Elucidation of architectural requirements from a spacer in supported proline-based catalysts of enantioselective aldol reaction

Article abstract of DOI:10.1002/adsc.200800734

In order to delineate the properties of the spacer architecture responsible for the strong positive dendritic effect exhibited by polymer-supported proline-based catalysts, we prepared two series of polystyrene-bound model catalysts. The first series was

Full text of DOI:10.1002/adsc.200800734

COMMUNICATIONS
DOI: 10.1002/adsc.200800734
Elucidation of Architectural Requirements from a Spacer in
Supported Proline-Based Catalysts of Enantioselective Aldol
Reaction
Kerem Goren,a,b Tzofit Kehat,a,b and Moshe Portnoya,*
a
School of Chemistry, Raymond and Beverly Sackler Faculty of Exact Sciences, Tel Aviv University, Tel Aviv, 69978, Israel
Fax : (+972)-3-6409293; e-mail: portnoy@post.tau.ac.il
The first two authors contributed equally to this work
b
Received: October 5, 2008; Revised: November 26, 2008; Published online: January 12, 2009
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/adsc.200800734.
Abstract: In order to delineate the properties of the
spacer architecture responsible for the strong posi-
tive dendritic effect exhibited by polymer-supported
proline-based catalysts, we prepared two series of
polystyrene-bound model catalysts. The first series
was based on a linear and partially dendritic spacers
(of reduced branching and valency) imitating the
length of the second generation spacer, while the
second series was based on the first generation den-
dron spacer with one functional (proline-terminat-
ed) and one non-functional arm. Comparative stud-
ies of the model and original (fully dendritic) cata-
lysts in the asymmetric aldol reaction of aromatic
aldehydes with acetone disclose the features charac-
teristic to the dendritic architecture, such as prox-
imity between the terminal catalytic units and en-
hanced branching, as crucial for inducing higher
yield and enantioselectivity in catalysis.
of various derivates,[3] in spite of a question mark on
the economic profitability of developing a successful
immobilized proline catalyst because of the low price
of this amino acid and the possibility of its easy recov-
ery from homogeneous reaction mixtures. Moreover,
immobilization frequently provides catalysts with an al-
tered reactivity profile and can indirectly afford valua-
ble mechanistic information. In earlier studies, unfortu-
nately, the immobilization of proline led to reduced se-
lectivity,[4] and only recently have a limited number of
covalently heterogenized proline-based catalysts, ex-
hibiting enantioselectivity approaching that of their
soluble analogues, been reported.[5] While these cata-
lysts demonstrated excellent results in the aldol reac-
tions with cyclic ketones, only a few supported proline-
containing peptides could match the performance of
proline in the reaction with acetone.[5a–c]
Recently, we observed that the introduction of a
short dendritic spacer between the polymer core and
the 4-hydroxyproline-derived catalytic units immobi-
lized via azide-alkyne “click” chemistry provides a
more active and remarkably more enantioselective
catalyst, as compared to the spacer-less non-dendritic
analogue. (Scheme 1).[6]
Keywords: aldol reaction; asymmetric organocataly-
sis; dendrimers; solid-phase synthesis; supported
catalysts
Although this was an extraordinary manifestation
of a positive dendritic effect, the first of its kind in
Over the past years a growing number of studies in supported organocatalysis,[7] the question of the possi-
the field of asymmetric catalysis were devoted to or- ble origin of the effect remains unresolved. Positive
ganic catalysts, metal-free small molecules, capable of dendritic effects were observed in the past in support-
promoting chemical transformation with high efficien- ed organometallic catalysis, but their possible explan-
cy and enantioselectivity.[1] Although heterogenized ations were based (in those cases discussed) on the
catalysts benefit from a number of economic, environ- proximity and cooperativity of the ligating sites in the
mental and technical advantages as compared to their metal coordination events, but not on cooperative in-
homogeneous analogues, reports of successful immo- teraction of two ligands or complexes with the sub-
bilization of enantioselective organocatalysts on solid strates.[8] In the case of organocatalysts such as pro-
supports remain relatively scarce.[2] l-Proline provides line, however, it is tempting to suggest cooperative
an excellent model for organocatalyst immobilization action of two proximal proline units on the substrates,
studies due to its simplicity and the ready availability in order to explain the aforementioned effect.
