Enantioselective synthesis of 2,2-disubstituted tetrahydrofurans: Palladium-catalyzed [3+2] cycloadditions of trimethylenemethane with ketones

Article abstract of DOI:10.1002/anie.201300616

O rings: An approach to the title compounds has been developed utilizing a cycloaddition of trimethylenemethane with aryl ketones. The products are formed in up to a 96 % yield with 95 % ee. The reaction is catalyzed by palladium in the presence of L1, which possesses a stereogenic phosphorus atom, and only a single epimer at the phosphorus atom yields the active catalyst. Cp=cyclopentadiene, TMS=trimethylsilyl. Copyright

Full text of DOI:10.1002/anie.201300616

Angewandte
Chemie
DOI: 10.1002/anie.201300616
Asymmetric Catalysis
Enantioselective Synthesis of 2,2-Disubstituted Tetrahydrofurans:
Palladium-Catalyzed [3+2] Cycloadditions of Trimethylenemethane
with Ketones**
Barry M. Trost* and Dustin A. Bringley
Found throughout nature and in synthetic products, tetrahy-
drofurans are a ubiquitous structural motif with wide-reach-
ing biological activity. Among the many approaches towards
their synthesis,[1] cycloaddition represents a powerful and
attractive strategy,[2] but catalytic enantioselective variants
remain rare.[3,4] We recently disclosed one such approach
involving the palladium-catalyzed cycloaddition of trimethy-
lenemethane (TMM) with aldehydes.[4c] Herein, we describe
the surprising success of this method in reactions with
aromatic ketones, a substrate class which was previously
unknown in Pd/TMM reactions using achiral ligands.
more difficult than with the corresponding aldehydes. Never-
theless, such a reaction would provide rapid access to
enantioenriched 2,2-disubstituted tetrahydrofurans, an impor-
tant structural component[14] that is found, for example, in the
antifungal drug posaconazole (1, Figure 1).[15]
First discovered over 30 years ago,[5] the palladium-
catalyzed TMM cycloaddition is a versatile method for the
synthesis of five-membered rings.[6] The recent discovery that
bulky phosphoramidites are effective chiral ligands has
stimulated a fresh examination of Pd/TMM chemistry, thus
leading to highly enantio- and diastereoselective cycloaddi-
tions with olefins,[7] imines,[8] and tropones.[9] A hallmark of
the phosphoramidites used in these studies is an improved
catalytic activity when compared to the achiral ligands used
previously, thus allowing reactions with highly substituted
acceptors such as tetrasubstituted olefins[7b] and ketimines[8b,c]
under mild reaction conditions. However, these substrates
required a TMM precursor bearing a cyano group to obtain
high ee values.[10] Given our earlier success with cycloaddi-
tions of aldehydes, we wondered whether more complex and
less reactive carbonyl groups might also function in the
asymmetric TMM cycloaddition, despite the fact that racemic
reactions with ketones are relatively challenging and quite
limited.[11,12] The reduced reactivity of the ketone coupled
with the presence of relatively acidic protons[13] undoubtedly
prevents a more general synthetic scope. Further, aliphatic
aldehydes were not tolerated under the asymmetric reaction
conditions, presumably because of the presence of a protons.
An additional challenge involves enantioselectivity, since
enantioface differentiation of ketones was expected to be
Figure 1. Structure of the antifungal agent posaconazole (1).
Despite these concerns, initial tests using acetophenone
under our previously optimized reaction conditions[4c] were
encouraging, thus giving the desired cycloadduct 4 in mod-
erate yield but with a reasonable ee value (Table 1, entry 1;
see Figure2 for ligand structures). Notably, the enantioselec-
tivity was roughly equivalent to that obtained in the reaction
of benzaldehyde, although the yield was significantly lower.
Also consistent with our earlier study, a Lewis acid cocatalyst
was not required (entry 2). While the product obtained from
these reactions was qualitatively clean as judged by 1H NMR
Table 1: Initial optimization of reaction conditions with acetophenone.
