A facile three-step synthesis of (±)-crispine A via an acyliminium ion cyclisation

Article abstract of DOI:10.1016/j.tet.2006.12.041

A high yielding cyclisation of the readily available N-(4,4-diethoxybutyl)-2-(3,4-dimethoxyphenyl)acetamide to 8,9-bis(methyloxy)-2,3,6,10b-tetrahydropyrrolo[2,1-a]isoquinolin-5(1H)-o ne is described. The latter can be reduced with either AlH₃

Full text of DOI:10.1016/j.tet.2006.12.041

Tetrahedron 63 (2007) 2053–2056
A facile three-step synthesis of ( )-crispine A via an
acyliminium ion cyclisation
,y
*
Frank D. King
Neurology and GI CEDD, GlaxoSmithKline, New Frontiers Science Park (North), Third Avenue, Harlow, Essex CM19 5AW, UK
Received 6 October 2006; revised 29 November 2006; accepted 14 December 2006
Available online 17 December 2006
Abstract—A high yielding cyclisation of the readily available N-(4,4-diethoxybutyl)-2-(3,4-dimethoxyphenyl)acetamide to 8,9-bis(methyl-
oxy)-2,3,6,10b-tetrahydropyrrolo[2,1-a]isoquinolin-5(1H)-one is described. The latter can be reduced with either AlH3 or BH3 to (ꢀ)-crisp-
ine A in an overall yield of 55%.
Ó 2006 Elsevier Ltd. All rights reserved.
1. Introduction
cyclisation of a-ethoxyamide 3a. TiCl4-mediated5 cyclisa-
tions of a-methoxyamides have been previously reported,
but no optimisation of this cyclisation had been undertaken.7
It was anticipated that 3a should be obtainable from the
amido-aldehyde diethylacetal 4.
Over the past 30 years, cyclic acyliminium ions have pro-
vided a fertile area for the synthesis of alkaloids and other
substituted azacycles.1 The most common method of synthe-
sising the cyclic acyliminium ions is by acid treatment of
a-alkoxyamides, in turn prepared either by the reduction
of imides or by the anodic oxidation of amides. Some time
ago we published the in situ generation of cyclic iminium
ions by the intramolecular condensation of an aminoalde-
hyde diethylacetal to give a facile, one-pot synthesis of indo-
lizidines and quinolizidines.2,3 However, in the intervening
time, this simple generation of iminium ions from readily
available starting materials has been little explored. In this
paper, we apply this methodology to the generation of a
cyclic acyliminium ion to give a simple, high yielding,
three-step synthesis of the alkaloid (ꢀ)-crispine A, 1, from
the commercially available 4-aminobutyraldehyde diethyl-
acetal (Scheme 1).4 Crispine A was selected as a test of the
methodology, as it had been reported that the anodic oxida-
tion of phenacylpyrrolidines with a p-methoxy substituent
oxidised on the benzylic position, rather than the amide
a-position.5 In addition, there has recently been an upsurge
of interest in the synthesis of crispine A due to the potential
use of the close analogue, crispine B, in oncology.6
Compound 4 was readily prepared from either 3,4-di-
methoxyphenacylchloride 5a, or the hydroxysuccinimide
(HOSu) ester 5b. The acetal 4 was found to be extremely
acid sensitive. Attempts to wash out any excess base from
reaction mixtures in either ethyl acetate or dichloromethane
with either 2 M HCl or 1 M H2SO4 resulted in partial cleav-
age of the acetal group. The crude 4 solidified on drying
under high vacuum and a portion of this solid was used to
seed the recrystallisation from Et2O/hexane. A slightly less
pure product was obtained from 5a, although in higher yield
(80% vs 72%), as the coloured impurities in the commer-
cially available material carried through into the product,
even after recrystallisation.
