An asymmetric pericyclic cascade approach to 3-alkyl-3-aryloxindoles: Generality, applications and mechanistic investigations

Article abstract of DOI:10.1039/c4ob02526a

The reaction of L-serine derived N-arylnitrones with alkylarylketenes generates asymmetric 3-alkyl-3-aryloxindoles in good to excellent yields (up to 93%) and excellent enantioselectivity (up to 98% ee) via a pericyclic cascade process. The optimization, scope and applications of this transformation are reported, alongside further synthetic and computational investigations. The preparation of the enantiomer of a Roche anti-cancer agent (RO4999200) 1 (96% ee) in three steps demonstrates the potential utility of this methodology.

Full text of DOI:10.1039/c4ob02526a

Organic &
Biomolecular Chemistry
View Article Online
View Journal
PAPER
An asymmetric pericyclic cascade approach to
3-alkyl-3-aryloxindoles: generality, applications
and mechanistic investigations†
Cite this: DOI: 10.1039/c4ob02526a
Edward Richmond,a Kenneth B. Ling,a Nicolas Duguet,a Lois B. Manton,a
Nihan Çelebi-Ölçüm,b,c Yu-Hong Lam,b Sezen Alsancak,c Alexandra M. Z. Slawin,a
K. N. Houk*b and Andrew D. Smith*a
The reaction of L-serine derived N-arylnitrones with alkylarylketenes generates asymmetric 3-alkyl-3-aryl-
oxindoles in good to excellent yields (up to 93%) and excellent enantioselectivity (up to 98% ee) via a
pericyclic cascade process. The optimization, scope and applications of this transformation are reported,
alongside further synthetic and computational investigations. The preparation of the enantiomer of a
Roche anti-cancer agent (RO4999200) 1 (96% ee) in three steps demonstrates the potential utility of this
methodology.
Received 2nd December 2014,
Accepted 3rd December 2014
DOI: 10.1039/c4ob02526a
www.rsc.org/obc
Introduction
Cascade reactions are highly desirable owing to the ability to
perform multiple sequential transformations without the
necessity for additional manipulation or introduction of
further reagents. Such approaches allow significant molecular
complexity to be rapidly assembled, provided each subsequent
transformation in the cascade unmasks a desirable, reactive
functionality.1 Pericyclic cascades are particularly attractive
given their predictable regio- and stereocontrol,2 coupled with
the potential to readily generate multiple carbon–carbon
bonds. Significant attention has focused on the expansion of
this field toward both carbocyclic and heterocyclic frame-
works.3 The 3,3-disubstituted oxindole scaffold is an appealing
target given the prevalence of naturally occurring species4 and
medicinal agents containing this core structure.5 Notably,
alkaloids 26 and 37 have both been prepared from 3,3-disubsti-
tuted oxindole precursors (Fig. 1).
Fig. 1 Oxindole medicinal agent
accessed synthetically from 3,3-disubstituted oxindoles.
1 and natural products 2 and 3
As a consequence of their wide-ranging biological pro-
perties, and given the synthetic community’s interest in devel-
oping novel approaches toward the preparation of molecules
with quaternary stereocentres,8 3,3-disubstituted oxindoles
have emerged as ideal frameworks on which to develop
new asymmetric methodologies.9 Typically, such approaches
employ anilides, isatins or suitably substituted oxindole
derivatives as starting materials (Fig. 2), although numerous
other standalone approaches have also been developed.9b,c,10–12
Asymmetric intramolecular anilide cyclizations13 or Heck
reactions14 typically employ a palladium catalyst in com-
bination with chiral ligands (a), and have found wide appli-
cation in synthesis.15 In similar systems, direct coupling
approaches, without the necessity for pre-activation have been
developed.16 However, these approaches have yet to be ren-
dered enantioselective. O-to-C transfer reactions (b) have
also been used to great effect including Trost’s asymmetric
allylic alkylation methodology,17 and Lewis base-catalyzed
O-to-C carboxyl transfer reactions.18 A plethora of catalytic
aEaStCHEM, School of Chemistry, University of St Andrews, North Haugh,
St Andrews KY16 9ST, UK. E-mail: ads10@st-andrews.ac.uk
bDepartment of Chemistry and Biochemistry, University of California, Los Angeles,
607 Charles E. Young Drive East, Los Angeles, CA 90095, USA.
E-mail: houk@chem.ucla.edu
cDepartment of Chemical Engineering, Yeditepe University, Istanbul, 34755, Turkey
†Electronic supplementary information (ESI) available: All spectroscopic data
and procedures for novel compounds, computational methodology, comparision
of B3LYP and MP2 geometries, full analysis of enantiomeric distributions,
B3LYP-D geometries and energies, distortion–interaction analysis, the Cartesian
coordinates of optimized structures, and the absolute energies and computed
corrections of the reported geometries. CCDC 1029035–1029037. For ESI and
crystallographic data in CIF or other electronic format see DOI: 10.1039/
c4ob02526a
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Fig. 3 Proposed mechanism.
