Zn-Catalyzed Enantio- and Diastereoselective Formal [4 + 2] Cycloaddition Involving Two Electron-Deficient Partners: Asymmetric Synthesis of Piperidines from 1-Azadienes and Nitro-Alkenes

Article abstract of DOI:10.1021/jacs.5b00033

We report a catalytic asymmetric synthesis of piperidines through [4 + 2] cycloaddition of 1-azadienes and nitro-alkenes. The reaction uses earth abundant Zn as catalyst and is highly diastereo- and regioselective. A novel BOPA ligand (F-BOPA) confers high reactivity and enantioselectivity in the process. The presence of ortho substitution on the arenes adjacent to the bis(oxazolines) was found to be particularly impactful, due to limiting the undesired coordination of 1-azadiene to the Lewis acid and thus allowing the reaction to be carried out at lower temperature. A series of secondary kinetic isotope effect studies using a range of ligands implicates a stepwise mechanism for the transformation, involving an initial Michael-type addition of the imine to the nitro-alkene followed by a cyclization event. The stepwise mechanism obviates the electronic requirement inherent to a concerted mechanism, explaining the successful cycloaddition between two electron-deficient partners. (Chemical Equation Presented).

Full text of DOI:10.1021/jacs.5b00033

Article
pubs.acs.org/JACS
Zn-Catalyzed Enantio- and Diastereoselective Formal [4 + 2]
Cycloaddition Involving Two Electron-Deficient Partners: Asymmetric
Synthesis of Piperidines from 1‑Azadienes and Nitro-Alkenes
John C. K. Chu, Derek M. Dalton,† and Tomislav Rovis*
Department of Chemistry, Colorado State University, Ft. Collins, Colorado 80523, United States
S
* Supporting Information
ABSTRACT: We report a catalytic asymmetric synthesis of piperidines through
[4 + 2] cycloaddition of 1-azadienes and nitro-alkenes. The reaction uses earth
abundant Zn as catalyst and is highly diastereo- and regioselective. A novel BOPA
ligand (F-BOPA) confers high reactivity and enantioselectivity in the process. The
presence of ortho substitution on the arenes adjacent to the bis(oxazolines) was
found to be particularly impactful, due to limiting the undesired coordination of 1-
azadiene to the Lewis acid and thus allowing the reaction to be carried out at lower
temperature. A series of secondary kinetic isotope effect studies using a range of
ligands implicates a stepwise mechanism for the transformation, involving an initial
Michael-type addition of the imine to the nitro-alkene followed by a cyclization
event. The stepwise mechanism obviates the electronic requirement inherent to a concerted mechanism, explaining the successful
cycloaddition between two electron-deficient partners.
INTRODUCTION
■
[4 + 2] Cycloadditions have shown extensive applications for
the construction of cyclohexenes in the syntheses of simple
biologically active molecules1 or complex natural products.2
They offer the advantages of a high level of stereochemical
control and atom-economy and lead to a rapid escalation of
molecular complexity. Unsurprisingly, tremendous attention
has been paid to explore catalytic and enantioselective [4 + 2]
cycloadditions of dienes and alkenes.3 Despite the prevalence of
piperidine derivatives in drug molecules (Figure 1), the aza-
analogue of the all-carbon [4 + 2] cycloaddition is less well
developed.4
Simple imines, 1-azadienes, and 2-azadienes5 are three
different classes of substrates. The use of 2-azadienes in
catalytic, enantioselective [4 + 2] cycloaddition is mainly for the
synthesis for polycyclic nitrogen heterocycles, as exemplified by
the Povarov reaction.6 Simple imines and 1-azadienes are more
challenging substrates, although they can be readily prepared by
the condensation of amines on aldehydes.7,8 This may be
attributed to the lack of a thermodynamic driving force since
Figure 1. Examples of pharmaceuticals containing 3-aminopiperidines
the breaking C−N π bonds are stronger and the forming C−N
and a consideration of bond enthalpies.
