Stereochemical surprises in the lewis acid-mediated allylation of isatins

Article abstract of DOI:10.1021/jo101420e

The BF₃·OEt₂-mediated allylation of isatin with an α-chiral allylic stannane is diastereo- and enantioselective. Conversely, allylation of any substituted isatin employing the identical protocol is not diastereoselective at all and o

Full text of DOI:10.1021/jo101420e

pubs.acs.org/joc
metal-catalyzed3-5 and organocatalyzed6,7 methods are
Stereochemical Surprises in the Lewis Acid-Mediated
Allylation of Isatins
available for this, including arylation,3 allylation,4,5 aldol,6
and Friedel-Crafts5,7 reactions.
Our laboratory recently accomplished the enantiospecific
preparation of a fragile R-chiral allylic stannane8 by copper-
catalyzed allylic substitution of an essentially enantiopure
allylic precursor with (Bu3Sn)2Zn9 [(S)-1 f (R)-2, Scheme
‡
,‡
Devendra J. Vyas,‡,† Roland Frohlich, and Martin Oestreich*
€
‡
€
Organisch-Chemisches Institut, Westfalische
Wilhelms-Universita€t Mu€nster, Corrensstrasse 40,
48149 Mu€nster, Germany, and †NRW Graduate School
of Chemistry
1].10 Subsequent BF3 OEt2-promoted allylation of alde-
3
hydes was syn diastereoselective and yielded homoallylic
alcohols with excellent enantiomeric excesses and chirality
transfers [(R)-2 f syn-3, Scheme 1].10 We reasoned that
activated carbonyl compounds such as isatin would also
participate in this highly stereocontrolled allylation. In anticipa-
tion of a rather systematic investigation employing a series of
functionalized isatins, we were surprised to learn that the parent
isatin behaves distinctly different from all substituted isatins
tested.12 In this Note, we share our unexpected observation that
the levels of diastereo- as well as enantioselection in this reagent-
controlled allylation are markedly dependent on any substitu-
tion of the isatin benzene core.
martin.oestreich@uni-muenster.de
Received July 19, 2010
(3) (a) Shintani, R.; Inoue, M.; Hayashi, T. Angew. Chem., Int. Ed. 2006, 45,
3353–3356. (b) Toullec, P. Y.; Jagt, R. B. C.; de Vries, J. G.; Feringa, B. L.;
Minnaard, A. J. Org. Lett. 2006, 8, 2715–2718. (c) Lai, H.; Huang, Z.; Wu, Q.;
Qin, Y. J. Org. Chem. 2009, 74, 283–288. (d)Tomita, D.;Yamatsugu, K.;Kanai,
M.; Shibasaki, M. J. Am. Chem. Soc. 2009, 131, 6946–6948. For methylation,
see: (e) Funabashi, K.; Jachmann, M.; Kanai, M.; Shibasaki, M. Angew. Chem.,
Int. Ed. 2003, 42, 5489–5492.
(4) (a) Itoh, J.; Han, S. B.; Krische, M. J. Angew. Chem., Int. Ed. 2009, 48,
6313–6316. (b) Kitajima, M.; Mori, I.; Arai, K.; Kogure, N.; Takayama, H.
Tetrahedron Lett. 2006, 47, 3199–3202.
(5) Hanhan, N. V.; Sahin, A. H.; Chang, T. W.; Fettinger, J. C.; Franz,
A. K. Angew. Chem., Int. Ed. 2010, 49, 744–747.
(6) (a) Luppi, G.; Cozzi, P. G.; Monari, M.; Kaptein, B.; Broxterman,
Q. B.; Tomasini, C. J. Org. Chem. 2005, 70, 7418–7421. (b) Luppi, G.; Monari,
^
M.; Correa, R. J.; Violante, F.; de, A.; Pinto, A. C.; Kaptein, B.; Broxterman,
Q. B.; Garden, S. J.; Tomasini, C. Tetrahedron 2006, 62, 12017–12024.
The BF3 OEt2-mediated allylation of isatin with an R-chiral
3
allylic stannane is diastereo- and enantioselective. Con-
versely, allylation of any substituted isatin employing the
identical protocol is not diastereoselective at all and only
enantioselective for the major diastereomer having syn
relative configuration. The anti isomer is, however, formed
in almost racemic form. Both absolute and relative config-
urations are unambiguously secured by X-ray analysis of
major isomers, and the stereochemical assignment of the
other 3-substituted 3-hydroxy oxindoles is based on similar
NMR spectroscopic characteristics. The remarkable obser-
vations are rationalized by an acyclic transition state model.
