Chew and spit: tree-feeding notodontid caterpillars anoint girdles with saliva
- First Online:
- Cite this article as:
- Dussourd, D.E., Peiffer, M. & Felton, G.W. Arthropod-Plant Interactions (2016) 10: 143. doi:10.1007/s11829-016-9416-1
Caterpillars of the notodontid Oedemasia leptinoides (formerly Schizura) use their mandibles to cut shallow girdles that encircle the petioles and stems of tree hosts. When girdles are complete, the larvae bathe the girdle surface with fluid. We test whether the fluid originates from the labial salivary glands or ventral eversible gland by blocking the openings to the glands and observing whether fluid is still released onto the girdles. Only larvae with functional labial salivary glands anointed girdles with fluid. Analysis of girdle rinses for a prominent salivary enzyme, glucose oxidase, confirmed that larvae apply saliva and documented that application occurs primarily at the end of girdling. We propose that girdling by notodontids, together with related furrowing and leaf-clipping behaviors exhibited by diverse caterpillar groups, serve at least in part to introduce salivary components to exposed vascular tissues; these compounds presumably function to suppress plant defensive responses normally elicited by caterpillar feeding.
KeywordsGirdling behavior Saliva Glucose oxidase Labial salivary gland Oedemasia leptinoides Notodontidae
Long viewed as passive victims of herbivory, plants are now known to possess complex multifaceted defenses (Voelckel and Jander 2014). Herbivore counterploys to these defenses have received less attention, although recent reviews document the breadth and sophistication of herbivore behavioral, physiological, and biochemical adaptations (Dussourd 1993; Karban and Agrawal 2002; Heckel 2014; Kant et al. 2015). Behavioral strategies include not just selective feeding on acceptable plant species and tissues, but also active manipulation of plants to disable defenses and block inducible responses. Mandibulate insect folivores, for example, commonly sever individual leaf veins, cut elongate trenches, or girdle stems and petioles (Dussourd 2009; Ganong et al. 2012). Cutting the xylem and phloem disrupts transport and profoundly alters leaf physiology (Delaney and Higley 2006; Delaney 2008). The cuts also rupture defensive canal systems such as laticifers that follow the vascular bundles (Fahn 1979; Metcalfe and Chalk 1983), thereby isolating and depressurizing the canals in a portion of the leaf where the insect feeds (Dussourd and Denno 1991). Severing leaf veins might also allow herbivores to introduce signaling molecules, toxins, or microbes into the vascular system to weaken the plant, alter plant nutrition, or impede defensive responses. However, such effects have not been previously associated with vein-cutting, trenching, or girdling by folivorous insects.
Alternatively, the fluid applied to girdles could originate from the ventral eversible gland (VEG), which opens just below the spinneret between the labium and first pair of legs (Fig. 1b, c). The VEG, also known as the adenosma, is presumed in notodontids to function in defense against predators (Detwiler 1922; Eisner et al. 1972). In O. leptinoides and related species, the VEG contains primarily formic acid (Roth and Eisner 1962, Attygalle et al. 1993). Simulated predatory attacks elicit discharge, which appears to the naked eye as a fine mist. Larvae accurately fire in any direction and can spray repeatedly before the VEG is depleted (movie 3 in Online Resources). VEG secretions can also impact plant defenses. Caterpillars of the notodontid Theroa zethus apply VEG fluid to the surface of veins and petioles of their latex-rich euphorb hosts (Dussourd 2015). The highly acidic secretion (pH~0.55) creates withered girdles and furrows that diminish latex outflow distal to the girdles and furrows where the larvae feed. Topical application of VEG secretion to leaf midribs alone sufficed to reduce latex exudation (Dussourd 2015). Spodoptera caterpillars (Noctuidae) also reportedly apply VEG fluids to the leaf surface together with oral secretions regurgitated during feeding. Larvae with an intact VEG triggered greater production of defense-related enzymes in hostplants and increased emission of volatiles relative to larvae with ablated VEG openings (Zebelo and Maffei 2012; Zebelo et al. 2014).
