Biosynthesis, evolution, metabolic engineering and transport of Glucosinolates

Glucosinolates are amino acid-derived natural plant products found in the order Capparales, which includes the agriculturally important rape, food vegetables such as broccoli and various cabbages, and the model Arabidopsis. Glucosinolates are hydrolyzed by β-thioglucosidases (myrosinases) to produce different products such as e.g. isothiocyanates, nitriles and thiocyanates, which have a wide range of biological activities. The glucosinolate/myrosinase system protects plants against herbivore attacks, and is implicated in host-plant recognition by specialized predators.

 

The presence of glucosinolates in the agricultural important rape (Brassica napus) is of economic importance as glucosinolates have been shown to reduce the feeding quality of rape seed meal. On the contrary, in some Brassica vegetables (e.g. cauliflower, Brussels sprouts, broccoli) the presence of glucosinolates has been shown to have cancer-preventive properties. Identification of the biosynthetic genes and their regulators provides molecular tools, which enable us to metabolically engineer plants with new glucosinolate profiles. This approach has great potential for the design of 'biotech Brassica crops' with improved pest resistance and increased nutritional value.

 

 
Figure 1. The presence of glucosinolates has many biological effects ranging from detrimental to highly desirable.
  

Biosynthesis
To date, more than 100 different glucosinolates have been identified. They are derived from protein amino acids and a number of chain-elongated methionines and phenylalanine. The high diversity is obtained by secondary modifications of the side chain and/or the glucose moiety. The synthesis of the core structure takes its starting point in an amino acid that is converted into an aldoxime and proceeds via an oxidation and conjugation with cysteine to an S-alkyl-thiohydroximate. The C-S bond is subsequently cleaved to form a thiohydroximic acid which undergoes glycosylation to form a desulphoglucosinolate that is finally sulfated to form the glucosinolate. All genes in the pathway of the core structure have been identified with the exception of the sulfur donor step. Interesting sulfur chemistry catalyzed by a tightly coupled S nanomachine is involved in synthesizing the sulfur-rich glucosinolates. The products from the aldoxime-metabolizing CYP83A1 and CYP83B1 have to conjugate, with a sulfur donor (likely cysteine) to produce the next intermediate, a cysteine-conjugate, that is the substrate for the C-S lyase. Our working hypothesis is that this step is catalyzed by a glutathione-S-transferase-type of enzyme. We have undertaken several approaches including bioinformatics, use of tagged enzymes introduced into the knockout backgrounds to identify the interacting sulfur-donating enzyme. Analogous S nanomachines are expected to be involved in the biosynthesis of camalexin and other indole phytoalexins.

Evolution
Glucosinolates are evolutionarily related to cyanogenic glucosides as both groups of natural plant products are derived from amino acids that are converted into oximes by cytochromes P450 that belong to the CYP79 family. Furthermore, CYP83B1 from Arabidopsis thaliana has been identified as the oxime-metabolizing enzyme in the biosynthetic pathway of glucosinolates by the combined use of a biochemical and a bioinformatics approach. Searching through the genome of A. thaliana for homologues of CYP71E1 (P450ox), the only known oxime-metabolizing enzyme in the biosynthetic pathway of cyanogenic glucosides, we identified CYP83B1 as the oxime-metabolizing enzyme in the glucosinolate pathway as evidenced by characterization of the recombinant protein expressed in Escherichia coli. The data are consistent with the hypothesis that the oxime-metabolizing enzyme in the cyanogenic pathway (P450ox) was mutated into a 'P450mox' that converted the oxime into a toxic compound that the plant detoxified into a glucosinolate. 

 

Metabolic engineering
There is a strong interest in altering levels of specific glucosinolates in crop plants as certain glucosinolates have desirable and others have undesirable properties. The identification and characterization of the Arabidopsis cytochromes P450 belonging to the CYP79 family, have provided an important tool for modulating the profile of glucosinolates derived from protein amino acids. Introduction of exogenous CYP79s have furthermore enabled us to produce new glucosinolates in Arabidopsis, as exemplified by a mustard- and capers-flavored Arabidopsis overexpressing the cyanogenic CYP79A1 and CYP79D2 from sorghum and cassava, respectively. Overexpression, downregulation, and knockouts of endogenous, possibly combined with the introduction of exogenous CYP79s makes it possible to generate custom designed glucosinolate profiles. This may ultimately result in crop plants with improved tolerance to pests, improved nutritional values including improved taste and improved cancer-preventing effects.

 
Figure 2. Schematic view of metabolic engineering glucosinolate profiles based on CYP79 derived from plants producing either cyanogenic glucosides or glucosinolates.
 
  

IAOx is a branching point between indole glucosinolates, camalexin and IAA
We have shown that superroot1 (SUR1) is a C-S lyase in glucosinolate biosynthesis. This is evidenced by selective metabolite profiling of sur1, which is completely devoid of aliphatic and indole glucosinolates and by accumulation of the C-S lyase substrate following in vivo feeding with radiolabelled p-hydroxyphenylacetaldoxime to the sur1 mutant. C-S lyase activity of recombinant SUR1 heterologously expressed in Escherichia coli was demonstrated using the C-S lyase substrate djenkolic acid. The abolishment of glucosinolates in sur1 shows that the SUR1 function is not redundant and thus that SUR1 constitutes a single gene family. The ´high-auxin´ phenotype of sur1 is caused by accumulation of endogenous C-S lyase substrates as well as aldoximes including indole-3-acetaldoxime that is channeled into the main auxin indole-3-acetic acid. Thereby sur1 partly resembles the sur2/cyp83b1 mutants, which are blocked in the IAOx-metabolizing step in biosynthesis of indole glucosinolates. 

 

We have shown that camalexin, the indole phytoalexin of Arabidopsis thaliana, is synthesized from tryptophan via indole-3-acetaldoxime (IAOx) in a reaction catalyzed by CYP79B2 and CYP79B3 (Glawischnig et al., 2004, PNAS). This is evidenced by the observation that cyp79B2/cyp79B3 double knockout mutant is devoid of camalexin, as it is also devoid of indole glucosinolates (Zhao et al., Genes. Dev., 16, 3100-3112, 2002), and by incorporation of isotope labeled IAOx into camalexin. This shows that only CYP79B2 and CYP79B3 contribute significantly to the IAOx pool from which camalexin and indole glucosinolates are synthesized. Furthermore, production of camalexin in the sur1 mutant devoid of glucosinolates excludes that camalexin is derived from indole glucosinolates. CYP79B2 plays an important role as transcript level of CYP79B2, but not CYP79B3, is increased upon induction of camalexin by silver nitrate as evidenced by microarray analysis and promoter-GUS data. The structural similarity between cruciferous indole phytoalexins suggests that these compounds are biogenetically related and synthesized from tryptophan via IAOx by CYP79B homologues. The data show that IAOx is a key branching point between several secondary metabolic pathways as well as primary metabolism, where IAOx previously has been shown to play a critical role in IAA homeostasis. We are currently taken several independent approaches to identify the IAOx-metabolizing enzymes in camalexin and auxin biosynthesis.

 
 

Figure 3. Schematic view of the important role of IAOx as key branching point between primary and secondary metabolism.
  

 

Identification of regulators of IAOx and indole glucosinolate biosynthesis
IAOx constitutes an important branching point (see above). The production of IAOx must be tightly regulated for maintaining IAA homeostasis, and for controlling the flow into the different pathways. IAOx is synthesized from CYP79B2/B3. My group has undertaken a molecular genetics approach to identify regulators of IAOx production. We have developed a screen using a bacterial dehalogenase as negative selection marker (Næsted et al., Plant J. 1995) with the aim of identifying positive regulators of CYP79B3. The marker works under the given conditions. Selected mutants have been taken through secondary screen, and mapping populations are currently being generated.

 

 

 

Figure 4. A negative selection screen aimed at identifying activators of IAOx biosynthesis.

Identification of transporters of natural products
Very little is known about transport of natural products within plants. We use glucosinolates as a model system to address this question. Glucosinolates are known to be transported around in the plant, e.g. from silique walls to the seeds. We have characterized biochemically the uptake of glucosinolates in rape leaf protoplasts, and provided evidence for the existence of a glucosinolate-specific symporter on the plasmamembrane. Recently, we have combined Xenopus oocytes expression and a functional genomics approach to identify novel Arabidopsid transporters of small molecules such as e.g. natural products and hormones. We have used bioinformatics to define a library of membrane proteins with 10-14 transmembrane domains (TMs), characteristic for secondary organic solute transporters. Currently, we have 238 cDNAs out of 350 for which an EST exists. The library was successfully used to clone a novel glucose transporter, AtSTP13 (At5g26340). The method is a powerful alternative to isolation of transporters from Arabidopsis by traditional methods such as functional complementation in yeast. As an LC-MS method has been developed for detection of substrates taken up by Xenopus oocytes, the library can be used to screen for transporters of any desired substrate.
 

 
Figure 5. As a proof of concept, AtSTP13 was identified in a screen for glucose uptake using the normalized Arabidopsis expression library of full-length transporter cDNAs.
  

    
 
Group leader
Barbara Ann Halkier
Associate Professor, Dr. scient.
E-mail:
Phone: +45 35 28 33 42
Fax: + 45 35 28 33 33

 

Group members
Morten Nørholm, postdoc
Jing Li, postdoc
Bjarne Gram Hansen, Ph.D. student
Majse Nafisi, Ph.D. student
Hussam Auis, Ph.D. student
Fernando Geu Flores, Ph.D. student
Niels Bjerg Nielsen, Ph.D. student
Martin Nybo Andersen, master student
Morten Thrane Nielsen, master student
Lis Byrsting Møller, technician

 

Previous members
Ilona Jonuskiene, Ph.D.
Michael Dalgaard Mikkelsen, Ph.D.,
Erich Glawischnig, EMBO fellow
Peter Naur, MSc.
Sixue Chen, postdoc
Ute Wittstock, postdoc
Beate Dedicova, postdoc
Bent Larsen Petersen, postdoc
Carsten Hørlev Hansen, Ph.D.
Søren Bak, Ph.D.
Liangcheng Du, Ph.D.