Adv. Synth. Catal. 2009, 351, 59 – 65
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
59

Kerem Goren et al.
COMMUNICATIONS
onstrated, particularly by List and Houk and Barbas,
that in solution only one proline molecule is associat-
ed with the substrates, intermediates and product
along the catalytic cycle.[9] Moreover, it was recently
demonstrated that all earlier reports claiming non-
linear effects in this reaction, originate from the non-
homogeneity of the reaction mixture with the proline
catalyst being in equilibrium between the solution
and solid phases.[11]
In the light of these reports, we wondered whether
the dendritic effect, which we had observed, was due
to the alternative bimolecular catalytic mechanism,
operative in supported catalysis only and benefiting
from the bi- or multi-valency of the dendritic spacers,
or whether it was a result of the other properties of
the dendritic architecture of the spacer and the mech-
anism of the catalysis is solely unimolecular. In this
manuscript we describe the synthesis of model pro-
line-based supported catalysts incorporating linear,
dendritic and partially dendritic spacers and the com-
parative catalytic study of these systems with the pre-
viously reported dendritic analogues, which may shed
light on the question presented above.
Scheme 1. Asymmetric aldol reaction with immobilized hy-
droxyproline-derived catalysts.[6]
While in the previous communication we reported
the synthesis of a linear analogue of the G1 catalyst,
On the other hand, over the years unequivocal ex- herein we focus on analogues of the G2 catalytic den-
perimental evidence was disclosed, supported by the- dron. Possible modes of dendritic arm truncation are
oretical study, for a unimolecular mechanism of the depicted in Scheme 2. Truncation of one of the lower
proline catalysis of the aldol reaction.[9,10] It was dem- branches (shown in green) leads to the structure
Scheme 2. Truncated spacer models of the G2 catalyst.
60
asc.wiley-vch.de
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 59 – 65

Architectural Requirements from a Spacer in Supported Proline-Based Catalysts
COMMUNICATIONS
G2(mono,di-Pro), which preserves the mutual posi- clearly points out that proximity between two proline
tions of the proline units (particularly the through- moieties is critical for achieving higher yield and
bonds distance), but reduces the branching in the den- enantioselectivity. The ee sharply decreases as we pro-
dritic spacer (the branched/linear unit ratio). Trunca- ceed from the bivalent structures [G2(mono,di-Pro),
tion of two upper branches (shown in red) leads to G2(di,mono-Pro)] to the monovalent catalyst
the structure G2(di,mono-Pro), which preserves a cer- G2(mono,mono-Pro). For the bivalent structures a
tain proximity between the two remaining proline shorter average distance between the proline units is
units, but is likely to increase the average distance be- clearly preferred [G2(mono,di-Pro) vs. G2(di,mono-
tween them, while reducing again the branching of Pro)]. While it would be difficult to estimate this dis-
the spacer. Truncation of all branches but one (shown tance due to the many degrees of freedom in the
in blue) leads to a linear-spacer catalyst G2(mono, structures, it is clear that the through-bonds distance
mono-Pro).
between the proline units will be the major contribu-
The synthesis of the three truncated spacers is de- ting factor to the distance parameter. Thus, in
picted in Scheme 3. As in the case of the perfect den- G2(mono,di-Pro) (as well as in all fully dendritic
dritic structures, the branching units of the spacers structures) there are 16 covalent bonds connecting
were derived from dimethyl-5-hydroxyisophthalate the two proline units. This must lead to a much short-
building block, which is immobilized via nucleophilic er through-space separation of the units than in the
substitution and activated by a reduction-chlorodehy- case of G2(di,mono-Pro) with 26 covalent bonds sep-
droxylation sequence.[12] The linear units were derived arating the two proline units. Formally, even
from 3-hydroxybenzyl alcohol, which was immobilized G2(mono,mono-Pro), G1ACTHNUGRTENNUG(mono-Pro) and G0ACHTUNGTRENNUNG(Pro)
via nucleophilic substitution and activated by a chloro- can be considered “multivalent” catalysts, since pro-
dehydroxylation reaction. The preparation of the line units in these polymers are covalently connected
azide-terminated spacers and the propargyl-carrying to the common polystyrene matrix. However, in these
protected hydroxyproline as well as the cycloaddition resins the amount of the bonds separating the prolines
and deprotection reactions, leading to the active sup- is much higher even in the optimal case: 56, 46 and 36
ported catalytic systems, were carried out by proce- bonds, respectively, and only if the linkers are at-
dures used for the perfect dendrons without substan- tached to the neighboring repeating units of the poly-
tial changes.[6]
styrene chain. On average the “separating bond
The three new supported catalytic systems were count” for these catalysts will be significantly higher,
tested in the aldol addition reaction of acetone to and thus, their multivalency is purely fictional.
