Journal of Protein Chemistry, Vol. 14, No. 4, 1995

Domain Organization of Bacillus thuringiensis CrylllA 6-Endotoxin Studied by Denaturation in Guanidine Hydrochloride Solutions and Limited Proteolysis Peter Ort, 1 Igor A. Zalunin, ~ Victor S. Gasparov, ~ Galina G. Chestukhina, 1 and Valentin M. Stepanov 1'~

Received January 17, 1995

Denaturation of Bacillus thuringiensis CryIIIA 6-endotoxin--an insecticidal protein, active against Coleoptera larvae--in concentrated guanidine hydrochloride solutions was pursued by fluorescence and circular dichroism spectroscopy and limited proteolysis. It was found that the protein consists of two fragments that differ by their stability to denaturation by guanidine hydrochloride at pH 3. The less stable fragment corresponds to the N-terminal a-helical domain limited by Leu-279; the more stable one starts with Ile-280, contains about 330 amino acid residues, and corresponds to the molecule C-terminal moiety that consist of its two/3-structural domains forming a superdomain. KEY WORDS: Bacillus thuringiensis ssp. tenebrionis; 6-endotoxin; domains; denaturation of 6endotoxin; limited proteolysis.

1. I N T R O D U C T I O N

limited proteolysis liberate the N-terminal 70-kDa moieties ("true" toxins), several 6-endotoxins are synthesized as 70-kDa proteins analogous to "true" toxins (Hoefte and Whiteley, 1989). The tertiary structure was established solely for the 67-kDa "true" toxin of CryIIIA 6-endotoxin, toxic for colorado potato beetle (Krieg et al., 1987), by X-ray crystallography (Li et al., 1991). The primary structure alignment of 6-endotoxins that belong to different classes and families reveals only a limited number of coinciding amino acid residues. Only five blocks of conservative residues were found when the amino acid sequences of several 6-endotoxins were compared (Hodgman and Ellar, 1990; Sanchis et al., 1989). In contrast, secondary structure predictions evidence similarity of 6endotoxin structural organization (Hodgman and Ellar, 1990). Therefore, it might be presumed that the CryIIIA spatial folding is characteristic for "true" toxins. This protein consists of N-terminal, predominantly c~-helical domain followed by two

Bacillus thuringiensis (BT) 3 produces 6-endotoxins, the proteins that form crystals within the bacterial cell simultaneously with spore formation. More than 40 known 6-endotoxins produced by several BT subspecies form a family of proteins toxic for larvae of various insects. According to the specificity of their action on large insect taxa, 6-endotoxins are divided into classes CryI to CryV (Hoefte and Whiteley, 1989). The latter are subdivided into families and subfamilies. Whereas the majority of these proteins have molecular mass around 130kDa and only after their activation by ~Laboratory of Protein Chemistry, Institute of Genetics and Selection of Industrial Microorganisms, 113545, Moscow, Russia. z To whom correspondence should be addressed. 3 Abbreviations: BT, Bacillus thuringiensis; Gdn-HC1, guanidine hydrochloride; PAGE, electrophoresis in polyacrylamide gel: SDS, sodium dodecylsulfate; CD, circular dichroism.

241 I)277 8033/95/0500 0241$07.50/0 9 1995 Plen um Publishing CoTporation

242

Oft et al.

/3-structural domains (Li et aL, 1991). These domains are thought to fulfill specific roles in the mechanism of toxin action (Li et al., 1991; Tyurin et al., 1987). The c~-helical domain, after reshuffling of its seven a-helixes, forms a transmembrane pore (or ion channel), whereas the second (and possibly the third) /3-structural domain(s) are of critical importance for the toxin interaction with a receptor (Li et al., 1991). Since their presumed functional roles involve the mutual rearrangement of structural domains as well as the reshuffling of structural elements within the domains, the study of 6-endotoxin unfolding in the presence of denaturing agents represents a promising way to analyze the dynamic aspects of 6-endotoxin action on the host cell. We decided to study the denaturation of CryIIIA 6-endotoxin in guanidine hydrochloride solutions using circular dichroism in the far-UV region and fluorescence of tryptophan residues to assess the process. Limited proteolysis of the partially unfolded protein was also applied to this end, taking advantage of the fact that pepsin remains remarkably stable and fully active in guanidine hydrochloride solutions.

