Synthesis of a heterocyclic amine and acid receptor

Article abstract of DOI:10.1016/S0040-4020(00)00806-1

The synthesis of a molecule able to bind both with an amine and a carboxylic acid is described. This new host consists in a pyrroloisoquinoline structure possessing a side ann with an acetamidopyridine moiety. The pyrroloisoquinoline part was obtained after improvement of the previously published route involving regioselective reduction of appropriate 3-arylalkylaminosuccinimides followed by ring closure using iminium chemistry. The tricyclic structure was then functionalized by cross-coupling reaction under Pd⁰ catalysis of its triflate derivative with the trimethylsilyl ketene acetal of methyl acetate. The so-obtained ester was then reacted with 2-amino-4,6-dimethylpyridine leading, to the targeted host. The association constants of this latter compound with some carboxylic acids and amines were determined by H NMR titration. NOESY experiments and molecular modeling calculations were performed in order to precise the actual host-guest interactions. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4020(00)00806-1

TETRAHEDRON
Pergamon
Tetrahedron 56 (2000) 8679±8688
Synthesis of a Heterocyclic Amine and Acid Receptor
*
Jean-Fran cË ois Bri eÁ re, Georges Dupas, Guy Qu e guiner and Jean Bourguignon
Laboratoire de Chimie Organique Fine et H e t e rocyclique de l'IRCOF, Institut National des Sciences Appliqu e es de Rouen,
6131 Mont Saint Aignan Cedex, France
7
Received 9 March 2000; revised 31 August 2000; accepted 8 September 2000
AbstractÐThe synthesis of a molecule able to bind both with an amine and a carboxylic acid is described. This new host consists in a
pyrroloisoquinoline structure possessing a side arm with an acetamidopyridine moiety. The pyrroloisoquinoline part was obtained after
improvement of the previously published route involving regioselective reduction of appropriate 3-arylalkylaminosuccinimides followed by
0
ring closure using iminium chemistry. The tricyclic structure was then functionalized by cross-coupling reaction under Pd catalysis of its
tri¯ate derivative with the trimethylsilyl ketene acetal of methyl acetate. The so-obtained ester was then reacted with 2-amino-4,6-dimethyl-
pyridine leading to the targeted host. The association constants of this latter compound with some carboxylic acids and amines were
determined by H NMR titration. NOESY experiments and molecular modeling calculations were performed in order to precise the actual
1
host±guest interactions. q 2000 Elsevier Science Ltd. All rights reserved.
Introduction
with the two N±H bonds would stabilize the transition
state of the reaction between an acid and an amine. We
decided then to synthesize this new heterocyclic system
using two different methodologies. We have published
recently these two routes, the key step of the ®rst one
being regioselective reduction of appropriate aminosuccini-
In a previous paper, we described the design and synthesis
of a heterocyclic molecule able to bind with various
amines. This host combined an isoquinoline structure
possessing a lone pair for hydrogen bonding with a N±H
group and a ¯exible arm terminated by an acetamido-
pyridine moiety ensuring the second binding point
1
2
mides and subsequent ring closure (pathway A) whereas
the second consisted in construction of the pyrroloisoquino-
line system followed by oxidation leading to the targeted
lactam (pathway B). The more convenient route was the
(
Scheme 1). We found it interesting to design also a
3
compound able to bind both with an acid and an amine.
For this purpose, we added a ®ve-membered ring possessing
a lactam group in order to allow supplementary hydrogen
bonding with a carboxylic acid.
®rst one and it afforded the tricyclic derivatives 4a±c
(Scheme 2). The aim of this paper is to achieve the synthesis
of the host molecules 2a,b and to study their binding proper-
ties with acids and amines. The main problems were the
introduction of the methyl group on the isoquinoline
nitrogen and the functional group interconversion leading
to the tricyclic ring system substituted by the side chain.
The resulting molecule 2a possesses two heterocyclic
nitrogen lone pairs. The CPK model of this host suggested
that these two heterocyclic nitrogen lone pairs combined
Scheme 1.
Keywords: molecular recognition; binding constants; molecular modeling; cross-coupling.
*
Corresponding author. Tel.: 133-2-35522474; fax: 133-2-35522962; e-mail: georges.dupas@insa-rouen.fr
0040±4020/00/$ - see front matter q 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(00)00806-1

8
680
J.-F. Bri eÁ re et al. / Tetrahedron 56 (2000) 8679±8688
Scheme 2.
Results and Discussion
be used without puri®cation. Starting from the appropriate
succinimide, this synthesis required four steps and three
puri®cations by ¯ash chromatography. We tried then to
improve the overall yield by decreasing the number of
puri®cations. The 4-aminosuccinimide 8 was thus regio-
selectively reduced under recently published conditions
Synthesis of the tricyclic phenol
2
We ®rst tried to introduce the methyl group by reductive
amination. This kind of reaction is usually carried out under
acidic conditions and classical conditions of the literature
led only to low yields of N±Me compound besides large
amounts of degradation products (NaBH CN/(HCHO) /
(
sodium borohydride in a mixture of methanol and dichloro-
7
methane). Under these conditions, the work-up was con-
siderably simpli®ed. The crude hydroxylactam 3d was not
converted into the methoxy derivative but cyclized under
p-toluenesulfonic acid catalysis leading to the benzyloxy-
pyrroloisoquinoline 4c in a 30% yield for the two steps
starting from 8 (Scheme 4).
3
n
4
CH COOH, NaBH /CF COOH/(HCHO) or Zn(BH4)2/
3
4
3
n
5
(
acidic medium and we decided to use the neutral conditions
HCHO) ). The tricyclic structure was rather unstable in
n
6
of Charles et al. with paraformaldehyde in re¯uxing
methanol (Scheme 3).
The benzyloxy derivative 4c was reacted with paraformal-
dehyde in methanol and the intermediate carbinolamine
subsequently reduced with hydrogen under palladium
dihydroxide catalysis. The Pearlman's catalyst was reactive
enough to cleave in the same pot the benzyloxy group. The
required phenol 7 was thus obtained in quantitative yield
starting from 4c and was again pure enough to be used
without further puri®cation. So, the tricyclic part was
obtained from 8 within three steps and one puri®cation in
a nice one-pot procedure.
