Additions of stabilised and semi-stabilised sulfur ylides to tosyl protected imines: Are they under kinetic or thermodynamic control?

Article abstract of DOI:10.1039/b107275g

Sulfur ylides react with imines, via betaines, to give aziridines. We sought to determine whether betaine formation was reversible in reactions of benzyl-, amide- and ester-stabilised ylides by carrying out cross-over experiments. Thus, the intermediate betaines were generated independently from the corresponding sulfonium salt in the presence of a more reactive imine (p-nitrobenzaldimine). It was found that no incorporation of the more reactive imine was observed in reactions with the benzyl-stabilised ylide, whilst >80% incorporation of the p-nitrobenzaldimine was observed from the ester- and amide-stabilised ylides. These results indicate that benzyl-stabilised ylides react irreversibly with imines but ester- and amide-stabilised react reversibly. Thus, the stereocontrolling step of the process is dependent on the type of ylide employed and the results are used to account for the different diastereoselectivities observed with the different ylides.

Full text of DOI:10.1039/b107275g

View Article Online / Journal Homepage / Table of Contents for this issue
Additions of stabilised and semi-stabilised sulfur ylides to tosyl
protected imines: are they under kinetic or thermodynamic control?
Varinder K. Aggarwal,*a Jonathan P. H. Charmant,a Cinzia Ciampi,b Jonathan M. Hornby,b
Christopher J. O’Brien,b George Hynda and Richard Parsonsb
1
a School of Chemistry, Cantocks Close, Bristol University, Bristol, UK BS8 1TS
b Department of Chemistry, University of Sheffield, Sheffield, UK S3 7HF
Received (in Cambridge, UK) 10th August 2001, Accepted 5th October 2001
First published as an Advance Article on the web 7th November 2001
Sulfur ylides react with imines, via betaines, to give aziridines. We sought to determine whether betaine formation was
reversible in reactions of benzyl-, amide- and ester-stabilised ylides by carrying out cross-over experiments. Thus, the
intermediate betaines were generated independently from the corresponding sulfonium salt in the presence of a more
reactive imine (p-nitrobenzaldimine). It was found that no incorporation of the more reactive imine was observed in
reactions with the benzyl-stabilised ylide, whilst >80% incorporation of the p-nitrobenzaldimine was observed from
the ester- and amide-stabilised ylides. These results indicate that benzyl-stabilised ylides react irreversibly with imines
but ester- and amide-stabilised react reversibly. Thus, the stereocontrolling step of the process is dependent on the
type of ylide employed and the results are used to account for the different diastereoselectivities observed with
the different ylides.
Introduction
The reaction of sulfur ylides with imines to furnish aziridines
provides a complementary process to alkene aziridination.1 We
have shown that this process can be rendered both catalytic and
asymmetric through a metal-mediated reaction of a chiral sul-
fide with a diazo compound (Scheme 1).2 We have also shown
Scheme 2 Catalytic asymmetric aziridination of imines with in situ
generation of diazo compounds.
ylides gave either no selectivity or predominantly trans substi-
tuted aziridines? In this paper we address these questions and
provide an understanding of the relevant factors governing the
origin of the diastereocontrol in reactions of sulfonium ylides
with imines.
Scheme 1 Catalytic asymmetric aziridination of imines.
that the diazo compounds can be generated in situ,3 thus lead-
ing to a practical catalytic asymmetric process for converting
imines into aziridines (Scheme 2). Although high enantioselec-
tivity has been achieved in this3 and the related epoxidation
process,4,5 the diastereoselectivity observed in imine aziridin-
ation is much poorer. Furthermore, the use of more stabilised
ylides (generated again from the corresponding diazo com-
pounds) led to even lower diastereoselectivity and eventually to
a reversal in favour of the cis isomer with ester-stabilised ylides.
Representative data from our lab and others6 are provided in
Table 1 and Scheme 3.
The variation in diastereoselectivity with different sulfur
ylides was intriguing and these results raised several questions.
Why was the diastereoselectivity much poorer in reactions of
benzyl-stabilised ylides with imines compared to the same reac-
tions with aldehydes (Scheme 4)? Why did ester-stabilised ylides
lead to predominantly cis aziridines whereas less stabilised
Reactions of sulfur ylides with aldehydes are believed to
occur via intermediate betaines, which subsequently ring close
to give epoxides. Indeed, we7 and others8,9 have shown that the
intermediate betaines can be accessed by deprotonation of
β-hydroxy sulfonium salts and this also leads to epoxides. In
addition, the mechanism has been further substantiated
through recent molecular modelling calculations, from which it
was found that end-on approach of sulfur ylides to aldehydes
was the favoured mode of addition and betaines were identified
as intermediates along the reaction pathway.10
In analysing the origin of diastereocontrol in epoxidation
reactions,7 we prepared the syn and anti β-hydroxy sulfonium
salts (3a and 3b, Scheme 5) of benzyl-stabilised ylides and gen-
erated the corresponding ylides by treatment with base. This
was carried out in the presence of a more reactive aldehyde
DOI: 10.1039/b107275g
J. Chem. Soc., Perkin Trans. 1, 2001, 3159–3166
This journal is © The Royal Society of Chemistry 2001
3159

View Article Online
Table 1 Reaction of N-tosylbenzaldimine with a variety of sulfonium
ylides
Entry
R1
R2
Ratio (trans : cis)b
Yield (%)
1a
2a
3a
4c
5d
6e
7e
8e
9e
10e
Ph
Ph
Ph
CONEt2
CONEt2
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
p-MeOC6H4
Ph
p-NO2C6H4
Ph
3 : 1
3 : 1
3 : 1
2 : 1
1 : 3
2 : 3
1 : 3
1 : 5
1 : 12
1 : 11
96
82
43
84
88
62
83
41
54
76
Ph
p-MeOC6H4
Ph
p-ClC6H4
p-NO2C6H4
C6H11
a All reactions were performed with 1 mol% Rh2(OAc)4, 1 equiv. of
dimethyl sulfide, 1.5 equiv. phenyldiazomethane in CH2Cl2 at RT.
b Diastereomeric ratios were determined by 1H NMR of the crude reac-
tion mixture. c Performed in THF at 60 ЊC with 1.5 equiv. of N,N-
diethyl diazoacetamide, 1 equiv. of tetrahydrothiophene and 1 mol%
Rh2(OAc)4. d Treatment of sulfonium salt with base in the presence
of imine (ref. 6). e All reactions were performed with 1 mol% Rh2(OAc)4,
1 equiv. of tetrahydrothiophene, 1.5 equiv. of ethyl diazoacetate in THF
at 60 ЊC.
