Base-catalyzed intramolecular hydroamination of cyclohexa-2,5-dienes: Insights into the mechanism through DFT calculations and application to the total synthesis of epi-elwesine

Article abstract of DOI:10.1002/chem.201403662

The base-catalyzed intramolecular hydroamination of 1-ethylaminocyclohexa-2,5-dienes is described. The transformation proceeds through isomerization of the cyclohexa- 1,4-dienyl fragment into the corresponding conjugated 1,3- diene prior to the hydroamina

Full text of DOI:10.1002/chem.201403662

DOI: 10.1002/chem.201403662
Full Paper
&
Organolithium Chemistry
Base-Catalyzed Intramolecular Hydroamination of Cyclohexa-2,5-
dienes: Insights into the Mechanism through DFT Calculations
and Application to the Total Synthesis of epi-Elwesine
Gꢀraldine Rousseau,[a] Raphaꢁl Lebeuf,[a] Kurt Schenk,[b] Frꢀdꢀric Castet,[a] Frꢀdꢀric Robert,*[a]
and Yannick Landais*[a]
Abstract: The base-catalyzed intramolecular hydroamination
of 1-ethylaminocyclohexa-2,5-dienes is described. The trans-
formation proceeds through isomerization of the cyclohexa-
1,4-dienyl fragment into the corresponding conjugated 1,3-
diene prior to the hydroamination step. Attaching a chiral
glycinol ether auxiliary on the amino group allows the proto-
nation to occur with complete diastereocontrol. The result-
ing lithium amide then adds onto the 1,3-dienyl moiety, af-
fording the desired fused pyrrolidine ring along with the
corresponding lithium allylic anion. Protonation of the latter
then proceeds with high regiocontrol to favor the resulting
allylic amines. In contrast, when the reaction was performed
on primary amines, fused pyrrolidines bearing a homoallylic
amino group were obtained. The stereochemical course of
the process and determination of the reaction pathways
were established based on calculations performed at the
DFT level. Finally, application of the methodology to the
enantioselective synthesis of (+)-epi-elwesine, a crinane alka-
loid, is described.
Introduction
tained by using the metal-catalyzed processes mentioned
above.[8] The addition of the alkali-amide moiety onto the
olefin is generally more efficient with “activated olefins”, in-
cluding styrenes, dienes, or enynes. Hydroamination of “unacti-
vated” olefins is also known but is generally more efficient if
performed in an intramolecular manner. This important reac-
tion has been incorporated with success in several synthesis of
complex natural targets.[9] In the course of our studies on the
reactivity and desymmetrization of 1-substituted cyclohexa-2,5-
dienes,[10,11] we communicated on a highly diastereoselective
intramolecular hydroamination of 1-ethylaminocyclohexa-2,5-
dienes 1 by using a substoichiometric amount of nBuLi as
a base.[12] We provide here a full description of this work, with
experimental evidence supporting our preliminary mechanistic
hypothesis. Calculations based on density functional theory
(DFT) were carried out, affording a full picture of the outcome
of this base-mediated cascade process. The study has also
been extended to the hydroamination using primary amines,
which surprisingly led to complementary regioselectivities. Fi-
nally, application of the methodology to the total synthesis of
(ent)-epi-elwesine, a crinine-type alkaloid, is reported.
Five- and six-membered ring nitrogen-containing heterocycles,
which are ubiquitous in Nature and in biologically relevant tar-
gets, may be accessed through a wide range of methodolo-
gies.[1] Among them, hydroamination of olefins is particularly
attractive because it is atom-economical and may be catalyzed
through a plethora of achiral and chiral reagents.[2] Organolan-
thanides have received particular attention in this respect.[3]
Transition-metal catalysts (Pd, Ru, Rh, Cu, and Fe), alkaline-
earth (Mg, Ca) and rare-earth metals (Sc, Y) have also been
shown to mediate inter- and intramolecular hydroamination
processes.[4] More recently, gold complexes[5] and Brçnsted
acids[6] have also been revealed to be efficient catalysts for this
transformation. Inter- and intramolecular based-catalyzed hy-
droamination, using lithium amides has also been reported, al-
though less often.[7] Organolithium bases with chiral additives
as well as homochiral lithium amide bases have been devised
and shown to provide the expected amines, although the
enantioselectivities remained modest compared with those ob-
[a] Dr. G. Rousseau, Dr. R. Lebeuf, Dr. F. Castet, Dr. F. Robert, Prof. Dr. Y. Landais
ISM, UMR 5255, University of Bordeaux and CNRS
351, Cours de la liberation, 33400 Talence Cedex (France)
E-mail: f.robert@ism.u-bordeaux1.fr
Results and Discussion
Base-catalyzed intramolecular hydroamination of a
secondary 1-ethylaminocyclohexa-2,5-diene 1
y.landais@ism.u-bordeaux1.fr
[b] Dr. K. Schenk
Laboratoire de Cristallographie, EPFL
Le cubotron, Dorigny, 1015 Lausanne (Switzerland)
We first tested the hydroamination process starting from the
achiral N-PMB precursor 1a,[13] which was readily available
through Birch reduction of biphenyl.[14] Treatment of the latter
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201403662.
Chem. Eur. J. 2014, 20, 14771 – 14782
14771
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
intermediate IV, protonation of the latter would then occur on
the least sterically hindered site (C3) to afford allylic amine 2.
Theoretical studies of the mechanism of the base-catalyzed
deprotonation–protonation–hydroamination cascade
Scheme 1. Base-catalyzed desymmetrization of aminocyclohexa-2,5-diene 1.
With the hypothesis described above in hand, we had an
almost complete picture that could rationalize the course of
the anionic cascade process. In the meantime, we started theo-
retical investigations at the DFT level to provide further ele-
ments that would support such assumptions and rationalize
the position of the remaining double bond. DFT calculations
were conducted at the B3LYP/6-31+G(d,p) level by using the
Gaussian program package.[18] Explicit solvation being crucial
for all calculations in which lithium is involved, two dimethyl
ether ligands[19] [1,2-dimethoxyethane (DME), used commonly
in place of “costly” THF, especially for large systems] coordinat-
ed to the lithium were added to keep the lithium adducts in
a tetrahedral environment.[20] All the calculations were per-
formed in the gas phase because with explicit ligands around
lithium no significant differences were observed with calcula-
tions performed using a polarizable continuum solvation
model.[10e] We started our calculations with the lithium amide
A-I, which is likely produced in the first instance by reaction of
the starting amine with nBuLi (the kinetic acidity of the NÀH
proton is certainly superior to that of a bisallylic CÀH proton).
A-I would also be regenerated at the end of the catalytic pro-
cess because the starting amine 1 is believed to protonate the
final lithium adduct A-IV. The lithium amide A-I could thus
evolve toward the more stable pentadienyl lithium compound
A-II. This complex is a plausible starting point for the process
because dienyllithium species are expected to be monomer-
ic.[21] The reaction pathway depicted in Scheme 3 showed two
energetically low transition-states for the proton transfer at C2
with a substoichiometric amount of nBuLi (30 mol%) as a base
in tetrahydrofuran (THF) as solvent led to the expected fused
pyrrolidine 2a in essentially quantitative yield as a single regio-
isomer (Scheme 1).
Addition of the amino group onto the olefinic linkage was
accompanied by isomerization of the remaining olefin. With
a direct and totally diastereoselective addition of the amino
group onto one of the nonconjugated olefin being unlikely
under the reaction conditions, we proposed instead that
a base-catalyzed isomerization of the 1,4-diene preceded the
intramolecular proton transfer and hydroamination step onto
a 1,3-diene, to lead to the observed allylic amine as a single
regio- and diastereomer.
Examination of the pKa of the secondary amine proton [ca.
36 in dimethyl sulfoxide (DMSO)] and that of allylic protons
(ca. 35 in DMSO)[15] in dienes 1 suggest that treatment of the
latter with a substoichiometric amount of nBuLi should lead to
an equilibrium between the corresponding lithium amide I
and a cyclohexadienyl lithium species II (Scheme 2).[11h,k] Depro-
Scheme 2. Possible pathway for the deprotonation–protonation
hydroamination cascade.
tonation at the nitrogen site is probably the fastest due to the
higher kinetic acidity of the NH, but formation of the penta-
dienyl lithium moiety should be favored on thermodynamic
grounds. The proton transfer between II and III is currently un-
known but must occur intramolecularly. An alternative inter-
molecular protonation through a second equivalent of amine
was ruled out because we observed that intermolecular reac-
tions of related cyclohexadienyl anions with electrophiles (MeI,
R3SiCl, H+) occurred exclusively at C4.[10e,16] We then envisioned
that the transfer of a proton to C2 (or C6) would generate
a 1,4 diene such as III. Addition of lithium amides to dienes is
well documented and appears to be much more favorable
than addition to nonconjugated olefins of the starting 1,4-
dienes.[17] With hydroamination at C6 leading to an allyl-lithium
Scheme 3. Free Gibbs energy profile [kcalmolÀ1] for the deprotonation–
protonation hydroamination cascade.
