JOURNAL OF BACTERIOLOGY, May 1995, p. 22762282 Vol. 177, No.

9
0021-9193/95/$04.0010
Copyright q 1995, American Society for Microbiology

Single Amino Acid Changes in Domain II of Bacillus thuringiensis

CryIAb d-Endotoxin Affect Irreversible Binding to Manduca sexta
Midgut Membrane Vesicles
FRANCIS RAJAMOHAN, EDWIN ALCANTARA, MI K. LEE, XUE JUN CHEN,
APRIL CURTISS, AND DONALD H. DEAN*
Department of Biochemistry, Ohio State University, Columbus, Ohio 43210
Received 28 November 1994/Accepted 16 February 1995

Deletion of amino acid residues 370 to 375 (D2) and single alanine substitutions between residues 371 and
375 (FNIGI) of lepidopteran-active Bacillus thuringiensis CryIAb d-endotoxin were constructed by site-directed
mutagenesis techniques. All mutants, except that with the I-to-A change at position 373 (I373A), produced
d-endotoxin as CryIAb and were stable upon activation either by Manduca sexta gut enzymes or by trypsin.
Mutants D2, F371A, and G374A lost most of the toxicity (400 times less) for M. sexta larvae, whereas N372A
and I375A were only 2 times less toxic than CryIAb. The results of homologous and heterologous competition
binding assays to M. sexta midgut brush border membrane vesicles (BBMV) revealed that the binding curves
for all mutant toxins were similar to those for the wild-type toxin. However, a significant difference in
irreversible binding was observed between the toxic (CryIAb, N372A, and I375A) and less-toxic (D2, F371A, and
G374A) proteins. Only 20 to 25% of bound, radiolabeled CryIAb, N372A, and I375A toxins was dissociated from
BBMV, whereas about 50 to 55% of the less-toxic mutants, D2, F371A, and G374A, was dissociated from their
binding sites by the addition of excess nonlabeled ligand. Voltage clamping experiments provided further
evidence that the insecticidal property (inhibition of short-circuit current across the M. sexta midgut) was
directly correlated to irreversible interaction of the toxin with the BBMV. We have also shown that CryIAb and
mutant toxins recognize 210- and 120-kDa peptides in ligand blotting. Our results imply that mutations in
residues 370 to 375 of domain II of CryIAb do not affect overall binding but do affect the irreversible association
of the toxin to the midgut columnar epithelial cells of M. sexta.

The Bacillus thuringiensis group of aerobic, spore-forming
soil bacteria has been used for the biological control of insect
pests for several decades. The d-endotoxins, which are insec-
ticidal crystal proteins are produced by this species during the
sporulation phase of growth. On the bases of target insect
specificity and DNA homology, the d-endotoxins have been
classified into at least five (CryI to CryV) major classes (11,
26). The CryI-type d-endotoxins are active against the larvae of
lepidopteran insects (11). Upon ingestion, the CryI-type d-en-
dotoxins are solubilized by the alkaline conditions existing in
the midgut of susceptible larvae and release a protoxin (130 to
135 kDa). The protoxin is further processed into a protease-
resistant active toxin (60 to 65 kDa) by the insect gut proteases
(12). This activated toxin is proposed to bind to a specific
receptor(s) located in the luminal plasma membrane of midgut
columnar epithelial cells, irreversibly integrates to the mem-
brane, and initiates formation of pores or ion channels (selec-
tive or nonselective) (4, 10, 25).
A three-domain structure for a related coleopteran-active
d-endotoxin, CryIIIA, has been determined by X-ray crystal-
lography (18). Domain I, a seven-a-helix bundle comprising
the N-terminal 250 amino acids, is proposed to participate in
ion channel formation. Domain II, consisting of three sets of
b-sheets, each terminating with a loop, is presumed to be
involved in membrane binding. Domain III, a b-sandwich
which includes the conserved C-terminal region of the toxin, is
now believed to play a role in ion channel function (2). Li et al.
(18) have suggested a similar structure for other insecticidal
d-endotoxins, since the core of the molecule encompassing all
* Corresponding author. Phone: (614) 292-8829. Fax: (614) 292-
3206. Electronic mail address: dean.10@osu.edu.
the domain interfaces is built on five highly conserved blocks of
amino acids common among most of the d-endotoxins.
Several in vitro membrane binding studies have been per-
formed with brush border membrane vesicles (BBMV) iso-
lated from susceptible insect larval midguts (10, 13, 17, 19, 28).
Often there is a positive correlation between either binding
affinity or number of binding sites and insect toxicity (17, 28).
However, there are certain exceptions to the above generali-
zation (4, 30). There is often more than one specific toxin
binding site on the BBMV for certain toxins (21, 28). The
toxin-binding proteins (receptors) of several insects have been
identified by the ligand blotting method (4, 21, 27). More
recently, Knight et al. (14) and Sangadala et al. (23) have
identified and purified a CryIAc toxin-binding protein from
Manduca sexta BBMV. This membrane receptor is aminopep-
tidase-N, which is a GalNAc-bearing glycoprotein (15).