In this paper, we determine whether caterpillars of the notodontid O. leptinoides bathe girdles with saliva or VEG secretions. Our approach is to test whether cauterizing the spinneret or blocking the VEG opening prevents fluid release from girdling caterpillars and to analyze rinses of finished girdles to determine whether an abundant salivary protein, glucose oxidase, is present. Given that the Notodontidae contains over 4400 species (Schintlmeister 2013), our results with O. leptinoides likely apply to numerous insect–plant associations.
Larvae of O. leptinoides from field-collected females were reared on excised leaves of pecan [Carya illinoensis (Wang.) K. Koch] until the third or fourth instar, and then sleeved outside on pecan trees until the mid-final instar. To minimize discharge of the VEG, larvae were collected from the field still attached to leaflets that were placed in a refrigerator and then chilled on ice. Larvae were not touched directly until they were too cold to respond. Chilled larvae were divided into three treatments with ten larvae per treatment: intact caterpillars, larvae with cauterized spinnerets, and larvae with blocked ventral eversible glands (VEG). Each larva was tested only once. Intact larvae were handled and chilled the same as the other larvae, but were otherwise unaltered. For the second treatment, larvae had their labial salivary glands blocked by cauterizing the spinneret opening with a heat pen (Iso-Tip, Altoona, Wisconsin, USA) as previously employed by Peiffer and Felton (2005). A sharpened ultra-tip was used to seal the spinneret opening under a dissecting microscope. Larvae were allowed to recover overnight; only larvae that had resumed feeding were tested the following morning. For the third treatment, the VEG opening was blocked by sealing it with a drop of paint (Testors gloss enamel 1114 yellow) using a fine pin under a dissecting microscope. Since the VEG seal was only temporary, larvae were tested immediately, as were the intact larvae. After a caterpillar in the blocked VEG treatment girdled, it was chilled again on ice and the VEG opening was examined under a microscope to verify that the paint remained in place.
Larvae in all three treatments were released in the field on the rachis of a pecan leaf using trees not previously employed for rearing. Each larva was restricted to a single compound leaf by a paper barrier at the base of the leaf. When a larva initiated girdling, the leaf still attached to the plant was gently pulled under a dissecting microscope and girdling was observed until completion at ~20×. Viewing under a microscope was required to determine whether the caterpillar bathed the girdle in fluid after finishing the girdle. The number of larvae in the three treatments that applied visible fluid was compared with Fisher exact tests (R Development Core Team 2014).
To test further whether caterpillars applied saliva to girdles, completed girdles from all three caterpillar treatments were rinsed with 100 µl 0.065 M Tris–HCl, pH 6.8 with 0.7 % SDS. A pipette was used to dispense the cold fluid over the girdle, and then the rinses were collected as they dripped off the girdle into a microcentrifuge tube that was immediately stored on ice. Rinses were tested for the salivary enzyme glucose oxidase using an immunoassay described below. Any glucose oxidase (GOX) detected in the rinses could have been applied by the caterpillars while cutting the girdle as part of their normal feeding procedure. Alternatively, the glucose oxidase could have been applied primarily or exclusively at the end of girdling when the caterpillar bathes the girdle in fluid. To distinguish these two possibilities, an additional group of caterpillars were stopped after 2 min of girdling before the girdles were bathed with fluid. These girdles were also rinsed in Tris buffer for GOX analysis.
Rinses of three controls were also collected. For each control, artificial girdles were cut in pecan rachises using a wire stripper to simulate natural girdles. To verify that GOX is not naturally present on pecan rachises or in PBS buffer, one group of artificial girdles was treated with 5 µl PBS and then rinsed with 100 µl Tris buffer. To verify that our rinsing procedure and subsequent analysis suffice to detect GOX, another group of artificial girdles was treated with 5 µl of 1 mg/mL fungal GOX (Type VII from Aspergillus niger, Sigma-Aldrich, St. Louis, Missouri, USA) in PBS and then rinsed with 100 µl Tris buffer. Finally to verify that O. leptinoides saliva contains glucose oxidase that can be removed from the plant surface with rinsing and detected with our antibodies, a third group of artificial girdles was treated with 5 µl homogenized labial salivary glands, freshly dissected from final instar O. leptinoides caterpillars. The paired glands from each larva were ground with a Biomasher II pestle in 30 µl PBS, and then 5 µL of the pulverized solution was applied to the girdle, which was rinsed with Tris buffer as described previously. For each treatment, rinses of ten girdles were analyzed, except for incomplete girdles (9 replicates) and for fungal GOX and PBS controls (5 replicates each).