σ bonds are weaker compared to the corresponding C−C
bonds (Figure 1).9
mates catalyzed by chiral phosphoric acid (Scheme 1, eq 2).13
Catalytic, enantioselective [4 + 2] cycloadditions with simple
Various groups have also reported the cycloaddition of N-
imines for the syntheses of monocyclic piperidines are mainly
restricted to the use of Danishefsky’s diene.10 There are limited
sulfonyl-1-azadienes with highly electron-rich alkenes generated
from the activation of aldehydes with organocatalysts, such as
examples of catalytic asymmetric [4 + 2] cycloadditions with 1-
NHCs,14 secondary amines,15,16 thioureas,17 and isothioureas.18
It is noteworthy that all these examples involve electron-rich
azadienes.11 In 2007, Arrayas and Carretero reported a Ni-
catalyzed enantioselective inverse-electron-demand Diels−
Alder reaction with N-sulfonyl-1-azadienes and enol ethers
(Scheme 1, eq 1).12 More recently, Masson disclosed an
asymmetric cycloaddition of N-aryl-1-azadienes and enecarba-
Received: January 7, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b00033
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 1. 1-Azadiene [4 + 2] Cycloadditions
RESULTS AND DISCUSSION
Reaction Discovery. As a result of a high-throughput
screen of first-row metal and Lewis acid catalysts, we identified
that Zn(II) salts enable a [4 + 2] cycloaddition of 1-azadiene 16
and nitro-alkene 17 to form tetrahydropyridine 18 (Table 1, eq
■
Table 1. ZnI2-Catalyzed [4 + 2] Cycloaddition of 1-
Azadienes and Nitro-Alkenes To Form Tetrahydropyridines
a
entry
equiv 17
T (°C)
conc (M)
additive
yield (%)
1
2
3
4
5
6
7
8
2
50
50
50
50
50
50
50
80
0.05
0.1
0.3
0.1
0.1
0.1
0.1
0.1
−
−
−
12
15
0
2
2
2
MgSO4
4 Å MS
4 Å MS
4 Å MS
4 Å MS
18
28
2
b
1.5
1
62 (78)
30
27
1.5
a
Yields determined by 1H NMR with 1,3,5-trimethoxybenzene as
b
internal standard. Reaction conducted in dioxane.
6).23 ZnI2 is the most effective catalyst identified. Remarkably,
only a single diastereomer and regioisomer of the product is
obtained. The reaction does not proceed without ZnI2. At high
reaction concentration (0.3 M), no product is observed (entries
1−3). As the enamine functionality is prone to hydrolysis, we
investigated the use of drying agent. The use of molecular
sieves gives an improved yield of the cycloadduct relative to
MgSO4 (entries 4, 5). Surprisingly, the cycloaddition is highly
sensitive to the number of equiv of 17. A lower amount of 17
(1.5 equiv) leads to an improved yield relative to 2 equiv
(entries 5, 6). With 2 equiv of 17, we observe the formation of
a 2:1 adduct (17:18).24 The formation of this undesired
product explains the decreased yield at higher concentration.
With an even lower equiv of 17, the reaction does not go to
completion and leads to decreased yield (entry 7). At higher
temperature, the yield drops significantly presumably due to
further reaction of 18 with 17 (entry 8). Dioxane is found to be
a better solvent than DME and provides the cycloadduct in
good yield (entry 6).
Initial Screen of BOPA Ligands. An initial screen of
commonly available ligands for Zn(II) resulted in low
enantioselectivity.25 Inspired by the work on Zn-catalyzed
enantioselective Michael addition to nitro-styrenes by Du and
co-workers, we explored the use of bis(oxazolinylphenyl)amide
(BOPA ligand)26 for rendering the cycloaddition asymmetric.
We were delighted to find that BOPA A gives promising
enantioselectivity under the reaction conditions reported by
Du, although the yield is low (Scheme 2).27,28 Ligand B, with a
cis-relationship of the phenyl groups, or ligand C, with phenyl
groups close to the oxazoline nitrogen, only provides a minor
improvement in yield but not enantioselectivity. BOPAs D and
E, which have significantly different steric environments, give
similar yields and ee’s. Gratifyingly, a much higher yield can be
obtained with F, and its improved reactivity allows the reaction
to be conducted at ambient temperature. Decreasing the
temperature increases enantioselectivity with no loss in yield.
alkenes. Considering the relative electron deficiency of 1-
azadienes, it can be expected that use of electron-poor alkenes
as the coupling partner would be challenging if the reaction
proceeds by a concerted mechanism.19
Current protocols for asymmetric [4 + 2] cycloadditions with
1-azadienes and electron-deficient olefins require chiral
auxiliaries. Ghosez demonstrated that the cycloaddition of
chiral hydrazones derived from enals with electron-deficient
alkenes proceeds with excellent diasteroselectivity (Scheme 1,
eq 3).20 Barluenga reported that N-aryl-1-azadiene undergoes a
highly diastereoselective [4 + 2] cycloaddition with a chiral
tungsten alkynyl(alkoxy)carbene complex (Scheme 1, eq 4).21
To the best of our knowledge, catalytic, asymmetric [4 + 2]
cycloadditions of 1-azadienes and electron-poor alkenes have
not been reported. Herein, we disclose the development of a
Zn-catalyzed regio-, diastereo-, and enantioselective [4 + 2]
cycloaddition of readily available 1-azadienes and nitro-alkenes
(Scheme 1, eq 5).22
B
DOI: 10.1021/jacs.5b00033
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 2. Investigation of BOPA Ligands
Scheme 4. Use of Spectator Imine and Mono-Oxazoline
symmetry of the BOPA ligand (mono-oxazoline H) gives a
significant erosion of ee (Scheme 4, eq 9). Both oxazolines are
necessary to achieve high levels of enantio-induction, and the
selectivity is sensitive to the addition of strongly ligating imines.
