8638 J. Am. Chem. Soc., Vol. 122, No. 36, 2000
Takayama et al.
CHCl3. The organic layer was dried over magnesium sulfate and
evaporated to give a crude alkaloidal fraction (2.56 g). The portion of
the crude base (1.10 g) was roughly separated by silica gel flash column
chromatography eluted with a CHCl3 to MeOH/CHCl3 gradient to give
the 10 fractions. The 2-5% MeOH/CHCl3 eluate was subjected to SiO2
medium-pressure liquid chromatography using 2% EtOH/CHCl3 to give
95 mg of pandamarilactonine-A (1), 10 mg of pandamarilactonine-B
(2), and 9 mg of pandamarilactone-1 (3).
of the final stage in the present synthesis was not satisfactory,
we succeeded in the first and biomimetic total synthesis of the
new alkaloids, which provided chemical support for the sug-
gested biogenetic route of 1 and 2 as well as proof for the
chemical structures deduced by spectroscopic analysis.
Optical Purity of Natural Pandamarilactonine-A and -B.
As described in the isolation section above, pandamarilacto-
nine-A (1) exhibited [R]23D +35.0° (c 4.37, CHCl3); in contrast,
the specific rotation of pandamarilactonine-B (2) was almost
zero. We were interested in the optical purity of these natural
products.14 With the synthetic racemates in hand, we first
examined the resolution of the enantiomers using chiral column
chromatography. Synthetic pandamarilactonine-A (1) exhibited
two peaks at 43.2 and 51.7 min when Chiralcel OB was used,
and synthetic pandamarilactonine-B (2) showed two peaks at
18.7 and 20.4 min by using Chiralcel OD. Then, natural
alkaloids were subjected to chiral HPLC analysis, which
established that the natural pandamarilactonine-A (1) contained
dominantly the (+)-enantiomer over the (-)-enantiomer in a
ratio of 63:37, while natural 2 existed as a racemate. On the
basis of the biosynthetic proposal and the chemical conversion
of 9 to 1 and 2, it is conceivable that pandamarilactonine-A
and pandamarilactonine-B undergo acid-catalyzed interconver-
sion during the isolation process and that pandamarilactonine-B
(2) is an artifact from pandamarilactonine-A (1). To test this
hypothesis, the natural alkaloids 1 and 2 were respectively
treated with 5% H2SO4 under the conditions used in the
partitioning of alkaloids and the recovered alkaloids were
carefully analyzed with 500 MHz NMR spectroscopy. In the
recovered 1 and 2, the presence of the stereoisomers and/or
pandamarilactonine-1 could not be observed. Further, recovered
pandamarilactonine-A exhibited same optical purity (63:37) by
chiral HPLC analysis, revealing that no racemization has
occurred. These experiments demonstrate that pandamarilacto-
nine-A and -B are not interconvertible under the isolation
conditions and that pandamarilactonine-B is not an artifact
produced during the isolation process.
In summary, two new alkaloids having a pyrrolidinyl R,â-
unsaturated γ-lactone moiety and a γ-alkylidene R,â-unsaturated
γ-lactone residue were isolated from a tropical medicinal plant,
P. amaryllifolius. Their structures were first deduced by
spectroscopic analysis including the new NMR technique PFG
J-HMBC 2D spectroscopy and then confirmed by biomimetic
total synthesis. Interestingly, it was found that one of the
diastereoisomers, pandamarilactonine-A (1) having the erythro
structure, was comprised of a mixture enriched with the (+)-
enantiomer, while another diastereomeric threo isomer, pan-
damarilactonine-B (2), occurred as a racemate in nature.
Further investigation of the minor constituents in this plant,
development of an efficient synthetic pathway of 1 and 2,
determination of the absolute configuration of (+)-pandama-
rilactonine-A, and biological evaluation of these alkaloids are
in progress in our laboratories.
