History sets the stage: Macroevolutionary influence on biotic interactions

2.1 What determines who associates with whom?

The earliest macroevolutionary contributions to the study of biotic interactions were from parasitologists. These studies showed that clades of specialist parasites consisted of species with patterns in character resemblance similar to patterns among their host species (Fahrenholz, 1913; von Ihering, 1891). These patterns in similarity were used to infer phylogenetic histories for the host species, because the simpler morphologies of the parasites relative to their hosts allowed trait-based phylogenies to be resolved more easily (Brooks, 1988). Such studies form the basis by which many study the evolution of ecological interactions, whether in mycorrhizal systems (Shefferson et al., 2007; Taylor, Bruns, Szaro, & Hodges, 2003), herbivory (Espinoza, Wiens, & Tracy, 2004; Poore, Hill, & Sotka, 2008) or disease (Streicker et al., 2010). However, such studies have led to the erroneous belief that one-species-to-one-species relationships are common among parasites and their hosts, and potentially even in mutualistic symbioses.

The typical biotic interaction most certainly involves two or more potential partners or host species for each species in question. While some interactions involve some degree of cophylogeny, or shared evolutionary history (Paterson et al., 1993), the majority do not seem to (Page, 2003). This suggests that in most interactions, a single host species is not sufficient to provide all of the nutritional, reproductive, or benefits that a species might need from it. The result may be generally asymmetric interactions, in which one clade of species needs another clade more than the reverse (Bascompte, Jordano, Melian, & Olesen, 2003), causing different evolutionary histories and dynamics in the two groups and leading to relatively broad interactions characterizing the Tree of Life.

What determines the breadth or assemblage of hosts that a species associates with? Three contributions to this special feature address this question. First, Chagnon et al. (2019) assessed patterns in specialization and host choice using the lichen genus Peltigera. The lichen is a symbiosis between a fungus and a photosynthetic microbial species, in this case a cyanobacterium. The authors found that fungal species associated with a core set of cyanobiont species, and that more closely related fungal species were more likely to share the same cyanobionts. Furthermore, fungal speciation led to the development of new combinations of fungus and cyanobiont, and the same cyanobionts became less likely to be shared between sister fungal species as speciation events became older.

Second, Jacquemyn and Merckx (2019) reviewed the evolutionary origins of mycoheterotrophy in plants. Mycoheterotrophy, or the utilization of mycorrhizal fungi as carbon sources, is an evolutionary puzzle because it involves the adaptive loss of photosynthesis in favour of mycorrhizal fungi as energy sources (Bidartondo, 2005; Merckx, 2013). Mycoheterotrophic plants typically associate with fungi that are ectomycorrhizal hosts of local trees and shrubs, making mycoheterotrophy usually a tri-partite interaction in which the mycoheterotrophic plant is a direct parasite of its mycorrhizal fungus, and an indirect parasite of a local photosynthetic plant species (Leake & Cameron, 2010). Although initially thought to be a curiosity of a small number of plant species, at least 10% of vascular plants experience at least one life stage in which they are effectively mycoheterotrophs (Leake & Cameron, 2010). Some studies on the evolution of this phenomenon have identified mutations coding for breakdown of different aspects of the photosynthetic apparatus (Barrett & Freudenstein, 2008), such as dysfunctional stomates (Roy et al., 2013). However, the authors note that most transitions to mycoheterotrophy have also involved shifts in host fungi, often away from saprotrophs to ectomycorrhizal hosts. Some photosynthesis-related mutations even influence the choice of mycorrhizal partner fungi (Abadie et al., 2006; Yamato, Ogura-Tsujita, Takahashi, & Yukawa, 2014).

Finally, Shefferson et al. (2019) asked whether evolutionary specialization can still occur in seemingly generalized interactions, using a community assemblage approach to assess the impacts of retaining or switching specific hosts. Mutualist interactions often form nested networks, in which one group of species specializes on a core set of hosts but also interacts with a broader suite of interactors (Bascompte et al., 2003). The authors explained this phenomena with the concept of “apparent generalism”, in which a small subset of host species are specialized on because of their unique abilities to fulfil important niche requirements in their partners, and a large set of other functionally redundant host species are associated with to fulfil other niche requirements. They found phylogenetic evidence of apparent generalism in the mycorrhizal interactions of members of the lady's slipper subfamily of the orchid family (family Orchidaceae, subfamily Cypripedioideae), identifying dominant fungal hosts found associating with almost all members of large clades of plants within this subfamily, while other fungi were also present but associated only with scattered species.

2.2 How does evolutionary history constrain adaptation in response to interactions?

Biotic interactions are subject to evolution via natural selection and present novel contexts under which selection can operate. Ecologists are typically familiar with the pioneering work of Paul Ehrlich and Peter Raven to understand co-evolution in plant–herbivore interactions (Ehrlich & Raven, 1964). This work ushered in a new era in microevolutionary ecology that grappled with evolutionary specialization and strict co-evolution, and has led to such developments as the biological market theory for co-evolution (Cowden & Peterson, 2009; Noë & Hammerstein, 1995), the theories of partner choice and fidelity (Sachs, Mueller, Wilcox, & Bull, 2004), and the geographic mosaic theory of co-evolution (Thompson, 1999).