Chem. Eur. J. 2014, 20, 14771 – 14782
www.chemeurj.org
14772
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
then hydroamination of the diene at C6, which should be
easily attained at room temperature. The starting pentadienyl
lithium A-II is the more stable form among all the computed
lithiated species, with the reaction thus being driven by the
exothermicity of the process (À8.9 kcalmolÀ1 was computed
for the free Gibbs energy difference between the starting
diene and the final allylic amine).
could nevertheless observe small enantiomeric excesses (less
than 10%) when the reaction was run in toluene in the pres-
ence of 1,2-diphenylethane-1,2-diol dimethyl ether as chiral
ligand around lithium.[24] These failures finally led us to consid-
er instead a diastereoselective approach.
Diastereoselective intramolecular hydroaminations of chiral
1-ethylaminocyclohexa-2,5-dienes 1
At the final allyl-lithium stage, a protonation should take
place regioselectively. This has been modeled by replacing one
of the two DME ligands by a dimethylamine ligand (used as
a model for the starting secondary amine) (Scheme 4).[22] Two
Dienes 1 are readily available through Birch reduction of the
corresponding arenes[13,14] followed by alkylation in situ with a-
bromo esters or nitriles. Reduction of the ester or nitrile func-
tions into the aldehyde, followed by introduction of the chiral
auxiliary through reductive amination, provided the desired
secondary amines 1b–f in good overall yields (see the Sup-
porting Information). Several chiral auxiliaries were selected, in-
cluding a-methylbenzylamine (Phe), cis-1-amino-2-methoxy-
indanol (Ind), and the O-protected phenylglycinol (Gly)
(Table 1). Preliminary attempts to cyclize amine 1b provided
Table 1. Cyclization of amines 1b–f.
Diene
R*
R
2
Yield [%][a]
d.r.[b]
67:33[c]
90:10
>98:2
>98:2
>98:2
1b
1c
1d
1e
1 f
Phe
Ind
Gly
Gly
Gly
Ph
Ph
Ph
2b
2c
2d
2e
2 f
93
26
95
46
82
Scheme 4. Free Gibbs energy profile [kcalmolÀ1] for the protonation step.
3,5-(OMe)2Ph
CH2OBn
complexes may then be drawn, depending on which DME is
replaced, one leading to the homoallylic amine after protona-
tion at C5, the second to the observed allylic amine (protona-
tion at C3). We considered the two complexes B-Ia and B-Ib to
be in equilibrium because only ligand exchange is involved.
From the calculations, it seems clear that the allylic product is
favored. The transition states leading to the two products are
quite similar in energy but the allylic amine–lithium complex
B-IIa is more stable, which is in line with the stability of the
final allylic (B-H3) and homoallylic (B-H5) amine products
(Scheme 4).
[a] Isolated yield after chromatography. [b] Estimated through 1H NMR
analysis of the crude reaction mixture. [c] Mixture of unidentified regio-
or diastereomers.
the bicyclic skeleton 2b after 4 h at room temperature in ex-
cellent yield but as a mixture of isomers (regio- or diastereo)
that could not be separated further. Introduction of an addi-
tional chelating group in precursors 1c–f proved to be benefi-
cial, leading to the bicyclic allylic amines 2c–f in good to excel-
lent yields and high levels of regio- and stereocontrol. The
best results were obtained with the simple and flexible O-
methyl protected phenylglycinol-derived chiral auxiliaries.
Lower yield and selectivity were observed with the cis-indanol
1c. The structure of the bicyclic amines was determined un-
equivocally through X-ray structure analysis of a crystal of 2e
(see the Supporting Information), confirming the position of
the double bond.
Having rationalized the complete process, enantioselective
and diastereoselective versions were then envisaged. Enantio-
selective protonation of lithiated species has received a lot of
attention, and has been recognized to be an efficient way to
deracemize chiral racemic compounds.[23] Our observations are
relevant in this context and the use of chiral ligands around
lithium was expected to open the way to a new enantioselec-
tive desymmetrization process. Unfortunately, all chiral amine-
based ligands (sparteine, ephedrine, bis-oxazolines) we tried
failed to afford satisfactory levels of enantioselectivities. We
In the proposed mechanism (Scheme 5), the proton transfer
would then proceed quantitatively at C2 rather than at C6.
This irreversible and stereochemically determining step, occur-
Chem. Eur. J. 2014, 20, 14771 – 14782
www.chemeurj.org
14773
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Hydroamination with protected phenylglycinols
Further experimental evidence that supported the proposed
mechanism were obtained. For instance, repeating the reaction
with precursor 1g, having a MOM-protected phenylglycinol,
led, as above, to the formation of the bicyclic structure (not
shown) as a 95:5 mixture of two isomers that could not be
separated at this stage (Scheme 7). Deprotection of the MOM
group under acidic conditions, however, afforded allylic amine
2g in 70% yield along with homoallylic isomer 2h in 3% yield,
proving that the final protonation could also take place at the
5-position with different lithium aggregates. More interestingly,
when precursor 1h, having an unprotected phenylglycinol
moiety as a chiral auxiliary, was treated with 1.2 equivalent of
nBuLi (1 equiv was used to deprotonate the free OH), allylic
amine 2g and homoallylic amine 2h were formed along with
bicyclic compound 2i, as a single diastereomer, the structure
of which was determined on the basis of X-ray diffraction stud-
ies. Only 8% of the starting amine 1h was also recovered,
showing that the process is high yielding.
Scheme 5. Hydroaminations of chiral 1-ethylaminocyclohexa-2,5-dienes.
ring with complete diastereocontrol, would set up the stereo-
chemistry of the quaternary center. The presence of a chelating
amino-ether chiral auxiliary is highly beneficial here, providing
sufficient rigidity to the lithium intermediate for efficient con-
trol of the selectivity during the stereochemically key proton
transfer.[25]
The presence of bicycle 2i is intriguing and should result
from a proton transfer at C6 instead of at C2, which would be
followed by an exclusive 6-exo cyclization instead of 5-exo, to
afford the bicyclo[3.3.1] system. Interestingly, the formation of
2i also confirmed the presence of the putative diene III
(Scheme 2) as an intermediate. Starting from the two comput-
ed dienes C-IIIa and C-IIIb (Scheme 6), the calculated reaction
pathways for the 6-exo cyclization exhibited interesting profiles
as depicted in Scheme 8.
Computational studies on the asymmetric deprotonation–
protonation–hydroamination cascade
DFT calculations were then extended to the chiral system,
which introduced additional challenges. Apart from the chiral
carbon on the ethylamino chain, the quaternary carbon center,
but also nitrogen and silicon atoms become chiral at some
point, thus multiplying the number of possible routes. Assum-
ing that the chirality at the lithi-
um center is labile and has a low
energy barrier, and that its inter-
conversion involves only solvent
molecules, we could find two
routes leading to the two possi-
ble
diastereomers
of
2d
(Scheme 6). The deprotonation
of C-I would then lead to two
possible pentadienyl complexes
C-IIa and C-IIb, bearing a differ-
ent chirality on the nitrogen
center, which can both undergo
hydrogen transfer, leading to C-
IIIa and C-IIIb. The 5-exo hydroa-
minations then give rise to two
possible allyllithium complexes
C-IVa and C-IVb, which, upon
protonation, give the two diaste-
reomers 2d and 2d’. The most
favorable path is clearly that
leading to the only observable
diastereomer 2d because the
two transition states, the final
complex C-IV and final product
are
all
energetically
Scheme 6. Free Gibbs energy profile [kcalmolÀ1] for the asymmetric deprotonation–protonation–hydroamination
favored.
cascade with phenylglycinol derivatives.
Chem. Eur. J. 2014, 20, 14771 – 14782
www.chemeurj.org
14774
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 9. Different reaction pathways depending on the protonation site.
at C6 would afford the 6-exo product, which has never been
isolated for this methoxy-protected phenylglycinol (Scheme 9).
Scheme 7. Cyclizations of amines 1g and 1h. X-ray structure of 2i.
