Latitudinal Gradients in Induced and Constitutive ...

J Chem Ecol DOI 10.1007/s10886-016-0735-6

Latitudinal Gradients in Induced and Constitutive Resistance against Herbivores Daniel N. Anstett 1,2 & Wen Chen 1 & Marc T. J. Johnson 1,2

Received: 18 March 2016 / Revised: 11 April 2016 / Accepted: 25 July 2016 # Springer Science+Business Media New York 2016

Abstract Plants are hypothesized to evolve increased defense against herbivores at lower latitudes, but an increasing number of studies report evidence that contradicts this hypothesis. Few studies have examined the evolution of constitutive and induced resistance along latitudinal gradients. When induction is not considered, underlying patterns of latitudinal clines in resistance can be obscured because plant resistance represents a combination of induced and constitutive resistance, which may show contrasting patterns with latitude. Here, we asked if there are latitudinal gradients in constitutive versus induced resistance by using genotypes of Oenothera biennis (Onagraceae) sampled along an 18° latitudinal gradient. We conducted two bioassay experiments to compare the resistance of plant genotypes against one generalist (Spodoptera exigua) and one specialist (Acanthoscelidius acephalus) herbivore. These insects were assayed on: i) undamaged control plants, ii) plants that had been induced with jasmonic acid, and iii) plants induced with herbivore damage. Additionally, we examined latitudinal gradients of constitutive and induced chemical resistance by measuring the concentrations of total phenolics, the concentration of oxidized phenolics, and the percentage of phenolics that were oxidized. Spodoptera exigua showed lower performance on plants from lower latitudes, whereas A. acephalus showed no latitudinal pattern. Constitutive total phenolics were greater in plants from lower Daniel N. Anstett and Alice Chen contributed equally to this paper. * Daniel N. Anstett [email protected]

1

Department of Biology, University of Toronto Mississauga, L5L 1C6, Mississauga, ON, Canada

2

Department of Ecology and Evolutionary Biology, University of Toronto, M5S 3B2, Toronto, ON, Canada

latitudes, but induced plants showed higher total phenolics at higher latitudes. Oxidative activity was greatest at higher latitudes regardless of induction. Overall, both latitude and induction have an impact on different metrics of plant resistance to herbivory. Further studies should consider the effect of induction and herbivore specialization more explicitly, which may help to resolve the controversy in latitudinal gradients in herbivory and defense. Keywords Chemical defense . Total phenolics . Bioassay . Oxidative capacity . Macroecology . Latitudinal gradients

Introduction Latitudinal gradients are characterized by changes in numerous abiotic and biotic factors that can influence the evolution of species. For example, longer growing seasons, higher temperatures, and increased diversity of species at lower latitudes are predicted to cause increased selection due to biotic interactions (Dobzhansky 1950; Schemske et al. 2009). In contrast, species found at higher latitudes experience stronger densityindependent selection from abiotic factors (Dobzhansky 1950). Herbivory and the evolution of plant defense is a classic model for testing the role of latitudinal gradients in evolution (Coley and Aide 1991; Moles et al. 2011a). Indeed, plants and insect herbivores represent more than half of the macroscopic diversity of life, and herbivory is the dominant ecological interaction in terrestrial ecosystems (Strong et al. 1984; Turcotte et al. 2014). Along latitudinal gradients, plants are hypothesized to evolve greater investment in defense at lower latitudes, where herbivory is expected to be higher (Coley and Aide 1991; Coley and Barone 1996; Levin 1976; Levin and York 1978; Pennings et al. 2009; Schemske et al. 2009). This

