Introduction

Plants are under constant attack by insects and often such insects co-occur on a host plant. Co-occuring phytophagous insects can compete among each other for the plant’s resources (Ali and Agrawal 2014; Denno et al. 1995). Several studies have shown that co-existing insect species on a plant show asymmetry in their performance and survival (Ali and Agrawal 2014; Davidson-Lowe et al. 2019; Pineda et al. 2017). This asymmetry is often due to differential plant responses to herbivores that differ in their mode of feeding. Insect herbivores can be broadly categorized into groups based on their feeding behavior, i.e., chewing insects with mandibular mouth parts that defoliate plants and insects with piercing sucking mouth parts that feed on plant phloem (Howe and Jander 2008). Although interspecific competition between phytophagous insects may primarily be among insects with similar feeding habits, such competition may also occur among insects from different feeding guilds (Denno et al. 1995).

Insect herbivores from different feeding guilds are also known to elicit antagonistic defense pathways in plants (Rodriguez-Saona et al. 2010; Thaler et al. 2012). Chewing insects, such as caterpillars, generally trigger jasmonic acid (JA) induced plant defenses, while aphids with piercing/sucking mouth parts predominantly induce salicylic acid (SA) mediated defenses in plants (Howe and Jander 2008; Thaler et al. 2012; Walling 2000). Some insect-related cues can potentially alter host plant chemistry making it conducive towards the insect’s survival, thereby outcompeting the other species in the process (Ali and Agrawal 2014; Davidson-Lowe, Szendrei & Ali,., 2019). An insect can induce the synthesis of plant volatiles that repel other potential herbivores, reducing competition or alter plant metabolites that reduce the performance of the secondary herbivore (Pineda et al. 2017; Soler et al. 2012). The performance of oleander aphids (Aphis nerii) was worse on common milkweed (Ascelpias syriaca) that were already infested with monarch caterpillars (Danaus plexippus) (Ali and Agrawal 2014). Alternately, monarch caterpillars gained more weight on common milkweed plants that were already infested with oleander aphids. In another study, herbivory by Colorado potato beetles (Leptinotarsa decemlineata) reduced the performance of green peach aphids (Myzus persicae) in potatoes (Davidson-Lowe et al. 2019). Therefore, it is evident that multiple or successive feeding by insects from different feeding guilds on the same plant shows asymmetry in performance of the herbivores involved. However, what is less understood is how elicitors and effectors contribute to such interactions.

In order to defend themselves, plants rely on a diverse array of chemical cues from herbivores and neighboring plants to mount a response (Acevedo et al. 2015; Bonaventure 2014; Erb et al. 2015; Markovic et al. 2018). Chemical cues that induce plant defensive compounds to confer resistance to the invading insect are known as elicitors (Felton and Tumlinson 2008). Elicitors present in insect oral secretion, saliva or oviposition fluids may trigger plants to synthesize proteins and metabolites that are directly detrimental to the insect herbivore (Acevedo et al. 2015; Felton and Tumlinson 2008; Hilker and Meiners 2006), or induce synthesis of volatile compounds in plants that attract natural enemies of the invading herbivore and may even prime neighboring plants of imminent attack (Engelberth et al. 2004; Erb et al. 2015). Conversely, chemical cues from insect secretions that suppress plant-induced defenses towards the herbivore are known as effectors and have been shown to be present in insect feces and oral secretions (Musser et al. 2002; Ray et al. 2015; Acevedo et al. 2017).

