Synthesis, analysis and rearrangement of novel unnatural glucosinolates

Article abstract of DOI:10.1016/S0008-6215(00)00308-6

As part of a structure activity study to examine the interaction of glucosinolates with leaf surfaces, a number of glucosinolates were synthesised bearing novel side chain functionalities. These included 7-carboxyheptyl, heptyl, and naphthyl side chains. For the carboxyheptyl glucosinolate, a novel intramolecular rearrangement reaction was observed during the final deprotection step, which generated an ester attached to the C-3 of glucose. Studies by ¹H NMR spectroscopy showed that the hydrophobic side chain associated with one face of the glucose ring and it was proposed that this was the driving force for the rearrangement. Similar hydrophobic interactions were also observed between the heptyl and naphthyl side chains and the glucose.

Full text of DOI:10.1016/S0008-6215(00)00308-6

www.elsevier.nl/locate/carres
Carbohydrate Research 330 (2001) 295–307
Synthesis, analysis and rearrangement of novel unnatural
glucosinolates
Nicola E. Davidson, Trevor J. Rutherford1, Nigel P. Botting*
School of Chemistry, Uni6ersity of St. Andrews, St. Andrews, Fife KY16 9ST, UK
Received 9 August 2000; received in revised form 19 October 2000; accepted 21 November 2000
Abstract
As part of a structure activity study to examine the interaction of glucosinolates with leaf surfaces, a number of
glucosinolates were synthesised bearing novel side chain functionalities. These included 7-carboxyheptyl, heptyl, and
naphthyl side chains. For the carboxyheptyl glucosinolate, a novel intramolecular rearrangement reaction was
1
observed during the final deprotection step, which generated an ester attached to the C-3 of glucose. Studies by H
NMR spectroscopy showed that the hydrophobic side chain associated with one face of the glucose ring and it was
proposed that this was the driving force for the rearrangement. Similar hydrophobic interactions were also observed
between the heptyl and naphthyl side chains and the glucose. © 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Glucosinolates; Cruciferae; brassicas
linked with consumption of brassicas by hu-
1. Introduction
mans.3 The anti-carcinogenic effects are
thought to arise from isothiocyanates which
Glucosinolates are a group of thioglycosides
that occur in all members of the Cruciferae,
including the brassica crops such as cabbage,
are glucosinolate breakdown products.4 In
plants, the breakdown products show anti-
fungal and anti-bacterial effects and deter at-
tack by non-adapted herbivores.5
Brussels sprouts and oilseed rape.1 Glucosino-
lates contain a common structure 5 and over
100 different examples have been isolated and
characterised,1 with a variety of substituents in
the side chain R, including allyl (sinigrin),
benzyl, indolyl and 4-hydroxybenzyl. Glucosi-
nolates possess a range of important biologi-
cal activities. Anti-nutritional and potentially
toxic effects have been observed following
consumption of high concentrations by mam-
mals,2 whereas reduced risk of cancer has been
However, in some cases glucosinolates play
a role in host–plant recognition5 and leaf
surface glucosinolates can act as ovi-position
(egg-laying) stimulants for brassica adapted
insects, e.g., cabbage and turnip root flies
(Delia radicum and Delia floralis).6,7 A combi-
nation of behavioural bioassays and electro-
physiology have been used to test 7–11
naturally occurring glucosinolates, essentially
the complete set of available compounds, in
comparative studies on the closely related
cabbage8 and turnip9 root flies. These studies
demonstrated that there were significant corre-
lations between overall length of the glucosi-
nolate side chain and the biological activity,
* Corresponding author. Tel.: +44-1334-463856; fax: +
44-1334-433808.
E-mail address: npb@st-andrews.ac.uk (N.P. Botting).
1 Present address: School of Chemistry, University of Birm-
ingham, Edgbaston, Birmingham B15 2TT, UK.
0008-6215/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(00)00308-6

296
N.E. Da6idson et al. / Carbohydrate Research 330 (2001) 295–307
measured by behavioural or electrophysiologi-
cal assays. It was also found that in model
systems, the wax coating of the leaf is vital if
any stimulatory activity is to be obtained.10
Interaction of the glucosinolate side chain
with the epicuticular waxes of the leaf surface
may thus be essential in presenting the glu-
cosinolate in the correct orientation for recog-
nition by the insect. Similar observations have
been reported for ovi-position by the dia-
mondback moth (Plutella xyostella) in re-
sponse to cabbage and sinigrin homogenates,11
where waxes were shown to significantly in-
crease the stimulatory effect of one glucosino-
late, sinigrin, but were ineffective on their
own. However, the structure activity study
employed a diverse range of side chains in-
cluding some aromatic and some indolyl, and
it was thus decided to prepare a series of
unnatural glucosinolates with simple alkyl side
chains of varying lengths to re-examine this
phenomenon in a more rational manner. As
part of the series, a side chain with a polar end
group was also required to determine whether
this would disrupt interactions with the leaf
surface and a glucosinolate with a 7-carboxy-
heptyl side chain was chosen as a suitable
synthetic target. The final target was the naph-
thylmethyl glucosinolate which would act as a
mimic for the natural indolyl glucosinolates
but lack the nitrogen atom and thus any pos-
sible recognition due to this functionality.
droximate 3. This is followed by sulfation with
chlorosulfonic acid in pyridine and deprotec-
tion using sodium methoxide in methanol
(Scheme 1).12 The oximyl chloride can be pre-
pared from either an aldehyde12 or a nitro-
alkane.13 A large number of naturally occur-
ring glucosinolates have been prepared using
this methodology, in particular by Rollin and
co-workers.14
For the alkyl glucosinolates the commer-
cially available aldehydes were converted to
the oximes, chlorinated using N-chlorosuccin-
imide in pyridine and the oximyl chlorides
used without further purification (Scheme 1).
Glucosinolates were prepared with propyl, bu-
tyl, pentyl, heptyl and nonyl side chains.
The naphthylmethyl glucosinolate 6 re-
quired 1-naphthylacetaldehyde as starting ma-
terial. This proved to be difficult to synthesise.
Reduction of 1-naphthylacetonitrile using
Raney nickel failed as did the oxidation of
1-naphthaleneethanol using chromic acid, lead
dioxide–sulfuric acid and lead tetraacetate.
The latter instead gave high yields of 1-naph-
thaldehyde, the over oxidised product.15 Oxi-
dation using both milder Swern conditions
and pyridinium chlorochromate (PCC) was
attempted, but in all cases, a mixture was
obtained of the desired acetaldehyde, 1-naph-
thaldehyde and starting material. Optimisa-
tion failed to give higher than 69% yield of the
1-naphthylacetaldehyde, which was difficult to
separate from the by-products. The problem
was finally solved by using an alternative re-
ductive route. 1-Naphthylacetic acid was con-
verted to its ethyl ester and then reduced to
1-naphthylacetaldehyde, in 94% yield, using
di-iso-butyl aluminium hydride (DIBAL-H).
Although this route required two steps, both
were efficient and gave no undesirable side
2. Results and discussion
The most convenient synthetic route to glu-
cosinolates 5 involves the coupling of an ox-
imyl chloride derivative 2 to tetraacetyl
thioglucose 1 to give the protected thiohy-
Scheme 1.

