‘rHE .JOURNAL (x:~
OF
HIOLOGICALCHEMISTRY
Vol. 266, No. 27, Issue of September 25, pp. 17954-17958, 1991 Printed in U.S . A .
1991 by The American Society for Riochemistry and Molecular Biology, Inc.
Functional Domains of Bacillus thuringiensisInsecticidal Crystal Proteins REFINEMENT OF H E L I O T H I S V I R E S C E N S
AND T R I C H O P L U S I A N I SPECIFICITYDOMAINS
ON CryIA(c)*
(Received for publication, April 22, 1991)
Albert Z. Ge$§, David RiversT, Ross Milne(1, and DonaldH. Dean**$$ From the **Departmentsof Biochemistry and Molecular Genetics, 1[Entomology, and$Microbiology, The Ohio State Uniuersity, Columbus, Ohio 43210 and the )IForestry Pest Management Institute, Sault Ste. Marie, Ontario P6A 5M7, Canada
Insecticidal crystal proteins (&endotoxins),CryIA(a) and CryIA(c), from Bacillusthuringiensis are 82% homologous. Despite this homology, CryIA(c) was determined to have10-fold more insecticidal activity toward Heliothis virescens and Trichoplusia ni than CryIA(a). Reciprocal recombinations between these two genes were performed by the homolog-scanning technique. The resultant mutants had different segments of their primary sequences exchanged. Bioassays with toxin proteins from these mutants revealed that amino acids 335-450 on CryIA(c) are associated with the activity against T. ni, whereas amino acids 335-615 on the same toxin are required to exchange full H.virescens specificity. One chimeric protein toxin, involving residues 450-612 from CryIA(c), demonstrated 30 times more activity against H. virescens than the native parental toxin, indicating that this region plays an important role in H. virescensspecificity. The structural integrity ofmutant toxin proteins was assessed by treatment with bovine trypsin. All actively toxic proteins formed a 65-kDA trypsin-resistant active toxic core, similar to the parental CryIA(c) toxin, indicating that toxin protein structure was not altered significantly. Contrarily, certain inactive mutant proteins were susceptible to complete protease hydrolysis, indicating that their lack of toxicity may have been dueto structural alterations.
CryIV, dipteran-active (Hofte and Whiteley, 1989). Understanding the specificity of these insecticidal toxins requires knowledge of their mechanism of action. The cryIA gene product (-135,000 Da) occurs as a dimer in the bipyramidal crystal (Holmes and Monro, 1965). The crystal and subunits are dissolved at thehigh alkaline pHof the lepidopteran midgut, cleaved by insect gut proteases, and reduced to the activated 65-kDa toxin (Huber and Luthy, 1981). The toxin binds to midgut columnarepithelial cells with high specificity (Hofmann et al., 1988a, 1988b) causing an influx of potassium ion (Harvey and Wolfersberger, 1979), disruption of cell homeostasis, and eventuallysis. In highly sensitive insects, paralysis and death occur with the &endotoxin alone. In less sensitiveinsects, damage to the midgut allows an avenue for invasion by B. thuringiensis from the germinated spore(Heimpeland Angus, 1959). The specificity of a 6 endotoxin seems to be determined by both extrinsic factors (insect midgut pH, proteases, and receptors) and intrinsic factors (toxin structure and functional domains, such as binding domain and cytolytic domain, within the toxic region of the protein) (Milne et al., 1990). Definition of a specificity domain implies that unique functional regions of major importance to specificity for a certain insect may be localized in