Asymmetric synthesis of the erythrinan alkaloid system using a chiral lithium amide base desymmetrisation as the key step

Article abstract of DOI:10.1016/j.tetlet.2003.08.078

A new asymmetric approach to the erythrinan alkaloid system is described, which involves chiral base desymmetrisation of a ring fused imide and a 6-exo-trig radical cyclisation as the key steps.

Full text of DOI:10.1016/j.tetlet.2003.08.078

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 7803–7807
Asymmetric synthesis of the erythrinan alkaloid system using a
chiral lithium amide base desymmetrisation as the key step
Christopher Gill, Daniel A. Greenhalgh and Nigel S. Simpkins*
School of Chemistry, The University of Nottingham, University Park, Nottingham NG7 2RD, UK
Received 30 July 2003; accepted 14 August 2003
Abstract—A new asymmetric approach to the erythrinan alkaloid system is described, which involves chiral base desymmetrisation
of a ring fused imide and a 6-exo-trig radical cyclisation as the key steps.
© 2003 Elsevier Ltd. All rights reserved.
The pyrrolo[2,1-a]isoquinoline structure is a significant
motif present in the erythrina alkaloid group,1 as well as
representing a biologically active system in its own right.2
Lete and co-workers have proposed that unsaturated
pyrrolo[2,1-a]isoquinolones of general structure 2 could
represent useful precursors to the complete erythrinan
skeleton 1 by cyclisation of the functionalised side chain
onto the lactam ring in a Michael type fashion.3–5
Although this research group has published a series of
studies aimed at achieving this objective, it has not been
accomplished to date, as far as we are aware.
We became interested in the asymmetric synthesis of
structures 2, since it seemed possible to access such
intermediates by our recently developed chiral lithium
amide base-mediated desymmetrisation of ring-fused
imides.6,7 Our approach is outlined in retrosynthetic form
in Scheme 1.
Scheme 1.
Keywords: chiral lithium amide base; erythrinan alkaloid; radical cyclisation.
* Corresponding author.
0040-4039/$ - see front matter © 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2003.08.078

7804
C. Gill et al. / Tetrahedron Letters 44 (2003) 7803–7807
Scheme 2.
At this stage the absolute configuration of 8 was
assigned by analogy to our previous work on the
corresponding N-phenyl imide,6 but this was later
confirmed by correlation with known compounds (see
below). Reaction of excess Grignard reagent with imide
8 occurred with very high regioselectivity (none of the
Access to the required intermediates 2 would be by
retro-Diels–Alder reaction of the cyclopentadiene
adduct 3, which in turn would arise by cyclisation of 4
via an N-acyliminium ion intermediate. Compound 4
would be prepared by addition of a suitable Grignard
reagent to imide 5, which would be available in high
levels of enantiopurity from the corresponding simple
meso-imide by application of our earlier chiral base
chemistry.
1
minor regioisomer could be observed by H NMR) to
give hydroxylactam intermediates 9 or 10. Cyclisation
of these intermediates was best accomplished after desi-
lylation, to give the pentacyclic products 11 and 12 in
good yields and with complete diastereocontrol. In the
case of 10 it was especially significant that clean cyclisa-
tion of the aromatic group occurred onto the N-
acyliminium intermediate, without interference from
the alkenyl side-chain that would lead to spiro-fused
products. This was a key aspect of our choice of a
relatively unreactive side chain, since Lete and co-work-
ers had previously demonstrated that compounds with
an alkenylsilane side chain show the unwanted (in this
context) spiro mode of cyclisation.
In the forward sense we expected that the stereochemi-
cal control would arise from initial Grignard reaction
distal to the controlling trimethylsilyl group of 5, fol-
lowed by exo-selective cyclisation of the aromatic
appendage onto the tricyclic nucleus. The final cyclisa-
tion of 2 to give 1 appeared constrained to give only a
single cyclohexane product with respect to the new ring
fusion. Since Lete and co-workers had demonstrated
the viability of this type of N-acyliminium ion cyclisa-
tion and retro Diels–Alder process,3–5 the key issues to
be addressed were the nature of the functionality FG
and choice of reaction mode for the final cyclisation.
The identity of intermediate 11 was then confirmed by
Diels–Alder cycloreversion by heating under reduced
pressure (Scheme 3).
