The scope and limitation of the [1,4]-SPh shift in the synthesis of allylic alcohols

Article abstract of DOI:10.1039/b005349j

Treatment of a series of 4-phenylsulfanyl-1,3-diols with toluene-p-sulfonyl chloride in pyridine gives stereospecifically substituted allylic alcohols in near quantitative yield via a [1,4]-SPh shift. We comment on the structural variation at both the mig

Full text of DOI:10.1039/b005349j

View Article Online / Journal Homepage / Table of Contents for this issue
The scope and limitation of the [1,4]-SPh shift in the synthesis
of allylic alcohols†
Varinder K. Aggarwal,a,b Jason Eames,*a,c Maria A. de las Heras,a Sara McIntyrea and
Stuart Warren*a
1
a University Chemical Laboratory, Lensfield Road, Cambridge, UK CB2 1EW
b School of Chemistry, University of Bristol, Bristol, UK BS8 1TS
c Department of Chemistry, Queen Mary, University of London, Mile End Road, London,
UK E1 4NS
Received (in Cambridge, UK) 4th July 2000, Accepted 31st October 2000
First published as an Advance Article on the web 30th November 2000
Treatment of a series of 4-phenylsulfanyl-1,3-diols with toluene-p-sulfonyl chloride in pyridine gives stereospecifically
substituted allylic alcohols in near quantitative yield via a [1,4]-SPh shift. We comment on the structural variation at
both the migration origin and terminus on the outcome of the title reaction and define its limits.
[1,4]-SR Participation by a sulfur atom via a five-membered
sulfonium ion intermediate is a well documented effect.1 The
majority of attention has been concerned with the rate of
acceleration of simple substitution reactions.2 Very little is
known3 about the acceptable substitution pattern for such
reactions or the product diversity and distribution. It is even
rarer that such reactions have been used in synthesis.4
We have previously shown that the treatment of the 4-PhS-
1,3-diol 1 with toluene-p-sulfonyl chloride (TsCl) in pyridine
gave instead of the expected tosylate 2, the allylic alcohol 4 in
97% yield, presumably via the sulfonium ion 3 and a [1,4]-SPh
shift (Scheme 1).5 We now report on the rearrangement of
and the unrearranged chloride 7 (substitution at what would
be the migratory terminus, but with no SPh migration)
(Scheme 2).
The required 4-phenylsulfanyl-1,3-diols 12, anti- and syn-15,
18, 21, 23, 25, 27, 30, anti- and syn-32, anti-34 and anti,anti-38,
syn and anti-47, 60, 61 and syn-62 for this study were all
prepared using known stereoselective aldol methodology with
α-PhS substituted aldehydes and most have been reported
previously.9
We first established, by the rearrangement of the alcohol 12,
that a secondary hydroxy group is unnecessary. This alcohol 12
was synthesised using a two step procedure: refluxing the stab-
ilised Wittig reagent 910 with the aldehyde 8 gave exclusively the
(E)-enoate 10 in excellent yield (Scheme 3). Non-regioselective
reduction with LiBH4 gave the required alcohol 12 in 24% and
the allylic alcohol 11 in 74% yield. Treatment of 12 with TsCl
in pyridine gave the cyclohexene 14 in 93% yield, by simple exo-
elimination via the sulfonium ion intermediate 13 (Scheme 3).
The rearrangement of more substituted cases, like anti- and
syn-1,3 diols 15 occurred as efficiently as the unsubstituted case
12 giving the anti- and syn-allylic alcohols 17 in excellent yield.
The [1,4]-SPh participation occurs independent of the develop-
ing stereochemistry within the sulfonium salt: anti-diol 15 gives
the anti-allylic alcohol 17 stereospecifically via the sulfonium
salt syn-16, while the syn-diol 15 gives the syn-allylic alcohol 17
via the sulfonium salt anti-16 (Scheme 4).
The effect of ring strain in such sulfonium intermediates as 3
was found to be more dependent on the size of the adjacent
spirocarbocyclic ring than on the substitution pattern of the
sulfonium ion: the five-membered ring compounds anti- and
syn-18, 21 and 23 rearranged in the same way as the six-
membered ring compounds to give the allylic alcohols anti- and
syn-20, 22 and 24. The reaction occurs irrespective of whether
there is a similar OH group 21, gem-dimethyl groups 23, or two
substituents (OH and Me) on adjacent carbon atoms arranged
anti- or syn- around the sulfonium salt 19 (Scheme 5). A larger
medium ring behaves in the same way; treatment of the
cyclodecane 25 gave the (E)-allylic alcohol 26 in good yield.
The (E)-stereochemistry was inferred from previous studies on
the toluene-p-sulfonic acid catalysed [1,2]-SPh rearrangement
of similar medium ring carbocycles.11
Scheme 1
related 4-PhS-1,3-diols with TsCl in pyridine6 with structural
variation at both the migration origin and terminus and the
synthesis of allylic alcohols. We comment on stereochemical
effects, structural variation and the Thorpe–Ingold effect7 on
the efficiency and outcome of the reaction. We also disclose
methods that allow the isolation of primary tosylate analogues.
The four possible products from this spirocyclic sulfonium
salt 3 are the allylic alcohol 4 (by elimination exo to the
sulfonium ring with [1,4]-SPh shift),8 the ketone 5 (by endo
elimination with [1,4]-SPh shift), the rearranged chloride 6
(substitution at the migratory origin with [1,4]-SPh shift),
† The Experimental section for this paper is available as supplementary
data. For direct electronic access see http://www.rsc.org/suppdata/p1/
b0/b005349j/
This exo-elimination pathway was disfavoured with a cyclo-
butane since formation of a cyclobutene is indeed disfavoured
4456
J. Chem. Soc., Perkin Trans. 1, 2000, 4456–4461
This journal is © The Royal Society of Chemistry 2000
DOI: 10.1039/b005349j

