solvent/binding/logP/R3N

[BAELL[ AT ]mel.dah.csiro.au (Jonathan Baell)]

 Dear All,
 A few questions:
 --------------------------------------------------------------------------
 A: Does anyone have a pointer to a good general discussion of approaches
 to incorporate environmental effects on molecular conformation such as
 might be experienced in a protein binding site and in an aqueous
 solution): e.g the pros and cons of simple distance-dependent dielectric
 methods (incorporating a constant of 1 or maybe 4?/using MNDO ESP
 charges?); more complex continuum methods; explicit solvent molecule
 calculations; treatment of highly basic amino groups in a protein binding
 site.
 B: Related to A; Didesmethylchlorpromazine binds much more poorly to some
 sites than chlorpromazine.  Its terminal amine (RCH2NH2) has been shown
 to be significantly more basic than that in chlorpromazine (RCH2NMe2)
 [which was unexpected as far as I'm concerned], so desolvation problems
 may account for its weaker activity, or perhaps loss of favourable
 binding by the two methyl groups; but has there been discussion in the
 literature on the general principles of alkyl-NH2 versus alkyl-NMe2
 receptor-ligand effects, two groups which are highly prevalent in
 biologically active molecules?
 C: I have seen a number of studies where a linear relationship has been
 reported between nematocidal activity of C. elegans and affinity at
 binding sites in C. elegans, suggesting an absence of a lipophilic effect
 on transport, which is surprising for a whole organism assay if
 membrane-transport is required to reach the active site, and one would
 usually expect this to be the case.  The dataset in these cases has a
 reasonable spread of logP values.  The nature of these studies is such
 that the nematode is subjected continuously to toxin over a long time
 (16h), rather than as a bolus-type dose.  Can anyone point me to a good
 discussion of the relationship between time/dosage regimen and
 logP/activity, and why the nature of an assay such as the C. elegans one
 cited above might minimize lipophilicity as being an observable modulator
 of toxicity?
 D (and potentially paradoxical c.f. question C): For a certain group of
 compounds, we have the following QSAR for larvicidal activity for a
 particular species of parasitic nematode:
 pLD50 = 2.11xpi - 0.782xpi^2 + 4.79
 95%CI     0.72      0.21           0.49
 T value   6.95      -8.92
 n=10, s=0.205, Table T-test 2.37 (95%)
 Y-mean=5.469, Y-variance = 0.833
 Variance explained by regression=95%
 F-ratio= 70.6, r=0.976, F(2,7,0.01)=9.55
 What concerns me is the slope of 2.11.  What does this mean biologically?
  I thought the slope in such equations was at most around unity for
 hydrophobically sensitive systems.  Could this value be explained by
 contributions of a substituent both to molecular lipophilicity (transport
 in vivo) and to binding to a localised hydrophobic pocket?
 The assay involves continuous exposure to active constituent in an agar
 environment over several days.
 --------------------------------------------------------------
 Any input would be greatly appreciated on any of these questions
 (including an internet address for a good QSAR discussion group, if there
 is one)
 Jonathan Ball
 email:    ball[ AT ]mel.dah.csiro.au or baell[ AT ]mel.dah.csiro.au