Palladium-catalyzed cascade reaction for the synthesis of substituted isoindolines

Article abstract of DOI:10.1002/anie.201008160

Arylate then cyclize: A palladium(II)-catalyzed cascade sequence has been developed to provide highly diastereomerically enriched cis-1-aryl-3-vinyl isoindolines (see scheme). The method uses commercially available aryl boronic acids and boroxine compounds containing a variety of electron-rich, -neutral, or -poor aromatic groups. Ts=4-toluenesulfonyl. Copyright

Full text of DOI:10.1002/anie.201008160

DOI: 10.1002/anie.201008160
Cascade Reactions
Palladium-Catalyzed Cascade Reaction for the Synthesis of Substituted
Isoindolines**
Florence J. Williams and Elizabeth R. Jarvo*
Isoindoline heterocycles have demonstrated potential in
medicinal chemistry as they exhibit activity across diverse
biological targets. They are present in molecules which act as
bronchodilaters, N-methyl-d-aspartate agonists, multidrug
resistance reversal agents, and fibrinogen receptor antago-
nists.[1] While several approaches to the synthesis of unsub-
stituted or monosubstituted isoindolines have been
reported,[2] few methods exist to produce disubstituted
isoindolines with high diastereoselectivity.[3] In addition,
there are no diastereoselective methods for the synthesis of
1,3-disubstituted isoindolines that allow for incorporation of
readily available boronic acids, which are practical building
blocks in medicinal chemistry. Synthetic methods for isoindo-
line synthesis that provide straightforward introduction of
substitutents on the heterocycle would enable preparation of
families of biologically significant compounds.
We designed a cascade sequence for isoindoline synthesis
that we anticipated could be catalyzed by a palladium(II)
complex and would utilize boronic acids as a starting material
(Scheme 1). The cascade reaction would initiate with aryla-
tion of imine 1.[4] The resultant sulfonamide would engage the
pendant allylic acetate by aminopalladation; b-acetoxy elim-
ination would release the isoindoline product.[5,6]
A major challenge was identification of a catalyst with the
appropriate electronic balance to facilitate all steps in the
catalytic cycle. While nucleophilic arylation of imines requires
electron-donating ligands,[4] migratory insertion is generally
promoted by palladium(II) catalysts with electrophilic char-
acter.[7] We selected phosphinite palladacycle 3, which is a
catalyst with demonstrated activity for arylation of imines,[4b]
with the thought that the p-accepting phosphonite would
balance s donation from the aryl group.[8] In practice, we have
found this complex to be an effective catalyst for our cascade
sequence (see below).
Scheme 1. Proposed synthesis of isoindoline derivatives. L=ligand,
Ts =4-toluenesulfonyl.
temperature, in the presence of catalyst 3, effective arylation
of 1 with phenyl boronic acid occurs (Table 1, entry 1).
Elevated reaction temperatures promote the cyclization
reaction (Table 1, entry 2; Method A). Notably, isoindoline
2 is generated as a single diastereomer under these reaction
conditions.
Table 1: Optimization of reaction conditions for cascade cyclizations.
We began our investigation with reaction conditions
similar to those employed for imine arylation.[4b] At room
Entry ArBX2 (equiv)
T [8C] Method t [h]
Yield [%][a]
2
4
5
[*] F. J. Williams, Prof. E. R. Jarvo
Department of Chemistry
1[b]
2
3
4
5
PhB(OH)2 (1)
PhB(OH)2 (1)
(PhBO)3 (0.5)
(PhBO)3 (0.5)
2-naphthyl-B(OH)2 (1)
2-naphthyl-B(OH)2 (1)
(2-naphthyl-BO)3 (0.5) 110
(m-BnOC6H4BO)3 (0.5) 110
(m-BnOC6H4BO)3 (0.5) 110
23
80
80
110
80
80
–
A
–
B
A
A
B
B
C
48
24
24
10
19
7
10
10
2
<5 85 <5
93 <5 7
University of California, Irvine
Irvine, CA 92697 (USA)
Fax: (+1)949-824-2210
E-mail: erjarvo@uci.edu
Homepage: http://chem.ps.uci.edu/~erjarvo/Jarvo_Group/
Home.html
43 23 >5
90 <5 <5
35 10 30
57 12 22
79 <5 16
6
7
8
23 <5
67 <5
7
5
[**] This work was supported by an NSF CAREER Award (CHE-0847273).
