Landscape genetics across the Andes mountains: Environmental variation drives genetic divergence in the leaf-cutting ant Atta cephalotes
Vanessa Muñoz-Valencia1, James Montoya-Lerma1, Perttu Seppä2 & Fernando Diaz3
1 Group of Agroecosystem Ecology and Natural Habitats, Department of Biology, Faculty of Natural Science, Universidad del Valle, Cali, Colombia.
2 Organismal and Evolutionary Biology Research Programme, Faculty of Biological and Environmental Sciences, University of Helsinki, Finland.
3 Biology Department, Colgate University, Hamilton, NY, USA.
Corresponding authors:
Vanessa Muñoz-Valencia, vanem28@gmail.com; James Montoya-Lerma, james.montoya@correounivalle.edu.co; Fernando Diaz, ferdiazfer@gmail.com

ABSTRACT

Distinguishing among the mechanisms underlying the spatial distribution of genetic variation resulting from the environmental or physical barriers from those arising due to simple geographic distance is challenging in complex landscapes. The Andean uplift represents one of the most heterogeneous habitats where these questions remain unexplored since multiple mechanisms may interact, confounding their relative roles. We explore this broad question in the leaf-cutting ant Atta cephalotes , a species that is distributed across the Andes mountains, using nuclear microsatellite markers and mtCOI gene sequences. We investigate spatial genetic divergence across the western range of the northern Andes in Colombia by testing the relative role of alternative scenarios of population divergence, including isolation by geographic distance (IBD), climatic conditions (IBE), and the physical barriers presented by the Andes mountains (IBB). Our results reveal substantial genetic differentiation among A. cephalotes populations for both types of markers, but only nuclear divergence followed a hierarchical pattern with multiple models of genetic divergence imposed by the western range. Model selection showed that the IBD, IBE (temperature and precipitation), and IBB (Andes mountains) models, often proposed as individual drivers of genetic divergence, interact and explain up to 33% of the genetic divergence in A. cephalotes . The IBE model remained significant after accounting for IBD, suggesting that environmental factors play a more prominent role than with IBB. These factors, in combination with the idiosyncratic dispersal patterns of ants, appear to determine the hierarchical patterns of gene flow. This study enriches our understanding of the forces shaping population divergence in complex habitat landscapes.

KEYWORDS

Andean uplift, western mountain range, spatial genetic structure, isolation by distance, isolation by barrier, model selection, isolation by environment.