This approach would enable a very rapid asymmetric
access to the key pyrrolo[2,1-a]isoquinolones, and
might allow us to incorporate a choice of side chains
appropriate for completion of the erythrinan skeleton.
Also, the chiral base approach would allow the prepa-
ration of either enantiomeric series. Here we describe
the successful implementation of this plan, which has
resulted in a short, completely stereocontrolled route to
a functionalised tetracyclic intermediate with great
potential for synthesis of naturally occurring alkaloids.
This gave the known unsaturated lactam 13, which had
spectroscopic data in accord with those published by
Lete and co-workers. We were also able to confirm our
Chiral base reaction of imide 6 using the bis-lithium
amide 7 gave the desired silylated product (−)-8 in 90%
yield and 91% ee, Scheme 2.8
Scheme 3.

C. Gill et al. / Tetrahedron Letters 44 (2003) 7803–7807
7805
Scheme 4.
assignment of absolute stereochemistry by comparison
of [h]D data.9
With the key stereochemical features of the synthesis
established we then transformed the lactam with the
appended unsaturated side chain as shown in Scheme 4.
Retro-Diels–Alder of 12 reaction proceeded to give 14,
the side chain of which was then cleaved by a conven-
tional two-step oxidation, to give aldehyde 15. Lete and
co-workers had previously conceived a ring closure of
this type of system via a dithiane anion, but we chose
instead to conduct a reductive cyclisation under free
radical conditions.
Considering the well-established repertoire of transfor-
mations established for intermediates related to these
compounds, we believe that the functional group ‘han-
dles’ installed in our intermediates should be attractive
for the synthesis of natural products and their rela-
tives.15 Thus, we anticipate that the strategy described
above should be a fruitful source of enantiomerically
pure alkaloids of this family. Extension of this desym-
metrisation strategy to include other types of alkaloid is
planned.
Thus, exposure of aldehyde 15 to tributyltin hydride,
under conditions described by Muller et al. for cyclisa-
tion of 5%-aldehyde nucleosides led efficiently to the
desired hydroxy lactam 16 as
a
mixture of
diastereomers at the newly formed carbinol centre.10–12
This outcome was especially welcome, since successful
6-exo-trig closures are rare compared to the 5-exo-trig
variants, and is probably due in part to the lack of a
competing 1-5 hydrogen atom abstraction process in
our system.
Acknowledgements
We thank the Engineering and Physical Sciences
Research Council (EPSRC) for support of C.D.G., and
we thank the University of Nottingham for support of
D.A.G.
In order to transform the mixture of alcohols 16 (ca.
3:1 mixture) into a single substance we also carried out
the oxidation of the separated compounds, each of
which gave the same ketone product 17.13,14
The formation of 16 represents the completion of the
full erythrinan skeleton in a highly stereocontrolled
fashion and a relatively short overall sequence (eight
steps from simple imide 6). Naturally occurring alka-
loids (which are of the opposite enantiomeric series to
the ones made here), often incorporate unsaturation in
the A and B rings, e.g. erysotrine 18 and erysotamidine
19, and may also possess C-11 oxygenation.1
References
1. For general reviews, see: (a) Bentley, K. W. The Isoquino-
line Alkaloids; Harwood Academic, 1998; (b) Bentley, K.
W. Nat. Prod. Rep. 2002, 19, 332. For leading synthesis
references, see: (c) Allin, S. M.; James, S. L.; Elsegood,
M. R. J.; Martin, W. P. J. Org. Chem. 2002, 67, 9464; (d)

7806
C. Gill et al. / Tetrahedron Letters 44 (2003) 7803–7807
Hosoi, S.; Nagao, M.; Tsuda, Y.; Isobe, K.; Sano, T.;
Ohta, T. J. Chem. Soc., Perkin Trans. 1 2000, 1505.
2. (a) For leading references, see Allin, S. M.; James, S. L.;
Martin, W. P.; Smith, T. A. D. Tetrahedron Lett. 2001,
42, 3943; (b) Zhang, Q.; Tu, G.; Zhao, Y.; Cheng, T.