(c) Malkov, A. V.; Kabeshov, M. A.; Bella, M.; Kysilka, O.; Malyshev,
ꢀꢁ
ꢀ
ꢁ
ꢀ
D. A.; Pluhackova, K.; Kocovsky, P. Org. Lett. 2007, 9, 5473–5476. (d) Chen.,
J.-R.; Liu, X.-P.; Zhu, X.-Y.; Li, L.; Qiao, Y.-F.; Zhang, J.-M.; Xiao, W.-J.
Tetrahedron 2007, 63, 10437–10444. (e) Nakamura, S.; Hara, N.; Nakashima,
H.; Kubo, K.; Shibata, N.; Toru, T. Chem.-Eur. J. 2008, 14, 8079–8081.
^
(f) Angelici, G.; Correa, R. J.; Garden, S. J.; Tomasini, C. Tetrahedron Lett.
2009, 50, 814–817. (g) Xue, F.; Zhang, S.; Liu, L.; Duan, W.; Wang, W.
Chem.-Asian J. 2009, 4, 1664–1667. (h) Itoh, T.; Ishikawa, H.; Hayashi, Y.
Org. Lett. 2009, 11, 3854–3857. (i) Hara, N.; Nakamura, S.; Shibata, N.; Toru,
T. Chem.-Eur. J. 2009, 15, 6790–6793.
(7) (a) Deng, J.; Zhang, S.; Ding, P.; Jiang, H.; Wang, W.; Li, J. Adv.
Synth. Catal. 2010, 352, 833–838. (b) Chauhan, P.; Chimni, S. S. Chem.-Eur.
J. 2010, 16, 7709–7713.
The oxindole ring is an ubiquitous substructure found in
natural and synthetic alkaloids. Structural variation is intro-
duced with substituents either at the benzene core (C-4-C-7
positions) or at the benzylic carbon atom (C-3 position).
Substitution at that saturated carbon atom usually renders
these molecules chiral, and the asymmetric syntheses of
oxindoles with a tertiary or quaternary stereogenic C-3
position is currently garnering significant attention.1,2 One
way of forming chiral oxindoles is addition of nucleophiles to
the electrophilic isatin carbonyl group, resulting in 3-sub-
stituted 3-hydroxy oxindoles.2 Several unique transition
(8) (a) The σ(C-Sn) bond is involved in hyperconjugation, e.g., stabiliz-
ing carbenium ions in the β position (β-tin effect): Davies, A. G. Organotin
Chemistry; Wiley-VCH: Weinheim, Germany, 2004; pp 35-41. (b) For the
relatively high reactivity of allylic stannanes, see: Mayr, H.; Kempf, B.; Ofial,
A. R. Acc. Chem. Res. 2003, 36, 66–77.
(9) Weickgenannt, A.; Oestreich, M. Chem.-Eur. J. 2010, 16, 402–412.
(10) (a) Schmidtmann, E. S.; Oestreich, M. Angew. Chem., Int. Ed. 2009,
48, 4634–4638. (b) We were able to assign the relative (cf. the Supporting
Information of ref 10a) but not the absolute configuration of the syn
homoallylic alcohols obtained from the BF3 OEt2-mediated allylation of
3
aldehydes (no chemical correlation possible and no X-ray analysis available).
For the related thermal allylation, the absolute configuration of anti homo-
allylic alcohols was assigned based on a cyclic transition state model.
(11) Schmidtmann, E. S.; Oestreich, M. Chem. Commun. 2006, 3643–
3645.
(12) Related but unexplained observations were made in organocatalyzed
aldol reactions. In one case, the enantiomeric excess obtained with the parent
isatin (84% ee) was substantially higher than those found for substituted
isatins (e.g., 49% ee for 7-chloro-substitued isatin).6g In another case, the
enantiomeric excess was excellent for 4,6-dibromo-substituted isatin (92%
ee) but only racemic material was formed with isatin itself (2% ee).6i
(1) (a) Trost, B. M.; Brennan, M. K. Synthesis 2009, 3003–3025.
