The role of steric and electronic interactions in the stereocontrol of the asymmetric 1,3-dipolar reactions of 5-ethoxy-3-p-(S)-tolylsulfinylfuran-2(5H)-ones with diazoalkanes: Theoretical calculations and experimental evidences

Article abstract of DOI:10.1021/jo0345603

The addition of diazomethane and diazoethane to (5S,SS)- and (5R,SS)-5-ethoxy-3-p-tolylsulfinylfuran-2(5H)-ones (1a and 1b) and their 4-methylderivatives (2a and 2b) proceeded in almost quantitative yields and complete regioselectivity. The observed π-facial selectivity is determined by the configurations at both C-5 and the sulfinyl group, the later being the most important. The syn adducts were almost exclusively obtained from 1a and 2a in apolar solvents but the π-facial selectivity was strongly decreased in more polar solvents. On the other hand, the major adducts from 1b and 2b were the anti ones and such predominance was slightly increased with solvent polarity. The exo-selectivity was complete in all the cases except for the anti approach to compounds 2a (in polar solvents) and 2b. The role of the sulfinyl group in this behavior was inferred by comparison of these results with those obtained in reactions of diazoalkanes with 5-methoxyfuran-2(5H)-one (3). Steric interactions seem to be the main ones responsible for the observed exo selectivity of reactions with diazoethane, but electronic factors, which can be modulated by the solvent, are also significant in the π-facial selectivity control. DFT computational methods are able to correctly predict the reactivity, regioselectivity, and π-facial selectivity exhibited by 5-alkoxyfuranones as well as their changes with the solvent polarity. A C-H...O hydrogen bond, involving the oxygen atom of the 5-alkoxy group at dipolarophiles and the endo-hydrogen atom at dipoles, seems to play a key role in the electronic interactions influencing the stereochemical course of these reactions.

Full text of DOI:10.1021/jo0345603

Th e Role of Ster ic a n d Electr on ic In ter a ction s in th e
Ster eocon tr ol of th e Asym m etr ic 1,3-Dip ola r Rea ction s of
5-Eth oxy-3-p-(S)-tolylsu lfin ylfu r a n -2(5H)-on es w ith Dia zoa lk a n es:
Th eor etica l Ca lcu la tion s a n d Exp er im en ta l Evid en ces
J ose´ L. Garc´ıa Ruano,*,† Alberto Fraile,† Gemma Gonza´lez,† M. Rosario Mart´ın,*,†
Fernando R. Clemente,‡ and Ruth Gordillo‡
Departamento de Quı´mica Orga´nica, Universidad Auto´noma, Cantoblanco, 28049-Madrid, Spain, and
Departamento de Quı´mica Orga´nica, Universidad de Extremadura, Avda. de Elvas, s/ n,
06071 Badajoz, Spain
joseluis.garcia.ruano@uam.es
Received April 30, 2003
The addition of diazomethane and diazoethane to (5S,SS)- and (5R,SS)-5-ethoxy-3-p-tolylsulfinyl-
furan-2(5H)-ones (1a and 1b) and their 4-methylderivatives (2a and 2b) proceeded in almost
quantitative yields and complete regioselectivity. The observed π-facial selectivity is determined
by the configurations at both C-5 and the sulfinyl group, the later being the most important. The
syn adducts were almost exclusively obtained from 1a and 2a in apolar solvents but the π-facial
selectivity was strongly decreased in more polar solvents. On the other hand, the major adducts
from 1b and 2b were the anti ones and such predominance was slightly increased with solvent
polarity. The exo-selectivity was complete in all the cases except for the anti approach to compounds
2a (in polar solvents) and 2b. The role of the sulfinyl group in this behavior was inferred by
comparison of these results with those obtained in reactions of diazoalkanes with 5-methoxyfuran-
2(5H)-one (3). Steric interactions seem to be the main ones responsible for the observed exo
selectivity of reactions with diazoethane, but electronic factors, which can be modulated by the
solvent, are also significant in the π-facial selectivity control. DFT computational methods are able
to correctly predict the reactivity, regioselectivity, and π-facial selectivity exhibited by 5-alkoxy-
furanones as well as their changes with the solvent polarity. A C-H‚‚‚O hydrogen bond, involving
the oxygen atom of the 5-alkoxy group at dipolarophiles and the endo-hydrogen atom at dipoles,
seems to play a key role in the electronic interactions influencing the stereochemical course of
these reactions.
In tr od u ction
Substituted vinyl sulfoxides have been widely used as
chiral dienophiles in asymmetric Diels-Alder reactions
due to the ability of the sulfinyl group to control their
π-facial selectivity, despite its influence on the reactivity
and endo/exo selectivity being not too large.1 In contrast,
the number of papers reporting the use of homochiral
vinyl sulfoxides as dipolarophiles is quite low,1,2 maybe
due to the easy desulfinylation of the resulting adducts.
Some years ago we initiated a program devoted to explore
the behavior of optically pure vinyl sulfoxides as dipo-
larophiles. Our first contribution3 concerned reactions of
diazomethane with (5S,SS)- and (5R,SS)-5-ethoxy-3-p-
tolylsulfinylfuran-2(5H)-ones, 1a and 1b, and their 4-
methyl derivatives, 2a and 2b (Figure 1). The results of
F IGURE 1. Dipolarophiles studied.
these reactions revealed that dipolarophilic features of
these cyclic sulfoxides were much better than their
dienophilic ones, with the sulfinyl group playing a quite
significant role in the reactivity and stereoselectivity of
the 1,3-dipolar reactions. In that first paper we proposed
that the π-facial selectivity was complete and exclusively
controlled by the sulfinyl configuration. Moreover, the
steric effects appeared as the only ones responsible for
the stereochemical course of the reactions. These studies
revealed that substrates 1 and 2 exhibited some advan-
tages as sulfinyl dipolarophiles in asymmetric synthesis
because the primary adducts were stable enough to be
isolated due to their bicyclic structure, which determined
that they were not too prone to desulfinylation. Conse-
* Corresponding authors. Fax: 3491 397 3966.
† Universidad Auto´noma.
‡ Universidad de Extremadura.
(1) Cid, M. B.; Garc´ıa Ruano, J . L. In Topics in Curent Chemistry.
Organosulfur Chemistry; Page, P. C. B., Ed.; Springer: Berlin,
Germany, 1999; pp 1-126.
(2) Gothelf, K. V.; J ørgersen, K. A. Chem. Rev. 1998, 98, 863-909.
(3) Garc´ıa Ruano, J . L.; Fraile, A.; Mart´ın M. R. Tetrahedron:
Asymmetry 1996, 7, 1943-1950.
10.1021/jo0345603 CCC: $25.00 © 2003 American Chemical Society
Published on Web 07/19/2003
6522
J . Org. Chem. 2003, 68, 6522-6534

The Role of Steric and Electronic Interactions
TABLE 1. Ad d ition s of Dia zoeth a n e a n d Dia zom eth a n e
to 1a
quently, we have also studied the reactions of 1a and 1b
with some other dipoles such as nitrile oxides,4,5 ni-
trones,6 and azomethine ylides.7
All these studies revealed to us that electrostatic
interactions also played an important role in the π-facial
selectivity control of many of these 1,3-dipolar reactions,
which contrasted with their apparent scarce significance
that we had proposed in the reactions with diazometh-
ane.3 Moreover, the presence of the sulfinyl group was
also decisive in determining the endo/exo selectivity. To
get some evidence about these influences in the reactions
of sulfinyl dipolarophiles with diazoalkanes, we initiated
the study of the reactions of 1 and 2 with diazoethane.
From the results obtained in this study, we inferred the
significant role of the electrostatic interactions in the
π-facial selectivity (by considering the influence of the
solvent polarity in the reaction course), thus questioning
the conclusions that we have presented in our first paper
on reactions with diazomethane.3 This prompted us to
revise the available data and perform new studies on
these reactions. In this paper we report a complete study
of the reactions of diazoethane and diazomethane with
(5S,SS)- and (5R,SS)-5-ethoxy-3-p-tolylsulfinylfuran-
2(5H)-ones, 1a and 1b (epimers at C-5), as well as with
their 4-methyl derivatives, 2a and 2b, under different
conditions. Additionally, we present the cycloadditions
of both diazoalkanes with 5-methoxyfuran-2(5H)-one (3),
which lacks the sulfinyl group at C-3. The comparison of
these results with those obtained from 1a and 1b allowed
us to know the role of the sulfinyl group at C-3 in the
features of these 1,3-dipolar reactions. Finally, we have
also studied these reactions from a theoretical point of
view, by using models able to provide a reasonable
estimate of the effect of solvent polarity on the energy
barriers. From these calculations, which correctly predict
most of the features of these reactions, it is possible to
know the nature of the different factors controlling them.
A new type of C-H‚‚‚O hydrogen bond, involving the
oxygen atom of the 5-alkoxy group of the dipolarophile
and the endo-hydrogen atom of the dipole, seems to play
a key role in the stereochemical outcome of these reac-
tions. This kind of interaction has been the subject of
study in recent computational papers,8 including effects
on the stereoselectivity of aldol8b and epoxidation8c reac-
tions.
entry
R
solvent
Et2O
Et2O
Et2O
CH3CN/Et2O
CH3CN/Et2O
Et2O
CH3CN/Et2O
CH3CN/Et2O
T (°C)
syn:anti
1a
2
3
Me
Me
Me
Me
Me
H
0
-40
-78
0
95:5
96:4
96:4
60:40
48:52
100:0
50:50
40:60
4
5
-40
6b
7
0
0
H
H
8c
-40
a
b
Isolated yield of syn-4a -exo 80%. These data are identical
with those reported in ref 3. c Isolated yield for the mixture syn-
5a + anti-5a 78%.
The results in the reactions of 1a with an excess of
diazomethane and diazoethane under different conditions
are collected in Table 1. The reaction of 1a with an
ethereal solution of diazoethane was complete in 5 min
1
at 0 °C, yielding a 95:5 mixture of ∆-pyrazolines syn-
4a -exo/anti-4a -exo11 in an almost quantitative yield (entry
1, Table 1). Reactions were also instantaneous at lower
temperatures, but diastereomeric ratios did not increase
significantly (entries 2 and 3, Table 1). The reactions
performed by using a 7.5:1 mixture of CH3CN/Et2O as
the solvent (entries 4 and 5) also took place in 5 min and
evolved with a clearly lower facial diastereoselection
yielding mixtures of syn and anti cycloadducts, the later
being slightly favored at -40 °C (entry 5). Only the exo-
adduct was obtained in all these reactions.11
Reactions with diazomethane showed a similar behav-
ior. The π-facial selectivity was complete in Et2O at 0 °C
(entry 6), which allowed the complete characterization
of syn-5a ,3 but strongly decreased in CH3CN (entry 7),
the anti-5a being the major adduct at -40 °C (entry 8).
The reactions of 1b12 with diazoethane (Table 2) were
also completely regioselective and exo-selective. The
π-facial selectivity in Et2O was significantly lower than
that observed for 1a . Thus, reactions conducted at 0 °C
afforded a 85:15 mixture of pyrazolines anti-4b-exo and
syn-4b-exo (entry 1, Table 2). Diastereomeric excesses
increased as the temperature decreased (entries 2 and
Resu lts a n d Discu ssion
We had previously reported the synthesis of compounds
1a and 1b,9 as well as that of their 4-methyl derivatives
2a and 2b.3 Compound 3 had been described by Schenck.10
(9) Carretero, J . C.; Garc´ıa Ruano, J . L.; Lorente, A.; Yuste, F.
Tetrahedron: Asymmetry 1993, 4, 177-180.
(10) Schenck, G. O. Liebigs Ann. 1953, 584, 156-176.
(4) Garc´ıa Ruano, J . L.; Fraile, A.; Mart´ın M. R. Tetrahedron 1999,
55, 14491-14500.
(11) The terminology syn/anti indicates the spatial arrangement of
the alkoxy group at C-5 and the pyrazoline ring at the furanone moiety.
It also indicates that the approach of the dipole has taken place either
to the face containing the alkoxy group (syn-approach) or to the
opposite one (anti-approach). Therefore, the syn/anti ratio will be
indicative of the π-facial selectivity of the reaction. The term exo
indicates the trans arrangement of the methyl group of the diazoethane
with respect to the furanone moiety at the pyrazoline ring. It is
characteristic of the exo approach of the dipole to the dipolarophile,
using as the reference the carbonyl group at the later.
(5) Garc´ıa Ruano, J . L.; Bercial, F.; Fraile, A.; Mart´ın M. R. Synlett
2002, 73-76.
(6) Garc´ıa Ruano, J . L.; Fraile, A.; Mart´ın Castro, A. M.; Mart´ın,
M. R. 19th International Symposium Organic Chemistry of Sulfur;
2000; Book of Abstracts, pp 21.
(7) Garc´ıa Ruano, J . L.; Fraile, A.; Mart´ın, M. R. 19th International
Symposium Organic Chemistry of Sulfur; 2000; Book of Abstracts, p
52.
(8) (a) Raymo, F. M.; Bartberger, M. D.; Houk, K. N.; Stoddart, J .
F. J . Am. Chem. Soc. 2001, 123, 9264-9267. (b) Bahmanyar, S.; Houk,
K. N. J . Am. Chem. Soc. 2001, 123, 12911-12912. (c) Washington, I.;
Houk, K. N. Angew. Chem., Int. Ed. 2001, 40, 4485-4488. (d)
Cannizzaro, C. E.; Houk, K. N. J . Am. Chem. Soc. 2002, 124, 7163-
7169.
(12) As the isolation of 1b in its diastereomerically pure form was
very difficult these reactions were performed starting from a 94:6
mixture of 1b and 1a . As a consequence, in all cases we also obtained
small amounts (∼6%) of the adduct syn-4a -exo, which could be easily
separable from the 4b adducts.
J . Org. Chem, Vol. 68, No. 17, 2003 6523