View Article Online
Scheme 2 Four possible products from the rearrangement of the diol 1 via the sulfonium ion 3.
Scheme 3
Scheme 4
by ring strain. Reaction of diol 27 under our usual con-
ation is very efficient, essentially as efficient as [1,2]-SPh
ditions gave the unrearranged chloride 29 in 89% yield with
no SPh migration. We believe that formation of the strained
sulfonium salt 28 indeed does occur as [1,4]-SPh particip-
migration which can occur even via a highly strained epi-
sulfonium ion.1 The chloride 29 must result from a substit-
ution reaction of 28 via a tight SN2 transition state with
Scheme 5
J. Chem. Soc., Perkin Trans. 1, 2000, 4456–4461
4457

View Article Online
attack occurring at the less hindered primary carbon of the
sulfonium intermediate (Scheme 6). Eliel has observed similar
behavour.12
It appears that the exo-elimination pathway was favoured
over all other pathways (Scheme 2) due to the near-perfect
alignment of the antiperiplanar β-hydrogen required for the
exo-E2 elimination process in these spirocyclic sulfonium
intermediates. The four-membered carbocycle 27 was the excep-
tion because the exo-elimination pathway was disfavoured due
to ring strain. For a more realistic comparison between these
exo- and endo-elimination pathways, we rearranged acyclic 4-
PhS-1,3-diols such as anti-34 and anti,anti-38. Treatment of the
acyclic diol 34 with TsCl in pyridine gave for the first time two
products: the allylic alcohol 36 (major) and ketone 37 (minor)
in a 85:15 ratio. The ketone 37 was presumably formed via an
enol, a tetra-substituted alkene, by endo-elimination of the
sulfonium salt syn-35. Nevertheless the allylic alcohol anti-36
was the major product.
The stereoselectivity in double bond formation in such elim-
ination processes was revealed by rearrangement of the diol
anti, anti-38 under our usual conditions. It gave the (E)-allylic
alcohol anti-40 as the major product [(E-):(Z-) 88:12—
determined by a 500 MHz NOESY spectrum], the most
thermodynamically stable of the three (E-40, Z-40 and the exo-
methylene compound 41) possible exo-elimination products.
This elimination pathway presumably occurs via the transition
state conformation A rather than conformation B due to
unfavourable 1,3-diaxial interactions (Scheme 8). The altern-
ative endo-elimination in 39 to give the ketone is not observed
presumably because exo-elimination to give (E)-40 is preferred.
The synthesis of ketones by elimination during [1,4]-PhS
migration can be achieved if the non-participating secondary
alcohol grouping is exo- to the sulfonium ring rather than
endo- as in the previous cases. The diols syn- and anti-47 were
synthesised by diastereoselective addition of MeMgBr to the
α-PhS substituted aldehyde 4514 (under Felkin control:
anti:syn = 71:29) followed by TBAF deprotection (Scheme 9).
