Diastereoselective synthesis of 10b-substituted hexahydropyrroloisoquinolines from L-tartaric acid. Creation of a quaternary carbon stereocentre via N-acyliminium ion cyclization

Article abstract of DOI:10.1016/j.tetlet.2004.06.036

A simple, three-step synthesis of 10b-substituted- hexahydropyrroloisoquinolines 12-17 starting from an L-tartaric acid derived imide is described. The methodology presented employs the addition of a Grignard reagent to the imide carbonyl group, followed by a one-pot acetylation- cyclization sequence. The crucial step, an acid-catalyzed carbon-carbon bond forming reaction via an N-acyliminium ion offers moderate to high stereoselectivity, which has been shown to be strongly dependent on the size of the R-substituent. The mixtures of pyrroloisoquinolines obtained can be separated as enantiomerically pure 2-silyloxy-derivatives.

Full text of DOI:10.1016/j.tetlet.2004.06.036

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 6011–6015
Diastereoselective synthesis of 10b-substituted
hexahydropyrroloisoquinolines from
quaternary carbon stereocentre via N-acyliminium ion cyclization
L
-tartaric acid. Creation of a
a
Danuta Mostowicz, Robert Wojcik, Grzegorz Dołega and Zbigniew Kałuza
b
c
a,
*
ꢀ
_
z
aInstitute of Organic Chemistry, Polish Academy of Sciences, 01-224 Warszawa, Poland
bFaculty of Chemistry, Warsaw University of Technology, 00-664 Warszawa, Poland
cInstitute of Chemistry, Podlasie University, 08-110 Siedlce, Poland
Received 19 April 2004; revised 1 June 2004; accepted 8 June 2004
Available online 24 June 2004
Abstract—A simple, three-step synthesis of 10b-substituted-hexahydropyrroloisoquinolines 12–17 starting from an L-tartaric acid
derived imide is described. The methodology presented employs the addition of a Grignard reagent to the imide carbonyl group,
followed by a one-pot acetylation–cyclization sequence. The crucial step, an acid-catalyzed carbon–carbon bond forming reaction
via an N-acyliminium ion offers moderate to high stereoselectivity, which has been shown to be strongly dependent on the size of the
R-substituent. The mixtures of pyrroloisoquinolines obtained can be separated as enantiomerically pure 2-silyloxy-derivatives.
ꢀ 2004 Elsevier Ltd. All rights reserved.
The pyrrolo[2,1-a]isoquinoline ring system is a key sub-
unit of Erythrina alkaloids,1 which have been widely
applied in folk medicine in tropical and sub-tropical
regions.2 Consequently, the stereoselective synthesis of
pyrroloisoquinolines has attracted significant attention
over recent years.3–8 The most common approach to the
asymmetric synthesis of 10b-substituted hexahydropyr-
roloisoquinolines involves Wittig olefination3–5 or addi-
tion of an organometallic reagent6;7 to the imide
carbonyl group of 1 (Scheme 1). The crucial step is
diastereoselective cyclization via N-acyliminium ion 4.
Steric control can be exerted by the groups already
present within the ring3–6 or as part of the nitrogen
substituent.7
variety of heterocycles bearing quaternary carbon ste-
reocentres next to the nitrogen atom,10–12 asymmetric
synthesis of 10b-substituted pyrrolo[2,1-a]isoquinolines
from a chiral dicarboxylic acid derived imide is rare,3–6
as far as we are aware.
Herein, we describe an efficient, stereocontrolled syn-
thesis of 10b-substituted hexahydropyrroloisoquinolines
starting from L-tartaric acid. Initially we planned to
prepare compounds 10a and 10b starting from the easily
available imide 64 using a procedure analogous to that
reported by Lete and co-workers9;10 (Scheme 2).
Silylation of 6 under standard conditions (TMS-Cl/
pyridine) gave bis-trimethylsilyl ether 7, which was
subsequently treated with phenylmagnesium bromide.
