2.1. Transgenic Arabidopsis synthesises benzoxazinoids Previous experiments to transfer GDIBOA biosynthesis to Arabidopsis by first generating individual Bx -gene transgenics and consecutive merging by crossing were not successful. Since in this approach the six essential genes of the pathway (Fig. 1) were driven by the cauliflower mosaic virus 35S promoter (p35S) gene silencing caused by the multiple use of the strong promoter might have occurred (Mlotshwa et al., 2010). To promote gene expression we employed Arabidopsis promoters and changed the codon usage of the G/C-rich maize genes to fit to the A/T preference of Arabidopsis. For the choice of promoters the criteria were moderate to high expression in rosette leaves and low expression in flower organs, the latter based on the finding that joining of Bx1 and Bx2 both expressed as p35 S-constructs generated female and male sterile plants (Supplemental Figure S12B). According to the Arabidopsis eFP browser developmental map (Winter et al., 2007) P450 genes of glucosinolate biosynthesis, Cyp71B7, Cyp79B2, Cyp83A1, and Cyp83B1 (SUPERROOT2) displayed a suitable expression pattern and the cDNA of the Bx- genes were integrated between the intrinsic start and stop codons (Supplemental Figure S1). To reduce the number of required transformation events, the genes Bx2 to Bx5 were merged in one T-DNA and the resulting transgenics were termed Cluster (the nomenclature of the transgenics is summarised Table 1). As another precautionary measure to guarantee efficiency of the maize P450s we isolated the maize gene model Zm00001d026483 that is annotated as P450 oxidoreductase (named ZmPor 2 in the following). According to the maize eFP browser (Stelpflug et al., 2016; Winter et al., 2007), ZmPor2 is highly expressed in seedlings. In the transgenics the gene is driven by p35S. The enzymatic activity of P450s in the transgenics was verified with isolated microsomes (Table 2A). For the substrates ION, and HION consecutive reactions yielding HBOA and DIBOA, respectively were detected (Table 2A). By crossing and selfing we obtained transgenic plants that feature in addition to the Cluster genes the UDPG: DI(M) BOA-glucosyltransferase Bx8 and ZmPor2, each transgene homozygously (Cluster+, Table 1). To complete the pathway Cluster + plants were crossed with transgenics harbouring Bx1 driven by promoters conferring different levels of expression, the strong promoters p35S and pSUR2 and the weak promoter pNos (Supplemental Figure S2A). We could verify GDIBOA biosynthesis in Arabidopsis when Bx1 was driven by pSUR2 or p35S. Quantification by LC-MS revealed concentrations of 3.7±0.7 nmol GDIBOA/g DW (Table 3) for p35S:: Bx1Cluster +(Bx1C+) and pSUR2::Bx1Cluster +plants. The result shows that high levels of BX1 are required to initiate BX biosynthesis. The concentrations achieved in the transgenics however are far below the level reached in grasses (e.g. up to 20 000 nmol/g DW in rye leaves, Copaja et al., 2006; Rice et al., 2005) and also below the effective concentration for defence (1 μmol/g FW; Bravo et al., 1997; Bravo and Lazo, 1996; Campos et al., 1989; Long et al., 1975).

Published as part of Abramov, Aleksej, Hoffmann, Thomas, Stark, Timo D., Zheng, Linlin, Lenk, Stefan, Hammerl, Richard, Lanzl, Tobias, Dawid, Corinna, Schon, Chris-Carolin, Schwab, Wilfried, Gierl, Alfons & Frey, Monika, 2021, Engineering of benzoxazinoid biosynthesis in Arabidopsis thaliana: Metabolic and physiological challenges, pp. 1-15 in Phytochemistry (112947) (112947) 192 on page 3, DOI: 10.1016/j.phytochem.2021.112947, http://zenodo.org/record/10126052