Base-catalyzed intramolecular hydroamination of
cyclohexa-2,5-dienes bearing a primary amine
Having found that removing the chiral auxiliary was difficult
because of the allylic nature of the tertiary amine (see below),
the study was then extended to 1-substituted cyclohexyl-2,5-
dienes possessing a primary amine. Base-catalyzed intramolec-
ular hydroamination of primary unsaturated amines by using
nBuLi to generate the starting lithium amide has been de-
scribed.[7d,h] The precursors were prepared through Birch reduc-
tive alkylation of suitable arenes or metallation of cyclohexa-
1,4-diene (Scheme 10).[11h,k,14] Cyclohexadienyl alcohol 3 was
Scheme 8. Free Gibbs energy profile [kcalmolÀ1] for the 5-exo versus 6-exo
aminations.
Scheme 10. Synthesis of dienes with primary amines.
For diene C-IIIa (arising from protonation at C2), the 5-exo
cyclization is kinetically favored (by 8 kcalmolÀ1) over the 6-exo
route, with the resulting C-IVa also being 9.2 kcalmolÀ1 more
stable than D-IVa. With the other diene C-IIIb, the situation is
reversed. Diene C-IIIb, resulting from protonation at C6, would
cyclize preferentially in a 6-exo fashion, with D-IVb also being
1.6 kcalmolÀ1 more stable than C-IVb. Apart from the energy
difference between 2d and 2d’, the different pathways fol-
lowed by the intermediate dienes may be mainly responsible
for the unique formation of 2d. Starting from C-II, protonation
at C2 would give the 5-exo product 2d, whereas protonation
obtained in two steps from benzoic acid (56%, see the Sup-
porting Information).[26] Protection of the alcohol function was
carried out by slow addition of NaH at 08C onto the substrate
to avoid isomerization of the diene.[27] Under these conditions,
4 was obtained in good yield. This isomerization is notable
and corroborates our hypothesis that these cyclohexadienes
may be easily deprotonated in the presence of strong bases.
The presence of the chelating CH2OH substituent likely favors
such a deprotonation as seen above with the amino-ether
Chem. Eur. J. 2014, 20, 14771 – 14782
www.chemeurj.org
14775
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
moiety. Reduction of the nitrile was then performed with
LiAlH4 and AlCl3, affording the amine in high yield.
and nBuLi] led to a deep orange-red coloration upon addition
of 9a at room temperature;[28] after workup, this reaction pro-
vided the secondary amine 12a in quantitative yield as
a single regioisomer. Interestingly, primary amines led to differ-
ent regioisomers compared with those obtained from sec-
ondary amines discussed previously. This proved to be a gener-
al trend because other amines (5, 9b and 9c) led predomi-
nantly to the same regioisomers in generally high yields. Quan-
The synthesis of the silylated precursor 8 proved to be more
troublesome. By using Woerpel’s procedure,[11h,j,k] cyclohexa-
1,4-diene was metallated then silylated to afford the 1-silyl-
cyclohexa-1,4-diene as an intermediate, which was not isolat-
ed, but instead metallated in situ with sBuLi and finally alkylat-
ed to give 7 in poor yield. Reduction as above gave 8 in good
yield. The hydroamination process on precursor 9a, carried out
by using a substoichiometric amount of nBuLi (20 mol%), af-
forded the desired fused pyrrolidine as a 7:3 mixture of two re-
gioisomers 12a and 12b (Table 2, entry 1 and Scheme 11), the
structures of which were assigned as discussed below. Repeat-
ing the reaction by using 20 mol% lithium diisopropylamide
[LDA; prepared from a 2:1 mixture of diisopropylamine (DIPA)
1
titative crude yields (based on H NMR analysis) were generally
obtained, suggesting that the lower isolated yields were due
to a loss of the polar amine on silica gel. In contrast to other
substrates, silylated diene 8 did not give the desired pyrrol-
idine because of extensive desilylation under the reaction con-
ditions. The determination of the structure of regioisomers in
the previous series, based solely on 1H NMR experiments,
proved to be inefficient; therefore, the position of the double
bond in the six-membered ring of fused pyrrolidines 10–14
was established by first preparing all three regioisomers of 12
using independent routes.
Table 2. Cyclization of amines 5, 8, and 9a–c.
Fused pyrrolidines having an allylic secondary amine were
obtained from compound 2g by removing the chiral auxiliary
(Scheme 11). Although numerous methods have been devel-
oped to cleave phenylglycinol auxiliaries, these generally rely
on the unique reactivity of the benzylic amine moiety.[29] Un-
fortunately, in our case, the amine is both benzylic and allylic,
thus restricting the number of methods available. After exten-
sive studies, we found that substitution of the b-hydroxy
group with a phenylselenyl substituent[30] allowed the removal
of the glycinyl fragment after oxidation of the selenium and b-
elimination, followed by hydrolysis of the enamine under
acidic conditions to provide pyrrolidine 12b. In parallel, the
bishomoallylic amine was obtained by using a three-step pro-
cedure from alcohol 16, involving a [3+2] cycloaddition of an
azide intermediate onto one of the olefins of the cyclohexa-
dienyl moiety.[31] The desired pyrrolidine 12c was thus ob-
tained in 25% overall yield from 16.
Diene[a]
Base
Product[b]
Yield [%][a,b]
Regio[c]
9a
9a
5
8
9b
9c
nBuLi[d]
LDA[e]
LDA[e]
LDA[e]
LDA[e]
LDA[e]
12a/12b
12a
10
11
13
95
80
71
–
92
73
67:33
>98:2
>98:2
–
85:15
>98:2
14
[a] Isolated yield after chromatography. [b] Quantitative crude yields were
generally obtained before chromatography. [c] Homoallylic amine/allylic
amine ratio estimated through 1H NMR analysis of the crude reaction mix-
ture. [d] 20 mol%. [e] 20 mol% of LDA prepared from a 2:1 mixture of
DIPA and nBuLi.
Comparison of the 13C NMR data of amines 10–14 with
those of 12b and 12c showed unambiguously that the hydro-
amination of primary amines 5 and 9a–c in the presence of
LDA generates the corresponding homoallylamines, without
traces of the corresponding allylic regioisomer. This assignment
was later supported by X-ray diffraction studies of 17 prepared
through tosylation of secondary amine 12a (Scheme 11). The
formation of the homoallylic amine as a unique regioisomer
suggests a mechanism closely related to that proposed before
for the cyclization of secondary amines, but with a final proto-
nation step proceeding through a different pathway. Our cal-
culations surprisingly showed that allylamine 12b was the
thermodynamic product (Scheme 12), indicating that the
proton transfer likely proceeds through low transition-state
barriers. A direct 1,3-proton transfer was envisioned: the close
proximity of the amine proton with the allylic system perhaps
explaining the reactivity difference between secondary and
tertiary amines. However, this hypothesis was ruled out be-
cause of the very high computed transition state E-TS
(+42.6 kcalmolÀ1).
Scheme 11. Preparation of pyrrolidine regioisomers 12b and 12c and
tosylamide 17.
Chem. Eur. J. 2014, 20, 14771 – 14782
www.chemeurj.org
14776
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Application of the protonation–hydroamination cascade to
the synthesis of crinane-type alkaloids: total synthesis of
(+)-epi-elwesine
The azabicyclo[4.3.0]nonane skeleton above (2a–e and 10–14)
is found in several classes of alkaloids, including those of Amar-
yllidaceae such as crinine 18 (Scheme 14),[33] but also in ter-
Scheme 12. Computed energies for the protonation–hydroamination of a
primary amine and energy profile for the direct 1,3-proton transfer.
The cyclizations of primary amines were carried out in the
presence of an excess of diisopropylamine. It was thus envi-
sioned that regioselective protonation could result from
a proton transfer from an amine complexed to the lithium,[32]
in a fashion similar to that computed for tertiary amines (see
Scheme 4). This difference in behavior would then arise from
a chelation by hydrogen bonding between the secondary
amine and DIPA that could direct the proton transfer towards
C5 instead of C3. A computed reaction pathway involving
a double proton transfer showed a much lower energetic pro-
file than for the protonation of tertiary amines (Scheme 13).
