A cornucopia of diversity—Ranunculales as a model lineage

Abstract

The Ranunculales are a hyperdiverse lineage in many aspects of their phenotype, including growth habit, floral and leaf morphology, reproductive mode, and specialized metabolism. Many Ranunculales species, such as opium poppy and goldenseal, have a high medicinal value. In addition, the order includes a large number of commercially important ornamental plants, such as columbines and larkspurs. The phylogenetic position of the order with respect to monocots and core eudicots and the diversity within this lineage make the Ranunculales an excellent group for studying evolutionary processes by comparative studies. Lately, the phylogeny of Ranunculales was revised, and genetic and genomic resources were developed for many species, allowing comparative analyses at the molecular scale. Here, we review the literature on the resources for genetic manipulation and genome sequencing, the recent phylogeny reconstruction of this order, and its fossil record. Further, we explain their habitat range and delve into the diversity in their floral morphology, focusing on perianth organ identity, floral symmetry, occurrences of spurs and nectaries, sexual and pollination systems, and fruit and dehiscence types. The Ranunculales order offers a wealth of opportunities for scientific exploration across various disciplines and scales, to gain novel insights into plant biology for researchers and plant enthusiasts alike.

Ranunculales are a morphologically hyperdiverse plant lineage. Here, we review their phylogeny, genetic resources, habitat range, and diversity in floral morphology to introduce this fascinating order to non-specialists.

Introduction

Ranunculales are the sister order to all other eudicots and have diverged before the core eudicots, which include approximately three-quarters of all angiosperms species (The Angiosperm Phylogeny Group, 2016). Studying Ranunculales can thus provide clues to the core eudicot’s ancestral states in terms of morphology and genetics. The order Ranunculales encompasses >4500 species and is composed of seven families: Ranunculaceae, Berberidaceae, Menispermaceae, Lardizabalaceae, Circeasteraceae, Papaveraceae, and Eupteleaceae (Fig. 1; The Angiosperm Phylogeny Group, 2016). They are remarkably diverse in terms of floral and fruit form, life history traits, leaf shape, growth shape, and their secondary metabolite composition. The flowers of Ranunculales are not only unusually diverse in their morphology, they are also unique in concentrating a variety of evolutionary transitions, such as changes in merism (number of floral organs), in phyllotaxy (whorled versus spiral) potentially leading to the emergence of organ fusion (in reproductive organs and perianth), and in the origin of novel organs. These transitions are only rarely observed in monocot or core eudicot model lineages. Further, Ranunculales exhibit a suite of homoplasious characters (shared character states that did not arise from a direct common ancestor, but independently via convergent evolution) such as transitions between sexual systems and pollination modes in closely related taxa, petal loss, spur formation, or transition to zygomorphy (Endress, 1995; Soza et al., 2012; Damerval and Becker, 2017; Becker et al., 2023). Homoplasies and the emergence of novel organs provide premier opportunities to study the molecular and genetic mechanisms involved in the origin of these special traits using species within Ranunculales as case studies.

Fig 1.

Simplified phylogeny of Ranunculales based on Wang et al. (2009), Ortiz et al. (2016), and Peng et al. (2023). Species for which major genomic resources are or will become available in the near future are next to their respective branches. Representative photos of Ranunculales flowers: (A) Aquilegia coerulea, (B) Thalictrum thalictroides, (C) Nigella damascena, (D) Staphisagria picta (Ranunculaceae), (E) Epimedium grandiflorum (Berberidaceae), (F) Pteridophyllum racemosum, (G) Capnoides sempervirens, (H) Eschscholzia californica, (I) Macleaya cordata, (J) Papaver somniferum (Papaveraceae). (Photo credit: A, D–G, H, J, Becker lab; B, Di Stilio lab; C, F, Jabbour; I, N, Pabón Mora.)

Aside from their morphological diversity, Ranunculales produce a multitude of secondary metabolites, many of them of pharmaceutical importance. Consequently, Ranunculales species have been used in traditional medicine since at least the early civilizations. Some species, such as Nigella sativa, were already mentioned in writing by, for example, Ayurveda, Siddha, Unani, Greek–Roman, Malay, Tibb-e-Nabwi, and Jewish civilizations (Heiss and Oeggl, 2005; Dabeer et al., 2022). The use of opium poppy (Papaver somniferum) as a narcotic drug dates even back to the Neolithic (Guerra-Doce, 2015; Yang et al., 2021). New World indigenous cultures used Ranunculales; for example, Navajos used Thalictrum fendleri tea during ceremonial war dance rites (Elmore, 1943) and Pomo women used Eschscholzia californica during infant weaning (Barrett, 1952). Ranunculales are a rich source of economically important phytochemicals, such as alkaloids, diterpenes, triterpenes, isoquinoline alkaloids, and cardiac as well as cyanogenic glycosides (Hao et al., 2015). These compounds contribute to a vast array of medicinal uses for different Ranunculales species, for example in ulcer treatment, and as antimicrobe and anti-inflammatory agents (Hao et al., 2015). Papaveraceae are notoriously known for their secondary metabolite diversity, and many of their compounds are essential pharmaceuticals of high economic value, including morphine, codeine, protopine, isocorydine, or berberine. Benzisoquinoline alkaloids (BIAs) in particular are well known for their analgesic, antitussive, antimicrobial, anticancer, and anti-inflammatory effect (Li et al., 2020; Avci et al., 2021; Becker et al., 2023). Of special pharmaceutical importance are morphine and codeine used as analgesics, the anticancer drug noscapine, and antibacterial compounds such as sanguinarine (Hagel and Facchini, 2013). Members of the other Ranunculales families synthesize unique and overlapping subsets of secondary metabolites (Hao et al., 2015). Consequently, a wide array of species is used as herbal extracts, even nowadays, for example in Chinese traditional medicine (Hao et al., 2015, 2017).

An informative phylogenetic position, combined with pharmacological relevance and stunning floral morphological diversity, has led to a strong research interest in the Ranunculales, resulting in the development of an array of genetic tools to aid in the investigation of gene function and regulation (Di Stilio, 2011; Becker et al., 2023). The powerful combination of genetic studies, comparative morphology, and secondary metabolite profiling will further enable the reconstruction of ancestral traits before the major core eudicot radiation.

In this review, we present an update on the phylogeny, fossil records, and ecology of Ranunculales, before adressing recent findings concerning the genetic origin, diversity, and evolution of floral and fruit traits. We also recapitulate the available omics resources and functional tools, and introduce the RanOmics project, aiming at selecting phylogenetically informative species to unravel the evolution of ecologically and economically important traits.

Genetic resources and functional tools for Ranunculales

In the ‘omics’ era, several genetic resources have been established for Ranunculales, mostly for mining genes related to secondary metabolite biosynthesis and regulation. The number of high-quality Ranunculales genomes, starting with the first sequenced genome from Macleaya cordata (Liu et al., 2017), has increased enormously in the past few years, allowing for comparative genome analysis (Fig. 1; Table 1). However, the suitability criteria for high quality reference genomes are unclear, hence we define them here as follows: the rate of Benchmarking Universal Single-Copy Orthologs (BUSCO; Manni et al., 2021) matches should be >95%. Table 1 shows that only two Ranunculales genomes match this criterion, these are P. somniferum (opium poppy) and Corydalis tomentella (Guo et al., 2018; Xu et al., 2022). Genomes with lower BUSCO values are available for Thalictrum thalictroides, Coptis chinensis, Aquilegia coerulea, and Aquilegia oxysepala (Ranunculaceae), Kingdonia uniflora (Circeasteraceae), Akebia trifoliata (Lardizabalaceae), Epimedium pubescens (Berberidaceae), Eschscholzia californica, Corydalis tomentella, Papaver somniferum, Papaver rhoeas, Papaver setigerum, and Macleaya chordata (Papaveraceae) (Liu et al., 2017, 2021; Filiault et al., 2018; Hori et al., 2018; Sun et al., 2020; Xie et al., 2020; Arias et al., 2021; Chen et al., 2021; Huang et al., 2021; Yang et al., 2021; Shen et al., 2022).

Table 1.