J Chem Ecol

Chemical Resistance

Inducibility

Constitutive Induced

a

Consitutive Resistance

30

40

50

Latitude

b

Constitutive Induced

Chemical Resistance

hypothesis has been called the BLatitudinal HerbivoryDefense Hypothesis^ (LHDH) (Johnson and Rasmann 2011). The LHDH was developed initially for comparing plant defense between species in tropical and temperate regions (Coley and Aide 1991; Coley and Barone 1996; Levin 1976; Levin and York 1978). Recent work has also found evidence for latitudinal clines in plant defense across temperate climates (Hallam and Read 2006; Pearse and Hipp 2012; Rasmann and Agrawal 2011; but see Moles et al. 2011b; Moreira et al. 2014). Recent studies have attempted to extend the LHDH to intraspecific variation within species, where plant defense also is predicted to be greater at lower latitudes because of greater herbivory pressure in warmer climates. However, at the individual species level, impact of other factors such as thermal tolerance of individual herbivores, as well as evolutionary and ecological processes associated with species range limits (Louthan et al. 2015), may cause deviations from traditional expectations. Indeed, while some single species studies do show higher resistance against herbivores at lower latitudes (Anstett et al. 2015; Lehndal and Ågren 2015; Więski and Pennings 2014), many studies report increased resistance at higher latitudes (Abdala-Roberts et al. 2016; Hiura and Nakamura 2013; Martz et al. 2009; Woods et al. 2012). Here, we focused on latitudinal gradients in herbivore resistance traits within species, which includes any trait that affects the preference or performance of herbivores. One reason why past studies might find variable support for latitudinal gradients in defense is because few studies distinguish between the effects of constitutive and induced resistance. Constitutive resistance is the baseline level of expression for an antiherbivore resistance trait in the absence of damage, whereas induced resistance is the level of resistance following induction by herbivore damage, such as increased levels of defensive compounds. We can determine the inducibility of a trait as the difference in the level of a resistance before and after induction (i.e., inducibility = induced trait value – constitutive trait value). Populations at low latitudes are expected to evolve greater constitutive resistance if they experience consistently high herbivory (Fig. 1). By contrast, populations at higher latitudes may evolve greater inducibility if herbivory is less frequent and unpredictable, since induced resistance can reduce metabolic costs of maintaining defenses (Fig. 1). These divergent patterns between constitutive and induced resistance with latitude may be further accentuated by the observation that constitutive resistance and inducibility frequently exhibit an evolutionary trade-off (Bingham and Agrawal 2010; Karban and Baldwin 1997; Kempel et al. 2011). It is unclear how commonly these trade-offs occur across latitudinal gradients. Only three studies have investigated how constitutive and induced resistance vary across latitudinal gradients (Moreira et al. 2014; Rasmann and Agrawal 2011; Więski and Pennings 2014). Among 18 Pinaceae species, constitutive resin was

30

40

50

Latitude

Fig. 1 Illustration of how constitutive and induced resistance can tradeoff along a latitudinal cline in response to varying selection pressures (a) Example of a trade-off between constitutive and induced chemical defenses across latitude. In this example, there is a strong latitudinal cline in constitutive resistance, but this cline is dampened following induced resistance because of a trade-off. Plants at higher latitudes have evolved lower constitutive defenses, but are still under some selection to be able to induce higher chemical resistance (albeit not as high as constitutive resistance at higher latitudes). The inset illustrates the resulting trade-off between inducibility and constitutive resistance. (b) An example where resistance is greater at lower latitudes, with no evidence of a trade-off between constitutive and induced resistance. Plants at higher latitudes have not experienced selection to increase inducible defenses

greater at higher latitudes, while induced resin increased at lower latitudes, suggesting a trade-off between constitutive and induced resistance (Moreira et al. 2014). By contrast, across 50 species of milkweed (Asclepias spp.), both constitutive and induced cardenolides increased at lower latitudes suggesting there was no trade-off (Rasmann and Agrawal 2011). Only one study has considered these trade-offs at the intraspecific level (Więski and Pennings 2014), and they found increased constitutive and induced resistance in low latitude marshes of the shrub species Iva frutescens, but no evidence of a trade-off. Here, we expand work on latitudinal gradients in trade-offs between constitutive and induced resistance within species using Oenothera biennis (Onagraceae, common evening primrose). We recently found that latitudinal patterns in herbivory varied depending on the identity and diet breadth of the herbivore species (Anstett et al. 2014), which showed that tests of LHDH within species exhibit more variable results than comparative studies conducted between species. When investigating genetically based gradients in plant resistance traits in O. biennis, phenolics and their oxidative activity were