Significance of BOPA N−H Bond. Eager to improve the
enantioselectivity beyond what was possible with modification
of 1-azadiene, we sought to understand how the structure of
BOPA affects enantioinduction. Modification of the nitrogen
bridging the aryls of BOPA had a surprising effect on both yield
and enantioselectivity (Scheme 5). Compared to the parent
Efforts to further enhance the enantioselectivity by lowering the
temperature are unsuccessful since only a trace amount of
product can be obtained at 0 °C. Modification of F by
installation of a benzyloxy group (ligand G) does not improve
yield and gives a slightly lower ee.
Electronic Effects of 1-Azadienes. With the hope that
substrate design would improve enantioselectivity, we altered
N-substituent electronics of the 1-azadienes (Scheme 3). No
Scheme 3. Modification of 1-Azadiene Electronics
Scheme 5. Significance of N−H Bonds in BOPA Ligands
BOPA F, the sulfur analogue I gives a significant loss of
enantioselectivity. Oxygen analogue J, despite having a similar
C−X bond and C−X−C bond angle, favors the formation of
the opposite enantiomer with low selectivity. Clearly, the N−H
is crucial for high levels of selectivity. To assess whether the
source of the enantioselectivity was the N or the H of the N−H
bond, N−Me analogue K was synthesized and revealed to
provide 5% ee, slightly favoring the opposite enantiomer. The
importance of the N−H bond in achieving high ee is consistent
with previously reported Zn·BOPA-catalyzed asymmetric
Michael additions with nitro-alkenes.29
Electronic Effects of Ligands. Given the impact of the
BOPA N−H bond on selectivity, we decided to study the effect
of acidity of the N−H bond on enantioselectivity. Astonish-
ingly, there seems to be no clear correlation of N−H acidity
with ee (Scheme 6). Both ligands L and M, which bear less
correlation between the electronics of the 1-azadienes and yield
is seen. However, we do observe that ee increases as N-
substituent electron density increases (21, 22, 23, and 20).
Unfortunately, addition of an electron-donating morpholine to
the N-aryl substituent (24) provides no further increase in
enantioselectivity.
Speculating that the presence of a more basic, more strongly
coordinating nitrogen found in the morpholinyl azadiene was
responsible for disappointing enantioselectivity found in 24, we
used spectator imine 25 to probe the effect that a competitive
coordinating group had on ee. We find a small erosion of ee
with the addition of imine 25 as a competitive coordinating
group (Scheme 4, eq 8). Furthermore, removal of the C2
C
DOI: 10.1021/jacs.5b00033
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 6. Electronic Effects of Ligands on ee
Scheme 7. Disappearing Electronic Effect with ortho-
Substitution
acidic and more acidic N−H bonds respectively, lead to
decreased ee. Therefore, we reason that the electronic effect of
the ligands is due to the change in electron density of the
oxazolines, instead of the N−H bonds. Indeed, the use of
ligands N and O, despite having a significant difference in the
acidity of the N−H bonds, gives a similar level of stereo-
chemical control with a reversal of enantioselectivity.
lines of Q are too electron deficient to coordinate zinc
effectively. The reaction with fluorinated ligand R gives the
same level of enantioselectivity as ligand F but with significantly
higher yield; surprisingly, the enantiomer obtained is not the
one predicted by the Hammett plot. Electronically different
ligands S, T, and U give various yields but the same ee’s,
suggesting that the bisoxazoline electronic effect on enantio-
selectivity disappears with ortho substitution.33
Mechanistic Discussion. The cycloaddition can proceed
by a concerted (Diels−Alder reaction, most likely asynchro-
nous) or stepwise (Michael addition followed by cyclization)
mechanism. We believe that the reaction occurs by a stepwise
mechanism, although we cannot discount a concerted reaction
mechanism.
