An enantiospecific synthesis of (+)-demethoxyerythratidinone from (S)-malic acid: key observations concerning the diastereocontrol in malic acid-derived N-acyliminium ion cyclisations

Article abstract of DOI:10.1016/j.tetlet.2007.06.111

The stereochemical outcome of N-acyliminium ion mediated cyclisations of malic acid derived lactams depend upon the nature of the protecting group on the lactam secondary alcohol, and also on the nature of the substituent at the reacting electrophilic cen

Full text of DOI:10.1016/j.tetlet.2007.06.111

Tetrahedron Letters 48 (2007) 5942–5947
An enantiospecific synthesis of (+)-demethoxyerythratidinone from
(S)-malic acid: key observations concerning the diastereocontrol
in malic acid-derived N-acyliminium ion cyclisations
Fengzhi Zhang, Nigel S. Simpkins* and Claire Wilson
School of Chemistry, The University of Nottingham, University Park, Nottingham NG7 2RD, UK
Received 18 May 2007; accepted 21 June 2007
Available online 24 June 2007
Abstract—The stereochemical outcome of N-acyliminium ion mediated cyclisations of malic acid derived lactams depend upon the
nature of the protecting group on the lactam secondary alcohol, and also on the nature of the substituent at the reacting electrophilic
centre. The origins of an unusual syn-selective cyclisation of a TIPS protected lactam are discussed, and the cyclisation is employed
as the key step in an asymmetric synthesis of 3-demethoxyerythratidinone (4).
Ó 2007 Elsevier Ltd. All rights reserved.
N-Acyliminium ions have proved to be especially useful
reactive intermediates for alkaloid synthesis, enabling
controlled C–C bond formation in both intra- and
inter-molecular modes with a range of nucleophilic spe-
cies.1 In connection with our programme of research
exploring new asymmetric routes to erythrinan alkaloids
we became interested in comparing our recently estab-
lished chiral base mediated (enantioselective) approach
to a related enantiospecific method using (S)-malic acid
as starting material.2 Previously, Lee and co-workers
had described a synthesis of pyrrolidinoisoquinoline
derivatives that used the cyclisation of hydroxylactam
1 to give the tricyclic product 3 as a single diastereoiso-
mer, Scheme 1.3
+
BF -OEt
3
2
N
N
N
O
O
O
HO
5
H
CH Cl , RT
2
2
AcO
AcO
AcO
1
3 74%
2
MeO
MeO
N
O
4
Scheme 1. Stereocontrolled N-acyliminium arylation.
Keywords: 3-Demethoxyerythratidinone; N-Acyliminium; Total synthesis; Cyclisation; Alkaloid.
*
Corresponding author. E-mail: nigel.simpkins@nottingham.ac.uk
0040-4039/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2007.06.111

F. Zhang et al. / Tetrahedron Letters 48 (2007) 5942–5947
5943
In this example, as in many related ones, the stereo-
chemical outcome results from nucleophilic attack anti
to the bystander acetoxy group, which may be influenc-
ing the outcome by neighbouring group participation.4
tricyclic lactam 7 in good yield but with a very modest
diastereomer ratio of 3:1. This outcome was disappoint-
ing and somewhat surprising based on the precedent
illustrated in Scheme 1. Lactam 7 could be converted
into the unsaturated counterpart 8 by treatment with
base and the enantiomeric excess of 47% of this product
reflected the low diastereomer ratio of the precursor.9
In this Letter, we describe how we have employed this
type of process in the synthesis of the naturally occur-
ring alkaloid (+)-3-demethoxyerythratidinone 4,5,6 and
also demonstrate that: (i) more highly substituted
hydroxylactams related to 1 may not give the high level
of control seen in Scheme 1; (ii) switching from acetate
protection to TIPS protection for the ring hydroxyl
can enable the opposite sense of diastereocontrol to that
shown in Scheme 1—but this depends upon the substitu-
ent present at C-5.