Tetrahedron 2002, 58, 6795. See also Ref. 4.
isoquinolinone 13 as a pale yellow oil (40 mg, 95%): [h]D20
+222 (c 0.39 in CHCl3), lit.3 [h]D20 +202.8 (c 1.43 in
CHCl3); wmax (CHCl3)/cm−1 2961, 2935 and 2852 (C-H),
1680 (CꢀO), 1613 and 1503 (Ar); lH (500 MHz) 1.61 (3H,
s, CH3), 2.68 (1H, dd, J 16.0, 4.3, CH2CHAHBAr), 2.94
(1H, ddd, J 16.0, 12.0, 6.5, CH2CHAHBAr), 3.24 (1H,
ddd, J 13.2, 12.0, 4.3, CHAHBCH2Ar), 3.85 (3H, s,
OCH3), 3.90 (3H, s, OCH3), 4.44 (1H, dd, J 13.2, 6.5,
CHAHBCH2Ar), 6.12 (1H, d, J 5.8, HCꢀCH), 6.61 (1H, s,
ArCH), 6.65 (1H, s, ArCH), 7.36 (1H, d, J 5.8, HCꢀCH);
lC (125.8 MHz) 26.8 (CH3), 29.3 (CH2CH2Ar), 34.7
(CH2CH2Ar), 55.9 (OCH3), 56.2 (OCH3), 65.6 (CNMe),
108.7 (ArCH), 111.9 (ArCH), 125.1 (ArC), 125.2
(HCꢀCH), 129.1 (ArC), 147.6 (ArC), 148.0 (ArC), 153.3
(HCꢀCH), 170.3 (CꢀO); MS (EI) m/z 259 (M+, 49%), 245
(45), 244 (M−CH3, 100), 228 (M−OCH3, 12), 200 (20),
122 (14), 57 (14) (HRMS: found M+, 259.1203.
C15H17NO3 requires M, 259.1209).
3. Gonzalez-Temprano, I.; Sotomayor, N.; Lete, E. Synlett
2002, 593.
4. Manteca, I.; Sotomayor, N.; Villa, M.-J.; Lete, E. Tetra-
hedron Lett. 1996, 37, 7841.
5. Manteca, I.; Etxarri, B.; Ardeo, A.; Arrasate, S.; Osante,
I.; Sotomayor, N.; Lete, E. Tetrahedron 1998, 54, 12361.
6. Adams, D. J.; Blake, A. J.; Cooke, P. A.; Gill, C. D.;
Simpkins, N. S. Tetrahedron 2002, 58, 4603.
7. Simpkins, N. S.; Gill, C. D. Org. Lett. 2003, 5, 535.
8. In previous work (Ref. 6) we used a simpler chiral base
(bis-phenethylamine) to access N-phenyl and N-benzyl
derivatives related to 8 in very high yields and ee. With
the modified nitrogen substituent required for the present
synthesis we were unable to obtain high chemical yields
using that procedure, and found that the dilithiated chiral
base 7 gave the best compromise in terms of workable
synthetic yield without too great a loss of ee.
The ee of this product was determined to be 85% by
HPLC (Chiracel OD column, 20% IPA in hexane, 0.4
mL/minute), the retention times were 25.8 min (minor)
and 31.7 min (major). We need to check further if the
apparent slight loss of stereochemical integrity (91% ee to
85% ee) is real.
Typical procedure for the silylation of 6 to give 8.
Chiral lithium amide base 7 was prepared from the
corresponding chiral amine (622 mg, 1.48 mmol) in THF
(6 mL) and n-BuLi (1.9 mL of a 1.5 M solution in
hexanes, 2.89 mmol). The resulting solution of the chiral
base 7 was then cooled to −100°C, before being added
dropwise via cannula over 15 min to a stirred solution of
the imide 6 (402 mg, 1.23 mmol) and freshly distilled
trimethylsilyl chloride (1.57 mL, 12.3 mmol) in THF (12
mL), maintaining a temperature of −100°C 1. The reac-
tion mixture was then allowed to warm to room tempera-
ture overnight, before quenching with saturated aqueous
NaHCO3 (10 mL) and extraction with EtOAc (3×50 mL).
The extracts were combined, washed with brine (100 mL),
dried (MgSO4), filtered and concentrated in vacuo. The
product was purified by flash column chromatography on
silica gel (gradient 10% EtOAc/petroleum ether to 40%
EtOAc/petroleum ether) to yield the imide 8 as a pale
yellow oil (443 mg, 90%): [h]2D0 −46 (c 1.71 in CHCl3);
HRMS (EI): found M+, 399.1854. C22H29NO4Si requires
M, 399.1866. The ee was determined as 91% by HPLC
(Chiracel OD column, 1% IPA in hexane, 0.8 mL/
minute), the retention times were 38.2 min (minor) and
40.5 minutes (major).
