Access to the akuammiline family of alkaloids: Total synthesis of (+)-scholarisine A

Article abstract of DOI:10.1021/ja3111626

The planning and implementation of an enantioselective total synthesis of (+)-scholarisine A is presented. Key tactics employed include a novel cyclization, consisting of a nitrile reduction coupled with concomitant addition of the resultant amine to an epoxide; a modified Fischer indolization; an oxidative lactonization of a diol in the presence of an indole ring; and a late-stage cyclization to complete the caged ring scaffold. The development of a possible "retro-biosynthetic" approach to other members of the akuammiline alkaloid family is also described.

Full text of DOI:10.1021/ja3111626

Article
pubs.acs.org/JACS
Access to the Akuammiline Family of Alkaloids: Total Synthesis of
(+)-Scholarisine A
Gregory L. Adams, Patrick J. Carroll, and Amos B. Smith, III*
Department of Chemistry, Laboratory for Research on the Structure of Matter, and Monell Chemical Senses Center, University of
Pennsylvania, Philadelphia, Pennsylvania 19104, United States
S
* Supporting Information
ABSTRACT: The planning and implementation of an enantioselective total
synthesis of (+)-scholarisine A is presented. Key tactics employed include a novel
cyclization, consisting of a nitrile reduction coupled with concomitant addition of
the resultant amine to an epoxide; a modified Fischer indolization; an oxidative
lactonization of a diol in the presence of an indole ring; and a late-stage cyclization
to complete the caged ring scaffold. The development of a possible “retro-
biosynthetic” approach to other members of the akuammiline alkaloid family is
also described.
provide an aldehyde at C-5. Double bond migration of the E-
ethylidine could then furnish an enamine, which could undergo
nucleophilic addition to the aldehyde. Finally, lactonization
between the resultant hydroxyl group and the methyl ester
would complete the caged scaffold of scholarisine A (1).1
Importantly, several members of the akuammiline alkaloid
class are known to have bioactivity. Derivatives of picraline
(3)20 are reported to be selective inhibitors of the receptor
SGLT2, a renal cortex membrane protein recently validated as a
target for type II diabetes treatment,21 while aspidophylline A
(9) and echitamine (14) respectively reverse drug resistance in
cancer cells8 and display in vivo anti-tumor activity in rodents.22
Earlier this year, we reported the first total synthesis and
assignment of the proposed absolute configuration of
(+)-scholarisine A (1).23 Herein we report a full account of
the rationale behind the planning of this synthetic venture, in
conjunction with implementation of the synthetic strategy, as
well as initial experiments that hold the promise of providing
access to other members of the akuammiline family via a “retro-
biosynthetic” fragmentation.
Structural Insight into the Akuammiline Family of
Alkaloids. After considering the proposed biosynthetic route
to (+)-scholarisine A (1), with specific attention focused on
aldehyde 2 (Figure 1), we began to view these alkaloids in a
structurally open format (cf. 15 and 16, Figure 2).
Inspired by indoline diol 15, a degradation product of
aspidodasycarpine (8) reported by the Djerassi group in 1964,7
we reasoned that if such a carbon scaffold could be accessed,
redox manipulation with functionalization might provide access
to the entire akuammiline class. For example, ester indolenine
aldehyde 16 could collapse to form picrinine (4). Reduction of
the aldehyde in 16 would also provide alcohol 18, which upon
cyclization to a furoindoline core could furnish desformoaspi-
INTRODUCTION
■
Scholarisine A (1), an akuammiline monoterpene indole
alkaloid isolated from the leaves of Alstonia scholaris by Luo
and co-workers, possesses an architecturally intricate cage-like
scaffold containing a bridged lactone, an aliphatic imine, and an
indolenine core (Figure 1).1 From the outset, the intrinsic
synthetic challenge involved in constructing this novel
architecture attracted our attention. The possibility of devising
a substrate that would also provide access to other members of
the akuammiline alkaloid family via a late-stage “retro-
biosynthetic” fragmentation further enhanced our interest in
this alkaloid class. We therefore initiated synthetic studies
toward the construction of (+)-scholarisine A (1) in 2008.
Alkaloids that derive biosynthetically via cyclization of
geissoschizine (5), forming a bond between C-7 and C-16,2
are considered members of the akuammiline alkaloid family
(Figure 1).3 Akuammiline (7),4 the progenitor from which this
class derives the name, was first characterized in 1932.5 Acetate
hydrolysis and deformylation is believed to give rise to the
strictamine congener (6),6 while hydrolysis of the amine moiety
would permit formation of the furoindoline core, observed in
aspidodasycarpine (8)7 and aspidophylline A (9),8 the latter the
subject of a recent elegant total synthesis by the Garg group.9
Akuammiline (7) could also give rise to picraline (3)10 upon
oxidation at C-5, while additional loss of the acetoxymethyl
moiety from C-16 would lead to picrinine (4).11 Further
functionalization of the picrinine core would result in formation
of lanciferine (10),12 nareline (11),13 and arbophylline (12),14
while vincorine (13)15 and echitamine (14)16 consist of a
scaffold containing a nitrogen shift.17 A racemic synthesis of
vincorine (13) was reported by Qin in 2009,18 while more
recently, in 2012, Ma reported an asymmetric total synthesis.19
Scholarisine A (1), the initial target of our efforts, is
proposed to arise biosynthetically via rearrangement of
picrinine (4) as illustrated in Figure 1. Opening of the
hemiaminal ether oxygen bridge between C-2 and C-5 would
Received: November 13, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja3111626 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 2. Open form view of akuammiline alkaloids.
Figure 1. Representative akuammiline alkaloids.
dodasycarpine (17).7 Supporting this concept, Garg et al.
recently demonstrated that the N-tosyl derivative of alcohol 18
spontaneously forms the furoindoline core during their
synthesis of apspidophylline A (9).9 Alternatively, activation
of the hydroxyl present in 18 could lead, via cyclization, to
strictamine (6). Furthermore, oxidation of the amine in 16 to
nitrone 19 would provide a substrate that might undergo a
[3+2] cycloaddition to provide nareline (11).13 Alternative
redox manipulation and functionalization could also provide the
open substrates 20 and 21, which upon collapse could lead to
lanciferine (10) and arbophylline (12), respectively.
