D. F. Corbett et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4014–4018
4015
the lower affinity of compounds 5 and 6. Varying the lipo-
philicity of the R3 substituent also had a dramatic effect on
affinity. Replacement of cinnamoyl in R3 by benzoyl while
retaining optimum R1 and R2 substituents, as in com-
pound 7, resulted in reduced affinity. The phenyl urea 8
had comparable affinity to the benzamide 7, although
most ureas were much less active than amides. The pipero-
noyl derivative 9 retained moderate affinity while the
trans-3-phenylcyclopropanoyl compound 10 was only
slightly less active than the cinnamoyl analogue 2. How-
ever, the 3-cyclopentylpropionyl analogue 11a gave
enhanced affinity at the 5-ht5A receptor with a pKi of 8.2.
Keeping the favourable R3 3-cyclopentylpropionyl sub-
stituent constant, the effects of varying R1 and R2 were
probed further. Analogues 12 and 13, where R1 is H and
indanyl, respectively, were significantly less active, but
the 2-(2-pyridyl)ethyl (14) and 2-phenoxyethyl (15)
derivatives retained good affinity. Maintaining R1 as
phenethyl and R3 as 3-cyclopentylpropionyl, a range
of R2 substituents containing basic groups afforded
compounds which retained good affinity, e.g., derivative
16, but none of these matched that of 11a. The sulfon-
amide 17, analogous to amide 2, had reduced affinity
with a pKi of 6.6.
Scheme 1. Reagents and conditions: (a) Wang resin (1.7 mmol/g, 150–
300 lm), di-2-pyridylcarbonate, Et3N, CH2Cl2; (b) N-R1-3- or
4-bromobenzylamine, CH2Cl2, rt, or 1,2-dichloroethane, 70 °C; (c)
2-, 3-, or 4-formylbenzeneboronic acid, Pd(Ph3P)4, Na2CO3, 1,2-
dimethoxyethane–water (9:1), 80 °C under argon; (d) R2NH2, Na2SO4,
AcOH, NaBH(OAc)3, 1,2-dichloroethane; (e) R3COCl, Et3N, CH2Cl2,
or R3NCO, CH2Cl2; (f) 20% CF3CO2H, 80% CH2Cl2.
split-mix directed sorting combinatorial process6 to af-
ford the products as single compounds. After cleavage
from the resin, the individual products were obtained
as trifluoroacetate salts in quantities of up to 35 lmol.
The robust nature of the solid-phase chemistry is dem-
onstrated by the isolation of 6625 compounds in >80%
LCMS purity from the planned 6912 component array.
Only 45% of these products required purification, which
was achieved using parallel preparative LC techniques.7
Receptor-binding affinities for a selection of the most
potent compounds against a range of relevant receptors
(5-HT, dopamine, and a-adrenoceptors) are shown in
Table 2. The compounds examined possess generally
good 5-ht5A selectivity when compared with binding
affinities at other 5-HT and dopamine receptor sub-
types, showing a range between 10- and 1000-fold. The
selectivity profile improves with increasing 5-ht5A bind-
ing affinity as demonstrated by compound 11a, which
is J 100-fold selective against the majority of the other
5-HT receptor subtypes. Similarly, amongst the ana-
logues tested, 11a shows the highest selectivity against
the adrenergic a1B receptor (ca. 50-fold). It does, howev-
er, possess significant affinity for the human serotonin
transporter.
A significant advantage of the high-throughput chemistry
approach is that analogues of screening hits are immedi-
ately available and SAR can be established quickly.
Although the array comprised six different biphenyl iso-
mer combinations, designed to allow the substituent
groups to be presented in different orientations, the com-
pounds with affinity for the 5-ht5A receptor consisted al-
most exclusively of the 4,4-isomers. Additionally, it was
clear that a basic group was required in the R2 substituent
and amides were generally much more active than ureas
(data not shown). The most potent compound from the
original array was derivative 2 (Table 1) with pKi 7.6.
The requirement for lipophilic groups in both R1 and R3
limits the opportunity to increase the hydrophilicity of
molecules in the active series. However, some com-
pounds lacking the R3 amide were found to possess
moderate affinity, e.g., the N-benzylpiperidinyl deriva-
tive 18. Interestingly, in this series the 4,3-isomer 19
maintained comparable affinity. However, amines of
the types 18 and 19 lacked selectivity against other
receptors in Table 2 (data not shown).
Affinity was confirmed in the [3H]LSD radioligand-
binding assay by re-preparation of the most potent com-
pounds, and a new focused array (969 compounds) was
synthesised based on the initial SAR, using the solid-
phase chemistry in Scheme 1 in both cases. The two ar-
rays were instrumental in probing further SAR and
selectivity data for the 5-ht5A receptor.
A selection of illustrative SAR data is summarised in
Table 1. Compounds 2–4 exemplify the requirement for
a basic group in the R2 substituent for good affinity.
Replacement of the N,N-dimethylaminoethyl group in 2
with the less basic 3-pyridylmethyl substituent in 3 result-
ed in comparable affinity, whereas compound 4 with a
non-basic substituent was more than 10-fold less active.
From a range of R1 variables, the phenethyl substituent
gave by far the most potent derivatives, as illustrated by
In another approach to reduce lipophilicity, analogues
20–23, in which each phenyl ring in turn is replaced by
a pyridyl moiety, were synthesised using variations of