Insect Biochemistry and Molecular Biology 30 (2000) 1069–1078www.elsevier.com/locate/ibmb

Binding of Bacillus thuringiensisd-endotoxins Cry1Ac andCry1Ba to a 120-kDa aminopeptidase-N ofEpiphyas postvittana

purified from both brush border membrane vesicles andbaculovirus-infected Sf9 cells

R.M. Simpson*, R.D. NewcombHorticulture and Food Research Institute of New Zealand Limited, Private Bag 92-169, Auckland, New Zealand

Received 10 November 1998; received in revised form 15 March 2000; accepted 27 March 2000

Abstract

A 120-kDa protein was purified from brush border membrane vesicles of the tortricid mothEpiphyas postvittana(Walker) basedboth on its activity as an aminopeptidase and the ability to bind theBacillus thuringiensisd-endotoxin Cry1Ac. The purified enzymehad a pI of 5.6 and was a leucine aminopeptidase, with some isoleucine, phenylalanine and tryptophan aminopeptidase activity.Further characterisation showed that the protein was also able to bind Cry1Ba. During purification, the molecular weight of theprotein decreased from 120 to 115 kDa due to the loss of a glycophosphatidinyl anchor. The protein was N-terminally sequencedand, using this information and conserved regions within other insect aminopeptidase-N (APN) sequences, redundant primers weredesigned to amplify the aminopeptidase coding sequence fromE. postvittanamidgut cDNA. The predicted protein sequence fromthe full-length cDNA was most closely related to the APN protein sequence fromHeliothis virescens(61% identity) and sharedother features of insect APNs including a Zn2+ binding site motif and four conserved cysteines. TheE. postvittanawas expressedin Sf9 cells using baculovirus, yielding a protein of molecular weight 130 kDa, but with unchanged N-terminal sequence. Purifiedrecombinant protein bound both Cry1Ac and Cry1Ba by ligand blot assays. However, despite the protein being expressed on theexternal surface of the Sf9 cells, it bound neither Cry1Ac nor Cry1Ba in vivo. 2000 Elsevier Science Ltd. All rights reserved.

Keywords:Aminopeptidase-N;Bacillus thuringiensis; Cry1Ac binding protein; Cry1Ba Cry1Ac binding protein; Baculovirus-induced proteinexpression;Epiphyas postvittana(Walker)

1. Introduction

During sporulation, the Gram-positive bacteria,Bacil-lus thuringiensisforms crystalline protein inclusions,which possess insecticidal activity. The mode of actionof these inclusion proteins ord-endotoxins has been par-tially elucidated (Gill, 1995). The crystals are solubilisedwithin the alkaline midgut of certain insects and cleavedby proteases to release an active toxin. This activatedtoxin then binds to specific high-affinity receptors on thesurface of the midgut epithelia. After binding, the toxinis thought to undergo a conformational change leading toformation of ion pores within the membrane of epithelial

* Corresponding author. Fax:+64-6-354-6731.E-mail address:rsimpson@hort.cri.nz (R.M. Simpson).

0965-1748/00/$ - see front matter 2000 Elsevier Science Ltd. All rights reserved.PII: S0965-1748 (00)00082-5

cells, eventually causing the death of the insect(Knowles and Dow, 1993).

Proteins which bind to one or more of thed-endo-toxins Cry1Aa, Cry1Ab, Cry1Ac and Cry1C have beenpurified and characterised from the midguts ofManducasexta (Knight et al., 1994; Valdamudi et al., 1995;Denolf et al., 1997),Heliothis virescens(Gill et al.,1995, Luo et al., 1997a),Lymantria dispar(Valaitis etal., 1995),Plutella xylostella(Denolf et al., 1997; Luoet al., 1997b) andBombyx mori(Yaio et al., 1997). In allbut one case, these proteins were found to be members ofthe aminopeptidase-N (APN) family (EC 3.4.11.2), theexception being a cadherin-like protein which boundCry1Aa inM. sexta(Valdamudi et al., 1995; Francis andBulla, 1997; Keeton and Bulla, 1997). APN is a 120- or170-kDa glycoprotein, with a glycosylphosphatidylinos-itol (GPI) anchor. It has been shown that incorporating

1070 R.M. Simpson, R.D. Newcomb / Insect Biochemistry and Molecular Biology 30 (2000) 1069–1078

a mixture of APN and an alkaline phosphatase into phos-pholipid vesicles enhances Cry1Ac-mediated efflux of86Rb+ up to 1000-fold (Sangadala et al., 1994). InTrich-oplusia ni brush border membrane vesicles (BBMVs),Cry1Ac pore-forming activity is dependent on theamount of APN activity present on the surface of theBBMV (Lorence et al., 1997). A 170-kDa APN fromH.virescenshas been shown by surface plasmon resonanceto bind Cry1Aa, Cry1Ab and Cry1Ac, and when the pro-tein was reconstituted in phospholipid vesicles, Cry-toxin-mediated86Rb+ efflux was related to toxicity of theendotoxin (Luo et al., 1997a).

