Allelic Variation of MYB10 Is the Major Force Controlling Natural Variation in Skin and Flesh Color in Strawberry (Fragaria spp.) Fruit

Abstract

The fruits of diploid and octoploid strawberry (Fragaria spp) show substantial natural variation in color due to distinct anthocyanin accumulation and distribution patterns. Anthocyanin biosynthesis is controlled by a clade of R2R3 MYB transcription factors, among which MYB10 is the main activator in strawberry fruit. Here, we show that mutations in MYB10 cause most of the variation in anthocyanin accumulation and distribution observed in diploid woodland strawberry (F. vesca) and octoploid cultivated strawberry (F ×ananassa). Using a mapping-by-sequencing approach, we identified a gypsy-transposon in MYB10 that truncates the protein and knocks out anthocyanin biosynthesis in a white-fruited F. vesca ecotype. Two additional loss-of-function mutations in MYB10 were identified among geographically diverse white-fruited F. vesca ecotypes. Genetic and transcriptomic analyses of octoploid Fragaria spp revealed that FaMYB10-2, one of three MYB10 homoeologs identified, regulates anthocyanin biosynthesis in developing fruit. Furthermore, independent mutations in MYB10-2 are the underlying cause of natural variation in fruit skin and flesh color in octoploid strawberry. We identified a CACTA-like transposon (FaEnSpm-2) insertion in the MYB10-2 promoter of red-fleshed accessions that was associated with enhanced expression. Our findings suggest that cis-regulatory elements in FaEnSpm-2 are responsible for enhanced MYB10-2 expression and anthocyanin biosynthesis in strawberry fruit flesh.

Figure 1.

QTL-Seq Identifies a Significant Region on Fvb1 Controlling Fruit Color in RV660 × WV596. (A) 𝚫SNP-index plot (Y-axis) over the seven F. vesca chromosomes. Line represents a sliding window of 3 Mb at 10-kb intervals. Significant SNPs are highlighted in blue (P < 0.03). (B) FvMYB10 PCR amplification using primers R/W-F and R/W-R designed flanking the 52 bp INDEL (c.278_279ins). In RV660, the expected fragment size was 181 bp according to the reference Hawaii-4 genome. See scheme in (C). (C) Schematic representation of FvMYB10-gypsy, the Long Terminal Repeat Transposable Element (LTR-TE) identified in the WV596 FvMYB10 coding sequence. In the scheme, primers used for population genotyping are highlighted (R/W-F, R/W-R, and ccm1, the latter at the 5′ end of the LTR-TE; Supplemental Figure 1).

Figure 2.

Effect of FvMYB10 Mutation on Structural Genes and Metabolites in F. vesca. (A) Expression analysis (by quantitative RT-PCR) of FvMYB10 and anthocyanin structural genes in ripe fruits from RF and WF F₂ pools. For FvMYB10, primers were designed upstream (left) and downstream (right) of the LTR-TE insertion point (Supplemental Data Set 11). Means (±se) of biological triplicates, consisting of pools of 20 to 25 fruits each, are represented in the graph. Asterisks indicate significant differences, as determined by Student’s t test (*P < 0.05; ** P < 0.01; *** P < 0.005). (B) Secondary metabolites with significant differences in levels between RF and WF pools. Means (±se) of biological triplicates from RF or WF pools relative to WF pools are represented in the graph. Asterisks indicate significant differences, as determined by Student’s t test (*P < 0.05; ** P < 0.01). (C) Transient RV660 FvMYB10 overexpression restored anthocyanin biosynthesis in WV596 fruits. Three representative fruits are shown. As a control, fruits were agroinfiltrated with the truncated fvmyb10-2 gene, which failed to induce anthocyanin accumulation.

Figure 3.