Chemical cues in insect secretions and excretions that alter plant defenses not only influence the survival and performance of the insect herbivore that deposits such cues, but may also affect the performance and preference of subsequent or co-existing herbivores on the plant by manipulating host chemistry. Oral secretion from several caterpillar species aboveground was shown to alter preference of rootworm larva belowground (Lu et al. 2016) indicating that cues from oral secretion of one chewing herbivore can stably influence the preference of another chewing herbivore that feed on two different plant organs. Elicitors associated with insect feeding can induce the synthesis of plant volatiles that repel other herbivores, including those from a different feeding guild to reduce competition (Pineda et al. 2017; Soler et al. 2012). We are only beginning to understand the effect of elicitor/effector mediated alteration of plant defenses by one insect and its effect on herbivory by another insect species. The alteration of plant chemistry by herbivore related cues from one guild of insect such as caterpillars may symmetrically increase the performance of a subsequent herbivore such as aphids or vice versa (Fig. 1). In a previous study, we have shown that plant chitinases Pr4 and EndochitinaseA (ChiA) in fall armyworm (FAW) frass can suppress herbivore-induced defenses in maize (Zea mays) (Ray et al. 2015, 2016). This led to increased performance of the caterpillar larvae on the plant over time. Given that plant response to caterpillar frass effectors lead to suppressed defenses that improved caterpillar performance, we hypothesized that insect herbivores from a separate feeding guild might show differential performance and preference on plants induced by frass effectors compared to non-induced plants (Fig. 1). Caterpillar frass mediated alteration in plant chemistry may be symmetric to herbivore performances of both chewing and piercing/sucking herbivores (Fig. 1 B1, C1) or their performances may be asymmetric (Fig. 1 B2, C2). In this study we investigated if caterpillar frass and frass effector mediated suppression of plant defenses in maize can be utilized by herbivores from a different feeding guild. We measured performance of corn leaf aphids (CLA; Rhopalosiphum maidis) on maize plants post FAW herbivory and treatment with frass or the frass chitinases. Frass effectors may also play a role in altering plant volatile blend that in turn may affect the preference of subsequent herbivore on the plant. Therefore, we measured plant volatiles emitted from maize plants post fall armyworm feeding and frass chitinase treatment. In an attempt to understand if aphids preferred volatiles emitted from plants that were treated with caterpillar frass effectors over undamaged plants, we performed choice tests in y-tube olfactometers with alate aphids.

Fig. 1
figure 1

Schematic representation showing the effect of caterpillar frass effector-altered plant chemistry on the performance of a subsequent group of herbivore, corn leaf aphids

Full size image

Materials and Methods

Plant and Insect Materials

Maize plants (inbred B73) were grown in Hagerstown loam in glasshouse conditions with 16 h of light and 8 h of dark cycle. When plants reached V3 stage (plants with three fully open leaves), they were used to conduct aphid performance and preference assays, and for volatile collections. The corn leaf aphid colony was maintained using barley host plants. Apterous, adult aphids were transferred from barley to maize plants three days prior to experiments, allowing them time to acclimate and aphid nymphs of the same cohort were chosen for performance assays. FAW eggs were purchased from Benzon Research (Carlisle, PA) and reared on artificial diet. FAW larvae in their 4th instar were used to damage maize plants in inverted clips cages as previously described (Ray et al. 2015).

Aphid Performance Assays

To determine the influence of prior FAW damage and frass deposition on corn leaf aphid performance, we allowed aphids to develop on plants either damaged by FAW, treated with or without frass, or treated with frass chitinases. Two 4th instar FAW larva were allowed to feed on each maize plant for 12 h in sleeve cages where frass from the caterpillar larva came in contact with the damaged tissue. Similarly, two larvae of the same age were localized in inverted clip cages on maize leaves for 12 h where they fed but their frass did not come in contact with the damaged tissue (Ray et al. 2015). Fifteen plants each were used for caterpillar damage with frass, caterpillar damage without frass and undamaged control plants. Following the FAW damage treatments, five corn leaf aphids that were two days old were added to each plant. After 7 days, the number of aphids was recorded. Frass chitinases Pr4 and ChiA were the chemical cues in FAW frass that suppressed caterpillar defenses in maize (Ray et al. 2016). Therefore, to understand the role of these frass chitinases in directly altering plant chemistry that may affect aphid performance, we fed V3 stage maize plants to four 4th instar FAW larvae for 12 h following which 20 μg of Pr4, ChiA or an equivalent volume of buffer were applied to the leaves surrounding the caterpillar damage. Twenty-four hours after application of chitinases or buffer, five aphids were applied on each plant and performance of the aphids measured as described above. Similar to the previous assay, fifteen plants were used for each treatment. ChiA and Pr4 proteins were heterologously expressed in E.coli bacteria with pET102 plasmid and purified using affinity column chromatography as described earlier (Ray et al. 2016).