There remains some debate over the role of APN asthe key binding site for endotoxin activity (Francis andBulla, 1997). It is a highly abundant protein, yet com-petitive binding assays suggest that a much loweramount of the binding proteins is present (van Rie et al.,1989; Garczynski et al., 1991). There are also inconsist-encies between competitive binding assays and ligandblotting experiments, the latter being used to assay bind-ing protein purification (Lee and Dean, 1996). Thecadherin-like receptor fromM. sexta has also beenexpressed in Sf9 cells, and the expressed protein has thesame ligand blotting and competitive binding assaycharacteristics as the native protein, not only forCry1Ab, but also for Cry1Aa and Cry1Ac (Francis andBulla, 1997; Keeton and Bulla, 1997). This has not yetbeen shown for APN.

The tortricid mothEpiphyas postvittana(Walker) is apolyphagous pest mainly affecting horticultural pro-duction and export from Australia and New Zealand(Wearing et al., 1991). Cry1Ac is an effective toxinagainstE. postvittana, while Cry1Ba has moderate insec-ticidal activity (Gleave et al., 1992; Simpson et al.,1997). Competitive binding studies have shown that eachof these toxins interact via two binding sites in themidgut (Simpson et al., 1997). Here we report thecharacterisation of Cry1Ac-binding proteins ofE. post-vittana and the purification of one of them, revealing itto be an APN. Furthermore, we describe the isolation ofthe gene encoding this APN and characterise the pro-tein’s d-endotoxin binding capacity, comparingendogenous and baculovirus-expressed APN.

2. Materials and methods

2.1. Insects

Epiphyas postvittanawere collected from Nelson,New Zealand, in 1975 and maintained in the Insect Rear-ing Facility, Mt Albert Research Centre, using con-ditions described previously (Singh, 1974; Clare et al.,1987). The midguts of fifth instar larvae were dissectedinto liquid nitrogen and stored at280°C.

2.2. Assays

Cry1Ac and Cry1Ba were prepared as described inSimpson et al. (1997). The ligand blot assay used todetect endotoxin binding was based on a western transferprotocol as described by Sanchis and Ellar (1993), modi-fied to include alkaline phosphatase-linked secondaryantibody, and colour development by NBT/BCIP (GibcoBRL). Alkaline phosphatase (AP) and leucine amino-peptidase (LAP) assays were performed as described byChristeller et al. (1989). Aminopeptidase activity usingvarious amino acid amide substrates was determinedusing the method described by Himmelhoch (1970).

Radiolabelled Cry1Ac and Cry1Ba were prepared asdescribed by Simpson et al. (1997). Competitive bindingassays were performed as described by van Rie et al.(1989), except that Sf9 cells were used instead of usingBBMVs. The optimal number of Sf9 cells for competi-tive binding assays was determined by a set of initialexperiments varying the number of control Sf9 cells.These were infected with baculovirus-containing pnlb,an external membrane-bound pectin lyase.

2.3. Protein purification

Brush border membrane vesicles (BBMVs) were pre-pared by the method of Wolfersberger et al. (1987).Approximately 1 ml of BBMV suspension was centri-fuged in a bench-top centrifuge at 14,000g for 5 min,then resuspended in 200µl of 1% w/v octylglucoside(OG) in 0.9% sodium chloride in 9.8 mM sodium phos-phate buffer, pH 7.4 (PBS) and briefly vortexed. Thiswas incubated at 4°C overnight, then centrifuged again.The supernatant was diluted to 1.2 ml using PBS, leftat 4°C for 2 h, then recentrifuged to remove any precipi-tate, and applied to a gel filtration column (Superose-12HR10/30; Pharmacia, Uppsala, Sweden; 300 mm long,10 mm i.d.), using 20 mM Tris–Hcl, 100 mM sodiumchloride, 0.15% w/v OG, pH 7.4. Fractions (200µl) con-taining a single Cry1Ac-binding protein were combinedand separated by anion exchange chromatography(Mono Q HR5/5; Pharmacia; 50 mm long, 5 mm i.d.),equilibrated with 20 mM Tris–HCl, 100 mM NaCl,0.15% OG, pH 7.4, and eluted with a gradient of 0.1 to1 M NaCl in 20 mM Tris–HCl, 0.15% OG, pH 7.4 in30 ml.

Sf9 cells expressing recombinantE. postvittanaAPN,prepared as described below, were briefly centrifuged,and an equal volume of PBS was added. The cells wereresuspended and 10% w/v OG added to give a finaldetergent concentration of 0.8%. This was left at 4°Covernight, and then recentrifuged. The supernatant wasdiluted to give a final OG concentration of 0.15% usingPBS, left at 4°C for 2 h, then centrifuged to remove anyprecipitate. The remainder of the purification scheme forrecombinant APN was as that used for the APN fromBBMVs.

1071R.M. Simpson, R.D. Newcomb / Insect Biochemistry and Molecular Biology 30 (2000) 1069–1078

2.4. PCR template preparation

Total RNA was prepared from midguts using a modi-fied protocol of Chirgwin et al. (1979). Approximately200 E. postvittanamidguts were thoroughly homogen-ised in liquid nitrogen using a mortar and pestle. Theresulting powder was resuspended in 15 ml of 4 M guan-idinium thiocyanate, 25 mM sodium citrate pH 7.0, 0.5%sarkosyl and 0.1 Mb-mercaptoethanol (solution D) and6 ml layered over 5 ml of 4.8 M CsCl, 10 mM EDTA,pH 8. The RNA was separated by centrifugation at150,000g for 16 h at 15°C. The RNA pellet was resus-pended in 400µl of diethylpyrocarbonate (DEPC)-treated dH2O. It was then precipitated by the addition of800 µl of ethanol and 10µl of 4 M sodium chlorideand stored under ethanol at220°C. The RNA pellet waswashed in 75% ethanol and air dried before resuspensionin DEPC-treated dH2O. PolyA+ RNA was prepared byaffinity chromatography on oligo-dT cellulose(Sambrook et al., 1989) using a QuickPrep Micro mRNApurification kit (Pharmacia). cDNA was made from 10 ofthe 400µl of polyA+ mRNA using reverse transcriptase(Superscript II, BRL) as per the manufacturer’s instruc-tions in a 20µl reaction volume using RoRidT16 (Ma etal., 1994) as a primer.

