Epistasis between genetic variations on MdMYB109 and MdHXK1 exerts a large effect on sugar content in apple fruit.

Many quantitative traits are controlled by multiple genetic variations with minor effects, making it challenging to resolve the underlying genetic network and to apply functional markers in breeding. Affected by up to a hundred quantitative trait loci (QTLs), fruit-soluble sugar content is one of the most complex quantitative traits in apple (Malus sp.). Here, QTLs for apple fruit sucrose and fructose content were identified via QTL mapping and bulked-segregant analysis sequencing (BSA-seq) using a population derived from a 'Jonathan' × 'Golden Delicious' cross. Allelic variations and non-allelic interactions were validated in the candidate genes within these defined QTL regions. Three single-nucleotide polymorphisms (SNPs) (SNP -326 C/T, SNP -705 A/G, and SNP -706 G/T) in the MdMYB109 promoter region affected the binding ability of the repressive transcription factor MdWRKY33, leading to increased MdMYB109 expression. MdMYB109 bound directly to the promoter of the sucrose transporter gene MdSUT2.2 and activated its expression, raising fruit sucrose content. A SNP (SNP1060 A/G) in the hexokinase gene MdHXK1 affected the phosphorylation of the transcription factor MdbHLH3, and phosphorylated MdbHLH3 interacted with MdMYB109 to co-activate MdSUT2.2 expression and increase fruit sucrose content. Adding the joint effects of the genotype combinations at the SNP markers based on the SNPs in MdMYB109 and MdHXK1 increased the prediction accuracy of a genomics-assisted prediction (GAP) model for total soluble solid content from 0.3758 to 0.5531. These results uncovered functional variations in MdMYB109 and MdHXK1 regulating apple fruit sucrose content. The updated GAP model with improved predictability can be used efficiently in apple breeding.