Stanna-Brook rearrangement of carboxylic acid derivatives. Synthetic utility and mechanistic studies

Article abstract of DOI:10.1021/ol049826m

(Equation presented) The reaction of R₃SnLi with carboxylic acid derivatives proceeds through a novel, very fast stanna-Brook rearrangement that generates α-alkoxyorganolithium compounds as intermediates. The outcome of these reactions depends on the nature of the carboxyl derivatives. Reaction of R₃SnLi with ester derivatives gives rise to coupled products through a novel C-C bond formation reaction. Experimental evidence of the detailed reaction mechanism is provided.

Full text of DOI:10.1021/ol049826m

ORGANIC
LETTERS
2004
Vol. 6, No. 6
1061-1063
Stanna-Brook Rearrangement of
Carboxylic Acid Derivatives. Synthetic
Utility and Mechanistic Studies
M. Rita Paleo, M. Isabel Calaza, Paula Gran˜a, and F. Javier Sardina*
Departamento de Qu´ımica Orga´nica, Facultad de Qu´ımica, UniVersidad de Santiago
de Compostela, 15782 Santiago de Compostela, Spain
qojskd@usc.es
Received January 29, 2004
ABSTRACT
The reaction of R SnLi with carboxylic acid derivatives proceeds through a novel, very fast stanna-Brook rearrangement that generates
3
r-alkoxyorganolithium compounds as intermediates. The outcome of these reactions depends on the nature of the carboxyl derivatives.
Reaction of R SnLi with ester derivatives gives rise to coupled products through a novel C−C bond formation reaction. Experimental evidence
3
of the detailed reaction mechanism is provided.
The reaction of organotin anions with carbonyl derivatives
has found an increasing number of applications in organic
synthesis.1 Stable R-alkoxy- and R-aminostannanes have been
prepared by addition of R3SnLi to aldehydes or ketones
followed by protection or displacement of the R-hydroxy-
stannane intermediate.1 These R-heterosubstituted stannanes
are convenient precursors to R-oxy- and R-aminoorgano-
lithium compounds, which are valuable nucleophiles.2 Sur-
prisingly, only a few studies on the reactions of R3SnLi with
carboxylic acid derivatives have appeared, despite the fact
that these additions could provide expedient access to
acylstannanes,3 important reagents for C-C bond formation.4
The reaction of R3SnLi with acyl chlorides5 is usually very
low-yielding and limited to the preparation of the acyl-
triphenyltin derivatives.6 Moderate yields of acylstannanes7
have also been obtained by reaction of trialkyltinlithium
reagents with esters,8 in the presence of excess BF3‚Et2O or
with thioesters.9 We wish to report that carboxylic acid
derivatives react with Bu3SnLi to form a new C-C bond
through a novel stanna-Brook rearrangement.
We were interested in preparing enantiomerically pure
R-amino-acylstannanes by an extension of our previously
reported method for the synthesis of R-amino ketones,10
which would involve the addition of Bu3SnLi to R-amino
acid-derived oxazolidinones 1.
Treatment of N-phenylfluorenyl-alanine-oxazolidinone (1a)
with Bu3SnLi (from Bu3SnH and LDA) in THF at -55 °C,
led to a product that did not show any Bu3Sn signals in its
1H NMR spectrum.11 Analysis of its spectroscopic data
suggested a dimeric structure 2a for the reaction product,
which was confirmed by X-ray crystallographic analysis. No
(6) Acyltins can be alternatively prepared by a Pd-catalyzed reaction from
acyl chlorides and hexaalkylditin compounds: (a) Mitchell, T. N.; Kwetkat,
K. Synthesis 1990, 1001. (b) Mitchell, T. N.; Kwetkat, K. J. Organomet.
Chem. 1992, 439, 127.
(1) (a) Davies, A. G. Organotin Chemistry; VCH: Weinheim, 1997. (b)
Sato, T. Synthesis 1990, 259.
