Dopaminergic 7-aminotetrahydroindolizines: ex-chiral pool synthesis and preferential D3 receptor binding.

Article abstract of DOI:10.1016/S0960-894X(01)00564-9

Starting from both isomers of enantiopure asparagine, heterocyclic bioisosteres of the preferential dopamine D3 receptor agonist (R)-7-OH-DPAT were investigated when SAR studies led to the 3-formyl substituted aminoindolizine (S)-1e (FAUC 54) displaying a K(i) value of 6.0 nM for the high affinity D3 binding site. In contrast, D3 affinity of the enantiomer (R)-1e was 300 fold lower.

Full text of DOI:10.1016/S0960-894X(01)00564-9

Bioorganic & Medicinal Chemistry Letters 11 (2001) 2863–2866
Dopaminergic 7-Aminotetrahydroindolizines: Ex-Chiral Pool
Synthesis and Preferential D3 Receptor Binding
Thomas Lehmann, Harald Hubner and Peter Gmeiner*
Department of Medicinal Chemistry, Emil Fischer Center, Friedrich-Alexander University, Schuhstraße 19,
D-91052 Erlangen, Germany
Received 12 June 2001; revised 2 August 2001; accepted 10 August 2001
Abstract—Startingfrom both isomers of enantiopure asparagine, heterocyclic bioisosteres of the preferential dopamine D3 receptor
agonist (R)-7-OH-DPAT were investigated when SAR studies led to the 3-formyl substituted aminoindolizine (S)-1e (FAUC 54)
displayinga Ki value of 6.0 nM for the high affinity D3 binding site. In contrast, D3 affinity of the enantiomer (R)-1e was 300 fold
lower. # 2001 Elsevier Science Ltd. All rights reserved.
The existence of two families of dopamine receptors has
been established by recent advances in classical phar-
macology and molecular biology.1 It is generally accep-
ted that the D2, D3, and D4 subtypes belongto the D2-
like family, while the D1-like family comprises D1 and
D5 receptors.2 There is strongevidence that D2 and D3
receptors exist postsynaptically and also as auto-
receptors controllingdopamine synthesis, release and
neuronal firing.1 The D3 receptor3 appears an impor-
tant target for the development of drug candidates since
it is selectively expressed in the brain limbic system,
which is thought to be involved in mood and cognitive
In this communication, we report on EPC synthesis
and subtype selective dopamine receptor bindingof
hydroxymethyl and formyl substituted 7-dipropylamino-
indolizines of type (R)-1 and (S)-1. In order to estimate
the influence of the hydrogen bond acceptor, 1-, 2- or 3-
methyl derivatives were evaluated, too.
4
disturbances as well as drugdependence. Most of the
recent SAR studies on preferential dopamine D3 recep-
tor agonists are based on (R)-7-OH-dipropylamino-
tetralin as a lead compound.5 In recent reports, we
described novel DPAT regioisomers6 and heterocyclic
bioisosteres includingenantiomerically pure dipropyl-
aminotetrahydroindolizines revealingautoreceptor
activity.7 Structural alignment of (R)-7-OH-DPAT and
tetrahydroindolizines led us to the assumption that a
hydrogen bond acceptor connected with one atom dis-
tance to the aromatic moiety of the aminoindolizine
framework could simulate the sp3 oxygen of (R)-7-OH-
DPAT. Dependingon the absolute stereochemistry of
the heterocyclic bioisostere, spatial overlap can
result from substitution in the positions 1 or 2
(superimposition A) or 3 (superimposition B).
The synthesis of the target compounds bearing (S)-con-
figuration relied on the employment of natural aspar-
agine as a chiral building block (Scheme 1). For the
preparation of the 1- and 2-substituted indolizines we
approached the formylpyrrole 3a as a central inter-
mediate. In practice, reductive N,N-dibenzylation and
subsequent treatment with boraneꢀ THF furnished the
primary amine 2 that could be reacted with 3-formyl-
2,5-dimethoxytetrahydrofuran.8 UsingHOAc as a sol-
vent for the Paal–Knorr reaction, the primary alcohol
3a was formed alongwith the acetate 3b that was
converted to 3a under aqueous basic conditions.