The insect specificity-determining regions of several toxins
have been identified by exchanging short peptide segments
between different toxins (5, 6). In an early study, Ge et al. (6)
transferred the Bombyx mori activity from a highly active toxin
(CryIAa) to a relatively inactive toxin (CryIAc) by exchanging
the amino acids 332 to 450. Later, a direct correlation between
toxicity and the membrane binding properties of those mutants
with B. mori BBMV was established (17). The initial (revers-
ible) binding of CryIAa toxin to B. mori BBMV was also
disrupted by the deletion of residues 365 to 371 of CryIAa
toxin (19).
In this work we evaluate the function of the putative loop
residues, 370PFNIGI375, of domain II, loop 2 of B. thuringiensis
d-endotoxin CryIAb with M. sexta larvae. We demonstrate that
deletion of amino acids 370 to 375 (D2) and alanine substitu-
tion of residue F-371 (F371A) and G374A significantly affect

2276
VOL. 177, 1995 DOMAIN II OF B. THURINGIENSIS CryIAb d-ENDOTOXIN 2277

larvicidal potency for M. sexta. We also compare the BBMV

binding and short-circuit current (Isc) inhibition properties of
the wild-type and mutant toxins to M. sexta. The irreversible
association of BBMV (receptor)-bound CryIAb toxin to M.
sexta is significantly reduced when residues F-371 and G-374
are replaced with alanine. Our results suggest that in addition
to initial receptor binding, irreversible integration of the d-en-
dotoxin into the midgut epithelial cells is a critical step for
toxicity and is affected by mutations in domain II, loop 2.

MATERIALS AND METHODS

Site-directed mutagenesis. The cryIAb(pSB033) gene was provided by Sandoz
Agro Inc., Palo Alto, Calif. pSB033 is the cryIAb gene cloned from HD133 into
pBluescript KS1 with the cryIAc terminator and cryIC promoter (constructed by
Sandoz Agro Inc.). Mutations in cryIAb were performed with the template
pSB033b. This was constructed by recloning the ApaI-to-NotI fragment (4.073
kb) of pSB033 into the corresponding enzyme-treated pBluescript KS2 vector.
Uracil-containing template was obtained by transforming Escherichia coli CJ236
(Bio-Rad) with pSB033b. Oligonucleotides used for this work were provided by
Sandoz Agro Inc. The site-directed mutagenesis procedure (Muta-Gene M13 in FIG. 1. (A) Alignment of the toxin protein sequence of CryIAb with that of
vitro mutagenesis kit; Bio-Rad) was as detailed in the manufacturers manual. a related beetle-active CryIIIA toxin, as given by Hodgman and Ellar (9). The
Single-stranded DNA sequencing was carried out by the method of Sanger et al. positions of the secondary structural elements of CryIIIA are given above its
(24) by following the manufacturers (United States Biochemical) instructions. sequence, as described by Li et al. (18). The positions of amino acid residues
All restriction enzymes were purchased from Boehringer Mannheim Biochemi- considered here are shown. (B) Summary of deletion (D2) and the numbers and
cals. amino acid residues substituted with alanine on CryIAb toxins. Numbers in
Purification, solubilization, and activation of d-endotoxin inclusions. Wild- parentheses indicate the corresponding amino acids deleted.
type and mutant proteins were expressed in E. coli MV1190, and the d-endotox-
ins were purified as previously described (17). The final pellet, referred to as the
crystal protein, was solubilized in crystal solubilization buffer (50 mM Na2CO3
Western blotting (immunoblotting). Gut enzyme digestion of the crystal pro-
NaHCO3, pH 9.5, 10 mM dithiothreitol) at 378C for 2 h. The yield of solubilized
tein was performed by incubating the protoxin with freshly prepared M. sexta gut
protein was estimated by the Coomassie protein assay reagent (Pierce Chemical
enzymes (the gut fluid vomited by the insect upon very gentle squeezing) in 50
Co.). The solubilized protoxin was digested with 2% (by mass) trypsin at 378C for
mM sodium carbonate buffer (pH 9.5) at 378C for 3 h. The enzyme reaction was
5 h and analyzed by sodium dodecyl sulfate (SDS)10% polyacrylamide gel
stopped by adding Laemmli sample buffer (16) and heating, resolved by SDS-
electrophoresis (PAGE).
PAGE, and transferred to a polyvinylidene difluoride membrane with a Bio-Rad
Iodination of d-endotoxins. Iodination of trypsin-activated toxin was carried
Trans Blot apparatus. The blot was processed and developed as described by Lu
out with iodobeads by following the manufacturers (Pierce) instructions. One
et al. (19). Anti-CryIAb serum (1:2,000 dilution), raised in a rabbit, was used as
iodobead was incubated with 1.0 mCi of Na125I (diluted in 50 ml of sodium
the primary antiserum in this experiment.
carbonate buffer) at room temperature for 5 min. A total of 25 mg of toxin
Ligand blot. A total of 50 mg of BBMV proteins was solubilized in Laemmli
solution was then added to the reaction tube, and the incubation was continued
sample buffer (16), separated by SDS7.5% PAGE, and transferred to a polyvi-
for another 10 min at room temperature. The iodination reaction was stopped by
nylidene difluoride membrane as described before (19). The blot was blocked
removing the reaction solution from the iodobead. Iodinated proteins were
with 3% nonfat dry milk (Kroger Co.) in TTBS (50 mM Tris-HCl, 150 mM NaCl,
separated from free iodine by passing the sample through a 2.0-ml Excellulose
0.05% Tween 20) for 1 h, washed three times with TTBS, and incubated with 2
column (Pierce Chemical Co.) according to the instructions provided by the
3 106 cpm of 125I-labeled toxin for 1 h at room temperature. The unbound toxin
manufacturer. The specific activities of the labeled CryIAb, D2, F371A, N372A,
was washed with TTBS (three times, 5 min each time), and the blot was dried and
G374A, and I375A proteins were 1.5, 1.1, 1.5, 1.4, 1.3, and 1.4 mCi/mg, respec-
exposed to Fuji RX film for 2 to 3 days.
tively.