Entry[a]
Catalyst
Ligand
T [8C]
Yield [%]
ee [%]
1[b]
2
3
[Pd(dba)2]
[Pd(dba)2]
L1
L1
L1
L1
L1
L2
L3
L4
50
50
50
23
50
50
50
50
48
60
91
70
50
<20
<20
0
80[c]
84[c]
78
80
78
71
55
–
[CpPd(h3-C3H5)]
[CpPd(h3-C3H5)]
[CpPd(h3-C3H5)]
[CpPd(h3-C3H5)]
[CpPd(h3-C3H5)]
[CpPd(h3-C3H5)]
4
[*] Prof. B. M. Trost, Dr. D. A. Bringley
Department of Chemistry, Stanford University
Stanford, CA 94305-5080 (USA)
5[d]
6
7
8
E-mail: bmtrost@stanford.edu
[**] We thank the NSF (CHE-1145236) and the NIH (GM 033049) for
their generous support of our programs. Additional financial
support was provided by the W. S. Johnson Graduate Fellowship
(D.A.B.). Palladium salts were a generous gift from Johnson
Matthey. We thank Dr. Allen Oliver at the University of Notre Dame
for X-ray crystal analyses.
[a] All reactions were conducted using 5 mol% Pd catalyst, 10 mol%
ligand, and 1.6 equiv of 2 at 0.15m in toluene unless stated otherwise.
Yields are for isolated products. The ee values were determined by HPLC
using a chiral stationary phase. [b] Reaction conducted in the presence of
10 mol% [In(acac)3]. [c] A minor impurity in the HPLC trace prevented
an accurate determination of the ee value. [d] Used only 5 mol% L1.
acac=acetylacetonate, Cp=cyclopentadiene, TMS=trimethylsilyl.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201300616.
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
Table 2: Substrate scope with aryl ketones.
Entry[a] Substrate
Product
Yield
[%]
ee
[%]
1[b]
70
71
88
80
86
95
Figure 2. Ligands used during optimization of reaction conditions.
2
spectroscopy, the HPLC traces contained a small impurity
which prevented an accurate determination of the ee value.
Reasoning that this impurity might stem from competitive
cycloaddition with dibenzylideneacetone (dba), we explored
alternative palladium sources. Gratifyingly, [CpPd(h3-C3H5)]
not only gave an improved purity profile, but also provided
the desired product in 91% yield (entry 3). The reaction
temperature could be reduced to 238C (entry 4), and a slight
loss in yield and no significant improvement in the ee value
indicated that reaction at 508C would be preferred. Finally,
cutting the amount of ligand in half produced a commensurate
reduction in yield (entry 5). While the yield was modest, the
observation that the selectivity remained unchanged provides
the first experimental evidence that the Pd/ligand ratio in the
active catalyst may be 1:1.
3
4[b,c]
60
82
47
5
6
69
74
(87)[d]
89
We briefly examined other ligands under these optimized
reaction conditions, but found that L1 is remarkably unique in
both catalytic activity and enantioselectivity. For example,
while the parent phosphoramidite L2 gave a slight decrease in
ee value, the product was formed in very poor yield, with
unreacted starting material as the predominant species
(entry 6). Simple 3-substituted binol derivatives such as L3
are likewise insufficient (entry 7). Finally, employing the
dimethylamine-derived phosphoramidite L4 gave a com-
pletely inactive catalyst (entry 8).
7
8
9
96
60
68
94
86
88
Having identified optimized reaction conditions, we next
evaluated the scope of the transformation (Table 2). Surpris-
ingly, reactions of ketones with larger aliphatic substituents,
such as propiophenone (entry 2) and 4’-methoxyisobutyro-
phenone (entry 3), gave slightly better ee values than the
corresponding acetophenone derivatives. Chemoselective
reaction with the carbonyl group of an enone was also
possible. Reaction of the isobutenyl ketone 9 at 508C
provided a 16:47:25 mixture of the bis(cycloadduct), tetrahy-
drofuran, and cyclopentane, respectively. Since the bis(cy-
cloadduct) necessarily arises by initial cycloaddition with the
olefin, this result constitutes only a slight preference for
cycloaddition with the carbonyl group under these reaction
conditions. However, the tetrahydrofuran was isolated in
60% yield and 82% ee when the reaction was conducted at
room temperature (entry 4), along with the cyclopentane in
20% yield and 79% ee and no detectable presence of the
bis(cycloadduct). The chemoselectivity in this example, while
relatively modest, is impressive considering that reactions of
enones employing L2 provided exclusive reaction with the
olefin.[7c] Compared to previously reported multistep synthe-
10
87
83
11
12
13
71
77
62
78
92
87
[a] All reactions were conducted for 3 h at 0.15m in toluene at 508C with
1.6 equivalents 1, 5 mol% [CpPd(h3-C3H5)], and 10 mol% L1 unless
stated otherwise. Yields are for isolated products. The ee values were
determined by HPLC using a chiral stationary phase. [b] Reaction
conducted at 238C for 24 h. [c] Used 3.0 equivalents of 1. [d] The
ee value obtained after a single recrystallization. The overall yield after
recrystallization was 31% from 11.
