Help meeting the eNTRy criteria

dayang-us commented 4 years ago

Hi all!

Some background info:

For my project I have managed to modify the AZ hit compound, AZ5595 by adding a primary amine to the ethanolamine moiety. I surmised that it's the best place to add a structure as to not interfere with the hydrogen bonds formed with the scaffold, and the t-butyl and phenyl ethanolamine moieties yielded the submicromolar potency against Pseudomonas aeruginosa.

Rationale: updatedinteractions (Edited 17/12/2020 with correct interactions)

Resulting molecule:

I'm trying to extend its spectrum of activity based on the eNTRy criteria, quoted from the paper below:

• Nonsterically encumbered ionisable Nitrogen (primary amines are the best) • Have low three-dimensionality (globularity < 0.25), defined as the ‘inverse condition number of the covariance matrix of atomic coordinates, 1 is a perfect sphere, 0 is a two or one-dimensional object. On the globularity scale, Benzene has a Glob of 0 while Adamantane has a Glob of 1. • Rigid (rotatable bonds ≤5), defined as the count of single bonds not in a ring, bound to a non-terminal heavy atom, excluding C-N amide bonds because of their high rotational energy barrier.

Additionally: • Has an increased amphiphilic moment (vSurf_A), a measure of vector pointing form the center of the hydrophobic domain to the center of the hydrophilic domain, with the vector length being proportional to the strength of the ampiphillic moment. Molecules with a strong ampiphillic moment will have a hydrophobic region across the molecule from the primary amine.

From their study, there is evidence that adding a primary amine is not enough to extend the spectrum of activity of the compound, e.g. Vancomycin (has a primary amine, 13 rotatable bonds, Glob=0.28, Gram-positive only)

Questions I need help with:

I am trying to put together some data on the molecules I'm studying, but encountered some roadblocks:

Submitting the SMILES notation to the eNTRyway website did not work. (I.e. It gave me an error status) I suspected it could be that my SMILES notation is wrong but it also did not work for SMILES notation extracted from database for known antibiotics such as Ampicillin and Penicillin G. Does anyone have any experience with the eNTRyway website, perhaps they require a specific type of SMILES notation?
The paper used MOE to describe the structures' Glob and VSurfA but I do not have access to the software (also was not listed on my university's database). I tried to find open source alternatives, the closest I could find for Glob is the Spherocity Index from RDKit's Descriptors3D module. I could not confirm if Glob and Spherocity Index are interchangeable. They do use the same scale, 1 = perfect sphere and 0 = benzene. Does anyone have any ideas and happy to help with this?
I tried to glean as much information from DataWarrior's Molecular Descriptors function, and came up with this table:

Name	IC50 (Pae)	MIC (Pae)	IC50 (Eco)	MIC (Eco)	Total Molweight	cLogP	H-Acceptors	H-Donors	Druglikeness	Rotatable Bonds	Basic Nitrogens	Spherocity Index
AZ5595	0.001	1.5	<0.005	3.1	406.492	2.3158	9	4	0.46819	7	3	0
AZ5595_DN1					421.507	1.2103	10	5	0.32681	7	4	0.0289497
AZ8074	1				354.417	1.4819	9	4	0.34658	5	3	0
AZ_9	0.013	0.188	>100	100	372.475	2.0493	9	4	0.303	7	3	0

_NB: AZ5595 is the hit compound, AZ5595DN1 is the molecule with primary amine added, AZ8074 is the chemical synthesis primer, and AZ9 is one of the compounds from the SAR studies with reduced potency.

AZ5595 has promising Spherocity Index but it has 7 rotatable bonds, which is more than the upper limit specified in the eNTRy criteria. The modified AZ5595 naturally will inherit the same number of rotatable bonds, and this time, with increased spherocity. I would really appreciate suggestions on alternative ways of adding the primary amine and meeting the eNTRy criteria from the trained eyes of OSA.

Thank you!

Relevant papers:

Hameed P, S.; Manjrekar, P.; Chinnapattu, M.; Humnabadkar, V.; Shanbhag, G.; Kedari, C.; Mudugal, N. V.; Ambady, A.; de Jonge, B. L. M.; Sadler, C.; Paul, B.; Sriram, S.; Kaur, P.; Guptha, S.; Raichurkar, A.; Fleming, P.; Eyermann, C. J.; McKinney, D. C.; Sambandamurthy, V. K.; Panda, M.; Ravishankar, S. Pyrazolopyrimidines Establish MurC as a Vulnerable Target in Pseudomonas aeruginosa and Escherichia coli. ACS Chemical Biology 2014, 9, 2274-2282. DOI: https://doi.org/10.1021/cb500360c
Richter, M. F.; Hergenrother, P. J. The challenge of converting Gram-positive-only compounds into broad-spectrum antibiotics. Annals of the New York Academy of Sciences 2019, 1435, 18-38. DOI: https://doi.org/10.1111/nyas.13598
Richter, M. F.; Drown, B. S.; Riley, A. P.; Garcia, A.; Shirai, T.; Svec, R. L.; Hergenrother, P. J. Predictive compound accumulation rules yield a broad-spectrum antibiotic. Nature 2017, 545, 299-304. DOI: https://doi.org/10.1038/nature22308

Update 28/10/2020:

Here's a table that summarises the molecules' structures and SMILES code generated from DataWarrior.