Total Synthesis of Scholarisine A: A Unified Route to
the Akuammiline Alkaloids. With the above considerations
in mind, we initially targeted indolenine ketone 23 (Figure 3)
as a versatile late-stage intermediate to devise a unified
approach to the akuammiline core. This substrate was
envisioned to provide synthetic access to scholarisine A (1),
as well as other members of the akuammiline family, via a series
of fragmentations and cyclizations (Figure 3).
For example, indolenine ketone 23 was envisioned to
undergo reduction followed by deprotection to provide
indoline diol amine 22, which could then be oxidized to either
scholarisine A (1) or diimine 25, that in turn might undergo an
aza retroaldol-type fragmentation to furnish aldehyde enamine
24, a scaffold reminiscent of the structures seen in Figure 2.
Alternatively, nitrogen deprotection and oxidation of the
amine in 23 could lead to diimine 26, which upon deprotection
of the oxygen might elicit a fragmentation reaction consisting of
the initially formed hydroxy ketone 29 collapsing to form lactol
Figure 3. Fragmentation approach to the akuammiline core.
28, which could then undergo fragmentation to render lactone
27.24
For each of the latter constitutional isomers, 27, 28, and 29,
the relative energies were calculated at the geometry-optimized
B3LYP/6-31G(d) level. The computationally determined
relative energies suggest lactone 27 to be the most energetically
stable isomer.
B
dx.doi.org/10.1021/ja3111626 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
We consider the proposed fragmentation processes (22 to 24
and 26 to 27) to comprise a reversal of the biosynthetic path to
scholarisine A (1) (Figure 1). Further manipulation of 24 and
27 to the oxidation states and functionalities illustrated in
Figure 2 might provide access to the entire akuammiline family.
Central to our initial synthetic analysis (Figure 4), indolenine
23 was envisioned to arise via a cyclization involving the acidic
available anhydride 35 in cold tetrahydrofuran,28 was subjected
to a modification of a known27 enzymatic asymmetrization
employing porcine pancreatic lipase in a reaction medium of
methyl acetate containing 4 Å sieves; the latter were introduced
to scavenge the methanol byproduct of enzyme acylation, thus
permitting a manageable amount of methyl acetate to be
employed as solvent upon reaction scale-up.29 Without
separation, the resultant mixture of monoacetate (+)-37 and
diacetate 38 (3.4:1) was directly treated with toluenesulfonyl
chloride in pyridine and dichloromethane (1:1 v/v) at room
temperature to furnish tosylate (−)-39. The reaction mixture
was then subjected to an aqueous workup after quenching the
excess tosyl chloride with N-methylpiperazine to create a water-
soluble sulfonamide.30 After workup, treatment of the product
mixture with potassium cyanide31 in hot dimethyl sulfoxide
furnished nitrile (+)-40. Basic hydrolysis was followed by
extraction with dichloromethane to remove impurities and diol
36, which formed upon saponification of diacetate 38, the latter
proving inert to the previous two reactions. Lactone (−)-32, in
turn, was obtained upon acidification in approximately 70%
yield from diol 36.
Figure 4. Initial retrosynthetic analysis.
decomposition of α-diazo ketone 3025 to install the quaternary
center.26 Cyclization substrate 30 in turn could be obtained via
carboxylate activation and homologation of the protected seco
acid of lactone 31, which was envisioned to arise from tricyclic
amine 34 after protection, oxidation, and Fischer indolization.
Amine 34, a central player in our proposed synthesis of
(+)-scholarisine A (1), would result via a selective nitrile
reductive cyclization of epoxide 33, which would derive after
functionalization from known lactone (−)-32.27 Importantly,
the convex−concave nature of lactone (−)-32 would dictate
the correct configuration at both the quaternary center and the
oxirane ring in lactone 33. Substrate control, once the correct
stereogenicity of (−)-32 was established, would thus dictate all
of the remaining stereogenic centers required to access the
(+)-scholarisine A (1) scaffold.
Epoxide (−)-33. With lactone (−)-32 in hand, the focus
shifted to the preparation of the cyclization substrate, epoxide
33. Toward this end, the anion of lactone (−)-32, derived by
reaction with 2.2 equiv of lithium diisopropyl amide in
tetrahydrofuran, was treated with cyanobenzotriazole32 at 0
°C to furnish nitrile (+)-4133 in 60% yield, after a workup
involving aqueous extraction (Scheme 2).
Scheme 2. Synthesis of Epoxide (−)-33
RESULTS AND DISCUSSION
■
Lactone (−)-32. At the outset, a synthetic route permitting
access to large amounts of lactone (−)-32 was developed and
implemented (Scheme 1). To this end, meso diol 36, obtained
via lithium aluminum hydride reduction of commercially
Scheme 1. Synthesis of Lactone (−)-32
Next, ethyl cyanolactone (−)-4233 was generated in 71%
yield upon addition of ethyl iodide in a warm slurry of
potassium carbonate and acetonitrile. As anticipated, alkylation
occurred from the convex face of cyanolactone (+)-41 to
deliver the requisite configuration at the quaternary center.
Epoxidation to prepare the scaffold for reductive cyclization was
then achieved via use of mCPBA in chloroform; a mixture of
diastereomeric epoxides, consistent with expectations based on
literature precedent, resulted.34 The desired epoxide (−)-3333
was formed in preference to (−)-4333 in a ratio of 3:1.
Crystallization from toluene provided pure (−)-33 in 58%
yield.