2.5. PCR and sequencing

cDNA encoding the 120-kDa Cry1Ac-binding proteinwas amplified from midgut cDNA using BoehringerMannheimTaq DNA polymerase following the manu-facturer’s instructions. Fifty-µl amplification reactionsincluded 2.5 UTaqDNA polymerase, 10 mM Tris–HCl,pH 8.3, 1.5 mM MgCl2, 50 mM KCl, 0.2 mM dNTPsand 100 pmol of each primer. All amplification reactionswere conducted in a Progene Thermal Cycler andincluded single primer and negative controls. RedundantPCR primers were designed from the N-terminal aminoacid sequence (APNf1) and conserved regions of otherinsect aminopeptidases (APNr1, APNr2) to amplify acentral region of the Cry1Ac-binding protein gene. Thesequences of all primers used in this study are given inTable 1. Amplification using 2/20µl of E. postvittanamidgut CDNA included a “hot start” of 94°C 3 min afterwhich theTaqDNA polymerase was added followed by30 cycles of 94°C 1 min, 50°C 1 min, 72°C 2 min. The39 end of the gene was amplified using a nested approachwith primers designed from sequence of the APNf1–APNr2 fragment (APNf2, APNf3) and primers (Ro, Ri;Ma et al., 1994) that bound within the cDNA synthesisprimer region (RoRidT16). Conditions for the PCR frommidgut cDNA (2/20µl) using the Ro and APNf2 primersincluded a hot start of 94°C 3 min followed by 35 cyclesof 94°C 1 min, 58°C 1 min, 72°C 3 min. Twoµl of thisreaction was used in the second round of PCR whichincluded a hot start of 94°C 3 min followed by 35 cycles

of 94°C 1 min, 65°C 30 s, 72°C 1 min. The major pro-duct (1.4 kb) was gel purified using Qiagen’s QIAquickGel Extraction Kit and cloned into a t-tailed pBluescriptSK2 plasmid vector (Stratagene, Marchuk et al., 1991)using standard techniques (Sambrook et al., 1989). ARapid Amplification of cDNA Ends (RACE) procedurewas employed to amplify the 59 end of the coding regionusing a modified version of Clonetech’s CapFinderPCR cDNA Library Construction Kit. cDNA was syn-thesised as above using the primer APNr3 in the pres-ence of 10 pmol of a primer that binds to the cap regionof mRNA (CAP). PCR using primers CAPCR andAPNr4 (Table 1) was then used to amplify the 59 regionof the APN gene. Conditions included a hot start of 94°C2 min followed by 35 cycles of 94°C 10 s, 60°C 30 s,72°C 1 min. The resulting PCR products were isolatedand cloned as above.

Primers designed to the extreme 59 and 39 ends of theAPN coding region (EpAPN59 and EpAPN39) were usedto amplify the entire coding region of theE. postvittanaAPN gene from 2µl of the midgut cDNA preparationusing High FidelityTaq DNA polymerase (BoehringerMannheim) according to the manufacturer’s instructions.This was cloned into pBSK2 as described above. Fourclones (pEpAPN1–4) were characterised on the basis ofrestriction enzyme cleavage sites. EpAPN1–2 contains asingle internalBamHI site while pEpAPN3–4 containstwo BamHI sites.

PCR products and cloned fragments (e.g. EpAPN1)were sequenced using dye-terminator chemistry andreactions run on an ABI 377 automatic sequenceraccording to the manufacturer’s instructions (AppliedBioSystems). Sequencing primers included those men-tioned above as well as APNf2, APNf4, APNf5, APNf6,APNr5, APNr6, APNr7 and APNr8, see Table 1.Sequence data were analysed using the GCG package,version 6 (Devereux et al., 1984) and Sequencher, ver-sion 3.0, 1995 (Gene codes Corporation, Inc.).

2.6. Baculovirus expression

The entire APN coding region of PEpAPN1 was sub-cloned into theSalI/XbaI sites of pFastbac of the Bac-to-Bac system (Life Technologies). A region containingthe APN coding region downstream of the baculoviruspolyhedrin promotor within pFastBac was transposedinto the baculovirus genome and used to transfect Sf9cells maintained in Sf900 II media (Life Technologies)according to the manufacturer’s instructions. Protein wasexpressed in 75-cm2 static culture vessels containing 15ml of SF900 II media seeded with 2 million cells andinfected with 2 million pfu of virus. Cultures containingbaculovirus expressing pnlb (a fungal pectin lyase) andno virus (cells only) were also raised as controls. Cellswere harvested 3–4 days after infection and collected bylow-speed centrifugation at 500g for 5 min. Recombi-

1072 R.M. Simpson, R.D. Newcomb / Insect Biochemistry and Molecular Biology 30 (2000) 1069–1078

Table 1Listing of primers used in amplification, cDNA synthesis and RACE reactions

