J. C. A. Hunt et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5222–5226
5225
choice of peripheral groups that were also compatible
with achieving high levels of antifungal activity.
(CHI).21 However, a structure-based prediction of logP
octanol using ClogP22 was 1.7 log units too low com-
pared to the CHI derived value. This discrepancy
proved to be reasonably consistent across a range of
analogues (Table 2). Similarly the aqueous solubility
of 1 was about an order of magnitude lower than the va-
lue of 0.9 lg/mL estimated from measured logP and
melting point using Yalkowsky’s general solubility
equation.23 This unexpectedly low solubility appears to
be a feature of this class of chemistry. Of the 600 com-
pounds in this series examined approximately 50% had
estimated logP octanol values (obtained by applying a
+1.7 log unit correction factor to the value calculated
using ClogP) in the preferred 1.5–5 range. However,
only about 5% were predicted to have aqueous solubility
greater than 1 lg/mL and those that did failed to satisfy
the SAR requirements for antifungal activity.
Simply increasing the bulk of the group (R1) on the pyr-
idine ring, with the aim of reducing the ability of the
molecules to stack, and hence their crystallinity, proved
unsuccessful due to increased lipophilicity. Thus dim-
ethylmorpholine 30 had a high melting point and low
aqueous solubility and showed reduced antifungal activ-
ity compared to the lead 1. Replacement of the mor-
pholine group with diethylamine (27) also failed to
increase solubility. Changing the amine to an ether
(34) made little difference to the melting point.
Increasing the size of the alkoxy group on the triazine ring
initially seemed more promising, with a significant (46 °C)
reduction in melting point observed on moving from
methoxy (1) to n-propoxy (12). This trend continued, with
the isopentyloxy analogue 15 having a melting point a fur-
ther 24 °C lower. However, as discussed above, this com-
pound was inactive, and both compounds were much
more lipophilic than the lead 1, more than offsetting any
benefit from reduced melting point with respect to solubil-
ity in water.
In conclusion, a novel area of chemistry with promis-
ing antifungal activity, particularly against wheat pow-
dery mildew, has been discovered. The development of
two complementary synthetic routes allowed an effi-
cient exploration of structure–activity relationships
around the pyridothieno-1,2,3-triazine 1. By careful
selection of substituents on the core ring system we
were able to design and prepare compounds with sig-
nificantly reduced melting point and increased water
solubility. However, we were unable to design com-
pounds which both satisfied the structural requirements
for good activity and possessed physical properties that
could allow the expression of this activity in a field
environment.
We therefore decided to explore the insertion of hetero-
atoms, especially oxygen, into the alkyl side chains of
the molecule. This is an approach that we successfully
developed in previous projects,20 which resulted in a
significant increase in solubility of the molecules
without exerting a detrimental effect on lipophilicity
or biological activity. We began by exploring the alk-
oxy group on the triazine, and found that replacing
the n-propoxy group in 12 with methoxyethoxy gave
a compound (36) with a slightly reduced melting point,
but significantly greater solubility than 1. Encourag-
ingly 36 also maintained good biological activity.
Applying the same approach to the substituent on
the pyridine ring also resulted in increased solubility,
without altering lipophilicity or melting point (compare
methoxyethylamine 33 with the morpholine 1). Replac-
ing the N–H in 33 with N–Me to give 32 dramatically
reduced the melting point and significantly increased
the water solubility. Extending this approach to the
bis(methoxyethyl)amine 37 resulted in an additional
increase in solubility. However, as expected from the
SAR discussed above, this compound was less active
due to the large size of the amine substituent on the
pyridine ring. Combining oxygen-containing side
chains of the preferred size at each end of the molecule
resulted in compound 38 which had very similar log P
Acknowledgments
We thank Paul Stanley and Susan Crosland for analyt-
ical support; and Beth Shirley for determining CHI
logP.
Supplementary data
Supplementary data, including typical synthetic experi-
mental procedures and details of the biological test
method used, are available in the online version of this
article. Supplementary data associated with this article
octanol to the original lead 1, but
a
melting
References and notes
point 120 °C lower and sixfold higher aqueous
solubility. Compound 38 also showed good antifungal
activity.
1. Youssefyeh, R. D.; Wilson, J. D. U.S. Patent 4,239,887,
1980; Chem. Abstr. 1981, 94, 139849.
2. Khandwala, A.; Van Inwegen, R.; Coutts, S.; Dally-
Meade, V.; Youssefyeh, R. D. Int. J. Immunopharmacol.
1983, 5, 491.
3. Youssefyeh, R. D.; Brown, R. E.; Wilson, J.; Shah, U.;
Jones, H.; Loev, B.; Khandwala, A.; Leibowitz, M. J.;
Sonnino-Goldman, P. J. Med. Chem. 1984, 27, 1639.
4. Quintela, J. M.; Peinador, C.; Veiga, M. C.; Botana, L.
M.; Alfonso, A.; Riguera, R. Eur. J. Med. Chem. 1998, 33,
887.
A common feature of the compounds made in this series
was that established prediction methods failed to cor-
rectly estimate their physical properties. Thus, for the
lead compound 1 the direct measurement of logP octa-
nol18 gave a value of 4.25 which was in good agreement
with the value of 4.06 determined indirectly from mea-
surements of chromatographic hydrophobicity index