Caterpillars on a phytochemical landscape: the case of alfalfa and the Melissa blue butterfly

Abstract

25 26 42 including alkaloids, phenolic glycosides and saponins. The saponins are represented in our data 43 by more than 25 individual compounds with beneficial and detrimental effects on caterpillars, 44 which highlights the value of metabolomic data as opposed to approaches that rely on total 45 concentrations within defensive classes. 46 47 48 49 50 51 52One of the conceptual pillars of trophic ecology is the idea that herbivores must overcome the 53 barrier of plant secondary chemistry before extracting the nutrients necessary for growth and 54 reproduction (Feeny et al., 1992). The success of this idea is reflected in areas of research that 55 include coevolution (Agrawal et al., 2012), ecological specialization (Dyer, 1995), and nutrient 56 flow in ecosystems (Olson, 1963). In most cases, progress has been made by chemical ecologists 57 focusing on small subsets of the secondary metabolites produced by plants and consumed by 58 herbivores. The focus on a few charismatic molecules or classes of compounds, such as 59 furanocoumarins (Berenbaum, 1983) or cardiac glycosides (Zalucki et al., 2001), was at least in 60 part necessitated by early methods in natural products chemistry that were targeted and not easily 61 optimized for the discovery of large suites of co-occurring primary and secondary metabolites 62 (Maag et al., 2015; Dyer et al., 2018). As technological limitations have dissipated, the 63 opportunity now exists for a deeper understanding of the challenges faced by herbivores, with 64 the possibility of discovering, among other things, novel compounds and synergistic interactions 65 among compounds (Prince & Pohnert, 2010; Richards et al., 2010; Sardans et al., 2011). More 66 generally, an important task is to quantify the phytochemical dimensionality of the antagonistic 67 interaction between plants and herbivores, with an eye towards understanding constraints on the 68 evolution of both players (Fordyce & Nice, 2008; Macel et al., 2010) and predicting the 69 formation of new plant-herbivore interactions (Erbilgin, 2018). Here we use the example of a 70 specialized herbivore and a novel host plant to investigate the phytochemical landscape from the 71 perspective of developing caterpillars. 72 The Melissa blue butterfly, Lycaeides melissa, specializes on larval host plants in the pea 73 family (Fabaceae), primarily in the genera Astragalus and Lupinus. Within the last 200 years, L. 74 melissa has colonized introduced alfalfa, Medicago sativa (Fabaceae), at least twice and 75 probably multiple times (Chaturvedi et al., 2018), forming a heterogeneous patchwork of 76 association throughout the range of the butterfly in western North America. M. sativa is a 77 suboptimal host, relative to native hosts that have been examined, and populations of the 78 butterfly that persist on M. sativa show evidence of loss of preference for native hosts (Forister et 79 al., 2012), reduced caterpillar performance on native hosts, and a slight increase in ability to 80 develop on the suboptimal novel host (Gompert et...

Predicting patch occupancy reveals the complexity of host range expansion

“…Previous experimental work and surveys of genetic variation in the field have suggested that antagonistic pleiotropy is not an important constraint on expansion of diet breadth for the Melissa blue (10,14). The effect of specific leaf area is also consistent with a previous experimental study (17) but is small compared to both positive and negative indirect effects associated with plant size and the density of flowers mediated through enemies and competitors (Fig. 2B).…”

“…For example, ants had the strongest effect on butterfly presence and absence (16) and one of the strongest effects on the abundance of tended herbivores. Specific leaf area was highly ranked for the Melissa blue (table S2) with a negative effect as previously observed in an experimental context (17), and specific leaf area had a negative effect on other (largely chewing) herbivores (table S2) but not on other ant-tended (sucking) herbivores, which likely interact differently with physical leaf traits.…”

“…Individual metabolites, for example, have a mix of positive and negative direct effects on the Melissa blue ( Fig. 3F), as observed in a previous rearing experiment (17), while the indirect effects of individual compounds are characterized more by positive effects mediated through numerous other species in the wild (Fig. 3G).…”

“…Each of the randomly selected focal plants was measured for size (as a box measurement of length, width, and height), and the number of flowering stems was counted. Three mature (but not senescent) leaves were selected haphazardly for leaf toughness measurements using a penetrometer (Chatillon 516 series) through the center of the middle leaflet (17).…”

“…For the latter (area and mass), five leaves were selected haphazardly from each envelope and weighed to the nearest tenth of a milligram on a microbalance and taped to a sheet of plain white paper that was then scanned. Leaf area (in cm 2 ) was taken using ImageJ software (version 1.52A) on the scanned images of individual leaves and used to calculate specific leaf area as area divided by mass (17). Because all of the analyses presented here focus on variation at the patch scale, plant measurements were averaged across plants within a location.…”