The MARDi cascade: A Michael-initiated domino-multicomponent approach for the stereoselective synthesis of seven-membered rings

Article abstract of DOI:10.1002/chem.200701708

The MARDi cascade is a recently invented three-component Michael-initiated condensation involving 1,3-dicarbonyl derivatives. It allows regio- and stereocontrolled access to a variety of functionalised and substituted seven-membered rings. The substitutio

Full text of DOI:10.1002/chem.200701708

DOI: 10.1002/chem.200701708
The MARDi Cascade: A Michael-Initiated Domino-Multicomponent
Approach for the Stereoselective Synthesis of Seven-Membered Rings
Yoann Coquerel,* Marie-HØl›ne Filippini, David Bensa, and Jean Rodriguez*[a]
Abstract: The MARDi cascade is a re-
cently invented three-component Mi-
chael-initiated condensation involving
1,3-dicarbonyl derivatives. It allows
regio- and stereocontrolled access to a
variety of functionalised and substitut-
ed seven-membered rings. The substitu-
tion array can be diastereoselectively
modulated by appropriate choice of
the reaction partners, and the reaction
allows the control of up to five newly
created stereocentres and a complete
chiral induction in the case of an opti-
cally active ketone precursor. The high
level of diastereoselectivity observed
has been attributed to total thermody-
namic control of the reaction. The at-
tractiveness of the present domino
three-component approach to seven-
membered rings resides in the diversity
of carbo- and heterocyclic structures
that can be accessed with total regio-
control and high stereocontrol by start-
ing from simple substrates, under user
and environmentally friendly condi-
tions, as now required in modern or-
ganic chemistry.
Keywords: domino
reactions
·
Michael addition · multicomponent
reactions · seven-membered rings ·
stereoselectivity
Introduction
chael-aldol–retro-Dieckmann) cascade. The development of
new approaches to substituted stereodefined cycloheptanes
constitutes a worthwhile synthetic pursuit because of their
ubiquitousness in naturally occurring products and their use-
fulness as synthetic building blocks.[10] In this article, we
report the full detailed study of the MARDi cascade.
The rapid creation of molecular complexity in a regio- and
stereodefined manner from simple substrates, combining
economical aspects[1,2] with environmental ones,[3] constitutes
a great challenge in modern organic chemistry from both
academic and industrial points of view. In this field, multi-
component reactions[4] (MCRs) involving domino process-
es[5] have emerged as powerful tools in diversity-oriented
synthesis (DOS) and have found multiple applications in the
discovery of new bioactive small molecules.[6] In the field of
MCRs, isocyanide-based transformations and metal-cata-
lyzed processes largely lead the way.[4] In our laboratory,
there is an ongoing interest in domino and multicomponent
reactions that involve 1,3-dicarbonyls and their derivatives.[7]
As part of this program, we discovered a new anionic
domino three-component two-carbon ring expansion for the
stereoselective preparation of functionalised seven-mem-
bered rings[8,9] that we named MARDi (an acronym for Mi-
Results and Discussion
The MARDi cascade with b-ketoesters: The discovery of
the MARDi cascade came out from a previous study on the
synthesis and reactivity of 2-hydroxybicyclo[3.2.1]octan-8-
A
ones.[11] The reaction of the Dieckmann ester 1a with acrole-
in (2a) or various b-substituted a,b-unsaturated aldehydes
2b–f in the presence of base (1.5 equiv of K2CO3 or 1,8-
diazabicyclo
ACHTREUNG
ene gives the substituted 2-hydroxybicyclo
AHCTREUNG
ones 3 by a domino Michael addition–intramolecular aldol
cyclisation (Scheme 1). The bicyclic products 3 are generally
obtained in high yields and with moderate diastereoselectivi-
ties. A study on the reactivity of these 2-hydroxybicyclo-
[a] Dr. Y. Coquerel, Dr. M.-H. Filippini, Dr. D. Bensa, Dr. J. Rodriguez
UniversitØ Paul CØzanne (Aix-Marseille III)
UMR CNRS 6178, Centre universitaire de St JØrôme
service 531, 13397 Marseille Cedex 20 (France)
Fax : (+33)491-28-8841
[3.2.1]octan-8-ones, pioneered by Stork–Landesman[12] and
Buchanan,[13] revealed that substituted cycloheptanols could
be obtained stereoselectively by a retro-Dieckmann frag-
mentation (i.e. 3b!4) promoted by K2CO3 or DBU
(1.0 equiv) in methanol (Scheme 1).[14] This two-step synthe-
sis of substituted cycloheptanols prompted us to surmise
E-mail: yoann.coquerel@univ-cezanne.fr
jean.rodriguez@univ-cezanne.fr
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
3078
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3078 – 3092

FULL PAPER
the same efficiency to provide
the di(methylcarboxylate)cyclo-
heptanols 4a and 5a (dr=
1.5:1), following a total transes-
terification, if the reaction is
conducted
in
methanol
(entry 3). As expected, the re-
action in ethanol provides the
di(ethylcarboxylate)cyclohepta-
nols 4b and 5b (dr=1.5:1) in
slightly lower yield, probably as
a result of the higher steric hin-
drance of the alkoxide nucleo-
phile in the final retro-Die-
ckmann step (entry 4).
Scheme 1. Synthesis and reactivity of 8-oxo-bicyclo[3.2.1]octanols.
A
that the overall cascade Michael addition–regioselective
aldol cyclisation–retro-Dieckmann fragmentation would be
possible in a one-pot process by starting from the Die-
ckmann ester 1a and a,b-unsaturated aldehydes 2 in MeOH
as the third component. Indeed, we were pleased to observe
that the cycloheptanol 4a was obtained stereoselectively in
36% yield (together with very polar material) from the Die-
ckmann ester 1a and acrolein (2a, 1.5 equiv) in the presence
of two equivalents of K2CO3 in methanol (0.4m) by means
Table 1. The MARDi cascade with the Dieckmann esters 1a,b and acro-
lein (2a).
Entry Substrate Conditions[a]
Product
Yield
[%][b]
1
2
3
4
1a
1a
1b
1b
K2CO3 (2.0 equiv), MeOH
(0.4m)
K2CO3 (0.5 equiv), MeOH
(0.4m)
K2CO3 (0.5 equiv), MeOH
(0.4m)
K2CO3 (0.5 equiv), EtOH
(0.4m)
4a
36
94
95
84
4a/5a
1.5:1
4a/5a
1.5:1
4b/5b
1.5:1
of the MARDi cascade (Scheme 2, Table 1, entry 1).[15]
A
very small amount of the unstable cycloheptene 6 was also
detected but could not be isolated. The structure of cyclo-
heptanol 4a has been confirmed by perfect matching of its
NMR spectroscopic data with those obtained previously in
the two-step synthesis of the same product.[11] When the
same reaction was performed with a substoichiometric
amount of K2CO3 (0.5 equiv), the diastereomeric cyclohep-
tanols 4a and 5a (dr=1.5:1) were obtained in 94% yield
(entry 2). These results suggest that in the presence of an
excess of base, the all-cis cycloheptanol 5a undergoes dehy-
dration to the cycloheptene 6, which decomposes rapidly in
the reaction mixture (Scheme 2). The less expensive Die-
ckmann ethyl ester 1b can also be used in this reaction with
[a] All reactions were conducted at room temperature with 1.5 equiv of
acrolein (2a). [b] Yield of isolated product.
We next turned our attention to the synthesis of more-
substituted cycloheptanols by using substituted acroleines
and substituted 2-oxo-cyclopentanecarboxylate methyl
esters (Scheme 3). The condensation between the Die-
ckmann ester 1a and crotonaldehyde (2b) was chosen for
the optimisation study (Table 2). Under the conditions pre-
viously developed for the condensation of acrolein
(0.5 equiv of K2CO3), the ex-
pected substituted cyclohepta-
nol 4c was obtained diastereo-
selectively (dr>25:1) in 54%
yield (entry 1). Increasing the
quantity of base and the use of
a
chelating additive had no
effect on both the yield and dia-
stereoselectivity (entries 2
and 3). Magnesium carbonate
left the starting material un-
changed (entry 4), while cesium
carbonate allowed the forma-
tion of 4c in modest yield
(entry 5), and the use of a che-
lating additive resulted in the
decomposition of the products,
even at low temperature (en-
tries 6 and 7). Alternatively,
Scheme 2. The MARDi cascade with the Dieckmann esters 1a,b and acrolein (2a).
Chem. Eur. J. 2008, 14, 3078 – 3092
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3079

