Scope and limitations of the [1,2]-alkylsulfanyl (SMe, SEt and SCH2Ph) and sulfanyl (SH) migration in the stereospecific synthesis of substituted tetrahydrofurans

Article abstract of DOI:10.1039/b007284m

Acid catalysed rearrangement of a series of 4-RS-1,3-diols (R = Me, Et, Bn and H) with toluene-p-sulfonic acid in dichloromethane gives stereospecifically substituted tetrahydrofurans via a [1,2]-SR shift in near quantitative yield. We comment on the structural variation of the migrating (RS) substituent and that of the migration origin and terminus on the outcome of the title reaction. We also report on the surprising similarity between an alkylsulfanyl (RS) and sulfanyl (SH) migrating group.

Full text of DOI:10.1039/b007284m

View Article Online / Journal Homepage / Table of Contents for this issue
Scope and limitations of the [1,2]-alkylsulfanyl (SMe, SEt and
SCH2Ph) and sulfanyl (SH) migration in the stereospecific
synthesis of substituted tetrahydrofurans†
1
Jason Eames,*a,b Nikolai Kuhnerta,c and Stuart Warren*a
a University Chemical Laboratory, Lensfield Road, Cambridge, UK CB2 1EW
b Department of Chemistry, Queen Mary, University of London, Mile End Road, London,
UK E1 4NS
c Department of Chemistry, University of Surrey, Guildford, Surrey, UK GU2 5XH
Received (in Cambridge, UK) 8th September 2000, Accepted 10th November 2000
First published as an Advance Article on the web 18th December 2000
Acid catalysed rearrangement of a series of 4-RS-1,3-diols (R = Me, Et, Bn and H) with toluene-p-sulfonic acid in
dichloromethane gives stereospecifically substituted tetrahydrofurans via a [1,2]-SR shift in near quantitative yield.
We comment on the structural variation of the migrating (RS) substituent and that of the migration origin and
terminus on the outcome of the title reaction. We also report on the surprising similarity between an alkylsulfanyl
(RS) and sulfanyl (SH) migrating group.
In a series of papers, we have reported numerous rearrange-
many of these 2-PhS-aldehyde precursors were easily synthesised
ments of substituted 1,3-diols (e.g. anti-3) involving [1,2]-SPh
from commercially available starting materials (PhSCl or
PhSCH2OMe).8,9 Furthermore, there was no danger in the loss
of the Ph group by nucleophilic attack on the sulfonium ion
intermediates such as the thiolanium ion syn-2 and the epi-
sulfonium ion syn-4 during the migration process. We now
report on the cyclisation of a new class of 1,3-diol with structural
variation at the migrating (RS) substituent (where R = Me, Et,
Bn and H) and comment on their relative performance, and in
particular the use of a sulfanyl (SH) migrating group.10
We required the 2-RS-aldehydes 8a–c, 11, 14, 17 and 20
for this study. These were synthesised by sulfenylation11 of the
silyl enol ethers 7, 10, 13, 16 and 19 (derived from the parent
aldehydes 6, 9, 12, 15 and 18) with RSCl—freshly prepared
by the addition of sulfonyl dichloride (SO2Cl2) to a stirred
solution of RSSR in CH2Cl2. The yields were excellent and this
procedure appears to be as efficient for simple alkyl groups
(e.g., R = Me and Et) as that previously reported1 for a phenyl
group (R = Ph) (see entry d in Table 1). The yield was slightly
reduced when using the more reactive benzylsulfanyl chloride
(BnSCl) (Scheme 2).
We synthesised the diol precursors using either the reliable
anti-stereoselective aldol reaction of the lithium (E)-enolate 22
of Heathcock’s ester (2,6-dimethylphenyl propionate 21)12,13 or
the syn-stereoselective aldol from the boron (Z)-enolate 24 of
Masamune’s ester (S-phenyl thiopropionate 23)14 with 2-RS-
aldehydes giving predictably single diastereoisomeric adducts
with greater than 98% stereocontrol (Schemes 3 and 5).