Genomic resources for RanOmics Ranunculales species

Species	Genome available	BUSCOa	Cell/callus culture system	Cell culture transformation system	Stable transformation	Virus-induced-gene silencing	Spontaneous mutants
Papaveraceae
Eschscholzia californica b	Hori et al. (2018)	Not given	Hauschild et al. (1998)	Fujii et al. (2007)	Park and Facchini, (2000b); Lotz et al. (2022)	Wege et al. (2007)	Lange et al., 2013; Wakelin et al. (2003); Barrell et al. (2010); Pollack et al. (2019); Conner and Barrell (2014)
Papaver somniferum b	Guo et al. (2018)	95.3	Eilert et al. (1985)	Belny et al. (1997)	Park and Facchini, (2000a); Chitty et al. (2003)	Hilemann et al. (2005)	Singh et al. (2017); Pathak et al. (2013); Dhawan et al. (2007); Belyaeva and Nevkrytaya (1979); Prajapati et al. (2001); Singh et al. (2014)
Macleaya cordata	Liu et al. (2017)	90.4	Franke and Böhm (1982)	No	Huang et al. (2017)	No
Papaver rhoeas	Yang et al. (2021)	92.8	No	No	No	No	O’Donnell et al. (1993); Foote et al., (1994)
Papaver setigerum	Yang et al. (2021)	94.5	No	No	No	no
Corydalis tomentella	Xu et al. (2022)	97.67	Holländer-Czytko et al. (1988) (Corydalis sempervirens)	No	No	No
Capnoides sempervirens b			No	No	No	Hidalgo et al. (2012)
Pteridophyllum racemosum b			No	No	No	No
Ranunculaceae
Coptis chinensis	Chen et al. (2021)	91.5	No	No	No	No
Aquilegia coerulea b	Filiaut et al. (2018)	Not given	No	No	No	Gould and Kramer (2007)	Cabin et al. (2022)
Aquilegia oxysepala	Xie et al. (2020)	93.2	No	No	No	No
Thalictrum thalictroides b	Arias et al. (2021)	84.5	Smolko and Peretti (1994)	Samanani et al. (2002) (Thalictrum flavum)	No	Di Stilio et al. (2010)	Martinez-Gómez et al. (2021); Galimba et al. (2012)
Nigella damascena b			Klimek-Chodacka et al. (2020)		No	Wang et al. (2015)	Goncalvez et al. (2013); Jabbour et al. (2021); Conde e Silva et al. (2023); Greyson and Raman (1975)
Staphisagria picta b			No	No	No	No
Hydrastis canadensis b			No	No	No	No
Cicaeasteraceae
Kingdonia uniflora	Sun et al. (2020)	90.6	No	No	No	No
Lardizabalaceae
Akebia trifoliata	Huang et al. (2021)	94.0	No	No	No	No
Berberidaceae
Epimedium pubescens	Shen et al. (2022)	89.9	No	No	No		Wang et al. (2017)
Epimedium grandiflorum b

^a Benchmarking Universal Single-Copy Orthologs (BUSCO) percentages are provided as a measure for genome completeness.

^b Reference genomes are being sequenced and/or expression atlases are being produced by the RanOmics group or are already available via Phytozome (for E. californica).

The available genomes already provide sufficient data for the inference of whole-genome duplications (WGDs) within the Ranunculales. When the genomes of P. somniferum, M. cordata, A. coerulea, and C. chinensis were analyzed in combination, one WGD was found to have probably occurred in the lineage leading to C. chinensis and A. coerulea, and another one in the lineage leading to P. somniferum and M. cordata (Liu et al., 2021). An additional WGD was identified in the lineage leading to P. somniferum and P. setigerum, which is not shared by P. rhoeas and M. cordata. Moreover, the P. setigerum genome shows an additional WGD (Yang et al., 2021), most probably contributing to its large genome size, which is almost double that of the closely related P. somniferum. These recent comparative genome studies suggest that the genome duplication history of Ranunculales is most likely to be as complex as those of the core eudicots, considering that the number of sequenced Ranunculales genomes is still relatively small.

Recently, the molecular evolution of morphine biosynthesis in the Papaveraceae was unraveled by comparative genomics: the final morphine biosynthesis steps, which require the STORR gene modules, was found to be <18 million years old (Li et al., 2020). The STORR gene, coding for the key enzyme converting morphinans to morphine, originated from a translational fusion of a cytochrome P450 and an oxidoreductase enzyme that occurred after the split of P. setigerum and P. somniferum from P. rhoeas and was then duplicated in the P. setigerum-specific WGD (Li et al., 2020; Yang et al., 2021).

The 1KP project (One Thousand Plant Transcriptomes Initiative, 2019) has provided transcriptomic data for a single or a few tissues of these Ranunculales species: the Lardizabalaceae A. trifoliata; the Menispermaceae Cocculus laurifolius; the Eupteleaceae Euptelea pleiosperma; the Berberidaceae Nandina domestica and Podophyllum peltatum; the Ranunculaceae Anemone hupehensis, Anemone pulsatilla, Cimicifuga racemosa, Hydrastis canadensis, and T. thalictroides; and the Papaveraceae Argemone mexicana, Capnoides sempervirens, Ceratocapnos vesicaria, Chelidonium majus, Corydalis linstowiana, E. californica, Hypecoum procumbens, Papaver bracteatum, P. rhoeas, P. setigerum, P. somniferum, and Sanguinaria canadensis.

Additional resources have been developed for species in the Ranunculales, with the aim of elucidating gene function and molecular processes, mainly for studies in evolutionary developmental genetics of flowers and secondary metabolite analysis. The latter may be studied by inducing their production in cell culture systems, providing the cultures with standardized substrates and analyzing their products. These cell culture systems were established for the Ranunculaceae T. thalictroides and Nigella damascena (Smolko and Peretti, 1994; Klimek-Chodacka et al., 2020), and the Papaveraceae E. californica, M. cordata, Corydalis sempervirens, and P. somniferum (Franke and Böhm, 1982; Eilert et al., 1985; Holländer-Czytko et al., 1988; Hauschild et al., 1998). Even more useful are stably transformed cell culture systems, allowing the careful analysis of genes involved in the regulation of biosynthesis. Protocols for these are available for Thalictrum flavum, N. damascena, E. californica, and P. somniferum (Belny et al., 1997; Samanani et al., 2002; Fujii et al., 2007; Mohammed and Masyab, 2020).

The analysis of developmental processes can be achieved only in growing plants, ideally using knockout mutants. However, stable transformation of plants and regeneration of the transgenics is a very challenging process, often requiring labor-intensive tissue culture steps, and has thus not been established for many Ranunculales species. The notable exceptions here are E. californica, M. cordata, and P. somniferum (Park and Facchini, 2000a, b; Huang et al., 2017; Lotz et al., 2022), but publications of gene function analysis based on regenerated Ranunculales transgenics do not exist to date.

An alternative approach to down-regulate gene expression in plants is virus-induced gene silencing (VIGS), which utilizes the plant’s immune system to repress viral transcript synthesis. Specific VIGS vector systems using modified plant viruses were developed to efficiently down-regulate target genes (Dinesh-Kumar et al., 2003; Liu et al., 2022). While this is a transient approach requiring careful analysis of the manipulated plants, tissue culture is not necessary, speeding up the process of gene function analysis considerably (Dommes et al., 2019; Rössner et al., 2022). This method is available for Aquilegia coerulea, T. thalictroides, T. clavatum (on dormant tubers), T. dioicum, N. damascena, Delphinium ajacis, E. californica, P. somniferum, and Cysticapnos vesicaria (Hileman et al., 2005; Gould and Kramer, 2007; Wege et al., 2007; Di Stilio et al., 2010; Hidalgo et al., 2012; Wang et al., 2015; Zhao et al., 2023) allowing for the comparative analysis of gene function among species and the assessment of functional conservation (Di Stilio, 2011)

Phylogeny of Ranunculales

Over the past several decades, tremendous progress has been made in delimiting and elucidating phylogenetic relationships among the infraordinal taxa within Ranunculales. As currently circumscribed by molecular data, the order consists of seven monophyletic families: Berberidaceae, Circaeasteraceae, Eupteleaceae, Lardizabalaceae, Menispermaceae, Papaveraceae, and Ranunculaceae (Wang et al., 2009; The Angiosperm Phylogeny Group, 2016).

Three major clades are recovered: Eupteleaceae, Papaveraceae, and the core Ranunculales (Kim et al., 2004; Wang et al., 2009). The relationships among these three clades are not well resolved, but the majority of phylogenetic analyses recognize the monogeneric Eupteleaceae as the earliest diverging lineage with weak to moderate support (e.g. Kim et al., 2004; Worberg et al., 2007; Wang et al., 2009; Sun et al., 2017; Peng et al., 2023). Within the core Ranunculales, Circaeasteraceae and Lardizabalaceae form a clade, and Menispermaceae, Berberidaceae, and Ranunculaceae form another clade, with Berberidaceae as sister to Ranunculaceae (Kim et al., 2004; Wang et al., 2009; Sun et al., 2017; Peng et al., 2023).

The Eupteleaceae include a single genus with two species only, Euptelea pleiosperma and Euptelea polyandra (Cao et al., 2016). Genome sequence or other resources are not available for this genus.

Papaveraceae sensu latu contain four subfamilies: Fumarioideae, Hypecoideae, Papaveroideae (including Chelidonieae, Eschscholzieae, and Papavereae), and Pteridophylloideae (Hoot et al., 2015). The position of Pteridophylloideae has been controversial (reviewed by Peng et al., 2023) Recently, a complete genus-level phylogeny was built for Papaveraceae, in which Papaveroideae form a clade, whereas Pteridophylloideae, Hypecoideae, and Fumarioideae form another clade, with Hypecoideae as sister to Fumarioideae; the relationships among 91% of all currently recognized genera in the family are well resolved (Peng et al., 2023).