A Hammett plot31 substantiates our proposal that the
electronics of the bisoxazolines, not the N−H bonds, govern
the stereoselectivity since it reveals a linear relation between log
(k1/k2) and the σ value of the substituents on the bisoxazolines
(Figure 2), where k1/k2 is relative rate for the formation of one
First, to probe if a concerted mechanism is possible, we
subjected 1,4-diphenylbutadiene, possessing a higher HOMO
than the 1-azadiene, to the reaction conditions and observed no
cycloadduct (Scheme 8, eq 10). Next, thiourea 28 is known to
activate nitro-olefins for the aza-Michael addition of imines
(Scheme 8, eq 11).34 Indeed, thiourea 28 also induces the [4 +
2] cycloaddition, although the yield and ee are low (Scheme 8,
eq 12), implicating that a stepwise mechanism is feasible. In
Scheme 8. Mechanistic Studies
Figure 2. Hammett plot indicating linear relation between electronics
of bisoxazolines and enantioselectivity.30
enantiomer to the other and is thus the enantiomeric ratio. The
implication of the linear relation is that a higher enantiose-
lectivity could be achieved with P, which bears seemingly
electron-donating OMe groups in the p-positions. Disappoint-
ingly, P gives lower selectivity than H-BOPA F and contradicts
the trend observed (Scheme 6). This irregularity can be
explained by the perpendicularity of the oxazolines to the
aromatic backbone for all ligands studied,32 rendering the p-
OMe group electron-withdrawing and thus similar to the m-
OMe substituent.
Impact of ortho-Substitution on BOPA. Since our efforts
to improve the enantioselectivity by rendering the bisoxazoline
more electron rich were unsuccessful, we attempted to make
the bisoxazoline highly electron deficient by installing strongly
electron-withdrawing groups at the o-position of the oxazolines
to favor the formation of the opposite enantiomer (Scheme 7).
The use of BOPA Q, bearing nitro groups, leads to a much
lower yield and no enantioselectivity. Presumably the bisoxazo-
D
DOI: 10.1021/jacs.5b00033
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
addition, we treated imine 29 with nitro-alkene 19 in the
presence of the Zn·F-BOPA catalyst and found that the aza-
Michael adduct is formed in excellent yield (Scheme 8, eq 13).
As a final test, we carried out kinetic studies with deuterated
nitro-alkene 19 and 1-azadiene 32.35 We found that an inverse
secondary kinetic isotope effect36 was observed with 19
(Scheme 9, eq 14) but not with 32 (Scheme 9, eq 15).
These results are consistent with a stepwise mechanism.
the Michael addition is responsible for controlling the degree of
enantioselectivity observed with 19.
Since the electronic effect on enantioselectivity is not present
with previously reported asymmetric Michael addition of indole
to nitro-alkenes,39 a change in coordination geometry of Zn
with different ligands is unlikely. On the other hand, the
substrate 1-azadiene 16 could potentially coordinate to Zn, as
suggested by the decreased ee in the presence of spectator
imine 25 (Scheme 4, eq 8). Therefore, we hypothesize that as
the bisoxazoline becomes less electron rich, 1-azadiene 16
becomes more competent to alter Zn coordination environ-
ment, leading to the observed electronic effect on enantiose-
lectivity. In the presence of 16, a more significant loss in yield
and ee for O compared to F is consistent with this hypothesis
(Scheme 11). Compared to O, the yield with R is partially
Scheme 9. Secondary Kinetic Isotope Effect
Scheme 11. Effect of Imine 16 as an Additive on Asymmetric
a
Michael Addition
Origin of the Electronic Effect on Enantioselectivity.
The change in the acidity of N−H bond is unlikely the origin of
the electronic effect on enantioselectivity since ligands N and O
give a similar level of stereochemical control.37 As the ligand
becomes more electron deficient, the Michael addition becomes
faster and the cyclization becomes slower due to stronger Lewis
acidity of Zn and more stabilized Zn nitronate, respectively.
Therefore, we also considered a possible switch in enantio-
determining step upon changing ligands. However, the
presence of inverse secondary kinetic isotope effect with 19
but not 32 for all ligands suggests that the first step of Michael-
addition is enantio-determining in all cases (Scheme 9, eqs 14
and 15). Further evidence that the establishment of the
absolute stereochemistry occurs at the Michael addition is
provided by the reaction with methyl nitro-ethylene 34
(Scheme 10, eq 16).38 Since no stereogenic center is generated
in Michael addition, we would expect a different major
enantiomer with ligand F and O if the second cyclization
step controlled the enantioinduction. In the event, low
enantioselectivity is seen, and the formation of the same
major enantiomer with ligands F and O from 34 suggests that
a
Yields and ee’s obtained in absence of 16 in parentheses.
rescued presumably because the ortho substitution disfavors the
coordination of imine 16 to Zn. The variation in yield and ee is
caused by imine 16 since ligands F, O, and R induce the same
level of efficiency and selectivity without imine 16.40
Optimization Studies with F-BOPA (Ligand R).
