Tempo and mode in coevolution of Agave sensu lato (Agavoideae, Asparagaceae) and its bat pollinators, Glossophaginae (Phyllostomidae)

Highlights

• Agave sensu lato appeared around 8 Mya.

• The best supported diversification model for Agavoidae shows two shifts around 8 Mya.

• Agave sensu lato speciation rate has been increasing towards the present.

• Nectarivorous bats from the Glossophaginae subfamily originated around 7 Mya.

• Agave and their pollinating bats share a diffuse coevolutionary history.

Abstract

The genus Agave sensu lato contains ca. 211 described species, many of which are considered keystone species because of their ecological dominance and the quantity of resources they provide with their massive, nectar-rich inflorescences. The large diversity of Agave species has been hypothesized as being related to their reproductive strategy (predominantly monocarpic) and diverse pollinators (e.g., bats, hummingbirds, hawkmoths). In particular, Agave species provide resources that a few genera of nectar feeding bats from the subfamily Glosophaginae are dependent upon. To explore a possible coevolutionary relationship between Agave and the bat species that pollinate them, we calibrated molecular phylogenies of both groups and looked for a correlation in their dates of divergence. One coding and two non-coding regions of the chloroplast genome were sequenced from 49 species of the Agavoideae (Asparagaceae), and the mitochondrial gene Cyt-b and nuclear coding gene RAG2 were either sequenced or obtained from gene bank for 120 Phyllostomid bats. Results from the analyses indicate that Agave sensu lato is a young genus (estimated crown age 2.7–8.5/stem age 4.6–12.3 Ma), with an increasing diversification rate, and the highest speciation rate among Agavoideae’s clades. The origin of the Glossophaginae bats (stem age 20.3–23.5 Ma) occurred prior to the stem age of Agave sensu lato, while the origin of the current pollinators of Agave species, members of the genera Glossophaga, Leptonycteris, Anoura, Choeronyscus, Musonycteris and Choeronycteris, was estimated to be around 6.3–16.2 Ma, overlapping with the stem age of Agave sensu lato, supporting the hypothesis of diffuse coevolution.

Introduction

Two clades of succulent plants define the landscape of most of Mexico and the Southwestern United States: Agavoideae and Cactaceae. Within these clades, two prominent lineages, Agave (ca. 211 species) and columnar cacti (Pachycereeae tribe, Cactaceae, ca. 58 species) are both mainly pollinated by bats, and are considered keystone species because of their local abundance and the high quantity of resources they provide (Good-Avila et al., 2006, Rocha et al., 2006, Hernández-Hernández et al., 2014). Both lineages exhibit asymmetric relationships with their pollinators (Slauson, 2000): while the nectarivorous bats (subfamily Glossophaginae) almost completely rely on these two plant resources for survival in arid and semi-arid regions (Arita and Humphrey, 1988, Rojas-Martinez et al., 1999, Trejo-Salazar et al., 2016), the degree of reliance of the agaves and columnar cacti on bat pollination depends on the abundance of the bat species in the region, but is rarely exclusive of them (Munguía-Rosas et al., 2009). In other regions of Mexico, in particular the dry tropical forests, Glossophaginae bats feed mainly on other tropical tree species, in particular, members of the Bombacoideae (Malvaceae) such as Ceiba and Pseudobombax (Álvarez and González Quintero, 1970, Arita and Ceballos, 1997). Consequently, it has been suggested that the relationship between columnar cacti, Agave and Glossophaginae bats may represent a complex case of diffuse coevolution, a shared evolutionary history between two lineages without coincident speciation and thus few species of one lineage influencing several species of the other lineage in a reciprocal manner (Valiente-Banuet et al., 1996, Valiente-Banuet, 2002, Good-Avila et al., 2006, Rocha et al., 2006, Trejo-Salazar et al., 2015, Trejo-Salazar et al., 2016). This hypothesis was proposed based on the observations of an overlapping distribution and a chiropterophilous pollination syndrome present in Agave and most columnar cacti, but it is also supported by field observations and experiments (Arizaga, 2000, Eguiarte et al., 2000, Molina-Freaner and Eguiarte, 2003, Rocha et al., 2006, Trejo-Salazar et al., 2015, Trejo-Salazar et al., 2016). To test if the relationship among these groups is an example of diffuse coevolution, a careful calibration of the time of origin and diversification is needed.

