Method for gene optimization

Abstract

The invention relates to method of modifying a coding sequence encoding a non-plant protein, comprising the steps of optimizing said coding sequence by codon substitution, thereby obtaining an optimized coding sequence which encodes said non-plant protein; and re-introducing at least one wild-type polyadenylation motif sequence at its position within said optimized gene sequence.

Description

Low transgene protein expression in transgenic plants can be attributable to several factors (see Lu et al (2015) for a review). These include:

• • low levels of transcription, attributable to linkage of genes to weak promoters.
  • aberrant processing of transcripts including mis-splicing and premature polyadenylation giving a transcript that lacks parts of the coding region.
  • low translational initiation caused by ATG initiation codons upstream of the start point and poor accessibility of the ribosome to the correct initiation site.
  • low translation rates due to presence of rarely used codons in the coding region.
  • post-transcriptional factors such as RNA stability, protein modifications and protein stability.

Often transgenes to be expressed in plants are derived from other plant species or non-plants. These genes are evolutionary adapted for expression in their host organism at the desired expression level in the host organism, but may be not adapted for expression in the transformed plant. Additionally some genes from the same plant as the plant to be transformed may not be adapted for high-level expression in that same plant. Different organisms have different DNA base compositions (AT % or GC %) as do different genomes within an eukaryotic cell (nucleus versus mitochondria (with also T to U) and in plant cells (plastids)). This difference in DNA base pair composition affects the frequency of the occurrence of degenerate codons coding for the same amino-acid (codon use frequency). The abundance of the cognate charged tRNAs is generally proportional to the frequency of the target codons in the genome. Thus for example genes that are rich in AT % are poorly translated in organisms that are GC % rich due to a lack of certain charged tRNAs. It is well known in the art that this problem can be overcome by recoding transgenes such that the codon usage reflects that used in the transgenic organism and if high expression is required, match the codon usage to that of highly expressed genes in that organism.

In addition to the question of codon-usage eukaryotic nuclear genes have transcript splicing and polyadenylation signals that may differ between eukaryotic organisms (eg between plants animals and insects or between dicotyledonous and monocotyledonous plants) and that may be absent in other organisms such as prokaryotes. Expression of a gene from exogenous species in a transgenic plant may thus lead to unwanted transcript processing such as mis-splicing and premature polyadenylation. In animals the polyadenylation signal has been found to be composed of 2 major elements, the AAUAAA motif positioning element (PE) located 10 to 30 bp upstream of the polyadenylation site (cleavage site (CS)) and a U or UG rich downstream element (DE) downstream of the CS (Colgan and Manley (1997)) Efforts have been made to identify DNA sequences that act as polyadenylation signals in plants. Joshi (1987) analysed 4 domains downstream of the coding region in 46 plant genomic sequences and identified putative consensus sequences upstream and downstream of the AAUAAA like motif. Graber et al (1999) compared polyadenylation signals in silico in yeast, Arabidopsis, rice, fruitfly, mouse and humans. They concluded that the use and conservation of the AAUAAA sequence varied between the 6 species with this signal being particularly weak in plants and yeast. They favoured a model where the polyadenylation signal consists of a series of elements where no one element is universally required. A lack of one element could be compensated by the presence of strong words in other elements. Graber et al (1999) proposed 5 sequence elements for plants, in order; the Upstream element (UE) (UUGUAU or UUGUAA), the PE (AAUAAA or AAUGAA=A rich), the U-rich (UUUUCU or UUUUUU or similar) the CS (UA or UC) and a second U-rich region. Thus plants in comparison to animals have an additional upstream element that contributes to the definition of the polyadenylation signal. Mogen et al. (1990) reported that deletions of upstream elements of the Cauliflower Mosaic Virus (CaMV) and the PeaRbsC polyadenylation regions reduced the efficiency of polyadenylation at the ‘correct’ site. A similar result was also reported by Sanfacon et al (2007), again on the CaMV polyadenylation signal.

Since the polyadenylation signals in plants are AT-rich and that prokaryotes lack these signals, genes from prokaryotes which are AT % rich frequently contain sequences that might be recognised as polyadenylation sequences. Thus presence of such ‘cryptic’ polyadenylation motifs in coding regions of transgenes has been attributed to poor expression of genes such as Bacillus thuringiensis genes in plants. Fischhoff et al (U.S. Pat. No. 7,741,118 B1) describes that removal of hexamer AATAAA-like motifs will improve gene expression. They provide a list of 16 potential polyadenylation motifs that should be reduced in frequency in the coding region of a transgene to improve expression in planta.

High-level expression of non-plant genes in plant is a critical agronomic issue. Therefore, there is a need to develop new methods to improve gene expression in plants notably by the provided method that introduces fewer modifications than proposed by some other gene modification methods known in the art.

DESCRIPTION OF THE INVENTION

The purpose of the invention is to provide a method for modifying a coding gene sequence, in particular when this coding gene sequence encodes a Bacillus thuringiensis insecticidal protein, in order to obtain in planta expression of the protein at a level significantly higher than the wild type gene sequence.

Another purpose of the invention is to provide a method for preparing a DNA construct comprising a modified gene sequence wherein the modified gene sequence is expressed at a level significantly higher than the wild type gene sequence in planta.

The present invention relates to a method of making a modified coding sequence encoding a non-plant protein, the method comprising:

• • a) identifying a coding sequence that encodes a non-plant protein;
  • b) identifying each polyadenylation motif sequence and its nucleic acid position in said coding sequence;
  • c) optimizing said coding sequence by codon substitution wherein the optimized coding sequence encodes for said non-plant protein; and
  • d) modifying said optimized coding sequence to obtain a modified gene sequence by introducing at least one polyadenylation motif sequence, as depicted in Table 1, in the optimized gene sequence, wherein the modified coding sequence comprises at least one polyadenylation motif and said modified coding sequence is encoding said non-plant protein.

The invention also relates to a method of modifying a coding sequence encoding a non-plant protein, comprising the steps of:

• • a) identifying each wild-type polyadenylation motif sequence and its position within said coding sequence;
  • b) optimizing said coding sequence by codon substitution, thereby obtaining an optimized coding sequence which encodes said non-plant protein; and
  • c) introducing at least one polyadenylation motif sequence as disclosed in Table 1 within said optimized gene sequence, thereby obtaining a modified coding sequence which comprises at least one polyadenylation motif and said modified coding sequence is encoding said non-plant protein.

In a preferred embodiment, and as disclosed below, the polyadenylation that is introduced in the optimized gene sequence is preferentially a weak polyadenylation motif, as defined below.

Furthermore, no strong polyadenylation, as defined below, is introduced in the optimized gene sequence.

Furthermore, the number of polyadenylation motif sequences introduced in the optimized gene sequence is such that the total number of polyadenylation motif sequences in the modified sequence is three or higher, but less than the number of polyadenylation motif sequence that were present in the wild-type sequence.

As will be described below, the modified sequence will thus contain a combination of at least three polyadenylation, identical to a polyadenylation combination that is present in the wild-type sequence, and the modified sequence will not comprise the combination of the totality of the polyadenylation motifs that are present in the wild-type sequence.

The purpose of the modification of the optimized gene sequence by reintroduction of polyadenylation motifs, as described herein, is that this would either improve or maintain expression of the sequence, as compared to the optimized sequence. The examples show, however, that re-introduction of polyadenylation motifs in the optimized sequence in order to obtain a modified sequence comprising all the polyadenylation motifs as present in the wild-type sequence may, in some instances, decrease the expression with regards to the optimized sequence. Introducing only some polyadenylation motifs, in particular the weak ones, in order to have at least 3 (and preferably at most 10 or at most 6) polyadenylation motifs in the modified sequence, but not all polyadenylation motifs shall thus make it possible to have a robust method that can be applicable to and repeated with virtually any sequence.

In this embodiment, the invention thus relates to a method of making a modified coding sequence encoding a non-plant protein, the method comprising:

• • a) identifying a coding sequence that encodes a non-plant protein;
  • b) identifying each polyadenylation motif sequence and its nucleic acid position in said coding sequence;
  • c) optimizing said coding sequence by codon substitution wherein the optimized coding sequence encodes for said non-plant protein; and
  • d) modifying said optimized coding sequence to obtain a modified gene sequence as disclosed in Table 1 by introducing at least one polyadenylation motif sequence in the optimized gene sequence, so as to obtain a modified coding sequence that comprises at least three polyadenylation motifs, but not all the polyadenylation motifs identified in step b), and said modified coding sequence is encoding said non-plant protein.

