Content uploaded by Marcus A Koch
Author content
All content in this area was uploaded by Marcus A Koch on Mar 17, 2021
Content may be subject to copyright.
Phytochemistry 185 (2021) 112668
0031-9422/© 2021 Elsevier Ltd. All rights reserved.
Review
Comparison of glucosinolate diversity in the crucifer tribe Cardamineae and
the remaining order Brassicales highlights repetitive evolutionary loss and
gain of biosynthetic steps
Niels Agerbirk
a
,
*
, Cecilie Cetti Hansen
a
, Christiane Kiefer
b
, Thure P. Hauser
a
,
Marian Ørgaard
a
, Conny Bruun Asmussen Lange
a
, Don Cipollini
c
, Marcus A. Koch
b
a
Department of Plant and Environmental Sciences, University of Copenhagen, Thorvaldsensvej 40, 1871, Frederiksberg C, Denmark
b
Department of Biodiversity and Plant Systematics, Centre for Organismal Studies, Heidelberg University, 69120, Heidelberg, Germany
c
Department of Biological Sciences, Wright State University, 3640 Colonel Glenn Highway, Dayton, OH, 45435, USA
ARTICLE INFO
Keywords:
Brassicales
Malphigiales
Sapindales
Brassicaceae
Tribe Cardamineae
Armoracia
Barbarea
Cardamine
Nasturtium
Planodes
Rorippa
Tribe Brassiceae
Brassica
Sinapis
Erucastrum
Coincya
Hemicrambe
Tribe Arabideae
Arabis
Tribe Camelineae
Arabidopsis
Resedaceae
Reseda
Moringaceae
Euphorbiaceae
Putranjivaceae
Violaceae
Rutaceae
Reliable glucosinolate proles
Phylogeny
Biosynthesis
Evolution
ABSTRACT
We review glucosinolate (GSL) diversity and analyze phylogeny in the crucifer tribe Cardamineae as well as
selected species from Brassicaceae (tribe Brassiceae) and Resedaceae. Some GSLs occur widely, while there is a
scattered distribution of many less common GSLs, tentatively sorted into three classes: ancient, intermediate and
more recently evolved. The number of conclusively identied GSLs in the tribe (53 GSLs) constitute 60% of all
GSLs known with certainty from any plant (89 GSLs) and apparently unique GSLs in the tribe constitute 10 of
those GSLs conclusively identied (19%). Intraspecic, qualitative GSL polymorphism is known from at least
four species in the tribe. The most ancient GSL biosynthesis in Brassicales probably involved biosynthesis from
Phe, Val, Leu, Ile and possibly Trp, and hydroxylation at the β-position. From a broad comparison of families in
Brassicales and tribes in Brassicaceae, we estimate that a common ancestor of the tribe Cardamineae and the
family Brassicaceae exhibited GSL biosynthesis from Phe, Val, Ile, Leu, possibly Tyr, Trp and homoPhe (ancient
GSLs), as well as homologs of Met and possibly homoIle (intermediate age GSLs). From the comparison of
phylogeny and GSL diversity, we also suggest that hydroxylation and subsequent methylation of indole GSLs and
usual modications of Met-derived GSLs (formation of sulnyls, sulfonyls and alkenyls) occur due to conserved
biochemical mechanisms and was present in a common ancestor of the family.
Apparent loss of homologs of Met as biosynthetic precursors was deduced in the entire genus Barbarea and was
frequent in Cardamine (e.g. C. pratensis, C. diphylla, C. concatenata, possibly C. amara). The loss was often
associated with appearance of signicant levels of unique or rare GSLs as well as recapitulation of ancient types
of GSLs. Biosynthetic traits interpreted as de novo evolution included hydroxylation at rare positions, acylation
at the thioglucose and use of dihomoIle and possibly homoIle as biosynthetic precursors. Biochemical aspects of
the deduced evolution are discussed and testable hypotheses proposed. Biosyntheses from Val, Leu, Ile, Phe, Trp,
homoPhe and homologs of Met are increasingly well understood, while GSL biosynthesis from mono- and
dihomoIle is poorly understood. Overall, interpretation of known diversity suggests that evolution of GSL
biosynthesis often seems to recapitulate ancient biosynthesis. In contrast, unprecedented GSL biosynthetic
innovation seems to be rare.
* Corresponding author.
E-mail address: nia@plen.ku.dk (N. Agerbirk).
Contents lists available at ScienceDirect
Phytochemistry
journal homepage: www.elsevier.com/locate/phytochem
https://doi.org/10.1016/j.phytochem.2021.112668
Received 17 August 2020; Received in revised form 5 January 2021; Accepted 9 January 2021
Phytochemistry 185 (2021) 112668
2
1. Introduction
The glucosinolates (GSLs) constitute a biochemically well-dened
group of metabolites based on an amino acid-derived backbone fun-
tionalized into a thiohydroximic acid, which combines an SH and an
N–OH functional group. In GSLs, these groups are connected to a glucose
residue and a sulfate residue, respectively, in a uniform way that denes
this class of secondary metabolites (Blaˇ
zevi´
c et al., 2020). It is also this
structural element that is mainly involved in biochemical trans-
formations of GSLs (Wittstock et al., 2016). The most fundamental
transformation is initiated by hydrolysis of the thioglucosidic bond
(Fig. 1). A dedicated enzyme, myrosinase (thioglucoside glucohy-
drolase, E.C. 3.2.1.147) catalyzes the hydrolysis after tissue disruption.
Tissue disruption is generally a prerequisite, as the enzyme is usually
conned to other compartments than GSLs, but exceptions are known
(Sugiyama and Hirai, 2019). Hydrolysis by myrosinase results in an
unstable thiohydroximate-O-sulfate that may fragment into a nitrile or
rearrange into an isothiocyanate, in both cases spontaneously (Wittstock
et al., 2016; Blaˇ
zevi´
c et al., 2020). However, at physiological pH,
spontaneous nitrile formation is a minor reaction. If the resulting iso-
thiocyanate carries a hydroxy group at the β-position, spontaneous
ring-closure results in an oxazolidine-2-thione. Both isothiocyanates and
oxazolidine-2-thiones are defensive molecules in various ways,
depending on the amino-acid derived side chain that generally distin-
guishes individual GSLs from each other (Wittstock et al., 2016; Müller
et al., 2018). Since the most ancient GSLs include side chains both with
and without β-hydroxy groups (Section 3.1.) and specier proteins
responsible for additional products are so far only known from one
derived family (see below), the archetypic glucosinolate-myrosinase
system may have been an activated defense system yielding iso-
thiocyanates and oxazolidine-2-thiones upon tissue disruption by her-
bivores (Fig. 1).
GSLs are apparently ubiquitous in the order Brassicales, with a few
scattered occurences in other orders (Rodman et al., 1998). The taxon-
omy of GSL-containing plants discussed in this review is summarized in
Table 1. Non-Brassicales occurrences include the recently reported
Luvunga scandens (Rutaceae: Sapindales) (Sirinut et al., 2017) and
Rinorea subintergifolia (Violaceae: Malphigiales) (Montaut et al., 2017)
as well as the classically recognized Putranjiva roxburghii (Putranjiva-
ceae: Malphigiales) (Kjær and Friis, 1962). An apparent consensus of
GSLs in the related genus Drypetes (Euphorbiaceae: Malphigiales)
(Rodman et al., 1998) is difcult to trace back to an authoritative
investigation, but GSLs were recently reported from Drypetes euryodes
and Drypetes gossweileri (Montaut et al., 2017).
Our understanding of the evolution of this defense system within the
Brassicales has progressed immensely in recent years. The basis of all
evolutionary scenarios is robust phylogenies, which are available at the
family level (Edger et al., 2018) and for individual families such as the
Brassicaceae (Nikolov et al., 2019; Huang et al., 2020; Walden et al.,
2020). Phylogenies can be based on one sequence only, suct as the
popular nuclear ribosomal DNA internal transcribed spacer (ITS)
sequence (Poczai and Hyv¨
onen, 2010). Phylogenies based on many
genes are preferred (e.g. One Thousand Plant Trancriptomes Initiative,
2019). However, the reliability of bifurcating phylogenies is debatable
because of the frequent occurrence of hybridization (Novikova et al.,
2016), suggesting the need for cautious interpretation in evolutionary
studies.
Numerous protein factors in the derived family Brassicaceae have
also been characterized (Chhajed et al., 2019), and shown to modify the
biological activity of the glucosinolate-myrosinase system, e.g. by
forming other products types such as nitriles in abundance, epithioni-
triles and organic thiocyanates (Wittstock et al., 2016). Even transport
systems and regulatory mechanisms are beginning to be elucidated
(Augustine and Bisht, 2017; Jørgensen et al., 2017; Xu et al., 2017), and
“atypical” myrosinases with functions in intact plant cells are being
discovered and characterized (Sugiyama and Hirai, 2019). A compre-
hensive catalog of Arabidopsis thaliana genes related to GSL biosynthesis
and general metabolism was recently published (Harun et al., 2020).
Finally, we have come far in understanding metabolic adaptations in
herbivores that often interfere with the basic reactions of the
GSL-myrosinase system in intricate ways (Jeschke et al., 2016; Beran
et al., 2018; Friedrichs et al., 2020; Malka et al., 2020). Biochemical
adaptations of other antagonists are also known (Chen et al., 2020).
The biosynthesis of GSLs is increasingly well understood, mainly
from studies in A. thaliana (Sønderby et al., 2010; Chhajed et al., 2020)
and evolutionary scenarios have been proposed. In particular, the
biosynthesis has repeatedly been suggested to have evolved in an
otherwise cyanogenic plant species, based on arguments including a
shared committed step catalyzed by CYP79 enzymes in GSL biosynthesis
and cyanogenic glycoside biosynthesis and biochemically similar re-
actions as the second step but with different products (e.g. Bak et al.,
1998; Bak et al., 1999; Hansen et al., 2001; Halkier and Gershenzon,
2006). However, CYP79s have since been identied in most angiosperm
orders (Nelson and Werck-Reichhart, 2011) and all of them catalyze the
conversion of an amino acid to the corresponding oxime (Sørensen et al.,
Fig. 1. An archetypic glucosinolate-myrosinase system, leading to an isothio-
cyanate (A) and an oxazolidine-2-thione (B) from myrosinase-catalyzed hy-
drolysis of two ancient glucosinolates, 11 and 31. The hydrolysis reactions are
unbalanced; water is an additional reactant and glucose, sulfate and hydrogen
ion are also released during the myrosinase-catalyzed hydrolysis. A rear-
rangement precedes the formation of isothiocyanate (Blaˇ
zevi´
c et al., 2020).
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
3
2018). Thus, most angiosperms seem to have the genetic potential to
produce amino acid derived oximes. Whereas some plant species
metabolize the reactive oximes to e.g. GSLs or cyanogenic glycosides
and thereby more stable products, other plants release volatile oximes
(e.g. poplar; Irmisch et al., 2013). Hence, the CYP79s in the rst step of
GSL biosynthesis could have other origins than in cyanogenic glycoside
biosynthesis.
It has been suggested that a nitrile oxide or aci-nitro compound
produced by a CYP83 once appeared in a common ancestor of the
Brassicales (Hansen et al., 2001). In a study of CYP83 mechanisms,
potential scenarios were discussed (Clausen et al., 2015). One of these
reactive metabolites is envisioned as an intermediate in GSL biosyn-
thesis (Sønderby et al., 2010; Chhajed et al., 2020), and regardless of the
reason for its original formation, it can be imagined that GSL biosyn-
thesis evolved based on general detoxication reactions, since the
involvement of glutathione followed by glycosylation and sulfation in
GSL biosynthesis resembles detoxication reactions (Halkier and Ger-
shenzon, 2006) (Fig. 2). Indeed, some of the involved enzymes seem to
have homologs that are very widespread in higher plants (Cannell et al.,
2020).
No matter the GSL evolutionary origin, we simply assume that GSL
biosynthesis in the Brassicales has a single origin that is different from
the origins in Malphigiales and Sapindales. This hypothesis is in line
Table 1
Glucosinolate-containing plant species discussed in the text of this paper. The
orders are listed alphabetically, so are families in orders and species in families.
Scientic name [Common name]
Order Brassicales
Family Brassicaceae:
Arabidopsis thaliana (L.) Heynh. [mouse ear cress]
Arabis hirsuta (L.) Scop. [hairy rockcress]
Arabis soyeri Reut. & A. Huet
Armoracia rusticana P. Gaertn., B. Mey & Scherb [horseradish]
Barbarea australis Hook f. [riverbed wintercress]
Barbarea grayi Hewson [native wintercress]
Barbarea vulgaris W.T. Aiton [wintercress, yellow rocket]
Boechera stricta (Graham) Al-Shehbaz [Drummond’s rockcress]
Brassica juncea (L.) Czern. [brown mustard, Indian mustard]
Brassica napus L. [oilseed rape]
Cardamine amara L. [large bittercress]
Cardamine asarifolia L. [asarum-leaved bittercress]
Cardamine diphylla (Michx.) A.W. Wood [two-leaved toothwort]
Cardamine concatenata (Michx.) O.Schwarz (syn. Dentaria laciniata Muhl. ex Willd.)
[cut-leaved toothwort]
Cardamine cordifolia A. Grey [heartleaf bittercress]
Cardamine hirsuta L. [hairy bittercress]
Cardamine kitaibelii Bech. [Kitaibel’s bittercress]
Cardamine pratensis L. [cuckoo ower]
Coincya rupestris ssp. leptocarpa (Gonz.-Albo) Leadlay
Erucastrum canariense Webb and Berthel.
Erysimum cheiranthoides L. [treacle-mustard)
Hemicrambe fruticulosa Webb
Iberis amara L. [rocket candytuft]
Nasturtium ofcinale W.T. Aiton [watercress]
Planodes virginica (L.) Greene (synonyms: P. virginicum, preferred by
worldoraonline, and Sibara virginica
(L.) Rollins) [Virginia rockcress]
Raphanus sativus L. [radish]
Rorippa amphibia (L.) Besser [water cabbage]
Rorippa indica (L.) Hiern [Indian yellowcress]
Rorippa sarmentosa (G.Forst. ex DC) J.F.Macbr.
Rorippa sylvestris (L.) Besser [creeping yellowcress]
Sinapis alba L. [white mustard]
Sinapis arvensis L. [charlock]
Sinapis pubescens L. (syn. S. boivinii Baillargeon) [pubescent mustard]
Family Gyrostemonaceae:
Gyrostemon ramulosus Desf. [corkybark]
Family Moringaceae
Moringa oleifera Lam. [horseradish tree]
Family Resedaceae:
Reseda alba L. [white mignonette]
Reseda lutea L. [yellow mignonette, wild mignonette]
Reseda luteola L. [dyer’s rocket; weld]
Reseda odorata L. [garden mignonette, common mignonette]
Family Tropaeolaceae
Tropaeolum majus L. [garden nasturtium]
Order Malphigiales
Family Euphorbiaceae
Drypetes euryodes (Hiern) Hutch.
Drypetes gossweileri S. Moore
Family Putranjivaceae
Putranjiva roxburghii Wall.
Family Violaceae
Rinorea subintegrifolia (P. Beauv.) Kuntze
Order Sapindales
Family Rutaceae
Luvunga scandens (Roxb.) Buch.-Ham. ex Wight. & Arn. [lavang lata]
Nomenclature within Brassicaceae follows https://brassibase.cos.uni-heid
elberg.de, and the remaining follow www.worldoraonline.org and www.ipni.
org.
Fig. 2. The essence of glucosinolate (GSL) biosynthesis as imagined for the
presumably ancient 2-methylpropylGSL and a β-hydroxylated derivative, 2-hy-
droxy-2-methylpropylGSL. The three steps between the CYP83 product and
thiohydroximic acid in general GSL biosynthesis involves glutathione, serving
as the donor of sulfur. The illustrated hypothetic pathway is based on the
known biosynthetic pathway of more recently evolved GSLs (Sønderby
et al., 2010).
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
4
with phylogeny (Rodman et al., 1998; Edger et al., 2015, 2018) and
multiple evolutionary investigations of GSL biosynthetic genes (Bend-
eroth et al., 2006; Hofberger et al., 2013; Edger et al., 2015; van den
Bergh et al., 2016; Cang et al., 2018; Barco and Clay, 2019; Abrahams
et al., 2020; Czerniawski et al., 2021). Some details of GSL biosynthesis
will be reviewed (Sections 3.5. and 3.6.) in an attempt to interpret
evolutionary patterns in biosynthetic terms. The GSL resulting from the
core structure biosynthesis, starting from a specic amino acid, is
termed the parent GSL from that amino acid (e.g. 2-methylpropylGSL
from Leu) and any additional reactions modifying a parent GSL are
termed (secondary) modications (Fig. 2).
Our focus in the complex evolutionary history of the GSL-myrosinase
system is the evolutionary history of the current structural GSL diversity.
The scientically demonstrated structural diversity was recently
reviewed, and it was concluded that 88 GSLs were known from nature
with high certainty and supported by solid spectroscopic data (Blaˇ
zevi´
c
et al., 2020). Since then, one more GSL has been conclusively reported,
making the number 89 (Montaut et al., 2020a). An additional 49 sug-
gested GSL structures were registered, most of which were suggested
decades ago with incomplete evidence (Blaˇ
zevi´
c et al., 2020). The evi-
dence backing up those GSLs range from faint (e.g. phenylGSL) to
considerable (e.g. 3-carboxypropylGSL from homoGlu), but it was
concluded that they all needed conrmation using modern methods.
Numerous additional GSLs have been suggested in recent years, based
on incomplete evidence and often quite supercial arguments, typically
as suggested identities of HPLC-MS peaks. It was concluded that this
myriad of recently suggested GSLs needed additional verication before
registration and discussion was meaningful. Indeed, the most recently
reported GSL involved revision of a previously suggested structure that
had to be abandoned (Montaut et al., 2020a). Although registration of
still un-identied GSLs can in principle be carried out in evolutionary
studies, any structural discussion is by denition impossible (unless
those GSLs are close to elucidated, such as GSLs numbered x1 and x2 for
which structures are NMR-elucidated except for the position of a methyl
group) (Fig. 3). However, it seems likely that numerous GSLs, possibly
hundreds, exist in plants in addition to the currently recognized number
of somewhere between 89 and 138. Still, the moderate numbers of
known and anticipated GSLs contrasts with the more than 8000 known
avonoids (Alseekh et al., 2020) and 14,000 known saponins (C´
ardenas
et al., 2020), making the GSLs an attractive group for evolutionary
studies.
The generally recognized GSLs are either known or anticipated to be
derived from the following standard amino acids, listed in approxi-
mately increasing complexity using standard three-letter abbreviations:
Ala, Val, Leu, Ile, Glu, Phe, Tyr, Met and Trp. Six of them are aliphatic, so
the corresponding GSLs are sometimes discussed collectively as aliphatic
GSLs (Fig. 4). GSLs from Phe and Tyr can be collectively termed ben-
zenic GSL. The former two terms are also used for GSLs derived from
chain elongated amino acids. The GSLs from Trp are also termed indole
GSLs Collectively, benzenic and indole GSLs can be termed aromatic GSL
(Fig. 2). That name also logically includes aliphatic GSLs esteried to
aromatic carboxylic acids (Blaˇ
zevi´
c et al., 2020) and is not informative
in an evolutionary context. The present authors suggest
GSL-classication by individual precursor amino acid in an evolutio-
nary/biosynthesis context (Figs. 3 and 4). Early investigators investi-
gated GSL biosynthesis in A. thaliana from n-homoMet, Phe and Trp, and
found distinct committed steps for each (Sønderby et al., 2010). GSL
biosynthesis is sometimes still presented as 2–3 parallel biosyntheses,
one departing from aliphatic amino acids or just Met and one from ar-
omatic amino acids or just Trp (Harun et al., 2020). This simplied view
seems to be based on the early focus on just a few GSLs in A. thaliana and
cultivated Brassica species. Considering that a gene for chain elongation
of aliphatic Met seems to also confer chain elongation of aromatic Phe
(Section 3.5.), and that a well-established CYP79 enzyme with aliphatic
n-homoMet as substrate seems to also accept the aromatic homoPhe in
vivo while Phe is metabolized by a separate CYP79 (Section 3.5.), the
present review will avoid discussing GSL biosynthesis in the apparently
too simplied groupings of aliphatic versus aromatic amino acids and
GSLs.
In most cases, chain elongation to form higher homologs can be
involved in GSL biosynthesis, as demonstrated for Met, Phe, Glu, and
anticipated for Ile and possibly Leu, Ala and Tyr (Blaˇ
zevi´
c et al., 2020).
Met is (almost) always subject to chain elongation before GSL biosyn-
thesis. Trp is never subject to chain elongation in GSL biosynthesis but
can be subject to chain-shortening before GSL biosynthesis as recently
discovered in Brassica napus (Pedras and Yaya, 2013; Pedras et al.,
2016). As the precursor amino acid has profound inuence on the
properties of the resulting GSL, an understanding of the evolutionary
history of amino acid use for GSL biosynthesis and chain elongation of
the various amino acids is warranted. Intermediate to late timing of two
key events, use of Trp and Met as precursors of GSLs, have been deduced
in the family-level phylogenetic tree (Edger et al., 2018), although the
deductions are not entirely in agreement with the phytochemical liter-
ature (Section 3.1.). As will be discussed in that section, use of several
amino acid precursors seem to predate the extensive use of Met in some
derived families today, such as Val, Leu and Ile without chain elongation
and Phe (and possibly Tyr) both with and without chain elongation.
Several reports have considered the associated gene sequence evolution
in subsets of the order (e.g. Benderoth et al., 2006; Hofberger et al.,
2013; Edger et al., 2015; van den Bergh et al., 2016; Barco and Clay,
2019; Abrahams et al., 2020).
The precursor amino acid, any chain elongation and various modi-
fying reactions can be deduced from GSL structure, so tracing evolu-
tionary events from GSL proles of present-day plants should be
possible, and this is a key ambition of the present paper. However, we do
acknowledge the need for taking into account also transcriptomic and
genomic data, and functional characterization of biosynthetic enzymes
(Sections 3.5. and 3.6.). We have elsewhere discussed the various ob-
stacles to overcome for reliable identication of GSL proles (Olsen
et al., 2016; Blaˇ
zevi´
c et al., 2020; Agerbirk et al., 2021). In essence, the
key to reliable GSL identication is NMR elucidation of structures or use
of authentic references in combination with highly discriminating
analytical methods such as HPLC-MS/MS or GC-MS. A major problem in
the current literature is general reliance on MS information only,
probably due to frequent unavailability of authentic references. A
detailed analysis of the literature (Agerbirk et al., 2021) identied only
ve major comparisons of reliably determined GSL proles in a phylo-
genetic context (Windsor et al., 2005; Agerbirk et al., 2008; Olsen et al.,
2016; Blaˇ
zevi´
c et al., 2017; Czerniawski et al., 2021).
