J Chem Ecol (2011) 37:1063–1070 DOI 10.1007/s10886-011-0018-1
Chemical Defense Across Three Trophic Levels: Catalpa bignonioides, the Caterpillar Ceratomia catalpae, and its Endoparasitoid Cotesia congregata Evan C. Lampert & Lee A. Dyer & M. Deane Bowers
Received: 24 May 2011 / Revised: 6 September 2011 / Accepted: 14 September 2011 / Published online: 24 September 2011 # Springer Science+Business Media, LLC 2011
Abstract Plant secondary chemistry can vary among plant tissues, individuals, and populations, and this variation has population-level consequences for upper trophic levels. In this study, we examined the multi-trophic consequences of variation in iridoid glycosides, which are a component of plant defense against generalist herbivores and also contribute to the unpalatability of sequestering herbivores to both vertebrate and invertebrate predators. Several populations of Catalpa bignonioides were located and examined for the presence of the Catalpa Sphinx, Ceratomia catalpae, a specialist herbivore of Catalpa. We quantified iridoid glycoside content in Catalpa Sphinx caterpillars and in damaged and undamaged C. bignonioides leaves. Overall, leaves of C. bignonioides that were damaged by Catalpa Sphinx caterpillars contained lower concentrations of two major iridoid glycosides, catalpol and catalposide, than leaves of undamaged trees from naturally occurring populations. Catalpa Sphinx caterpillars sequester only catalpol, and increasing catalpol and catalposide concentrations in leaves were associated with increased catalpol
E. C. Lampert : M. D. Bowers (*) University of Colorado Museum of Natural History and Department of Ecology and Evolutionary Biology, UCB 334, University of Colorado, Boulder, CO 80309, USA e-mail:
[email protected] L. A. Dyer Biology Department, University of Nevada, Reno, NV 89557, USA Present Address: E. C. Lampert Biology Department, Gainesville State University, Gainesville, GA 30509, USA
sequestration by caterpillars. The parasitoid Cotesia congregata develops successfully inside catalpol-sequestering Catalpa Sphinx caterpillars, and we examined parasitoid larvae for the presence of catalpol. Parasitoid larvae dissected from caterpillars contained catalpol, but at lower concentrations than their host caterpillars. The variation in chemical defense documented here has rarely been documented over multiple trophic levels, but such resolved systems are ideal for examining competing hypotheses about the effects of plant secondary metabolites on higher trophic levels. Key Words Catalpa Sphinx . Catalpol . Catalposide . Iridoid glycoside . Sequestration . Bignoniaceae . Lepidoptera . Sphingidae . Hymenoptera . Braconidae
Introduction Secondary metabolites produced by plants influence interactions of plants with their pathogens, herbivores, and competitors. Concentrations of these compounds vary within and among populations, among different plant tissues, and over the course of a plant’s lifetime (Krischik and Denno, 1983) as a result of genetic variation as well as phenotypic plasticity. Changes in resource availability and attack by herbivores and pathogens can cause either increased or decreased levels of such compounds (Harvell, 1990; Karban and Baldwin, 1997; Agrawal et al., 1999; Barton, 2008), with changes occurring within a few hours or days (Baldwin, 1987; Fuchs and Bowers, 2004; Dyer et al., 2004) or as long as weeks or months (Haukioja, 2006; Barton, 2008). As a result of variation in plant genotype, resource availability, herbivore or pathogen attack, the intensity of that attack, and the direction and time course
1064
of the plant response, different populations of a single plant species may provide highly variable targets for prospective herbivores. For insects that sequester defensive compounds produced by their host plants, variation in these compounds may have substantial consequences for these insects and their interactions with their natural enemies. For example, individual insects that feed on plants with higher levels of these compounds may suffer fitness consequences (Camara, 1997; Fordyce and Nice, 2008; Smilanich et al., 2009). Yet, at the same time, sequestered compounds may be important in protecting insects from their own enemies, and higher levels of these compounds may provide enhanced protection (Skelhorn and Rowe, 2006). While the efficacy of sequestered defensive compounds has been demonstrated for predators, less is known about how the compounds affect parasitoids. There is little reason to assume that consumers at higher trophic levels are unable to