INTRODUCTION
Phytochemical diversity, or the richness and abundance of the secondary compounds produced by plants, is a key axis of the functional phenotype that affects plant survival within its biotic and abiotic environment (Jones & Firn 1991; Romeo et al. 1996; Hunter 2016). Ecologists still struggle to understand, not only the origin of phytochemical diversity, but also to quantify the consequences of ecologically-relevant dimensions of phytochemical diversity (Richardset al. 2015), and how the functional axis of phytochemical diversity relates to the functional axis of plant growth form (Díazet al. 2016; Durán et al. 2019). The overall assumption is that a plant’s phytochemical make-up is the result of its evolutionary history (Becerra 1997; Futuyma & Agrawal 2009b), as well as its adaptations to the environment (Coley et al. 1985; Fine et al. 2004; Defossez et al. 2018). Several ecological and evolutionary hypotheses have been proposed for explaining variation in phytochemical diversity, including the co-evolutionary hypothesis (Ehrlich & Raven 1964), the screening hypothesis (Firn & Jones 2003) and the resource availability hypothesis (Coley et al. 1985). The aim of this study is to merge these hypotheses in order to explain patterns of variation in the coupled plant growth form-phytochemical phenotypes related to anti-herbivore defences across closely related species that together have colonized large-scale climatic gradients.
From a co-evolutionary perspective, the concept of an arms race between plants and herbivores has been proposed for explaining the ever increasing diversity of plant secondary compounds over evolutionary times (Ehrlich & Raven 1964). The idea being that herbivores, in particular insects, impose strong selection pressure on plants to evolve novel key adaptations for escaping their enemies. Therefore, a phylogenetic escalation for more, and more potent, phytochemical defence traits should be observed as lineages diversify (Vermeij 1994; Farrell & Mitter 1998). For instance, it was shown that parsnip plants evolved more complex angular forms of furanocoumarins from more simple linear furanocoumarins (Berenbaum & Feeny 1981), or that more complex forms of cardenolides in milkweeds (Asclepias spp.) have emerged from more simple forms as the consequence of the co-evolution with their associated cerambicid beetles in the genus Tetraopes (Farrell & Mitter 1998). Accordingly, it is predicted that first, the presence of diverse forms of toxic phytochemicals in plants should depend, at least partially, on the species evolutionary history, with more recently-derived species to bear more complex levels of phytochemical diversity compared to ancestrally-derived species, and second, that more closely related species should be more similar in their phytochemical make-up than distantly-related species. In other words, we should observe a phylogenetic signal for phytochemical diversity across species (Agrawal et al. 2009).
Along the same lines, the screening hypothesis proposes that phytochemical diversity is maintained because it increases plants’ resistance against both generalist and specialist herbivores (Lewinsohn & Gijzen 2009; Ali & Agrawal 2012). Accordingly, Richards et al. (2015) showed that within the genus Piperaceae high phytochemical diversity is associated with high diversity of herbivores, but also with lower herbivore damage, indeed highlighting a positive effect of phytochemical diversity in increasing resistance against herbivores. Two mechanisms have been proposed for how phytochemical diversity could favour plant resistance against herbivores. First, with high phytochemical production, a plant is more likely to contain a potent compound that is effective against a major herbivore, cumulatively creating a selective advantage within a population (Firn & Jones 1996). For instance, only a few of the 100-plus gibberellins have a known biological activity, but those few that are active are potent at nano molar amounts (Fischbach & Clardy 2007). However, Berenbaumet al. (1991) found that furanocoumarins in Pastinaca sativa are all equally and effectively toxic to a wide variety of herbivores. Second, high levels of phytochemical diversity might result in effective combinations of compounds that work synergistically against herbivores (Berenbaum & Neal 1985; Rasmann & Agrawal 2009; Richardset al. 2012), such as when the impact of nicotine on the generalist Spodopotera exigua caterpillars is enhanced by proteinase inhibitors in leaves of wild tobacco plants (Steppuhn & Baldwin 2007).
Altogether, the screening hypothesis indicates that selection should favour higher levels of phytochemical diversity, particularly in habitats where herbivore pressure is high. Within this framework, it has been long postulated that because warmer and more stable tropical or lowland environments generate higher levels of plant-herbivore interactions (Dobzhansky 1950; Schemske 2009), it should lead to increased defence mechanisms compared to colder and less stable environments such as temperate locations or high elevation (Coley & Barone 1996). Nevertheless, reviews on the topic have also shown contrasting patterns of defence investment along both latitude (Moleset al. 2011) and elevation gradients (Rasmann et al.2014b). This could be explained by other factors also influencing a plant defensive phytochemical make-up. For instance, the resource availability hypothesis (Coley et al. 1985) states that environmental resources, such as soil nutrients, dictate how much a plant can invest in growth and in defences. Specifically, it was predicted, and later shown, that tropical plants growing in resource-poor sandy soils, grow more slowly and are more defended compared to their congeners that live in the nearby resource-rich clay soils (Fine et al. 2004). Similarly, alpine Cardamine(Brassicaceae) species, living in resource poor soils, produce more secondary metabolites (glucosinolates) than their low-elevation congeners (Pellissier et al. 2016; Defossez et al. 2018). Therefore, a holistic approach that encompasses environmental gradients, and their biotic and abiotic correlates, within a phylogenetic comparative framework is needed to tease out the intricate processes generating chemical diversity in plants.
To this end, we performed comparative analyses of severalCardamine species growing along the elevation gradient of the Alps. All Cardamine plants have been shown to produce a wide array of glucosinolates (hereafter refereed to GSLs) (Pellissieret al. 2016). GSLs are sulphur- and nitrogen-containing plant secondary metabolites that, upon tissue disruption, undergo a myrosinase-catalysed hydrolysis generating a variety of by-products, including nitriles, isothiocyanates, thiocyanates, oxazolidine-2-thione, and indole, that are toxic to both specialist and generalist insect herbivores (van Dam et al. 2009). We measured diversity of GSLs across species, generated novel indices that functionally characterize GSL diversity, and used a previously-collected set of growth-related traits affiliated with each species’ growing habit (Defossez et al. 2018). This natural system allowed asking: 1) is variation in GSL diversity across species correlated with species’ phylogenetic distance? We predicted to observe phylogenetic signal for glucosinolates diversity, meaning that that closely related species are more similar in their phytochemical make-up than distant-related species. 2) Does variation in GSL diversity simultaneously converge with plant species adaptation to their specific environment? Since along the elevation gradient of the Alps, similar habitats should generate similar types and levels of herbivory (Hodkinson 2005), we predicted that adaptation to a specific environment, not only shapes the plant growth phenotype, but also structures a unique chemical phenotype. 3) How are different metrics of phytochemical diversity related to plant-herbivore interaction? Because each metric of phytochemical diversity can only capture a fraction of the chemical complexity, we predicted that not all metrics of phytochemical diversity similarly predict plant resistance against specialist and generalist herbivores (Wetzel & Whitehead 2020). With this work, we thus expand on the ecological and evolutionary processes that drive and maintain phytochemical diversity across space and time, and integrate the functional axis of phytochemical diversity with the functional axis of plant growth forms.