We have concentrated on members of the tribe Cardamineae as an
evolutionary study system because the tribe is of general ecological and
even agronomical interest, as exemplied elsewhere (Olsen et al., 2016;
Agerbirk et al., 2021). Particular attention has been paid to the genus
Barbarea and the highly variable species Cardamine pratensis, due to
their atypical GSL proles. C. pratensis forms a polyploid species com-
plex (Melich´
arkov´
a et al., 2020), and this species as well as a group of
related species is referred to as the C. pratensis alliance in the rest of this
paper.
1.1. Comprehensive numbers and abbreviations for glucosinolates and
precursor amino acids
The Brown-Windsor system of GSL abbreviations (Brown et al., 2003;
Windsor et al., 2005) has been further adapted in order to cover all
relevant structures, using a system as close to general organic chemistry
nomenclature as possible but abbreviated and simplied (Agerbirk et al.,
2021). Most abbreviations (Figs. 3 and 4) should be intuitively recog-
nizable by users of the Brown-Windsor abbreviations, but two differ
signicantly and had to be derived from trivial names: BAR for gluco-
barbarin, (S)-2-hydroxy-2-phenylethylGSL, and EBAR for epi-
glucobarbarin, (R)-2-hydroxy-2-phenylethylGSL. The corresponding
Brown-Windsor abbreviation for BAR, (S)2OH-2PE, was too unhandy
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
5
Fig. 3. All glucosinolates (GSLs) derived from aromatic amino acids known from the tribe Cardamineae. The constant part of the GSLs is abbreviated GSL in most
structures and exemplied in case of 11, a similar system is used for 6
′-isoferuloylated GSLs as exemplied for 129. Abbreviations of individual GSLs follow a
comprehensive system systematically explained in an accompanying paper (Agerbirk et al., 2021); spaces have occasionally been inserted in some long names and
abbreviations for easier reading. The semisystematic name of “glucobarbarin” is (S)-2-hydroxy-2-phenylethylGSL, and for “epiglucobarbarin” it is
(R)-2-hydroxy-2-phenylethylGSL.
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
6
for the many further substituted structures that occur in tribe Carda-
mineae, such as the known para-hydroxy 6′-isoferuloyl derivative of
EBAR (Fig. 3).
Pioneers in cataloging botanical GSL diversity used a comprehensive
numbering system (Fahey et al., 2001, Agerbirk and Olsen, 2012;
Blaˇ
zevi´
c et al., 2020). This system assigns a xed number to each
scientically accepted GSL and is a key to major tabulated compilations
of botanical diversity (Fahey et al., 2001; Olsen et al., 2016). An example
is 62 for 2-methylpropylGSL (2mPr). We use this numbering system in
tables, gures and occasionally in text to simplify comparison with the
remaining literature. For indicating desulfated GSLs, which are both
relevant as biosynthetic precursors and as derivatives in analytical
chemistry, we precede the GSL number with a “d”, e.g. d62 is desulfo 62
(desulfo 2mPr). Numbers shown in brackets, e.g. [104], indicate
tentatively suggested GSLs whose existence in nature has not yet been
conrmed by NMR-spectroscopy (Blaˇ
zevi´
c et al., 2020). In this case
hypothetic 4-phenylbutylGSL (Fig. 3) that would correspond to a
recently reported minor isothiocyanate in horseradish, Armoracia
Fig. 4. All glucosinolates (GSLs) derived from aliphatic amino acids known from the tribe Cardamineae. The constant part of the GSLs is abbreviated GSL in most
structures and exemplied in case of 107. Abbreviations of individual GSLs follow a comprehensive system explained in the text (Section 1.1.); spaces have oc-
casionally been inserted in some long names and abbreviations for easier reading. BCAA; branched chain amino acid.
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
7
rusticana (Deki´
c et al., 2017).
For amino acids, standard three-letter abreviations are used, and for
homologs we add the prex ‘homo’, e.g. homoIle for homoisoleucine.
For higher homologs, we use Arabic numbers rather than Greek prexes
for brevity, e.g. 2homoIle for dihomoisoleucine, and for indicating an
unspecied higher homolog or a range of chain lengths, we replace the
number with ‘n’ or the relevant numeral range, e.g. 4-6homoMet for
tetra-to hexahomomethionine. Branched-chain amino acids (BCAA)
comprise Val, Leu, Ile and their homologs (Fig. 4).
1.2. Scope and aims
The general scope of this review is to characterize GSL evolution in
structural detail, such that suggested but structurally undened “GSL-
like” metabolites could not be included. In accordance with a recent
revision of accepted GSL structures (Blaˇ
zevi´
c et al., 2020), we limit our
discussion to the range of structures accepted or partially accepted
(“bracketed”) by that revision, and hence ignore various putative GSLs
with little to no scientic support. However, in the Supplementary Table
S1 listing reported GSLs by species, reports of some additional putative
GSLs are included.
The review-work presented here had three aims. The rst aim was to
obtain an updated molecular phylogeny of the tribe Cardamineae, based
on ITS-sequences from GenBank. The second aim was to critically review
the knowledge of GSL diversity in the tribe by including attention to
analytical quality in the reviewing process. For a comparative perspec-
tive, GSL diversity in a subset of the remaining Brassicaceae and a few
Resedaceae were reviewed. The third aim was to propose testable hy-
potheses about evolutionary processes, based on additional review of
the literature on GSL biosynthetic enzymes and genes.
2. Results
2.1. Updated phylogeny of the tribe Cardamineae
A molecular phylogeny based on ITS sequences was obtained for 171
taxa out of the total 386 species (44%) in the tribe (Koch et al., 2017).
The phylogeny (Supplementary Fig. S1) generally agrees with an earlier
phylogeny based on fewer species (Olsen et al., 2016). The genus
Aplanodes is resolved as sister to all other species of the tribe included in
the phylogeny. However, there is either low or no bootstrap support for
the backbone of the phylogeny, which might indicate rapid radiation,
conicting phylogenetic signal due to reticulate evolution (high number
of polyploids may be seen as an indicator) or a lack of sequence infor-
mation due to a limited number of single nucleotide polymorphisms
(SNPs) analyzed. We tentatively conclude that the genus Cardamine in
its current taxonomic circumscription is not monophyletic and that most
other genera or larger alliances are resolved as monophyletic with high
bootstrap support. These genera or larger clades are Leav-
enworthia/Selenia (87% bootstrap support), the C. pratensis alliance
(97% bootstrap support), the genus Barbarea (100% support) including
two chemotypes of B. vulgaris (Agerbirk et al., 2021), and Nasturtium
(100% support). The C. pratensis alliance and the genus Barbarea are a
focus of this work because of their deviating GSL proles, but species in
less well-resolved and less well-supported positions in the tree are also
considered. Non-Cardamine genera are indicated with grey boxes, and
species discussed in some detail are indicated in bold and are marked
with grey boxes, too (Supplementary Fig. S1).
2.2. Glucosinolate diversity in the tribe Cardamineae
2.2.1. Critical evaluation of literature
We previously reviewed GSL proles of all members of the tribe
Cardamineae (Olsen et al., 2016). Although that review provided a
detailed overview of the known diversity, several genera and multiple
species were not suitably investigated. In an accompanying paper
(Agerbirk et al., 2021), we tested GSL occurrences considered ques-
tionable in the previous review, and specically tested several cases of
“not analyzed” for cases where presence or absence of a particular
biosynthetic group had never been reliably tested for an entire genus. In
the accompanying paper, we provide evidence for Trp-derived GSLs and
various biosynthetic groups in Nasturtium, Rorippa and Planodes,
conclusive evidence for several questioned GSLs from Armoracia, lack of
thioGlc-acylated GSLs in the tribe apart from Barbarea, lack of aliphatic
GSLs in two types of B. vulgaris, and lack of Phe-derived GSLs (not chain
elongated) in Nasturtium. Including the accompanying paper, we
brought our previous literature review up to date. For this purpose, we
inspected relevant papers (until around 2019) that were published since
the former review (and a single paper from 1983 that had been over-
looked), and compiled GSL diversity by species (Louda and Rodman,
1983; Giallourou et al., 2016; Pellissier et al., 2016; Deki´
c et al., 2017;
Ciska et al., 2017; Jeon et al., 2017; Bakhtiari et al., 2018, 2019; Liang
et al., 2018; Cuong et al., 2019; Okamura et al., 2019; Badenes-P´
erez
et al., 2020; Montaut et al., 2020b). A recent claim without evidense of
identication (Hussain et al., 2020), concerning proposed aliphatic GSLs
in B. vulgaris, was ignored for reasons detailed in the accompanying
paper (Agerbirk et al., 2021). We also re-checked and occasionally
revised our scoring of GSLs in older references as either conclusive,
tentative or not sufciently evidenced (Supplementary Table S1A).
In total, the newly included papers concerned horseradish
(A. rusticana), multiple Cardamine spp., watercress (N. ofcinale), Ror-
ippa sarmentosa and an erratum on Rorippa austriaca. Poorly evidenced
suggestions of GSLs never convincingly characterized from any plant
were ignored, relying on the recent compilation by Blaˇ
zevi´
c et al.
(2020). As done previously, we judged the reports based on the criterion
that conclusive compound identication should be based on use of
reference compounds, validated reference materials or NMR structure
elucidation. We also required suitably specic peak detection by means
of an MS detector (or a diode array detector in case of some Trp-derived
GSLs), and when relevant one additional spectral characteristic, either
complete MS (typically MS2) or a characteristic UV spectrum (in case of
Trp-derived GSLs only). In the case of indirect identication by GC-MS
of GSL products, we also regarded identication by reliable retention
index (RI) compared to databases as proof of identity because of the
general power of GC-MS as an identication tool when using both RI and
MS spectral comparison. Reports that did not meet the criteria were
registered as tentative by putting the relevant GSLs in parentheses in the
table. For our own recent results (Agerbirk et al., 2021), registration as
tentative typically resulted from lack of an authentic standard but
reasonable t
R
compared to homologs, or agreement of t
R
and m/z with an
authentic standard but with too low levels to check MS2. However,
criteria for registration as tentative or conclusive could not always be as
strict as for our own results from inspection of the general literature,
because of lack of details in reports. Detailed comments for individual
reports are given in Supplementary Table S1, which also contains a
supplement with examples of our arguments.
In the majority of recent papers, comparison with previous studies of
the same species were notsystematic or frequently missing completely,
possibly due to the difculty in locating previous studies in standard
literature search (Section 4.2.). This was the case for 3moBZ suggested
from C. cordifolia with faint evidence (Humphrey et al., 2016), without
discussion of a previous suggestion with moderate evidence of an isomer
(BAR or EBAR) in the same species (Louda and Rodman, 1983). Neither
suggestion was accepted. Similarly, the analytical basis of surprising
structures was typically not elaborated on or discussed, making it
difcult to judge whether several surprising structures should be
accepted as existing in the tribe. This was the case for, e.g., unsaturated
Met-derived GSLs such as 4mSbuen and 4mSObuen reported by two
inconclusive studies (Pellissier et al., 2016; Ciska et al., 2017) and
suggested 5BzOp and S2hPeen (Ciska et al., 2017; Bakhtiari et al., 2018).
Furthermore, 2-3homoMet and 5homoMet derived GSLs were reported
from Cardamine amara by Pellissier et al. (2016) based on HPLC-UV
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
8
Table 2
Diversity of well-established glucosinolates (GSLs) in the tribe Cardamineae, ordered by genus and by amino acid precursor (except the category of glucose 6′-acylated (“6′-Acyl”) that is collected for all precursors).
Well-identied (and suggested) glucosinolates by apparent amino acid precursor or substitution (and alkyl chain length for n-homoMet-derived GSLs).
Genus
(inv./acc.)
a
1-3HomoMet
(3–5C)
4-6HomoMet
(6–8C)
7-8HomoMet
(9–10C)
Trp Phe HomoPhe
/2–4homoPhe
Val Leu Ile 1-2HomoIle 6′-
Acyl
Armoracia
b
(1/3 =33%)
12, 73, 94, 95, 101,
107 (82, 84)
66, 67, 69, 88
(87)
n.d. 28, 43, 48 11 105
2-4homoPhe: [104], [106],
[156]
(56) 62 61 (54) NA
Barbarea
c
(9/29 =31%)
n.d. n.d. n.d. 28, 43, 47, 48,
138
n.d. 40S, 40R, 105, 139S,
139R, 140, 142R, 148
n.d. n.d. n.d. n.d. 129,
130,
131S,
131R,
132R
Cardamine
d
(22/239 =9%)
12, 24R, 64, 72, 84, 94,
101 (24S, 38S, 107)
66, 69, 92
(67, 87)
n.d. 28, 43, 47, 48, 138 11, 23, 46 105 56, 57 31, 62 30, 61 29, 54, 58, 141, 149 n.d.
Leavenworthia
(2/8 =25%)
101 n.d. n.d. NA n.d. 105 57 n.d. n.d. – NA
Nasturtium
e
(2/5 =40%)
12, 72 (24R, 64,
73, 84, 101)
66, 67, 69, 80, 87,
92 (88)
n.d. 28, 43, 48 n.d. 40S, 40R, 105 n.d. n.d. n.d. n.d. n.d.
Planodes
f
(1/2 =50%)
64, 72 (94) 66, 67, 69, 87,
92 (80, 88)
(68, [89]) 43 n.d. 40S, 40R, 105, 139R n.d. n.d. n.d. n.d. n.d.
Rorippa
(9/87 =10%)
12, 24R, 64,
72, 101, 107
66, 67, 69, 80
87, 92
68, 77, 79, [89] (65) 28, 43, 47, 48, 138 n.d. 105 n.d. 62 61 n.d. n.d.
Selenia
(1/5 =20%)
n.d. n.d. n.d. – 11 40, 105 n.d. n.d. n.d. – NA
For space-considerations, GSLs are indicated by numbers only. GSLs in parentheses, (), are tentatively known in the indicated genus. GSLs in square brackets, [], are of generally uncertain status in nature as such, and are
not in parentheses as well if the evidence in this genus is as high as in any other genus. For the following genera (# of accepted species) not mentioned in the table, no data were found in the literature search: Andrzeiowskia
(1), Aplanodes (2), Bengt-jonsellia (2), Iodanthus (1), Ornithocarpa (2), Pteroneurum (0) and Sisymbrella (2). 6′-Acyl: Glucosinolates with any kind of 6′-acylation. Those listed all had an isoferuloyl moeity and represented
GSLs derived from Trp or homoPhe. n.d., searched for but not detected as judged by an overall evaluation of the analytical conditions employed in representative species and relevant organs (seeds and calibrated HPLC-MS
in case of the category 6′Acyl). NA, not analyzed using suitable methodology. A dash (-) in an otherwise empty eld indicates a situation intermediate between NA and n.d.: no reliable reports for presence or absence were
found, but NA cannot be concluded either, as applied methods or organs used for analysis might or might not have allowed detection of the group of GSLs in question.
a
Number of species investigated and reviewed here/Number of accepted species for the genus according to Brassibase.
b
A number of additional putative GSLs have been suggested based on tentative interpretation of trace GC-MS peaks in A. rusticana. These controversial suggestions do not t into the established biosynthetic framework
and are in other aspects so uncertain that they have been left out from Fig. 1 and from the structures considered for review here. They include straight chain aliphatics ([13], [102]), suggested GSLs from chain elongated
Leu ([52], [55]), long chain unsaturated and various hydroxy derivatives ([25], [37]), and discontinued “19”, “39”, “41”, “86” and never included in numbering system, hypothetic “3-hydroxy-pent-4-enyl GSL” and
“iso28”. Similarly, tentative suggestions of HSbu (133) and either BAR or EBAR (40) have been challenged here and elsewhere and is omitted. Finally, suggested 4mSObuen (63) only based on HPLC-UV t
R
in one report is
ignored.
c
A controversial report of aliphatic GSLs (95, 56, 107) in Barbarea (Cole, 1976) has been challenged convincingly (Windsor et al., 2005; Supplementary Table S1A; Agerbirk et al., 2021) and is omitted here.
d
Suggested 4mSbuen (83) solely based on t
R
in HPLC-UV, and suggested and not reliably identied BAR/EBAR (40) and 3moBZ (45) and
ω
-hydroxyalkyls and numerous rare or never characterized putative structures,
all suggested by only one source each, have been omitted due to insufciency of the evidence.
e
Tentative suggestions of Phe derived BZ (11) and 4hBZ (23) in N. ofcinale has here and elsewhere been challenged convincingly (Supplementary Table S1A) and is omitted here.
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
9
only, without discussion of a previous report using HPLC-MS (Windsor
et al., 2005) that reported absence of n-homoMet-derived GSLs in this
species. Lack of discussion was even the case for an unexpected GSL
structure (“veratryl GSL” =“phenylvinyl GSL”) (Bakhtiari et al., 2018)
that had never been suggested from nature before and might be an
artifact (a loss of water MS-fragment of BAR or EBAR). Similarly, many
incomprehensive names had to be ignored (“trimethoxy GSL”,
“hydroxybenzyl-methylether GSL”) as well as some with imprecise
names (“hydroxypropyl GSL”, “butyl GSL”, “hydroxymethylbutyl GSL?”,
“2-hydroxymethylpropyl GSL”) (Bakhtiari et al., 2018), the latter
apparently supposed to be a known GSL although no known GSL could
be named in this way. Several additional reports needed discussion
before GSL proles could be listed as either conclusive or tentative
(Supplementary Table S1A). This general tendency for lack of precision,
lack of comparison with the previous literature and lack of discussion of
analytical uncertainty is in contrast to solid, classic work such as
Yamane et al. (1992) using NMR for conclusive peak identication in
ecological work, and work by HPLC-MS pioneers (Grifths et al., 2001),
who explicitly concluded that suggested structures based on
HPLC-MS/MS without authentic standards are tentative. Despite the
mentioned uncertainty and need for critical evaluation, almost all
considered papers, including those criticized for details above, mainly t
into the general picture for the relevant genus or species.
All in all, by systematically applying strict analytical chemistry
principles as well as conservative judgement in cases of unclear reports,
we classied each reported GSL as either conclusively identied,
tentatively identied (in parentheses) or insufciently evidenced
(labelled in red print) (Supplementary Table S1A).
2.2.2. Results of the literature survey
The GSL diversity by genus is summarized in Table 2, and by species
in Supplementary Table S1A. The majority of species in the tribe Car-
damineae are not yet investigated for GSLs, and investigation of many
species is only supercial. Apart from some small genera never inves-
tigated, the large genera Cardamine and Rorippa are relatively less well
investigated (9% and 10%, respectively, of all accepted species and with
a tendency for inconclusive methods applied). The two crops in the
tribe, N. ofcinale and A. rusticana have frequently been investigated,
but thorough investigations are few and much uncertainty remains on
the complex prole of A. rusticana, while the prole of N. ofcinale is
well established. The chemical defense-model genus Barbarea is very
well investigated and the analytical quality is high. Few studies
considered multiple organs, but it was apparent that this practice
increased the number of detected GSLs and the chance of obtaining
conclusive identication in at least one organ (e.g. Agerbirk et al., 2001;
van Leur et al., 2006; Agerbirk et al., 2010a; Agerbirk and Olsen, 2011;
Agneta et al., 2014; Agerbirk et al., 2015; Deki´
c et al., 2017; Agerbirk
et al., 2021).
For well-established GSLs known from nature with high certainty, 53
are conclusively identied in the tribe Cardamineae, which is 60% of the
89 known in all plants. The GSLs in the tribe can formally be derived
from 1-8homoMet (21), homoPhe (14), Trp (6), 1-2homoIle (5), Phe (or
Tyr) (3), Val (2), Ile (2) and Leu (2) and show a wide range of side-chain
modications (Figs. 3 and 4). An additional three well-established GSLs
from homologs of Met have been tentatively identied in the tribe
(Fig. 4). Of less well-established GSLs, recognized as tentative by
Blaˇ
zevi´
c et al. (2020), four are known from the tribe with as high cer-
tainty as in any plant, and three of them are so far only known from the
tribe: those apparently derived from 2-4homoPhe in horseradish
(Fig. 3).
Of the 53 well-established GSLs in the tribe, ten are unique in the
plant kingdom (19%) to our knowledge. Of these, ve are a unique type
of acyl derivative of rare or generally known GSLs (Fig. 3), three are
homoPhe-derived with unique hydroxylation patterns (3hPE, 4hBAR
and 3hEBAR) (Fig. 3), and two are 2homoIle-derived with unique hy-
droxylation pattern (3hmPe and 2h3mPe) (Fig. 4). These unique GSLs
were from the genus Barbarea (8 GSLs) and C. pratensis (2 GSLs). This is a
high proportion considering that the number of species in the tribe is
small (ca. 7%) compared to the order Brassicales, but this may be the
result of bias from quite intense investigation of this tribe in recent
decades. If we tentatively include so far unaccepted but suggested GSLs,
the proportion of unique GSLs would probably still be high, as a large
number of potentially novel GSLs have been reported from the tribe (e.g.
Grob and Matile, 1980; Lin et al., 2014; Bianco et al., 2014; Olsen et al.,
2016; Deki´
c et al., 2017).
While the tribe Cardamineae is (by denition) monophyletic, the
genus Cardamine most likely is not (Section 2.1.) and can metaphorically
be described as what remains in the tribe after assigning separate genus
names to a number of well-separated, monophyletic groups within the
tribe (Supplementary Fig. S1). The tribe Cardamineae is among those
tribes with signicant species diversication rate shifts resulting in the
high number of species (Huang et al., 2020). Furthermore, 63% of the
species are polyploids (Hohmann et al., 2015). The number of accepted
species names is about 386, and the crown group age and start of
diversication of the tribe is estimated to be approximately 12.6 million
years ago (Walden et al., 2020). In principle, it would be relevant to
subdivide the genus Cardamine in several clades for the tribal overview
(Table 2). However, of the 20 Cardamine species identied with any GSL
prole data at all, ten were only investigated using HPLC-UV of insuf-
cient reliability and a further three were not in the phylogeny (Sup-
plementary Table S1). Of the remaining seven species, three were found
in poorly resolved positions of the phylogeny. Another published tribal
phylogeny mainly based on different marker genes from the plastid
genome (Pellissier et al., 2016) did not show any signicant resolution
compared to other phylogenies (Olsen et al., 2016; Supplementary Fig.
S1), and the phylogeny presented deviated in some respects, in partic-
ular concerning the position of C. pratensis. In other respects, all phy-
logenies agree, including a well-supported clade containing both
C. concatenata and C. diphylla. Considering the problems with data
quality and phylogenetic resolution, we abstain from GSL diversity ta-
bles for subsections of the genus Cardamine. However, a few individual
species are discussed, including C. hirsuta, C. pratensis and C. impatiens,
that according to all available phylogenies are as distantly related as the
monophyletic genera in the tribe, as well as the somewhat related
C. concatenata and C. diphylla. (Sections 3.1., 3.2.).
Of the characteristic monophyletic genera in this rapidly expanding
tribe, only Barbarea, Nasturtium and Rorippa are well-investigated for
GSLs. While the GSL-pool in Nasturtium is mainly similar to the pool of
Cardamine, the pool of GSLs in Rorippa is somewhat deviating, with a
tendency for additional chain elongation and complete S-oxidation of
polyhomoMet derivatives, resulting in
ω
-(methylsulfonyl)alkyl GSLs.