Construction of Ketones (+)-44 and (+)-45: Fischer
Indole Substrates. Orchestration of what proved to be a
C
dx.doi.org/10.1021/ja3111626 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
novel reductive cyclization of epoxide (−)-33 to access
(+)-4433 (Scheme 3) was achieved via nitrile hydrogenation,
followed by in situ epoxide ring-opening to generate tricylic
amine (−)-34.33,35 The reaction was performed employing a
Scheme 5. Initial Indole Synthesis
Scheme 3. Synthesis of Ketone (+)-44
pressure bomb at ca. 100 psi of hydrogen, with rhodium on
alumina as the catalyst. The commonly employed nitrile
hydrogenation medium of alcoholic ammonia could not be
used in this case due to decomposition of lactone (−)-33.
Tetrahydrofuran was thus employed as the solvent, but heating
was required, since the reaction did not progress at room
temperature. Even with heating, the reaction proved to be slow.
Ketone (+)-44, employed as the initial substrate for the
Fischer indole synthesis, was prepared via protection of amine
(−)-34 with di-tert-butyl dicarbonate followed by Dess−Martin
oxidation.36 Improved cyclization conditions were subsequently
discovered (Scheme 4), wherein a mixture of acetic acid and
formed, but only decomposition followed. Although the
products of decomposition were not isolated, a possible route
for the decomposition of 46 was envisioned to entail proton
abstraction, coupled with the protected nitrogen acting as a
leaving group, as illustrated in Scheme 5.
To determine if this decomposition pathway pertained, the
possibility of proton abstraction was removed via substitution
of the involved hydrogen with a benzoyl group. A stable
hydrazone 47 could then be obtained upon subjecting (+)-44
to benzoylphenylhydrazine39 in pyridine containing aqueous
hydrochloric acid (Scheme 5).40 The weakly acidic reaction
medium of pyridine/hydrochloric acid was chosen to retain the
nitrogen protection. Under these conditions the hydrazone
proved stable, even upon heating at 100 °C for 5 days.
However, upon raising the temperature to 180 °C for 5 h,
employing microwave radiation, two different indole products
were obtained; indole amide (−)-49 resulted in a 37% yield,
with indole (−)-5033 isolated as a minor product. We reason
that during the course of this reaction, the secondary alkyl
nitrogen undergoes exchange of protection via thermal
decomposition of the carbamate, concurrent with the Fischer
annulation process, to form first the benzoyl indole 48 as a
transient intermediate41 that then undergoes benzoyl transfer
either bimolecularly or in an intramolecular fashion.42 Indole
(−)-50 results via competing hydrolysis of benzoyl indole 48.43
After we demonstrated that indole (−)-49 was a viable
synthetic intermediate (vide infra), an alternate route to (−)-49
more amenable to scale-up was devised, which directly
introduced benzoyl protection by employing ketone (+)-45.
Subjecting this ketone to benzoyl phenylhydrazine in a solution
of acetic acid and trifluoroacetic acid (TFA) (1:1) for 1 week at
75 °C furnished (−)-49 in 57% yield (Scheme 6).44
Scheme 4. Synthesis of Ketone (+)-45
tetrahydrofuran (1:1) was employed.37 Under these conditions
the reaction proceeds at room temperature at a reasonable rate.
Tetrahydrofuran, which is known to suppress epoxide hydro-
genolysis when used as a solvent for catalytic hydrogenation of
alkenes, is required in this case as a co-solvent to achieve a clean
reaction product.38 Uncharacterized side products were
observed when pure acetic acid was used as the solvent.
The product from this hydrogenation was next subjected,
without purification, directly to protection of the secondary
amine in this case with benzoyl chloride, employing
diisopropylethylamine (DIEA) as an acid scavenger (Scheme
4). After an aqueous workup, oxidation with Dess−Martin
periodinane furnished ketone (+)-45,33 which was sufficiently
pure to use directly in the next step.
Construction of the Indole Ring. With ketones (+)-44
and (+)-45, possessing the 2-azabicyclo[3.3.1]nonan-8-one ring
system, in hand, we explored installation of the indole. Initial
attempts at Fischer annulation, employing phenylhydrazine and
ketone (+)-44, proved unsuccessful (Scheme 5). Following the
reaction by LCMS, however, indicated that the hydrazone was
While this route provided sufficient indole (−)-49 to advance
the synthesis, the acidic conditions and long reaction time
appeared to cause partial degradation of the product as the
reaction progressed. Carbazole (−)-51,33 observed as a minor
byproduct, probably forms via protonation of the amide moiety,
followed by an E1 elimination and oxidation. Involvement of
the nitrogen atom in this elimination pathway lends further
support to the proposed hydrazone decomposition pathway
outlined in Scheme 5. Later studies established that benzyl
D
dx.doi.org/10.1021/ja3111626 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
group, resulted only in lactone ring closure. Attempts to
convert the lactone directly to the corresponding Weinreb
amide using either basic or acidic conditions, or treatment of
the lactone with potassium trimethylsilanol in tetrahydrofuran,
also proved unproductive.
Scheme 6. Synthesis of Indole (−)-49
Lactone (−)-49 could however be reduced with lithium
aluminum hydride to furnish diol (+)-54. Equally important,
the less hindered hydroxyl of the resultant diol (+)-54 could be
protected selectively as the TBDPS ether46 to furnish (+)-55.
Oxidation of Alcohol (+)-55 to the Carboxylic Acid:
Another Difficult Transformation. Indole alcohol (+)-55
was next subjected to oxidation with tetrapropylammonium
perruthenate, employing N-methylmorpholine N-oxide as the
co-oxidant (TPAP/NMO);47 a mixture of products resulted
(Scheme 9).
instead of benzoyl protection of the phenylhydrazine would
prove beneficial. To this end, treatment of ketone (+)-45 with
benzyl phenylhydrazine45 employing the slightly acidic medium
of pyridine/hydrochloric acid at 110 °C overnight furnished the
benzyl-protected indole (−)-52 in greater than 70% yield
(Scheme 7).
