Taste detection of the non-volatile isothiocyanate moringin results in deterrence to glucosinolate-adapted insect larvae

Article abstract of DOI:10.1016/j.phytochem.2015.08.007

Isothiocyanates (ITCs), released from Brassicales plants after hydrolysis of glucosinolates, are known for their negative effects on herbivores but mechanisms have been elusive. The ITCs are initially present in dissolved form at the site of herbivore feeding, but volatile ITCs may subsequently enter the gas phase and all ITCs may react with matrix components. Deterrence to herbivores resulting from topically applied volatile ITCs in artificial feeding assays may hence lead to ambiguous conclusions. In the present study, the non-volatile ITC moringin (4-(α-l-rhamnopyranosyloxy)benzyl ITC) and its glucosinolate precursor glucomoringin were examined for effects on behaviour and taste physiology of specialist insect herbivores of Brassicales. In feeding bioassays, glucomoringin was not deterrent to larvae of Pieris napi (Lepidoptera: Pieridae) and Athalia rosae (Hymenoptera: Tenthredinidae), which are adapted to glucosinolates. Glucomoringin stimulated feeding of larvae of the related Pieris brassicae (Lepidoptera: Pieridae) and also elicited electrophysiological activity from a glucosinolate-sensitive gustatory neuron in the lateral maxillary taste sensilla. In contrast, the ITC moringin was deterrent to P. napi and P. brassicae at high levels and to A. rosae at both high and low levels when topically applied to cabbage leaf discs (either 12, 120 or 1200 nmol moringin per leaf disc of 1 cm diameter). Survival of A. rosae was also significantly reduced when larvae were kept on leaves treated with moringin for several days. Furthermore, moringin elicited electrophysiological activity in a deterrent-sensitive neuron in the medial maxillary taste sensillum of P. brassicae, providing a sensory mechanism for the deterrence and the first known ITC taste response of an insect. In simulated feeding assays, recovery of moringin was high, in accordance with its non-volatile nature. Our results demonstrate taste-mediated deterrence of a non-volatile, natural ITC to glucosinolate-adapted insects.

Full text of DOI:10.1016/j.phytochem.2015.08.007

Phytochemistry xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Phytochemistry
journal homepage: www.elsevier.com/locate/phytochem
Taste detection of the non-volatile isothiocyanate moringin results
in deterrence to glucosinolate-adapted insect larvae
b,
c
d
e
Caroline Müller a, , Joop van Loon , Sara Ruschioni , Gina Rosalinda De Nicola , Carl Erik Olsen ,
⇑
⇑
Renato Iori d, Niels Agerbirk e,
⇑
a Chemical Ecology, Bielefeld University, Universitätsstr. 25, 33615 Bielefeld, Germany
b Laboratory of Entomology, Wageningen University, Droevendaalsesteeg 1, 6708 PB Wageningen, The Netherlands
c Dipartimento di Scienze Agrarie, Alimentari ed Ambientali, Università Politecnica delle Marche, Via Brecce Blanche, 60131 Ancona, Italy
d Consiglio per la ricerca in agricoltura e l’analisi dell’economia agraria, Centro di ricerca per le colture industriali (CRA-CIN), Via di Corticella 133, 40128 Bologna, Italy
e Copenhagen Plant Science Center and Plant Biochemistry Laboratory, Department of Plant and Environmental Sciences, University of Copenhagen, Thorvaldsensvej 40,
1871 Frederiksberg C, Denmark
a r t i c l e i n f o
a b s t r a c t
Article history:
Isothiocyanates (ITCs), released from Brassicales plants after hydrolysis of glucosinolates, are known for
their negative effects on herbivores but mechanisms have been elusive. The ITCs are initially present in dis-
solved form at the site of herbivore feeding, but volatile ITCs may subsequently enter the gas phase and all
ITCs may react with matrix components. Deterrence to herbivores resulting from topically applied volatile
ITCs in artificial feeding assays may hence lead to ambiguous conclusions. In the present study, the non-
Received 17 February 2015
Received in revised form 11 August 2015
Accepted 17 August 2015
Available online xxxx
volatile ITC moringin (4-(a-L-rhamnopyranosyloxy)benzyl ITC) and its glucosinolate precursor gluco-
Keywords:
moringin were examined for effects on behaviour and taste physiology of specialist insect herbivores of
Brassicales. In feeding bioassays, glucomoringin was not deterrent to larvae of Pieris napi (Lepidoptera:
Pieridae) and Athalia rosae (Hymenoptera: Tenthredinidae), which are adapted to glucosinolates.
Glucomoringin stimulated feeding of larvae of the related Pieris brassicae (Lepidoptera: Pieridae) and also
elicited electrophysiological activity from a glucosinolate-sensitive gustatory neuron in the lateral maxillary
taste sensilla. In contrast, the ITC moringin was deterrent to P. napi and P. brassicae at high levels and to A.
rosae at both high and low levels when topically applied to cabbage leaf discs (either 12, 120 or 1200 nmol
moringin per leaf disc of 1 cm diameter). Survival of A. rosae was also significantly reduced when larvae
were kept on leaves treated with moringin for several days. Furthermore, moringin elicited electrophysio-
logical activity in a deterrent-sensitive neuron in the medial maxillary taste sensillum of P. brassicae, provid-
ing a sensory mechanism for the deterrence and the first known ITC taste response of an insect. In simulated
feeding assays, recovery of moringin was high, in accordance with its non-volatile nature. Our results
demonstrate taste-mediated deterrence of a non-volatile, natural ITC to glucosinolate-adapted insects.
Ó 2015 Elsevier Ltd. All rights reserved.
Isothiocyanate
Glucosinolate
HPLC–MS/MS
NMR
Taste
Deterrent
Stimulant
Sensory physiology
Neuron
Specialist herbivores
Brassicales
1. Introduction
against many herbivores, a number of Brassicales specialist insects
can usually avoid ITC formation in the gut thanks to various
Glucosinolates are defensive chemicals in the plant order
Brassicales. Upon tissue disruption they are converted to isothio-
cyanates (ITCs) or other products due to endogenous enzymes,
the myrosinases (b-thioglucoside glucohydrolases; E.C. 3.2.1.147)
(Agerbirk and Olsen, 2012). The ITCs, commonly known as mustard
oils, are reactive, pungent chemicals. Hence, the glucosinolate–
myrosinase system has been aptly named the mustard oil bomb
(Matile, 1980). While the mustard oil bomb is an effective defense
biochemical mechanisms (Müller, 2009; Winde and Wittstock,
2011; Opitz et al., 2011; Beran et al., 2014).
A chewing herbivore will initially come into contact with an ITC
in the water-dissolved form, and dissolved ITC concentrations could
serve to guide its feeding response. Only subsequently, volatile ITCs
will enter the gas phase. Remarkably, excitation of a taste neuron by
an ITC has never been demonstrated. In contrast, in insect-plant
research, a few volatile (commercially available) ITCs have mainly
been used for behavioural studies and olfactory research (Li et al.,
2000; Agrawal and Kurashige, 2003; Barker et al., 2006). However,
much uncertainty regarding suitable methods for toxicity and deter-
rency testing of ITCs is apparent from the literature. First, volatile
⇑
Corresponding authors.
E-mail addresses: caroline.mueller@uni-bielefeld.de (C. Müller), joop.vanloon@
wur.nl (J. van Loon), nia@plen.ku.dk (N. Agerbirk).
http://dx.doi.org/10.1016/j.phytochem.2015.08.007
0031-9422/Ó 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Müller, C., et al. Taste detection of the non-volatile isothiocyanate moringin results in deterrence to glucosinolate-
adapted insect larvae. Phytochemistry (2015), http://dx.doi.org/10.1016/j.phytochem.2015.08.007

2
C. Müller et al. / Phytochemistry xxx (2015) xxx–xxx
ITCs quickly evaporate, making the actual concentrations present in
bioassays difficult to estimate. Second, chemical reactions with
matrix components based on addition of ITCs to various substrates
can occur in these bioassays which typically last hours or longer.
In a pioneering study on ITC effects on insects, allyl ITC was added
to a synthetic diet offered to a glucosinolate-adapted moth, and LC50
The aims of this work were (1) to isolate and characterize a side
chain glycosylated glucosinolate and its corresponding natural,
non-volatile ITC, (2) to test the effects of these compounds on per-
formance, feeding preference and taste physiology of glucosino-
late-adapted insect larvae, and (3) to critically examine the
recovery of the ITC under feeding assay conditions. As glucosino-
late-adapted insects, we chose larvae of the model insects Pieris
napi and Pieris brassicae (Lepidoptera: Pieridae) and Athalia rosae
(Hymenoptera: Tenthredinidae), that represent two different mus-
tard oil bomb avoidance strategies, metabolic diversion in Pieris
spp. (Wittstock et al., 2004; Agerbirk et al., 2006, 2010b; Stauber
et al., 2012) and glucosinolate sequestration with desulfation/sul-
fation in A. rosae (Müller et al., 2001; Opitz et al., 2011;
Abdalsamee et al., 2014), respectively. P. brassicae was included
since it is particularly amenable to electrophysiological experi-
ments (Schoonhoven and van Loon, 2002). For comparison in some
experiments, we included a glucosinolate without the critical gly-
cosylation, p-hydroxybenzylglucosinolate (sinalbin) (3). We
demonstrate that the non-volatile ITC moringin is a suitable model
to test taste-mediated ITC effects on insects due to its lack of
volatility, stimulation of a taste neuron, deterrent effects and high
recovery in bioassays.
values in the low lmol/g diet fresh wt. range were reported (Li et al.,
2000). In a later paper (Agrawal and Kurashige, 2003), it was argued
that studies employing addition of ITCs to synthetic diets had failed
because allyl ITC is volatile and will be released to fumigate the larvae.