Y. Coquerel, J. Rodriguez et al.
The structure of 4c has been
deduced from the strong NOEs
observed in NOEDIFF experi-
ments between the protons a to
the methyl, the hydroxyl (1,3-
cis for the methyl and the hy-
droxyl groups) and the two car-
boxyl groups (1,4-cis for the
Scheme 3. The MARDi cascade: synthesis of cycloheptanols.
two carboxylates). The 1,2-trans
relationship between the hy-
Table 2. Optimization study for the MARDi cascade with the Die-
ckmann ester 1a and crotonaldehyde (2b).
droxyl and the methylcarboxyl groups was deduced from
the large 3J
(H,H) coupling constant (10 Hz) observed be-
A
Entry
Conditions[a]
Yield of 4c [%][b]
tween the protons a to these substituents. The structure of
4c was secured later by X-ray diffraction analysis
(Scheme 4).[16]
1
K2CO3 (0.5 equiv), 20 h
54
2
K2CO3 (1 equiv), 21 h
54
3
K2CO3 (1 equiv), [18]crown-6, 23 h
54
With an efficient protocol for the stereoselective synthesis
of the substituted cycloheptanol 4c in hand, we explored the
scope of the MARDi cascade, first, with a variety of b-sub-
stituted acroleines. The results are summarized in Table 3. It
appears that the MARDi cascade is a general reaction that
can be performed in good to high yield with alkyl (en-
tries 1,2) and aryl (entries 3–5) b-substituted aldehydes with
excellent diastereoselectivity. By starting from the optically
active 5-methyl-2-oxo-cyclopentanecarboxylate methyl ester
(+)-1c derived from (+)-pulegone,[17] the MARDi cascade
with crotonaldehyde (2b) provided the optically active cy-
cloheptanol (+)-4h, which contains five stereogenic centres,
with total chiral induction (entry 6).
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
C
no reaction
37
decomposition
decomposition
63
74
no reaction
0[c]
0[c]
0[c]
49
48
52
50
94
52
78
32
The excellent diastereoselectivity observed for the synthe-
sis of cycloheptanols 4c–h by the MARDi cascade is in
sharp contrast with the low selectivities observed for the
[a] Unless otherwise mentioned, all reactions were conducted in dry
MeOH (0.4m) at room temperature with 1.5 equiv of crotonaldehyde
(2b). [b] Except for entries 1 and 18 in which 4c was isolated by flash
chromatography, yields were determined from the 1H and/or 13C NMR
spectra of the crude mixtures (dr>25:1). [c] The bicyclo[3.2.1]octanols
3a and 3b (R=Me) were isolated (dr=1:1.5).
preparation of the intermediate bicyclo[3.2.1]octanones 3
N
from the same substrates. The interrupted reaction of the
Dieckmann ester 1a with crotonaldehyde (2b) in the pres-
ence of 1.5 equivalents of DBU in methanol at 08C for 4 h
G
gave
the
methyl
2-hydroxy-4-methyl-8-oxo-bicyclo-
A
40:60 mixture of epimers at C4 (Scheme 4).[11] The high dia-
stereoselectivity of the MARDi cascade can thus only be ex-
plained by the selective retro-Dieckmann fragmentation of
sodium methoxide and potassium hydroxide proved superior
to potassium carbonate (entries 8 and 9). Non-ionic bases
such as pyridine left the starting material unchanged even
the equatorial 4-methyl bicyclo
the expected cycloheptanol 4c, while the axial 4-methyl
bicyclo[3.2.1]octanone 3c (kinetic product, see the Support-
ing Information) suffered a tandem retro-aldol–retro-Mi-
chael ring opening to give back 1a and 2b, followed by a
thermodynamically controlled ring reconstitution that led to
A
after
diazabicyclo
methylaminopyridine (DMAP) resulted in the formation of
the corresponding bicyclo[3.2.1]octanols 3a and 3b (R=Me,
entries 11–13, respectively). On the other hand, tetramethyl-
guanidine (TMG), 1,5-diazabicyclo[4.3.0]non-5-ene (DBN,
a
prolonged reaction time (entry 10), and 1,4-
ACHTREUNG
A
AHCTREUNG
equatorial 4-methylbicyclo[3.2.1]octanone 3d, precursor of
U
N
the cycloheptanol 4c. Evidence for total thermodynamic
control of the MARDi cascade came from a cross experi-
ment in the presence of cinnamaldehyde (2d). Indeed, when
after a prolonged reaction time) and DBU gave the desired
cycloheptanol 4c with a comparable efficiency to K2CO3
(entries 14–17). Finally, 1.0 equiv of DBU was found to be
efficient to give 4c in high yield and with very good diaste-
reoselectivity (dr>25:1, entry 18), while an excess of DBU
or additives had a deleterious impact on the yield (en-
tries 19–21), probably favouring the dehydration of the cy-
cloheptanol to the corresponding unstable cycloheptene as
in the case of acrolein.
the axial 4-methylbicyclo[3.2.1]octanone 3c is allowed to
G
react with cinnamaldehyde (2d) under the optimised
MARDi cascade conditions for b-substituted aldehydes, a
mixture of cycloheptanols 4c (40%) and 4e (10%) is ob-
tained, accompanied by unreacted 3c (Scheme 4). This spe-
cific behaviour can be rationalised by invoking the steric
hindrance of the axial methyl substituent on the b-face
3080
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3078 – 3092

The MARDi Cascade
FULL PAPER
Scheme 4. The diastereoselectivity of the MARDi cascade with aldehydes 2b–f.
Table 3. The MARDi cascade with b-substituted acroleines.[a]
layer. After acidification with
HCl and extraction with diethyl
ether or ethyl acetate, we were
pleased to isolate the clean, dia-
stereomerically pure, cyclohep-
tenic acids 8 in good to high
yield.[18] The acid 8a was select-
ed for the optimisation study
(Table 4) and one equivalent of
Entry
Substrate
Aldehyde
Product
Yield [%][b]
1
2
3
4
5
6
1a
1a
1a
1a
1a
(+)-1c
2b: R2 =Me
2c: R2 =nPr
2d: R2 =Ph
4c: R1 =H, R2 =Me
4d: R1 =H, R2 =nPr
4e: R1 =H, R2 =Ph
94
60
96
64
83
43
2e: R2 =2-furyl
2 f: R2 =o-anisyl
2b: R2 =Me
4 f: R1 =H, R2 =2-furyl
4g: R1 =H, R2 =o-anisyl
(+)-4h: R1 =R2 =Me
[a] All reactions were conducted at room temperature in dry MeOH (0.3–0.4m) by using 1 equiv of DBU and
1.5 equiv of aldehyde. [b] Yields are given for the isolated homogeneous products obtained after silica-gel
flash chromatography.
which prevents the retro-Dieckmann fragmentation. Finally,
Table 4. Optimization study for the MARDi cascade between the Die-
for the optically active pentasubstituted cycloheptanol 4h
obtained from crotonaldehyde (2b), the initial Michael addi-
tion step occurs with high diastereofacial control as a result
of the presence of the 3-methyl substituent in the b-ketoest-
er 1c. This results in the exclusive 1,3-trans relationship be-
tween the two methyl substituents in the cycloheptanol (+)-
4h.
With a-substituted acroleines (no b-substituent), the issue
of the MARDi cascade is somewhat different. In early ex-
periments, the cycloheptanols 7 (R3 =Me) could be obtained
from the Dieckmann ester 1a and methacrolein (2g), but in
moderate yield (36% with 0.5 equiv of K2CO3) and with
poor diastereoselectivity as a result of the lack of control
over the newly created stereocentre which is formed during
the Michael addition (Scheme 3). A brief study aiming at
optimising the nature and amount of base established that
cycloheptanols 7 could not be obtained in a yield higher
than 40% and, more importantly, that increasing the
amount of base resulted in dramatic drop of the yield (e.g.
<5% with 1.0 equiv DBU). However, the crude mixtures
were clean, containing essentially the cycloheptanols 7,
which prompted us to examine the content of the aqueous
ckmann ester 1a and methacrolein (2g).
Entry
Conditions[a]
Yield of 8a [%][b]
1
2
3
4
K2CO3 (0.5 equiv), 48 h
K2CO3 (1.5 equiv), 22 h
KOH (0.5 equiv), 6 h
KOH (1 equiv), 17 h
46
82
39
65
73
32
60
44
90
76
62[c]
5
6
7
8
9
10
11
KOH (2.7 equiv), 7 h
DBU (0.25 equiv), 51 h
DBU (0.5 equiv), 9 h, reflux
DBU (0.5 equiv), 47 h
DBU (1.0 equiv), 18 h
DBU (1.0 equiv), 72 h
DBU (2.0 equiv), 24 h
[a] Unless otherwise mentioned, all reactions were conducted in MeOH
(0.3–0.4m) at room temperature with 1.5 equiv of aldehyde. [b] Yields are
given for the clean homogeneous, diastereomerically pure, crude prod-
ucts obtained after acidic workup. [c] 18% of the corresponding diacid
was also present in the crude mixture.
DBU in 0.4m methanol was found to give the best result
(entry 9, Scheme 5). While the overall transformation is not
much affected by the nature of the base (compare entries 2,
5 and 9), prolonged reaction times result in lower yields,
Chem. Eur. J. 2008, 14, 3078 – 3092
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3081