Scheme 14. Pictet–Spengler reaction on model substrate 14.
pene indole alkaloids such as aspidospermine and more com-
plex strychnine.[34] All these natural products have in common
a quaternary center and a cis-ring junction, which still consti-
tutes a challenging synthetic problem. Desymmetrization of cy-
clohexadienes flanked with an aryl substituent and an ethyl-
amino group such as those described above offers an attrac-
tive and straightforward route through which to access the cri-
nine-type alkaloid skeleton in a limited number of steps. This
strategy, called “BRAD” (Birch reductive alkylation–desymmetri-
zation) allows the preparation of simple building blocks, open-
ing a unified access to several classes of alkaloids, including
those mentioned above.[14a] For instance, with the azabicyclo-
[4.3.0]nonene skeleton in hand, a simple Pictet–Spengler pro-
cess should allow the formation of the B-ring and enable
access to the complete skeleton of Amaryllidaceae alkaloids of
the crinine family.[31,33] Several syntheses of alkaloids of this
family have been reported in which the B-ring is formed
through a Pictet–Spengler transformation.[31,33,35] This approach
was tested on a simple model compound 14, which, upon
treatment with the Eschenmoser salt,[34b] effectively afforded
the desired tetracyclic crinine skeleton 19 in good yield
(Scheme 14).[36] Interestingly, the olefin was not affected under
these conditions. When the same reaction was carried out with
unsymmetrical analogue 13, no trace of the desired tetracycle
was observed, suggesting that in this case the arene is not suf-
ficiently nucleophilic to react with the iminium intermediate.
Under forcing conditions (formaldehyde, HCl, MeOH–H2O,
608C, 15 h), the desired product was however observed but in
a mixture of products that could not be separated further.
Based on these results, we embarked into the enantioselec-
tive synthesis of epi-elwesine 20, which belongs to the broad
family of crinine alkaloids. epi-Elwesine 20 and (+)-elwesine 21,
also called dihydrocrinine and dihydroepicrinine, respectively,
have been isolated in small quantities from a giant snowdrop
Galanthus elwesii Hook,[37] which also contains, as major com-
ponent, galanthamine, which is used in the treatment of Alz-
heimer’s disease. No enantioselective total synthesis of 20 and
21 has been reported to date.[38] However, both have been ob-
Scheme 13. Free Gibbs energy profile [kcalmolÀ1] for the protonation step
with hydrogen shuttle.
The occurrence of hydrogen bonding between the two amines
in F-II and F-TS2 lowers the transition state for a proton trans-
fer at C3 (compare +8.9 kcalmolÀ1 for F-TS1 in Scheme 13
with 13.4 kcalmolÀ1 for B-TSb in Scheme 4). One can assume
that a transition state for a protonation at C5 would be in the
same range as for B-TSa because of the absence of hydrogen
bonding at this position.
Chem. Eur. J. 2014, 20, 14771 – 14782
www.chemeurj.org
14777
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tained in enantiomerically pure form from dihydro-oxocrinine,
a closely related alkaloid, through resolution with tartaric acid
followed by reduction of the ketone function at C4.[39] We plan-
ned to rely, as a final step, on the Pictet–Spengler reaction of
a suitably functionalized azabicyclo[4.3.0]nonene skeleton.[31,33]
The latter was expected to be generated by applying the
above diastereo- and regioselective hydroamination cascade,
followed by introduction of the alcohol function to the C-ring.
The required cyclohexadiene precursor 24 was not accessible
by using our previously reported Birch reduction strategy[14]
because of the low stability of the methylenedioxy group
under these conditions. We therefore focused on the simple
precursor 25, which is available from cyclohexene-1,3-dione by
using a high-yielding five-step sequence (Scheme 15).[40] Com-
With 23 in hand, we then studied the functionalization of
ring C through direct introduction of the hydroxyl group at C-
3. Hydroboration was first tested by using various boranes.
Whereas the use of 9-borabicyclo[3.3.1]nonane (9-BBN-H) or
(Cyclohex)2BH led essentially to recovered starting material,
BH3-SMe2 provided the undesired regioisomer as its acetate 27
in quantitative yield (Scheme 17). The complete regio- and ste-
Scheme 15. Retrosynthetic analysis of crinine-type alkaloids.
Scheme 17. Functionalization of the C-ring olefin in 23.
pound 25 would then deliver 24 through oxidation of the
olefin into an aldehyde followed by a reductive amination to
install the chiral auxiliary.
reocontrol suggests that the allylic amino group probably co-
ordinates the boron center, directing the hydroboration onto
the syn face and on the vicinal olefinic center. Attempts at re-
versing the regioselectivity through oxymercuration using
Hg(OAc)2 or Hg(OCOCF3)2 were also tested but this led essen-
tially to starting compound 23. We then turned our attention
to a two-step alternative involving epoxidation followed by re-
ductive opening of the epoxide. Epoxidation of 23 with m-
chloroperoxybenzoic acid (m-CPBA) led to 28a as a single ste-
reoisomer, the syn-relative configuration of which was assigned
based on X-ray diffraction studies of a derivative (i.e., 29, see
below). Similarly, model compound 2d led to the correspond-
ing epoxide 28b with the same stereoselectivity. The high syn-
stereocontrol could arise either through a directed epoxidation
from the corresponding ammonium[42] or from the N-oxide
formed in situ (Scheme 17).[43] The latter would be formed by
oxidation of the amine before epoxidation and then reduction
with Na2S2O5, thus regenerating the tertiary amine. Epoxide
28a was then treated with various hydride sources [LiAlH4, di-
isobutylaluminum hydride (DIBAL-H), LiBEt3H], but essentially
no reaction or degradation of the starting material was ob-
served. The results above indicate that the sterically hindered
chiral auxiliary prevents reagents from approaching the olefinic
moiety or the epoxide. Removal of the phenylglycinol methyl
Based on this disconnection, 25 was converted into 24
through a four-step, two-pot sequence involving a Lemieux–
Johnson oxidation[41] followed by reductive amination
(Scheme 16). Compound 24 was used directly in the hydroami-
nation step without further purification to provide, using a sub-
stoichiometric amount of nBuLi, azabicyclo[4.3.0]nonene inter-
mediate 23 (R*=Gly) in 50% overall yield from 25.
Scheme 16. Synthesis of azabicyclo[4.3.0]nonene 23.
Chem. Eur. J. 2014, 20, 14771 – 14782
www.chemeurj.org
14778
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ether was thus carried out through hydrogenolysis, leading to
the secondary amine 29 in good yield. X-ray diffraction studies
on 29 secured its relative configuration and those of 28a–b.
All our efforts to open the epoxide of 29 met with failure, irre-
spective of the hydride source used.
Although our efforts to restore the unsaturation on ring C in
29 were not successful (see above), the sequence was repeat-
ed with the Boc-protected amine 30, and we were pleased to
observe that treatment of the latter with Ph3P-I2 in N,N-dime-
thylformamide (DMF) led to the formation of the correspond-
ing iodohydrins as a mixture of regioisomers, which were di-
rectly submitted to the elimination step using zinc dust. This
afforded the desired olefin 32a in a satisfactory 35% overall
yield from 2d (four steps; Scheme 19). The same strategy was
Previous work by Overman et al.[38c] on the synthesis of race-
mic epi-elwesine showed that an intermediate such as 23,
devoid of chiral auxiliary, could be converted in two steps into
the natural product. It was thus decided to employ the epox-
ide as an olefin protective group that could then be unmasked
later in the synthesis. Epoxide ring opening to form the corre-
sponding iodohydrin,[44] followed by zinc-mediated elimina-
tion[45] would thus be used, respectively, to protect then re-
store the olefinic appendage. Treatment of 29 with a mixture
of PPh3 and I2 led effectively to epoxide ring opening and for-
mation of the iodohydrin as a mixture of two regioisomers,
albeit in 30% yield. Unfortunately, addition of zinc dust to this
mixture led to unreproducible results, indicating that the free
amino group reacted under these conditions. This was solved
by removing the chiral auxiliary through hydrogenolysis with
protection of the secondary amine in situ as a carbamate. By
using this strategy, carbamate 30, as a model compound, was
obtained in moderate yield (the modest yield does not reflect
the efficiency of the two-step, one-pot process, due to the loss
of material upon purification; Scheme 18). Epoxide ring open-
Scheme 19. Completion of the total synthesis of (+)-epi-elwesine 20.
applied to the epi-elwesine precursor 28a, providing olefin
32b in 33% yield from 23. The Boc protective group was then
removed under acidic conditions [trifluoroacetic acid (TFA)],
leading to secondary amine 33. Introduction of the hydroxyl
group at C-3 was then performed through oxy-mercuration fol-
lowing Overman’s precedent,[38c] giving rise to the alcohol 34
in 55% yield. Pictet–Spengler reaction onto 34 using formalde-
hyde finally afforded enantiomerically pure (+)-epi-elwesine
20, the natural product enantiomer.
Conclusion
Scheme 18. Attempts at opening the epoxide of 30.
