Palladium-catalyzed [3,3] sigmatropic rearrangement of (allyloxy) iminodiazaphospholidines: Allylic transposition of C-O and C-N functionality

Article abstract of DOI:10.1002/anie.200353284

A P^V=N to P^V=O interconversion is the thermodynamic driving force for the title reaction. Iminodiazaphospholidines 1 give phosphoramides 2, which are subsequently hydrolyzed under mild acidic conditions to yield primary or tosylamines 3. R = alkyl, phenyl; Ts = p-toluenesulfonyl.

Full text of DOI:10.1002/anie.200353284

Angewandte
Chemie
Sigmatropic Rearrangement
Palladium-Catalyzed [3,3] Sigmatropic
Rearrangement of
(Allyloxy)iminodiazaphospholidines: Allylic
À
À
Transposition of C O and C N Functionality**
Ernest E. Lee and Robert A. Batey*
The principle of driving reactions thermodynamically through
the conversion of PIII reagents into PV O products is well
=
established, as exemplified by the Wittig and Mitsunobu
reactions, and the [2,3] sigmatropic rearrangement of allyl
phosphites into allyl phosphonates.[1] Herein we describe a
novel [3,3] sigmatropic rearrangement in which allylic trans-
V
position is driven by a PV N to P O interconversion
=
=
(Scheme 1). We envisaged a process whereby conversion of
Scheme 1. Proposed route to allylic amines based on the [3,3] sigma-
tropic rearrangement of phospholidines 3.
an allylic alcohol 1 into a phosphoramidite 2, followed by a
Staudinger reaction[2] would generate a phospholidine 3. A
[3,3] sigmatropic 3-aza-2-phospha-1-oxa-Cope[3] rearrange-
ment of 3 would then generate a phosphoramide 4, which on
deprotection would lead to the transposed allylic amine 5.[4]
The overall process is analogous to the aza variants of the
Cope [3,3] sigmatropic rearrangement,[5] the most important
example of which is the well-known Overman rearrangement
of allylic imidates into allylic amides.[6] The estimated
thermodynamic driving force for a phospholidine–phosphor-
amide interconversion,[7] such as would occur in a sigmatropic
rearrangement, is approximately 25 kcalmolÀ1.[8]
The feasibility of this approach was tested by using
(allyloxy)iminodiazaphospholidines 6 and 7 as substrates.
[*] E. E. Lee, Prof. R. A. Batey
Department of Chemistry, University of Toronto
80 St. George Street, Toronto, Ontario, M5S 3H6 (Canada)
Fax: (+1)416-978-5059
E-mail: rbatey@chem.utoronto.ca
[**] The Natural Science and Engineering Research Council (NSERC) of
Canada funded this research. E.E.L. thanks the Ontario Ministry of
Training for funding in the form of an Ontario Graduate Scholarship.
R.A.B. gratefully acknowledges receipt of a Premier's Research
Excellence Award. We also thank Dr. Alex Young for mass spectral
analysis and Dr. Tim Burrow for NMR assistance.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2004, 43, 1865 –1868
DOI: 10.1002/anie.200353284
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1865