citrate buffer, p H 3 or 3.5, at 20~ for 30 min; then the undissolved material was centrifuged off. The solutions for CD measurements were obtained directly by dissolving of crystals in 2-6 M Gdn-HC1, containing 100 mM citrate buffer, p H 3.0. All buffer solutions used in these experiments contained i mM sodium ethylenediaminetetraacetate.

2. MATERIALS AND METHODS

2.4. Spectrofluorimetric Measurements

2.1. Materials Bacillus thuringiensis ssp. tenebrionis was used. The cells were cultivated on trypcasin medium (Chestukhina et al., 1977). 6-Endotoxin crystals isolated by the two-phase distribution system p-xylene-water (Chestukhina et al., 1977) contained less than 0.5% of spores. Guanidine hydrochloride (Gdn-HC1), biochemistry grade, was purchased from Merck. Pepsin was purified in this laboratory by ion-exchange chromatography on aminosilochrom. All other reagents were of analytical grade.

Fluorescence measurements were performed on a Hitachi-MPF-2A spectrofluorimeter with an excitation wavelength of 297 nm. We used 5-nm bandwidth slits in the excitation and emission beams. Cells with 10-mm excitation and emission path length were used. The cell holder was maintained at 20~ Protein concentration was 1.0/xM. To monitor the protein denaturation, the ratio of the emission intensities at 380 and 320nm--the wavelengths flanking the emission maxima of the native (335 nm) and the completely denatured (352nm) protein--was taken as a structure-dependent parameter.

2.2. Dissolving of the Crystals

2.5. Circular Dichroism Measurements

For denaturation experiments at alkaline p H the crystals were suspended in 50 mM NaOH (final protein concentration 10-20/xM) at 20~ for 5 min. Undissolved material was centrifuged off at 6000 rpm. Then the supernatant was adjusted to p H 10 or 11 with 100 mM sodium carbonate buffer, or to 3.5 with 100 mM sodium citrate buffer. For experiments at acidic p H the crystals were suspended in 2 M Gdn.HC1 that contained 100 mM

The CD spectra were recorded on a Jasco J-5000TL spectropolarimeter in the far-UV region at 20~ in 0.1 ram-path-length demountable quartz cells. The protein concentration was 20/~M. The obtained data were expressed as molar ellipticity (0) calculated per averaged amino acid residue. The mean residue molecular mass 113.5 was calculated for the protein from its amino acid composition (Hoefte et al., 1987).

2.3. Denaturation with Guanidine Hydrochloride

6-Endotoxin stock solutions in 100 mM sodium carbonate buffer, pH10.0-11.0, or in 2 M Gdn.HC1, p H 3.0-3.5, were diluted with Gdn.HC1 solutions to establish appropriate Gdn.HC1 concentration (from 0 to 7 M) and pH, then incubated at 20~ for 30rain. Denaturation extent was assessed by the measurement of protein fluorescence and circular dichroism as well as by limited proteolysis. Specially designed P A G E experiments failed to detect any degradation of CryIIIA that might be caused by the eventual presence of an admixture of bacterial proteinases to the 6endotoxin preparation.