Fig. 1 Targeted sulfonium salts for mechanistic studies.
In addition to answering the questions raised above, the ques-
tion of whether betaine formation is reversible has important
consequences in terms of enantiocontrol. If betaine formation
is irreversible, non-bonded interactions in the transition state
leading to its formation will be responsible for the stereocontrol
(diastereo- and enantio-control). However, if betaine formation
is reversible, it will be non-bonded interactions in the transition
state of the ring-closing step that will control the relative and
absolute stereochemistry of the aziridine product.
As we have employed benzyl-, amide- and ester-stabilised
ylides in aziridination processes, we decided to test whether the
reaction of each of these ylides with N-tosyl imines leading to
intermediate betaines was reversible and so we required the
corresponding diastereomerically pure sulfonium salts 4a, 4b,
5a, 5b, 6a and 6b (Fig. 1).
Scheme 3
Synthesis of sulfides and sulfonium salts
The syn and anti sulfonium salts 4a and 4b were prepared as
shown in Scheme 6. Reaction of the N–Ts imine derived from
Scheme 4 Comparison of diastereoselectivity in reactions of sulfur
ylides with aldehydes and imines.
Scheme 6 Reagents and conditions: i, Rh2(OAc)4 (1 mol%), BnEt3-
NϩClϪ (10 mol%), 1,4-dioxane, tetrahydrothiophene (20 mol%), 40 ЊC,
84% (3 : 1, trans : cis); ii, NaSMe, EtOH, reflux 6 h, 56%; iii, NaSMe,
EtOH, reflux 1 h, 97%.
Scheme 5 Cross-over reactions with epoxides.
(cross-over experiment) to determine whether betaine form-
ation was reversible. As no incorporation of the more reactive
aldehydes was observed starting from the anti β-hydroxy sulfon-
ium salt, we concluded that formation of the anti betaine was
irreversible. In contrast, partial incorporation of the more
reactive aldehyde was observed when we started from the syn
β-hydroxy sulfonium salt indicating that syn betaine formation
was partially reversible. The degree of reversibility was depend-
ent on the R group: aromatic groups (leading back to aromatic
aldehydes) showed a greater tendency to reversibility than
aliphatic groups. The same types of cross-over experiment have
also been conducted to analyse whether betaine/oxaphos-
phetane formation in the Wittig reaction is reversible.11 We
therefore decided to employ the same methods to determine
whether ylide additions to imines leading to intermediate
betaines were reversible, as this could influence the diastereo-
selectivity of the product.
benzaldehyde with the benzaldehyde tosylhydrazone salt in the
presence of tetrahydrothiophene and Rh2(OAc)4 gave a 3 : 1
separable mixture of aziridines in good yield. Although these
cis and trans aziridines could be made stereospecifically from
reaction of the cis and trans stilbenes with PhINTs,12 we found
the above method to be more convenient. Ring opening of the
aziridines was effected with NaSMe and interestingly was found
to be substantially faster and more efficient for the trans isomer
compared to the cis isomer. We believe that this is due to the
greater relief of ring strain inherent in the trans isomer (vide
infra). Alkylation of the sulfide 8b was carried out with MeI
and the salt 4b was partially characterised although we found it
more convenient to prepare the salts and use them directly in
subsequent studies.
The syn and anti sulfides 11a and 11b were prepared directly
by an aldol reaction and furnished a separable 2 : 1 mixture of
isomers (Scheme 7). The relative stereochemistry of the syn 11a
3160
J. Chem. Soc., Perkin Trans. 1, 2001, 3159–3166

View Article Online
Scheme 7 Formation of the ester aldol adduct.
and anti 11b isomers was determined by X-ray crystallography
(Fig. 2). Alkylation of sulfides 11a and 11b with Meerwein’s
reagent gave the sulfonium salts 5a and 5b. As these salts were
hygroscopic, they were not isolated but generated immediately
before use in the cross-over experiments.
The same strategy was applied to the synthesis of syn and
anti sulfides 13a and 13b using an aldol reaction with amide
1213 (Scheme 8). In the aldol process involving the amide we
Scheme 8 Formation of the amide aldol adduct.
found that LiCl was essential to obtain reasonable yields of the
aldol adducts.14 However, in this case we were unable to separ-
ate the syn and anti isomers. We therefore attempted to directly
convert the syn and anti ester aldols adducts 11a and 11b into
the corresponding amide aldols 13a and 13b using HNEt2–
Me3Al (Scheme 9).15 However both the syn and anti ester aldols
Fig. 2 Solid state molecular structures of 11a and 11b showing one of
the four crystallographically independent molecules in each unit cell.
Ellipsoids are drawn at the 50% probability level.
Et2NAlMe2. This procedure thereby provides ready access to the
syn amide 13a but the pure anti isomer 13b remained elusive.
Nevertheless from cross-over reactions involving pure syn 13a
and a 1 : 1 mixture of syn and anti isomers, it would be possible
to calculate the results of the pure anti isomer.
Results of cross-over experiments
Treatment of syn and anti sulfonium salts 4a and 4b (formed
from 8a and 8b by methylation with MeI) resulted in stereo-
specific ring closure to give exclusively cis and trans aziridines,
respectively (Scheme 10). The fact that no scrambling of stereo-
chemistry was observed indicated that no base-catalysed epi-
merisation occurred and that the intermediate betaines did not
revert back to the corresponding ylide and imine. This was con-
firmed by carrying out the same reaction in the presence of
Scheme 9 Reaction of esters 11a and 11b with Me3Al–HNEt2.