The GSL-pool of Barbarea deviates the most from other tribe members by
the lack of GSL derived from aliphatic amino acids and non-elongated
Phe/Tyr, by the dominance of homoPhe-derived GSLs and the high
number of derivatives of those, by the occurrence of 6′-acylated GSLs
and by the presence of the disubstituted Trp-derived 1,4moIM (138).
However, 138 in C. pratensis and Rorippa amphibia and 139R in Planodes
virginica constitute a structural connection of Barbarea to the remaining
tribe. In more general terms, GSLs derived from homoPhe and Trp are
commonplace in the tribe and the entire family. The pool of GSLs in
Selenia may be equally deviating, but the single study so far is insuf-
cient for discussion. A group of Cardamine spp. commonly named
“C. pratensis and its allies” appears as well-separated as some of the
accepted genera. Based on detailed investigation of C. pratensis itself, the
clade seems to have a highly distinct GSL pool, lacking n-homoMet-
derived and with a broad diversity of GSLs derived from branched chain
amino acids, including chain elongated homologs of Ile (Olsen et al.,
2016). The only other species from this group that has been analyzed for
GSLs appears to be Cardamine rivularis (Pellissier et al., 2016), but the
analysis was inconclusive due to use of HPLC-UV only.
The genus Barbarea seems well-investigated in most of its European
range, but specic geographic areas warranting further research have
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
10
been pointed out (including the Caucasus) (Christensen et al., 2014;
Agerbirk et al., 2015). The “C. pratensis and its allies” clade warrants
additional attention using conclusive analytical methods. A recent study
screening several species with uniform methods generally conrmed
C. pratensis and Barbarea as deviating, but discovered the mentioned
similarities to other tribe members (Agerbirk et al., 2021).
Intraspecic qualitative polymorphism may be important for evo-
lution of diversity and is known for four species in the tribe: B. vulgaris,
C. pratensis, C. hirsuta and N. ofcinale. Although we would expect all
four polymorphisms to be genetically based, this is formally only proven
for B. vulgaris. The polymorphism of N. ofcinale may be due to plant
breeding possibly involving intraspecic crosses, as the polymorphism
was most pronounced for cultivated material and was accompanied with
chromosome numbers not usually found for the species (Agerbirk et al.,
2014). The several polymorphisms in B. vulgaris (van Leur et al., 2006;
Agerbirk et al., 2015) are increasingly well-characterized genetically
(van Leur et al., 2006; Byrne et al., 2017; Liu et al., 2019) and involve
multiple genes and loci. The polymorphism in C. pratensis has been little
investigated but can be expected to involve a number of genes due to the
complex chemical proles (Agerbirk et al., 2010a), and the poly-
morphism in C. hirsuta is just recently reported (Agerbirk et al., 2021).
2.3. Glucosinolates in other Brassicaceae species and Reseda
(Resedaceae) representing basal families
For comparative reasons, GSL proles of some non-Cardamineae
members of the Brassicaceae family were needed. For this reason, we
compiled GSL proles for A. thaliana, Arabis hirsuta, Arabis soyeri, Sinapis
alba, Sinapis arvensis, Brassica nigra, Brassica rapa, Brassica oleraceae,
Hemicrambe fruticulosa, Erucastrum canariense, Coincya rupestris ssp.
leptocarpa and Sinapis pubescens (syn. S. boivinii). These reference species
were selected because of the availability of high quality reports for each
of them, and because of the known presence of GSLs resembling or
contrasting with those occurring in tribe Cardamineae. An exhaustive
literature survey was not attempted, but several high-quality in-
vestigations of each species were preferentially used, although not
possible for the latter four species (Kjær and Gmelin, 1958; Kjær and
Schuster, 1972; Sang et al., 1984; Daxenbichler et al., 1991; Wittstock
and Halkier, 2000; Agerbirk et al., 2001, 2008, 2010b; Reichelt et al.,
2002; Bennett et al., 2004a; Padilla et al., 2007; Lee et al., 2013; Baek
et al., 2016; Pfalz et al., 2016; Hanschen and Schreiner, 2017; Klopsch
et al., 2017; Andini et al., 2019; Parpazian et al., 2019). The analysis is
presented in a similar way as the analysis of tribe Cardamineae members
(Supplementary Table S1B). Briey, the analyzed literature revealed
that the selected species exhibited a variety of features resembling those
seen in tribe Cardamineae (Fig. 5C). Met-derived GSLs were conclusively
detected in all of these. Two species had conclusively been shown to
contain Ile-derived 1mPr (61) (B. rapa and C. rupestris) and two other
species 2homoIle-derived 3mPe (58) (E. canariense and H. fruticulosa).
To provide a perspective on the diversity in the family Brassicaceae,
Reseda luteola and Reseda odorata (family: Resedaceae) were investi-
gated because the literature suggested a remarkable resemblance of
their GSL proles to that of B. vulgaris (Supplementary Table S1B). The
presence of BAR in R. luteola is well-established (Kjær and Gmelin, 1958;
Bennett et al., 2004a), despite the great evolutionary distance to Bar-
barea. Two independent reports of IM in the species (Bennett et al.,
2004a; Agerbirk et al., 2021) are also of sufcient quality, and PE and
EBAR were also detected with certainty (Agerbirk et al., 2021). How-
ever, hydroxy or methoxy derivatives of PE, BAR or EBAR were not
detected (Agerbirk et al., 2021). Similarly, putative constituents like BZ,
1mEt, 1mPr/2mPr and isomers of mBu and mPe have been searched for
in both leaves and seeds but not detected (Agerbirk et al., 2021).
However, an apparent isomer of hydroxybutylGSL was detected, sug-
gesting that use of amino acid precursors for GSL biosynthesis in
R. luteola is not restricted to Trp and homoPhe, as in B. vulgaris, but
include an aliphatic precursor that could be Leu or Ile. The isomeric
hydroxybutylGSL was not identied. In contrast, there is conclusive
evidence of 2h2mPr (31) in Reseda alba (Gmelin et al., 1970), proving
that Leu is a GSL precursor in the genus.
A literature report (Bennett et al., 2004a) of R. odorata being similar
in GSL prole to R. luteola could not be conrmed (Agerbirk et al.,
2021). In contrast, R. odorata was dominated by 2RhaOBZ (109)
(Agerbirk et al., 2021). A recent reliable report of the GSL prole of
Reseda lutea found BZ, 3hBZ, 2RhaOBZ, IM and an HPLC-UV trace of PE
(Pagnotta et al., 2020). That report also listed all so far reported GSLs in
the genus (=the above-mentioned).
While several Reseda species accumulate IM, neither of the usual IM
derivatives for tribe Cardamineae (Fig. 3) have been detected, although
they were specically searched for by Agerbirk et al. (2021). There is
only one report of a root GSL prole of a Reseda species (Pagnotta et al.,
2020), and no IM derivative was detected. However, roots of more
species should be investigated before conclusions on presence or
absence of substituted indole GSLs in this family can be made.
3. Discussion
3.1. Glucosinolate variation in a phylogenetic perspective
We applied a wide evolutionary perspective to distinguish the orig-
inal GSL genetic heritage of the tribe Cardamineae from GSLs originated
by biosynthetic innovation, based on a recent phylogeny of the order
Brassicales (Edger et al., 2018). It is generally agreed (Blaˇ
zevi´
c et al.,
2017; Edger et al., 2018) that GSL biosynthesis from Phe/Tyr and
standard (non-chain elongated) BCAAs is ancient (Fig. 5A), with
well-established occurrences even in the most basal families Tropaeo-
laceae and Akaniaceae (Blaˇ
zevi´
c et al., 2017). The GSLs from Trp and
homoPhe are also considered ancient, with well-established occurrences
in several families (Blaˇ
zevi´
c et al., 2017) including Resedaceae. The
Trp-derived GSLs were suggested to have evolved after the split of
Limnanthaceae from a remaining set of 11 derived families (Edger et al.,
2018), but that suggestion did not take into account a report of
considerable levels of IM in Tropaeolum majus from Tropaeolaceae
(Ludwig-Müller et al., 2002). In agreement with the latter report, Ben-
nett et al. (2004a) reported traces of IM in T. majus seeds. A renewed
search for Trp-derived GSLs may be needed in basal families to resolve
whether they have a monophyletic origin.
Concerning side chain modication, we note widespread occurrence
of the β-hydroxylated 2h2mPr (31) in, e.g., Tropaeolaceae (Dax-
enbichler et al., 1991), Akaniaceae (Montaut et al., 2015), Limnantha-
ceae (Daxenbichler and VanEtten, 1974), and Gyrostemonaceae
(Daxenbichler et al., 1991), as well as BAR and EBAR in Resedaceae
(Section 2.3) (Fig. 5A). We conclude that β-hydroxylation appears to be
as ancient as GSL biosynthesis itself. The distribution of side
chain-modied Trp-derived GSLs in basal families is poorly known,
except for occurrence of N-hydroxylation and subsequent methylation in
two basal families, Salvadoraceae (Bennett et al., 2004b) and Tovar-
iaceae (Schraudolf and B¨
auerle, 1986), so there is some basis for sug-
gesting N-methoxylation to be rather ancient. Concerning substituted
Trp-derived GSLs, Mithen et al. (2010) claimed to detect 4moIM in
Tovariaceae, but the illustrated chromatogram was compatible with
1moIM but not 4moIM in this family, in agreement with the previous
report (Schraudolf and B¨
auerle, 1986). The relative age of substitution
at the 1- and 4-positions illustrated (4-substitution more recent than
1-substitution) reects the simplest hypothesis according to the current
insufcient data (Fig. 5A).
Likewise, it seems generally agreed that recruitment of chain elon-
gated Met happened in a common ancestor of a few derived families in
Brassicales (after a whole genome duplication). It was proposed that
three derived families exclusively shared this group of GSLs: the Bras-
sicaceae, Cleomaceae and Capparaceae (Mithen et al., 2010; Edger et al.,
2015), along with a characteristic set of side chain modications related
to the thio bridge (formation of sulnyls, sulfonyls and alkenyls)
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
11
(caption on next page)
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
12
(Fig. 5B). In a newer report comparing MAM-genes considered to be
responsible for Met-chain elongation for GSL biosynthesis, data were
presented that could suggest independent recruitments of chain elon-
gated Met in the Cleomaceae and Brassicaceae (Abrahams et al., 2020).
However, a few reports from basal families, concerning 2homoMet
derived 4mSOb and Buen in Gyrostemon ramulosus (Gyrostemonaceae)
(Bennett et al., 2004a) and 3mSOp and 4mSOb in Moringa oleifera
(Moringaceae) (Maldini et al., 2014), contrasts with the hypothesis of
one or two recent origins of n-homoMet derived GSLs. These surprising
ndings were not discussed and the identity of the G. ramulosus seeds
were not further tested (Bennett et al., 2004a). If these reports from
Gyrostemonaceae (Bennett et al., 2004) and Moringaceae (Maldini
et al., 2014) can be reproduced, the origin of n-homoMet derived GSLs in
the evolution of the Brassicales would seem much more complex. In
general, the number of high-quality studies of the non-Brassicaceae
families is limited and often not focused on testing lack of specic
GSLs (e.g. Grifths et al., 2001; Montaut et al., 2015). In addition, many
studies lack either MS-detection or use of authentic standards (e.g.
Matth¨
aus and ¨
Ozcan, 2002; Bor et al., 2009; Mithen et al., 2010). Taken
together, the exact timing of the gain of n-homoMet derived GSLs, the
AOP2 gene needed for the formation of Buen, and many other features
warrants further research. In the following, we will accept the proposed
recent origin of n-homoMet-derived GSLs in derived families (Edger
et al., 2015; Abrahams et al., 2020), although we note the occasional
conicting literature reports. In more basal species, chain elongation of
Phe to form homoPhe and eventually PE is well-established in Reseda-
ceae, while structurally inconclusive signs of GSLs from chain elongated
BCAAs or Ala were reported from the related Tovariaceae and Gyro-
stemonaceae (Mithen et al., 2010).
The two known GSLs derived from homoIle would appear to be of
similar intermediate age as the n-homoMet-derived. This is the case for
2h2mBu (29S) with the trivial name glucocleomin (Fig. 5), which is
known from Brassicaceae (e.g. Olsen et al., 2016); Capparaceae (Dax-
enbichler et al., 1991) and Cleomaceae (Kjær and Thomsen, 1962). From
a biosynthetic argument, we place the apparent parent GSL 2mBu (54)
in the same group of intermediate age GSLs (Fig. 5). MethylGSL derived
from Ala seems to be restricted to derived families other than Brassi-
caceae (Mithen et al., 2010), while the existence in nature of GSLs
potentially derived from chain elongated Ala needs conrmation
(Blaˇ
zevi´
c et al., 2020). Neither group of Ala-related GSL is discussed
further. Finally, as recently evolved GSLs, we dene all those that are not
known from basal families and are not widely present in the three most
derived families.
In order to present the essence of our literature review (Supple-
mentary Table S1), we selected a representative variety of tribe Carda-
mineae species based on their contrasting and well-investigated GSL
proles. One species was included for its complementary GSL prole
only, being the only known occurrence of 2h2mBu (29S) and the only
conclusive occurrence of 2h2mPr (31) in the tribe: C. concatenata
(Daxenbichler et al., 1991). The identications of 29S and 31 were
reliable, but remaining diversity in C. concatenata not well described.
We compiled presence, circumstantial evidence of presence or
apparent absence of specic GSLs (Fig. 5C). Apparent absence (“Tested,
not reported”) was dened as lack of report after analysis using methods
that we judged ought to have revealed the particular GSL (Section 4.2.).
From the literature review, it was rarely possible to conclude “not
found” or “absent” for a particular GSL, because few authors explicitly
indicated what was not detected. We have elsewhere exemplied
(Agerbirk et al., 2021) how specic search using extracted ion HPLC-MS
chromatograms for a given GSL often results in its detection as a minute
peak not otherwise apparent. Some authors may even deliberately
ignore peaks considered to be insignicant in certain contexts. Hence,
lack of reporting a GSL can probably not in general be interpreted as “not
found”.
We noticed that the selection of representative species concerning
GSL diversity created a bias for species with unusual GSL composition,
including those anticipated to be a result of recent innovation. The
compilation revealed that a large proportion of GSLs known from basal
families in the order also occurred in one or often several species in the
tribes Cardamineae and Brassiceae. However, there was a tendency for a
more reduced diversity of the ancient types in the tribe Brassiceae
(Fig. 5C). We concluded those scattered GSLs also known from basal
families to be of ancient or recapitulated nature (mechanisms of “reca-
pitulation” are discussed below). In contrast, despite the selection bias
for species with rare GSLs, several GSLs not known from basal families
showed a highly scattered occurrence in the tribes Cardamineae and
Arabideae, with one of them, 3mPe (58) derived from 2homoIle, also
found scattered in Brassiceae. We concluded these to be probable results
of recent biosynthetic innovation (Fig. 5D).
3.2. Glucosinolate evolutionary history in a biochemical perspective
We will inspect two aspects of GSL biosynthesis separately: forma-
tion of parent GSLs and subsequent modication (Fig. 2). Parent GSLs
include IM from Trp, BZ from Phe, PE from homoPhe or 3mPe from
2homoIle. Subsequent modication usually concerns the side chain (e.g.
hydroxylation of the parent GSL PE to the modied BAR), although the
thioglucose moiety can also be modied.
In order to be able to discuss the deduced evolution of GSL biosyn-
thesis in tribe Cardamineae in a phylogenetic perspective, we compiled a
phylogenetic tree comparing some of the investigated species and tribes
in Brassicaceae with Reseda as an outgroup (Fig. 6). For a better
phylogenetic overview over the entire family, the interested reader is
referred to Huang et al. (2020). The few selected species were attempted
to be representative for the entire dataset (Supplementary Table S1)
with respect to either GSL diversity or phylogenetic position.
The resulting diagram (Fig. 6A) conrms the well-known fact that
GSLs derived from Trp, homoPhe and n-homoMet are widespread in the
Brassicaceae. Those from Trp seems to be almost universally present,
while those from homoPhe and/or n-homoMet are occasionally absent.
When those from n-homoMet are absent, one or more other, non-Trp
derived GSLs are consistently present (Figs. 5 and 6). Concerning
those derived from n-homoMet, the diagram suggests some sub-
structure in different chain length proles and secondary loss of this
group. There is a tendency for short side chain n-homoMet derived GSLs
in tribe Brassiceae and longer side chains in Cardamineae. An exception
Fig. 5. Structural redundancy and innovation in glucosinolate (GSL) biodiversity. A. Representative GSL structures categorized as ancient due to presence in non-
Brassicaceae members of the order Brassicales. The poorly known status for a substituted Trp-derived is indicated (see text). B. Representative GSLs from three
derived families (Capparaceae, Cleomaceae and Brassicaceae) with a simplied indication of the biosynthetic connections of n-homoMet derived GSLs. C. Distri-
bution of three groups of GSLs in selected members of the tribes Cardamineae, Arabideae and Brassiceae. The rst group, those illustrated in panel A, seem to be due
to ancient or recapitulated biosynthesis. The second group seem to be of intermediate age, the n-homoMet derived are pooled for space considerations. Possibly, the
4-substituted Trp derived 4moIM (48) and homoIle derived 54 and 29 also belongs to this group. The third group is deduced to represent recently evolved bio-
syntheses, as discussed in text, and the GSLs are illustrated in panel D. The category “Present” in panel C indicates one or more conclusive demonstrations of the
relevant GSL, while “Tested, not reported” means that relevant organs have been tested using relevant methods, yet the GSL was not reported, although explicit
search for the GSL was not necessarily reported. Hence we could not conclude the GSL to be “not found”, although this would be the simplest interpretation. The
category “Circumstantial evidence” means that reasonable but not conclusive evidence for the relevant GSL has been published, while the category “Insufcent or
missing data” means that relevant organs (roots for substituted Trp-derived and seeds for SGlc-acylated) have not been sufciently investigated using methods with
demonstrated ability to reveal the GSL in question.
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
13
(caption on next page)
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
14
for the former tribe is S. arvensis, with polymorphism for either short or
long side chains. Of the ancient carbon skeletons reecting Phe and
standard BCAAs as precursors, there is a scattered but regular occur-
rence of the ancient precursor amino acids (when taking into account
the over-representation of the genus Barbarea in the diagram). The GSLs
derived from chain elongated Ile (1-2homoIle) occur in a highly scat-
tered pattern (Figs. 5 and 6).
Concerning side chain modication, the deduced presence of
CYP81F and IGMT genes, believed to be responsible for hydroxylation
and subsequent methylation of Trp-derived GSLs in the entire Brassi-
caceae, was very common in this family, and the two features always co-
occurred. Recent analysis of the C. hirsuta genome supports our scoring
of these genes based on GSL proles (Czerniawski et al., 2021). Simi-
larly, the scoring of hydroxylation at beta (OH at β), irrespective of
precursor amino acid, illustrates how widespread this type of modi-
cation is. In contrast, all the remaining deduced types of side chain
modication seem to be rare, possibly associated with biosynthetic
innovation. The occurrence of two types of modication, however,
cannot be reliably scored: hydroxylation at delta (OH at δ) can only
reasonably be scored if the relevant substrate 3mPe is present, which is
in itself extremely rare as it requires 2homoIle as precursor. Likewise,
absence of methoxylation of hydroxyl groups of benzenic GSL can only
be deduced reliably if the phenolic precursor is present. In contrast, the
very rare nature of the deduced hydroxylation at the meta-position (OH
at meta) is obvious (Fig. 6A). While acylation of the thioglucose moeity
occurs in Barbarea and otherwise appears to be rare, many reports have
not presented a reliable search.
The simplest evolutionary hypothesis to explain this pattern is that a
common ancestor of the family exhibited those biosynthetic traits that
are both known in non-Brassicaceae and are widespread within the
family. Likewise, traits that are widespread in the family and in tribe
Cardamineae can be assumed to have been present in a common
ancestor of the tribe. Hence, from the compilation of GSL proles (Figs. 5
and 6, Supplementary Table S1) we suggest that common ancestors of
both the family Brassicaceae and the tribe Cardamineae exhibited the
entire genetic repertoire of GSL biosynthesis from Phe, possibly Tyr (see
below), non-chain elongated BCAAs (Val, Leu, Ile), Trp, homoPhe and n-
homoMet. Possibly, GSLs from homoIle were also present in a common
ancestor. This could be argued from their presence in the sister families
Capparaceae and Cleomaceae, but their extreme scarcity in tribe Car-
damineae and the entire family Brassicaceae suggest that they were not
present in a common ancestor but are a result of independent de novo
evolution in the lineage leading to C. concatenata. Concerning secondary
modication, the compilation suggests that the common ancestors of
family Brassicaceae and tribe Cardamineae exhibited β-hydroxylation of
aliphatic chains (OH at β), hydroxylation and subsequent methylation in
Trp-derived GSLs, and modication related to the thio bridge in the n-
homoMet derived GSLs (Fig. 5B). A similarly broad repertoire of pre-
cursors is currently possibly realized in A. rusticana, while proles of, e.
g., R. amphibia, A. thaliana, C. diphylla and C. hirsuta approach this di-
versity. It is worth noting that most of the repertoire may accumulate at
trace levels only, such as most GSLs in A. rusticana (Agerbirk et al.,
2021). We will use the term “recapitulation” for such occasional pres-
ence of ancient types of GSLs in distal and apparently isolated positions
of an evolutionary tree, irrespective of the genetic/biochemical basis.
Biosynthesis from Tyr cannot easily be distinguished from
biosynthesis from Phe combined with para-hydroxylation (although the
entire palette of structures in a species can provide a hint) (Pagnotta
et al., 2017), so these features are shown in an intermediate position in
Fig. 6A and B. As such biosynthesis from Phe/Tyr is very frequent and
also occurs in basal families, it may well represent an original trait in the
tribe Cardamineae and family Brassicaceae. Comparison of S. alba and
S. arvensis suggests that GSL biosynthesis is via Phe in S. alba but directly
from Tyr in S. arvensis (Fig. 6A). However, both occurrences can be
regarded as recapitulation of ancient diversity.
3.3. Glucosinolates resulting from biosynthetic innovation
From the present investigation, other GSLs and features seems to
occur as results of recent evolutionary innovation of biosynthetic traits
(Fig. 5D), with innovation meaning de novo biosynthetic evolution
leading to biosynthetic steps (=enzymes) previously unprecedented in
the evolutionary line in question (“gain of synthetic ability in the given
plant lineage”, Pichersky and Lewinsohn, 2011).