the primarysequence of a toxin. We have concentrated our effortson elucidating thespecificity determinants on the CryIA toxins, particularly CryIA(a) and CryIA(c),which we previously termed IcpA and IcpC (Ge et al., 1989). CryIA(a) is highly active toward Bombyx mori (silkworm), whereas CryIA(c) isvery active toward Heliothis virescens (tobacco budworm)and Trichoplusia n i (cabbage Bacillus thuringiensis is a Gram-positive,spore-forming been looper) (Hofte et al., 1988; Ge et al., 1989).Ithas bacterium of considerable economicimportance asa biological proposed that this variability in specificity of b-endotoxins is pesticide (Rowe and Margaritis, 1987). The insecticidal comencoded by the conserved hypervariable regions of these toxponents, termed &endotoxins, are protoxin proteins, deposited as crystalsin the cytoplasm during the stationary phase ins (Geiser et al., 1986). We previously used a protein-engineering technique called of growth. Many strainsof B. thuringiensis can produce more homolog-scanning mutagenesis (Cunningham et al., 1989) to than one type of &endotoxin, each of which has its own locate a functional domain for B. mori specificity within the insecticidal spectrum. Based on theirsequence homology and hypervariable region of the cryIA(a) gene by performing a specificities, b-endotoxins havebeen classified into four major series of reciprocal recombinations between crylA(a) and groups: CryI, lepidopteran-active; CryII, active against both cryIA(c). This allowed us to identify a limited region of the Lepidoptera and Diptera; CryIII,coleopteran-active;and, gene that would transfer virtually all of the B. mori activity * This work was supported by Grant R01 A129092 from the Na- (specificity) from an active gene to an inactive homologous gene (Ge et al., 1989). In this paper, we use the same set of tional Institutes of Health (to D. H. D.) and by a grant from the Forestry Canada, Sault Ste. Marie, Ontario (to R. M.). The costs of reciprocal recombinationmutantsto locatespecificitydopublication of this article were defrayed in part by the payment of mains on the CryIA(c) protein for two economicallyimportant page charges. This article must therefore be hereby marked “aduer- insect pests, H . virescens and T. ni. tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Current address: Dept. of Pathology, L-235, Stanford University Medical School, Stanford, CA 94305-5324. $$ To whom correspondence should be addressed. Tel.: 614-2928829.
MATERIALS AND METHODS
CloningandOverexpression of cryIA(a),cryIA(c),andHybrid 6Endotoxin Genes-The plasmid pKK223-3 was obtained from Pharmacia LKB Biotechnology Inc.; E. coli host strain dM103 was from
17954
Specificity Domains of B. thuringiensis &Endotoxin .J. Messing (Rutgers, The State University, New Brunswick, NJ). B.
17955
on the supernatantusing the Bio-Rad dye-binding method. Samples were made to 0.5 mg/ml in solubilizing buffer and aliquots taken for sodium dodecyl sulfate-polyacrylamide gel analysis (Fig. 2).The solubilized protoxins were further treated by adding bovine trypsin (Sigma, T-8642, tosylphenylalanyl chloromethyl ketone-treated) at 0.5 mg/ml and incubated for 18 h a t room temperature. The digestion was stopped by adding a 10-fold molar excess of diisopropyl fluorophosphate (Sigma, D-0879, diluted in anhydrous isopropyl alcohol). Aliquots were taken for sodium dodecyl sulfate-polyacrylamide gel analysis to determine protease stabilityof the toxins (Fig. 3).