The rearrangement of the simple cyclic 1,3-diols anti-26a–c
(R = Me, Et and Bn) with a symmetrical migration origin was
studied (Scheme 4), primarily to see whether there were any
unusual effects on the rearrangement upon changing from
a phenylsulfanyl (PhS) migrating substituent, since we have
previously observed significant changes in the mechanistic
pathway in related diols when investigating [1,4]-SR shifts.10
These diols were synthesised from the 2-RS-aldehyde 8a–c
(R = Me, Et and Bn) and the lithium (E)-enolate 22 giving
diastereoisomerically pure aldol anti-25a–c (R = Me, Et and
Bn) in excellent yield (Table 1). Subsequent reduction (LiAlH4
in ether) gave the diols anti-26a–c required for the rearrange-
ment study. Treatment of these under our usual conditions1
(catalytic TsOH in refluxing CH2Cl2 for 5 minutes) gave stereo-
specifically the spirocyclic tetrahydrofurans anti-28a–c in high
Scheme 1
shift1 (to give tetrahydrofuran anti-5), [1,3]-,2 [1,4]-SPh shift3,4
(to give allylic alcohols anti-1) and [2,3]-SPh shifts to give a
variety of stereoisomerically pure heterocyclic6,7 and allylic3
derivatives involving stereospecific C–O, C–N, C–S and C–C
bond formation (Scheme 1). Apart from minor use of aromatic
derivatives, such as 4-Me-2 and 4-MeO-PhS5 groups we have
always used the phenylsulfanyl (PhS) group. This was chosen
due to its UV activity and ease of removal,7 but also because
† Experimental details are available as supplementary data. For direct
electronic access see http://www.rsc.org/suppdata/p1/b0/b007284m/
138
J. Chem. Soc., Perkin Trans. 1, 2001, 138–143
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b007284m

View Article Online
Table 1 The yield for the synthesis of 4-RS-1,3-diols anti-26a–c and THF anti-28a–c
Series R
RS-aldehyde
Yield (%)
Aldol
Yield (%)
Diol
Yield (%)
THF’s
Yield (%)
a; Me
8a
8b
8c
8d
97
98
83
98
anti-25a
anti-25b
anti-25c
anti-25d
89
93
93
84
anti-26a
anti-26b
anti-26c
anti-3
89
94
95
89
anti-28a
anti-28b
anti-28c
anti-5
99
99
99
92
b; Et
c; CH2Ph
d; Ph1
Scheme 2
Scheme 3 Reagents and conditions: a LDA, THF, Ϫ78 ЊC; b 9-BBN-
OTf, toluene, Ϫ30 ЊC.
yield, presumably via a hybrid 6-endo/5-exo-tet cyclisation1
of the episulfonium ion anti-27 with an overall [1,2]-SR shift.
These reactions were essentially as good as those with the
SPh group,1,15 and the yields were near quantitative, giving
very little room for any dealkylation of the intermediate epi-
sulfonium ion 27.
For the remainder of this study, we chose to use a benzyl-
sulfanyl (BnS) migrating group since this would be a better
substituent16 for observing any possible SN2 dealkylation of
intermediate episulfonium ions and would also give access to
the more interesting free sulfanyl (SH) group by reductive
removal of the benzyl group.17 We synthesised all the possible
structural combinations of the acyclic 4-BnS-1,3-diols 31,
anti- and syn-33 and 37 to probe the effect of stereochemistry
on the [1,2]-RS shift. These diols were made by the addition
of the lithium enolate of ethyl acetate, 2,6-dimethylphenyl
propionate or isobutyrate as well as addition of the syn-
stereoselective boron enolate of Masamune’s thioester S-
phenyl thiopropionate14 to the aldehyde 11, followed by simple
Scheme 4
LiAlH4 reduction in ether (Scheme 5). The relative stereo-
chemistry obtained using the stereoselective aldol reaction with
2,6-dimethylphenyl propionate and S-phenyl thiopropionate is
very reliable and the corresponding phenylsulfanyl derivatives
have previously been reported.23 Subsequent rearrangement of
these diols 31, syn- and anti-33 and 37 with TsOH in CH2Cl2
gave stereospecifically the tetrahydrofurans 38, syn- and anti-39
J. Chem. Soc., Perkin Trans. 1, 2001, 138–143
139

View Article Online
Table 2 The yield for the synthesis of 3-BnS THF’s 38, 39 and 40 and THF’s 42, 44 and 46
Yield
(%)
Yield
(%)
Yield
(%)
Yield
(%)
Yield
(%)
RS-aldehyde
Aldol
Diol
3-BnS-THF’s
Thiols
3-HS-THF’s
11
11
11
11
30
95
90
65
73
31
93
92
92
87
38
96
99
99
99
41
70
79
68
75
42
96
98
95
98
anti-32
syn-34
36
anti-33
syn-33
37
anti-39
syn-39
40
anti-43
syn-43
45
anti-44
syn-44
46
Scheme 5
Scheme 6
and 40 in near quantitative yield (the lowest yield being
96%). Their behaviour is identical to that of the PhS analogues
(Table 2).1 These reactions occur cleanly and cyclisation onto
the episulfonium ion is stereospecific: the anti-diol 33 gives the
anti-tetrahydrofuran 39, whilst the syn-diol 33 gives the syn-
tetrahydrofuran 39 (Scheme 6).
