A practical synthesis of 7-azaindolylcarboxy-endo-tropanamide (DF 1012)

Article abstract of DOI:10.1021/op025570t

An optimised cost-effective synthesis of the new antitussive drug, DF1012, is herewith reported. The new synthetic route to the key intermediate DF1005 is based on the unusual deprotection step of the 1-tert-butyl-3-cyano-7-azaindole intermediate, which can also be regarded as a convenient way for the industrial production of the expensive 7-azaindole 1. The second key intermediate, endo-tropanamine 6, was obtained in high yield by a novel one-pot stereoselective process using a Pd-catalysed reductive amination procedure.

Full text of DOI:10.1021/op025570t

Organic Process Research & Development 2003, 7, 209−213
A Practical Synthesis of 7-Azaindolylcarboxy-endo-tropanamide (DF 1012)
Marcello Allegretti,*,† Roberto Anacardio,† M. Candida Cesta,† Roberto Curti,† Marco Mantovanini,† Giuseppe Nano,†
Alessandra Topai,† and Giuseppe Zampella‡
Chemistry Section, Dompe´ S.p.A. Research Centre, Via Campo di Pile, 67100 L’Aquila, Italy, and Biotechnologies and
Biosciences Department of Milano-Bicocca UniVersity, Piazza della Scienza 2, 20126 Milan, Italy
Scheme 2
Abstract:
An optimised cost-effective synthesis of the new antitussive drug,
DF1012, is herewith reported. The new synthetic route to the
key intermediate DF1005 is based on the unusual deprotection
step of the 1-tert-butyl-3-cyano-7-azaindole intermediate, which
can also be regarded as a convenient way for the industrial
production of the expensive 7-azaindole 1. The second key
intermediate, endo-tropanamine 6, was obtained in high yield
by a novel one-pot stereoselective process using a Pd-catalysed
reductive amination procedure.
Introduction
starting point for scaling-up studies due to the extremely high
cost of the starting materials 1 and 3.
7-Azaindolylcarboxy-endo-tropanamide (DF 1012) (Fig-
ure 1) is the selected candidate drug in a new class of
nonnarcotic antitussive compounds and is actually under
investigation in phase II clinical trials.
Considering the number of competitors in the antitussive-
active drug area, the final price of the new product is one of
the main issues to be evaluated because it could play a crucial
role for its commercial success. The active principle high
cost could then become, in this scenario, a serious drawback
for an antitussive drug once on the market.
Results and Discussion
Preparation of 7-Azaindolyl-3-carboxylic Acid (2). The
proposed pathway for the industrial preparation of 2 is
represented in Scheme 3.
Figure 1. DF 1012.
DF 1012 and the class-related compounds were synthe-
sised from the common intermediate 7-azaindolyl-3-car-
boxylic acid 2 via the general route outlined in Scheme 1.
The cheap starting materials, succinonitrile and ethylfor-
mate, were condensed in the presence of sodium methoxide
to give the corresponding salt of 2-hydroxymethylenebuty-
ronitrile the latter of which was treated, in the reaction
medium, with tert-butylamine to obtain the enaminonitrile
7. The base-catalyzed internal condensation of 7 led to 1-tert-
butyl-2-amino-4-cyanopyrrole 8 in high yield. In the fourth
step, the 7-azaindole structure was easily built up by reaction
of the intermediate 8 with 1,1,3,3-tetramethoxypropane in
the presence of catalytic p-toluenesulfonic acid. Treatment
of 9 with AlCl3 (3 equiv) allowed the removal of the tert-
butyl protecting group; the acidic hydrolysis of the cyano
functionality in controlled conditions (step 6a) led to the
desired 7-azaindolyl-3-carboxylic acid 2.