Likewise, the invention also relates to a method of modifying a coding sequence encoding a non-plant protein, comprising the steps of:

• • a) identifying each wild-type polyadenylation motif sequence and its position within said coding sequence;
  • b) optimizing said coding sequence by codon substitution, thereby obtaining an optimized coding sequence which encodes said non-plant protein; and
  • c) introducing at least one polyadenylation motif sequence as disclosed in Table 1 within said optimized gene sequence, so as to obtain a modified coding sequence that comprises at least three polyadenylation motifs, but not all the polyadenylation motifs identified in step a), and said modified coding sequence is encoding said non-plant protein.

These methods are preferably performed with one or more of the following characteristics, that can be implemented independently or through any combination:

• • the polyadenylation motif sequence introduced in step d) (or c) is not a strong polyadenylation motif sequence (as described below)
  • the polyadenylation motif sequence introduced in step d) (or c) is a weak polyadenylation motif sequence (as described below)
  • the polyadenylation motif sequence is one that has been identified in step b) and is reintroduced at its position identified in step b) (or a)
  • the modified sequence contains at most 6 polyadenylation motif sequences
  • the modified sequence contains at most 10 polyadenylation motif sequences.
  • All of the at least 3 polyadenylation motif sequences in the modified sequence were initially present in the wild-type sequence and are localized at the location they had in the wild-type sequence.

In a preferred embodiment, said polyadenylation motif sequence introduced within said optimized gene sequence has been identified in step a), and in this case, it is further preferred when said polyadenylation motif sequence identified in step a) is introduced at its position within said optimized gene sequence.

In a further embodiment, when more than one polyadenylation motif is introduced within the optimized sequence, each polyadenylation motif introduced within said optimized sequence is a wild-type polyadenylation motif identified in step a), which is introduced at a nucleic acid position corresponding to its position within the coding sequence.

The present invention encompasses the identification of a coding gene sequence. This coding gene sequence is the wild-type coding sequence which is isolated or identified from the organism where the protein is naturally expressed.

Specifically, the invention encompasses the use of exogenous coding sequences encoding non-plant proteins. Preferably, the non-plant protein is an insecticidal protein encoded by Bacillus thuringiensis.

The coding gene sequence can be a fragment of the wild-type coding gene sequence. For example, the coding gene sequence can be a sequence encoding for the toxin fragment of a Bacillus thuringiensis protein. Also the coding sequence can encode for a fusion between two protein fragments obtained from different wild-type proteins.

Where appropriate, the coding sequence may be optimized for increased expression in the transformed plant. There are a number of optimizations that can be performed at the DNA level, without changing the protein sequence, by conservative codon exchanges which replace one codon by another codon encoding the same amino acid.

The parameters that can be optimized are for example, codon usage, local GC content, absence of splice sites, mRNA secondary structure, polyadenylation motifs.

More specifically, the genes can be synthesized using plant-preferred codons for improved expression, or may be synthesized using codons at a plant-preferred codon usage frequency. As a consequence, the GC content of the gene will often be increased. Methods to achieve such optimization for expression are well known in the art and are notably described in Campbell and Gowri (1990) for a discussion of host-preferred codon usage and more specifically for synthesizing plant-preferred genes (WO91/16432, and Murray et al. (1989)). WO 91/16432 describes in particular a process for modifying a Bt ICP gene to improve its expression in a plant cell, transformed with the gene; the process comprising the step of: changing A and T sequences in a plurality of translational codons of the gene to corresponding G and C sequences encoding the same amino acids, so as to improve the gene's transcription to an mRNA, the nuclear accumulation of the mRNA and/or the nuclear export of the mRNA, particularly the gene's transcription, in the plant cell.

In order to perform such optimization, different algorithms are available to predict the position of polyadenylation motifs in plant genes. For example, Ji et al (2015) have developed the algorithm PASPA (PolyA Site Prediction in Plants and Algae; http://bmi.xmu.edu.cn/paspa). Other algorithms are also available like PAC (poly(A) site classifier (Wu et al 2012) and polyA-iEP (Tzanis et al (2011). The baseline of this system is that motifs are identified with a given level of probability to represent a polyadenylation motif. The polyadenylation motifs harboring a high level of probability to represent a polyadenylation sites in the wild type sequence can then be removed from the optimized sequence.

It is intended here that polyadenylation motifs in the present invention consist of the following 16 motif sequences: AAAATA, AACCAA, AAGCAT, AATAAA, AATAAT, AATACA, AATCAA, AATTAA, ATAAAA, ATACAT, ATACTA, ATATAA, ATGAAA, ATTAAA, ATTAAT and CATAAA. Each polyadenylation motif can be referred by a specific sequence as described by the polyA code in Table 1. Furthermore, each polyadenylation motif comprised in a coding sequence can be characterized not only by its sequence but also by its nucleic acid position in the coding sequence, for instance “located between nucleotide X and nucleotide Y of the coding sequence”, or “whereas the first nucleotide of the polyadenylation motif is located at position X within the coding sequence”.

Consequently, a coding sequence encoding for a non-plant protein can be characterized by a combination of polyadenylation motifs wherein a polyA sequence and a nucleic acid position within the coding sequence can be assigned to each polyadenylation motif identified in the coding sequence.

Optimization of a coding sequence by the methods described above can lead to the removal of one or more polyadenylation motifs as listed in Table 1. To some extent, all the polyadenylation motifs can be removed so that the optimized coding sequence is free of any polyadenylation motifs. However, complete removal of polyadenylation motifs can impose great constraints on other sequence variables, notably on the amino acid sequence (i.e. a loss of identity of the protein obtained from the optimized sequence with regards to the wild-type protein obtained from the non-optimized wild-type sequence).

The applicant has shown that surprisingly, it is possible to reintroduce polyadenylation motifs in the optimized gene sequence and that one will still observe a reduction in the level of polyadenylation motifs initially calculated and predicted from the wild type sequence for most of the polyadenylation motifs.

Without being bound by this theory of the mechanism sustaining this phenomenon, the inventors suppose that the optimization of the coding gene sequence is modifying the surrounding motifs acting in combination with polyadenylation motifs to reduce gene expression so that when the polyadenylation motifs are reintroduced in the optimized coding gene sequence, they lose their ability to reduce gene expression.

It is to be noted that the polyadenylation site(s) that is (are) introduced within the optimized sequence was (were) not necessarily present in the wild-type sequence. Furthermore, such site(s) is (are) not necessarily introduced at the same location of the wild-type polyadenylation sites in the wild-type sequence.

The person skilled in the art can determine which and how many polyadenylation motifs with a specific polyA code can introduced in the optimized coding sequence to perform the invention, while maintaining the wild type protein sequence.

In a preferred embodiment, each polyadenylation motif that is introduced in the optimized coding sequence is introduced at a nucleic acid position identical to a wild-type polyadenylation site position as identified in the wild type coding sequence. In a most preferred embodiment, the polyadenylation motif that is introduced within the optimized sequence is the wild-type polyadenylation motif, introduced at its natural (wild-type) position.

In these two embodiments, the protein encoded by the modified coding sequence is the wild-type protein, encoded by the wild-type, non-optimized sequence.

Most preferably, the method of the invention relates to a method of modifying a coding sequence wherein all the introduced polyadenylation motifs are introduced at a nucleic acid position identical to their positions as identified in the wild type coding sequence and wherein the resulting modified coding sequence comprises the same combination of polyadenylation motifs as identified in the wild type coding sequence.

In this embodiment, wild-type polyadenylation motifs as identified within the coding sequence are introduced at nucleic acid positions identical to their positions within the coding sequence, so as to obtain a modified coding sequence comprising the same combination of polyadenylation motifs as the wild type coding sequence.

Such a modified sequence is encoding the wild-type non-plant protein and is able to be expressed at a level significantly higher compared with the non-optimized wild-type sequence even though it comprises all the polyadenylation motifs as present in the wild-type coding sequence.

The method of the invention also relates to a method of modifying a coding sequence wherein all the introduced polyadenylation motifs are introduced at a nucleic acid position identical to their position as identified in the wild type coding sequence, the resulting modified coding sequence shall thus comprise a combination of one, two, or three polyadenylation motifs as identified in the wild type coding sequence. In such cases, the modified coding sequence is expected to not comprise the combination of all the polyadenylation motifs as identified in the wild-type coding sequence.

In a specific embodiment, wild-type polyadenylation motifs as identified within the coding sequence are introduced in step c) at nucleic acid position identical to their position within the coding sequence, so as to obtain a modified coding sequence comprising one polyadenylation motif present in the wild type sequence.

