The ‘mustard oil bomb’: not so easy to assemble?! Localization, expression and distribution of the components of the myrosinase enzyme system
Glucosinolates are plant secondary metabolites that are hydrolysed by the action of myrosinases into various products (isothiocyanates, thiocyanates, epithionitriles, nitriles, oxazolidines). Massive hydrolysis of glucosinolates occurs only upon tissue damage but there is also evidence indicating metabolism of glucosinolates in intact plant tissues. It was originally believed that the glucosinolate–myrosinase system in intact plants was stable due to a spatial separation of the components. This has been coined as the ‘mustard oil bomb’ theory. Proteins that form complexes with myrosinases have been described: myrosinase-binding proteins (MBPs) and myrosinase-associated proteins (MyAPs/ESM). The roles of these proteins and their biological relevance are not yet completely known. Other proteins of the myrosinase enzyme system are the epithiospecifier protein (ESP) and the thiocyanate-forming protein (TFP) that divert the glucosinolate hydrolysis from isothiocyanate production to nitrile/epithionitrile or thiocyanate production. Some glucosinolate hydrolysis products act as plant defence compounds against insects and pathogens or have beneficial health effects on humans. In this review, we survey and critically assess the available information concerning the localization, both at the tissular/cellular and subcellular level, of the different components of the myrosinase enzyme system. Data from the model plant Arabidopsis thaliana is compared to that from other glucosinolate-producing Brassicaceae in order to show common as well as divergent features of the ‘mustard oil bomb’ among these species.
KeywordsGlucosinolate hydrolysis Plant defence Secondary metabolites Myrosin cell Plant innate immunity
35S promoter of the cauliflower mosaic virus
Green fluorescent protein
Inducible myrosinase-associated protein
Myrosinase-binding protein-related protein
Transmission electron microscope
- Myrosin grain
Myrosin cell protein body
Glucosinolates (formerly known as mustard oil glucosides) are a group of sulphur- and nitrogen-containing secondary plant metabolites present in the order Capparales that contains the Brassicaceae family which includes crop plants such as oilseed rape (Brassica napus), white mustard (Sinapis alba), broccoli and cabbage (Brassica oleracea) as well as the model plant Arabidopsis thaliana. Glucosinolates are β-thioglucoside N-hydroxysulfates with a sulphur-linked β-d-glucopyranose moiety and a variable side chain that is derived from one of eight amino acids (for review: Daxenbichler et al. 1991; Fahey et al. 2001). Hydrolysis of glucosinolates upon tissue damage is catalyzed by glycosylated thioglucosidases called myrosinases (EC 22.214.171.124; glycoside hydrolase family 1) and results in a range of products (isothiocyanates, epithionitriles, nitriles, thiocyanates). The nature of these products depends on the structure of the glucosinolate in question, the conditions (e.g., pH) and the presence of additional proteins and cofactors. Isothiocyanates are the most common hydrolysis products whereas epithionitriles or nitriles are produced in the presence of epithiospecifier protein (ESP). In some cases thiocyanates are produced in the presence of a thiocyanate-forming factor (for review: Bones and Rossiter 2006).
In vitro glucosinolate hydrolysis assays with recombinant myrosinases or purified myrosinases have shown that myrosinases can act upon glucosinolates on their own. No additional proteins need to be present in order to produce isothiocyanates. However, it has been shown that some myrosinases can form complexes with two groups of proteins, called myrosinase-binding proteins (MBPs) and myrosinase-associated proteins (MyAP), in some species (for review: Rask et al. 2000). Whether these complexes are present in planta or are only formed following tissue damage has not yet been established. Knowing this would however help in answering the questions if and how the formation of these complexes affects myrosinase activity and/or function and what the biological relevance of these complexes is.
Whereas glucosinolates are considered as biologically inactive, their degradation products have toxic effects on fungi and bacteria (Brader et al. 2006). Some glucosinolate hydrolysis products are believed to have anticarcinogenic properties in humans (Holst and Williamson 2004; Fimognari and Hrelia 2007) while others have goitrogenic and antinutritional effects in animals (Fahey et al. 2001). Glucosinolate hydrolysis products have toxic effects on insect pests or serve as attractants for specialist insects (for review: Wittstock et al. 2003). Some specialist insects have evolved counteradaptive mechanisms to overcome this plant defence. Detoxification of glucosinolates by a sulfatase in the diamondback moth Plutella xylostella (Ratzka et al. 2002) or by a nitrile-specifier protein in the cabbage white butterfly Pieris rapae (Wittstock et al. 2004) have been described. Other insects such as the turnip sawfly Athalia rosae (Müller et al. 2001) or the cabbage aphid Brevicoryne brassicae (Bridges et al. 2002; Kazana et al. 2007) tolerate glucosinolates by sequestering them.
Enzyme and substrate are localized in the same subcellular compartment of the same cells but in ‘inactive’ forms,
They are localized in different subcellular compartments of the same cell,
They are localized in different cells, either in the same or in different subcellular compartments.
The question concerning the localization of the glucosinolate–myrosinase system has a long history and has been studied by several research groups using a panoply of different approaches and techniques (Andreasson and Jørgensen 2003). Although much of the early research was done on B. napus and S. alba, studying the myrosinase enzyme system has been facilitated by the fact that the model plant species A. thaliana belongs to the Brassicaceae family and, therefore, is a host to this system. The picture is however far from complete, even in A. thaliana, and results do not always seem to be consistent. This is mostly due to the different experimental designs and tools used by different research groups. But it also reflects the possibility that the composition, spatial distribution and functioning of the glucosinolate–myrosinase system differs among species. In addition, the composition of the myrosinase enzyme system is different between plant organs and changes during a plant’s growth cycle.
Although one may think that most is known about glucosinolate metabolism, new genes/proteins involved in the biosynthesis (Hansen et al. 2007) and hydrolysis (Burow et al. 2007a) of glucosinolates and the regulation (Gigolashvili et al. 2007a, b) of the system have only been recently identified and/or characterized. The true complexity of the system and its interaction with other biosynthetic pathways may yet be revealed.
The biosynthesis and transport of glucosinolates as well as the regulation of glucosinolate metabolism will not be discussed here and we refer the reader to recent reviews covering these aspects (Grubb and Abel 2006; Halkier and Gershenzon 2006; Yan and Chen 2007). In the following review we focus on the localization of glucosinolates and the enzymes and proteins involved in their degradation, the myrosinase enzyme system. The aim is to give a comprehensive survey of the current knowledge as well as the missing pieces in different species. Although we want to render the complexity of the system, not all of the available information could be discussed and some generalisations had to be made in order to make this review both informative and readable.