Y. Coquerel, J. Rodriguez et al.
Scheme 5. The MARDi cascade with aldehydes 2g–m.
owing to partial decomposition (entry 10), and an excess of
DBU results in partial double saponification to provide a
substantial amount of the corresponding diacid (entry 11). A
variety of a-substituted acroleines were tried, and the results
are summarised in Table 5. Good to excellent yields were
COOH (d=2.84 ppm), indicating that they are on the same
face of the seven-membered ring (1,3-trans relationship be-
tween Me and COOH). The HMBC spectrum revealed a
3
crosspeak as a result of a J
A
the vinyl proton (d=7.10 ppm) and the carbon atom of the
conjugated ester (d=
168.6 ppm), indicating that che-
moselective saponification of
the non-conjugated ester had
occured. The structure of 8a
has recently been confirmed by
X-ray diffraction analysis of the
dihydroxylated derivative 9 ob-
tained in 83% yield from 8a
(Scheme 5).[16]
Table 5. The MARDi cascade with a-substituted acroleines.[a]
Entry
Substrate
Aldehyde
t [h]
Product
Yield [%][b]
1
2
3
4
5
6
7
8
9
1a
1a
1a
1a
1a
1a
1a
1d
1d
1a
2g: R=Me
2h: R=Et
2i: R=nBu
2j: R=Ph
18
50
22
19
50
48[c]
48
9
8a: R=Me
8b: R=Et
8c: R=nBu
8d: R=Ph
90
96
75
91
71
87
84
98
62
80
ꢁ
ꢁ
2k: R=C CSiMe3
8e: R=C CH
2l: R=(CH2)2OBn
2m: R=(CH2)2CO2Me
2g: R=Me
8 f: R=(CH2)2OBn
8g: R=(CH2)2CO2Me
8h: R=Me
8i: R=(CH2)2OBn
8j
Our previous work on the
2l: R=(CH2)2OBn
2g: R=Me
24
10
96[d]
synthesis
[3.2.1]octanes
of
A
bicyclo-
(Scheme 6)
A
[a] Unless stated otherwise, all reactions were conducted at room temperature in dry MeOH (0.4m) by using
1.0 equiv of DBU and 1.5 equiv of aldehyde. [b] Yields are given for the clean homogeneous, diastereomerical-
ly pure, crude products obtained after acidic workup. [c] Reflux. [d] Reaction performed in dry EtOH.
from the Dieckmann ester 1a
and methacrolein (2g) showed
that this carbocyclisation was
poorly diastereoselective,[11] and
obtained regardless of the nature of the substituent, and
functionalised alkyl chains can be introduced (entries 6, 7,
and 9). As expected,[19] in the case of aldehyde 2k, the reac-
no steric argument can be used here for the chemoselective
retro-Dieckmann fragmentation of one isomer of A. Thus,
during the reaction, the cycloheptanols B are formed as a
mixture of isomers, as proven by our initial attempts on this
reaction (see above). One can then postulate that a stereo-
selective proton capture of the putative last intermediate C
of the cascade occurs, followed by chemoselective saponifi-
cation of the non-conjugated ester with the water produced
on dehydration or upon hydrolysis of the reaction mixture
(Scheme 6a). However, thermodynamic considerations
cannot account for the 1,3-trans diastereoselectivity ob-
served for the MARDi cascade with a-substituted alde-
hydes, the 1,3-trans cycloheptenic acids 8 being the less-
stable isomers, as corroborated by simple theoretical calcu-
lations.[20] Furthermore, when the reaction of entry 1
(Table 5) is run in the presence of 3 Š molecular sieves, the
ꢀ
tion proceeds with concomitant C SiMe3 bond cleavage
leading to 8e (entry 5). The bicyclic b-ketoester 1d gave
similar results (entries 8 and 9), allowing the facile construc-
tion of the corresponding functionalised bicyclo-
[5.4.0]undecene ring system found in many natural products.
When performing the reaction in ethanol and starting from
1a and methacrolein (2g), an additional trans-esterification
occurred to give the corresponding ethyl ester 8j (entry 10).
The cycloheptenes 8a–j are obtained with excellent dia-
stereoselectivity, and their structure has been determined by
extensive NMR spectroscopic studies. The NOESY spec-
trum of 8a clearly shows a correlation spot between the Me
group (d, d=1.20 ppm) and the hydrogen atom a to the
3082
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3078 – 3092

The MARDi Cascade
FULL PAPER
study, we found that 1) DBU is
not satisfactory in the case of b-
ketosulfones, leaving substantial
amounts of the intermediate
bicyclo
A
unfrag-
mented, 2) a stoichiometric
amount of K2CO3 gives the best
result in most cases and 3) dilu-
tion of the reaction mixture and
addition of a co-solvent (THF
or toluene), which prevents the
direct retro-Dieckmann cleav-
age of the starting material, has
a pronounced beneficial impact
on both the yield and the selec-
Scheme 6. The diastereoselectivity of the MARDi cascade with aldehydes 2g–m.
cycloheptenic acid 8a is isolated in 80% yield, indicating
that the presence of water is not necessary for the saponifi-
cation step. Therefore, the high diastereoselectivity and the
regioselective saponification might be better explained by
invoking an intramolecular lactonization–elimination se-
quence of the isomer D which drives the equilibrium, result-
ing only in the formation of the 1,3-trans cycloheptenic acids
8 (Scheme 6b).
The MARDi cascade with b-ketosulfones: To further
expand the scope of the MARDi cascade, we studied its fea-
sibility with easily accessible thiofunctionalised precursors.[21]
The sulfides 10a–c[22] were prepared by using modified Trost
conditions[23] (1,3-dimethyl-3,4,5,6-tetrahydro-2(1H)-pyrimi-
dinone (DMPU) advantageously replaces hexamethylphos-
phoramide (HMPA)), the optically active sulfoxide 11a
(mixture of two diastereomers) was prepared from the cy-
clopentanone N-phenylimine and (ꢀ)-menthyl p-toluenesul-
finate[24] and the sulfones 12a–c were obtained by oxidation
of the corresponding sulfides with meta-chloroperbenzoic
acid (mCPBA, Scheme 7).[25]
Scheme 7. Thiofunctionalised precursors tested in the MARDi cascade.
tivity. Indeed, under the optimised conditions (1.0 equiv
K2CO3, 1:1 THF/methanol, 0.06m, RT, 20 h), the reaction of
b-ketosulfone 12a with acrolein (2a) yields 93% of the sul-
fonyl cycloheptanol 15a as a 5:1 inseparable diastereomeric
mixture (entry 1). The same reaction conducted with croto-
naldehyde (2b) provides, stereoselectively, the correspond-
ing substituted cycloheptanol 15b in high yield (entry 2). In
the case of 2-substituted aldehydes, such as methacrolein
(2g) and 2-butyl acrolein (2i), the corresponding cyclohepta-
nols 15c and 15d are also obtained stereoselectively (en-
tries 3 and 4, respectively), but together with various
amounts of the corresponding carboxylic acids. The saponifi-
Preliminary experiments showed that in the presence of
base (K2CO3 or DBU) and methanol, b-ketosulfide 10a
reacts with a,b-unsaturated aldehydes 2a,b,g to give the 8-
oxo-bicyclo[3.2.1]octanols 13 in good yields and with low
A
stereoselectivities (Scheme 8). The final retro-Dieckmann
step of the MARDi cascade does not occur, probably due to
a lack of stabilization of the negative charge on the carbon
atom that bears the sulfide group. The same reactivity was
observed with the optically active b-ketosulfoxide 11a with
slightly better stereoselectivity.
We next examined the reac-
tion of the b-ketosulfone 12a
with representative a,b-unsatu-
rated aldehydes and were
pleased to find that sulfonyl-
substituted cycloheptanols could
be obtained stereoselectively by
means of the MARDi cascade
(Table 6). After an optimisation Scheme 8. Reactivity of b-keto-sulfides and -sulfoxides.
Chem. Eur. J. 2008, 14, 3078 – 3092
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3083

Y. Coquerel, J. Rodriguez et al.
Table 6. The MARDi cascade with b-ketosulfones.[a]
Entry
Substrate
Aldehyde
Product[b]
Yield [%][c] (major)
dr[d]
5:1
1
12a
2a
93
2
12a
12a
12a
2b
2g
2i
90 (51)
95 (62)
99 (79)
4:1:1:1
8:2:1:1
8:1:1
3[e]
4
5
12b
2g
73 (40)
2:1
6
12c
2a,b
nd[f]
[a] Unless otherwise stated, reactions were performed with 1.5 equiv of aldehyde and 1 equiv of K2CO3 in 1:1 MeOH/THF (0.06m) at room temperature
for 20 h. [b] The represented diastereomer is the major product. [c] Yields are given for the clean crude mixture of diastereomers. Yields in parenthesis
are given for the isolated major diastereomers obtained by single recrystallisation for 15b, 15c and 15e, or silica gel flash chromatography for 15d.
[d] Diastereomeric ratios were determined from the crude 1H and 13C NMR spectra. [e] Reaction performed with 2 equiv of K2CO3. [f] nd=not deter-
mined. The reaction was monitored by mass spectroscopy (ESI+).
cation probably occurred upon hydrolysis of the reaction,
and the acids were converted back to the methyl esters by
treatment with (trimethylsilyl)diazomethane prior to isola-
tion.[26] The MARDi cascade can also be performed with the
bicyclic b-ketosulfone 12b, which reacts with methacrolein
zylic position of the hydroxyl group, leading to a thermody-
namically stable conjugated system. However, by starting
from the bicyclic b-ketosulfone 12c, the retro-Dieckmann
fragmentation of the starting material was the only reaction
observed to give 16, probably due to the favourable steric
strain release compared to the highly demanding formation
of the corresponding bridged intermediate (entry 6).
(2g) to provide, stereoselectively, the bicyclo[5.4.0]undecane
A
15e (entry 5). In this case, the dehydration of the intermedi-
ate cycloheptanol seems to be facilitated by the homoben-
3084
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3078 – 3092