Consistent with our proposal that the ortho substitution in R
disfavors undesired coordination of 1-azadiene to Zn, the
cycloaddition with ligand R is more efficient than O, allowing
the [4 + 2] cycloaddition to be conducted at lower
temperatures (Table 2, eq 17). Since 1-azadiene 32 is more
soluble in PhMe than 16, it was used in the optimization. While
the temperature has no effect on the yield (>90% NMR yields
with 1,3,5-trimethoxybenzene as internal standard in all cases),
it does affect reaction rate and enantioselectivity in that rate
decreases and ee increases with decreasing temperature (Table
2, entries 1−4). We were excited to find that the cycloaddition
affords piperidine 23 in excellent yield and enantioselectivity41
at −38 °C, albeit with a long reaction time (entry 4). Increasing
catalyst loading (entry 5), reaction concentration (entry 6), and
the equivalents of nitro-alkene (entry 7) can be used to
decrease reaction time. With 20 mol % of catalyst loading, 3
equiv of nitro-alkene42 and a concentration of 0.18 M, reaction
time can be effectively shortened to 24 h without diminishing
yield or ee.
Scheme 10. Same Enantiomer from 34 with F and O
1-Azadiene Scope. With the optimized conditions in hand,
we investigated the substrate scope of the reaction (Scheme
12). Regardless of the electronics of the aromatic ring on
E
DOI: 10.1021/jacs.5b00033
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Compared to 23 bearing 4-phenoxyphenyl, both more electron-
rich 20 and less electron-rich 22 are obtained with slightly
lower ee’s. This is consistent with our proposal that highly
electron-rich 1-azadienes competitively coordinate to zinc and
lead to an erosion of stereochemical control. It is noteworthy
that 16 is not electron rich enough to cause a dramatic decrease
in enantioselectivity at rt. This is not unreasonable because
coordination of azadienes to Zn entails a higher entropic
penalty at higher temperature. Different protecting groups on
nitrogen are used in subsequent studies, and the one giving a
higher solubility of the 1-azadiene is chosen in each case.
The electronic properties of the aryl ring at the 4-position of
the 1-azadienes (39−41) also have no apparent effect on yields
and ee’s remain high in all cases. The enantioselectivity is
higher with more electron-poor aryl groups, following the trend
predicted by the potential coordination of 1-azadienes to zinc.
Ortho substitution (42) is tolerated, although the reaction
affords the cycloadducts in slightly lower yield and
enantioselectivity. On the other hand, the reaction yielding
furan-bearing 43 is sluggish and gives moderate enantioselec-
tivity. Aliphatic enal-derived azadiene is not a competent
substrate and gives the cycloadduct (44) in modest yield and
ee. The reaction of more conjugated azatriene does not go to
completion even in the presence of a large excess of the nitro-
alkene and more catalyst and gives the cycloadduct (45) in
moderate yield and enantioselectivity.
Table 2. Optimization Studies with Ligand R
entry
x
T (°C)
conc (M)
t (h)
ee (%)
1
2
3
4
5
6
7
10
10
10
10
20
20
20
rt
3
0.09
0.09
0.09
0.09
0.09
0.18
0.18
<24
<24
48
61
73
78
89
89
89
−18
−38
−38
−38
−38
120
72
36
a
24
89
a
3 equiv of 19, 82% isolated yield.
a
Scheme 12. Scope of 1-Azadienes
The [4 + 2] cycloaddition of chalcone-derived azadiene
requires a higher catalyst loading and the use of a large excess of
nitro-alkenes and gives cycloadduct 46 in excellent yield but
moderate enantioselectivity. When the reaction is carried out at
ambient temperature, no enantioselectivity is observed.43
Nitro-Alkene Scope. Compared to 19, the [4 + 2]
cycloadditions of sterically more-demanding nitro-alkenes
require either a longer reaction time or the use of more
nitro-alkene and afford the cycloadducts (47 and 48) in
synthetically useful yield and slightly decreased enantioselec-
tivity (Scheme 13). An increased catalyst loading is needed for
the sterically demanding isopropyl nitro-alkene, and a more
drastic decrease in selectivity is observed (49). The silyl-
protected alcohol (50) is tolerated and provides a handle in the
cycloadduct for further functionalization. Nitro-ethylene shows
reactivity under the reaction conditions, but the cycloadduct
(51) is difficult to isolate due to its instability. α-Methyl nitro-
ethylene also reacts, albeit with moderate yield and
enantioselectivity (52). More conjugated nitro-alkenes are
inert under the reaction conditions and afford no cycloadduct
(53 and 54), although a higher electrophilicity is expected from
the extended conjugation.