Spatial distribution of the myrosinase enzyme system
Approximately 120 glucosinolates have been described to date but each glucosinolate-containing plant species contains only a subset of these (Fahey et al. 2001). Glucosinolates are detected in organs throughout the plant (from roots to flowers) but both the nature and quantity of glucosinolates depend on the organ, the developmental stage and external factors (see for example: Brown et al. 2003; Shelton 2005; Velasco et al. 2007).
Are glucosinolates localized in a specific type of cells?
Kelly et al. (1998) analysed the cellular localization of glucosinolates in Brassica juncea cotyledons of imbibed seeds by using an antibody raised against 2-propenylglucosinolate (sinigrin). With exception of myrosin cells (see later) which showed no label, all ground tissue cells were labelled indicating that glucosinolates are distributed in all cotyledon cells at this mature seed stage. The cellular localization of glucosinolates in A. thaliana flower stalks was inferred by Koroleva et al. (2000) from sulphur measurements. One cell type had particular high sulphur content, hence referred to as S-cells, and myrosinase hydrolysis of the sap collected from these cells indicated the presence of glucosinolates. S-cells are positioned in groups between the endodermis and the phloem cells of each vascular bundle (Koroleva et al. 2000) and are close to and sometimes in direct contact with myrosin cells (Bones et al. 1991; Andreasson et al. 2001a; Husebye et al. 2002). The groups of S-cells are connected by a thin ring of cells that also reveal a high sulphur content and a third cell type characterized by high sulphur content were epidermal cells. The two latter types of cells were not analysed as to the presence of glucosinolates (Koroleva et al. 2000). Whereas the above mentioned studies detected glucosinolates in cell types that were different from myrosin cells, Thangstad et al. (2001) observed a slightly different situation in excised immature seeds of B. napus. Upon feeding of a radiolabelled glucosinolate precursor to pods, label localizes to specific cells in radicle and cotyledons of the developing embryo. The pattern and number of labelled cells matches the distribution of myrosin cells, indicating a possible localization of myrosinases and desulphoglucosinolates/glucosinolates at the examined stage in the same or adjacent cells (Thangstad et al. 2001). Whether the cells accumulating glucosinolates contain the machinery for de novo biosynthesis of glucosinolates or rely on import of glucosinolates is at present not known. For information about localization of the enzymes involved in the biosynthesis and the transport of glucosinolates and/or desulphoglucosinolates the reader is referred to recent reviews by Grubb and Abel (2006) and Halkier and Gershenzon (2006) and references cited therein.
Glucosinolates are contained in vacuoles
As to the subcellular compartment in which glucosinolates are stored, the vacuole is the most likely candidate. A localization of glucosinolates in vacuoles of horseradish (Armoracia lapathifolia = A. rusticana) and oilseed rape (B. napus) has been reported (Grob and Matile 1979; Matile 1980; Helmlinger et al. 1983; Yiu et al. 1984). Only one immunohistochemical localization of glucosinolates has been reported (Kelly et al. 1998). Using anti-glucosinolate antibodies, 2-propenylglucosinolate (sinigrin) was found to be localized in protein bodies/vacuoles of B. juncea cotyledons from imbibed seeds. However, 100 h after start of imbibition the sinigrin staining was restricted to the proximity of the cell wall (Kelly et al. 1998). No analysis of the subcellular localization of glucosinolates in A. thaliana has been reported.
In the three species where the localization of glucosinolates has been assessed, these seem to show a different distribution pattern. Glucosinolates are however difficult to localize and none of the techniques that were used provides unequivocal proof. In addition these studies focused on a particular organ and more extensive studies are required to conclude as to the localization of glucosinolates in other organs of the studied plants. As to the subcellular localization of glucosinolates, the evidence obtained so far seems to point quite consistently toward a vacuolar localization.
Most of the published studies on the expression and localization of the glucosinolate-degrading thioglucosidases, called myrosinases, have been performed on the three Brassicaceae species B. napus, S. alba and A. thaliana notably with the use of a diverse set of antibodies. Although this has been instrumental in understanding the system, the antibodies’ difference in specificity or their lack thereof complicates the interpretation and comparison of the results.
Complexity of myrosinase expression patterns in organs and tissues
Early purifications of B. napus seed myrosinases identified several isoenzymes (Lönnerdal and Janson 1973; Bones and Slupphaug 1989; Lenman et al. 1990). With polyclonal (K505) and monoclonal (3D7) antibodies raised against B. napus myrosinases, three distinct myrosinase polypeptides were detected in B. napus seed extracts (Lenman et al. 1990). These were called M75, M70 and M65, according to their apparent molecular masses of 75, 70 and 65 kDa, respectively. Several minor additional myrosinase bands with intermediate molecular masses were subsequently detected with these antibodies (3D7 and K505) on B. napus seed extracts (Lenman et al. 1993a). The M70 and M65 form, at least in vitro, complexes of high molecular masses with a group of proteins called myrosinase-binding proteins (MBPs) (Lenman et al. 1990). Although the M75 member was thought to form homodimers but not complexes with MBPs (Lenman et al. 1990; Geshi et al. 1998; Eriksson et al. 2002), a recent report indicates that even the M75 may form complexes (Bellostas et al. 2008).
The number of myrosinase encoding genes in B. napus is estimated to be around 20, which are distributed in the three subfamilies MA, MB and MC (Xue et al. 1992; Lenman et al. 1993b; Thangstad et al. 1993; Falk et al. 1995a; Rask et al. 2000). It has been shown that the B. napus M75, M70 and M65 seed myrosinases are encoded by MA, MC and MB genes respectively (Lenman et al. 1993a; Falk et al. 1995a). In mature B. napus seeds high myrosinase activity and transcript levels are observed (Bones 1990; Falk et al. 1992; Lenman et al. 1993a). A differential expression of the different myrosinase forms and their encoding gene families between organs and tissues has been shown in B. napus. Using probes that were specific to the myrosinase gene families MA and MB it was shown that their expression follows a similar temporal profile during embryogenesis in B. napus. It should be noted that only a few of the more than 20 myrosinase genes in B. napus have been sequenced and specificity has to be proven. In situ hybridizations on B. napus developing embryos show that MA and MB are both expressed in cotyledons and axis but that MB genes are preferentially expressed in cotyledons whereas MA and MC genes are mainly expressed in the axis (Lenman et al. 1993a; Falk et al. 1995a). The expression of different myrosinase proteins differs also between cotyledons, hypocotyls and roots in early B. napus seedling development (James and Rossiter 1991; Lenman et al. 1993a).