The heterocyclic amines 4a±c afforded the intermediate
ethers 5a±c isolated as yellow solids. The compounds
5a±c were in situ reduced in nice yields with hydrogen
under Raney nickel catalysis.
Various systems were tried in order to cleave the methoxy
group of 6a,b (NaI/ClSiMe , Cl Si/NaI, BCl , BBr ). What-
3
4
3
3
ever the conditions used, only degradation products were
formed and it must be pointed out that when the methoxy
group was at C-9, no reaction occurred. Alternatively, the
benzyloxy derivative 6c underwent clean cleavage
promoted by Pearlman catalyst leading to the desired phenol
Functionalization of the benzene ring
7. The crude phenol obtained after ®ltration of the catalyst
and subsequent removal of the solvent was pure enough to
In order to introduce the side chain with a coupling reaction,
we selected the tri¯ate group offering large possibilities of
Scheme 3.

J.-F. Bri eÁ re et al. / Tetrahedron 56 (2000) 8679±8688
8681
Scheme 4. One-pot synthesis of the tricyclic structure.
8
functionalization of an aromatic ring. We tried ®rst
the required tri¯ate was obtained in a 84% yield in DMF
with triethylamine as a base.
9
classical conditions with tri¯ic anhydride but only degra-
dation products were observed probably due to the acidity of
the medium. Alternatively, when pyridine was used as a
solvent, the tri¯ate was obtained in a 61% yield but it
could not be puri®ed. In order to use more basic conditions,
we tried ditri¯imide because it is necessary to activate ®rst
the phenol as a phenolate. The tri¯ate 9 was easily isolated
This tri¯ate was reacted with allyltributylstannane under
10
Stille conditions and led to the allyl derivative in a 75%
yield (Scheme 5). This allyl group was subsequently
oxidized with ozone and a mixture of the acetal 11 and
the ester 12b (with a small amount of the acid 12a) was
isolated. This reaction was not optimized because we
found it better to use other coupling reactions. Whereas
(LHMDS, THF, rt) but the yields were not reproducible
owing to the low solubility of the phenolate in THF. Finally,
2 3 2 2 3 2 3 2 3 3 4
Scheme 5. (i) 2.2 PhN(Tf) /Et N, DMF/rt. (ii) CH vCH±CH ±SnBu /LiCl, DMF/PdCl (PPh ) , 958C/4 h. (iii) O /HCl/MeOH. (iv) 13b/Pd(PPh ) /AcOLi/
THF re¯ux.
Scheme 6.

8
682
J.-F. Bri eÁ re et al. / Tetrahedron 56 (2000) 8679±8688
(
7.85 ppm for H3 instead of 6.69 ppm for H5 , numbering of
0 0
Scheme 7) of the pyridine moiety indicating that this proton
is close to the carbonyl group. This shift allowed us to
assume that the acetylaminopyridine preferred a trans
conformation. Secondly, the chemical shift of the proton
of the lactam group appeared to be very dependent on the
concentration. This concentration induced shift could be the
consequence of a dimerization process. When a dilution
17
titration was performed over the range of concentrations
2
1
studied (about 0.02 M), a value of K ˆ12 M for the
Scheme 7. Comprehensive numbering of host 2b for the discussion.
d
dimerization was measured. This low value was very
interesting because a dimerization process could lead to
smaller values for the association constants with guests.
However, with binding experiments performed at 0.02 M,
this value means roughly that 20±25% of the receptor is
present as a dimer or in a higher form. In order to obtain
more information on the conformation of 2b in solution, we
performed NOESY experiments on a 500 MHz spectro-
meter. Lots of cross peaks were observed (Fig. 1, left) and
the tri¯ate did not react cleanly under the Reformatski
conditions de®ned by Orsini, the trimethylsilyl ketene
1
1
1
acetal of methyl acetate afforded the ester 12b in variable
2
1
3
yields (50 to 65%). An increasing of the amounts of
reagent or catalyst did not improve this yield.
The last step of the synthesis involved the conversion of the
ester 12b into the targeted host.
one of them showed a weak correlation between H of the
8
benzene ring and one proton of the 6-methylpyridine group.
This correlation may be due to a bowl shape of the host. To
have an idea of the ¯exibility of host 2b, MM2 calcula-
Our previous results in this ®eld showed that this conversion
is not so easy. So we decided to check several methods
described in the literature (Scheme 6). The complex LAH/
1
18
tions were carried out with systematic variation of the
two dihedral angles u and u and assuming a trans confor-
1
4
15
-aminopyridine or BBr activation of the ester 12b led
2
only to tarry material. However, the complex Me Al/2-
3
1
2
mation of the acetamidopyridine moiety (Scheme 7). A
number of 362 conformations was obtained with about 20
in the range of 4±4.2 kcal mol and a global minimum
3
1
6
aminopyridine afforded the host 2a in low yield (21%).
This compound was not soluble enough in chloroform to
undergo NMR binding studies. Fortunately, the more lipo-
philic 2-amino-4,6-dimethylpyridine led to the more soluble
host 2b in a better yield (57%) by increasing the reaction
time. It must be emphasized that an excess of complex
2
1
2
1
having an energy of 3.92 kcal mol was found (Fig. 1,
right). Among the twenty conformers possessing energies
close to the global minimum, most of them are folded
enough to explain the vicinity of H8 and the methyl
(2.4 mol) was required because the isoquinoline nitrogen
is probably involved in the complexation process.
connected to C6 . However the measured distance between
0
these protons, was too high to explain the observed NOE's.
In fact, it is impossible to exclude the contribution of the
auto-association process to account this observation.
Binding experiments
1
13
We ®rst examined the H and C NMR spectra of 2b in
deuteriochloroform. The main features were the following:
First of all, a down®eld shift was observed for the b proton
Binding experiments with some acids and amines were then
undertaken using the H NMR titration method. It must be
pointed out that although dimerization of the receptor is not
1
19
21
Figure 1. Left, observed NOE's in host 2b and right, more stable conformer of 2b (Eˆ3.94 kcal mol ). Intensity of correlations: wˆweak, sˆstrong,
mˆmedium.