A remaining question was whether an SH grouping itself
could be used in this [1,2]-SR reaction. It offered particular
interest as simple loss of a proton might occur easily in the
[1,2]-SH shift. The next step was to remove the benzyl group
to give the free sulfanyl (SH) group.17 This was easily achieved
by reduction with sodium in liquid ammonia in good yield
(60–90%) to give the thiols 41, anti- and syn-43, 45 and anti-47
as pleasant smelling liquids. Rearrangement of this series of
4-HS-1,3-diols 41, anti- and syn-43, 45 and anti-47 gave the
tetrahydrofurans 42, anti- and syn-44, 46 and anti-49 in good
yield (Table 2). Though the yields of the sulfanyl tetrahydro-
furans are as high as the phenyl- and alkylsulfanyl cases, the
reaction is much slower. One hour’s refluxing with catalytic
TsOH in CH2Cl2 is necessary rather than just a few minutes.
The reaction certainly does proceed with inversion at the
migration terminus and therefore via a protonated episulfide
such as syn-48 (Scheme 7). We have never observed any epi-
sulfide formation under these conditions, presumably showing
that proton loss from 48 must be slower than capture of the
episulfonium ion with the OH group. The longer reaction time
indicates, as might be expected, that the sulfanyl (SH) group is
less nucleophilic than a comparable alkyl or arylsulfanyl group.
To establish whether this [1,2]-SBn shift occurs stereo-
specifically with clean inversion at both the migration origin
and terminus, we synthesised a series of 4-BnS-1,3-diols with an
unsymmetrical acyclic tertiary migration origin. These diols
were synthesised from the chiral 2-BnS-aldehyde 17 by reaction
with Heathcock’s enolate (E)-22 and with acetone (Schemes 8
and 9). The C(3,4)-Felkin–Anh selectivity18 was good (Table 3),
whereas the C(2,3)-stereocontrol (>98:<2) in anti,anti-50
140
J. Chem. Soc., Perkin Trans. 1, 2001, 138–143

View Article Online
Table 3 The yield for the synthesis of 4-BnS 3-hydroxyl esters-50, 54,
59 and 62
Aldol C(3,4)
selectivity
(Felkin–Anh)
RS-
aldehyde
Aldol C(2,3)
selectivity
Yield
(%)
Aldol
17
17
14
20
anti,anti-50
anti-54
>98:<2
—
—
89:11
86:14
67:33
75:25
85
79
91
92
anti-59
anti,anti-62
>98:<2
Scheme 7
Scheme 8
was very well controlled by (E)-enolate geometry.13 Simple
reduction (LiAlH4 in ether) of the ester anti,anti-50 gave the
1,3-diol anti,anti-51, which has the all-important stereogenicity
at the migration origin and what would become the migration
terminus in the tetrahydrofurans. Acid catalysed rearrangement
of this diol (TsOH in CH2Cl2) gave the corresponding tetra-
hydrofuran anti,anti-53 as a single diastereoisomer. Evidently,
the cyclisation via the episulfonium ion 52 was stereospecific
with inversion at both the migration origin and terminus
(determined by NOE differences). Clean inversion occurs at a
tertiary stereocentre, and thus the cyclisation must be occurring
via a rather loose SN2–tight SN1 transition state to account for
the observed stereochemical outcome.