Fig. 2 Typical approaches toward asymmetric 3,3-disubstituted
oxindoles.
rate the oxindole 8 with excellent enantiocontrol and regener-
ate chiral aldehyde 9 (Fig. 3).29
methodologies has been developed over the past decade
employing isatins (c) as starting materials,19 giving access to
3-substituted-3-hydroxyoxindoles that serve as convenient syn-
thetic intermediates.20 Latterly, both stoichiometric and cataly-
tic asymmetric alkylation approaches (d) have been reported to
access 3,3-disubstituted oxindole species.21 This manuscript
details the asymmetric cycloaddition cascade reaction between
nitrones and ketenes (e), allowing direct access to the unpro-
tected 3,3-disubstituted oxindole motif. This method contrasts
the commonly employed approaches that require protection of
the amide functionality, thereby generating N-protected
oxindoles.
The ability of this methodology to generate highly substi-
tuted quaternary stereocentres at the oxindole C(3)-position
with excellent enantiocontrol (up to 90% ee), coupled with the
low cost of the starting materials, warranted further develop-
ment of this reaction manifold. To this end, this manuscript
describes our studies devoted to the optimization of the levels
of enantioselectivity in this transformation, alongside compu-
tational and experimental mechanistic studies of this process.
The full scope and limitations of the optimized process are
delineated, as well as its application to a target Roche anti-
cancer agent (RO4999200).
Previous studies and mechanism
Results and discussion
Stereodirecting group optimization
The hetero-Claisen approach to oxindoles using N-phenyl-
nitrones and diphenylketene was first reported by Staudinger22
and subsequently investigated by Lippman23 and Taylor.24 To explore the necessary structural requirements for generat-
Despite its synthetic potential, an asymmetric variant of this ing high enantiocontrol in this reaction manifold, a range of
process was overlooked until we developed an asymmetric enantiopure N-aryl nitrones 11–16 was synthesized from
route to 3,3-disubstituted oxindoles (up to 90% ee) using readily available chiral starting materials. These nitrones were
Garner’s aldehyde derived N-aryl nitrones and disubstituted then evaluated in the pericyclic cascade process with ethyl-
ketenes.25 Subsequent studies extended this methodology to phenylketene (Fig. 4).30 Initially, Naproxen-derived nitrone 11
the construction of asymmetric 3,3-spirocarbocyclic oxi- was synthesized and evaluated, generating oxindole 10 in a poor
ndoles,26 and computational studies led to a revised mechanis- 27% ee. Mannitol-derived nitrone 12 proved more successful,
tic rationale for these processes.27 The mechanistic pathway is providing oxindole 10 in 78% yield and 70% ee after treatment
consistent with a 3 + 2 cycloaddition across the ketene CvO with ethylphenylketene. An α-oxygenated series of nitrones
bond, with preferential anti-addition with respect to the aryl 13–16, derived from (S)-ethyl lactate, was also synthesized and
portion of the ketene. Facial selectivity in this cycloaddition is tested. These nitrones proved difficult to isolate and were con-
controlled by the preferred arrangement of large and electrone- sequently prepared and evaluated in situ.31 A general trend of
gative allylic groups, and 1,3-allylic strain28 within the enantio- increasing enantioselectivity with increasing substituent size
pure nitrone chiral auxiliary 4, generating a stereodefined was observed, with O-TIPS-substituted nitrone 15 delivering 10
five-membered intermediate 5. Subsequent [3,3]-sigmatropic in 80% ee.
rearrangement yields 6, which undergoes rearomatization and
Subsequent studies prepared and evaluated a series of
tautomerization to give imino acid 7. Each of these steps was chiral nitrones bearing a protected nitrogen atom at the α-posi-
established by computational studies of the reaction transition tion (Fig. 5). The acyclic, α-dibenzylamino nitrone 17 provided
states.27 Acidic hydrolysis and concomitant cyclization gene- the oxindole in 70% ee, but in poor yield. The N-Boc nitrone 4,
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 4 Variation of the chiral nitrone. aIsolated yield of oxindole 10 after
purification by column chromatography. bDetermined by chiral HPLC
analysis.
derived from Garner’s aldehyde, gave oxindole 10 in good yield
and 84% ee. This prompted us to evaluate a series of struc-
tural analogues of 4 in which the N-substituent is varied.
Upon treatment with ethylphenylketene, the N-benzyl nitrone
18 gave the desired oxindole with poor enantiocontrol,
suggesting that structural rigidity or restricted rotation at this
position may be crucial to engendering high levels of enantio-
selectivity. As a consequence, a sulfonamide substituent was
investigated. With N-tosyl nitrone 19, the oxindole was
obtained in good yield and excellent enantiocontrol (75%
yield, 91% ee). Increasing the size of the sulfonamide group
was found to improve the enantioselectivity, as the use of
N-TIPBS (2,4,6-triisopropylbenzenesulfonyl) nitrone 20 furn-
ished oxindole 10 in 86% yield and 96% ee. A single-crystal
X-ray structure of 20 was obtained, and gives an excellent
representation as to the steric impact of this TIPBS residue.32
Further studies allowed for preparation of 20 on gram scale
(see ESI† for details).
Fig. 5 Optimization of stereodirecting groups on the nitrone and rep-
resentation of the X-ray crystal structure of 20. aIsolated yield of oxi-
ndole 10 after purification by column chromatography. bDetermined by
chiral HPLC analysis.
is determined at the 3 + 2 cycloaddition step of the pericyclic
cascade simultaneously by the facial selectivity of the nitrone
and the direction of attack on the ketene.27 Our previous calcu-
lations have shown that the approach of ketene from the
Re face of nitrone is strongly disfavored as it places the ring
methylene group at the sterically more demanding inside posi-
tion. Indeed, nitrones 4, ent-12 and 20 all displayed signifi-
cantly high π-facial selectivities with the Si face attack
contributing more than 97% to the diastereomeric cycloadduct
distribution (see ESI†), and the competing stereodetermining
factor was the orientation of the unsymmetrically substituted
ketene at the TS (Fig. 6).
