Initial and Final Products, Nitriles, and Ascorbigens Produced in Myrosinase-Catalyzed Hydrolysis of Indole Glucosinolates

Article abstract of DOI:10.1021/jf9708498

Micellar electrokinetic capillary chromatography (MECC) was used to follow the myrosinase (β-thioglucoside glucohydrolase EC 3.2.3.1)-catalyzed transformation of glucobrassicin (indol-3-ylmethylglucosinolate, 1a) and neoglucobrassicin (N-methoxyglucobrassicin, 1b) into nitriles, ascorbigens, and other products. The influence of pH, ascorbic acid, and Fe(II) ions was investigated. In the presence of ascorbic acid, (5 mM), thiocyanate ion and ascorbigens were the dominating products from 1a and 1b. In the presence of Fe(II) ions (2.5 mM), nitriles were the dominating products between pH 4 and 6-7. During hydrolysis of 1b in neutral or weakly basic solution, an unstable intermediate was detected by MECC. Comparisons of the rate of ascorbigen formation from 1a, 1b, and indol-3-ylcarbinol showed that ascorbigens were formed directly from ascorbate and unstable products of the hydrolysis of indole glucosinolates and that indol-3-ylcarbinols were not important intermediates. Structures of 1a, 1b, and products of 1b were confirmed by ¹H NMR, MS, and UV spectroscopy.

Full text of DOI:10.1021/jf9708498

J. Agric. Food Chem. 1998, 46, 1563−1571
1563
In itia l a n d F in a l P r od u cts, Nitr iles, a n d Ascor bigen s P r od u ced in
Myr osin a se-Ca ta lyzed Hyd r olysis of In d ole Glu cosin ola tes
Niels Agerbirk,† Carl Erik Olsen, and Hilmer Sørensen*
Chemistry Department, Royal Veterinary and Agricultural University, Thorvaldsensvej 40,
DK-1871 Frederiksberg C, Denmark
Micellar electrokinetic capillary chromatography (MECC) was used to follow the myrosinase (â-
thioglucoside glucohydrolase EC 3.2.3.1)-catalyzed transformation of glucobrassicin (indol-3-
ylmethylglucosinolate, 1a ) and neoglucobrassicin (N-methoxyglucobrassicin, 1b) into nitriles,
ascorbigens, and other products. The influence of pH, ascorbic acid, and Fe(II) ions was investigated.
In the presence of ascorbic acid, (5 mM), thiocyanate ion and ascorbigens were the dominating
products from 1a and 1b. In the presence of Fe(II) ions (2.5 mM), nitriles were the dominating
products between pH 4 and 6-7. During hydrolysis of 1b in neutral or weakly basic solution, an
unstable intermediate was detected by MECC. Comparisons of the rate of ascorbigen formation
from 1a , 1b, and indol-3-ylcarbinol showed that ascorbigens were formed directly from ascorbate
and unstable products of the hydrolysis of indole glucosinolates and that indol-3-ylcarbinols were
not important intermediates. Structures of 1a , 1b, and products of 1b were confirmed by 1H NMR,
MS, and UV spectroscopy.
Keyw or d s: Indole glucosinolate; myrosinase; ascorbigen; isothiocyanate; nitrile
INTRODUCTION
5a and ascorbic acid in vitro (Kiss and Neukom, 1966),
indol-3-ylcarbinols could possibly be the initial products
of the hydrolysis of indole glucosinolates, forming ascor-
bigens and indole oligomers (Grose and Bjeldanes, 1992,
Agerbirk et al., 1996) by further reactions. However,
intermediacy of 5a in the formation of 4a during
autolysis has not been proved, and it has been suggested
that indol-3-ylmethyl isothiocyanate (3a ) is the reactive
intermediate during autolysis (Preobrazhenskaya et al.,
1993; Bjergegaard et al., 1994).
It is well established that indol-3-ylacetonitriles are
important products of the nonenzymatic degradation of
indole glucosinolates during cooking (Slominski and
Campbell, 1989), but their importance as products of
myrosinase-catalyzed hydrolysis has been debated. In
vitro myrosinase-catalyzed hydrolysis of 1a gave a high
molar yield of indol-3-ylacetonitrile (2a ) near pH 3
(Gmelin and Virtanen, 1961; Latxague et al., 1991), but
the molar yield of 2a at pH values relevant for plant
tissue (pH 5-6) was 1% or below (Schraudolf and
Weber, 1969; Searle et al., 1982). While nitriles were
the dominant glucosinolate products from autolysis (pH
5-6) of aliphatic glucosinolates in cabbage (Daxen-
bichler et al., 1977), only low molar yields of indol-3-
ylacetonitriles were reported from autolysis of several
Brassica vegetables (Slominski and Campbell, 1989).
The high yield of nitriles from autolysis of aliphatic
glucosinolates in fresh vegetables was presumably due
to the effect of endogenous Fe(II) ion (Van Etten and
Tookey, 1979). A stimulation of nitrile formation from
myrosinase-catalyzed hydrolysis of 1a was reported, but
the stimulation was surprisingly weak, as the presence
of Fe(II) ions increased the molar yield of 2a to only 3%
at pH 4-6 (Searle et al., 1982).
The products of the myrosinase (â-thioglucoside glu-
cohydrolase, EC 3.2.3.1)-catalyzed hydrolysis of indole
glucosinolates (indol-3-ylmethylglucosinolates) have at-
tracted attention for a number of reasons: as precursors
of ascorbigens and thiocyanate ion in crushed Brassica
vegetables (Kuta´cek et al., 1960; Gmelin and Virtanen,
1961, 1962; Kiss and Neukom, 1966); as precursors of
phytoalexins (Monde et al., 1994) or plant hormones
(Rausch et al., 1983) in infected plants; and as possible
antinutritional and anticarcinogenic constituents of
oilseed rape and Brassica vegetables (Bradfield and
Bjeldanes, 1987a,b; McDanell et al., 1987, 1988; Sø-
rensen, 1990; J ensen et al., 1991; Loft et al., 1992).
Glucobrassicin (indol-3-ylmethylglucosinolate, 1a ) and
neoglucobrassicin (N-methoxyglucobrassicin, 1b) are
common constituents of rape and various cabbage and
kale species (McDanell et al., 1988).
There is conflicting evidence of the relative abundance
of the several types of hydrolysis products of indole
glucosinolates (Bjergegaard et al., 1994) (Figure 1).
While some authors found that indol-3-ylcarbinols were
the dominant products of indole glucosinolates from
autolysis of crushed Brassica vegetables (Bradfield and
Bjeldanes, 1987a; Slominski and Campbell, 1989), other
authors found that ascorbigens were the dominant
products (McDanell et al., 1987; Preobrazhenskaya et
al., 1993). The poor stability of indol-3-ylcarbinol (5a )
in acidic solution could be one explanation of the
conflicting evidence (Bradfield and Bjeldanes, 1987a),
and use of extraction methods unsuitable for ascorbig-
ens could be another (Preobrazhenskaya et al., 1993).
As ascorbigen (4a ) may be formed from reaction between
The myrosinase-catalyzed hydrolysis of the substi-
tuted indole glucosinolate 1b seems to follow the same
pattern as known for 1a , including formation of ascor-
bigens (Gmelin and Virtanen, 1962; Hanley et al., 1990),
* Corresponding author (tel., +45 35 28 24 32; fax, +45 35
28 20 89).
