Abstract
One major goal in plant evolutionary ecology is to address how and why tritrophic interactions mediated by phytochemical plant defences vary across species, space, and time. In this study, we tested three classical hypotheses about plant defences: (i) the resource-availability hypothesis, (ii) the altitudinal/elevational gradient hypothesis and (iii) the defence escalation hypothesis. For this purpose, predatory soil nematodes were challenged to hunt for root herbivores based on volatile cues from damaged or intact roots of 18 Alpine Festuca grass species adapted to distinct climatic niches spanning 2000 meters of elevation. We found that adaptation into harsh, nutrient-limited alpine environments coincided with the production of specific blends of volatiles, highly attractive for nematodes. We also found that recently-diverged taxa exposed to herbivores released higher amounts of volatiles than ancestrally-diverged species. Therefore, our model provides evidence that belowground indirect plant defences associated with tritrophic interactions have evolved under two classical hypotheses in plant ecology. While phylogenetic drivers of volatile emissions point to the defence-escalation hypothesis, plant local adaptation of indirect defences is in line with the resource availability hypothesis.
Similar content being viewed by others
Introduction
The predominant function of plant secondary chemistry is to increase tissue toxicity against herbivores1. However, it can also enhance plants’ ability to attract natural enemies of the herbivores2. Volatile organic compounds (VOCs), for instance, can serve as attractive cues for foraging predators and parasitoids aboveground3, as well as entomopathogenic nematodes (EPNs) belowground4. When VOCs, by attracting predators, benefit plants by reducing herbivore performance, they act as indirect defence traits and should be under selection5,6. While most plants release VOCs constitutively, most VOCs are, as a cost-saving strategy7, only induced upon herbivore feeding. Qualitative and quantitative changes in plant volatile emissions following herbivore attacks are assumed to enhance predator attraction8. The magnitude, diversity and intensity of the chemical blend have been shown to vary widely across species and also depend on the local biotic and abiotic conditions8,9. Therefore, the unique phytochemical defence phenotype of a plant is the result of several interdependent factors relating to the evolutionary history of plant species and their adaptation to the biotic and abiotic environment10,11,12. A current major goal in plant defence ecology is to disentangle the contributions of ecological and phylogenetic drivers of the natural variation in plant defence expression13.
Plant defence theories relating to niche occupation suggest that adaptation and speciation into different habitats result in unique plant chemical phenotypes optimized for the local biotic and abiotic conditions14. Within this framework, the abiotic-grounded resource availability hypothesis15 states that plant defences vary substantially across resource gradients, in which nutrient-poor habitats should select for higher investment in defence to protect precious tissue build-up14,16. The biotic-centred latitudinal/elevational gradient hypothesis suggests that plant defence investment should be greater in warmer and more stable regions as biotic interactions such as herbivory are thought to be stronger in such climates17,18,19. Several tests of this hypothesis have indeed shown that plants invest more in defences in tropical regions20,21,22 as well as at low elevation sites23,24 compared to temperate and high elevation regions, respectively. Yet, reviews of such tests have yielded mixed results, both along latitude25 and elevation26, suggesting that the relationship between climate, herbivore pressure, and plant defence level is not constant and linearly-dependent across scales27. These difficulties in predicting plant defence expressions based exclusively on local resources, latitude or elevation call for integrative approaches accounting for highly tractable ecological niches and evolutionary histories of plants. Accordingly, co-evolutionary theory stresses the importance of novel key adaptations in driving phylogenetic diversification through time in response to herbivore pressure28,29. Ehrlich and Raven’s30 co-evolutionary theory proposes a defence escalation hypothesis, in which the evolution of novel traits that promote speciation, such as novel and more potent defence traits, is incremental (and directional) through the diversification of plant clades. Thus, a phylogenetic escalation for more, and more potent defence traits as lineages diversify should be observed31,32.
We here used the natural elevational gradient of the Alps that provides a range of climate, and natural macro-evolutionary processes to test (i) herbivore effects on VOCs-mediated tritrophic interactions, and (ii) three major hypotheses of classic plant defence theory (Fig. 1). We hypothesized that while classical plant defence theory has been developed for aboveground defence traits directly acting on herbivores (e.g. physical defences such as spines, or chemicals such toxins), its predictions should also hold for indirect defence traits (such as VOCs attracting natural enemies of the herbivores), if these traits confer benefit for the plant5, as was recently shown on similar systems6. First, we asked if herbivores induce VOCs that facilitate predator recruitment5. If the cost of volatile induction remains limited, the ability to induce VOCs after herbivory should be widespread across the genus but the degree of inducibility might vary between species because of their speciation into unique biotic and abiotic environments33. Second, we asked if plant speciation within specific local abiotic and biotic conditions shapes VOCs production and nematode recruitment15,16. Because litter decomposition and resource availability are reduced under cold temperatures at high elevation, according to the resource availability hypothesis, alpine species should invest more in VOCs production and nematode recruitment. However, if herbivores are the main drivers of VOCs production and subsequent EPN attraction, according to the latitudinal/elevational gradient hypothesis, sites generating higher biotic interactions, such as warmer and more stable low elevation sites, should favour increased VOCs production17,24,34. Finally, we asked whether there is evidence for defence escalation as plant lineages diversify30,32. If diversification implies a co-evolutionary arms race between plants and herbivores that include predator recruitment, then according to the defence escalation hypothesis, more recently diverged species should produce higher amounts of attractive VOCs involved in tritrophic interactions.