As discussed above several myrosinases ranging from 65 to 75 kDa were detected during B. napus embryogenesis and seedling development. Only one band of 72 kDa was detected on mature organs (leaf, stem, pedicel, flower, silique) of flowering B. napus plants, using the same antibodies (K505 and 3D7) (Lenman et al. 1993a). On the mRNA expression level, MC expression was shown to be restricted to developing seeds in B. napus (Falk et al. 1995a) and no MA transcripts were detected by northern blot in any adult B. napus tissue tested (Lenman et al. 1993a). However, from later studies it is clear that expression of MA type genes is not restricted to seeds. Using an MA type promoter (MYR1) Thangstad et al. (2004) showed that this promoter also directed expression to some adult tissues. MB on the other hand was expressed in all adult B. napus tissues tested (Falk et al. 1992; Lenman et al. 1993a). The presence of MB transcripts in B. napus roots was however not consistently observed and may indicate the presence of an additional family of myrosinases in this organ (Falk et al. 1992; Lenman et al. 1993a). It has become evident, mostly through transcriptional profiling performed on A. thaliana, that genes in the myrosinase enzyme system are responsive to abiotic factors (e.g., Bones et al. 1994; Nikiforova et al. 2003; Rizhsky et al. 2004; Jost et al. 2005). Some of the variation observed could, therefore, be related to different growth conditions or stages (see e.g., Bones et al. 1994).
With an estimated number of ten, S. alba presents fewer myrosinase encoding genes than B. napus (Eriksson et al. 2001). S. alba seed extracts have been shown to contain several myrosinases of different sizes (Björkman and Janson 1972; Björkman and Lönnerdal 1973; Xue et al. 1992, Bones et al. 1994). The antibody 3D7 raised against B. napus myrosinase recognizes water soluble myrosinases with apparent molecular masses of 72 to 75 kDa and of 78 kDa. These are encoded by MA genes, whereas 59 and 62 kDa myrosinases extractable in the presence of a denaturing agent are encoded by MB genes (Eriksson et al. 2001). Expression of myrosinases at the transcript level is however very low or absent in mature S. alba seeds (Xue et al. 1993). During S. alba embryo development MA and MB expressions follow a similar temporal profile (Xue et al. 1993). In situ hybridizations with myrosinase MA and MB family specific probes on developing embryos showed that both were expressed in cotyledons and axis. MB genes are however preferentially expressed in cotyledons whereas MA genes are mainly expressed in the axis (Xue et al. 1993).
MA gene expression seems to be embryo specific in S. alba as no transcripts were detected in cotyledons, hypocotyls, roots or emerging leaves of developing S. alba seedlings. MB transcripts on the other hand were detected during S. alba seedling development. Their expression is higher in hypocotyls than cotyledons but decreases in both tissues over time whereas that in leaves increases. In seedling roots MB transcript levels were hardly detected (Xue et al. 1993). A differential expression of myrosinases between S. alba organs has also been shown at the protein level (Bones 1990; Bones et al. 1994; Eriksson et al. 2001).
In A. thaliana the situation regarding myrosinase expression is better characterized and seems to be less complex than in B. napus and S. alba. Only six myrosinase genes have been identified in A. thaliana where they are called TGG for thioglucoside glucohydrolase (a synonym for thioglucosidase) (Chadchawan et al. 1993; Xue et al. 1992). Expression analyses of the genes TGG1 (At5g26000) and TGG2 (At5g25980) and the proteins they encode have been published. The use of antibodies that discriminate between TGG1 and TGG2 allowed Ueda et al. (2006) to compare the expression patterns of these two myrosinases in A. thaliana Col-0. Western blot analysis shows that TGG2 is mainly expressed in flowers and siliques whereas TGG1 is highly expressed in rosette leaves, flowers and less in siliques, flower stalks and seedlings. None of these two myrosinase proteins is detected in roots or seeds (Ueda et al. 2006). This corroborates TGG1 promoter expression studies (Husebye et al. 2002; Thangstad et al. 2004) although some TGG1 transcripts had been detected earlier by in situ hybridization of mature seeds (Xue et al. 1995). Sequence analysis of the third myrosinase gene of A. thaliana, TGG3 (At5g48375), indicates that it is a pseudogene (Zhang et al. 2002) and likely has no myrosinase activity. It is noteworthy that its expression at the transcript level is limited to floral tissues, more specifically to petal and stamen (Zhang et al. 2002). A regulatory role of the pseudogene should, therefore, not be excluded.
Three additional myrosinase genes, called TGG4, TGG5 and TGG6 have been identified (Andersson et al. 2004; Xu et al. 2004), but their characterization is limited. TGG4 (At1g47600) and TGG5 (At1g51470) have been shown to encode functional myrosinases and seem to have root specific expression patterns (Andersson et al. 2004). Publicly available transcript expression data (Toufighi et al. 2005; Zimmermann et al. 2004) seems to corroborate these findings. This expression pattern is also consistent with the fact that A. thaliana roots present myrosinase activity despite the absence of TGG1 and TGG2 expression. In addition, tgg1tgg2 double mutant plants present similar root myrosinase activity as the wild-type plants (Barth and Jander 2006). No myrosinase activity has yet been reported for TGG6 (At1g51490). Its expression at the mRNA level seems to be flower specific with the highest expression levels in pollen (Toufighi et al. 2005).
Although unlikely, one cannot exclude the presence of other, so far unidentified, root myrosinases among the members of glycoside hydrolase family 1 (Xu et al. 2004). A gene, called PYK10 (At3g09260), had been identified some years ago as a ‘root myrosinase’ (Nitz et al. 2001) but it is unlikely that it represents a genuine myrosinase. Indeed, β-d-glucosidase and β-d-fucosidase activities have been shown for PYK10 (Matsushima et al. 2003, 2004).
