Toward creation of a universal NMR database for stereochemical assignment: The case of 1,3,5-trisubstituted acyclic systems

Article abstract of DOI:10.1002/1522-2675(20000906)83:9<2562::AID-HLCA2562>3.0.CO;2-Z

Using the diastereoisomeric triols 1a-d (Fig. 1) and examples summarized in Fig. 2, the central C-atom of acyclic 1,3,5-triols is demonstrated to exhibit a distinctive chemical shift that is dependent on the 1,3- and 3,5-relative configuration, but is independent of the functionalities present outside of this structural motif. These NMR characteristics are then used to predict the relative configuration of several natural products (Fig, 7). In addition, an example is given to show the possibility of assembling an NMR database for a larger array of functional groups from NMR databases of smaller arrays of functional groups.

Full text of DOI:10.1002/1522-2675(20000906)83:9<2562::AID-HLCA2562>3.0.CO;2-Z

2562
Helvetica Chimica Acta ± Vol. 83 (2000)
Toward Creation of a Universal NMR Database for Stereochemical
Assignment: The Case of 1,3,5-Trisubstituted Acyclic Systems
by Yoshihisa Kobayashi, Choon-Hong Tan, and Yoshito Kishi*
Department of Chemistry and Chemical Biology, Harvard University, 12 Oxford Street, Cambridge,
Massachusetts 02138, USA
Dedicated to Professor Albert Eschenmoser on the occasion of his 75th birthday
Using the diastereoisomeric triols 1a ± d (Fig. 1) and examples summarized in Fig. 2, the central C-atom of
acyclic 1,3,5-triols is demonstrated to exhibit a distinctive chemical shift that is dependent on the 1,3- and 3,5-
relative configuration, but is independent of the functionalities present outside of this structural motif. These
NMR characteristics are then used to predict the relative configuration of several natural products (Fig. 7). In
addition, an example is given to show the possibility of assembling an NMR database for a larger array of
functional groups from NMR databases of smaller arrays of functional groups.
Through our work on palytoxin [1], AAL toxins/fumonisins [2], maitotoxin [3], and
universal NMR databases [4], we have experimentally demonstrated that steric and/or
stereoelectronic interactions between the structural motifs connected either directly or
with a one-CH2 bridge are significant, whereas interactions between the structural
motifs connected with a two- or more-CH2 bridge are almost negligible. On this basis,
we have noticed that 1,3,5-trisubstituted acyclic compounds, represented by A, may
exhibit a unique characteristic: the central C-atom may exhibit a distinctive chemical
shift which is dependent on the X1/X2- and X2/X3-relative configuration, but is
independent of the functionalities present outside of this structural motif1). One might
be concerned that, through the substituents X1 and X3, functional groups present
outside of the 1,3,5-trisubstituted moiety might indirectly affect the chemical shifts of
these nuclei. However, upon a close examination of the cases previously studied in our
laboratories, we speculate that such indirect steric and/or stereoelectronic effects on the
chemical shifts in question is negligibly small [2 ± 4].
To test this hypothesis, we have chosen a 1,3,5-triol system (X1 ˆ X2 ˆ
X3 ˆ OH in A) as an example, synthesized the four possible diastereoisomers
1
)
This analysis may be applicable for the chemical shift of the H-atom attached to the central C-atom.

