Cascading effects of host plant inbreeding on ... - Wiley Online Library

Functional Ecology 2015, 29, 328–337

doi: 10.1111/1365-2435.12358

Cascading effects of host plant inbreeding on the larval growth, muscle molecular composition, and flight capacity of an adult herbivorous insect Scott L. Portman*,1, Rupesh R. Kariyat2, Michelle A. Johnston1, Andrew G. Stephenson1 and James H. Marden1 1

Department of Biology, The Pennsylvania State University, 208 Mueller Laboratory, University Park, Pennsylvania 16802, USA; and 2Department of Environmental Systems Science, ETH Zürich8092, Zürich, Switzerland

Summary 1. A primary function of adult winged insects is dispersal. Limiting larval dietary intake (partial starvation) has been shown to affect the flight muscle metabolism of adult moths reared on artificial diet, but a more ecologically relevant question is whether natural variation in host plant quality can lead to differences in the flight capacity of adult insects. 2. Recent studies have shown that inbreeding compromises plant antiherbivore defences. We created inbred and outbred progeny from locally collected horsenettle (Solanum carolinense L.) and examined how host plant inbreeding affects the growth, development, and flight muscle physiology of tobacco hornworm (Manduca sexta L.), a specialist herbivore on Solanaceae. We tested the hypothesis that within population genetic variation in host plant quality, resulting from inbreeding, can create significant changes to the larval development and flight physiology of an adult insect. 3. We found that Manduca larvae reared on inbred horsenettle plants grew faster and developed into larger pupae compared to larvae reared on outbred plants. Adult flight metabolic rate was greater in adults reared on inbred plants compared to outbred plants, and this elevation was independent of body mass when we excluded one plant family that produced small, low metabolic rate moths regardless of breeding regime. Differences in mass-specific flight metabolism were associated with changes in alternative splicing of troponin t, a flight muscle protein that regulates muscle contraction. 4. These results show that host plant inbreeding can create effects that cascade through larval and pupal development to affect dispersal-related traits of the adult stage. Hence, plant inbreeding may also impact herbivore population dynamics, particularly their ability to spread away from, and possibly into, isolated patches of inbred plants creating increased herbivore pressure on these plant populations. More generally, our findings reveal that changes in population biology at one trophic level can affect the metabolic physiology and flight capacity of an animal at a higher trophic level. Key-words: dispersal, flight, gene expression, horsenettle, inbreeding, Manduca sexta, metabolism, Solanum carolinense, Troponin t, tobacco hornworm Introduction The ability to fly has contributed immensely to the evolutionary success of insects (Roff and Fairbairn 1991). Insects use flight to evade predators (Chai & Srygley 1990), locate mates (Langellotto & Denno 2001) and colonize new habitat (Haag et al. 2005); thus, improved flight capability has the potential to increase survival, *Correspondence author. E-mail: [email protected]

reproductive success and geographical distribution. Flight capacity strongly correlates with the size and power output of flight muscles (Marden 1987; Hill, Thomas & Blakeley 1999; Berwaerts, Van Dyck & Aerts 2002); yet, large flight muscles are metabolically costly to build and maintain (Zera & Denno 1997), even during rest (Zera, Potts & Kobus 1998). In addition, physiological conditions that affect gut function, such as parasite infections, result in altered flight muscle development and decreased performance (Marden & Cobb 2004; Schilder & Marden 2007).