Myrosinases are localized in myrosin cells
The use of monoclonal and polyclonal antibodies (UNI288 and K089) raised respectively against B. napus and S. alba myrosinases made it possible to show that myrosinases were indeed localized in myrosin cells of B. napus cotyledons and radicles of imbibed seeds (Thangstad et al. 1990, 1991; Bones et al. 1991). A myrosin cell localization of myrosinases was independently confirmed by Höglund et al. (1991) using the anti-myrosinase antibodies 3D7 and K505 on B. napus developing embryos. The expression occurs in a time-dependent manner and is only observed in embryonic cells (cotyledon and axis). Epidermal cells, meristematic cells of the axis and provascular cells of the cotyledons were not labelled (Höglund et al. 1991). Although the distribution of myrosin cells is usually considered to be random, Thangstad et al. (1990) and Bones et al. (1991) noted that a large proportion of these cells are localized in the second-outmost cell layer (Fig. 1b) of the embryo axis and cotyledon of B. napus and S. alba.
During B. napus seed germination myrosinases were detected by immunohistochemistry in some parenchyma and vascular cells of the axis and in the guard cells and a small number of parenchyma cells of the cotyledon (Bones et al. 1991; Höglund et al. 1991; Andreasson et al. 2001a).
In hypocotyls and young rosette leaves of adult B. napus plants, myrosinases were detected in scattered parenchyma cells of the mesophyll and the vascular tissue as well as guard cells (Bones et al. 1991; Höglund et al. 1991; Andreasson et al. 2001a). Expression in guard cells was reported to disappear in older leaves (Höglund et al. 1991). Myrosin cells in B. napus siliques are localized both in the mesophyll and associated with the vascular tissue (Höglund et al. 1991; Andreasson et al. 2001a). Whereas Höglund et al. (1991) detect myrosin cells in B. napus petals both in the mesophyll and associated with the vascular tissue, Andreasson et al. (2001a) detect them only in the phloem parenchyma. In stems myrosinases are localized in the xylem and some cortical cells according to Höglund et al. (1991) whereas Andreasson et al. (2001a) consider myrosinase expression to be phloem specific in this tissue. In B. napus roots, mainly cortical cells were labelled with anti-myrosinase antibodies (Höglund et al. 1991; Andreasson et al. 2001a). In a more recent study, a myrosinase promoter from B. napus (the MYR1 promoter; Fig. 1c) was shown to lead to myrosinase expression in myrosin cells localized in the cortex of B. napus radicles (Thangstad et al. 2004).
In developing S. alba embryos myrosinase transcripts are also localized in myrosin cells, as was shown by in situ hybridization (Xue et al. 1993). The distribution pattern of these cells in S. alba embryos is similar to the one observed by Höglund et al. (1991) in B. napus embryos. A myrosin cell-restricted localization of myrosinase in cotyledons and radicles of imbibed S. alba seeds was shown with the polyclonal antibody K089 raised against S. alba myrosinase (Thangstad et al. 1991). In contrast to myrosinase protein detection in B. napus leaf guard cells and vascular tissue (Höglund et al. 1991; Andreasson et al. 2001a), no labelling of these cell types was observed in S. alba leaves by in situ hybridization (Xue et al. 1993).
Hara et al. used the tissue printing technique to localize myrosinase to the epidermis and the vascular cambium of Raphanus sativus, Brassica campestris and Wasabia japonica roots and at lower levels to the cortex of Raphanus sativus roots. The exact nature of the cells could however not be identified (Hara et al. 2000; 2001). In cotyledons and radicles of imbibed seeds of R. sativus and B. oleracea myrosinase expression was also shown to be restricted to myrosin cells by the use of the polyclonal antibody K089 raised against S. alba myrosinase (Thangstad et al. 1991). Similarly, immunohistochemical analysis with an antibody raised against B. napus myrosinase (James and Rossiter 1991) showed myrosinase expression in myrosin cells of cotyledons of imbibed B. juncea seeds (Kelly et al. 1998).
In A. thaliana, in situ hybridizations with TGG1 and TGG2 specific probes showed expression in scattered cells of seedling leaf tissue. No labelling of epidermis or xylem was observed. In flowers, signals for the two myrosinases were detected in sepal, petal and gynoecium but not in stamen or pollen grains. The positive cells in flowers were close to but not in the xylem (Xue et al. 1995). Extensive localization analyses of A. thaliana TGG1 and TGG2 at the transcript level by promoter expression experiments (Fig. 1d, e; Husebye et al. 2002; Thangstad et al. 2004; Barth and Jander 2006) and at the protein level by immunohistochemistry with anti-myrosinase antibodies (Andreasson et al. 2001a; Husebye et al. 2002; Ueda et al. 2006) have been reported. The results show TGG1 and TGG2 related staining of cells situated in the phloem parenchyma of all tested organs of seedlings and mature plants except in roots, developing seeds and mature seeds. Staining was additionally observed in guard cells and this was specifically related to TGG1 expression. Surprisingly the Arabidopsis TGG1 promoter also led to guard cell specific expression in tobacco plants, a plant with no endogenous myrosinase or glucosinolates (Thangstad et al. 2004). The use of anti-myrosinase antibody K089 showed also staining of cells in the cortex, the endodermis and the xylem of A. thaliana stems (Husebye et al. 2002), which was not observed in the other studies.
A more detailed analysis of the myrosinase-positive phloem cells in stems revealed them to be specific phloem parenchyma cells. These cells differentiate themselves from the sieve element-companion cell complex by a larger diameter and length. The myrosin phloem cells face the endodermis and are situated in proximity to and sometimes in direct contact with the glucosinolate containing S-cells (Andreasson et al. 2001a; Husebye et al. 2002; Ueda et al. 2006).
A complementary approach, in which the promoter of the B. napus myrosinase MYR1 (a MA type myrosinase) gene was used instead of the A. thalianaTGG1 promoter, gave a similar staining of myrosin cells and guard cells in transformed A. thaliana plants. Interestingly enough, the same construct gave a staining restricted to myrosin cells in developing B. napus seeds and seedlings, indicating differences in the regulation mechanisms (Thangstad et al. 2004). That a difference in myrosinase localization exists among plant species is also shown by the fact that myrosin cells of B. napus and S. alba are observed in the ground tissue whereas this is not the case in A. thaliana. Another difference already mentioned between B. napus and S. alba on one hand and A. thaliana on the other is at the mature dry seed level: myrosinases are abundant in B. napus and S. alba seeds (Bones 1990) but absent or expressed at very low levels in A. thaliana seeds (Gallardo et al. 2001; Ueda et al. 2006).
Are different myrosinases expressed in different myrosin cells?