We thank Markus Dꢀrr for preparation of isoindoline 2i. We thank
Frontier Chemicals for donation of boronic acids and Heraeus Metal
Processing for donation of palladium complexes.
9[c]
[a] Determined by 1H NMR spectroscopy of aliquots taken from the
reaction mixture utilizing Ph2SiMe2 as an internal standard. [b] Relative
ratio of products to starting material reported. [c] 2 equiv of CsF added.
Bn=benzyl.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201008160.
Angew. Chem. Int. Ed. 2011, 50, 4459 –4462
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4459

Communications
Table 2: Scope of isoindoline synthesis.
To develop robust conditions for the cascade reaction, we
optimized the reaction conditions for two additional boronic
acid partners.[9] We found that the initial conditions (Meth-
od A) worked well for electron-rich aryl boronic acids.
However, reactions of electron-neutral and electron-poor
aryl boronic acids suffered from competitive hydrolysis of the
imine. We utilized aryl boroxines, the anhydrous trimer of
boronic acids, to minimize competitive hydrolysis (Table 1,
entry 7; Method B). Despite reduced hydrolysis, certain
electron-poor aryl boroxines still provided slow reaction
rates and, under prolonged reaction times, decomposition of
the isoindoline occurred (Table 1, entry 8). We examined a
series of additives thought to accelerate transmetalation
events and found that CsF accelerated reactions of partic-
ularly sluggish boroxines and avoided the formation of
decomposition products (Table 1, entry 9; Method C).
Good yields of isoindoline compounds incorporating a
variety of substituted boronic acid derivatives were obtained
with excellent diastereoselectivity (Table 2). For electron-rich
aryl boronic acids as well as certain electron-neutral aryl
boronic acids, our original conditions at 808C were quite
successful (Method A; Table 2, entries 1–3). For example,
ether-substituted boronic acids reacted smoothly under these
conditions. When hydrolysis was a problem with an electron-
neutral aryl boronic acid, the corresponding boroxine was
used, and the reaction was run at an elevated temperature
(Method B; Table 2, entries 4, 5, and 7). Finally, if an electron-
poor aryl boron partner was needed, the corresponding
boroxine was used with added cesium fluoride to increase the
reaction rate (Method C; Table 2, entries 6, 8, and 9). Halide-
and trifluoromethyl-substituted boroxines afforded good
yields of product under these reaction conditions.
Entry
Ar
Method[a]
t
[h]
2
Yield d.r.[c]
[%][b]
1
2
3
4
5
6
R=OMe
A
A
A
B
B
C
24 2a 74
11 2b 73
12 2c 66
>20:1
>20:1
>20:1
20:1
>20:1
18:1
R=OBn
R=CH2OtBu
R=H
R=Ph
R=CF3
6
6
1
2d 82
2e 60
2 f 54
7
8
9
B
C
C
10 2g 83
>20:1
11:1
2
2
2h 72
2i 64
14:1
[a] All reactions were performed in sealed vials with [1]=66 mm.
Method A: 1 equiv of ArB(OH)2, 808C; Method B: 0.5 equiv of
(ArBO)3, 1108C; Method C: 0.5 equiv of (ArBO)3, 2 equiv of CsF,
1108C. [b] Yield of isolated product after column chromatography on
silica gel. [c] Determined by 1H NMR spectroscopy.
and b-acetoxy elimination steps. We found that addition of
[PdCl2(CH3CN)2] and P(2-furyl)3 resulted in good yields of
the desired tetrahydroisoquinoline 7 [Eq. (1)].[13] The reaction
All of the isoindoline products shown in Table 2 were
formed with high diastereoselectivity for the cis isomer. We
hypothesized that the cis isoindoline was lower in energy than
the trans isoindoline as that isomer minimized steric inter-
actions between the sulfonamide group and isoindoline
substituents.[10] To determine the relative stabilities of the
diastereomers we performed DFT calculations using B3LYP/
6-311G(d) to identify an energy difference between the
lowest energy cis product conformer and the lowest energy
trans product conformer.[11] The cis-2a product was calculated
to be 3 kcalmolÀ1 lower in energy than the trans-2a. There-
fore, the major product formed is indeed the more stable
diastereomer. Resubjection of the trans diastereomer of 2h to
the reaction conditions resulted in partial isomerization back
to the cis diastereomer. These results indicate thermodynamic
control of product distribution.[12]
We sought to determine whether or not this reaction
would be capable of preparing tetrahydroisoquinoline com-
pounds. Imine 6 was designed to undergo cascade cyclization
to provide 1-aryl-3-vinyl-tetrahydroisoquinoline 7. Under our
standard reaction conditions, arylation of the imine preceded
smoothly, however, the reaction stalled and cyclization was
not observed. We hypothesized that while the palladacycle
may be capable of catalyzing migratory insertion to form a
five-membered ring, formation of a six-membered ring would
be more challenging and could require a more electrophilic
catalyst. We examined alternative catalysts for the cyclization
could be performed in one reaction flask by simply adding the
second catalyst directly to the reaction after the arylation step
was complete. While the diastereoselectivity of this reaction is
modest, formation of the more challenging six-membered
ring heterocycle is noteworthy.