J.-F. Bri eÁ re et al. / Tetrahedron 56 (2000) 8679±8688
8683
Figure 2. Job plots of host 2b with two carboxylic acids and two amines.
negligible (vide supra), this process was neglected in the
calculations. For this reason, the computed values are
probably overestimated. The acids used in complexation
studies were acetic and benzoic acids. While increasing
the concentration of the guest, the signals of the two NH
groups of 2b were shifted down®eld until saturation was
achieved. (for the carboxamide NH, these saturations
occurred at 10.9 ppm in the case of acetic acid and
Good values of Ka were obtained (Table 1, 2nd column).
For benzoic acid, the association constants with 2b and
21
2-acetamidopyridine are similar (Table 1, 3rd column)
indicating that the pyridine lone pair and the carboxamide
NH are probably involved in the binding process. However,
in the case of acetic acid, the measured Ka value is about
four fold higher showing the contribution of a supplemen-
tary interaction due to the pyrroloisoquinoline unit. The
difference between the two acids may be attributed to the
bulkiness of the phenyl ring.
11.5 ppm for benzoic acid). The complex stoichiometry
was determined by Job's method of continuous variation
20
(
Fig. 2). Whereas a 1/1 stoichiometry could be assumed
with acetic acid (maximum at xˆ0.5), an other geometry
could not be discharged with benzoic acid. For the
calculation of the associations constants, we found
better to use the observed shifts for the signal corre-
sponding to H10 not involved in self association of the host
On 1/1 complexation, the proximity of benzoic acid into the
binding pocket of the receptor 2b was supported by NOESY
experiments where the aromatic protons of this acid and
both the methyl pyridine protons and H11 of the methylene
group showed cross peaks (Fig. 3, lower part and Fig. 4). It
must be pointed out that no cross peaks were observed with
acetic acid and 2b, under the same NMR conditions. MM2
(no shift was observed with dilution of the host without
guest).
Table 1. Association constants of 2b with acids and amines
21
21
Kass (M ) with 15 or 1
Guest
Kass (M ) with 2b
Dd max (ppm)
a
Acetic acid
Benzoic acid
Benzylamine
4-
Aminomethylpyridine
350
240
470
190
80
230
1.94
2.66
0.62
0.34
a
b
42
250
b
n-Propylamine
45
±
0.44
a
21
Binding with 2-acetamidopyridine 15.

8
684
J.-F. Bri eÁ re et al. / Tetrahedron 56 (2000) 8679±8688
Figure 3. Intermolecular observed NOE's and molecular modeling of the binding of benzoic acid with 2b.
Figure 4. 1/1 Complexation of benzoic acid with 2b. Selected parts of the NOESY spectrum showing cross-peaks.
calculations clearly showed the participation of the pyridine
amide in forming the complex as well as the vicinity of the
above mentioned protons groups (Fig. 3, high).
supported by the low value of Ka with benzylamine with
the simpli®ed receptor 1. On 1/1 complexation, the
proximity of benzylamine into the binding pocket of recep-
tor 2b was also supported by NOESY experiments. Various
cross peaks (Fig. 5, high) were observed indicating that the
amine group is close to the pyrroloisoquinoline moiety, the
methylene group connected to the amide and the pyridine
ring. This vicinity was supported by MM2 calculations
(Fig. 5). Two possibilities were computed: (1) hydrogen
bonding between the lone pair of the amine with the NH
amide and between one amine NH with the pyridine lone
pair (Fig. 5, left); (2) hydrogen bonding of the ®rst amine
NH with the pyridine lone pair and of the second amine NH
with the isoquinoline lone pair (Fig. 5, right). However, it
was impossible to determine the actual complexation
process since the energies of the two possibilities were
closely related.
Similarly, the complexation of three amines was studied:
benzylamine, 4-aminomethylpyridine and n-propylamine.
While no signi®cative shift of the signals corresponding to
the host could be observed, it was possible to monitor the
shift of the NH protons of the guest by increasing its
2
concentration. The Job plot with 4-aminomethylpyridine
was in agreement with a 1/1 complex (Fig. 2, right) and
an other geometry could not be discharged with benzyl-
amine. In every case, the calculation process showed a
good convergence and Ka values are reported in Table 1
(2nd column). These values are compared to those measured
with a simpli®ed amine host (3rd column) previously
1
described by us. With 4-aminomethylpyridine, the Ka is
slightly lower than that observed with the simpli®ed
receptor 1. On the contrary, with benzylamine, the associa-
tion with 2b was more than ten fold higher than that with 1.
This difference might be attributed to a supplementary inter-
action between the aromatic parts. This assumption is
Conclusion
In this paper, the synthesis of the new heterocyclic host 2b

J.-F. Bri eÁ re et al. / Tetrahedron 56 (2000) 8679±8688
8685
Figure 5. Observed intermolecular NOE's and modeling of the two binding modes of benzylamine with 2b.
has been described. This host possesses a pyrroloisoquino-
line structure connected to an acetamidopyridine at the end
of a ¯exible arm. Good association constants were obtained
both with some carboxylic acids or amines and NMR
experiments associated to molecular modeling seemed to
prove that the guests are bound into the pocket resulting
from the ¯exibility of the receptor. We intend now to use
the results described in this paper and in our previous
monitored by thin layer chromatography (TLC) with silica
plates (Merck, Kieselgel 60 F254).
1a,3a,4,5-Tetrahydro-N-methyl-7-methoxy-1H,3H-pyr-
rolo[3,2-c]isoquinolin-2-one (6a). A solution of the above
compound 4a (0.28 g, 1.27 mmol, obtained after several
2
recrystallizations from a mixture ethyl acetate/cyclo-
hexane), paraformaldehyde (0.2 g, 6.4 mmol) in methanol
(12 mL) was heated to re¯ux for 45 min. After cooling to
room temperature, the resulting solution was transferred
into a Paar apparatus, under an argon atmosphere. Raney
nickel W2 (0.14 g) was then added and the bumb was
1
,2
reports to design a receptor able to promote within a
supramolecular process, the reaction between an amine
and a carboxylic acid derivative.
¯ushed with hydrogen. The mixture was then vigorously
shaked under hydrogen (7 bar) for 5 h. The catalyst was
®ltered on a celite pad and the ®ltrate evaporated under
reduced pressure. The crude product was puri®ed by ¯ash
chromatography on a silica column (AcOEt/EtOH: 1/1).
