Molecular structure of lathyritol, a galactosylbornesitol from Lathyrus odoratus seeds, by NMR

Article abstract of DOI:10.1016/j.carres.2005.01.013

The molecular structure of galactosyl-d-(-)-bornesitol, a novel compound isolated from sweet pea seeds, was determined to be α-d-galactopyranosyl- (1→3)-1-O-methyl-1d-myo-inositol by 1D and 2D NMR spectroscopy and is assigned the trivial name lathyritol.

Full text of DOI:10.1016/j.carres.2005.01.013

Carbohydrate
RESEARCH
Carbohydrate Research 340 (2005) 1441–1446
Note
Molecular structure of lathyritol, a galactosylbornesitol from
Lathyrus odoratus seeds, by NMR
Ralph L. Obendorf,a,* Christine E. McInnis,a Marcin Horbowicz,b
d
Ivan Keresztesc and LesławB. Lahuta
aSeed Biology, Department of Crop and Soil Sciences, Cornell University, Ithaca, NY 14853-1901, USA
bResearch Institute of Vegetable Crops, Skierniewice, Poland
cNMR Facility, Department of Chemistry and Chemical Biology, Cornell University, Ithaca, NY 14853, USA
dDepartment of Plant Physiology and Biotechnology, University of Warmia and Mazury, Oczapowskiego 1A, 10-957 Olsztyn, Poland
Received 12 December 2004; accepted 6 January 2005
Available online 19 February 2005
Abstract—The molecular structure of galactosyl-D-(ꢀ)-bornesitol, a novel compound isolated from sweet pea seeds, was determined
to be a-D-galactopyranosyl-(1!3)-1-O-methyl-1D-myo-inositol by 1D and 2D NMR spectroscopy and is assigned the trivial name
lathyritol.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Keywords: Sweet pea seed; Galactosyl-D-(ꢀ)-bornesitol; Galactosylcyclitol; NMR
OH
4'
6'
Seeds of many species accumulate galactosylcyclitols
as part of their normal seed maturation storage
compounds.1–4 Structural characterization is the first
step to understand the metabolism and physiological
role of galactosylcyclitols in seeds. Sweet pea (Lathyrus
odoratus L.) seeds accumulate several soluble carbohy-
drates including the novel compound lathyritol (galac-
tosyl-D-(ꢀ)-bornesitol; a-D-galactopyranosyl-(1!3)-1-
O-methyl-1D-myo-inositol) (1) and free D-(ꢀ)-bornesitol
(1D-1-O-methyl-myo-inositol) (2). The bornesitol moiety
of galactosylbornesitol and free bornesitol from sweet
pea seeds were confirmed to be the D configuration, as
determined by their NMR spectra and optical rotation.
The identification of free L-bornesitol in sweet pea flow-
ers was probably incorrectly labeled.5 Herein we report
the molecular structure and absolute configuration of
1, a novel compound.
CH2OH
O
H
H
5'
H
H
HO
H
H
5
2'
3'
1'
OH
3
HO
H
4
OH
O
H
1
H
H
OH
OH
1
2
6
O
OH
H
CH3
H
H
5
3
HO
4
HO
OH
H
H
H
2
1
2
6
O
OH
H
CH3
chromatogram peaks corresponding to 1 and 2 in seed
extracts, in selected fractions during purification, and
in the analysis of hydrolysis products of 1. Purified 1
was white powder and 2 was white crystalline solid. Acid
hydrolysis of 1 produced D-galactose and 2 in a 1:1 mole
ratio. Compound 1 was hydrolyzed by a-D-galactosi-
dase, but not by b-D-galactosidase, demonstrating an
a-anomeric linkage for the D-galactopyranosyl residue.
Analysis of NMR spectra determined that the linkage
Analysis by high-resolution gas chromatography of
trimethylsilyl derivatives allowed the identification of
*
Corresponding author. Tel.: +1 607 227 9313; fax: +1 607 255 2644;
e-mail: rlo1@cornell.edu
0008-6215/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2005.01.013

1442
R. L. Obendorf et al. / Carbohydrate Research 340 (2005) 1441–1446
must be between carbon C-10 on the galactopyranosyl
ring and carbon C-3 on the bornesitol ring.
hydrogen (dH 5.00) were useful for assigning resonances
to the D-galactopyranosyl ring. The C-60 on the galacto-
pyranosyl ring was shielded the most and identified as
the resonance at dC 61.45, the second furthest upfield.
