Caterpillar attack triggers accumulation of the toxic maize protein RIP2



  • Some plant-derived anti-herbivore defensive proteins are induced by insect feeding, resist digestion in the caterpillar gut and are eliminated in the frass. We have identified several maize proteins in fall armyworm (Spodoptera frugiperda) frass that potentially play a role in herbivore defense. Furthermore, the toxicity of one of these proteins, ribosome-inactivating protein 2 (RIP2), was assessed and factors regulating its accumulation were determined.
  • To understand factors regulating RIP2 protein accumulation, maize (Zea mays) plants were infested with fall armyworm larvae or treated with exogenous hormones. The toxicity of recombinant RIP2 protein against fall armyworm was tested.
  • The results show that RIP2 protein is synthesized as an inactive proenzyme that can be processed in the caterpillar gut. Also, caterpillar feeding, but not mechanical wounding, induced foliar RIP2 protein accumulation. Quantitative real-time PCR indicated that RIP2 transcripts were rapidly induced (1 h) and immunoblot analysis indicated that RIP2 protein accumulated soon after attack and was present in the leaf for up to 4 d after caterpillar removal. Several phytohormones, including methyl jasmonate, ethylene, and abscisic acid, regulated RIP2 protein expression. Furthermore, bioassays of purified recombinant RIP2 protein against fall armyworm significantly retarded caterpillar growth.
  • We conclude that the toxic protein RIP2 is induced by caterpillar feeding and is one of a potential suite of proteins that defend maize against chewing herbivores.


Plants are frequently attacked by voracious insect herbivores, and consequently have evolved a broad range of defense mechanisms to combat these pests. Some of the most effective plant defense mechanisms combine physical traits such as trichomes, thorns and cuticles (Levin, 1973) with chemical substances that deter feeding and ‘poison’ the insects. Often these defenses are constitutively produced and always available to protect the plant from herbivory. Plants also employ inducible defenses that are activated in response to herbivore attack. Induced defenses include indirect defenses such as production of green leaf volatiles and other volatile organic compounds that attract natural enemies of the attacking insects, and direct defenses that rely on the accumulation of secondary metabolites, defensive proteins, and other toxic compounds (Howe & Jander, 2008). Toxic secondary metabolites include glucosinolates, benzoxazinoids, alkaloids, and phenolics and other compounds. In addition, herbivore-defensive proteins are typically induced in response to insect feeding. A major class of defensive proteins includes the protease inhibitors (PIs) that rapidly accumulate in potato and tomato after insect feeding (Green & Ryan, 1972). In the insect gut, PIs interact with gut proteases and thereby inhibit digestion. Transgenic plants expressing PI genes have detrimental effects on herbivore growth; however, insects often counter this defense by producing more or different digestive proteases (Chen, 2008; Howe & Jander, 2008; Zhu-Salzman et al., 2008). Plant oxidative enzymes such as polyphenol oxidases (PPOs), peroxidases, lipoxygenases, and ascorbate oxidases are also involved in plant defenses against insect herbivores (Chen, 2008). They oxidize essential nutrients or form electrophilic products that react with the nucleophilic amino acid side chains and lower plant nutritive value (Felton et al., 1994; Constabel, 1999; Chen, 2008). For example, it has been shown that a maize anionic peroxidase suppresses caterpillar growth when it was ectopically expressed in callus (Dowd et al., 2010). Besides these proteins, some proteases, such as maize insect resistance 1-cysteine protease (Mir1-CP) and leucine aminopeptidase (LAP), have been shown to defend plants against insect herbivores (Chao et al., 1999; Chen, 2008; Zhu-Salzman et al., 2008).

Herbivore- or jasmonate (JA)-inducible proteins have been identified by microarray and proteomic analyses of insect-infested plants (Reymond et al., 2000; Collins et al., 2010). However, it has been shown that several plant defense proteins can successfully survive digestion in the insect gut and be eliminated in the frass, whereas proteins that do not have defensive functions, such as Rubisco, are rapidly degraded and not detected in frass. For example, it has been demonstrated that several jasmonate-inducible proteins (JIPs) actually retain enzymatic activity in the insect midgut and can be deleterious to the herbivore in the digestive tract (Chen et al., 2005). This finding implies that proteomic analysis of frass protein components is a novel approach for identifying plant proteins that are resistant to digestion and potentially function in herbivore defense (Chen et al., 2007). Using this technique, several JIPs were found in Manduca sexta frass that reduced the insect's ability to obtain essential nutrients from the plant (Chen et al., 2007). One of these proteins was a threonine deaminase isoform 2 (TD2; Chen et al., 2007). TD2 is synthesized as proenzyme, and when insects ingest tomato leaves, TD2 is converted to the processed form in their guts, where it reduces the concentrations of threonine, an essential amino acid for phytophagous insects (Chen et al., 2007). In this study, we used proteomic analysis of fall armyworm (FAW, Spodoptera frugiperda) frass to determine if there are possible herbivore-induced defensive proteins in maize that resist digestion. One of the predominant proteins found in this analysis was identified as ribosome-inactivating protein 2 (RIP2).

Ribosome-inactivating proteins (RIPs) are enzymes that have site-specific RNA N-glycosidase activity that arrest translation (Bass et al., 2004). They block translational elongation by depurinating residues on the large ribosomal RNA component (Endo & Tsurugi, 1987; Endo et al., 1987; Nielsen & Boston, 2001). In 1925, it was reported that pokeweed RIPs inhibit viral infection (Duggar & Armstrong, 1925; Irvin, 1983; Nielsen & Boston, 2001). Since then it has been shown that some RIPs function in defense against viruses, pathogens, and insects (Nielsen et al., 2001; Peumans et al., 2001; Bertholdo-Vargas et al., 2009). Based on their protein structure, RIPs are classified into three types (Nielsen & Boston, 2001). Type 1 RIPs consist of a single polypeptide chain with an approximate molecular mass of 30 kDa (Nielsen & Boston, 2001). Pokeweed antiviral protein, soapwort saporin, and barley translation inhibitor are type 1 RIPs (Nielsen & Boston, 2001). Type 2 RIPs have two polypeptide subunits that have an enzymatic domain and galactose-binding domain linked by disulfide bonds with an approximate molecular mass of 60 kDa (Nielsen & Boston, 2001). Ricin and abrin are well-known, highly toxic type 2 RIPs (Nielsen & Boston, 2001). Type 3 RIPs are synthesized as inactive precursors (proRIPs) that require proteolytic modification to form processed RIP (Nielsen & Boston, 2001). This processing step requires removal of c. 25 amino acids from the middle of the protein precursor coupled with additional processing at the N- and C-termini (Nielsen & Boston, 2001). Barley JIP60 and two maize RIP proteins belong to type 3 RIPs (Nielsen & Boston, 2001). The precursor form of the RIP1 protein is c. 32 kDa and after processing it comprises 16.5 and 8.5 kDa polypeptides that associate as a heterodimer (Bass et al., 2004). Unprocessed RIP2 protein is c. 30 kDa and the size of the processed subunits has not yet been reported.

