Highly Regioselective Palladium-Catalyzed Carboxylation of Allylic Alcohols with CO2

Article abstract of DOI:10.1002/chem.201503359

Various allylic alcohols were carboxylated in the presence of a catalytic amount of PdCl₂ and PPh₃ using ZnEt₂ as a stoichiometric transmetalation agent under a CO₂ atmosphere (1atm). This carboxylation proceeded in a highly regioselective manner to afford branched carboxylic acids predominantly. The β,γ-unsaturated carboxylic acid thus obtained was successfully converted into an optically active γ-butyrolactone, a known intermediate of (R)-baclofen.

Full text of DOI:10.1002/chem.201503359

DOI: 10.1002/chem.201503359
Communication
&
Carboxylation
Highly Regioselective Palladium-Catalyzed Carboxylation of Allylic
Alcohols with CO2
Tsuyoshi Mita,*[a] Yuki Higuchi,[a] and Yoshihiro Sato*[a, b]
Abstract: Various allylic alcohols were carboxylated in the
presence of a catalytic amount of PdCl2 and PPh3 using
ZnEt2 as a stoichiometric transmetalation agent under
a CO2 atmosphere (1 atm). This carboxylation proceeded
in a highly regioselective manner to afford branched car-
boxylic acids predominantly. The b,g-unsaturated carboxyl-
ic acid thus obtained was successfully converted into an
optically active g-butyrolactone, a known intermediate of
(R)-baclofen.
Scheme 1. Umpolung strategy for p-allylpalladium.
nucleophilic, low-valent metals such as Zn,[4] InI,[5] SnCl2,[6]
[8]
[9]
SmI2,[7] and alkyl metal reagents including ZnEt2 and BEt3
Since carbon dioxide (CO2) is abundantly available on Earth, it
is an ideal C1 building block in organic synthesis; therefore,
many efforts to develop a method for catalytic/noncatalytic in-
corporation of CO2 have been made over the past decade.[1]
However, compared to other electrophiles such as aldehydes,
CO2 is much more stable and rather difficult to react with nu-
cleophiles with suppression of side reactions (e.g., protonation)
and/or decarboxylation that may occur at a high temperature.
One solution by our group to overcome these problems is the
use of reactive fluorostannate and silicate nucleophiles. By em-
ploying stannane and silane reagents in combination with CsF,
we have achieved one-pot synthesis of a-amino acids from
CO2 and imines, in which imino carbon, which possesses a posi-
tive charge in classical organic chemistry, smoothly reacts with
CO2.[2] This success has motivated us to develop another um-
polung carboxylation with CO2 using an appropriate nucleo-
philic metallic intermediate.
have been employed, and their use has resulted in coupling
with electrophiles such as aldehydes and their equivalents
(Scheme 1, [Eq. (2)]).[10] However, catalytic carboxylation of allyl-
ic alcohol derivatives with CO2, according to the above-men-
tioned umpolung strategy, has not yet been developed,[11] al-
though a nickel-catalyzed process[12] as well as a cobalt-cata-
lyzed[13] propargylic carboxylation were recently achieved.[14]
We herein disclose the first example of carboxylation of allylic
alcohol derivatives in the presence of a palladium catalyst and
ZnEt2. As a bonus, this catalytic system could utilize non-pro-
tected free allylic alcohols. Because a direct transformation of
allylic alcohols to b,g-unsaturated carboxylic acids using CO2
has not been achieved thus far, our method will provide a new
synthetic option to prepare these compounds in a single step
without preactivation of allylic alcohols (Scheme 2).
We then focused on the use of allylic alcohols as starting
materials. Allylic alcohols and their derivatives are known to
convert into p-allylpalladium intermediates in the presence of
Pd0.[3] Since this species generally acts as an electrophile
(Scheme 1, [Eq. (1)]),[3] switching of the polarity should be nec-
essary for reaction with CO2. To make p-allylpalladium species
Scheme 2. Our current work.