Multiple FvMYB10 Variants Were Detected in a Panel of 11 White-Fruited F. vesca Ecotypes. (A) FvMYB10-gypsy insertion (fvmyb10-2) was not detected by PCR analysis in FvMYB10 from other white-fruited F. vesca ecotypes. (B) Schematic representation of FvMYB10 showing the three alleles identified in white F. vesca ecotypes. fvmyb10-2 and fvmyb10-3 were identified in this study, whereas fvmyb10-1 was previously described by Zhang et al. (2015) and Hawkins et al. (2016). (C) Outline of the protein products encoded by FvMYB10 alleles represented in (B) and list of the accessions in which they were found. The two repeats (R2R3) of the structurally conserved DNA binding domain are shaded in gray. The conserved motif KPRPR[S/T]F found in anthocyanin MYB regulators is shown in red. (D) FIN12 has a large deletion in Fvb1 resulting in the loss of 7 predicted genes (Supplemental Figure 3). FvMYB10 is among them and is highlighted in the scheme. Primers ccm52 and ccm49 were designed flanking the region in Hawaii-4. Both primers are 100 Kb apart, resulting in no PCR product. In FIN12, a >10 kb band was obtained. The picture on the right shows FvMYB10 CDS amplification with primers FvMYB10-F and FvMYB10-R only in Hawaii-4 but not in FIN12. (E) FvMYB10 transient overexpression assays showing complementation of GER2 and FIN12 white fruit. Two representative fruits are shown.

Figure 4.

SNPs and QTLs Associated with Fruit Color Traits in Octoploid Strawberry. (A) Manhattan plots of a GLM, MLM, and MLMM for white fruit in UF breeding population 17.66. Plots for different chromosomes are shown in different colors, which follow the order: chromosome 1-1 to chromosome 7-4. Green line represents the significance threshold. (B) Positions of QTLs on LG 1-2 controlling fruit color detected in the cv Senga Sengana × F. chiloensis ssp. lucida USA2 F₂ population. Thick and thin bars represent 1- and 2-LOD QTL intervals, respectively, and are drawn to the right of LG 1-2. Names of QTLs as described in Supplemental Data Set 4. Markers in the pick of the QTL and in the 2-LOD interval are highlighted in red and pink, respectively. The estimated position of FvMYB10 is indicated with an asterisk based on the position of flanking SNPs in F. vesca reference genome v.4. (C) LOD scores of flesh QTL are plotted along the x axis (dotted line indicates the 4.5 LOD threshold), while genetic distances are plotted along the y axis.

Figure 5.

MYB10 Promoter Variants Across Subgenomes and Accessions in Octoploid Fragaria spp. (A) Schematic representation of FvMYB10 showing the positions of primers used for PCR amplification of MYB10_pro from octoploid Fragaria spp (Supplemental Data Set 11). (B) MYB10_pro PCR amplification yielded different sized promoter alleles from each accession. White-fleshed accessions USA1, USA2, and FC157 contained an additional 2.1- or 2.8-kb band compared with red-fleshed ones. Red arrows point to bands that were excised from the gel and cloned for sequencing. (C) Sequences obtained from (B) were aligned with the respective sequences of the four FaMYB10 homoeologs from cv Camarosa retrieved from the reference genome and analyzed using the Neighbor-Joining method with 1000 bootstrap replicates for homology tree construction. Sequences were grouped in three main clades that revealed their chromosomal origin: (1) Magenta clade contains promoter sequences from all Fvb1-3A homoeologs; (2) green clade from Fvb1-3B and Fvb1-1 homoeologs, which are identical; and (3) in blue, all alleles from Fvb1-2. On the right, a schematic representation of a promoter allele representative of each clade is shown. Three different versions of Fvb1-2 MYB10_pro were found: (1) a 23 kb allele from cv Camarosa containing four large INDELs, (2) a 2.1 kb allele from USA2, and (3) a 2.8 kb allele from FC157. A color code was used for each homoeolog copy of MYB10_pro in the tree and their respective hallmark sequences in the schemes: Magenta for MYB10-3A_pro, green for MYB10-1_pro and MYB10-3B_pro (they are identical), and blue for MYB10-2_pro. Bootstrap values >50% are shown in each node. Scale bar indicates substitution rate of nucleotides per site. The branch length is proportional to the nucleotide substitution per site.

Figure 6.

The FaEnSpm-2 Insertion in the Upstream Regulatory Region of MYB10-2 Is Associated with Internal RF Color. Different mutations in the coding sequence lead to completely white fruits lacking anthocyanins in the epidermis. (A) FaEnSpm-2 cosegregates with red flesh color in the SS×FcL F₂ population. The Fvb1-2-specific codominant IFC-1 marker was developed using a combination of three primers. Primer ccm107 binds to the junction sequence of FaEnSpm-2 and together with ccm109 amplify a 317-bp band associated with the red-flesh phenotype, whereas the 645 bp product (no TE insertion) associates with the white flesh phenotype. Colored circles indicate flesh color: cream color = 1, orange = 2, and red = 3. Circles with two colors represent different scores from different seasons (Supplemental Figure 5). (B) IFC-1 marker in F. chiloensis accessions. Only the red-fleshed FC154 carried the FaEnSpm-2 TE at pMYB10-2. Among white-fleshed F. chiloensis, two different genotypes were found: (1) F. chiloensis ssp. lucida (USA1 and USA2) and pacifica (FC285) accessions both contain the smaller promoter allele version corresponding to the 2.1-kb pMYB10 variant cloned from USA2, whereas F. chiloensis ssp. chiloensis accessions (FC156, FC157, FC160, and FC187) harbor the larger 2.8-kb promoter allele identified in FC157. (C) Data from RNA-seq showing the expression level (FPKM value) of each MYB10 homoeolog in cv Senga Sengana and six white-fleshed F. chiloensis accessions. Expression level dominance is biased toward MYB10-2 in all accessions tested. (D) Scheme representing the fcmyb10-1 allele, which contains a G/T SNP in the coding sequence leading to a predicted truncated protein. Accessions homozygous for this variant in Fvb1-2, that is, FC187, are completely white. FC157 is heterozygous fcmyb10-1/FcMYB10 in chromosome Fvb1-2 (Supplemental Data Set 5) and has a light pink epidermis. The two repeats (R2R3) of the structurally conserved DNA binding domain are shaded in gray. The conserved motif KPRPR[S/T]F found in anthocyanin MYB regulators is shown in red.

Figure 7.

RNA-seq Expression Data of Structural Genes from Anthocyanin Biosynthetic Pathway in cv Senga Sengana and White-Fleshed F. chiloensis Accessions. Mean FPKM values from the biological replicates for each homoeolog of the same gene were added up, and the resulting value represented only if above 10 FPKM. To avoid naming all cv Camarosa homoeologs, each gene was labeled using the corresponding F. vesca gene code. Enzymes encoded by genes that are dramatically downregulated in all white-fleshed accessions are highlighted in red. FPKM values and the cv Camarosa gene code for each homoeolog are provided in Supplemental Data Set 6.

Figure 8.

Prediction of Putative Regulatory cis-Elements in the Promoter Region of cv Camarosa FaMYB10-2 Not Found in MYB10-2_pro from the White-Fleshed Accessions USA2 or FC157. Functional validation of MYB10 as the causal agent under the internal fruit color QTL. (A) Schematic representation of selected regulatory cis-elements exclusively predicted in the cv Camarosa MYB10-2_pro allele but not present in the white-fleshed accessions USA2 or FC157. Additional information on these elements can be found in Supplemental Data Set 8. (B) Complementation of the white flesh phenotype in F. chiloensis accessions by transient FvMYB10 overexpression. A representative fruit from each assay is shown.

Figure 9.

High-throughput Markers for Fruit Skin and Flesh Color. (A) HRM analysis of the WS_CID_01 marker associated with white fruit color using 102 breeding accessions. The blue and red peaks are associated with white and red color in strawberry, respectively. (B) Scatterplots of IFC-2 KASP assay results in octoploid strawberry accessions reveal the clustering of white-fleshed lines on the X- (FAM) axis (yellow dots), homozygous red-fleshed lines on the Y- (HEX) axis (blue squares), and heterozygous lines in between (green triangles).