The aldehyde 45 was easily synthesised from the protected
ketone 43 using the de Groot and Jansen rearrangement.9,15
Treatment of syn- and anti-diols 47 separately with TsCl in
pyridine gave the same ketone 50 in 96% and 95% respec-
tively, presumably via exo-elimination of the sulfonium ions
syn- and anti-48 and subsequent formation of an enol 49
(Scheme 10).
Scheme 6
The incorporation of a heteroatom substituent in the ring
(X = O or S) did not interfere with the reaction pathway. The
five and six-membered ring diols 30, syn- and anti-32 (X = O)13
and anti-32 (X = S) gave only the expected allylic alcohols 31,
syn- and anti-33 (X = O) and anti-33 (X = S) although in
reduced chemical yield. It appears that competitive 1,5- and
1,6-participation by the heteroatom X does not occur as it
is less efficient than the observed [1,4]-SPh participation
(Scheme 7).
The normal exo-elimination pathway is disfavoured if the
migration origin is secondary rather than tertiary. The 1,3-diols
60–62 were synthesised9b using our previously reported aldol
and reduction procedure as shown in Scheme 11. Addition of
the enolates 54, (E)-5516 and 56 to the 2-PhS-aldehyde 53
gave yields of aldols 57–59 which depended on the degree of
Scheme 7
Scheme 8
4458
J. Chem. Soc., Perkin Trans. 1, 2000, 4456–4461

View Article Online
Scheme 9
Scheme 10
Scheme 11
Table 1 Stereoselectivity in the aldol reactions
sponding diols anti- and syn-60, anti,anti- and anti,syn-61 and
syn-62 in excellent yield.
3,4-anti:syn
(Felkin)
Treatment of the anti- and syn-diols 60 with TsCl in pyridine
gave an inseparable regioisomeric mixture of the unrearranged
chlorides 64 (major) and rearranged chlorides 65 (minor) in
excellent yield: no allylic alcohols were observed (Scheme 12).
Presumably these reactions occur via a tight SN2 displacement
at the primary carbon of the sulfonium intermediate 63 giving
the unrearranged chloride 64 with no SPh migration as the
major product. The minor component must come from an SN2
displacement at the secondary centre in the sulfonium salt to
give the rearranged chloride 65 via a [1,4]-SPh migration. The
displacement is stereospecific because each stereoisomer of 60
gives a single diastereoisomer of the minor rearranged product
65 and we assume the reaction occurs cleanly with inversion.18
Aldol
2,3-anti:syn
Yield (%)
60
61
62
—
>98:2
—
75:25
67:33
>98:2
45
89
87
competitive enolisation of the aldehyde 53 under the reaction
conditions. The C(3,4)-Felkin–Anh17 selectivity was much
better (Table 1) for the more sterically demanding9b enolate 56,
than for the unsubstituted 54 or mono-substituted (E)-55,
whereas the C(2,3)-selectivity in the aldol 58 was controlled by
the enolate geometry.16 Reduction (LiAlH4) gave the corre-
J. Chem. Soc., Perkin Trans. 1, 2000, 4456–4461
4459

