Saccharide libraries as potential templates for regio- and chiroselective introduction of two functional groups into [60]fullerene

Article abstract of DOI:10.1021/jo9822277

This paper reports regio- and chiroselective introduction of two boronic acid groups into [60]fullerene controlled by a saccharide used as a template molecule. The double [4 + 2] cycloadditions between [60]fullerene and 1:2 saccharide-boronic acid complexes 6-9 afforded [60]fullerene-bisadducts 12. Their structures were identified on the basis of ¹H and ¹³C NMR, UV-vis, and CD spectroscopy, mass spectrometry, and chiral HPLC analysis. When 3-O- methyl-D-glucose (4) was used as the template molecule, high regioselectivity was achieved which gave trans-4 isomer 12a as a main isomer in 72.5% yield. The chiro- as well as regioselective preparation of e isomer 12c was attained in 81.4% ee from the 55.7% yield racemic mixture by the reaction using the D- mannitol-3,4-carbonate template (3). When the enantiomers, D-threitol (D-2) and L-threitol (L-2) were used as the templates, cis-3 isomers 12b and ent- 12b with opposite chirality were yielded in 44.2 and 45.2% ee, respectively. On the other hand, 1-O-methyl-α-D-mannopyranoside template (5) featured nonselective cycloaddition.

Full text of DOI:10.1021/jo9822277

984
J . Org. Chem. 1999, 64, 984-990
Sa cch a r id e Libr a r ies a s P oten tia l Tem p la tes for Regio- a n d
Ch ir oselective In tr od u ction of Tw o F u n ction a l Gr ou p s in to
[60]F u ller en e
Tsutomu Ishi-i,† Kazuaki Nakashima,† Seiji Shinkai,*,† and Atsushi Ikeda‡
Chemotransfiguration Project, J apan Science and Technology Corporation (J ST), 2432 Aikawa,
Kurume, Fukuoka 839-0861, J apan, and Department of Chemistry and Biochemistry,
Graduate School of Engineering, Kyushu University, 6-10-1 Hakozaki, Higashi-ku, 812-8581, J apan
Received November 6, 1998
This paper reports regio- and chiroselective introduction of two boronic acid groups into [60]fullerene
controlled by a saccharide used as a template molecule. The double [4+2] cycloadditions between
[60]fullerene and 1:2 saccharide-boronic acid complexes 6-9 afforded [60]fullerene-bisadducts
12. Their structures were identified on the basis of 1H and 13C NMR, UV-vis, and CD spectroscopy,
mass spectrometry, and chiral HPLC analysis. When 3-O-methyl-D-glucose (4) was used as the
template molecule, high regioselectivity was achieved which gave trans-4 isomer 12a as a main
isomer in 72.5% yield. The chiro- as well as regioselective preparation of e isomer 12c was attained
in 81.4% ee from the 55.7% yield racemic mixture by the reaction using the D-mannitol-3,4-carbonate
template (3). When the enantiomers, D-threitol (D-2) and L-threitol (L-2) were used as the templates,
cis-3 isomers 12b and ent-12b with opposite chirality were yielded in 44.2 and 45.2% ee, respectively.
On the other hand, 1-O-methyl-R-D-mannopyranoside template (5) featured nonselective cycload-
dition.
In tr od u ction
is expected. It is not yet possible to selectively obtain one
of these regioisomers, because of the lack of potential
templates which can cover all of the possible spacer
lengthes between the two [6,6] junction bonds on the [60]-
fullerene surface. Furthermore, chiroselective bifunction-
alization of [60]fullerene has scarcely been attained
except for a few limited reports.2,3,5 Is there any template
family as a practical library that can overcome the
problems described above? We recently noticed that the
saccharide family could be a potential candidate, because
naturally abundant saccharides can be used and their
inherent chirality would be very useful for chiro- as well
as regioselective bifunctionalization of [60]fullerene. This
methodology will be achieved by utilizing the saccharide-
boronic acid interaction,10 in which boronic acids react
with 1,2- or 1,3-diols in saccharide molecules to form
cyclic boronate esters. Thus, we chose saccharides as
templates and boronic acids as functional groups for
regio- and chiroselective bifunctionalization of [60]-
fullerene. Furthermore, when the saccharide-boronic
acid interaction is used to make a memory for the original
template saccharide on the [60]fullerene surface, one can
expect that removal and rebinding of the saccharide
templates can occur reversibly (Scheme 1). This revers-
ible system between bifunctionalized [60]fullerene de-
rivatives and template molecules has never been re-
ported.
Recently, regioselective introduction of two functional
groups into [60]fullerene has been of much concern.1-8
The principle idea is to use a template molecule by which
two functional molecules are arranged at the desired
angle and distance. For example, when bifunctionaliza-
tion (i.e., cyclopropanation,2-4 [3+2] cycloaddition,5,6 and
[4+2] cycloaddition7) occurs on the [6,6] junction bonds
between the two six-membered rings on [60]fullerene
surface, the formation of eight different regioisomers (cis-
1, cis-2, cis-3, e, trans-1, trans-2, trans-3, and trans-4)9
* To whom correspondence should be addressed. Fax: +81 942 39
9012. Phone: +81 942 39 9011. E-mail: seijitcm@mbox.nc.kyushu-
u.ac.jp.
† Chemotransfiguration Project, J apan Science and Technology
Corporation (J ST).
‡ Department of Chemistry and Biochemistry, Graduate School of
Engineering, Kyushu University.
(1) (a) Isaacs, L.; Haldimann, R. F.; Diederich, F. Angew. Chem.,
Int. Ed. Engl. 1994, 33, 2339-2342. (b) Isaacs, L.; Diederich, F.;
Haldimann, R. E. Helv. Chim. Acta 1997, 80, 317-342.
(2) Nierengarten, J .-F.; Gramlich, V.; Cardullo, F.; Diederich, F.
Angew. Chem., Int. Ed. Engl. 1996, 35, 2101-2103.
(3) Nierengarten, J .-F.; Habicher, T.; Kessinger, R.; Cardullo, F.;
Diederich, F.; Gramlich, V.; Gisselbrecht, J .-P.; Boudon, C.; Gross, M.
Helv. Chim. Acta 1997, 80, 2238-2276.
(4) (a) Nierengarten, J .-F.; Herrmann, A.; Tykwinski, R. R.; Ru¨tti-
mann, M.; Diederich, F.; Boudon, C.; Gisselbrecht, J .-P.; Gross, M. Helv.