Communications
Table 1: Preparation of (allyloxy)iminodiazaphospholidines.
These compounds were cleanly prepared in a one-pot process
by the sequential treatment of the corresponding allylic
alcohols with the phospholidine 8,[9] as described by Alexakis
et al.,[10] followed by tosyl azide and diphenylphosphoryl
azide (DPPA),[11] respectively (Table 1).[12] The reaction was
monitored by 31P NMR spectroscopy to ensure complete
conversion of the intermediate phosphoramidite (d ꢀ
130 ppm for 2, whereas d ꢀ 24 ppm for 6 and d ꢀ 24 and
À10 ppm for 7). The iminodiazaphospholidines were then
purified by chromatography on silica gel, with Et3N as an
additive to prevent acid-promoted decomposition.
R1OH
Yield [%][a]
R1OH
Yield [%][a]
6a
7a
92
87
6g
7g
94
91
6b
7b
93
90
6h
7h
92
84
Initial attempts at the thermal rearrangement of com-
pounds 6 were unsatisfactory and led to products arising from
pathways of both the desired [3,3] and formal [1,3] sigma-
tropic rearrangement. As PdII and HgII catalysts are known to
catalyze the [3,3] sigmatropic rearrangement of allylic imi-
dates,[6] a variety of PdII catalysts were screened for the
rearrangement of 6a into 9a, but only [PdCl2(MeCN)2] was
found to be an active catalyst.[13] In the presence of
[PdCl2(MeCN)2] (5 mol%), the rearrangement of both 6
and 7 proceeded smoothly at room temperature to yield only
the products of [3,3] rearrangement 9 and 10, respectively
(Table 2). In the reactions of the DPPA-derived substrates 7,
the addition of 4-Š molecular sieves was required to ensure
complete conversion into 10. The rearrangements were
[b,c]
6c
7c
95
92
6i
7i
–
[b,c]
–
6d
7d
91
89
6j
7j
92
87
6e
7e
89
89
6k
7k
86
78
6 f
7 f
93
86
6l
7l
87
80
[a] Yield of isolated product, 0.6-mmol scale. [b] These compounds could
not be purified by column chromatography on silica gel and were used
crude in subsequent transformations. [c] Reaction conducted in [D6]ben-
zene. Ts=p-toluenesulfonyl.
conveniently
monitored
spectroscopy
by
(d ꢀ
31P NMR
Table 2: Pd-catalyzed [3,3] sigmatropic rearrangement of (allyloxy)iminodiazaphospholidines and
subsequent hydrolysis.
20 ppm for 9, and d ꢀ 20 and
À4 ppm for 10). The phosphora-
mides 9 and 10 were cleaved under
acidic conditions to yield the allylic
tosylamines 11 and free allylic
amines 12, respectively.[14] As
mild, acidic conditions are used
for the final hydrolysis, this overall
process complements the Overman
rearrangement of allylic imidates.
Strongly basic conditions (3–5m
NaOH) are employed for the
hydrolysis of the intermediate tri-
chloroacetamides in the Overman
protocol.
A variety of substitution pat-
terns are tolerated on the allylic
substrates 6 and 7, including sub-
stitution in the allylic group a, b,
and g to the oxygen atom. Notably,
the reaction worked well for sub-
strates substituted at the b position
(6c and 7c), as previous attempts
at metal-catalyzed rearrangements
of the corresponding allylic imi-
dates have had mixed success.[15]
Substrates 6 f and 7 f both under-
went rearrangement in good yield
to afford only the E isomers 9 f and
10 f. The reaction of the substrates
6d and 7d, derived from a simple
secondary allylic alcohol, to give
R1
R3
Yield [%][a]
Yield [%][a]
6a
7a
6b
7b
6c
7c
6d
7d
6e
7e
6 f
7 f
9a
95
90
93
89
95
91
91
86
88
trace
90
93
11a
88
81
88
81
97
87
93
85
90
–
10a[b]
9a
12a[f]
11a
10a[b]
9c
12a[f]
11c
10c[b]
9d
12c[f]
11d
10d[b]
9e[c,d]
10e[b]
9 f[c]
12d[f]
11e
12e[f]
11 f
83
79
10 f[b]
12 f[f]
6g
7g
6h
7h
9g[c,d]
10g[b]
9h[c,d]
10h[b]
75
11g
85
–
n.r.[e]
76
12g[f]
11h
80
–
n.r.[e]
12h[f]
6i
7i
9i[c]
80
11i
78
–
10i[b]
n.r.[e]
12i[f]
6j
7j
6k
7k
9j
n.r.[e]
n.r.[e]
90
–
–
–
–
82
78
10j[b]
9k
11k
12k[f]
10k[b]
84
6l
7l
9l
88
83
11 l
90
79
10l[b]
12 l[f]
[a] Yield of isolated product, 0.6-mmol scale. [b] Reaction conducted in the presence of 4-Š molecular
sieves. [c] Reaction conducted in toluene. [d] Reaction conducted at 458C. [e] Only starting material was
observed by 31P NMR spectroscopy. [f] Reaction conducted with HCl (1m) in MeOH.
1866
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 1865 –1868