[a] Dr. T. Mita, Y. Higuchi, Prof. Dr. Y. Sato
Faculty of Pharmaceutical Sciences
Hokkaido University
Nishi 6, Kita 12, Kita-ku, Sapporo 060-0812 (Japan)
Fax: (+81)11-706-3722
E-mail: tmita@pharm.hokudai.ac.jp
biyo@pharm.hokudai.ac.jp
We first employed a typical allylic alcohol derivative, allylic
acetate 1a, to determine the feasibility of the planned umpo-
lung carboxylation using PdCl2 (10 mol%) and PPh3 (20 mol%)
in THF at RT for 16 h under 1 atm of CO2. Several low-valent re-
ductants (Mg, Mn, Zn, InI, SnCl2, and SmI2) and alkyl metal re-
agents (AlEt3, ZnMe2, ZnEt2, and BEt3) were examined,[15]
among which only ZnEt2 (3.0 equiv) promoted carboxylation to
afford the branched product 3a exclusively in 94% yield (96%
Homepage: http://gouka.pharm.hokudai.ac.jp/FSC/jpn/page/top_page.htm
[b] Prof. Dr. Y. Sato
ACT-C, Japan Science and Technology Agency (JST), Sapporo 060-0812
(Japan)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201503359.
Chem. Eur. J. 2015, 21, 16391 – 16394
16391
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
exclusively (Figure 1). g,g-Disubstituted allylic alcohol 2b and
b,g-disubstituted 2c smoothly underwent carboxylation to
afford branched carboxylic acids. Branched allylic alcohols (2d-
2i) with various functional groups such as an unprotected hy-
droxyl group, a TBSO group, and an ester were examined, and
all of them were carboxylated well to afford the corresponding
branched carboxylic acids in high yields. Notably, aryl bromide
(2i) could survive during this palladium-catalyzed process.[16] b-
Aryl-substituted allylic alcohols bearing electronically different
substituents (donating OMe, neutral Ph, and withdrawing Cl,
CF3 groups) at the para-position (2j–2m) were also applicable.
In addition, b-alkyl-substituted allylic alcohol 2n was moder-
ately reactive, and allyl alcohol 2o exhibited high reactivity.
Moreover, carboxylations of 2a and 2l were conducted on 6.0
and 4.0 mmol scales, respectively, affording the corresponding
carboxylic acids 4a and 4l in high yields without esterification
with TMSCHN2. In the reactions of 2a–2i, not even a trace
amount of linear carboxylic acids was observed in contrast to
the case of the nickel system.[12]
Table 1. Condition screening.
Entry
Substrate
Catalyst
[mol%]
Ligand
[mol%]
Yield [%][a]
rec 1a
3a
or 2a
1
1a
1a
1a
1a
1a
1a
1a
1a
2a
2a
2a
PdCl2 (10)
PdCl2 (5)
Pd(OAc)2 (10)
Pd(PPh3)4 (10)
PdCl2 (10)
PdCl2 (10)
PdCl2 (10)
NiCl2·DME (10)
PdCl2 (10)
PdCl2 (5)
PPh3 (20)
PPh3 (10)
PPh3 (20)
–
PCy3 (20)
PnBu3 (20)
dppe (10)
PPh3 (20)
PPh3 (20)
PPh3 (10)
–
94 (96)
–
2[b]
3
91 (92)
–
75
94
–
52
–
–
–
–
94
22
96
–
4
5
6
7
8
9[b,c]
10[c,d]
11[c]
93 (92)
95 (95)
84
–
–
–
Pd(PPh3)4 (10)
[a] Yields were determined by 1H NMR analysis using 1,1,2,2-tetrachloro-
ethane as an internal standard. Isolated yields are given in parentheses.
[b] Reaction time: 48 h. [c] 3.5 equiv of ZnEt2 was used. [d] Reaction time:
72 h. ZnEt2: 1m solution in hexane.
Taking advantage of the accessibility of allylic alcohols from
aldehydes, a one-pot carboxylation was performed, employing
3-phenylpropanal (5) as a starting material (Scheme 3, [Eq. (3)]).
After the generation of allylic alkoxide 6d by treatment of
vinyl Grignard reagent with 5 at RT, PdCl2, PPh3, and ZnEt2
were added accordingly under a CO2 atmosphere. The expect-
ed one-pot carboxylation thus proceeded smoothly to afford
3d in 88% yield.