Bacillus thuringiensis CrylIIA ~-Endotoxin 2.6. Limited Proteolysis and N-Terminal Sequence Analysis Endotoxin CryIIIA solutions in 100 mM citrate buffer, p H 3 , at appropriate concentrations of Gdn.HC1 were kept for 30 min at 20~ then pepsin was added and the mixture was incubated for 30 min at 37~ Enzyme-toxin ratio varied from 1 : 100 to 1 : 10. The hydrolysis was stopped either (i) by adding cold acetone to a 66% final concentration or (ii) by raising p H to 6.5 with 1 M citrate buffer followed by dialysis against 10 mM citrate buffer, p H 6.5, and adding 35% trichloroacetic acid to 7% final concentration. The precipitates were collected and analyzed by SDS-PAGE. For N-terminal sequence analysis the proteins after S D S - P A G E were immediately transferred to polyvinylidene difluoride membranes. An Applied Biosystems model 470A automated gas-phase sequencer was used.

2.7. Polyacrylamide Gel Electrophoresis PAGE was performed on 10% PAG slabs in the presence of 0.1% SDS at 50mA and 150V as described by Laemmli (1970). The protein samples (lmg/ml) were dissolved in 8 M urea containing 10 mM sodium phosphate buffer, p H 7.5, 1% SDS, and 1% mercaptoethanol. Then 10-50-/xl aliquots of the solutions were introduced into the wells. Protein bands were stained by 0.25% Coomassie blue R-250 with subsequent washing of the gels with 7% acetic acid at 100~ As molecular mass standards, bovine serum albumin (66 kDa), ovalbumin (45kDa), glyceraldehyde-3-phosphate dehydrogenase (36 kDa), carbonic anhydrase (29 kDa), soybean trypsin inhibitor (20.1 kDa), and cytochrome c (12.3 kDa) were used.

243 with 1 M H C I and the necessary amount of guanidine hydrochloride solution was added to attain its necessary concentration. As shown in Fig. 1, at p H 10 a denaturation transition was observed at 4-5 M concentration of Gdn.HC1. At p H 11 the transition appears at lower Gdn-HC1 concentration, 1.5-3.7 M, apparently due to ionization of tyrosine residues and deprotonation of lysines. The fluorescence spectrum of the denatured protein was analogous to that of tryptophan in solution, indicating that practically all tryptophan residues of the unfolded protein became available to the solvent (Kronman and Holmes, 1971). We presume that at least the first (a-helical) and the second (/3-structural) domains, which contain respectively six and four tryptophan residues, are denatured under these conditions. The third (/3-structural) domain does not contain tryptophans, hence its denaturation cannot be pursued by fluorescence measurements. Obviously, the necessity to use exceptionally high p H values to bring 6-endotoxin in solution severely handicaps the study of its denaturation in alkaline solutions. It appears that the 6-endotoxin structure under these conditions, especially at p H 11, becomes rather unstable and vulnerable to chemical deterioration (e.g., desulfurization, etc.). Therefore, we turned to study 6-endotoxin denaturation at pH3.0-3.5, although it was necessary to add 2 M Gdn.HC1 to the mixture to

Ia&o/Iszo 1

2.8. Protein Assay Protein concentration was determined assuming that the A2s0 value of its 0.1% solution was equal to 1.6.

3. RESULTS 3.1. Denaturation of CrylIIA ~-Endotoxin as Assessed by Optical Methods To study endotoxin denaturation at alkaline pH, its crystals were dissolved at p H 12, then the p H value of the solution was adjusted to 10 or 11

gJ

J

y

7

I

J

2 ,,7 f 8a//7 x ~ C/,J /M)

]

t

4'

5

6.

Fig. 1. Fluorescence of CryIIIA endotoxin in Gdn-HCI solutions at different p H values. The 1/xM CryllIA endotoxin solutions containing Gdn.HCI at (1) p H 11, (2) p H 3.0, (3) p H 3.5, and (4) p H 10 were prepared as described in Materials and Methods, Section 2. The ratio of the emission intensities at 380 and 320 nm is plotted on the ordinate.