We provide here a full account on our research on lithium
base-catalyzed intramolecular hydroamination of 1-substituted
cyclohexa-2,5-dienes. When a secondary amine is present on
the ethylamino substituent at C-1, azabicyclo[4.3.0]nonenes
bearing an allylic amine are generated. In contrast, primary
amines linked to the quaternary center provide the corre-
sponding homoallylic amines. A mechanism has been pro-
posed that rationalizes the different observations. Isomeriza-
tion of the cyclohexa-2,5-diene system prior to the hydroami-
ing was tested with LiBEt3H, but again this led to the unde-
sired regioisomer 31. The use of the Lewis acidic DIBAL-H also
afforded 31 in 46% yield. The regioselectivity of the epoxide
ring opening follows the Fꢃrst–Plattner rules,[46] favoring
a trans-diaxial mode and the formation of the isomer in a chair
conformation instead of the other approach that leads to the
more energetic twist-boat conformation.
Chem. Eur. J. 2014, 20, 14771 – 14782
www.chemeurj.org
14779
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
at RT overnight, the solution was filtered twice through a pad of
Celite and concentrated in vacuo. I2 (1.02 g, 4 mmol, 2 equiv) was
added to a solution of crude epoxide (2 mmol, 1 equiv) and PPh3
(1.05 g, 4 mmol, 2 equiv) in DMF (15 mL). The reaction mixture was
stirred at RT for 1.5 h. Zn dust (650 mg, 10 mmol, 5 equiv) was
added and the resulting reaction mixture was stirred at RT for 2 h.
Et2O was added and the organic layer was washed with H2O (3ꢄ),
brine, dried over MgSO4, filtered, and concentrated in vacuo. Purifi-
cation by flash chromatography (silica gel; petroleum ether/EtOAc,
9:1) gave alkene 32b (195 mg, 33% over four steps) as a colorless
oil. 1H NMR (300 MHz, CDCl3): d (two rotamers)=6.78 (br s, 1H),
6.72 (br s, 2H), 6.23–6.21 (m, 0.5H), 6.24–6.21 (m, 0.5H), 5.93 (s,
2H), 5.86–5.83 (m, 1H), 4.39 (br s, 0.5H), 4.26 (br s, 0.5H), 3.46–3.32
(m, 2H), 2.18–1.85 (m, 6H), 1.50 (s, 4.5H), 1.45 ppm (s, 4.5H);
13C NMR (62.5 MHz, CDCl3): d (two rotamers)=154.6, 147.6, 135.8,
140.4, 128.7, 128.24, 127.1, 126.8, 118.8, 107.9, 106.8, 100.9, 79.5,
58.9, 47.6, 47.0, 43.7, 34.7, 31.6, 28.6, 22.1 ppm; IR (neat, NaCl): n˜ =
3051, 2976, 2929, 1689, 1507, 1489, 1434, 1402, 1366 cmÀ1; HRMS
(LSIMS): m/z: calcd for C20H25NO4: 343.1783 [M]+; found: 343.1779.
nation step was suggested, which is supported both by the
isolation and the structure determination of minor byproducts,
as well as by DFT calculations. These computational studies
have also shed light on the origin of the enantioselectivity and
have provided useful information on the organolithium inter-
mediates involved in such processes. Finally, the value of the
methodology was demonstrated by the facile access to cri-
nane-type alkaloids and the completion of the first enantiose-
lective synthesis of the enantiomer of naturally occurring epi-
elwesine, an alkaloid of the crinine family.
Experimental Section
General procedure for hydroamination reactions
The amine (1.0 equiv.) was dissolved in THF (1m), and nBuLi (2.5m
in hexane, 0.2 equiv) is added dropwise at RT. The reaction was
stirred for 4 h during which the solution turned from orange to
yellow. A drop of water was added and the organic layer was dilut-
ed with diethyl ether, washed with saturated Na2CO3 solution, and
dried over sodium sulfate. After removal of the solvents, the prod-
uct was obtained as a colorless oil in nearly quantitative yield.
3a-(Benzo[d][1,3]dioxol-5-yl)-2,3,3a,4,5,7a-hexahydro-1H-indole
(33): At 08C, TFA (5.4 mL, 72.7 mmol, 135 equiv) was added to a so-
lution of protected amine 32b (186 mg, 0.54 mmol, 1 equiv) in
CH2Cl2 (27 mL) and the reaction mixture was stirred at this temper-
ature for 30 min. Aq. NaOH (3m) was added until the pH remained
basic. The reaction mixture was then stirred at RT for 50 min. After
separation of the two layers, the aqueous layer was extracted with
CH2Cl2 (2ꢄ). The combined organic layers were washed with brine,
dried over MgSO4, filtered, and concentrated in vacuo. Purification
by flash chromatography (silica gel; CH2Cl2/MeOH/28% aq NH4OH:
9:1:0.2) gave amine 33 (112 mg, 85%) as a white solid. Analytical
data were in good agreement with those reported in the litera-
ture.[38] [a]D +57.9 (c=0.01 in CHCl3).
3a-(Benzo[d][1,3]dioxol-5-yl)-1-((S)-2-methoxy-1-phenylethyl)-
2,3,3a,4,5,7a-hexahydro-1H-indole (23): [a]D = +120.6 (c=0.02
1
in CHCl3); H NMR (300 MHz, CDCl3): d=7.30–7.23 (m, 5H), 6.75 (m,
1H), 6.70–6.66 (m, 2H), 6.09 (m, 1H), 5.96 (m, 1H), 5.90 (s, 2H),
4.19–4.10 (m, 1H), 3.78 (m, 2H), 3.34 (s, 3H), 3.10 (m, 1H), 2.96 (m,
1H), 2.56 (m, 1H), 2.06–1.85 (m, 4H), 1.65–1.57 ppm (m, 2H);
13C NMR (75.4 MHz, CDCl3): d=147.1, 145.1, 142.3, 137.9, 137.9,
131.3, 128.7, 128.1, 126.0, 119.3, 107.4, 100.7, 75.1, 62.2, 61.1, 58.8,
46.3, 45.7, 34.6, 34.6, 23.5 ppm; IR (neat, NaCl): n˜ =3024, 2922,
3a-Benzo[1,3]dioxol-5-yl-octahydroindol-6-ol (34). A solution of
alkene 33 (112 mg, 0.46 mmol, 1 equiv) in THF (0.5 mL) was added
to a mixture of mercuric acetate (291 mg, 0.91 mmol, 2 equiv) in
H2O (1.8 mL) and THF (0.5 mL). The reaction mixture was stirred for
24 h at RT before NaBH4 (0.5m in 3N NaOH, 1.8 mL, 0.46 mmol)
and 3N NaOH (0.5 mL) were added. After 30 min, the mixture was
quenched with solid potassium carbonate and extracted with Et2O
(2ꢄ12 mL). The combined organic extracts were dried over MgSO4,
filtered, and concentrated in vacuo. Purification by flash chroma-
tography (silica gel; CH2Cl2/MeOH/28% aq NH4OH, 9:1:0.2) gave al-
cohol 34 (66 mg, 55%) as a white solid. Mp 87–938C; [a]D +63.4
(c=0.009 in CHCl3); 1H NMR (250 MHz, CDCl3): d=6.84–6.71 (m,
3H), 5.92 (s, 2H), 4.02 (br. s, 1H), 3.95 (br. s, 1H), 3.71 (br. s, 1H),
3.19–2.95 (m, 2H), 2.25 (td, J=14.3, 3.4 Hz, 1H), 2.06–1.65 (m, 6H),
1.37 ppm (tt, J=13.1, 2.4 Hz, 1H); 13C NMR (62.5 MHz, CDCl3): d=
147.7, 145.5, 139.2, 119.3, 107.8, 107.4, 100.9, 66.6, 60.5, 46.3, 42.7,
41.7, 31.4, 29.3, 26.5 ppm; IR (neat, NaCl): n˜ =3308, 1610, 1505,
1487, 1434, 1232 cmÀ1; HRMS (LSIMS): m/z: calcd for C15H20NO3:
262.1443 [M+H]+; found: 262.1446.
(+)-epi-Elwesine (20):[38] Aqueous formaldehyde (37%, 0.5 mL,
2.7 mmol, 17 equiv) was added to a solution of 34 (43 mg,
0.16 mmol, 1 equiv) in MeOH (0.2 mL). The reaction mixture was
stirred for 15 min, then poured into 6N HCl (7 mL). After stirring for
12 h, the reaction solution was basified with 28% aq NH4OH and
extracted with CHCl3 (2ꢄ). The combined organic extracts were
washed with H2O, dried over MgSO4, filtered, and concentrated in
vacuo. Purification by flash chromatography (silica gel; CH2Cl2/
MeOH/28% aq NH4OH, 9:1:0.2) gave epi-Elwesine 20 (38 mg, 85%)
as a white solid. [a]D +15.7 (c=0.7 in CHCl3); 1H NMR (250 MHz,
CDCl3): d=6.68 (s, 1H), 6.45 (s, 1H), 5.88 (s, 2H), 4.32 (d, J=
16.8 Hz, 1H), 3.74 (d, J=16.8 Hz, 1H), 3.67–3.55 (m, 1H), 3.39–3.29
1685, 1600, 1191, 1109 cmÀ1
; HRMS (LSIMS): m/z: calcd for
C24H28NO3: 378.2069 [M+H]+; found: 378.2068;.