© 2014 The Authors. Functional Ecology © 2014 British Ecological Society

Inbred host plants promote enhanced insect growth and flight capacity 329 Changes in the pre-mRNA splicing of troponin t (Tnt), a contractile regulatory protein, are associated with variation in insect flight muscle performance. Quantitative variation in the relative abundance of Tnt spliceforms correlates with differences in flight muscle power output in dragonflies (Marden et al. 2001) and flight metabolic rate in Lepidoptera (Marden et al. 2008), and a large body of experimental work has shown that different Tnt spliceforms affect the calcium sensitivity of muscle activation and force output (e.g. Gomes et al. 2004). Larval nutrition also affects the relative abundance of the different Tnt spliceforms expressed in moth flight muscles (Marden et al. 2008). That experiment involved partial starvation of larvae during rearing on an artificial diet, so the results are difficult to extrapolate to more ecologically realistic variation. Considering that relatively small changes in peak flight muscle metabolic rate (approximately 15–20%) strongly affect colonization of new habitat patches (Haag et al. 2005) and movement of individual butterflies in the field (Niitep~ old et al. 2009), it is possible that variation in nutrient acquisition during immature stages cascades through altered gene expression to influence flight ability and dispersal. Plants defend themselves against herbivores with a variety of structural (Hanley et al. 2007) and chemical defences (Chen 2008). Plants producing a greater number of structural defences and/or high levels of toxic secondary metabolites have lower food quality because these defences reduce the capability of insect herbivores to acquire nutrients from the plant’s tissues. Reductions in nutrient intake rate have negative effects on the growth and development of herbivorous insects (Scriber 1981; Johnson et al. 1989; Boggs & Ross 1993; Leclaire & Brandl 1994; Haviola et al. 2007). Some studies have shown that certain co-evolved herbivores grew better on artificial diet when it contained a specific secondary plant metabolite (Bowers 1983, 1984). However, living host plants produce complex arrays of antiherbivore defences; hence, these results may not equate to insect performance on zoetic hosts. Inbreeding is common in flowering plants (Barrett & Eckert 1990) and frequently results in decreased fitness (inbreeding depression: Charlesworth & Charlesworth 1987; Husband & Schemske 1996). Inbred plants attract more insect herbivores and fewer predaceous insects, suffer greater levels of herbivory, and have greater exposure to herbivore-transmitted diseases than outbred plants (Stephenson et al. 2004; Bello-Bedoy & Nunez-Farfan 2011; Kariyat et al. 2012a). Moreover, herbivores consume more and grow faster when reared on inbred plants (Carr & Eubanks 2002; Delphia et al. 2009a). These studies suggest that plant inbreeding may have broad impacts on herbivore populations and community level processes in natural and agricultural systems, but the effects of host plant inbreeding on larval development and flight of subsequent adult insects have not been examined. A fuller understanding of how host plant inbreeding affects adult flight capability is important because adult mobility greatly

influences the geographical distribution and demographics of insect populations (Roff 1994; Hanski 2011). Since inbreeding negatively affects defence traits and renders plants more vulnerable to attack by insect herbivores, we hypothesized that Manduca sexta (Sphingidae) reared on inbred horsenettle (Solanum carolinense) plants would (i) exhibit increased larval growth rate and body mass; (ii) develop into larger adults; (iii) produce adults with higher-performing flight muscles; and (iv) flight muscle functionality would be accompanied by molecular changes in Tnt. We tested these hypotheses by measuring larval growth, adult body size, adult flight metabolism and the relative abundance of Tnt isoforms in the adult flight muscle from insects reared on paired genets of both inbred and outbred horsenettle plants.