2
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
ses of the 2,2,4-trisubstituted tetrahydrofuran core of pos-
aconazole (1),[16] rapid access was obtained by employing the
commercially available triazolyl derivative 11 (entry 5).
While the selectivity was modest, the ee value was improved
to 87% after a single recrystallization, thus providing the
cycloadduct 12 in 31% overall yield from 11.
of the dihydrofuran may form a secondary interaction with
the palladium, which may help explain the improved catalytic
performance of L1 relative to L2. These secondary interac-
tions should also favor the formation of a monoligated
palladium/phosphoramidite complex, as was hypothesized
from the optimization studies (Table 1).
The reaction also proved to be relatively insensitive to the
aromatic portion of the ketone, including substitution pattern
and electronic nature of the substituent (Table 2, entries 6–
12). Heterocycles were also tolerated (entry 13). An X-ray
crystal structure of 22 unambiguously allowed the determi-
nation of the absolute stereochemistry (Figure 3). It should be
In summary, we have demonstrated a novel palladium-
catalyzed [3+2] cycloaddition of trimethylenemethane with
ketones. This reaction demonstrates novel reactivity which
has not been previously observed and provides access to
highly enantioenriched tetrahydrofurans bearing a tetrasub-
stituted stereocenter. An example utilizing an a,b-unsatu-
rated ketone also demonstrates a modest preference (3:1) for
reaction with the carbonyl group. A critical factor enabling
this reaction was the development of a C1-symmetric phos-
phoramidite which demonstrates uniquely high activity under
these reaction conditions, wherein the epimer at phosphorus
relative to the chiral scaffold that was uniquely reactive was
identified as the R,R,R,SP isomer.
Experimental Section
Figure 3. X-ray-based ORTEP drawing of cycloadduct 22. Thermal
Representative procedure for the synthesis of (R)-2-(4’-bromobi-
phenyl-4-yl)-2-methyl-4-methylenetetrahydrofuran (22): Toluene
(1.0 mL) was added to an argon-purged vial of 4-acetyl-4’-bromo-
ellipsoids are drawn at the 50% probability level.
biphenyl
(41.3 mg,
0.15 mmol),
[CpPd(h3-C3H5)]
(1.6 mg,
noted that the sense of chirality is consistent with that
observed in both aldehyde and nitroalkene cycloaddi-
tions.[4,7d]
0.0075 mmol), and L1 (10.2 mg, 0.015 mmol) and the solution stirred
for 2 min before 2-[(trimethylsilyl)methyl]allyl acetate (50 mL,
0.24 mmol) was added. The vial was immediately immersed in an
oil bath set to 508C and stirred for 3 h. It was then purified directly by
flash chromatography (4% ethyl acetate in hexanes) to give a white
solid (42.7 mg, 87% yield). 1H NMR (400 MHz, CDCl3): d = 7.55–
7.43 (m, 8H), 4.97 (quintet, J = 2.2 Hz, 1H), 4.87 (quintet, J = 2.2 Hz,
1H), 4.51 (d, J = 13.4 Hz, 1H), 4.39 (d, J = 13.4 Hz, 1H), 2.94 (d, J =
15.3 Hz, 1H), 2.77 (d, J = 15.3 Hz, 1H), 1.57 ppm (s, 3H). 13C NMR
(100 MHz, CDCl3): d = 148.2, 146.4, 140.1, 138.7, 132.2, 128.9, 127.1,
The unique efficiency of L1 in this transformation remains
impressive. Because the binol derivative in L1 is unsymmetric,
the ligand is chiral at phosphorus (RP and SP). In our previous
work,[4] we found that only a single diastereomer of L1 was
active. Thus, the chirality at phosphorus serves as a unique
control element for the reactivity of the catalyst, a phenom-
enon which to the best of our knowledge has not been
previously described in the chemical literature. By appropri-
ate selection of the reaction conditions,[4] we were able to
synthesize L1 in < 1: > 20 d.r., and the major (and active)
diastereomer appears to possess the R,R,R,SP configuration
based on its 31P NMR spectrum, since the major signal (at d =
144.2 ppm) appears upfield of the minor signal (at d =
148.0 ppm).[17] Although we were unsuccessful in crystallizing
L1 directly, this assignment is corroborated by X-ray crystal
analysis of L4 (see the Supporting Information), which clearly
depicts the SP configuration and also shows the major
diastereomer (at d = 148.9) upfield of the minor diastereomer
(at d = 149.6) in the 31P NMR spectrum.