The selectivity issue in the cyclisation leading to 7 could
not be resolved by the use of lower reaction tempera-
tures since the reaction proved very sluggish below
ꢀ40 °C, and the use of Lewis acids such as BF3–OEt2
gave similarly modest selectivity to TIPSOTf. We rea-
soned that exchanging the ring acetate substituent for
a rather bulky TIPS protecting group might enhance
the facial selectivity in the N-acyliminium reaction.
However, a Grignard addition reaction of an O-silyl
imide counterpart of 5 had previously been shown to
occur with low regioselectivity, and so starting with a
silicon protected imide was unattractive.10 Thus, the
combined addition and in situ hydroxyl deprotection
represented by the conversion of 5 into 6 now proved
helpful in combining the high regiocontrol of the acetate
protected series with the opportunity to swap for a silyl
protected derivative in the pursuit of high diastereocon-
trol in the subsequent cyclisation. With this plan in place
the modified synthetic sequence is shown in Scheme 3.
Our synthesis began with the addition of but-3-enylmag-
nesium bromide to the readily available enantiopure
imide 5,7 which gave hydroxylactam 6 in a completely
regiocontrolled fashion, Scheme 2.
This outcome was expected based on a closely related
Grignard addition described previously in which the
ring acetate was retained in the product.8 In our case,
the use of a large excess of Grignard reagent (5 equiv)
gave optimal yields but also deacylated the secondary
alcohol. This proved of little consequence at this stage
since acetylation of 6 re-installed the secondary acetate
without affecting the tertiary hydroxyl function. This
hydroxylactam, on treatment with TIPSOTf, gave the
Protection of 6 as the corresponding secondary TIPS
ether proved straightforward and exposure of this
MeO
MeO
MeO
(i) Ac O, Et N
2
3
DMAP, CH Cl
MgBr
2
2
HO
MeO
N
N
O
O
O
(ii) TIPSOTf
THF
o
o
CH Cl , -40
C
-78 C to RT
2
2
AcO
HO
6 61%
5
MeO
MeO
MeO
MeO
NaH, THF
N
O
N
O
AcO
7 82% (3:1)
Scheme 2. N-acyliminium arylation with modest stereocontrol.
(-)-8 95%, 47% ee
(i) TIPSOTf, CH Cl
2
MeO
MeO
MeO
MeO
2
(i) TBAF, THF
(ii) Ac O, CH Cl
2
2,6-lutidine, RT
N
N
6
O
O
2
2
(ii) BF -OEt CH Cl
2
3
2,
2
Et N, DMAP
3
o
-78 C to RT
TIPSO
9 80% >9:1
(iii) NaH, THF
(+)-8 65%, >99% ee
(from pure 9)
Scheme 3. N-acyliminium arylation with high and reversed stereocontrol.

5944
F. Zhang et al. / Tetrahedron Letters 48 (2007) 5942–5947
product to BF3–OEt2, as described by Lee (Scheme 1),3
but initially at low temperature, then gave a cyclised lac-
tam in good yield and in at least 9:1 selectivity. To our
initial surprise, the product of this reaction was 9, hav-
ing a pseudo-diastereomeric relationship to 7 from the
acetate series, in which aromatic attack on the putative
N-acyliminium intermediate occurred from the same side
of the lactam ring as the OTIPS substituent.
anti alignment of the existing C–O r bond and the
emerging r* orbital at the reacting centre. In contrast,
the reactions of 10 involving enol silanes and other
silicon-based nucleophiles seem to give rather modest
selectivities in most cases.13
This, slightly mysterious, ‘syn-effect’ appears not to have
been observed before in intramolecular closures of the
type we have described, or indeed with any systems
involving aromatics as nucleophiles.14 If stereoelectron-
ics were responsible for our cyclisation outcome we
expected to be able to generalise the result and achieve
similar selectivities using lactams with other
substituents.