(2) (a) Still, W. C. J. Am. Chem. Soc. 1977, 99, 4836. (b) Still, W. C. J.
Am. Chem. Soc. 1978, 100, 1481.
(3) Still, W. C.; Sreekumar, C. J. Am. Chem. Soc. 1980, 102, 1201.
(4) Wyatt, P. B. Sci. Synth. 2003, 5, 423.
(7) Peddle, G. J. D. J. Organomet. Chem. 1968, 14, 139.
(8) Acylstannanes can also be prepared by reaction of excess aldehyde
with R3SnMgCl or with R3SnLi and an oxidant: (a) Quintard, J.-P.;
Elissondo, B.; Mouko-Mpegna, D. J. Organomet. Chem. 1983, 251, 175.
(b) Marshall, J. A.; Welmaker, G. S.; Gung, B. W. J. Am. Chem. Soc. 1991,
113, 647.
(5) (a) Shirakawa, E.; Yamamoto, Y.; Nakao, Y.; Tsuchimoto, T.;
Hiyama, T. Chem. Commun. 2001, 1926 and references therein. (b)
Shirakawa, E.; Nakao, Y.; Yoshida, H.; Hiyama, T. J. Am. Chem. Soc. 2000,
122, 9030.
(9) Capperucci, A.; Degl’Innocenti, A.; Faggi, C.; Reginato, G.; Ricci,
A.; Dembech, P.; Seconi, G. J. Org. Chem. 1989, 54, 2966.
(10) Paleo, M. R.; Calaza, M. I.; Sardina, F. J. J. Org. Chem. 1997, 62,
6862.
10.1021/ol049826m CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/21/2004

other diastereomer was observed in the crude reaction
mixture. With the structure of the product firmly established,
we studied the generality of this surprising C-C bond-
forming process. After some experimentation we found that
when Boc-protected oxazolidinones 1b-e were reacted with
Bu3SnLi, dimers 2 were always obtained as the main products
(Scheme 1), along with small amounts of alcohols 3 (6-8%
proceed via rearrangement of the initially formed stannyl
adduct 8 to give R-oxy-carbanion 9 (Scheme 2). To our
Scheme 2. Proposed Mechanism
Scheme 1. Reactions of Carboxylic Acid Derivatives with
Bu3SnLi
knowledge, this is the first example of a rearrangement where
the stannyl group migrates from carbon to oxygen13 in what
constitutes a tin analogue of the Brook rearrangement.14
The reaction takes a different course when silyl-lithium
reagents are used as nucleophiles. Thus, reaction of oxazo-
lidinones 1a,b with PhMe2SiLi in THF led to the corre-
sponding acylsilanes in 69 and 72% yield, respectively. No
dimers were detected in the crude reaction mixtures.
Valuable mechanistic information was derived from a
series of experiments on amide substrates. Reaction of Bu3-
SnLi with N-methoxyamides 10a-c gave primary amides
13a-c in good to excellent yields,15 with loss of the
N-methoxy group (Scheme 3, entries 1-3). This supports
yield).12 The reactions were very stereoselective, and mainly,
or exclusively, one diastereoisomer of dimer 2 was obtained.
N-Boc-Alanine derivative 1b led to the corresponding
product as an 8:1 ratio of diastereoisomers. N-Boc-Valine-
derived oxazolidinone (1c) led to the C2-symmetric diaste-
reoisomer 2c in 72% yield, whose structure was unambigu-
ously established by X-ray crystallography. The dimer 2c
was obtained in an improved 82% yield when Bu3SnLi
prepared from hexabutylditin and nBuLi was reacted with
1c. Phenylalanine derivative 1d led to an inseparable mixture
of diastereoisomers, which was equilibrated at 330 K in
CDCl3, to provide only the symmetrical stereoisomer 2d.