J Chem Ecol

important predictors of damage by multiple herbivore species (Anstett et al. 2015). Their concentration generally increased at lower latitudes as predicted by LHDH. However, these results were from field experiments that did not disentangle the contributions of constitutive versus induced resistance. Oenothera biennis can induce local and systemic resistance following damage by a diversity of herbivores (McArt et al. 2013; McGuire and Johnson 2006). Therefore, determining the roles of induced and constitutive resistance in latitudinal gradients of defense represents an important gap in our understanding of the evolution of plant-herbivore interactions on O. biennis. In this study, we investigated genetically based patterns of induced and constitutive resistance from genotypes sampled along a latitudinal gradient within the native range of Oenothera biennis. We characterized resistance using herbivore bioassays with the generalist caterpillar Spodoptera exigua (Lepidoptera, Noctuidae), and the specialist weevil Acanthoscelidius acephalus (Coleoptera, Curculionidae), both of which are leaf chewers. We also examined constitutive and induced resistance (using both herbivory and hormonal induction with jasmonic acid) in total phenolics and the oxidative capacity of phenolics. Specifically, we asked: (1) Do constitutive and induced resistance vary with latitude? (2) Do these latitudinal patterns differ between herbivores that vary in diet breadth (i.e., specialist vs. generalist)? (3) Is there a tradeoff between constitutive and induced chemical resistance? (4) Do these trade-offs drive latitudinal gradients in plant resistance? Answering these questions will help determine how plant-herbivore interactions influence the evolution of complex patterns of defense evolution in a biogeographic context.

Methods Study System Oenothera biennis L. (Onagraceae) is a eudicot forb often found in disturbed habitats such as roadsides, old fields, and along bodies of water. The native range of O. biennis extends over much of eastern North America, including north to the boreal forest of Canada and south to the Gulf coast of the USA. This species has a semelparous lifehistory, forming a rosette before bolting at the end of the first or second year of growth. Oenothera biennis has a genetic system called permanent translocation heterozygosity, which renders plants functionally asexual because all of the seeds from a single plant are genetically identical to one another and the parent plant (Cleland 1972; Rauwolf et al. 2008). Oenothera biennis is a good model for the study of plantherbivore interactions (Johnson 2011). A wide variety of heritable traits of O. biennis influence resistance against a diverse community of generalist and specialist arthropod herbivores (Johnson et al. 2009a, 2014; Johnson and Agrawal 2005). Phenolic secondary metabolites within Oenothera have been

shown to be particularly important in defense against herbivores (Agrawal et al. 2012; Anstett et al. 2015). These phenolics include flavonoids, caffeic acid derivatives, and ellagitannins (Johnson et al. 2009a, 2014), which gain their defensive function by binding proteins (Feeny 1970; Hagerman and Butler 1991) and creating oxidative stress (Appel 1993; Salminen and Karonen 2011; Salminen et al. 2011). Six clonal genotypes were selected in total, two from populations at each of low, medium, and high latitudes across the range of O. biennis (Castleberry, Alabama, USA: 31.1°N, −87.1°W; Hot Springs, Arkansas, USA: 34.5°N, −93.1°W; Galeton, Pennsylvania, USA: 41.7°N, −77.6°W; Marienville, Pennsylvania, USA: 41.5°N, −79.1°W; Notre Dame, Quebec, Canada: 47.6°N, −79.5°W, Slocan River, British Columbia, Canada: 49.7°N, −117.5°W). In a previous common garden field experiment, levels of phenolics and other traits were quantified from these genotypes, and we selected genotypes that accurately captured the observed latitudinal variation in phenolics, but without any knowledge of how constitutive and induced resistance contributed to this variation (Anstett et al. 2015). We measured resistance by using insect bioassays and measurements of phenolics. Bioassays were performed with one generalist caterpillar, S. exigua, and one specialist weevil, A. acephalus. Spodoptera exigua is a leaf-chewing herbivore with a broad host range (Berdegué et al. 1998), which has been utilized extensively as a model organism of resistance to generalist herbivores on Oenothera (Johnson and Agrawal 2005; Johnson et al. 2009b; McGuire and Johnson 2006). Acanthoscelidius acephalus feeds on the leaves at the apical portion of bolting O. biennis plants (Johnson and Agrawal 2007). Spodoptera exigua were acquired as third instars (Benzon Research Inc., Carlisle, PA, USA), while A. acephalus were collected from the field in Mississauga (43.52°N, 79.60°W) and Newmarket (44.03°N, 79.52°W), Ontario, Canada, from mid-June to early July 2014.

Growing Conditions and Treatments All experiments were conducted in a Conviron CMP6050 (Winnipeg, Canada) walk-in growth chamber. Throughout the experiment we used a light intensity of 500 μmol/m 2 /s on a 16:8 h day:night cycle and 25 °C day: 20 °C night temperature regime with a ramp rate of 1 °C/h to simulate natural conditions. Seeds were germinated on moistened filter paper in petri dishes sealed with Parafilm. Seedlings were planted in 0.5 L square plastic pots filled with Pro-Mix® BX Mycorrhizae™ (Premier Tech Horticulture Ltd). One week after the seedlings were planted, ~0.4 g of slow release nutrient pellets (Nutricote® Total 13–13-13, Chiiso-Asahi Fertilizer Co. Ltd.) were placed in each pot.