The stereochemical assignment of the lactam 9 was
secured following an X-ray structure determination,
Figure 1.11
We chose to examine the parent system analogous to
that shown in Scheme 1, in which the C-5 substituent
is simply a hydrogen, Scheme 4.
For direct comparison with the result in Scheme 1 we
first converted 5 into 13 by reduction with NaBH4 in
CH2Cl2–MeOH, and then carried out the cyclisation
using BF3–OEt2 in CH2Cl2 between 0 °C and room
temperature. In accord with previous results, including
that in Scheme 1, only a single diastereomeric product,
lactam 15, resulting from the expected anti-attack was
obtained (in 56% yield for two steps).
We then prepared the OTIPS derivative 14 in order to
test for the ‘syn-effect’, which required routine conver-
sions of 5 into 11 by acidic hydrolysis (86%), protection
to give 12 and then NaBH4 reduction to give 14. Cycli-
sation of this compound using BF3–OEt2 in CH2Cl2 was
carried out at the lowest viable temperature, ꢀ40 °C to
rt, so as to mimic the conditions in Scheme 3, and max-
imise selectivity. The cyclisation gave two diastereomeric
products 16 and 18 in which the former compound, aris-
ing from anti-addition predominated. This is the same
sense of selectivity as the acetate series, and demon-
strates that whatever is responsible for the syn-selectivity
leading to the butenyl derivative 9 does not seem to
Figure 1. X-ray structure of 9 illustrating attachment of the aromatic
group syn to the O-silyl substituent. Only one of four crystallograph-
ically independent molecules is shown, and the ellipsoids are drawn at
the 30% probability level. The iPr substituents on the TIPS group, and
most of the hydrogen atoms, have been omitted for clarity.
Compound 9 was accompanied by small amounts (up to
10% but usually less) of the corresponding minor diaste-
reomer, which was easily separated by chromatography,
and both of which were used in subsequent chemistry
(vide infra). By removing the silicon protecting group,
acetylating the so-formed alcohol and elimination of
acetic acid as before, diastereomerically pure 9 was con-
verted into enantiomerically pure (+)-8.
MeO
MeO
MeO
MeO
steps
HO
N
N
O
O
O
see text
H
RO
RO
5
R = Ac
13 R = Ac
11 R = H
14 R = TIPS
12 R = TIPS
Bn
N
AcO
O
TBSO
MeO
2
MeO
MeO
BF -OEt
3
10
+
N
N
O
O
MeO
CH Cl
2
2
H
H
see text for
conditions
High syn-selectivity has been reported previously for
intermolecular reactions of the acetoxylactam 10 with
allyl, propargyl and allenylsilanes, and stannanes, under
Lewis acid catalysis, with product ratios ranging from
3.8:1 to >100:1.12 This ‘remarkable cis-selectivity’ was
attributed to a stereoelectronic effect involving preferred
RO
RO
R = Ac 56% from 5
15 R = Ac
100 : 0
5 : 1
17 R = Ac
R = TIPS 47% from 11
16 R = TIPS
18 R = TIPS
Scheme 4. N-acyliminium cyclisation of the parent lactam.

F. Zhang et al. / Tetrahedron Letters 48 (2007) 5942–5947
5945
operate for the parent system where the C-5 substituent
is simply a proton.
with LiAlH4–AlCl3 served to remove the unwanted
amide carbonyl and effect removal of the silicon protect-
ing group. Diol 21 had been prepared previously by
Wasserman and Amici in their synthesis of racemic
3-demethoxyerythratidinone,6d and application of the
final two steps from their synthesis to 21 involved dou-
ble Swern oxidation to give 22 and then aldol cyclisation
to give (+)-3-demethoxyerythratidinone 4. The spectro-
scopic data for this compound were consistent with the
structure shown and previous published data.17
A more detailed analysis of this system must await fur-
ther results, but at present a most likely explanation for
the selective formation of 9 is based on a conformational
relay effect as illustrated in Figure 2.15
H
O
H
+
H
N
As anticipated, the use of 19 as starting material resulted
5
OSiR
3
in the formation of the final alkaloid as its natural anti-
26
pode—[aꢁD +316 (c, 0.4, CHCl3), comparison data5—
20
[aꢁD +325 (c, 0.25, CHCl3). When we applied the same
sequence of reactions shown in Scheme 5 to the diaste-
reomeric starting lactam 9 we obtained very similar
yields for each step (in parentheses) and were able to
access the natural product as the unnatural (ꢀ)-antipode.
