Synthesis and Biological activity of kappa opioid receptor agonists. Part 2: Preparation of 3-aryl-2-pyridone analogues generated by solution- and solid-phase parallel synthesis methods

Article abstract of DOI:10.1016/S0960-894X(03)00033-7

New analogues of the previously described 3-aryl pyridone KOR agonists have been synthesised by parallel synthetic methods, both in solution- and with solid-phase chemistry, making use of the well known and versatile Mitsunobu, Suzuki and Buchwald reactions. Opioid receptor binding data for the compounds produced is reported.

Full text of DOI:10.1016/S0960-894X(03)00033-7

Bioorganic & Medicinal Chemistry Letters 13 (2003) 1141–1145
Synthesis and Biological Activity of Kappa Opioid Receptor
Agonists. Part 2: Preparation of 3-Aryl-2-pyridone Analogues
Generated by Solution- and Solid-Phase Parallel Synthesis
Methods
Graeme Semple,a,* Britt-Marie Andersson,b Vijay Chhajlani,b Jennie Georgsson,a
Magnus J. Johansson,a Asa Rosenquista and Lars Swansona
aDepartment of Medicinal Chemistry, AstraZeneca R&D Molndal, S-431 83, Sweden
¨
bDepartment of Biochemistry and Cell Biology, AstraZeneca R&D Molndal, S-431 83, Sweden
¨
Received 27 August 2002; accepted 17 December 2002
Abstract—New analogues of the previously described 3-aryl pyridone KORagonists have been synthesised by parallel synthetic
methods, both in solution- and with solid-phase chemistry, making use of the well known and versatile Mitsunobu, Suzuki and
Buchwald reactions. Opioid receptor binding data for the compounds produced is reported.
# 2003 Elsevier Science Ltd. All rights reserved.
We reported recently on the design and synthesis of
potent and selective pyridone-based agonists of the
kappa opioid receptor (KOR) based on our under-
standing of the mode of binding of the classical N–C–
Our original synthesis of the 3-aryl-pyridin-2-one series
of KORagonists made use of a sodium hydride medi-
ated alkylation of 3-aryl- or 3-bromo-pyridin-2-one
with 1-(2-chloro-2-phenylethyl) pyrrolidine in DMF.1
For the first stage of our initial SARexpansion, we
employed the same synthetic step to investigate the
effect of incorporating additional nitrogen atoms into
the central heterocyclic portion of the molecule, as
shown in Figure 1. In each case, the incorporation of a
single nitrogen atom reduces the calculated logP of the
resultant compound by 1.5–2 log units compared to the
corresponding pyridone. The pyridones 4a,b (Fig. 1,
X,Y,Z=C) have been described previously.1 The synth-
esis of the analogous pyrazinone 4e (Y,Z=C; X=N)
was achieved in two steps by condensation of phenyl
glycinamide with glyoxal,4 followed by alkylation of the
resultant heterocycle 6 with 1-(2-chloro-2-phenylethyl)
pyrrolidine regiospecifically and in good yield, using the
same conditions as for the pyridone. The pyridazinone
4f (X,Y=C; Z=N) was prepared by alkylation of
5-phenyl-4,5-dihydropyridazin-3-one 7,5 again under
the same conditions. The major product of the reaction
was the pyridazinone resulting from alkylation and
concomitant oxidation of the heterocycle. Two minor
non-oxidised alkylation products were also isolated, and
HPLC-MS analysis indicated that they had arisen from
direct alkylation of the dihydro- ring, suggesting that
1
C–N–CO KORagonist pharmacophore. As part of
our continuing examination of compounds of this type,
we now wish to describe the synthetic strategy used to
prepare, and the opioid receptor affinity of, a selection
of related analogues. A key driver in these endeavours
was to introduce more polar substituents to reduce
lipophilicity and increase the polar surface area of the
resultant compounds, the overall objective being to
reduce their ability to cross the blood–brain barrier
whilst hopefully maintaining efficacy at the receptor.2
Such compounds may still act at peripherally located
KORs and thus may be useful for the treatment of
visceral pain, avoiding as they would the well-docu-
mented central side effects of previously known KOR
agonists.3 To this end, we have now developed highly
flexible parallel synthesis methods amenable to the pre-
paration of a wide variety of substituted analogues for
further SARexploration.