Chim. Acta 1997, 80, 293-316. (b) Nierengarten, J .-F.; Schall, C.;
Nicoud, J .-F. Angew. Chem., Int. Ed. Engl. 1998, 37, 1934-1936. (c)
Nierengarten, J .-F.; Felder, D.; Nicoud, J .-F. Tetrahedron Lett. 1998,
39, 2747-2750. (d) Nierengarten, J .-F.; Schall, C.; Nicoud, J .-F.;
Heinrich, B.; Guillon, D. Tetrahedron Lett. 1998, 39, 5747-5750.
(5) Nakamura, E.; Isobe, H.; Tokuyama, H.; Sawamura, M. Chem.
Commun. 1996, 1747-1748.
In a preliminary communication, we reported the
regioselective double [4+2] cycloaddition between [60]-
fullerene and saccharide-3,4-bis(bromomethyl)phenyl-
boronic acid (1:2) complexes, which can generate o-quin-
odimethane species11 to react with the two [6,6] junction
bonds in [60]fullerene (Scheme 1).12 The regiospectrum
(6) Isobe, H.; Tokuyama, H.; Sawamura, M.; Nakamura, E. J . Org.
Chem. 1997, 62, 5034-5041.
(7) Taki, M.; Sugita, S.; Nakamura, Y.; Kasashima, E.; Yashima,
E.; Okamoto, Y.; Nishimura, J . J . Am. Chem. Soc. 1997, 119, 926-
932.
(8) (a) Kampe, K.-D.; Egger, N.; Vogel, M. Angew. Chem., Int. Ed.
Engl. 1993, 32, 1147-1176. (b) Shin, L.-L.; Chien, K.-M.; Liu, T.-Y.;
Lin, T.-I.; Her, G.-R.; Luh, T.-Y. J . Chem. Soc., Chem. Commun. 1995,
1159-1160. (c) Dong, G.-X.; Li, J .-S.; Chan, T.-H. J . Chem. Soc., Chem.
Commun. 1995, 1725-1726. (d) Shen, C. K.-F.; Chien, K.-M.; J uo, C.-
G.; Her, G.-R.; Luh, T.-Y. J . Org. Chem. 1996, 61, 9242-9244.
(9) Naming of [60]fullerene-bisadducts: Hirsch, A.; Lamparth, I.;
Karfunkel, H. R. Angew. Chem., Int. Ed. Engl. 1994, 33, 437-438.
(10) For comprehensive reviews for saccharide-boronic acid interac-
tions, see: (a) J ames, T. D.; Sandanayake, K. R. A. S.; Shinkai, S.
Angew. Chem., Int. Ed. Engl. 1996, 35, 1910-1922. (b) J ames, T. D.;
Linnane, P.; Shinkai, S. Chem. Commun. 1996, 281-288.
10.1021/jo9822277 CCC: $18.00 © 1999 American Chemical Society
Published on Web 01/16/1999

Saccharide Libraries as Potential Templates
J . Org. Chem., Vol. 64, No. 3, 1999 985
Sch em e 1
Sch em e 2
was closely related to a structure of the saccharide used
as the template molecule. Here, we report a full account
of this study including the new findings of the chirose-
lective introduction of two boronic acid groups into [60]-
fullerene, which is controlled by the inherent chirality
of saccharide templates.
Resu lts a n d Discu ssion
P r epar ation of 1:2 Sacch ar ide-Bor on ic Acid Com -
p lexes. To access 1:2 saccharide-boronic acid complexes
6-9, we selected five saccharides, D-threitol (D-2), L-
threitol (L-2), D-mannitol-3,4-carbonate (3), 3-O-methyl-
D-glucose (4), and 1-O-methyl-R-D-mannopyranoside (5),
as templates. The complexes were prepared from 2,4,6-
[3,4-bis(bromomethyl)phenyl]boroxin 113 and the corre-
sponding saccharides in refluxing toluene with azeotropic
removal of water (Scheme 2). The products 6-9 were
identified by means of spectral data and elemental
analyses. For the complex obtained from 1 and threitol
2, one can propose two different cyclic structures, a 1,2-
adduct with two five-membered rings and a 1,3-adduct
with two six-membered rings. Previously, we confirmed
on the basis of the X-ray crystallographic study that in
6 the two boronic acids react with the 1,3-diols in threitol
to form the two six-membered rings.13 Extensive 1H NMR
[60]fullerene connected with a saccharide-bisboronate
moiety, we adopted the dilute-concentration conditions
at 1 × 10-3 mol dm-3. To simplify the product analysis,
the saccharides in the product 10 were removed by
treatment with aqueous 1.2 mol dm-3 HCl solution. Then,
the boronic acid groups in 11 were protected with 2,2-
dimethylpropane-1,3-diol. Finally, the desired bisadduct
12 was isolated by column chromatography (silica gel,
toluene/chloroform) in 46-58% yields as a regioisomeric
mixture.
Isom er Distr ibu tion of Bisa d d u cts. To analyze the
isomer distribution in 12, the regioisomeric mixture of
12 was subjected to HPLC analysis using a COSMOSIL
5PBB column eluting with n-hexane/toluene (3:7 (v/v),
1.0 mL/min). The results are summarized in Table 1. In
HPLC analysis, seven or eight peaks were always ob-
servable in the bisadduct region and the relative intensity
changed depending on the saccharide template. Com-
pounds 6, ent-6, 7, and 8 afforded a major peak for peak
7 (47.3%), peak 7 (47.8%), peak 8 (55.7%), and peak 6
(72.5%), respectively, indicating that a single regioisomer
was selectively formed in each case. On the other hand,
compound 9 featured a rather nonselective product
1
and H-1H COSY studies of 8 showed that the saccharide
moiety for 8 adopts the furanose form: the coupling
constants of the saccharide moiety were very similar to
those of 1,2:5,6-bisarylglucofuranose reported previ-
ously.14
Rea ction betw een [60]F u ller en e a n d 1:2 Sa cch a -
r id e-Bor on ic Acid Com p lexes. The intramolecular
double additions between [60]fullerene and 6-9 were
performed in refluxing toluene in the presence of 18-
crown-6 and KI under a nitrogen atmosphere (Scheme
3). To avoid the intermolecular double addition giving bis-
(11) Belik, P.; Gu¨gel, A.; Spickermann, J .; Mu¨llen, K. Angew. Chem.,
Int. Ed. Engl. 1993, 32, 78-80.