Using microencapsulated allyl ITC molecules in synthetic diets, a toxic
effect on larvae of a glucosinolate-adapted butterfly was demon-
strated, with LC50 around 2 lmol/g diet fresh wt. (as judged from pre-
sented graphs) (Agrawal and Kurashige, 2003). A more recent
approach employed dilute solutions of benzyl ITC sprayed every
48 h on test leaves offered to caterpillars (Bejai et al., 2012). In this
study, evaporation of the ITC was suggested to be prevented by using
closed Petri dishes sealed with cellophane tape. However, levels of
benzyl ITC on the test leaves (and in the atmosphere in the Petri
dishes) were not experimentally controlled. From all three reports
(Li et al., 2000; Agrawal and Kurashige, 2003; Bejai et al., 2012) it is
evident that an experimental measurement of the actual levels of ITCs
during biological testing is needed for quantitative conclusions, and
that both volatility and reactivity of the ITC should be considered. A
non-volatile ITC offers a useful opportunity for testing whether ITCs
as such in the liquid phase are deterrent and/or growth-inhibiting
to insect larvae and if they are perceived at the level of taste neurons.
We expected the rare class of side chain glycosylated glucosino-
lates (Kjær et al., 1979; Olsen and Sørensen, 1979; Olsen et al.,
1981; Gueyrard et al., 2000; Bennett et al., 2003; Kim et al.,
2004) to form particularly non-volatile ITCs. A representative
2. Results and discussion
2.1. Isolation and characterization of glucomoringin, moringin and
desulfoglucomoringin
Glucomoringin (1) and moringin (2) were isolated at the
Bologna laboratory (CRA-CIN) from Moringa oleifera seeds using
established methods (Section 3.2), and their structures confirmed
by one and two-dimensional NMR spectroscopic analyses (Sections
3.8 and 3.9). The preparations were essentially pure as judged from
1H NMR spectra and HPLC chromatograms (Section 3.2). The chem-
ical shifts essentially agreed with previously published data
(Gueyrard et al., 2010; de Graaf et al., 2015). For moringin, how-
ever, a possible moderate difference of the chemical shift of the
quaternary ITC carbon was observed as compared to a recent
report using a different solvent (de Graaf et al., 2015). The signal
was very weak as expected for an ITC (Glaser et al., 2015), and
was not detected in the one-dimensional 13C NMR spectrum. The
value reported here (133 ppm extracted from the HMBC spectrum
due to correlation with the methylene H) agreed with the calcu-
lated value using a standard NMR prediction tool (www.nmrdb.
glucosinolate from this structural class is 4-(a-L-rhamnopyranosy-
loxy)benzylglucosinolate (1) (Kjær et al., 1979; Gueyrard et al.,
2000), which is traditionally known from trees of the tropical
genus Moringa but was more recently discovered in herbs from
temperate climates (de Graaf et al., 2015). The common name glu-
comoringin was recently suggested for this compound (Brunelli
et al., 2010), in accordance with the principles of classical glucosi-
nolate nomenclature (Kjær, 1960). According to the same nomen-
clature, the corresponding ITC (4-(a-L-rhamnopyranosyloxy)
benzyl ITC) (2) would be named moringin. Recently, antibiotic
properties as well as promising physiological activities in mam-
mals have been reported for moringin (Galuppo et al., 2013;
Waterman et al., 2014; Giacoppo et al., 2015). However, effects
of side chain glycosylated glucosinolates and their ITCs on insect
herbivores have to our knowledge so far not been investigated.
org). The L-configuration of the rhamnose residue could not be con-
cluded from NMR but was known from a published comparison of
the M. oleifera compound with an authentic, synthetic standard
(Gueyrard et al., 2000).
The ion trap mass spectrum of glucomoringin after analytical
desulfation was of interest because the analytical chemistry of side
chain glycosylated glucosinolates is poorly known. It was mea-
sured along with control spectra of desulfoderivatives of related
glucosinolates, namely benzylglucosinolate, its p-hydroxyl deriva-
tive 3 and the p-methoxyl derivative (glucoaubrietin) (Fig. S1).
These controls served to evaluate the structural basis of specific
cleavages observed for desulfo glucomoringin. The electrospray
ionization with NaCl-spiked eluents produced solely sodium
adducts, whereas proton adducts were not observed. The ion trap
mass spectra of the three control glucosinolates provided evidence
of very similar fragmentation as reported for many other glucosi-
nolates (Agerbirk et al., 2014, 2015). They were dominated by
sodium adducts of thioGlc (base peak, m/z 219) and anhydroGlc
(m/z 185) while the ‘‘type c” adduct of the aglucone after loss of
anhydroGlc was less abundant (16–25% of base peak) (Fig. S1).
The ion trap mass spectrum of the desulfo glucomoringin Na+
Glucomoringin (1)
Moringin (2)
Sinalbin (3)
Please cite this article in press as: Müller, C., et al. Taste detection of the non-volatile isothiocyanate moringin results in deterrence to glucosinolate-
adapted insect larvae. Phytochemistry (2015), http://dx.doi.org/10.1016/j.phytochem.2015.08.007

C. Müller et al. / Phytochemistry xxx (2015) xxx–xxx
3
adduct at m/z 514 differed markedly from the controls. The base
peak at m/z 368 was interpreted as a sinalbin Na+ adduct due to
loss of anhydroRha (loss of 146). The second most intense ion at
m/z 352 was interpreted as a ‘‘type c” Na+ adduct of the aglucone
after loss of anhydroGlc, while the usual fragment adducts at m/z
219 and 185 were of lower abundance. It was concluded that the
rhamnoside bond is very susceptible to cleavage under the ion trap
conditions resulting in an informative fragment ion, in contrast to
the ether bond in the control desulfo glucoaubrietin. Furthermore,
the ‘‘type c” aglucone fragment due to loss of anhydroGlc has a
high affinity for Na+ compared to type c fragments from the less
hydrophilic controls, as expected for a hydrophilic side chain glu-
cosinolate (Agerbirk et al., 2014). As also noticed in the latter
paper, the ion trap mass spectrum of a desulfated glucosinolate
Na+ adduct was different from the electrospray ionization mass
spectrum of the corresponding intact glucosinolate (Bennett
et al., 2003).
Moringin was potentially unstable in aq. MeOH solvent, but
direct HPLC analysis was considered and found feasible. Less than
5% of the moringin was lost during 24 h incubation in 20% aq.
MeOH at room temperature, and the phenol-like chromophore
resulted in good sensitivity in HPLC–UV at around 220 nm. Simi-
larly, HPLC–MS was feasible, and the ion trap mass spectrum was
highly characteristic, suggesting an interesting fragmentation and
rearrangement of the sodium adduct involving (Na+-catalyzed?)
nucleophilic attack of the SCN moiety on the glycoside C in the
ion trap (Fig. 1). The proposed, surprising rearrangement in the
ion trap would explain the detected major fragment ions at m/z
228 and 129 but cannot be considered proven. Similar apparent
rearrangement products have been observed from some cyano-
genic glucosides in ion trap MS/MS, in support of our interpretation
glucosinolate-treated versus control cabbage leaf discs in all three
test concentrations of both glucomoringin and sinalbin (Fig. 2).
The test leaves of B. rapa also contained glucosinolates, but other
than sinalbin and glucomoringin (gluconapin, glucobrassicin and
1-methoxy as well as 4-methoxyderivatives of the latter, mean
31 nmol per leaf disc). The endogenous glucosinolates must have
been already sufficient to stimulate a pronounced feeding response
in P. napi and A. rosae, whereas additional glucosinolates did nei-
ther lead to an increased nor decreased feeding of the specialists.
However, sinalbin in the highest concentration (30 mM) had by
trend a negative effect on feeding by A. rosae larvae (Wilcoxon
matched-pairs test, P = 0.09; Fig. 2). In essence, these results show
that the unusual rhamnoside moiety of glucomoringin did not
confer any deterrence.
