Knowledge of the diet of generalist insect herbivores is critical for understanding insect feeding preferences regarding different plants, as well as for detecting and predicting plant–insect interactions in natural communities. This becomes especially important when the insects of interest are agricultural pests, such as grasshoppers. Grasshoppers cause significant damage to crops and rangelands resulting in serious economic losses in the United States and worldwide. For example, in 17 western U.S. states, grasshoppers annually consume 25% of available rangeland forage, which averages about $1 billion per year (Hewitt and Onsager, 1983). Because of their important role in accelerating nutrient cycling, grasshoppers can influence plant community composition and, in particular, alter the abundance and species richness of plant species (Belovsky and Slade, 2000). Consequently, knowledge of the feeding preferences of grasshoppers can be important for control efforts and effective restoration of damaged areas (Branson and Sword, 2009).

The first step in any study on feeding preferences of insect herbivores is an accurate confirmation of food that is consumed. Among various techniques available for food identification (including direct observation, feeding trails, and microscopic gut content analysis), PCR assays have been shown to be an accurate and relatively quick method for detecting ingested plants, features that are especially important for large-scale studies (e.g., Jurado-Rivera et al., 2009; Garcia-Robledo et al., 2013). In particular, plant DNA sequences extracted from insect gut contents can provide information about insect feeding choices occurring under natural conditions, which can be hidden from direct observations of insects on plants, or may contradict feeding preferences of insects observed in laboratory feeding trials (e.g., Garcia-Robledo et al., 2013). Therefore, potentially erroneous plant–insect interactions can be corrected.

Previous studies on plant DNA detection from insect guts have been conducted on beetles (e.g., Jurado-Rivera et al., 2009; Wallinger et al., 2013), moths (Miller et al., 2006), flies (Junnila et al., 2011), and hemipterans (Matheson et al., 2008), but only Matheson et al. (2008) included one grasshopper in their study, dissecting it 4 h post-ingestion (PI). Studies that used small insects or insect larvae often obtain whole-body DNA extracts (e.g., Staudacher et al., 2011). The extraction of plant DNA from relatively large insects is complicated by the presence of excessive amounts of nontarget DNA of the herbivore; in this case, isolating the digestive system and preventing contamination of gut contents with possible plant material from the outside surface of the insect (e.g., Matheson et al., 2008) is critical for increasing the yield of target plant DNA. Grasshoppers that reach large sizes as adults are among the most important agricultural pests, with enormous economic costs (Hewitt and Onsager, 1983); therefore, information about their food consumption and, in particular, on tissue preparation and subsequent detection of plant DNA from their gut contents is much needed.

In addition, the availability of a protocol for plant DNA extraction from different parts of an insect gut has many advantages in terms of exploring new aspects of herbivore feeding, and is especially useful for insects of relatively large size. It can allow the researcher to “follow” the plant DNA during food consumption and, for example, (1) to determine the approximate time of food consumption from its location in each compartment of the insect digestive system, or in the case of mixed diet, (2) to infer the sequence of ingestion of different plant species.

This study provides, for the first time, (1) an optimized step-by-step protocol for DNA extraction and PCR assay for detecting plant food in grasshoppers; (2) evidence of detectability of ingested plant DNA in nymphs and adult grasshoppers via feeding experiments; and (3) a step-by-step protocol for dissection and plant DNA detection in different sections of grasshopper guts to follow up the digestive pathway through the gut.

Fig. 1.

PCR amplification of three fragments of the trnL gene (primers c-d, e-f, and g-h) and the ITS gene from four test plants (A) and from ingested plants within gut contents of four grasshopper individuals (B). Ci: Cichorium intybus; Ms: Miscanthus sinensis; Pl: Plantago lanceolata; Tr: Trifolium repens. Each group of four lanes for each plant species represents one plant individual. Grl : Melanoplus differentialis, Gr2: nymph Melanoplus spp. grasshopper; Gr3–4: M. femurrubrum grasshoppers; M: molecular marker (1-kb DNA ladder). Each group of four lanes for each grasshopper species represents one individual. Primers c-d successfully amplified fragments of the chloroplast trnL (UAA) gene in several other test plants (C) and in ingested plants in gut contents of several nymph individuals of the Melanoplus spp. grasshoppers (D). Each lane represents a different plant individual of Poaceae, Asteraceae, Fabaceae, and Plantaginaceae families (C) and a different grasshopper individual (D).

Table 1.

The feeding experiments used in the study for plant DNA detection from grasshopper gut contents.