N.E. Da6idson et al. / Carbohydrate Research 330 (2001) 295–307
297
terification took place to give the methyl ester
1
10b. Analysis by H and 13C NMR spec-
troscopy (in D2O) showed no signals for O-
acetyl groups (ꢀ1.9–2.1 ppm) or the ethyl
ester (ꢀ4.0–4.3 and 1.25 ppm), but a methyl
signal was observed at 3.61 ppm, with a corre-
Scheme 2.
13
sponding C resonance at 55.1 ppm. Negative
products. The aldehyde was then used without
purification for the preparation of 1-naphthy-
lacetaldoxime in 52% yield. Chlorination of
the oxime did not appear to go to completion
ion ESMS confirmed that the ethyl ester had
undergone transesterification, and had been
completely converted to the methyl ester 10b.
Further hydrolysis was thus required and so
the methyl ester 10b was treated with
methanolic potassium hydroxide at room tem-
perature for 24 h and the product purified by
C18 reverse phase column chromatography.
1H NMR at 500 MHz confirmed the presence
of three glucose containing species in the aq
fractions, one of which was shown to be a free
glucoside. The peak for the anomeric proton
resonance at 4.67 ppm could not be integrated
due to overlap with the HOD signal, but by
recording the spectra at 55 °C, the HOD sig-
nal was moved to 4.47 ppm, and the propor-
tion of the two remaining species (2.5:1.0) was
estimated from the H-1 peak integrals ratios.
Resonance assignments for the major product
were consistent with the desired acidic glucosi-
nolate 11 (Table 1).
1
by H NMR, but the mixture of oxime and
oximyl chloride was used for the coupling step
without purification. This gave the thiohy-
droximate in 87% yield after recrystallisation.
Sulfation and deprotection were then carried
out as before to give the final glucosinolate 6,
but in this case in fairly poor 33% overall
yield.
Synthesis of (7-carboxyheptyl)glucosinolate
began with 8-bromooctanoic acid 7. This was
converted to the ethyl ester using thionyl bro-
mide and ethanol. It was necessary to use
thionyl bromide as the normal thionyl chlo-
ride method which gave a mixture of the
8-bromo and 8-chloro derivatives. Oxidation
to the aldehyde, ethyl 8-oxooctanoate 8, was
accomplished in 61% yield by the iodide
modified Kornblum oxidation procedure re-
ported by Bauer and Macomber.16 Reaction
with hydroxylamine hydrochloride gave the
oxime, which was treated with N-chlorosuc-
cinimide to give ethyl 8-chloro-8-hydroxyimi-
nooctanoate 9 in 86% yield (Scheme 2).
1H–1H spin-coupling constants for each of
the three glycoside moieties in the product
mixture are shown in Table 2. Large (9–10
3
Hz) JHH values indicated a diaxial relation-
ship between vicinal ring protons, consistent
with the b-glucose configuration, and 4C1
chair conformation. JH-5,H-6 and JH-5,H-6 spin-
coupling constants indicated that the favoured
orientation of the hydroxymethyl group was
the gauche–gauche rotamer.17
Further NMR studies were undertaken to
characterise the rearrangement product.
Haverkamp et al.18 previously investigated
The coupling and sulfation were carried out
as before to give 10a in 72% yield (Scheme 3).
However, during deprotection of 10a transes-
1
changes in H chemical shifts for glycosides
Scheme 3.

298
N.E. Da6idson et al. / Carbohydrate Research 330 (2001) 295–307
Table 1
1H and 13C chemical shift assignments (ppm) for the glucose moiety of glucosinolates and related compounds in D2O
Compound
H-1
H-2
H-3
H-4
H-5
H-6
H6%
C-1
C-2
C-3
C-4
C-5
C-6
1
16
6
5.00
5.03
4.60
4.67
4.53
3.44
3.49
3.31
3.82
3.04
3.56
3.59
3.11
5.11
3.4
3.45
3.51
3.32
3.62
3.4
3.54
3.57
2.71
3.59
3.4
3.88
3.92
3.48
3.92
3.87
3.71
3.76
3.41
3.76
3.68
84.5
74.6
79.7
71.8
82.8
63.3
84.6
91.6
86.9
74.5
72.9
79.6
83.3
71.3
70.9
82.5
82.8
62.8
63.3
64.0
14 or 15
Glc-1-SH a
a b-1-thioglucose (sodium salt) from Ref. 23.
Table 2
1H–1H spin-coupling constants (Hz) for the glucose moiety of glucosinolates in D2O
Compound
JH-1,H-2
JH-2,H-3
JH-3,H-4
JH-4,H-5
JH-5,H-6
JH-5,H-6%
JH-6,H-6%
1
16
6
9.9
9.8
9.9
8.8
9.0
9.8
2.3
1.6
4.7
5.6
5.3
2.4
12.6
12.6
12.7
a
a
a
8.9
9.2
10.0
a Unresolved due to strong coupling.
upon acetylation, and reported characteristic
a-effects (resonance frequency changes for the
proton attached to the carbon which bears the
O-acyl group) of 1–1.5 ppm downfield, and
b-effects (for protons attached to adjacent car-
bon atoms) of approximately 0.2 ppm
downfield. Chemical shifts for the glucose por-
tion of the unknown product showed
downfield shift changes of 1.55 ppm on H-3
(relative to the major product), and 0.38 and
0.17 ppm on H-2 and H-4, respectively, con-
sistent with O-acylation on C-3. The transes-
terification was specific to the 3-position of the
glucose ring, indicative of an intramolecular
process.
In an HMBC experiment, H-3 showed a
long range correlation (through two or three
bonds) to a quaternary 13C signal at 175.9
ppm. The same carbon was also correlated
with proton resonances at one end of the
aliphatic chain. The chemical shift of this qua-
ternary carbon was upfield of the carbonyl
carbon resonance of the free carboxylic acid
(186.7 ppm). From these data we could infer
that the side chain was connected to the glu-
cose C-3 via an ester linkage. There were no
correlations in the HMBC experiment be-
tween the opposite end of the aliphatic chain
and the anomeric resonances of the glycoside.
However, protons a and a´ı showed correla-
tions to a quaternary 13C resonance at 125.0
ppm, although the functionality of this carbon
was not identified. A signal in the negative ion
electrospray mass spectrum at m/z 348, and a
small IR absorbance at 2252 cm−1 were taken
as tentative evidence of isocyanate formation.
NOESY spectra (with 500 ms mixing time)
showed strong NOEs between H-1 and pro-
tons attached to the first two carbons proxi-
mal to the thiooxime in the major product 11.
Weak NOEs were observed from the H-5 and
the high-field H-6 protons to protons of the
alkyl chain. No significant NOEs were ob-
served connecting the aliphatic protons with
either H-2 or H-4, indicating that the side
chain associates exclusively with the B-face of
the glucose ring. Based on NOE data for
glucobrassicin (a natural glucosinolate bearing
an indole side chain) Prestera et al. also in-
ferred an association between a hydrophobic
side chain and the B face of the glycoside.19
The association, ascribed to an aromatic
stacking interaction, is clearly also observed
here with non-aromatic hydrophobic moieties.
A mechanism can be proposed to explain
the observations as shown in Scheme 4. Firstly
transesterification of 10b takes place first,
which is assisted by the conformation of the
glucosinolate in solution which positions the
alkyl side chain underneath the glucose ring.
This will be in competition with the straight-
forward hydrolysis of the ester 10b to give the