benzaldehyde or 4-nitrobenzaldehyde and compared
to the previously obtained catalysts (Table 1). The G2ACHTUNGTRENNUNG
One can erroneously conclude that the tetravalent
(Pro) is preferred over the bivalent catalyst. How-
trend in the G2 series (entries 3–6 and 9–12) very ever, the comparison with the G1 series points out
Scheme 3. Synthesis of the G2 catalyst models. Reagents and conditions: a) dimethyl 5-hydroxyisophthalate, LiH, TBAI,
DMF, 608C, overnight; b) LiBH4, B(OMe)3, THF, 608C, overnight; c) PPh3, C2Cl6, THF, room temperature, overnight; d)
AHCTUNGTRENNUNG
NaN3, TBAI, DMF, 608C, 24 h; e) A, sodium ascorbate, CuSO4, DMF, 508C, overnight; f) 0.2% TFA in DCM, room temper-
ature, 5 min; g) LiOH, THF/H2O, 40 8C, 4 h; h) 3-hydroxybenzyl alcohol, LiH, TBAI, DMF, 608C, overnight.
Adv. Synth. Catal. 2009, 351, 59 – 65
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
61

Kerem Goren et al.
Table 1. Aldol reaction with G1 and G2 catalysts incorporating dendritic, partially dendritic and linear spacers.[a]
COMMUNICATIONS
Entry
Catalyst
R
Conversion [%]
Yield [%][b]
ee [%][c]
1
2
3
4
5
6
7
8
G1
G1
G2
N
H
H
H
H
H
H
NO2
NO2
NO2
NO2
NO2
NO2
73
61
52
36
58
44
43
35
95
79
94
95
94
71
68
28
68
56
45
24
85
23
84
61
51
25
100
100
100
69
100
91
100
100
100
71
G2(mono,di-Pro)
G2(di,mono-Pro)
G2(mono,mono-Pro)
G1
G1
G2
G2(mono,di-Pro)
G2(di,mono-Pro)
G2(mono,mono-Pro)
9
10
11
12
[a]
Reaction conditions: 1 equiv. of benzaldehyde/4-nitrobenzaldehyde, 27 equiv. of acetone and 0.3 equiv. of catalyst in
DMSO for 4 days at room temperature.
NMR yield.
[b]
[c]
The ee was determined by HPLC, using Chiralpak AD (R=H) or Chiralcel OJ (R=NO2) columns.
that the differences between G2
G2(mono,di-Pro) are due to the reduced branching/ amount of by-products was formed (Table 2, entry 4).
increased flexibility of the spacer, since G1(Pro), Although the difference between G1(OPh,Pro) and
which is bivalent, is almost as good as G2(Pro). The G1(Cp,Pro) can be explained by the lower steric bulk
ACHTUNGTREN(NUNG Pro) and the expense of the catalyst chemoselectivity as a large
A
ACHTUNGTRENNUNG
A
ACHTUNGTRENNUNG
deterioration in the performance upon reduction of of the non-functional arm of the former, the influence
the “branching per length unit” parameter of the of the triazole cannot be ruled out.[5f]
spacer is further evidenced by comparison of
G2(mono, mono-Pro) to G1(mono-Pro) in the ben- previously obtained it seems that the branching
zaldehyde reaction (entries 2 and 6). nature of the spacers contributes to the improved se-
From the comparison of these results with those
AHCTUNGTRENNUNG
Having established these trends, we decided to pre- lectivity of the dendritic catalysts, but the major con-
pare two additional G1-analogues, which preserve the tribution comes from the multi- or bivalency of the
branching nature of the spacer, but contain only one dendritic catalyst and relative proximity of the proline
proline-functionalized arm, while the other arm lacks units. This proximity can potentially be translated into
the proline unit or any other nucleophilic or acidic interaction of the units with each other and/or the
site (Scheme 4). Both structures were based on the substrates, possibly through hydrogen bonding. Al-
methyl 3-hydroxy-5-hydroxymethylbenzoate hetero- though at the moment we lack direct proof, indirect
functional branching unit that was immobilized on support for such interactions can also be provided by
Wang Bromo polystyrene via nucleophilic substitution a substantial decrease in the dendritic effect in the re-
(similar to the other phenolic modules used in the action of benzaldehyde with acetone upon substitu-
synthesis of perfect or truncated dendrons, vide tion of 15% of DMSO by MeOH, a polar protic sol-