All attempts to convert 4 directly into 3a via acid catalysis
were unsuccessful. However, 4 could be converted to 3b in
effectively quantitative yield by rapidly stirring an Et2O
solution of 4 with 2 M HCl. Surprisingly, much of 3b was ex-
tracted into the 2 M HCl solution and advantage was taken of
this for purification. The Et2O layer as diluted with an equal
volume of hexane, and 3b was extracted into H2O. The com-
bined aqueous extracts were then basified with solid K2CO3
and 3b back-extracted into CH2Cl2. Consistent with the lit-
erature on related N-acyl 2-hydroxypyrrolidines, the NMR
spectrum of 3b indicates that it exists in equilibrium with
about 10% of the ring-opened amido-aldehyde, which prob-
ably prevents crystallisation. It was noted that 3b is also
acid-labile, with the CDCl3 NMR solution decomposing over
a period of few weeks, presumably due to the presence of
a small amount of acid. Compound 3b has also been reported
2. Results and discussion
The retrosynthetic strategy was to aim for the lactam 2,
which should be obtainable from the acid-catalysed
Keywords: (ꢀ)-Crispine A; Acyliminium; Cyclisation.
* Tel.: +44 (0)207 6794583; e-mail: fd.king@btinternet.com
Present address: Department of Chemistry, University College London,
20 Gordon Street, London WC1H 0AJ, UK.
y
0040–4020/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2006.12.041

2054
F. D. King / Tetrahedron 63 (2007) 2053–2056
OEt
MeO
MeO
O
O
MeO
MeO
(a)
(d)
OEt
OEt
+
H2N
OEt
X
HN
4 R = Cl 80%
5a R = Cl
R = OSu 72%
(b), (c)
5b R = OSu
MeO
MeO
O
MeO
MeO
MeO
MeO
O
(d)
(e)
N
N
OR
N
( ) crispine A 1 80%
3a R = Et ~100%
3b R = H ~100%
2 90% from 3a
87% from 4
Scheme 1. (a) Either X¼Cl, 1.1 equiv Et3N/CH2Cl2 or X¼OSu, EtOAc; (b) 2 M HCl, Et2O for 30 min, add hexane, extract product into H2O, then K2CO3,
˚
CH2Cl2 (see text); (c) 0.1 equiv TFA, 3 A molecular sieves, Et2O; (d) 1.5 equiv AlCl3, CH2Cl2 2 h at rt and (e) LAH, 0.5 equiv concd H2SO4, THF.
as a non-isolated intermediate, prepared by ozonolysis of an
olefin and treatment with dimethyl sulfide. Subsequent cyc-
lisation with formic acid gave 2 in 60% yield.8 Treatment of
our isolated 3b with formic acid did, indeed, give a cyclic
product. LC/MS of the crude product indicated w25% of a
dimer, which probably prevented efficient crystallisation.
However, column chromatography (SiO2; CH2Cl2/2%
MeOH) resulted in the isolation of 2 as a crystalline solid
with the quoted literature yield.
purification of the product required other than Et2O tritura-
tion. In addition, formic acid cyclisation of 4 afforded 2 in
exactly the same manner as with 3a.
Finally, the reduction of 2 to 1 was investigated. Using LAH,
the addition of 2 gave a rapid evolution of gas, presumably
hydrogen, and subsequent reduction was slow, with multiple
products being formed. Presumably under the basic condi-
tions, the enolate was rapidly formed and deactivated the
system to reduction. Diborane had previously been used
on closely related lactams.10 Following this reported proce-
dure, a 77% yield of 1 was obtained, but refluxing acid cleav-
age of the borane adduct of the product was required. We had
previously used alane, formed from adding ½ molar equiva-
lents of concd H2SO4 to LAH in THF,11 for the reduction of
oxime to avoid anion formation. Reduction of 2 with alane
gave a high yield (80%) of 1, but with a much simpler aque-
ous base treatment of the reaction mixture.
The a-ethoxyamide 3a was obtained from 3b in effectively
quantitative yield by treatment with a catalytic quantity of
˚
acid in the presence of EtOH and 3 A molecular sieves to
capture the water. The H NMR spectrum of 3a showed
1
two equal triplets for the ethyl CH3, each integrating for
1.5 protons, and two doublets for the CHOEt proton, each
integrating for 0.5 protons, consistent with a restricted rota-
tion of the amide bond with a 1:1 ratio of rotamers. The 13C
spectrum similarly showed a doubling of most of the signals.
