Differential alteration of two aminopeptidases N associated with resistance to Cry1Ac in cabbage looper

Kasorn Tiewsiri and Ping Wang1

Department of Entomology, Cornell University, New York State Agricultural Experiment Station, Geneva, NY 14456

Edited* by May R. Berenbaum, University of Illinois at Urbana–Champaign, Urbana, IL, and approved July 22, 2011 (received for review February 19, 2011) The soil bacterium Bacillus thuringiensis (Bt) is the most success- “Mode 1” type resistance is the most common type of re- fully used biopesticide in agriculture, and its insecticidal protein sistance to Bt observed in insects. It is characterized by a high are the primary transgenes used for insect control in trans- level of resistance (>500-fold) to at least one Cry1A toxin but not genic crops. However, evolution of insect resistance to Bt to Cry1C toxins, recessive inheritance, and reduced binding of threatens the long-term future of Bt applications. To date, cases of one or more Cry1A toxins to the midgut brush border membrane resistance to Bt toxins have been reported in agricultural situa- (8). Mode 1 type resistance has been associated with mutations tions in six insect species, but the molecular basis for these cases of the midgut cadherin in three Lepidoptera: H. virescens, of resistance remains unclear. Here we report that the resistance Pectinophora gossypiella, and Helicoverpa armigera (9–11). The to the Bt toxin Cry1Ac in the cabbage looper, Trichoplusia ni, midgut cadherin has been suggested to serve as the first Cry toxin evolved in greenhouses, is associated with differential alteration receptor in the sequential interactions of Cry toxins with midgut of two midgut aminopeptidases N, APN1 and APN6, conferred by brush border membrane proteins (12, 13). Consequently, muta- a trans-regulatory mechanism. Biochemical, proteomic, and molec- tions in the cadherin gene may result in Bt resistance in insects ular analyses showed that in the Cry1Ac-resistant T. ni, APN1 (9–11). The field- and greenhouse-evolved resistance to Cry1Ac was significantly down-regulated, whereas APN6 was significantly in Plutella xylostella and T. ni is of the typical mode 1 type. up-regulated. The Cry1Ac resistance was correlated with down- However, the cadherin gene in P. xylostella or T. ni is not in- regulation of APN1 but not with the up-regulation of APN6. The volved in the resistance mechanism (14, 15), indicating that the concurrent up-regulation of APN6 and down-regulation of APN1 mode 1 type resistance selected in agricultural systems involves might play a role in compensating for the loss of APN1 to minimize a different but yet to be known molecular genetic mechanism. the fitness costs of the resistance. Along with identifying reduced The Cry1Ac resistance that has evolved in T. ni populations in expression of APN1 as the molecular basis of Bt resistance selected commercial greenhouses is monogenic and recessive in in- in an agricultural setting, our findings demonstrate the impor- heritance and is conferred by loss of toxin-binding sites in the tance of APN1 to the mode of action of Bt toxin Cry1Ac. larval midgut brush border membrane (16, 17). In this study, the greenhouse-derived Cry1Ac-resistant T. ni (16) was used as he soil bacterium Bacillus thuringiensis (Bt) has long been a unique system to identify the alteration of Cry1Ac-binding Tused as a biopesticide for insect control in agriculture and proteins and its association with the resistance to Cry1Ac. This public health (1, 2). Bt genes coding for insecticidal toxins are study reports the identification of the molecular basis of insect the primary transgenes in current transgenic crops (Bt crops) resistance to Bt toxins evolved in an agricultural system. conferring insect resistance (3). With the increasing scale and prolonged planting of Bt crops, evolution of insect resistance to Results Bt toxins in agricultural systems has become the foremost threat Alteration of Midgut Brush Border Membrane Vesicle Proteins in SCIENCES AGRICULTURAL to the long-term future of Bt biopesticides and Bt crops. To date, Cry1Ac-Resistant Larvae Identified by SDS/PAGE Separation. Proteins cases of insect resistance to Bt toxins in open fields or green- from the midgut brush border membrane vesicle (BBMV) pro- houses have been reported in six species (2). teins of the susceptible Cornell strain and the resistant GLEN- fi Cry toxins are the major insecticidal proteins in Bt. Once Cry1Ac-BCS strain exhibited highly similar protein pro les on ingested by insects, the Cry protoxins are activated by the insect SDS/PAGE analysis, except for the absence of a protein band at digestive proteases. The active toxins penetrate through the insect 110 kDa in BBMVs from resistant larvae (Fig. 1). In this band, APN1 was the primary protein with 8 and 46 minor proteins, midgut peritrophic membrane and reach the midgut brush border – membrane, where they interact with specific binding sites, leading respectively, determined by nano liquid chromatography tan- to cell lysis after a multistep cascade that is incompletely un- dem mass spectroscopy (nano LC-MS/MS) analyses with two derstood. The Bt pathogenesis pathway is complex, and thus different mass spectrometry systems, the Synapt HDMS mechanisms of Bt resistance can be diverse. To date, numerous (Waters) and the LTQ Orbitrap Velos (Thermo-Fisher Scien- tific). The relative abundance of APN1 in this 110-kDa band was insect strains resistant to Bt toxins have been established by se- estimated as 77 and 42 mol %, respectively, with the data lection with Bt toxins under laboratory conditions (4), and various obtained from the two mass spectrometry systems using expo- resistance mechanisms have been reported in laboratory-selected nentially modified protein abundance index (emPAI) values of Bt-resistant insects, including alterations of midgut digestive pro- teases, decreased peritrophic membrane permeability, height- ened immune response, enhanced esterase production, and re- Author contributions: K.T. and P.W. designed research; K.T. performed research; K.T. and duced Cry toxin binding (1, 2, 5). More recently, mutation of an P.W. analyzed data; and K.T. and P.W. wrote the paper. ABC transporter in Heliothis virescens was found to correlate with The authors declare no conflict of interest. resistance to Cry1A toxins (6). However, it has become clear that *This Direct Submission article had a prearranged editor. resistance mechanisms in laboratory-selected Bt-resistant insects Data deposition: The DNA sequences reported in this paper have been deposited in the do not necessarily represent Bt resistance mechanisms evolved in GenBank database (accession nos. JF303656–JF303663). fi eld populations (7). Molecular mechanisms of Bt resistance 1To whom correspondence should be addressed. E-mail: [email protected] or pingwang@ evolved in agriculture have not yet been reported. cornell.edu.