The structural nature of the cyclising alcohol (OH group)
was also found to be unimportant on the outcome of the
reaction. A secondary or tertiary OH group behaved identically
to a cyclising primary OH group to give similar tetrahydro-
furans in high yield. The diols anti,syn- and anti,anti-55 con-
taining the required secondary OH group were synthesised
using the reliable syn19 and anti-1,3-stereoselective20 reduction
developed by Prasad and Evans and their co-workers as shown
in Scheme 9. Treatment of these diols anti,syn- and anti,anti-55
with TsOH in CH2Cl2 gave stereospecifically the tetrahydro-
furans anti,syn- and anti,anti-57 in excellent yield. Inversion
at the migration origin and retention of configuration at the
cyclising secondary alcohol were observed. The rearrangement
of the diol 56 (synthesised by the addition of MeMgCl to
the ketone anti-54) with a cyclising tertiary OH group also
proceeded cleanly to give the tetrahydrofuran anti-58 in near
quantitative yield (Scheme 9). The reaction appears to be
Scheme 9
insensitive to the substitution pattern and the relative stereo-
chemistry of the cyclising nucleophile.
Another variant was the rearrangement of 1,3-diols such as
syn- and anti-60, anti,anti- and syn,anti-63 with a secondary
migration origin. These diols were synthesised by our usual
aldol and reduction methodology as shown in Schemes 10 and
11. However, this case was slightly different since the addition
of the pre-cooled aldehyde 14 and 20 was required to prevent
enolisation of the aldehyde (by the enolate) in the aldol step.
The Felkin selectivity18 C(3,4) observed was moderate, and was
significantly lower than in previous cases involving the sterically
demanding tertiary 2-BnS-aldehyde 17 (Table 3). The relative
C(2,3) stereochemistry was well controlled (always >98:<2).
Acid catalysed rearrangement of these 4-BnS-1,3-diols gave
stereospecifically the tetrahydrofurans anti- and syn-61, anti,
anti- and syn,anti-64 in quantitative yield (Schemes 10 and 11).
The reaction was stereospecific since a different diastereoiso-
meric diol anti- and syn-63 leads to a different diastereoisomeric
tetrahydrofuran 64 in both cases (Table 4). The yields were
slightly lower than those of the corresponding diols with a ter-
tiary migration origin, and the reaction times were at least one
order of magnitude longer. Clearly, the rate determining form-
J. Chem. Soc., Perkin Trans. 1, 2001, 138–143
141

View Article Online
Table 4 The yield for the synthesis of 4-BnS diols and THF’s 53, 57,
58, 61 and 62
ation of the intermediate episulfonium ion is slower and this is
presumably due to the presence of the less substituted migra-
tion origin, a manifestation of the exo-component of the
Thorpe–Ingold effect.21,22
Yield
(%)
Yield
(%)
Ester/ketone
Diols
3-BnS-THF’s
We rearranged a series of 1,3-diols with a primary migration
origin in an attempt to probe this Thorpe–Ingold effect.22 These
were synthesised by the addition of the appropriate lithium
enolate (derived from 21, 29 and 35) to the commercially avail-
able BnS-ketone 65 (Scheme 12). LiAlH4 reduction gave the
4-BnS-1,3-diols 67, anti- and syn-69 and 71 in excellent yield
(Table 5). Treatment of these diols (e.g. 69) with TsOH in
refluxing CH2Cl2 gave no tetrahydrofuran formation, even on
prolonged reflux (24 hours) and simply gave recovered starting
material. If the reaction was forced by prolonged reflux in a
higher boiling solvent, such as toluene, these diols simply
decomposed to give unidentified products. Presumably, the rate
determining cyclisation step to form such an episulfonium
anti,anti-50
anti-54
anti,anti-51
anti,syn-55
anti,anti-55
anti-56
88
91
80
94
91
88
93
91
anti,syn-53
anti,syn-57a
anti,anti-57b
anti-58
98
91
90
85
98
99
99
98
anti-54
anti-54
anti-59
anti-60
anti-61
syn-59
syn-60
anti,anti-63
syn,anti-63
syn-61
anti,anti-64
syn,anti-64
anti,anti-62
syn,anti-62
a Selectivity of diastereoselective reduction: anti,syn-57–anti,anti-57;
>98:<2. b Selectivity of diastereoselective reduction: anti,anti-57–
anti,syn-57; 92:8.