The MARDi Cascade
FULL PAPER
Although the MARDi cascade is less diastereoselective in
the sulfonyl series, the diastereomerically pure major iso-
mers could be isolated (except for entry 1) by a very effi-
cient recrystallisation in most cases. The relative configura-
tions of the sulfonyl cycloheptanols 15a–d and cycloheptene
15e have been determined by standard high-field 2D NMR
spectroscopic techniques and
The hetero-MARDi cascade was first tested (see struc-
tures depicted below) in the oxygen series with furanone
17a, obtained from methyl glycolate and methyl acrylate,[31]
and the isomeric furanones 17b–g, formed by rhodium-car-
benoid insertion into the OH bond of the corresponding d-
hydroxy-b-ketoesters[32] or d-hydroxy-b-ketosulfone[33]. The
the structures of 15b and 15c
have been secured by X-ray dif-
fraction analysis.[16] The com-
parison of the products of the
MARDi cascade in the sulfonyl
and the corresponding carboxyl
series (vide supra) highlights
that stereocontrol and chemical
differentiation of the final cy-
cloheptane might be achieved
by appropriate choice of the
stabilising group on the starting
cyclopentanone (compare, for
example, compounds 4c with
15b, and 8a with 15c).
The retro-Dieckmann frag-
mentation is easier for a b-keto-
phenylsulfone than for a b-ketomethylester (the proton a to
a phenylsulfonyl group is about 1.5 pKa units lower than the
same proton a to a methyl ester group).[27] This can argue
for an easier fragmentation of the intermediate sulfonyl-sub-
study was initiated with furanone 17a and the representative
aldehydes 2a,b and 2g, which were allowed to react under
the previously optimised conditions. The desired oxepane
could not be obtained, and the open-chain products of type
20, resulting from a Michael addition–retro-Dieckmann
fragmentation sequence, were the only products isolated
(Table 7, entry 1). In this case, the intramolecular aldolisa-
tion does not occur due to the presence of the oxygen atom
a to the aldolisation site destabilizing the enolate, thus set-
ting the stage for the fragmentation.[34] Further experimental
evidence for this explanation came from our trials to synthe-
stituted bicyclo[3.2.1]octane of the MARDi cascade. Thus,
A
the diastereomeric ratio of the alkyl and hydroxyl substitu-
ents in cycloheptanol 15b approximately reflects the diaste-
reomeric ratio of the corresponding intermediate bicyclo-
[3.2.1]octane (axial methyl at C4 major), the relative config-
uration of the sulfonyl-substituted carbon atom probably re-
sults from a final thermodynamically controlled protonation.
In the case of 2-substituted aldehydes, the above mentioned
sise the corresponding bicyclo[3.2.1]octanones (K2CO3 or
A
intramolecular
lactonization–elimination
mechanism
DBU in acetone or THF) which resulted only in the forma-
tion of the Michael adduct (no ring-closing aldolisation)
even after a prolonged reaction time. We next examined the
reactivity of the isomeric furanone 17b in the MARDi cas-
cade, and were pleased to find that the expected oxepanes
could be obtained stereoselectively. The reactions were con-
ducted at various concentrations and a substoichiometric
amount of K2CO3, again, proved to be the most efficient
base for the transformation, giving the best yields in a rela-
tively diluted medium. Actually, these cycloheptanols are
relatively unstable in the reaction mixture, particularly in
concentrated basic media. For example, at a concentration
higher than 0.2m, the furanone 17b reacted with acrolein
(2a) to give the corresponding hydroxyoxepanes 21a in very
low yield as a 1.5:1 mixture of epimers (entry 2). Lowering
the concentration to 0.1 and 0.04m (entries 3 and 4) had no
effect on diastereoselectivity, but allowed the formation of
21a in 24 and 46% yields, respectively. Unfortunately, this
compound suffered degradation upon silica-gel chromatog-
raphy and only a small quantity of 21a could be isolated.
Thus, the crude product 21a was silylated prior to purifica-
(Scheme 6b), which provides evidence for the complete dia-
stereoselectivity observed in the b-ketoester series, cannot
be invoked here and thus the seven-membered rings 15c–e
are also obtained as mixtures of isomers.
The MARDi cascade in the heterocyclic series: Heterocyclic
seven-membered rings constitute the core or a key fragment
of a number of bioactive compounds, many of which are iso-
lated from natural sources.[28] The known biological proper-
ties of these compounds and the huge potential in drug dis-
covery of this nuclei renders desirable the development of
simple and general methodologies for their regio- and ste-
reoselective synthesis. Although the preparation of thie-
panes is less documented, the number of methodologies
made available for the preparation of azepanes and oxe-
panes has steadily increased in the past decades.[29] From
our previous results, we surmised that the MARDi cascade
could allow a quick access to a variety of heterocyclic seven-
membered rings.[30]
Chem. Eur. J. 2008, 14, 3078 – 3092
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3085

Y. Coquerel, J. Rodriguez et al.
Table 7. The MARDi cascade for the preparation of oxepanes.[a]
Entry
Substrate
Aldehyde
Product[b]
Yield [%][c]
1[d]
17a
2a,b,g
nd[n]
2[e]
17b
2a
14
3[g]
4[h]
17b
17b
2a
2a
21a
21a
24
46
5[h]
6[i]
7[i]
8[l]
17b
17b
17b
17b
2a
2b
2n
2g
42
28
21
42
9[d]
17c–f
2a,b,g
nd[n]
[a] Unless stated otherwise, all reactions were performed in dry MeOH (0.04m) with 1.5 equiv of aldehyde at room temperature for 20 h. [b] For 21a–d,
the major isomer is the thermodynamically favoured isomer (a-OH). The relative configurations of compounds 21b–e have been established by
2D NMR spectroscopic techniques. [c] Isolated, except for entries 2–4 for which the yields were estimated from the crude 1H and 13C NMR spectra.
[d] K2CO3 or DBU (0.5 or 1 equiv). [e] K2CO3 (0.25m, 0.5 equiv). [f] dr=1.5:1. [g] K2CO3 (0.1m, 0.5 equiv). [h] K2CO3 (0.5 equiv). [i] DBU (1 equiv).
[j] dr=6:1. [k] dr=4:1. [l] DBU (0.5 equiv). [m] drꢂ10:1. [n] nd=not determined. The reaction was monitored by mass spectroscopy (ESI+).
3086
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3078 – 3092