METHODS AND RESULTS

Sample collection—Adult Melanoplus femurrubrum and M. differentialis grasshoppers (Acrididae: Orthoptera), and nymphs of Melanoplus spp. grasshoppers (i.e., M. differentialis and M. bivittatus) were collected at the Western Maryland Research and Education Center (Keedysville, Maryland, USA) and the Cincinnati Center for Field Studies (New Haven, Ohio, USA). In addition, 40 different plant species of Poaceae, Asteraceae, Fabaceae, and Plantaginaceae families were collected from the study plots and used for the feeding experiments described below. Among these species, Trifolium repens L., Cichorium intybus L., Plantago lanceolata L., and Miscanthus sinensis Andersson were used for testing primers; voucher specimens for these plants (AA-0001, AA-0002, AA-0003, and AA-0004, respectively) have been deposited at the herbarium of the University of Cincinnati (CINC). Furthermore, Bouteloua curtipendula (Michx.) Torr, and Bothriochloa bladhii (Retz.) S. T. Blake used in the feeding experiments described below were grown at the University of Cincinnati greenhouse from seeds obtained from Prairie Moon Nursery (Winona, Minnesota, USA) and Plant World Seeds (Newton Abbot, Devon, United Kingdom), respectively.

Protocol development—To first obtain plant DNA in grasshopper guts, a step-by-step protocol was developed. Following are the most important steps of this protocol; more details are provided in Appendices 1 and 2.

Step 1: Dissection and tissue preparation—After collection, grasshoppers' bodies and plant leaves (1–2 leaves from each plant species) were immediately frozen separately at −20°C. On the day of dissection, four frozen grasshoppers (two adult M. femurrubrum, one adult M. differentialis, and one nymph) were removed from the freezer and immediately rinsed with 70% ethanol to wash off all possible large, nonhost plant debris from the exterior of the insects. The grasshopper tissues were relatively soft and easy to dissect, so additional time for thawing was not needed. The hind legs and wings were then removed using fine forceps and fine scissors from a standard dissecting set. The exoskeleton of each grasshopper was then cut along the side and the digestive system was extracted. Whole guts were then stored in 1.5 mL microcentrifuge tubes with 70% ethanol overnight before the DNA extraction (Appendix 1,  Video 1).

Step 2: DNA extraction—Plant DNA was extracted from four samples of grasshopper gut contents and from T. repens, C. intybus, P. lanceolata, and M. sinensis, representing grasshopper host plants (prepared in Step 1 above); both plants and grasshoppers were collected from the same study plot. DNA extraction was conducted with QIAGEN DNeasy Plant Mini Kit (cat. no. 69104; QIAGEN, Culver City, California, USA) according to QIAGEN guidelines. Although this kit is generally used for DNA extraction from standard plant tissue, the kit was recommended by QIAGEN Technical Service as useful for isolating plant material inside the insect gut. After isolation, DNA from plants and grasshopper guts was stored at −20°C for further PCR amplification.

Step 3: Primer testing and PCR amplification—DNA barcodes amplifying the chloroplast trnL (UAA) gene and the nuclear ITS 1–2 region were chosen for screening of plant DNA obtained from grasshopper guts because these primers proved successful for detecting ingested plant DNA in a wide range of insect herbivores (e.g., Jurado-Rivera et al., 2009; Staudacher et al., 2011; Pumarino et al., 2011). In contrast, primers suggested by Matheson et al. (2008) that targeted the rbcL region did not work in initial screens for this study and were not pursued further. Four sets of universal primers were tested separately on plants and grasshopper gut contents: three sets for noncoding regions of the chloroplast trnL (UAA) gene (Taberlet et al., 1991, 2007) and one set for the nuclear ITS region (White et al., 1990). The primer mix was prepared for each primer pair (2 µM of each forward and reverse primer). Each PCR reaction (of 10 µL volume) consisted of the following: 5 µL of QIAGEN Master Mix (QIAGEN), 1 µL of primer mix, 3.8 µL of dH₂O, and 0.2–0.3 µL of DNA. Although other PCR-based protocols sometimes use larger amounts of DNA (e.g., Matheson et al., 2008), the smaller amounts used here were sufficient, as evidenced below. Samples were amplified under the following thermocycler conditions: denaturation of 95°C for 15 min; followed by 35 cycles of 95°C for 15 s, 57°C for 90 s, and 72°C for 60 s; followed by a final extension of 60°C for 30 min. PCR products were then separated in a 1 % agarose gel and visualized under a UV transilluminator (Fig. 1A–B).