Lete and co-workers9;10 have reported recently on the
synthesis of racemic 10b-substituted pyrroloisoquino-
lines from succinimide employing a tandem carbophilic
addition of the organolithium reagent-N-acyliminium
ion cyclization sequence. Although this methodology
has been successfully applied for the preparation of a
After an aqueous workup, filtration through a silica gel
pad and recrystallization from hexane, the hydroxy-
lactam 8 was isolated in 75% yield. Both trimethylsilyl
groups in lactam 8 seem stable under the reaction con-
ditions as well as during the isolation procedure, prob-
ably due to the steric hindrance. Treatment of 8 with
trifluoroacetic acid (TFA) in refluxing dichloromethane
for approx. 1 h led to dihydroxypyrroloisoquinolines
10a and 10b isolated in trace amounts only. Further
increase in the reaction time resulted in the formation of
Keywords: N-Acyliminium ion cyclization; Diastereoselective synthe-
sis; Erythrinan alkaloids.
* Corresponding author. Tel.: +48-22-6323221; fax: +48-22-6326681;
e-mail: zkicho@icho.edu.pl
0040-4039/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.06.036

6012
D. Mostowicz et al. / Tetrahedron Letters 45 (2004) 6011–6015
yield of 10a and 10b. Available literature reports6;9;11;13
O
suggest that the synthesis of nitrogen heterocycles via
N-acyliminium formation of a quaternary carbon atom
usually requires prolonged heating of hydroxy-lactams
with protic acids. For example, the racemic 1,2-unsub-
stituted analogue of 10 was obtained from the respective
hydroxy-lactam after 36 h heating with TFA.9 Com-
pound 8, when treated with TFA gave desilylated 9a and
its open chain tautomer 9b, both of which apparently
are not stable under the reaction conditions.
N R1
R
Wittig
R2M
olefination
O
1
R2
R2
OH
N R1
N R1
R
R
O
A+
O
A+
2
3
To facilitate the synthesis of the 10b-substituted pyr-
roloisoquinolines we prepared the triacetate 11a. We
assumed that the presence of an acetoxy group at C-1
would speed up the slowest step of the reaction, namely
the N-acyliminium ion formation.14 Furthermore, we
also expected to see improved stereoselection in favor
of trans-addition of the nucleophile, due to the known
ability of acetoxy groups to bridge adjacent cationic
centres.15 Treatment of 8 with acetic anhydride (4 equiv)
and DMAP (1.1 equiv) in acetonitrile resulted in clean
conversion to a mixture of acetates 11a and 11b in a
R2
N R1
R1 =
R3
R
OMe
OMe
O
4
OMe
OMe
R2
1
ratio of ꢀ11:1, as established by H NMR analysis of
R
the crude reaction mixture. However, pure 11a could be
isolated by chromatography, on silica gel but in only
15% yield. The 1H NMR experiment indicated that 11a,
on standing in chloroform solution, slowly cyclized to
give 12. Consequently, the crude reaction mixture of
acetates 11a and 11b was treated with BF3ÆEt2O
(4 equiv) to give hexahydropyrroloisoquinolines 12a and
12b in 94% overall yield and in a 3:1 ratio, respectively.16
We found that the mixture of 12a and 12b was very
difficult to separate by chromatography. However,
treatment of this mixture with NaOMe in anhydrous
N
O
R3
5
Scheme 1.
tar-like products. Several other protic and/or Lewis
acids were tested as catalysts for the cyclization, but no
significant improvement was achieved in respect to the
HO
HO
TMSO
TMSO
TMSO
TMSO
OH
Ph
O
O
R
i
ii
(L)-Tartaric
acid
N
N
75%
97%
N
ref. 4
R
R
O
O
O
O
iii
R =
6
7
iv
8
O
AcO
AcO
OAc
Ph
HO
HO
OH
Ph
O
O
O
O
N
N
R
R
HO
HO
Ph
HO
HO
Ph
N
O
O
9a
11a
N
+
+
Ph
O
O
8% <
O
O
Ph
HO
HO
10a
vii
10b
AcO
AcO
R
R
vi
vii
vi
N
H
N
H
O
O
O
O
9b
11b
O
O
AcO
AcO
Ph
N
AcO
AcO
Ph
v
94%
N
+
O
O
12b
12a
Scheme 2. Reagents and conditions: (i) TMS-Cl, pyridine, 0 ꢁC gradually to rt, 1 h; (ii) PhMgBr (2 equiv), THF, 0 ꢁC gradually to rt, 2 h; (iii)
CF3CO2H (4 equiv), CH2Cl2 reflux, 1 h; (iv) Ac2O (4 equiv), DMAP (1.1 equiv), MeCN, rt, 4 h; (v) BF3ÆEt2O (4 equiv), rt, 10 min then semi-satd
NaHCO3/H2O; (vi) MeONa, MeOH; (vii) Ac2O, pyridine.