Helvetica Chimica Acta ± Vol. 83 (2000)
2563
1a ± d, and measured the 13C chemical shifts in (D6)DMSO and (D4)methanol
(Fig. 1)2)3).
Fig. 1. 13C-NMR Chemical shifts (ppm; in (D6)DMSO and CD3OD) of C(5) of 1a ± d
A large number of natural products and synthetic compounds containing the partial
structure of ÀCH(OH)CH2CH(OH)CH2CH(OH)À in an acyclic system are recorded
in the literature. However, few of these are stereochemically well-defined, with the
assigned chemical shifts being reported in (D6)DMSO or CD3OD3). Fig. 2 summarizes
the distribution of the chemical shifts for the central C-atom of each stereochemical
2
)
The triols 1a ± d were synthesized in 13 steps, i.e., 1. Me(CH2)2CHO ‡ CH2ˆCHCH2MgBr, THF, 08, 2.
MPM-Cl, NaH, Bu4NI, DMF, 3. OsO4, N-methylmorpholine N-oxide (NMO), THF/H2O, followed by
Pb(OAc)4, benzene, 4. CH2ˆCHCH2MgBr, THF, 08, 5. Ac2O, Py, 4-(dimethylamino)pyridine (DMAP), 6.
2,2-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ), CH2Cl2/H2O, followed by chromatographic separa-
tion. 7. KOH, MeOH, 8. BnBr, NaH, Bu4NI, DMF, 9. OsO4, NMO, THF/H2O, followed by Pb(OAc)4,
benzene, 10. CH2ˆCHCH2MgBr, THF, 08, 11. OsO4, NMO, THF/H2O, followed by Pb(OAc)4, benzene, 12.
NaBH4, MeOH, 13. H2, PdÀC, MeOH, followed by chromatographic separation. The relative
configuration of the products obtained at the step 4 was established from the 13C-NMR spectra of the
acetonide of 1,3-diols, whereas that at the step 10 was deduced from the 13C-NMR spectra of their
hydrogenation/hydrogenolysis products, cf. the symmetry/asymmetry present in IIIa ± c. 13C-NMR
(CD3OD, 100 MHz): 1a: 60.0 (C(1)); 40.9 (C(2)); 68.8 (C(3)); 45.4 (C(4)); 70.4 (C(5)); 45.1 (C(6)); 71.0
(C(7)); 41.0 (C(8)); 19.7 (C(9)); 14.5 (C(10)); 1b: 60.2 (C(1)); 41.5 (C(2)); 66.7 (C(3)); 45.9 (C(4)); 68.4
(C(5)); 45.7 (C(6)); 71.1 (C(7)); 40.9 (C(8)); 19.7 (C(9)); 14.5 (C(10)); 1c: 60.0 (C(1)); 40.9 (C(2)); 68.8
(C(3)); 46.1 (C(4)); 68.2 (C(5)); 45.7 (C(6)); 68.7 (C(7)); 41.5 (C(8)); 19.9 (C(9)); 14.5 (C(10)); 1d: 60.2
(C(1)); 41.4 (C(2)); 66.9 (C(3)); 46.5 (C(4)); 66.3 (C(5)); 46.2 (C(6)); 68.9 (C(7)); 41.4 (C(8)); 19.9 (C(9));
14.5 (C(10)). 13C-NMR ((D6)DMSO, 100 MHz): 1a: 58.0 (C(1)); 40.4 (C(2)); 66.2 (C(3)); 44.8 (C(4)); 67.8
(C(5)); 44.5 (C(6)); 68.2 (C(7)); 39.6 (C(8)); 18.3 (C(9)); 14.2 (C(10)); 1b: 58.1 (C(1)); 41.0 (C(2)); 64.2
(C(3)); 45.4 (C(4)); 66.0 (C(5)); 45.1 (C(6)); 68.4 (C(7)); 39.6 (C(8)); 18.3 (C(9)); 14.2 (C(10)); 1c: 58.0
(C(1)); 40.4 (C(2)); 66.5 (C(3)); 45.3 (C(4)); 65.9 (C(5)); 45.1 (C(6)); 66.1 (C(7)); 40.2 (C(8)); 18.5 (C(9));
14.2 (C(10)); 1d: 58.2 (C(1)); 41.0 (C(2)); 64.5 (C(3)); 45.8 (C(4)); 63.9 (C(5)); 45.6 (C(6)); 66.3 (C(7));
40.2 (C(8)); 18.5 (C(9)); 14.2 (C(10)).
3
)
The solvent-effect studies on the two contiguous propionate NMR databases indicated no significant
intramolecular interactions present in (D6)DMSO (the residual solvent signal was set at 39.5 ppm) and
CD3OD (the residual solvent signal was set at 49.0 ppm) but suggested formation of intramolecular H-
bonding networks in CDCl3 [4a].