www.pnas.org/cgi/doi/10.1073/pnas.1102555108 PNAS | August 23, 2011 | vol. 108 | no. 34 | 14037–14042 Downloaded by guest on September 26, 2021 A B could bind to multiple proteins on the membrane blot, and that the patterns for the samples from the resistant and susceptible MW (kDa) larvae were similar, except for the lack of a positive band at 110 (APN6) 250 kDa from the resistant larvae (Fig. 2A). Identification of the 0.03% 150 0.2% 0.03% (APN6) proteins in this 110-kDa band by nano LC-MS/MS confirmed 1.1% 100 APN1 (77%, 42%) 1.3% that APN1 was the primary protein (77 mol %). The positive 2.1% 75 2.7% band at 200 kDa was not visible on the SDS/PAGE gel stained with Coomassie blue, but analysis of the gel slice corresponding 50 to the position of the positive band on the membrane blot showed that the band contained cadherin. The primary protein 37 in the >200-kDa positive band was identified as polycalin, which accounted for 26 mol % of the proteins in the band by emPAI-

25 based estimation. Western blot analysis of the midgut BBMV proteins confirmed that the 110-kDa Cry1Ac-binding protein in the midgut BBMVs was APN1, and that this 110-kDa APN1 was lacking in the resistant larvae (Fig. 2B). Fig. 1. SDS/PAGE analysis of midgut BBMV proteins from Cry1Ac-resistant and -susceptible larvae. (A) BBMV proteins from the resistant larvae lacked Transcript Levels of APN1 and APN6 Genes Are Differentially Altered a 110-kDa protein band. APN1 was identified to be the primary protein in in Cry1Ac-Resistant Larvae. Quantitative RT-PCR (qRT-PCR) the band, accounting for 77 and 42 mol %, respectively, by two nano LC-MS/ analysis revealed a significantly lower transcript level of apn1 in MS systems. (B) APN6 was identified in 6 of the 8 protein bands from the the resistant strain, with a ratio of 0.026 to the transcript level of resistant larvae with relative abundance ranging from 0.03 to 2.7 mol % and apn1 in the susceptible strain. In contrast, the transcript level of in 1 of the 17 protein bands from the susceptible larvae with a relative apn6 in the resistant strain was significantly up-regulated, with abundance of 0.03 mol %. Arrows indicate the protein bands analyzed by a ratio of 39 to that in the susceptible strain. The transcript levels nano LC-MS/MS for protein identification. of apn2, apn3, apn4, and apn5 did not differ significantly between the susceptible and resistant larvae (Fig. 3). the identified proteins [protein content (mol %) = emPAI/ ∑(emPAI) × 100] (18). Further identification of proteins from APN1 and APN6 Genes Show No Genetic Linkage with Cry1Ac 17 protein bands ranging from 33 to >250 kDa from the sus- Resistance. For linkage analysis of apn1 and apn6 with Cry1Ac ceptible strain and 8 matching bands ranging from 88 to >250 resistance, a single-pair cross between a male from the GLEN- Cry1Ac-BCS strain (apn1Gapn1G apn6Gapn6G) and a female kDa from the resistant strain showed that APN6 was widely from the Benzon strain (apn1Bapn1B apn6Bapn6B) was done present in the protein bands analyzed from the resistant strain, in to produce F progeny (apn1Gapn1B apn6Gapn6B), after which abundances ranging from 0.03 to 2.7 mol %. However, among 1 a female F progeny was backcrossed with a male from the GLEN- the 17 protein bands analyzed from the susceptible strain, only 1 Cry1Ac-BCS strain to produce backcross progenies. Genotyping one band minimally visible on the gel at 150 kDa was found to of 12 Cry1Ac-selected larvae and 12 Cry1Ac-nonselected larvae contain APN6 at low abundance (0.03 mol %) (Fig. 1B). from the backcross family showed that in both the Cry1Ac- Alteration of Midgut BBMV Proteins in Cry1Ac-Resistant Larvae selected and -nonselected groups, six individuals had the genotype apn1Gapn1B apn6Gapn6B, and the other six individuals had the Identified by Liquid-Based Quantitative Proteomic Analysis. Com- apn1Gapn1G apn6Gapn6G apn1Gapn1G parative proteomic analysis of midgut BBMV proteins from the genotype . The ratios of to apn1Gapn1B and of apn6Gapn6G to apn6Gapn6B were 1:1, as susceptible and resistant larvae by isobaric tagging for relative and would be expected with perfect random assortment (P > 0.10 by absolute quantification (iTRAQ) identified 1,464 and 1,440 pro- the χ2 test), in both the Cry1Ac-selected and -nonselected groups, teins at a 95% confidence interval in two independently prepared demonstrating that apn1 and apn6 did not cosegregate with the sample sets, including 908 proteins identified in both sample sets. resistance trait. This indicates that there is no genetic linkage Of these 908 proteins, those containing at least one unique pep- between apn1 and apn6 with Cry1Ac resistance. Neither genotype tide (496 proteins) were used for quantitative analysis to ensure G G G B G B G G apn1 apn1 apn6 apn6 nor apn1 apn1 apn6 apn6 was quantitative data reliability. Based on the variations observed present in the 24 individuals analyzed, indicating that apn1 and between the two sample sets [Δlog2 ratio = │log2 ratio(sample set 1) − log2 ratio(sample set 2)│], the iTRAQ quantitative ratio [│log2 (115 + 117)/(114 + 116)│] cutoff point was determined to be 1.5 (i.e., 95% of the proteins identified fell within a deviation A B of 1.5). With a significance cutoff threshold of 1.5, only two pro- teins, APN1 and APN6, were identified to be significantly differ- MW MW ent in quantity between the susceptible and the resistant strains kDa Polycalin kDa in both sample sets. The log2ratio of APN1 in the BBMV proteins between the resistant and susceptible strains was found to be −3.1 200 Cadherin 200 and −3.3 in the two sample sets. Therefore, the ratio of APN1 between the resistant and susceptible strains was 0.11 (quantitative 115 APN1 115 APN1 ratio = 2[log2ratio(sample set 1) + log2ratio(sample set 2)]/2). In contrast, the quantity of APN6 was higher in the BBMV proteins of the resistant strain. The log2ratio of APN6 between the resistant and susceptible strains was 2.6 in both sample sets; therefore, the ratio of APN6 Fig. 2. Cry1Ac toxin overlay binding analysis and identification of toxin- between the resistant and susceptible strains was 6.0. binding proteins from midgut BBMV proteins. (A) Three Cry1Ac-binding proteins from the midgut BBMV proteins were identified as polycalin, cad- Midgut BBMV Proteins from Cry1Ac-Resistant Larvae Lack the 110- herin, and APN1 by nano LC-MS/MS. The 110-kDa Cry1Ac-binding protein kDa Cry1Ac- Binding Protein APN1. A Cry1Ac toxin overlay binding APN1 was lacking in the resistant strain. (B) Western blot analysis with analysis of the midgut BBMV proteins showed that Cry1Ac antibodies specifictoT. ni APN1.

