Total synthesis of bacilosarcins A and B

Article abstract of DOI:10.1002/anie.200804571

(Chemical Equation Presented) I want to ride my bicycle: The thermodynamic stability of nitrogen-containing heterocyclic ring systems is exploited in the first enantioselective total synthesis of bacilosarcins A and B, which has been achieved in simple tw

Full text of DOI:10.1002/anie.200804571

Communications
DOI: 10.1002/anie.200804571
Natural Product Synthesis
Total Synthesis of Bacilosarcins A and B**
Masaru Enomoto and Shigefumi Kuwahara*
In the course of screening for plant growth regulators of
microbial origin, Igarashi and co-workers discovered two
novel herbicidal substances, bacilosarcins A and B, in the
culture broth of the bacterium Bacillus subtilis TP-B0611
isolated from intestinal contents of the sardine Sardinops
melanostica. Extensive spectroscopic analyses and chemical
conversions allowed the elucidation of their structures as 1
and 2, respectively (Scheme 1).[1] The growth inhibitory
products.[4] The structural novelty of these two natural
products and the potent herbicidal activity of 1 prompted us
to embark on the total synthesis of 1 and 2, as well as
structure–activity relationship studies to develop a new type
of herbicide, which would also lead to the elucidation of their
action mechanism. We describe herein the first total synthesis
of bacilosarcins A (1) and B (2) together with a new efficient
synthesis of AI-77-B (3).
Bacilosarcin A (1), which has the unusual heterobicyclic
ring system, can be traced retrosynthetically to two simpler
precursors, 4 (amicoumacin A)[5] and 2,3-butanedione, by
dissecting the hemiacetal and N,N-acetal bonds of the bicyclic
ring portion (Scheme 2). We expected that the formation of
Scheme 1. Structures of Bacilosarcins A (1), B (2), and AI-77-B (3).
activity of bacilosarcin A (1) was remarkably more potent
against barnyard millets than that of herbimycin A (an
ansamycin antibiotic with potent herbicidal activity),[2] while
the herbicidal activity of 2 was weak. To date, the mode of
action of 1 and 2 remains entirely unknown. From a structural
viewpoint, both 1 and 2 belong to a small family of natural
products that is characterized by a dihydroisocoumarin ring
system linked to an unusual amino acid, as represented by AI-
77-B (3; known as an antiulcerogenic substance produced by
Bacillus pumilus).[3] At first glance, the molecular architec-
tures of bacilosarcins A and B captured our interest, since the
3-oxa-6,9-diazabicyclo[3.3.1]nonane ring system incorporated
in 1 was totally unprecedented both in natural and synthetic
compounds, and the 2-hydroxymorpholine substructure con-
tained in 2 was a very rare structural motif in natural
[*] M. Enomoto, Prof. Dr. S. Kuwahara
Laboratory of Applied Bioorganic Chemistry
Graduate School of Agricultural Science, Tohoku University
Tsutsumidori-Amamiyamachi, Aoba-ku, Sendai 981-8555 (Japan)
Fax: (+81)22-717-8783
Scheme 2. Retrosynthetic analysis of 1 and 2.
E-mail: skuwahar@biochem.tohoku.ac.jp
[**] We thank Prof. Yasuhiro Igarashi (Toyama Prefectural University) for
providing spectra of the bacilosarcins and valuable discussions.
Financial support for this work was provided by the Ministry of
Education, Culture, Sports, Science, and Technology (Japan) (no.
19380065).
the heterobicyclic ring system of 1, in which the hydroxy
group at the C3 position and the amino groups at the C4 and
C6 positions of 4 participate, could take place preferentially
by simply mixing 4 and 2,3-butanedione under appropriate
acidic conditions, since molecular modeling studies showed
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200804571.