Angewandte
Chemie
the corresponding primary allylic phosphoramides occurred
in excellent yields at room temperature. However, reactions
of the 2-cyclohexenyl substrates were far more sluggish, with
6e requiring heating at 458C and 7e yielding only trace
amounts of products after 48 h at 808C. A similar trend was
observed with substrates 6g–i and 7g–i. It is apparent that
substrates with sterically demanding substituents react more
slowly in the case of tosyl-derived, and are unreactive in the
case of DPPA-derived substrates. The steric limitations of this
reaction are further emphasized by the lack of reactivity of
the substrates 6j and 7j, which are derived from a g,g-
disubstituted allylic alcohol.
In conclusion, a novel palladium(ii)-catalyzed rearrange-
ment of (allyloxy)iminodiazaphospholidines has been devel-
oped for the synthesis of allylic amines and tosylamines.
Investigations into diastereo- and enantioselective variants
are currently underway in our laboratory.
Received: November 10, 2003 [Z53284]
Keywords: allylic amines · azides · homogeneous catalysis ·
palladium · sigmatropic rearrangement
.
The transposition of the enantioenriched E substrates 6k
and 7k produced only the E phosphoramides 9k and 10k with
clean transfer of chirality. The Z substrates 6l and 7l under-
went rearrangement to the E products 9l and 10l, albeit with
diminished enantiomeric excess.[16] The [3,3] sigmatropic
rearrangement presumably proceeds through intramolecular
attack on the palladium-coordinated double bond by the lone
[1] T. Janecki, R. Bodalski, Synthesis 1990, 799 – 801, and references
therein.
[2] Y. G. Gololobov, L. F. Kasukhin, Tetrahedron 1992, 48, 1353 –
1406.
[3] Classification based on the allylic imidate rearrangement
described in: F. Vögtle, E. Goldschmitt, Chem. Ber. 1976, 109,
1 – 40.
[4] Allylic amines are important synthetic intermediates as well as
targets. For a review of their synthesis, see: M. Johannsen, K. A.
Jørgensen, Chem. Rev. 1998, 98, 1689 – 1708.
pair of electrons on the nitrogen atom of PV N, followed by
=
rearrangement of the resulting phosphonium intermediate.
For example, in the case of 6k the reaction proceeds via the
p complex 13 and phosphonium ion 14 in a fashion analogous
to that proposed for the rearrangement of allylic imidates
(Scheme 2).[6a,d] The absolute configuration[17] and olefin
geometry of the products in both cases are consistent with
this mechanism.
[5] a) K. Ritter in Houben-Weyl. Stereoselective Synthesis, Vol. E 21e
(Eds.: G. Helmchen, R. W. Hoffmann, J. Mulzer, E. Schau-
mann), Thieme, Stuttgart, 1996, pp. 5677 – 5699; b) R. P. Lutz,
Chem. Rev. 1984, 84, 206 – 247.
[6] a) L. E. Overman, J. Am. Chem. Soc. 1976, 98, 2901 – 2910;
b) L. E. Overman, Acc. Chem. Res. 1980, 13, 218 – 224; c) L. E.
Overman, Angew. Chem. 1984, 96, 565 – 573; Angew. Chem. Int.
Ed. Engl. 1984, 23, 579 – 586; d) T. G. Schenck, B. Bosnich, J.
Am. Chem. Soc. 1985, 107, 2058 – 2066; e) M. Calter, T. K. Hollis,
L. E. Overman, J. Ziller, G. G. Zipp, J. Org. Chem. 1997, 62,
1449 – 1456; f) T. Nishikawa, M. Asai, N. Ohyabu, M. Isobe, J.
Org. Chem. 1998, 63, 188 – 192; g) Y. Uozumi, K. Kazuhiko, T.
Hayashi, Tetrahedron: Asymmetry 1998, 9, 1065 – 1072; h) I.
Savage, E. J. Thomas, P. D. Wilson, J. Chem. Soc. Perkin Trans. 1
1999, 3291 – 3303; i) T. Donde, L. E. Overman, J. Am. Chem.
Soc. 1999, 121, 2933 – 2934; j) L. E. Overman, C. E. Owen, M. M.
Pavan, C. J. Richards, Org. Lett. 2003, 5, 1809 – 1812; k) C. E.
Anderson, L. E. Overman, J. Am. Chem. Soc. 2003, 125, 12412–
12413.
Scheme 2. Proposed mechanism for the Pd-catalyzed reaction, as
exemplified by the conversion of 6k into 9k.
Comparison of the results of the PdII-catalyzed [3,3]
sigmatropic rearrangement at ambient temperatures with the
thermal rearrangement of substrates 6 clearly demonstrates
the advantages of metal catalysis to facilitate clean rearrange-
ments. For example, the thermal rearrangement of the
diazaphospholidine 6a at 1308C led to the [3,3] product 9a
and [1,3] product 15a in a ratio of 3.5:1 (Scheme 3).
Furthermore, the thermal rearrangement of 6g only yielded
the [1,3] product 15g, whereas that of 6m only yielded the
[3,3] product 9m. In the last two examples, only the
thermodynamically more stable allylic phosphoramide was
formed. These results suggest that ionization and subsequent
recombination is competitive with the [3,3] sigmatropic
rearrangement under thermal conditions.
[7] For the use of phospholidine–phosphoramide interconversion as
a thermodynamic driving force in other rearrangements, see:
a) T. A. Mastryukova, N. V. Mashchenko, I. L. Odinets, P. V.
Petrovskii, M. I. Kabachnik, Russ. J. Gen. Chem. 1988, 58, 1756 –
1761; b) E. J. Cabrita, C. A. M. Afonso, A. Gil de Oliveira
Santos, Chem. Eur. J. 2001, 7, 1455 – 1467.
[8] The thermodynamic driving force for the [3,3] sigmatropic
rearrangement of 3 into 4 was estimated by comparison with
=
the analogous conversion of (NH2)2(MeO)P NH into
=
(NH2)2(MeNH)P O, a formal [1,3] sigmatropic rearrangement
which involves the same overall bonding reorganization. Geom-
etry optimizations, single-point energies, and vibrational analysis
were calculated at the B3LYP/6-311G* level. For comparison,
=
=
the driving force of a C NH to C O transposition can be
estimated by the energy difference between the imidate
=
=
Me(MeO)C NH and the amide Me(MeNH)C O, the latter
calculated to be 18.6 kcalmolÀ1 lower in energy at the B3LYP/6-
311G*level. (Calculations were performed on a Dual 2-GHz
Power PC G5 by using Spartan'02, Version 1.0.4e, Wavefunction
Inc., Irvine, CA).
[9] The phospholidine 8 was prepared as described in: S. Hanessian,
Y. L. Bennani, Y. Leblanc, Heterocycles 1993, 35, 1411 – 1424.
[10] A. Alexakis, S. Mutti, P. Mangeney, J. Org. Chem. 1992, 57,
1224 – 1237.
Scheme 3. Thermal rearrangements of phospholidines 6.
Angew. Chem. Int. Ed. 2004, 43, 1865 –1868
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1867

