Skip to main content
NCBI home page
Annals of Botany logoLink to Annals of Botany
. 2015 Sep 30;117(1):37–50. doi: 10.1093/aob/mcv138

Phylogeography and modes of reproduction in diploid and tetraploid halophytes of Limonium species (Plumbaginaceae): evidence for a pattern of geographical parthenogenesis

Ana Sofia Róis 1, Flávio Sádio 1, Octávio S Paulo 2, Generosa Teixeira 3, Ana Paula Paes 1, Dalila Espírito-Santo 1,4, Timothy F Sharbel 5, Ana D Caperta 1,4,*
PMCID: PMC4701142  PMID: 26424783

Abstract

Background and Aims The genus Limonium (Plumbaginaceae) has long been recognized to have sexual and apomictic (asexual seed formation) modes of reproduction. This study aimed to elucidate phylogeographical patterns and modes of reproduction in diploid and tetraploid Limonium species, namely three putative sexual diploid species with morphological affinities (L. nydeggeri, L. ovalifolium, L. lanceolatum) and three related, probably apomict tetraploid species (L. binervosum, L. dodartii, L. multiflorum).

Methods cpDNA diversity and differentiation between natural populations of the species were investigated using two chloroplast sequence regions (trnL intron and trnL–trnF intergenic spacer). Floral heteromorphies, ovule cytoembryological analyses and pollination and crossing tests were performed in representative species of each ploidy group, namely diploid L. ovalifolium and tetraploid L. multiflorum, using plants from greenhouse collections.

Key Results and Conclusions Genetic analyses showed that diploid species have a higher haplotype diversity and a higher number of unique (endemic) haplotypes than tetraploid species. Network analysis revealed correlations between cpDNA haplotype distribution and ploidy groups, species groups and geographical origin, and haplotype sharing within and among species with distinct ploidy levels. Reproductive biology analyses showed that diploid L. ovalifolium mainly forms meiotically reduced tetrasporic embryo sacs of Gagea ova, Adoxa and Drusa types. Limonium multiflorum, however, has only unreduced, diplosporic (apomictic) embryo sacs of Rudbeckia type, and autonomous apomictic development seems to occur. Taken together, the findings provide evidence of a pattern of ‘geographical parthenogenesis’ in which quaternary climatic oscillations appear to be involved in the geographical patterns of coastal diploid and tetraploid Limonium species.

Keywords: Apomixis, cytoplasmic markers, cpDNA, female gametophyte development, floral biology, geographical parthenogenesis, Limonium, phylogeography, polyploidy, reproductive modes

INTRODUCTION

Phylogeographical studies based on cpDNA haplotype variation and diversity have provided a great opportunity to reconstruct major processes in evolutionary history and biogeography in flowering plants (angiosperms) (Avise, 2000; Petit and Vendramin, 2007). Some of these works have helped to clarify the complicated reticulate evolutionary patterns observed in plant genera driven by sexual and/or apomictic reproductive modes (e.g. Hörandl et al., 2009; Kiefer et al., 2009; Majeský et al., 2012). Apomixis (reproduction via asexually formed seed) (Asker and Jerling, 1992) is present in 293 genera (∼2·2 %) of flowering plants and found to be taxonomically widespread with no tendency to occur in specific groups, and occurs with sexuality at all taxonomic levels (Hojsgaard et al., 2014). Most apomictic plants are polyploids and exhibit facultative sexuality, but there are cases of apomixis at the diploid level, including some species in the genera Boechera (Koltunow and Grossniklaus, 2003) and Paspalum (Siena et al., 2008). In recent years, biogeographical studies of some apomictic taxa (e.g. Arnica alpina, Paspalum simplex) and apomictic complexes (e.g. Taraxacum officinale agg., Ranunculus auricomus agg.) have shown patterns typical of ‘geographical parthenogenesis’, in which apomicts have larger distributions in higher latitudes than their sexual relatives, in addition to having differentially populated previously glaciated areas (Bierzychudek, 1985; Kearney, 2005; Hörandl et al., 2008). However, other apomictic taxa (e.g. Boechera) do not show evidence of geographical parthenogenesis, as recurrent hybridization has led to diffuse geographical patterns (Koch et al., 2003; Kiefer et al., 2009).

In Limonium (Plumbaginaceae), a cosmopolitan species-rich genus of annuals and perennial herbs, shrubs and lianas, often adapted to extreme coastal environments (Kubitzki, 1993), the effects of hybridization, polyploidy and apomixis appear to have all combined to shape their radiation (Lledó et al., 2005). Multiple series of complex aggregates of sexual diploid species and asexual polyploid hybrids have been described (Erben, 1978, 1993). Some authors have used various molecular markers to infer the importance of genetic structuring in defining Limonium interspecies delimitation (Palacios et al., 2000; Palop-Esteban et al., 2007; Róis et al., 2013). However, the majority of these studies covered parts of the distribution ranges or only a small part of the distribution of the species, and for most apomitic Limonium taxa biogeographical analyses have not yet been conducted.

Baker (1948, 1953a, b, 1966), who was influential in the development of evolutionary and biogeographical hypotheses for Plumbaginaceae, described floral heteromorphies, including heterostyly and pollen–stigma dimorphisms, linked to a sporophytic self-incompatibility system, which prevent mating between individuals of the same floral morph while allowing mating between morphs. In this system, dimorphic pollen–stigma populations, with plants featuring flowers with distinct pollen grains types (coarsely reticulate, A type; or finely reticulate, B type) and distinct stigma types (polygonal or papillate cells) appear to be outcrossing. In contrast, monomorphic self-incompatible populations showing only one pollen–stigma combination seem to form seeds through apomixis (Baker, 1966). In Limonium spp. various types of tetrasporous (meiotic) reduced embryo sacs have been reported, including Adoxa, Fritillaria, Drusa and Penea types (Dahlgren, 1916; D’Amato, 1940, 1949; Hjelmqvist and Grazi, 1964). Also, in polyploid Limonium vulgare (also known as Statice limonium L.; Statice, nom. rej. vs. Armeria [Greuter et al., 2000]) Fritillaria-type female gametophyte development occurs (D’Amato, 1940). Furthermore, facultative gametophytic apomicts like triploid (2n =   3x = 27) Statice oleaefolia var. confusa form meiotic, tetrasporous reduced Adoxa-type embryo sacs in parallel with apomictic, unreduced, diplosporous Erigeron-type embryo sacs (D’Amato, 1949). Nevertheless, for most Limonium apomictic taxa, data on patterns of female gametophyte formation are lacking.

Geographically, the Iberian Atlantic coast is between the North Atlantic shore and Mediterranean Iberia, where Limonium species showing a typical European Atlantic distribution meet species distributed in the south-west Mediterranean. This is the case for L. vulgare and related species with pinnately arranged leaves (Cortinhas et al., 2015), as well as species with parallel-nerved leaves such as diploids of the Limonium ovalifolium complex, like L. ovalifolium and Limonium lanceolatum (Franco, 1984; Erben, 1993; EUNIS, 2014), and tetraploids of the Limonium binervosum complex, which includes, among others, L. binervosum, Limonium dodartii and Limonium multiflorum (Erben, 1978; Ingrouille and Stace, 1986). Inferences about the primary reproductive strategies of these species have been based upon analyses of flower heteromorphies (Erben, 1978, 1999; Ingrouille and Stace, 1985, 1986), which have shown that L. nydeggeri and L. ovalifolium are likely sexual, whereas tetraploids are all putative agamospermous species (Ingrouille and Stace, 1985). Flow cytometric seed screening, which has the potential to determine the routes of seed formation based on the proportional DNA content of embryo and endosperm nuclei (Matzk et al., 2000), failed in the elucidation of L. ovalifolium and L. multiflorum reproductive modes using mature seeds (Róis et al., 2012). Thus, even if in this species sexuality and/or agamospermy is suspected, it has never been directly tested. Notably, members of the L. ovalifolium species complex have imperfect but distinguishable morphological differentiation and genetic and epigenetic co-variation, as revealed by methylation-sensitive amplification polymorphism (MSAP) analysis (Róis et al. 2013). Conversely, the tetraploid L. binervosum species complex contains even lower levels of overall genetic diversity, but higher levels of epigenetic variability (Róis et al., 2013).

In this work we investigated (1) the phylogeographical patterns of diploid and tetraploid Limonium spp. which grow in sympatry as well as in allopatry using the maternally inherited chloroplast trnL intron and trnL–trnF intergenic spacers (Petit et al., 2003; Petit and Vendramin, 2007); and (2) potential correlations of geographical patterns with species differentiation and reproductive modes.