Materials and methods THE STUDY SYSTEM

Solanum carolinense is a perennial weed, native to the eastern United States; it grows in crop fields, pastures and early successional habitats. After initial establishment, the plant spreads via horizontal (rhizome-like) roots that extend up to 1 m from the parent stem. Roots over-winter and produce new ramets in the late spring (Ilnicki et al. 1962). Although uncommon in weeds, horsenettle exhibits a typical Solanaceous type, ribonuclease-mediated gametophytic self-incompatibility (GSI) system controlled by the multiallelic S-locus (Richman et al. 1995). However, the GSI system in S. carolinense is leaky, being influenced by flower age, prior fruit production (Stephenson et al. 2003; Travers, MenaAlı & Stephenson 2004) and the presence of certain S alleles (Mena-Ali & Stephenson 2007). This leakiness allows the plants to produce selfed seeds when cross pollen is scarce or when there are few S alleles in the population. Horsenettle exhibits a variety of antiherbivore defences such as constitutive and induced structural traits (spines and trichomes; Kariyat et al. 2013a). In addition, all parts of the plant contain toxic secondary compounds (e.g., glycoalkaloids), especially the fruits (Cipollini & Levey 1997). Despite these defences, many herbivores feed on the leaves, fruits, flowers or roots of horsenettle (Cipollini & Levey 1997; Wise 2007). Tobacco hornworm larvae (Manduca sexta L.; Sphingidae) have been found feeding on horsenettle in the area from which our laboratory populations were collected (Delphia et al. 2009a).

PLANT MATERIAL

The plant material used for this study had been previously (2002) derived from a natural population located near State College, PA, USA (40° 470 29″ N, 77° 510 31″ W). Sixteen plants were randomly selected and collected from this field site (maternal parents). Two ramets were propagated from each the 16 fieldcollected parent plants by taking 10 cm cuttings from the horizontal roots. Flowers produced on one ramet were cross-pollinated (outbred), while flowers from a second ramet were self-pollinated (inbred). Six selfed seeds (inbred) and six outcrossed seeds from each of the 16 maternal parents were collected and germinated in a greenhouse (see: Mena-Ali & Stephenson 2007). These plants have been maintained via root cuttings following cold treatments (4 °C) and used for subsequent greenhouse and field studies of the effects of inbreeding on plant growth and reproduction (Mena-Ali, Keser & Stephenson 2008; Kariyat et al. 2011). Field, greenhouse

© 2014 The Authors. Functional Ecology © 2014 British Ecological Society, Functional Ecology, 29, 328–337

330 S. L. Portman et al. and laboratory studies using ramets from subsets of these genets have shown that inbreeding affects (i) resistance to herbivores (Delphia et al. 2009a; Kariyat et al. 2012a,b) and pathogens (Kariyat et al. 2012c), (ii) volatile mediated defences and physical defences (Delphia et al. 2009b; Kariyat et al. 2012a, 2013a), (iii) herbivore feeding and oviposition (Delphia et al. 2009a; Kariyat et al. 2013b) and (iv) transcriptome-wide expression of defencerelated genes (Kariyat et al. 2012b). These studies consistently revealed that inbreeding compromises the resistance of horsenettle to natural enemies. The scope of our study was constrained by the availability of greenhouse space necessary for the propagation of plants and colocation of insect rearing cages; therefore, we chose to use inbred plants and outbred plants from three (designated B1, B3 and B4) of the original 16 maternal families. Cold treated (4 °C for 6–8 weeks) horizontal roots from the six different genets (three families 9 two breeding types) were cut into approximately 10-cm sections and allowed to sprout in flatbeds containing a peat-based potting soil (Pro-Mix, Premier Horticulture Inc., Quakertown, PA, USA). Sprouts were maintained in a greenhouse (16:8 L:D; 25:22 °C; 65% RH) and watered on alternate days. After 2 weeks, the sprouts were transplanted into 2-L pots and moved to an insect-free greenhouse to establish a nursery. Non-flowering 6- to 8-week-old plants, from each of the six genets, were separated into different rearing cages (71 cm 9 57 cm 9 66 cm, L 9 W 9 H). Each rearing cage contained 12 individual ramets from a single genet.

LARVAL REARING

Manduca sexta eggs from an unknown number of females (the suppliers characterized the number as ‘several’; Boyce Thompson Institute, Ithaca, NY, USA) were hatched in petri dishes (90 mm 9 15 mm; Becton Dickinson & Co., Lincoln Park, NJ, USA) on moist Whatmanâ filter paper in a growth chamber (16:8 L:D; 25°C; 65% RH). After most of the eggs hatched, 72 neonate larvae were moved to the greenhouse (16:8 L:D; 25:22°C; 65% RH). Single larvae were assigned randomly to the horsenettle plants. Twelve neonate larvae were placed in each of the six rearing cages (12 caterpillars/cage/genet). In order to keep track of individuals, each caterpillar was uniquely marked with a small dot of coloured paint (Crayola Inc., Easton, PA, USA). Marks were reapplied after each larval moult.