Transformation of the carbaldehyde function into a
methyl group was accomplished by Wolff–Kishner
reduction resultingin formation of the cyclization
*Correspondingauthor. Tel.: +49-9131-8529383; fax: +49-9131-
8522585; e-mail: gmeiner@pharmazie.uni-erlangen.de
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00564-9

2864
T. Lehmann et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2863–2866
AgOTf was envisaged to form a reactive triflate giving
access to the indolizines 7,8. In fact, bromide formation
was observed in 71% yield. Careful NMR studies
clearly proved the structure of the secondary bromide
1010 which was obviously formed by rearrangement via
an aziridinium intermediate. Similar reactions have been
observed and described by us in detail.11 Anticipating
the reversibility of the migration, we subjected the bro-
mide 10 to AgOTf. In fact, formation of the indolizines
11 and 12 as a 1:2 mixture of regioisomers was
observed. Subsequent deprotection led to the carbalde-
hydes 7 and 8 which were readily separable by flash
chromatography.12 Employingone-pot conditions, the
reaction sequence afforded a 39% yield of the main
regioisomer 8 and 20% of the 2-substituted isomer 7.
The optical integrity of the synthesis was proved by
couplingof the carbaldehyde 8 with (R)- and (S)-ala-
nine methyl ester under reductive conditions
Scheme 1. (a) HOAc/NaOAc, 70ꢁ, 75 min (3a: 48%; 3b: 10%); (b)
K2CO3, H2O, rt, 24 h (100%).
precursor 4 in 70% yield (Scheme 2). In accordance to
our recently described observations on stereoelec-
tronically controlled cationic cyclizations,9 formation of
a six-membered ringwas induced by activation of the
terminal alcohol group. Thus, treatment of 4 with tri-
fluoromethanesulfonic anhydride gave the 2-methyl
indolizine 5 and the 1-methyl substituted heterocycle 6
as a 1:10 mixture of regioisomers reflecting the influence
of the electron donatingproperies of the methyl group
onto the site-selectivity of the electrophilic attack. 5-Exo
attack of the intermediately formed aziridinium salt
could not be detected.
1
(NaBH3CN, MeOH) when H and 13C NMR investi-
gations of the resulting secondary amines indicated an
isomeric excess >95%. Exchange of the N,N-dibenzyl
protectinggroups by the pharmacophoric N,N-dipropyl
substitution pattern was intended by hydrogenolysis
and subsequent reductive alkylation (Scheme 3). Thus,
the methyl substituted dibenzylaminoindolizines could
be readily transformed into the primary amines upon
hydrogenation in presence of Pearlman’s catalyst. Sub-
sequent treatment with an excess of propionaldehyde
and NaBH3CN afforded the final products (S)-1a and
(S)-1b. Subjectingthe carbaldehydes 7 and 8 to identical
hydrogenation conditions resulted in debenzylation
and reduction of the formyl group into a mixture of
hydroxymethyl and methyl derivatives when careful
optimization of the reaction time and the solvent
accomplished the 1- and 2-hydroxymethyl substituted
aminoindolizines in satisfactory yield. Finally, reductive
dipropylation was performed to give the target
compounds (S)-1c and (S)-1d.
Due to the electron acceptingeffect of the carbaldehyde
function, Tf2O activation of the pyrrole 3a did not
result in cationic p-cyclization. In order to increase the
electron density of the aromatic region, the protected
derivative 9, which was readily available by acid cata-
lyzed acetalization, was selected as a cyclization pre-
cursor. However, Tf2O induced sulfonylation of 9
caused decomposition that we put down to impurities of
TfOH beingable to activate the acetal function for side
reactions. Seekingfor a smooth alternative, we tried to
transform the alcohol function into a bromide using
neutral Appel conditions. Subsequent treatment with
Modification of the 3-position of the indolizine frame-
work was planned by regiocontrolled formylation of
the unsubstituted N,N-7-dipropylaminoindolizine 13
(Scheme 4) which was synthesized from the asparagine
derived buildingblock 2 applyinga previously described
reaction sequence.8 Subjecting 13 to Vilsmeier–Haack
reaction conditions gave preferential formylation in
position 3 when 42% of pure carbaldehyde (S)-1e was
isolated.13 As a minor regioisomer, the 1-formyl deriva-
tive (S)-1f was obtained (5% yield). Reductive mod-
ification of the 3-formylindolizine (S)-1e was
Scheme 2. (a) Tf2O, CH2Cl2, rt, 48 h (5: 5%, 6: 50%); (b) hydrazine
hydrate, diethylene glycol, 130ꢁ, 1 h; +KOH, 170ꢁ, 4 h (70%); (c) 2,2,-
.