Our development of the isoindoline-forming reaction was
influenced by our mechanistic rationale (Scheme 1), however,
we recognized that alternative mechanisms could also be
viable. In the proposed mechanism, the catalyst remains in the
palladium(II) oxidation state throughout the transformation
(Scheme 2, Pathway A). However, alternative mechanisms
involving oxidative addition of a palladium(0) catalyst with
the allylic acetate could also lead to product formation. Two
likely alternative mechanisms are outlined (Pathway B and
C).[14] In Pathway B, palladium-catalyzed imine arylation
provides sulfonamide 4. Subsequent formation of a p-
allylpalladium(II) intermediate and intramolecular attack
by the sulfonamide generates the heterocycle. Pathway C
involves formation of a p-allylpalladium(II) intermediate,
attack of the imine nitrogen to form an iminium ion, and
capture by a nucleophilic aryl species.
4460 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 4459 –4462

branched allylic acetate 8 [Scheme 3, Eq. (3)]. If Pathway B
or C is operative, and the reaction proceeds through a p-
allylpalladium intermediate, both 1 and 8 should provide
similar yields of the desired isoindoline 2. However, isoindo-
line 2 cannot be formed from 8 according to Pathway A.
Upon subjecting branched allylic acetate 8 to our standard
reaction conditions employing catalyst 3, less than 5%
isoindoline 2 was formed [Scheme 3, Eq. (3)]. This data is
inconsistent with Pathways B and C and supports reaction via
Pathway A.
To obtain additional evidence to distinguish between
alternative mechanisms we designed a positive control where
reaction according to Pathway A would provide a different
product than reaction along Pathway B or C [Scheme 3,
Eq. (4)]. Imine 9 would undergo cascade cyclization accord-
ing to Pathway A to provide a mixture of E- and Z-olefin
isomers (10 and 11) because both aminopalladation and b-
acetoxy elimination proceed stereospecifically.[5,6] However,
if Pathway B or C is operative, attack on the p-allylpalladiu-
m(II) intermediate would result in exclusive formation of the
thermodynamically favored olefin isomer, 10.[13] Under the
standard reaction conditions, imine 9 afforded a 4.7:1 mixture
of E- and Z-olefin isomers.[16] Although the Z isomer was
formed as the minor product, its appearance provides
evidence that Pathway A is operative. In addition to provid-
ing insight into the proposed mechanism, this reaction
demonstrates that substitution on the allylic acetate is
tolerated in the reaction.
Scheme 2. Possible mechanism for annulation.
To distinguish between the proposed mechanism of the
cascade sequence (Pathway A) and alternative mechanisms
involving oxidative addition into the allylic acetate (Path-
way B and C), we designed
a series of experiments
(Scheme 3). While no one experiment conclusively rules out
In conclusion, we report a palladium-catalyzed cascade
reaction that affords cis-1,3-disubstitued isoindolines. A range
of electron-rich and electron-poor aryl boronic acid deriva-
tives participate. In addition, isoindoline products contain a
new terminal olefin, thus providing a functional handle for
future derivatization. The method is also amenable to
reactions of substituted allylic acetates and the synthesis of
tetrahydroisoquinoline compounds. Mechanistic experiments
À
are consistent with palladium(II) catalysis where the key C N
bond-forming event occurs by aminopalladation of a pendant
olefin.
Experimental Section
Scheme 3. Mechanistic experiments. dba=trans,trans-dibenzylidene-
Standard procedure for Method A: In a glove box, a vial equipped
with a stir bar and septum was charged with 1 (36 mg, 0.10 mmol,
1.0 equiv), catalyst 3 (4.4 mg, 0.0046 mmol, 0.046 equiv), K3PO4
(21 mg, 0.10 mmol, 1.0 equiv), and BaO (30 mg, 0.20 mmol,
2.0 equiv). Outside of the glove box aryl boronic acid (0.20 mmol,
2.0 equiv) was added to the vial, followed by an inlet for positive
pressure of N2. Toluene (1.5 mL) was added, the N2 inlet was removed
and the vial was placed in an 808C bath. Once the reaction was
complete (as evident by TLC), the reaction mixture was directly
purified by column chromatography on silica gel (0–15% EtOAc/
hexanes).