Scheme 1
This process is based on an already existing method.2
Broderick proposed a general preparation method for 1-sub-
stituted-2-aminopyrroles and pointed to their use as inter-
mediates in the synthesis of 1-substituted-3-cyano-7-
azaindoles. Even if 1-H-7-azaindoles were not accessible by
means of this procedure, Broderick showed the synthesis of
the attractive intermediate N-benzyl-3-cyano-7-azaindole.
endo-Tropanamine 6 is not a commercially available reagent
and therefore was prepared, in the discovery phase, from
commercial tropinone 3, according to an established synthetic
method1 described in Scheme 2.
The synthetic process was a satisfactory general strategy
for a laboratory-scale preparation, but it was not a suitable
* Author for correspondence. E-mail: marcello.allegretti@dompe.it.
† Dompe´ S.p.A. Research Centre.
‡ Milano-Bicocca University.
(2) Brodrick, A.; Wibberley, D. G. J. Chem. Soc. Perkin Trans. I 1975, 19,
(1) Bagley, J. R.; Riley, T. N. J. Heterocycl. Chem. 1977, 14, 599.
1910.
10.1021/op025570t CCC: $25.00 © 2003 American Chemical Society
Published on Web 02/21/2003
Vol. 7, No. 2, 2003 / Organic Process Research & Development
•
209

Scheme 3
Scheme 4
to 1 was in fact obtained. Considering this result, the overall
process from succinonitrile can thus be considered both as
an advantageous synthetic method for the intermediate 2
preparation (step 6a) and as a new alternative convenient
method for the synthesis of the expensive compound 1.
The overall high yield obtained at kilogram-lab scale,
together with the low cost of the starting materials and the
ease of purification, makes this process an economic and
efficient way to the 3-carboxy-7-azaindole 2 synthesis.
Preparation of endo-Tropanamine 6. The most widely
used method1 for the preparation of endo-tropanamine 6 is
the catalytic debenzylation (PtO2, EtOH) of the Schiff base
4 derived from tropinone and benzylamine coupling (Scheme
2). The iminic intermediate 4 reduction leads to a disap-
pointing 8:2 ratio of endo/eso isomers; however, the pure
endo-tropanamine 6 can be easily isolated in the final step
by crystallisation of the corresponding bis-hydrochloride.
Few years ago Eli-Lilly’s researchers studied how to increase
the stereoselectivity in the endo-tropanamine 6 synthesis.5
They found that the use of bulky reducing agents such as
sodium tris[(2-ethylhexanoyloxy)borohydride] enabled the
set up of a new, highly stereoselective (endo/eso > 50:1)
process for the preparation of the endo-N-benzyl-tropan-
amine. This method is surely appealing in terms of yield
and stereoselectivity; however, it makes necessary a careful
recovery of the valuable reagent.
Therefore, in the first phase of our work, we tried to
synthesise compound 2 through N-benzyl-3-cyano-7-azain-
dole, but any attempt3,4 (Pd/C- or PtO2-catalysed hydrogena-
tion, Na/liq.NH3, Na/naphthalene, ethylchloroformate/CF3-
COOH) to remove the benzyl group completely failed. These
results were in agreement with the reported low reactivity
of N-benzyl pyrroles in debenzylation conditions.
The use of tert-butylamine in the first step of the synthesis
allowed us to perform the final deprotection step using a
Lewis acid reagent with surprisingly high yield.
The tert-butyl group is not known as a protecting group
for indole or pyrrole nitrogen; the cyano group-activating
effect in position 3 is essential to determine the unusual high
reactivity of our intermediate 9. In fact, the reaction
proceeded in very low yield, as the 3-carboxy or 3-carboxy-
amide 9-analogues were the starting materials.
At some stage in our initial attempts to scale-up the
Scheme 2 method we faced unexpected difficulties.
The higher 4 concentration, due to the necessary solvent
reduction at kilogram-lab scale, favoured the catalyst poison-
ing, thus decreasing the hydrogenolysis rate. Even a second
addition of the same amount of catalyst did not afford the
complete transformation of the intermediate N-benzyltro-
panamines.