† Current adress: Danisco Biotechnology, Langebrogade 1,
DK-1001 Copenhagen K, Denmark.
S0021-8561(97)00849-2 CCC: $15.00 © 1998 American Chemical Society
Published on Web 04/03/1998

1564 J. Agric. Food Chem., Vol. 46, No. 4, 1998
Agerbirk et al.
F igu r e 1. Products of the myrosinase-catalyzed hydrolysis of glucobrassicin (1a ) and neoglucobrassicin (1b). The numbering
system in reported NMR spectra is indicated. Key to numbers: 2a , indol-3-ylacetonitrile; 3a , indol-3-ylmethyl isothiocyanate; 4a ,
ascorbigen (A-stereoisomer is shown); 5a , indol-3-ylcarbinol; 6a , di(indol-3-yl)methane; 7a , brassinin; 2b, N-methoxy-2a , etc.; 6b,
di(N-methoxyindol-3-yl)methane; 7b, N-methoxybrassinin. The formation of 2,3-bis(indol-3-ylmethyl)indole from 5a and 6a is
not shown (Agerbirk et al., 1996).
but there are no reports of detailed variation of the
reaction conditions.
in the formation of ascorbigens at the pH values of
crushed plant tissue (pH 5-6) (Schraudolf and Weber,
1969; Daxenbichler et al., 1977).
We have previously developed analytical techniques
for the determination of intact indole glucosinolates
(Michaelsen et al., 1992; Feldl et al., 1994a), as well as
both their monomeric and oligomeric hydrolysis prod-
ucts (Feldl et al., 1994a; Agerbirk et al., 1996) and the
thiocyanate ion (Bjergegaard et al., 1995b).
In the present study, the stimulation of nitrile forma-
tion by Fe(II) ions was reinvestigated. The stimulation
was found to be significantly higher than previously
reported by Searle et al. (1982).
In addition, the yields of indol-3-carbinols, oligomers,
and ascorbigens from myrosinase-catalyzed hydrolysis
of 1a and 1b were determined in the presence or
absence of ascorbic acid. Ascorbigens were found to be
the dominant products. It was shown that indol-3-
ylcarbinols do not transiently accumulate during the
formation of ascorbigens from indole glucosinolates, and
that indol-3-ylcarbinols are not important intermediates
MATERIALS AND METHODS
Refer en ce Com p ou n d s a n d Rea gen ts. Potassium thio-
cyanate (photographic grade), ammonium iron(II)sulfate, N-
acetylcysteine, 5a , and 2a were obtained from Sigma (St.
Louis, MO). Allyl isothiocyanate was from Aldrich-Chemie
(Steinham, Germany). L-Ascorbic acid was from Bie og Bern-
tsen (Rødovre, Denmark). Allylglucosinolate (sinigrin) and
but-3-enylglucosinolate (gluconapin) were from the collection
of reference compounds of this laboratory (Sørensen, 1990).
6a was synthesized as previously described (Agerbirk et al.,
1996). All other reagents were from the laboratory collection
of pro analysi grade chemicals.
Potassium thiocyanate was dried in a vacuum before weigh-
ing. UV: 193 nm (max), 215 nm (shoulder), ꢀ220 ) 3.1 × 103
M-1 cm-1
.
4a was synthesized by reaction of 5a with L-ascorbic acid.
L-Ascorbic acid (1.2 g) and 5a (1.0 g) were dissolved in 250

Myrosinase-Catalyzed Hydrolysis of Indole Glucosinolates
J. Agric. Food Chem., Vol. 46, No. 4, 1998 1565
Ta ble 1. In flu en ce of Neoglu cobr a ssicin (1b)
mL of water and left for 2 h at room temperature. Soluble 4a
was separated from precipitated 6a by filtration. The aqueous
phase was neutralized with NaOH and extracted with 3 × 75
mL of ethyl acetate. After evaporation of the solvent, the
product was dried in a vacuum. Yield: 0.61 g (29%) of 4a (red,
amorphous solid) with a trace of ascorbic acid. Structure
confirmed by 1H NMR, UV, and MS (vide infra). MECC
showed that only one dominant stereoisomeric ascorbigen was
formed (Feldl et al., 1994b).
Sp ectr oscop y. UV spectra were measured in quartz
cuvettes on a Shimadzu MPS-2000 UV-visible light spectro-
photometer. UV data are reported in the following way: λmax
(relative intensity). 1H NMR spectra were recorded at 250
MHz on a Bruker AC250P NMR spectrometer. In deuterated
chloroform and methanol, chemical shifts (δ) are relative to
internal TMS; in water, chemical shifts are relative to dioxane
(3.75 ppm) unless otherwise stated. NMR data are reported
in the following way: chemical shift δH/ppm (relative intensity,
multiplicity, coupling constant J , assignment).
Con cen tr a tion a n d Ascor bic Acid on th e P r od u cts of
Myr osin a se-Ca ta lyzed Hyd r olysis of 1b a t p H 5.6a
conditions
products (% of initial [1b])
sum of indolyl groups
in 5b and 6b
[1b]
(mM)
asc.
ac
5b
6b
12
26
0.6
2.1
17
42
4
4
73
67
41
94
5
0.60
0.43
0.35
0.33
0.13
+
+
2
7
4
8
87
75
a
The analysis method included indole oligomers, but ascorbig-
ens could not be determined quantitatively. Organic solvent (1-
propanol and acetonitrile, see text) was added after the complete
hydrolysis of 1b, and the mixture was analyzed by the MECC-
cholate method. In some experiments, 5 mM ascorbic acid was
present during the hydrolysis, as indicated with a “+” in the
column “asc. ac.”.
Mass spectra were recorded on a J EOL AX505W mass
spectrometer. EI spectra were measured at 70 eV and at a
source temperature of 240 °C. MS data are reported in the
following way: m/z value, assignment (relative abundance).
The direct inlet mass spectra of 4a and 4b and to a lesser
degree 2b revealed the presence of major impurities.
From MECC and NMR, 1b contained 7% 1a and 2% 4-meth-
oxyglucobrassicin, and 1a contained 5% 1b and 4% 4-meth-
oxyglucobrassicin. The apparent ꢀ280 around 5000 M-1 cm-1
of the indole glucosinolate preparations indicated a low content
of inorganic salts etc. (Feldl et al., 1994a).
Myrosinase was purified, characterized, and assayed by use
of previously described procedures (Michaelsen et al., 1991)
from seeds of yellow mustard (Sinapis alba L. cv. Albatros),
obtained from Trifolium A/S (Roskilde, Denmark). The my-
rosinase for use in glucosinolate degradation experiments was
dissolved in the con A elution buffer and stored until use at 4
°C.
Glu cosin ola te Hyd r olysis. Purified glucosinolate (10-
100 nmol) was dissolved in 200 µL of the appropriate buffer,
and 2-4 µL of myrosinase solution (6 units/mL) was added.
The mixture was analyzed by MECC before the addition of
myrosinase and by repeated injections after the addition of
myrosinase. The reported yields of the thiocyanate ion were
corrected for a trace of thiocyanate ion in the glucosinolate
preparation.