(b) Galliford, C. V.; Scheidt, K. A. Angew. Chem., Int. Ed. 2007, 46, 8748–
8758. (c) Marti, C.; Carreira, E. M. Eur. J. Org. Chem. 2003, 2209–2219.
(2) Zhou, F.; Liu, Y.-L.; Zhou, J. Adv. Synth. Catal. 2010, 352, 1381–
1407.
6720 J. Org. Chem. 2010, 75, 6720–6723
Published on Web 09/08/2010
DOI: 10.1021/jo101420e
r
2010 American Chemical Society

Vyas et al.
JOCNote
SCHEME 1. Stereoselective Allylation of Aldehydes with an
R-Chiral Allylic Stannanea,b
TABLE 2. Allylation of Unprotected Substituted Isatinsa
aThe absolute configuration of (R)-2 is clearly determined by the stereo-
chemical course of the substitution, therefore its enantiomeric excess is
unchanged.11 bThe absolute configuration of syn-3 is not assigned.10b
TABLE 1. Diastereo- and Enantioselective Allylation of Isatina
entry isatin
R
oxindole yield (%)b drc ee (%)d,e ct (%)f
1
2
3
4
5
6
6a
7a
8a
9a
5-Me
5-OMe
5-F
12a
13a
14a
15a
16a
17a
79
75
89
80
57g
77
87:13 94 (5)
89:11 96 (-)
80:20 95 (-)
75:25 n.d.
n.d.
70:30 88 (-)
98
99
99
5-Br
10a 4-Br
11a 7-Cl
99h
99
89
aAll reactions were conducted with BF3 OEt2 (1.1 equiv) and (R)-2
3
(∼2.0 equiv) with a substrate concentration of 0.5 M in CH2Cl2 at
-78 °C f rt. bIsolated yield of analytically pure material after flash
chromatography. cSyn:anti ratio determined by 1H NMR spectroscopy
prior to purification by integration of the baseline-separated resonance
signals of the diastereomers (cf. the Supporting Information). dDetermined
by HPLC analysis on chiral Daicel Chiralpak and Chiralcel columns.
eEnantiomeric excess of minor diastereomer (anti) in parentheses. fChirality
transfer based on allylic stannane (R)-2 (96% ee). gNot full conversion,
entry isatin PGb oxindole yield (%)c
drd
ee (%)e ct (%)f
1
2
3
4
4a
4b
4c
4d
H
5a
5b
5c
5d
78
85
84
traces
>95:5
>95:5
>95:5
96
92
92
99
96
96
Me
Bn
Boc
aAll reactions were conducted with BF3 OEt2 (1.1 equiv) and (R)-2
3
h
and only isolated yield of major diastereomer (syn). Enantiopure allylic
stannane (R)-2 (99% ee) used. n.d. = not determined.
(∼2.0 equiv) with a substrate concentration of 0.5 M in CH2Cl2 at
-78 °C f rt. bPG = protective group with Bn = benzyl and Boc = tert-
butoxycarbonyl. cIsolated yield of analytically pure material after flash
chromatography. dSyn:anti ratio determined by 1H NMR spectroscopy
prior to purification by integration of the baseline-separated resonance
the substituent even worsened the situation. Allylation of a
4-substituted isatin was sluggish (Table 2, entry 5), and a
substituent in the C-7 position resulted in poor diastereo-
control and chirality transfer (Table 2, entry 6).
e
signals of the diastereomers (cf. the Supporting Information). Deter-
mined by HPLC analysis on chiral Daicel Chiralpak and Chiralcel
columns. fChirality transfer based on allylic stannane (R)-2 (96% ee).
The same but more pronounced trends were seen with
isatins benzylated at the nitrogen atom (6c-11c, Table 3).
We were able to confirm the partial loss of stereochemical
information in the minor anti isomer. The reaction of the
7-substituted benzylated isatin was particularly remarkable
because the diastereoselectivity was reversed, now favoring
anti rather than syn relative configuration (Table 3, entry 6).
The chirality transfer was again poor but enantiomeric
excesses were moderate for both diastereomers.
