Identification of Climate and Genetic Factors That Control Fat Content and Fatty Acid Composition of Theobroma cacao L. Beans

Abstract

The main ingredients of chocolate are usually cocoa powder, cocoa butter, and sugar. Both the powder and the butter are extracted from the beans of the cacao tree (Theobroma cacao L.). The cocoa butter represents the fat in the beans and possesses a unique fatty acid profile that results in chocolate's characteristic texture and mouthfeel. Here, we used a linkage mapping population and phenotypic data of 3,292 samples from 420 progeny which led to the identification of 27 quantitative trait loci (QTLs) for fatty acid composition and six QTLs for fat content. Progeny showed extensive variation in fat levels and composition, with the level of palmitic acid negatively correlated to the sum of stearic acid, oleic acid, and linoleic acid. A major QTL explaining 24% of the relative level of palmitic acid was mapped to the distal end of chromosome 4, and those higher levels of palmitic acid were associated with the presence of a haplotype from the "TSH 1188" parent in the progeny. Within this region of chromosome 4 is the Thecc1EG017405 gene, an orthologue and isoform of the stearoyl-acyl carrier protein (ACP) desaturase (SAD) gene in plants, which is involved in fatty acid biosynthesis. Besides allelic differences, we also show that climate factors can change the fatty acid composition in the beans, including a significant positive correlation between higher temperatures and the higher level of palmitic acid. Moreover, we found a significant pollen donor effect from the variety "SIAL 70" which was associated with decreased palmitic acid levels.

Keywords: QTL; SNP; cacao; fat content; fatty acid composition; heritability; linkage mapping; weather.

Figure 1

Predicted FA biosynthesis pathway in cacao with candidate genes. Cocoa bean fat is comprised primarily of palmitic acid, stearic acid, and oleic acid. FAs are synthesized in plastids, prior to transport to the cytosol, then the endoplasmic reticulum for modification and lipid assembly. Candidate genes in other plant species have been identified, and fatty acid-related genes exist (Argout et al., 2011)

Figure 2

Harvest of MP01 progenies for FA analysis. (A) Distribution of number of MP01 progenies (y-axis) by month of harvest (x-axis). (B) Number of replicates (pods) per genotype (x-axis) harvested per harvest date (y-axis). (C) Skew of parental haplotype combinations (T1C1, T1C2, T2C1, and T2C2) based on harvest date, based on Tcm004s00992801 marker.

Figure 3

Distribution of FA content for MP01 progenies for both periods.

Figure 4

Negative correlation between palmitic acid (C16:0) and the sum of stearic acid (C18:0), oleic acid (C18:1), and linoleic acid (C18:2). Axes represent percent of either palmitic acid relative to total fat, or the sum of stearic, oleic, and linoleic acids relative to total fat.

Figure 5

QTLs from best linear unbiased predictors (BLUPs) calculated using the 12 sampling dates spanning August 2010 to November 2011. The -log₁₀ p is the statistical magnitude of the significance of the peak maker associated to the trait. The coordinates are based on the “Matina1-6” genome V1.1. The threshold of significance (red dash) is a multiple comparison adjustment at 3.688.

Figure 6

QTLs from the best linear unbiased predictors (BLUPs) from the first harvest season (6 sampling dates) spanning August 2010 to January 2011. The -log₁₀ p is the statistical magnitude of the significance of the peak maker associated to the trait. The coordinates are based on the “Matina1-6” genome V1.1. The threshold of significance (red dash) is a multiple comparison adjustment at 3.688.

Figure 7

QTLs from the best linear unbiased predictors (BLUPs) from the second harvest season (6 sampling dates) spanning May 2011 to November 2011. The -log₁₀ p is the statistical magnitude of the significance of the peak maker associated to the FA. The coordinates are based on the “Matina1-6” genome V1.1. The threshold of significance (red dash) is a multiple comparison adjustment at 3.688.

Figure 8

Neighbor joining tree for haplotypes from 200 diversity panel individuals. Zoomed in section for the palmitic haplotype from TSH1188 H2 haplotype and the cluster with Iquitos haplotypes with support values are on the right figure.