Garc´ıa Ruano et al.
TABLE 2. Ad d ition s of Dia zoeth a n e a n d Dia zom eth a n e
to 1b
entry
R
solvent
Et2O
Et2O
Et2O
CH3CN/Et2O
CH3CN/Et2O
Et2O
Et2O
Et2O
T (°C)
syn:anti
F IGURE 2. Coupling constants of adduct 4 and 5.
1a
2
3
Me
Me
Me
Me
Me
H
H
H
H
H
0
-40
-78
0
15:85
8:92
7:93
8:92
5:95
17:83
9:91
7:93
15:85
9:91
TABLE 3. Ad d ition s of Dia zoeth a n e a n d Dia zom eth a n e
to 2a
4
5
-40
6b
7
0
-40
-78
0
8
9
10
CH3CN/Et2O
CH3CN/Et2O
-40
a
Isolated yield of syn-4b-exo 10%. b Isolated yield of syn-5b 13%.
entry
R
solvent
time (h) T (°C) syn:anti (exo:endo)
1
2
Me Et2O
Me CH3CN/Et2O
Me CH3CN/Et2O
1.5
1
5.5
1.5
1
0
0
100:0
84:16 (12:4)
82:18 (13:5)
100:0
75:25
74:26
3), reaching a 86% de at -78 °C. Contrasting with the
results obtained from 1a , the π-facial selectivity slightly
increased when the reactions from 1b were conducted in
CH3CN/Et2O mixtures (compare entries 4 and 5 with 1
and 2, respectively).
3
-40
4a
5
H
H
H
Et2O
CH3CN/Et2O
CH3CN/Et2O
0
0
6
6.5
-40
a
This result is identical with that reported in ref 3.
Similar results were obtained in reactions with diazo-
methane: the adduct syn-5b was formed in a similar
proportion to that of syn-4b-exo in its mixture13 (compare
entries 6 and 1, Table 2). The influence of the tempera-
ture (compare entries 2 and 3 with 7 and 8) is identical
for both diazoalkanes, whereas the influence of the
polarity of the solvent (compare entries 6 and 1 with 9
and 4, respectively) is scarcely significant in the reactions
with diazomethane.
All the initial purification trials performed on com-
pounds 4 with use of chromatography on silica gel or
crystallization were unsuccessful, due to the easy decom-
position of the pyrazolines. The use of a silica gel
previously treated with Et3N for 24 h for the column
chromatography avoided the decomposition of the com-
pounds syn-4a -exo and syn-4b-exo, allowing their separa-
tion from their respective mixtures as well as their
complete spectroscopic characterization. The adducts
anti-4 were characterized from the NMR spectra of their
mixtures along with the syn-adducts, once all the signals
of the later had been assigned. The adducts 5, derived
from diazomethane, are stable in silica gel, but their
chromatographic separation was only possible for adducts
5b. Compound syn-5a , which had been obtained as a sole
diastereoisomer under the conditions of entry 6 (Table
1), could be easily characterized, whereas the spectro-
scopic parameters for its anti isomer were obtained from
the spectra of the purified mixtures of syn-5a + anti-5a ,
once the signals of the first one had been completely
assigned.
The syn or anti stereochemistry of the pyrazolines11
was deduced from the values observed for J 3a,4 (5.9-6.7
and 1.5-2.8 Hz, respectively, see Figure 2), whereas the
exo character11 of all the cycloadducts was established
from the values of J 3,3a (they ranged between 2.5 and 4.4
Hz). As a reference we used the values of the coupling
constants measured for pyrazolines derived from diazo-
methane (Figure 2), whose configurational assignment
had been unequivocally established.3 The absolute con-
figurations of compounds 4 and 5 were assigned as
indicated in Tables 1 and 2 on the assumption that the
configurations at C-5 and sulfur of the starting sulfinyl
furanones 1a and 1b remained unaltered in the course
of these reactions.
The reaction of 2a with diazoethane, in a nonpolar
solvent (entry 1, Table 3) evolved with complete regio-
and stereoselectivity, affording syn-6a -exo as the sole
diastereoisomer in quantitative yield. When a 13.5:1
mixture of CH3CN/Et2O was used as the solvent, two new
compounds, both of them with the anti-stereochemistry
(anti-6a -exo and anti-6a -endo), were detected (entries 2
and 3). These cycloadditions required 90 and 60 min for
completion at 0 °C, which reveals that, as expected, the
dipolarophilic reactivity of the methyl derivative 2a is
lower than that of 1a and 1b. The reactions of 2a with
diazomethane afforded similar results, showing a com-
plete stereoselectivity in ether3sonly yielding the adduct
syn-7a swhich strongly decreased when acetonitrile was
used as the solvent (entries 5 and 6).
(13) The data corresponding to this reaction in ref 3 indicate that
anti-5b was exclusively formed under the conditions indicated in entry
6. However, taking into account that compound syn-4b-exo resulted
from the approach of the diazoethane to the face of the dipolarophile
bearing the OEt group at C-5, the absence of a similar adduct (syn-
5b) in the reaction of 1b with the less hindered diazomethane was
difficult to explain. Therefore, we have repeated the later reaction and
thoroughly studied the reaction crude, with the results indicated in
entry 6 of Table 2. This mistake was responsible for the wrong
conclusions reported in ref 3.
The reaction of 2b with an excess of diazoethane
(ethereal solution) at 0 °C afforded a 2:1 mixture of two
6524 J . Org. Chem., Vol. 68, No. 17, 2003

