Lewis acid-lewis base catalyzed enantioselective hetero-diels-alder reaction for direct access to δ-lactones

Article abstract of DOI:10.1021/ol800742d

A complex formed In situ from Er(OTf)₃ and a simple commercially available norephedrine ligand promotes an unprecedented [4 + 2] cycloaddition of α,β-unsaturated acid chlorides with a broad range of aromatic and heteroaromatic aldehydes by a cooperative bifunctional Lewis acid-Lewis base catalytic mode of action providing valuable δ-lactone building blocks with excellent enantioselectivity.

Full text of DOI:10.1021/ol800742d

ORGANIC
LETTERS
2008
Vol. 10, No. 10
2019-2022
Lewis Acid-Lewis Base Catalyzed
Enantioselective Hetero-Diels-Alder
Reaction for Direct Access to
δ-Lactones
Paolo S. Tiseni and Rene´ Peters*
ETH Zu¨rich, Laboratory of Organic Chemistry, Wolfgang-Pauli-Str. 10,
Ho¨nggerberg HCI E 111, CH-8093 Zu¨rich, Switzerland
peters@org.chem.ethz.ch
Received April 2, 2008
ABSTRACT
A complex formed in situ from Er(OTf)3 and a simple commercially available norephedrine ligand promotes an unprecedented [4 + 2] cycloaddition
of r,ꢀ-unsaturated acid chlorides with a broad range of aromatic and heteroaromatic aldehydes by a cooperative bifunctional Lewis acid-Lewis
base catalytic mode of action providing valuable δ-lactone building blocks with excellent enantioselectivity.
δ-Lactone moieties constitute an exceptionally widespread
structural motif in natural and synthetic compounds which
display high efficacy, e.g., as anticancer1 or cholesterol-
lowering agents. For instance, the majority of statin drugs
such as Lipitor and Zocor, the world’s biggest selling drugs
in the last years, contain a ꢀ-hydroxy-δ-lactone moiety or
the corresponding open-chain carboxylate form.2
ditions and due to their inherent instability (rapid dimerization
and polymerization).7,8 However, in our preceding studies
we demonstrated that these species can be trapped in situ
and at the same time activated as diene components for [4
+ 2] cycloadditions by an enantiopure nucleophilic tertiary
amine, thus forming zwitterionic dienolates 1 (Figure 1)
which are able to undergo an enantioselective HDA reaction
with the highly activated aldehyde chloral.
Hetero-Diels-Alder (HDA) reactions3 of Brassard’s diene
(1,3-dimethoxy-1-(trimethylsilyloxy)butadiene) and alde-
hydes have been developed to synthesize ꢀ-methoxy-
substituted R,ꢀ-unsaturated δ-lactones.4,5 Recently, we have
reported a direct concept using R,ꢀ-unsaturated acid chlorides
as substrates which is based on the in situ formation of
vinylketenes.6 These intermediates were previously not useful
for catalytic asymmetric Diels-Alder reactions as a result
of their tendency to preferentially undergo [2 + 2] cycload-
Nonactivated aldehydes, e.g. benzaldehyde or even p-
nitrobenzaldehyde, did not furnish the targeted products due
(4) (a) Fan, Q.; Lin, L.; Liu, J.; Huang, Y.; Feng, X. Eur. J. Org. Chem.
2005, 3542. (b) Lin, L.; Fan, Q.; Qin, B.; Feng, X. J. Org. Chem. 2006, 71,
4141. (c) Du, H.; Zhao, D.; Ding, K. Chem. Eur. J. 2004, 10, 5964. (d)
Lin, L.; Chen, Z.; Yang, X.; Liu, X.; Feng, X. Org. Lett. 2008, 10, 1311.
For an alternative enantioselective concept using allenic esters, see: (e)
Oisaki, K.; Zhao, D.; Kanai, M.; Shibasaki, M. J. Am. Chem. Soc. 2007,
129, 7439
.
(5) (a) Reviews about δ-lactones: Collins, I. J. Chem. Soc., Perkin Trans.
1 1999, 1377. (b) Boucard, V.; Broustal, G.; Campagne, J. M. Eur. J. Org.
Chem. 2007, 225
.
(1) Shaw, S. J.; Sundermann, K. F.; Burlingame, M. A.; Myles, D. C.;
Freeze, B. S.; Xian, M.; Brouard, I.; Smith, A. B., III J. Am. Chem. Soc.
2005, 127, 6532.
(6) Tiseni, P. S.; Peters, R. Angew. Chem., Int. Ed. 2007, 46, 5325
.
(7) Tidwell, T. T. Ketenes; John Wiley & Sons: Hoboken, NJ, 2006.
(8) 2-(Trimethylsilyl)vinylketene has been reported to be a remarkably
stable vinylketene and has been used for non-enantioselective HDA
reactions: Bennett, D. M.; Okamoto, I.; Danheiser, R. L. Org. Lett. 1999,
1, 641.
(2) Istvan, E. S.; Deisenhofer, J. Science 2001, 292, 1160.
(3) General review on HDA reactions: Jørgensen, K. A. Angew. Chem.,
Int. Ed. 2000, 39, 3558.
10.1021/ol800742d CCC: $40.75
Published on Web 04/23/2008
2008 American Chemical Society

should be advantageous to achieve a rigid cycloaddition
transition state. The assumption of a cooperative bifunctional
Lewis acid-Lewis base activation mechanism was initially
supported by the observation that Er(OTf)3 did not destroy
the nucleophilicity of pyridine in the model reaction pre-
sented in Table 1 (entry 1), while no reaction took place in
the absence of either nucleophile or Lewis acid.
Figure 1. Zwitterionic dienolates 1 as reactive dienes for enanti-
oselective HDA reactions.
to poor reactivity of the organocatalysts. A bifunctional
Lewis acid-Lewis base catalyst by which the reactivity of
both the dienolate and aldehyde could be controlled was
therefore required.9,10 The combination of Lewis acids and
Lewis bases successfully united in one catalytic system has
recently found numerous applications in asymmetric catalysis
due to synergistic activation of both the electrophilic and
nucleophilic substrates, often allowing high reaction rates
and excellent chirality transfer.11
Table 1. Development of the Title Reactiona
The present studies were based upon the hypothesis that
a Lewis acid such as a lanthanide salt, which offers an
exceptionally high number of coordination sites, would be
advantageous to bind both aldehyde and dienolate plus
additional ligands to control reactivity and stereoselectivity.12
The development of lanthanide complexes acting as bifunc-
tional catalysts was pioneered by the seminal work of
Shibasaki et al.13
To implement a cycloaddition transition state with a high
level of organization the nucleophilic catalyst should be
directly connected to the Lewis acid template. We envisaged
that an oxophilic lanthanide would tightly bind to an
alcoholate moiety while a tertiary amino group would
undergo a hemilabile coordination still permitting a sufficient
reactivity to nucleophilically trap a vinylketene intermediate.
Although Er(III) complexes with aliphatic ꢀ- or γ-amino
alcohols possessing a tertiary amino group have to our
knowledge never been previously described in the literature,
Er(OTf)3 was chosen for these investigations as lanthanide
source owing to the combination of (a) a comparatively low
price of Er which is linked to its importance for telecom-
munication industry14 and (b) a relatively small ionic radius
(as a consequence of the lanthanide contraction)15 which
a Compound 2a was slowly added by syringe pump over 120 min (1:1
stoichiometry of both substrates). Stirring was continued for an additional
150 min. b NMR yields using MeNO2 as internal standard. c Determined
by chiral column HPLC. d Compound 2a was added over 30 min. e T )
-10 °C. f 1.5 equiv of Er(OTf)3. g 2.5 equiv of DIPEA. h 0.2 equiv of 4b.
i 0.1 equiv of 4b. j 0.05 equiv of 4b.
(9) In our previous work, Sn(OTf)2 was used as cocatalyst but was not
directly involved in the cycloaddition step itself, see ref 6.
With N-methylephedrine 4a, δ-lactone 5a was formed with
a promising ee of 74% (entry 2), yet the yield was low.
Replacing the NMe2 group by a pyrrolidine unit not only
enhanced the reactivity (yield ) 35%) but also resulted in
an ee value of 95% (entry 3). The nucleophilicity of the
tertiary amino group is essential as entries 4 and 5 demon-
strate, in which the steric accessibility and the electron
density of the amino group are diminished with the conse-
quence of reduced or no reactivity. Whereas in the case of
a tertiary or primary alcohol moiety the title reaction was
retarded (entries 6 and 7), a methyl-protected hydroxyl
impeded high enantioselectivity (entry 8), while TMS
protection gave no product at all. Entry 9 demonstrates that
(10) For the application of bifunctional catalysts in [2 + 2] cycloadditions
of ketenes, see, e.g.: (a) LiClO4 in combination with O-protected cinchona
alkaloids: Zhu, C.; Shen, X.; Nelson, S. G. J. Am. Chem. Soc. 2004, 126,
5352. (b) Metal triflates including Er(OTf)3 in combination with O-protected
cinchona alkaloids: Calter, M. A.; Tretyak, O. A.; Flaschenriem, C. Org.
Lett. 2005, 7, 1809. (c) In(OTf)3 in combination with O-protected cinchona
alkaloids: France, S.; Shah, M. H.; Weatherwax, A.; Wack, H.; Roth, J. P.;
Lectka, T. J. Am. Chem. Soc. 2005, 127, 1206. (d) Oxazaborolidine catalyst:
Gnanadesikan, V.; Corey, E. J. Org. Lett. 2006, 8, 4943. (e) A hybrid
cinchona alkaloid/salene-Co complex: Lin, Y.-M.; Boucau, J.; Li, Z.;
Casarotto, V.; Lin, J.; Nguyen, A. N.; Ehrmantraut, J. Org. Lett. 2007, 9,
567
.
(11) Dual activation catalysis review: Ma, J.-A.; Cahard, D. Angew.
Chem., Int. Ed. 2004, 43, 4566.
(12) Review about the use of lanthanides in asymmetric catalysis:
Mikami, K.; Terada, M.; Matsuzawa, H. Angew. Chem., Int. Ed. 2002, 41,
3554.
(13) (a) Shibasaki, M.; Sasai, H.; Arai, T. Angew. Chem., Int. Ed. Engl.
1997, 36, 1236. (b) Shibasaki, M.; Kanai, M.; Funabashi, K. Chem.
Commun. 2002, 1989. (c) Kanai, M.; Kato, N.; Ichikawa, E.; Shibasaki,
M. Synlett 2005, 1491. (d) Shibasaki, M.; Matsunaga, S. Chem. Soc. ReV.
2006, 35, 269.
(14) Bellemare, A. Prog. Quant. Electron. 2003, 27, 211. Er(OTf)3 is,
e.g., ca. 4-5 times less expensive than Yb(OTf)3.
(15) Shannon, R. D. Acta Crystallogr. 1976, A32, 751.
2020
Org. Lett., Vol. 10, No. 10, 2008

O-acylated amino alcohols as in 4h cannot be significantly
involved in the catalytic cycle. Compound 4h has also never
been detected in the reaction mixture or crude product using
4b as ligand. This indicates that (a) the oxygen atom binds
to the Er(III) ion and (b) that this bond must be inert under
the reaction conditions. The complexation of 4b proceeds
very rapidly since identical results were obtained by either
preformation of the catalyst for 15 or 150 min at rt or by
adding the ligand to the reaction mixture without any
precoordination time. Even if the alcohol moiety is first
deprotonated by NaH at rt for 1 h the outcome is identical
in terms of both yield and enantioselectivity.
To enhance the yield to a synthetically useful level, it was
necessary to decrease the addition time of 2a from 150 to
30 min (entry 10)16 to raise the reaction temperature from
-15 to -10 °C (entry 11), the amount of Er(OTf)3 from 1.1
to 1.5 equiv (entry 12) and the amount of DIPEA from 2.0
to 2.5 equiv (entry 13), while the amount of the chiral
nucleophilic ligand could be decreased to 10-20 mol %
without seriously impacting the reaction outcome (entries
14-15). Amino alcohol loadings lower than 10 mol %
resulted in reduced yield and enantioselectivity (entry 16).
In contrast to the organocatalyst of our previous studies,6
the novel Lewis acid-Lewis base catalyst system provided
excellent enantioselectivities irrespective of the size of the
substituent R1 at the 3-position of acid chloride 2 (Table 2,
entries 1-6). Unbranched alkyl groups such as Me or Et,
R- or ꢀ-branched alkyls as in i-Pr and i-Bu, alicyclic
substituents like cyclohexyl or aromatic groups such as Ph
all furnished ee values g94% using PhCHO as test substrate.
While the yield was low with 3,3-dimethylacryloyl chloride
2b, which is notoriously highly sensitive toward polymeri-
zation under basic reaction conditions, the yields were
synthetically useful in all other cases employing 3a as
dienophile.
Table 2. Scope and Limitations of the Title Reactiona
no.
5
R1
R2
4b/equiv yieldb/% eec/%
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
a
iPr
Me
Et
iBu
cHex Ph
Ph
Ph
Ph
Ph
0.2
0.2
0.2
0.2
0.2
0.1
0.2
0.2
0.2
0.1
0.2
0.1
0.2
0.1
0.2
0.2
0.1
0.2
0.1
0.2
0.2
56
24
62
54
65
64
26
30
55
77
78
71
77
70
91
72
87
62
23
40
46
95
95
95
98
96
94
94
94
95
88
93
92
95
93
91
88
93
92
94
95
96
b
c
d
e
f
g
h
i
j
k
l
m
n
o
p
q
r
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
o-MeOC6H4
p-MeC6H4
2-naphthyl
o-ClC6H4
m-ClC6H4
p-ClC6H4
m-BrC6H4
p-BrC6H4
o-O2NC6H4
p-O2NC6H4
m-F3CC6H4
p-O2NC6H4CHdCH
2-furyl
s
t
u
2-thiophenyl
2-(3-Br)thiophenyl
a Compound 2 was slowly added by syringe pump over 30 min (1:1
stoichiometry of both substrates). Stirring was continued for an additional
120 min. b Isolated yield. c Determined by chiral column HPLC.
The reaction generally provided excellent enantioselec-
tivities with all kinds of aromatic aldehydes regardless of
their electronic or steric nature (entries 7-21, ee )
88-96%).17 While electron donors such as o-OMe or p-Me
resulted in lower yields (entries 7 and 8), electron-withdraw-
ing substituents such as Cl, Br, NO2, or CF3 enhanced the
reactivity (entries 10-17) as compared to PhCHO and also
promoted the use of R,ꢀ-unsaturated enals (entry 18). The
substitution pattern of the aldehyde is less important, since
all ortho-, meta-, or para-substituted systems were well
tolerated. Even employing electron-rich heterocycles such
as furan or thiophene (entries 19-21) was successful while
the catalyst system is not yet amenable to aliphatic alde-
hydes.18
1), which is further supported by the absence of a nonlinear
effect indicating that higher aggregates are most likely not
involved. The results presented in Table 1 are in accordance
with a mechanism in which the reversibly binding Lewis
basic site forms a nucleophilic dienolate which strongly binds
to the metal ion in 7,20 resulting in a highly organized
Scheme 1. Proposed Catalytic Cycle
Er(III) is known to prefer high coordination numbers,
typically 7-10.19 We therefore assume that the ligand and
both substrates all bind to the same metal center (Scheme
(16) Slow addition of the acid chloride substrates over 30 min is
recommended to maintain a low vinylketene concentration so as to minimize
di- and oligomerization.
(17) The absolute configuration was determined by chemical correlation
(see the Supporting Information).
(18) The enolizable dihydrocinnamaldehyde and cyclohexylcarbaldehyde
as well as the nonenolizable pivaldehyde resulted in complete aldehyde
decomposition.
Org. Lett., Vol. 10, No. 10, 2008
2021

transition state for a vinylogous aldol addition reaction.21
Turnover is achieved by an intramolecular acylation.
Cl ions generated from 2 are assumed to be the reason
for the required use of stoichiometric amounts of Er(OTf)3,
since the coordination of Cl- might deactivate the catalyst
species and additional Er(III) might be necessary as a Cl-
trap. This is supported by the fact that ErCl3 cannot catalyze
the title reaction, while mixtures of Er(OTf)3 and ErCl3 result
in considerably decreased reactivity. To render the process
catalytic in the lanthanide, an alternative Cl- trap or an
alternative leaving group would be required.
In conclusion, we have developed a novel bifunctional
Lewis acid-Lewis base catalyst system which enables the
[4 + 2] cycloaddition of R,ꢀ-unsaturated acid chlorides and
a broad range of aromatic and heteroaromatic aldehydes,
providing direct access to δ-lactone building blocks with
generally excellent enantioselectivity. Our results show that
(a) Er(III) and the amino alcohol ligand form an inert Er-O
bond precluding O-acylation, (b) the Lewis acid22 as well
as the nucleophilic amino moiety are essential for product
formation, and (c) the catalyst is most likely a monomeric
species. A key characteristic of the catalyst system is its
simplicity, since the commercially available nucleophilic
amino alcohol ligand can be prepared from inexpensive
norephedrine23 in a single step.24 This catalyst should also
be attractive for alternative reactions which rely on a
synergistic activation mechanism. Studies along these lines
are underway.
Acknowledgment. This work was financially supported
by F. Hoffmann-La Roche. We thank Priv.-Doz. Dr. Martin
Karpf and Dr. Paul Spurr (both F. Hoffmann-La Roche,
Synthesis and Process Research) for carefully reading this
manuscript and Nicolas Dietl (TU Berlin) for technical
assistance (semester internship at ETHZ).
Supporting Information Available: Experimental pro-
1
cedures, full characterization data for all new products, H/
13C NMR spectra, and HPLC data. This material is available
free of charge via the Internet at http://pubs.acs.org.
OL800742D
(19) (a) Selected examples: Ren, Y.; Chen, S.; Xie, G.; Gao, S.; Shi, Q.
Inorg. Chim. Acta 2006, 359, 2047. (b) Freire, R. O.; do Monte, E. V.;
Rocha, G. B.; Simas, A. M. J. Organomet. Chem. 2006, 691, 2584.
(20) Er(III) complexes are paramagnetic thus precluding NMR investiga-
tions. Unfortunately, mass spectroscopic investigations did not lead to the
identification of significant catalyst species, and attempts to get X-ray quality
crystals have failed so far.
(22) Similar results as with Er(OTf)3 were also obtained with Yb(OTf)3
possessing a similar ionic radius (R1, R2 ) Ph: yield ) 63%; ee ) 96%),
while the larger Ce(III) (yield ) 22%; ee ) 34%) and Pr(III) (yield )
11%; ee ) 34%) gave significantly lower reactivity and enantioselectivity.
(23) Both enantiomers of norephedrine or 4b are commercially available
at a comparable price.
(24) (a) Kang, J.; Lee, J. W.; Kim, J. I. J. Chem. Soc., Chem. Commun.
1994, 17, 2009. (b) Zhao, D.; Chen, C.-Y.; Xu, F.; Tan, L.; Tillyer, R.;
Pierce, M. E.; Moore, J. R. Org. Synth. 2000, 77, 12.
(21) Denmark, S. E.; Heemstra, J. R., Jr.; Beutner, G. L. Angew. Chem.,
Int. Ed. 2005, 44, 4682.
2022
Org. Lett., Vol. 10, No. 10, 2008