10. Muller, E.; Gasparutto, D.; Jaquinod, M.; Romieu, A.;
Cadet, J. Tetrahedron 2000, 56, 8689.
11. Radical cyclisation of aldehyde 15.
To a stirred solution of aldehyde 15 (145 mg, 0.46 mmol)
in dry, degassed, benzene (20 mL) under Ar, at reflux,
was added a solution of nBuSnH (0.25 mL, 0.93 mmol)
and AIBN (76 mg, 0.46 mmol) in dry, degassed benzene
(20 mL) over 6 h via a syringe pump. The reaction
mixture was heated at reflux for a further 16 h then
allowed to cool to room temperature and concentrated in
vacuo. Purification by flash silica column chromatogra-
phy (CH2Cl2 then 1 to 4% MeOH gradient) gave a
partially separable mixture of epimeric alcohols 16 in the
ratio 3:1, in addition to traces of non-cyclised alcohol
(ꢀ5%), in a combined yield of 114 mg (78%).
Data for major diastereoisomer; oil; [h]2D9 +88.5 (c 0.8 in
CHCl3); wmax (CHCl3)/cm−1 3617 (OH), 2937 (CH), 1681
(CꢀO), 1464, 1359, 1122, 1052, 1004, 894; lH (500 MHz)
1.42–1.46 (1H, m), 1.56–1.63 (1H, m), 1.71 (1H, ddd, J
18.1, 9.0, 4.4), 1.77 (1H, ddd, J 14.7, 11.3, 3.7), 2.00 (2H,
app. dt, J 13.5, 4.3), 2.35–2.37 (2H, m), 2.49 (1H, ddd, J
6.8, 6.8, 4.6), 2.57 (2H, dd, J 17.1, 4.1), 3.06 (1H, ddd, J
16.4, 11.2, 7.3), 3.18 (1H, ddd, J 13.3, 11.2, 5.5), 3.72-3.76
(1H, m), 3.83 (3H, s, OMe), 3.86 (3H, s, OMe), 4.18 (1H,
ddd, J 13.3, 7.3, 1.5), 6.52 (1H, s, ArH), 7.04 (1H, s,
ArH); lC (125 MHz) 18.2 (CH2), 26.3 (CH2), 31.8 (CH2),
35.0 (CH2), 35.6 (CH2), 35.8 (CH2), 44.9 (CH), 55.9
(CH3), 56.2 (CH3), 64.9 (C), 71.1 (CH), 109.8 (CH), 111.8
(CH), 125.8 (C), 133.6 (C), 147.5 (C), 147.9 (C), 176.2
(CꢀO); EIMS (m/z) 317 (M+, 56%), 274 (100%), 244
(79%); (HRMS M+ 317.1621, C18H23NO4 requires
317.1627).
9. Preparation of (+)-13
Isoquinoline derivative 11 (53 mg, 0.16 mmol) was heated
in a round bottomed flask by a Bunsen burner under a
high vacuum (0.25 mmHg) for approximately 30 s. The
1
progress of the reaction was monitored by H NMR and
the procedure repeated until evolution of cyclopentadiene
ceased to be observed. The dark coloured crude dihy-
dropyrrolo[2,1-a]isoquinolinone
was
dissolved
in
dichloromethane (10 mL), decolourising charcoal was
added and the resultant reaction mixture heated to reflux
with stirring. After 10 min the mixture was filtered hot,
concentrated in vacuo and the procedure repeated twice
more until the crude product had become a dark yellow
colour. The crude product was then purified via flash
column chromatography using silica gel (80% EtOAc/
petroleum ether) to yield dihydropyrrolo[2,1-a]-
Data for minor diastereoisomer; oil; [h]3D0 +66.0 (c 1.0 in
CHCl3); wmax (CHCl3)/cm−1 3507 (OH), 2938 (CH), 1672
(CꢀO), 1463, 1361, 1049, 908; lH (500 MHz) 1.47–1.70
(4H, m), 1.92–1.98 (1H, m), 2.01–2.05 (1H, m), 2.40 (1H,
br.s, disappears on D2O shake, OH), 2.45 (1H, dd, J 17.1,
9.5), 2.63 (1H, dd, J 17.1, 10.2), 2.74 (1H, app. dt, J 16.2,
4.6), 2.82 (1H, ddd, J 9.6, 9.6, 5.6), 2.92 (1H, ddd, J 16.2,

C. Gill et al. / Tetrahedron Letters 44 (2003) 7803–7807
7807
9.7, 6.6), 3.21 (1H, ddd, J 13.0, 9.7, 5.6), 3.86 (3H, s,
OMe), 3.87 (3H, s, OMe), 4.06 (1H, ddd, J 13.0, 6.7, 4.0),
4.52 (1H, m), 6.61 (1H, s, ArH), 6.78 (1H, s, ArH); lC
(125 MHz) 18.8 (CH2), 27.8 (CH2), 28.2 (CH2), 31.5
(CH2), 35.1 (2×CH2), 45.0 (CH), 55.9 (CH3), 56.4 (CH3),
64.2 (C), 68.7 (CH), 108.6 (CH), 112.0 (CH), 125.6 (C),
134.4 (C), 147.6 (C), 148.0 (C), 172.3 (CꢀO); EIMS (m/z)
317 (M+, 28%), 274 (100%), 245 (35%); (HRMS M+
317.1618, C18H23NO4 requires 317.1627).