14038 | www.pnas.org/cgi/doi/10.1073/pnas.1102555108 Tiewsiri and Wang Downloaded by guest on September 26, 2021 100 Susceptible strain and APN1 was detected by Western blot analysis. The ratio Resistant strain * between the numbers of individuals of the two groups was 8:9 in both families, which was statistically similar to the predicted 2 10 random assortment ratio 1:1 (P > 0.10 by the χ test). In contrast, all of the larvae in the Cry1Ac-selected groups from both backcross families (22 larvae in all) had a significantly reduced < 1 apn1 transcript level with a ratio 0.1 to the susceptible larvae (Fig. 4), demonstrating the linkage (cosegregation) between the significantly reduced apn1 transcript level (apn1 transcript level < < χ2 0.1 0.1) and Cry1Ac resistance (P 0.001, test). APN6 gene

Relative transcript level transcript levels in both Cry1Ac-selected and -nonselected larvae * were similarly up-regulated, with a ratio to that of the susceptible strain ranging from 15 to >75, with no association with resistance 0.01 apn1 apn2 apn3 apn4 apn5 apn6 (Fig. 4). Although the unselected backcross individuals in both APN genes backcross families exhibited two distinct groups with differing apn1 transcript levels and differing presence or absence of APN1 Fig. 3. Relative transcript levels of APN genes in the Cry1Ac-resistant (Fig. 4), these two groups of larvae had similar apn6 transcript and -susceptible larvae by qRT-PCR analysis. Error bars indicate SEMs from levels (P > 0.10, t test of the means of apn6 transcript levels in the analysis of five individuals. Asterisks indicate that transcript levels in the the two groups). Therefore, there was no correlation between the resistant larvae were significantly different from the susceptible control levels of apn1 and apn6 transcripts in the larvae. larvae by the t test (P < 0.01; n =5). Discussion In this study, comparative biochemical and proteomic analyses apn6 are in the same linkage group, as is known in other Lepi- have shown that the Cry1Ac-resistant T. ni strain differs from its doptera (19). near-isogenic susceptible strain by two proteins in the midgut brush border membrane, APN1 and APN6. In the Cry1Ac-re- Association Between Resistance and Transcription of APN Genes. Two sistant strain, APN1 expression was significantly reduced to reciprocal F backcross families were generated between the 2 a ratio of 0.11 at the protein level and a ratio of 0.026 at the GLEN-Cry1Ac-BCS strain and its near-isogenic Cornell strain transcript level to the susceptible strain. In contrast, APN6 was and treated with or without Cry1Ac selection. qRT-PCR analysis significantly increased, to a ratio of 6.0 at the protein level and of apn1 transcript levels in larval midgut showed that in both a ratio of 39 at the transcript level. Furthermore, the 110-kDa – backcross families, the non Cry1Ac-selected individuals clearly APN1 protein was lacking in the resistant larvae (Figs. 1 and 2). exhibited two distinct groups in apn1 transcript level (Fig. 4). The differential changes of APN1 and APN6 proteins in the One group demonstrated a significantly reduced apn1 transcript resistant strain were not conferred by mutations of the APN1 level to a ratio of no more than 0.1 to the susceptible strain and and APN6 genes, but instead were regulated at the transcription no detectable 110-kDa APN1 by Western blot analysis. The level by an as-yet unidentified trans-regulatory mechanism (Fig. 3). other group had an apn1 transcript level ratio of >0.4 to that of The down-regulation of APN1 expression was linked to the the susceptible strain, which was similar to that in F1 individuals, resistance to Cry1Ac in T. ni (Fig. 4).