Various acid-catalysed cyclisation conditions of 3a were
investigated. Most BrTnsted acids attempted (TFA, TFMSA,
acetic acid, H2SO4 and HCl) did not give the expected
cyclisation, but a product resulting from dimerisation of
the acyliminium ion was obtained. However, the Lewis acids
AlCl3 and TiCl4 were both successful. More than 1 equiv of
Lewis acid was required for successful cyclisation. Optimi-
sation with AlCl3 showed that more than 2 equiv resulted
in the formation of an oily precipitate and, finally, a cyclic
product contaminated with demethylated phenol product
(which could be removed with a 2 M NaOH wash);
1.5–2 equiv consistently gave a relatively clean reaction,
with the product being obtained in >95% purity from a sim-
ple Et2O trituration of the worked-up crude product inw90%
yield. TiCl4 gave similar results and it was noted that the
dropwise addition of a CH2Cl2 solution of TiCl4 could
be made at 0 ꢁC, rather than the reported ꢂ78 ꢁC.5 Also
1.5 equiv was found to be as effective as the reported
2.8 equiv.
In conclusion, a facile, three step, high yielding (55% over-
all) synthesis of (ꢀ)-crispine Avia an acyliminium ion is de-
scribed. The scope of this acyliminium ion formation is
being further investigated.
3. Experimental
3.1. General
1
Melting points are uncorrected. H NMR spectra were re-
corded on a Bruker AM-400 spectrometer using Me4Si as
an internal standard. 13C NMR spectra were recorded on a
Bruker AMX300 spectrometer. All solvents and reagents
were of commercial grade and used without purification.
All solvent evaporations were carried out under reduced
pressure. All compounds and crude reaction mixtures were
analysed by LC/MS using an HP1050 LC system operating
under reverse phase using a 3.3 cmꢃ4.6 mm ID, 3 mm
ABZ+PLUS column with a 5 min run, starting with 0.1%
formic acid+10 mM ammonium acetate 95% and finishing
with acetonitrile+0.05% formic acid, with a UV detection
range of 215–330 nm. MS detection used a Waters ZQ mass
spectrometer and all products were >98% pure. HRMS mea-
surements were performed with a Micromass Q-Tof 2 hybrid
quadrupole time-of-flight mass spectrometer, equipped with
a Z-spray interface and the elemental composition was cal-
culated using MassLynx v4.0.
Having confirmed the cyclisation of 3a, the Lewis-acid
mediated cyclisation was attempted directly on 4.9 Surpris-
ingly the cyclisation of 4 was as effective as that of 3a. Op-
timisation of the cyclisation of 4 using AlCl3 gave similar
results to 3a. Using more than 2 equiv again resulted in de-
methylation. More than 1 equiv was required, but amounts
close to 1 equiv again gave a much slower, less clean reac-
tion. Approximately 1.5 equiv was highly effective, with no

F. D. King / Tetrahedron 63 (2007) 2053–2056
2055
3.1.1. Synthesis of N-(4,4-diethoxybutyl)-2-(3,4-dimeth-
oxyphenyl)acetamide (4). Method (A). DCC (7.3 g,
0.035 mol) was added in one portion to a stirred solution
of (3,4-dimethoxyphenyl)acetic acid (6.9 g, 0.035 mol) and
HOSu (4.2 g, 0.036 mol) in EtOAc (250 mL), at ambient
temperatures under argon. After stirring for 3 h, a solution
of 4-aminobutyraldehyde diethylacetal (ex. Aldrich, 90%
pure) (6 mL, 0.035 mol) in EtOAc (25 mL) was added in
one portion. The reaction mixture was stirred over night.
Et2O (200 mL) was then added and the precipitated DCU
collected by filtration, and washed with Et2O (100 mL).
The combined filtrates were washed with 2 M NaOH
(50 mL), H2O (50 mL) and brine (50 mL), and then dried
(MgSO4), filtered and concentrated by rotary evaporation.