MATERIALS AND METHODS

Study species and population sampling

Here we studied (1) parallel-veined species of the genus Limonium, including three diploid species of the L. ovalifolium complex: L. ovalifolium sensu stricto, L. nydeggeri and L. lanceolatum (Erben, 1978, 1993, 1999), with both diploid and aneuploid cytotypes (2n = 2x = 15, 16) (Róis et al., 2012); and (2) three agamospermous tetraploid species from the agamospermous L. binervosum complex: L. binervosum sensu stricto, L. dodartii and L. multiflorum (Erben, 1978, 1993; Ingrouille and Stace, 1986), exhibiting tetraploid and aneuploid cytotypes (2n =  4x = 35 [Erben, 1993]; 2n =  4x = 32, 34, 35, 36 [Róis et al., 2012]). The L. ovalifolium species complex, which thrives on maritime cliffs and in salt marshes, reaches its northern limits on the Atlantic coast of France in the Rance estuary (Lahondère and Biorét, 1995, 1996) and has its southern limits in Morocco (Erben, 1978, 1993, 1999) (Fig. 1). However, the three diploid species have distinct distribution ranges: L. ovalifolium has the largest distribution area, extending from Atlantic France, south-west of the Iberian Peninsula to Morocco (Erben, 1978, 1993); L. lanceolatum is found in both the north-west and south-west coasts of Iberian Peninsula (Franco, 1984; EUNIS, 2014); and L. nydeggeri is a Lusitanian endemic from the west and south-west coasts of Portugal (Erben, 1999; Espírito-Santo et al., 2012). The L. binervosum species complex is a group of Atlantic coastal agamospermous microspecies that grow on maritime rocks and saltmarshes (Stace, 2010). Geographically, L. binervosum sensu stricto is distributed from Scotland to Portugal (Pignatti, 1971; Erben, 1978; Ingrouille and Stace, 1986), although it is not mentioned for Portugal in Flora Iberica (Erben, 1993). Limonium dodartii has been recognized on the French Atlantic coast, north-western Spain and western Atlantic Portugal coasts (Erben, 1993; Lahondère and Biorét, 1995, 1996) as well as on the Atlantic coasts of France (Lahondère and Biorét, 1996). Finally, L. multiflorum is a Lusitanian endemic on the western coast of Portugal (Erben, 1978, 1993; Espírito-Santo et al., 2012; Caperta et al., 2014).

Fig. 1.

Fig. 1.

Open in a new tab

Distribution of diploid and tetraploid Limonium spp. on Atlantic and Mediterranean coasts. Diploid species are L. lanceolatum, L. nydeggeri and L. ovalifolium; and tetraploid species are L. binervosum, L. dodartii, and L. multiflorum.

Populations were surveyed along the Portuguese coast in Beira Litoral (west), Estremadura (west), Alentejo (south-west) and Algarve (south), where diploid and tetraploid species show overlapping distributions. The two Lusitanian endemic species, L. nydeggeri and L. multiflorum, were sampled over their whole distribution ranges. Also, an L. multiflorum subpopulation of the Cabo Raso (CRm) population was investigated in this study in order to have equivalent numbers of individuals in the populations analysed. All populations were recorded using a global positioning system (Table 1), Google Earth 6·0.2 was used for georeferencing and geographical mapping of the populations was performed using ArcGIS Desktop 10 (ESRI).

Table 1.

Natural Limonium populations surveyed in Portuguese provinces. Species were sampled in Beira Litoral (west), Estremadura (west), Alentejo (southwest) and Algarve (south)

Species Ploidy level Habitat Province/site location Population acronym Geographical coordinates Census
L. lanceolatum 2x (A S. Róis & A. D. Caperta, unpubl. res.) Salt marshes Alentejo/Odemira, Vila Nova de Milfontes VNMF 37·7277/–8·7709 25
Collectors: ADC, APP, ASR
Algarve/Tavira, Sapal do Barril SB 37·0861/–7·6627 50
Collectors: ADC,APP, ASR
L. nydeggeri 2x (Róis et al. 2012) Sea cliffs Estremadura/Peniche, Ilha do Baleal B 39·3789/–9·3409 25
Collectors: ADC, APP, ASR
Estremadura/Peniche, Nossa Sra dos Remédios NSR 39·3699/–9·3957 25
Collectors: ADC, APP, ASR
Estremadura/Cascais, Cabo Raso CRn 38·7100/–9·4858 50
Collectors: ADC, APP, ASR
2x (A S. Róis & A. D. Caperta, unpubl. res.) Algarve/Aljezur, Pontal da Carrapateira C 37·1950/–8·9111 25
Collectors: ADC, APP, ASR
Algarve/Vila do Bispo, Cabo de São Vicente CSV 37·0226/–8·9965 25
Collectors: ADC, ASR
L. ovalifolium 2x (Róis et al., 2012) Sea cliffs Algarve/Lagos, Praia da Luz PL 37·0874/–8·7290 25
Collectors: ADC, ASR
Algarve/Vila do Bispo, Cabo de Sagres CS 36·9942/–8·9487 30
Collectors: ADC, APP, ASR
L. binervosum 4x (A S. Róis & A. D. Caperta, unpubl. res.) Sea cliffs Beira Litoral/Cabo Mondego, Figueira da Foz CM 40·1911/–8·9052 25
Collectors: ADC, APP, ASR, SM, FS
Beira Litoral/Marinha Grande, Praia da Concha C 39·7552/–9·0322 20
Collectors: ADC, APP, ASR, SM
Beira Litoral/Aveiro, São Jacinto SJ 40·3944/–8·4350 25
Collectors: ADC, APP, SM
L. dodartii 4x (A S. Róis & A. D. Caperta, unpubl. res.) Sea cliffs Alentejo/Odemira, Cabo Sardão CSo 37·5986/–8·8160 50
Collectors: ADC, APP, ASR, SM, PA
Alentejo/Odemira, Porto Covo PCo 37·8516/–8·7919 50
Collectors: ADC, APP, ASR, SM, PA
Alentejo/Sines, Praia da Vieirinha PV 37·8967/–8·7978 30
Collectors: ADC, FS, ASR, SM, APP
L. multiflorum 4x (Róis et al. 2012) Sea cliffs Estremadura/Cascais, Cabo Raso CRm 38·7100/–9·4858 25
Collectors: ADC, APP, ASR
Estremadura/Lourinhã, Vale dos Frades VF 39·2765/–9·3358 20
Collectors: ADC, APP, ASR
Estremadura/Mafra, Foz do Lizandro FL 38·9415/–9·4152 50
Collectors: ADC, APP, ASR
Open in a new tab

Abbreviations of collectors from LEAF, ISA, University of Lisboa: ADC, Ana D. Caperta; APP, Ana P. Paes; ASR, Ana S. Róis; FS, Flávio Sádio; PA, Pedro Arsénio; SM, Sérgio Martins.

DNA isolation, PCR amplification and sequencing

A total of 128 plants were included in this study (Table 2) and three young leaves per individual were sampled from all plants at the same developmental stage. Leaves were cut, placed in paper envelopes and dried immediately at room temperature in sealed containers with abundant silica gel. Total genomic DNA was extracted using an Agencourt Chloropure kit (Beckman Coulter). Two regions of the chloroplast genome (trnL intron and trnL–trnF intergenic spacer [Clegg et al., 1994]) were amplified and sequenced using the universal PCR primers of Taberlet et al. (1991; combinations c/d and e/f). PCR reactions 10 µl were performed containing 50 ng of template DNA, 10× NH4 PCR buffer, 50 mm MgCl2, 100 mm dNTPs, 10 mm of each primer and 5 U/µl Taq DNA polymerase (Bioline). The PCR amplifications began with an initial cycle of 5 min at 95 °C, 30 cycles of 30 s at 95 °C, 30 s at 55 °C and 1·40 min at 72 °C, and a final step of 10 min at 72 °C. The PCR product was verified on an agarose 1·5 % Tris–acetate–EDTA (TAE) gel, and the products were purified using a Qiagen PCR purification kit and sequenced directly using the ABI Prism™ BigDye Ready Reaction Terminator Cycle Sequencing Kit on an ABI 3700 genetic analyser. Sequences were generated from at least two separate PCR amplifications per individual, and both strands of all fragments were sequenced.

Table 2.

cpDNA haplotype diversity in Limonium ovalifolium and L. binervosum species complexes

Species Diploids
 Tetraploids
 No. of individuals/ haplotype
L. lanceolatum
 L. nydeggeri
 L. ovalifolium
 L. binervosum
 L. dodartii
 L. multiflorum
Population VNMF SB B NSR CRn C CSV PL CS CM PC SJ CSo PCo PV CRm VF FL
H1 5 1 6 12
H2 1 1 1 2 3 1 3 5 9 27
H3 5 4 9
H4 8 8
H5 1 1
H6 1 1
H7 1 1
H8 1 1
H9 1 1
H10 1 1
H11 3 4 2 2 1 6 18
H12 1 1
H13 6 2 7 15
H14 1 1
H15 1 1
H16 1 1
H17 1 1
H18 1 1
H19 2 3 1 6
H20 2 2
H21 4 4
H22 1 1
H23 1 1
H24 1 1
H25 1 1
H26 1 1
H27 1 1
H28 10 10
Ni 10 10 8 7 6 9 7 9 6 5 4 6 6 4 8 6 7 10
Hn 4 11 9 2 3 7
Hu 2 7 6 1 2 6
Hd (s.d.) 0·684 (0·070) 0·865 (0·029) 0·829 (0·085) 0 0·111 (0·096) 0·462 (0·128)
π 0·0156 0·0079 0·0054 0 0·0001 0·002
Open in a new tab

H#, haplotype number; Ni, number individuals analysed; Hn, number of haplotypes; Hu, number of unique haplotypes; Hd (s.d.), haplotype diversity and standard deviation; π, nucleotide diversity.