Within the Chiroptera, only two families contain nectarivorous bats, one in the Old World (Pteropodidae) and the other from the New World (Phyllostomidae). Among new world bats only Glossophaginae and Lonchophyllinae exhibit nectarivory, representing an example of dietary convergence based on morphological, behavioral, ecological and physiological strong resemblances. These independent adaptations to a nectar-feeding lifestyle in bats originated during the Miocene at (17–25) and (8–19) Ma respectively (Datzmann et al., 2010, Rojas et al., 2011, Rojas et al., 2012). In particular, the bat species that pollinate agave and cacti belong to the Subfamily Glossophaginae (sensu Baker et al., 2016), which contains 13 genera, most of them (eight) with representatives in Mexico (Arita and Santos del Prado, 1999). Within Glossophaginae, members of Leptonycteris, Glossophaga and Choeronycteris are considered to be the most specialized, feeding exclusively on nectar and pollen (Arizmendi et al., 2002). Moreover, the three species within the Leptonycteris genus feed primarily from agave and columnar cacti nectar and pollen (Arita and Humphrey, 1988, Rojas-Martinez et al., 1999, Cole and Wilson, 2006, Trejo-Salazar et al., 2016). Bats are able to travel long distances if they are provided with an abundant and easily accessible source of highly energetic, nocturnally released reward (see review in Fleming et al., 2009). The preference of Leptonycteris bats for agave and columnar cacti flowers arises from their needs for both resource accessibility (Muchhala, 2003, Fleming et al., 2009) and a high nectar volume. In addition, Agave nectar contains dimethylsulfides (de la Fuente et al., 2007), aromatic compounds to which nectar-feeding bats are strongly attracted (von Helversen et al., 2000). These compounds are found in a wide range of unrelated bat-pollinated plants, including agaves and columnar cacti, probably as a result of convergent evolution (Knudsen and Tollsten, 1995).

Agave is highly adapted to arid and semi-arid conditions exhibiting Crassulacean acid metabolism (CAM), and producing a characteristic massive inflorescence (Eguiarte et al., 2000, Eguiarte et al., 2013, Rocha et al., 2006; see Heyduk et al., 2016 for a detailed analysis on the evolution of CAM within the Agavoidae), which is followed by the death of the vegetative rosette, in contrast to iteroparous columnar cacti. Agave has traditionally been divided into two subgenera based on the branching of the inflorescence: Littaea which has an unbranched spike or racemose inflorescence (including eight morphological groups Berger 1921 (in Gentry, 1982)) and Agave with a branched or paniculate inflorescence (including 12 morphological groups; Berger 1921 (in Gentry, 1982)). A molecular estimate of the time of origin of Agave sensu lato (consisting of the monophyletic lineages of Agave sensu stricto, Manfreda, Polianthes and Prochnyanthes) indicated that the genus is young, 7.8–10.6 Ma (Good-Avila et al., 2006), and has undergone rapid diversification, potentially driven by an adaptive radiation.

Our main aim was to answer two central questions: Does the current ecological interaction between Glossophaginae bats and Agave sensu lato species represent a diffuse coevolutionary process? If so, could bat pollination have driven, at least in part, the rapid diversification of Agave?

In order to answer these questions, we extended the study mentioned previously in two ways: first, by including more representatives of herbaceous species, and second, by determining the most likely diversification model for Agave sensu lato and the rest of the sampled members of the family. In addition, to assess the potential of a shared evolutionary history between these two lineages, we addressed the monophyly of the nectarivorous groups within the Glossophaginae bats and estimated their date of divergence.