supra). Then the non-functional arm was assembled vent, which usually disrupts hydrogen bonding. In this
first via chlorodehydroxylation-phenoxydechlorina- solvent mixture, under conditions otherwise equal to
tion or chlorodehydroxylation-azidodechlorination the abovementioned, the dendritic catalysts exhibited
“click” cycloaddition chemistry. Once the non-func- lower chemoselectivity (in spite of high aldehyde con-
tional arm was in place, the ester was reduced, thus sumption) as compared to the non-dendritic analogue
again forming the hydroxymethyl group for the func- [yields of 14, 8 and 4% for G0
ACHUTGTNRENN(GU Pro), G1CAHTUNGTRENN(UGN Pro) and
tional arm assembly via the five-step sequence as in G2(Pro) respectively]. Moreover, the improvement in
AHCTUNGTRENNUNG
the preceding synthetic schemes.
The new model catalyst G1
marginally better than G1(mono-Pro) (Table 2, catalysts, respectively). Although the change in the re-
the enantioselectivity due to dendronization was mini-
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
entry 3 vs. 2). The model G1ACHTUNTRGNEUNG(Cp,Pro) was a somewhat action media can influence the aldol reaction in many
more active and enantioselective catalyst (46% ee) in different ways,[13] these findings fall in line with the
this reaction, but this improvement was achieved at model studies described above.
62
asc.wiley-vch.de
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 59 – 65

Architectural Requirements from a Spacer in Supported Proline-Based Catalysts
COMMUNICATIONS
In recent years a number of homogeneous cata-
lysts with two proline-derived units in the catalyst
molecule were prepared and studied in the asym-
metric aldol reaction.[14] Of particular interest could
be their comparison with the mono-proline ana-
logues. However, such an analogue with a truly
non-functional moiety replacing the second proline
unit was only once reported by Benaglia.[14a] PEG
with two O-tethered hydroxyproline units was ex-
amined alongside with MeO-PEG functionalized
with one such unit. Interestingly, when compared
under similar conditions (DMSO, 20–24 h) and
equal loading of proline moieties, the performance
of the bis-proline catalyst was notably better than
that of the mono-proline analogue (67 vs. 36% yield
and 74 vs. 60% ee). Although the two proline units
in this PEG-based catalyst are separated by a signif-
icantly longer spacer than in our systems, such coin-
cidence with our findings may imply a general phe-
nomenon of enhanced activity/enantioselectivity in
bivalent proline-based systems and will be a subject
for our further investigation.
In conclusion, we have demonstrated using
simple models of supported dendritic catalysts,
based on linear or more sophisticated partially den-
dritic spacers, that the proximity of the two proline
units is crucial and responsible for achieving higher
yield and enantioselectivity in the aldol reaction.
High branching of the spacer probably provides a
benefit of a smaller magnitude, while an incremen-
tal positive contribution of the triazole connecting
unit is also a possibility.
Whether these findings imply that in the general
case of bivalent or polyvalent proline-decorated cat-
alysts the mechanism of the reaction deviates from
the monomolecular pathway, established in the liter-
ature, remains to be investigated.
Experimental Section
General Procedure for 1,3-Dipolar Cycloaddition
Propargylated derivative (5 equiv. per azidobenzyl unit),
sodium ascorbate (1 equiv. per azidobenzyl unit, 1M in
DMF) and copper(II) sulfate pentahydrate (0.25 equiv.
per azidobenzyl unit) were added to a suspension of poly-
styrene azidobenzyl-terminated resin (1 equiv.) in DMF
(10 mL per 1 g resin). The suspension was heated to 508C
overnight. The resin was washed with DMF/water, DMF,
THF/water, THF, DCM and then dried under vacuum.