Because lactols are regarded as masked aldehydes, they can
be converted into allylic alcohol derivatives by the attack of al-
kenyl nucleophiles. In fact, the reaction of 7 with vinyl
Grignard reagent produced allylic alkoxide 8 f, which under-
went the desired carboxylation to afford 3 f in 61% yield
isolated yield) after methyl esterification with TMSCHN2
(Table 1, entry 1). The reaction also proceeded with lower cata-
lyst loading (5 mol%), though a longer time was required
(48 h; entry 2). Although the use of Pd(OAc)2 instead of PdCl2
resulted in a lower yield (entry 3), the catalyst activity of
Pd(PPh3)4 was as good as that of PdCl2/PPh3 (entry 4). Lower re-
activity or no reaction was observed when other phosphine li-
gands including PCy3, PnBu3, and bidentate dppe were used
(entries 5 to 7; dppe=1,2-bis(diphenylphosphino)ethane).
When a nickel catalyst was em-
ployed, 1a was completely con-
sumed, but the desired 3a was
not obtained at all (entry 8). In-
terestingly, free allylic alcohol 2a
underwent carboxylation as well
using 3.5 equiv of ZnEt2 for 48 h,
and 3a was obtained in compa-
rable yield (93% yield) (entry 9).
Although
a longer time was
needed, the carboxylation was
completed in 72 h to afford 3a
in high yield in the presence of
5 mol% of PdCl2 (entry 10).
Pd(PPh3)4 was inferior to PdCl2/
PPh3 in the case of direct use of
allylic alcohols (entry 11).
Next, substrate scope for sev-
eral allylic alcohols was investi-
gated under the optimal condi-
tions using 10 mol% of PdCl2 Figure 1. Substrate scope of regioselective carboxylation of allylic alcohols (0.2 mmol). Isolated yields are shown
unless otherwise noted. [a] 6.0 mmol scale. Allylic alcohol 2a was recovered in 17% yield. [b] The reaction was
and 20 mol% of PPh3. Linear and
performed at 608C. [c] Allylic alcohol 2b was recovered in 18% yield. [d] The reaction was performed at 408C.
branched allylic alcohols were all
[e] 4.5 equiv of ZnEt2 was used. [f] Allylic alcohol 2j was recovered in 10% yield. [g] 4.0 mmol scale. [h] Allylic alco-
1
carboxylated regioselectively to
hol 2n was recovered in 13% yield. [i] Determined by H NMR analysis. (The isolated yield (77%) was diminished
afford branched carboxylic acids due to the volatility.)
Chem. Eur. J. 2015, 21, 16391 – 16394
www.chemeurj.org
16392
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
Scheme 5. Umpolung reactivity of p-allylpalladium toward CO2.
Based on the above experimental results and previous re-
ports, we propose a plausible catalytic cycle (Figure 2). First, al-
lylic alcohol 2 would be either activated by Lewis acidic ZnEt2
directly or by carbonate formation with CO2,[21] leading to the
generation of h3-allylpalladium II, which is in equilibrium to the
terminal h1-allylpalladium III. Either II or III would then be
transmetalated with ZnEt2. The resulting nucleophilic h1-allyl-
palladium IV possessing ethyl and phosphine ligands would
attack CO2 at the g-position.[10] When optically active branched
alcohol 2d (69% ee) was subjected to the carboxylation, the
product 3d was obtained as a racemic form,[15] indicating that
the carboxylation occurs from the terminal h1-allylpalladium IV
that contains no chiral information. Recent studies on palladi-
um-catalyzed enantioselective allylations of aldehydes using
chiral phosphine ligands[22] have suggested that h1-allylethyl-
palladium species[8h] bearing a chiral ligand are involved in the
stereoselective allylation rather than h1-allylzinc species V origi-
nally proposed by Tamaru.[8a] On the basis of this information,
our carboxylation may also proceed from h1-allylethylpalladium
IV rather than from h1-allylzinc V, because we employed similar
reaction conditions[8,22a,c,d,f] except for the use of CO2 instead of
aldehydes. After the nucleophilic attack to CO2, palladium car-
boxylate VI would be formed, and it could be reduced by
ZnEt2 to regenerate Pd0 along with the release of zinc carbox-
ylate VII.