More, in our hands, the formation of undesired by-
products, mainly N-ethyltropanamines, unexpectedly in-
creased. The combination of these two factors lowered the
final yield of 6 (52% vs 75%) and compromised its isolation.
To bypass the deprotection step we studied and developed
a novel one-step reductive amination procedure using am-
monium formate both as nitrogen and hydrogen source (see
Scheme 4).6
We could then argue that the 3-cyano-group coordination
by 1 equiv of AlCl3 is crucial for the activation of the Lewis
acid-mediated C-N bond cleavage (as depicted in Figure
2).
The new method, if compared with the classical Borch
reductive amination procedure,7,8 led to very good results
not only in terms of yield but also in terms of stereoselec-
Figure 2. Effect of the Lewis acid in the deprotection step.
(5) Burks, J. E., Jr.; Espinosa, L.; LaBell, E. S.; McGill, J. M.; Ritter, A. R.;
Speakman, J. L.; Williams, M.; Bradley, D. A.; Haehl, M. G.; Schmid, C.
R. Org. Process Res. DeV. 1997, 1, 198.
(6) Allegretti, M.; Berdini, V.; Cesta, M. C.; Curti, R.; Nicolini, L.; Topai, A.
Tetrahedron Lett. 2001, 42, 4257.
The final hydrolytic step to 2 requires carefully controlled
conditions to avoid contemporary product decarboxylation-
(step 6b). Under reflux the quantitative transformation of 9
(7) Borch, R. F.; Bernstein, M. D.; Durst, H. D. J. Am. Chem. Soc. 1971, 93,
2897.
(3) Lim, M.-I.; Klein, R. S.; Fox, J. J. J. Org. Chem. 1979, 44, 3826.
(4) Anderson, H. J.; Groves, J. K. Tetrahedron Lett. 1971, 34, 3165.
(8) Yoon, N. M.; Kim, E. G.; Son, H. S.; Choi, J. Synth. Commun. 1993, 23,
1595.
210
•
Vol. 7, No. 2, 2003 / Organic Process Research & Development

Table 1. a
10% Pd-C
(% w/w)
molar ratio
6
entry no.
solvent
T (h)
HCO2-NH4+/3 (yield%)
1
2
3
4
5
6
7
8
9
EtOH
MeOH
MeOH
MeOH
MeOH
MeOH/H2O, 9:1
MeOH/H2O, 9:1
MeOH/H2O, 9:1
MeOH/H2O, 8:2
MeOH/H2O, 9:1
MeOH/H2O ,9:1
MeOH/H2O, 9:1
24
24
24
24
30
24
30
24
24
24
24
24
3
3
3
3
3
3
3
3
3
10
5
2
12
42
58
43
40
65
55
50
66
79
81
10
30
10
10
10
20
20
20
20
10
10
11
12
6
12
12
a Note: All reactions were carried out at 25 °C.
tivity. No traces of the eso isomer 5 were detected in the
reaction mixture before the crystallisation.
As shown in Table 1, we tested a number of conditions
to optimise yield, reaction time, and catalyst amount. Initially
the reaction was performed in methanol and led to the pure
6 product in low yield but good stereoselectivity (entry 2).
Optimisation in alcoholic media did not proceed beyond fair
yields (entries 3-5). A surprisingly favourable effect on the
reaction course due to water addition was observed (entries
6-12). In the optimised conditions an hydro alcoholic
(MeOH/H2O, 9:1) mixture was the chosen solvent.
The catalyst amount was found to be a crucial parameter
in reducing the reaction time. As detailed in the Experimental
Section, we preferred to use a quite high catalyst/substrate
ratio (1:4 w/w) to ensure reaction completion in 3 h, as the
catalyst can be recovered in excellent yield. The reductive
amination reaction is generally supposed to proceed through
an iminic intermediate
The unexpected reactivity of ketones in the hydro
alcoholic mixture led us to the hypothesis of the hemi-aminal
species as the reaction key intermediates. At this point it is
also important to keep in mind that the observed stereo-
chemistry was never affected by relative changes in the
optimisation process (GC-MS endo/eso ratio >99:1).
