Divergent total synthesis of the tricyclic marine alkaloids lepadiformine, fasicularin, and isomers of polycitorols by reagent-controlled diastereoselective reductive amination

Article abstract of DOI:10.1002/chem.201404316

We describe a flexible and divergent route to the pyrrolo-/pyrido[1,2-j]quinoline frameworks of tricyclic marine alkaloids via a common intermediate formed by the ester-enolate Claisen rearrangement of a cyclic amino acid allylic ester. We have synthesize

Full text of DOI:10.1002/chem.201404316

DOI: 10.1002/chem.201404316
Full Paper
&
Organic Chemistry
Divergent Total Synthesis of the Tricyclic Marine Alkaloids
Lepadiformine, Fasicularin, and Isomers of Polycitorols by
Reagent-Controlled Diastereoselective Reductive Amination
Jinkyung In, Seokwoo Lee, Yongseok Kwon, and Sanghee Kim*[a]
Abstract: We describe a flexible and divergent route to the
pyrrolo-/pyrido[1,2-j]quinoline frameworks of tricyclic marine
alkaloids via a common intermediate formed by the ester–
enolate Claisen rearrangement of a cyclic amino acid allylic
ester. We have synthesized the proposed structure of polyci-
torols and demonstrated that the structure of these alkaloids
requires revision. In addition to asymmetric formal synthe-
ses, stereoselective and concise total syntheses of (À)-lepadi-
formine and (À)-fasicularin were also accomplished from
simple, commercially available starting materials in a com-
pletely substrate-controlled manner. The key step in these
total syntheses was the reagent-dependent stereoselective
reductive amination of the common intermediate to yield
either indolizidines 55a or 55b. Aziridinium-mediated
carbon homologation of the hindered C-10 group to the
homoallylic group facilitated the synthesis.
Introduction
There has been considerable interest in the synthesis of tricy-
clic marine alkaloids with a perhydropyrrolo[1,2-j]quinoline or
perhydropyrido[2,1-j]quinoline ring system. The first members
of this structurally unique natural product family were the cy-
lindricines (1, Figure 1), which were isolated from the ascidian
Clavelina cylindrica in the early 1990s.[1] Interestingly, pyrrolo-
quinoline cylindricine A (1a) and pyridoquinoline cylindricine B
(1b) can interconvert, presumably via aziridinium intermediate
2. Another notable alkaloid of this general class is lepadifor-
mine (3a). This marine alkaloid, which is now known as lepadi-
formine A, was isolated from the tunicate Clavelina lepadiformis
in 1994 and exhibits strong cardiovascular effects as well as
moderate cytotoxicity.[2] The originally proposed structure for
lepadiformine A has been shown to be incorrect, and the cor-
rect relative and absolute stereochemistries were determined
by its total synthesis.[3] The revised structure of lepadiformine
includes a trans-1-azadecalin A/B ring system instead of the
cis-1-azadecalin framework of cylindricines. X-ray crystallo-
graphic analysis of the hydrochloride salt of synthetic 3a indi-
cated that its B-ring exists in an unusual twist-boat form, as
shown in Figure 1.[3a]
Figure 1. Structures of tricyclic marine alkaloids.
age at C-2. Lepadiformine C (3c) is another congener and lacks
the hydroxymethyl group at C-13.[4] In 1997, fasicularin (4) was
isolated from the ascidian Neptheis fasicularis.[5] This alkaloid
exhibits modest cytotoxic activity against Vero cells and selec-
tive activity against a DNA repair-deficient yeast.[5] The structur-
al features of fasicularin are similar to those of cylindricine B
(1b), but the A/B ring system is a trans-1-azadecalin similar to
that of lepadiformines. In addition, unlike most cylindricines,
this alkaloid lacks the carbonyl group at C-4. More recently,
two new tricyclic alkaloids were isolated from a marine ascidi-
an of the family Polycitoridae and named polycitorols A and B
by Tanaka and co-workers.[6] Based on spectral analysis, the
structures of the polycitorols were proposed to be those
Since the isolation of cylindricines and lepadiformine A,
a number of their congeners have been identified. A marine al-
kaloid that is very closely related to lepadiformine A is lepadi-
formine B (3b), which bears a butyl instead of a hexyl append-
[a] J. In, S. Lee, Y. Kwon, Prof. S. Kim
College of Pharmacy
Seoul National University
San 56-1, Shilim, Kwanak, Seoul, 151-742 (Korea)
E-mail: pennkim@snu.ac.kr
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201404316.
Chem. Eur. J. 2014, 20, 1 – 11
1
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
shown in 5a and 5b. The proposed structures of the polycitor-
ols are similar to those of the cylindricines, but they have
a butyl group rather than a hexyl group at C-2 and lack the
C-4 oxygenation found in most cylindricines.
by annulation of the C-10 carboxylic group with the C-5 propa-
nol moiety and piperidine B ring formation involving the nitro-
gen and C-5 vinyl group would lead to the formation of the
cis-fused A/B ring system of the polycitorols (5). With cyclic
amino acid 6b, a trans-fused A/B ring system would be ob-
tained by ring closure through annulation of the C-10 carbox-
ylic group with the C-5 propanol moiety and piperidine ring
formation involving the nitrogen and the C-5 vinyl group.
The ester–enolate Claisen rearrangement of amino acid allyl-
ic ester 7 was envisioned to deliver the densely functionalized
cyclic amino acids 6a or 6b. The stereochemical outcome of
the Claisen rearrangement of 7 can be determined from the
enolate geometry and transition-state geometry. The Claisen
rearrangement proceeds via a chairlike transition state.[10] How-
ever, at the beginning of our studies, there were no examples
in the literature describing the control of the ester–enolate ge-
ometry in cyclic a-amino acid ester compounds possessing an
exocyclic N-carbonyl group such as 7.[11] Therefore, we were
uncertain about the stereochemistry of the major product
formed from the rearrangement of 7. However, the stereo-
chemical outcome was not a concern if high selectivity was
achieved because both C-5 epimer 6a and 6b could be trans-
formed to either the cis- or trans-1-azadecalin A/B ring as
stated above.
Due to their bioactivities and structural features, these tricy-
clic marine alkaloids have been the subject of numerous syn-
thetic studies.[7,8] In this context, we have previously reported
a formal synthesis of lepadiformine A.[9] Herein, we describe
the full details of synthetic work that includes the substrate-
controlled total syntheses of (À)-lepadiformine A (3a) and
(À)-fasicularin (4) via a common intermediate as well as the
synthesis of the proposed structure of the polycitorols (5). One
of the key features of these total syntheses was the reagent-
dependent reversal of diastereofacial selectivity in the B-ring
formation of the tricyclic marine alkaloids. This key step as well
as the appropriate combination of functional groups in the
A/B ring formation enabled us to selectively synthesize the
above-mentioned alkaloids from the same intermediate.