Scheme 10
Scheme 11
Scheme 12
142
J. Chem. Soc., Perkin Trans. 1, 2001, 138–143

View Article Online
Table 5 The yield for the synthesis of 4-BnS-1,3-diols 67, 69 and 71
from a primary migration origin does not occur and decom-
position of the diol is observed. The yields for all these
cyclisations are near quantitative (the lowest being 85%).
The cyclisation occurs cleanly with complete stereochemical
control giving virtually all the possible combinations of stereo-
chemistry and substitution pattern within the tetrahydrofuran
framework.
RS-
ketone
Aldol C(2,3)
selectivity
Yield
(%)
Yield
(%)
Aldol
Diol
65
65
66
—
75:25
93
91
67
92
65
22
93
anti-68–
syn-68
70
anti-69
syn-69
71
65
—
95
Experimental
Experimental details are available as supplementary data. For
direct electronic access see http://www.rsc.org/suppdata/p1/b0/
b007284m/.
ion 72 is no longer favourable and an alternative unknown
decomposition pathway is now favoured. This is not entirely
surprising since it has been reported that the rate of a cyclis-
ation can be improved by up to a million fold by the use of a
neo-pentyl nucleophile.22 Previous studies within our laboratory
involving the PhS group have shown the cyclisation is also
dependent on both the substituent pattern at the migration
origin (tertiary cyclise more efficiently than secondary) and the
anti- developing stereochemistry is preferred over syn- within
the tetrahydrofuran framework.23
Acknowledgements
We thank the EPSRC for a grant (to J. E.), Zeneca Process
Technology Department, Grangemouth for a CASE award
(to J. E.), RTL and DFG (to N. K.) for grants.
Rate enhancement of substitution by participation with a PhS
group is well documented.24 Participation by an arylsulfanyl
group has also been reported, though much less frequently.2,5
The use of alkylsulfanyl groups is even rarer. Within this
area, MeS is the most common whereas the use of the more
reactive BnS is very rare. However, there are cases where a
benzylsulfanyl group is known to participate through a four-
membered ring.24 The sulfanyl (SH) group is a commonly used
nucleophile in cyclisations resulting in cyclic sulfides, even in
acidic conditions,25,26 but is rarely seen as a participating group
which accelerates a reaction, and also remains intact as SH at
the end of the reaction. We believe our results report the first
preparatively useful [1,2]-SH shift. In a destructive sense,
participation by SH is believed to be the cause of the failure of
HS(CH2)2OH as a protecting group in a lysozyme synthesis.27
By comparison, we have already reported that [1,4]-SH
participation28 during PhS migration, as in the acid catalysed
rearrangement of 74, leads to thiolane formation 76 rather than
[1,4]-SH migration (Scheme 13). Whereas the episulfonium
References
1 V. K. Aggarwal, I. Coldham, S. McIntyre and S. Warren, J. Chem.
Soc., Perkin Trans. 1, 1991, 451.
2 P. Brownbridge and S. Warren, J. Chem. Soc., Perkin Trans. 1, 1976,
2125.
3 M.-J. Villa and S. Warren, J. Chem. Soc., Perkin Trans. 1, 1994,
2953.
4 J. Eames, M. A. de las Heras, R. V. H. Jones and S. Warren,
Tetrahedron Lett., 1996, 37, 1117.
5 R. C. Hartley, I. C. Richards and S. Warren, J. Chem. Soc., Perkin
Trans. 1, 1994, 359.
6 K. Chibale and S. Warren, J. Chem. Soc., Perkin Trans. 1, 1995,
2411.
7 I. Coldham and S. Warren, J. Chem. Soc., Perkin Trans. 1, 1993,
1637.
8 A. de Groot and B. J. M. Jansen, Synthesis, 1985, 434.
9 A. de Groot and B. J. M. Jansen, Tetrahedron Lett., 1981, 22,
887.
10 Preliminary communication: J. Eames, R. V. H. Jones and
S. Warren, Tetrahedron Lett., 1996, 27, 4823.
11 S. M. Tuladhar and A. G. Fallis, Can. J. Chem., 1987, 65, 1833.
12 C. H. Heathcock, C. T. Buse, W. A. Kleschick, M. C. Pirrung,
J. E. Sohn and J. Lampe, J. Org. Chem., 1980, 45, 1066.
13 C. H. Heathcock, M. C. Pirrung, S. H. Montgomery and J. Lampe,
Tetrahedron, 1981, 37, 4087.
14 S. Masamune, L. D. L. Lu, D. S. Garvey and M. Hirama,
Tetrahedron Lett., 1975, 1225.
15 L. Djakovitch, J. Eames, R. V. H. Jones, S. McIntyre and S. Warren,
Tetrahedron Lett., 1995, 36, 1723.
Scheme 13
16 A. J. Kirby, in Stereoelectronic Effects, Oxford Chemistry Primer,
Oxford University Press, Oxford, 1996, p. 39.
17 W. H. Hartung and R. Simonoff, Org. React., 1953, 7, 263.
18 (a) M. Cherest, H. Felkin and N. Prudent, Tetrahedron Lett.,
1968, 2199; (b) N. T. Anh and O. Eisenstein, Nouv. J. Chim., 1977,
1, 77.
19 K.-M. Chen, G. E. Hardtmann, K. Prasad, O. Repic and M. J.
Shapiro, Chem. Lett., 1987, 10, 1923.
20 K. T. Chapman, D. A. Evans and E. M. Carreira, J. Am. Chem. Soc.,
1988, 110, 3560.
21 P. G. Sammes and D. J. Weller, Synthesis, 1995, 1205.
22 A. J. Kirby, Adv. Phys. Org. Chem., 1981, 17, 183.
23 V. K. Aggarwal, J. Eames, M.-J. Villa, S. McIntyre, F. H. Sansbury
and S. Warren, J. Chem. Soc., Perkin Trans. 1, 2000, 533.
24 E. L. Eliel, W. H. Pearson, L. M. Jewell, A. G. Abatjoglou and
W. R. Kenan, Tetrahedron Lett., 1980, 21, 331.
25 C. D. Maycock and R. J. Stoodley, J. Chem Soc., Perkin Trans. 1,
1979, 1852.
26 T. C. Owen and J. K. Leone, J. Org. Chem., 1992, 57, 6585.
27 J. J. Sharp, A. B. Robinson and M. D. Kamen, J. Am. Chem. Soc.,
1973, 95, 6097.
28 J. Eames, R. V. H. Jones and S. Warren, Tetrahedron Lett., 1996, 37,
707.
ion syn-48 does not lose a proton from sulfur, but rather con-
tinues SH migration to give the tetrahydrofuran anti-49, the
thiolanium ion 75 does lose a proton under the same acidic
conditions to give the thiolane 76, presumably because both
reactions are under thermodynamic control.
The effect of alkyl-S participation has been estimated as 30
times that of O-alkyl and 103 times that of alkyl participation.29
A comparision of the solvolysis of anti-2-chlorocyclohexanol
and anti-2-chlorocyclohexanethiol revealed that an SH group
was about 104 times more efficient than OH as a participating
group.30 We cannot compare RS or SH participation with RO
or OH participation, but it is clear that SH is at least an order
of magnitude less effective than alkyl-S or aryl-S.
In conclusion, we have shown that migrating substituents RS
(R = Me, Et, Bn and H) in diols like anti-26 are as efficient
as the previously reported PhS cases.23 These [1,2]-RS shifts
are indeed stereospecific with inversion observed at both the
migration origin and terminus. The relative reaction rates of
cyclisation are all similar, except for those with the sulfanyl
(SH) migrating group, which are at least a magnitude slower.
Furthermore, the cyclisation occurs more efficiently for a
tertiary migration origin than a secondary, whereas migration
29 H. Morita and S. Oae, Tetrahedron Lett., 1969, 1347.
30 P. Crouzet, E. Laurent-Dieuzeide and J. Wylde, Bull. Soc. Chim. Fr.,
1968, 1463.
J. Chem. Soc., Perkin Trans. 1, 2001, 138–143
143