Dormant versus evolving aminopalladated intermediates: Toward a unified mechanistic scenario in PdII-catalyzed aminations

Article abstract of DOI:10.1002/chem.201302744

Pd^II-catalyzed alkene aminopalladation and allylic Ci￡H activation are two competing reaction sequences sharing the same reaction conditions. This study aimed at understanding the factors that bias one or the other path in the intramolecular oxidative cyclization of two types of N-tosyl amidoalkenes. The results obtained are in accord with the initial generation of a high-energy cyclic (5- or 6-membered) aminopalladated intermediate. However, this latter species can evolve only if the following specific conditions are met: the availability of distocyclic β-H elimination pathway, the presence of a strong terminal oxidant, or the availability of a carbopalladation pathway. Conversely, the cyclic alkylpalladium complex is only a latent species in equilibrium with the initial substrate and cannot evolve. Such a reactivity hurdle leaves the way open for alternative reactivities such as allylic Ci￡H activation of the olefinic substrate to generate a η³-allyl complex followed by its interception by the nitrogen nucleophile, [3,3]-sigmatropic rearrangement, or decomposition. This study proposes a unifying mechanistic picture that connects these competing mechanisms. Asleep or awake? Unsaturated N nucleophiles react through two competing pathways under Pd^II catalysis: Ci￡H allylic activation and aminopalladation. New data are in accord with the initial generation of a high-energy cyclic aminopalladated intermediate (API) that can either evolve along different pathways such as β-H elimination, oxidation_, or carbopalladation, or lay dormant, the latter leading to alternative types of reactivity, such as allylic Ci￡H activation or [3,3]-sigmatropic rearrangement, gaining the upper hand (see scheme; Ts=p-toluenesulfonyl). Copyright

Full text of DOI:10.1002/chem.201302744

DOI: 10.1002/chem.201302744
Full Paper
&
Palladium Catalysis
Dormant versus Evolving Aminopalladated Intermediates: Toward
a Unified Mechanistic Scenario in PdII-Catalyzed Aminations
Jamshid Rajabi,[a, b] Mꢀlanie M. Lorion,[a, b] Vu Linh Ly,[a, b] Frꢀdꢀric Liron,[a, b] Julie Oble,[a, b]
Guillaume Prestat,[a, b, c] and Giovanni Poli*[a, b]
Abstract: PdII-catalyzed alkene aminopalladation and allylic
CꢀH activation are two competing reaction sequences shar-
ing the same reaction conditions. This study aimed at under-
standing the factors that bias one or the other path in the
intramolecular oxidative cyclization of two types of N-tosyl
amidoalkenes. The results obtained are in accord with the in-
itial generation of a high-energy cyclic (5- or 6-membered)
aminopalladated intermediate. However, this latter species
can evolve only if the following specific conditions are met:
the availability of distocyclic b-H elimination pathway, the
presence of a strong terminal oxidant, or the availability of
a carbopalladation pathway. Conversely, the cyclic alkylpalla-
dium complex is only a latent species in equilibrium with
the initial substrate and cannot evolve. Such a reactivity
hurdle leaves the way open for alternative reactivities such
as allylic CꢀH activation of the olefinic substrate to generate
a h3-allyl complex followed by its interception by the nitro-
gen nucleophile, [3,3]-sigmatropic rearrangement, or decom-
position. This study proposes a unifying mechanistic picture
that connects these competing mechanisms.
amino-[1b] and oxy-[3] palladation processes, including the very
popular Wacker process, follow this path. Alternatively, dehy-
dropalladation can be avoided in the presence of a nucleophilic
oxidizing agent that can oxidatively cleave the PdꢀC bond of
II,[4] in some cases passing through a PdIV intermediate III
(path a2).[5,6] On the other hand, if the substrate bears an allylic
hydrogen atom, its removal from I (path b) can generate an h3-
allyl palladium complex IV, which can be in situ trapped by
a nucleophilic species to deliver the final compound.[7] This
latter path is usually followed by cyclic or terminal alkenes.[8]
Finally, if the PdII-complexed alkene carries a juxtaposed unsa-
turated moiety, a PdII-catalyzed [3,3]-sigmatropic process can
take place, this time isohypsic for palladium (path c).
Introduction
Literature covering the oxidative PdII-catalyzed addition of nu-
cleophiles to olefins is rich and shows that a number of new
synthetically useful compounds can be built-up starting from
simple unsaturated substrates.[1] Despite their considerable
synthetic utility, the mechanism of these processes are far from
being clear, as different and sometimes hardly distinguishable
mechanistic paths can be operative (Scheme 1).[2] Indeed, after
activation of the unsaturated moiety by the PdII catalyst to
give complex I, a nucleopalladation reaction can take place
(path a), thereby affording s-alkyl palladium complex II.
This latter can evolve, usually through dehydropalladation
(path a1), affording the final product with the concomitant re-
duction of PdII to Pd0. Oxidation of this latter species by a ter-
minal oxidizing agent closes the catalytic cycle. A number of
Despite the documented description of these mutually ex-
cluding reactivity patterns, a study aimed at elucidating the
rules governing path selection was, to the best of our knowl-
edge, still wanting.
[a] Dr. J. Rajabi, M. M. Lorion, V. L. Ly, Dr. F. Liron, Dr. J. Oble,
Prof. Dr. G. Prestat, Prof. Dr. G. Poli
Results
Sorbonne Universitꢀs, UPMC Univ Paris 06, UMR 7201
Institut Parisien de Chimie Molꢀculaire, 75005 Paris (France)
FR2769 Institut de Chimie Molꢀculaire
Choice of substrates and methods
E-mail: giovanni.poli@upmc.fr
Accordingly, we decided to investigate the Pd-catalyzed oxida-
tive cyclization of several linear nitrogen-containing derivatives
featuring different tethers between the nitrogen atom and the
unsaturated moiety, and by using different terminal oxidizing
agents.