Circaeasteraceae consists of two monotypic genera, Circaeaster and Kingdonia, and it is the sister group to Lardizabalaceae (Wang et al., 2009; Sun et al., 2017). Within Lardizabalaceae (the sister family of Circaeasteraceae), Sargentodoxa, Decaisnea, and Sinofranchetia are successive sister taxa to the other genera (Wang et al., 2009, 2020).

Within Menispermaceae, two subfamilies are recognized: Chasmantheroideae and Menispermoideae (Ortiz et al., 2016). Chasmantheroideae comprises Coscinieae and Burasaieae, and Menispermoideae comprises eight tribes, among which Menispermeae is the earliest diverging, followed by Anomospermeae, then Limacieae. Cebatheae, Cissampelideae, Pachygoneae, Spirospermeae, and Tiliacoreae form a clade with strong support, but the relationships among these five tribes are not resolved because they might have diversified rapidly over a period of <6 million years (Wang et al., 2017; Lian et al., 2020).

Berberidaceae contain three subfamilies, Podophylloideae, Berberidoideae, and Nandinoideae, corresponding to the chromosome base numbers x=6, 7 and 8, or 10, respectively (Wang et al., 2007, 2009; Sun et al., 2018). Recently, Hsieh et al. (2021) further updated the classification system for this family at the tribal and generic levels.

Ranunculaceae (Tamura, 1965, 1993) consists of five subfamilies: Coptidoideae, Glaucidioideae, Hydrastidoideae, Ranunculoideae, and Thalictroideae (Wang et al., 2009). Most studies support Glaucidioideae as sister to the remaining taxa of the family, followed by Hydrastidoideae, then Coptidoideae (e.g. Kim et al., 2004; Wang et al., 2009, 2016; Cossard et al., 2016; Zhai et al., 2019), whereas other studies place Glaucidioideae as sister to Hydrastidoideae (Hoot et al., 1999; Soltis, 2000). Thalictroideae and Ranunculoideae are characterized by the T- and R-type chromosomes, respectively (with R-type being metacentric and T-type telocentric, with only one arm; Wang et al., 2009), but the monophyly of Ranunculoideae remains controversial. Based on eight DNA loci from three genomes, Cossard et al. (2016) placed Thalictroideae in Ranunculoideae, as sister to Adonideae. That was confirmed by a plastid phylogenomic analysis (Zhai et al., 2019), whereas a phylotranscriptomic analysis strongly supports the monophyletic Ranunculoideae (He et al., 2022). He et al. (2022) suggest that the different positions of Adonideae in the nuclear and plastid trees could result from ancient hybridization and/or subsequent introgression events. The currently recognized Ranunculoideae contains 10 tribes, which together with Thalictroideae appear to have diversified rapidly over a period of <14 million years, and perhaps in as little as 1–2 million years (Wang et al., 2016).

Fossil record

The fossil record of the Ranunculales includes nearly 800 occurrences (Xing et al., 2016), but most of them should be considered with caution. In particular, few reliable fossils have been described from the Cretaceous period (Friis et al., 2011). Three northern hemisphere fossils could illustrate the early diversification of Ranunculales during this period. The flower of Teixeiraea lusitanica from the Cretaceous [~113 million years ago (Ma)] of Portugal is considered to be part of the stem or crown of the Ranunculales without family assignment von Balthazar et al., 2005). Also, from Portugal and with similar age, the flower Kajanthus lusitanicus is the first Cretaceous occurrence of Ranunculales assigned to the family Lardizabalaceae (Mendes et al., 2014). However, a new study considers this flower as more confidently assigned to the crown group of Ranunculales, making it undefined at the family level (Schönenberger et al., 2020). The anatomy of the stem of the liana Atli mornii Smith, Little, Cooper, Burnham, and Stockey from the Late Cretaceous (77–74 Ma) of Canada allows for the identification of Ranunculales without family affinity, and reinforces the early presence of Ranunculales in Laurasia (Smith et al., 2013). However, the recent description of Santaniella lobata based on fruits and stems from the Cretaceous (Barremian/Aptian, ~125 Ma) of Brazil related to Ranunculales (Gobo et al., 2022) along with the leaf with unknown affinity but close to Ranunculales in shape, named Baderadea pinnatissecta described from the same region (Pessoa et al., 2021), could indicate a Lower Cretaceous origin of the Ranunculales in Gondwana rather than Laurasia. Nevertheless, additional data from S. lobata indicate that this fossil belongs to angiosperms without certainty about the order (Pessoa et al., 2023).

With the exclusion of the monotypic family Circaeasteraceae, the other families are represented in the fossil record (Xing et al., 2016). Although the families of Ranunculales appear to have diverged early on, no Cretaceous fossil can be confidently assigned to any extant family. The unequivocal fossils assigned to a particular family are mostly fruits, seeds, leaves, wood, and pollen from the Paleogene (Friis et al., 2011).

The family Berberidaceae is represented by ~100 fossils from the Oligocene to the Pliocene, mainly from North America and Europe, but also from Asia (Friis et al., 2011; Xing et al., 2016; Chen et al., 2020). Fossils of Berberidaceae are represented by only two genera, Mahonia and Berberis, based on leaves and seeds (Xing et al., 2016). Fossils of the Eupteleaceae family are scarce; however, Friis et al. (2011) indicate the presence of a few fossils from this family in the Northern Hemisphere from the Paleocene to the Miocene.

The fossil record of Lardizabalaceae was recently reviewed (Wang et al., 2020). During the Cenozoic, fossils attributed to this family come from the Eocene to Miocene of Europe and the USA, as well as from the Miocene of Japan and the Cenozoic of South America. Most of the fossils belong to the genus Sargentodoxa, with the exception of Decaisnea seeds from the Oligocene of Germany and a liana attributed to the family level (Wang et al., 2020).

The very diverse woody family Menispermaceae has a very abundant fossil record compared with other Ranunculales families, with many fossil fruits, leaves, and wood having been described (Jacques, 2009; Xing et al., 2016). Several Cretaceous fossils may be credible, such as the morphological genus Menispermites, but need revision (Jacques, 2009). Characteristic endocarps named ‘moonseed’ are traditionally found in North America and Europe (Jacques, 2009), and were also recently found in South America (Herrera et al., 2011; Jud et al., 2018) and Asia (Han et al., 2018, 2020) as early as the Paleocene. Within this family, a total of 44 genera have been found in the fossil record, of which 17 are extant and 27 are extinct (Jacques, 2009). This fossil record attests to a rapid and universal diversification of Menispermaceae during the Paleogene as well as a complex migration of flora during this period.

Reliable fossil record of Papaveraceae is meager and is represented, to our knowledge, only by a Corydalis from the Pliocene of Italy (Mai, 1995). The fossil record of the Ranunculaceae family, mostly based on fruits, was revised by Pigg and Devore (2005). Most of these fossils are distributed in Europe and North America, from the Paleocene to the Pliocene, and some seeds were recently found in the Pliocene of China (Huang et al., 2021).

It is noteworthy that the fossil record of Ranunculales is relatively sparse in comparison with the present diversity of the order and knowing its ancient evolutionary history. A large part of the extant diversity is represented by plants with herbaceous or climbing habitus, which have low fossilization potential (Friis et al., 2011). Moreover, the potential Ranunculales fossils from the Cretaceous are also difficult to distinguish from indirectly related early-diverging eudicot lineages (e.g. Sun et al., 2011; Pessoa et al., 2021). The Ranunculales fossil record also illustrates a well-known bias in collecting and studies in paleobotany, namely the historical focus on Europe and North America (Xing et al., 2016). Recent discoveries, particularly from South America and Asia, may strengthen the fossil record of the order in the future.

Distribution and ecological niches

The order Ranunculales comprises ~4500 species, primarily occupying temperate areas of the world, with few members cosmopolitan or reaching into the tropics. Namely, the two species of Euptelea (Eupteleaceae) occur in Japan between 400 m and 1500 m (E. polyandra) and from India to China between 900 m and 3600 m (E.3 pleiosperma; Endress, 1993). In contrast, the Papaveraceae with ~430 species in 42 genera are primarily distributed in the northern hemisphere with few exceptions, including Papaver aculeatum in South Africa, and the genus Bocconia that reaches central and South America (Kadereit, 1993). The Fumarioideae are concentrated in the Sino-Himalayan and Mediterranean regions, with occurrences in South Africa and North America (Lidén, 1993a). Both Papavereae and Chelidonieae (Papaveroideae) contain Old and New World genera. Most Papavereae in the New World inhabit western North America, while the Old World genera are concentrated in southwest and central Asia, and the Mediterranean. The genus Papaver is broadly distributed in the Old and the New World. The Eschscholzieae (Papaveroideae) are found in the New World and almost exclusively in Pacific North America. Hunnemania is present in the east of Mexico. Most Papavereae and Eschscholzieae are found in open vegetation arid and warm climates, with a few exceptions that have colonized arctic areas. Conversely, the Chelidonieae of the New World occupy regions in Northeast America with the exceptions of Bocconia (Central and South America) and Glaucium and Dicranostigma (West and Central Asia). They can inhabit dry open areas (Glaucium, Dicranostigma, and Macleaya) or deciduous forests (Hylomecon, Sanguinaria, and Stylophorum). The only species of Pteridophyllum, Pteridophyllum racemosum, is a Japanese endemic (Lidén, 1993b).

Species of Papaveraceae selected as part of the RanOmics project include: Corydalis tomentella, Capnoides sempervirens, Eschscholzia californica, Macleaya cordata, Papaver rhoeas, Papaver setigerum, Papaver somniferum, and Pteridophyllum racemosum. Corydalis tomentella is a perennial, native to China, that grows in rock crevices, between 700 m and 1000 m. The plant itself reaches 15–20 cm, it has characteristic golden yellow flowers in dense inflorescences, and it can tolerate freezing temperatures (http://www.efloras.org/flora_page.aspx?flora_id=2). Capnoides sempervirens (pale corydalis or rock harlequin) is a biennial plant from the mid-latitudes of North America where it grows on exposed ridges and rocky outcrops (Sprengelmeyer and Rebertus, 2015), and it produces monosymmetric flowers, which are exceptional in that they are in a terminal position (Hidalgo and Gleissberg, 2010). Eschscholzia californica is a small herb able to grow as annual or perennial with native ranges from Northern California to Southwestern Mexico, with cymose inflorescences and flowers with deciduous sepals and characteristic yellow petals (Becker et al., 2023). Macleaya cordata is a herbaceous, perennial native to China, Japan, and Taiwan, unusual in that it can reach sizes of up to 3 m, it spreads by rhizomes, and it has massive inflorescences of showy but apetalous flowers (Kadereit, 1993; Arango-Ocampo et al., 2016). Macleaya cordata is the source of alkaloids with broad uses as detoxifiers, antimicrobials, and insecticidals (Liu et al., 2017). All Papaver species are herbs with cymes carrying large showy flowers. Papaver somniferum is the source of opium, and its center of domestication was the Mediterranean basin (Salavert et al., 2020; Hong et al., 2022). Numerous biochemical accounts with emphasis on the production of BIAs are available for different landraces (Pei et al., 2021), and two features have been linked to domestication, namely changes in capsule dehiscence and seed size (Zohary et al., 2012). Interestingly, morphine, codeine, and thebaine are lacking in capsules of the closely related and geographically overlapping P. setigerum (La Valva et al., 1985). Albeit the two species were thought to be part of the same taxonomic unit, P. somniferum is 30–150 cm high, self-pollinated, and diploid, while P. setigerum is 60 cm high, a field weed occurring in disturbed grounds that can be diploid or tetraploid (Hammer 1977; Jesus et al., 2021). Papaver rhoeas, the red poppy, is a remarkable species with exceptional beauty that has reproduced and expanded its native range across the Mediterranean as an agricultural weed (Colledge et al., 2004). Papaver rhoeas is a self-incompatible herb, currently pollinated by bees, flies, and beetles (McNaughton and Harper, 1960; Foote, 1994). Finally, the rare P. racemosum is a herb with leaves of astonishing shape convergent to those of ferns, is only found in Japan, and it grows between 1000 m and 2000 m in coniferous forests. It shares with the rest of Papaveraceae the caducous sepals and the dimerous floral organization, despite the unusual leaf phenotypes (Lidén, 1993a, b).

The Lardizabalaceae (35 species in eight genera) are primarily present in Japan, the Sino Himalayan mountains, Central and East China, and Vietnam. All genera are woody vines in subtropical evergreen forests or warm temperate green forests. Only Lardizabala and Boquila are endemic to temperate forests of Central and South Chile (Cheng-Yih and Kubitzki, 1993). Their most prominent member is Akebia trifoliata, a deciduous to evergreen twining vine, reaching up to 10 m in height with functionally unisexual flowers. Its berries are a rich source of vitamin C and pectin, and the seeds contain a high percentage of unsaturated fatty acids; the species is widely advertised as a new fruit crop. Akebia trifoliata consists of three subspecies, all with different but overlapping distributions ranging from subtropical to temperate regions from 20 m up to 2800 m in elevation in China and Taiwan (Zhang et al., 2021; Zou et al., 2022). Only A. trifoliata is a member of the RanOmics project.

Conversely, the Circaeasteraceae (two species) are herbs. Two genera are recognized: Circaeaster is present in India, Nepal, and China, and grows in moist coniferous forests between 1200 m and 5000 m. Kingdonia uniflora, on the other hand, is endemic to China (provinces of Shaanxi, Sichuan, Gansu, and Yunnan) between 2800 m and 3200 m (Cheng-Yih and Kubitzki, 1993) and is selected as the representative species of Circaeasteraceae for the RanOmics project.

The Menispermaceae consist of ~450 species in 71 genera, including many woody climbers, and rarely trees, shrubs, or herbs. They are of cosmopolitan distribution, mostly confined to the tropical lowlands in the Old and the New World. They are extremely diverse in their habitats and found in Africa and Southeast Asia (Stephania), extra tropical North America (Cocculus and Menispermum), the Mediterranean (Cocculus), Japan (Cocculus and Stephania), and South America (Abuta and Chondrodendron) (Kessler, 1993; Ortiz et al., 2007).

Species of Berberidaceae selected as part of the RanOmics project include Epimedium grandiflorum and Epimedium pubescens. The Berberidaceae include ~650 species organized in 14–17 genera. They are herbs or woody shrubs, often a component of mesophytic forests in East Asia, Northeast America (Achlys, Diphylleia, Jeffersonia, Podophyllum, and Sinopodophyllum), Andean South America (Berberis); even desert xerophytes are found in Southwest Asia. Members of Berberis are also found in South America, from Colombia to Chile, Juan Fernandez Islands, and Argentina (Loconte, 1993). In addition, a few species of Berberis have become invasive in North America and South Africa (Keet et al., 2016). The pharmaceutically and horticulturally relevant Epimedium genus includes only herbaceous species growing mainly in woodlands. Its center of diversity is East Asia, with most species native to China. However, some species grow in the Alps, the Balkan region, Algeria, Caucasia, Japan, east Russia, and Kashmir (Zhang et al., 2022). Epimedium pubescens is native to the Chinese provinces Anhui, Jiangxi, and Sichuan (Stearn et al., 2002). Epimedium grandiflorum, a species with large flowers comprising curved nectar spurs, grows in Japan, North Korea, and South China, and varies greatly in flower color between white, light yellow, and purple-pink (Stearn et al., 2002).

The Ranunculaceae is a cosmopolitan family with ~2500 speciess in 59 genera. With large preferences for temperate or cool climates, they are a rare element in the tropics (Chartier et al., 2016). The most broadly distributed elements in northern and southern hemispheres include Anemone, Caltha, Clematis, Myosurus, Ranunculus, and Thalictrum. A total of 44 genera are present in East Asia, 24 in Europe, with few genera in temperate North America and in Highlands in South America (Tamura, 1993). Species of Ranunculaceae selected as part of the RanOmics project include Coptis chinensis, Aquilegia coerulea, Aquilegia oxysepala, Thalictrum thalictroides, Nigella damascena, Staphisagria picta, and Hydrastis canadensis.

In many phylogenies, H. canadensis or goldenseal is the sister species to all remaining Ranunculaceae. It is native to the eastern deciduous forests of North America. It grows in dense patches resulting from clonal growth via rhizome and lateral root formations (Sanders and McGraw, 2005). The rhizomes of this species are highly prized as a food supplement and as a traditional remedy for diverse conditions, including wound healing, digestive disorders, and cancer, with berberine as the pharmacologically most active ingredient (Mandal et al., 2020). Several H. canadensis populations are under serious threat caused by commercial and private harvesting of natural populations (Albrecht and McCarthy, 2006).

Sister to the Thalictroideae and Ranunculoideae are the Coptidoideae, with Coptis chinensis as a RanOmics species representative. The species has an at least 2000 year long history as traditional Chinese medicine, with berberine also as the dominant alkaloid. The rhizomes of C. chinensis are harvested, and it is cultivated in several Chinese provinces in shady, moist, and cool mountainous regions between 1200 m and 1800 m (Chen et al., 2021). Coptis chinensis is endangered in the wild and its remaining populations are found in the woodlands of central China at altitudes of 500–2000 m. This species, like H. canadensis, suffers from harvesting of the rhizomes (He et al., 2007).

Nigella damascena (commonly known as love-in-the-mist) is an annual herbaceous weedy species growing throughout the Mediterranean. As a popular ornamental plant, it was most probably distributed by seeds along ancient trade routes (Heiss and Oeggl, 2005). Interestingly, a mutant that lost petal identity and has numerous petaloid tepals was described as early as in 1601 (Clusius, 1601).

Staphisagria picta is a species endemic to Corsica, Sardinia, and Majorca, growing between 150 m and 600 m in open grasslands (Orellana et al., 2009). Aquilegia oxysepala is broadly found throughout Southeastern China and grows in open patches, along roadsides and forest margins at low altitudes (Li et al., 2014). Aquilegia coerulea (also described as Aquilegia caerulea) also has a large area of distribution, stretching across the Southern and central Rocky Mountains of western North America from 2100 m to 3700 m altitude (Miller, 1981). For genetic studies, mainly the commercially available, fast cycling cultivar ‘Origami’ is used (Sharma and Kramer, 2013). Thalictrum thalictroides (Ranunculaceae, also known as Anemonella thalictroides, commonly called rue anemone) is a spring ephemeral growing on streams and open woods in the Eastern USA (Lubbers and Christensen, 1986).

Taken together, the Ranunculales species for which genomic resources of various kinds are available occupy diverse habitats that range from dry Mediterranean islands (S. picta) over high altitudes (A. coerulea), to damp temperate forests (P. racemosum). Some species are abundant (N. damascena) or even invasive (E. californica), but several Ranunculales are rare and threatened in the wild (H. canadensis, C. chinensis, S. picta, and P. racemosum).

Floral diversity in Ranunculales

Floral structure and perianth in families of Ranunculales

Like floral phyllotaxis and symmetry, perianth organ identity, development, and function(s) are extremely diverse in Ranunculales, and range from absent to undifferentiated tepals, or more or less differentiated and petaloid sepals and modified and nectariferous petals (Fig. 2). For instance, flowers of the monotypic Eupteleaceae are perianthless (Ren et al., 2007), whereas those in both monotypic genera of Circaeasteraceae typically have tepals, which in Kingdonia co-occur with modified nectariferous petals interpreted as staminodial in origin (Ren et al., 2004; Tian et al., 2005). In contrast, flowers in Menispermaceae and Lardizabalaceae typically have persistent and more or less petaloid sepals and, when present, nectariferous petals (Endress, 1995). Papaveraceae have caduceus (easily detached) sepals and nectarless petals in the Papaveroideae, and more or less persistent petaloid sepals and spurred and nectar-collecting petals in Fumarioideae (Sauquet et al., 2015). In Berberidaceae and Ranunculaceae, flowers can also be perianthless or have a perianth differentiated into more or less caducous or petaloid and persistent sepals, and more or less modified and nectariferous petals traditionally referred to as ‘Nektarblätter’ (Hiepko, 1965; Terabayashi, 1985; Endress, 1995).

Fig. 2.

Simplified phylogeny of Ranunculales showing ancestral floral traits of the Ranunculales families.

As sister to all other eudicots, Ranunculales are thus pivotal to understanding the evolution of perianth and petaloidy in the largest clade of flowering plants. Previous comparative studies have shown that Ranunculales petals can 2be more similar to sepals in position, development, structure, and function(s) (e.g. petaloidy), and thus referred to as petaloid sepals of bracteopetalous origin, or more similar to stamens called Nektarblätter or nectariferous staminodia of andropetalous origin (Hiepko, 1965; Terabayashi, 1985; Endress, 1995). As in other eudicot lineages, the line between bracteo- and andropetals is usually defined by a set of developmental, structural, and functional traits which are thought to have evolved independently several times, including in the Ranunculales (Ronse De Craene and Brockington, 2013). However, a comparative study of gene expression patterns and floral organ identity challenged this view by suggesting that petals are deeply homologous and correlate with duplications and subfunctionalizations of B-class MADS box genes (Rasmussen et al., 2009, and see below).

Ancestral floral characters

Based on current ancestral character reconstructions of floral traits, the most recent common ancestor (MRCA) of Ranunculales had a differentiated perianth with at least three series (or whorls) of organs, assumed to be petaloid (Fig. 2; Carrive et al., 2020). The question of how different these whorls were remains unanswered, as does the question of which of the two outer whorls was lost in the families with only two whorls of perianth organs. Reconstructions of other perianth characters are consistent with the eudicot ancestor of Sauquet et al. (2017); that is, the androecium would have been composed of more than two whorls of stamens, and the gynoecium would have consisted of a few free carpels.

Unfortunately, such ancestral character reconstructions are sometimes hampered by the confusion surrounding the identity of perianth organs, and the definition of petaloidy. In Ranunculales, for instance, highly modified nectariferous petals described in previous literature as staminodia may be misinterpreted as belonging to the androecium, and as a result nectaries would be coded as present on the androecium, whereas these organs are more likely to be homologous to the other, less modified petals of andropetalous origin in other taxa. In addition, the distinction in the perianth between outermost sepals that may be caducous or become gradually petaloid, and the innermost ones persisting as bracteopetals that are regularly associated with modified and nectariferous andropetals, as in Berberidaceae or Ranunculaceae, has received little attention (e.g. Terabayashi, 1985). Such a re-evaluation of older literature would show that in Berberidaceae, for instance, nectaries always differentiate on more or less modified petaloid organs of androecial origin that were likely to have not only been present in their MRCA, but also associated with a distinct series of persistent petaloid organs of bracteal origin and an outermost one more or less caducous and/or petaloid. This interpretation is supported by the similarities observed in the recent reconstruction of the Berberidaceae and Ranunculales MRCA floral Bauplan. It suggests that the occurrence of petals of both bracteal and androecial origin in flowers of the MRCA of Berberidaceae, surrounded by an outermost series of more or less caducous and/or petaloid sepals, may be ancestral for the order as a whole.

Floral organ identity and petaloidy

The ABC homeotic genes define a strict model of organ identity in angiosperms (Coen and Meyerowitz, 1991). The identity of perianth organs is determined based on the expression of the A- and B-class genes, with the B-class giving the petaloid character of the organs. Class A genes are specific to angiosperms, characterized by the presence of sterile organs forming the perianth, surrounding the reproductive organs. Positive self-regulatory loops and antagonistic relationships among members of the ABC-class genes can modulate the timing of accumulation of the products of the different homeotic genes (Schwarz-Sommer et al., 1992; Halfter et al., 1994; Jack et al., 1994; Causier et al., 2010; Conde E Silva et al., 2023). The expression and functional evaluation of the ABC model of flower development genes has provided valuable insight into the evolution of flower patterning in Ranunculaceae Aquilegia (Kramer et al., 2003, 2004), Thalictrum (Di Stilio et al., 2005; Galimba et al., 2012, 2018; Larue et al., 2013; Galimba and Di Stilio, 2015; Soza et al., 2016; Martínez-Gómez et al., 2021), N. damascena (Wang et al., 2015), and Delphinium ajacis (Zhao et al., 2023), and in the Papaveraceae P. somniferum (Drea et al., 2007; Pabón-Mora et al., 2012) and E. californica (Yellina et al., 2010; Lange et al., 2013).

In core eudicots, an antagonistic relationship between classes A and C restricts their mutual expression (Causier et al., 2010). However, in Ranunculales, duplication events and subfunctionalization of members of the different gene classes suggest that the well-characterized Arabidopsis core eudicot model does not strictly apply. For instance, the role of A-class homologs (FUL-like genes) in sepal identity has only been demonstrated in E. californica and P. somniferum (Pabón-Mora et al., 2012). Knockdown of FUL function in E. californica or P. somniferum by VIGS reveals slight defects in petal shape and color, but petal identity is not lost. In Aquilegia coerulea, A-class genes have been recruited primarily in the proper patterning of leaves and have no function in perianth identity (Pabón-Mora et al., 2013). In N. damascena, FUL homologs have no role in floral organ identity, and an AGAMOUS-Like gene (AGL6) promotes sepal identity (Wang et al., 2016). All these results support the idea that FUL homologs do not have a strict A function in basal eudicots as they do in core eudicot models (Litt, 2007).

B-class genes, particularly AP3 homologs, have duplicated locally in Ranunculales, allowing for subfunctionalization and independent loss of petal identity genes (AP3-3) without affecting stamen identity factors (AP3-1 and AP3-2), and resulting in apetalous flowers independently (Zhang et al., 2013; Arango-Ocampo et al., 2016).

In Thalictrum, one such genus with apetalous flowers, certain B-class genes are expressed in the sepals only when they are petaloid, as in T. thalictroides (Galimba et al., 2018). In this species, E-class genes are also involved in the petaloidy of sepals, and have been suggested to keep the boundaries between either sepal and stamen zones or stamen and carpel zones by interacting with B- and C-class genes (Soza et al., 2016). Perianth organ identity in T. thalictroides would therefore be controlled by a sliding boundary model, with a shift towards sepals in the expression of B-class genes (Larue et al., 2013).

Members of the B-class genes are positive regulators of the expression of the C-class genes in E. californica, A. coerulea, and N. damascena, which in turn restrain the expression of the B-class genes (Yellina et al., 2010; Lange et al., 2013; Sharma and Kramer, 2013). The balance between the expression of the different paralogs of each gene class in the transition zones between floral organs is essential to maintain full organ identity and the proper number of each organ type. Flexibility in perianth organ identity may therefore result from the extension or restriction of B- or C-class genes, by modulating the interactions between ABC genes during species evolution. These mechanisms vary among species and individuals, depending on environmental conditions, particularly in flowers with spiral phyllotaxis.

Given the widespread occurrence of petaloidy in sepals or tepals in Ranunculales, and the potential that the ancestral flower had a perianth with whorled phyllotaxy (Sauquet et al., 2017; Sokoloff et al., 2018), perianth organ identity in the MRCA of Ranunculales may have been controlled by a sliding boundary model of floral organ identity. Either one, two, or even the three whorls (in the perianthless Eupteleaceae) would have been lost, and shifts in petaloidy could have occurred in the remaining whorls, depending on taxa, resulting in a strict ABC model of perianth identity evolving independently in some Papaveraceae (such as Eschscholzia, Chanderbali et al., 2010) and Berberidaceae, similar to the ABC model at play in core eudicots.

The petaloid appearance of sepals in different members of the Ranunculaceae (Fig. 2) has a different genetic basis. In Thalictrum, petaloid sepals express B- and E-class genes, and their targeted silencing or mutation leads to green leafy sepals (Soza et al., 2016; Galimba et al., 2018; Martínez-Gómez et al., 2021), whereas in Aquilegia the B gene AP3-1 controls the novel identity of the staminodium, and contributes to color but not papillate cell types in the sepals (Kramer et al., 2007; Sharma and Kramer, 2017). A ‘B’ gene paralog product of a Ranunculales-specific duplication, APETALA3-3, has become subfunctionalized to petal identity in Aquilegia (Sharma et al., 2011). This B gene is expressed in petals across other Ranunculales (Kramer et al., 2003) and has been secondarily lost in apetalous taxa such as Thalictrum (Di Stilio et al., 2005; Zhang et al., 2013).

Loss-of-function mutations in Thalictrum B-class genes, as found in natural and horticultural mutants (Martínez-Gómez et al., 2021) or by VIGS, result in female (carpellate) flowers, suggesting a recapitulation of one step in unisexual flower evolution (Larue et al., 2013). This hypothesis has played out in recent findings that B-class MADS box genes are involved in sex determination in other taxa, such as cycads and the rubber tree (Guo et al., 2022; Liu et al., 2022). Ovule identity is induced by the ‘D’ gene lineage (STK-like genes), based on studies in Petunia (Angenent et al., 1995). D-class and C-class genes originated from a gene duplication preceding the diversification of angiosperms (Kramer et al., 2004). Studies in T. thalictroides led to the finding that of the two AG paralogs, one performed the typical C function (in stamen and carpel identity, and floral determinacy) while the other subfunctionalized, taking on a D function role in ovule identity (Di Stilio et al., 2005; Zahn et al., 2006; Galimba et al., 2012; Galimba and Di Stilio, 2015). In Ranunculaceae no D-class genes were found, but a family-wide C lineage duplication was recorded (RanAG1/2, Kramer et al., 2004). In the Papaveraceae E. californica, a D lineage gene is found and an independent duplication occurred in the C lineage, resulting in two AG paralogs (Zahn et al., 2006; Yellina et al., 2010).

Ranunculales in the evolution of sexual and pollination systems

Most Ranunculales species are insect pollinated, some are hummingbird pollinated (Aquilegia), and the fly pollination syndrome (small, dull-colored, open flowers with nectariferous petals) is present in at least one genus in each family, whereas wind pollination syndrome (apetalous flowers with drooping stamens and filiform stigmas) is present in Eupteleaceae, Papaveraceae, and Ranunculaceae (Endress, 2010). Among American Aquilegia species, there is directionality in the evolution of pollination mode: substantially showier flowers with spurred petals and petaloid sepals are ancestrally pollinated by bees, with spurs getting longer with multiple transitions to hummingbird pollination and then to moth pollination (wind pollination is not known in this genus) (Whittall and Hodges, 2007).

Petals have been secondarily lost independently in Thalictrum and Enemion, the latter with flowers that resemble Thalictrum thalictroides and that are visited by small pollen-collecting bees. Thalictrum flowers are pollinated by small generalist insects, wind pollinated, or both (Kaplan and Mulcahy, 1971; Pellmyr, 1995). Very few systems lend themselves to the study of transitions between insect and wind pollination among closely related taxa at the genus level, and Thalictrum is one of them (Timerman and Barrett, 2019). From insect-pollinated, diploid, and hermaphrodite ancestors, Thalictrum species have transitioned at least eight times to wind pollination (Wang et al., 2019) in association with polyploidy and unisexual flowers (dioecy, cryptic dioecy, andromonoecy, and gynomonoecy, Soza et al., 2012, 2013). The search for pollination syndromes in Thalictrum by multivariate analysis of flower morphology identified four distinct flower morphotypes: ‘petaloid sepal’, ‘showy stamens’, and ‘small unisexual’, associated with insect pollination in the first two and wind pollination in the third. An ‘intermediate’ type that included a known mixed-pollinated (ambophilous) species was also identified, and the pattern held after considering phylogeny (Martínez-Gómez et al., 2023). These data broadly support the existence of detectable flower morphotypes from convergent evolution underlying the pollination mode in Thalictrum, presumably via different paths (petaloid sepals or showy stamens) from an ancestral mixed pollination state. Thus, pollination mode in Thalictrum is best described as a continuum between insect (the ancestral state) and wind pollination. An interesting research avenue would be to apply a comparable analysis of flower morphotypes to the direct outgroups and to other sister genera and families of Ranunculales. This approach would enable a deeper understanding of the evolutionary trajectory of flower morphologies in relation to pollinators and the sexual system at a broader phylogenetic scale.

Floral phyllotaxis and symmetry

Floral phyllotaxis (the arrangement of organs on the floral receptacle) may be whorled, spiral, and/or irregular (Endress, 2011). In spiral phyllotaxis, there is a delay (plastochron) between the initiation of two subsequent organs, whereas in whorled phyllotaxis, there is a marked plastochron only between whorls of organs belonging to different categories. The ancestral flower of Ranunculales was reconstructed as having a whorled phyllotaxis at anthesis, a condition that is observed today in most families of the order except Circaeasteraceae and some Ranunculaceae (Carrive et al., 2020). Although many members of this latter family have flowers with an apparently whorled perianth at anthesis, the initiation of perianth organs may follow a spiral pattern (Ren et al., 2011; Zhao et al., 2012). Reproductive organs are usually spirally arranged (Jabbour et al., 2009; Zhao et al., 2012), except in Aquilegia (Tucker and Hodges, 2005). However, together with their regular increase in numbers, especially in Ranunculaceae, their phyllotaxis may become more or less irregular with the insertion of incomplete parastichies on the onset of the androecium (Zhao et al., 2012).

With the exception of the perianthless monotypic family Eupteleaceae, almost all Ranunculales have flowers with at least one series of perianth organs (tepals, sepals, and/or petals), and the vast majority of these species have actinomorphic (i.e. polysymmetric) flowers. Reconstructing the ancestral state for the perianth is somewhat problematic in this order because Eupteleaceae are sister to the remaining six families of the order. However, it is very likely that the ancestral flower of all Ranunculales had a perianth that was actinomorphic (Damerval and Nadot, 2007; Carrive et al., 2020), as well as the ancestral flower of Berberidaceae, Ranunculaceae, Menispermaceae, Circaeasteraceae, and Lardizabalaceae. The flowers of Papaveraceae, and those of the genus Epimedium (Berberidaceae) were ancestrally dimerous and therefore dissymmetric, even if the corolla of Pteridophyllum (strongly supported as sister to Hypecoideae+Fumarioideae, Peng et al., 2023) and Papaveroideae, but also Epimedium, is visually actinomorphic (Sauquet et al., 2015; Carrive et al., 2020; Guo et al., 2022). Zygomorphy (i.e. monosymmetry) evolved once within Ranunculaceae, in the ancestral flower of the speciose tribe Delphinieae, once in Menispermaceae, in the ancestor of Antizoma, Cissampelos, and Cyclea (Ortiz et al., 2016), and probably twice within the subfamily Fumarioideae (Papaveraceae) (Hoot et al., 2015; Sauquet et al., 2015). Interestingly, zygomorphy evolved from dissymmetry in Papaveraceae, in which the dimerous ancestral state itself evolved from an actinomorphic state, a highly uncommon situation in angiosperms. Zygomorphy in Fumarioideae is created by the morphological differentiation of the two symmetry planes (e.g. in Lamprocapnos and Dicentra) followed by the formation of a single spur in the transverse plane during floral development (Damerval et al., 2013). Before anthesis, there is a 90° rotation of the pedicel (resupination) leading to a secondary vertical orientation of the symmetry plane (Endress, 1999; Hidalgo and Gleissberg, 2010). In Ranunculaceae and Menispermaceae, zygomorphy evolved from actinomorphy as in the vast majority of angiosperms (Reyes et al., 2016).

While zygomorphy has evolved independently in Papaveraceae and Ranunculaceae, their genetic bases could rely on CYCLOIDEA-Like (CYL) genes, as in several other angiosperm groups (for a review, see Hileman, 2014). The CYL lineage has probably undergone a duplication in the Ranunculales after the divergence of the Eupteleaceae (Damerval et al., 2022). In Fumarioideae, an asymmetric expression has been observed at late developmental stages in the zygomorphic flower of C. sempervirens (Damerval et al., 2013). CYL silencing by VIGS in the zygomorphic flower of Cysticapnos vesicaria reveals a role in sepal and petal identity and a possible involvement in zygomorphy (Zhao et al., 2018). In Ranunculaceae, additional duplications took place in both CYL lineages in the common ancestor of the zygomorphic tribe Delphinieae (Jabbour et al., 2014). Asymmetric expression of some paralogs was observed in the perianth (sepals and/or petals) of several species (Jabbour et al., 2014; Zhao et al., 2023). Silencing of CYL2 paralogs in Delphinium ajacis reveals a role for these genes in the sepal and primordia number, and in the dorsal identity for CYL2b or latero-ventral identity for CYL2a. It has been suggested that these identity roles were achieved through regulatory interactions with APETALA3-3 for CYL2b, and AGAMOUS-Like6-1a and DIVARICATA1 for CYL2a (Zhao et al., 2023).

3D morphogenesis of petals

Petal shape refers to the 3D structure of the organs from the inner whorl of the perianth (the andropetals). As in most angiosperm flowers, the petals of Ranunculales were ancestrally leaf shaped, with a flat blade and clawed at the base. This shape, combined with bright or colorful cues, is commonly referred to as ‘petaloid’ (Carrive et al., 2020). Among the six families of Ranunculales that have flowers bearing a perianth (Eupteleaceae are perianthless), such petaloid petals are observed in Menispermaceae, Circaeasteraceae, and in the subfamily Papaveroideae (Papaveraceae). In the other three families, the regular development of nectaries on the petals changes the shape, which becomes three dimensional due to the development of more or less pronounced nectar-storing invaginations ‘Nektarblätter’ (Hiepko, 1965); they are considered as having an elaborate form compared with flat petals and have evolved several times in Ranunculales. The ancestor of Lardizabalaceae already had strongly reduced and nectariferous petals (Zhang and Ren, 2011). In the subfamily Fumarioideae of Papaveraceae, petals are highly elaborate, and fused at the top.

In Ranunculaceae, some species have flat and regular petals with only a scale at the base protecting the nectary (e.g. Ranunculus), while other species have petals of various and elaborate shapes (tube shaped in Eranthis and Helleborus, spatula shaped in Actaea, including long stalks in Aconitum). The development of nectaries on these elaborate petals has been reviewed by Zhao et al. (2018) and will be discussed in the next section.

Ancestral state reconstruction of petal shape in Ranunculaceae showed that petals were ancestrally flat with a clawed base (Delpeuch et al., 2022). Elaborate, 3D shapes evolved independently from this ancestral petal by differential elongation of organ regions, depending on species. Recently, petal 3D morphogenesis was studied in the genus Staphisagria, which belongs to the only zygomorphic clade of Ranunculaceae (Zalko et al., 2021). Here, the complex petal shape seems to be the result of synorganization in the whole flower.

Flowers in the Berberidaceae family also have nectary-bearing petals ranging from flat to fan or funnel shaped. This morphological diversity results from developmental heterochrony and differential thickening (Su et al., 2021). Interestingly, these elaborate petals were probably ancestral in the family. In Epimedium (Berberidaceae), the co-occurrence of a spur with nectary development is responsible for the complex petal shape (Xie et al., 2022). Morphogenesis of simple and elaborate petals in angiosperms in general has been recently reviewed elsewhere (Fu et al., 2022).

Spurs

Spurs are tridimensional structures borne on the perianth, most often on petals, and occur frequently in Ranunculales. They are present in Ranunculaceae, where they have three independent origins [in Delphinieae, Myosurus and Aquilegia (Carrive et al., 2020), in Berberidaceae with a single origin in the common ancestor of Vancouveria and Epimedium (Sun et al., 2018; Guo et al., 2022), and in Papaveraceae with a single origin in the ancestor of all Fumarioideae]. In contrast to the rest of the angiosperms, where spurred flowers are most often zygomorphic (Jabbour et al., 2008; Citerne et al., 2010), in Ranunculales spurs are observed in flowers with various types of symmetry (Damerval and Nadot, 2007; Damerval and Becker, 2017; Carrive et al., 2020). In actinomorphic flowers, spurs are borne on each of the petals (Aquilegia—Ranunculaceae, Vancouveria—Berberidaceae). In dissymmetric flowers, the spurs are also borne on petals but their number varies depending on the degree of differentiation among the two whorls of petals. In Epimedium (Berberidaceae), spurs are borne on each of the four petals, whereas in dissymmetric flowers of Fumarioideae (Papaveraceae), only the outermost petals (two in number) are spurred (Endress, 1999). In zygomorphic flowers of Fumarioideae, a spur is borne on the outermost petal that is secondarily dorsal after resupination of the floral pedicel (Endress, 1999). The situation is more complex in Delphinieae (Ranunculaceae) where the number of spurs varies among genera and also among organ categories. All Delphinieae flowers have a spur (or hood in Aconitum, Gymnaconitum, and in some representatives of Delphinium subg. Consolida) that develops on the dorsalmost sepal, with a single exception (D. turcicum with peloric flowers devoid of corolla; Espinosa et al., 2017). Depending on the lineage, one (in the species included in Delphinium subg. Consolida) or two (in Staphisagria, the remaining species of Delphinium, Aconitum, and Gymnaconitum) spurred and nectariferous petals are nested within the dorsal sepal (Jabbour and Renner, 2012; Zalko et al., 2021).

The genetic origin of spurs was investigated in Aquilegia. Several transcription factors have been identified in the formation and elongation of the cup of the spur, some of which involved auxin signaling (Yant et al., 2015; Ballerini et al., 2020; Zhang et al., 2020). Whether the same or different genetic mechanisms have been recruited in the several independent evolutionary occurrences of spurs in Ranunculales is still unknown.

Nectary development

The production and secretion of nectar is a key innovation in flowering plants that attracts pollinators and facilitates sexual reproduction. In many taxa of angiosperms, various floral organs develop such as secreting tissues to offer sugary rewards to pollinators in exchange for their service in pollen transfer. These floral nectaries are believed to have evolved many times independently in angiosperms and may be located on various organs of the flower (Erbar, 2014). They may be located on the adaxial side of inner perianth organs or on members of the androecium (stamens or staminodes), as in some basal angiosperms, monocots (Liliales), and eudicots (e.g. Oxalidales and Caprifoliaceae). The monocot orders Asparagales and Zingiberales are characterized by septal nectaries. Receptacular nectaries often develop between the androecium and gynoecium, in association with the filament bases (Bernardello, 2007); the nectaries may be located on the receptacle (as in many rosids) or on the gynoecium (as in many asterids). The floral nectaries in Ranunculales exhibit a great diversity (Fig. 3). All families of Ranunculales, except for Eupteleaceae, have species that develop floral nectaries. A recent study focusing on the ancestral traits of Ranunculales flowers indicated that these nectaries are likely to have evolved many times independently (Carrive et al., 2020), reflecting the various floral organs that bear the nectaries in different families.

Fig. 3.

SEM images of mature nectaries from four Ranunculales species: (A) Aquilegia coerulea (Ranunculaceae), (B) Epimedium grandiflorum (Berberidaceae), (C) Corydalis aurea (Papaveraceae), (D) Lamprocapnos spectabilis (Papaveraceae). Enlarged views of the nectary cells for Aquilegia and Epimedium are shown in insets. N, nectary; S, spur; St, stamen. Asterisks indicate swelling epidermal cells, and red arrowheads indicate active secreting cells. Scale bars: 100 μm (insets 10 μm).

Most Ranunculaceae species develop floral nectaries on their petals, associated with nectar-storing invaginations of various shapes, such as spurs (e.g. Aquilegia and Aconitum), funnels (e.g. Helleborus and Eranthis), urns (e.g. Nigella), or cups (e.g. Coptis). However, nectaries can also be found on other floral organs in this family, including stamens and carpels. In Clematis alpina, which re-evolved petals after the petal loss in the common ancestor of Clematis, the nectary is not present on petals but on carpels instead (Erbar, 2014). In several apetalous genera, including Caltha and Anemone, the nectary probably re-evolved and is also present on carpels (Peterson et al., 1979; Erbar and Leins, 2013). It is worth noting that in the wind-pollinated genera Thalictrum, floral nectaries and petals were lost, possibly due to the relaxation of selection pressure to maintain these costly structures for pollinator attraction.

The close association between nectaries and elaborate petaloid organs in Ranunculaceae has been hypothesized to facilitate the diversification of pollinator interaction for Ranunculaceae species. In Aquilegia, a recently established model for nectary development, nectaries develop inside the tips of the spurs (Fig. 3A), and secrete sucrose- and hexose-abundant nectar to be stored in spurs. The amount of secreted nectar and the length/curvature of the spurs are highly diverse in different Aquilegia species (Puzey et al., 2012; Edwards et al., 2021). Together, these traits limit nectar access to a specific type of pollinator and can function as reproductive barriers among Aquilegia species.

In the closely related family Berberidaceae, nectaries are also commonly found on the perianth (Su et al., 2021). Many of these nectariferous organs are historically considered staminodes due to their locations and developmental origins. However, gene expression profiling and phylogenetic analysis showed that these nectariferous organs from several genera, including Berberis and Epimedium, express the petal identify B-class gene AP3-3 (Kramer et al., 2003; Rasmussen et al., 2009). In Berberis, the inner two perianth whorls bear elliptical, markedly protruding nectaries that embrace the fertile stamens (Erbar, 2014). In Epimedium, each petal develops a 3D spur, similar to Aquilegia, and bears a nectary at each tip (Xie et al., 2022) (Fig. 3B).

In the Papaveroideae subfamily of the Papaveraceae, floral nectaries are absent, while they are usually present in the members of the Fumarioideae clade (Wang et al., 2023), for example Dicentra, Corydalis, Capnoides, and Fumaria, and these nectaries are likely to be homologous (Carrive et al., 2020). Interestingly, the perianths from these genera also develop spurs that hold nectar, but the sites of nectar production and secretion are shifted to the bases of the stamens. In the bisymmetric flower of Dicentra (Lamprocapnos), six stamens are organized as two triplets, and the filaments of each triplet are basally fused. At the abaxial base of the central filament of each stamen triplet, a ball-shaped nectary develops and is completely enclosed by the petal spur (Zhang and Zhao, 2018) (Fig. 3C). In the zygomorphic flower of Corydalis, one out of four petals forms a nectar spur. A ‘stalklet’ develops from the base of stamen bundles, is inserted into the spur, and bears a nectary at the free end (Erbar, 2014) (Fig. 3D).

In recent years, many comparative studies have surveyed the organization and gene expression profiles of floral nectaries in Ranunculales and reported distinct cellular and molecular mechanisms of nectary development and nectar secretion (Vesprini et al., 1999; Damerval et al., 2013; Erbar and Leins, 2013; Erbar, 2014; Antoń and Kamińska, 2015; Zhang and Zhao, 2018; Min et al., 2019; Xie et al., 2022). While most nectaries in core eudicots employ nectary stomata or secretory trichomes to release nectar, such structures are typically absent in Ranunculales nectaries, with a potentially notable exception in Fumarioideae (i.e. A. asiatiaca) (Fig. 3A–D; Wang et al., 2023). Instead, nectar secretion by ruptured epidermis or cuticle micro-channels was proposed. At the molecular level, the YABBY family transcription factor gene CRABS CLAW (CRC) was required for nectary development in several asterid and rosid lineages (Bowman and Smyth, 1999; Lee et al., 2005). However, CRC expression was not detected in Ranunculaceae nectaries, and nectary development in Aquilegia is instead directed by the STYLISH (STY) family of transcription factor genes (Min et al., 2019). Expression of STY genes has also been reported in the nectariferous petals of Delphinium exaltatum and Epimedium (Min et al., 2019). In contrast, expression of CRC orthologs was observed at the nectariferous base of the stamen filaments in the Papaveraceae C. sempervirens and Lamprocapnos spectabilis (Damerval et al., 2013), potentially reflecting the independent evolution of nectaries in Ranunculaceae and Papaveaceae. Future functional studies are required to fully elucidate the cellular and developmental mechanisms of nectar production and nectary formation in Ranunculales.

Fruit morphology and dehiscence types

Gynoecium and fruit type vary greatly in Ranunculales (Fig. 4). The ancestral condition was identified for the entire order after careful character optimization and found to be a multicarpellate, apocarpous gynoecium. However, different morphologies have become fixed in different families. Whereas Papaveraceae sensu lato (including former Fumariaceae) have a syncarpous gynoecium, the apocarpous condition is common in Eupteleaceae and is a synapomorphy for Lardizabalacaeae [Menispermaceae [Berberidaceae+Ranunculaceae]]. Members of the Ranunculaceae have predominantly an apocarpous gynoecium, but the carpels are frequently described as being connate to different degrees in some genera (such as Nigella and Glaucidium for instance). Berberidaceae are unique in that all members regularly possess a unicarpellate gynoecium, which, as in some Ranunculaceae, is probably derived from an ancestral multicarpellate and apocarpous condition, that is also distinctly entirely ascidiate (versus more or less plicate in all other Ranunculales, e.g. Endress, 1995) (Fig. 4)

Fig. 4.

Ancestral state reconstruction of gynoecium (left) and fruit (right) characters based on a phylogeny using rbcL as the marker gene. Trait descriptions are from Cheng-Yih and Kubitzki (1993), Endress (1993), Kadereit (1993), Lidén (1993a, b), Loconte (1993), and Tamura (1993).

In terms of fruit type, the ancestral condition is the presence of dry dehiscent fruits. Within that category, indehiscent samaras (a winged achene with the wing developing from the ovary wall) are predominant in Eupteleaceae. In contrast, longitudinally dehiscent fruits, whether derived from a syncarpous gynoecium (capsules) or from an apocarpous gynoecium (follicles), are plesiomorphic for the rest of the families in the order. Capsules are typical in Papaveraceae and Ranunculaceae (Fig. 4). Fleshy fruits have been independently acquired in Lardizabalaceae (in the genus Sinofranchetia) and many Berberidaceae, as well as in Hydrastis (Ranunculaceae). Drupaceous fruits, also indehiscent, are characteristic of the Menispermaceae (Fig. 4). Achenes, dry indehiscent fruits, were independently acquired in Circaeasteraceae and some Ranunculaceae.

Regarding the genetic bases for fruit development, there are a number of genes whose function seems to be maintained in both Papaveraceae and Arabidopsis. They include FRUITFULL (FUL) genes largely expressed in the fruit wall in E. californica and P. somniferum. When FUL genes are down-regulated, fruit defects include premature rupture of the fruit wall and numerous cell proliferation defects, especially in the endocarp (Pabón-Mora et al., 2012). APETALA2 (AP2) genes are, on the other hand, very different. The two copies show overlapping expression only in the commissural tissue, and one of the homologs is also expressed in the fruit wall. Very important is the fact that both copies are absent from the dehiscence zone (DZ). These expression patterns suggest a role for AP2 genes in fruit wall development, most probably acting as repressors of DZ-specific genes (Zumajo-Cardona et al., 2021). Further, the E. californica homolog of CRC (EcCRC) is required for adaxial gynoecium tissue development, and down-regulation leads to a complete abolishment of the DZ (Orashakova et al., 2009).

Genes probably controlling the formation of the DZ in Papaveraceae are SPATULA/ALCATRAZ homologs specifically restricted to those layers (Zumajo-Cardona et al., 2017), acting together with REPLUMLESS genes, which were observed in the DZ not only in Bocconia, but also in Papaver, suggesting that this is a common putative role for many Papaveraceae (Zumajo-Cardona et al., 2018). In A. thaliana, INDEHISCENT and SHATTERPROOF1 and 2 are essential for the formation of the DZ. However, as their orthologs do not exist in Ranunculales (Zahn et al., 2006; Pabón-Mora et al., 2014), the dry dehiscent fruits predominant in the Ranunculales require a gene regulatory network very different from that of A. thaliana.

Conclusions

This review has highlighted the Ranunculales as an emerging model lineage for comparative analysis of morphological and metabolic traits in angiosperms, pointing out recent developments in the field of genomics and genetic manipulation of several members from diverse families. The amazing morphological diversity of Ranunculales raises the question of the underlying genetic bases (particularly concerning convergent traits), still largely unexplored, but also the question of floral integration (whether traits evolve independently from each other or in a correlated manner). Addressing these questions in Ranunculales, an order with a key phylogenetic position, may contribute to a better understanding of the drivers of morphological evolution in angiosperms as a whole. Combining a solid phylogeny and fossils for its calibration, molecular tools and genetic resources, together with high morphological diversity, convergent evolution of characters, frequent switching between reproductive systems, and developmental trajectories and functions of perianth organs, the Ranunculales order offers new avenues for investigations into plant evolution and adaptation.

Author contributions

AB: conceptualization; CD, FJ, VdS, and AB: review and editing. All authors participated in writing the original draft,

Conflict of interest

The authors declare no conflict of interest.

Funding

Work in AB’s group on E. californica and on genomic resources of Ranunculales was continuously funded by the DFG (German Research Foundation, grants BE2547/3-1; 6-1; 6-2; 7-2; 14-1; 24-1, the RanOmics project is funded by 27-1). VD was funded by the National Science Foundation (USA), Division of Environmental Biology (Opportunities for Promoting Understanding through Synthesis—Mid-Career Synthesis) grant no. 1911539. YG is supported by National Science Foundation (USA) Postdoctoral Research Fellowships in Biology Program under grant no. 2305493. YG and EMK are supported by National Science Foundation (USA) EDGE Award IOS no. 2128195.