Sixteen GSLs in tribe Cardamineae, eleven specied in Fig. 5D plus
ve 6’iFe derivatives, are suggested here to result from de novo evolu-
tion in the tribe or in the family Brassicaceae (Table 3). Five are BCAA-
derived. Two are simply β-hydroxy derivatives of ancient parent GSLs,
which is the case for 1hmEt (57R) and 1hmPr (30). Three belong to the
recently discovered class of 2homoIle-derived GSL: 3mPe (58), 2h3mPe
(141) and 2h3mPe (149), all requiring two chain elongations of Ile. The
δ-substituted 3hmPe (141) would seem to be the most radical innova-
tion among the aliphatic GSLs, and could potentially give rise to a new
type of cyclic product (a seven-membered ring analogous to the ve-
membered ring in Fig. 2). Hydroxylation at δ is not simply an addi-
tional product of a common enzyme for hydroxylation at β (Olsen et al.,
2016). Some of this innovation (30, 57R and formation of 149 from 58)
can be characterized as recapitulation of an existing chemistry in GSL
biosynthesis (OH at β), while use of chain elongated Ile and δ-substitu-
tion are more radical novelties. The highly scattered distribution of
1-2homoIle derived GSLs suggest multiple independent incidents of
biosynthetic innovation, or possibly recapitulaton in case of those
derived from homoIle, a group with poorly understood history (Section
3.1.).
If the very frequent Trp-derived 4moIM (48) is due to recent inno-
vation, it must be in a common ancestor of the entire family Brassicaceae
or a couple of derived families. In contrast, the disubstituted Trp-derived
1,4moIM (138) is quite rare according to the present investigation. Only
one laboratory has ever reported this GSL, so species selection may be
biased, but the distribution is dispersed: three isolated cases in the tribe
Cardamineae (R. amphibia, C. pratensis and B. vulgaris P-type), and one
isolated case in another Brassicaceous tribe (Brassiceae) at minute levels
(Agerbirk et al., 2008). Detailed inspection suggests two types of recent
biosynthetic evolution (Section 3.6).
Five phenolic GSLs were apparently de novo evolved (4hPE (140),
3hPE (148), 4hBAR (139S), 3hEBAR (142R) and 4hEBAR (139R)) and
are meta- and para-derivatives of ancient GSLs (Fig. 5D). A further two
(x1, x2) await nal structural elucidation but are known to involve
combined meta- and para-hydroxylation and methylation in one of the
two positions (Fig. 3). These motifs are similar to apparent innovation in
a distinct lineage, tribe Arabideae (Fig. 5C and D). Similar to the situ-
ation in Sinapis discussed above, the innovation in Arabis is probably due
Fig. 6. Aspects of glucosinolate (GSL) evolution in the order Brassicales with focus on the tribe Cardamineae. A. Phylogeny matched with GSL structural or
biosynthetic features. Structural and biochemical features of GSL proles of the respective species are indicated as deduced precursor amino acids and deduced
modication of parent GSLs from the various precursors. Categories are based on GSL proles as in Fig. 5C, but interpreted in a biosynthetic context. Presence of para-
hydroxylated phenyl groups can potentially be due to either use of a specic precursor amino acid (Tyr or homoTyr) or a specic modication (para-hydroxylation),
and is hence shown in an intermediate position, with the relevant backbones shown in B. Panel C shows the deduced modication steps. For the phylogeny in A,
phylogenetic relationships based on Bayesian inference (MrBayes) of ITS regions were calculated for a subset of species from Brassicaceae and using Reseda
(Resedaceae) as outgroup. Labels for B. vulgaris (group 3, group 7) refer to the ITS sequence pools listed in Agerbirk et al., (in review). Bootstrap values from 1000
replicates are shown for Bayesian and maximum-likelihood inference, respectively. Side-chain modication exclusively known from n-homoMet derived GSLs
(Fig. 5B) is left out for space-considerations. For GSL prole data, group 7 of B. vulgaris was assumed to represent ssp. vulgaris.
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
15
to biosynthesis from Tyr, because non-phenolic counterparts were
lacking, while the innovation in tribe Cardamineae can be due to sec-
ondary modication, because of the meta-substituted structures and the
cooccurence of non-phenolic counterparts. In any case, the structural
motif (phenolic GSLs) was present in the ancient set of GSLs, so the
innovation can be viewed as “recapitulation-like”. Indeed, two meta-
substituted GSLs based on non-chain elongated Phe, 3hBZ and 3moBZ,
are known in basal families of the order Brassicales (Limnanthaceae;
Resedaceae etc.) and in Lepidium (Daxenbichler et al., 1991; Pagnotta
et al., 2017, 2020) but not known from tribes discussed here (Fig. 5).
Formation of phenols and methylation could also be recruited from e.g.
avonoid biosynthesis (Alseekh et al., 2020). However, a different class
of hydrolysis product from 4hBAR and 4hEBAR contributes an ecolog-
ical potential to this otherwise trivial innovation (Section 3.6.).
The ve 6’iFe derivatives are obviously dependent on the same type
of parent GSL modication. Concerning 6′-acylation, it is only known
from very few genera (reviewed by Andini et al., 2019), as conrmed by
systematic screening by Agerbirk et al. (2021) and in each genus mainly
with different aromatic acids (isoferulic acid in Barbarea, benzoic acid in
Arabidopsis, sinapic acid in Sinapis and Raphanus). However, minor
peaks compatible with acylation with sinapic acid were revealed in both
A. thaliana (Kliebenstein et al., 2007) and B. vulgaris (Bianco et al.,
2013), uniting these phenotypes with the recent report from Sinapis
(Andini et al., 2019) and the original report from Raphanus (Linscheid
et al., 1980). Hence, acylation with e.g. sinapic acid at very low levels,
usually not detected, may be widespread, with the cases shown in Fig. 6
being a result of recent quantitative evolution (and diversication to use
of different aromatic acids). Although this hypothesis may seem spec-
ulative, it should be kept in mind because many past and present labo-
ratories fail to carry out analyses in ways that allow detection of acyl
derivatives.
Although the relevant GSLs are known from many other tribes, those
derived from 7-8homoMet in Rorippa, Planodes, and Sinapis (Fig. 6),
could be considered as due to recent mutation, this interpretation should
be discussed in a biochemical context (Section 3.5.). The same is the case
if putative GSLs derived from 2-4homoPhe in A. rusticana can be
conrmed (Table 2).
Innovations in GSL biosynthesis are not restricted to the tribe Car-
damineae, but discussion is difcult because most reports do not sys-
tematically analyze related species to place potential innovations in a
phylogenetic context. Analysis of a previous paper (Windsor et al., 2005)
using the same logic as applied here, would suggest
ω
-hydroxyalkylGSL
as novelties in A. thaliana and some related species, controlled by the
AOP3 gene. MethoxyarylalkylGSL in Arabis (Fig. 5) and benzoic acid
esters in A. thaliana also appear to be due to recent biosynthetic inno-
vation (Reichelt et al., 2002; Czerniawski et al., 2021; Agerbirk et al.,
2021).
3.4. Association of loss and innovation
Ancient types of GSLs that are not derived from n-homoMet (Fig. 5)
are seen in two contexts: Co-occuring with n-homoMet derived GSLs, as
in e.g. A. rusticana, C. impatiens, and C. hirsuta, and without n-homoMet
derived GSLs in at least 5–6 taxa: Barbarea, C. diphylla, C. pratensis, C.
contatenata, C. cordifolia and possibly C. amara (but contradicted by
Pellissier et al., 2016) (Supplementary Table S1A). Inconclusive data
based on HPLC-UV furthermore suggested lack of n-homoMet derived
GSLs in C. asarifolia and C. kitaibelii, in the former also associated with
ancient type GSLs (Supplementary Table S1A) (Pellissier et al., 2016).
Like previous authors (Windsor et al., 2005), we interpret lack of
n-homoMet derived GSLs as a result of secondary loss of their biosyn-
thesis. Such loss is otherwise quite rare (Windsor et al., 2005; Agerbirk
et al., 2008) so the frequency in Cardamine seems high. Interestingly, for
3–4 of the species (C. diphylla, C. concatenata, C. asarifolia and possibly
C. amara), the widespread ability to form GSLs derived from homoPhe
also seemed to be lost (Fig. 6C), in agreement with a common biosyn-
thesis of n-homoMet and homoPhe derived GSLs (Section 3.5.). Several
GSLs interpreted as the result of de novo evolution or recapitulated
Table 3
Glucosinolates from the tribe Cardamineae tentatively concluded (or suggested, bottom of table) to have evolved de novo in the tribe Cardamineae or generally in the
family Brassicaceae.
Precursor amino
acid
Glucosinolate Key structural characteristic Deduced biochemistry General comments
Val 1hmEt (57R) Non-ancient β-hydroxylation Possibly GS-OH like? Widespread in family
Ile 1hmPr (30) Do. Do. Do.
2HomoIle 3mPe (58) Precursor Twice chain elongation of Ile Also in tribe Brassiceae
2HomoIle 2h3mPe (149) Do. Do. Unique for C. pratensis.
2HomoIle 3hmPe (141) Precursor and atypical position of
hydroxylation (δ)
Twice chain elongation of Ile, atypical GS-
OH?
Unique for C. pratensis.
HomoPhe 3hPE (148) Phenolic Ring-oxidation (meta) Unique for Barbarea
HomoPhe 3hEBAR (142R) Do. Do. Do.
HomoPhe/
homoTyr
4hPE (140) Do. Tyr as precursor or ring-oxidation (para) In Barbarea and also in tribe Arabideae
HomoPhe/
homoTyr
4hBAR (139S)
4hEBAR (139R)
Do. Do. In Barbarea, 139R also in Planodes and tribe
Arabideae
HomoPhe/
homoTyr
3/4hmoPE (x2)
3/4hmoEBAR (x1)
Di-substituted with further methyl Ring-oxidation twice or Tyr as precursor,
methylation
Unique for Barbarea, position of methyl not
claried
Trp 1,4moIM (138) Combined 1- and 4-substitution CYP81F +IGMT Also in tribe Brassiceae
Phe 6’iF PE (129)
6’iF BAR (131S)
6’iF EBAR (131R)
6′-Isoferuloyl group Acylation Unique for Barbarea but 6′-acylation
widespread
Phe/Tyr 6’iF 4hEBAR
(132R)
Do. Do Do.
Trp 6’iF IM (130) Do. Do Do.
Tentative de novo evolved:
HomoIle 2mBu (54) Precursor Chain elongation of Ile Also in other families and non-Brassicales
HomoIle 2h2mBu (29S) Do. Do. Do.
7-8HomoMet 65, 68, 77, 79,
[89]
Highly elongated MAM characteristics Widespread in family
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
16
biosynthesis seem to be associated with apparent loss of n-homoMet
derived GSLs (Figs. 5C and 6). Residual levels of n-homoMet derived
GSLs were carefully searched for in these cases (Windsor et al., 2005;
Olsen et al., 2016; Agerbirk et al., 2021), as a conrmation of the pre-
dicted evolution. With steadily increasing analytical sensitivity, we nd
it likely that n-homoMet-derived GSLs will one day be detected in those
species. Meanwhile, characterization of the biosynthetic genes serves
the same purpose and is complementary evidence (Liu et al., 2016,
Wang et al., 2021). Even if residual traces are one day revealed, loss of
n-homoMet-derived GSLs in the sense of a quantitative decline of several
magnitudes, from major metabolites to trace level metabolites or less,
may be equivalent to absolute loss in an ecological sense and in terms of
being an evolutionary factor leading to increased biosynthesis of other
GSLs.
In Barbarea, loss of n-homoMet derived GSLs was apparently
accompanied by very high levels of PE, BAR and/or EBAR and occa-
sionally phenolic derivatives (Windsor et al., 2005; Agerbirk et al.,
2015). In a few taxa (especially in P-type B. vulgaris), derivatives occur
that are probably due to biosynthetic innovation in the tribe Cardami-
neae or in the genus Barbarea. However, the usual dominant GSLs in
Barbarea spp. (PE, BAR or EBAR) are also present at high levels in several
tribe Cardamineae members that retain n-homoMet-derived GSLs (e.g.
N. ofcinale, P. virginica), and a phenolic derivative was detected in
P. virginica, so it is easy to imagine a stepwise evolution ending at the
deviating GSL prole of Barbarea. It is conspicious that the only tribe
members with high levels of phenolic derivatives of homoPhe-derived
GSLs are in the genus Barbarea. This pattern (and the seasonal regula-
tion and unique reactivity of 4hEBAR) suggests an ecological role of
4hPE or 4hBAR/4hEBAR in Barbarea, possibly substituting for or com-
plementing effects of n-homoMet-derived GSLs (Section 3.6).
Deduced loss of n-homoMet-derived GSLs in C. pratensis, C. diphylla,
C. concatenata and C. amara seems to be associated with GSL biosyn-
thesis based on standard BCAAs (and Phe/Tyr in C. pratensis and
C. amara), but in particular with development of rare GSLs based on 1-
2homoIle (Fig. 5C). Interestingly, the dominant of these in C. diphylla
and C. concatenata are derived from homoIle (possibly classifying as
recapitulation), while the dominant in C. pratensis are derived from
2homoIle, suggesting independent evolution. Use of 2homoIle as pre-
cursor is to our knowledge not known from any member of the order
Brassicales outside the family Brassicaceae, as opposed to ancient bio-
syntheses based on non-chain elongated BCAAs (Fig. 5), and the
occurrence of GSLs derived from homoIle in other families that may be
due to independent evolution (Section 3.1.). The same pattern of com-
bined loss and gain was likewise observed in Coincya rupestris that had
apparently lost or nearly lost the biosynthesis of 4hBZ (Agerbirk et al.,
2008); these species were called “hot spots” of GSL evolution. In some
agreement with this observation, an HPLC-MS signal attributed to the
apparently de novo evolved 3-methoxycarbonylpropylGSL in Erysimum
was only found in few species, and was found at high levels only in
species with low levels of n-homoMet derived GSLs (Züst et al., 2020).
In the suggested evolution combining loss and gain of biosynthetic
capabilities, genetic polymorphism would obviously have to occur in the
population at the time of initial spread of a mutation. However, after
initial loss of a biosynthetic step followed by compensatory gain of
another, additional genetic polymorphism would be expected due to the
possibility of more than one new biosynthesis taking over. Indeed, in
two groups that may have lost Met-derived GSL biosynthesis (Barbarea
and “C. pratensis and its allies”), polymorphism concerning the deduced
derived GSLs has been discovered. In the genus Barbarea, the poly-
morphism mainly concerns the differences between the P-type and G-
type and is limited to secondary modication of parent GSLs. In
C. pratensis and its allies, in contrast, polymorphism includes a broad
range of parent GSLs (derived from Val, Leu, Ile, 1-2homoIle, Phe and
possibly Tyr), as well as a variety of secondary modications (Agerbirk
et al., 2010a). These cases seem to support the hypothesis of evolution
happening by a combination of loss and gain of biosynthetic characters.
Isolated occurrence of a GSL derived from 2homoIle was also re-
ported from two species in tribe Brassiceae that have not lost n-homoMet
derived GSLs (Fig. 5C). GSLs derived from 2homoIle have so far only
been reported by three laboratories (Agerbirk et al., 2008, 2010a;
Montaut et al., 2010; Olsen et al., 2016; Badenes-P´
erez et al., 2020), and
some authors might have confused detected 3mPe with isomers such as
the very poorly evidenced, putative “hexyl GSL” ([20]) listed by Fahey
et al. (2001) in one species and unreliably claimed by Bennett et al.
(2004a) in C. pratensis. Even if we include reported isomers as in-
dications of 3mPe, this GSL seems to be very rare in the family, as ex-
pected from a result of rare incidents of biosynthetic innovation rather
than recapitulation.
We discussed above whether GSLs derived from homoIle should be
classied as intermediate age or recent, due to their existence in Cap-
paraceae and Cleomaceae (Section 3.1.). However, 2mBu (54) and
2h2mBu (29S) derived from homoIle are known from a non-Brassicales
plant, P. roxburgii (Putranjivaceae) (Kjær and Friis, 1962), along with
1mEt (56) and 1mPr (61). In agreement with this nding, 2h2mBu was
identied in the related D. euryodes, while GSLs apparently derived from
Phe and Tyr were found in other members of the family (Montaut et al.,
2017). The GSLs in Putranjivaceae (order Malpighiales) (Table 1) are
considered an independent occurrence of GSLs from the GSLs in the
order Brassicales (Rodman et al., 1998), based on the huge phylogenetic
distance, and show that remarkable similarities can occur independently
in plant biochemical evolution (Pichersky and Lewinsohn, 2011; Takos
et al., 2011; Huang et al., 2016; Hansen et al., 2018). Likewise, an un-
identied “n-methylbutylGSL” (that must require chain elongation,
either of Leu, Ile or Ala) was found in multiple Erysimum species (Züst
et al., 2020). Biosynthetic information is needed to distinguish recapit-
ulation and innovation (Hansen et al., 2018), or even lateral gene
transfer among distantly related plants (Dunning and Christin, 2020);
distribution data as presented here can only provide hypotheses
(Pichersky and Lewinsohn, 2011).
To summarize sections 3.1 to 3.4., we stress the difculties in
extracting reliable GSL diversity data from the phytochemical literature
and in interpreting biosynthesis from the diversity. With due consider-
ation of these difculties, we suggest a wide structural diversity in a
common ancestor-population of the family Brassicaceae, and later evo-
lution by loss of function, recapitulation and occasional biosynthetic
innovation. These three categories seemed to be correlated, based on a
moderate number of cases, possibly suggesting a causal connection. In
order to better understand the genetic composition of the suggested
common ancestor, and how the syndromes of multiple, apparent
biosynthetic innovations in e.g. Barbarea, C. diphylla, C. pratensis and
C. contatenata could have evolved, we nally discuss genetic and
biosynthetic information in relation to our hypothesis. Since side chain
modication has signicant ecological potential, ecological aspects will
be included briey.
3.5. Biosynthesis of parent glucosinolates from amino acids
The genetic basis of evolution that mainly recapitulates itself could
either be a limited number of enzymes with varying expression or a set
of exible enzymes with a limited range of specicities obtainable by
simple back-and-forth mutation of active-site amino acid residues, as
previously reviewed (Olsen et al., 2016). Since our last review of this
theme, much additional evidence has appeared.
The committed step for biosynthesis of simple (non-chain elongated)
GSLs is conversion of an
α
-amino acid to an oxime by a cytochrome P450
enzyme, abbreviated CYP (Fig. 7A). The committed step is followed by
the remaining core structure biosynthesis pathway that leads to the
parent GSL (reviewed by Sønderby et al., 2010; Chhajed et al., 2020). An
example would be the conversion of Ile to 1mPr via the oxime (Fig. 7A).
Some amino acid precursors are initially chain elongated, a yet hypo-
thetic example is chain elongation of Ile to 2homoIle, followed by the
action of a CYP79 to form the corresponding oxime and the remaining
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
17
core structure pathway to give the parent GSL 3mPe (Fig. 7B).
The GSLs derived from non-chain elongated standard amino acids
are a minority among GSLs, although the group of Phe/Tyr-derived is in
itself a major group. They are derived from a limited number of amino
acids, including Phe, Val, Leu, Ile, Trp, and possibly Tyr in the tribe, and
probably including Ala and the non-standard amino acid 4-methoxyin-
dol-3-ylglycine in some other plants (Blaˇ
zevi´
c et al., 2020). The
known enzymology behind this broad range of non-chain elongated
precursors was recently limited to CYP79A converting Phe to an oxime,
CYP79B converting Trp to an oxime (reviewed by Sønderby et al., 2010),
and natural variant CYP79F enzyme in Boechera stricta converting either
Val or Ile to oximes (Prasad et al., 2012). Our interpretation of BZ and
4hBZ in Sinapis spp. to be recapitulation of an ancient biosynthesis is
supported by an analysis of the transcriptome of S. alba; a homolog of
the A. thaliana CYP79A was identied and suggested to be responsible
for the biosynthesis of BZ and 4hBZ (Zhang et al., 2016). The variant
CYP79F enzymes in B. stricta were products of a polymorphic
Branched-Chain-Methionine-Allocation locus causing variable levels of
GSLs derived from Val or Ile (and n-homoMet) (Schranz et al., 2009;
Prasad et al., 2012), which we interpret as mutation of substrate spec-
icity in CYP79F genes and enzymes. Concerning the concluded
frequent recapitulation of ancient GSL biosynthesis (Section 3.4.), this
polymorphism in B. stricta suggests that just a few mutations (Prasad
et al., 2012) in a gene involved in biosynthesis of n-homoMet-derived
GSLs may result in a changed specicity to include more ancient use of
amino acids. This could be one mechanism leading to the apparent
recapitulation of ancient biosynthesis deduced from phylogenetic com-
parisons (Figs. 5 and 6). However, the actual evolutionary history
leading to the polymorphism in B. stricta is not known.
Recently, Wang et al. (2020) demonstrated that a fourth type of
enzyme from A. thaliana, CYP79C, can convert several amino acids,
including Phe, Leu and Ile, into their corresponding oximes. Hence, we
are approaching a range of CYP specicities corresponding to the known
GSL precursors, although a few are still missing (e.g. Ala). Interestingly,
although both CYP79A and CYP79C members are functional in
A. thaliana, the genes are of very low expression and expression of
CYP79C genes is only detectable in some oral parts. Furthermore, their
downstream GSL products (which are known from homologous or het-
erologous overexpression) have so far not been demonstrated with
certainty in A. thaliana (Fig. 6A) (Wang et al., 2020). All in all, the
A. thaliana CYP79A and CYP79C genes exemplify that functional
biosynthetic genes for ancient types of GSLs can exist in a species even if
the resulting GSLs have so far escaped analytical detection. Increased
and possibly also less tissue-specic expression (Czerniawski et al.,
2021) of such genes is a second possible mechanism for apparent reca-
pitulation in an evolutionary line. According to evolutionary theory,
completely inactive genes would undergo pseudogenization and there-
fore not be expected to remain functional for long, so low or very
tissue-specic expression associated with some natural selection would
be required for any such unapparent ancient gene to later be subject to
increased expression or more general expression. Observation of higher
levels of some usually undetectable GSLs upon overexpression of cab-
bage transcription factors in N. ofcinale (Cuong et al., 2019) is addi-
tional support for the model. Indeed, one of the GSLs that only was
detectable after overexpression of transcription factors was claimed to
be Ile-derived 1mPr that is often interpreted to represent recapitulation
of ancient biosynthesis, but unfortunately, conclusive evidence for the
identity was not provided. Advances in molecular genetics of
Fig. 7. Stages in the biosynthesis of parent glucosinolates (GSLs) without (A) or with (B) chain elongation of the precursor standard amino acid. A CYP79 enzyme
catalyzes the rst reaction in the known (cytosolic) core structure biosynthesis pathways, followed by six enzymatic steps constituting the remaining core structure
biosynthesis pathway, abbreviated “r. csb”. For GSLs without chain elongation (A), the CYP79 catalyzed reaction is the committed step. For GSLs needing chain
elongation (B), however, the chain elongation machinery as well as transport (“T”) across the chloroplast membrane and reversible amino transferase reactions
collectively constitute the committed step, illustrated as a box-like reaction arrow containing the individual reactions.
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
18
N. ofcinale (Voutsina et al., 2016; Mand´
akov´
a and Lysak, 2019) should
allow mutation studies or overexpression studies to examine the genetic
basis of the many minor GSLs (Agerbirk et al., 2021) in this species.