(12) Ishi-i, T.; Nakashima, K.; Shinkai, S. Chem. Commun. 1998,
1047-1087.
(13) Ishi-i, T.; Nakashima, K.; Shinkai, S.; Araki, K. Tetrahedron
1998, 54, 8679-8686. Previously, we imaged a structure with two five-
membered rings for the D-threitol complex linked to porphyrins via
boronic acid groups: Takeuchi, M.; Chin, Y.; Imada, T.; Shinkai, S.
Chem. Commun. 1996, 1867-1868. Although it was very difficult to
discriminate between these two structures, this complex may also
adopt a structure with two six-membered rings.
(14) (a) Wood, P. J .; Siddiqui, I. R. Carbohydr. Res. 1974, 36, 247-
256. (b) Norrild, J . C.; Eggert, H. J . Am. Chem. Soc. 1995, 117, 1479-
1484.

986 J . Org. Chem., Vol. 64, No. 3, 1999
Ishi-i et al.
Sch em e 3
isolation of peak 7 (12b from 6, ent-12b from ent-6) and
peak 8 (12c) was achieved only by a HPLC separation
method using a COSMOSIL 5PBB column eluting with
n-hexane/toluene (3:7 (v/v), 0.5 mL/min). Finally, 12a ,
12b, ent-12b, and 12c can be obtained in 16%, 10%, 10%,
and 11% isolated yield from 8, 6, ent-6, and 7, respec-
tively. The regioisomers isolated, 12a , 12b, ent-12b, and
12c, showed a single peak, which is consistent with peak
6, peak 7, peak 7, and peak 8, in their HPLC chromato-
grams, respectively.
Id en tifica tion of Bisa d d u cts. The structures of 12a ,
12b, ent-12b, and 12c were identified to be trans-4, cis-
3, cis-3, and e, respectively, on the basis of the molecular
1
symmetry deduced from H and 13C NMR spectra.15 The
comparison of their UV-vis spectra with those reported
in the preceding reports9,16 also supported these assign-
ments. The derivation of 12 to Nishimura’s phenol
17
derivative 137 by treatment with H2O2 also afforded
very helpful information for the structure identification
of 12a , 12b, ent-12b, and 12c (Scheme 4). The details of
the assignment processes are described below.
1
In the H NMR spectrum (toluene-d8, 90 °C) of 12a ,
one could observe one CH3 proton peak (0.71 ppm, 12H)
and one CH2 proton peak (3.50 ppm, 8H) for the protec-
tive groups and four CH2 proton peaks (4.04, 4.04, 4.12,
and 4.19 ppm, 2H each) and three ArH proton peaks
(7.42, 8.13, and 8.17 ppm, 2H each) for the o-xylenyl
groups. This splitting pattern is commensurate with
either C2 symmetry (cis-3, trans-2, or trans-3) or Cs
symmetry (cis-1, cis-2, trans-1, or trans-4).9,16 Among
them, the size of the saccharide template is actually too
small or too large to give cis-1, trans-1, trans-2, and trans-
3; therefore, these isomers can be excluded. The residual
possible isomers are cis-2, cis-3, and trans-4. The solubil-
ity of 12a into deuterated NMR solvents was not so high
as to obtain a satisfactory 13C NMR spectrum, so that
we converted 12a to more soluble 13a by treatment with
H2O2.17 Compound 13a did not coincide either with cis-2
or with cis-3, reported by Nishimura et al.7 Furthermore,
the 13C NMR spectrum of 13a (DMSO-d6, 120 °C) gave
30 sp2 carbon peaks and 2 sp3 carbon peaks for the [60]-
fullerene moiety. This splitting pattern is commensurate
with Cs symmetrical cis-2 and trans-4.9,16 As the summary
of the results mentioned above, 12a can be identified to
be trans-4, which satisfies all of the above-mentioned
information.
Ta ble 1. Resu lts of HP LC An a lysisa for 12
product percentageb in 12 obtained from
peak
no.
retention
time/min
6
ent-6
7
8
9
1
2
3
4
5
6
7
8
39.6
41.5
43.0
46.5
49.7
51.2
52.9
54.1
1.5
2.3
5.4
2.9
5.3
8.5
5.8
8.1
10.7
47.8
10.9
4.2
4.3
3.4
10.5
7.7
14.2
0
0.2
1.5
3.0
3.7
3.6
72.5
0
5.6
7.5
11.4
15.2
23.7
15.6
0
7.8
12.3
11.3
47.3
12.1
Bisadducts 12b and ent-12b showed the same 1H NMR
spectrum (C2D2Cl4, 90 °C), and the splitting pattern was
commensurate with either C2 symmetry or Cs sym-
metry.9,16 In the 13C NMR spectra (C2D2Cl4, 90 °C) of 12b
and ent-12b, the splitting pattern with 28 sp2 carbon
peaks and 2 sp3 carbon peaks for the [60]fullerene moiety
indicates the C2 symmetrical cis-3, trans-2, or trans-3.18
J udging from the size of the template saccharides, only
55.7
15.5
21.0
a
Conditions: stationary phase, COSMOSIL 5PBB (10 × 250
mm); eluent, n-hexane/toluene (3:7 (v/v)); flow rate, 1.0 mL/min.
b
Computed from the peak area obtained by the chromatogram
followed at 350 nm.
distribution. The results imply that the saccharide
changes the orientation between the two bis(bromom-
ethyl)phenyl groups in 6-9 and the addition reaction
occurs with the different [6,6] junction bonds on the [60]-
fullerene surface.
Isola tion of Bisa d d u cts. It seems difficult to isolate
eight different isomers of bisadduct 12 and to identify
the structure of each isomer. We thus decided to chal-
lenge the isolation of main regioisomers of peak 7 for 6,
peak 7 for ent-6, peak 8 for 7, and peak 6 for 8. We found
that peak 6 (12a ) can be easily isolated by recrystalliza-
tion from dichloromethane/n-hexane. In contrast, the
(15) To obtain the clear spectra, the 1H and 13C NMR spectra of 12
and 13 were measured at high temperature (90-120 °C), since the
spectra were broadened at room temperature because of the slow
flipping motions around the two cyclohexene rings in 12 and 13. For
the flipping motion in methano[1,2]benzenomethano[60]fullerenes,
see: Nakamura, Y.; Minowa, T.; Tobita, S.; Shizuka, H.; Nishimura,
J . J . Chem. Soc., Perkin Trans. 2 1995, 2351-2357.