Our investigation commenced with the allylation of four
isatins, either unprotected or protected at the nitrogen atom
(4a and 4b-d, Table 1). To our delight, allylic stannane (R)-2
(96% ee) reacted cleanly with 4a-c with superb stereocon-
trol (Table 1, entries 1-3). Oxindoles 5a-c were formed as
almost single stereoisomers with syn relative configuration13
(vide infra). Only Boc protection was not compatible with
the reaction conditions (Table 1, entry 4).
Encouraged by these results, we continued testing func-
tionalized, unprotected isatins applying the identical proto-
col (6a-11a, Table 2). Any donor or acceptor substituent in
the C-5 position was detrimental to diastereoselectivity, and
substantial amounts of the anti isomer were formed (Table 2,
entries 1-4). Aside from that deterioration, the enantiomeric
purity of the individual diastereomers held another surprise.
The transfer of chirality was immaculate for syn isomers but
close to none for anti isomers. We were able to verify this
intriguing finding for the 5-Me-substituted isatin (6a f syn-
12a and anti-12a, Table 2, entry 1). Changing the position of
The absolute and relative configurations were assigned by
X-ray analysis of (3R,10S)-13c, (3R,10S)-14c, and (3S*,10S*)-
17c. Separation of the diastereomeric oxindoles by flash
chromatography on silica gel was possible in the benzyl-
protected series, and the pure major isomers were highly
crystalline. The molecular structure of enantiopure (3R,10S)-
13c is depicted in Figure 1 (for the molecular structures of
(3R,10S)-14c and (3S*,10S*)-17c, see the Supporting Infor-
mation). The syn and anti relationships of all oxindoles were
1
deduced from diagnostic H and 13C NMR characteristics:
The chemical shifts of the hydrogen atom at C-10 and its
3J coupling constant as well as those of the stereogenic
carbon atoms (C-3 and C-10) were clearly assignable to the
diastereomers (for a tabulated list of chemical shifts and
coupling constants, see the Supporting Information).
(13) All oxindoles are depicted in the usual (yet improper) way, and the
syn and anti descriptors cannot be assigned based on these drawings. If the
longest carbon chain including the carboxyl group is extended in a zigzag
conformation, the 3R,10S absolute configuration corresponds to syn relative
configuration.
J. Org. Chem. Vol. 75, No. 19, 2010 6721

Vyas et al.
JOCNote
TABLE 3. Allylation of Benzyl-Protected Substituted Isatinsa
FIGURE 2. Suggested acyclic transition state for the syn selective
allylation, and the role of substituents at the isatin benzene core.
The transition state TS might provide a few clues as to why
substituents at the isatin benzene core erode the diastereos-
electivity. A substituent in the C-4 position dramatically
diminishes reactivity, not even reaching full conversion.
Steric interaction between R4 at C-4 and the phenyl group
of 2 might account for that. Along the same lines, R5 at C-5 is
not effecting reactivity but diastereoselectivity. The influence
of R7 at C-7 is particularly difficult to explain because it even
reverses the diastereoselectivity. We think that a substituent
in the C-7 position gears the spatial orientation of the benzyl
group, thereby bringing about steric conflict with the tin
moiety in 2. This effect is less pronounced if the nitrogen
atom is unprotected.
The reagent-controlled allylation of isatin using our en-
antioenriched allylic stannane produced the 3-substituted
3-hydroxy oxindole with high diastereoselectivity (dr > 95:5)
and superb chirality transfer (99%). However, these excellent
levels of stereoselection were only obtained for the parent
isatin, none of the functionalized isatins afforded compar-
able selectivities. The configurational assignment allowed
for the development of a simple model, helping to under-
stand these unexpected findings. We hope to show here that
seemingly minor changes in a reactant might well result in
dramatic changes in the reaction outcome. For isatin allyla-
tion, it is certainly advisable to include more examples than
that of the parent compound.12
entry isatin
R
oxindole yield (%)b drc ee (%)d,e,f ct (%)g
1
2
3
4
5
6
6c
7c
8c
9c
5-Me
5-OMe
5-F
12c
90
91
85
79
38i
86
81:19 98 (36)
67:33 98 (21)
62:38 96 (-)
60:40 94 (-)
99
99
97
95
98
80
13ch
14ch
15c
5-Br
10c 4-Br
11c 7-Cl
16c
17ch
n.d.