Primer Primer sequence (59 to 39) 59 Position

APNf1 AAYGTIGAYCCIGCIYTITAYMGNYTMCC 157APNr1 CKRTAIGTIARIARCCCCARTTYTCCAT 1046APNf2 ACGCCTATAATGTCAACGTACCTCTT 757APNr2 TYRTCRTARTTNACNCKRTARAANCC 1880APNf3 ATTGAGGTCGCCAACACTAAGCC 1747APNr3 CTGGAACTGAGTAGTGGCCATCCA 585APNf4 AGCTAGCTCCAATCATGT 2231APNr4 CCCCCCAAAGAACGTCGTAGT 226APNf5 CTTCCCATCAACCTCTTCGAA 1489APNr5 AGGTGAAGACATCAGCAGTA 2392APNf6 AGGTTAACGCTTATCTCCAAT 2696APNr6 CAGGACGGCAGCGTTGCTGTT 2742APNr7 CAGCAGTCCCCAGTTTTCCAT 1038APNr8 GTAGTTAACTCTGTAGTAAC 1875CAP TACGGCTGCGAGAAGACGACAGAAGGGCAPCR TACGGCTGCGAGAAGACGACAGAARo ATCGATGGTCGACGCATGCGGATCCRi GGATCCAAAGCTTGAATTCGAGCTCRoRidT16 ATCGATGGTCGACGCATGCGGATCCAAAGCTTGAATTCGAGCTCT16

nantE. postvittanaAPN protein purification is describedin the protein purification section.

3. Results

3.1. Binding protein purification

The d-endotoxin ligand blot assay revealed threemajor Cry1Ac binding proteins inE. postvittanaBBMVsof molecular weights 120, 180 and 200 kDa (Fig. 1b,Lane 1); binding of Cry1Ac was not altered in the pres-

Fig. 1. Silver stained 7.5% SDS–polyacrylamide gel (a), ligand blot (b) and IEF (c) analysis of purification of APN. In both gels Lane 1 isBBMV proteins, Lane 2 solubilised BBMV proteins, Lane 3 after gel filtration and Lane 4 after anion exchange chromatography. The ligand blotwas probed with Cry1Ac. For the IEF gel, a pH 3–10 Bio-Rad ReadyGel was used, and only the fully purified protein was focused.

ence of 200 mM N-acetylgalactosamine (data notshown). All three major Cry1Ac-binding proteins werealso able to bind Cry1Ba. The binding proteins weresolubilised using 1% OG (Fig. 1a, Lane 2), although onsolubilisation the molecular weight of the 120-kDa bind-ing protein decreased to 115 kDa. Purification of the115-kDa protein by gel filtration followed by ionexchange chromatography yielded a single peak withboth Cry1Ac-binding and leucine aminopeptidaseactivity. IEF and SDS–PAGE showed that the proteinwas purified to homogeneity (Fig. 1a, Lane 4 and Fig.1c, Lane 1). The pI of the protein was 5.6 as determined

1073R.M. Simpson, R.D. Newcomb / Insect Biochemistry and Molecular Biology 30 (2000) 1069–1078

from the IEF gel. The N-terminal amino acid sequencewas determined by gas-phase sequencing asXXNVDPALYRLPT (single letter amino acid code, Xas any amino acid). Ligand blot assays demonstrated thatthis protein was able to bind Cry1Ba toxin in additionto Cry1Ac (Fig. 4). This contains the YRLPT sequence,which is conserved in all Cry1Ac-binding APNs (Medofet al., 1996).

The relative rates of hydrolysis of amino acid amidesby the enzyme are given in Table 2. The enzyme wasmost active with leucinamide, but also was able tohydrolyse isoleucinamide, tryptophanamide, phenylalan-inamide, valinamide and histidinamide to a lesser extent.There was no significant hydrolysis of serinamide, proli-namide, alaninamide, glycinamide and argininamide.

3.2. cDNA sequence analysis

Sequencing of pEposAPN and the 59 and 39 RACEproducts were aligned to form a contig of 3450 bp inlength (Fig. 2); GenBank accession number AF276241).This sequence contained an open reading frame of 1007amino acids (EposAPN) and includes the N-terminalamino acid sequence attained from the purified 115-kDaCry1Ac-binding protein described above. The predictedstart codon is embedded in a consensus Kozac trans-lation initiation sequence (UCAACUCAAGAUGGC,nts -10-5, Fig. 2) Cavener and Ray, 1991). There is aputative polyadenylation signal (AAUAAA) 14 bpbefore the polyA tail. Analysis of the predicted proteinEposAPN for protein localisation by PSORT predictswith p=0.919 possibility that the protein is located inthe plasma membrane (Nakai and Kanehisa, 1992). Theinferred protein contains a putative signal peptide(MAALKLLVFALACYCASS, residues 1–18) and aGPI anchor including a GPI signal sequence (GSA, resi-dues 985–987, Fig. 3) and GPI anchor hydrophobic tail(SMTLLGLAALFNFVL, residues 993–1007).Epos-APN contains four cysteine residues which are con-

Table 2Relative rates of amino acid amide hydrolysis

Amino acid amide Relative rate of hydrolysis(Leu=100)