1144
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 1144 –1148

Angewandte
Chemie
that any other mode of cyclization would lead to less stable
bicyclic rings. The desired ring formation that leads to 1
should be favored by the thermodynamically preferred
orientations of the substituents on the bicyclic ring of 1—
the axially oriented hydroxy group at the anomeric C2’
position as well as the three equatorially oriented alkyl groups
at the C3, C2’, and C3’ positions (see the conformational
diagram in Scheme 2). The cyclization precursor 4 could be
obtained from 6 (amicoumacin C)[3d] by selective ammonol-
ysis of its g-lactone. On the other hand, the 2-hydroxymor-
pholine ring of 2 should be formed spontaneously from 5,
which could also be elaborated from 6 through N-alkylation
of the amino group on the g-lactone ring followed by
ammonolysis of the resulting product. Since the stereochem-
ical control at the nitrogen-bearing C3’ stereogenic center of 5
seemed to be difficult because of its proclivity for enolization,
we decided to conduct the N-alkylation in a nonstereoselec-
tive manner. Instead we envisaged that the enolizability at the
C3’ position of 5 could be exploited for the stereoconvergent
formation of 2 by equilibration of the two epimers of 5, since
the formation of the 3’b-epimer of 2 seemed to be difficult
because of severe 1,3-diaxial steric repulsion between its 3’b-
methyl group and 4b-substituent. The common amide inter-
mediate 6 was further divided into amine segment 7 and acid
segment 8. To date, various synthetic routes to 7 (or its non-
methylated derivative) and analogues of 8 have been
developed by many research groups in their efforts toward
the total synthesis of AI-77-B (3),[6–8] most of which success-
fully converted natural amino acids or sugars into the amine
and acid segments. Any of these routes, however, seemed to
leave room for improvement in reproducibility, selectivity,
and conciseness. We planned to prepare 7 and 8 in a more
efficient manner by utilizing the intermolecular epoxide ring
opening of 10 with the aromatic nucleophile 9 and the
intramolecular epoxide ring opening of 11, respectively.
Our synthesis of the amine segment 7 began with the
transformation of epoxy alcohol 12 into azido epoxide 10 with
inversion of configuration (Scheme 3);[9] 12 in turn was
readily produced in high yields (> 90%) on multigram
scales (> 10 g) by the Sharpless kinetic resolution of 5-
methyl-1-hexen-3-ol.[10] The epoxide ring of 10 was opened
under the conditions reported by Ganem and co-workers with
the lithium anion generated from 13.[11,12] Treatment of the
resulting azido alcohol 14 with acid afforded lactone 15 (96%
ee, as determined by HPLC analysis using a chiral stationary
phase). Finally, 15 was reduced to 7 by catalytic hydro-
genation. Thus, our new synthesis of 7 was accomplished in
27% overall yield from 12 in only four highly reproducible
steps.
Scheme 4 shows our six-step synthesis of the acid segment
8 from the previously reported aldehyde 16.[13] Asymmetric
epoxidation of 16 under the organocatalytic conditions
reported by Cꢀrdova and co-workers proceeded smoothly
to give 18.[14] Chain elongation of 18 by Andoꢁs protocol
Scheme 4. Preparation of acid segment 8. Reagents and conditions:
a) aq H2O2, 17, CHCl3, RT, 91%; b) (PhO)2P(O)CH2CO2tBu, NaH,
THF, ꢀ788C!RT, 69%; c) TFA, CH2Cl2, RT, 75%; d) NaN3, aq AcOH,
RT, 51% (71% based on recovered starting material); e) H2, PtO2,
Boc2O, EtOH, RT, 82%; f) H2, Pd/C, EtOH, RT, quantitative. TFA=tri-
fluoroacetic acid, Boc=tert-butoxycarbonyl.
afforded a 7:1 mixture of 11 and its E isomer.[15] Treatment of
the mixture with TFA in CH2Cl2 brought about a smooth
intramolecular epoxide ring-opening reaction to give hydroxy
lactone 19,[16] which was then reacted with sodium azide in
aqueous acetic acid.[17] The resulting conjugate adduct, which
was obtained in 51% yield (71% based on recovered starting
material) as a 5:1 mixture of 20 and its epimer at the azide-
bearing stereogenic center was subjected to catalytic hydro-
genation in the presence of Boc2O to furnish 21 (97% ee, as
determined from the NMR spectra of the corresponding (R)-
and (S)-MTPA esters (MTPA = a-methoxy-a-(trifluorome-
thyl)phenylacetyl). Finally, hydrogenolysis of the benzyl ether
resulted in the quantitative formation of 8. Thus, the synthesis
of 8 was achieved in an overall yield of 20% from 16 in a
concise six-step sequence.