evolve adaptations to dietary toxins, such as compounds sequestered by herbivores, yet relatively few studies have documented the transfer of sequestered compounds to higher trophic levels (e.g., McDougall et al., 1988; Rossini et al., 2000; Bowers, 2003; Reudler Talsma, 2007). Endoparasitoid larvae spend their entire larval development inside a single host where they often are exposed continuously to compounds sequestered by the host (Ode, 2006), and they are likely to ingest substantial amounts of these compounds. Because the midand hind-gut are not connected in hymenopteran endoparasitoids until emergence or pupation (Quicke, 1997), larvae of these species are incapable of egestion, and any ingested compounds not metabolized remain in their bodies. For instance, tomatidine, the aglycone of the alkaloid αtomatine, was detected in larvae of the ichneumonid endoparasitoid, Hyposoter exiguae, that were developing within Helicoverpa zea caterpillars reared on alkaloidspiked artificial diet (Campbell and Duffey, 1979). Several other studies have detected plant-produced or derived secondary metabolites in adult parasitoids, with concentrations magnified in adults of some species (McDougall et al., 1988; Campos et al., 1990; Reudler Talsma, 2007) and minute in others (Rossini et al., 2000; Bowers, 2003; Reudler Talsma, 2007). In this study, we examined variation of catalpol and catalposide (a derivative of catalpol), two iridoid glycosides produced by Southern Catalpa trees, Catalpa bignonioides Walter (Bignoniaceae). We examined effects of these two defensive compounds on the primary herbivore of Catalpa, the Catalpa Sphinx, Ceratomia catalpae Boisduval (Lepidoptera: Sphingidae) and its primary endoparasitoid, Cotesia congregata Say (Hymenoptera: Braconidae). Iridoid glycosides (IGs) are monoterpene-derived secondary compounds that mediate interactions between plants, their
J Chem Ecol (2011) 37:1063–1070
herbivores and herbivore natural enemies (Bowers, 1993; Nishida, 2002 and references therein). Iridoid glycosides are sequestered by four orders of herbivores, including the Catalpa Sphinx (Bowers, 2003), and sequestered IGs can deter predators (e.g., Bowers, 1980; Bowers and Farley, 1990; Dyer and Bowers, 1996; Camara, 1997; Theodoratus and Bowers, 1999; Rayor and Munson, 2002). This study addressed three questions about variation in IGs in Catalpa trees, Catalpa Sphinx caterpillars, and C. congregata parasitoids: 1) Does IG content vary within and among populations of Catalpa trees, and is this variation associated with the presence of Catalpa Sphinx larvae? 2) Is the IG content of Catalpa Sphinx caterpillars correlated with the IG content of the trees on which they fed? 3) Are larvae of the parasitoid, Cotesia congregata, able to sequester IGs from their caterpillar hosts, and, if so, what is the relationship between IG concentrations in parasitoid larvae and caterpillar hosts?
Methods and Materials Study System Catalpa bignonioides, the Southern Catalpa, is a widespread deciduous tree found throughout North America as an ornamental (Sibley, 2009). This species contains primarily two IGs, catalpol and catalposide, although a few others have been recorded (von Poser et al., 2000). The Catalpa Sphinx, Ceratomia catalpae, is a specialist on Catalpa spp. (Baerg, 1935; Nayar and Fraenkel, 1963; Bowers, 2003); larvae are gregarious and brightly colored, black, yellow and white, and sequester high concentrations of the IG catalpol from their host trees (Bowers, 2003). Adults are cryptically colored, grayishbrown; they do not contain IGs and are palatable to birds (Bowers and Farley, 1990). A major natural enemy of Catalpa Sphinx larvae is the gregarious koinobiont parasitoid Cotesia congregata, which parasitizes larvae of a number of sphingid species, including Catalpa Sphinx (Harwood et al., 2002; Bowers, 2003). Cotesia congregata oviposit into 2nd-4th instar sphingid larvae and develop over ~2 wk, whereupon they emerge through the cuticle, spin silken cocoons, and pupate; adults emerge approximately 10 d later. Concentrations of catalpol in Catalpa Sphinx hemolymph, the sole source of nutrients for developing C. congregata larvae, can reach up to 50% dry weight (Bowers, 2003), and thus parasitoids developing in the hemocoel will be exposed to very high levels of IGs. Catalpol in the caterpillar host appears to have little negative impact on C. congregata fitness when reared from Catalpa Sphinx (Crocker, 2008; Lampert et al., 2010) despite evidence that catalpol is absorbed by developing parasitoid larvae (Bowers, 2003).