The yield was 0.26 g (90%) of a white solid. Mp 1928C
Experimental
The infra-red spectra were recorded on a Beckmann IR 4250
1
13
spectrometer. The H and C NMR spectra were recorded
either on a 200 MHz, a 400 MHz or a 500 MHz Bruker
apparatus. Spectra were recorded in deuteriochloroform or
(AcOEt/cyclohexane). TLC: R
IR (cm ): 3177 (NH lactame); 1696 (CvO lactame); 1276
ˆ0.25 (AcOEt/EtOH: 1/1).
f
2
1
1
in hexadeuteriodimethylsulfoxide (DMSO-d ). NOESY
6
(C±N). H NMR (CDCl ): 2.40±2.62 (m, 2H); 2.45 (s, 3H);
3
experiments were performed on deuteriochloroform solu-
tions at 258C with the following acquisition parameters:
shift width for F and F 12 ppm; F 2048 dots, F 512
3.40±3.60 (m, 1H); 3.49 (d, 1H, Jˆ15.0 Hz); 3.74 (d, 1H,
Jˆ15.0 Hz); 3.78 (s, 3H); 4.77 (d, 1H, Jˆ6.5 Hz); 6.62 (d,
1H, Jˆ2.6 Hz); 6.84 (dd, 1H, Jˆ8.5 and 2.6 Hz); 6.92 (m,
1
2
2
1
dots; scan number 32; relaxation time 2 s; mixing time
1H); 7.15 (d, 1H, Jˆ8.5 Hz). Anal. calcd for C13
H N O : C,
16 2 2
2
s. Processing parameters: F function sin , phase angle
1
9
67.21; H, 6.96; N, 12.06. Found: C, 67.7; H, 6.3; N, 11.9.
2
2
08, 2048 dots; F function sin , phase angle 908, 1024
1
dots. Chemicals were purchased from Aldrich Co. and
Janssen Co. and, unless otherwise stated, were used without
further puri®cation. Flash chromatographies were
performed with silica 60 (70±230 mesh from Merck) and
1a,3a,4,5-Tetrahydro-7-benzyloxy-1H,3H-pyrrolo[3,2-
c]isoquinolin-2-one (4c). Sodium borohydride (0.49 g,
12.7 mmol) was added in one portion to a ice-cooled solu-
tion of 3-N-[(3-benzyloxyphenyl)methyl]aminopyrrolidine-

8
686
J.-F. Bri eÁ re et al. / Tetrahedron 56 (2000) 8679±8688
2
,5-dione 8 (3.95 g, 12.7 mmol) in a mixture of dichloro-
methane (170 mL) and methanol (85 mL). The solution was
stirred at 08C for 90 min. Two other similar amounts of
sodium borohydride were added followed by, respectively
2
trated in vacuo. The resulting solid (0.145 g, quantitative
yield) was not puri®ed and could be used for the following
steps.
90 min and 3 h stirring at 08C. Water (150 mL was added
and the aqueous layer extracted with dichloromethane
Starting from the N±H compound 4c: A solution of the
benzyloxy derivative 4c (2.02 g, 6.90 mmol), paraformalde-
hyde (1.11 g, 36.2 mmol) in methanol (260 mL) was heated
to re¯ux for 45 min. After cooling at room temperature,
palladium hydroxide (1.28 g) was added. The mixture was
then stirred vigorously at rt under an hydrogen atmosphere
for 16 h. The catalyst was ®ltered on a celite pad and the
®ltrate concentrated in vacuo. The resulting solid was not
puri®ed and could be used for the following steps. The yield
(5£90 mL). After removal of the solvent, the white solid
(intermediate compound 3d) was taken in dichloromethane
(225 mL). After addition of p-toluene sulfonic acid mono-
hydrate (4.96 g, 26.1 mmol), the mixture was heated to
re¯ux for 18 h. The resulting solution was cooled to room
temperature and washed with an aqueous saturated solution
of sodium hydrogencarbonate. The aqueous layer was
extracted with dichloromethane (50 mL). The collected
organic layers were dried on magnesium sulfate and the
solvent was removed under reduced pressure. The crude
solid was puri®ed by ¯ash chromatography on silica gel
2
1
was 1.5 g (100%) of a yellow solid. IR (cm ): 3141 (NH
lactam, OH phenol); 1682 (CvO lactame); 1301 (C±N). H
1
NMR (DMSO-d ): 2.20±2.30 (m, 2H); 2.25 (s, 3H); 3.19±
6
3.30 (m, 1H); 3.27 (d, 1H, Jˆ15.3 Hz); 3.56 (d, 1H,
Jˆ15.3 Hz); 4.53 (d, 1H, Jˆ6.6 Hz); 6.45 (d, 1H,
Jˆ2.4 Hz); 6.61 (dd, 1H, Jˆ8.4 and 2.4 Hz); 7.10 (d, 1H,
Jˆ8.4 Hz); 8.13 (m, 1H); 9.50 (m, 1H). Anal. calcd for
C H N O : C, 66.04; H, 6.47; N, 12.84. Found: C, 66.0;
(
(
ethyl acetate/ethanol: 1/1, R ˆ0.15). The yield was 1.19 g
f
30%) of a white solid. Mp 2088C (ethyl acetate/cyclo-
2
1
hexane). IR (cm ): 3185 (NH amine and lactam); 1696
(
1
CvO lactam). H NMR (CDCl ; 200 MHz): 1.68 (m,
3
12 14
2
2
1
H); 2.32 (dd, 1H, Jˆ1.2 and 17.3 Hz); 2.86 (dd, 1H, Jˆ7
H, 6.2; N, 12.5.
and 17.3 Hz); 3.80 (m, 1H); 3.93 (s, 2H); 4.56 (d, 1H,
Jˆ4.9 Hz); 5.07 (s, 2H); 5.89 (m, 1H); 6.70 (d, 1H,
Jˆ2.5 Hz); 6.90 (dd, 1H, Jˆ2.5 and 8.5 Hz); 7.18 (d, 1H,
1a,3a,4,5-Tetrahydro-7-tri¯uoromethylsulphonyl-N-
methyl-1H,3H-pyrrolo[3,2-c]isoquinolin-2-one (9).