Scheme 9. Synthesis of Aldehyde (+)-56
Scheme 7. Synthesis of Indole (−)-52
Lactone Ring-Opening and Hydroxyl Protection: A
Difficult Transformation. Once indole (−)-49 was obtained,
attempts were made to construct seco acid 53 (Scheme 8).
The desired aldehyde (+)-56 was obtained along with cyclic
ether (−)-58; the ratio was 2:1. The mixture however proved
very difficult to separate by chromatography. We envision that
cyclic ether (−)-58 arises via competing indole oxidation as
illustrated in Scheme 9.48 Fortunately, pure aldehyde (+)-5633
could be obtained via Parikh−Doering oxidation49 in a
moderate yield of 50%.
Scheme 8. Lactone Opening and Hydroxyl Protection
Multiple conditions were next employed to oxidize aldehyde
(+)-56, but the desired carboxylate product 53 (Scheme 8) was
not observed. The only products that could be observed
resulted from indole oxidation in a fashion similar to that
illustrated in Scheme 9 [cf. (−)-58].48c
We reasoned that the sterically encumbered environment of
the aldehyde carbonyl in (+)-56 led to our inability to obtain
the carboxylic acid. Presumably, formation of the requisite
aldehyde hydrate required for oxidation was unfavorable due to
an increase in steric interactions. In support of this scenario,
aldehyde (+)-56 did not form a cyanohydrin or bisulfite adduct.
Attempts at methylenation of aldehyde (+)-56 also proved
unsuccessful. Aldehyde (+)-56 did however react readily with
alkyllithium reagents. A decision was thus made to define the
stereochemical outcome of this reaction. Obtaining the correct
stereogenicity upon addition of a carbon nucleophile to the
aldehyde would orchestrate the same oxidation state and
configuration required for (+)-scholarisine A (1). To explore
this stereochemical issue we selected aldehyde 61 (Scheme 10),
where both nitrogens were similarly protected with benzyl
groups.
Carboxylate activation in 53 followed by reaction with
diazomethane would provide α-diazoketone 30, a central
intermediate in our synthetic analysis (see Figure 4). To
achieve this transformation, opening of the lactone ring in
(−)-49 with selective protection of the resulting hydroxyl
group would be required.
Unfortunately, all efforts to achieve this transformation failed.
Upon treatment with aqueous sodium hydroxide, lactone ring-
opening could be observed by LCMS, but all further
manipulations of the resultant hydroxy carboxylate, including
addition of trialkyl silyl chlorides to capture the free hydroxyl
E
dx.doi.org/10.1021/ja3111626 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 10. Synthesis of Diol (+)-62
Figure 5. NOE-derived model of the reactivity of (+)-61.
Figure 6. A new plan forward.
Reaction of Aldehyde (+)-61 with Benzyloxymethyl-
lithium. To arrive at aldehyde (+)-61, indole (−)-52 was
reduced employing lithium aluminum hydride to furnish diol
(+)-59;33 the nitrogen benzoyl group was also reduced to the
benzyl amine. Diol (+)-59 was then selectively protected as the
TBDPS ether to furnish (+)-60. This three-step sequence
proved highly effective, proceeding in a 70% yield. Oxidation of
alcohol (+)-60 with TPAP/NMO completed construction of
aldehyde (+)-61. In this case protection of the indole nitrogen
in (+)-60 blocks competing indole ring oxidation.
With aldehyde (+)-61 in hand, we examined the stereo-
chemical outcome upon addition of an alkyllithium species.
Pleasingly, in situ generation of benzyloxymethyllithium in
tetrahydrofuran at −78 °C employing the Still protocol,50
followed by addition of aldehyde (+)-61 to the solution,
afforded diol (+)-62 in 63% yield after hydrolytic workup to
remove the TBDPS group. The undesired diastereomer, diol
(+)-63, was also obtained in 11% yield. That the required
stereogenicity of (+)-62 for (+)-scholarisine A (1) was
obtained, was initially based on NOESY NMR studies that
indicated that the average conformation of the aldehyde
carbonyl in (+)-61 is as depicted in Figure 5. Nucleophilic
attack occurring from the external face of the carbonyl would
lead to the formation of diol (+)-62. The NMR-defined
aldehyde conformation was also observed in the X-ray structure
of the unprotected indole (+)-56 (see Scheme 9). Con-
firmation of this stereochemical assignment was subsequently
achieved both by NMR NOE correlations observed in lactone
(−)-65 and by X-ray analysis of (+)-74 (vide infra).
performed, followed by protecting group exchange, to yield
indole alcohol 64 (Figure 6). Activation of the hydroxyl group
would permit cyclization to form the indolenine quaternary
center. Deprotection followed by amine oxidation would then
complete the total synthesis of (+)-scholarisine A (1).
Lactonization of Diol (+)-62. Although initial attempts to
oxidize diol (+)-62, employing TPAP/NMO as the oxidant,
proved successful to furnish the desired lactone (−)-65, the
yield (36%) was only modest (Scheme 11).
To understand this oxidation, the reaction was interrupted
after 1 h, and the products were isolated via preparative TLC.
Lactol (−)-67 (15%) and lactone (−)-65 (36%) were isolated
as the expected products that would occur if the reaction
progressed as desired. Also present, however, was aldehyde
ether (−)-69 (22%), possessing an internal ether linkage, along
with formate (−)-71 (21%). When pure lactol (−)-67 was
treated with TPAP/NMO, only lactone (−)-65 was observed
by LCMS; none of the other byproducts were observed. We
reason that byproducts (−)-69 and (−)-71 were likely the
result of oxidation of indole aldehyde 66 prior to lactol
formation, as shown in Scheme 11.51
A two-step protocol was therefore developed (Scheme 12)
that employed first a Corey−Palani52 oxidation to arrive at
lactol (−)-67, employing 2-iodoxybenzoic acid (IBX) in
tetrahydrofuran and dimethyl sulfoxide (1:1); TPAP/NMO
oxidation then led to desired lactone (−)-65 in 57% yield for
the two steps, along with lactol (−)-72 (11%). Lactol (−)-72,
the only observed side product, presumably results from
oxidation of the more hindered secondary hydroxyl group in
the first step.