Insect bioassay. Toxicity assays were performed with newly hatched larvae. A
Midgut isolation and BBMV purification. M. sexta eggs were kindly supplied
total of 100 ml of toxin dilutions (diluted in phosphate-buffered saline [8 mM
by Michael Jackson, U.S. Department of Agriculture (North Carolina). The eggs
Na2HPO4, 2 mM KH2PO4, 150 mM NaCl, pH 7.4]) was layered on artificial diet
were hatched and raised to the fifth instar on artificial diet (Bio Serve). The
(2 cm2) and air dried. A total of 10 to 15 larvae were placed on each dilution, and
dissection procedure was as described previously (2). BBMV were prepared by
four to five dilutions were used to determine the slope. Mortality was scored after
the differential magnesium precipitation method as modified by Wolfersberger et
5 days. Effective dose estimates (50% lethal concentration of toxin [LC50] and
al. (33). The final BBMV pellet was resuspended in binding buffer (8 mM
95% fiducial limits) were obtained by probit analysis (22).
NaHPO4, 2 mM KH2PO4, 150 mM NaCl, pH 7.4), and the protein concentration
Voltage clamp analysis. Voltage clamp analysis was performed as described by
was measured with the Coomassie protein assay reagent (Pierce).
Chen et al. (2). The dissection of insects and mounting of midgut membrane
Competition binding assay. A total of 100 mg of BBMV per ml was incubated
were as described by Harvey et al. (8) After 15 min of stabilization, 50 ng of
with 1 nM 125I-labeled toxin in the presence of increasing concentrations (0 to
trypsin-activated toxin per ml was injected into the lumen side of the chamber.
1,000 nM) of the appropriate nonlabeled toxin in 100 ml of binding buffer (with
The volume of the lumen chamber is 3.75 ml. The Isc was tracked with a Kipp
0.1% bovine serum albumin) at room temperature for 1 h. The unbound toxin
and Zonen recorder, and data were collected with the MacLab data acquisition
was separated from the BBMV-bound toxin by centrifugation at 13,500 3 g for
system. The slope of falling Isc was measured.
15 min in a Fisher Microfuge. The pellet which contained the bound toxin was
washed three times with binding buffer, and the radioactivity in the final pellet
was measured in a gamma counter (Beckman). Binding data were analyzed by RESULTS
using the LIGAND computer program (20). In homologous competition assays,
one labeled protein was inhibited by the same nonlabeled protein, whereas in Construction of mutants. Mutants with deletions or alanine
heterologous competition assays, labeled CryIAb was inhibited by nonlabeled
mutant toxins.
substitutions at the predicted loop 2 region of domain II of
Dissociation binding assay. 125I-labeled toxin (1.0 nM) was incubated with 100 CryIAb toxin were constructed. An amino acid alignment of
mg of BBMV per ml in 100 ml of binding buffer at room temperature for 2 h CryIAb loop 2 residues with CryIIIA is shown in Fig. 1A. This
(association binding) to achieve saturation binding. In one experiment, 100 nM alignment agrees with that of Hodgman and Ellar (9). The
(final concentration) nonlabeled toxin (in a 10-ml volume) was added to each
sample tube at 2 h of association reaction, and the reaction was stopped at
positions of amino acid residues deleted or replaced in CryIAb
different time points (10 min to 2 h) by centrifugation. In another experiment, at toxin are shown in Fig. 1B.
2 h of association binding, increasing concentrations (1 to 500 nM) of the Analysis and activation of crystal proteins. Crystal proteins
corresponding nonlabeled toxin were added and incubation was continued for purified from the wild type and mutants were solubilized, ac-
another 2 h. The samples were centrifuged, pellets were washed three times with
binding buffer, and the radioactivity in the final pellet was measured as men-
tivated by trypsin, and analyzed by SDS-PAGE. Our results
tioned before. The counts at 2 h of association reaction were considered to be (Fig. 2A) showed that the mutants produced protoxins (135
100% binding. kDa) and trypsin-activated toxins (65 kDa) similar to those of
2278 RAJAMOHAN ET AL. J. BACTERIOL.

FIG. 2. (A) Coomassie blue-stained SDS10% PAGE gel comparing the
yield of protoxins (lanes 1 to 6) and trypsin-activated toxins (lanes 7 to 12). Lane
1, CryIAb; lane 2, D2; lane 3, F371A; lane 4, N372A; lane 5, G374A; lane 6,
I375A; lanes 7 to 12, corresponding (lanes 1 to 6, respectively) trypsin-activated
toxins. Each lane contained 5.0 mg of toxin, except lanes 5 and 6, which contained
3.0 mg of toxin. (B) Western blot analysis of the stability of wild-type and mutant
proteins after digestion with M. sexta gut enzymes. Lanes 1, 3, 5, 7, 9, and 11,
protoxins of CryIAb, D2, F371A, N372A, G374A, and I375A, respectively; lanes
2, 4, 6, 8, 10, and 12, gut enzyme-activated toxins of CryIAb, D2, F371A, N372A,
G374A, and I375A, respectively. Molecular mass marker positions (in kilodal-
tons) are indicated to the left.