Results
Herbivore-induced VOCs facilitate predator recruitment
To address our aims, we studied VOCs production and EPN attraction on plants with and without root herbivore attack across 18 species of Festuca that grow in the Alps (Figs. 1, S1). On average, across all species, we found that Melolontha herbivory decreased root biomass by about 40% (see supplementary material for description of the statistical methods; Fig. 2a; average 0.8 effect size decrease; and Fig. S2a; MCMCglmm for root biomass; post.mean = −657.9, l–95% CI = −1090.2, u-95% CI = −265.9, sampling = 1000, pMCMC = 0.002), and increased entomopathogenic nematodes (EPNs) recruitment by about 30% when compared to healthy plants (Fig. 2b; 0.48 effect size increase; and Fig. S2b; MCMCglmm for nematodes; post.mean = 96.49 l–5% CI = 31,81, u-95%CI = 160,60, sampling = 868, [31.81, 160.60], pMCMC(868) = 0.002). Predator recruitment is likely mediated by the abundance and identity of herbivore-induced root VOCs released in rhizosphere, which show strong species-level variations (Fig. S3; PERMANOVA; species effect: F17,148 = 3.16, p = 0.001). However, we found no effect of the herbivore treatment on volatile profiles (herbivore treatment effect; F1,148 = 0.81, p = 0.60; root biomass effect; F1,148 = 2.08, p = 0.04, and species by treatment interaction; F17,148 = 0.86, p = 0.88), indicating that, although larvae ate abundantly on the roots, root VOCs induction by M. melolontha larvae, across the multivariate VOCs matrix, was weak. Herbivore effect on VOCs induction was however strong when assessing the total amount of VOCs produced (Table S4), as well as when assessing individual compounds such as the sesquiterpenoids E-β-caryophyllene and isosativene (Table S4, Fig. S4). Finally, we observed a positive association between the total constitutive production of VOCs and their inducibility (Fig. S5; correlation corrected for spurious correlations: r = 0.83, p < 0.001; PGLS, F1,16 = 35.82, p < 0.001).
Plant species adapted to high elevation, cold and humid habitats, recruit more predatory nematodes
The climate of each species’ home site from along the elevational gradient was an important factor determining EPN recruitment. Root herbivore induced-species originating from higher elevation, colder and more humid habitats (Fig. S7) produced almost twice more VOCs and attracted about twice as many nematodes than their lowland congeners from warmer and drier habitats (Fig. 3; see PCA1 axis ranging from negative to positive values, and indicating variation from cold and humid high elevation habitats to warm and dry low elevation habitats, Table 1). Multivariate phylogenetic-corrected regression models of the different climatic variables indicated that the number of precipitation days per growing season (r = 0.7, F1,16 = 15.54, p = 0.001), humidity (r = 0.60, F1,16 = 9.21, p = 0.01) and number of frost days(r = 0.52, F1,16 = 5.85, p = 0.03), which are all intimately associated with elevation, were the best predictors of nematode attraction. Furthermore, we explored the relationship between VOCs production and EPNs attraction with soil variables (Fig. S8, Table S3) and found a significant relationship between herbivore-induced VOCs and soil variables (Fig. S9, distance-based redundancy analysis (dbRDA): F8,9 = 1.67, p = 0.01). The soil effect was mainly driven by soil relative humidity (HR), which was also positively associated with the terpenes β-pinene and E-β-caryophyllene, while soils richer in organic matter (high CEC, pH and CN values) were negatively associated with the production of the terpenes aromadendrene, β-vatirenene and isosativene. On the other hand, none of the soil variables could explain EPN recruitment measured in the bioassays (overall linear model; r = 0.36, F8,9 = 0.18 and p = 0.99).
Phylogenetic escalation for VOCs production and EPNs recruitment
The radiation of Festuca along elevation gradients and across climatic niches carries phylogenetic signal that is expected under a Brownian motion model of trait evolution (K = 0.5, p = 0.04; K = 0.4, p = 0.06 respectively; Figs. S10 and S11, Table S5). Furthermore, our data was remarkably consistent with classic theory: we found phylogenetic escalation in the total amount and diversity of inducible VOCs (Table 1, Fig. 3). Specifically, after 7–9 divergence events from the split that separated the predominantly high altitudinal and latitudinal fine-leaved Festuca species from their broad-leaved congeners, between 13 and 16 million years ago, the plants’ inducible VOCs production and their diversity increased by an average of 4 folds when induced by herbivory. Additionally, we found a weak negative correlation (r = 0.43, estimate = −52.12, F1,16 = 3.79, p = 0.07) between branch length and the first axis of the principal component analysis of the VOCs, representing the axis of maximal scent variation across species (Fig. S10). This indicates that more ancestrally-diverged species tend to have similar chemical composition characterized by monoterpenes and low-weight hydrocarbon chains (VOCs 4, 5, 6, 9, 10, 13, 15), while more recently-diverged species tend to produce additional compounds, which, most importantly, included sesquiterpenes.
Discussion
We tested three major predictions of classic plant defence theory regarding plant direct defences aboveground for the potential to explain a plants’ ability to recruit predators (i.e. indirect defences through tritrophic interactions) belowground. On one hand, findings showing that Festuca species adapted to high elevation cold and humid habitats had high predator recruitment after herbivore attack and chemical defence induction, provided evidence for the resource availability theory. On the other hand, we showed that increasingly-derived Festuca species were associated with highest emissions and diversity of volatile organic compounds (VOCs), thereby supporting the defence-escalation theory.
First, we found that root herbivory induced plants to be more attractive to soil-dwelling entomopathogenic nematodes (EPNs), in general, across 18 grass species of Festuca, when compared to undamaged plants. Aboveground herbivore-induced VOCs were shown to attract predatory arthropods (e.g., wasps and mites) in several dozens of systems and in a variety of habitats2. Later, root herbivores were also shown to induce root VOCs for attracting soil-dwelling EPNs35, but the effect of VOCs induction on predator recruitment were never tested across several plant species simultaneously. In summarizing work on chemically-mediated belowground tritrophic interactions, terpene-based compounds seem to play a major role in root defence36. For instance, corn plants genetically modified for enhanced E-β-caryophyllene production were more protected by EPNs against Diabrotica virgifera corn root larvae in the field37. Nonetheless, several other VOC classes can also induce attraction - as well as repulsion - of EPNs35. Because we found no direct correlation between total VOC production and EPN attraction, the unique identity of the VOC blends should be the main driver of EPN behaviour. Accordingly, previous work on Festuca rubra showed that EPN recruitment relies on the balance between the emissions of three attractants (α-curcumene, 1-undecene, nonadecatriene) and one deterrent (decane) compound38. Contrary to theory based on resource allocation within plants39, we observed a positive association between the total constitutive production of VOCs and their inducibility. Such positive association is not unprecedented21,40,41,42,43. For instance, positive correlations between constitutive and inducible defences were observed across >50 species of milkweeds (Asclepias) and across 21 wild cotton (Gossypium) species. Therefore, we consider our findings as a common pattern resulting from repeated evolution independently of the phylogenetic history, the geography of diversification, and of the cost of induction across species.
Second, our findings show that species growing at high elevation colder and more humid habitats, which are generally less productive44, harbour enhanced emission of root-herbivore induced VOC and better predatory recruitment power. These observations point to the prominent role of the resource availability hypothesis and refute the latitudinal/elevational gradient hypothesis. Our most strongly-supported climatic and soil predictors of induced VOCs production and EPN recruitment were related to temperature and humidity, and organic matter rich soils, respectively. Speciation into more climatically-stressed and resource-poor high elevations has thus been followed by a generalized adaptation to produce more attractive VOC blends. In the case of Festuca, we observed a relatively strong phylogenetic signal for elevation and climatic niche, indicating that more closely-related species tend to share more similar altitudinal zones (e.g., species of the F. pratensis clade in the phylogeny; Fig. S12)45. Nonetheless, we inferred several phylogenetically-independent transitions to high elevation niches across the phylogeny, as shown by the observation that representative alpine species span multiple clades in the phylogeny (Amphigenes: F. pulchella; Eskia-Dimorpha: F. quadrifolia; Exaratae: F. violacea; Festuca: F. alpina, F. halleri; Fig. 2, Figs. S12–14), thus having independently colonized the alpine zone regardless of phylogenetic history. Altogether, our results indicate that fescue radiation into alpine niches from different ancestors has resulted in optimal adaptation towards the production of highly attractive VOCs blends. This suggests that the resource-availability hypothesis is associated with a convergent evolutionary pattern. However, parallel to the plant adaptation perspective, we cannot exclude that insect adaptation may have also driven differences in VOC production along elevation. At low elevation, where plants and root herbivores co-occur frequently, selective pressures may have promoted insect adaptations enabling the control of plant VOC emissions. Along this line, the composition of oral secretions of adapted herbivores has been shown to limit VOC induction46. At the opposite end of the elevation gradient, a relaxation in plant–herbivore interactions may have hampered such insect adaptation process, thereby increasing VOC emissions at high elevation.
Third, we provide the first evidence of defence escalation for plant traits mediating tritrophic interactions between plants herbivores and predators. The continuous model of evolution we used indicates that defence escalation has occurred gradually via continuous character displacements in VOCs, but also according to subsequent speciation events; indeed we found strong correlation between VOCs (I, I-C, H) and the patristic distance and the number of intervening nodes from common ancestor (Table 1). Festuca ancestors likely originated in the lowland southern Palaearctic region and then diversified preferentially upwards into high mountains and also into high latitudinal Northern and Southern Hemisphere regions and colder climates47,48,49. Nonetheless, even when accounting for this directional radiation, we still found consistently stronger inducibility in species that inhabit higher altitudes. To date, other tests of the defence escalation hypothesis include interactions between plants and herbivores (i.e. direct defences). For instance previous work highlighted an increase in more toxic angular furanocoumarins in wild parsnip species50, an increase in tolerance (i. e., the ability to regrow) in milkweeds51 or an increase in alkaloid diversity in Ficus species52. While this pattern might apply to a wide range of defence traits, several studies failed to find correlation between phylogenetic branch lengths or speciation events and defence traits. For instance, the total amount of cardenolides decreases with increasing phylogenetic node numbers in milkweeds51. In the genus Streptanthus, while the total amount of glucosinolates varies independent of the phylogeny, the production of aliphatic glucosinolates increases over evolutionary time, while glucosinolate diversity declines53. Thus, our findings align with previous studies in that not all traits show consistent patterns of defence escalation. Indeed, while induction of VOCs in Alpine Festuca species correlates with phylogenetic branch length, the constitutive production of VOCs varies independently of the phylogeny.
Altogether, we have shown that for high-elevation Festuca species, herbivores induce consistently more and more diverse VOC emissions and attract more EPNs than their low-elevation counterparts. Moreover, our results suggest that higher EPN recruitment in harsh conditions relies on specific VOC profiles rather than on total amount of VOC. This is in line with the assumption that the cost of herbivory depends on the plant’s opportunities to acquire resources. In this case, the short growing season of alpine environments makes the period of resource acquisition brief and intermittent during the growing season. While high-elevation soils contain lower diversity and abundance of arthropods and predators6 we cannot completely exclude moderate amounts of root herbivory by different guilds, and thus we expect selection of the induction of optimally-attractive VOC blends upon unexpected herbivore attack. Finally, with high altitude conditions, we might also expect pleiotropy of plant defence traits, in which the production of specific secondary metabolites can ameliorate plant fitness in harsh environments. Indeed, a recent review has indicated that VOC emission rates for phototrophs in extreme environments are frequently higher than that of organisms from less stressful environments54. Therefore, the reason for widespread VOC emissions from different extreme environments and its association with herbivore pressure and phylogenetic history deserves further attention, as these compounds may have important roles in stress resistance and adaptation to extreme habitats.
Material and Methods
Study system
With more than 500 species described, the genus Festuca represents a major evolutionary line of the tribe Poeae. These tufted grasses are herbaceous perennials with a height range of 10–200 cm and a cosmopolitan distribution, growing on every continent except Antarctica47,55. Within the entire European Alpine range, there are about 50–70 species depending on their classification56. From this range, we collected different taxa at 31 sites (Fig. S1, Table S1), ultimately selecting 18 confirmed species with which we were able to perform belowground olfactometer behavioral assays (see below). The plants were collected from pristine sites that represent the optimal niche of each species; www.infoflora.ch) (Fig. S11). At each site, we excavated 20–30 individuals with their own soil and potted them in plastic pots (20 cm diameter, 25 cm high) by completing the native soil with standard soil mixture composed of 2/3 regular potting soil (5 parts compost, 4 parts peat substitute based on wood fiber and 1 part topsoil, pH in water = 7.45, conductivity = 2 mS cm−1, Ricoter, Aarberg, Switzerland) and 1/3 sand in garden plastic pots (25 cm diameter, 30 cm high) in an outdoor common garden at the Botanical Garden of Neuchâtel, Switzerland. For phylogenetic reconstruction of the Festuca species see details in supplementary methods, Tables S1–S2, and Figs. S12–S14.
At each collection site, we also collected soil samples to characterize edaphic variables associated with each species. We sampled about 1 Kg of soil from 10–30 soil cores (5 cm wide and 20 cm deep). Soils samples were homogenized, dried at 40 °C for 48 hours, sieved at 2 mm, and ground using agate mortars for subsequent physicochemical analyses. Specifically, we measured (1) relative humidity (HR) by weighting soils before and after drying at 105 °C for 3 days; (2) organic matter content (OR) through loss on ignition by weighting before and after burning 10 gr of soil at 450 °C for 2 hrs; (3) pH in 2.5 soil volume of deionized water; (4) cation exchange capacity (CEC) following the cobaltihexamine chloride method; (5) carbon to nitrogen ratio (CN) using an elemental analyser (FLASH2000, Thermo Fisher Scientific, Waltham, Massachusetts, United States); (6) total carbonates (CaCO3) using a calcimeter (Bernard type), and (7) bioavailable phosphorous (P_bio) using the Kjeldahl digestion method57. Additionally, for each species we extracted ecological indicator values for soil fertility (Soil_N) based on Landolt et al.58. To measure EPNs infectivity potential of each site we ran an infectivity test shortly after soil collection (maximum one week) to avoid loss of potential EPNs present in soil. Three soil subsamples (200 g) were added with five larvae of the greater wax moth (Galleria mellonella) and stored in the dark59. After six days we checked for dead larvae, which were placed in White traps for nine days and monitored for EPNs emergence. The experiment was run twice on the same soils to corroborate initial results. The average values of each soil variable are summarized in Table S3.
Root herbivores and predatory nematodes
In order to study Festuca interspecific variation in belowground tritrophic interactions along environmental gradients, we used the dietary generalist herbivore, Melolontha melolontha (Coleoptera: Scarabaeidae). The larvae develop in soils by feeding on roots for up to three years, causing severe damage to plants, particularly to Poaceae in low-elevation grasslands, where herbivore population densities can reach 100 individuals/m2, but the beetle larvae practically disappear in the Alpine zone60. M. melolontha larvae were collected in pastures around Switzerland and stored at 10 °C in potting soil. Although M. melolontha is generally found between 300 and 600 m above sea level (asl), Scarabaeidae root herbivores can easily cause damages in mountainous zones61. While we used a single herbivore species in order to compare herbivore effects on different plant species, we assumed that the generalist diet of M. melolontha enables to extrapolate herbivore effects of closely-related Scarabaeidae larva damaging mountainous zones. As soil predators, we used Heterorhabditis megidis (Rhabditida: Heterorhabditidae) entomopathogenic nematodes (EPNs) (Andermatt Biocontrol, Andermatt, Switzerland). EPNs mode of action can be summarized in three steps: i) penetration into the arthropod host by the third instar resistant stage juveniles, ii) release of symbiotic bacteria in the hemolymph of the host and iii) EPNs feeding on the newly-formed bacterial soup for 2–3 generations inside the host62. Species of EPNs in the genera Steinernema and Heterorhabditis, including H. megidis, are found in most habitats, including the Swiss Alps. In a previous survey of nematodes along elevation gradients, we collected H. megidis up to 2000 m asl, and we observed that EPNs, in general, range from 500 to 2750 m. asl63. Most EPNs actively seek their hosts in the soil and, while CO2 can serve as a general cue for host searching, more specific plant-associated volatile compounds have been shown to inform EPNs of the presence of a root insect herbivore in the surroundings4,64. H. megidis relies on host-plant cues for several systems such as coniferous plants, grasses and other angiosperms36. Additionally, nematodes belonging to Heterorhabditis genera cause relatively fast larval mortality (i.e., after 7 days) for different Scarabaeidae species65.
Belowground olfactometer bioassays
After a period of about 10 months (from summer 2016 to spring 2017), including an overwintering period of about 6 months in which all aboveground parts of the plants were cut to the ground, 5 plants per species were randomly selected from the initial pool and divided into two equally-sized tufts for performing olfactometer bioassays. For three species with wide distribution ranges (F. heterophylla, F. ovina, F. rubra), we doubled the sample size in order to include two distinct collection sites per species. Individual tufts of each species were separated in half to obtain unique clonal pairs of plants, which were potted in custom-made two-arm belowground olfactometer pots as described in Kergunteuil et al.6, one pair of clones per olfactometer arm and filled with the same soil mixture as previously described. One month later, one plant per olfactometer was infested with three M. melolontha L2 larvae that were previously starved for three days. Two days after the plant infestation, the central pieces of the olfactometers were filled with 10% moist sand (Neogard, Gontenschwil, Switzerland) and assembled with each pair of plants. Three days after the initial damage, 2000 six-day-old H. megidis EPNs were released in the central chamber of the olfactometers and allowed to move toward either uninfested or infested plants for 24 hours. The behavioural choices of nematodes were recorded by counting the individuals collected in the arms connected to ‘uninfested’ and ‘infested’ treatments after the Baermann funnel extraction4.