D. Mostowicz et al. / Tetrahedron Letters 45 (2004) 6011–6015
6013
of compounds 12–17 we applied the modified procedure,
consisting of the use of crude hydroxy-lactams for the
one-pot acetylation–cyclization reaction sequence.16
O
O
H
H
7.6
4.7
H
11.7
6.3
H
9.5
O
O
0
H
H
14
19.4
6.4
H
H
10
10
AcO
4.5
AcO
H
H
1
1
3
As is evident from Table 1, the overall yields of 12–17,
based on the starting imide 7, were high, except for entry
5. Interestingly, even the pyrroloisoquinoline 17, bearing
the phenylethynyl substituent at C10b was obtained in a
high yield. Previous attempts to synthesise C10b-phen-
ylethynyl pyrroloisoquinolines were unsuccessful.10
2
2
6
6
3
AcO
AcO
N
5
N
5
4
4
O
O
12a
12b
Figure 1.
methanol followed by separation on silica gel of the
resulting respective dihydroxy derivatives 10a,b and
subsequent re-acetylation gave pure 12a and 12b. NO-
ESY experiments, carried out with epimeric 12a and
12b, indicated an interaction between the ortho-phenylic
protons with the protons at C-2 of 12a and C-1 of 12b
(Fig. 1). Interestingly, the C-1 and C-10 protons of 12a
showed a very strong positive NOE effect. The respec-
tive spin interaction in the case of compound 12b was
substantially weaker, indicating that pyrroloisoquino-
lines 12a and 12b possess 10bS and 10bR-configurations,
respectively.
Although the diastereomers 12–17 or their respective
dihydroxy derivatives 10, 18–22 can be separated by
careful chromatography on silica gel, we have found
that the 2-t-butyldimethylsilyloxy-pyrroloisoquinolines
23ab–28ab are much easier to purify due to improved
differentiation in the polarity of the 10b epimers (Scheme
4). The reaction sequence––alkaline hydrolysis of the
acetates 12–17 and silylation (TBS-Cl, imidazole, DMF)
of the crude diols followed by their separation, on silica
gel––resulted in the isolation of enantiomerically pure
pyrroloisoquinolines 23ab–28ab in approx. 90% yields.
The NOESY experiments, carried out with 24ab–27ab
showed an interaction between the protons of the R
substituent at the carbon atom bonded to C-10b with
the protons at C-2 and the OH group of the C-10bS
epimers. For the C-10bR epimers the respective protons
of the R substituent only interact with the proton at C-1.
The configuration at C-10b in pyrroloisoquinolines 28ab
possessing a phenylethynyl substituent could not be di-
rectly established on the basis of NOESY experiments,
due to the lack of diagnostic protons able to interact
with protons at C-1 or C-2. However, the epimeric
mixture of pyrroloisoquinolines 22 when subjected to
palladium catalyzed cyclization18 [0.2 equiv Pd(OAc)2,
TEA, THF, rt, 2 h], gave dihydrofuranyl derivative 29a
in 82% yield. The formation of 29a indicates that the
major diastereomer of 22 possesses the S-configuration
at C-10b. Compound 29b was not detected in the crude
reaction mixture, however, a small amount (ꢀ4%) of
unreacted pyrroloisoquinoline 10b(R) 22 was isolated.
Prolonging the reaction time resulted in complete
decomposition of 10b(R) 22. While both processes––
5- and 6-endo-dig cyclization, according to Baldwin’s
rules are favorable––the latter was not observed, prob-
ably due to the rigidity of the pyrroloisoquinoline
skeleton.