There are several other important differences between the two RIP protein isoforms, RIP1 and RIP2, identified in maize (Bass et al., 2004). First, the amino acid sequences of RIP1 and RIP2 proteins are c. 73% similar (Bass et al., 1995). Secondly, the RIP1 protein is expressed in kernel, where it is believed to protect the seed from pathogen infection (Nielsen et al., 2001), whereas the RIP2 protein is expressed throughout the plant from the leaves to the tassel, but not the kernel (Bass et al., 2004). Thirdly, the RIP1 gene maps on chromosome 8 (bin 8.05) of the maize genome, while the RIP2 gene is located on chromosome 7 (bin 7.04), where there is a strong quantitative trait locus for caterpillar resistance (Bass et al., 1995, 2004; Brooks et al., 2005). These studies suggest that the RIP2 protein may play an important defensive role in vegetative tissues, as the RIP1 protein does in the kernel, but this has not been confirmed by directly testing the toxicity of RIP2 on insect herbivores.

The first objective of this study was to determine if putative herbivore defense proteins expressed in maize could survive digestion in the FAW gut and be excreted in the frass as was the case for M. sexta fed on tomato (Chen et al., 2005). The second objective was to determine if one of the putative defensive proteins, RIP2, identified in frass was induced by caterpillar feeding and toxic to FAW larvae.

Materials and Methods

Plant materials, insect rearing, and frass collection

Maize (Zea mays L.) genotypes resistant (Mp708) and susceptible (Tx601) to fall armyworm (Spodoptera frugiperda J. E. Smith) feeding were obtained from Dr W. Paul Williams (USDA-ARS Corn Host Plant Resistance Research Laboratory, Mississippi State University, USA). These genotypes were used because of their differences in insect resistance and both were fed to caterpillars for the frass proteomic analysis. Unless otherwise noted, Tx601 was used for the experiments reported here. Seeds were sown in 18 l black plastic pots (two plants per pot) filled with topsoil (Hagerstown loam) in the Crop and Soil Sciences glasshouse at The Pennsylvania State University (University Park, PA, USA). Plants were grown with supplemental light (photosynthethically active radiation ranged from 500 to 1200 μmol m−2 s−1), the average temperature was 28°C, a 16 h photoperiod was used and plants were watered as needed. Maize plants were grown until the V8 stage (eight fully emerged leaves; Ritchie et al., 1986). For RNA and protein isolation (see later), tissue (c. 5 mm) surrounding the wounded area near the whorl was collected after caterpillar feeding or wounding with a paper punch (6 mm diameter, 10 times per plant). Samples were stored at −80°C after immediately freezing in liquid nitrogen.

Fall armyworm eggs were obtained from the USDA-ARS Corn Host Plant Resistance Research Laboratory, Mississippi State University, USA. Larvae were reared on an artificial diet (Peiffer & Felton, 2005) in a 27°C incubator with a 16 h photoperiod until they molted to the fifth instar. For frass collection, plants were infested with three fifth-instar FAW larvae per plant for 24 h. Leaf tissue was collected in an c. 1 cm radius around the feeding site, placed in diet cups and fed to naïve fifth-instar larvae. Frass was collected within 24 h after feeding and stored at −80°C after immediately freezing in liquid nitrogen.

Immunoblot analysis

The leaf tissue and caterpillar frass were frozen in liquid nitrogen and homogenized using the Geno/Grinder 2000 (SPEX CertiPrep, Metuchen, NJ, USA). These samples were then extracted with sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer (Laemmli, 1970). Protein concentrations were determined by RC-DC Protein assay (Bio-Rad). Equal amounts of protein were loaded in each gel lane and separated by SDS-PAGE. Gels were blotted onto nitrocellulose membranes using the Panther Semi-Dry Electroblotter (Thermo Scientific, Owl, MA, USA). The blots were reversibly stained with Ponceau S (Salinovich & Montelaro, 1986) before immunoblot analysis to visualize protein transfer.

Polyclonal antibody to RIP protein was obtained from Dr Rebecca Boston (North Carolina State University, Raleigh, NC, USA). It should be pointed out that the anti-RIP antibody reacts with both RIP1 and RIP2 proteins, but the RIP1 protein is slightly larger than the RIP2 protein and the two forms can be differentiated by size on SDS-PAGE. (Bass et al., 1992, 2004). Western blots were carried out using 1 : 10 000 diluted anti-RIP antibody and 1 : 10 000 diluted anti-rabbit secondary antibody conjugated with horseradish peroxidase (Thermo Fisher Scientific, Rockford, IL, USA). Immunoreacting proteins were detected by chemiluminescence (West Femto Maximum Sensitivity Substrate, Thermo Scientific).

RIP2 bioassays and RIP2 digestion

Recombinant RIP2 (rRIP2) protein was produced and purified as described by Bass et al. (2004). Processed recombinant RIP2 (prRIP2) protein was prepared by treating the recombinant protein with immobilized papain (Thermo Fisher Scientific). The papain beads–prRIP2 protein mixture was centrifuged and the supernatant containing the prRIP2 protein was collected. For toxicity bioassays, the appropriate amount of BSA, rRIP2 protein, or prRIP2 protein (0.8 μg protein (100 mg)–1 artificial diet) was placed into 24-well multiwell plate (BD Falcon, Franklin Lakes, NJ, USA) containing artificial diet (Peiffer & Felton, 2005). An equal volume of phosphate-buffered saline served as the control. The plates were air-dried and a single larva (1 d after hatching) was placed in each well and each treatment had 24 larvae. The plates were covered with Breath Easy™ (USA Scientific Inc., Ocala, FL, USA) and incubated at 26°C for 5 d with a 16 h photoperiod before the final larval weight was determined. The results were analyzed by SAS statistical software (SAS Institute, Cary, NC, USA).