thuringiensis &endotoxin genes, cryIA(a) and cryIA(c) from strains HD-1 (Schnepf and Whiteley, 1981) and HD-244 (McLinden et al., 1985), respectively, were subcloned into pKK223-3. Oligonucleotidedirected, site-specific mutagenesis was performed by the method of Kunkel(1985),as describedpreviously (Ge et al., 1989). Overexpressed &endotoxins formed distinct crystals within E. coli cells (Ge cJt al., 1990). Chimeric protein toxin genes were generated and overexpressed inE. coli as previously described (Ge etal., 1989, 1990) and are illustrated in Fig. 1. Purification of Recombinant 6-Endotoxins-All &endotoxin genes were overexpressed in E. coli host strain JM103 (Ge et al., 1990). 6endotoxin was purified from E. coli by a modification of the method of Hofte et al. (1986). Cells were grown in 500 ml of LB medium in shake flasks with 50 pg/ml ampicillin added. After 48 h, cells were centrifuged (7000 X g for 15 min). The cell pellet was incubated at 4 "C overnight in 50 ml of buffer (50 mM Tris-HC1, pH 8, 50 mM EDTA, 15% sucrose, and 2mg/mllysozyme). The cell lysate was subsequently sonicated on ice (3 X 5 min, large tip, Fisher Sonic Dismembrator model 300). The suspension was centrifuged a t 15,000 X g for 15 min, and the supernatant was discarded. The pellet was washed three times with 0.5 M NaC1, 2% Triton X-100, five times with 0.5 M NaC1, and two times with distilledwater. The crystal protein was solubilized in bicarbonate buffer (50 mM Na,CO:,/HCl, pH 9.5, and 10 mM dithiothreitol). The solubilized proteins were examined by polyacrylamide gel electrophoresis. Protein concentrations for insectbioassays were determined byCoomassie Protein Assay Reagent (Pierce Chemical Co.) and for protease studies by the dye-binding assay (Bio-Rad). Toxicity Assay-H. uirescens and 2'. ni eggs were kindly supplied by Lillian Moug (United States Departmentof Agriculture, Western Cotton Research Laboratory,Phoenix, AZ). Bioassays were performed essentially as described by McLinden et al. (1985). Diluted proteins were applied to artificial diet disks (1.5-cm diameter), and each first instar larvawas placed on the diet ina container. Five days after feeding, insect mortalities were recorded. Twenty insects (LC,, unit was pg/ml) were used for each data point.Five data points were used to calculate the LC,,, values. Bioassays were repeated at least three times for T. ni and five times for H . uirescens. The LC,,, values and the 95% confidence intervals were calculated according to the method of Raymond (1985). Protease Digestion-Crystal inclusionbodies were suspended at approximately 2 mg/ml in solubilizing buffer (0.1 M 3-[cyclohexylaminol-1-propanesulfonic acid, pH 10.5, 0.025 M dithiothreitol;both were purchased from Sigma) and incubated for 1 h a t room temperature. The suspensionwas centrifuged and protein levels determined
RESULTS
A Specificity Domain for T. ni on CryIA(c)-As shown in Table I, LC5ovalues for strain 4101 (CryIA(a)) (LCso = 5.11 pg) and 4201 (CryIA(c)) (LC5,, = 0.43 pg) demonstrate the greater toxicity of CryIA(c) for T. ni. As described by Ge et al. (1989), site-directed mutagenesis was used to remove the first and third EcoRI sites and introduce anXhoI site to the cryIA(a), creating 4102, and to remove the first E c o R I site from cryIA(c), creating 4202. Thesemutationsalteredthe following amino acid sequences: removal of the first EcoRI site from 4101 and 4201 did not alter the amino acids involved, but removal of the third E c o R I site from 4101 altered isoleucine 531 to valine (Ile"'" -+ Val), creating4103, and the further introduction of the XhoI to 4101, creating 4102, altered phenylalanine 612 to leucine (PheG"-+ Leu). The results in Table Ishow thatthesechangesdidnot significantly alterthe toxicity of mutant proteins toward T. ni. Homolog-scanning mutagenesis between the twogenes, cryZA(a) (4102) and cryZA(c) (4202), resulted in a set of chimeric proteins (41044112 and 4204-4212). Among these, mutant proteins 4105 (encompassing the hypervariableregion) and 4110 (the NH2terminal portion of the hypervariableregion) have significantly improved activity (lower LC5ovalues) relative to the recipient parent 4101, and approximately the same as 4202 (the T. ni-active donor parent). Corresponding losses in activity were observed in the reciprocal exchange mutants 4205 and 4210. The chimeric protein 4109, the COOH-terminal region of the hypervariable region, has a slightly improved LCsovalue, but thereis no corresponding loss in thereciprocal exchange mutant 4209.