2564
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 2. Chemical-shift distribution of the central C-atom of acyclic 1,3,5-triol systems. The panels a and b show the
~
distribution of chemical shifts in CD3OD and (D6)DMSO, respectively.
represents the data taken from
aflastatin A, with indication of the C-atom number. Likewise, & from the degradation product I of aflastatin A,
&
*
*
from IIa,b, from IIIa ± c, from 1a ± d.
array in question4). Three sub-groups are easily recognized in both solvents, with all
the anti/anti cases clustering in the range of 66.3 Æ 0.5 ppm (CD3OD) and 63.9 Æ
0.5 ppm ((D6)DMSO), all the anti/syn and syn/anti in the range of 68.6 Æ 0.5 ppm
(CD3OD) and 66.2 Æ 0.5 ppm ((D6)DMSO), and all the syn/syn in the range of 70.7 Æ
0.5 ppm (CD3OD) and 68.2 Æ 0.5 ppm ((D6)DMSO). This exercise demonstrates that
the chemical shift of the central C-atom depends primarily on the C(1)/C(3)- and C(3)/
C(5)-relative configuration but is independent of the rest of the functionalities present
in the molecule.
4
)
A SciFinder sub-structure search (July, 2000) gave 420 hits containing this partial structure. Among them,
aflastatin A and one of its degradation products, I, met with these conditions [5]. However, related to the
work of universal NMR databases, we have synthesized and characterized IIa,b and IIIa ± c, which did add
valuable data points to construct this graph.

Helvetica Chimica Acta ± Vol. 83 (2000)
2565
We have noted the possibility that an NMR database for a larger array of functional
groups could be assembled from NMR databases of smaller arrays of functional groups
[4b]. In connection with this objective, the NMR database of a 1,3-diol such as 2a,b5) is
intriguing (Fig. 3)6). The 13C chemical shift indicated for the syn-diol 2a is downfield by
ca. 2 ppm than that for the anti-diol 2b (for a related observation, see [6]). Interestingly,
this phenomenon occurs in each syn/anti 1,3-diol pair studied, including every syn/anti
pair of each 1,3-diol system present in the structures shown in Figs. 1 and 4. Upon compa-
rison of the two databases summarized in Figs. 1 and 3, one can recognize a series of
additivity increments to predict the chemical shift for the central carbon of 1,3,5-triols
based on the 1,3-diol database; the chemical shift is found ca. 4 ppm downfield for 1a
(syn ‡ syn ˆ 2 ‡ 2), and ca. 2 ppm downfield for 1b (anti ‡ syn ˆ 0 ‡ 2) and 1c (syn ‡
anti ˆ 2 ‡ 0), in reference to the corresponding chemical shift for 1d (anti/anti)7).
Fig. 3. 13C-NMR Chemical shifts (ppm; in (D6)DMSO and CD3OD) of C(5) of 1,3-Diols 2a and 2b
During the course of this study, we recognized an interesting trend with regard to
the chemical shift of the C-atom indicated by an arrow in 3a,b; the effect of a primary
alcohol on the chemical shift is approximately equal in magnitude (ca. 2 ppm) to
that of an anti secondary alcohol (Fig. 4) (cf. C(5) of 1b and 1d vs. C(3) of 3a
and 3b). This trend is observed consistently in 1a ± d, as well as in the examples
known in the literature8). Similarly, we have examined an effect of an olefinic
bond on the 13C chemical shift of the 1,3-diol system. Among the three olefin-
5
)
The diols 2a,b were synthesized in a similar route to the one given for the synthesis of 1a ± d. 13C-NMR
(CD3OD, 100 MHz): 2a: 62.9 (C(1)); 33.7 (C(2)); 22.9 (C(3)); 38.4 (C(4)); 71.5 (C(5)); 44.9 (C(6)); 71.4
(C(7)); 40.9 (C(8)); 19.7 (C(9)); 14.5 (C(10)); 2b: 62.9 (C(1)); 33.7 (C(2)); 23.1 (C(3)); 39.0 (C(4)); 69.2
(C(5)); 45.6 (C(6)); 69.0 (C(7)); 41.5 (C(8)); 20.0 (C(9)); 14.5 (C(10)). 13C-NMR ((D6)DMSO, 100 MHz):
2a: 60.8 (C(1)); 32.7 (C(2)); 21.6 (C(3)); 37.3 (C(4)); 68.9 (C(5)); 44.3 (C(6)); 68.5 (C(7)); 39.3 (C(8)); 18.2
(C(9)); 14.2 (C(10)); 2b: 60.8 (C(1)); 32.7 (C(2)); 21.9 (C(3)); 37.9 (C(4)); 66.6 (C(5)); 44.9 (C(6)); 66.3
(C(7)); 40.2 (C(8)); 18.5 (C(9)); 14.2 (C(10)).
This NMR database, originally created for the stereochemical assignment of the 1,3-diol moiety present in
the desertomycin/oasomycin class of natural products, should have a wide application. A SciFinder sub-
structure search (July, 2000) gave 1,143 hits containing the partial structure of 1,3-diol, i.e., A with X1 ˆ H
and X2 ˆ X3 ˆ OH. For example, based on the chemical shifts reported for the C(10) (69.0 ppm,
(D6)DMSO), C(22) (68.8), C(49) (66.8), and C(53) (66.6) C-atoms of theonezolide A [7i], this NMR
database predicts its relative configuration at C(8)/C(10), C(22)/C(24), C(47)/C(49), and C(53)/C(55) to
be syn, syn, anti, and anti, respectively.
6
)
7
8
)
To formulate an empirical rule, we clearly need to determine and incorporate a proper coefficient for this
additivity.
A SciFinder sub-structure search (July, 2000) gave 74 hits containing the partial structure of 3a,b. The
application of 1,3,5-triol database 3a,b for the degradation product of shurimycin A, described as 3 in [7c],
revealed the relative configuration of C(21)/C(23) (anti), based on the NMR chemical shift of C(21)
(66.3 ppm in CD3OD; X1 ˆ H).
)