INSECT MEASUREMENTS

To monitor changes in larval growth over time, body mass measurements were taken for each caterpillar at two discrete time points during each instar (early instar and late instar). Late instar mass measurements were taken during the quiescent period prior to moulting known as ‘moult sleep’ (Reinecke, Buckner & Grugel 1980); early instar measurements were taken on the day following moulting. Larval body mass measurements began when 1st instar larvae reached their late instar time point (moult sleep). Pupae were weighed, and their length and thoracic width were measured with a digital calliper (Mitutoyo American Corp., Aurora, IL, USA). Thorax width was measured at the widest point of the thorax, which contains the developing flight muscles.

ADULT FLIGHT METABOLISM

After eclosing, adult males and females were housed in separate growth chambers (24:0 L:D; 25 °C; 65% RH) and held in constant light to suppress flight activity. This helped to prevent wing damage and motivated them to begin flying shortly after exposure to low-light conditions. Adults were fed, every 2nd day, 200 lL of a solution that mimicked the sugar content of floral nectar (Baker

& Baker 1983), made from lemon-lime Gatoradeâ, sucrose (Gatorade + added sucrose = 054 M sucrose), fructose (025 M) and glucose (026 M). Adults were not offered food on days they were used for the flight metabolic assay. The number of adults available for the flight metabolism assay was significantly reduced because 9 pupae died, 3 adults eclosed with deformed wings, 3 more damaged their wings trying to escape confinement and 1 adult died prematurely. Moths with deformed or damaged wings were excluded from subsequent flight assays. Five to seven days after eclosing, adults were flown for 5 min in a 10-L glass jar attached to a positive pressure (push system), flow-through respirometer. Flight assays required multiple days to complete; each day, both inbred fed and outbred fed adults were selected for testing to avoid sampling bias. Data reported here are the initial flight assay for each moth. A subsequent flight assay was performed for a subset of moths after 24 hrs of carrying an attached load that added 33% additional mass to the moth’s body weight (data not shown). Immediately after completion of the final flight assay, the weights were removed for a 2–3 day recovery period- after which the moths were flash frozen for RNA isolation (see below). Incurrent CO2-free dry air came from a FT-IR purge gas generator (Whatman International Ltd., Maidstone, UK) at a flow rate measured by a mass flow meter (mean = 155 L min1) at air temperatures ranging from 24 to 29°C. The jar was completely flushed of atmospheric CO2 before the moth was stimulated to fly; after which the jar was shaken gently as needed to stimulate continuous flight. After the 5 min flight period, stimulation ceased and the recording continued until the CO2 signal returned to baseline (i.e. resting metabolic rate). Subsamples of the excurrent air (100 mL min1) were passed through a KClO4 filter to remove water vapour, and then through a LiCor 6262 infrared gas analyser calibrated with a certified span gas of known CO2 concentration. Analogue voltage outputs of the mass flow meter and gas analyser were converted to digital using a Sable Systems UI2 and recorded on a computer using Expedata 1.0.3 software (Sable Systems International, Las Vegas, NV, USA). Voltage data were imported into Igor Pro 4.02 (Wavemetrics Inc., Portland, OR, USA) and converted to ml CO2 using a multinomial function for the gas analyser provided by the manufacturer. We subtracted the baseline (resting) CO2 production rate and used a Z-transformation (Bartholomew, Vleck & Vleck 1981) to remove temporal autocorrelation and estimate the instantaneous metabolic cost of flight. From the Z-transformed data, we determined the peak metabolic rate and, by integration, the total CO2 emitted during 5 min of flight.