244

Ort et al.

keep the protein in solution at these pH values. According to the data collected at higher Gdn.HC1 concentrations, 6-endotoxin tertiary structure might be considered as intact or at least being very similar to the native state in 2 M Gdn.HC1. F i g u r e ' l shows the denaturation curve for CrylIIA protein at pH 3.5. Only one transition in 3.2-5 M Gdn.HC1 was observed coinciding with that found at pH 10. At pH 3.0 two transitions can be discerned on the denaturation curve; the first transition takes place in the 2 - 3 M interval, the second at 4 - 5 M Gdn-HC1. Hence, at p H 3 , 6-endotoxin behaves like a structure composed of two cooperative domains, the first one denaturing at 2-3 M Gdn.HC1, the second one stable under these conditions but losing its compactness in 5 M Gdn.HC1. It appears that at pH 3.5 and 10 both these domains denature simultaneously, being equally vulnerable to denaturing agent. It might happen that the alkaline treatment induces loosening of the domain structure (Fig. 1), thus influencing its stability at pH 3.5. To check this assumption, we dissolved CrylIIA protein at pH 12, then transferred this solution into pH3.5 citrate buffer. This sample shows two denaturation transitions, at 2.7-3.5 M and 3.9-4.6M Gdn.HC1 (Fig. 2), whereas only one transition was observed when the protein was directly dissolved at pH 3.5. These data indicate that alkaline pretreatment can indeed influence the stability of 6-endotoxin. Circular dichroism measurements of #endotoxin solutions in Gdn-HC1 at pH 3.0 allowed us to study the predominant secondary structure of the less stale domain. As shown in Fig. 3, the CD

ZY80 /1320 1,j

7,o

9 o,j

Z

Y I

|

2

J

0

-zo

-//~

~, - 5 o

-80

ZOJ

2)15

l 22J

235

) 2#5

J 288

~., H m Fig. 3. CD spectra of CryIIIA endotoxin at different Gdn.HCI

concentrations, CD spectra were recorded for CryIIIA endotoxin (20/xM) at p H 3 in (1) 2M, (2) 4M, and (3) 6M Gdn.HCI solutions.

spectrum of CrylIIA 6-endotoxin in 2 M Gdn.HC1 has a minimum near 222 nm, which is peculiar for proteins with a rather high a-helical content. The CD spectrum in 4MGdn.HC1 is profoundly different in both shape and magnitude of O at all wavelengths: the minimum near 222nm has disappeared, which is characteristic for the destruction of c~-helixes. The CD curve measured in 6MGdn.HC1 solution indicates further disorganization of the secondary structure. These data indicate that the o<-helical domain (the N-terminal in CryIlIA and, presumably, in other 6-endotoxins) is less stable to denaturation caused by Gdn.HC1 than are the two /3-structural domains comprising the C-terminal moiety of the molecule.

3.2. CrylIIA 6-Endotoxin Denaturation with Guanidine Hydrochloride as Assessed by Limited Proteolysis with Pepsin I

I

#

d

8

EC~/z . HGL] CM)

Fig, 2. Fluorescence of CrylHA endotoxin dissolved at pH 12 and transferred into Gdn.HC1 solutions at pH 3.5. The ratio of the emission intensities at 380 and 320nm is plotted on the ordinate.

The action of a proteinase on a protein submitted to denaturation would lead to extended degradation of its denatured domains. Insofar as the denatured protein being transferred from Gdn.HC1 solution in water might regain its compactness, it was necessary to perform the

Bacillus thuringiensis CrylIIA 6-Endotoxin

245 CryIIIA hydrolysis with pepsin in 3 M or 4 M Gdn.HC1 solutions at pH3.0 leads to the accumulation of the major 37-kDa band (presumably a doublet) and minor bands of 48, 31, 29.5, 29, 20, and 16 kDa. Predominant formation of the 37-kDa fragment (or two fragments with similar molecular masses, but possessing "ragged ends") indicates that in 3-4MGdn-HC1 a part of the molecule corresponding to approximately 26 kDa becomes open to proteolysis and suffers extensive degradation by pepsin, whereas the remnant 37-kDa fragment remains sufficiently stable and apparently keeps its native structure. The 37-kDa fragment was isolated by SDSPAGE and its N-terminal sequence Ile-Ala-LeuPhe- was established. Hence, according to the CryIIIA primary structure (Hoefte et al., 1987), the 37-kDa fragment appears due to the cleavage of the Leu-279-Ile-280 peptide bond by pepsin. It should