the wild-type CryIAb, except in the case of I373A. Since the
mutant I373A did not yield toxin (data not shown), it was not
considered for further studies. To evaluate whether the muta-
tions caused any structural alterations to the mutant protein,
the protoxins were subjected to M. sexta gut juice. The Western
blot analysis showed that the mutants yielded stable 60-kDa
toxins similar to those of the wild type (Fig. 2B).
Biological activity. The biological activities of the mutant
and wild-type proteins on first-instar M. sexta larvae were com-
pared and are reported in Table 1. The LC50s showed a two-
fold difference in potency between the wild type and N372A or
I375A (LC50s of 65, 125, and 120 ng of toxin per cm2, respec-
tively). In contrast, D2, F371A, and G374A lost much (400
times) of their larvicidal activity (LC50s of .30, 29, and 20 mg
of toxin per cm2, respectively).
Competition membrane binding study. In homologous com-
petition binding assays, labeled toxins were put into competi-
tion with the corresponding nonlabeled toxin to evaluate the
binding affinity and binding site concentrations on M. sexta
BBMV. Our studies showed that the wild-type and mutant
proteins were able to bind to M. sexta BBMV with similar
binding affinities and binding site concentrations (Fig. 3A and
Table 1). The binding affinities (Kd) estimated for the toxins
were between 0.4 and 1.2 nM (Table 1). Heterologous compe-
tition experiments were performed to evaluate whether the
mutant proteins recognize the same binding site as that of the
wild-type toxin. When labeled CryIAb toxin was put into com-
petition with nonlabeled mutant proteins, all the mutants com-
peted as efficiently for the CryIAb binding site(s) as did non-
labeled CryIAb (Fig. 3B).
Dissociation of membrane-bound toxins. Since all the toxins
showed similar binding affinities to M. sexta BBMV, we tested
the stability of the BBMV-bound toxins. The toxins were first
allowed to bind to the BBMV and were then chased with the
corresponding nonlabeled toxins. Our results showed that 75
to 80% of the BBMV-bound wild-type toxin and mutant toxins
N372A and I375A were not displaced by the addition of non-
labeled ligands (i.e., they were irreversibly associated). In con-
trast, only 45 to 60% of the mutant toxins D2, F371A, and
G374A was able to associate irreversibly with BBMV (Fig.
4A). The continued decrease in binding for the defective mu-
tants (Fig. 4A) was not due to toxin breakdown or release of
labeled fragments, since we could recover the intact labeled
toxins after incubation with BBMV for 4 h (data not shown).
When the extent of dissociation was analyzed as a function of
increasing concentrations of nonlabeled toxin, the maximum
dissociation occurred at 100 nM nonlabeled ligand. Also, in
agreement with the previous experiment, 75 to 80% of the wild
type, N372A, and I375A and 45 to 50% of D2, F371A, and
G374A remained associated with the BBMV (Fig. 4B) after
incubation with nonlabeled ligand.
Identification of the toxin-binding polypeptide. To deter-
mine the toxin-binding receptor molecule(s), M. sexta BBMV
proteins were separated by SDS7.5% PAGE and transferred
to a polyvinylidene difluoride membrane. The membrane was
then probed with iodine-labeled CryIAb, D2, F371A, and
N372A toxins. Our results showed that all the toxins bound
preferentially to a 210-kDa major peptide and to some extent
to a diffuse 120-kDa peptide (Fig. 5A). A similar result was
observed when the BBMV proteins were probed with biotin-

TABLE 1. Biological activity, binding efficiency, and voltage sensing of CryIAb loop 2 mutant proteins in experiments with M. sexta
Isc inhibition rate sloped
Toxin LC50a (ng/cm2) Kdb (nM) Bmaxc (pmol/mg)
(mA/cm2/min)

CryIAb 62 (3691) 0.61 6 0.11 2.1 6 0.29 222.8 6 3.3

D2 .25,000e 0.98 6 0.13 2.0 6 0.21 23.1 6 1.0
F371A 29,000 (19 3 10335 3 103) 1.2 6 0.11 1.9 6 0.17 23.5 6 0.97
N372A 125 (100150) 1.2 6 0.18 2.1 6 0.29 212.8 6 2.11
G374A 21,600 (15 3 10330 3 103) 1.5 6 0.12 2.0 6 0.15 23.6 6 1.07
I375A 150 (90210) 0.97 6 0.16 2.0 6 0.38 210.1 6 2.98
a
95% confidence limits are given in parentheses.
b
Kd, dissociation constant (determined from homologous competition binding). Each value is the mean of three experiments.
c
Bmax, binding site concentration (determined from homologous competition binding). Each value is the mean of three experiments.
d
Each value is the mean of three experiments.
e
Insufficient mortality to calculate the LC50.