Phytohormone treatments

The effect of exogenous ethylene (ET) application was tested by spraying plants with 3 mM 2-chloroethylphosphonic acid (ethephon; Sigma-Aldrich). The effect of exogenous methyl jasmonate (MeJA; Bedoukian Research Inc., Danbury, CT, USA) was tested by spraying with 0.01% MeJA. MeJA was first diluted 10-fold with ethanol to become 10% MeJA stock solution (v/v). The fresh working solution of 0.01% MeJA was prepared with sterile water. In addition, the effect of methyl salicylate (MESA; MP Biomedicals Inc., Solon, OH, USA) was tested by spraying 1 mM MESA on plants. The ethephon, MeJA, and MESA concentrations used were similar to those used on maize in a previous study (Ankala et al., 2009). ABA (Sigma-Aldrich) was first dissolved as 50 mM solution in ethanol and then diluted in sterile water to 300 μM ABA and sprayed on plants. Each plant hormone (c. 50 ml per plant) was sprayed on two individual plants for both wounded and unwounded treatments. For hormones that were dissolved in ethanol, controls (wounded and unwounded plants depending on the experiment) were treated with the same amount of 0.1% ethanol in sterile water. Leaf tissue was collected after 24 h.

Analysis of developmental and systemic expression

Maize plants of different vegetative stages (V2, V4, V6, V8, and V10; Ritchie et al., 1986) were infested with two or three third- or fifth-instar larvae per plant based on plant size. In each stage, two to six plants were used for control and feeding treatments and tissue adjacent to the feeding sites was collected and immediately frozen in liquid nitrogen.

For systemic analysis of RIP2 expression five fifth-instar larvae were placed in the center of unexpanded 11th leaf in the whorl of V8 stage plants for 24 h. A small, perforated plastic bag (Ziploc sandwich bag; SC Johnson, Racine, WI, USA) was placed on the tip of leaf blade around the larva and gently taped to the leaf to limit caterpillar movement. Leaf samples from the feeding site (0 cm) and distal to the site (3 and 5 cm) were collected. Samples were also taken from other leaves (unexpanded ninth and 10th leaves) of the same plant. Tissues from four to five different plants were harvested for control and fed plants.

Insect feeding

Plants were infested with three newly molted fifth-instar larvae for various time points. Unwounded plants were used as controls. A paper punch was used to mechanically wound the plants (multiple wounding sites per plants). Leaf tissue adjacent to the wound sites was collected and immediately frozen in liquid nitrogen and stored at −80°C.

RIP2 protein persistence after larval feeding

Plants were infested with newly molted fifth-instar larvae (one larva per plant) for 24 h and then removed. This was designated as 0 d and samples were collected at several time points (1–6 d) after larval removal. Leaf tissue adjacent to the feeding site was collected and immediately frozen in liquid nitrogen and stored at −80°C.

Quantitative reverse transcription polymerase chain reaction (qRT-PCR)

Total RNA from leaf tissues was isolated using the TRIzol Reagent (Invitrogen) and DNase (Promega) following the manufacturer's instructions. The first-strand cDNA was synthesized with the High Capacity cDNA Reverse Transcription Kit (ABI, Foster City, CA, USA) with oligo-(dT) 20 primers following the manufacturer's instructions. qRT-PCR was carried out in an ABI 7500 Fast Real Time PCR System. The primers were designed by Primer Express software for real-time PCR (version 3.0; ABI). The PCR conditions were as follows: step 1, 50°C for 2 min and 95°C for 10 min; step 2, 95°C for 15 s and 60°C for 1 min repeated 40 cycles; step 3, 72°C for 10 min; step 4, dissociation stage. Relative quantification of gene expression was analyzed by ABI 7500 Fast SDS Software (version 1.4; ABI). The data set was normalized using actin as a control. Gene-specific forward (F) and reverse (R) primers used to generate data present in this study were as follows: ACTIN-F (accession number: U60511.1) 5′- GGA GCT CGA GAA TGC CAA GAG CAG-3′, ACTIN-R 5′- GAC CTC AGG GCA TCT GAA CCT CTC-3′, RIP2-F (L26305) 5′-GAG ATC CCC GAC ATG AAG GA-3′, RIP2-R 5′-CTG CGC TGC TGC GTT TT-3′, RIP1-F(M83926) 5′-TGT GAT CCC CGA CAT GCA-3′, RIP1-R 5′-CGA TCC TCG CTG CTT CGT-3′. The result was analyzed by SAS statistical software (SAS).

Multidimensional protein identification technology (MudPIT) and frass protein identification