A APN1 B APN1 1.2 1.6 SCIENCES AGRICULTURAL 1.4 1 1.2 0.8 1 0.6 0.8 0.6 0.4 0.4 0.2 0.2 Relative apn1 transcript level Relative apn1 transcript level 0 0 F1 backcross backcross F1 backcross backcross (rs) untreated (rs:rr) treated (rr) (rs) untreated (rs:rr) treated (rr)

80 80

60 60

40 40 apn6 transcript level 20 20 Relative Relative apn6 transcript level 0 0 F1 backcross backcross F1 backcross backcross (rs) untreated (rs:rr) treated (rr) (rs) untreated (rs:rr) treated (rr)

Fig. 4. Transcript levels of APN1 and APN6 genes in F1, Cry1Ac-selected and -nonselected larvae from backcross family a (A) and family b (B) analyzed by qRT- PCR. The transcript levels are relative to the transcript levels in the susceptible Cornell strain. The expression of APN1 in nonselected larvae was also examined by Western blot analysis (Upper).

Tiewsiri and Wang PNAS | August 23, 2011 | vol. 108 | no. 34 | 14039 Downloaded by guest on September 26, 2021 Recent studies have indicated that Cry1A toxins interact with Spodoptera frugiperda, Helicoverpa zea, and P. gossypiella, have at least two types of receptors on the midgut brush border developed resistance to Bt crops (2, 3). However, the molecular membrane (12, 13, 20, 21). Activated Cry1A toxins bind to the mechanisms of the field- and greenhouse-evolved resistance have first receptor (the midgut cadherin) with high affinity, and the not yet been identified. Here we report the identification of interaction with cadherin facilitates oligomerization of the toxins a molecular mechanism of Bt resistance evolved in agricultural via a proteolytic process. The Cry1A oligomers have a high systems. With alteration of APNs as the molecular basis for binding affinity to the secondary receptor (APN or alkaline Cry1Ac resistance, the greenhouse-evolved T. ni not only is phosphatase) and thus bind to the secondary receptor, eventually highly resistant to Cry1Ac on an artificial diet, but also is highly leading to insertion of the oligomers into the midgut cell mem- resistant to Cry1Ac (16). Interestingly, different alleles brane, with resulting cell lysis. It also has been proposed that and differing transcript levels of APN genes have been reported binding of Cry1A toxins to the cadherin may activate a cellular in laboratory-selected Cry1A toxin-resistant strains of several signaling pathway leading to cell death without the involvement lepidopterans (27, 28). However, genetic linkages between the of APN (21). Nevertheless, in the sequential toxin-binding different APN alleles or differential transcript levels and the fi events, cadherin has been suggested to serve as the rst receptor resistance in these insects have not been established to confirm fi for Cry1A toxins. Loss-of-function (i.e., binding af nity to the correlation with resistance. Cry1A) mutations of the cadherin gene are expected to confer The heterozygous individuals and the homozygous Cry1Ac- mode 1 type resistance as observed in H. virescens, P. gossypiella, fi – resistant individuals clearly showed signi cantly different levels and H. armigera (9 11); however, the mode 1 type resistance in of apn1 expression (unselected individuals in Fig. 4). However, T. ni is associated with a loss of APN1. In previous studies, we the expression level of apn6 in these same individuals did not found that specific binding of Cry1Ab or Cry1Ac to the midgut demonstrate a correlated pattern (Fig. 4). Therefore, the up- BBMVs from the Cry1Ac-resistant T. ni larvae was undetectable, regulated apn6 expression is unlikely to be a response to the and that cadherin was not involved in the resistance (15, 16). reduction of apn1 expression, but may be coregulated with apn1 Therefore, results from our studies on Cry1Ac resistance in T. ni expression. The factor or factors regulating the differential indicate that cadherin alone without APN1 is not sufficient to apn1 apn6 fi serve as an efficient binding site for Cry1Ab or Cry1Ac toxins in transcription of and remain to be identi ed. The the midgut of T. ni, and that the toxicity of Cry1Ab or Cry1Ac in possibility that two different but genetically linked pathways T. ni requires crucial involvement of APN1. regulate the expression of apn1 and apn6 cannot be excluded. APN1 has been well documented as a Cry1A-binding protein Although up-regulation of apn6 transcription is not directly associated with toxicity of Cry1A toxins in lepidopterans (20). linked to Cry1Ac resistance, the up-regulation of apn6 expres- fi The Cry1Ac overlay binding analysis in this study confirmed sion could play a role in minimizing possible tness costs asso- Cry1Ac binding to APN1 of T. ni (Fig. 2). More recently, APN6 ciated with the resistance by compensating for the loss of APN1 has been identified from lepidopteran larvae, and affinity puri- with APN6. Indeed, midgut aminopeptidase activities were similar fication of H. armigera midgut proteins using Cry1Ac as a ligand in the resistant and susceptible T. ni (16). indicated that APN6 was not among the Cry1Ac-binding pro- The genetic basis for the alteration of Cry1A-binding sites in teins (22). Similarly, results of the present study indicate that mode 1 type resistance is reportedly associated with target site APN6 in the midgut BBMVs from the Cry1Ac-resistant T. ni is (cadherin) gene mutations (9–11). Our present results show that significantly more abundant than that from the Cry1Ac-suscep- Bt toxin-binding site alteration can be conferred by a trans-reg- tible strain (Fig. 1), but no specific binding of Cry1Ac or Cry1Ab ulatory mechanism as well. Therefore, the molecular bases for to the midgut BBMVs from resistant larvae could be detected mode 1 type high-level Bt resistance are not limited to binding (16), nor additional midgut BBMV protein bands with Cry1Ac site mutations, and the genetic bases for Bt resistance are more binding affinity were detected from resistant larvae by the diverse than previously thought. Cry1Ac overlay binding assay (Fig. 2). Therefore, APN6 does not appear to be a Cry1A-binding protein. In a recent report, Materials and Methods Pacheco et al. (23) proposed a “ping-pong” binding mechanism Insects. A highly inbred laboratory strain of T. ni (Cornell strain) was used as based on observations on the binding of Cry1Ab mutants to the susceptible strain (17). A Cry1Ac-resistant strain of T. ni near-isogenic to Manduca sexta cadherin and APN. In this ping-pong binding the susceptible Cornell strain, GLEN-Cry1Ac-BCS, was derived from a Bt-re- model, the Cry toxin first binds to the abundant but low-affinity sistant population collected from commercial greenhouses (16). The GLEN- APN, facilitating concentration of the toxin at the midgut brush Cry1Ac-BCS strain had been backcrossed with the Cornell strain eight times at the time that study was carried out. In addition, a susceptible T. ni strain border membrane for the consequent binding to cadherin. The fi from Benzon Research (Benzon strain) was used for genetic linkage analysis Cry toxin binds to the APN with a high af nity after oligomeri- of APN genes with resistance to Cry1Ac in T. ni. zation on interaction with cadherin (23). Our results indicate that binding of Cry1Ab or Cry1Ac to the midgut brush border Preparation of Bt Toxin Cry1Ac. Cry1Ac protoxin was prepared from the Bt membrane in T. ni requires APN1, because no specific binding kurstaki strain HD-73, as described by Kain et al. (17). The Cry1Ac protoxin sites for Cry1Ab or Cry1Ac could be detected in the midgut was activated by incubation with N-p-tosyl-L-phenylalanine chloromethyl