Chemical analyses of VOCs
Immediately after the olfactometer bioassays described above, all the plants were removed from the pots, and roots were carefully washed and ground to a fine powder in liquid nitrogen. The root material was then stored at −80 °C before chemical analyses of volatile organic compounds. All analyses were done with a gas chromatograph (GC) (Agilent 7890A) coupled with a mass spectrometer detector (MSD) (Agilent 5975C). The samples were prepared and introduced into the GC with the use of robotic multipurpose samplers (MPS, Gerstel GmBH). The compounds were separated on Agilent HP-5MS columns (30 m length × 0.25 mm i.d., and 0.25 µm film thickness). In all cases, the MSD transfer line temperature was set at 280 °C and the ion source and quadrupole temperatures were set at 230 °C and 150 °C, respectively. The electron impact (EI) mode was used with a scanning over the mass range of 33–250 m/z. The chemical compounds were trapped through the use of solid-phase microextraction (SPME) headspace technique. The samples were incubated in a 20 ml glass vial for 3 minutes at 35 °C before inserting a 100 µm polydimethylsiloxane (PDMS) coated fiber (Supelco, Bellefonte, USA) into the headspace for 20 minutes. Afterward, the compounds were thermally desorbed from the fiber for 210 seconds (splitless mode, 250 °C, 6.5 psi pressure, 210 mL/min purge flow, helium carrier gas) before injection onto the GC column. The initial column temperature of 50 °C was held for 1 min, then it was ramped at 6 °C/min until 250 °C (hold time 1 min), and finally a 3 min post run at 260 °C. The helium flow rate was 0.9 mL/min (constant flow mode). The analysis of root-emitted volatiles was based on 26 compounds that were not detected in blank samples. While 3 volatiles were identified using pure standards, the additional volatiles were tentatively identified by comparing the mass spectra with the NIST05 mass spectra library (Table S4). The peak area for each detected compound was divided by the peak area of the internal standard (Tetralin®, Sigma-Aldrich, St. Gallen, Switzerland), and the root emissions of volatiles were finally given as tetralin equivalent nanograms of compound released by gram of root biomass collected66. For subsequent statistical analyses, we calculated the total amount of VOCs produced per species, and the Shannon diversity index H (function diversity from the package vegan67). We did not include the number of individual peaks (i.e. peak richness), since it is highly correlated with both the total amount (t1,158 = 2.92, p < 0.001), and the Shannon diversity (t1,158 = 6.30, p < 0.001).
Statistical analyses
Effect of herbivory on root biomass, nematode recruitment and VOCs production
To address the impact of the herbivore treatment, we first calculated standardized effect sizes for root biomass and EPNs recruited between herbivore-infested plants minus the control plants using Hedge’s g statistics (function cohen.d in package effsize in R68). We next used Bayesian phylogenetic mixed models (BPMMs), as implemented in the R package MCMCglmm69, to test the effect of the herbivore treatment on average root biomass, EPNs recruitment, total and all individual VOCs emitted across all species, while taking into account phylogenetic relationships among Festuca species. Differences of VOCs production across species and herbivore treatment were also assessed with a permutational multivariate analysis of variance (PERMANOVA) based on Bray-Curtis distance matrices (function adonis from the package vegan67), and average species and treatment differences were visualized using non-metric multidimensional scaling (NMDS) (function metaMDS in vegan) (Fig. S3). All VOCs data were log-transformed to normalize their distribution (function decostand in vegan). To test for potential negative correlations between constitutive VOCs (species means) and their inducibility (i.e., the difference in mean VOCs values for each species between control and damaged plants), we conducted tests to control for potential spurious correlations70 in Matlab (version 7.5.0.342 – R2007b, MathWorks Inc., USA). We also performed phylogenetically controlled analyses to tests for correlations between mean constitutive VOCs levels and their inducibility.
Phylogenetic signals and tests of the defence escalation hypothesis
We first estimated the phylogenetic signal (Blomberg’s K) for all individual VOCs, their sum and their Shannon diversity, and the climatic niche (PCA1) using species-level averages with the function phylosig of the package phytools in R71. A K value of 1 indicates that trait values are consistent with the tree topology and a random walk model (i.e., trait similarity is directly proportional to the extent of shared evolutionary history). A K value close to 0 indicates no influence of shared ancestry on trait values (i.e., phylogenetic independence). When K > 1, the traits have more phylogenetic signal than expected from a random walk model on the given phylogeny45. Next, to test for the defence escalation hypothesis, we calculated the patristic distance from root to tip (i.e. the number of apomorphic step changes separating two taxa on a cladogram), for determining the amount of divergence of each species from their common ancestor. We then performed correlation analyses between the patristic distance and the constitutive, induced and inducibility of VOCs and nematode recruited trait values of the 18 species of Festuca. We also regressed the first axis of a principal component analysis (PCA) of all VOCs (function dudi.pca, packages ade472) against patristic distance. Finally, in addition to using branch length, we also regressed traits values against the number of intervening nodes from root to tip for each species (e.g. as in Agrawal and Fishbein51), although the two measures are highly correlated; r = 0.88, t1,16 = 6.89, p < 0.001).