IV
FIG. 1. Homolog-scanningmutagenesis of c r y I A ( a ) ( Aa)n d cryIA(c) ( B )genes. Native genes, 4101 and 4201, were altered by oligonucleotide-directed mutagenesis to remove EcoRI sites (4102,4103, and 4202) and introduce an XhoI site (4102). Regions of the genes were exchanged to locate functional domains within the conserved region, residues 90-330 (4104 and 4204), or hypervariable region, residues 332-612 (4105 and 4205). Subdivisions of the hypervariable region were exchanged,including residues 332-428 (4106 and 4206); 429522 (4107 and 4207); 522-612 (4108 and 4208); 450-612 (4109 and 4209); 332-450 (4110 and 4210); 428-450 (4111 and 4211); and 332-526 (4112 and 4212). Thesemutants were cloned intothe expression vectorpKK223-3(modified t o remove its EcoRI site). E , EcoRI; C, ClaI; S, SstI; RV, EcoRV; K , KpnI, X, XhoI.
CryIA(c)
e
L:
F
F
E
cs
CS
."
RV
I
X
K
K
of B. thuringiensis &Endotoxin CryIA(c)
Specificity Domains
17956
toxins (4111 and 4211, Table I), as did exchanges between amino acids 332-526 (4112 and 4212, Table I). A Specificity Domain forH. virescens on CryIA(c)-As with T. ni, changing amino acids Ile”:” 4Val and Phefi” + Leu on cryIA(a) (4102 and 4103) and cryIA(c) (4202) in order to add or remove restriction enzyme sites did not affect the activity of their proteins significantly against H. uirescens (Table 11). Bioassays of homolog-scanning mutants revealed that a H.uirescens specificity domain onCryIA(c) spans most of the hypervariable domain (amino acids 332-612). Hybrid protein 4105, with amino acids 332-612 being replaced with FIG. 2. Coomassie Blue stained 12%sodium dodecyl sulfatepolyacrylamide gels showingsolubilized protoxinsfrom 4100 the homologous region CryIA(c),seemsto be as toxic as and 4200 series homolog-scanning mutants (see “Protease CryIA(c) (4201) toward H.uirescens (Table 11). Subdivisions Digestion” for the method). A, Lane 1, molecular weight standards; of this region and subsequent homolog-scanning generated Lane 2, 4101; Lane 3, 4102; Lane 4, 4103; Lane 5, 4104; Lane 6, 4105; interesting results. Replacement of amino acids 450-612 on Lane 7, 4106; Lane 8, 4109; Lane 9, 4110; Lane IO, 4111. R, Lane I, CryIA(a) by corresponding amino acids from CryIA(c) (cremolecular weight standards;Lane 2,4201; Lane 3,4202; Lane 4,4204; ating 4109) was much more toxic against H. virescens than Lane 5, 4205; Lane 6, 4206; Lane 7, 4209; Lane 8, 4210; Lane 9,4211. eitherCryIA(a)andCryIA(c)(Table 11). It improved the potency of the less active &endotoxin, CryIA(a) (4101), by A B 300-fold, and of the more active &endotoxin, CryIA(c) (4201), k D a 1 2 3 4 5 6 7 8 8 1 0 1 2 3 4 5 6 7 8 9 by 30-fold. On the other hand, the reciprocal exchange (4209) 97 . unexpectedly maintained its original activity compared with 66. CryIA(c) (Table 11). Contrary to the results withB. mori and ” ” ” “ 45 T. ni, reciprocal switching of amino acids 332-450 between CryIA(a) and CryIA(c)(4110 and 4210) does not alter activi31 ties of these toxins relative to their parent toxins(4101 and 4201, Table 11). Exchanging the region between amino acids 21 90-330 did not affect toxic activity appreciably. Replacement of aminoacids 428-450 (hybridprotein 4211) seemed to FIG.3. Coomassie Blue stained12%sodium dodecyl sulfate- decrease the activity of both CryIA(a) and CryIA(c), similar polyacrylamide gels showing trypsin-activated toxins from to what we previously observed with B. mori (Ge et al., 1989) 4 100 and 4200 series homolog-scanning mutants (see “Proand T. ni (Table I). tease Digestion” for the method). A, Lane 1, molecularweight Protease Sensitivity of Certain Mutants-Perturbation of standards; Lane 2, 4101; Lane 3,4102; Lane 4, 4103; Lane 5, 4104; protein structuremay bedetected by examining thesensitivity Lane 6,4105; Lane 7,4106; Lane 8,4109; Lane 9,4110; Lane IO; 4111. proteases,relative to the native H, Lane I , molecular weight standards; Lane 2, 4201; Lane 3, 4202; of the mutant protein to protein (Pace and Barrett,1984). In thecase of &endotoxins, Lane 4,4204; Lane 5,4205; Lane 6,4206; Lane 7,4209; Lane 8,4210; it is well known that the 130-kDa protoxin is processed to a Lane 9, 4211. 