VOL. 177, 1995 DOMAIN II OF B. THURINGIENSIS CryIAb d-ENDOTOXIN
FIG. 3. Binding of 125I-labeled toxins to M. sexta BBMV. (A) M. sexta BBMV
(100 mg/ml) were incubated with 1.0 nM 125I-labeled CryIAb, D2, F371A,
N372A, G374A, and I375A toxins in the presence of the corresponding nonla-
beled toxins. (B) M. sexta BBMV (100 mg/ml) were incubated with 1.0 nM
125
I-labeled CryIAb toxin in the presence of nonlabeled CryIAb, D2, F371A,
N372A, G374A, and I375A toxins. Binding is expressed as a percentage of the
amount bound upon incubation with labeled toxin alone. On M. sexta vesicles,
these amounts were 3,700, 3,215, 3,543, 3,098, 3,671, and 3,775 cpm for CryIAb,
D2, F371A, N372A, G374A, and I375A, respectively.
labeled CryIAb, D2, F371A, and N372A toxins (data not
shown). The binding of the 125I-labeled CryIAb toxin to 210-
and 120-kDa peptides was inhibited nearly completely when
incubated with an excess (250-fold) amount of nonlabeled
CryIAb protein (Fig. 5B).
Voltage clamping study. We examined the inhibition of Isc
in response to the addition of wild-type and mutant toxin (50
ng/ml) to the lumen side of the midgut. Isc measures the active
transport of ions from the hemolymph side of the midgut
2279
membrane to the lumen side. The slopes of Isc inhibition of
mutants N372A and I375A were comparable to that for the
wild type, whereas the slopes of inhibition of Isc for the mu-
tants D2, F371A, and G374A were six to seven times less than
that for the wild type (Fig. 6; Table 1).
DISCUSSION
We have previously reported (19) that the amino acid resi-
dues 365 to 371 (predicted loop 2) of CryIAa toxin are essen-
tial for membrane binding and toxicity to B. mori. In that case,
we observed that deletion of or alanine substitution at residues
365 to 371 of domain II significantly affected initial binding as
measured by competition binding assays. In this study we have
continued our investigation on the predicted receptor binding
domain (domain II) of another toxin, CryIAb, and analyzed
the function of putative loop 2 amino acid residues (370 to
375), PFNIGI, with M. sexta. We have made the unexpected
observation that deletion or modification of some residues in
loop 2 of domain II greatly affects the irreversible binding as
measured by dissociation assays, although the initial binding as
measured by competition binding experiments is not affected.
Insect specificity of several lepidopteran-active CryIA type
d-endotoxins is often correlated with the hypervariable region
between amino acid residues 280 and 550 (5, 6, 29). The struc-
tural alignment of CryIA toxins with the secondary structure of
CryIIIA beetle-active toxin, determined by X-ray crystallogra-
phy, showed that the hypervariable region corresponds to do-
main II (18) of CryIIIA. Domain II (amino acid residues 291 to
500), composed of three b-pleated sheets, each ending in a
loop structure, is proposed to participate in receptor binding
and host specificity (18). Although CryIIIA and CryIA share
only 33% amino acid identity (9), Yamamoto and Powell (34)
have proposed a very strong similarity of domains II and III of
CryIIIA toxin to CryIA-type toxins on the basis of computer-
predicted secondary structure analysis. These studies allowed
us to align the loop 2 residues of CryIIIA and CryIAb (Fig. 1).
This is in agreement with the alignment of Hodgman and Ellar
(9). Using site-directed mutagenesis techniques, we have de-
leted residues 370 to 375 (D2) and individually replaced resi-
dues 371FNIGI375 with alanine. Alanine substitutions were
chosen because they might not cause gross structural pertur-
bation in the folding of the mutant proteins (3).
All the mutants, except I373A, express d-endotoxin at a level
comparable to that of the wild-type CryIAb toxin. Our bioassay
data show that the mutants D2, F371A, and G374A have sub-
stantially lost the larvicidal activity (more than 400-fold) for M.
sexta larvae (Table 1). Three lines of evidence suggest that the
loss of larvicidal activity is not a result of the protein being
grossly affected or misfolded. First, these proteins can be ex-
pressed at levels comparable to those for wild-type CryIAb
(Fig. 2A). Poor expression can be correlated with an unstable
or misfolded protein (1). Second, both mutant and wild-type
protoxins yielded a 65-kDa fragment after trypsin activation
(Fig. 2A). Finally, the wild-type and mutant protoxins are sim-
ilarly processed into a stable 60-kDa toxin on treatment with
M. sexta gut juice enzymes (Fig. 2B).
Our radiolabeled-toxin binding assay data showed that the
binding of mutant proteins to M. sexta BBMV is highly specific
and saturable, like that of the wild-type toxin (data not shown).