In another embodiment, wild-type polyadenylation motifs as identified within the coding sequence are introduced in step c) at nucleic acid position identical to their position within the coding sequence, so as to obtain a modified coding sequence comprising a combination of two polyadenylation motifs present in the wild type sequence.

It is to be noted that, in order to obtain a modified coding sequence comprising a combination of two polyadenylation motifs present in the wild type sequence, one will introduce one or two polyadenylation motifs present in the wild type sequence within the optimized sequence, depending on whether said optimized sequence already comprises one or zero of such polyadenylation motifs present in the wild type sequence.

In another embodiment, wild-type polyadenylation motifs as identified within the coding sequence are introduced in step c) at nucleic acid position identical to their position within the coding sequence, so as to obtain a modified coding sequence comprising a combination of three polyadenylation motifs present in the wild type sequence.

It is to be noted that, in order to obtain a modified coding sequence comprising a combination of three polyadenylation motifs present in the wild type sequence, one will introduce one, two or three polyadenylation motifs present in the wild type sequence within the optimized sequence, depending on whether said optimized sequence already comprises two, one or zero of such polyadenylation motifs present in the wild type sequence.

It is expected that polyadenylation motifs have different effects on the level of gene expression whatever the gene sequence considered. Polyadenylation motifs can be featured as strong or weak motifs. Weak motifs are expected to be more tolerated for the expression of the coding gene sequence than strong motifs. Advantageously, the polyadenylation motifs introduced in the optimized gene sequence are the polyadenylation motifs identified as weak motifs.

Polyadenylation motifs have been ranked from the weakest to the strongest regarding their occurrence in maize coding gene sequences and regarding their occurrence in optimized gene sequences shown to be well expressed in plants.

Preferably the weak polyadenylation motifs are chosen amongst the AAAATA, AAGCAT, AATCAA, and ATGAAA sequences. The strong motifs are chosen amongst the AATTAA, ATACAT, ATACTA, ATATAA, ATTAAA, ATTAAT and CATAAA sequences.

More preferably, the rank amongst the weak polyadenylation motifs are from the weaker to the less weak: ATGAAA, AATCAA, AAAATA and AAGCAT. The rank amongst the strong polyadenylation motifs are from the stronger to the less strong: ATACTA, ATTAAT, AATTAA, ATTAAA, CATAAA, ATATAA, ATACAT.

Most preferably, the rank from the weaker to the stronger polyadenylation motifs is provided in Table 1.

Upon comparison of the coding gene sequences from monocotyledons and dicotyledons, one can note that the common strongest polyadenylation motifs are the AATTAA, ATACTA, ATATAA, ATTAAA, ATTAAT and CATAAA motifs.

Another weak polyadenylation motif (poladenylation motif AACCAA) can also be identified.

In one embodiment of the invention, none of the strongest polyadenylation motifs AATTAA, ATACTA, ATATAA, ATTAAA, ATTAAT and CATAAA are added in the optimized sequence. It is, however, to be noted that the modified sequence may contain one (or more of these polyadenylation motifs, if they are present in the optimized sequence, after the native (wild-type) sequence has been optimized by codon substitution)

Most preferably, the polyadenylation motifs that are added to the optimized sequence are chosen in the group consisting of ATGAAA, AATCAA, AAAATA, AACCAA and AAGCAT. Consequently only these weak polyadenylation motifs are added in the optimized sequence.

In a further embodiment of the invention, the final modified sequence comprises a total of three to a few more polyadenylation motifs wherein at least three polyadenylation motifs are corresponding to a combination of three polyadenylation motifs identified in the wild-type coding sequence, and wherein the modified coding sequence does not comprise all the polyadenylation motifs identified and present in the wild-type coding sequence.

Preferably, the modified optimized sequence comprises three to ten polyadenylation motifs.

Preferably, the modified optimized sequence comprises three to six polyadenylation motifs.

The step of reintroducing polyadenylation motifs might create changes in the amino acid at the edges of the polyadenylation motifs.

It can also create additional motifs that might reduce gene expression. Additional motifs can include for example cryptic splice sites GGTAAG, GGTGAT, GTAAAA and GTAAGT and/or polyA or polyT and/or repeats of 7 or more base pairs.

The present method may thus comprises further steps of modifications of the modified gene sequence in order to make sure that the protein sequence encoded by the modified sequence is identical to the one coded by the wild-type coding gene sequence. The person skilled in the art knows the different type of motifs to be checked and, when appropriate, will modify the modified gene sequence so that the function of the protein encoded by the modified coding gene sequence is not altered compared to the function of the protein encoded by the wild type coding sequence. More preferably, the amino-acid sequence of the protein encoded by the modified coding gene sequence is identical to the amino-acid sequence of the protein encoded by the wild type coding sequence.

The present invention encompasses a modified coding sequence obtainable according to the method of making a modified coding sequence wherein polyadenylation motifs are introduced at a nucleic acid position identical to their position as identified in the wild type coding sequence and wherein the resulting modified coding sequence comprises the same combination of polyadenylation motifs as identified in the wild-type coding sequence.

It is intended in the present invention that all the steps of the method of making a modified coding sequence can be made by combining in silico sequence designing with the preparation of synthetic coding gene sequence. Most preferably, all the steps of the methods described above can be made in silico. Therefore, in such a case, the method of making an expression cassette as described below will require as a first step the synthesis of the corresponding modified coding gene sequence.

The invention thus encompasses a method making a nucleic acid molecule, comprising the steps of performing the methods as disclosed above and synthetizing the modified nucleic acid harboring the modified sequence as obtained.

Another embodiment of the invention is a method of making an expression cassette comprising a modified coding sequence encoding for a non-plant protein, the method comprising the steps of making a modified coding sequence encoding for said protein according to any method as described above and operably linking a promoter and a terminator to said modified coding sequence to obtain a construct for expression in plant.

More specifically, the invention relates to a method of making an expression cassette wherein polyadenylation motifs are introduced at a nucleic acid position identical to their position as identified in the wild type coding sequence and wherein the resulting modified coding sequence comprises the same combination of polyadenylation motifs as identified in the wild type coding sequence.

These cassettes may be obtained in silico or actually synthetized.

More preferably, the coding sequence is encoding a Bacillus thuringiensis insecticidal protein.

The term “operably linked” as used herein means that the promoter and the modified coding gene sequence are oriented such that the promoter directs expression of the coding sequence, generally in the 5′ to 3′ direction.

The constructs may also contain enhancers (such as introns) and terminators at the 3′ end of the coding sequence.

A promoter “active in plants” is a promoter that is able to drive expression of a gene operably linked thereto in a plant cell.

For being expressed, a coding sequence may be present under the control of a constitutive, tissue specific, developmentally regulated or inducible promoter.

Promoters may come from the same species or from another species (heterologous promoters). Although some promoters may have the same pattern of regulation when they are used in different species, it is often preferable to use monocotyledonous promoters in monocotyledons and dicotyledonous promoters in dicotyledonous plants.

In a preferred embodiment, said construct is under the control of a constitutive promoter.

Examples of constitutive promoters useful for expression include the 35S promoter or the 19S promoter (Kay et al., 1987, Science, 236:1299-1302), the rice actin promoter (McElroy et al., 1990, Plant Cell, 2:163-171), the pCRV promoter (Depigny-This et al., 1992, Plant Molecular Biology, 20:467-479), the CsVMV promoter (Verdaguer et al., 1998, Plant Mol Biol. 6:1129-39), the ubiquitin 1 promoter of maize (Christensen et al., 1996, Transgenic. Res., 5:213), the regulatory sequences of the T-DNA of Agrobacterium tumefaciens, including mannopine synthase, nopaline synthase, octopine synthase.

Other suitable promoters could be used. It could be an inducible promoter or a tissue-specific promoter such as, for example, a leaf-specific promoter or a seed-specific. Numerous tissue-specific promoters are described in the literature and any one of them can be used. One can cite the promoters disclosed in US 20130024998.

In one embodiment, the protein encoded by the modified coding sequence is targeted to the chloroplast for expression. In this manner, the expression cassette will additionally contain a nucleic acid encoding a transit peptide to direct the protein to the chloroplasts. Such transit peptides are known in the art. See, for example, Von Heijne et al. (1991) Plant Mol. Biol. Rep. 9:104-126; Clark et al. (1989) J. Biol. Chem. 264:17544-17550; Della-Cioppa et al. (1987) Plant Physiol. 84:965-968; Romer et al. (1993) Biochem. Biophys. Res. Commun. 196:1414-1421; and Shah et al. (1986) Science 233:478-481.

