Some mechanistic observations on the borohydride mediated reductive cyclisation of tosylhydrazones

Article abstract of DOI:10.1039/b101053k

The previously described, highly stereoselective ring-closure of δ-unsaturated tosylhydrazones upon reduction with borohydride may not be a radical process but rather an enetype concerted transformation of the intermediate mono-substituted diimide.

Full text of DOI:10.1039/b101053k

Some mechanistic observations on the borohydride mediated reductive
cyclisation of tosylhydrazones
Luis D. Mirandaa and Samir Z. Zard*ab
a Institut de Chimie des Substances Naturelles, CNRS, 91198 Gif-Sur-Yvette, France.
E-mail: zard@icns.cnrs-gif.fr
b Laboratoire de Synthèse Organique associé au CNRS Ecole Polytechnique, 91128 Palaiseau, France
Received (in Cambridge, UK) 30th January 2001, Accepted 27th March 2001
First published as an Advance Article on the web 22nd May 2001
The previously described, highly stereoselective ring-closure
of d-unsaturated tosylhydrazones upon reduction with
borohydride may not be a radical process but rather an ene-
type concerted transformation of the intermediate mono-
substituted diimide.
abstraction from the stannane by the primary radical 5. In the
present study, we adduce evidence indicating that the radical
cyclisation is indeed irreversible under the stannane reduction
conditions and that the cyclopentane formation in the case of the
hydrazone precursor may in fact not be a radical process at
all.
Reduction with borohydride of tosylhydrazones derived from
ketones may lead to the formation of radical species through
loss of molecular nitrogen as shown in Scheme 1.1 Various
other related and important transformations such as the
reduction of diazonium salts,2 the Wolff–Kishner reaction,3 the
Wharton rearrangement,4 sometimes proceed through a similar
radical pathway.
Our approach hinges on the separate generation of the
cyclised radical 5 and showing that this species does not
undergo ring opening under dilute tributylstannane reduction
conditions. This was accomplished by exploiting the properties
of the xanthate transfer reaction.7 In this case, the cyclised
product is itself a xanthate and therefore allows the regeneration
of the cyclised radical. The requisite starting xanthate 6c was
readily obtained in 66% overall yield by reacting crude bromide
6b, itself derived from alcohol 6a, with commercially available
potassium O-ethyl xanthate. Heating this xanthate with a small
amount of lauroyl peroxide in 1,2-dichloroethane (0.1 M)
resulted in the smooth formation of cyclopentylmethyl xan-
thates 7a and 7b in a 3+7 ratio and in 78% combined yield
(Scheme 3). It is interesting to note that the ratio of the cis and
trans isomers did not change when the reaction was conducted
at 1 M concentration. This is a first indication that the
cyclisation is not so readily reversible under these conditions.
The two isomers could be separated using preparative thin-
layer chromatography, albeit with some difficulty. Reductive
dexanthylation of the pure trans isomer 7b using tributyl-
stannane in benzene at 0.001 M concentration gave the
corresponding pure trans methyl cyclopentanes 3b in 70%
yield. Perhaps more importantly, exposure of an almost pure
sample (95+5 cis+trans) of the thermodynamically less stable
cis isomer 7a to the same conditions gave methylcyclopentanes
3a and 3b in 85+15 ratio, i.e. with only a very slight modification
of the initial relative stereochemistry. Clearly, under such
highly dilute conditions, a roughly 3+7 mixture of the two
diastereoisomers should have been obtained in both cases if the
equilibrium between radicals 4 and 5 was fast in comparison to
hydrogen atom abstraction from the stannane.
Scheme 1
In a pioneering study in this area, Taber and colleagues5
described a remarkably diastereoselective cyclisation reaction
starting with hydrazone 1 (Scheme 2). This transformation was
presumed to involve ring closure of a benzyl type radical 4, and
the unusually high diastereoselectivity (97+3) was ascribed to
the non-reversibility of the cyclisation step under these specific
conditions. The corresponding tributylstannane mediated ring-
closure starting from the imidazole thiocarbamate derivative 2
is much less stereoselective, leading to a 3+7 ratio of the same
diastereoisomers 3, and the possible reversibility of the ring-
forming step was invoked as the cause of the erosion in
selectivity. While reversible 5-exo cyclisations involving
benzylic and other stabilised carbon centred radicals are
known,6 it seemed to us that the ring-opening, reverse step, must
be too slow in this case to compete with hydrogen atom
The above findings constitute strong evidence against the
operation of a radical mechanism in the transformation of the
tosylhydrazone depicted in Scheme 2. The high diaster-
eoselectivity cannot be explained by a kinetically controlled,
Scheme 2
Scheme 3
1068
Chem. Commun., 2001, 1068–1069
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b101053k