The Role of Steric and Electronic Interactions
F IGURE 3. NOE effects of adducts 6.
TABLE 4. Ad d ition s of Dia zoeth a n e a n d Dia zom eth a n e
to 2b
entry
R
solvent
time (h) T (°C) syn:anti (exo:endo)
1
Me Et2O
Me CH3CN/Et2O
3
3
3
3
0
-40
0
0:100 (66:34)
0:100 (78:22)
0:100
2
3a
4
H
H
Et2O
CH3CN/Et2O
-40
5:95
a
This result was identical with that reported in ref 3.
1∆-pyrazolines, anti-6b-exo and anti-6b-endo (entry 1,
Table 4), and required 3 h for completion, thus revealing
that both reactivity and mainly exo selectivity of 2b were
lower than those for 2a , whereas the π-facial selectivity
was complete but opposite for both substrates. The
increase in the solvent polarity had scarce influence. We
must remark that the formation of the endo-adducts from
compounds 2 only took place for the anti approaches of
the dipole to 2a and 2b, whereas the exo-adducts were
exclusively formed in the syn addition mode. The use of
acetonitrile as the solvent scarcely modified the exo/endo
selectivity. The reactions of 2b with diazomethane (en-
tries 3 and 4, Table 4) exclusively evolved into the adduct
anti-7b in ethyl ether (entry 3, Table 4), but yielded a
95:5 mixture of anti-7b and syn-7b when performed in a
7.5:1 mixture of acetonitrile/ether (entry 4, Table 4).
The separation and purification of adducts 6 or 7 were
not possible by column chromatography, but their char-
acterization was made by NMR analysis of the reaction
crude. The unequivocal configurational assignment of
compounds shown in Figure 3 was performed on the basis
of the observed NOE effects, which also confirmed the
regiochemistry of the adducts.
The stereochemical results from 1a and 1b can only
be explained by assuming a significant role of both
electrostatic and steric factors, which in turn are related
to the configurations at both the sulfur atom and C-5.14
Electrostatic repulsion between sulfinyl and carbonyl
oxygens completely shifts the conformational equilibria
around the C-S bond toward the s-cis rotamer A (Figure
4), which arranges the p-tolyl group toward one of the
diastereotopic faces of the dipolarophile. From a steric
point of view, this orientation makes difficult the ap-
proach of the dipole to such a face. The influence of this
factor must be identical for 1a and 1b, both exhibiting
the same configuration at sulfur. By contrast, the steric
F IGURE 4. Conformational equilibrium of sulfinylfuranones
1 and 2 and substituent effects on diastereofacial selectivity.
influence of the EtO groupsmuch lower than that of the
p-tolyl groupsin the π-facial selectivity must be opposite
that for both compounds, differing at the configuration
at C-5. Thus, we expect a more stereoselective evolution
for compound 1b (the anti approach would be favored by
both chiral centers) than for 1a (each center favors a
different approach, which would counterbalance their
influence). The experimental results show that is not the
case in ethyl ether (at 0 °C, de was higher than 90% for
1a but lower than 70% for 1b), thus suggesting that other
contributions should be considered. The fact that the use
of a more polar solvent (CH3CN) lowers the proportion
of the syn adducts, decreasing the de value in reactions
from 1a but increasing it in those from 1b, suggests an
electrostatic character for these interactions that favors
the syn-approaches in apolar solvents, although the
nature of such interactions is not so obvious.15
The decrease in the π-facial selectivity observed in
reactions of 1a with the change of the solvent polarity is
more marked that the increase observed in reactions of
1b. This can be explained as a consequence of the shifting
of the conformational equilibria toward rotamers B
(Figure 4) in polar solvents, where the electrostatic
repulsion between the sulfinyl and carbonyl oxygens is
lower. As we can see in Figure 4, in the case of compound
1a , the steric effects clearly favor the anti approach to
rotamer B, whose population will only be significant in
polar solvents, which minimizes the importance of elec-
trostatic interactions. As the anti approach to rotamer
A was also favored in polar solvents both tendencies
(15) Taking into account that the carbonated end at the dipole
supports a negative charge, initially we thought that the alkoxy oxygen
could exhibit a possitive charge as a consequence of the strong anomeric
effect present in 5-alkoxy-furan-2-(5H)-ones, but further theoretical
calculations revealed that this was not the case.
(14) This contrasts with the previously reported proposal in ref 3,
which assumed that the π-facial selectivity was exclusively controlled
by the sulfinyl configuration through the steric effects present in the
most stable conformation around the C-S bond.
J . Org. Chem, Vol. 68, No. 17, 2003 6525

Garc´ıa Ruano et al.
TABLE 5. Ad d ition s of Dia zoeth a n e a n d Dia zom eth a n e to 3
entry
1b
solvent
Et2O
time (h)
1.5
T (°C)
-10
R
adducts 8 or 10a
adducts 9a
anti:syn
55:45
regioisomeric ratio
92:8
Me
anti-8-exo (40)
anti-8-endo (7)
syn-8-exo (45)
anti-8-exo (59.6)
anti-8-endo (9.2)
syn-8-exo (29)
anti-8-exo (67)
anti-8-endo (8)
syn-8-exo (24.2)
anti-10 (50)
anti-9-exo (5.5)
anti-9-endo (2.5)
2
3
CH3CN
CH3CN
1.5
3
-10
-40
Me
Me
anti-9-exo (1.4)
anti-9-endo (0.8)
71:29
97.8:2.2
98.7:1.3
anti-9-exo (0.5)
anti-9-endo (0.3)
75.5:24.5
4
5
Et2O
12
12
-10
-10
H
H
-
-
50:50
73:27
100:0
100:0
syn-10 (50)
anti-10 (73)
syn-10 (27)
CH3CN
a
b
Relative proportion determined by 1H NMR of the reaction crude. Combined yield after column chromatography 90% (41% of syn-
8-exo and 33% of anti-8-exo as pure isomers).
superimposed and strongly reduced the facial diastereo-
selectivity of 1a in acetonitrile (Table 1). On the contrary,
in the case of 1b, the evolution of rotamer B must be
scarcely selective (the face sterically favored by each
chiral center is different), which partially compensates
for the higher stereoselective evolution of rotamer A
expected in polar solvents (see above). Consequently, the
facial stereoselectivity of the reactions from 1b is scarcely
affected by the solvent polarity.
The results with 2a and 2b can be explained on the
basis of the lower reactivitysand, therefore, the higher
selectivitysof these substrates, by assuming a model
similar to that used for compounds 1a and 1b, with the
only difference that conformations A must be predomi-
nant in any solvent, because of the strong steric Me/p-
tolyl interaction unstabilizing rotamers B. Thus, both
compounds evolved in a completely stereoselective way
in Et2O. The increase in the solvent polarity destabilized
the syn approaches causing an increase in the proportion
of the anti-adducts, which is only possible for 2a .
At this point, the study of the reactions of 5-alkoxy-
furan-2(5H)-ones with diazoalkanes in solvents of differ-
ent polarity appeared doubly interesting. First, it would
allow us to establish the role of the 5-alkoxy group on
the course of these cycloadditions, thus confirming or not
the conclusions deduced from the above results concern-
ing the relative importance of the steric and electrostatic
grounds in the stereoselectivity control. Second, the
comparison between the results from this study with
those of the sulfinylfuranones 1a and 1b would provide
direct evidence about the role of the sulfinyl group in
these 1,3-dipolar reactions. To our knowledge, the only
reported data are those concerning reactions of diazo-
methane with 5-methoxyfuran-2(5H)-ones16 and 5-(l)-
menthyloxyfuran-2(5H)-one17 in apolar solvents. More-
over, although different stereoselectivities (70/3016 and
55/4517anti/syn ratios) were reported in both papers, they
did not include any comment about them. This prompted
us to study the reactions of diazoethane and diazo-
methane with 5-methoxyfuran-2(5H)-one (3) in ether and
acetonitrile.18 The results are shown in Table 5.
The reaction of diazoethane with 3 required 1.5 h at
-10 °C, affording a mixture of five isomers (entry 1); two
of them, anti-9-endo and anti-9-exo, exhibited the op-
posite regiochemistry to that of the adducts formed from
1 and 2. The major adducts, anti-8-exo and syn-8-exo,
isolated in 33% and 41% yields, were fully characterized.
Both regioselectivity (92:8) and facial stereoselectivity
(55:45) of this reaction were improved in acetonitrile
(entries 2 and 3). The major adduct anti-8-exo obtained
under the conditions of entry 3 was isolated in 60% yield.
The reactions of 3 with diazomethane in ether and
acetonitrile were completely regioselective (entries 4 and
5), whereas the stereoselectivity was almost nonexistent
in Et2O but increased up to 46% de in acetonitrile, with
anti-10 being the major adduct. Regioisomers 8 and 9
were easily separated by chromatography as mixtures
of stereoisomers. Compounds anti-9-exo and anti-9-endo
(16) Farin˜a, F.; Mart´ın, M. V.; Sa´nchez, F. Heterocycles 1986, 24,
2587-2592.
(17) Rispens, M. T.; Keller, E.; de Lange, B.; Zijlstra, R. W. J .;
Feringa, B. L. Tetrahedron: Asymmetry 1994, 5, 607-624 and refer-
ences therein.
(18) These data suggest that the results published in ref 16,
concerning the π-facial selectivity of the reaction of 5-methoxy-2(5H)-
furanone with diazomethane at -10 °C (30:70 syn:anti), could be
wrong. On the other hand, the stereoselectivity reported in ref 17 for
reactions of 5-(l)-menthyloxy-2(5H)-furanone with diazomethane (40:
60 syn:anti) is not consistent with that of ref 16 from a steric point of
view (as the menthyloxy group is bulkier than the methoxy one, a
higher proportion of the anti-adduct must be expected in the former
case).
6526 J . Org. Chem., Vol. 68, No. 17, 2003

The Role of Steric and Electronic Interactions
SCHEME 1
from a steric point of view, clearly favoring the anti
approaches.20 The increase in the π-facial selectivity (anti/
syn ∼ 3:1) observed in acetonitrile suggests that some
type of electrostatic interaction is able to counterbalance
the steric repulsion for the syn approach. Finally, the
high exo/endo ratio observed in reaction with diazoethane
(∼12:1 for the major regioisomer) scarcely modified by
the solvent polarity suggests that the steric factors favor
the exo addition mode (vide infra).
The comparison of the results obtained in the reactions
of diazomethane and diazoethane with compounds 1
(Tables 1 and 2) and 3 (Table 5) allowed us to deduce
the role of the sulfinyl group in the course of these 1,3-
dipolar reactions. It can be summarized as follows:
(a) It strongly increases the reactivity of the dipolaro-
phile. Thus, reactions of 3 with diazoethane and diazo-
methane at -10 °C required 1.5 and 12 h, respectively,
to reach completion (Table 5), whereas they are instan-
taneous from 1a and 1b (Tables 1 and 2).
were spectroscopically characterized from their mixture.
It was also possible for anti-8-endo to be characterized
from the mixture containing the three adducts 8, once
the two major exo adducts had been independently
characterized.
(b) It improves the regiochemistry. Reactions of diazo-
ethane with 1a and 1b (Tables 1 and 2) evolved with
complete regioselectivity, whereas mixtures of regioiso-
mers were obtained starting from 3 (Table 5).
(c) It is the main group responsible for the π-facial
selectivity in low-polarity solvents (it was very high or
even complete for compounds 1 and 2 but scarcely
significant for 3), where the sulfur configuration deter-
mines the face of the furanone ring, which is favored for
the approach of the dipole, regardless the configuration
at C-5.
The most significant parameters in the stereochemical
and regiochemical assignment of adducts 8 and 9 are
depicted in Scheme 1.
The different regiochemistry of the adducts was un-
equivocally established by chemical and spectroscopical
methods. Thus regioisomers 8 exhibit H-3a coupled to
the three protons of the bicyclic fragment, whereas the
equivalent proton at 9 (now designed as H-6a) only shows
one apparent coupling constant with H-3a (J 3a,6a ) 8.1
or 7.5 Hz), because the trans relationship between H-6
and H-6a determines an almost zero value for J 6,6a. As a
confirmation of the so assigned regiochemistry, com-
pounds anti-8-exo and syn-8-exo were independently
transformed into 4-ethyl-5-methoxyfuran-2(5H)-one19 in
almost quantitative yield by heating in toluene for 1 h.
The same compound was obtained as the sole product
starting from the anti-8-exo/syn-8-exo/anti-8-endo mix-
ture. These results unequivocally indicate that the carbon
of the dipole is joined to C-4 at the dipolarophile in
compounds 8. By contrast, under the same reaction
conditions, the anti-9-exo/anti-9-endo mixture remained
unaltered for longer periods of time. This stability toward
the extrusion of the nitrogen could be a consequence of
the change in the regiochemistry. The stereochemistry
syn or anti assigned to compounds 8 could be deduced
from the values of J 3a,4 (<2 Hz for the anti-adducts and
6.4 Hz for the syn-adduct), which are similar to those
previously reported for compounds 10.16 The endo or exo
character was deduced from the value of J 3a,3 (g9 and
e4 Hz, respectively). These rules allowed us to assign
the stereochemistry of compounds 9.
(d) It favors the exo addition mode in reactions with
diazoethane (the exo selectivity is complete for com-
pounds 1a and 1b, whereas a small amount of endo
adduct was formed starting from 3).
The problem of the endo/exo selectivity observed in
reactions with diazoethane deserves additional com-
ments. The reactions of compounds 1a and 1b were
completely exo-selective (Tables 1 and 2). This was also
the case in the syn-approaches of the dipole to compounds
2a , 2b, and 3 (Tables 3-5). The endo-adducts were only
formed as the minor ones in the anti approaches of the
dipole to these later furanones (exo/endo ratios ranged
between 2 and 3 for 2a and 2b, being ∼7 for 3). To explain
these results we assume that the most important factor
controlling the exo-selectivity of these reactions must be
mainly related to the steric interactions of the susbstitu-
ents at C-4 of the furanone ring (C-5 and H for 1 and 3
and C-5 and CH3 for 2) with those of the diazoethane
(Me and H) in the transition state corresponding to each
approach of the dipole to the dipolarophile. During the
approach of the dipole to the face displaying the OR
group, only the exo-adducts were formed from every
substrate (syn-endo adducts were never detected). This
can be explained by taking into account that the steric
interaction of Me/C5-OR, present in the endo modes for
any substrate, is more destabilizing than that for Me/H
(compounds 1 and 3) or Me/C-H (compounds 2) of the exo
modes. During the attack of the dipole to the opposite
face of that supporting the alkoxy group (anti approach)
As can be deduced from Table 5, the regioselectivity
observed in reactions with diazoethane (92:8, entry 1) was
lower than that with diazomethane (100:0, entry 4), as
could be expected from a steric point of view. On the other
hand, the reactivity of MeCHN2 at -10 °C (1.5 h were
required to reach completion) was higher than that of
CH2N2 (12 h), which was not unexpected by FMO theory.
Moreover, the π-facial selectivity, measured as the syn/
anti ratio of the adducts 8, was similar for both dipoles
(45/47 for diazoethane, entry 1, and 50:50 for diazo-
methane, entry 4), and quite lower than that expected
(20) The steric differentiation of the diastereotopic faces at the
furanone ring produced by the OMe group at C-5 was able to control
the π-facial selectivity of their reactions with cyclopentadiene (Feringa,
B. L.; de J ong, J . C. J . Org. Chem. 1988, 53, 1125-1127).
(19) Hoffmann, N.; Buschmann, H.; Raabe, G.; Scharf, H. Tetrahe-
dron 1994, 50, 11167-11186.
J . Org. Chem, Vol. 68, No. 17, 2003 6527

Garc´ıa Ruano et al.
F IGURE 6. Regiochemical pathways.
TABLE 6. Globa l Electr op h ilicity (ω in eV) a n d Loca l
Softn ess (s in eV-1) for Dip oles a n d Dip ola r op h iles
ω
siR
sjR
diazomethane
diazoethane
1a
1b
3
1.405
1.245
1.821
1.821
1.692
0.386a
0.390a
0.044c
0.044c
0.040c
0.573b
0.497b
0.087d
0.087d
0.066d
a
b
d
R ) -, i ) N. R ) -, j ) C. c R ) +, i ) 3. R ) +, j ) 4.
Table 6 presents the most significant global and local
indices for the diazoalkanes and the dipolarophiles 1a ,
1b, and 3 computed at the B3LYP/6-31G(d) level. The ω
values indicate that the 1,3-DC of diazoalkanes and the
5-alkoxyfuranones can be classified as Sustmann Type I
processes24 in which a net charge transfer from the dipole
to the dipolarohile occurs in the cycloaddition. The
difference in the reactivity of 1 and 3 toward diazoal-
kanes could be explained in terms of electrophilicity (ω).
Higher ω values make the dipolarophile more electro-
philic and therefore more reactive. The opposite is also
true for the different reactivity of diazoalkanes, that is,
lower ω values imply higher nucleophilicity of the dipole
and consequently higher reactivity (diazoethane > diazo-
methane).
F IGURE 5. Approach of diazoethane to furanones 1-3.
the exo mode is more stable than the endo one in the
case of compounds 1 and 3 (Me/H and H/C5-H interac-
tions are preferred with respect to Me/C5-H and H/H,
Figure 5), whereas the interactions of both modes are
similar for compounds 2 (Me/C-H and H/C5-H or H/C-H
and Me/C5-H). The preference for the exo-adducts in the
later case could also be due to the higher rigidity of the
cyclic C5-H fragment with respect to the exocyclic CH3.
As on the basis of the experimental data we have not
been able to establish the nature of the electrostatic
interactions that have some influence on the π-facial
selectivity of these reactions, we decided to explore these
reactions from a theoretical point of view, to obtain some
answer to this problem. Moreover, the increase in the
reactivity and regioselectivity produced by the sulfinyl
group experimentally observed for all these reactions
could also be evaluated by analyzing the global and local
indices of the reactants. The density functional theory
formulation of Parr, Pearson, and Yang (PPY-DFT)21 has
introduced a number of global descriptors of electronic
structure, such as the electronic chemical potential (µ),
chemical hardness (η) and chemical softness (s), and
electrophilicity index (ω),22 providing modern chemistry
with more quantitative concepts.
Recently, Domingo et al. have created a global elec-
trophilicity scale,23 based on the ω values, for dienes and
dienophiles commonly used in Diels-Alder reactions
(DA),23a as well as dipoles and dipolarophiles used in 1,3-
DC reactions.23b Within this scale, 1,3-DC reactions
between dipoles with a low ω value (marginal electro-
philes or nucleophiles) and dipolarophiles with high ω
values (strong electrophiles) correspond with normal
electron demand cycloaddition (NED) processes and,
accordingly, with a charge transfer from the dipole to the
dipolarophile. The opposite interactions (electrophilic
dipoles with nucleophilic dipolarophiles) are classified as
inverse electron demand (IED) 1,3-DC reactions. We have
analyzed our 1,3-dipolar reactions on the same basis.
Besides the definition of global descriptors, PPY-DFT
provided local reactivity indices derived directly from the
ground-state electron density. The local softness, skR
(local softness at atom k with R being either -, for an
electrophilic attack, or +, for a nucleophilic one), is one
of the most common descriptors of site reactivity.
In the case of a diene or dipole (A) that reacts with a
dienophile or dipolarophile (B), two regiochemical path-
ways are possible (Figure 6). The regioselectivity in
cycloaddition reactions with two bond-forming interac-
tions has been recently studied by means of the local hard
and soft acids and bases (HSAB) principle.25 In these
studies, successful prediction of the regioselectivity was
obtained by a local softness matching square sum crite-
rion, which can be expressed as follows: pathway I is
preferred over pathway II when
(sA- - sB+ )2 + (sA- - s+B )2 < (sA- - sB+ )2 +
1
1
2
2
1
2
-
-
+
-A2
B2
-A2
B2 B1
(sA- - sB+ )2 w ∆sBA + < ∆sA
(1)
1
1
+
+
2
1
1
(24) (a) Sustmann, R. Tetrahedron Lett. 1971, 2717-2720. (b)
Sustmann, R.; Trill, H. Angew. Chem., Int. Ed. Engl. 1972, 11, 838-
840. (c) Sustmann, R. Pure Appl. Chem. 1974, 40, 569-593.
(25) (a) Damon, S.; Van de Woude, G.; Me´ndez, F.; Geerlings, P. J .
Phys. Chem. A 1997, 101, 886-893. (b) Sengupta, D.; Chandra, A. K.;
Nguyen, M. T. J . Org. Chem. 1997, 62, 6404-6406. (c) Chandra, A.
K.; Nguyen, M. T. J . Comput. Chem. 1998, 19, 195-202. (d) Chandra,
A. K.; Nguyen, M. T. J . Phys. Chem. A 1998, 102, 6181-6185. (e) Le,
T. N.; Nguyen, L. T.; Chandra, A. K.; De Proft, F.; Geerlings, P.;
Nguyen, M. T. J . Chem. Soc., Perkin Trans. 2 1999, 1249-1255. (f)
Chandra, A. K.; Uchimaru, T.; Nguyen, M. T. J . Chem. Soc., Perkin
Trans. 2 1999, 2117-2121.
(21) (a) Parr, R. G.; Yang, W. Density Functional Theory of Atoms
and Molecules; Oxford University Press: New York, 1989. (b) Parr, R.
G.; Pearson, R. G. J . Am. Chem. Soc. 1983, 105, 7512-7516.
(22) Parr, R. G.; von Szentpa´ly, L.; Liu, S. J . Am. Chem. Soc. 1999,
121, 1922-1924.
(23) (a) Domingo, L. R.; Aurell, M. J .; Pe´rez, P.; Contreras, R.
Tetrahedron 2002, 58, 4417-4423. (b) Pe´rez, P.; Domingo, L. R.; Aurell,
M. J .; Contreras, R. Tetrahedron 2003, 59, 3117-3125.
6528 J . Org. Chem., Vol. 68, No. 17, 2003

The Role of Steric and Electronic Interactions
TABLE 7. Regioch em ica l An a lysis in Ter m s of
TABLE 8. En er gy Ba r r ier s for th e Tr a n sition
Str u ctu r es a n d Cycloa d d u cts 10 a t Differ en t Levels of
Ca lcu la tion a s w ell a s th e En er gy Differ en ce An ti-Syn
in th e Tr a n sition Str u ctu r es (k ca l/m ol)
∆s Va lu es (in 10-3 eV-1
)
pathway
diazomethane
2.61
diazoethane
2.47
-C-
4
I (∆sN3
)
+
1a
+
syn-10
anti-10
TS10
-C-
3
II (∆sN4
I-II
)
)
)
5.98
-3.36
2.59
4.51
ꢀa
TS adduct Ts adduct (anti-syn)
+
+
-2.04
-C-
B3LYP/6-31G(d)
1.00 13.2 -24.2 14.6 -24.7
4.20 10.0 -35.4 10.9 -35.9
35.94 9.8 -37.7 10.4 -38.1
1.00 15.1 -20.7 16.4 -21.6
4.20 12.0 -31.9 12.7 -32.9
35.94 11.8 -34.2 12.1 -35.3
1.4
0.9
0.6
1.3
0.7
0.3
1.2
0.7
0.3
1.3
0.5
-0.4
1.0b
0.9
0.8
I (∆sN3
)
+
1b
3
2.45
+
4
-C-
II (∆sN4
I-II
5.94
-3.35
4.84
4.48
+
+
3
-2.03
B3LYP/6-31++G(d,p)
-C-
I (∆sN3
)
+
4.33
+
4
-C-
II (∆sN4
I-II
6.89
-2.05
5.57
+
+
3
B3LYP/6-31++G(d,p)// 1.00 15.2 -20.6 16.4 -21.5
B3LYP/6-31G(d)
-1.24
4.20 12.1 -32.0 12.8 -32.9
35.94 12.0 -34.6 12.3 -35.3
1.00 16.9 -15.4 18.2 -16.2
4.20 11.4 -27.9 11.9 -26.0
35.94 10.1 -30.7 9.7 -27.6
1.00
B3LYP/cc-pVTZ//
B3LYP/6-31G(d)
for a NED cycloaddition process or when
(sA+ - sB- )2 + (sA+ - s-B )2 < (sA+ - sB- )2 +
MP2/6-31++G(d,p)
1
1
2
2
1
2
4.20
35.94
+
+
+A2
B2
+A2
B2 B1
(sA+ - sB- )2 w ∆sBA - < ∆sA
(2)
1
1
-
-
-
2
1
1
a
ꢀ ) dielectric constant (1.00, gas phase; 4.20, Et2O; 35.94,
b
CH3CN). At the HF/6-31++G(d,p) level, TS10(syn - anti) ) 1.2
kcal/mol.
for an IED one.
In the case of our 1,3-DC, we name pathway I for the
regiochemical orientation in which N(dipole)-C3(dipolarophile)
and C(dipole)-C4(dipolarophile) bonds are created and, therefore,
pathway II for the orientation with an opposite bond
forming pattern. The regiochemical preference can be
predicted with the aid of eq 1 and the results presented
in Table 7. In all cases, pathway I is preferred over
pathway II, as indicated by the lower ∆s values of the
former. The presence of the sulfoxide group in position 3
of the dipolarophile enhances the regiochemical prefer-
ence for pathway I and, in addition, a methyl substituent
in the dipole (diazoethane) leads to lower preference for
the same pathway. Hence, the reaction of diazoethane
with 3 presents the lowest of all values for the regio-
chemical preference for pathway I (pathway I-II in Table
7) in agreement with the experimental results that
correspond with the only case in which regioisomers from
pathway II were detected.
The global and local PPY-DFT reactivity indices have
been shown as very useful tools in the study of the
reactivity and regioselectivity of these 1,3-DC. However,
the stereochemical outcome must be studied by examina-
tion of the stationary points along the reaction path.
Diastereoselectivity in 1,3-DC of 1 and 2 with diazoal-
kanes is strongly dependent on the polarity of the solvent,
thus suggesting a significant contribution of some kind
of electrostatic interactions involving the alkoxy group
at C-5 favoring the syn-approaches more than expected
from a steric point of view. To shed light on the nature
and scope of this effect, we started our study with the
reaction of diazomethane with furanone 3 (see Table 5)
for which this is the only factor affecting π-facial selectiv-
ity of the cycloaddition.
with the increase in solvent polarity but this effect is
higher for the cycloadducts (higher exothermicity) and,
therefore, the energy barriers for the reverse process
raise. Solvent effects are also involved in the stereochem-
istry of the addition, thus the syn approach is less favored
with increasing polarity of the medium, in agreement
with the experimental results. However, at this level of
calculation, syn-10 is always the most favored adduct
whereas ∼1:1 and ∼3:1 anti:syn ratios were experimen-
tally obtained in Et2O and CH3CN, respectively. In
conclusion, the B3LYP/6-31G(d) method reproduces the
observed directing effect of the alkoxy group in position
5 but its magnitude is overestimated.
The results with a higher basis set that includes diffuse
and polarization functions on first row and hydrogen
atoms (6-31++G(d,p)) exhibit a similar trend but with a
slightly lower syn preference in the gas phase and a
slightly higher solvent effect. However, in the most polar
solvent (CH3CN) the syn-adduct is still favored by 0.3
kcal/mol over the anti-adduct.
The lower stabilization of the transition structures and
adducts with respect to the reactants found with this
higher basis set suggests that a basis set superposition
error (BSSE) could be significant and responsible for the
overestimation of the 5-alkoxy directing effect. We then
decided to use a more complete basis set but through
single-point energy calculations. As a previous step, we
carried out single-point energy calculations at the B3LYP/
6-31++G(d,p) level (for which we have the results of the
full optimization) on the B3LYP/6-31G(d) optimized
geometries and used the zero-point energy correction
(ZPEC) at this latter level. This procedure allowed us to
validate B3LYP/6-31G(d) geometries for use in single-
point energy calculations at higher levels. The obtained
electronic energies plus ZPEC of all the structures are
between 0.1 and 0.6 kcal/mol higher than those obtained
with the B3LYP/6-31++G(d,p) geometries. These values
are slightly lower in solution, from 0.0 to 0.2 kcal/mol in
Et2O, and from 0.0 to 0.4 kcal/mol in CH3CN. However,
and most significantly, the energy differences (TS-
Table 8 summarizes the results of our study in the
reaction of diazomethane with 3. At all the DFT levels
of calculation, the reaction is predicted to be exothermic
with 32.3 kcal/mol as the lowest energy barrier for the
reverse process and, because the experimental results
were taken at temperatures below 0 °C, these are 1,3-
DC under kinetic control. At the B3LYP/6-31G(d) level,
there is an important preference in the gas phase for the
syn stereochemistry over the anti one (1.4 kcal/mol).
Energy barriers for the transition structures are reduced
J . Org. Chem, Vol. 68, No. 17, 2003 6529

Garc´ıa Ruano et al.
TABLE 9. Sign ifica n t P a r a m eter s (ch a r ges, q, in e;
d ista n ces, d , in Å; a n d bon d or d er s, BO) for TS-syn -10
a n d TS-a n ti-10
complexes reveals the presence of this kind of interaction
controlling the stereochemistry of the cycloaddition. It
should be noted that the direct comparison of the TS with
the isolated reactants can lead to more profound changes
than expected because in our case, both molecules are
interacting in a cycloaddition process involving the atoms
to which H and O are bonded. Hence, in this case, an
optimal agreement is accounted for by comparing syn and
anti approaches and even Hendo and Hexo in the same syn
approach. In the above-mentioned paper, Scheiner et al.
reported a C-H bond contraction in the CH‚‚‚O com-
plexes between 0.2 and 2.9 mÅ and, in our case, the
C-Hendo distance in TS-syn-10 is ca. 3-4 mÅ shorter
than the other C-H distances in the dipole fragment of
the TS’s. Additionally, CH‚‚‚O complexes exhibit en-
hancements in the natural charges of the H and O atoms,
between 16 and 33 me and between -1 to -23 me,
respectively. Hendo in TS-syn-10 present a natural charge
value ∼25 me higher than the other H atoms in the dipole
fragment of the TS’s, whereas the O atom at C-5 of the
dipolarophile in TS-syn-10 presents a natural charge 11
me lower than that in TS-anti-10. This result refers to
the concept of induction; the proximity of a negative
charge to the H atom in diazomethane induces a polar-
ization in the molecule that rises its positive charge,
whereas the proximity of a positive charge to the O atom
at C-5 of the furanone has the opposite effect. In contrast,
the trend in ChelpG charges is just the opposite that in
natural charges, which could easily be understood if this
type of charge is visualized as the electrostatic effect
measured at a certain distance from the nucleus. There-
fore, the proximity of a charge with different sign tends
to neutralize the effects, and accordingly, ChelpG charges
should be lower in absolute value when H and O atoms
are involved in CH‚‚‚O hydrogen bonding.
An NBO analysis of the transition structures also
shows the presence of the CH‚‚‚O hydrogen bond by the
weakening of the C-H bond (lower bond order in ∼0.011),
appreciable H‚‚‚O bond order (0.009), and the significant
value for the nO f σ/C-H second-order interaction, 1.6
kcal/mol, which is very close to the TS10 anti - syn
energy difference of 1.4 kcal/mol (Table 8).
The reactions of diazomethane with 1a ,b (Scheme 2)
were studied following the methodology that provided the
best results in the preceding case, i.e. full geometry
optimization and frequency analysis of the reactants,
transition structures, and cycloadducts at the B3LYP/6-
31G(d) level and subsequent single-point energy calcula-
tions at the B3LYP/cc-pVTZ level. The results of these
calculations are presented in Table 10. The energy
barriers for the 1,3-DC are lower than those found for
the 3-unsubstituted furanone 3 and the reactions are
more exothermic, thus showing the higher reactivity of
the 3-sulfinylfuranones.
TS-syn-10
TS-anti-10
NNCH2
H
3
O-5 Hendo Hexo O-5 Hendo Hexo O-5
a
qChelpG
qnatural
dH-C
0.287 -0.386 0.072 0.092 -0.283 0.096 0.136 -0.442
0.243 -0.570 0.281 0.254 -0.588 0.256 0.258 -0.577
b
1.080
1.085 1.089
0.878 0.899
2.235
1.088 1.089
0.899 0.898
2.387
BOH-C 0.904
c
dH-X5
d
BOH-O
qtransfer
0.009 0.000
0.210 0.184
0.001 0.000
e
a
b
ChelpG charges (see Computational Methods). Natural
charges. c X5 ) substituent at C5 in 3 (O in the syn approach and
d
H in the anti one). Intermolecular Hdipole - Odipolarophile
.
e Charge
transfer from the dipole to the dipolarophile in the transition
structures (natural charges).
reactants and adducts-reactants) and stereochemical
preferences are almost the same with both procedures
(Table 8).
Our choice for a more complete basis set was the
Dunning’s triple-ú one, cc-pVTZ, which includes polariza-
tion functions on all the atoms. In accordance with the
observation in the case of 6-31++G(d,p), a more complete
basis set raises the energies of transition structures and
products with respect to the reactants. Finally, the most
important consequence is present in the solvation effects
on the stereoselectivity of the process. Now, the formation
of adduct syn-10 is favored in Et2O, but with the lowest
preference of all (0.5 kcal/mol), whereas the formation
of anti-10 is favored in CH3CN by 0.4 kcal/mol. These
are the theoretical results in best agreement with the
experimental ones, but this methodology still slightly
overestimates the directing effect of the 5-alkoxy group
vicinal to the CC double bond.
Experimental results are successfully reproduced by
DFT methods with a sufficiently large basis set. At this
point, we decided to test ab initio methods versus DFT
ones toward the problem of the diastereoselectivity of the
cycloaddition. Full geometry optimization and frequency
analysis of TS-syn-10 and TS-anti-10 were carried out
at the HF/6-31++G(d,p) level, and taking these geom-
etries as the starting point, subsequent optimization at
the correlated MP2/6-31++G(d,p) level was performed.
The results of these calculations are presented in Table
8. The values for the stereochemical preference of the
reaction are very similar to those found with the DFT
methods but with a slightly lower syn preference in the
gas phase and lower solvent effects.
To establish the nature of the 5-alkoxy directing effect,
we have explored different magnitudes in the reactants
and the transition structures calculated at the B3LYP/
6-31G(d) level (Table 9). Hydrogen atoms in diazometh-
ane present notable positive charges while the oxygen
atom at C-5 in 3 is negatively charged. These two atoms
are the closest nondirectly bonded ones in TS-syn-10 so
an attractive Coulombic interaction could justify the
5-alkoxy effect.
The syn addition is greatly favored over the anti one
(3.0-3.4 kcal/mol) in the reaction of diazomethane with
1a in agreement with the high stereoselectivity found in
the experiments when the reaction was carried out in
the least polar solvent. On the other hand, in the reaction
of diazomethane with 1b, the stereochemical preference
is significantly lower: syn addition is slightly favored
over anti (0.3 kcal/mol) with the 6-31G(d) basis set, while
anti addition is the favored one (0.3 kcal/mol) with the
cc-pVTZ. In this case, again calculations with the shortest
Comparison of the values presented in Table 9 with
the ones reported by Scheiner et al.26 for several CH‚‚‚O
(26) Gu, Y.; Kar, T.; Scheiner, S. J . Am. Chem. Soc. 1999, 121, 9411-
9422.
6530 J . Org. Chem., Vol. 68, No. 17, 2003

The Role of Steric and Electronic Interactions
SCHEME 2
TABLE 10. En er gy Ba r r ier s for th e Tr a n sition
Str u ctu r es a n d Cycloa d d u cts 5a ,b a t Differ en t Levels of
Ca lcu la tion (k ca l/m ol)
interaction. As in the case of TS-10, several pieces of
evidence support this hydrogen bonding (vide supra): (a)
The higher absolute values of the natural charges for the
H and O atoms in the syn-TS while ChelpG charges
exhibit lower absolute values; (b) shortening and weak-
ening of the C-H bond in the syn-TS over the anti one;
and (c) appreciable H‚‚‚O bond order.
In addition, the NBO analysis revealed significant
values for the nO f σ/C-H interactions, 1.16 kcal/mol in
TS-syn-5a and 0.79 kcal/mol in TS-syn-5b.
B3LYP/cc-pVTZ//
B3LYP/6-31G(d)
B3LYP/6-31G(d)
syn-5a
anti-5a
syn-5b
anti-5b
TS
adduct
TS
adduct
TS
adduct
TS
10.3
-25.4
13.3
-24.8
11.6
13.8
-16.3
17.2
-16.2
15.7
-24.9
11.9
-15.8
15.4
It should be noted that these calculations where carried
out only for the gas-phase processes since the two factors
controlling the stereochemistry are affected by solvent
effects. On one hand, increasing solvent polarity stabi-
lizes charges and therefore weakens CH‚‚‚O interactions.
On the other hand, in high-polarity solvents, the confor-
mational preference of the rotamers with the S f O bond
in an s-cis arrangement with respect to the dipolarophilic
double bond (rotamer A in Figure 4) must be lower, thus
increasing the population of the other possible conformers
around the C3-S (rotamers B in Figure 4) that exhibit
a less significant influence of the bulky p-tolyl group. This
fact complicates the calculations of the reactions of
diazomethane with 1a or 1b since, to take into account
all the effects, three conformations should be studied for
each one of the structures presented in Scheme 2. Due
to the size of the problem and the high number of
structures involved, the required computational cost
exceeds the aim of this work. Nevertheless, the solvent
polarity effects on the outcome of these reactions could
be rationalized from a qualitative point of view as
previously indicated (see above).
adduct
-25.2
-16.2
basis set tend to overestimate the magnitude of the
5-alkoxy directing effect. To sum up, calculations at the
B3LYP/cc-pVTZ//B3LYP/6-31G(d) reveal that the stere-
ochemistry of the addition is mainly governed by the
configuration of the S atom in position 3 (steric effects,
SE) and it is modulated by the configuration in C5
(CH‚‚‚O hydrogen bond, CHO). Assuming that these two
effects are operating independently, the anti-syn energy
difference in TS-5a consists of the sum of both effects
favoring the syn addition
∆E(TS-5a )anti-syn ) SE + CHO ) 3.4
(3)
whereas in TS-5b, steric effects favor the anti-adduct and
the 5-alkoxy directing effect favors the syn one
∆E(TS-5b)anti-syn ) SE - CHO ) 0.3
(4)
which yields a value of 1.85 kcal/mol for the steric effects
(SE) and 1.55 kcal/mol for the 5-alkoxy directing effect
(CHO). The latter is close to the anti-syn energy differ-
ence for TS-10 in the gas phase (1.3 kcal/mol).
Con clu sion s
Table 11 presents the most significant parameters in
TS-5a and TS-5b allowing the analysis of the CH‚‚‚O
We can conclude that the reactions of diazoalkanes
with 5-alkoxy-3-p-tolylsulfinylfuran-2(5H)-ones are quite
J . Org. Chem, Vol. 68, No. 17, 2003 6531

Garc´ıa Ruano et al.
TABLE 11. Sign ifica n t P a r a m eter s (ch a r ges, q, in e; d ista n ces, d , in Å; a n d bon d or d er s, BO) for TSs 5a a n d 5b
TS-syn -5a
TS-a n ti-5a
TS-syn-5b
TS-anti-5b
NNCH2
H
1a
O-5
1b
O-5
Hendo
O-5
Hendo
O-5
Hendo
O-5
Hendo
O-5
a
qChelpG
qnatural
DH-C
0.287
0.243
1.080
0.904
-0.389
-0.562
0.085
0.280
1.084
0.880
2.302
0.006
0.288
-0.268
-0.598
0.141
0.260
1.088
0.899
2.504
0.000
0.251
-0.490
-0.585
-0.455
-0.576
0.073
0.281
1.084
0.882
2.331
0.005
0.284
-0.247
-0.599
0.141
0.258
1.087
0.898
2.507
0.000
0.262
-0.546
-0.585
b
BOH-C
c
dH-X5
d
e
BOH-O
qtransfer
a
b
ChelpG charges (see Computational Methods). Natural charges. c X5 ) substituent at C5 in 3 (O in the syn approach and H in the
d
anti one). Intermolecular Hdipole - Odipolarophile
charges).
.
e Charge transfer from the dipole to the dipolarophile in the transition structures (natural
(3R,3a S,4S,6a R,SS)-4-E t h oxy-3-m et h yl-6a -[(4-m et h yl-
p h en yl)su lfin yl]-3,3a ,4,6a -tetr a h yd r o-6H-fu r o[3,4-c]p yr a -
zol-6-on e (syn -4a -exo). syn-4a -exo was obtained from 1a and
diazoethane, after 5 min of reaction time. It was isolated as a
white solid by column chromatography (silica gel was previ-
ously treated with triethylamine), 50:45:5 hexane/dichlo-
interesting in asymmetric synthesis because they evolve
in high yields under mild conditions affording bicyclic
pyrazolines with complete regioselectivity, exo-selectivity,
and very high π-facial selectivity, which can be modu-
lated, becoming almost complete, with the solvent polar-
ity. The sulfinyl group increases the dipolarophilic reac-
tivity of the butenolides, as well as their regioselectivity
and exo-selectivity. Moreover it is the main group re-
sponsible for the π-facial selectivity, but electrostatic
interactions between dipoles and the alkoxy group at C-5
are also significant in apolar solvents. Finally, the steric
interactions between the substituents at diazoethane and
at C-4 of the furanone rings are the main reasons for the
observed exo-selectivity. The results obtained in the
theoretical study of these 1,3-dipolar reactions, making
use of the density functional theory of Parr, Pearson, and
Yang (PPY-DTF), are consistent with the experimentally
observed behavior and suggest that the stabilizing hy-
drogen bonds between the negatively charged alkoxy
oxygen at C-5 of the furanone and the positively charged
hydrogen atoms at the dipole almost compensate for the
steric interactions associated with the syn approaches of
both reagents.
romethane/diethyl ether. Yield 80%; mp 91-93 °C dec. [R]20
D
1
+416.0 (c 0.5, acetone); IR (KBr) 1756, 1595, 1084, 1060; H
NMR δ 7.60 and 7.38 (AA′BB′ system, 4H), 5.40 (qd, J ) 7.3
and 3.7, 1H), 4.68 (d, J ) 6.7, 1H), 3.68 (m, 1H), 3.44 (m, 1H),
2.73 (dd, J ) 6.7 and 3.7, 1H), 2.45 (s, 3H), 1.57 (d, J ) 7.3,
3H), 1.13 (t, J ) 7.1, 3H); 13C NMR δ 163.5, 143.7, 134.1, 130.4,
124.7, 122.3, 102.2, 88.0, 66.8, 41.7, 21.4, 17.9, 14.4. Anal.
Calcd for C15H18N2O4S: C, 55.88; H, 5.63; N, 8.69; S, 9.95.
Found: C, 55.75; H, 5.39; N, 8.61; S, 10.41.
(3S,3a R,4S,6a S,SS)-4-E t h oxy-3-m et h yl-6a -[(4-m et h yl-
p h en yl)su lfin yl]-3,3a ,4,6a -tetr a h yd r o-6H-fu r o[3,4-c]p yr a -
zol-6-on e (a n ti-4a -exo). anti-4a -exo was obtained along with
syn-4a -exo by addition of diazoethane to a solution of 1a in
acetonitrile. It could not be isolated as a pure compound.
Spectroscopic data for anti-4a -exo correspond to those of the
crude reaction mixture (entries 4 and 5, Table 1). IR (film)
1771, 1595, 1493, 1088, 1061; 1H NMR δ 7.53 and 7.33 (AA′BB′
system, 4H), 5.19 (d, J ) 1.6, 1H), 4.84 (qd, J ) 7.4 and 3.7,
1H), 3.92 (m, 1H), 3.66 (m, 1H), 2.59 (dd, J ) 3.7 and 1.6, 1H),
2.41 (s, 3H), 1.28 (t, J ) 7.1, 3H), 0.93 (d, J ) 7.4, 3H); 13C
NMR δ 164.0, 143.5, 134.2, 129.6, 125.9, 118.1, 107.1, 93.4,
66.2, 46.1, 21.4, 17.2, 14.7.
Exp er im en ta l Section
(3aR,4S,6aS,SS)-4-Eth oxy-6a-[(4-m eth ylph en yl)su lfin yl]-
3,3a ,4,6a -tetr a h yd r o-6H-fu r o[3,4-c]p yr a zol-6-on e (a n ti-
5a ). anti-5a was obtained along with syn-5a by addition of
diazomethane to a solution of 1a in acetonitrile. It could not
be separated from its stereoisomer by column chromatography
(60:35:5 hexane/dichloromethane/diethyl ether). The data were
obtained from a 59:41 mixture of anti-5a and syn-5a (entry 8,
Gen er a l Meth od s. Flash chromatography was performed
with silica gel 60 (230-400 mesh), and silica gel F254 plates
were used for preparative TLC. The IR spectra frequencies
are gives in cm-1. NMR spectra were determined in CDCl3
solutions at 300 (or 200) and 75.5 (or 50.3) MHz for 1H and
13C NMR, respectively; chemical shifts (δ) were reported in
ppm and J values are gives in hertz. Optical rotations were
obtained at ambient temperature in the solvent and concen-
tration (g/100 mL) indicated in each case. Diazomethane and
diazoethane were synthesized according to the procedure
described in ref 27.
Table 1). [R]20 +411.6 (c 1.15, CHCl3); IR (film) 1773, 1595,
D
1
1492, 1087, 1061; H NMR δ 7.46 and 7.27 (AA′BB′ system,
4H), 5.03 (d, J ) 2.8, 1H), 4.75 (dd, J ) 19.2 and 2.5, 1H),
3.92 (m, 1H), 3.88 (dd, J ) 19.2 and 9.1, 1H), 3.68 (m, 1H),
3.12 (dt, J ) 9.1 and 2.6, 1H), 2.41 (s, 3H), 1.27 (t, J ) 7.1,
3H); 13C NMR δ 163.9, 143.5, 133.6, 129.6, 125.6, 118.7, 107.8,
84.2, 66.6, 38.8, 21.5, 14.7. Anal. Calcd for C14H16N2O4S: C,
54.53; H, 5.23; N, 9.08; S, 10.40. Found: C, 54.65; H, 5.41; N,
8.99; S, 10.95.
Gen er a l P r oced u r e for th e Ad d ition of Dia zoa lk a n es
to F u r a n -2(5H)-on es. To a solution of furanone (0.40 mmol)
in diethyl ether or acetonitrile (11 mL for 1a , 1b, 2b, and 3 or
20 mL for 2a ), cooled at the temperature indicated in Tables
1-5, was added an ethereal solution of diazoethane or diazo-
methane (1.5 mL, containing 0.6 mmol/mL). The reaction was
kept at the same temperature during the period indicated in
each case. The solvent was removed and the residue was
(3S,3a R,4R,6a S,SS)-4-E t h oxy-3-m et h yl-6a -[(4-m et h yl-
p h en yl)su lfin yl]-3,3a ,4,6a -tetr a h yd r o-6H-fu r o[3,4-c]p yr a -
zol-6-on e (syn -4b-exo). syn-4b-exo was obtained as the minor
product by reaction of 1b with diazoethane (entry 1, Table 2).
It was isolated as an oil by column chromatography (silica gel
was previously treated with triethylamine), 60:35:5 hexane/
dichloromethane/diethyl ether, and could not be thoroughly
purified. Yield 10%. IR (film) 1775, 1595, 1493, 1088, 1059;
1H NMR δ 7.44 and 7.33 (AA′BB′ system, 4H), 5.61 (d, J )
6.6, 1H), 5.22 (qd, J ) 7.4 and 4.4, 1H), 3.89 (m, 1H), 3.67 (m,
1H), 2.65 (dd, J ) 6.6 and 4.4, 1H), 2.41 (s, 3H), 1.21 (t, J )
1
analyzed by H NMR (Tables 1-5) and purified as indicated
in each case.
The data for adducts syn-5a , anti-5b, syn-7a , and anti-7b
have been reported previously in ref 3 and the data for the
adducts 10 in ref 16.
(27) Arndt, F. Organic Syntheses; Wiley: New York, 1941; Collect.
Vol. II, pp 165-167.
6532 J . Org. Chem., Vol. 68, No. 17, 2003

The Role of Steric and Electronic Interactions
7.1, 3H), 0.70 (d, J ) 7.4, 3H); 13C NMR δ 164.7, 143.7, 133.9,
(3aR,4R,6aS,SS)-4-Eth oxy-3a-m eth yl-6a-[(4-m eth ylph en -
yl)su lfin yl]-3,3a ,4,6a -tetr a h yd r o-6H-fu r o[3,4-c]p yr a zol-6-
on e (syn -7b). syn-7b was obtained as the minor adduct by
addition of diazomethane to 2b in acetonitrile. The spectro-
scopic data for syn-7b correspond to those of the crude reaction
129.7, 125.4, 120.1, 102.5, 86.9, 67.1, 42.8, 21.5, 17.0, 14.7.
(3R,3a S,4R,6a R,SS)-4-Eth oxy-3-m eth yl-6a -[(4-m eth yl-
p h en yl)su lfin yl]-3,3a ,4,6a -tetr a h yd r o-6H-fu r o[3,4-c]p yr a -
zol-6-on e (a n ti-4b-exo). anti-4b-exo was obtained as the
major adduct by reaction of 1b with diazoethane, although it
could not be isolated diastereoisomerically pure. This com-
pound was characterized by its spectroscopic data from the
crude reaction mixture (entry 3, Table 2). IR (film) 1768, 1596,
1
mixture. H NMR δ 7.70 and 7.37 (AA′BB′ system, 4H), 5.22
and 4.23 (AB system, J ) 18.1, 2H), 5.10 (s, 1H), 3.71 (m, 1H),
3.51 (m, 1H), 2.45 (s, 3H), 1.78 (s, 3H), 1.10 (t, 3H); 13C NMR
δ 164.5, 143.7, 133.0, 129.1, 128.2, 109.7, 106.7, 85.6, 66.0, 50.1,
21.6, 15.5, 14.5.
1
1493, 1086, 1061; H NMR δ 7.59 and 7.34 (AA′BB′ system,
4-Meth oxy-3-m eth yl-3,3a ,4,6a -tetr a h yd r o-6H-fu r o[3,4-
c]p yr a zol-6-on e (a n ti-8-exo). anti-8-exo was obtained by
reaction of 3 with diazoethane. It was isolated by column
chromatography (1:2 ethyl acetate/hexane) as an oil. IR (film)
4H), 5.04 (d, J ) 1.5, 1H), 4.97 (qd, J ) 7.2 and 3.4, 1H), 3.31
(m, 2H), 2.44 (dd, J ) 3.4 and 1.5, 1H), 2.42 (s, 3H), 1.60 (d,
J ) 7.2, 3H), 0.83 (t, J ) 7.1, 3H); 13C NMR δ 163.3, 143.1,
134.0, 129.9, 125.4, 122.7, 107.0, 93.8, 65.4, 44.4, 21.3, 18.2,
14.3.
1
1778, 1555, 1176, 1112; H NMR δ 5.73 (dd, J ) 8.8 and 2.4,
1H), 5.04 (d, J ) 1.6, 1H), 4.83 (m, 1H), 3.44 (s, 3H), 2.37 (ddd,
J ) 8.8, 4.1, and 1.6, 1H), 1.46 (d, J ) 7.3, 3H);13C NMR δ
167.1, 108.1, 92.5, 91.7, 57.1, 46.6, 18.5. Anal. Calcd for
C7H10N2O3: C, 49.41; H, 5.92; N, 16.46. Found: C, 49.11; H,
5.69; N 16.19.
4-Meth oxy-3-m eth yl-3,3a ,4,6a -tetr a h yd r o-6H-fu r o[3,4-
c]p yr a zol-6-on e (a n ti-8-en d o). anti-8-endo was obtained as
the minor product by reaction of 3 with diazoethane and could
not be fully purified by column chromatography (1:2 ethyl
acetate/hexane). 1H NMR (from a 80:20 mixture of anti-8-endo/
anti-8-exo) δ 5.64 (dd, J ) 9.1 and 1.3, 1H), 5.13 (d, J ) 2.0,
1H), 4.96 (m, 1H), 3.51 (s, 3H), 2.81 (td, J ) 9.2 and 2.0, 1H),
1.61 (d, J ) 7.6, 3H); 13C NMR δ 168.0, 104.2, 93.2, 88.1, 57.3,
41.9, 13.6.
(3aR,4R,6aS,SS)-4-Eth oxy-6a-[(4-m eth ylph en yl)su lfin yl]-
3,3a ,4,6a -t et r a h yd r o-6H -fu r o[3,4-c]p yr a zol-6-on e (syn -
5b). syn-5b was obtained as the minor product by reaction of
1b with diazomethane (entry 6, Table 2). It was isolated as a
white solid by column chromatography (35:60:5 hexane/dichlo-
romethane/diethyl ether). Yield 13%; mp 123-125 °C dec
(hexane/ethyl acetate); [R]20D +383.2 (c 0.25, CHCl3). IR (KBr)
1771, 1594, 1492, 1086, 1059; 1H NMR δ 7.40 and 7.29 (AA′BB′
system, 4H), 5.51 (d, J ) 5.9, 1H), 5.07 (m, 1H), 3.83 (m, 1H),
3.59 (m, 1H), 3.31 (m, 2H), 2.41 (s, 3H), 1.16 (t, J ) 7.1, 3H);
13C NMR δ 165.5, 143.6, 133.3, 129.7, 125.4, 119.1, 101.6, 79.5,
66.5, 35.2, 21.5, 14.5. Anal. Calcd for C14H16N2O4S: C, 54.53;
H, 5.23; N, 9.08; S, 10.40. Found: C, 54.09; H, 5.09; N, 9.13;
S, 10.59.
4-Meth oxy-3-m eth yl-3,3a ,4,6a -tetr a h yd r o-6H-fu r o[3,4-
c]p yr a zol-6-on e (syn -8-exo). syn-8-exo was obtained by reac-
tion of 3 with diazoethane. It was isolated as an oil by column
chromatography (1:2 ethyl acetate/hexane). IR (film) 1781,
(3R,3aS,4S,6aR,SS)-4-Eth oxy-3,3a-dim eth yl-6a-[(4-m eth -
ylph en yl)su lfin yl]-3,3a,4,6a-tetr ah ydr o-6H-fu r o[3,4-c]pyr a-
zol-6-on e (syn -6a -exo). syn-6a -exo was obtained as a colorless
oil after 90 min of reaction time from 2a and diazoethane in
1
1555, 1180, 1144; H NMR δ 5.55 (dd, J ) 9.0 and 2.3, 1H),
quantitative yield (entry 1, Table 3); [R]20 +423.0 (c 1.6,
D
5.42 (d, J ) 6.4, 1H), 5.38 (m, 1H), 3.44 (s, 3H), 2.75 (ddd, J )
9.0, 6.4, and 2.5, 1H), 1.33 (d, J ) 7.3, 3H); 13C NMR δ 168.0,
103.6, 92.6, 86.7, 57.7, 43.2, 17.7. Anal. Calcd for C7H10N2O3:
C, 49.41; H, 5.92; N, 16.46. Found: C, 49.21; H, 6.26; N, 16.54.
6-Meth oxy-3-m eth yl-3,3a ,6,6a -tetr a h yd r o-4H-fu r o[3,4-
c]p yr a zol-4-on es (a n ti-9-exo a n d a n ti-9-en d o). anti-9-exo
and anti-9-endo were obtained by reaction of 3 with diazoeth-
ane and could not be fully purified by column chromatography
(1:2 ethyl acetate/hexane). Both isomers of 9 were character-
ized from their NMR data from a 32:68 mixture of anti-9-endo/
anti-9-exo. anti-9-exo: 1H NMR δ 5.76 (s, 1H), 5.26 (m, 2H),
3.61 (s, 3H), 2.78 (dd, J ) 7.4 and 1.8, 1H), 1.39 (d, J ) 7.2,
3H). 1H NMR (C6D6) δ 5.28 (s, 1H), 4.83 (m, 1H), 4.54 (dd, J )
7.5 and 2.3, 1H), 2.96 (s, 3H), 1.89 (dd, J ) 7.5 and 1.8, 1H),
0.61 (d, J ) 7.3, 3H). 13C NMR δ 175.3, 104.1, 92.9, 89.8, 57.2,
42.3, 18.3. anti-9-endo: 1H NMR δ 5.71 (s, 1H), 5.26 (m, 1H),
4.78 (m, 1H), 3.60 (s, 3H), 3.11 (dd, J ) 9.1 and 8.2, 1H), 1.69
1
acetone). IR (film) 1775, 1595, 1493, 1083, 1053; H NMR δ
7.75 and 7.42 (AA′BB′ system, 4H), 5.42 (s, 1H), 4.67 (q, J )
7.4, 1H), 3.90 (m, 1H), 3.69 (m, 1H), 2.46 (s, 3H), 1.70 (d, J )
7.4, 3H), 1.48 (s, 3H), 1.23 (t, J ) 7.1, 3H); 13C NMR δ 161.1,
143.2, 133.6, 129.4, 126.7, 111.9, 107.2, 89.0, 67.5, 54.5, 21.6,
14.7, 12.9, 12.0.
(3aR,4S,6aS,SS)-4-Eth oxy-3a-m eth yl-6a-[(4-m eth ylph en -
yl)su lfin yl]-3,3a ,4,6a -tetr a h yd r o-6H-fu r o[3,4-c]p yr a zol-6-
on e (a n ti-7a ). anti-7a was obtained as the minor adduct by
addition of diazomethane to 2a in acetonitrile. It decomposed
by column chromatography. The spectroscopic data for anti-
7a correspond to those of the crude reaction mixture (entry 5,
1
Table 3). H NMR δ 7.76 and 7.37 (AA′BB′ system, 4H), 4.82
(s, 1H), 4.78 and 4.58 (AB system, J ) 18.3, 2H), 3.91 (m, 1H),
3.65 (m, 1H), 2.45 (s, 3H), 1.64 (s, 3H), 1.27 (t, J ) 7.0, 3H);
13C NMR δ 162.2, 143.7, 133.6, 129.2, 128.2, 110.1, 106.7, 89.1,
67.4, 50.8, 21.6, 14.8, 11.4.
1
(d, J ) 7.5, 3H). H NMR (C6D6) δ 5.22 (d, J ) 0.6, 1H), 4.48
4-Eth oxy-3,3a -d im eth yl-6a -[(4-m eth ylp h en yl)su lfin yl]-
3,3a ,4,6a -tetr a h yd r o-6H-fu r o[3,4-c]p yr a zol-6-on e (a n ti-
6b-exo a n d a n ti-6b-en d o). Compounds anti-6-endo and anti-
6-exo were obtained by reaction of 2b with diazoethane (entries
1 and 2, Table 4) and could not be separated. A 43:57 mixture
of anti-6-endo/anti-6-exo was obtained by washing the crude
reaction with diethyl ether. White solid; IR (KBr) 1769, 1594,
1496, 1085, 1055. Anal. Calcd for C16H20N2O4S: C, 57.12; H,
5.99; N, 8.33; S, 9.53. Found: C, 57.10; H, 5.96; N, 8.39; S,
9.50. (3R,3a S,4R,6a R,SS)-a n ti-6b-exo (major isomer): 1H
NMR δ 7.72 and 7.37 (AA′BB′ system, 4H), 4.96 (s, 1H), 4.48
(q, J ) 7.5, 1H), 3.81 (m, 1H), 3.55 (m, 1H), 2.44 (s, 3H), 1.66
(d, J ) 7.5, 3H), 1.32 (s, 3H), 1.15 (t, J ) 7.1, 3H); 13C NMR δ
162.1, 142.8, 134.3, 129.3, 126.9, 112.1, 105.4, 94.6, 66.6 (or
66.3), 54.2, 21.5, 14.5, 12.9, 9.5. (3S,3a S,4R,6a R,SS)-syn -6b-
en d o (minor isomer): 1H NMR δ 7.71 and 7.37 (AA′BB′
system, 4H), 4.96 (s, 1H), 4.81 (q, J ) 7.6, 1H), 3.81 (m, 1H),
3.55 (m, 1H), 2.44 (s, 3H), 1.49 (d, J ) 7.6, 3H), 1.43 (s, 3H),
1.16 (t, J ) 7.1, 3H); 13C NMR δ 162.7, 142.8, 134.0, 129.3,
126.9, 112.4, 103.1, 96.7, 66.3 (or 66.6), 51.6, 21.5, 14.6, 13.5,
12.2.
(ddd, J ) 8.1, 2.0, and 0.6, 1H), 3.82 (m, 1H), 2.94 (s, 3H),
2.12 (dd, J ) 9.1 and 8.3, 1H), 1.44 (d, J ) 7.5, 3H). 13C NMR
δ 172.8, 103.5, 95.2, 87.7, 57.1, 39.1, 14.3.
Com p u ta tion a l Meth od s
The Gaussian 98 program28 was used to perform all of the
calculations. All the geometries of the reactants, transition
structures, and products were fully optimized at the well-
(28) Frisch, M. J .; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J . R.; Zakrzewski, V. G.; Montgomery, J . A.,
J r.; Stratmann, R. E.; Burant, J . C.; Dapprich, S.; Millam, J . M.;
Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J .;
Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo,
C.; Clifford, S.; Ochterski, J .; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J . B.; Cioslowski, J .; Ortiz, J . V.; Baboul, A. G.; Stefanov,
B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.;
Martin, R. L.; Fox, D. J .; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
J ohnson, B.; Chen, W.; Wong, M. W.; Andres, J . L.; Head-Gordon, M.;
Replogle, E. S.; Pople, J . A. Gaussian 98, Revision A.7; Gaussian,
Inc.: Pittsburgh, PA, 1998.
J . Org. Chem, Vol. 68, No. 17, 2003 6533

Garc´ıa Ruano et al.
established B3LYP/6-31G(d) level and characterized as minima
or first-order saddle points by frequency analysis. The reported
values of energy barriers in the gas phase include zero-point
energy corrections (ZPEC) scaled by a factor of 0.98.29
The same scale factor was used for the calculations at the
B3LYP/6-31++G(d,p), which include ZPEC at the same level,
and for the calculations with the B3LYP/6-31++G(d,p)//
B3LYP/6-31G(d) and B3LYP/cc-pVTZ//B3LYP/6-31G(d) model
chemistries, which include ZPEC at the same level as the
optimized geometries (i.e. B3LYP/6-31G(d)). Ab initio calcula-
tions consisted of the geometry optimization and frequency
analysis at the HF/6-31++G(d,p) level. Subsequent geometry
optimization was performed at the correlated second-order
MP2/6-31++G(d,p) level starting from the HF geometries. The
reported energy values at this level include ZPEC at the HF/
6-31++G(d,p) level scaled by a factor of 0.92.29 Solvation effects
were evaluated by means of single-point energy calculations
with the polarizable continuum model (PCM) of Tomasi et al.30
with dielectric constants of 4.20 (Et2O) and 35.94 (CH3CN).31
While these calculations are still approximate because they
do not include explicit solvent molecules, they do provide
reasonable estimates of the effect of solvent polarity on the
energy barriers. We expect a change in the geometries upon
solvation, but the size of the system, the number of different
stereochemical approaches to take into account, and the
general difficulties of convergence of the geometry optimiza-
tion involving the PCM method preclude a full optimization
in solution at the levels of calculations used in this paper.
Charges on individual atoms were computed by using the
natural population32 and the ChelpG33 schemes. Wiberg bond
indices and the atomic orbital overlap matrix (for the calcula-
tion of the Fukui functions) were obtained from a natural bond
orbital (NBO)32,34 analysis with the BNDIDX and SAO key-
words, respectively.
Ack n ow led gm en t. This work was supported by
research funds donated by the Direccio´n de Investiga-
cio´n Cient´ıfica y Te´cnica, CAICYT (Projects BQU2000-
0246, BQU2000-0248, and PB98-0102). R.G. thanks the
J unta de Extremadura-Fondo Social Europeo for a
predoctoral fellowship.
Su p p or tin g In for m a tion Ava ila ble: A brief explanation
of the DTF global and local properties used and Cartesian
coordinates of all of the structures with their computed total
energies. This material is available free of charge via the
Internet at http://pubs.acs.org.
J O0345603
(32) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. Rev. 1988, 88,
899-926.
(29) Scott, A. P.; Radom, L. J . Phys. Chem. 1996, 100, 16502-16513.
(30) Cossi, M.; Barone, V.; Cammi, R.; Tomasi, J . Chem. Phys. Lett.
1996, 255, 327-335.
(31) Reichardt, C. Solvents and Solvent Effects in Organic Chemistry;
VCH: New York, 1990; pp 408-410.
(33) (a) Chirlian, L. E.; Francl, M. M. J . Comput. Chem. 1987, 8,
894-905. (b) Breneman, C. M.; Wiberg, K. B. J . Comput. Chem. 1990,
11, 361-373.
(34) Reed, A. E.; Weinstock, R. B.; Weinhold, F. J . Chem. Phys. 1985,
83, 735-746.
6534 J . Org. Chem., Vol. 68, No. 17, 2003