When I'm designing small molecules, I generally expect each nitrogen atom to form 3 covalent bonds to other atoms. This can be like NH3 in ammonia or -NH- in protein peptide bonds. Nevertheless, sometimes the small molecule design tool makes a nitrogen atom form 4 covalent bonds to other atoms, like -NH3 at the end of a long chain or -NH2- when the nitrogen is next to -O-. Why does the design tool do this? What are some rules that nitrogen follows in the design tool? Also, is -NH3 positively charged? Is there a way to show the charge on atoms within the design tool or in the small molecule properties window?
A related question is how many donors and how many acceptors does each of -NH-, -NH3, and -NH2- contain? I'm thinking -NH- has 1 donor and 1 acceptor, but what about the others? Do -NH3 and -NH2- each contain 1 acceptor? Does -NH2- contain 1 or 2 donors? Does -NH3 contain 2 or 3 donors? Why do they behave the way they do? Can you recommend a reference that explains it all?
The change from three to four bonding entries has to do with the fact that the lone pair electrons on the nitrogen is basic at physiological pH, and will often pick up a hydrogen ion and become charged, depending on the exact molecular context it's found it. Foldit has special-casing code to make sure that the molecule built is always in the charge and protonation state that would be relevant in your body. So -NH3 would indeed be positively charged, though there isn't any display in Foldit to show that (but it is tracked internally).
It's the lone pair on nitrogen which is the hydrogen bond acceptor, so a nitrogen will only be an hbond acceptor if it has only three bonds to it. But sometimes those lone pair electrons will be tied up in delocalized bonds, so roughly speaking a nitrogen is only a hbond acceptor if it has one single bond and one double bond.
Hbond donors are easier – pretty much any hydrogen which is attached to a nitrogen will be an hbond donor, and there's no distinction between the hydrogens on the atom. – That said, the counting of total donors on a per-molecule basis may be different from this, as sometimes for medicinal chemistry purposes you don't count the number of potential hydrogen bonds, but instead count the number of heavy atoms which can form hydrogen bonds. So each nitrogen may contribute "1" to the hbond donor count, regardless of how many hydrogens it has (well, assuming it has at least one).
I think the small molecule properties window treats the -NH3 groups as -NH2 groups when it calculates the polar surface area and the molecular weight of the ligand.
I noticed the polar surface area problem when trying to fix one of my structures for Puzzle 2278 (Je11-2 20217.736 +7310.9 med 811pm 3/23/23). This structure had for its Ligand TPSA objective a bonus of 310.85 out of 1000. Foldit said its TPSA value of 139.69 was over 120, so I should remove polar atoms to fix it. Meanwhile, in the small molecule properties window, it said the PSA was 111. I tried removing some polar atoms and got TPSA=137.04 when the PSA was 108. Then I got TPSA=124.15 when the PSA was 95. Finally, when the PSA was 83 or 70, I got the full 1000 point TPSA bonus, so I think the TPSA was below 120 then. Are TPSA and PSA the same thing? If the ligand has fewer polar atoms (like less hydrogens attached to nitrogens), its polar surface area should be smaller. This makes me think TPSA is calculated using -NH3 groups while PSA is calculated using -NH2 groups. What do you think?
I noticed the molecular weight problem on a different 2278 solution called Je16-7 (look for shares from March 24 2023 with scores like 17946.437 +7250 or 18413.581 +7400). I took some snapshots of it with all hydrogens showing. Then I sketched it and counted all of its atoms. It had 11 C, 3 N, 3 O, and 23 H atoms. Using the molecular weights 12 14 16 & 1 for C N O & H respectively, I calculated a molecular weight of 245 for the ligand. Meanwhile, the small molecule properties window said the molecular weight was 243, as if 2 of the ligand's H's were missing. The 3D image of the ligand in the Foldit GUI showed it had 2 -NH3 groups, so I think the small molecule properties window was treating these 2 -NH3 groups as if they were 2 -NH2 groups.
Yes, for consistency with the "standard" way of doing calculations outside of Foldit, we often change the molecule back to the neutral form before computing properties. It's been a while since the ligand properties panel was touched, so it may be that it's not doing the proper neutralization first, so you may have that discrepancy with the objectives.