TNT ISOFORM PROFILING OF MANDUCA SEXTA

Immediately after flight, total adult mass was measured. After the final flight assay, whole adults were flash-frozen in liquid nitrogen and stored at –80°C. Subsequently, whole thoraxes were separated and weighed from the frozen insect. An approximately 02-g section of frozen dorsal longitudinal flight muscle was dissected on dry ice and used for RNA isolation and characterization of Tnt isoforms (see: Marden et al. 2008). RNA isolations for five inbred fed individuals failed, and the reactions were not repeated because the muscle tissue samples had degraded after freeze-thawing. Fulllength Tnt cDNA sequence from fall armyworm (Spodoptera frugiperda J.E. Smith) and from M. sexta transcriptome data (NCBI SRA: PRJNA79369) were blasted against the M. sexta genome (Agricultural Pest Genomics Resources, http://agripestbase.org) to locate Tnt and identify exons. Nucleotide sequences flanking the 50 alternatively spliced exon region of Manduca Tnt were nearly identical to the corresponding sequence for which we already possessed fluorescently labelled primers for fall armyworm Tnt (TntAltF 5-56FAM-CACCCG TGCG AC -ATTAAATAAAC-3, TntAltR 5-GCGCCATTCGTTGATGTATTC-3). These primers


Inbred host plants promote enhanced insect growth and flight capacity 331 hybridized with constitutively spliced regions on both sides of the 50 alternatively spliced region and successfully amplified the target region of Tnt from Manduca cDNA. The PCR products were sequenced and confirmed to align to the exons of Manduca Tnt evident in genomic sequence. STATISTICAL ANALYSIS

To examine larval growth and development time across all larval instars (2nd–5th), including measurements taken both early and late within instars, the data were first log10 transformed to achieve normality. Repeated-measures ANOVA models that nested individual larvae (random factor) within plant maternal family and plant breeding type were used to compare the growth and development rate of larvae that fed on inbred vs. outbred host plants. ANOVA models included maternal family (host plant), breeding type (host plant), individual (insect), instar (insect), time point (insect), maternal family*breeding type and instar*breeding type as predictor variables. Post hoc comparisons (Tukey’s honest significant difference) were carried out for significant interaction terms. Analysis of variance was also used to compare pupae mass, adult body mass and adult thorax mass for insects that developed on inbred vs. outbred host plants. ANOVA models for pupae and adult body size comparisons included maternal family, breeding type, gender (insect) and maternal family*breeding type as predictor variables. Data reported for Tnt isoforms are based on relative abundance values calculated by dividing individual isoform peak heights by the sum of all isoform peak heights present in that insect. Isoform relative abundances were arcsine transformed to achieve normality. Analysis of variance was used to compare Tnt isoform relative abundance differences in adult moths that developed on inbred vs. outbred host plants. All Tnt ANOVA models included breeding type and gender as predictor variables, but not body mass (which in contrast to Fall Armyworm moths showed no relationship with Tnt isoform abundances; Marden et al. 2008). Many factors can affect an insect’s flight metabolic rate; thus, we used a stepwise ANCOVA model to examine the variation in adult flight metabolism. Predictor variables originally included in the stepwise procedure were maternal family, breeding type, gender, body mass, thorax mass, relative abundances of Tnt isoforms A, B, C, D, E, F and maternal family*breeding type interaction,. The stepwise procedure for adult peak flight metabolic rate resulted in an ANCOVA model that included only adult body mass and Tnt E relative abundance as significant predictors. When plant maternal family B1 was excluded from the analysis, the stepwise procedure resulted in an ANCOVA model that included maternal family, breeding type, body mass and the relative abundance of Tnt E as significant predictor variables (P ≤ 005). All analyses were carried out in JMP v. 10 (SAS Institute, Cary, NC, USA).

fed larvae. Post hoc comparisons of larval development time by instar and breeding showed that this acceleration in development occurred during the 1st, 3rd and 5th instars (P < 00001, P = 0002, P = 0001; Fig. 1). Growth rates (g h1) across all instars (ANOVA, P < 00001, R2 = 095; log-transformed data; Table S2) showed overall effects from host plant breeding type (P = 0011) instar (P < 00001) and time point (P < 00001). On average, the growth rate of inbred fed larvae was 92% higher than outbred fed larvae. This is consistent with the shorter development times recorded in the inbred fed larvae. Differences in larval body mass for early instars (ANOVA P < 00001, R2 = 096; log-transformed data; see Table S3) and late instars (ANOVA P < 00001, R2 = 099; log-transformed data; Table S4) showed no independent effects from host plant breeding (P = 028, P = 017), and hence, it appears that inbred fed larvae reached similar critical masses (Reinecke, Buckner & Grugel 1980) more rapidly. PUPAE AND ADULT BODY SIZE

Body mass of pupae showed significant effects from breeding type (P < 00001), but not maternal family, gender or maternal family*breeding type interaction (ANOVA P < 00001, R2 = 035; Fig. 2a). Pupae that developed from inbred fed larvae (N = 32) had a 203% higher mean body mass than outbred fed pupae (N = 27). Considering that late instar larval body mass showed no host plant effects, the effects of host plant breeding type on pupal mass was a surprising result - due perhaps to differences in larval water content and/or ability to synthesize adult tissue during metamorphosis. Similar host plant breeding type effects were also found for differences in pupal length (ANOVA P = 0006, R2 = 013) and diameter (ANOVA P < 00001, R2 = 035). Inbred fed pupae were 48% longer and 84% wider than outbred fed pupae (Fig. S1 in Supporting information).

Results LARVAL GROWTH

Over the course of larval development, larvae reared on inbred host plants exhibited more rapid growth. Development time (hrs) across all instars (ANOVA P < 00001, R2 = 097; log-transformed data; Table S1 in Supporting information) showed significant effects of breeding type (P < 00001), maternal family (P < 00001), instar (P < 00001), time point (P < 00001), along with maternal family*breeding type (P = 0022) and instar*breeding type (P < 00001) interactions. On average, inbred fed larvae reached the 5th instar 154 h (65%) earlier than outbred

Fig. 1. Average (mean SE) development times for Manduca sexta larvae reared on inbred vs. outbred horsenettle plants. Bars represent average development times for each larval instar. Asterisks designate significant differences in average values for inbred fed vs. outbred fed larvae (P ≤ 005, Tukey’s HSD).


332 S. L. Portman et al. (a)

(b)

Fig. 2. Average (mean SE) pupal mass (a) and adult body mass (b) of Manduca sexta adults reared on inbred vs. outbred horsenettle plants from three maternal plant families (B1, B3, B4). Numbers in bars indicate sample sizes. Asterisks designate significant differences in average values of inbred fed vs. outbred fed moths (Bonferroni cut-off: P ≤ 0017).

Differences in pupae size carried over to the adult stage. Adult body mass differed (ANOVA P = 00004, R2 = 044; Table S5) according to breeding type (P = 0018), maternal family (P = 002) and gender (P = 0008). On average, females (N = 21) were heavier than males (N = 19), and inbred fed moths (N = 24) were 384% heavier than outbred fed moths (N = 16; Fig. 2b). Thorax mass (which correlates closely with flight muscle mass: Marden 1987; Srygley & Chai 1990) varied with body mass (P = 0006) and gender (P = 001), but did not show independent effects of plant breeding type or maternal family (ANCOVA P = 0009, R2 = 023), suggesting that larval diet had no significant effects on flight muscle mass outside of the effects on overall body size. ALTERNATIVE SPLICING OF TNT IN FLIGHT MUSCLE

To further examine the effects of plant breeding on flight and muscle physiology, we quantified the relative abundances of alternatively spliced transcripts of Tnt. Consistent with previous work on this gene in Lepidoptera (Marden et al. 2008), we found 6 Tnt fragment sizes corresponding to the known spliceforms (A–F) of Tnt (Fig. S2). The relative abundance of isoform E was 258% lower in inbred fed moths (ANOVA P = 0011, R2 = 023 breeding type P = 0049; gender P = 0008; Table 1), but no independent significant differences in relative abundances were found for the other 5 Tnt isoforms (Fig. S3). These results indicate that host plant Table 1. ANOVA table for relative abundance of Tnt E, from adult flight muscle, showing effects of plant breeding type and gender of the insect. Abundance data for Tnt were arcsine transformed to achieve data normality. Significant factors are shown in bold type ANOVA

Relative Abundance of Tnt E

Source

d.f.

Sum of Sqs

Mean Sq

Model Error Total Plant breeding Insect gender

2 34 36 1

0017 0054 0071 0007

0008 0002

1

0013

F Ratio

Prob > F

R2

516

001

023

417

005

788

001

inbreeding affects this phenotypically plastic feature of the molecular composition of adult insect flight muscles. ADULT FLIGHT METABOLISM

On average, inbred fed moths had a 165% higher peak metabolic rate (P = 0014) and 146% higher total metabolic output (P = 0029) than outbred fed moths (Fig. 3). Total CO2 output was tightly correlated with peak metabolic rate (r2 = 091; compare Fig 3a and b) because the moths maintained a consistent level of flight effort throughout the 5 min assay. Since these two variables are so tightly related, we present further only the results for peak flight metabolic rate. Host plant effects on insect development and body size can potentially influence flight metabolic rate in a number of ways, either as direct negative effects of diet composition (i.e. toxins), correlated effects of size (the benefit of a larger thorax and/or the energetic cost of lifting a heavier body), or via developmental changes in the molecular composition of the flight muscles (e.g. alternative splicing of Tnt that affect cross-bridge activation and power output). For this reason, a stepwise model was appropriate to determine which set of variables most affected peak metabolic rate during flight. The stepwise model showed that adult body mass (P < 00001) and the relative abundance of Tnt E (P = 003) had significant effects on peak metabolic rate (ANCOVA P < 00001, R2 = 044; Table 2). Body mass had a positive effect, which may reflect beneficial physiological effects of feeding on inbred plants and/or the increased metabolic cost of lifting a heavier load (Fig. 4a). Excluding plant family B1, which produced moths with uniformly low metabolic rates (Fig. 3), moths reared on inbred plants had a 154% increase in their body mass-adjusted average peak metabolic rate (ANCOVA P = 00001, R2 = 057; Fig. 4b). Tnt E had a negative effect on body mass-adjusted flight metabolism (P = 003, R2 = 014), which is consistent with the lower relative abundance of this isoform in flight muscles of moths reared on inbred plants (Fig. 5).

Discussion The ability to fly allows insects to escape predation (Chai & Srygley 1990), locate mates (Langellotto & Denno


Inbred host plants promote enhanced insect growth and flight capacity 333 (a)

(b)

Fig. 3. Boxplots of peak flight metabolic rate (a) and total CO2 emitted (b) during 5 mins of flight by Manduca sexta adult moths reared on inbred (N = 25) compared to adults reared on outbred (N = 17) horsenettle plants from three maternal plant families (B1, B3, B4).

Table 2. ANOVA table for peak flight metabolic rate showing effects of body mass and abundance of Tnt isoform E. Significant factors are shown in bold type ANOVA

Peak Flight Metabolic Rate

Source

d.f.

Sum of Sqs

Mean Sq

Model Error Total Body mass (g) Abundance Tnt E

2 33 35 1

833498 1073775 190273 725113

416749 32539

1

160918

F Ratio

Prob > F

R2

1281