Having succeeded in the efficient short-step syntheses of
the amine segment 7 and the acid segment 8, we turned our
attention to their condensation and subsequent elaboration
into bacilosarcin A (1; Scheme 5). Condensation of 7 and 8
proceeded smoothly to give amide 22.[18] Unmasking of the
methyl-protected phenolic hydroxy and Boc-protected amino
groups of 22 were effected in one pot by treating 22 with BBr3/
CH2Cl2 in the presence of anisole to afford 6 (amicoumacin C,
isolated as its hydrochloride salt by treatment of 6 with HCl/
MeOH).[19] The hydrochloride salt of 6 was treated with
methanolic ammonia to give 4 (amicoumacin A), which was
then mixed with 2,3-butanedione under various sets of acidic
Scheme 3. Preparation of amine segment 7. Reagents and conditions:
a) (PhO)2P(O)N3, DEAD, Ph3P, THF, RT, 69%; b) nBuLi, BF3OEt2,
toluene, ꢀ788C, 48%; c) CSA, toluene, 1008C, 87%; d) H2, 10% Pd/C,
EtOH, RT, 94%. DEAD=diethyl azodicarboxylate, CSA=camphorsul-
fonic acid.
Angew. Chem. Int. Ed. 2009, 48, 1144 –1148
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1145

Communications
Scheme 5. Completion of the synthesis of bacilosarcin A (1). Reagents
and conditions: a) EDC·HCl, DMAP, HOBt, CH2Cl2/DMF, ꢀ108C,
62%; b) BBr3, anisole, CH2Cl2, ꢀ78 to ꢀ58C, 73%; c) NH3(7m)/
MeOH, ꢀ58C; d) NH4Cl, CH3CN/MeOH/H2O, RT, 57% (2 steps);
e) aq NaOH (0.02m, pH 9) then aq HCl (0.02m, pH 6.5), EtOH/H2O,
90%. EDC=1-ethyl-3-(3-dimethylaminopropyl)carbodiimide,
DMAP=4-dimethylaminopyridine, HOBt=1-hydroxybenzotriazole.
Scheme 6. N alkylation of model substrate 23 and morpholine ring
formation from the resulting products, 25a and 25b.
conditions. Among the reaction conditions attempted, treat-
ment of 4 with a catalytic amount of NH4Cl in CH3CN/MeOH
containing a small amount of water gave the best result, and
furnished the desired cyclization product 1 in 57% yield from
6·HCl. The spectral data of 1 were identical to those of the
natural product.[1] Thus, the first total synthesis of bacilosar-
cin A was achieved in 5.1% overall yield from 16 in 10 steps.
The hydrochloride salt of 6 was also converted into AI-77-B
(3) by following a reported procedure.[6] The new synthesis of
AI-77-B was accomplished in 8.0% overall yield from 16 in
9 steps.
For the synthesis of bacilosarcin B (2), we needed to
install a 3-oxobut-2-yl substituent at the amino group of 6.
This N-alkylation reaction was examined by using model
substrate 23, which was obtained as its TFA salt by
deprotection of 21 with TFA (Scheme 6). After several
unsuccessful attempts (see the Supporting Information), we
found that the N-alkylation could be achieved by simply
treating 23·TFAwith 3-hydroxy-2-butanone in the presence of
MgSO4 and NaHCO3 in CH3CN, to give a 1:1 epimeric
mixture of 25a and 25b in an excellent yield of 92% after
column chromatography. It is proposed that this reaction,
known as the Amadori rearrangement,[20] proceeds via the
imine intermediate 24, followed by tautomerization to the
two epimers. Fortunately, the epimers were found to be
separable by careful preparative TLC, which provided us with
the opportunity to investigate the aforementioned possibility
that both epimers of 5 (Scheme 2) could be transformed
stereoconvergently into the thermodynamically more stable
stereoisomer 2 by epimerization at the C3’ position of 5.