Some non-host plants that are consistently rejected for feeding
by specialist herbivorous insects can be rendered acceptable by
surface application of token stimuli for host-plant recognition
(Schoonhoven et al., 2005). We used this approach to test whether
glucomoringin stimulates feeding of P. brassicae larvae. We found
that they exhibited a significant preference for leaf discs of pea
(Pisum sativum L.), a non-host plant, painted with glucomoringin
solution over control discs painted with the solvent. All 10 larvae
tested fed from the glucomoringin-treated leaf discs (median
41.8 mm2). Five larvae did not feed at all from the control discs,
the other five fed small amounts that were on average less than
10% (median 2.9 mm2) of the consumption from glucomoringin-
treated discs (Wilcoxon matched-pairs test, one-sided, T = 4,
P < 0.005). These results demonstrate that glucomoringin acts as
a feeding stimulant to P. brassicae.
In contrast, the ITC moringin (2) deterred larvae of all three
investigated species when applied in different concentrations on
Brassica spp. leaves. In the case of P. napi and P. brassicae, caterpil-
lars were only significantly deterred from feeding by the highest
ITC concentration. In the case of A. rosae, a deterrent effect was
observed at all three test concentrations spanning two orders of
magnitude (Fig. 3). Thus, moringin (or a compound formed as an
experimental artifact due to our addition of moringin) must have
been deterrent to larvae of all three species, in particular to A.
rosae. Due to the high recovery of intact moringin at our conditions
(Section 2.5), we assume for simplicity that the deterrence was a
direct consequence of the added ITC. However, like in any other
experiment involving the addition of an ITC to a complex biological
matrix, effects of reaction products (Agerbirk et al., 2015) or bio-
logically induced gene products (Hara et al., 2013 and references
cited in that work) cannot be excluded until the entire biochemical
mechanism is unraveled. Our investigations of electrophysiological
effects of intact moringin in a plant cell free solution (Section 2.3)
were intended as a first step in this direction. Whether the
observed difference in response to added moringin is a general
ˇ
of the m/z 228 fragment (Picmanová et al., 2015). Due to the quan-
titative robustness of the UV-detector, this detection mode was
selected for quantitative analysis (Section 2.5). Other authors have
likewise reported on the stability of moringin in aq. solution and
have presented LC–MS analysis conditions based on detection of
the acetate adduct in negative mode (Waterman et al., 2014).
2.2. The isothiocyanate moringin acts as deterrent and lowers
performance in contrast to the precursor glucosinolate
Larvae of two Brassicaceae specialists belonging to different
orders, the butterfly P. napi (Lepidoptera: Pieridae) and the sawfly
A. rosae (Hymenoptera: Tenthredinidae), were offered cabbage leaf
discs (Brassica rapa) topically treated with the glucosinolates glu-
comoringin (1) and sinalbin (3) in different concentrations (0.3, 3
or 30 mM corresponding to 12 nmol, 120 nmol or 1200 nmol
added per leaf disc, respectively). The paired-choice
assays revealed no significant differences in consumption on
Intens.
+MS, 10.2min #1183
8
a (Moringin)
x10
334
[M+Na]+
3
2
1
0
7
x10
+MS2(334), 10.2min #1184
b
- HSCN
loss of
0.8
0.6
0.4
0.2
0.0
129
228
169
230 (34S effect)
188
107
100
147
150
200
250
300
350
m/z
Fig. 1. Ion trap mass spectroscopy of moringin (a-b). (a) ESI mass spectrum of moringin at the HPLC–MS peak apex showing almost exclusive formation of the Na+ adduct
from NaCl doped eluent. (b) Ion trap (MS2) mass spectrum of the Na+ adduct, with interpretation of the fragment ions.
Please cite this article in press as: Müller, C., et al. Taste detection of the non-volatile isothiocyanate moringin results in deterrence to glucosinolate-
adapted insect larvae. Phytochemistry (2015), http://dx.doi.org/10.1016/j.phytochem.2015.08.007

4
C. Müller et al. / Phytochemistry xxx (2015) xxx–xxx
P. napi / glucomoringin
P. napi / sinalbin
b
30
25
20
15
10
5
a
30
25
20
15
10
5
Test
Control
0
0
30 mM
3 mM
0.3 mM
30 mM
3 mM
0.3 mM
c
d
A. rosae / glucomoringin
A. rosae / sinalbin
P = 0.09
30
25
20
15
10
5
30
25
20
15
10
5
0
0
30 mM
3 mM
0.3 mM
30 mM
3 mM
0.3 mM
Fig. 2. Mean feeding amount (with standard errors) on Brassica rapa pekinensis leaf discs treated with different amounts of the glucosinolates sinalbin (a, c) and
glucomoringin (b, d) in 80% aq. MeOH and leaf discs treated with the solvent only by larvae of Pieris napi (a, b) and Athalia rosae (c, d). Differences in feeding amounts on test
and control discs were tested by Wilcoxon matched-pairs tests. All differences are not significant (P > 0.05). N = 15.
difference in ITC sensitivity between the hymenopteran and lepi-
dopteran species or relates to the hydrophilic moringin in particu-
lar cannot be distinguished. The stronger response of A. rosae
larvae to moringin could be due to detection by chemoreceptors
or rapid post-ingestive feedback. Adults of A. rosae can detect var-
ious volatile ITCs by chemoreceptors, and experienced adults are
even attracted to them (Barker et al., 2006). Further studies are
necessary to determine how A. rosae larvae sense ITCs and whether
volatile and non-volatile ITCs are detected in similar ways by the
sawfly.
Since A. rosae was most sensitive in the feeding experiment, the
long-term effect of moringin on growth performance of A. rosae
was studied in more detail. Larvae did not grow on leaf discs trea-
ted with 30 mM moringin (Fig. 4) and half of them died. These lar-
vae presumably died from starvation, because moringin (or a
moringin-dependent compound) deterred them from feeding
(Fig. 3) and they consumed only very little. However, additional
toxic effects of moringin or degradation products of moringin on
the larvae cannot be ruled out. With decreasing concentrations of
moringin, growth of larvae increased. Larvae fed with control leaf
discs gained the highest mass. In this control treatment and in
the 0.3 mM moringin treatment, only one larva per treatment
(6.7%) died within seven days.
The high sensitivity of A. rosae towards the ITC is particularly
interesting, since larvae of A. rosae can perform almost equally well
on isogenic lines of Brassica juncea with low or high glucosinolate
concentrations and low or high myrosinase activities (Müller and
Sieling, 2006), with the high glucosinolate/high myrosinase line
potentially producing high levels of volatile ITCs. Thus, they are
tolerant to variation in the glucosinolate–myrosinase system and
can potentially deal well with high levels of both substrate and
enzyme, but must effectively circumvent the formation of ITCs,
since they seem highly sensitive to ITCs (Fig. 3). Indeed, the sawfly
A. rosae is a typical example of a glucosinolate-adapted Brassicales-
specialist herbivore in that the larvae avoid formation of ITCs. The
present demonstration of a high sensitivity of A. rosae larvae to a
deterrent ITC confirms that the larvae are dependent on an effi-
cient recognition of ITCs and avoidance of ITC formation during
feeding. Similar conclusions can be drawn for both Pieris species,
which are also glucosinolate-adapted herbivores avoiding ITC for-
mation, even though they seem to be not as strongly deterred by
moringin as A. rosae. Besides biochemical detoxification mecha-
nisms, there is also evidence that many specialized herbivores
exhibit complex feeding behaviours ensuring optimal detoxifica-
tion of plant defences (e.g. Wittstock and Gershenzon, 2002;
Pentzold et al., 2014). Data in the following section suggest that
effects of ITCs on feeding behaviour may be mediated by taste
recognition.
2.3. Electrophysiological activity of moringin and glucomoringin on
taste neurons
In order to provide an independent test of the sensory proper-
ties of moringin and glucomoringin (as opposed to potential reac-
tion products formed on leaf discs or induced leaf disc
metabolites), sensory physiology was conducted. Because the sen-
sory physiology of P. brassicae larvae is well established, this spe-
cies was chosen for investigation despite the modest sensitivity
to moringin. Taste sensilla present on the maxillary galea of P. bras-
sicae caterpillars were stimulated with moringin and gluco-
moringin to investigate the effect on eight gustatory neurons of
which the reaction spectra have been characterized (Schoo
nhoven and van Loon, 2002). Remarkably, moringin activated a
neuron in the medial sensillum styloconicum. The stimulation
was only observed at 30 mM, the highest concentration tested,
causing relatively low response intensity (Fig. 5a). This is the first
report on an ITC, moringin, triggering electrophysiological activity
in a taste neuron. The typical tonic nature of the firing activity of
this medial neuron (Fig. 5a) suggests it to originate in the deter-
rent-sensitive neuron, consistent with the behavioural deterrence
observed (Fig. 3b). Activity of the medial deterrent-sensitive
neuron in Pieris caterpillars is known to give rise to a prominent
Please cite this article in press as: Müller, C., et al. Taste detection of the non-volatile isothiocyanate moringin results in deterrence to glucosinolate-
adapted insect larvae. Phytochemistry (2015), http://dx.doi.org/10.1016/j.phytochem.2015.08.007

C. Müller et al. / Phytochemistry xxx (2015) xxx–xxx
5
inhibition of feeding behaviour on leaf material of a highly pre-
ferred host-plant (Luo et al., 1995; Zhou et al., 2009). Thus far only
olfactory neurons have been documented that respond to ITCs. The
highly volatile allyl ITC (derived from sinigrin) is detected by olfac-
tory neurons located in the antenna of larval P. brassicae (Visser,
1983). Olfactory neurons in the antennae of the Vetch aphid
Megoura viciae and of the Black bean aphid Aphis fabae (both Hemi-
ptera) respond to volatile ITCs, which act as repellents to the latter
species (Visser and Piron, 1995; Nottingham et al., 1991). It would
be relevant in future studies to investigate structure–activity rela-
tionships of taste neuron responses to ITCs as has been done for
olfactory responses (Barker et al., 2006).
