Changing substrate specificity and iteration of amino acid chain elongation in glucosinolate biosynthesis through targeted mutagenesis of Arabidopsis methylthioalkylmalate synthase 1

Abstract

Methylthioalkylmalate synthases catalyse the committing step of amino acid chain elongation in glucosinolate biosynthesis. As such, this group of enzymes plays an important role in determining the glucosinolate composition of Brassicaceae species, including Arabidopsis thaliana Based on protein structure modelling of MAM1 from A. thaliana and analysis of 57 MAM sequences from Brassicaceae species, we identified four polymorphic residues likely to interact with the 2-oxo acid substrate. Through site-directed mutagenesis, the natural variation in these residues and the effect on product composition were investigated. Fifteen MAM1 variants as well as the native MAM1 and MAM3 from A. thaliana were characterised by heterologous expression of the glucosinolate chain elongation pathway in Escherichia coli Detected products derived from leucine, methionine or phenylalanine were elongated with up to six methylene groups. Product profile and accumulation were changed in 14 of the variants, demonstrating the relevance of the identified residues. The majority of the single amino acid substitutions decreased the length of methionine-derived products, while approximately half of the substitutions increased the phenylalanine-derived products. Combining two substitutions enabled the MAM1 variant to increase the number of elongation rounds of methionine from three to four. Notably, characterisation of the native MAMs indicated that MAM1 and not MAM3 is responsible for homophenylalanine production. This hypothesis was confirmed by glucosinolate analysis in mam1 and mam3 mutants of A. thaliana.

Keywords: Escherichia coli; biotechnology; enzyme activity; structural biology.

Figure 1. Schematic view of the chain elongation pathway in glucosinolate biosynthesis

The first step is deamination by a branched-chain aminotransferase (BCAT), followed by a three-step cycle: condensation by a MAM, isomerisation by an isopropylmalate isomerase (IPMI) and oxidative decarboxylation by an isopropylmalate dehydrogenase (IPMDH). At this point the intermediate will either go through another cycle or exit by transamination by a BCAT. The pathway intermediates are (1) amino acid, (2) 2-oxo acid, (3) 2-alkylmalic acid, (4) 3-alkylmalic acid and (5) chain-elongated amino acid.

Figure 2. Design of expression constructs containing the chain elongation pathway of A. thaliana

(A) The two constructs used in combination to express the biosynthetic genes. The MAM gene is replaced with native or mutated variants of MAM1 and MAM3 as well as IPMS2. Different T7 promoters were used as indicated in the arrows. Ribosomal binding sites, genes and terminators are indicated as dark grey, light grey and black boxes, respectively. (B) Promoters used in this work. Sequence differences are highlighted.

Figure 3. Homology model of MAM1 from A. thaliana

(A) The model of MAM1 was based on the crystal structure of N. meningitis IPMS (PDB-ID 3RMJ). β sheets (purple) and α helixes (turquoise) comprise the (β/α)₈ catalytic barrel characteristic of IPMS/MAM family. The binding of IPMS substrate is shown in yellow. Green residues represent the five residues in MAM1 predicted to be within 8 Å of the substrate. Grey residues represent two His residues conserved in the IPMS/MAM family. (B) Simplified presentation of the residues in MAM1, which is predicted to be in close proximity of the substrate. Colours as in (A).

Figure 4. Levels of protein amino acid substrates in E. coli expressing the chain elongation pathway with different MAM variants

GA, GAG and GAM represent MAM variants with multiple mutations; GA, T257G/G259A; GAG, T257G/G259A/A290G; GAM, T257G/G259A/L186M. Levels of leucine, phenylalanine and valine are statistically unchanged between the constructs when comparing to the average mean of all cultures. Methionine levels vary between cultures, although the differences are small. Data represent average and standard deviation of three biological replicates each grown in three technical replicates.

Figure 5. Production of chain-elongated products in all MAM variants

Products are grouped according to precursor amino acid: methionine (purple), phenylalanine (grey) and leucine (green). Methionine-derived products include methionine elongated from one to six cycles, leucine-derived products include leucine elongated once or twice and phenylalanine-derived product is only elongated once. Table shows the percentage of total production within each group of compounds. Data represent average and standard deviation of three biological replicates each grown in three technical replicates.

Figure 6. Production of amino acids elongated with more than two methylene groups

MAM3 is the only MAM variant that produces five and six times elongated products. HexaHM levels are below the limit of quantification and thus not included in the graph. Data represent average and standard deviation of three biological replicates each grown in three technical replicates.

Figure 7. Production of HPhe in E. coli by MAM1 variants and the effect of MAM mutants on HPhe-derived glucosinolates in A. thaliana

(A) Production of HPhe in MAM variants with changed specificity toward phenylalanine. Data represent average and standard deviation of three biological replicates each grown in three technical replicates. Tukey’s HSD (Honestly Significant Difference) was used for statistical analysis. Variants marked with asterisk denotes significant changes: P-values: 0.01–0.05 = *; P-values: 0.001–0.01 = **; P-values: 0–0.001 = ***. (B) Glucosinolate analysis in A. thaliana ecotype Col-0 and knockout mutants mam1 and mam3. Data represent average and standard deviation of 3–7 biological replicates.

Figure 8. Protein levels in E. coli strains expressing the chain elongation pathway genes

(A) BCAT4, (B) MAM1, (C) LSU1, (D) MAM3, (E) SSU3, (F) IPMS2 and (G) IPMDH1. A representative peptide for each protein was used for relative quantification. Data represent average and standard deviation of three biological replicates.