97
43:57 79 (81)
aAll reactions were conducted with BF3 OEt2 (1.1 equiv) and (R)-2
3
(∼2.0 equiv) with a substrate concentration of 0.5 M in CH2Cl2 at
-78 °C f rt. bIsolated yield of analytically pure material after flash
chromatography. cSyn:anti ratio determined by 1H NMR spectroscopy
prior to purification by integration of the baseline-separated resonance
signals of the diastereomers (cf. the Supporting Information). dDetermined
by HPLC analysis on chiral Daicel Chiralpak and Chiralcel columns.
eEnantiomeric excess of minor diastereomer (anti) in parentheses. fEnantio-
pure allylic stannane (R)-2 (99% ee) used. gChirality transfer based on
allylic stannane (R)-2 (96% ee). hX-ray analysis of the major diastereomer.
iNot full conversion, and only isolated yield of major diastereomer (syn).
n.d. =not determined.
Experimental Section
General Procedure for BF3 OEt2-Promoted Allylation of Isa-
tins. The indicated isatin (0.20 mmol) is dissolved under inert
3
atmosphere in dry CH2Cl2 (3.0 mL) at -78 °C. BF3 OEt2
3
(0.22 mmol, 1.1 equiv, 1.0 M in CH2Cl2) is added to this
solution, and the solution is briefly brought to rt. Cooling to
-78 °C is followed by the addition of a solution of freshly
prepared enantioenriched allylic stannane (R)-2 (∼0.40 mmol,
∼2.0 equiv, 96% ee or 99% ee) in dry CH2Cl2 (1.0 mL). The
reaction mixture is then allowed to warm to rt. At full conver-
sion (TLC monitoring), Me3Al (0.50 mmol, 2.5 equiv, 2 M in
toluene) is added at 0 °C,14 and the reaction mixture is main-
tained at rt for an additional 2 h. H2O is added (dilute aqueous
HCl must not be used to quench the reaction as epimerization of
the oxindole is observed), and the aqueous phase is extracted
with CH2Cl2 (3 ꢀ 5.0 mL). The combined organic layers are
back-extracted with H2O (3 ꢀ 5.0 mL), washed with brine, and
dried over Na2SO4. Evaporation of the solvents under reduced
pressure affords the crude oxindoles as a mixture of diastereo-
mers, followed by purification by flash chromatography on
FIGURE 1. Molecular structure of (3R,10S)-13c.
The relative and absolute configuration is secured for the
major oxindole diastereomer, and the configuration of the
allylic stannane is also known.10,11 Moreover, there is no
ambiguity in the facial selectivity (CdO of isatin and CdC of
allyl fragment) because of the excellent enantiomeric excesses of
the syn isomer. We therefore propose an antiperiplanar acyclic
transition state to rationalize the R configuration at C-3 and S
configuration at C-10 of the syn oxindoles (TS, Figure 2). We
cannot predict transition states for the anti oxindoles as these are
formed in almost racemic form. With no information about the
facial selectivity in that diastereomeric pathway, we might only
hypothesize that the relative energies of the stereoisomeric
transition states yielding enantiomeric, that is racemic, anti
oxindoles must be similar.
^
(14) Renaud, P.; Lacote, E.; Quaronta, L. Tetrahedron Lett. 1998, 39,
2123–2126.
6722 J. Org. Chem. Vol. 75, No. 19, 2010

Vyas et al.
JOCNote
silica gel with cyclohexane (150 mL) to remove tin impurities
and then cyclohexane/tert-butyl methyl ether mixtures.