Standard procedure for Method B: In a glove box, a vial
equipped with a stir bar and septum was charged with 1 (36 mg,
0.10 mmol, 1.0 equiv), catalyst 3 (4.4 mg, 0.0046 mmol, 0.046 equiv),
boroxine (0.050 mmol, 0.50 equiv), K3PO4 (21 mg, 0.10 mmol,
1.0 equiv), and BaO (30 mg, 0.20 mmol, 2.0 equiv). The vial was
brought out of the glove box, and toluene (1.5 mL) was added under
positive pressure of N2. The N2 inlet was removed and the vial was set
to stir in a 1108C bath and monitored by TLC. Once the reaction was
acetone.
alternative Pathways B or C, taken as a whole, the experi-
ments in Scheme 3 are most consistent with Pathway A and
make mechanisms involving p-allylpalladium intermediates
unlikely. First, we performed a series of reactions using a
palladium(0) catalyst and either imine 1 or sulfonamide 4 as
starting materials [Scheme 3, Eq. (2)].[15] All reactions were
run under our standard reaction conditions. In contrast to
reactions using catalyst 3, when [Pd2(dba)3] was employed as
the catalyst no discernable isoindoline 2 was generated. These
results are consistent with mechanisms that do not involve
oxidative addition by a palladium(0) catalyst.
For further information concerning the viability of p-
allylpalladium intermediates in the reaction, we synthesized
Angew. Chem. Int. Ed. 2011, 50, 4459 –4462
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4461

Communications
complete (as evident by TLC), the reaction mixture was directly
purified by column chromatography on silica gel (0–15% EtOAc/
hexanes).
e) P. J. Kukkola, N. A. Bilci, T. J. Ikeler, Tetrahedron Lett. 1996,
37, 5065 – 5068.
[4] a) K. Ueura, S. Miyamura, T. Satoh, M. Miura, J. Organomet.
Chem. 2006, 691, 2821 – 2826; b) P. He, Y. Lu, Q.-S. Hu,
Tetrahedron Lett. 2007, 48, 5283 – 5288; c) G.-N. Ma, T. Zhang,
M. Shi, Org. Lett. 2009, 11, 875 – 878.
[5] Aminopalladation: a) J. D. Neukom, N. S. Perch, J. P. Wolfe, J.
Am. Chem. Soc. 2010, 132, 6276 – 6277, and references therein;
b) T. E. Mꢀller, K. C. Hultzsch, M. Yus, F. Foubelo, M. Tada,
Chem. Rev. 2008, 108, 3795 – 3892.
[6] Examples of b-acetoxy elimination occurring in preference to b-
hydride elimination: a) A. G. Steinig, A. Meijere, Eur. J. Org.
Chem. 1999, 1333 – 1334; b) H. Ohmiya, Y. Makida, T. Tanaka,
M. Sawamura, J. Am. Chem. Soc. 2008, 130, 17276 – 17277; c) M.
Lautens, E. Tayama, C. Herse, J. Am. Chem. Soc. 2005, 127, 72 –
73; d) H. Ohmiya, Y. Makida, D. Li, M. Tanabe, M. Sawamura, J.
Am. Chem. Soc. 2010, 132, 879 – 889.
Standard procedure for Method C: In a glove box, a vial
equipped with a stir bar and septa cap was charged with 1 (36 mg,
0.10 mmol, 1.0 equiv), catalyst 3 (4.4 mg, 0.0046 mmol, 0.046 equiv),
boroxine (0.050 mmol, 0.50 equiv), K3PO4 (21 mg, 0.10 mmol,
1.0 equiv), BaO (30 mg, 0.20 mmol, 2.0 equiv), and CsF (30. mg,
0.20 mmol, 2.0 equiv). The vial was brought out of the glove box, and
toluene (1.5 mL) was added under positive pressure of N2. The N2
inlet was removed and the vial was set to stir in a 1108C bath and
monitored by TLC. Once the reaction was complete (as evident by
TLC), the reaction mixture was directly purified by column chroma-
tography on silica gel (0–15% EtOAc/hexanes).