(16) (a) Schick, G.; Hirsch, A.; Mauser, H.; Clark, T. Chem. Eur. J .
1996, 2, 935-943. (b) Djojo, F.; Herzog, A.; Lamparth, I.; Hampel, F.;
Hirsch, A. Chem. Eur. J . 1996, 2, 1537-1547.
(17) (a) Yarboro, T. L.; Karr, C., J r. J . Org. Chem. 1959, 24, 1141-
1143. (b) Kidwell, R. L.; Murphy, M.; Darling, S. D. Organic Synthesis;
Wiley: New York, 1973; Collect. Vol. V, pp 918-921.

Saccharide Libraries as Potential Templates
J . Org. Chem., Vol. 64, No. 3, 1999 987
Sch em e 4
the formation of cis-3 is possible as the structures of 12b
and ent-12b.
characteristic absorption maxima were observed at 643
and 708 nm for 12a , at 661 and 731 nm for 12b, at 661
and 731 nm for ent-12b, and at 423 nm for 12c,
respectively (Figure 1). These maxima are consistent with
those in the previous reports.9,16
The spectral data of phenol derivatives 13b (ent-13b)
and 13c coincide with those of cis-3 and e reported by
Nishimura et al., respectively.7 In contrast, 13a shows
different spectral behavior compared to cis-2, cis-3, and
e reported by Nishimura et al.7 As a summary of the
foregoing findings deduced from the spectral data and
the derivation, 12a , 12b (ent-12b), and 12c are unequivo-
cally identified to be trans-4, cis-3, and e, respectively.20
Op tica lly Active Bisa d d u cts. Since C2 symmetrical
1
The H and 13C NMR spectra (C2D2Cl4, 90 °C) of 12c
gave the two sets of proton and carbon peaks for the two
boronic acid groups, respectively, suggesting that the two
boronic acid groups in 12c are chemically unequivalent.19
Since the spectral behavior is observable only for C1
symmetrical e, 12c was identified to be e.9,16
The UV-vis spectra of 12a , 12b, ent-12b, and 12c are
also supportive of their proposed structure. It is already
known that the UV-vis spectra of [60]fullerene-bisad-
ducts depend on the addition pattern rather than on the
nature of the substituent.9,16 In the UV-vis spectra of
12a , 12b, ent-12b, and 12c in dichloromethane, the
(18) Although the theoretical number of sp2 carbon peaks for the
[60]fullerene moiety as well as the benzene moieties in 12b and ent-
12b should be 34, the peak number observed was 33. Probably, a peak
of the adjacent carbon to the boron atom would be missing.
(19) Theoretically, the total number of sp2 carbon peaks for the [60]-
fullerene moiety as well as the benzene moieties in 12c should be 68.
However, it was very difficult to count all of the sp2 carbon peaks due
of the overlapping of some peaks.

988 J . Org. Chem., Vol. 64, No. 3, 1999
Ishi-i et al.
Ta ble 2. Resu lts of Ch ir a l HP LC An a lysisa a n d
En a n tiom er ic Excess (% ee) of 13a , 13b, en t-13b, a n d 13c
retention time/min
(peak intensity/%)
% ee
13a
13b
29.1 (100)b
39.3 (72.1)
51.6 (27.9)
39.3 (27.4)
51.6 (72.6)
29.4 (9.3)b
72.8 (90.7)b
44.2
45.2
81.4
ent-13b
13c
a
Conditions: stationary phase, CHIRALCEL OD-H (4.6 × 250
b
mm); eluent, n-hexane/2-propanol (9:1 (v/v)); flow rate, 1.0 mL/
min. n-Hexane/2-propanol (7:3 (v/v)).
F igu r e 1. Absorption spectra of 12a , 12b, ent-12b, and 12c
in dichloromethane.
F igu r e 3. CD spectrum of 12c in dichloromethane (2 × 10-4
mol dm-3, 1 cm cell).
Nishimura and Okamoto’s groups.7,22 The chromatograms
of 13b and ent-13b showed the two peaks with different
intensity, and the intensity ratio of the two peaks was
exchanged in each case of 13b and ent-13b (Table 2). On
the basis of the results of the chiral HPLC analysis, the
enantiomeric excesses of 12b and ent-12b were estimated
to be 44.2% ee and 45.2% ee, respectively (Table 2).23
Although 12c with C1 symmetry was also CD-active in
dichloromethane, the ∆ꢀ values of 12c were smaller by
1 or 2 orders of magnitude compared to those of 12b and
ent-12b with C2 symmetry (Figure 3). The similar weak
Cotton effects in e isomer were also reported by Diederich
et al.3 We assign the weak Cotton effects to induced
circular dichroism24 originating from the perturbation of
the achiral fullerene chromophore in 12c by the local
asymmetry based on the direction of the B(OCH2C(CH3)2-
CH2O) group in the benzene ring. Bisadduct 12c showed
the high enantiomeric excess of 81.4% ee, which was
deduced from chiral HPLC analysis of 13c (Table 2).23
On the other hand, 12a with achiral Cs symmetry was
CD-silent in dichloromethane, and as expected, 13a gave
a single peak in HPLC analysis with a chiral packed
column (Table 2).
F igu r e 2. CD spectra of 12b and ent-12b in dichloromethane
(2 × 10-5 mol dm-3, 1 cm cell).
cis-3 and C1 symmetrical e addition patterns are chiral,
one can expect that 12b, ent-12b, and 12c are chirose-
lectively yielded from 6, ent-6, and 7 because of the
inherent chirality of saccharide templates, D-2, L-2, and
3, respectively. Thus, two pairs of enantiomers 12bA/
12bB and 12cA/12cB are available, and one of the two
possible enantiomers should be preferentially formed in
each reaction of 6, ent-6, and 7 (Scheme 4). This chirose-
lectivity was investigated on the basis of the CD (circular
dichroism) spectra and HPLC analysis with a chiral
packed column.
As expected, 12b and ent-12b were CD-active in
dichloromethane, and the spectral shape was symmetri-
cal to each other (Figure 2). The molar circular-dichroic
absorptions (∆ꢀ) of 12b and ent-12b are as large as those
of chiral cis-3 isomers reported previously.2,3,5 The strong
Cotton effects observed in these spectra are mainly due
to strong chiroptical contributions from the chiral fullerene
chromophore in 12b and ent-12b with C2 symmetry.21
The optical purities of these compounds were estimated
by a HPLC method (CHIRALCEL OD-H) developed by
(21) (a) Hawkins, J . M.; Meyer, A.; Nambu, M. J . Am. Chem. Soc.
