A disulfide bond A-like oxidoreductase is a strong candidate gene for self-incompatibility in apricot (Prunus armeniaca) pollen

Abstract

S-RNase based gametophytic self-incompatibility (SI) is a widespread prezygotic reproductive barrier in flowering plants. In the Solanaceae, Plantaginaceae and Rosaceae gametophytic SI is controlled by the pistil-specific S-RNases and the pollen S-locus F-box proteins but non-S-specific factors, namely modifiers, are also required. In apricot, Prunus armeniaca (Rosaceae), we previously mapped two pollen-part mutations that confer self-compatibility in cultivars Canino and Katy at the distal end of chromosome 3 (M-locus) unlinked to the S-locus. Here, we used high-resolution mapping to identify the M-locus with an ~134 kb segment containing ParM-1-16 genes. Gene expression analysis identified four genes preferentially expressed in anthers as modifier gene candidates, ParM-6, -7, -9 and -14. Variant calling of WGS Illumina data from Canino, Katy, and 10 self-incompatible cultivars detected a 358 bp miniature inverted-repeat transposable element (MITE) insertion in ParM-7 shared only by self-compatible apricots, supporting ParM-7 as strong candidate gene required for SI. ParM-7 encodes a disulfide bond A-like oxidoreductase protein, which we named ParMDO. The MITE insertion truncates the ParMDO ORF and produces a loss of SI function, suggesting that pollen rejection in Prunus is dependent on redox regulation. Based on phylogentic analyses we also suggest that ParMDO may have originated from a tandem duplication followed by subfunctionalization and pollen-specific expression.

Keywords: DsbA oxidoreductase; M-locus; Prunus; gametophytic self-incompatibility; modifier; pollen-part mutation.

Fig. 1.

Genetic and physical maps of the apricot M-locus. (A) BAC clones from the self-incompatible cultivar Goldrich were used for de novo M-locus reference sequence assembly. BAC clones corresponding to the M₁- and M₂-haplotypes have a blue and yellow background, respectively. (B) Final aM-supercontig comprising major contigs aM-contig-1, -2, and -3 and the unresolved GAPs -13 and -15. Positions of peach (PGS3) and apricot (AGS) SSRs and SNPs are shown. (C) Genetic maps of the Canino and Katy M-loci were refined using new recombinants (left) and markers (dashed and dotted lines) delimiting a physical region of ~134 kb. (D) The ~134 kb region with the annotated apricot genes; green arrows, ParM-1–16. ParMDO corresponds to ParM-7. Grey arrows, ParM-17 to -19. ORFs identified in P. persica and P. mume but not supported by P. armenica expression data.

Fig. 2.

Gene expression analysis. (A) Heat map illustrating RNA-Seq differential expression data. Pairwise comparisons are shown for each apricot cultivar (columns). Blue, positive log fold-change (log FC) indicates higher expression in the first tissue compared with the second; red, negative log FC. Cultivars and RNA samples are as follows: G, Goldrich; C, Canino; K, Katy; Ant, Anther; St, Style; Lf, Leaf. Non-significant differences with P-values>0.001 are indicated (n.s.). (B) RT-PCR analysis comparing expression in different tissues from CC-77 (MM): Lf, leaf; Pt, petal; Ov, ovary; St, style; Po, pollen. Housekeeping genes actin and sand-like as well as style- and pollen-specific genes, S-RNase and SFB, respectively, were used as controls.

Fig. 3.

Identification of the variant causing SC in Canino and Katy pollen. (A) Variant filtering. SNVs and SVs within the aM-supercontig were called for Canino, Katy, and 10 self-incompatible cultivars (see Supplementary Table 1) and filtered as shown. (B) The insertional variant in ParM-7. ParM-7 is transcribed from right to left. Grey, coding exons; white, UTR regions; black, ParM-6 shown for reference; dashed line, 358 bp FaSt MITE insertion site. (C) ParM-7 coding sequence and predicted amino acid sequence for m- and M-alleles. FaSt MITE insertion (orange) leads to a premature stop codon (asterisk) in the m-allele. FaSt MITE target site duplications and 5′ and 3′terminal inverted repeats are shown in red and underlined, respectively. The dicysteine redox motif CPWC conserved in DsbA-like proteins is boxed. (D) PCR-genotyping of ParM-7 in mapping recombinants and self-compatible and self-incompatible apricot cultivars. Recombinants from G×C and K×K populations and homozygous controls CC-67 (mm) and CC-77 (MM) are included.

Fig. 4.

Ortholog analysis of ParMDO. (A) M-locus syntenic blocks between P. persica, M. domestica, S. lycopersicum, and A. thaliana. Black rectangles within circular genome regions represent gene annotation in scale. Red triangles indicate a scale change. Putative ParMDO and ParM-8 orthologs are shown. (B) Clustering analysis of plant DsbA-like proteins. Three main clusters group DsbA-like proteins from Rosaceae (blue dashed line), Solanaceae (red), and Brassicaceae (green). The Rosaceae cluster is subdivided into subgroups I (ParMDO, their Prunus orthologs, and the Fragaria gene04224v1.0 protein) and II (all the rest including ParM-8). Bootstrap values>50% with 1000 replications are shown on the branches. The scale bar indicates the number of amino acid substitutions per site.