A
1
3
Jˆ8.5 Hz); 7.20±7.50 (m, 5H). C NMR (CDCl ): 39.4
solution of the crude tricyclic phenol (0.7 g, 3.2 mmol) 7,
triethylamine (1.1 mL, 8 mmol) and dimethylformamide
(21 mL, recently distilled over calcium hydride) was cooled
to 08C under an argon atmosphere. Ditri¯imide (2.53 g,
7 mmol) was then added at this temperature. The mixture
was stirred at 08C for 30 min, and then at room temperature
for 4 h. After addition of a small amount of water, the
solvent was removed under reduced pressure (about
3
3
7
4 or 5
4 or 5
14
(
1
C ); 46.35 (C ); 53.1 (C
); 114.7 (C
); 54.1 (C
); 70.55 (C );
); 127.8
9
or 11
9
or 11
aromatic
12.6 (C
); 124.4 (C
); 129.1 (C ); 130.8 (C
); 175.9 (C ). A signal corre-
aromatic
aromatic
aromatic
aromatic
(
C
); 128.5 (C
);
aromatic
aromatic
2
138.5 (C
); 159.5 (C
sponding to a quaternary carbon atom could not be
observed. Anal. calcd for C H N O ´H O: C, 69.21; H,
1
8
18
2
2
2
6.45; N, 8.97. Found: C, 69.4; H, 6.15; N, 9.1.
1
Torr). Water (36 mL) was then added and the product
1a,3a,4,5-Tetrahydro-7-benzyloxy-N-methyl-1H,3H-pyr-
rolo[3,2-c]isoquinolin-2-one (6c). A solution of compound
extracted with dichloromethane (4£30 mL). The organic
layers were dried on magnesium sulfate and the solvent
removed under reduced pressure. The resulting crude
product was puri®ed by ¯ash chromatography (silica gel
ethyl acetate/ethanol: 5/1). The yield was 0.95 g (84%) of
a colorless oil changing in a white solid after storage in
2
4c (0.05 g, 0.17 mmol), paraformaldehyde (0.03 g,
1.9 mmol) in methanol (12 mL) was heated to re¯ux for
45 min. After cooling to room temperature, the resulting
mixture was transferred in a Paar apparatus with Raney
nickel W2 (0.08 g). The mixture was then vigorously shaked
under hydrogen (7 bar) for 5 h. The catalyst was ®ltered on a
celite pad and the ®ltrate evaporated under reduced
pressure. The crude product was puri®ed by ¯ash chromato-
refrigerator. Mp 1178C. TLC: R ˆ0.15 (AcOEt/EtOH:
f
5/1). IR: 3254 (NH lactame); 1697 (CvO lactame); 1423,
1219 (C±N); 1141. NMR H (CDCl , 200 MHz): 2.40±2.60
1
3
(m, 2H); 2.47 (s, 3H); 3.54 (d, 1H, Jˆ15.5 Hz); 3.50±3.70
(m, 1H); 3.78 (d, 1H, Jˆ15.5 Hz); 4.80 d, 1H, Jˆ6.7 Hz);
7.04 (d, 1H, Jˆ2.4 Hz); 7.18 (dd, 1H, Jˆ2.4 and 8.6 Hz);
7.38 (d, 1H, Jˆ8.6 Hz); 7.79 (m, 1H). Anal. calcd for
C H F N O S C, 44.57; H, 3.75; N, 8.00. Found: C,
graphy on a silica column (R ˆ0.15, AcOEt/EtOH: 3/2).
f
The yield was 0.047 g (89%) of a white solid. Mp 2028C
(
2
ethyl acetate/cyclohexane). IR (cm ): 3183 (NH lactam);
1
1
698 (CvO lactam); H NMR (CDCl ): 2.45 (s, 3H); 2.46
1
3
13 13
3
2
4
(
m, 1H); 2.58 (dd, 1H, Jˆ5.3 and 16.8 Hz); 3.49 (d, 1H,
44.2; H, 3.5; N, 7.8.
Jˆ15.7 Hz); 3.41±3.56 (m, 1H); 3.76 (d, 1H, Jˆ15.7 Hz);
4
1
1
.76 (d, 1H, Jˆ6.1 Hz); 5.06 (s, 2H); 6.18 (m, 1H); 6.71 (d,
H, Jˆ2.6 Hz); 6.90 (dd, 1H, Jˆ8.5 and 2.6 Hz); 7.12 (d,
H, Jˆ8.4 Hz); 7.2±7.4 (m, 5H). Anal. calcd for
1-Methoxy-1-trimethysilyloxyethylene (13a) and ethyl 2-
12
trimethylsilyl acetate (13b). A solution of n-butyllithium
in hexane (2.5 M solution, 8 mL, 20 mmol) was added drop-
wise to a cooled mixture (2788C) of freshly distilled diiso-
propylamine (2.7 mL, 20.5 mmol) in THF (15 mL) under an
argon atmosphere. The mixture was stirred for 45 min at
2788C. Methyl acetate (1.6 mL, 20.2 mmol) was then
progressively added and the resulting solution stirred for
30 min. Finally, trimethylsilyl chloride (5 mL, 39.4 mmol)
was added and the solution stirred again for 30 min at rt. The
white solid was ®ltered and the solvent removed under
reduced pressure. The residue was taken in anhydrous
diethyl ether, and the solvent was removed after ®ltration.
C H N O : C, 73,99; H, 6,55; N, 9,09. Found: C, 73,5;
H, 6,3; N, 9,0.
19
20
2
2
1
a,3a,4,5-Tetrahydro-7-hydroxy-N-methyl-1H,3H-pyrrolo-
3,2-c]isoquinolin-2-one (7). Starting from the N-methyl
compound 6c: To a solution of the N-methyl compound
c (0.205 g, 0.66 mmol) in methanol (25 mL), palladium
[
6
hydroxide (0.15 g) was added. The mixture was then stirred
vigorously at rt under an hydrogen atmosphere for 16 h. The
catalyst was ®ltered on a celite pad and the ®ltrate concen-

J.-F. Bri eÁ re et al. / Tetrahedron 56 (2000) 8679±8688
8687
The so-obtained liquid was immediately distilled under
reduced pressure (38±428C/15 Torr), and must be used
quickly. It must be pointed out that, owing to the volatility
of this compound, it was very dif®cult to evaluate the yield
of this reaction. A large amount was lost during the removal
of the solvents and the ®nal product was a mixture of the
required O-silylated compound with the C-silylated one.