~
125.8, 121.7, 105.1, 84.8, 70.4, 46.4, 29.4 ppm. IR (thin film): n ¼3074,
D
2911, 2857, 1662, 1429, 1384 cmꢀ1. ½aꢁ23¼ꢀ43.2 (c = 1.70, CHCl3).
HPLC: Chiralpak AD-H, 0.8 mLminꢀ1, 1% iPrOH in heptane, l =
254 nm, tR,minor = 9.1 min, tR,major = 12.4 min. Elemental analysis (%):
calcd for C18H17BrO: C 65.67, H 5.20, Br 24.27; found: C 65.42,
H 5.47, Br 24.37.
Received: January 24, 2013
Published online: && &&, &&&&
Keywords: asymmetric catalysis · cycloadditions · heterocycles ·
.
palladium · phosphoramidite ligands
For a possible explanation as to why only the SP isomer of
L1 is catalytically active, we performed AM1 semi-empirical
calculations using Spartan to determine the lowest-energy
conformations for the respective diastereomers. The results
suggest that the ground-state geometries are substantially
different, and that only the SP isomer is poised to achieve
secondary bonding interactions between the palladium metal
and a naphthyl group from the pyrrolidine,[18] as has been
previously observed in p-allylpalladium/phosphoramidite
complexes.[19] The need for such interactions would be in
accordance with the observation that L4 is unreactive in the
present system. The models also suggest that the oxygen atom
[1] a) J. P. Wolfe, M. B. Hay, Tetrahedron 2007, 63, 261; b) G. Jalce,
X. Franck, B. Figadꢀre, Tetrahedron: Asymmetry 2009, 20, 2537.
[2] a) Cycloaddition Reactions in Organic Synthesis (Eds.: S.
Kobayashi, K. A. Jørgensen), Wiley-VCH, Weinheim, 2002;
b) S. Muthusamy, J. Krishnamurthi, Top. Heterocycl. Chem.
2008, 12, 147.
[3] For catalytic, enantioselective syntheses of bicyclic tetrahydro-
furans, see: a) H. Suga, K. Inoue, S. Inoue, A. Kakehi, J. Am.
Chem. Soc. 2002, 124, 14836; b) D. M. Hodgson, R. Glen, G. H.
Grant, A. J. Redgrave, J. Org. Chem. 2003, 68, 581; c) N.
Shimada, M. Anada, S. Nakamura, H. Nambu, H. Tsutsui, S.
Hashimoto, Org. Lett. 2008, 10, 3603; d) K. Ishida, H. Kusama,
N. Iwasawa, J. Am. Chem. Soc. 2010, 132, 8842; e) N. Z. Burns,
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
M. R. Witten, E. N. Jacobsen, J. Am. Chem. Soc. 2011, 133,
14578.
originally used as evidence for the intermediacy of Pd/TMM. See
B. M. Trost, D. M. T. Chan, J. Am. Chem. Soc. 1983, 105, 2326.
[14] a) R. Baker, N. R. Curtis, J. M. Elliott, T. Harrison, G. J.
Hollingworth, P. S. Jackson, J. J. Kulagowski, N. M. Rupniak,
E. M. Seward, C. J. Swain, B. J. Williams, U.S. Patent 6,613,765,
Sep 2, 2003; b) C. J. Dutton, B. J. Banks, C. B. Cooper, Nat. Prod.
Rep. 1995, 12, 165 – 181.
[4] For catalytic, enantioselective syntheses of monocyclic tetrahy-
drofurans, see: a) B. M. Trost, C. Jiang, J. Am. Chem. Soc. 2001,
123, 12907; b) A. T. Parsons, J. S. Johnson, J. Am. Chem. Soc.
2009, 131, 3122; c) B. M. Trost, D. A. Bringley, S. M. Silverman,
J. Am. Chem. Soc. 2011, 133, 7664.
[5] B. M. Trost, D. M. T. Chan, J. Am. Chem. Soc. 1979, 101, 6429.
[6] a) S. Yamago, E. Nakamura, Org. React. 2003, 61, 1; b) also see
Ref. [3].
[7] a) B. M. Trost, J. P. Stambuli, S. M. Silverman, U. Schwçrer, J.