proteolysis under the conditions used for the unfolding of 8-endotoxin, which requires that the proteinase applied should be stable enough in the presence of a denaturing agent under the experimental conditions. Since swine pepsin has been found to retain its activity in concentrated Gdn-HC1 solutions at p H 3 (data to be published elsewhere), this enzyme was selected to perform limited proteolysis of partially denatured CrylIIA endotoxin. Figure 4 shows the pattern of CrylIIA hydrolysis with pepsin in Gdn.HC1 as assessed by SDS-PAGE. Pepsin failed to attack CrylIIA &endotoxin dissolved in 2 M Gdn.HC1, insofar as only one 63-kDa band corresponding to that of nonhydrolyzed protein was detected after this treatment. In accordance with these data tryptophan fluorescence indicated that CrylIIA protein structure remained stable in 2 M Gdn.HC1 solution at p H 3.

Control Gdn.HCI:

2M

4M

Pepsin 2M

3M

4M

5M

6M

6.7M

66kDa

45kDa

36kDa 29kDa 20kDa

~2kDa

Fig. 4. SDS-PAGE of CrylIIA endotoxin pepsin digests obtained at different concentrations of Gdn.HCI. The CryIIIA endotoxin (1/xM) in Gdn.HC1 solutions containing 0.1 M citrate buffer, p H 3 , was incubated with pepsin (enzyme: endotoxin ratio 1:100) at 20~ for 30min. Aliquots of these mixtures were submitted to electrophoresis. Two lanes correspond to control solutions of the endotoxin, respectively, in 2 and 4M Gdn.HCI, to which no pepsin has been added, while the others correspond to pepsin hydrolysates obtained in the presence of indicated Gdn.HCI concentrations. Positions of the molecular mass markers are as indicated.

246

Ort et al.

be noted that this bond satisfies fairly well pepsin specificity (Zinchenko et al., 1976). In the tertiary structure of CryIIIA these residues are located within the seventh a-helix near its C-terminus (Li et at., 1991): 259

;

285

YESWVNFNRYRREMTLTVLDLIALFPL It appears that this helix is unfolded along with other a-helixes of the unstable N-terminal domain, thus making the proteolytic attack feasible. Thus,

Fig. 5. Time course of pepsin digestion of CrylIIA endotoxin in 4M Gdn-HC1. The CryIIIA endotoxin was incubated with pepsin (enzyme: endotoxin ratio 1:100) in 4M Gdn.HCI that contained 0.1 M citrate buffer,pH 3. Aliquots were withdrawnat the following time intervals: Lane 1:100 min; lane 2:60 min; lane 3:30 rain; lane 4:16 min; lane 5:9 rain; lane 6:4 rain. Lane 7 contains the molecular mass markers: bovine serum albumin (66kDa), ovalbumin (45 kDa), glyceraldehyde-3-phosphatedehydrogenase (36kDa), carbonic anhydrase (29kDa), soybean trypsin inhibitor (20.1kDa) and cytochrome c (12.3 kDa).

