Simultaneous Quantification of Organic Acids in Tamarillo (Solanum betaceum) and Untargeted Chemotyping Using Methyl Chloroformate Derivatisation and GC-MS

Abstract

Sixteen organic acids were quantified in peel and pulp of Amber, Laird's Large and Mulligan cultivars of tamarillo using GC-MS. Fourteen of these compounds had not previously been quantified in tamarillo. An untargeted metabolomics approach was used in parallel to identify and quantify 64 more metabolites relative to the internal standard, indicating abundances of glutamic acid, pro-line, aspartic acid and γ-aminobutyric acid as well as lower concentrations of several other essential fatty acids and amino acids. The main findings were that total organic acid concentration was significantly higher (p < 0.05) in pulp than in peel, with the highest concentration seen in Mulligan pulp (219.7 mg/g DW). Remarkably, after citric acid, the potent bactericide itaconic acid was the second most abundant organic acid. At least 95% of organic acids in tamarillo were one of these two acids, as well as cis-aconitic, malic and 4-toluic acids. Differences between cultivar chemotypes were as substantial as differences between tissues. These results suggest that the bitter flavour of the peel does not result from organic acids. The combination of targeted and untargeted metabolomics techniques for simultaneous qualitative and quantitative investigation of nutrients and flavours is efficient and informative.

Keywords: citric acid; fruit; itaconic acid; metabolite profiling; metabolomics; ripening.

Figure 1

Example TIC chromatograms for (a) a 200 mg/L organic acid standard, and (b) a sample of Mulligan tamarillo pulp overlaid above the EIC chromatograms for each of the sixteen organic acids quantified here. The name of the organic acid for each EIC is followed by the m/z of the ion used for quantification.

Figure 2

Boxplots showing the concentration of sixteen organic acids absolutely quantified in the peel and pulp of Amber, Laird’s Large and Mulligan tamarillo cultivars using MCF derivatization and GC-MS. The middle line of each box shows the median. Y-axis units are mg of organic acid per gram dry weight of tamarillo sample and are scaled logarithmically. Asterisks after the name indicate acids whose concentration differs significantly between cultivars and tissues. Single asterisks indicate p < 0.05 and triple asterisks p < 0.001.

Figure 3

Statistical analysis and visualization of the combined data set. (A) shows Principal Component Analysis of the data; (B) is a heat map showing relative differences in each metabolite across the different sample types. Hierarchical Clustering has been applied to both axes. Note that the sample clustering resolves the different cultivars and tissue types perfectly; (C) is a pie chart showing the proportion of different metabolite class identities annotated here.

Figure 4

Relative peak area of the twelve compounds that contributed to the greatest differences on Principle Component 1, so discriminating between tamarillo pulp and peel. Y-axis units are peak area relative to recovery of the d₄-alanine internal standard and are scaled logarithmically. The asterisks after the acid name indicates that the concentration of all compounds differed significantly between cultivars and tissues with p < 0.001.

Figure 5

Relative peak area of the twelve compounds that contributed to the greatest differences on Principle Component Two, and so discriminating between the tissues of the Laird’s Large cultivar and the Amber and Mulligan ones. Y-axis units are peak area relative to recovery of the d₄-alanine internal standard and are scaled logarithmically. Asterisks after the name indicate acids whose concentration differs significantly between cultivars and tissues. Single asterisks indicate p < 0.05, double indicates p < 0.01 and triple asterisks p < 0.001.