Download New Phytol

RESEARCH New Phytol. (2000), 146, 343–352
Differences in glucosinolate patterns and
arbuscular mycorrhizal status of
glucosinolate-containing plant species
H. V I E R H E I L I G", R. B E N N E T T#, G. K I D D L E$, M. K A L D O R F%
   J. L U D W I G - M U$ L L E R&*
" Institut fuW r Phytopathologie, Christian-Albrechts-UniversitaW t,
Hermann-Rodewald-Str. 9, D-24118 Kiel, Germany
# Cellular Metabolism and Enzymology Group, Institute of Food Research,
Norwich Research Park, Colney, Norwich, Norfolk NR4 7UA, UK
$ Biochemistry and Physiology Department, IACR–Rothamsted, Harpenden, Hertfordshire
AL5 2JQ, UK
% Institut fuW r OW kologie, Lehrbereich Umweltwissenschaften, Friedrich-Schiller-UniversitaW t,
Dornburgerstr. 159, D-07743 Jena, Germany
& Institut fuW r Botanik, Technische UniversitaW t Dresden, Zellescher Weg 22,
D-01062 Dresden, Germany
Received 20 August 1999 ; accepted 17 January 2000

Under defined laboratory conditions it was shown that two glucosinolate-containing plant species, Tropaeolum
majus and Carica papaya, were colonized by arbuscular mycorrhizal (AM) fungi, whereas it was not possible to
detect AM fungal structures in other glucosinolate-containing plants (including several Brassicaceae).
Benzylglucosinolate was present in all of the T. majus cultivars and in C. papaya it was the major glucosinolate.
2-Phenylethylglucosinolate was found in most of the non-host plants tested. Its absence in the AM host plants
indicates a possible role for the isothiocyanate produced from its myrosinase-catalysed hydrolysis as a general AM
inhibitory factor in non-host plants. The results suggest that some of the indole glucosinolates might also be
involved in preventing AM formation in some of the species. In all plants tested, both AM hosts and non-hosts,
the glucosinolate pattern was altered after inoculation with one of three different AM fungi (Glomus mosseae,
Glomus intraradices and Gigaspora rosea), indicating signals between AM fungi and plants even before root
colonization. The glucosinolate induction was not specifically dependent on the AM fungus. A time-course study
in T. majus showed that glucosinolate induction was present during all stages of mycorrhizal colonization.
Key words : arbuscular mycorrhiza, Brassicaceae, Glomus, glucosinolates, Tropaeolum majus.

Glucosinolates are amino acid-derived secondary
plant products synthesized by members of the
Brassicaceae including crops such as Brassica napus
(oilseed rape), Brassica rapa (Chinese cabbage)
and Brassica juncea (brown mustard), and also by
other plant families such as Arabidaceae, Capparaceae, Caricaceae, Resedaceae and Tropaeolaceae
(Rodman, 1991 ; Wallsgrove et al., 1998). So far
100 glucosinolates have been described (Daxenbichler et al., 1991 ; Sørensen, 1991), which can be
grouped into three general classes based on the
*Author for correspondence (tel j49 351 463 3939 ; fax j49 351
463 7032 ; e-mail jutta.ludwig-mueller!mailbox.tu-dresden.de).
precursor amino acids – aliphatic\alkenyl (derived
from L-Met homologues), aromatic (derived from
L-Phe, L-Phe homologues and L-Tyr) and indolylglucosinolates (derived from L-Trp) (Bennett &
Wallsgrove, 1994 ; Wallsgrove et al., 1998). Upon
tissue disruption such as fungal infection or insect
feeding, glucosinolates are catabolized by myrosinases (thioglucosidases ; EC 3.2.3.1) to produce a
variety of bioactive compounds (dependent on the
parent glucosinolate) including isothiocyanates, thiocyanates, nitriles, oxazolidenethiones and epithioalkanes (for review see Wallsgrove et al., 1998).
Some of these compounds have been shown to be
fungitoxic\fungistatic (Greenhalgh & Mitchell,
1976 ; Bennett & Wallsgrove, 1994 ; Wallsgrove et
344
RESEARCH H. Vierheilig et al.
al., 1998). Others such as the indolylglucosinolates
play a role in the biosynthesis of indole-3-acetic acid
(IAA) in the Brassicaceae and in the formation of
clubroot disease (Plasmodiophora brassicae ; LudwigMu$ ller & Hilgenberg, 1988 ; Ludwig-Mu$ ller et al.,
1990, 1997, 1999a,b). Non-Brassica plants such as
Tropaeolum majus and Carica papaya have benzylglucosinolate as their major glucosinolate (Bennett et
al., 1996, 1997).