A Plan Forward. The discovery that diol (+)-62 could be
obtained in good yield with the appropriate stereogenicity led
rapidly to a path forward, as outlined in retrospective fashion in
Figure 6. Indole (−)-52 would first be converted to diol (+)-62
as described above. Oxidative lactonization would then be
Completion of the Total Synthesis of (+)-Scholarisine
A (1). With ample quantities of lactone (−)-65 available,
treatment with aluminum chloride in toluene,53 employing
F
dx.doi.org/10.1021/ja3111626 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 11. Initial Oxidative Lactonization
Scheme 13. Total Synthesis of (+)-Scholarisine A (1)
Scheme 12. Two-Step Oxidative Lactonization
previous step, then furnished a sulfonamide indolenine
substrate. The trifluoroacetyl moiety, however, proved
preferable to sulfonamide protection of the nitrogen due to
superior lability during subsequent removal.
Removal of the trifluoroacetyl group in (+)-76 was then
readily achieved employing a mixture of saturated aqueous
K2CO3 and methanol (1:2) to furnish the corresponding free
amine. Oxidation with iodosobenzene (PhIO)56 in methylene
chloride completed the synthesis of (+)-scholarisine A (1);33
the yield for the final two steps was 77%. Totally synthetic
(+)-scholarisine A (1) was identical in all respects to the natural
product [i.e., 1H and 13C NMR (500 and 125 MHz,
respectively), HRMS parent ion identification, and chiroptic
properties].1 The structure of synthetic (+)-scholarisine A (1)
was also defined via X-ray analysis. Taken together, these
results confirmed the structure and absolute configuration of
(+)-scholarisine A (1). Approximately 70 mg of totally
synthetic (+)-scholarisine A (1) has been prepared to date.
“Retrobiosynthetic” Fragmentation of Scholarisine A
(1)A First Step To Access Other Members of the
Akuammiline Alkaloid Family. Having achieved the total
synthesis of (+)-scholarisine A (1), we turned our attention to
devising a substrate that would hold the promise of providing
access to other members of the akuammiline alkaloid family via
a late-stage “retro-biosynthetic” strategy. We focused on
utilizing synthetic (+)-scholarisine A (1) to access aldehyde
79 (Scheme 14), a carboxylate congener of aldehyde 2 (Figure
1).
sonication, selectively led to release of the oxygen and indole
nitrogen benzyl groups to furnish alcohol (−)-73 in 83% yield
(Scheme 13). Multiple attempts however at hydroxyl activation
in (−)-73 resulted only in scaffold degradation. Cyclization
involving nucleophilic attack by the tertiary alkyl amine, as
opposed to the indole ring, was suspected. Exchange of
nitrogen protection was therefore explored as a means to
remove the basicity of the aliphatic nitrogen.
The required protecting group exchange was achieved via
transfer hydrogenolysis first to render amine (+)-74;33
acylation employing trifluoroacetylimidazole54 in tetrahydrofur-
an then furnished trifluoroacetamide (−)-75. Pleasingly, single-
crystal X-ray analysis of (+)-74 confirmed the C(5)-stereo-
genicity, previously assigned on the basis of NMR NOE
correlations in lactone (−)-65.
Activation of the hydroxyl in (−)-75 with toluenesulfonyl
chloride, followed by treatment of the stable sulfonate ester
with tert-butyliminotri(pyrrolidino)phosphorane (BTPP),55
effected the strategically planned cyclization involving indole
deprotonation to furnish indolenine (+)-76 in a 64% yield over
the four steps. On the other hand, direct treatment of amine
(+)-74 with a variety of sulfonyl chlorides resulted only in
selective substitution at both the oxygen and secondary amine
to furnish the corresponding sulfonate ester sulfonamide.
Deprotonation of the indole nitrogen, which was inert in the
Toward this end, synthetic (+)-scholarisine A (1) was treated
overnight at 65 °C with aqueous sodium hydroxide (1 N) in
isopropanol (Scheme 14). The solution was then acidified with
aqueous hydrochloric acid (1 N), the solvent removed in vacuo,
and the residue purified via supercritical fluid chromatography
(SFC) to afford (+)-77 in approximately 50% yield. The
G
dx.doi.org/10.1021/ja3111626 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 14. “Retro-biosynthetic” Fragmentation
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental details, spectra, and X-ray crystallography. This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
smithab@sas.upenn.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors dedicate this article to Professor Carl Djerassi on
the occasion of his upcoming 90th birthday, friend, scholar, and
gentleman. Financial support was provided by the National
Institutes of Health (National Cancer Institute) through CA-
19033. We also thank Drs. George Furst and Rakesh Kohli for
help respectively obtaining the high-resolution NMR and mass
spectral data. Prof. William Dailey is also thanked for his
guidance while we performed the computational studies.
assigned structure was confirmed employing multiple 1D and
2D NMR experiments. Particularly diagnostic was the similarity
1
of the HNMR spectra of (+)-scholarisine A (1) and the
scaffold hydrogens in (+)-77, except for the differences in the
chemical shifts and coupling constants for the hydrogens on C-
5 and C-6, the epimerized carbon and the vicinal hydrogens,
respectively.
REFERENCES
■
(1) Cai, X. H.; Tan, Q. G.; Liu, Y. P.; Feng, T.; Du, Z. Z.; Li, W. Q.;
Luo, X. D. Org. Lett. 2008, 10, 577−580.
(2) (a) Smith, G. F. Chem. Ind. 1961, 1120. (b) Wenkert, E.;
Wickberg, B. J. Am. Chem. Soc. 1965, 87, 1580−1589.