From the results presented above it is clear that in all plants that were studied myrosin cells contained myrosinase. Different myrosin cell types were detected, although no difference in ultrastructure was revealed between the myrosin cells of the phloem and the ground tissue in B. napus (Andreasson et al. 2001a and references therein). Myrosinases within a species show sequence divergence, different extents of glycosylation and different organ/tissue expression. An interesting question is therefore: Do different myrosin cells contain different myrosinases?
In situ experiments on B. napus embryos with MA and MB specific probes showed that a differential expression seems indeed to exist as some myrosin cells were only labelled by the probes for one of the gene families. However, most myrosin cells accumulated transcripts of MA and MB genes (Lenman et al. 1993a). Subsequently, Falk et al. (1995a) showed that MB and MC genes are expressed in the same myrosin cells of developing B. napus seeds. Similarly, in situ experiments performed on S. alba embryo axis with MA and MB specific probes showed that some myrosin cells showed either MA or MB transcripts whereas the majority were labelled by both probes (Xue et al. 1993). Sequence information was only available for a few S. alba myrosinase genes and cross-reaction between MA and MB family-specific probes can not be totally excluded.
The initial results on A. thalianaTGG1 and TGG2 expression analysis by in situ hybridization presented by Xue et al. (1995) did not allow to discern whether the labelled cells in organs where a signal was obtained for both TGG1 and TGG2 contain both transcripts, and such a possibility was not discussed by the authors. It was recently shown that TGG1 and TGG2 can be detected in the same myrosin phloem cell (Ueda et al. 2006), but whether myrosin phloem cells are capable of expressing specifically one of the two myrosinases is unresolved. In guard cells of A. thaliana however myrosinase expression is due to TGG1 but not TGG2 expression (Husebye et al. 2002; Thangstad et al. 2004; Barth and Jander 2006).
Myrosinases are vacuolar proteins
For a review on early attempts to determine the subcellular localization of myrosinase the reader is referred to Bones and Iversen (1985) and Bones and Rossiter (1996). More recently the vacuolar localization of myrosinases has been consistently shown in several plant species by a variety of techniques. Immunohistochemical analysis on B. napus with the monoclonal anti-myrosinase antibody UNI288 allowed Thangstad et al. (1990) to propose an association of myrosinase with the tonoplast-like membrane surrounding globular protein grains, the ‘myrosin grains’, in myrosin cells. Höglund et al. (1991) additionally noted a uniform distribution of the label in the cytoplasm. A more detailed investigation of the myrosinase localization showed that myrosinase was distributed in myrosin cell protein bodies or vacuoles (Bones et al. 1991). Myrosin cells were found both in ground tissue and in vascular tissue. A subcellular localization in myrosin cell protein bodies/vacuoles was proven using immunogold-EM on B. napus, B. oleracea, S. alba and R. sativus using the polyclonal anti-myrosinase antibody K089 (Thangstad et al. 1991). A uniformly distributed labelling over the myrosin cell protein bodies (‘myrosin grains’) in cotyledons and radicles of imbibed seeds was observed (Thangstad et al. 1991). This was later confirmed in cotyledons of B. napus imbibed seeds (Geshi et al. 1998) and in developing embryos of S. alba (Höglund et al. 1992) with the anti-myrosinase antibody 3D7. Grob and Matile (1979) also detected a substantial proportion of myrosinase activity in a vacuolar fraction from Armoracia lapathifolia (= A. rusticana) root. In A. thaliana, Andreasson et al. (2001a) detected myrosinase in the vacuoles of myrosin cells with the anti-myrosinase antibody 3D7. Confocal microscopy revealed however a heterogeneous staining of a structure forming a continuous reticular system, named ‘myrosin body’, rather than individual myrosin grains. Immunogold-EM analysis with TGG1 and TGG2 specific antibodies confirmed that both TGG1 and TGG2 localize to vacuoles (Ueda et al. 2006). Recently, two proteomic studies of A. thaliana vacuoles also reported identification of TGG1 and TGG2 in A. thaliana vacuoles (Carter et al. 2004; Chen et al. 2006). However, Jaquinod et al. (2007) did not detect myrosinases in their proteomic study of vacuoles from A. thaliana suspension cells and hypothesized that a difference in starting material was the cause. A. thaliana suspension cells seem indeed to contain only very low levels of myrosinase proteins and activity in comparison to other plant tissue (Alvarez et al. 2008). This is supported by earlier results investigating myrosinase activity in cell and tissue cultures and regenerant plants of Brassica napus and Sinapis alba (Bones 1990). The myrosinase activity was found to be very low in, e.g., callus which has a low level of tissue organisation and in cell cultures. More intriguingly however is a recent report that identifies both TGG1 and TGG2 in the proteome of A. thaliana leaf peroxisomes (Reumann et al. 2007).
Brassicaceae species exhibit a high complexity regarding the number of myrosinases, differences in their expression among organs, changes in expression during the life cycle of a plant and the absence or presence of protein complexes. Myrosinases are localized in the vacuoles of idioblastic cells called myrosin cells whose tissular distribution may vary between the same organs of different species and between organs of the same plant. In species such as B. napus, with as many as 20 potential myrosinases, we may currently only have got a glimpse of the complexity of the situation. The sequencing of the Brassica genome will undoubtedly provide us with a valuable tool in reevaluating the information that was gathered so far. Although the use of antibodies and family-specific probes have been helpful, more refined techniques will be necessary to get a more detailed knowledge of the expression and localization of myrosinases in B. napus. Even in A. thaliana that has been used extensively in glucosinolate research the last 10 years and that possesses a much more restricted number of functional myrosinases, many open questions remain. Especially the characterization of root myrosinases is awaited with much interest.
Myrosinase-binding proteins (MBPs) were originally identified in B. napus as they co-purified with seed myrosinases. As mentioned before these MBPs form complexes of high molecular masses with the M70 and M65 myrosinases but they exist also as free polypeptides. Originally two proteins, called MBP50 and MBP52 because of their apparent molecular masses of 50 and 52 kDa, were characterized (Lenman et al. 1990). Subsequently, two additional proteins of 80 and 100 kDa (immunoreacting with the monoclonal anti-MBP antibody 34:14) were identified in B. napus seeds. These, however, are not able to form complexes with myrosinases and were therefore referred to as MBPRPs, which stands for MBP related proteins (Falk et al. 1995b). Geshi and Brandt (1998) characterized two MBPs (called later MBP70 and MBP97; Geshi et al. 1998) from B. napus seedlings that form complexes with the M70 and M65 myrosinases. That MBPs are required for complex formation of myrosinases in B. napus has been shown by the absence of complexes in antisense MBP plants (Eriksson et al. 2002). MBP sequences are characterized by repeats of different nature (Taipalensuu et al. 1997c), one of them being jacalin-like lectin domains (PF01419). Lectin activity was experimentally shown for the B. napus seed MBP50/52 as well as the seedling MBP 70p (Taipalensuu et al. 1997b; Geshi and Brandt 1998).
The A. thaliana genome possesses 17 genes encoding putative myrosinase-binding proteins (Rask et al. 2000) but no experimental evidence of a complex formation between these proteins and myrosinases in A. thaliana has been presented. The designation of myrosinase-binding protein (MBP) relies, therefore, solely on sequence homology to the earlier characterized B. napus MBPs.
Expression of myrosinase-binding proteins
The highest MBPs/MBPRPs protein levels in B. napus were detected (with the antibody 34:14) in seeds, pistils and sepals, whereas they were almost absent in stems and anthers (Falk et al. 1995b). In addition, the four members differ in their expression pattern: MBP50 and MBP52 are only detected in seeds, MBPRP80 is only present in seeds and sepals whereas MBPRP100 is present in most tested organs (Falk et al. 1995b). The MBP70 identified by Geshi and Brandt (1998) is absent in mature seed but appears during germination (after 2 days), where it is mainly expressed in the hypocotyl and the root but barely in the cotyledon. The MBP97 is present in mature seeds and accumulates further during germination, together with polypeptides of higher apparent molecular masses (Geshi et al. 1998).
The information concerning the expression of A. thaliana MBPs is sparse and only three have been characterized in some detail on the transcript level. Takechi et al. (1999) identified an MBP-encoding cDNA that showed flower specific expression by Northern blot. More detailed analysis by in situ hybridization showed expression in floral meristems, young pistils and stamens at an early stage of flower development. In immature flowers a signal was detected in petals and ovules and weakly in filaments. Furthermore this MBP transcript was detected in the style just before anthesis and in the embryo and placenta tissue after fertilisation. No signal was detected in developing seeds (Takechi et al. 1999).
Capella et al. (2001) also described two flower specific MBPs from A. thaliana, called MBP1 and MBP2, with MBP2 having a high sequence similarity to the MBP described by Takechi et al. (1999). In situ hybridization of MBP1 on floral tissue detects expression in male (stamens, filament and tapetal cells) and female (ovule, ovary and style) organs, petals and pedicels. No signal was observed in pollen grains or sepals (Capella et al. 2001). The flower specific expression of MBP1 and MBP2 argues against a requirement of these MBPs for TGG1 function, as the latter is expressed in several other plant tissues (Capella et al. 2001).
The use of antibodies raised against the MBPs of B. napus has not led to the detection of MBPs in A. thaliana. Andreasson et al. (2001a) even questioned the presence of a functional MBP homolog in A. thaliana and speculated that MBPs may be correlated to the presence of ground tissue myrosin cells. Proteins encoded by MBP1 and other putative myrosinase binding proteins have been identified in A. thaliana in independent proteomic studies (Carter et al. 2004; Casasoli et al. 2007; Chen et al. 2006; Huttlin et al. 2007) but whether these proteins have any function in glucosinolate metabolism or not and what this function could be is yet unknown.
Do myrosinase-binding proteins colocalize with myrosinases?
Immunocytochemical analyses with anti-myrosinase antibody 3D7 and anti-MBP antibody S4C6 showed that myrosinases and MBP97 colocalize in myrosin cells of hypocotyls and cotyledons of 7 day old B. napus seedlings (Geshi et al. 1998; Andreasson et al. 2001a, b). Whereas Geshi et al. (1998) also observe a colocalization in cotyledons of B. napus seeds imbibed for 2 days, Andreasson et al. (2001a) do not observe this in 2 day old B. napus seedlings. In the latter case MBPs are detected in all ground tissue cells, but not in myrosin cells, the epidermis and vascular tissue (Andreasson et al. 2001a). Localization of MBPs and myrosinases in different cells was also observed in mature (later than 40 days after pollination) B. napus embryos (Eriksson et al. 2002). The colocalization of MBPs and myrosinases in the same cells at certain developmental stages allows at least theoretically the formation of complexes in planta. The localization of myrosinases and MBPs in different cells of B. napus at other stages argues in favour of a complex formation solely after tissue rupture.
Myrosinase-binding proteins may be vacuolar proteins too
Differential centrifugation indicated that the B. napus MBP70 is either membrane-associated or part of a large insoluble protein complex (Geshi and Brandt 1998). Immunogold labelling showed localization of MBP70 and MBP97 to vacuolar myrosin grains in cotyledons, roots and hypocotyls of B. napus seedlings (Geshi and Brandt 1998; Geshi et al. 1998).
In an analysis of the proteome of vacuoles derived from A. thaliana rosette leaf protoplasts, Carter et al. (2004) identified three ‘jacalin lectin’ proteins that correspond to MBP1 and two other putative myrosinase-binding proteins. Chen et al. (2006) also reported identification of myrosinase-binding proteins in A. thaliana vacuoles and Huttlin et al. (2007) identified peptides of MBPs in a proteomics study of the organellar fraction extracted from 10 day old plants. Based on the absence of predicted signal peptides in MBP sequences some authors regard it as unlikely that these proteins are localized to the vacuoles (Ueda et al. 2006). Casasoli et al. (2007) detect two putative MBPs in the nuclear proteome of A. thaliana.
As their name indicates myrosinase-binding proteins form complexes with myrosinases. These have been shown in B. napus but not yet in A. thaliana, and the role of these complexes has not yet been fully established. MBPs are most probably vacuolar proteins. A cellular colocalization with myrosinases was also observed in some contexts.
The first myrosinase-associated protein (MyAP or MAP) was identified as a 40 kDa glycoprotein in myrosinase complexes from B. napus seeds, hence its name (Falk et al. 1995a). Subsequently additional MyAPs have been identified in B. napus (Taipalensuu et al. 1996). Like MBPs/MBPRPs, MyAPs have been found to form complexes with the M70 and M65 seed myrosinases of B. napus. However, as antisense MBP plants do not present complexes, MyAPs are not responsible for establishing them (Eriksson et al. 2002). MyAPs are more prevalent as myrosinase-free forms (Taipalensuu et al. 1996) and it has not been shown if the complexes exist in planta.
In A. thaliana a handful of genes are currently annotated as encoding putative myrosinase-associated proteins (TAIR annotation V7). This annotation is based on sequence similarity to the B. napus MyAPs described above and not on published experimental data. The only one of these A. thaliana putative MyAPs that has been characterized in some detail is the epithiospecifier modifier 1 (ESM1; encoded by At3g14210) (Zhang et al. 2006). Although its biochemical activity is largely unknown, it was shown to affect negatively the nitrile to isothiocyanate ratio in glucosinolate degradation and its name is derived from the fact that this happens in combination with the epithiospecifier protein (ESP; Zhang et al. 2006). Although the possibility of a direct interaction of ESM1 with ESP and/or myrosinases was mentioned (Zhang et al. 2006), this remains purely speculative. MyAPs are structurally unrelated to MBPs/MBPRPs. Their sequences present a GDSL-lipase like domain (PF00657; Akoh et al. 2004) as common feature, but no lipase activity has yet been reported for any of the MyAPs.
Expression analysis of myrosinase-associated proteins
At the transcript level B. napus MyAPs are expressed in all organs with the highest expression in immature seeds, where they are expressed at the later stages of development. In 3-week-old plants expression is higher in root and hypocotyl than in leaf and cotyledon. In flowering plants, expression is hardly visible in any of the tested organs (Taipalensuu et al. 1996). A myrosinase-associated protein called iMyAP because its transcript was inducible by treatments such as wounding and methyljasmonate application has also been described in B. napus (Taipalensuu et al. 1997a; Andreasson et al. 1999). Promoter expression studies of iMyAP from B. napus in transgenic A. thaliana plants in an uninduced state, showed expression in the hypocotyl of 4 day old seedlings but no consistent expression in 7 day old seedlings. In flowering plants expression at the uninduced state was restricted to the silique abscission zone (Andreasson et al. 1999). This expression pattern however does not seem to be representative of other B. napus MyAP genes (Taipalensuu et al. 1996).
Expression analysis of the only characterized A. thaliana MyAP, the epithiospecifier modifier 1 (ESM1), was not reported (Zhang et al. 2006). Publicly available transcript expression data indicates however that ESM1 expression is largely restricted to rosette leaves, whereas the other putative MyAP genes show quite different expression patterns (Zimmermann et al. 2004; Toufighi et al. 2005). Whether the differences in expression pattern between MyAPs are indicative of different functional roles is not known.
The cellular localization of myrosinase-associated proteins is not yet known
Even less is known about the cell type in which MyAPs are expressed. In situ hybridizations on the iMyAP did not allow to identify the cell type in which transcripts were detected at the uninduced state but showed that expression appeared in virtually all cells upon induction (Taipalensuu et al. 1997a). A similar unspecific expression pattern after induction was observed in transgenic plants transformed with an iMyAP promoter:GUS construct (Andreasson et al. 1999). The data published for epithiospecifier modifier 1 (ESM1) did not include localization data (Zhang et al. 2006). Nothing in the published reports point to a myrosin cell specific expression of MyAPs.
MyAPs subcellular localization may be vacuolar too
According to Zhang et al. (2006) homology and sequence analysis suggests that ESM1 is membrane-attached and resides in the endoplasmic reticulum. Experimental evidence so far suggests otherwise, but is not conclusive in itself either. In a study of the A. thaliana vegetative proteome, ESM1 was described as representing the bulk of vacuolar protein content. In addition five other putative myrosinase-associated proteins were detected in the vacuoles (Carter et al. 2004). Three of these were also identified in an independent A. thaliana tonoplast proteomics study (Shimaoka et al. 2004) and Chen et al. (2006) also reported identification of myrosinase-associated proteins in A. thaliana vacuoles. In another proteomics study Huttlin et al. (2007) identified peptides for ESM1 in the organellar fraction but other putative MyAPs in both the organellar and microsomal fraction. Recently, ESM1 was however also identified as a peroxisomal protein in a proteomics study of A. thaliana (Reumann et al. 2007). Hence, although most studies point toward a vacuolar localization of MyAPs a localization to ER and peroxisomes have also been reported.
Myrosinase-associated proteins (MyAPs) are another group of proteins present in the myrosinase complexes. ESM1, an A. thaliana MyAP, has been shown to affect the outcome of glucosinolate hydrolysis. The types of cells in which MyAPs are expressed have not yet been identified and the subcellular localization has not yet been established unequivocally.
Epithiospecifier protein (ESP) and thiocyanate-forming protein (TFP)
A more detailed analysis of epithiospecifier protein (ESP) and thiocyanate-forming protein (TFP) will be presented elsewhere in this issue (Burow and Wittstock, this issue). ESPs have been purified from and characterized in different Brassicaceae species like B. napus (Bernardi et al. 2000; Foo et al. 2000), B. oleracea (Matusheski et al. 2006) and A. thaliana (Lambrix et al. 2001; Zabala et al. 2005). As mentioned above not all Brassicaceae species possess a functional ESP and even within the A. thaliana species, ecotypes with or without ESP activity have been identified. When present, ESP diverts the myrosinase-catalyzed glucosinolate hydrolysis from isothiocyanates to either nitriles or epithionitriles, although the exact mechanism is not yet known (Lambrix et al. 2001). The formation of thiocyanate from glucosinolates was only observed in a few plants and is restricted to three glucosinolates (for review: Bones and Rossiter 2006). Recently, a thiocyanate-forming protein (TFP), showing 68% sequence similarity to A. thaliana ESP, has recently been identified in Lepidium sativum (Burow et al. 2007a) but the information about the expression and localization of thiocyanate-forming factors is still very limited.
Expression analysis of epithiospecifier protein (ESP)
Expression analysis of ESP has only been reported for A. thaliana. Analysis at the transcript, protein and activity level shows expression of ESP in all organs, except roots and anthers, with levels changing during the plant’s growth cycle (Burow et al. 2007b).
Cellular localization of epithiospecifier protein (ESP)
The most noteworthy results of an immunocytochemical analysis with an anti-peptide antibody against A. thaliana ESP are a labelling of epidermal cells but not the guard cells in most tissues and labelling of large cells adjacent to the phloem that the authors identified as S-cells in leaves and stems (Burow et al. 2007b). Although no direct experimental comparison with myrosinase localization was performed, ESP seems therefore to localize to different (but sometimes adjacent) cells than the ones expressing myrosinase in A. thaliana. The localization of ESP to S-cells, the cells identified in stems and described as containing glucosinolates by Koroleva et al. (2000), indicates at least a partial overlap of ESP and glucosinolate localization in some tissues.