Utility of the Cycloadducts. The PMP or 4-phenox-
yphenyl (POP) group of the cycloadducts can be easily
removed by CAN to afford free secondary amine 55 (Scheme
14, eq 18). The medicinally significant structural motif 3-
aminopiperidine can be obtained by the reduction of the nitro
group with nickel boride44 (Scheme 14, eq 19).
Stereochemical Model. On the basis of our mechanistic
investigations noted above, we propose that the reaction
proceeds in a stepwise fashion (Figure 3). The aza-diene adds
to the Zn-coordinated nitro-alkene in Michael fashion, likely via
its s-trans conformer. The relative stereochemistry is confirmed
by X-ray analysis. A bond rotation then is required for the
nitronate to cyclize onto the α,β-unsaturated iminium. This
may proceed in clockwise fashion placing the phenyl group and
the nitro on the same side of the piperidine ring (blue path) or
a
b
All reported yields are isolated unless otherwise stated. Reaction
c
d
conducted at 0.09 M. 30 mol % catalyst, 5 equiv of 19. 30 mol %
catalyst, 6 equiv of 19. rt, 10 mol % catalyst, 1.5 equiv of 19, NMR
yield with 1,3,5-trimethoxybenzene as internal standard.
e
nitrogen of the 1-azadienes, excellent isolated yields of the
cycloadducts are obtained. The reaction time does not reflect
the intrinsic reactivity of the 1-azadienes since they have
different solubility at low temperature, affecting reaction rate.
F
DOI: 10.1021/jacs.5b00033
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
intermediate and leaves the butyl group in a pseudoaxial
orientation on the piperidinyl half-chair. The red path results in
a conformational switch placing the butyl substituent in a
pseudoequatorial position and is disfavored by reason for A1,3-
strain.
Scheme 13. Nitroalkene Scope
The electrostatic interaction between the nitronate and the
positively charged iminium may also play a significant role since
we observe the same levels of diastereoselectivity when nitro-
ethylene is used as a substrate (Scheme 13, 51). In addition, the
cis relationship cannot be explained by the sterics of the nitro
group, as, in the case of the cycloaddition with α methyl nitro-
ethylene, the phenyl group is also cis to the nitro group (52).
Finally, the ligand environment does not induce a cis
relationship because this orientation also occurs with ZnI2 in
the absence of a ligand. It is noteworthy that no matter which
direction this bond rotates, an anti relationship between the
butyl group and the nitro group is expected.
CONCLUSION
■
We developed a catalytic enantioselective [4 + 2] cycloaddition
of readily prepared 1-azadienes and nitro-alkenes. The
cycloaddition requires inexpensive earth-abundant Zn as
catalyst and is highly regio- and diastereoselective. Mechanistic
evidence suggests that the reaction proceeds with a stepwise
mechanism. An electronic effect of the bisoxazoline ligand on
the enantioselectivity is observed and can be explained by the
deleterious coordination of 1-azadiene to Zn. The key to
achieving high enantioselecitivity of the reaction is the use of a
novel fluorinated BOPA ligand, whose ortho substitution
disfavors the undesired coordination of 1-azadienes to Zn.
The utility of the cycloadducts is demonstrated by the easy
reduction of the nitro group to give medicinally interesting 3-
aminopiperidines.
a
b
c
See footnote a, Scheme 12. 5 equiv nitro-alkene. 3- mol % catalyst,
d
6 equiv nitro-alkene. NMR yield, no reductive workup, ee not
determined. rt, 10 mol % catalyst, 61% ee at −38 °C. rt.
e
f
Scheme 14. Utility of cycloadducts
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental details and characterization of all new compounds
and ligands (G, H, J−U). This material is free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*rovis@lamar.colostate.edu
■
Present Address
†Department of Chemistry, University of California, Berkeley,
California 94720, United States.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank NIGMS (GM80442) for support. J.C.K.C. thanks the
Croucher Foundation, and D.M.D. thanks the NIH Ruth
Kirschstein Predoctoral Fellowship for support.
REFERENCES
■
(1) Funel, J.-A.; Abele, S. Angew. Chem., Int. Ed. 2013, 52, 3822−
3863.