General Procedure for Deprotection of Proline
Cleavage of trityl protecting group: The protected proline
resin (1 equiv.) was washed 3 times for 5 min with a solu-
tion of 0.1% TFA and 1% H2O in DCM (10 mL per g
resin).
Adv. Synth. Catal. 2009, 351, 59 – 65
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
63

Kerem Goren et al.
COMMUNICATIONS
Table 2. Aldol reaction with G1 catalysts with two and one functional arms.[a]
Entry
Catalyst
G1(Pro)
R
Conversion [%]
Yield [%][b]
ee [%][c]
1
2
3
4
5
6
7
G
H
H
H
H
NO2
NO2
NO2
73
61
70
100
100
91
52
36
36
21
95
79
78
68
28
31
46
85
23
24
G1(mono-Pro)
G1(OPh,Pro)
G1(Cp,Pro)
G1(Pro)
G1(mono-Pro)
G1(OPh,Pro)
T
E
N
R
AHCTUNGTRENNUNG
C
92
[a]
Reaction conditions: 1 equiv. of benzaldehyde/4-nitrobenzaldehyde, 27 equiv. of acetone and 0.3 equiv. of catalyst in
DMSO for 4 days at room temperature.
NMR yield.
[b]
[c]
The ee was determined by HPLC, using Chiralpak AD (R=H) or Chiralcel OJ (R=NO2) columns.
Hydrolysis of methyl ester: Lithium hydroxide (5 equiv.
[2] a) M. Benaglia, A. Puglisi, F. Cozzi, Chem. Rev. 2003,
103, 3401–3430; b) F. Cozzi, Adv. Synth. Catal. 2006,
348, 1367–1390.
[3] a) B. List, Tetrahedron 2004, ##60##58, 5573–5590;
b) G. Guillena, C. Nꢂjera, D. J. Ramꢃn, Tetrahedron:
Asymmetry 2007, 18, 2249–2293.
[4] a) K. Kondo, T. Yamano, K. Takemoto, Makromol.
Chem. 1985, 186, 1781–1785; b) D. Dhar, I. Beadham,
S. Chandrasekaran, Proc. Indian Acad. Sci. Chem. Sci.
2003, 115, 365–372.
per methyl ester unit) was dissolved in THF/H2O (10:1)
(10 mL per 1 g resin) and then added to a suspension of the
proline methyl ester-terminated resin (1 equiv.) in THF
(10 mL per 1 g resin). The suspension was heated to 408C
for 4 h. The resin was washed with THF/H2O, THF, DCM
and then dried under vacuum.
General Procedure for the Aldol Reaction
The catalytic resin (0.3 mmol of proline units, 0.3 equiv.) was
added to a mixture of DMSO:acetone 4:1 (8 mL:2 mL). The
suspension was stirred for 5 min at room temperature and
then the aldehyde (1 mmol, 1 equiv.) was added. The sus-
pension was mixed at room temperature for 4 days. The
resin was separated from the solution by filtration and
washed with ethyl acetate. Water (10 mL) and saturated
aqueous NH4Cl solution (10 mL) were added to the com-
bined filtrate. The mixture was extracted with ethyl acetate
(3ꢁ10 mL). The organic phase was dried on MgSO4. The
solvent was evaporated, and the crude material was ana-
lyzed to determine conversion and yield, and then chroma-
tographed on a silica gel column (1:9 EtOAc:hexanes up to
3:7 EtOAc:hexanes) to yield the pure product as a yellow
oil. The product ee was determined by HPLC, using Chiral-
pak AD (benzaldehyde product) or Chiralcel OJ (4-nitro-
benzaldehyde product) columns.
[5] a) J. D. Revell, D. Gantenbein, P. Krattiger, H. Wen-
nemers, Biopolymers 2006, 84, 105–113; b) M. R. M.
Andreae, A. P. Davis, Tetrahedron: Asymmetry 2005,
16, 2487–2492; c) K. Akagawa, S. Sakamoto, K. Kudo,
Tetrahedron Lett. 2005, 46, 8185–8187; d) D. Font, C.
Jimeno, M. A. Pericꢄs, Org. Lett. 2006, 8, 4653–4655;
e) D. Font, A. Bastero, S. Sayalero, C. Jimeno, M. A.
Pericꢄs, Org. Lett. 2007, 9, 1943–1946; f) D. Font, S.