Alaninamide 5Argininamide 0.1Glycinamide 0.2Histidinamide 6Isoleucinamide 16Leucinamide 100Phenylalaninamide 12Prolinamide 0.1Serinamide 0.1Tryptophanamide 18Valinamide 6

served in all insect APNs isolated to date (residues 750,757, 785, 821). The inferredEposAPN protein containsno predicted N-linked glycosylation sites, but has manypossible O-linked glycosylation sites. It has a predictedmolecular weight of 108 kDa once the predicted signaland other leader peptides are removed (amino acids 51–990) and assuming no glycosylation. Blast searchesusing protein sequence against the NCBI database revealthat EposAPN is most similar to other insect glutamylaminopeptidases fromH. virescens(61% identical; Gillet al., 1995) andM. sextaAPN1 (42% identical; Knightet al., 1995).EposAPN is less related toM. sextaAPN2and P. xylostellaAPN (Denolf et al., 1997).EposAPNshares conserved Zn2+ binding and catalytic residueswith other metallopeptidases (amino acid residues 374,375, 378, 397; Rawlings and Barrett, 1995).

3.3. Baculovirus expressed EposAPN

Sf9 cells infected by baculovirus expressingEposAPNfrom the clone pEposAPN1 had significantly increasedAPN activity compared to uninfected and pnlb-infectedSf9 cells (Table 3). When the cells were lysed, there wasno significant increase in APN activity, suggesting thatnearly all the enzyme was expressed on the surface ofthe cells.EposAPN was purified from Sf9 cells using amethod that was essentially the same as the purificationof APN from midgut material. The purified recombinantEposAPN protein had a molecular weight of 130 kDa,and was able to bind both Cry1Ac and Cry1Ba in ligandblot assays (Fig. 4). The N-terminal amino acid sequencewas determined by gas-phase sequencing asXDNVDPALYRLPTT.

The results of competitive binding assays of Sf9 cellsare shown in Fig. 5. Cells expressing the pectin lyase,pnlb, were able to saturably bind both Cry1Ac andCry1Ba (filled symbols). Using Sf9 cells expressingrecombinantEposAPN showed no change in the bindingof either Cry1Ac or Cry1Ba as compared to cellsexpressing the pectin lyase pnlb (the hollow symbols).

4. Discussion

Competitive binding assays using125I-labelled d-endotoxins ofB. thuringiensisshow thatE. postvittanaBBMVs possess at least two binding sites for Cry1Acand Cry1B (Simpson et al., 1997). One binding site had ahigh affinity with both Cry1Ac and Cry1Ba (dissociationconstant in the high pM range) while there was a distinctlow-affinity site (dissociation constants in the high nMrange) for each toxin. Ligand blot binding assaysrevealed three binding proteins fromE. postvittanaBBMVs with molecular weights 120, 180, and 200 kDawhich all bound both Cry1Ac and Cry1Ba. The 120-kDaCry1Ac/Cry1Ba-binding protein was purified, using the

1074 R.M. Simpson, R.D. Newcomb / Insect Biochemistry and Molecular Biology 30 (2000) 1069–1078

Fig. 2. Nucleotide and inferred amino acid (below) sequence ofEposAPN. The amino acid sequenced N-terminus is marked by the singleunderlining and the putative polyadenylation signal is doubly underlined. Potential O-linked glycosylation sites are marked by plus signs.

ligand blot assay to follow purification. During purifi-cation the molecular weight, by SDS–PAGE, dropped to115 kDa. This decrease in size is probably due to the lossof a glycosylphosphatidylinositol (GPI) anchor. Furtherevidence for the presence of a GPI anchor was our obser-vation of ethanolamine in low molecular weight gel fil-tration fractions (data not shown, Ferguson and Willi-ams, 1988). The purified binding protein hasaminopeptidase activity and the gene encoding the bind-ing protein is most similar to other lepidopteran APNgenes isolated to date.

The inferred protein contains conserved features ofaminopeptidases such as the conserved Zn2+-bindingcysteines and catalytic residues. Similar to the lepidop-teran APNs fromH. virescensand M. sexta (Lu andAdang, 1996; Luo et al., 1997b),EposAPN contains aGPI anchor; during membrane solubilisation this iscleaved resulting in the 115-kDa protein. In contrast, it

has no sites for N-linked glycosylation, although thereare many potential sites for O-linked glycosylation, includ-ing a large number of serine and threonine residues nearthe C-terminus. If these C-terminal sites are glycosylated,the carbohydrate chains are probably involved in holdingthe enzyme proud from the membrane.

After purification, baculovirus-expressed recombinantEposAPN retained its aminopeptidase activity. TherecombinantEposAPN had a higher molecular weightthan BBMV-purified protein (130 kDa compared with120 kDa); a result similar to the findings of Denolf etal. (1997), who found evidence for altered glycosylationon the expressed protein. N-terminal sequencing con-firmed the identity of the recombinant protein asEpos-APN, thus the higher molecular weight could be due toeither the lack of cleavage of the GPI anchor, extra gly-cosylation of the protein or a combination of both in Sf9cells as compared with BBMVs.

1075R.M. Simpson, R.D. Newcomb / Insect Biochemistry and Molecular Biology 30 (2000) 1069–1078

Fig. 3. Alignment of the amino acid sequence of aminopeptidase-N fromE. postvittana(this study),H. virescens(Gill et al., 1995),M. sextaAPN1 (Knight et al., 1995),M. sextaAPN2 (Denolf et al., 1997) andP. xylostella(Denolf et al., 1997). Amino acids with dark backgrounds areconserved residues of the Zn2+ binding motif, potential signal peptide sequences are underlined, sequenced N-terminus ofEposAPN is boldlyunderlined, highly conserved residues have a grey background, and residues potentially part of the GPI anchor signal sequence and hydrophobiccore sequence are boxed.