*Corresponding author at current address: Arena Pharmaceuticals,
6166 Nancy Ridge Drive, San Diego, CA 92121, USA. Fax: +1-858-
453-7210; e-mail: gsemple@arenapharm.com; http://www.arenapharm.
com
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00033-7

1142
G. Semple et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1141–1145
Emax>95% relative to U-69593) in our in vitro func-
tional test.8 However, due to their 10–20-fold weaker
activity when compared to 4b, these structures were not
pursued further.
Whilst the alkylation reactions outlined above proceed
cleanly and in adequate yield, there are limitations to
the range of chloro-amines that can be prepared for use
in the alkylation step. This component was obtained by
treatment of 2-pyrrolidino-1-phenylethanol with thionyl
chloride under reflux,9 and dehydration became a sig-
nificant competing process when we tried to apply these
conditions to other related examples. Moreover, such
compounds are mustard agents and must be handled
with extreme caution. Thus, as an alternative to the key
C–N bond formation step we investigated the use of the
Mitsunobu reaction, which can readily tolerate a wide
range of functional groups and generally proceeds
under mild conditions. The reaction has also been
adapted to both solution and solid phase multiple par-
allel synthesis, and has also been described using solid
supported reagents.10 In addition, a wide variety of b-
amino-alcohols, the considerably more benign building
blocks required to couple to the pyridones, can be
readily prepared in one or two steps from commercially
available styrene oxides or aryl bromo-ketones as out-
lined in Figure 2. We found the latter route to be the
most expedient, avoiding as it does the issue of regios-
electivity in the epoxide ring opening, which can change
significantly depending on the substituents on the Ar
group.
Figure 1. Reagents and conditions: (i) H2O2, AcOH, 80 ꢁC, 18 h, 90%;
(ii) Ac2O, 80 ꢁC, 43%; (iii) (a) NaH, DMF, 0 ꢁC, 1 h; (b) 1-(2-chloro-2-
phenylethyl) pyrrolidine, rt 55–60%; (iv) CHOCHO, methanol, 1 M
NaOH, 69%; (v) (a) NaH, DMF, 0 ꢁC, 1 h (b) 1-(2-chloro-2-phenyl-
ethyl) pyrrolidine, rt 32%; (vi) CH3CO3H, H2SO4, acetone, rt to
reflux, 2 h, 88%; (vii) ArB(OH)2, Pd(OAc)2, dioxan, NaHCO3, 80 ꢁC,
4 h 60–70%.
the oxidation event may be secondary to alkylation. Nei-
ther of these purported dihydro- compounds showed any
biological activity and were not further characterised.
The pyrimidinone analogues 4c,d ((X,Z=C; Y=N)
were prepared in three steps, the aryl group being
appended in the final step by Suzuki coupling. Com-
mercially available 5-bromopyrimidine 8 was converted
to the bromopyrimidone 9 in one step by treatment with
peracetic acid and sulfuric acid as described previously.6
Alkylation of this bromo-heterocycle with 1-(2-chloro-
2-phenylethyl) pyrrolidine to provide 10, again pro-
ceeded in good yield. Suzuki coupling of the requisite
boronic acid was carried out under the conditions pre-
viously described.1 Purification of all the analogues in
this phase of the study was by conventional column
chromatography and the data provided in Table 1
represents mean IC50 values for the fully purified com-
pounds.7 It can be seen from these data that the inclu-
sion of an additional nitrogen atom in the central
heterocyclic portion of the molecule the parent struc-
ture, results in a decrease in affinity for the cloned
human KORby at least one order of magnitude. It
appears however, that additional substituents to the
terminal ring may provide a similar SARto that for the
pyridone, as the 2,4-dichloro substituted pyrimidone 4d,
showed a significant improvement in binding over 4c. In
addition, 4d maintained good selectivity over the other
opioid receptors tested and showed a clear full agonist
effect (ED50=615ꢀ66 nM cf. 24ꢀ1.8 nM for U-69,593.