the hydrolysis with pepsin proceeds practically on the borderline between the N-terminal a-helical domain and /3-structural domains forming the C-terminal part of the CryIIIA molecule (Li et al., 1991). The molecular mass of the CrylIIA 280-644 fragment would correspond to 41.4kDa, which is 4.4kDa higher than the molecular mass of the 37-kDa fragment. We presume that the protein C-terminal structure, comprised of approximately 40 amino acid residues, also becomes partially unfolded and degraded. This fragment would embrace the last four /3-strands of the third (/3-structural) domain. Hence, the 37-kDa fragment, stable toward pepsin attack in 3 - 4 M Gdn.HC1, consists of the second domain and a major part of the third /3-structural domain (Li et al., 1991). It has been shown that the increase of the incubation time of CryIIIA with pepsin in 4 M Gdn.HC1 from 6 to 100 min leads to accumulation of the 37-kDa fragment (Fig. 5). Analogous results were obtained when the amount of pepsin was increased from 1 : 100 to 1 : 10 enzyme-cryIIIA ratio (data not shown). Further increase of the denaturing agent concentration, to 5 and 6 M Gdn-HC1, gave rather blurred results. Rather unexpectedly, 20, 27, and 4 4 k D a were observed (Fig. 4), the last two of which were not registered at lower Gdn.HC1 concentrations. In our opinion it would be premature to interpret these data in structural terms. Tryptophan fluorescence measurement indicated that at least the first and the second domains were denatured in 5 M Gdn-HC1, but the absence of complete degradation of the protein with pepsin suggests that the protein is not yet completely unfolded. The contribution of the pepsin partial inactivation to the observed effect also seems quite plausible. In 6.7 M Gdn.HC1 no hydrolysis with pepsin was observed, due to its inactivation, and solely the 63-kDa band appeared on S D S - P A G E .

4. DISCUSSION

The molecular structure of Bacillus thuringiensis 6-endotoxins is characterized by the presence of structurally autonomous domains (Chestukhina et aL, 1982, 1990). The domain organization seems to be involved in the mechanism of entom6cidal

Bacillus thuringiensis CryIIIA 6-Endotoxin activity of these proteins, with each of the domains responsible for a definite stage in 3-endotoxin interaction with the host cell membrane (Li et al., 1991; Tyurin et al., 1987). It is conceivable that the dynamics of 3-endotoxin domain organization is of crucial importance for understanding the behavior of these structures in the course of the protein action on the host cells. To shed light on the problem, we studied the unfolding of CryIIIA 3-endotoxin produced by Bacillus thuringiensis subspecies tenebrionis, the only member of this protein family for which the tertiary structure has been established by X-ray crystallography (Li et al., 1991). To this end we studied CryIIIA denaturation with increasing Gdn.HC1 concentrations. Two main approaches were applied to follow CryIIIA 6-endotoxin denaturation. The measurement of tryptophan residue fluorescence by spectrofluorimetry allowed us to study the unfolding of two of the three domains present in the CryIIIA structure, the first, an a-helical domain, and the second, a /3-structural one, whereas the third domain, being devoid of tryptophan residues, escapes observation. CD measurement allowed us to follow the behavior of the a-helical and /3-structural domains, insofar as the contribution of a-helixes and /3-structures to the CD spectrum is characteristic enough. Another approach consisted in the application of limited proteolysis of CrylIIA 6-endotoxin with pepsin under conditions that favor protein denaturation, in concentrated Gdn.HC1 solutions. It ought to be stressed that the latter method allows us to detect even temporary unfolding of the domains that would be completely degraded by the proteinase; therefore its results would not necessarily agree with that obtained by optical measurements. The data on CryIIIA denaturation by increasTable I. Denaturation Transitions Observed for CryIIIA 6Endotoxin p H at which the endotoxin was brought in solution

Final p H at which fluorescence was measured

Gdn.HCI concentrations at which denaturation transitions were observed (M)