General synthetic strategy
The main structural difference between tricyclic marine alka-
loids 3–5 was the stereochemistry of the A/B ring fusion. To
synthesize these alkaloids, we adopted a flexible approach that
could be amended to prepare both cis and trans-1-azadecalin
A/B ring systems. We envisioned that either cyclic amino acid
6a or its C-5 epimer 6b (Scheme 1) could potentially lead to
Results and Discussion
Synthesis of key intermediate 6
As per the general synthetic plan, our synthesis started with
the preparation of protected cyclic amino acid 8 as a synthetic
precursor for Claisen rearrangement substrate 7 (Scheme 2).
Cyclic amino acid 8 was envisioned to be derived from com-
mercially available (S)-methyl-N-Boc-pyroglutamate (9), which
was the only source of chirality in our synthesis.
Although the stereochemistry at the a-position of the car-
bonyl group of 8 would be destroyed upon deprotonation
during the ester–enolate Claisen rearrangement, it was pre-
pared as a single configuration product from N-Boc-pyrogluta-
mate (9). The DIBAL-H reduction of 9 was followed by in situ
treatment with acidic methanol, affording aminal 10. The hy-
droxyl group of 10 was protected as its benzyl ether to yield
known compound 11.[12] The nucleophilic addition of a vinyl
group to the N-acyliminium ion generated in situ from 11 was
achieved selectively by employing a Grignard-derived vinylcop-
per reagent to produce 12 as the only detectable isomer in
92% yield. The vinyl moiety of 12 was transformed to a carbox-
ylic acid group by ruthenium tetraoxide-mediated oxidative
cleavage to yield amino acid 8 in 83% yield.[13] Esterification of
acid 8 with known allylic alcohol 13[14] under Steglich’s DCC
coupling[15] conditions provided the desired allylic ester 7 in
high yield.
Scheme 1. General retrosynthetic analysis. Boc=tert-butoxycarbonyl.
TBDPS= tert-butyldiphenylsilyl.
both cis and trans-1-azadecalin A/B ring systems by the appro-
priate combination of functional groups. For example, a trans-
1-azadecalin A/B ring system of 3 and 4 could be constructed
from cyclic amino acid 6a through a ring closure involving the
formation of a bond between the C-5 vinyl group (natural
product numbering) and the C-10 carboxylic group along with
the formation of the piperidine B-ring by joining the nitrogen
function to the C-5 propanol moiety. Alternatively, ring closure
With multigram quantities of 7 in hand, we explored the
ester–enolate Claisen rearrangement. As mentioned earlier,
either rearrangement product 6a or 6b could be used for the
total synthesis if high selectivity was achieved in the reaction.
We surveyed a series of reaction conditions involving changes
&
&
Chem. Eur. J. 2014, 20, 1 – 11
www.chemeurj.org
2
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Due to the clear predominance of the chair conformation in
the transition state of the Claisen rearrangement, the outcome
of 6a and 6c indicated the exclusive formation of the (E)-
ketene acetal as an intermediate during the reaction. This se-
lective enolization might be due to the unfavorable steric in-
teraction between the bulky N-Boc group and the alkyl group
of the ester moiety in the transition state leading to the (Z)-
ketene acetal.[16] Another plausible explanation for this selectiv-
ity is chelation control between the (E)-ester enolate oxygen
atom and the heteroatom of the N-Boc group.[17]
The stereochemistry of rearrangement product 6a was es-
tablished by its conversion to the final natural product (vide
infra), and the stereochemistry of 6c was determined by NOE
experiments after derivatization to a rigid bicyclic com-
pound.[18] The major product 6a resulted from rearrangement
on the less-hindered face opposite the C-13 substituent. The
facial selectivity of this process was dependent on the nature
of the bases. LiHMDS yielded better results as the base than
LDA (Table 1, entry 1 vs. 3 and 6 vs. 9). The reactions conduct-
ed with LiHMDS occurred with higher selectivity than those
conducted with NaHMDS or KHMDS (entry 6 vs. entries 7 and
8). The reasons for this selectivity are not yet clear.
The best reaction conditions in terms of both yield and se-
lectivity were as follows. When substrate 7 was added to a solu-
tion of LiHMDS in THF at À788C and gradually warmed to
room temperature, a very high diastereofacial selectivity was
achieved in the rearrangement to afford 6a and 6c in a 10:1
ratio and 90% combined yield (Table 1, entry 1). Similarly high
yields and selectivity were obtained when the mixture of 7
and TBSCl was treated with LiHMDS in THF at À788C followed
by gradual warming to room temperature (entry 6).
Scheme 2. Synthesis of key intermediate 6a by an ester–enolate Claisen re-
arrangement: a) DIBAL-H, CH2Cl2, À788C to RT, then 2.0m HCl (aq.), MeOH,
08C, 70%; b) NaH, BnBr, DMF, 08C to RT, 85%; c) CuI, vinylmagnesium bro-
mide, BF3·OEt2, THF, À788C to RT, 92%; d) NaIO4, RuCl·H2O, CH2Cl2, CH3CN,
H2O, RT, 83%; e) DCC, DMAP, CH2Cl2, RT, 88%; DIBAL-H=diisobutylaluminum
hydride, DMF=dimethylformamide, DCC=N,N’-dicyclohexylcarbodiimide,
DMAP=4-(dimethylamino)pyridine.
Table 1. Reaction conditions for the rearrangement of 7.
Entry
Base
(3 equiv)
Additive
(1.5 equiv)
Solvent
Yield [%],[a]
6a/6c[b]
Formal synthesis of (À)-lepadiformine A with rearrangement
product 6a
1[c]
2[c]
3[c]
4[c]
5[c]
6[d]
7[d]
8[d]
9[d]
LiHMDS
KHMDS
LDA
LiHMDS
LDA
LiHMDS
NaHMDS
KHMDS
LDA
–
–
–
–
THF
THF
THF
THF/HMPA
THF/HMPA
THF
THF
THF
90, 10:1
80, 3:1
60, 3:1
60, 5:1
40, 3:1
88, 8:1
85, 4:1
70, 5:1
50, 2:1
The tricyclic amino nitrile 14 (Scheme 3) was a key advanced
intermediate in Weinreb’s total synthesis of lepadiformine A.[3b]
We envisioned that this intermediate could be readily synthe-
sized from the functionalized cyclic amino acid 6a, which is
the major product of the Claisen rearrangement of 7. As previ-
ously mentioned, ring closure by bonding between the C-5
vinyl group and the C-10 carboxylic group of 6a would yield
the desired trans-fused A/B ring system of the lepadiformines.
We planned to use a ring-closing metathesis (RCM) reaction[19]
for this ring closure.
–
TBSCl
TBSCl
TBSCl
TBSCl
THF
[a] Isolated yields. [b] Determined by 1H NMR spectroscopic analysis of
the crude reaction mixture. [c] Substrate was added to a solution of base
at À788C. Solvent (0.05m). [d] Base was added to a solution of substrate
and additive at À788C. Solvent (0.02m). LiHMDS=lithium bis(trimethylsi-
lyl)amide, KHMDS=potassium bis(trimethylsilyl)amide, LDA=lithium dii-
sopropylamide, HMPA=hexamethylphosphoramide, TBSCl=tert-butyldi-
methylsilyl chloride.