Comins and co-workers have previously described con-
ditions (triphenyl phosphine and diethylazodicarboxy-
late at rt in THF) under which they observed selective
O-alkylation or mixtures of O- and N-alkylated pro-
ducts from the reaction of 2-pyridone with primary or
secondary benzyl alcohols.11 Undeterred, we investigated
Figure 2. Reagents and conditions: (i) neat amine, reflux, or amine in
methanol, reflux 15–90%; (ii) (a) amine, DIEA, DCM, rt; (b) NaBH4,
MeOH, water, rt 40–75%; (iii) Bu3P, TMAD, THF, 18 h, 40–80%;
(iv) Bu3P, TMAD, THF, 18 h, 88%; (v) Suzuki, ArB(OH)2, Pd(OAc)2,
dioxan, NaHCO3, 80 ꢁC, 4 h, 40–90% or Buchwald, Pd2(dba)3,
R-BINAP, dioxan, 90 ꢁC, 18 h, 40–90%.
Table 1.
No.
Ar
X
Y
Z
k-Binding IC50 (nMꢀSEM)a
d-Binding IC50 (nM)
m-Binding IC50 (nM)
4a
4b
4c
4d
4e
4f
Ph
2,4-Cl2Ph
Ph
2,4-Cl2Ph
Ph
Ph
C
C
C
C
N
C
C
C
N
N
C
C
C
C
C
C
C
N
56ꢀ18
2.3ꢀ0.2
1760ꢀ260
41ꢀ7
>10,000
>10,000
n.d.
>10,000
n.d.
n.d.
>10,000
2100
n.d.
9600
n.d.
n.d.
2450ꢀ520
550ꢀ70
Values for d and m binding are from single determinations. n.d., Value not determined.
aValues for k binding are the arithmetical mean of at least three determinations ꢀ SEM.

G. Semple et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1141–1145
1143
this, and a variety of alternative procedures and found
that good yields for N-alkylation of 3-arylpyridin-2-
ones with amino alcohols could be achieved using Bu3P
and 1,10-azobis(N,N-dimethylformamide (TMAD) as
the catalytic dehydrating system, with N versus O
selectivities in general between 85:15 and 99:1. In addi-
tion, the combination of these reagents, and the pre-
sence of a basic amino group in our products allowed us
to apply a rapid purification procedure using simple fil-
tration, followed by ion-exchange on pre-packed SCX
columns, facilitating the deployment of this chemistry in
a parallel format. Having previously demonstrated in
principle, that the pyridone series had functional agonist
properties, we examined only the binding affinities of
the new compounds for opioid receptors at this stage.
After analysis by HPLC-MS, those mixtures that were
of >85% purity (i.e., <15% O-alkylated product) were
screened in the KORbinding assay at a single concen-
tration of either 1 or 10 mM. In a typical run, a batch of
40 parallel Mitsonobu reactions resulted in only four
examples where this criterion was not met [specifically
1-(2-hydroxy-3,4-dimethylbutyl] pyrrolidine in place of
2-pyrrolidino-1-arylethanols gave almost exclusive O-
alkylation). First estimates of the IC50 for most com-
pounds were made by determination of five-point dose–
response curves at this stage. Selected examples, chosen
to answer specific medicinal chemistry questions, were
then further purified to homogeneity (>96% purity,
leaving essentially no O-alkylated product) by parallel
preparative HPLC. Full dose–response curves were then
determined for binding to kappa, delta and mu opioid
receptors and the data provided in Table 2 represents
mean IC50 values for the fully purified material.
