8/14/2019 EP0431829A1

1/29

JEuropaisches PatentamtEuropean Patent OfficeOffice europeen des brevets 0 4 3 1 8 2 9

A 100 Publication number:

EUROPEAN PATENT A P P L I C A T I O N

int.cl.5:C12N 15/82, C12N 15/12,A01H 5/00, A01N 6 3 /0 2 Application number:

90312944.3 Date of filing: 28.11.90

Applicant: AGRACETUS, INC.8520 University GreenMiddleton, Wisconsin 53562(US) Inventor: Barton, Kenneth A.1718 Aurora Street

Middleton, Wl 53562(US)Inventor: Miller, Michael J.1401 Main StreetCross Plains, Wl 53528(US)

Priority: 29.11.89 US 443425 Date of publication of application:12.06.91 Bulletin 91/24 Designated Contracting States:AT BE CH DE DK ES FR GB GR IT LI LU NL SE

Representative: Bentham, Stephen et alJ.A. Kemp & Co. 14 South Square Gray's InnLondon, WC1R5LX(GB)

Insecticidal toxins in plants.

Transgenic plants have been created which ex-press an insect-specific toxin from a scorpion. Thechimeric inheritable trait produced conditions of tox-icity in the plant cells of toxicity to certain insectsupon ingestion of plant tissues. The inheritable traithas also been cross-bred to plants transgenic to theBacillus thuringiensis delta-endotoxin to produceplants having two independent insect-specific toxintraits. Insect feeding trials revealed additive toxiceffects. A generalized approach for developing otherinsecticidal toxins as candidates for insertion intotransgenic plants is also presented.

CM00

LLJ

Xerox Copy Centre

8/14/2019 EP0431829A1

2/29

EP 0 431 829 A1

INSECTICIDAL TOXINS IN PLANTS

The present invention relates to the generalfield of the genetic engineering of plants and re-lates, in particular, to the genetic engineering ofplants to confer to plants a trait of toxicity tospecific insect predators. 5The technology of recombinant DNA manipula-tion has evolved to a point which now permits thegenetic engineering of some higher organisms. Inthe area of plant technology, it has become possi-ble to genetically engineer many crop plant spe- wcies, including several commercially importantcrops. Now that it is possible to transfer exogenousgenetic traits into plants, a logical area of investiga-tion is to identify traits which can be added to theplants which will increase their agricultural value. 75Since the biochemical mechanisms responsible forplant vigor and growth are poorly characterized andunderstood, and since the technology of the ge-netic transfer of exogenous traits into plants per-mits only a relatively few genes to be transferred sointo plants at any one time, it would appear difficultto envision, given the present state of the art, howdramatic changes in inherent plant growth or vigorcharacteristics can be obtained through geneticengineering. However, since actual crop yields in 25normal field situations are often dictated more bypredatory factors on the plants, such as disease orinsect predation, it is possible to increase the ac-tual yield of crop plants under field conditions notby making the plants grow better, but by adversely 30effecting the limiting predatory factors which wouldotherwise decrease the effective yield from theplants.The present techniques for controlling insectpredation of crop plants are based on use of syn- 35thetic chemical insecticides, which have been anaccepted component of the cultivation practices formajor agricultural crops, at least in the developedcountries, for several decades. Most of the majoragricultural chemicals utilized for insecticidal pur- 40poses are toxic to a relatively broad spectrum ofinsect pests and many persist in the environmentgiving rise to adverse environmental conse-quences. There have therefore been many effortsto develop pesticidal compounds that are uniquely 45toxic to specific insects and which also do notpersist in the environment.Biological insecticidal agents can meet severalof these criteria. For example, there have beenseveral products based on the use of various forms 50of the delta-endotoxin produced by the soil dwell-ing microorganism Bacillus thuringiensis (B.t.) asinsecticidal agents. This polypeptide toxin hasbeen found to be specifically toxic to Lepidopteraninsects, and has been used for many years com-

mercially as a foliar applied insecticide. It has alsorecently been found that various forms of the B.t.toxin can be toxic to insects, when expressed li>side the tissues of plants on which the insectsfeed. This is the first, and so far only, example of anatural biological insecticide being expressed inthe cells of transgenic plants.It has also been a feature of the history of theuse of insecticide in agricultural applications thatthe insects sometimes become resistant to theapplied pesticides. This phenomenon occursthrough natural selection of the most insecticidalresistant members of the insect population follow-ing continual application of a single insecticidalagent year after year. While no broad spectrumresistance in insects has yet been shown to havedeveloped to a biological toxin, such as the B.t.delta-endotoxin, the possibility of the developmentof such a resistant population must be consideredin the long-term planning for the development ofinsect resistant plants. One way to minimize thepossibility of the development of any such insecti-cidal resistant populations is to impose the insectpredators to a regimen of at least two toxins, eachof which has an independent mode of activity. Theuse of two agents, either simultaneously or sequen-tially imposed on the target populations, dramati-cally decreases the statistical possibility that resis-tant insect populations could develop. To be usefulas a biological agent to be expressed in plants,such toxins should preferably be polypeptides thatcan be introduced into plant cells through singlegene traits. One place to look for a source for sucha toxin is animals which are insect predators.One class of organism which is able to predateon insects is the scorpion. The scorpion is anArthropod predator which has relatively poorly de-veloped senses, but which has developed the abil-ity to create a venom which rapidly immobilizes itsinsect prey, which is otherwise more mobile thanthe scorpion. The venom of the scorpion serves adual function, as a defense against possible scor-pion predators and to procure its own prey. Thevenom is therefore composed of a complex cock-tail of toxins, many of which are pharmacologyactive proteins. The activity of the individual pro-teins from the scorpion toxin has been found to berelatively selective toward specific classes of or-ganisms, such as mammals, insects, or crusta-ceans.Several insect-selective neurotoxins have beenisolated from scorpions and have been character-ized and sequenced. One such neurotoxin, thepeptide AalT which has been isolated from theNorth African scorpion Androctonus australis Hec-

8/14/2019 EP0431829A1

3/29

EP 0 431 829 A1

coding region for the AalT toxin and theoligonucleotides used to construct it.Fig. 3 is a sequence listing of the entire scor-pion toxin expression cassette from pAMVSTI .5 Fig. 4 is a schematic illustration of the con-struction of the plasmid pTV4AMVST1 .Fig. 5 is a sequence listing of the syntheticcoding region for the BelTI toxin and theoligonucleotides used to construct it.

w Fig. 6 is a schematic illustration of the con-struction of pTV4AMVBelT1 .Fig. 7 is a sequence listing of the syntheticcoding region for the BelT2 toxin and theoligonucleotides used to construct it.75 Fig. 8 is a schematic illustration of the con-struction of pTV4AMVBelT2.

Fig. 9 is a graphical representation of insectfeeding trials conducted with transgenic plants.Fig. 10 is a sequence listing of the synthetic20 coding region for the AgalV toxin and theoligonucleotides to construct it.Fig. 11 is a schematic illustration of the con-struction of pTV4AMVAgalV.

Fig. 12 is a sequence listing of the synthetic25 coding region for the SfIT toxin and theoligonucleotides to construct it.Fig. 13 is a schematic illustration of the con-struction of pTV4AMVSflT.In accordance with the present invention, it has30 been discovered that a gene for one of the con-stituents in the venom of an Arthropod insect pred-ator can be synthesized and then genetically en-

gineered into plants to create plants which will haveunique and specific toxicity when ingested by cer-35 tain insect pests. It has been discovered that aprotein toxin natively produced by an insect preda-tory animals such as scorpions and spiders can besuccessfully expressed in the cells of plants, with-out damage or visible morphological change to the40 plant, even while the plant is imbued with a toxicitywhen ingested by certain insects. Based on thesefindings, not only are specific genes disclosedherein which have utility for the introduction intovarious plant species to increase the resistance of45 the plants to predation by insect pests, but a newclass of possible insecticidal agents is disclosedwhich can be genetically engineered into plants toincrease their resistance to predation by insects in

general.so One particular toxin which has been discoveredhere and is described in further detail below is a

polypeptide toxin which was discovered as a con-stituent of the venom produced by a North Africanscorpion Androctonus australis. This toxin has been55 one which was found to be toxic to insects in bothin vitro and in vivo tests. Other toxins illustratedherein are two" toxins isolated from a central Asianscorpion Buthus epeus, and two toxins isolated

tor, is a highly charged poiypeptide consisting ofseventy amino acids. It has been reported, basedon in vitro studies, that the specificity of AalTpeptlde is toward a synaptosomal membrane vesi-cle of insects, and that the protein shows no affinitytoward other insect tissues, or for any mammalian,arachnid or crustacean tissues. The AalT polypep-tide binds specifically, reversibly and with highaffinity to a single class of non-interactive bindingsites on insect neural membranes.One difficulty incumbent in the genetic en-gineering of plants to express toxic compounds isthat the toxicity of such compounds to plants cellscan adversely effect the ability to recover trans-formed plants. It appears, for example, that plantcells imbued with the trait to produce the full-lengthamino acid sequence of the B.t. delta endotoxin arenegatively selected during present transformationtechniques, perhaps due to toxicity of the full-length toxin on the plant cells themselves. It istherefore the case, at least at present, that the invivo toxicity to plant cells of peptides toxic toinsects is not predictable. It is also not possible topredict the toxicity of insect toxins which are nor-mally injected into the insect when such toxins areadministered to insects by ingestion.The present invention is summarized in thattransgenic plants have been created which effec-tively express in their cells an insect specific toxinof an insect predator in an amount sufficient so asto cause toxicity to selective insects which ingesttissues of the plant.It is another object of the present invention toprovide a method for finding novel strategies forimbuing plants with insect resistance which com-prises screening polypeptides produced by insectpredators for their effectiveness once inserted intoplant cells in imbuing the plants with insect resis-tant properties.It is another object of the present inventionthus to provide genetically engineered plants whichare toxic to insect predators thus lessening theneed for artificial agricultural chemicals to protectfield crops.It is yet another object of the present inventionto provide alternative insect-selective peptide toxinswhich can be introduced into plants, in conjunctionwith B.t. toxins, so as to minimize the possibilitiesof developing toxin resistant insect populations.Other objects, advantages, and features of thepresent invention will become apparent from thefollowing specification when taken in conjunctionwith the accompanying drawings.

Fig. 1 is a chart of codon usage developed toillustrate codon usage in plant cells and used inconstruction of the oligonucleotides in the exam-ples of the present invention.

Fig. 2 is a sequence listing of the synthetic

8/14/2019 EP0431829A1

4/29

EP 0 431 829 A1

from two species of spider. Therefore, from themethodology of the selection and testing fromthese toxins, and by the logic by which thesetoxins were isolated and tested, it should be appar-ent that other similar toxins can be found by similar 5screening and testing. In particular, the phenom-enon of insect toxicity uncovered and found to beeffective with the toxin disclosed herein was devel-oped by a rational plan to respond to a perceivedneed to find other traits which could be introduced winto transgenic plants to provide protection againstinsect predation. It has previously been demon-strated, notably with the toxins produced by thesoil dwelling microorganism Bacillus thuringiensis,that insecticidal toxins can successfully be pro- 75duced in plant cells so as to render those plantstoxic to insects which ingest them. To find othersuitable insecticidal toxins, the inquiry was directedtoward predators which rely on insecticidal chemi-cals likely to have unique toxicity to insects. A wide 20variety of predators do utilize toxins in preyingupon insects, most notably Arthropods such asspiders, scorpions, centipedes and predatory in-sects such as wasps. The scorpion in particularseemed a preferable candidate from which insecti- 25cides could be uncovered. In nature the scorpion isa relatively slow animal with relatively poorly devel-oped visual sense and a poor sense of smell. Thescorpion therefore requires a swift acting and po-tent venom to serve the function of rapidly immo- 30bilizing the prey of the scorpion. In addition, signifi-cant effort had been expended by others in theisolation, characterization and sequencing of scor-pion toxins, which aided in the selection of a par-ticular protein the toxicity of which could be 35screened. Accordingly, the scorpion toxin repre-sented a class of agents which could be investi-gated for possible insecticidal toxin activity. Othertarget predators of insects include any animalswhich rely on a peptide toxin to either incapacitate 40or kill their insect prey.It must be remembered that to be a candidatefor genetic insertion into plants, a toxin shouldideally meet several constraints. One constraint, atleast at present given the level of skill in the art of 45genetically engineering plants, is that the toxinshould preferably be a peptide which can be syn-thesized by a single gene trait which can be in-serted into plant cells. As the technology of plantgenetic engineering improves, and the ability to 50insert larger number of genes becomes developed,it may also be possible to insert genes coding forenzymes which catalyze the synthesis of non-pep-tide toxins. Another constraint is that the toxinshould be selected so as to be relatively specific in 55its activity. Many toxins are active broadly againstmost animals. Such toxins may not be preferredcandidates for genetic engineering into plants to be

used for human or animal food. However, the de-veloping capacity to construct chimeric genes toexpress peptides in plants in a tissue-specific man-ner raises the possibility of using broader spectrumtoxins, since it is becoming possible to produceplants which will produce the toxins only in planttissues that will not be used as animal or humanfood. Nevertheless, optimal toxin candidate wouldbe a toxin which is uniquely toxic to insects, butwhich is minimally or not at all toxic to any otherclasses of animals which might feed on the plants.In particular, it is desired that there be no toxicityto mammals, so that the insertion of the toxin intothe cells of plants still results in plants which haveunchanged nutritive value to humans or to domes-tic animals.Suitable toxins can then be screened both forin vitro and in vivo activity both to determine effec-tiveness of the toxin and to determine the selectiv-ity of the toxic effect. Such in vitro screening canbe done using toxins purified from the animals inquestion, or using toxins synthesized by gene ex-pression systems used in procaryotic organisms toproduce significant quantities of the toxin, or bytransgenic introduction into plant cells which canbe used in feeding trials on the target insects.An example of a suitable toxin, which wasisolated from the North African scorpion Androc-tonus australis hector, has been designated AalT.The insect selective neurotoxin, AalT, is a polypep-tide consisting of 70 amino acids. It has beenpreviously suggested that the 8 cystine residues inAalT contribute to 4 disulphide bridges of the ma-ture AalT protein which combine to give the proteina unique secondary structure which contributes to-ward the selectivity of the toxin toward insects,relative to other scorpion toxins which are activetoward vertebrate life forms. Darbon et al., Intern. J.Peptide Protein Res., 20 p. 320 (1982). Previouiwork by others has indicated that the AalT peptidehas specific affinity toward the synaptosomal mem-brane vesicles of insects and has little or no affinityfor non-innervated insect tissues, or for tissues ofother animal forms, such as mammals, arachnids,or crustaceans. It has been reported that the AalTpolypeptide binds specifically, reversibly, and witha high affinity to a single class of non-interactivebinding sites on the insect neural membrane. Oth-ers have previously demonstrated that the AalTpeptide has toxicity against various insects whenapplied topically, when injected, or when ingestedat high doses by certain insect larvae. Similar re-search has been reported establishing the toxicityand specificity of two toxins, designated BelT1 andBelT2 isolated from the Asian scorpion Buthusepeus.Once a toxin, such as AalT, has been selected,it is then necessary to prepare a chimeric expres-

8/14/2019 EP0431829A1

5/29

EP 0 431 829 A1

Once the chimeric plant expression cassette isconstructed, it is then necessary to transform theconstructed gene into plant cells. There exist atpresent several mechanisms for the insertion of5 foreign genes into plants. One method, such asutilized here, involves a plant pathogenic bacteriumknown as Agrobacterium tumefaciens. Wild-typeAgrobacterium tumefaciens harbors a plasmid, re-ferred to as the Ti plasmid, which has the native10 ability to transfer a portion of its DNA, referred toas T-DNA, into the genomic DNA of infected plantcells. By removing the genes which code for tu-morous activity normally contained within the T-DNA of an Agrobacterium, and by substituting for

75 the Ti plasmid with a foreign expression cassettesynthetically constructed, it is possible to use thenatural genetic engineering ability of Agrobacteriumtumefaciens to transform any desired foreign geneinto plant cells. Other techniques for genetic en-20 gineering of plants have also been developed,which obviate some of the disadvantages of Ag-robacterium transformation, and provide the theo-retical ability to transform many, if not all, plantspecies. Such techniques include the electropora-25 tion and regeneration of plant protoplasts, the di-rect injection of naked DNA into plant flowers, and,most notably, the newly developed technique ofparticle-mediated acceleration plant transformation.In accordance with the techniques of genetically30 engineering plants using particle-mediated planttransformation, naked DNA containing copies of a

plant expression cassette is coated onto extremelysmall carrier particles. The carrier particles canthen be physically accelerated by an explosive

35 force, such as achieved through an electric shockdischarge, into the living cells of plant tissue. Tis-sue types which can be readily transformedthrough particle-mediated transformation tech-niques include the growing meristem of plants,40 plant embryos, and plant germ cells such as pol-len. It is envisioned that any of the above planttransformation techniques may be used within thescope of the present invention. Based on the ex-45 perience to date in the genetic engineering ofplants, there appears to be little difference in theexpression of genes, once inserted into plant cells,attributable to the method of plant transformationitself. Since some transformation techniques can

50 result in more than one copy being inserted, andsince the site of foreign DNA insertion appears tobe random using any technique, the activity of theforeign gene inserted into plant cells seems moredependent on the expression characteristics of the55 individual inserted genes resulting both from thechimeric control regions (promoters, polyadenyla-tion regions, enhancers, etc.) and from the influ-

ence of indigenous plant DNA adjacent the

sion cassette suitable for expressing the peptide inthe cells of target plants. There are a number ofways in which such a chimeric expression cassettecan be constructed, as is known to those of or-dinary skill in the art. At a minimum, the construc-tion of such an expression cassette requires apromoter which initiates transcription of a down-stream coding region in the cells of plants. Down-stream of the coding region in the plant it is re-quired that there be a translation or transcriptionterminator, such as a poiyadenylation sequence,several examples of which are known to be usefulin the genetic engineering of plants. Between thepromoter and the coding region, it is also possibleto use a 5' non-coding translational enhancer se-quence. One such preferred sequence which hasbeen effectively utilized is the 5' non-coding se-quence from the alfalfa mosaic virus coat proteingene.

The coding sequence itself can be created in anumber of ways. One strategy for the creation ofsuch a coding sequence would be to isolate themRNA of the toxin gene, as expressed in the cellsof the insect predator, and then to create a cDNAcoding region corresponding to the coding se-quence in the organism which is the native host tothe toxin. However, since it has also been foundthat certain coding sequences work better thanothers in the expression of proteins in plant cells, apreferred strategy would be to discover the aminoacid sequence of the peptide toxin, and then tosynthesize a coding region customized so as tomaximize the expression of the toxin protein inplant cells. If the amino acid sequence is of areasonable size, it is well within the scope of rea-sonable skill in the art at the present time tosynthesize such coding regions, as has been donehere. In order to optimize the translational effi-ciency of such systems, it is possible to analyzethe pattern of codon usage by plant genes whichare normally expressed at abundant levels. By ex-amining the frequency of codon usage by effi-ciently or very actively expressing natural plantgenes, it is possible to determine which codonsare, in essence, preferred within the translationalsystems normally present in plant cells. It is there-fore possible to construct a synthetic coding regionfor any peptide which includes, as the codon foreach amino acid in the peptide, the codon which ismost preferred by plant cells, at least in the nativegenes most actively expressed in those plant cells.Using such preferred usage codons, it is possibleto construct a protein coding sequence which willresult in a significantly enhanced level of tran-slational efficiency compared to what would beachieved by taking the coding sequence directlyfrom the organism which natively produces thetoxin peptide.

8/14/2019 EP0431829A1

6/29

EP 0 431 829 A1 10

chimeric insert, and by copy number, rather thanby the technique of inserting the genes into plants.The process of creating insect resistant plants inaccordance with this method cannot be consideredcomplete once a coding sequence for the foreign 5gene has been inserted into plant cells and theplants regenerated into whole plants. It is thennecessary to test the transgenic plants thus cre-ated through appropriate biological assays by in-sect feeding. Only by empirically testing the toxic- wity of such plants to insects themselves can it betruly determined if the desired toxicity has beenobtained.Example 1 isToxin Expression Cassette

The synthesis of the expression cassette forthe expression of the scorpion toxin AalT in plant 20cells was commenced with the alteration of a pre-viously constructed plant expression cassette con-structed for the expression of the B.t. delta-en-dotoxin in plant cells. This plasmid, which hadpreviously been designated pAMVBTS, included a 25promoter sequence derived from the cauliflowermosaic virus 35S gene (CaMV 35S), followed by a5' non-coding region which had been syntheticallyconstructed to correspond to the 5' non-codingregion from the alfalfa mosaic virus coat protein, 30which had previously been found to enhance theefficient expression of synthetic genes in plant tis-sues. Following the 5' non-coding translational en-hancer was a shortened toxin coding sequence forthe B.t. delta-endotoxin, followed by convenient 35cleavage sites for restriction enzymes including aPst I site, followed by a polyadenylation regionderived from the nopaline synthase gene derivedfrom the Ti plasmid of Agrobacterium tumefaciens.Also on the plasmid pAMVBTS, reading in the 40opposite direction, was a selectible marker for am-picillin resistance constructed to be expressive inthe cells of procaryotic hosts. The plasmidpAMVBTS has previously been deposited with theAmerican Type Culture Collection, Rockville, Mary- 45land, as Accession No. 53637 deposited June 24,1987. Details of this plasmid may be found inBarton et al., Plant Physiol., 85, pp. 1103-1109(1987) and published PCT patent application WO89/04868 published June 1, 1989, the disclosure of 50both documents being hereby incorporated hereinby reference.The seventy amino acid protein sequence forthe AalT insect toxin polypeptides had previouslybeen reported by Darbon et al., Int. J. Pep. Res., 5520, pp. 320-330 (1982). Since the" intent was toconstruct an expression cassette for optimal effi-ciency in expressing the protein in plant cells, as

opposed to arthropod cells, the coding region ofthe toxin was not cloned from its native host, but itwas instead synthesized so as to enhance thetranslational effectiveness of the protein coding re-gion in plants. An artificial toxin coding DNA se-quence was therefore first derived based on thepublished amino acid sequence. To develop such asynthetic coding sequence, reference was had to atable, previously developed, listing the frequency ofexpression of the codons utilized by native plantgenes in the expressions of efficiently producednative plant proteins. Fig. 1 herewith contains theamino acid codon usage frequency table which hadbeen ascertained by the investigators here basedon publicly available sequence data on sequencedplant genes which had been determined to beefficiently and highly expressed in plant cells. Thusfor each amino acid position in the desired peptide,a codon was selected representing the most popu-lar codon used in native plant cells to express thatparticular amino acid in native plant genes. Thepreferred codons for each amino acid are indicatedby outlines in Fig. 1. Based on this rule of pre-ferred codon usage, a 210 base pair syntheticnucleotide sequence to code for the same 70 ami-no acid polypeptide produced in the scorpion wasderived. In order to synthesize a DNA sequence ofthat length, six overlapping and homologous syn-thetic oligonucleotides were created correspondingto the overall sequence which was desired. Theoverall protein coding sequence with the sixoligonucleotides, designated MM62 through MM67,is shown in Fig. 2. The total expression cassette forthe AalT toxin in plant cells, derived from theplasmid pAMVBTS by substituting the AalT peptidecoding region for the B.t. region, with the AalTtoxin coding region extending from nucleotides 482through 694, is shown in Fig. 3, and is designatedpAMVSTI .To construct the expression cassette forpAMVSTI, the six overlapping syntheticoligonucleotides were initially annealed in pairs oftwo to form a double-stranded molecule. The initialcenter pair (MM63 and MM66) was then kinased toadd phosphate groups to each 5' end. The threedouble stranded oligonucleotides were ligated andthen inserted into a plasmid pUC18 which hadbeen previously digested Hind III and PST I. Theresulting plasmid, designated pUCSTI, wassequenced to ensure that the insert was the de-sired sequence as illustrated in Fig. 2.The synthetic DNA sequence was designed toincorporate a 5' restriction enzyme site for theenzyme BspMI immediately adjacent to the initiat-ing methionine codon (ATG) of the insect toxincoding region as noted in Fig. 2. The plasmidpUCSTI was first digested with Pst I to release the3' end of the insect toxin coding region, and also to

8/14/2019 EP0431829A1

7/29

EP 0 431 829 A1 121

ton et al., supra. Seeds of tobacco (Nicotianatabacum variety Havana 425) were surface steril-ized and germinated on MS medium so that ascep-tically grown leaves and stems could be used for5 plant transformation procedures. The ascepticallygrown tobacco tissues were inoculated with Ag-robacterium containing the transformation vectors,and the plants were regenerated following selectionfor kanamycin-resistant transformants in a manner10 well known to those skilled in the art. The resultingtransformant plants, designated as RO plants, wereanalyzed for appropriate toxin DNA, RNA, and pro-tein components within the cells of those plants asdescribed below.

75 Toxicity StudiesThe first toxicity studies conducted with the

transgenic plants were conducted with tobacco hor-20 nworm (Manduca sexta), which had been pur-chased as eggs from Carolina Biological Supply.Initial feedings of tobacco homworms on RO plantsderived from the regenerates failed to demonstrate

any ascertainable toxicity to the insects.25 In order to verify that the insensitivity of thefeeding insects was due to ineffectiveness of thetoxin rather than a failure of the transformationtechniques, RNA screens were then conducted onapproximately 40 RO plants which had been cre-30 ated. This screen was done by slot-blot hybridiza-tions conducted on RNA from each of the transfor-mants. Northern gels were run to determine theintegrity of the hybridizing RNA in the transfor-mants. The procedures used for such slot-blots and35 Northern gels are discussed in Barton et al., supra.Several of the RO plants which had the highestlevels of hybridizing RNA were then self-pollinatedto provide seed for R1 seedling analysis. Oneparticular plant was initially identified to have a

40 very active transcription of the gene encoding forthe AalT protein. This plant, designated plantT2636, was determined to have approximately0.001% of its total mRNA homologous to the AalTprobes, while other regenerates expressed at var-45 ious lower levels. Such variation and activity ofgene expression from heterologous genes insertedinto plants is quite common following Agrobac-terium transformation, or other known techniquesfor inserting genes into plants. The most highlyso expressing plants were then assayed further withthe principal focus being on plant 2636 and theprogeny created from its self-pollination. The tran-script analysis of the RNA from these plantsshowed hybridizing mRNA of the expected size of55 530 nucleotides including the poly-A tail. Measure-ment of relative levels of the transcript from theAalT and the APH-II (Kanamycin resistance) codingregions showed a consistent 10-fold difference in

destroy the BspMI site located within the polylinkerderived from pUC18. The linear DNA fragment thuscreated was then digested with BspMI to releasethe 5' end of the coding region with a "CATG"sticky end, which would be compatible in ligationwith an Nco I restriction site, a site which wasconveniently available in the plasmid pAMVBTS.The plasmid pAMVBTS was then digested withNco I and Pst I, and the expression cassette por-tion of the vector isolated from the BTS codingregion. The synthetic scorpion toxin coding regionwas then ligated into the vector. This process isillustrated in Fig. 4. The resulting plasmid, des-ignated pAMVSTI, incorporates the plant expres-sion cassette in its entirety derived from the plas-mid pAMVBTS. The expression cassette thus con-sists of a CaMV 35S transcriptional promoter, a 5'non-coding translational enhancer having a se-quence similar to that of the alfalfa mosaic viruscoat protein mRNA non-coding region, a syntheticDNA fragment encoding the protein sequence forthe AalT insect toxin, and the polyadenylation re-gion for nopaline synthase. A selectible marker ofampicillin resistance was also included in the plas-mid pAMVSTI for selection in bacteria.To enable the transformation of plant cells withAgrobacterium, the expression cassette pAMVAalTwas cointegrated into the Agrobacteriumtumefaciens binary vector pTV4 as described inBarton et al., and WO 89/04868, both incorporatedby reference above, with reference to the plasmidpAMVBTSH. The deposited plasmidpTV4AMVBTSH, ATCC accession No. 53636 is aconvenient source of plasmid pTV4 by digestingthe cointegrate plasmid pTV4AMVBTSH with Xho I,by ligating the digestion products to reclose thetwo resultant plasmids pTV4 and pAMVBTS, andby transformation into E. coli with selection forsulfadiazene resistance. A~mTrTiprep will confirm theisolation of pTV4. The same procedure, with selec-tion for ampicillin resistance, can be used to derivepAMVBTS for use as an expression vector inplants.To cointegrate pAMVSTI into pTV4, both plas-mids ware digested with Xho I, the linearized DNAswere combined and religated. The product wastransformed into E. coli and selected for both sul-fadiazene and ampicillin resistance, and then con-firmed by miniprep analysis.In order to do in vitro toxicity studies using thetoxin, supplies of AalT toxin were obtained fromSigma.Plant Transformations

The plasmid pAMVSTI was cointegrated intothe binary vector pTV4, and conjugated into anAgrobacterium strain EHA101 as described by Bar-

8/14/2019 EP0431829A1

8/29

EP 0 431 829 A13 14

concentration of AalT RNA compared to the APH-IIRNA, which might be due to the difference inpromoter regions (the stronger CAMV 35S pro-moter on the AalT gene versus the nopaline syn-thase promoter on the APH-II gene) but this is notcertain.Continued feeding trials with tobacco horn-worms on tissues of R1 and R2 plants resulted inno mortality to the insects and insignificant reduc-tion in feeding, relative to controls. However, insecttoxicity assays were also conducted with othercandidate insects and the results are illustrated inFig. 9. In particular, insect toxicity studies wereconducted with the cotton bollworm (Heliothis zea)and the beet armyworm (Spodoptera exigua).When supplied tissue from the R1 and R2 progenyof plant T2636, both bollworms and armywormswere each subject to very high mortality and dra-matically reduced feeding behaviors. The samefindings were found on other plants which were the

progeny of original RO plants which had previouslybeen determined to express lower levels of theAalT transcripts than the plant T2636.To verify the effect of the AalT peptide ontarget insects, separately from the peptide pro-duced in the plant cells, the commercially availablepreparation of crude venom of Androctonusaustralis (Sigma Chemical) was introduced boththrough topical and ingested applications. For topi-cal use, the AalT peptide was dissolved in 50%DMSO at a concentration of approximately 5 milli-grams per milliliter, and approximately 1 microliterof this solution was directly applied to the skin ofthe target insect. Paralysis, behavior modifications,and mortality relative to the controlled treatmentswere monitored for up to 24 hours following ap-plication. It was discovered that the toxin is active iftopically applied to homworms, even though itseemed relatively ineffective in an ingested form inplant tissues. Reduced activity of treated hornwormlarvae, as well as occasional spastic responses,were observed for periods of time after topicalapplication, generally followed by complete recov-ery. Injection of toxin resulted in a prolonged pe-riod of reduced activity, but not in mortality. Otherorganisms, such as the cotton bollworm, were sen-sitive to the toxin whether the root of introductionwas topical application, injection, or ingestion.Fig. 9 summarizes the feeding trial of the var-ious insect larvae on the progeny of tobacco plantT3376, an R1 progeny of the primary RO regener-ate plant T2636. To generate the plants used forthe tests summarized in Fig. 9, the plant T3376was propagated in three ways, by self-pollination,by out crossing as either a pollen donor or amaternal pollen recipient to plant T3219, an R2plant which had within it a homozygous insertion ofthe single BTS gene expressing the Bacillus thurin-

giensis delta-endotoxin in insect toxic doses, and inaddition plant T3219 was self-pollinated. Prior tofeeding trials all of the progeny of these crosseswere analyzed by polymerase chain reaction (PCR)5 to confirm the presence of the appropriate AalT orB.t. sequences. Self progeny of the plant T3219were expected to be homozygous B.t., while out-crosses of T3219 and T3376 have been isolatedwhich contained a single copy of both of the rel-10 evant genes. Progeny of plant T3376, which did notcontain an AalT gene, were excluded from thestudy.In the graphs of Fig. 9, "St" represents larvaefed tissues from the self-pollinated progeny of plant75 T2636, "H425" represents larvae fed wild-type(nontransgenic) Havana 425 tobacco tissues as acontrol, "Bt" represents larvae fed tissues of plantsapparently heterozygous with a single copy of B.t.gene (resulting from crossing of T3219 with20 T3376), "B.t. + B.t." represents larvae fed tissuesfrom apparently homozygous B.t. gene-containingtobacco plants (from self-pollinated T3219), while"B.t. + S.t." represents larvae fed tissues of plantsderived from crosses between a B.t. gene contain-25 ing parent (T3219) and a parent containing an AalTgene (T3376). The three graphs show survival ofManduca sexta, Heliothis and Spodoptera over 0-8days feeding on such tissues. It can be seen thatthe AalT gene had no obvious effect on Manduca30 in this experiment, but AalT caused significant mor-tality relative to controls to both Heliothis andSpodoptera. Furthermore, the presence of both B.t.and AalT in a plant provides at least as muchtoxicity to insects as either toxin individually, and a35 possible additive effect is indicated. Thus, the pres-ence of both toxins would be expected to serve asa deterent to the development of resistance toeither individual toxin.The studies revealed that the chimeric genetic40 construction is stable through at least the R2 gen-eration and that the toxicity to susceptible insectscontinues and is an inheritable trait which can bepassed to progeny plants by Mendelian inheri-tance. This observation is consistent with previous

45 experiments in the insertion of transgenic genesinto the germ line of plants.Example 2

so An insecticidal polypeptide possessing neuro-toxic activity toward insects (Insectotoxin 11, orBelT1 ) was isolated from the venom of the CentralAsian scorpion Buthus epeus Zhdanova et al.,Bioorganicheskaya Khimiya, 3 pp. 485-493 (1977),55 Grishin, Int. J. Quant. Chem., 19, pp. 291-298(1981). Following purification of the peptide andtests for bioactivity, amino acid sequence was ob-tained and shown to be (in single letter amino acid

8/14/2019 EP0431829A1

9/29

EP 0 431 829 A1 165

complete the Nco I recognition sequenceCCATGG). BspMI was therefore used on the BelTfragment to generate the compatible "CATG"sticky end, which will ligate to Nco I generated5 sites, although the Nco I recognition site will bedestroyed by the ligation. At the carboxy-terminusof the coding region, immediately adjacent to theBam HI site, is a recognition sequence for Pst I(CTGCAG), the enzyme used to excise the codingw sequence from pUC18 for ligation of the carboxy-terminus into the expression plasmid for chimericgene construction.As illustrated in Fig. 6, to construct a chimericgene capable of expression in plants, the plasmid75 pAMVBTS was digested to completion with en-zymes Nco I and Pst I, and the vector was purifiedfrom agarose gels following electrophoresis. TheBspMI and Pst I digested oligonucleotide encodingBelT1 was purified from pUC18 in a similar man-

20 ner, the two DNA fragments were combined andligated. E. coli was transformed and selected forampicillin resistance, and the properly constructedplasmid pAMVBelTI was identified by minipreps.This plasmid is essentially identical to pAMVSTI,25 except for the different amino acid coding regionsspecifying the two different insect toxins. From 5'to 3', the expression cassette consists of the CaMV35S promoter, the mRNA 5' noncoding translationalenhancer region corresponding to that of alfalfa30 mosaic virus coat protein mRNA, the synthetic cod-ing sequence for BelT1, and the poiyadenylationregion from nopaline synthase. In order to move

pAMVBelTI into plant cells, the plasmid was di-gested at a unique Xho I endonuclease site imme-35 diately 5' to the CaMV 35S promoter, and theplasmid was cointegrated by ligation with Xho Idigested Agrobacterium vector plasmid pTV4. Fol-lowing transformation of E. coli, and selection forboth sulfadiazene and ampicillin resistance, the40 properly cointegrated plasmids to pTV4AMVBelT1were identified by DNA minipreps. These plasmidswere then conjugated into Agrobacterium and usedin transformation of plant tissues as described pre-viously in Example 1 and Barton et al. (Barton,45 Whiteley and Yang).Following regeneration of plants transformedwith AMVBelTI, approximately 50 plants will bescreened using mRNA slot-blots to identify thoseplants that express the BelT1 gene most strongly.50 The four plants with the most BelT1 mRNA will beallowed to flower and set seed, and the progenywill be analyzed in insect feeding trials. As withAalT, toxicity is expected to be present againstsome species of insects. In addition to self-pollina-55 tion of plants containing BelT1 , the four plants withhighest expression will be outcrossed to plantspreviously identified to express significant levels ofeither AalT (Example 1), B.t. delta-endotoxin, or

code):MCMPCFTTRPDMAQQCRACCKGRGKCFGPQCL-CGYDAs in Example 1 demonstrating synthesis of agene encoding the AalT peptide, a chimeric genewas constructed to enable expression of BelT1 inplants including a synthetic coding region. Basedon the most frequently used codons in plants (Fig.1), four oligonucleotides designated MM83 throughMM86 were synthesized to span the amino acidcoding region of the peptide. Illustrated in Figure 5is both the synthetic coding region and theoligonucleotides.Oligonucleotides MM83 and MM85 are complimen-tary, and represent the amino-terminal portion ofthe peptide coding sequence. OligonucleotidesMM84 and MM86 are complimentary, representingthe carboxyl-terminal portion of the coding region.To construct the functional gene for expression inplants, the four synthetic oligonucleotides were ini-tially combined in complimentary pairs and an-nealed, then the two sets of pairs were combinedand the overlaps annealed. The oligonucleotideswere then treated with polynucleotide kinase in thepresence of ATP to provide 5'-monophosphates atthe hydroxyl termini of the oligonucleotides to fa-cilitate ligation. The annealed oligonucleotides werethen combined with pUC18 plasmid vector pre-viously digested with Hind III and Bam HI. DNAligase was then provided to ligate the pairs to-gether, and to join the synthetic BelT1 codingregion to the vector pUC18. E. coli was trans-formed to ampicillin resistance, and appropriateplasmids were identified by DNA minipreps. Theinsert DNA was sequenced to confirm the se-quence as shown in Figure 5.As may be seen by reference to Fig. 5, the 5'end of the duplex oligonucleotide has a fournucleotide sticky end (overlap) compatible with thatgenerated by the endonuclease Hind III, while the3' end has a sticky end compatible with that gen-erated by endonuclease Bam HI. Immediately adja-cent to the Hind III site in the synthetic insert is arecognition site for cleavage with BspMI, an en-donuclease that cleaves distally from the hex-anucleotide recognition site, which in thisoligonucleotide will result in a 4-nucleotide stickyend compatible with that of endonuclease Nco I.Nco I is the enzyme recognition site that will beused to join the BelT1 coding sequence to regula-tory DNA sequence from pAMVBTS for construc-tion of the functional chimeric gene, where thecentral "ATG" in the Nco I recognition "CCATGG"will represent the first codon of the BelT1 codingsequence. Because the second codon of BelT1 iscystine (Cys, TGC) it is not possible to provide anNco I cleavage site, since that would require thatthe first nucleotide of the codon be a "G" (to

8/14/2019 EP0431829A1

10/29

EP 0 431 829 A17 18

BelT2 (Example 3). Progeny of these outcrosseswill be subjected to insect toxicity bioassays, and itis expected that each toxin activity will be func-tional in the progeny and further that additive ef-fects will be apparent between some of the toxins,resulting in additional levels of insect resistance. Itis expected that the development of resistance tothese toxins by susceptible insects is significantlydelayed by the presence of two different toxinswithin the same plant.Example 3

An insecticidal polypeptide (Insectotoxin 12, orBelT2) possessing neurotoxic paralytic activity to-ward insects has been isolated from the venom ofthe Central Asian scorpion Buthus epeus Grishin etal., Bioorganicheskaya Khimiya, 5, pp. 1285-1294(1979); Grishin, Int. J. Quant. Chem., 19, pp. 291-298 (1981). Following purification of the peptideand tests for bioactivity, amino acid sequence wasobtained and shown to be (in single letter aminoacid code):MADGYVKGKSGCKISCFLDNDLCNADCKYYGGKL-NSWCIPDKSGYCWCPNKGWNSIKSETNTCAs in Examples 1 and 2 demonstrating synthe-sis of genes encoding the AalT and BelT1 pep-tides, respectively, a chimeric toxin gene was con-structed to enable expression of BelT2 in plants.Based on the most frequently used codons inplants (Figure 1), six oligonucleotides were syn-thesized to span the amino acid coding region ofthe peptide. These oligonucleotides, as they annealto form the complete coding sequence, are shownin Figure 7. Oligonucleotides MM74 and MM77,MM75 and MM78, and MM88 and MM89 are com-plimentary pairs, and represent the amino-terminalportion, the central portion, and the carboxy-termi-nal portion of the synthetic BelT2 coding sequence.To construct the functional gene for expression inplants, the six synthetic oligonucleotides were ini-tially combined in 3 separate reactions as com-plimentary pairs and then annealed. The reactioncontaining the central pair of oligonucleotides(MM75 and MM78) was then treated with poly-nucleotide kinase in the presence of ATP to pro-vide 5'-monophosphates at the hydroxyl termini ofthese two oligonucleotides, to facilitate ligation tothe other two pairs. The three pairs of annealedoligonucleotides were then combined into one re-action, which included pUC18 (previously digestedwith Hind III and Bam HI), and the overlaps wereallowed to anneal. DNA ligase was then provided toligate the three pairs together, and to insert thesynthetic BelT2 coding sequence into the pUC18vector. E. coli was transformed and selected forampicillin resistance, and appropriate plasmids

were identified by DNA minipreps. The insert DNAwas sequenced to confirm the sequence as shownin Figure 7.The 5' end of the complete oligonucleotide for5 BelT2, as shown in Figure 7, had a sticky end(overlap) compatible with that generated by theendonuclease Hind III, although this site was de-stroyed in ligation to the pUC18 vector. The 3' endof the synthetic DNA has a sticky end compatiblew with that generated by endonuclease Bam HI, withthis site retained in the pUC-derived plasmid. Im-mediately adjacent to the Hind Ill-compatible site inthe synthetic insert is a recognition site for en-donuclease Nco I. This is the enzyme recognition75 site that was used to join the BelT2 coding se-quence to regulatory DNA sequences derived frompAMVBTS for construction of the functionalchimeric gene, where the central "ATG" in the NcoI recognition sequence "CCATGG" will represent20 the first codon of the BelT2 coding sequence. Atthe carboxy-terminus of the coding region, betweenBam HI site and the termination codons is a rec-ognition sequence for Pst I (CTGCAG), the enzymeused to excise the coding sequence from pUC1825 for ligation of the carboxy-terminus into the expres-sion plasmid for chimeric gene construction.As shown in Fig. 8, to construct a chimericgene capable of expression in plants, the plasmidpAMVBTS was digested to completion with en-30 zymes Nco I and Pst I, and the vector was purifiedfrom agarose gels following electrophoresis. TheNco I and Pst I digested duplex oligonucleotideencoding BelT2 was purified from pUC18 in asimilar manner, after which the two DNA fragments35 were combined and ligated. E. coli was trans-formed and selected for ampicillin resistance, andthe properly constructed plasmid pAMVBelT2 wasidentified by minipreps. This plasmid is essentiallyidentical to pAMVSTI , except for the different ami-40 no acid coding regions specifying the different in-secticidal toxins. From 5' to 3', the expressioncassette consists of the CaMV 35S promoter, themRNA 5'-noncoding translational enhancer regioncorresponding to that of alfalfa mosaic virus coat

45 protein mRNA, the synthetic coding sequence forBelT2, and the polyadenylation region fromnopaline synthase.In order to move pAMVBelT2 into plant cells,the plasmid was digested at a unique Xho I en-50 donuclease site immediately 5' to the CaMV pro-moter, and the plasmid was cointegrated by liga-tion with DNA of Xho I digested Agrobacteriumvector plasmid pTV4 by the method of Barton, etal. as described in Example 1. Following trans-55 formation of E. coli and selection for both sul-fadiazene and ampicillin resistance, the properlycointegrated plasmids pTV4AMVBelT2 were iden-tified by DNA minipreps. These plasmids were then

10

8/14/2019 EP0431829A1

11/29

EP 0 431 829 A1 209

reaction and annealed. This generates a duplexoligonucleotide with Nco I and Pst I compatibleends at the amino-terminal and carboxy-terminalends of the coding sequence, respectively. In a5 separate reaction, the plasmid pAMVBTS may bedigested to completion with enzymes Nco Iand PstI, and the vector may be purified from agarose gelsfollowing electrophoresis. The annealedoligonucleotides encoding AgalV can then be com-10 bined with the pAMVBTS vector fragment, andDNA ligase added to ligate the vector and codingregion together. E. coli may be transformed toampicillin resistance, and appropriate plasmid,pAMVAgalV, can be identified by DNA minipreps.75 The insert DNA may be sequenced to confirm thesequence as shown in Figure 10. From 5' to 3', theexpression cassette would consist of the CaMV35S promoter, the mRNA 5' noncoding region cor-responding to that of alfalfa mosaic virus coat pro-

20 tein mRNA, the synthetic coding sequence forAgalV, and the polyadenylation region fromnopaline synthase.In order to move pAMVAgalV into plant ceils,the plasmid can be digested at a unique Xho I25 endonuclease site immediately 5' to the CaMV 35Spromoter, and the plasmid can be cointegrated byligation with Xho I digested Agrobacterium vectorplasmid pTV4. Following transformation of E. coliand selection for both sulfadiazene and ampicillin30 resistance, the properly cointegrated plasmidspTV4AMVAgalV can be identified by DNAminipreps. These plasmids can be then conjugated

into Agrobacterium and used in transformation ofplant tissues as described previously in Example 135 and Barton et al., supra.Following regeneration of plants transformedwith AMVAgalV, approximately 50 plants are to bescreened using mRNA slot-blots to identify thoseplants that express the AgalV gene most strongly.40 The four plants with the most AgaiV mRNA can beallowed to flower and set seed, and the progenyanalyzed in insect feeding trials. As with AalT,toxicity is expected against some species of in-sects. In addition to self-pollination of plants con-

45 taining AgalV, the four plants with highest expres-sion can be outcrossed to plants previously iden-tified to express significant levels of either AalT(Example 1), B.t. delta-endotoxin, BelT1 (Example2), BelT2 (Example 3) or other insecticidal toxins.50 Progeny of these outcrosses may be subjected tobioassays, and it may be observed that each toxinactivity was functional in the progeny and furtherthat additive effects are apparent between some ofthe toxins, resulting in additional levels of insect55 resistance. It may be found in subsequent experi-ments that the development of resistance to thesetoxins by susceptible insects was significantly de-

layed by the presence of two different toxins within

conjugated into Agrobacterium and used in trans-formation of plant tissues as described previouslyin Example 1 and Barton et al.Following regeneration of plants transformedwith pTV4AMVBelT2, approximately 50 plants willbe screened using mRNA slot-blots to identifythose plants that express the BelT2 gene moststrongly. The four plants with the most BelT2mRNA will be allowed to flower and set seed, andthe progeny will be analyzed in insect feedingtrials. As with AalT, toxicity will be present againstsome species of insects. In addition to self-pollina-tion, the four plants with highest expression ofBelT2 genes will be outcrossed to plants previouslyidentified to express significant levels of either AalTpeptide, B.t. delta-endotoxin, or BelT1 peptide(Example 2). Progeny of these outcrosses will besubjected to bioassays, and it is expected thateach independent toxin activity will be functional inthe progeny, and further that additive effects will beapparent between some of the toxins resulting inadditional levels of insect resistance. It is expectedthat in subsequent experiments that the develop-ment of resistance to these toxins by susceptibleinsects will be significantly delayed by the pres-ence of two different toxins within the same plant.Example 4

A series of insecticidal polypeptides possess-ing neurotoxic activity toward insects (agatoxins Ithrough VI) were isolated from the venom of thefunnel web spider, Agelenopsis aperta Skinner etal., J. Biol. Chem., 264, pp. 2T5CP2i55 (1989).Following purification of the peptide and tests forbioactivity, amino acid sequences were obtainedfor each of the six peptides, which were found tobe highly homologous. The amino acid sequencefor agatoxin IV (AgalV), which was demonstrated tobe insecticidal when injected into several insectspecies, is shown below: (in single letter aminoacid code):ACVGENQQCADWAGPHCCDGYYCTCRYFPKCIC-RNNNAs in Example 1 demonstrating synthesis of agene encoding the AalT peptide, a gene may beconstructed to enable expression of AgalV inplants. Based on the most frequently used codonsin plants (Figure 1), two complimentaryoligonucleotides (KB152 and KB153) may be syn-thesized to span the amino acid coding region ofthe peptide. The synthetic nucleotide sequence toexpress the AgalV toxin is shown in Figure 10. Themanipulations for constructing a gene for expres-sion in plants and transfer of that gene into plantcells are shown in Figure 11. To construct thefunctional gene for expression in plants, the 2 syn-thetic oligonucleotides are initially combined in one

11

8/14/2019 EP0431829A1

12/29

EP 0 431 829 A11 22

the same plant.Example 5

An insecticidal polypeptide possessing neuro-toxic activity toward insects (termed in this exam-ple as "SfIT") was isolated from the venom of thecellular spider, Segestria florentina, Sagdiev et al.,Bioorganicheskaya Khimiya, 13, pp. 1013-1018(1987). Following purification of the peptide andtests for bioactivity, amino acid sequence was ob-tained. The amino acid sequence for SfIT, whichhas a molecular mass of 3988 daltons, and iscomprised of 35 amino acids, (excluding an amino-terminal methionine which is added in the examplebelow), is shown below in single letter amino acidcode:RQDMVDESVCYITDNNCNGGKCLRSKACHADPW-EL

As in Example 1 demonstrating synthesis of agene encoding the AalT peptide, a chimeric genecan be constructed to enable expression of SfIT inplants. Based on the most frequently used codonsin plants (Figure 1), two complimentaryoligonucleotides (KB154 and KB155) can be syn-thesized to span the amino acid coding region ofthe peptide; an additional codon for methionine,ATG, is included at the amino-terminus of the cod-ing region to enable appropriate translational initi-ation (Figure 12). The manipulations of theseoligonucleotides to construct an expression cas-sette and transfer it into plants are illustrated inFigure 13. To construct the functional gene forexpression in plants, the two syntheticoligonucleotides can be initially combined in onereaction and annealed. This generates a duplexoligonucleotide with Nco I and Pst I compatibleends at the amino-terminal and carboxy-terminalends of the coding sequence, respectively. In aseparate reaction, the plasmid pAMVBTS can bedigested to completion with enzymes Nco Iand PstI, and the vector can be purified from agarose gelsfollowing electrophoresis. The annealedoligonucleotides encoding SfIT can then be com-bined with the pAMVBTS vector fragment, and theDNA ligase can be added to ligate the vector andcoding region together. E. coli can be transformedand selected for ampicillin resistance, and appro-priate plasmids, pAMVSfIT, containing expressioncassettes can be identified by DNA minipreps. Theinsert DNA would be sequenced to confirm thesequence as shown in Figure 12. From 5' to 3', theexpression cassette consists of the CaMV 35S pro-moter, then the mRNA 5' noncoding region cor-responding to that of alfalfa mosaic virus coat pro-tein mRNA, the synthetic coding sequence for SfIT,and the polyadenylation region from nopaline syn-thase.

In order to move pAMVSfIT into plant cells, theplasmid may be digested at a unique Xho I en-donuclease site immediately 5' to the CaMV 35Spromoter, and the plasmid may be cointegrated by5 ligation with Xho I digested Agrobacterium vectorplasmid pTV4. Following transformation of E. coliand selection for both sulfadiazene and ampicillinresistance, the properly cointegrated plasmidspTV4AMVSflT can be identified by DNA minipreps.w These plasmids may then be conjugated into Ag-robacterium and used in transformation of planttissues as described previously in Example 1 andBarton et al., supra.Following regeneration of plants transformed75 with AMVSfIT, approximately 50 plants arescreened using mRNA slot-blots to identify thoseplants that express the SfIT gene most strongly.The four plants with the most SfIT mRNA can beallowed to flower and set seed, and the progeny20 analyzed in insect feeding trials. As with AalT,toxicity is expected to be present against somespecies of insects. In addition to self-pollination ofplants containing SfIT, the four plants with highestexpression can be outcrossed to plants previously25 identified to express significant levels of either AalT(Example 1), B.t. delta-endotoxin, BelT1 (Example2), BelT2 (Example 3) or other insecticidal toxins.Progeny of these outcrosses may be subjected tobioassays, and it will be observed that each toxin30 activity is functional in the progeny and further thatadditive effects are apparent between some of thetoxins, resulting in additional levels of insect resis-tance. It would be apparent in subsequent experi-ments that the development of resistance to these35 toxins by susceptible insects is significantly de-layed by the presence of two different toxins withinthe same plant.Claims

40 1. A plant comprising in its genome an inheritablechimeric genetic construction including a pro-moter effective to promote expression of adownstream coding sequence in plant cells, a45 coding region coding for the expression inplant cells of an insect polypeptide toxin na-tively produced by an insect predatory Ar-thropod, and a termination sequence effectiveto terminate the transcription or translation ofso the genetic construction product in plant cells,the genetic construction effective to express inthe cells of the plant sufficient amounts of thetoxin to be lethal to specific insects upon in-gestion of the plant tissue.55 2. A plant according to claim 1 wherein the toxinis an insect specific toxin.

12

8/14/2019 EP0431829A1

13/29

EP 0 431 829 A1 243

toxicity to insects ingesting the plant tissues.14. A transgenic plant comprising in its genometwo inheritable genetic constructions insertedinto the germ line of the plant by other thanMendelian inheritance, the two inheritable ge-

netic constructions each coding for the expres-sion of separate insecticidal toxins, the twotoxins having different modes of toxicity, andeach toxin being expressed in tissues of theplant in sufficient levels to be toxic to specificinsects upon ingestion.15. A transgenic plant according to claim 14wherein one toxin is a toxin natively produced

by an insect predatory animal.

3. A plant according to claim 2 wherein the toxinproduced is Aalt, BelT1 , BelT2, AgalV or SfITtoxin.4. A plant according to claim 1, 2 or 3 whereinthe coding region is a synthesized sequence

selected to include for each codon of the se-quence a codon which is preferentially utilizedby native plant genes which express well inplants cells.5. A plant according to any one of the precedingclaims wherein between the promoter and the

coding region is a 5' untranslated sequencefrom a plant virus which enhances the transla-tion of the downstream coding sequence inplant cells.6. A plant according to any one of claims 2 to 5

wherein the coding region codes for a proteinhomologous to a protein coded by the proteincoding region of a plant expression cassette ofpAMVAalT, pAMVBelTI, pAMVBelT2, pAM-VAgalV, or pAMVSflT.7. A plant according to any one of the precedingclaims wherein the plant also comprises in its

genome a second chimeric genetic construc-tion effective to express in the plant cells asecond polypeptide toxin selectively toxic tospecific insects, and with a mode of toxicityindependent from the first toxin.8. A plant according to claim 7 wherein the sec-ond toxin is derived from the delta-endotoxin ofBacillus thuringiensis.9. A plant according to any one of claims 2 to 8wherein the genome comprises a coding re-gion coding for an insect specific polypeptidetoxin natively produced by an insect predatoryscorpion.10. Seeds of a plant as claimed in any one of thepreceding claims which carry the inheritablechimeric genetic construction.11. A plant comprising in its genome a chimericplant expression cassette comprising the plantexpression cassette of pAMVAalT,pAMVBelTI, pAMVBelT2, pAMVAgalV, orpAMVSflT.12. Seeds of a plant as claimed in claim 11.13. A transgenic plant expressing in its tissues asufficient amount of a toxin natively produced

in an insect predatory animal so as to have

10

15

16. A transgenic plant according to claim 14 or 15wherein one toxin is a delta-endotoxin from20 Bacillus thuringiensis.

17. Seeds from a plant as claimed in any one ofclaims 14 to 16.25 18. A method of creating plants with improvedresistance to insect predation comprising the

steps of(a) selecting from toxins of insect predatoryanimals a polypeptide toxin;30 (b) synthesizing a nucleotide sequence cod-ing for the amino acid sequence of thepolypeptide toxin;(c) inserting the nucleotide sequence into aplant expression plasmid vector having a35 promoter effective in plant cells and a termi-nator effective in plant cells to create achimeric genetic construction for the ex-

pression of the toxin in plant cells;(d) transforming the chimeric genetic con-40 struction into the cells of plants and re-covering whole fertile transgenic plants in-cluding in their genome the chimeric ge-netic construction;(e) feeding tissues of progeny of the trans-45 genie plants to insects to test for the toxicityof the transgenic plant tissues to insects;and(f) selecting those transgenic plants whichare effectively toxic to susceptible insects

50 upon ingestion of the transgenic plant tis-sue.19. A method according to claim 18 wherein thetoxin is selected in step (a) from the toxin

55 constituents of scorpion toxin, or from toxinsproduced by spiders.

20. A method according to claim 18 or 19 wherein

13

8/14/2019 EP0431829A1

14/29

25 EP 0 431 829 A1 26

the toxin selected in step (a) is specificallytoxic only to at least some insects.21. A method according to claim 18, 19 or 20further comprising the step of self-pollinating 5the selected transgenic plants and again test-

ing the progeny produced for insect toxicity.22. A method according to any one of claims 18 to21 further comprising the step of cross-breed- 10ing the selected transgenic plants to plantsincluding an inheritable trait for expression ofanother toxin lethal to insects and selecting for

progeny plants carrying both of the traits fortoxin production. 7523. A method according to any one of claims 18 to22 for the production of plants or seeds asclaimed in any one of claims 1 to 17. 2024. A plant seed comprising in its genome achimeric genetic construction effective to ex-

press in tissues of a plant grown from the seedan insecticidal toxin natively produced by ananimal insect predator. 2525. A plant seed according to claim 24 wherein thetoxin is from a scorpion.26. A plant seed according to claim 25 wherein the 30toxin is the insect neurotoxin AalT.27. A plant seed according to claim 25 wherein thetoxin is homologous to a protein coded by theprotein coding sequence of the plant expres- 35sion cassette of pAMVAalT, pAMVBelTI,pAMVBelT2, PAMVAgalV or pAMVSflT.

40

45

50

55

14

8/14/2019 EP0431829A1

15/29

EP 0 431 829 A1

F i g . 1

fraction AmAcid Codon Number /1CCCmAcid Codon Numbar /1000 FraccGly GGG ISO. 00 10.23 0.12Sly GGA 323.00 20.66 0.25Sly GOT 355.00 22.70 0.28

.-:iz~..z-Tvr TAT 122.30 7.80 :.:

8/14/2019 EP0431829A1

16/29

EP 0 431 829 A1

F i g . 2

MM62 / " initiation codon 'ATG1

V Hind III BspMI J m k k n g y a v d s s5' AGCTTACCTGCGTCACATGAAGAAGAACGGCTACGCCGTGGACAGCAGCG

^ m 3' ATGGACGCAGTGCACTTCTTCTTGCCGATGCGGCACCTGTCGTCGC

G K A P E C L L S N Y C N N Q C TGCAAGGCCCCAGAGTGCCTCCTCAGCAACTACT [ GCAACAACCAGTGCACCCGTTCCGGGGTCTCACGGAGGAGTCGTTGATGACGTTGTTGJ GTCACGTGG

,MM63 K V H Y A D K G Y C C L L S C Y C^ -@ AAGGTGCACTACGCCGACAAGGGCTACTGCTGCCTCCTCAGCTGCTACTG

_ _ ^TTCCACGTGATGCGGGTGTTCCCGATGACGACGGAGGAGTCGACGATGACMM66 F G L N D D K K V L E I S D T R

CTTCGGCCTCAACGA [cGACAAGAAGGTGCTTGAGATCAGCGACACCAGGAGAAGCCGGAGTTGCTGCTGTTCTJTCCACGAACTCTAGTCGCTGTGGTCCT

K S Y C D T T I I N * * P S t I MM64AGAGCTACTGCGACACCACCATCATCAACTAATAGCTGCA 3 'TCTCGATGACGCTGTGGTGGTAGTAGTTGATTATCG 5 ' ^ -

MM67

16

8/14/2019 EP0431829A1

17/29

8/14/2019 EP0431829A1

18/29

EP 0 431 829 A1

pAMVBTSXholAaiT synthetic coding region

mTTiniTiiinimnnniiiiiiNco I \ Pst I(compatible) \

digest Nco I + Pst Ipurify vec tor

XholNco RB

pAMVSTI digest p la s mid swith Xho Icombine, iigateselect Sur and AprXho

pTV4AMVST1

LBF i g . 4

18

8/14/2019 EP0431829A1

19/29

8/14/2019 EP0431829A1

20/29

EP 0 431 829 A1

Bern synthetic coding regioninnillllllniiiiiiniiniTTrNcol(compatible) Pst

digest Nco I + Pst Ipurify vector

Xho(Nco Idestroyed) R8

XhopAMVBern digest plasmidswith Xho Icombine, llgateselect Sur and AprXho

pTV4AMVBelT1

F i g . 6

20

8/14/2019 EP0431829A1

21/29

8/14/2019 EP0431829A1

22/29

EP 0 431 829 A1

pAMVBTSXholBelT2 synthetic coding region

TnTmTTTirnmimTmTTTTrNCO I \ Pst I

digest Nco I + Pst Ipurify vector Pst I

XhoNCO RB

XholpAMVBeiT2 digest plasmldswith Xho Icombine, ligateselect Sur and AprXhol

pTV4AMVBelT2

F i g . 8

22

8/14/2019 EP0431829A1

23/29

8/14/2019 EP0431829A1

24/29

8/14/2019 EP0431829A1

25/29

EP 0 431 829 A1

AgalV synthetic coding region(synthetic oligonucleotidesKBi52andKB153)pAMVBTS

Xhol'"@@"" @ "@Nco I \ Pst I

digest Nco I + Pst Ipurify vector Pst I

XholXho I

RB

Pst IpAMVAgaiV digest plasmidswith Xho Icombine, ligateselect Sur and AprXhol

pTV4AMVAgalV

F i g . 1 1

25

8/14/2019 EP0431829A1

26/29

8/14/2019 EP0431829A1

27/29

EP 0 431 829 A1

Sf IT synthetic coding region(synthetic ollgonucleotidesKB 154 and KB1 55)pAMVBTS

Xholnimniiimiiii iNCOl \(compatible) \ Pstl

digest Nco I + Pst Ipurify vector Pstl

XholXhoNco I (destroyed) RB

Xhol@PstlpAMVSfIT digest plasm idswith Xho icombine, llgateselect Sur and AprXhol

PTV4AMVSHT

F i g . 1 3

27

8/14/2019 EP0431829A1

28/29

JEuropeanPatent Office Application Number

EP 90 31 2 9 4 4EUROPEAN S E A R C H

R E P O R T

DOCUMENTS CONSIDERED TO BE R E L E V A N TCitation of document with indication, where appropriate, Relevant CLASSIFICATIONOFTHECategory of relevant passages to claim APPLICATION(Int. CI.5)

Y J.M. WALKER et al.: "Molecular Biology 1-3,6-10, C12N& Biotechnology", 2nd edition, 1988, pages 117-147; chap- 13-26 15/82ter 7, M.G.K. JONES et al.: "Plant biotechnology" C 12 N 15/12*Page 142, paragraph 1 * A 01 H 5/00A 01 N 63/02D,Y JOURNAL OF BIOLOGICAL CHEMISTRY, vol. 264, no. 32, 1-3,6,9,15th November 1989, pages 19259-1 9265, US; P.E. 10,1 8-26BOUGIS et al.: "Precursors of Androctonus australis scor-

pion neurotoxins"*Whole document *Y NATURE, vol. 328, no. 6125, 2nd July 1987, pages 12-13, 7,8,14-17Macmillan Journals Ltd; R. SHIELDS: "Agricultural geneticstowards insect-resistant plants"*Column 2, paragraph 3 *Y GENE, vol. 73, no. 2, 20th December 1988, pages 409-418, 18,19Elsevier Science Publishers B.V. (Biomedical Division), Am-

sterdam, NL; L.F. CARBONELL et al.: "Synthesis of a genefor an insect-specific scorpion neurotoxin and attempts toexpress it using baculovirus vectors" technical fieldsWhnlArinn..mnf SEARCHED(Int. C..5)

Y JOURNAL OF BIOLOGY CHEMISTRY, vol. 264, no. 4, 5th 3,19 a 01 HFebruary 1989, pages 2150-2155, The American Society for a 01 NBiochemistry and Molecular Biology, Inc., US; W.S. SKIN-NER et al.: "Purification and characterization of two classesof neurotoxins from the funnel web spider, Agelenopsisaperta"*Figure 3 *

Y CHEMICAL ABSTRACTS, vol. 107, 1987, page 267, abstract 3no. 170906v, Columbus, Ohio, US; N.Zh. SAGDIEV et al.:"Study of venom toxic components of the cellular spiderSegestria florentina",&BIOORG. KHIM. 1987, 13(8), 1013-18*Abstract *

- / -

The present search report has been drawn up for all claimsPlace of search Date of completion of search ExaminerThe Hague 26 February 91 MADDOX A.D.

CATEGORYOFCITEDDOCUMENTS E: earlier patent document, but published on, or afterX: particularly relevant if taken alone the filing dateY: particularly relevant if combined with another D: document cited in the applicationdocument of the same catagory L: document cited for other reasonsA: technological backgroundO: non-written disclosure &: member of the same patent family, correspondingP: intermediate document documentT: theory or principle underlying the invention

8/14/2019 EP0431829A1

29/29

Page 2

JEuropeanPatent Office Application NumberEUROPEAN S E A R C H

R E P O R TEP 90 31 2 9 4 4

DOCUMENTS C O N S I D E R E D TO BE R E L E V A N TCLASSIFICATIONOFTHEAPPLICATION(Int. CI.5)

itation of document with Indication, where appropriate,of relevant passages

Relevantto claimategory 1-4,6-10,13-26EP-A-0 374 753 (CIBA-GEIGY)*Page 31, lines 5-57; page 35 - page 37, line 4; page 14,lines 30-55 *

WO-A-8 904 371 (LOUISIANA STATE UNIV.)*Page 11, line 31 - page 17; example 1 *WO-A-8 904 868 (AGRACETUS)*Whole document *EP-A-0 339 009 (MONSANTO)" Pages 5-7 *

P.X

7,8,14-17

TECHNICALFIELDSSEARCHED(Int. CI.5)

The present search report has been drawn up for all claimsExaminer

MADDOX A.D.Date of completion of search

26 February 91Place of searchThe Hague

E: earlier patent document, but published on, or afterthe filing dateD: document cited in the applicationL: document cited for other reasonsCATEGORYOFCITEDDOCUMENTSX: particularly relevant if taken aloneY: particularly relevant if combined with anotherdocument of the same catagoryA technological background