Unlike the APN expressed by Denolf et al. (1997),ligand blot assays with recombinantEposAPN showedthat it was able to bind both Cry1Ac and Cry1Ba. How-ever, in competitive binding assays using radiolabelledtoxins Sf9 cells expressingEposAPN showed no differ-ence in the ability to bind either Cry1Ac or Cry1Ba ascompared to control Sf9 cells. Inconsistencies between

ligand blot and competitive binding assays have beenreported previously (Lee and Dean, 1996), but this resultis more distinct: a protein which binds endotoxins by theligand blot assay does not bind at all in competitive bind-ing assay. This suggests that binding to aminopeptidasethat is not bound to a lipid bilayer is artifactual. Thereare a number of possible explanations for the difference

1076 R.M. Simpson, R.D. Newcomb / Insect Biochemistry and Molecular Biology 30 (2000) 1069–1078

Table 3Leucine aminopeptidase activities in Sf9 cells

Cells APN activity (nmol min21 mgcell protein21)

Uninfected cells 9pnlb-infected cells 11EposAPN-infected cells 1720LysedEposAPN-infected cells 1744

Fig. 4. Ligand blot analysis of BBMV-purifiedEposAPN and recom-binantEposAPN. The BBMV-purified protein is in Lanes 1 and 3, therecombinant protein from Sf9 cells in Lanes 2 and 4. Lanes 1 and 2were probed with Cry1Ac, Lanes 3 and 4 with Cry1Ba.

between the two assay results. Thed-endotoxin bindingsite on the APN may lie close to the GPI anchor or beone of the C-terminal O-linked carbohydrate side chains,and when the APN is attached to a membrane, the siteis sterically protected. Or the binding site may consistof two carbohydrate or peptide motifs within a certaindistance of each other, and the native structure of APNdoes not produce this while the various denaturing stepsof the ligand blot mean that an artifactual binding siteis formed. Alternatively, the binding seen in the ligandblot may be due to nonspecific interactions between thetoxins and sugar groups on the APN, which are not seenin the competitive binding assay.

APN has been implicated in the mode of action ofB.thuringiensisd-endotoxins. Initial evidence came fromSangadala et al. (1994) who showed that a mixture of

Fig. 5. Competitive binding curves for Sf9 cells infected with baculo-virus containing the geneEposAPN1 (hollow symbols) and pnlb (filledsymbols) using radiolabelled Cry1Ac (squares) and Cry1Ba (circles).

APN and alkaline phosphatase increased Cry1Ac-induced efflux of Rb2+ from phospholipid vesicles. Mas-son et al. (1995) showed, using surface plasmon reson-ance, thatM. sextaAPN was able to specifically bindCry1Ac with a KD approximately the same as thatdetermined by competitive binding assay. However, thestudy showed the existence of two binding sites on APN,whereas competitive binding studies showed only one(van Rie et al., 1989). These results have been criticisedas being inconsistent and unqualitative (Schuck andMinton, 1996). Lee et al. (1996) showed that adding pur-ified APN inhibited binding of Cry1Ac toL. disparBBMV, but again free APN was used. Other studieshave tried to show the role of APN by indirect methods,such as showing that pore-forming activity is related tothe amount of APN activity, which was varied by cleav-age of the GPI anchor (Lorence et al., 1997), or thatbinding of radiolabelled toxin decreases with incubationwith enzymes that cleave GPI anchors or N-linkedcarbohydrate side chains (Denolf et al., 1997). Thesestudies only show that the binding protein probably hasa GPI anchor, and that binding involves N-linked carbo-hydrates, not that APN is the binding protein.

The competitive binding assay results suggest that theE. postvittana120-kDa APN is not the specific receptorfor d-endotoxins. Other evidence which suggests that the120-kDa APN does not act as a specific endotoxin recep-tor is given by Luo et al. (1997b), who found that theligand blot binding of Cry1Ac to APN fromP. xylostellawas essentially unchanged between susceptible larvaeand ones with over one thousand-fold resistance toCry1Ac. In contrast, the cadherin-like receptor fromM.sextahas been expressed in Sf9 cells, and both ligandblot and competitive binding assay results of the recom-binant protein gave the same results as native protein(Francis and Bulla, 1997). This shows that expression ofputative receptors tod-endotoxins in Sf9 cells is able to

1077R.M. Simpson, R.D. Newcomb / Insect Biochemistry and Molecular Biology 30 (2000) 1069–1078

yield proteins able to bind the toxins. Thus the lack ofbinding seen by Denolf et al. (1997) and in this studyis likely to be caused by an inherent lack of specificbinding by APN rather than problems caused byexpression within the Sf9 cells.

The role APN plays in the mode of action of endo-toxin toxicity remains questionable. Although it is clearthat d-endotoxins do bind to at least some APNs in thegut of Lepidoptera it is not obvious whether this bindingresults in pore formation. APN might be part of a morecomplex array of proteins that together allow bindingand pore formation, or perhaps the endotoxins are trans-ferred to and bind other less-abundant proteins after firstbeing associated with the carbohydrate side chains ofAPN. Or possibly APN has no role at all in the modeof action ofB. thuringiensisd-endotoxins.