The observed reaction pathway is presumed to proceed by
initial formation of alcohol 78 via hydrolytic opening of the
scholarisine A lactone bridge. The protonated form of 78 was
readily observed by LCMS upon treatment of scholarisine A
(1) with alcoholic aqueous sodium hydroxide at room
temperature. Acidification of the sample at room temperature
resulted in immediate reversion to scholarisine A (1), again
observed by LCMS. However upon heating, the hydroxy imine
carboxylate (78) likely undergoes an aza retroaldol fragmenta-
tion to provide enamine aldehyde 79 as a transient reaction
intermediate that can then undergo cyclization via enamine
attack at the aldehyde carbonyl to re-form 78, or alternatively to
form hydroxy imine 80, where the hydroxyl carbon has
undergone epimerization. Under the basic reaction conditions,
80 appears to be thermodynamically more stable than 78, as
complete conversion to 80 was observed to occur by LCMS.
Acidification at room temperature provided carboxylic acid
(+)-77. Formation of carboxylic acid (+)-77 demonstrates that
in fact a “retro-biosynthetic” fragmentation of this system is
possible. Exploration of this dynamic system, we believe, holds
the promise of access to other members of the akuammiline
alkaloid family.
Summary. The first total synthesis of (+)-scholarisine A (1)
has been achieved, employing a longest linear reaction
sequence of 25 steps from commercially available anhydride
35. Key synthetic tactics include a novel cyclization, comprising
nitrile reduction coupled with concomitant addition of the
resultant amine to an epoxide; a modified Fischer synthetic
protocol; an oxidative lactonization of a diol in the presence of
an indole ring; and a late-stage cyclization to complete the
caged ring scaffold of (+)-scholarisine A (1). A “retro-
biosynthetic” fragmentation of totally synthetic (+)-scholarisine
A (1) has also been achieved. Studies are currently underway to
exploit this fragmentation to gain access to related members of
the akuammiline family.
(3) For an overview of the akuammiline alkaloid family, see:
(a) Ramirez, A.; Garcia-Rubio, S. Curr. Med. Chem. 2003, 10, 1891−
1915. (b) Llopart, C. C.; Joule, J. A. Arkivoc 2004, x, 20−38. For a
review of the alkaloids found in Alstonia species, see: (c) Rahman, A.
U.; Qureshi, M. M.; Ali, S. S. J. Chem. Soc. Pak. 1990, 12, 355−406.
(4) Olivier, L.; Levy, J.; Le Men, J.; Janot, M. M.; Budzikiewicz, H.;
Djerassi, C. Bull. Soc. Chim. Fr. 1965, 868−876.
(5) Henry, T. A. J. Chem. Soc. 1932, 2759−2768.
(6) (a) Schnoes, H.; Biemann, K.; Mokry, J.; Kompis, I.; Chatterjee,
A.; Ganguli, G. J. Org. Chem. 1966, 31, 1641−1642. (b) Ahmad, Y.;
Fatima, K.; Rahman, A.; Occolowitz, J.; Solheim, B.; Clardy, J.;
Garnick, R.; Le Quesne, P. J. Am. Chem. Soc. 1977, 99, 1943−1946.
(7) (a) Joule, J. A.; Ohashi, M.; Gilbert, B.; Djerassi, C. Tetrahedron
1965, 21, 1717−1734. (b) Ohashi, M.; Joule, J. A.; Djerassi, C.
Tetrahedron Lett. 1964, 5, 3899−3905.
(8) Subramaniam, G.; Hiraku, O.; Hayashi, M.; Koyano, T.;
Komiyama, K.; Kam, T. S. J. Nat. Prod. 2007, 70, 1783−1789.
(9) Zu, L.; Boal, B. W.; Garg, N. K. J. Am. Chem. Soc. 2011, 133,
8877−8879.
(10) (a) Britten, A. Z.; Smith, G. F. J. Chem. Soc. 1963, 3850−3854.
(b) Britten, A.; Smith, G.; Spiteller, G. Chem. Ind. 1963, 1492−1493.
(11) (a) Chatterjee, A.; Mukherjee, B.; Ray, A. B.; Das, B.
Tetrahedron Lett. 1965, 6, 3633−3637. (b) Ghosh, R.;
Roychowdhury, P.; Chattopadhyay, D.; Iitaka, Y. Acta Crystallogr.,
Sect. C: Cryst. Struct. Commun. 1988, 44, 2151−2154.
(12) Lewin, G.; Poisson. J. Bull. Soc. Chim. Fr. 1980, II-400−II-404.
(13) Morita, Y.; Hesse, M.; Schmid, H.; Banerji, A.; Banerji, J.;
Chatterjee, A.; Oberhansli, W. E. Helv. Chim. Acta 1977, 60, 1419−
1434.
(14) Lim, K. H.; Kam, T. S. Tetrahedron Lett. 2006, 47, 8653−8655.
(15) (a) Mokry, J.; Dubravkova, L.; Sefcovic, P. Experientia 1962, 18,
564−565. (b) Das, B. C.; Cosson, J. P.; Lukacs, G.; Potier, P.
Tetrahedron Lett. 1974, 15, 4229−4302. (c) Mamatas-Kalamaras, S.;
Sevenet, T.; Thal, C.; Potier, P. Phytochemistry 1975, 14, 1637−1639.
(16) (a) Goodson, J. A.; Henry, T. A. J. Chem. Soc. 1925, 127, 1640−
1648. (b) Manohar, H.; Ramaseshan, S. Tetrahedron Lett. 1961, 2,
814−816.
(17) The echitamine core has been proposed to arise directly from
geissoschizine (5) (see ref 2a). This core has also been shown to
H
dx.doi.org/10.1021/ja3111626 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
derive from strictamine (6), via reduction with zinc in methanolic
sulfuric acid (see ref 6a).