J Chem Ecol

Ten days after fertilization, all plants were randomized and subjected to one of the three treatments: unmanipulated control, 1 mM jasmonic acid (JA) induction, or S. exigua herbivory induction. Each treatment was replicated across 10 plants per genotype. Jasmonic acid is a master regulator of induced resistance (Creelman and Mullet 1997; Memelink 2009; Thaler et al. 1996), and exogenous application of JA has been shown to increase resistance against herbivores in O. biennis (McArt et al. 2013). The JA treatment was utilized to complement the herbivory treatment to test the effects of induced resistance in the absence of tissue loss due to herbivore damage. The leaves of all JA treated plants were sprayed with 3 ml of 1 mM JA 2.5 weeks after planting. Caterpillar damaged plants received 2–4 Spodoptera exigua larvae per plant depending on the size of the plant. Caterpillars were removed once they removed ~5 % of the plant’s total leaf area within 2–3 d. All plants, regardless of treatment were enclosed within spunpolyester bags, which effectively contained caterpillars and standardized the effects the bags. Bioassay Experiment Six days after the herbivore induction treatments were applied, four fully expanded rosette leaves were removed from each plant with a razor blade. The leaf area was measured from each plant, and two leaves per genotype (one for each assay insect species) were individually placed into separate 90 mm Petri dishes lined with moistened filter paper; the other two leaves were used in chemical assays (see below). The number of replicates for each bioassay ranged from 8 to 10 leaves for each of the six genotypes for each treatment combination (total = 341 replicates across both bioassay species). In one half of the Petri dishes we placed a freshly hatched 1st instar S. exigua caterpillar, and in the other half we placed a single weevil. On the 5th day of the bioassay, all insects were removed. Caterpillar performance was quantified by assessing the mortality and the wet mass of each surviving larva measured on a 10−7 g UMX2 Ultra Microbalance (Mettler Toledo Canada, Mississauga, ON, CA). Caterpillars were placed in a tube for 6 h without food prior to measurements to allow them to void their gut contents. We did not weigh the weevils because adults exhibit determinate growth. We measured the amount of leaf area removed (mm2) by both caterpillars and weevils as a measure of resistance. Extraction of Phenolics We quantified total phenolics and oxidative activity from leaf tissue taken from the same plants used in the bioassays (see above). Two leaves collected after induction were taken from all plants, flash-frozen, and stored at −80 °C in sealed bags to prevent phenolic oxidation. Leaves were freeze-dried (Labconco ® FreeZone 4.5 L Console Freeze Dry System, Kansas City, MO, USA) for two nights, and subsequently ground to a powder using a CryoMill (RETSCH, Haan, Germany). Leaf powder (20 mg)

was measured into each tube and macerated in 1400 μl 80 % V/V acetone in 2 ml microcentrifuge tubes, vortexed for 5 min, and stored at 4 °C overnight. Macerated samples then were vortexed for 5 min and placed in a shaker for 3 h (22 °C; 280 rpm). Samples were centrifuged for 5 min at 18,000 g (Centrifuge 5430R, Eppendorf, Hamburg, Germany) and decanted into new microcentrifuge tubes. The remaining liquid was capped with punctured lids and placed in a SpeedVac Concentrator (SVC100H, Refrigerated Concentration Trap RT100 and High Vacuum Pump VP100, Savant Instruments Inc., UK) for 50 min to draw out the acetone. The remaining solids were macerated in an acetone/water solution (80:20 V/ V), centrifuged for 5 min at 18,000 g, and decanted again with the same procedure, except the supernatant was poured into the tube containing the previous supernatant. Extracts were subsequently dried in the concentrator for 48 h. Before vortexing for 5 min, 1000 μl of dH2O were added to each tube, followed by centrifugation for 5 min at 16,000 g (Spectrafuge 24D, Labnet, Mandel, Edison, NJ, USA). The supernatant was kept as the stock phenolic extract for each sample and used in downstream analyses. An aliquot of each extract (100 μl) was diluted to 20 % V/V in water for quantification of phenolics. Solution Preparation for Phenolic Quantification The following solutions were used for phenolic quantification: (1) 0.6 % formic acid (Buffer A, Sigma Aldrich), (2) 50 mM pH 10 carbonate buffer (Buffer B), (3) Buffer C (9:5 ratio per volume of buffer A to buffer B), (4) 20 % m/v Na2CO3 (Sigma Aldrich, Japan), (5) 2 mg/ml gallic acid (GA) (Sigma Aldrich), and (6) 1 N Folin-Ciocalteu (FC) reagent (SigmaAldrich). Buffer B was prepared with NaHCO3 (Sigma Aldrich) and adjusted to pH 10 with NaOH (Fisher Scientific) using a pH meter (sympHony SB70P, VWR). Gallic acid was dissolved with a few drops of 95 % EtOH and diluted to 2 mg/ml with dH2O. GA standards of 1.0 mg/ ml, 0.5 mg/ml, and 0.1 mg/ml were prepared from the 2 mg/ml GA. Quantification of Phenolics Total phenolic concentration and oxidative capacity of phenolics were estimated using a modified version of Salminen and Karonen (2011). Briefly, 20 μl of each diluted sample were loaded into 96 - well plates in triplicate for both the total phenolics and oxidative capacity measurements. Buffer A (180 μl) was added to 20 μl of extract to each well to conduct the oxidative capacity measurements, while 280 μl of Buffer C were added to each well to conduct the total phenolics measurements. The plate was incubated in a spectrophotometer (Multiskan GO, Thermo Fisher Scientific Oy, Finland) at 25 °C for 60 min (10 s of shaking every 10 min) monitored by SKANIT Software 3.2 (Thermo Scientific). Buffer B (100 μl) was added to the Buffer A mixture immediately after incubation to stop

J Chem Ecol

phenolic oxidation, and 50 μl of each buffer-extract mixture were transferred to a new 96 well plate that also contained the GA standards for the calibration curve. The FC reagent (50 μl) was added to each well before shaking the plate for 10 s and adding 100 μl of 20 % m/v Na2CO3 to each sample. The plate was incubated at 25 °C for 30 min (10 s of shaking every 10 min), and absorbance was read at 730 nm. Three replicates for each gallic acid standard and buffer-extract mixture were used. The absorbance readings were compared to the calibration curve, giving phenolic concentrations in mg/g equivalents. Statistical Analyses All statistical analyses were carried out in R (R Core Development Team 2012). We determined the effect of latitude, induction, and their interaction on herbivory by both insect species, Spodoptera mass, and phenolic chemistry using Type III sums-of-squares implemented using the anova function in the CAR package. The glm function with family = binomial was used to carry out a logistic regression on the S. exigua mortality data with the factors identified above. In these models, we set contrasts to be orthogonal with each induction treatment compared to the control. Evidence for trade-offs between constitutive and induced defenses were assessed with a linear regression of inducibility versus constitutive resistance, fit using the lm function in R. We measured inducibility as: trait value in the induced state (induced plants) – trait value in the constitutive state (control plants) for all chemically related traits. For insect herbivory related metrics, the above calculation was performed, except that the results were multiplied by −1 to account for the fact that herbivore related metrics are in terms of herbivore benefit and, thus, the reciprocal of plant chemical resistance.

Results Herbivore Performance Latitude and the induction treatment both predicted the performance of the generalist caterpillar (Table 1). Herbivory by the generalist caterpillar was greater on genotypes from higher latitudes (i.e., constitutive resistance decreased with increasing latitude). There was no effect of induction, and no interaction between latitude and induction (Table 1, Fig. 2A). Latitude and the induction treatment interacted to affect caterpillar mass (Table 1, Fig. 2B). Control plants showed an increase in caterpillar mass with increasing latitude, while caterpillar induced plants exhibited induced resistance at low latitudes and induced susceptibility at high latitudes (Fig. 2B). Herbivore induced plants showed the strongest positive relationship with latitude, and their slope was marginally significantly different from the control plants (t164 = −1.89, P = 0.06). The slope of JA treated plants was negative but not significantly different from the

Table 1 Effects of induction treatment and latitude of origin of the plant genotype on the performance of herbivores and secondary chemistry of Oenothera biennis SS

df

F-value

P-value

Spodoptera exigua Herbivory Treatment

3.08

2

1.70

0.19

Latitude Treatment x Latitude

13.99 3.63

1 2

15.44 2.00