2566
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 4. 13C-NMR Chemical shifts (ppm; in (D6)DMSO and CD3OD) of C(3) of 1a ± d and 3a,b, and of C(4) of
the olefinic 1,3-diols 4a and 4b
substitution patterns often found in polyketide natural products9), we have used
4a,b10) as an example (Fig. 4). The 13C chemical shifts indicated by an arrow again
match well with the values reported in the literature11).
9
)
Two additional olefin-substitution patterns represented by V and VI are often recognized in polyketide
natural products. A SciFinder sub-structure search (July, 2000) gave 484, 96, and 17 hits containing an
olefin(s) of types 4, V and VI, respectively.
10
)
13C-NMR (CD3OD, 100 MHz): 4a: 63.0 (C(1)); 131.1 (C(2)); 134.5 (C(3)); 71.7 (C(4)); 45.3 (C(5)); 70.4
(C(6)); 41.1 (C(7)); 19.7 (C(8)); 14.5 (C(9)); 4b: 63.1 (C(1)); 129.9 (C(2)); 135.4 (C(3)); 69.6 (C(4)); 45.6
(C(5)); 68.7 (C(6)); 41.3 (C(7)); 19.9 (C(8)); 14.5 (C(9)). 13C-NMR ((D6)DMSO, 100 MHz): 4a: 61.0
(C(1)); 129.3 (C(2)); 133.2 (C(3)); 69.0 (C(4)); 44.9 (C(5)); 67.6 (C(6)); 39.7 (C(7)); 18.3 (C(8)); 14.2
(C(9)); 4b: 61.0 (C(1)); 128.4 (C(2)); 134.2 (C(3)); 67.1 (C(4)); 45.1 (C(5)); 66.0 (C(6)); 40.0 (C(7)); 18.4
(C(8)); 14.1 (C(9)).
Based on the chemical shifts of C(37) (72.0 ppm in CD3OD) and C(47) (70.3), the 4a,b-NMR database
predicts the C(35)/C(37) and C(45)/C(47) relative configuration of linearmycin A [7b] to be syn and anti,
respectively. Likewise, the C(35)/C(37) relative configuration of oasomycin B [8] is predicted from the
chemical shift of C(37) (66.8 ppm in (D6)DMSO).
11
)