The invention also encompasses a vector containing the expression cassette (nucleic acid construct) of the invention.

A vector, such as a plasmid, can thus be used for transforming host cells. The construction of vectors for transformation of host cells is within the capability of one skilled in the art following standard techniques.

The decision as to whether to use a vector for transforming a cell, or which vector to use, is guided by the method of transformation selected, and by the host cell selected.

Possible vectors include the Ti plasmid vectors, shuttle vectors designed merely to maximally yield high numbers of copies, episomal vectors containing minimal sequences necessary for ultimate replication once transformation has occurred, transposon vectors, including the possibility of RNA forms of the gene sequences. The selection of vectors and methods to construct them are commonly known to persons of ordinary skill in the art and are described in general technical references (Mullis, K B (1987), Methods in Enzymology).

For transformation methods within a plant cell, one can cite methods of direct transfer of genes such as direct micro-injection into plant embryos, vacuum infiltration or electroporation, direct precipitation by means of PEG or the bombardment by gun of particles covered with the DNA of interest.

It is preferred to transform the plant cell with a bacterial strain, in particular Agrobacterium, in particular Agrobacterium tumefaciens. In particular, it is possible to use the method described by Ishida et al. (Nature Biotechnology, 14, 745-750, 1996) for the transformation of Monocotyledons.

The cell can be transformed by any transient expression system, some of which are well known from the person skilled in the art.

In a specific embodiment, said expression cassette is stably integrated within the genome of said host cell. This embodiment is particularly interesting for plant host cells. Stable integration within the genome means that the expression cassette can be transmitted to the progeny of said host cell upon division.

The invention also encompasses a plant containing at least one cell containing the expression cassette as defined above, preferably stably integrated within its genome.

A part of a transgenic plant, in particular fruit, seed, grain or pollen, comprising such a cell or generated from such a cell is also encompassed by the invention. The host cell as mentioned above, part of the invention, is also considered as being a part of a transgenic plant.

It is reminded that a whole plant can be regenerated from a single transformed plant cell. Thus, in a further aspect the present invention provides transgenic plants (or parts of them) including the expression cassette according to the invention. The regeneration can proceed by known methods.

The seeds which grow by fertilization from this plant, also contain this transgene in their genome.

Said plant or part of a plant according to the invention can be a plant or a part of it from various species, notably an Angiosperm, Monocotyledons as Dicotyledons.

Said plant is preferably selected from the group consisting of maize wheat, barley, rape, sugar beet and sunflower. In a preferred embodiment, said plant is maize.

DESCRIPTION OF THE FIGURES

FIG. 1A, FIG. 1B, and FIG. 1C illustrate the steps of various embodiments to perform a wild type gene modification to express the protein at high level in plant.

FIG. 2A, FIG. 2B, FIG. 2C, and FIG. 2D illustrate the application of the method on Bacillus thuringiensis Axmi028 coding sequence.

FIG. 3A, FIG. 3B, FIG. 3C, and FIG. 3D illustrate the PASPA results on Axmi028

FIG. 4A, FIG. 4B, FIG. 4C, and FIG. 4D illustrate the application of the method on Bacillus thuringiensis Axmi100 coding sequence.

FIG. 5A, FIG. 5B, FIG. 5C, and FIG. 5D illustrate the PASPA results on Axmi100

FIG. 6A and FIG. 6B show luciferase expression levels of the Axmi028-LUC fusions (FIG. 6A) and of the Axmi100-LUC fusions (FIG. 6B) in transient assays in maize leaves.

FIG. 7A and FIG. 7B show Western Blots analysis of independent Axmi028 maize transformants using a C Myc Tag antibody (FIG. 7A) and independent Axmi100 maize transformants using a polyclonal antibody (FIG. 7B). In FIG. 7A, lane 1, molecular size markers; lane 2, non-transformed maize; lanes 3-4, 028-WT+GUS transformants; lanes 5-6, 028-opt+GUS transformants; lanes 7-8, 028-opt+pA+GUS transformants and lanes 9-10; 028-opt+3pA+GUS transformants. 10 μg protein was loaded for each plant sample.

In FIG. 7B, lanes 1, 26, 27 and 52, molecular size markers; lanes 2-11, 100-WT+GUS transformants; lanes 12 and 38, non-transformed maize; lanes 13 and 39; protein extract from a Axmi100-expressing bacterial strain; lanes 14-23, 100-opt+pA+GUS transformants; lanes 28-37, 100-opt+3pA+GUS transformants and lanes 40-49, 100-opt+GUS transformants. 5 μg protein was loaded for each plant sample.

FIG. 8 depicts the relative expression intensity of the full length protein Axmi028 in maize protoplast transient assay. Within every lane the intensity of the detected c-Myc and His tags signals was set into relation to the nptII signal as calculated via the software Image Lab 5.2.1, BioRad.

FIG. 9A and FIG. 9B show the RNA expression level by QRT-PCR of Axmi028 and Axmi100 optimized genes in maize transformants. FIG. 9A) 3′ primer pair, FIG. 9B) 5′ primer pair.

EXAMPLES

Example 1

Sequences for Improvement of Gene Expression Via Gene Optimization of Regions Flanking Polyadenylation Motifs

Axmi028

The wild-type coding region of an AT rich gene, Axmi028 from Bacillus thuringiensis (U.S. Pat. No. 8,314,292 B2) lacking the C-terminal crystal domain, was analysed for the polyadenylation motifs as listed in Table 1.

This truncated wild-type coding sequence (028-WT; SEQ ID NO: 1) contains 31 such sites (FIG. 2A). First a codon-optimized Axmi028 sequence was synthesized, as described in U.S. Pat. No. 8,314,292 B2, which lacks 30 of the 31 polyA sites as described in SEQ ID NO: 2 (028-opt FIG. 2B). This sequence was optimized for maize expression. A further sequence (028-opt+pA SEQ ID NO: 3; FIG. 2C) was designed where the 30 putative polyadenylation motif sequences were reintroduced into this optimized 028-opt sequence such that the amino-acid sequence of the Axmi028 sequence is conserved.

In two cases the reintroduction of the polyadenylation motif resulted in a change of the amino-acid sequence. This was corrected by introducing 3 more base pairs of wild-type sequence 5′ of these polyadenylation motifs. In addition the sequence was examined for the presence of additional sequences that might reduce expression which may have been created by the juxtaposition of the optimized and polyadenylation motifs (cryptic splice sites GGTAAG, GGTGAT, GTAAAA and GTAAGT and/or polyA and polyT sequences and/or 7 or more repeated base pairs). No such motifs were found in the 028-opt+pA sequence. This process is outlined in FIG. 1A.

Several attempts have been made in the art to predict the position of polyadenylation sites in plant genes. Ji et al (2015) have developed the algorithm PASPA (PolyA Site Prediction in Plants and Algae; http://bmi.xmu.edu.cn/paspa). This algorithm was applied to the 028-WT, 028-opt and 028-opt+pA sequences using parameters defined for Rice. FIGS. 3A through 3D show the output of this analysis. It can be seen that the likelihood of premature polyadenylation as predicted by PASPA is greatest for the 028-WT sequence (FIG. 3A) and least for the 028-opt sequence (FIG. 3B). The 028-opt+pA sequence (FIG. 3C) has a significant reduction in the likelihood of premature polyadenylation compared to the 028-WT sequence.

Axmi100

A second AT rich Bacillus thuringiensis gene, Axmi100 (US20100005543) was also modified. As for Axmi028, for expression in plants, the C-terminal crystal domain was removed. This truncated wild-type coding sequence, as described in SEQ ID NO: 4 (100-WT, FIG. 4A), contains 25 polyadenylation motifs. A codon-optimized Axmi100 sequence was synthesized (US20100005543) which lacks 24 of the 25 wild-type polyadenylation motifs (100-opt or SEQ ID NO: 5, FIG. 4B). This sequence, optimized for maize expression, however has 5 additional polyadenylation motifs not present in the wild-type sequence. A further sequence, as described in SEQ ID NO: 6 (100-opt+pA FIG. 4C), was synthesized where the 24 putative polyadenylation motifs were reintroduced into this optimized 100-opt sequence at the same nucleic position as identified in the wild type coding sequence. This sequence conserves the amino-acid sequence of Axmi100 and contains no cryptic splice site, polyA or polyT motifs (cryptic splice sites GGTAAG, GGTGAT, GTAAAA and GTAAGT and/or polyA and polyT sequences and/or 7 or more repeated base pairs).