Figure 2.
Examination of molecular models suggests that the
arrangement shown is a reasonable representation of
the reactive conformation of the intermediate N-acyl-
iminium ion. The presence of the very bulky TIPS group
(R = iPr) pushes the butenyl group to the opposite (top)
face of the lactam ring, which in turn engenders attack
of the aromatic group from the same side as the OTIPS
substituent. Reaction in the opposite sense requires a
more congested arrangement.
In summary, these new results help to clarify the stereo-
chemical issues related to N-acyliminium ion reactions
of malic acid derived lactams, and particularly the unu-
sual syn-effect seen with a TIPS protected compound.
The synthesis of (+)-3-demethoxyerythratidinone repre-
sents only the third asymmetric access to this com-
pound, and at eight steps from commercial material is
one of the shortest of any of the routes established to
date.
In the simpler system, where the C-5 substituent is a pro-
ton, the relay effect is lost and reaction is more facile anti
to the OTIPS group (although by no great measure since
the selectivity is modest).
Acknowledgement
With an efficient and selective access to 9 available we
were then in a position to conclude our proposed new
access to 3-demethoxyerythratidinone 4. Initially, in
order to match our synthetic material with the alkaloid
in the natural enantiomeric series we used the minor
product generated from the cyclisation shown in Scheme
3— that is, the diastereomer of 9 (19), Scheme 5.16
We gratefully acknowledge the University of Notting-
ham and the ORS scheme for support of this work
through a studentship to F.Z.
References and notes
Wacker oxidation of diastereomerically pure 19 gave
ketolactam 20 in good yield and subsequent reduction
1. (a) Maryanoff, B. E.; Zhang, H.-C.; Cohen, J. H.; Turchi,
I. J.; Maryanoff, C. A. Chem. Rev. 2004, 104, 1431; (b)
MeO
MeO
MeO
MeO
MeO
N
N
N
PdCl , CuCl
2
O
O
LiAlH , AlCl
4
MeO
3
THF-Et O
2
O , DMF-H O
2
2
O
HO
21 48% (60%)
TIPSO
19
TIPSO
HO
20 86% (84%)
MeO
MeO
MeO
MeO
(COCl) , DMSO
2
20% KOH (aq)
N
N
o
MeOH
reflux
Et N, CH Cl , -78
C
3
2
2
O
O
O
22 67% (69%)
(+)-4 50%
Scheme 5. Synthesis of (+)-3-demethoxyerythratidinone.

5946
F. Zhang et al. / Tetrahedron Letters 48 (2007) 5942–5947
Huang, P.-Q. Synlett 2006, 1133; (c) Speckamp, W. N.;
12 Hz, 1H, NCHAHB), 3.75 (s, 3H, OCH3), 3.77 (s, 3H,
OCH3), 4.31 (dd, J = 5, 9 Hz, 1H, NCHAHB), 4.41 (t,
J = 9.6 Hz, 1H, CHOTIPS), 6.48 (s, 1H, ArH), 6.80 (s,
1H, ArH). 13C NMR (100.6 MHz, CDCl3): d = 12.4
(SiCH), 17.6 (CH3), 17.7 (CH3), 17.8 (CH3), 17.9 (CH3),
28.2 (ArCH2), 29.8 (COCH3), 30.0 (CCH2), 34.9
(ArCH2CH2), 38.9 (CH2COMe), 40.6 (NCOCH2), 55.6
(OCH3), 55.7 (OCH3), 66.2 (C), 75.9 (CHOSi), 107.8
(ArCH), 111.4 (ArCH), 124.6 (ArC), 132.2 (ArC), 147.83
(ArC), 147.81 (ArC), 169.4 (C@O), 207.1 (C@O). HR-
EIMS: m/z Calcd for C27H43NO5Si(M++H): 490.2989;
found, 490.2990.