Single crystal X-ray analysis of 12a confirmed our
deduction and provided an unequivocal proof of the
absolute configuration of both epimers (Fig. 2).17
In order to study the scope of this new methodology,
several other 10b-substituted pyrroloisoquinolines 12–17
were obtained (Scheme 3; Table 1). For the preparation
Figure 2.
O
TMSO
TMSO
O
Unequivocal establishment of the configuration at
C-10b of all pyrroloisoquinolines 12–17 allowed for the
rationalization of the observed stereoselectivity of
cyclization (Table 1). The C-10b(S) diastereomer is
created via trans-addition of the nucleophile (dimeth-
oxyphenyl moiety) with respect to the acetoxy group
involved in bridging the adjacent cationic centre (Fig. 3).
O
AcO
AcO
R
N
i, ii, iii
N
O
O
7
12 : R = Ph
O
13 : R = i-Pr
14 : R = c-hexyl
15 : R = vinyl
16 : R = allyl
O
Consequently, the C-10b(R) epimer is a product of syn-
addition. The stereochemical outcome of cyclizations
described in this paper can be rationalized by assuming
the importance of the differences in steric interactions of
the approaching nucleophile with the R substituent
versus the acetoxy group. The prevalent trans-addition
of the nucleophile was observed for the relatively small
17 : R =
Ph
Scheme 3. Reagents and conditions: (i) RMgX (1.5–3 equiv), 0 ꢁC
gradually to rt, (R ¼ Ph, i-Pr, c-hexyl, PhCBC), or )78 ꢁC gradually
to 0 ꢁC, (R ¼ allyl, vinyl), THF or Et2O (see Table 1); (ii) Ac2O
(4 equiv), DMAP (1.1 equiv), MeCN, rt, 2–4 h; (iii) BF3ÆEt2O (4 equiv),
rt, 10–30 min then satd NaHCO3/H2O.

6014
D. Mostowicz et al. / Tetrahedron Letters 45 (2004) 6011–6015
Table 1
Entry
RMgX/solvent
Producta
Yield [%]b
S:Rc
1
2
3
4
5
6
PhMgBr/THF
i-PrMg2/Et2O
12
13
14
15
16
17
82
80
50
62
38
81
3:1
1:6.3
1:10
1:1.2
4.6:1
9.5:1
c-HexylMgBr/Et2O
AllylMgCl/THF
VinylMgBr/THF
PhCBCMgBr/THF
a All reactions were carried out at 1 mmol scale, applying the standard procedure, see Scheme 3.
b Isolated yields of the mixture of diastereomers calculated for three steps.
c Diastereomeric ratio was determined by 1H NMR analysis.
R groups (phenyl, vinyl, phenylethynyl, entries 1, 5 and
6, Table 1). In contrast, when bulky R groups (i-propyl,
c-hexyl, allyl; entries 2, 3 and 4, Table 1) are present,
syn-addition occurs predominantly.
OMe
OMe
Ph
Ph
OMe
OMe
O
OH
O
(S)
(R)
N
HO
N
In summary, we have successfully developed a simple,
three-step synthesis of the C-10b-substituted-1,2-diacet-
oxy-hexahydropyrroloisoquinolines 12–17, starting
O
O
29a
82%
29b
Ph
R =
from the imide 6, derived from
L-tartaric acid. The
presented methodology employs the addition of a
Grignard reagent to the imide carbonyl group of 9,
followed by an one-pot acetylation–cyclization
sequence. The crucial step, an acid-catalyzed carbon–
carbon bond forming reaction via an N-acyliminium ion
offers moderate to high stereoselectivity, which, it has
been shown, is strongly dependent on the size of the R
substituent. The mixtures of pyrroloisoquinolines ob-
tained can be separated as enantiomerically pure 2-si-
lyloxy-derivatives. Additionally, the C-10b-substituted
pyrroloisoquinolines, obtained using this methodology,
can be regarded as valuable synthons for the prepara-
tion of Erythrina alkaloids.
iii
OMe
10: R = Ph
18: R = i-Pr
OMe
HO
R
N
19: R = c-hexyl
20: R = allyl
21: R = vinyl
i
12-17
HO
22: R =
Ph
O
ii
OMe
OMe
OMe
OMe
HO
O
HO
O
R
(S)
R
(R)
+
TBSO
TBSO
N
N
References and notes
23a: R = Ph
24a: R = i-Pr
23b: R = Ph
1. Dyke, S. F.; Quessy, S. N. In The Alkaloids; Rodrigo,
R. G. A., Ed.; Academic: New York, 1981; Vol. 18, p 1.
2. Tanaka, H.; Tanaka, T.; Etoh, H.; Goto, S.; Terada, Y.
Heterocycles 1999, 51, 2759–2764.