Helvetica Chimica Acta ± Vol. 83 (2000)
2567
Fig. 5. 13C-NMR Chemical shifts expected for unknown 1,3-polyol systems. Expected values in (D6)DMSO for
the C-atom indicated by an arrow are shown. Those in CD3OD are ca. 70 ppm (syn/syn), 68 (anti/syn and
syn/anti), and 66 (anti/anti).
It is tempting to apply these patterns to predict the configuration of unknown
compounds. Fig. 5 summarizes a possible outcome for the case of an unknown
compound containing a 1,3,5-triol system. Based on the chemical shift of the central C-
atom, an anti/anti, syn/anti or anti/syn, or syn/syn diastereoisomer can be identified.
However, this approach alone can not differentiate an anti/syn diastereoisomer from a
syn/anti12). Knowing the chemical shifts of two central C-atoms, it is possible to deduce
the relative configuration of 6 out of the 8 possible diastereoisomers for a 1,3,5,7-
tetrol system employing this approach (Fig. 5)12). We should add that this simple
approach has successfully been used to assign the relative configuration of the OH
groups at C(23), C(25), C(27), and C(29), as well as C(33), C(35), and C(37) of
desertomycins/oasomycins (Fig. 6) [8]. Similarly, we can suggest the relative config-
uration of 1,3-polyol systems found in the literature; some of these are summarized in
Fig. 7.
The universal NMR databases we have created contain configurational as well as
conformational information, including the dynamic properties of a given system. Thus
far, we have treated the structural portions of the desertomycin/oasomycin class of natural
12
)
For some cases, discrimination of these (cf., anti/syn vs. syn/anti in the triol series, and anti/syn/anti vs. syn/
anti/syn in the tetrol series) should be possible; for example, the cases where one of the two ends of an
unknown compound corresponds to the functional group shown in Fig. 4, or to the functional group of
X1 ˆ H (Fig. 3).

2568
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 6. 13C-NMR Chemical shifts ((D6)DMSO) of C(25), C(27), and C(35) of desertomycins/oasomycins
Fig. 7. Chemical-shift distribution of the central C-atom of 1,3,5-triol systems with unknown configuration. The
*
panels a and b show the distribution of chemical shifts in CD3OD and (D6)DMSO, respectively [7]. a)
~
represents the data taken from quinolidomicin A, with indication of the C-atom number. Likewise,
from
!
linearmycin A, from shurimycin A, & from monazomycin, from malolactomycin A,
from brasilinolide
^
´
&
*
A, from mathemycin A. b) represents the data taken from blasticidin A, with indication of the C-atom
!
number. Likewise,
from theonezoline A, & from kanchanamycin A.
products as if the macrocyclic lactone rings did not significantly affect their NMR
properties. However, we have noticed that the chemical shifts reported for the central
C-atoms of the 1,3,5-triols present in desertomycins/oasomycins differ slightly from

Helvetica Chimica Acta ± Vol. 83 (2000)
2569
those observed for 1a ± d (Fig. 8). In this context, the 1,3-diol systems found in the
polyene macrolide antibiotics are enlightening. Among a large number of polyene
macrolide antibiotics, mycoticin A (known also as flavofungin) [9] is especially
intriguing, because complete assignments of the 1H- and 13C-NMR in (D6)DMSO were
accomplished, and also because an X-ray analysis was performed on the structurally
closely related roxaticin [10]. The chemical shifts reported for the C(17) (69.28 ppm in
(D6)DMSO), C(19) (67.53), C(21) (62.16), C(23) (62.63), and C(25) (66.28) C-atoms
of mycoticin A still follow the general trend, i.e., three sub-groups are recognized with
the order of chemical shifts (downfield-to-upfield) being syn/syn ! anti/syn and syn/
anti ! anti/anti. However, the absolute values of chemical shifts reported for mycoticin
A differ more significantly from those found in 1a ± d than those for desertomycins/
oasomycins (Fig. 8).
Fig. 8. Chemical-shift distribution of the central C-atom of 1,3,5-triol systems. The panels a, b, and c show the
13
&
distribution of C chemical shifts in the acyclic (Fig. 2), desertomycins/oasomycins and mycoticin A ( ), and
^
filipin III ( ) series in (D6)DMSO, respectively, with indication of the C-atom number.
In our view, there are at least two reasons for this deviation in the absolute values of
chemical shifts. First, it is well-documented that the polyene and 1,3-polyol chains
interact with each other transannularly [9b][10], and one can imagine that an
anisotropic effect from the polyene chain on the 1,3-polyol chain might have resulted in
a deviation on the chemical shifts in question. In this context, the NMR data of filipin
III are informative [11]; the chemical shifts for the C(13) (64.7 ppm in (D6)DMSO) and
C(11) (68.3) C-atoms deviate significantly from the expected standard value (Fig. 8).
Interestingly, the chemical shift of the proton attached to the C(13) is also abnormally
shifted upfield13). Using the abnormalities of chemical shifts noticed for HÀC(13) and
C(13) of filipin III as a reference, we speculate the magnitude of anisotropic effects to
be relatively small for the case of desertomycins/oasomycins and mycoticin A. Second,
13
)
The chemical shifts of HÀC(13) and HÀC(11) of filipin III were found to be 3.04 ppm ((D6)DMSO) and
3.77, respectively [11]. These are significantly upfield-shifted, in reference to the chemical shifts for the
proton attached to the central C-atom of 1a (3.80 ppm, (D6)DMSO), 1b (3.81), 1c (3.80), and 1d (3.86).