To prepare the metathesis precursor, we first focused on
conversion of the carboxylic acid group of 6a into a homoallyl
group in 15. To this end, carboxylic acid 6a was reduced with
LiAlH4 to the corresponding primary alcohol 16 (94%). Unfortu-
nately, all of our attempts to convert the primary alcohol
group in 16 to the homoallyl group in 15 were unsuccessful,
most likely due to a high degree of steric hindrance.
in base, additive, temperature, and solvent to achieve high ste-
reoselectivity by selective enolization. Some typical results are
summarized in Table 1.
The above failure forced us to devise an alternative RCM
substrate. We envisioned that two-carbon-atom-homologation
of the C-5 vinyl group of 6a and conversion of a C-10 carbonyl
function to a carbon–carbon double bond would provide the
appropriate RCM substrate, such as diene 17. The successful
Under all the reaction conditions tested, the major product
was rearrangement product 6a, and the minor one was 6c.
The other two possible isomers (6b and 6d) were not detect-
ed, even in the reactions with HMPA (Table 1, entries 4 and 5).
Chem. Eur. J. 2014, 20, 1 – 11
www.chemeurj.org
3
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
tionalized lepadiformine B-ring to complete the formal synthe-
sis. Deprotection of silyl ether 23 with TBAF followed by oxida-
tion of resultant alcohol 24 with Dess–Martin periodinane[23]
yielded aldehyde 25 in 95% overall yield. Treatment of 25 with
p-TsOH in aqueous acetone under reflux led to the removal of
the N-Boc protecting group and subsequent cyclization to the
unstable tricyclic enamine 26, which has been previously re-
ported by the Weinreb group.[3b] Under conditions analogous
to those used by Weinreb, enamine 26 was converted in situ
to the more stable a-amino nitrile 14 in 70% overall yield. Our
spectroscopic data for 14 were identical to those previously re-
ported. Therefore, we have accomplished the formal synthesis
of (À)-lepadiformine A and can confirm the relative stereo-
chemistry of 6a.
Formal synthesis of (À)-fasicularin
Next, we focused our attention on the synthesis of (À)-fasicu-
larin (4) from aldehyde 25. We envisioned that the reduction
of iminium species 27 (Scheme 4) from the stereoelectronically
Scheme 3. Formal synthesis of (À)-lepadiformine A (3a): a) LiAlH4, THF, 08C,
94%; b) MeI, K2CO3, acetone, reflux, 93%; c) i) 9-BBN, TlOEt, THF, 508C; ii) vin-
yl bromide, [PdCl2(dppf)], AsPh3, THF/DMF, 408C, 73%; d) LiAlH4, THF, 08C,
91%; e) (COCl)2, DMSO, Et3N, CH2Cl2, 08C; f) (CH3Ph3P)+IÀ, nBuLi, THF, 84%
(over two steps); g) 21, CH2Cl2, 408C, 98%; h) H2, Pd/C, Et3N, MeOH, RT, 95%;
i) TBAF, THF, RT; j) Dess–Martin periodinane, CH2Cl2, RT, 95% (over 2 steps);
k) p-TsOH, acetone/H2O, reflux; l) KCN, 1n HCl, acetone/H2O, RT, 70% (over
2 steps). 9-BBN=9-borabicyclo[3.3.1]nonane, TBAF=tetrabutylammonium
fluoride, Dess–Martin periodinane=1,1,1-tris(acetyloxy)-1,1-dihydro-1,2-ben-
ziodoxol-3-(1H)one, p-TsOH=p-toluenesulfonic acid.
Scheme 4. Formal synthesis of (À)-fasicularin: a) C6H13MgBr, THF, À788C,
92%; b) Dess–Martin periodinane, CH2Cl2, RT, 93%; c) TFA/CH2Cl2 (3:5), 08C
to RT; d) NaCNBH3, MeOH, À408C, 85% from 30 e) l-Selectride, THF, À788C,
80% from 30; TFA=trifluoroacetic acid, DEAD=diethyl azodicarboxylate, l-
Selectride=lithium tri-sec-butylborohydride.
route to revised RCM substrate 17 began with methylation of
the carboxylic acid moiety of Claisen rearrangement product
6a to afford ester 18. A one-pot tandem sequence involving
regioselective hydroboration of the terminal olefin in 18 with
9-BBN followed by coupling of the resulting borane under
Suzuki–Miyaura conditions ([PdCl2(dppf)], AsPh3) (dppf=1,1’-
bis(diphenylphosphino)ferrocene) with vinyl bromide yielded
19 in 73% overall yield.[20] Treatment of 19 with LiAlH4 effected
reduction of the ester to afford alcohol 20 in 91% yield. Swern
oxidation of 20 followed by a Wittig reaction of the resulting
hindered aldehyde successfully provided the desired RCM sub-
strate 17 in 84% two-step yield. The RCM of diene 17 was suc-
cessfully performed with second-generation Grubbs’ catalyst
21[21] in CH2Cl2 at 408C, yielding the desired azaspiro-cyclohex-
ene derivative 22 in excellent yield (98%). Hydrogenation of
the olefinic bond of 22 using H2 and 10% Pd/C in the presence
of Et3N as a catalyst poison afforded 23 without effecting hy-
drogenolysis of the O-benzyl protecting group.[22]
preferred face would yield the C-2 stereochemistry of fasicular-
in. The expected product 28 has been converted into pyrido-
quinoline alkaloid 4 by Kibayashi and co-workers.[3a,24]
To continue the synthesis of (À)-fasicularin (4) from aldehyde
25, the hexyl group was first introduced using hexylmagnesi-
um bromide to afford alcohol 29 as a 1:1 mixture of diastereo-
mers in 92% yield. Secondary alcohol 29 was oxidized with the
Dess–Martin periodinane to yield ketone 30 (93%). For annula-
tion of the B-ring by reductive amination, the N-Boc protecting
group of 30 was removed with TFA to afford iminium salt 27,
and the resultant salt was treated with NaCNBH3 in MeOH at
À408C to give 28 as the only isomer in high overall yield. Pref-
erence for the formation of 28 could be rationalized by Ste-
vens’ stereoelectronic principle,[25] The spectroscopic data for
compound 28 synthesized by this route were identical to
those previously reported.[24] The remaining steps to complete
the synthesis were accomplished by employing the same reac-
After achieving carbocyclic A-ring formation, we directed
our efforts toward the construction of the cyano-group-func-
&
&
Chem. Eur. J. 2014, 20, 1 – 11
www.chemeurj.org
4
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
tion conditions as those previously described in the litera-
ture.[24]
synthesis of lepadiformine and fasicularin, we planned to con-
struct the piperidine B ring prior to the A ring.