Am. Chem. Soc. 2006, 128, 13328; b) B. M. Trost, N. Cramer,
S. M. Silverman, J. Am. Chem. Soc. 2007, 129, 12396; c) B. M.
Trost, S. M. Silverman, J. P. Stambuli, J. Am. Chem. Soc. 2011,
133, 19483; d) B. M. Trost, D. A. Bringley, P. S. Seng, Org. Lett.
2012, 14, 234.
[8] a) B. M. Trost, S. M. Silverman, J. P. Stambuli, J. Am. Chem. Soc.
2007, 129, 12398; b) B. M. Trost, S. M. Silverman, J. Am. Chem.
Soc. 2010, 132, 8238; c) B. M. Trost, S. M. Silverman, J. Am.
Chem. Soc. 2012, 134, 4941.
[9] B. M. Trost, P. J. McDougall, O. Hartman, P. Wathen, J. Am.
Chem. Soc. 2008, 130, 14960.
[15] a) F. Bennett, A. K. Saksena, R. G. Lovey, Y. T. Liu, N. M. Patel,
P. Pinto, R. Pike, E. Jao, V. M. Girijavallabhan, A. K. Ganguly,
D. Loebenberg, H. Wang, A. Cacciapuoti, E. Moss, F. Menzel,
R. S. Hare, A. Nomeir, Bioorg. Med. Chem. Lett. 2006, 16, 186;
b) C. A. Kauffman, A. N. Malani, C. Easley, P. Kirkpatrick, Nat.
Rev. Drug Discovery 2007, 6, 183.
[16] a) A. K. Saksena, V. M. Girijavallabhan, R. G. Lovey, R. E. Pike,
J. A. Desai, A. K. Ganguly, R. S. Hare, D. Loebenberg, A.
Cacciapuoti, R. M. Parmegiani, Bioorg. Med. Chem. Lett. 1994,
4, 2023; b) A. K. Saksena, V. M. Girijavallabhan, R. G. Lovey,
R. E. Pike, H. Wang, A. K. Ganguly, B. Morgan, A. Zaks, M. S.
Puar, Tetrahedron Lett. 1995, 36, 1787; c) A. K. Saksena, V. M.
Girijavallabhan, H. Wang, Y. T. Liu, R. E. Pike, A. K. Ganguly,
Tetrahedron Lett. 1996, 37, 5657; d) B. Morgan, D. R. Dodds, A.
Zaks, D. R. Andrews, R. Klesse, J. Org. Chem. 1997, 62, 7736;
e) R. G. Lovey, A. K. Saksena, V. M. Girijavallabhan, P. Blun-
dell, H. Guzik, D. Loebenberg, R. M. Parmegiani, A. Caccia-
puoti, Bioorg. Med. Chem. Lett. 2002, 12, 1739.
[10] Unpublished results from our laboratories. Reactions of alkyli-
dene oxindoles and ketimines with the parent donor 2 gave low
ee values.
[11] B. M. Trost, S. A. King, T. Schmidt, J. Am. Chem. Soc. 1989, 111,
5902.
[17] M. T. Reetz, J. A. Ma, R. Goddard, Angew. Chem. 2005, 117,
416; Angew. Chem. Int. Ed. 2005, 44, 412.
[12] a) B. M. Trost, S. Sharma, T. Schmidt, J. Am. Chem. Soc. 1992,
114, 7903; b) B. M. Trost, S. Sharma, T. Schmidt, Tetrahedron
Lett. 1993, 34, 7183.
[18] See the Supporting Information for details.
[19] I. S. Mikhel, H. Rꢁegger, P. Butti, F. Camponovo, D. Huber, A.
Mezzetti, Organometallics 2008, 27, 2937.
[13] The Pd/TMM intermediate can also serve as a base. Deproto-
nation and subsequent methallylation of acetophenone was
4
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Asymmetric Catalysis
B. M. Trost,*
D. A. Bringley
&&&& — &&&&
Enantioselective Synthesis of 2,2-
Disubstituted Tetrahydrofurans:
Palladium-Catalyzed [3+2]
Cycloadditions of Trimethylenemethane
with Ketones
O rings: An approach to the title com-
pounds has been developed utilizing
a cycloaddition of trimethylenemethane
with aryl ketones. The products are
formed in up to a 96% yield with 95% ee.
The reaction is catalyzed by palladium in
the presence of L1, which possesses
a stereogenic phosphorus atom, and only
a single epimer at the phosphorus atom
yields the active catalyst. Cp=cyclopen-
tadiene, TMS=trimethylsilyl.
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!