Video 1.

Demonstration of grasshopper dissection detailing the procedure for isolating the gut and preparing the foregut and combined midgut+hindgut parts for DNA extraction. This video is an MP4 file and can be viewed  here with QuickTime or Windows Media Player, or can be viewed from the  Botanical Society of America's YouTube channel.

Step 4: DNA sequencing and final primer selection—To confirm the presence and identity of plant DNA isolated from grasshopper guts, PCR products obtained from grasshoppers and from known plant species (from Step 1 above) were sequenced using Sanger sequencing at the Beckman Coulter Genomics facility (Danvers, Massachusetts, USA). Sequences were then edited in BioEdit (Hall, 1999) and BLASTed against the National Center for Biotechnology Information (NCBI) GenBank database ( http://www.ncbi.nlm.nih.gov/genbank/) for plant identification using 98–100% match identity. Following Chen et al. (2010), the quality of sequences for both plants and grasshopper gut contents was estimated using CodonCode Aligner 4.2.5.0 (CodonCode Corporation, Centerville, Massachusetts, USA) for low, middle, and high quality levels. The highest-quality sequences (quality values higher than 30) were observed for primers c-d (Taberlet et al., 1991); consequently, these primers were chosen to demonstrate the utility of this protocol.

Table 2.

Plant DNA detectability in grasshopper gut contents across several time intervals post-ingestion in different grasshopper species. A single grasshopper was tested for each time point in all experiments.

Fig. 2.

PCR amplification of the trnL gene at different time intervals post-ingestion (PI) after feeding experiments with grasshoppers. The numbers correspond to hours post-ingestion (h PI). One grasshopper individual has been dissected at each time point. “−”: negative control (DNA from grasshopper's leg muscle tissue); “+”: positive control (plants offered for feeding). Each lane (A–C) represents a different grasshopper individual. (A) No-choice feeding experiment with adult Melanoplus femurrubrum grasshoppers and Bothriochloa bladhii plants. The plant DNA was present in the grasshopper guts up to 12 h PI. (B) Choice feeding experiment with adult M. femurrubrum grasshoppers and a mixture of plants. The plant DNA was present in the grasshopper guts up to 8 h PI. (C) Choice feeding experiment with nymph Melanoplus spp. grasshoppers. The plant DNA was present in the grasshopper guts up to 12 h PI. (D) Choice feeding experiment with adult M. differentialis grasshoppers. Bb: Bothriochloa bladhii (positive control 1); Be: Bouteloua curtipendula (positive control 2); Neg: negative control; fg: foregut; mhg: combined midgut+hindgut. Each of the two lanes (foregut and combined midgut+hindgut) at each time point represents the same grasshopper individual. The plant DNA was present in both foregut and combined midgut+hindgut parts up to 22 h PI.

To confirm the utility of primers c-d for a wide range of grasshoppers' potential host plants and grasshopper gut content samples, DNA extraction, amplification, and sequencing were repeated with the remaining 36 collected plant species of Poaceae, Asteraceae, Fabaceae, and Plantaginaceae families and also with 26 nymphs of the Melanoplus spp. grasshoppers collected from the same study plots. High-quality sequences with quality values higher than 30 (Chen et al., 2010) were obtained for all 36 study plants (100%, P < 0.0001, binomial test) and for 18 out of 26 (69%) grasshopper guts (P = 0.03, binomial test). For this analysis, grasshopper gut contents with only single plant DNA were considered. Thus, these results demonstrated that the 500-bp region of the chloroplast trnL (UAA) gene, amplified by primers c-d, can be reliably detected in grasshopper guts and their potential host plants (Fig. 1C–D).

Testing the protocol —To further demonstrate the effectiveness of this protocol and to determine how long plant DNA remains detectable in the digestive system of grasshoppers of different sizes, three choice experiments and one no-choice feeding experiment with Melanoplus grasshoppers were conducted (Table 1, Appendix 2). In no-choice experiments, grasshoppers were fed a single plant species, while in choice experiments grasshoppers were provided with two or more plant species. Grasshoppers were originally collected in the field and their weights ranged from 0.11–1.66 g. Following Siemann and Rogers (2003), grasshoppers were starved for 24 h prior to all feeding experiments to make sure that no previously digested plants were present in the gut. Nymph grasshoppers (Experiment 1, choice) and adult M. differentialis grasshoppers (Experiment 2, choice) were offered leaves from both Bouteloua and Bothriochloa grasses for 3.5 h, M. femurrubrum grasshoppers were fed a mixture of plants for two days (Experiment 3, choice), and additionally, another group of M. femurrubrum grasshoppers were fed leaves from Bothriochloa bladhii grass for 3.5 h (Experiment 4, no-choice). After feeding, grasshoppers were transferred to new containers that did not contain food. Grasshoppers were then frozen separately at −20°C at several time intervals after feeding (one grasshopper at each time point); plant DNA was then extracted and sequenced from each grasshopper using the protocol described above.

The results demonstrated that plant DNA can be detectable up to 12 h PI in the guts of nymph Melanoplus spp. grasshoppers (Table 2, Fig. 2A) and adult M. femurrubrum grasshoppers (Table 2, Fig. 2B, C), as well as up to 22 h PI in M. differentialis grasshoppers, which were the largest in size (Table 2, Fig. 2D). Because of the difference in size and, consequently, weight of grasshoppers, the DNA extraction step (Step 2) of the protocol was adjusted. To meet the requirements for sample weight according to the QIAGEN kit (≤100 mg wet weight), the following were used in this study: whole bodies of nymph grasshoppers, whole guts of M. femurrubrum grasshoppers, or two parts of a gut of M. differentialis grasshoppers (foregut and combined midgut+hindgut, Appendix 1).

Choice feeding experiments with M. differentialis grasshoppers were also used to illustrate the utility of the proposed protocol for detection of plant DNA in different parts of the grasshopper digestive system. In this case, the tissue preparation step (Step 1) of the described protocol was also adjusted: after isolating the grasshopper gut from the body, the foregut and combined midgut+hindgut parts were separated (Appendix 1,  Video 1). These parts were then stored separately in 70% ethanol, and plant DNA was then extracted from each section of the digestive system. The results of PCR amplification and obtained sequences of ingested plant DNA demonstrated that a researcher can “follow” the plant DNA in the process of food consumption up to 22 h PI and can make conclusions about the feeding behavior of an insect—specifically, on the order of ingested plants. For example, in this study, the pattern of PCR amplification for foregut and combined midgut+hindgut sections at 3 h PI (Fig. 2D) suggested that M. differentialis grasshoppers consumed different plant species sequentially, and did not switch often between grasses offered in the choice experiments.

CONCLUSIONS

Considering the high agricultural significance of grasshoppers (e.g., Hewitt and Onsager, 1983) and their impact on plant communities (e.g., Belovsky and Slade, 2000), there is a major need for an effective protocol for detecting grasshopper interactions with host plants. The utility of the chloroplast trnL (UAA) gene for detecting plant DNA from some coleopteran species has been demonstrated in similar studies (e.g., Jurado-Rivera et al., 2009; Staudacher et al., 2011). The developed protocol also demonstrated the utility of the chloroplast trnL (UAA) gene for PCR-based work with grasshoppers; 500-bp fragments of ingested plant DNA were successfully amplified and sequenced within grasshopper guts across multiple time intervals postingestion. The developed protocol was also effective for detecting plant DNA from different sections of grasshopper guts, which has not yet been reported as previous studies on large insects used whole guts for plant DNA extraction (e.g., Matheson et al., 2008).

The protocol described here has many applications. For example, researchers can sacrifice a small subsample of grasshoppers to accurately determine the time of starvation needed to make sure that no other previously digested plant fragments are present in gut contents. In addition, researchers can follow the “movement” of plant DNA during the food consumption process to better understand the feeding behavior of insect herbivores.

The main advantages of this protocol are as follows: (1) it includes a relatively quick DNA extraction step (less than 3 h); (2) it results in high resolution of the trnL gene for plant identification at the genus and, often, at the species level; and (3) it capitalizes on the low cost of PCR and sequencing procedures, which are advantageous for small laboratories without access to next-generation sequencing technologies. Potential difficulties of using this protocol include the following: (1) occasionally low resolution of the trnL in species discrimination (three out of 40 cases in this study), and (2) detection of multiple plant DNA in some gut contents (six out of 26 samples in this study). When critical, the former can be addressed by amplifying additional loci; the latter requires additional molecular techniques, such as cloning (Garcia-Robledo et al., 2013), or less labor-intensive methods, such as computational analysis of mixed sequencing chromatograms (Kommedal et al., 2008; Chang et al., 2012). Overall, this is a convenient protocol for detecting plant–insect interactions, and although it was developed specifically for grasshoppers, it can potentially be extended to other plant and insect species to explore different aspects of insect herbivory.