The Angiosperm Terrestrial Revolution is tightly linked to key evolutionary novelties in the reproductive and vegetative biology of flowering plants (Benton et al. 2022). Among the reproductive innovations, not only flowers, but also the embryo sac and double fertilization leading to the formation of nutritive endosperm have diversified to an extent unseen in any other lineage of land plants (Friedman 1998; Benton et al. 2022). Although plant reproductive patterns have been studied extensively (Ebert and Greilhuber 2005; Caparelli et al. 2006), little is known about the variation in embryo sac and endosperm formations within plant families (Rangan 2020).
Variation in the type of embryo sac across angiosperm lineages includes developmental and structural differences of the female gametophyte. These have resulted in the delimitation of numerous types of embryo sac (e.g. Polygonum-type, Oenothera-type, Allium-type, Drusa-type, Penaea-type, etc.; Maheshwari 1937; Friedman et al. 2008). These are recognized based on the number of megaspore nuclei that initiate the gametophyte (i.e. gametophyte of a monosporic, bisporic, or tetrasporic origin), on the final number of nuclei in the mature embryo sac (i.e. gametophyte 4, 8 or 16-nucleate), and on the final disposition and function of its cells (Maheshwari 1937; Friedman et al. 2008). This variation in angiosperm gametophytes ultimately creates variation in endosperm genetics and ploidy (e.g. the degree of heterozygosity, maternal to paternal genomic ratios, etc.), which is likely subject to selection and is therefore of evolutionary importance (Friedman et al. 2008).
Approximately 80% of angiosperm species in 239 families have a Polygonum-type embryo sac (Kordyum and Mosyakin 2020), i.e. monosporic, 8-nucleate and with a micropillar pole with a three-celled egg apparatus, a chalazal pole with three antipodals and a central cell with two nuclei, giving rise to a triploid endosperm upon fertilization. The prevailing hypothesis, based on phylogeny, is that the Polygonum-type embryo sac has evolved only once in angiosperm evolution and was the ancestral condition of crown angiosperms (Friedman et al. 2008; Kordyum and Mosyakin 2020; but see Williams and Friedman 2002) for alternative hypotheses).
Over more than a hundred years, embryological observations have resulted in description of different types of embryo sac. These have been named after particular families or genera where they were first observed, but were later found also in unrelated families (Maheshwari 1937). For example, the tetrasporic Penaea-type embryo sac was first discovered in the family Penaeaceae (Stephens 1909) and later found in the Apiaceae (Håkansson 1923, 1927), Euphorbiaceae (Subba Rao 1937) and Malphighiaceae (Fagerlind 1938). These independent origins of embryo sac types support the notion that the developmental and structural characteristics of the female gametophyte are highly homoplasious among the angiosperms and therefore potentially adaptive (Friedman et al. 2008). Although the evolutionary importance of the embryo sac and the type of endosperm seems unquestionable, information on their variation within families or genera is scarce.
In the Apiaceae, species of three of its four subfamilies (i.e. Apioideae, Saniculoideae and Mackinlayoideae) have been studied embryologically, and all of them have been found to have a Polygonum-type embryo sac and therefore a 3n endosperm, with no other embryo sac type reported for these subfamilies (Jurica 1922; Håkansson 1923, 1927, 1952; Tseng 1967; Henwood and Hart 2001). However, within the Azorelloideae, the few species studied display a tetrasporic, 16-nucleate, Drusa-type embryo sac and 3n endosperm (in Bowlesia incana and Drusa oppositifolia) or a tetrasporic, 16-nucleate, Penaea-type embryo sac and 5n endosperm (in Azorella trifurcata and A. spinosa; Håkansson 1923, 1927, 1952; Tseng 1967; Henwood and Hart 2001). This variation within the Azorelloideae makes this subfamily a good model group for studying the evolutionary importance of the embryo sac and endosperm in its diversification, and particularly within Azorella, the richest genus within the subfamily.
Azorella species are dominant cushion-forming plants in the Andes and in Patagonia. Moreover, their nursing effect on other high-elevation plant species makes them key in the maintenance of the biodiversity of these regions (Martinez 1989; Nuñez et al. 1999; Sklenář 2009; Calviño et al. 2016; Fernández et al. 2016, 2017a). Based on phylogeny, the genus is currently treated in a broad sense, that is, encompassing the former smaller genera Huanaca Cav., Laretia Gilliesh & Hook., Mulinum Pers., Schizeilema (Hook.f.) F. Muell. and Stilbocarpa (Hook.f.) A. Gray (Plunkett and Nicolas 2017; Fernández and Calviño 2019; Fernandez et al. 2020). Azorella sensu Fernández and Calviño (2019) comprises 58 species divided between two main monophyletic lineages formalized as subgenera: the Austral lineage (subgenus Azorella), which includes 23 species distributed in Austral regions of the Southern Hemisphere, including Patagonia, Australia, New Zealand and the subantarctic islands, and the Andean-Patagonian lineage (subgenus Andinae) comprising 35 species mainly of South American distribution in high-elevation regions of the Andes and Patagonia (Fernández and Calviño 2019). The genus is further divided into ten sections (corresponding to the main sub-lineages; Plunkett and Nicolas 2017) that represent monophyletic groups or species of uncertain position (Andersson et al. 2006; Nicolas and Plunkett 2009, 2012; Fernández et al. 2017a, b; Plunkett and Nicolas 2017; Fernández and Calviño 2019). Morphological synapomorphies support the monophyly and diversification of subgenus Andinae, specifically a woody cushion habit (Fernández et al. 2017a). However, no morphological synapomorphies have been described for the whole genus (Fernández et al. 2016, 2017c). Given the above-mentioned embryological observations within the Azorelloideae, a tetrasporic, 16-nucleate embryo sac is suspected as synapomorphic for the whole subfamily, although this has not been examined empirically for most of its genera, including its most speciose genus Azorella. This and the phylogenetic positions of the azorelloids (Nicolas and Plunkett 2009) studied embryologically so far led us to hypothesize that the Penaea-type and therefore a higher ploidy level (5n) of endosperm are synapomorphies of the genus Azorella.
In this study, we sampled nineteen species of Azorella and two species of Pozoa. This amounts to a five-fold increase in the number of species screened for embryo sac and endosperm variation within the Azorelloideae. We combined, for the first time, classical embryological techniques with quantitative flow cytometric seed screening to analyse the type of embryo sac and endosperm ploidy level in each species. The resulting data allowed us to test our aforesaid hypothesis regarding the evolution of the embryo sac and endosperm within the Azorelloideae.