Unexpected formation of dimethylthioketene cycloadducts in the reaction of 1,3-diphenylaziridine-2,2-dicarboxylate with cyclobutanethione derivatives

Article abstract of DOI:10.1002/1522-2675(200209)85:9<2644::AID-HLCA2644>3.0.CO;2-V

The reaction of 2,2,4,4-tetramethyl-3-thioxocyclobutanone (1) with cis-1-alkyl-2,3-diphenylaziridines 5 in boiling toluene yielded the expected trans-configured spirocyclic 1,3-thiazolidines 6 (Scheme 1). Analogously, dimethyl trans-1-(4-methoxyphenyl)aziridine-2,3-dicarboxylate (trans-7) reacted with 1 and the corresponding dithione 2, respectively, to give spirocyclic 1,3-thiazolidine-2,4-dicarboxylates 8 (Scheme 2). However, mixtures of cis- and trans-derivatives were obtained in these cases. Unexpectedly, the reaction of 1 with dimethyl 1,3-diphenylaziridine-2,2-dicarboxylate (11) led to a mixture of the cycloadduct 13 and 5-(isopropylidene)-4-phenyl-1,3-thiazolidine-2,2-dicarboxylate (14), a formal cycloadduct of azomethine ylide 12 with dimethyl-thioketene (Scheme 3). The regioisomeric adduct 16 was obtained from the reaction between 2 and 11. The structures of 6b, cis-8a, cis-8b, 10, and 16 have been established by X-ray crystallography.

Full text of DOI:10.1002/1522-2675(200209)85:9<2644::AID-HLCA2644>3.0.CO;2-V

2644
Helvetica Chimica Acta Vol. 85 (2002)
Unexpected Formation of Dimethylthioketene Cycloadducts in the Reaction
of 1,3-Diphenylaziridine-2,2-dicarboxylate with Cyclobutanethione
Derivatives
by Grzegorz Mloston¬* and Katarzyna Urbaniak1)
¬
¬
¬
¬
University of Ëodz, Narutowicza 68, PL-90136 Ëodz
and
Anthony Linden and Heinz Heimgartner*
Organisch-chemisches Institut der Universit‰t Z¸rich, Winterthurerstrasse 190, CH-8057 Z¸rich
Dedicated to Professor Albert Padwa on the occasion of this 60th birthay
The reaction of 2,2,4,4-tetramethyl-3-thioxocyclobutanone (1) with cis-1-alkyl-2,3-diphenylaziridines 5 in
boiling toluene yielded the expected trans-configured spirocyclic 1,3-thiazolidines 6 (Scheme 1). Analogously,
dimethyl trans-1-(4-methoxyphenyl)aziridine-2,3-dicarboxylate (trans-7) reacted with 1 and the corresponding
dithione 2, respectively, to give spirocyclic 1,3-thiazolidine-2,4-dicarboxylates 8 (Scheme 2). However, mixtures
of cis- and trans-derivatives were obtained in these cases. Unexpectedly, the reaction of 1 with dimethyl 1,3-
diphenylaziridine-2,2-dicarboxylate (11) led to a mixture of the cycloadduct 13 and 5-(isopropylidene)-4-
phenyl-1,3-thiazolidine-2,2-dicarboxylate (14), a formal cycloadduct of azomethine ylide 12 with dimethyl-
thioketene (Scheme 3). The regioisomeric adduct 16 was obtained from the reaction between 2 and 11. The
structures of 6b, cis-8a, cis-8b, 10, and 16 have been established by X-ray crystallography.
Introduction. 1,3-Dipolar cycloadditions with thiocarbonyl compounds have been
elaborated extensively in recent years [1][2]. Due to the convenient access and their
remarkable stability, thiocarbonyl derivatives of 2,2,4,4-tetramethylcyclobutane-1,3-
dione are among the most frequently explored representatives. In the case of the
−monothione× 1, stable [3 ‡ 2] cycloadducts were obtained from reactions with CH2N2
[3], nitrones [4], thiocarbonyl ylides [5][6], thiocarbonyl S-oxides (sulfines) [7][8],
and thiocarbonyl S-sulfides (thiosulfines) [9]. When reactions were carried out with
imidazole N-oxides, which are structural analogues of nitrones, a rearrangement of the
primarily formed cycloadduct led to a S-transfer to the heterocyclic system to give
imidazole-2-thiones [10]. Most of these cycloadditions occurred in a regioselective
manner. Unlike 1, the analogous −dithione× 2 has not found many applications, although
some reactions with nitrones [4] and diazo compounds [11], have also been described.
Azomethine ylides are 1,3-dipoles with great importance for the stereocontrolled
synthesis of five-membered N-heterocycles and pyrroles [12][13]. In comparison with
C,C dipolarophiles, reactions with heterodipolarophiles are less well-known, and,
for CˆS dipolarophiles, only few examples have been reported. However, some
1
¬
¬
Part of the planned Ph.D. thesis of K. U., University of Ëodz.
)

Helvetica Chimica Acta Vol. 85 (2002)
2645
functionalized 1,3-thiazolidines, which are accessible by this reaction, are interesting as
biologically active compounds. The recently published new approaches to the penam
skeleton, based on the cycloaddition of azomethine ylides with CˆS dipolarophiles,
deserve special attention [14].
In our previous work, some reactions of azomethine ylides, generated by different
methods, with 1,3-thiazole-5(4H)-thiones [15][16] and aromatic thioketones [17 19]
have been studied. Thermal reactions of aziridine-2,3-dicarboxylates 3 with thiocar-
bonyl compounds offer a convenient stereocontrolled approach to 1,3-thiazolidine-2,4-
dicarboxylates. Nevertheless, this approach has been explored only to a limited extent,
and thioketones 1 and 2 have not been used so far in reactions with 3. An additional
question emerges in thermal reactions with 1 and 2, as the thermal dissociation of 2
leads to dimethylthioketene [20], which is expected to react easily with 1,3-dipolar
species [21]. However, no reaction of a thioketene with an azomethine ylide has been
reported so far, and penicillin 4, which has a 5-isopropylidene-1,3-thiazolidine moiety,
and, therefore, is a formal cycloadduct of an azomethine ylide and dimethylthioketene,
was synthesized by other methods [22].
The goal of the present work was to study the behavior of 1 and 2 in thermal
reactions with aziridine carboxylates.
Results and Discussion. The only known reaction of an azomethine ylide with 1
was carried out by heating cis-1-methyl-2,3-diphenylaziridine (5a) in toluene [17].
Based on the well-documented conrotatory course of the thermal ring opening of
aziridines [23], the structure of the only product isolated was tentatively formulated as
the trans-6,8-diphenyl derivative 6a (Scheme 1). In an extension of this preliminary
experiment, the analogous reaction with cis-1-isopropyl-2,3-diphenylaziridine 5b2) was
performed, leading to 6b as the sole cycloadduct. Compared with 5a, the reaction with
5b was remarkably slow and accompanied by significant decomposition of the starting
materials. The isolated product 6b showed similar spectroscopic properties as 6a. In the
1H-NMR spectrum, the 4 Me groups of the cyclobutanone moiety show resonances at
1.60, 1.40, 1.17, and 0.67 ppm for 6a, and at 1.66, 1.39, 1.07, and 0.63 ppm for 6b. On the
2
)
The cis-configuration of 5b has been established by X-ray crystallography [24].