Moolenaar, M. J. Tetrahedron 2000, 56, 3817; (d) Speck-
amp, W. N.; Hiemstra, H. Tetrahedron 1985, 41, 4367.
2. Blake, A. J.; Gill, C.; Greenhalgh, D. A.; Simpkins, N. S.;
Zhang, F. Synthesis 2005, 19, 3287.
3. Lee, Y. S.; Kang, D. W.; Lee, S. J.; Park, H. J. Org. Chem.
1995, 60, 7149.
4. Louwrier, S.; Ostendorf, M.; Boom, A.; Hiemstra, H.;
Speckamp, W. N. Tetrahedron 1996, 52, 2603.
5. Isolation: Barton, D. H. R.; Gunatilaka, A. A. L.; Letcher,
R. M.; Lobo, A. M. F. T.; Widdowson, D. A. J. Chem.
Soc., Perkin Trans. 1 1973, 874.
6. Synthesis: (a) Gao, S.; Tu, Y. Q.; Hu, X.; Wang, S.; Hua,
R.; Jiang, Y.; Zhao, Y.; Fan, X.; Zhang, S. Org. Lett.
2006, 8, 2373; (b) Wang, Q.; Padwa, A. Org. Lett. 2006, 8,
601; (c) Danishefsky, S. J.; Panek, J. S. J. Am. Chem. Soc.
1987, 109, 917; (d) Wasserman, H. H.; Amici, R. M. J.
Org. Chem. 1989, 54, 5843, and references cited therein;
For an overview of pioneering work towards erythrinan
alkaloids using N-acyliminium chemistry see Ref.1d.
Asymmetric synthesis: (e) Allin, S. M.; Streetley, G. B.;
Slater, M.; James, S. L.; Martin, W. P. Tetrahedron Lett.
2004, 45, 5493; (f) Hosoi, S.; Ishida, K.; Sangai, M.;
Tsuda, Y. Chem. Pharm. Bull. 1992, 40, 3115.
7. Pilli, R. A.; Russowsky, D. J. Org. Chem. 1996, 61, 3187.
8. Tomihisa, O.; Sojiro, S.; Rumiko, S.; Nozoe, S. Tetrahe-
dron Lett. 1990, 31, 7329.
9. The ee was determined as 47% by HPLC (Chiracel OD
Column, 20% IPA in hexane, 0.4 mL/min); the retention
times were 23.4 min (minor) and 31.2 min (major).
10. (a) He, B.-Y.; Wu, T.-J.; Yu, X.-Y.; Huang, P.-Q.
Tetrahedron:Asymmetry 2003, 14, 2101; For a relevant
computational study, see: (b) Ye, J.-L.; Huang, P.-Q.; Lu,
X. J. Org. Chem. 2007, 72, 35.
11. The crystal structure has been deposited at the Cambridge
Crystallographic Data Centre; Deposition Number
CCDC 646549.
12. Keum, G.; Kim, G. Bull. Korean Chem. Soc. 1994, 15, 278.
13. (a) Othman, R. B.; Bousquet, T.; Othman, M.; Dalla, V.
Org. Lett. 2005, 7, 5335; (b) Othman, R. B.; Bousquet, T.;
Fousse, A.; Othman, M.; Dalla, V. Org. Lett. 2005, 7,
2825.