Mycorrhizas are symbiotic associations between
plants and root-colonizing fungi. The complex
cellular relationship between host roots and arbuscular mycorrhizal (AM) fungi requires a continuous
exchange of signals (reviewed by Vierheilig et al.,
1998) which leads to the development of specific AM
fungal structures in the roots of host plants. Most
higher plants are able to form AM symbiosis with
fungi of the order Glomales ; however there are
contradictory reports about the mycorrhizal status of
plants in the Brassicaceae (Medve, 1983 ; Harley &
Harley, 1987 ; Newman & Reddell, 1987 ; Koide &
Schreiner, 1992). In general the Brassicaceae are
known as AM non-host plants, however recent
papers report AM colonization in wild crucifers
including Capsella bursa-pastoris, Coronopus didymus, Hesperis matronalis, Matthiola incana and
Sisymbrium irio (DeMars & Boerner, 1994, 1995 ;
Kapoor et al., 1996). Tommerup (1984) observed
that appressoria of Glomus caledonium become firmly
attached to roots of B. napus, but only a few
penetration pegs were formed. Glenn et al. (1985)
examined a number of Brassica cultivars and detected penetration of roots by AM fungi. However
microscopical analysis revealed that mycorrhizal
fungal penetration occurred only in dead cortical
cells, which would have very low\zero glucosinolate
concentrations and therefore minimal capacity to
produce fungitoxic isothiocyanates. In addition,
hyphae growing near healthy cells showed retracted
cytoplasm, indicating the presence of inhibitory
compounds in the root exudates (Glenn et al., 1985).
In none of these studies, however, was a functional
AM association characterized by the formation of
arbuscules observed.
Several hypothesis have been suggested to explain
the inability of AM non-host plants to form
mycorrhizas (Koide & Schreiner, 1992 ; Vierheilig et
al., 1998). Whereas the Chenopodiaceae and lupins
appear to lack factors essential for AM mycorrhiza,
in other non-hosts such as Brassicaceae two other
mechanisms have been proposed : (1) the differences
in phosphate acquisition\scavenging systems compared with mycorrhizal species (Murley et al., 1998) ;
and (2) the formation of fungitoxic\fungistatic
breakdown products from the glucosinolates (Vierheilig & Ocampo, 1990a,b ; Koide, 1991 ; Koide &
Schreiner, 1992 ; Schreiner & Koide, 1993a,b).
In this study we tested the correlation between
constitutive endogenous glucosinolate concentra-
tions and\or patterns, in a variety of glucosinolatecontaining plants, with the ability\inability of these
plants to form AM symbiosis. Glucosinolate concentrations were measured in the roots of AM-inoculated and non-inoculated plants in order to determine
if suppression and\or induction of these metabolites
occurs, and hence might affect mycorrhizal colonization.
  
Plant material and inoculation procedure
The sources and cultivars of plant material used in
this study are summarized in Table 1. Seeds were
surface-sterilized in 50% commercial bleach for
5 min, rinsed several times with tap water and
germinated in autoclaved (40 min ; 120mC) vermiculite. After 8 d the seedlings were transferred to
a steam-sterilized (40 min, 120mC) mixture of silicate
sand, TurFace2 (baked clay substrate which is
mechanically broken to a diameter of 2–5 mm ;
Applied Industrial Materials, Buffalo Grove, IL,
USA) and soil (2 : 2 : 1, v\v\v). Plants were inoculated
in a growth chamber (day : night cycle 16 h, 22mC :
8 h, 20mC ; 50% rh) using the compartment system
developed by Wyss et al. (1991) consisting of three
compartments. A central compartment (20i10i2
cm) contained beans (Phaseolus vulgaris L. cv. Sun
Gold), inoculated with one of the three AM fungi
tested. This was separated by a nylon screen (60 µm
mesh size), which is penetrated by hyphae but not by
roots, from the lateral compartments which in turn
were subdivided into five small subcompartments
(3n3i10i2 cm). Test plants were grown in the
subcompartments. With AM host plants this design
results in AM colonization of the plants in the lateral
subcompartments within 2 wk. Thus a period of 4n5
wk, as used in most of our experiments, ensures that
if plants are susceptible to AM fungi root colonization will occur. For the time-course experiment,
plants were harvested 11, 21, 31 and 105 d after
joining the lateral compartments with the central
compartment and thus exposing the plants to the
AM fungi. Roots of three plants per species were
harvested and a subsample taken for determination
of infection following clearing and staining (Phillips
& Hayman, 1970). Stained roots were inspected with
a Leitz Laborlux 12 light microscope, and the
percentage of root colonization was determined by
counting the fungal structures (hyphae, arbuscules,
vesicles) in infected roots according to a modification
of Newman’s (1966) method.
Several different AM fungi were tested – Gigaspora rosea Nicolson & Schenck (Bago et al., 1998),
formerly wrongly classified as Gigaspora margarita
Becker & Hall (DAOM 194757 ; Department of
Agriculture, Ottawa, Canada) ; Glomus mosseae
(Nicolson & Gerdemann) Gerd. & Trappe (BEG 12 ;
RESEARCH Glucosinolates and AM colonization
345
Table 1. Plant material and suppliers used in this study
Plant species
Cultivar
Family
Supplier
Barbarea praecox
Wild type
Arabidaceae
Barbarea vulgaris
Variegata
Arabidaceae
Brassica napus
Brassica napus
Brassica nigra
Carica papaya
Lepidium sativum
Lepidium sativum
Nasturtium officinalis
Reseda alba
Reseda lutea
Reseda luteola
Sinapis alba
Tropaeolum majus
Tropaeolum majus
Tropaeolum majus
Low seed GSL
High seed GSL
Wild type
Unknown
Curled
Garden\plain
Wild type
Wild type
Wild type
Wild type
Wild type
Wild type
Wild type
Nanum
Brassicaceae
Brassicaceae
Brassicaceae
Caricaceae
Brassicaceae
Brassicaceae
Brassicaceae
Resedaceae
Resedaceae
Resedaceae
Brassicaceae
Tropaeolaceae
Tropaeolaceae
Tropaeolaceae
Kings (E. W. Kings & Co. Ltd), Monks Farm,
Kelvedon, Essex, UK
Chiltern Seeds, Bortree Stile, Ulverston,
Cumbria, UK
Norddeutsche Pflanzenzucht, Holtsee, Germany
Norddeutsche Pflanzenzucht, Holtsee, Germany
IACR-Rothamsted, yearly fresh stock
Ripe fruits (local supermarket)
Kings
Chiltern Seeds
Chiltern Seeds
Botanical Garden, Frankfurt
Botanical Garden, Frankfurt
Botanical Garden, Frankfurt
IACR-Rothamsted, yearly fresh stock
Botanical Garden, Frankfurt
Samenhandlung Knutzen, Kiel, Germany
Dehner Gartencenter, Rain am Lech, Germany
The names of the cultivars are given when known. Mycorrhizal plants are given in bold.
