Sulfonamides as novel terminators of cationic cyclisations

Article abstract of DOI:10.1039/b207755h

Trifluoromethanesulfonic (triflic) acid is an excellent catalyst for inducing overall 5-endo cyclisation of homoallylic sulfonamides [e.g. 4] to give pyrrolidines [e.g. 5]. In competitive experiments, pyrrolidines or homopiperidines are formed in preference to piperidines, even when the latter would be obtained by trapping a tertiary carbocation. Cationic cascades terminated by a sulfonamide group are viable for the efficient formation of polycyclic systems.

Full text of DOI:10.1039/b207755h

Sulfonamides as novel terminators of cationic cyclisations
Charlotte M. Haskins and David W. Knight*
Chemistry Department, Cardiff University, P.O. Box 912, Cardiff, UK CF10 3TB.
E-mail: knightdw@cf.ac.uk
Received (in Cambridge, UK) 7th August 2002, Accepted 23rd September 2002
First published as an Advance Article on the web 23rd October 2002
Trifluoromethanesulfonic (triflic) acid is an excellent cata-
lyst for inducing overall 5-endo cyclisation of homoallylic
sulfonamides [e.g. 4] to give pyrrolidines [e.g. 5]. In
competitive experiments, pyrrolidines or homopiperidines
are formed in preference to piperidines, even when the latter
would be obtained by trapping a tertiary carbocation.
Cationic cascades terminated by a sulfonamide group are
viable for the efficient formation of polycyclic systems.
reflux to achieve complete reaction; the excellent yields of the
resulting pyrrolidines 9 were however retained. spiro-Pyrroli-
dines can also be prepared using this chemistry: the ylidenecy-
clohexane derivative 10 gave an excellent return of the expected
product 11 under the mildest set of conditions. Given that the
reaction pathway involves alkene protonation and carbocation
generation, it was not surprising to find that the isomeric
cyclohexene derivative 12 gave the related spiro-pyrrolidine 13
under similar conditions. This observation will provide an
additional degree of flexibility in terms of precursor synthesis.
Aldehyde functions, at least when not having an a-proton,
survive the acidic conditions, as shown by the very clean
conversion of the N-tosyl alaninal derivative 14 into the
corresponding pyrrolidine carboxaldehyde 15. However, suita-
bly positioned hydroxyl groups compete with the sulfonamide
as carbocation traps. Thus, the alaninol derivative 16 gave a ca.
1+1 mixture of the pyrrolidinemethanol 17 and the aminopyran
18, indicating a not unexpected limitation of this method-
ology.
Despite being disfavoured by Baldwin’s rules,1 5-endo-trig
cyclisations involving electron-deficient intermediates such as
iodonium ions2 have made a considerable contribution to
synthetic methodology; it is likely that these are not exceptions
to the rules, being electrophile rather than nucleophile-driven.
Our own studies have uncovered a highly stereocontrolled
synthesis of pyrrolidines using such cyclisations of (E)-
homoallylic sulfonamides 1 which, when exposed to iodine in
the presence of potassium carbonate, are converted rapidly into
the corresponding 2,5-trans-iodopyrrolidines 2 in excellent
yields, probably via a chair-like transition state (Scheme 1).3 In
contrast, omission of a base results in exclusive formation of the
corresponding 2,5-cis isomers 3, indicative of an isomerization
of the initially-formed kinetic isomers 2 to the thermodynamic
isomers 3. This idea was confirmed when examples of
2,5-trans-pyrrolidines 2 were exposed to iodine–hydrogen
iodide mixtures when conversion into the cis-isomers 3
occurred rapidly.3 Selenocyclisations of sulfonamides 1 show
similar features.4 This chemistry led us to speculate that it might
be possible to trigger such cyclisations using acid. Herein, we
report the successful outcome of our initial studies in this
area.
We then examined the stereochemical features of the
cyclisation, using the cinnamyl derivatives 6 (Table 2). In the
case of the glycinate 6a, our initial experiments (Table 1) gave
Table 1
The required precursors [4,6 etc.; Table 1] were readily
obtained in good overall yield using the Stork procedure5
followed by N-protecting group exchange. Examples of the
corresponding aldehydes 14 and alcohols 16 were then obtained
by reduction. Initial experiments using the prenyl derivative 4b
indicated the viability of the idea in that the desired product 5b
was observed upon exposure to a variety of proton sources.
Following some experimentation, we discovered that using
trifluoromethanesulfonic (triflic) acid in chloroform repre-
sented optimum conditions for inducing cyclisation. Some
definitive results are presented in Table 1. Both prenyl
derivatives 4 underwent rapid and very clean cyclisation when
exposed to 0.4 equivalents of triflic acid at 0 °C to give
essentially quantitative yields of the pyrrolidines 5. In contrast,
the cinnamyl derivatives 6 required slightly more demanding
conditions (0.6 eq. TfOH, 25 °C, 4 h) before similar yields of the
5-phenyl derivatives 7 were obtained. Understandably, cyclisa-
tion of the corresponding crotyl analogue 8 required heating to
Scheme 1
2724
CHEM. COMMUN., 2002, 2724–2725
This journal is © The Royal Society of Chemistry 2002