The different reactivity of the two possible hemi-aminals
in the Pd-catalysed hydrogenolytic step justifies the absolute
stereoselectivity toward the endo-tropanamine, according to
the mechanism shown (Figure 3).
The new reductive amination procedure was applied to a
wide number of substrates and was shown to be a general,
straightforward, and mild route for the synthesis of primary
amines from ketones. The overall results of the mechanistic
and reactivity studies have been extensively discussed in our
recent work.9
The reported results of the kilogram-lab scale-up trials
confirmed the proposed process as a simple, high-yielding,
one-pot alternative to the procedures of Borch7,8 and Leuck-
art10,11 for primary amines synthesis. The observed absolute
stereoselectivity referring to the tropanonic ring functionality
Figure 3. Proposed reaction mechanism of the endo-tropan-
amine stereoselective synthesis via hemi-aminal.
makes this process, from our point of view, extremely
advantageous for the industrial preparation of the endo-
tropanamine 6 as confirmed by the excellent results at
kilogram-lab scale (see Experimental Section).
Preparation of DF1012 (N-endo-8-Methyl-8-azabicyclo-
[3.2.1.]oct-3-yl-1H-pyrrolo[2,3-b]pyridin-3-yl-carboxam-
ide). The condensation of the two key-intermediates 2 and
6 to DF 1012 was generally performed using the DCC/HOBT
system. Despite the satisfactory yield obtained in these con-
ditions, several attempts were made to avoid the use of ex-
pensive (HOBT) and toxic (DCC) reagents in the final step.
The need for an alternative condensation route prompted
us to study the acyl chloride coupling. However, the low
solubility of the formed acyl chloride hydrochlorides strongly
limited the choice of the solvent, and in most cases, a very
low yield was obtained.
The best results were obtained using dioxane at room
temperature in the presence of triethylamine to give DF 1012
with 82% yield.
Experimental Section
General Procedures. Ten percent palladium on activated
charcoal was purchased of “purissimum” grade from Fluka.
Other commercial reagent grade solvents and chemicals were
used as obtained unless otherwise noted. Melting points were
obtained on a Buchi 530 apparatus and are uncorrected. Thin-
layer chromatography was carried out with Macherey-Nagel-
DURASIL-25 silica gel plates. Nuclear magnetic resonance
(NMR) spectra were recorded on a Bruker ARX-300 MHz.
GC-MS analyses were performed on a Varian 3400 chro-
matograph (analytical column: Supelwax-10 fused silica
capillary column) and an Incos 50 XL (Finnigan Mat)
spectrometer. Mass spectra were recorded on a Finnigan TSQ
700, a triple-quadrupole mass spectrometer, equipped with
electrospray ionisation (ESI) source (Thermo Finnigan, San
Jose, CA).
(9) Allegretti, M.; Berdini, V.; Cesta, M. C.; Curti, R.; Di Bello, N.; Nano,
G.; Nicolini, L.; Topai, A. Tetrahedron 2002, 58, 5669.
(10) Moore, L. M. Org. React. 1949, 5, 301.
(11) Carlson, R.; Lejon, T.; Lundsted, T.; Le Clouerec, E. Acta Chim. Scand.
1993, 47, 1046.
Vol. 7, No. 2, 2003 / Organic Process Research & Development
•
211

N-tert-Butylaminomethylenesuccinonitrile (7). A 250-L
steel reactor in nitrogen atmosphere was loaded with sodium
methoxide (4 kg, 74 mol) and anhydrous toluene (60 L).