J Chem Ecol (2011) 37:1063–1070
1065
Leaf and Caterpillar Collections Leaves were collected 24–30 July 2008, from 14 separate Catalpa bignonioides stands at locations from New Jersey to Georgia (Table 1); trees in five of these stands were attacked by Catalpa Sphinx larvae. Trees in each population were sampled according to the same protocol. Approximately 10 leaves were removed from a single branch of caterpillar-free trees. To compare damaged and undamaged leaves in trees infested with Catalpa Sphinx larvae, approximately 10 leaves were collected from an undamaged branch and 10 from a damaged branch on each tree. At least 5 and up to 30 trees were sampled per site in this manner, depending on the number of trees in a stand. All leaves were dried for 2 d in an oven at 50°C, after which they were ground into fine powder using a coffee grinder. In trees with larvae, all Catalpa Sphinx larvae within reach (~3.0 m) were removed. Larvae ranged from 2nd to 5th instar depending on the site. Larvae were stored alive in a plastic cooler until they could be processed in the laboratory in Boulder, CO, USA, where 5th instars (as determined by head capsule size) were immediately weighed to the nearest 0.1 mg and then frozen at −80°C for chemical analysis. Remaining larvae were used to collect parasitoids or to establish a laboratory colony. Parasitoid Collections To determine the IG concentrations of parasitoid larvae and their caterpillar hosts, a separate set of Catalpa Sphinx larvae and larval Cotesia congregata were analyzed. Catalpol was measured in both caterpillar hemolymph and parasitoid larvae. For these data, Catalpa Sphinx larvae from a heavily parasitized population were collected in August 2008, from a mixture of catalpa trees at the USDA South Farm in Beltsville, MD, USA; the iridoid glycoside content of these trees was not measured. Tree species included Catalpa ovata, C. x galleana (a hybrid Table 1 Locations from which Catalpa bignonioides leaves were collected, dates when leaves were collected during the summer of 2008, the number of trees sampled and the number of trees attacked by Catalpa Sphinx caterpillars in each population
Population
between C. ovata and C. speciosa), and C. x erubescens (hybrid of C. ovata x C. bignonioides). Twenty-four late 4th instars and 9 early 5th instars were included in chemical analyses. Larvae were anesthetized with CO2, the abdominal horns were clipped, and 5 μl of hemolymph drained from these wounds were collected using microcapillary tubes and immediately placed in 5 ml methanol to extract catalpol. Caterpillars were dissected completely lengthwise and examined for Cotesia congregata larvae, which were removed, counted, washed with clean methanol, and dried with a paper towel. All members of a single C. congregata brood were pooled and weighed to the nearest 0.01 mg. Cotesia congregata larvae then were placed directly into methanol for catalpol extraction. Tree Chemistry and Analyses A subsample of 25 mg of ground leaves was used for iridoid glycoside analysis. Sample preparation and analysis by gas chromatography (GC) followed previously described methods (Bowers and Stamp, 1993; Bowers, 2003). Briefly, samples were extracted overnight in methanol, filtered and evaporated to dryness, partitioned between water and ether to remove nonpolar compounds, and the water layer (containing the IGs) was evaporated. Phenyl-ß-D-glucopyranoside (PBG) was added as an internal standard. Standards composed of purified PBG, catalpol, and catalposide were used for instrument calibration. For GC analysis, iridoid glycosides were derivatized using Tri-sil Z™ (Pierce Chemical Company). GC analyses were performed on an Agilent 7890 system (Agilent Technologies) equipped with an Agilent DB-1 column (30 m, 0.320 mm, 0.25 μm particle size), and data were processed with Agilent ChemStations software (version B-03-01). Paired t-tests were used to compare catalpol and catalposide content between damaged and undamaged leaves of caterpillar-infested trees. A
Latitude
Longitude
Date
Trees present
Trees attacked
Cecil Co., MD
39°41'38.76″
76°06′48.72″
25 July
6
4
Adams Co., PA Frederick Co., MD Franklin Co., GA Madison Co., GA Clarke Co., GA Johnston Co., NC Botetourt Co., VA Cumberland Co., VA Caroline, Co., MD New Castle Co., DE Cumberland Co., NJ Cape May Co., NJ Ocean Co., NJ
39°46′51.50″ 39°37′24.18″ 34°17′20.46″ 34°07′52.98″ 33°52′59.40″ 35°41′22.74″ 37°37′11.04″ 37°42′44.38″ 38°42′27.78″ 39°24′20.22″ 39°23′58.50″ 39°14′25.01″ 40°00′38.64″
77°15′37.63″ 77°25′19.38″ 83°05′14.82″ 83°12′57.60″ 83°25′12.06″ 78°23′55.80″ 79°58′58.14″ 78°10′43.14″ 75°54′14.46″ 75°33′18.30″ 75°04′13.98″ 74°49′44.56″ 74°12′58.74″
25 25 26 26 26 27 27 28 29 29 29 30 30
15 13 ~25 7 7 15 8 19 7 15 8 15 12
0 3 0 0 5 0 3 12 0 0 0 0 0
July July July July July July July July July July July July July
1066
random effects MANOVA was utilized to test for variation in catalpol and catalposide across populations, and profile analysis allowed for a comparison of variation of catapol vs. catalposide across populations. To examine the relationships between presence of caterpillars and leaf content of catalpol and catalposide in sampled trees, we utilized MANOVA and profile analysis, with population means of catalpol and catalposide as response variables and caterpillar presence as a fixed independent variable with three levels: trees from caterpillar-free populations, caterpillarinfested trees, and caterpillar-free trees from populations with caterpillars present. Residuals for both dependent variables were not significantly different from normal (Shapiro-Wilk test for normality, P>0.3 in either case). To avoid pseudoreplication, population means were utilized for these three levels of the independent variable. All linear contrasts between levels of independent variables were examined for the overall MANOVA model, and profile analysis tested the hypothesis that caterpillar presence affected catalpol differently from catalposide. Caterpillar Chemical Analyses Fifth instar Catalpa Sphinx were removed from the freezer and immediately ground in methanol, with sand, to help macerate the insects. Iridoid glycoside extraction and analysis by GC were performed as described above. Fresh-dry weight conversions were obtained by weighing, drying, and re-weighing separate sets of larvae. The relationship between chemistry of leaves of individual trees and the catalpol sequestration by Catalpa Sphinx larvae was examined by using least-squares regression. Sequestration was compared among populations using ANOVA, with the percent dry mass of catalpol as the response variable. Because natural enemies such as vertebrate predators may consume an entire insect before finding it unpalatable, we also compared total catalpol amounts sequestered among populations by using analysis of covariance (ANCOVA), treating caterpillar dry weight as a continuous covariate and the quantity of catalpol (in milligrams) as the response variable. All statistical analyses were performed with SPSS and SAS 9.2. Caterpillar Hemolymph and Parasitoid Chemical Analyses Capillary tubes of Catalpa Sphinx hemolymph were extracted for 72 h and C. congregata larvae were macerated with sand and extracted in 5 ml methanol for 48 h. Sample preparation and GC analysis were as described above. Because all C. congregata larvae from a single brood were ground and homogenized together, amounts of catalpol represented that absorbed by the entire parasitoid brood. Independent 2-tailed t-tests were performed to determine whether hemolymph catalpol content, parasitoid clutch size (number of larvae inside the host), parasitoid clutch dry mass, and parasitoid catalpol concentration (% dry mass)
J Chem Ecol (2011) 37:1063–1070
differed between 4th and 5th instar host larvae. Leastsquares regression analyses were performed to determine if catalpol concentration in host hemolymph significantly influenced both the amount and concentration of catalpol in parasitoid larvae. Because all individuals from each brood were combined, we also performed multiple regressions to determine whether number and weight of parasitoid larvae and caterpillar catalpol sequestration influenced the amount and concentration of catalpol detected inside parasitoid larvae.
Results Tree Chemistry Leaf iridoid glycoside content varied from