1993, 115, 9844-9845. (b) Herrmann, A.; Ru¨ttimann, M.; Thilgen, C.;
Diederich, F. Helv. Chim. Acta 1995, 78, 1673-1704.
(22) Okamoto, Y.; Yashima, E. Angew. Chem., Int. Ed. Engl. 1998,
37, 1020-1043.
(23) Although we attempted the optical resolution of 12b, ent-12b,
and 12c, it was not achieved by using any chiral stationary phases.
(24) The Cotton effects due to induced circular dichroism were found
in some [60]fullerene derivatives having a chiral substituent. For
example: (a) Bianco, A.; Maggini, M.; Scorrano, G.; Toniolo, C.;
Marconi, G.; Villani, C.; Prato, M. J . Am. Chem. Soc. 1996, 118, 4072-
4080. (b) Wilson, S. R.; Lu, Q.; Cao, J .; Wu, Y.; Welch, C. J .; Schuster,
D. I. Tetrahedron 1996, 52, 5131-5142.
(20) Three isomers (in-in, out-in, and out-out) in 12a , 12b, and ent-
12b are available, in principle, owing to the direction of the
B(OCH2C(CH3)2CH2O) group in the benzene ring. J udging from the
size of the template saccharides and the splitting patterns of the 1H
and 13C NMR spectra, the out-in and out-out isomers can be excluded.
Thus, 12a , 12b, and ent-12b isolated here are the in-in isomer. Also,
two isomers (in and out) are available in 12c. Similarly, 12c is
identified as in isomer. For an explanation of the concept of the in-in,
out-in, and out-out isomers and the in and out isomers, see refs 2-7.

Saccharide Libraries as Potential Templates
J . Org. Chem., Vol. 64, No. 3, 1999 989
mg, 1.0 mmol) and L-2 (183 mg, 1.5 mmol) in dry toluene (150
mL) was heated at the reflux temperature with azeotropic
removal of water (Dean-Stark) for 2 h under a nitrogen
atmosphere. After the reaction mixture was cooled to room
temperature, it was evaporated in vacuo to dryness. The solid
residue was purified by recrystallization from dichloromethane/
n-hexane to give ent-6 in 72% yield (720 mg, 1.08 mmol) as
colorless prisms. The spectral data of ent-6 coincided with those
of 6:13 mp 176-177 °C (6; 176-177 °C).13
As a summary of the foregoing findings with the CD
spectra and chiral HPLC analyses, it is clear that the
double additions between [60]fullerene and 6, ent-6, and
7 chiroselectively afforded the optically active 12b, ent-
12b, and 12c, respectively. The stereospectrum depends
on the inherent chirality of saccharides used as the
template molecules. In particular, 6 and ent-6 afforded
the opposite enantiomers (12bA/12bB) preferentially, as
deduced from the sign of the Cotton effects in CD spectra
(Figure 2) and the peak intensity in chiral HPLC analysis
(Table 2).
D-Ma n n itol-1,2:5,6-bis[3,4-bis(br om om eth yl)p h en ylbo-
r on a te]-3,4-ca r bon a te (7). According to the same procedure
used for the synthesis of ent-6, 7 was obtained by complexation
of 1 (869 mg, 1.0 mmol) with 326 (312 mg, 1.5 mmol) in dry
toluene (230 mL) for 3 h in 95% yield (1.07 g, 1.42 mmol):
colorless prisms (dichloromethane/n-hexane); mp 90-91 °C;
IR (KBr) νmax 2973, 2909, 1828, 1798, 1610, 1412 (νBC), 1356
(νBO), 1217, 1098, 785, 681 cm-1; 1H NMR (CDCl3) δ 4.32 (dd,
J ) 6.0, 9.9 Hz, 2H, CH2), 4.50 (d, J ) 10.6 Hz, 2H, CH), 4.62
(ddd, J ) 9.9, 10.6, 13.6 Hz, 2H, CH), 4.65, 4.66 (s, each 4H,
CH2Br), 4.77 (dd, J ) 6.0, 13.6 Hz, 2H, CH2), 7.31, 7.60 (d, J
) 7.6 Hz, each 2H, ArH), 7.71 (s, 2H, ArH). Anal. Calcd for
Con clu d in g Rem a r k s
In this study, we have demonstrated that saccharides
are very useful as a versatile template for regio- and
chiroselective introduction of two functional groups into
[60]fullerene. The addition patterns in the double addi-
tions between [60]fullerene and 1:2 saccharide-boronic
acid complexes highly reflect the structure of saccharides
used as the template molecules. In particular, the inher-
ent chirality of saccharide templates plays a decisive role
in the chiroselective double additions leading to chiral
[60]fullerene-bisadducts. It is now possible to chirose-
lectively prepare each enantiomer of chiral [60]fullerene-
bisadducts by utilizing either a D-saccharide or a L-sac-
charide as a template molecule.
In the [60]fullerene-bisadducts obtained here, the
arrangement of the two boronic acid groups would be
complementary to the angle and the distance of the
original saccharide used as the template molecule. When
saccharide-boronic acid interaction is utilized, rebinding
of the bifunctionalized [60]fullerene with the original
saccharide template should occur selectively. This idea
would serve to create the concept of “memory storage for
the template molecule” as found in the molecular im-
printing technique.25 This implies that the molecular
imprinting can be realized even in homogeneous solution
using reactive CdC double bonds in the solubilized [60]-
fullerene surface. We believe that this study is applicable
not only to the regio- and chiroselective preparation of
various bifunctional [60]fullerene derivatives but also to
a new technique for molecular imprinting.
C
23H22B2Br4O7: C, 36.75; H, 2.95. Found: C, 37.12; H, 3.14.