(18) Zhang, M.; Huang, X.; Shen, L.; Qin, Y. J. Am. Chem. Soc. 2009,
synthetic (+)-scholarisine A (1) were confirmed by single-crystal X-ray
analysis. Crystallization of (+)-59 with chloroform led to co-
crystallization of chloroform in the unit cell, which in turn permitted
confirmation of the absolute configuration of (+)-59 by X-ray analysis,
exploiting the heavy-atom anomalous dispersion of chlorine. See the
Supporting Information for CIF files of (+)-44, (−)-50, (−)-51, and
(+)-56. See ref 23 for CIF files of (−)-33, (−)-34, (+)-41, (−)-42,
(−)-43, (+)-45, (+)-59, (+)-74, and totally synthetic (+)-scholarisine
A (1).
131, 6013−6020.
(19) Zi, W.; Xie, W.; Ma, D. J. Am. Chem. Soc. 2012, 134, 9126−
9129.
(20) Arai, H.; Hirasawa, Y.; Rahman, A.; Kusumawati, I.; Zaini, N. C.;
Sato, S.; Aoyama, C.; Takeo, J.; Morita, H. Bioorg. Med. Chem. 2010,
18, 2152−2158.
(21) Meng, W.; Ellsworth, B. A.; Nirschl, A. A.; McCann, P. J.; Patel,
M.; Girotra, R. N.; Wu, G.; Sher, P. M.; Morrison, E. P.; Biller, S. A.;
Zahler, R.; Deshpande, P. P.; Pullockaran, A.; Hagan, D. L.; Morgan,
N.; Taylor, J. R.; Obermeier, M. T.; Humphreys, W. G.; Khanna, A.;
Discenza, L.; Robertson, J. G.; Wang, A.; Han, S.; Wetterau, J. R.;
Janovitz, E. B.; Flint, O. P.; Whaley, J. M.; Washburn, W. N. J. Med.
Chem. 2008, 51, 1145−1149.
(22) (a) Kamarajan, P.; Sekar, N.; Mathuram, V.; Govindasamy, S.
Biochem. Int. 1991, 25, 491−498. (b) Jagetia, G. C.; Baliga, M. S.;
Venkatesh, P.; Ulloor, J. N.; Mantena, S. K.; Genebriera, J.; Mathuram,
V. J. Pharm. Pharmacol. 2005, 57, 1213−1219.
(23) Adams, G. L.; Carroll, P. J.; Smith, A. B., III J. Am. Chem. Soc.
2012, 134, 4037−4040.
(24) Garg recently reported the use of a similar lactone (see ref 9).
We initially considered lactone (27) while contemplating a Fischer
annulation approach inspired by the work of Levy and Dolby. See:
(a) Laronze, J. Y.; Laronze, J.; Patigny, D.; Levy, J. Terahedron Lett
1986, 27, 489−492. (b) Dolby, L. J.; Esfandiari, Z. J. Org. Chem. 1972,
37, 43−46. (c) Dolby, L. J.; Nelson, S. J. J. Org. Chem. 1973, 38,
2882−2887.
(25) For an overview on acid-promoted decomposition of
unsaturated α-diazo ketones, see: Smith, A. B., III; Dieter, R. K.
Tetrahedron 1981, 37, 2407−2439.
(26) Attempts to form the indolenine quaternary center of the
akuammiline family have proven difficult. See: (a) Bennasar, M. L.;
Zulaica, E.; Lopez, M.; Bosch, J. Tetrahedron Lett. 1988, 29, 2361−
2364. (b) Bosch, J.; Bennasar, M. L. Synlett 1995, 587−596.
(c) Bennasar, M. L.; Zulaica, E.; Ramirez, A.; Bosch, J. J. Org. Chem.
1996, 61, 1239−1251 Also see refs 24b and 24c..
(27) Danieli, B.; Lesma, G.; Mauro, M.; Palmisano, G.; Passarella, D.
J. Org. Chem. 1995, 60, 2506−2513. Also see: Danieli, B.; Lesma, G.;
Mauro, M.; Palmisano, G.; Passarella, D. Tetrahedron 1994, 50, 8837−
8852. Danieli, B.; Lesma, G.; Mauro, M.; Palmisano, G.; Passarella, D.
Tetrahedron: Asymmetry 1990, 1, 793−800.
(28) Ochiai, H.; Ohtani, T.; Ishida, A.; Kishikawa, K.; Yamamoto, S.;
Takeda, H.; Obata, T.; Nakai, H.; Toda, M. Eur. J. Med. Chem. 2004,
39, 555−571.
(29) For use of 4 Å sieves during lipase transesterification with
methyl ester solvent, see: Ottolina, G.; Carrea, G.; Riva, S. J. Org.
Chem. 1990, 55, 2366−2369. For an overview of enzymatic
asymmetrization of meso-compounds, see: Danieli, B.; Lesma, G.;
Passarella, D.; Riva, S. Chiral Synthons via Enzyme-Mediated
Asymmetrization of Meso-Compounds. In Advances in the Use of
Synthons in Organic Chemistry; Dondoni, A., Ed.; JAI Press Inc.:
Greenwich, CT, 1993; Vol. 1, pp 143−219.
(30) For use of a diamine to quench the reaction of tosyl chloride,
see: Morita, J.; Nakatsuji, H.; Misaki, T.; Tanabe, Y. Green Chem. 2005,
7, 711−715.
(31) Extreme caution should be employed when using potassium
cyanide, especially when in a DMSO solution. See: Potassium
Cyanide. e-EROS Encyclopedia of Reagents for Organic Synthesis
[Online]; Wiley & Sons, Posted April 15, 2001; http://onlinelibrary.
wiley.com/o/eros/articles/rp208/frame.html (accessed November 5,
2012).
(32) (a) Hughes, T. V.; Hammond, S. D.; Cava, M. P. J. Org. Chem.
1998, 63, 401−402. (b) Katritzky, A. R.; Akue-Gedu, R.; Vakulenko, A.