Scheme 3. One-pot reaction from an aldehyde or a lactol.
under optimal conditions without the need to purify 8 f
(Scheme 3, [Eq. (4)]).
We next turned our attention to the synthesis of g-butyro-
lactones, which are structural scaffolds for a variety of biologi-
cally active natural products as well as useful building blocks
in organic synthesis.[17] Because b,g-unsaturated carboxylic
acids have two functional groups, a carboxyl group and a vinyl
group, at appropriate positions, cationic cyclizations would be
possible for preparing these compounds. The product 4l was
thus converted into butenolide 9 by treatment of PhI(OTf)2 de-
rived from PhI(OAc)2 and TMSOTf in situ (Scheme 4).[18] Enantio-
selective conjugate reduction of 9 catalyzed by Stryker’s re-
agent in combination with (R)-DTBM-SEGPHOS proceeded effi-
ciently to afford optically active g-butyrolactone 10 quantita-
tively with 99% ee.[19] Lactone 10 is a known intermediate for
the practical asymmetric synthesis of (R)-baclofen,[20] a GABAB
agonist used for treatment of spasticity caused by disease of
the spinal cord.
To confirm the nucleophilicity of p-allylpalladium, [(h3-al-
ly)PdCl]2 was treated with ZnEt2 in the absence or presence of
PPh3 under a CO2 atmosphere (Scheme 5). Without PPh3, the
carboxylation did not proceed at all. In contrast, with PPh3, the
desired carboxylation proceeded to afford 4o in 36% yield,
suggesting that a phosphine ligand is necessary for the car-
boxylation.
In summary, we have developed the first palladium-cata-
lyzed carboxylation of allylic alcohols. The successful use of al-
lylic alcohols enhanced the practicality of this carboxylation.
The b,g-unsaturated carboxylic acid thus obtained could be
converted into an optically active g-butyrolactone. Efforts
Scheme 4. Synthetic application of the product.
Figure 2. A plausible catalytic cycle.
Chem. Eur. J. 2015, 21, 16391 – 16394
www.chemeurj.org
16393
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
[9] a) M. Kimura, I. Kiyama, T. Tomizawa, Y. Horino, S. Tanaka, Y. Tamaru, Tet-
rahedron Lett. 1999, 40, 6795; b) M. Kimura, T. Tomizawa, Y. Horino, S.
Tanaka, Y. Tamaru, Tetrahedron Lett. 2000, 41, 3627; c) M. Shimizu, M.
Kimura, T. Watanabe, Y. Tamaru, Org. Lett. 2005, 7, 637.
toward the development of an asymmetric variant are actively
ongoing.
[10] For reviews of nucleophilic p-allylpalladium species, see: a) Y. Tamaru, J.
Organomet. Chem. 1999, 576, 215; b) J. A. Marshall, Chem. Rev. 2000,
100, 3163; c) K. J. Szabó, Chem. Eur. J. 2004, 10, 5268; d) K. J. Szabó, Syn-
lett 2006, 811; e) G. Zanoni, A. Pontiroli, A. Marchetti, G. Vidari, Eur. J.
Org. Chem. 2007, 3599.
Acknowledgements
This work was financially supported by Grants-in-Aid for Scien-
tific Research (C) (No. 26410108) and Grants-in-Aid for Scientific
Research (B) (No. 26293001) from JSPS and also by a Grant-in-
Aid for Scientific Research on Innovative Areas “Molecular Acti-
vation Directed toward Straightforward Synthesis (No.
25105701)” from MEXT.
[11] h1-Allylpalladium species are known to react with CO2. From allylstan-
nane and boron, see: a) M. Shi, K. M. Nicholas, J. Am. Chem. Soc. 1997,
119, 5057; b) R. Johansson, O. F. Wendt, Dalton Trans. 2007, 488; c) D. P.
Hruszkewycz, J. Wu, N. Hazari, C. D. Incarvito, J. Am. Chem. Soc. 2011,
133, 3280; d) J. Wu, N. Hazari, Chem. Commun. 2011, 47, 1069; e) H. A.
Duong, P. B. Huleatt, Q.-W. Tan, E. L. Shuying, Org. Lett. 2013, 15, 4034.
From bis(h3-allyl)palladium by the attack of a monodentate ligand, see:
f) T. Hung, P. W. Jolly, G. Wilke, J. Organomet. Chem. 1980, 190, C5; g) J.
Wu, J. C. Green, N. Hazari, D. P. Hruszkewycz, C. D. Incarvito, T. J. Schmei-
er, Organometallics 2010, 29, 6369. From hydropalladation of 1,2- and
1,3-dienes by the palladium pincer complex, see: h) J. Takaya, N. Iwasa-
wa, J. Am. Chem. Soc. 2008, 130, 15254; i) J. Takaya, K. Sasano, N. Iwasa-
wa, Org. Lett. 2011, 13, 1698; j) H.-W. Suh, L. M. Guard, N. Hazari, Chem.
Sci. 2014, 5, 3859.
Keywords: allylic
compounds
·
carbon
dioxide
·
carboxylation · carboxylic acids · umpolung
[1] For recent reviews on CO2 incorporation reactions, see: a) T. Sakakura,
J.-C. Choi, H. Yasuda, Chem. Rev. 2007, 107, 2365; b) M. Mori, Eur. J. Org.
Chem. 2007, 4981; c) S. N. Riduan, Y. Zhang, Dalton Trans. 2010, 39,
3347; d) M. Cokoja, C. Bruckmeier, B. Rieger, W. A. Herrmann, F. E. Kühn,
Angew. Chem. Int. Ed. 2011, 50, 8510; Angew. Chem. 2011, 123, 8662;
e) Y. Tsuji, T. Fujihara, Chem. Commun. 2012, 48, 9956; f) L. Zhang, Z.
Hou, Chem. Sci. 2013, 4, 3395; g) N. Kielland, C. J. Whiteoak, A. W. Kleij,
Adv. Synth. Catal. 2013, 355, 2115; h) X. Cai, B. Xie, Synthesis 2013, 45,
3305; i) C. Maeda, Y. Miyazaki, T. Ema, Catal. Sci. Technol. 2014, 4, 1482;
j) B. Yu, L.-N. He, ChemSusChem 2015, 8, 52.
[2] a) T. Mita, J. Chen, M. Sugawara, Y. Sato, Angew. Chem. Int. Ed. 2011, 50,
1393; Angew. Chem. 2011, 123, 1429; b) T. Mita, M. Sugawara, H. Hase-
gawa, Y. Sato, J. Org. Chem. 2012, 77, 2159; c) T. Mita, J. Chen, M. Suga-
wara, Y. Sato, Org. Lett. 2012, 14, 6202; d) T. Mita, Y. Higuchi, Y. Sato,
Chem. Eur. J. 2013, 19, 1123; e) T. Mita, Y. Sato, J. Synth. Org. Chem. Jpn.
2013, 71, 1163.
[3] a) J. Tsuji, Palladium Reagents and Catalysts: New Perspectives for the
21st Century, Wiley, Chichester, 2004; b) B. M. Trost, M. L. Crawley, Chem.
Rev. 2003, 103, 2921.
[4] Y. Masuyama, N. Kinugawa, Y. Kurusu, J. Org. Chem. 1987, 52, 3702.
[5] a) S. Araki, T. Kamei, T. Hirashita, H. Yamamura, M. Kawai, Org. Lett. 2000,
2, 847; b) T. Hirashita, T. Kamei, M. Satake, T. Horie, H. Shimizu, S. Arai,
Org. Biomol. Chem. 2003, 1, 3799.
[12] T. Moragas, J. Cornella, R. Martin, J. Am. Chem. Soc. 2014, 136, 17702.
[13] K. Nogi, T. Fujihara, J. Terao, Y. Tsuji, Chem. Commun. 2014, 50, 13052.
[14] In- and Zn-promoted carboxylations of allylic halides were also devel-
oped by Ma, see: a) B. Miao, S. Ma, Chem. Commun. 2014, 50, 3285;
b) B. Miao, S. Ma, Org. Chem. Front. 2015, 2, 65.
[15] See the Supporting Information for details.
[16] Carboxylation of aryl bromides catalyzed by Pd/tBuXPhos using ZnEt2
has already been reported, see: A. Correa, R. Martin, J. Am. Chem. Soc.
2009, 131, 15974.
[17] a) Y. S. Rao, Chem. Rev. 1964, 64, 353; b) Y. S. Rao, Chem. Rev. 1976, 76,
625.
[18] a) K. Kiyokawa, S. Yahata, T. Kojima, S. Minakata, Org. Lett. 2014, 16,
4646; b) F. V. Singh, J. Rehbein, T. Wirth, ChemistryOpen 2012, 1, 245.
[19] B. H. Lipshutz, J. M. Servesko, B. R. Taft, J. Am. Chem. Soc. 2004, 126,
8352.
[20] a) A. Schoenfelder, A. Mann, S. L. Coz, Synlett 1993, 63; b) C. Mazzini, J.
Lebreton, V. Alphand, R. Furstoss, Tetrahedron Lett. 1997, 38, 1195.
[21] M. Sakamoto, I. Shimizu, A. Yamamoto, Bull. Chem. Soc. Jpn. 1996, 69,
1065.
[22] a) G. Zanoni, S. Gladiali, A. Marchetti, P. Piccinini, I. Tredici, G. Vidari,
Angew. Chem. Int. Ed. 2004, 43, 846; Angew. Chem. 2004, 116, 864; b) S.-
F. Zhu, Y. Yang, L.-X. Wang, B. Liu, Q.-L. Zhou, Org. Lett. 2005, 7, 2333;
c) G. P. Howell, A. J. Minnaard, B. L. Feringa, Org. Biomol. Chem. 2006, 4,
1278; d) W. Wang, T. Zhang, M. Shi, Organometallics 2009, 28, 2640;
e) X.-C. Qiao, S.-F. Zhu, W.-Q. Chen, Q.-L. Zhou, Tetrahedron: Asymmetry
2010, 21, 1216; f) J.-J. Jiang, D. Wang, W.-F. Wang, Z.-L. Yuan, M.-X.
Zhao, F.-J. Wang, M. Shi, Tetrahedron: Asymmetry 2010, 21, 2050; g) S.-F.
Zhu, X.-C. Qiao, Y.-Z. Zhang, L.-X. Wang, Q.-L. Zhou, Chem. Sci. 2011, 2,
1135.
[6] a) Y. Masuyama, J. P. Takahara, Y. Kurusu, J. Am. Chem. Soc. 1988, 110,
4473; b) J. P. Takahara, Y. Masuyama, Y. Kurusu, J. Am. Chem. Soc. 1992,
114, 2577.
[7] T. Tabuchi, J. Inanaga, M. Yamaguchi, Tetrahedron Lett. 1986, 27, 1195.
[8] a) K. Yasui, Y. Goto, T. Yajima, Y. Taniseki, K. Fugami, A. Tanaka, Y. Tamaru,
Tetrahedron Lett. 1993, 34, 7619; b) Y. Tamaru, A. Tanaka, K. Yasui, S.
Goto, S. Tanaka, Angew. Chem. Int. Ed. Engl. 1995, 34, 787; Angew. Chem.
1995, 107, 862; c) M. Shimizu, M. Kimura, S. Tanaka, Y. Tamaru, Tetrahe-
dron Lett. 1998, 39, 609; d) M. Kimura, Y. Ogawa, M. Sihimizu, M. Sueishi,
S. Tanaka, Y. Tamaru, Tetrahedron Lett. 1998, 39, 6903; e) T. Sakamoto, K.
Takahashi, T. Yamazaki, T. Kitazume, J. Org. Chem. 1999, 64, 9467; f) M.
Kimura, M. Shimizu, K. Shibata, M. Tazoe, Y. Tamaru, Angew. Chem. Int.
Ed. 2003, 42, 3392; Angew. Chem. 2003, 115, 3514; g) M. Kimura, M. Shi-
mizu, S. Tanaka, Y. Tamaru, Tetrahedron 2005, 61, 3709; h) A. Flahaut, K.
Toutah, P. Mangeney, S. Roland, Eur. J. Inorg. Chem. 2009, 5422.
Received: August 25, 2015
Published online on September 29, 2015
Chem. Eur. J. 2015, 21, 16391 – 16394
www.chemeurj.org
16394
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim