CA1340002C - Obtaining proteins which are pathogenic to insects and microorganisms ofthe bacillus thuringiensis type - Google Patents

Abstract

Bacillus thuringiensis DSM
4082, its mutants which synthesize proteins pathogenic to insects, and a process for obtaining, by protoplast fusion, two or more strains having different toxins are described.

Description

I ~ 4 !'~

Obtaining proteins which are pathogenic to insects and microorganisms of the Bacillus thuringiensis type In addition to synthetic insecticides and insec-ticides isolated from certain plants, bacterial insec-ticides are known; these are proteins which are formedin the wild type and in various mutants of 3acillus thuringiensis (B.t.), which has been known since 1915.
Under unfavorable environmental conditions, bacillus species form surviving stages (endospores) which are heat-resistant and resistant to drying out. However, during the sporulation, Bacillus thuringiens is addition-ally forms a parasporal inclusion body which is clearly visible under the optical microscope and, because of its crystal-like structure, is si0ply referred to as a cry-stal. These protein crystals are responsible for thetoxicity of 8.t. (Angus 1954). After oral intake, they dissolve in the alkaline medium of the intestine of the target insects. The action of intestinal proteases (Tojo et aL. 1985) and crystal-associated proteaseS (Turley et al. 1985) result in the formation of toxic cleavage pro-ducts (Delta-endotoxins), which are still not known ex-actly and which finally lead to lysis of intestinal epi-thelial cells. The molecular mechanism of the Delta-endotoxin action is not yet known exactly.
The many 3.t. strains isolated are now usually classified by comparison of the flagellum antigens (De ~arjac and Bonnefoi 1973; De Barjac 1981). On this basis, it has been possible to distinguish between 2Z H-antigen groups to date. Moreover, the strains are divided into pathotypes on the basis of their action spectrum. More than 95~ of the B.t. isolates found to date form ortho-rhombic crystals and are effective against butterfly lar-vae (pathotype A). Pathotype B includes varieties having activity against mosquito larvae and comprising round, irregular crystals, while pathotype C includes strains having activity against beetle larvae and comprising lamellar crystals. For some crystal-forming B.t. subspecies q n ~

(eg. tochigiensis, kumantoensis), no insecticidal activ--ity is known (pathotype ND). Furthermore, strains which are pathogenic to butterflies and have Low mosquitocidal activity have also been described. These isolates form S not only lepidopter-specific crystals (protein P1: molecu-lar weight about 130,000) but also a small oval inclusion body (protein P2: molecular weight about 65,000) which has a weak action against butterflies and mosquitos (lizuka and Yamamoto 1983).
There is no correlation between the serotype and pathotype. For example, pathotype A, 9 and C strains having the same flagellum antigens (H8) has been isolated.
In all 9.t. strains, from 2 to 12 plasmids having molecular weights of from 1.5 to ~ 150 Md are detectable (Carlton and Gonzalez 1985a). Curing experiments (Gonzalez et al. 1981), crossing experiments (Gonzalez et al. 1982) and cloning experiments (McLinden et al. 1985) have shown that the ability to form crystals is plasmid-determined in many strains. The genes involved may be located on one or more plasmids (30-150 Md). Hybridization experi-ments with cloned crystalline genes indicated great simi-larity with the Delta-endotoxin genes within the patho-type A group (Lereclus et al. 1982; Kronstadt et al.
1983). By cloning a Delta-endotoxin gene, it has there-fore been possible to find further crystalline genes of the same pathotype very simply by hybridization with the specific sample. In this manner, it has been possible to detect, for example, 3 or more Delta-endotoxin genes in B.t. kurstaki HD-1 (Kronstad and whiteley 1986).
The function of most of the plasmids in ~.t. is unknown. Apart from the crystalline protein genes, two further plasmid-coded properties have now been described.
Tam and Fitz-James (1986) have been able to assign the formation of phage-like particles in 9.t. israelensis to a 68 Md plasmid. These particles are also formed during sporulation and are visible as further small inclusion bodies, in addition to the Delta-endotoxin crystal.

Finally, Ozawa and Iwahana (1986) demonstrated the coup-ling of crystal formation and 3-exotoxin formation (Sebesta et al. 1981) in a 62 Md plasmid in B.t. darm-stadiensis.
B.t. strains are physiologically indistinguish-able fror 8acillus cereus (Baumann et al. 1984). All H-antigens characterized in B.t. to date have been found in B. cereus strains isolated from soil samples (Ohba and Aizawa 1986). Thus, the only difference between B.t.
and 3. cereus is the formation of the Delta-endotoxin crystal during sporulation. However, this property may easily be lost through plasmid curing. Hence, B.t. must be regarded as a variety of B. cereus.
The action spectra of various strains also dif-fer greatly within the same pathotype (Krieg and Langenbruch 1981). For optimum insect control, it is therefore practical to use different B.t. strains. Since the growth properties of various B.t. varieties are very different, fermentation has to be optimized separately ;'0 for each strain. To date, this has been impossible owing to the high costs entailed. One possible method of im-proving this situation would be to use a production strain into which the genetic information for the formation of various Delta-endotoxins is subsequently introduced as required. The transfer of crystalline plasmids has been possible to date only through the conjugation-like mechan-ism discovered by Gonzalez et al. (1982). This method has made it possible for plasmids to be transferred from strains which are active against butterflies at a higher ;0 rate (up to 80~) to cry strains and even to B. cereus.
However, the transfer of plasmids is greatly dependent on the crossing partners. Transfer of crystal formation was cletectable only in certain hybrids.
Molecular genetic handling of the Delta-endotoxins 'iS is facilitated by virtue of the fact that the genes in-volved are generally localized on plasmids. Several crystalline genes in E. coli or B. subtilis have already 139~00~

been cloned and sequenced (Schnepf and Whiteley 1981;
~eld et al. 1982; Klier et al. 1982; Shibano et al. 1985;
Schnepf et al. 1985; Adang et al. 1985). The high ex-pression (up to 30% of the sporulated cells) is charac-teristic of the Delta-endotoxin production in 8.t. In E. coli and B. subtilis, the expression with cloned genes is significantly poorer. This may be explained, for exam-ple, in terms of the different regulation mechanisms in these organisms. In order to avoid the problems, it would be desirable to investigate the Delta-endotoxins directly in B~ thuringiensis and B. cereus. Since there was no suitable transformation method in B.t., it has been im-possible to date to carry out such experiments (Carlton and (ionzalez 1985b; Aronson et al. 1986).
It is an object of the present invention to pro-vide an efficient transformation system which, as far as possible, permits all Bacillus thuringiensis and Bacillus cereus strains to be transfor0ed, forming new strains which have the ability to produce a plurality of toxins in the same organism. In the absence of natural compet-ence, one possible method to induce absorption of DNA is protoplast transformation (Bibb et al. 1978). This method was the basis of the novel transformation process. The protoplast transformation was developed from protoplast fusion (Schaeffer et al. 1976; Fodor and Alfoldi 1976).
In thle case of cell fusion, the cell contents of both parent strains are combined by fusion of the cell mem-branes, and recombination of genes and redistribution of plasmids may occur.
We have found that this object is achieved by a process for obtaining microorganisms of the ~acillus thuringiensis type, which are genetically fixed and have the ability to form endotoxins, by protoplast fusion of two ar more strains having different toxins, wherein the cell wall is degraded in a first step, fusion is induced in a second step and finally regeneration of the cell wall is induced.

1 ~40002 The present lnventlon relates ln partlcular to a novel straln of Baclllus thurlnglensls having hlgh actlvlty agalnst larvae of butterfly pests and beetles.
The straln has the name Baclllus thurlnglensls F 8-1 and was deposlted under No. 4082 at the Deutsche Sammlung fur Mlkroorganlsmen, 1)-3400 Gottingen on 10 Aprll 1987.
In order to achieve protoplast fusion of the cells, the cell wall must flrst be degraded. This ls achieved by lncubatlng the cells ln an osmotlcally stablllzed buffer with lysozyme. Transformatlon or fusion is lnduced by hlgh concentratlons of polyethylene glycol. The actlon of the polyethylene glycol ls not known exactly. After the actual transformation or fusion, an lntact cell wall must be bullt up agaln by the protoplast (regeneratlon~. An lmportant step ls the selectlon of transformed or fused clones. There are two possibilitles for this in lndlrect selectlon, the protoplasts are f~rst regenerated on a sultable medlum. Thls ls foll~wed by selectlon, for example by superposition on a selecti~n medium. Direct selection is more advantageous.
Here, the choice of sultable conditlons can ensure that only selected clones grow on the regeneratlon medlum.
The currently known B.t. lsolates which are toxlc to lnsects can be dlvlded lnto a plurallty of pathotypes on the basls oE the actlon spectra of their Delta-endotoxlns.
Stralns whlch are active against Lepidoptera are assigned to pathotyl?e A, those whlch are actlve against Diptera are assigned to pathotype B and those which are active against Coleoptera are assigned to pathotype C.

13~0002 TAPLE ~
Serotype Name Pathotype Strain (eg.) 1 thuringiensis A HD-1, HD-20 3a 3b kurstaki A HD-1, HD-73 S 5a Sb galleriae A HD-29, HD-950 6 entomocidus A HD-967 7 aizawai A HD-135, HD-980 8a 8b morrisoni A, B, C HD-12, ~I 256-82 (tenebrionis PG 14) 11 kyushuensis C
14 israelensis C A60-, HD-567 Table 1 gives an overview with selected examples;
serotypes and pathotypes cannot be correlated. For exam-ple, pathotype A, P and C strains having the same flagel-Lum antigens (H8) have been isolated.
The preparations commercially available to date are based on a few pathotype A strains for controlling some butterfly species which cause damage in vegetable and fruit cultivation and in forestry, and on a pathotype H strain for controlling mosquito and blackfly larvae.
Particularly when used in agricultural crops, the known strains and toxins are not always completely satisfactory. In order to provide the user with a feas-;'S ible alternative to known, in particular synthetic activeingredients, it is necessary to find novel strains hav-ing improved properties, ie. higher toxicity for certain pests, an extended action spectrum, for example against many Lepidoptera species and/or a combined spectrum, for example against the larvae of Lepidoptera and Coleoptera.
Since each individual isolate has only a narrow action spectrum, it should be possible to create novel strains by combining a large number of insecticidal specl:ra.
i5 With regard to production and application, the properties of the combination strains must be superior .' .) {~ Z

to those of a straightforward mixture of two individual strains.
In all El.t. strains, from 2 to 12 plasmids having molecular weights of from 1.5 to > 150 Md are detectable (Carlon, Gonzalez 1985 in Molecular Eliology of Bacilli, Academic Press). Curing experiments (Gonzalez et al.
1981, Plasmid 5, 351 et seq), crossing experiments (Gonz-alez et al. 1982, Proc. Natl. Acad. Sci. USA 79, 6951 et seq) and cloning experiments (McLinden et al. 1985, Appl.
Environm. Microbiol. 50, 620 et seq) have shown that the ability to forT crystals is plasmid-coded in many strains.
The 3enes involved are located on one or more plasmids (30-150 Md).
Hybridization experiments with cloned crystalline gene; indicated great similarity of Delta-endotoxin genes within the pathotype A group (Lereclus et al. 1982, Mol.
Gen. Genetic 186, 391 et seq; Kronstad et al. 1983, J.
Elacteriol. 154, 419 et seq).
E~y cloning a gene, it is possible to find further 0 cryscalline genes of the same pathotype by hybridization with the specific sample.
In this way, for example, three or more Delta-endol:oxin genes have been detected in El.t. var. kurstaki (HD-I) (Kronstad, Whiteley 1986, Gene 43, 29 et seq).
E~.t. strains are physiologically indistinguish-able from El. cereus (Baumann et al. 1984). In B. cereus straiins isolated from soil sa~ples, it has been possible to find the H-antigens characterized to date for E~.t.
(Ohba, Aizawa 1986, J~ Elasic Microbiol. 26, 185 et seq).
iO The c\nly difference between 3.t~ and El.c. is thus the formation of the Delta-endotoxin crystal during sporulation.
Transfer of plasmids isolated from E~.t. has been described in various publications. In most cases, the host in which the genes or gene fragments are expressed are E. coli (U.S. Patent 4,448,885, European Patent 63,039 and European Patent Applications 93,062 and 186,379). How-ever, the expression of E~ti toxin is also possible, for 1340~02 examPle~ in a. megaterium and ~. subtilis (European Patent 195,285).
Further work on this subject includes the follow-ing: Schnepp, Whiteley 1981; Proc. Natl. Acad. Sci. USA
S 78, 2893 et seq, Kronstad et al. 1983, J. ~acteriol. 154, 419 et seq, Held et al. 1982, Proc. Natl. Acad. Sci. USA
79, 6065 et seq, Klier et al. 1982 and EMBO Journal 1, 791 et seq.
However, expression of the cloned genes is as a rule poor in E. coli and the other hosts, in contrast to that in the original host P.t., for which a high Delta-endotoxin production of up to 30~ of the sporulated cel.
is typical.
A protoplast/liquid system which permits bidirec-tional transfer of genetic information is described below.The primary fusion product is a diploid cell in which the total DNA from both parent strains is present. After segregation in stable haploid cells, mixtures of the plasmids from both parents are present. There is a high frequency of clones in which the ability to form crystals has been transferred from one parent to the other.
A precondition for the fusion is protoplastization of the cells (Shaeffer et al. 1976, Proc. Natl. Acad.
Sci. USA 79, 2151 et seq, foder, Alfoldi 1976, Proc. Natl.
Acad. Sci. USA 73, 2142 et seq). Degradation of the cell wall is achieved by incubation of the cells in an osmoti-call~ stabilized buffer with lysozyme.
Fusion is induced by high concentrations of poly-ethyLene glycol (PEG). When fusion is complete, an in-S0 tact cell wall must be built up again from the protoplast (regeneration).
An important step is the selection of fused clones; there are two possibilities: in indirect selec-tion" the protoplasts are first regenerated on a suitable ~SS medium, after which selection is effected, for examPle by superposition on a selection medium. Direct selec-tion is more advantageous; here, by choosing suitable 13400~2 conditions, it is possible to ensure that only selected colonies grow on the regeneration medium.
To simplify the selection, auxotrophic and apatho-genic cry mutants were constructed, as well as, by means of transformation, phenotypes containing antibiotic markers.
The technique described can in principle be used to construct all possible Delta-endotoxin combinations, which may lead to altered, broader, more active proper-ties from the biological point of view.

? ~ ~

The following strains can, for example, act as donors:
No.
1 Bacillus thuringiensis var. thuringiensis eg. HD 2, HD 20 2 var. finiti~us --S 3a 3b var. kurstaki eg. HD 1 3a var. alesti 4a 4b var. sotto HD-772 4a 4b var. dendrolimus 4a 4b var. kenyae 10 Sa 5b var. galleriae HD 29, HD ~S0 6 var. subtoxicus var. entomocidus HD 967 7 var. aizawai HD 980, HD 135 8a 8ib var. morrisoni HD 12 var. tenebrionis 8i-256-82 9 var. tolworthi var. darmstadiensis 11a 11b var. toumanoffi 20 11a 11c var. kyushuensis 12 var. thompsoni 13 var. pakistani 14 var. israelensis A60, HD 567, 25 15 var. dakota 16 var. indiana 17 var. tohokuensis 18 var. kumamotoensis 19 var. yunnanensis 'iO Z1 var. colmeri - var. wuhanensis 8acillus sphaericus ATCC 29 203 8acillus popilliae ATCC 14 706 These strains are generally available.
The abovement-ioned strains are also suitable as reciplients; however, it is also possible to use, for 13~0002 example, the following:
Bacillus subtilis DS 11 402 Bacillus cereus ATCC 21 281, DSM 31 351 E. coli 5 Bacillus megaterium Another possibility of reducing to a minimum the ecotoxicological pollution of the biotope to be treated is to grow asporous mutants of these strains obtained by protoplast fusion. Examples of the lack of spores are described in, for example, EP-A 59 460 or EP-D-178 151.

The Table describes a selection of constructed strains in which various properties were transferred by protoplast fusion. The list is restricted to strains tS with crystal formation.
Strain Biological activity Pathotype A Pathotype C
HD-2 cry thur (wild t~pe) +
HD-2 cry Z0 HD-Z cry x HD-2 cry thur. +
HD-2 cry x HD-73 cry kurstaki +
B1 2'i6-82 cry ten (wild type) - +
HD-2 cry B1 256-82 c-y ten (F32--13) _ +
F32-13 x HD-2 cry thur. (F8-1) + +
HD-2 cry x HD-8 cry ~al.
(F19-.30) +
8. cereus HM7-1 B. cereus HM7-1 x HD-8 cry gal.
(F12-1) +
B. cereus HM7-1 HD-2 cry thur.
(F4-3) +
. ce!reus HM7-1 HD-73 cry kurstaki (F10) The Table gives an overview of the sensitivity 1 r ~

of B.t. strains with regard to protoplastization in the different media STML, MML and SMMP M/L.
STML MML SMMP M/L
BI 256-82 ++ ++A +A
5 A 60 ++ - -PG 14 ++ +A -/+
HD 8 PC 194 ++ ++A
HD 70 30 ++ ++ +
HD 1 ++ ++ -/+
lO HD 2 D Z6 ++ + ++
B.C. DSM 31 ++ ++ -/+
A = Aggregation The Table gives an examDle of homologous fusion~S between auxotrophic mutants of the strain HD 2.
TABLE
Homologous fusion between auxotrophic mutants of the strain HD 2 (var. thuringiensis) Parents:
Reci~ient D6-4 met , arg , Smr, TcS, crys Donor D12 (pBC 16) met , arg , SmS, Tcr, crys Fusion rates: 2.0X of TcrSmr colonies 1.2X of prototrophic colonies Analysis of the TcrSm~ colonies 25 met arg cryspercentage - + + 25~6 _ + - 3~6 + + + 48~8 + + - 1~2 - - + 9~8 - - - 1~2 + _ + 9.8 The procedure for protoplast fusion is described briefly below:

l3~n~l~n~

1) Harvest the cells from 5 ml of culture (autolysis stage or overnight cuLture'l Z) Resuspend cells in 1 ml of protoplastization buffer 3) Incubate for 2 hours at Z8~C

4) Wash protoplasts twice in SMMP 0edium S) Resuspend orotoplasts in 0.5 ml of SMMP medium 6) Prepare fusion batch:
0.2 ml of protoplast donor + 0.2 ml of protoplast recipient 0.3 ml of 50~ strength PEG in SMM buffer 7) Incubate for S minutes at room temperature 8) Add 5 ml of SMMP medium ana remove protoplasts by cen-trifuging 9) Resuspend protoplasts in 1.5 ml of SMMP medium 10) Incubate protoplasts for 5 hours at 28~C
11) Prepare dilution series in SMM buffer and plate out aliquots on regeneration agar.

The Table gives an example of fusion between Z0 heterologous strains, where the crystal-forming property is transferred.
TABLE
Fusi~n between heterologous strains Fusi~n F6 Donor: HD 73-30 (pBC 16) met , arg , Sts, Tcr, crys Recil~ient: HD 2 D6-4 met , arg , Str, Tcs, crys Fusion rate: O.S~ of Str, Tcr colonies Str, Tcr colonies per plate: 10 iO The Table gives an example of fusion in which two HD-2 mutants with different crystal-forming properties were combined.
Donor: F 32-19 met~, pur~, cry+ten Reci~ient: HD-2 D 10 met , pur , cry thur.
i5 Fusion rate: 0.7~
Prototrophic colonies: > 100/plate Microscopic testing of colonies (76) Pathotype A crystal 26.3%
Pathotype C crystal 7.9X
A + C crystals 65.8 rhe Table gives an example of fusion of E;.t.
protoplasts with B.c. protoplasts.
Fusion F12 Donor: E. thuringiensis HD 8 K4-3 (paC 16) Recipient: ~. cereus UM 7-1 (pCM 194) his , nic , Sm , Tc , Cm , crys Analysis sf the TcrSmr colonies his , nic , Cmr, crys 4 his , nic , Cm5, crys 5 15 his , nic , CmS, crys 2 his , nic , Cmr, crys 2 EXAMPLE OF USE
The biological action of various fused strains is summarized in the Table.
The experiments were carried out as follows:
a) PLutella maculipennis (diamondback moth) Ingestion and contact test Young kohlrabi leaves are immersed for 3 seconds .'5 in an aqueous preparation of the test substance and placed on a circular filter moistened with 0.5 ml of ~ater, in a plastic beaker. 10 larvae in the 2nd to 3rd stage are then placed on the leaf.
Consumption and mortality are assessed 3 days :50 after the beginning of the test.
b) Si~odoptera littoralis (Egyptian cotton owlet moth) 3reeding test on synthetic nutrient medium ~ reeding is carried out in 100 ml plastic beakers, on about 50 ml of the standard nutrient ~edium, into which, ;5 in the liquid state, the active ingredient is carefully mixed. For each concentration, 10 beakers containing a 3 ~

larY,3 (L3) of 10-12 mm in length are prepared. Observa-tion continues as a rule until the moth hatches.
c) Aedes aegypti (yellow-fever mosquito) Contact and ingestion test S 250 Tl plastic beakers are filled with 200 ml of tap water at 23~C, and 20 Aedes larvae (L2) are intro-duced. Thereafter, the test substance is added to the vessel, and the mortality is determined after 24 hours.
d) Leptinotarsa decemlineata (Colorado potato beetle) Ingestion and contact test Pieces punched from young potato leaves are im-mersed for 3 seconds in an aqueous preparation of the test substance and then placed in palette compartments (1 punched piece per compartment); a previously weighed potal:o beetle larva (L3) is then used per compartment;
the number of replications n is 10. Consumption, weight change of the larvae (mg) and mortality are evaluated 3 days after the beginning of the test.
~ABLE
PlutellaSpodoptera Aedes Leptinotarsa % mort. % mort. % mort. ~ weight (~mg) at 100 ppmat 1,000 ppm at 100 ppm at 0.1%
HD-8 K4-3 75 100 0 +202 accorcing to inventian B.c. D1 0 20 0 +194 F12-l 70 100 0 +202 F4-3 65 60 0 +189 LI 2'i6-82 0 0 O - 6 F8-1 85 60 0 + 2 accorcing to inventicn Control 10 0 0 +251 (Untreated) HD-8 K4-3 = B.t. galleria + pBC 16 + pc 194 (pathotype A) B.c. D1_ = 8. cereus his strr F12-1 = B. cereus fused with HD8 cry gal (pathotype A) F4-3 = B. cereus fused with HD-2 cry thur (pathotype A) BI 256-82 = B.t. tenebrionis (pathotype C) ~- 3 ~ f~ 2 F8-1 = F32-13 cry ten fused with HD-2 D10 cry thur.
(pathotype A ~ C) Data on the nutrient solutions used Examples of formulations for protoplastization buffers are the following:
1) STM: O.S M sucrose, 30 mM tris, 5 mM MgClz . 6H20 pH 8.0 2) ';TML: STM with 0.5 mg/ml of lysozyme 3) ';MM: 0.5 M sucrose, 20 mM of Na2 maleate, 20 mM
of MgCl;7 pH 6.50 4) MM: 20 mM of Na2 maleate, 20 mM MgCl2; pH 6.5 5) MML: MM with 5 mg/ml of lysozyme 6) 'iMMP: Same volume of 2 x SSM and 4 x Antibiotic lS Medium 3 were combined after autoclaving.

7) 'iMMP M~L: SMMP with 1,000 U/ml of mutanolysine and 1 mg/ml of lysozyme An example of the composition of a medium for 20 regenerating protoplasts is the following:
Casein peptone 10 9 Yeast extract 5 9 Glucose 5 9 NaCl 2 9 Na citrate 2 9 MgCl2 ~ 6H2~ 0.5 9 CaCl2 . 2H20 Sucrase 340 9 pH 7.0 Gelatin 25 9 Starch 25 9 Agar 25 9 RC1: R with 1 llg/ml of chloramphenicol RC2: R with 2 ug/ml of chloramphenicol RTC1: R with 15 llg/ml of tetracyclin and 1 ug/~l of chloramphenicol L 3 ~

Defatted soybean meal 1C.0 g/l Potato starch 5.0 g/l Yeast autolysate 2.0 g/l 5 Anhydrous K2HP04 l.0 g/l MgS04 . 7H20 0.3 g/l CaCl2 . 6H20 0.08 g/l MnCl2 - 4H20 CuCl O.Oû5 g/l lO ZnCl;~ 0.005 g/l FeCl~ ~ 9 For use as an insecticide, the insecticidal pre-paration or toxin obtained according to the invention is mixed in a conventional manner with the usual additives (carriers, adhesives, wetting agents, etc.) and converted to a form suitable for use. The insecticide formulated in this manner can be used in the form of a wettable pow-der, a suspension, granules or the like.

Claims (5)

1. Bacillus thuringiensis DSM 4082 or a mutant thereof which synthesizes a protein pathogenic to insects.

2. Bacillus thuringiensis DSM 4082.

3. A process for obtaining microorganisms of the Bacillus thuringiensis type, which are genetically fixed and have the ability to form endotoxins, by protoplast fusion of two or more strains having different toxins, wherein the cell wall is degraded in a first step, fusion is induced in a second step and finally regeneration of the cell wall is induced.

4. The use of Bacillus thuringiensis 4082 or a mutant thereof obtainable by the process as claimed in claim 3 or of a toxin obtainable therefrom for controlling insects.

5. An insecticidal composition comprising as active compound Bacillus thuringiensis DSM 4082 or a mutant thereof obtainable by the process as claimed in claim 3 or of a toxin obtainable therefrom, in admixture with a suitable carrier.