Chloroplastic and nuclear diversity of wild beets at a large geographical scale: Insights into the evolutionary history of the Beta section

Abstract Historical demographic processes and mating systems are believed to be major factors in the shaping of the intraspecies genetic diversity of plants. Among Caryophyllales, the Beta section of the genus Beta, within the Amaranthaceae/Chenopodiaceae alliance, is an interesting study model with species and subspecies (Beta macrocarpa, Beta patula, Beta vulgaris maritima and B.v. adanensis) differing in geographical distribution and mating system. In addition, one of the species, B. macrocarpa, mainly diploid, varies in its level of ploidy with a tetraploid cytotype described in the Canary Islands and in Portugal. In this study, we analyzed the nucleotide diversity of chloroplastic and nuclear sequences on a representative sampling of species and subspecies of the Beta section (except B. patula). Our objectives were (1) to assess their genetic relationships through phylogenetic and multivariate analyses, (2) relate their genetic diversity to their mating system, and (3) reconsider the ploidy status and the origin of the Canarian Beta macrocarpa.


| INTRODUCTION
The nature of forces that shape genetic diversity of species is a longstanding question in evolutionary biology (Leffler et al., 2012). Both historical demographic process that occurred during glaciation periods and life history traits are generally admitted to be the major factors influencing the present intraspecies genetic diversity. In plants, mating systems are believed to be of main importance, in particular the frequent transition to self-fertility that is expected to affect both neutral diversity and the efficacy of selection (Glémin, 2007;Glémin, Bazin, & Charlesworth, 2006). Empirical studies in a set of species have partially confirmed these theoretical expectations (reviewed in Glémin &Galtier, 2012 andin Castric, Billiard, &Vekemans, 2013). Another evolutionary mechanism influencing plant species diversification is polyploidization. This can occur after interspecies hybridization (allo-polyploidy) or intraspecific genome duplication (autopolyploidy) (reviewed by Soltis, Marchant, Van de Peer, & Soltis, 2015). It has been generally believed that allopolyploids are more frequent than autopolyploids thanks to the expected gain in fitness of hybrids combining two diverged genomes and thus enlarging their ability of conquering new environments (Abbott et al., 2013). However, autopolyploid occurrence seems to have been underestimated as it appears to be as frequent as allopolyploids, partly due to the difficulty in phenotypically distinguishing them from their diploid counterparts (Barker, Arrigo, Baniaga, Li, & Levin, 2016).
Mediterranean Basin (Aegean islands, Turkey and Syria). In addition, subspecies of the Beta section differ in their mating system: B.v. maritima is allogamous and self-incompatible, while B. macrocarpa and B. v. adanensis have been described as self-compatible (Bruun et al., 1995;Letschert, 1993).

| DNA amplification and sequencing
The DNA extraction from dried leaf tissue was carried out with a Nucleospin ® 96Plant kit (Macherey-Nagel) on a Microlab ® Star robot (Hamilton).

| cpDNA sequences
Four cpDNA regions were selected for sequencing: the trnK intron (K1K2) including the matK gene, the trnD-trnT intergenic spacer (DT), the trnL-trnF intergenic spacer (LF), and the 5′ part of the intergenic spacer HK ranging between trnH and psbA. On account of its size (about 1,900 base pairs [bp]), the K1K2 region was amplified in two overlapping fragments.

| Nuclear DNA sequences
PCR products were directly sequenced for the autogamous diploid species B.v adanensis and B. macrocarpa and for the Corollinae species. For the outcrossers B. v. maritima and tetraploid B. macrocarpa, PCR products were cloned into pCR2.1-TOPO using TOPO TA Cloning Kit (Invitrogen, Carlsbad, CA) before sequencing. A minimum of six clones was sequenced to reliably identify both haplotypes and examine PCR-generated errors due to nucleotide misincorporation and/or recombination.
All sequences generated in the present study have been registered in Genbank (KP747713-KP748171).

| Phylogenetic and haplotype network reconstructions
The alignment resulted in a dataset of 3,742 bp for the chloroplas- considered: they correspond to the intronic and exonic regions of the adh and cab11 genes and to ITS. For the ML analyses, datasets, concatenated or not, were considered as one partition.
Analyses with MrBAYES were done as follows: two runs of four Markov chains were calculated simultaneously for 1,000,000 to 5,000,000 generations depending on the dataset, with initial equal probabilities for all trees and a random starting tree. Trees were sampled each 100 generations, and the consensus tree with posterior probabilities (PP) was calculated after removal of the first 25% to 50% (according to the analysis) of the total number of generated trees (according to the analysis). The average standard deviation of split frequencies between the two independent runs was lower than 0.01.

| Principal component analysis
In order to assess the existence of genetic clusters within the Beta section, we conducted a principal component analysis (PCA) on the concatenated nuclear sequences of all individuals except for the samples B. v. maritima 6 and B. v. adanenis a10 (adh sequence was missing for 6, and cab11 sequence for a10) using adegenet R package (Jombart, 2008; R Core Team Development 2014).