BBMVs from the Cry1Ac-resistant T. ni (16). ketone-treated trypsin at a ratio of 1: 20 (wt/wt) in 50 mM Na2CO3 (pH 10.0) It has been well documented that Cry1Ab or Cry1Ac can bind at 37 °C for 1–16 h, and completion of digestion was examined by SDS/PAGE to isolated cadherin protein without APN (12, 24, 25). Cry1Ac analysis. Activated Cry1Ac toxin was purified by anion-exchange chroma- also can bind to the cadherin from both the Cry1Ac-susceptible tography on a UNO Q column (Bio-Rad) with a linear gradient of 0–1 M NaCl and the Cry1Ac-resistant T. ni larvae, as demonstrated by the in 20 mM Na2CO3 (pH 10.0). toxin overlay binding assay (Fig. 2), in disagreement with the fi previous observation that Cry1Ac does not bind to the BBMVs Midgut BBMV Preparation. Mid- fth instar larvae were dissected in cold dis- from the resistant T. ni (16). Apparently, the binding of toxin to section buffer [17 mM Tris-HCl (pH 7.5), 5 mM EGTA, 300 mM mannitol, 1 mM PMSF] to isolate the midgut epithelium. Midgut BBMV proteins were pre- cadherin in the midgut brush border membrane differs from that pared as described by Wolfersberger et al. (29). The protein concentrations of of the separated proteins in vitro. fi fi the BBMV protein preparations were determined using the Bio-Rad Protein Since the rst report of eld-evolved insect resistance to Bt in Assay Kit II. The activities of brush border membrane marker enzymes al- 1990 (26), populations of two insect species, P. xylostella, and kaline phosphatase and aminopeptidase in the BBMV protein preparations T. ni, have developed resistance to Bt sprays in open-field or and their initial midgut tissue homogenates were determined (30) to con- commercial greenhouses, and four other species, Busseola fusca, firm the enrichment of brush border membranes in the BBMV protein