Acknowledgements

We thank the University of Waikato DNA sequencingfacility for DNA sequencing, Catriona Knight at Auck-land University for N-terminal amino acid sequencingand David Greenwood and Andrew Gleave for theircomments on the manuscript.

References

Cavener, D.R., Ray, S.C., 1991. Eukaryotic start and stop translationsites. Nucleic Acids Res. 19, 3185–3192.

Chirgwin, J.M., Przybyla, A.E., MacDonald, R.J., Rutter, W.J., 1979.Isolation of biologically active ribonucleic acid from sourcesenriched in ribonuclease. Biochemistry 18, 5294–5299.

Christeller, J.T., Shaw, B.D., Gardiner, S.E., Dymock, J., 1989. Partialpurification and characterization of the major midgut proteases ofgrass grub larvae (Costelytra zealandica, Coleoptra: Scaribaeidae).Insect Biochem. Mol. Biol. 19, 221–231.

Clare, G.G., Ochieng-Odero, J.P.R., Dalzell, S.J., Singh, P., 1987. Apractical rearing method for leafroller (Lepidoptera: Tortricidae).N. Z. J. Zool. 14, 597–601.

Denolf, P., Hendrickx, K., van Damme, J., Jansens, S., Peferoen, M.,Degheele, D. et al., 1997. Cloning and characterization ofManducasexta and Plutella xylostellamidgut aminopeptidase-N enzymesrelated to Bacillus thuringiensistoxin-binding proteins. Eur. J.Biochem. 248, 748–761.

Devereux, J., Haeberli, P., Smithies, O., 1984. A comprehensive setof sequence analysis programs for the VAX. Nucleic Acids Res.12, 387–395.

Ferguson, M.A.J., Williams, A.F., 1988. Cell-surface anchoring of pro-teins via glycosyl-phosphatidylinositol structures. Annu. Rev.Biochem. 57, 285–320.

Francis, B.R., Bulla, L.A. Jr., 1997. Further characterisation of BT-R1,the cadherin-like receptor for Cry1Ab toxin in tobacco hornworm(Manduca sexta) midguts. Insect Biochem. Mol. Biol. 27, 541–550.

Garczynski, S.F., Crim, H.W., Adang, M.J., 1991. Identification ofputative insect brush border membrane-binding molecules specificto Bacillus thuringiensisd-endotoxin by protein blot analysis. Appl.Environ. Microbiol. 57, 2816–2820.

Gill, S.S., 1995. Mechanism of action ofBacillus thuringiensistoxins.Mem. Inst. Oswaldo Cruz 90, 69–74.

Gill, S.S., Cowles, E.A., Francis, V., 1995. Identification, isolation,and cloning of aBacillus thuringiensisCry1Ac toxin-binding pro-tein from the midgut of the lepidopteran insectHeliothis virescens.J. Biol. Chem. 270, 27277–27282.

Gleave, A.P., Hedges, R.J., Broadwell, A.H., Wigley, P.J., 1992. Clon-ing and nucleotide sequence of an insecticidal crystal protein fromBacillus thuringiensisDSIR732 active against three species of leaf-roller (Lepidoptera: Tortricidae). N. Z. J. Crop Hort. Sci. 20, 27–36.

Himmelhoch, S.R., 1970. Leucine aminopeptidase from swine kidney.Methods Enzymol. XIX, 508–521.

Keeton, T.P., Bulla, L.A. Jr., 1997. Ligand specificity and affinity ofBT-R1, theBacillus thuringiensisCry1A toxin receptor fromMan-duca sexta, expressed in mammalian and insect cell cultures. Appl.Environ. Microbiol. 63, 3419–3425.

Knight, P.J.K., Crickmore, N., Ellar, D.J., 1994. The receptor forBacil-lus thuringiensisCryIA(c) delta-endotoxin in the brush bordermembrane of the lepidopteranManduca sextais aminopeptidaseM. Mol. Microbiol. 11, 429–436.

Knight, P.J.K., Knowles, B.H., Ellar, D.J., 1995. Molecular cloning ofan insect aminopeptidase-N that serves as a receptor forBacillusthuringiensisCryIA(c) toxin. J. Biol. Chem. 270, 17765–17770.

Knowles, B.H., Dow, J.A.T., 1993. The crystald-endotoxins ofBacil-lus thuringiensis: models of their mechanism of action on the insectgut. BioEssays 15, 469–476.

Lee, M.K., Dean, D.H., 1996. Inconsistencies in determiningBacillusthuringiensistoxin binding sites relationship by comparing compe-tition assays with ligand blotting. Biochem. Biophys. Res. Com-mun. 220, 575–580.

Lee, M.K., You, T.H., Young, B.A., Cotrill, J.A., Valaitis, A.P., Dean,D.H., 1996. Aminopeptidase N purified from gypsy moth brushborder membrane vesicles is a specific receptor forBacillus thurin-giensisCryIAc toxin. Appl. Environ. Microbiol. 62, 2845–2849.

Lorence, A., Darszon, A., Bravo, A., 1997. Aminopeptidase dependentpore formation ofBacillus thuringiensisCry1Ac toxin onTricho-plusia ni membranes. FEBS Lett. 414, 303–307.