Climate effect on VOCs production and nematode recruitment
To measure the effect of climate on constitutive, induced and inducibility of VOCs production and EPN attraction, we first constructed the climatic niche optima for each species based on occurrence data, and the first axis of the ordination of the climatic variables (Figs. S6 and S7). We then regressed this first PCA axis with each response variable using phylogenetic corrected analyses (PGLS) (function pgls, package caper73). In all cases, to reduce Type I errors, we optimized branch length thorough maximum likelihood74. We next dissected the importance of each individual climatic variable by running a multivariate PGLS regression model.
Soil effect on VOCs production and nematode recruitment
To address potential correlations between the soil physico-chemical variables and the whole VOCs matrix we mapped the soil variables (Table S3, Fig. S8) on the VOCs ordination based on a distance-based redundancy analysis (dbRDA) (function cca in the package vegan). We performed the analysis using both Euclidean and Bray-Curtis distances, and obtained very similar results, and thus used the Euclidean distance results to show. Finally, we performed a multivariate regression analysis for testing the effect of all soil variables on EPNs attraction.
Data availability
Data underlying this article can be accessed on Dryad Digital Repository at http://dx.doi.org/XXX/doi.goes.here, and used under the Creative Commons Attribution licence.
References
Fraenkel, G. S. The raison d'être of secondary plant substances. Science 129, 1466–1470 (1959).
Turlings, T. C. J. & Erb, M. Tritrophic interactions mediated by herbivore-induced plant volatiles: mechanisms, ecological relevance, and application potential. Annu. Rev. Entomol. 63, 433–452, https://doi.org/10.1146/annurev-ento-020117-043507 (2018).
Turlings, T. C. J., Tumlinson, J. H. & Lewis, W. J. Exploitation of herbivore-induced plant odors by host-seeking parasitic wasps. Science 250, 1251–1253 (1990).
Rasmann, S. et al. Recruitment of entomopathogenic nematodes by insect-damaged maize roots. Nature 434, 732–737, https://doi.org/10.1038/Nature03451 (2005).
Kessler, A. & Heil, M. The multiple faces of indirect defences and their agents of natural selection. Funct. Ecol. 25, 348–357, https://doi.org/10.1111/j.1365-2435.2010.01818.x (2011).
Kergunteuil, A., Röder, G. & Rasmann, S. Environmental gradients and the evolution of tri-trophic interactions. Ecol. Lett. 22, 292–301, https://doi.org/10.1111/ele.13190 (2019).
Agrawal, A. A. Induced responses to herbivory and increased plant performance. Science 279, 1201–1202 (1998).
Dicke, M. & Baldwin, I. T. The evolutionary context for herbivore-induced plant volatiles: beyond the ‘cry for help’. Trends Plant Sci. 15, 167–175, https://doi.org/10.1016/j.tplants.2009.12.002 (2010).
Pichersky, E., Noel, J. P. & Dudareva, N. Biosynthesis of plant volatiles: Nature’s diversity and ingenuity. Science 311, 808–811, https://doi.org/10.1126/science.1118510 (2006).
Gould, S. J. & Lewontin, R. C. Spandrels of San-Marco and the Panglossian Paradigm - a Critique of the Adaptationist Program. Proceedings of the Royal Society of London Series B-Biological Sciences 205, 581–598 (1979).
Agrawal, A. A. Macroevolution of plant defense strategies. Trends Ecol. Evol. 22, 103–109 (2007).
Futuyma, D. J. & Agrawal, A. A. Evolutionary history and species interactions. Proceedings of the National Academy of Sciences of the United States of America 106, 18043–18044, https://doi.org/10.1073/pnas.0910334106 (2009).
Stamp, N. Out of the quagmire of plant defense hypotheses. Q. Rev. Biol. 78, 23–55 (2003).
Defossez, E., Pellissier, L. & Rasmann, S. The unfolding of plant growth form-defence syndromes along elevation gradients. Ecol. Lett. 21, 609–618, https://doi.org/10.1111/ele.12926 (2018).
Coley, P. D., Bryant, J. P. & Chapin, F. S. Resource availability and plant antiherbivore defense. Science 230, 895–899 (1985).
Fine, P. V. A., Mesones, I. & Coley, P. D. Herbivores promote habitat specialization by trees in amazonian forests. Science 305, 663–665 (2004).
Schemske, D. W., Mittelbach, G. G., Cornell, H. V., Sobel, J. M. & Roy, K. Is there a latitudinal gradient in the importance of biotic interactions? Annual Review of Ecology Evolution and Systematics 40, 245–269 (2009).
Galmán, A. et al. A global analysis of elevational gradients in leaf herbivory and its underlying drivers: Effects of plant growth form, leaf habit and climatic correlates. J. Ecol. 106, 413–421, https://doi.org/10.1111/1365-2745.12866 (2018).
Zhang, S., Zhang, Y. & Ma, K. Latitudinal variation in herbivory: hemispheric asymmetries and the role of climatic drivers. J. Ecol. 104, 1089–1095, https://doi.org/10.1111/1365-2745.12588 (2016).
Coley, P. D. & Barone, J. A. Herbivory and plant defenses in tropical forests. Annu. Rev. Ecol. Syst. 27, 305–335 (1996).
Rasmann, S. & Agrawal, A. A. Latitudinal patterns in plant defense: evolution of cardenolides, their toxicity and induction following herbivory. Ecol. Lett. 14, 476–483, https://doi.org/10.1111/j.1461-0248.2011.01609.x (2011).
Pearse, I. S. & Hipp, A. L. Global patterns of leaf defenses in oak species. Evolution 66, 2272–2286, https://doi.org/10.1111/j.1558-5646.2012.01591.x (2012).
Moreira, X. et al. Trade-offs between constitutive and induced defences drive geographical and climatic clines in pine chemical defences. Ecol. Lett. 17, 537–546, https://doi.org/10.1111/ele.12253 (2014).
Pellissier, L. et al. Shifts in species richness, herbivore specialization, and plant resistance along elevation gradients. Ecology and Evolution 2, 1818–1825, https://doi.org/10.1002/ece3.296 (2012).
