Toward creation of a universal NMR database for the stereochemical assignment of acyclic compounds: The case of two contiguous propionate units

Article abstract of DOI:10.1021/ol9903786

matrix presented Using triol 1 as a representative example of natural products containing two contiguous propionate units, ¹³C and ¹H NMR databases for the stereochemical assignment of acyclic compounds have been created. Chemical sh

Full text of DOI:10.1021/ol9903786

ORGANIC
LETTERS
1
999
Vol. 1, No. 13
177-2180
Toward Creation of a Universal NMR
Database for the Stereochemical
Assignment of Acyclic Compounds:
The Case of Two Contiguous Propionate
Units
2
Yoshihisa Kobayashi, Jinhwa Lee, Kenichi Tezuka, and Yoshito Kishi*
Department of Chemistry and Chemical Biology, HarVard UniVersity,
12 Oxford Street, Cambridge, Massachusetts 02138
kishi@chemistry.harVard.edu
Received November 29, 1999
ABSTRACT
Using triol 1 as a representative example of natural products containing two contiguous propionate units, 13C and H NMR databases for the
stereochemical assignment of acyclic compounds have been created. Chemical shift increments due to the presence of additional functional
groups as well as solvent effects are discussed.
1
1
Work on the stereochemical assignments of palytoxin, AAL
compound in question are (1) inherent to the specific
stereochemical arrangements of (small) substituents on its
carbon backbone and (2) independent from the rest of the
molecule. We have therefore suggested the possibility that
fatty acids and related compounds bearing (small) substit-
uents on their backbones have the capacity to create unique
structural motifs, to carry specific information, and to serve
as functional materials. These experimental facts also
suggest the possibility that, given a universal database(s) for
various classes of functionalities, the stereochemical assign-
2
3
toxins/fumonisins, and maitotoxin in our laboratory has
yielded a vast volume of experimental data on the structural
properties of fatty acids and related compounds. These
studies have shown that the structural properties of a
(
1) Cha, J. K.; Christ, W. J.; Finan, J. M.; Fujioka, H.; Kishi, Y.; Klein,
L. L.; Ko, S. S.; Leder, J.; McWhorter, W. W., Jr.; Pfaff, K.-P.; Yonaga,
M.; Uemura, D.; Hirata, Y. J. Am Chem. Soc. 1982, 104, 7369-7371 and
preceding papers.
3a
(
2) For the stereochemical assignment of AAL toxins and fumonisins
from this laboratory, see: (a) Boyle, C. D.; Harmange, J.-C.; Kishi, Y. J.
Am. Chem. Soc. 1994, 116, 4995-4996. (b) Boyle, C. D.; Kishi, Y.
Tetrahedron Lett. 1995, 36, 5695-5698 and references therein. For the
work from other laboratories, see: (c) Oikawa, H.; Matsuda, I.; Kagawa,
T.; Ichihara, A.; Kohmoto, K. Tetrahedron 1994, 50, 13347-13368. (d)
Hoye, T. R.; Jim e´ nez, J. I.; Shier, W. T. J. Am. Chem. Soc. 1994, 116,
(3) For the work from this laboratory, see: (a) Zheng, W.; DeMattei, J.
A.; Wu, J.-P.; Duan, J. J.-W.; Cook, L. R.; Oinuma, H.; Kishi, Y. J. Am.
Chem. Soc. 1996, 118, 7946-7968. (b) Cook, L. R.; Oinuma, H.; Semones,
M. A.; Kishi, Y. J. Am. Chem. Soc. 1997, 119, 7928-7937. For the work
from the laboratories at Tokyo and Tohoku Universities, see: (c) Sasaki,
M.; Matsumori, N.; Maruyama, T.; Nonomura, T.; Murata, M.; Tachibana,
K.; Yasumoto, T. Angew. Chem., Int. Ed. Engl. 1996, 35, 1672-1675. (d)
Nonomura, T.; Sasaki, M.; Matsumori, N.; Murata, M.; Tachibana, K.;
Yasumoto, T. Angew. Chem., Int. Ed. Engl. 1996, 35, 1675-1678.
9
409-9410. (e) ApSimon, J. W.; Blackwell, B. A.; Edwards, O. E.;
Fruchier, A. Tetrahedron Lett. 1994, 35, 7703-7706. (f) Poch, G. K.;
Powell, R. G.; Plattner, R. D.; Weisleder, D. Tetrahedron Lett. 1994, 35,
7
707-7710. (g) Blackwell, B. A.; Edwards, O. E.; ApSimon, J. W.;
Fruchier, A. Tetrahedron Lett. 1995, 36, 1973-1976.
1
0.1021/ol9903786 CCC: $18.00 © 1999 American Chemical Society
Published on Web 12/23/1999