3. Lee, Y. S.; Lee, J. Y.; Park, H. Synth. Commun. 1997, 27,
2799–2812.
4. Lee, J. Y.; Lee, Y. S.; Chung, B. Y.; Park, H. Tetrahedron
1997, 53, 2449–2458.
5. Lee, Y. S.; Lee, J. Y.; Kim, D. W.; Park, H. Tetrahedron
1999, 55, 4631–4636.
24b: R = i-Pr
25b: R = c-hexyl
26b: R = allyl
27b: R = vinyl
25a: R = c-hexyl
26a: R = allyl
27a: R = vinyl
28a: R =
28b: R =
Ph
Ph
Scheme 4. Reagents and conditions: (i) MeONa, MeOH; (ii) TBS-Cl,
imidazole, DMF, rt, 24 h; (iii) Pd(OAc)2 (0.2 equiv), TEA, THF, rt,
2 h.
6. Gill, Ch; Greenhalgh, D. A.; Simpkins, N. S. Tetrahedron
Lett. 2003, 44, 7803–7807.
7. Garcia, E.; Arrasate, S.; Ardeo, A.; Lete, E.; Sotomayor,
N. Tetrahedron Lett. 2001, 42, 1511–1513.
8. Padwa, A.; Henning, R.; Kappe, C. O.; Reger, T. S. J.
Org. Chem. 1998, 63, 1144–1155, and references cited
therein.
9. Manteca, I.; Sotomayor, N.; Villa, M.-J.; Lete, E. Tetra-
hedron Lett. 1996, 37, 7841–7844.
Me
O
O
O
N
10. Collado, M. I.; Manteca, I.; Sotomayor, N.; Villa, M.-J.;
Lete, E. J. Org. Chem. 1997, 62, 2080–2092.
11. (a) Hitchings, G. J.; Helliwell, M.; Vernon, J. M. J. Chem.
Soc., Perkin Trans. 1 1990, 83–87, and references cited
therein; (b) Bailey, P. D.; Morgan, K. M.; Smith, D. I.;
Vernon, J. M. Tetrahedron Lett. 1994, 35, 7115–7118; (c)
AcO
R
R1
Nu
Figure 3.

D. Mostowicz et al. / Tetrahedron Letters 45 (2004) 6011–6015
6015
Bailey, P. D.; Morgan, K. M.; Smith, D. I.; Vernon, J. M.
Tetrahedron 2003, 59, 3369–3378.
12. Othoman, M.; Pigeon, P.; Decroix, B. Tetrahedron 1997,
53, 2495–2504.
13. Shoemaker, H. E.; Speckamp, W. N. Tetrahedron 1980,
36, 951–958.
14. Hiemstra, H.; Speckamp, W. N. In Comprehensive Organic
Synthesis; Trost, B. M., Fleming, I., Eds.; Pergamon:
Oxford, 1991; Vol. 2, pp 1047–1082.
Selected data for 12b; mp 211–212 ꢁC (EtOAc/hexane);
½aꢁ )27.0 (c 1, CH2Cl2); IR (CH2Cl2): 3066, 1755,
D
1708 cmꢂ1
;
1H NMR (500 MHz, CDCl3) (d, ppm): 1.90
and 2.02 (2s, 6H), 2.64 (m, 1H), 2.98 (m, 1H), 3.18 (m,
1H), 3.83 and 3.88 (2s, 6H), 4.21(m, 1H), 5.22 (d, 1H,
J ¼ 2:7 Hz), 5.99 (d, 1H, J ¼ 2:7 Hz), 6.64 and 6.81 (2s,
2H), 7.27–7.36 (m, 5H); 13C NMR (125.76 MHz, CDCl3)
(d, ppm): 20.51, 20.92, 27.50, 36.55, 55.80, 55.95, 69.12,
75.35, 76.91, 110.72, 111.58, 125.97, 127.02, 127.19,
128.11, 128.55, 141.83, 146.90, 148.36, 167.20, 169.29,
169.66; HRMS (LSIMS+) m=z: (M + Hþ) calcd for
C24H26O7N: 440.1709. Found: 440.1713.