3a-Benzo[1,3]dioxol-5-yl-6-(2-methoxy-1-phenylethyl)octahydro-
1-oxa-6-azacyclopropa[e]indene (28a): At 08C, m-CPBA (70%)
(74 mg, 0.3 mmol, 3 equiv) was added to a solution of alkene 23
(38 mg, 0.1 mmol, 1 equiv) in CH2Cl2 (1 mL). The reaction mixture
was stirred at 08C for 3 h and at RT overnight, then washed three
times with 10% Na2S2O5 aqueous solution, three times with a 10%
NaHCO3 aqueous solution, and, finally, brine. The organic solution
was dried over MgSO4, filtered, and concentrated in vacuo. Purifica-
tion by flash chromatography (silica gel; petroleum ether/EtOAc,
9:1) gave epoxide 28a (22 mg, 56%) as a yellow oil. [a]D +70.5
1
(c=0.02 in CHCl3); H NMR (250 MHz, CDCl3): d=7.35–7.25 (m, 5H),
6.70–6.51 (m, 3H), 5.90 (s, 2H), 4.38 (t, J=6.6 Hz, 1H), 3.92 (dd, J=
9.8, 6.4 Hz, 1H), 3.80 (dd, J=9.8, 7.2 Hz, 1H), 3.41–3.34 (m, 1H),
3.36 (s, 3H), 3.23 (br s, 1H), 3.19–3.17 (m, 1H), 2.96–2.89 (m, 1H),
2.50–2.43 (m, 1H), 1.97–1.75 (m, 4H), 1.34–1.14 ppm (m, 2H);
13C NMR (62.5 MHz, CDCl3): d=147.4, 145.2, 141.7, 138.0, 128.7,
128.2, 127.3, 119.5, 107.6, 107.5, 100.8, 75.1, 62.6, 61.0, 58.7, 52.7,
51.1, 45.7, 45.4, 39.0, 31.6, 21.5 ppm; IR (neat, NaCl): n˜ =2924, 1676,
1490, 1437, 1341, 1234, 1114, 1039 cmÀ1; MS (EI): m/z (%): 348 (100)
[MÀC2H5O]+; HRMS (LSIMS): m/z: calcd for C24H28NO4: 394.2018
[M+H]+; found: 394.2030.
3a-Benzo[1,3]dioxol-5-yl-2,3,3a,4,5,7a-hexahydroindole-1-car-
boxylic acid tert-butyl ester (32b): In a three-necked flask, Pd/C
(430 mg, 0.41 mmol, 20 mol%) and HClO4 (two drops) were added
to a solution of epoxide 28a (800 mg, 2.03 mmol, 1 equiv) and
Boc2O (1.33 g,6.1 mmol, 3 equiv) in EtOH/H2O (2:1, 40 mL). The re-
action mixture was degassed with three freeze-pump-thaw cycles
and H2 was introduced with a balloon (pump–H2, 3ꢄ). After stirring
Chem. Eur. J. 2014, 20, 14771 – 14782
www.chemeurj.org
14780
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[5] a) G. Kovꢅcs, G. Ujaque, A. Lledos, J. Am. Chem. Soc. 2008, 130, 853–
864; b) D. S. Nathan, V. Rauniyar, L. H. Gregory, J. Wu, D. D. Toste, Nature
2011, 470, 245–249.
(m, 1H), 2.94 (dd, J=12.0, 5.1 Hz, 1H), 2.87–2.76 (m, 1H), 2.64
(br. s, 1H), 2.40–2.35 (m, 1H), 2.25–2.13 (m, 1H), 2.09–1.93 (m, 2H),
1.79–1.46 (m, 3H), 1.30 ppm (q, J=12.2 Hz, 1H); 13C NMR (50 MHz,
CDCl3): d=146.1, 145.6, 141.1, 125.8, 106.1, 103.3, 100.6, 68.8, 66.4,
62.1, 51.9, 42.1, 37.8, 36.7, 30.8, 26.6 ppm; IR (neat, NaCl): n˜ =2390,
1642, 1488, 1243 cmÀ1; HRMS (ESI): m/z: calcd for C16H20NO3:
274.1437 [M+H]+; found: 274.1444.
[6] B. Schlummer, J. F. Hartwig, Org. Lett. 2002, 4, 1471–1474.
[7] Intermolecular: a) C. S. Barry, N. S. Simpkins, Tetrahedron Lett. 2007, 48,
8192–8195; b) C. G. Hartung, C. Breindl, A. Tillack, M. Beller, Tetrahedron
2000, 56, 5157–5162; c) J. A. Seijas, M. P. Vasquez-Tato, C. M. M. Entenza
Martinez, M. G. Onega, S. Veiga, Tetrahedron Lett. 1998, 39, 5073–5076;
Intramolecular: d) A. Ates, C. Quinet, Eur. J. Org. Chem. 2003, 1623–
1626; e) S. Tsuchida, A. Kaneshige, T. Ogata, H. Baba, Y. Yamamoto, K.
Tomioka, Org. Lett. 2008, 10, 3635–3638; f) W. Zhang, J. B. Werness, W.
Tang, Org. Lett. 2008, 10, 2023–2026; g) C. Quinet, P. Jourdain, C. Her-
mans, A. Ates, I. Lucas, I. E. Marko, Tetrahedron 2008, 64, 1077–1087;
h) C. Quinet, L. Sampoux, E. Marko, Eur. J. Org. Chem. 2009, 1806–1811;
i) W. A. L. van Otterlo, R. Pathak, C. B. De Koning, M. A. Fernandes, Tetra-
hedron Lett. 2004, 45, 9561–9563; j) W. Zhang, J. B. Werness, W. Tang,
Tetrahedron 2009, 65, 3090–3095.
Acknowledgements
CNRS and MENRT are gratefully acknowledged for PhD grants
to G.R. and R.L. Dr. B. Kauffmann (IECB, University Bordeaux-1)
is thanked for diffraction studies.
[8] a) T. Ogata, A. Ujihara, S. Tsuchida, T. Shimizu, A. Kaneshige, K. Tomioka,
Tetrahedron Lett. 2007, 48, 6648–6650; b) P. H. Martinez, K. C. Hultzsch,
F. Hampel, Chem. Commun. 2006, 2221–2223; c) J. Deschamp, C. Olier,
E. Schulz, R. Guillot, J. Hannedouche, J. Collin, Adv. Synth. Catal. 2010,
352, 2171–2176; d) J. Deschamp, J. Collin, J. Hannedouche, E. Schulz,
Eur. J. Org. Chem. 2011, 3329–3338; e) T. Ogata, T. Kimachi, K. Yamada,
Y. Yamamoto, K. Tomioka, Heterocycles 2012, 86, 469–485.
[9] a) K. A. Parker, D. Fokas, J. Am. Chem. Soc. 1992, 114, 9688–9689; b) J.
Mulzer, G. Durner, D. Trauner, Angew. Chem. 1996, 108, 3046–3048;
Angew. Chem. Int. Ed. Engl. 1996, 35, 2830–2832; c) D. Trauner, J. W.
Bats, A. Werner, J. Mulzer, J. Org. Chem. 1998, 63, 5908–5918; d) H.
Nagata, N. Miyazawa, K. Ogasawara, Chem. Commun. 2001, 1094–1095;
e) B. M. Trost, W. Tang, F. D. Toste, J. Am. Chem. Soc. 2005, 127, 14785–
14803; f) M. Varin, E. Barrꢀ, B. Iorga, C. Guillou, Chem. Eur. J. 2008, 14,
6606–6608.
[10] a) Y. Landais, Chimia 1998, 52, 104–111; b) N. Abd Rahman, Y. Landais,
Curr. Org. Chem. 2002, 6, 1369–1395; c) R. Beniazza, J. Dunet, F. Robert,
K. Schenk, Y. Landais, Org. Lett. 2007, 9, 3913–3916; d) J. Dunet, F.
Robert, K. Schenk, Y. Landais, Org. Lett. 2010, 12, 2178–2181; e) G. Rous-
seau, F. Robert, Y. Landais, Chem. Eur. J. 2009, 15, 11160–11173; f) G.
Rousseau, F. Robert, K. Schenk, Y. Landais, Org. Lett. 2008, 10, 4441–
4444.