Other basic groups were tolerated however, with the 3-
hydroxy-1-pyrrolidine replacement being the most pro-
mising, and providing some useful increase in polar
surface area. Non-cyclic dialkyl amine groups in this
position (e.g., 14, R,R’’=Et) resulted in significant loss
in binding affinity (data not shown) as did incor-
poration of the less basic 1-imidazole moiety. In keeping
with our previous observations, the SARhere is similar
to that described for the classical ethylene diamine series
12
of KORagonists.
The majority of substituents we looked at for our initial
investigation of substitution on the Ar group, did not
increase the polar surface area of the molecules. How-
ever, some of the examples that did are also shown in
Table 2. Of these, the 3-carboxymethyl ether substituted
analogue 14i had disappointingly poor activity given
that the same substitution has been used to confer per-
ipheral selectivity in the classical series.13 Again, there
were no substituents tested which were better than the
parent compound (with the exception of chloro sub-
stituted analogues which increased lipophilicity-data not
shown), but the benzomorpholinone analogue 14j, was
only 7-fold poorer in binding affinity to the KORthan
its parent compound and was considerably less lipo-
phillic, giving us another useful pathway into more
complex structures.
Crucially, the Mitsunobu procedure could also be
applied to the bromopyridone 15 to provide inter-
mediates of the type 16. By the addition of a second
parallel step to incorporate the substituent R, using our
previously described Suzuki coupling methodology,1
followed by a second parallel ion-exchange purification
step, a wide variety of compounds of the type 14, could be
prepared in a matrix format whereby the SARof sub-
stituents at three different positions could be examined
Manipulation of the pyrrolidine group did not provide
any improvement in the binding affinities of the resul-
tant compounds when compared to the parent series.
Table 2.
No.
RAr
NR0R00
k-Binding IC50
(nMꢀSEM)a
d-Binding IC50
(nM)
m-Binding IC50
(nM)
2
14a (=4a)
14b
14c
14d
14e
14f
14g
14h
14i
Ph
Ph
Ph
Ph
Ph
Ph
Ph
1-Pyrrolidine
1-Pyrrolidine
1-Morpholine
1-Imidazole
1-Diazepine
56.3ꢀ18
8.7ꢀ3.8
700ꢀ400
>10,000
1400b
>10,000
>10,000
>10,000
n.d.
>10,000
>10,000
>10,000
>10,000
5600
>10,000
2716
>10,000
n.d.
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
4-CF3–Ph
4-CF3–Ph
4-CF3–Ph
4-CF3–Ph
Ph
4-CF3–Ph
4-CF3–Ph
4-CF3–Ph
4-CF3–Ph
1-(3-Hydroxy pyrrolidine)c
1-Pyrrolidine
1-Pyrrolidine
470ꢀ140
4-(NEt2)-Ph
4-MeO-Ph
3-(OCH2CO2H)-Phd
3080b
320ꢀ150
420ꢀ190
57ꢀ15
1-Pyrrolidine
1-Pyrrolidine
14j
>10,000
14k
14l
14m
14n
Ph–NH–
4-Me–Ph–NH–
4-CF3–Ph–NH–
4-Me–Ph–CH(CH3)NH–
—
Ph
Ph
Ph
Ph
—
1-Pyrrolidine
1-Pyrrolidine
1-Pyrrolidine
1-Pyrrolidine
—
>1,000
94ꢀ29
n.d.
n.d.
n.d.
>10,000
>10,000
250ꢀ24
>10,000
>10,000
>10,000
5800ꢀ1420
82ꢀ22
330ꢀ110
2.0ꢀ0.1
U-69,593
n.d., Value not determined.
aValues for k binding are the arithmetical mean of at least two determinations ꢀ SEM except where marked b, which are from single determinations.
All values for d and m binding are from single determinations.
cThe hydroxy group of the 3-hydroxy pyrrolidine was protected as the TBDMS ether prior to reaction with styrene oxide. The final compound was
isolated by treatment of the partially purified product (after SCX purification) with TFA in DCM for 24 h, followed by preparative HPLC.
dThe acid group was protected as the methyl ester during the synthesis. The final compound was isolated by treatment of the partially purified
product (after SCX purification) with LiOH in dioxan/water for 4 h, followed by preparative HPLC.