conditions, and the eventual presence of initiators, monoalk-
yldimides (monoalkyldiazenes) may react by an ionic, a radical,
or a concerted pericyclic-type mechanism. Whatever the exact
mechanism operating in the present case, the observations of
Taber and co-workers3,5 will certainly have important conse-
quences for organic synthesis.
Notes and references
1 D. F. Taber, Y. Wang and S. J. Stachel, Tetrahedron Lett., 1993, 34,
6209; A. G. Myers, M. Movassaghi and B. Zheng, J. Am. Chem Soc.,
1997, 119, 8572; A. G. Myers, M. Movassaghi and B. Zheng,
Tetrahedron Lett., 1997, 38, 6569. For related reactions which may
involve radicals, see L. Caglioti, F. Gasparini, D. Misti and G. Palmieri,
Tetrahedron, 1978, 34, 135; F. S. Guziec and D. Wei, Tetrahedron Lett.,
1992, 33, 7465; A. J. Bloodworth and D. Korkodilos, Tetrahedron Lett.,
1991, 32, 6953.
Scheme 4
irreversible radical ring closure. The radical cyclisation in such
systems is inherently poorly diastereoselective: a similar 1+2.5
cis+trans ratio was recently reported by Studer,8 who generated
the analogous 1-phenylhex-5-enyl radical by heating the
TEMPO derived precursor at 130 °C. Cyclisations involving
non-stabilised secondary radicals also take place with a modest
preference for the trans isomer.9 The very high cis-selectivity
observed by Taber and his colleagues in the present case is
probably the result of an intramolecular, ene-type pericyclic
process, as shown in Scheme 4. Placing the aryl group in the less
hindered pseudo-equatorial position leads to the cis isomer.
Such a concerted mechanism parallels that proposed for
reductions with diimide itself,10 the rigid, highly organised
transition state being more consonant with the observed high cis
selectivity. Thus, depending on the structure, the experimental
2 C. Galli, Chem. Rev., 1988, 88, 765.
3 D. F. Taber and S. J. Stachel, Tetrahedron Lett., 1992, 33, 903; D. F.
Taber and J. M. Anthony, Tetrahedron Lett., 1980, 2779.
4 G. Stork and P. G. Williard, J. Am. Chem Soc., 1977, 99, 7067.
5 D. F. Taber, Y. Wang and T. F. Pahutski, Jr., J. Org. Chem., 2000, 65,
3861.
6 C. Walling and A. Cioffari, J. Am. Chem. Soc., 1972, 94, 6064. See also:
D. P. Curran and C.-T. Chang, J. Org. Chem., 1989, 54, 3140.
7 S. Z. Zard, Angew. Chem., Int. Ed. Engl., 1997, 36, 672.
8 A. Studer, Angew. Chem., Int. Ed. Engl., 2000, 39, 1108.
9 A. L. J. Beckwith and C. H. Schiesser, Tetrahedron, 1985, 41, 3925.
10 D. J. Pasto, in Comprehensive Organic Synthesis, B. M. Trost, I.
Fleming, eds., Pergamon Press, Oxford, 1991, vol. 8, pp. 471–476.
Chem. Commun., 2001, 1068–1069
1069