This mixture was used after evaluation of the molar ratio
by integration of the NMR spectrum of the crude product
and assuming that the C-silylated product did not react.
Colorless liquid. Bp 38±428C (15 Torr).
2-[7-(1a,3a,4,5-Tetrahydro-N-methyl-1H,3H-pyrrolo-
0
[3,2-c]isoquinolin-2-one)]-N-(2 -pyridyl) acetamide (2a).
A solution of trimethylaluminum (0.23 mL of 2 M solution,
0.46 mmol) in heptane was added to 2-aminopyridine
(0.039 g, 0.42 mmol) in anhydrous tetrahydrofurane
(2 mL) under an argon atmosphere. The mixture was then
stirred at rt for 3 h. This mixture was then slowly added to a
solution of the ester 12b (0.053 g, 0.19 mmol) in anhydrous
THF (3 mL). The ¯ask contents were heated to re¯ux for
18 h and, after cooling, 1 M aqueous hydrochloric acid
(0.46 mL) was carefully added. After one hour stirring at
rt, a saturated aqueous solution of sodium hydrogencarbo-
nate (7 mL) was added and the aqueous layer extracted with
dichloromethane. The collected organic layers were dried
on magnesium sulfate and the solvent removed under
reduced pressure. The residue was puri®ed by ¯ash chroma-
tography on silica gel (the remaining 2-aminopyridine was
eluted with pure ethyl acetate and the product with ethyl
1
3a: H NMR (CDCl , 200 MHz): 0.23 (s, 9H); 3.11 (d, 1H,
1
3
Jˆ2.8 Hz); 3.2 (d, 1H, Jˆ2.8 Hz); 3.55 (s,3H).
1
3b: H NMR (CDCl , 200 MHz): 0.12 (s, 9H); 1.91 (s,
H); 3.63 (s, 3H).
1
2
3
Methyl 2-(1a,3a,4,5-tetrahydro-N-methyl-1H,3H-pyr-
rolo[3,2-c]isoquinolin-2-one)-7-yl acetate (12b). In a
three-necked ¯ask were dissolved the tri¯ate 9 (0.835 g,
acetate/ethanol: 5/1, R ˆ0.1). The yield was 0.013 g (21%)
f
2
of a white solid. IR (cm ): 3422, 3191 (NH lactame,
1
amide); 1683 (CvO lactame, amide). MS (I.E): 333
[M ]. H NMR (CDCl ): 2.46 (s, 3H); 2.50 (m, 2H); 3.52
1
z
1
2
0
.4 mmol), tetrakis triphenylphosphine palladium (0.43 g,
.36 mmol) and lithium acetate (0.49 g, 7.3 mmol) in tetra-
3
(d, 1H, Jˆ15.3 Hz); 3.57 (m, 1H); 3.73 (s, 2H); 3.76 (d, 1H,
hydrofurane (45 mL) under an argon atmosphere. The
preceding compound (0.7 g, 4.8 mmol, OSi/CSiˆ3,
Jˆ15.3 Hz); 4.80 (d, 1H, Jˆ6.5 Hz); 6.04 (m, 1H); 7.04 (m,
1H); 7.11 (s, 1H); 7.27 (m, 2H); 7.70 (m, 1H); 8.19 (m, 3H).
NMR H (DMSO-d ): 2.29 (m, 2H); 2.48 (s, 3H); 3.67(s,
1
1.5 equiv. in OSi) was added and the reaction mixture was
6
heated to re¯ux for 16 h. After cooling to rt, an other amount
of tetrakis triphenylphosphine palladium (0.056 g,
2H); 4.64 (d, 1H, Jˆ6.8 Hz); 7.06 (s, 1H); 7.06 (m, 1H);
7.19 (d, 1H, Jˆ7.8 Hz); 7.29 (d, 1H, Jˆ7.9 Hz); 7.73 (m,
1H); 8.02 (d, 1H, Jˆ8.3 Hz); 8.29 (s, 2H); 10.67 (m, 1H).
0.05 mmol) was added and the mixture again heated to
re¯ux for 5 h. The solvent was removed under reduced pres-
sure and the residue was chromatographed on a column
2-[7-(1a,3a,4,5-Tetrahydro-N-methyl-1H,3H-pyrrolo-
0
0
0
containing 42 g of silica gel (R ˆ0.1, ethyl acetate/ethanol:
[3,2-c]isoquinolin-2-one)]-N-(4 ,6 -dimethyl-2 -pyridyl)-
acetamide (2b). A solution of trimethylaluminum (2 M,
1.3 mL, 2.6 mmol) in heptane was added on a stirred solu-
tion of 2-amino-4,6-dimethylpyridine (0.308 g, 2.5 mmol)
in anhydrous tetrahydrofurane under an argon atmosphere.
The mixture was then stirred for 3 h at rt. The resulting
mixture was slowly added on a solution of ester 12b
(0.307 g, 1.1 mmol) in tetrahydrofurane (17 mL). The reac-
tion mixture was heated to re¯ux for 72 h. After cooling to
rt, aqueous hydrochloric acid (1 M solution, 2.6 mL) was
added. After one hour stirring at rt, a saturated aqueous
solution of sodium hydrogencarbonate (30 mL) was added
and the aqueous layer extracted with dichloromethane
(5£16 mL). The collected organic layers were dried on
magnesium sulfate and the solvent removed under reduced
pressure. The residue was puri®ed by ¯ash chromatography
on silica gel (the remaining 2-amino-4,6-dimethylpyridine
was eluted with pure ethyl acetate and the product with ethyl
f
5/1). The yield was 0.33±0.43 g (50±65%) of a white solid.