On further drying on a high vacuum line, the oil solidified.
The solid was triturated, initially with Et2O (50 mL), and
subsequently hexane (150 mL) was added in portion. The
white solid was collected, washed with hexane (100 mL)
and dried to give 4 as a white crystalline solid (8.5 g,
(400 MHz, CDCl3): d 1.7–2.2 (m, 4H), 3.59 (s, 2H), 3.87
(br s, 6H), 4.12 (br s, 1H), 5.66 (br s, 1H), 6.77–6.90 (m,
3H), 9.74 (br s, 0.1H). 13C NMR+DEPT (75 MHz, CDCl3)
d¼23.23 (CH2), 32.04 (CH2), 41.63 (CH2), 46.67 (CH2),
55.87 (CH3), 55.90 (CH3), 81.90 (CH), 111.24 (CH),
112.16 (CH), 121.14 (CH), 126.55 (C), 148.04 (C), 149.07
(C), 171.70 (C). The crude 3b (6.7 g) was dissolved in
CH2Cl2 (150 mL) and stirred at ambient temperatures with
˚
EtOH (4 mL), 3 A molecular sieves (25 g) and TFA
(0.5 mL) for 3.5 h. Et3N (1 mL) was added and the reaction
mixture filtered through a pad of Kieselghur, washing the
collected solids with CH2Cl2 (2ꢃ100 mL). The combined
organics were washed with H2O (50 mL) and brine
(50 mL), and dried (MgSO4). Concentration of the filtered
organics under reduced pressure gave 3a as a colourless
oil (6.9 g,w100%). 1H NMR (400 MHz, CDCl3): d 1.16 (t,
J¼7.2 Hz, 1.5H), 1.27 (t, J¼7.2 Hz, 1.5H), 1.64–2.21 (m,
4H), 3.31–3.72 (m, 6H), 3.86 (s, 6H), 5.10 (d, J¼4.8 Hz,
0.5H), 5.55 (d, J¼4.8 Hz, 0.5H), 6.77–6.88 (m, 3H). 13C
NMR+DEPT (75 MHz, CDCl3) d¼15.37 (CH3), 15.44
(CH3), 21.06 (CH2), 23.08 (CH2), 31.55 (CH2), 31.88
(CH2), 40.70 (CH2), 41.76 (CH2), 45.86 (CH2), 46.31
(CH2), 55.85 (CH3), 55.91 (CH3), 62.08 (CH2), 64.46
(CH2), 85.94 (CH), 87.51 (CH), 111.18 (CH), 111.23
(CH), 112.19 (CH), 112.33 (CH), 121.15 (CH), 121.30
(CH), 127.04 (C), 127.66 (C), 147.92 (C), 148.97 (C),
149.03 (C), 171.02 (C), 171.32 (C). HRMS: calcd for
C16H23NO4Na [M+Na]+ 316.1525, found 316.1530.
1
72%); mp 56–57 ꢁC. H NMR (400 MHz, CDCl3): d 1.17
(t, J¼7.2 Hz, 6H), 1.49–1.60 (m, 4H), 3.23 (m, 2H), 3.40–
3.48 (m, 2H), 3.51 (s, 2H), 3.55–3.63 (m, 2H), 3.88 (s,
6H), 4.43 (t, J¼5.2 Hz, 1H), 5.53 (br s, 1H), 6.77 (s, 1H),
6.78 (d, J¼8.4 Hz, 1H), 6.84 (d, J¼8.4 Hz, 1H). 13C
NMR+DEPT (75 MHz, CDCl3) d¼12.27 (CH3), 24.51
(CH2), 30.79 (CH2), 39.25 (CH2), 43.37 (CH2), 55.86
(CH2), 61.09 (CH3), 61.26 (CH3), 102.49 (CH), 111.50
(CH), 112.42 (CH), 121.56 (C), 127.46 (C), 148.26 (C),
149.23 (C), 171.31 (C). HRMS: calcd for C18H29NO5Na
[M+Na]+ 362.1943, found 362.1949. The mother liquors
from the crystallisation contained 4 (w2 g), which could be
used for the subsequent hydrolysis reaction.