11a and 11b gave the same syn amide aldol 13a in moderate
yield; none of the anti amide aldol was obtained. As the reac-
tion mixture also contained ester 10 and imine 9 but none of the
amide 12, it is reasonable to conclude that the two ester aldols
11a and 11b equilibrate via the aluminium enolate 14 to give the
syn aldol 11a (Scheme 9). Attempts to generate the aluminium
enolate 14 directly from 10 using various aluminium-based
reagents to prove this mechanism were unsuccessful. This aldol
is subsequently converted into the syn amide 13a using
p-NO C H CH᎐NTs, but as expected, no incorporation of the
᎐
2
6
4
more reactive imine was observed. This indicated that reaction
of benzyl-stabilised ylides with PhCH᎐NTs leading to the syn
᎐
and anti betaine intermediates was irreversible (Scheme 11). The
diastereoselectivity of the reaction is therefore set up in the
betaine forming step. The preferred formation of the trans
aziridine can be readily accounted for by analysing the relevant
J. Chem. Soc., Perkin Trans. 1, 2001, 3159–3166
3161

View Article Online
Table 2 Cross-over experiments with ester aldol adducts
Entry
Solvent
Aldol adduct
14 (%)b
15 (%)b
Ratioc (14 : 15)
16 (%)b
17 (%)b
Ratioc (16 : 17)
1
2
3
4
5
6
7
8
Dioxane
Dioxane
THF
11a
11b
11a
11b
11a
11b
11a
11b
4.0
4.0
2.5
5.2
10.8
4.7
4.0
2.8
2.0
1.8
1.3
2.4
4.7
2.1
2
2 : 1
2 : 1
2 : 1
2 : 1
2 : 1
2 : 1
2 : 1
2 : 1
29.0
25.0
27.1
28.0
20.0
17.0
24.5
28.9
5.4
5.7
4.5
3.2
—
—
4.5
3.2
5 : 1
5 : 1
5 : 1
5 : 1
—
—
5 : 1
5 : 1
THF
CH3CN
CH3CN
CH2Cl2
CH2Cl2
1.5
a All sulfonium salts were formed in CH2Cl2 at RT using 1 equiv. of Me3OBF4. The CH2Cl2 was then removed under high vacuum and the solvent of
choice added, followed by 1 equiv. of imine and KOtBu. b Yields were determined using 1,3-benzodioxole as a 1H NMR internal standard.
c Diastereomeric ratios were determined by 1H NMR.
Fig. 3 End-on addition of sulfur ylide to imine.
Scheme 10 Ring closure experiments with semi-stabilised sulfur ylide
precursors.
Scheme 12 Cross-over experiments with ester-stabilised sulfur ylide
precursors.
Scheme 11 Mechanism for benzyl-stabilised sulfur ylide aziridination.
transition states (Fig. 3). Assuming an end-on approach of the
sulfur ylide to the imine, this will give rise to the two transition
states A and B, which lead to the trans and cis aziridines,
respectively. Clearly transition state A is less sterically encum-
bered than transition state B, which accounts for the preferred
formation of the trans aziridine.
Scheme 13 Mechanism for ester-stabilised sulfur ylide aziridination.
event, treatment of the sulfide 11a with Meerwein’s reagent
As syn and anti sulfonium salts 4a/4b and 5a/5b (ester and
amide) could more readily equilibrate through base-catalysed
epimerisation, we decided to carry out the cross-over experi-
ments on these pairs of substrates rather than to determine
whether ring closure was stereospecific. The cross-over experi-
ments were conducted in several different solvents (Scheme 12,
Table 2 and Scheme 14, Table 3) to determine whether solvent
polarity affected the outcome of the overall process. In the
followed by base in the presence of p-NO C H CH᎐NTs led to
᎐
2 6 4
a 5 : 1 mixture of cis : trans aziridines derived from the
p-nitrobenzaldimine 16 and 17 with only a small amount of
aziridines 14 and 15 derived from direct ring closure. As both
syn and anti sulfonium salts 5a and 5b (derived from 11a and
11b, respectively) essentially gave the same results, this implied
that rapid base-catalysed epimerisation of the two diastereo-
isomers 5a and 5b occurred. The cross-over experiments were
3162
J. Chem. Soc., Perkin Trans. 1, 2001, 3159–3166

View Article Online
Table 3 Cross-over experiments with amide aldol adducts
Entry
Solvent
Aldol adduct
18 (%)b
19 (%)b
Ratioc (18 : 19)
20 (%)b
21 (%)b
Ratioc (20 : 21)
1
2
3
4
5
6
7
8
Dioxane
Dioxane
THF
13a
Trace
3.4
3.1
2.8
2.2
7.2
6.8
8.6
1.3
3.8
4.5
4.1
7.0
10.9
20.3
18.3
—
1 : 1
1 : 1.5
1 : 1.5
1 : 3
1 : 1.5
1 : 3
16.6
35.6
13.4
28.9
17.8
30.4
16.0
23.7
8.3
22.7
7.8
17.4
20.6
35.3
18.2
25.4
2 : 1
1.6 : 1
1.7 : 1
1.7 : 1
1 : 1
1 : 1
1 : 1
13a : 13b (1 : 1)
13a
THF
13a : 13b (1 : 1)
13a
13a : 13b (1 : 1)
13a
13a : 13b (1 : 1)
CH3CN
CH3CN
CH2Cl2
CH2Cl2
1 : 2
1 : 1
a All sulfonium salts were formed in CH2Cl2 at rt using 1 equiv. of Me3OBF4. The CH2Cl2 was then removed under high vacuum and the solvent of
choice added, followed by 1 equiv. of imine and KOtBu. b Yields were determined using 1,3-benzodioxole as a 1H NMR internal standard.
c Diastereomeric ratios were determined by 1H NMR.
ketones19 give cis aziridines predominantly. The thermo-
dynamic preference for the cis aziridine can be understood
in terms of steric hindrance. The largest group on the three
membered ring is the Ts group and this will prefer to be anti to
the other substituents to minimise 1,2-steric interactions. Thus,
the remaining two groups are cis to each other. We can now
understand why the trans aziridine reacted more rapidly with
NaSMe (Scheme 6) than the cis isomer: opening of the trans
aziridine resulted in greater relief of ring strain than opening of
the corresponding cis isomer.