Monitoring the reaction by TLC showed that when the less
polar epimer 25a was subjected to ammonolysis conditions, a
clean reaction took place, from which 27a was isolated in
43% yield.[21] On the other hand, exposure of the more polar
epimer 25b to the same conditions gave a complex mixture in
which, however, we could detect a considerable amount of
27a by 1H NMR spectroscopy. These results suggest that
firstly, the less polar epimer, to which we assigned the
structure 25a, was first converted into 26a under the
ammonolysis conditions; this was then converted smoothly
into 27a, which possesses the equatorially oriented 3’a-
methyl group; and secondly, in the case of the more polar
epimer 25b, the ring formation of ammonolysis product 26b
into 27b was precluded because of the severe steric repulsion
between the C4 carbamoylmethyl group and the axially
oriented C3’ methyl substituent in 27b. This repulsion
resulted in the observed formation of 27a by enolization of
26b to 28 followed by tautomerization to 26a.
With the encouraging results in the model studies in hand,
we proceeded to the final stage of the total synthesis of
bacilosarcin B (2; Scheme 7). When 6·HCl was subjected to
the Amadori rearrangement conditions, 29 was formed in
72% yield.[22] Finally, treatment of 29 with NH3/EtOH (2m)
furnished 2 in 39% yield (isolated as its hydrochloride salt by
treatment of 2 with HCl/MeOH) via intermediate 5. In this
case, we could not obtain any direct evidence for the
1146
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 1144 –1148

Angewandte
Chemie
[1] M. Azumi, K. Ogawa, T. Fujita, M. Takeshita, R. Yoshida, T.
Furumai, Y. Igarashi, Tetrahedron 2008, 64, 6420.
[2] For herbimycin A, see: a) S. Omura, Y. Iwai, Y. Takahashi, N.
Sadakane, A. Nakagawa, J. Antibiot. 1979, 32, 255; b) S. Omura,
A. Nakagawa, N. Sadakane, Tetrahedron Lett. 1979, 20, 4323.
[3] a) J. Itoh, T. Shomura, S. Omoto, S. Miyado, Y. Yuda, U. Shibata,
S. Inouye, Agric. Biol. Chem. 1982, 46, 1255; b) Y. Shimojima, H.
Hayashi, T. Ooka, M. Shibukawa, Agric. Biol. Chem. 1982, 46,
1823; c) Y. Shimojima, H. Hayashi, T. Ooka, M. Shibukawa, Y.
Iitaka, Tetrahedron Lett. 1982, 23, 5435; d) J. Itoh, S. Omoto, N.
Nishizawa, Y. Kodama, S. Inouye, Agric. Biol. Chem. 1982, 46,
2659; e) Y. Shimojima, H. Hayashi, T. Ooka, M. Shibukawa, Y.
Iitaka, Tetrahedron 1984, 40, 2519.
[4] For examples of natural products containing a 2-hydroxymor-
pholine core, see: a) H. P. Zhang, Y. Kamano, H. Kizu, H.
Itokawa, G. R. Pettit, C. L. Herald, Chem. Lett. 1994, 2271;
b) R. P. Maskey, I. Grꢂn-Wollny, H. H. Fiebig, H. Laatsch,
Angew. Chem. 2002, 114, 623; Angew. Chem. Int. Ed. 2002, 41,
597.
Scheme 7. Completion of the synthesis of bacilosarcin B (2). Reagents
and conditions: a) MgSO4 (5.0 equiv), NaHCO3 (1.0 equiv), CH3CN,
758C, 72%; b) NH3(2m)/EtOH, 39%.
[5] J. Itoh, S. Omoto, T. Shomura, N. Nishizawa, S. Miyado, Y. Yuda,
U. Shibata, S. Inouye, J. Antibiot. 1981, 34, 611.
[6] For total syntheses of 3, see: a) Y. Hamada, A. Kawai, Y. Kohno,
O. Hara, T. Shioiri, J. Am. Chem. Soc. 1989, 111, 1524; b) S. D.
Broady, J. E. Rexhausen, E. J. Thomas, J. Chem. Soc. Chem.
Commun. 1991, 708; c) Y. Hamada, O. Hara, A. Kawai, Y.
Kohno, T. Shioiri, Tetrahedron 1991, 47, 8635; d) R. A. Ward, G.
Procter, Tetrahedron Lett. 1992, 33, 3359; e) J.-M. Durgant, P.
Vogel, Helv. Chim. Acta 1993, 76, 222; f) R. A. Ward, G. Procter,
Tetrahedron 1995, 51, 12301; g) S. D. Broady, J. E. Rexhausen,
E. J. Thomas, J. Chem. Soc. Perkin Trans. 1 1999, 1083; h) H.
Kotsuki, T. Araki, A. Miyazaki, M. Iwasaki, P. K. Datta, Org.
Lett. 1999, 1, 499; i) A. K. Ghosh, A. Bischoff, J. Cappiello, Org.
Lett. 2001, 3, 2677; j) A. K. Ghosh, A. Bischoff, J. Cappiello, Eur.
J. Org. Chem. 2003, 821.
[7] For syntheses of 7 or its non-methylated derivative, see: a) H.
Kotsuki, A. Miyazaki, M. Ochi, Chem. Lett. 1992, 1255; b) L.
Bertelli, R. Fiaschi, E. Napolitano, Gazz. Chim. Ital. 1993, 123,
669; c) S. Superchi, F. Minutolo, D. Pini, P. Salvadori, J. Org.
Chem. 1996, 61, 3183; d) A. K. Ghosh. J. Cappiello, Tetrahedron
Lett. 1998, 39, 8803.
[8] For syntheses of the analogs of 8, see: a) A. Kawai, O. Hara, Y.
Hamada, T. Shioiri, Tetrahedron Lett. 1988, 29, 6331; b) J. P.
Gesson, J. C. Jacquesy, M. Mondon, Tetrahedron Lett. 1989, 30,
6503; c) N. Ikota, A. Hanaki, Chem. Pharm. Bull. 1989, 37, 1087;
d) Y. Hamada, A. Kawai, T. Matsui, O. Hara, T. Shioiri,
Tetrahedron 1990, 46, 4823; e) K. Shinozaki, K. Mizuno, H.
Wakamatsu, Y. Masaki, Chem. Pharm. Bull. 1996, 44, 1823; f) C.
Mukai, M. Miyakawa, M. Hanaoko, J. Chem. Soc. Perkin Trans. 1
1997, 913; g) G. Davies, A. T. Russell, Tetrahedron Lett. 2002, 43,
8519; h) T. Muramatsu, S. Yamashita, Y. Nakamura, M. Suzuki,
N. Mase, H. Yoda, K. Takabe, Tetrahedron Lett. 2007, 48, 8956.
[9] B. Lal, B. N. Pramanik, M. S. Manhas, A. K. Bose, Tetrahedron
Lett. 1977, 18, 1977.
stereoconvergent formation of 2 from both the 3’a and 3’b
epimers of 5, since the epimeric mixture 29 could not be
separated. However, the detection of a trace amount of 4
(amicoumacin A, Scheme 5) in the crude product mixture by
NMR spectroscopy was considered to indirectly indicate the
intervention of equilibrium reactions (Scheme 6). This is
because 4, which lacks the 3-oxobut-2-yl substituent, would
probably be generated by isomerization of 5 to the corre-
sponding a-hydroxy imine intermediate and its hydrolysis
with a trace amount of water contained in the reaction
solvent. The spectral data of 2·HCl were identical with those
of natural bacilosarcin B.[23] The overall yield of 2 from 16 was
2.5% (10 steps).
In conclusion, the first total syntheses of bacilosarcins A
(1) and B (2) that contain unusual heterocyclic ring systems in
their molecular architectures have been accomplished by the
thermodynamically controlled ring-forming reaction between
cyclization precursor 4 (amicoumacin A) and 2,3-butane-
dione for 1, and the nonstereoselective N-alkylation of 6
(amicoumacin C) with 3-hydroxy-2-butanone under Amadori
rearrangement conditions coupled with stereoconvergent
cyclization under equilibrium conditions for 2. The common
intermediates 7 and 8 were prepared by highly reproducible
short-step sequences involving inter- and intramolecular
epoxide ring-opening reactions, respectively. The improved
accessibility to 7 and 8 enabled us to achieve a new efficient
synthesis of AI-77-B (3), and will also facilitate the prepara-
tion of analogues of 1 and 2 for structure–activity relationship
studies to develop new types of herbicides.
[10] M. Bessodes, M. Saꢃah, K. Antonakis, J. Org. Chem. 1992, 57,
4441.
[11] M. J. Eis, J. E. Wrobel, B. Ganem, J. Am. Chem. Soc. 1984, 106,
3693.
[12] For the preparation of 13, see a) R. Casas, C. Cavꢄ, J. dꢁAngelo,
Tetrahedron Lett. 1995, 36, 1039; b) A. V. Kalinin, J. F. Bower, P.
Riebel, V. Snieckus, J. Org. Chem. 1999, 64, 2986.
Received: September 17, 2008
Revised: November 30, 2008
Published online: December 30, 2008
[13] Compound 16 was obtained in 60–80% yields by the ozonolysis
of benzyl sorbate: a) J. Anaya, D. H. R. Barton, M. C. Caballero,
S. D. Gero, M. Grande, N. M. Laso, J. I. M. Hernando, Tetrahe-
dron: Asymmetry 1994, 5, 2137; b) T. Veysoglu, L. A. Mitscher,
J. K. Swayze, Synthesis 1980, 807.
Keywords: alkaloids · epoxidation · heterocycles ·
natural products · total synthesis
.
Angew. Chem. Int. Ed. 2009, 48, 1144 –1148
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1147

Communications
[14] a) H. Sundꢄn, I. Ibrahem, A. Cꢀrdova, Tetrahedron Lett. 2006,
ascribed to the difficulty in the isolation of 27a because of its
47, 99; b) M. Marigo, J. Franzꢄn, T. B. Poulsen, W. Zhuang, K. A.
Jørgensen, J. Am. Chem. Soc. 2005, 127, 6964.
extremely high polarity.
[22] The H NMR spectrum of 29 indicated that a small degree of
1
[15] K. Ando, J. Org. Chem. 1999, 64, 8406.
asymmetric induction took place in the Amadori reaction of
6·HCl, which give 29 as a 3:2 epimeric mixture, instead of a 1:1
mixture, as in the case of the model compound 25.
[16] For ZnCl2-promoted hydroxylactonization, see: a) R. M. Owen,
W. R. Roush, Org. Lett. 2005, 7, 3941; b) K. Nacro, M. Baltas, J.-
M. Escudier, L. Gorrichon, Tetrahedron 1997, 53, 659.
[17] H. Ok Kim, H. W. Baek, H. R. Moon, D.-K. Kim, M. W. Chun,
L. S. Jeong, Org. Biomol. Chem. 2004, 2, 1164.
[18] At this stage, a small amount of the C4 epimer of 22 that
originated from the moderate stereoselectivity (5:1) in the
conversion of 19 into 20 (Scheme 3) could be removed com-
pletely by column chromatography.
[19] The addition of anisole was essential for the successful outcome.
For Lewis acid-promoted deprotection in the presence of
anisole, see: T. Tsuji, T. Kataoka, M. Yoshioka, Y. Sendo, Y.
Nishitani, S. Hirai, T. Maeda, W. Nagata, Tetrahedron Lett. 1979,
20, 2793.
[20] a) V. A. Yaylayan, Food Sci. Technol. Res. 2003, 9, 1; b) A. F.
Jalbout, M. A. H. Shipar, J. L. Navarro, Food Chem. 2007, 103,
919.
[23] The 1H and 13C NMR spectra (in CD3OD) of 2·HCl synthesized
by us were identical to those of natural bacilosarcin B, while
those of our synthetic sample of 2 (free amine) were not. We
were informed of the possibility that the natural bacilosarcin B
that was studied by NMR spectroscopy was not the free amine,
but instead an ammonium salt formed from the free amine and
KH2SO4, since the final purification of natural bacilosarcin B
was performed by reverse-phase HPLC using MeCN/0.15%
KH2PO4 (pH 3.5) as eluent. This possibility was supported by
our observation that the 1H NMR spectrum of
a sample
prepared by mixing the free amine 2 with aqueous KH2PO4
and then concentrating the resulting mixture showed very good
agreement with that of natural bacilosarcin B as well as with that
of 2·HCl, but was significantly different from that of the free
amine (see the Supporting Information).
[21] The moderate yield (43%) of the conversion of 25a into 27a
despite the smooth transformation observed by TLC can be
1148
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 1144 –1148