Moles, A. T. et al. Putting plant resistance traits on the map: a test of the idea that plants are better defended at lower latitudes. New Phytol. 191, 777–788, https://doi.org/10.1111/j.1469-8137.2011.03732.x (2011).
Rasmann, S., Pellissier, L., Defossez, E., Jactel, H. & Kunstler, G. Climate-driven change in plant–insect interactions along elevation gradients. Funct. Ecol. 28, 46–54, https://doi.org/10.1111/1365-2435.12135 (2014).
Johnson, M. T. J. & Rasmann, S. The latitudinal herbivory-defence hypothesis takes a detour on the map. New Phytol. 191, 589–592, https://doi.org/10.1111/j.1469-8137.2011.03816.x (2011).
Mitter, C., Farrell, B. & Futuyma, D. J. Phylogenetic studies of insect-plant interactions: Insights into the genesis of diversity. Trends Ecol. Evol. 6, 290–293, https://doi.org/10.1016/0169-5347(91)90007-k (1991).
Weiblen, G. D., Webb, C. O., Novotny, V., Basset, Y. & Miller, S. E. Phylogenetic dispersion of host use in a tropical insect herbivore community. Ecology 87, S62–S75 (2006).
Ehrlich, P. R. & Raven, P. H. Butterflies and plants - a study in coevolution. Evolution 18, 586–608, https://doi.org/10.2307/2406212 (1964).
Farrell, B. D. & Mitter, C. The timing of insect/plant diversification: might Tetraopes (Coleoptera: Cerambycidae) and Asclepias (Asclepiadaceae) have co-evolved? Biol. J. Linn. Soc. 63, 553–577 (1998).
Vermeij, G. J. The evolutionary interaction among species: selection, escalation, and coevolution. Annu. Rev. Ecol. Syst. 25, 219–236 (1994).
Karban, R. & Baldwin, I. T. Induced responses to herbivory. 319 (The University of Chicago Press, 1997).
Godschalx, A. L., Rodríguez-Castañeda, G. & Rasmann, S. Contribution of different predator guilds to tritrophic interactions along ecological clines. Current Opinion in Insect Science 32, 104–109, https://doi.org/10.1016/j.cois.2019.01.002 (2019).
Rasmann, S. & Turlings, T. C. J. Root signals that mediate mutualistic interactions in the rhizosphere. Curr. Opin. Plant Biol. 32, 62–68, https://doi.org/10.1016/j.pbi.2016.06.017 (2016).
Rasmann, S., Ali, J., Helder, J. & van der Putten, W. Ecology and evolution of soil nematode chemotaxis. J. Chem. Ecol. 38, 615–628, https://doi.org/10.1007/s10886-012-0118-6 (2012).
Degenhardt, J. et al. Restoring a maize root signal that attracts insect-killing nematodes to control a major pest. Proceedings of the National Academy of Sciences of the United States of America 106, 13213–13218 (2009).
Kergunteuil, A., Humair, L., Münzbergová, Z. & Rasmann, S. Plant adaptation to different climates shapes the strengths of chemically mediated tritrophic interactions. Funct. Ecol. 33, 1893–1903, https://doi.org/10.1111/1365-2435.13396 (2019).
Heil, M. & Baldwin, I. T. Fitness costs of induced resistance: emerging experimental support for a slippery concept. Trends Plant Sci. 7, 61–67 (2002).
Agrawal, A. A., Conner, J. K. & Rasmann, S. In Evolution After Darwin: the First 150 Years (eds Bell, M. A., Futuyma, D. J., Eanes, W. F. & Levinton, J. S.) 243–268 (Sinauer, 2010).
Thaler, J. S. & Karban, R. A phylogenetic reconstruction of constitutive and induced resistance in. Gossypium. Am. Nat. 149, 1139–1146 (1997).
Agrawal, A. A. & Fishbein, M. Plant defense syndromes. Ecology 87, S132–S149 (2006).
Steward, J. L. & Keeler, K. H. Are there trade-offs among antiherbivore defenses in Ipomoea (Convolvulaceae)? Oikos 53, 79–86, https://doi.org/10.2307/3565666 (1988).
Nagy, L. & Grabherr, G. The biology of alpine habitats. (Oxford University Press, 2009).
Blomberg, S. P., Garland, T., Ives, A. R. & Crespi, B. Testing for phylogenetic signal in comparative data: behavioral traits are more labile. Evolution 57, 717–745, https://doi.org/10.1554/0014-3820(2003)057[0717:tfpsic]2.0.co;2 (2003).
Sobhy, I. S., Miyake, A., Shinya, T. & Galis, I. Oral secretions affect HIPVs induced by generalist (Mythimna loreyi) and specialist (Parnara guttata) herbivores in rice. J. Chem. Ecol. 43, 929–943, https://doi.org/10.1007/s10886-017-0882-4 (2017).
Catalán, P. In Plant Genome: Biodiversity and evolution Vol. 1D (eds A. K. Sharma & A. Sharma) Ch. 10, 255-303 (Enfield, 2006).
Inda, L. A., Segarra-Moragues, J. G., Muller, J., Peterson, P. M. & Catalan, P. Dated historical biogeography of the temperate Loliinae (Poaceae, Pooideae) grasses in the northern and southern hemispheres. Mol. Phylogen. Evol. 46, 932–957, https://doi.org/10.1016/j.ympev.2007.11.022 (2008).
Minaya, M. et al. Contrasting dispersal histories of broad- and fine-leaved temperate Loliinae grasses: range expansion, founder events, and the roles of distance and barriers. J. Biogeogr. 44, 1980–1993, https://doi.org/10.1111/jbi.13012 (2017).
Berenbaum, M. & Feeny, P. Toxicity of angular furanocoumarins to swallowtail butterflies - escalation in a co-evolutionary arms-race. Science 212, 927–929 (1981).
Agrawal, A. A. & Fishbein, M. Phylogenetic escalation and decline of plant defense strategies. Proceedings of the National Academy of Sciences of the United States of America 105, 10057–10060, https://doi.org/10.1073/pnas.0802368105 (2008).
Volf, M. et al. Community structure of insect herbivores is driven by conservatism, escalation and divergence of defensive traits in Ficus. Ecol. Lett. 21, 83–92, https://doi.org/10.1111/ele.12875 (2018).
Cacho, N. I., Kliebenstein, D. J. & Strauss, S. Y. Macroevolutionary patterns of glucosinolate defense and tests of defense-escalation and resource availability hypotheses. New Phytol. 208, 915–927, https://doi.org/10.1111/nph.13561 (2015).
Rinnan, R., Steinke, M., McGenity, T. & Loreto, F. Plant volatiles in extreme terrestrial and marine environments. Plant Cell Environ 37, 1776–1789, https://doi.org/10.1111/pce.12320 (2014).
Barkworth, M. E., Capels, K. M., Long, S., Anderton, L. K. & Piep, M. B. In Flora of North America: North of Mexico. Vol. 24 389–392 (Oxford University Press, Oxford, UK, 2007).
Aeschimann, D., Lauber, K., Moser, D. M. & Theurillat, J.-P. Flora Alpina. (Haupt 2004).
Allen, S. E., Grimshaw, H. M., Parkinson, J. A. & Quarmby, C. Chemical analysis of ecological materials. (Blackwell Scientific Publications., 1974).
Landolt, E. Flora indicativa. 1. edn, (Haupt Verlag, 2010).
Caroli, L., Glazer, I. & Gaugler, R. Entomopathogenic nematode infectivity assay: comparison of penetration rate into different hosts. Vol. 6 (1996).
Lewis, G. C. & Clements, R. O. In Pest and Disease Management Handbook (ed David V. Alford) 84-122 (Blackwell Science Ltd, 2008).
Hann, P., Trska, C., Wechselberger, K. F., Eitzinger, J. & Kromp, B. Phyllopertha horticola (Coleoptera: Scarabaeidae) larvae in eastern Austrian mountainous grasslands and the associated damage risk related to soil, topography and management. Springerplus 4, 139, https://doi.org/10.1186/s40064-015-0918-6 (2015).
Dillman, A. R. & Sternberg, P. W. Entomopathogenic Nematodes. Curr. Biol. 22, R430–R431, https://doi.org/10.1016/j.cub.2012.03.047 (2012).
Kergunteuil, A., Campos-Herrera, R., Sánchez-Moreno, S., Vittoz, P. & Rasmann, S. The abundance, diversity and metabolic footprint of soil nematodes is highest in high elevation alpine grasslands. Frontiers in Ecology and Evolution 4, 1–12, https://doi.org/10.3389/fevo.2016.00084 (2016).
Turlings, T., Hiltpold, I. & Rasmann, S. The importance of root-produced volatiles as foraging cues for entomopathogenic nematodes. Plant Soil 358, 51–60, https://doi.org/10.1007/s11104-012-1295-3 (2012).
Koppenhofer, A. M. & Fuzy, E. M. Effect of white grub developmental stage on susceptibility to entomopathogenic nematodes. J. Econ. Entomol. 97, 1842–1849 (2004).
Rasmann, S., Erwin, A. C., Halitschke, R. & Agrawal, A. A. Direct and indirect root defences of milkweed (Asclepias syriaca): trophic cascades, trade-offs and novel methods for studying subterranean herbivory. J. Ecol. 99, 16–25 (2011).
vegan: Community Ecology Package v. 2.0-10 http://vegan.r-forge.r-project.org/ (2013).
effsize: Efficient Effect Size Computation (R package version 0.7.4, 2018).
Hadfield, J. D. MCMC methods for multi-response generalized linear mixed models: The MCMCglmm R Package. Journal of Statistical Software 33, 1–22 (2010).
Morris, W. F., Traw, M. B. & Bergelson, J. On testing for a tradeoff between constitutive and induced resistance. Oikos 112, 102–110 (2006).
Revell, L. J. phytools: an R package for phylogenetic comparative biology (and other things). Methods in Ecology and Evolution 3, 217–223, https://doi.org/10.1111/j.2041-210X.2011.00169.x (2012).
Dray, S. & Dufour, A. B. The ade4 package: Implementing the duality diagram for ecologists. Journal of Statistical Software 22, 1–20 (2007).
Orme, D. et al. caper: Comparative Analyses of Phylogenetics and Evolution in R. (2013).
Mazel, F. et al. Improving phylogenetic regression under complex evolutionary models. Ecology 97, 286–293, https://doi.org/10.1890/15-0086.1 (2016).
Acknowledgements
We thank Maiann Suhner for helping with Festuca collection and identification. This work was supported by Swiss National Science Foundation grants (179481 and 159869) to SR, the Spanish Ministry of Economy and Competitivity CGL2012-39953-C02-01grant to PC. MFM was funded by a University of Zaragoza/Santander PhD fellowship. Moreover, PC and MFM were partially funded by a Spanish Aragón Government and European Social Fund Bioflora A01-17 grant.
Author information
Authors and Affiliations
Contributions
A.K. and S.R. conceived the study. A.K., L.H., A.-L.M. collected the data, M.F.M.-A. generated sequence data, P.C. built the phylogeny, S.R., and A.K. analysed the data, S.R., A.K. and A.G. wrote the draft manuscript, all authors contributed to the final version of the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Kergunteuil, A., Humair, L., Maire, AL. et al. Tritrophic interactions follow phylogenetic escalation and climatic adaptation. Sci Rep 10, 2074 (2020). https://doi.org/10.1038/s41598-020-59068-2
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-020-59068-2
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.