The suspension was cooled to 5 °C under stirring, and a
solution of ethyl formate (6.17 kg, 83.3 mol) and succino-
nitrile (5.60 kg, 69.9 mol) in anhydrous toluene (60 L) was
added dropwise. The addition rate was controlled to avoid
the inner temperature exceeding 20 °C. Once the addition
was done, the mixture was stirred for 2.5 h at room
temperature, still under nitrogen atmosphere. tert-Butylamine
(5.22 kg, 71.4 mol) and acetic acid (5.0 kg, 83.3 mol) were
subsequently added to the reaction mixture, ensuring that
the temperature did not exceed 40 °C.
The mixture was refluxed for 2.5 h and then cooled to
room temperature; an 18% NaCl solution (17 L) was added,
and the two phases were separated. The aqueous layer was
extracted twice with toluene (7.5 L), and the collected organic
extracts were evaporated under reduced pressure to give 7
(10.28 kg, 63.0 mol) as a brown solid (E/Z isomers mixture
2:1). Typically, the obtained solid was subsequently used
without further purification. An analytically pure sample was
recrystallised from acetone/diethyl ether: mp 118-119 °C;
1H NMR (CDCl3) δ 1.35 (s, 9H), 3.18 (s, 2H), 5.0 (bs, 1H),
6.88 (d, 1H, (E isomer), J ) 14.2 Hz), 6.99 (d, 1H, (Z
isomer), J ) 14.5 Hz); IR (Nujol) 3300, 2170, 1650, 1210
cm-1. Anal. Calcd for C9H13N3: C, 66.23; H, 8.03; N, 25.74.
Found: C, 65.98; H, 8.01; N, 25.81. ESI-MS m/z 164 [M +
H]+.
was added to the residue, and then it was refluxed and stirred
for 15 min. The latter mixture was cooled again to room
temperature, and the solid was filtered off and washed with
isopropyl ether. The organic layers were collected, and the
solvent was distilled at reduced pressure. A solid from the
latter residue was crystallised from n-heptane (18 L),
separated by centrifugation and desiccated at 40 °C to give
9 (6.35 kg, 31.9 mol) as yellow powder in 80% yield (mp
84-92 °C). The product was subsequently used without
further purification. An analytically pure sample was recrys-
tallised from cyclohexane: mp 91-92 °C; 1H NMR (CDCl3)
δ 1.92 (s, 9H), 7.15-7.22 (m, 1H), 7.85 (s, 1H), 8.05 (d,
1H, J ) 8 Hz), 8.45 (m, 1H); IR (Nujol) 2190, 1590, 1515,
760 cm-1. Anal. Calcd for C12H13N3: C, 72.33; H, 6.58; N,
21.09. Found: C, 72.16; H, 6.61; N, 21.11. ESI-MS m/z 200
[M + H]+.
3-Cyano-7-azaindole (10). In nitrogen atmosphere at
room temperature, 9 (6.00 kg, 30.1 mol) was added por-
tionwise to a suspension of AlCl3 (90 mol in 50 L of
chlorobenzene). The mixture was stirred in a 250-L glass-
lined reactor. At the end of additions the reaction mixture
was refluxed for 8 h. The resulting mixture was cooled to
room temperature and by means of a dosing pump added to
precooled 2 N HCl (120 L) in a 250-L glass-lined reactor
and then stirred for 2 h. Celite (4 kg) was added to the
mixture, which was then transferred through a filter into a
250-L glass-lined reactor. The phases were separated, and
the organic one was washed with 2 N HCl (20 L): the
collected aqueous layers were washed again with chloroben-
zene (20 L). The strongly acidic aqueous solution was
adjusted to pH ) 2.2-2.4 with 50% aqueous NaOH while
cooling. The mixture was stirred while a precipitate formed;
the solid was separated from mother liquors by centrifuga-
tion, washed with water, and desiccated at T ) 70 °C to
give 10 (3.93 kg, 27.5 mol) in 91.2% yield, as a slightly
violet powder. The powder was subsequently used without
further purification. An analytically pure sample was recrys-
2-Amino-1-tert-butyl-4-cyanopyrrole (8). In a 250-L
steel reactor 85% KOH (7.35 kg, 111.3 mol) was dissolved
in denatured ethanol (60 L) at 50 °C ,and the opalescent
solution was then cooled at room temperature. To the latter,
a solution of 7 (10.00 kg, 61.3 mol) in denatured ethanol
(17 L) was added while stirring. This mixture was continu-
ously stirred for 4 h at room temperature and then concen-
trated under reduced pressure until a solid started to separate.