Name	Structure	SMILES from DataWarrior
AZ5595		`CC(C)(C)c1cc(Nc2c(cn[nH]3)c3nc(N[C@H](CO)c3ccccc3)n2)nn1C`
AZ5595_DN1		`CC(C)(C)c1cc(Nc2c(cn[nH]3)c3nc(NC(C(N)O)c3ccccc3)n2)nn1C`
AZ8074		`OCC(CCCC1)N1c1nc(Nc2n[nH]c(C3CC3)c2)c(cn[nH]2)c2n1`
AZ9		`CC(C)[C@H](CO)Nc1nc(Nc2nn(C)c(C(C)(C)C)c2)c(cn[nH]2)c2n1`

The SMILES notations seems to communicate well with CHEMBl's MarvinJs Import structure function and OpenBabel did detect them and converted into sdf files. However, I think I've generated 4 different SMILES notation for AZ5595_DN (generated by DataWarrior, ChemSketch and ChemBL search query). Not sure if there's an error with the sketched structure that is causing the problems? Tautomeric states and chirality perhaps?

mattodd commented 4 years ago

Very nice work. Odd that the SMILES input didn't seem to work for you. Are the strings working in anything else? Want to post them here? For the molecule suggestion, did you think about just replacing the OH with NH2? Synthetically probably easier, I would guess? Could you also please just clarify (maybe by just adding a figure to the above) what AZ8074 and AZ9 look like?

Andy-Wilkinson commented 4 years ago

I've had a look at the eNTRyway website and it seems to be giving an error with any submission at the moment (including some that previously were fine) so it seems to be an issue there.

You can install the eNTRyway tool locally from https://github.com/HergenrotherLab/entry-cli. Just to test this I've run the proposed compound successfully,

> python calc_props.py -s "CN1N=C(NC2=NC(N[C@@H](C(N)O)C3=CC=CC=C3)=NC3=C2C=NN3)C=C1C(C)(C)C"
..tot conformations = 186624
..tot confs tested = 100000
..below energy threshold = 40509
Properties for CN1N=C(NC2=NC(N[C@@H](C(N)O)C3=CC=CC=C3)=NC3=C2C=NN3)C=C1C(C)(C)C
--------------------------
Mol. Wt.:       421.498780
Formula:        C21H27N9O
RB:             7
Glob:           0.106897
PBF:            1.080287
primary_amine:  True

Regarding the MOE 'Glob' descriptor, my understanding that the eNTRyway tool's globularity is designed to be a freely-available mimic of this.

dayang-us commented 4 years ago

Very nice work. Odd that the SMILES input didn't seem to work for you. Are the strings working in anything else? Want to post them here? For the molecule suggestion, did you think about just replacing the OH with NH2? Synthetically probably easier, I would guess? Could you also please just clarify (maybe by just adding a figure to the above) what AZ8074 and AZ9 look like?

Hi Prof @mattodd, I've updated the Issue with the chemical structures and SMILES notation. I agree with replacing OH with NH2, I'll look into this for the 2nd iteration of the project! Thank you!

dayang-us commented 4 years ago

I've had a look at the eNTRyway website and it seems to be giving an error with any submission at the moment (including some that previously were fine) so it seems to be an issue there...

Hi @Andy-Wilkinson this is really great, thank you so much for helping out and linking me to the internal eNTRyway tool! Will definitely make good use of it.

Update:

python calc_props.py -s "CC(C)(C)c1cc(Nc2c(cn[nH]3)c3nc(N[C@H](c3ccccc3)N)n2)nn1C"
..tot conformations = 62208
..tot confs tested = 62208
..below energy threshold = 31362
Properties for CC(C)(C)c1cc(Nc2c(cn[nH]3)c3nc(N[C@H](c3ccccc3)N)n2)nn1C
--------------------------
Mol. Wt.:       391.472800
Formula:        C20H25N9
RB:             6
Glob:           0.094911
PBF:            0.992900
primary_amine:  True

Tried it with a different molecule and the program worked perfectly, thanks again 👍

jhjensen2 commented 3 years ago

@dayang-us I have made an RDKit implementation of vsurf_A. I couldn't find the details of the MOE implementation so I just used what is in RDKit and scaled the values empirically to values that are somewhat similar to those in the Richter paper.

dayang-us commented 3 years ago

@dayang-us I have made an RDKit implementation of vsurf_A. I couldn't find the details of the MOE implementation so I just used what is in RDKit and scaled the values empirically to values that are somewhat similar to those in the Richter paper.

This is really great, @jhjensen2! Thank you very much for this.

juliocoll commented 2 years ago

just to mention, Data Warrior has an option to calculate globularity.

In case you have not used, the option is in DataWarrior/Chemistry/From chemical structure/Calculate Properties/3D/ It gives two globilarity calculation options. I copied both down:

Globularity (flat/linear <0.5 <spherical) using singular value decomposition of 3D-atom coordinates
Globularity(non-spherical <0.9 <spherical) from molecule volume volume and surface (VDW-radii, 1.4 A probe) As I found out only the first one, called SVD in the resulting calculation column, shows correlation with the globularity values reported by the Ritcher's 2017 paper.

opensourceantibiotics / murligase