Frass proteins were extracted using a phenol-based extraction method described elsewhere (Chen et al., 2007) and quantified with a Bradford assay (Bradford, 1976). For mass spectrometry analysis, 100 μg of total protein were run 1 cm into a 10% denaturing polyacrylamide gel (Chen et al., 2005). The gel was stained with Coomassie Brilliant Blue and destained overnight. The piece of gel containing the proteins was excised and subjected to in-gel trypsin digestion according to Shevchenko's method with modifications (Mattiacci et al., 1995). Briefly, gel bands were dehydrated using 100% acetonitrile and incubated with 10 mM dithiothreitol in 100 mM ammonium bicarbonate, pH 8, at 56°C for 45 min, dehydrated again and incubated in the dark with 50 mM iodoacetamide in 100 mM ammonium bicarbonate for 20 min. Gel bands were washed with ammonium bicarbonate and dehydrated again. Sequencing grade, modified trypsin was prepared to 0.01 μg μl−1 in 50 mM ammonium bicarbonate and c. 50 μl of this was added to each gel band so that the gel was completely submerged. Bands were then incubated at 37°C overnight. The extracted peptides were resuspended in a solution of 2% acetonitrile/0.1% trifluoroacetic acid to 20 μl. From this, 10 μl were automatically injected by a Waters nanoAcquity Sample Manager ( and loaded for 5 min onto a Waters Symmetry C18 peptide trap (5 μm, 180 μm × 20 mm) at 4 μl min−1 in 2%ACN/0.1% formic acid. The bound peptides were then eluted using a Waters nanoAcquity UPLC (buffer A, 99.9% water/0.1% formic acid; buffer B, 99.9% acetonitrile/0.1% formic acid) onto a Michrom MAGIC C18AQ column (3 μ, 200A, 100 U × 150 mm, and eluted over 240 min with a gradient of 5% B to 30% B in 210 min at a flow rate of 1 μl min−1. Eluted peptides were sprayed into a ThermoFisher LTQ-FT Ultra mass spectrometer ( using a Michrom ADVANCE nanospray source. Survey scans were taken in the FT (25 000 resolution determined at m/z 400) and the top 10 ions in each survey scan were then subjected to automatic low-energy collision-induced dissociation (CID) in the LTQ. The resulting MS/MS spectra were converted to peak lists using BioWorks Browser v3.3.1 (ThermoFisher) using the default parameters and searched against an appropriate protein database using the Mascot searching algorithm, v. 2.2 or 2.2.03 with a fragment ion mass tolerance of 0.80 Da and a parent ion tolerance of ± 10.0 ppm allowing for up to two missed tryptic sites. Oxidation of cysteine (carbamidomethyl cysteine) as well as variable modification of methionine oxidation were specified in Mascot ( The Mascot output was analyzed using Scaffold (version Scaffold-01_07_00; Proteome Software Inc., Portland, OR, USA; to probabilistically validate protein identifications using the ProteinProphet computer algorithm (Lawrence & Novak, 2006). At the time of analysis, corn had no comprehensive proteome information available and therefore putative peptides were generated by six frame translation of the TIGR plant transcript assembly sequences (Kramer & Muthukrishnan, 1998; Database issue: D846-51. Protein identifications were accepted if they could be established at > 95.0% probability and contained at least two identified peptides.


Identification of maize proteins in insect frass

Since previous analysis of the M. sexta frass proteome focused on identification of herbivore defense proteins in the dicot tomato (Chen et al., 2007), we were interested in determining if there were potential defensive proteins in frass from caterpillars that fed on the monocot maize. Two maize genotypes were used: Tx601 is susceptible to FAW feeding (Williams et al., 1989), whereas Mp704 is highly resistant to a number of lepidopteran pests, including FAW (Williams & Davis, 1982). These plants were infested with FAW larvae for 24 h to trigger defenses and induced plants were used as the diet for a naïve set of caterpillars. Frass from these caterpillars was used for proteome determination.

Table 1 shows the most abundant proteins in frass from FAW reared on Tx601 or Mp708. In most cases, similar proteins were detected in each genotype. The complete list of proteins identified is shown in the Supporting Information, Table S1. The number of peptide counts in Tables 1 and S1 represents the number of peptides for a particular protein in the frass that were identified by mass spectrometry. These results supported the findings of Chen et al. (2007), demonstrating that potential plant defensive proteins can resist digestive proteases in the gut and be eliminated in the frass. One of the proteins found in the frass was a maize RIP protein. Two RIP genes are in the maize genome: RIP1 and RIP2. Since the RIP1 protein is kernel-specific and FAW larvae were only fed on leaves, it is likely that protein found in frass was the RIP2 protein isoform that is expressed in vegetative tissues (Bass et al., 2004). A comparison of the RIP2 protein amino acid sequence aligned with peptides found in frass confirmed this (Fig. S1).

Table 1. Maize proteins identified in frass of fall armyworm (FAW, Spodoptera frugiperda) fed on two maize genotypes (Tx601 and Mp704)
Protein name aAccession no.bNo. peptide (Tx601)No. peptide (Mp704)
  1. a

    This is a partial list of all of proteins identified. The complete list of protein identified is shown in Table S1.

  2. b

    The peptide sequences are blasted against the TIGR database.

Beta-D-glucosidase precursor (Zea mays (maize))TA105145_4577_c7975
Allene oxide synthase (Zea mays (maize))TA95323_4577_b3339
Lipoxygenase 1 (Oryza sativa (rice))AY110012_b3028
Ribosome-inactivating protein (Zea mays (maize))TA132043_4577_c2618
Putative fruit protein (Oryza sativa cv japonica)TA116740_4577_a2717
Putative endo-1,3-beta-glucanase (Oryza sativa cv japonica)TA133226_4577_b1415
Peroxidase (Zea mays (Maize))TA112628_4577_b2013
GDSL-like lipase/acylhydrolase family protein, expressed (Oryza sativa cv japonica)BG837878_c812
Chitinase (Zea mays ssp. parviglumis (Balsas teosinte))TA95600_4577_b1510
Leucine aminopeptidase 2, chloroplast precursor (Oryza sativa (Rice))TA112645_4577_b122

Two maize RIP genes respond to insect herbivory differently

Because the two maize RIP genes had differences in tissue-specific expression (Bass et al., 2004), we examined the accumulation of RIP1 and RIP2 transcripts in plants infested with FAW larvae using qRT-PCR and gene-specific primers. RIP2 transcripts were highly expressed in caterpillar-fed leaves, while those for RIP1 were undetectable in both control and 24 h-fed plants (Fig. 1). This provides additional evidence that the RIP2 protein is the isoform identified in frass (Table S1).

Figure 1.

Expression analysis of ribosome-inactivating protein 1 (RIP1) and RIP2 transcripts in maize (Zea mays) leaves in response to caterpillar feeding. Plants were infested with fall armyworm (FAW, Spodoptera frugiperda) larvae (three larvae per plant) for 24 h. Undamaged plants were used as the control. Total RNA was isolated from the feeding site and relative expression levels were determined using quantitative reverse transcription polymerase chain reaction (qRT-PCR). Gene expression was normalized to that of actin (= 3–4; error bar, + SE). Letters indicate significant differences by least significant difference test (< 0.05). ND, not detected.