GSL, glucosinolate.
Table 2. Semi-systematic, trivial and abbreviated names of glucosinolates, with major myrosinase-catalysed
breakdown products
Semi-systematic
name
Aliphatic\Alkenyl
(R) 2-OH-3-Butenyl
(R) 2-OH-3-Pentenyl
2-Propenyl
3-Butenyl
4-Pentenyl
Aromatic
Benzyl
p-OH-Benzyl
2-Phenylethyl
(R\S) 2-OH-2-Phenylethyl
Indole
3-Indolylmethyl
N-MeO-3-Indolylmethyl
4-OH-3-Indolylmethyl
4-MeO-3-Indolylmethyl
Trivial name
Abbreviation
in Tables
and Figures
Major breakdown product
Volatility
of product
Progoitrin
Gluconapoleiferin
Sinigrin
Gluconapin
Glucobrassicanapin
2-H3B
2-H4P
2-P
3-B
4-P
5-Vinyloxazolidene-2-thione
5-Allyloxazolidene-2-thione
Allylisothiocyanate
3-Butenylisothiocyanate
4-Pentenylisothiocyanate
Non-volatile
Non-volatile
Volatile
Volatile
Volatile
Glucotropaeolin
Sinalbin
Gluconasturtin
R l glucosibarin
S l glucobarbarin
Benz
H-Benz
2-PE
Benzylisothiocyanate
p-OH-benzylthiocyanate
Phenylethylisothiocyanate
Volatile
Non-volatile
Volatile
2-HPE
5-Phenyloxazolidene-2-thione Non-volatile
Glucobrassicin
Neoglucobrassicin
4-OH-Glucobrassicin
4-MeO-Glucobrassicin
3-IM
1-MeO
4-OH
4-MeO
Indole-3-acetonitrile
Unidentified (nitrile ?)
Unidentified (nitrile ?)
Unidentified (nitrile ?)
La Banque Europe! enne des Glomales, International
Institute of Biotechnology, Kent, UK) and two
isolates of Glomus intraradices Smith & Schenck
(DAOM 197198 and INVAM Sy 167 ; the latter was
kindly provided by S. Reinhard, Institut fu$ r Pflanzenerna$ hrung, Universita$ t Hohenheim, Germany).
The experiments were performed once with three
replicates per treatment.
Glucosinolate analysis
A subsample of roots from each of the three plants
was washed with tap water and dried between filter
papers, and the fresh weight of each sample was
Non-volatile
Non-volatile
Non-volatile
Non-volatile
recorded. The plant material was frozen in liquid N
#
before freeze-drying. Dry samples were milled to a
fine powder before glucosinolate analyses. Glucosinolate extraction (from 3i40 mg d. wt per sample)
and determinations were performed as previously
described (Porter et al., 1991 ; Bennett et al., 1996 ;
Kiddle et al., 1999) using sinigrin as an extraction
standard. Separation and detection of desulphoglucosinolates was performed using a Waters 996
photodiode array HPLC (Waters, Milford, MA,
USA), and identifications were achieved using authentic standards which were previously identified by
NMR ("$C and "H) and chemical ionization mass
spectrometry (G. Kiddle et al., unpublished). All
RESEARCH H. Vierheilig et al.

Mycorrhizal status of tested plants
AM fungal structures were observed only in C.
papaya and the T. majus cultivars. Successful
colonization was defined as the formation of arbuscules and\or vesicles as functional mycorrhiza. When
the T. majus cultivars were inoculated with G. rosea
arbuscules were clearly visible in the roots (not
shown), and in plants inoculated with G. mosseae or
G. intraradices arbuscules and vesicles were observed
(Fig. 1a,b). In C. papaya inoculated with G. mosseae,
roots were extensively colonized and many arbuscules were formed (Fig. 1c). No AM fungal structures attached to or in the roots, or any other
alterations of the roots, were observed in the other
glucosinolate-containing species (Tables 1, 3).
Glucosinolates in AM-inoculated and non-inoculated
T. majus plants
Slightly different glucosinolate profiles were found
in uninoculated roots of the different T. majus
(a)
2
35
T. majus ‘Nanum’
30
25
20
1
15
10
5
0
50
0
2
T. majus ‘BGF’
40
30
1
20
Benzylglucosinolatei1000 (nmol g–1 d. wt)
analyses were done in triplicate for each sample. The
semi-systematic and trivial names of glucosinolates
identified in this study, and the abbreviations used in
the tables and figures, are given in Table 2.