3.1.3. 8,9-Bis(methyloxy)-2,3,6,10b-tetrahydropyr-
rolo[2,1-a]isoquinolin-5(1H)-one (2). Method (A). AlCl3
(3.5 g, 0.026 mol) was added to a stirred solution of 3a
(5.0 g, 0.017 mol) in dry CH2Cl2 (100 mL) at ambient
temperatures under argon. After 2 h, ice (50 g) was added
and the reaction stirred for a further 15 min. The organic
layer was separated and the aqueous layer extracted
with CH2Cl2 (50 mL). The combined organics were dried
(MgSO4), filtered and concentrated in vacuo. Trituration
with Et2O (50 mL) gave 2 (3.8 g, 90%); mp 164–165 ꢁC
Method (B). A solution of (3,4-dimethoxyphenyl)acetyl
chloride (ex. Aldrich) (10 g, 0.047 mol) in anhydrous
CH2Cl2 (20 mL) was added, dropwise, over 20 min to stirred
solution of 4-aminobutyraldehyde diethylacetal (7.8 mL,
0.045 mol) and Et3N (6.5 mL, 0.047 mol) in anhydrous
CH2Cl2 (200 mL) at 0 ꢁC. The reaction mixture was stirred
and allowed to warm to ambient temperatures over 2 h.
The CH2Cl2 was removed by rotary evaporation and re-
placed with Et2O (300 mL). After washing with H2O
(2ꢃ50 mL) and brine (50 mL), the separated organic layer
was dried (MgSO4), filtered and concentrated by rotary
evaporation. The resultant oil was re-dissolved in Et2O
(200 mL) and hexane added until the mixture was slightly
cloudy. A seed crystal from Method (A) was added to initiate
crystallisation. Additional hexane was added to approxi-
mately 1:1 Et2O/hexane and the solid collected, washed
with hexane (2ꢃ100 mL) and dried to give 4 (12.2 g, 80%)
as a beige-coloured solid.
1
(lit.8 163–165 ꢁC). H NMR (400 MHz, CDCl3): d 1.83–
1.93 (m, 1H), 1.96–2.09 (m 1H), 2.12–2.19 (m, 1H), 2.58–
2.64 (m, 1H), 3.46–3.71 (m, 4H), 3.88 (s, 3H), 3.89 (s,
3H), 4.57–4.62 (m, 1H), 6.67 (s, 1H), 6.68 (s, 1H). 13C
NMR+DEPT (75 MHz, CDCl3) d¼23.09 (CH2), 31.81
(CH2), 38.14 (CH2), 44.72 (CH2), 56.06 (CH3), 56.19
(CH3), 59.56 (CH), 107.69 (CH), 111.27 (CH), 124.93 (C),
127.99 (C), 147.88 (C), 148.56 (C), 167.60 (C). HRMS:
calcd for C14H17NO3 [MH]+ 248.1287, found 248.1291.
Method (B). The method was identical to Method (A) except
that 4 (1.0 g, 0.003 mol) was used to give 2 (0.65 g, 87%),
identical to that obtained from Method (A).
3.1.2. 1-{[3,4-Bis(methyloxy)phenyl]acetyl}-2-(ethyloxy)-
pyrrolidine (3a). A solution of 4 (8.4 g, 0.025 mol), dis-
solved in EtOAc (20 mL) and Et2O (100 mL) was stirred
vigorously at ambient temperatures with aqueous 2 M HCl
(50 mL) for 1 h. Hexane (100 mL) was then added and the
aqueous layer separated. The organics were washed with
H2O (2ꢃ50 mL) and the combined aqueous extracts basified
with solid K2CO3. The product (3b) was extracted into
CH2Cl2 (3ꢃ100 mL), dried (MgSO4) and concentrated
under reduced pressure to give 3b as an oil in equilibrium
withw10% of the amidoaldehyde (6.7 g,w100%). 1H NMR
3.1.4. 8,9-Bis(methyloxy)-1,2,3,5,6,10b-hexahydropyr-
rolo[2,1-a]isoquinoline [( )-crispine A] (1). Concd
H2SO4 (0.13 mL, 0.0025 mol) was added, dropwise over
5 min, to a stirred solution of LAH (5 mL of a 1 M solution,
0.005 mol) at 0 ꢁC under argon. After stirring at 0 ꢁC for
15 min, 2 (0.33 g, 0.0013 mol) was added in one portion
and the reaction stirred at room temperature for 14 h. The re-
action mixture was then cooled to 0 ꢁC and 2 N NaOH
(0.6 mL) was then carefully added and the reaction mixture
stirred for 15 min. The solid was collected and washed with

2056
F. D. King / Tetrahedron 63 (2007) 2053–2056
4. For recent syntheses, see: KnTlker, H.-J.; Agarwal, S.
Tetrahedron Lett. 2005, 46, 1173–1175 and Meyer, N.;
Opatz, T. Eur. J. Org. Chem. 2006, 3997–4002.
Et2O (2ꢃ30 mL). The combined organics were then evapo-
rated in vacuo and the residue purified by column chromato-
graphy (Al2O3, CH2Cl2/2% MeOH). Crystallisation from
hexane afforded the pure 1 (0.25 g, 80%); mp 88–89 ꢁC
(lit.12 88–89 ꢁC). 1H NMR (400 MHz, CDCl3): d 1.63–
1.98 (m, 3H), 2.28–2.36 (m, 1H), 2.50–2.76 (m, 3H), 2.99–
3.11 (m, 2H), 3.16–3.21 (m, 1H), 3.41 (br t, J¼8 Hz), 3.83 (s,
3H), 3.84 (s, 3H), 6.57 (s, 1H), 6.61 (s, 1H).
5. Moeller, K. D.; Wang, P. W.; Tarazi, S.; Marzabadi, M. R.;
Wong, P. L. J. Org. Chem. 1991, 56, 1058–1067.
6. Szawka1o, J.; Zawadzka, A.; Wojtasiewicz, K.; Leniewski, A.;
Drabowicz, J.; Czarnocki, Z. Tetrahedron: Asymmetry 2005,
16, 3619–3621 and references therein; Kapat, A.; Sen, S. K.;
Baskaran, S. Annual IIT Madras Chemistry Symposium; July
12–13, 2006; 40 p; Zhang, Q.; Tu, G.; Zhao, Y.; Cheng, T.
Tetrahedron 2002, 58, 6795–6798.
Acknowledgements
7. Moeller, K. D., personal communication.
8. Kano, S.; Yuasa, Y.; Shibuya, S. Synth. Commun. 1985, 15,
883–889.
9. For a similar reaction on solid phase, see: Veerman, J. J. N.;
Klein, J.; Aben, R. W. M.; Scheeren, H. W.; Kruse, C. G.;
van Maarseveen, J. H.; Rutjes, F. P. J. T.; Hiemstra, H. Eur. J
Org. Chem. 2002, 3133–3139.
I am grateful to the GSK Neurology and GI CEDD for fund-
ing this research, to Prof. S. Caddick for the recent use of
his laboratory and to many of my colleagues for sharing my
enthusiasm.
References and notes
10. Maryanoff, B. E.; McComsey, D. F.; Gardocki, J. F.; Shank,
R. P.; Costanzo, M. J.; Nortey, S. O.; Schneider, C. R.; Setler,
P. E. J. Med. Chem. 1987, 30, 1433–1454.
11. King, F. D.; Martin, R. T. Tetrahedron Lett. 1991, 32, 2281–
2284.
12. Bremner, J. B. Aust. J. Chem. 1984, 37, 1203–1215.
1. Speckamp, W. N.; Moolenaar, M. J. Tetrahedron 2000, 56,
3817–3856 and references therein.
2. King, F. D. Tetrahedron Lett. 1983, 24, 3281–3282.
3. King, F. D. J. Chem. Soc., Perkin Trans. 1 1986, 447–453.