The evolutionary study of interactions has many microevolutionary predictions that are thought to apply universally. However, tests of predictions such as partner choice are typically performed only on model systems, due to the difficulty in designing and carrying out selection experiments on key interactions. When tested more broadly, some predictions seem clade-specific, suggesting some degree of context dependence. For example, the myrmecophytic interaction between species of ants and their acacia hosts was long thought to be mutualistic, but experimental studies suggest that the high cost of tolerating the ant may lead to decreased fitness in some cases (Stanton & Palmer, 2011). Likewise, the legume–rhizobia interaction shows a great deal of variation in benefit to both partners, and this is influenced to some extent by phylogenetic relationships among the rhizobia (Barrett, Zee, Bever, Miller, & Thrall, 2016).

Two contributions to our special feature further our understanding of this topic. First, Coley et al. (2019) combined transcriptomics, field ecology and chemical analysis with phylogenetic perspectives to analyse the evolutionary history and adaptive context behind the production of a key secondary metabolite, tyrosine, which acts as an anti-herbivore compound and a precursor to other anti-herbivore compounds. They present a reconstruction of the evolutionary history of this trait across a clade of legumes showing that overexpression itself evolved once, but was often followed by the production of other secondary metabolites. This is strong evidence that overexpression of tyrosine has evolved as an adaptive means of defending against herbivores, and that it is an important step in the development of derived defensive compounds.

Second, Foisy et al. (2019) test Ehrlich and Raven (1964)’s “escape-and-radiate” hypothesis linking herbivore diversification rates to adaptation to novel defences evolving in plants, a hypothesis originally thought to apply universally. They focus on the evolution of latex and resin canals using a global dataset involving plants from 300 different families based on a now classic study conducted about 30 years earlier (Farrell, Dussourd, & Mitter, 1991). The authors used both sister-taxa comparisons and clade-specific analyses, and found no support for an overall relationship. However, the evolutionary origin of laticifers, or tissues containing latex, in the Papaveraceae corresponded to increased diversification rates, suggesting a relationship specific to only one clade.

2.3 How does evolutionary history influence community, ecosystem, and global ecological processes and patterns?

Darwin imagined that natural selection working slowly over time would yield small differences across related taxa that would compound over time (Darwin, 1859). This is the underlying expectation in assessments of phylogenetic signal and inertia, although it is no longer expected that natural selection is the only mechanism yielding this pattern (Blomberg et al., 2003; Münkemüller et al., 2012). However, rapid evolution is commonplace and likely affects ecological processes occurring at all biological levels (Meester et al., 2019; Shefferson & Salguero-Gómez, 2015). Given this knowledge, how do ecological processes and rapid evolution interact with evolutionary history to determine population, community and ecosystem dynamics (Yguel et al., 2016)?

Evolutionary history has become of particular interest to community ecologists in the last two decades. The field of community evolution has developed as a means of exploring the factors distributing biodiversity across regions at all geographical scales via ecological and evolutionary mechanisms, primarily involving tests of the influences of ecological interactions, such as competition and facilitation, of the abiotic environment as a filter, and of random processes such as dispersal (Emerson & Gillespie, 2008; Kraft, Cornwell, Webb, & Ackerly, 2007; Pashirzad, Ejtehadi, Vaezi, & Shefferson, 2019). Recent research has tied evolutionary processes both to community diversity and ecosystem functioning, finding that natural selection can cause dramatic changes in plant community productivity and implying that speciation and extinction may have similarly dramatic effects on ecosystem functioning in general (van Moorsel et al. 2018). However, the impacts of speciation have only been explored historically (Harmon et al., 2009), leading to uncertainty in the degree to which rapid speciation might affect higher ecological processes.

Three contributions address this theme. First, Pérez-Izquierdo et al. (2019) ask whether genetic variation within a tree species influences the phylogenetic composition of soil microbial communities within the root zone. They used three long-term common gardens using three genetic variants of the maritime pine, Pinus pinaster, and found impacts on microbial phylogenetic structure of plant genetic variation, as well as of elevation and soil conditions. They further found that phylogenetically different microbial communities translated to ecosystem differences, including impacts on phosphorus cycling, and that these differences were primarily linked to differences in mutualist and decomposer communities.

Second, Shroeder et al. (2019) used a phylogenetic perspective to understand how host trees structure soil microbial communities via plant–soil feedbacks. The authors focused on the impacts of spatial distance and plant host species relationships to determine the key determinants of microbial community composition. They found that the degree of phylogenetic relatedness between host trees in a tropical rainforest in Mexico significantly predicted the rhizosphere microbial community. This was particularly true of soil pathogens, suggesting that biological compatibility between host and pathogen exhibits phylogenetic signal.

Third, Goud et al. (2019) developed a metric useful to link phylogenies to ecosystem function. They termed this metric the “integrated metabolic strategy” (IMS), which uses stable isotope ratios to assess how plants deal with fundamental trade-offs involving carbon gain and water loss. They assessed how IMS varies both within and across species of Asclepias grown under controlled conditions, finding much variability at both levels. IMS varied across species in ways highlighting the degree of aridity in their native habitats, suggesting further importance to evolutionary history in shaping leaf-level traits and suggesting a possible relationship to herbivory.