A highly enantioselective zirconium catalyst for intramolecular alkene hydroamination: Significant isotope effects on rate and stereoselectivity

Article abstract of DOI:10.1002/anie.201006163

Zirconium catalysts sparkle: A new chiral zirconium complex has been used to catalyze hydroamination reactions to cyclize aminopentenes into 2-methylpyrrolidines (see scheme). The rate law supports a mechanism involving a reversible substrate-catalyst interaction that precedes the rate-determining step. A new mechanism for zirconium-catalyzed hydroamination has been proposed based on kinetic isotope effects and the significant effect of isotopic substitution on enantioselectivity.

Full text of DOI:10.1002/anie.201006163

DOI: 10.1002/anie.201006163
Asymmetric Hydroamination
A Highly Enantioselective Zirconium Catalyst for Intramolecular
Alkene Hydroamination: Significant Isotope Effects on Rate and
Stereoselectivity**
Kuntal Manna, Songchen Xu, and Aaron D. Sadow*
In memory of Victor S.-Y. Lin
Asymmetric olefin hydroamination/cyclization is a promising
method for the synthesis of optically active cyclic amines.[1] To
date, the most selective catalysts are C2-symmetric rare earth
and zirconium complexes.[2,3] These C2-symmetric catalysts
contrast with seminal C1-symmetric ansa-lanthanidocenes
that epimerize under catalytic conditions.[4] Despite advances,
current asymmetric hydroamination/cyclization catalysts are
highly sensitive to substrate substitution patterns.
This unusually high activity motivated the preparation of
optically active oxazolinylborate analogues that might pro-
vide highly reactive, robust, and non-epimerizable complexes.
Herein, we report a chiral catalyst that provides pyrrolidine
derivatives with excellent enantiopurity. Our data, including a
unique deuterium isotope effect on enantioselectivity, unam-
biguously rules out both pathways of Scheme 1.
Precatalyst preparation is outlined in Equations (1) and
Two distinct pathways have been proposed for C N bond
(2).[8] Deprotonation of the oxazoline substrate 2H-OxiPr,Me
2
ꢀ
formation in zirconium-catalyzed processes:[1b] olefin inser-
by nBuLi and subsequent treatment with 0.5 equivalents of
PhBCl2 provided chiral borane 2.[9] A crude sample of 2 was
then treated with NaC5H5 in THF and gave 3-H2 as a mixture
of C5H5 isomers; one isomer is shown in Equation (1).
[5]
ꢀ
tion into a M NR2 bond (Scheme 1, Mechanism A) and a-
abstraction followed by [2p + 2p] cycloaddition (Mecha-
nism B).[3,6] These two pathways can be difficult to distinguish
because of conflicting observations, including large
deuterium isotope effects and first- or second-order
rate laws.
Our research group has recently described an
achiral catalyst [{PhB(C5H4)(OxMe )2}Zr(NMe2)2]
2
(1; OxMe = 4,4-dimethyl-2-oxazolinyl) that cyclizes
2
aminoalkene compounds at room temperature.[7]
Reaction of [Zr(NMe2)4] and 3-H2 afforded 4 in excellent
yield of isolated product [Eq. (2)]. One n˜CN band at 1565 cmꢀ1
in the IR spectrum of 4 suggests that both oxazoline units are
coordinated to zirconium (for comparison, 2H-OxiPr,Me : n˜CN
=
2
1632 cmꢀ1).
ꢀ
Scheme 1. Mechanism A: olefin insertion into a M N bond; Mecha-
nism B: a-abstraction and [2p+2p] cycloaddition.
A catalytic amount of 4 (2–10 mol%) and primary
aminoalkenes 5a–9a rapidly yielded pyrrolidines 5b–9b
with enantiomeric excesses ranging from 89% to 98% (see
Table 1). Although the reaction rate is sensitive to catalyst
and substrate concentration, the enantioselectivity is not. The
catalyst can be recycled once without loss of activity or
enantioselectivity. The secondary aminoalkene 5a-NMe was
not cyclized by 4, even upon heating for 12 hours at 80, 120,
140, or 1708C. However, upon addition of nPrNH2, 20%
conversion into 5b was observed at room temperature.