2570
Helvetica Chimica Acta ± Vol. 83 (2000)
mycoticin A is known to be conformationally relatively rigid, and its preferred
conformation is similar to the solid-state conformation of roxaticin, in which the 1,3-
polyol chain exists in an extended form [10]. As previously stated, the universal NMR
databases contain conformational information, including the distribution of conformer
population for a given system. The small observed discrepancy may thus be due to a
difference in the population of conformers. This analysis may explain why larger
deviations in chemical shifts from the expected values were observed in mycoticin A
than in desertomycins/oasomycins. Because of a smaller ring size and also because of
the presence of a conjugated polyene, the macrolactone embedding the 1,3,5-triol
systems of mycoticin A is conformationally less flexible than that of desertomycins/
oasomycins, which is reflected in the 13C chemical shifts.
In summary, it has been recognized, and experimentally proven, that the central C-
atom of a 1,3,5-triol system exhibits a distinctive chemical shift which is dependent on
the 1,3- and 3,5-relative configuration, but is independent of the functionalities present
outside of this structural motif. A similar phenomenon has been observed for the polyol
systems described in Fig. 4. We believe that this phenomenon can be extended to
more generalized structures such as B and the corresponding structures bearing the
terminal functionalities shown in Fig. 4. We expect that many structural motifs with
unique structural characteristics independent of the rest of the functionalities present
within a molecule will be identified. In this context, we have recently recognized that
the chemical shift of the Me group in the partial structure represented by C shows a
unique property, and we are currently completing its NMR database.