[11] For related work on 1-substituted cyclohexa-2,5-dienes, see: a) A.
Studer, F. Schleth, Synlett 2005, 3033–3041; b) M. C. Elliott, J. S. Paine,
Org. Biomol. Chem. 2009, 7, 3455–3462; c) M. S. Maji, R. Frohlich, A.
Studer, Org. Lett. 2008, 10, 1847–1850; d) R. Umeda, A. Studer, Org. Lett.
2007, 9, 2175–2178; e) M. Butters, M. C. Elliott, J. Hill-Cousins, J. S.
Paine, J. K. E. Walker, Org. Lett. 2007, 9, 3635–3638; f) G. Errasti, C.
Koundꢀ, O. Mirguet, T. Lecourt, L. Micouin, Org. Lett. 2009, 11, 2912–
2915; g) R. Umeda, A. Studer, Org. Lett. 2008, 10, 993–996; h) C. W. Ro-
berson, K. A. Woerpel, J. Am. Chem. Soc. 2002, 124, 11342–11348; i) D.
Chappell, M. G. B. Drew, S. Gibson, L. M. Harwood, A. T. Russell, Synlett
2010, 517–520; j) A. Studer, S. Amrein, Angew. Chem. 2000, 112, 3196–
3198; Angew. Chem. Int. Ed. 2000, 39, 3080–3082; Angew. Chem. 2000,
112, 3196–3198; k) A. Studer, S. Amrein, F. Schleth, T. Schulte, J. C.
Walton, J. Am. Chem. Soc. 2003, 125, 5726–5733.
[12] R. Lebeuf, F. Robert, K. Schenk, Y. Landais, Org. Lett. 2006, 8, 4755–
4758.
[13] Compound 1a is readily available through Birch reduction of biphenyl
followed by alkylation in situ with a-bromo acetonitrile, reduction of
the nitrile function into an amine, followed by reductive amination with
p-anisaldehyde (see the Supporting Information).
[14] a) R. Lebeuf, J. Dunet, R. Beniazza, D. Ibrahim, G. Bose, F. Robert, Y.
Landais, J. Org. Chem. 2009, 74, 6469–6478; b) R. Lebeuf, F. Robert, Y.
Landais, Org. Lett. 2005, 7, 4557–4560.
[15] a) F. G. Bordwell, G. E. Drucker, H. E. Fried, J. Org. Chem. 1981, 46, 632–
635; b) E. M. Arnett, K. G. Venkatasubramanian, J. Org. Chem. 1983, 48,
1569–1578.
[16] R. B. Bates, D. W. Gosselink, J. A. Kaczynski, Tetrahedron Lett. 1967, 8,
199–204.
[17] a) R. J. Schlott, J. C. Falk, K. W. Narducy, J. Org. Chem. 1972, 37, 4243–
4245; b) K. Takabe, T. Yamada, T. Katagiri, J. Tanaka, Org. Synth. 1989, 67,
44–47.
Keywords: alkaloids
reaction mechanisms · rearrangements
·
hydroamination
·
isomerization
·
[1] a) Asymmetric Synthesis of Nitrogen Heterocycles (Ed.: J. Royer), Wiley-
VCH, Weinheim, 2009 and references cited therein; b) T. M. Miller, D. W.
Cleveland, Science 2005, 307, 361–362; c) S. France, D. J. Guerin, S. J.
Miller, T. Lectka, Chem. Rev. 2003, 103, 2985–3012.
[2] For reviews on hydroamination of olefins, see: a) T. E. Mꢃller, M. Beller,
Chem. Rev. 1998, 98, 675–703; b) J. Seayad, A. Tillack, C. G. Hartung, M.
Beller, Adv. Synth. Catal. 2002, 344, 795–812; c) T. E. Mꢃller, K. C.
Hultzsch, M. Yus, F. Foubelo, M. Tada, Chem. Rev. 2008, 108, 3795–3892;
d) K. C. Hultzsch, Org. Biomol. Chem. 2005, 3, 1819–1824; e) I. Aillaud, J.
Collin, J. Hannedouche, E. Schulz, Dalton Trans. 2007, 5105–5118;
f) K. C. Hultzsch, Adv. Synth. Catal. 2005, 347, 367–391.
[3] a) S. Hong, T. J. Marks, Acc. Chem. Res. 2004, 37, 673–686; b) P. W.
Roesky, T. E. Muller, Angew. Chem. 2003, 115, 2812–2814; Angew. Chem.
Int. Ed. 2003, 42, 2708–2710; c) D. Riegert, J. Collin, A. Meddour, E.
Schulz, A. Trifonov, J. Org. Chem. 2006, 71, 2514–2517; d) J. Collin, J.-C.
Daran, O. Jacquet, E. Schulz, A. Trifonov, Chem. Eur. J. 2005, 11, 3455–
3462; e) M. R. Gagne, L. Brard, V. P. Conticello, M. A. Giardello, C. L.
Stern, T. J. Marks, Organometallics 1992, 11, 2003–2005; f) M. A. Giardel-
lo, V. P. Conticello, L. Brard, M. R. Gagnꢀ, T. J. Marks, J. Am. Chem. Soc.
1994, 116, 10241–10254; g) X. Yu, T. J. Marks, Organometallics 2007, 26,
365–376; h) C. Quinet, A. Ates, E. Marko, Tetrahedron Lett. 2008, 49,
5032–5035.
[4] With Rh: a) M. Beller, H. Trauthwein, M. Eichberger, C. Breindl, J. Herwig,
T. E. Mꢃller, O. R. Thiel, Chem. Eur. J. 1999, 5, 1306–1319; with Pd:
b) B. M. Cochran, F. M. Michael, J. Am. Chem. Soc. 2008, 130, 2786–2792.
With Pt: c) H. Qian, R. A. Widenhoefer, Org. Lett. 2005, 7, 2635–2638.
With Ir: d) R. Dorta, P. Egli, F. Zurcher, A. Togni, J. Am. Chem. Soc. 1997,
119, 10857–10858. With Cu: e) H. Ohmiya, T. Moriya, M. Sawamura, Org.
Lett. 2009, 11, 2145–2147. With Fe: f) K. Komeyama, T. Morimoto, K.
Takaki, Angew. Chem. 2006, 118, 3004–3007; Angew. Chem. Int. Ed.
2006, 45, 2938–2941; Angew. Chem. 2006, 118, 3004–3007. With Zr
and Ti: g) P. D. Knight, I. Munslow, P. N. O’Shaughnessy, P. Scott, Chem.
Commun. 2004, 894–895; h) M. C. Wood, D. C. Leitch, C. S. Yeung, J. A.
Kozak, L. L. Schafer, Angew. Chem. 2007, 119, 358–362; Angew. Chem.
Int. Ed. 2007, 46, 354–358; Angew. Chem. 2007, 119, 358–362; i) A. L.
Gott, A. J. Clarke, G. J. Clarkson, P. Scott, Chem. Commun. 2008, 1422–
1424; j) C. Mꢃller, W. Saak, S. Doye, Eur. J. Org. Chem. 2008, 2731–2739.
With Sc: k) D. V. Gribkov, K. C. Hultzsch, F. Hampel, J. Am. Chem. Soc.
2006, 128, 3748–3759. Yttrium: l) D. V. Gribkov, K. C. Hultzsch, Chem.
Commun. 2004, 730–731; m) D. V. Gribkov, K. C. Hultzsch, F. Hampel,
Chem. Eur. J. 2003, 9, 4796–4810; n) P. N. O’Shaughnessy, P. Scott, Tetra-
hedron: Asymmetry 2003, 14, 1979–1983; o) J. Y. Kim, T. Livinghouse,
Org. Lett. 2005, 7, 1737–1739. With Mg and Ca: p) M. R. Crimmin, M. Ar-
rowsmith, A. G. M. Barrett, I. J. Casely, M. S. Hill, P. A. Procopiou, J. Am.
Chem. Soc. 2009, 131, 9670–9685; q) X. Zhang, T. J. Emge, K. C.
Hultzsch, Angew. Chem. 2012, 124, 406–410; Angew. Chem. Int. Ed.
2012, 51, 394–398; Angew. Chem. 2012, 124, 406–410.
Chem. Eur. J. 2014, 20, 14771 – 14782
www.chemeurj.org
14781
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[18] Gaussian 09, Revision A.1, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Men-
nucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian,
A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara,
K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro,
M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Ko-
bayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyen-
gar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B.
Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,
O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg,
S. Dapprich, A. D. Daniels, ꢆ. Farkas, J. B. Foresman, J. V. Ortiz, J. Cio-
slowski, and D. J. Fox, Gaussian, Inc., Wallingford CT, 2009.