2646
Helvetica Chimica Acta Vol. 85 (2002)
other hand, a remarkable difference appears with respect to the chemical shifts of
HÀC(6) and HÀC(8): whereas in 6a both absorb at 4.73ppm [17], two singlets at 5.37
and 5.09 ppm appear in the case of 6b. This difference could indicate that the two
products have different configurations. Finally, an X-ray crystal-structure determina-
tion of 6b was performed and confirmed the proposed trans-substitution in the 1,3-
thiazolidine (Fig. 1).
Scheme 1
Fig. 1. ORTEP Plot [25] of the molecular structure of 6b (50% probability ellipsoids, arbitrary numbering of atoms)
The reaction of 1 with dimethyl trans-1-(4-methoxyphenyl)aziridine-2,3-dicarbox-
ylate (trans-7) in toluene at 1108, as well as in xylene at 1308, afforded mixtures of two
stereoisomeric cycloadducts of type 8a in a ratio of 4 :1 and 1:1, respectively
(Scheme 2). The reactions with trans-7, in contrast to those with 5b, occurred smoothly
and gave the products in good yields.
1
A characteristic difference between cis-8a and trans-8a appears in the H-NMR
spectrum. The major product formed in toluene showed two singlets at 5.35 and

Helvetica Chimica Acta Vol. 85 (2002)
2647
Scheme 2
5.07 ppm (Dd ˆ 0.28), whereas, for the minor product, the signals appear at 5.27 and
5.13ppm ( Dd ˆ 0.14). Crystallization from MeOH gave the major component in pure
form. The X-ray crystal-structure determination of this compound confirmed the
expected cis-configuration (cis-8a, Fig. 2,a).
The minor component trans-8a was separated from the mother liquor by means of
chromatography. Except for the Dd mentioned above, there were only slight
differences in the spectrosopic properties of cis-8a and trans-8a.
The formation of cis-8a is in line with the expected reaction course, i.e., a
conrotatory electrocyclic ring opening of trans-7, leading to the azomethine ylide
(E,E)-9, followed by a suprafacial [3 ‡ 2] cycloaddition with the CˆS group of 1. It is
well known, however, that, at elevated temperatures, (E,E)-9 is in equilibrium with
(E,Z)-9, and both ylides react with dipolarophiles at comparable rates [23][26]. For
this reason, the formation of both stereoisomeric cycloadducts in temperature-
dependent ratios is easy to explain. Attempts to isomerize cis-8a in boiling toluene
showed that, under these conditions, no secondary isomerization takes place.
When the analogous reaction was carried out with 1 and cis-7 in boiling xylene for
3h, trans-8a was obtained as the sole product.
Similarly, a mixture of dithione 2 and trans-7 was heated in toluene and in xylene,
and, in analogy to the experiments with 1, mixtures of cis-8b and trans-8b (Scheme 2)
were detected in approximately same ratios in the crude mixture. The major product
cis-8b crystallized from MeOH solution as orange needles. The characteristic
absorption at 279.6 ppm (CˆS) in the 13C-NMR spectrum confirmed the presence of
a cyclobutanethione fragment. The cis-configuration has been established by X-ray
crystallography (Fig. 2,b). The crystal structure of the thioxo compound cis-8b is very
similar to that of the O-analogue cis-8a.
In analogy to the reaction with 1, heating 2 with cis-7 led to the formation of trans-
8b, and none of the isomeric cis-1,3-thiazolidine was detected in the mixture.
Earlier attempts to react 1 with dimethyl 1,3-diphenylaziridine-2,2-dicarboxlate
(11), generated in situ from dimethyl 4,5-dihydro-1,5-diphenyl-1H-1,2,3-triazole-4,4-
dicarboxylate (10; Scheme 3) in toluene, were in vain [17]. The structure of 10, which

2648
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 2. ORTEP Plot [25] of the molecular structure of a) cis-8a and b) cis-8b (50% probability ellipsoids,
arbitrary numbering of atoms)
was prepared by treatment of dimethyl benzylidenemalonate with PhN3 [27], has now
been established by X-ray crystallography (Fig. 3). When the reaction with 1 was
carried out in boiling xylene for 60 h, leading to the formation of the intermediate
azomethine ylide 12, no 1 remained according to TLC analysis. Chromatographic
separation of the mixture yielded two products in a ratio that was dependent upon the

Helvetica Chimica Acta Vol. 85 (2002)
2649
Scheme 3
Fig. 3. ORTEP Plot [25] of the molecular structure of 10 (50% probability ellipsoids, arbitrary numbering of atoms)