Pandamarilactonine-A (1). An amorphous powder. Rf value 0.5
[SiO2, solvent system 5% MeOH in CHCl3]. [R]23 +35.0° (c 4.37,
D
CHCl3). UV (MeOH): λmax (nm) (log ꢀ) 275 (2.43), 232 (sh), 207
(1.82). IR (neat): νmax (cm-1) 1750 (lactone). FABMS (NBA): m/z
318 [M + H]+. HR-FABMS (NBA): calcd for C18H24NO4 318.1704,
1
found 318.1721. H NMR (500 MHz, CDCl3): δ 7.09 (1H, dd, J )
1.5 and 1.8 Hz, H-16), 6.99 (1H, d-like, J ) 1.5 Hz, H-4), 5.18 (1H,
dd, J ) 7.9 and 7.9 Hz, H-6), 4.80 (1H, ddd, J ) 1.8, 1.8 and 5.5 Hz,
H-15), 3.12 (1H, dd, J ) 6.7 and 7.6 Hz, H-11), 2.88 (1H, ddd, J )
4.0, 7.9 and 12.9 Hz, H-9), 2.83 (1H, m, H-14), 2.45 (1H, m, H-9),
2.43 (2H, dd, J ) 7.3 and 15.0 Hz, H2-7), 2.21 (1H, m, H-11), 1.99
(3H, d-like, J ) 0.9 Hz, H3-21), 1.93 (3H, dd, J ) 1.5 and 1.8 Hz,
H3-20), 1.70-1.80 (2H, m, H-12 and H-13), 1.59-1.70 (3H, m, H2-8
1
and H-13), 1.42 (1H, m, H-12). H NMR (400 MHz, acetone-d6): δ
7.27 (1H, dd, J ) 1.5 and 2.9 Hz, H-16), 7.26 (1H, ddd, J ) 1.7, 1.7
and 3.2 Hz, H-4), 5.33 (1H, dd, J ) 7.8 and 8.1 Hz, H-6), 4.86 (1H,
ddq, J ) 4.9, 2.9 and 2.0 Hz, H-15), 3.10 (1H, ddd-like, J ) 2.0, 7.1
and 8.5 Hz, H-11), 2.90 (1H, ddd, J ) 8.1, 8.1 and 11.7 Hz, H-9),
2.85 (1H, ddd, J ) 4.9, 4.9 and 9.3 Hz, H-14), 2.44 (1H, ddd, J ) 4.9,
6.8 and 11.7 Hz, H-9), 2.38 (2H, ddd, J1 ) J2 ) J3 ) 7.6 Hz, H-7),
2.19 (1H, ddd, J ) 6.1, 8.8 and 10.2 Hz, H-11), 1.90 (3H, dd-like, J
) 0.5 and 1.5 Hz, H-21), 1.83 (3H, dd, J ) 1.5 and 2.0 Hz, H-20),
1.57-1.77 (5H, m, 2 × H-8, H-12, and 2 × H-13), 1.42 (1H, m, H-12).
13C NMR (125 MHz, CDCl3): δ 174.3 (C-18), 171.1 (C-2), 148.6 (C-
5), 147.0 (C-16), 137.7 (C-4), 131.2 (C-17), 129.1 (C-3), 114.1 (C-6),
83.4 (C-15), 65.3 (C-14), 55.0 (C-9), 54.2 (C-11), 28.3 (C-8), 25.7
(C-12), 24.0 (C-7), 23.8 (C-13), 10.7 (C-20), 10.5 (C-21).
Pandamarilactonine-B (2). An amorphous powder. Rf value 0.5
[SiO2, solvent system 5% MeOH in CHCl3]. [R]23D 0° (c 0.20, CHCl3).