Epithiospecifier protein (ESP) is detected in cytosol and nucleus
Both immunocytochemical analysis with a peptide specific antibody against A. thaliana ESP (Burow et al. 2007b) and expression of a CaMV35S:GFP-ESP construct in A. thaliana protoplasts (Miao and Zentgraf 2007) show a cytoplasmic and nuclear localization of ESP. Hence, ESP does not only seem to be localized to different cells than myrosinase but also to different subcellular compartments than any of the other identified components of the myrosinase enzyme system.
So far the cellular localization of ESP has only been reported for A. thaliana, where ESP proteins seem to be present in epidermal cells and the glucosinolate-containing S-cells. These localizations are unexpected and intriguing, and confirmation by independent techniques would seem a necessity. In addition, the localization of ESP in other species such as B. napus is awaited with much expectation. Interestingly, ESP is detected in the cytoplasm and nucleus, which distinguishes it from the other proteins of the myrosinase enzyme system.
When taking into consideration all the known components of the hydrolysis part (glucosinolates, myrosinases, MBPs, MyAPs, ESP/TFP) of the glucosinolate-myrosinase system, it is quite clear from the data presented in this review that the system is complex and that care must be taken when generalisations are made. Indeed, a complete picture of the system is not yet available for any of the studied plants and the system even shows differences within each plant at the organ and tissue level. In addition, the system is not static but expression of its components is changing during the plant’s growth cycle and may be affected by additional factors (not described here). Some plants may not even possess all parts of the system (Thangstad et al. 2004). Whether this has an impact on the system is yet unknown as a function or role has not yet been identified for each component. It is still unknown if all members of the same protein/gene family have the same or similar activity and function. For some of the components there is not even evidence for a role in the myrosinase enzyme system yet.
Although different species seem to contain variations of the ‘mustard oil bomb’ with slightly different compositions and spatial distributions of the myrosinase-enzyme system, some general features can nevertheless be mentioned. Glucosinolates, myrosinases, MBPs and MyAPs seem to be present in vacuoles. Myrosinases are expressed in myrosin cells that are different but in close proximity of the cells accumulating glucosinolates. The expressions of MBPs and MyAPs do not seem to be restricted to one particular cell type.
‘Omics’-approaches in combination with single cell technology will undoubtedly be valuable tools in developing our understanding of the spatial distribution of the glucosinolate–myrosinase system. As we discussed before, the results provided by some proteomics studies contradict ‘established’ knowledge.
Possible roles and functions of these complexes and the myrosinase interacting proteins therein have been proposed (Rask et al. 2000) but so far experimental evidence is limited. To study the functional mechanism of these complexes their reassembly in vitro with recombinant or purified proteins or the extraction of intact complexes from plants (Bellostas et al. 2008) need to be considered.
If we want to take advantage on an economical scale of the plant defence aspects and human health benefits that are attributed to the glucosinolate–myrosinase system, we also need to address the question about how far this system in the model plant A. thaliana is equivalent to the one in agricultural crops such as B. napus. The knowledge about the spatial distribution of the hydrolysis part of glucosinolate-myrosinase in A. thaliana and other Brassicaceae that is described in this review seems to indicate that the information cannot be transposed unconditionally from the former to the latter.
Understanding the spatial distribution, from the organ down to the subcellular level, of the myrosinase enzyme system components is in our view a prerequisite in answering the questions about the biological and physiological relevance of this system. This has to be complemented by the glucosinolate biosynthetic pathway and its regulation, aspects that were out of scope of the current review.
This work was supported by the Biotechnology and Functional genomics (FUGE) programmes of the Norwegian Research Council and ERA-PG through grants NFR 182897, 185173 and 175691.
- Andersson D, Chakrabarty R, Zhang J et al (2004) A new myrosinase gene family in Arabidopsis thaliana. In: Abstracts of the 15th International Conference on Arabidopsis Research, Berlin, 11–14 July 2004Google Scholar
- Bones AM, Iversen T-H (1985) Myrosin cells and myrosinase. Isr J Bot 34:351–375Google Scholar
- Bones A, Slupphaug G (1989) Purification, characterization and partial amino acid sequencing of β-thioglucosidase from Brassica napus L. J Plant Physiol 134:722–729Google Scholar
- Burow M, Wittstock U (2008) Regulation and function of specifier proteins in plants. Phytochem Rev (this issue). doi:10.1007/s11101-008-9113-5
- Chen S, Collum RP, Sarry JE et al (2006) Vacuolar proteomics revealed important properties of the myrosinase-glucosinolate system. In: Abstracts of the First International Conference on Glucosinolates, Jena, 10–14 September 2006Google Scholar
- Fimognari C, Hrelia P (2007) Sulforaphane as a promising molecule for fighting cancer. Mutat Res–Rev Mut Res 635:90–104Google Scholar
- Foo HL, Gronning LM, Goodenough L et al (2000) Purification and characterisation of epithiospecifier protein from Brassica napus: enzymic intramolecular sulphur addition within alkenyl thiohydroximates derived from alkenyl glucosinolate hydrolysis. FEBS Lett 468:243–246. doi:10.1016/S0014-5793(00)01176-5 PubMedCrossRefGoogle Scholar
- Jost R, Altschmied L, Bloem E et al (2005) Expression profiling of metabolic genes in response to methyl jasmonate reveals regulation of genes of primary and secondary sulfur-related pathways in Arabidopsis thaliana. Photosynth Res 86:491–508. doi:10.1007/s11120-005-7386-8 PubMedCrossRefGoogle Scholar
- Lönnerdal B, Janson J-C (1973) Studies on myrosinases II. Purification and characterization of a myrosinase from rapeseed (Brassica napus L.). Biochim Biophys Acta 315:421–429Google Scholar
- Luthy B, Matile P (1984) The mustard oil bomb: rectified analysis of the subcellular organization of the myrosinase system. Biochem Physiol Pflanz 179:5–12Google Scholar
- Matile P (1980) “Die Senfölbombe”: Zur Kompartimentierung des Myrosinasesystems. Biochem Physiol Pflanz 175:722–731Google Scholar
- Wittstock U, Kliebenstein DJ, Lambrix V et al (2003) Glucosinolate hydrolysis and its impact on generalist and specialist insect herbivores. In: Romeo JT (ed) Integrative phytochemistry: from ethnobotany to molecular ecology. Recent advances in phytochemistry, vol 37. Elsevier, Amsterdam, pp 101–125CrossRefGoogle Scholar