14040 | www.pnas.org/cgi/doi/10.1073/pnas.1102555108 Tiewsiri and Wang Downloaded by guest on September 26, 2021 preparations. The enrichment of the two enzyme activities was typically 5- to Toxin Overlay Binding Assay and Western Blot Analysis. Toxin overlay binding 6-fold and 7- to 10-fold, respectively. assays of Cry1Ac to T. ni larval midgut BBMV proteins were conducted as described by Bravo et al. (32). The binding of Cry1Ac to the BBMV proteins iTRAQ Labeling and Quantitative Proteomic Analysis of Midgut BBMV Proteins. on the membrane blot was detected by probing with a Cry1Ac-specific BBMV proteins were solubilized in 0.5 M Hepes (pH 7.4) with 5 mM EGTA, 0.3 rabbit antibody, followed by the secondary anti-rabbit IgG antibody conju- M mannitol, and 1% SDS and then reduced with 5 mM Tris-(2-carboxyethyl)- gated with alkaline phosphatase, and then visualized with a colorimetric phosphine at 37 °C for 1 h. Subsequently, the thiol groups were blocked with reaction with nitroblue tetrazolium/bromochloroindolyl phosphate. Dupli- 8 mM methyl methanethiosulfonate at room temperature for 10 min. The cate BBMV protein samples from the susceptible and resistant strains were protein samples were then digested with sequencing-grade modified trypsin included in the SDS/PAGE, and the gel was stained with Coomassie brilliant at 37 °C for 16 h, and the resulting tryptic peptides were labeled using iTRAQ blue R250. The protein bands matching the positive bands from the toxin reagent (Applied Biosystems). The sample from the Cornell strain was la- overlay binding assay were excised for protein identification by nano LC-MS/ beled separately with reporter ion reagents 114 and 116, and that from the MS analysis as described above. GLEN-Cry1Ac-BCS strain was labeled with reporter ion reagents 115 and 117 For Western blot analysis of APN1 in T. ni larvae, proteins from midgut as technical repeats. The labeled samples were combined and fractionated tissue or midgut BBMV proteins were separated by SDS/PAGE and trans- by OFFGEL isoeletric focusing electrophoresis using the Agilent 3100 OFFGEL ferred onto Immobilon-P membrane (Millipore). The membrane was treated Fractionator (Agilent) with Immobiline DryStrip (pH 3–10; 24 cm) (GE in 5% nonfat milk and 0.5% Tween-20 and then incubated with a polyclonal Healthcare). The fractions collected were pooled into 10 final fractions and antibody specifictoT. ni APN1 at 4 °C overnight, followed by incubation analyzed by nano LC-MS/MS analysis after desalting by solid-phase extrac- with the alkaline phosphatase–conjugated secondary antibodies after sev- tion using the Sep-Pak C18 cartridge (Waters). eral washes of the membrane with PBS. Positive antibody reactions were Nano LC-MS/MS analysis was performed using the LTQ Orbitrap Velos mass visualized by a colorimetric reaction as described above. spectrometer equipped with nano ion source with high-energy collisional dissociation at Cornell University’s Proteomics and Mass Spectrometry Core qRT-PCR. Total RNA was isolated from individual midguts of fifth instar larvae Facility. Samples were injected onto a PepMap C18 trap column (5 μm; 300 using the Qiagen RNeasy Mini Kit coupled with an on-column digestion μm × 5 mm) (Dionex) for online desalting and then separated on a PepMap procedure with DNase, following the manufacturer’s instructions, and then C18 RP nano column (3 μm; 75 μm × 15 cm) (Dionex). The eluted peptides used for cDNA synthesis with the Promega ImProm-II reverse-transcription were detected in the LTQ Orbitrap Velos through a nano ion source with system. The cDNA preparations from individual larvae were used for qRT- a 10-μm analyte emitter (New Objective). Data were acquired with Xcalibur PCR analysis. Real-time PCR samples were prepared in iQ SYBR Green 2.1 software (Thermo Fisher Scientific). Supermix (Bio-Rad), and reactions were carried out using the iQ5 RT-PCR The MS/MS raw spectra from iTRAQ experiments were processed with detection system (Bio-Rad). qRT-PCR included an initial hot start at 95 °C for fi Proteome Discoverer 1.1 (Thermo Fisher Scienti c), and a subsequent data- 3 min, followed by 40 cycles of amplification at 95 °C for 15 s, 58 °C for 30 s, base search was performed using Mascot Deamon version 2.2.04 (Matrix and 72 °C for 30 s. Relative transcript levels of six T. ni APN genes in the test Science) with a T. ni protein sequence database containing 15,536 sequence larvae compared with the control Cornell strain were calculated using the entries generated by combining 12,457 sequences (including 12,294 ESTs) ΔΔCT method, with an actin gene transcript as an internal control for downloaded from GenBank (http://www.ncbi.nlm.nih.gov/genbank/)on sample normalization. November 20, 2009, and 3,079 sequences (including 2,992 ESTs) generated in fi our laboratory. For the iTRAQ quantitative analysis and protein identi ca- Genetic Linkage Analyses. For linkage analysis of APN1 and APN6 genes with tion, peptide mass tolerance and fragment mass tolerance values were 10 Cry1Ac resistance, a backcross family between the GLEN-Cry1Ac-BCS strain fi ppm and 30 mDa, respectively. The signi cance threshold was set at a 95% and the Benzon strain was prepared by a single-pair cross of a female F fi fi 1 con dence interval, and only those peptides passing this lter were used for progeny with a male from the GLEN-Cry1Ac-BCS strain. Individuals from the fi fi protein identi cation. Furthermore, proteins identi ed as containing at backcross family were reared on an artificial diet (non–Cry1Ac-selected) or < least two peptides with a P value 0.001 determined by Mascot probability treated with 500 μg/mL of Cry1Ac for 7 d as described by Wang et al. (16) to analysis were analyzed further. Intensities of the reporter ions (114, 115, select Cry1Ac-resistant individuals (Cry1Ac-selected). Larval genomic DNA fi 116, and 117) from iTRAQ tags on fragmentation were used for quanti - was prepared from individual larvae using a rapid DNA isolation method cation, and the relative protein ratios were normalized at the median ratio (33). A 665-bp genomic DNA fragment for an APN1 gene and a 406-bp ge- for the fourplex in each set of experiments. nomic DNA fragment for an APN6 gene were amplified from each larva by Internal errors for the iTRAQ analysis were examined with the two PCR and subsequently sequenced to determine the allele types. technical replicates for each BBMV protein sample included in the iTRAQ runs For linkage analysis of APN1 and APN6 gene transcript levels with Cry1Ac SCIENCES to ensure the quality of the analysis. The internal error was defined such that resistance, a single-pair cross was prepared between a male from GLEN- AGRICULTURAL for 95% of the identified proteins, the difference in protein ratio between Cry1Ac-BCS and a female from its near-isogenic Cornell strain to generate F the two technical replicates [│log (117/115)│ and │log (116/114)│] was below 1 2 2 progeny. An F female was backcrossed with a male from the GLEN-Cry1Ac- this value (31). For the two sample sets analyzed in this study, the internal 1 BCS strain, and an F male was backcrossed with a female from the GLEN- errors of the iTRAQ analyses were 0.25 and 0.27 (log ratio), respectively. 1 2 Cry1Ac-BCS strain to generate two backcross families, known as backcross Similarly, biological variation was determined with the data from two in- family a and backcross family b, respectively. The progenies from each dependently prepared sets of midgut BBMV samples [Δlog ratio = │log ratio 2 2 backcross family were divided into two groups with and without selection (sample set 1) − log ratio(sample set 2)│]. The significance cutoff point was 2 with 500 μg/mL of Cry1Ac toxin as described above. Fifth instar larvae from set such that the variations of 95% of the identified proteins between the two the Cry1Ac-selected and -nonselected groups were dissected to isolate the sample sets was below this value (31). midgut tissue, and the midgut tissue was processed for Western blot analysis of APN1 protein and qRT-PCR analysis for APN gene transcription as SDS/PAGE Analysis of BBMV Proteins and Protein Identification by Nano LC-MS/ described above. MS. BBMV proteins (50 μg) were separated by 10% SDS/PAGE and then stained with Coomassie brilliant blue R250. Selected protein bands were ACKNOWLEDGMENTS. We thank Wendy Kain for excellent technical excised and subjected to in-gel digestion with trypsin for nano LC-MS/MS assistance, Sheng Zhang for critical advice and help with the proteomic analysis with the LTQ Orbitrap Velos or the Synapt HDMS MS system at analysis, and Xin Zhang for the antibodies used in this study. This study was ’ Cornell University s Proteomics and Mass Spectrometry Core Facility. MS/ supported by Agriculture and Food Research Initiative Competitive Grant MS data analysis for protein identification was performed as described 2008-35302-18806 from the US Department of Agriculture’s National Insti- above. tute of Food and Agriculture.