then 1% MeOH in CH2Cl2) gave ketone 17 as a colour-
less oil (10 mg, 63%); [h]2D3 +52.2 (c 1.0 in CHCl3); wmax
(CHCl3)/cm−1 2938 (CH), 1720, 1682 (CꢀO), 1464, 1361,
1122, 907, 869; lH (500 MHz) 1.86–1.97 (3H, m), 2.18–
2.23 (1H, m), 2.49–2.55 (1H, m), 2.63 (1H, dd, J 17.5,
10.4), 2.69-2.77 (2H, m), 2.78 (1H, dd, J 17.5, 7.8), 2.97
(1H, ddd, J 16.7, 10.9, 7.0), 3.16 (1H, app. t, J 9.4), 3.26
(1H, ddd, J 13.1, 10.9, 5.1), 3.83 (3H, s, OMe), 3.86 (3H,
s, OMe), 4.21 (1H, ddd, J 13.1, 7.0, 2.6), 6.45 (1H, s,
ArH), 6.58 (1H, s, ArH); lC (125 MHz) 18.7 (CH2), 27.1
(CH2), 34.3 (CH2), 35.3 (CH2), 36.1 (CH2), 37.8 (CH2),
53.3 (CH), 56.0 (CH3), 56.1 (CH3), 66.2 (CH), 108.0
(CH), 111.7 (CH), 125.2 (C), 133.1 (C), 148.1 (C), 148.4
(C), 171.6 (CꢀO), 210.6 (CꢀO); EIMS (m/z) 315 (M+,
39%), 272 (39%), 245 (100%), 151 (19%); (HRMS M+
315.1470, C18H23NO4 requires 315.1471).
12. This mixture of epimeric alcohols has been reported
previously in racemic form, see: Mondon, A.; Hansen, K.
F.; Boehme, K.; Faro, H. P.; Nestler, H. J.; Vilhuber, H.
G.; Bottcher, K. Chem. Ber. 1970, 103, 615.
13. Oxidation of the minor diastereomer of 16 to give ketone
17
To a stirred suspension of Dess–Martin periodinane (31
mg, 0.07 mmol) in CH2Cl2 at room temperature under N2
was added a solution of alcohol 16 (19 mg, 0.06 mmol) in
CH2Cl2 (1 mL). The resulting mixture was stirred at
room temperature for 30 min then quenched with satu-
rated aqueous sodium thiosulfate solution (5 mL) and
extracted with CH2Cl2 (2×10 mL). The organic phases
were dried (MgSO4), filtered and concentrated in vacuo.
Purification by flash silica column chromatography (0.5
14. This ketone has been reported previously in racemic
form, see: (a) Mondon, A.; Bottcher, K. Chem. Ber. 1970,
103, 1512; (b) Kametani, T.; Higashiyama, K.; Honda,
T.; Otomasu, H. Heterocycles 1984, 22, 569.
15. See for example: Isobe, K.; Mohri, K.; Takeda, N.;
Suzuki, K.; Hosoi, S.; Tsuda, Y. Chem. and Pharm. Bull.
1994, 42, 197.