Mp (dec) 1838C. IR (cm ): 3410, 3180 (NH lactam); 1736
2
1
1
CvO ester); 1698 (CvO lactam). H NMR (CDCl ): 2.45
dd, 1H); 2.45 (s, 3H); 2.57 (dd, 1H); 3.4±3.6 (m, 1H); 3.50
d, 1H); 3.59 (s, 2H); 3.68 (s, 3H); 3.74 (d, 1H); 4.76 (d, 1H);
(
(
(
3
7,01 (s, 1H); 7.16 (d, 1H); 7.22 (d, 1H); 7,49 (m, 1H). Anal.
calcd for C H N O : C, 65.66; H, 6.63; N, 10.21. Found:
C, 65.7; H, 6.4; N, 9.8.
1
5
18
2
3
1
[
a,3a,4,5-Tetrahydro-7-allyl-N-methyl-1H,3H-pyrrolo-
3,2-c]isoquinolin-2-one (10). To a mixture of the tri¯ate 9
(
0.078 g, 0,22 mmol), lithium chloride (0.027 g) and tetra-
kis triphenylphosphine palladium (0.003 g, 0.004 mmol) in
freshly distilled dimethylformamide (1.5 mL) under an
argon atmosphere, allyltributyltin (85 mL, 0.27 mmol) was
added. The solution was heated to re¯ux for 3 h. After cool-
ing to rt, water (6 mL) was added and the product was
extracted with dichloromethane. After drying on magne-
sium sulfate, the solvent was removed under reduced pres-
sure. The residue was puri®ed by ¯ash chromatography on
acetate/ethanol: 5/1, R ˆ0.3). The yield was 0.228 g (57%)
f
of a white solid. Mp 1478C (dec). R ˆ0.15 (AcOEt/EtOH:
f
2
5/1). IR (cm ): 3287 (NH lactam, amide); 1700 (CvO
1
1
lactam, amide). NMR. H (CDCl ,400 MHz): 2.27 (s, 3H);
silica gel (R ˆ0.1, ethyl acetate/ethanol: 5/1). The yield was
f
3
2
1
.04 g (75%) of a white solid. Mp 1708C. IR (cm ): 3171,
0
3
2.33 (s, 3H); 2.42 (s, 3H); 2.43 (dd, 1H, Jˆ7.0 and 17.0 Hz);
2.50 (dd, 1H, Jˆ5.0 and 17.0 Hz); 3.47 (d, 1H, Jˆ15.0 Hz);
3.52 (m, 1H); 3.64 (s, 2H); 3.71 (d, 1H, Jˆ15.0 Hz); 4.74 (d,
1H, Jˆ7.0 Hz); 6.63 (m, 1H); 6.69 (s, 1H); 7.04 (s, 1H);
7.15 (d, 1H, Jˆ8.0 Hz); 7.19 (dd, 1H, Jˆ8.0 and 1.0 Hz);
1
080 (NH lactame); 1699 (CvO lactame). NMR
H
(
CDCl ): 2.44 (dd, 1H, J hidden by the 6-methyl group);
3
2
.46 (s, 3H); 2.56 (dd, 1H, Jˆ5.4 and 16.9 Hz); 3.35 (d, 2H,
Jˆ6.5 Hz); 3.50 (d, 1H, Jˆ15.3 Hz); 3.50±3.60 (m, 1H);
13
3
2
.76 (d, 1H, Jˆ15.3 Hz); 4.78 (d, 1H, Jˆ6.7 Hz); 5.08 (m,
7.85 (s, 1H); 8.11 (m, 1H). NMR C (numbering of Scheme
7, CDCl ): 21.23 (C ); 23.18 (C ); 31.64 (C ); 42.79 (C );
H); 5.93 (m, 1H); 6.50 (m, 1H); 6.94 (s, 1H); 7.09 (d, 1H,
3
14
13
3
12
Jˆ9.4 Hz); 7.15 (d, 1H, Jˆ7.5 Hz). Anal. calcd for
44.10 (C ); 53.56 (C ); 53.77 (C ); 59.96 (C ); 111.91
11 5 1a 3a
C H N O: C, 74.33; H, 7.50; N, 11.56. Found: C, 74.25;
2
(C3
(C or C ); 131.65 (C ); 133.78 (C ); 134.33 (C ); 150.57
); 120.51 (C5 ); 127.25 (C ); 128.41 (C or C ); 128.54
0 0
15
18
6 8 9
H, 7.4; N, 11.4.
8
9
9a
7
6a

8
688
J.-F. Bri eÁ re et al. / Tetrahedron 56 (2000) 8679±8688
(
1
6
C2
0
); 150.73 (C4
76.32 (C ). Anal. calcd for C H N O : C, 69.20; H,
0
); 155.71 (C6
0
); 169.61 (CvO amide);
References
2
21 24
4
2
.65; N, 15.37. Found: C, 69.12; H, 6.67; N, 15.55
1. Charpentier, P.; Bri eÁ re, J. F.; Dupas, G.; Qu e guiner, G.;
Bourguignon, J. Tetrahedron 1996, 52, 10441.
Calculation of binding constants. Prior to these experi-
ments, CDCl was stored on anhydrous potassium carbonate
in order to remove any residual acidity from this solvent.
. Briere, J. F.; Charpentier, P.; Dupas, G.; Queguiner, G.;
Á Â
2
Bourguignon, J. Tetrahedron 1997, 53, 2075.
3
3
. Briere, J. F.; Dupas, G.; Queguiner, G.; Bourguignon, J.
Heterocycles 2000 (in press).
. Gribble, G. W.; Nutaitis, C. F. Synthesis 1987, 709.
Á
Â
Dimerization constant of host 2b. A stock solution of the
host (0.018235 g) in CDCl (1 mL) was prepared with a
4
3
5. Bhattacharyya, S.; Catterjee, A.; Duttachowdhury, S. K.
J. Chem. Soc., Perkin Trans. 1 1994, 1.
2
1
great accuracy (0.05003 mol L ). An NMR tube (5 mm
diameter) was charged with CDCl3 (0.4 mL). Small
amounts of the stock solution were then successively
6. (a) Sondengam, B. L.; Hentchoya Hemo, J.; Charles, G.