Electrophysiological responses of larval maxillary taste neurons
to glucosinolates have been demonstrated in several Pieris species
(Schoonhoven and van Loon, 2002) as well as in other glucosino-
late-adapted Brassicales-specialist insects (Mitchell and Gregory,
1981; van Loon et al., 2002) and are known to provide the neural
input resulting in strong stimulation of feeding behaviour in these
species (Schoonhoven and Blom, 1988). In our tests on P. brassicae,
glucomoringin and sinigrin, the latter included as known stimulant
to this species, elicited responses from a single neuron in the lateral
sensillum styloconicum only (Fig. 5c and d). Dose response curves
were shallow within the range tested (Fig. 6). Mixtures of the two
glucosinolates also elicited activity in a single neuron (not shown),
indicating that the same neuron responds to both compounds with
differential sensitivity. Glucosinolates differing in the side chain
have also previously been found to stimulate feeding and oviposi-
tion behaviour of herbivorous insects to different extents (David
and Gardiner, 1966; Du et al., 1995; Barker et al., 2006).
a
b
c
25
20
15
10
5
P. napi
Test
Control
0
30 mM
3 mM
5 mM
3 mM
0.3 mM
35
30
25
20
15
10
5
P. brassicae
0
1 mM
30 mM
60
50
40
30
20
10
0
A. rosae
2.4. Recovery of glucosinolates
In feeding assays, in which compounds are artificially applied
on a leaf substrate, it is important to control for the actual amount
30 mM
0.3 mM
Fig. 3. Mean feeding amount (with standard errors) on leaf discs treated with
different amounts of the ITC moringin (2) in 80% aq. MeOH and leaf discs treated
with the solvent only by larvae of Pieris napi (a), P. brassicae (b) and Athalia rosae (c).
Differences in feeding amounts on test and control discs were tested by Wilcoxon
matched-pairs tests (*P < 0.05, ***P < 0.001, other differences are not significant,
P > 0.05). N = 9–15. Leaf discs were taken from Brassica rapa pekinensis in (a) and (c)
and B. oleracea in (b).
25
30 mM
3 mM
a b c c
0.3 mM
control
a b b c
20
15
10
5
a b c c
a b b b
a b b b
n.s.
n.s.
0
d1
d2
d3
d4
d5
d6
d7
Fig. 4. Growth of larvae (mean body mass with standard errors) of Athalia rosae
reared on Brassica rapa pekinensis leaf discs treated with different amounts of the
ITC moringin in 80% aq. MeOH and leaf discs treated with solvent only. Three-day
old larvae were used (d1). In the beginning, 15 larvae per treatment were prepared.
N at d7: 6–14. Lower case letters indicate significant differences between treatment
groups within a day. n.s. – not significant. d – days. Kruskal–Wallis tests followed by
multiple comparisons for each day separately.
Fig. 5. Recordings of electrophysiological activity during the first 1000 ms after
contacting the sensillum tip pore produced by taste neurons in styloconic sensilla
on the maxillary galea of Pieris brassicae final instar larvae to (a) 30 mM moringin,
medial sensillum; (b) solvent (tap water), medial sensillum; (c) 2 mM sinigrin,
lateral sensillum; (d) 2 mM glucomoringin, lateral sensillum; (e) solvent (tap
water), lateral sensillum.
Please cite this article in press as: Müller, C., et al. Taste detection of the non-volatile isothiocyanate moringin results in deterrence to glucosinolate-
adapted insect larvae. Phytochemistry (2015), http://dx.doi.org/10.1016/j.phytochem.2015.08.007

6
C. Müller et al. / Phytochemistry xxx (2015) xxx–xxx
of compound that remains on the leaf over the course of the exper-
iment. The recovery of glucomoringin (1) applied to leaf discs in
simulated bioassays varied between 42% and 72% with no obvious
decline with time or dependency of the applied glucosinolate
amount (Fig. 7a). Although the glucomoringin was free from sinal-
bin contamination according to NMR (Section 3.8), a small amount
of desulfosinalbin (3–7% of the amount of desulfo glucomoringin)
was present in the desulfated extracts of glucomoringin-spiked
leaves as well as in control samples of pure glucomoringin desul-
fated as control, suggesting that a low proportion of the spiked glu-
comoringin had been hydrolyzed at the rhamnoside bond as an
artifact during the analytical desulfation step (Förster et al.,
2015) rather than as a result of plant enzymes. We did not correct
the recovery for this minor and uncertain error. The recovery of
sinalbin applied to leaf discs in simulated bioassays varied
between 57% and 83%, and the recovery was slightly reduced over
time (N = 5 per time-point and treatment) (Fig. 7b). Hence, the
presence and approximate intended levels of both glucosinolates
were confirmed under our bioassay conditions.
100
80
60
40
20
0
a
1,200 nmol glucomoringin
120 nmol glucomoringin
12 nmol glucomoringin
0
1
2
3
4
5
Time (h)
100
80
60
40
20
0
b
2.5. Recovery of the non-volatile moringin (2) was high in feeding
assays
To test whether moringin was stable at the feeding assay condi-
tions (and not lost due to reaction with leaf constituents), a recov-
ery experiment with determination of the free moringin by direct
HPLC was carried out using the usual evaporation of solvent fol-
lowed by incubation at feeding assay-like conditions for 3 h
(Fig. 8). Two levels tested, 1200 and 120 nmol per leaf disc, did
not give significantly different relative recoveries, and the mean
relative recovery for the combined experiment was 63% (SD 21%,
N = 11). We concluded that moringin was relatively stable after
spiking to leaf discs and could be recovered as free, non-conjugated
moringin. Additional experiments, aimed at including potential
reactive adducts such as dithiocarbamates, showed slightly higher
recoveries, suggesting that a minor fraction of the applied morin-
gin could be present as reactive conjugates that might contribute
to the deterrence (Agerbirk et al., 2015). However, the dominance
of free moringin and the sensory physiological activity of free mor-
ingin (Section 2.3) suggested this compound to be important for
the deterrence observed.
1,200 nmol sinalbin
120 nmol sinalbin
12 nmol sinalbin
0
1
2
3
4
5
Time (h)
Fig. 7. Recovery (mean and standard errors) of the glucosinolates glucomoringin (a)
and sinalbin (b) at different time-points after application on Brassica rapa leaf discs
at simulated bioassay conditions. Amounts applied per leaf disc were either
1200 nmol (squares), 120 nmol (diamonds) or 12 nmol (triangles). N = 5.
ITC moringin, tested in the present study, shows a high stability
and its biological activity as a deterrent could be demonstrated
for several herbivores (Section 2.2). As reactive molecules, all ITCs
will decrease in concentration in the presence of nucleophiles such
as cysteine and proteins (Nakamura et al., 2009; Brown and
Hampton, 2011; Hanschen et al., 2012). In a companion paper, an
analysis method capable of including reactive dithiocarbamate
adducts of ITCs was devised (Agerbirk et al., 2015). Preliminary
results suggest that a significant part of recovered volatile benzyl
ITC was in the form of a reactive adduct (possibly with macro-
molecules), while the majority of the non-volatile, hydrophilic
moringin was in the free form even after extended incubation on
leaf discs. Hence, classical investigations of deterrence and/or
repellence of ITCs to glucosinolate-adapted insects should possibly
be reconsidered, as reactive adducts may be responsible for the
effects, at least in part. In conclusion, moringin would seem to be
a useful compound for investigating responses of insects towards
ITCs and their detection by insect taste organs due to its lack of
volatility and apparently low tendency to form adducts on leaf sur-
faces. For this research, moringin was chosen for its absolute lack
of volatility. Certainly, other glycosylated ITCs could be equally
useful, and it is possible that ITCs carrying sulfinyl or sulfonyl func-
tionalities could be used similarly. Indeed, it would be very rele-
vant to investigate whether the deterrent and taste neuron
stimulating effects of moringin reported here can be considered
general effects of ITCs or are related to the particular structure of
moringin.
The ITCs previously tested in chemo-ecological studies were
volatile, and a significant fraction of ITCs tested in feeding assays
(Li et al., 2000; Agrawal and Kurashige, 2003; Bejai et al., 2012)
thus must have been in the gas phase. In contrast, the non-volatile
80
70
60
50
40
30
Glucomoringin
Sinigrin
20
10
0
0
5
10
Concentraꢀon (mM)
Fig. 6. Electrophysiological activity (means and standard errors) in a single taste
neuron in the lateral maxillary sensillum styloconicum in final instar larvae of Pieris
brassicae to three concentrations of sinigrin and glucomoringin (1). N = 5.