We initially considered the 2,2-dimethyl-1,3-dioxolanyl
auxiliary (Fig. 6, X = O) to study the stereoinduction in the
absence of steric interactions from the protecting group. The
Computational studies – role of the N-substituent in
determining enantioselectivity
In order to understand the origin of the effect of the size of calculations predict a 1.4 kcal mol−1 difference in free energy
the N-substituent on the enantioselectivity of the reaction, the between TSE-O and TSZ-O favoring the trans addition with
competing stereoisomeric transition structures (TSs) for respect to the phenyl portion of the ketene. The cycloaddition
nitrones 4, ent-12, and 20 were located using Gaussian 0933 at occurs in the plane of ketene substituents, and the observed
the M06-2X/6-311+G(d,p)//B3LYP/6-31G(d) level. In the pro- trans-phenyl selectivity can be explained by the attack of nitrone
posed mechanism, the stereochemical outcome of the reaction from the least hindered side in the plane of ketene. Steric
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
The inclusion of N-protected chiral auxiliaries leads to an
increase in the E/Z selectivity by introducing a higher degree of
steric hindrance at the transition structures with the Ph group
cis (Fig. 6, X = NBoc and NTIPBS). While the additional contri-
bution of N-PG to the selectivity is found to be minimal for
N-Boc nitrone, the steric effects become significant with the
increasing size of the protecting group, disfavoring TSE-
NTIPBS compared with TSZ-NTIPBS with an energetic cost that
amounts to 4.0 kcal mol−1. Steric interactions between the
N-substituent and the ketene substituents result in substantial
deviations in the antiperiplanarity of the C–N bond at the TS.
The O–C–C–N dihedral angle (τ) decreases with increasing size
of the N-substituent and correlates well with the activation
energy (R2 = 0.94, see ESI†) suggesting a combination of
steric and electronic effects in determining the selectivity of
the reaction. A distortion-interaction analysis shows that both
distortion and interaction energies favorably contribute to the
stabilization of TSE-NTIPBS with respect to TSZ-NTIPBS (see
ESI†). The edge-to-face π–π interaction displayed in the X-ray
structure of nitrone 20 is well reproduced by the calculations
(shown highlighted in Fig. 6).36
Scope and limitations
With N-TIPBS nitrone 20 established as the optimum chiral
auxiliary for this asymmetric oxindole forming methodology,
reactions with a range of alkylarylketenes were undertaken to
demonstrate the scope and limitations of this transformation.
To provide a direct comparison between the N-TIPBS and
N-Boc nitrone chiral auxiliaries, the reactions with the same
series of alkylarylketenes were selected, the results of which
are summarized in Fig. 7. In all cases, an improvement in ee
Fig. 6 Relative Gibbs free energies (in kcal mol−1) of stereoselectivity-
determining 3 + 2 cycloaddition transition structures for nitrones 4, ent-
12 and 20, calculated with M06-2X/6-311+G(d,p)(THF)//B3LYP/6-31G(d).
The hydrogen atoms, except those around the Newman projections, are
omitted for clarity.
effects of ketene substituents for nucleophilic additions at the
ketene LUMO in the plane of substituents are well documented,
and have been proposed to be responsible for high levels of E/Z
selectivities obtained in the reactions of unsymmetrically substi-
tuted ketenes with carbon and oxygen nucleophiles.34,35
a
Fig. 7 Asymmetric oxindole synthesis using nitrone 20; ee values in
parentheses represent those obtained using N-Boc nitrone 4.25
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
was observed with N-TIPBS nitrone 20, whilst yields remained
high and a variety of aryl and alkyl substitution patterns at the
C(3)-position of the oxindole were successfully incorporated
(21–26).
Using the optimised N-TIBPS nitrone 20, an extensive
ketene screen showed that numerous alkylarylketenes were
well tolerated in this process, yielding the respective oxindoles
in good yields although varying ee (Fig. 8).37 β-Branching of
the alkyl group of the ketene component was well tolerated
(27) and reaction with halo-arylketenes yielded oxindoles 28,
29 and 30 in excellent ee. The transformation is also compati-
ble with a C(3)-indolyl substituent, which is common to many
natural products,38 as oxindoles 31 and 32 were furnished in
88% and 84% ee, respectively. Ketenes bearing a 2-substituted
phenyl ring (33 and 34) or a 1-naphthyl substituent (35) ring
resulted in diminished enantiocontrol, yet these reactions still
provided the desired oxindoles in moderate to excellent yields.
α-Branching of the alkyl substituent resulted in a racemic
mixture, as illustrated by the reaction generating oxindole 36.
Attempted extension to C(3)-tertiary asymmetric oxindoles
using monosubstituted ketenes (generated in situ) allowed the
isolation of oxindole 37 in a promising 72% yield and 50% ee.
However, this species proved configurationally unstable, with
slow racemization observed over time.39 Oxindole 38 was gen-
erated using a solution of methylketene in 54% yield, but was
racemic.
Fig. 9 Asymmetric oxindole synthesis using 4-(39 and 40) or 2-substi-
tuted (43 and 44) N-aryl nitrones.
Regioselectivity
Attention next turned to substitution of the N-aryl nitrone ring
to probe the effect upon regio- and stereoselectivity of this
process (Fig. 9). 4-Tolyl- and 4-chlorophenyl N-aryl nitrones 39
and 40 each gave the 5-substituted oxindole as a single regio-
isomer, in excellent ee and acceptable yields. Similarly, 2-tolyl
nitrone 43 and 2-chlorophenyl nitrone 44 gave 7-substituted
oxindoles 45 and 46, respectively, as single regioisomers in
good yield although with lower enantioselectivities.
However, treatment of 3-substituted N-aryl nitrones 47 and
48 (Fig. 10) with ethylphenylketene gave a 40 : 60 regioisomeric
mixture of the 4- and 6-substituted oxindole isomers res-
pectively, in good yields. Using 3-tolyl nitrone 47, the ee of the
inseparable 4- and 6-regioisomers was 91%, while 3-chloro-
phenyl nitrone 48 gave the separable 4- and 6-chlorooxindoles
in 86% ee.
The computed regioselectivity-determining transition struc-
tures are shown in Fig. 11 for 3-tolyl nitrone 53 and methyl-
phenylketene as the model reactants. According to the
proposed pericyclic cascade mechanism, the regioselectivity is
dictated by the hetero-Claisen rearrangement step of the 3 + 2
Fig. 10 Regioselectivity trends in the asymmetric oxindole synthesis
using 3-substituted N-aryl nitrones. aoxindoles 49 and 50 were isolated
and analyzed as a mixture of regioisomers in 88% combined yield: the
reported yields refer to mol% fraction of total isolated material.
Fig. 8 Asymmetric oxindole synthesis; generality.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
with ethylphenylketene (Fig. 10) where a minimal preference
for the 6-isomer is observed in the case of both chloro- and
methyl- substitution.
Mechanistic validation
In our recent communication, the computational exploration
of the possible reaction pathways for the formation of the key
imino acid intermediates in the reactions of N-phenyl nitrones
with ketenes were evaluated at the M06-2X/6-311+G(d,p)//
B3LYP/6-31G(d) level (Scheme 1).27 These studies described a
novel 1,3-dipolar cycloaddition and hetero-[3,3]-rearrangement
cascade (Scheme 1, Pathway A) involving a chirality transfer
between the highly asynchronous but concerted pericyclic
steps. Here, we extend our calculations to the SCS-MP2/6-31G
(d)//MP2/6-31G(d) level of theory (Fig. 12), which provided the
best results in an equivalent key [3,3]-sigmatropic rearrange-
ment step of the acid-promoted Fischer indole reaction, where
B3LYP optimizations were shown to yield highly asynchronous
and dissociative transition states.40
In the first step of the pericyclic cascade mechanism, the
3 + 2 cycloaddition between nitrone and ketene across the CvO
bond forms the substituted 1,2,4-dioxazolidine. The 3 + 2
cycloaddition transition structure, TS-(3 + 2), is predicted to be
16.7 kcal mol−1 uphill from the starting nitrone and ketene
(Fig. 12). TS-(3 + 2) features two forming C–O bonds of 1.57 Å
and 2.12 Å. The geometry and electronic features are in good
agreement with the previously described 3 + 2 cycloaddition
TS obtained using B3LYP.27 This cycloaddition step sets the
stereochemistry of the 5-membered ring in the cycloadduct
involving an unsymmetrically substituted exocyclic alkylidene
group. The cycloadduct int-1 is 6.9 kcal mol−1 more stable rela-
tive to the separated reactants, and smoothly undergoes an
aromatic hetero-[3,3]-sigmatropic shift via TS-[3,3] with an acti-
vation free energy of 2.3 kcal mol−1 furnishing the quaternary
stereocenter. The subsequent stereospecific hetero-[3,3]-
rearrangement step transfers the stereochemical information
Fig. 11 Regioselectivity-determining transition structures (SCS-MP2/
6-31G(d)(THF)//MP2/6-31G(d)(THF)) involving meta-substituted nitrones.
cycloadduct. This [3,3]-sigmatropic rearrangement is highly
asynchronous since the C–C bond is barely formed when the
N–O bond is being cleaved at the transition state. The pertur-
bation by any substituent at either the 4- or the 6-position is
therefore likely to be minimal. Indeed, the free energies of the
transition structures for the 4-isomer and the 6-isomer were
found to differ only by 0.1 kcal mol−1 in favor of the 6-isomer.
This is essentially replicated in experimental investigations
Scheme 1 Mechanistic proposals for the reactions of N-phenylnitrones with disubstituted ketenes.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Fig. 12 Pericyclic cascade (Pathway A, Scheme 1) and competing heterolytic cleavage pathway (Pathway B, Scheme 1; TS’-[3,3]) (R = i-Pr). The rela-
tive free energies (ΔG, kcal/mol), calculated using SCS-MP2/6-31G(d)//MP2/6-31G(d), are given with respect to the separated reactants.
of the cycloadduct to intermediate int-2, installing the desired
quaternary stereocenter. TS-[3,3] is early and highly asynchro-
nous; the breaking N–O bond is 1.75 Å as the forming C–C
bond is 2.89 Å. The critical distances in the MP2 optimized
transition structure are significantly shorter compared with
those predicted using B3LYP (dC–C(B3LYP) = 3.47 Å, dN–O(B3LYP) =
1.95 Å). The intrinsic reaction coordinate (IRC) path from TS-
[3,3] indicates no intermediates, and connected the cyclo-
adduct int-1 to intermediate int-2, which has a relative free
energy of −36.6 kcal mol−1 with respect to the separated reac-
tants. Cleavage of the hemiaminal linkage and rearomatization
form the aromatic imino acid.
An analogous [3,3]-rearrangement transition structure, TS′-
[3,3], in a previously proposed nucleophilic addition and
hetero-Claisen rearrangement mechanism (Scheme 1, pathway
B) led to dissociation rather than rearrangement (red inset,
Fig. 13 Competing [3,3]-rearrangement and cleavage pathways (a) in
Fig. 12). This pathway has a higher free energy of activation
the Fischer indole synthesis;39 (b) in pathway B; (c) in pathway A.
(ΔG‡ = 25.8 kcal mol−1) than the pericyclic cascade (ΔG‡
=
16.7 kcal mol−1
experimentally.
)
and is, therefore, not observed
desired rearrangement.40 Instead, heterolysis of the N–N(+)
bond occurs, giving an aniline and a resonance-stabilized car-
bocation. Our computations showed that these attempted
rearrangements fail because the electron-donating groups on
the terminal aliphatic carbon of the rearranging system signifi-
cantly lower the bond dissociation enthalpy of the N–N(+)
bond, and favor the heterolysis of this bond over the intended
[3,3]-sigmatropic rearrangement. Similarly, 3-aza,4-oxa-Cope
rearrangement transition states of pathways A and B contain a
disubstituted alkylidene group derived from the ketene reac-
tant (Fig. 13b and 13c). In accord with the previous findings,40
the located transition structures, TS-[3,3] and TS′-[3,3] are both
early and highly asynchronous, indicative of significant weak-
ening of the O–N bond (Fig. 12). Our calculations predict that
Further mechanistic insight can be gleaned by comparing
the sigmatropic rearrangement step in the acid-promoted
Fischer indole reaction, the originally proposed, nucleophilic
addition/hetero-Claisen rearrangement pathway (Pathway B,
Scheme 1), and the pericyclic cascade pathway (Pathway A,
Scheme 1). The key steps common to these three pathways are
laid out in Fig. 13. The key step of the Fischer indole synthesis
is an aromatic hetero-Claisen rearrangement (3,4-diaza-Cope
rearrangement, Fig. 13a); this corresponds to a 3-aza,4-oxa-
Cope rearrangement involved in both pathways A and B in the
present study (Fig. 13b and c). Our previous work found that
Fischer indolization substrates with certain substitution pat-
terns, such as the one shown in Fig. 13a, do not undergo the
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
TS-[3,3] in pathway A gives the rearranged product (Fig. 13c),
but the analogous TS′-[3,3] in pathway B leads to heterolysis of
the O–N(+) bond akin to the heterolytic cleavage of the N–N(+)
bond in the Fischer indole reaction (Fig. 13b). The formation
of the non-charge separated arylimine and α-lactone products
(ΔG = −40.5 kcal mol−1) favors heterolysis over rearrangement
in pathway B. On the other hand, the ketene oxygen is involved
in the hemiaminal linkage in the cycloadduct making the
rearranging system in pathway A uncharged. With no formal
positive charge on nitrogen, the subsequent [3,3]-shift pro-
ceeds via TS-[3,3] in a concerted manner. This rearrangement
is facile, presumably driven by both the electronic effects that
weaken the N–O bond and the release of the ring strain. Thus,
the competing high-energy-barrier to heterolysis of the O–N+
bond disfavors the operation of the nucleophilic addition-
hetero-Claisen pathway. By contrast, our proposed pathway
proceeds first by a 3 + 2 cycloaddition, yielding an uncharged
system, which can undergo facile [3,3]-rearrangement. Com-
parison of pathway B (Fig. 13b) and the case of a failed Fischer
indolization attempt (Fig. 13a) suggests the importance of elec-
tronic effects in hetero-Claisen rearrangements.
Scheme 3 Isolation of oxazolidinone 59 and representation of its X-ray
crystal structure.
[3,3]-sigmatropic rearrangement, was investigated. Treatment
of N-tert-butyl nitrone 56 with one equivalent of ethyl(para-
chlorophenyl)ketene 57 gave oxazolidinone 59 in 3 : 1 crude dr
(Scheme 3). The major diastereoisomer was isolated in 41%
yield with N-tert-butyl imine also formed in this reaction.42 Iso-
lation of 59 provides indirect evidence for the computed reac-
tion mechanism via initial 3 + 2 cycloaddition across the
ketene CvO bond to furnish transient intermediate 58 which,
via radical or ionic cleavage of the N–O bond, rearranges to
generate the 5-membered oxazolidinone 59.43 An alternative
dissociative decomposition of intermediate 58 can be envi-
saged to account for the generation of N-tert-butyl benz-
aldehyde imine 60 via N–O bond cleavage without
concomitant C–C bond formation, resulting in elimination of
the parent arylbutyric acid 61.
Attempted intermediate isolation
Experimental validation of the computationally predicted
cycloaddition pathway by trapping anticipated reaction inter-
mediates was probed. Initially, in an attempt to preclude the
proposed rearomatization step, N-xylyl nitrone 54 was prepared
and treated with one equivalent of ethylphenylketene. However,
rather than the expected seven-membered intermediate, dearo-
matized imino-lactone 55 was produced as a 3.5 : 1 mixture of
diastereoisomers. The major diastereoisomer was isolated in
66% yield after crystallization from MeOH, with the relative
configuration within 55 proven by X-ray crystallography
(Scheme 2).41
Next, replacement of the nitrone N-aryl substituent with a
saturated alkyl substituent, thereby removing the potential for
Application to a target compound: anti-cancer agent
Finally, to demonstrate the efficiency of this pericyclic method-
ology and its potential utility in synthesis, Roche p53 inhibi-
tor44 (RO4999200) 1 was selected as a target, with ent-1
synthesized in three simple steps from commercially available
starting materials.45 Recently, Kündig and co-workers reported
the first asymmetric approach toward this species based upon
a palladium-catalyzed intramolecular α-arylation,46 while the
Roche route relies upon chiral HPLC separation.47 Feng and
co-workers also recently reported an expedient catalytic prepa-
ration of 1.48 Our synthesis began with alkylation of commer-
cially available m-anisylacetic acid 62, followed by conversion
to the acid chloride 63, which was isolated in 87% yield over
two steps. Dehydrohalogenation provided ketene 64, which
was used without isolation as a crude solution in THF
(Scheme 4). In congruence with our regioselectivity studies,
treatment of the N-3-chlorophenyl N′-TIPBS nitrone 48 with 64
provided a 60 : 40 mixture of 6- and 4-regioisomeric oxindoles,
which could be readily separated by column chromatography
over silica. Roche p53 inhibitor ent-1 was isolated in 52% yield
Scheme 2 Formation of lactone 55 and representation of its X-ray
crystal structure.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
Acknowledgements
We thank the Royal Society for a URF (ADS), Cancer Research
UK (ER), the Leverhulme Trust (ND) and EPSRC (KBL) for
funding. We are also grateful to the National Science Foun-
dation (USA) for financial support of the research at UCLA
(KNH). We thank J. Douglas and S. M. Leckie for preparation
of previously reported ketenes and Andrew Thomas (Roche) for
data related to oxindole 1. We also thank the EPSRC National
Mass Spectrometry Service Centre (Swansea).
We also thank the UCLA Institute for Digital Research
and Education (IDRE), the XSEDE resources provided by the
XSEDE Science Gateways Program (TG-CHE040013N) of the
National Science Foundation (USA), and TUBITAK ULAKBIM
High Performance and Grid Computing Center (TRUBA,
Turkey) for generous computer time.
Notes and references
1 (a) K. C. Nicolaou, D. J. Edmonds and P. G. Bulger, Angew.
Chem., Int. Ed., 2006, 45, 7134–7186, and references
therein; (b) K. C. Nicolaou and J. S. Chen, Chem. Soc. Rev.,
2009, 38, 2993–3009; (c) A. Padwa, Chem. Soc. Rev., 2009,
38, 3072–3081; (d) B. B. Toure and D. G. Hall, Chem. Rev.,
2009, 109, 4439–4486; (e) E. A. Anderson, Org. Biomol.
Chem., 2011, 9, 3997–4006.
Scheme 4 Application of this methodology to the synthesis of Roche
anti-cancer agent ent-1, and novel analogue 65.
2 (a) K. N. Houk, J. Gonzalez and Y. Li, Acc. Chem. Res., 1995,
28, 81–90; (b) P. H.-Y. Cheong, C. Legault, J. M. Um,
N. Çelebi-Ölçüm and K. N. Houk, Chem. Rev., 2011, 111,
5042–5137. For a related pericyclic cascade approach to
this manuscript that utilizes a nitrone 3 + 2 cycloaddition
followed by [3,3]-sigmatropic rearrangement see: D.-L. Mo,
D. J. Wink and L. L. Andersen, Chem. – Eur. J., 2014, 20,
13217–13225.
3 (a) G. I. Elliott, J. Velicky, H. Ishikawa, Y. Li and
D. L. Boger, Angew. Chem., Int. Ed., 2006, 45, 620–622;
(b) J. Poulin, C. M. Grise-Bard and L. Barriault, Chem. Soc.
Rev., 2009, 38, 3092–3101.
4 For reviews, see: (a) C. Marti and E. M. Carreira, Eur. J. Org.
Chem., 2003, 2209–2219; (b) C. V. Galliford and
K. A. Scheidt, Angew. Chem., Int. Ed., 2007, 46, 8748–8758.
5 M. K. Christensen, K. D. Erichsen, C. Trojel-Hansen,
J. Tjørnelund, S. J. Nielsen, K. Frydenvang, T. N. Johansen,
B. Nielsen, M. Sehested, P. B. Jensen, M. Ikaunieks,
A. Zaichenko, E. Loza, I. Kalvinsh and F. Björkling, J. Med.
Chem., 2010, 53, 7140–7145.
and 96% ee, whilst the novel regioisomer 65 was isolated in
36% yield and 79% ee.
Conclusions
In summary, the optimization of the asymmetric hetero-
Claisen approach to oxindoles has been demonstrated. By vari-
ation of the chiral nitrone structure, N-TIPBS nitrone 20 was
identified as an excellent transmitter of chiral information in
reaction with a variety of alkylarylketenes (>15 examples). A
range of substituted aryl N-TIPBS nitrones was also prepared
and used in a survey of the regioselectivity of this transform-
ation (8 examples). These synthetic studies are consistent with
the previous computational examination of this reaction
system, and the proposed pericyclic cascade mechanism
accounts for the observed levels of enantioselectivity.
Further computational investigation revealed that the tra-
ditionally invoked mechanistic rationale for this process is
energetically improbable under the reaction conditions, and
instead results in a dissociative, non-productive fragmentation
event. In contrast, the proposed 3 + 2, [3,3]-cascade mechan-
ism is found to be a facile, energetically downhill process, and
is substantiated by the isolation of oxazolidinone 59. As a dem-
onstration of synthetic utility, this methodology was used in a
concise, asymmetric preparation of Roche p53 inhibitor ent-1
in 96% ee, along with novel regioisomer 65. Current efforts
within our laboratory are focused on application of this asym-
metric methodology in complex molecule synthesis.
6 T. D. Bagul, G. Lakshmaiah, T. Kawabata and K. Fuji, Org.
Lett., 2002, 4, 249–251.
7 (a) L. E. Overman and Y. Shin, Org. Lett., 2007, 9, 339–341;
(b) J. E. DeLorbe, S. Y. Jabri, S. M. Mennen, L. E. Overman
and F.-L. Zhang, J. Am. Chem. Soc., 2011, 133, 6549–6552.
8 J. Christoffers and A. Mann, Angew. Chem., Int. Ed., 2001,
40, 4591–4597.
9 (a) B. M. Trost and M. K. Brennan, Synthesis, 2009, 3003–
3025; (b) F. Zhou, Y.-L. Liu and J. Zhou, Adv. Synth. Catal.,
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
2010, 352, 1381–1407; (c) L. M. Repka and S. E. Reisman, 30 A brief solvent and temperature screen was undertaken
J. Org. Chem., 2013, 78, 12314–12320.
(see ESI† for details). The use of THF as solvent at a reac-
tion temperature of −78 °C remained optimal for both
yield and enantioselectivity of the product oxindole.
31 The obtained silyl-nitrones were found to hydrolyze on
attempted purification by column chromatography, and
were not amenable to storage. Consequently, these nitrones
were prepared and used immediately as crude residues and
only limited characterization data was obtained. Their
instability, however, rendered them impractical as potential
nitrone chiral auxiliaries.
10 P. B. Alper, C. Meyers, D. R. Lerchner, D. R. Siegel and
E. M. Carreira, Angew. Chem., Int. Ed., 1999, 38, 3186–
3189.
11 S. Edmonson, S. J. Danishefsky, L. Sepp-Lorenzino and
N. Rosen, J. Am. Chem. Soc., 1999, 121, 2147–2155.
12 R. He, C. Ding and K. Maruoka, Angew. Chem., Int. Ed.,
2009, 48, 4559–4561.
13 (a) S. Lee and J. F. Hartwig, J. Org. Chem., 2001, 66, 3402–
3415; (b) J. E. M. N. Klein and R. J. K. Taylor, Eur. J. Org.
Chem., 2011, 6821–6841.
32 CCDC 1029035 contains the supplementary crystallo-
graphic data for 20.
14 A. B. Douney, K. Hatanaka, J. J. Kodanko, M. Oestreich,
L. E. Overman, L. A. Pfeifer and M. M. Weiss, J. Am. Chem. 33 M. J. Frisch, et al., Gaussian 09, revision A.2, Gaussian, Inc.,
Soc., 2003, 125, 6261–6271. Wallingford, CT, 2009. See ESI† for the full citation.
15 (a) L. E. Overman and M. D. Rosen, Angew. Chem., Int. Ed., 34 H. R. Seikaly and T. T. Tidwell, Tetrahedron, 1986, 42, 2587–
2000, 39, 4596–4599; (b) A. B. Douney and L. E. Overman,
Chem. Rev., 2003, 103, 2945–2964.
16 (a) P. E. Kündig and Y.-X. Jia, Angew. Chem., Int. Ed., 2009,
2613.
35 C. E. Cannizzaro, T. Strassner and K. N. Houk, J. Am. Chem.
Soc., 2001, 123, 2668–2669.
48, 1636–1639; (b) R. J. K. Taylor and A. Perry, Chem. 36 π-π interactions displayed in the X-ray structure of nitrone
Commun., 2009, 3249–3251.
20 suggest the importance of dispersion interactions in
determining the selectivity at the cycloaddition step.
Optimizations using dispersion corrected density func-
tional theory gave similar geometries and relative energies;
the results of calculations at the B3LYP-D3/6-311+G(d,p)//
B3LYP-D/6-31G(d) level are given in the ESI.†
17 (a) B. M. Trost and M. Osipov, Angew. Chem., Int. Ed., 2013,
52, 9176–9181; (b) B. M. Trost and M. K. Brennan, Org.
Lett., 2006, 8, 2027–2030.
18 I. D. Hills and G. C. Fu, Angew. Chem., Int. Ed., 2003, 42,
3921–3924.
19 G. S. Singh and Z. Y. Desta, Chem. Rev., 2012, 112, 6104– 37 For all reactions described herein, racemic oxindole
6155.
samples were obtained from reaction of the requisite
achiral diarylnitrone and alkylarylketene (see ESI† for
Experimental details). For the enantioenriched oxindoles,
the absolute configuration is assigned by analogy to that
determined by derivatization in our earlier publication (ref.
25).
20 (a) K. Zheng, C. Yin, X. Lin, L. Lin and X. Feng, Angew.
Chem., Int. Ed., 2011, 50, 2573–2577; (b) B. Tan,
G. Hernandez-Torres and C. F. Barbas III, J. Am. Chem. Soc.,
2011, 133, 12354–12357.
21 (a) G. Lakshmaiah, T. Kawabata, M. Shang and K. Fuji,
J. Org. Chem., 1999, 64, 1699–1704; (b) S. Ma, X. Han, 38 J. E. DeLorbe, D. Horne, R. Jove, S. M. Mennen, S. Nam,
S. Krishnan, S. C. Virgil and B. M. Stoltz, Angew. Chem., Int.
Ed., 2009, 48, 8037–8041.
F.-L. Zhang and L. E. Overman, J. Am. Chem. Soc., 2013,
135, 4117–4128.
22 H. Staudinger and G. Miescher, Helv. Chim. Acta, 1919, 2, 39 After fourteen days, an ee value of 30% was obtained when
554–582. the HPLC sample of oxindole 37 was reanalyzed.
23 C. H. Hassall and A. E. Lippman, J. Chem. Soc., 1953, 1059– 40 N. Çelebi-Ölçüm, B. W. Boal, A. D. Huters, N. K. Garg and
1063. K. N. Houk, J. Am. Chem. Soc., 2011, 133, 5752–5755.
24 M. Hafiz and G. A. Taylor, J. Chem. Soc., Perkin Trans. 1, 41 Stokes and Taylor have described the isolation of a very
1980, 1700–1705.
25 N. Duguet, A. M. Z. Slawin and A. D. Smith, Org. Lett.,
2009, 11, 3858–3861.
26 E. Richmond, N. Duguet, A. M. Z. Slawin, T. Lebl and
A. D. Smith, Org. Lett., 2012, 14, 2762–2765.
27 N. Çelebi-Ölçüm, Y.-h. Lam, E. Richmond, K. B. Ling,
similar lactone species using the same xylylnitrone starting
material: D. P. Stokes and G. A. Taylor, J. Chem. Soc. C,
1971, 2334–2336. The relative configuration within 55 was
confirmed by single crystal X-ray diffraction. CCDC
1029036 contains the supplementary crystallographic data
for 55.
A. D. Smith and K. N. Houk, Angew. Chem., Int. Ed., 2011, 42 The relative configuration within 59 was confirmed by
50, 11478–11482.
28 R. W. Hoffmann, Chem. Rev., 1989, 89, 1841–1860.
single crystal X-ray diffraction. CCDC 1029037 contains the
supplementary crystallographic data for 59.
29 In our previous studies (ref. 26), the corresponding alde- 43 The generation of oxazolidinones in reactions of symmetric
hyde has been isolated from a crude reaction mixture and
retreated with PhNHOH to regenerate TIPBS nitrone 20.
The nitrone has then been shown to be an effective
chiral auxiliary in a second cycle of asymmetric oxindole
synthesis.
ketenes and nitrones has been previously observed and
investigated; for selected examples see: R. N. Pratt,
D. P. Stokes and G. A. Taylor, J. Chem. Soc., Perkin Trans. 1,
1975, 498–503; M. A. Abou-Gharbia and M. M. Jouillie,
J. Org. Chem., 1979, 44, 2961–2966. Through an 17O label-
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Organic & Biomolecular Chemistry
Paper
ing study, Evans and Taylor provided experimental evi-
dence for the proposed 3 + 2 cycloaddition/[1,3]-rearrange-
ment pathway for the formation of oxazolidinones from
diarylnitrones and ketenes. The possibility of α-lactone for-
chemotherapy and in protecting healthy non-cancerous
cells from side effects associated with chemotherapy. For
further details, see: A. V. Gudkov and E. A. Komarova,
Biochem. Biophys. Res. Commun., 2005, 331, 726–736.
mation in this transformation via oxygen atom transfer to 45 Using the L-serine derived chiral nitrone, the furnished
ketenes was also probed. However, a subsequent compet-
ing deoxygenation-recombination mechanism involving
the addition of α-lactones to nitriles and imines was con-
cluded to be an unlikely mechanistic pathway for this
oxindole is the enantiomer of the reported Roche com-
pound. The absolute configuration of 1 prepared in this
manuscript was assigned by HPLC analysis by analogy to
ref. 46. Please see ESI† for further information.
transformation. For further details, see: A. R. Evans and 46 D. Katayev and E. P. Kündig, Helv. Chim. Acta, 2012, 95,
G. A. Taylor, J. Chem. Soc., Perkin Trans. 1, 1987, 567–569. 2287–2295.
44 p53 is a tumour suppressor protein that is lost or sup- 47 K. C. Luk, S. S. So, J. Zhang and Z. Zhang and
pressed in the majority of cancer cells. Inhibition of F. Hoffmann-La Roche, AG patent, WO2006/136606, 2006.
p53 has been shown to have two main therapeutic appli- 48 J. Guo, S. Dong, Y. Zhang, Y. Kuang, X. Liu, L. Lin and
cations: in sensitizing cancer cells to other forms of
X. Feng, Angew. Chem., Int. Ed., 2013, 52, 10245–10249.
This journal is © The Royal Society of Chemistry 2014
Org. Biomol. Chem.