Tetrahedron Lett. 1973, 29, 261. (b) Kapnang, H.; Charles, G.;
Sondengam, B. L.; Hentchoya Hemo, J. Tetrahedron Lett. 1977,
1
added into the NMR tube and the corresponding H NMR
spectrum was recorded. The changes in the chemical shifts
of host were monitored and the dimerization constant was
computed according to the method described by Horman
and Dreux for caffeine and use of Excel-97 . The added
volume of stock solution (mL) and the observed chemical
shift of the N±H lactam group (ppm) were the following:
1
8, 3469.
. Yong Sup Lee; Dong Wook Kang; Sook Ja Lee; Hokoon Park
7
J. Org. Chem. 1995, 60, 7149.
1
7
w
8
9
1
1
. Ritter, K. Synthesis 1993, 735.
. Szabo-Pusztay, K.; Szabo, L. Synthesis 1982, 85.
0. (a) Echavarren, A. M.; Stille, J. K. J. Am. Chem. Soc. 1987,
09, 5478. (b) Tilley, J. W.; Sarabu, R.; Wagner, R.; Mulkerins, K.
0
6
6
.1, 5.93; 0.01, 6.02; 0.01, 6.10; 0.01, 6.19; 0.01, 6.24; 0.02,
.34; 0.02, 6.43; 0.02, 6.50; 0.02, 6.565; 0.05, 6.68; 0.05,
.77; 0.05, 6.84; 0.1, 6.95; 0.1, 7.02; 0.1, 7.07; 0.1, 7.11
J. Org. Chem. 1990, 55, 906. (c) For a comprehensive review see:
Farina, V.; Krishnamurthy, V.; Scoot, W. J. Org. React. 1997, 50,
1
1
1
1
1
2
1
5
1
1
1
1
1
1
.
2
0
Job plots. The continuous variation method of Job was
used in order to determine the stochiometry of the
complexes. Two solutions in deuteriochloroform (2 mL)
of the host and the guest were prepared (the concentrations
1. Orsini, F.; Pelizzoni, F. Synth. Commun. 1987, 1, 1389.
2. Ainsworth, C.; Chen, F.; Kuo, Y. N. J. Organomet. Chem.
972, 46, 59.
3. Carfagna, C.; Musco, A.; Sallese, G. J. Org. Chem. 1991, 56,
61.
2
1
must be the same, in the range 0.013 mol L ). NMR tubes
were then charged with amounts of these two solutions in
order to obtain a total volume of 0.5 mL. Volumes (mL) of
host and guest solutions were as follows: 0.5, 0; 0.4, 0.1; 0.3,
0.2; 0.25, 0.25; 0.2, 0.3; 0.1, 0.4; 0.5. A signal correspond-
ing to the NH amide of 2b for binding of acids and NH for
4. Solladi e -Cavallo, A.; Benchegroun, M. J. Org. Chem. 1992,
7, 5831.
5. Yazawa, H.; Tanaka, K.; Kariyone, K. Tetrahedron Lett. 1974,
5 (46), 3995.
2
6. Basha, A.; Lipton, M.; Weinreb, S. M. Tetrahedron Lett. 1977,
8 (48), 4171.
binding of amines was then monitored. The curves are
depicted in Fig. 2.
7. Horman, I.; Dreux, B. Helv. Chim. Acta 1984, 67, 754.
w
8. Calculations were carried out with the PCMODEL imple-
Association constants. An amount of host 2b was accurately
weighed and diluted in deuteriochloroform (0.5 mL) in
order to obtain a concentration exactly known of about
0.02 mol L . The NMR spectrum of 2b alone was then
recorded. The amine or acid guest were accurately weighed
w
mentation of MM2 force ®eld. PCMODEL was purchased from
Serena Software, Box 3076, Bloomington, IN 47402-3076, USA.
w
Similar calculations were carried out with Cerius2 (MSI software,
2
1
9685 Scranton Road, San Diego, CA 92121-3752, USA) and the
CVFF force ®eld. The results were similar.
and diluted in CDCl (5 mL) in order to obtain a concentra-
3
2
1
19. (a) Rotger, M. C.; Gonzalez, J. F.; Ballester, P.; Deya, P. M.;
Costa, A. J. Org. Chem. 1994, 59, 4501. (b) Bhattacharyya, S.;
Chatterjee, A.; Duttachowdhury, S. K. J. Chem. Soc., Perkin Trans.
tion exactly known of about 0.3 mol L . Aliquots of the
guest stock solution were successively added and the corre-
sponding NMR spectra recorded: Aliquots of 0.01 mL until
achieving host±guest,1, then 0.02 mL aliquots until a total
added volume of 0.1 mL, 0.04 mL until a total added
volume of 0.3 mL, 0.1 mL aliquots until a total added
volume of 1 mL and ®nally 0.2 mL aliquots until a total
added volume of 2 mL. The titration experiment may be
stopped when saturation is achieved before adding 0.2 mL
aliquots. The chemical shifts of selected protons were
measured. Dilution experiments were also performed in
order to be sure that the observed shifts were only due to
complexation of guest.
1
1
1994, 1. (c) Kelly, T. R.; Kim, M. H. J. Am. Chem. Soc. 1994,
16, 7072. (d) Connors, K. A. Binding Constants; Wiley: New
York, 1989.
0. (a) Job, A. Ann. Chim. (10th series) 1928, 9, 113. For recent
and comprehensive examples of Job plots see: (b) Blanda, M. T.;
2
Horner, J. H.; Newcomb, M. J. Org. Chem. 1989, 54, 4626.
(
c) Coteron, J. M.; Hacket, F.; Schneider, H. J. J. Org. Chem.
1
996, 61, 1429.
1. For binding of acids with the 2-acetamidopyridine moiety see
2
for example: (a) Yang, J.; Fan, E.; Geib, J.; Hamilton, A. D. J. Am.
Chem. Soc. 1993, 115, 5314. (b) Garcia-Tellado, F.; Goswami, S.;
Chang, S. K.; Geib, A. D.; Hamilton, A. D. J. Am. Chem. Soc.
1990, 112, 7393. (c) Goswami, S.; Ghosh, K.; Dasgupta, S.
Tetrahedron 1996, 52, 12223. For a review concerning molecular
recognition of acids see: Seel, C.; Galan, A.; de Mendoza, J. Top.
Cur. Chem. 1995, 175, 101±132.
Acknowledgements
Financial support from the R e gion Haute Normandie is
gratefully acknowledged.