View Article Online
Table 2 Allylic alcohols from the rearrangement of 4-phenylsulfanyl-
1,3-diols with TsCl in pyridine
Diol
Allylic alcohol
Yield (%)
anti-15
anti-17
syn-17
94
96
95
93
93
95
90
50
60
84
79
85
90
syn-15
anti-18
anti-20
syn-18
syn-20
22
24
(E)-26
31
anti-33; X = O
anti-33; X = S
syn-33; X = O
anti-36
21
23
25
30 (1:1)
anti-32; X = O
anti-32; X = S
syn-32; X = O
anti-34
anti,anti-38
(E)-anti-40:anti-41
reactions of diol 61. The regioselective substitution of these
types of sulfonium intermediates is clearly sensitive to the C3
substituent and the stereochemistry in the sulfonium salts
69a ϩ b.
Attempts to isolate the original tosylate 73 under non-basic
conditions proved unsuccessful. Addition of n-BuLi (1 eq.) to a
solution of diol anti-15 in THF at Ϫ78 ЊC and reaction with
TsCl gave only the allylic alcohol 17 in near quantitative yield
(Scheme 14). Isolation of the analogous tosylate was achieved
by inhibiting sulfur participation by steric and electronic
factors. Reaction of the sulfone anti-74 with TsCl in pyridine
gave the primary tosylate anti-75 in 95% yield. Participation
by a sulfone is known to be much slower than the more
nucleophilic sulfide by at least four orders of magnitude.19
Reaction of the anti-diol 76 containing two secondary alcohols,
with TsCl in pyridine gave a 96% yield of a mixture of
secondary tosylates 77 and 78 in a ratio 88:12. No particip-
ation of SPh occurred with these more hindered secondary
tosylates.
In conclusion we have shown that the substituted 4-
phenylsulfanyl-1,3-diols fall into the following three categories.
1. Those with a tertiary origin and primary terminus (e.g.
anti-15)—all these rearrange to give allylic alcohols in excellent
yield (Table 2), except when exo-elimination is inhibited, as with
cyclobutane 27.
Scheme 12
These cases are different from previous ones because cleavage of
the much weaker C–S in the tertiary carbon in the sulfonium
ions like 16 and syn-35 derived from 15 and anti-34 is evidently
preferred rather than direct substitution. In comparison the
rearrangement of the remaining diols anti,anti-61 and syn-62,
gave a reversal of selectivity giving the rearranged chlorides syn,
anti-68 and anti-72 as the major product (Scheme 13). Presum-
ably, substitution of sulfonium intermediate (e.g. 69a) at the
primary carbon is now disfavoured due to the adjacent methyl
group. It appears that direct substitution via a loose SN2
reaction at the more substituted secondary centre is evidently
preferred. Indeed the substitution at the secondary centre is
stereospecific, as we obtain two different stereoisomers in both
Scheme 13
4460
J. Chem. Soc., Perkin Trans. 1, 2000, 4456–4461