12 12 12 3.5 3.0

11 10 3.5 3.5 3.0

1.5-3.7 4.0-5.0 2.7-3.5; 3.9-4.6 3.2-5.0 2.0-3.0; 4.0-5.0

247 ing Gdn.HC1 concentrations are presented in Fig. 1 and summarized in Table I. At p H 3 . 5 and 10, CryIIIA 6-endotoxin revealed only one denaturation transition at 3.2-5M and 4-5 M Gdn.HC1, respectively. This means that the o~-helical and both /3-structural domains became denatured at the same conditions. Apparently, rather tight interaction between the domains under these conditions leads to the denaturation of all three domains as a cooperative structure. At p i l l 1 still one transition was observed, which proceeded at lower (1.5-3.7M) Gdn.HC1 concentration, indicating substantial loosening of the structure of all domains. This might be explained by the unfavorable influence of dissociation of tyrosine phenol groups as well as deprotonation of lysine residues on the overall stability of the structure. Thirty-one lysine residues are present in the CryIIIA structure, 13 in the first a-helical domain and 9 each in the second and the third/3-structural domains (Hoefte et al., 1987; Li et al., 1991). The tyrosine residues are distributed as follows: 13 in the first, 15 in the second, and 9 in the third domain. It appears that at p H 3 the interaction between the domains becomes loose, inasmuch as at this p H the denaturation curve reveals two transition steps, between 2-3 M and 4-5 M Gdn.HC1. Hence, one of the domains suffered the unfolding of its spatial structure at a substantially lower (2-3M) concentration of the denaturing agent. Destabilization at this p H value might be a result of aspartic and glutamic acid w-carboxyl group protonation. It appears that the remaining part of the molecule retained its stability, undergoing denaturation within the same Gdn-HC1 concentration range as observed at p H 3.5 or 10. As evidenced by its N-terminal sequence, Ile-Ala-Leu-Phe-, the fragment with the molecular mass of 37 kDa, a major one isolated after limited proteolysis at p H 3 in 4 M Gdn.HC1, i.e., under the conditions that lead to selective denaturation of one domain of CryIIIA protein, appears due to a cleavage of the Leu-279-fle-280 peptide bond by pepsin. Hence, it is the first domain limited by Leu-279 that suffers unfolding and extended digestion. It corresponds fairly well to the ahelical domain as described by Li et aI. (1991), which starts at Asp-58 and proceeds until Leu-290, whereas the last, the seventh a-helix, becomes cleaved in the middle. This conclusion is supported by the disappearance of the traits characteristic for

248

a-helical structures observed by CD measurement in 4 M Gdn.HC1 solution of CrylIIA protein at p H 3.0. The fragment that starts with Ile-280 and proceeds until approximately the 600th residue, thus comprising the second [from Tyr-291 to Phe-500, according to Li et aI. (1991)] and the major part of the third (from Phe-501 to Asn-644) domains, on the contrary, remains relatively stable to the further attack by pepsin. It might be presumed that these two /3-structural domains (with the exception of approximately 40 amino acid residues cleaved off from the C-terminal) form a cooperative structure, a kind of a "superdomain." The data on CryIIIA protein domain organization acquired by study of its denaturation with Gdn.HC1, followed by fluorescence spectroscopy and limited proteolysis with pepsin, are consistent with the structure of this 6-endotoxin established by X-ray crystallography. The denaturation behavior of CryIIIA endotoxin might give some hints on the dynamic properties of the 6-endotoxin molecule relevant to its function. The pattern of CryIIIA 6-endotoxin denaturation with Gdn.HC1 at p H 3.5, 10, and 11 indicates that this molecule behaves as a unitary cooperative structure within a rather broad set of conditions, evidently due to tight interdomain interactions. The functional importance of the latter is stressed by the fact that two conservative blocks of amino acid residues--blocks II and III (Sanchis et al., 1989)--are located on the interfaces between, respectively, the first (o~-helical) and the second (/3-structural) domains and the second and the third /3-structural domains. On the other hand, under certain conditions, remarkable differences were observed in the stability of the CryIIIA 6-endotoxin domains. Thus, at p H 3 in concentrated Gdn-HC1 solutions CryIIIA 6-endotoxin suffers an abrupt structural transition due to destabilization of the N-terminal a-helical domain that becomes vulnerable for proteolytic degradation. The same domain is thought to rearrange in the course of 6-endotoxin interaction with the host cellular membrane, which leads to the formation of a transmembrane ion channel (Li et al., 1991). Our observation that the a-helical domain is inclined to loose its structure selectively seems to be in line with the latter assumption. It is conceivable that the available data would not allow us to identify the cause of

Oft et al.