The PASPA algorithm was applied to the 100-WT, 100-opt and 100-opt+pA sequences using parameters defined for Rice. FIGS. 5A through 5D show the output of this analysis. It can be seen that the likelihood of premature polyadenylation as predicted by PASPA is greatest for the 100-WT sequence (FIG. 5A) and least for the 100-opt sequence (FIG. 5B). However the 100-opt+pA sequence (FIG. 5C) has a significant reduction in the likelihood of premature polyadenylation compared to the 100-WT sequence.

Example 2

Transient Expression Testing of Genes Optimized in Regions Flanking Polyadenylation Motifs in Maize and Tobacco

Axmi028

Two transient systems were employed. The first was an indirect assay system where the three different Axmi028 sequences (028-WT, 028-opt and 028-opt+pA) were fused in frame to the reporter firefly luciferase gene (LUC) and placed under the control of the constitutive maize Ubiquitin promoter (SEQ ID NO: 7, SEQ ID NO: 8 and SEQ ID NO: 9). The rationale is that any premature polyadenylation in the Axmi028 sequence will terminate the transcript preventing the possibility to create a transcript containing the full Axmi028+Luc fusion. A reduction in Luc signal from the 028-WT-Luc or 028-opt-pA-Luc fusion genes compared to the 028-opt-Luc control may then be attributed to an increased occurrence of premature polyadenylation. Plasmids containing these fusions are co-bombarded into maize leaf tissue with a control 35S-Renilla luciferase construct. 24 hrs later the luminescence of the firefly and Renilla luciferases is measured and the signal from the firefly luc normalised using the control Renilla luc signal. The normalised firefly luc signal from the 028-WT-Luc gene is then compared to that from the 028-opt-Luc and 028-opt-pA-Luc genes (FIG. 6A).

The second transient system is by agro-infiltration of binary plasmid constructs containing Axmi028 versions into the tobacco N. benthamiana. The 028-WT, 028-opt and 028-opt+pA genes driven from the constitutive viral CsVMV promoter (Verdaguer et al (1996)) are cloned into an SB11-derived binary vector (Komari et al (1996)) that also contains the fluorescent reporter gene AnCyan (CloneTech) expressed from the constitutive maize Ubiquitin promoter forming the plasmids 028-WT+Cyan, 028-opt+Cyan and 028-opt+pA+Cyan. These three binary vectors plus the empty SB11+Cyan control are transferred into the agrobacterium strain LBA4404 (pSB1)) according to Komari et al (1996). Agro-infiltration is performed with these 4 strains essentially as described by Leckie and Steward (2011). Four leaves of five plants are infiltrated, each leaf being infiltrated with the four strains in different parts of the leaf. After 3 days the zones expressing AnCyan are visualised, then excised. The zones infiltrated with the same agrobacterial strain in each plant are pooled and frozen in liquid nitrogen. Samples are taken for the measurement of transcript levels of the Axmi28 gene and the AnCyan gene by QRT-PCR and for Western analysis using antibodies against Axmi028 and AnCyan. Primer pairs for QR-PCR analysis are designed in the 3′ region of the coding sequences of the Axmi028 gene sequences. The transcript expression of 028-WT/AnCyan is then compared to that of 028-opt/AnCyan and to that of 028-opt+pA/AnCyan in order to determine the effect of the termination of transcription by the use of cryptic polyadenylation motifs prior to the position of the primers used for the QRT-PCR reaction.

Similar results can be obtained when the level of Axmi028 protein, normalised for AnCyan protein expression, is compared between the three Axmi028 constructs.

Additional gene constructs were made where Axmi028 versions each have an additional N-terminal His TAG and a C-terminal C-Myc TAG allowing visualization of the Axmi028 proteins in Western blots using HisTAG or C-MycTAG antibodies. These Axmi028 versions are 028-h(WT)m (SEQ ID NO: 17), 028-h(opt)m (SEQ ID NO: 18) and 028-h(opt+pA)m (SEQ ID NO: 19). The 028-h(WT)m, 028-h(opt)m and 028-h(opt+pA)m genes driven from the constitutive viral CsVMV promoter (Verdaguer et al (1996)) were cloned into an SB11-derived binary vector (Komari et al (1996)) that also contains the beta glucuronidase (GUS) reporter gene (Jefferson et al, 1987)) expressed from the constitutive maize Ubiquitin promoter forming the plasmids 028-WT+GUS, 028-opt+GUS and 028-opt+pA+GUS. These three binary vectors plus the empty SB11+GUS control were transferred into the agrobacterium strain LBA4404 (pSB1)) according to Komari et al (1996). As described above transient assays are performed in N. benthamiana. Protein and RNA Samples are also extracted from 20 to 25 immature maize embryos co-cultivated for 7 days with the agrobacterial strains containing the different Axmi028+GUS constructs. To compensate for potential differences in T-DNA delivery during co-cultivation between the different samples GUS fluorimetrical activity assays using 4-methylumbelliferyl-beta-D-glucuronide (MUG) were performed on each protein sample. Protein amounts used in Westerns were then adjusted to give an equal GUS activity per sample. As for the transient expression analysis in N. benthamiana, analysis of these samples allows the comparison of expression of the different Axmi028 versions.

Axmi100

In an identical fashion as described above for Axmi028 the 100-WT, 100-opt and 100-opt+pA sequences are tested by transient assays in maize (SEQ ID NO: 10, SEQ ID NO: 11 and SEQ ID NO: 12) and tobacco. The expression of 100-WT is thus compared to that of 100-opt and to that of 100-opt+pA (FIG. 6B). Western blot analysis was made on the same samples by using a polyclonal antibody raised against Axmi100 protein. The result are in line with the result depicted in FIGS. 6A and 6B: the presence of polyadenylation signals makes it possible to obtain good expression of the protein, better than the expression of the optimized protein in which no polyadenylation have been added, and the expression of the optimized protein in which all wild-type polyadenylation have been added, whereas the wild-type protein is not properly expressed (data not shown).

The Axmi100 versions are also expressed as N-terminal His-Tag and C-terminal C-Myc Tag versions; 100-h(WT)m, 100-h(opt)m and 100-h(opt+pA)m (SEQ ID NO: 20, SEQ ID NO: 21 and SEQ ID NO: 22) in transient assays in tobacco and in immature maize embryos. The expression of 100-h(WT)m is thus compared to that of 100-h(opt)m and to that of 100-h(opt+pA)m.

Example 3

Stable Expression in Maize of Genes Optimized in Regions Flanking Polyadenylation Motifs

Axmi028

The strains described in example 2 (028-WT+GUS, 028-WT+Cyan, 028-opt+GUS, 028-opt+Cyan, 028-opt+pA+GUS and 028-opt+pA+Cyan) are transformed into maize essentially as described by Ishida et al (1996). A minimum of 10 individual, single copy transformants with an intact T-DNA, are produced for each construct. QRT-PCR and Western analyses are performed on TO leaf material. Leaf Axim028 expression and protein levels of the 028-WT plants are compared to the 028-opt and 028-opt+pA transformants as in the previous example (FIG. 7A).

Axmi100

As described above for Axmi028, the different versions of Axmi100 are transformed into maize. Leaf Axmi100 expression and protein levels of the 100-WT plants are compared to the 100-opt and 100-opt+pA as in the previous example (FIG. 7B).

Example 4

Identification of Weak Polyadenylation Motifs that can Remain in Codon-Optimized Sequences

A further improvement to the above procedure is to leave only weak polyadenylation motifs in the optimized sequence. Although reintroducing all polyadenylation motifs identified in Table 1 in the optimized sequence significantly improves expression to levels similar to that obtained by a fully optimized sequence the procedure may not be optimal in all cases. This is since as the number of polyadenylation motifs increases in the wild-type sequence the more of the sequence cannot be optimized and the more potential exists for undesirable sequences created at the junctions of optimized and polyadenylation sequences. An in silico approach was used to identify weak and strong polyadenylation sequences in maize. This approach is based on the idea that strong polyadenylation motifs will be under-represented in the coding sequences of maize genes and particularly so in highly expressed genes. Conversely weak polyadenylation motifs should not be under-represented. However the occurrence of a motif may also be dependent on the amino-acids it can encode. Motifs that ‘encode’ amino-acids used frequently and with codons frequently used for that amino-acid will be overrepresented. Thus keeping the ‘weak’ motifs that are the most over-represented compared to the theoretical calculation should select motifs that both:

• • a) Are not strong polyadenylation signals
  • b) Are frequently used since they encode amino-acids that are frequently used/or codons that are frequently used in maize.

TABLE 1
occurrence of polyadenylation motifs in maize coding sequence
Motif Occurrence in Maize CDS v3
Motif	PolyA code	Theoretical	Real	% Real
ATGAAA	polyA8	10081	21568	214%
AATCAA	polyA5	10081	16116	160%
AAAATA	polyA12	8265	12978	157%
AAGCAT	polyA9	12297	18669	152%
AACCAA	polyA3	12297	15522	126%
ATAAAA	polyA7	8265	9660	117%
AATAAT	polyA2	8265	9297	112%
AATAAA	polyA1	8265	9276	112%
AATACA	polyA15	10081	10519	104%
ATACAT	polyA11	10081	9945	99%
ATATAA	polyA4	8265	7009	85%
CATAAA	polyA16	10081	7884	78%
ATTAAA	polyA13	8265	6364	77%
AATTAA	polyA14	8265	6167	75%
ATTAAT	polyA10	8265	5771	70%
ATACTA	polyA6	10081	6638	66%

a) Analysis of CDSs in Maize CDS Database v3:

First the entire predicted coding regions of maize were analysed (maize CDS database v3, ftp://ftp.ensemblgenomes.org/pub/release-27/plants/fasta/zea_mays/cds/). The predicted number of each polyadenylation motif was determined in this dataset using the observed size of the dataset (63279365 bp) and the base-pair composition of this dataset (54.95% GC). Then the actual number of occurrences was determined and the ratio of real/predicted occurrences calculated (see Table 1). Results show that some polyadenylation motifs are significantly under-represented and some significantly overrepresented.

In the maize CDSv3 dataset 4 motifs are 150% or more over-represented. Those that are over-represented are candidates for sequences that are weak polyadenylation motifs and sequences that allow good gene expression. These polyadenylation sequences can be left within optimized sequences with a low probability that they will compromise gene expression. This protocol is outlined in FIG. 1B.

b) Analysis of CDS in Monocotyledons:

A crop-specific search for polyadenylation motifs as listed in table 1 was performed to define those motifs that occur with high levels in the CDS of the respective crop of interest. For that purpose CDS of 2 defined corn lines (B73, AGPv3.22) were analyzed postulating that, as in the previous example, the CDS contain codons that are frequently used to encode certain amino acids in the crop. Polyadenylation motifs that are present within codons of the CDS are not strong, but only minor-functional or likely non-functional. So those naturally occurring motifs can remain in any transgene as they would not influence its stable expression in the crop of interest and will be referred to as crop-specific weak motifs.

First the presence of polyadenylation motifs was analyzed in those different corn datasets. For count checks, OligoCounter (http://webhost1.mh-hannover.de/davenport/oligocounted, Tümmler laboratory at Hannover Medical School, Germany) and J Browse (Skinner et al, Genome Res. 2009. 19: 1630-1638) were used to check motif distributions in the genomes. Those counts were normalized to the total number of predicted transcript per CDS dataset.

Table 2 is showing the percentage of CDS that contain the given polyadenylation motif in for B73 and AGPv3.22 data set. To facilitate the comparison, the column “motif occurrence maize CDSv3” from table 1 is added. It represents in percentage the actual motif occurrence in the entire maizev3 dataset divided by the theoretical occurrence in the dataset. The distribution of polyadenylation motifs was very similar in all corn datasets with little variations in their rankings. It is interesting to note that the top 5 polyadenylation motifs ATGAAA, AAGCAT, AACCAA, AATCAA and AAAATA occur in corn transcripts with relatively high frequency (>10% of CDS containing these motifs in dataset B73). This result is consistent with the frequencies found in experiment described above in section a).

Even though this analysis is not considered as providing an exhaustive list of all the weak polyadenylation motifs, it can be concluded that the 5 polyadenylation motifs identified are confirmed as weak motifs in corn.

In addition to the corn datasets, another monocotyledon crop Sorghum bicolor (Sbicolor_255_v2.1) was analyzed (Table 2). Analogously to corn these CDS sets were analyzed for their total abundance of polyadenylation motifs counts (see table 2). Those counts were normalized to the total number of predicted transcript per CDS dataset. Remarkably, the two different monocot crops show very similar relative abundance of all polyadenylation motifs with the top five most abundant motifs being exactly the same.

According to those data the polyadenylation motifs ATGAAA, AAGCAT, AACCAA, AATCAA and AAAATA can remain in any transgene expressed in monocotyledons as they would not influence its stable expression in the crop of interest and will be referred to as monocot-specific weak polyadenylation motifs.

Interestingly, the six strongest ATAAAA, CATAAA, ATACTA, ATTAAA, AATTAA, and ATTAAT polyadenylation motifs are also consistent amongst monocotyledons.

c) Comparison between monocotyledons and dicotyledons weak polyadenylation motifs:

In addition to the monocotyledons datasets, a dicotyledon crop Beta vulgaris (RefBeet-1.2) was analyzed. Table 2 shows that B. vulgaris presents a similar distribution of motifs from the weakest to strongest motifs. One motif AATAAT was found more frequently in the CDS dataset then in those of monocotyledons CDSs. However, there is a clear overlap in the most abundant motifs between all crop datasets analyzed.

These data suggest that the three polyadenylation motifs ATGAAA, AAGCAT and AATCAA can likely remain in any transgene expressed in flowering plants.

The overall data shows that the identification of the five motifs ATGAAA, AAGCAT, AACCAA, AATCAA and AAAATA as weak polyadenylation motifs is robust in the plant kingdom.

TABLE 2
Occurrence of polyadenylation motifs in monocotyledons and
dicotyledons. The results shown in column “motif occurrence in maize
CDSvs” is an extraction from table 1.
		AGP	maize		S. bicolor		Ref Beet
	B73	v 3.22	CDSv3		255v2.1		1.2.
ATGAAA	13.72	34.10	214.00	ATGAAA	44.40	ATGAAA	78.70
AAGGAT	12.24	29.52	152.00	AAGCAT	38.65	AAGCAT	50.34
AACCAA	12.02	24.54	126.00	AACCAA	31.81	AACCAA	44.36
AATCAA	11.82	25.48	160.00	AATCAA	34.04	AATCAA	63.11
AAAATA	10.88	20.52	157.00	AAAATA	26.06	AAAATA	44.16
ATAAAA	8.12	15.27	117.00	ATAAAA	19.71	ATAAAA	36.30
ATACAT	8.05	15.73	99.00	ATACAT	20.31	ATACAT	28.90
AATACA	7.84	16.63	104.00	AATACA	21.14	AATACA	32.22
AATAAA	7.59	14.67	112.00	AATAAA	19.06	AATAAA	33.00
AATAAT	7.02	14.70	112.00	AATAAT	20.11	AATAAT	45.21
ATATAA	5.93	11.08	85.00	ATATAA	14.29	ATATAA	27.18
CATAAA	5.85	12.47	78.00	CATAAA	15.61	CATAAA	24.80
AATTAA	5.15	9.75	75.00	AATTAA	12.29	AATTAA	29.23
ATTAAA	4.87	10.06	77.00	ATTAAA	12.60	ATTAAA	30.86
ATACTA	4.71	10.50	66.00	ATACTA	13.63	ATACTA	24.49
ATTAAT	4.59	9.13	70.00	ATTAAT	11.02	ATTAAT	28.55

d) Analysis of Polyadenylation Motifs in Transgenes Expressed in Planta

In order to assess the presence of polyadenylation motifs in transgenes expressed in planta, 21 bacterial gene sequences and 5 gene sequences from eukaryotic organisms were analyzed. Those genes were shown to be expressed in planta. The total number of polyadenylation motifs was counted to identify those motifs in transgenes that were described as functional and/or expressed in planta (demonstrated either via analysis of transgene expression levels or via new phenotypes detected in transgenic plants). Based on those numbers of polyadenylation motifs counted in the transgenes, the abundance of any motif was calculated (number of motifs/number of genes analyzed). According to their calculated abundance across genes, species of origin & expressing crop the polyadenylation motifs were grouped: Signals with high abundance (≥50% in all genes analyzed) are rated as non-functional or very weak polyadenylation signals, those signals with medium abundance (≥25% in all genes analyzed) are rated as minor functional or weak polyadenylation signals. Signals with low abundance (≥0% in all genes analyzed) are rated as functional or strong polyA signals.

To assess if specific motifs were deleted in genes optimized for transgene expression in planta with higher frequency, the percentage of polyadenylation motifs that remained after optimization was calculated. Even though this analysis is not considered to provide an exhaustive identification of all weak polyadenylation motifs existing in plants, the retention of motifs in addition to high abundance in this variety of genes is a valuable indication for their weak impact on transcript stability. The two motifs AATCAA and ATGAAA show highest abundance across all genes and remained even in optimized sequences to a high percentage. In addition to those two motifs, the motif AAAATA also shows high abundance in the set of genes analyzed.

According to these data these polyadenylation motifs are unlikely to influence transgene expression in any crop of interest.

Example 5

Transient Expression Testing of Genes that are Maize Codon-Optimized but Contain a Minimum of Three Polyadenylation Motifs at the Wild-Type Position

a) Transient Expression System as Performed in Example 2: Axmi028

The Amxi028 wild-type but C-terminus truncated sequence was examined for the presence of weak polyadenylation motifs that can remain in the optimized sequence. Two ATGAAA and four AATCAA motifs were found, these are the most overrepresented motifs in the maize CDSv3 database. The optimized sequence already has an AATACA motif present at its wild-type position of 626 bp. This motif is neither underrepresented nor overrepresented in the maize CDS v3 dataset (104% of real/theoretical). Two of the four AATCAA motifs were introduced in the optimized sequence (positions 1036 bp and 1232 bp) giving in total three weak motifs that are identified in the wild-type position in the modified optimized coding sequence. This sequence 028opt+3pA as described in SEQ ID NO: 13 (FIG. 2D), does not comprise any of the strongest ATAAAA, CATAAA, ATACTA, ATTAAA, AATTAA, and ATTAAT polyadenylation motifs. It was analyzed by PASPA (FIG. 3D) and found to give an identical probability curve of polyadenylation as the fully optimized sequence (FIG. 3B). This example shows that the addition of weak motifs at their wild-type position in the optimized coding sequence has no effect on the level of predictability of the polyadenylation motifs as compared to that of the modified optimized sequence.

This 028-opt+3pA sequence is analyzed in the maize (SEQ ID NO: 15) and tobacco and maize embryo transient testing systems (SEQ ID NO: 23) as described in examples 2 and 3. The level of RNA and protein expression obtained from the 028-opt+3pA sequence is then compared to that obtained from the 028+opt sequence.

b) Transient Expression System of the Axmi028 Gene in Maize Protoplasts:

Different Axmi028 versions were transformed into maize protoplasts by transient transfection. The 028-WT, 028-opt, 028-optpA and 028-opt3pA genes driven from the doubled version of constitutive viral 35S promoter (Guilley et al. 1982) are cloned into an pD35-derived vector (http://www.dna-cloning.com/) that adds a N-terminal His tag and a C-terminal Myc tag to each of the Axmi028 versions. It also contains the nptII neomycin phosphotransferase gene referring resistance to kanamycin expressed from the constitutive nos promoter (Depicker et al. 1982) forming the plasmids pD35-nH-cM-Axmi028-wt, pD35-nH-cM-Axmi028-opt, pD35-nH-cM-Axmi028-optpA, pD35-nH-cM-Axmi028-opt3pA (SEQ ID NO: 29, SEQ ID NO: 30, SEQ ID NO: 31, SEQ ID NO: 32, respectively). These four vectors plus a control vector expressing no Axmi028 but a reporter gene td-tomato (SEQ ID NO: 33) are transfected into corn protoplasts of corn line A188 according to the protocol of PEG-mediated transformation of plant protoplasts (Sheen, 2002). 48 h post transfection protoplasts were harvested by centrifugation, total protein was analyzed for the presence of Axmi028 versions by Western analysis by using a primary antibody anti-6X His IgG labeled with fluorescent dye CF680 (N-terminal His-tag detection via filter Alexa 680), a primary antibody anti-c-Myc-Cy3 labeled with fluorescent dye Cy3 (C-terminal Myc-tag detection via filter Alexa 546) and a nptII antibody with an secondary HRP-conjugated antibody for nptII detection as an internal control. The level of Axmi028 protein, normalised for nptII protein expression, is compared between the four Axmi028 constructs. FIG. 8 depicts the relative expression intensity of Axmi028 protein. Detection of the full-length Axmi028 versions is made via the N-terminal His tag (detection: fluorescence filter Alexa680) and the C-terminal Myc tag (detection: fluorescence filter Alexa546). Detection of nptII (detection: chemiluminescence) as internal transformation control was possible in all samples. Within every lane the intensity of the detected c-Myc and 6His signals was set into relation to the nptII signal as calculated via the software Image Lab 5.2.1, BioRad.

Conclusions on Axmi028 Transient Assays:

FIG. 6A shows that in the transient 028-Luc maize assay the expression of luciferase from the 028-WT-Luc sequence is 10% of that from the 028-opt-Luc sequence. The addition of 3 or all polyA motifs to the optimized sequence did not reduce expression of luciferase compared to that obtained from the 028-opt-Luc sequence.

FIG. 8 shows the protein expression of Axmi028-WT, 028-opt, 028-optpA and 028-opt3pA in corn protoplasts after transient transformation. Signal detection was only possible for the optimized versions but not for the WT. The addition of 3 polyadenylation motifs to the optimized sequence did not reduce expression compared to that obtained from the 028-opt sequence. In this assay, the addition of 3 polyadenylation motifs gives higher protein expression levels than the expression of 028-opt and 028-opt-pA. In contrast the expression of 028-opt-pA was lower than that of 028-opt.

Axmi100

The Ami100 wild-type but C-terminus truncated sequence was examined for the presence of weak polyadenylation motifs that can remain in the optimized sequence. Three AATCAA motifs were found, these are overrepresented motifs in the maize CDSv3 database. The optimized sequence already has an AAGCAT motif present at its wild-type position of 1788 bp. This motif is overrepresented in the maize CDS v3 dataset (152% of real/theoretical). The first two of the three AATCAA motifs were introduced in the optimized sequence (positions 13 bp and 1192 bp) giving three weak polyadenylation motifs identified in the optimized coding sequence at the wild-type position. This sequence 100opt+3pA as described in SEQ ID NO: 14 (FIG. 4D) finally comprises 6 weak polyadenylation motifs and does not comprises any of the strongest ATAAAA, CATAAA, ATACTA, ATTAAA, AATTAA, and ATTAAT polyadenylation motifs. It was analyzed by PASPA (FIG. 5D) and as for Axmi028+3pA, found to give an identical probability curve of polyadenylation as the fully optimized sequence (FIG. 5B).

This 100-opt+3pA sequence is analyzed in the maize (SEQ ID NO: 16) and tobacco and maize embryo transient testing systems (SEQ ID NO: 24) as described in examples 2 and 3. The level of RNA and protein expression obtained from the 100-opt+3pA sequence is then compared to that obtained from the 100-opt sequence.

Western blot analysis was made on the same samples by using a polyclonal antibody raised against Axmi100 protein. The result are in line with the result depicted in FIGS. 6A and 6B: the presence of polyadenylation signals makes it possible to obtain good expression of the protein, better than the expression of the optimized protein in which no polyadenylation have been added, and the expression of the optimized protein in which all wild-type polyadenylation have been added, whereas the wild-type protein is not properly expressed (data not shown).

Conclusion on Axmi100 Expression in Transient Assay:

FIG. 6B shows that in the transient Axmi100-Luc assay the expression of luciferase from the 100-WT-Luc sequence is 17% of that from the 100-opt-Luc sequence. The addition of 3 polyA motifs to the optimized sequence did not reduce expression of luciferase compared to that obtained from the 100-opt-Luc sequence. However the addition of all the polyA motifs reduced luciferase expression to 11% of the 100-opt-Luc sequence.

Western Blot, which looks at the protein quantity rather than at the activity, confirms the results shown in FIG. 6B.

Example 6

Stable Expression Testing of Genes that are Maize Codon-Optimized but Contain a Minimum of Three Polyadenylation Motifs

Axmi028

The strains described in example 2 and 5 (028-WT+GUS, 028-WT+Cyan, 028-opt+GUS, 028-opt+Cyan, 028-opt+pA+GUS, 028-opt+pA+Cyan, 028-opt+3pA+GUS and 028-opt+3pA+Cyan) are transformed into maize essentially as described by Ishida et al (1996). A minimum of 10 individual, single copy transformants with an intact T-DNA, are produced for each construct. QRT-PCR and Western analyses are performed on TO leaf material. Leaf Axim028 expression and protein levels of the 028-WT plants are compared to the 028-opt and 028-opt+3pA transformants,

The Western blot analyses on maize leaf protein extracts from plants transformed with Axmi028+GUS constructs (FIG. 7A) were performed with an antibody against the C-Myc TAG ([9E10]—Chip Grade ab32″ (abcam)). No Axmi028 protein could be detected in plants transformed with the 028-WT gene. Axmi028 protein expression could be observed in plants transformed with the 028-opt, 028-opt+3pA and 028-opt+pA sequences. This expression seemed highest in plants transformed with 028-opt+3pA.