Aphid Preference Assay

To determine how prior FAW herbivory and exposure to frass chitinases influence corn leaf aphid preference, we recorded choices made by adult, alate corn leaf aphids in glass y-tube olfactometers. For behavioural assays, V3-stage maize plants were damaged by 4th instar FAW caterpillars in inverted clip cages until they consumed leaf tissue of 3 cm diameter (~2 h). Following feeding by caterpillars, plants were treated with ChiA or buffer to the damaged area as described above. Aphid choice tests were performed with undamaged plants versus damaged plants with ChiA or buffer. Plants were placed in isolated glass chambers and charcoal-filtered, humidified air (500 ml min−1) was blown over each plant and into the arm of a glass Y-tube olfactometer. A single alate aphid was introduced at the base of the tube. Aphid choice was recorded for no more than 30 mins per individual after which the assay was terminated. Five aphids were used for each plant set and 30 aphid choices were recorded for each treatment pair. The glass Y-tubes olfactometers were rinsed with acetone after each aphid choice and the air flow entering the olfactometer arms were switched between the two plant treatments after each aphid choice to prevent directional bias. A new pair of treated plants were used after 5 aphid choices were recorded.

Collection and Analysis of Plant Volatiles

Four 4th instar FAW caterpillars were allowed to feed on V3-stage maize plants in inverted clip cages till they consumed 3 cm diameter of leaf area (~2 h) as before. The plants were then treated with 20 μg of ChiA, Pr4 or equivalent volume of buffer on to the feeding site. Following the chitinase treatments on damaged plants, foliar volatiles were collected for 8 h under 180 μmol m−2 s−1 of light using dynamic headspace sampling. Plant leaves were enclosed in individual glass chambers (20 cm diameter and 30 cm height) resting on teflon guillotine-style bases. Activated charcoal purified air was pushed into each chamber at 2 L min−1 and air containing plant volatiles was pulled from each chamber at 1 L min−1, over volatile filter traps containing HayeSepQ (Sigma Aldrich, USA). Volatiles were collected from 8 plants per treatment. Leaf volatiles were also collected from plants that were artificially wounded and treated with ChiA, Pr4 and buffer. Each plant was mechanically wounded four times using methods described in (Ray et al. 2015) and 20 μg of frass chitinases (ChiA/Pr4) or buffer was applied to each wound site. Following the wounding and chitinase treatments, maize volatiles were collected from eight plants per treatment, following the methods described above for 8 h. Volatile filter traps were eluted with 150 μl of dichloromethane and a solution of nonyl acetate (80 ng μL−1) and n-octane (40 ng μL−1) was added to each sample (5 μL) as internal standards. Samples were analyzed using an Agilent 6890 gas chromatograph and 5973 mass spectrometer with a splitless injector held at 250 °C. After sample injection (1 μL), the column (Rxi®-1 ms, 30 m, 0.25 mm id, 0.25 μm film thickness; Restek, USA) was maintained at 40 °C for 2 min after which temperature was increased 10 °C at a time until it reached 190 °C and then 12 °C at a time until it reached 280 °C. Tentative identification of target compounds were made by comparison of mass spectra and retention times with published data (NIST14 mass spectral library). Compounds identified with >90% fidelity by ChemStation (Agilent, USA) were quantified relative to the nonyl acetate standard. Retention indices of the identified compounds were calculated and is reported in Table S2 (Guiochon 1964).

Statistical Analyses

Aphid survival data were normalized and transformed by box-cox transformation and one-way ANOVA was run in R Studio using Anova function (Boston, USA). Post-hoc analyses and means separation was performed with Tukey test at P < 0.05. For the aphid preference tests, chi-square analysis goodness of fit test was run with the null hypothesis that aphids would choose each treatment in 1:1 ratio at P < 0.05. Multivariate random forest analysis was performed on the volatile data that predicted the volatile compounds that may be most different among treatments (Ranganathan and Borges 2010). Volatile compounds that were predicted to be different among treatments by random forest analysis (Fig. S2) were further analyzed for differences by one-way ANOVA and means separations were performed using Tukey test at P < 0.05. The p-values and F statistics for these analyses are provided in supplemental Table 1.