The two nearly coincident H-60 hydrogens (dH 3.58
and 3.60) were characteristic of distal galactopyranosyl
rings.7–9 The remaining resonances could be assigned
by coupling constants and 2D COSY and NOESY
experiments. The hydrogen at 3.45 ppm was attached
to the downfield carbon (dC 75.81) with the resonance
shifted +4.3 ppm downfield in comparison to the corre-
sponding resonance for free D-(ꢀ)-bornesitol 2 (Table
1), thereby identifying the point of galactose substitu-
tion on the bornesitol ring.7–9 HMBC experiments
showed clear interactions between the anomeric hydro-
gen (d 5.00; H-10) on the galactopyranosyl ring and
the carbon at d 75.81 (C-3) on the cyclitol ring, confirm-
ing the point of linkage. The small coupling constant
13C NMR spectra identified seven carbon resonances
and 1H NMR spectra identified nine hydrogen reso-
nances for 2 (Table 1). The 1H NMR spectra of 2
showed a characteristic intense singlet at dH 3.28 corre-
sponding to the O-methyl group. A highly deshielded
triplet at dH 4.16 corresponded to the only equatorial
hydrogen with the smallest coupling constants (J1,2
;
J2,3) characteristic of an Hax–Heq–Hax system, and was
assigned H-2 in accordance with published spectra for
O-methyl ethers of myo-inositol.6,7 The remaining
hydrogens exhibited large coupling constants character-
istic of Hax–Hax–Hax transaxial systems on the cyclitol
ring. The furthest downfield resonance in the 13C
NMR spectra at dC 80.80 (C-1) indicated the site of
methylation.1,6,7 By contrast, the methyl carbon was
the furthest upfield, corresponding to dC 57.10. Assign-
ments for the five remaining carbon (d 68–75) and
hydrogen resonances on the cyclitol ring were
determined by 2D COSY and HSQC experiments (Table
1).
13C NMR spectra identified 13 carbon resonances and
1H NMR spectra identified 16 hydrogen resonances for
1 (Table 1). The 1H NMR spectra for 1 showed the char-
acteristic intense singlet at dH 3.29 and the furthest up-
field carbon (dC 57.22) resonance in the 13C NMR
spectrum corresponding to the O-methyl group on the
cyclitol ring. A highly deshielded triplet at dH 4.38 sug-
gested that H-2 was equatorial. The small coupling con-
stants corresponding to J1,2 and J2,3 confirmed that H-2
was indeed equatorial on 1 and 2. The characteristically
downfield anomeric carbon (dC 95.72) and anomeric
0
0
J1 ,2 3.9 verified the a-anomeric linkage.
2D NOSEY experiments showed interactions between
the anomeric hydrogen (dH 5.00; H-10) of the D-galacto-
pyranosyl ring and hydrogens at dH 3.45 (H-3) and dH
4.38 (H-2) of the cyclitol ring, consistent with the as-
signed linkage. In addition, the observed correlation be-
tween H-50 (dH 4.03) of the D-galactopyranosyl ring and
H-3 (dH 3.45) of the cyclitol ring indicated close contact
between those hydrogens. Molecular mechanics calcula-
tions predicted considerably shorter interatomic dis-
0
˚
tance between H-5 and H-3 for D-bornesitol (3.08 A)
˚
than the L isomer (4.26 A), which suggested that the
cyclitol ring is in the D configuration.
Coupling constants between adjacent hydrogens were
calculated from the 1D 1H signals. The derived coupling
Table 1. 1H and 13C NMR chemical shifts and proton–proton coupling constants of lathyritol (1) and D-(ꢀ)-bornesitol (2) in D2O at 25 ꢁC
Position (no)
Chemical shifts (ppm)
Coupling constants (J, Hz)
1
2
No
1
2
dH
dC
dH
dC
Cyclitol ring
1
3.05
4.38
3.45
3.60
3.18
3.54
3.29
80.70
63.98
75.81
71.30
74.68
71.98
57.22
3.06
4.16
3.35
3.47
3.13
3.49
3.28
80.80
68.00
71.50
72.70
74.80
72.00
57.10
J1,2
J2,3
J3,4
J4,5
J5,6
J6,1
2.7
2.7
2.8
2.9
2
3
10.0
9.4
10.0
9.4
4
5
9.4
10.0
9.4
10.0
6
OCH3
Galactopyranosyl ring
10
0
J1 ,2
0
0
0
0
0
5.00
3.70
3.80
3.86
4.03
3.58
3.60
95.72
68.74
69.79
69.63
71.33
61.45
3.9
10.4
3.3
20
0
J2 ,3
30
0
J3 ,4
40
1.3
7.7a
4.8a
0
J4 ,5
50
J5 ,6 a
0
60a
60b
0
J5 ,6 b
0
ꢀ11.6a
0
J6 a,6 b
0
a Determined by simulation.