Because the C-2 epimer of 28 could be converted to lepadi-
formine A in a single step, attempts were made to invert the
diastereoselectivity to develop a new selective route to lepadi-
formine A. We envisioned that a bulky reducing agent would
invert the selectivity because the reduction of iminium ion 27
by a bulky reducing agent could occur through a twisted-boat
conformation (equatorial attack) to avoid the severe 1,3-diaxial
interactions between the bridge carbon atom (C-11) and the
incoming agent. However, after attempts with various hydride
reagents, the diastereofacial preference of the reductive amina-
tion of iminium salt 27 could not be reversed. When the reduc-
tion was performed with H2 and Pd/C in the presence of Et3N,
the diastereoselectivity was reduced to 3:1 but not inverted.
When iminium salt 27 was treated with the bulky l-Selectride,
the reduction product was not the product of reductive amina-
tion but secondary alcohol 31 (ca. 1:1 mixture).
To this end, we first masked the carboxylic acid moiety of
Claisen rearrangement product 6a, and the vinyl group was
converted to the corresponding aldehyde by ozonolysis. Resul-
tant aldehyde 32 was subjected to a Horner–Emmons olefina-
tion[26] with 33 to yield a,b-unsaturated carbonyl compound
34 in 70% yield. Catalytic hydrogenation of 34 over Pd/C re-
sulted in the reduction of the double bond and simultaneous
O-benzyl group cleavage to afford reductive amination sub-
strate 35 (85%).
For the intramolecular reductive amination, the N-Boc pro-
tecting group of 35 was removed with TFA, and the crude
compound was subjected to typical reductive amination condi-
tions. The reaction with NaCNBH3/AcOH in CH3CN at À788C re-
sulted in good diastereoselectivity to yield indolizidines 36a
and 36b in a ratio of 5:1. In other solvents, such as MeOH,
THF, and CH2Cl2, the selectivity decreased to approximately
3:1. The C-2 stereochemistry of the two diastereomers was de-
termined by NOESY analysis.[18] The preference for the forma-
tion of 36a can be rationalized by Stevens’ stereoelectronic
principle[25] and/or substituent-directing effects. Intermediate
iminium salt 37 would exist predominantly in the half-chair
conformation as shown in Scheme 5. The preferred axial attack
by the hydride reagent is validated by Stevens’ stereoelec-
tronic principle. In addition, in this case, the hydride donor
might be delivered to the stereoelectronically preferred face
by a nearby directing group, such as a hydroxyl group.
Synthesis of the proposed structure of polycitorols with re-
arrangement product 6a
The total synthesis of the proposed structure of polycitorols
has not been previously reported. The A/B ring fusion of poly-
citorols was proposed to be cis, similar to that of cylindricines.
As mentioned above, the cis-1-azadecalin A/B ring system of 5
would be accessible from the major rearrangement product
(6a, Scheme 5) by a ring-closing annulation of the carboxylic
group with the C-5 propanol moiety and piperidine B-ring for-
mation involving the nitrogen and vinyl group. The formation
of the B ring was expected to be produced by a reductive ami-
nation of the appropriate carbonyl group derived from the
vinyl group of 6a. Unlike the cascade sequence for the formal
Next, we focused our efforts on the construction of the A
ring with the C-10 carboxylic group and the C-5 propanol
moiety of 36a. First, the hydroxyl group of 36a was internally
protected as the lactone by treatment with NaH (Scheme 6).
Scheme 6. Synthesis of polycitorol A and B: a) NaH, DMF, 08C to RT, 90%;
b) TBAF, THF, 08C, 88%; c) Ph3P, I2, imidazole, CH2Cl2, 08C, 97%; d) tBuLi,
Et2O, À1008C, 60%; e) hydrazine hydrate, KOH, ethylene glycol, 1608C, 1 h,
then 2108C, 5 h, 65%; f) TFAA, THF, 1208C, sealed tube, 90% brsm;
TFAA=trifluoroacetic anhydride.
Scheme 5. Synthesis of reductive amination product 36: a) MeI, K2CO3, ace-
tone, reflux; b) O3, MeOH/CH2Cl2, À788C, then Ph3P, 0 8C, 81% (over two
steps); c) 33, NaH, THF, 08C to RT, 70%; d) H2, Pd/C, MeOH, RT, 85%;
e) i) TFA/CH2Cl2 (3:5), 08C to RT; ii) NaCNBH3, AcOH, CH3CN, À788C, 60%.
Chem. Eur. J. 2014, 20, 1 – 11
www.chemeurj.org
5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Deprotection of silyl ether 38 followed by Appel reaction of
the resultant alcohol 39 yielded iodide 40, which was subject-
ed to halogen–metal exchange to effect anionic cyclization.[27]
Treatment of 40 with tBuLi at very low temperature produced
the desired tricyclic product 41 in 60% yield. Finally, com-
pound 41 was converted to the proposed structure (i.e., poly-
citorol A (5a)) by Wolff–Kishner–Huang reduction[28] of its hin-
dered carbonyl group at C-9. The transformation of the ob-
tained compound 5a into the polycitorol B (5b) proposed
structure was realized using the protocol developed by Cossy
and Pardo.[29] When 5a was heated in a sealed tube with a cata-
lytic amount of TFAA in THF, ring-expanded product 5b was
formed presumably via aziridinium intermediate 42. In this re-
action, only one stereoisomer was observed as the final prod-
uct. The spectral data for 5a and 5b were in good agreement
with the assigned structures. The identity of synthetic com-
pound 5a was further confirmed by comparison of the
1H NMR spectra of its hydrochloride salt with the literature
spectra of the hydrochloride salt of 43[30] (Figure 2), which
differ only in the length of the alkyl chain (i.e., hexyl vs.
butyl).[18]
work toward elucidating the correct structure of the polycitor-
ols was temporarily suspended.
Revised synthetic approach to (À)-lepadiformine and (À)-fa-
sicularin
The ultimate goal of the second round of synthesis was to de-
velop a more concise and selective synthesis of both (À)-lepa-
diformine A (3a) and (À)-fasicularin (4) from the same inter-
mediate. In our preceding synthesis, the reductive amination
of tricyclic substrate 29 resulted in the predominant formation
of pyrroloquinoline 27b (Scheme 4). Based on these results,
we concluded that formation of the azaspirocyclic skeleton
prior to the annulation of the B-ring by reductive amination
does not appear to be an adequate synthetic route for control-
ling the C-2 stereochemistry of both epimers. Therefore, in the
new approach, the formation of the A-ring was postponed
until the B-ring of the tricyclic alkaloid was constructed by re-
ductive amination.
Model system for a unified route to both (À)-lepadiformine
(3a) and (À)-fasicularin (4)
Before starting the second round of total synthesis, an appro-
priate model system was developed to determine if the revised
sequence would provide adequate control of the stereochem-
istry at C-2. Therefore, we chose pyrrolidine derivatives
46a–c[18] (Table 2) as the reductive amination/cyclization pre-
cursors. The first model compound, 46a, has no substituent at
C-10 (lepadiformine numbering), while the second one, 46b,
has an ester substituent, similar to the real system. Compound
46c has a hydroxyl methyl group protected with a nonchelat-
ing bulky silyl group at the C-10 position.
Figure 2. Structures of compounds 5a’–a’’ and 43–45.