UV (MeOH): λmax (nm) (log ꢀ): 275 (2.49), 232 (sh), 207 (1.72). IR
(neat): νmax (cm-1) 1758 (lactone). FABMS (NBA): m/z 318 [M +
H]+. HR-FABMS (NBA): calcd for C18H24NO4 318.1704, found
1
318.1704. H NMR (500 MHz, CDCl3): δ 7.05 (1H, dd, J ) 1.5 and
1.7 Hz, H-16), 7.00 (1H, d-like, J ) 1.5 Hz, H-4), 5.18 (1H, dd, J )
7.8 and 8.0 Hz, H-6), 4.71 (1H, ddd, J ) 1.7, 2.0 and 5.9 Hz, H-15),
3.12 (1H, m, H-11), 2.73 (1H, m, H-9), 2.70 (1H, m, H-14), 2.42-
2.48 (2H, m, H-7 and H-9), 2.36 (1H, m, H-7), 2.25 (1H, m, H-11),
1.99 (3H, d-like, J ) 0.7 Hz, H3-21), 1.93 (3H, dd, J ) 1.7 and 1.7
Hz, H3-20), 1.73-1.87 (4H, m, H2-12 and H2-13), 1.59-1.67 (2H, m,
H2-8). 1H NMR (400 MHz, acetone-d6): δ 7.28 (1H, dd-like, J ) 1.5
and 2.9 Hz, H-16), 7.26 (1H, ddd, J ) 1.5, 1.7 and 3.2 Hz, H-4), 5.32
(1H, dd-like, J ) 7.8 and 8.1 Hz, H-6), 4.81 (1H, ddq, J ) 3.9, 2.9
and 2.0 Hz, H-15), 3.08 (1H, m, H-11), 2.80 (1H, ddd, J ) 8.1, 8.1
and 12.0 Hz, H-9), 2.77 (1H, ddd, J ) 3.9, 5.1 and 8.8 Hz, H-14),
2.45 (1H, m, H-9), 2.33 (2H, ddd, J ) 7.8, 7.8 and 15.6 Hz, H-7),
2.24 (1H, m, H-11), 1.92 (3H, dd-like, J ) 0.7 and 1.5 Hz, H-21),
1.83 (3H, dd-like, J ) 1.5 and 2.0 Hz, H-20), 1.59-1.77 (6H, m, 2 ×
H-8, 2 × H-12, and 2 × H-13). 13C NMR (125 MHz, CDCl3): δ 174.3
(C-18), 171.1 (C-2), 148.5 (C-5), 147.5 (C-16), 137.7 (C-4), 130.8 (C-
17), 129.1 (C-3), 114.1 (C-6), 83.4 (C-15), 66.3 (C-14), 55.8 (C-9),
54.2 (C-11), 28.4 (C-8), 27.1 (C-12), 24.0 (C-7), 24.0 (C-13), 10.8
(C-20), 10.5 (C-21).
Experimental Section15
Extraction and Isolation of Alkaloids. Fresh young leaves (1.2
kg) of P. amaryllifolius purchased at flower market in Bangkok
(Thailand) were macerated with ethyl alcohol (6 L) three times, and
filtered. The combined filtrate was concentrated under reduced pressure
to give a crude extract (105 g), which was then partitioned between
Et2O and 5% aqueous H2SO4. The water-soluble fraction was alkalinized
with concentrated NH4OH until pH 10 and exhaustively extracted with
Experimental Conditions for PFG J-HMBC 2D Spectroscopy.
All NMR experiments were performed at 303 K for solution of ca. 10
mg of alkaloid dissolved in 0.5 mL of acetone-d6 on a JEOL LA600
spectrometer equipped with a 5 mm NALORAC HX inverse probe.
Fifteen J-HMBC 2D spectra were acquired with 16 scans for a 1024
(F2) × 128 (F1) data matrix and with 300 ms of constant time and
varying evolution time (∆) from 10 to 290 ms in 20 ms steps. The
HMBC signal intensity is given by a sine function of the evolution
time (∆) and coupling constant (JHX). Therefore, the J value (coupling
constant) can be obtained from a least-squares approximation by fitting
(14) Takayama, H.; Kurihara, M.; Kitajima, M.; Said, I. M.; Aimi, N. J.
Org. Chem. 1999, 64, 1772-1773.
(15) See the Supporting Information for general procedures.