MECC. MECC of glucosinolates, thiocyanate ion, and
indolylic monomeric hydrolysis products was performed as
described by Feldl et al. (1994a) but with UV detection at 220
nm. The pH of the buffer was adjusted to 7.0 before the
addition of 1-propanol, at 80% of the final volume. In some
experiments the pH of the buffer was adjusted to 8.0 instead
of 7.0, to slow the hydrolysis of 3b during the electrophoresis
time. To separate 1a and 2a , the temperature was in some
cases increased to 60 °C. MECC of indolyl oligomers was
performed as described by Agerbirk et al. (1996), with UV
detection at 220 nm. Satisfactory peak shape of 4a and 4b
was only accomplished at one set of conditions: DTAB 30 °C.
Resp on se F a ctor s. The response factor of thiocyanate ion
was determined by MECC (DTAB, 30 °C) of solutions of pure
KSCN. Molar response factors of indole derivatives at 220 nm
in MECC (DTAB, 30 °C) relative to 2a were measured in three
different ways: (1) from the relative magnitude of ꢀ220
,
Glucosinolate hydrolysis at all the reported conditions was
followed by repeated MECC analyses. The temperature of the
hydrolysis mixture was 30 °C, which was the temperature of
the autosampler. Before the measurements reported in Table
1, but after the complete hydrolysis of 1b, 1-propanol-
acetonitrile (3:2) was added to 16%, in an attempt to keep the
oligomers in solution. In Table 1, the analysis result of the
first measurement (10 min after the addition of myrosinase)
is reported, and it was confirmed by additional measurements
that the concentrations of both 5b and 6b were unchanged
during longer incubation. Formation of the N-methoxyindol-
3-ylmethylpropyl ether (neoIPE; Agerbirk et al., 1996) was
insignificant when 1-propanol was added after the complete
hydrolysis of 1b.
The analysis result after 30-60 min, or after the complete
disappearance of the unstable intermediate 3b, is reported in
Table 2. Some experiments were repeated, and representative
results are reported. For the analysis of indole oligomers,
addition of organic solvent was necessary to keep the hydro-
phobic oligomers in solution (Agerbirk et al., 1996).
The buffers for glucosinolate hydrolysis were either 200 mM
Tris-HCl (pH 7.2), 200 mM Na2HPO4-NaH2PO4 (pH 7 (100
mM) or pH 8 (200 mM)), 50 mM maleic acid-sodium maleate
(pH 6), 100 or 200 mM acetic acid-sodium acetate (pH 5.6 or
4), or 100 mM formic acid-sodium formiate (pH 3.5). The Fe-
(II) ion was added as FeSO4 or (NH4)2Fe(SO4)2. Phosphate
buffer was avoided whenever iron salts were added. Tris
buffer was avoided whenever reaction with isothiocyanates
could occur. Cysteine was added as the hydrochloride. Buffers
containing reducing compounds were prepared fresh. The final
pH of the hydrolysis buffers with added reagents was mea-
sured and adjusted by titration with NaOH.
assuming identical ꢀ280 of monomeric indole compounds (Feldl
et al., 1994a); (2) from UV spectroscopy (280 nm) and MECC
of pure compounds; and (3) from the normalized areas of
substrate and product peaks in MECC, using an argument of
stoichiometry. Standards of 5a and 5b were dissolved in
phosphate buffer pH 8, and the compounds were stable at this
pH. All reported yields of monomeric indole derivatives
obtained with the DTAB method were calculated using the
response factor found for 2a . With the cholate method, yields
of 5b and 6b were calculated from response factors obtained
with standards of 5a and 6a (Agerbirk et al., 1996).
Determination of SCN- in the presence of Fe(II) and Fe-
(III) ions. The influence of the reaction conditions on the
quantitative determination of thiocyanate ion by MECC was
tested. Neither Fe(II) ions, Fe(III) ions, ascorbic acid, the
various buffer solutions used for glucosinolate hydrolysis (pH
3.5, 4, 5.6, or 7), nor strongly acidic or strongly basic conditions
influenced the determination.
Glu cosin ola tes. 1a and 1b were purified from frozen
broccoli (Brassica oleracea L. var. botrytis L. subvar. cymosa
Lam.) bought in a food store, by a procedure modified from
Thies (1988) and Visentin et al. (1992). While nitrate, formi-
ate, and the aliphatic glucosinolate 4-(methylsulfinyl)butyl-
glucosinolate were eluted from the ion exchange column with
the salt front, 1b, 4-methoxyglucobrassicin and 1a adsorbed
more strongly to the column material and were eluted in the
order mentioned. Fractions pure in 1b and in 1a were
obtained and desalted, and they contained 22 µmol (34% yield)
of 1a and 10 µmol (20% yield) of 1b, as potassium salts
(Agerbirk, 1997). The purity was tested by 1H NMR and
quantitative UV spectroscopy, by HPLC and MECC of des-
ulfoglucosinolates (EU, 1990; Bjergegaard et al., 1995a), and
by MECC of intact glucosinolates (Michaelsen et al., 1992).
P u r ifica tion of Hyd r olysis P r od u cts: 2b. 1b (1 µmol)
was hydrolyzed by myrosinase in aqueous 5 mM cysteine, 2.5

1566 J. Agric. Food Chem., Vol. 46, No. 4, 1998
Agerbirk et al.
Ta ble 2. Yield s of Th iocya n a te Ion a n d Mon om er ic
In d olyl Com p ou n d s fr om Myr osin a se-Ca ta lyzed
Hyd r olysis of 1a a n d 1b in th e P r esen ce of 5 m M
Ascor bic Acid a
Ch a r a cter iza tion of Myr osin a se. Myrosinase used
in this study was isolated from seeds of Sinapis alba.
The native myrosinase behaved like the dimeric myro-
sinase isolated from seeds of Brassica napus during gel
filtration. IEF indicated that the myrosinase prepara-
tion consisted of a number of isoenzymes, with pI values
from 5.6 to 6.2. The specific activity (Michaelsen et al.,
1991) of the purified enzyme was 5 units/mg. The
molecular weight of the subunits was 73 kDa as
determined by reducing SDS-PAGE. The lower pH
limit of activity of the enzyme was near pH 3.5. It
hydrolyzed both allyl- and but-3-enylglucosinolate to the
corresponding isothiocyanates at neutral pH, as shown
by conversion of the products of hydrolysis into the
corresponding substituted thiourea compounds in 70-
80% of the theoretical yields (results not shown). In the
experiments with 1a and 1b reported below, total
hydrolysis of the glucosinolates was always observed at
the end of each incubation (e.g. Figure 2A).
In itia l P r od u cts fr om In d ole Glu cosin ola te Hy-
d r olysis. In the presence of Fe(II) ions at 2.5 mM,
which were protected from oxidation by 5 mM cysteine
added to the buffers, the indol-3-ylacetonitriles were the
dominant products formed at pH 4 and 5.6, and were
formed to a considerable extent even in neutral solution
(Figures 2 and 3). The yield of thiocyanate ion was
lowered quantitatively in response to nitrile formation,
except for a few deviating observations (Figure 3). The
formation of the indol-3-ylacetonitriles was dependent
on both myrosinase and Fe(II) ions, as no degradation
of the glucosinolate was detectable during incubations
of the glucosinolates in the reaction buffer with cysteine
and Fe(II) ions, but without myrosinase.
The identity of 2b was confirmed by various spectro-
scopic methods (Table 3), while the identities of 2a and
thiocyanate ion were confirmed by spiking with refer-
ence compounds.
In the absence of Fe(II) ions, the yield of nitriles was
low at pH higher than 4. Extraction of the isolated
preparation of 1a with ethyl acetate prior to the hy-
drolysis resulted in an apparent lowering of the yield
of 2a between pH 4 and 7, suggesting that the majority
of the yield of 2a and 2b reported here (Figure 3), in
the absence of Fe(II) ions, was due to contamination of
the glucosinolate preparations with small amounts of
the corresponding nitrile.
Thiocyanate ion, 5a , 5b, or indolyl oligomers were in
general detected immediately following the hydrolysis
of the indole glucosinolates in the absence of ascorbic
acid (Figure 4B; Table 1). 6b was formed even at pH
5.6, where 5b was stable. The identities of 5a and 6a
were confirmed by spiking with reference compounds,
while the identity of 5b and 6b were tentatively
confirmed by MS, UV spectroscopy, and instability of
5b in dilute acid (results not shown). Taken together,
the yields of 5b and 6b accounted for the majority of
the hydrolyzed glucosinolate (except for the highest
concentration, where precipitation probably occurred),
but the relative amounts depended on the initial con-
centration of glucosinolate (Table 1).
An Un sta ble In ter m ed ia te fr om Hyd r olysis of
Neoglu cobr a ssicin (1b). Under certain conditions, we
saw exceptions to the immediate formation of thiocy-
anate ion and 5b from 1b in the absence of ascorbic acid.
In phosphate buffer at pH 7, a trace of an unstable
intermediate was detected by MECC (DTAB 30 °C). The
intermediate was not detected at lower pH. In phos-
phate buffer pH 8, the intermediate was more promi-
conditions
products
(% of initial [glucosinolate])
glucosinolate
(concn/mM)
pH
SCN-
2a
4a
5a
7.0
5.6
4.0
1a (0.67)
1a (0.33)
1a (0.63)
61
73
59
tr.
tr.
2.7
27
37
36
2.5
3.0
2.7
2b
4b
5b
7.0
5.6
4.0
1b (1.30)
1b (1.30)
1b (0.58)
80
94
84
<1.5
<1.2
<4.3
34
38
36
13
8
9
a
The pH was varied as indicated. At pH 7, phosphate buffer
was used, at pH 5.6 and 4, acetate buffer was used. MECC
conditions: DTAB (30 °C).
mM FeSO4, 100 mM AcOH-NaOH (pH 4), yielding SCN-, 5b,
and 0.7 µmol of 2b. Upon incubation overnight, 5b disap-
peared from the solution. After centrifugation, the superna-
tant was neutralized with Na2HPO4, and iron phosphate
removed by centrifugation. 2b was extracted with diethyl
ether, yielding 0.3 µmol of 2b. The product was pure by
MECC, NMR, and MS, except for some aliphatic hydrocarbons
detectable in MS and NMR, and a small impurity giving rise
to an ion at 530 m/z in the EI mass spectrum.
4b. 1b (0.7 µmol) was hydrolyzed by myrosinase in a pH 4
acetate buffer with 5 mM L-ascorbic acid. Extraction with
ethyl acetate gave 0.2 µmol of 4b, which was characterized by
NMR and MS.
Un k n ow n P r od u cts 1 a n d 2. The unknown reaction
products of N-acetylcysteine and hydrolysis products of 1b
were prepared as described in Results and loaded on a
Sephadex DEAE A-25 column (acetate). The unknowns were
eluted with 0.2 M potassium sulfate, and individual fractions
were analyzed by MECC and UV spectroscopy.
RESULTS
Figure 1 illustrates the myrosinase-catalyzed hy-
drolysis of the indole glucosinolates, at the various
conditions used in the present work. The mechanisms
of formation of ascorbigens (4a and 4b) and of 6b, which
were deduced from the time course experiments (vide
infra), are shown.
Resp on se F a ctor s of Glu cosin ola tes a n d Th eir
P r od u cts (Con d ition s: DTAB, 30 °C). The molar
response factor of thiocyanate ion was 0.10 relative to
the molar response factor of IAN.
From their nearly similar ꢀ220, the molar response
factors of most monomeric indole derivatives were
assumed to be similar. When the response factors of
the glucosinolates were calculated from the amount of
thiocyanate ion produced, the indole glucosinolates
seemed to have the same response factors as the other
investigated indole derivatives. When the response
factors were measured by MECC and UV spectroscopy
of purified compounds, 4a and an impure preparation
of 5b had significantly lower molar response factors
than 2a , while pure 5a had exactly the same response
factor as 2a . A lower response factor for 4a than for
other indole derivatives was possibly due to poor
chromatographic peak shape (Figure 2). Thus, the data
as a whole suggested that the response factors of
monomeric indole derivatives, with the possible excep-
tion of 4a and 4b, were roughly similar. All reported
yields in Table 2 were calculated using the response
factor found for 2a , meaning that reported yields of
ascorbigens may be moderately underestimated.

Myrosinase-Catalyzed Hydrolysis of Indole Glucosinolates
J. Agric. Food Chem., Vol. 46, No. 4, 1998 1567
F igu r e 2. MECC chromatograms of neoglucobrassicin (1b) and its hydrolysis products, showing the effect of the reaction conditions
on the products formed. Small, unlabeled peaks represent the corresponding products from hydrolysis of contaminating 1a . (A-
D) Reaction conditions as indicated.
F igu r e 3. Influence of pH and Fe(II) ions on molar yields of (A) nitriles A and (B) thiocyanate ion from hydrolysis of glucobrassicin
(1a ) and neoglucobrassicin (1b). Data from both 1a and 1b hydrolyses were used.
nent. If the pH of the MECC buffer was likewise
changed to pH 8, the peak representing the unstable
intermediate was further enlarged (Figure 2B). A study
of the time course of the formation of the carbinol, the
thiocyanate ion, and the unstable intermediate (Figure
4A) suggested that the intermediate could be N-meth-
oxyindol-3-ylmethyl isothiocyanate (3b). 1a was also
degraded at pH 8 to investigate whether an unstable
intermediate would appear during the degradation of
this glucosinolate, but in this case the carbinol was the
first product detected (Figure 4B).
In Tris buffer (400 mM, pH 8), the intermediate was
not observed, the molar yield of 5b was reduced from
31 to 11%, while the yield of thiocyanate ion was
unaltered. This result indicated a reaction between the
intermediate and the amine Tris. Similarly, in the
presence of the thiol N-acetylcysteine (at 10, 40, or 200
mM), two new products were formed instead of 5b
(Figure 2C). After purification, the two unknown
products were characterized by UV spectroscopy. The
UV spectrum of the unknown product 2 was similar to
the reported UV spectrum of 7b (Takasugi et al., 1988).
F or m a tion of Ascor bigen s. When ascorbic acid
was present in the hydrolysis mixture at 5 mM, the
yields of both 5b and 6b were reduced 10-fold or more
(Table 1). When the hydrolysis into monomeric indole
derivatives was followed by MECC with DTAB as
surfactant, the yields of the ascorbigens from both 1a
and 1b was found to be higher than the yields of 5a
and 5b (Table 2). Formation of oligomers was not
quantitatively detected in this experiment, as organic
solvent was not added to the reaction mixture. Yields
of oligomers are therefore not reported in Table 2. The
identity of 4a was confirmed by spiking with an
authentic standard, while the identity of 4b was ten-
tatively confirmed by MS, NMR, and UV spectroscopy
(Table 3). The molar yields of the ascorbigens were
always ca. 3-10-fold higher than the molar yields of
indol-3-ylcarbinols, even at pH values where the latter

1568 J. Agric. Food Chem., Vol. 46, No. 4, 1998
Agerbirk et al.
Ta ble 3. Sp ectr oscop ic Da ta of In d ole Glu cosin ola tes a n d P u r ified Hyd r olysis P r od u cts of 1b (Sp ectr a of Au th en tic
Hyd r olysis P r od u cts of 1a Ar e Sh ow n in Ad d ition for Com p a r ison )
glucobrassicin (1a )
UV (H2O): 219 nm (100), 270 nm (shoulder, 15), 279 nm (16), 286 nm (shoulder, 13)
1H NMR (D2O, int. st. H2O at 4.75 ppm): 2.94 (1 H, d tr, J 5′-4′ ) 10 Hz, J 5′-6′ ) 4 Hz, 5′-H), 3.15-3.4 (3 H, m, 2′-H, 3′-H, 4′-H),
3.57 (2H, d, 4 Hz, 6′-H), 4.18 (1H, d, 16 Hz, 1-Ha or 1-Hb), 4.29 (1H, d, 16 Hz, 1-Ha or 1-Hb), 4.81 (1H, d, 10 Hz, 1′-H),
7.20 (1H, tr, 8 Hz, 5′′-H), 7.28 (1H, tr, 8 Hz, 6′′-H), 7.34 (1H, s, 2′′-H), 7.54 (1H, d, 8 Hz, 7′′-H), 7.75 (1H, d, 8 Hz, 4′′-H)
neoglucobrassicin (1b)
UV (H2O): 221 nm (100), 274 nm (14), 288 nm (14)
1H NMR (D2O): 2.99 (1H, d tr, J 5′-4′ ) 10 Hz, J 5′-6′ ) 4 Hz, 5′-H), 3.15-3.4 (3H, m, 2′-H, 3′-H, 4′-H), 3.55 (2H, irregular doublet,
4 Hz, 6′-H), 4.11 (3H, s, OCH3), 4.16 (1H, d, 16 Hz, 1-Ha or 1-Hb), 4.25 (1H, d, 16 Hz, 1-Ha or 1-Hb), 4.81 (1H, d, 10 Hz, 1′-H),
7.25 (1H, tr, 8 Hz, 5′′-H), 7.37 (1H, tr, 8 Hz, 6′′-H), 7.53 (1H, s, 2′′-H), 7.59 (1H, d, 8 Hz, 7′′-H), 7.77 (1H, d, 8 Hz, 4′′-H)
indol-3-ylacetonitrile (2a )
UV (H2O): 216 nm (100), 270 nm (16), 277 nm (17), 286 nm (12)
1H NMR (CDCl3): 3.83 (1H, d, 1 Hz, 1-H), 7.18 (1H, tr, 8 Hz, 5′′-H), 7.25 (1H, s, 2′′-H), 7.26 (1H, tr, 8 Hz, 6′′-H), 7.40 (1H, d, 7′′-H,
8 Hz), 7.59 (1H, d, 4′′-H, 8 Hz), 8.2 (1H, broad s, NH)
N-methoxyindol-3-ylacetonitrile (2b)
UV (H2O): 219 nm (100), 270 nm (18), 283 nm (18), 295 (shoulder)
EI-MS m/z (%): 186, M+ (87); 155, (M - OCH3)+ (100); 128, (M - OCH3 - HCN)+ (49)
1H NMR (CDCl3): 3.81 (2H, s, 1-H), 4.09 (3H, s, OCH3), 7.17 (1H, tr, 8 Hz, 5′′-H), 7.31 (1H, tr, 8 Hz, 6′′-H), 7.32 (1H, s, 2′′-H),
7.45 (1H, d, 8 Hz, 7′′-H), 7.56 (1H, d, 8 Hz, 4′′-H), and additional long-range couplings between aromatic protons
ascorbigen (4a )
EI-MS m/z (%): 305, M+ (10); 130, (M-C6O6H7)+ (55); 116, (M-C6O6H7 -CH2)+ (100)
1H NMR (CD3OD): 3.22 (1H, d, J ab ) 14 Hz, 1-Ha or 1-Hb), 3.40 (1H, d, J ab ) 14 Hz, 1-Ha or 1-Hb), 3.77 (1H, s, (J LX < 1 Hz),
HX), 3.98 (1H, dd, J AB ) 10 Hz, J AL ) 3.1 Hz, HA), 4.12 (1H, dd, J AB ) 10 Hz, J BL ) 5.6 Hz, HB), 4.20 (1H, dd, J BL ) 5.6 Hz,
J AL ) 3.1 Hz, (J LX < 1 Hz), HL), 6.97 (1H, tr, 8 Hz, 5′′-H), 7.06 (1H, tr, 8 Hz, 6′′-H), 7.19 (1H, s, 2′′-H), 7.31 (1H, d, 8 Hz, 7′′-H),
7.62 (1H, d, 8 Hz, 4′′-H), and additional long-range couplings in aromatic system and 2′′-H and 1-Ha/Hb
neoascorbigen (4b)
EI-MS m/z (%): 335, M+ (13); 160, (M-C6O6H7)+ (100)
1H NMR (CD3OD, int. std. TMS): 3.20 (1H, d, 14 Hz, 1-Ha or 1-Hb), 3.89 (1H, s, HX), 4.05 (3H, s, OCH3), 7.02 (1H, tr, 8 Hz, 5′′-H),
7.17 (1H, tr, 8 Hz, 6′′-H), 7.33 (1H, s, 2′′-H), 7.36 (1H, d, 8 Hz, 7′′-H), 7.61 (1H, d, 8 Hz, 4′′-H), some peaks from impurities were
also present and additional long-range couplings between aromatic protons were seen; signal at 3.20 ppm (1-H) showed coupling
to a proton further downfield, which seemed to be covered by the signal of the solvent at 3.3 ppm
unknown product 1
UV (H2O): ca. 220 nm (shoulder, 100), 270 nm (18), 290 nm (shoulder, 14)
unknown product 2
UV (H2O): 218 (100), 250 (shoulder, 36), 269 (39)
F igu r e 4. Time course of hydrolysis of (A) neoglucobrassicin (1b) and (B) glucobrassicin (1a ) at pH 8, illustrating the formation
and disappearance of an unstable intermediate from 1b hydrolysis. No such intermediate was detected from hydrolysis of 1a .
MECC conditions: DTAB 30 °C, pH 8.
DISCUSSION
are perfectly stable (5a , pH 7; 5b: pH 5.6 and 7;
Agerbirk et al. (1996) and unpublished results).
The formation of ascorbigens from 1a and 1b was
followed by repeated analysis of the reaction mixture
during the glucosinolate hydrolysis. The ascorbigens
were formed immediately after glucosinolate hydro-
lysis, without prior accumulation of the indol-3-yl-
carbinols as intermediates (Figure 5A-C). To test
whether 5a could be short-lived intermediate in ascor-
bigen formation, the formation of 4a from pure 5a and
ascorbic acid was investigated (Figure 5D). 4a was not
formed from 5a and ascorbic acid at pH 7, but it was
formed relatively slowly at pH 5.6, and was formed
rapidly at pH 4.
Myr osin a se. Myrosinase used in the present inves-
tigation was isolated from seeds of S. alba and tested
by standard procedures (Michaelsen et al., 1991). From
the subunit molecular weight and extraction method,
it seemed to be a member of the so-called MA or Myr 1
family of myrosinases (Xue, 1994; Bones and Rossiter,
1996).
Id en tifica tion of Glu cosin ola tes a n d Hyd r olysis
P r od u cts. The identifications of 1a and 1b by chro-
matography and spectroscopy were unambiguous and
were in agreement with previously reported NMR
spectra of indole glucosinolates (Hanley et al., 1990;

Myrosinase-Catalyzed Hydrolysis of Indole Glucosinolates
J. Agric. Food Chem., Vol. 46, No. 4, 1998 1569
F igu r e 5. Time course of the formation of ascorbigen (4a ) from glucobrassicin (1a ) (A-C) or from indol-3-ylcarbinol (5a ) (D) at
varying pH, showing that 5a was not an important intermediate in the formation of 4a .
Viaud et al., 1992). The identification of 2b was
likewise unambiguous. The position of the methoxy
group at the heteroatom was supported by the absence
of an NH signal in the NMR spectrum of 2b compared
to the spectrum of 2a .
a nitrile without rearrangement (Ettlinger and Lun-
deen, 1956, 1957; Ettlinger et al., 1961; Miller, 1965;
Saarivirta, 1973; Uda et al., 1986; Iori et al., 1996). The
formation of either nitrile or isothiocyanate, depending
on pH and the presence or absence of Fe(II) ions, is
consistent with this mechanism.
The identification of 4b was tentative and was based
on the mass spectrum, analogy with the formation of
4a , and an NMR spectrum showing major impurities.
In support of the proposed structure, a doublet repre-
senting one of the C(1)H2 protons was detected, while
the other seemed to be masked by the solvent signal.
The coupling constant of this doublet (14 Hz) was in
agreement with the corresponding coupling constant of
4a and a published NMR spectrum of synthetic 4b
(Muchanov et al., 1994). Signals of the N-methoxyindol-
3-yl group were unambiguously identified. A singlet
representing HX of the ascorbic acid residue was de-
tected with δ 3.89 (δ 3.79 reported by Muchanov et al.
(1994)), while the expected multiplets of HA, HB, and
HL of the ascorbic acid residue were not clearly identi-
fied, due to the very low amount of 4b analyzed.
Gen er a l Rea ction Mech a n ism a n d F or m a tion of
Nitr iles. The well-established general mechanism of
glucosinolate hydrolysis involves hydrolysis of the â-thio-
glucoside bond by myrosinase. Nonenzymatic degrada-
tion of the aglucone, a thiohydroxamate-O-sulfonate or
thiohydroxamic acid-O-sulfonate depending on pH, yields
either an isothiocyanate via Lossen rearrangement, or
In the case of the indole glucosinolates, rearrange-
ment to the isothiocyanate was proved by isolation of
3b from myrosinase-catalyzed hydrolysis of 1b at “low
water conditions” (Hanley et al., 1990) and by trapping
3a by methyl thiolate, forming the phytoalexin 7a
(Monde et al., 1994). The reported increased formation
of nitrile at low pH was also in agreement with hydroly-
sis of indole glucosinolates to the thiohydroxamate-O-
sulfonate, while the atypical, low stimulation of nitrile
formation by Fe(II) ions was unexpected. The strong
stimulation of nitrile formation from myrosinase-
catalyzed hydrolysis of 1a and 1b reported here (Figure
4), at pH 4-7 in the presence of Fe(II) ions, supports
that the primary product of the myrosinase-catalyzed
hydrolysis of a (substituted) indole glucosinolate is a
(substituted) indol-3-ylmethylthiohydroxamate-O-sul-
fonate, analogous to the hydrolysis of aliphatic glucosi-
nolates.
The role of the Fe(II) ions in the formation of nitriles
is not fully understood, but interference with the Lossen
rearrangement caused by affinity of Fe(II) ions for the
sulfide group of the primary product is one possible

1570 J. Agric. Food Chem., Vol. 46, No. 4, 1998
Agerbirk et al.
explanation. Control experiments without myrosinase
showed that the formation of nitriles was dependent on
both Fe(II) ions and myrosinase, excluding Fe(II)-
catalyzed hydrolysis of the glucosinolates to be signifi-
cant at our conditions.
conditions were 5b is stable (Table 1). The isothio-
cyanatessand not the carbinolssmay also be the true
precursors of the carbonium ion intermediate during
ascorbigen formation, as suggested from the time course
of the formation of 4a and 4b (vide supra).
The very low stimulation of 2a formation by Fe(II)
ions reported by Searle et al. (1982), can be attributed
to the use of a citrate-phosphate buffer, which will
complex-bind Fe(II) ions. Due to a probable contamina-
tion of our glucosinolate preparations with trace amounts
of the corresponding nitriles, the yields reported here
in the absence of Fe(II) ions are maximum estimates
but do not contradict much lower yields reported by
other workers (Schraudolf and Weber, 1969; Searle et
al., 1982).
Formation of so-called abnormal hydrolysis products
(Benn, 1977) depends on additional proteins or factors
(Hasapis and MacLeod, 1982; Cole, 1980). Formation
of organic thiocyanates (known for allyl-, benzyl-, and
4-(methylthio)butylglucosinolates only) seems to be as-
sociated with formal ability to form a resonance-
stabilized carbonium ion upon formal loss of thiocyanate
ion (Lu¨thy and Benn, 1977). The indole glucosinolates
could thus a priori be able to form organic thiocyanates.
The fact that allylglucosinolate formed the correspond-
ing isothiocyanate in high yield upon hydrolysis cata-
lyzed by our myrosinase preparation (result not shown)
showed that the purified myrosinase did not contain
factors with ability to change the product to organic
thiocyanates. Involvement of indol-3-ylmethyl thiocy-
anates in the formation of the various products of indole
glucosinolate hydrolysis can thus be excluded in our
experiments.
F or m a tion of Ascor bigen s. The time course in-
vestigations of ascorbigen formation from either 5a or
from myrosinase-catalyzed hydrolysis of 1a or 1b,
suggested that 5a and 5b were not important as
intermediates in the formation of 4a and 4b (Figure 5).
Instead, the ascorbigens may have been formed from
reaction between either the isothiocyanate (vide infra)
or the carbonium ion, in competition with the formation
of indol-3-ylcarbinols from reaction with water (Figure
1). This reaction mechanism was also suggested on a
theoretical basis (Preobrazhenskaya et al., 1993; Bjerge-
gaard et al., 1994). At lower pH, where the isothiocy-
anate is less stable and the carbinol more reactive, it
could not be concluded from the time course experiments
whether 5a was important as a reaction intermediate.
In any case, neither 5a nor 5b accumulated during
glucosinolate hydrolysis at low pH (pH 4) prior to
ascorbigen formation (Figure 5 C).
Un sta ble In ter m ed ia te fr om Neoglu cobr a ssicin
(1b). On the basis of the previous isolation of 3b from
hydrolysis of 1b at “low water conditions” (Hanley et
al., 1990) and the formation of allyl isothiocyanate, not
allyl thiocyanate, from allylglucosinolate by our myro-
sinase preparation, we suggest that the reaction inter-
mediate observed during hydrolysis of 1b at pH 8 could
be the isothiocyanate (Figures 2 and 5A). The high
reactivity of the intermediate with nucleophilic com-
pounds supports this suggestion. The stabilizing influ-
ence of the N-methoxy group of 3b (Hanley et al., 1990)
explains our inability to detect the corresponding isothio-
cyanate from 1a (Figure 4).
ABBREVIATIONS USED
DTAB, dodecyltrimethylammonium bromide; EI,
electron impact, MECC, micellar electrokinetic capil-
lary chromatography; Tris, tris(hydroxymethyl)ami-
nomethane.
ACKNOWLEDGMENT
Sandro Palmieri is gratefully acknowledged for a stay
of one of us (N.A.) at Istituto Sperimentale per le Colture
Industriali, Bologna, Italy. The skilled technical as-
sistance of S. Cinti, K. Mortensen, J . Hallick, and P.
Møller is acknowledged.
LITERATURE CITED
Agerbirk, N. Products of the myrosinase catalysed hydrolysis
of glucosinolates in Brassica vegetables. Thesis, Royal
Veterinary and Agricultural University, Frederiksberg,
Denmark, 1997; pp 159-174.
Agerbirk, N.; Bjergegaard, C.; Olsen, C. E.; Sørensen, H.
Kinetic investigations of the transformations of indol-3-
ylcarbinol into oligomeric compounds based on micellar
electrokinetic capillary chromatography. J . Chromatogr. A
1996, 745, 239-248.
Benn, M. Glucosinolates. Pure Appl. Chem. 1977, 49, 197-
210.
Bjergegaard, C.; Li, P. W.; Michaelsen, S.; Møller, P.; Otte, J .;
Sørensen, H. Glucosinolates and their transformation prod-
ucts-compounds with a broad biological activity. Proc. Int.
Euro Food Tox IV Conference; H. Kozlowska, H., Fornal, J .,
Zdunczyk, Z., Eds.; 1994; Vol. I, pp 1-15 (ISBN 83-901475-
0-5).
Bjergegaard, C.; Michaelsen, S.; Møller, P.; Sørensen, H.
Separation of desulphoglucosinolates by micellar electroki-
netic capillary chromatography based on a bile salt. J .
Chromatogr. A 1995a , 717, 325-333.
Bjergegaard, C.; Møller, P.; Sørensen, H. Determination of
thiocyanate, iodide, nitrate and nitrite in biological samples
by micellar electrokinetic capillary chromatography. J .
Chromatogr. A 1995b, 717, 409-414.
Bones, A. M.; Rossiter, J . T. The myrosinase-glucosinolate
system, its organisation and biochemistry. Physiol. Plant.
1996, 97, 194-208.
Bradfield, C. A.; Bjeldanes, L. F. High-performance liquid
chromatographic analysis of anticarcinogenic indoles in
Brassica oleracea. J . Agric. Food Chem. 1987a , 35, 46-49.
Bradfield, C. A.; Bjeldanes, L. F. Dietary modification of
xenobiotic metabolism: contribution of indolylic compounds
present in Brassica oleracea. J . Agric. Food Chem. 1987b,
35, 896-900.
Cole, R. Volatile components produced during ontogeny of some
cultivated crucifers. J . Sci. Food Agric. 1980, 31, 549-557.
Daxenbichler, M. E.; Van Etten, C. H.; Spencer, G. F. Glu-
cosinolates and derived products in cruciferous vegetables.
Identification of organic nitriles from cabbage. J . Agric. Food
Chem. 1977, 25, 121-124.
Ettlinger, M. G.; Lundeen, A. J . The structures of sinigrin and
sinalbin; an enzymatic rearrangement. J . Am. Chem. Soc.
1956, 78, 4172-4173.
Ettlinger, M. G.; Lundeen, A. J . First synthesis of a mustard
oil glucoside; the enzymatic Lossen rearrangement. J . Am.
Chem. Soc. 1957, 79, 1764-1765.
Ettlinger, M. G.; Dateo, G. P., J r.; Harrison, B. W.; Mabry, T.
J .; Thompson, C. P. Vitamin C as a coenzyme: The hydroly-
sis of mustard oil glucosides. Proc. Natl. Acad. Sci. U.S.A.
1961, 47, 1875-1880.
The existence of the short-lived isothiocyanate inter-
mediate also at lower pH is suggested from the observed
immediate formation of 6b from 1b hydrolysis, at

Myrosinase-Catalyzed Hydrolysis of Indole Glucosinolates
J. Agric. Food Chem., Vol. 46, No. 4, 1998 1571
EU. Determination of glucosinolates in oilseeds by liquid
chromatography (HPLC). J . Eur. Communities 1990, L 170,
(03.07), 27-34.
Feldl, C.; Møller, P.; Otte, J .; Sørensen, H. Micellar electro-
kinetic capillary chromatography for determination of in-
dolyl glucosinolates and transformation products thereof.
Anal. Biochem. 1994a , 217, 62-69.
Feldl, C.; Møller, P.; Olsen, C. E.; Otte, J .; Sørensen, H.
Structure and properties of ascorbigens and other transfor-
mation products of indolyl glucosinolates; potential anti-
cancerogens. GCIRC Bull. 1994b, 10, 128-133.
Gmelin, R.; Virtanen, A. I. Glucobrassicin, der Precursor von
SCN-, 3-Indolylacetonitril und Ascorbigen in Brassica ol-
eracea Species. Ann. Acad. Scientiarum Fennicae, Ser. A,
II. Chem. 1961, 107, 1-25.
Gmelin, R.; Virtanen, A. I. Neoglucobrassicin, ein zweiter
SCN--Precursor vom Indoltyp in Brassica-Arten. Acta Chem.
Scand. 1962, 16, 1378-1384.
Grose, K. R.; Bjeldanes, L. F. Oligomerisation of indole-3-
carbinol in aqueous acid. Chem. Res. Toxicol. 1992, 5, 188-
193.
Hanley, A. B.; Parsley, K. R.; Lewis, J . A.; Fenwick, G. R.
Chemistry of indole glucosinolates: Intermediacy of indol-
3-ylmethyl isothiocyanates in the enzymic hydrolysis of
indole glucosinolates. J . Chem. Soc., Perkin Trans. 1, 1990,
2273-2276.
Hasapis, X.; MacLeod, A. J . Benzyl glucosinolate degradation
in heat treated Lepidium sativum seeds and detection of a
thiocyanate forming factor. Phytochemistry 1982, 21, 1009-
1013.
Miller, H. E. The aglucone of sinigrin. Thesis, Rice University,
Houston, Texas, 1965.
Monde, K.; Takasugi, M.; Ohnishi, T. Biosynthesis of crucif-
erous phytoalexins. J . Am. Chem. Soc. 1994, 116, 6650-
6657.
Muchanov, V. I.; Kaganski, M. M.; Sorokin, A. A.; Antonian,
S. G.; Tananova, G. V.; Mikhailevskaya, L. L.; Kinsirski, A.
S.; Preobrazhenskaya, M. N. Neoascorbigen and its ana-
logues: Synthesis and study. Khim.-Farm. Zh. 1994, 28 (7),
6-10.
Preobrazhenskaya, M. N.; Bukhman, V. M.; Korolev, A. M.;
Efimov, S. A. Ascorbigen and other indole-derived com-
pounds from Brassica vegetables and their analogues as
anticarcinogenic and immunomodulating agents. Pharmac.
Ther. 1993, 60, 301-313.
Rausch, T.; Butcher, D. N.; Hildenberg, W. Indol-3-methyl-
glucosinolate biosynthesis and metabolism in clubroot dis-
eased plants. Physiol. Plant. 1983, 58, 93-100.
Saaravirta, M. The formation of benzylcyanide, benzylthiocy-
anate, benzylisothiocyanate and benzylamine from benzyl-
glucosinolate in Lepidium. Planta Med. 1973, 24, 112-119.
Schraudolf, H.; Weber, H. IAN-Bildung aus Glucobrassicin:
pH Abha¨ngigkeit und wachstumsphysiologische Bedeutung.
Planta 1969, 88, 136-143.
Searle, L. M.; Chamberlain, K.; Rausch, T.; Butcher, D. N. The
conversion of 3-indolylmethylglucosinolate to 3-indolylac-
etonitrile by myrosinase, and its relevance to the clubroot
disease of the Cruciferae. J . Exp. Bot. 1982, 33, 935-942.
Slominski, B. A.; Campbell, L. D. Formation of indole glucosi-
nolate breakdown products in autolyzed, steamed and
cooked Brassica vegetables. J . Agric. Food Chem. 1989, 37,
1297-1302.
Sørensen, H. Glucosinolates: StructuresPropertiessFunction.
In Canola and Rapeseed: Production, Chemistry, Nutrition
and Processing Technology; Shahidi, F., Ed.; Van Nostrand
Reinhold: New York, 1990; ISBN 0-442-00295-5.
Takasugi, M.; Monde, K.; Katsui, N.; Shirata, A. Novel Sulfur-
Containing Phytoalexins from the Chinese Cabbage Bras-
sica campestris L. ssp. pekinensis (Cruciferae). Bull. Chem.
Soc. J pn. 1988, 61, 285-289.
Iori, R.; Rollin, P.; Streicher, H.; Thiem, J .; Palmieri, S. The
myrosinase-glucosinolate interaction mechanism studied
using some synthetic competitive inhibitors. FEBS Lett.
1996, 385, 87-90.
J ensen, S. K.; Michaelsen, S.; Kachlicki, P.; Sørensen, H.
4-Hydroxyglucobrassicin and degradation products of glu-
cosinolates in relation to unsolved problems with the quality
of double low oilseed rape. Proc. 8th. Int. Rapeseed Congr.
1991, 1359-1364.
Kiss, G.; Neukom, H. U¨ ber die Struktur des Ascorbigens. Helv.
Chim. Acta 1966, 49, 989-992.
Thies, W. Isolation of sinigrin and glucotropaeolin from
cruciferous seeds. Fat Sci. Technol. 1988, 90, 311-314.
Kuta´cek, M.; Procha´zka, Z.; Gru¨nberger, D. Biogenesis of
ascorbigen, 3-indolyl acetonitril and indole-3-carboxylic acid
from D,L-tryptophan-3-14C in Brassica oleraceae L. Nature
1960, 187, 61-62.
Latxague, L.; Gardrat, C.; Coustille, J . L.; Viaud, M. C.; Rollin,
P. Identification of enzymatic degradation products from
synthesised glucobrassicin by gas chromatography-mass
spectrometry. J . Chromatogr. 1991, 585, 166-170.
Loft, S.; Otte, J .; Poulsen, H. E.; Sørensen, H. Influence of
intact and myrosinase-treated indolyl glucosinolates on the
metabolism in vivo of metronidazole and antipyrine in the
rat. Food Chem. Toxicol. 1992, 30, 927-935.
Lu¨thy, J .; Benn, M. H. Thiocyanate formation from
glucosinolates: a study of the autolysis of allylglucosinolate
in Thlaspi arvense L. seed flour extracts. Can. J . Biochem.
1977, 55, 1028-1031.
McDanell, R.; McLean, A. E. M.; Hanley, A. B.; Heaney, R.
K.; Fenwick, G. R Differential induction of mixed-function
oxidase (MFO) activity in rat liver and intestine by diets
containing processed cabbage: Correlation with cabbage
levels of glucosinolates and glucosinolate hydrolysis prod-
ucts. Food Chem. Toxicol. 1987, 25, 5, 363-368.
McDanell, R.; McLean, A. E. M.; Hanley, A. B.; Heaney, R.
K.; Fenwick, G. R Chemical and biological properties of
indole glucosinolates (glucobrassicins): A review. Food
Chem. Toxicol. 1988, 26, 59-70.
Uda, Y.; Kurata, T.; Arakawa, N. Effects of pH and ferrous
ion on the degradation of glucosinolates by myrosinase.
Agric. Biol. Chem. 1986, 50, 2735-2740.
Van Etten, C. H.; Tookey, H. L. Chemistry and Biological
Effects of Glucosinolates. In Herbivores: Their interaction
with secondary plant metabolites; Rosenthal, G. A., J anzen,
D. H., Applebaum, S. W., Eds.; Academic Press: San Diego,
CA, 1979.
Viaud, M.-C.; Rollin, P.; Latzague, L.; Gardrat, C. Synthetic
studies on indole glucosinolates. Part 1. Synthesis of glu-
cobrassicin and its 4- and 5-methoxy derivatives. J . Chem.
Res. (miniprint) 1992, 1664-1672.
Visentin, M.; Tava, A.; Iori, R.; Palmieri, S. Isolation and
identification of trans-4-(methylthio)-3-butenyl glucosinolate
from radish roots (Raphanus sativus L.). J . Agric. Food
Chem. 1992, 40, 1687-1691.
Xue, J . Molecular studies of myrosinase in Brassicaceae-from
protein to gene and gene to protein. Dissertation, Sveriges
Lantbruksuniversitet, Uppsala, Sweden, 1994; ISBN 91-
576-4848-1.
Received for review October 1, 1997. Revised manuscript
received J anuary 8, 1998. Accepted February 17, 1998. This
research work was partially supported from the Commission
of the European Communities, Agriculture and Fisheries
(FAIR) specific RTD Program CT95-0260, High Quality Oils,
Proteins and Bioactive Products for Food and Non-Food
Purposes Based on Biorefining of Cruciferous Oilseed Crops.
Michaelsen, S.; Mortensen, K.; Sørensen, H. Myrosinases in
Brassica: Characterisation and properties. Proc. 8th Int.
Rapeseed Congr. 1991, 905-910.
Michaelsen, S.; Møller, P.; Sørensen, H. Factors influencing
the separation and quantitation of intact glucosinolates and
desulphoglucosinolates by micellar electrokinetic capillary
chromatography. J . Chromatogr. 1992, 608, 363-374.
J F9708498