However, the spectroscopic data for synthetic compounds
5a and 5b did not match those reported for the natural prod-
ucts.[6] These observations implied that the structures of the
natural polycitorols had been assigned incorrectly. We envi-
sioned that the stereochemical identity of natural polycitorol A
could be deduced by comparison of its NMR spectral data
with those for the diastereomers of lepadiformine A because
the pattern of the NMR spectra of the butyl group is similar to
that of the hexyl group. Thus, we compared the NMR spectro-
scopic data for polycitorol A with those for the reported iso-
mers of lepadiformine A (3a). Of the eight possible diastereo-
mers of 3a, six have been synthesized and disclosed by several
groups as part of their efforts toward the structural revision of
the originally proposed lepadiformine A structure. Only two
diastereomers (i.e., 44 and 45 in Figure 2) have not been re-
ported. A thorough examination of the reported data of the
six isomers with those of polycitorol A revealed only the differ-
ences. This implied that the configuration of one of the two
previously unreported isomers is most likely that of natural
polycitorol A. Our speculated structures of natural polycitorol A
(i.e., 5a’ and 5a’’) might be assembled from Claisen rearrange-
ment product 6c by the appropriate combination of functional
groups at C-10 and C-5. However, compound 6c was the
minor product of the reaction. Unfortunately, we were unable
to reverse the diastereoselectivity. Therefore, our synthetic
For the intramolecular reductive amination, the N-Boc pro-
tecting groups of the model compounds were removed with
TFA to afford iminium salt 47, and the resulting salt was sub-
jected to various reduction conditions to yield indolizidines 48.
The results are summarized in Table 2. Regardless of the nature
of the reducing agent, compound 46a, which has no substitu-
ent at C-10, led to the exclusive formation of cis-48a, in which
the C-10 and C-2 hydrogen atoms have a cis arrangement
(Table 2, entries 1–5). However, when the reduction was per-
formed with 46b, which has a substituent at C-10, reagent-
dependent stereoselectivity was observed. For example, upon
addition of NaCNBH3 at room temperature, indolizidine cis-48b
was obtained as the major reductive amination product
(entry 6). Higher selectivity was obtained at lower tempera-
tures (entries 7 and 8). Reduction in THF yielded a higher selec-
tivity than that in MeOH (entries 8 vs. 9). When NaBH4 and Na-
(OAc)3BH were employed as the reducing agent, a high diaster-
eofacial selectivity for cis-48b was also observed (entries 10
and 11). However, when iminium salt 47b was treated with the
bulky reducing agent l-Selectride, the diastereofacial selectivity
was completely reversed (entry 12). Under these reduction
conditions, indolizidine trans-48b was the major isomer and
was obtained in a ratio of 1:11 with a combined yield of 75%.
The reductive amination under catalytic hydrogenation condi-
&
&
Chem. Eur. J. 2014, 20, 1 – 11
www.chemeurj.org
6
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Table 2. Model systems for reductive amination.
Entry
Substrate
Reducing agent
Solvent
T [8C]
Yield [%],[a]
cis/trans
1
2
3
4
5
6
7
8
46a
46a
46a
46a
46a
46b
46b
46b
46b
46b
46b
46b
46b
46b
46c
46c
46c
46c
46c
46c
NaCNBH3
NaBH4
Na(OAc)3BH
l-Selectride
Pd/C
NaCNBH3
NaCNBH3
NaCNBH3
NaCNBH3
NaBH4
Na(OAc)3BH
l-Selectride
Pd/C
LiEt3BH
NaCNBH3
NaBH4
Na(OAc)3BH
l-Selectride
LiEt3BH
MeOH
MeOH
MeOH
THF
MeOH
MeOH
MeOH
MeOH
THF
MeOH
MeOH
THF
MeOH
THF
MeOH
MeOH
MeOH
THF
THF
MeOH
À40
À40
À40
À78
RT
78, 1:0
85, 1:0
85, 1:0
80, 1:0
90, 1:0
83, 5:1
80, 7:1
92, 20:1
76%, 74:1
80, 9.4:1
85, 7.3:1
75, 1:11
73, 1:13
95, 1:2.4
89, 2.8:1
91, 5:1
Figure 3. Two half-chair conformations of iminium salt 47 and reaction path-
ways to 48.
RT
0
À40
À40
À40
À40
À78
RT
À78
À40
À40
À40
À78
À78
RT
(Figure 3, path b) through a twisted-boat conformation, which
would lead to the formation of trans-48, might be more steri-
cally favored in the case of a bulky hydride donor and C-10
substituted substrate.
9
10
11
12
13
14
15
16
17
18
19
20
Another possible mechanistic path via the less stable confor-
mer 47’B cannot be completely ruled out. If the reaction rate
through conformer 47A is considerably slower than the inter-
conversion rate of the two conformers, the reduction of the
iminium ion by the bulky reducing agent would occur from
the less stable conformer 47’B, as stated by the Curtin–Ham-
mett principle.[32] Therefore, it is not unreasonable to postulate
that reduction by l-Selectride can also occur from the stereo-
electronically preferred face of conformer 47’B (Figure 3,
path c) to yield indolizidine trans-48.
89, 2.4:1
80, 1:7.1
90, 1:2
Pd/C
93, 1:22
[a] Yields and ratio were determined by GC analysis.
tions also yielded trans-48b as a major isomer with high selec-
tivity (entry 13), while the reaction with LiEt3BH afforded low
diastereoselectivity (entry 14). Substrate 46c also exhibited re-
agent-dependent stereoselectivity (entries 15–20), and the dia-
stereofacial selectivity pattern was the same as that observed
for 46b. However, the level of selectivity was generally lower,
most likely due to the difference in the nature of the C-10 sub-
stituents, with the exception of the reaction involving catalytic
hydrogenation (entry 20). This hydride reagent-dependent re-
versal of diastereofacial selectivity in the formation of indolizi-
dine and similar ring systems has not been previously reported
in the literature. This complete reversal of stereochemistry in
the reductive amination reaction could be explained by both
steric and stereoelectronic factors. Iminium salt 47 could exist
in one of two half-chair conformations, 47A and 47’B, as
shown in Figure 3. Our theoretical calculations predicted that
47A was more stable than 47’B regardless of the substituent
at C-10.[31] Therefore, we expected that iminium salt 47 would
exist predominantly in conformation 47A at reaction tempera-
ture. According to the stereoelectronic principles delineated by
Stevens,[25] the addition of the hydride reagent to the iminium
moiety should occur from the stereoelectronically preferred
face of 47A (Figure 3, path a) via a chairlike transition state to
yield cis-48. However, in the case of a bulky reducing agent,
“axial attack” (Figure 3, path a) would be disfavored due to
severe steric interactions between the incoming bulky reagent
and the C-10 substituent group. Therefore, “equatorial attack”
Synthesis of the B/C ring system of (À)-lepadiformine and
(À)-fasicularin
With the opportunity to access both C-2 epimers, we initiated
the second round of total synthesis. In our new synthesis, in-
stead of utilizing Claisen rearrangement product 6a, a more
elaborate rearrangement product 49 (Scheme 7) was devised
for the concise synthesis. Therefore, our synthesis began with
the preparation of allylic alcohol 50 for the new Claisen rear-
rangement substrate 51. For the synthesis of 51, we utilized
known ketone 52,[33] which was readily synthesized in three
steps from commercially available g-decanolactone. The
ketone group of 52 was protected as the ethylene ketal, and
the ester group was then reduced with DIBAL-H to provide al-
lylic alcohol 50 in 77% overall yield.