15. Bernardi, A.; Micheli, F.; Potenza, D.; Scolastico, C.;
Villa, R. Tetrahedron Lett. 1990, 31, 4949–4952.
16. All compounds gave satisfactory spectroscopic and ana-
lytical data. Typical procedure for 12a and 12b: To a
solution of imide 7 (439 mg, 1 mmol) in dry THF (5 mL),
freshly prepared PhMgBr (2 mmol) in THF (5 mL) was
added at 0 ꢁC, and gradually warmed up to rt. The
resulting mixture was stirred at rt until TLC indicated the
disappearance of 7 (approx. 0.5 h), then was poured into
ice-cold semi-saturated aqueous NaHCO3 (30 mL) and
extracted with t-butyl methyl ether (3 · 20 mL). The
combined extracts were washed with ice-cold water, dried
(MgSO4), filtered and evaporated in vacuo. The crude
hydroxy-lactam 8 obtained was dissolved in dry MeCN
(5 mL), and after cooling to 0 ꢁC, dimethylaminopyridine
(135 mg, 1.1 mmol) and Ac2O (378 lL, 4 mmol) were
added. The cooling bath was removed and stirring was
continued at rt for 2 h, then BF3ÆEt2O (507 lL, 4 mmol)
was added in one portion. The mixture was stirred at rt for
10 min then cooled to 0 ꢁC, quenched with saturated
aqueous NaHCO3 (5 mL) and extracted with CH2Cl2
(3 · 20 mL). The combined extracts were washed with
water (2 · 20 mL), dried (MgSO4), filtered and evaporated
in vacuo. The product was purified by flash column
chromatography on silica gel (EtOAc/hexane ¼ 1:1) to
yield pyrroloisoquinolines 12a and 12b as a 3:1 mixture in
82% yield. The mixture of 12a and 12b (220 mg, 0.5 mmol)
obtained was dissolved at rt in dry MeOH (10 mL)
containing MeONa (16 mg, 0.3 mmol). The solution was
stirred until TLC indicated the disappearance of the
substrate (ꢀ0.5 h), then the reaction was quenched by the
addition of a small piece of dry ice and evaporated in
vacuo. The product was purified by flash column chro-
matography on silica gel (EtOAc/hexane/MeOH, 7.5:4:1)
to give 10a (107 mg, 60% yield) and 10b (32 mg, 18%
yield). Dihydroxy-pyrroloisoquinolines 10a and 10b were
acetylated (Py/Ac2O) and crystallized to yield analytically
pure acetates 12a and 12b.
Typical procedure for the preparation of pyrroloisoquin-
olines 23a and 23b: To a stirred solution of the crude
reaction mixture of 10a and 10b obtained as above from
12a, 12b (220 mg, 0.5 mmol), imidazole (102 mg, 1.5 mmol)
in DMF (3 mL) was added followed by tert-butyldimeth-
ylchlorosilane (150 mg, 1 mmol) and the mixture was
stirred for 24 h at room temperature. The mixture was
poured into water, extracted with ethyl acetate
(2 · 10 mL), washed with water and brine then dried over
MgSO4, and evaporated under reduced pressure. The
residue was purified by flash column chromatography
(EtOAc/hexane, 35:65) to give 23a, (167 mg, 71% yield)
and 23b (54 mg, 23% yield).
Selected data for 23a: oil; ½aꢁ )120.2 (c 1.1, CH2Cl2); IR
D
(CH2Cl2): 3683, 1705, 1608 cmꢂ1
.