dimethylpropan-1,3-diol, TsOH H2O, benzene, reflux, 4 h (85%); (d)
CBr4, PPh3, CH2Cl2, rt, 16 h (71%); (e) AgOTf, CH2Cl2, 0ꢁ, 75 min;
(f) CF3COOH, MeOH, 0 ꢁC, 75 min (7: 20%, 8: 39%, based on 9).
Scheme 3. (a) (1) H2, Pd(OH)2/C, MeOH–EtOAc 1:1, rt, 6 h; (2) pro-
pionaldehyde, MeOH, 0ꢁ, 90 min for 5, 6, 180 min for 7, 8 (19–41%).

T. Lehmann et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2863–2866
2865
DPAT as an internal standard, similar affinities could
be determined when compared to cloned human D2
receptors stable expressed in CHO cells (Kihigh=10 for
D2long and Kihigh=3.3 for D2short). The investigation of
the D3 receptor bindingof the test compounds indi-
cated Ki values ranging only in micromolar or sub-
micromolar concentrations.
As expected, the
methylindolizines 1a,b,h showed bindingproperties that
were similar to those of the unsubstituted 7-aminoindoli-
zine 13. It is interestingto note, that the ( R)-enantio-
mers of the 2- and 1-hydroxymethyl substituted ligands
1c and 1d displayed higher D3 (and also D2) affinity
than its optical antipodes. This is contrary to the 3-
hydroxymethyl derivative 1g revealingsnigificantly
higher binding of the (S)-enantiomer. Comparison of
the 1-formylindolizine 1f with the 3-formyl regioisomer
1e shows the same relationship clearly corroboratingan
analogy of binding modes as indicated in the schematic
superimpositions A and B. (S)-1e (FAUC 54) exhibited
the best bindingproperties and a Hill coefficient
nH=À0.68 that was identical to that we obtained for
the D3 receptor agonist 7-OH-DPAT. The biphasic
bindingcurves clearly indicated agonist properties for
(S)-1e when differentiation between a high affinity site
representingthe ternary receptor G-protein complex
and a low affinity bindingsite is typical. Detailed ana-
lysis of the bindingcurves of ( S)-1e provided a
Kihigh=6.0Æ1.3 nM and a Kilow=150Æ20 nM.
Scheme 4. (a) DMF, POCl3, 0ꢁ to rt, 4 h [(S)-1e: 42%, (S)-1f: 5%]. (b)
NaBH4, isopropanol, rt, 19 h (89%); (c) LiAlH4, THF, 0ꢁ to reflux, 48
h (44%).
Table 1. Receptor bindingdata for the target compounds ( R)-1 and
(S)-1 compared to the D2 autoreceptor agonist 13 and the D3 agonist
(R)-7-OH-DPAT employingbovine D1 and D2 and human D3
receptors. Ki values [nM] are given based on the means of 2–8 experi-
ments each performed in triplicate, the results of which did not vary
more than 25%, except for (R)1b (35%). For (S)-1e and (R)-7-OH-
DPAT, a clear differentiation between a high affinity binding site
(representingthe ternary complex) and a low affinity bindingsite was
observed for D2 and D3 when labeled with [3H]spiperone
Compd
R
bD1
[3H]
SCH23390
bD2high
[3H]
pram.
bD2
[3H]spip.
hD3
[3H]spip.