Esterification of acid 8 with allylic alcohol 50 provided allylic
ester 51 (88%). The ester–enolate Claisen rearrangement was
accomplished using the previously described reaction condi-
tions. Treatment of 51 with LiHMDS in THF at À788C and grad-
ual warming to room temperature afforded, after esterification
with TMSCHN2, the desired rearrangement product 53 and
a minor amount of its stereoisomer in 88% combined yield
and 10:1 diastereoselectivity.
For the intramolecular reductive amination, the N-Boc and
ketal protecting groups of 53 were simultaneously removed
with TFA to yield iminium salt 54. In light of the model system
Chem. Eur. J. 2014, 20, 1 – 11
www.chemeurj.org
7
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Scheme 8. Completion of the synthesis of (À)-fasicularin: a) LiAlH4, THF, À78
to 08C, 96%; b) MsCl, Et3N, Et2O, 08C; c) allylmagnesium bromide, CuI, THF,
À788C to RT, 60% from 56b; d) Grubbs II catalyst 21, CH2Cl2, reflux, 90%;
e) H2, Pd/C, MeOH, RT, 92%; f) NH4SCN, DEAD, Ph3P, CH2Cl2, RT, 88%.
MsCl=methanesulfonyl chloride.
rived from allylmagnesium bromide and CuI. The allyl cuprate
regioselectively attacked the reactive aziridinium ion at the
methylene carbon atom to form desired RCM substrate 57b in
60% yield over two steps. Although aziridinium ions are well-
established intermediates,[35] their reactions with a carbon nu-
cleophile are much less general and have been less widely ap-
plied in synthesis.[36] Therefore, we believe that this successful
aziridinium-mediated carbon homologation at a sterically hin-
dered position will increase the utility of the aziridinium ion in
the synthesis of complex nitrogen-containing compounds.
The RCM of diene 57b was successfully performed, furnish-
ing the desired tricyclic compound 59b in high yield (90%).
Catalytic hydrogenation of 59b over Pd/C resulted in the re-
duction of the olefinic bond and removal of the benzyl pro-
tecting group to yield tricyclic compound 60 in 92% yield. The
spectroscopic data for compound 60 synthesized by this route
were identical to those previously reported.[24] Completion of
the synthesis was accomplished by employing the same reac-
tion conditions as those previously described in the litera-
ture.[24] The spectral and optical rotation data for synthetic
(À)-fasicularin (4) were in good agreement with the reported
values.[5,24]
Scheme 7. Synthesis of indolizidines 55a and 55b by using the ester–eno-
late Claisen rearrangement and reductive amination: a) ethylene glycol,
PPTS, toluene, reflux; b) DIBAL-H, THF, À788C, 77% (over two steps);
c) EDCI, DMAP, CH2Cl2, RT, 88%; d) LiHMDS, TBSCl, THF, À788C to RT;
e) TMSCHN2, MeOH, RT, 88% from 51; f) TFA/CH2Cl2 (3:5), 08C to RT; g) con-
ditions A: NaCNBH3, MeOH, À788C, 87% from 53, 55a/55b=12:1; condi-
tions B: l-Selectride, THF, À788C, 91% from 53, 55a/55b=1:24. PPTS=pyri-
dinium p-toluenesulfonate, EDCI=N-(3-dimethylaminopropyl)-N’-ethylcarbo-
diimide hydrochloride, TMSCHN2 =(trimethylsilyl)diazomethane.
study, the resultant salt was treated with hydride reagents.
When the reduction was performed with NaCNBH3 at À788C,
reductive amination product 55a was obtained along with
a minor amount of its stereoisomer 55b in 87% combined
yield and high diastereoselectivity (12:1).[34] When iminium salt
54 was treated with l-Selectride at À788C, the diastereofacial
selectivity was completely inverted, which afforded indolizidine
55b as the major isomer in a ratio of 1:24 and 91% combined
yield.
With the synthesis of (À)-fasicularin (4) accomplished, we
next turned to the synthesis of (À)-lepadiformine A (3a) from
indolizidine 55a (Scheme 9). To construct the A-ring, we used
the same methodology as for fasicularin. Therefore, the ester
of 55a was reduced to alcohol 56a (97%). Again, aziridinium
ion-mediated allylation provided the corresponding RCM sub-
strate, 57a, in 65% overall yield.[37] The RCM of diene 57a fol-
lowed by hydrogenation yielded (À)-lepadiformine A (80%),
the spectral data of which (1H and 13C) were in complete
agreement with those reported in the literature.[2a,24] The opti-
cal rotation of synthetic lepadiformine was also in agreement
with that found in the literature.
Total synthesis of (À)-lepadiformine and (À)-fasicularin
After accomplishing the stereoselective formation of diastereo-
isomers 55a and 55b, we focused our efforts on the construc-
tion of the A-ring through an RCM reaction to complete the
synthesis. For the total synthesis of (À)-fasicularin (4), the ester
of 55b was reduced with LiAlH4 to the corresponding primary
alcohol, 56b, in 96% yield (Scheme 8). Despite considerable
steric hindrance, we successfully converted the hydroxymethyl
group of 56b to the homoallyl group in 57b with the assis-
tance of the tertiary amine within the molecule. Treatment of
56b with MsCl in the presence of Et3N resulted in the unstable
aziridinium salt 58b, which was treated with allyl cuprate de-
&
&
Chem. Eur. J. 2014, 20, 1 – 11
www.chemeurj.org
8
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Grimaud, J. F. Biard, M. P. Sauviat, M. Nabil, J. F. Verbist, J. Y. Petit, Toxicon
2001, 39, 1231–1237.
[3] a) H. Abe, S. Aoyagi, C. Kibayashi, J. Am. Chem. Soc. 2000, 122, 4583–
4592; b) P. Sun, C. Sun, S. M. Weinreb, J. Org. Chem. 2002, 67, 4337–
4345.
[4] M.-P. Sauviat, J. Vercauteren, N. Grimaud, M. Jugꢁ, M. Nabil, J.-Y. Petit, J.-
F. Biard, J. Nat. Prod. 2006, 69, 558–562.
[5] A. D. Patil, A. J. Freyer, R. Reichwein, B. Carte, L. B. Killmer, L. Faucette,
R. K. Johnson, D. J. Faulkner, Tetrahedron Lett. 1997, 38, 363–364.
[6] H. H. Issa, J. Tanaka, R. Rachmat, A. Setiawan, A. Trianto, T. Higa, Mar.
Drugs 1995, 3, 78–83.