Figure 3. Proposed stepwise mechanism and diastereoselectivity
model (Lewis acid not shown for clarity).
(2) Nicolaou, K. C.; Snyder, S. A.; Montagnon, T.;
Vassilikogiannakis, G. Angew. Chem., Int. Ed. 2002, 41, 1668−1698.
(3) For selected reviews: (a) Kagan, H. B.; Riant, O. Chem. Rev.
1992, 92, 1007−1019. (b) Corey, E. J. Angew. Chem., Int. Ed. 2002, 41,
1650−1667. (c) Merino, P.; Marques-Lopez, E.; Tejero, T.; Herrera,
by counterclockwise rotation placing the substituents opposite.
The blue path proceeds by minimalist reorganization of the
G
DOI: 10.1021/jacs.5b00033
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
R. P. Synthesis 2010, 1−26. (d) Jiang, X.; Wang, R. Chem. Rev. 2013,
113, 5515−5546.
(30) The ee values for the Hammett plot were obtained from the
average values of three experiments with each ligand.
(31) For a review on Hammett plot, see: Hansch, C.; Leo, A.; Taft, R.
W. Chem. Rev. 1991, 91, 165−195.
(4) For a review on catalytic enantioselective aza-[4 + 2]
cycloadditions, see: Masson, G.; Lalli, C.; Benohoud, M.; Dagousset,
G. Chem. Soc. Rev. 2013, 42, 902−923.
(32) The perpendicularity of the bisoxazoline to the aromatic
backbone is supported by the crystal structure of the sulfur analogue
(S, instead of NH, as the linker) of a BOPA ligand, see: Kooijman, H.;
Spek, A. L.; Zondervan, C.; Feringa, B. L. Acta Crystallogr., Sect. E:
Struct. Rep. Online 2002, 58 (8), m429−431. In the crystal structure of
a Ca(BOPA)2 complex, the bisoxazoline and the aromatic backbone is
only slightly out of plane, see: Nixon, T. D.; Ward, B. D. Chem.
Commun. 2012, 48, 11790−11792. However, the N−H in the BOPA
ligand in this case is deprotonated by a strong base and might require
the bisoxazoline to be in plane to stabilize the negative charge on the
N.
(5) For a review on the use of azadienes in [4 + 2] cycloaddition, see:
Boger, D. L. Tetrahedron 1983, 39, 2869−2939.
(6) For selected reviews: (a) Kouznetsov, V. V. Tetrahedron 2009, 65,
2721−2750. (b) Fochi, M.; Caruana, L.; Bernardi, L. Synthesis 2014,
135−157.
(7) For a review on the use of 1-azadienes in [4 + 2] cycloadditions,
see: Behforouz, M.; Ahmadian, M. Tetrahedron 2000, 56, 5259−5288.
(8) Daniels, P. H.; Wong, J. L.; Atwood, J. L.; Canada, L. G.; Rogers,
R. D. J. Org. Chem. 1980, 45, 435−440.
(9) Jung, M. E.; Shapiro, J. J. J. Am. Chem. Soc. 1980, 102, 7862−
7866.
(10) For a review on the use of simple imines in [4 + 2]
cycloadditions: see Jorgensen, K. A. Angew. Chem., Int. Ed. 2000, 39,
3558−3588.
(11) Chiral auxiliaries may also be used for the reaction of this class;
see: Clark, R. C.; Pfeiffer, S. S.; Boger, D. L. J. Am. Chem. Soc. 2006,
128, 2587−2593.
(12) Esquivias, J.; Arrayas, R. G.; Carretero, J. C. J. Am. Chem. Soc.
2007, 129, 1480−1481.
(13) He, L.; Laurent, G.; Retailleau, P.; Folleas, B.; Brayer, J.-L.;
Masson, G. Angew. Chem., Int. Ed. 2013, 52, 11088−11091.
(14) (a) He, M.; Struble, J. R.; Bode, J. W. J. Am. Chem. Soc. 2006,
128, 8418−8420. (b) Zhao, X.; Ruhl, K. E.; Rovis, T. Angew. Chem.,
Int. Ed. 2012, 51, 12330−12333. (c) Jian, T.-Y.; Shao, P.-L.; Ye, S.
Chem. Commun. 2011, 47, 2381−2383.
(15) Han, B.; He, Z.-Q.; Li, J.-L.; Li, R.; Jiang, K.; Liu, T.-Y.; Chen,
Y.-C. Angew. Chem., Int. Ed. 2009, 48, 5474−5477.