[b] Dr. J. Rajabi, M. M. Lorion, V. L. Ly, Dr. F. Liron, Dr. J. Oble,
Prof. Dr. G. Prestat, Prof. Dr. G. Poli
CNRS, UMR 7201, Institut Parisien de Chimie Molꢀculaire
75005 Paris (France)
[c] Prof. Dr. G. Prestat
Present address:
Our recent results on direct allylic amination in AcOH repre-
sented the starting point of the present project.[9] Indeed, we
recently reported that treatment of N-tosyl carbamates 1a and
1b with Pd(OAc)2 (0.1 equiv), White’s disulfoxide ligand PhS(O)-
(CH2)2S(O)Ph (0.15 equiv), and phenylbenzoquinone (PhBQ;
Universitꢀ Paris Descartes, UMR 8601 CNRS
Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques
75006 Paris (France)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201302744.
Chem. Eur. J. 2014, 20, 1539 – 1546
1539
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 1. Possible paths for the oxidative PdII-catalyzed addition of nucleophiles to alkenes. Color codes for paths: a/a1, aminopalladation/dehydropallada-
tion: blue; a2, oxidative coupling: green; b, allylic CꢀH activation: red; c: [3,3]-sigmatropic rearrangement: magenta. Z and QX2 are the terminal stoichiometric
oxidants.
1.07 equiv) in AcOH at 458C (Table 1, method A), gave the cor-
responding 4-vinyl-oxazolidinone 2a and oxazinanone 2b, re-
spectively (Table 1, entries 1 and 2). On the other hand, allyl
carbamate 1c gave allyl N-tosyl amine 2c’ (Table 1, entry 3).[10]
N-Tosyl carboxamides were next considered. Pent-4-enoic acid
N-tosyl amide 1e gave only degradation products, whereas
the analogous treatment of the homologous hex-5-enoic acid
N-tosyl amide 1 f and hept-6-enoic acid N-tosyl amide 1g re-
stored explicit reactivities. In particular, whereas 1 f afforded
a virtually inseparable 71:29 mixture of the direct 5-exo amina-
tion product, 5-vinyl-pyrrolidin-2-one 2 f, and 6-methylidene-pi-
peridin-2-one 2 f’ obtained from a 6-exo aminopalladation/de-
hydropalladation path (Table 1, entry 6), 1g gave exclusively
the 6-exo direct allylic amination product, 6-vinyl-pyrrolidin-2-
one 2g (Table 1, entry 7). Finally, (E)-hex-4-enoic acid N-tosyl
amide 1h gave 5-vinyl-pyrrolidin-2-one 2 f in a moderate 54%
yield (Table 1, entry 8), the same compound as previously ob-
tained from 1 f. These results indicate a clear CꢀH activation/al-
lylation (direct allylic amination) type of reactivity for the case
of 1a, 1b, 1 f, and 1g, whereas a likely aminopalladation/dehy-
dropalladation type of reactivity for 1h and, to a small extent,
1 f. Subsequent treatment of the same substrates, under iden-
tical conditions as above, except for the use of PhI(OAc)2 in-
stead of PhBQ as oxidizing agent, gave very interesting results.
Indeed, carbamates 1a and 1c, as well as amides 1e and 1 f,
gave exclusively the corresponding products of aminoacetoxy-
lation, 1a and 1 f, with random endo/exo selectivity. Carbamate
1h gave a 62:38 mixture of vinyl- and b-styryl pyrrolidones 2 f
and 3h. Finally, carbamates 1b and 1d, as well as amide 1g,
gave only intractable material.
Discussion
Although seemingly erratic at a first glance, the above results
become coherent assuming the setting up of a rapid equilibri-
um[11] between the substrate and the corresponding cyclic ami-
nopalladated intermediate (API),[12] coupled with a totally or
partially inhibited b-H elimination of the API, involving the
cyclic hydrogen atom (proxicyclic b-H elimination).[13,14]
Thus, in the absence of this way of evolution (see below),
and owing to the known reversibility of the aminopalladation
step, the substrate can take alternative reaction paths.[15] In
this case, the cyclic aminopalladated intermediate (API) is just
a dormant (off-cycle) intermediate in the catalytic cycle, being
gradually drained into the active stream during the reaction
course (equation 1).
Chem. Eur. J. 2014, 20, 1539 – 1546
www.chemeurj.org
1540
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
allylic amination (Scheme 2). No-
tably, although cyclic API from
carbamate 1a appears to be to-
tally dormant, that from carboxa-
mide 1 f is consumed through
a minor, yet clear pathway, that
is, competitive proxicyclic dehy-
dropalladative “leakage” to give
2 f’.
Table 1. Behavior of unsaturated N-tosyl carbamates under oxidative PdII catalysis.[a]
Entry
Substrate
Product(s)[b]
Product(s)[c]
method B
method A and/or A’
1
Homologous substrates 1b
and 1g also afford the corre-
sponding products of direct allyl-
ic amination, 2b and 2g, respec-
tively, likely due to a less favored
7-exo aminopalladation step.
[d]
2
3
–
In the case of substrate 1c,
we can also assume a similar
rapid and reversible equilibrium
involving a dormant cyclic API.