The mass was then diluted with water (50 L) and stirred at
room temperature for another hour. The precipitate was
separated by centrifugation, washed with water (15 L), and
desiccated at 50 °C to give 8 (6.73 kg, 41.2 mol) as brown
solid. The yield based on succinonitrile was 60.6%. Typi-
cally, the solid was used in subsequent reactions without
further purification. An analytically pure sample was recrys-
1
tallised from isopropyl alcohol: mp ) 259-260 °C; H
NMR (DMSO-d6) δ 7.30 (dd, 1H, J1 ) 8.0 Hz, J2 ) 4.7
Hz), 8.12 (dd, 1H, J1 ) 8.0 Hz, J2 ) 1.5 Hz), 8.41 (dd, 1H,
J1 ) 4.7 Hz, J2 ) 1.6 Hz), 8.45 (s, 1H), 12.8 (bs, 1H); IR
(Nujol) 3070, 2190, 1580, 1280, 760 cm-1. Anal. Calcd for
C8H5N3: C, 67.12; H, 3.52; N, 29.35. Found: C, 67.09; H,
3.50; N, 29.32. ESI-MS m/z 144 [M + H]+.
1
tallised from isopropyl ether: mp 112-113 °C; H NMR
(CDCl3) δ 1.60 (s, 9H), 3.15 (bs, 2H), 5.76 (d, 1H, J ) 2.2
Hz), 6.96 (d, 1H, J ) 2.2 Hz); IR (Nujol) 3380, 3320, 2190,
1640, 1200, 785, 715 cm-1. Anal. Calcd for C9H13N3: C,
66.23; H, 8.03; N, 25.74. Found: C, 66.06; H, 8.04; N, 25.76.
ESI-MS m/z 164 [M + H]+.
1-tert-butyl-3-cyano-7-azaindole (9). To a solution of 8
(6.50 kg, 39.8 mol) in toluene (65 L) in a 250-L glass-lined
reactor were added 1,1,3,3-tetramethoxypropane (7.11 kg,
43.3 mol) and then p-toluenesulfonic acid monohydrate (758
g, 3.98 mol). The reaction mixture was gently heated to reflux
temperature, thus removing methanol (reaction by-product)
by azeotropic distillation. After 1 h the reaction mixture was
cooled at 40 °C, transferred into a 100-L steel reactor and
evaporated under reduced pressure. Isopropyl ether (45 L)
7-Azaindol-3-carboxylic acid (2). A 250-L glass-lined
reactor was loaded with 10 (3.21 kg, 22.4 mol) and 32%
HCl (32 L). The suspension was heated to 70 °C to give a
homogeneous solution, which was stirred at 70 °C for 20 h;
water (25 L) and charcoal (320 g) were then added. The
resulting mixture was stirred at 70 °C for 1 h; Celite (320
g) was then added, and the mixture was left stirring for an
additional hour at 70 °C. The warm solution was transferred
into a 250-L glass-lined reactor through a filter, the filtrate
cooled by brine to 8-10 °C, and the pH adjusted to 2.5 with
50% NaOH. A white precipitate formed; after the suspension
stirred for 2 h at room temperature, the solid was separated
by centrifugation, washed with copious water till disappear-
ance of chloride ions, and dried at T ) 70 °C to give 2 (3.43
212
•
Vol. 7, No. 2, 2003 / Organic Process Research & Development

kg, 21.15 mol) in 94.3% yield as a white solid. An
analytically pure sample was recrystallised from ethanol/2
N HCl, 4:1: mp 245-246 °C; 1H NMR (DMSO-d6) δ 7.18-
7.25 (m, 1H), 8.13 (s, 1H), 8.29-8.33 (m, 2H) 11.7-12.5
(bs, 2H); IR (Nujol) 3460, 2920, 2570, 1685, 1350, 1195,
1140, 1040,805, 750 cm-1. Anal. Calcd for C8H6N2O2: C,
59.26; H, 3.73; N, 17.28. Found: C, 59.12; H, 3.74; N, 17.25.
