Distyly is an intriguing floral adaptation that increases pollen transfer precision and restricts inbreeding. It has been a model system in evolutionary biology since Darwin. Although the Slocus determines the long-and short-styled morphs, the genes were unknown in Turnera. We have now identified these genes.We used deletion mapping to identify, and then sequence, BAC clones and genome scaffolds to construct S/s haplotypes. We investigated candidate gene expression, hemizygosity, and used mutants, to explore gene function.The s-haplotype possessed 21 genes collinear with a region of chromosome 7 of grape. The S-haplotype possessed three additional genes and two inversions. TsSPH1 was expressed in filaments and anthers, TsYUC6 in anthers and TsBAHD in pistils. Long-homostyle mutants did not possess TsBAHD and a short-homostyle mutant did not express TsSPH1.Three hemizygous genes appear to determine S-morph characteristics in T. subulata. Hemizygosity is common to all distylous species investigated, yet the genes differ. The pistil candidate gene, TsBAHD, differs from that of Primula, but both may inactivate brassinosteroids causing short styles. TsYUC6 is involved in auxin synthesis and likely determines pollen characteristics. TsSPH1 is likely involved in filament elongation. We propose an incompatibility mechanism involving TsYUC6 and TsBAHD.

Heterostyly employs distinct hermaphroditic floral morphs to enforce outbreeding. Morphs differ structurally in stigma/anther positioning, promoting cross-pollination, and physiologically blocking self-fertilization. Heterostyly is controlled by a self-incompatibility (S)-locus of a small number of linked S-genes specific to short-styled morph genomes. Turnera possesses three S-genes, namely TsBAHD (controlling pistil characters), TsYUC6, and TsSPH1 (controlling stamen characters). Here, we compare pistil and stamen transcriptomes of floral morphs of T. subulata to investigate hypothesized S-gene function(s) and whether hormonal differences might contribute to physiological incompatibility. We then use network analyses to identify genetic networks underpinning heterostyly. We found a depletion of brassinosteroid-regulated genes in short styled (S)-morph pistils, consistent with hypothesized brassinosteroid-inactivating activity of TsBAHD. In S-morph anthers, auxin-regulated genes were enriched, consistent with hypothesized auxin biosynthesis activity of TsYUC6. Evidence was found for auxin elevation and brassinosteroid reduction in both pistils and stamens of S- relative to long styled (L)-morph flowers, consistent with reciprocal hormonal differences contributing to physiological incompatibility. Additional hormone pathways were also affected, however, suggesting S-gene activities intersect with a signaling hub. Interestingly, distinct S-genes controlling pistil length, from three species with independently evolved heterostyly, potentially intersect with phytochrome interacting factor (PIF) network hubs which mediate red/far-red light signaling. We propose that modification of the activities of PIF hubs by the S-locus could be a common theme in the evolution of heterostyly.

Heterostyly is a breeding system that promotes outbreeding through a combination of morphological and physiological floral traits. In Turnera these traits are governed by a single, hemizygous S-locus containing just three genes. We report that the S-locus gene, BAHD, is mutated and encodes a severely truncated protein in a self-compatible long homostyle species. Further, a self-compatible long homostyle mutant possesses a T. krapovickasii BAHD allele with a point mutation in a highly conserved domain of BAHD acyl transferases. Wild type and mutant TkBAHD alleles were expressed in Arabidopsis to assay for brassinosteroid (BR) inactivating activity. The wild type but not mutant allele caused dwarfism, consistent with the wild type possessing, but the mutant allele having lost, BR inactivating activity. To investigate whether BRs act directly in self-incompatibility, BRs were added to in vitro pollen cultures of the two mating types. A small morph specific stimulatory effect on pollen tube growth was found with 5 µM brassinolide, but no genotype specific inhibition was observed. These results suggest that BAHD acts pleiotropically to mediate pistil length and physiological mating type through BR inactivation, and that in regard to self-incompatibility, BR acts by differentially regulating gene expression in pistils, rather than directly on pollen.

In heterostylous, self-incompatible Turnera species, a member of the YUCCA gene family, YUC6, resides at the S-locus and has been hypothesized to determine the male mating type. YUCCA gene family members synthesize the auxin, indole-3-acetic acid, via a two-step process involving the TAA gene family. Consequently, it has been speculated that differences in auxin concentration in developing anthers are the biochemical basis underlying the male mating type. Here, we provide empirical evidence that supports this hypothesis. Using a transgenic knockdown approach, we show that YUC6 acts pleiotropically to control both the male physiological mating type and pollen size, but not the filament length dimorphism associated with heterostyly in Turnera. Using qPCR to assess YUC6 expression in different transgenic lines, we demonstrate that the level of YUC6 knockdown correlates with the degree of change observed in the male mating type. Further assessment of YUC6 expression through anther development, in the knockdown lines, suggests that the male mating type is irreversibly determined during a specific developmental window prior to microsporogenesis, which is consistent with the genetically sporophytic nature of this self-incompatibility system. These results represent the first gene controlling male mating type to be characterized in any species with heterostyly.

A majority of Turnera species (Passifloraceae) exhibit distyly, a reproductive system involving both self-incompatibility and reciprocal herkogamy. This system differs from self-incompatibility in Passiflora species. The genetic basis of distyly in Turnera is a supergene, restricted to the S-morph, and containing three S-genes. How supergenes and distyly evolved in Turnera, and the other Angiosperm families exhibiting distyly remain largely unknown. Unraveling the evolutionary origins in Turnera requires the generation of genomic resources and extensive phylogenetic analyses. Here, we present the annotated draft genome of the S-morph of distylous Turnera subulata. Our annotation allowed for phylogenetic analyses of the three S-genes’ families across 56 plant species ranging from non-seed plants to eudicots. In addition to the phylogenetic analysis, we identified the three S-genes’ closest paralogs in two species of Passiflora. Our analyses suggest that the S-locus evolved after the divergence of Passiflora and Turnera. Finally, to provide insights into the neofunctionalization of the S-genes, we compared expression patterns of the S-genes with close paralogs in Arabidopsis and Populus trichocarpa. The annotation of the T. subulata genome will provide a useful resource for future comparative work. Additionally, this work has provided insights into the convergent nature of distyly and the origin of supergenes.