However, a [3,3]-sigmatropic re-
arrangement[16] can concurrently
take place, affording, after decar-
boxylation, the observed allyl N-
tosyl amine 2c’ (Scheme 3).[17]
[d]
4
–
5
6
Method A appears ineffective
for internal alkene 1d. Inhibition
of CꢀH allylic activation was ex-
pected, owing to the known se-
lectivity for terminal alkenes of
the disulfoxide system. Less ex-
pected was the inhibition
toward aminopalladation and
[3,3]-sigmatropic rearrangement.
However, the use of harsher re-
action conditions are more suc-
cessful (see below).
[d]
7
8
–
Treatment of carboxamide 1e,
by using method A, leads to
only intractable material, in ac-
cordance with the postulated
equilibrium involving a dormant
cyclic API and final degradation.
On the other hand, carboxamide
1h, bearing an extra methyl
group with respect to 1e, can
generate a 5-membered API that
can easily evolve through disto-
cyclic b-H elimination, whereas
CꢀH allylic activation is automat-
ically inhibited owing to the internal position of the alkene
and [3,3]-sigmatropic rearrangement is in this case no longer
an option. Therefore, formation of 5-vinyl-pyrrolidin-2-one 2 f
through the aminopalladation pathway is not surprising
(Scheme 4).
[a] Color codes for path origin of products: direct allylic amination: red; aminopalladation/b-H elimination:
blue; aminopalladation/acetoxylation: green; [3,3]-sigmatropic rearrangement: magenta. The percentage
values in parentheses refer to yields upon isolation. [b] Starting material was recovered. [c] Method B: Pd(OAc)2
(0.1 equiv), PhS(O)(CH2)2S(O)Ph (0.15 equiv), AcOH, PhI(OAc)2 (2.1 equiv), 458C. [d] Degradation. [e] The reaction
was performed in CH2Cl2. [f] Traces of double-bond isomers were also detected. [g] 2 f and 3h were obtained
as an inseparable mixture in the presence of traces of other side products.
Method A
Accordingly, in the case of treatment of carbamate 1a and car-
boxamide 1 f by using method A, we can assume the genera-
tion of a dormant 6-membered API, coupled with a slower, yet
irreversible, allylic CꢀH activation of the substrate to give the
corresponding h3-allyl derivative, the final interception of
which by the nitrogen nucleophile gives the product of direct
The above results imply that the simple formal slippage of
the unsaturation in the substrate chain is associated with
a switching of the operating mechanism from CꢀH allylic acti-
Chem. Eur. J. 2014, 20, 1539 – 1546
www.chemeurj.org
1541
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 6. Activation of the cyclic APIs by PhI(OAc)2.
Scheme 2. Allylic CꢀH activation of substrates involving a dormant cyclic
API.
reversible generation of an inherently dormant cyclic API, and
its activation by PhI(OAc)2 that derails the original CꢀH activa-
tion reactivity (Scheme 6).[18] Notably, PhI(OAc)2 either switches
the reactivity toward aminoacetoxylation or prevents any pos-
sibility of allylic CꢀH activation that might otherwise occur,
thus affording degradation products, as in the case of 1b and
1g.[19]
Treatment of 1h according to method B produced a mixture
of 5-vinyl-pyrrolidone 2 f and 5-styryl-pyrrolidone 3h, the latter
very likely derived from Mizoroki–Heck-type coupling between
2 f and in situ generated PhI.[20] Independent of this initially un-
expected[21] pseudodomino[22] reactivity of 2 f, these results
suggest that in this case, b-H elimination dominates over po-
tential PhI(OAc)2-promoted oxidative coupling (Scheme 7).
Scheme 3. [3,3]-Sigmatropic rearrangement of allylic carbamate 1c.
Scheme 7. Behavior of carboxamide 1h under method B.
Scheme 4. Behavior of carboxamide 1h under method A.
Method A’
In the hope of improving upon some of results obtained with
method A, we repeated some experiments using an aerobic
method (method A’), which was previously used by Larock and
Scheme 5. Isomeric substrates afford the same product under identical reac-
tions conditions, by different mechanisms.
co-workers for the oxidative cyclization of N-tosyl amines[23]
:
Pd(OAc)2 (0.05 equiv), NaOAc (2.0 equiv), DMSO, O2 (1 atm),
808C ; Table 1. In the event, treatment of 1c under method A’
led to an improved yield of N-tosyl allylamine 2c’ (Table 1,
entry 3, 65%; compare with 54% obtained by using meth-
od A), whereas 1d, unreactive under method A, gave a mixture
of 4-vinyl oxazolidinone 2a, N-tosyl 3-aminobutene 2d’ and N-
tosyl 4-acetoxy-3-aminobutene 2d’’ (Table 1, entry 4). This reac-
tivity reflects the competition between the [3,3]-sigmatropic
rearrangement path, giving the allylic N-tosyl amine 2d, and
the distocyclic b-H elimination path, affording the oxazolidi-
none 2a. This latter species, owing to the high temperature
used for method A’, suffers partial ring opening at C-4 by ace-
tate anion followed by decarboxylation, thus affording the
aminoacetoxy derivative 2d’’ (Scheme 8).
vation to aminopalladation. Indeed, isomeric substrates 1 f and
1h afford the same product 2 f under identical reaction condi-
tions, yet through different mechanisms (Table 1, entries 6 and
8; Schemes 2 and 4; Scheme 5).