3-O-Meth yl-D-glu cofu r an ose-1,2:5,6-bis[3,4-bis(br om om -
eth yl)p h en ylbor on a te] (8). According to the same procedure
used for the synthesis of ent-6, 8 was obtained by complexation
of 1 (87 mg, 0.1 mmol) with 4 (29 mg, 0.15 mmol) in dry toluene
(30 mL) for 13 h in 76% yield (84 mg, 0.114 mmol): white
powder (dichloromethane/n-hexane); mp 154-155 °C; IR (KBr)
ν
max 2980, 2951, 1647, 1437 (νBC), 1368 (νBO), 1220, 1080, 679,
608 cm-1; H NMR (CDCl3) δ 3.57 (s, 3H, OMe), 3.98 (d, J )
3.2 Hz, 1H, H-3), 4.18 (dd, J ) 3.2, 6.1 Hz, 1H, H-4), 4.42 (d,
J ) 7.2 Hz, 2H, H-6), 4.65 (s, 8H, CH2Br), 4.87 (dt, J ) 6.1,
7.2 Hz, 1H, H-5), 4.93 (d, J ) 4.5 Hz, 1H, H-2), 6.25 (d, J )
4.5 Hz, 1H, H-1), 7.36 (d, J ) 7.8 Hz, 1H, ArH), 7.39 (d, J )
7.8 Hz, 1H, ArH), 7.71 (dd, J ) 1.2, 7.8 Hz, 1H, ArH), 7.74
(dd, J ) 1.2, 7.8 Hz, 1H, ArH), 7.77 (d, J ) 1.2 Hz, 1H, ArH),
7.81 (d, J ) 1.2 Hz, 1H, ArH). Anal. Calcd for C23H24B2Br4O6:
C, 37.45; H, 3.28. Found: C, 37.81; H, 3.45.
1
Met h yl-r-D-m a n n op yr a n osid e-2,3:4,6-b is[3,4-b is(b r o-
m om eth yl)p h en ylbor on a te] (9). According to the same
procedure used for the synthesis of ent-6, 9 was obtained by
complexation of 1 (869 mg, 1.0 mmol) with 5 (305 mg, 1.5
mmol) in dry toluene (300 mL) for 14 h in 60% yield (660 mg,
0.895 mmol): white powder (dichloromethane/n-hexane); mp
190-192 °C; IR (KBr) νmax 2957, 2911, 1611, 1439 (νBC), 1374
(νBO), 1331 (νBO), 1210, 1082, 683, 613 cm-1; 1H NMR (CDCl3)
δ 3.57 (s, 3H, OMe), 3.87-3.99 (m, 3H), 4.26-4.34 (m, 1H),
4.58 (d, J ) 7.2 Hz, 1H), 4.65-4.75 (m, 1H), 4.67, 4.68, 4.69,
4.71 (s, each 2H, CH2Br), 5.13 (s, 1H, H-1), 7.37 (d, J ) 7.5
Hz, 1H, ArH), 7.42 (d, J ) 7.5 Hz, 1H, ArH), 7.78 (dd, J ) 1.1,
7.5 Hz, 1H, ArH), 7.80 (dd, J ) 1.1, 7.5 Hz, 1H, ArH), 7.84 (d,
J ) 1.1 Hz, 1H, ArH), 7.89 (d, J ) 1.1 Hz, 1H, ArH). Anal.
Calcd for C23H24B2Br4O6: C, 37.45; H, 3.28. Found: C, 38.09;
H, 3.34.
Gen er a l P r oced u r e for th e Rea ction betw een [60]-
F u ller en e a n d Sa cch a r id e-Bisbor on a tes 6-9. To a solu-
tion of [60]fullerene (72 mg, 0.1 mmol), 18-crown-6 (1.06 g,
4.0 mmol), and pottasium iodide (166 mg, 1.0 mmol) in dry
toluene (100 mL) was added 6-9 (0.1 mmol) under a nitrogen
atmosphere, and the mixture was heated at the reflux tem-
perature for 40 h. After the reaction mixture was cooled to
room temperature, it was evaporated in vacuo to dryness. The
residue was washed with water and dissolved in THF (20 mL).
Hydrochloric acid (1.2 mol dm-3, 10 mL) was added to the THF
solution at room temperature under an argon atmosphere, and
the mixture was stirred for 12 h. The reaction mixture was
washed with 1% sodium bisulfite brine solution, washed with
brine, dried over anhydrous magnesium sulfate, and evapo-
rated in vacuo to dryness. The residue was washed with
methanol to give crude 11 as a brown solid. The crude 11 and
2,2-dimethylpropane-1,3-diol (21 mg, 2.0 mmol) in dry toluene
Exp er im en ta l Section
Gen er a l Con sid er a tion s. All melting points are uncor-
1
rected. IR spectra were measured as KBr pellets. H and 13C
NMR spectra were recorded on a 300 MHz spectrometer in
CDCl3, toluene-d8, C2D2Cl4, or DMSO-d6 with tetramethylsi-
lane as the internal standard. Mass spectra (negative SIMS)
were measured in 2-nitrobenzyl alcohol. UV-vis and CD
spectra were measured in dichloromethane. The isomer dis-
tribution was determined by HPLC using a COSMOSIL 5PBB
column (10 × 250 mm) eluting with n-hexane/toluene (3:7 (v/
v), 1.0 mL/min). The enantiomeric excess was determined by
HPLC using a CHIRALCEL OD-H column (4.6 × 250 mm)
eluting with n-hexane/2-propanol (7:3-9:1 (v/v), 1.0 mL/min).
Column chromatography was carried out on silica gel (Wako
C-300). Compound 3 was prepared according to a method
described previously.26 The preparations of 1 and 6 were
already reported.13 D-Threitol (D-2), L-threitol (L-2), and methyl-
R-D-mannopyranoside (5) were purchased from SIGMA Co.
Ltd. 3-O-Methyl-D-glucose (4) was purchased from Wako Pure
Chemical Industries, Ltd.
L-Th r eitol-1,3:2,4-bis[-3,4-bis(br om om eth yl)p h en ylbo-
r on a te] (en t-6) (Gen er a l P r oced u r e). A suspension of 1 (869
(25) For comprehensive reviews for the molecular imprinting tech-
nique, see: (a) Wulff, G. Angew. Chem., Int. Ed. Engl. 1995, 34, 1812-
1832. (b) Kriz, D.; Ramstro¨m, O.; Mosbach, K. Anal. Chem. 1997, 69,
345A-349A.
(26) Hough, L.; Priddle, J . E.; Theobald, R. S. J . Chem. Soc. 1962,
1934-1938.

990 J . Org. Chem., Vol. 64, No. 3, 1999
Ishi-i et al.