1H NMR (500 MHz,
CDCl3) (d, ppm): 0.14 and 0.20 (2s, 6H), 0.93 (s, 9H), 1.64
(d, 1H, J ¼ 10:0 Hz), 2.45 (m, 1H), 2.72 (m, 1H), 3.57 (m,
1H), 3.71 (m, 1H), 3.86 and 3.98 (2s, 6H), 4.23 (d, 1H,
J ¼ 8:3 Hz), 4.57 (dd, 1H, J ¼ 8:3, 10.0 Hz), 6.62 and 7.34
(2s, 2H), 7.16–7.20 (m, 2H), 7.29–7.38 (m, 3H); 13C NMR
(125.76 MHz, CDCl3) (d, ppm): 18.37, 25.76, 26.26, 37.75,
55.96, 56.41, 66.39, 76.60, 82.10, 108.94, 111.35, 126.6,
127.54, 128.21, 128.86, 132.85, 137.77, 147.87, 148.53,
170.24. HRMS (EI) m=z: (Mþ) calcd for C26H35O5NSi:
469.2287. Found: 469.2285.
Selected data for 23b: mp 216–218 ꢁC; ½aꢁ )4.4 (c 1.2,
D
1
CH2Cl2); IR (CH2Cl2): 3559, 1705, 1611 cmꢂ1. H NMR
(500 MHz, CDCl3) (d, ppm): 0.15 and 019 (2s, 6H), 0.88 (s,
9H), 2.00 (d, 1H, J ¼ 3:9 Hz), 2.27 (m, 1H), 2.89 (m, 1H),
3.10 (m, 1H), 3.87 and 3.88 (2s, 6H), 4.11 (m, 1H), 4.20 (d,
1H, J ¼ 3:1), 4.55 (dd, 1H, J ¼ 3:1, 3.9 Hz), 6.67 and 7.08
(2s, 2H), 7.22–7.29 (m, 3H), 7.35–7.39 (m, 2H); 13C NMR
(125.76 MHz, CDCl3) (d, ppm): 18.11, 25.64, 27.46, 35.78,
55.87, 56.27, 69.54, 77.70, 80.89, 110.61, 112.34, 126.67,
127.57, 127.64, 128.23, 128.75, 142.75, 147.50, 148.61,
171.03. HRMS (EI) m=z: (Mþ) calcd for C26H35O5NSi:
469.2287. Found: 469.2279.
Selected data for 12a; mp 230–231 ꢁC (ethanol); ½aꢁ )57.2
D
1
(c 2.4, CH2Cl2); IR (CH2Cl2): 3055, 1753, 1713 cmꢂ1; H
NMR (500 MHz, CDCl3) (d, ppm): 1.86 and 2.13 (2s, 6H),
2.47 (m, 1H), 2.82 (m, 1H), 3.43 (m, 1H), 3.87 (m, 1H),
3.88 and 3.91 (2s, 6H), 5.72 (d, 1H, J ¼ 6:8 Hz), 5.99 (d,
1H, J ¼ 6:8 Hz), 6.65 and 7.14 (2s, 2H), 7.07–7.10 (m, 2H),
7.27–7.34 (m, 3H); 13C NMR (125.76 MHz, CDCl3) (d,
ppm): 20.60, 20.67, 26.12, 37,59, 55.94, 56.20, 67.86, 74.59,
78.27, 109.19, 111.68, 126.63, 127.42, 128.02, 128.31,
131.06, 138.45, 147.75, 147.68, 166.18, 169.82, 170.17;
HRMS (EI) m=z: (Mþ) calcd for C24H25O7N: 439.1631.
Found: 439.1642.
17. The crystallographic data for compound 12a has been
deposited with the Cambridge Crystallographic Data
Center as supplementary publication number CCDC
222580. Copies of the data can be obtained, free of
charge, on application to CCDC, 12 Union Road,
Cambridge, CB2 1EZ, UK [fax: +44(0)-1223-336033 or
e-mail: deposit@ccdc.cam.ac.uk].
18. Saito, T.; Suzuki, T.; Morimoto, M.; Akiyama, C.;
Ochiai, T. J. Am. Chem. Soc. 1998, 120, 11633–
11644.