Glucosinolatesi 1000 (nmol g–1 d. wt)
346
10
0
0
2-H3B
3-B
3-IM
Glucosinolate
Benz
Fig. 2. Glucosinolates in roots of two different cultivars of
Tropaeolum majus after inoculation with Glomus mosseae.
The analyses were carried out 4n5 wk after inoculation.
The infection rate was 86p6 and 78p5% for ‘ nanum ’ and
‘ Botanical Garden Frankfurt ’ (BGF), respectively. Open
bars, uninoculated ; closed bars, inoculated. Values are
given as meanspSE, n l 3. See Table 2 for abbreviations
of glucosinolates.
(c)
(b)
Fig. 1. Infection of Tropaeolum majus by Glomus intraradices with the formation of arbuscules (a) and vesicles
(b). (c) Heavy infection (arbuscule formation) of Carica papaya by Glomus mosseae.
347
1000
50
800
40
Control
600
30
G. mosseae
G. intraradices
400
20
Gigaspora ssp.
200
10
0
2-H3B
3-B
3-IM
Glucosinolate
Benz
Benzylglucosinolatei1000 (nmol g–1 d. wt)
Glucosinolates (nmol g–1 d. wt)
RESEARCH Glucosinolates and AM colonization
0
Fig. 3. Glucosinolate induction in Tropaeolum majus after inoculation with specific AM fungi. T. majus
‘ Botanical Garden Frankfurt ’ was inoculated with Glomus mosseae, Glomus intraradices and Gigaspora rosea.
Plants were harvested and analysed 4n5 wk after inoculation. The infection rates were 78p5% for G. mosseae,
63p5% for G. intraradices, and 90p7% for Gigaspora rosea. Values are given as meanspSE, n l 3. See Table
2 for abbreviations of glucosinolates.
60
200
Leaves
50
150
40
100
30
20
50
10
Benzylglucosinolatei 1000 (nmol g–1 d. wt)
Benzylglucosinolatei 1000 (nmol g–1 d. wt)
Roots
nd
0
11
20
31
105
11
20
31
105
0
Days after inoculation
Fig. 4. Time-course of glucosinolate induction in roots and leaves of Tropaeolum majus ‘ Botanical Garden
Frankfurt ’ by Glomus mosseae. Infection rates were 57p7, 81p11 and 85p5% for plants 11, 20 and 31 d after
inoculation, respectively. nd, not determined. Shaded bars, uninoculated ; closed bars, inoculated. Values are
given as meanspSE, n l 3.
cultivars, but in all of them benzylglucosinolate was
present at high concentration (Fig. 2). In T. majus
‘ Nanum ’ and ‘ Botanical Garden Frankfurt ’ (BGF),
traces of alkenylglucosinolates and 3-indolylmethylglucosinolate (3-IM) were also found (Fig. 2),
whereas in T. majus ‘ Samenhandlung Knutzen Kiel,
Germany (SKK) ’ there were only traces of 3-IM,
and none of the alkenylglucosinolates (data not
shown).
After inoculation with a Glomus species there was
no effect on the benzylglucosinolate concentration in
T. majus ‘ nanum ’, whereas increases in 2-OH-3butenyl-, 3-butenyl- and 3-indolylmethylglucosinolates were measured. The T. majus cultivar from
the Botanical Garden Frankfurt showed a marked
increase in benzylglucosinolate and 3-IM. In summary, all roots in which G. mosseae formed arbuscules and vesicles showed increases in several
glucosinolates compared with the uninoculated controls.
To determine whether the induction of glucosinolates in T. majus was dependent on AM fungal
RESEARCH H. Vierheilig et al.
348
20
8
20
L. sativum plain
S. alba
15
15
6
10
10
4
5
5
2
0
0
0
20
12
6
B. vulgaris
N. officinalis
5
10
4
8
3
6
2
4
1
2
0
0
0
2HP
E
HBe
nz
Be
nz
2PE
3IM
1M
eO
4M
eO
4OH
B. praecox
2HP
E
HBe
nz
Be
nz
2PE
3IM
1M
eO
4M
eO
4OH
Glucosinolatesi 1000 (nmol g–1 d. wt)
L. sativum curled
15
10
2HP
E
HBe
nz
Be
nz
2PE
3IM
1M
eO
4M
eO
4OH
5
Glucosinolate
Fig. 5. Induction of glucosinolates in different crucifer species inoculated with Glomus mosseae. No colonization
was observed in any of the species. Plants were harvested 4n5 wk after inoculation. Open bars, uninoculated ;
closed bars, inoculated. Values are given as meanspSE, n l 3. See Table 2 for abbreviations of glucosinolates.
Glucosinolates in different control and G. mosseaeinoculated species
AM colonization by G. mosseae of glucosinolatecontaining species from the families Arabidaceae,
Brassicaceae and Resedaceae was also tested (Table
1). Even without successful establishment of the
mycorrhizal symbioses in these species, a consistent
change was observed in the glucosinolate content in
inoculated versus non-inoculated plants. However,
the glucosinolate accumulation pattern was not the
same for all species. This can be seen as an increase
in several classes of glucosinolates (Figs 5, 6). In the
10
Glucosinolatesi1000
(nmol g–1 d. wt)
(a)
(b)
8
6
4
2
3B
4P
2PE
3I
1- M
M
4- eO
M
e
4- O
O
H
2H3
2- B
H4
P
3-
B
4P
2PE
3I
1- M
M
4- eO
M
e
4- O
O
H
0
2H3
2- B
H4
P
species and\or genera, T. majus ‘ Botanical Garden
Frankfurt ’ was inoculated with three different AM
fungi. The two major glucosinolates (benzylglucosinolate and 3-IM) were induced by all three
isolates to approximately the same extent, but
induction of 3-butenylglucosinolate was only observed after inoculation with G. intraradices (Fig. 3).
A time-course experiment (Fig. 4) showed that the
induction of benzylglucosinolate in roots colonized
by G. mosseae occurred in early (young roots, 11 d
after inoculation, dai), intermediate (20 and 31 dai)
and later (105 dai) phases of colonization. In contrast
to roots, no consistent differences of benzylglucosinolate were observed in leaves (one exception
was a decrease found at 20 dai in leaves of inoculated
plants). The effects on glucosinolate induction
therefore appear to be root-localized and not systemic.
Glucosinolate
Fig. 6. Induction of glucosinolates in roots of two different
Brassica napus varieties having (a) low (0n4 µmol g−") or (b)
high (3n1 µmol g−") seed glucosinolate content, after
inoculation with Glomus mosseae. Plants were harvested
3 wk after inoculation. Open bars, uninoculated ; closed
bars, inoculated. Values are given as meanspSE, n l 3.
See Table 2 for abbreviations of glucosinolates.
species of Arabidaceae the aromatic and indolylglucosinolates were affected : in Barbarea vulgaris, in
addition to 2-phenylethylglucosinolate (2-PE), four
indolylglucosinolates were induced, whereas in
Barbarea praecox the concentration of 3-IM was
reduced but 2-PE showed increases similar to those
found in B. vulgaris (Fig. 5). In Nasturtium officinalis
(Brassicaceae), slight induction of 2-phenylethyland 4-OH-3-indolylmethylglucosinolate was found.
In both Lepidium sativum cultivars 2-PE was induced
in inoculated roots, whereas the major glucosinolate
RESEARCH Glucosinolates and AM colonization
i (traces)
i (traces)*
i
i
i
i
i
i (traces)
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i*
i
i
i
2-PE
H-Benz
Benz
i

i
i
i
i
i (traces)
i
i
i
i
i
4-P
T. majus ‘ SKK ’
T. majus Nanum
T. majus ‘ BGF ’
C. papaya
L. sativum
S. alba
B. praecox
B. vulgaris
N. officinalis
B. napus (low)
B. napus (high)
B. nigra
R. alba
R. lutea
R. luteola
*Strong induction in inoculated roots.
AM
AM
AM
AM
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Plant species
i
i
i
2-H4P
i
2-P
3-B
i (traces)
i
i
i
i
i
i
i
i (traces)
i
i (traces)
i
i (traces)*
i
0-α-rhamnopyranosyl
-benzyl
C8-\ C9methylsulphinyl
AM
status
2-H3B
(benzylglucosinolate) was essentially unaltered. In
Sinapis alba only 4-MeO-3-indolylmethylglucosinolate was induced, and other glucosinolates were
reduced in inoculated roots compared with controls.
Two B. napus cultivars with different seed glucosinolate contents were also investigated. One single
low cultivar (low erucic acid, high total glucosinolate,
3n1 µmol g-" seed, according to the distributor) and
one double low cultivar (low erucic acid and low total
glucosinolate, 0n4 µmol g−" seed, according to the
distributor) were inoculated. No AM colonization
was found in the roots of either cultivar. As with the
other non-mycorrhizal crucifers tested, there appears
to be some effect on the plant of the AM fungus –
seen again as increases in specific glucosinolates in
the inoculated roots (Fig. 6). The patterns for
induction were essentially the same in both cultivars.
The root glucosinolate concentration and pattern
differed only slightly between cultivars, although the
seed glucosinolate levels were different. No systemic
changes\induction of glucosinolates in leaves was
observed.
Table 3 summarizes the results and presents the
glucosinolate patterns in other plants tested during
this study, but not described in detail (e.g. Brassica
nigra, Reseda species).
i
i (traces)
i
i
i
i
i
4-OH
4-MeO
1-MeO
3-IM
2-HPE
Indole glucosinolates
Aromatic glucosinolates
Aliphatic glucosinolates
Table 3. Glucosinolate pattern in roots of mycorrhizal and non-mycorrhizal plant species
349
The colonization by AM fungi of glucosinolatecontaining plant species such as certain Brassica and
Reseda species has been reported ; however a functional mycorrhizal association has never been proven
(Medve, 1983 ; Harley & Harley, 1987 ; Newman &
Reddell, 1987 ; Tester et al., 1987 ; Koide &
Schreiner, 1992). Most studies have been performed
on field material where the unequivocal identification
of individual fungi is often difficult unless arbuscules
and vesicles are formed (Harley & Harley, 1987 ;
Tester et al., 1987) and therefore hyphae attributed
to AM fungi might in fact be those of nonmycorrhizal fungi (DeMars & Boerner, 1995).
In our study defined AM fungal isolates were used
under controlled conditions, and some of our results
contradict earlier reports (Medve, 1983 ; Harley &
Harley, 1987 ; Newman & Reddell, 1987 ; Tester
et al., 1987 ; Koide & Schreiner, 1992). No root
colonization or even AM fungal hyphal attachment
to roots was seen in species of Arabidaceae, Brassicaceae or Resedaceae. Therefore we suggest that the
term non-mycorrhizal should be maintained for
these plants.
To our knowledge there are no previous comprehensive data that link AM colonization with the
quantitative\qualitative glucosinolate content of
roots. Where glucosinolate analyses have been performed, often only seed glucosinolate content was
analysed, and rarely any other tissue (e.g. Daxenbichler et al., 1991).
350
RESEARCH H. Vierheilig et al.
In the present study, high root concentrations of
benzylglucosinolate were found in the AM non-host
plant L. sativum and in the AM hosts T. majus and
C. papaya. In C. papaya cyanogenic glucosides are
also present (Bennett et al., 1996, 1997). However,
the formation of arbuscules and vesicles in C. papaya
indicates that the AM fungus can deal with these
metabolites. Since myrosinases are known to occur
in all glucosinolate-containing plants (Wallsgrove
et al., 1998), and thus benzylglucosinolate can be
readily hydrolysed to a potentially bioactive compound, this indicates that benzylglucosinolate might
not be the sole factor responsible for the non-host
status of L. sativum. The presence of 2-PE in the
roots of L. sativum might be a more important factor.
2-PE, and its myrosinase-produced isothiocyanate,
is a good candidate to explain the AM non-host
status of L. sativum and the other glucosinolatecontaining plants. It was detected only in non-AM
plants and not in glucosinolate-containing AM host
plants. It was not detected in non-inoculated roots of
the AM non-host B. napus, but it was present in
AM-inoculated roots. In roots of Reseda lutea and
Reseda alba the major glucosinolates are o-alpha(rhamnopyranosyloxy)benzyl and 2-phenylethyl, respectively (R. Bennett, unpublished results). The
modified benzylglucosinolate also yields an isothiocyanate upon myrosinase hydrolysis (Olsen & Sørensen, 1979).
Analysis of changes in the much smaller concentrations of aliphatic glucosinolates did not give any
clear correlations with AM inhibition. 2-Hydroxy-3butenyl glucosinolate and 3-butenylglucosinolate
were detected in all T. majus cultivars and in several
AM non-hosts, whereas both compounds were
absent in C. papaya and a range of other non-hosts.
Other aliphatic glucosinolates such as 2-hydroxy-4pentenylglucosinolate, 2-propenylglucosinolate and
7-methylsulfinylheptyl and 8-methylsulfinyloctyl
glucosinolates were detected sporadically in low
concentrations in some non-hosts. The presence of
the same aliphatic glucosinolates in similar concentrations in the AM host and non-host plants makes it
unlikely that aliphatic glucosinolates are responsible
for the inability of the non-host plants to become
colonized by AM fungi.
A different pattern was observed with the indole
glucosinolates. 3-IM was detected in the mycorrhizal
T. majus cultivars and in most of the tested AM nonhosts (except L. sativum). In the mycorrhizal plant
T. majus, similar levels of 3-IM were detected to
those in some of the non-host plants (Fig. 5 ; N.
officinalis, S. alba, B. vulgaris), indicating that the
concentration of 3-IM is not a regulating factor for
AM status. Apart from 3-IM, no other indole
glucosinolates were detected in the mycorrhizal
plants, but several other indole glucosinolates were
found only in roots of AM non-hosts (Figs 5 and 6),
so these indole glucosinolates and their breakdown
products might be additional factors inhibiting
mycorrhizal colonization of these plants. However
their absence, or presence in very low concentrations,
in L. sativum, R. lutea and R. luteola exclude them as
more general factors.
There are few data on the specific cellular
localization (e.g. epidermis versus cortex) of glucosinolates\myrosinases in the roots of the species
investigated. The initial theory that myrosin cells
were the sole site of glucosinolates and myrosinase is
unlikely, and therefore it cannot be avoidance of
these cells that explains the successful colonization
of T. majus and C. papaya (Kelly et al., 1998 ;
Wallsgrove et al., 1998). Probably there is synergistic
inhibition by various root metabolites leading to
inhibition of colonization in the other glucosinolatecontaining species. The degradation of glucosinolates to isothiocyanates was proposed to be one of the
reasons responsible for the non-mycotrophic status
of the Brassicaceae, as the latter compounds are
fungitoxic. Several studies support this hypothesis.
Root colonization with AM fungi was reduced in
mycorrhizal plants when co-cultivated with crucifers
(Hayman et al., 1975), and exudates of crucifers
reduced AM fungal spore germination (El-Atrach
et al., 1989) and hyphal spreading in soil (Vierheilig
et al., 1995). Moreover, the volatile and soluble
fractions of cabbage root extracts (Vierheilig &
Ocampo, 1990a) and of sinigrin in combination with
myrosinase (Vierheilig & Ocampo, 1990b) exhibited
an inhibitory effect on spore germination of G.
mosseae. Some glucosinolate hydrolysis products
might inhibit spore germination and\or hyphal
growth, whereas others might be inactive. Schreiner
& Koide (1993a) isolated three antifungal compounds derived from glucosinolates from Brassica
kaber. The predominant antifungal compound was
identified as p-OH-benzylisothiocyanate, which is
derived from p-OH-benzylglucosinolate. In this
study p-OH-benzylglucosinolate was present only in
S. alba and is therefore unlikely to be generally
responsible for inhibition of colonization ; benzylglucosinolate would seem unlikely as two species
with high root concentrations were colonized (Table
3). The major breakdown product of 2-PE is the
volatile 2-phenylethylisothiocyanate (Table 2), but it
is not known whether this compound inhibits AM
fungus growth and\or development. However the
general high chemical\biological reactivity of the
isothiocyanates indicates antifungal activity (Wallsgrove et al., 1998). This hypothesis is currently
under investigation in our laboratory.
The formation of the AM symbiosis requires a
continuous exchange of signals between plant and
fungus. This exchange of signals starts even before a
direct contact between both symbiotic partners
occurs. There is abundant information about signals
released by roots toward AM fungi (reviewed by
Vierheilig et al., 1998), but scarcely any data are
RESEARCH Glucosinolates and AM colonization
available about signals from the fungi toward plants.
After inoculation with an AM fungus, changes in
glucosinolate levels were observed in all the plant
species investigated in this study, even when no
hyphae were attached to the roots, indicating that
signals were released by the fungus. A precolonization signal in the AM interaction has already
been suggested in a study of chitinase activity in B.
napus roots inoculated by G. mosseae (Vierheilig et
al., 1994).
In summary, our results show : the aromatic
glucosinolate 2-phenylethylglucosinolate as a possible general factor responsible for the non-susceptibility to AM infection of some plants ; a
plant–fungus interaction as shown by changes in
endogenous root glucosinolate concentrations in AM
host and non-host plants ; and localization of the
effects of inoculation in the roots, with no systemic
increase of glucosinolates.

Bago B, Bentivenga SP, Brenac V, Dodd JC, Piche Y, Simon
L. 1998. Molecular analysis of Gigaspora (Glomales, Gigasporaceae). New Phytologist 139 : 581–588.
Bennett RN, Kiddle G, Hick AJ, Dawson GW, Wallsgrove
RM. 1996. Distribution and activity of microsomal NADPHdependent monooxygenases and amino acid decarboxylases in
cruciferous and non-cruciferous plants and their relationship to
foliar glucosinolate content. Plant, Cell & Environment 19 :
801–812.
Bennett RN, Kiddle G, Wallsgrove RM. 1997. Biosynthesis of
benzylglucosinolate, cyanogenic glucosides and phenylpropanoids in Carica papaya. Phytochemistry 45 : 59–66.
Bennett RN, Wallsgrove RM. 1994. Secondary metabolites in
plant defence mechanisms. New Phytologist 127 : 617–633.
Daxenbichler ME, Spencer GF, Carlson DG, Rose GB,
Brinker AM, Powell RG. 1991. Glucosinolate composition of
seeds from 297 species of wild plants. Phytochemistry 30 :
2623–2638.
DeMars BG, Boerner REJ. 1994. Vesicular–arbuscular mycorrhizal fungi colonisation in Capsella bursa-pastoris (Brassicaceae). American Midland Naturalist 132 : 377–380.
DeMars BG, Boerner REJ. 1995. Arbuscular mycorrhizal
development in three crucifers. Mycorrhiza 5 : 405–408.
El-Atrach F, Vierheilig H, Ocampo JA. 1989. Influence of nonhost plants on vesicular–arbuscular mycorrhizal infection
of host plants and on spore germination. Soil Biology and
Biochemistry 21 : 161–163.
Glenn MG, Chew FS, Williams PH. 1985. Hyphal penetration
of Brassica (Cruciferae) roots by a vesicular–arbuscular mycorrhizal fungus. New Phytologist 99 : 463–472.
Greenhalgh JR, Mitchell ND. 1976. The involvement of
flavour volatiles in the resistance to downy mildew of wild and
cultivated forms of Brassica oleracea. New Phytologist 77 :
391–398.
Harley JL, Harley EL. 1987. A check-list of mycorrhiza in the
British flora. New Phytologist 105 : 1–102.
Hayman DS, Johnson AM, Ruddesdin I. 1975. The influence
of phosphate and crop species on Endogone spores and
vesicular–arbuscular mycorrhiza under field conditions. Plant
and Soil 43 : 489–495.
Kapoor R, Pathak A, Mago P, Mukerji KG. 1996. VAM in two
crucifers. Cruciferae Newsletter 18 : 116–117.
Kelly PJ, Bones A, Rossiter JT. 1998. Sub-cellular immunolocalization of the glucosinolate sinigrin in seedlings of Brassica
juncea. Planta 206 : 370–377.
Kiddle GA, Bennett RN, Hick AJ, Wallsgrove RM. 1999. C–S
lyase activities in leaves of crucifers and non-crucifers. Partial
purification and characterisation of C–S lyase activities in
oilseed rape. Plant, Cell & Environment 22 : 433–445.
351
Koide RT. 1991. Tansley Review No. 29. Nutrient supply,
nutrient demand and plant responses to mycorrhizal infection.
New Phytologist 117 : 365–386.
Koide RT, Schreiner RP. 1992. Regulation of the vesicular–
arbuscular mycorrhizal symbiosis. Annual Review of Plant
Physiology and Plant Molecular Biology 43 : 557–581.
Ludwig-Mu$ ller J, Hilgenberg W. 1988. A plasma membranebound enzyme oxidizes L-tryptophan to indole-3-acetaldoxime. Physiologia Plantarum 74 : 240–250.
Ludwig-Mu$ ller J, Ihmig S, Bennett RN, Kiddle G, Ruppel
M, Hilgenberg W. 1999a. The host range of Plasmodiophora
brassicae and its relationship to endogenous glucosinolate
content. New Phytologist 141 : 443–458.
Ludwig-Mu$ ller J, Pieper K, Ruppel M, Epstein E, Cohen JD,
Kiddle G, Bennett RN. 1999b. Indole glucosinolates and
auxin biosynthesis in Arabidopsis thaliana glucosinolate mutants
and the development of clubroot disease. Planta 208 : 409–419.
Ludwig-Mu$ ller J, Rausch T, Lang S, Hilgenberg W. 1990.
Plasma membrane-bound high pI peroxidase isoenzymes
convert tryptophan to indole-3-acetaldoxime. Phytochemistry
29 : 1397–1400.
Ludwig-Mu$ ller J, Schubert B, Pieper K, Ihmig S, Hilgenberg W. 1997. Glucosinolate content in susceptible and tolerant
Chinese cabbage varieties during the development of the
clubroot disease. Phytochemistry 44 : 407–414.
Medve RJ. 1983. The mycorrhizal status of crucifers. American
Midland Naturalist 109 : 406–408.
Murley VR, Theodorou ME, Plaxton WC. 1998. Phosphate
starvation-inducible pyrophosphate-dependent phosphofructokinase occurs in plants whose roots do not form symbiotic
associations with mycorrhizal fungi. Physiologia Plantarum
103 : 405–414.
Newman EI. 1966. A method of estimating the total length of
root in a sample. Journal of Applied Ecology 3 : 139–145.
Newman EI, Reddell P. 1987. The distribution of mycorrhizas
among families of vascular plants. New Phytologist 106 :
745–751.
Olsen O, Sørensen H. 1979. Isolation of glucosinolates and the
identification of o(α-L-rhamnopyranosyloxy)benzylglucosinolate from Reseda odorata. Phytochemistry 18 : 1547–1552.
Phillips JM, Hayman DS. 1970. Improved procedures for
clearing roots and staining parasitic and vesiscular–arbuscular
mycorrhizal fungi for rapid assessment of infection. Transactions of the British Mycological Society 55 : 158–160.
Porter AJR, Morton AM, Kiddle G, Doughty KJ, Wallsgrove
RM. 1991. Variations in the glucosinolate content of oilseed
rape leaves. I. Effect of leaf age and position. Annals of Applied
Biology 118 : 461–467.
Rodman JE. 1991. A taxonomic analysis of glucosinolateproducing plants, I : phenetics. Systematic Botany 16 : 598–618.
Schreiner RP, Koide RT. 1993a. Antifungal compounds from
the roots of mycotrophic and non-mycotrophic plant species.
New Phytologist 123 : 99–105.
Schreiner RP, Koide RT. 1993b. Mustards, mustard oils and
mycorrhizas. New Phytologist 123 : 107–113.
Sørensen H. 1991. Glucosinolates : structure, properties, function. In : Shahidi F, ed. Canola and rapeseed. New York, USA :
Van Nostrand Reinhold, 149–172.
Tester M, Smith SE, Smith FA. 1987. The phenomenon of
‘ nonmycorrhizal ’ plants. Canadian Journal of Botany 65 :
419–431.
Tommerup IC. 1984. Development of infection by a vesicular–
arbuscular mycorrhizal fungus in Brassica napus L. and
Trifolium subterraneum L. New Phytologist 98 : 487–495.
Vierheilig H, Alt M, Ma$ der P, Boller T, Wiemken A. 1995.
Spreading of Glomus mosseae, a vesicular–arbuscular mycorrhizal fungus, across the rhizosphere of host and non-host
plants. Soil Biology and Biochemistry 27 : 1113–1115.
Vierheilig H, Alt M, Mohr U, Boller T, Wiemken A. 1994.
Ethylene biosynthesis and activities of chitinase and β-1,3glucanase in the roots of host and non-host plants of
vesicular–arbuscular mycorrhizal fungi after inoculation with
Glomus mosseae. Journal of Plant Physiology 143 : 337–343.
Vierheilig H, Bago B, Albrecht C, Poulin MJ, Piche Y. 1998.
Flavonoids and arbuscular mycorrhizal fungi. In : Manthey J,
Buslig B, eds. Flavonoids in the living system. New York, USA :
Plenum, 9–33.
Vierheilig H, Ocampo JA. 1990a. Role of root extracts and
352
RESEARCH H. Vierheilig et al.
volatile substances of non-host plants on vesicular–arbuscular
mycorrhizal spore germination. Symbiosis 9 : 199–202.
Vierheilig H, Ocampo JA. 1990b. Effect of isothiocyanates on
germination of spores of G. mosseae. Soil Biology and Biochemistry 22 : 1161–1162.
Wallsgrove RM, Doughty KJ, Bennett RN. 1998. Gluco-
sinolates. In : Singh BJ, ed. Plant amino acids. Biochemistry and
biotechnology. New York, USA : Marcel Dekker, 523–562.
Wyss P, Boller T, Wiemken A. 1991. Phytoalexin response is
elicited by a pathogen (Rhizoctonia solani) but not by a
mycorrhizal fungus (Glomus mosseae) in soybean roots. Experientia 47 : 395–399.