14. Some directing effects, leading to modest to high syn-
selectivity in reductions using Et3SiH have been proposed,
see: (a) Huang, J.-M.; Hong, S.-C.; Wu, K.-L.; Tsai,
Y.-M. Tetrahedron Lett. 2004, 45, 3047; (b) Huang, P. Q.;
Chen, Q. F.; Chen, C. L.; Zhang, H. K. Tetrahedron:
Asymmetry 1999, 10, 3827, see also Ref. 10a.
Diketone 22 via diol 21: A solution of AlCl3 (1.5 g,
11 mmol) in Et2O (10 mL) was added dropwise to a
solution of LiAlH4 (11 mL, 1 M in THF) at ꢀ15 °C, then
the resulting solution was stirred at rt for 1 h. Then the
above solution was added dropwise to a solution of lactam
20 (600 mg, 1.2 mmol) in THF (10 mL) at ꢀ15 °C. The
resulting reaction mixture was stirred overnight before
quenching by careful addition of 5% NH3 (aq). The
product was extracted into EtOAc (200 mL) and washed
with water (10 mL) and brine (10 mL). The organic extract
was dried over anhydrous MgSO4 and concentrated. The
crude product was purified by flash chromatography on a
silica column (eluent: PE/AcOEt = 4:1 then 2:1) to give
the diol 21 (189 mg, 48%) as a colourless oil. This epimeric
mixture of products was used directly in the next step.
A solution of (COCl)2 (95 mg, 0.75 mmol) in DCM (2 mL)
under a nitrogen atmosphere was cooled to ꢀ78 °C, and
then DMSO (0.11 mL, 1.5 mmol) was added dropwise,
and the resulting solution was stirred at this temperature
for 30 min. Diol 21 (60 mg, 0.19 mmol) in DCM (2 mL)
was added dropwise, then the reaction mixture was stirred
at ꢀ78 °C for 2 h. Et3N (0.38 mL, 2.8 mmol) was added
dropwise, the reaction mixture was warmed to room
temperature, and then water (5 mL) was added. The
product was extracted into DCM (30 mL), the organic
phase was washed with water (5 mL) and brine (5 mL),
dried over anhydrous MgSO4 and then concentrated. The
crude product was purified by flash chromatography on a
silica column (eluent: PE/AcOEt = 1:2) to give 22 as a
25
colourless oil (40 mg, 67%). ½aꢁD 50.2 (c 0.75 in CHCl3).
IR (CHCl3): 2936, 2850, 1746, 1709 cmꢀ1 1H NMR
.
(400 MHz, CDCl3): d = 2.11 (s, 3H, CH3CO), 2.11–2.26
(m, 3H, CH2CHAHB), 2.35–2.43 (m, 3H, CH2CO,
ArCHAHB), 2.48–2.54 (m, 1H, CH2CHAHB), 3.00–3.10
(m, 4H, ArCHAHBCHAHBNCH2), 3.14–3.22 (m, 1H,
ArCH2CHAHBN), 3.82 (s, 3H, OCH3), 3.84 (s, 3H,
OCH3), 6.51 (s, 1H, ArH), 6.90 (s, 1H, ArH). 13C NMR
(100.6 MHz, CDCl3): d = 21.1 (CH2), 30.2 (CH3), 32.6
(CH2), 36.3 (CH2), 39.2 (CH2), 41.3 (CH2), 43.5 (CH2),
55.8 (OCH3), 55.9 (OCH3), 68.1 (C), 109.6 (ArCH), 111.6
(ArCH), 125.0 (ArC), 126.0 (ArC), 147.6 (ArC), 148.2
(ArC), 208.0 (C@O), 215.9 (C@O). HR-EIMS: m/z Calcd
for C18H23NO4Na(M++Na): 340.1515; found, 340.1519.