6-endotoxin rearrangement in vivo, but the demonstration of the intrinsic differential stability of the 6-endotoxin domains is relevant for further elucidation of its mechanism of action. Earlier Convents and co-workers (1990, 1991) studied Gdn.HC1 denaturation of "true" toxins formed from CryIA(b) and CryIC 6-endotoxins. within a broad p H range. At p H 8.0 CryI "true" toxin denatured as a unitary structure, although at p H 4 and 11 two denaturation transitions were observed, which indicated the presence of two domains presumably analogous to those found in CryIIIA endotoxin. On the other hand, the relative stability of CryI "true" toxin domains toward denaturation with Gdn.HC1 differed substantially from that revealed by CryIIIA protein. Thus, it seems that the N-terminal a-helical domain of CryI "true" toxin is more stable to Gdn.HC1 unfolding at p H 4 than the/3-structural domain (or domains). The reason for this difference is unclear. It might be caused by individual differences in the relative stability of the domains, although one cannot exclude an artifactual influence of alkaline pretreatment of "true" toxins in the course of their dissolution. Thus, it might be presumed that CryIIIA protein and CryI "true" toxins share a common type of molecular organization that consists of two cooperative structures, a domain that roughly corresponds to the N-terminal moiety and a "superdomain" comprising the C-terminal half of the molecule, although the relative stability of these domains toward p H and Gdn.HC1 concentration is different for these proteins.

ACKNOWLEDGEMENT

Support by the Russian Foundation for Basic Research under grant 93-04-20263 is gratefully acknowledged.

REFERENCES Chestukhina, G. G., Kostina, L. I., Zalunin, I. A., Kotova, T. S., Katrukha, S. P., Kuznetsov, Yu. S., and Stepanov, V. M. (1977). Biokhimiya 42, 1660-1667. Chestukhina, G. G., Kostina, L. I., Mikhailova, A. L., Tyurin, S. A., Klepikova, F. S., and Stepanov, V. M. (1982). Arch. Microbiol. 132, 159-162. Chestukhina, G. G., Tyurin, S. A., Kostina, L. I., Osterman, A. L., Zalunin, I. A., Khodova, O. M., and Stepanov, V. M. (1990). J. Protein Chem. 9, 501-507.

Bacillus thuringiensis CrylIIA 6-Endoloxin Convents, D., Houssier, C., Lasters, I., and Lauwereys, M. (1990). J. Biol. Chem. 265, 1369-1374. Convents, D., Cherlet~ M., van Damme, J., Lasters, I., and Lauwereys, M. (1991). Eur. J. Biochem. 195, 631-635. Hodgman, T. C., and Ellar, D. J. (1990). J. DNA Sequence l, 97-106. Hoefte, H., and Whiteley, H. R. (1989). Microbiol. Rev. 53, 242-255. Hoefte, H., Seurinck, J., Van Houtven, A., and Vaeck, M. (1987). Nucleic Acids Res. 15, 7183. Krieg, A., Schnetter, W., Huger, A. M., and Langenbruch, G. A. (1987). System. Appl. Microbiol. 9, 138-141.

249 Kronman, M. J., and Holmes, L. G. (1971). Photochem. Photobiol. 14, 113-134. Laemmli, U. K. (1970). Nature 227, 680-685. Li, J., Carroll, J., and Ellar, D. J. (1991). Nature 353, 815-821. Sanchis, V., Lereclus, D., Menou, G., Chaufaux, J., Guo, S., and Lecadet, M.-M. (1989). Mol. Microbiol. 3, 229-238. Tyurin, S. A., Chestukhina, G. G., Osterman, A. L., Khodova, O. M., Timokhina, E. A., Klepikova, F. S., and Stepanov, V. M. (1987). Biokhimiya 52, 918-926. Zinchenko, A. A., Roomsch, L. D., and Antonov, V. K. (1976). Bioorg. Khirn. 2, 803-810.