Please cite this article in press as: Müller, C., et al. Taste detection of the non-volatile isothiocyanate moringin results in deterrence to glucosinolate-
adapted insect larvae. Phytochemistry (2015), http://dx.doi.org/10.1016/j.phytochem.2015.08.007

C. Müller et al. / Phytochemistry xxx (2015) xxx–xxx
7
biochemical difference between sinalbin and glucomoringin, con-
ferred by the Rha moiety, would hence be the formation of a stable
and very hydrophilic ITC from glucomoringin versus the formation
of a highly labile and moderately lipophilic ITC from sinalbin.
Natural acetylation of the Rha moiety results in a more lipophilic
ITC, 4-acetylmoringin, which also has an interesting bioactivity
(Park et al., 2011; Waterman et al., 2014). Despite the stability of
moringin, the ITC functional group showed the usual reactivity
with a nucleophile (Agerbirk et al., 2015). The stability, reaction
with a nucleophile and biological deterrence of moringin reported
here suggests that a major ecological effect of converting sinalbin
to glucomoringin in some plants is to obtain a conventional
mustard oil bomb producing an ITC product with usual defensive
chemical reactivity but high hydrophilicity and lack of volatility.
Whether this ‘‘novel weapon” (Barto et al., 2010) has ecological
consequences needs to be tested, but the deterrent effects of
moringin described here would seem to support this hypothesis.
2
a
mAU
/nm
b
mAU
2
3. Experimental
/nm
3.1. Chemicals and general conditions
The HPLC–MS/MS instrument was an Agilent 1100 Series HPLC
(Agilent Technologies, Germany) coupled to a Bruker HCT-Ultra ion
trap mass spectrometer (Bruker Daltonics, Bremen, Germany) fit-
ted with a Zorbax C18 column (Agilent; 1.8
l
m, 2.1 ꢀ 50 mm),
c
and was operated exactly as described elsewhere (Agerbirk et al.,
2015). Instruments for HPLC-PDA and NMR were as previously
described (Agerbirk and Olsen, 2012) unless otherwise indicated
(Section 3.6). The HPLC–PDA instrument was equipped with a Luna
phenylhexyl column (5
l
m, 250 ꢀ 4.5 mm) (Phenomenex) with
PDA detection at 210–370 nm, gradient programmes and detection
wavelength for quantitation (band width 8 nm) depended on the
specific analyte as detailed in the relevant sections. Chemical
shifts in NMR for D2O as solvent were relative to those of dioxane
(dH = 3.75, dC = 67.4) and for CD3OD as solvent were relative to
those of TMS at 0 ppm. Sinalbin was prepared as previously
reported (Baasanjav-Gerber et al., 2011). Benzylglucosinolate
for desulfation were from Merck Eurolab (Darmstadt, Germany),
and desulfo 4-methoxybenzylglucosinolate (glucoaubrietin) was
isolated as previously reported (Agerbirk et al., 2008).
Fig. 8. High recovery of native moringin (2) from a leaf disc after incubation at
feeding assay-like conditions as measured by direct HPLC–PDA at 221 nm. (a)
Moringin control corresponding to applied amount. (b) Recovered moringin from
leaf disc. (c) Control extract of leaf disc without moringin spike. Each insert shows
the PDA UV spectrum at the peak apex.
2.6. Chemical ecology of glucomoringin as compared to sinalbin
The Rha moiety of glucomoringin is a rare structural element in
glucosinolate biochemistry, and our investigation of its activities to
insects was first and foremost an attempt to investigate general
effects of the ITC hydrolysis product of a glucosinolate in the
solid/liquid phase. Any relevance for the natural defense of Moringa
trees was not our purpose and cannot be concluded from this
study. However, the recent discovery of glucomoringin in a poten-
tial host plant of the tested insect species (de Graaf et al., 2015)
means that the deterrent activities to the tested species may be
of ecological significance.
The biosynthesis of glucomoringin occurs probably by rhamno-
sylation of sinalbin. A recent report demonstrates an interesting
polymorphism in a wild Brassicaceae species, Noccaea caerulescens,
with some populations dominated by sinalbin and others by gluco-
moringin (de Graaf et al., 2015). Sinalbin has originated de novo
repeatedly in the evolution of Brassicaceae (Agerbirk et al., 2008),
and the biochemical basis of the ecological success of sinalbin-
containing species, as well as complex interactions with glucosino-
late-adapted herbivores, have been discussed elsewhere (Müller
et al., 2002; Müller and Brakefield, 2003; Brader et al., 2006;
Agerbirk et al., 2007, 2008, 2010a). In contrast to the stability of
the ITC from glucomoringin, the labile ITC from sinalbin is hydro-
lyzed relatively quickly to thiocyanate ion and 4-hydroxybenzyl
alcohol or an ascorbic acid adduct (Buskov et al., 2000; Agerbirk
et al., 2008) (although the ITC could be isolated from plants
using water-free solvents) (Pedras and Smith, 1997). The key
3.2. Isolation and purity of glucomoringin and moringin
Glucomoringin was isolated from M. oleifera Lam. (Moringaceae)
as previously reported (Brunelli et al., 2010). Briefly, the glucosino-
late was purified in two sequential steps, isolation through anion
exchange followed by gel filtration. The identity was confirmed by
NMR spectroscopy (Fig. S2). The purity was confirmed by inspection
of the 1H NMR spectrum for traces of other glucosinolates (Fig. S2). In
particular, sinalbin (aromatic signals at slightly lower chemical shift
than for glucomoringin) was searched for but not detected in the iso-
lated preparation. An additional purity estimate (99.4% of total HPLC
area at 229 nm) was obtained from HPLC of intact glucosinolates
under the conditions of Agerbirk et al. (2010b), based on a chro-
matogram with peak absorption of glucomoringin around 0.8 ‘‘AU”
to optimize detection of any trace contaminants.
Moringin was produced via myrosinase-catalyzed hydrolysis of
glucomoringin in phosphate buffer (pH 6.5) at 37 °C and purified
by reverse-phase chromatography as reported (Brunelli et al.,
2010). The identity and purity was confirmed by NMR spec-
troscopy (Fig. S3), and an additional purity estimate (99.3% of total
HPLC area at 221 nm) was obtained from HPLC under the condi-
tions of Section 3.7, based on a chromatogram with peak absorp-
tion of moringin around 1 ‘‘AU” to optimize detection of any
trace contaminants.
Please cite this article in press as: Müller, C., et al. Taste detection of the non-volatile isothiocyanate moringin results in deterrence to glucosinolate-
adapted insect larvae. Phytochemistry (2015), http://dx.doi.org/10.1016/j.phytochem.2015.08.007

8
C. Müller et al. / Phytochemistry xxx (2015) xxx–xxx
3.3. Insects
arately, the body mass of larvae reared on different diets was com-
pared by Kruskal–Wallis tests followed by multiple comparisons.
Adults of P. napi L. were collected in the field in Blankenrode
(Germany) and offered Brassica rapa var. pekinensis for oviposition.
The resulting larvae were maintained in a growth chamber at
20 °C, 70% relative humidity and a photoperiod of L16:D8 until
use in bioassays. Larvae of P. brassicae L. were taken from a colony
maintained on cabbage plants (Brassica oleracea L. var. gemmifera
cv. Cyrus) in an air-conditioned room at 25 2 °C, 60–70% relative
humidity and a photoperiod of L16:D8. Larvae of P. brassicae for
feeding and electrophysiological assays were reared from neonate
to the fourth or fifth instar on cabbage plants grown in an air-con-
ditioned greenhouse under long day conditions at 18–22 °C. Exper-
imental larvae had finished ecdysis 24–48 h before testing and had
been starved for 2 h. Larvae of A. rosae L. were taken from a colony
originally collected in the field in Bielefeld (Germany) and kept for
several generations on a mixture of Sinapis alba and B. rapa
pekinensis in a greenhouse at 20 °C and a photoperiod of L16:D8.
3.5. Electrophysiology
The tip recording technique was used to record responses from
taste neurons in the sensilla styloconica on the maxillary galea of P.
brassicae final instar larvae (van Loon and Schoonhoven, 1999).
Excised head capsules were mounted on a silver wire electrode
which was connected to the input of a pre-amplifier (Syntech Taste
Probe DTP-1, Hilversum, The Netherlands). Stimulus solutions of
0.4, 2 and 10 mM glucomoringin and sinigrin (2-propenyl glucosi-
nolate; Janssen Pharmaceuticals, Beerse, Belgium) or moringin (0.3,
3 and 30 mM) were filled into glass micropipettes with a tip diam-
eter of ca. 30 lm. Amplified signals were digitized by an A/D-inter-
face (Syntech IDAC-4) and sampled into a personal computer.
Electrophysiological responses were quantified by counting the
number of spikes in the first 1000 ms after the start of stimulation,
testing in total five larvae. Spikes were counted visually by the
experimenter with the aid of Autospike v. 3.7 software (Syntech,
Hilversum, The Netherlands).