1144
G. Semple et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1141–1145
simultaneously. Furthermore, coupling of amines using
the palladium-mediated coupling conditions described
by Buchwald and coworkers14 allowed preparation of
amino substituted pyridones 14 (R=NH-aryl or NH-
alkyl), a new class of potential kappa agonists. Again,
purification was accomplished in parallel by ion
exchange as outlined above. Representative examples of
this latter series are shown in Table 2. In general the
compounds were poorer KORligands than the parent
series, but 4-substitution of the aryl ring in this case
again provided an improvement in binding affinity,
suggesting that they too bind in a manner similar to the
classical series, and that the receptor pocket can
accommodate substituents where the aryl ring is dis-
placed from the pyridone ring. It was even more sur-
prising to us that 14n, in which these two rings are
further separated, also retained weaker but significant
binding affinity for the KOR.
readily prepared in this way. Having demonstrated that
the Mitsunobu reaction of the amino alcohols with 3-
bromo-2-pyridone gave predominantly N-alkylation in
solution, we were pleased to find, following cleavage of
a small portion of the resin produced here, that the
same was true when the reaction was performed on
solid phase, albeit with some minor modifications to the
reaction conditions. Finally, Suzuki coupling of a com-
mercially available set of 80 boronic acid derivatives,
and acid catalysed cleavage of the final resin bound
product, provided the 3-aryl-2-pyridone series 22. The
overall synthesis closely mirrored the optimised solu-
tion-phase route, but had the advantage of removing
the two parallel purification stages required for the full
application of that method. After analysis of the cleaved
products by HPLC-MS (typical purity of mixtures of
diastereomers>85%, with the major impurities in each
case arising from O-alkylation in the Mitsunobu reac-
tion) and screening at 1 mM in the KORbinding assay
(68 of 80 examples showed>50% inhibition of ligand
binding at this concentration) selected examples were
purified to homogeneity by parallel preparative HPLC.
The data provided in Table 3, where we again report
only selected results mainly from the more hydrophilic
substituents, represents mean IC50 values for the fully
purified material.
At this stage, we opted to generate a small library of
compounds with a hydroxyl group on the pyrrolidine
ring. We already had an indication that this modifi-
cation could be tolerated and we firmly believed that the
SARof the pyridone series and classical KORagonist
series, where such an addition can often provide sig-
nificant improvements in potency,15 are very closely
related. However, as an alternative to protection of the
alcohol function with TBDMS with the addition of a
final deprotection step under our solution conditions
described above, we decided to investigate the use of a
solid phase approach, wherein the hydroxyl group was
protected in a similar manner, but by a resin-bound silyl
group (Fig. 3). Thus, 3-(S)-hydroxy pyrrolidine15 was
attached via the hydroxyl group to a Wang resin-bound
benzylpropyl silyl linker, by treatment of the resin with
the neat amine.
In general, it appeared that most substitutions on the
aryl ring had positive effects, although those sub-
stituents that provided the largest improvement in affi-
nity in the unsubstituted pyrrolidine series,1 were not
necessarily those that were most beneficial here. The
inclusion of oxygen containing groups was of clear
benefit in the case of 2-formyl, 3-acetyl or 3-hydroxy-
methyl, but had a detrimental effect when the group was
3-carboxy (20m). 3- or 4-Trifluoromethoxy substitutions
also gave good improvements in binding affinity. As
with the unsubstituted pyrrolidine series however,
incorporation of two substituents onto the aromatic
ring provided a further improvement in affinity for the
KOR. In particular, the 2-methoxy substituent, whilst
having only a modest effect on its own (20b), was able to
assist in providing good affinity for the KORin combi-
nation with additional substituents (20d–g) although
there was a notable reduction in receptor selectivity in a
couple of examples. A similar combination of sub-
stituent effects was seen with the 4-Fluoro analogue 20j,
the affinity of which was improved 4-fold by addition of a
3-methyl group (20k). The most potent compound of this
type was 20e, with an affinity for the KORcomparable to
that of the literature standard compound, U-69,593 in our
hands, and a significant reduction in lipophilicity com-
pared to the pyridone 4a. In addition, we were able to
confirm that the compound was a full agonist at the KOR
(ED50=32ꢀ9nM in the GTPgS binding assay8).