Significantly, the ee values obtained with 4 as the catalyst
are the highest to date for unprotected pyrrolidines 5b, 6b,
[*] K. Manna, S. Xu, Prof. A. D. Sadow
Department of Chemistry and Ames Laboratory
Iowa State University, 1605 Gilman Hall, Ames, IA (USA)
Fax: (+1)515-294-0105
E-mail: sadow@iastate.edu
Homepage: http://www.public.iastate.edu/~sadow/
[**] Dr. A. Bakac is thanked for valuable discussions about kinetics and
Dr. M. Jeffries-EL is thanked for assistance with HPLC analysis. The
U.S. DOE Office of Basic Energy Science (DE-AC02CH11358) is
acknowledged for support of the mechanistic studies. A.D.S. is an
Alfred P. Sloan Research Fellow.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201006163.
Angew. Chem. Int. Ed. 2011, 50, 1865 –1868
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1865

Communications
Table 1: Cyclization of aminoalkenes catalyzed by 4.
Cyclization occurs upon addi-
Product[a]
Solvent
t [h]
Conversion [%]
ee [%][b]
tion of aminoalkenes to 4 without a
Substrate
detectable
induction
period;
HNMe2 and 5b–9b are observed
within approximately two minutes.
ln[6a] varies linearly with time for
5a
5b
5b
5b
5b
5b
5b
C6D6
C6D6
CD2Cl2
[D8]THF
[D8]THF
[D8]THF
1.25
6
5
5
12
>95, 93[c]
96
>95
>95
>95
>95
93
93
94
about two half-lives to provide kobs
.
5a[d]
5a
A linear relationship between kobs
and [4] provides the empirical rate
law ꢀd[6a]/dt = k’obs[4][6a], which
is consistent with reversible sub-
strate–catalyst association preced-
ing the turnover limiting step
(TLS). To test this hypothesis, initial
cyclization rates were measured
over a 9.8–154 mm concentration
range with respect to 6a ([4] =
5.4 mm, 238C). Indeed, the rate
increases with [6a] until saturation
is observed (Figure 1). At higher
concentrations of 6a, the rate
decreases slightly as a result of
inhibitory association of another
equivalent of substrate. The plot
also contains a nonzero x-intercept
that coincides with [4]. A nonlinear
least-squares regression analysis of
the data provides good correlation
with Equation (3), corresponding to
the reaction mechanism shown in
Equation (4) that describes the ini-
tial stage of the reaction.[11]
5a
95
5a[e]
5a[f]
96[g]
98[g]
5days
6a
6b
6b
6b
C6D6
C7D8
[D8]THF
1.25
6.5
11
>95, 88[c]
>95
93
90
90
94
6a[d]
6a[e]
C6D6
C6D6
4
7
88
89
92
89
5a-NMe[h]
5a-NMe
5b-NMe
5b-NMe
C6D6
C6D6 +nPrNH2
24
1
0
20
n.d.
n.d.
[i]
9a
9b
C6D6
C6D6
0.5
>95, d.r.=1.1:1
93, 92
n.d.
ꢀd½6aꢁ
k2ð½6aꢁꢀ½4ꢁÞ½catalyst Aꢁ
2
¼
10a[j]
10b
5
24
K0 þ ð½6aꢁꢀ½4ꢁÞ þ ð½6aꢁꢀ½4ꢁÞ KSI
ð3Þ
dt
The three parameters obtained
from the fit are k2 (1.7 ꢂ 0.3 ꢀ
10ꢀ2 mꢀ1 sꢀ1), K’ (2.8 ꢂ 0.7 ꢀ 10ꢀ2 m;
11a
12a
11b
12b
C6D6
C6D6
30
40
65
48
46
(kꢀ1 + k2)/k1), and KSI (5 ꢂ 2mꢀ1
;
[catalyst·6a2]/[6a][catalyst·6a]; sub-
strate inhibition). The curve does
not pass through the origin because
one equivalent of substrate is
required to form the active catalyst,
thus giving the terms ([6a]–[4]; i.e.,
corrected substrate concentration).