V. Arkivoc 2007, iii, 5−12. (c) Katritzky, A. R.; Brzezinski, J. Z.; Lam, J.
N. Rev. Roum. Chim. 1991, 36, 573−580.
(33) The structures of (−)-33, (−)-34, (+)-41, (−)-42, (−)-43,
(+)-44, (+)-45, (−)-50, (−)-51, (+)-56, (+)-59, (+)-74, and totally
(34) (a) Kasai, T.; Watanabe, H.; Mori, K. Bioorg. Med. Chem. 1993,
1, 67−70. (b) OBrien, P.; Tournayre, J. Tetrahedron 1997, 53, 17527−
17542.
(35) Rhodium on alumina has been shown to tolerate the oxirane
moiety. See: (a) Tarbell, D. S.; West, R. L.; McNally, J. G.;
McCorkindale, N. J.; Kunstmann, M.; Carman, R. M.; Cremer, S. E.;
Cross, A. D.; Huffman, K. R.; Varino, H. L.; Rosowsky, A.; Chapman,
D. D. J. Am. Chem. Soc. 1961, 83, 3096−3113. For reduction of
nitriles to primary amines using rhodium on alumina, see:
(b) Freifelder, M. J. Am. Chem. Soc. 1960, 82, 2386−2389. Bridged
azapolycyclic alcohols have been constructed via intramolecular
expoxide opening by an amide. See: (c) Schultz, R.; Staas, W.;
Spurlock, L. J. Org. Chem. 1973, 38, 3091−3093. Also see ref 24c.
(36) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155−4156.
(37) For use of acetic acid as a solvent for nitrile hydrogenation
employing rhodium on alumina, see: Galan, A.; De Mendoza, J.;
Prados, P.; Rojo, J.; Echavarren, A. M. J. Org. Chem. 1991, 56, 452−
454.
(38) Sajiki, H.; Hattori, K.; Hirota, K. Chem.Eur. J. 2000, 6, 2200−
2204.
(39) Benzoyl phenylhydrazine has been used in an attempt to avoid
pyrazole or pyrazolone formation during Fischer indole syntheses with
acyclic 1,3-dicarbonyl compounds. See: Mills, K.; Alkhawaja, K.;
Alsaleh, F.; Joule, J. A. J. Chem. Soc., Perkin Trans. 1 1981, 636−641.
(40) Welch, W. M. Synthesis 1977, 645−646.
(41) N-Aroylindoles have been isolated upon Fischer syntheses with
aroyl phenylhydrazines. See: Yamamoto, H. J. Org. Chem. 1967, 32,
3693−3695.
(42) For similar acyl migration, see: Kardos-Balogh, Z.; Soti, F.;
Incze, M.; Kajtar-Peredy, M.; Radics, L.; Szantay, C. Heterocycles 1989,
28, 303−312.
(43) Possible benzoyl migration involving formation and decom-
position of a 1,3-benzodiazapine has not been ruled out. See:
Plieninger, H.; Nogradi, I. Chem. Ber 1955, 88, 1965−1967. Chen, F.
M. F.; Forrest, T. P. Can. J. Chem. 1973, 51, 881−884.
(44) It is possible that the corresponding benzoyl indole forms and
then undergoes hydrolysis under the reaction conditions. Unprotected
indole products were observed by Joule and co-workers during their
Fischer indole studies with benzoyl phenylhydrazine (see ref 39).
(45) Perni, R. B.; Gribble, G. W. Org. Prep. Proced. Int. 1982, 14,
343−346.
(46) Hanessian, S.; Lavallee, P. Can. J. Chem. 1975, 53, 2975−2977.
(47) Ley, S. V.; Norman, J.; Griffith, W. P.; Marsden, S. P. Synthesis
1994, 639−666.
(48) For reports of similar oxidations, see: (a) Oikawa, Y.;
Yonemitsu, O. J. Org. Chem. 1977, 42, 1213−1216. (b) Pinchon, T.
M.; Nuzillard, J. M.; Richard, B.; Massiot, G.; Le Men-Olivier, L.;
Sevenet, T. Phytochemistry 1990, 29, 3341−3344. (c) Wang, T.; Xu,
Q.; Yu, P.; Liu, X.; Cook, J. M. Org. Lett. 2001, 3, 345−348. (d) Yu, J.;
Liao, X.; Cook, J. M. Org. Lett. 2002, 4, 4681−4684.
(49) Parikh, J. R.; Doering, W. E. J. Am. Chem. Soc. 1967, 89, 5505−
5507.
(50) Still, W. C. J. Am. Chem. Soc. 1978, 100, 1481−1487.
(51) N-Methylmorpholine N-oxide is known to oxidize activated
halides to the carbonyl oxidation state. See: Griffith, W. P.; Jolliffe, J.
M.; Ley, S. V.; Springhorn, K. F.; Tiffin, P. D. Synth. Commun. 1992,
22, 1967−1971.
(52) (a) Corey, E.; Palani, A. Tetrahedron Lett. 1995, 36, 7945−7948.
(b) Corey, E.; Palani, A. Tetrahedron Lett. 1995, 36, 3485−3488.
I
dx.doi.org/10.1021/ja3111626 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(53) (a) Watanabe, T.; Kobayashi, A.; Nishiura, M.; Takahashi, H.;
Usui, T.; Kamiyama, I.; Mochizuki, N.; Noritake, K.; Yokoyama, Y.;
Murakami, Y. Chem. Pharm. Bull. 1991, 39, 1152−1156. (b) Murakami,
Y.; Watanabe, T.; Kobayashi, A.; Yokoyama, Y. Synthesis 1984, 738−
740.
(54) Staab, H.; Walther, G. Chem. Ber. 1962, 95, 2070−2072.
(55) Schwesinger, R.; Willaredt, J.; Schlemper, H.; Keller, M.;
Schmitt, D.; Fritz, H. Chem. Ber. 1994, 127, 2435−2454.
(56) (a) Muller, P.; Gilabert, D. Tetrahedron 1988, 44, 7171−7175.
(b) Larsen, J.; Jorgensen, K. J. Chem. Soc., Perkin Trans. 2 1992, 1213−
1217.
J
dx.doi.org/10.1021/ja3111626 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX