IE83748B1 - Synthetic plant genes and method for preparation - Google Patents

Description

SYNTHETIC PLANT GENES AND METHOD FOR PREPARATIQN relates to genetic invention particularly to The present engineering and more plant transformation in which a plant is transformed to express a heterologous gene.

Although great progress has been made in recent years with respect to transgenic plants which express foreign proteins such as herbicide resistant enzymes and viral coat proteins, very little is known about the majori factors affecting expression «of foreign genes in plants. Several potential factors could be responsible in varying degrees for the level of protein expression from a particular coding sequence.

The level of a particular mRNA in the cell is certainly a critical factor.

The potential causes of low steady state levels of mRNA due to the nature of the coding sequence are full length RNA synthesis might not This could, for example, many. First, occur at a high frequency. be caused by the premature termination of RNA during transcription or due to unexpected mRNA processing full length RNA could polyA Second, be produced but then processed (splicing, addition) in the nucleus in a fashion that creates a nonfunctional mRNA. If the ~RNA is properly synthesized, terminated and polyadenylated, it then can move to the cytoplasm for translation. In the cytoplasm, mRNAs have distinct half lives that are determined by their sequences and by the cell type in A Some RNAS are very short- In addtion, which they are expressed. lived and some are much more long—lived. ‘there is an effect, whose magnitude is uncertain, of translational efficiency on mRNA vhalf—life. In addition, every RNA molecule folds into a particular structure, or perhaps family of sturctures, which is determined by its sequence. The particular structure of any RNA might lead to greater or lesser stability Structure per se is probably also a the in the cytoplasm. determinant of mRNA processing in nucleus.

Unfortunately, it is impossible to predict, and nearly the structure of any’ RNA it is impossible to determine, in vitro or in vivo. However, (except for tRNA) likely that dramatically changing the sequence of an RNA will have a large effect on its folded structure.

It is likely that structure per se or particular structural features also have a role in determining RNA stability.

Some particular sequences and signals have been identified in RNAS that have the potential for having a specific effect on RNA stability. This section summarizes what is known about these sequences and signals. These identified sequences often are A+T rich, and thus are more likely to occur in an A+T rich coding sequence such as a B.t. gene. The sequence motif ATTTA (or AUUUA as it appears in RNA) has been implicated as a destabilizing sequence in mammalian cell mRNA {Shaw and Kamen, 1986); No analysis of the function of this sequence in plants has been done.

Many short lived mRNAs have A+T rich 3' untranslated and these regions often have the .AITTA sequence, sometimes present in mutiple copies or as multimers (e.g., ATTTATTTA...). Shaw and Kamen showed that the transfer of the 3' end of an unstable mRNA to a stable RNA kglobin or VA1) decreased the stable RNA's half life dramatically. They further showed that a pentamer of ATTTA had a profound destabilizing and that this signal could end regions, effect on a stable message, exert its effect whether it was located at the 3' or within the coding sequence. However, the number of ATTTA sequences and/or the sequence context in which they occur also appear to be important in determining whether they function as destabilizing sequences.

Shaw and Kamen showed that a trimer of ATTTA had much less effect than a pentamer on mRNA stability and a dimer or a monomer had no effect on stability (Shaw and Kamen, 1987). Note that multimers of ATTTA such as a pentamer automatically create an A+T rich region. shown to be a cytoplasmic effect, not In other unstable mRNAs, the ATTTA sequence but it is often This nuclear. may be present in only a single copy, contained in an A+T rich region. From the animal cell data collected to date, it appears that ATTTA at least win some contexts is important in stability, but it is not yet possible to predict which occurences of ATTTA are destabiling elements or whether any’ of these .effects are likely to be seen in plants.

Some studies on mRNA degradation in animal cells also indicate that RNA degradation may begin in som cases with nucleolytic attack in 5+? rich regions. it is not clear if these cleavages occur at ATTTA sequences. There are.also examples of mRNAs that have differential stability depending on the cell type in which they are expressed or on the stage within the cell cycle at which they are expressed. For example, histone mRNAs are stable during DNA synthesis but unstable if DNA synthesis is disrupted. The 3' end of some histone mRNAs seems to be responsible for this effect (Pandey and Marzluff, 1987‘). It does not appear to be mediated by ATTTA, nor is it clear what controls the ‘differential stability" of this mRNA.

Another example is the differential stability of IgG mRNA in B lymphocytes during B cell maturation (Genovese and Milcarek, 1988). A final example is the instability of a mutant beta-thallesemic globin mRNA.

In bone marrow cells, where this gene is normally expressed, the mutant mRNA is unstable, while the wild- type mRNA is stable. When_ the mutant gene is expressed in HeLa or L cells in vitro, the mutant mRNA shows no instability (Lim et al., i988). These examples all provide evidence that mRNA stability can be mediated by cell type or cell cycle specific Furthermore this type of instability is not Given these factors. yet associated with specific sequences. uncertainties, it is not possible to predict which RNAS are likely to be unstable in a given cell. In addition, even the ATTTA motif may act differentially depending on the nature of the cell in which the RNA is present. Shaw and Kamen (1987) have reported that activation of protein kinase C can block degradation mediated by ATTTA. ‘mature 3"terminus.

The addition of a polyadenylate string to the 3' end is common to mostweucaryotic mRNAs, both plant and animal. The currently accepted view of polyA addition is that the nascent transcript extends beyond the Contained within this transcript are signals for polyadenylation and proper 3.1 end formation. This processing at the 3' end involves cleavage of the mRNA and addition of polyA to the mature 3’ end. By searching for consensus sequences near the polyA tract in both plant and animal mRNAs, it has been possible to identify consensus sequences that apparently are involved in polyh addition and 3’ The same consensus sequences seem to be These signals end cleavage. important to both of these processes. are typically a variation on the sequence AATAAA. In animal cells, some variants of this sequence that are functional have been identified; in plant cells there seems to be an extended range of functional sequences (Wickens and Stephenson, 1984; Dean et al., 1986).

Because all of these consensus sequences are variations on AATAAA, they all are A+T rich sequences.

This sequence is typically found 15 to 20 bp before the polyA tract iJ1 a mature mRNA. Experiments in animal cells indicate that this sequence is involved in both polyp. addition and» 3' maturation. Site directed mutations in this sequence can disrupt these functions (Conway and Wickens, 1988; Wickens et al., 1987). However, it has also been observed that sequences up to S0 to 100 bp 3' to the putative polyA signal are also required; i.e., a gene that has a normal AATAAA but has been replaced or disrupted downstream does not get properly polyadenylated (Gil and Proudfoot, 1984; Sadofsky and Alwine, 1984; McDevitt et al., 1984). That is, the polyA signal itself is not sufficient for' complete and proper processing: It is not yet known what specific downstream sequences are required in addition Ep the polyA signal, or if there is a specific sequence that has this function. Therefore, sequence analysis can only identify potential polyn signals.

In. naturally? occuring umNAsn that are normally polyadenylated, it has been observed that disruption of this process, either by altering the polyh signal or other sequences in the mRNA, profound effects can be obtained in the level of functional mRNA. This has been observed in several naturally occuring mRNAs, with results that are gene specific so far. There are no general rules that can be derived yet from the study of mutants of these natural genes, and no rules that can be applied to heterologous genes. Below are four examples: . In a globin gene, absence of a proper polyA site leads to improper termination of transcription.

It is likely, but not proven, that the improperly terminated RNA is nonfunctional and unstable (Proudfoot et al., 1987).

. In a globin gene, absence of a functional polyA signal can lead to a 100—fold decrease in the level of mRNA accumulation (Proudfoot et al., 1987).

. A globin gene polyA site was placed into the 3' ends of two different histone genes. The histone genes contain a secondary structure (stemrloop) near their 3' The amount of properly polyadenylated histone mRNA producedvfrom these chimeras decreased as the distance between the stem—1oop and the polyA site Also, the two histone genes produced ends. increased. ‘ greatly different levels of properly polyadenylated mRNA. site and other sequences on the mRNA that can modulate mRNA accumulation (Pandy and Marzluff, 1987).

. The soybean leghemoglobin gene has been cloned into HeLa cells, and it has been determined that this plant gene contains a "cryptic" polyadenylation signal that is active in animal cells, but is not utilized in This leads to the production of a new This again .This suggests an interaction between the polyA plant cells. polyadenylated mRNA that is nonfunctional. shows that analysis of a gene in one cell type cannot cell types behavior in alternative ). predict its (Wiebauer et al., From these examples, mRNAs proper polyadenylation is ihmortant in mRNA accumulation, and that disruption of this process can effect mRNA However, insufficient knowledge exists to predict the effect of In a heterologous gene, it is clear that in natural levels significantly. changes in a normal gene. where we do not know if the putative polyA sites (consensus sequences) are functional, it is even harder to predict the consequences. possible that the putative. sites identified are That is, these sites may not act as but instead function as aberrant However, it is disfunctional. proper po1yA sites, sites that give rise to unstable mRNAs.

In animal cell systems, AATAAA is by far the most common signal identified ix: mRNAs upstream of the polyA, but at least four variants have also been found (Wickens and Stephenson, 1984). In plants, not nearly ‘so much analysis has been done, but it is clear that multiple sequences similar to AATAAA can be used. The plant sites below called major or minor refer only to the study of Dean et al. (1986) uhich analyzed only three types of’ plant gene. The designation of polyadenylation sites as major or minor refers only to the frequency of their occurrence as functional sites in naturally occurring genes that have been analyzed.

In the case of plants this is a very limited database.

It is hard to predict with any certainty that a site designated major or minor is more or less likely to 11'] a function partially’ or completely when found heterologous gene such as B.t.

PA AATAAA Major consensus site PIA AATAAT Major plant site PZA AACCAA Minor plant site P 3A ATATAA " P4A AATCAA " PSA ATACTA " P6A ATAAAA " P7A ATGAAA " P8A -AAGCAT " P9A ATTAAT " PIOA ATACAT " P11A AAAATA " PIZA ATTAAA Minor animal site P13A AATTAA " P14A AATACA " PISA CATAAA " Another type of RNA processing that occurs in the nucleus is intron splicing. Nearly all of the work on intron processing has been done in animal cells, but some data is emerging from plants. Intron processing depends on proper 5' and 3' splice junction sequences.

Consensus sequences for these junctions have been derived for both animal and plant mRNAs, but only a few nucleotides are known to be invariant. Therefore, it is hard to predict with any certainty whether a putative splice junction is functional or partially functional based solely on sequence analysis. In particular, the only invariant nucleotides are GT at the 5' end of the intron and AG at the 3' end of the intron. In plants, at every nearby position, either within the intron or in the exon flanking the intron, all four nucleotides can be found, although some positions show some nucleotide preference (Brown, 1986; Hanley and Schuler, 1988).

A plant intron has been moved from a patatin gene into a GUS gene. To do this, site directed mutagenesis was performed to introduce new restriction sites, and this mutagenesis changed several nucleotides in the intron and exon sequences flanking the GT and AG. This intron still functioned properly, indicating the importance of the GT and AG and the flexibility at other nucleotide positons. There are -10.. of course many occurences of GT and AG in all genes that do not function»as intron splice junctions, so there must be some other sequence or structrual features that identify splice junctions. In plants, one such feature appears to be base composition per se. Wiebauer et al. (1988) and Goodall et al. (1983) have analyzed plant introns and exons and found that exons have ~50% A+'1‘ while introns have ~70% A+T.

Goodall et al. (1988) also created an artificial plant intron that has consensus S‘ and 3' splice junctions and a random A+T rich internal sequence. This intron was spliced correctly in plants. when the internal segment was replaced by a G+C rich sequence, splicing efficiency was drastically reduced. These two examples demonsatrate that intron recognition in plants may depend on very general features —— splice junctions that have a great deal of sequence diversity and A+T richness of the intron itself. This, of course, makes it difficult to predict from sequence alone whether any particular sequence is likely to function as an active or partially active intron for RNA processing.

B.t. genes being A+T rich contain stretches of various lengths that have 70% or greater A+T. The number of such stretches identified by sequence analysis depends on the length of sequence numerous scanned.

As for polyadenylation described above, there are complications in predicting what sequences might be utilized as splice sites in any given gene. First, many naturally occuring genes have alternative splicing pathways that_create alternative combinations of exons in the final mRNA (Gallega and Nadal4Ginard, .1988; Helfman and Ricci, 1988; Tsurushita and Korn, 1989). That is, some splice junctions are apparently recognized under some circumstances or in certain cell types, but not in others. The rules governing this In addition, there can be an such that are not understood. interaction between processing paths utilization of a particular polyadenylation site can interfere with splicing at a nearby splice site and vice versa (Adami and Nevins, 1988; Brady and Wold, 1988; Marzluff and Pandey, 1988). Again no predictive rules are available. Also, sequence changes in a gene can drastically alter the utilization of particular splice junctions. For example, in a bovine growth hormone gene, small deletions in an exon a few hundred bases downstreanx of an intron cause the splicing efficiency of the intron to drop from greater than 95% to less than 2% (essentially nonfunctional). Other deletions however have essentially no effect (Hampson and Rottman, 1988). Finally, a variety of in vitro and in vivo experiments indicate that mutations that disrupt normal splicing lead to rapid degradation of the RNA in the nucleus. Splicing is a multistep process in the nucleus and mutations in normal splicing can lead to blockades in the process at a Any of these blockades can then Studies of variety of steps. lead to an abnormal and unstable RNA. mutants of normally processed (polyadenylation and relevant to_ the study of genes might splicing) genes are heterologous genes such as B.t. B;t. __12_ contain functional signals that lead to the production of aberrant nonfunctional mRNAs, and these mRNAs are likely to be unstable. But the B.t. genes are perhaps even more likely to contain signals that are analogous to mutant signals in a natural gene. As shown above these mutant signals are very likely to cause defects in the processing pathways whose consequence is to produce unstable mRNAs.

It is not known with any certainty what signals RNA transcription termination in plant or animal cells.

Some studies on ‘animal genes that indicate that stretches of sequence rich in T cause termination by calf thymus RNA polymerase II in vitro. These studies have shown that the 3' ends of in vitro terminated transcripts often lie within runs of T such as T5, T6 or T7- Other identified sites have not been composed solely of T, but have had one or more ‘other nucleotides as well. Termination has been found to occur within the sequences TATTTTTT, ATTCTC, TTCTT (Dedrick et al., 1987; Reines et al., 1987)- In the case of these latter two, the context in which the. sequence is found has been C+T rich as well. It is not known if this is essential. Other studies have implicated stretches of A as potential transcriptional terminators. An interesting example from SV4O illustrates the uncertainty in defining terminators based on sequence alone. One potential terminator in SV40 was‘ identified as being A rich and having a region of dyad symmetry (potential stem-loop) 5’ to the A rich stretch. However, a second terminator identified experimentally downstream in the same gene ...13... was not A rich and included no potential secondary structure (Kessler et al., 1988). Of course, due to the A+T content of B.t. genes, they are rich in runs of A or T that could act as terminators. The ‘importance of termination to stability of the mENA is shown by the globin gene example described above.

Absence of a normal polyA site leads to a failure in proper termination with a consequent decrease in mRNA.

There is also an effect on mRNA stability due the translation of tin: mRNA. Premature translational termination in human triose phosphate isomerase leads to instabilityv of the mRNA (Daar et al., 1988).

Another example is the beta-thallesemic globin mRNA described above that is specifically unstable in bone marrow cells (Lim et al., 1988). The defect in this mutant gene is a single base pair deletion at codon 44 that leads to translational termination (a nonsense codon) at codon 60. Compared to properly translated normal globin mRNA, this mutant RNA is very unstable.

These results indicate that an improperly translated mRNA is unstable. Other work in yeast indicates that proper but poor translation can have an effect on mRNA levels. A heterologous gene was modified to convert certain codons to more yeast preferred codons. An overall 10-fold increase in protein production was achieved, but there was also about a 3-fold increase in mRNA floekema et al., 1987). This indicates that more efficient translation can lead to greater mRNA stability, and that the effect of codon usage can be at the RNA level as well as the translational level.

It is not clear from codon usage studies which codons _14- lead to poor translation, or how this is coupled to mRNA stability.

EP-A-Q 359 472 discloses modifying B.t. sequences to render them more planttlike. The sequence is modifiied so that the codon usage in the sequence is approximately the same as the codon usage in a plant. In contrast, the claimed invention is related to a specific methodology for incree-s1-ng the ‘expression of the gene in a plant by removing the occurrence of particular DNA sequences- Therefore, it is an object of the present invention to provide a method for preparing synthetic plant genes which express their respective proteins at relatively high levels when compared to uild—type genes. It is yet another object. of the present invention to provide synthetic plant genes which express the ‘crystal protein toxin of Bacillus thuringiensis at relatively high levels.

BE15E_DEscRIPTIoN OF THE DRAWINGS illustrates the steps employed in gene to Figure increase expression modifying a wild—type efficiency in plants.

A Figure 2 illustrates a comparison of the modified B.tvk. HD—1 sequence of Example 1 versus the wild—type sequence of Bjc.k_ (upper line). the changes in (lower line) HD—l which encodes the crystal protein toxin Figure 3 illustrates A comparison of the changes in the synthetic B.t.k. HD-1 sequence of Example 2 (lower line) versus the wild-type sequence of Bet.k. HD—l which encodes the crystal protein toxin (upper line).

Figure 4 illustrates a conparison of the changes in ED-73 sequence of Sxanple 3 the synthetic B.t.k- type sequence of B.c.k. (lower line) versus the wild- HD-73 (upper linel.

Figure 5 represents a plasmid map of intermediate slant transformation vector Cassette PMON393» ..]_5__ Figure 6 represents a plasmid map of intermediate plant transformation vector cassette pMON900; Figure 7 represents a map for the disarmed T—DNA of A. tumefaciens ACO.

Figure 8 illustrates a comparison of the changes in the synthetic truncated B.t.k. ED-73 gene (Amino acids 29~615 with an N—terminal Met—Ala) of Example 3 (lower line) versus the wild-type sequence of B.t.k. HD—73 (upper line).

Figure 9 illustrates a comparison of the changes in the synthetic/wild—type full length- B.t.k. HD—73 (lower line) versus the wild— HD-73 (upper sequence of Example 3 type full—length sequence of B.t.k. line).

Figure 10 illustrates a comparison of the changes in the. synthetic/modified full length B-t.k. HD-7.3 (lower line) versus the wild- (upper sequence of Example 3 type full—length sequence of B.t.kp HD-73 line).

Figure 11 illustrates a comparison of the changes fully synthetic fulltlength B.t.k. HD—73 versus the wild— (upper in the sequence of Example 3 (lower line) type full—length sequence of .B.t.k. HD—73 line).

Figure 12 illustrates a comparison of the changes sequence of Example 5 (lower which in the synthetic B.t.t. line) versus the wild-type sequence of B.t.t. encodes the crystal protein toxin (upper line).

Figure 13 illustrates a comparison of the changes in the synthetic B.t. P2 sequence of Example 6 (lower line) versus the wild—type sequence of B.t.k. -1 6_ HD-1 which encodes the P2 protein toxin (upper line).

Figure 14 illustrates a comparison of the changes in the synthetic B.t- Example 7 of B.t. entomocidus sequence of (lower line) versus the wild—type sequence entomocidus which encodes the Btent protein toxin (upper line).

Figure 15 illustrates a plasmid map for plant expression cassette vector pMON744.

Figure 16 illustrates a comparison of the changes in the (PLRV) coat synthetic potato leaf roll virus protein sequence of Example 9 (lower line) versus the wild—type coat protein sequence of PLRV (upper line).

SIBIEMEEI_QE_IHE_1EEEflIIQfl The present invention provides a method for modifying a wild—type Structural gene sequence which encodes an insecticidal protein of Bacillus thuringiensis to enhance the.expression of said protein in plants which comprises: identifying regions within said sequence with greater than four consecutive adenine or thymine nucleotides; modifying the regions of step (a) which have two or more polyadenylation signals within a ten base sequence to remove said signals while maintaining a gene sequence which encodes said protein; and modifying the 15-30 base regions surrounding the regions of step (a) to remove major plant polyadenylation sig- nals, consecutive sequences containing more than one minor polyadenylation signal and consecutive seq"e"Ces containing more than one ATTTA sequence while maintaining agenesequenggwhmh encodes said orotein.

The present in‘/9I1CiOn‘further provides a method for modifying a wild—type structural gene sequence which encodes an insecticidal protein of Bacillus thuringiensis to enhance the expression of said protein in plants which comprises: a) removing polyadenylation signals contained in said wild- type gene while retaining a sequence whfch encodes said protein; and b) removing ATTTA sequences contained in said wild—type gene while retaining a sequence which encodes said protein.

According to a further embodiment a method for improving the ex ression of a heterolo ous ene in lants is rovided wherein P ‘J 9 P P said‘ gene comprises a modified chimeric gene containing a promoter which functions in plant cells operably linked to a and a 3' non«translaLed region coding sequence structural containing a polyadenylation signal which functions in plants to cause the addition of polyadenylate nucleotides to the 3' end of the RNA, wherein said structural coding sequence encodes an insecticidal protein at least a portion of which was derived from a Bacillus thurlngiensis protein; wherein said method comprises modifying said structural coding sequence so that said sequence has a DNA sequence which differs from the naturally occurring DNA sequence encoding said Bacillus thuringiensls protein and said structural coding sequence does not contain more than S consecu~ of either adenine or thymine tive nucleotides consisting residues.

As a further embodiment of the present invention a method for improving the expression of a hcterologous gene in plants is provided wherein said gene comprises a modified chimeric gene containing a promoter which functions in plant cells operably linked to a structural coding sequence and a 3’ non-translated region containing a polyadenylation signal which fimctions in plants to cause the addition of polyadenylatc nucleotides to the 3’ end of the RNA, wherein said structural coding sequence encodes an insecticidal protein at least a portion of which was derived from a Bacillus thuringiensis protein, wherein said method comprises modifying said structural coding sequence so that said sequence has a DNA sequence whichdiffers from the naturally occurring DNA sequence encoding said Bacillus thuringiensis protein and has the following characteristics: said structural coding sequence has a region which is complernentary to the following sequence: GGCIT GATTCCT AGCGAACT CTT CGATTCT CT GGTTGATGAGCT GTTC 1 5 IO 15 20 25 30 35 40 45 said region in said coding sequence having eliminated 2 AACCAA and l AATTAA sequence.

The present invention provides a method for -Pfiepafing synthetic plant genes which encode the crystal protein toxin of Bacillus thuringiensis (B.t.) . Suitable B.t. subspecies include, but are not limited to, B.t. kurstaki HD—1, B.t. kurstaki HD—73, B.t. sotto, B.t. berli1er, B.t. thuringiensis, B_t:. tolworthi, B.t. denruolimus, B.t. alesti, B. t. _19_ galleriae, B.t. aizawai, B.t. subtoxicus, B_t_ entomocidus, B.t¢ tenebrionis and B.t. san diego The expression of B.t. genes in plants is problematic. Although the expression of B.t. genes in plants at insecticidal levels has been reported, this accomplishment hast not been’ straightforward- In Vparticulas, the expression of a full-length Iepidopteran specific B.t; gene (comprising DNA from a B.t.k; isolate) has been reported to.be unsuccessful ‘in yielding insecticidal levels of expression in some plant species (vaeck et al., 198? and Barton et al., 1987).

It has been reported that expression of the full- length gene from B.t-k- HD-1 was detectable in tomato plants but that truncated genes led to :1 higher frequency of insecticidal plants with an overall higher level of expression. Truncated genes of B.t. .berlinez' also led to a higher frequency of insecticidal plants in tobacco (Vaeck et al., 1987).

On the other hand, insecticidal plants were provided from lettuce transformants using a full—length gene.

It has also been reported that the full length gene from B.t.k. HU—73 gave some insecticidal effect in tobacco (Adang et al., 1987). However, the B.t. mRNA detected in these plants was only 1.7 kb compared to 38-2l(1OSl5)A is biologically active. Therefore, the low level of B.t. levels of B.t. mRNA.

Messenger RNA levels are determined by the rate of It is the balance toxin expression may be the result of the low synthesis and rate of degradation. between these two that determines the steady state level of mRNA. The rate of maximized by the use of the CaMV 35S promoter, a strong constitutive plant expressible promoter. The synthesis has been use of other plant promoters such as nopaline synthase (NOS), (MAS) and ribulose bisphosphatecarboxylase small subunit (RUBISCO) mannopine synthase have not led to dramatic changes in the levels of B.t toxin protein expression indicating that the effects toxin protein levels are promoter that the determining B.t. independent. These data imply coding sequences of DNA genes encoding B.t. toxin proteins are somehow responsible for the poor expression level, and that this effect is manifested by a low level of accumulated stable mRNA.

Lower than expected levels of mRNA have been observed with four different lepidopteran (two from B.t.k. HD—l; B.t. berliner and B.t.k. specific genes HD—73) as from the coleopteran that for well as the gene B.t. It appears tenebrionis. type B.t. strongly in the truncated coding sequences. specific these effects are full length coding lepidopteran genes manifest more in the sequences than These effects are seen across plant species although their magnitude seems greater in some plant species such as tobacco. are protected against all important lepidopteran pests (or against Colorado potato beetle in the case of-B.t. tenebrionis), and in addition to have a level of B.t. _ expression that provides an additional safety margin over and above the efficacious protection level. It is also important to devise plant genes which function reproducibly from species to species, so that insect resistant plants -can be obtained in a predictable fashion.

In order to achieve these goals, it is important to understand the nature of the poorer than expected The level of than genes in plants. much lower expression of B.t. stable B.t. mRNA That is, compared to other coding sequences in plants is expected. driven by the same promoter, the level of B.t. mRNA measured by Northern analysis or nuclease protection experiments is much lower. For example, tomato plant 337 (Fischhoff’et al., 1987) was selected as the best expressing plant with pMON9711 which :contains the B.t.k. HD-1 KpnI fragment driven by the C3MV 35S promoter and contains the NOS-NPTII*NOS selectable marker gene. In this plant the level of B.t. mRNA is between 100 to 1000 fold lower than the level of NPTII mRNA, even though the 35S promoter is approximately 50- fold stronger than the NOS promoter (Sanders et al., 1987). indicating that the toxin protein produced in plants 'synthesis and rate of degradation. is biologically active. Therefore, the low level of B.t. toxin expression may be the result of the low levels of B.t. mRNA.

Messenger RNA levels are determined by the rate of It is the balance between these two that determines the steady state level of mRNA. The rate of synthesis has ‘been maximized by the use of the CaMV 35S promoter, a strong constitutive plant expressible promoter. The use of other plant promoters such as nopaline synthase (NOS), mannopine (MAS) and‘ ribulose bisphosphatecarboxylase small subunit (RUBISCO) have isynthase not led'to dramatic changes in the levels of B.t. toxin protein expression indicating that the effects toxin protein levels are promoter imply that the determining" B.t. independent. These data sequences of DNA genes encoding B.t. toxin proteins are somehow responsible for the poor expression level, and that this effect is manifested by a low level of coding accumulated stable mRNA.

Lower than expected levels observed with four different lepidopteran specific genes (two from B.t.k. HD~l; B.t. berliner and B.t.k.

HD-73) as well as the gene from the coleopteran specific B.t. It appears that for these effects are of mRNA have been tenebrionis. lepidopteran type B.t. manifest more strongly ;h1 the full length coding sequences than in the truncated coding’ sequences. genes These effects are seen across plant species although their magnitude seems greater in some plant species such as tobacco.

The nature of the coding sequences of B.t. genes distinguishes them from plant genes as well as many other heterologous genes expressed in plants. In particular, B.t. genes are very rich (~62%) in adenine (A) and thymine (T) while plant genes and most bacterial genes which have been expressed in piants are on the order of 45-55% A+T. The A4T content of the genomes (and thus the genes) of any organism are features of that organism and reflect its evolutionary While within any one organism genes have the A+T content can vary history. similar A+T content, tremendously from organism to organism. For example, some Bacillus species have among the most A+T rich genomes while some Steptomyces species are among the least A+T rich genomes (~30 to 35% A+T).

Due to the degeneracy of the genetic code and the limited number of codon choices for any amino acid, most of the "excess" A+T of the structural coding sequences of some Bacillus species are found in the third position of the codons. That is, genes of some Bacillus species have A or T as the third nucleotide Thus A+T in part can In addition, in many codons. content determine codon usage bias. it is clear that genes evolve for maximum function in the organism This means that particular from one in which they evolve. nucleotide sequences where they may play no role except to code have the found in a gene organism, for a particular stretch of amino acids, potential to berrecognized as gene control elements in another organism (such as transcriptional promoters or terminators, polyA addition sites, intron splice sites, or specific mfiNA degradation signals). It is perhaps surprising that such misread signals*are not a more common feature of heterologous gene expression, but-this can be explained in part by the relatively homogeneous A+T content (~50%) of umny' organisms.

This A+T content plus the nature of the genetic code put clear constraints on the likliehood of occurence of any particular oligonucleotide sequence. Thus, a gene from E. coli with a 50% h+T content is much less likely to contain any particular A+T rich segment than a gene from B. thuringiensis.

As described above, the expression of Blt. toxin protein in plants has been problematic. Although the observations made in other systems described above offer the hope of a means to elevate the expression level of B.t. toxin proteins in plants, the success obtained by the present method is quite unexpected.

Indeed, inasmuch as it has been recently reported that expression of the full—length B.t.k. toxin protein in tobacco makes callus tissue necrotic (Barton et al., 1987); one would reasonably expect that high level expression of B.t. toxin protein to be unattainable due to the reported toxicity effects.

In its most rigorous application, the method of the present invention involves the modification of an ("structural existing structural coding sequence gene").which codes for a particular protein by removal sequences and putative polyadenylation site directed mutagenesis of the DNA It is most preferred of ATTTA signals by comprising the structural gene., that substantially all the polyadenylation signals and ATTTA removed although enhanced expression levels are observed with onlyi partial removal of either of the above identified sequences. sequences are gA1ternately if 21 synthetic gene is prepared which codes ford the expression of ‘the subject protein, codons are selected to avoid the ATTTA sequence and putative polyadenylation signals. For purposes of the present invention putative polyadenylation signals include} but are not necessarily limited to, AATAAA, AATAAT, AACCAA, ATATAA; -AATCAA, ATACTA, ATAAAA, ATGAAA, AAGCAT, ATTAAT, ATACAT, AAAATA, ATTAAA, AATTAA, AATACA and CATAAA. In replacing the ATTTA sequences and polyadenylation signals, preferably utilized which avoid the codons which are codons are rarely found in plant genomes.

Another embodiment of the present represented in the flow diagram of Figure 1, employs a method for the modification of an existing structural gene or alternately the de novo synthesis of a structural gene which method is somewhat less rigorous Referring to invention, than the method first described above.

Figure 1, the selected DNA sequence is scanned to identify regions with greater than four consecutive adenine (A) nucleotides. The A+T regions are plant polyadenylation signals. or more consecutive A'or T nucleotides eliminates most plant polyadenylation signals, if there are more than one of the minor polyadenylation signals identified of each other, then the or thymine (T) scanned for potential Although the absence of five within ten nucleotides nucleotide sequence of this region is preferably Vdependent upon (1) _25_ altered to remove these signals while maintaining the original encoded amino acid sequence.

The second step is to consider the 15 to 30 enucleotide regions surrounding the A+T rich region identified in step one. If the A+T content of the surrounding region is less than 80%, the region should be examined for polyadenylation'signals. Alteration of the region based on polyadenylation signals is I the number of polyadenylation signals present and (2) presence of a major plant polyadenylation signal.

The extended region is examined for the presence of plant polyadenylation signals. The polyadenylation signals are removed by site-directed mutagenesis of the DNA The extended region is also examined for multiple copies of the ATTTA sequence sequence. which are also removed by mutagenesis.

It is also preferred that regions comprising many consecutive A+T bases or G+C bases are disrupted since these regions are predicted to have a higher likelihood to form hairpin structure due to self— insertion of complementarity. Therefore, heterogeneous base pairs would reduce the likelihood of self—complementary secondary structure formation which are known to inhibit transcription and/or translation in some organisms. In most cases, the adverse effects may be minimized by using sequences which do not contain more than five consecutive A+T or G+C.

The oligonucleotides used in the mutagenesis are - designed to maintain the proper amino acid sequence and reading frame and preferably to not introduce common restriction sites such as BglII, HindIII, £acI, KpnI, ECQRI, NcoI, PstI and San into the modified gene. These restriction sites are found in multi- linker insertion sites of cloning vectors such as plasmids pUCi18 and pMON7258. Of course, the introduction of new polyadenylation signals, ATTTA sequences or consecutive stretches of more than five A+T or G+C, should also be avoided. The preferred size for the oligonucleotides is around 40~50 bases, but fragments ranging from 18 to 100 bases have been utilized. In most cases, a minimum of 5 to 8 base pairs of homology to the template DNA on both ends of the synthesized fragment are maintained to insure proper hybridization of the primer to the template.

The oligonucleotides should avoid sequences longer than five base pairs A+T or G+C. Codons used in the replacement of wild—type codons should preferably avoid the TA or CG doublet wherever possible. Codons are selected from a plant preferred codon table (such as Table I below) so as to avoid codons which are rarely found in plant genomes, and efforts should be made to select codons to preferably adjust the G+C content to about 50%. -28 _ Table I CGA C GC CGG CGU AGA 11.66 C UA CUC CUG C UU UUA UUG UCA UCC UCG UCU AGC AGU ACA ACC ACG ACU CCA CCC CCG CCU GCA GCC GCG GCU Percent Usage 29 23 28 3 21 21 41 7 19 Table I — continued Amin9_A2id QQdQn GL1’ GGA GGG GGG GGU ADA AUG 7 A00 GUA GUC GUG GUU AAA AAG AAC AAU CAA CAG CAC CAU GAA GAG GAC GAU UAC UAU UGC UGU Percent Usage 11 45 43 28 Table I - continued Preferredicodon Usage in Plants Percent Usage CQdQfl PHE UUC 56 UUU 44 MET AUG 100 TRP UGG 100 Regions with many consecutive A+T bases or G+C bases are predicted to have a higher likelihood to form hairpin structures due to self—complementarity.

Disruption of these regions by the insertion of heterogeneous base pairs is preferred and should reduce the likelihood of the formation of self= complementary secondary structures such as hairpins which are known in some organisms to inhibit transcription (transcriptional terminators) and translation (attenuators). However, it is difficult to predict the biological effect of a potential hairpin forming region.

It is evident to those skilled_in the art that while the above description is directed toward the modification of the DNA sequences of wild-type genes, the present method can be used to construct a completely synthetic gene for‘ a given amino acid sequence. Regions with five or more consecutive A+T or G+C nucleotides should be avoided. Codons should be selected avoiding the TA and CG doublets in-codons whenever possible. Codon usage can be normalized against a plant preferred codon usage table (such as Table I) and the G+C content preferably adjusted to about 50%. The resulting sequence should be examined to ensure‘ that there are minimal putative plant polyadenylation signals and ATTTA sequences Restriction sites found :h1 commonly used cloning also preferably avoided. However, restriction -vectors are placement of several unique sites throughout the gene is useful for analysis of gene expression or construction of gene variants.

The expression of ea plant gene which exists in double—stranded DNA form involves transcription of messenger RNA (mRNA) from one strand of the DNA by RNA and the subsequent processing of This polymerase enzyme, the mRNA primary transcript inside the nucleus. processing involves a 3' non—translated region which adds polyadenylate nucleotides to the 3' end of the RNA. Transcription of DNA into mRNA is regulated by a region of DNA usually referred to as the "promoter." The promoter region contains a sequence of bases that signals RNA polymerase to associate with the DNA and to initiate the transcription of mRNA using one of the DNA strands as a template to make a corresponding strand of RNA. I A number of promoters which are active in plant cells have been described in the literature. These include the nopaline synthase (NOS) and. octopine synthase (OCS) promoters (which are carried on tumor- _32_ inducing plasmids of_Agrobacterium tumefaciens), the Cauliflower Mosaic Virus (CaMV) 19S and 35S promoters, the light—inducible promoter from the small subunit of ~ribulose bis-phosphate carboxylase (ssRUBISCO, a very abundant plant polypeptide) and the mannopine synthase (MAS) (Velten— et al. 1984 and Velten & Schell, 1985); All of these promoters have been used to create various types of DNA constructs which have (see e.g., PCT publication promoter been expressed in plants W084/02913 (Rogers et al., Monsanto).

Promoters which are known or are found to cause transcription of RNA in plant cells can be used in the present invention. Such promoters may be obtained from plants or plant viruses and include, but are not limited to, the CaMV3SS promoter and promoters isolated from plant genes such as ssRUBISCO genes. As described below, it is preferred that the particular promoter should be capable of causing sufficient expression to result in the production of selected an effective amount of protein.

The promoters used in the DNA constructs (i.e. chimeric plant genes) of the present invention may be to affect their control the CaMV35S promoter modified, if desired, characteristics. For example, may be ligated to the portion of the ssRUBISCO gene that represses the expression of SSRUBISCO in the absence of light, to create a promoter which is active The resulting chimeric described herein. For the phrase "CaMV35S" in leaves but not in roots. promoter may be used as purposes of this description, promoter thus includes variationswof CaMV3SS promoter, _33_ e.g., promoters derived by means of ligation with operator regions,‘ ‘random or controlled -mutagenesis, etc. Furthermore, the" promoters may be altered to contain multiple "enhancer sequences" to assist in elevating~ -gene expression .

The RNA produced by a DNA construct of the present invention also contains a 5' non-translated leader sequence. This sequence can be derived from the promoter selected to express the gene, and can be specifically modified "so as to increase translation of the mRNA. The 5' non-translated regions can also be obtained from viral RNA's, from suitable eukaryotic genes, or from a synthetic gene sequence. The present invention is not limited to constructs, as presented in the following examples. Rather, the non—translated leader sequence can be part- of the 5' end of the non- translated region of the coding sequence for the virus coat protein, or part of the promoter sequence, or can be derived from an unrelated promoter or coding sequence. In any case, it is preferred that the sequence flanking the initiation site conform to the translational consensus sequence rules for enhanced translation initiation reported by Kozak (1984).

The DNA construct of the present invention also contains a modified or fully-synthetic structural coding sequence encoding the crystal toxin protein of Bacillus thuringiensis; ‘which has been changed to enhance, the performance Ofi‘ Ehe gene in plants.

The structural genes of t_:‘h'e present invention may optionally encode a fusion protein comprising an amino-terminal chlorop;'l.~ast ftra-Ilsit Peptide 01‘ secretory signal sequence (see-for instance, Examples and 11). i 38-2l(lO515)A EcoRV fragment (extending from -343 to -90) and inserted into the same plasmid between the HindIII and The enhanced CaMV35S promoter thus -343 PstI sites. contains and -90 (Kay et al., a duplication of sequences between ).

The 3' end of the 75 gene is derived from the 75 gene contained on the clone designated 17.1 (Schuler et al., 1982). This 3' end fragment, which includes the polyadenylation signals, extends from an AvaII site located about 30 bp upstream of the termination codon for the beta-conglycinin gene in clone 17.1 to an EcoRI site located about 450 bp downstream of this termination codon.

The remainder of pMON893 contains a pBR322 which provides segment of an origin of replication in E. coli and a region for homologous recombination with disarmed T-DNA iJ1 Agrobacterium strain ACO the oriV region from the broad host the (described below); plasmid RK1; streptomycin/spectinomycin range the resistance gene from Tn7; and a chimeric NPTII gene, and the nopaline containing the CaMV3SS promoter (NOS) 3' resistance in transformed plant cells. synthase end, which provides kanamycin vector Figure 6, transformation is a derivative of pMON893.

Referring to plasmid pMON900 enhanced CaMV35S promoter of pMON893 has been replaced (MAS) The other segments are the same with the 1.Skb mannopine synthase promoter (Velten et al. 1984). as_plasmid pMON893. construct into plasmid vector pMON893 or pMON900, After incorporation of a DNA the intermediate vector is introduced into A. tumefaciens Herrera-Estrella (19832, Bevan (1983), Klee (1985) and EPO publication 120,316 (Schilperoort et al.). In addition to plant transformation vectors derived from the Ti or root-inducing (Ri) plasmids) of Agrobacterium, alternative methods can be used to insert the DNA constructs of this invention into plant cells. Such methods may involve, for example, the use of liposomes, electroporation, chemicals that increase free DNA_uptake, free DNA delivery via microprojectile bombardment, and transformation using viruses or pollen.

A particularly useful Ti plasmid cassette vector for transformation of dicotyledonous plants is shown in Figure 5. Referring to Figure 5, the expression cassette pMON893 consists of the enhanced CaMV35S promoter (EN 35S) and the 3' end including polyadenylation signals from a soybean gene encoding the alpha—prime subunit of beta-conglycinin. Between these two elements is a multilinker containing multiple restriction sites for the insertion of genes.

The enhanced CaMV35S promoter was constructed as A fragment of the CaMV35S promoter extending and +9 was previously (1985). This follows. between position -343 constructed in pUC13 by Odell et al. segment contains a region identified by Odell et al. (1985) as being necessary for maximal expression of It was excised as a ClaI~ with DNA the CaMV3SS promoter.

HindIII fragment, made blunt ended polymerase I (Klenow fragment) and inserted into the HincII site of pUC18. This upstream region of the 35S promoter was excised from this plasmid as a HindIII— EcoRV fragment (‘extending from -343 to -90) and inserted into the'sameH plasmid between the HindIII» and The enhanced CaMV35S promoter thus -343 PstI sites. contains a duplication of sequences between and -90 (Kay et al., 1987).

The 3' end of the 7S_ gene is derived from the 78 genecontained on the clone designated 17.1 (Schuler et a1., 1982). This 3' end fragment, which includes the polyadenylation signals, extends from an AvaII site located about 30 bp upstream of the termination codon for the beta~cong1ycinin gene in clone 17.1 to an EcoRI site located about 450 bp downstream of this termination codon.

The remainder of pMON893 contains a segment of pBR322 which provides an origin of replication in E. coli and a region for homologous recombination with strain ACO the disarmed 'I‘—DNA in Agrobacterium (described below); the oriv region from the broad host range plasmid RK1; the streptomycin/spectinomycin resistance gene from Tn7; and a chimeric NPTII gene, containing the CaMV35S promoter and the nopaline synthase (NOS) 3' end, which provides kanamycin resistance in transformed plant cells.

Referring to Figure 6, transformation vector plasmid pMON90O is a derivative of pMON893. The enhanced CaMV35S promoter of pMON893 has been replaced with the 1.5kb mannopine synthase (MAS) promoter (Ve1_ten et al. 1984). The other segments are the same asgplasmid pMON893. construct into plasmid vector pMQN893 or pMON900, the intermediate vector is ‘introduced into A. tumefaciens After incorporation of a DNA strain ACO which contains a disarmed Ti plasmid.

Cointegrate Ti plasmid vectors are selected and used to transform dicotyledonous plants.

Referring to Figure 7, A. tumefaciens ACO is a disarmed strain similar to pTiB6SE described by Fraley For construction of ACO the starting strain was the strain A208 which et al. (1985).

Agrobacterium contains a nopaline-type Ti plasmid. ‘The Ti plasmid was disarmed in a manner similar to that described by Fraley et al. (1985) so that essentially all of the native T~DNA was removed except for the left border and a few hundred base pairs of T—DNA inside the left border. The remainder of the T-DNA extending to a point just beyond the right border was replaced with a novel piece of DNA including (from left to right) a the oriV region from plasmid RK2, segment of pBR322, The and the kanamycin resistance gene from Tn601. pBR322 and oriV segments are similar to the segments in pMON893 and provide a region of ‘homology for cointegrate formation.

The following examples elucidate the practice of the present invention and are provided to better should not be interpreted in any way to limit the Those skilled in the scope of the present invention. modifications, art will recognize that various truncations etc. can be made to the methods and genes described herein while not departing from the spirit and scope of the present invention. -38_ Referring to Figure 2, the wild-type B.t.k.lfiD-1 gene is known to be expressed poorlf in plants as a The G+C full length gene or as a truncated gene. containing content of the B.t.k. gene is low (37%) many A+T rich regions, potential polyadenylation sites (18 sites; see"Table II for the list of sequences) and numerous ATTTA sequences.

Table II I. E S I . 1 E I I J . . J AATAAA* AAGCAT AATAAT* ATTAAT AACCAA ATACAT ATATAA AAAATA' AATCAA ATTAAA** ATACTA AATTAA** ATAAAA AATACA** ATGAAA CATAAA** * indicates a potential major plant polyadenylation site. ** indicates a potential minor animal polyadenylation site.

All others are potential minor plant polyadenylation sites.

Table III lists the synthetic oligonucleotides designed and syntheeized for the site~directed mutagenesis of the B.t.k. HD—1 gene.

Table III Erimen Length_1h2L sequence BTKIBS 18 TCCCCAGATA ATATCAAC BTK24O 48 GGCTTGATTC CTAGCGAACT CTTCGATTCT CTGGTTGATG AGCTGTTC BTK462 54 CAAAACTGAG AGGTGGAGGT TGGCAGCTTG AACGTACACG GAGAGGAGAGGAAC BTK669 48 AGTTAGTGTA AGCTCTCTTC TGAACTGGTT GTACCTGATC CAATCTCT BTK930 39 AGCCATGATC TGGTGACCGG ACCAGTAGTA TTCTCCTCT BTK1110 32 AGTTGTTGGT TGTTGATcc¢ GATGTTAAAA GG _40_ Table III - continued Erimer LengIh_Jh2i sequence GTGATGAAGG GATGATGTTG TTGAACTCAG CACTACG BTK1380A 37 CAGAAGTTCC AGAGCCAAGA TTAGTAGACT TGGTGAGTGG GATTTGGGTG ATTTGTGATG AAGGGATGAT GTTGTTGAAC TCAGCACTAC GATGTATCCA BTK1380T 100 TGATGTGTGG AACTGAAGGT TTGTGGT BTKl600 27 The B.t.k. HD—1 gene (Bglll fragment from pMON992l encoding amino acids 29~607 with a Met-Ala at the N- terminus) was cloned into pMON7258 (pUC118 derivative which contains a BglII site in the multilinker cloning region) at the B9111 site resulting in pMON5342. The orientation of the B.t.k. gene was chosen so that the opposite strand (negative strand) was synthesized in filamentous phage particles for the mutagenesis. The of Kunkle (1985) was used for _the using plasmid pMON5342 as i procedure mutagenesis stafiting material.

The regions for mutagenesis were selected in the following manner. All regions of the DNA sequence of the B.t.k. gene were identified which contained five or more consecutive base pairs which were A or T.

These were ranked in terms of length and highest percentage of A+T in the surrounding sequence over a ¥30 base‘pair region. The DNA was then analysed for regions which might contain polyadenylation sites (see Table II above) or ATTTA sequences. Oligonucleotides were designed which maximized the elimination of A+T which contained one or more or ATTTA consecutive regions polyadenylation sites potential plant polyadenylation sites were rated more critical (see Table II) based on published reports. did sequences . TWO Codons were selected which increased G+C content, not generate restriction sites for enzymes useful for cloning and assembly of the modified gene (BamHI, BglII, SacI, Ncol, EcoRV) and did not contain the reported to be plants. The doublets TA or GC which have been infrequently found in codons in oligonucleotides were at least 18 bp long ranging up to 100 base pairs and contained at least 5-8 base pairs of direct homology to native sequences at the ends of the fragments for efficient hybridization and priming in site-directed mutagenesis Figure 2 compares the ‘wild-type B.t.k. sequence with the sequence which resulted from the HD—1 gene modifications by site-directed mutagenesis.

The end result of these changes was to increase the G+C content of B.t.k. gene from 37% to 41% while also decreasing the potential plant polyadenylation sites from 18 to 7 and decreasing the ATTTA regions from 13 to 7. Specifically, the mutagenesis changes from reactions. amino (5') terminus to the carboxy (3') terminus are as follows: T BTK185 is an 18-mer used to eliminate a plant ppolyadenylation site in the midst of a nine base pair region of A+T.

BTK240 is a 48-mer. by this oligonucleotide to eliminate three potential polyadenylation sites (2 AACCAA, 1 AATTAA). Another region close to the region altered by BTK240, starting at bp’ 312, had a high A+'l‘ content (13 of 15 base pairs) and an ATTTA region. _ However, it did not contain a potential polyadenylation site and its longest string of uninterrupted A+T was seven base .Seven base pairs were changed pairs.

BTK462 is a The first six changes were to reduce the A+T —mer introducing 13 base pair changes. richness of the gene by replacing wild—type codons with codons containing G and C while avoiding the CG doublet. .The next seven changes made by BTK462 were used to eliminate an A+T rich region (13 of 14 base pairs were A or T) containing two ATTTA regions.

BTK669 is a 48—mer making nine individual base pair changes eliminating three possible polyadenylation sites (ATATAA, AATCAA, and AATTAA) and a single ATTTA site.

BTK930 is a 39—mer designed to increase the G+C content and to eliminate a potential polyadenylation site.(AATAAT — a major site). This region did contain a nine base pair region of consecutive A+T sequence.

One of the base pair changes was a G to A because a G at this position would have created a G+C rich region (CCGG(G)C). Since'seguencing reactions indicate that there can be difficulties generating sequence through G+C consecutive bases, it was thought to be prudent to avoid generating potentially problematic regions even if they were problematic only in vitro.

BTK11-10 is a 32—mer- designed to introduce five changes in the wild—type gene.

(AATAAT - a major site) was eliminated in the midst of an A+T rich region (19 of 22 base pairs).

BTK1380A and BTK1380T are responsible for 14 individual base pair changes. The first region (1380A) has 17 consecutive A+T base pairs. In this region is an ATTTA and a potential polyadenylatimu (1380T) The large size of this contains all the site (AATAAT). The 100-mer changes dictated by 1380A. primer was in-part an experiment to determine if it was feasible to utilize large oligonucleotides for (over 60 bases in length). A second was that the 100—mer which had previously been The original primer ordered to mutagenesis was used to consideration mutagenize a template mutageneized by 1380A. mutagenize the region downstream and adjacent to 1380A did not anneal efficiently to the desired site as indicated by an inability to obtain clean sequence utilizing the primer. The large region of homology of 1380T did assure proper annealing. The extended size of 138OT was more of a: convenience rather than a The second region adjacent to 1380A (22 of 29 necessity. covered by 1380T has a high A+T content bases are A or T).

One potential site’ ,VIII) in tobacco.

BTK16OO is a 27-mer responsible for five individual An ATTTA. region and 21 plant identified and —the base fpair changes. site were polyadenylation "appropriate changes engineered.

A total of 62 bases were changed by site-directed mutagenesis. The G+C content increased by 55 base pairs, the potential polyadenylation reduced front 18 to seven and the ATTTA sequences The changes in the DNA sites were decreased from 13 to seven. sequence resulted in changes in 55 of the 579 codons B. t-k. gene in pMON5342 in the truncated (approximately 9.5%).

Referring to Table IV modified B.t.k. contained all of the above subsets of HD~1 genes were constructed that (pMON5370) or various modifications individual modifications. These genes were inserted intc> pMON893 for plant transformation plants containing these genes were analyzed. with the individual and tobacco of ‘tobacco plants undertaken for analysis several reasons. modifications was Expression of the wild type truncated gene in tobacco is very poor, resulting in infrequent identification of plants toxic to THW. feeding assays as at least 60% mortality of tobacco hornworm neonate larvae with a damage rating of 1 or Toxicity is defined by leaf less (scale is (J to 4; O is equivalent to total protection, 4 total damage). The modified HD—1 gene (pMON5370) shows a large increase in expression (estimated to be approximately 100-fold; see Table Therefore, increases in expression of the wild-type gene due to indidvidual modifications would he apparently atlarge increase in the frequency of toxic tobacco plants and the presence of detectable B.t.k. protein. Results are shown in the following -table: Table IV Relative effects of Reqional Modifications xithin the B.t.k. Gene E ti! Ivfiv 1 I ‘E K E T 0 Canztznnn pMON5370 185,24o,669,930, ll0,l380a+b,1600 38 22 pMONl0707 l85,240,462,669 48 19 pMONlO706 930,1110,138oa+b,16oo 43 1 pMONlOS39 185 55 2 pMONl0537 240 57 17 pMON10540 185,240 88 23 pMON10705 462 . 47 1 The effects of each individual oligonucleotides' changes on expression did reveal some overall trends. ‘half were incorporated into pMON10707 designed. to identify‘ the key regions. The nine different oligonucleotides were divided in half by their position on the gene. Changes in the N-terminal (185g240; C-terminal half changes were incorporated The results ,669). into pMON10706 (930, 1110, l380a+b, 1600). of analysis of plants with these two constructs indicate that pMON10707 produces a substantial number Protein from these plants pMON10706 plants of toxic plants (19 of 48)- is detectable by ELISA analysis. were rarely identified as insecticidal (1 of 43) and the levels of B.t.kp were barely’ detectable by immunological analysis. Investigation of the N- terminal changes in greater detail was done with 4 pMON constructs; 10539 (185 alone), 10537 (240 alone); 10540 (185 and 240? and 10705 (462 alone). The results indicate that the presence of the changes in 240 were required to generate a substantial number of toxic plants (pMONl0S40; 23 of 88, pMONlO537; 17 of ). The absence of the 240 changes resulted in a low frequency of toxic plants with low B.t-k. protein levels, identical to results with the wild type gene.

These results indicate that the changes in 240 are in B.t.k. wild~type responsible for a substantial increase expression levels over an analogous construct in tobacco. Changes in additional regions (l8S,462,669) in conjunction with 240 may result in increases in B.t.k. expression (>2 fold). However, .changes at the 240 region of the N—terminal portion of the gene do result in dramatic increases in expression.

Despite the importance of the alteration of the 240 region ix: expression of modified genes, increased expression can be achieved by alteration of other regions. Hybrid genes, part wild—type, part (synthetic, were generated to determine the effects of the levels of B.t.k. synthetic gene segments on expression. A hybrid gene was generated with a synthetic N—terminal third (base pair 1 to 590 of Figure 2: to the XbaI site) with the C—terminal wild type B.t.k. HD~1 (pMON5378) Plants transformed with this vector were as toxic as plants transformed with the modified HD—1 gene (pMON5370). This is consistent with the alteration off the 240 region- pMON10538, a hybrid with a wild—type N-terminal third (wild type gene for the first 600 base pairs, to the However, second Xbal site) and a synthetic C-terminal last two~ thirds (base pair 590 to 1845 of Figure 3 was used to transform tobacco and resulted in a dramatic increase in expression. The levels of expression do not appear to be as high as those seen with the synthetic gene, but are comparable to the modified gene levels. These results indicate that modification of the 240 segment expression since A fully synthetic is not essential to increased pMONl0538 has an intact 240 region. gene is, in most cases, superior for expression levels of B.t.k.‘ (See Example 2.) Example,2 —~ Fullv Svnthetic B.t.k. HD—1 Gene A synthetic B.t.k. HD—l gene was designed using the preferred plant codons listed in Table V below.

Table V lists the codpns and frequency of use in plant genes of dicotyledonous —p1ants compared to .the frequency of their use in the wild type B.t.k. HD-1 igene (amino acids 1-615) and the synthetic gene of this example. The total number of each amino acid in this segment of the gene is listed in the parenthesis under the amino acid designated.

E . E .1 : ! ARG (43) LEU (49) SER (64) Codon in Usaae Svnthetic B:t.k, HD:1 Gene CGA CGC CGG , CGU AGA AGG CUR CUC CUG CUU UUA UUG UCA UCC UCG UCU AGC AGU -49_ Table V Percent Usage in Elants/Wt B.t.k,/Svn 50 ' 32 THR (42) PRO (34) (31) GLY (46) <46)‘ Codon in Usaqe Svnthetic B.t.k. HD-1 Gene ACA ACC ACG ACU CCA CCC CCG CCU GCA GCC GCG GCU GGA GGC GGG GGU AUA AUC AUU Tab%e V — continued Percent Usage in 3 11 45 19 14 ll S3 0 12 3 Table V — continued Codon in Usaqe Synthetic B.t.k. HD—1 Gene Percent Usage in Plants/Wt B.t.k./Svn E07 E01 VAL GUA .9 4 5 3 (38) GUC 20 5 16 GUG 28 11 37 GUU 43 39 45 LYS AAA 36 100 33 (3) AAG 64 o 67 ASN AAC 72 27 80 (44) AAU 28 73 20 GLN CAA 64 77 61 <31) CAG 36 23 39 HIS CAC 65 0 so (10) can 35 100 20 GLU GAA 48 37 50 (30) GAG 52 '13 50 ASP GAC 48 17 55 (23) GAU 52 33 Table V — Continued thetic B.t.k. my-1 Gene Percent Usage in AminQ_Agid canon Rlants/Wt B.t.k./Svn TYR uac 68 20 72 <25) UAU . 32 so 28 cys UGC 78 so 100 (2) UGU 22 so 0 par uuc 56 17 83 (36) UUU 44 83 17 MET AUG 100 100 100' (9) TR? UGG 100 100 100 (W The resulting synthetic gene lacks ATTTA sequences, contains only one potential polyadenylation site and has a G+C content of 48.5%. Figure 3 is a comparison of the wild—type HD~1 sequence to the synthetic gene seguenee for amino 1-615. There is approximately 77% DNA homology between the synthetic acids gene and the wi1d;type gene and 356 of the 615 codons have been changed (approximately 60%).

Example 3' -~ Svnthetic B.t:.k. HD-73 Gene ‘The crystal protein toxin from B.t.ku HD-73 exhibits.a higher unit activity against some important The toxin protein of HD-1 and HD- in the N- agricultural pests. 73 exhibit substantial homology (~90%) terminal 450 amino acids, but differ substantially in the amino acid region 451—6l5. Fusion proteins comprising amino acids 1-450 of HD—1 and 451-615 of HD- 73 exhibit the insecticidal properties of the wild~ type HD—73. The strategy employed was to use the 5': two thirds of the synthetic HD~l gene (first 1350 bases, up to the SacI site) and to dramatically modify the final 590 bases (through amino acid 645) of the HD— in a manner consistent with the algorithm used to design the synthetic HD—l gene. Table VI below lists the oligonucleotides used to modify the HD—73 gene in the order used in the gene from S‘ to 3' end. Nine oligonucleotides were used in a 590 base pair region, each nucleotide ranging in size from 33 to 60 bases.

The only regions left unchanged were areas where there were no long consecutive strings of A or T bases (longer than six). All polyadenylation sites and ATTTA sites were eliminated.

Erimsz K1363 Kl437 K1471 K1561 K1642 K1675 K1741 Table VI Mntaggngsjs Ezimgns fgz 5,5,3, up-73 Icngmh_ihn; Scgncnnc AATACTATCG TGTTGTTGAA CGGTGTATCC TCCTGAAATG TGAAGAGAAA ATTTCCACTG TAACGAGGTC AATCCTGG GTGAATAGGG CATACCTCAC TATCTGGTAG TGTAGCTGGA AGAAGATGGA TTCAAAGTAA TGGATTGGAG GAGGTAGC ACTAAAGTTT CGATGTTACC GATGCGATGA CTCAGCACTA A ACAGAACCGT GTT CTGTTGAGTC TCCACCAGTG GTCACAGAAG ACGAACTCTA ATGTTGGATGG ACTGTATTGG TGA CCGAAATCGC ATTATCCAAG CTAACACCCA GAGTGAAGA Table VI - continued Mntagenesjs Ezjmgps 1:91; a,;;,g, 1112-73 Rfimfl Kl797 36 AACTGGAATG AACTCGAATC.

TGTCGATAAT CACTCC KTERM 54 GGACACTAGA TCTTAGTGAT AATCGGTCAC-ATTTGTCTTG AGTCCAAGCT GGTT The resulting gene has two potential polyadenylation sites (compared to 18 in the WT) and no ATTTA sequence (12 in the WT). The G+C content has increased from 37% to 48%. A total of 59 individual base pair Changes were made using the primers in Table VI. Overall, there is 90% DNA homology between the region of the HD—73 gene_modified by site directed mutagenesis and the wild—type sequence of the analogous region of HD—73. The synthetic HD~73 is a hybrid of the first 1360 bases from the synthetic HD—1 and the next 590 bases or so modified HD-73 sequence.

Figure 4 is a comparison of the above~described synthetic B.t.k. HD-73 and the wild-type B.t.k. HD—73 encoding amino acids 1-645. In the modified region of the HD—73 gene 44 of the 170 codons (25%) were changed as a result of the site-directed mutagenesis changes resulting from the oligonucleotides found in Table VI.

Overall, approximately 50% of" the codons in the synthetic .B.t.ku HD:73 differ from the analogous segment of the wild—type and HD—73 gene.

A one base pair deletion in the synthetic HD—73 Vgene was detected in the course of sequencing the 37 end at base pair 1890. This results in a frame—shift mutation at amino acid 625 with a premature stop codon at amino acid 640 (pMON5379). Table VII below compares the codon usage of the wild—type gene of‘B.t.k. HD—73 versus the synthetic gene of this example for amino acids 451-645 and codon usage of naturally occurring genes of dicotyledonous plants. The total number of each amino acid encoded in this segment of the gene is found in the parentheses under the amino acid designation.

Table VII codon Usaqe in Synthetic B.t.k. HD—73 Gene Percent Usage in Blants/Wt HD—73/Svn ARG CGA 7 10 0 (10) CGC 11 0 CGG 5 10 0 CGU 25 20 23 AGA 29 60 62 AGG 23 0 8 Table VII — continued E I H _. 5 1 . 2 E "[_73 3 Percent Usage in LEU CUA 8 25 8 (12) cuc 20 17 53 cps 10 17 8 CUU 23 8 UUA 5 33 UUG 30 0 17 syn uca 14 24 18 (21) ucc 26 10 27 ucs 3 10 0 ucu 21 24 18 AGC 21 0 14 AGU 15 33 23 THR ACA 21 47 38 (15) ACC 41 13 31 ACG 7 13 0 ACU 31 27 31 PRO CCA 45 71 71 (7) ccc 19 0 cce 9 14 ccu 26 14 29 Tablg VII - continued Codon Usaae in Svnthetic B.t.k. HD—73 Gene Percent Usage in ‘ Cgdsm Elantszmz H12—Z3LSzz1 ALA GCA 23 29 31 (14) GCC 32 7 8 GCG 3 21 15 GCU 41 43 GLY GGA 32 33 43 (15) GGC 20 0 0 GGG 11 27 14 GGU 37 40 ILE AUA 12 33 7 (15) AUG 45 7 40 AUU 43 so 53 VAL GUA 9 40 7 (15) GUC 20 0 7 GU_G 28 20 36 GUU 43 40 50 LYS AAA 36 67 100 (3) .AAC; 64 - 33 0 ASN AM: 72 20 53 (20) AAU 28 80 47 3o GLN (5) HIS ' TYR (5) CYS (0) (13) TRP (2) CAR CAG CAC CAU GAA GAG GAC GAU UAC UAU UGC UGU Table VII — continued 4 Codon Usaae in Svnthetic B.t.k, HD—73 Gene Percent Usage in Emma P_1.am:s_ -7 SO SO _5o_ truncated synthetic HD—73 gene was constructed. The sequence of this synthetic HD-73 gene is identical to that of the above synthetic HD—73 Another ’gene in the region in which they overlap (amino acids -615), ‘and it also encodes Met-Ala at the N- terminus. Figure 8 shows "a comparison of this truncated synthetic HD-73 gene with the N-terminal Met- Ala versus the wild—type HD-73 gene.

While the previous examples have been directed at the preparation of synthetic and modified genes encoding" truncated B.t.k. proteins, synthetic or modified genes can also be prepared which encode full length toxin proteins.

One full length .B.t.k. gene consists of the synthetic HD—73 sequence of Figure 4 from nucleotide 1= 1845 plus wild—type HD—73 sequence encoding amino acids 616 to the C—terminus of the native protein.

Figure 9 shows a comparison of this synthetic/wild- type full length HD—73 gene versus the wild—type full length HD—73 gene.

Another full length B.t.k. gene consists of the synthetic HD~73_sequence of Figure 4 from nucleotide 1— 1845 plus a modified HD—73 sequence ending amino acids 616 to the C-terminus of the native protein. The C- terminal portion has been modified by site-directed mutagenesis to remove putative polyadenylation signals and ATTTA sequences according to the algorithm of Figure 1. Figure 10 shows a comparison of this synthetic/modified full length HD-73 gene versus the wild—type full length HD-73 gene.

Another full length B.t,k. gene consists of a fully synthetic HD-73‘ sequences which incorporates the synthetic HD—73 sequence of Figure 4 from nucleotide 1- .1845 plus a synthetic sequence encoding amino acids to the C-terminus of the native protein. The C- terminal synthetic portion has been designed to eliminate putative polyadenylation signals and ATTTA sequences and to include plant preferred codons.

Figure 11 shows a comparison of this fully synthetic full. length HD~73 gene versus the wild-type full length HD~73 gene.

Alternatively, consists of a fully synthetic sequence comprising base (Figure 3) and base pairs another full length B.t.kp gene pairs 1-1830 of B.t.k. HD-1 1834-3534 of B.t_k_ HD-73 (Figure 11).

Example 4 -~ Expression of Modified and Svnthetic Bi£.k. HD—1 and Svnthetic HDs73 A number of plant transformation vectors for the expression of B.t.k. constructed by incorporating the structural coding sequences of the genes were previously described genes into plant transformation cassette vector pMON893. The respective intermediate transformation vector is inserted into a suitable disarmed Agrobacterium vector such as A.

Tissue explants are cocultured with the tumefaciens ACO, supra. disarmed Agrobacterium vector and plants regenerated under selection for kanamycin resistance using known ‘protocols: tobacco (Horsch et al., 1985); tomato (McCormick et al., 1986) and cotton (Trolinder et al., ). ‘a) Tobacco.

The level of B.t.k. HD-1 protein in transmudc tobacco plants containing pMON9921 (wild type pMON5370 (modified HD-1, Example 1, Figure truncated), Figure 3) ) and pMON5377 (synthetic HD—1, Example 2, were analyzed by Western analysis. Leaf tissue was frozen in liquid nitrogen, ground to a fine powder and (wt:volume) of SDS—PAGE sample then ground in a 1:2 frozen on dry ice, then buffer. Samples were incubated for 10 minutes in a boiling water bath and microfuged for 10 minutes. The protein concentration was determined by the method of Fifty ug of of the supernatant Bradford (Anal. Biochem. 72:248—254). protein was run per lane on 9% SDS—PAGE gels, the protein transferred to nitrocellulose and the B.t_k.

HD~l protein visualized using" antibodies produced against B.t.k. HD—1 protein as the primary antibody and alkaline phosphatase conjugated second antibody as described by the manufacturer (Promega, Madison, WI).

Purified ’HD—l tryptic fragment was used as the control. Whereas the B.t.k. protein from tobacco plants containing pMON9921 was below the level of detection, the B.t.k. protein from plants containing the modified (pMON5370) and synthetic (pMON5377) genes was easily detected. The B.t.k. protein from plants containing pMON9921 remained undetectable, even with fold longer incubation times. of B.t.k. HD-1 protein in these plants is estimated in The relative levels ‘relative mRNA levels (see below).

Table VIII. Because the protein from plants containing pMON9921 was not observed, the level of estimated from, the protein in these plants was Plants containing the modified gene (pMON5370) expressed approximately fold more B.t.k. protein than plants containing the wild-type gene (pMON9921). Plants containing the fully synthetic B.t.k. HD—1 gene (pMONS377) expressed approximately five fold xmore protein than plants containing" the modified gene. The modified gene contributes the majority of the increase in B.t.k. expression observed. The plants used to gknerate the above data are tflua best representatives front each construct based either on a tobacco hornworm bioassay or on data derived from previous Western analysis.

Table VIII Expression of B.c.k. HD—l Protein . . 1 Fold Increase Gene B.c.k. Protein‘ in B-t.k.

Dnznnhufinn ybntnr Canmmnaatinn Exnnmuunn Wild type pMON992l 10 1 Modified pMON537O 1000 100 Synthetic pMON$377 S000 500’ * B.t.k. protein concentrations are expressed in ng/mg of total soluble protein. The level of B.t.k. protein for plants containing the wild type gene are estimated from mRNA levels.

Plants containing these genes were tested for bioactivity to determine whether the increased quantities of protein observed by Western analysis result inua corresponding increase in bioactivity.

Leaves from the same plants used for the Western data in Table l were tested for bioactivity against two A detached leaf bioassay was first done extremely sensitive insects. tobacco hornworm, an Beaves from all three transgenic using lepidopteran insect. totally protected and 100% (see Table IX tobacco plants were mortality of tobacco hornworm observed A much less sensitive insect, beet armyworm, Beet below). was then used in another detached leaf bioassay. armyworm is approximately 500 fold less sensitive to B.t-k. HD-1 protein than tobacco hornworm. The difference in sensitivity of these two insects was determined using purified HD-1 protein in a diet incorporation assay (see below)- Plants containing (pMON992l) showed only minimal whereas plants the wi1d—type gene protection against beet armyworm, containing the modified gene showed almost complete protection and plants containing the fully synthetic gene were totally protected against beet armyworm damage. The results of these bioassays confirm the levels of B.t.k.

Western analysis and demonstrates that the increased levels of .B.t.k. HD-1 protein correlates with increased insecticidal activity.

HD-1 expression observed in the ~55- Table IX Erotection of Tobacco Plants from Gene Tobacco Hornworm Beet Armyworm Deacrintian SESLQK Damaqe* ______ Damaae* None None NL NL Wild type pMON9921 0 3 Modified pMQN5370 0 Synthetic pMON5377 O O * Extent of insect damage was rated: 0, no damage; , slight; 2, moderate; 3, severe; or NL, no leaf left.

The bioactivity of the B.t.k_ HD-1 protein produced by these transgenic plants was further investigated to more accurately quantitate the relative activities.

Leaf tissue from tobacco plants containing the wild~ type, modified and synthetic genes were ground in 100 mM sodium carbonate buffer, pH 10 at a 1:2 (wtzvol) Particulate material was removed by centrifugation. The supernatant was~incorporated into a synthetic diet similar to that described by Marrone et al. (1985). The diet medium was prepared the day of the test with the solutions incorporated-in place of the 20% water component. One ml of the diet was aliquoted into 96 well plates. ratio- plant extract After the diet dried, one neonate tobacco budworm larva was added to each well. Sixteen insects were tested with each plant sample. incubated at 27°C. After seven days, the larvae from each treatment were combined zuui weighed (fll an The-average weight per insect was The plants were analytical balance. calculated and compared to a standard curve relating .B.t.k. protein concentrations to average larval Insect weight was inversely proportional (in weight. to the relative increase in a logarithmic manner) B.t.k. protein concentration. The amount of B.t.k. HD— protein, based on the extent of larval growth inhibition was determined for two different plants containing each of the three genes. The specific (ng of B.t.k. HD—1 per mg of plant protein) activity tPlants containing the was determined for each plant. modified HD-1 gene (pMON5370) (1200 and 1600 ng) of B_t.k. HD—l per mg of This value compares closely averaged approximately ng plant extract protein. with the 1000 ng of B.t.k. plant extract protein as B.t.k. HD—1 concentrations for HD-1 protein per mg of determined by Western analysis (Table I). the plants containing the synthetic HD-1 gene averaged approximately 8200 ng (7200 and 9200 ng) of B.t.k. HD— protein per mg of plant extract protein. This number compares well to the 5000 ng of HD-1 protein per mg of plant extract protein estimated by Western analysis. Likewise, plants containing the synthetic gene showed approximately a six~fold higher specific activity than the corresponding plants containing the modified gene for these bioassays. In the Western A67- analysis the ratio was approximately 10 fold, again both are in good agreement. The level of B.t.k. protein in plants containing the wi1d~type HD~1 gene (pMON9921) was too low to give a significant decrease in larval weight and hence was below a level that could be quantitated in this assay. In conclusion, the levels of B.t.k. HD-1 protein determined by both the bioassays and the Western analysis for these plants containing the modified and synthetic genes agree, which demonstrates that the B.t.k. HD-1 protein produced by these plants is biologically active- The levels of mRNA were determined in the plants containing the wild—type B.t-k- HD—l gene (pMON9921) and the modified gene (pMON5370) to establish whether the increased_levels of protein production result from increased transcription or translation- mRNA from plants containing the synthetic gene could not be analyzed directly with the same DNA probe as used for because of the mRNA the wild—type and modified ~genes numerous changes made in the coding sequence. was isolated and hybridized with a single-stranded probe homologous to approximately the 5' 90 bp of the wild~type or modified gene coding sequences. The were digested with S1 nuclease and the hybrids analyzed by gel protected electrophoresis. excess cnf probe and long hybridization time, amount of protected probe is proportional to the amount of B.t.k. mRNA present in the sample. Two plants expressing the modified gene (pMON5370) were fragments Because the procedure used a large the probe found to produce up to ten-fold more RNA than a plant expressing the wiidvtype gene (pMON9921).

The increased mRNA level from the modified gene is consistent with the expected from the modifications introduced into this gene. this 10 fold increase in mRNA with the modified gene compared to the wild—type gene is in contrast to the 100 fold increase in B.t.k. protein from these genes in tobacco plants. If the two mRNAs were equally well translated then a 10 fold increase in stable mRNA result However, would be expected to yield a 10 fold increase in protein. The higher increase in protein indicates that the modified gene mRNA is translated at about a fold higher efficiency than wild—type. Thus, about half of the total effect on gene expression can be explained by changes in mRNA levels and about half to changes in translational efficiency. This increase in translational efficiency is striking in that only about 9.5% of the codons have been changed in the this effect is clearly not due modified gene; that is, The increased to wholesale codon usage changes. translational efficiency could be due to changes in mRNA secondary structure that affect translation or to the removal of specific translational blockades due to specific codons that were changed.

The increased expression seen with the synthetic 1 gene was also seen with a synthetic HD~73 gene B.t.k. HD—73 was undetected in extracts HD~ in tobacco. of tobacco plants containing the wild-type truncated HD— whereas B.t.k. HD~73 protein was gene (pMON5367), from tobacco plants easily detected in extracts ,69_ containing the synthetic HD—73 gene of Figure 4 (pMON5383). Approximately 1000 ng of 8.t.k. HD~73 .protein was detected per mg of total soluble plant protein.

As described in Example 3 above, protein encoded :h1 pMON5383 contains a terminal extension of amino acids not encoded in the These extra amino acids had the B.t.k. HD-73 small C- wild—type H5-73 protein. no effect on insect toxicity or on increased plant synthetic HD—73 gene was and expression. A second constructed as described in Example 3 (Figure 8) used to transform tobacco (pMON5390). Analysis of plants containing pMON539O showed that this gene was expressed at levels comparable to that of pMON5383 and that these plants had similar insecticidal efficacy.

In tobacco plants the synthetic HD—l gene was expressed at approximately a 5-fold higher level than the synthetic HD—73 gene. However, this synthetic HD— gene still was expressed at least lCO—fold better than the wild—type HD—73 gene. The HD—73 protein is approximately 5~fold more toxic to many insect pests than the HD—1 protein, so both synthetic HD~1 and HD» genes provide approximately comparable insecticidal efficacy in tobacco.

The full length B.t.k. HD~73 genes described in also incorporated into the plant vector pMON893 so that they were expressed from the En 35S synthetic/wild-type full length HD-73 gene of Figure 9 was incorporated into pMON893 to create pMON10505.

The synthetic/modified full length HD-73 gene of Example 3 'were transformation promoter. The ,70_ Figure 10 was incorporated into pMON893 to create pMON10526. The fully synthetic HD—73 gene of Figure 11 was incorporated into pMON893 to create pMON10518.

These vectors were used to obtain transformed tobacco and the plants were analyzed for insecticidal plants, HD—73 protein levels by efficacy and for B.t.k.

Western blot or ELISA immunoassay.

Tobacco plants containing all three of these full length B.t.k. genes produced detectable B.t.k‘ protein and showed 100% mortality of tobacco hornworm. This result is surprising in light of previous reported attempts to express the full length B.t.k. genes in transgenic plants. Vaeck et al. (1987) reported that a full length B,t.k. berliner gene similar to our HD~1 gene could not be detectably expressed in tobacco.

Barton et al. (1987) reported a similar result for another full length gene from B-t.k. HD-1 (the so called. 4.5 kb gene), and further indicated that tobacco callus containing this gene became necrotic, indicating that the full length gene product was toxic Fischhoff et al. (1987) reported that to plant cells.

HD—1 gene in tomato was poorly the full length B.t.k- expressed compared to a truncated gene, and no plants that were fully toxic to tobacco hornworm could be recovered. All three of the above reports indicated much higher expression levels and recovery of toxic plants if the respective B.t.k. genes were truncated.

Adang et al. reported that the full length HD~73 gene yielded a few tobacco plants with some biological activity (none were highly toxic) against hornworm and barely detectable B.t.k. protein. It was also noted ‘functional toxin. -7]- by them that the major B.t.k. mRNA in these plants was a truncated 1.7 kb species that would not encode a This indicated improper expression of the gene in tobacco. In contrast to all of these reports, the three full length B.t.k. HD—73 genes described above all lead to relatively high levels of protein and high levels of insect toxicity.

B.t.k. protein and mRNA levels in tobacco plants are shown in Table X for these three vectors. As can be seen from the table, the synthetic/wild-type gene (pMON10506) produces B.t-k. protein as about 0.01% of the synthetic/modified gene 0.02% soluble total soluble protein; B,t.k_ as about of total and the fully synthetic gene produces B-t.k.

B.t.k. mRNA produces protein; as about 0.2% of total soluble protein. was analyzed in these plants by Northern blot analysis synthetic half of the genes as a using the common 5' the increased protein pro‘3. As shown in Table X, levels can largely be attributed to increased mRNA Compared to the truncated modified and levels. this could indicate that the major synthetic genes, contributors to increased translational efficiency are in the 5' half of the gene while the 3’ half of the gene contains mostly determinants of mRNA stability.

The increased protein levels also indicate that increasing the amount of the full length gene that is synthetic or.modified increases B.t.k. protein levels.

Compared to the truncated synthetic B.t.k. HD—73 genes (pMON5383 or pMON5390), the fully synthetic gene (pMON10S18) produces as much or slightly more B.t.k. protein demonstrating that the full length genes are capable of being expressed at high levels in plants.

These tobacco plants with high levels of full length HD—73 protein show no evidence of abnormality and are fully fertile. The B.t.k; protein levels in these plants also produce the expected levels of insect toxicity based on feeding studies with beet armyworm or diet incorporation assays of plant extracts with tobacco budworm. The B.t.k. protein detected by Western blot analysis in these tobacco plants often contains a varying amount of protein of about 80 kDa which is apparently a proteolytic fragment of the full The C—terminal half of the full length protein. proteolytically length protein is sensitive, and similar proteolytic fragments are seen coli and B-t. itself.

The Northern known to be from the full length gene in E.

These fragments are fully insecticidal. analysis indicated that essentially all of the mRNA from these full length genes was of the expected full length size. There is no evidence of truncated mRNAs that could give rise to the 80 kDa protein fragment. it is possible that the fragment is not and is merely due to In addition, present in intact plant cells proteolysis during extraction for immunoassay.

Table X Full Length B.t.k. HD-73 Protein and MEI III .I] Ell B.t.k. protein Relative B.t.k.

Gene description Vector concentration mRNA level Synthetic/wild type pMON1050-6 >100 0.5 Synthetic/modified pMON10526 400 1 Fully synthetic pMON10518 >2000 40 Thus, there is no serious impediment to producing high levels of B.t.k- HD—73 protein in plants from synthetic genes, and this is expected to be true of other full length lepidopteran active genes such as B_t.k. HD—1 or B.t. entomocidus. The fully synthetic B_t-k- HD—1 gene of Example 3 has been assembled in plant transformation vectors such as pMON89_3. in pMONl0Sl8 was also The fully synthetic gene in another plant vector Although the CaMV35S promoter is utilized and analyzed in tobacco plants. generally a high level constitutive promoter in most plant tissues, the expression level of genes driven the CaMV35S promoter is low in floral tissue relative to the levels seen in leaf tissue. Because the economically important targets damaged by some insects are the floral parts or derived from floral parts (e.g., cotton squares and bolls, tobacco buds, tomato it may‘ be advantageous to increase protein in these tissues over buds and fruit), the expression of B.t. that obtained with the CaMV35S promoter. [transformation vector analogous to pMON893.

The 35S promoter of Figwort Mosaic Virus (FMV) is analogous to the CaMV35S promoter. This promoter_has been isolated and engineered into a plant Relative to the-CaMV promoter, the FMV 35S promoter is highly expressed in the floral tissue, while still providing similar high levels of gene expression in other tissues such as leaf. A plant transformation vector, pMON10517, was constructed in which the full length synthetic B.t.k. HD~73 gene of Figure 11 was driven by the FMV 35S promoter. This vector is identical to pMONlO518 of Example 3 except that the FMV promoter is substituted for the CaMV promoter. Tobacco plants transformed with pMONl05l7 and pMON10518 were obtained and compared for expression of the B.t.k. protein by Western blot or ELISA immunoassay in leaf and floral tissue. This analysis showed that pMONl05l7 containing the FMV promoter expressed the full length HD—73 protein at higher levels in floral tissue than pMON10518 containing the CaMV promoter. ,Expression of the full length B-t.k- HD—73 protein from pMON10517 in leaf tissue is comparable to that seen with the most expressing plants containing when floral tissue was analyzed, tobacco highly However, plants containing pMON10Sl8 that had high levels of B.t.k. protein in leaf tissue did not have detectable B.t.k. protein in the flowers. On the other hand, flowers of tobacco plants containing pMON10517 had levels of B.t.k. protein nearly as high as the levels in leaves at approximately 0.05% of total soluble protein. This analysis showed that the FMV promoter pMON105l8_ could be used to produce relatively high levels of B.t.k. protein in floral tissue compared to the.CaMV promoter. genes tested in tobacco were introduced into other plants [to demonstrate the broad utility of this Transgenic tomatoes were produced which show that the invention. contain these three genes. Data increased expression observed-with the modified and synthetic gene in tobacco also extends to tomato.

Whereas the B_t:.k. HD—] protein is only barely detectable in plants containing the wild type HD—1 gene-(pMON9921), B.t.k- HD-1 was readily detected and determined for plants containing the genes. the levels modified (pMoN53‘7o) or synthetic (pMONS377) Expression levels for the plants containing the wild- modified and synthetic HD~1" genes were 100 and 500 ng per mg of total plant The increase in B_t.k. tYP€r approximately 10, extract see Table XI below).

HD-1 protein for the modified gene accounted for the majority of increase observed; 10 fold higher than the plants containing the wild—type gene, compared to only an additional five—fold increase for plants containing the synthetic gene. Again the site-directed changes made in the modified gene are the major contributors to the increased expression of B.t.k. HD—1.

Table XI Fold Increase Gene B.t.k. Protein* in B.t.k.

Dnznzinxinn yecnaz Cnncnntnatign Exnnefifiiou Wild type pMON9921 10 1 Modified pMONS370 100 10 Synthetic pMON5377 500 50 * B.t.k. HD-l protein concentrations are expressed in ng/mg of'total soluble plant protein. Data for plants containing the wild—type gene are estimates from mRNA levels and protein levels determined by ELISA.

These differences in B-t.k- HD~1 expression were confirmed with bioassays against tobacco hornworm and beet armyworm. Leaves from tomato plants containing each of these genes controlled tobacco hornworm damage and produced 100% nwrtality. with beet armyworm, leaves from plants containing the wild-type HD—1 gene (pMON9921) showed significant damage, leaves from plants containing the modified gene (pMON5370) showed less damage and leaves from plants containing the synthetic gene (pMON5377) were completely protected (see Table XII below).

Table XII mfl2m Gene Tobacco.Hornworm Beet Armyworm Dnnnnhndan EEQLQL Dnm&Qa£________ Damam:L_____ None None NL NL Wild type pMON9921 0 Modified pMON537O 0 Synthetic pMON5377 O * Damage was rated as shown in Table IX.

The generality of the synthetic gene approach was extended in tomato with a synthetic B.t-k. HD~73 gene.

In tomato, extracts from plants containing the wild- (pMON5367) showed no Extracts from plants type truncated HD-73 gene detectable HD—73_ protein. containing the synthetic HD~73 gene (pMON5383) showed high levels of B.t.k- HD-73 protein, approximately 2000 ng per mg of plant extract protein. These data clearly demonstrate that the changes made in the synthetic HD—73 gene lead to dramatic increases in the expression of the HD—73 protein in tomato as well as in tobacco In contrast to tobacco, the synthetic HD—73 gene in tomato is expressed at approximately 4~fold to 5—fo1d_ higher levels than the synthetic HD—1 gene. Because the HD—73 protein is about 5-fold more active than the ,78, HD—l protein againstv many insect pests including Heliothis ‘species, the increased expression of synthetic HD—73 compared to synthetic HD—1 corresponds to about a 25~fold increased insecticidal efficacy in tomato.

In order to determine the mechanisms involved in the increased expression of modified and synthetic B.t.k. HD—1 genes in tomato, S1 nuclease analysis of mRNA levels from transformed tomato plants was performed. As indicated above, a similar analysis had been performed with tobacco plants, and this analysis showed that the modified gene produced up to 10~fo1d more mRNA than the wild—type gene- The analysis in tomato utilized a different DNA probe that allowed the analysis of wild—type (pMON9921), modified (pMON5370) and synthetic (pMON5377) HD-l genes with the same probe. This probe was derived from the 5' untranslated region of the CaMV3SS promoter in pMON893 that was common to all three of these vectors (pMON992l, pMoN537o and pMON5377). mRNA levels from the modified This 51 analysis indicated that B.t-k. gene were 3 to 5 fold higher than for the wild—type gene, and that mRNA levels for the synthetic gene were about 2 to 3 fold higher than for the modified gene.

Three independent transformants were analyzed for each gene. Compared to the fold increases in B.t.k. HD—1 protein from these genes in tomato shown in Table XI, these mRNA increases can explain about half of the total protein increase as-was seen in tobacco for the wild-type and modified genes. For tomato the total mRNA increase from wild-type to synthetic is about 6 ‘extends to the synthetic gene as well. to 15 fold compared to a protein increase of about 50 fold. This result is similar to that seen for tobacco in comparing the wild-type and modified genes, and it . That is, about half of the total fold increase in B.t.k- protein from wild-type to modified genes can be explained by mRNA increases and about half to enhanced translational The same is also true in comparing the Although there this mRNA efficiency. modified gene to the synthetic gene. is an additional increase in RNA levels, increase can explain only about half of the total protein increase.

The full length B.t.k. genes described above were also used to transform tomato plants and these plants analyzed for B.t.k. protein and insecticidal The results of this analysis are shown in were efficacy.

Table XIII. gene (pMON10506) produce the B-t.k. HD—73 protein at levels of about 0.01% of their total soluble protein. synthetic/modified gene Plants containing the synthetic/wild—type containing the produce about 0.04% B.t.k. protein, and (pMONl0518) Plants (pMON10526) plants containing the fully synthetic gene produce about 0.2% B.E.k. protein. These results are very similar to the tobacco plant results for the same genes. mRNA levels estimated by Northern blot analysis in tomato also increase in parallel with the protein level increase. As for tobacco with these three genes, most of‘ the protein increase can be attributed to increased mRNA with a small component of translational efficiency increase indicated for the fully synthetic gene.‘ The highest levels «of full lpMous39o) . length B.t.k. protein (from pMON10518) are comparable to or just slightly lower than the highest levels observed for the truncated HD—73 genes (pMON5383_and Tomato plants expressing these full length genes have the insecticidal activity expected for the observed protein levels as «determined by feeding assays with beet armyworm or by diet incorporation of plant extracts with tobacco hornworm.

Table XIII Full Length B.t.k.rHD~73 Protein and N V . . m B.t-k. protein Relative B.t.k.

Gene description Vector concentration mRNA level Synthetic/wild‘ type pudmosoa 100 1 Synthetic/modified pMONl0526 400 2-4 Fully synthetic pMON105l8 2000 10 c) Cotton.

The generality of the increased expression of B.t.k. HD—1 and B.t.k. HD-73 by use of the modified and synthetic genes was extended to cotton.

Transgenic calli were produced which contain the wild type (pMON9921) and. the synthetic HD—l (pMON5377) genes. Here again the B.t.k. HD-1 protein produced ifronn calli containing the wild—type gene was not /detected, whereas calli containing the synthetic HD—1 -81, gene expressed the 39-1 protein at easily detectable levels. The HD-1 protein was produced at approximately 1000 ng/mg of plant calli extract protein. Again, to ensure that the protein produced by the transgenic cotton calli was biologically active and that the increased expression observed with the synthetic gene translated to increased. biofogical activity, extracts of cotton calli were made in similar manner as described for tobacco plants, except that the calli was first dried between Whatman filter paper to remove as much of the water as possible. The dried calli were then ground in liquid nitrogen and ground in 100 mM sodium carbonate buffer, pH 10.

Approximately 0.5 ml aliquotes of this material was applied to tomato leaves with a paant brush. After the leaf dried, five tobacco hornworm larvae were applied to each of two leaf samples. Leaves painted with extract from Control calli were completely destroyed. Leaves painted with extract from calli containing the wild-type HD~1 gene (pMON9921) showed severe damage. Leaves painted with extract from calli containing the synthetic HD—1 gene (pMONS377) showed no damage (see Table XIV below).

Table XIV Brotection against Tobacco Hornworm_hx_IQmat9_Leaxes Painted with Extracts,Prepared from Cotton Calli Containing a Control. the Wild-Tvpe B.t.k. HD—1 Gene.

Svnthetic HD—1 Gene or Synthetic HD-73 Gene Tobacco Hornworm Gene _ Description lento: Damaae:__.i_____ Control Control NL Wild type HD-1 pMON9921 3 Synthetic HD~1 pMONS377 Synthetic HD—73 pMON5383 0 * Damage was rated as shown in Tablelx.

Cotton calli were also produced containing another synthetic: gene, a gene encoding B.t.k; HD—73. The preparation of this gene is described in Example 3.

Calli containing the synthetic HD—73 gene produced the corresponding HD-73 protein at even higher levels than the calli which contained the synthetic HD-1 gene.

Extracts made from calli containing the HD—73 synthetic gene (pMONS383) showed complete control of tobacco hornworm when painted onto tomato leaves as described above for extracts containing the HD—1 protein. (See Table XIV).

Transgenic cotton plants containing the synthetic B.t.k. HD-1 gene (pMON5377) or the synthetic B.t.k. HD- /73 gene (pMON5383) have also been examined. These plants produce the HDél or HD—73 proteins at levels comparable to that seen in cotton callus with the same genes and comparable to tomato and tobacco plants with these genes. For either synthetic truncated HD—1 or HD«73—genes; cotton plants expressing B.t.k. protein at 1000 to 2000 ng/mg total protein (0.1% to 0.2%) were recovered at a high frequency. .Insect feeding assays were performed with leaves from cotton plants expressing the synthetic HD—1 or HD—73 genes. These leaves showed no damage (rating of 0) when challenged with larvae of cabbage looper (Trichoplusia ni), and only slight damage when challenged with larvae of beet armyworm (Spodoptera exigua). Damage ratings are as defined in TablelX above. This demonstrated that Cotton plants as well as calli expressed the synthetic HD—1 or HD—73 genes at high levels and that those plants were protected from damage by Lepidopteran [insect larvae.

Transgenic cotton plants containing either the (pMON5377) or the (pMON5383) were also synthetic truncated HD—1 gene synthetic truncated HD-73 gene assessed for protection against Cotton bollworm at the whole plant level in the greenhouse. This is a more realistic: test of the ability’ of these plants to produce an agriculturally acceptable level of control.

The cotton bollworm (Heliothis zea) is a major pest of cotton that produces economic damage by destroying terminals, squares and bolls, and protection of these fruiting bodies as well as the leaf tissue will be important for effective insect control and adequate crop protection. To test the protection afforded to -34, whole plants, R1 progeny of cotton plants expressing high levels of either s.c.;;. no-1 (pMON5377) or B.t:.k.

HD-73 (pMON5383) were assayed by applying 10-15_eggs of cotton bollwornx per’ boll or square to the 20 uppermost squares or bolls on each plant. At least 12 plants were analyzed per treatment. The batch rate of the eggs was approximately 70%. This corresponds to very high insect pressure compared to numbers of larvae per plant seen under typical field conditions.

Under these conditions 100% of the bolls on control cotton plants were destroyed by insect damage. For significant boll. protection was (HD~1) had 70- the transgenics, observed. Plants containing pMON5377 75% of the bolls survive the intense pressure of this Plants containing pMON5383 (HD-73) had 80%.to This is likely to be a assay. 90% boll protection. consequence of the higher activity of HD-73 protein against cotton bollworm compared to HD-l protein. In cases where the transgenic plants were damaged by the insects, the surviving larvae were delayed in their development by at least one instar. / Therefore, the increased expression obtained with the modified and synthetic genes is not limited to any tobacco, tomato and cotton calli and cotton one crop; plants all showed drastic increases in B.t.k, expression when. the plants/calli were produced containing the modified or synthetic genes. _Likewise, the utility of changes made to produce the modified and synthetic B.t.k. HD—1 gene is not limited to the HD-1 gene. The synthetic HD—73 gene in all three species also showed drastic increases in expression.

In summary, it has"been demonstrated that: (1) the genetic changes made in the HD—1 modified gene lead to very significant increases in B.t.k. HD—1 expression; (2) production of a totally synthetic gene lead to a further five-fold increase in B.t.k. HD-1 expression; (3) the changes incorporated into the modified HD—1 gene accounted for the majority of the increased B.t.k. expression observed with the synthetic gene; (4) the increased expression was demonstrated in three different plants —- tobacco plants, tomato plants and cotton calli and cotton plants; (5) the increased observed by Western also expression as analysis correlated. with similar increases in bioactivity, showing that the B.t.k. HD—1 proteins produced were comparably active; (6) when the method of the present invention used to design the synthetic HD~1 gene was employed to design a synthetic HD—73 gene it also was tomato and ‘with expressed at much higher levels in tobacco, wild—type equivalent gene in bioactivity; (7) a cotton than the increases fully consequent synthetic full length B.t.k. levels comparable to synthetic truncated genes. gene was expressed at Example 5 —- Svnthetic B.t. tenebrionis Gene in Referring to Figure 12, a synthetic gene encoding a Coleopteran active toxin is prepared by making the. indicated changes in the wild—type gene of B.t. ‘tenebrionis or de novo synthesis of the synthetic structural gene. The synthetic gene is inserted into an intermediate plant transformation vector such as pMON893: Plasmids pME>N893 containing the synthetic B.t.t. gene is then inserted into a suitable disarmed Agrobacterium strain such as A. tumefaciens ACO.

Iransformation and Regeneration of Potato Sterile shoot cultures of Russet Burbank are maintained in "vials containing 10 ml of PM medium (Murashige and Skoog (MS) inorganic salts, 30 g/l surcose, 0.17 g/l NaH2PO4HfiL, 0.4 mg/l thiamine-HCl, and 100 mg/l myo-inositol, solidified with 1. g/l Gelrite at pH 6.0). When shoots reached approximately cm in length, stem internode segments of 7~1O mm are excised and smeared at the cut ends with a disarmed Agrobacterium_ tumefaciens vector containing the synthetic B.t:.t. gene from a The stem explants are co-cultured for three four day old plate culture. days at 23°C on a sterile filter paper placed over 1.5 ml of a tobacco cell feeder layer overlaid on 1/10 P medium (1/10 strength MS inorganic salts and organic addenda without casein as in Jarret et al. (1980), 30 g/l surcose and 8.0 g/l agar). Following co-culture the explants are transferred to full strength P-1 medium for callus induction, composed of MS inorganic salts, organic additions as in Jarret et al. (1980) with the exception of casein, 3.0 mg/l benzyladenine and 0.01 mg/l naphthaleneacetic acid (NAA) Carbenicillin (500 mg/l) is (BR), (Jarret, et al., 1980). included to inhibit bacterial growth, and 100 mg/1 kanamycin is added to select for transformed cells.

After four weeks the explants are transferred to medium of the same compos it ion but with 0.3 mg/l gibberellic acid (GA3) replacing the BA‘ and. NAA (Jarret et al., 1981) tc) promote shoot formation.

Shoots begin to develop approximately two weeks after transfer to shoot induction medium; these are excised and transferred to vials of PM medium for rooting.

Shoots are tested for kanamycin resistance conferred by’ the enzyme neomycin phosphotransferase II, by placing a section of the stem onto callus induction medium containing MS organic and inorganic salts, 30 g/l surcrose, 2.25 mg/l BA, 0.186 mg/l NAA, 10 mg/1 GA3 (Webb, et al., 1983) and 200 mg/1 kanamycin to select for transformed cells.

The synthetic B.t.t. gene described in figure 12, was placed into a plant expression vector as descibed in example 5. The plasmid has the following synthetic BglII fragment having inserted into characteristics; a approximately 1800 base pairs was pMON893 in such a manner that the enhanced 35S promoter would express the B.C.t. gene. This pMON1982, was used to transform both Tobacco plants, selected as construct, tobacco and tomato. kanamycin resistant plants were screened with rabbit anti—B.t.t. antibody. Cross—reactive material was detected at levels predicted to be suitable to cause mortality to CPB. These target insects will not feed on tobacco, but the transgenic tobacco plants do ,demonstrate_ that the synthetic gene does improve expression of this protein to detectable levels.

Tomato plants with the pMON1982 construct were determined to produce B;t.t. protein at levels insecticidal to CPB. In initial studies, the leaves of four plants (5190, 5225, 5328 and 5133) showed little or no damage when exposed to CPB larvae (damage rating of 0-1 on a scale of 0 to 4 with 4 as no leaf Under these conditions the control leaves Immunological analysis of remaining). were completely eaten. these plants confirmed the presence of material cross~ reactive with anti—B.t.t. antibody.f Levels of protein expression in these plants were estimated ‘at aproximately 1 to 5 ng of B.t.t. protein in 50 ug of total extractable protein. A total of 17 tomato plants (17 of 65 tested) have been identified which demonstrate protection of leaf tissue from CPB (rating of 0 or 1) and show good insect mortality.

Results similar to those seen in tobacco and tomato with pMON1982 were seen with pMON1984 in the same pMON1984. is identical to pMON1982 plant species.

(CMTI) is except that the synthetic protease inhibitor fused upstream of the native proteolytic cleavage site. Levels of expression in tobacco were estimated to be similar to pMON1982, between 10-15 ng per 50ug of total soluble protein.

Tomato plants expressing pMON1984 have identified which protect the leaves from ingestion by been CPB. The damage rating was 0 with 100% insect mortality.

Potato was transformed as described in example 5 with a vector similar tn) pMON1982 containing the enhanced CaMV35S/synthetic: B.t.t. gene. Leaves of (were totally .protected when challenged, potato plants transfiormed with. this vector, were screened by CPB insect bioassay. Of the 35 plants tested, leaves from 4 plants, 16a, 13c, 13d, and_23a Insect bioassays with leaves from three other plants, 13e, la, and 13b, recorded damage levels of 1 on a scale of O to 4 with 4 being total devestation of the leaf material. Immunological analysis confirmed the presence of B;t.t. cross-reactive material in the leaf tissue. The level of B.t.t. protein in leaf tissue of plant 16a (damage rating~of 0) was estimated at 20-50 ng of B.t.t. protein/50 ug of total soluble protein.

The levels of B.t.t. protein seen in 16a tissue was consistent with its biological Immunological analysis of l3e and 13b (tissue which scored 1 in damage rating) reveal less protein (5-10 ng/S0 ug of total soluble protein) than in plant 16a.

Cuttings of plant 16a were challenged with 50 to 200 eggs of CPB :h1 a whole plant assay. Under these conditions l6a showed no damage and 100% mortality of insects while control potato plants were heavily damaged.

Examnlg 5 -- Smmmm E c K 22 gram", Gang The P2 protein is a distinct insecticidal protein produced by some strains of B.t. including B.t.k. HD— . It is characterized by its activity against both lepidopteran and dipteran insects (Yamamoto and Iizuka, 1983). Genes encoding_the P2 protein have been isolated. and characterized (Donovan et a1., activity.

The P2 proteins encoded by these genes are in length. AThese with_ the ). approximately 600 amino eacids proteins share only limited homology lepidopteran specific P1 type proteins, such as the B.t.k. HD-l and HD-73 proteins described in previous examples.

The P2 proteins have substantial activity against a variety of lepidopteran larvae including cabbage looper, tobacco hornworm and tobacco budworm. Because they are active against agronomically important insect, pests, the P2 proteins are a desirable candidate in the production of insect tolerant transgenic plants either alone or in combination with the other B.t. toxins described i1: the above~ examples. In some plants, expression of the P2 protein alone might be sufficient to provide protection against vdamaging insects. In addition, the P2 proteins might provide protection against agronomically important dipteran pests. In other cases, expression of P2 together with the B.t.k. HD-1 or HD-73 protein might be preferred.

The P2 proteins should provide at least an additive level of insecticidal activity when combined with the crystal protein toxin of B.t.k. HD~1 or ED-73, and the combination may even provide a synergistic activity.

Although the mode of action of the P2 protein is unknown, its distinct amino acid sequence suggests that it functions differently from the B.t.k. HD-1 and HD—73 type of proteins. Production of tuo insect tolerance proteins with different modes of action in the same plant would minimize the potential for development of insect resistance to B.t. proteins in ‘of B.c. bacteria. plants. The lack of substantial DNA homology between P2 genes and the HD~1 and HD-73 genes minimizes-the potential for recombination between nmltiple insect Vtolerance genes in the plant chromosome.

The genes encoding the P2 protein although distinct in sequence from the B.t.k} HD—1 and HD—73 genes share many common features with these genes. In particular, the P2 protein genes have a high A+T content (65%), multiple potential polyadenylation signal sequences (26) and numerous ATTTA sequences (10). Because of its overall similarity to the poorly expressed wild- type B.t.k. HD-1 and HD—73 genes, the same problems are expected in expression of the wild—type P2 gene as were encountered with the previous examples. Based on the above—described method for designing the synthetic B.t. genes, a synthetic P2 gene has been designed which gene should be expressed at adequate levels for protection in plants. A comparision of the wi1d—type and synthetic P2 genes is shown in Figure 13.

Example 7 —- Svnthetic B.t..Entomocidus Gene The B.t. entomocidus ("Btent") protein is a distinct insecticidal protein produced_by some strains It is characterized by its high level of activity against some lepidopterans that are relatively insensitive to B.t.k. HD-1 and HDr73 such as Spodoptera species including beet armyworm (Visser -et al., 1988). Genes encoding the Btent protein have been isolated and characterized_4Honee et al, 1988).

The Btent proteins encoded by’ these genes are approximately the same length as B.t.k. HD-1 and HD- 73. These proteins share only 68% amino acid"homology with the B.t.k. HD-1 and HD—73 proteins. It is likely that only the N—terminal half of the Btent protein is required for insecticidal activity as is the case for HD-1 and HD-73. Over the first 625 amino acids, Btent shares only 38% amino acid homology with HD-1 and HD- 73.

Because of their higher activity against Spodoptera species that are relatively insensitive to RD-1 and HD- 73, the Btent proteins are a desirable candidate for the production of insect tolerant transgenic plants either alone or in combination with the other B.t. toxins described in the above examples. In some plants production of’ Btent alone might be sufficient to control the agronomically important pests. In other plants, the production of two distinct insect tolerance proteins would provide protection against a wider array of insects. Against those insects where both proteins are active, the combination of the B.t.k. HD-1 or HD—73 type protein plus the Btent protein should provide at least additive insecticidal efficacy, and may even provide a synergistic activity.

In addition, because of its distinct amino acid sequence, the Btent protein may have a different mode of action than HD-1 or HD—73. Production of two insecticidal proteins in the same plant with different modes of action would nninimize ‘the potential for vdevelopment of insect resistance to B.t. proteins in plants. The relative lack of DNA sequence homology with the B.t.k. type genes minimizes the potential for recombination between multiple insect tolerance genes in the plant chromosome.

The genes encoding the Btent protein although distinct in sequence from the B.t.k. HD-1 and HD—73- genes share many common features with these genes. In particular; the Btent protein genes have a high A+T content‘ (62%), multiple ‘potential polyadenylation signal sequences (39 in the full length_ coding sequence and 27 in the.first 1875 nucleotides that is likely to encode the active toxic fragment) and. numerous ATTTA sequences (16 in the full length coding sequence and 12 in the first 1875 nucleotides).

Because of its overall similarity to the poorly expressed wild type B.t.k. HD-1 and HD-73 genes, the wild—type Btent genes are expected to exhibit similar problems in expression as were encountered with the wild-type HD-1 and HD-73 genes. Based on the above- described method used for designing the ‘other synthetic B.t. genes, a synthetic Btent gene has been designed which gene should be.expressed at adequate levels for protection in plants. A comparision of the wild type and synthetic Btent genes is shown in Figure 14.

Exampleofl -- Svnthetic B.t.k. Genes for Expression I" G .

High level expression of heterologous genes in corn cells has been shown to be enhanced by the presence of a corn gene intron (Callis et a1., 1987). Typically these introns have been located in the 5' untranslated _94_ region of the chimeric gene. It has been shown that the CaMV35S promoter and ‘the NOS 3' end‘ function efficiently in the expression of heterologous genes in corn cells (Fromm et al., 1986).

.Referring to Figure 15, a plant expression cassette vector (pMON744) was constructed that contains these sequences. Specifically the expression cassette contains the enhanced CaMV 35S promoter followed by intron 1 of the corn Adhi gene (Callis et a1., 1987).

This is followed by a multilinker_cloning site for insertion of coding sequences; this multilinker contains a BglII site among others. Following the multilinker is the NOS 3' end- pMON744 also contains the selectable marker gene 35S/NPTII/NOS 3' for kanamycin selection of "transgenic corn cells. In addition, pMON744 has an E- coli origin of replication and an ampicillin resistance gene for selection of the plasmid in E. coli.

Five B.t.k. coding sequences described in the previous examples were inserted into the BglII site of pMON744 for corn cell expression of B.t.k. The coding sequences inserted and resulting vectors were: . Wild type B.t.k. HD—1 from pMON9921 to make pMON8652.

. Modified .B.t.k. HD—1 ‘from pMON5370 to make pMON8642.

. Synthetic B.t.k. HD¥1 from pMONS377 to make pMON3643.

. Synthetic B.t.k. HD~73 firom pMON5390 to make pMON8644.

. Synthetic full length B.t.k. HD-73 from pMON10518 to make‘ pMON10902.r pMON8652 (wild-type .B.t.k. HD-1) was used to transform corn cell protoplasts and stably transformed kanamycin resistant callus was isolated. B.t.k. mRNA in the corn cells was analyzed by ‘nuclease 31 protection and found to be present at a level comparable to that seen with the same wild—type coding sequence (pMON9921) in transgenic tomato plants; pMON8652 and pMON8642 (modified HD-1) were used to transform corn cell protoplasts in a transient The level of B.t.k. mRNA was The modified HD-1 expression system. analyzed by nuclease 51 protection. gave rise to a several fold increase in B.t.k. mRNh compared to the wild—type coding’ sequence in the transiently transformed corn cells. This indicated that the modifications introduced into the B.t.k. HD~1 gene are capable of enhancing B.t.k. expression in monocot cells as was demonstrated for dicot plants and cells. pMON8642 (modified HD~1) and pMON8643 (synthetic HD~1) were used to transform Black Mexican Sweet (EMS) corn cell protoplasts by PEG-mediated DNA uptake, and stably transformed corn callus was selected by growth oni kanamycin containing -plant growth medium.

Individual callus colonies that were derived from single transformed cells were isolated and propagated separately on kanamycin containing medium.

To assess the expression of_the B.tik. genes in these cells, callus samples were tested for insect toxicity by bioassay against tobacco hornworm larvae; For each vector; 96 callus lines were tested by bioassay. Portions of each callus were placed on lsterile water agar plates, and five neonate tobacco hornworm larvae were added and allowed to feed for 4 days. For pMON8643, 100% of the larvae died after feeding on 15 of the 96 calli and these calli showédt little feeding damage. _For pMON8642, only 1 of the 96' calli was toxic to the larvae. This showed that the B.t.k. gene was being expressed in these samples at insecticidal levels. The observation that significantly more calli containing pMON8643 were toxic than for pMON8642 showed that significantly higher levels of expression were obtained when the synthetic HD—1 coding sequence was contained in corn cells than when the modified HD-1 coding sequence was used, similar to the previous examples with dicot plants. A semiquantitative immunoassay showed that the pMON8643 toxic samples had significantly higher B.t.k. protein levels than the pMON8642 toxic sample.

The 16 callus samples that were toxic to tobacco hornworm were also tested_ for activity against European corn borer. European corn borer is approximately 40-fold less sensitive to the HD-1 gene product than is tobacco hornworm. Larvae of European corn borer were applied to the callus samples and allowed to feed for 4 days. Two of the 16 calli tested, both of which contained pMdN8643 (synthetic HD- ,1), were toxic to European corn borer larvae.

To assess the expression of the B.t.k. genes in differentiated corn tissue, another method of DNA delivery was used. Young leaves were excised from ‘corn plants, and DNA samples were delivered into the leaf tissue by microprojectile bombardment. In this system, the DNA on the microprojectiles is transiently expressed in the leaf cells afiter bombardment. Three DNA samples were used, and each DNA was tested in triplicate. . pMON744, the corn expression vector with no B.t.k. gene. . pMON8643 (synthetic HD-1)- . pMON752, a corn expression vector for the GUS gene, no B.t.k. gene.

The leaves were incubated at room temperature for 24 hours. The pMON752 samples were stained with a zubstrate that allows visual detection of the GUS gene product. This analysis showed that over one hundred spots in each sample were expressing the GUS product and the the triplicate samples showed very similar levels of GUS expression. For the pMON744 and pMON8643 samples 5 larvae of tobacco hornworm were added to each leaf and allowed to feed for 48 hours.

All three samples bombarded with pMON744 showed extensive feeding damage and no larval mortality. All three samples bombarded with pMON8643 showed no evidence of feeding damage and 100% larval mortality.

The samples were also assayed for the presence of‘ B.t.k. protein by a qualitativelimmunoassay. All of the pMON8643 samples had detectable B.t.k. protein.

These results denonstrated ‘that the the synthetic B.t.k. gene was expressed in differentiated corn plant tissue at insecticidal levels.

Exam21e_2 '-- Expression of Svnthetic B.t. Genes RUBISCO (SSU) are_ often_ highly expressed, light regulated. and sometimes show tissue specificity- These expression properties are largely due to the promoter sequences of these genes. It has been possible to use SSU promoters to express heterologous genes in transformed plants. Typically a plant will Contain multiple SSU genes, and the expression levels and tissue specificity of different S58 genes will be different; The SSU pgoteins are encoded in the nucleus and synthesized in the cytoplasm as precursors that contain an bkterminal extension known as the chloroplast transit peptide (CTP). The CTP directs the precursor to the chloroplast and promotes the uptake of the SSU protein into the chloroplast, In this process, the CT? is cleaved from the SSU protein.

These CTP sequences have. been used to direct heterologous proteins into chloroplasts of transformed plants. p M The SSU promoters might have seyeral advantages for expression of B.t.ku genes 31: plants. some SSU promoters are very highly expressed and could give ‘rise to expression levels as high or higher than thosep _99_ observed with_ the CaMV3$S promdter; _The tissue distribution of expression from. SSU 'piomoters is differént from that of the CaMY35S promoter, so-for control of some insect pests, it may be advéntafieous. to direct the expression of B.t.k. to those cells in which SSU is most highly expressed. For example,» although relatively constitutive, in the leaf the CaMV35S promoter is more highly expressed in vascular tissue than in some other parts of the leaf, while most SSU promoters a‘re—most highly expressed in the mesophyll cells of the leaf. Some SSU,promoters also are more‘ highly tissue specific, so it could be possible to utilize a specific SSU promoter to express B.t.Ak. in only 'a subset of ‘plant tissues, if for ‘example B.t. expression in certain cells was found to be deleterious to those cells. For example, for control of Colorado potato beetle in potato, it may be advantageous to use-SSU promoters to direct B.t.t. expression to the leaves butgnot to the edible tubers- Utilizing SSU CTP sequences to localize B.t. proteins to the chloroplast might also be advantageous. Localization of the lB.t- to the chloroplast could protect the protein from proteases found in the cytoplasm. This could stabilize the B.t:. protein and lead to higher levels of accumulation of active protein. B.t. ‘genes containing the CTP could be used in combination with the SSU promoter or with other promoters such as CaMV35S.

A variety of plant transformation vectorswere constructed for the expression of B.t.k,. genes utilizing SSU promoters and SSU CTPs. The promoters and CTPs utilized were from the petunia SSU11a gene described by Tumer et ‘ail. (1986) and from ‘the Arabidopsis atsililgene (an SSU gene) described DY Krebbers et al. (i989) and by Elionor et a1. (1989).

The petunia SSU11a promoter was contained on a;DNA fragment that extended approximately 800 bp upstream of_the SSU coding sequence. The Arabidopsis ats1A prooter was contained on a DNA fragment that extended appro1'cimatelyV1.8- kb upstream of the SSU coding- sequence. At the upstream end convenient sites from the multilinker of pUCl8 were used to move these promoters into plant transformation vectors such as pMON893. These promoter fragments extended to the start of the SSU coding sequence at which point an NcoI restriction site was engineered to allow insertion of the B.t- coding sequence, replacing the SSU coding sequence.

When SSU promoters were used in combination with their CTP, the DNA fragments extended through the coding sequence of the CTP and a small portion of the mature SSU coding sequence at which point an NcoI restriction site was engineered by standard techniques to allow the in frame fusion of B.t. coding sequences with the CTP. In particular, for the petunia SSU11a CTP,.B.t. coding sequences were fused to the SSU sequence after amino acid 8 of the mature SSU sequence at which point the NcoI site was placed. The 8 amino acids of mature SSU sequence were included because preliminary in vitro chloroplast uptake experiments indicated that uptake was of B.t.k. was observed only if this segment of mature SSU was included. For the Arabidopsis ats1A CTP, the complete CTP was included plus 24 amino acids of nature SSU sequence plus the sequence gly-gly-arg—va1-asn—cys—met—gln—ala—met; terminating in an Ncol site for B.t. fusion. This short sequence reiterates the native SSU CT? cleavage site (between the cys and met) plus a short segment surrounding the cleavage site. This sequence was included in order to insure proper uptake_ into B.t. coding sequences were fused to In vitro uptake chloroplasts. this ats1A CTP after the met codon. experiments with this CTP construction and other (non- B.t.) coding sequences showed that this CTP did target proteins to the chloroplast.

When CTPS were used in combination with the CaMV 35S promoter, the same CTP segments were used. They were excised just upstream of the ATG start sites of the CTP by engineering of BglII sites, and placed downstream of the CBMVBSS promoter in pMON893, as Bg1II7to NcoI fragments. B.t. coding sequences were fused as described above.

The wild type B.t.k. HD—l coding sequence of pMON9921 (see Figure 1) was fused to the ats1A promoter to make pMON1925 or the ats1A promoter plus CTP to make pMON1921. These vectors were used to transform tobacco plants, and the plants were screened for activity against tobacco hornworm. No toxic plants were recovered. This is surprising in light of the fact that toxic plants could be recovered, albeit at a low frequency, after transformation with pMON9921 in which the.B.t.k. coding sequence was expressed from the enhanced CaMV3SS, promoter in pMON893, and in light of the fact that Elionor et al. (1989) report that the ats1A [promoter itself ism comparable in strength to the CaMV3SS promoter"and approximately 10- fold stronger when the CT}? sequence is included. At least for the wiId—type B.t.,k. HD-1 coding sequence, this does not appear to he the case. _A variety of plant transformation vectors were constructed -utilizing either the truncated synthetic HD—’l3 coding sequence of Figure 4 or-the full length B.t.k. HD-73 coding sequence of Figure 11. These are listed in the table below.

Table XV Gene Constructs with CTPS yentnr Exnmter CIR Bl£.&_Jflt13 Qmfinaihamume pMON10806 En 35$ atalh truncated pMON1081-1 En3SS SSU11a full length pMON10811 SSUl1a SSU11a truncated pMON10819 SSU11a none truncated pMON10815 atsla none truncated pMON1081'l atslh. atsln truncated pHON10821 En 35$ ats1A truncated pMONl0822 En 355 atsllk full length pMON10838 SSU1la SSUlla full length pMONl0839 atsla atslh full length All of the above vectors were used to transform tobacco plants- For all of the vectors containing truncated B.t.k. genes, leaf tissue from these plants has been analyzed for toxicity to insects and B.t.k. protein levels by immunoassay. pMON10806, 10811, 10819 and 10821 produce levels of B.t.k. protein comparable tn) pMON5383 and pMONS390 which contain synthetic B.t.k. HD-73 coding sequences driven by the Sn 35S promoter itself with no CTP. These plants also have the insecticidal activity expected for the B.t.k; protein levels detected. For pMON108l5 and pMON10817 (containing the ats1A promoter), the level of B.t.k. .protein is about 5-fold higher than that found in ,plants containing pMON5383 or 5390. These plants also have higher insecticidal activity. Plants containing and 10817 contain up in) 1% of their total soluble leaf protein as B.t.k. ED-73. This is the highest level of B.t.k. protein yet obtained with any of the synthetic genes.

This result is surprising in two respects. as noted above, the wild type coding sequences fused to the ats1A promoter and CTP did not show any evidence of higher levels of expression than for En 35S, and in fact had lower expression based on the Second, Elionor First, absence of any insecticidal plants. et al. (1989) show that for two other genes, the ats1A CTP can increase expression from the ats1A promoter by about 10-fold. For the synthetic B.t.k. HD—73 gene, there is no consistent increase seen by including the CTP over and above that seen for the atslA promoter alone.

Tobacco plants containing the full length synthetic ao—73 fused to the SSUIIA CTP and driven by the En 35S promoter produced levels of B.t.k. protein and insecticidal activity comparable to pMON1518 which contains does not include the CTP. In addition, for pMONl0518 the B.t.k. protein extracted from plants was observed by gel electrophoresis to contain multiple forms less than full length, apparently due the cleavage of the C—terminal portion (not required for toxicity) in the cytoplasm. For pMON108l4, the majority of the protein appeared to be intact full length indicating that the protein has been stabilized from proteolysis by targeting to the chloroplast. 'compartmentalization.

~The:B.t. proteins produced from the synthetic genes described here are localized to the cytoplasm of the plant cell} and this cytoplasmic localization results in plants that are insecticidally effective. It may be advantageous for some purposes to direct the B.t. proteins to othericompartments of the plant cell.

Localizing Bpt. proteins in compartments other than the cytoplasm may result in less exposure of the B.t- proteins to cytoplasmic proteases leading to greater enhanced protein yielding Extracellular localization accumulation of the insecticidal activity. could lead to more efficient exposure of’ certain to the B.t. efficacy. If a B.t. deleterious to plant cell function, then localization insects proteins leadinge to greater protein were found tx) be to a noncytoplasmic compartment could protect these cells from the . protein.

In plants as well as other eucaryotes, proteins that are destined to be localized either extracellularly or in several specific compartments are typically synthesized with an N-terminal amino acid extension known as the signal peptide. This signal peptide directs the protein to enter the compartmentalization pathway, and it is typically cleaved from the mature protein as an early step in For an extracellular protein, the secretory pathway typically involves cotranslational insertion into the endoplasmic reticulum with cleavage of the signal peptide occuring at this stage. The mature protein then passes thru the Golgi body into vesicles that fuse with the plasma membrane ithus releasing the protein into the extracellular space. Proteins destined for other compartments follow a similar pathway. For example, proteins that are destined for the endoplasmic reticulum or the Golgi body follow this scheme, but, they are specifically retained in the appropriate compartment. In plants, some proteins» are‘ also targeted to the vacuole, another membrane bound compartment in the cytoplasam of many plant cells.

Vacuole targeted proteins diverge from the above pathway at the Golgi body where they enter vesicles that fuse with the vacuole.

A common feature of this protein targeting is the signal peptide that initiates the compartmentalization process. Fusing a signal peptide to a protein will in many cases lead to the targeting of that protein to the endoplasmic reticulum. The efficiency of this step may depend on the sequence of the mature protein itself as well. The signals that direct a protein to a specific compartment rather than to the extracellular space are not as clearly defined. It appears that many of the signals that direct the protein to specific compartments are contained within the amino acid sequence of the mature protein. This has been shown for some vacuole targeted proteins, but it is not yet possible to define these sequences precisely. It appears that »secretion into the extracellular space is the "default" pathway for a protein that contains a signal sequence but no other- compartmentalization signals. Thus, a strategy. to idirect B.t. proteins out of the cytoplasm is to fuse the genes for synthetic B.t. genes to DNA sequences encoding known plant signal peptides. These fusion genes will give rise to B.t. proteins that enter the secretory pathway, and lead to extracellualar secretion »or targeting to the vacuole or other compartments, Signal sequences for several plant genes have been One such sequence is for the tobacco related protein PR1b described by described. pathogenesis Cornelissen et al. The PR1b protein is normally ‘localized to the extracellular space. Another type of signal peptide is contained on seed storage proteins These proteins are localized to the like of legumes. protein body of seeds, which is a vacuole compartment found in seeds; A signal peptide DNA sequence for the beta subunit of the 7S storage protein of common bean (Phaseolus vulgaris)/ PvuB has been described by Doyle et al. Based on the published these published sequences, genes were synthesized by chemical synthesis of oligonucleotides that encoded the signal peptides for PRlb and PvuB. The synthetic genes for these signal peptides corresponded exactly to the reported DNA sequences. Just upstream of the translational intiation codon of each signal peptide a eBamHI and BglII site were inserted with the BamHI site at the 5' end. This a1Iowed_the insertion of the signal peptide encoding segments into the Bg1II site —l09— of pMON893 for expression from the En 35S promoter.

In ‘some cases to achieve secretion . or compartmentalization of heterologous proteins, it_has proved necessary to include some amino acid sequence beyond the normal cleavage site of the signal peptide; This may be necessary to insure proper cleavage of the signal peptide. For PR1b the synthetic DNA sequence also included the first 10 amino acids of mature PR1b.

For PvuB the synthetic DNA sequence included the first 13 amino acids of mature PvuB. Both synthetic signal peptide encoding segments ended with Ncol sites to allow fusion in frame to the methionine initiation codon of the synthetic B.t. genes.

Four vectors encoding synthetic B.t.k. HD—73 genes were constructed containing these signal peptides.

The synthetic truncated HD—73 gene from pMON5383 was fused with the signal peptide sequence of PvuB and incorporated into pMON893 to create pMON10827. The synthetic truncated HD—73 gene from pMONS383 was also fused with the signal peptide sequence of PR1b to create pMON10824. The full length synthetic HD—73 gene from pMON10518 was fused with the signal peptide sequence.of PvuB and incorporated into pMON893 to create pMON10828. The full length synthetic HD~73 gene from pMON1OS18 was also fused with the signal peptide sequence of PR1b and incorporated into pMON893 to create pMON1082S.

These vectors were used to transform tobacco plants and the plants were assayed for expression of the B.t.k. protein’by Western blot analysis and for insecticidal efficacy. pMONl0824 and pMON10827 produced amounts of &.t.k. protein in leaf comarable to the truncated an-73 ‘vectors, pMON5383 ‘and’ pMoN_s39'o. pMON1082S and pMON10823 produced full length B,t.k. ‘ protein in amounts comparable to pMON10S18. In all caées, the plants were insecticidally active against tobacco hornworm. -1I1— BIBLIQGBAEHI Adami, G. and Nevins, J. (1988) RNA Processing, Cold ‘Spring Harbor Laboratory, p. 26.

Adang, et al., V._ 3 Ergnggnign (1987) pp. 345-353, Alan R. Liss, Inc.

Barton, K. A. et al., £1anL_Eny5iQ1* (1987), 85,1103- .

Bevan, M. et al., uannrg (1983) 304:184.

Brady, H. and Wold, w. (1988), RNA Processing, c¢1d Spring Harbor Laboratory, p. 224.

Brown, John w., flflQ]gjg ggjgs Bgseazgh (1986) Vol. 14, No. 24, p. 9549.

Conway, L. and Wickens, M. (1988), RNA Processing, Cold Spring Harbor Laboratory, p. 40.

Cornellssen, B.J.C., et al., fiMBQ_J. (1986) Vol. 5, No. 1; 37-40.

Daar, I. O. et al. (1988), RNA Processing, Cold Spring Harbor Laboratory, p.'45.

, No. 5, p. 2229.

Dedrick, R., et al., Qhgmiagxx (1987), Vol. 262, No. 19, pp. 9098-1106.

Donovan, W. P. et al., (1988), Vol. 263, N0. 1, pp. 561-567.

Doyle, J.J., et al., J. Biol. Qhgm.

No. 20, 9228-9236. (1986), Vol. 261, (1989), 218:78~ 86.

Fischboff, D. A. et a1., fiigfimgchnglggy (1987), Vol. , p. 807.

Fraley, R. T. et a1., Eigzxgghnglggx (1985) 3:629—635.

From, M., Taylor, L. P. and Walbot, V., uaggxg (1985), 319:791-793.

Gallego, N. E. and Nadal—Ginard B. (1988), RNA Processing, Cold Spring Harbor Laboratory, p. 61.

Genovese, C. and Milcarek, C. (1988), RNA Processing, Cold Spring Harbor Laboratory, p. 62. ' Gil, A. and Proudfoot; N. J., Nature (1984), Vol- 312, p. 473.

Goodall, G. et al. (1988), RNA Processing, Cold Spring Harbor Laboratory, p. 63.

Gross, et al. (1988), RNA Processing, Cold Spring Harbor Laboratory, p. 128.

Hampson, R. K. and Rottman, F. M. (1988), RNA Processing, Cold Spring Harbor Laboratory, p. 68.

Helfman, D. M. and Ricci, W. M. (1988), RNA Processing, Cold Spring Harbor Laboratory, p. 219.

Herrera—Estrella, L. et al., Nature (1983), 303:209.

Hoekema, A. et a1., (1987), Vol. 7, pp. 2914-2924.

Honee. G. et a1., (1988)., Vol.

, No. 13.

Horsch, R. B. et a1., science (1985), 227:1229. —l13- Jarret, R. L. et g1.,"1n_yi;;Q (1981), 17;825.

Kay,'R. et al., science (1987), 236:1299-1302.

Kessler, M. et al. (1988), RNA Processing, Cold Spring Harbor Laboratory, p. 85.

.Klee, H. J. et al., aigzigghnglggg (1985), 3:637—642.

Kozak, M., Eafinpg (1984), 308:241—246.

Krebbersy E.. et al., E1anL1M1enn1ar_Bin12g¥ (1988), :745~759.

Kunkel, T. A.. Er9c1_naL11_acad1_$2i1_fl$A (1985). Vol. , pp. 488-492.

Marrone et al., J. E993, EnLQmQ;, (1985), 78—290—293.

Marzluff, W. and Pandey, N. (1988), RNA Processing, Cold Spring Harbor Laboratory, p. 244.

McCormick. 8. et al., E1anL_£e11_Be2QrLs (1986), 5:81- MCDevitt, M. A. et al., $211 (1984), V0] 37, pp. 993- 999.

Murashiqe. T. and Skooq. F.. Rhxs1g11_E1anL (1962), :473.

Odell, J. et al., uagnng (1985), 313:810.

Pandey, N. B. and Marzluff, W. F. (1987), RNA Processing, Cold Spring Harbor Laboratory, p. 133.

Proudfoot, N. J. et al. (1987), RNA Processing, Cold Spring Harbor Laboratory, p. 17.

Reines, 0., et al., J*_MQl4_fliQl* (1987) 196:299~312. ~114— »Vo1. 10, No. 24, pp. 8225-8244.

Shaw, G. 8 Kamen, R., cell (1986), 46:6S9‘667.

Shaw, G. and Kamen, R. (1987), RNA Processing, Cold Spring Harbor Laboratory, p. 220.

Trolinder, N. L. and Goodin, J. R., R1anL_Qe1l Beggrng (1987), 6:231-234.

Tsurushita, N. and Korn, L. J. (1987), RNA Processing, Cold Spring Harbor Laboratory, p. 215.

Tumer. N.E., et al., Hnc1eic_AciQs_Beg. 14:8, 3325. (1986), Vol.

Vaeck, M. et al., flafiurg (1987), Vol. 328, p. 33.

Velten et al., fifiQ_Q, (1984), 322723-2730.

Velten & Schell. Hu212i2_AsiQ:_Beaear2h (1985). :6981-6998. . et: al., 1491. Gen. Genet. (1988), 212:219— Visser, B. 224.

Wickens, M. and Stephenson, P., sgienge (1984), Vol. 226, p. 1045.

Wickens, M. et al. (1987), RNA Processing, Cold Spring Harbor Laboratory, p. 9.

Wiebauer. K. et al., (1988), Vol. 8, No. 5, pp. 2042-2051.

Yamamoto, T. (1933), Vol. 227, No. 1, pp. 233-241.

Claims (2)

1.CLAIMS J. A method for modifying a wild—type structural gene sequence which encodes an insecticida1_protein of Bacillus thuriagji ensis to enhance the expression of said protein in p1an:5 35 which comprises: identifying regions within said sequence with greater Vthan four consecutive adenine or thymine nucleotides; modifying the regions of step (a) which have two or More polyadenylation signals within a ten base sequence to remove ‘said signals while maintaining a sequence which encodes said protein; and modifying the 15-30 base regions surrounding the regions of step (a) to remove major plant polyadenyla— tion signals, consecutive sequences containing more than one minor polyadenylation signal and consecutive sequences containing more than one ATTTA sequence while maintaining a gene sequence which encodes said protein. A method for modifying a wild-type structural gene sequence which encodes an insecticidal protein of Bacillus thuringi— en§is to enhance the expression of said protein in plants which comprises: signals contained in said removing polyadenylation wi1d~type gene while retaining a sequence which encodes said protein; and removing ATTTA sequences contained in said wild-type gene while retaining a sequence which encodes said pro—. cein. gene’ A method of claim 2 further comprising the removal of selfc complementary segnences and replacement of such sequences; with nonself—~comp1vementary DNA cotngrising plant preferred codons while retaining a structural gene sequence ‘encoding- said protein . A method of clains _1 to 3 further comprising the use of planttpceterted sequences in the removal of the polyadenyla- tion signals ‘and sequences- A—l‘:et.hod of 1 ‘to 3 in which. the plant polyadenylation signals are selected the group consisting of Aacrana, Aarapii, Mocha, mm, Anna, AAGCQT, ATTAAT, ATACB1‘, AABATA, A1‘-‘LEAR, AATTAA, RAIACA and CATAAA. A method for improving the expression of a heterologous gene in plants wherein said gene comprises a modified chimeric gene containing a promoter which functions in plant cells ope;-ably linked to a structural coding sequence and a 3' ncm—ti-anslated region containing a polyadenylation signal which functions in plants to cause the addition of polyadenylate nucleotides to the 3’ end of the RNA, wherein said structural coding sequence encodes an insecticidal protein at lieastm a portion of which was derived from a Bacillus churingiensis protein, wherein said method com- prises modifying said structural coding sequence so that said sequence has a DNA sequence which differs from the naturally occurring DNA sequence encoding said Bacillus thuringiensis protein and said structural coding sequence does not contain more than 5 consecutive_ nucleotides consisting of either adenine or thy-nine residues. A method for improvind the expression of a hetegologous gene in plants wherein said gene comprises a modified Chimeric gene containing a promoter which functidns in plant cells opetably linked to a structural coding sequence and a 3” non—tzans1ated region containing a pblyadenylatian signal which functions in plants to cause the addition of po;yadenylate nucleotides to the 3' end of the RNA, wherein said structural coding sequence encodes an. insecticidal prptefin at least a portion of vmdch was derived from a Bacillus thuringiensis prdtein, whefein said method com‘. prises nodifying said structuraf coding sequence so that said sequence has a DNA sequence which'dif£ers frum the uatmdra-ilya Dim ‘sequence vencoding said Bacilius cbuzingieasis protein and has fhe following—cha:acteristics: said structural coding sequence has a region which is Complementary to the following sequence: GGCTTGATTC CTAGCGAACTC TTCGATTC TC TGGT'l‘GATGAGC'1‘GTTC 1 5 10 15 20 25 30 35 40 45 said region in said coding sequence having eliminated

2.AACCAA and 1 AATTAA sequence. 1% nethod according to claim 7, wherein said structural coding sequence encodes an insecticidal protein at least a portion of which was derived from a Bacillus thuringiensis kurstakis HD-1. A method according to claim 7 or 8, wherein the plant is a tobacco plant. A modified chimeric gene containing a promoter whxch functions in plant cells operably linked to 2: structural coding Sequence and a 3' r.on——translated region Containing a. polyadenylation signal which functions in plants to cause the addition of polyacienylate nuclgotides to the 3' said structural ~coding sequence encodes an L-xse<_:tic.='.da1 protein at least; a portion of which was derivg-‘d from a Bacillus thuringiensis protein, wherein said stzruczural coding sequence has a DNA sequence which differs from the naturaily occ:'._1r2:ing BRA sequence encoding said Bacillus tlzttriagieasis protein and is Sé1'ected from: 7 V which encodes an T 4 A ‘I31 structutalyene ‘ ED-1 having. the