N.E. Da6idson et al. / Carbohydrate Research 330 (2001) 295–307
299
glucosinolate 11. The bicyclic system 12 pro-
duced will be fairly strained and so is hy-
drolysed under the basic conditions employed
to give a O-sulfated hydroxamic acid interme-
diate 13. A Lossen-type rearrangement can
then take place, analogous to that catalysed
by myrosinase in the metabolism of glucosino-
lates, which would give the isocyanate 14. The
isocyanate should be very susceptible to nucle-
ophilic attack and thus very unstable. How-
ever, there is some evidence for its presence in
the IR spectrum of the product mixture and a
molecule of this molecular weight is observed
in the ESMS. As there is a strongly nucle-
ophilic thiol group in close proximity to the
isocyanate, this could result in a rapid in-
tramolecular cyclisation to give 15 and the
product could exist as an equilibrium mixture
of 14 and 15.
the glucose ring (H-1, H-3 and H-5) than for
those on the A-face. NOEs involving the
anomeric proton were unresolved from those
involving the C-8 methylene protons using
conventional 2D NOESY experiments. It was
not clear whether the strong NOEs from the
C-8 protons were masking significant weaker
NOEs from H-1. A one-dimensional NOE
spectrum, selective for H-1 (Fig. 1(b)), was
obtained using the double-selective 1D
TOCSY-NOESY pulse sequence described by
Uhrin and Barlow.20 The first pulse was selec-
tive for band of resonance frequencies encom-
passing protons H-2 through H-5, and
magnetisation was transferred to H-1 by
means of a 50 ms DIPSI-2 spin–lock se-
quence. The second selective pulse, beginning
the NOESY portion of the pulse sequence,
gave clean excitation of H-1 without excita-
tion of the C-8 methylene protons. Similarly,
1D NOE spectra were obtained selectively for
the combined H-2 and H-4 signals, H-3 and
H-5. The resulting spectra (Fig. 1) indicate
that the naphthyl side chain has a loose asso-
ciation with exclusively the B-face of the glu-
cose ring.
NMR studies were also carried out to inves-
tigate the solution conformations of the hep-
tylglucosinolate
5
(R=(CH2)6CH3) and
naphthylmethylglucosinolate 6. The heptylglu-
cosinolate gave a similar set of NOEs connect-
ing the H-1, H-3 and H-5 with the protons
attached to the two carbon atoms proximal to
the thiooximoyl functionality, implying that
the side chain was also positioned under the
glucose ring. For the naphthylmethylglucosi-
nolate aromatic ring currents reduced the pro-
ton chemical shifts of the glucose moiety, as
previously observed for aromatic glucosino-
lates.19 The chemical shift differences were
more pronounced for protons on the B-face of
The accessible conformational space for the
glucosinolates spans a wide range, as there are
several bonds with low energy barriers of rota-
tion. The experimental evidence indicates that
the hydrophobic moiety is associated with one
face of the glucose ring for a significant pro-
portion of the time, but few of the alternative
conformations could be excluded since the
Scheme 4.

300
N.E. Da6idson et al. / Carbohydrate Research 330 (2001) 295–307
Fig. 1.
3. Conclusions
solution state structure is under-defined by
NMR data.
In conclusion it can be seen that during the
synthesis of a novel 7-carboxyheptylglucosino-
late an intramolecular rearrangement reaction
has been observed. The first step was transes-
terification, resulting in attachment of the side
chain via an ester linkage to the 3-hydroxy
group of the glucose. This reaction is facili-
tated by the solution conformation of the
glucosinolate where there is an interaction of
the alkyl side chain one face of the glucose
ring. It then appears that further reaction
takes place, which cleaves the thiohydroxi-
mate moiety and probably produces an isocy-
anate after Lossen-type rearrangement. 1H
An approximate model was constructed by
a combination of molecular mechanics (AM-
BER) and semi-empirical (AM1) energy calcu-
lations, although a full conformational search
was not conducted due to the low accuracy
available for modelling the energy of the hy-
drophobic interactions. The model depicted in
Fig. 2 shows the carboxylate in proximity with
the C-3 hydroxyl of glucose. It is not clear on
the basis of this model why intramolecular
acylation should occur exclusively at the C-3
position, as the C-2 and C-4 hydroxyl groups
appear to be equally accessible.

N.E. Da6idson et al. / Carbohydrate Research 330 (2001) 295–307
301
NMR spectroscopy and molecular modelling
calculations (AM1 and AMBER) on the 7-
carboxyheptylglucosinolate, along with the
novel heptyl and naphthylmethyl derivatives
also synthesised, indicate that in all three cases
the hydrophobic side chains associate with
one face of the glucose ring.
layer, the remaining organics and aq layer
were extracted with EtOAc (65 mL). The or-
ganic layers were combined, dried (MgSO4)
and concentrated under reduced pressure to
afford a brown solid. The product was ex-
tracted as before and recrystallised from
EtOAc–petroleum ether (bp 40–60 °C) to af-
ford the title compound as a white solid (2.86
g, 69%): mp 127 °C; [h]2D5 −17.1° (c 0.14,
CHCl3); IR (Nujol); w 3300 (OH), 1730 (CO),
1600 cm−1 (CꢀN); 1H NMR (200 MHz,
CD3OD) 5.40 (1 H, t, J3,4 9 Hz, H-3), 5.28 (1
H, d, J1,2 10 Hz, H-1), 5.05 (2 H, m, H-2,4),
4.92 (1 H, br s, OH), 4.25 (1 H, dd, J5,6b 7,
J6a,6b 12 Hz, H-6b), 4.13 (1 H, dd, J5,6a 2, J6a,6b
12 Hz, H-6a), 3.99 (1 H, m, H-5), 2.55 (2 H, t,
J 6 Hz, CH2(CH2)5CH3), 2.07, 2.04, 2.05, 2.00
(12 H, 4×s, CH3COO), 1.66 (2 H, t, J 6 Hz,
CH2(CH2)4CH3), 1.35 (8 H, br s, (CH2)4CH3),
0.93 (3 H, t, J 7 Hz, CH3); 13C NMR (50.31
MHz, CD3COCD3) 171.00, 170.67, 170.39,
170.08 (CO), 151.66 (CꢀN), 80.34 (C-1), 76.58
(C-5), 74.61 (C-3), 71.35 (C-2), 69.61 (C-4),
63.51 (C-6), 32.96, 30.63, 29.85, 28.36, 21.03,
20.97 (CH2), 20.92 (CH3COO), 14.73 (CH3);
m/z (CI) 506 (MH+, 7%), 331 (100,
C6H7O(OAc)+4 ) and 271 (40, C6H7O(OAc)+3 );
Anal. Calcd for C22H35NO10S: C, 52.26; H,
6.98; N, 2.77. Found: C, 52.39; H, 7.12; N,
2.68; (HRMS) C22H36NO10S requires MH+
506.6871. Found MH+ 506.2060.)
4. Experimental
General methods.—Routine 1H and 13C
NMR spectra were obtained using a Varian
2000 f.t spectrometer (1H, 300 MHz;13C, 75.42
MHz) and a Varian Gemini f.t spectrometer
(1H 200 MHz; 13C 50.31 MHz). Mass spectra
were recorded on an A.E.I MS-902 spectrome-
ter. Optical rotations were measured at rt
using an Optical Activity Ltd. AA 1000 polar-
imeter with 20 cm path-length cells. IR spectra
were obtained on a Perkin–Elmer 1420 instru-
ment or
spectrometer.
Heptyl glucosinolate.—2,3,4,6-Tetra-O-ace-
tyl-1-thio-b- -glucopyranose (3 g, 8.23 mmol)
a
Perkin–Elmer 1710 FT-IR
D
was dissolved in dry THF (100 mL) before
1-chorooctaldoxime (1.46 g, 8.23 mmol), dis-
solved in dry THF (100 mL), was added. To
this, dry triethylamine (7.81 g, 77.2 mmol) was
added and the resulting solution was stirred,
under a N2 atmosphere, at rt for 22 h. Diethyl
ether (65 mL) was then added and the solu-
tion washed with 1 M H2SO4 (65 mL). After
removal of approximately 60% of the organic
A solution of chlorosulfonic acid (3.03 g,
26.0 mmol) in dry CH2Cl2 (30 mL) was added
to a stirred mixture of dry pyridine (38.6 g,
488.3 mmol) and dry CH2Cl2 (30 mL) main-
tained at 0 °C and under a N2 atmosphere.
2,3,4,6-Tetra-O-acetyl-b- -glucopyranosyl-
D
heptyl thiohydroximate (1.3 g, 2.6 mmol) dis-
solved in dry CH2Cl2 (20 mL) was then added
and the resulting mixture stirred as given pre-
viously. Potassium hydrogen carbonate (1.67
g, 16.7 mmol) was added in water (100 mL)
and this mixture allowed to stir for 30 min
before partial concentration under reduced
pressure followed by co-evaporation with tol-
uene. Further aq potassium hydrogen carbon-
ate (5.0 g, 50.1 mmol) was added and the
two-phase system partially concentrated under
reduced pressure to afford a brown solid.
Purification by chromatography, on silica gel,
eluting with petroleum ether (bp 40–60 °C):
Fig. 2.

