Exploiting natural variation of secondary metabolism identifies a gene controlling the glycosylation diversity of dihydroxybenzoic acids in Arabidopsis thaliana

Abstract

Plant secondary metabolism is an active research area because of the unique and important roles the specialized metabolites have in the interaction of plants with their biotic and abiotic environment, the diversity and complexity of the compounds and their importance to human medicine. Thousands of natural accessions of Arabidopsis thaliana characterized with increasing genomic precision are available, providing new opportunities to explore the biochemical and genetic mechanisms affecting variation in secondary metabolism within this model species. In this study, we focused on four aromatic metabolites that were differentially accumulated among 96 Arabidopsis natural accessions as revealed by leaf metabolic profiling. Using UV, mass spectrometry, and NMR data, we identified these four compounds as different dihydroxybenzoic acid (DHBA) glycosides, namely 2,5-dihydroxybenzoic acid (gentisic acid) 5-O-β-D-glucoside, 2,3-dihydroxybenzoic acid 3-O-β-D-glucoside, 2,5-dihydroxybenzoic acid 5-O-β-D-xyloside, and 2,3-dihydroxybenzoic acid 3-O-β-D-xyloside. Quantitative trait locus (QTL) mapping using recombinant inbred lines generated from C24 and Col-0 revealed a major-effect QTL controlling the relative proportion of xylosides vs. glucosides. Association mapping identified markers linked to a gene encoding a UDP glycosyltransferase gene. Analysis of Transfer DNA (T-DNA) knockout lines verified that this gene is required for DHBA xylosylation in planta and recombinant protein was able to xylosylate DHBA in vitro. This study demonstrates that exploiting natural variation of secondary metabolism is a powerful approach for gene function discovery.

Keywords: Arabidopsis; UDP glycosyltransferase (UGT); dihydroxybenzoic acid glycosides; natural variation; secondary metabolism.

Figure 1

Comparison of leaf metabolite profile of Col-0 and C24. HPLC chromatograms of leaf extract from Col-0 and C24 clearly show the differential accumulation of compounds 1 and 2 (indicated by arrowheads) between these two accessions.

Figure 2

Compounds 1 and 2 are resistant to base hydrolysis but labile to acid hydrolysis. The two compounds were subjected to no treatment, base hydrolysis, or acid hydrolysis, and subsequently analyzed by LC-MS as described in Materials and Methods. The m/z value of the molecular ion for a compound is shown to the right.

Figure 3

Four DHBA glycosides identified in this study. (A) HPLC chromatograms of leaf extract from C24 and selected Col-0 × C24 F₂ plants, showing the retention times of the four DHBA glycosides. (B) UV spectra of the four DHBA glycosides. (C) Chemical structures of the four DHBA glycosides.

Figure 4

Accumulation patterns of DHBA glycosides in Col-0 × C24 F₂ population. The levels of different DHBA glycosides accumulated in 290 individual F₂ plants were plotted against each other. (A) The 2,3-DHBAG vs. 2,5-DHBAG; (B) 2,3-DHBAX vs. 2,5-DHBAX; (C) 2,3-DHBA glycosides (sum of 2,3-DHBAG and 2,3-DHBAX) vs. 2,5-DHBA glycosides (sum of 2,5-DHBAG and 2,5-DHBAX); (D) 2,5-DHBAX vs. 2,5-DHBAG; (E) 2,3-DHBAX vs. 2,3-DHBAG; (F) DHBAX (sum of 2,3-DHBAX and 2,5-DHBAX) vs. DHBAG (sum of 2,3-DHBAG and 2,5-DHBAG); and (G) distribution pattern of the relative abundance value of DHBAX in the F₂ population.

Figure 5

QTL analysis of molar percentage of DHBA xylosides from the Col-0 × C24 RIL population. The y-axis corresponds to the negative log₁₀ of the P-value as determined by MQM mapping in the R/QTL package. The x-axis indicates the chromosome, and the positions of each marker are indicated by the upward tick marks on the axis line. Only one position, at the top of chromosome 5, exceeded a 1000-permutation-estimated threshold of 0.05.

Figure 6

Identification and localization of genetic variation responsible for variation in the xyloside proportion of DHBA glycosides. Association tests were carried out using a genome-wide collection of 177,088 SNPs and DHBA glycoside measurements from 96 accessions of A. thaliana. Each point represents an individual SNP. The y-axis is the negative log of the P-value for the test for association between the segregation of that SNP and the proportion of DHBA xylosides present in leaf tissue (see Materials and Methods). The x-axis corresponds to the position of each SNP in the genome. The five chromosomes of A. thaliana are indicated by alternating black and gray color for each. The red and blue horizontal lines correspond to P-values of 10⁻⁵ and 10⁻⁷, respectively.

Figure 7

Col-0 UGT89A2 has a 2,5-DHBA xylosyltransferase activity. Crude protein extract from E. coli expressing Col-0 UGT89A2 or harboring an empty pET30a vector was tested for 2,5-DHBA xylosyltransferase activities using a HPLC-based assay.

Figure 8

Disruption of UGT89A2 in Col-0 results in accumulation of 2,5-DHBAG instead of 2,5-DHBAX. (A) Schematic representation of the positions of the T-DNA insert in two independent UGT89A2 knockout lines, SALK_147583C and SALK_081110C. (B) HPLC chromatograms of the leaf extract from Col-0 and the two T-DNA knockout lines.