| Statistical analyses-nucleotide diversity parameters
For each species/subspecies of the Beta section, we estimated the nucleotide diversity both as π, the average number of nucleotide differences per site between a pair of randomly chosen sequences (Nei, 1987), and as Watterson's θ w (Watterson, 1975). Among species/subspecies of the Beta section, we calculated shared and fixed polymorphisms and the nucleotide divergence (Dxy), using DnaSP version 5 (Librado & Rozas, 2009).

| Phylogenetic analyses
The concatenated chloroplastic sequences from the 57 samples of the   Figures S2-S4).
Overall, chloroplastic and nuclear phylogenetic trees showed that

| Principal component analysis
The

| Nucleotide diversity of the Beta section
The representative distribution of the sampling enabled us to measure the overall nucleotide diversity of the members of the Beta section, at both chloroplastic and nuclear levels ( In parallel of the phylogenetic and PCA analyses, the level of divergence between the members of the Beta section can be described by assessing the number of private and shared polymorphisms among members, as well as the number of fixed differences (Table 3).

Accordingly, B. macrocarpa represents a distinct genetic pool from
Beta vulgaris, as it exhibits fixed differences at both genomic compartments with B.v. maritima and B.v. adanensis, while B.v. maritima and B.v. adanensis exhibit none.
The same pattern is less obvious when considering the nucleotide divergence among Beta section members (Dxy, at the chloroplastic level, but is higher at the nuclear loci, especially at the ITS locus with a level of nucleotide divergence that is 5 time as T A B L E 2 Species diversity of the Beta section. At each locus, chloroplastic (cp) and nuclear loci (Adh, Cab11, and ITS) and for each species/ subspecies are given: the number of populations per species (Pop) and sequences (Seq), number of haplotypes, number of segregating sites, diversity per site estimated from the total number of mutations (Θ w ), diversity as the average number of nucleotide differences per site between a pair of randomly chosen sequences (π) with standard deviation (SD)

| DISCUSSION
The present study aimed to survey the chloroplastic and nuclear genetic diversities of Beta species (Beta section) and explore the phylogenetic relationships among them.
Accordingly with former studies (Andrello et al., 2016(Andrello et al., , 2017Kadereit et al., 2006;Letschert, 1993;Romeiras et al., 2016) v. adanensis (Bruun et al., 1995). Further studies on a larger sampling and including a population level, in particular by contrasting para- In previous studies, Beta macrocarpa has been described as two cytotypes: one diploid cytotype widely distributed from Portugal to Turkey, along the Mediterranean Basin, and a tetraploid cytotype first found in the Canary Islands (Buttler, 1977) Earlier studies on this tetraploid cytotype have suggested a hybrid origin of the taxon between B.
v. maritima and B. macrocarpa: (1) cytological observations revealed a complete diploidised meiosis as expected for an alloploid (Lange & de Bock, 1989), (2) genetic analyses on nuclear allozyme loci showed B.v. maritima and B. macrocarpa alleles-like (Abe & Tsuda, 1987;Letschert, 1993), and (3) a maritima-like chloroplastic haplotype was found in a Canarian individual (Kishima, Mikami, Hirai, Sigiura, & Kinoshita, 1987). Nevertheless, the occurrence of tetraploid individuals does not seem to be restricted to the Canary Islands as formerly believed: recent studies localized 4× individuals on another Macaronesian island, Santo Porto (Madeira Archipelago) (Leys et al., 2014) but also in continental populations from Southern Portugal (Castro et al., 2013). The present study confirms the hybrid origin of 4× B. macrocarpa from two Canary Islands (Gran Canaria and Tenerife): at the nuclear level each individual bears a maritima-like allele and a macrocarpa-like allele with the exception of ITS where only one allele, belonging to the B.v. maritima clade, was found. This is most likely due to concerted evolution as observed in allopolyploid Gossypium species (Wendel, Schnabel, & Seelanan, 1995), rice (Bao, Wendel, & Ge, 2010), or tobacco (in Bao et al., 2010. At the chloroplastic level, both 4× Canarian individuals shared the same haplotype with B.v. maritima individuals. This suggests that the initial maternal parent of the hybrid was B.v. maritima, and thus B. macrocarpa was the pollen donor. The hybridization between selfincompatible B.v. maritima and self-compatible B. macrocarpa led to an alloploid species, described as self-compatible in early studies (Buttler, 1977). Our results suggest that 4× individuals mainly reproduce by selfing, as we did not find any heterozygosity at the homeologous loci.
It must be noted that if the present study confirms the allopoly- Islands where they were until now considered as absent. It remains to know the relative occurrence of the two forms in the Canary Islands as well as the geographical origin of 4× macrocarpa populations: whether the hybridization occurred in the islands or in the continent followed by long-distance dispersal (Linder & Barker, 2014). Further studies are needed to describe the phenotypic characteristics and the ecological preferences of the different macrocarpa cytotypes in order to better distinguish them taxonomically but also to understand how the two types coexist in the Macaronesian archipelago and the adjacent regions.

ACKNOWLEDGMENTS
This work was supported by grants from the Région Nord-Pas de Calais and the European Community (European Regional Development Fund). We wish to warmly thank Lothar Frese, Brian