Scheme 1
ment of an unknown compound may be achieved, in contrast
to our previous cases,1-3 without synthetic efforts. In this
Letter, we report our first step toward the creation of a
universal NMR database through analysis of two contiguous
propionate units as an example.
To test the feasibility and reliability of this approach, we
selected a partial structure typically found in the polypro-
pionate class of natural products. Specifically, we chose triol
Figure 1. Difference in carbon chemical shifts between the average
and the values of 1a-h (100 MHz, CD OD). The x and y axes
3
represent carbon number and ∆δ (δ1a-h - δave) in ppm, respectively,
for all the graphs in this paper.
experimental values with predicted values should form a
basis to estimate the chemical shift increments due to
additional functional groups present in an unknown com-
pound.
1
as a representative example of natural products containing
two contiguous propionate units. Our strategy was then to
synthesize the eight possible diastereomers 1a-h to create
1
13
universal H and C NMR databases for this structural motif.
Several comments on this plan are in order. First, any
compound accommodating two contiguous propionate units
could be used for the present data collection, but we prefer
that the system meet the following criteria: (1) the two side
chains should be different, thereby enabling the 13C and H
chemical shifts to be precisely measured and (2) the two
side chains should not bear a substituent(s) which may
influence the propionate units via unusual electronic and/or
steric effects. Second, in the previous cases,1 we compared
the chemical shifts of synthetic diastereomers with the
chemical shifts of the natural product and used a degree of
chemical shift deviation as the indicator for match/mismatch
judgments. For the current purpose, we will use a deviation
of chemical shift from the average value of the eight synthetic
diastereomers as the reference point. Third, to correlate the
NMR data of an unknown compound with a universal
database, we need to estimate the effects on chemical shifts
due to an additional functional group or groups present in
an unknown compound. The left side chain of 1 should allow
us to install representative functional groups on the backbone
and to determine their effects on the database. Degrees of
such effects could also be estimated via use of an empirical
formula for predicting 13C chemical shifts. Comparison of
1
-3
Among numerous synthetic methods available for the
iterative construction of polypropionates, we elected to adopt
the Brown crotylboration protocol for the preparation of the
eight diasteromers possible for 1. In the first chain elongation
(Scheme 1), both the expected syn and anti products were
obtained in approximately 93% and 75% enantiomeric
4
(
4) (a) Brown, H. C.; Bhat, K. S. J. Am. Chem. Soc. 1986, 108, 5919-
5
923. (b) Brown, H. C.; Bhat, K. S.; Randad, R. S. J. Org. Chem. 1989,
5
4, 1570-1576.
(5) The enantiomeric excess was estimated from the 19F NMR of its (R)-
Mosher ester. It is worthwhile adding that 4 diastereomers 4b, 4d, 4e, and
h obtained stereospecifically in the second crotylboration were found to
be optically pure.
4
2178
Org. Lett., Vol. 1, No. 13, 1999

Figure 2. Difference in proton chemical shifts between the average
and the values of 1a-h (500 MHz, CD OD).
3
Figure 3. Difference in carbon chemical shifts of each of 1a-h,
∆
δ ) δ((CD ) SO) - δ(CD OD) in ppm (100 MHz).
3
2
3
5
excess, respectively. In the second chain elongation, it
chemistry at C.5 and C.6 as well as C.7 and C.8. Second,
each diastereomer was converted to the corresponding
C.5,C.7-acetonide, the NMR analyses of which allowed us
should be noted that (1) four diastereomers, 4b, 4d, 4e, and
h, out of eight were obtained stereospecifically, (2) three
5
9
4
diastereomers, 4a (stereoselectivity ) 2:1), 4f (2:1), and 4g
1:1), were formed as a mixture of diastereomers, and (3)
to conclude the relative stereochemistry at C.5 and C.7.
Third, the eight diastereomers were prepared in an optically
active form, the absolute configuration of which was
(
diastereomer 4c could not be obtained by this method.
Diastereomer 4c was instead prepared via zinc-mediated
determined by the chirality of (Ipc)
2
BOMe used in the first
6
4
crotylation (stereoselectivity ) 5:1), which was found also
crotylboration.
1
3
1
to be a stereoselective (5:1) means for the preparation of
The C and H NMR spectra of each diastereomer were
recorded in three commonly used NMR solvents [CD OD,
(CD SO, and CDCl ], and the chemical shift assignments
were established through COSY, HMQC, and DEPT experi-
diastereomer 4g. In addition, application of the tin-mediated
3
7
crotylation developed by Keck improved the stereoselec-
3
)
2
3
8
tivity in the formation of 4a (5:1) and 4f (5:1). Finally,
1
3
1
hydrogenation/hydrogenolysis [H
a-h furnished 1a-h, respectively.
The stereochemistry of 1a-h was assigned on the basis
2
/Pd(OH)
2
on C/MeOH] of
ments. The C and H databases were created and shown
as a deviation in chemical shift for each nucleus of a given
diastereomer from the average chemical shift of the nucleus
in question (Figures 1 and 2).
4
of the following facts. First, it is well established that (E)-
4
and (Z)-crotylboronates yield anti and syn products, respec-
As anticipated from our previous work, each diastereomer
exhibited distinct and differing NMR spectroscopic properties
tively, thereby allowing assignment of the relative stereo-
from each other. It should be noted that the diastereomers
(
(
(
6) For examples similar to the present case, see ref 3a.
7) Keck, G. E.; Abbott, D. E. Tetrahedron Lett. 1984, 25, 1883-1886.
8) For the case of 4a, although the stereoselectivity (4a:4b) could be
13
1f and 1h exhibit relatively small differences in the C NMR
database in CD
H NMR database. Interestingly, this trend, i.e., the comple-
mentary nature of the H and C NMR databases, was also
noticed for several cases in our maitotoxin studies.3a
Upon comparison of the chemical shifts found for each
3
OD but exhibit significant differences in the
1
improved up to 5:1, a significant amount of the linear product (3:2
regioselectivity) was also formed.
1
13
9) The 3C chemical shifts of two acetonide methyl groups are known
1
(
to be diagnostic in differentiating syn- and anti-1,3-diol acetonides: (a)
Rychnovsky, S. D.; Skalitzky, D. J. Tetrahedron Lett. 1990, 31, 945-948.
b) Evans, D. A.; Reiger, D. L.; Gage, J. R. Tetrahedron Lett. 1990, 31,
(
7
nucleus in CD
solvent effects was evident. However, it is important to note
that, upon changing the solvent from CD OD to (CD SO,
3 3 2 3
OD, (CD ) SO, and CDCl , the presence of
099-7100. In addition, NOE studies were conducted on the acetonides to
confirm the stereochemical assignment at C.5 and C.7.
10) CS ChemNMR Pro version 1.0, Renate Buergin Schaller, Develop-
(
3
3 2
)
ment Centre, Bergstrasse 114, Zurich, Switzerland, installed in CS Chem-
Draw Pro version 4.5, was used.
each nucleus of the eight diastereomers 1a-h was found to
Org. Lett., Vol. 1, No. 13, 1999
2179