(R)-1a
(S)-1a
(R)-1b
(S)-1b
(R)-1c
(S)-1c
(R)-1d
(S)-1d
(R)-1e
(S)-1e
2-CH3
2-CH3
1-CH3
1-CH3
2-CH2OH >100,000
2-CH2OH >100,000
1-CH2OH
1-CH2OH >100,000
3-CHO
3-CHO
83,000
44,000
71,000
68,000
800
640
1000
380
1000
13,000
200
8000
1700
21
17,000
9600
12,000
4000
1300
770
2100
750
1400
8600
330
9700
34,000
>100,000
6100
55,000
21,000
(high) 90
In conclusion, the ability of the hydroxyl function of
(R)-7-OH DPAT interactingwith serine residues in
TM5 of the D3 receptor protein as an H-bond acceptor
may be best adopted by a formyl group when positioned
at C3 of the (S)-7-dipropylaminotetrahydroindolizine
core structure. Based on FAUC 54 [(S)-1e], further SAR
studies, functional experiments and investigations of the
bioactive conformation will be published without delay.
100,000
47,000
26,000
1800
(high) 6,0
(low) 6500 (low) 150
(R)-1f
(S)-1f
(R)-1g
(S)-1g
(R)-1h
(S)-1h
13
1-CHO
1-CHO
3-CH2OH >100,000
3-CH2OH
3-CH3
3-CH3
H
80,000
>100,000
380
1600
4700
87
430
150
150
3,5
12,000
5900
37,000
5500
1300
3800
4100
n.d.
510
3600
7400
420
13,000
12,000
29,000
590
430
560
(high) 0,8
(low) 9,5
>100,000
n.d.
Acknowledgements
(R)-7-OH-DPAT
We thank Dr. J.-C. Schwartz and Dr. P. Sokoloff
(INSERM, Paris) for providingD3 receptor expressing
cell lines.
accomplished employingthe complex metal hydrides
NaBH4 and LiAlH4 to afford the alcohol (S)-1g and the
methyl derivative (S)-1h, respectively.
References and Notes
Employingthe identical procedures, we synthesized the
optical antipodes (R)-1a–h startingfrom unnatural ( R)-
asparagine.
1. Neve, K. A., Neve, R. L. Eds.; The Dopamine Receptors.
Humana Press: Totowa, New Jersey, 1997.
2. Kebabian, J. W.; Tarazi, F. I.; Kula, N. S.; Baldessarini,
R. J. Drug Discovery Today 1997, 2, 333.
3. Sokoloff, P.; Giros, P.; Martres, M.-P.; Bouthenet, M.-L.;
Schwartz, J.-C. Nature 1990, 347, 146.
´
4. For review, see: Sokoloff, P.; Griffon, N.; Sautel, F.; Lev-
esque, D.; Pilon, C.; Schwartz, J.-C.; Ridray, S.; Diaz, J.;
Simon, P.; Costentin, J.; Mann, A.; Wermuth, C. G.; Caine, S.
B.; Parsons, L. H.; Koob, G. F. Biomed. Health Res. 1998, 19
(Dopamine Receptor Subtypes), 118.
5. Gonzalez, A. M.; Sibley, D. R. Eur. J. Pharmacol. 1995,
272, R1. For examples, see: Glase, S. A.; Corbin, A. E.;
Pugsley, T. A.; Heffner, T. G.; Wise, L. D. J. Med. Chem.
The final products of type 1 were evaluated in vitro for
their abilities to displace [3H]spiperone from cloned
human dopamine D3 receptors14 beingstably expressed
in CHO cells (Table 1).15 D1 and D2 affinities were
determined by employingbovine striatal membrane
preparations and the D1 selective antagonists [3H]SCH
23390 and [3H]spiperone, respectively. Additionally, the
selective dopamine D2 autoreceptor agonist [3H]prami-
pexole was utilized for competition experiments at the
high affinity binding site.16,17 Employing( R)-7-OH

2866
T. Lehmann et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2863–2866
23
80.05, H 6.88, N 7.99. 8: [a]D =À168.0ꢁ (c=1.0, CHCl3); H
NMR (CDCl3, 360 MHz): d (ppm)=2.05 (dddd, J=13.0, 12.5,
12.5, 5.5 Hz, 1H, H-6ax), 2.17–2.25 (m, 1H, H-6eq), 3.03 (dd,
J=17.0, 11.5 Hz, 1H, H-8ax), 3.14 (dddd, J=12.5, 11.5, 5.0,
2.5 Hz, 1H, H-7), 3.47 (ddd, J=17.0, 5.0, 1.5 Hz, 1H, H-8eq), 3.68
(d, J=14.0 Hz, 2H, NCH2ar), 3.78 (d, J=14.0 Hz, 2H, NCH2ar),
3.83 (ddd, J=12.5, 12.5, 4.5 Hz, 1H, H-5ax), 4.10 (ddd, J=12.5,
5.5, 2.0 Hz, 1H, H-5eq), 6.41 (d, J=3.0 Hz, 1H, H-3), 6.53 (d,
J=3.0 Hz, 1H, H-2), 7.20–7.25 (m, 2H, p-ar), 7.27–7.33 (m, 4H,
m-ar), 7.39 (m, 4H, o-ar), 9.78 (s, 1H, CHO). HRMS (EI) calcd
for C23H24N2O (M+): 344.1888; found: 344.1882.