The resultant resin-bound amine could be heated with
neat styrene oxide in a similar manner to the analogous
solution phase reaction. Cleavage of a small portion of
the resin showed that the ring opening of styrene oxide
itself was completely regiospecific. Again, alkylation of
aryl bromoketones followed by reduction with sodium
borohydride was a more generally applicable route to
ensure the required regiochemistry of the resultant
amino-alcohol, and substituted analogues could be
Incorporation of either thiophene or fused aromatic
rings lead to a significant decrease in affinity in com-
parison to the most potent substituted phenyl analogues
(e.g., 20, R=2-thiophene showed only 31% inhibition
of radioligand binding @ 1 mM). Reasonable activity
could be rescued in this case though by incorporation of
a 5-acetyl substituent (20p).
Figure 3. Reagents and conditions: (i) neat amine, 80 ꢁC; (ii) styrene-
oxide, 80 ꢁC or (a) ArCOCH2Br, DIEA, rt; (b) NaBH4, MeOH, water,
rt; (iii) Bu3P, TMAD, THF, 18 h; (iv) ArB(OH)2, Pd(PPh3)4, dioxan,
NaHCO3, 80 ꢁC; (v) DCM, TFA.

G. Semple et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1141–1145
1145
Table 3.
No.
R
k-Binding IC50 (nMꢀSEM)a
d-Binding IC50 (nM)
m-Binding IC50 (nM)
20a (=14f)
20b
20c
20d
20e
20f
20g
20h
20i
20j
20k
20l
20m
20n
20o
Ph
2-OMe–Ph
3-NH2–Ph
2-OMe, 5-Cl–Ph
2-OMe, 5-F–Ph
2,6-(OMe)2–Ph
2,5-(Ome)2–Ph
4-OCF3–Ph
465ꢀ135
94ꢀ29
>10,000
>10,000
>10,000
>10,000
460ꢀ230
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
n.d.
>10,000
n.d.
2910
190ꢀ23
>10,000
3910
25ꢀ19
53ꢀ11
5.5ꢀ1.3
9.8ꢀ7.8
6.9ꢀ7.1
7.6ꢀ5.7
22ꢀ13
2220
>5000
>5000
>10,000
>5000
>5000
n.d.
3-OCF3–Ph
4-F–Ph
33ꢀ27
3-Me, 4-F–Ph
3-(CH2OH)–Ph
3-(CO2H)–Ph
4-(COCH3)–Ph
2-(CHO)–Ph
8.4ꢀ6.9
9.6ꢀ2.6
41% @ 1 mM
14ꢀ5.8
13ꢀ10
>10,000
>10,000
1750
n.d.
20p
38ꢀ11
>10,000
>10,000
250ꢀ24
U–69,593
—
2.0ꢀ0.1
5800ꢀ1420
n.d., Value not determined.
aValues for k binding are the arithmetical mean of at least two determinations ꢀ SEM except where marked b, which are from single determinations.
All values for d and m binding are from single determinations.
These data clearly demonstrate the advantage of run-
ning multiple parallel reactions in lead expansion, as the
SARwas far from predictable. Indeed data from the
unsubstituted pyrrolidine series inferred that the chloro
analogue, 20d would be most potent here, when in fact
it was one order of magnitude weaker in binding affinity
to the KORthen 20e.
5. Breukelman, S. P.; Meakins, G. D.; Roe, A. M. J. Chem.
Soc., Perkin Trans. 1 1985, 1627.