increment, respectively. The chemical shift increments thus
estimated were found to compare well with the experimen-
tally obtained values both in the 1d and A1-5 or B1 and
B2-3 series (Table 1). This exercise demonstrates that
a
10
Table 1. Experimental (Upper Line) and Predicted (Lower
Line) Increments (ppm) of 13C NMR Chemical Shifts
Figure 4. Difference in carbon chemical shifts of each of 1a-h,
∆
δ ) δ(CDCl3) - δ(CD OD) in ppm (100 MHz).
3
exhibit approximately same magnitude of solvent effects
Figure 3). Consequently, the overall profile of the NMR
databases in CD OD and (CD SO became virtually identi-
cal. On the other hand, upon changing the solvent from CD
OD to CDCl , each nucleus of 1a-h was found to exhibit a
(
3
3 2
)
3
-
3
different magnitude of the solvent effects among the eight
diastereomers. However, a similarity is noticeable in the
overall profile of solvent effects between 1a/1c, 1b/1d, 1e/
1
3
necessary adjustment(s) to the C NMR database due to the
presence of a new array of functional groups can be made
using the established empirical formula.
1g, and 1f/1h (Figure 4). Interestingly, the diastereomers in
The reliability and usefulness of our NMR database for
the stereochemical assignment of acyclic compounds are
discussed in the following paper using the oasomycin class
of natural products as an example.
these pairings share the same relative configuration at the
C.5, C.6, and C.7 positions. These observations may suggest
that a six-membered intramolecular hydrogen bonding array
plays a major role in determining the overall structural
properties of these diastereomers in CDCl
not play a role in CD OD and (CD SO.
3
, whereas it does
Acknowledgment. We gratefully acknowledge Dr. Laura
Cook-Blumberg for her contributions at the early phase of
this program. Financial support from the National Institutes
of Health (NS 12108) is gratefully acknowledged.
3
3 2
)
To determine chemical shift increments on the 13C NMR
database due to the presence of additional functional groups,
two series of derivatives of 1d were prepared, i.e., A1-5
and B1-2. Although chemical shift increments were obtained
Supporting Information Available: Experimental pro-
cedures summarized in Scheme 1, table listing the chemical
1
13
also for the H NMR spectra, we focused only on the
C
shifts of acetonides derived from 1a-h, database graphs in
NMR spectra because such incremental values could be
estimated via the empirical formula known in the literature.
1
(
CD
3
)
2
SO and CDCl
3
, and tables listing complete H and
1
3
C NMR assignments of 1a-h in CD
3
OD, (CD
3 2
) SO, and
1
0
Using the program developed by Renate Buergin Schaller,
the chemical shifts were estimated for each carbon of A1-
, B1-3, and 1d. The chemical shift difference between a
13
CDCl
3
, and C NMR assignments of A1-5 and B1-3. This
material is available free of charge via the Internet at
http://pubs.acs.org.
5
given carbon in 1d or B1 and the corresponding carbon in
A1-5 or B2-3 was assumed to represent the chemical shift
OL9903786
2180
Org. Lett., Vol. 1, No. 13, 1999