INTRODUCTION

Population divergence can be determined by the spatial distribution of individuals, where geographic proximity modulates genetic similarity, leading to a pattern of isolation by distance – IBD (Lee & Mitchell-Olds, 2011; Shafer & Wolf, 2013; Slatkin, 1993; Wright, 1943). However, populations in complex landscapes are often exposed to environmental variation and physical barriers that can also contribute to genetic divergence (Manel & Holderegger, 2013; Manel, Schwartz, Luikart, & Taberlet, 2003; Shafer & Wolf, 2013). Populations could adapt to their local environment, maximizing the fitness of individuals under local conditions while decreasing the fitness of immigrants from alternative environments (Carro, Quintela, Ruiz, & Barreiro, 2019; Sobel, 2014; Wang & Bradburd, 2014; Wang, Glor, & Losos, 2013). This adaptive reduction in gene flow can produce a pattern of isolation by environment – IBE (Sexton, Hangartner, & Hoffmann, 2014; Shafer & Wolf, 2013; Wang & Bradburd, 2014). Alternatively, gene flow could be restricted by allopatric scenarios of genetic differentiation mediated by physical barriers in the landscape, generating patterns of isolation by barrier – IBB (De Queiroz, Torrente-Vilara, Quilodran, da Costa Doria, & Montoya-Burgos, 2017; Haffer, 2008; Rull, 2011; Turchetto-Zolet, Pinheiro, Salgueiro, & Palma-Silva, 2013). With increasing landscape complexity, gene flow is likely to be influenced by a combination of geographical and ecological factors in which these isolating mechanisms are not mutually exclusive (Crispo, Bentzen, Reznick, Kinnison, & Hendry, 2006; Edwards, Keogh, & Knowles, 2012; Noguerales, Cordero, & Ortego, 2016; Wang et al., 2013).
The Andean mountain ranges not only represent one of the most unexplored environments, but also offer a great complexity of landscapes, promoting the diversification of a wide range of taxa (Salgado-Roa et al., 2018). Across these mountains, restricted gene flow has been reported in multiple organisms, including birds (Cadena, Pedraza, & Brumfield, 2016), plants (Lagomarsino, Condamine, Antonelli, Mulch, & Davis, 2016; Luebert & Weigend, 2014; Pérez-Escobar et al., 2017), mammals (Antonelli et al., 2009; Hoorn et al., 2010), insects (Antonelli et al., 2009; De-Silva et al., 2017; Hoorn et al., 2010), and other arthropods (Salgado-Roa et al., 2018). However, organisms from contrasting populations in these habitats are often exposed to a combination of distance, environmental, and physical barriers to dispersal, challenging the investigation of the relative roles of different isolating mechanisms (James, Coltman, Murray, Hamelin, & Sperling, 2011; Meirmans, 2015; Noguerales et al., 2016). It remains unclear how the Andean uplift has modulated patterns of gene flow and the evolution of several of the most ecologically important groups in the Neotropics, including social insects. For example, although the entire evolution of Neotropical ants occurs across the Andes (Mueller et al., 2017), the interplay between their population structure and environmental variation relative to the effect of these mountains on isolated populations remains largely unknown.
The leaf-cutting ant A. cephalotes is a major urban and agricultural pest in the Neotropics, colonizing a wide spectrum of environments (Della Lucia, Gandra, & Guedes, 2014; Fernández, Castro-Huertas, & Serna, 2015; Hölldobler & Wilson, 2011). In Colombia, its distribution overlaps with the maximum complexity of the Andean uplift, ranging from 0 to 2100 m.a.s.l. (Fernández et al., 2015), with a vertical thermal gradient of 0.6 °C/100 m (Hermelin, 2015). The northern section of these mountains in Colombia splits into three main branches: the western, central, and eastern ranges. Populations ofA. cephalotes are separated by these mountains while simultaneously being exposed to complex combinations of topographical and environmental variation (Kattan, Franco, Rojas, & Morales, 2004; Pérez-Escobar et al., 2017; Salgado-Roa et al., 2018). Such conditions provide a tremendous climatic spectrum for local adaptation, with an interplay between population dynamics and species-specific dispersal patterns (Hakala, Seppä, & Helanterä, 2019). Evolution under such environmental heterogeneity could act to shape patterns of gene flow (De Queiroz et al., 2017; Lee & Mitchell-Olds, 2011; Noguerales et al., 2016; Wang et al., 2013), which often produces more complex scenarios than genetic divergence due to IBD alone (Slatkin, 1993; Wright, 1943). For example, we recently found that the eastern range of the Andes in Colombia plays a major role as a geographic barrier to historical gene flow, restricting the dispersion of A. cephalotes from north to south (Muñoz-Valencia, Vélez-Matínez, Montoya-Lerma, & Díaz, 2021). Although this initial study demonstrates the significant influence of the Andes on population divergence in the leaf-cutting ant at the phylogeographic scale, the role of local adaptation occurring at more regional scales across the Andes remains untested.
This study focuses on a finer and more complex geographic distribution scale of A. cephalotes : that of the western range of the Andean uplift in Colombia. We use a landscape genetic approach to investigate the role of geographic features and environmental variation in the definition of patterns of spatial genetic structure in A. cephalotes . Using nuclear (microsatellites) and mitochondrial (mtCOI ) markers, we test the relative roles played by geographic distance, climate variation, and a major dispersal barrier (the western range) in modulating patterns of gene flow. As a monogynous (single-queen) species, A. cephalotes presumably undertakes long-distance nuptial flights that can potentially overcome isolating barriers (Cherrett, 1968; Helms, 2018; Moser, 1967). Our results demonstrate that gene flow is limited by a complex interaction of the three isolating mechanisms (IBD, IBE, and IBB) rather than IBD alone, while IBE appears to play a stronger role than IBB. Investigating the spatial genetic structure of a species in an exceptionally heterogeneous environment helps to elucidate the evolution and diversity of this ecologically dominant group of ants in the Neotropics.