This analysis separates the TLS (k2)
from the substrate binding constant
31[g]
[a] Reaction conditions: 10 mol% of 4 and room temperature unless otherwise noted. [b] The ee values
(ꢂ0.5%) were determined by 1H and/or 19F NMR spectroscopy of Mosher amide derivatives.
Assignments of absolute configuration were based on literature reports.[2,3,8] [c] Yield of isolated
product. [d] 2 mol% of 4. [e] 08C. [f] ꢀ308C. [g] The ee values were determined by HPLC on a chiral
stationary phase. [h] 1708C. [i] 10–30 mol%. [j] 1108C. n.d.=not determined.
7b, and 9b.[10] Catalytic species derived from 1 and 4 are
uncommonly active below room temperature compared to
reported zirconium(IV) catalysts, including related
[(CGC)Zr(Me)Cl]
[(dpm)Zr(NMe2)2]
(CGC = Me2Si(C5Me4)(NtBu)),[5c]
(dpm = 5,5-dimethyldipyrrolylme-
thane),[5d] and [Cp*(salicyloxazolinato)Zr(NMe2)2] (Cp* =
pentamethylcyclopentadienyl) complexes.[6d] C1-symmetric 4
produces pyrrolidines at 08C and even at the low temperature
of ꢀ308C with improved ee values of up to 98%.
(K’), thus allowing the measurement of the isotope effect (kH/
kD) for k2. A nonlinear least-squares fit (R = 0.99) of
1866
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 1865 –1868

Table 2: The ee values for proteo- and deutero-pyrrolidines obtained by
cyclization of aminoalkenes catalyzed by 4.[a]
Substrate
Product
ee [%][b]
5a
93
[D2]-5a
6a
95
90
[D2]-6a
97, 98[c]
8a
89
92
[D2]-8a
Figure 1. Plot of initial cyclization rate (ꢀd[6a]/dt)ini versus [substra-
&
[a] Reaction conditions: 238C, C6D6, 1.25–12 h, quantitative yield. [b] The
ee values were determined by HPLC on a chiral stationary phase.
[c] ꢀ308C in THF.
*
te]ini for 6a ( ) and [D2]-6a ( ), measured in [D8]toluene at 238C. The
curves represent nonlinear least-squares fits [Eq. (3)].
(ꢀd[[D2]-6a]/dt)ini vs. [[D2]-6a]ini provides a curve with the
values: k2(D) = 7.2 ꢂ 0.5 ꢀ 10ꢀ3 mꢀ1 sꢀ1, K’(D) = 1.5 ꢂ 0.2 ꢀ 10ꢀ2 m,
CH2} species because an NH group is not present in the
imido moiety. Thus, the accumulated data including the rate
law, the KIE, isotopic perturbation of enantioselectivity, and
the KIE for the two enantiotopic pathways eliminate olefin
insertion and [2p + 2p] cycloaddition as possible mechanisms
and KSI(D) = 3.2 ꢂ 0.9mꢀ1
.
The composition of the catalytic species can be partly
assessed from this kinetic analysis. First, interaction of 6a and
4 in a 1:1 ratio provides the active catalyst A that is likely
[(3)Zr(NHR)(NMe2)] (catalyst initiation; NHR = amido-
alkene). The equilibrium step (k1/kꢀ1) involves reversible
interaction of catalyst A with 6a to give [(3)Zr(NHR)2] and
HNMe2. A secondary amide ligand (i.e. NMe2 or pyrrolide) is
ꢀ
for C N bond formation.
ꢀ
C N bond formation establishes the configuration of the
ꢀ
new stereocenter, whereas the new C H bond is not attached
to the stereogenic carbon atom. However, the stereochemical
ꢀ
ꢀ
relationship between the N H and the C N bond suggests
ꢀ
ꢀ
ꢀ
not a sufficient coligand for cyclization of an amidoalkene.
that C N and C H bond formation and cleavage of the N H
bond occur in a concerted fashion during the cyclization step.
(H)
The kinetic isotope effects (KIE) from initial rate plots k2
/
/
(H)
k2(D) (2.3 ꢂ 0.5) and from second-order rate constants, k’obs
Given that an N H bond is broken during the TLS and the
ꢀ
(D)
ꢀ
k’obs = 3.5 indicate that an N H bond is broken during the
TLS. These isotope effects are inconsistent with 1,2-olefin
catalytic intermediate contains two NHR ligands, we propose
ꢀ
a six-center transition state in which N H transfer from one
ꢀ
insertion into the Zr N bond. A KIE in olefin hydroamina-
amide group to the terminal methylene unit of the other
ꢀ
tion/cyclization was noted by Marks and co-workers for the
[Cp*2LnR] (Ln = lanthanoid) and [(CGC)U(NMe2)2] sys-
tems—in those cases a proton-assisted insertion is propo-
sed.[5c,12]
amidoalkene is concerted with intraligand C N bond for-
mation (Figure 2). The two participating ligands are proposed
to be two amido groups because kinetics indicate that two
substrates interact with the catalyst in the turnover limiting
step, and the addition of a third substrate, presumably as a
coordinated amine, inhibits the cyclization. Finally, this
mechanism is consistent with the observation that secondary
aminoalkenes are cyclized only in the presence of a primary
amine. Presumably, a primary amido ligand is formed that
transfers a proton to the cyclizable amidoalkene ligand.
Some of the observations reported here have been
previously observed in zirconium(IV)-, rare-earth-, and
organoactinide-mediated hydroamination reactions; includ-
ing a second-order rate law,[13] substantial KIE,[2,5c,6d,12,13c] and
isotope effects on diastereoselectivity.[12] Here, saturation
kinetics provide a mechanistic connection between zero-
order and first-order substrate dependence, as illustrated by
Importantly, the ee values for deuteron-pyrrolidines ([D2]-
5b, [D2]-6b, and [D2]-8b) are systematically and significantly
higher than the values for the corresponding proteo-pyrroli-
dines (Table 2). In the most dramatic example, the ee value
for cyclization increases from 90% (DDG° = 1.7 kcalmolꢀ1)
for 6b to more than 97% (DDG° = 2.5 kcalmolꢀ1) for [D2]-
6b. By using the rate constants k2(H) and k2(D) and the ratio of
enantiomers, the KIE for formation of each stereoisomer is
calculated: major R enantiomer: kHR/kDR = 2.2 ꢂ 0.5; minor
S enantiomer: kHS/kDS = 7.7 ꢂ 0.1.
The H (or D) atom from the amine is central to the step
that determines stereochemistry, thus ruling out intramolec-
=
=
ular [2p + 2p] cycloaddition of a {Zr NCH2CR2CH2HC
Angew. Chem. Int. Ed. 2011, 50, 1865 –1868
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1867

Communications
[1] a) S. Hong, T. J. Marks, Acc. Chem. Res.
2004, 37, 673 – 686; b) T. E. Mꢁller, K. C.
Hultzsch, M. Yus, F. Foubelo, M. Tada,
Chem. Rev. 2008, 108, 3795 – 3892.
[2] D. V. Gribkov, K. C. Hultzsch, F. Hampel, J.
Am. Chem. Soc. 2006, 128, 3748 – 3759.
[3] 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.
[4] a) M. R. Gagnꢂ, L. Brard, V. P. Conticello,
M. A. Giardello, C. L. Stern, T. J. Marks,
Organometallics 1992, 11, 2003 – 2005;
b) M. A. Giardello, V. P. Conticello, L.
Brard, M. R. Gagnꢂ, T. J. Marks, J. Am.
Chem. Soc. 1994, 116, 10241 – 10254.
[5] a) D. V. Gribkov, K. C. Hultzsch, Angew.
Chem. 2004, 116, 5659 – 5663; Angew.
Chem. Int. Ed. 2004, 43, 5542 – 5546;
b) P. D. Knight, I. Munslow, P. N. O’Shaugh-
nessy, P. Scott, Chem. Commun. 2004, 894 –
895; c) B. D. Stubbert, T. J. Marks, J. Am.
Chem. Soc. 2007, 129, 6149 – 6167; d) S.
Figure 2. Proposed new catalytic cycle for hydroamination of aminoalkenes as catalyzed by 4.
Majumder, A. L. Odom, Organometallics 2008, 27, 1174 – 1177.
[6] a) J. A. Bexrud, J. D. Beard, D. C. Leitch, L. L. Schafer, Org.
Lett. 2005, 7, 1959 – 1962; b) H. Kim, P. H. Lee, T. Livinghouse,
Chem. Commun. 2005, 5205 – 5207; c) D. A. Watson, M. Chiu,
R. G. Bergman, Organometallics 2006, 25, 4731 – 4733; d) A. L.
Gott, A. J. Clarke, G. J. Clarkson, P. Scott, Chem. Commun.
2008, 1422 – 1424.
[7] K. Manna, A. Ellern, A. D. Sadow, Chem. Commun. 2010, 46,
339 – 341.
[8] See the Supporting Information.
our observations on an oxazolinylborate magnesium hydro-
amination system.[14] Still, the involvement of the mechanism
proposed here in other early transition metal catalyzed
hydroamination reactions requires further experiments, as
considerable evidence supports the insertion mechanism for
many olefin hydroaminations, including those catalyzed by
[Cp2ZrMe]+ (Cp = cyclopentadienyl), [(CGC)Zr(Me)Cl],
and [(dpm)Zr(NMe2)2].[5a,c,d] Furthermore, computational
ꢀ
studies have provided support for olefin insertion into La
[9] J. F. Dunne, K. Manna, J. W. Wiench, A. Ellern, M. Pruski, A. D.
Sadow, Dalton Trans. 2010, 39, 641 – 653.
=
N bonds as well as neutral {Zr NR/alkene} [2p + 2p] cyclo-
addition reactions in other systems.[15] The unusually high
reactivity and enantioselectivity available with 4 as a preca-
talyst in hydroaminations may relate to the ability of the {3}Zr
moiety to stabilize the proposed six-center transition state.
Studies are currently underway that will test our proposed
mechanism and develop a model that rationalizes the
observed enantioselectivity through variation of the ancillary
ligands.
[10] The ee values from Ref. [3]: 5b: 95%; 6b: 85%; 8b: 73%; from
Ref. [4]: 5b: 74%; 6b: 82%; 7b: 88%; 8b: 93%.
[11] A. Cornish-Bowden, Fundamentals of Enzyme Kinetics, 3rd ed.,
Portland Press, London, 2004, pp. 137 – 141.
[12] M. R. Gagnꢂ, C. L. Stern, T. J. Marks, J. Am. Chem. Soc. 1992,
114, 275 – 294.
[13] a) R. K. Thomson, J. A. Bexrud, L. L. Schafer, Organometallics
2006, 25, 4069 – 4071; b) C. Mꢁller, W. Saak, S. Doye, Eur. J. Org.
Chem. 2008, 2731 – 2739; c) A. L. Reznichenko, K. C. Hultzsch,
Organometallics 2010, 29, 24 – 27.
[14] J. F. Dunne, B. Fulton, A. Ellern, A. D. Sadow, J. Am. Chem. Soc.
2010, 132, 17680 – 17683.
Received: October 1, 2010
Published online: January 11, 2011
[15] a) A. Motta, G. Lanza, I. L. Fragala, T. J. Marks, Organometal-
lics 2004, 23, 4097 – 4104; b) B. F. Straub, R. G. Bergman, Angew.
Chem. 2001, 113, 4768 – 4771; Angew. Chem. Int. Ed. 2001, 40,
4632 – 4635; c) C. Mꢁller, R. Koch, S. Doye, Chem. Eur. J. 2008,
14, 10430 – 10436.
Keywords: enantioselectivity · hydroamination · isotope effects ·
oxazolines · pyrrolidines
.
1868
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 1865 –1868