RIP2 protein accumulates in maize leaves and FAW frass

Because the RIP2 protein is synthesized as an inactive proenzyme (Bass et al., 2004), the form of RIP2 protein present in FAW-fed leaves and caterpillar frass was determined using immunoblot analysis. The RIP2 proenzyme with a predicted size of c. 30 kD was detected in the whorl (Fig. 2), but in the frass there was a 12 kD immunoreacting polypeptide that was most probably the processed form of RIP2 (pRIP2; Fig. 2). This band was absent when larvae were fed on an artificial diet, which indicated that it must have been ingested from the maize leaves.

Figure 2.

Immunoblot analysis of ribosome-inactivating protein 2 (RIP2) in maize (Zea mays) leaves. Plants were infested with fall armyworm (FAW, Spodoptera frugiperda) larvae (three larvae per plant) for 24 h. 1, proteins extracted from the feeding site; 2, proteins extracted from frass of S. frugiperda fed on induced plant material. Equal amounts of protein (5 μg) were loaded in each well. Proteins were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and analyzed on immunoblots using anti-RIP antibody. The polypeptide labeled RIP2 corresponds to the 30 kD proenzyme form of RIP2. The polypeptide labeled pRIP2 corresponds to 12 kD processed form of RIP2.

To determine if the proenzyme form of the RIP2 protein is commonly induced in insect-fed maize, 13 inbred lines and two teosinte subspecies were tested for RIP2 protein expression. The proenzyme form of RIP2 was detected in all genotypes (Fig. S2). However, the size of the RIP2 proenzyme varied among the inbreds, which could be due to allelic differences in protein sequences or post-translational modifications. Although control samples were not analyzed for each genotype, the RIP2 protein was induced by FAW feeding B73, Mp708 and Tx601. These results suggest that RIP2 protein expression in the whorl in response to insect feeding is a widespread phenomenon in maize.

Insect herbivory, not mechanical wounding, induces RIP2 protein

Because herbivore feeding is a combination of mechanical wounding and deposition of oral secretions, we examined the induction of RIP2 transcripts and RIP2 protein expression in response to mechanical wounding and FAW feeding. Maize plants were infested with FAW larvae for 24 h or mechanically wounded with a paper punch. Fig. 3(a) shows that RIP2 transcripts increased c. 100-fold in abundance after herbivore attack. However, plants without insect feeding or that were mechanically wounded did not accumulate RIP2 transcripts. Immunoblot analysis showed that the RIP2 protein also accumulated abundantly in response to herbivory, but far less in response to wounding (Fig. 3b). Therefore it appears that caterpillar feeding is needed to significantly trigger RIP2 protein expression but mechanical wounding is not.

Figure 3.

Ribosome-inactivating protein 2 (RIP2) transcript and protein expression in maize (Zea mays) leaves in response to various treatments. Plants were wounded with a paper punch (W) or infested with fall armyworm (FAW, Spodoptera frugiperda) larvae (F) (three larvae per plant) for 24 h. Undamaged plants were used as a control (C). (a) To determine RIP2 transcript abundances, total RNA was isolated from the wound site on the leaves. Relative expression levels were determined using quantitative reverse transcription polymerase chain reaction (qRT-PCR) and data were normalized to the expression of the reference gene, actin (= 4, error bars, + SE). Letters indicate significant differences by least significant difference test (< 0.05). (b) To determine the accumulation of RIP2 protein, leaf proteins were isolated and analyzed on immunoblots using anti-RIP antibody. Equal amounts of protein (50 μg) were loaded in each well. In addition to detection of RIP2 in the leaf, RIP1 was also detected in the kernel (K). A Ponceau S stained membrane showing the 55 kD Rubisco large subunit was the loading control.

Characterization of RIP2 transcripts and RIP2 protein induction in response to FAW feeding

Since FAW feeding triggered the accumulation of both RIP2 transcripts and RIP2 protein, we determined the rapidity of this response. RIP2 transcript abundance increased slightly 30 min after insect feeding, but were not significantly higher than the control. However, there was a significant increase in transcript abundance between 1 and 4 h and they remained high at 24 h (Fig. 4a). Immunoblot analysis indicated that the RIP2 protein was present in low amounts before FAW feeding and increased in abundance at 1 h and remained relatively constant up to 24 h (Fig. 4b). These data indicated that continuous caterpillar feeding induced the accumulation of both RIP2 transcripts and RIP2 protein.

Figure 4.

Time course of ribosome-inactivating protein 2 (RIP2) expression in response to feeding by fall armyworm (FAW, Spodoptera frugiperda) larvae (one larva per plant). (a) RIP2 transcript abundances were determined by quantitative reverse transcription polymerase chain reaction (qRT-PCR). Total RNA was isolated from feeding sites and unwounded maize (Zea mays) plants (0 h, control). Gene expression was normalized to that of actin (= 4; error bars, + SE). Letters indicate significant differences by least significant difference test (< 0.05). (b) RIP2 protein accumulation was determined by immunoblot analysis using anti-RIP antibody. Proteins were extracted from leaves at the larval feeding sites. Uninfested plants were collected as a control (0 h). Each time point represents four biological samples. Equal amounts of protein (50 μg) were loaded in each well. A Ponceau S stained membrane showing the 55 kD Rubisco large subunit was the loading control.

To test the stability of RIP2 protein in the leaf, plants were infested with FAW larvae (one larva per plant) for 24 h to induce RIP2 protein expression. Then the larva was removed and leaf samples were collected at subsequent time points. After the initial 24 h of feeding, RIP2 protein was abundant in the leaf up to 4 d after caterpillar removal, but dramatically decreased at 5 d (Fig. 5). These results indicate that the RIP2 protein is highly stable in the plant even in the absence of herbivore feeding.

Figure 5.

Immunoblot analysis of ribosome-inactivating protein 2 (RIP2) in maize (Zea mays) leaves after larval removal. Maize plants were infested with fall armyworm (FAW, Spodoptera frugiperda) larvae (one larva per plant) for 24 h. The larvae were removed and this was designated as 0 d. Samples from other time points (up to 6 d) were collected. Proteins were extracted from leaves at the feeding sites. Uninfested plants were collected as controls. The control samples (C0 and C6) were collected at the same time at 0 and 6 d. Each time point represents two biological samples. Equal amounts of protein (50 μg) were loaded in each well. A Ponceau S stained membrane showing the 55 kD Rubisco large subunit was the loading control.

RIP2 protein is expressed locally during maize vegetative development

In dicot plants, such as tomato, there is often systemic induction of defense proteins such as PIN1 (Montesano et al., 2003). To determine if FAW feeding could systemically induce RIP2 protein accumulation, larvae were placed in ‘cages’ on the maize leaves to limit their movement. Several leaf samples were collected distally from the feeding site and non-fed leaves. Immunoblots showed that RIP2 protein accumulation was induced only near the feeding site (Fig. S3) and there was no widespread systemic induction of RIP2 protein.

The vegetative portion of the maize life cycle can be roughly divided into two stages, juvenile and adult. The juvenile stage persists from germination until the plants have five to six leaves, when there is a transition to the adult stage (Freeling, 1992). Maize plants that undergo this transition earlier in their life cycle are more resistant to caterpillar feeding (Williams et al., 1998; Brooks et al., 2007). To determine if there were temporal differences in RIP2 protein expression during maize development, samples were collected from whorls at several vegetative stages and analyzed for RIP2 protein expression. Fig. S4 shows that RIP2 protein accumulation was induced by FAW feeding in the juvenile stages (V2 and V4), during the transition from juvenile to adult (V6), and in the adult (V8, V10) vegetative stages. This indicates that the RIP2 protein can protect maize against herbivory during all of the vegetative stages tested.

Phytohormones regulating RIP2 protein expression

Several phytohormones, salicylic acid (SA), JA, and ET are known to regulate plant responses to biotic stresses (Park et al., 2002). Since FAW feeding induces RIP2 protein expression, it is likely that one or more of these phytohormones is involved in its regulation. When unwounded plants were treated with MeSA, MeJA, and ethephon (to generate ET), there was no induction of RIP2 protein expression (data not shown). To determine if physical damage was required in addition to phytohormone treatment, plants were mechanically wounded before hormone application. In some cases, the wounded control plants showed a slight induction of RIP2 protein expression, but this generally occurred when 0.1% ethanol was included in the buffer as in the MeJA treatment. Plants that were wounded and treated with ethephon accumulated RIP2 protein (Fig. 6a). Although JA is usually involved in regulating defenses against chewing insect herbivores, MeJA treatment after wounding at a concentration of 0.01% failed to trigger RIP2 protein expression (Fig. 6b). However, the combination of wounding, MeJA and ethephon induced RIP2 protein accumulation. These results suggest that ET is involved RIP2 protein expression and the combination of these two phytohormones could not induce RIP2 protein expression in the absence of mechanical wounding.

Figure 6.

Effect of exogenous phytohormones (ethylene and methyl jasmonate) on ribosome-inactivating protein 2 (RIP2) expression in maize (Zea mays) leaves after 24 h treatment. (a) Ethylene: maize plants were damaged by a paper punch (wounded only, W). Wounded plants were sprayed with 3 mM 2-chloroethylphosphonic acid (ethephon) (wounded + ethephon, WE). Undamaged plants without ethephon treatment were the controls (control, C). Leaf samples were collected after 24 h of treatment. (b) Methyl jasmonate (MeJA): maize plants were damaged by paper punch (W). Wounded plants were sprayed with 0.01% MeJA and 3 mM ethephon (wounded + MeJA, WJ; wounded + MeJA + ethephon, WJE). Undamaged plants were also treated with 0.01% MeJA and 3 mM ethephon (MeJA + ethephon, JE). Undamaged plants with no treatment acted as controls (C). Each treatment represents two biological samples. Equal amounts of protein (50 μg) were loaded in each well. Ponceau S stained membranes showing the 55 kD Rubisco large subunit were the loading controls.

Analysis of the RIP2 promoter sequence from the maize genome database (Schnable et al., 2009) using PLANTCARE (Lescot et al., 2002) predicted the presence of several cis-acting regulatory elements that should respond to MeJA. A previous study reported that the same MeJA concentration (0.01%) used here induced the expression of another maize insect defense protein, Mir1-CP, in the genotype Mp708 (Ankala et al., 2009). However, Mp708 has constitutively elevated concentrations of JA and application of a lower exogenous MeJA concentration was sufficient to stimulate defense protein accumulation in this inbred (Shivaji et al., 2010). Therefore, MeJA concentrations from 0.03 to 0.1% were used to determine its effect on RIP2 protein expression. Increasing the MeJA concentration to 0.03% slightly enhanced RIP2 protein expression in wounded plants; however, only plants treated with the highest concentration of MeJA (0.1%) induced RIP2 protein expression in the absence of wounding (Fig. S5a). MeSA did not induce RIP2 protein expression, even in combination with mechanical wounding (data not shown). As a previous study reported that RIP2 protein could be induced by water deficit (Bass et al., 2004), plants were treated with ABA, which signals drought stress (Jeffers et al., 2005). ABA with wounding treatment induced RIP2 protein expression (Fig. S5b) and confirmed the results of Bass et al. (2004).

Effect of heterologously expressed RIP2 on FAW larval growth

Because the RIP1 protein has been shown to reduce the growth of several insect pests and fungal pathogens, including Aspergillus flavus (Dowd et al., 1998, 2012; Nielsen et al., 2001), we determined whether the RIP2 protein was toxic to FAW. To do this, we heterologously expressed and purified recombinant RIP2 (rRIP2) protein. As previous studies indicated that RIP1 protein is processed into two polypeptides (Nielsen & Boston, 2001) and that rRIP2 protein is processed by papain (Bass et al., 2004), purified rRIP2 protein was treated with papain-coated beads to generate the processed form, prRIP2. To determine if the rRIP2 protein was processed in the caterpillar midgut, FAW larvae were fed on a diet containing rRIP2 protein and their frass was collected. Fig. 7 shows immunoblot analysis of various forms of the rRIP2 protein. Lane 1 shows rRIP2 protein with an apparent molecular mass of 37 kD, which is c. 6.5 kD larger than the plant form because the rRIP2 protein is a fusion protein containing 34 histidine residues and 11 additional amino acids in N-terminal leader sequence (Bass et al., 2004). Lane 2 shows purified rRIP2 protein treated with papain-coated beads and indicates that rRIP2 protein is cleaved into two proteins of c. 15 and 12 kD. Sequencing of these polypeptides using mass spectrometry indicated that the upper and lower bands corresponded to the N- and C-terminal polypeptides, respectively (Fig. S6). When frass proteins from larvae-fed purified rRIP protein were analyzed by immunoblot, only the lower 12 kD band was apparent (Fig. 7, lane 3). Hence, it appears that rRIP2 protein can be processed by the FAW digestive system in a manner similar to the papain-coated beads. However, it is not clear why the 15 kD band was not detected on the immunoblot. It is possible that this polypeptide is degraded in the midgut into peptides that are too small to be retained in the SDS-PAGE gel. This could also occur when native leaf RIP2 protein moves through the FAW midgut, because the upper band was not detected in this immunoblot (Fig. 2, lane 2).

