Organic Letters
Letter
and C10 positions of parvistemoamide largely hinders the
synthetic and pharmacological studies on parvistemoamide and
medium-sized-ring-containing Stemona alkaloids.
successfully achieved, parvistemoamide (3) could be synthe-
sized and, more importantly, the stereochemistry of
parvistemoamide and its transformation relationship with
stemoamide (4) would be clearly revealed, which may further
promote the total syntheses of other more complex
stemoamide-type alkaloids.
At same time as reporting the isolation of parvistemoamide,
Xu simultaneously reported that the 10-membered lactam in
parvistemoamide (3) may come from the pyrrolo[1,2-a]-
azepine nucleus in stemoamide (4) by C9a−N oxidative
cleavage.4e Afterward, Pilli and co-workers revealed their
synthetic study on parvistemoamide in 2005 and proposed a
possible biomimetic conversion of parvistemoamide to
stemoamide via oxidation at C9a, a nucleophilic attack by an
activated form of the amide nitrogen, and the following
dehydration and reduction of the intermediate N-acyliminium
ion,5a which is different from Xu’s proposal4e (Figure 1).
However, in Pilli’s work only a 5:1 mixture of two lactams with
C9a−α-OH and C10−α-Me and C9a−β-OH, C10−β-Me was
achieved, and no any NMR data were provided.5a More
importantly, neither Xu’s nor Pilli’s viewpoint on the
transformation relationship between parvistemoamide and
stemoamide has been studied by experiments until now,
although many elegant formal or total synthesis studies toward
stemoamide have been revealed.6 Therefore, the relative
configuration of parvistemoamide and its transformation
relationship with stemoamide are still unsolved mysteries.
In this regard, accomplishing the total syntheses of the
possible structures of parvistemoamide may help to confirm its
stereochemistry and, more importantly, promote the study of
the transformation relationship of Stemona alkaloids. In the
continuation of our research interest in Stemona alkaloids,7 we
herein (A) report the total syntheses of four stereoisomers of
parvistemoamide (3) via macrolactamization and (B) reveal
the transformations of the possible structures of parvistemoa-
mide (3) to stemoamide (4) and 9a-epi-stemoamide by
transannular cyclization or Pilli’s transformation, although the
unambiguous structural determination for parvistemoamide
was not accomplished in this work.
With the above retrosynthetic analysis in mind, we
commenced the synthesis of stereoisomers 3a−3d, which
encompass the two structures proposed for parvistemoamide
(3) by Xu and co-workers4e,5c (Scheme 2). Starting from
ketone 6,8 the protection of the ketone carbonyl group with
1,3-dimercaptopropane and the hydrolysis of the methyl ester
with LiOH generated carboxylic acid 7. Under the promotion
of TMSI, the carboxybenzyl group (Cbz) in 7 was deprotected
to render the unprotected amine, which was then subjected to
the Corey−Nicolaou macrolactamization reaction to construct
the 10-membered lactam 8 in a 50% yield in a rather high-
dilution solvent (c = 1.0 × 10−3 M). Then, the above-generated
unprotected amide was protected by Boc to deliver 9 in a 71%
yield, which was identified by X-ray analysis. Sequentially, 9
underwent the oxidative deprotection of the 1,3-dithiane
protecting group by PIFA, affording ketone 10 in a 72% yield.
We subsequently studied the stereoselective reduction of
ketone 10. When subjecting ketone 10 to reducing reagents
such as sterically hindered LiAl(tBuO)3H, the Boc-protected
10-membered lactam in the desired product was susceptible to
nucleophilic ring opening, and an unexpected lactam−lactone
exchange product formed via the ring opening of lactam and
the lactonization of the amide carbonyl group with the newly
This is mainly due to the existence of Boc weakening the p−π
conjugation effect between the amide nitrogen atom and the
amide carbonyl group, thus weakening the C−N bond and
increasing the electrophilicity of amide carbonyl group.9
Pleasingly, when the reaction mixture was flash-quenched
with 3 M HCl, only a single diastereoisomer with C9a−α-OH
was afforded in a 73% yield. The subsequent sequential
C9a−α-OH protection by TES and the stereospecific
methylation at C10 rendered 11a smoothly. When 11a was
treated with BF3·OEt2, TES was removed prior Boc, the
lactam−lactone exchange byproduct was formed (see the SI).
After exhaustive attempts, Mg(ClO4)2, which was compatible
with the sensitive TES, was considered as the most appropriate
reagent to remove Boc.
Based on the structural characters of parvistemoamide, we
propose the following retro-synthetic strategy (Scheme 1). The
Scheme 1. Retro-Synthetic Analysis of Stemoamide and
Parvistemoamide
Then, TES was deprotected by CAN, accomplishing the
synthesis of 3a. After the attempted introduction of a double
bond in C10 failed (see the SI), the most commonly employed
chiral adjustment method in the synthetic studies on Stemona
alkaloids was used. As a result, 3a (β-Me) and 3b (α-Me) were
obtained in a 39% total yield in a diastereoselective ratio of 1:1
when 3a was subjected to K2CO3/MeOH. The structures of 3a
by X-ray crystallographic analysis, and 3b was found to be
identical to the structure proposed by Xu.4e,5b However, the
1H and 13C NMR data of neither 3a nor 3b are the same as
those of the natural sample. One of the most significant
differences in the1H NMR data is that there is an additional set
10-membered lactam ring in parvistemoamide is proposed to
be constructed via the macrolactamization of the amphoteric
compound 5, which can be further traced back to ketone 6 via
the protection of the keto-carbonyl group and the deprotection
of the ester and amine. Ketone 6 is a known compound and
could be easily obtained.8 On the other hand, parvistemoamide
may be derived from stemoamide via the C9a−N oxidative
cleavage proposed by Xu,4e or stemoamide could come from
parvistemoamide via the C−N bond formation proposed by
Pilli5a or transannular cyclization. If the above proposal is
of peaks at 2.80−3.0 ppm in our reported data. As for the 13
C
NMR data, it is obvious that both 3a and 3b are not the
naturally occurring parvistemoamide.
Thus, we have to resort to the epimers at the C9a position.
After several trials, NaBH4 gave a 3.8:1 ratio of two
diastereoisomers in favor of β-OH (see the SI). Following a
6223
Org. Lett. 2021, 23, 6222−6226