The MARDi Cascade
FULL PAPER
tion, allowing the isolation of the corresponding silyl ethers
21b in 42% yield (dr=1.5:1) from 17b (entry 5). This en-
couraging result prompted us to study the evolution and the
diastereoselectivity of the reaction with substituted alde-
hydes. In this case, DBU gave the best results under the op-
timised concentration conditions. The reaction of furanone
17b with crotonaldehyde (2b) gave the expected substituted
hydroxyoxepanes 21c (dr=6:1, Table 7, entry 6). Also, the
reaction of 17b with cyclopentene carboxaldehyde (2n)
gave the corresponding fused bicyclic compounds 21d (dr=
4:1, entry 7) stereoselectively with an all-trans relationship
of five contiguous stereogenic carbon atoms. By analogy
with the carbocyclic version of the MARDi cascade, the re-
action of 17b with methacrolein (2g) provided the 1,3-trans
carboxylic acid 21e in 42% yield with high diastereoselec-
tivity (drꢂ10:1, entry 8). In this case, the diastereoselectiv-
ity is not as high as in the parent carbocyclic reaction. This
result indicates that at least a fraction of the all-cis diaste-
reomer of the intermediate dimethyl ester oxepanol also un-
dergoes the intramolecular lactonization–dehydration se-
quence (see Scheme 6b) to give a very minor amount of the
diastereomeric 1,3-cis oxepine carboxylic acid, which could
not be separated from the major 1,3-trans isomer. The rest
of the material is essentially composed of degradation prod-
ucts, but trace amounts of the unfragmented bicyclo-
creasing the quantity of K2CO3 to 1 and 1.5 equivalents fav-
oured the formation of the dehydrated product 25b (R=
Me) without significant alteration of the global yield (en-
tries 3 and 4). However, in the latter case, the crude product
was very clean and contained only the dehydrated product
25b (R=Me), so the aqueous layer was acidified and ex-
tracted again to provide 51% of the acid 25b (R=H). The
structure of 25b (R=H) was confirmed by conversion to its
methylester by treatment with trimethylsilyldiazomethane.[26]
Not surprisingly, the reaction of pyrrolidone 18b with cyclo-
pentene carboxaldehyde (2n) provided the hydroxyazepane
25c and the azepine 25d (dr=1.6:1) with a yield and ratio
hydroxyazepane/azepine similar to those obtained in the re-
action with acrolein (compare entries 3 and 5). The bicyclo-
A
substituents on the seven-membered ring as proven by X-
ray diffraction analysis.[16] As expected, the reaction of 18b
with methacrolein (2g) provided the carboxylic acid 25e,
albeit in very low yield (entry 6), the major product being
the corresponding Michael–retro-Dieckmann aldehyde. Fi-
nally, in the sulfur series from the commercially available
thiafuranone 19, the MARDi cascade was also successful
under the standard conditions. Thiepines 26a–c were ob-
tained with aldehydes 2a,b,n, respectively, following dehy-
dration of the intermediate hydroxythiepane (entries 7–9),
and the carboxylic acids 26d and 26 f could be isolated with
aldehydes 2g and 2i, respectively (entries 10 and 11). How-
ever, for the reaction with methacrolein (entry 10), 17% of
the corresponding dimethylester 26e was also isolated as a
57:43 inseparable mixture of isomers (1,3-cis major). This
gives a further argument for the proposed tandem intramo-
lecular lactonization–elimination mechanism in the case of
2-substituted acroleins, leading to the dehydrated seven-
membered ring carboxylic acids (Scheme 6b), as the water
produced on dehydration of the dimethylester hydroxythie-
pane does not lead to the saponification of the non-conju-
gated ester group in the 1,3-trans dimethyl ester thiepine. In
this case, the dehydration of the transient hydroxy thiepane,
resulting from a poorly diastereoselective MARDi cascade,
to give 26e is in competition with the intramolecular lacto-
nization–elimination mechanism proposed for the formation
of 26d.
A
methyl ester oxepine could also be detected by mass spec-
troscopy. Our attempts to synthesize highly substituted oxe-
panes from the substituted furanones 17c–f only resulted in
the formation of the Michael–retro-Dieckmann products 22
and/or the bicyclo[3.2.1]octanols 23, together with considera-
R
ble amounts of degradation products (entry 9). To explain
this behaviour, one can argue that the nucleophilic attack of
the methoxide anion on both faces of the ketone in the in-
termediate bicyclo[3.2.1]octanone is sterically disfavoured.
T
Thus the retro-Dieckmann fragmentation cannot occur, and
the material partially decomposes under the reaction condi-
tions. Finally, by starting from the furanic b-ketosulfone 17g
and representative aldehydes 2a and 2g, the Michael adduct
was the major product and no sulfonyl-substituted oxepanol
could be detected.
With the feasibility of the hetero-MARDi cascade having
been established in the case of oxepanes and oxepines, we
logically tried to extend the method to the synthesis of aze-
panes (or azepines) and thiepanes (or thiepines). Under the
now well established optimised conditions of the cascade,
the pyrrolidone 18a[35] resulted almost exclusively in the for-
mation of the diester 24 by retro-Dieckmann ring opening
of the starting material (Table 8, entry 1). Actually, this
result is not surprising considering that the MARDi cascade
is performed under thermodynamic conditions, and the pyr-
rolidone 18a is the kinetic product of the Dieckmann cycli-
sation of the diester 24 (the thermodynamic product is the
isomeric pyrrolidone 18b).[35] Indeed, the reaction between
the pyrrolidone 18b and acrolein provided a 1:1 mixture of
the expected hydroxyazepanes 25a (dr=3.5:1) and the cor-
responding azepine 25b (R=Me) in 36% yield (entry 2). In-
When the MARDi cascade is performed with pyrrolidone
18b or thiofuranone 19, the presence of the heteroatom b to
the hydroxyl group in the cycloheptanol clearly favours the
dehydration. In almost every case, the best yield of hetero-
MARDi product was obtained with thiofuranone 19. This
can be rationalised by the relative higher acidity of the
proton a to the sulfur atom compared to nitrogen, which
might accelerate the dehydration process, thus avoiding un-
wanted over reaction of the cycloheptanol.
Conclusion
We have demonstrated that the MARDi cascade (up to five
steps in the domino process) allows regio- and stereocon-
Chem. Eur. J. 2008, 14, 3078 – 3092
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3087

Y. Coquerel, J. Rodriguez et al.
Table 8. The MARDi cascade for the preparation azepanes, azepines and thiepines.[a]
Entry
Substrate
Aldehyde
Product[b]
Yield [%][c]
1[d]
18a
2a,b and 2g
64–70
2[e]
18b
2a
19 (25a), 17 (25b: R=Me)
3[g]
4[h]
18b
18b
2a
2a
25a and 25b
25a and 25b
8 (25a), 25 (25b: R=Me)
0 (25a) 31 (25b: R=Me) 51 (25b: R=H)
5[i]
18b
2n
10 (25c) 22 (25d)
6[k]
7[e]
8[i]
18b
2g
2a
2b
7
19
55
19
58
3088
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3078 – 3092

The MARDi Cascade
Table 8. (Continued)
FULL PAPER
Entry
Substrate
Aldehyde
Product[b]
Yield [%][c]
9[i]
19
2n
28
10[k]
19
2g
51 (26d) 17 (26e)
11[k]
19
2i
42
[a] Unless stated otherwise, all reactions were performed in dry MeOH (0.04m) with 1.5 equiv of aldehyde at room temperature for 20 h. [b] For 25a,
25d, 26b, 26c and 26e, the major isomer is the thermodynamically favoured isomer (a-OH or b-CO2Me). The relative configurations of compounds 25a,
25c–e and 26b–f have been established by 2D NMR spectroscopic techniques and the structure of 25c has been secured by X-ray diffraction analysis
(see reference [16]). [c] Isolated. [d] K2CO3 or DBU (0.5 or 1 equiv). [e] K2CO3 (0.5 equiv). [f] dr=3.5:1. [g] K2CO3 (1 equiv). [h] K2CO3 (1.5 equiv).
[i] DBU (1 equiv). [j] dr=1.6:1. [k] DBU (0.5 equiv). [l] dr=5:1. [m] dr=1.3:1. [n] dr=17:1. [o] dr=10:1.
and distilled under an argon atmosphere. Petroleum ether refers to the
trolled access to a variety of functionalised and substituted
seven-membered rings. This new multicomponent reaction is
a condensation and thus no byproducts are formed, except
for water when dehydration is observed. The substitution
array can be diastereoselectively modulated by appropriate
choice of the reaction partners, and the reaction allows the
control of up to five newly created stereocentres and a com-
plete chiral induction in the case of an optically active
ketone precursor in a single operation. The high level of dia-
stereoselectivity observed has been attributed to total ther-
modynamic control of the reaction. The attractiveness of the
present three-component domino approach to seven-mem-
bered rings resides in the diversity of carbo- and heterocy-
clic structures that can be accessed with total regiocontrol
and high stereocontrol by starting from simple substrates,
under user and environmentally friendly conditions, as now
required in modern organic chemistry. Our work now focus-
es on the asymmetric version of the MARDi cascade and its
applications to the total synthesis natural products.
fraction of petroleum ether that was distilled between 40 and 658C. Alde-
hydes 2 and the Dieckmann esters 1a,b were distilled prior use. The reac-
tions were monitored by TLC, which was performed on Merck 60F254
plates and visualised with an ethanolic solution of p-anisaldehyde and
sulfuric acid or an ethanolic solution of molybdophosphoric acid. Flash
chromatography was performed with Merck 230–400 mesh silica gel.
NMR spectroscopic data were recorded on a Brüker AC 200, Avance 300
or Avance 400 spectrometer in CDCl3 or (CD3)2CO, and chemical shifts
(d) are given in ppm relative to the residual non-deutered signal for
1H NMR spectra (CHCl3: d=7.26 ppm) and relative to the deuterated
solvent signal for 13C NMR spectra (CDCl3: 77.0 ppm, (CD3)2CO:
29.8 ppm); coupling constants (J) are in Hertz, and the classical abbrevia-
tions are used to describe the signal multiplicity; peak assignment and
relative configurations have been established from standard COSY,
NOESY, HMQC and HMBC experiments. Mass spectra were recorded
on an API III Plus Sciex spectrometer equipped with an electrospray ion-
isation source and a triple quadripole detector or on a Brüker Esquire
6000 spectrometer equipped with an electrospray ionisation source and
an ion-trap detector. Optical rotations were recorded on a Perkin–Elmer
241 polarimeter at 258C, path length=10 cm, by using a sodium lamp as
the light source. Melting points are uncorrected and were measured with
a Büchi B-540 apparatus. Elemental analyses were performed on a
Thermo Finnigan EA 1112 analyser.
General procedure for compounds of type 4: The procedure for (+)-4h
is representative for all compounds of type 4.
Experimental Section
Compound 4h: An oven dried, two-necked, round-bottomed flask under
an argon atmosphere and equipped with a magnetic Teflon-coated stir-
ring bar was charged at room temperature with b-ketoester (+)-1c
(498 mg, 3.19 mmol) and MeOH (10 mL) in that order. At 08C, DBU
(480 mL, 3.22 mmol) was added and the mixture was stirred for 10 min,
then crotonaldehyde (2b) (400 mL, 4.83 mmol) was added. The reaction
mixture was slowly warmed to room temperature and stirred at that tem-
perature for 21 h. Most of the methanol was then removed under reduced
pressure and water (20 mL) was added; the resulting aqueous medium
General: All reactions were performed in oven-dried glassware under an
argon atmosphere. All reagents were obtained from commercial sources
and used as supplied unless otherwise stated. MeOH was dried by reflux-
ing with magnesium and then distilled under an argon atmosphere.
K2CO3 and Cs2CO3 were dried by prolonged storage at 1408C in an
oven. Pyridine and triethylamine were dried over solid KOH and distilled
under an argon atmosphere. HMPA was dried by refluxing with CaH2
Chem. Eur. J. 2008, 14, 3078 – 3092
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3089