1
1995, 38, 3132. Van Vliet, L. A.; Tepper, P. G.; Dijkstra, D.;
Damsma, G.; Wikstrom, H.; Pugsley, T. A.; Akunne, H. C.;
Heffner, T. G.; Glase, S. A.; Wise, L. D. J. Med. Chem. 1996,
39, 4233. Murray, P. J.; Helden, R. M.; Johnson, M. R.;
Robertson, G. M.; Scopes, D. I. C.; Stokes, M.; Wadman, S.;
Whitehead, J. W. F.; Hayes, A. G.; Kilpatrick, G. J.; Large,
C.; Stubbs, C. M.; Turpin, M. P. Bioorg. Med. Chem. Lett.
1996, 6, 403. Dubuffet, T.; Newman-Tancredi, A.; Cussac, D.;
Audinot, V.; Loutz, A.; Millan, M. J.; Lavielle, G. Bioorg.
Med. Chem. Lett. 1999, 9, 2059.
6. Gmeiner, P.; Hummel, E.; Mierau, J. Bioorg. Med. Chem.
Lett. 1994, 4, 2589.
7. Gmeiner, P.; Mierau, J.; Hofner, G. Arch. Pharm. (Wein-
heim) 1992, 325, 57.
13. Analytical data: (S)-1e: [a]2D3=À17.3ꢁ (c=0.1, CHCl3); 1H
NMR (CDCl3, 360 MHz): d (ppm)=0.88 (t, J=7.3 Hz, 6H,
NCH2CH2CH3), 1.45 (sext., J=7.3 Hz, 4H, NCH2CH2CH3),
1.86 (dddd, J=13.0, 12.1, 11.9, 5.5 Hz, 1H, H-6ax), 2.07–2.16
(m, 1H, H-6eq), 2.40–2.50 (m, 4H, NCH2CH2CH3), 2.72 (dd,
J=16.1, 10.6 Hz, 1H, H-8ax), 2.98 (ddd, J=16.1, 4.7, 2.1 Hz,
1H, H-8eq), 3.05 (dddd, J=11.9, 10.6, 4.7, 2.8 Hz, 1H, H-7),
4.03 (ddd, J=13.5, 13.0, 4.8 Hz, 1H, H-5ax), 4.84 (ddd,
J=13.5, 5.5, 2.4 Hz, 1H, H-5eq), 5.98 (d, J=3.9 Hz, 1H, H-1),
6.87 (d, J=3.9 Hz, 1H, H-2), 9.40 (s, 1H, CHO). Anal. calcd
for C15H24N2O (248.37): C 72.54, H 9.74, N 11.28; found: C
72.42, H 9.77, N 11.21.