2650
Helvetica Chimica Acta Vol. 85 (2002)
concentration of the reactants.3) The more polar crystalline product 13 (6%) was the
expected [3 ‡ 2] cycloadduct (cf. [28]).
1
From the less polar fraction, a compound was isolated for which the H-NMR
spectrum showed only two signals for MeC groups at 1.92 and 1.80 ppm, as well as two
signals for MeO at 3.88 and 3.70 (intensity 1:1:1:1). In the 13C-NMR spectrum, four
signals for quaternary C-atoms appeared at 144.4, 142.1, 127.7, and 126.6 ppm. Whereas
the two signals at lower field can be attributed to Cq (arom.), the other two originate
from a novel structural fragment. Both, the 13C-NMR and IR spectra confirmed the
presence of two ester CˆO groups, but there were no signals for a cyclobutanone CˆO
‡
group4). Finally, the ESI-MS revealed [M ‡ Na] at m/z 420, corresponding to the
molecular formula C22H23NO4S. Compared with the cycloadduct 13, the molecular mass
differed by 70 mass units, indicating the loss of dimethylketene. On the basis of these
spectroscopic data and the correct elemental analysis, structure 14 (7%; Scheme 3),
which is a formal cycloadduct of 12 and dimethylthioketene, was proposed. This
conclusion was confirmed by an X-ray crystal-structure determination [28].
The unexpected formation of 14 could be explained by three pathways. The first
interpretation involves a primary [2 ‡ 2] cycloreversion of 1 to equimolar amounts of
dimethylketene and dimethylthioketene. Subsequently, the latter could intercept 12 to
give the product 14. As there are no data available indicating a thermal dissociation of
1, we consider this mechanism rather unlikely. As a second possibility, a thermal
extrusion of dimethylketene from the primarily formed cycloadduct 13 is conceivable.
To test this hypothesis, a sample of 13 was heated in xylene at 1308 for 60 h. After
1
evaporation of the solvent, the H-NMR analysis showed that no 14 was formed, and
only 13, along with small amounts of decompositon products, was present. A third
mechanistic explanation via a multistep process could also be considered. Thioketones
are electron-rich dipolarophiles, and the 1,3-dipole 12 with two electron-withdrawing
groups is considered an electron-poor intermediate. Thus, the fundamental require-
ments for a stepwise mechanism of cycloaddition are fulfilled, i.e., steric hindrance at
one terminus of the dipolar species and a large gap between the frontier orbitals of the
reaction partners [29 31]. However, it is rather difficult to formulate a convincing
zwitterionic intermediate, which, after elimination of dimethylketene, would subse-
quently convert to 14. Finally, the observed influence of the concentration of the
reagents on the ratio 13/14 may suggest that elimination of dimethylketene from 13
could be supported by some components of the reaction mixture5).
According to [20], thermal decomposition of dithione 2 yields dimethylthioketene,
a highly reactive heterocumulene. To test its cycloaddition with 12 to give 14, a mixture
of 10 and 2 in xylene was heated until 12 completely disappeared (TLC). In this case,
the conversion was much faster than with 1 (5 h). The 1H-NMR analysis of the mixture
showed that the expected 14 has not been formed. After chromatographic workup, the
less polar orange substance was identified as the known dithiolactone 15 (Scheme 3; cf.
3
)
)
)
In an experiment with 2.2 mmol of 1 and 2.0 mmol of 10 in 1.5 ml of xylene, the ratio 13/14 was ca. 1:2,
whereas, in 4.5 ml of xylene, it was 2:1.
Typically, cycloclobutanone derivatives show the CˆO absorption at ca. 1780 cmÀ1 (IR) and at 220
4
5
218 ppm (13C-NMR).
At the moment, an unambigous reaction pathway leading to 13 and 14 cannot be formulated. Further
studies are needed for additional arguments.

Helvetica Chimica Acta Vol. 85 (2002)
2651
[32]). The compound from the more polar main fraction showed similar patterns of
1
signals in the H- and 13C-NMR spectra as 14. However, in the IR spectrum, the
fingerprint region of 14 and the new compound were significantly different. The MS
and elemental analysis confirmed that the two compounds have isomeric structures.
This conclusion was unambiguously confirmed by an X-ray crystal-structure determi-
nation, which established structure 16 (Fig. 4) for the product from the reaction
between 11 and 2.
Fig. 4. ORTEP Plot [25] of the molecular structure of 16 (50% probability ellipsoids, arbitrary numbering of
atoms)
The formation of the regioisomer 16 in the reaction of 2 with 12 suggests that, in this
case, a different reaction mechanism must be formulated. It is plausible that a 1,3-
dipolar cycloaddition of 12 with in situ generated dimethylthioketene leads to 16. An
additional evidence for the appearence of dimethylthioketene is the formation of the
isomer 15. This isomerization has been reported earlier (cf. [20]).
In conclusion, we have shown that thermal additions of symmetrically substituted
aziridine 7 with both −monothione× 1 and −dithione× 2 occur smoothly to give mixtures of
diastereoisomers of the expected 1,3-thiazolidine-dicarboxylates. The ratio of the
stereoisomers reflects an isomerization of the initially formed azomethine ylide (E,E)-
9. The reaction of the asymmetrically substituted aziridine dicarboxylate 11 with 1 is
proposed to occur stepwise via a zwitterionic intermediate, which eliminates
dimethylketene to give 14 or undergoes a 1,5-cyclization affording 13. In boiling xylene,
−dithione× 2 dissociates, and the highly reactive dimethylthioketene either dimerizes to

2652
Helvetica Chimica Acta Vol. 85 (2002)
give 15 or is intercepted by azomethine ylide 12 to yield 16. These results show that 2
can be exploited as a source of dimethylthioketene. To the best of our knowledge, the
formation of 16 is the first example of a 1,3-dipolar cycloaddition with dimethyl-
thioketene, and, on the other hand, it is the first reaction of an azomethine ylide with a
thioketene.
We thank the analytical sections of our institutes for spectra and elemental analyses and Miss Jovita Cavegn
for her assistance with the determination of the crystal structures. Financial support of the Polish State
Committee for Scientific Research (Grant No. 3TO9A 00716), the Swiss National Science Foundation, and F.
Hoffmann-La Roche AG, Basel, is gratefully acknowledged.
Experimental Part
1. General. See [19].
2. Starting Materials. 2,2,4,4-Tetramethyl-3-thioxocyclobutanone (1) and 2,2,4,4-tetramethylcyclobutane-1,3-
dithione (2) were prepared by thionation of 2,2,4,4-tetramethylcyclobutane-1,3-dione with P4S10 in pyridine as
described in [33]. cis-1-Isopropyl-2,3-diphenylaziridine (5b) was synthesized by Gabriel×s method according to
[34], cis- and trans-1-(4-methoxyphenyl)aziridine-2,3-dicarboxylates (cis- and trans-7) were obtained by a
modified procedure based on the reaction of 4-methoxyphenyl azide with dimethyl fumarate and subsequent
thermal decomposition of the primarily formed 4,5-dihydro-1H-1,2,3-triazole derivative as reported by Szeimies
and Huisgen [35]. Dimethyl 4,5-dihydro-1,5-diphenyl-1H-1,2,3-triazole-4,4-dicarboxylate (10) was obtained
¬
from dimethyl benzylidenemalonate and phenyl azide according to Texier and Carrie [27].
3. Reaction of 1 with 5b. A soln. of 5b (521 mg, 2.2 mmol) and 1 (312 mg, 2.0 mmol) in 5 ml of freshly
purified toluene was heated under reflux for 30 h. After evaporation of the solvent, the crude mixture was
analyzed by 1H-NMR spectroscopy, and complete conversion of 1 was evidenced by the absence of the
characteristic s of 1 at 1.35 ppm. Separation of the mixture was achieved by CC (SiO2; petroleum ether with
increasing amounts of CH2Cl2). The main fraction isolated with petroleum ether/CH2Cl2 7:3was identified as
6b. Anal. pure 6b was obtained by recrystallization from MeOH.
trans-7-Isopropyl-1,1,3,3-tetramethyl-6,8-diphenyl-5-thia-7-azaspiro[3.4]octan-2-one (6b). Yield: 240 mg
(30%). Colorless crystals. M.p. 114 1168 (MeOH). IR (KBr): 2966s, 1776vs (CˆO), 1460m, 1365m, 1171m,
1034m, 723m, 702m. 1H-NMR: 0.41, 1.05 (2d, J ˆ 6.7, Me2CH); 0.63, 1.07, 1.39, 1.66 (4s, 4 Me); 2.47 (sept., J ˆ 6.7,
Me2CH); 5.09, 5.37 (2s, HÀC(6), HÀC(8)); 7.24 7.46 (m, 8 arom. H); 8.50 8.53( m, 2 arom. H). 13C-NMR:
18.3, 20.7, 21.5, 23.6, 23.7, 25.5 (6 Me); 48.2 (Me2CH); 61.0, 66.6 (C(1), C(3)); 67.1, 70.2 (C(6), C(8)); 68.3 (C(4));
127.6, 127.8, 127.9, 128.2, 128.4, 128.8, 131.7 (10 arom. CH); 140.4, 141.3 (2 arom. C); 222.4 (CˆO). ESI-MS
‡
(NaI): 416 (100, [M ‡ Na] ). Anal. calc. for C25H31NOS (393.60): C 76.29, H 7.94, N 3.56, S 8.15; found: C 75.88,
H 7.53, N 3.41, S 7.53.
4. Reaction of 1 with trans-7. A soln. of trans-7 (265 mg, 1 mmol) and 1 (156 mg, 1 mmol) in 3ml of toluene
was heated under reflux, until the red color of 1 completely disappeared (14 h). A similar experiment carried
out in boiling xylene was complete after 5.5 h. After evaporation of the toluene, the crude mixture was analyzed
by 1H-NMR and a 4 :1 ratio for cis-8/trans-8 was established based on the intensities of the s attributed to
HÀC(6)/HÀC(8) in each product (the ratio in xylene was 1:1). A first crop of cis-8 (180 mg) was obtained after
crystallization from MeOH. The mother liquor, after evaporation of the solvent, was separated by means of CC
(SiO2; petroleum ether with increasing amounts of CH2Cl2). The reported yields refer to total amounts of
isolated products.
Dimethyl cis-7-(4-Methoxyphenyl)-1,1,3,3-tetramethyl-2-oxo-5-thia-7-azaspiro[3.4]octane-6,8-dicarboxylate
(cis-8a). Yield: 210 mg (50%). Colorless crystals. M.p. 135 1378 (MeOH). IR (KBr): 1778vs (CˆO, ketone),
1753vs (CˆO, ester), 1730vs (CˆO, ester), 1516vs, 1437m, 1275s, 1248s, 1205s, 1178s, 1038m, 1016m. 1H-NMR:
1.19, 1.28, 1.32, 1.38 (4s, 4 Me); 3.73, 3.76, 3.77 (3s, 3MeO); 5.07, 5.35 (2 s, HÀC(6), HÀC(8)); 6.84, 6.75
(AA'BB', J ˆ 9.2, 4 arom. H). 13C-NMR: 21.6, 21.9, 22.1, 22.3(4 Me); 52.0, 52.8, 55.5 (3MeO); 60.1, 66.3(C(1),
C(3)); 61.7, 65.1 (C(6), C(8)); 68.7 (C(4)); 113.7, 115.0 (4 arom. CH); 137.5, 152.8 (2 arom. C); 169.8, 169.9
‡
(2 CˆO, ester); 218.6 (CˆO, ketone). ESI-MS (NaI): 444 (100, [M ‡ Na] ). Anal. calc. for C21H27NO6S
(421.51): C 59.84, H 6.46, N 3.32, S 7.61; found: C 59.71, H 6.27, N 3.25, S 7.49.
Dimethyl trans-7-(4-Methoxyphenyl)-1,1,3,3-tetramethyl-2-oxo-5-thia-7-azaspiro-[3.4]octane-6,8-dicarbox-
ylate (trans-8a). Yield: 45 mg (11%). Colorless crystals. M.p. 120 1228 (MeOH). IR (KBr): 1782s (CˆO,
ketone), 1745s (br.) (CˆO, ester), 1514vs, 1435m, 1246vs, 1200vs, 1171vs, 1038s, 820m. 1H-NMR: 1.26, 1.27, 1.36,

Helvetica Chimica Acta Vol. 85 (2002)
2653
1.52 (4s, 4 Me); 3.57, 3.74, 3.75 (3s, 3MeO); 5.13, 5.27 (H ÀC(6), HÀC(8)); 6.80, 6.74 (AA'BB', J ˆ 8.2, 4 arom.
H). 13C-NMR: 19.3, 21.4, 23.2, 24.2 (4 Me); 51.6, 52.5, 55.3 (3 MeO); 60.3, 67.2 (C(1), C(3)); 61.9, 69.4 (C(6),
C(8)); 65.4 (C(4)); 114.7, 118.2 (4 arom. CH); 137.2, 154.3 (2 arom. C); 169.8, 170.7 (2 CˆO, ester); 219.3(C ˆO,
‡
ketone). CI-MS (NH3): 422 (100, [M ‡ 1] ). Anal. calc. for C21H27NO6S (421.51): C 59.84, H 6.46, N 3.32, S 7.61;
found: C 59.78, H 6.38, N 3.15, S 7.48.
5. Reaction of 2 with trans-7. A soln. of 2 (172 mg, 1 mmol) and trans-7 (265 mg, 1 mmol) in 3ml of abs.
toluene was heated under reflux for 15 h. Analysis of the mixture was accomplished as described in Sect. 4. The
ratio cis-8b/trans-8b (1H-NMR) was ca. 4 :1. The crude mixture was dissolved in 2 ml of MeOH and stored
overnight in the refrigerator; then, 150 mg of colorless crystals of cis-8b were filtered. The mother liquor was
evaporated, and the residue was separated (CC, SiO2; petroleum ether with increasing amounts of CH2Cl2) to
give trans-8b (40 mg) as the less polar fraction. The more polar cis-8b (60 mg) was isolated as a thick oil, which
crystallized at r.t.
Dimethyl cis-7-(4-Methoxyphenyl)-1,1,3,3-tetramethyl-2-thioxo-5-thia-7-azaspiro[3.4]octane-6,8-dicarbox-
ylate (cis-8b). Yield: 210 mg (48%). Orange crystals. M.p. 143 146 8 (MeOH). IR (KBr): 2953m, 1755vs
(CˆO), 1726vs (CˆO), 1514vs, 1464m, 1437m, 1277vs, 1218vs, 1203vs, 1174vs, 1041s (CˆS), 818m. 1H-NMR:
1.24, 1.36, 1.40, 1.45 (4s, 4 Me); 3.73, 3.76, 3.77 (3s, 3MeO); 5.18, 5.36 (2 s, HÀC(6), HÀC(8)); 6.77, 6.84
(AA'BB', J ˆ 9.2, 4 arom. H). 13C-NMR: 25.6, 26.1, 26.2, 26.3(4 Me); 52.0, 52.8, 55.3(3MeO); 61.8, 65.4 (C(6),
C(8)); 63.1, 68.4 (C(1), C(3)); 70.6 (C(4)); 113.7, 115.0 (4 arom. CH); 137.6, 152.7 (2 arom. C); 169.7, 169.9 (2
‡
CˆO); 279.6 (CˆS). ESI-MS (NaI): 460 (100, [M ‡ Na] ). Anal. calc. for C21H27NO5S2 (437.57): C 57.64,
H 6.22, N 3.20, S 14.65; found: C 57.52, H 6.03, N 3.06, S 14.57.
Dimethyl trans-7-(4-Methoxyphenyl)-1,1,3,3-tetramethyl-2-thioxo-5-thia-7-azaspiro[3.4]octane-6,8-dicar-
boxylate (trans-8b). Yield: 50 mg (11%). Orange crystals. M.p. 128 1308 (MeOH). IR (KBr): 2968m, 1738vs
1
(CˆO), 1514vs, 1362m, 1269vs, 1248vs, 1200vs, 1174vs, 1036vs (CˆS), 985m, 83 3m. H-NMR: 1.33, 1.34, 1.45,
1.59 (4s, 4 Me); 3.57, 3.74, 3.75 (3s, 3MeO); 5.21, 5.29 (2 s, HÀC(6), HÀC(8)); 6.75, 6.80 (AA'BB', J ˆ 9.2, 4
arom. H). 13C-NMR: 23.1, 25.3, 27.4, 28.6 (4 Me); 51.6, 52.5, 55.4 (3 MeO); 62.0, 69.7 (C(6), C(8)); 63.2, 69.6
(C(1), C(3)); 69.4 (C(4)); 114.7, 118.2 (4 arom. CH); 137.3, 154.3 (2 arom. C); 170.0, 170.7 (2 CˆO); 280.3
‡
(CˆS). CI-MS (NH3): 438 (100, [M ‡ 1] ). Anal. calc. for C21H27NO5S2 (437.57): C 57.64, H 6.22, N 3.20,
S 14.65; found: C 57.48, H 6.12, N 3.11, S 14.51.
6. Reactions of cis-7 with 1 and 2. Solns. of cis-7 (88 mg, 0.3mmol) and 0.3mmol of 1 or 2 in 1 ml of abs.
xylene were heated under reflux for 3h. In both experiments, TLC tests showed complete conversion of cis-7.
After evaporation of the solvent, the residual thick oils were examined by 1H-NMR spectroscopy, and trans-8a
or trans-8b, respectively, were identified as exclusive products. After crystallization, pure trans-8a (71 mg, 56%)
and trans-8b (81 mg, 62%), identical (m.p., 1H-NMR, IR) with the products of Sect. 4 and 5, were obtained.
7. Reaction of 10 with 1. A soln. of 10 (678 mg, 2.0 mmol) and 1 (345 mg, 2.2 mmol) in 1.5 ml (or 4.5 ml) of
xylene was heated under reflux for 60 h. The initially red color of the mixture changed to brown. After
evaporation of the solvent, the partially decomposed residues were analyzed by 1H-NMR spectroscopy, and the
ratio 13/14 was established based on the intensities of s at 0.67 ppm (3H) for 13 and s at 1.92 ppm (3H) for 14.
CC (SiO2; petroleum ether/CH2Cl2 4 :6) gave a fraction containing 14, and the more polar 13 was eluted with
CH2Cl2. Anal. pure samples were obtained by recrystallization.
Dimethyl 1,1,3,3-Tetramethyl-7,8-diphenyl-5-thia-7-azaspiro[3.4]octan-6,6-dicarboxylate (13). Yield: 60 mg
(6%, from the less concentrated mixture). Colorless crystals. M.p. 196 1988 (hexane/CH2Cl2). IR (KBr):
2958m, 1778vs (CˆO), 1741vs (CˆO), 1597s, 1504vs, 1454s, 1335s, 1254vs, 1213s, 1171s, 1049m, 1034s, 748m,
704m. 1H-NMR: 0.67, 1.28, 1.42, 1.53(4 s, 4 Me); 3.48, 3.74 (2s, 2 MeO); 6.00 (s, HÀC(8)); 6.77 7.26 (m, 10 arom.
H). 13C-NMR: 20.4, 23.6, 23.7, 24.1 (4 Me); 52.9, 53.4 (2 MeO); 61.4, 67.3 (C(1), C(3)); 68.5 (C(4)); 69.8 (C(8));
75.8 (C(5)); 116.8, 119.5, 128.1, 128.2, 128.6 (10 arom. CH); 138.9, 142.1 (2 arom. C); 168.6, 169.0 (2 CˆO);
‡
220.5 (CˆO, cyclobutanone). ESI-MS (NaI): 420 (100, [M ‡ Na] ). Anal. calc. for C26H29NO5S (467.59):
C 66.79, H 6.25, N 2.99, S 6.86; found: C 66.82, H 6.38, N 3.06, S 7.01.
Dimethyl 5-Isopropylidene-3,4-diphenyl-1,3-thiazolidine-2,2-dicarboxylate (14). Yield: 55 mg (7% from the
more concentrated mixture). Colorless crystals. M.p. 116 1188 (hexane/CH2Cl2). IR (KBr): 1755s (CˆO),
1
1740s (CˆO), 1597m, 1504s, 1433m, 1240s (br.), 1055m, 1041m, 748m, 712m. H-NMR: 1.80, 1.92 (2s, 2 Me);
3.70, 3.88 (2s, 2 MeO); 5.81 (s, HÀC(4)); 6.61-7.60 (m, 10 arom. H). 13C-NMR: 22.0, 23.7 (2 Me); 53.2 (br.,
2 MeO); 71.4 (C(4)); 76.9 (C(2)); 116.0, 119.5, 126.8, 127.3, 128.4, 128.6 (10 arom. CH); 126.6, 127.6 (CˆC);
‡
124.1, 144.4 (2 arom. C); 167.2, 169.5 (2 CˆO). ESI-MS: 420 (100, [M ‡ Na] ), 302 (20). Anal. calc. for
C22H23NO4S (397.49): C 66.48, H 5.83, N 3.52, S 8.07; found: C 66.27, H 5.60, N 3.47, S 7.91.
8. Reaction of 10 with 2. A soln. of 10 (678 mg, 2 mmol) and 2 (344 mg, 2 mmol) in 2 ml of abs. xylene was
heated under reflux for 5 h. After this time, the TLC analysis showed complete conversion of 2. Xylene was

2654
Helvetica Chimica Acta Vol. 85 (2002)
Table. Crystallographic Data of 6b, cis-8a, cis-8b, 10, and 16
6b
cis-8a
cis-8b
10
16
Crystallized from
Empirical formula
Formula weight
MeOH
C25H31NOS
393.59
MeOH
C21H27NO6S
421.50
MeOH/CH2Cl2
C21H27NO5S2
437.57
MeOH/CH2Cl2
C18H17N3O4
339.35
MeOH/CH2Cl2
C22H23NO4S
397.49
[g molÀ1
]
Crystal color, habit
colorless,
plate
colorless,
prism
orange,
needle
0.02 Â 0.05
 0.25
colorless,
prism
colorless,
prism
Crystal dimensions
[mm]
T [K]
Crystal system
Space group
Z
0.05 Â 0.15
 0.22
160(1)
monoclinic
P2l/c
0.07 Â 0.12
0.10 Â 0.22
 0.28
160(1)
monoclinic
P2l/n
0.15 Â 0.17
 0.22
 0.20
160(1)
orthorhombic
P2l2l2l
160(1)
orthorhombic
P2l2l2l
160(1)
triclinic
≈
Pl
2
4
4
4
4
Reflections for cell de-
termination
2q Range for cell de-
termination [8]
Unit-cell parameters
a [ä]
b [ä]
c [ä]
a [8]
b [8]
5212
3531
52251
5042
5662
4
55
4
60
4
50
4
60
4
60
11.1313(2)
12.7310(2)
15.3727(2)
90
93.1073(7)
90
2175.30(6)
1.202
0.164
8.0291(1)
12.9780(2)
20.3603(3)
90
90
90
2121.58(5)
1.319
0.189
f and w
60
8.0658(1)
12.9688(2)
21.0148(3)
90
90
90
2198.23(5)
1.322
0.274
8.2591(1)
15.2263(2)
13.8455(2)
90
106.5616(6)
90
1668.91(4)
1.350
0.09730.191
f and w
60
7.8680(1)
11.4916(2)
12.0825(2)
106.0120(7)
103.8056(7)
98.3705(7)
993.24(3)
1.329
g [8]
V [ä3]
Dx [g cmÀ3
]
m (MoKa) (mmÀ1
Scan type
)
f and w
55
w
f and w
60
2q(max) [8]
50
Transmission factors
(min; max)
0.940; 0.994
Total reflections mea-
sured
Symmetry independent
reflections
Reflections used
[I > 2s(I)]
81085
4980
3470
53589
6212
4471
28596
3875
3368
45127
4863
3568
27274
5787
4416
Parameters refined
Final R
254
264
264
227
254
0.0435
0.0440
0.012
0.0437
0.0416
0.015
0.0368
0.0339
0.005
0.0463
0.0447
0.005
0.0437
0.0459
0.005
wR
Weights: p in w ˆ
[s2(Fo) ‡ (pFo)2]À1
Goodness-of-fit
Secondary extinction
coefficient
1.865
1.3(2) Â 10À6
1.420
1.1(1) Â 10À6
1.807
1.1(1) Â 10À6
2.888
2.6(4) Â 10À6
2.454
1.5(6) Â 10À6
Final Dmax/s
D1 (max; min) [e äÀ3
0.0004
0.23; À0.26
0.0004
0.26; À0.28
0.0005
0.23; À0.22
0.0006
0.26; À0.22
0.0003
0.29; À0.25
]

Helvetica Chimica Acta Vol. 85 (2002)
2655
evaporated, and the residue was separated by means of CC (SiO2; petroleum ether with increasing amount of
CH2Cl2). As the less polar fraction, an orange oil was obtained, which was identified as 15. After increasing the
polarity of the eluent, 16 was isolated.
4-Isopropylidene-3,3-dimethylthietane-2-thione (15). Yield: 150 mg (43%). Orange oil ([32]: b.p. 112-1168/
20 Torr). 1H-NMR: 1.40 (s, 2 Me); 1.67 (s, Me); 1.80 (s, Me). 13C-NMR: 19.9 (Me); 21.7 (Me); 25.0 (2 Me); 73.9
(C(3)); 122.3, 130.7 (CˆC); 243.0 (CˆS).
Dimethyl 5-Isopropylidene-2,3-diphenyl-1,3-thiazolidine-4,4-dicarboxylate (16). Yield: 210 mg (26%).
Colorless crystals. M.p. 188 1908 (MeOH/CH2Cl2). IR (KBr): 1763vs (CˆO), 1597w, 1498m, 1456m, 1254vs,
1221s, 1053vs, 768m, 742m, 694m. 1H-NMR: 1.83( s, 2 Me); 3.50, 3.88 (2s, 2 MeO); 6.43( s, HÀC(2)); 6.75 7.68
(m, 10 arom. H). 13C-NMR: 21.5, 25.5 (2 Me); 52.5, 53.1 (2 MeO); 69.5 (C(2)); 80.2 (C(4)); 121.9, 122.7, 127.9,
128.5, 128.7, 128.8, 129.4, 129.6 (10 arom. CH); 127.3, 128.1 (CˆC); 139.6, 143.8 (2 arom. C); 168.6, 169.9
‡
(2 CˆO). CI-MS (NH3): 398 (100, [M ‡ 1] ). Anal. calc. for C22H23NO4S (397.49): C 66.48, H 5.83, N 3.52,
S 8.07; found: C 66.47, H 5.67, N 3.41, S 7.82.
9. Attempted Thermal Elimination of Dimethylketene from 13. A soln. of 13 (20 mg, 0.04 mmol) in 1 ml of
abs. xylene was heated under reflux for 60 h. The solvent was evaporated, and the residue was analyzed by
1H-NMR spectroscopy, which confirmed that only partially decomposed starting material was present.
10. X-Ray Crystal-Structure Determination of 6b, cis-8a, cis-8b, 10, and 16 (see the Table and Figs. 1 4)6).
All measurements were made on a Nonius KappaCCD diffractometer [36] with graphite-monochromated
MoKa radiation (l 0.71073ä) and an Oxford Cryosystem Cryostream 700 cooler. Data reduction was performed
with HKL Denzo and Scalepack [37]. The intensities were corrected for Lorentz and polarization effects, and, in
the case of cis-8b, an absorption correction based on the multiscan method [38] was applied. Equivalent
reflections were merged; except for cis-8a and cis-8b, where Friedel pairs were not merged. Data collection and
refinement parameters are given in the Table, and views of the molecules are shown in Figs. 1 4. Each structure
was solved by direct methods by means of SIR92 [39], which revealed the positions of all non-H-atoms. The non-
H-atoms were refined anisotropically. All of the H-atoms were fixed in geometrically calculated positions
(d(CÀH) ˆ 0.95 ä), and each was assigned a fixed isotropic displacement parameter with a value equal to
1.2Ueq of its parent atom. Refinement of each structure was carried out on F by means of full-matrix least-
squares procedures, which minimized the function Sw(j Fo jÀj Fc j )2. Corrections for secondary extinction were
applied. For 6b, cis-8a, cis-8b, 10, and 16, three, two, two, two, and three reflections, resp., whose intensities were
considered to be extreme outliners, were omitted from the final refinement. The compounds cis-8a and cis-8b
crystallize in a polar space group, and refinement of the absolute structure parameter [40] yielded values of
0.18(6) and 0.08(5), resp. These values are not particularly precise [41], but are suggestive of a partial inversion
twin for cis-8a and probably an enantiomerically pure cis-8b, although enantiomeric purity or some degree of
inversion twinning cannot be excluded for either compound. As enantiomerically pure reaction products were
not expected, some degree of spontaneous resolution may have taken place. Neutral-atom-scattering factors for
non-H-atoms were taken from [42a], and the scattering factors for H-atoms were taken from [43]. Anomalous
dispersion effects were included in Fc [44]; the values for f' and f'' were those of [42b]. The values of the mass
attenuation coefficients are those of [42c]. All calculations were performed by means of the teXsan crys-
tallographic software package [45].
REFERENCES
[1] P. Metzner, Top. Curr. Chem. 1999, 204, 127.
[2] M. T. Molina, M. Yanez, O. Mo, R. Notario, J.-L. Abboud, in −The chemistry of double-bonded functional
groups×, Suppl. A3, Ed. S. Patai, John Wiley & Sons, New York, 1997, p. 1355.
¬
[3] R. Huisgen, J. Penelle, G. Mloston, A. Buyle Padias, H. K. Hall Jr., J. Am. Chem. Soc. 1992, 114, 266.
[4] D. S. C. Black, K. G. Watson, Austr. J. Chem. 1973, 26, 2491.
¬
[5] G. Mloston, R. Huisgen, K. Polborn, Tetrahedron 1999, 55, 11475.
¬
[6] R. Huisgen, X. Li, G. Mloston, C. Fulka, Eur. J. Org. Chem. 2000, 1695.
6
)
Crystallographic data (excluding structure factors) for the structures of 6b, cis-8a, cis-8b, 10, and 16 have
been deposited with the Cambridge Crystallographic Data Centre as supplementary publications No.
CCDC-185696 to CCDC-185700. Copies of the data can be obtained, free of charge, on application to the
CCDC, 12 Union Road, Cambridge CB21EZ, UK (fax: ‡44-((0)1223-336033; e-mail: deposit@ccdc.ca-
m.ac.uk).

2656
Helvetica Chimica Acta Vol. 85 (2002)
¬
[7] R. Huisgen, G. Mloston, K. Polborn, J. Org. Chem. 1996, 61, 6570.
¬
[8] R. Huisgen, G. Mloston, K. Polborn, R. Sustmann, W. Sicking, Liebigs Ann. Recl. 1997, 179; R. Huisgen, G.
¬
Mloston, K. Polborn, F. Palacios-Gambra, Liebigs Ann. Recl. 1997, 187.
¬
[9] G. Mloston, H. Heimgartner, Helv. Chim. Acta 1995, 78, 1293.
¬
[10] G. Mloston, T. Gendek, H. Heimgartner, Helv. Chim. Acta 1998, 81, 1585.
¬
[11] M. K‰gi, A. Linden, G. Mloston, H. Heimgartner, Helv. Chim. Acta 1996, 79, 855; 1998, 81, 285.
[12] L. Harwood, in −The Chemistry of Heterocyclic Compounds, Vol. 59: Synthetic Applications of 1,3-Dipolar
Cycloaddition Chemistry Toward Heterocycles and Natural Products×, Eds. A. Padwa, W. H. Pearson, John
Wiley & Sons, New York, 2002, p. 315; J. W. Lown, in −1,3-Dipolar Cycloaddition Chemistry×, Vol. 1, Ed. A.
Padwa, John Wiley & Sons, New York, 1984, p. 653.
[13] E. Vedejs, in −Advances in Cycloaddition×, Vol. 3, Ed. D. P. Curran, JAI Press, London, 1993, p. 33.
[14] T. Gallagher, J. Heterocycl. Chem. 1999, 36, 1365.
¬
¬
[15] G. Mloston, A. Linden, H. Heimgartner, Pol. J. Chem. 1997, 71, 32; G. Mloston, A. Linden, H. Heimgartner,
Helv. Chim. Acta 1998, 81, 558.
¬
[16] A. Gebert, A. Linden, G. Mloston, H. Heimgartner, Heterocycles 2002, 56, 393; A. Gebert, H.
Heimgartner, Helv. Chim. Acta 2002, 85, 2073.
¬
[17] G. Mloston, Z. Skrzypek, Bull. Soc. Chim. Belg. 1990, 99, 167.
[18] A. Gebert, A. Linden, H. Heimgartner, Heterocycles 2001, 54, 691.
¬
[19] K. Urbaniak, G. Mloston, H. Heimgartner, Helv. Chim. Acta 2002, 85, 2056.
[20] G. Seybold, Tetrahedron Lett. 1974, 555, and refs. cited therein.
[21] E. Schaumann, Tetrahedron 1988, 44, 1827.
[22] N. F. Osborne, J. Chem. Soc., Perkin Trans. 1 1982, 1429.
[23] R. Huisgen, in −1,3-Dipolar Cycloaddition Chemistry×, Vol. 1, Ed. A. Padwa, John Wiley & Sons, New York,
1984, p. 1; R. M. Kellogg, Tetrahedron 1976, 32, 2165.
¬
[24] T. Olszak, T. Skarzynski, G. Mloston, Acta Crystallogr., Sect. C 1988, 44, 205.
[25] C. K. Johnson, −ORTEP II×, Report ORNL-5138, Oak Ridge National Laboratory, Oak Ridge, TN, 1976.
[26] G. Gonzalez, M. V. Martin, M. C. Paredes, Heterocycles 2000, 52, 23 7.
¬
¬
¬
[27] F. Texier, R. Carrie, Bull. Soc. Chim. Fr. 1971, 4119; D. Gree, R. Carrie, Can. J. Chem. 1984, 62, 939.
¬
[28] M. Domagala, K. Urbaniak, G. Mloston, Acta Crystallogr., Sect. C, submitted.
¬
[29] R. Huisgen, G. Mloston, E. Langhals, Helv. Chim. Acta 2001, 84, 1805.
¬
[30] R. Huisgen, G. Mloston, H. Giera, E. Langhals, Tetrahedron 2002, 58, 507.
¬
[31] R. Huisgen, E. Langhals, G. Mloston, T. Oshima, J. Rapp, J. Heterocycl. Chem. 1987, 9, S-1.
[32] K. Muthuramu, B. Sundari, V. Ramamurthy, Tetrahedron 1983, 39, 2719.
¬
¬
[33] G. Mloston, J. Romanski, A. Linden, H. Heimgartner, Helv. Chim. Acta 1993, 76, 2147.
¬
[34] R. Bartnik, G. Mloston, S. Lesniak, Pol. J. Chem. 1979, 5, 53 7.
[35] G. Szeimies, R. Huisgen, Chem. Ber. 1966, 99, 491.
[36] R. Hooft, KappaCCD Collect Software, Nonius BV, Delft, The Netherlands, 1999.
[37] Z. Otwinowski, W. Minor, in −Methods in Enzymology, Vol. 276, Macromolecular Crystallography×, Part A,
Eds. C. W. Carte Jr., R. M. Sweet, Academic Press, New York, 1997, p. 307.
[38] R. H. Blessing, Acta Crystallogr., Sect. A 1995, 51, 3 3 .
[39] A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi, M. C. Burla, G. Polidori, M. Camalli, SIR92, J.
Appl. Crystallogr. 1994, 27, 435.
[40] a) H. D. Flack, Acta Crystallogr., Sect. A 1983, 39, 876; b) G. Bernardinelli, H. D. Flack, Acta Crystallogr.,
Sect. A 1985, 41, 500.
[41] H. D. Flack, G. Bernardinelli, J. Appl. Crystallogr. 2000, 33, 1143.
[42] a) E. N. Maslen, A. G. Fox, M. A. O×Keefe, in −International Tables for Crystallography×, Ed. A. J. C.
Wilson, Kluwer Academic Publishers, Dordrecht, 1992, Vol. C, Table 6.1.1.1, p. 477; b) D. C. Creagh, W. J.
McAuley, in −International Tables for Crystallography×, Ed. A. J. C. Wilson, Kluwer Academic Publishers,
Dordrecht, 1992, Vol. C, Table 4.2.6.8, p. 219; c) D. C. Creagh, J. H. Hubbell, in −International Tables for Crys-
tallography×, Ed. A. J. C. Wilson, Kluwer Academic Publishers, Dordrecht, 1992, Vol. C, Table 4.2.4.3, p. 200.
[43] R. F. Stewart, E. R. Davidson, W. T. Simpson, J. Chem. Phys. 1965, 42, 3175.
[44] J. A. Ibers, W. C. Hamilton, Acta Crystallogr. 1964, 17, 781.
[45] −teXsan: Single Crystal Structure Analysis Software×, Version 1.10, Molecular Structure Corporation, The
Woodlands, TX, 1999.
Received June 5, 2002