To test whether O. leptinoides salivary glands or VEG contained GOX activity, dissected glands were homogenized in PBS buffer and total protein quantified by Bradford assay. The sample was combined with 2 × native sample buffer (85 mM Tris HCl pH 6.8, 33 % glycerol, and 0.1 % bromphenol blue), and 10 µg was loaded onto an 8 % acrylamide gel and run under native conditions. An in-gel activity stain for GOX was done by immersing the gel in 0.1 M phosphate buffer, pH 7.0, containing 2.5 % ethanol, 0.5 mg/ml dianisidine, 7.5 µg/ml horseradish peroxidase, and 19 mg/ml glucose. After 20 min, the reaction was stopped with water.
In previous experiments, we found that to recover GOX from the plant surface, SDS is required (Peiffer and Felton 2005). However, after time on the plant and treatment with SDS, the GOX enzyme is no longer active and we are not able to do activity assays on the washes. Therefore, to test for the presence or absence of GOX on the girdles, we used immunoassays with an antibody specific to GOX. Each girdle rinse was applied to a Nanosep 10 K centrifugal filter (Pall Life Sciences, Ann Arbor MI) and centrifuged at 10,000×g, 4 °C, for 10 min. The concentrated samples were then recovered and combined with 10 µl of 2× SDS sample buffer (85 mM Tris HCl pH 6.8, 33 % glycerol, 2.7 % SDS, 0.6 % DTT, and 0.1 % bromphenol blue) and incubated in a boiling water bath for 5 min. Proteins were separated on a 12 % acrylamide gel and then transferred to 0.1 μm nitrocellulose (Pall Life Sciences, Ann Arbor MI). Western blots were blocked with 5 % dry milk in TTBS with 0.5 % gelatin, probed with Anti-GOX (produced by a synthetic peptide based on H. zea GOX sequence derived from the cDNA) diluted 1:5,000, and detected with the Vector ABC kit (Vector Laboratories, Burlingame, CA) and Vector DAB kit.
For the incomplete girdles, rinses produced faint bands at most indicating that saliva with GOX is applied primarily at the end of girdling (Fig. 4d). As expected, rinses of artificial girdles bathed with ground O. leptinoides salivary glands or fungal GOX all contained detectable GOX (Fig. 4e, f). The multiple bands present in the salivary gland rinses could be due to salivary proteases breaking GOX into multiple pieces. None of the girdles bathed just with PBS contained GOX confirming the absence of any proteins in the PBS buffer or on the plant surface that react with our antibody (Fig. 4e).
Oedemasia leptinoides caterpillars with blocked ventral eversible glands anointed girdles with fluid, whereas larvae with cauterized spinnerets did not. The fluid therefore must originate from the labial salivary glands and not from the digestive system (regurgitant), VEG, or other secretory structures such as the mandibular salivary glands. The presence of glucose oxidase in girdle rinses from intact and VEG-blocked larvae further confirmed that saliva was deposited on the girdles. Since rinses of incomplete girdles contained only trace amounts of GOX at most, saliva must be applied primarily at the end of girdling when the larva bathes the girdle with fluid. Another notodontid, Schizura ipomoeae, similarly wipes its labium over girdles in river birch at the end of girdling, thereby coating the girdle with fluid, presumably saliva from the labial salivary glands (movie 4 in online resources). In contrast, larvae of Theroa zethus (Notodontidae) use VEG acid to create withered girdles and furrows in leaf veins (Dussourd 2015). However, T. zethus larvae also wipe their labium over the girdle surface leaving behind visible fluid, presumably saliva (Dussourd 2015). These observations suggest that saliva application is widespread in notodontids that create girdles and furrows.
How O. leptinoides girdles and saliva affect pecan defenses is not known. Pecan leaves produce several polyphenols including juglone (Gueldner et al. 1994), which is known to inhibit the growth of nonadapted caterpillar species (Lindroth et al. 1990; Thiboldeaux et al. 1994). The leaves also emit complex mixtures of volatile organic compounds including monoterpenes and sesquiterpenes (Farag 2008; Corella-Madueño et al. 2011; Vargas-Arispuro et al. 2013; Zhu et al. 2013). In other tree species, caterpillar feeding increases emission of volatile compounds (Arimura et al. 2004; Copolovic et al. 2011), which are known to prime defenses in undamaged tissues (Girón-Calva et al. 2014) and to attract parasitoids and predators (Turlings and Wäckers 2004; Degenhardt 2008; McCormick et al. 2012).
Plant production of defensive chemicals and emission of volatiles can be suppressed by effectors in caterpillar saliva (Delphia et al. 2006; Weech et al. 2008; Felton et al. 2014). For example, glucose oxidase in the saliva of H. zea (Noctuidae) larvae prevents the inducible increase of nicotine in tobacco normally triggered by feeding (Musser et al. 2002). Glucose oxidase has a widespread distribution in insect saliva being found in 22 families of caterpillars, as well as in sawflies and aphids (Harmel et al. 2008; Eichenseer et al. 2010). All ten notodontid species examined by Eichenseer et al. (2010) produced substantial quantities, including Coelodasys unicornis (formerly Schizura), another girdler (Ganong et al. 2012). As documented in this paper, the saliva that O. leptinoides paints on pecan girdles contains glucose oxidase, which may similarly suppress pecan defensive responses. In addition to GOX, caterpillars may deploy a suite of other salivary effectors to inhibit host defenses. For example, multiple ATPases in the saliva of H.zea were shown to suppress induced plant defenses (Wu et al. 2012).
Oedemasia leptinoides caterpillars not only produce girdles, they also sometimes cut furrows in leaf midribs or sever leaf petioles after only partially consuming a leaf, a behavior called leaf or petiole clipping (Ganong et al. 2012). Leaf-clipping has a widespread distribution, being reported in 40 species in 12 families of caterpillars and sawflies (Heinrich and Collins 1983; Risley and Crossley 1988; Edwards and Wanjura 1989; Weinstein 1990; Scriber 1996; Akino 2005; Ganong et al. 2012). An O. leptinoides larva photographed clipping a birch leaf spent four minutes rubbing its labium over the severed petiole stub (Ganong et al. 2012), apparently anointing the stub with saliva. O. leptinoides larvae observed under a microscope in this study while cutting furrows in pecan leaflets applied fluid, presumably saliva, to the furrows (Madalyn Van Valkenburg pers. comm.). Thus, girdling, leaf-clipping, and furrowing may all serve the common function of introducing salivary components into vascular tissues, presumably to suppress plant defensive responses. Whether girdling by ovipositing beetles on pecan and other trees (Coppedge 2011; Paro et al. 2014) or vein-cutting and trenching by over 19 families of folivorous caterpillars, beetles, katydids and sawflies (Dussourd 2009), likewise introduce saliva into the cuts is not known. Our results with notodontids suggest that saliva application is likely widespread, not just during feeding (Felton et al. 2014), but also during behavioral manipulation of hostplants before and after feeding.
Many thanks to Madalyn Van Valkenburg for assistance in the laboratory and field, to Daniel Champion for assistance editing movies, to two anonymous reviewers for helpful suggestions, and to Jim Miller and David Wagner for providing useful information on notodontids. Financial support was provided by US Department of Agriculture–National Institute of Food and Agriculture Grant 2011-67013-30352 and National Science Foundation Grant IOS-1256326 (G.W.F), the University of Central Arkansas Research Council (D.E.D.), and the Arkansas Center for Plant-Powered Production (P3). The P3 Center is funded through the RII: Arkansas ASSET Initiatives (AR EPSCoR) I (EPS-0701890) and II (EPS-1003970) by the National Science Foundation and the Arkansas Science and Technology Authority.
Movie 2 Closeup of a final instar Oedemasia leptinoides larva completing a girdle in a pecan rachis. The white fibrous girdle appears dry during the cutting procedure. At the end, the larva moves its head side to side as it rubs its spinneret back and forth over the surface, painting the girdle with saliva (mpg 121514 kb)