R. L. Obendorf et al. / Carbohydrate Research 340 (2005) 1441–1446
1443
constants (Table 1) allowed confirmation of the assign-
ments made for cyclic bornesitols. The coupling con-
stants for 1 indicated that the cyclitol ring is in the
same configuration as 2. The presence of small coupling
constants (J1,2 and J2,3) for H-2 on 1 and 2 indicated the
presence of the Hax–Heq–Hax system, while the large
coupling constants for the remaining hydrogens indi-
cated they were all Hax–Hax–Hax transaxial systems
+96.5) for 1 is expected compared to those for galactinol
20
D
(a-D-galactopyranosyl-(1!1)-1L-myo-inositol; ½aꢁ +135.6)11
and galactosylononitol (a-D-galactopyranosyl-(1!3)-
20
1D-4-O-methyl-myo-inositol; ½aꢁ +129.6).7
D
The GC mass spectrum of Me3Si ethers of 2 (Fig. 1,
top) was characteristic for bornesitol.12 Mass spectra
of Me3Si ethers O-methyl inositols exhibit several char-
acteristic ions12 (present in Fig. 1) that are different than
their corresponding inositols. The major fragment ions
(m/z abundance relative to m/z 217) for 2 were m/z 89
(21), 159 (42), 217 (100), 247 (21), 260 (26), 374 (7.8),
375 (7.9), 432 (5.9), 433 (10), and 449 (4.4), and the
abundance ratios, m/z to m/z, were 260/247 (1.24), 374/
375 (0.99), and 432/433 (0.59). The GC–MS spectrum
of Me3Si ethers of 1 (Fig. 1, bottom) exhibited high
abundance ions at m/z 191, 204, 217, and 361 from
characteristic of the myo-inositol ring in bornesitol.
23
D
The optical rotation of 1 was ½aꢁ +96.5 (c 0.76, H2O)
23
D
and of 2 was ½aꢁ ꢀ28.4 (c 1.45, H2O). Compound 2 was
assigned the D configuration because of levorotation.10
This configuration was also inferred from the NMR
data for 1 (see above). The optical rotation value
23
30
(½aꢁ ꢀ28.4) for 2 is similar to published values (½aꢁ
D
D
18
20
23
D
ꢀ26.4; ½aꢁ ꢀ32.05; ½aꢁ ꢀ32.6),4,7,8 and the value (½aꢁ
D
D
Figure 1. Mass spectrum of Me3Si-D-(ꢀ)-bornesitol (2) (top) and of Me3Si-a-D-galactopyranosyl-(1!3)-1D-1-O-methyl-myo-inositol (1) (bottom).

1444
R. L. Obendorf et al. / Carbohydrate Research 340 (2005) 1441–1446
fragmentation of the Me3Si ethers of the galactopyrano-
syl ring, while ions at m/z 159, 247, 375, and 449 resulted
from fragmentation of the Me3Si ethers of the O-methyl
cyclitol ring.
Estimated molecular weights determined from nega-
tive-ion electrospray ionization mass spectra (ESI MS)
were 356 for 1 and 194 for 2.
45 min, and analyzed2,8,9 on a Hewlett–Packard 6890
GC (Agilent Technologies, Palo Alto, CA, USA)
equipped with a flame-ionization detector, split-mode
injector (1:50), and a HP-1MS capillary column
(15 m · 0.25 mm i.d., 0.25 lm film thickness). The GC
instrument was operated with a programmed initial tem-
perature of 150 ꢁC, adjusted to 200 ꢁC at 3 ꢁC minꢀ1
,
Compound 1 is a-D-galactopyranosyl-(1!3)-1-O-
methyl-1D-myo-inositol and assigned the trivial name
of lathyritol, a novel galactosyl D-bornesitol that accu-
mulates in Lathyrus odoratus L. seeds during late seed
maturation and desiccation. Lathyritol accumulation
began at 32 days after pollination (DAP; maximum seed
fresh weight), increased to 1.5 mg gꢀ1 dry weight at 34
DAP (maximum seed dry weight) and to 6.6 mg gꢀ1
dry weight at 38 DAP when seeds had dried to 20%
moisture. Lathyritol represents another methylated
derivative of galactinol, the other compound being
galactosyl ononitol.7
adjusted to 325 ꢁC at 7 ꢁC minꢀ1, and held at 325 ꢁC
for 20 min. The injector port was operated at 335 ꢁC
and the detector at 350 ꢁC. The carrier gas was nitrogen
at 2.5 mL minꢀ1. Typical retention times for trimethylsi-
lyl derivatives were 23.8 min for 1 and 6.6 min for 2.
Fractions 14–17 containing 2 (eluted with EtOH–
water, 2:98, v/v) were freeze dried, pooled, loaded onto
a carbon–Celite column (25 mm · 900 mm), and eluted
with 2 L of water followed by 2 L of EtOH–water
(2:98, v/v). Fractions containing 2 (90% pure) were
pooled, freeze dried, and dissolved in hot (80 ꢁC)
MeOH. Crystals collected after evaporative cooling
yielded 50 mg of 2 (99% pure). A second crystallization
from MeOH did not increase purity.
1. Experimental
Compound 1 eluted from the primary column with
EtOH–water (4:96, 5:95, 6:94, v/v). Fractions 23–28 con-
taining 1 were freeze dried, and dry residues were pooled
(1785 mg). This pooled sample containing 1 was dis-
solved in water, loaded onto a Dowex 50Wx8 cation ex-
change column (25 mm · 600 mm), and eluted with
water to remove contaminants. Fractions enriched in 1
were pooled and freeze dried. A sample containing 1
(266 mg, 85% pure) was loaded onto a carbon–Celite
column (25 mm · 900 mm) and eluted stepwise with
water (2 L) followed by EtOH–water (4 L of 3:97, 2 L
of 3.5:96.5, 1 L of 4:96, v/v). Fractions 281–304 (eluted
with EtOH–water, 3.5:96.5, v/v) yielded 95 mg of 1
(97% pure) that was taken for analysis by NMR, GC–
MS, LC–MS, acid, and enzymatic hydrolysis, and opti-
cal rotation.
1.1. Extraction and purification
Sweet pea (Lathyrus odoratus L.) seeds were purchased
from Outsidepride.com, Inc. (Salem, OR, USA). Soluble
carbohydrates in embryo tissues (cotyledons; axis) ex-
pressed as mg gꢀ1 dry weight SE for three replicate
samples were: D-(ꢀ)-bornesitol (2) (2.0 0.2;
3.6 1.1), lathyritol (1) (2.0 0.2; 6.4 1.6), myo-inosi-
tol (0.2 0.0; 0.9 0.2), galactinol (0.6 0.0; 2.1 0.4),
sucrose (16.0 0.2; 39.0 6.5), raffinose (3.8 0.2;
10.1 1.7), stachyose (20.9 1.0; 75.3 14.0), and ver-
bascose (1.4 0.1; 6.1 0.8). Approximately 300 g of
embryos from de-coated seeds were pulverized to a fine
powder in a coffee grinder, extracted with 2.5 L of
EtOH–water (1:1, v/v) at 80 ꢁC for 30 min, and cooled
to 22 ꢁC. Solids were removed by filtration through
cheese cloth and centrifugation (25 samples). Clear
supernatants were diluted 1:4 (v/v) with water, frozen
at ꢀ80 ꢁC, and freeze dried. Dry residues were dissolved
in water, and 1 and 2 were purified by preparative chro-
matography on a stationary phase of carbon (Darco
G60; J. T. Baker, Phillipsburg, NJ) and Celite 545-AW
(Supelco, Bellefonte, PA), 1:1 (w/w).13 Freeze-dried sam-
ples were dissolved in minimal water for loading onto a
water slurry packed column (100 mm · 200 mm), and
soluble carbohydrates were eluted at 4 ꢁC with stepwise
increments of EtOH–water, v/v, (4 L of water; 4 L of
2:98; 4 L of 4:96; 2 L of 5:95; 2 L of 6:94) and collected
in 500 mL fractions. Fractions were sampled (200 lL
fraction + 25 lg of phenyl a-D-glucoside as internal
standard), dried in a stream of nitrogen gas and stored
overnight above P2O5, derivatized with 100 lL of N-
trimethylsilylimidazole–pyridine (1:1, v/v) at 80 ꢁC for
1.2. NMR analysis
Purified samples (60 mg of 1, 100 mg of 2) were dis-
solved in 1 or 2 mL of 99.96% D2O (Cambridge Isotope
Labs, Andover, MA) for NMR analysis. Hydrogen and
2D spectra were recorded on a Varian INOVA spec-
trometer operating at 499.92 and 125.72 MHz for
hydrogen and carbon observation, respectively, using a
Varian inverse triple resonance probe head equipped
with a z-axis gradient coil. Hydrogen spectra were refer-
enced relative to residual HOD at 4.63 ppm as an inter-
nal standard. Carbon spectra were recorded on a Varian
INOVA spectrometer operating at 100.54 MHz for
carbon observation using a Varian auto-switchable
broadband probe and referenced relative to dioxane at
67.4 ppm as an external standard.
2D Spectra for 2 were collected with sweep widths of
1 kHz in the hydrogen and 5 kHz in the carbon dimen-

R. L. Obendorf et al. / Carbohydrate Research 340 (2005) 1441–1446
1445
sion. For 1, sweep widths of 1.2 and 6.3 kHz were used.
1.4. Optical rotation analysis
Gradient-enhanced DQ-COSY spectra were acquired in
phase-sensitive mode with 64 or 128 complex points in
the indirectly detected dimension. One transient of 512
or 1k data points were collected for each increment.
The data were zero filled to 2k complex data points
and multiplied with 45ꢁ shifted squared sinusoidal win-
dowfunctions in both dimensions prior to Fourier
transform. NOESY spectra were acquired in phase-sen-
sitive mode with 200 complex points in the indirectly
detected dimension and 500 ms mixing time. Sixteen
transients of 512 or 1k data points were collected for
each increment. The data were zero filled to 2k complex
data points and multiplied with Gaussian window func-
tions in both dimensions prior to Fourier transforma-
tion. Gradient enhanced HSQC spectra were acquired
in phase sensitive mode with 128 complex points in
the indirectly detected dimension and 3.57 ms evolution
time. Two or four transients of 512 or 1k data points
were collected for each increment. The data were zero
filled to 2k complex data points and multiplied with
Gaussian window functions in both dimensions prior
to Fourier transformation. Gradient enhanced HMBC
spectra were collected using a band selective constant
time sequence14 in absolute value mode with 400 points
in the indirectly detected dimension and 62.5 ms evolu-
tion time for multiple bond couplings. Two or four
transients of 512 or 1k data points were collected for
each increment. Selective inversion of the carbon reso-
nances was accomplished with Q3 Gaussian cascade
pulses. The data were zero filled to 2k complex data
points and multiplied with shifted sinusoidal window
functions in both dimensions prior to Fourier
transformation.
Optical rotation values for 1 (90 mg) and 2 (45 mg) dis-
solved in water were assayed with a Autopol III polarim-
eter (Rudolph Research Analytical, Flanders, NJ, USA)
at 589 lm and 23 ꢁC using a 100 mm path length cell.
1.5. Acid and enzymic hydrolysis
Compound 1 (200 lg) was hydrolyzed with 1 M
CF3CO2H for 16 h at 80 ꢁC and evaporated to dryness.
The mole ratio of D-galactose to 2 was calculated after
GC analysis of the trimethylsilylated products. Com-
pound 1 (200 lg) was incubated with 1.25 units of de-
salted
green-coffee-bean
a-D-galactosidase
(EC
3.2.1.22) in 200 lL of water at 22 ꢁC or with 0.5 units
of bovine liver b-D-galactosidase (EC 3.2.1.23) in
200 lL of water at 37 ꢁC for 24 h. Enzyme protein was
removed by filtration (10,000 MW cut-off filter), the
sample was dried, and products assayed by GC analysis
of trimethylsilyl derivatives. Raffinose and lactose
(100 lg) were hydrolyzed by a- or b-D-galactosidase,
respectively, confirming that both enzymes were active.
Acknowledgements
We thank F. A. Loewus for providing authentic L-(+)-
bornesitol, A. M. Condo for collecting the NMR spec-
tra, W. Wiczkowski for collecting LC–MS spectra, and
B. A. Lewis for assistance with the optical rotation mea-
surements. This research was conducted as part of Mul-
tistate Projects W-168 (NY-C 125-423) and W-1168
(NY-C 125-802) and was supported in part by a grant
from the PresidentÕs Council of Cornell Women to
C.E.M., a fellowship from The Kosciuszko Foundation
to M.H., and Cornell University Agricultural Experi-
ment Station federal formula funds received from Coop-
erative State Research, Education and Extension
Service, US Department Agriculture. Any opinions,
findings, conclusions, or recommendations expressed in
this publication are those of the authors and do not nec-
essarily reflect the viewof the US Department of
Agriculture.
1.3. GC–MS and LC–MS analysis
GC–MS analyses of trimethylsilyl-derivatives of 1 and 2
(prepared as above) were with a 5971A Mass Selective
Detector set at SCAN mode (ion source 70 eV) on a
Hewlett–Packard 5890 GC. Separation was on an HP-
1 capillary column (12 m · 0.2 mm i.d., 0.33 lm film
thickness) using helium as carrier gas at 50 cm/s flow
rate. Injection port temperature was 250 ꢁC, split 1:10.
Detector temperature was 280 ꢁC. Column oven temper-
ature was programmed at 150 ꢁC for 2 min and then ad-
justed 8 ꢁC/min to 300 ꢁC.
References
Electrospray ionization mass spectra (ESI MS) in neg-
ative mode were performed on 1 and 2 using a Shimadzu
HPLC-MS—QP 8000a system (Shimadzu, Kyoto,
Japan). The curve desolvation line (CDL) temperature
was 200 ꢁC with ꢀ70 V CDL voltage, ꢀ5.0 kV probe
voltage, ꢀ45 V defragmentation voltage, 2.8 mL minꢀ1
nebulizer gas flow, 0.2 mL minꢀ1 column flowrate,
and a 5 lL injection loop. Samples were eluted with ace-
tonitrile–formic acid–water, 40:0.25:59.75 (v/v/v).
1. Schweizer, T. F.; Horman, I. Carbohydr. Res. 1981, 95,
61–71.
2. Horbowicz, M.; Obendorf, R. L. Seed Sci. Res. 1994, 4,
385–405.
3. Obendorf, R. L. Seed Sci. Res. 1997, 7, 63–74.
4. Szczecinski, P.; Gryff-Keller, A.; Horbowicz, M.; Lahuta,
L. B. J. Agric. Food Chem. 2000, 48, 2717–2720.
5. Ichimura, K.; Kohata, K.; Mukasa, Y.; Yamaguchi, Y.;
Goto, R.; Suto, K. Biosci. Biotechnol. Biochem. 1999, 63,
189–191.

1446
R. L. Obendorf et al. / Carbohydrate Research 340 (2005) 1441–1446
6. Angyal, S. J.; Odier, L. Carbohydr. Res. 1983, 123, 23–
29.
10. Bien, S.; Ginsburg, D. J. Chem. Soc. 1958 (Sep), 3189–3194.
11. Brown, R. J.; Serro, R. F. J. Am. Chem. Soc. 1953, 75,
7. Richter, A.; Peterbauer, T.; Brereton, I. J. Nat. Prod.
1997, 60, 749–751.
1040–1042.
12. Binder, R. G.; Haddon, W. F. Carbohydr. Res. 1984, 129,
8. Obendorf, R. L.; Steadman, K. J.; Fuller, D. J.; Hor-
bowicz, M.; Lewis, B. A. Carbohydr. Res. 2000, 328, 623–
627.
9. Steadman, K. J.; Fuller, D. J.; Obendorf, R. L. Carbohydr.
Res. 2001, 331, 19–25.
21–32.
13. Whistler, R. L.; Durso, D. F. J. Am. Chem. Soc. 1950, 72,
677–679.
14. Claridge, T. D. W.; Perez-Victoria, I. Org. Biomol. Chem.
2003, 1, 3632–3634.
´