View Article Online
Table 3 Alkyl chlorides from the rearrangement of 4-phenylsulfanyl-
1,3-diols with TsCl in pyridine
Experimental
The Experimental section for this paper is available as sup-
plementary data. For direct electronic access see http://
www.rsc.org/suppdata/p1/b0/b005349j
Unrearranged
chloride
Yield
(%)
Rearranged
chloride
Diol
Ratio
27
anti-60
syn-60
syn,anti-61
anti,anti-61
syn-62
29
anti-64
syn-64
syn,anti-67
anti,anti-67
syn-71
89
84
90
90
95
95
—
syn-65
anti-65
anti,anti-68
syn,anti-68
anti-72
—
Acknowledgements
76:24
86:14
81:19
17:83
20:80
We thank the EPSRC for a grant (to V. K. A, J. E. and S. A. M.),
Zeneca Process Technology Department, Grangemouth for
a CASE award (to J. E.) and the Spanish Ministerio de
Educación y Ciencia and Comisión Interministerial de Ciencia
y Technología (CICYT) for support (M. A. H.).
References
1 E. L. Eliel, W. H. Pearson, L. M. Jewell and A. G. Abatjoglou,
Tetrahedron Lett., 1980, 21, 331.
2 J. Eames, N. Kuhnert, F. H. Sansbury and S. Warren, Synlett, 1999,
1211.
3 M.-J. Villa and S. Warren, J. Chem. Soc., Perkin Trans. 1, 1994, 2954.
4 J. Eames, N. Kuhnert and S. Warren, Synlett, 1999, 1215.
5 (a) L. Djakovitch, J. Eames, R. V. H. Jones, S. McIntyre and
S. Warren, Tetrahedron Lett., 1995, 36, 1723; (b) L. Djakovitch,
J. Eames, D. J. Fox, F. H. Sansbury and S. Warren, J. Chem. Soc.,
Perkin Trans. 1, 1999, 2771.
6 Preliminary communication: J. Eames, M. A. de las Heras, R. V. H.
Jones and S. Warren, Tetrahedron Lett., 1996, 37, 1117.
7 (a) A. J. Kirby, Adv. Phys. Org. Chem., 1981, 17, 183; (b)
L. Mandolini, Adv. Phys. Org. Chem., 1986, 22, 1.
8 What is exo- to the sulfonium ring in the spirocyclic intermediate 13
is endo- to the cyclohexyl ring and vice versa.
9 (a) V. K. Aggarwal, I. Coldham, S. McIntyre and S. Warren,
J. Chem. Soc., Perkin Trans. 1, 1991, 451; (b) V. K. Aggarwal,
J. Eames, M.-J. Villa, S. McIntyre, F. H. Sansbury and S. Warren,
J. Chem. Soc., Perkin Trans. 1, 2000, 533.
10 M. S. Wadia, R. S. Mali, S. G. Tilve and V. J. Yadav, Synthesis, 1987,
401.
11 M. Hannaby, P. Judson and S. Warren, J. Chem. Soc., Perkin
Trans. 1, 1992, 2609.
12 E. L. Eliel, R. O. Hutchins, R. Mebane and R. L. Willer, J. Org.
Chem., 1976, 41, 1052.
13 J. Eames, D. J. Fox, M. A. de la Heras and S. Warren, J. Chem. Soc.,
Perkin Trans. 1, 2000, 1903.
14 S. McIntyre, F. H. Sansbury and S. Warren, Tetrahedron Lett., 1991,
32, 5409.
15 (a) A. de Groot and B. J. M. Jansen, Tetrahedron Lett., 1981, 22, 887;
A. de Groot and B. J. M. Jansen, Synthesis, 1985, 434; (b) J. Otera,
Synthesis, 1988, 95.
Scheme 14
2. Those with a tertiary migratory origin and secondary
terminus—no rearrangement is observed, instead a regio-
selective mixture of secondary tosylates was isolated (e.g. anti-
76).
3. Those with a secondary migratory origin and primary
terminus—all these rearrange to give a regioisomeric mixture
of rearranged and unrearranged chlorides, by substitution on
the intermediate sulfonium salt in excellent yield (Table 3).
16 C. H. Heathcock, M. C. Pirrung and J. E. Sohn, J. Org. Chem., 1979,
44, 4294.
17 (a) M. Chérest, H. Felkin and N. Prudent, Tetrahedron Lett., 1968,
2199; (b) N. T. Anh and O. Eisenstein, Nouv. J. Chim., 1977, 1, 77.
18 This was calculated by integration of the CH2Cl signal of the
unrearranged chloride and the CH2SPh signal in the rearranged
chloride within the proton NMR spectrum.
19 P. Brownbridge, P. K. G. Hodgson, R. Shepherd and S. Warren,
J. Chem. Soc., Perkin Trans. 1, 1976, 2024.
J. Chem. Soc., Perkin Trans. 1, 2000, 4456–4461
4461