65-kDa active toxinby trypsin-like midgut proteases and that bovine trypsin will mimic this process (Luthy and Ebersold, TABLE I 1981). There are, however, numerous arginines within the Mortality of R. thuringiensis &endotoxins toward neonate larvae of T. ni folded 65-kDa toxin that would be substrates for hydrolysis by trypsin if partial unfolding occurred. We solubilized the LC,
-
”
”
Proteins
Recipient Amino acids exchanged
(95% confidence interval) Pg
4101 4102
CrylA(a) CrylA(a)
4103 4104 4105 4106 4109 4110 4111 4112 4201 4202 4204 4205 4206 4209 4210 4211 4212
CrylA(a) CrylA(a) CrylA(a) CrylA(a) CrylA(a) CrylA(a) CrylA(a) CrylA(a) CrylA(c) CrylA(c) CrylA(c) CrylA(c) CrylA(c) CrylA(c) CrylA(c) CrylA(c) CrvlA(c)
Ile”:”+ Val, phe612 --f Leu Ile”:” + Val 90-330 332-612 332-428 450-612 332-450 428-450 332-526 90-330 332-612 332-428 450-612 332-450 428-450 332-526
5.11 (3.44-9.27) 3.26 (2.37-4.06) 5.51 (1.07-8.30) 4.76 (2.97-9.76) 1.02 (0.47-1.42) 5.76 (4.44-8.07) 2.63 (1.89-3.72) 1.14 (0.75-1.52) >10.24 >10.24 0.43 (0.07-0.80) 1.23 (0.38-2.12) 2.50 (2.00-3.09) X.12 1.60 (0.79-2.20) 0.49 (0.17-0.76) 9.70 (6.83-25.2) 1.43 (0.68-2.33) X.12
Exchanges of amino acids 332-428 did not affect theoriginal toxicity of CryIA(a) and CryIA(c) significantly (hybrid proteins 4106 and 4206, Table I), whereas exchanges between amino acids 428-450 resulted in a loss in activity of both
TABLE I1 Mortality of B. thuringiensis &endotoxins toward neonate larvae of H . virescens LC, Proteins Recipient Amino acids exchanged (95% confidence PR
4101 4102
CrylA(a) CrylA(a)
4103 4104 4105 4106 4109 4110 4111 4112 4201 4202 4204 4205 4206 4209 4210 4211 4212
CrylA(a) (2.45-3.86) 3.06 CrylA(a) >3.0 90-332 CrylA(a) 332-612 (0.24-0.45) 0.33 CrylA(a) 332-428 (2.50-4.26) 3.16 CrylA(a) (0.007-0.018) 0.01 450-612 CrylA(a) 332-450 (1.23-4.12) 2.20 CrylA(a) 428-450 >6.0 CrylA(a) 332-526 >6.0 CrylA(c) (0.08-1.12) 0.3 CrylA(c) (0.33-0.46) 0.39 CrylA(c)(0.12-0.43) 0.2990-332 CrylA(c) (2.44-19.1) 4.21 332-612 CrylA(c) (0.64-2.70) 1.05 332-428 CrylA(c) (0.22-0.35) 0.28 450-612 CrylA(c) (0.22-1.00) 0.37 332-450 CrylA(c) (0.57-1.39) 0.78 428-450 CrylA(c) 332-526 >6.0
(1.75-2.98)
Ile“Y:’+ Val, Phe6” + Leu Ile”’:’ + Val
2.68 (1.96-7.22) 2.97
Specificity Domains of B. thuringiensis &Endotoxin CryIA(c) protoxin protein of each of the proteins expressed by the native and mutant genes shown in Fig. 2 and subjected them t o hydrolysis with bovine trypsin (Fig. 3). In allcases, protoxins of all native and mutant genes were processed to 65-kDa toxic cores, except for mutant proteins 4107,4207,4108,4208, or 4112, and 4212, which were eitherunstableinstorage completely hydrolyzed by trypsin to small peptides undetectable on thepolyacrylamide gels (data not shown). DISCUSSION
Operationally, we have identified specificity domains of dendotoxins as the smallest regions that can be exchanged between two homologous cry genes that result in the exchange of virtually all of the toxin activityfor a specific insect (Ge et al., 1989). This implies that one of the reciprocal pairs will receive virtually all of the activity of the more potent parent, and the other will concomitantly lose activity to become as impotentasthe less activeparent.The specificity of an insecticidal protein is the resultof several functions, including the proteolytic processing by insect midgut proteases (Haider and Ellar,1989), receptor binding (VanRie et al., 1989,1990), and/or cytolytic functions (Wolfersberger,1990). Exactly which of these functions has the most influence on a particular &endotoxin is not certain at this time. Different functions may have relatively more or less importance, depending on the &endotoxin and the insect. We found that a T. n i specificity-determining region on CryIA(c) is located between amino acids 332-450 within the hypervariable domain (Fig. 1 and Table I). Thisregion is the complement of a B. mori specificity domain located on the CryIA(a) toxin (Ge et al., 1989). It also partially overlaps the dipteran specificity domainon CryIIA toxin(Widnerand Whiteley, 1990). We observed, however, that this region does not represent the specificity domain for H . virescens on the CryIA(c). The smallest contiguous region that reciprocally transfers and replaces H. virescens toxicity is the total hypervariable region from amino acids 332-612 (Fig. 1 and Table 11). The LCso value of the hybrid CryIA(a) (hybrid protein 4105) is improved about eight times from 2.68-0.33 pg. As expected, the reciprocal hybrid CryIA(c) (hybrid protein 4205) seems to be less potent than the original CryIA(c). Accordingly, its LCs0 value increases from 0.3-4.21 pg/ml. Within this domain, there are 145 amino aciddifferencesbetween these two proteins. One hybrid protein yielded unexpected results. Hybrid protein 4109 demonstrated more potency against H. virescens than eitherof its parental toxins. The toxicity of this protein toward H. uirescens is about 30 times higher than CryIA(c), which previously was the most potent toxin reported against H. virescens (Hofte and Whiteley, 1989).At first glance, this region might be considered as the specificity domain toward H. virescens. However, the replacement of this region on the reciprocal mutant 4209 does not seem to affect the LCsovalue significantly. Our interpretation of these results is that the combination of the CryIA(c) region from 450-612 and the CryIA(a) region from 332-450 creates a stronger functional domain, such as a midgut receptor-binding region, than any other natural or synthetic construct identified to date. Trypsin digestion of the homolog-scanning mutants has allowed us to verify the structural integrity of the mutants usedinthisstudyandtodetectstructuralalterationsin certain of the mutants (4107, 4207, 4108, 4208, 4112, and 4212) thatrenderedthemsensitiveto hydrolysis by this protease. Thissuggests a basis for the lack of toxicity of these toxins. From preliminary studies with insect gut juice proteases (data not shown), we conclude that these toxins are
17957
destroyed in the insect gut and rendered inactive.All of these genetic exchanges involvedan EcoRV site in aregion encoding a conserved tract of amino acids, SQRYRVRVRYAS, found in allinsecticidal crystal proteinsfrom B. thuringiensis (Hofte and Whiteley,1989). Application of the predictive algorithms of Chou and Fasman (1974) and Gamieret al. (1978) indicates that this region may have a strong secondary structure as an
[email protected],thestructuralproperties that surround this amino acidsequence are conserved for CryIA(a) or CryIA(c), butmay not be coupled with the structures of the other without altering the structure so as to expose trypsin-sensitivesites somewhere ontheprotein. Thesestudiesindicatetheimportance of determiningthe structural integrity of all mutant forms of Cry proteins, especially those that reduce activity. As we were preparing this manuscript for publication, a very similar paper by Schnepf et al. (1990) appeared that reported on partial homolog-scanning mutagenesis for the cryZA(a) and cryIA(c) genes. In their study, transfers were made in only one direction so that portions of the cryZA(a) gene were substituted into the cryZA(c) gene. Their set is essentially equivalent to our 4205,4206,4208,4209, 4210. and None of these recombinants led to an increase in insecticidal activity, but on the basis of loss of activity, they concluded that the whole hypervariable region of CryIA(c) is required for H. virescens specificity. Our work is in agreement with this conclusion, but we provide two types of positive data to supportthis conclusion. First, we observe an increasein insecticidal activity in the corresponding homolog-scanning mutant, 4105. Partial transfer of this region does not improve the toxicity of CryIA(a) 4106 or 4110 proteins. Second, we demonstrate that the loss of activity observed in the substitution of this region (creating 4205) is not due to structural alteration resulting in protease sensitivity(Fig. 3). However, our data do not indicate a progressive reduction of toxicity of chimeric proteins (4206, 4210, and 4205) on H. virescens as suggested by Schnepf et al. (1990). However, we do observe a corresponding progressive pattern of increase of toxicity of chimeric proteins(4106,4110, and 4105). Schnepf et al. (1990) alsoconclude that the significant loss of activitytoward Manduca sexta, upon exchange of the small region 428-450 (4211), identifies it as a specificity-determining region. In all of our studies onhomolog exchanges of this region, including T. n i (Table I),H . virescens (Table 11), and B. mori (Ge et al., 1989), we have observed loss of virtually all activity in oneof the homolog pairs and greatreduction of activity in the other. Further single site-directed mutations in this region result in structural abnormalities of the protein, causing either protease sensitivity or altered thermal denaturation curves (the homolog-scanning mutants 4111 and 4211 fall in the latter category).’ It is possible that the loss of activity observed by of the chimeric proteins Schnepf et al. is due to the sensitivity to the gut proteases. Our results differ with those of Schnepf et al. (1990) in one other aspect. They constructed a recombinant thatis virtually the same as our 4109, but whereas they observed an 8-fold decrease in H. virescens activity, we observed a 30-fold increase. These two constructs are apparently from identical genes (cryZA(c) from B. thuringiensis strainHD-73and cryZA(a) from B. thuringiensis strain HD-l), and unless secondary mutations were spontaneously introduced into either of these genes, thereisnopresentexplanation for these differences. Ourfindingsare, however, in agreement with those by Edwards et al. (1987) that describe a fusion between a cryZA(a) gene and a cryIA(c) at the SstI site that has a
’ B. D. Almond and D. H. Dean, manuscript in preparation.
-121
17958
Specificity Domains of B. thuringiensis &Endotoxin CryIA(c)
significant increase in H.virescens activity. B. thuringiensk insecticidal crystal proteins associate with receptorson insect midgut epithelial cells before the action can takeplace(Hofmann et al., 1988a,1988b;Hofte et al., 1988; Van Rie et al., 1990). The heterogeneity and concentraOf receptors may determine the Of an insect to the toxin (Van Rie et al., 1989). It is possible that hybrid protein 4109 has acquired a super binding domain.We are currently testing this hypothesis. It will be interesting to know which amino acids On C q 1 A ( a ) 9 from contributed to the increased activity of the hybridProtein4109 against H. uirescens. However, the finding that a hybrid protein is much toxic to an insect than its parental proteins has providedus with a new approachto improve the activities of insecticidal protein toxins. 332-4509
Acknowledgments-We wish to thank T. Wright for excellent technical assistance.