(16) Li, J.-L.; Liu, T.-Y.; Chen, Y.-C. Acc. Chem. Res. 2012, 45, 1491−
1500.
(33) There is likely a steric and electronic component to the impact
of the ortho substituent given that H, Cl, MeO, and Me provide nearly
identical enantioselectivities (68−71% ee).
(34) Lykke, L.; Monge, D.; Nielsen, M.; Jørgensen, K. A. Chem.
Eur. J. 2010, 16, 13330−13334.
(35) Since azadiene 16 has a poor solubility in PhMe, 32 was used for
the kinetic isotope effect studies. Azadiene 32 behaves similarly in all
respects to azadiene 16.
(36) Van Sickle, D. E.; Rodin, J. O. J. Am. Chem. Soc. 1964, 86,
3091−3094. (b) Taagepera, M.; Thornton, E. R. J. Am. Chem. Soc.
1972, 94, 1168−1177. (c) Gajewski, J. J.; Peterson, K. B.; Kagel, J. R.;
Huang, Y. C. J. J. Am. Chem. Soc. 1989, 111, 9078−9081.
(37) The pKa’s of 4-methoxyaniline and 3-methoxyaniline differ by
more than 1 unit.
(38) The reactions at rt gave no enantioselectivity.
(39) Liu, H.; Li, W.; Du, D.-M. Sci. China, Ser. B: Chem. 2009, 52,
1321−1330.
(40) This observation is consistent with ref 39.
(41) For the determination of absolute configuration, see Supporting
Information.
(17) Jiang, X.; Shi, X.; Wang, S.; Sun, T.; Cao, Y.; Wang, R. Angew.
(42) Unlike the situation when ZnI2 is used as catalyst in absence of
BOPA (see Table 1), stirring azadiene 16 with 3 equiv of nitro-alkene
17 with 10 mol % Zn(OTf)2 and F-BOPA ligand at rt for 16 h did not
lead to the formation of the 2:1 adduct as judged by MS.
(43) There is no reaction at rt without ZnII, so an uncatalyzed
pathway can be ruled out. The reaction is also irreversible at rt since
no nitro-alkene exchange takes place when the cycloadduct (no
reductive workup) is subject to the reaction conditions with another
nitro-alkene.
Chem., Int. Ed. 2012, 51, 2084−2087.
(18) Simal, C.; Lebl, T.; Slawin, A. M. Z.; Smith, A. D. Angew. Chem.,
Int. Ed. 2012, 51, 3653−3657.
(19) Houk, K. N. Acc. Chem. Res. 1975, 8, 361−369.
(20) Beaudegnies, R.; Ghosez, L. Tetrahedron: Asymmetry 1994, 5,
557−560.
(21) Barluenga, J.; de la Rua, R. B.; de Saa, D.; Ballesteros, A.; Tomas
́
,
M. Angew. Chem., Int. Ed. 2005, 44, 4981−4983.
(22) For an elegant catalytic enantioselective synthesis of 3-
nitropiperidines with a different substitution pattern, see: Lin, H.;
Tan, Y.; Liu, W.-J.; Zhang, Z.-C.; Sun, X.-W.; Lin, G.-Q. Chem.
Commun. 2013, 49, 4024−4026.
(44) Nose, A.; Kudo, T. Chem. Pharm. Bull. 1981, 29, 1159−1161.
(23) For results of the high-throughput screen, please see Supporting
Information.
(24) The 2:1 adduct is observed by 1H NMR of the unpurified
reaction mixture and by MS but has, to date, resisted isolation and
characterization.
(25) For example, other bisoxazoline ligands such as BOX or PYBOX
gives low enantioselectivity. See Supporting Information for details.
(26) BOPA ligands have been used in Zn-catalyzed enantioselective
Michael addition to nitro-alkenes; see: Lu, S. F.; Du, D.-M.; Xu, J. Org.
Lett. 2006, 8, 2115−2118.
(27) Due to the sensitive nature of the tetrahydropyridine, we sought
to adopt a reductive workup procedure to reduce the enamine
functionality and found that the crude tetrahydropyridine can be
reduced to stable 3-nitropiperidine with ZnCl2 and NaBH3CN in
MeOH.
(28) In constrast to the racemic reaction, ZnI2 does not give any
product. Although DME and dioxane are the optimal solvents for the
racemic reaction, their use gives lower ee than PhMe. See Supporting
Information for the results of solvent screen.
(29) Liu, H.; Lu, S.-F.; Xu, J.; Du, D.-M. Chem.Asian J. 2008, 3,
1111−1121.
H
DOI: 10.1021/jacs.5b00033
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX