Highly enantioselective α-aminoxylation of aldehydes and ketones with a polymer-supported organocatalyst

Article abstract of DOI:10.1021/ol070526p

The first catalytic enantioselective α-aminoxylation of aldehydes and ketones using an insoluble, polymer-supported organocatalyst (1) derived from trans-4-hydroxyproline is reported (ee: 96-99%). Reaction rates in the aminoxylation of cyclic ketones with 1 are higher than those reported with L-proline. The insoluble nature of 1 simplifies workup conditions and allows catalyst recycling without an apparent decrease in enantioselectivity or yield.

Full text of DOI:10.1021/ol070526p

ORGANIC
LETTERS
2007
Vol. 9, No. 10
1943-1946
Highly Enantioselective
of Aldehydes and Ketones with a
Polymer-Supported Organocatalyst
r-Aminoxylation
Daniel Font,†,‡ Amaia Bastero,† Sonia Sayalero,† Ciril Jimeno,† and
Miquel A. Perica`s*,†,‡
Institute of Chemical Research of Catalonia (ICIQ), AV. Pa¨ısos Catalans, 16,
43007 Tarragona, Spain, and Departament de Qu´ımica Orga`nica, UniVersitat de
Barcelona (UB), 08028 Barcelona, Spain
mapericas@iciq.es
Received March 2, 2007
ABSTRACT
The first catalytic enantioselective
from trans-4-hydroxyproline is reported (ee: 96
reported with -proline. The insoluble nature of 1 simplifies workup conditions and allows catalyst recycling without an apparent decrease in
r
-aminoxylation of aldehydes and ketones using an insoluble, polymer-supported organocatalyst (1) derived
−99%). Reaction rates in the aminoxylation of cyclic ketones with 1 are higher than those
L
enantioselectivity or yield.
Optically active R-hydroxycarbonyl compounds and 1,2-diols
are interesting building blocks for the construction of
biologically active products. As a consequence, a wide
variety of diastereoselective and enantioselective methods
for their preparation have been developed. Among these
methods, the R-aminoxylation of carbonyl compounds has
been the subject of much interest in recent times.1 The first
catalytic enantioselective introduction of an aminoxyl group
at the R-position of ketones, involving nitrosobenzene as the
electrophile, was reported in 2003 by Yamamoto using tin
enolates and a silver complex as the catalyst.2 Also in 2003,
the direct proline-catalyzed R-aminoxylation of aldehydes
with nitrosobenzene was independently reported by Zhong,3
MacMillan,4 and Hayashi.5a One year later, the scope of the
proline-catalyzed reaction was extended to ketones simul-
taneously by the groups of Hayashi5b and Co´rdova.6 After
this initial breakthrough, however, little has been reported
concerning the use of new organocatalysts7 or substrates8 in
this process, and most additional effort has been devoted to
the elucidation of the mechanism in the proline-catalyzed
reaction (Scheme 1).9
Scheme 1. Proline-Mediated R-Aminoxylation of Carbonyl
Compounds
Associated with this process but normally overlooked, the
formation of azoxybenzene usually accompanies that of the
† ICIQ.
‡ UB.
(1) (a) Merino, P.; Tejero, T. Angew. Chem., Int. Ed. 2004, 43, 2995.
(b) Janey, J. M. Angew. Chem., Int. Ed. 2005, 44, 4292.
(2) Momiyama, N.; Yamamoto, H. J. Am. Chem. Soc. 2003, 125, 6038.
(3) Zhong, G. Angew. Chem., Int. Ed. 2003, 42, 4247.
(4) Brown, S. P.; Brochu, M. P.; Sinz, C. J.; MacMillan, D. W. C. J.
Am. Chem. Soc. 2003, 125, 10808.
(5) (a) Hayashi, Y.; Yamaguchi, J.; Hibino, K.; Shoji, M. Tetrahedron
Lett. 2003, 44, 8293. (b) Hayashi, Y.; Yamaguchi, J.; Sumiya, T.; Shoji,
M. Angew. Chem., Int. Ed. 2004, 43, 1112.
(6) Bøgevig, A.; Sunde´n, H.; Co´rdova, A. Angew. Chem., Int. Ed. 2004,
43, 1109.
10.1021/ol070526p CCC: $37.00
© 2007 American Chemical Society
Published on Web 04/17/2007

R-aminoxylation product.5b Barbas and co-workers proposed
that azoxybenzene arises from the reaction of the R-ami-
noxylation product with nitrosobenzene.10 Working with an
excess of this reagent, these authors could perform ami-
noxylation and O-N bond heterolysis in a tandem manner
for the direct, enantioselective R-hydroxylation of cyclohex-
anones (Scheme 2). From a practical perspective, it becomes
found in the literature for the use of polymer-supported
catalysts in the R-aminoxylation of carbonyl compounds.14
We report here the first asymmetric R-aminoxylation of
aldehydes and ketones using 1, a readily available, recover-
able, and reusable organocatalyst.
Our initial goal in this study was the optimization of the
reaction conditions for the suppression of azoxybenzene
formation in the R-aminoxylation of ketones using the
supported catalyst 1 and cyclohexanone (2a) as a model
substrate (Table 1). It must be pointed out here that, although
Scheme 2. Tandem Aminoxylation and O-N Bond
Heterolysis of Cyclohexanones
Table 1. R-Aminoxylation of Cyclohexanone (2a) with
Polymer-Supported Catalyst 1a
addition
time (h)
entry
solvent
yield (%)b
ee (%)c
1
2
3
4
5
CHCl3
CHCl3
DMF
DMF
DMF
0d
0f
0h
5i
3i
21e
nd
nd
96
98
98
0g
clear that the detection of azoxybenzene in aminoxylation
reaction crudes is at the expense of R-aminoxylation yield.
Although organocatalysis incorporates important charac-
teristics of environmentally benign practices through the
avoidance of toxic metal contamination both in waste and
in reaction products, the possibility of separating the catalyst
by simple physical methods and even reusing it would
represent an additional bonus in view of a potential large-
scale operation. Thus, the development of polymer-supported
organocatalysts suitable for work in a broad variety of
solvents without deterioration of the catalytic performance
of their monomeric counterparts is the subject of much
current interest.11 We have recently reported a new strategy
for supporting trans-4-hydroxyproline onto Merrifield-type
resins through Cu(I)-catalyzed Huisgen 1,3-dipolar cyclo-
addition12 (click chemistry) and have shown that the resulting
resin (1) behaves as a highly active enantioselective and
diastereoselective, yet reusable, organocatalyst for the direct
aldol reaction in water.13 Up to now, no reference can be
25
48
60 (74)e
a
Reaction conditions: resin 1 (0.05 mmol; f: 0.6 mmol/g); 1 mL of
solvent; variable amounts of 2a and nitrosobenzene depending on the
experiment; 23 °C. Isolated yield of 3a. Determined by HPLC; see
Supporting Information. d 1 (10 mol %); 2a (10 equiv); nitrosobenzene (1
equiv); 2 h. e Estimated by 1H NMR with mesitylene as the internal standard.
b
c
f
g
1 (20 mol %); 2a (1 equiv); nitrosobenzene (3 equiv); 24 h; 4 °C.
A
1:3.3 mixture of 2-hydroxycyclohexanone and azoxybenzene was formed
in low yield. h 1 (20 mol %); 2a (10 equiv); nitrosobenzene (1 equiv); 2 h.
i
1 (20 mol %), 2a (2 equiv); nitrosobenzene (1 equiv; 0.5 M solution).
most of the work reported on this reaction uses L-proline as
the catalyst, very different reaction conditions have been
used, which is somewhat intriguing.15
Thus, a clear difference exists between Hayashi’s proce-
dure which involves reaction at low temperature (0 °C) and
slow addition of nitrosobenzene to an excess of ketone as
key features to suppress undesired R,R′-bisaminoxylation and
azoxybenzene formation5b and that of Co´rdova,6 who reported
reactions at room temperature in chloroform with addition
of the reactants in one portion, without the observation of
any side reaction. We first checked the addition of reactants
at once to the suspension of resin 1, and because resin
swelling is critical for the successful use of insoluble
polymer-supported catalysts, we paid attention to identify
solvents with good swelling properties for 1. When the
reaction was conducted in chloroform, at room temperature
(7) Hayashi, Y.; Yamaguchi, J.; Hibino, K.; Sumiya, T.; Urushima, T.;
Shoji, M.; Hashizume, D.; Koshino, H. AdV. Synth. Catal. 2004, 346, 1435.
(8) Kim, S.-G.; Park, T.-H. Tetrahedron Lett. 2006, 47, 9067.
(9) Mechanistic studies: (a) Cheong, P. H.-Y.; Houk, K. N. J. Am. Chem.
Soc. 2004, 126, 13912. (b) Mathew, S. P.; Iwamura, H.; Blackmond, D. G.
Angew. Chem., Int. Ed. 2004, 43, 3317. (c) Mathew, S. P.; Klussmann, M.;
Wells, D. H., Jr.; Armstrong, A.; Blackmond, D. G. Chem. Commun. 2006,
4291.
(10) Ramachary, D. B.; Barbas, C. F., III. Org. Lett. 2005, 7, 1577.
(11) For reviews on immobilized organocatalysts, see: (a) Benaglia, M.;
Puglisi, A.; Cozzi, F. Chem. ReV. 2003, 103, 3401. (b) Cozzi, F. AdV. Synth.
Catal. 2006, 348, 1367. For approaches to proline immobilization, see: (c)
Benaglia, M.; Cinquini, M.; Cozzi, F.; Puglisi, A.; Celentano, G. AdV. Synth.
Catal. 2002, 344, 533. (d) Benaglia, M.; Cinquini, M.; Cozzi, F.; Puglisi,
A.; Celentano, G. J. Mol. Catal. A: Chem. 2003, 204-205, 157. (e) Kondo,
K.; Yamano, T.; Takemoto, K. Makromol. Chem. 1985, 186, 1781. (f)
Sakthivel, K.; Notz, W.; Bui, T.; Barbas, C. F. J. Am. Chem. Soc. 2001,
123, 5260.
(13) Font, D.; Jimeno, C.; Perica`s, M. A. Org. Lett. 2006, 8, 4653.
(14) For a different strategy towards catalyst recovery in this reaction,
involving the use of ionic liquids, see: (a) Guo, H.-M.; Niu, H.-Y.; Xue,
M.-X.; Guo, Q.-X.; Cun, L.-F.; Mi, A.-Q.; Jiang, Y.-Z.; Wang, J.-J. Green
Chem. 2006, 8, 682. (b) Huang, K.; Huang, Z.-Z.; Li, X.-L. J. Org. Chem.
2006, 71, 8320.
(12) (a) Rostovtsed, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B.
Angew. Chem., Int. Ed. 2002, 41, 2596. (b) Tornøe, C. W.; Christensen,
C.; Meldal, M. J. Org. Chem. 2002, 67, 3057.
(15) For a comment on this discrepancy, see: Hayashi, Y.; Yamaguchi,
J.; Sumiya, T.; Hibino, K.; Shoji, M. J. Org. Chem. 2004, 69, 5966.
1944
Org. Lett., Vol. 9, No. 10, 2007

for 2 h, with a large excess of 2a (10 equiv) relative to
nitrosobenzene, the desired R-aminoxylation product 3a was
obtained in only 21% yield, along with 2-hydroxycyclohex-
anone (12%) and azoxybenzene (entry 1). In an attempt to
improve the selectivity of the reaction and, hence, to increase
the yield, we used an excess of nitrosobenzene with respect
to the ketone (3:1). In this manner, we expected to induce
the O-N bond heterolysis after the aminoxylation step (see
Scheme 2) as reported by Barbas.10 Indeed, a mixture of
2-hydroxycyclohexanone and azoxybenzene (1:3.3 ratio) was
Table 2. R-Aminoxylation of Ketones with the
Polymer-Supported Catalyst 1a
1
exclusively detected in the reaction crude by H NMR
spectroscopy, although conversion was very low (entry 2).