Axmi100

As described above for Axmi028, the different versions of axmi100 are transformed into maize. Levels of Lepidopteran resistance in 100-opt and 100-opt+3pA transformed plants are compared to levels of 100-WT transformants in leaf feeding assays.

Western blot analysis on maize leaf protein extracts from plants transformed with Axmi100+GUS constructs (FIG. 7B) were performed with an Axmi100 polyclonal antibody. No Axmi100 protein could be detected in plants transformed with the 100-WT gene. Axmi100 protein expression could be observed in plants transformed with the 100-opt, 100-opt+3pA and 100-opt+pA sequences. Expression seemed least robust in 100-opt+pA transformants where only 5 out of 10 transformed plants expressed significant levels of Axmi100 protein. RT-QPCR analyses using 3′ and 5′ primer pairs (SEQ ID 25-28) also showed that transcript levels of 100-opt+pA in maize transformants were lower than those obtained in 100-opt and 100-opt+3pA transformants (FIGS. 9A and 9B).

It can be concluded from these results that gene optimization is necessary to obtain Axmi028 and Axmi100 protein expression. The addition of weak polyadenylation motifs does not impair protein expression. The presence of 3 or a few more weak polyadenylation motifs in the optimized sequences does not impair protein expression or can improve expression compared to the optimized gene sequence. However the re-introduction of all the polyA motifs into the optimized sequence can reduce the chance of obtaining a protein expression level equivalent to that obtained from the optimized gene.

REFERENCES

• Campbell and Gowri (1990). Codon usage in higher plants, green algae and cyanobacteries. Plant Physiol. 92:1-11.
• Colgan D F, Manley J L. (1997). Mechanism and regulation of mRNA polyadenylation. Genes Dev. 11(21):2755-66.
• Guilley H, Dudley R K, Jonard G, Balazs E, Richards K E: Transcription of Cauliflower mosaic virus DNA: detection of promoter sequences, and characterization of transcripts. Cell. 1982, 30: 763-773.
• Lu A, Diehn S and Cigan M (2015). Maize Protein Expression. In Recent Advancements in Gene Expression and Enabling Technologies in Crop Plants. Editors; Kasi Azhakanandam, Aron Silverstone, Henry Daniell and Michael R. Davey. Springer ISBN 978-1-4939-2201-7; ISBN 978-1-4939-2202-4 (eBook); DOI 10.1007/978-1-4939-2202-4.
• Graber J H, Cantor C R, Mohr S C, Smith T F. (1999) In silico detection of control signals: mRNA 3′-end-processing sequences in diverse species. Proc Natl Acad Sci USA. 96:14055-60.
• Ishida Y, Saito H, Ohta S, Hiei Y, Komari T, Kumashiro T. (1996) High efficiency transformation of maize (Zea mays L.) mediated by Agrobacterium tumefaciens. Nature Biotechnol. 14, 745-50.
• Jefferson R A, Kavanagh, T A and Bevan, M W (1987). GUS fusions: beta-glucuronidase as a sensitive and versatile gene fusion marker in higher plants. EMBO J. 6: 3901-3907.
• Ji G, Li L, Li Q Q, Wu X, Fu J, Chen G, Wu X. (2015) PASPA: a web server for mRNA poly(A) site predictions in plants and algae. Bioinformatics 31:1671-3.
• Joshi C P. (1987) Putative polyadenylation signals in nuclear genes of higher plants: a compilation and analysis. Nucleic Acids Res. 15(23):9627-40.
• Komari T, Hiei Y, Saito Y, Murai N, Kumashiro T. (1996). Vectors carrying two separate T-DNAs for co-transformation of higher plants mediated by Agrobacterium tumefaciens and segregation of transformants free from selection markers. Plant J. 10:165-74.
• Leckie B M, Neal Stewart C Jr. (2011). Agroinfiltration as a technique for rapid assays for evaluating candidate insect resistance transgenes in plants. Plant Cell Rep. 30(3):325-34.
• Mogen B D, MacDonald M H, Graybosch R, Hunt A G. (1990) Upstream sequences other than AAUAAA are required for efficient messenger RNA 3′-end formation in plants. Plant Cell. 2(12):1261-72.

Murray E E et al. (1989) Codon usage in plant genes. Nucleic Acids Res. 17:477-498

• Sanfacon H, Brodmann P, Hohn T. (1991) A dissection of the cauliflower mosaic virus polyadenylation signal. Genes Dev. 5(1):141-9.
• Sheen, J. 2002, A transient expression assay using Arabidopsis mesophyll protoplasts. http://genetics.mgh.harvard.edu/sheenweb/
• Tzanis G et al (2011). PolyA-iEP: A data mining method for the effective prediction of polyadenylation sites. Expert Syst. Appl. 38(10) 12398-12408.
• Verdaguer B, de Kochko A, Beachy R N, Fauquet C. (1996). Isolation and expression in transgenic tobacco and rice plants, of the cassava vein mosaic virus (CVMV) promoter. Plant Mol Biol. 31:1129-39
• Wu X et al (2012). Comprehensive recognition of messenger RNA polyadenylation pattern in plants. African journal of biotechnology, vol 11(14), pp 3215-3234.

Claims

1. A method of making an expression cassette comprising a nucleic acid molecule having a coding sequence encoding a non-plant protein for transforming a plant, comprising the steps of: a) identifying a coding sequence that encodes a non-plant protein;b) identifying each polyadenylation motif sequence from among AAAATA, AACCAA, AAGCAT, AATAAA, AATAAT, AATACA, AATCAA, AATTAA, ATAAAA, ATACAT, ATACTA, ATATAA, ATGAAA, ATTAAA, ATTAAT and CATAAA, and its nucleic acid position in said coding sequence;c) optimizing said coding sequence by codon substitution, including by recoding the nucleic acid molecule such that the codon usage reflects that used in the plant, wherein the optimized coding sequence encodes for said non-plant protein, wherein said optimization step provides an optimized coding sequence in which one or more polyadenylation motif sequences have been removed, and wherein said optimization step does not result in adding to said coding sequence any polyadenylation motif from among the AATTAA, ATACTA, ATATAA, ATTAAA, ATTAAT and CATAAA motifs;d) modifying said optimized coding sequence to obtain a modified gene sequence by introducing in the optimized gene sequence at least one polyadenylation motif sequence selected from among AAAATA, AACCAA, AAGCAT, AATAAA, AATAAT, AATACA, AATCAA, ATAAAA, ATACAT and ATGAAA, so as to obtain a modified coding sequence that comprises at least three but not all polyadenylation motifs identified in step b), wherein said modifying step does not result in adding to said coding sequence any polyadenylation motif from among the AATTAA, ATACTA, ATATAA, ATTAAA, ATTAAT and CATAAA motifs, resulting in said modified coding sequence comprising more polyadenylation motifs than the optimized sequence, and said modified coding sequence is encoding said non-plant protein;e) synthesizing a nucleic acid molecule having the modified coding sequence as obtained in step d; andf) operably linking a promoter and a terminator to said nucleic acid molecule to obtain an expression cassette for expression in plant.

2. The method of claim 1 wherein each polyadenylation motif introduced within said optimized sequence in step d) is a wild-type polyadenylation motif identified in step b), and is introduced at a nucleic acid position corresponding to its position within the coding sequence as identified in step b).

3. The method of claim 1 wherein the polyadenylation motif sequence introduced in step d) is chosen amongst motifs ATGAAA, AAGCAT, AACCAA, AATCAA or AAAATA.

4. The method of claim 2 wherein the polyadenylation motif sequence introduced in step d) is chosen amongst motifs ATGAAA, AAGCAT, AACCAA, AATCAA or AAAATA.

5. The method of claim 1 wherein the modified coding sequence comprises three to ten polyadenylation motifs.

6. The method of claim 2 wherein the modified coding sequence comprises three to ten polyadenylation motifs.

7. The method of claim 1, wherein said coding sequence is encoding a prokaryotic protein.

8. The method of claim 1, wherein said coding sequence is encoding an insecticidal Bacillus thuringiensis protein.

9. The method of claim 1, further comprising the steps of transforming a plant cell with the expression cassette and regenerating a plant from the transformed plant cell.

10. The method of claim 2, further comprising the steps of transforming a plant cell with the expression cassette and regenerating a plant from the transformed plant cell.