[29] a) M. D. Ennis, R. L. Hoffman, N. B. Ghazal, D. W. Old, P. A. Mooney, J.
Org. Chem. 1996, 61, 5813–5817; b) C. Agami, F. Couty, G. Evano, Tetra-
hedron Lett. 1999, 40, 3709–3712; c) K. Higashiyama, H. Inoue, T. Ya-
mauchi, H. Takahashi, J. Chem. Soc. Perkin Trans. 1 1995, 111 – 115 ; d) M.
Mehmandoust, C. Marazano, B. C. Das, J. Chem. Soc. Chem. Commun
1989, 1185–1187.
[30] R. A. Bragg, J. Clayden, M. Bladon, O. Ichihara, Tetrahedron Lett. 2001,
42, 3411–3414.
[31] J. M. Schkeryantz, W. H. Pearson, Tetrahedron 1996, 52, 3107–3116.
[32] D. B. Collum, A. J. McNeil, A. Ramirez, Angew. Chem. 2007, 119, 3060–
3077; Angew. Chem. Int. Ed. 2007, 46, 3002–3017; Angew. Chem. 2007,
119, 3060–3077 and references cited therein.
[33] S. F. Martin, The Alkaloids (Ed.: A. Brossi) Academic Press, New York,
1987, Vol. 30, p. 251–376.
[19] a) L. M. Pratt, R. Mu, C. Cartera, B. Woodford, Tetrahedron 2007, 63,
1331–1338; b) A. Streitwieser, J. R. Reyes, T. Singhapricha, S. Vu, K.
Shah, J. Org. Chem. 2010, 75, 3821–3830; c) E. E. Kwan, D. A. Evans, Org.
Lett. 2010, 12, 5124–5127; d) P. Rçnnholm, J. Grꢇfenstein, P.-O. Norrby,
G. Hilmersson, S. O. Nilsson Lill, Org. Biomol. Chem. 2012, 10, 2807–
2814; e) O. LarraÇaga, A. de Cꢈzar, F. M. Bickelhaupt, R. Zangi, F. P.
Cossꢉo, Chem. Eur. J. 2013, 19, 13761–13773.
[20] a) C. C. Pye, Int. J. Quantum Chem. 2000, 76, 62–76; b) A. Streitwieser, J.
Org. Chem. 2009, 74, 4433–4446.
[21] a) M. Stꢇhle, M. Schlosser, J. Organomet. Chem. 1981, 220, 277–283;
b) H. J. Reich, Chem. Rev. 2013, 113, 7130–7178.
[34] a) M. Hesse, Alkaloids: Nature’s Curse or Blessing? Wiley, VHCA, Zꢃrich,
2002, p.413; b) J. E. Saxton, in The Alkaloids, Alkaloids of the Aspidosper-
mine Group (Ed.: G. A. Cordell) Academic Press, New York, 1998, Vol. 51,
p.1; c) “Monoterpenoid Indole Alkaloids”, W. A. Creasey, in The Chemis-
try of Heterocyclic Compounds, Supplement to Vol. 25, Part 4 (Eds.: E. C.
Taylor, J. E. Saxton), Wiley, New York, 1994, pp. 715–754.
[35] a) S. F. Martin, C. L. Campbell, J. Org. Chem. 1988, 53, 3184–3190;
b) G. E. Keck, R. R. Webb, J. Am. Chem. Soc. 1981, 103, 3173–3177.
[36] Compound 19 has been described previously (see ref. [14b]), but its
structure was incorrectly assigned because of the lack of X-ray analysis
at that time.
[22] Lithium dimethylamide has been used as an LDA surrogate when
studying reactivity, see: a) K. J. Singh, D. B. Collum, J. Am. Chem. Soc.
2006, 128, 13753–13760; b) J. C. Riggs, K. J. Singh, M. Yun, D. B. Collum,
J. Am. Chem. Soc. 2008, 130, 13709–13717.
[23] a) L. Duhamel, P. Duhamel, J.-C. Launay, J.-C. Plaquevent, Bull. Soc. Chim.
Fr. 1984, 421–430; b) C. Fehr, Angew. Chem. 1996, 108, 2726–2748;
Angew. Chem. Int. Ed. Engl. 1996, 35, 2566–2587; Angew. Chem. 1996,
108, 2726–2748; c) A. Yanagisawa, H. Yamamoto, Protonation of eno-
lates in Comprehensive Asymmetric Catalysis, Eds. E. N. Jacobsen, A.
Pfaltz, H. Yamamoto, Springer, Berlin, 1999, Vol. 3, 1295–1306.
[24] a) M. Suzuki, Y. Kawamoto, T. Sakai, Y. Yamamoto, K. Tomioka, Org. Lett.
2009, 11, 653–655; b) S. Harada, T. Sakai, K. Takasu, K.-i. Yamada, Y. Ya-
mamoto, K. Tomioka, Tetrahedron 2013, 69, 3264–3273.
[37] H. G. Boit, W. Dçpke, Naturwiss. Unterr. Chem. 1961, 48, 406–407.
[38] For total synthesis of racemic epi-elwesine and elwesine, see: a) I. H.
Sanchez, F. J. Lopez, H. J. Flores, M. I. Larraza, Heterocycles 1983, 20,
247–254; b) I. H. Sanchez, F. J. Lopez, J. J. Soria, M. I. Larraza, H. J.
Flores, J. Am. Chem. Soc. 1983, 105, 7640–7643; c) L. E. Overman, R. M.
Burk, Tetrahedron Lett. 1984, 25, 5739–5742; d) R. V. Stevens, L. E. Du-
pree, Jr., J. Chem. Soc. Chem. Commun. 1970, 1585–1586.
[39] H. Irie, S. Uyeo, A. Yoshitake, J. Chem. Soc. C 1968, 1802–1804.
[40] G. Rousseau, F. Robert, Y. Landais, Synthesis 2010, 1223–1228.
[41] R. Pappo, J. D. Allen, R. Lemieux, W. Johnson, J. Org. Chem. 1956, 21,
478–479.
[42] M. B. Brennan, T. D. W. Claridge, R. G. Compton, S. G. Davies, A. M.
Fletcher, M. C. Henstridge, D. S. Hewings, W. Kurosawa, J. A. Lee, P. M.
Roberts, A. K. Schoonen, J. E. Thomson, J. Org. Chem. 2012, 77, 7241–
7261.
[25] D. Kuck, J. Schneider, H.-F. Grꢃtzmacher, J. Chem. Soc. Perkin Trans. 2
1985, 689–696.
[26] a) K. Soai, S. Yokoyama, K. Moshida, Synthesis 1987, 647–648; b) W. C.
Guida, E. E. Entreken, A. R. Guida, J. Org. Chem. 1984, 49, 3024–3026.
[27] When the diene was metallated at 08C and the temperature was raised
to 208C before addition of BnBr, 38% of 5 and 37% of the correspond-
ing cyclohexa-1,3-diene isomer were formed.
[28] Interestingly, we observed in certain cases that LDA-mediated hydroa-
mination of 9a led to a mixture of 12c and 12a, depending on the
source of nBuLi used to prepare LDA. Green-blue coloration of the reac-
tion mixture was observed instead of the orange-red coloration de-
scribed above for the formation of homoallylic amines. A radical pro-
cess involving an aminyl radical issued from the oxidation of the lithi-
um amide might explain this observation. For precedent, see: M. New-
comb, M. T. Burchill, T. M. Deeb, J. Am. Chem. Soc. 1988, 110, 6528–
6535.
[43] a) C. V. De Amicis, P. R. Graupner, J. A. Erickson, J. W. Paschal, H. A. Kirst,
L. C. Creemer, P. E. Fanwick, J. Org. Chem. 2001, 66, 8431–8435; b) A. H.
Hoveyda, D. A. Evans, G. C. Fu, Chem. Rev. 1993, 93, 1307–1370.
[44] a) N. Iranpoor, H. Firouzabadi, M. Chitsazi, A. Ali Jafari, Tetrahedron 2002,
58, 7037–7042; b) A. Garcꢉa-Granados, P. E. Lopez, E. Melguizo, J. N.
Moliz, A. Parra, Y. Simeo, J. A. Dobado, J. Org. Chem. 2003, 68, 4833–
4844.
[45] S.-H. Yu, S.-K. Chung, Tetrahedron: Asymmetry 2005, 16, 2729–2747.
[46] A. Fꢃrst, P. A. Plattner, Helv. Chim. Acta 1949, 32, 275–283.
Received: June 2, 2014
Revised: July 10, 2014
Published online on September 15, 2014
Chem. Eur. J. 2014, 20, 14771 – 14782
www.chemeurj.org
14782
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim