Using integrative taxonomy and multispecies coalescent models for phylogeny reconstruction and species delimitation within the “Nastanthus–Gamocarpha” clade (Calyceraceae)

Highlights

• Diverse lines of evidence provide rationale for species delimitation.

• Coalescent-based methods could be applied in a wide range of species evolving systems.

• Phenology facilitates the recognition of biologically meaningful species.

Abstract

The Calyceraceae (47 spp.) is a small family of plants that is sister to the Asteraceae (∼ 25,000 spp.), one of the largest families of angiosperms. Most members of Calyceraceae are endemic to the Andes and Patagonia, representing an excellent model within which to study diversification patterns in these regions. The single phylogenetic study of Calyceraceae conducted to date revealed that the boundaries of most genera and several species of this family require further analyses, especially the “Nastanthus–Gamocarpha” clade. In this study, we reconstructed the phylogeny of the “Nastanthus–Gamocarpha” clade using multispecies coalescent models under BPP and StarBeast2 programs, sampling 63 individuals from 13 of the 14 species recognized to date. We then used this phylogenetic framework to delimit species using BFD and the A11 method implemented in BPP. Species limits suggested through a coalescent approach were then re-evaluated in the light of morphology, geography, and phenology. Coalescent-based methods indicated that most putative lineages could be recognized as distinct species. Morphological, geographical, ecological, and phenological data further supported species delimitation. Necessary taxonomic changes are proposed. Namely, the paraphyletic Nastanthus is synonymized under Gamocarpha, while five species of Boopis are transferred into Gamocarpha. We used an integrative taxonomic approach to recognize 13 species and one subspecies within the newly circumscribed genus Gamocarpha.

Introduction

Robust phylogenetic frameworks and accurate species delimitation are essential for ecological and evolutionary studies, biodiversity assessments, conservation planning, and biological control (Bortolus, 2008). Species delimitation is a critical and dynamic task in systematics, in which different species concepts, data types, and methodologies have prevailed through time. Over the past decade, emphasis has been placed on statistical rigor for species delimitation, leading to the development of new methodological approaches based on coalescent theory (Carstens et al., 2013).

Species delimitation is especially problematic within recently diverged lineages that have had little time to accumulate morphological differences. Apart from instances of rapid radiation, low morphological divergence may also result from phenotypic convergence (Niemiller et al., 2012). In such cases, delimiting species solely based on morphology may lead to the recognition of broadly defined taxa, underestimating species numbers.

The unified species concept (de Queiroz, 2007), which distinguishes between what a species is (i.e., a separately evolving metapopulation lineage), from the criteria used to recognize and delimit such metapopulation lineages (e.g., morphology, breeding barriers, ecology, and so forth), has prompted researchers to develop new methods based on coalescent theory to accurately define and circumscribe meta-populations. These methods use multi-locus data and multi-individual sampling per taxon to identify speciation events, understand the processes leading to speciation, and quantify the probability of lineage independency (Knowles and Carstens, 2007, O’Meara, 2010, Carstens and Dewey, 2010, Yang and Rannala, 2010, Zhang et al., 2011, Carstens et al., 2013, Grummer et al., 2014). Coalescence-based methods emphasize incomplete lineage-sorting as a source of inconsistency between gene trees and the species tree (Rannala and Yang, 2003, Liu et al., 2009, Heled and Drummond, 2010), allowing the accurate estimation of the species tree. Data derived from coalescent-based methods are particularly useful when interpreted in the light of additional sources of data such as geography, morphology, ecology, and behavior (Weins and Penkrot, 2002, Schlick-Steiner et al., 2010, Gratton et al., 2016, Fernández et al., 2017).

Calyceraceae includes 47 species traditionally arranged in six genera: Acicarpha Juss. (4 spp.), Boopis Juss. (16 spp.), Calycera Cav. (9 spp.), Gamocarpha DC. (5 spp.), Moschopis Phil. (7 spp.), and Nastanthus Miers (6 spp.) (Zavala-Gallo et al., 2010, Zavala-Gallo, 2013, Denham et al., 2014). Most species grow along the southernmost Andes, as far north as Bolivia and extending to the Patagonian steppe in Argentina and Chile. A few species grow in northern and central Argentina, southern Brazil, and Uruguay, reaching the Atlantic sea-shore. One species is endemic to the Malvinas Islands (Zavala-Gallo, 2013, Denham et al., 2016). This family belongs to the MGCA clade within Asterales, which includes Menyanthaceae, Goodeniaceae, Calyceraceae, and Asteraceae (Lundberg and Bremer, 2003, APG, 2009). Calyceraceae is monophyletic and sister to Asteraceae, one of the five largest families of angiosperms. Although Calyceraceae is diagnosed by a series of morphological features, its genera lack diagnostic characters, and taxonomic revisionary efforts are still needed. The most comprehensive phylogeny of the family to date (41 of 47 species sampled) indicated that the traditionally recognized genera are not monophyletic, except for Acicarpha (Denham et al., 2016). This phylogenetic study recovered seven well-supported clades and two main larger clades that could be diagnosed by different basic chromosome numbers.

One of the seven strongly supported clades within Calyceraceae is the “Nastanthus–Gamocarpha” clade, which includes 14 species (Denham et al., 2016). This clade includes all six species of Nastanthus, three of the five species of Gamocarpha currently recognized (G. dentata Phil. and G. angustifolia Phil. fall outside this clade), and five of the 16 species of Boopis recognized (Table 1). These phylogenetic results supported the synonymy among Boopis, Gamocarpha, and Nastanthus proposed decades earlier based on morphology (Reitz, 1988, Hellwig, 2007). The species of the “Nastanthus–Gamocarpha” clade grow along the Southern portions of the Central Andes, Southern Andes, and Patagonia in Argentina and Chile, from 25° to 54° S (Denham et al., 2016). This is the most recently diverged clade in the family, with a crown-group that emerged at around 5 Myr and diversified when peaks of intensity in the southern Andes uplift and glacial and interglacial periods were prevalent (Denham et al., 2016). The most distinctive feature of the “Nastanthus–Gamocarpha” clade is the central inflorescence (cephalioid) that not only produces cymose groups surrounded by the involucre (inner cymose groups) as expected, but also often produces more or less detached cymose groups surrounding the involucre (outer cymose groups) that is unique in this family (Pozner et al., in prep.). Even though there is strong molecular support for this clade, relationships within it remain largely unresolved (Denham et al., 2016). Because this clade has diverged very recently, a deeper understanding of relationships among its members can be achieved through coalescent approaches.

This paper aims to explore relationships among members of the “Nastanthus–Gamocarpha” clade through multispecies coalescent models, using a multi-locus genetic dataset and extensive multi-individual sampling across the distributional range of each putative species. Specifically, we aim to: (i) test current species boundaries using an integrative taxonomic approach; (ii) reconstruct relationships among taxa; and (iii) reevaluate generic boundaries within Calyceraceae.