1. Schnepf E, et al. (1998) Bacillus thuringiensis and its pesticidal crystal proteins. Mi- 5. Heckel DG, et al. (2007) The diversity of Bt resistance genes in species of Lepidoptera. crobiol Mol Biol Rev 62:775–806. J Invertebr Pathol 95:192–197. 2. Bravo A, Likitvivatanavong S, Gill SS, Soberón M (2011) Bacillus thuringiensis: A story 6. Gahan LJ, Pauchet Y, Vogel H, Heckel DG (2010) An ABC transporter mutation is corre- of a successful bioinsecticide. Insect Biochem Mol Biol 41:423–431. lated with insect resistance to Bacillus thuringiensis Cry1Ac toxin. PLoS Genet 6:e1001248. 3. Tabashnik BE, Van Rensburg JB, Carrière Y (2009) Field-evolved insect resistance to Bt 7. Tabashnik BE, et al. (2003) Insect resistance to transgenic Bt crops: Lessons from the crops: Definition, theory, and data. J Econ Entomol 102:2011–2025. laboratory and field. J Econ Entomol 96:1031–1038. 4. Ferré J, Van Rie J (2002) Biochemistry and genetics of insect resistance to Bacillus 8. Tabashnik BE, et al. (1998) Insect resistance to Bacillus thuringiensis: Uniform or di- thuringiensis. Annu Rev Entomol 47:501–533. verse? Philos Trans R Soc Lond B Biol Sci 353:1751–1756.

Tiewsiri and Wang PNAS | August 23, 2011 | vol. 108 | no. 34 | 14041 Downloaded by guest on September 26, 2021 9. Gahan LJ, Gould F, Heckel DG (2001) Identification of a gene associated with Bt re- armigera: Identification of proteins binding the δ-endotoxin, Cry1Ac of Bacillus sistance in Heliothis virescens. Science 293:857–860. thuringiensis. Insect Biochem Mol Biol 38:685–696. 10. Morin S, et al. (2003) Three cadherin alleles associated with resistance to Bacillus 23. Pacheco S, et al. (2009) Domain II loop 3 of Bacillus thuringiensis Cry1Ab toxin is in- thuringiensis in pink bollworm. Proc Natl Acad Sci USA 100:5004–5009. volved in a “ping-pong” binding mechanism with Manduca sexta aminopeptidase-N 11. Xu X, Yu L, Wu Y (2005) Disruption of a cadherin gene associated with resistance to and cadherin receptors. J Biol Chem 284:32750–32757. Cry1Ac δ-endotoxin of Bacillus thuringiensis in Helicoverpa armigera. Appl Environ 24. Hua G, Jurat-Fuentes JL, Adang MJ (2004) Bt-R1a extracellular cadherin repeat 12 Microbiol 71:948–954. mediates Bacillus thuringiensis Cry1Ab binding and cytotoxicity. J Biol Chem 279: 12. Bravo A, et al. (2004) Oligomerization triggers binding of a Bacillus thuringiensis 28051–28056. Cry1Ab pore-forming toxin to aminopeptidase N receptor leading to insertion into 25. Flannagan RD, et al. (2005) Identification, cloning and expression of a Cry1Ab cad- membrane microdomains. Biochim Biophys Acta 1667:38–46. herin receptor from European corn borer, Ostrinia nubilalis (Hubner) (Lepidoptera: 13. Jurat-Fuentes JL, Adang MJ (2006) Cry toxin mode of action in susceptible and re- Crambidae). Insect Biochem Mol Biol 35:33–40. sistant Heliothis virescens larvae. J Invertebr Pathol 92:166–171. 26. Tabashnik BE, Cushing NL, Finson N, Johnson MW (1990) Field development of re- 14. Baxter SW, et al. (2005) Novel genetic basis of field-evolved resistance to Bt toxins in sistance to Bacillus thuringiensis in diamondback moth (Lepidoptera: Plutellidae). Plutella xylostella. Insect Mol Biol 14:327–334. J Econ Entomol 83:1671–1676. 15. Zhang X (2007) Sequence variation in cadherin alleles from the cabbage looper, 27. Yang Y, et al. (2010) Molecular characterization and RNA interference of three midgut Trichoplusia ni. MSc thesis (Cornell University, Ithaca, NY). aminopeptidase N isozymes from Bacillus thuringiensis-susceptible and -resistant 16. Wang P, et al. (2007) Mechanism of resistance to Bacillus thuringiensis toxin Cry1Ac in strains of sugarcane borer, Diatraea saccharalis. Insect Biochem Mol Biol 40:592–603. a greenhouse population of the cabbage looper, Trichoplusia ni. Appl Environ Mi- 28. Zhang S, et al. (2009) Mutation of an aminopeptidase N gene is associated with crobiol 73:1199–1207. Helicoverpa armigera resistance to Bacillus thuringiensis Cry1Ac toxin. Insect Biochem 17. Kain WC, et al. (2004) Inheritance of resistance to Bacillus thuringiensis Cry1Ac toxin Mol Biol 39:421–429. in a greenhouse-derived strain of cabbage looper (Lepidoptera: Noctuidae). J Econ 29. Wolfersberger M, et al. (1987) Preparation and partial characterization of amino Entomol 97:2073–2078. acid–transporting brush border membrane vesicles from the larval midgut of the 18. Ishihama Y, et al. (2005) Exponentially modified protein abundance index (emPAI) for cabbage butterfly(Pieris brassicae). Comp Biochem Physiol 86A:301–308. estimation of absolute protein amount in proteomics by the number of sequenced 30. Jurat-Fuentes JL, Adang MJ (2004) Characterization of a Cry1Ac receptor alkaline peptides per protein. Mol Cell Proteomics 4:1265–1272. phosphatase in susceptible and resistant Heliothis virescens larvae. Eur J Biochem 271: 19. Baxter SW, Zhao JZ, Shelton AM, Vogel H, Heckel DG (2008) Genetic mapping of Bt- 3127–3135. toxin binding proteins in a Cry1A toxin-resistant strain of diamondback moth Plutella 31. Redding AM, Mukhopadhyay A, Joyner DC, Hazen TC, Keasling JD (2006) Study of xylostella. Insect Biochem Mol Biol 38:125–135. nitrate stress in Desulfovibrio vulgaris Hildenborough using iTRAQ proteomics. Brief 20. Pigott CR, Ellar DJ (2007) Role of receptors in Bacillus thuringiensis crystal toxin ac- Funct Genomics Proteomics 5:133–143. tivity. Microbiol Mol Biol Rev 71:255–281. 32. Bravo A, Miranda R, Gómez I, Soberón M (2002) Pore formation activity of Cry1Ab 21. Zhang X, Candas M, Griko NB, Taussig R, Bulla LA, Jr. (2006) A mechanism of cell toxin from Bacillus thuringiensis in an improved membrane vesicle preparation from death involving an adenylyl cyclase/PKA signaling pathway is induced by the Cry1Ab Manduca sexta midgut cell microvilli. Biochim Biophys Acta 1562:63–69. toxin of Bacillus thuringiensis. Proc Natl Acad Sci USA 103:9897–9902. 33. Kikkert JR, et al. (2006) Detection of Contarinia nasturtii (Diptera: Cecidomyiidae) in 22. Angelucci C, et al. (2008) Diversity of aminopeptidases, derived from four lepidop- New York, a new pest of cruciferous plants in the United States. J Econ Entomol 99: teran gene duplications, and polycalins expressed in the midgut of Helicoverpa 1310–1315.

14042 | www.pnas.org/cgi/doi/10.1073/pnas.1102555108 Tiewsiri and Wang Downloaded by guest on September 26, 2021