(20 mL) was heated at the reflux temperature with azeotropic
removal of water (Dean-Stark) for 2 h under a nitrogen
atmosphere. After the reaction mixture was cooled to room
temperature, it was evaporated in vacuo to dryness. The
residue was purified by silica gel column chromatography
eluting with toluene to give a trace amount of [60]fullerene
and monoadduct and eluting with toluene/chloroform (1:1 (v/
v)) to give 12 as a regioisomeric mixture.
146.54, 146.57, 147.10, 147.36, 148.10, 148.14, 148.16, 148.19,
149.15, 150.45, 150.91, 154.66, 154.68, 154.75, 154.78, 155.06,
155.50, 155.62, 155.71, 161.74, 161.75, 163.76 (sp2 C); MS
(negative SIMS, NBA) m/z 1152 (M-); UV-vis (dichlo-
romethane) λmax (ꢀ) 423 (4812), 400 (sh, 5459), 315 (40 351),
232 (119 550) nm; CD (dichloromethane) λmax (∆ꢀ) 730 (1.45),
667 (0.49), 609 (-0.74), 460 (-1.55), 421 (-0.86) nm. Anal.
Calcd for C86H34B2O4‚0.2(CH2Cl2): C, 88.50; H, 2.96. Found:
C, 88.58; H, 3.11.
65,72-Bis(4,4-d im eth yl-2,6-d ioxa -1-bor a cycloh exa n yl)-
1,2:34,35-b is(m et h a n o[1,2]b en zen om et h a n o)[60]fu ller -
en e (12a ). In the reaction of 8, the isolation of main isomer
12a from the regioisomeric mixture was performed by recrys-
tallization from dichloromethane/n-hexane: dark brown solid;
mp >450 °C; IR (KBr) νmax 2955, 2930, 2880, 1476, 1422 (νBC),
65,72-Dih yd r oxy-1,2:34,35-b is(m et h a n o[1,2]b en zen o-
m eth a n o)[60]fu ller en e (13a ) (Gen er a l P r oced u r e). To a
solution of 12a (57 mg, 0.05 mmol) in acetic acid (0.5 mL) and
THF (5 mL) was added 30% H2O2 (0.5 mL) at 0 °C, and the
mixture was stirred for 50 h at room temperature under an
argon atmosphere. The reaction mixture was extracted with
toluene and washed with 1% ammonium sulfate aqueous
solution and water. The organic phase was dried over anhy-
drous magnesium sulfate and evaporated in vacuo to dryness.
The solid residue was purified by silica gel column chroma-
tography eluting with toluene/ethyl acetate (19:1 (v/v)) to give
13a in 56% yield (27 mg, 0.028 mmol). An analytical sample
was obtained as a dark brown solid by recrystallization from
dichloromethane/n-hexane: mp >450 °C; IR (KBr) νmax 3442
(νOH), 2923, 1615, 1501, 1456, 1354, 1269, 1217, 1150, 770, 695,
527 cm-1; 1H NMR (DMSO-d6, 120 °C) δ 4.21 (br s, 4H, CH2),
4.35, 4.44 (d, J ) 13.8 Hz, each 2H, CH2), 6.86 (dd, J ) 2.4,
8.0 Hz, 2H, ArH), 6.94 (d, J ) 2.4 Hz, 2H, ArH), 7.42 (d, J )
8.0 Hz, 2H, ArH), 8.96 (s, D2O-exchange, 2H, OH); 13C NMR
(DMSO-d6, 120 °C) δ 42.65, 42.69 (CH2), 63.73, 64.39 (fullerene-
sp3 C), 113.83, 114.43 (sp2 CH), 127.79 (sp2 C), 127.89 (sp2 CH),
129.23, 133.67, 134.75, 137.29, 137.55, 138.41, 140.05, 140.09,
140.23, 140.67, 141.14, 141.41, 142.32, 143.49, 143.75, 143.80,
144.50, 145.13, 145.24, 145.30, 146.56, 146.96, 147.19, 147.21,
148.44, 148.55, 150.29, 152.21, 152.33, 154.16, 156.30, 156.50
(sp2 C); MS (negative SIMS, NBA) m/z 960 (M-); UV-vis
(cyclohexane/2-propanol ) 1:1 (v/v)) λmax (ꢀ) 709 (756), 643
(921), 450 (sh, 3994), 309 (52 903), 266 (100 000), 227 (139
240) nm. Anal. Calcd for C76H16O2‚1.1(C6H14): C, 93.97; H,
3.00. Found: C, 93.58; H, 2.58.
65,73-Dih yd r oxy-1,2:16,17-b is(m et h a n o[1,2]b en zen o-
m eth a n o)[60]fu ller en e (13b). According to the same proce-
dure used for the synthesis of 13a , 13b7 was prepared from
12b. An analytical sample was obtained as a dark brown solid
by silica gel column chromatography eluting with toluene: mp
>450 °C; 1H NMR (DMSO-d6, 120 °C) δ 3.79, 4.00 (d, J ) 13.7
Hz, 2H, CH2), 4.18, 4.28 (d, J ) 13.7 Hz, 2H, CH2), 6.85 (dd,
J ) 2.3, 8.0 Hz, 2H, ArH), 6.90 (d, J ) 2.3 Hz, 2H, ArH), 7.30
(d, J ) 8.0 Hz, 2H, ArH), 8.61-9.27 (br s, 2H, OH).
65,73-Dih yd r oxy-1,2:16,17-b is(m et h a n o[1,2]b en zen o-
m eth a n o)[60]fu ller en e (en t-13b). According to the same
procedure used for the synthesis of 13a , ent-13b7 was prepared
from ent-12b. An analytical sample was obtained as a dark
brown solid by silica gel column chromatography eluting with
toluene: mp >450 °C.
1
1341, 1320 (νBO), 1293, 1246, 1181, 1123, 668, 527 cm-1; H
NMR (toluene-d8, 90 °C) δ 0.71 (s, 12H, Me), 3.50 (s, 8H, CH2),
4.04 (s, 4H, CH2), 4.12, 4.19 (d, J ) 14.4 Hz, each 2H, CH2),
7.42, 8.17 (d, J ) 7.4 Hz, each 2H, ArH), 8.13 (s, 2H, ArH);
MS (negative SIMS, NBA) m/z 1152 (M-), 1151 [(M-H)-];
UV-vis (dichloromethane) λmax (ꢀ) 708 (530), 643 (730), 450
(sh, 3522), 307 (47 739), 264 (92 000), 233 (127 910) nm. Anal.
Calcd for C86H34B2O4: C, 89.60; H, 2.97. Found: C, 89.18; H,
3.08.