[7] For reviews, see: a) S. M. Weinreb, Chem. Rev. 2006, 106, 2531–2549;
b) P. Schꢂr, S. Cren, P. Renaud, Chimia 2006, 60, 131–141; c) C. Kibaya-
shi, Chem. Pharm. Bull. 2005, 53, 1375–1386; d) C. Kibayashi, S. Aoyagi,
H. Abe, Bull. Chem. Soc. Jpn. 2003, 76, 2059–2074; e) S. M. Weinreb,
Acc. Chem. Res. 2003, 36, 59–65.
Scheme 9. Completion of the synthesis of (À)-lepadiformine: a) LiAlH4, THF,
À78 to 08C, 97%; b) MsCl, Et3N, Et2O, 08C; c) allylmagnesium bromide, CuI,
THF, À788C to RT, 65% from 56a; d) Grubbs II catalyst 21, CH2Cl2, reflux,
90%; e) H2, Pd/C, MeOH, RT, 89%.
[8] For selected recent synthesis, see: a) D. M. Dalton, T. Rovis, Org. Lett.
2013, 15, 2346–2349; b) M. A. Perry, M. D. Morin, B. W. Slafer, S. D. Rych-
novsky, J. Org. Chem. 2012, 77, 3390–3400; c) K. Weidner, A. Giroult, P.
Panchaud, P. Renaud, J. Am. Chem. Soc. 2010, 132, 17511–17515; d) H.
Liang, M. A. Ciufolini, Chem. Eur. J. 2010, 16, 13262–13270; e) M. Fujita-
ni, M. Tsuchiya, K. Okano, K. Takasu, M. Ihara, H. Tokuyama, Synlett 2010,
822–826; f) S.-L. Mei, G. Zhao, Eur. J. Org. Chem. 2010, 1660–1668;
g) A. M. Meyer, C. E. Katz, S.-W. Li, D. Vander Velde, J. Aube, Org. Lett.
2010, 12, 1244–1247; h) P. Vital, M. Hosseini, M. S. Shanmugham, C. H.
Gotfredsen, P. Harris, D. Tanner, Chem. Commun. 2009, 1888–1890; i) B.
Lygo, E. H. M. Kirton, C. Lumley, Org. Biomol. Chem. 2008, 6, 3085–
3090; j) X. M. Zhang, M. Wang, Y. Q. Tu, C. A. Fan, Y. J. Jiang, S. Y. Zhang,
F. M. Zhang, Synlett 2008, 2831–2835; k) A. C. Flick, M. J. A. Caballero, A.
Padwa, Org. Lett. 2008, 10, 1871–1874; l) J. J. Caldwell, D. Craig, Angew.
Chem. Int. Ed. 2007, 46, 2631–2634; Angew. Chem. 2007, 119, 2685–
2688; m) T. Shibuguchi, H. Mihara, A. Kuramochi, S. Sakubara, T. Ohshi-
ma, M. Shibasaki, Angew. Chem. Int. Ed. 2006, 45, 4635–4637; Angew.
Chem. 2006, 118, 4751–4753.
Conclusion
The planning and implementation of flexible, divergent total
syntheses of pyrrolo-/pyrido[1,2-j]quinoline tricyclic marine al-
kaloids is presented. A functionally and stereochemically en-
riched intermediate was concisely and stereoselectively assem-
bled by employing an ester–enolate Claisen rearrangement re-
action of the cyclic amino acid allylic ester possessing an exo-
cyclic N-carbonyl group. This common intermediate was readily
converted to (À)-lepadiformine A, (À)-fasicularin, and the pro-
posed structure of polycitorols A and B in a substrate-con-
trolled manner. In these studies, we have demonstrated that
the structure of polycitorols requires revision. With the more
elaborate ester–enolate Claisen rearrangement product, the
more concise and selective synthesis of both lepadiformine A
and fasicularin was also achieved in a substrate- and reagent-
controlled manner. One of the key features employed in these
total syntheses was a hydride reagent-dependent reversal of
diastereofacial selectivity in the formation of the indolizidine,
which we believe is the first example of this occurrence. In ad-
dition, aziridinium-mediated carbon homologation of the hin-
dered C-10 group into a homoallylic group greatly facilitated
the synthesis.
[9] M. Lee, T. Lee, E. Y. Kim, H. Ko, D. Kim, S. Kim, Org. Lett. 2006, 8, 745–
748.
[10] For reviews and articles on ester–enolate Claisen rearrangements, see:
a) R. E. Ireland, R. H. Mueller, J. Am. Chem. Soc. 1972, 94, 5897–5898;
b) R. E. Ireland, R. H. Mueller, A. K. Willard, J. Am. Chem. Soc. 1976, 98,
2868–2877; c) A. M. Martꢃn Castro, Chem. Rev. 2004, 104, 2939–3002;
d) Y. Chai, S. Hong, H. A. Lindsay, C. McFarland, M. C. McIntosh, Tetrahe-
dron 2002, 58, 2905–2928.
[11] a) K. Sakaguchi, M. Yamamoto, Y. Watanabe, Y. Ohfune, Tetrahedron Lett.
2007, 48, 4821–4824; b) K. Sakaguchi, M. Ayabe, Y. Watanabe, T. Okada,
K. Kawamura, T. Shinada, Y. Ohfune, Tetrahedron 2009, 65, 10355–
10364.
[12] T. Katoh, Y. Nagata, Y. Kobayashi, K. Arai, J. Minami, S. Terashima, Tetra-
hedron 1994, 50, 6221–6238.
[13] Alternatively, amino acid 8 could be obtained as a diastereoisomeric
mixture (2:1) from 9 through Terashima’s five-step sequence and Pin-
nick oxidation of the reported aldehyde intermediate; see references [9]
and [12].
Acknowledgements
[14] a) D. Kim, Y. M. Jang, I. O. Kim, S. W. Park, J. Chem. Soc. Chem. Commun.
1988, 760–761; b) J. A. Murphy, F. Rasheed, S. J. Roome, K. A. Scott, N.
Lewis, J. Chem. Soc. Perkin Trans. 1 1998, 2331–2340.
[15] B. Neises, W. Steglich, Angew. Chem. Int. Ed. Engl. 1978, 17, 522–524;
Angew. Chem. 1978, 90, 556–557.
[16] a) M. B. Smith, in Organic Synthesis 3rd ed., McGraw-Hill, New York,
2010, pp. 790–792; b) D. A. Evans, in Asymmertic Synthesis Vol. 3 (Ed:
J. D. Morrison), Academic Press, New York, 1984, p. 19; c) A. I. Meyers,
E. S. Synder, J. J. H. Akerman, J. Am. Chem. Soc. 1978, 100, 8186–8189.
[17] R. E. Ireland, P. Wipf, J. D. Armstrong III, J. Org. Chem. 1991, 56, 650–
657.
This work was supported by the Mid-Career Researcher Pro-
gram (No. 2013R1A2A1A01015998) of the National Research
Foundation of Korea (NRF) grant funded by the Korean gov-
ernment (MSIP). We are grateful to Prof. J. Tanaka (University of
the Ryukyus) for providing copies of 1H NMR spectroscopic
data for proposed polycitorol A (5a).
[18] For details, see the Supporting Information.
