose the NH to improve permeability, suggestions for what to put on the amide N rather than Me. Would NCH2C#N given enough of a dipole dipole interaction to put the C=O and the NCH2C#N trans?
Hi Ed, my cursory examination of the situation points to sterics (as opposed to electrostatics) and the usual fix is to form a ring (I see that we also have piperazines that have been assayed and some azetidines have also been submitted as designs). I’ll take a look in the CSD after I return from the north coast tomorrow and will hopefully get back to you tomorrow.
Hi Ed, I’ve not found anything in the CSD that adds signifcantly to whaty I was saying previously. The piperazine-linked inhibitors ERI-UCB-a0b0dbcb-2 and ERI-UCB-a0b0dbcb-12 are actually a bit more potent than MAT-POS-916a2c5a-2. Neither ERI-UCB-a0b0dbcb-6 (pyrrolidine-linked) nor ERI-UCB-a0b0dbcb-8 (azetidine-linked) appears to show any real activity. It might just be possible to target the catalytic cysteine by linking an electrophilic warhead to the piperazine. A substituent on either carbon bonded to the amidic N of the piperazine is likely to have an axial preference.
I’d just wanted to point out that MAT-POS-54c4bf04-3 is more potent than both ERI-UCB-a0b0dbcb-2 and ERI-UCB-a0b0dbcb-12. Is something fundamentally wrong with isoquinolines?
1 Like
I’m not aware of specific problems with isoquinolines although the benzenoid ring is potentially vulnerable to CYP-mediated metabolism (as would also be the case for isoquinolines) and, as noted by @JSPEN, the aromatic nitrogen makes the isoquinolines potential substrates for aldehyde oxidase. That said, there may be specific problems associated with the isoquinoline substructure when it is incorporated into one or more of the structural series that form the basis of project MedChem. I believe that there is microsomal stability data available for some project compounds although I’ve not seen any of this and I don’t know whether there are any plans to make the data more widely available.
The IC50 ratios for the matched molecular pair MAT-POS-54c4bf04-3 / ERI-UCB-a0b0dbcb-2 are both below the MSR cutoffs of 3.6 (fluorescence assay) and 3.1 (RapidFire assay) proposed by @Wal-Ward. That said, it’s always good to have options since these give the project team room to manoeuvre. The ‘unmatched’ hydrogen bond donor of the quinolone may weaken affinity and PET-UNK-82544c07-1 has been ordered. The ‘unmatched’ HB donor is absent in this 4-pyridone analog and I’d also expect the the carbonyl oxygen to be a very strong HB acceptor (as discussed in this article). What is not clear right now is how much of the potency benefits of the benzenoid ring of quinolones (and isoquinolines) are due to protein-ligand contact (as opposed to conformational biasing).
@miko_a, on this issue. I believe it’s more of a diversity strategy. Going all in on the isoquinolines would be troublesome if we find something of that substructure that can’t be worked out as @pwkenny mentioned. So far we do have solubility data that suggests the quinolones are more soluble than the isoquinolones, and of course Cyp inhibition is a worry – though the data is just coming back. On the other end, the quinolones may have some cell permeability issues.
@pwkenny, we do have initial ADME data, the issue is just getting it all into formats which are presentable. It has come back in a scattered fashion from multiple places, so the aggregation is happening now.