Results

Plant Exposure to Caterpillar Frass Following Herbivory Reduced Performance of Corn Leaf Aphids

Our results indicate that corn leaf aphid performance was highest on plants that were damaged by caterpillars in inverted clip cages and had no exposure to caterpillar frass. Comparatively, aphid performance was nearly 50% lower on plants fed on by caterpillars in sleeve cages where caterpillar frass accumulated (P < 0.001, F = 12.38, df = 2) (Fig. 2a). Similar to these trends, aphid performance on undamaged plants was reduced by nearly 50% compared to those that were damaged by FAW larvae without frass accumulation(Fig. 2a).

Fig. 2
figure 2

Corn leaf aphid performance on maize plants treated with fall armyworm frass. Abundance of corn leaf aphids was measured on maize plants after they were fed on by fall armyworm for  12 h and whole frass (a), or frass proteins Pr4 and EndochitinaseA was applied (b). Aphid abundance was also measured on maize plants that were treated with frass proteins after artificial wounding (c). Means separation was done using Tukey test and different letters signify different aphid numbers at p < 0.05

Full size image

Plant Exposure to Chitinases from Caterpillar Frass Following Wounding Reduced Performance of Corn Leaf Aphids

Similar to our findings for plant exposure to frass, plants treated with frass effector ChiA showed lower aphid performance compared to FAW-damaged, buffer-treated controls (Fig. 2b). Following the 7-day period of introduction of aphids to plants, aphid number was reduced by two-thirds on ChiA-treated plants compared to plants damaged by FAW with no exposure to frass chitinase. Undamaged control plants had 50% fewer aphids compared to FAW-damaged, buffer-treated plants but had a similar number of aphids compared to plants treated with ChiA (P < 0.001, F = 8.38, df = 3). Plants treated with Pr4 had an intermediate number of aphids, but that number was not statistically different from the other treatments (Fig. 2b).

While feeding on plant tissues, lepidopteran larvae deposit oral secretions that contain a variety of elicitor and effector compounds that can alter plant defenses. To isolate the effect of frass chitinases on plant responses and their influence on aphid performance, we conducted a follow up experiment where we measured aphid performance on artificially wounded maize plants with or without frass chitinase application. Corn leaf aphid performance was reduced by 40–50% on plants treated with chitinases, ChiA or Pr4, and on undamaged plants compared to artificially wounded plants treated with buffer (P < 0.001, F = 10.75, df = 3) (Fig. 2c).

Caterpillar Frass Alters Emission of Volatile Organic Compounds in Maize Plants

In order to determine the influence of frass chitinases on herbivore-induced plant volatiles, we measured volatiles from plants with frass chitinases post FAW herbivory. We were able to identify and measure several plant volatiles, including several terpenes from both plants that were fed with FAW caterpillars in inverted cages without feces and from FAW-fed plants that were treated with frass chitinases after such damage. However, there were only two volatile compounds that were differentially emitted from these plants. Indole, a heterocyclic plant volatile that is implicated in plant-plant communication and priming as well as reduction of herbivore performance by intoxication (Erb et al. 2015; Veyrat et al. 2016), was higher in plants damaged by FAW larvae compared to undamaged plants. Plants treated with ChiA showed lower levels of indole compared to FAW-fed plants treated with buffer controls, but such decrement was not statistically significant (Fig. 3a). Another plant volatile differentially expressed in frass-treated plants compared to control plants was (E)-β-farnesene (EBF), a sesquiterpene that is synthesized by plants and is an alarm pheromone for aphids (Bhatia et al. 2015; Joachim and Weisser 2015). In our study, plants that were damaged by FAW and then treated with ChiA showed lower levels of EBF compared to FAW-fed, buffer-treated plants but again, this was not statistically significant (Fig. 3b).

Fig. 3
figure 3

Volatile organic compounds released from maize leaves damaged by fall armyworm caterpillars and treated with frass proteins Pr4 and EndochitinaseA. Plants were fed by caterpillar larva for  2h and volatiles were collected for 8 h during the day. Means separation was done using Tukey test and different letters signify differential emissions of plant volatiles at p < 0.05

Full size image

We also measured plant volatiles from plants that were treated with frass chitinases Pr4 and ChiA post artificial wounding in order to understand the isolated effects of frass effectors only on plant volatile emissions. We saw an increase in several plant volatiles such as indole, EBF, (E)-β-caryophyllene, linalool and acetic acid phenyl ester in response to ChiA treatment compared to the buffer-treated plants (Fig. 4a–f). Pr4 treatment on the other hand, failed to elicit plant volatiles that were higher than the buffer-treated controls in any of the compounds that were measured.

Fig. 4
figure 4

Volatile organic compounds released from maize leaves post artificial wounding and treatment of frass proteins Pr4 and EndochitinaseA. Plants were fed on by FAW and volatiles were collected for 8 h during the day. Means separation was done using Tukey test and different letters signify different levels of volatile emissions at p < 0.05

Full size image

Alate Corn Leaf Aphids Marginally Prefer Plants Treated with EndochitinaseA

Changes in the volatile emissions from frass chitinase ChiA-treated plants prompted us to investigate the preference of corn leaf aphids for plants treated with ChiA over plants that were buffer-treated or undamaged. Alate corn leaf aphids marginally preferred ChiA-treated plants over buffer-treated plants (α = 0.094, χ2 = 2.793, df = 1) post FAW herbivory. However, they showed no preference between ChiA-treated and undamaged plants (α = 0.715, χ2 = 0.133, df = 1) in y-tube olfactometer assays (Fig. 5).

Fig. 5
figure 5

Aphid preference to plants treated with EndochitinaseA post fall armyworm herbivory. Winged corn leaf aphid preference was observed in a Y-tube olfactometer with aphids subjected to odors emitted by an undamaged plant or a plant treated with buffer/EndochitinaseA post fall armyworm damage for 2 h. Asterisk shows significantly different choices as determined by chi-square goodness of fit test at p < 0.05.

Full size image

Discussion

Plants induce dynamic responses to herbivory from insects of different feeding guilds by inducing alternate pathways of plant defenses and when such insect herbivores are present together or in succession, they may alter plant chemistry to favor one insect species over the other (Ali and Agrawal 2014; Pineda et al. 2017; Davidson-Lowe et al. 2019). Although induction of plant defenses by elicitors and effectors in insect secretion and excretion are well documented, their role in interspecific interactions of insect herbivores on a host plant is poorly understood. Suppression of induced plant defenses by effector molecules in insect secretions and excretions are known to increase performance of the herbivore depositing such cues on the plant (Acevedo et al. 2015, 2018; Ray et al. 2016). However, the implications of such suppressed plant defenses in an ecological perspective can be more consequential whereby subsequent herbivores may hone in on a vulnerable plant. Our previous study demonstrated that FAW frass contains plant chitinases that act as effectors and suppress caterpillar induced defenses in maize (Ray et al. 2016). While frass and frass-related effectors were known to increase caterpillar performance by suppressing plant defenses, the possible implications of a secondary herbivore performance on the same plant was unknown.

In this study, we show that aphids performed best on plants damaged by FAW caterpillars that had no exposure to frass. In contrast, aphid populations were reduced by nearly 50% on FAW-damaged maize plants where caterpillar frass accumulated in whorls, compared to plants without frass accumulation (Fig. 2a). Such suppression of aphid performance may be due to the presence of accumulated frass itself on the plant, which may emit volatiles or harbor microbes and noxious chemicals that may be detrimental to aphid growth. To eliminate all such possibilities, we fed FAW caterpillars in cages so that frass could not accumulate on the plants, followed by application of frass chitinases ChiA and Pr4. We observed that maize plants treated with frass protein ChiA post FAW herbivory had one third of the number of aphids compared to plants that were treated with buffer. Taken together, these results suggest that caterpillar frass, which is known to alter plant defenses to benefit caterpillar performance, can also reduce performance of a subsequent herbivore from an alternative feeding guild. This may potentially be adaptive, in that subsequent herbivores may compete for host resources. We also link frass protein ChiA to this phenomenon, showing that this frass molecule can independently mimic the effects of whole caterpillar frass on plants reducing performance of subsequent aphid herbivory. We observed that aphids performed better on damaged plants compared to undamaged plants (Fig. 2a–c); however, caterpillar frass or frass effector ChiA can alter the chemistry of a damaged plant such that aphid performance decreases compared to a damaged plant without frass. These results directly supports the second scenario of our hypothesis (Fig. 1 B2,C2) where effector(s) from one insect herbivore alter plant defenses asymmetrically by enhancing the performance of one herbivore while reducing the performance of a subsequent herbivore. This not only adds clarity to our understanding of plant resistance involving multiple pest communities that may co-occur or sequentially occur on the plants, but also broadens our understanding of elicitor/effector-mediated plant defense responses and their role in plant insect interactions involving multiple species of insect herbivores.

Volatile organic compounds that are emitted from plants both above and belowground have been shown to alter host preference for insect herbivores and natural enemies of these insect pests (Ali and Agrawal 2017; Felton and Tumlinson 2008; Jiménez-Martínez et al. 2004; Walling 2000). Elicitors and effectors from insect oral secretion are widely known to alter plant volatile profiles. We observed that application of frass protein ChiA induced emission of indole post artificial wounding (Fig. 4c). However, indole production was suppressed in plants that were treated with ChiA post FAW herbivory compared to buffer treated controls. This suggests that elicitors from oral secretion may interact with frass effectors such as ChiA to suppress indole synthesis. Indole has been implicated in plant-plant communication and therefore ChiA mediated suppression of indole emission can be of larger ecological relevance (Erb et al. 2015).

Frass ChiA induced the emission of aphid alarm pheromone (E)-β-farnesene which is known to be involved in deterring aphids from settling on a plant (Gibson and Pickett 1983; Joachim and Weisser 2015). ChiA also induces the synthesis of several other herbivore-induced plant volatiles such as linalool, (E)-β-caryophyllene and acetic acid phenyl ester (Fig. 4). Induction of (E)-β-caryophyllene by ChiA, in particular has been shown to suppress the effect of (E)-β-farnesene and therefore may explain why aphids do not preferentially choose buffer-treated plants over ChiA-treated plants (Fig. 5) (Dawson et al. 1984). Linalool, the terpene alcohol that is induced by ChiA has been shown to affect preference of insect herbivores and pollinators alike (Cunningham 2004; Turlings et al. 1990). Therefore, frass protein ChiA mediated induction of these HIPVs (Herbivore-Induced Plant Volatiles) may be involved in altering the preference of secondary herbivores such as aphids on the host plant.

However, in our choice test assays we also observed that winged (alate) aphids showed marginal preference to undamaged control plants over plants with caterpillar damage but without frass effector (Fig. 5). This result is confounding since aphid performance was lowest on undamaged control plants compared to plants that were damaged either by caterpillars or artificially. Insect preference to a plant is controlled by several factors and alteration of volatiles by frass protein effector ChiA may not show a strong correlation to aphid preference but may have ecological relevance in attracting or repelling other insect pests and in plant-plant communication. While we focused on how effectors in caterpillar frass altered plant chemistry and in turn altered aphid preference, frass odors and/or visual cues from accumulated frass may also play a role in attracting or repelling aphids to maize plants with caterpillar herbivory.

The complexity of plant-insect dialogs encompasses multiple trophic levels, plant induced defenses by insect elicitors/effectors and asymmetry in host responses to insect herbivory in multi-species host-herbivore interactions. The interaction among these insects and plant species is modulated by highly complex chemical interactions. This study shows that proteinaceous effectors from caterpillar frass, which are known to suppress caterpillar-induced plant defenses and increase caterpillar performance, can also alternatively lead to changes in the plant that suppress performance of a secondary herbivore. Therefore, frass effector-mediated plant responses fails to benefit the performance of a subsequent herbivore on the plant from another feeding guild such as aphids.