Genus-wide sequencing supports a two-locus model for sex-determination in Phoenix

Abstract

The date palm tree is a commercially important member of the genus Phoenix whose 14 species are dioecious with separate male and female individuals. To identify sex determining genes we sequenced the genomes of 15 female and 13 male Phoenix trees representing all 14 species. We identified male-specific sequences and extended them using phased single-molecule sequencing or BAC clones. We observed that only four genes contained sequences conserved in all analyzed Phoenix males. Most of these sequences showed similarity to a single genomic locus in the closely related monoecious oil palm. CYP703 and GPAT3, two single copy genes present in males and critical for male flower development in other monocots, were absent in females. A LOG-like gene appears translocated into the Y-linked region and is suggested to play a role in suppressing female flowers. Our data are consistent with a two-mutation model for the evolution of dioecy in Phoenix.

Fig. 1

Identification of male-specific kmers in the genus Phoenix. a Sixteen base pairs kmers identified only in the male (blue) or female (red) individual for each species were tested for their presence in other species. As expected for an XY sex determination system, kmers specific to males maintained counts through all species tested while female-specific kmers disappeared after requiring presence in more than eight species. b Kmers specific to male Phoenix enumerated against the male-specific BAC contigs. Kmers specific to males in at least 12 species are fully covered by the sequenced BACs in this study (dark blue—% kmers matching BACs, light blue—% kmers not matched to BACs). Only a female was available for Phoenix pusilla comparisons (14F)

Fig. 2

Male-specific BAC and scaffold sequences contain varying numbers of male-specific kmers. Percent of kmers matching in six species were used to normalize to 100% for size of the sequence contig. The dpBGPAT3like contig (purple) maintained the largest percentage of kmers through the 14 species, followed by dpBCYPlike (light green), dpBLOGlike (yellow), and the dpB3Y scaffold (blue). For comparison, three scaffolds (green, orange, red) containing kmers conserved in only a few males of the genus were included (dpS6Y, dpB1Y, dpB2Y). These are likely due to the spread of non-recombination on the Y chromosome after speciation as they do not maintain kmers in all 14 species. For comparison, see genes synonymous mutation rates in the text. Only a female was available for Phoenix pusilla comparisons (14F)

Fig. 3

The sex determination region in Phoenix. a Phoenix male-specific kmers/bp in date palm scaffolds based on number of species compared (specs: species). Darker boxes indicate higher densities of kmers/bp. Scaffolds with male-specific kmers had similarity to a single locus in oil palm. The only gene not deriving from the region was LOG-like. b Sequencing of representatives of both male (DP_MALE) and female (DP_FEMALE) alleles in date palm showed coverage of significant sections of the oil palm scaffold NW_011550905.1. Each date palm scaffold is connected by a dotted line to a colored rectangle showing span of similarity with oil palm. c Comparison of the date palm X and Y alleles to the orthologous oil palm region revealed a deletion of the GPAT3 and CYP703 gene in the X chromosome. A similar GPAT3 deletion is found in the Y chromosome although the GPAT3 gene does exist elsewhere on Y. The region showed a fusion/inversion (red dotted lines) with respect to the oil palm locus joining the region downstream of MYB315L to the region upstream of BAG-like in date palm. Note shift of sequence coordinates in dpB2Y with respect to oil palm reveals inversion. Cytidine deaminase resides in this region in oil palm. d Analysis of male-specific kmers present in all species of Phoenix identified only four scaffolds dpBGPATlike containing GPAT3, dpBCYPlike containing CYP703, dpB3Y containing cytidine deaminase like (CytDeaminase), and dpBLOGlike containing LOG. The most widely conserved kmers were focused in gene and apparently gene regulatory regions (#species track). Sequence coverage of each genome was normalized to 2N and plotted as a heatmap. Blue rows represent males; red rows represent females from the various Phoenix species. Green rows represent hermaphrodite palms Brahea and Livistona. Males showed one copy and hermaphrodite palms showed two copies per genome in most exonic regions (N copies, by genome track) with some variation (see text). The average coverage for all males indicated these regions were likely at 1N except for the LOG gene that has an autosomal paralog and cytidine deaminase that has an X-linked copy (N copies, average track). Gene expression (mRNA track) showed that GPAT3, CYP703, and LOG were all expressed in male flower (Mfl) and was either very low or undetectable in male leaf (Mlf), female flower (Ffl), and female leaf (Flf)

Fig. 4

A model for the development of dioecy in the genus Phoenix. The genes in the region as found in female scaffold dpS2X, oil palm scaffold NW_011550905.1, and male contig dpB2Y. The fact that the male-specific sequences are focused in this region suggests that it may be the origin of sex determination in Phoenix. a The first step would be the deletion (white pentagons with red x) of GPAT3 and/or CYP703, two genes known to be critical to male flower formation, leading to gynodioecy and formation of a proto-X chromosome. b An inversion (gray triangles) of the corresponding region on the normal chromosome moving the cytidine deaminase (CYT DA) gene would create a proto-Y chromosome with recombination arrest. c This would be followed by a duplication and translocation (red arrow) into the region of the LOG gene creating the final Y chromosome with suppression of female flower development. NCBI GENE IDs as follows AARF: LOC105059738, GPI: LOC105059739, MY315L: LOC105059740, MAP-1: LOC105059742, CYT DA: LOC105059743, MYB-A: LOC105059783, GPAT3: LOC105059961, GPAT3tr: truncated GPAT3, TIF2: LOC105059784, CYP703: LOC105059962, BAG: LOC105059785. The intervening gene between CYP703 and BAG was not present in either male or female alleles of date palm. A short, truncated 3′ end of GPAT3 remains in both X and Y alleles (white pentagon with no red x)