We checked that an equilibration was indeed taking place
by warming a CDCl3 solution of the pure unsymmetrical
isomer (obtained by precipitation of the mixture from CH2-
Cl2-hexane) at 50 °C for 24 h, a procedure that yielded the
symmetric isomer 2d quantitatively. Reaction of 1e gave a
mixture of two methionine-derived dimers (63% yield) that
was quantitatively transformed into pure 2e by warming at
50 °C for 6 h. It thus appears that the dimerization is very
stereoselective, and even in the less selective cases the
thermodynamically favored isomer can be obtained as the
sole product after equilibration of the crude reaction mixture.
Other types of esters underwent this C-C bond-forming
reaction as well, both in inter- and intramolecular fashion
(Scheme 1). We propose that these coupling reactions
Scheme 3. Reaction of Bu3SnLi with Amides
the intermediacy of carbanion 12, the product of the Sn-
Brook rearrangement, in the reaction, since such a species
should undergo fast elimination of the â-methoxy group to
give imidate 15, which should yield the observed products
upon aqueous workup. In fact, the presence of a benzimidate
as an intermediate in this reaction was substantiated by
1062
Org. Lett., Vol. 6, No. 6, 2004

analysis of the 119Sn and 13C NMR spectra of the reaction
mixture of 10a and Me3SnLi in deuterated THF before
quenching (13C NMR of the imidate carbon, δ 168.0 ppm;
119Sn NMR of the Sn-O moiety, δ 103.2 ppm).16
We did not detect any ring-opened products from the
reaction of cyclopropylamide 10c with Bu3SnLi, a result that
strongly argues against the intermediacy of free radicals in
this transformation, since cyclopropyl-substituted methyl
radicals readily experience ring opening.17
was slowly added as a quench, proving that the protonation
of the carbanionic intermediate was happening during the
quench and that all the steps leading to anion 12 are equilibria
faster than the protonation of the stannyl-lithium nucleophile.
When the reaction was conducted with the aliphatic amide
11b, the corresponding aldehyde was formed in good yield
without the need to resort to a slow aqueous quench, since
iPr2NH, present in the reaction medium as a byproduct from
Bu3SnLi generation, is acidic enough to protonate the
intermediate anion 12 (R ) Ph(CH2)2-, a more basic anion
than its benzylic counterpart). Thus, treatment of 11b with
Bu3SnLi (320 mol %) at 0 °C to room temperature for 48 h
provided aldehyde 14b in 54% yield.19 When the reaction
This N-demethoxylation process should produce methox-
ide anion as a byproduct. That this is in fact the case was
shown by quenching the reaction of 10a and Bu3SnLi with
benzoyl chloride: 1H NMR analysis of the crude reaction
mixture showed the formation of an equimolar amount of
methyl benzoate. This result rules out an alternative mech-
anism for this reaction, namely, the abstraction of one of
the N-methoxy hydrogens by a base, followed by expulsion
of a benzimidate anion, since formaldehyde should be formed
in this process instead of methoxide.18
A very different outcome, but also supportive of the
stanna-Brook rearrangement mechanism, was obtained from
the reactions of N,N-dialkylamides. Treatment of N,N-
dimethylbenzamide (11a) with Bu3SnLi at 0 °C for 6 h led
mainly to recovered amide when carefully dried amide 11a
was used; however, when moisture was not removed from
the starting material, we observed the presence of ap-
proximately 5% benzaldehyde in the crude reaction mixture,
presumably due to protonation of the small amount of anion
12 (R ) phenyl, R1 ) Me) present at equilibrium. When
the reaction was repeated with thoroughly dried 11a, but now
slowly adding a solution of THF-H2O for 3 h as a quench,
benzaldehyde was isolated in 30% yield.19 Deuterated
benzaldehyde (14a, R2 ) D) was obtained when THF-D2O
1
was run in deuterated THF and followed by H and 13C
NMR, the formation of aminal 16, derived from protonation
of intermediate anion 12 (R ) phenethyl, R1 ) Me), was
assessed by the presence of a triplet at δ 4.51 ppm in the 1H
NMR spectrum (which disappeared after aqueous quenching)
and a CH signal at δ 93.5 ppm in the 13C NMR spectrum.20
All these observations taken together point to a common
mechanism for the reaction of R3SnLi with carboxylic acid
derivatives: an initial nucleophilic addition to the carbonyl
carbon followed by a fast, reversible stanna-Brook rear-
rangement that generates an R-oxycarbanion whose fate
depends on the structure of the substrate and the reaction
conditions.