ESI-MS m/z 163 [M + H]+.
N-endo-8-Methyl-8-azabicyclo[3.2.1.]oct-3-yl-1H-pyr-
rolo[2,3-b]pyridin-3-yl-carboxamide (DF1012). In a 25-L
glass reactor 2 (1.3 kg, 8.02 mol) and toluene (10 L) were
mixed. The suspension was stirred and heated to T ) 70
°C. Thionyl chloride (0.75 L, 10.2 mol) was dripped into
the mixture slowly, and this mixure was then refluxed for 2
h. After cooling at room temperature the solid was separated
by vacuum filtration, washed with toluene, and then dried
at 60 °C in a vacuum oven to give 2 chloride hydrochloride
as a white solid (1.4 kg, 6.45 mol).
6‚bis-hydrochloride (900 g, 4.22 mol), placed in 25-L
glass reactor, was dissolved in water (1 L) and cooled at T
) 5 °C. Dichloromethane (2 L) was added, and 32% aqueous
NaOH (1 L; 10.8 mol) was slowly dripped under stirring
and cooling. The two phases were separated, and the aqueous
one was re-extracted with dichloromethane (2 L). Solvent
was distilled off under reduced pressure to give 6 (0.55 kg,
3.92 mol) as a yellow oil.
7-Azaindole (1). A 250-L glass-lined reactor was loaded
with 10 (3.21 kg, 22.4 mol) and HCl 32% (32 L). The
suspension was refluxed for 24 h. The solution was cooled
to 70 °C, and additional HCl 32% (32 L) was poured into
the mixture; the mass was then reheated to reflux for an
additional 24 h. Finally, the mixture was cooled to room
temperature and diluted with water (40 L), and the pH was
adjusted to 2.5 with 50% aqueous NaOH. Charcoal (320 g)
was added, and the mixture was stirred for 1 h at room
temperature. Celite (320 g) was added to the mixture which
was then stirred for 1 h at the same temperature and then
transferred into a 250-L steel reactor through a filter. The
filtrate was cooled by brine to 8-10 °C; methylene chloride
was added (25 L), and the pH was adjusted to 12 with 50%
aqueous NaOH. The phases were separated, and the aqueous
layer was extracted again with methylene chloride (25 L).
Solvent was distilled under reduced pressure to give 1 (1.93
kg, 16.3 mol) in 72.9% yield as a white solid: mp ) 105-
The compound 2 chloride hydrochloride (0.705 kg, 3.25
mol) was suspended in a 25-L glass reactor with dioxane (5
L), then triethylamine (0.724 kg, 7.15 mol) was added under
stirring. A solution of 6 (0.55 kg, 3.92 mol) in dioxane (5
L) was added dropwise and the mixture kept under stirring
overnight. The reaction mixture was concentrated under
reduced pressure, water (10 L) was added to the residue,
and the pH was adjusted to 1.5 with 2 N HCl. The aqueous
solution was first washed with dichloromethane (2 × 5 L),
and then 50% aqueous NaOH was slowly added while
stirring and cooling until pH > 10. The mixture was stirred
overnight at T ) 5 °C; the precipitate was vacuum-filtered,
washed with water (10 L), and dried under vacuum at T )
60 °C to give a pale-pink solid (0.74 kg, 2.6 mol). The crude
product was resuspended in ethanol (3 L) and refluxed for 2
h; the mixture was then cooled at room temperature and
refrigerated at T ) 5 °C overnight. The resulting solid was
vacuum-filtered, washed with cooled ethanol (3 × 0.25 L),
and dried under vacuum at T ) 80 °C to give DF1012 (0.718
kg, 2.53 mol) as a white solid in 78% yield: mp ) 281-
1
106 °C, H NMR (CDCl3) δ 6.55 (d, 1H, J ) 3.0 Hz), 7.1
(dd, 1H J1 ) 8.0 Hz, J2 ) 4.7 Hz), 7.45 (d, 1H, J ) 3.0
Hz), 7.98 (m, 1H), 8.32 (bs, 1H), 9.7 (bs, 1H); IR (Nujol)
3040, 1580, 1280, 1120, 890, 760 cm-1. Anal. Calcd for
C7H6N2: C, 71.17; H, 5.12; N, 23.71. Found: C, 71.20; H,
5.11; N, 23.69. ESI-MS m/z 119 [M + H]+.
3-Endo-amino-8-methyl-8-azabicyclo[3.2.1.]octane Bis-
hydrochloride (6‚Bis-hydrochloride). In a 50-L glass
reactor under nitrogen atmosphere 3 (600 g, 4.31 mol) was
dissolved in 90% aqueous MeOH (15 L) and stirred with
ammonium formate (2500 g, 39.64 mol). After complete
dissolution, 10% Pd/C (0.15 kg, 0.14 mol) was added, and
the reaction mixture was stirred for 3 h. The catalyst was
filtered under vacuum on a Celite panel and washed with
90% MeOH. The filtrate was concentrated under reduced
pressure in a 25-L glass reactor. The so-obtained oil was
dissolved in absolute ethyl alcohol (10 L), and 37% HCl
(0.75 L) was added dropwise while stirring. The solution
was seeded and left stirring at room temperature for 1 h and
for 5 h at 4 °C.
1
282 °C; H NMR (DMSO-d6) δ 1.9-2.2 (m, 8H), 2.25 (s,
3H) 3.12 (bs, 2H), 4.05 (m, 1H), 7.25 (dd, 1H, J1 ) 7.9, J2
) 4.6), 7.50 (d, 1H, J ) 4.6 Hz), 8.20 (s, 1H), 8.43 (d, 1H),
8.5 (d, 1H), 12.1-12.3 (bs, 1H); IR (Nujol) 3300, 3100,
1620, 1530, 1290, 1200 cm-1. Anal. Calcd for C16H20N4O:
C, 67.58; H, 7.09; N, 19.70. Found: C, 67.60; H, 7.07; N,
19.71. ESI-MS m/z 285 [M + H]+.
The precipitate was filtered, washed with absolute ethyl
alcohol, and desiccated at 40 °C in a vacuum oven to give
6‚bis-hydrochloride (760 g, 3.56 mol) as a white solid in
Acknowledgment
1
82.7% yield: mp > 250 °C; H NMR (DMSO-d6) δ 2.4-
We sincerely thank Professor Carmelo Gandolfi in whose
memory this paper is dedicated.
2.05 (m, 6H), 2.8-2.55 (m, 5H), 3.7-3.5 (m, 1H), 4.0-3.8
(bs, 2H), 8.7-8.2 (bs, 3H), 11.2-11.0 (bs, 1H). IR (Nujol)
3300, 3210, 3150, 2700, 2560, 1550, 1040 cm-1. Anal. Calcd
for C8H18N2Cl2: C, 45.08; H, 8.51; N, 13.14; Cl, 33.26.
Found: C, 45.09; H, 8.50; N, 13.12; Cl, 33.26. ESI-MS m/z
141 [M + H]+.
Received for review July 18, 2002.
OP025570T
Vol. 7, No. 2, 2003 / Organic Process Research & Development
•
213