A genome-wide metabolomic resource for tomato fruit from Solanum pennellii

Abstract

Tomato and its processed products are one of the most widely consumed fruits. Its domestication, however, has resulted in the loss of some 95% of the genetic and chemical diversity of wild relatives. In order to elucidate this diversity, exploit its potential for plant breeding, as well as understand its biological significance, analytical approaches have been developed, alongside the production of genetic crosses of wild relatives with commercial varieties. In this article, we describe a multi-platform metabolomic analysis, using NMR, mass spectrometry and HPLC, of introgression lines of Solanum pennellii with a domesticated line in order to analyse and quantify alleles (QTL) responsible for metabolic traits. We have identified QTL for health-related antioxidant carotenoids and tocopherols, as well as molecular signatures for some 2000 compounds. Correlation analyses have revealed intricate interactions in isoprenoid formation in the plastid that can be extrapolated to other crop plants.

Figure 1. Metabolite changes associated with the altered colour phenotype of IL3-2.

(A) Carotenoid profiles, obtained by HPLC-UV/Vis analysis, recorded at 450 nm; (i), M82; (ii) IL3-2. The chromatographic peaks and UV/Vis spectra are 1-lutein, 2-β-carotene and 3-lycopene. (B) Quantitative changes associated with pathway components. The position of the proposed pathway block (the fruit- specific phytoene synthase-1) is indicated. (C) Physical map of chromosome 3, showing the position of IL3-2 and Psy-1. (D) PCA of the metabolomic dataset for IL3-2; (i), scatter diagram of the score values, highlighting the separation based on variance in chemical composition between IL3-2 (brown dots) and the M82 control (red dots). The % contribution of each component to the variance is shown (ii), loadings plot of variables with the identity of some of the variables annotated. Only variables contributing to a significant difference (p-value of < 0.05) are shown.

Figure 2. Metabolite changes associated with the altered colour phenotype of IL6-3.

(A) Carotenoid profiles, obtained by HPLC-UV/Vis analysis, recorded at 450 nm; (i), M82; (ii), IL6-3. The chromatographic peaks and UV/Vis spectra are 1-lutein, 2-β-carotene and 3-lycopene. (B) Quantitative changes associated with pathway components. The position of the proposed step in the pathway up-regulated (the fruit specific lycopene β-cyclase, CYC-B) is indicated. (C) Physical map of chromosome 6, showing the position of IL6-3 and CYC-B. (D) PCA of the metabolomic dataset for Il6-3; (i) scatter diagram of the score values, highlighting the separation based on variance in chemical composition between IL6-3 (brown dots) and the M82 control (red dots). The % contribution of each component to the variance is shown (ii), loadings plot of variables with the identity of some of the variables annotated. Only variables contributing to a significant difference (p-value of < 0.05) are shown.

Figure 3. Metabolite changes associated with the altered colour phenotype of IL12-2.

(A) Carotenoid profiles, obtained by HPLC-UV/Vis analysis, recorded at 450 nm; (i), M82; (ii), IL12-2. The chromatographic peaks and UV/Vis spectra are 1-lutein, 2-β-carotene and 3-lycopene. (B) Quantitative changes associated with pathway components. The position of the proposed up-regulation in the pathway (lycopene ε-cyclase, Lyc-E) is indicated. (C) Physical map of chromosome 12, showing the position of Il12-2 and Lyc-E. (D) PCA of the metabolomic dataset for IL12-2; (i), scatter diagram of the score values, highlighting the separation based on variance in chemical composition between IL12-2 (brown dots) and the M82 control (red dots).

Figure 4. The identification of candidate genes from the metabolomic dataset.

(A) Changes in molecular features (metabolites) associated with Chr8 ILs. Each green dot is an IL, red dot a molecular feature (putative metabolite). The black connecting lines show an increase, the blue a decrease. The cluster of features circled represents significant metabolite changes common to both 8-2 and 8-2-1, of which a reduction in γ-tocopherol and increase in α-tocopherol was observed. (B) LC-MS profile validating the identity of tocopherols. (C) Quantitative changes (± s.e.) in tocopherols (γ and α) found in IL8-2 and IL8-2-1, relative to their M82 comparator. (D) Co-localization of the candidate gene with the change in tocopherol metabolites in IL8-2 and 8-2-1, as well as the position of the candidate gene product within the tocopherol biosynthetic pathway.

Figure 5. A correlation network constructed from metabolites involved in chloroplast isoprenoid biosynthesis.

Pearson coefficients (r 0.6 to 0.8 with a p-value significance < 0.05) derived from the metabolite levels associated with all 76 ILs of the S. pennellii collection. Metabolites associated with the Calvin cycle are shown as blue dots, the chloroplast located 2-C-methyl-D-erythritol-4-phosphate (MEP) pathway purple dots, chlorophyll degradation green dots, phytyl synthesis khaki coloured, tocopherol red coloured, prenyl lipid biosynthesis yellow, and carotenoid formation orange. Green connecting lines represent positive correlations and red connecting lines negative. Putative hubs derived from the number of connections are circled and the putative metabolites annotated. Collectively, the networks' strengths varied from 0.4 to 0.8 with the average being 0.45. Abbreviations: GGPP, geranylgeranyl diphosphate; IPP, isopentenyl diphosphate; DMAPP, dimethylallyl diphosphate; MEP, methylerythritol pathway.