Method B
We next repeated all the experiments of Table 1, replacing
PhBQ with PhI(OAc)2 as the terminal oxidizing agent (Table 1,
method B). Quite interestingly, carbamates 1a, 1c, as well as
carboxamides 1e and 1 f, all gave 1,2-aminoacetoxylation
products, clearly the result of oxidative evolution of the corre-
sponding cyclic APIs. These experiments are in accord with the
Chem. Eur. J. 2014, 20, 1539 – 1546
www.chemeurj.org
1542
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
equimolar mixture of 1a and Pd(OAc)2/PhS(O)(CH2)2S(O)Ph at
358C in CD2Cl2 over time (Figure 1). The spectrum cleanly
shows the progressive disappearance of substrate 1a and the
concomitant appearance of product 2a, without apparent for-
mation of either the corresponding cyclic API or the h3-allylpal-
ladium complex. This result is thus consistent with thermody-
namically unfavored cyclic API as well as h3-allylpalladium com-
plex.
Accordingly, for the evolution of model substrate 1a into
either 2a (method A) or into 3a+3a’ (method B), one can
draw a qualitative energetic profile (Figure 2). This profile well
illustrates all of the crucial features of the system: a) the lack
of accumulation (and thus lack of detection) of the cyclic API
as well as the h3-allylpalladium complex; b) the lack of accessi-
bility of the proxicyclic dehydropalladation product; c) the acti-
vation of the cyclic API in the presence of PhI(OAc)2.
Scheme 8. Competition between [3,3]-sigmatropic rearrangement and disto-
cyclic b-H elimination of allylic carbamate 1d.
Trapping of the cyclic API through carbopalladation
We next thought that a way to further confirm the involve-
ment of the elusive cyclic APIs could involve a built-in juxta-
posed unsaturated moiety with which an otherwise inherently
dormant API complex would react through a carbopalladation/
dehydropalladation sequence. Accordingly, we submitted N-
tosyl 3-vinylhept-5-enamide 4 to the reaction conditions char-
acterized by method A. Indeed, 4 can be formally considered
as 1e (which proved to be unreactive under method A) bear-
ing a juxtaposed extra unsaturated moiety. In the event, the
planned domino reaction did take place, affording fused pyrro-
lidin-2-one 5 as a single diastereoisomer in 10% or 39% yield,
Finally, carboxamide 1e, also unreactive under method A,
gave, under method A’, dihydropyridinone 2e (20% yield),
a result that may be understood in terms of a 6-endo amino-
palladation followed by a b-H elimination.[24]
Spectroscopic study
To gain further information about the postulated hidden equi-
librium between the substrate and the corresponding cyclic
1
API, we analyzed the evolution of an H NMR spectrum of an
1
Figure 1. Time evolution of the olefinic portion of an H NMR spectrum of an equimolar solution of 1a and Pd(OAc)2/PhS(O)(CH2)2S(O)Ph at 358C in CD2Cl2.
Chem. Eur. J. 2014, 20, 1539 – 1546
www.chemeurj.org
1543
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ized, transformations. Thus, for
example, the present study clari-
fies that intramolecular direct al-
lylic amination of unsaturated N-
tosyl carbamates can be ach-
ieved only if several conditions
are met at the same time: a) the
alkene has to generate a dor-
mant cyclic API that permits the
(slower) allylic CꢀH activation
step to take place; b) the nitro-
gen atom has to be of a carba-
mate or amide nature, proxicy-
clic b-H elimination of which is
forbidden; c) allowed ring clo-
sures are of a 5- or a 6-exo type.
The so-far perplexing behavior
of N-trichloroacetyl allylic carba-
mates[27] under PdII catalysis in
the absence or presence of CuII
can be now easily rationalized
according to a [3,3]-sigmatropic
Figure 2. Qualitative energetic profile showing the evolution of substrate 1a into 2a (method A, right) or into 3a
(method B, left). Structure 3a’ is omitted for clarity.
when using 10 mol% or 26 mol% Pd(OAc)2, respectively. Thus,
despite an inefficient Pd turnover, this result is in total accord
with the involvement of the postulated elusive cyclic API,
thereby validating our concept. However, to render this trans-
formation of synthetic relevance, Pd catalysis had to be
reached to some extent. We reasoned that the inefficient Pd
turnover was likely due, among other things, to an unfavorable
equilibrium in the reductive-elimination step, AcO[Pd]HQ-
[Pd0]+AcOH, under the acidic conditions of method A’.[25]
Indeed, when applying the basic conditions of method A’, the
yield of 5 could be improved to 40% by using 10 mol%
Pd(OAc)2 (Scheme 9). Although further improvements in yield
Scheme 9. Activation of cyclic API by intramolecular carbopalladation. Con-
ditions i (method A): Pd(OAc)2 (0.1 equiv), [PhS(O)(CH2)]2 (0.15 equiv), AcOH,
PhBQ (1.1 equiv), 458C, 18 h; from 1e: only decomposition (see also Table 1,
entry 5); from 4: (10%); Conditions i’ (method A): Pd(OAc)2 (0.26 equiv),
[PhS(O)(CH2)]2 (0.36 equiv), AcOH, PhBQ (1.1 equiv), 458C, 18 h; from 4:
(39%); Conditions ii (method A’): Pd(OAc)2 (0.1 equiv), NaOAc (2 equiv),
DMSO, O2, 808C, 24 h; from 4: (40%).
are still desirable, this transformation under method A’ entails
63% yield per step on average and represents an original cata-
lytic pure domino sequence.[22]
Conclusion
In conclusion, the present study sheds light on the mechanistic
relationship between competing types of PdII-catalysis involv-
ing unsaturated N-tosyl carbamates and carboxamides, so far
only disjointedly reported. The work is founded on the recog-
nition of the reversible formation of cyclic APIs, and rationaliz-
es the observed reactivities on the basis of either cyclic API ac-
tivation[26] (through distocyclic b-H elimination, carbopallada-
tion, or oxidation with PhI(OAc)2), or the involvement of alter-
native irreversible reactivities such as allylic CꢀH activation or
decarboxylative [3,3]-sigmatropic rearrangement. Because such
competing paths originate from the same starting substrate,
inhibition of a fast, yet reversible, path activates the next more
facile one.
reactivity in the former case and a CuCl2-based cyclic API acti-
vation in the latter one (Scheme 10). Future studies will ad-
dress, among other things, stereochemical aspects of the
above transformations.[28]
Experimental Section
General procedure for direct allylic amination (method A)
A round-bottom flask under air was charged with the substrate
(1.0 equiv), Pd(OAc)2 (0.1 equiv), disulfoxide ligand (0.15 equiv),
phenylbenzoquinone (1.07 equiv), and AcOH (0.2m). The reaction
mixture was allowed to stir at 458C for 24 h. The reaction mixture
was hydrolyzed with a saturated aqueous solution of NH4Cl
(10 mL) and extracted with AcOEt (3ꢂ10 mL). The combined or-
ganic layers were dried over MgSO4, filtered, and concentrated in
Application of the above concepts allows an understanding
of path selection for reported, yet mechanistically non rational-
Chem. Eur. J. 2014, 20, 1539 – 1546
www.chemeurj.org
1544
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
13C NMR (101 MHz, CDCl3): d=173.5, 170.2, 145.1, 135.5, 129.4,
128.2, 65.4, 57.6, 30.9, 22.5, 21.5, 20.4 ppm; IR (film): n˜ =3070,
2955, 1740, 1596, 1452, 1358, 1223 cmꢀ1; HRMS (ESI): m/z calcd for
C14H17NNaO5S: 334.07251 [M+Na]+; found: 334.07243.
Acknowledgements
Scheme 10. Behaviour of N-trichloroacetyl allylic carbamates with or without
cyclic API activation. Conditions: i) Pd(OAc)2 (5 mol%), LiCl (2 equiv), THF,
608C (87%). ii) Pd(OAc)2 (5 mol%), CuCl2 (4 equiv), LiCl (2 equiv), THF, 258C
(70%).
CNRS and UPMC are acknowledged for financial support. V.L.L.
thanks the international exchange program JCEMolChem for fi-
nancial support. We thank FranÅois Hoang–Aoiro Studio for de-
signing the cover of this article.
vacuo. The residue was taken up in CH2Cl2 (45 mL) and treated
with an equal volume of 1N KOH for 30 min. After decantation and
phase separation, the organic layer was dried over MgSO4 and con-
centrated in vacuo. Purification by silica gel column chromatogra-
phy (cyclohexane/AcOEt, 80:20) afforded the desired product.
Keywords: [3,3]-sigmatropic
rearrangement
·
aminopalladation · beta-hydride elimination · CꢀH activation ·
palladium
General procedure for direct allylic amination (method A’)
[1] For books and reviews, see: a) J. Tsuji in Palladium Reagents and Cata-
lysts, 2nd ed., Wiley, New York, 2004; b) R. I. McDonald, G. Liu, S. S.
Stahl, Chem. Rev. 2011, 111, 2981–3019; c) T. Hoosokawa in Handbook
of Organopalladium Chemistry for Organic Synthesis; Aminopalladation
and Related Reactions Involving Other Group 15 Atom Nucleophiles; Ami-
nopalladation-Dehydropalladation and Related Reactions (Ed.: E.-i. Ne-
gishi), John Wiley & Sons, Inc., New York, 2003; d) A. Minatti, K. MuÇiz,
Chem. Soc. Rev. 2007, 36, 1142–1152; e) J. L. Jeffrey, R. Sarpong, Chem.
Sci. 2013, 4, 4092–4106; f) J. P. Wolfe, J. D. Neukom, D. H. Mai in Cata-
lyzed Carbon-Heteroatom Bond Formation (Ed.: A. K. Yudin), Wiley-VCH,
Weinheim, 2010, pp. 1–34.
Into a round-bottom flask were added the substrate (1.0 equiv),
Pd(OAc)2 (0.05 equiv), NaOAc (2 equiv), and DMSO (0.05m). The
flask was flushed with O2 (balloon). The reaction mixture was
stirred at 808C for the appropriate reaction time. The mixture was
then cooled to room temperature, diluted with ether (10 mL), and
hydrolyzed with saturated aq. NH4Cl solution (10 mL). The organic
layer was washed with saturated aq. NaCl solution (10 mL). The
layers were separated and the aqueous phase was extracted with
ether (3ꢂ30 mL). The combined organic layers were washed with
10% aq. NaCl solution (20 mL) and dried over MgSO4. The solvent
was evaporated in vacuo. Purification by silica gel column chroma-
tography (cyclohexane/AcOEt, 80:20), afforded the desired prod-
uct.
[2] Throughout this article, brackets around palladium atom intend to
render implicit the dative ligands.
[3] S. Cacchi, A. Arcadi in Handbook of Organopalladium Chemistry for Or-
ganic Synthesis (Ed.: E.-i. Negishi), John Wiley & Sons, 2002, 2289.
[4] For oxidative cleavage of the CꢀPd bond by using CuII salts, see: a) G.
Broggini, V. Barbera, E. M. Beccalli, U. Chiacchio, A. Fasana, S. Galli, S.
Gazzola, Adv. Synth. Catal. 2013, 355, 1640–1648; b) A. Lei, X. Lu, G. Liu,
Tetrahedron Lett. 2004, 45, 1785–1788; c) M. R. Manzoni, T. P. Zabawa,
D. Kasi, S. R. Chemler, Organometallics 2004, 23, 5618–5621; d) J.-E.
Bꢃckvall, Acc. Chem. Res. 1983, 16, 335–342.
[5] For reviews, see: a) K. MuÇiz, Angew. Chem. 2009, 121, 9576–9588;
Angew. Chem. Int. Ed. 2009, 48, 9412–9423; b) P. Sehnal, R. J. K. Taylor,
I. J. S. Fairlamb, Chem. Rev. 2010, 110, 824–889; c) T. W. Lyons, M. S. San-
ford, Chem. Rev. 2010, 110, 1147–1169; d) N. R. Deprez, M. S. Sanford,
Inorg. Chem. 2007, 46, 1924–1935.
General procedure for oxidative aminoacetoxylation
(method B)
A round-bottom flask under air was charged with the substrate
(1.0 equiv), Pd(OAc)2 (0.1 equiv), disulfoxide ligand (0.15 equiv), io-
dobenzene diacetate (2.1 equiv), and AcOH (0.2m). The reaction
mixture was allowed to stir at 458C for the appropriate reaction
time. The reaction mixture was hydrolyzed with a saturated aque-
ous solution of NH4Cl (10 mL) and extracted with AcOEt (3ꢂ
20 mL). The combined organic layers were dried over MgSO4, fil-
tered, and concentrated in vacuo. Purification by column chroma-
tography (cyclohexane/AcOEt, 80:20) afforded the desired product.
[6] E. J. Alexanian, C. Lee, E. J. Sorensen, J. Am. Chem. Soc. 2005, 127,
7690–7691.
[7] For books and reviews, see: a) G. Liu, Y. Wu, Top. Curr. Chem. 2009, 292,
195–209; b) B. ꢄkermark, K. Zetterberg in Handbook of Palladium
Chemistry for Organic Synthesis (Ed.: E.-i. Negishi), John Wiley & Sons,
2002, 1875–1885; c) T. Jensen, P. Fristrup, Chem. Eur. J. 2009, 15, 9632–
9636; d) C. J. Engelin, P. Fristrup, Molecules 2011, 16, 951–969; e) J.
Muzart, Bull. Soc. Chim. Fr. 1986, 65–77; f) H. Greenberg, J.-E. Bꢃckvall in
Transition Metals for Organic Synthesis (Eds.: M. Beller, C. Bolm), Wiley-
VCH, 2004, pp. 243–265.
[8] For examples of Pd-catalyzed CꢀH activation followed by O-nucleophile
trapping, see: a) P. E. Gormisky, M. C. White, J. Am. Chem. Soc. 2011, 133,
12584–12589; b) K. J. Fraunhoffer, N. Prabagaran, L. E. Sirois, M. C.
White, J. Am. Chem. Soc. 2006, 128, 9032–9033; c) E. M. Stang, M. C.
White, Nat. Chem. 2009, 1, 547–551; d) E. M. Stang, M. C. White, Angew.
Chem. 2011, 123, 2142–2145; Angew. Chem. Int. Ed. 2011, 50, 2094–
2097; e) N. A. Vermeulen, J. H. Delcamp, M. C. White, J. Am. Chem. Soc.
2010, 132, 11323–11328; f) D. J. Covell, N. A. Vermeulen, N. A. Labenz,
M. C. White, Angew. Chem. 2006, 118, 8397–8400; Angew. Chem. Int. Ed.
2006, 45, 8217–8220; g) K. J. Fraunhoffer, D. A. Bachovchin, M. C. White,
Org. Lett. 2005, 7, 223–226; h) W. H. Henderson, C. T. Check, N. Proust,
J. P. Stambuli, Org. Lett. 2010, 12, 824–827; i) E. Thiery, C. Aouf, J. Belloy,
D. Harakat, J. Le Bras, J. Muzart, J. Org. Chem. 2010, 75, 1771–1774;
j) L. T. Pilarski, P. G. Janson, K. J. Szabꢅ, J. Org. Chem. 2011, 76, 1503–
3-(p-Tolylsulfonyl)-4-vinyl-1,3-oxazinan-2-one (2b)
1H NMR (400 MHz, CDCl3): d=7.90 (d, J=8.1 Hz, 2H), 7.29 (d, J=
8.0 Hz, 2H), 5.87 (ddd, J=16.9, 10.5, 5.1 Hz, 1H), 5.41 (dd, J=10.3,
1.2 Hz, 1H), 5.40 (dd, J=16.9, 1.4 Hz, 1H), 5.25–5.23 (m, 1H), 4.34–
4.27 (m, 2H), 2.42 (s, 3H), 2.32–2.22 (m, 1H), 2.02–1.91 ppm (m,
1H); 13C NMR (50 MHz, CDCl3): d=148.3, 144.9, 135.6, 135.0, 129.3,
129.1, 118.8, 64.5, 56.4, 27.3, 21.6 ppm; IR (thin film): n˜ =2925,
1723, 1596, 1482, 1410, 1345, 1290, 1266, 1154 cmꢀ1; HRMS (ESI):
m/z calcd for C13H15NNaO4S: 304.06140 [M+Na]+; found:
304.06168.
[5-Oxo-1-(p-tolylsulfonyl)pyrrolidin-2-yl]methyl acetate (3e)
1H NMR (400 MHz, CDCl3): d=7.90 (d, J=8.4 Hz, 2H), 7.29 (d, J=
8.6 Hz, 2H), 4.62–4.56 (m, 1H), 4.47 (dd, J=11.8, 4.4 Hz, 1H), 4.16
(dd, J=11.8, 2.8 Hz, 1H), 2.56 (ddd, J=13.3, 10.5, 5.9 Hz, 1H), 2.40
(s, 3H), 2.36–2.23 (m, 2H), 2.06–1.91 (m, 1H), 1.88 ppm (s, 3H);
Chem. Eur. J. 2014, 20, 1539 – 1546
www.chemeurj.org
1545
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
1506; k) D. J. Covell, M. C. White, Angew. Chem. 2008, 120, 6548–6551;
Angew. Chem. Int. Ed. 2008, 47, 6448–6451; l) B.-L. Lin, J. A. Labinger,
J. E. Bercaw, Can. J. Chem. 2009, 87, 264–271; m) C. Le, K. Kunchithapat-
ham, W. H. Henderson, C. T. Check, J. P. Stambuli, Chem. Eur. J. 2013, 19,
11153–11157; for examples of Pd-catalyzed CꢀH activation followed by
N-nucleophile trapping, see: n) E. M. Beccalli, G. Broggini, G. Paladino,
A. Penoni, C. Zoni, J. Org. Chem. 2004, 69, 5627–5630; o) K. J. Fraun-
hoffer, M. C. White, J. Am. Chem. Soc. 2007, 129, 7274–7276; p) X. Qi,
G. T. Rice, M. S. Lall, M. S. Plummer, M. C. White, Tetrahedron 2010, 66,
4816–4826; q) G. Liu, G. Yin, L. Wu, Angew. Chem. 2008, 120, 4811–
4814; Angew. Chem. Int. Ed. 2008, 47, 4733–4736; r) L. Wu, G. Qiu, G.
Liu, Org. Lett. 2009, 11, 2707–2710; s) S. A. Reed, M. C. White, J. Am.
Chem. Soc. 2008, 130, 3316–3318; t) S. A. Reed, A. R. Mazzotti, M. C.
White, J. Am. Chem. Soc. 2009, 131, 11701–11706; u) M. R. Luzung, C. A.
Lewis, P. S. Baran, Angew. Chem. 2009, 121, 7159–7163; Angew. Chem.
Int. Ed. 2009, 48, 7025–7029; v) I. I. Strambeanu, M. C. White, J. Am.
Chem. Soc. 2013, 135, 12032–12037; for examples of Pd-catalyzed CꢀH
activation followed by C-nucleophile trapping, see: w) A. J. Young, M. C.
White, J. Am. Chem. Soc. 2008, 130, 14090–14091; x) A. J. Young, M. C.
White, Angew. Chem. 2011, 123, 6956–6959; Angew. Chem. Int. Ed.
2011, 50, 6824–6827; y) B. M. Trost, M. M. Hansmann, D. A. Thaisri-
vongs, Angew. Chem. Int. Ed. 2012, 51, 4950–4953; for examples of Pd-
catalyzed CꢀH activation followed by miscellaneous trapping, see:
z) J. M. Larsson, T. S. N. Zhao, K. J. Szabꢅ, Org. Lett. 2011, 13, 1888–1891;
aa) E. M. Stang, M. C. White, J. Am. Chem. Soc. 2011, 133, 14892—14895;
for a recent mechanistic study, see: ab) C. Engelin, T. Jensen, S. Rodri-
guez-Rodriguez, P. Fristrup, ACS Catal. 2013, 3, 294–302.
[17] In contrast to the other cases, this transformation is isohypsic for the
palladium catalyst, which remains in oxidation state II. As a conse-
quence, in this case, the terminal oxidizing agent used in the experi-
ment becomes ineffectual.
[18] A referee suggested that, in this case, an alternative rationalization
might be that PhI(OAc)2 is generating a different catalytic palladium
species than that generated under method A, and this species may pro-
mote aminopalladation rather than allylic CꢀH cleavage. We favor the
API-based rationalization for the following reasons: it mimics the same
scenario as observed in the transformation 1h!2 f, with the only dif-
ference being that the step that drives the reaction thermodynamically
is now the API oxidation/reductive elimination rather than API dehydro-
palladation; the PhI(OAc)2 reagent is compatible with CꢀH activation;
for example, see: a) L. T. Pilarski, N. Selander, D. Bçse, K. J. Szabꢅ, Org.
Lett. 2009, 11, 5518–5521; and b) ref. [8z]; a control experiment verified
that PhI(OAc)2 does not affect (oxidize) the disulfoxide ligand; the reac-
tion conditions needed to trigger a [3,3]-sigmatropic rearrangement
versus an aminochlorination in N-trichloroacetyl allylic carbamates (see
discussion of Scheme 10) are virtually identical, except for the presence
of the copper oxidant; the allylic CꢀH activation step is slow compared
to the oxidation of an alkylpalladium(II) complex by PhI(OAc)2 and its
subsequent reductive elimination. So, the likelihood that aminoacetoxy-
lation overwhelms CꢀH allylic oxidation, owing the massive difference
in the rates of the two potentially competing reactions, seems very
plausible; the successful transformation, 4!5, further strengthens this
hypothesis.
[19] This result suggests that the 7-membered aminopalladated intermedi-
ates are not generated, possibly owing to unfavorable ring size, and
that direct interaction between the substrate and the strong oxidant,
PhI(OAc)2 only gives degradation products. In general, the experiments
in the presence of PhI(OAc)2 were consistently less clean and had
poorer yields than those run in the presence of PhBQ as the oxidizing
agent.
[20] This Mizoroki–Heck coupling reaction variant has precedents: a) A. Ro-
driguez, W. J. Moran, Eur. J. Org. Chem. 2009, 1313–1316; b) X. Qu, P.
Sun, T. Li, J. Mao, Adv. Synth. Catal. 2011, 353, 1061–1066; c) N. M. Evdo-
kimov, A. Kornienko, I. V. Magedov, Tetrahedron Lett. 2011, 52, 4327–
4329.
[21] We thank Dr. Kevin Cariou (ICSN, Gif-sur-Yvette, France) for having
drawn our attention to this originally unexpected reactivity.
[22] For a definition of pseudo- and pure domino sequences, see: a) G. Poli,
G. Giambastiani, J. Org. Chem. 2002, 67, 9456–9459; b) G. Prestat, G.
Poli, Chemtracts: Org. Chem. 2004, 17, 97–103.
[23] R. C. Larock, T. R. Hightower, L. A. Hasvold, K. P. Peterson, J. Org. Chem.
1996, 61, 3584–3585.
[24] The alternative involvement of allylic CꢀH activation, though not com-
pletely ruled out, seems much less likely because a high-energy anti h3-
allyl complex would have to be involved.
[25] The reductive elimination step AcO[Pd]HQ[Pd0]+AcOH in the Mizoro-
ki–Heck reaction is known to be promoted by a base; for example,
see: C. Amatore, A. Jutand, G. Meyer, I. Carelli, I. Chiarotto, Eur. J. Inorg.
Chem. 2000, 1855–1859.
[26] In light of this study, carbonylative trapping of cyclic APIs, as developed
by Tamaru and co-workers, can be considered as another mode of
cyclic-API activation: H. Harayama, A. Abe, T. Sakado, M. Kimura, K.
Fugami, S. Tanaka, Y. Tamaru, J. Org. Chem. 1997, 62, 2113.
[9] F. Nahra, F. Liron, G. Prestat, C. Mealli, A. Messaoudi, G. Poli, Chem. Eur.
J. 2009, 15, 11078–11082.
[10] Crotyl carbamate 1d, and other carbamates featuring ꢁ 5 C atoms be-
tween N and the unsaturation did not cyclize under Method A.
[11] For precedents of reversible aminopalladation, see ref. [6] and: a) C.
Hahn, P. Morvillo, A. Vitagliano, Eur. J. Inorg. Chem. 2001, 419–429; b) C.
Hahn, P. Morvillo, E. Herdtweck, A. Vitagliano, Organometallics 2002, 21,
1807–1818; c) M. E. Cucciolito, A. D’Amora, A. Vitagliano, Organometal-
lics 2005, 24, 3359–3361; d) B. M. Cochran, F. E. Michael, J. Am. Chem.
Soc. 2008, 130, 2786–2792; e) V. I. Timokhin, S. S. Stahl, J. Am. Chem.
Soc. 2005, 127, 17888–17893; f) G. L. J. Bar, G. C. Lloyd-Jones, K. I.
Booker-Milburn, J. Am. Chem. Soc. 2005, 127, 7308–7309.
[12] These cyclic aminopalladated intermediates (APIs) are expected to be
initially and reversibly generated as protonated species in acetic acid.
[13] At present, we ignore the exact origin of such dehydropalladation in-
hibition of the resulting s-alkylpalladium complexes. For a previous
proposal of impeded b-H elimination from a cyclic API, see: a) A. Balꢆzs,
A. Hetꢀnyi, Z. Szakonyi, R. Sillanpꢃꢃ, F. Fꢇlçp, Chem. Eur. J. 2009, 15,
7376–7381. Experimental as well as theoretical studies are in progress
in order to investigate this intriguing point. Literature precedents (see
ref. [23]) as well as experiments in our laboratory indicate that the cor-
responding unsaturated N-tosyl amines (TsNHCH2R) do not experience
such a difficulty.
[14] To straightforwardly distinguish between a b-H elimination involving
a hydrogen atom of a cyclic structure and an alternative elimination in-
volving a hydrogen atom in a linear fragment, we propose the adjec-
tives ‘distocyclic’ and ‘proxicyclic’, respectively.
[15] As to the role of the disulfoxide ligand, we can anticipate that its pres-
ence will favor the allylic CꢀH activation step, when this route is at
work. On the other hand, we do not have information about its role, if
any, when the alternative reactivities become dominant.
[16] a) A. Lei, X. Lu, Org. Lett. 2000, 2, 2357–2360; b) J. Dauvergne, A. M.
Happe, V. Jadhav, D. Justice, M.-C. Matos, P. J. McCormack, M. R. Pitts,
S. M. Roberts, S. K. Singh, T. J. Snape, J. Whittall, Tetrahedron 2004, 60,
2559–2567; c) M. Ciclosi, C. Fava, R. Galeazzi, M. Orena, J. Sepulveda-
Arques, Tetrahedron Lett. 2002, 43, 2199–2202; d) S. R. Mellegaard-
Waetzig, D. K. Rayabarapu, J. A. Tunge, Synlett 2005, 2759; e) A. Joosten,
A. K. ꢄ. Persson, R. Millet, M. T. Johnson, J.-E. Bꢃckvall, Chem. Eur. J.
2012, 18, 15151–15157; f) A. M. Martꢈn Castro, Chem. Rev. 2004, 104,
2939–3002.
[27] S. D. R. Christie, A. D. Warrington, C. J. Lunniss, Synthesis 2009, 148–
154.
[28] The present study did not address the stereochemistry of the steps in-
volved in these transformations. For recent papers addressing in part
this intriguing topic, see: a) ref [18e]; b) C. Martꢈnez, Y. Wu, A. B. Wein-
stein, S. S. Stahl, G. Liu, K. Muniz, J. Org. Chem. 2013, 78, 6309–6315
and references therein.
Received: July 14, 2013
Revised: October 9, 2013
Published online on January 8, 2014
Chem. Eur. J. 2014, 20, 1539 – 1546
www.chemeurj.org
1546
ꢁ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim