RDC-enhanced NMR spectroscopy in structure elucidation of sucro-neolambertellin

Article abstract of DOI:10.1002/anie.200705037

(Graph Presented) A new method based on NMR spectroscopy is introduced to determine the absolute configuration of natural compounds and has been successfully applied to sucro-neolambertellin. The method relies on the measurement of residual dipolar coupli

Full text of DOI:10.1002/anie.200705037

Communications
DOI: 10.1002/anie.200705037
Glycoconjugates
RDC-Enhanced NMR Spectroscopy in Structure Elucidation of Sucro-
Neolambertellin**
Anne Schuetz, Takanori Murakami, Noboru Takada, Jochen Junker, Masaru Hashimoto,* and
Christian Griesinger*
In the course of mechanistic investigations of mycoparasitism
by Lambertella species the novel glycoside sucro-neolamber-
tellin (1) was found. In this work, the configuration of 1
(Figure 1) was determined using NMR spectroscopy. Whereas
dipolar couplings (RDCs) together with a structure calcula-
tion protocol that uses floating chirality to unequivocally
establish the configuration of 1. RDCs are inherently differ-
ent restraints from J couplings and NOEs, and therefore
restrict conformation and configuration more than NOEs and
J couplings alone. While this difference could be shown
recently for examples on cyclic systems with naturally
restricted conformations, the expansion to open-chain sys-
tems underpins the general applicability of the approach.[1]
With the established configuration in the central furanose
ring, sucro-neolambertellin is the first natural product iden-
tified to have a sucrose moiety that is oxidized at the 5’-carbon
atom of the fructose group.
Sucro-neolambertellin was discovered within culture
broth of Lambertella sp. 1346[2] during studies of the relation-
ship between the production of lambertellols,[3] candidate
substances of mycoparasitism on apple fruit, and the culture
conditions.[4]
Figure 1. Sucro-neolambertellin (1) with correct configuration of the
central five-membered ring(C22 to C25 as indicated in red: RSSR).
The dihedral angles around the glycosidic linkage are defined as f:
H31-C31-O30-C22 and y: C31-O30-C22-O27. The existence of the
chromophore neolambertellin (3) in the molecule was deduced from
the UV spectrum.
Negative fast atom bombardment mass spectroscopy
(negative-FABMS) of 1 provided a pseudomolecular ion
signal [MÀH]À at m/z 581.1525, thus implying that the
molecular formula is C26H29O16. The photo diode array
the conventional NMR spectroscopic parameters, namely J
couplings and NOEs, did not suffice to determine the
configuration, we report herein the additional use of residual
1
(PDA) spectrum of 1 as well as H and 13C NMR spectra
(Supporting Information) identify neolambertellin (3[5]
,
Figure 1) as the aglycon.
The 1H NMR spectra reveal a unique hexofuranose
moiety carrying an a-d-glucopyranoside unit at C22 (the
numbering is shown in Figure 1). The a-d-glucopyranoside
unit was identified from the characteristic J coupling pattern
of the protons from C31 to C36. The absolute configuration
was established by acidic hydrolysis, perbenzoylation, and
circular dichroism (CD) spectroscopy. The constitution of the
linkage between the aglycon and the hexofuranose moiety
[*] T. Murakami, N. Takada, M. Hashimoto
Faculty of Agriculture and Life Sciences
Hirosaki University
3-Bunkyo-Cho, Hirosaki, 036-8561 (Japan)
Fax: (+81)172-39-3782
E-mail: hmasaru@cc.hirosaki-u.ac.jp
A. Schuetz, J. Junker,[+] Prof. C. Griesinger
Abteilungfür NMR-basierte Strukturbiologie
Max-Planck-Institut für Biophysikalische Chemie
Am Fassberg11, 37077 Göttingen (Germany)
Fax: (+49)551-201-2202
À
was established in the following way: The C7 OH group was
assigned by observing HMBC correlations between this
exchangeable proton in [D6]DMSO at d = 10.28 ppm and
E-mail: cigr@nmr.mpibpc.mpg.de
13
À
C resonances for C6a and C7. As there is no C6 OH
[+] Present address: Fundç¼o Oswaldo Cruz—CDTS
Rua Sizenando Nabuco 100, CEP 21040-250 Rio de Janeiro, RJ
(Brazil)
1
resonance in the H NMR spectra, the hexofuranose moiety
must be connected through the C6 oxygen atom to the
neolambertellin aglycon.
[**] This work was supported by the Max Planck Society, the DFG (GRK
782), and the Fonds der Chemischen Industrie. Part of this work was
also supported by a grant-in-aid for scientific research (no.
17580090) from the Ministry of Education, Culture, Sports, Science
and Technology of Japan. We also thank the Suntory Institute for
Bioorganic Research for the SUNBOR Grant. We thank Prof. Jun
Kawabata and Dr. Eri Fukushi of Hokkaido University for mass
spectrometry measurements, and Dr. Edward d’Auvergne for
insightful discussions and careful reading.
More information concerning the configuration of the
five-membered ring could not be derived owing to signal
overlap. Therefore, 1 was converted into its nona-acetate 2,
which considerably reduced the signal overlap in the proton
dimension. All NMR spectroscopic analyses and structure
calculations were then performed on the peracetate 2
(Figure 1). At this stage the absolute configuration of the
glucose unit and the connectivity of the three moieties
neolambertellin, hexofuranose, and a-d-glucose are clear,
Supportinginformation for this article is available on the WWW
under http://www.angewandte.org or from the author.
2032
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 2032 –2034

Angewandte
Chemie
whereas the configuration of the stereocenters of the
hexofuranose unit need to be established.
The trans configuration of the ring protons H23 and H24
(3JH,H = 9.7 Hz) in combination with the similar chemical
shifts of the carbon atoms and protons, which are related by a
À
pseudo C2-symmetry axis through the middle of the C23 C24
bond and O27, is compatible only with four configurations
RRRR, RSSR, SRRS, and SSSS (where for example, RRRR
means C22/R,C23/R,C24/R,C25/R).[6] Thus, the challenge was
to find the correct configuration for these stereocenters that
are remote from the known stereocenters of the a-d-glucose
unit. Therefore, long-range structural restraints relating
conformation and configuration of the hexofuranose moiety
to the known a-d-glucopyranoside unit were required. Such
restraints are provided by quantitative NOEs as well as by
RDCs. The latter reflect orientations of bond vectors in
molecules relative to a global-alignment tensor.[7] They have
only recently become accessible for natural compounds
through the availability of alignment media that are compat-
ible with organic solvents such as DMSO.[8]
Figure 2. Energy penalties associated with NOE and RDC violations.
The configurations that are in agreement with the H23 and H24 3JH,H
couplingand the symmetry of chemical shifts are given in bold letters.
The x denotes that for this configuration, the RDC energy minimization
failed. The optimal configuration is highlighted within a box.
Quantitative NOE build-up curves for eight different
mixing times varying from 55 to 1000 ms were derived from a
NOESY experiment that suppresses zero-quantum artifacts
almost completely.[9] The build-up curves for protons with
fixed distances were compatible with correlation times that
varied by less than 20%. This would be either compatible with
more or less free motion of all three moieties with respect to
each other or a rather rigid molecule. Since there are 18
specific NOEs between the 3 moieties, it was concluded that
the molecule shows little internal dynamics. A total of 49
NOEs were integrated. Owing to signal overlap and strong
coupling, 22 were treated as unambiguous and the rest as
ambiguous restraints.
that is also used for protein structure calculations. The derived
RDC energy penalty for each configuration is given as open
bars in Figure 2.
The three configurations SRRR, SRSR, and RSSR have
the lowest RDC energies of less than 10 kcalmolÀ1. Only the
RSSR configuration is among those possible as determined
from the chemical-shift symmetry and the trans configuration
of H23 and H24. It should be mentioned that the RRRR and
SRRS configurations are incompatible with the RDCs,
whereas they could not be excluded by the NOE analysis.
The RSSR configuration suggests that the hexofuranose ring
is derived from d-fructose.
The 10 lowest-energy structures found for the RSSR
configuration have an average pairwise root mean square
deviation (rmsd)[17] of only 0.106 Š (Figure 3), which is not
surprising because of the similar correlation times and the
abundant NOEs between the three moieties. For the con-
Although only the four configurations for the hexofur-
anose ring mentioned above need to be distinguished,
structure calculations with floating chirality[10–13] for all
centers in the furanose ring were performed by using X-
PLOR NIH[14] with the above-mentioned ambiguous and
unambiguous NOEs (see also the Supporting Information).
Although floating chirality protocols are frequently used for
proteins, so far they have rarely been applied to small
molecules. Whereas the stereocenters in the hexofuranose
ring were left floating, those of the a-d-glucopyranoside
moiety were fixed. For each of the 16 possible configurations
the energy penalty associated with NOE violations (averaged
over the 10 lowest-energy structures) is shown as filled bars in
Figure 2. The NOE violations are not significant enough to
single out one configuration. Therefore, the structures for all
16 configurations were subjected to cross-validation filtering
by RDCs.
À
formation of the C22 C31 glycosidic linkage we find f=
À70.18 and y = À63.78, which matches extremely well with
the values f= À558and y = À588 obtained for sucrose in
DMSO.[18] The H31 and C22 3JC,H coupling constant measured
by HMBC[19] suggests f= Æ 558 or Æ1178 using a Karplus
calibration.[20]
In conclusion, the conformation and the absolute config-
uration of sucro-neolambertellin could be determined from
NMR spectroscopy, chemical degradation, and CD spectros-
copy. From the chemical shifts and the trans configuration of
the H23 and H24 protons, only four configurations are
feasible: RRRR, RSSR, SRRS, and SSSS. The quantitative
NOE analysis does not distinguish between these. The cross-
validation filtering against RDCs (Figure 4) clearly, and with
a large margin, isolates RSSR among the other possibilities.
This result is also biosynthetically sensible because this
moiety could then be derived from the natural d-fructose.
We believe that with the described methodology, the config-
uration of a wide range of natural compounds can be
determined and the requirement for stereoselective synthesis
to prove configuration can be reduced. In addition, the cross-
validation filtering of conformations and configurations by
RDCs D(CiHi)exp, which are the dipolar couplings between
carbon and proton i, were obtained by aligning 2 in three
different polymer-based media (listed in the Supporting
1
Information). RDCs were extracted from t2-coupled H and
13C HSQC spectra by superimposing and fitting w2 traces from
isotropic and anisotropic spectra. The RDCs were used to
cross validate the above-mentioned 10 structures for the 16
different configurations using the SANI module[15,16] from X-
PLOR with the standard force constant of 50 kcalmolÀ1 HzÀ2
Angew. Chem. Int. Ed. 2008, 47, 2032 –2034
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2033

Communications
[1] C. Aroulanda, V. Bouchard, F. Guibe, J. Courtieu, D. Merlet,
Chem. Eur. J. 2003, 9, 4536 – 4539; D. L. Yan, F. Delaglio, A.
Kaerner, A. D. Kline, H. P. Ho, M. J. Ahapiro, T. A. Smitka,
G. A. Stephenson, E. R. Zartler, J. Am. Chem. Soc. 2004, 126,
5008 – 5017; D. L. Yan, E. R. Zartler, Magn. Reson. Chem. 2005,
43, 53 – 64; C. M. Thiele, A. Marx, R. Berger, J. Fischer, M. Biel,
A. Giannis, Angew. Chem. 2006, 118, 4566 – 4571; Angew. Chem.
Int. Ed. 2006, 45, 4455 – 4460.
[2] Y. Harada, M. Sasaki, Rep. Tottori Mycol. Inst. 1990, 28, 275 –
285.
[3] T. Murakami, Y. Morikawa, M. Hashimoto, T. Okuno, Y.
Harada, Org. Lett. 2004, 6, 157 – 160.
[4] Cultivation was carried out at 208C whereas lambertellols A, B,
C, and neolambertellin were obtained by culturing at 258C.
Interestingly, this condition does not provide 1.
[5] T. Murakami, M. Hashimoto, T. Okuno, Bioorg. Med. Chem.
Lett. 2005, 15, 4185 – 4188.
[6] L. Butera, S. Englard, J. S. Blanchard, Carbohydr. Res. 1986, 148,
179 – 188.
[7] N. Tjandra, A. Bax, Science 1997, 278, 1111 – 1114.
[8] a) J. C. Freudenberger, S. Knꢀr, K. Kobzar, D. Heckmann, T.
Paulutat, H. Kessler, B. Luy, Angew. Chem. 2005, 117, 427 – 430;
Angew. Chem. Int. Ed. 2005, 44, 423 – 426; b) P. Haberz, J.
Farjon, C. Griesinger Angew. Chem. 2005, 117, 431–433; Angew.
Chem. Int. Ed. 2005, 44, 427 – 429; Angew. Chem. Int. Ed. 2005,
44, 427 – 429; c) G. Kummerlowe, J. Auernheimer, A. Lendlein,
B. Luy, J. Am. Chem. Soc. 2007, 129, 6080 – 6081.
Figure 3. The 10 conformers of 2 (RSSR configuration) with lowest
energies found in floating chirality calculations (acetate moieties are
not shown). The structures have been superimposed on the carbon
atoms of the furanose ring. The average pairwise rmsd[17] is 0.106 ꢀ.
[9] M. J. Thrippleton, J. Keeler, Angew. Chem. 2003, 115, 4068 –
4071; Angew. Chem. Int. Ed. 2003, 42, 3938 – 3941.
[10] A. Pardi, D. R. Hare, M. E. Selsted, D. Mierke, R. D. Morrison,
D. A. Bassolino, A. C. Bach, J. Mol. Biol. 1988, 201, 625 – 636.
[11] T. A. Holak, M. Nilges, H. Oschkinat, FEBS Lett. 1989, 242,
218 – 224.
[12] R. H. A. Folmer, C. W. Hilbers, R. N. H. Konings, M. Nilges, J.
Biomol. NMR 1997, 9, 245 – 258.
[13] M. Reggelin, M. Kꢀck, K. Conde-Frieboes, D. Mierke, Angew.
Chem. 1994, 106, 822 – 824; Angew. Chem. Int. Ed. Engl. 1994,
33, 753 – 755.
[14] a) C. D. Schwieters, J. J. Kuszewski, N. Tjandra, G. M. Clore, J.
Magn. Reson. 2003, 160, 65 – 74; b) C. D. Schwieters, J. J.
Kuszewski, G. M. Clore, Prog. Nucl. Magn. Reson. Spectrosc.
2006, 48, 47 – 62.
Figure 4. Schematic representation of the procedure used for RDC-
enhanced structure elucidation of sucro-neolambertellin (1).
RDCs is not limited to rather rigid molecules but can also be
applied to more flexible molecules.[21] Also, in light of
reference [22], RDCs may also prove to be a sensitive
NMR spectroscopic parameter to detect differences in
configurations of synthetic and natural products with very
similar chemical shifts.
[15] a) G. M. Clore, A. M. Gronenborn, N. Tjandra, J. Magn. Reson.
1998, 131, 159 – 162; b) G. M. Clore, A. M. Gronenborn, A. Bax,
J. Magn. Reson. 1998, 133, 216 – 221.
[16] J. Klages, C. Neubauer, M. Coles, H. Kessler, B. Luy, Chem-
BioChem 2005, 6, 1672 – 1678.
[17] The rmsd was calculated after superimposing the structures on
C22, C23, C24, and C25. It was calculated over all atoms except
those in the acetate groups (starting from the carbonyl atom)
and the methyl group for lack of experimental restraints.
[18] S. Bagley, M. Odelius, A. Laaksonen, G. Widmalm, Acta Chem.
Scand. 1994, 48, 792 – 799.
Experimental Section
Experimental procedures regarding the isolation of 1, its derivatiza-
tion, CD and NMR spectroscopic measurements, preparation of
alignment media, and details of NOE analysis as well as structure
calculation are given in the Supporting Information.
[19] L. Verdier, P. Sakhaii, M. Zweckstetter, C. Griesinger, J. Magn.
Reson. 2003, 163, 353 – 359.
[20] I. Tvaroska, M. Hricovini, E. Petrakova, Carbohydr. Res. 1989,
189, 359 – 362.
[21] A. Schuetz, J. Junker, A. Leonov, O. F. Lange, T. F. Molinski, C.
Griesinger, J. Am. Chem. Soc. 2007, 129, 15114 – 15115.
[22] G. Pattenden, N. J. Ashweek, A. G. Baker-Glenn, G. M. Walker,
J. G. K. Yee, Angew. Chem. 2007, 119, 4437 – 4441; Angew.
Chem. Int. Ed. 2007, 46, 4359 – 4363.
Received: October 31, 2007
Published online: January 31, 2008
Keywords: glycoconjugates · natural products ·
.
NMR spectroscopy · residual dipolar couplings ·
structure elucidation
2034 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 2032 –2034