Y. Coquerel, J. Rodriguez et al.
was extracted three times with diethyl ether and the combined organic
layers were washed with brine, dried with anhydrous sodium sulfate, fil-
trated and then concentrated to give the crude product. The crude prod-
uct was purified by flash chromatography on silica gel with diethyl ether
in petroleum ether as the eluent to give 357 mg (43%) of 4h as slightly
yellow oil. Rf =0.37 (Et2O/petroleum ether 9:1); [a]D =+28.3 (c=1.15 g
per 100 mL in CH2Cl2); 1H NMR (400 MHz, CDCl3): d=3.92 (ddd, 3J-
1.66–2.05 (m, 4H), 1.21 (t, 3J
A
N
7.2 Hz, 3H); 13C NMR (50 MHz, CDCl3): d=181.0 (C), 168.0 (C), 148.9
(CH), 133.4 (C), 60.6 (CH2), 42.1 (CH), 34.5 (CH2), 30.3 (CH), 27.0
(CH2), 24.1 (CH2), 21.7 (CH3), 14.1 ppm (CH3); MS (ESI-) m/z (%): 225
(100) [MꢀH]+; elemental analysis calcd (%) for C12H18O4: C 63.70, H
8.02; found: C 63.55, H 7.74.
Cycloheptandiol 9: A solution of osmium tetroxide in toluene (0.10m,
4.3 mL, 0.430 mmol) was added to a solution of 8a (84 mg, 0.396 mmol)
in diethyl ether/pyridine (5:1, 6 mL) at 08C. The reaction mixture was
stirred at 08C for 30 min and then 40% NaHSO3 (4 mL) was added. The
reaction mixture was vigorously stirred at room temperature for 13 h and
then concentrated HCl (1 mL) was added. The organic layer was separat-
ed and the remaining aqueous layer was extracted four times with ethyl
acetate. The combined organic layers were dried with anhydrous sodium
sulfate, filtrated and then concentrated to give the crude product which
was recrystallised from ethyl acetate/hexane by slow evaporation at room
temperature to provide 81 mg (83%) of 9 as white needles suitable for
X-ray diffraction analysis: Rf =0.46 (AcOEt); m.p. 1378C; 1H NMR
ACHTREUNG
A
ACHTREUNG
AHCTREUNG
A
ACHTREUNG
the hydrogen atom of the hydroxyl group was not detected; 13C NMR
(100 MHz, CDCl3): d=176.2 (C), 175.8 (C), 73.5 (CH), 53.2 (CH), 51.7
(CH3), 51.5 (CH3), 43.2 (CH), 33.7 (CH2), 27.5 (CH), 25.0 (CH), 24.0
(CH2), 22.6 (CH3), 22.6 ppm (CH3); MS (ESI+): m/z (%): 281 (100)
[M+Na]+; elemental analysis calcd (%) for C13H22O5: C 60.45, H 8.58;
found: C 60.34, H 8.49.
The characterization data for the cycloheptanols 4a, 5a, 4c and 4e–g
have been reported previously (see reference [11a]).
(300 MHz, CDCl3): d=3.81 (s, 3H), 3.49 (brs, 1H), 3.44 (d, 3J
9.9 Hz, 1H), 2.67 (dddd, 3J
(H,H)=10.5, 6.7, 6.7, 3.1 Hz, 1H), 2.20 (dddd,
3J(H,H)=14.3, 10.9, 10.8, 2.7 Hz, 1H), 1.70–2.05 (m, 6H), 1.59 (ddd, 3J-
(H,H)=15.6, 9.1, 6.8 Hz, 1H), 1.06 ppm (d, J
Cycloheptanols 4b and 5b (1.5:1 mixture of isomers): 13C NMR
(50 MHz, CDCl3): d=175.5 (C), 175.4 (C), 175.0 (C), 174.9 (C), 73.1
(CH), 72.1 (CH), 60.0 (CH2), 59.9 (CH2), 59.8 (CH2), 59.8 (CH2), 53.7
(CH), 52.9 (CH), 43.9 (CH), 42.1 (CH), 33.4 (CH2), 32.3 (CH2), 28.8
(CH2), 27.1 (CH2), 25.2 (CH2), 24.6 (CH2), 23.8 (CH2), 23.6 (CH2), 13.6
(CH3), 13.6 ppm (CH3); MS (ESI+): m/z (%): 281 (100) [M+Na]+; ele-
mental analysis calcd (%) for C13H22O5: C 60.45, H 8.58; found: C 60.81,
H 8.89.
AHCTREUNG
AHCTREUNG
3
A
U
peak masked; 13C NMR (75 MHz, (CD3)2CO): d=20.5 (CH3), 22.9
(CH2), 32.7 (CH), 33.2 (CH2), 35.1 (CH2), 41.4 (CH), 52.6 (CH3), 79.9
(C), 81.2 (CH), 177.3 (C), 177.5 ppm (C); MS (ESI-) m/z (%): 245 (100)
[MꢀH]ꢀ.
General procedure for compounds of type 15: The procedure for 15b is
representative for compounds of type 15.
Cycloheptanol 4d: Rf =0.36 (Et2O/petroleum ether 9:1); 1H NMR
3
(200 MHz, CDCl3): d=3.95 (dt, J
3.65 (s, 3H), 2.65 (d, J
A
Compound 15b: An oven dried, two-necked, round-bottomed flask
under an argon atmosphere and equipped with a magnetic Teflon-coated
stirring bar was charged at 08C with b-ketosulfone 12a (401 mg,
1.79 mmol), 1:1 MeOH/THF (30 mL) and K2CO3 (246 mg, 1.79 mmol) in
that order. The resulting mixture was stirred at 08C for 5 min and then
crotonaldehyde (2b) (220 mL, 2.66 mmol) was added. The reaction mix-
ture was slowly warmed to room temperature, stirred at this temperature
for 20 h (including ramp-up time), filtered through a pad of Celite and
then concentrated. The solid obtained was partially dissolved in dichloro-
methane, filtrated over Celite and concentrated again to give 524 mg
(90%) of the clean crude product. The major diastereomer was precipi-
tated from diethyl ether and recrystallised from ethanol to give 295 mg
(51%) of pure 15b as white needles suitable for X-ray diffraction analy-
sis. Rf =0.74 (AcOEt); m.p. 1448C; 1H NMR (300 MHz, CDCl3): d=
7.84–7.89 (m, 2H), 7.61–7.69 (m, 1H), 7.52–7.59 (m, 2H), 3.99 (brt, 3J-
3
ACHTREUNG
(m, 6H), 1.05–1.65 (m, 4H), 0.82 ppm (t, 3J
drogen atom of the hydroxyl group was not detected; 13C NMR (50 MHz,
CDCl3): d=176.3 (C), 175.5 (C), 73.2 (CH), 54.1 (CH), 51.4 (CH3), 51.2
(CH3), 49.0 (CH), 35.1 (CH), 38.3 (CH2), 38.2 (CH2), 27.6 (CH2), 23.1
(CH2), 19.3 (CH2), 13.8 ppm (CH3); MS (ESI+): m/z (%): 295 (100)
[M+Na]+, 273 [M+H]+; elemental analysis calcd (%) for C14H24O5: C
61.74, H 8.88; found: C 61.92, H 9.02.
General procedure for compounds of type 8: The procedure for 8a is
representative for compounds of type 8.
Compound 8a: An oven dried, two-necked, round-bottomed flask under
an argon atmosphere and equipped with a magnetic Teflon-coated stir-
ring bar was charged at room temperature with b-ketoester 1a (561 mg,
3.95 mmol), MeOH (10 mL) and DBU (590 mL, 3.95 mmol) in that order.
The resulting pale-yellow mixture was stirred at room temperature for
10 min, and then methacroleine (2g) (490 mL, 5.93 mmol) was added and
the reaction mixture was stirred at the same temperature for 18 h. Most
of the methanol was then removed under reduced pressure and water
(25 mL) was added; the resulting basic aqueous medium was extracted
twice with diethyl ether and then acidified to pH 1 with HCl (1n) solu-
tion. The acidic aqueous layer was extracted three times with diethyl
ether and the combined organic layers were dried with anhydrous
sodium sulfate, filtrated and concentrated to give the crude clean product
as a colourless oil suitable for analysis. 1H NMR (400 MHz, CDCl3): d=
(H,H)=9.6 Hz, 1H), 3.68 (s, 3H), 2.76 (ddd, 3J
ACHTREUNG
3
ACHTREUNG
1H), 2.49–2.66 (m, 2H), 2.37 (ddd, J
2.24 (m, 2H), 1.93 (ddd, 3J
AHCTREUNG
H
ACHTREUNG
A
ACHTREUNG
3H); 13C NMR (75 MHz, CDCl3): d=175.1 (C), 137.6 (C), 133.8 (CH),
129.2 (2CH), 128.9 (2CH), 70.7 (CH), 68.2 (CH), 54.4 (CH), 52.0 (CH3),
39.5 (CH2), 27.4 (CH), 26.5 (CH2), 24.8 (CH2), 22.8 ppm (CH3); MS
(ESI+): m/z (%): 344 (100) [M+NH4]+, 327 [M+H]+; elemental analysis
calcd (%) C16H22O5S: C 58.87, H 6.79, S 9.82; found: C 58.47, H 6.66, S
9.62.
6.83 (d, 3J
9.0 Hz, 1H), 2.70–2.80 (m, 1H), 2.61 (dd, 3J
2.52 (dd, 3J(H,H)=16.0, 9.0 Hz, 1H), 2.03 (ddd, 3J
2.7 Hz, 1H), 1.98 (ddd, 3J
(H,H)=14.0, 5.0, 2.2 Hz, 1H), 1.66–1.83 (m,
2H), 1.20 ppm (d, 3J(H,H)=7.2 Hz, 3H); 13C NMR (100 MHz, CDCl3):
A
N
The characterization data for the cycloheptanols 15a,c–e have been re-
ported previously (see reference [21]).
AHCTREUNG
A
ACHTREUNG
AHCTREUNG
The experimental protocol and the characterization data for the hetero-
cyclic seven-membered rings 21a–e, 25a–e, 26a–d and 26 f have been re-
ported previously (see reference [30]). The structures of compounds 13,
14, 20, 22 and 23 were determined by 13C NMR and mass spectroscopy of
the crude material and the products were not further characterised.
ACHTREUNG
d=181.8 (C), 168.6 (C), 149.3 (CH), 133.3 (C), 51.9 (CH3), 42.3 (C), 34.6
(CH2), 30.5 (CH), 27.1 (CH2), 24.3 (CH2), 21.8 ppm (CH3); MS (ESI-):
m/z (%): 211 (100) [MꢀH]+; elemental analysis calcd (%) for C11H16O4:
C 62.25, H 7.60; found: C 62.09, H 7.75.
Thiepine trans-26e: This compound has been prepared by treatment of
the corresponding carboxylic acid 26d with TMSCHN2 as described in
reference [30]. Rf =0.56 (AcOEt/petroleum ether 3:7); 1H NMR
The characterization data for the cycloheptenic acids 8b–i have been re-
ported previously (see reference [18]). These data are actually for the
products with the trans relative configuration, and not cis as published at
the time.
(300 MHz, CDCl3): d=6.93 (d, 3J
(s, 3H), 3.26 (dd, 3J
ACHTREUNG
A
ACHTREUNG
Cycloheptenic acid 8j: 1H NMR (200 MHz, CDCl3): d=6.82 (d, 3J-
4.4, 4.1, 4.0 Hz, 1H), 2.87–2.98 (m, 1H), 2.87 (dd, 3J
ACHTREUNG
A
G
1H), 1.91–2.09 (m, 2H), 1.16 ppm (d, 3J
ACHTREUNG
3090
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3078 – 3092

The MARDi Cascade
FULL PAPER
(75 MHz, CDCl3): d 173.9 (C), 165.9 (C), 148.9 (CH), 130.2 (C), 52.6
(CH3), 52.0 (CH3), 43.2 (CH), 33.8 (CH2), 33.1 (CH2), 31.3 (CH),
22.2 ppm (CH3); MS (ESI+): m/z (%): 362 (100) [M+NH4]+, 245
[M+H]+; elemental analysis calcd (%) for C11H16O4S: C 54.08, H 6.60, S
13.12; found: C 54.37, H 6.76, S 12.94.
[9] For other reports on two-carbon ring expansion of five-membered
rings see: a) M. Miesch, G. Mislin, M. Franck-Neumann, Tetrahe-
dron Lett. 1998, 39, 6873–6876; b) T. L. Dzwiniel, J. M. Stryker, J.
Am. Chem. Soc. 2004, 126, 9184–9185; c) S. R. Schulz, S. Blechert,
Angew. Chem. 2007, 119, 4040–4044; Angew. Chem. Int. Ed. 2007,
46, 3966–3970.
[10] For reviews see: a) L. Yet, Tetrahedron 1999, 55, 9349–9403 and L.
Yet, Chem. Rev. 2000, 100, 2963–3007; b) E. J. Kantorowski, M. J.
Kurth, Tetrahedron 2000, 56, 4317–4353; c) M. E. Maier, Angew.
Chem. 2000, 112, 2153–2157; Angew. Chem. Int. Ed. 2000, 39, 2073–
2077.
[11] a) M.-H. Filippini, R. Faure, J. Rodriguez, J. Org. Chem. 1995, 60,
6872–6882; b) M.-H. Filippini, J. Rodriguez, Chem. Rev. 1999, 99,
27–76.
Acknowledgements
We thank Dr. M. Giorgi (UniversitØ Paul CØzanne) for quality X-ray
structure determinations, Dr. R. Faure (UniversitØ Paul CØzanne) for
helpful assistance with NMR spectroscopic structure determinations, Dr.
P. Crucianni and Dr. M. Malacria (UniversitØ Paris VI) for a generous
gift of aldehyde 2k, Dr. J. Viala (UniversitØ Paul CØzanne) for assistance
with artwork, and Dr. J.-M. Pons and Dr. M. Rajzmann (UniversitØ Paul
CØzanne) for theoretical calculations. Financial support from the French
Research Ministry, the CNRS and the UniversitØ Paul CØzanne
(UMR 6178) is also gratefully acknowledged.
[12] a) G. Stork, H. K. Landesman, J. Am. Chem. Soc. 1956, 78, 5129–
5130; b) J. B. Hendrickson, R. K. Boeckman, Jr., J. Am. Chem. Soc.
1971, 93, 1307–1308.
[13] a) W. G. Dauben, J. W. MacFarland, J. Am. Chem. Soc. 1960, 82,
4245–4248; b) C. A. Grob, J. Hostynek, Helv. Chim. Acta 1963, 46,
2209–2225; c) G. L. Buchanan, C. Maxwell, W. Henderson, Tetrahe-
dron 1965, 21, 3273–3276; d) G. L. Buchanan, G. W. McLay, Tetra-
hedron 1966, 22, 1521–1525; e) G. L. Buchanan, G. A. R. Young, J.
Chem. Soc. Chem. Commun. 1971, 643–645; f) R. Chakraborty,
M. K. Basu, B. C. Ranu, Tetrahedron 1992, 48, 8849–8854.
[14] For a recent related radical-mediated fragmentation of bicyclo-
[1] Step economy: a) P. A. Wender, F. C. Bi, G. G. Gamber, F. Gosselin,
R. D. Hubbard, M. J. C. Scanio, R. Sun, T. J. Williams, L. Zhang,
Pure Appl. Chem. 2002, 74, 25–31; b) P. A. Wender, J. L. Baryza,
S. E. Brenner, M. O. Clarke, G. G. Gamber, J. C. Horan, T. C.
Jessop, C. Kan, K. Pattabiraman, T. J. Williams, Pure Appl. Chem.
2003, 75, 143–155; c) P. A. Wender, J. L. Baryza, S. E. Brenner,
M. O. Clarke, M. L. Craske, J. C. Horan, T. Meyer, Curr. Drug Dis-
covery Technol. 2004, 1, 1–11; d) P. A. Wender, G. G. Gamber, R. D.
Hubbard, S. M. Pham, L. Zhang, J. Am. Chem. Soc. 2005, 127,
2836–2837.
[2] Atom economy: a) B. M. Trost, Science 1991, 254, 1471–1477; b) B.
M. Trost, Angew. Chem. 1995, 107, 285–287; Angew. Chem. Int. Ed.
1995, 34, 259–281; c) B. M. Trost, Acc. Chem. Res. 2002, 35, 695–
705.
[3] For a recent special issue on green chemistry see Chem. Rev. 2007,
107, 2167–2820.
[4] a) For a recent monograph see: Multicomponent Reactions (Eds.: J.
Zhu, H. BienaymØ), Wiley-VCH, Weinheim, 2005; b) for a special
issue on MCRs see Tetrahedron 2005, 61, 11299–11519; for some
recent reviews of MCRs see: c) H. BienaymØ, C. Hulme, G. Oddon,
P. Schmitt, Chem. Eur. J. 2000, 6, 3321–3329; d) A. Dçmling, I. Ugi,
Angew. Chem. 2000, 112, 3300–3344; Angew. Chem. Int. Ed. 2000,
39, 3168–3210; e) J. Zhu, Eur. J. Org. Chem. 2003, 1133–1144;
f) R. V. Orru, M. de Greef, Synthesis 2003, 1471–1499; g) D. J.
Ramón, M. Yus, Angew. Chem. 2005, 117, 1628–1661; Angew.
Chem. Int. Ed. 2005, 44, 1602–1634; h) A. Dçmling, Chem. Rev.
2006, 106, 17–89.
[5] For representative reviews see: a) L. F. Tietze, Chem. Rev. 1996, 96,
115–136; b) L. F. Tietze, M. E. Lieb, Curr. Opin. Chem. Biol. 1998,
2, 363–371; c) J. Rodriguez, Synlett 1999, 505–518; d) for a recent
monograph see: L. F. Tietze, G. Brasche, K. M. Gericke, Domino
Reactions in Organic Synthesis, Wiley-VCH, Weinheim, 2006.
[6] For reviews see: a) S. L. Schreiber, Science 2000, 287, 1964–1969;
b) A. Ulaczyk-Lesanko, D. G. Hall, Curr. Opin. Chem. Biol. 2005, 9,
266–276; c) J. Gerencser, G. Dorman, F. Darvas, QSAR Comb. Sci.
2006, 25, 439–448; d) D. Tejedor, F. Garcia-Tellado, Chem. Soc. Rev.
2007, 36, 484–491.
[3.2.1]octanes see: T. K. Pradhan, A. Hassner, Synlett 2007, 1071–
1074.
[15] For another ring-expansion strategy of cyclic b-ketoesters for the
preparation of medium-size rings initiated by conjugated addition
see: I. Hachiya, W. Maehara, Y. Yamada, T. Kamiki, M. Shimizu,
Synlett 2006, 3271–3274.
[16] CCDC 620411 (4c), 615710 (9), 615708 (15b), 615709 (15c) and
620412 (25c) contain the supplementary crystallographic data for
this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/
data_request/cif
[17] a) S. A. Achmad, G. W. K. Cavill, Aust. J. Chem. 1963, 16, 858–868;
b) J. Wolinsky, D. Chan, J. Org. Chem. 1965, 30, 41–43; c) J. N.
Marx, L. R. Norman, J. Org. Chem. 1975, 40, 1602–1606.
[18] A part of these results has been reported in a preliminary communi-
cation: M.-H. Filippini, J. Rodriguez, J. Org. Chem. 1997, 62, 3034–
3035; the structure of the cycloheptenic acid 8 has been revisited
and confirmed to be as depicted in Scheme 5. The incorrect 1,3-cis
stereochemistry previously reported for 8 is a result of the misinter-
pretation of an artifact on the NOESY spectrum, which indeed
clearly shows the 1,3-trans stereochemistry (see discussion).
[19] R. S. Dickson, H. P. Kirch, Aust. J. Chem. 1972, 25, 1815–1818.
[20] The enthalpies of formation, which were kindly provided by one of
the reviewers of reference [18] and calculated with PCMODEL for
both 1,3-cis and 1,3-trans isomers of 8a, corroborate our own results
obtained from AM1/RHF (AMPAC) semiempirical calculations and
showed a preference of 1.1–2.2 kcal in favour of the 1,3-cis-diequa-
torial isomer, which corresponds to an approximate 80:20 to 95:5
mixture (1,3-cis major) if thermodynamic equilibrium is achieved.
See the Supporting Information.
[21] A part of these results has been reported in a preliminary communi-
cation: a) Y. Coquerel, D. Bensa, V. Moret, J. Rodriguez, Synlett
2006, 2751–2754; b) Y. Coquerel, D. Bensa, V. Moret, J. Rodriguez,
Synlett 2006, 3368–3368.
[22] For the synthesis of 9c see: a) S. H. Bertz, J. M. Cook, A. Gawish,
U. Weiss, Org. Synth. 1986, 64, 27–37; b) M. Harmata, P. Rashatasa-
khon, Tetrahedron Lett. 2002, 43, 3641–3644.
[23] B. M. Trost, T. N. Salzmann, K. Hiroi, J. Am. Chem. Soc. 1976, 98,
4887–4902.
[24] M. C. CarreÇo, J. L. Garcꢁa Ruano, C. Pedregal, A. Rubio, J. Chem.
Soc. Perkin Trans. 1 1989, 1335–1337.
[7] a) C. Simon, T. Constantieux, J. Rodriguez, Eur. J. Org. Chem. 2004,
4957–4980; b) F. LiØby-Muller, C. Simon, T. Constantieux, J. Rodri-
guez, QSAR Comb. Sci. 2006, 25, 432–438; c) F. LiØby-Muller, T.
Constantieux, J. Rodriguez, J. Am. Chem. Soc. 2005, 127, 17176–
17177; d) H. Habib-Zahmani, J. Viala, S. Hacini, J. Rodriguez, Syn-
lett 2007, 1037–1042; e) F. LiØby-Müller, T. Constantieux, J. Rodri-
guez, Synlett 2007, 1323–1325.
[25] a) G. A. Kraus, J. Hansen, D. Vines, Synth. Commun. 1992, 22,
2625–2634; b) B. M. Trost, D. P. Curran, Tetrahedron Lett. 1981, 22,
1287–1290.
[8] M.-H. Filippini, J. Rodriguez, M. Santelli, J. Chem. Soc. Chem.
Commun. 1993, 1647–1648.
Chem. Eur. J. 2008, 14, 3078 – 3092
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3091

Y. Coquerel, J. Rodriguez et al.
[26] a) N. Hashimoto, T. Aoyama, T. Shiori, Chem. Pharm. Bull. 1981,
29, 1475–1478; for the mechanism see: b) E. Kühnel, D. D. P.
Laffan, G. C. Lloyd-Jones, T. Martꢁnez del Campo, I. R. Shepperson,
J. L. Slaughter, Angew. Chem. 2007, 119, 7205–7208; Angew. Chem.
Int. Ed. 2007, 46, 7075–7078.
[27] The pKa values of a-substituted acetones MeCOCH2G measured in
DMSO are 12.5 for G=SO2Ph, and 14.2 for G=CO2Et. See:
a) F. G. Bordwell, J. A. Harrelson, Jr., Can. J. Chem. 1990, 68, 1714–
1718; b) F. G. Bordwell, Acc. Chem. Res. 1988, 21, 456–463.
T. Fletcher, D. J. Garland, W. Huang, C. Jones, K. J. Koeller, S. A.
Kolodziej, J. Li, R. E. Manning, M. W. Mahoney, R. E. Miller, D. A.
Mischke, N. P. Rath, E. J. Reinhard, M. B. Tollefson, W. F. Vernier,
G. M. Wagner, S. R. Rapp, J. Beaudry, K. Glenn, K. Regina, J. R.
Schuh, M. E. Smith, J. S. Trivedi, D. B. Reitz, J. Med. Chem. 2005,
48, 5853–5868.
[29] For an overview of the previously developed synthetic approaches
to heterocyclic seven-membered rings see references [2–17] cited in
reference [30] of the present article.
[28] For oxepanes see: a) T. Yasumoto, M. Murata, Chem. Rev. 1993, 93,
1897–1909; b) D. J. Faulkner, Nat. Prod. Rep. 1998, 15, 113–158; for
azepanes see, for example: c) A. R. Carrol, E. Hyde, J. Smith, R. J.
Quinn, G. Guymer, P. I. Forster, J. Org. Chem. 2005, 70, 1096–1099;
d) M. C. De la Fuente, S. E. Pullan, I. Biesmans, D. Domꢁnguez, J.
Org. Chem. 2006, 71, 3963–3966; e) H. Li, Y. BlØriot, J.-M. Mallet,
E. Rodriguez-Garcia, P. Vogel, Y. Zhang, P. Sinaꢂ, Tetrahedron:
Asymmetry 2005, 16, 313–319; for thiepanes see, for example: f) S.
Dubaele, W. Jahnke, J. Schoepfer, J. Fuchs, P. Chꢃne, Bioorg. Med.
Chem. Lett. 2006, 16, 923–927; g) S. J. Tremont, L. F. Lee, H.-C.
Huang, B. T. Keller, S. C. Banerjee, S. R. Both, A. J. Carpenter, C.-
C. Wang, D. J. Garland, W. Huang, C. Jones, K. J. Koeller, S. A. Ko-
lodziej, J. Li, R. E. Manning, M. W. Mahoney, R. E. Miller, D. A.
Mischke, N. P. Rath, T. Fletcher, E. J. Reinhard, M. B. Tollefson,
W. F. Vernier, G. M. Wagner, S. R. Rapp, J. Beaudry, K. Glenn, K.
Regina, J. R. Schuh, M. E. Smith, J. S. Trivedi, D. B. Reitz, J. Med.
Chem. 2005, 48, 5837–5852; h) H.-C. Huang, S. J. Tremont, L. F. Lee
B. T. Keller, A. J. Carpenter, C.-C. Wang, S. C. Banerjee, S. R. Both,
[30] A part of these results has been reported in a preliminary communi-
cation: Y. Coquerel, D. Bensa, A. Doutheau, J. Rodriguez, Org.
Lett. 2006, 8, 4819–4822.
[31] P. Dowd, S.-C. Choi, Tetrahedron 1991, 47, 4847–4860.
[32] a) M. P. Moyer, P. L. Feldman, H. Rapoport, J. Org. Chem. 1985, 50,
5223–5230; b) M. A. Calter, P. M. Sugathapala, C. Zhu, Tetrahedron
Lett. 1997, 38, 3837–3840; c) K. Jones, T. Toutounji, Tetrahedron
2001, 57, 2427–2431.
[33] F. Lacrampe, F. LØost, A. Doutheau, Tetrahedron Lett. 2000, 41,
4773–4776.
[34] This behaviour has previously been reported when no aldolisation
was allowed after the Michael addition: M.-H. Filippini, J. Rodri-
guez, Synth. Commun. 1995, 25, 245–252.
[35] M. McHugh, G. R. Proctor, J. Chem. Res. Miniprint 1984, 8, 2230–
2253.
Received: October 30, 2007
Published online: February 6, 2008
3092
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3078 – 3092