Keywords: aziridinium
amination · substrate and reagent control · total synthesis
·
diastereoselectivity
·
reductive
[19] For reviews, see: a) R. H. Grubbs, S. J. Miller, G. C. Fu, Acc. Chem. Res.
1995, 28, 446–452; b) J. B. Brenneman, S. F. Martin, Curr. Org. Chem.
2005, 9, 1535–1549; c) J. C. Conrad, D. E. Fogg, Curr. Org. Chem. 2006,
10, 185–202.
[20] S. R. Chemler, S. J. Danishefsky, Org. Lett. 2000, 2, 2695–2698 and refer-
ences therein.
[1] a) A. J. Blackman, C. Li, Tetrahedron 1993, 49, 8645–8656; b) C. Li, A. J.
Blackman, Aust. J. Chem. 1994, 47, 1355–1361.
[2] a) J. F. Biard, S. Guyot, C. Roussakis, J. F. Verbist, J. Vercauteren, J. F.
Weber, K. Boukef, Tetrahedron Lett. 1994, 35, 2691–2694; b) M. Jugꢁ, N.
[21] M. Scholl, S. Ding, C. W. Lee, R. H. Grubbs, Org. Lett. 1999, 1, 953–956.
Chem. Eur. J. 2014, 20, 1 – 11
www.chemeurj.org
9
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
[22] For examples of the use of amines as catalyst poisons, see: a) B. P.
Czech, R. A. Bartsch, J. Org. Chem. 1984, 49, 4076–4078; b) H. Sajiki, K.
Hirota, Tetrahedron 1998, 54, 13981–13996; c) I. Kadota, H. Takamura, K.
Sato, A. Ohno, M. Masayuki, Y. Yamamoto, J. Am. Chem. Soc. 2003, 125,
11893–11899.
a) J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865–
3868; b) B. Delley, J. Chem. Phys. 2000, 113, 7756–7764.
[32] a) J. I. Seeman, Chem. Rev. 1983, 83, 83–134; b) E. L. Eliel, S. H. Wilen, in
Stereochemistry of Organic Compounds, John Wiley and Sons, New York,
1994, pp. 648–655.
[23] a) D. B. Dess, J. C. Martin, J. Org. Chem. 1983, 48, 4155–4156; b) D. B.
Dess, J. C. Martin, J. Am. Chem. Soc. 1991, 113, 7277–7287.
[24] H. Abe, S. Aoyagi, C. Kibayashi, J. Am. Chem. Soc. 2005, 127, 1473–
1480.
[25] a) R. V. Stevens, Acc. Chem. Res. 1984, 17, 289–296; b) P. Deslong-
champs, in Stereoelectronic Effect in Organic Chemistry, Pergamon, New
York, 1983, pp. 209–221.
[26] a) W. S. Wadsworth, W. D. Emmons Jr, J. Am. Chem. Soc. 1961, 83, 1733–
1738; b) For review, see: B. E. Maryanoff, A. B. Reitz, Chem. Rev. 1989,
89, 863–927.
[33] a) K. Nagasawa, N. Hori, H. Koshino, T. Nakata, Heterocycles 1999, 50,
919–928; b) T. Kitahara, K. Mori, M. Matsui, M. Iwamoto, Y. Takagi, Y.
Warita, Agric. Biol. Chem. 1982, 46, 1369–1375.
[34] The stereochemistry of these two isomers was assigned by NOE differ-
ence experiments of the corresponding alcohols 56a and 56b; see the
Supporting Information.
[35] a) W. H. Pearson, B. W. Lian, S. C. Bergmeier, K. M. Lokantha Rai, A. Hass-
ner, in Comprehensive Heterocyclic Chemistry II, Vol. 1 A (Ed.: A. Padwa),
Pergamon, Oxford, 1996, pp. 1–96; b) A. K. Yudin, in Aziridines and Ep-
oxides in Organic Synthesis Wiley-VCH, Weinheim, 2006.
[27] a) M. P. Cooke Jr, I. N. Houpis, Tetrahedron Lett. 1985, 26, 4987–4990;
b) D. Kim, S. Kim, J. J. Lee, H. S. Kim, Tetrahedron Lett. 1990, 31, 4027–
4028; c) K. Ohtsuki, K. Matsuo, T. Yoshikawa, C. Moriya, K. Tomita-
Yokotani, K. Shishido, M. Shindo, Org. Lett. 2008, 10, 1247–1250.
[28] a) M. Huang-Minlon, J. Am. Chem. Soc. 1946, 68, 2487–2488; b) M.
Huang-Minlon, J. Am. Chem. Soc. 1949, 71, 3301–3303; c) A. R. Matlin,
C. F. George, S. Wolff, W. C. Agosta, J. Am. Chem. Soc. 1986, 108, 3385–
3394.
[29] a) J. Cossy, C. Dumas, D. G. Pardo, Eur. J. Org. Chem. 1999, 1693–1699;
b) T.-X. Mꢁtro, D. G. Pardo, J. Cossy, J. Org. Chem. 2007, 72, 6556–6561.
[30] W. H. Pearson, Y. Ren, J. Org. Chem. 1999, 64, 688–689.
[31] The energy-minimized conformations (Materials Studio 6.1, Systematic
grid scan) were further optimized by the density functional theory
method (PBE-GGA) at the DNP level (Materials Studio 6.1, DMol3), see:
[36] a) A. Gayet, P. G. Andersson, Adv. Synth. Catal. 2005, 347, 1242–1246;
b) D. R. Andrews, V. H. Dahanukar, J. M. Eckert, D. Gala, B. S. Lucas, D. P.
Schumacher, I. A. Zavialov, Tetrahedron Lett. 2002, 43, 6121–6125;
c) M. A. Graham, A. H. Wadsworth, M. Thornton-Pett, C. M. Rayner,
Chem. Commun. 2001, 966–967; d) D. Tanner, P. Somfai, Tetrahedron
1986, 42, 5657–5664; e) D.-H. Yoon, P. Kang, W. K. Lee, Y. Kim, H.-J. Ha,
Org. Lett. 2012, 14, 429–431.
[37] In this process, unlike the aziridinium salt 58b, the aziridinium inter-
mediate was rather stable.
Received: July 9, 2014
Published online on && &&, 0000
&
&
Chem. Eur. J. 2014, 20, 1 – 11
www.chemeurj.org
10
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
FULL PAPER
From one to four: The title marine nat-
ural products were synthesized in a com-
pletely substrate-controlled manner via
a common intermediate derived from
the amino acid ester–enolate Claisen re-
arrangement. The key step involves a re-
agent-dependent stereoselective reduc-
tive amination (see scheme).
&
Organic Chemistry
J. In, S. Lee, Y. Kwon, S. Kim*
&& – &&
Divergent Total Synthesis of the
Tricyclic Marine Alkaloids
Lepadiformine, Fasicularin, and
Isomers of Polycitorols by Reagent-
Controlled Diastereoselective
Reductive Amination
Chem. Eur. J. 2014, 20, 1 – 11
www.chemeurj.org
11
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