302
N.E. Da6idson et al. / Carbohydrate Research 330 (2001) 295–307
1-Naphthylacetic acid ethyl ester.—1-
Naphthylacetic acid (20 g, 107.4 mmol) was
dissolved in dry EtOH (500 mL) before thio-
nyl chloride (14.06 g, 118.2 mmol) was added.
The resulting solution was heated under reflux
for 3.5 h before concentration under reduced
pressure to afford the title compound as a
yellow oil (22.95 g, 100%): IR (neat); w 3050
(aromatic) 2900–3000 (alkyl), 1730 (CO),
1520, 1600 (aromatic), 1150–1275 cm−1
EtOAc (1:1) with 10% (v/v) MeOH, allowed
recovery of the title compound as a white
solid (697 mg, 43%): mp 107–109 °C; [h]D25
−16.4° (c 0.14, MeOH); IR (Nujol); w 1750
1
(CO), 1580 (CꢀN), 1250 cm−1 (COOC); H
NMR (200 MHz, CD3OD) 5.40 (1 H, t, J3,4
9
Hz, H-3), 5.36 (1 H, d, J1,2 10 Hz, H-1), 5.03
(2 H, 2×t, J 10, 9 Hz, H-2,4), 4.24 (2 H, dd,
J5,6b 4, J6a,6b 12 Hz, H-6b), 4.13 (2 H, dd, J5,6a
2, J6a,6b 12 Hz, H-6a), 4.03 (1 H, m, H-5),
2.66 (2 H, t, J 8 Hz, CH2(CH2)5CH3), 2.07,
2.04, 1.99 (12 H, 3×s, CH3COO), 1.85 (2 H,
br m, CH2(CH2)4CH3), 1.34 (8 H, br s,
(CH2)4), 0.91 (3 H, t, J 6 Hz, CH3); 13C
NMR (50.31 MHz, CD3OD) 172.51, 171.82,
171.55, 171.25 (CO), 159.94 (CꢀN), 81.18 (C-
1), 77.03 (C-5), 75.31 (C-3), 71.72 (C-2),
69.80 (C-4), 63.68 (C-6), 33.98, 33.25, 30.54,
30.49, 28.78, 24.02 (CH2), 21.02, 20.88
(CH3COO) 14.78 (CH3); m/z (ES−) 584
1
(COOC); H NMR (200 MHz, CDCl3) 8.06 (1
H, d, J 8 Hz, H-2), 7.88 (2 H, m, H-4,8), 7.54
(4 H, m, H-3,5,6,7), 4.20 (2 H, q, J 7 Hz,
OCH2CH3), 4.11 (2 H, s, CH2COCH2CH3),
1.27 (3 H, t, J 7 Hz, CH3); 13C NMR (50.31
MHz, CDCl3) 172.19 (CO), 134.36 (ArC-10),
132.67 (ArC-9), 131.25 (ArC-1), 129.29,
128.61, 126.89, 126.34, 126.09, 126.01, 124.41
(ArC), 61.54 (OCH2CH3), 39.85 (CH2CO-
CH2CH3), 14.79, 14.73 (OCH2CH3).
([M−K]−,
100%);
Anal.
Calcd
for
1-Naphthylacetaldehyde.—1-Naphthylacetic
acid ethyl ester (10 g, 46.7 mmol) was dis-
solved in dry toluene (200 mL) and the solu-
tion cooled to −78 °C before DIBAL-H (1.5
M in toluene; 12.74 g, 89.6 mmol) was
added. The resulting solution was stirred at
−78 °C for 2 h under a N2 atmosphere be-
fore dry MeOH (100 mL) was added and the
solution allowed to warm slowly to rt. A
solution of potassium sodium tartrate (60 g
in 100 mL H2O) was added and the biphasic
system stirred overnight. The organic layer
was removed and the aq layer washed with
Et2O (2×150 mL). The organic layers were
combined, dried (MgSO4) and concentrated
under reduced pressure to yield a yellow oil
C22H34KNO13S2: C, 42.36; H, 5.49; N, 2.25.
Found: C, 42.08; H, 5.52; N, 1.98.
Finally a solution of potassium methoxide
was prepared by adding a catalytic amount
of potassium (up to 100 mg) to dry MeOH
(1 mL) at rt and under a N2 atmosphere. A
small volume of this solution was added to
2,3,4,6-tetra-O-acetyl-heptyl
glucosinolate
(460 mg, 0.74 mmol), dissolved in dry MeOH
(10 mL), until the pH reached 8–9. The re-
sulting solution was stirred overnight under a
N2 atmosphere before being concentrated un-
der reduced pressure to afford a gold
coloured solid (337, 100%): [h]2D5 −14.0°
(c 0.2, H2O); IR (Nujol) w 3400 cm−1 (OH);
1H NMR (200 MHz, CD3OD) 4.83 (1 H, d,
J1,2 10, H-1), 3.85 (1 H, d, J6a,6b 10 Hz,
H-6b), 3.65 (1 H, dd, J5,6a 4, J6a,6b 10 Hz,
H-6a), 3.34 (4 H, m, H-2,3,4,5), 2.69 (2 H, t,
J 7, Hz CH2(CH2)5CH3), 1.71 (2 H, br m,
CH2-
1
(7.43 g, 94%): H NMR (200 MHz, CDCl3)
9.79 (1 H, t, J 2 Hz, CHO), 7.90 (3 H, m,
ArH), 7.46 (4 H, m, ArH), 4.11 (2 H, d, J 2
Hz, CH2); 13C NMR (50.31 MHz, CDCl3)
200.28 (CHO), 134.43 (ArC-10), 132.80 (ArC-
9), 48.89 (CH2), 129.47, 129.03, 128.89,
127.26, 126.63, 126.22, 126.15, 124.10(ArC).
1 - Naphthylacetaldoxime.—1 - Naphthylace-
taldehyde (6.65 g, 39.1 mmol) was dissolved
in EtOH (70 mL) and pyridine (7 mL) before
hydroxylamine hydrochloride (5 g, 71.9
mmol) was added. The resulting solution was
heated under reflux, with stirring, for 2.5 h
before being concentrated under reduced
pressure. Water was then added resulting in
(CH2)4CH3), 1.33 (8 H, br m, (CH2)4), 0.91 (3
H, t, J 7 Hz, CH3); 13C NMR (50.31 MHz,
CD3OD) 162.66 (CꢀN), 84.01 (C-1), 82.59
(C-5), 79.83 (C-3), 74.44 (C-2), 71.39 (C-4),
62.90 (C-6), 33.99, 33.25, 30.65, 30.47, 29.08,
24.02 (CH2), 14.77 (CH3); m/z (ES−) 416
([M−K]−, 100%); Anal. Calcd for C14H26-
KNO9S2: C, 36.91; H, 5.75; N, 3.07. Found:
C, 37.12; H, 5.89; N, 2.84.