REFERENCES Chou, p. y.*and Fasman, G. D. (1974) Biochemistry 139 222-245 Cunningham, B. c., Jhurani, p., Ng, p., and Wells, J. A. (1989) Science 243,1330-1336 Edwards, D. L.,Herrnstadt, C., Wilcox, E. R., and Wow,S.-Y. (1987) European Patent Application 0 228 838 A2 Garnier, J., Osguthorpe, D. J., and Robson, B. (1978) J. Mol. Bid. 120,97-120 Ge, A. Z., Shivarova, N. I., and Dean, D. H. (1989) Proc. Natl. Acad. Sci. U. S. A . 86,4037-4041 G ~A., z,, pester, R. M., and D ~D. H. ~ ( 1~9 9 ~G~~~ ,) 9 3 , 49-54 Geiser, M., Schweitzer, S., and Grimm, C. (1986) Gene 4 8 , 109-118 Haider, M. Z., and Ellar, D. J. (1989) J. Mol. Biol. 2 0 8 , 83-194 Harvey, w.R.9 and Wolfersberger, M. G . (1979) J . EXP. B i d 839 293-304 Heimpel, A. M., and Angus, T. A. (1959) J. Insect Pathol. 1,152-170
Hofmann, C., Luthy, P., Hutter, R., and Pliska, V. (1988a) Eur. J . Biochem.c., 173,85-91 Hofmann, Vanderbriiggen, H., Hofte, H., Van Rie, J., Jansens, s., and Van Mellaert, H. (198813) Proc. Natl. Acad.Sci. U. S. A. 8 5 , 7844-7848 Hofte, H., and Whiteley, H. R. (1989) Microbiol. Rev. 53,242-255 Hofte, H., deGreve, H., Seurinck, J., Jansens, S., Mahillon, J., Ampe, C., Vandekerckhove, J., Vanderbruggen, H., van Montagu, M., &beau, M., and Vaeck, M. (1986) Eur. J. Bhchem. 161,273-280 Hofte, H., Van Rie, J., Jansens, S., Van Houtven, A., Vanderbriiggen, H., and Vaeck, M. (1988) A d . Enuiron. Microbiol. 54,2010-2017 Holmes, K. C., and Monro, R. E. (1965) J. Mol. Biol. 14,572-581 Huber, H. E,, and Luthy, p. (1981) in Pathogenesis of Invertebrate Microbial Diseuses (Davidson, E., ed) pp. 207-234, Allanheld Osmun, Totowa, NJ Kunkel, T. A. (1985) Proc. Natl. Acad. Sci. U. S. A. 8 2 , 488-492 Luthy, P., and Ebersold, H. R. (1981) in Pathgenesis of Invertebrate Microbial Diseuses @avidson, E., ed) pp, 235-268, Allanheld Osmun, Totowa, NJ McLinden, J. H., Sabourin, J. R., Clark, B.D., Gender, D. R., Workman, W.E., and Dean, D.H. (1985) Appl. Enuiron. Microbiol. 50,623-628 Milne, R., Ge, A. Z., Rivers, D., and Dean, D. H. (1990) in Analytical Chemistry of Bacillus thuringiensis (Hickle, L. and Fitch, B., eds), pp. 22-35, American Chemical Society, Washington, D. C. Pace, C.N., and Barrett, A. J. (1984) Biochem, J , 2 1 9 , 411-417 Raymond, M. (1985) Cab. ORSTOM Ser, Entomol, Med. parusitol. Rowe, G. E., and Margaritis, A. (1987) Crit. Reu. Biotechnol. 6, 87127 Schnepf, H.E., and Whiteley, H.R. (1981) Proc. Natl. Acad.Sci. U. S. A . 78,2893-2897 SchnePf, H. E., Tomczak, K, Ortega, J. and WhiteleY, H. R. (1990) J. Biol. Chem. 265,20923-20930 Van Rie, J., Jansens, s.,Hofte, H., Degheele, D., and Van Mellaert, H. (1989) Eur. J. Biochem. 1 8 6 , 239-247 van ~ iJ., ~ J , ~ s., oft^, ~ H,, ~~ ~ ~~D., and h ~van ~ Mellaert, ~~ l ~, H. (1990) Appl. Environ. Microbiol. 56, 1378-1385 Widner, W.R., and Whiteley, H.R. (1990) J . Bacteriol. 172, 28262832 Wolfersberger, M. G. (1990) Experientia (Busel) 46,475-477 p.3
,