Lu, Y.-J., Adang, M.J., 1996. Conversion ofBacillus thuringiensisCry1Ac-binding aminopeptidase to a soluble form by endogenousphosphatidylinositol phospholipase C. Insect Biochem. Mol. Biol.226, 33–40.

Luo, K., Sangadala, S., Masson, L., Mazza, A., Brousseau, M., Adang,M.J., 1997a. TheHeliothis virescens170 kDa aminopeptidase func-tions as “Receptor A” by mediating specificBacillus thuringiensisCry1A d-endotoxin binding and pore formation. Insect Biochem.Mol. Biol. 27, 735–743.

Luo, K., Tabashnik, B.E., Adang, M.J., 1997b. Binding ofBacillusthuringiensisCry1Ac toxin to aminopeptidase in susceptible andresistant diamondback moths (Plutella xylostella). Appl. Environ.Microbiol. 63, 1024–1027.

Ma, P.W.K., Knipple, D.C., Roelofs, W.L., 1994. Structural organiza-tion of the Helicoverpa zeagene encoding the precursor proteinfor pheromone biosynthesis-activating neuropeptide and other neu-ropeptides. Proc. Natl. Acad. Sci. USA 91, 6506–6510.

Marchuk, D., Drumm, M., Saulino, A., Collins, F.S., 1991. Construc-tion of T-vectors, a rapid and general system for direct cloning ofunmodified PCR products. Nucleic Acids Res. 19, 1154.

Masson, L., Lu, Y.-J., Mazza, M., Brousseau, R., Adang, M.J., 1995.The CryIA(c) receptor purified fromManduca sextadisplays mul-tiple specificities. J. Biol. Chem. 270, 20309–20315.

Medof, M.E., Nagarajan, S., Tykocinski, M.L., 1996. Cell-surfaceengineering with GPI-anchored proteins. FASEB J. 10, 574–586.

Nakai, K., Kanehisa, M., 1992. A knowledge base for predicting pro-tein localization sites in eukaryotic cells. Genomics 14, 897–911.

Rawlings, N.D., Barrett, A.J., 1995. Evolutionary families of metallo-peptidases. Methods Enzymol. 248, 183–228.

Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular cloning: alaboratory manual. Cold Spring Harbor Laboratory Press, NewYork.

1078 R.M. Simpson, R.D. Newcomb / Insect Biochemistry and Molecular Biology 30 (2000) 1069–1078

Sanchis, V., Ellar, D.J., 1993. Identification and partial purification ofa Bacillus thuringiensisd-endotoxin binding protein fromSpodop-tera littoralis gut membranes. FEBS Lett 316, 264–268.

Sangadala, S., Walters, F.S., English, L.H., Adang, M.J., 1994. A mix-ture of Manduca sextaaminopeptidase and phosphatase enhancesBacillus thuringiensis insecticidal CryIA(c) toxin binding and86Rb+–K+ efflux in vitro. J. Biol. Chem. 269, 10088–10092.

Schuck, P., Minton, A.P., 1996. Kinetic analysis of biosensor data:elementary tests for self-consistency. Trends Biol. Sci. 21, 458–460.

Simpson, R.M., Burgess, E.P.J., Markwick, N.P., 1997.Bacillus thur-ingiensisd-endotoxin binding sites in two Lepidoptera,Wiseanaspp. andEpiphyas postvittana. J. Invert. Pathol. 70, 136–142.

Singh, P., 1974. A chemically-defined medium for rearingEpiphyaspostvittana(Lepidoptera: Tortricidae). N. Z. J. Zool. 1, 241–243.

Valaitis, A.P., Lee, M.K., Rajamohan, F., 1995. Brush border mem-brane aminopeptidase-N in the midgut of the gypsy moth serves asthe receptor for the CryIA(c)d-endotoxin ofBacillus thuringiensis.Insect Biochem. Mol. Biol. 25, 1143–1151.

Valdamudi, R.K., Weber, E., Ji, I., Ji, T.H., Bulla, L.A. Jr., 1995.

Cloning and expression of a receptor for an insecticidal toxin ofBacillus thuringiensis. J. Biol. Chem. 270, 5490–5494.

van Rie, J., Jansens, S., Ho¨fte, H., Degheele, D., van Mellaert, H.,1989. Specificity ofBacillus thuringiensisd-endotoxins. Impor-tance of specific receptors on the mid-gut of target insects. Eur. J.Biochem. 186, 239–247.

Wearing, C.H., Thomas, W.P., Dugdale, J.S., Danthanarayana, W.,1991. Tortricid pests of pome and stone fruits, Australian and NewZealand species. In: van der Greet, L.P.S., Evenhuis, H.H. (Eds.).Tortricid Pests: Their Biology, Natural Enemies and Control, vol.5. Elsevier, Amsterdam, pp. 453–472.

Wolfersberger, M., Luethy, P., Maurer, A., Parenti, P., Sacchi, F.V.,Giordana, B., Hanozet, G.M., 1987. Preparation and partial charac-terization of amino acid transporting brush border membrane ves-icles from the larval midgut of the cabbage butterfly (Pierisbrassicae). Comp. Biochem. Physiol. 86A, 301–308.

Yaio, K., Kadotani, T., Kuwana, H., Shinkawa, A., Takahashi, T., Iwa-hana, H., Sato, R., 1997. Aminopeptidase N fromBombyx moriisa candidate for the receptor ofBacillus thuringiesisCry1Aa toxin.Eur. J. Biochem. 246, 652–657.