6. Kress, T. J. J. Org. Chem. 1985, 50, 3073.
7. All opioid receptor binding affinities were determined by
classical filter binding methods, measuring the displacement of
the appropriate radiolabelled ligand from membranes of
HEK293 cells overexpresing the requisite human opioid
receptor. Ligands used were [125I]-(D-Pro10)-Dynorphin A(1-
(k), [125I]-Enkephalin (m) and [125I]-Deltorphin (d).
8. Functional activity was determined using a radiolabelled
[35S]-GTPgS assay, again using membranes of HEK-293 cells
overexpressing the human kappa opioid receptor.
9. Shapiro, S. L.; Soloway, H.; Freedman, L. J. Am. Chem.
Soc. 1958, 80, 6060.
10. Pelletier, J. C.; Kincaid, S. Tetrahedron Lett. 2000, 41,
797. Barrett, A. G. M.; Roberts, R. S.; Schroder, J. Org. Lett.
2000, 2, 2999.
11. Comins, D. L.; Jianhua, G. Tetrahedron Lett. 1994, 35,
2819.
12. Costello, G. F.; James, R.; Shaw, J. S.; Slater, A. M.;
Stutchbury, N. C. J. J. Med. Chem. 1991, 34, 181.
13. Shaw, J. S.; Carroll, J. A.; Alcock, P.; Main, B. G. Br. J.
Pharmacol. 1989, 96, 986.
14. Wolfe, J. P.; Wagaw, S.; Buchwald, S. L. J. Am. Chem.
Soc. 1996, 118, 7215. In a typical procedure, the bromopyr-
idone derivative 16, the required amine (0.95 equiv) and the
catalyst components, Pd2(dba)3 and R-BINAP, were placed in
each well of a dry Robbins block. Dry dioxan was added and
the block sealed and agitated under argon for 18 h at 90 ꢁC.
The reaction mixtures were filtered thorough silica plugs
(eluting with 0.2% ammonia in EtOAc), and then purified by
ion exchange on pre-packed SCX columns.
In summary, we have prepared a number of new analo-
gues of our previously described 3-aryl pyridone KOR
agonists. We have concentrated our efforts on changes
to the structure aimed at reducing the lipohilicity and
increasing the polar surface area of the resultant mole-
cules. To expand the range of accessible derivatives we
have developed parallel synthetic procedures, both in
solution and on solid phase, making use of the well
known and versatile Mitsunobu, Suzuki and Buchwald
reactions. Although we have herein reported on the
individual chemistry steps developed, the serial use of
such parallel methods clearly provides a powerful tool
for the exploration of the SARof the resultant KOR
agonist series. Using these techniques, we have been
able to uncover interesting new structures that have
served to direct our further work in this area and these
efforts will be reported in due course.
References and Notes
15. For example our own unpublished data for asimadoline16
and its des-hydroxy analogue showed that the former is
around 10-fold more potent in both binding and function at
the KOR. In addition it is the S-isomer at the hydroxyl posi-
tion which is most active.
1. Semple, G.; Andersson, B. M.; Chhajlani, V.; Georgsson,
J.; Johansson, M.; Lindschoten, M.; Ponten, F.; Rosenquist,
´
A.; Sorenson, H.; Swanson, L.; Swanson, M. Bioorg. Med.
¨
Chem. Lett. 2002, 12, 197.
2. Clark, D. E. J. Pharm. Sci. 1999, 88, 815.
3. Rimoy, G. H.; Wright, D. M.; Bhaskar, N. K.; Rubin, P. C.
Eur. J. Clin. Pharmacol. 1994, 46, 203.
16. Barber, A.; Bartoszyk, G. D.; Bender, H. M.; Gotts-
chlich, R.; Greiner, H. E.; Harting, J.; Mauler, F.; Minck,
K. O.; Murray, R. D.; Simon, M. Br. J. Pharmacol. 1994,
113, 1317.
4. Jones, R. G. J. Am. Chem. Soc. 1949, 71, 78.