Communications
[11] Although we have experienced no problems with either of these
azides, appropriate safety measures should be taken.
[12] a) J. Bellan, M. Sanchez, M. R. Marre-Mazi›res, A. M. Beltran,
Bull. Soc. Chim. Fr. 1985, 3, 491 – 495; b) M. R. Marre, M.
Sanchez, J. F. Brazier, R. Wolf, J. Bellan, Can. J. Chem. 1982, 60,
456 – 468.
[13] The use of the following catalysts resulted in complete recovery
of 6a: PdCl2, [PdCl2(PPh3)2], [PdCl2(PCHX3)2], [Pd2Cl2(allyl)2],
[PdCl2(cod)]. cod = 1,5-cyclooctadienone.
[14] V. Mizrahi, T. A. Modro, J. Org. Chem. 1983, 48, 3030 – 3037.
[15] Depending on the allylic imidate used, either prolonged reaction
time was required or no reaction was observed: P. Metz, C. Mues,
A. Schoop, Tetrahedron 1992, 48, 1071 – 1080, and references
therein.
[16] Compounds 6k and 7k (95% ee) underwent rearrangement to
give 9k and 10k with 91% ee, whereas the rearrangement of 6l
and 7l (95% ee) gave 9l and 10l with 70% ee (determined by
HPLC on a chiral phase).
[17] The optical rotation of the methyl ester prepared by the
ozonolysis of 11k was compared to that of a previously reported
authentic sample; see Supporting Information for details.
1868
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 1865 –1868