Received: December 23, 2010
Published online: April 7, 2011
[7] a) J. Tsuji, Palladium Reagents and Catalysts, Wiley, Chichester,
2004, pp. 11 – 13; b) R. H. Crabtree, The Organometallic
Chemistry of Transition Metals, 4th ed., Wiley, Hoboken, NJ,
2005, pp. 141 – 197.
[8] a) C. A. Tolman, Chem. Rev. 1977, 77, 313 – 348; b) R. B. Bed-
ford, S. L. Hazelwood, P. N. Horton, M. B. Hursthouse, Dalton
Trans. 2003, 4164 – 4174; c) J. Dupont, C. S. Consorti, J. Spencer,
Chem. Rev. 2005, 105, 2527 – 2571.
[9] Under these reaction conditions, aryl trifluoroborate salts and
vinyl pinacol boronic ester do not react and hydrolysis of starting
material 1 predominates.
[10] For discussion of cis selectivity in a similar five-membered
oxazolidine system, see: J.-G. Shim, Y. Yamamoto, Heterocycles
2000, 52, 885 – 895.
Keywords: cascade reactions · diastereoselectivity ·
imine arylation · isoindolines · palladium
.
[1] a) M. Berglund, M. F. Dalence-Guzman, S. Skogvall, O. Sterner,
Bioorg. Med. Chem. 2008, 16, 2513 – 2528; b) O. F. Funk, V.
Kettmann, J. Drimal, T. Langer, J. Med. Chem. 2004, 47, 2750 –
2760; c) D. Berger, R. Citarella, M. Dutia, L. Greenburger, W.
Hallet, R. Paul, D. Powell, J. Med. Chem. 1999, 42, 2145 – 2161;
d) A. A. Krysko, B. M. Chugunov, O. L. Malovichko, S. A.
Andronati, T. A. Kabanova, T. L. Karaseva, A. V. Kiriyak,
Bioorg. Med. Chem. 2004, 14, 5533 – 5535.
[2] Representative examples of syntheses of isoindolines: a) M.
Bao, H. Nakamura, A. Inoue, Y. Yamamoto, Chem. Lett. 2002,
158 – 159; b) O. Gaertzen, S. L. Buchwald, J. Org. Chem. 2002,
67, 465 – 475; c) C. A. Strassert, J. Awruch, Monatsh. Chem.
2006, 137, 1499 – 1503; d) T. M. Barnard, G. S. Vanier, M. J.
Collins, Org. Process Res. Dev. 2006, 10, 1233 – 1237; e) A.
Mꢀller, K. Polborn, K. T. Wanner, J. Heterocycl. Chem. 2007, 44,
575 – 590; f) K. N. Clary, M. Parvez, T. G. Back, J. Org. Chem.
2010, 75, 3751 – 3760.
[11] For experimental details and references concerning the calcu-
lations, see the Supporting Information.
[12] For experimental details, see the Supporting Information.
[13] This combination has been used for synthesis of piperazines by
carboamination of acyclic precursors, see: J. S. Nakhla, J. P.
Wolfe, Org. Lett. 2007, 9, 3279 – 3282.
[14] a) B. M. Trost, D. L. VanVranken, Chem. Rev. 1996, 96, 395 –
422; b) J. Tsuji, Palladium Reagents and Catalysts, Wiley, Chi-
chester, 2004, chap. 4.
[3] a) S. Fustero, J. Moscardꢁ, M. Sꢂnchez-Rosellꢁ, E. Rodrꢃguez, P.
Barrio, Org. Lett. 2010, 12, 5494 – 5497; b) D. Enders, A. A.
Narine, F. Toulgoat, T. Bisschops, Angew. Chem. 2008, 120,
5744 – 5748; Angew. Chem. Int. Ed. 2008, 47, 5661 – 5665; c) M.
Bꢀchert, S. Meinke, A. H. G. P. Prenzel, N. Depperman, W.
Maison, Org. Lett. 2006, 8, 5553 – 5556; d) C. Wang, X.-H. Chen,
S.-M. Zhou, L.-Z. Gong, Chem. Commun. 2010, 46, 1275 – 1277;
[15] In the presence or absence of phosphonite ligand, isoindoline 2
was formed in lower yields when using a palladium(0) catalyst
than when using catalyst 3. Representative experiments are
shown in Scheme 3, see the Supporting Information for more
detail.
[16] The major diastereomers of 10 and 11 were assumed to be cis in
analogy to isoindolines 2.
4462 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 4459 –4462