3.4. Feeding and performance assays
To test the activities of sinalbin, glucomoringin and moringin on
feeding behaviour of larvae, paired-choice feeding assays were car-
ried out. Two leaf discs (1 cm diameter) of B. rapa pekinensis (var.
Nagaoka, obtained commercially) for P. napi and A. rosae. For P.
brassicae, moringin bioassays were carried out using B. oleracea
(var. gemmifera cv. Cyrus), whereas glucomoringin was applied
on pea, P. sativum L. (cv. Oregon sugar snap). Leaf discs were cut
with a cork borer and placed in a Petri dish (55 mm diameter for
P. napi and A. rosae and 90 mm for P. brassicae), lined with moist-
ened filter paper, approximately 15 mm apart. One disc was trea-
ted with 40 lL of the relevant compound in 80% aq. MeOH in
solutions of 30, 3, or 0.3 mM (corresponding to 1200, 120 and
12 nmol per leaf disc of about 28 mg fresh wt.) for P. napi and A.
rosae, or 20 lL solutions of 30 mM glucomoringin and 30, 5 and
1 mM moringin, all in 0.5% Tween 20 in tap water, for P. brassicae.
The other disc was treated with the same amount of solvent as
control. Third instar larvae of P. napi and A. rosae and fourth instar
(moringin) or fifth instar (glucomoringin) larvae of P. brassicae
were starved for 2 h and then placed individually in the Petri
dishes once the solvent had evaporated. Larvae were allowed to
feed for 5 h (P. napi), 2 h or 6 h (P. brassicae, moringin and gluco-
moringin respectively) and 3.5 h (A. rosae), due to their different
feeding rates. Afterwards, the amount of feeding on each disc
was determined by measuring the area eaten in mm2. The
paired-choice assays were replicated 15 times per compound and
concentration for P. napi and A. rosae and 9 times (moringin) or
10 times (glucomoringin) for P. brassicae. In order to test for differ-
ences in feeding amounts on test versus control discs within each
treatment, Wilcoxon matched-pairs tests were carried out.
3.6. Recovery of glucosinolates under bioassay conditions
To evaluate the recovery of glucosinolates applied artificially on
leaves in the bioassays, leaf discs of B. rapa pekinensis were treated
with solutions of sinalbin and glucomoringin in different concen-
trations as in the feeding bioassay, and placed in a Petri dish lined
with moistened filter paper. After evaporation of the solvent (about
1 h under a fume hood), a subset of leaf discs was immediately
taken and frozen. Further leaf discs were taken after 2 h and 5 h,
respectively (N = 5 per time-point, glucosinolate and concentra-
tion). After lyophilisation, glucosinolate amounts on the leaves
were quantified by extraction with 80% MeOH, conversion to
desulfoglucosinolates and analysis by HPLC–PDA as previously
reported, except that sinigrin was used as internal standard
(Abdalsamee and Müller, 2013).
3.7. Recovery of moringin under bioassay conditions
To test the recovery of moringin under bioassay conditions,
white cabbage was obtained at a Copenhagen grocery store. Leaf
discs (1 cm diameter) were treated with 40 lL of moringin solution
in 80% aq. MeOH, either 30 mM, 3 mM or 0 mM (blank). After evap-
oration of the solvent (80 min), the leaf discs were transferred to
simulated feeding assays in glass Petri dishes with lids (9 cm diam-
eter, 1.6 cm inner height) containing moist filter paper, and incu-
bated for 3 h. For quantification, (linear) standard curves were
produced, using 0, 10, 20, 30 and 40 lL of each moringin solution.
Control experiments consisted of solvent blanks using leaf discs
and the same amounts of moringin evaporated in HPLC vials and
incubated in parallel. After the incubation, moringin was carefully
To test the long-term effects of moringin on growth and sur-
vival of A. rosae, which responded most sensitively in the feeding
assays, 3-day old larvae were kept individually in Petri dishes
and fed with leaf discs of cabbage treated with different concentra-
tions of moringin. Food was prepared by cutting leaf discs (1 cm
extracted from leaf discs or control vials in 2 ꢀ 750
MeOH and quantitated by HPLC on the same day. The HPLC injec-
tion volume was 10 L for samples from the 30 mM experiment
and 20 L for samples from the 3 mM experiment. Eluents were
lL 20% aq.
l
l
diameter) and treating each with 40
l
L of 30, 3, or 0.3 mM morin-
H2O (A) and MeOH (B), and the gradient program was: 0–2 min,
isocratic with A, 2–40 min, linear gradient from pure A to pure B,
40–46 min, 100% B, 46–50 min, linear gradient from B to A, 50–
57 min, isocratic A. The tR of 2 was 36 min.
gin in 80% MeOH or 80% MeOH only as control. The solvent was
allowed to evaporate for 45 min. Afterwards, larvae were offered
leaf discs in Petri dishes (55 mm diameter), lined with moistened
filter paper. Every day, the body mass of each larva was deter-
mined, starting off with 15 larvae per concentration. The food
was replaced daily by freshly prepared food. At day 7 for the
0.3 mM moringin treatment and at days 6 and 7 for the control
treatment, two leaf discs per larva were offered as most larvae
had consumed 100% of the food the day before. For each day sep-
3.8. Characterization of glucomoringin (1)
NMR spectra were initially run in CD3OD, but for comparison
with literature data NMR was repeated with D2O solvent. 1H
NMR (400 MHz, D2O) data for the isolated glucomoringin
Please cite this article in press as: Müller, C., et al. Taste detection of the non-volatile isothiocyanate moringin results in deterrence to glucosinolate-
adapted insect larvae. Phytochemistry (2015), http://dx.doi.org/10.1016/j.phytochem.2015.08.007

C. Müller et al. / Phytochemistry xxx (2015) xxx–xxx
9
Agerbirk, N., Olsen, C.E., Chew, F.S., Ørgaard, M., 2010a. Variable glucosinolate
profiles of Cardamine pratensis (Brassicaceae) with equal chromosome numbers.
J. Agric. Food Chem. 58, 4693–4700.
Agerbirk, N., Olsen, C.E., Poulsen, E., Jacobsen, N., Hansen, P.R., 2010b. Complex
metabolism of aromatic glucosinolates in Pieris rapae caterpillars involving
nitrile formation, hydroxylation, demethylation, sulfation, and host plant
dependent carboxylic acid formation. Insect Biochem. Mol. Biol. 40,
126–137.
Agerbirk, N., Olsen, C.E., Cipollini, D., Ørgaard, M., Linde-Laursen, I., Chew, F.S., 2014.
Specific glucosinolate analysis reveals variable levels of epimeric
glucobarbarins, dietary precursors of 5-phenyloxazolidine-2-thiones, in
watercress types with contrasting chromosome number. J. Agric. Food Chem.
62, 9586–9596.
Agerbirk, N., Olsen, C.E., Heimes, C., Christensen, S., Bak, S., Hauser, T., 2015.
Multiple hydroxyphenethyl glucosinolate isomers and their tandem mass
spectrometric distinction in a geographically structured polymorphism in the
crucifer Barbarea vulgaris. Phytochemistry 115, 130–142.
Agerbirk, N., De Nicola, G.R., Olsen, C.E., Müller, C., Iori, R., 2015. Derivatisation of
isothiocyanates and their reactive adducts for chromatographic analysis.
Phytochemistry. http://dx.doi.org/10.1016/j.phytochem.2015.06.004.
Agrawal, A.A., Kurashige, N.S., 2003. A role for isothiocyanates in plant resistance
against the specialist herbivore Pieris rapae. J. Chem. Ecol. 29, 1403–1415.
Baasanjav-Gerber, C., Monien, B.H., Mewis, I., Schreiner, M., Barillari, J., Iori, R., Glatt,
H., 2011. Identification of glucosinolate congeners able to form DNA adducts
and to induce mutations upon activation by myrosinase. Mol. Nutr. Food Res.
55, 783–792.
Barker, A.M., Molotsane, R., Müller, C., Schaffner, U., Städler, E., 2006. Chemosensory
and behavioural responses of the turnip sawfly, Athalia rosae, to glucosinolates
and isothiocyanates. Chemoecology 16, 209–218.
Barto, E.K., Powell, J.R., Cipollini, D., 2010. How novel are the chemical weapons of
garlic mustard in North American understories? Biol. Invasions 12, 3465–3471.
Bejai, S., Fridborg, I., Ekbom, B., 2012. Varied response of Spodoptera littoralis against
Arabidopsis thaliana with metabolically engineered glucosinolate profiles. Plant
Physiol. Biochem. 50, 72–78.
(Section 3.1.) essentially agreed with data for the synthetic com-
pound reported by Gueyrard et al. (2010), including the surprising
observation of a multiplet for the anomeric thioGlc H also reported
by those authors for the rhamnoside, but not for the glucoside or
galactoside. COSY showed coupling not only to thioGlc H2 but also
long range (H–C–C–C–H) coupling to thioGlc H3, providing an
explanation for the surprising multiplicity, possibly a higher order
multiplet due to steric effects of the rhamnoside residue on the
configuration of the thioGlc residue. Inspection of 2D spectra
(COSY, HSQC, HMBC) in CD3OD solvent further confirmed the
structure, the proton spectrum in this solvent is shown as Fig. S2.
13C NMR (100 MHz, D2O) agreed with data reported by Gueyrard
et al. (2010) except that all observed values were 0.7 ppm lower
than values in that paper. This difference was attributed to uncer-
tainty concerning the internal standard used with D2O solvent by
Gueyrard et al. (2010). The desulfated derivative was prepared as
previously reported (Agerbirk and Olsen, 2012). Despite absence
of sinalbin in the intact glucomoringin, the eluate from desulfation
was not pure but included desulfo sinalbin apparently due to
hydrolysis of the rhamnoside bond during incubation with the
semi pure sulfatase preparation (Förster et al., 2015). However,
the HPLC separated the sinalbin contaminant from desulfo-
moringin before UV and MS. HPLC–UV of desulfated derivative:
(kmax/nm) (rel. intensity): 223 (100%), 272 (7%), 280 (sh, 5%).
HPLC–MS/MS of the desulfated derivative: see Fig. S1.
Bennett, R.N., Mellon, F.A., Foidl, N., Pratt, J.H., Dupont, M.S., Perkins, L., Kroon, P.A.,
2003. Profiling glucosinolates and phenolics in vegetative and reproductive
tissues of the multipurpose tress Moringa oleifera L. (horseradish tree) and
Moringa stenopetala L. J. Agric. Food Chem. 51, 3546–3553.
Beran, F., Pauchet, Y., Kunert, G., Reichelt, M., Wielsch, N., Vogel, H., Reinecke, A.,
Svatos, A., Mewis, I., Schmid, D., Ramasamy, S., Ulrichs, C., Hansson, B.S.,
Gershenzon, J., Heckel, D.G., 2014. Phyllotreta striolata flea beetles use host plant
defense compounds to create their own glucosinolate-myrosinase system. Proc.
Natl. Acad. Sci. U.S.A. 111, 7349–7354.
Brader, G., Mikkelsen, M.D., Halkier, B.A., Palva, E.T., 2006. Altering glucosinolate
profiles modulates disease resistance in plants. Plant J. 34, 1311–1321.
Brown, K.K., Hampton, M.B., 2011. Biological targets of isothiocyanates. Biochim.
Biophys. Acta 1810, 888–894.
Brunelli, D., Tavecchio, M., Falcioni, C., Frapolli, R., Erba, E., Iori, R., Rollin, P., Barillari,
J., Manzotti, C., Morazzoni, P., D’Incalci, M., 2010. The isothiocyanate produced
from glucomoringin inhibits NF-kB and reduces myeloma growth in nude mice
in vivo. Biochem. Pharmacol. 79, 1141–1148.
3.9. Characterization of moringin (2)
1H NMR (400 MHz; CD3OD) (Fig. S3), d (assignment, #H, multi-
plicity, J/Hz): 7.30 (2 + 6, 2H, d, 8.9), 7.10 (3 + 5, 2H, d, 8.9), 5.4 (10,
1H, d, 1.9), 4.69 (CH2, 2H, s), 3.99 (20, 1H, dd, 3.6/1.8), 3.83 (30, 1H,
dd, 9.5/3.7), 3.62 (50, 1H, m), 3.45 (40, 1H, t, 9.5), 1.22 (60, 3H, d, 6.1).
13C NMR (100 MHz, CD3OD) d 157.8 (4), 133 (AN@C@S, only
observed as HMBC correlation with methylene H), 130.1 (1),
129.7 (2 + 6), 117.9 (3 + 5), 99.9 (10), 73.9 (40), 72.3 (30), 72.0 (20),
70.8 (50), 49.0 (CH2, hidden under solvent signal, revealed in DEPT),
18.1 (60). HPLC–UV: see Fig. 8. HPLC–MS/MS: see Fig. 1.
Acknowledgements
Buskov, S., Hasselstrom, J., Olsen, C.E., Sørensen, H., Sørensen, J.C., Sørensen, S., 2000.
Supercritical fluid chromatography as
determination of 4-hydroxybenzylglucosinolate degradation products. J.
Biochem. Biophys. Methods 43, 157–174.
a method of analysis for the
We thank two anonymous reviewers for their helpful
comments.
David, W.A.L., Gardiner, B.O., 1966. Mustard oil glucosides as feeding stimulants for
Pieris brassicae larvae in a semi-synthetic diet. Entomol. Exp. Appl. 9, 247–255.
de Graaf, R.M., Krosse, S., Swolfs, A.E.M., Te Brinke, E., Prill, N., Leimu, R., van Galen,
P.M., Wang, Y., Aarts, M.G.M., van Dam, N.M., 2015. Isolation and identification
Appendix A. Supplementary data
of 4-a-rhamnosyloxy benzyl glucosinolate in Noccaea caerulescens showing
intraspecific variation. Phytochemistry 110, 166–171.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.phytochem.2015.
08.007. These data include MOL files and InChiKeys of the most
important compounds described in this article.
Du, Y.-J., Van Loon, J.J.A., Renwick, J.A.A., 1995. Contact chemoreception of
oviposition stimulating glucosinolates and an oviposition deterrent
cardenolide in two subspecies of Pieris napi. Physiol. Entomol. 20, 164–174.
Förster, N., Ulrichs, C., Schreiner, M., Müller, C.T., Mewis, I., 2015. Development of a
reliable extraction and quantification method for glucosinolates in Moringa
oleifera. Food Chem. 166, 456–464.
Galuppo, M., Nicola, G.R., Iori, R., Dell’utri, P., Bramanti, P., Mazzon, E., 2013.
Antibacterial activity of glucomoringin bioactivated with myrosinase against
two important pathogens affecting the health of long-term patients in hospitals.
Molecules 18, 4340–4348.
References
Abdalsamee, M.K., Müller, C., 2013. Effects of indole glucosinolates on performance
and sequestration by the sawfly Athalia rosae and consequences of feeding on
the plant defense system. J. Chem. Ecol. 38, 1366–1375.
Abdalsamee, M.K., Giampà, M., Niehaus, K., Müller, C., 2014. Rapid incorporation of
glucosinolates as a strategy used by a herbivore to prevent activation by
myrosinases. Insect Biochem. Mol. Biol. 52, 115–123.
Agerbirk, N., Olsen, C.E., 2012. Glucosinolate structures in evolution.
Phytochemistry 77, 16–45.
Agerbirk, N., Müller, C., Olsen, C.E., Chew, F.S., 2006. A common pathway for
metabolism of 4-hydroxybenzylglucosinolate in Pieris and Anthocaris
(Lepidoptera: Pieridae). Biochem. Syst. Ecol. 34, 189–198.
Agerbirk, N., Olsen, C.E., Topbjerg, H.B., Sørensen, J.C., 2007. Host plant dependent
metabolism of 4-hydroxybenzylglucosinolate in Pieris rapae: substrate
specificity and effects of genetic modification and plant nitrile hydratase.
Insect Biochem. Mol. Biol. 37, 1119–1130.
Agerbirk, N., Warwick, S.I., Hansen, P.R., Olsen, C.E., 2008. Sinapis phylogeny and
evolution of glucosinolates and specific nitrile degrading enzymes.
Phytochemistry 69, 2937–2949.
Giacoppo, S., Galuppo, M., De Nicola, G.R., Iori, R., Bramanti, P., Mazzon, E., 2015. 4-
(a-L-Rhamnosyloxy)-benzyl isothiocyanate, a bioactive phytochemical that
attenuates secondary damage in an experimental model of spinal cord injury.
Bioorg. Med. Chem. 23, 80–88.
Glaser, R., Hillebrand, R., Wycoff, W., Camasta, C., Gates, K.S., 2015. Near-silence of
isothiocyanate carbon in 13C NMR spectra: a case study of allyl isothiocyanate. J.
Org. Chem. 80, 4360–4369.
Gueyrard, D., Barillari, J., Iori, R., Palmieri, S., Rollin, P., 2000. First synthesis of an O-
glycosylated glucosinolate isolated from Moringa oleifera. Tetrahedron Lett. 41,
8307–8309.
Gueyrard, D., Iori, R., Tatibouët, A., Rollin, P., 2010. Glucosinolate chemistry:
synthesis of O-glycosylated derivatives of glucosinalbin. Eur. J. Org. Chem.,
3657–3664
Hanschen, F.S., Brüggemann, N., Brodehl., A., Mewis, I., Schreiner, M., Rohn, S., Kroh,
L.W., 2012. Characterization of products from the reaction of glucosinolate-
Please cite this article in press as: Müller, C., et al. Taste detection of the non-volatile isothiocyanate moringin results in deterrence to glucosinolate-
adapted insect larvae. Phytochemistry (2015), http://dx.doi.org/10.1016/j.phytochem.2015.08.007

10
C. Müller et al. / Phytochemistry xxx (2015) xxx–xxx
derived isothiocyanates with cysteine and lysine derivatives formed in either
model systems or broccoli sprouts. J. Agric. Food Chem. 60, 7735–7745.
Hara, M., Harazaki, A., Tabata, K., 2013. Administration of isothiocyanates enhances
Park, E.J., Cheenpracha, S., Chang, L.C., Kondratyuk, T.P., Pezzuto, J.M., 2011.
Inhibition of lipopolysaccharide-induced cyclooxygenase-2 and inducible
nitric oxide synthase expression by 4-[(20-O-acetyl-
a-L-rhamnosyloxy)benzyl]
heat tolerance in Arabidopsis thaliana. Plant Growth Regul. 69, 71–77.
Kim, S.-J., Jin, S., Ishii, G., 2004. Isolation and structural elucidation of 4-(b-D-
glucopyranosyldisulfanyl)butyl glucosinolate from leaves of rocket salad (Eruca
sativa L.) and its antioxidant activity. Biosci. Biotechnol. Biochem. 68, 2444–2450.
Kjær, A., 1960. Naturally derived isothiocyanates (mustard oils) and their parent
glucosides. Fortschr. Chem. Org. Naturst. 18, 122–176.
isothiocyanate from Moringa oleifera. Nutr. Cancer 63, 971–982.
Pedras, M.S.C., Smith, K.C., 1997. Sinalexin, a phytoalexin from white mustard
elicited by destruxin B and Alternaria brassicae. Phytochemistry 46, 833–837.
Pentzold, S., Zagrobelny, M., Roelsgaard, P.S., Møller, B.L., Bak, S., 2014. The multiple
strategies of an insect herbivore to overcome plant cyanogenic glucoside
defence. PLoS ONE 9, e91337.
ˇ
Kjær, A., Malver, O., El-Menshawi, B., Reisch, et al., 1979. Isothiocyanates in myrosinase-
treated seed extracts of Moringa peregrina. Phytochemistry 18, 1485–1487.
Li, Q., Eigenbrode, S.D., Stringham, G.R., Thiagarajah, M.R., 2000. Feeding and growth
of Plutella xylostella and Spodoptera eridania on Brassica juncea with varying
glucosinolate concentrations and myrosinase activities. J. Chem. Ecol. 26, 2401–2419.
Luo, L.-E., Van Loon, J.J.A., Schoonhoven, L.M., 1995. Behavioural and sensory
responses to some neem compounds by Pieris brassicae larvae. Physiol. Entomol.
20, 134–140.
Picmanová, M., Neilson, E.H., Motawia, M.S., Olsen, C.E., Agerbirk, N., Gray, C.J.,
Flitsch, S., Meier, S., Silvestro, D., Jørgensen, K., Sánchez-Pérez, R., Møller, B.L.,
Bjarnholt, N., 2015. A recycling pathway for cyanogenic glycosides evidenced by
the comparative metabolic profiling in three cyanogenic plant species. Biochem.
J. 469, 375–389.
Schoonhoven, L.M., Blom, F., 1988. Chemoreception and feeding behaviour in a
caterpillar: towards a model of brain functioning in insects. Entomol. Exp. Appl.
49, 123–129.
Matile, P., 1980. The mustard oil bomb – compartmentation of the myrosinase
system. Biochem. Physiol. Pflanzen 175, 722–731.
Mitchell, B.K., Gregory, P., 1981. Physiology of the lateral galeal sensillum in red
Schoonhoven, L.M., Van Loon, J.J.A., 2002. An inventory of taste in caterpillars: each
species its own key. Acta Zool. Acad. Sci. Hung. 48, 215–263.
Schoonhoven, L.M., Van Loon, J.J.A., Dicke, M., 2005. Insect-Plant Biology, 2nd ed.
Oxford University Press, Oxford, U.K.
turnip beetle larvae (entomoscelis-American brown)
– responses to NaCl,
glucosinolates and other glycosides. J. Comp. Physiol. 144, 495–501.
Müller, C., 2009. Interactions between glucosinolate- and myrosinase-containing
plants and the sawfly Athalia rosae. Phytochem. Rev. 8, 121–134.
Müller, C., Brakefield, P.M., 2003. Analysis of a chemical defense in sawfly larvae: easy
bleeding targets predatory wasps in late summer. J. Chem. Ecol. 29, 2683–2694.
Müller, C., Sieling, N., 2006. Effects of glucosinolate and myrosinase levels in
Brassica juncea on a glucosinolate-sequestering herbivore – and vice versa.
Chemoecology 16, 191–201.
Müller, C., Agerbirk, N., Olsen, C.E., Boeve, J.-L., Schaffner, U., Brakefield, P.M., 2001.
Sequestration of host plant glucosinolates in the defensive hemolymph of the
sawfly Athalia rosae. J. Chem. Ecol. 27, 2505–2516.
Müller, C., Boeve, J.L., Brakefield, P.M., 2002. Host plant derived feeding deterrence
towards ants in the turnip sawfly Athalia rosae. Entomol. Exp. Appl. 104, 153–157.
Nakamura, T., Kawai, Y., Kitamoto, N., Osawa, T., Kato, Y., 2009. Covalent
modification of lysine residues by allyl isothiocyanate in physiological
conditions: plausible transformation of isothiocyanate from thiol to amine.
Chem. Res. Toxicol. 22, 536–542.
Stauber, E.J., Kiczka, P., van Ohlen, M., Vogt, B., Janowitz, T., Piotrowski, M., Beuerle,
T., Wittstock, U., 2012. Turning the ‘mustard oil bomb’ into a ‘cyanide bomb’:
aromatic glucosinolate metabolism in a specialist insect herbivore. PLoS ONE 7,
e35545.
Van Loon, J.J.A., Schoonhoven, L.M., 1999. Specialist deterrent chemoreceptors
enable Pieris caterpillars to discriminate between chemically different
deterrents. Entomol. Exp. Appl. 91, 29–35.
Van Loon, J.J.A., Wang, C.Z., Nielsen, J.K., Gols, R., Qiu, Y.T., 2002. Flavonoids from
cabbage are feeding stimulants for diamondback moth larvae additional to
glucosinolates: chemoreception and behaviour. Entomol. Exp. Appl. 104, 27–34.
Visser, J.H., 1983. Differential sensory perception of plant compounds by insects.
ACS Symp. Ser. 208, 215–230.
Visser, J.H., Piron, P.G.M., 1995. Olfactory antennal responses to plant volatiles in
apterous virginoparae of the vetch aphid Megoura viciae. Entomol. Exp. Appl. 77,
37–46.
Waterman, C., Cheng, D.M., Rojas-Silva, P., Poulev, A., Dreifus, J., Lila, M.A., Raskin, I.,
2014. Stable, water-extractable isothiocyanates from Moringa oleifera leaves
attenuate inflammation in vitro. Phytochemistry 103, 114–122.
Winde, I., Wittstock, U., 2011. Insect-herbivore counteradaptations to the plant
glucosinolate-myrosinase system. Phytochemistry 72, 1566–1575.
Wittstock, U., Gershenzon, J., 2002. Constitutive plant toxins and their role in
defense against herbivores and pathogens. Curr. Opin. Plant Biol. 5, 1–8.
Wittstock, U., Agerbirk, N., Stauber, E.J., Olsen, C.E., Hippler, M., Mitchell-Olds, T.,
Gershenzon, J., Vogel, H., 2004. Successful herbivore attack due to metabolic
diversion of a plant chemical defense. Proc. Natl. Acad. Sci. U.S.A. 101, 4859–
4864.
Nottingham, S.F., Hardie, J., Dawson, G.W., Hick, A.J., Pickett, J.A., Wadhams, L.J.,
Woodcock, C.M., 1991. Behavioral and electrophysiological responses of aphids
to host and nonhost plant volatiles. J. Chem. Ecol. 17, 1231–1242.
Olsen, O., Sørensen, H., 1979. Isolation of glucosinolates and the identification of o-
(a-L-rhamnopyranosyloxy)benzylglucosinolate from Reseda odorata. Phytochemistry
18, 1547–1552.
Olsen, O., Rasmussen, K.W., Sørensen, H., 1981. Glucosinolates in Sesamoides
canescens and S. pygmaea: identification of 2-a-L-arabinopyranosyloxy)-2-
phenylethylglucosinolate. Phytochemistry 20, 1857–1861.
Opitz, S.E.W., Mix, A., Winde, I.B., Müller, C., 2011. Desulfation followed by
sulfation: metabolism of benzylglucosinolate in Athalia rosae (Hymenoptera:
Tenthrenidae). ChemBioChem 12, 1252–1257.
Zhou, D.S., Wang, C.Z., Van Loon, J.J.A., 2009. Chemosensory basis of behavioural
plasticity in response to deterrent plant chemicals in the larva of the small
cabbage white butterfly Pieris rapae. J. Insect Physiol. 55, 788–792.
Please cite this article in press as: Müller, C., et al. Taste detection of the non-volatile isothiocyanate moringin results in deterrence to glucosinolate-
adapted insect larvae. Phytochemistry (2015), http://dx.doi.org/10.1016/j.phytochem.2015.08.007