Figure 7.

Immunoblot analysis of recombinant ribosome-inactivating protein 2 (rRIP2) digested by papain-coated beads and fall armyworm (FAW, Spodoptera frugiperda) larvae. 1, rRIP2 (0.25 μg); 2, rRIP2 incubated with papain-coated beads (0.15 μg); 3, FAW frass from larvae fed on an artificial diet containing rRIP2 (1.9 μg). The resulting protein was separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and analyzed on an immunoblot using anti-RIP antibody. The 37 kD polypeptide corresponds to rRIP2 (lane 1). The two small polypeptides correspond to processed forms of rRIP2.

Before feeding bioassays, the amount of RIP2 protein expressed in plants exposed to larval feeding was estimated by analyzing serial dilutions of plant extracts on immunoblots (data not shown). We determined that the RIP2 protein concentration after 24 h of FAW feeding was c. 0.8 μg (100 mg)–1 FW. This concentration of either rRIP2 or prRIP2 protein was used for bioassays. After 5 d both rRIP2 and prRIP2-fed larvae had lower weights than those fed BSA or buffer (Fig. 8). A comparison of RIP2 protein-fed larvae with controls indicated that their weight was suppressed by 26%. These results indicate that both forms of the RIP2 protein are toxic to FAW larvae at a concentration typically found in caterpillar-fed whorls.

Figure 8.

Effect of recombinant ribosome-inactivating protein 2 (rRIP2) on fall armyworm (FAW, Spodoptera frugiperda) larval growth. BSA, phosphate-buffered saline (PBS; buffer), rRIP2, and RIP2 processed with papain (prRIP2, see the 'Materials and Methods' section) were placed into an artificial diet at a ratio of 0.8 μg protein (100 mg)–1 diet. At 1 d after hatching, neonate larvae were fed with protein/diet for 5 d and the final weight of larvae was determined. Letters indicate significant differences by least significant difference test (< 0.05) (= 21–24; error bars, + SE).


Plants have evolved with a number of defense mechanisms to protect themselves against insect herbivory. Because herbivores consume foliage that is used for growth and development, plants respond to herbivory by synthesizing a number of antinutritional substances (Berenbaum, 1995; Felton & Gatehouse, 1996; Felton, 2005; Zhu-Salzman et al., 2008). When insects ingest this cocktail of antinutritional proteins, it causes ‘indigestion’, limits their ability to fully utilize plant nutrients and impairs their growth (Felton, 2005). Studies have shown that some ingested plant defensive proteins remain intact in the insect gut and are eliminated in the frass (Chen et al., 2005, 2007; Jeffers et al., 2005; Zhu-Salzman et al., 2008).

Some of the maize proteins identified in FAW frass in this study have been implicated in plant defense against herbivores and pathogens. Beta-D-glucosidase has been shown to activate DIMBOA, a secondary metabolite toxic to the European corn borer (Ostrinia nubilalis; Yu et al., 2009). Lipoxygenase (LOX) and allene oxide synthase (AOS) catalyze steps in the JA biosynthetic pathway and the production of JA in response to herbivory triggers many plant defenses against insects (Howe & Jander, 2008). These results are supported by a previous study demonstrating that transcripts for two maize LOX genes (ZmLOX1 and ZmLOX3) and AOS increased in abundance when Mp708 and Tx601 plants are subjected to FAW feeding (Shivaji et al., 2010). In addition, maize 9-LOX (ZmLOX3) plays a defensive role against nematodes (Meloidogyne incognita) and fungi (Aspergillus flavus and Aspergillus nidulan; Gao et al., 2008, 2009). A putative fruit protein with unknown function that shares similarity with a pathogen-responsive oxidoreductase (drd-1) in potato (Montesano et al., 2003) was also present in the frass proteome. Another protein was an endo-beta-glucanase. In rice, an isoform of this enzyme responded to wounding, MeJA, and ethephon (Akiyama et al., 2009), which suggests its possible function in herbivore defense. Another protein that was identified was peroxidase. In maize, peroxidases have been shown to be associated with disease and caterpillar resistance (Dowd & Johnson, 2005; Dowd et al., 2010). The growth rates of two major maize pests (Helicoverpa zea and Lasioderma serricorne) were retarded when they were fed the maize peroxidase px11 (Chen et al., 2008). In addition to these proteins, a GDSL-like lipase (GLIP1) present in the frass proteome is similar to one in Arabidopsis that is involved in defense against the necrotrophic fungus Alternaria brassicicola (Oh et al., 2005). GLIP1, which is regulated by ET, triggers systemic resistance signaling in plants after fungal infection (Oh et al., 2005). Another Arabidopsis GDSL-like lipase (GLIP2) plays a role in plant defense against pathogens (Lee et al., 2009). Chitinases, which also were found in frass, have been shown to be defensive proteins in plants. For example, the overexpression of poplar chitinase (WIN6) in tomato retarded the development of Colorado potato beetle (Lawrence & Novak, 2006) and ectopic expression of a rice chitinase in peanuts enhanced resistance to Cercospora arachidicola (Iqbal et al., 2012). In addition, transcripts of the chitinase gene Prm3 have been shown to increase in response to FAW feeding in maize (Shivaji et al., 2010). One possible function of chitinases is to catalyze the disruption of the chitin-rich peritrophic matrix of the attacking herbivore. We also identified several peptides of leucine aminopeptidase in frass (Table 1). Tomato leucine aminopeptidase A (LapA) increases in response to wounding, exogenous MeJA, pathogen infection, and insect feeding (Chao et al., 1999; Chen et al., 2005, 2007; Zhu-Salzman et al., 2008). LapA protein has been detected in the midgut and frass of M. sexta (Chen et al., 2005, 2007), and its overexpression in tomato delays M. sexta growth and development (Lee et al., 2009). These proteomic studies indicated that a rich cocktail of putative maize herbivore defense proteins was present in FAW frass.

RIP genes are ubiquitous in the plant kingdom and more than 130 RIP genes have been identified (Girbes et al., 2004; Jiang et al., 2008). Expression and accumulation of these RIP-like proteins is regulated by a plethora of developmental, abiotic and biotic factors (Nielsen & Boston, 2001). In maize, the kernel isoform of the RIP1 protein has been extensively studied (Bass et al., 1992) and transgenic plants overexpressing this gene have been shown to inhibit insect growth (Dowd et al., 2003). In a more recent study, Dowd et al. (2012) reported that maize transformed with a single construct containing the coding sequences for the RIP1 protein, wheat germ agglutinin and tobacco hornworm chitinase inhibited the growth of H. zea and FAW larvae. But this construct is clearly more complicated than the native RIP2 gene sequence. Because the RIP1 protein is only expressed in the kernel, it is unlikely that it defends maize against leaf-feeding herbivores. The RIP2 protein, on the other hand, is induced by FAW feeding and inhibits growth of the herbivore. In addition, RIP1 and RIP2 share only 73% protein sequence similarity (Bass et al., 1995), so it is possible that these differences could result in differential degrees of toxicity between the two proteins.

Here we demonstrate that both the RIP2 protein and its transcripts are induced by FAW feeding in the whorl. Furthermore, we have shown that mechanically wounded plants treated with ET and JA, phytohormones known to be involved in signal transduction pathways leading to caterpillar defenses in maize (Harfouche et al., 2006; Ankala et al., 2009; Shivaji et al., 2010), expressed enhanced levels of RIP2. These results, in conjunction with its indigestibility, strongly suggested that it could be involved in herbivore defense. To test this, FAW larvae were fed both unprocessed and processed recombinant RIP2 protein in a physiologically relevant concentration. The results indicated that both forms inhibited FAW growth by c. 25%. In addition, peptide fragments of the recombinant RIP2 protein detected in the frass were similar in size to those present in frass from caterpillars that fed on maize. Only a few studies have investigated the mechanism of RIP protein toxicity to insects. SNA-I from elderberry caused cell apoptosis in the gut tissues of Acyrthosiphon pisum and Spodoptera exigua (Shahidi-Noghabi et al., 2010b). SNA-I and SNA-II also induced caspase-3 like activity in the midgut cell line of Lepidoptera (Shahidi-Noghabi et al., 2010a). These findings suggest that RIP proteins are toxic to insect herbivores because they trigger apoptosis in the midgut; however, the mechanism of RIP2 protein action on the FAW midgut has not yet been investigated.

The rapid induction of RIP2 transcripts and RIP2 protein in response to caterpillar feeding could be an early deterrent against insect herbivory. In the caterpillar-resistant maize inbred Mp708, another defensive enzyme, the cysteine protease Mir1-CP, accumulates within 1 h after insect attack (Pechan et al., 2000). Other genes that are rapidly induced in response to wounding in maize include wound-induced proteinase inhibitor (WIP1), as early as 30 min after wounding (Rohrmeier & Lehle, 1993), and maize proteinase inhibitor (MPI) at 20 min after wounding (Tamayo et al., 2000). Therefore, it appears that herbivore feeding rapidly activates a suite of defensive proteins in maize.

In addition to being rapidly induced, the RIP2 protein persists in the leaves up to four d. One explanation for its long half-life is that it responds to the insect feeding behavior. Caterpillars do not continuously feed on plants and eating bouts are often separated by multiple gaps (Reynolds et al., 1986). During these gaps, they need time to digest leaf tissue or prepare to molt. Because molting in the field usually takes 1 or 2 d, plants need to maintain their defenses for the next insect attack; therefore the RIP2 protein could be deployed for long-term defense. In addition, it could protect the plant from the occurrence of fungal infection in the open wound sites caused by herbivory. Because the kernel RIP1 protein inhibits fungal growth, it is possible that the foliar RIP2 protein might function in a similar manner (Nielsen et al., 2001).

We also observed that the size of unprocessed RIP2 proteins varied among the inbred lines tested (Fig. S2). However, a comparison of the cDNA sequences of RIP2 from Mp708, Tx601 and B73 indicated that the derived amino acid sequences were nearly identical. So, it seems unlikely that minor changes in the amino acid sequence would significantly alter their size. Therefore, size differences could be due to some type of post-translational modification.

As mentioned previously, RIP1 and RIP2 map to different physical locations on the maize genome. RIP2 is on chromosome 7 whereas RIP1 is located on chromosome 8 (Bass et al., 1995, 2004). Maize chromosomes 1, 5, 7, and 9 contain major loci for insect resistance to FAW and the southwestern corn borer (Diatraea grandiosella; Brooks et al., 2007). The RIP2 gene lies within strong quantitative trait locus for caterpillar resistance in maize (Brooks et al., 2005, 2007). This, in conjunction with its ability to impair FAW growth by c. 25%, suggests that the RIP2 protein is an important member of a group of herbivore-induced resistance proteins in maize.


We would like to thank Ms Paige Byrns and Penn State University College of Medicine Mass Spectrometry and Proteomics Facility for their technical support, Dr Paul W. Williams for generously supplying maize seeds, and Dr Sarah M. Assmann for ABA reagents. We are especially grateful to Dr Rebecca Boston for supplying anti-RIP antibody and the RIP2 expression vector. This work was supported by grants from the National Science Foundation (IOS-0641219 awarded to D.S.L.) and the US Department of Agriculture (2007-35604-17791 awarded to G.A.H.).