In view of the negative results recorded with chloroform,
we turned our attention to DMF. In this solvent, when
nitrosobenzene was added to the reaction mixture in one
portion (entry 3), 3a was obtained in 25% yield and 96% ee
(entry 3).16 The effect of slow addition of nitrosobenzene to
the reaction mixture was next investigated.5b To our delight,
both the yield and enantioselectivity of the reaction were
significantly improved with this modification (entries 4 and
5).
The reaction worked most nicely when nitrosobenzene was
added over a 3 h period to a mixture of 2a and 1 in DMF,
providing the desired 3a in 74% yield and 98% ee. Very
interestingly, the insoluble nature of 1 allowed for a highly
simplified product isolation by filtration of the resin followed
by removal of solvents and unreacted 2a under reduced
pressure (see Supporting Information). Neither the R,R′-
bisaminoxylation product nor 2-hydroxycyclohexanone was
observed in the crude product, which was of satisfactory
purity for synthetic purposes. Analytically pure 3a could be
isolated by flash column chromatography, but at the expense
of some yield decrease (60%, entry 5), because the chro-
matographic process induced always partial O-N cleavage.
To assess the performance of 1, the optimized reaction
conditions were then applied to a set of six-membered ring
cyclic ketones. In analogy with what is observed for 2a,
excellent enantioselectivities were recorded with substrates
2b-e, as shown in Table 2. Yields were generally high,
although with tetrahydro-4H-pyran-4-one (2d) and 1-methyl-
4-piperidone (2e) they were somewhat lower (entries 4 and
5). A similar behavior has been reported with L-proline.5
It is interesting to note that 1, in spite of its heterogeneous
nature, is able to promote faster reactions than proline.5 As
an additional bonus derived from its insoluble nature, the
R-aminoxylated ketone (3a) can be converted to the corre-
sponding R-hydroxy ketone through a greatly simplified
procedure involving filtration of the catalyst (what effectively
stops the R-aminoxylation process) and addition of excess
nitrosobenzene to the reaction mixture to selectively cleave
the O-N bond (see Supporting Information).10 In this
manner, the two steps involved in the enantioselective
conversion of a ketone substrate into the corresponding
R-hydroxyketone can be separated in time, with an important
gain in chemoselectivity.
a
Reaction conditions: catalyst 1 (20 mol %; f: 0.6 mmol/g); 1 mL of
DMF; ketone (0.5 mmol); nitrosobenzene (0.25 mmol); 0.5 M in DMF;
addition, 3 h; 23 °C. b Calculated by 1H NMR with mesitylene as the internal
standard. In parentheses: isolated yield. c Determined by chiral HPLC (see
Supporting Information).
Because the reaction of aldehydes with nitrosobenzene is
known to be faster than that of ketones, we reasoned that
the balance between R-aminoxylation and the subsequent
reaction of nitrosobenzene with the aminoxylated product
would be even more favorable with these substrates. To test
the suitability of 1 for this reaction, we simply mixed
polymer-supported catalyst 1 (10 mol %) with nitrosobenzene
and propanal (3 equiv), taken as a representative substrate,
at low temperature in chloroform.17 Under these conditions,
propanal (4a) was R-aminoxylated in only 1.5 h, at 0-4 °C,
with 5a being isolated in 81% yield (Table 3, entry 1). Very
interestingly, no aldol product could be detected in the
reaction crude. So, it is clear that, in the presence of 1,
aldehydes react much faster with nitrosobenzene than with
themselves18 and also faster than ketones do. After this
encouraging result, a representative set of aldehydes (4b-
h) was also tested under these reaction conditions (addition
of nitrosobenzene in one portion, 0-4 °C, 3 equiv of
aldehyde, 10 mol % of catalyst). These results are collected
in Table 3. Because the R-aminoxylated aldehydes are not
stable,3,4 they were isolated as the R-aminoxylated primary
alcohols (6a-h) after in situ reduction with sodium boro-
(16) DMSO was also tried and afforded the aminoxylated ketone in 20%
yield. It must be noted that the 1H NMR spectra of the reaction crudes (in
DMF and DMSO) showed small signals due to the formation of the R,R′-
bisaminoxylation product which was not observed using chloroform.