Data analyses

The intron trnL and intergenic spacer trnL–trnLF sequences were aligned using BioEdit 7.1.3.0 (Hall, 1999) and concatenated using Concatenator V 1.1.0 (Pina-Martins and Paulo, 2008). Chloroplast haplotypes were defined using Network 4.6.1.1 (Bandelt et al., 1999; www.fluxus-engineering.com/), and a media-joining network was calculated under default parameters (weights = 10 and ε = 0). Genetic diversity parameters, such as haplotype diversity (Hd) and nucleotide diversity (π) and genetic differentiation (GST), and neutrality tests (Tajima’s D) were calculated using the program DnaSP 4·0 (Rozas, et al. 2003).

The Bayesian phylogenetic inference was implemented with MrBayes v. 3.2.2 (Ronquist et al., 2012). The posterior probabilities of the phylogenetic trees were estimated by a Metropolis-coupled Markov chain Monte Carlo sampling algorithm (MCMCMC). The stationary phase was obtained with a total of 1 500 000 generations that were sampled every 100 generations with a burn-in of 1000. The priors for the model of sequence evolution was estimated with the Akaike information criterion (AIC) using MrModeltest2 (Nylander, 2004).

Pollen exclusion and crossing experiments

A total of ten fresh flowers were used for floral heteromorphism determinations in L. ovalifolium and L. multiflorum growing in experimental collections at Instituto Superior de Agronomia. From these plants, stigma and pollen types (A/B pollen and cob-like/papillate stigmas; Baker, 1948; Erben, 1978) were determined. These organs were dissected and covered with a drop of water, and stigma preparations were observed by optical light microscopy (Leitz HM-LUX 3) with × 40 magnification. Pollen viability was scored using Alexander’s stain test (Alexander, 1969). The total number of pollen grains was determined and pollen viability estimates were performed by one person using ∼20 flowers per plant; the number was counted under an optical light microscope (Leitz HM-LUX 3) at × 20 magnification.

A total of 18 plants, 12 self-incompatible L. ovalifolium individuals (with A or B combination) and six male-sterile L. multiflorum individuals (Róis et al., 2012) were subjected to pollen exclusion experiments to test for apomixis, as described in Khan et al. (2012), with some modifications. No flower emasculation prior to anther dehiscence was performed in inflorescences (scapes) as the plants were self-incompatible or male-sterile.

Pollination treatments included the following: (1) in each plant, an intact inflorescence was left unbagged to test for natural pollination (control); (2) intact inflorescences were bagged with vegetable cellulose paper that did not allow the passage of pollen or floral visitors, to test for apomixis; (3) pollen from conspecifics was used to test for cross-compatibility between three self-incompatible plants (A and B combinations). In the last treatment, manual pollination was performed once for each inflorescence by brushing dehiscing anthers against the receiving stigmas until pollen could be seen on the stigma surface. Pollen exclusion bags were immediately placed over the inflorescences used in each cross-pollination to prevent undesired pollination events. After 9 weeks, bags were removed and the number of produced seeds was counted. The average number of seeds produced per scape (either unbagged or bagged) was calculated from measurements of six plants of each type. Seeds were then germinated as described in Róis et al. (2012). Germination was considered to have occurred on the emergence of radicles, and the number of seeds that germinated after 1 week was recorded.

Non-parametric Kruskal–Wallis analyses of variance of data were performed (Sokal and Rohlf, 2012) using SPSS 20 for Windows.

Electron microscopy analyses

Stigma, pollen and ovules were analysed using flowers from the above L. ovalifolium and L. multiflorum plants. Flowers were fixed in a 2·5 % glutaraldehyde solution in 0·1 m sodium phosphate buffer, pH 7·2, for 5 h at 4 °C as described in Hayat (1981). The material was then dehydrated in a graded ethanol series (30, 50, 75 and 100 % ethanol for 30 min each). Flowers were dried on a BioRad Polaron E3500 critical point drier, and coated with a thin layer of gold on a Jeol JFC-1200. Observations were carried out on a JSM-5220 LV scanning electron microscope equipped with a direct image acquisition system.

Cytohistological analyses of megasporogenesis and gametogenesis

Flower buds at distinct floral stages ranging from 2 to 5 mm in length were collected from some of the above-mentioned L. ovalifolium and L. multiflorum plants. Pistils were dissected from flower buds, followed by clearing and staining following the protocol of Stelly et al. (1984), with some modifications. Briefly, the tissues were fixed in FAA (3·7 % formalin, 5 % acetic acid, 70 % ethanol) overnight at 4 °C and then dehydrated in 50, 75 and 100 % progressive water/ethanol solutions for 30 min each. Then, samples were stained with pure Mayer’s haemalum for 48 h, and placed in 0·1 % chloral hydrate (Wallis, 1957). Ovules were digested with 2 % cellulase (Sigma), 2 % cellulase ‘Onozuka R-10’ (Serva) and 2 % pectinase enzyme solution in Enzyme Buffer (40 mL of 0·1 m citric acid-1-hydrate and 60 mL of 0·1 m sodium citrate dihydrate; pH 4·8) for 20 min at 37 °C, as described in Caperta et al. (2008). Finally, ovules were mounted in multi-well slides in a drop of chloral hydrate, under a coverslip. Cytohistological observations were made using a fluorescence microscope (Zeiss Axioskop 2) using differential interference contrast (DIC) optics, under × 20 and × 100 magnification. Images were collected with an AxioCam 289 MRc5 digital camera (Zeiss) and further processed using 290 Adobe Photoshop 5·0 (Adobe Systems, Mountain View, CA, USA).

To test for differences between L. ovalifolium and L. multiflorum, χ2 tests of independence were performed (Sokal and Rohlf, 2012) using SPSS 20 for Windows.

RESULTS

cpDNA haplotype distribution and diversity

The sequencing of the intron trnL and intergenic spacer trnL–trnLF resulted in a final matrix of 128 individuals for a 646-bp concatenated fragment (Supplementary Data Table S1). Sixty-two single-nucleotide polymorphisms and four indels were detected.

A total of 28 (H1–H28) haplotypes were identified in the 18 species populations analysed (Table 2). Diploid species showed a higher number of haplotypes (n =  18) than tetraploid species (n =  10), and most species presented more than three haplotypes. All species showed populations with unique haplotypes but the number of endemic haplotypes was higher in diploid than in tetraploid species. Limonium nydeggeri and L. multiflorum had the highest number of haplotypes and the highest number of unique haplotypes of all diploid and tetraploid species respectively.

In both species complexes, populations with just one haplotype were found: haplotype H28 in L. lanceolatum SB, haplotype H1 in L. binervosum CM and SJ, haplotype H3 in L. dodartii CSo and PCo, and haplotype H4 in the PV population. A common haplotype, H11, was detected in all diploid species, albeit not in all their populations. Haplotype H2 was shared by diploid and tetraploid species, as observed in L. nydeggeri NSR and CRn, L. ovalifolium CS, L. binervosum PC, L. dodartii CSo and L. multiflorum CRM, VF and FL populations. cpDNA haplotype (Hd) and nucleotide diversity (π) was higher in diploid than tetraploid species. In these latter species, cpDNA haplotype diversity ranged from none (L. binervosum) to moderate (L. multiflorum). Finally, considering all species, the genetic differentiation coefficient (GST) among populations was 0·5015 and FST was 0·6103. Moreover, Tajima’s D neutrality tests were negative but non-significant (–1·5105; P >  0·10).

Network relationships among haplotypes

The network evidenced clustering for ploidy levels in which specific haplotypes were found (Fig. 2A). Apart from haplotype H2, which was shared by both diploid and tetraploid species, diploid (H11–H28) and tetraploid (H1, H3–H10) specific haplotypes were evident. Tetraploid haplotypes seem to cluster around H2, with H2 the putative ancestral (shared with diploid species) and H1, H3, H4, H5, H6, H8 and H10 as derived haplotypes. However, some tetraploid haplotypes that did not cluster with the mentioned group (e.g. H7 and H9) seemed to be derived from haplotypes present in diploid species rather than from tetraploids.

Fig. 2.

Fig. 2.

Open in a new tab

Median-joining network of the cpDNA sequence regions (intron trnL and intergenic spacer trnL–trnLF) in Limonium ovalifolium and Limonium binervosum species complexes. Representation of haplotypes by ploidy level (2n or 4n) (A) and species (B): L. lanceolatum, L. nydeggeri and L. ovalifolium (diploids) and L. binervosum, L. dodartii and L. multiflorum (tetraploids). Circles represent populations and circle size is proportional to haplotype frequency. Each small red circle represents a (predicted) mutation that was not detected in the individuals sampled.