The amide sulfonium salts behaved in a similar way to the
ester derivatives (Scheme 14, Table 3). As with the esters, both
the syn and the 1 : 1 syn : anti mixture of amide sulfonium salts
6a and 6b gave similar results indicating that rapid base-
catalysed epimerisation occurred. As with the ester-stabilised
ylides, the major aziridine product incorporated the p-nitro
benzaldimine indicating that the betaine fragmented to the
ylide and imine more rapidly than it underwent direct ring
closure. Thus, as with ester-stabilised ylides, ring closure
of the betaine rather than betaine formation is again the
step which controls the stereochemistry of aziridinations with
amide-stabilised ylides (Scheme 15).
Scheme 14 Cross-over experiments with amide-stabilised sulfur ylide
precursors.
carried out in a range of solvents but the outcome of the pro-
cess was similar in each case (Scheme 12, Table 2). Thus in all
the solvents, following betaine formation, reversion to the imine
and ylide was faster than direct ring closure and the ylide was
subsequently trapped by the more reactive p-nitrobenz-
aldimine.16 Reversion of the betaine to give the imine and ylide
is favoured in this case because the ylide is stabilised by the ester
group. Indeed this ylide can, unlike the benzyl- or amide-
stabilised ylide, be isolated and stored.17 There is considerable
evidence for the high thermodynamic stability and low reac-
tivity of the ester-stabilised ylide: Ratts17 found that this ylide
does not react with simple aldehydes, only 1,2-dicarbonyl com-
pounds, and indeed Dai was unable to couple the same ylide
with N-tosyl imines.6
As betaine formation is reversible, it is the step involving ring
closure of the betaine which controls the stereochemistry of the
product aziridine (Scheme 13). The cis diastereoselectivity
observed indicates that ring closure of the syn betaine is more
rapid than the ring closure of the anti betaine. The relative rates
of ring closure are determined by non-bonded interactions in
the transition state. In related epoxidation, the trans isomer is
favoured as the product-like transition state for ring closure of
the anti betaine is less hindered than the transition state leading
to the cis isomer. As the same factors will also be relevant to
ring closure of the syn and anti betaines, 5a and 5b, the pre-
ferred formation of the cis aziridine implies that this isomer
suffers fewer steric interactions than the trans isomer.
Scheme 15 Mechanism for amide-stabilised sulfur ylide aziridination.
The lower diastereoselectivity observed with the amide-
stabilised ylides compared to reactions of ester-stabilised ylides
is probably due to competing steric interactions. As in the case
of ester-stabilised ylides, the two substituents on the aziridine
will prefer to occupy a position anti to the bulky tosyl group.
However, in this case the bulky amide group will prefer to
occupy a position anti to the other two substituents. These two
competing steric interactions lead to low diastereocontrol.
Conclusions
Indeed, it has been shown experimentally that cis aziridines
are thermodynamically more stable then the corresponding
trans isomers: palladium-catalysed isomerisation of unsatur-
ated aziridines18 and base-catalysed isomerisation of aziridinyl
The addition of benzyl-stabilised sulfur ylides to imines is an
irreversible process and therefore the selectivity is determined
by the relative rates of formation of the anti and syn betaines.
J. Chem. Soc., Perkin Trans. 1, 2001, 3159–3166
3163

View Article Online
This is in contrast to the epoxidation process where formation
of the syn-betaine is partially reversible.7 Amide- and ester-
stabilised ylides react reversibly with imines to give the inter-
mediate betaines which subsequently ring close to give the
aziridines. In these cases ring closure of the betaine is therefore
the stereocontrolling step.
The change in stereocontrolling step according to the type of
ylide employed has an important consequence in the design of
chiral sulfonium ylides for asymmetric aziridination. For semi-
stabilised (benzyl) ylides, interactions in the transition state
leading to betaine formation control the stereochemistry of the
aziridine whilst for stabilised (ester/amide) ylides it is inter-
actions in the transition state for ring closure of the betaine
that will influence the stereochemistry. It is therefore likely that
different sulfides will be required to achieve high stereocontrol
for the different classes of ylides.
stirred at rt for 60 h. The excess methyl iodide was removed
in vacuo and the solid residue triturated with diethyl ether
(10 cm3). The resultant solid 4a was dissolved in dichlorometh-
ane (3 cm3) and 50% aqueous NaOH (0.5 cm3) and the solution
stirred at rt for 20 h before being diluted with dichloromethane
(7 cm3). After being washed with water (5 cm3), the organic
fraction was dried over MgSO4 and evaporated in vacuo to give
the title compound as a white solid (28 mg, 40%). Data (1H
NMR and 13C NMR) were identical to those reported.19,20
anti-[1-Ethoxycarbonyl-2-phenyl-2-(toluene-4-sulfonamido)-
ethyl]dimethylsulfonium tetrafluoroborate 5b
Ethyl anti-3-(4-methylphenylsulfonylamino)-2-(methylthio)-3-
phenylpropanoate (0.12 g, 0.305 mmol) and trimethyloxonium
tetrafluoroborate (0.045 g, 0.305 mmol) were stirred overnight
in dichloromethane (4 cm3) at rt. The resultant solution was
frozen in liquid nitrogen and the solvent removed under high
vacuum, to yield a very air-sensitive pale brown solid (0.15 g,
assuming 100% yield), which was not isolated (pure anti by
Experimental
1
General
crude H NMR); δH(250 MHz; DMSO) 0.87 (3H, t, J 7.0 Hz,
OCH2CH3), 2.25 (3H, s, ArCH3), 3.08 (3H, s, SCH3), 3.11 (3H,
s, SCH3), 4.00 (2H, q, J 7.0 Hz, OCH2CH3), 4.72 (1H, m,
NCH), 5.01 [1H, m, CHS(CH3)2], 7.05–7.17 (7H, m, Ar), 7.39–
7.47 (2H, m, Ar), 9.08 (1H, d, J 3.0 Hz, NH); δF(235.5 MHz;
DMSO) Ϫ148 [BF4Ϫ ϩS(CH3)2].
Infrared spectra were recorded on a Perkin-Elmer Spectrum
One FT-IR spectrometer and only selected absorbancies (νmax
)
are reported. Mass spectra (m/z) were recorded on a Micromass
Analytical Autospec spectrometer. Microanalyses were per-
formed using a Carlo Erba EA1108. Melting points were
determined on a Kofler hot stage. Nuclear magnetic resonance
(NMR) were recorded at the field strength shown and on a
JEOL Delta GX270, GX400 eclipse400 or Alpha500 instru-
ment. Chemical shifts (δH and δC) are quoted in parts per
million (ppm), referenced to TMS or the appropriate solvent
peak. TLC was performed on aluminium backed silica plates
(60 F254) which were developed using standard visualising
agents. Flash chromatography was performed on silica gel
(Merck Kieselgel 60 F254 230–400 mesh). All commercially
available reagents and solvents were purified and dried accord-
ing to standard procedures and all experiments were carried
out under an inert nitrogen atmosphere. Literature procedures
were used to prepare the following compounds: N-benzyl-
idenetoluene-p-sulfonamide,2b trans-N-(p-tolylsulfonyl)-2,3-di-
phenylaziridine,3 cis-N-(p-tolylsulfonyl)-2,3-diphenylaziridine,3
N,N-diethyl-2-methylsulfanylacetamide,13 and trimethyloxonium
tetrafluoroborate.20
syn-N-(1,2-Diphenyl-1-methylthioethyl)toluenesulfonamide 8a
NaSMe (60 mg, 0.86 mmol) was added to a stirred suspension
of cis-N-(p-tolylsulfonyl)-2,3-diphenylaziridine (150 mg, 0.43
mmol) in ethanol (5 cm3). The mixture was heated at reflux for
6 h before being allowed to cool to rt. The mixture was then
diluted with water (5 cm3) and extracted with ethyl acetate (4 ×
10 cm3). The combined organic fractions were dried over
MgSO4 and evaporated in vacuo to furnish a yellow oil, which
solidified on standing. The residue was purified by column
chromatography eluting with 85 : 15, petroleum ether–EtOAc
to give the required sulfide 8a (96 mg, 56%) as an off-white solid
mp 133.0–134.5 ЊC (petroleum ether–EtOAc); νmax(Attenuated
total reflection, ATR)/cmϪ1 3268, 3062, 3029, 2917, and 2891;
δH(400 MHz; CDCl3) 1.82 (3H, s, SCH3), 2.32 (3H, s, ArCH3),
3.94 (1H, d, J 8.4 Hz, SCHPh), 4.60 (1H, dd, J 5.1 and 8.4 Hz,
NCHPh), 5.62 (1H, d, J 5.1 Hz, NH), 7.15–6.85 (12H, m, Ar),
7.46 (2H, d, J 9 Hz, Ar); δC(100 MHz; CDC13) 14.5, 21.5, 58.1,
61.5, 126.9, 127.3, 127.4, 127.6, 127.7, 128.3, 128.5, 128.9,
136.8, 137.2, 137.4, 142.7; m/z (EI) 349 (Mϩ Ϫ SCH3, 54%), 260
(100), 194 (33), 155 (61) (Found: Mϩ Ϫ SCH3, 350.1216.
C21H20NO2S requires Mϩ, 350.1215).
anti-[1,2-Diphenyl-2-(toluene-4-sulfonamido)ethyl]dimethyl-
sulfonium iodide 4b
anti-N-(1,2-Diphenyl-l-methylthioethyl)toluenesulfonamide
(100 mg, 0.25 mmol) was dissolved in methyl iodide (1 cm3)
and the mixture stirred at rt for 80 h. Excess methyl iodide
was removed in vacuo and the residue solid was triturated with
diethyl ether (5 × 15 cm3) to afford sulfonium salt 4b as an off-
white solid (85 mg, 63%), δH(250 MHz; d6DMSO) 2.25 (3H, s,
ArCH3), 2.75 [6H, s, S(CH3)2], 5.16 (1H, d, J 8.0 Hz, CH), 5.32
(1H, m, CH), 7.47–6.97 (14H, m, Ar), 8.73 (1H, d, J 10 Hz,
NH).
anti-N-(1,2-Diphenyl-1-methylthioethyl)toluene-4-sulfonamide
8b
NaSMe (32 mg, 0.45 mmol) was added to a stirred suspension
of trans-N-(p-tolylsulfonyl)-2,3-diphenylaziridine (80 mg, 0.23
mmol) in EtOH (3 cm3). The mixture was heated to reflux for
45 min before being allowed to cool to rt. The resultant suspen-
sion was diluted with water (5 cm3) and the EtOH was removed
in vacuo. The remaining solution was extracted with ethyl
acetate (5 × 5 cm3) and the combined organic extracts dried
over MgSO4. The solvent was removed in vacuo to furnish the
title compound 8b (90 mg, 97%) as an off-white solid, Rf 0.33
(80 : 20 petroleum ether–ethyl acetate); mp 139–141 ЊC;
νmax(ATR)/cmϪ1 3268, 3062, 3029, 2917, and 2891; δH-
(400 MHz; CDCl3) 1.69 (3H, s, SCH3), 2.30 (3H, s, ArCH3),
3.92 (1H, d, J 7.0 Hz, SCHPh), 4.55 (1H, t, J 7.0 Hz, NCHPh),
5.34 (1H, d, J 7.0 Hz, NH), 6.86 (2H, m, Ar), 6.96–7.20 (10H,
m, Ar), 7.36 (2H, d, J 8 Hz, Ar); δC(100 MHz; CDCl3) 15.2,
21.5, 58.6, 61.9, 127.2, 127.8, 127.9, 128.0, 128.6, 128.8, 129.3,
137.0, 137.4, 137.5, 142.9; m/z (EI) 349 (Mϩ ϪSCH3, 54%), 260
(100), 194 (12), 155 (63) (Found: Mϩ Ϫ SCH3, 350.1207.
C21H20NO2S requires Mϩ, 350.1215).
trans-N-( p-Tolylsulfonyl)-2,3-diphenylaziridine (Scheme 10)
To a stirred suspension of 4b (60 mg, 0.11 mmol) in dichloro-
methane (2 cm3) was added 50% aqueous NaOH (0.5 cm3). The
mixture was stirred at rt before being diluted with dichloro-
methane (3 cm3) and partitioned with water (5 cm3). The
organic fraction was dried over MgSO4 and the solvent evapor-
ated in vacuo to give the title compound as a white solid (40 mg,
100%). Data (1H NMR and 13C NMR) were identical to those
reported.21,22
cis-N-( p-Tolylsulfonyl)-2,3-diphenylaziridine (Scheme 10)
syn-N-(l,2-Diphenyl-l-methylthioethyl)toluenesulfonamide (79
mg, 0.2 mmol) was dissolved in methyl iodide (1 cm3) and
3164
J. Chem. Soc., Perkin Trans. 1, 2001, 3159–3166

View Article Online
2-Methylsulfanyl-3-phenyl-3-(toluene-4-sulfonamido)propionic
acid ethyl ester syn-11a and anti-11b
resultant mixture was stirred for 1 h. After this time the reaction
mixture was cooled to Ϫ78 ЊC. N,N-Diethyl-2-methylsulfanyl-
acetamide(0.5g, 3.1mmol)dissolvedinTHF(10cm3)wasadded
dropwise to the cooled solution over a period of 30 min and the
mixture was stirred for 1 h. Finally N-benzylidenetoluene-p-
sulfonamide (0.96 g, 3.7 mmol) dissolved in THF (6 cm3) was
added dropwise to the solution (at Ϫ78 ЊC) and stirring was
continued for 40 min. 2 M ethanolic HCl (10 cm3) was added
with rapid stirring, ether (30 cm3) was then added and the mix-
ture stirred for 10 min. The vessel and its contents were then
allowed to warm to rt. The organic phase was washed with
water (40 cm3), saturated aqueous NaHCO3 (2 × 10 cm3) and
brine (20 cm3) until neutral. The organic phase was then dried
with MgSO4 and the solvent removed in vacuo to yield a yellow
To diisopropylamine (0.98 cm3, 6.97 mmol) in THF (10 cm3)
was added n-butyllithium (5.02 cm3, 1.3 M, 6.53 mmol) and the
resultant mixture was stirred for 1 h. After this time the reaction
mixture was cooled to Ϫ78 ЊC. Ethyl (methylthio)acetate
(0.41 cm3, 3.1 mmol) dissolved in THF (10 cm3) was added
dropwise to the cooled solution over a period of 30 min, and
stirred for 1 h. Finally N-benzylidenetoluene-p-sulfonamide
(0.96 g, 3.7 mmol) dissolved in THF (6 cm3) was added drop-
wise to the solution at Ϫ78 ЊC and stirred for 40 min. 2 M
ethanolic HCl (10 cm3) was added with rapid stirring, diethyl
ether (30 cm3) was then added and the mixture stirred for
10 min. The vessel and its contents were allowed to warm to rt.
The organic phase was washed with water (40 cm3), saturated
aqueous NaHCO3 (2 × 10 cm3) and brine (20 cm3). The organic
phase was then dried with MgSO4 and the solvent removed
in vacuo to yield a yellow solid (2 : 1 anti : syn, ratio by crude
1H NMR). The solid was purified by flash chromatography (85
: 15, petroleum ether–EtOAc), giving two separate colour-
less solid diastereoisomers: anti: [0.58 g, 40%, Rf 0.43 (80 : 20,
petroleum ether–EtOAc)] as a white solid: mp 127–128 ЊC (from
petroleum ether–EtOAc) (Found: C, 58.22; H, 5.76; N, 3.56.
C19H23NO4S2 requires C, 57.99; H, 5.89; N, 3.56%); νmax(ATR)/
cmϪ1 3257, 2925, 1728, 1330 and 1154; δH(400 MHz; CDCl3)
0.99 (3H, t, J 7.0 Hz, OCH2CH3), 2.02 (3H, s, SCH3), 2.36 (3H,
s, ArCH3), 3.39 (1H, d, J 10.2 Hz, CHSCH3), 3.93 (2H, q, J 7.0
Hz, OCH2CH3), 4.54 (1H, dd, J 10.2 and 3.3 Hz, NCH), 5.70
(1H, d, J 3.3 Hz, NH), 7.05–7.13 (7H, m, Ar), 7.42–7.51 (2H,
m, Ar); δC(100 MHz; CDCl3) 12.1, 13.8, 21.5, 52.5, 55.8, 61.4,
127.3, 128.1, 128.2, 128.3, 128.4, 129.3, 137.1, 137.2, 143.2,
168.7; m/z (EI) 393 (Mϩ, 100%); syn: [0.30 g, 21%, Rf 0.40
(80 : 20, petroleum ether–EtOAc)]; as a white solid: mp 134–
136 ЊC (from petroleum ether–EtOAc) (Found C, 58.14; H,
5.96; N, 3.61. C19H23NO4S2 requires C, 57.99; H, 5.89; N,
3.56%); νmax(ATR)/cmϪ1 3257, 3190, 1712, 1333 and 1156;
δH(400 MHz; CDCl3) 1.11 (3H, t, J 7.0 Hz, OCH2CH3), 2.09
(3H, s, SCH3), 2.33 (3H, s, ArCH3), 3.47 (1H, d, J 6.0 Hz,
CHSCH3), 4.01–4.15 (2H, m, OCH2CH3), 4.84 (1H, dd, J 6.0
and 9.0 Hz, NCH), 6.09 (1H, d, J 9.0 Hz, NH), 7.06–7.17 (7H,
m, Ar), 7.52–7.56 (2H, m, Ar); δC(100 MHz CDCl3) 14.2, 15.6,
21.4, 53.9, 59.0, 61.6, 126.7, 127.2, 127.9, 128.5, 129.2, 137.8,
138.0, 139.9, 143.0, 170.8; m/z (EI) 393 (Mϩ, 100%).
1
solid (1 : 1 anti : syn ratio by crude H NMR). The solid was
purified by flash chromatography (Rf 0.41, 60 : 40, petroleum
ether–EtOAc), and by recrystallisation, yielding a white solid
containing both diastereoisomers, which could not be separated
(0.93 g, 72%), as a white powder; mp 149–150 ЊC (from petrol-
eum ether–EtOAc of 1 : 1 syn : anti); νmax(ATR)/cmϪ1 3355,
3261, 2981, 1616, and 1614; δH(400 MHz; CDCl3) 0.62syn (3H, t,
J 7.0 Hz, NCH2CH3), 0.72anti (6H, m, 2 × NCH2CH3), 0.89syn
(3H, t, J 7.0 Hz, NCH2CH3), 1.77anti (3H, s, SCH3), 2.07syn (3H,
s, SCH3), 2.27anti (1H, s, ArCH3), 2.33syn (1H, s, ArCH3), 2.80–
3.20anti&syn [8H, m, 2 × N(CH2CH3)2], 3.37anti (1H, d, J 10.2 Hz,
CH3SCH), 3.56syn (1H, d, J 4.0 Hz, CH3SCH), 4.52anti (1H, dd,
J 10.2 and 2.2 Hz, NHCH), 4.74syn (1H, dd, J 8.0 and 4.0 Hz,
NHCH), 5.78anti (1H, d, J 2.2 Hz, NH), 7.03–7.77 (18H, m, Ar);
δC(100 MHz; CDCl3) 10.5, 12.8, 14.1, 14.3, 14.7, 21.4, 21.5,
41.0, 41.1, 42.2, 42.5, 47.8, 48.5, 55.6, 60.2, 126.8, 127.1, 127.5,
127.5, 127.6, 127.8, 128.0, 128.3, 128.4, 129.0, 129.4, 136.5,
138.3, 138.5, 139.2, 142.5, 143.3, 166.2, 168.8; m/z (El) 421 (Mϩ,
10%), 373 (24), 300 (55), 161 (100) (Found: Mϩ, 420.1528.
C21H28N2O3S2 requires Mϩ, 420.1541).
syn-N,N-Diethyl-2-methylsulfanyl-3-phenyl-3-(toluene-4-
sulfonamido)propionamide 13a
A 1.0 M solution of trimethylaluminium in hexane (1.0 cm3,
1.0 mmol) was slowly added at rt to a solution of diethylamine
(103 µl, 1.0 mmol) in 2.5 cm3 of dry dichloromethane. The
mixture was stirred at rt for 15 min and 11a or 11b was added in
one portion (393 mg, 1.0 mmol). The mixture was warmed to
40 ЊC until TLC indicated that the reaction had gone to com-
pletion. The reaction mixture was diluted with dichlorometh-
ane (15 cm3) and quenched with dilute HCl (1 M), separated
and the organic layer dried (MgSO4). The solvent was removed
in vacuo to afford the crude amide as a pale yellow solid. The
solid was purified by flash chromatography (60 : 40, petroleum
ether–EtOAc), and by recrystallisation (hexane–EtOAc), to
yield the product (106 mg, 25%) as a white solid: mp 161.0–
163.5 ЊC (from petroleum ether–EtOAc); νmax(ATR)/cmϪ1 3355,
3259, 2980, 1613, 1324, and 1301; δH(400 MHz; CDCl3) 0.62
(3H, t, J 7.0 Hz, NCH2CH3), 0.89 (3H, t, J 7.0 Hz, NCH2CH3),
2.07 (3H, s, SCH3), 2.33 (3H, s, ArCH3), 2.80 [1H, dq, J 14.0
and 7.0 Hz, N(CHHCH3)2], 3.05 [1H, dq, J 14.0 and 7.0 Hz,
N(CHHCH3)2], 3.22 [2H, q, J 7.0 Hz, N(CH2CH3)2], 3.56 (1H,
d, J 4.0 Hz, CHSCH3), 4.74 (1H, dd, J 7.3 and 4.0 Hz, NCH),
(NH not visible), 7.06–7.17 (7H, m, Ar), 7.52–7.56 (2H, m, Ar);
δC(100 MHz; CDCl3) 12.8, 14.1, 14.7, 21.5, 41.0, 42.5, 48.5,
60.2, 126.5, 126.8, 127.1, 127.6, 128.3, 129.0, 129.8, 138.5,
139.2, 142.4, 143.3, 168.8; m/z (El) 421 (Mϩ, 12%), 373 (27), 300
(54), 161 (100) (Found: Mϩ, 420.1548. C21H28N2O3S2 requires
Mϩ, 420.1541).
Crystal structure of 11a†
Crystal data for C19H23NO4S2; M = 393.50; crystal dimensions
0.30 × 0.20 × 0.20 mm3. Monoclinic, a = 22.104(3), b =
18.785(3), c = 20.095(3) Å, U = 7979.2(19) Å3, Z = 16, Dc = 1.310
Mg mϪ3, space group P2(1)/c, Mo-Kα radiation (λ = 0.71073
Å), µ(Mo-Kα) = 0.290 mmϪ1, F(000) = 3328.
Crystal structure of 11b†
Crystal data for C19H23NO4S2; M = 393.50; crystal dimensions
0.75 × 0.50 × 0.50 mm3. Monoclinic, a = 10.3330(14), b =
28.469(3), c = 26.751(4) Å, U = 7857.9(18) Å3, Z = 16, Dc = 1.330
Mg mϪ3, space group P2(1)/c, Mo-Kα radiation (λ = 0.71073
Å), µ(Mo-Kα) = 0.295 mmϪ1, F(000) = 3328.
N,N-Diethyl-2-methylsulfanyl-3-phenyl-3-(toluene-4-sulfon-
amido)propionamide syn-13a anti-13b
To a mixture of LiCl (0.8 g, 18.6 mmol) and diisopropylamine
(0.98 cm3, 6.97 mmol) in THF (10 cm3) under nitrogen was
added n-butyllithium (5.02 cm3, 1.3 M, 6.53 mmol) and the
General method for crossover experiments
The aldol adduct (11a/11b and 13a/13b) (0.076 mmol) and tri-
methyloxonium tetrafluoroborate (11.3 mg, 0.076 mmol) were
stirred overnight in dichloromethane (1 cm3) at rt under nitro-
gen. The resultant solution was frozen in liquid nitrogen and
† CCDC reference number 169025. See http://www.rsc.org/suppdata/
p1/b1/b107275g/ for crystallographic files in .cif format.
J. Chem. Soc., Perkin Trans. 1, 2001, 3159–3166
3165

View Article Online
4 V. K. Aggarwal, G. J. Ford, S. Fonquerna, H. Adams, R. V. H. Jones
and R. Fieldhouse, J. Am. Chem. Soc., 1998, 120, 8328.
5 V. K. Aggarwal, E. Alonso, G. Hynd, K. M. Lydon, M. J. Palmer,
M. Porcelloni and J. R. Studley, Angew. Chem., Int. Ed., 2001, 40, 1430.
6 Y.-G. Zhou, A.-H. Li, X.-L. Hou and L.-X. Dai, Tetrahedron Lett.,
1997, 38, 7225.
7 V. K. Aggarwal, S. Calamai and J. G. Ford, J. Chem. Soc., Perkin
Trans. 1, 1997, 593.
8 T. Durst, R. Viau, R. Van den Elzen and C. H. Nguyen, J. Chem.
Soc., Chem. Commun., 1971, 1334.
9 M. Yoshimine and M. Hatch, J. Am. Chem. Soc., 1967, 89, 5831.
10 M. K. Lindvall and A. M. P. Koskinen, J. Org. Chem., 1999, 64, 4596.
We have carried out higher level calculations and found that in fact
the syn addition pathway is the lowest energy pathway. Full details
will be reported in due course.
11 B. E. Maryanoff, A. B. Reitz, M. S. Mutter, R. R. Whittle and R. A.
Olofson, J. Am. Chem. Soc., 1986, 108, 7664.
the solvent removed under high vacuum, to yield a very hydro-
scopic pale brown solid (assumed 100% yield). To the sulfonium
salt (0.076 mmol) was added dry solvent (1 cm3) and N-
(nitrobenzylidene)toluene-p-sulfonamide (23 mg, 0.076 mmol)
and the mixture was stirred for 5 min. KOtBu (76 µl, 0.076
mmol, of a 1 M solution in THF) was then added dropwise
with rapid stirring. The resultant solution was stirred for 1 h at
rt then diluted with dichloromethane (5 cm3) and the organic
layer extracted with water (2 × 5 cm3). The organic phase was
then dried with MgSO4 and the solvent removed in vacuo to
yield a pale yellow solid. To the solvent-free mixture was added
1,3-benzodioxole (internal standard, 2.6 µl, 0.0226 mmol) and
the whole reaction mixture was subjected to NMR analysis to
determine yields based on the internal standard.
12 D. A. Evans, M. M. Faul and M. T. Bilodeau, J. Org. Chem., 1991,
56, 6744.
13 K. W. Ratts and A. N. Yao, J. Org. Chem., 1968, 33, 70.
14 LiCl is often employed as an additive in amide aldol reactions to
break up aggregates of the lithium enolate: see A. G. Myers, B. A.
Yang and H. Chen, Org. Synth., 1999, 77, 22.
Acknowledgements
We thank Dr M. Murray for performing the NMR spectro-
scopy and the Universities of Sheffield and Bristol for financial
support.
15 A. Basha, M. Lipton and S. M. Weinreb, Tetrahedron Lett., 1977, 18,
4171.
16 One reviewer suggested that upon scission of the sulfonium salt to
ylide/imine, recombination with themselves rather than cross-
coupling with p-nitrobenzaldimine could occur. We found that
References
1 (a) D. Tanner, Angew. Chem., Int. Ed. Engl., 1994, 33, 599; (b)
H. M. I. Osborn and J. B. Sweeney, Tetrahedron: Asymmetry, 1997,
8, 1693; (c) E. N. Jacobsen, Comprehensive Asymmetric Catalysis II,
eds. E. N. Jacobsen, A. Pfaltz and H. Yamamoto, Springer, Berlin,
1999, p. 607; (d ) R. S. Atkinson, Tetrahedron, 1999, 55, 1519; (e)
A.-H. Li, L.-X. Dai and V. K. Aggarwal, Chem. Rev., 1997, 97, 2341;
( f ) for a recent review on the synthetic utility of chiral aziridines see:
W. McCoull and F. A. Davis, Synthesis, 2000, 1347.
reaction of the same ylide with a mixture of PhCH᎐NTs and
᎐
p-NO2C6H4CH᎐Ts (competition experiments) gave a 15 : 1 ratio in
᎐
favour of reaction with the more reactive p-NO2-derivative. Thus,
upon scission, cross-coupling is the most likely event.
17 K. W. Ratts and A. N. Yao, J. Org. Chem., 1966, 31, 1689.
18 N. Mimura, T. Ibuka, M. Akaji, Y. Miwa, T. Taga, K. Nakai,
H. Tamamura, N. Fujii and Y. Yamamoto, J. Chem. Soc., Chem.
Commun., 1996, 351.
2 (a) V. K. Aggarwal, A. Thompson, R. V. H. Jones and M. C. H.
Standen, J. Org. Chem., 1996, 61, 8368; (b) V. K. Aggarwal,
M. Ferrara, C. J. O’Brien, A. Thompson, R. V. H. Jones and
R. Fieldhouse, J. Chem. Soc., Perkin Trans. 1, 2001, 1635.
3 V. K. Aggarwal, E. Alonso, G. Fang, M. Ferrara, G. Hynd and
M. Porcelloni, Angew. Chem., Int. Ed., 2001, 40, 1433.
19 R. E. Lutz and A. B. Turner, J. Org. Chem., 1968, 33, 516.
20 T. J. Curphey, Org. Synth., 1971, 51, 142.
21 D. A. Evans, M. T. Bilodeau and M. M. Faul, J. Am. Chem. Soc.,
1994, 116, 2742.
22 T. P. Seden and R. W. Turner, J. Chem. Soc., Chem. Commun., 1968,
876.
3166
J. Chem. Soc., Perkin Trans. 1, 2001, 3159–3166