(17) An experiment in DMF with the conditions of entry 1 (Table 2)
showed, after 7 h, no progress of the reaction.
(18) For examples of self-aldolization of aldehydes, see: Northrup, A.;
MacMillan, D. W. C. J. Am. Chem. Soc. 2002, 124, 6798.
Org. Lett., Vol. 9, No. 10, 2007
1945

obtained for 4h (entry 8) can be attributed to a decreased
reactivity of the rather hindered, enamine intermediate
(compare entries 1-7 with entry 8). Substrates containing
olefinic double bonds or the thioether functionality can be
also selectively aminoxylated (entries 2, 3, and 5).
Table 3. R-Aminoxylation of Aldehydes with
Polymer-Supported Catalyst 1 (10 mol %)a
Because one of the main advantages of using an insoluble
catalyst is the possibility of its recycling and reuse, we
explored the possibility of a repeated use of 1. To this end,
we filtered the resin after the aminoxylation reaction, washed
it several times with dichloromethane, and dried it in vacuum
before reuse. In this way, resin 1 could be recycled three
times without an appreciable loss in enantioselectivity or
yield (see Supporting Information).
In summary, we have developed suitable reaction condi-
tions for the use of 1 in the highly enantioselective R-ami-
noxylation of aldehydes and ketones. To the best of our
knowledge, this study reports the first example of an
insoluble and recyclable catalyst which is operative for the
considered reaction. Key to this behavior, resin 1 exhibits a
high rate balance between the catalytic R-aminoxylation and
the noncatalytic, subsequent reaction with nitrosobenzene
leading to O-N cleavage. From a practical perspective, the
use of 1 as a catalyst for the aminoxylation reaction allows
for significantly simplified workup conditions, the product
being obtained in most favorable cases after simple filtration
of the catalyst and elimination of the solvents. Moreover, if
the R-hydroxycarbonyl compound is the ultimate target of
the reaction, the insoluble nature of 1 is again advantageous
because the catalytic R-aminoxylation step and the noncata-
lytic O-N cleavage can be separated at will with the result
of increased selectivity.
a
Reaction conditions: resin 1 (10 mol %; 77.8 mg; f: 0.6 mmol/g);
b
Acknowledgment. We thank MEC (Grant CTQ2005-
02193/BQU), DURSI (Grant 2005SGR225), Consolider
Ingenio 2010 (Grant CSD2006-0003), and ICIQ Foundation
for financial support. D.F. is grateful to MEC for a
fellowship, and A.B. is grateful to MEC for a “Juan de la
Cierva” research contract.
aldehyde (3 equiv); nitrosobenzene (1 equiv); 2 mL of CHCl3; 0-4 °C.
Yield of isolated product. Determined by chiral HPLC (see Supporting
Information). Determined by H NMR.
c
d
1
hydride. It is assumed that the reduction step takes place
with quantitative yield and that the enantiomeric purity of 6
exactly reflects that of 5.1 As shown in Table 3, aldehydes
4a-h are selectively converted within 1-3 h to the corre-
sponding aminoxylated products 5a-h. This correlates well
with the higher reactivity expected for aldehydes than for
ketones (compare Table 3 with Table 2). Aminoxylated
primary alcohols (6a-g) are obtained generally in high yield
and with high enantioselectivities (96-99%). The lower yield
Supporting Information Available: Experimental pro-
cedures, characterization of new aminoxylated products, and
HPLC data. This material is available free of charge via the
Internet at http://pubs.acs.org.
OL070526P
1946
Org. Lett., Vol. 9, No. 10, 2007