Resonance hybrids + Contribution of delocalization to logP

Hello





I'm using Marvin 5.0 on a Mac and I love it.





I work with lipophilic weak acids that remain membrane permeable when ionized because the negative charge is so well delocalized over a large part of the molecule. I am interested in quantifying the importance of this charge delocalization in my structure-activity relationships.





First, I am wondering if Marvin can generate a resonance hybrid (an average of all resonance structures). This intuitively would be more useful than the canonical form, which I assume to be the most important contributor and not an average.





Second, I am wondering whether the logD calculator takes into account the extensiveness of charge delocalization? I am surprised that the calculated logP of my molecules typically falls from around +3 for the neutral molecule to around -1 for the ionized form. This appears inconsistent with the known conserved lipophilicity.





We can use phenol as a simple example. Its logP drops from +1.76 to -0.82 upon deprotonation. I believe the effect of negative charge delocalization throughout the ring is underestimated in the calculation.





Cheers


Louis


Canada

Hello again





In my first post (above), I asked whether extensive electron delocalization was taken into account when calculating logP of ionized molecules. I would also like to know whether the negative effect of non-planarity on lipophilicity is also taken into account. I have a feeling that it is not or that it is underestimated.





Cheers


Louis


Canada

Hi,

Quote:

First, I am wondering if Marvin can generate a resonance hybrid (an average of all resonance structures). This intuitively would be more useful than the canonical form, which I assume to be the most important contributor and not an average.

Do you mean that the graphical representation of a molecule need to show anyway the delocalizable "islands" of atoms which are belongs to the same pi-system? Or do you mean something else on the generation of the resonance hybrid?





In Marvin the graphical representation of a delocalizable pi-system only available for aromatic rings.

Quote:

Second, I am wondering whether the logD calculator takes into account the extensiveness of charge delocalization? I am surprised that the calculated logP of my molecules typically falls from around +3 for the neutral molecule to around -1 for the ionized form. This appears inconsistent with the known conserved lipophilicity.

logP increment of the ionized atoms depends, of course, on the delocalizability of the environment of the charged atom. This is considered in the logP calculator. For example see the attached picture of an aliphatic and an aromatic alcohol. I think don't need to explain why the O(-) atoms have different logP increments. Could you send an example when you faced with an inconsistency?

Quote:

We can use phenol as a simple example. Its logP drops from +1.76 to -0.82 upon deprotonation. I believe the effect of negative charge delocalization throughout the ring is underestimated in the calculation.

Is this the "inconsistency" that you mentioned above? Or do you believe that the "DAMPING" effect of the delocalization on the logD decrease should be more enhanced? Do you have an experimental data for the pH-partition of phenol?

Quote:

I would also like to know whether the negative effect of non-planarity on lipophilicity is also taken into account.

What do you mean on this? Please show an example.








Jozsi