Homologous competition binding studies suggested that the
binding affinity (Kd) and binding site concentrations (Bmax) of
both toxic and less-toxic mutants and the wild-type toxin to M.
sexta BBMV do not differ significantly (Fig. 3A; Table 1). The
heterologous competition experiments, performed to investi-
gate whether the mutant and wild-type toxins recognize the
2280 RAJAMOHAN ET AL.
FIG. 4. Dissociation of bound 125I-labeled toxins from M. sexta BBMV. (A)
M. sexta BBMVs (100 mg/ml) were incubated (association reaction) with 1 nM
125
I-labeled CryIAb, D2, F371A, N372A, G374A, and I375A toxins. At 120 min
after initiation of the association reaction, 100 nM concentrations of the corre-
sponding nonlabeled toxins were added to the test samples and incubation was
continued (postbinding). (B) At 120 min after initiation of the association reac-
tion, increasing concentrations of the corresponding nonlabeled toxins were
added and incubation was continued for another 120 min. Binding is expressed
as a percentage of the amount bound compared with the amount bound at 0 h
postincubation.
same binding site, reveal that all the mutants successfully com-
pete with the wild-type toxin for the binding to M. sexta vesicles
(Fig. 3B). These data show that the reduced larvicidal activity
of D2, F371A, and G374A is not caused by changes in the
membrane binding property. Since the deletion of residues 370
to 375 (D2) does not alter the initial binding, it suggests a
possible involvement of other segments of domain II, such as
loops 1 and 3, in initial receptor binding function. In contrast,
J. BACTERIOL.
FIG. 5. Binding of iodine-labeled toxins to protein blots of M. sexta BBMV
proteins. (A) BBMV proteins (50 mg) were blotted and probed with 2 3 106 cpm
of iodine-labeled toxins. Lane 1, CryIAb; lane 2, D2; lane 3, F371A; lane 4,
N372A. (B) BBMV proteins (50 mg) were blotted and probed with 2 nM 125I-
labeled CryIAb in the absence of nonlabeled CryIAb (lane 1) and the presence
of 500 nM nonlabeled CryIAb toxin (lane 2). The arrow indicates the position of
the 210-kDa protein.
the deletion of loop 2 residues 365 to 371 of CryIAa toxin
removed all toxicity and considerably reduced the binding to B.
mori BBMV (19). This comparison illustrates the complexity of
interaction between different toxins and different insect midgut
receptors.
Our ligand blot experiments provided additional evidence
that the wild-type and mutant (toxic and less-toxic) toxins rec-
ognize a common 210-kDa peptide (Fig. 5A), as previously
observed by Vadlamudi et al. (27). We have also noticed bind-
ing to a 120-kDa protein in most of our preparations (Fig. 5A
and B). The binding of the labeled CryIAb toxin to 210- and
120-kDa peptides is specific and could be inhibited in the
presence of excess nonlabeled CryIAb toxin (Fig. 5B). These
experiments suggested that the mutants D2, F371A, and
N372A bind to the same receptor proteins with affinities sim-
ilar to that of CryIAb, although they have lost most of their
toxicity for M. sexta.
Recently, several reports have indicated that the initial bind-
ing is not the only factor to determine toxicity. The CryIAc
toxin binds with high affinity to Spodoptera frugiperda BBMV
but is not toxic to the larvae (4). In another case, Wolfers-
berger (30) has observed an inverse correlation between mem-
brane binding and toxicity to Lymantria dispar for the toxins
CryIAb and CryIAc. In addition, insect toxin resistance is not
caused by reduced toxin binding in some cases (7). These
experiments suggest that the late binding events, such as irre-
versible membrane association of the toxin, could also play a
major role in determining the toxicity. Ihara et al. (13) have
identified irreversible and reversible components of receptor-
bound CryIAa and CryIAb toxins with B. mori. They have
concluded that CryIAa has a greater irreversible component
with B. mori BBMV than does CryIAb, and hence CryIAa is
more toxic to the insect.
Interestingly, our dissociation experiments showed a striking
difference between toxic and less-toxic mutants. About 75 to
80% of wild-type toxins and toxic mutants are irreversibly
associated with the BBMV, whereas only 45 to 50% of the
VOL. 177, 1995 DOMAIN II OF B. THURINGIENSIS CryIAb d-ENDOTOXIN
FIG. 6. Inhibition of Isc across M. sexta midgut. A total of 50 ng of CryIAb,
D2, F371A, N372A, G374A, I375A, and CryIIIA toxins per ml were injected in
separate experiments into the lumen side of the chamber, and the drop in Isc was
measured.
nontoxic mutants is irreversibly associated with the BBMV
under similar conditions (Fig. 4). This considerable reduction
in the integration of the receptor-bound toxin could account
for the 400-fold difference in toxicity. It has been suggested
that after receptor binding, the toxin undergoes a major con-
formational change which enables the a-helix 5 of domain I to
penetrate the membrane (31, 32). Our results suggest that the
amino acid residues F-371 and G-374 of loop 2 directly or
indirectly play an important role in holding the CryIAb toxin
bound to the receptor during the transition state in M. sexta.
Also, the irreversible association may be one of the rate-de-
termining steps in the mechanism of action of d-endotoxins.
These present results signify that the membrane binding
might be a two-step process (at least in this case): (i) recogni-
tion of the target (receptor) molecule and (ii) irreversible
association with midgut membrane. Apparently, not all the
toxins that initially recognize the receptor(s) go on to the
irreversible associated state. This study extends the knowledge
of Cry toxin-membrane interaction beyond the analysis of ini-
tial binding. In this context, it should be noted that the previ-
ous observations (4, 30) in which toxicity was not correlated
with binding should be interpreted with caution. Those con-
clusions were drawn from simple protein blotting and homol-
ogous and heterologous competition binding studies which do
not distinguish the reversible and irreversible steps. The anal-
ysis of the ratio of reversible to irreversible components in
those toxins to the corresponding insects would be necessary
before making any conclusion on the relationship between
toxicity and receptor binding.