View Article Online
Table 2
x eq. TOH
y h
19a+20a
x eq. TfOH y h
19b+20b
0.6
1.0
2.0
5.0
4.5
2.0
1.5
1.0
1
1
> 20
> 20
1
1
1
1
0.6
1.5
2.0
5.0
4.0
2.0
1.5
1.0
2.9
3.5
4.0
9.0
1
1
1
1
a 1+1 mixture of the cis- and trans-pyrrolidine carboxylates 19a
and 20a.6 Increasing the amount of triflic acid to two
equivalents resulted in the almost exclusive formation of the
cis-isomer 19a. That this thermodynamically more stable
isomer is obtained predominantly under these conditions
suggests that a similar isomerization process can take place to
that observed during the iodocyclisations (Scheme 1).3,4
Similarly, treatment of the alanine-derived sulfonamide 6b with
excess triflic acid also led to a synthetically useful preponder-
ance of the ‘cis’-isomer 19b although to obtain this, five rather
than two equivalents of acid were required.
When we attempted to extend the methodology to syntheses
of piperidine derivatives [e.g. 23], we found that formation of
such six-membered rings is particularly disfavoured. For
example, exposure of the homoprenyl derivative of alaninate 21
to triflic acid gave only the pyrrolidines 22 (cis:trans ~ 3+2)
(Scheme 2),6 despite the implication that a less stable secondary
carbocation is involved. Similarly, treatment of the cyclohex-
enyl derivative 24 with catalytic triflic acid did not give the
spiro-piperidine 25 but rather the 6/7 ring system 26 (Scheme
3). In all the above examples, isolated yields were in excess of
90%.
Scheme 4
tions to give the azasteroid derivatives 32 in 80–83% isolated
yields, although as a gross mixture of stereoisomers (Scheme
4).8
In conclusion, we have established that acid-catalysed overall
5-endo-trig cyclisations of homoallylic sulfonamides are useful
for the rapid and efficient assembly of pyrrolidines, and are
likely to be of some generality but which will clearly be limited
to substrates having other functionality which is stable to triflic
acid. Similarly general may be an isomerization process which
leads to a considerable predominance of the more thermody-
namically stable isomers, using an excess of the acid. Finally,
the method has some potential for the elaboration of polycyclic
systems by cation cyclisation cascades terminated by the
sulfonamide function although, again, other functional group
sensitivity to triflic acid will have to be kept in mind in future
applications.9
Finally, we were intrigued by the possibility that similar
chemistry could be applied to related polyene systems using a
sulfonamide group to terminate the cascade, in a further
development of the classical biomimetic studies of Johnson.7
We were delighted to find that the geranyl derivatives 27
underwent rapid cyclisation at 0 °C to give ca. 90% isolated
yields of the trans-annulated pyrrolidines 28, both as 3+2
mixtures, epimeric at the amino-ester stereogenic centre
(Scheme 4). This assignment of structure was confirmed by a
single crystal X-ray crystallographic determination of a sepa-
rated sample of the major isomer of the glycine derivative 28a.
Under the same conditions, the corresponding farnesyl deriva-
tives 29 also gave excellent returns of the tricyclics 30, but now
in isomer ratios of ca 3+3+1+1; the major products were the
trans-fused pair, epimeric at the amino-ester centre. The two
geranylgeranyl derivatives 31 also underwent smooth cyclisa-
We are grateful to Professor G. Pattenden for a generous
supply of geranylgeraniol, to Dr K. M. A. Malik for the X-ray
data which will be published elsewhere, to Dr N. J. Haskins and
the EPSRC Mass Spectrometry Service, University College,
Swansea for mass spectral data and to the EPSRC for financial
support.
Notes and references
1 J. E. Baldwin, J. Chem. Soc., Chem. Commun., 1976, 734; J. E. Baldwin,
J. Cutting, W. Dupont, L. Kruse, L. Silberman and R. C. Thomas, J.
Chem. Soc., Chem. Commun., 1976, 736.
2 D. W. Knight, Prog. Heterocycl. Chem., 2002, 14, in press.
3 A. D. Jones, D. W. Knight and D. E. Hibbs, J. Chem. Soc., Perkin Trans.
1, 2001, 1182.
4 A. D. Jones, D. W. Knight, A. L. Redfern and J. Gilmore, Tetrahedron
Lett., 1999, 40, 3267.
5 G. Stork, A. Y. W. Leong and A. M. Touzin, J. Org. Chem., 1976, 41,
3491.
6 Structural and stereochemical assignments were made by comparisons
with authentic compounds, obtained by deiodination of the correspond-
ing iodopyrrolidines whose structures have been determined by extensive
NMR studies and confirmed by X-ray crystallography (ref. 3).
7 J. K. Sutherland, Comp. Org. Synth., 1991, 3, 341. For a recent
contribution, see: W. S. Johnson, W. R. Bartlett, B. A. Czeskis, A.
Gantier, C. H. Lee, R. Lemoine, E. J. Leopold, G. R. Luedtke and K. J.
Bancroft, J. Org. Chem., 1999, 64, 9587.
Scheme 2
8 Detailed mass spectral analyses of cyclised products 28, 30 and 32
provided excellent additional evidence for these structural assignments.
We are grateful to Dr N. J. Haskins for these data which will be published
elsewhere.
9 Full analytical and spectroscopic data consistent with the proposed
structures have been obtained for all structures reported herein.
Scheme 3
CHEM. COMMUN., 2002, 2724–2725
2725