In summary, we report a new and very fast stanna-Brook
rearrangement of carboxylic acid derivatives upon reaction
with R3SnLi. This addition-rearrangement pathway seems
to be quite common since we have observed it in reactions
of several types of esters and amides with trialkyltinlithium
reagents. We are currently further exploring the synthetic
applications of these transformations, in particular for the
preparation of the core part of a series of HIV-protease21
inhibitors from amino-diols 2.
(11) Kosugi, M.; Naka, H.; Sano, H.; Migita, T. Bull. Chem. Soc. Jpn.
1987, 60, 3462.
(12) For characterization purposes, the reduced products 3 were inde-
pendently prepared by treatment of the corresponding oxazolidinones with
LiAlH(tBuO)3 in THF.
(13) (a) A different kind of anionic 1,2-stannyl rearrangement was
reported: Iwamoto, K.; Chatani, N.; Murai, S. J. Organomet. Chem. 1999,
574, 171. (b) We have suggested that a similar rearrangement may be taking
place in the reaction of R-aminoaldehydes and 3-oxo-azepines with Bu3-
SnLi, although we could not provide any experimental mechanistic
evidence: Calaza, M. I.; Paleo, M. R.; Sardina, F. J. J. Am. Chem. Soc.
2001, 123, 2095.
(14) (a) Brook, A. G. J. Am. Chem. Soc. 1958, 80, 1886. (b) Brook, A.
G. Acc. Chem. Res. 1974, 7, 77.
(15) Addition of PhMe2SiLi to amides 10a,b led to the corresponding
secondary amides 13a,b, indicating that Weinreb amides undergo Brook
rearrangement on reaction with Si anions.
(16) (a) A similar compound, 4,5-dihydro-4,4-dimethyl-2-phenyloxazole,
showed a signal at δ 162.2 ppm in 13C NMR: Ashburn, S. P.; Coates, R.
M. J. Org. Chem. 1985, 50, 3076. (b) See ref 1a, p 21.
(17) Castaing, M.; Pereyre, M.; Ratier, M.; Blum, P. M.; Davies, A. G.
J. Chem. Soc., Perkin Trans. 2 1979, 287.
(18) Graham, S. L.; Scholz, T. H. Tetrahedron Lett. 1990, 31, 6269.
(19) Material balance is composed of starting amide.
Acknowledgment. Financial support from the CICYT
(Grant BQU2002-01368 and a fellowship to P.G.) and the
Xunta de Galicia (Grant PGIDIT03PXIC20910PN and a
fellowship to M.I.C.) is gratefully acknowledged.
Supporting Information Available: Complete experi-
mental procedures, spectroscopic and analytical data, and
X-ray crystal structure data for 2a and 2c. This material is
available free of charge via the Internet at http://pubs.acs.org.
OL049826M
(20) (a) Myers, A. G.; Kung, D. W.; Zhong, B.; Movassaghi, M.; Kwon,
S. J. Am. Chem. Soc. 1999, 121, 8401. (b) Cattoe¨n, X.; Sole´, S.; Pradel, C.;
Gornitzka, H.; Miqueu, K.; Bourissou, D.; Bertrand, G. J. Org. Chem. 2003,
68, 911.
(21) Babine, R. E.; Bender, S. L. Chem. ReV. 1997, 97, 1359.
Org. Lett., Vol. 6, No. 6, 2004
1063