Some GSLs can only be derived from standard amino acids after
chain elongation in the chloroplast, meaning that the “chain elongation
machinery” constitute the committed step, with the potential to control
structural specicity and ux (Fig. 7B) (reviewed by Sønderby et al.,
2010; Chhajed et al., 2020). This is the case for only three groups of GSLs
known with certainty in nature and corresponding to 1-8homoMet,
homoPhe and 1-2homoIle. Despite the small number of precursors,
these groups are collectively the majority of known GSLs (Blaˇ
zevi´
c et al.,
2020). Of these groups, the biosynthesis of the n-homoMet–derived GSLs
is extremely well characterized in A. thaliana and in Brassica crops
(Sønderby et al., 2010; Augustine and Bisht, 2017; Harun et al., 2020;
Kittipol et al., 2019; Chhajed et al., 2020). It is well established that
chain elongation is carried out by a set of enzymes including methyl-
thioalkylmalate synthase, genetically abbreviated “MAM”, that leads to
varying extent of chain elongation (Windsor et al., 2005; Textor et al.,
2007). A MAM homolog is present even in a basal family, and the origin
of the present diversity in the Brassicaceae is believed to be gene
duplication followed by diversication (Barco and Clay, 2019). As chain
elongation of Phe to homoPhe is the only chain elongation needed to
form GSLs known from basal families (Fig. 5), this is the predicted
function of the original MAM in GSL biosynthesis. Using biotechnolog-
ical approaches, just a few rationally selected point mutations in a MAM
from Brassica juncea resulted in changed chain-length prole of the
products (Kumar et al., 2019). In agreement with this plasticity, the
broad palette of chain lengths including highly chain elongated GSLs in
Nasturtium, Rorippa and Planodes (Agerbirk et al., 2021) is paralleled in
many other species outside the tribe, such as the common weed
S. arvensis (Grifths et al., 2001; Agerbirk et al., 2008), and this varia-
tion could be a result of mutation in the degree of processive elongation
by MAM enzymes. This hypothesis would be testable using established
methods (e.g. Kumar et al., 2019). A specic clade of MAM genes were
found to be correlated with long side chains in the tribe Camelineae
(Czerniawski et al., 2021).
An ortholog of A. thaliana MAM1 is also highly expressed in
B. vulgaris (Liu et al., 2016) although this species does not accumulate
any n-homoMet-derived GSL (Agerbirk et al., 2021). Given the predicted
original function of MAM in basal families, a changed substrate speci-
city of MAM from B. vulgaris (Wang et al., 2021) would classify as
recapitulation of an ancient biosynthesis, rather than innovation.
Recently, it was reported that mutant A. thaliana mam1 (lacking a
functional MAM1) does not accumulate homoPhe-derived PE, in
contrast to the wild type, suggesting that AtMAM1 accepts both Phe and
n-homoMet for chain elongation leading to GSL biosynthesis (Petersen
et al., 2019). A similar model was proposed for oilseed rape, Brassica
napus, based on correlation of gene expression (Kittipol et al., 2019). In
conclusion, the original function of MAM may have been chain elon-
gation of Phe. In the evolution to chain elongated Met in the common
ancestor of three derived families, the specicity for Phe was retained,
explaining the wide distribution of homoPhe-derived GSLs in the Bras-
sicaceae. Accumulation of GSLs from homoPhe but not n-homoMet in e.
g. Barbarea can be viewed as recapitulation of the ancient more narrow
specicity of the entire chain elongation machinery. Functional testing
of the properties of B. vulgaris enzymes showed that the more narrow
specicity realized in this species is inuenced by BvMAM1 properties
but cannot be entirely explained by MAM1 specicity (Wang et al.,
2021).
In addition to the recruitment of a gene from primary metabolism
leading to MAM genes in the Brassicaceae, an independent recruitment
seems to have led to a group of MAM-like genes in the sister-family
Cleomaceae (Abrahams et al., 2020). No functional investigation of
this group of genes has been presented, but it is assumed that the
MAM-like genes in Cleomaceae are also involved in chain elongation of
amino acids for GSL biosynthesis (Abrahams et al., 2020). They could be
involved in chain elongation of Met, to form n-homoMet derived GSLs
known from this derived family. However, they could also be involved in
a single chain elongation of Ile, forming the homoIle derived 2h2mBu
(29S, “glucocleomin”), which is characteristic for the family. Hence,
functional investigation of MAM enzymes in the Cleomaceae would be
relevant.
The enzymology of chain elongation of Ile is otherwise completely
unknown, but recent results show that the chain elongation machinery
from A. thaliana has several main products when expressed in a bacterial
host, resulting in chain elongation of not only Met and Phe as in the
native plant, but also of Leu not known to happen in the native plant
(Petersen et al., 2019). When tobacco was the heterologous host, Leu
was also chain elongated (Wang et al., 2020). These ndings suggest
that a ne balance of interactions controls the selection of amino acids to
be chain elongated, and that the chain elongation machinery for Met can
accept Leu but not Ile at certain conditions. It seems possible that a
mutation would allow a change of substrate from Leu to Ile and that such
mutation could simultaneously exclude Met and sometimes Phe from
chain elongation, resulting in the phenotypes reported for C. diphylla,
C. pratensis and C. concatenata (Fig. 5). Testing this hypothesis requires a
combination of descriptive studies (metabolomics, genomics) and
functional studies, employing e.g. knock-out, over-expression or heter-
ologous expression.
All in all, we suggest to experimentally test whether biosynthetic
evolution from GSLs derived from n-homoMet +homoPhe to 1-2homo-
Ile (as in C. diphylla +C. concatenata), from homoMet to 1-2homoIle
with retention of homoPhe (as in C. pratensis) and from n-homoMet +
homoPhe to solely homoPhe (as in Barbarea) in the tribe Cardamineae,
could be due to changes of the specicity of the chain elongation ma-
chinery and/or a subsequent CYP79-catalyzed step discussed below.
After chain elongation of Met, enzymes from the CYP79F subfamily
are known to catalyze the entry of n-homoMet precursors to the GSL core
structure biosynthesis (Sønderby et al., 2010), but a gene of this family,
CYP79F6, is also highly expressed in B. vulgaris (Liu et al., 2016). On this
background it has been suggested that the specicity of CYP79F en-
zymes may include homoPhe (Byrne et al., 2017), and experiments to
test this hypothesis are reported separately (Wang et al., 2021). Again,
the corresponding enzymology concerning 1-2homoIle is completely
unknown.
From experimental manipulation of CYP79 expression in A. thaliana,
it is well established that the remaining core structure biosynthesis
pathway is rather exible in terms of side chains (e.g. Petersen et al.,
2001). Two core structure pathways have been envisioned and termed
the aliphatic and aromatic pathways (Sønderby et al., 2010), but in
terms of side-chain specicity this distinction has later been challenged
(Wang et al., 2020; Wang et al., 2021).
3.6. Parent glucosinolate modication: biochemistry and ecology
A well understood GSL modication is modication of the parent
indole GSL: IM. In this type of modication, hydroxylation and subse-
quent methylation leads to either of two substituted Trp-derived GSLs
that are almost ubiquitous in the Brassicaceae (Fig. 6D) (Pfalz et al.,
2016). The ecological signicance of this type of modication is
increasingly well understood, as repeatedly reviewed ((Pastorczyk and
Bednarek, 2016; Blaˇ
zevi´
c et al., 2020). The rare di-substituted
Trp-derived GSL 1,4moIM has been found in three monophyletic
groups in the tribe Cardamineae (the C. pratensis allies, Rorippa and
Barbarea). As the biosynthetic steps involved (“CYP81F +IGMT” in
Fig. 6A) are seen throughout the family, the distribution could be due to
occasional change of substrate specicity of a CYP81F enzyme. In two
cases, 1,4moIM is present at low levels and co-occurs with the mono-
substituted 1moIM and 4moIM, suggesting a broadening of substrate
specicity. But in the third case, the P-type of B. vulgaris, 1moIM is
usually lacking and the resulting 1,4moIM is a major root GSL, which
could be due to a changed but equally narrow substrate specicity of a
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
19
CYP81F. As 1moIM is a major root GSL in the corresponding G-type, we
hope that the increasingly well-understood genetics of B. vulgaris will
lead to characterization of the responsible gene and any ecological
effects.
Another well-understood side chain modication is partial or full
oxidation of the side chain sulfur in n-homoMet derived GSLs, and
elimination to form GSLs with terminal unsaturation, e.g. from 4mSOb
to Buen (Fig. 5B). These typical Brassicaceae side chain modications
have been discussed by multiple authors and repeatedly reviewed
(Windsor et al., 2005; van den Bergh et al., 2016; Kliebenstein and
Cacho, 2016; Velasco et al., 2017; Cang et al., 2018; Barco and Clay,
2019). There is little doubt that their almost ubiquitous presence in the
family Brassicaceae is due to a common biochemical mechanism with
shared genetic background. The coherent phylogenetic distribution also
in tribe Cardamineae, suggests that this syndrome of side chain modi-
cation of n-homoMet-derived GSls could be due to a shared mechanism
also in tribe Cardamineae. The presence of homologs of one of these
genes (FMO
GS-OX
) in B. vulgaris has been interpreted as support of the
suggested evolution from an ancestor with n-homoMet-derived GSLs
(Byrne et al., 2017).
Side chain modication has ecological consequences by altering the
reactivity of GSL hydrolysis products, and this is particularly evident for
β-hydroxylation (Fig. 2). The oxazolidine-2-thiones (OATs) have
different defensive properties than isothiocyanates, as recently reviewed
(Müller et al., 2018). In Arabidopsis, this important biosynthetic step
(from Buen to 2hBuen) is controlled by the gene GS-OH, which is of
considerable ecological and agronomic importance (Hansen et al., 2008;
Kliebenstein and Cacho, 2016) (Fig. 8A).
In the genus Brassica, biosynthesis of 2hBuen from Buen is likewise
known (James and Rossiter, 1990; Velasco et al., 2017), although the
stereospecicity is higher than in A. thaliana. A homolog of GS-OH (in
Brassica known as GSL-OH) was suggested to exist in B. rapa (Zang et al.,
2009). The gene was reported to show similar position of two introns,
one small apparent deletion and 85% nucleotide sequence identity, all
compared to the A. thaliana gene (Zang et al., 2009). In the currently
available B. rapa GSL related gene database (http://brassicadb.org
/brad/glucoGene.php) (accessed Oct 28, 2020), three B. rapa GSL-OH
genes are listed (Bra021670, Bra 021671 and Bra022920) with the
A. thaliana GS-OH discovery listed as reference. Correlation studies of
GSL-OH genes in Brassica oleracea have been performed, by comparing
GSL levels and proles with gene expression in inbred lines. The study
found that high expression of a B. oleracea GSL-OH was correlated with
high levels of 2hBuen, while low expression of GSL-OH was correlated
with high levels of GSLs upstream of the hydroxylation step (Robin et al.,
2016). We are not aware of functional studies conrming the role of
these genes in Brassica species.
In the genus Barbarea, 2hBuen is not found but two epimeric β-hy-
droxylated GSLs, BAR and EBAR, are formed. Homologs of GS-OH were
also reported from B. vulgaris (Liu et al., 2016), with a function of cor-
responding genes in hydroxylation of PE to form BAR or EBAR supported
by genetic mapping (Byrne et al., 2017). However, the map used has
later been questioned after comparison with a high-resolution map (Liu
et al., 2019). No functional studies have demonstrated the role of the
B. vulgaris GS-OH genes in GSL hydroxylation, and this should be a future
priority. From the frequent occurrence of BAR/EBAR in the tribe and
basal families, the needed hydroxylation could be either a conserved
trait or represent a recapitulation by change of substrate specicity of
GS-OH.
A coherent phylogenetic distribution of a trait suggests a shared
biochemical basis. From the deduced almost ubiquitous occurrence of
the β-hydroxylation step in the Brassicales (Fig. 5; Fig. 6), it is tempting
to suggest that all β-hydroxylation in the order is simply controlled by
one gene family that has adapted to the various available side chains in
various species. This hypothesis would explain the otherwise isolated
evolution of β-hydroxylation of a range of BCAA derived backbones in
C. pratensis after combined loss of n-homoMet derived GSLs and gain of
BCAA-derived GSLs. Understanding GSL side chain hydroxylation in
basal species of the order is needed for a general understanding of the
evolution of this apparently near-coherent trait.
The OAT formed from BAR undergoes the same metabolism (to an
oxazolidin-2-one, abbreviated OAO) (Fig. 8B) in both R. luteola,
N. ofcinale and B. vulgaris (Agerbirk et al., 2018). This apparent con-
servation of down-stream metabolism was suggested to support a single
mechanism of general β-hydroxylation conserved in the order Brassi-
cales, since conservation of the metabolic gene would only make sense if
the β-hydroxylation had also been conserved in each evolutionary line.
The heat sensitive catalyst accepted a wide structural range of OATs in
agreement with general ability to metabolize the various OATs from the
various β-hydroxylated GSLs (Fig. 5).
An even more radical suggestion would be that hydroxylation of
aliphatic backbones at other positions would also be controlled by the
same gene family, such as δ-hydroxylation in C. pratensis and the sug-
gested γ-hydroxylation in some species of Erysimum, which has been
suggested based on indirect evidence (Kjær and Schuster, 1973). The
latter hypothesis is in agreement with the presence of a possible ortholog
of GS-OH in Erysimum cheiranthoides despite lack of the conventional
product 2hBuen (Züst et al., 2020).
Alternatively, multiple specic genes catalyzing β-hydroxylation
could have evolved independently multiple times and currently coexist
(Kliebenstein and Cacho, 2016). Two recent reviews discuss the limited
evidence on GS-OH evolution (Kliebenstein and Cacho, 2016; Barco and
Clay, 2019). One concludes that the GS-OH locus is polymorphic in
multiple genera, based on literature reports of GSL diversity (Klieben-
stein and Cacho, 2016). According to personal communication with the
rst author (D. Kliebenstein), the usually proposed homology between
GSL-OH genes in Brassica and GS-OH in A. thaliana is not a straightfor-
ward interpretation. The other, a meta-analysis of publicly available
genomes, suggests that origins of GS-OH homologs in three representa-
tive species are unclear, and do not discuss GS-OH further (Barco and
Clay, 2019). According to personal communication with the rst author
(B. Barco), a possible ortholog to AtGS-OH was found in each tested
species, but relationships were unclear. We conclude that the present
Fig. 8. Biochemical aspects of aliphatic side chain oxidation of glucosinolates
(GSLs). A. Biosynthesis of three well-investigated GSLs, all involving enzymes of
the class “2-oxoglutarate-dependent dioxygenases”, although the case of BAR
biosynthesis is still tentative (Byrne et al., 2017). B. Conserved metabolism of
an OAT into the corresponding oxazolidine-2-one (OAO) in three Brassicales
species (Barbarea vulgaris, Nasturtium ofcinale and Reseda luteola). MYR, myr-
osinase; GS-OH, glucosinolate hydroxylating enzyme; GRS, glucor-
aphasatin synthase.
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
20
knowledge of the enzymology, genetics and evolution of β-hydroxyl-
ation is very limited.
Support for evolutionary plasticity of function of GS-OH came from
radish, Raphanus sativus. A gene with approximately 50% identity of
predicted amino acid sequence with that of A. thaliana GS-OH was re-
ported to be responsible for introduction of unsaturation in aliphatic
GSL biosynthesis (Fig. 8A) (Kakizaki et al., 2017), an oxidation reaction.
The relevant class of enzymes (variously named “
α
-ketoglutar-
ate-dependent mononuclear non-heme iron enzymes” or more
commonly “2-oxoglutarate-dependent dioxygenases”) are extremely
versatile catalysts responsible for a wide range of oxidations in nature
(Gao et al., 2018; Mittchell and Wang, 2019). We conclude that func-
tional studies and detailed investigations of the homology and phylog-
eny of the various enzymes oxidizing aliphatic carbons in GSL
biosynthesis are much desired. Such studies could possibly shed light on
factors controlling substrate specicity, regioselectivity and reaction
mechanism, as well as revealing the evolutionary history of GSL hy-
droxylation. In case of the tribe Cardamineae, this would include the
apparently novel hydroxylation at β as well as γ of a likewise de-novo
evolved 3-methylpentyl side chain in C. pratensis.
It seems likely that the apparent innovations in the tribe Cardami-
neae have ecological functions. Indeed, both B. vulgaris and C. pratensis
are invasive plants in the U.S.A. In C. pratensis, the two hydroxy de-
rivatives of 3mPe are expected to form different products (Olsen et al.,
2016). In Barbarea, 3hEBAR forms a more hydrophilic OAT than EBAR,
while 4hEBAR forms a completely different kind of product, a
thiazolidin-2-one (Agerbirk and Olsen, 2015). Further support of an
ecological function of hydroxylation is strong seasonal regulation of
4hEBAR (induction in fall and winter). However, so far ecological in-
vestigations gave mixed results concerning ecological effects of 2-hy-
droxylation (van Leur et al., 2008; Heimes et al., 2016; Müller et al.,
2018), while the effect of the phenol group is not tested except for a
general effect on detoxication reactions (Agerbirk et al., 2006, 2007).
Mutants and overexpression constructs developed for functional testing
of evolutionary hypotheses would be equally useful for testing ecolog-
ical and agronomic roles of GSL biosynthesis traits.
A nal novel feature in Barbarea is acylation of seed GSLs. The bio-
logical function of this coupling to a quite unusual phenolic acid, iso-
ferulic acid, is yet to be discovered. The coupling only happens late in
silique ontogeny (Heimes et al., 2016) and the resulting esters disappear
quickly during germination, suggesting associated hydrolytic enzymes
(Agerbirk and Olsen, 2011).
Suggesting specic evolutionary mechanisms at the population-
genetics level for the deduced and suggested processes would be
beyond the scope of this review. Roles of whole genome duplications at
the family level have been discussed repeatedly (e.g. van den Bergh
et al., 2016; Edger et al., 2018), but do not seem to be directly related to
the recent evolution in the tribe Cardamineae discussed here, since the
identied innovations did not show associations to recent poly-
ploidization events (Olsen et al., 2016). Hybridization can lead to
reticulate phylogenies, in which genes through their evolution can
spend time in two or more genomes. This phenomenon is common in the
family Brassicaceae including tribe Cardamineae (Marhold and Lihova,
2006). Hybridization has been characterized in detail in several species
in the tribe, including Cardamine occulta and relatives (Mand´
akov´
a et al.,
2019) and Barbarea vulgaris (Christensen et al., 2014, 2016). Interest-
ingly, the chromosomal structure seems to be rather stable in the tribe
(Mand´
akov´
a and Lysak, 2019). Characterization of GSL proles in
groups of species with known hybridization history might shed light on
an involvement of hybridization in structural GSL evolution.
3.7. Concluding remarks
We have extracted and critically assessed GSL proles from the
phytochemical literature in multiple-species comparisons. From com-
parison of the tribe Cardamineae with the remaining family and order,
and with the general literature on GSL biosynthesis, we demonstrate that
apparent recapitulation of ancient biosyntheses is frequent in the tribe
and other Brassicaceae. We also identify cases that so far would seem to
represent recent de novo evolution of GSL biosynthesis. Finally, we
conclude that both kinds of evolution, recapitulation and de-novo evo-
lution, tend to be associated with loss of GSLs derived from (homologs
of) Met in tribe Cardamineae.
In the discussion, we propose three kinds of testable hypotheses that
would explain specic observed diversity patterns. We propose, based
on examples, that ancient genes with generally low expression could
exist undetected in an evolutionary line followed by, e.g., increased
expression in a distal species, thereby causing apparent recapitulation of
ancient biosynthesis. Alternatively, back-and-forth mutation of speci-
city of biosynthetic genes over evolutionary time could explain reca-
pitulation. Finally, mutation to unprecedented activity or recruitment of
other genes could explain de novo-evolution in GSL biosynthesis.
We agree with multiple previous authors (cited in Sections 1.1. and
3.5.) that functional studies of GSL biosynthetic genes and enzymes,
mutant- and tracer studies for concluding biosynthetic pathways, as well
as phylogenetics and comparative genomics, are indispensable for
testing and further developing robust hypotheses about past GSL prole
evolution. With the present review, we have exemplied the difculties
in obtaining conclusive GSL diversity tables, as well as the usefulness of
such diversity tables for interpreting the genomic and biosynthetic data
that are rapidly appearing.
The literature review revealed the huge variation in the analytical
quality behind published GSL proles. Published GSL proles needed
detailed considerations of quality before being used for phylogenetic
comparisons. Use of authentic references and MS2 detection was found
to be indispensable quality requirements for accepting reported GSLs as
reliably identied, and future publication of screens not backed up by
standards or NMR or based on HPLC-UV is discouraged. After a pre-
liminary investigation (Olsen et al., 2016), we critically examined key
species with focus on potentially occuring GSLs, with a particular focus
of reporting both negative and positive results with analytical details
(Agerbirk et al., 2021). The combination of literature review and
experimental study has provided a reasonably dense data set for a pre-
liminary characterization of the tribe.
Even in the single tribe analyzed in detail, a large proportion of all
GSLs known to science were identied (Figs. 3 and 4) and should be
considered by future investigators, rather than the too common ap-
proaches of searching for twenty or so commercially available GSLs or
not taking into account known isomers before suggesting a structure. For
the process of creating overviews of presence and absence of selected
GSL, the analysis of the literature was haunted with problems of inter-
pretation. Uncertain reliability of positive results was a major problem,
and the usually uncertain experimental basis of negative results was an
even larger problem. These problems resulted in schemes of GSL di-
versity with many uncertainties, but still considered suitable for dis-
cussing and concluding trends in the evolutionary history of GSL
biosynthesis.
In a phylogenetic perspective, GSLs were for simplicity divided into
three sets (Fig. 5). The rst and ancient set were those more or less
widely occurring in basal families in the order Brassicales. In contrast to
some other authors, we included Trp-derived GSLs in this set based on
two literature reports. The second “intermediate” set, not known from
basal families, were apparently derived GSLs commonly found in three
derived families; Capparaceae, Cleomaceae and Brassicaceae. This set
was dominated by GSLs derived from chain elongated Met (n-homoMet),
but also contained GSLs derived from homoIle and the 4-substituted Trp-
derived GSLs. The third “recently evolved” set was composed of a
considerable number of GSLs apparently conned to one or a few genera
within the family Brassicaceae. This set includes 2homoIle derived and
numerous substituted GSLs derived from Phe, homoPhe, Trp and
branched chain amino acids. A similar logic suggested
ω
-hydrox-
yalkylGSLs from A. thaliana, biosynthesized from 1-2homoMet, to be
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
21
recently evolved.
Even in the tribe Brassiceae, a wider variety of the ancient GSLs
occurs than often appreciated, sometimes at very low levels. Although n-
homoMet-derived GSLs often dominate in the family Brassicaceae, there
is evidence to suggest that they have been lost occasionally. In those
cases, both GSLs from both the ancient and the recently evolved sets can
accumulate at high levels, equivalent to the levels of the lost n-homo-
Met-derived GSLs. Obviously, both structural and quantitative changes
have been factors in the evolution of GSLs in the family Brassicaceae,
exemplied by the tribes Cardamineae, Arabideae and Brassiceae.
In terms of GSL diversity, the tribe Cardamineae would seem to be
among the best-characterized phylogenetic groups of the Brassicaceae
and contains three genetically well-characterized species, B. vulgaris
(Byrne et al., 2017; Liu et al., 2019), C. hirsuta (Hay et al., 2014; Gan
et al., 2016) and N. ofcinale (Voutsina et al., 2016; Mand´
akov´
a and
Lysak, 2019). Similar characterization of C. pratensis would seem
particularly relevant. We invite biologists from all relevant disciplines to
join us in continued investigation of the impressive and
well-characterized GSL diversity in this tribe.
4. Experimental
4.1. Phylogeny reconstructions
A tribal-wide analysis was performed for Cardamineae to provide the
respective background to discuss the occurrence of the various and
diverse GSLs. Taxon sampling was comprehensive and comprised 171
taxa out of the total 386 species (44%) belonging to the tribe according
to Brassibase (Koch et al., 2017), two outgroup species (Brassica napus L.
and Clausia aprica (Stephan ex Willd.) Korn-Trotsky) and the newly
sequenced Barbarea accessions. Taxon sampling and outgroup selection
followed Huang et al. (2020). The alignment (comprising the internal
transcribed spacer 1 and 2 of nuclear encoded ribosomal RNA and the
intervening 5.8 S rDNA gene) had a total length of 680 bp (Supple-
mentary Table S2). In a rst step GTR +I +G was determined as the best
molecular evolutionary model using PartitionFinder2 (Lanfear et al.,
2017). Phylogenetic analysis resulting in a Maximum Likelihood to-
pology was then performed using RAxML-ng (Kozlov et al., 2019),
setting the molecular evolutionary model to GTR +I +G, setting branch
lengths to linked and starting the analysis from 20 parsimony trees
generated within RAxML-ng. Number of bootstrap replicates was set to
1000Bootstrap support above 50% are indicated on the branches of the
Maximum Likelihood tree resulting from the RAxML analysis. GenBank
accession codes are given with species names (Supplementary Fig. S1).
For a cartoon-like presentation of phylogenetic relationships among
taxa within Brassicaceae for which relevant information on GSLs is
available, we also conducted a Bayesian phylogenetic analysis focusing
on 27 accessions from Brassicaceae and using Reseda (Resedaceae) as
outgroup. The alignment had a length of 674 bp and is shown in Sup-
plementary Table S3. The best molecular evolutionary model was
selected using PartitionFinder2 (Lanfear et al., 2017). Phylogenetic
analysis was done using the program MrBayes (Huelsenbeck and Ron-
quist 2001; Ronquist and Huelsenbeck 2003). The molecular evolu-
tionary model was set to GTR +I +G and four runs with 5,000,000
generations were completed. Sampling frequency was set to 5000 and
temperature was set to 0.01. The four runs were inspected for quality in
Tracer1.7.1 (Rambaut et al., 2018). The rst 25% of the sampled trees
were discarded as burn in using MrBayes’ sumt and sump functions.
4.2. Compiling glucosinolate diversity papers on the tribe Cardamineae
We searched for GSL prole data for all currently accepted and some
formerly used genera of the tribe (as listed in Table 2 incl. notes) in Web
of Sciences by searching for “[genus name] AND glucosinolate*” as
topic. We scanned abstracts and when relevant full texts to reveal
whether signicant analytical data were included. We generally
included papers with relevant analytical data, but in a few cases we
excluded papers that we regarded as not focused on providing novel GSL
prole data but mainly on reporting variation in the levels of well-
established GSLs, typically as a function of a manipulated variable.
The latter group of papers not included mainly concerned HPLC-UV
analysis of major GSLs in watercress and some early papers on
C. cordifolia.
The described search strategy did not reveal all relevant papers. For
example, the highly relevant screens by Daxenbichler et al. (1991),
Bennett et al. (2004a) and Windsor et al. (2005) were not revealed.
Hence, whenever we were aware of additional relevant literature, it was
included. We also searched literature in less systematic ways and
inspected relevant reference lists and reviews, e.g. Blaˇ
zevi´
c et al. (2020).
4.3. Compiling glucosinolate diversity papers on other species
Based on previous knowledge of reliable analytical studies, we
selected the included twelve species (Section 2.3.) of Brassicaceae and
searched for additional high-quality analytical studies in Web of Sci-
ence, using “[species name] AND glucosinolate*” as topic. For the genus
Reseda, we mainly relied on the reviews presented by Blaˇ
zevi´
c et al.
(2017) and Pagnotta et al. (2020), but in each case checking the original
report. For the overview of GSLs known from ancestral families, we
mainly relied on Daxenbichler et al. (1991), Fahey et al. (2001), Mithen
et al. (2010) and Blaˇ
zevi´
c et al. (2017), in each case checking the
original reports and taking into account more recent changes in botan-
ical nomenclature.
4.4. Compiling literature data for Figs. 6C and 7A
For concluding absence of specic GSLs, we applied our own sys-
tematic searches as well as interpretation of literature screens (Supple-
mentary Table S1). The GSLs concluded to be present in this
supplementary table were indicated as such in Fig. 5, while those
tentatively accepted from high quality evidence were illustrated with
the category “Circumstantial evidence”. Aliphatic GSLs only suggested
based on HPLC-UV, even including comparison with a limited number of
authentic standards, were not accepted for this category, as this evi-
dence is of too low reliability. We listed as “tested, not reported” those
explicitly searched for and not found in own papers, but also some that
from a holistic inspection of the literature was concluded to be “tested,
not reported”, i.e. if the same author had reported the GSL in question
from another species using the same method. When the appropriate
organ was not included (e.g. seeds for 6′-iF conjugated GSLs and roots
for substituted Trp-derived GSLs), or when the GSL was not available as
a standard (e.g. the dihomoIle-derived 58, 141 and 149 by Daxenbichler
et al., 1991) or the type of GSL probably not detectable or distinguish-
able with the used method (e.g. lack of distinction of 4hBZ and
Trp-derived GSLs by Daxenbichler et al., 1991), we concluded that data
were “insufcient or missing”. For our own data, published before the
discovery of 141 and 149 but using the same methods, we also deemed it
correct to conclude those to be “tested, not reported”.
Presence or absence of deduced biosynthetic characters (Fig. 6A) was
concluded similarly, in this case considering all GSLs of a particular
structural type as indicated in the gure (Fig. 6B and C). As ring-
methoxy groups are conclusively reported from P-type B. vulgaris (in
x1 and x2, Fig. 3), the P-type was concluded to be positive for the
character “Methox-Ph”. As hydroxylation at β was detectable for a
6homoMet-derived GSL in R. amphibia, this character was concluded to
be tested, not reported for R. sylvestris containing the relevant precursor,
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
22
8mSOo (69).
When citing papers that did not distinguish diastereomers, notably
Daxenbichler et al. (1991), we merely indicated the number (e.g. 40
means either 40R or 40S or both, not specied).
Author contributions
Conceived study: NA. ITS sequencing and phylogeny: MAK, CK.
Reviewed GSL diversity literature: NA. Wrote manuscript draft: NA,
CCH, MAK. Discussions and nal writing: All.
Declaration of competing interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
Acknowledgements
We thank M. E. Schranz, B. Barco, N. Clay, P. Velasco Pazos, B. L.
Møller and D. Kliebenstein for stimulating discussions and/or corre-
spondence, and ve anonymous reviewers for numerous helpful com-
ments to earlier versions of this text, comments that signicantly
improved the paper.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.phytochem.2021.112668.
References
Abrahams, R.S., Pires, J.C., Schranz, M.E., 2020. Genomic origin and diversication of
the glucosinolate MAM locus. Front. Plant Sci. 11, 711. https://doi.org/10.3389/
fpls.2020.00711.
Agerbirk, N., Petersen, B.L., Olsen, C.E., Halkier, B.A., Nielsen, J.K., 2001. 1,4-
Dimethoxyglucobrassicin in Barbarea and 4-hydroxyglucobrassicin in Arabidopsis
and Brassica. J. Agric. Food Chem. 49, 1502–1507. https://doi.org/10.1021/
jf001256r.
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. Systemat. Ecol. 34, 189–198. https://doi.org/10.1016/j.
bse.2005.09.005.
Agerbirk, N., Olsen, C.E., Topbjerg, H.B., Sørensen, J.C., 2007. Host-plant dependent
metabolism of 4-hydroxybenzylglucosinolate in Pieris rapae: substrate specicity and
effects of genetic modication and plant nitrile hydratase. Insect Biochem. Mol. Biol.
37, 1119–1130. https://doi.org/10.1016/j.ibmb.2007.06.009.
Agerbirk, N., Warwick, S., Hansen, P.R., Olsen, C.E., 2008. Sinapis phylogeny and
evolution of glucosinolates and specic nitrile degrading enzymes. Phytochemistry
69, 2937–2949. https://doi.org/10.1016/j.phytochem.2008.08.014.
Agerbirk, N., Olsen, C.E., Chew, F., Ørgaard, M., 2010a. Variable glucosinolate proles of
Cardamine pratensis (Brassicaceae) with equal chromosome numbers. J. Agric. Food
Chem. 58, 4693–4700. https://doi.org/10.1021/jf904362m.
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. https://doi.org/
10.1016/j.ibmb.2010.01.003.
Agerbirk, N., Olsen, C.E., 2011. Isoferuloyl derivatives of ve seed glucosinolates in the
crucifer genus Barbarea. Phytochemistry 72, 610–623. https://doi.org/10.1016/j.
phytochem.2011.01.034.
Agerbirk, N., Olsen, C.E., 2012. Glucosinolate structures in evolution. Phytochemistry
77, 16–45. https://doi.org/10.1016/j.phytochem.2012.02.005.
Agerbirk, N., Olsen, C.E., Cipollini, D., Ørgaard, M., Linde-Laursen, I., Chew, F.C., 2014.
Specic glucosinolate analysis reveals variable levels of glucobarbarins, dietary
precursors of 5-phenyloxazolidine-2-thiones, in watercress types with contrasting
chromosome number. J. Agric. Food Chem. 62, 9586–9596. https://doi.org/
10.1021/jf5032795.
Agerbirk, N., Olsen, C.E., 2015. Glucosinolate hydrolysis products in the crucifer
Barbarea vulgaris include a thiazolidine-2-one from a specic phenolic isomer as well
as oxazolidine-2-thiones. Phytochemistry 115, 143–151. https://doi.org/10.1016/j.
phytochem.2014.11.002.
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. https://doi.org/10.1016/j.
phytochem.2014.09.003.
Agerbirk, N., Matthes, A., Erthmann, P.Ø., Ugolini, L., Cinti, S., Lazaridi, E., Nuzillard, J.-
M., Müller, C., Bak, S., Rollin, P., Lazzeri, L., 2018. Glucosinolate turnover in
Brassicales species to an oxazolidin-2-one, formed via the 2-thione and without
formation of thioamide. Phytochemistry 153, 79–83. https://doi.org/10.1016/j.
phytochem.2018.05.006.
Agerbirk, N., Hansen, C.C., Olsen, C.E., Kiefer, C., Hauser, T.P., Christensen, S.,
Jensen, K.R., Ørgaard, M., Pattison, D.I., Lange, C.B.A., Cipollini, D., Koch, M.A.,
2021. Glucosinolate Proles and Phylogeny in Barbarea Comared to Other Tribe
Cardamineae (Brassicaceae) and Reseda (Resedaceae), Based on a Library of Ion Trap
HPLC-MS/MS Data of Reference Desulfoglucosinolates. Phytochemistry in press.
Agneta, R., Lelario, F., De Maria, S., M¨
ollers, C., Bufo, S.A., Rivelli, A.R., 2014.
Glucosinolate prole and distribution among plant tissues and phenological stages of
eld-grown horseradish. Phytochemistry 106, 178–187. https://doi.org/10.1016/j.
phytochem.2014.06.019.
Alseekh, S., de Souza, L.P., Benina, M., Fernie, A.R., 2020. The style and substance of
plant avonoid decoration: towards dening both structure and function.
Phytochemistry 174, 112347. https://doi.org/10.1016/j.phytochem.2020.112347.
Andini, S., Dekker, P., Gruppen, H., Araya-Cloutier, C., Vincken, J.-P., 2019. Modulation
of glucosinolate composition in Brassicaceae seeds by germination andf fungal
elicitation. J. Agric. Food Chem. 67, 12770–12779. https://doi.org/10.1021/acs.
jafc.9b05771.
Augustine, R., Bisht, N., 2017. Regulation of glucosinolate metabolism: from model plant
Arabidopsis thaliana to Brassica crops. In: M´
erillon, J.-M., Ramawat, K.G. (Eds.),
Glucosinolates. Ref. Ser. Phytochem, pp. 163–199. https://doi.org/10.1007/978-3-
319-25462-3_3.
Badenes-P´
erez, F.R., Gershenzon, J., Heckel, D.G., 2020. Plant glucosinolate content
increases susceptibility to diamondback moth (Lepidoptera: plutellidae) regardsless
of its diet. J. Pest. Sci. 93, 491–506. https://doi.org/10.1007/s101340-019-01139-z.
Baek, S.-A., Jung, Y.-H., Lim, S.-H., Park, S.U., Kim, J.K., 2016. Metabolic proling in
Chinese cabbage (Brassica rapa L. supsp. pekinensis) cultivars reveals that
glucosinolate content is correlated with carotenoid content. J. Agric. Food Chem. 64,
4426–4434. https://doi.org/10.1021/acs.jafc.6b01323.
Bak, S., Nielsen, H.L., Halkier, B.A., 1998. The presence of CYP79 homologues in
glucosinolate-producing plants shows evolutionary conservation of the enzymes in
the conversion of amino acid to aldoxime in the biosynthesis of cyanogenic
glucosides and glucosinolates. Plant Mol. Biol. 38, 725–734. https://doi.org/
10.1023/A:1006064202774.
Bak, S., Olsen, C.E., Petersen, B.L., Møller, B.L., Halkier, B.A., 1999. Metabolic
engineering of p-hydroxybenzylglucosinolate in Arabidopsis by expression of the
cyanogenic CYP79A1 from Sorghum bicolor. Plant J. 20, 663–671. https://doi.org/
10.1046/j.1365-313x.1999.00642.x.
Bakhtiari, M., Glauser, G., Rasmann, S., 2018. Root JA induction modies glucosinolate
proles and increases subsequent aboveground resistance to herbivore attack in
Cardamine hirsuta. Front. Plant Sci. 9, 1230. https://doi.org/10.3389/
fpls.2018.01230.
Bakhtiari, M., Formenti, L., Caggia, V., Glauser, G., 2019. Variable effects on growth and
defense traits for plant ecotypic differentiation and phenotypic plasticity along
elevation gradients. Ecol. Evol. 9, 3740–3755. https://doi.org/10.1002/ece3.4999.
Barco, B., Clay, N.K., 2019. Evolution of glucosinolate diversity via whole-genome
duplications, gene rearrangements, and substrate promiscuity. Annu. Rev. Plant Biol.
70, 585–604. https://doi.org/10.1146/annurev-arplant-050718-100152.
Benderoth, M., Textor, S., Windsor, A.J., Mitchell-Olds, T., Gershenzon, J., Kroymann, J.,
2006. Positive selection driving diversication in plant secondary metabolism. Proc.
Natl. Acad. Sci. U.S.A. 103, 9118–9123. https://doi.org/10.1073/pnas.0601738103.
Bennett, R.N., Mellon, F.A., Kroon, P.A., 2004a. Screening crucifer seeds as sources of
specic intact glucosinolates using ion-pair high-performance liquid
chromatography negative ion electrospray mass spectrometry. J. Agric. Food Chem.
52, 428–438. https://doi.org/10.1021/jf030530p.
Bennett, R.N., Mellon, F.A., Rosa, E.A.S., Perkins, L., Kroon, P.A., 2004b. Proling
glucosinolates, avonoids, alkaloids, and other secondary metabolites in tissues of
Azima tetracantha L. (Salvadoraceae). J. Agric. Food Chem. 52, 5856–5862. https://
doi.org/10.1021/jf040091+.
Beran, F., Sporer, T., Paetz, C., Ahn, S.J., Betzin, F., Kunert, G., Shekhov, A., Vass˜
ao, D.G.,
Bartram, S., Lorenz, S., Reichelt, M., 2018. One pathway is not enough: the cabbage
stem ea beetle Psylliodes chrysocephala uses multiple strategies to overcome the
glucosinolate-myrosinase defense in its host plants. Front. Plant Sci. 9, 1754. https://
doi.org/10.3389/fpls.2018.01754.
Bianco, G., Agerbirk, N., Losito, I., Cataldi, T.R.I., 2014. Acylated glucosinolates with
diverse acyl groups investigated by high resolution mass spectrometry and infrared
multiphoton dissociation. Phytochemistry 100, 92–102. https://doi.org/10.1016/j.
phytochem.2014.01.010.
Blaˇ
zevi´
c, I., Montaut, S., Burˇ
cul, F., Rollin, P., 2017. Glucosinolates: novel sources and
biological potential. In: M´
erillon, J.-M., Ramawat, K.G. (Eds.), Glucosinolates.
Ref. Ser. Phytochem, pp. 3–60. https://doi.org/10.1007/978-3-319-26479-0_1-1.
Blaˇ
zevi´
c, I., Montaut, S., Burˇ
cul, F., Olsen, C.E., Burow, M., Rollin, P., Agerbirk, N., 2020.
Glucosinolate structural diversity, identication, chemical synthesis and metabolism
in plants. Phytochemistry 168, 112100. https://doi.org/10.1016/j.
phytochem.2019.112100.
Bor, M., Ozkur, O., Ozdemir, F., Turkan, I., 2009. Identication and characterization of
the glucosinolate-myrosinase system in caper (Capparis ovata Desf.). Plant Mol. Biol.
Rep. 27, 518–525. https://doi.org/10.1007/s11105-009-0117-0.
Brown, P.D., Tokuhisa, J.G., Reichelt, M., Gershenzon, J., 2003. Variation of
glucosinolate accumulation among different organs and developmental stages of
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
23
Arabidopsis thaliana. Phytochemistry 62, 471–481. https://doi.org/10.1016/s0031-
9422(02)00549-6.
Byrne, S., Erthmann, P.Ø., Agerbirk, N., Bak, S., Hauser, T.P., Nagy, I., Paina, C., Asp, T.,
2017. The genome sequence of Barbarea vulgaris facilitates the study of ecological
biochemistry. Sci. Rep. 7, 40728. https://doi.org/10.1038/srep40728.
Cang, W., Sheng, Y.-X., Evivie, E.R., Kong, W.-W., Li, J., 2018. Lineage-specic evolution
of avin-containing monooxygenases involved in aliphatic glucosinolate side-chain
modication. J. Systemat. Evol. 56, 92–104. https://doi.org/10.1111/jse.12289.
Cannell, N., Emms, D.M., Hetherington, A.J., MacKay, J., Kelly, S., Dolan, L.,
Sweetlove, L.J., 2020. Multiple metabolic innovations and losses are associated with
major transitions in land plant evolution. Curr. Biol. 30, 1783–1800. https://doi.
org/10.1016/j.cub.2020.02.086.
C´
ardenas, P.D., Almeida, A., Bak, S., 2020. Evolution of structural diversity of terpenoids.
Front. Plant Sci. 10, 1523. https://doi.org/10.3389/fpls.2019.01523.
Chen, J., Ullah, C., Reichelt, M., Beran, F., yang, Z.-L., Gershenzon, J.,
Hammerbacher, A., Vass˜
ao, D.G., 2020. The phytopathogenic fungus Sclerotina
sclerotiorum detoxies plant glucosinolate hydrolysis products via an isothiocyanate
hydrolase. Nat. Commun. 11, 3090. https://doi.org/10.1038/s41467-020-16921-2.
Chhajed, S., Misra, B.B., Tello, N., Chen, S.X., 2019. Chemodiversity of the glucosinolate-
myrosinase system at the single cell type resolution. Front. Plant Sci. 10, 618.
https://doi.org/10.3389/fpls.2019.00618.
Chhajed, S., Mostafa, I., He, Y., Abou-Hashem, M., El-Domiaty, M., Chen, S., 2020.
Glucosinolate biosynthesis and the glucosinolate-myrosinase system in plant
defense. Agronomy 10, 1786. https://doi.org/10.3390/agronomy10111786.
Christensen, S., Heimes, C., Agerbirk, N., Kuzina, V., Olsen, C.E., Hauser, T.P., 2014.
Different geographical distributions of two genotypes of Barbarea vulgaris that differ
in resistance to insects and a pathogen. J. Chem. Ecol. 40, 491–501. https://doi.org/
10.1007/s10886-014-0430-4.
Christensen, S., Sørensen, H., Munk, K.R., Hauser, T.P., 2016. A hybridisation barrier
between two evolutionary lineages of Barbarea vulgaris (Brassicaceae) that differ in
biotic resistances. Evol. Ecol. 30, 887–904. https://doi.org/10.1007/s10682-016-
9858-z.
Ciska, E., Horbovicz, M., Rogowska, M., Kosson, R., Drabi´
nska, N., Honke, J., 2017.
Evaluation of seasonal variations in the glucosinolate content in leaves and roots of
four European horseradish (Armoracia rusticana) landraces. Pol. J. Food Nutr. Sci.
67, 301–308. https://doi.org/10.1515/pjfns-2016-0029.
Clausen, M., Kannangara, R., Olsen, C.E., Blomstedt, C.K., Gleadow, R.M., Jørgensen, K.,
Bak, S., Motawie, M.S., Møller, B.L., 2015. The bifurcation of the cyanogenic
glucoside and glucosinolate biosynthetic pathways. Plant J. 84, 558–573. https://
doi.org/10.1111/tpj.13023.
Cole, R.A., 1976. Isothiocyanates, nitriles and thiocyanates as products of autolysis of
glucosinolates in Cruciferae. Phytochemistry 15, 759–762. https://doi.org/10.1016/
S0031-9422(00)94437-6.
Cuong, D.M., Park, C.H., Bong, S.J., Kim, N.S., Kim, J.K., Park, S.U., 2019. Enhancement
of glucosinolate production in watercress (Nasturtium ofcinale) hairy roots by
overexpressing cabbage transcription factors. J. Agric. Food Chem. 67, 4860–4867.
https://doi.org/10.1021/acs.jafc.9b00440.
Czerniawski, P., Piasecka, A., Bednarek, P., 2021. Evolutionary changes in the
glucosinolate biosynthetic capacity in species representing Capsella, Camelina and
Neslia genera. Phytochemistry 181, 112571. https://doi.org/10.1016/j.
phytochem.2020.112571.
Daxenbichler, M.E., VanEtten, C.H., 1974. 5,5-Dimethyloxazolidine-2-thione formation
from glucosinolate in Limnanthes alba Benth. seed. J. Am. Oil Chem. Soc. 51,
449–450. https://doi.org/10.1007/BF02635152.
Daxenbichler, M.E., Spencer, G.F., Carlson, D.G., Rose, G.B., Brinker, A.M., Powell, R.G.,
1991. Glucosinolate composition of seeds from 297 species of wild plants.
Phytochemistry 30, 2623–2638. https://doi.org/10.1016/0031-9422(91)85112-D.
Deki´
c, M.S., Radulovic, N.S., Stojanovic, N.M., Randjelovic, P.J., Stojanovic-Radic, Z.,
Najman, S., Stojanovic, S., 2017. Spasmolytic, antimicrobial and cytotoxic activities
of 5-phenylpentyl isothiocyanate, a new glucosinolate autolysis product from
horseradish (Armoracia rusticana P. Gaertn., B. Mey. & Scherb., Brassicaceae). Food
Chem. 232, 329–339. https://doi.org/10.1016/j.foodchem.2017.03.150.
Dunning, L.T., Christin, P.-A., 2020. Reticulate evolution, lateral gene transfer, and
innovation in plants. Am. J. Bot. 107, 541–544. https://doi.org/10.1002/ajb2.1452.
Edger, P.P., Heidel-Fischer, H.M., Bekaert, M., Rota, J., Gloeckner, G., Platts, A.E.,
Heckel, D., Der, J.P., Wafula, E.K., Tang, M., Hofberger, J.A., Smithson, A., Hall, J.C.,
Blanchette, M., Bureau, T.E., Wright, S.I., dePamphilis, C.W., Schranz, M.E.,
Barker, M.S., Conant, G.C., Wahlberg, N., Vogel, H., Pires, J.C., Wheat, C.W., 2015.
The buttery plant arms-race escalated by gene and genome duplications. Proc. Natl.
Acad. Sci. U.S.A. 112, 8362–8366. https://doi.org/10.1073/pnas.1503926112.
Edger, P.P., Hall, J.C., Harkess, A., Tang, M., Coombs, J., Mohammadin, S., Schranz, M.
E., Xiong, Z., Leebens-Mack, J., Meyers, B.C., Sytsma, K.J., Koch, M.A., Al-Shehbaz, I.
A., Pires, J.C., 2018. Brassicales phylogeny inferred from 72 plastid genes: a
reanalysis of the phylogenetic localization of two paleopolyploid events and origin of
novel chemical defenses. Am. J. Bot. 105, 463–469. https://doi.org/10.1002/
ajb2.1040.
Fahey, J.W., Zalcmann, A.T., Talalay, P., 2001. The chemical diversity and distribution of
glucosinolates and isothiocyanates among plants. Phytochemistry 56, 5–51. https://
doi.org/10.1016/s0031-9422(00)00316-2.
Friedrichs, J., Schweiger, R., Geisler, S., Mix, A., Wittstock, U., Müller, C., 2020. Novel
glucosinolate metabolism in larvae of the leaf beetle Phaedon cochleariae. Insect
Biochem. Mol. Biol. 124, 103431. https://doi.org/10.1016/sj.ibmb.2020.103431.
Gan, X., Hay, A., Kwantes, M., Haberer, G., Hallab, A., Ioio, R.D., Hofhuis, H., Pieper, B.,
Cartolano, M., Neumann, U., Nikolov, L.A., Song, B., Hajheidari, M., Briskine, R.,
Kougioumoutzi, E., Vlad, D., Broholm, S., Hein, J., Meksem, K., Lightfoot, D.,
Shimizu, K.K., Shimizu-Inatsugi, R., Imprialou, M., Kudrna, D., Wing, R., Sato, S.,
Huijser, P., Filatov, D., Mayer, K.F.X., Mott, R., Tsiantis, M., 2016. The Cardamine
hirsuta genome offers insight into the evolution of morphological diversity. Nature
Plants 2, 16167. https://doi.org/10.1038/nplants.2016.167.
Gao, S.-S., Naowarojna, N., Cheng, R., Liu, X., Liu, P., 2018. Recent examples of
α
-ketoglutarate-dependent mononuclear non-haem iron enzymes in natural product
biosyntheses. Nat. Prod. Rep. 35, 792–837. https://doi.org/10.1039/C7NP00067G.
Giallourou, N., Oruna-Concha, M.J., Harbourne, N., 2016. Effects of domestic processing
methods on the phytochemical content of watercress (Nasturtium ofcinale). Food
Chem. 212, 411–419. https://doi.org/10.1016/j.foodchem.2016.05.190.
Gmelin, R., Kjær, A., Schuster, A., 1970. 2-Hydroxy-2-methylpropyl glucosinolate in
Reseda alba. Phytochemistry 9, 599–600. https://doi.org/10.1016/S0031-9422(00)
85698-8.
Grifths, D.W., Deighton, N., Birch, A.N.E., Patrian, B., Baur, R., St¨
adler, E., 2001.
Identication of glucosinolates on the leaf surface of plants from the Cruciferae and
other closely related plants. Phytochemistry 57, 693–700. https://doi.org/10.1016/
s0031-9422(01)00138-8.
Grob Jr., K., Matile, P., 1980. Capillary GC of glucosinolate-derived horseradish
constituents. Phytochemistry 19, 1789–1793. https://doi.org/10.1016/S0031-9422
(00)83814-5.
Halkier, B.A., Gershenzon, J., 2006. Biology and biochemistry of glucosinolates. Annu.
Rev. Plant Biol. 57, 303–333. https://doi.org/10.1146/annurev.
arplant.57.032905.105228.
Hanschen, F.S., Schreiner, M., 2017. Isothiocyanates, nitriles, and epithionitriles from
glucosinolates are affected by genotype and developmental stage in Brassica oleracea
varieties. Front. Plant Sci. 8, 1095. https://doi.org/10.1016/S0031-9422(00)83814-5.
Hansen, C.H., Du, L., Naur, P., Olsen, C.E., Axelsen, K.B., Hick, A.J., Pickett, J.A.,
Halkier, B.A., 2001. CYP83B1 is the oxime-metabolizing enzyme in the glucosinolate
pathway in Arabidopsis. J. Biol. Chem. 276, 24790–24796. https://doi.org/10.1074/
jbc.M102637200.
Hansen, B.G., Kerwin, R.E., Ober, J.A., Lambrix, V.M., Mitchell-Olds, T., Gershenzon, J.,
Halkier, B.A., Kliebenstein, D.J., 2008. A novel 2-oxoacid-dependent dioxygenase
involved in the formation of the goiterogenic 2-hydroxybut-3-enyl glucosinolate and
generalist insect resistance in Arabidopsis. Plant Physiol. 148, 2096–2108. https://
doi.org/10.1104/pp.108.129981.
Hansen, C.C., Sørensen, M., Veiga, T.A.M., Zibrandtsen, J.F.S., Heskes, A.M., Olsen, C.E.,
Boughton, B.A., Møller, B.L., Neilson, E.H.J., 2018. Recongured cyanogenic
glucoside biosynthesis in Eucalyptus cladocalyx involves a cytochrome P450
CYP706C55. Plant Physiol. (Wash. D C) 178, 1081–1095. https://doi.org/10.1104/
pp.18.00998.
Harun, S., Rehda, Abdullah-Zawawi, M.-R., Goh, H.-H., Mohamed-Hussein, Z.-A., 2020.
A comprehensive gene inventory for glucosinolate biosynthetic pathway in
Arabidopsis thaliana. J. Agric. Food Chem. 68, 7281–7297. https://doi.org/10.1104/
pp.18.00998.
Hay, A.S., Pieper, B., Cooke, E., Mand´
akova, T., Cartolano, M., Tattersall, A.D., Ioio, R.D.,
McGowan, S.J., Barkoulas, M., Galinha, C., Rast, M.I., Hofhuis, H., Then, C.,
Plieske, J., Ganal, M., Mott, R., Martinez-Garcia, J.F., Carine, M.A., Scotland, R.W.,
Gan, X., Filatov, D.A., Lysak, M.A., Tsiantis, M., 2014. Cardamine hirsuta: a versatile
genetic system for comparative studies. Plant J. 78, 1–15. https://doi.org/10.1111/
tpj.12447.
Heimes, C., Agerbirk, N., Sørensen, H., van M¨
olken, T., Hauser, T.P., 2016. Ecotopic
differentiation of two sympatric chemotypes of Barbarea vulgaris (Brassicaceae) with
different biotic resistances. Plant Ecol. 217, 1055–1068. https://doi.org/10.1007/
s11258-016-0631-8.
Hofberger, J.A., Lyons, E., Edger, P.P., Pires, J.C., Schranz, M.E., 2013. Whole genome
and tandem duplicate retention facilitated glucosinolate pathway diversication in
the mustard family. Genome Biol. Evol. 5, 2155–2173. https://doi.org/10.1093/
gbe/evt162.
Hohmann, N., Wold, E., Lysak, M., Koch, M.A., 2015. A time-calibrated road map of
Brassicaceae species radiation and evolutionary history. Plant Cell 27, 2770–2784.
https://doi.org/10.1105/tpc.15.00482.
Huang, R., O’Donnell, A.J., Barboline, J.J., Barkman, T.J., 2016. Convergent evolution of
caffeine in plants by co-option of exapted ancestral enzymes. Proc. Natl. Acad. Sci. U.
S.A. 113, 10613–10618. https://doi.org/10.1073/pnas.1602575113.
Huang, X.C., German, D., Koch, M.A., 2020. Temporal patterns of diversication in
Brassicaceae demonstrate decoupling of rate shifts and mesopolyploidization events.
Ann. Bot. 125, 29–47. https://doi.org/10.1093/aob/mcz123.
Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of phylogeny.
Bioinformatics 17, 754–755. https://doi.org/10.1093/bioinformatics/17.8.754.
Humphrey, P.T., Gloss, A.D., Alexandre, N.M., Villalobos, M.M., Fremgen, M.R.,
Groen, S.C., Meihls, L.N., Jander, G., Whiteman, N.K., 2016. Aversion and attraction
to harmful plant secondary compounds jointly shape the foraging ecology of a
specialist herbivore. Ecol. Evol. 6, 3256–3268. https://doi.org/10.1002/ece3.2082.
Hussain, M., Gao, J., Bano, S., Wang, L., Lin, Y., Arthurs, S., Qasim, M., Mao, R., 2020.
Diamondback moth larvae trigger host plant volatiles that lure its adult females for
oviposition. Insects 11, 725. https://doi.org/10.3390/insects11110725.
Irmisch, S., McCormick, A.C., Boeckler, G.A., Schmidt, A., Reichelt, M., Schneider, B.,
Block, K., Schnitzler, J.P., Gershenzon, J., Unsicker, S.B., Kollner, T.G., 2013. Two
herbivore-induced cutochrome P45o enzymes CYP79D6 and CYP79D7 catalyze the
formation of volatile aldoximes involved in poplar defense. Plant Cell 25,
4737–4754. https://doi.org/10.1105/tpc.113.118265.
Jeon, J., Bong, S.J., Park, J.S., Park, Y.-K., Arasu, M.V., Al-Dhabi, N.A., Park, S.U., 2017.
De novo transcriptome analysis and glucosinolate proling in watercress (Nasturtium
ofcinale R. Br.). BMC Genom. 18, 401. https://doi.org/10.1186/s12864-017-3792-
5.
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
24
Jeschke, V., Gershenzon, J., Vass˜
ao, D.G., 2016. Insect detoxication of glucosinolates
and their hydrolysis products. Adv. Bot. Res. 80, 199–245. https://doi.org/10.1016/
bs.abr.2016.06.003.
Jørgensen, M.E., Xu, D., Crocoll, C., Ramírez, D., Motawia, M.S., Olsen, C.E., Nour-
Eldin, H.H., Halkier, B.A., 2017. Origin and evolution of transporter substrate
specicity within the NPF family.
Kakizaki, T., Kitashiba, H., Zou, Z., Li, F., Fukino, N., Ohara, T., Nishio, T., Ishida, M.,
2017. A 2-oxoglutarate-dependent dioxygenase mediates the biosynthesis of
glucoraphasatin in radish. Plant Physiol. 173, 1583–1593. https://doi.org/10.1104/
pp.16.01814.
Kittipol, V., He, Z., Wang, L., Doheny-Adams, T., Langer, S., Bancroft, I., 2019. Genetic
architecture of glucosinolate variation in Brassica napus. J. Plant Physiol. 240,
152988. https://doi.org/10.1016/j.jplph.2019.06.001.
Kjær, A., Gmelin, R., 1958. Isothiocyanates XXXIII. An isothiocyanate glucoside
(glucobarbarin) of Reseda lueteola L. Acta Chem. Scand. 12, 1693–1694. https://doi.
org/10.3891/acta.chem.scand.12-1693.
Kjær, A., Friis, P., 1962. Isothiocyanates XLIII. Isothiocyanates from Putranjiva roxburghii
Wall. including (S)-2-methylbutyl isothiocyanate, a new mustard oil of natural
derivation. Acta Chem. Scand. 16, 936–946. https://doi.org/10.3891/acta.chem.
scand.16-0936.
Kjær, A., Schuster, A., 1972. Glucosinolates in seeds of Arabis hirsuta (L.) Scop.: some
new, naturally derived isothiocyanates. Acta Chem. Scand. 26, 8–14. https://doi.
org/10.3891/acta.chem.scand.26-0008.
Kjær, A., Schuster, A., 1973.
ω
-Methylthioalkylglucosinolates and some oxidized
congeners in seeds of Erysimum rhaeticum. Phytochemistry 12, 929–933. https://doi.
org/10.1016/0031-9422(73)80705-8.
Kjær, A., Thomsen, H., 1962. Isothiocyanates XLII. Glucocleomin, a new natural
glucoside, furnishing (-)-5-ethyl-5-methyl-2-oxazolidinethione on enzymic
hydrolysis. Acta Chem. Scand. 16, 591–598. https://doi.org/10.3891/acta.chem.
scand.16-0591.
Kliebenstein, D.J., D’Auria, J.C., Behere, A.S., Kim, J.H., Gunderson, K.L., Breen, J.N.,
Lee, G., Gershenzon, J., Last, R.L., Jander, G., 2007. Characterization of seed-specic
benzoyloxyglucosinolate mutations in Arabidopsis thaliana. Plant J. 51, 1062–1076.
https://doi.org/10.1111/j.1365-313X.2007.03205.x.
Kliebenstein, D.J., Cacho, N.I., 2016. Nonlinear selection and a blend of convergent,
divergent and parallel evolution shapes natural variation in glucosinolates. Adv. Bot.
Res. 80, 31–55. https://doi.org/10.1016/bs.abr.2016.06.002.
Klopsch, R., Witzel, K., Artemyeva, A., Ruppel, S., Hanschen, F., 2017. Genotypic
variation of glucosinolates and their breakdown products in leaves of Brassica rapa.
J. Agric. Food Chem. 66, 5481–5490. https://doi.org/10.1021/acs.jafc.8b01038.
Koch, M.A., German, D.A., Kiefer, M., Franzke, A., 2017. Database taxonomics as key to
modern plant biology. Trends Plant Sci. 23, 4–6. https://doi.org/10.1016/j.
tplants.2017.10.005.
Kozlov, A.M., Darriba, D., Flouri, T., Morel, B., Stamatakis, A., 2019. RAxML-NG: a fast,
scalable, and user-friendly tool for maximum likelihood phylogenetic inference.
Bioinformatics 35, 4453–4455. https://doi.org/10.1093/bioinformatics/btz305.
Kumar, R., Lee, S.G., Augustine, R., Reichelt, M., Vass˜
ao, D.G., Palavalli, M.H., Allen, A.,
Gershenzon, J., Jez, J.M., Bisht, N.C., 2019. Molecular basis of the evolution of
methylthioalkylmalate synthase and the diversity of methionine-derived
glucosinolates. Plant Cell 31, 1633–1647. https://doi.org/10.1105/tpc.19.00046.
Lanfear, R., Frandsen, P.B., Wright, A.M., Senfeld, T., Calcott, B., 2017. PartitionFinder 2:
new methods for selecting partitioned models of evolution for molecular and
morphological phylogenetic analyses. Mol. Biol. Evol. 34, 772–773. https://doi.org/
10.1093/mlbev/msw260.
Lee, J.G., Bonnema, G., Zhang, N., Kwak, J.H., de Vos, R.C.H., Beekwilder, J., 2013.
Evaluation of glucosinolate variation in a collection of turnip (Brassica rapa)
germplasm by the analysis of intact and desulfo glucosinolates. J. Agric. Food Chem.
61, 3984–3993. https://doi.org/10.1021/jf400890p.
Liang, X., Lee, H.W., Lu, Y., Zou, L., Ong, C.N., 2018. Simultaneous quantication of 22
glucosinolates in 12 Brassicaceae vegetables by hydrophilic interaction
chromatography-tandem mass spectrometry. ACS Omega 3, 15546–15553. https://
doi.org/10.1021/acsomega.8b01668.
Lin, L.-Z., Sun, J., Chen, P., Zhang, R.-W., Fan, X.-E., Li, L.-W., Harnly, J.M., 2014.
Proling of glucosinolates and avonoids in Rorippa indica (Linn.) hiern. (Cruciferae)
by UHPLC-PDA-ESI/HRMS(n). J. Agric. Food Chem. 62, 6118–6129. https://doi.
org/10.1021/jf405538d.
Linscheid, M., Wendisch, D., Strack, D., 1980. The structures of sinapic acid esters and
their metabolism in cotyledons of Raphanus sativus. Z. Naturforsch. 35c, 907–914.
https://doi.org/10.1515/znc-1980-11-1206.
Liu, T., Zhang, X., Yang, H., Agerbirk, N., Qiu, Y., Wang, H., Shen, D., Song, J., Li, X.,
2016. Aromatic glucosinolate biosynthesis pathway in Barbarea vulgaris and its
response to Plutella xylostella infestation. Front. Plant Sci. 7, 83. https://doi.org/
10.3389/fpls.2016.00083.
Liu, T.-J., Zhang, Y.-J., Agerbirk, N., Wang, H.-P., Wei, X.-C., Song, J.-P., He, H.-J.,
Zhao, X.-Z., Zhang, X.-H., Li, X.-X., 2019. A high-density genetic map and QTL
mapping of leaf traits and glucosinolates in Barbarea vulgaris. BMC Genom. 20, 371.
https://doi.org/10.1186/s12864-019-5769-z.
Louda, S.M., Rodman, J.E., 1983. Ecological patterns in the glucosinolate content of a
native mustard, Cardamine cordifolia, in the Rocky Mountains. J. Chem. Ecol. 9,
397–422. https://doi.org/10.1007/BF00988458.
Ludwig-Müller, J., Bennett, R.N., Garcia-Garrido, J.M., Pich´
e, Y., Vierheilig, H., 2002.
Reduced arbuscular mycorrhizal root colonization in Tropaeolum majus and Carica
papaya after jasmonic acid application can not be attributed to incerased
glucosinolate levels. J. Plant Physiol. 159, 517–523. https://doi.org/10.1078/0176-
1617-00731.
Maldini, M., Maksoud, S.A., Natella, F., Montoro, P., Petretto, G.L., Foddai, M., De
Nicola, G.R., Chessa, M., Pintore, G., 2014. Moringa oleifera: a study of phenolics and
glucosinolates by mass spectrometry. J. Mass Spectrom. 49, 900–910. https://doi.
org/10.1002/jms.3437.
Malka, O., Easson, M.L.A.E., Paetz, C., G¨
otz, M., Reichelt, M., Stein, B., Luck, K.,
Staniˇ
si´
c, A., Juravel, K., Santos-Garcia, D., Mondaca, L.L., Springate, S., Colvin, J.,
Winter, S., Gershenzon, J., Morin, S., Vass˜
ao, D.G., 2020. Glucosylation prevents
plant defense activation in phloem-feeding insects. Nat. Chem. Biol. 16, 1420–1426.
https://doi.org/10.1038/s41589-020-00658-6.
Mand´
akov´
a, T., Lysak, M.A., 2019. Healthy roots and leaves: comparative genome
structure of horseradish and watercress. Plant Physiol. 179, 66–73. https://doi.org/
10.1104/pp.18.01165.
Mand´
akov´
a, T., Zozomov´
a-Lihova, J., Kudoh, H., Zhao, Y., Lysak, M.A., Marhold, K.,
2019. The story of promiscuous crucifers: origin and genome evolution of an
invasive species, Cardamine occulta (Brassicaceae), and its relatives. Ann. Bot. 124,
209–220. https://doi.org/10.1093/aob/mcz019.
Marhold, K., Lihova, J., 2006. Polyploidy, hybridization and reticulate evolution: lessons
from the Brassicaceae. Plant Systemat. Evol. 259, 143–174. https://doi.org/
10.1007/s00606-006-0417-x.
Matth¨
aus, B., ¨
Ozcan, M., 2002. Glucosinolate composition of young shoots and ower
buds of capers (Capparis species) growing wild in Türkey. J. Agric. Food Chem. 50,
7323–7325. https://doi.org/10.1021/jf020530+.
Melich´
arkov´
a, A., ˇ
Slenker, M., Zozomov´
a-Lihov´
a, J., Skokanov´
a, K., ˇ
Singliarov´
a, B.,
Kaˇ
cm´
arov´
a, T., Caboˇ
nov´
a, M., Kempa, M., ˇ
Sr´
amkov´
a, G., Mand´
akov´
a, T., Lys´
ak, M.
A., Svitok, M., M´
artonov´
a, L., Marhold, K., 2020. So closely related and yet so
different: strong contrasts between the evolutionary histories of species of the
Cardamine pratensis popyploid species complex in Central Europe. Front. Plant Sci.
11, 588856. https://doi.org/10.3389/fpls.2020.588856.
Mithen, R., Bennett, R., Marquez, J., 2010. Glucosinolate biochemical diversity and
innovation in the Brassicales. Phytochemistry 71, 2074–2086. https://doi.org/
10.1016/j.phytochem.2010.09.017.
Mittchell, A.J., Weng, J.-K., 2019. Unleashing the synthetic power of plant oxygenases:
from mechanism to application. Plant Physiol. (Wash. D C) 179, 813–829. https://
doi.org/10.1104/pp.18.01223.
Montaut, S., Bleeker, R.S., Jacques, C., 2010. Phytochemical constituents of Cardamine
diphylla. Can. J. Chem. 88, 50–55. https://doi.org/10.1139/V09-153.
Montaut, S., Zhang, W.-D., Nuzillard, J.-M., De Nicola, G.R., Rollin, P., 2015.
Glucosinolate diversity in Bretschneidera sinensis of Chinese origin. J. Nat. Prod. 78,
2001–2006. https://doi.org/10.1021/acs.jnatprod.5b00338.
Montaut, S., De Nicola, G.R., Agnaniet, H., Issembe, Y., Rollin, P., Menut, C., 2017.
Probing for the presence of glucosinolates in three Drypetes spp. (Drypetes euryodes
(hiern) Hutch., Drypetes gossweileri S. Moore, Drypetes laciniata Hutch.) and two
Rinorea spp. (Rinorea subintegrifolia O. Ktze and Rinorea woermanniana (Buttner)
Engl.) from Gabon. Nat. Prod. Res. 31, 308–313. https://doi.org/10.1080/
14786419.2016.1236099.
Montaut, S., Read, S., Blaˇ
zevi´
c, I., Nuzillard, J.-M., Roje, M., Harakat, D., Rollin, P.,
2020a. Investigation of the glucosinolates in Hesperis matronalis L. And Hesperis
laciniata all.: Unveiling 4’-O-β-D-apiofuranosylglucomatrolin. Carbohydr. Res. 488,
107898. https://doi.org/10.1016/j.carres.2019.107898.
Montaut, S., Raharivelomanana, P., Butaud, J.-F., Lehartel, T., Rollin, P., 2020b.
Glucosinolates of the only three Brassicales indigenous to French Polynesia. Nat.
Prod. Res. 34, 2847–2851. https://doi.org/10.1080/14786419.2019.1591401.
Müller, C., Schulz, M., Pagnotta, E., Ugolini, L., Yang, T., Matthes, A., Lazzeri, L.,
Agerbirk, N., 2018. The role of the glucosinolate-myrosinase system in mediating
greater resistance of Barbarea verna than B. vulgaris to Mamestra brassicae larvae.
J. Chem. Ecol. 44, 1190–1205. https://doi.org/10.1007/s10886-018-1016-3.
Nelson, D., Werck-Reichhart, D., 2011. A P450-centric view of evolution. Plant J. 66,
194–211. https://doi.org/10.1111/j.1365-313X.2011.04529.x.
Nikolov, L.A., Shushkov, P., Nevado, B., Gan, X., Al-Shehbaz, I.A., Filatov, D., Bailey, C.
D., Tsiantis, M., 2019. Resolving the backbone of the Brassicaceae phylogeny for
investigating trait diversity. New Phytol. 222, 1638–1651. https://doi.org/10.1111/
nph.15732.
Novikova, P.Y., Hohmann, N., Nizhynska, V., Tsuchimatsu, T., Ali, J., Muir, G.,
Guggisberg, A., Paape, T., Schmid, K., Fedorenko, O.M., Holm, S., S¨
all, T.,
Schl¨
otterer, C., Marhold, K., Widmer, A., Sese, J., Shimizu, K.K., Weigel, D.,
Kr¨
amer, U., Koch, M.A., Nordborg, M., 2016. Sequencing of the genus Arabidopsis
identies a complex history of non-bifurcating speciation and abundant trans-
specic polymorphism. Nat. Genet. 48, 1077–1082. https://doi.org/10.1038/
ng.3617.
Okamura, Y., Tsuzuki, N., Kuroda, S., Sato, A., Sawada, Y., Hirai, M.Y., Murakami, M.,
2019. Interspecic differences in the larval performance of Pieris butteries
(Lepidoptera: pieridae) are associated with differences in the glucosinolate proles
of host plants. J. Insect Sci. 19, 1–9. https://doi.org/10.1093/jisesa/iez035.
Olsen, C.E., Huang, X.-C., Hansen, C.I.C., Cipollini, D., Ørgaard, M., Matthes, A., Geu-
Flores, F., Koch, M.A., Agerbirk, N., 2016. Glucosinolate diversity within a
phylogenetic framework of the tribe Cardamineae (Brassicaceae) unraveled with
HPLC-MS/MS and NMR-based analytical distinction of 70 desulfoglucosinolates.
Phytochemistry 132, 33–56. https://doi.org/10.1016/j.phytochem.2016.09.013.
One Thousand Plant Trancriptomes Initiative, 2019. One thousand plant transcriptomes
and the phylogenomics of green plants. Nature 574, 679–685. https://doi.org/
10.1038/s41586-019-1693-2.
Padilla, G., Cartea, M.E., Velasco, P., de Haro, A., Ord´
as, A., 2007. Variation of
glucosinolates in vegetable crops of Brassica rapa. Phytochemistry 68, 536–545.
https://doi.org/10.1016/j.phytochem.2006.11.017.
Pagnotta, E., Agerbirk, N., Olsen, C.E., Ugolini, L., Cinti, S., Lazzeri, L., 2017. Hydroxyl
and methoxyl derivatives of benzylglucosinolate in Lepidium densiorum with
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
25
hydrolysis to isothiocyanates and non-isothiocyanate products: substitution governs
product type and mass spectral fragmentation. J. Agric. Food Chem. 65, 3167–3178.
https://doi.org/10.1021/acs.jafc.7b00529.
Pagnotta, E., Montaut, S., Matteo, R., Rollin, P., Nuzillard, J.M., Lazzeri, L., Bagatta, M.,
2020. Glucosinolates in Reseda lutea L.: distribution in plant tissues during owering
time. Biochem. Systemat. Ecol. 90, 104043. https://doi.org/10.1016/j.
bse.2020.104043.
Parpazian, S., Girdwood, T., Wessels, B.A., Poelman, E.H., Dicke, M., Moritz, T.,
Albrectsen, B.R., 2019. Leaf metabolic signatures induced by real or simulated
herbivory in black mustard (Brassica nigra). Metabolomics 15, 130. https://doi.org/
10.1007/s11306-019-1592-4.
Pedras, M.S.C., Yaya, E.E., 2013. Dissecting metabolic puzzles through isotope feeding: a
novel amino acid in the biosynthetic pathway of the cruciferous phytoalexins
rapalexin A and isocyalexin. A. Org. Biomol. Chem. 11, 1149–1166. https://doi.org/
10.1039/C2OB27076E.
Pastorczyk, M., Bednarek, P., 2016. The function of glucosinolates and related
metabolites in plant innate immunity. Adv. Bot. Res. 80, 171–198. https://doi.org/
10.1016/bs.abr.2016.06.007.
Pedras, M.S.C., To, Q.H., Schatte, G., 2016. Divergent reactivity of an indole
glucosinolate yields Lossen or Neber rearrangement products: the phytoalexin
rapalexin A or a unique β-D-glucopyranose fused heterocycle. Chem. Commun. 52,
2505–2508. https://doi.org/10.1039/C5CC09822J.
Pellissier, L., Moreira, X., Danner, H., Serrano, M., Salamin, N., van Dam, N.,
Rasmann, S., 2016. The simultaneous inducibility of phytochemicals related to plant
direct and indirect defences against herbivores is stronger at low elevation. J. Ecol.
104, 1116–1125. https://doi.org/10.1111/1365-2745.15580.
Petersen, B.L., Andr´
easson, E., Bak, S., Agerbirk, N., Halkier, B.A., 2001. Characterization
of transgenic Arabidopsis thaliana with metabolically engineered high levels of p-
hydroxybenzylglucosinolate. Planta 212, 612–618. https://doi.org/10.1007/
s004250000429.
Petersen, A., Hansen, L.G., Mirza, N., Crocoll, C., Mirza, O., Halkier, B.A., 2019.
Changing substrate specicity and iteration of amino acid chain elongation in
glucosinolate biosynthesis through targeted mutagenesis of Arabidopsis
methylthioalkylmalate synthase 1. Biosci. Rep. 39, BSR20190446 https://doi.org/
10.1042/BSR20190446.
Pfalz, M., Mukhaimar, M., Perreau, F., Kirk, J., Hansen, C.I.C., Olsen, C.E., Agerbirk, N.,
Kroymann, J., 2016. Methyl transfer in glucosinolate biosynthesis mediated by
indole glucosinolate O-methyltransferase 5. Plant Physiol. (Wash. D C) 172,
2190–2203. https://doi.org/10.1104/pp.16.01402.
Poczai, P., Hyv¨
onen, J., 2010. Nuclear ribosomal spacer regions in plant phylogenetics:
problems and prospects. Mol. Biol. Rep. 37, 1897–1912. https://doi.org/10.1007/
s11033-009-9630-3.
Prasad, K.V.S.K., Song, B.-H., Olson-manning, C., Anderson, J.T., Lee, C.-R., Schranz, M.
E., Windsor, A.J., Clauss, M.J., Manzaneda, A.J., Naqvi, I., Reichelt, M.,
Gershenzon, J., Rupasinghe, S.G., Schuler, M.A., Mitchell-Olds, T., 2012. A gain-of-
function polymorphism controlling complex traits and tness in nature. Science 337,
1081–1084. https://doi.org/10.1126/science.1221636.
Rambaut, A., Drummond, A.J., Xie, D., Baele, G., Suchard, M.A., 2018. Posterior
summarization in Bayesian phylogenetics using Tracer 1.7. Syst. Biol. 67, 901–904.
https://doi.org/10.1093/sysbio/syy032.
Reichelt, M., Brown, P.D., Schneider, B., Oldham, N.J., Stauber, E., Tokuhisa, J.,
Kliebenstein, D.J., Mitchell-Olds, T., Gershenzon, J., 2002. Benzoic acid
glucosinolate esters and other glucosinolates from Arabidopsis thaliana.
Phytochemistry 59, 663–671. https://doi.org/10.1016/s0031-9422(02)00014-6.
Robin, A.H.K., Yi, G.-E., Laila, R., Yang, K., Park, J.-I., Kim, H.R., Nou, I.-S., 2016.
Expression proling of glucosinolate biosynthetic genes in Brassica oleracea L. var.
capitata inbred lines reveals their association with glucosinolate content. Molecules
21, 787. https://doi.org/10.3390/molecules21060787.
Rodman, J.E., Soltis, P.S., Soltis, D.E., Sytsma, K.J., Karol, K.G., 1998. Parallel evolution
of glucosinolate biosynthesis inferred from congruent nuclear and plastid gene
phylogenies. Am. J. Bot. 85, 997–1006. https://doi.org/10.2307/2446366.
Ronquist, F., Huelsenbeck, J.P., 2003. Mrbayes 3: Bayesian phylogenetic inference under
mixed models. Bioinformatics 19, 1572–1574. https://doi.org/10.1093/
bioinformatics/btg180.
Rossiter, J.T., James, D.C., 1990. Biosynthesis of (R)-2-hydroxybut-3-enylglucosinolate
(progoitrin) from [3,4-
3
H] but-3-enyIgIucosinoIate in Brassica napus. J. Chem. Soc.
Perkin Trans. 1, 1909–1913. https://doi.org/10.1039/P19900001909.
Sang, J.P., Minchinton, I.R., Johnstone, P.K., Truscott, R.J.W., 1984. Glucosinolate
proles in the seed, root and leaf tissue of cabbage, mustard, rapeseed, radish and
swede. Can. J. Plant Sci. 64, 77–93. https://doi.org/10.1007/10.1007/
BF00988082.
Schranz, M.E., Manzaneda, A.J., Windsor, A.J., Clauss, M.J., Mitchell-Olds, T., 2009.
Ecological genomics of Boechera stricta: identication of a QTL controlling the allocation
of methionine- vs branched-chain amino acid-derived glucosinolates and levels of insect
herbivory. Heredity 102, 465–474. https://doi.org/10.1038/hdy.2009.12.
Sugiyama, R., Hirai, M.Y., 2019. Atypical myrosinase as a mediator of glucosinolate
functions in plants. Front. Plant Sci. 10, 1008. https://doi.org/10.3389/
fpls.2019.01008.
Sønderby, I.E., Geu-Flores, F., Halkier, B.A., 2010. Biosynthesis of glucosinolates – gene
discovery and beyond. Trends Plant Sci. 15, 283–290. https://doi.org/10.1016/j.
tplants.2010.02.005.
Sørensen, M., Neilson, E.H.J., Møller, B.L., 2018. Oximes: unrecognized chameleons in
general and specialized plant metabolism. Mol. Plant 11, 95–117. https://doi.org/
10.1016/j.molp.2017.12.014.
Takos, A.M., Knudsen, C., Lai, D., Kannangara, R., Mikkelsen, L., Motawia, M.S., Olsen, C.
E., Sato, S., Tabata, S., Jørgensen, K., Møller, B., Rook, F., 2011. Genomic clustering of
cyanogenic glucoside biosynthetic genes aids their identication in Lotus japonicus and
suggests the repeated evolution of this chemical defense pathway. Plant J. 68,
273–286. https://doi.org/10.1111/j.1365-313X.2011.04685.x.
Textor, S., de Kraker, J.-W., Hause, B., Gershenzon, J., Tokuhisa, J.G., 2007. MAM3
catalyzes the formation of all aliphatic glucosinolate chain lengths in Arabidopsis.
Plant Physiol. 144, 60–71. https://doi.org/10.1104/pp.106.091579.
van den Bergh, E., Hofberger, J.A., Schranz, E., 2016. Flower power and the mustard
bomb: comparative analysis of gene and genome duplications in glucosinolate
biosynthetic pathway evolution in Cleomaceae and Brassicaceae. Am. J. Bot. 103,
1–11. https://doi.org/10.3732/ajb.1500445.
van Leur, H., Raaijmakers, C., van Dam, N.M., 2006. A heritable polymorphism within
natural populations of Barbarea vulgaris. Phytochemistry 67, 1214–1223. https://
doi.org/10.1016/j.phytochem.2006.04.021.
van Leur, H., Vet, L.E.M., van der Putten, W., van Dam, N.M., 2008. Barbarea vulgaris
glucosinolate phenotypes differentially affect performance and preference of two
different species of lepidopteran hervivores. J. Chem. Ecol. 34, 121–131. https://doi.
org/10.1007/s10886-007-9424-9.
Velasco, P., Rodríguez, V.M., Francisco, M., Cartea, M.E., Soengas, P., 2017. Genetics and
breeding of Brassica crops. In: M´
erillon, J.-M., Ramawat, K.G. (Eds.), Glucosinolates,
Reference Series in Phytochemistry. Springer, Cham, Switzerland, pp. 61–86.
https://doi.org/10.1007/978-3-319-25462-3_2.
Voutsina, N., Payne, A.C., Hancock, R.D., Clarkson, G.J.J., Rothwell, S.D., Chapham, M.
A., Taylor, G., 2016. Characterization of the watercress (Nasturtium ofcinale R. Br.;
Brassicaceae) transcriptome using RNASeq and identication of candidate genes for
important phytonutrient traits linked to human health. BMC Genom. 17, 378.
https://doi.org/10.1186/s12864-016-2704-4.
Walden, N., German, D.A., Wolf, E.M., Kiefer, M., Rigault, P., Huang, X.C., Kiefer, C.,
Schmickl, R., Franzke, A., Neuffer, B., Mummenhoff, K., Koch, M.A., 2020. Nested
whole-genome duplications coincide with diversication and high morphological
disparity in Brassicaceae. Nat. Commun. 11, 3795. https://doi.org/10.1038/s41467-
020-17605-7.
Wang, C., Dissing, M.M., Agerbirk, N., Crocoll, C., Halkier, B.A., 2020. Characterization
of Arabidopsis CYP79C1 and CYP79C2 by glucosinolate pathway engineering in
Nicotiana benthamiana shows substrate specicity toward a range of aliphatic and
aromatic amino acids. Front. Plant Sci. 11, 57. https://doi.org/10.3389/
fpls.2020.00057.
Wang, C., Dissing, M.M., Agerbirk, N., Crocoll, C., Halkier, B.A., 2021. Engineering and
optimization of the 2-phenylethylglucosinolate production in Nicotiana
benthamiana by combining biosynthetic genes from Barbarea vulgaris and
Arabidopsis thaliana. Plant J. https://doi.org/10.1111/tpj.15212 inpress.
Windsor, A.J., Reichelt, M., Figuth, A., Svatos, A., Kroymann, J., Kliebenstein, D.J.,
Gershenzon, J., Mitchell-Olds, T., 2005. Geographic and evolutionary diversication
of glucosinolates among near relatives of Arabidopsis thaliana (Brassicaceae).
Phytochemistry 66, 1321–1333. https://doi.org/10.1016/j.
phytochem.2005.04.016.
Wittstock, U., Halkier, B.A., 2000. Cytochrome P450 CYP79A2 from Arabidopsis thaliana
L. catalyzes the conversion of L-phenylalanine to phenylacetaldoxime in the
biosynthesis of benzylglucosinolate. J. Biol. Chem. 275, 14659–14666. https://doi.
org/10.1074/jbc.275.19.14659.
Wittstock, U., Kurzbach, E., Herfurth, A.-M., Stauber, E.J., 2016. Glucosinolate
breakdown. Adv. Bot. Res. 80, 125–169. https://doi.org/10.1016/bs.
abr.2016.06.006.
Xu, D., Hanschen, F.S., Witzel, K., Nintemann, S.J., Nour-Eldin, H.m, Schreiner, M.,
Halkier, B.A., 2017. Rhizosecretion of stele-synthesized glucosinolates and their
catabolites requires GTR-mediated import in Arabidopsis. J. Exp. Bot. 68,
3205–3214. https://doi.org/10.1093/jxb/erw355.
Yamane, A., Fujikura, J., Ogawa, H., Mizutani, J., 1992. Isothiocyanates as allelopathic
compounds from Rorippa indica Hiern. (Cruciferae) roots. J. Chem. Ecol. 18,
1941–1954. https://doi.org/10.1007/BF00981918.
Zang, Y.-X., Kim, H.U., Kim, J.A., Lim, M.-H., Jin, M., Lee, S.C., Kwon, S.J., Lee, S.I.,
Hong, J.K., Park, T.-H., Mun, J.-H., Seol, Y.-J., Hong, S.-B., Park, B.-S., 2009.
Genome-wide identication of glucosinolate synthesis genes in Brassica rapa. FEBS J.
276, 3559–3574. https://doi.org/10.3389/fpls.2016.00259.
Zhang, X., Liu, T., Duan, M., Song, J., Li, X., 2016. De novo transcriptome analysis of
Sinapis alba in revealing the glucosinolate and phytochelatin pathways. Front. Plant
Sci. 7, 259. https://doi.org/10.1111/j.1742-4658.2009.07076.x.
Züst, T., Strickler, S.R., Powell, A.F., Mabry, M.E., An, H., Mirzaei, M., York, T.,
Holland, C.K., Kumar, P., Erb, M., Petschenka, G., G´
omez, J.-M., Perfectti, F.,
Müller, C., Pires, J.C., Mueller, L.A., Jander, G., 2020. Independent evolution of
ancestral and novel defenses in a genus of toxic plants (Erysimum, Brassicaceae).
eLife 9, e51712. https://doi.org/10.7554/eLife.51712.
N. Agerbirk et al.
Phytochemistry 185 (2021) 112668
26
Niels Agerbirk (1967) is an Associate Professor at the Uni-
versity of Copenhagen and teaches biochemistry and chemistry.
He graduated in Biology from Aarhus University (1994) and
received his Ph.D. (1997) in Biochemistry from the Royal Vet-
erinary and Agricultural University, Denmark. Doctoral and
postdoctoral training (1994–2000) in natural product chemis-
try, enzymology and chemical ecology was in the laboratories
of H. Sørensen and J.K. Nielsen (Copenhagen), S. Palmieri
(Bologna) and J.A.A. Renwick (Ithaca), with a year in industry
(yeast group, B. Rønnow, Danisco Biotechnology). His research
is mainly on glucosinolates and related metabolites, with a
broad perspective ranging from chemistry and enzymology to
ecology and evolution.
Cecilie Cetti Hansen is a PhD fellow in plant biochemistry and
biotechnology, University of Copenhagen. Her current research
investigates structural diversity, biosynthesis, role and evolu-
tion of specialized metabolites in plants. Specic focus includes
functional studies of cytochrome P450 enzymes, NADPH-
dependent cytochrome P450 oxidoreductases, and UDP-
glycosyltransferases, and their key roles plant metabolism.
Christiane Kiefer (1979) is a Research Assistant at the Uni-
versity of Heidelberg. She graduated in Biology from Heidel-
berg University (2005) and received her Ph.D. (2008) from the
same University before receiving postdoctoral training at the
Max Planck Institute for Plant Breeding Research (Cologne,
Germany) in the laboratory of George Coupland. Since 2017
she works in at the COS, Heidelberg in the Department of
Marcus Koch. Her research focuses mainly on genomic analyses
in the Brassicaceae including association studies and phyloge-
netic analyses.
Thure P. Hauser (1959) is an Associate Professor at Univer-
sity of Copenhagen, with a PhD in Biology from Aarhus Uni-
versity (1994). Thure is specialized in plant ecology and
evolution, with a focus on reproduction, genetic variation and
interactions among plants, herbivores and associated mi-
crobes. He has previously been employed at the former
Research Center Risø and Botanical Institute, University of
Copenhagen. He teaches courses at various levels in plant
ecology.
Marian Ørgaard (1958) is an Associate Professor at the Uni-
versity of Copenhagen with a PhD in Experimental Plant Sys-
tematics from The Royal Veterinary and Agricultural
University of Copenhagen (1992). Doctoral and postdoctoral
training in molecular systematics was obtained in Prof J. S.
Heslop-Harrisons group, John Innes Centre, Norwich, Jodrell
lab in Kew, Leicester University, UK. Her research is focused on
plant speciation and evolution. She teaches botany – ecology
and plant systematics - at various levels.
Conny Bruun Asmussen Lange (1964) is an Associate Pro-
fessor at the University of Copenhagen and teaches Botany and
Plant Science. Conny graduated in Biology from Aarhus Uni-
versity (1992) and received her Ph.D. (1995) in Plant Sys-
tematics, also from University of Aarhus, Denmark. Doctoral
and postdoctoral training (1992–2000) in molecular system-
atics was in the laboratories of S.S. Renner (then Aarhus), A.
Liston (Oregon), O. Seberg (Copenhagen), J.J. Doyle (Ithaca)
and M.W. Chase (Kew, London). Her research is mainly on
phylogeny, DNA barcoding and molecular identication,
focusing on Palms, Legumes and native Danish plants.
Don Cipollini is a Professor of Biological Sciences at Wright
State University in Dayton, Ohio and Director of Wright State’s
Interdisciplinary Environmental Sciences PhD program. He has
BS and MS degrees in Biology from Indiana University of
Pennsylvania (1990, 1993) and a PhD in Ecology from Penn
State University (1997), advised by Dr. Jack Schultz. He con-
ducted postdoctoral training under Dr. Joy Bergelson at the
University of Chicago from 1997 to 1999. Dr. Cipollini’s
research focuses on the physiology and ecology of plant de-
fenses to herbivores and pathogens and the ecology, impacts,
and management of invasive plants and insects. His work in-
cludes studies of the chemical ecology of many ecologically and
agronomically important mustards, including Arabidopsis
thaliana, Brassica spp., and Alliaria petiolata, a notorious
invader of North America.
Marcus A. Koch (1967) is a Full Professor of Plant Systematics
and Evolution at the University of Heidelberg. He graduated in
Biology from Osnabrück University (1992) and received his Ph.
D. (1995) in Plant Evolutionary Biology. Afterwards he joined
Max-Planck-Institutes in Jena and Cologne, and was appointed
associate professor in Vienna. At Heidelberg University he is
acting as managing director of the Botanical Garden and Her-
barium. His research is covering a broad spectrum of organ-
ismal plant biology and environmental science with some
particular focus on Brassicaceae evolutionary history, system-
atics and taxonomy.
N. Agerbirk et al.