on silica gel (cyclohexane and cyclohexane/tert-butyl methyl
ether = 70/30) afforded the analytically pure title compound as
a mixture of diastereomers [65 mg, 84%, dr>95:5, 92% ee for
(3R)-3-Hydroxy-3-[(1S,2E)-1-phenylpent-2-enyl]indolin-2-one
(syn-5a, Table 1, entry 1). syn-5a was prepared from isatin (4a,
30 mg, 0.20 mmol) according to the general procedure. Purifica-
tion by flash column chromatography on silica gel (cyclohexane
and cyclohexane/tert-butyl methyl ether = 60/40) afforded the
analytically pure title compound as a mixture of diastereomers
[47 mg, 78%, dr > 95:5, 96% ee for (3R,10S)-5a] as a white solid:
[R]20D þ62.9 (c 1.05, CHCl3); mp 150 °C; the enantiomeric excess
was determined by HPLC analysis with a Chiralcel OD-H
column (90/10 n-heptane:i-PrOH; flow rate 0.8 mL/min; λ =
(3R,10S)-5c] as a white solid: [R]20 þ25.8 (c 3.90, CHCl3); mp
D
156 °C; the enantiomeric excess was determined by HPLC analysis
with a Chiralcel OD-H column (90/10 n-heptane:i-PrOH; flow rate
0.8 mL/min; λ = 210 nm; tmajor = 11.4 min, tminor = 9.7 min); 1H
NMR (400 MHz, CDCl3) δ1.02(t, J= 7.5 Hz, 3H), 2.10-2.19 (m,
2H), 3.25 (s, 1H), 3.90 (d, J = 10.2 Hz, 1H), 4.53 (d, J = 15.9 Hz,
1H), 4.88 (d, J = 15.9 Hz, 1H), 5.93 (dt, J= 15.2, 6.2 Hz, 1H), 6.13
(dd, J = 15.2, 10.3 Hz, 1H), 6.42 (d, J = 7.7 Hz, 1H), 6.90-6.98
(m, 2H), 6.99-7.24 (m, 11H); 13C NMR (100 MHz, CDCl3) δ
13.7, 25.9, 43.9, 57.5, 78.0, 109.4, 122.8, 124.1, 124.5, 127.1, 127.1,
127.5, 128.1, 128.7, 128.8, 129.2, 129.7, 135.3, 137.7, 139.8, 143.2,
177.4; IR (ATR) ν 3347, 2967, 2362, 1697, 1613, 1493, 1469, 1383,
1352, 1176, 976 cm-1; HRMS (ESI) calcd for C26H25NO2Na [M þ
Na]þ 406.1778, found 406.1779.
1
210 nm; tmajor = 23.8 min, tminor = 12.8 min); H NMR (300
MHz, CDCl3) δ 1.00 (t, J = 7.5 Hz, 3H), 2.04-2.17 (m, 2H), 3.29
(s, 1H), 3.81 (d, J = 9.6 Hz, 1H), 5.86 (dt, J = 15.2, 5.9 Hz, 1H),
5.98 (dd, J = 15.2, 9.7 Hz, 1H), 6.64 (d, J = 7.5 Hz, 1H),
6.98-7.05(m, 1H), 7.07-7.20(m, 7H), 7.77 (br s, 1H); 13C NMR
(75 MHz, CDCl3) δ 13.7, 25.9, 57.5, 78.5, 109.9, 110.1, 122.7,
123.9, 125.0, 127.2, 128.0, 129.2, 129.7, 137.6, 139.4, 140.7,
179.4; IR (ATR) ν 3458, 3191, 2960, 2360, 2337, 1695, 1622,
1469, 1347, 1298, 1267, 1221, 1153, 974 cm-1; HRMS (ESI)
calcd for C19H19NO2Na [M þ Na]þ 316.1308, found 316.1306.
Anal. Calcd for C19H19NO2 (293.36): C, 77.79; H, 6.53; N, 4.77.
Found: C, 77.45; H, 6.46; N, 4.61.
Acknowledgment. D.J.V. thanks the NRW Graduate School
of Chemistry for a predoctoral fellowship (2008-2011). We also
thank Birgit Wibbeling for assistance with an X-ray measurement.
Supporting Information Available: List of diagnostic chemi-
cal shifts and coupling constants, molecular structures, experi-
mental procedures, and characterization data, as well as 1H and
13C NMR spectra for all compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
(3R)-1-Benzyl-3-hydroxy-3-[(1S,2E)-1-phenylpent-2-enyl]indo-
lin-2-one (syn-5c, Table 1, entry 3). syn-5c was prepared from
1-benzylindoline-2,3-dione (4c, 47 mg, 0.20 mmol) according to
the general procedure. Purification by flash column chromatography
J. Org. Chem. Vol. 75, No. 19, 2010 6723