Sayalero, A. Bastero, C. Jimeno, M. A. Pericꢄs, Org.
Lett. 2008, 10, 337–340; g) F. Giacalone, M. Gruttadau-
ria, A. M. Marculescu, R. Noto, Tetrahedron Lett. 2007,
48, 255–259; h) M. Gruttadauria, F. Giacalone, A. M.
Marculescu, P. L. Meo, S. Riela, R. Noto, Eur. J. Org.
Chem. 2007, 4688–4698; i) F. Calderꢃn, R. Fernꢂndez,
F. Sꢂnchez, A. Fernꢂndez-Mayoralas, Adv. Synth. Catal.
2005, 347, 1395–1403; j) E. G. Doyagꢅez, F. Calderꢃn,
R. Fernꢂndez, F. Sꢂnchez, A. Fernꢂndez-Mayoralas, J.
Org. Chem. 2007, 72, 9353–9356.
[6] T. Kehat, M. Portnoy, Chem. Commun. 2007, 2823–
2825.
Acknowledgements
[7] Although supported dendritic amino alcohol catalysts
for Et2Zn addition to aldehydes were reported, the cat-
alytic cycle involved organometallic complexes:
a) Y. M. Chung, H. K. Rhee, Chem. Commun. 2002,
238–239; b) Y. M. Chung, H. K. Rhee, Compt. Rend.
Chim. 2003, 6, 695–705.
This research was supported by The Israel Science Founda-
tion (Grant No. 960/06).
[8] a) A. Dahan, M. Portnoy, Chem. Commun. 2002, 2700–
2701; b) A. Dahan, M. Portnoy, J. Am .Chem. Soc.,
2007, 129, 5860–5869.
References
[1] a) P. I. Dalko, L. Moisan, Angew. Chem. 2004, 116,
5248–5286; Angew. Chem. Int. Ed. 2004, 43, 5138–
5175; b) H. Pellissier, Tetrahedron 2007, 63, 9267–9331.
[9] a) L. Hoang, S. Bahmanyar, K. N. Houk, B. List, J. Am.
Chem. Soc. 2003, 125, 16–17; b) S. Bahmanyar, K. N.
64
asc.wiley-vch.de
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 59 – 65

Architectural Requirements from a Spacer in Supported Proline-Based Catalysts
COMMUNICATIONS
Houk, J. Am. Chem. Soc. 2001, 123, 11273–11283; K.
Sakthivel, W. Notz, T. Bui, C. F. Barbas III, J. Am.
Chem. Soc. 2001, 123, 5260–5267.
[13] For solvent effects on proline catalysis in solution see,
for instance: a) B. List, R. A. Lerner, C. F. Barbas III,
J. Am. Chem. Soc. 2000, 122, 2395–2396; b) B. List, P.
Pojarliev, C. Castello, Org. Lett. 2001, 3, 573–575.
[14] a) M. Benaglia, M. Cinquini, F. Cozzi, A. Puglishi, G.
Celentano, Adv. Synth. Catal. 2002, 344, 533–542;
b) M. Benaglia, G. Celentano, F. Cozzi, Adv. Synth.
Catal. 2001, 343, 171–173; c) S. Guizzetti, M. Benaglia,
L. Pignataro, A. Puglishi, Tetrahedron: Asymmetry
2006, 17, 2754–2760; d) G. Guillena, M. C. Hita, C.
Najera, S. F. Viozquez, Tetrahedron: Asymmetry 2007,
18, 2300–2304.
[10] a) M. E. Jung, Tetrahedron 1976, 32, 3–31; b) K. L.
Brown, L. Damm, J. D. Dunitz, A. Eschenmoser, R.
Hobi, C. Kratky, Helv. Chim. Acta 1978, 61, 3108–
3135; c) D. Rajagopal, M. S. Moni, S. Subramanian, S.
Swaminathan, Tetrahedron: Asymmetry 1999, 10, 1631–
1634.
[11] M. Klussmann, H. Lwamura, S. P. Mathew, D. H. Wells
Jr, U. Pandya, A. Armstrong, D. G. Blackmond, Nature
2006, 441, 621–623.
[12] A. Mansour, T. Kehat, M. Portnoy, Org. Biomol.
Chem. 2008, 6, 3382–3387.
Adv. Synth. Catal. 2009, 351, 59 – 65
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
65