The relationship between toxicity and irreversible mem-
brane association is further strengthened by the results of volt-
age clamping experiments with M. sexta midgut membrane.
Our toxicity data and dissociation binding data are consistent
with the voltage clamping data. The toxins that have a greater
irreversible association with the membrane inhibit the Isc more
efficiently than the toxins with reduced membrane association
(Fig. 6). These results suggest that the irreversible binding of
2281
the toxins to BBMV could be used to measure the relationship
between toxicity and binding.
B. thuringiensis d-endotoxin potency requires a number of
factors, including solubilization and prompt proteolytic activa-
tion of the toxin (12), recognition of the specific membrane
receptors present in the susceptible insects, irreversible asso-
ciation with the membrane, and the ability to create ion chan-
nels causing solute leakage leading to membrane disruption.
These results open interesting perspectives for further studies
on the mode of action of B. thuringiensis d-endotoxins. Further
analysis of the irreversible association of Cry toxins to different
insect (susceptible and resistant) BBMVs will possibly answer
the question of whether the relationship between toxicity and
irreversible association to the midgut epithelial cells applies as
a general rule.
ACKNOWLEDGMENTS
We thank Takashi Yamamoto, Gary Powell, and Daniel R. Zeigler
for their critical evaluation and David Matthews (Sandoz Agro Inc.)
for preparing the mutagenic oligonucleotides. We are grateful to
David Stetson for the use of the voltage clamp apparatus.
This work was supported by grants from Sandoz Agro Inc. and from
the NIH (RO1 AI 29092) to D.H.D.
REFERENCES
1. Almond, B. D., and D. H. Dean. 1993. Structural stability of Bacillus thurin-
giensis d-endotoxin homolog-scanning mutants determined by susceptibility
to protease. Appl. Environ. Microbiol. 59:24422448.
2. Chen, X. J., M. K. Lee, and D. H. Dean. 1993. Site-directed mutations in a
highly conserved region of Bacillus thuringiensis d-endotoxin affect inhibition
of short circuit current across Bombyx mori midguts. Proc. Natl. Acad. Sci.
USA 90:90419045.
3. Cunningham, B. C., P. Jharani, P. Ng, and J. A. Wells. 1989. Receptor and
antibody epitopes in human growth hormone identified by homolog-scan-
ning mutagenesis. Science 243:13301336.
4. Garczynski, L. F., J. W. Crim, and M. J. Adang. 1991. Identification of
putative insect brush border membrane-binding molecules specific to Bacil-
lus thuringiensis d-endotoxin by protein blot analysis. Appl. Environ. Micro-
biol. 57:28162820.
5. Ge, A. Z., D. Rivers, R. Milne, and D. H. Dean. 1991. Functional domains of
Bacillus thuringiensis insecticidal crystal proteins: refinement of the Heliothis
virescens and Trichoplusia ni specificity domains on CryIAc. J. Biol. Chem.
266:1795417958.
6. Ge, A. Z., N. I. Shivarova, and D. H. Dean. 1989. Location of the Bombyx
mori specificity domain on a Bacillus thuringiensis delta-endotoxin protein.
Proc. Natl. Acad. Sci. USA 86:40374041.
7. Gould, F., A. Martinez-Ramirez, A. Anderson, J. Ferre, F. J. Silva, and W. J.
Moar. 1992. Broad-spectrum resistance to Bacillus thuringiensis toxins in
Heliothis virescens. Proc. Natl. Acad. Sci. USA 89:79867990.
8. Harvey, W. R., D. N. Crawford, and D. D. Spaeth. 1990. Isolation, voltage
clamping and flux measurements in lepidopteran midgut. Methods Enzymol.
192:599608.
9. Hodgman, T. C., and D. J. Ellar. 1990. Models for the structure and function
of the Bacillus thuringiensis d-endotoxins determined by compilational anal-
ysis. DNA Sequence J. DNA Seq. Mapp. 1:97106.
10. Hofmann, C., P. Luthy, R. Hutter, and V. Pliska. 1988. Binding of the
delta-endotoxin from Bacillus thuringiensis to brush-border membrane vesi-
cles of the cabbage butterfly (Pieris brassica). Eur. J. Biochem. 173:8591.
11. Hofte, H., and H. R. Whiteley. 1989. Insecticidal crystal proteins of Bacillus
thuringiensis. Microbiol. Rev. 53:242255.
12. Huber, H. E., and P. Luthy. 1981. Bacillus thuringiensis delta-endotoxin:
composition and activation. In E. D. Davidson (ed.), Pathogenesis of inver-
tebrate microbial diseases. Allanheld, Osmum and Co. Publishers, Inc., To-
towa, N.J.
13. Ihara, H., E. Kuroda, A. Wadano, and M. Himeno. 1993. Specific toxicity of
d-endotoxins from Bacillus thuringiensis to Bombyx mori. Biosci. Biotechnol.
Biochem. 57:200204.
14. Knight, P. J. K., N. Crickmore, and D. J. Ellar. 1994. The receptor for
Bacillus thuringiensis CryIA(c) delta-endotoxin in the brush border mem-
brane of the lepidopteran M. sexta is aminopeptidase N. Mol. Microbiol.