Helvetica Chimica Acta ± Vol. 83 (2000)
2571
We gratefully acknowledge Dr. Kenichi Tezuka for recording 2D-NMR spectra. Financial support from the
National Institutes of Health (NS 12108) is gratefully acknowledged.
REFERENCES
[1] J. K. Cha, W. J. Christ, J. M. Finan, H. Fujioka, Y. Kishi, L. L. Klein, S. S. Ko, J. Leder, W. W. McWhorter,
Jr., K.-P. Pfaff, M. Yonaga, D. Uemura, Y. Hirata, J. Am. Chem. Soc. 1982, 104, 7369, and preceding papers.
[2] a) C. D. Boyle, J.-C. Harmange, Y. Kishi, J. Am. Chem. Soc. 1994, 116, 4995; b) C. D. Boyle, Y. Kishi,
Tetrahedron Lett. 1995, 36, 5695 and refs. cit. therein; c) H. Oikawa, I. Matsuda, T. Kagawa, A. Ichihara, K.
Â
Kohmoto, Tetrahedron 1994, 50, 13347; d) T. R. Hoye, J. I. Jimenez, W. T. Shier, J. Am. Chem. Soc. 1994,
116, 9409; e) J. W. ApSimon, B. A. Blackwell, O. E. Edwards, A. Fruchier, Tetrahedron Lett. 1994, 35, 7703;
f) G. K. Poch, R. G. Powell, R. D. Plattner, D. Weisleder, Tetrahedron Lett. 1994, 35, 7707; g) B. A.
Blackwell, O. E. Edwards, J. W. ApSimon, A. Fruchier, Tetrahedron Lett. 1995, 36, 1973.
[3] a) W. Zheng, J. A. DeMattei, J.-P. Wu, J. J.-W. Duan, L. R. Cook, H. Oinuma, Y. Kishi, J. Am. Chem. Soc.
1996, 118, 7946; b) L. R. Cook, H. Oinuma, M. A. Semones, Y. Kishi, J. Am. Chem. Soc. 1997, 119, 7928;
c) M. Sasaki, N. Matsumori, T. Maruyama, T. Nonomura, M. Murata, K. Tachibana, T. Yasumoto, Angew.
Chem., Int. Ed. 1996, 35, 1672; d) T. Nonomura, M. Sasaki, N. Matsumori, M. Murata, K. Tachibana, T.
Yasumoto, Angew. Chem., Int. Ed. 1996, 35, 1675.
[4] a) Y. Kobayashi, J. Lee, K. Tezuka, Y. Kishi, Org. Lett. 1999, 1, 2177; b) J. Lee, Y. Kobayashi, K. Tezuka, Y.
Kishi, Org. Lett. 1999, 1, 2181.
[5] a) S. Sakuda, M. Ono, K. Furihata, J. Nakayama, A. Suzuki, A. Isogai, J. Am. Chem. Soc. 1996, 118, 7855;
b) M. Ono, S. Sakuda, H. Ikeda, K. Furihata, J. Nakayama, A. Suzuki, A. Isogai, J. Antibiot. 1998, 51, 1019;
c) H. Ikeda, N. Matsumori, M. Ono, A. Suzuki, A. Isogai, H. Nagasawa, S. Sakuda, J. Org. Chem. 2000, 65,
438.
[6] R. W. Hoffmann, U. Weidmann, Chem. Ber. 1985, 118, 3980.
[7] a) Quinolidomicin A: Y. Hayakawa, K. Shin-ya, K. Furihata, H. Seto, J. Antibiot. 1993, 46, 1563;
b) linearmycin A: S. Sakuda, U. Guce-Bigol, M. Itoh, T. Nishimura, Y. Yamada, Tetrahedron Lett. 1995, 36,
2777; c) shurimycin A; degradation product 3: S. Kumazawa, Y. Asami, K. Awane, H. Ohtani, C. Fukuchi,
T. Mikawa, T. Hayase, J. Antibiot. 1994, 47, 688; d) monazomycin; degradation product 7: H. Nakayama, K.
Furihata, H. Seto, N. Otake, Tetrahedron Lett. 1981, 22, 5217; e) malolactomycin A: H. Koshino, K.
Kobinata, J. Uzawa, M. Uramoto, K. Isono, H. Osada, Tetrahedron 1993, 49, 8827; f) brasilinolide A: H.
Shigemori, Y. Tanaka, K. Yazawa, Y. Mikami, J. Kobayashi, Tetrahedron 1996, 52, 9031; g) mathemycin A:
T. Mukhopadhyay, E. K. S. Vijayakumar, S. R. Nadkarni, H.-W. Fehlhaber, H. Kogler, S. Petry, J. Antibiot.
1998, 51, 582; h) blasticidin A: S. Sakuda, M. Ono, H. Ikeda, Y. Inagaki, J. Nakayama, A. Suzuki, A. Isogai,
Tetrahedron Lett. 1997, 38, 7399; i) theonezolide A: J. Kobayashi, K. Kondo, M. Ishibashi, M. R. Wälchli, T.
Nakamura, J. Am. Chem. Soc. 1993, 115, 6661; j) kanchanamycin A: H. Stephan, C. Kempter, J. W.
Metzger, G. Jung, O. Potterat, C. Pfefferle, H.-P. Fiedler, J. Antibiot. 1996, 49, 765.
[8] a) Y. Kobayashi, C.-H. Tan, Y. Kishi, submitted for publication; b) C.-H. Tan, Y. Kobayashi, Y. Kishi,
submitted for publication.
[9] a) S. L. Schreiber, M. T. Goulet, J. Am. Chem. Soc. 1987, 109, 8120 and Tetrahedron Lett. 1987, 28, 6001;
Â
Â
b) L. Szilagyi, P. Sandor, Magn. Reson. Chem. 1990, 28, 963.
[10] H. Maehr, R. Yang, L.-N. Hong, C.-M. Liu, M. H. Hatada, L. J. Todaro, J. Org. Chem. 1989, 54, 3816.
[11] P. Rochet, J.-M. Lancelin, Magn. Reson. Chem. 1997, 35, 538.
Received July 31, 2000