Convergent domestication of bitter apples and pears by selecting mutations of MYB transcription factors to reduce proanthocyanidin levels

Abstract

Fruit domestication has long aimed to reduce bitterness, yet the molecular mechanisms behind this trait remain only partially understood. Wild apples and pears naturally accumulate high levels of bitter proanthocyanidins (PAs), also known as condensed tannins. In this study, a convergent domestication process was identified in both fruits, involving the selection of weak alleles of MYB transcription factors that regulate PA biosynthesis. In apples, domestication targeted the MYB-Tannin-Tamer (MdMYBTT) gene. A 411-base pair transposable element inserted into the third exon of this gene in cultivated varieties produced a truncated, non-functional protein unable to activate the PA biosynthetic gene Anthocyanidin Reductase 1 (ANR1). The resulting mdmybtt allele led to reduced PA levels and was fixed in domesticated apples through positive selection. Likewise, in pears, a 57-base pair insertion in the promoter of the MYBPA1 gene suppressed its expression in cultivated varieties, limiting PA production. This insertion created the mybpa1 allele, which was similarly fixed during pear domestication. These findings highlight a shared evolutionary strategy to reduce fruit bitterness by selecting mutations that suppress PA synthesis. These findings offer valuable insights into the molecular basis of domestication and inform breeding efforts to optimize both flavor and nutritional quality.

Keywords: Malus; Pyrus; Comparative genomics; Domestication; Proanthocyanidin; Selective sweep.

Fig. 1

Identification of the MYB genes with a potential role in apple domestication. a illustrates proanthocyanidin (PA) content in mature fruits of 31 wild accessions and 14 apple cultivars. b-d selection sweep analysis involving comparison of M. domestica with M. sieversii, M. sylvestris, and other wild Malus accessions. The horizontal red line indicates the top 5% Pi ratio. e a phylogenetic tree of MYBs identified through selection sweep analyses. f DNA sequence alignment of MD15G1051400 and its alleles identified from different reference genomes of four wild species (M. baccata, M. sieversii, M. sylvestris, and M. fusca) and eight M. domestica cultivars (Hanfu, Gala, Honeycrisp, WA38, Antonovka, Fuji, M9, and MM106). g a snapshot of IGV view demonstrates DNA and RNA sequence reads of M. sieboldii (GSI) and M. domestica “Huashuo” (HS) mapped to the GDDH13 reference genome. h PCR analysis of six wild accessions and three cultivars of apples revealed two DNA bands, with the upper bands containing the 411-bp insert. W58, M. sieversii; W3, M. baccata; W77, M. toringoides; W32, M. micromalus; W43, M. robusta; W36, M. prunifolia; A145, HS, and A214 are M. domestica. The following GenBank or Arabidopsis TAIR accession numbers or Gene ID in GDR were used: M. domestica MdMYB9 (MD08G1070700), MdMYB11 (MD09G1184000), MdMYB12 (MD15G1215500), MdMYBPA1 (MD07G1153200), MdMYB22 (MD03G1297100), MdMYB10 (MD09G1278600), MdMYB110a (MD17G1261000); Arabidopsis thaliana AtMYB012 (At2g47460), AtMYB113 (At1g66370), AtMYB075 (At1g56650), AtMYB090 (At1g66390); Vitis vinifera VvMYBPA2 (ACK56131), VvMYBPA1 (CAJ90831); Vaccinium uliginosum VuMYBPA1 (KR106180); Fragaria ananassa FaMYB9 (AFL02460); Populus trichocarpa PtMYB115 (Potri.002G173900.1)

Fig. 2

Overexpression of MdMYBTT and mdmybtt in tomato plants and apple callus. a photographs show one wild-type (WT) and six transgenic “Micro-Tom” tomato plants [OETT-6, OETT-15, and OETT-16 for MdMYBTT overexpression and OEtt-1, OEtt-7, and OEtt-11 for mdmybtt overexpression]. b relative transcript levels of MdMYBTT and mdmybtt were determined using qRT-PCR. c total PA levels were estimated using the HCl-vanillin method. d relative transcript levels of MdMYBTT and mdmybtt were assessed using qRT-PCR. e photographs show one WT and six transgenic callus lines of M. domestica “Wanglin” [T-1, T-2, and T-3 for MdMYBTT overexpression and t-1, t-7, and t-11 for mdmybtt overexpression. f photographs show DMACA staining of one WT and six transgenic callus lines. g total PA levels were quantified using the HCl-vanillin method. Error bars represent standard deviation of three biological replicates. Based on Fisher’s least significant difference (LSD) test, significant difference at the 0.05 level is indicated by different lowercase letters

Fig. 3

MdMYBTT binds to and activates the ANR1 promoter. a yeast cells co-transformed with the constructs named on the left were cultured on SD/–Leu/– aureobasidin A (AbA²⁰⁰) non-selective medium (left panel) and SD/–Leu/ + AbA²⁰⁰ selective medium (right panel), in a dilution series of 1, 10^–1, 10^–2, 10^–3, and 10^–4 (i-v). Yeast growth on selective medium demonstrates the physical interaction between MdMYBTT and mdmybtt with the MdANR (MD10G1311100) promoter. b-c a dual-luciferase reporter assay was performed by transforming Nicotiana benthamiana leaves with the combinations of two plasmid constructs (b) and analyzing the Luc/Ren ratio post-transformation (c). 778, pSAK778; ANR1, MdANR1 promoter; T, MdMYBTT, t, mdmybtt. d shows the 375-bp sequence upstream of MdANR1 ATG, containing two MYBCORE binding motifs (highlighted in orange). e, EMSA shows MdMYBTT and mdmybtt binding to the MYBCORE cis-regulatory elements of the MdANR1 promoter in vitro. The symbols + and – indicate the presence and absence of the corresponding proteins and probes. f, N. benthamiana leaves were co-transformed with the combinations of two plasmid constructs. 778, pSAK778; 375/375 m, pGreenII 0800-pANR1-375/375 m, T, pSAK778-MYBTT; t, pSAK778-mybtt. g the Luc/Ren ratio of the transformation experiments described in f. Error bars represent standard deviation of three biological replicates. Based on Fisher’s LSD test, significant difference at the 0.05 level is indicated by different lowercase letters

Fig. 4

Relationship between PA content in apple fruit flesh and MdMYBTT genotypes. a, PA content in the fruit flesh of 12 MdMYBTT homozygous (TT), 18 heterozygous (Tt), and 15 mdmybtt homozygous (tt) apple accessions. b-l relative amount of individual (b-k) and total (l) PAs in the fruit flesh of 4 MdMYBTT homozygous (TT), 13 heterozygous (Tt), and 136 mdmybtt homozygous (tt) apple accessions

Fig. 5

Mutation of the MYBPA1 promotor reduces PA content in pear fruit flesh. a DMACA staining of pear fruit collected from three wild accessions (Betulifolia-1, Calleryana, and Betulifolia-2) and three cultivars (Hongxiangsu, Dangshansuli, and Yali) at 120 days after full bloom (DAFB). b PA content in the fruit flesh of the three wild accessions and three cultivars at 30 and 120 DAFB. c, Selective sweep analysis of 32 wild accessions and 40 cultivars of Asian pears based on Pi values. The horizontal red line indicates the top 5% Pi ratio. d A phylogenetic tree of the MYBs identified from the selective analysis and MYBs with known functions. e MYBPA1 transcript levels (FPKM) in fruit flesh collected from two wild accessions (Betulifolia and Calleryana) and two cultivars (Dangshansuli and Hongxiangsu) at 30 and 120 DAFB. f Alignment of the DNA sequences of MYBPA1 downloaded from six different pear reference genomes. The gene IDs are derived from the reference genes of one wild (P. betulifolia “Shanxiduli”), three Asia pear cultivars (P. pyrifolia “Nijisseiki,” “Dangshansuli,” and “Yunhong No. 1”) and two Europe pear cultivars (P. communis “d’Anjou” and “Bartlett”). g, PCR analysis of three wild accessions and six cultivars of pear revealed two DNA bands, with the upper band containing the 57-bp insert. W8 and W9, P. betulifolia; W12, P. calleryana; C12, C14, and C15, P. pyrifolia; C249, C250, and C254, P. bretschneideri. h-j the promoter activities of MYBPA1 and mybpa1 were compared using a dual-luciferase reporter assay with two reporter constructs (h) transformation of N. benthamiana leaves (i), and estimation of the Luc/Ren ratio (j). k relative expression level of MYBPA1 in one WT and three transgenic callus lines (PA1-1, PA1-2, and PA1-3) of M. domestica “Wanglin”. l-m photographs show callus of one WT and three transgenic callus lines of M. domestica “Wanglin.” The callus in m was stained with DMACA. n Total PA levels were determined using the HCl-vanillin method. Error bars represent standard deviation of three biological replicates. Based on Fisher’s LSD test, significant difference at the 0.05 level is indicated by different lowercase letters