11:429436.
15. Knowles, B. H., P. J. K. Knight, and D. J. Ellar. 1991. N-acetyl galactosamine
is part of the receptor in insect gut epithelia that recognises an insecticidal
protein from Bacillus thuringiensis. Proc. R. Soc. London B 245:3135.
16. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of
the head of bacteriophage T4. Nature (London) 227:680685.
2282 RAJAMOHAN ET AL.
17. Lee, M. K., R. E. Milne, A. Z. Ge, and D. H. Dean. 1992. Location of Bombyx
mori receptor binding region of a Bacillus thuringiensis delta-endotoxin. J.
Biol. Chem. 267:31153121.
18. Li, J., J. Carroll, and D. J. Ellar. 1991. Crystal structure of insecticidal delta
endotoxin from Bacillus thuringiensis at 2.5 resolution. Nature (London)
353:815821.
19. Lu, H., F. Rajamohan, and D. H. Dean. 1994. Identification of amino acid
residues of Bacillus thuringiensis d-endotoxin CryIAa associated with mem-
brane binding and toxicity to Bombyx mori. J. Bacteriol. 176:55545559.
20. Munson, P. J., and D. Rodbard. 1980. Data analysis and curvefitting for
ligand binding experiments. Anal. Biochem. 107:220239.
21. Oddou, P., H. Hartmann, and M. Geiser. 1991. Identification and charac-
terization of Heliothis virescens midgut membrane proteins binding Bacillus
thuringiensis d-endotoxins. Eur. J. Biochem. 202:673680.
22. Raymond, M. 1985. Presentation dun programme danalysis logprobit pour
micro-ordinnateur Cah. ORSTROM Ser. Entomol. Mer. Parasitol. 22:117
121.
23. Sangadala, S., F. S. Walters, L. H. English, and M. J. Adang. 1994. A
mixture of Manduca sexta aminopeptidase and phosphatase enhances Bacil-
lus thuringiensis insecticidal CryIA(c) toxin binding and 86Rb1-K1 efflux in
vitro. J. Biol. Chem. 269:1008810092.
24. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing with
chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:54635467.
25. Schwartz, J. L., L. Garneau, L. Masson, and R. Brousseau. 1991. Early
response of cultured lepidopteran cells to exposure to delta-endotoxin from
Bacillus thuringiensis: involvement of calcium and anionic channels. Biochim.
Biophys. Acta 1065:250260.
26. Tailor, R., J. Tippett, G. Gibb, S. Pells, D. Pike, L. Jordan, and S. Ely. 1992.
Identification and characterization of a novel Bacillus thuringiensis d-endo-
toxin entomocidal to coleopteran and lepidopteran larvae. Mol. Microbiol.
6:12111217.
J. BACTERIOL.
27. Vadlamudi, R. K., T. H. Ji, and L. A. Bulla, Jr. 1993. A specific binding
protein from Manduca sexta for the insecticidal toxin of Bacillus thuringiensis
subsp. berliner. J. Biol. Chem. 286:1233412340.
28. Van Rie, J., S. Jansens, H. Hofte, D. Degheele, and H. Van Mellaert. 1989.
Specificity of Bacillus thuringiensis delta-endotoxins: importance of specific
receptors on the brush border membrane of the midgut of target insects.
Eur. J. Biochem. 186:239247.
29. Visser, B., D. Bosch, and G. Honee. 1993. Domain-function studies of Ba-
cillus thuringiensis crystal proteins: a genetic approach. In P. F. Entwistle,
J. S. Cory, M. J. Bailey, and S. Higgs (ed.), Bacillus thuringiensis: an envi-
ronmental biopesticide. Theory and practice. John Wiley & Sons, Ltd.,
Chichester, United Kingdom.
30. Wolfersberger, M. G. 1990. The toxicity of two Bacillus thuringiensis delta-
endotoxins to gypsy moth larvae is inversely related to the affinity of binding
sites on midgut brush border membranes for the toxins. Experientia 46:475
477.
31. Wolfersberger, M. G. 1991. Inhibition of potassium-gradient-driven phenyl-
alanine uptake in larval Lymantria dispar midgut by two Bacillus thuringiensis
delta-endotoxins correlated with the activity of the toxins as gypsy moth
larvicides. J. Exp. Biol. 161:519525.
32. Wolfersberger, M. G., C. Hofmann, and P. Luthy. 1986. Interaction of
Bacillus thuringiensis delta-endotoxin with membrane vesicles isolated from
lepidopteran larval midgut. Zentralbl. Bakteriol. Mikrobiol. Hyg. Abt. 1
Suppl. 15:237238.
33. Wolfersberger, M. G., P. Luthy, A. Maure, P. Parenti, F. V. Sacchi, B.
Giordana, and G. M. Hanozet. 1987. Preparation of brush border membrane
vesicles (BBMV) from larval lepidopteran midgut. Comp. Biochem. Bio-
physiol. 86A:301308.
34. Yamamoto, T., and G. K. Powell. 1993. Bacillus thuringiensis crystal proteins:
recent advances in understanding its insecticidal activity. In L. Kim (ed.),
Advanced engineered pesticides. Marcel Dekker, Inc., New York.