A metabolic perspective of selection for fruit quality related to apple domestication and improvement

Abstract

Background: Apple is an economically important fruit crop. Changes in metabolism accompanying human-guided evolution can be revealed using a multiomics approach. We perform genome-wide metabolic analysis of apple fruits collected from 292 wild and cultivated accessions representing various consumption types.

Results: We find decreased amounts of certain metabolites, including tannins, organic acids, phenolic acids, and flavonoids as the wild accessions transition to cultivated apples, while lysolipids increase in the "Golden Delicious" to "Ralls Janet" pedigree, suggesting better storage. We identify a total of 222,877 significant single-nucleotide polymorphisms that are associated with 2205 apple metabolites. Investigation of a region from 2.84 to 5.01 Mb on chromosome 16 containing co-mapping regions for tannins, organic acids, phenolic acids, and flavonoids indicates the importance of these metabolites for fruit quality and nutrition during breeding. The tannin and acidity-related genes Myb9-like and PH4 are mapped closely to fruit weight locus fw1 from 3.41 to 3.76 Mb on chromosome 15, a region under selection during domestication. Lysophosphatidylethanolamine (LPE) 18:1, which is suppressed by fatty acid desaturase-2 (FAD2), is positively correlated to fruit firmness. We find the fruit weight is negatively correlated with salicylic acid and abscisic acid levels. Further functional assays demonstrate regulation of these hormone levels by NAC-like activated by Apetala3/Pistillata (NAP) and ATP binding cassette G25 (ABCG25), respectively.

Conclusions: This study provides a metabolic perspective for selection on fruit quality during domestication and improvement, which is a valuable resource for investigating mechanisms controlling apple metabolite content and quality.

Keywords: Apple; Fruit weight; Metabolome; Quality; Storage; Taste; mGWAS.

Fig. 1

Genome-wide relationship in cultivated apple and wild species. a Geographic distribution of apple accessions. The circle represents the sample size. b Principal component analysis (PCA) of apple accessions. c The phylogenetic tree and population structure of apple accessions. “RJ” represents “Ralls Janet,” “SD” represents “Starking Delicious,” “GD” represents “Golden Delicious” and “Jona” represents “Jonathan.” The accession orders and positions on the X-axis are the same. d Linkage disequilibrium (LD) decay patterns in M. sieversii, M. sylvestris, other wild, heirloom, and cultivar groups. e Nucleotide diversity (π) between different groups. Lowercase letters indicate the significant differences based on Fisher’s least significant difference (LSD) test at p ≤ 0.01

Fig. 2

Profiling of metabolites among different apple accessions. a PCA analysis of all metabolites in other wild accessions, M. sieversii, heirlooms, and cultivars. b Differentiation of metabolites between M. sieversii/ heirloom and heirloom/ cultivar comparisons. c Box plots for the contents of flavonoids (quercetin 7-O-rutinoside, quercetin-3-O-rutinoside, and quercetin-3-O-robinobioside) and phenolic acids (isochlorogenic acid A, isochlorogenic acid B, and isochlorogenic acid C) with large fold changes between M. sieversii and heirloom groups. d PCA analysis between the “Ralls Janet” and “Golden Delicious” pedigrees, and e between high-weight (≥ 300) and low-weight (≤ 90) domesticated apples. t[1] and t[2] represent the first and second principal component. R2x[1] and R2x[2] indicate the proportion of variance explained by t[1] and t[2], respectively. The colored dots indicate different apple accessions and the ellipses indicate the 95% confidence interval for each data set

Fig. 3

Genome distribution of significant SNPs related to all metabolites. The circles from outside to center represent amino acids and nucleic acids (a), alkaloids (b), flavonoids (c), lipids (d), organic acids (e), others (f), phenolic acids (g), soluble tannins (h), and unknown metabolites (i). The unit for chromosome length is Mb. The P-value of all significant SNPs associated with metabolites are listed in Additional file 2: Table S9

Fig. 4

Identifying metabolites and genes potentially involved in apple domestication. a The contents of tannins with significant differential accumulation in M. sieversii/ heirloom and heirloom/ cultivar comparisons. b Manhattan plots of procyanidin A2 by GWAS. c A phylogenetic tree for MD15G1051400 and other R2R3 MYB members. d,e qRT-PCR analysis of Myb9-like and LAR (leucoanthocyanidin reductase) in high- and low-procyanidin content apples. f,g Results of transient overexpression of Myb9-like and LAR in apple fruits than empty vector (pGreenII 62-SK) injections. h Nucleotide polymorphism identified in the promoter of LAR. i Box plot for procyanidins contents correlated with C and T in Chr16:3404850. j qRT-PCR analysis of LAR in Myb9-like overexpressing apple fruit. k GUS transactivation assay in apples. The GUS reporter vector contains the LAR promoter, while the effector vector contains Myb9-like. l A simplified metabolic biosynthetic pathway of procyanidins in fruit. ANR, anthocyanidin reductase; Cys-EC, 4β-(S-cysteinyl)-epicatechin. Data are expressed as mean ± sd. “*” indicate p ≤ 0.05 by t-test

Fig. 5

Functional interpretation of GWAS for lysophosphatidylethanolamine (LPE) 18:1 content in apple. a Lysolipid contents of apples in the “Ralls Janet” (RS) and “Golden Delicious” (GD) pedigrees. b Manhattan plots of LPE 18:1 content by GWAS. c An unrooted phylogenetic tree for MD12G1057600 and Arabidopsis FAD members. d The firmness of apples in “RS” and “GD” pedigrees. e Correlation analysis (R²) between firmness and LPE 18:1 in domesticated apples. f qRT-PCR analysis of FAD2 transcripts in domesticated fruits with high and low LPE 18:1. g Transient overexpression of FAD2 in apple fruits. h,i The firmness and respiration rate of apples treated with LPE 18:1 during storage. j Nucleotide polymorphisms identified in the promoter and coding sequence of FAD2. k Box plots of LPE 18:1 correlated with identified nucleotide polymorphisms. l Nucleotide polymorphisms in the “RS” and “GD” pedigrees. The data are expressed as mean ± sd. “*” indicates p ≤ 0.05 by t-test. The correlation was evaluated based on the Pearson correlation coefficient

Fig. 6

Functional analysis of GWAS for SA and ABA content related to apple fruit weight. Fruit weight (a), SA (b), and ABA (c) contents in high- (H) and low- (L) weight domesticated fruits. d,e Manhattan plots of SA and ABA contents by GWAS. f An unrooted phylogenetic tree was constructed for MD15G1051400 and other NAC members. Fruit weight (g), SA (h), and ABA (i) contents during fruit development. j,k qRT-PCR analysis of NAP (MD16G1069500) and ABCG25 (MD03G1222900) in H and L apple groups. l,m Results of transient overexpression of NAP and ABCG25 in apple fruits. n Nucleotide polymorphisms identified in the promoter and coding sequence of NAP. o Box plot of SA and ABA correlated with identified nucleotide polymorphism. p Nucleotide polymorphisms in H and L apple groups. DAFB, days after full bloom. Data are expressed as mean ± sd. “*” indicate p ≤ 0.05 by t-test