8. Gmeiner, P.; Lerche, H. Heterocycles 1990, 31, 9.
9. Lehmann, T.; Gmeiner, P. Heterocycles 2000, 53, 1371.
10. Analytical data: 10: [a]2D3=+21.5ꢁ (c=1.0, CHCl3); 1H
NMR (CDCl3, 200 MHz): d (ppm)=0.76 (s, 3H, CH3), 1.25 (s,
3H, CH3), 1.73 (dddd, J=14.9, 8.3, 8.1, 2.7 Hz, 1H, H-3), 2.35
(dddd, J=14.9, 10.2, 5.6, 4.7 Hz, 1H, H-3), 2.66 (dd, J=13.0,
9.0 Hz, 1H, H-1), 2.76 (dd, J=13.0, 6.0 Hz, 1H, H-1), 3.47 (d,
J=13.4 Hz, 2H, NCH2ar), 3.57 (d, J=11.0 Hz, 2H, CH2O),
3.60 (d, J=13.4 Hz, 2H, NCH2ar), 3.68 (d, J=11.0 Hz, 2H,
CH2O), 3.69–3.83 (m, 1H, H-2), 3.80–3.93 (m, 2H, H-4), 5.35
(s, 1H, OCH), 6.20 (dd, J=2.5, 1.8 Hz, 1H, H-4-pyrrole), 6.52
(dd, J=2.5, 2.5 Hz, 1H, H-5-pyrrole), 6.73 (dd, J=2.5, 1.8 Hz,
1H, H-2-pyrrole), 7.18–7.42 (m, 10H, ar). Anal. calcd for
C28H35N2O2Br: C 65.75, H 6.90, N 5.48; Found: C 65.65, H
6.96, N 5.45.
14. Sokoloff, P.; Andrieux, M.; Besancon, R.; Pilon, C.;
Martres, M.-P.; Giros, B.; Schwartz, J.-C. Eur. J. Pharmacol.
1992, 225, 331.
15. Receptor bindingstudies were carried out as described in
ref 16. In brief, competition experiments with the bovine
dopamine receptors were run with striatal membranes at a
final protein concentration of 100 mg/mL and the radioligand
[3H]SCH 23390 at 0.3 nM (Kd=0.27 nM) for D1 and with 90
mg/mL and the antagonist [3H]spiperone at 0.5 nM (Kd=0.65–
0.85 nM) for D2. Competition experiments specifically label-
ingthe high affinity bindingsite of the bovine D2 receptor
were done with the agonist [3H]pramipexole at 0.5 nM
(Kd=1.2–2.8 nM) and a protein concentration of 200 mg/mL.
Preparations of membranes from CHO cells expressingthe
human D3 receptor (30–70 mg/mL) were employed with
[3H]spiperone at 0.5 nM (Kd=0.1–0.4 nM).
11. Gmeiner, P.; Junge, D.; Kartner, A. J. Org. Chem. 1994,
59, 6766. Weber, K.; Kuklinkski, S.; Gmeiner, P. Org. Lett.
2000, 2, 647.
12. Analytical data: 7: [a]2D3=À132.8ꢁ (c=1.0, CHCl3); 1H
NMR (CDCl3, 360 MHz): d (ppm)=2.00 (dddd, J=12.0, 12.0,
11.5, 5.5 Hz, 1H, H-6ax), 2.16–2.24 (m, 1H, H-6eq), 2.83 (ddd,
J=15.5, 11.5, 1.0 Hz, 1H, H-8ax), 3.03 (ddd, J=15.5, 5.0, 1.0
Hz, 1H, H-8eq), 3.10 (dddd, J=11.5, 11.5, 5.0, 2.5 Hz, 1H, H-
7), 3.67 (d, J=14.0 Hz, 2H, NCH2ar), 3.74 (d, J=14.0 Hz,
2H, NCH2ar), 3.82 (ddd, J=12.5, 12.0, 4.5 Hz, 1H, H-5ax),
4.14 (ddd, J=12.5, 5.5, 2.0 Hz, 1H, H-5eq), 6.28 (brs, 1H, H-
1), 7.08 (d, J=1.5 Hz, 1H, H-3), 7.19–7.25 (m, 2H, p-ar), 7.27–
7.33 (m, 4H, m-ar), 7.38 (m, 4H, o-ar), 9.65 (s, 1H, CHO).
Anal. calcd for C23H24N2O: C 80.20, H 7.02, N 8.13; found: C
16. Ohnmacht, U.; Trankle, C.; Mohr, K.; Gmeiner, P. Phar-
mazie 1999, 54, 294.
17. Hubner, H.; Haubmann, C.; Utz, W.; Gmeiner, P. J. Med.
Chem. 2000, 43, 756.