MATERIALS AND METHODS

Sampling

Ant sampling was conducted in the Colombian Pacific and Andean regions, which are separated by the western mountain range of the Colombian Andes (Figure 1). The Pacific region is classified as a tropical rainforest with an extremely humid climate and an annual average temperature of 27 °C. The Andean region is further divided into two groups: Andean 1 (800 - 1050 m.a.s.l.) and Andean 2 (1300 - 2200 m.a.s.l.), with highly variable climatic conditions. The inner valleys in Andean 1 are climatically classified as tropical savanna and tend to be dry, with an annual temperature of 25 °C, while the range summits in Andean 2 are more humid, with a temperate climate, tropical monsoons, and an annual temperature of 21 °C (Supplementary table ST1) (Chen & Chen, 2013; Hernández-Camacho, 1992; Kattan et al., 2004; Köppen, 1884; Peel, Finlayson, & McMahon, 2007).
Environmental variation across the three regions was characterized by differences in temperature, humidity, and precipitation, measured as five-year averages of the annual temperature (°C), relative humidity (%), and precipitation (mm), respectively, for each location (IDEAM, 2019). In addition, a climate classification was represented using four categories (1 to 4) of different climatic conditions mediated by the tropical Andes. Variation in topography was estimated by elevation above sea level. A dummy variable was used to evaluate the Andean uplift as a major geographic barrier to gene flow. The code 0 was used for populations from the western side of the western range (Pacific region), and 1 for populations from the eastern side of these mountains (Supplementary table ST1).
Worker ants from nine to twenty nests in ten locations (total of 153 nests) were sampled in the period 2017-2018 (Table 1). The distance between nests in each location was at least 1.5 km, ensuring that the sampled nests were independent colonies. Three, two, and five locations were sampled from the Pacific, Andean 1, and Andean 2 regions, respectively (Table 1).

Molecular methods

DNA extraction and PCR amplification of microsatellite markers

Total DNA extraction was carried out for five workers from each nest (total 765 workers) using TNES lysis buffer (Tris 50mM, NaCl 0.4M, EDTA 100mM, SDS 0.5%), pH 7.5, and chloroform:isoamyl alcohol (24:1), following Wasko et al. (2003), with minor modifications as described by Muñoz-Valencia et al. (2020).
Thirteen microsatellite loci developed for A. cephalotes(Muñoz-Valencia et al., 2020) were used (Supplementary Table ST2). PCR reactions were carried out following Muñoz-Valencia et al. (2020) in a 10 µL volume containing 10 ng of DNA, 1 X Phusion Flash PCR Master Mix (Thermo Fisher Scientific), and 2 µM of each labeled primer. The thermal profile was: 98 °C for 1 min, followed by 34 cycles of 98 °C for 1 s, annealing temperature for 15 s, and 72 °C for 20 s, followed by a final extension step at 72 °C for 1 min. The fluorescent amplified fragments were visualized using an automated DNA sequencer ABI 3130 Genetic Analyzer (Applied Biosystems), and allele sizes were estimated using GeneMapper version 4.0 (Thermo Fisher Scientific).

Sequencing of the mtCOI gene

A 368 bp fragment of the mitochondrial cytochrome oxidase subunit I gene (mtCOI ) was sequenced from one sample per nest, obtaining a total of 146 sequences after discarding failed amplifications and sequencing (GenBank accession numbers: MW245066 - MW245211). PCR amplification was carried out using the universal primers Ben and Jerry, following Kronauer et al. (2004) and Simon et al. (1994). PCR reactions were performed in a 10 µl volume, containing 10 ng of DNA, 1X GoTaq® Master Mix (PROMEGA), and 2 µM of each primer. The thermal cycling profile was 94 °C for 2 min followed by 30 cycles of 94 °C for 1 min, 58 °C for 1 min, and 72 °C for 1 min, with a final extension step at 72 °C for 10 min. PCR products were confirmed by electrophoresis on 1 % agarose gels. All PCR products were sequenced by Psomagen, Maryland, USA.

Population genetics analyses