(+)-3-demethoxyerythratidinone 4: A solution of diketone
22 (34 mg, 0.107 mmol) and 20% KOH (1.5 mL) in MeOH
(30 mL) was heated at 120 °C under a nitrogen atmo-
sphere for 10 h. The mixture was cooled and then
concentrated and extracted with DCM (20 mL), the
organic phase was then washed with water (5 mL) and
brine (5 mL). The organic phase was dried over anhydrous
MgSO4 and concentrated. The crude product was purified
by flash chromatography on a silica column (eluent:
DCM/MeOH = 10:1) to give (+)-4 as a yellow oil
15. Pilli has proposed a somewhat similar explanation for
reactions of related regioisomeric lactams, see: Schuch, C.
M.; Pilli, R. A. Tetrahedron: Asymmetry 2002, 13, 1973.
16. Experimental and compound data for Scheme 5: Ketolac-
tam 20: PdCl2 (61 mg, 0.34 mmol) and CuCl (171 mg,
1.7 mmol) were dissolved in a mixture of DMF (8 mL) and
water (1.5 mL), the reaction mixture was stirred at rt for
2 h. Then lactam 19 (818 mg, 1.7 mmol) in DMF (3 mL)
was added and the resulting solution was stirred under an
O2 atmosphere for 20 h. The product was then extracted
with ether (150 mL) and washed with water (10 mL) twice
and brine (10 mL). The organic extract was dried over
anhydrous MgSO4 and concentrated. The crude product
was purified by flash chromatography on a silica column
(eluent: PE/AcOEt = 1:1) to give the methyl ketone 20
26
(731 mg, 86%) as a sticky oil [aꢁD ꢀ78.6 (c 1.55 in CHCl3)
IR (CHCl3): 2936, 2868, 1714, 1682, 1457, 1361, 1113,
26
1069, 996, 883 cmꢀ1
.
1H NMR (400 MHz, CDCl3):
(15.9 mg, 50%). [aꢁD +316 (c 0.4 in CHCl3). IR (CHCl3):
1
d = 1.05–1.10 (m, 21H, CH(CH3)2), 2.00 (s, 3H, CH3CO),
2.18–2.24 (m, 3H, COCHAHBCH2), 2.54–2.78 (m, 5H,
ArCH2, CH2CON, COCHAHBCH2), 2.85 (td, J = 4,
2936, 2852, 2399, 1666, 1463, 1360, 1107 cmꢀ1. H NMR
(400 MHz, CDCl3): d = 2.24–2.33 (m, 2H, CCH2), 2.43–
2.68 (m, 4H, CH2CO, CHAHBC@), 2.68–2.92 (m, 2H,

F. Zhang et al. / Tetrahedron Letters 48 (2007) 5942–5947
5947
NCHAHBCHAHBC@), 3.03–3.12 (m, 2H, ArCHAHBCH2-
NCHAHB), 3.26 (dd, 1H, J = 7.6, 14.4 Hz, NCHAHB),
3.45–3.52 (m, 1H, NCHAHB), 3.76 (s, 3H, OCH3), 3.88 (s,
3H, OCH3), 6.12 (s, 1H, CH@), 6.56 (s, 1H, ArH), 6.66 (s,
1H, ArH). 13C NMR (100.6 MHz, CDCl3): d = 21.5
(ArCH2), 28.6 (CH2C@), 32.8 (CH2CO), 35.9 (CCH2),
40.1 (ArCH2CH2N), 45.7 (NCH2CH2C@), 55.9 (OCH3),
56.0 (OCH3), 63.7 (C), 110.2 (ArCH), 112.8 (ArCH),
123.9 (CH@C), 124.4 (ArC), 125.4 (ArC), 146.9 (ArC),
148.4 (ArC), 168.5 (C@), 199.3 (C@O). HR-EIMS:
m/z Calcd for C18H22NO3(M++H): 300.1594; found,
300.1586.
17. Despite the natural product having been prepared a number
of times there are scant data readily available in the
literature because so many reports concern formal synthe-
ses that converge on late Tsuda intermediates, see Ref. 6.