N.E. Da6idson et al. / Carbohydrate Research 330 (2001) 295–307
303
CH2), 3.94 (1 H, dd, J5,6b 5, J6a,6b 12 Hz,
H-6b), 3.79 (1 H, d, J6a,6b 12 Hz, H-6a), 3.05
(1 H, m, H-5), 1.91, 1.90, 1.85 (12 H, 3×s,
the appearance of a yellow solid which was
recrystallised from EtOH to afford the title
compound (3.78 g, 52%) as a mixture of iso-
mers (40% anti:60% syn): mp 127–128 °C
(lit.,15 123–124 °C); 1H NMR (200 MHz,
CDCl3) 8.80 (1 H, br s, NOH), 8.08–7.38 (7
H, m, ArH), 6.84 (1 H, t, J 5 Hz, CHꢀN), 4.20
(2 H, d, J 5 Hz, CH2 (syn)), 4.02 (2 H, d, J 5
Hz, CH2 (anti)); 13C NMR (50.31 MHz,
CDCl3) 151.49 (CHꢀN (syn) 151.15 (CHꢀN
(anti)), 134.38, 133.31, 129.52, 129.34, 128.41,
128.29, 127.61, 127.43, 127.36, 127.29, 127.08,
126.94, 126.72, 126.63, 126.44, 126.11, 126.06,
124.22, 124.11, 124.02 (ArC (syn, anti)), 33.95,
30.19 (CH2); m/z (CI) 186 (MH+, 100%), 141
(52, [M−CH2NOH]+).
13
CH3); C NMR (50.31 MHz, CDCl3) 171.34,
170.76, 169.98, 169.88(COO), 150.17 (CꢀN),
134.18, 132.33, 131.67, 129.47, 128.33, 126.89,
126.37, 126.02, 125.90, 123.12, (ArC), 80.05
(C-1), 75.92 (C-5), 74.06 (C-3), 70.34 (C-2),
68.13 (C-4), 62.36 (C-6), 35.57 (CH2), 20.89,
20.78(CH3); m/z (CI) 548 (MH+, 8%), 331
(100, C6H7O(OAc)+4 ) and 271 (33, C6H7O-
(OAc)+3 ); Anal. Calcd for C26H29NO10S·0.5
H2O: C, 56.11; H, 5.43; N, 2.53. Found: C,
56.10; H, 5.16; N, 2.48; HRMS C26H30NO10S
requires MH+ 548.5880. Found MH+
548.1590.
Chloro-1-naphthylacetaldoxime.—1-Naph-
thylacetaldoxime (6.5 g, 35.1 mmol) was
added to a stirred solution of CHCl3 (50 mL)
and pyridine (1.39 g, 17.6 mmol) before N-
chlorosuccinimide (4.69 g, 35.1 mmol) was
added slowly. The resulting solution was
stirred at rt for 4 h then poured onto ice-wa-
ter. The organic layer was separated and
washed with water (50 mL) and satd NaCl
solution (50 mL) while the aq layer was ex-
tracted with Et2O (75 mL). The organics were
combined, dried (MgSO4) and concentrated
under reduced pressure. This afforded a mix-
ture of the (1:2) oxime and oximyl chloride
2,3,4,6 - Tetra - O - acetyl - 1 - naphthylmethyl
glucosinolate.—Sulfonation of 2,3,4,6-tetra-
O - acetyl - b -
D
- glucopyranosyl - 1 - naphthyl-
methyl thiohydroximate (3.50 g, 6.4 mmol)
was carried out as for the heptyl derivative.
Purification by flash chromatography on silica
gel, eluting with (1:1) petroleum ether (bp
40–60 °C)–EtOAc then (9:1) EtOAc–MeOH
afforded the protected glucosinolate as a
brown solid (2.54 g, 60%): mp (dec.) 110–
112 °C; [h]2D5 −27.5° (c 0.14, H2O); IR (Nu-
jol); w 3000 (aromatic), 1750 (CO), 1225 cm−1
1
(COOC); H NMR (200 MHz, CD3OD) 8.19
(1 H, d, J 8 Hz, H-2%), 7.96 (1 H, d, J 8 Hz,
H-8%), 7.86 (1 H, d, J 8 Hz, H-4%), 7.56 (4 H,
m, H-3%,5%,6%,7%), 4.89 (4 H, m, H-1,2,3,4), 4.53
(2 H, s, CH2), 4.01 (1 H, dd, J5,6a 5, J6a,6b 12
Hz, H-6b), 3.75 (1 H, d, J6a,6b 12 Hz, H-6a),
3.30 (1 H, m, H-5), 1.95 (12 H, m, CH3COO);
13C NMR (50.31 MHz, CD3OD) 172.42,
171.64, 171.32, 171.07 (COO), 158.87 (CꢀN),
124.49–135.66 (ArC), 81.36 (C-1), 77.12 (C-5),
75.13 (C-3), 71.30 (C-2), 69.31 (C-4), 63.27
(C-6), 36.51 (CH2), 20.91, 20.76 (CH3); m/z
(ES−) 626 ([M−K]−, 100%); Anal. Calcd for
C26H28KNO13S2: C, 46.91; H, 4.24; N, 2.10.
Found: C, 47.12; H, 4.54; N, 2.07.
1
(5.53 g, 48% oximyl chloride): H NMR (200
MHz, CDCl3) 8.86 (1 H, br s, NOH), 8.09–
7.69 (3 H, m, H-2,4,8), 7.32–7.61 (4 H, m,
13
H-3,5,6,7), 4.24 (2 H, s, CH2CꢀN); C NMR
(50.31 MHz, CDCl3) 140.07 (CꢀN), 122.87-
133.83 (ArC), 40.20 (CH2).
2,3,4,6-Tetra-O-acetyl-i- -glucopyranosyl-
D
1-naphthylmethyl thiohydroximate.—This was
prepared as for the heptyl derivative using
2,3,4,6-tetra-O-acetyl-1-thio-b- -glucopyra-
D
nose (4.32 g, 11.9 mmol) and crude chloro-1-
naphthylacetaldoxime (3.65 g, representing ca.
11 mmol of the oxime). Recrystallisation from
EtOH afforded the title compound as a white
1-Naphthylmethyl glucosinolate (6).—De-
protection of 2,3,4,6-tetra-O-acetyl-1-naph-
thylmethyl glucosinolate (500 mg, 0.75 mmol)
was carried out in dry MeOH (10 mL) with a
catalytic amount of potassium metal as for the
heptyl derivative. Purification by flash chro-
matography, on silica gel, eluting initially with
EtOAc then (4:1) EtOAc–MeOH afforded the
solid (5.62 g, 87%): mp 208–210 °C; [h]D25
+
4.64° (c 0.14, MeOH); IR (Nujol); w 3300
(OH), 1720, 1750 (CO), 1620 (CꢀN), 1225,
1
1250 cm−1 (COOC); H NMR (200 MHz,
CDCl3) 7.80 (3 H, m, H-2%,4%,8%), 7.40 (4 H, m,
H-3%,5%,6%,7%), 4.84 (3 H, m, H-2,3,4), 4.64 (1 H,
d, J1,2 9 Hz, H-1), 4.26 (2 H, 2×d, J 18 Hz,

304
N.E. Da6idson et al. / Carbohydrate Research 330 (2001) 295–307
for 11 h before being rapidly cooled and
poured onto ice-cold brine. The product was
extracted with Et2O (2×300 mL) and the
organic layer washed with water (150 mL),
brine (150 mL), satd NaHCO3 (150 mL) and
more brine (150 mL) before being dried
(MgSO4) and concentrated under reduced
pressure. The crude product was purified by
flash chromatography, on silica gel, eluting
with (9:1) petroleum ether (bp 40–60 °C)–
EtOAc. This afforded the title compound as a
colourless oil (2.27 g, 61%): IR (neat); w 2722,
2860, 2937 (alkyl), 1735 (CO), 1349–1466 (al-
product as a gold-coloured foam (198 mg,
53%): [h]2D5 −3.8° (c 0.2, H2O); IR (Nujol); w
3390 cm−1 (OH); 1H NMR (200 MHz,
CD3OD) 8.22 (1 H, d, J 8 Hz, H-2%), 7.91 (1
H, d, J 8 Hz, H-8%), 7.81 (1 H, d, J 8 Hz,
H-4%), 7.41–7.65 (4 H, m, H-3%,5%,6%,7%), 4.82 (1
H, d, J1,2 6 Hz, H-1), 4.37 (2 H, m, CH2), 3.74
(1 H, dd, J5,6b 2, J6a,6b 12 Hz, H-6b), 3.56 (1 H,
dd, J5,6a 6, J6a,6b 12 Hz, H-6a), 3.17 (2 H, t, J
9 Hz, H-2,4), 2.93 (1 H, t, J3,4 9 Hz, H-3), 2.82
(1 H, m, H-5); 13C NMR (50.31 MHz,
CD3OD) 162.03 (CꢀN), 135.59, 133.54,
132.98, 130.20, 129.12, 128.03, 127.40, 127.10,
126.85, 124.65 (ArC), 83.81, 83.78 (C-1), 82.47
(C-5), 79.40 (C-3), 74.26 (C-2), 71.10 (C-4),
62.76, 62.67 (C-6), 36.35 (CH2); m/z (ES−)
458 ([M−K]−, 100%); Anal. Calcd for
C18H20KNO9S2·0.5 H2O: C, 42.68; H, 4.18; N,
2.76. Found: C, 42.48; H, 4.73; N, 2.40.
1
kyl), 1182 cm−1 (COOC); H NMR (200
MHz, CDCl3) 9.71 (1 H, t, J 2 Hz, CHO),
4.07 (2 H, q, J 7 Hz, COOCH2), 2.39, 2.38 (2
H, 2×t, J 7 Hz, CH2CO), 2.24 (2 H, t, J 7
Hz, CH2COO), 1.46 (4 H, m, (CH2)2), 1.30–
13
1.15 (7 H, m, (CH2)2, CH3); C NMR (50.31
MHz, CDCl3) 203.12 (CHO), 174.13 (COO),
60.65 (OCH2CH3), 44.21 (CH2CHO), 34.64
(CH2COO), 29.25, 29.20, 25.14, 22.28 (CH2),
14.68 (CH3); m/z (CI) 187 (MH+, 86%), 171
(5, [M−CH3]+), 141 (100, [M−OCH2CH3]+).
Ethyl 8-hydroxyiminooctanoate.—Ethyl 8-
oxooctanoate (2.26 g, 12.11 mmol) was dis-
solved in EtOH (45 mL) and pyridine (4.5
mL) before hydroxylamine hydrochloride
(1.55 g, 22.3 mmol) was added and the result-
ing solution heated under reflux for 2.5 h.
Concentration under reduced pressure and ad-
dition of water afforded a white solid which
was removed by filtration under reduced pres-
sure. This was identified as the oxime existing
as a mixture of (1:1) syn and anti isomers,
(2.06 g, 85%): mp 28 °C; IR (Nujol); w 3050–
2200 (OH), 2820, 2900 (alkyl), 1700 cm−1
Ethyl
8-bromooctanoate.—8-Bromooc-
tanoic acid (5.78 g, 25.9 mmol) was dissolved
in dry EtOH (125 mL) before thionyl bromide
(5.92 g, 28.5 mmol) was added and the result-
ing solution heated under reflux with stirring
for 2 h under a N2 atmosphere. Concentration
under reduced pressure afforded a brown oil
which was purified by flash chromatography,
on silica gel, eluting initially with petroleum
ether (bp 40–60 °C) then (9:1) petroleum
ether–EtOAc. This yielded the title compound
as a pale yellow oil (5.26 g, 81%); IR (neat)
2858, 2935 (alkyl), 1736 (CO), 1373–1465 (al-
1
kyl), 1183 cm−1 (COOC); H NMR (200
MHz, CDCl3) 4.12 (2 H, q, J 7 Hz, OCH2),
3.40 (2 H, t, J 7 Hz, BrCH2), 2.29 (2 H, t, J 7
Hz, CH2COO), 1.85 (2 H, quintet, J 7 Hz,
BrCH2CH2), 1.62 (2 H, quintet, J 7 Hz,
CH2CH2COO), 1.39 (6 H, m, (CH2)3), 1.25 (3
H, t, J 7 Hz, CH3); 13C NMR (50.31 MHz,
CDCl3) 174.26 (COO), 60.70 (OCH2CH3),
34.76, 34.41, 33.18, 29.39, 28.89, 28.45, 25.31
(CH2), 14.72 (CH3); m/z (CI) 252, 254 (MH+,
6%), 251, 253 (55, 56, M+), 205, 207 (5,
[M−C2H5OH]+).
Ethyl 8-oxooctanoate (8).—Following the
procedure of Bauer and Macomber,21 KI (3.33
g, 20.1 mmol) and Na2CO3 (2.11 g, 20.1
mmol) were suspended in Me2SO (150 mL) at
80–85 °C before ethyl 8-bromoooctanoate (5
g, 20.1 mmol) was added with stirring. This
mixture was stirred at the above temperature
1
(CO); H NMR (200 MHz, CDCl3) 7.92 (1 H,
br s, NOH), 7.40 (1 H, t, J 6 Hz, CHꢀN
(anti)), 6.72 (1 H, t, J 5 Hz, CHꢀN (syn)), 4.11
(2 H, q, J 8 Hz, OCH2), 2.28 (2 H, t, J 8 Hz,
CH2COO), 2.18 (2 H, q, J 7 Hz, CH2CH),
1.61 (2 H, m, CH2), 1.48 (2 H, m, CH2), 1.34
(4 H, m, (CH2)2), 1.24 (3 H, t, J 7 Hz, CH3);
13C NMR (50.31 MHz, CDCl3) 174.39 (CO),
152.51 (CHꢀN (syn)), 152.45 (CHꢀN (anti)),
60.76 (OCH2), 34.75 29.87, 29.47, 29.25, 29.17,
26.80, 26.29, 25.40, 25.25 (CH2), 14.73 (CH3);
m/z (CI) 202 (MH+, 100%), 184 (11, [M−
OH]+), 138 (18, [M−C2H7O2]+); Anal. Calcd
for C10H19NO3: C, 59.68; H, 9.51; N, 6.96.
Found: C, 59.62; H, 9.59; N, 6.87.

N.E. Da6idson et al. / Carbohydrate Research 330 (2001) 295–307
305
Ethyl
8-chloro-8-hydroxyiminooctanoate
2,3,4,6-Tetra-O-acetyl-(7-ethoxycarbonyl-
heptyl) glucosinolate.—Chlorosulfonic acid
(9.1 g, 78.1 mmol) was added to dry CH2Cl2
(65 mL) and the resulting solution was added
slowly to an ice-cold solution of dry CH2Cl2
(65 mL) and dry pyridine (58.6 g, 741 mmol)
maintained under a N2 atmosphere. The thio-
hydroximate (2.2 g, 3.9 mmol) was added as a
solution in dry CH2Cl2 (45 mL) at rt and the
resulting solution was stirred under a N2 at-
mosphere for 24 h. Potassium hydrogen car-
bonate (11.7 g, 117 mmol) was added in water
(100 mL) and the resulting two-phase system
was concentrated under reduced pressure to
allow removal of the organics. This yielded a
substantial quantity of salts which were re-
moved by filtration before the aq layer was
extracted with EtOAc. The organic layer was
dried (MgSO4) and concentrated under re-
duced pressure to afford a cream-coloured
foam. This was purified by flash chromatogra-
phy, on silica gel, eluting with (1:1) petroleum
ether (bp 40–60 °C)–EtOAc then (4:1)
EtOAc–MeOH. This afforded the title com-
pound as a cream-coloured solid (1.91 g,
72%): mp (dec.) 109–111 °C; [h]2D5 −11.4° (c
0.14, H2O); IR (Nujol); w 1700 (CO), 1560
(9).—The oxime (2 g, 10.1 mmol) was dis-
solved in CHCl3 (32 mL) and pyridine (397
mg, 5 mmol) before N-chlorosuccinimide
(1.34 g, 10.1 mmol) was added slowly at 0 °C.
The resulting solution was stirred at rt for 3.5
h before being poured onto ice-water. The
organic layer was separated and washed with
water (30 mL) then brine (30 mL) and the aq
layer was extracted with Et2O (2×30 mL).
The organic layers were combined, dried
(MgSO4) and concentrated under reduced
pressure to afford a pale yellow oil (2.05 g,
1
86%): H NMR (200 MHz, CDCl3) 4.10 (2 H,
q, J 7 Hz, OCH2), 2.46 (2 H, t, J 7 Hz,
CH2COO), 2.27 (2 H, t, J 7 Hz, CH2CꢀN),
1.61 (4 H, m, (CH2)2), 1.47–1.19 (7 H, m,
13
(CH2)2, CH3); C NMR (50.31 MHz, CDCl3)
174.55 (CO), 144.34 (CꢀN), 60.85 (COOCH2),
25.20, 26.51, 28.57, 29.11, 34.72, 36.92 (CH2),
14.70 (CH3).
2,3,4,6-Tetra-O-acetyl-i- -glucopyranosyl-
D
(7-ethoxycarbonylheptyl) thiohydroximate.—
The synthesis was carried out as for the heptyl
derivative using 2,3,4,6-tetra-O-acetyl-1-thio-
b- -glucopyranose (2.24 g, 6.14 mmol) and
D
the oximyl chloride (2.03 g, 8.59 mmol) in dry
THF (65 mL). Recrystallisation from a mix-
ture of hexane and EtOAc afforded the title
compound as a white solid (2.57 g, 74%): mp
90 °C; [h]2D5 −12.1° (c 0.14, CHCl3); IR (Nu-
jol): w 3240 (OH), 1685, 1715 (CO), 1225
cm−1 (COOC); 1H NMR (200 MHz, CD3OD)
5.39 (1 H, t, J3,4 10 Hz, H-3), 5.28 (1 H, d, J1,2
10 Hz, H-1), 5.05, 4.98 (2 H, 2×t, J 10, 10
Hz, H-2,4), 4.83 (1 H, br s, NOH), 4.22–4.07
(4 H, m, OCH2 and H-6a,6b), 4.00 (1 H, m,
H-5), 2.55 (2 H, t, J 7 Hz, CH2CꢀN), 2.33 (2
H, t, J 7 Hz, CH2COO), 2.06, 2.03, 1.99, 1.98
(12 H, 4×s, CH3COO), 1.65 (4 H, m,
(CH2)2), 1.40 (4 H, m, (CH2)2), 1.25 (3 H, t, J
1
(CꢀN), 1200 cm−1 (COOC); H NMR (200
MHz, CD3OD) 5.41 (1 H, t, J3,4 10 Hz, H-3),
5.37 (1 H, d, J1,2 10 Hz, H-1), 5.06, 4.99 (2 H,
2×t, J 10, 10 Hz, H-2,4), 4.30-3.88 (5 H, m,
OCH2, H-5,6a,6b), 2.67 (2 H, t, J 7 Hz,
CH2CꢀN), 2.34 (2 H, t, J 7 Hz, CH2COO),
2.07, 2.04, 2.02, 1.99 (12 H, 4×s, CH3COO),
1.67 (4 H, m, (CH2)2), 1.41 (4 H, m, (CH2)2),
1.25 (3 H, t, J 7 Hz, CH3); 13C NMR (50.31
MHz, CD3OD) 175.89 (COOCH2CH3),
172.50, 171.80, 171.52, 171.23 (COO), 159.77
(CꢀN), 81.19 (C-1), 77.03 (C-5), 75.33 (C-3),
71.75 (C-2), 69.86 (C-4), 63.69 (C-6), 61.72
(OCH2), 35.30, 33.83, 30.09, 29.97, 28.45,
26.13 (CH2), 21.03, 20.87 (CH3COO), 14.89
(CH3); m/z (ES−) 642 ([M−K]−, 100%);
Anal. Calcd for C24H36KNO15S2: C, 42.28; H,
5.32; N, 2.05. Found: C, 42.53; H, 5.41; N,
2.29.
13
7 Hz, CH3); C NMR (50.31 MHz, CD3OD)
174.37 (COOCH2CH3), 171.00, 170.37,
170.03, 169.78 (CH3COO), 151.06 (CꢀN),
79.50 (C-1), 75.48 (C-5), 73.95 (C-3), 70.54
(C-2), 68.49 (C-4), 62.27 (C-6), 60.25 (OCH2),
33.81 31.86, 28.64, 28.50, 27.11, 24.69, 19.54,
19.45, 19.39 (CH2), 13.41 (OCH2CH3); m/
z(CI) 564 (MH+, 2%), 331 (100, C6H7O-
(OAc)+4 ), 271 (33, C6H7O(OAc)+3 ); Anal.
Calcd for C24H37NO12S: C, 51.15; H, 6.62; N,
2.49. Found: C, 51.19; H, 6.72; N, 2.49.
(7-Methoxycarbonylheptyl)
glucosinolate
(10b).—2,3,4,6 - Tetra - O - acetyl - (7 - ethoxy-
carbonylheptyl) glucosinolate (750 mg, 681.8
mmol) was dissolved in dry MeOH (10 mL)
before a catalytic quantity of potassium

306
N.E. Da6idson et al. / Carbohydrate Research 330 (2001) 295–307
NMR studies.—NMR spectra were ac-
quired using a Varian Unity+ spectrometer,
operating at 500.3 MHz (1H frequency), with
a nominal probe temperature of 30 °C. One-
methoxide was added with stirring until the
pH reached 8–9. The resulting solution was
stirred at rt, under a N2 atmosphere, for 24 h.
Concentration under reduced pressure af-
forded a cream-coloured foam (540 mg, 98%):
[h]2D5 −12° (c 0.2, H2O); IR (Nujol); w 3400
(OH), 2850, 2930 (alkyl), 1720 (CO), 1250
1
dimensional H spectra were recorded with 20
s relaxation delay, to ensure full relaxation
between pulses and accurate quantification
from peak integrals. Prior to transformation,
all data were apodised with a cosine-bell func-
tion followed by zero-filling.
1
cm−1 (COOC); H NMR (200 MHz, D2O)
4.93 (1 H, d, J1,2 9 Hz, H-1), 3.81 (1 H, dd,
J5,6b 2, J6a,6b 11 Hz, H-6b), 3.63 (1 H, m, H-6a),
3.61 (3H, s, OCH3), 3.42 (4 H, m, H-2,3,4,5),
2.63 (2 H, t, J 7 Hz, CH2CꢀN), 2.32 (2 H, t, J
7 Hz, CH2COO), 1.57 (4 H, m, (CH2)2), 1.30
(4 H, m, (CH2)2); 13C NMR (50.31 MHz,
D2O) 180.60 (COO), 167.53 (CꢀN), 84.73 (C-
1), 83.02 (C-5),7 9.99 (C-3), 74.87 (C-2), 72.01
(C-4), 63.52 (C-6), 55.05 (OCH3), 36.57, 34.99,
30.77, 30.67, 29.63, 27.10 (CH2); m/z (ES−)
460 ([M−K]−, 100%); Anal. Calcd for
C15H26KNO11S2·0.5 H20 requires C, 35.42; H,
5.35; N, 2.75. Found: C, 35.26; H, 5.15; N,
2.65.
Attempted synthesis of (7-carboxyheptyl)
glucosinolate (11).—(7-Methoxycarbonylhep-
tyl) glucosinolate (100 mg, 0.20 mmol) was
dissolved in MeOH (600 mL) before 1 M
potassium hydroxide (300 mL, 0.30 mmol) was
added and the resulting solution was stirred
overnight at rt. Analysis by TLC ((4:1)
EtOAc–MeOH) showed the reaction to be
incomplete and more potassium hydroxide
(100 mL, 0.10 mmol) was added. The resulting
solution was stirred as before for several hours
prior to concentration under reduced pressure.
Methanol was then added and the remaining
solid was removed by filtration before purifi-
cation, on C18 reverse-phase silica, eluting
with water then MeOH. NMR analysis of the
aq fractions showed (7-carboxyheptyl) glucosi-
nolate as a mixture with two side-products; 1H
NMR (500 MHz, D2O) 5.00 (H-1), 3.88 (H-
6b), 3.71 (H-6a), 3.56 (H-3), 3.54 (H-5), 3.45
(H-4), 3.44 (H-2), 2.69 (H-1%), 2.17 (H-6%), 1.71
(H-2%), 1.55 (H-5%), 1.40 (H-3%), 1.33 (H-4%); 13C
NMR (125.8 MHz, D2O) 186.68 (COOH),
167.63 (CꢀN), 84.50 (C-1), 82.80 (C-5), 79.74
(C-3), 74.63 (C-2), 71.81 (C-4), 63.26 (C-6),
40.23 (C-6%), 34.80 (C-1%), 31.00, 30.67, 29.59,
28.43 (C-2%,3%,4%,5%); m/z (ES−) 222 ([M−2K]2-,
100%).
To suppress zero-quantum correlations in
NOESY experiments, the mixing time was
incremented on each pulse of the 16-step
phase cycle.20 The total variation of the mix-
ing time was 95%. 1D TOCSY-NOESY
spectra were acquired using the implementa-
tion described by Uhrin and Barlow.20 All
selective pulses used the quiet-sneeze pulse
shape, spanning a single bandwidth. Spectra
were acquired with 400 ms NOE mixing time,
a recycle time of 5.4 s per transient, and a
digital resolution after transformation of 0.2
Hz/point. After 1000 scans the limit of detec-
tion of a signal above the noise was estimated
,
to be an NOE of 0.2% (equivalent to 4.9 A
1
time-averaged H–1H distance).
J-coupling constants were measured from
homonuclear J-resolved spectra, with a digital
resolution of 0.25 Hz/point in f1 after
transformation.
Molecular modelling.—Minimum energy
models of the acidic glucosinolate 11 were
computed using either the AM1 semi-empiri-
cal parameter set implemented within the MO-
PAC software package, or the AMBER
molecular mechanics force field, implemented
within the DISCOVER package (MSI, San
Diego). All calculations were run on a Silicon
Graphics O2 computer, with R5000 processor.
Initial models were constructed from b- -glu-
D
cose (optimised using AMBER), with the hy-
drophobic acyl chain extended away from the
saccharide. Following an initial AM1 optimi-
sation to refine the geometry of the thiooxime
functionality the structure was modified by
restrained molecular mechanics minimisation
to place the alkyl chain into an orientation
consistent with experimental NOE data. Po-
2
,
tentials of 10 kcal/mol/A (up to a maximum
of 10 kcal/mol) were applied to restrain the
protons of the two methylene groups closest

N.E. Da6idson et al. / Carbohydrate Research 330 (2001) 295–307
307
2. Griffiths, G. W.; Birch, A. N. E.; Hillman, J. R. J. Hort.
Sci. Biotech. 1998, 73, 1–18.
3. Heaney, R. K.; Fenwick, G. R. Natural Toxins 1995, 3,
233–237.
4. (a) Zhang, Y.; Kensler, T. W.; Cho, C.- G.; Posner. G.
H.; Talalay, P. Proc. Natl. Acad. Sci. USA 1994, 91,
3147–3150. (b) Fahey, J. W.; Zhang, Y.; Talalay, P.
Proc. Natl. Acad. Sci. USA 1997, 94, 10367–10372.
5. Louda, S.; Mole, S. In The Chemical Participants; Rosen-
thal, G. A.; Berenbaum, M. R., Eds. Herbivores: their
interactions with secondary plant metabolites. Academic:
New York, 1991; Vol. 1, pp. 123–164.
6. Bauer, R.; Birch, A. N. E.; Hopkins, R. J.; Griffiths, D.
W.; Simmonds, M. S. J.; Sta¨dler, E. Entomol. Exp. Appl.
1996, 78, 61–75.
7. Hopkins, R. J.; Birch, A. N. E.; Griffiths, D. W.; Baur,
R.; Sta¨dler, E.; McKinlay, R. G. J. Chem. Ecol. 1997, 23,
629–643.
8. Roessingh, P.; Sta¨dler, E.; Fenwick, G. R.; Lewis, J. A.;
Nielsen, J. K.; Hurter, J.; Ramp, T. Entomol. Exp. Appl.
1992, 65, 267–282.
9. Hopkins, R. J.; Wright, F.; McKinlay, R. G.; Birch, A.
N. E. Entomol. Exp. Appl. 1996, 80, 93–96.
10. Roessingh, P.; Sta¨dler, E. Entomol. Exp. Appl. 1990, 57,
93–100.
,
to the thiooxime to within 5 A of Glc H-1 and
,
to within 5 A of Glc H-5. As the conforma-
tion was under-defined by experimental data
at the distal end of the carboxyheptyl chain an
additional weak restraint (up to 2 kcal/mol
maximum) was applied between Glc C-3 and
the carbonyl carbon of the carboxylate. Since
there are no reliable molecular mechanics
parameters available for a thiooxime, crude
parameters for bond lengths and bond angles
were added to the AMBER force field library,
based upon the AM1 optimised geometry.
Bond lengths for the thiooxime were consis-
tent with the corresponding values in the crys-
tal structure of sinigrin,22 with the exception
,
of b(NꢁO) which was 1.27 A by AM1 and
1.44 A in sinigrin. Most bond angles were also
,
similar in the AM1 and crystal structure mod-
els, but significant differences were observed
for t(SꢁC-7ꢁC-8) (109.1° by AM1 vs. 125.3° in
sinigrin), t(NꢁOꢁS) (119.0° by AM1 vs. 109.4°
in sinigrin) and t(C-7ꢁNꢁO) (120.1° by AM1
vs. 110.7° in sinigrin). Models are thus pre-
sented as a qualitative description only. The
values used for the sulfate group parameters
were those reported by Huige and Altona.24
Torsional potentials for the thiooxime were
set to zero. The resulting structure was sub-
jected to a final AM1 minimisation to com-
pensate for the unrefined AMBER parameters
in the thiooxime.
11. Spencer, J. L. Entomol. Exp. Appl. 1996, 81, 165–173.
12. Benn, M. H. Can. J. Chem. 1963, 41, 2836.
13. Viaud, M. C.; Rollin, P.; Latxague, L.; Gardrat, C. J.
Chem. Res (S) 1992, 207.
14. (a) Cassel, S.; Casenave, B.; Deleris, G.; Latxague, L.;
Rollin, P. Tetrahedron 1998, 54, 8515–8524. (b)
Brochard, L.; Joseph, B.; Viaud, M. C.; Rollin, P. Synth.
Commun. 1994, 24, 1403–1414.
15. Jensen, K. A.; Dynesen, E. Acta Chem. Scand. 1950, 4,
692–702.
16. Bauer, D. P.; Macomber, R. S. J. Org. Chem. 1975, 40,
1990.
17. Nishida, Y.; Hori, H.; Ohrui, H.; Meguro, H. J. Carbo-
hydr. Chem. 1987, 7, 239–250.
18. Haverkamp, J.; van Halbeek, H.; Dorland, L.; Vliegen-
thart, J. F. G.; Pfeil, R.; Schauer, R. Eur. J. Biochem.
1982, 122, 305–311.
19. Prestera, T.; Fahey, J. W.; Holtzclaw, W. D.; Abeygu-
nawardana, C.; Kachinski, J. L.; Talalay, P. Anal.
Biochem. 1996, 239, 168–179.
Acknowledgements
20. Uhrin, D.; Barlow, P. N. J. Magn. Reson. 1997, 126,
We thank the Wellcome Trust for grant
040331 and the University of St Andrews for a
studentship (to NED).
248–255.
21. Kornblum, N.; Larson, H. O.; Blackwood, R. K.;
Mooberry, D. D.; Oliveto, E. P.; Graham, G. E. J. Am.
Chem. Soc. 1956, 78, 1497–1501.
22. Marsh, R. E.; Waser, J. Acta Crystallogr. 1970, 26,
1030–1037.
1
23. Pouchert, C. J.; Behnke, J. Aldrich Library of 13C and H
References
FT NMR Spectra, Aldrich, 1993.
24. Huige, C. J. M.; Altona, C. J. Comput. Chem. 1995, 16,
56–79.
1. Fenwick, G. R.; Heaney, R. K.; Mullin, W. J. CRC Re6.
Food Sci. Nutr. 1983, 18, 123–201.
.