There was also grouping with respect to species (Fig. 2B). Globally, L. lanceolatum, L. nydeggeri and L. ovalifolium had haplotypes that were more interrelated than related to tetraploid species. Diploid species were mainly structured around haplotypes H11, H13 and H19. Nevertheless, in L. nydeggeri and L. ovalifolium some haplotypes were derived from haplotype H2 (e.g. haplotypes H15 and H17, H18–H21, H25 and H26). Limonium binervosum, L. dodartii and L. multiflorum populations were more homogeneous than those of diploid species (Supplementary Data Figure S1). The first two species were clearly differentiated from L. multiflorum, which had two haplotypes, H7 and H9, derived from haplotypes that were frequent in L. nydegerii (e.g. H11 and H19).

Although only L. nydeggeri and L. multiflorum samples were collected considering their total distribution range (both endemic species) (Fig. 1), we identified haplotypes specific to particular geographical regions (Table 2 and Supplementary Data Figure S1). In diploid species, haplotypes H20–H27 (L. nydeggeri [CSV]; L. ovalifolium [PL and CS]) and haplotype H28 (L. lanceolatum [SB]) were only found in southern Portugal. Moreover, a geographical pattern was observed for haplotype H1 in Beira Litoral (L. binervosum [CM, PC and SJ]), haplotypes H3 and H4 in south-west Alentejo (L. dodartii [CSo, PCo and PV]) and haplotypes H5–H10 on the west coast (L. multiflorum [CRm, VF, FL]).

The Bayesian phylogenetic analysis corroborated the results found using the network, with several polytomies resulting from recent divergences (Supplementary Data Figure S2).

Reproductive studies in diploids and tetraploids

In both L. ovalifolium and L. multiflorum, regular, hypogynous and pentamerous flowers were found. Each flower was characterized by a synsepalous calyx, a sympetalous corolla, five antepetalous stamens and a superior pentacarpellary unilocular gynoecium. There were differences between species in terms of pollen–stigma morphs (Fig. 3). In L. ovalifolium plants featuring flowers showing stigmas with polygonal cells (cob) and pollen with macroreticulate exine (self-incompatible A combination [Baker, 1948]; Fig. 3A, C), or stigmas with papillate (pap) cells and pollen with microreticulate exine (self-incompatible B combination; Fig. 3B, D) were detected. In contrast, in L. multiflorum, stigmas with cob or pap cells were present (Fig. 3E, F) but pollen grains exhibited an abnormal morphology without the typical exine patterns (Fig. 3G, H). Therefore, in this latter species we were unable to attribute a specific pollen–stigma morph. Furthermore, in L. ovalifolium, empty anthers were seldom found, whereas L. multiflorum had many flowers with empty anthers (two or three) and sometimes flowers with no pollen at all were detected. Limonium ovalifolium pollen grains showed very high viability (∼97 %), but only low to moderate levels of viability (0–69 %) were found for L. multiflorum plants.

Fig. 3.

Fig. 3.

Open in a new tab

Scanning electron microscope photographs of Limonium ovalifolium and L. multiflorum stigmas and pollen grains. Each image is oriented so that the micropyle is towards the top of the page and the chalaza is towards the bottom. (A–D) Limonium ovalifolium. (A, B) Papillate and cob-type stigma, respectively. (C) Pollen grain with coarsely reticulate exine surface with a large reticulum (polar view) (D) Pollen grain with finely reticulate exine surface (polar view). (E–H) L. multiflorum. (E, F) Papillate and cob-like stigmas, respectively. (G) Bilateral pollen grains showing perforate exine surface with colpi position denounced. (H) Pollen grains with verrucate exine surface.

Pollination experiments in both self-incompatible A-type and B-type L. ovalifolium plants showed that scapes produced on average fewer seeds (<2 seeds, n = 12 plants) than L. multiflorum (>100 seeds, n = 6 plants) (Supplementary Data Table S2). No significant differences in seed production were found between L. ovalifolium plant types considering both control and bagged plants (Supplementary Data Table S2; P > 0·929, α < 0·05) or between L. multiflorum bagged and unbagged plants (Supplementary Data Table S2; P > 0·873, α < 0·05). However, significant differences between L. ovalifolium and L. multiflorum for the number of seeds produced either in control or bagged scapes were detected (Supplementary Data Table S2; P = 0·044, α < 0·05). The percentage of seed germination was similar (∼55 %) in the two species.

Cross-pollination experiments between pairs of self-incompatible L. ovalifolium plants (A × B combinations) resulted in the production of a small amount of seeds on average (Supplementary Data Table S2), with a moderate percentage of germination (62 %).

Ovule development and gamete formation

In both L. ovalifolium and L. multiflorum unopened flowers (<2·5 mm) exhibited five common primordia, which were initiated simultaneously on a flattened apex after calyx initiation (data not shown). Then, the gynoecial primordia attained continuity and developed subsequently into five distinct styles and stigmata, and thereafter the remaining apex was transformed into the single ovule primordium (Supplementary Data Fig. S3A). During development of the unique basal ovule, a long and slender funicle developed around it (Supplementary Data Fig. S3B, C) and the ovule underwent curvature of 360 ° (ovule circinotropous) (Supplementary Data Fig. S3D). In the mature ovule, the micropyle pointed away from the basal placenta towards the top of the ovary and was in close contact with the obturator (Supplementary Data Fig. S3E, F). This transmitting tissue protruded from the roof of the ovary into the unilocular superior ovary (data not shown). In general, a single ovary yielded just a single basal ovule, but ovaries with two ovules were occasionally found.

During the development of ovules, megasporogenesis and megagametogenesis took place to generate female gametophytes of sexual and/or apomictic origin. For both L. ovalifolium and L. multiflorum ∼700 ovules were examined, and there were significant differences between species for the number of ovules of sexual and apomictic origin produced (χ2 =1270·701; P < 0·05). The first meiosis was always observed in unopened flowers smaller than 2·5 mm. In L. ovalifolium, tetrasporic embryo sac development of the Gagea ova type was the most common (>85 %) (Fig. 4; Table 3). In brief, at the beginning of meiotic interphase, megasporocytes showed nuclei with two nucleoli (Fig. 4A) but soon progressed to prophase I, where only one nucleolus was visible (Fig. 4B). Then, female gametophytes exhibited two nuclei in division, not partitioned into separate cells (Fig. 4C), and following meiosis II a linear tetrad was formed (Fig. 4D). At the beginning of gametogenesis, the coenomegasporocyte attained a bipolar 1 + 3 nuclear arrangement, with one nucleus in a micropylar-most position, with the other three located towards the chalazal pole (Fig. 4E). At this stage the gametophyte exhibited prominent vacuoles between each nucleus, and no cytokinesis or nucleus degeneration occurred. At the micropylar pole the single megaspore nucleus underwent a nuclear mitotic division, while at the chalazal pole the three megaspore nuclei fused to produce a triploid restitution nucleus which further divided into a secondary tetranucleate embryo sac (Fig. 4F). Thereafter, the nucleus at the micropylar-most pole divided again, originating the cells of the egg apparatus, the synergids and egg cell, while the chalazal pair of nuclei did not divide further. Thus, the mature embryo sacs were six-nucleate (Fig. 4G). Other tetrasporic-origin embryo sacs were also produced, namely of the Adoxa (Fig. 4H, I) and Drusa types (Fig. 4J). Occasionally, unreduced, diplosporic (apomictic) embryo sacs of Rudbeckia type were formed (Table 3).

Fig. 4.

Fig. 4.

Open in a new tab

Megaspore and embryo sac formation in Limonium ovalifolium (differential interference contrast). Each image is oriented so that the micropyle is towards the top of the page and the chalaza is towards the bottom. (A–D) Plant with accession number 2009I4SR (Sra Remédios, Peniche). (E, F) Plant 2009I1PL (Praia da Luz, Lagos). (G–I) Plant 2009I2CS (Cabo de Sagres, Vila do Bispo). (J) Plant 2009I4SR (Sra Remédios, Peniche). (A) Megaspore mother cell at interphase with two nucleoli (arrowed). (B) Megaspore mother cell in meiosis I (prophase I) showing a single nucleolus at the chalazal pole. (C) Two nuclei in division (meiosis II, anaphase II) not partitioned into separate cells. (D) Coenomegaspore with four reduced nuclei, each nucleus with a single nucleolus (arrowed) (four-nucleated stage; tetrasporic development). (E) Megaspores in a bipolar 1 + 3 arrangement (arrowed) with the embryo sac showing extended vacuolization (v). (F) Embryo sac at secondary tetranucleate stage upon division of the restitution nucleus (arrows indicate mononucleolated nuclei [micropylar pole] and trinucleolated nuclei [chalazal pole]); (G) Six-nucleate embryo sac; the lowest nuclei of the secondary tetranucleate stage do not further divide (Gagea ova type). (H) Adoxa-type embryo sac. Arrows indicate egg cell and antipodals. (I) Detail of an embryo sac of Adoxa type with superior and inferior polar nuclei arrowed. (J) Drusa-type embryo sac. Dotted oval indicates cells at the chalazal pole. Each image is oriented so that the micropyle is towards the top of the page and the chalaza is towards the bottom. Abbreviations: ant, antipodals; egg, egg cell; ii, integuments; n, nucleus; nl, nucleolus; nc, nucellus; syn, synergid; v, vacuole.

Table 3.

Ovule development in Limonium ovalifolium and Limonium multiflorum

Taxon Accession numbera Originb Chromosome numberc No. of ovules analysed (%)d No. of embryos (%)e No. of sexual embryo sac types (%)
 No. of apomictic, Rudbeckia-type embryo sacs (%)
Gagea ova Adoxa Drusa
L. ovalifolium 2009I4SR Sra Remédios, Peniche, Portugal 16 41·7 (304) 6·2 (45) 31·3 (228) 2·2 (16) 1·6 (12) 0·4 (3)
2009I1PL Praia da Luz, Lagos, Portugal 16 28 (204) 2·1 (15) 22·1 (161) 2·3 (17) 1·5 (11)
2009I2CS Cabo de Sagres, Vila do Bispo, Portugal 16 30·3 (221) 3·1 (23) 23·3 (170) 1·8 (13) 1 (7) 1·1 (8)
Frequency of ovules (%)f 11·4 76·7 6·3 4·1 1·5
Total number of ovules 729 83 559 46 30 11
L. multiflorum 2009I4VF Vale dos Frades, Lourinhã, Portugal 35 40·1 (298) 3·5 (26) 36·6 (272)
2009I9CR Cabo Raso, Cascais, Portugal 32 35·2 (262) 4·2 (31) 31·0 (231)
2010I5FL Foz do Lizandro, Mafra, Portugal 36 24·7 (184) 2·5 (19) 22·2 (165)
Frequency of ovules (%)f 10·2 89·8
Total number of ovules 744 76 668
Open in a new tab

aAccessions from the ISA collections.

bPopulation origin.

cChromosome numbers were counted from root tip squashes (Róis et al., 2012).

dFrequency of ovules analysed = number of ovules analysed in each accession/total number of ovules analysed × 100. Numbers in parentheses are numbers of ovules/embryo sacs examined.

eFrequency of embryos = number of embryos/number of ovules analysed × 100. Numbers in parentheses are numbers of ovules/embryo sacs examined.

fFrequency of ovules = number of ovules with sexual or apomictic origin/ number of ovules analysed × 100.

By contrast, L. multiflorum only produced diplosporous, meiotically unreduced eggs of the Rudbeckia type (Table 3). During meiosis I, prophase cells showed a single nucleolus (Fig. 5A) and in metaphase plates bivalent and univalent chromosomes scattered over the spindle were observed (Fig. 5B). After anaphase I, a reconstitution nucleus was formed, irregular in shape, slightly elongated and somewhat constricted in the middle (Fig. 5C). This was followed by progression of the restitution nucleus through a mitotic anaphase, which in most cases formed a tripolar spindle (Fig. 5D). Embryo sacs at the two-nucleate stage showed one nucleus with one large nucleolus positioned at the micropylar end and another two-nucleolated nucleus at the most chalazal end (Fig. 5E). After a karyokinetic nuclear division at the two-nucleated stage, a four-nucleate gametophyte was produced (Fig. 5F). In the micropylar region, the egg cell apparatus (egg cell; two synergids; upper polar nucleus) was formed after a second mitotic division of the these nuclei (Fig. 5G). The central part of the embryo sac contained more than one protoendospermatic cell (four in the case illustrated; Fig. 5G), and at the chalazal-most region a variable number of antipodal cells with two nucleoli each were usually present (Fig. 5H). The ultimate development of the egg cell appeared to occur autonomously, as no pollen tubes were observed at the micropyle or at the extension of the gametophyte, and the first developmental stages of the embryo proceeded with perfectly intact synergids and polar nuclei (Fig. 5I).

Fig. 5.

Fig. 5.

Open in a new tab

Megaspore and embryo sac formation in Limonium multiflorum (differential interference contrast). Each image is oriented so that the micropyle is towards the top of the page and the chalaza is towards the bottom. (A–D) Plant with accession number 2009I4VF (Vale dos Frades, Lourinhã). (E, F) Plant 2009I9CR (Cabo Raso, Cascais) (G) Plant 2009I4VF (Vale dos Frades, Lourinhã). (H) Plant 2010I5FL (Foz do Lizandro, Mafra). (I) Plant 2009I4VF (Vale dos Frades, Lourinhã). (A) Megasporocyte within ovule showing one nucleolus. (B) First metaphase exhibiting laggard chromosome (arrowed). (C) Restitution nuclei after first meiotic division (meiotic diplospory). Nucleoli with distinct dimensions within the same nucleus are arrowed. (D) Tripolar spindle after restitution nucleus formation (rn). (E) Two-nucleate stage embryo sac resulting from mitotic-like division of megasporocyte, showing one nucleus with one nucleolus at the micropylar-most position and another nucleus with two nucleoli at the chalazal pole (arrowed). (F) Embryo sac at four-nucleate stage resulting from mitotic-like division of megasporocyte; arrows indicate spindles. (G) Detail of egg cell apparatus (in dotted oval) in a mature embryo sac; antipodal cell is arrowed. In this composite image the area within the dotted oval, taken at a different light exposure, was copied into the background of the figure. (H) Mature embryo sac of Rudbeckia type. (I) Parthenogenetic embryo formation within a mature sac. Abbreviations: e, embryo; ii, integuments; rn, restitution nucleus; nc, nucellus; syn, synergid.

DISCUSSION

Phylogeographical patterns

In order to compare cpDNA diversity in diploid and tetraploid plant species growing in sympatry as well as allopatry on the continental Portuguese coast, we analysed cp haplotypes in 18 populations of six Limonium species. Our results reveal that haplotype and nucleotide diversity are higher in diploids than in tetraploids, where moderate to no diversity at all is found. Remarkably, these data are concordant with previous MSAP studies on populations of these tetraploid species, in which little to no genetic and epigenetic variation is detected (Róis et al., 2013). The frequency of endemic, unique haplotypes is ≥50 % for both species complexes (Table 2), which could be interpreted as lineage differentiation within and between complexes.

Among diploid species, L. nydeggeri revealed the highest haplotype and nucleotide diversity, and in the tetraploid species complex L. multiflorum presented the highest values for these parameters. Noticeably, both species represent two Lusitania endemics, the only ones for which the entire distribution range was sampled. It is interesting to note that L. multiflorum, which has a restricted distribution on the west coast of the Iberian Peninsula, has lower haplotype diversity than L. nydeggeri, which is more widespread and is distributed in both the west and south-west coasts of Portugal (Erben, 1978, 1993, 1999). For the remaining species, haplotype diversity is most probably underestimated as they also grow on the Atlantic coasts of the British Isles and France, and on south-west Atlantic and Mediterranean coasts (Erben, 1993; EUNIS, 2014). Considering the centrally located cp haplotypes for both L. nydeggeri (H2, H11, H13 and H19) and L. multiflorum (H2), our data suggest that the Lusitania region represents the centre of diversity for these species (Figs 1 and 2).

Our network analyses demonstrated differentiation between diploid and tetraploid lineages, some species-specific lineages and suprahaplotypes that can be shared by up to five taxa. Suprahaplotype H11 is shared by the three diploid taxa, and L. lanceolatum haplotypes appear to be derived from diploid lineages. In contrast, all tetraploid species share the suprahaplotype H2, from which derived, species-specific haplotypes were found, including haplotypes H3 and H4 in L. dodartii and haplotype H1 in L. binervosum (Fig. 2). The suprahaplotype H2 is the most frequent and seems to be a central haplotype that is shared by both diploid and tetraploid species. Nevertheless, its derived haplotypes are present in distinct diploid or tetraploid lineages, with haplotype H2 as the likely common ancestral haplotype (Fig. 2). In apomictic complexes, common ancestry followed by sequence divergence has been demonstrated in e.g. Boechera (Sharbel and Mitchell-Olds, 2001; Dobeš et al., 2004; Kiefer et al., 2009) and Ranunculus (Paun et al., 2005; Hörandl et al., 2009). Moreover, our data suggest that tetraploids were generated independently four times: haplotypes H2 and connected haplotypes (see haplotypes coloured yellow in Fig. 2A), H7 and H9.

Reproductive findings in diploid and tetraploids

Our results on floral heteromorphies, female gametophyte development and pollination experiments further demonstrate distinct reproductive modes in diploid and tetraploid Limonium species. Although both L. ovalifolium and L. multiflorum plants exhibit the regular pentamerous flowers commonly found in Limonium (de Laet et al., 1995), L. ovalifolium displayed dimorphic flowers, confirming earlier descriptions (Erben, 1978). By contrast, L. multiflorum exhibited regular stigmas but abnormal pollen grains without the typical exine patterns, and are therefore male-sterile (Róis et al., 2012). As reported in other biological systems, poor quality of pollen due to mutation accumulation (Muller’s ratchet) relaxes selection for a male function and will result in many of the progeny sired by asexuals being weak or inviable (Mogie et al., 2007).

Here we show for the first time that L. ovalifolium mostly forms meiotically reduced tetrasporic embryo sacs (Table 3) of the G. ova type (Romanov, 1957), although tetrasporous female gametophytes of Adoxa and Drusa types are also found, as earlier reported for other Limonium spp. (Dahlgren, 1916; D’Amato, 1949; Hjelmqvist and Grazi, 1964). Moreover, a very low frequency of diplosporic embryo sacs of the Rudbeckia type (Battaglia, 1946) were also produced in this species (Table 3). Also, pollen exclusion tests in bagged self-incompatible L. ovalifolium plants result in the production of a very small number of seeds in the absence of self-compatible pollen, suggesting apomixis as a hypothetical reproductive event occurring at the diploid level. As rarely reported in literature, cases of apomixis known at the diploid level include the genus Boechera (Koltunow and Grossniklaus, 2003) and some species of Paspalum (Siena et al., 2008). At present our data are insufficient to provide solid evidence of apomixis at diploid level and need further conclusive support from genetic analyses, which are ongoing with SSRs (simple sequence repeats).

In contrast, in L. multiflorum meiotically unreduced, diplosporous embryo sacs of the Rudbeckia type (Battaglia, 1951) are produced and autonomous apomixis was observed. In other polyploid apomicts (e.g. Statice oleaefolia var. confusa), the formation of reduced embryo sacs of the Adoxa type occurs in parallel with unreduced apomictic embryo sacs of the Erigeron type (facultative apomixis; D’Amato, 1949). Also, polyploid Limonium transwallianum forms reduced tetrasporous Fritillaria and Drusa types and unreduced diplosporous embryo sacs (Hjelmqvist and Grazi, 1964). However, tetraploid L. multiflorum only originates apomictic embryo sacs. Moreover, pollen exclusion tests reveal that this species produces a moderate to high number of viable seeds per scape, even in the presence of pollen grains with various morphological abnormalities. Therefore, we hypothesize that endosperm formation is pollen-independent, as has been found in other apomicts (Aliyu et al., 2010). Cytohistological and scanning electron microscopy studies of mature seeds of L. ovalifolium and L. multiflorum showed that the embryo was enveloped by an endosperm tissue that was formed exclusively of polygonal starch grains with a well-defined hilum in the centre, often with radiating clefts (Róis et al., 2012).

Interestingly, a common origin of apomixis and tetraspory has been hypothesized based on the identification of similar cytological characteristics during ovule development (Carman et al., 1991; Carman, 1997; Peel et al., 1997a, b). For example, both tetrasporic and diplosporic embryo sacs form precociously due to ‘precocious gametophytization’ of the megaspore mother cell (MMC) (Battaglia, 1989), meaning that MMCs undergo vacuolization, which normally occurs later in the surviving megaspore (Crane and Carman, 1987), as we also detected here.

Conclusions: linking cpDNA distribution patterns with reproductive modes – evidence of geographical parthenogenesis

Phylogeographical analysis of diploid and tetraploid Limonium spp. shows that diploids are more widespread along the Portuguese coast than tetraploids (Figs 1 and 2). Interestingly, tetraploids of agamospermous L. binervosum have a broad distribution in higher latitudes, extending to the coasts of north-west France and the British Isles (Ingrouille and Stace, 1985, 1986; Lahondère and Biorét, 1995, 1996). Even if northern populations of L. binervosum species complex (British Isles) are all tetraploid, diploid L. ovalifolium extends north to the coast of France and overlaps there only with tetraploid L. dodartii. Hence, it seems that the observed geographical pattern is not a strictly exclusive pattern of ploidy levels. Moreover, for the Limonium coastal species studied it appears that latitudinal ice sheet limits, but no altitudinal ones, during the quaternary glaciations (Hewitt, 2011) are involved in the observed geographical pattern by permitting the northern expansion of southern populations/species to suitable but non-optimal habitats, with the subsequent differentiation/adaptation of these front-edge populations.

Indirect estimates of Limonium breeding systems based on pollen–stigma dimorphisms (Baker, 1953c; Erben, 1978), in addition to our data, show that L. ovalifolium are outcrossers with hypothetically very rare apomictic events, whereas male-sterile L. multiflorum plants (Róis et al., 2012) reproduce through autonomous apomixis. In general, the majority of species of the L. binervosum group are characterized by self-incompatible monomorphic pollen–stigma combinations, and pollen is either not produced or is of very low stainability (0–38 %), although well-stained large pollen grains are also observed (Ingrouille and Stace, 1985). Even though it has been suggested species in this group are agamosperms (Ingrouille and Stace, 1985), tetraploid L. transwallianum (belonging to the L. binervosum complex) is a facultative apomict that produces both reduced tetrasporous and unreduced diplosporous embryo sacs (Hjelmqvist and Grazi, 1964). Thus, it appears that in this species complex both obligate and facultative apomixis may occur. Unfortunately, flow cytometric seed screening studies in diploid and tetraploid Limonium species did not enable the determination of their modes of reproduction since each mature seed profile was only characterized by a single embryo peak (Róis et al., 2012). Despite our use of high-throughput methods to measure seed formation, we were unable to identify sexual versus apomixis seed production, in addition to autonomous apomixis.

The occurrence of autonomous apomixis, where the need for pollen is completely eliminated, is taxonomically infrequent and occurs in species characterized by ‘geographical parthenogenesis’, as in many Asteraceae (Noyes, 2007; Hörandl et al., 2008). Autonomous apomixis appears to have a role in colonization scenarios (Baker, 1955, 1967; Baker and Stebbins, 1965) and genetic bottlenecks (Hörandl, 2010). Here, the phylogeographical patterns of some tetraploid Limonium species suggest the formation of clones restricted to single sites, e.g. L. binervosum (populations CM, SJ), L. dodartii (populations PCo, PV) or L. multiflorum (population FL). In these populations it is possible that uniparental reproduction might alleviate the reduced fecundity resulting from pollen limitation and compensate for the lack of suitable pollinators in the colonized region (Rambuda and Johnson, 2004) after long-distance dispersal of seeds by birds, wind and water along the coast. In conclusion, our phylogeographical and reproductive biology findings suggest a pattern of geographical parthenogenesis in diploid and tetraploid coastal Limonium species.

SUPPLEMENTARY DATA

Supplementary data are available online at www.aob.oxfordjournals.org and consist of the following. Table S1: the intron trnL and intergenic spacer trnL–trnLF sequence haplotypes of L. ovalifolium and L. binervosum species complexes. Table S2: pollen exclusion tests and crossing experiments in L. ovalifolium and L. multiflorum. Figure S1: median-joining network of cpDNA sequence regions in populations of L. ovalifolium and L. binervosum species complexes. Figure S2: Bayesian phylogenetic analysis of haplotypes. Figure S3: scanning electron microscopy photographs of floral and ovule development in L. multiflorum and L. ovalifolium.

Supplementary Data
supp_117_1_37__index.html (928B, html)

ACKNOWLEDGEMENTS

The authors wish to thank the Instituto da Conservação da Natureza e Florestas (ICNF) for permission to collect samples, Lúcia Gomes for help in reproductive studies, and Joana Firmino for help with molecular work, Telmo Nunes (CBA) for technical support on scanning electron microscope images and Pedro Arsénio (LEAF) for mapping with ArcGIS, and the helpful comments of S. J. Hiscock and two anonymous reviewers. The Portuguese Research Foundation (FCT) provided financial support for our work under projects Pest-OE/AGR/UI0240/2011 and Incentivo/AGR/UI0240/2014 and provided a grant to A.S.R. (SFRH/BD/62542/2009). The authors declare that they have no competing interests. A.D.C. designed and coordinated the study. A.D.C., A.S.R., A.P.P. and F.S. performed the plant sampling. A.S.R. and F.S. performed the molecular genetic studies. A.S.R. and O.S.P. processed the raw data and carried out the statistical analysis. A.S.R. performed the cytoembryological studies. G.T. performed the electron microscopy work. F.S. and A.P.P. performed the flower heteromorphies analyses, pollen exclusion and crossing tests and pollen studies. A.D.C., A.S.R., O.S.P. and T.F.S. drafted the manuscript. D.E.S. revised the manuscript. All authors read and approved the manuscript.

LITERATURE CITED

  1. Alexander MP. 1969. Differential staining of aborted and non-aborted pollen. Stain Technology 44: 117–122. [DOI] [PubMed] [Google Scholar]
  2. Aliyu OM, Schranz ME, Sharbel TF. 2010. Quantitative variation for apomictic reproduction in the genus Boechera (Brassicaceae). American Journal of Botany 97: 1719–1731. [DOI] [PubMed] [Google Scholar]
  3. Asker SE, Jerlin L. 1992. Apomixis in plants. Boca Raton: CRC Press. [Google Scholar]
  4. Avise JC. 2000. Phylogeography: the history and formation of species. Harvard University Press. [Google Scholar]
  5. Baker HG. 1948. Dimorphism and monomorphism in the Plumbaginaceae: I. A survey of the family. Annals of Botany 12: 207–219. [Google Scholar]
  6. Baker HG. 1953a. Dimorphism and monomorphism in the Plumbaginaceae: II. Pollen and stigmata in the genus Limonium. Annals of Botany 17: 433–446. [Google Scholar]
  7. Baker HG. 1953b. Dimorphism and monomorphism in the Plumbaginaceae: III. Correlation of geographical distribution patterns with dimorphism and monomorphism in Limonium. Annals of Botany 17: 615–628. [Google Scholar]
  8. Baker HG. 1953c. Race formation and reproductive method in flowering plants. Symposia of the Society for Experimental Biology 7: 114–145. [Google Scholar]
  9. Baker HG. 1966. The evolution, functioning and breakdown of heteromorphic incompatibility systems. I. The Plumbaginaceae. Evolution 20: 349–368. [DOI] [PubMed] [Google Scholar]
  10. Baker HG. 1967. Support for Baker’s law – as a rule. Evolution 21: 853–856. [DOI] [PubMed] [Google Scholar]
  11. Baker HG, Stebbins GL. 1965. The evolution of colonizing species. New York: Academic Press. [Google Scholar]
  12. Bandelt HJ, Forster P, Röhl A. 1999. Median-joining networks for inferring intraspecific phylogenies. Molecular Biology and Evolution 16: 37–48. [DOI] [PubMed] [Google Scholar]
  13. Battaglia E. 1946. Ricerche cariologiche e embriologiche sul genere Rudbeckia (Asteraceae). VII: Apomissa in Rudbeckia laciniata L. ed amfimissia nella sua varieta’ a fiori doppi. Nuovo Giornale Botanico Italiano 53: 437–482. [Google Scholar]
  14. Battaglia E. 1951. Development of angiosperm embryo sacs with non-haploid eggs. American Journal of Botany 1: 1–47. [Google Scholar]
  15. Battaglia E. 1989. Embryological questions. 14. The evolution of the female gametophyte of angiosperms: an interpretative key. Annals of Botany 47: 7–144. [Google Scholar]
  16. Bierzychudek P. 1985. Patterns in plant parthenogenesis. Experientia 41: 1255–1264. [DOI] [PubMed] [Google Scholar]
  17. Caperta AD, Rosa M, Delgado M, et al. 2008. Distribution patterns of phosphorylated Thr 3 and Thr 32 of histone H3 in plant mitosis and meiosis. Cytogenetic and Genome Research 122: 73–79. [DOI] [PubMed] [Google Scholar]
  18. Caperta AD, Santo DE, Silva V, et al. 2014. Habitat specificity of a threatened and endemic, cliff-dwelling halophyte. AoB PLANTS 6: plu032; doi:10.1093/aobpla/plu032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Carman JG. 1997. Asynchronous expression of duplicate genes in angiosperms may cause apomixis, bispory, tetraspory, and polyembryony. Biological Journal of the Linnean Society 61: 51–94. [Google Scholar]
  20. Carman JG, Crone CF, Riera-Lizarazu O. 1991. Comparative histology of cell walls during meiotic and apomeiotic megasporogenesis in two hexaploid Australasian Elymus species. Crop Science 31: 1527–1532. [Google Scholar]
  21. Clegg MT, Gaut BS, Learn GH, Morton BR. 1994. Rates and patterns of chloroplast DNA evolution. Proceedings of the National Academy of Sciences of the USA 91: 6795–6801. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Cortinhas AL, Erben M, Paes AP, Espírito-Santo D, Guara-Requena M, Caperta AD. 2015. Taxonomic complexity in the halophyte Limonium vulgare and related taxa (Plumbaginaceae): insights from analysis of morphological, reproductive and karyological data. Annals of Botany 115: 369–383. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Crane CF, Carman JG. 1987. Mechanisms of apomixis in Elymus rectisetus from eastern Australia and New Zealand. American Journal of Botany 74: 477–496. [Google Scholar]
  24. Dahlgren KVO. 1916. Zytologische und embryologische Studien uber die Reihen Primulales und Plumbaginales. Kungliga Svenska Vetenskapsakademiens Handlingar 56: 1–80. [Google Scholar]
  25. D’Amato F. 1940. Contributo all'embriologia delle Plumbaginaceae. Nuovo Giornale Botanico Italiano 47: 349–382. [Google Scholar]
  26. D’Amato F. 1949. Triploidia e apomissia in Statice oleaefolia Scop. var. confusa Godr. Caryologia 2: 71–84. [Google Scholar]
  27. Dobeš CH, Mitchell-Olds T, Koch MA. 2004. Extensive chloroplast haplotype variation indicates Pleistocene hybridization and radiation of North American Arabis drummondii, A. × divaricarpa, and A. holboellii (Brassicaceae). Molecular Ecology 13: 349–370. [DOI] [PubMed] [Google Scholar]
  28. Erben M. 1978. Die Gattung Limonium im südwestmediterranen Raum. Mitteilungen der Botanischen Staatssammlung München 14: 361–626. [Google Scholar]
  29. Erben M. 1993. Limonium Mill. In: Castroviejo S, Aedo C, Cirujano S, et al. eds. Flora Iberica, Vol. 3 Madrid: Real Jardín Botânico, CSIC, 2–143. [Google Scholar]
  30. Erben M. 1999. Limonium nydeggeri – eine neue Art aus Sudwestportugal. Sendtnera 6:103–107. [Google Scholar]
  31. Espírito-Santo MD, Caperta AD, Alves HN, Moreira I. 2012. Plantas endémicas do litoral de Portugal Continental. In: Monteiro A, Gomes da Silva F, Jorge R, eds. Gestão e conservação da flora e da vegetação de Portugal e da África Lusófona. In Honorium do Professor Catedrático Emérito Ilídio Rosário dos Santos Moreira. Lisbon: IsaPress, 267–302. [Google Scholar]
  32. EUNIS [European Nature Information System] . 2014. Species mentioned in legal texts. http://eunis.eea.europa.eu/species/4039 (28 October 2014). [Google Scholar]
  33. Franco JA. 1984. Nova flora de Portugal (Continente e Açores), Vol. II, Clethraceae–Compositae. Lisbon: Author’s edition. [Google Scholar]
  34. Greuter W, McNeill J, Barrie FR, et al. eds. 2000. International Code of Botanical Nomenclature (Saint Louis Code). Regnum Vegetabile 138. Königstein: Koeltz Scientific Books. [Google Scholar]
  35. Hall TA. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symposium Series 41: 95–98. [Google Scholar]
  36. Hayat MA. 1981. Fixation for electron microscopy. New York: Academic Press. [Google Scholar]
  37. Hewitt GM. 2011. Quaternary phylogeography: the roots of hybrid zones. Genetica 139: 617–638. [DOI] [PubMed] [Google Scholar]
  38. Hjelmqvist H, Grazi F. 1964. Studies on variation in embryo sac development. Botaniska Notiser 117:141–166. [Google Scholar]
  39. Hörandl E. 2010. The evolution of self-fertility in apomictic plants. Sexual Plant Reproduction 23: 73–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Hörandl E, Cosendai A-C, Temsch EM. 2008. Understanding the geographic distributions of apomictic plants: a case for a pluralistic approach. Plant Ecology and Diversity 1: 309–320. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Hörandl E, Greilhuber J, Klímová K, et al. 2009. Reticulate evolution and taxonomical concepts in the Ranunculus auricomus complex (Ranunculaceae): insights from analysis of morphological, karyological and molecular data. Taxon 58: 1194–1215. [PMC free article] [PubMed] [Google Scholar]
  42. Hojsgaard D, Klatt S, Baier R, Carman JG, Hörandl E. 2014. Taxonomy and biogeography of apomixis in angiosperms and associated biodiversity characteristics. Critical Reviews in Plant Sciences 33: 414–427. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Ingrouille MJ, Stace CA. 1985. Pattern of variation of agamospermous Limonium (Plumbaginaceae) in the British Isles. Nordic Journal of Botany 5: 113–125. [Google Scholar]
  44. Ingrouille MJ, Stace CA. 1986. The Limonium binervosum aggregate (Plumbaginaceae) in the British Isles. Biological Journal of the Linnean Society 92: 177–217. [Google Scholar]
  45. Kearney M. 2005. Hybridization, glaciation and geographical parthenogenesis. Trends in Ecology & Evolution 20: 495–502. [DOI] [PubMed] [Google Scholar]
  46. Khan Z, Santpere G, Traveset A. 2012. Breeding system and ecological traits of the critically endangered endemic plant Limonium barceloi (Gil and Llorens) (Plumbaginaceae). Plant Systematics and Evolution 298: 1101–1110. [Google Scholar]
  47. Kiefer C, Dobeš C, Sharbel T, Koch MA. 2009. Phylogeographic structure of the chloroplast DNA gene pool in North American Boechera – a genus and continental-wide perspective. Molecular Phylogenetics and Evolution 52: 303–311. [DOI] [PubMed] [Google Scholar]
  48. Koch MA, Dobeš C, Mitchell-Olds T. 2003. Multiple hybrid formation in natural populations: concerted evolution of the internal transcribed spacer of nuclear ribosomal DNA (ITS) in North American Arabis divaricarpa (Brassicaceae). Molecular Biology and Evolution 20: 338–350. [DOI] [PubMed] [Google Scholar]
  49. Koltunow AM, Grossniklaus U. 2003. Apomixis: a developmental perspective. Annual Review of Plant Biology 54: 547–574. [DOI] [PubMed] [Google Scholar]
  50. Kubitzki K. 1993. Plumbaginaceae. In: Kubitzki K, Rohwer JG, Bittrich V, eds. The families and genera of vascular plants. Vol. 2 Berlin: Springer, 523–530. [Google Scholar]
  51. de Laet J, Clinckemaillie D, Smets E. 1995. Floral ontogeny in the Plumbaginaceae. Journal of Plant Research 108: 289–304. [Google Scholar]
  52. Lahondère C, Bioret F. 1995. Contribution à l'étude morphologique, chorologique et phytosociologique des espèces à nervation parallèle du genre Limonium du littoral atlantique, de la Baie du Mont Saint-Michel à la frontière espagnole. Bulletin de la Société Botanique du Centre-Ouest Nord 26: 337–364. [Google Scholar]
  53. Lahondère C, Biorét F. 1996. Le genre Limonium Miller sur les côtes armoricaines. E.R.I.C.A. 8: 1–22. [Google Scholar]
  54. Lledó MD, Crespo MB, Fay MF, Chase MW. 2005. Molecular phylogenetics of Limonium and related genera (Plumbaginaceae): biogeographical and systematic implications. American Journal of Botany 92: 1189–1198. [DOI] [PubMed] [Google Scholar]
  55. Matzk F, Meister A, Schubert I. 2000. An efficient screen for reproductive pathways using mature seeds of monocots and dicots. Plant Journal 21: 97–108. [DOI] [PubMed] [Google Scholar]
  56. Majeský L', Vašut RJ, Kitner M, Trávníček B. 2012. The pattern of genetic variability in apomictic clones of Taraxacum officinale indicates the alternation of asexual and sexual histories of apomicts. PLoS ONE 7: e41868. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Mogie M, Britton NF, Stewart-Cox JA. 2007. Asexuality, polyploidy and the male function. In: Hörandl E, Grossniklaus U, van Dijk P, Sharbel TF, eds. Apomixis, evolution, mechanisms and perspectives. Regnum Vegatabile 147. Ruggell, Liechtenstein: A. R. Gantner, 195–214. [Google Scholar]
  58. Nylander JAA. 2004. Bayesian phylogenetics and the evolution of gall wasps. Comprehensive summaries of Uppsala Dissertations from the Faculty of Science and Technology. ISSN 1104-232X, 937. [Google Scholar]
  59. Noyes RD. 2007. Apomixis in the Asteraceae: diamonds in the rough. Functional Plant Science and Biotechnology 1: 207–222. [Google Scholar]
  60. Palacios C, Rosselló JA, González-Candelas F. 2000. Study of the evolutionary relationships among Limonium species (Plumbaginaceae) using nuclear and cytoplasmic molecular markers. Molecular Phylogenetics and Evolution 14: 232–249. [DOI] [PubMed] [Google Scholar]
  61. Palop-Esteban M, Segarra-Moragues JG, González-Candelas F. 2007. Historical and biological determinants of genetic diversity in the highly endemic triploid sea lavender Limonium dufourii (Plumbaginaceae). Molecular Ecology 16: 3814–3827. [DOI] [PubMed] [Google Scholar]
  62. Paun O, Lehnebach C, Johansson JT, Lockhart PJ, Hörandl E. 2005. Phylogenetic relationships and biogeography of Ranunculus and allied genera (Ranunculaceae) in the Mediterranean region and in the European Alpine System. Taxon 54: 911–930. [Google Scholar]
  63. Peel MD, Carman JG, Leblanc O. 1997a. Megasporocyte callose in apomictic buffelgrass, Kentucky bluegrass, Pennisetum squamulatum Fresen, Tripsacum L., and weeping lovegrass. Crop Science 37: 724–732. [Google Scholar]
  64. Peel MD, Carman JG, Liu ZW, Wang RR-C. 1997b. Meiotic anomalies in hybrids between wheat and apomictic Elymus rectisetus (Nees in Lehm.) A. Löve and Connor. Crop Science 37: 717–723. [Google Scholar]
  65. Petit RJ, Vendramin GG. 2007. Plant phylogeography based on organelle genes: an introduction. In: Weiss S, Ferrand N, eds. Phylogeography of Southern European refugia. Evolutionary perspectives on the origins and conservation of European biodiversity. Springer, 23–97. [Google Scholar]
  66. Petit RJ, Aguinagalde I, de Beaulieu J-L, et al. 2003. Glacial refugia: hotspots but not melting pots of genetic diversity. Science 300: 1563–1565. [DOI] [PubMed] [Google Scholar]
  67. Pignatti S. 1971. Studi sui Limonium, VIII. In: Heywood VH, eds. Florae Europaea. Notulae systematicae ad flora Europaeam spectantes 11. Botanical Journal of the Linnean Society 64: 353–381. [Google Scholar]
  68. Pina-Martins F, Paulo OS. 2008. Concatenator: sequence data matrices handling made easy. Molecular Ecology Resources 8: 1254–1255. [DOI] [PubMed] [Google Scholar]
  69. Rambuda TD, Johnson SD. 2004. Breeding systems of invasive alien plants in South Africa: does Baker’s rule apply? Diversity and Distributions 10: 409–416. [Google Scholar]
  70. Róis AS, Teixeira G, Sharbel TF, et al. 2012. Male fertility versus sterility, cytotype, and DNA quantitative variation in seed production in diploid and tetraploid sea-lavenders (Limonium sp., Plumbaginaceae) reveal diversity in reproduction modes. Sexual Plant Reproduction 25: 305–318. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Róis AS, López CMR, Cortinhas AL, et al. 2013. Epigenetic rather than genetic factors may explain phenotypic divergence between coastal populations of diploid and tetraploid Limonium spp. (Plumbaginaceae) in Portugal. BMC Plant Biology 13: 205. [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Romanov ID. 1957. Embryo sac in the genus Tulipa. Comptes Rendus (Doklady) de l'Académie des Sciences de l'URSS 115: 1025–1027. [Google Scholar]
  73. Ronquist F, Teslenko M, van der Mark P, et al. 2012. MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Systematic Biology 61: 539–542. [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Rozas J, Sánchez-Del Barrio JC, Messeguer X, Rozas R. 2003. DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics 19: 2496–2497. [DOI] [PubMed] [Google Scholar]
  75. Sharbel TF, Mitchell-Olds T. 2001. Recurrent polyploid origins and chloroplast phylogeography in the Arabis holboellii complex (Brassicaceae). Heredity 87: 59–68. [DOI] [PubMed] [Google Scholar]
  76. Siena LA, Sartor ME, Espinoza F, Quarin CL, Ortiz JPA. 2008. Genetic and embryological evidences of apomixis at the diploid level in Paspalum rufum support recurrent auto-polyploidization in the species. Sexual Plant Reproduction 21: 205–215. [Google Scholar]
  77. Sokal RR, Rohlf FJ. 2012. Biometry: the principles and practice of statistics in biological research, 4th edn New York: Freeman. [Google Scholar]
  78. Stace C. 2010. New flora of the British Isles, 3rd edn Cambridge: Cambridge University Press. [Google Scholar]
  79. Stelly M, et al. 1984. Mayer's hemalum-methyl salicylate: a stain-clearing technique for observations within whole ovules. Stain Technology 59: 155–161. [DOI] [PubMed] [Google Scholar]
  80. Taberlet P, Gielly L, Patou G, Bouvet J. 1991. Universal primers for amplification of three noncoding regions of chloroplast DNA. Plant Molecular Biology 17: 1105–1109. [DOI] [PubMed] [Google Scholar]
  81. Wallis T.E. 1957. Analytical microscopy. London: J & A Churchill. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Data
supp_117_1_37__index.html (928B, html)
supp_mcv138_aob-15228-s01.pdf (552.5KB, pdf)

Articles from Annals of Botany are provided here courtesy of Oxford University Press

RESOURCES