65,73-Bis(4,4-d im eth yl-2,6-d ioxa -1-bor a cycloh exa n yl)-
1,2:16,17-b is(m et h a n o[1,2]b en zen om et h a n o)[60]fu ller -
en e (12b). In the reaction of 6, the isolation of main isomer
12b from the regioisomeric mixture was performed by HPLC
(COSMOSIL 5PBB column) eluting with n-hexane/toluene (3:7
(v/v), 0.5 mL/min). An analytical sample of 12b was obtained
as a dark brown solid by recrystallization from dichloromethane/
n-hexane: mp >450 °C; IR (KBr) νmax 2955, 2926, 2880, 1476,
1422 (νBC), 1341, 1320 (νBO), 1293, 1246, 1183, 1123, 664, 525
1
cm-1; H NMR (C2D2Cl4, 90 °C) δ 1.04 (s, 12H, Me), 3.79 (s,
8H, CH2), 3.95, 4.02 (d, J ) 13.3 Hz, each 2H, CH2), 4.23, 4.30
(d, J ) 13.8 Hz, each 2H, CH2), 7.45, 7.87 (d, J ) 7.3 Hz, each
2H, ArH), 7.88 (s, 2H, ArH); 13C NMR (C2D2Cl4, 90 °C) δ 23.12
(CH3), 32.90 (C), 43.99, 45.75 (CH2), 61.93, 65.67 (fullerene-
sp3 C), 73.58 (OCH2), 128.04 (sp2 CH), 130.36 (sp2 C), 134.53,
134.55 (sp2 CH), 134.69, 136.09, 137.22, 137.41, 139.48, 140.91,
141.92, 142.04, 142.84, 142.92, 143.23, 143.40, 145.94, 145.97,
146.23, 146.30, 146.68, 147.04. 147.20, 147.22, 147.54, 148.57,
148.90, 149.85, 150.18, 150.59, 150.82, 152.47, 154.98 (sp2 C);
MS (negative SIMS, NBA) m/z 1152 (M-); UV-vis (dichlo-
romethane) λmax (ꢀ) 731 (595), 661 (725), 395 (sh, 6500), 261
(98 750), 232 (118 000) nm; CD (dichloromethane) λmax (∆ꢀ)
726 (-23.48), 658 (-11.25), 608 (5.51), 461 (17.48), 442 (-4.40),
399 (10.67), 341 (-7.70), 287 (67.57), 264 (-83.00), 246 (14.61)
nm. Anal. Calcd for C86H34B2O4‚0.75(C6H14): C, 89.28; H, 3.68.
Found: C, 88.91; H, 3.49.
65,73-Bis(4,4-d im eth yl-2,6-d ioxa -1-bor a cycloh exa n yl)-
1,2:16,17-b is(m et h a n o[1,2]b en zen om et h a n o)[60]fu ller -
en e (en t-12b). According to the same procedure used for the
synthesis of 12b, ent-12b was isolated and purified. The
spectral data of ent-12b coincided with those of 12b except
those for CD spectra: CD (dichloromethane) λmax (∆ꢀ) 726
(26.70), 658 (14.45), 608 (-3.76), 461 (-18.90), 442 (4.51), 399
(-11.83), 341 (7.50), 287 (-73.23), 264 (87.07), 246 (-16.35)
nm.
64,72-Dih yd r oxy-1,2:18,36-b is(m et h a n o[1,2]b en zen o-
m eth a n o)[60]fu ller en e (13c). According to the same proce-
dure used for the synthesis of 13a , 13c7 was prepared from
12c. An analytical sample was obtained as a reddish brown
solid by silica gel column chromatography eluting with toluene/
64,72-Bis(4,4-d im eth yl-2,6-d ioxa -1-bor a cycloh exa n yl)-
1,2:18,36-b is(m et h a n o[1,2]b en zen om et h a n o)[60]fu ller -
en e (12c). According to the same procedure used for the
synthesis of 12b, 12c was isolated and purified: reddish brown
solid; mp >450 °C; IR (KBr) νmax 2957, 2928, 2880, 1476, 1422
1
ethyl acetate (19:1 (v/v)): mp >450 °C; H NMR (DMSO-d6,
120 °C) δ 3.81-3.99 (br m, 2H, CH2), 4.10-4.29 (br m, 6H,
CH2), 6.82, 6.85 (dd, J ) 2.4, 8.4 Hz, each 1H, ArH), 6.91, 7.00
(d, J ) 2.4 Hz, each 1H, ArH), 7.34, 7.36 (d, J ) 8.4 Hz, each
1H, ArH), 8.93, 9.00 (s, each 1H, OH).
(νBC), 1341, 1320 (νBO), 1293, 1246, 1181, 1121, 666, 527 cm-1
;
1H NMR (C2D2Cl4, 90 °C) δ 1.02, 1.06 (s, each 6H, Me), 3.76,
3.81 (s, each 4H, OCH2), 3.80-4.31 (m, 8H, CH2), 7.47 (d, J )
7.4 Hz, 1H, ArH), 7.48 (d, J ) 7.4 Hz, 1H, ArH), 7.82, 7.85 (s,
each 1H, ArH), 7.89 (d, J ) 7.4 Hz, 2H, ArH); 13C NMR (C2D2-
Cl4, 90 °C) δ 22.00, 22.02 (CH3), 31.78, 31.81 (C), 44.95, 45.04,
45.23, 45.26 (CH2), 64.43, 64.52, 64.81, 64.92 (fullerene-sp3 C),
72.46, 72.51 (OCH2), 126.90, 126.96, 132.98, 133.03, 133.52,
133.58 (sp2 CH), 134.84, 135.75, 136.37, 136.41, 136.74, 136.77,
137.12, 137.16, 137.50, 137.54, 140.29, 140.54, 140.75, 141.09,
141.64, 141.86, 141.90, 142.46, 142.99, 143.84, 144.53, 144.69,
144.77, 145.01, 145.17, 145.91, 145.93, 146.18, 146.21, 146.26,
Ack n ow led gm en t . We thank Dr. Tony D. J ames
(University of Birmingham) for helpful discussion on the
saccharide-boronic acid interaction, Dr. Masahito Sano
(Chemotransfiguration Project, J ST) for writing the
manuscript, and Miss Ritsuko Iguchi (Chemotransfigu-
ration Project, J ST) for the measurement of mass
spectrometry.
J O9822277